THE J O U R N A L OF BIOLOGICAL CHEMISTRY Vol. 258, No. 22, Issue of Novemher 25, pp. 1:1597-13605,1983 Printed LR U.S.A

Time-dependent Decreases in Binding Affinity of Agonists for @- Adrenergic Receptors of Intact S49 Lymphoma Cells A MECHANISM OF DESENSITIZATION*

(Received for publication, July 15, 1983)

Paul A. Insel$, Lawrence C. Mahan, Harvey J. Motulsky, Lloyd M. Stoolman$, and Arlene M. Koachman From the Division of Pharmacology, “013 H, Department of Medicine, University of California, San Diego, La Jolla, California 92093

Studies of @-adrenergic receptors on several types of intact target cells indicate that agonists, but not antag- onists, show prominent KDIK,,~ discrepancies-lower affinities (KD) in equilibrium binding studies (Le. in competition with antagonist radioligands) than in functional assays (Kact). In this report we show that intact S49 lymphoma cells initially bind @-adrenergic agonists in a high affinity manner but that this high affinity binding rapidly converts to a low affinity state. This time course is similar to that of agonist-mediated desensitization in 549 cells. In competitive binding studies conducted with [1251]iodocyan~pindolol or [1251] iodohydroxybenzylpindolol, IC50 values for 8-adrener- gic agonists increased 13-200-fold between 1 and 60 min, whereas antagonists yielded similar ICsO values at the two time points. The binding of antagonists, but not agonists, could be simulated by a computer model based on the law of mass action. In contrast with results in intact S49 cells, crude membrane fractions yielded similar ICSo values for agonists in 1- and 60-min in- cubations with [1Z51]iodocyanopindolol. Moreover, time-dependent decreases in apparent affinity of the agonist (-)-isoproterenol were observed not only in wild type S49 cells but also in several S49 variants (UNC, cyc-, H21a) having defects in N,, the guanine nucleotide binding protein that couples receptors to activation of adenylate cyclase, and in Kin-, an S49 variant with absent CAMP-dependent protein kinase activity. These results show that &adrenergic recep- tors of intact 549 cells demonstrate a prominent time- dependent decrease in apparent affinity for agonists and that this decrease in affinity does not require cAMP generation, the N, components defective in sev- eral S49 variants or CAMP-dependent protein kinase. The rapid conversion of agonist binding from high to low affinity can account for the rapid desensitization of intact cells to catecholamines and probably explains previously reported KD/KaCt discrepancies of intact cells.

@-Adrenergic receptors are probably the best studied of the cell surface receptors coupled to adenylate cyclase. Consider-

* This work was supported in part by grants from the National Science Foundation, National Institutes of Health, and the American Cancer Society. 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.

To whom correspondence should be directed. I Present address, Department of Laboratory Medicine, University

of California, San Francisco, San Francisco, CA 94143.

able recent progress has been made in defining the molecular components of these receptors and mechanisms by which receptor occupancy by agonists triggers the activation of adenylate cyclase (for reviews see Refs. 1-5). Many previous studies have emphasized properties of receptors in solubilized or native membranes and the interaction of receptors with NS.’ Experimental findings in several systems have indicated, for example, that GTP bound to N can allosterically decrease the affinity of agonists in binding to receptors and that GTP is a critical cofactor in the hormonal stimulation of adenylate cyclase (1-6). Such conclusions derive from work undertaken in numerous cell types, but many of the recent, key results on the P-adrenergic receptor/adenylate cyclase system have been those obtained using membranes derived from murine S49 lymphoma cells. S49 cells have been especially useful in these studies because of the availability of numerous clonal variants having lesions in the pathway of cAMP generation, particu- larly those S49 variants having lesions in N,” (7-10).

An alternative approach to studies of receptors in target cell membranes is to explore the mechanism of receptor dynamics in intact target cells. Studies in intact cells permit receptors to be probed in their native membrane environment where they are subject to regulation by cellular events, organ- elles, and constituents. Previous work on P-adrenergic recep- tors in this physiological milieu has suggested that these receptors can be readily identified using radioligand binding techniques (for review see Ref. 11). Although certain proper- ties of receptors are quite similar in membranes and intact cells, a vexing problem observed for P-adrenergic agonist interaction with intact cells has been the striking discrepancy between KO and K,,,. Typically, the KIj of an agonist deter- mined in competitive binding assays with radiolabeled antag- onists is about 2 orders of magnitude greater than its K,,, (I 1- 14), i.e. agonists appear to have a much lower affinity in the binding studies. In S49 lymphoma cells, for example, we have observed a prominent Ku/Ksct discrepancy for agonists inter-

’ The abbreviations used are: N., the guanine nucleotide binding component that mediates stimulation of adenylate cyclase; Hepes, N- Z-hydroxyethyIpiperazine-N’-2-ethanesulfonic acid; KD, the equilib- rium dissociation constant obtained in binding studies; K,,,, the half- maximally effective concentration of an agonist in cAMP accumula- tion studies; ICw, the concentration of drug producing 50% inhibition of specific binding; K,, the half-maximally effective concentration of an antagonist in cAMP accumulation studies; (’251]ICYP, [12sI]iodo- cyanopindolol; [”SI]IHYP, [‘2sI]iodohydroxybenzylpindolol.

N, is used here as a generic term to refer to the guanine nucleotide binding component that mediates stimulation of adenylate cyclase (also called G/F, G, or N). Although purified N. from liver and turkey erythrocytes appears to contain multiple subunits (%lo), the number of subunits, their stoichiometry, and the definitive, functional role of each in S49 cells is, as yet, ill-defined, as discussed in the text.

13597

by guest on July 24, 2020http://w

ww

.jbc.org/D

ownloaded from

13598 Time-dependent P-Adrenergic Binding to S49 Cells

acting with &adrenergic receptors on intact cells (12, 15). The existence of spare receptors on intact cells has been suggested in some cell systems to explain Ku/KaCt discrepan- cies (13). We have previously shown, however, that in the S49 cell spare receptors do not account for this phenomenon (12). Even though agonists yield high Ku values in equilibrium binding studies with S49 cells, we have found in preliminary studies that agonists may interact transiently with high affin- ity binding sites in these cells (16). We have thus examined this possibility in more detail by comparing competitive bind- ing studies with P-adrenergic agonists and antagonists at early time points (nonequilibrium binding conditions) with com- petitive binding studies at times closer to steady state of radioligand binding. We find that when intact wild type S49 cells and several S49 variants having lesions in N or with absent CAMP-dependent protein kinase activity are exposed to agonists, the receptors show a prominent and rapid de- crease in binding affinity for agonists. This decrease in bind- ing affinity closely parallels the kinetics of agonist-mediated desensitization of intact S49 cells. The results provide evi- dence for the dynamic regulation of P-adrenergic receptor affinity for agonists as an early event that does not require CAMP generation, CAMP-dependent protein kinase, or N, component(s) defective or absent in S49 variants.

EXPERIMENTAL PROCEDURES

Cell Lines and Cell Culture-Wild type (24.3.2) S49 lymphoma cells and three S49 variants lacking hormonal responsiveness but retaining 8-adrenergic receptors were used UNC, uncoupled cells which possess N units that are structurally and functionally altered (17, 18); cyc- cells (94.15.1), cells that appear to lack N activity, as defined func- tionally or immunologically (1, 19); H21a, a variant whose 8-adrener- gic receptors retain regulation by guanine nucleotides but whose N is unable to restore functional activity to cyc- or UNC in complemen- tation assays with isolated membranes or in somatic cell hybrids (20). In addition, we also used an S49 variant with absent CAMP-dependent protein kinase activity (clone number 24.4.6, Kin-, Ref. 21). All cells were grown at 37 "C in T flasks in Dulbecco's modified Eagle's medium containing 10% heat-inactivated horse serum (growth me- dium) in a 10% CO,:90% air environment. Cells were used for exper- iments only when growing logarithmically (51 X lofi cells/ml) and when viability (assessed by trypan blue exclusion) was >go%.

Cellular cAMP Accumulation-Cells in growth medium were cen- trifuged a t 200 X g for 5 min at room temperature and resuspended in fresh growth medium at a density of -1 X lo6 cells/ml. Unless otherwise indicated, aliquots of cells were then incubated a t 37 "C for 15 min in the presence of 100 p~ Ro 20-1724 (a potent phosphodi- esterase inhibitor for these cells), 1 mM ascorbic acid (to block oxidation of catecholamines), and various other drugs. Two basic protocols were used. For agonist activity, we incubated cells with the adrenergic agents and estimated Keet values graphically. For assess- ment of antagonist activity, we incubated adrenergic agents together with 100 nM (-)-isoproterenol. Incubations were terminated by pi- petting 1 ml of the cell suspensions into a 1.5-ml plastic microfuge tube and pelleting the cells by centrifugation in a Beckman microfuge for 20 s. The medium was aspirated, 50 mM sodium acetate buffer (pH 4.0) containing 0.2 mM isobutylmethylxallthine was added, sam- ples were immediately boiled for 3 min, and removed to an ice bath. Aliquots of the boiled extract were then assayed for cAMP using a competitive binding protein method as previously described (22). Kt values for each antagonist were corrected for (-)-isoproterenol con- centration.

b-Adrenergic Receptor Binding Assay-Cells in growth medium were centrifuged at 200 X g for 5 min at room temperature and resuspended a t a density of 1.0-2.0 X lo6 cells/ml in Dulbecco's modified Eagle's medium (pH 7.4) containing 20 mM Hepes (in lieu of NaHCOJ and 1 mg/ml of bovine serum albumin (binding medium). The cells were then incubated for 5 rnin a t 37 "C and reactions were

warmed (37 "C) binding medum that contained ['2511]ICYP or [12sI] initiated by adding aliquots of cells (generally 4-8 X lo5 cells) to

IHYP and varying concentrations of competing agents in a total volume of 0.5 ml. Catecholamines were protected from oxidation by using either 1 mM ascorbic acid or 10 pg/ml each of superoxide

dismutase and catalase (23). In general samples were placed in a shaking water bath as shaking improved the consistency between replicates (data not shown). Reactions were terminated after 1 min or 60 min by addition of 10 ml of 37 "C buffer (5 mM potassium phosphate, p H 7.0, 1 mM MgS04) followed by rapid filtration and washing over glass fiber filters (Gilman A/E filters for [12sII]IHYP and Whatman GF/C filters for ['z51]ICYP). Reactions stopped in this manner showed stable values of specifically bound radioactivity over the short time (51 min) that preceded filtration over GF/C filters and washing with an additional 10 ml of buffer. The dried filters were counted in a gamma counter (82% efficiency) or in a liquid scintilla- tion counter (62% efficiency). Nonspecific binding was determined by the inclusion of radioligand plus 1 ~ L M (-)-propranolol or 1 mM (-)-isoproterenol; these two methods yielded similar values for non- specific binding. Specific binding of both radioligands was linear with cell number. The radioligands were iodinated as described previously (12,22); both radioligands remained undegraded during the course of the binding experiments. For computer modeling of competitive bind- ing of subtype selective antagonists, binding data was analyzed using the LIGAND program of Munson and Rodbard (24). Receptor con- centrations were maintained 5 3 PM, and thus binding of the radioli- gands would be expected to represent almost exclusively interaction of the (-)enantiomer with the receptor (25,26).

Computer Modeling of Radioligand Binding Data-A computer model was used to compare data observed in our kinetic studies of competitive binding with results predicted by the law of mass action. This model assumed that both radioligand and competitor bind reversibly to the receptors with specified association and dissociation rate constants. We solved the resulting differential equations to determine the amount of radioligand binding as a function of time. The simulations were based on parameters matching those of the radioligand binding experiments: 1) we used the known association and dissociation rate constants for ICYP binding to the P-adrenergic receptors of S49 cells (Fig. 1); 2) the unlabeled drug had a known equilibrium KD; and 3) the ligand and receptor concentrations were those of the experiment. This receptor concentration was low enough so that over 95% of the radioligand and inhibitor were unbound a t all times. Details of this computer modeling will be presented else- where.3 Both experimental and simulated data were expressed as specific binding in the presence of competing drug relative to specific binding of the radioligand alone.

Another computer program was used to numerically simulate a model incorporating an agonist-induced state change. At intervals of 5 x min, the program calculated the extent of each binding reaction and recalculated the concentration of each molecular species.

Materials-Horse serum and Dulbecco's modified Eagle's medium were obtained from Gibco. (3H]cAMP was from New England NU- clear. Superoxide dismutase and catalase were purchased from Sigma. Several compounds were gifts as follows: (-)- and (+)-isoproterenol and (-)-epinephrine, Sterling-Winthrop; (-)- and (+)-propranolol, Ayerst; Ro 20-1724, Hoffman-LaRoche; cyanopindolol and hydrox- ylbenzylpindolol, Sandoz; butoxamine, Ayerst; terbutaline, Astra; IC1 118,551 andpractolol, ICI; sotalol, Regis; fenoterol, Dr. K. Minneman, Emory University; hydroxybenzylisoproterenol, Dr. R. J. Lefkowitz, Duke University; and betaxolol, Synthelabo.

RESULTS

Validation of f'251]ICYP As a Probe for /3-Adrenergic Recep- tors in Intact S49 Cells-In previous studies we have shown that ['251]IHYP and [3H]dihydroalprenolo1 can be used to identify @-adrenergic receptors on intact S49 cells (12). In the current studies, we initially performed several experiments to show that ['"I]ICYP could also be used as a probe of these receptors on S49 cells. These data are shown in the Mini- print.4 We found that ['251]ICYP bound with appropriate

Motulsky, H. J., and Mahan, L. C. (1984) Mol. Pharrnacol., in press.

Portions of the data (Figs. 1,2,4, and 6) are presented in miniprint at the end of this paper. Miniprint can be easily read with the aid of a standard magnifying glass. Full size photocopies are available from the Journal of Biological Chemistry, 9650 Rockville Pike, Bethesda, MD 20814. Request Document No. 83M-486, cite the authors, and include a check or money order for $1.60 per set of photocopies. Full size photocopies are also included in the microfilm edition of the Journal that is available from Waverly Press.

by guest on July 24, 2020http://w

ww

.jbc.org/D

ownloaded from

Time-dependent &Adrenergic Binding to S49 Cells 13599

kinetics (Fig. I), reversibility (Fig. I), stereoselectivity (data not shown), and saturability (Fig. 2) to a single class (Hill coefficient = 0.98 k 0.02) of sites on intact wild type S49 cells. Specific binding achieved equilibrium by 60-90 min. Routinely, for convenience, most subsequent experiments with ['2sI]ICYP were conducted for 60 min. Compared to [1251] IHYP (12), ['"I]ICYP yielded similar values of Kl, (18 f 2 PM) and B,,, (1220 f 67 sites/cell, n = 9, mean f SE) but much lower nonspecific binding. In equilibrium binding stud- ies, agonists competed with ['251]ICYP with a typical p2- adrenergic rank order (isoproterenol > epinephrine >> nor- epinephrine); &- and &selective agents (zinterol, IC1 118,551, and practolol) competed for ['2sII]ICYP sites in a manner compatible with a single class of &-adrenergic recep- tors on these cells (data not shown).

Competition of Agonists and Antagonists for Radioligand Binding Sites in I - m i n and 60-min Incubations with Wild Type 5'49 Cells-The ICso of the P-antagonist (-)-propranolol in competing for ["'I]ICP or ['"I]IHYP sites on S49 cells was similar after either a 1-min or 60-min incubation (Fig. 3). In experiments conducted for an additional 3 h (Fig. 4, top), the apparent ICso value for (-)-propranolol increased slightly (Sbfold), probably reflecting differences in approach to equi- librium of radioligand and competitive drug (27). Similar results were obtained using competition of pindolol for [12sI]

It I

751 IHYP IHYP

501 0 -12 -It -10 -9 -8 -7 -6 -5

log [I ( - 1- PropranololIl, M FIG. 3. Competition of (-)-propranolol for ['261]ICYP and

['261]IHYP binding sites on intact wild type ,949 cells: 1-min uersus 60-min incubations. Wild type S49 cells were incubated with varying concentrations of (-)-propranolol and 83 p~ ['251]ICYP (top) and 130 pM ['251]IHYP (bottom) for 1 min (A-A) or 60 min (U). Binding reactions were terminated as described under "Experimental Procedures." The data shown are similar to those obtained in two experiments, each run in triplicate, with ['*'I]ICYP and ['251]IHYP.

IHYP binding sites for several hours (data not shown). This similarity in ICso for antagonists tested at 1 min and 60 min ([IC,,, 60 min]/[IC,,, 1 min] I 4) was noted for antagonists whose apparent affinity in competing for ['2sII]ICYP sites varied over 4 orders of magnitude (Table I).

By contrast, @-adrenergic agonists competed for ['"I]ICYP or ["'I]IHYP sites with much lower ICso values in 1-min compared to 60-min incubations (Fig. 5 , Table I). Agonists were 13-200-fold more potent in the 1-min than in the 60- min incubations. Conducting incubations for up to 4 h yielded no further increase of the ICso values for agonists (Fig. 4, bottom). To determine whether this increase was character- istic of agonists regardless of equilibrium Kn, we performed identical experiments with several agonists and all yielded similar results (Table I).

We have previously shown that cells remain viable and drugs remain stable during the incubation periods required in these experiments (12). Because in many of the current stud- ies we used a new method (including 10 pg/ml each of super- oxide dismutase or catalase) to prevent oxidation of catechol- amines (23), we confirmed the stability of the drugs by incu- bating cells with isoproterenol for 60 min and using the media from those cells in an incubation with fresh cells for 1 min. This preincubated isoproterenol was as effective as fresh isoproterenol in binding studies with these cells (data not shown).

To assess whether intact cells are required for the time- dependent decrease in ability of agonists to compete for ['zsI] ICYP binding sites, we prepared a washed membrane fraction. Competitive binding experiments conducted with these mem- branes, ['"I]lICYP, and varying concentrations of isoproter- enol indicated little (<3-fold) increase in the ICsO value for

TABLE I Binding and functional activity of @-adrenergic agonists and

antagonists in intact S49 lymphoma cells Intact wild type S49 cells were used in studies of cAMP accumu-

lation and radioligand binding as described under "Experimental Procedures." ['2'I]ICYP (50-100 pM) was used in all radioligand binding studies in this table. All experiments were conducted at least twice and yielded similar results on each occasion. Binding data are represented as mean IC50 values, where IC,, represented the concen- tration inhibiting specific binding by 50% after either 1-min or 60- min incubations. K,,, and KI values, for agonists and antagonists, respectively, are mean values obtained in at least 2 cAMP accumu- lation studies for each compound. KD, 60 min values are ICso, 60 min values corrected for the concentration of ['251]ICYP in each experi- ment bv the method of Chenz and Prusoff (41).

Compound

Agonists (-)-Isoproterenc (-)-Epinephrinc (-)-Norepineph Terbutaline Fenoterol Hydroxybenzyl-

isoproterenol Antagonists

Betaxolol (-)-Propranolol

Zinterol Sotalol Butoxamine Atenolol Practolol

1 -

Radioligand binding

~~ ~~~,

cAMP

L 1 min 60 min

0.096 2.5

12.5 1.5 2.6 0.008

.001f 0.26 0.27 2.86 0.78 3.57

15.3

P

1

M

9.2 283

2555 130 32.5 0.9

,0009 0.29 0.21 1.02 3.11

14.2 20.6

i t

2.0 57.8

13.8

564 49.5

0.3

96 115 204 87 13

112

0.012 0.025 1.8 0.57 0.016 .0015

by guest on July 24, 2020http://w

ww

.jbc.org/D

ownloaded from

13600 Time-dependent @-Adrenergic Binding to S49 Cells

I

1 I I I I 1 I

0 -10 -9 -8 -7 -6 -5 -4 -3 log E( - )- IsoproterenolI, M

FIG. 5. Competition of (-)-isoproterenol for [12sI]ICYP and ['261]IHYP binding sites on intact wild type S49 cells: 1-min versus 60-min incubations. Wild type S49 cells were incubated with varying concentrations of (-)-isoproterenol and 56-89 PM [1251] ICYP (top) and 100 pM ('"I]IHYP (bottom) for 1 min (A-A) or 60 min (c-".). Binding reactions were terminated as described under "Experimental Procedures." The data shown are the mean ? SE ( n = 4) for ['251]ICYP and similar to those obtained in two experiments with ['251]IHYP. Using LIGAND, data at 60 min were described by a two-site model (p < 0.001 two-site versus one-site) with 20% of the receptors having a K, of 3 nM and 80% with a Kr of 4 pM.

(-)-isoproterenol between 1 and 60 min (Fig. 6). Thus, intact cells appear to be necessary to demonstrate transient high affinity agonist binding.

Kinetics of Desensitization of C A M P Generation and of Con- version of @-Adrenergic Receptor Binding of Agonists from High to Low Affinity-In order to define more precisely the time course over which receptors convert from high to low apparent affinity, we conducted kinetic experiments in which we used a single concentration of agonist or antagonist in competing for ['251]ICYP binding sites on intact S49 cells. Data from such experiments, which represent the differing rates of binding of [12sI]ICYP alone or in the presence of competing drug, were expressed as percentage of specific binding observed at each time point. Specific radioligand binding in the presence of a competitor under nonequilibrium conditions is dependent on both the association and dissocia- tion rate constants for each compound. These constants are rarely known for unlabeled compounds, and multiple kinetic constants may define any particular Kl). Therefore one cannot easily intuit the patterns of radioligand binding in the pres- ence of competitor predicted by the law of mass action, and a computer model was written to mathematically determine this binding. As shown in Fig. 7, 1 nM (-)-propranolol com-

(-)ISOPROTERENOL

, ............. A

\i (-)PROPRANOLOL

+% i , c ,..... .......

0 1 2 3 4 60 70 TIME (min)

FIG. 7. Kinetic analysis of agonist and antagonist competi- tion for ['2sI]ICYP binding to intact wild type S49 cells. The experimental data were obtained as follows. 549 cells were incubated with (lzsI]ICYP alone or with ['251]ICYP and either (-)-isoproterenol (top) or (-)-propranolol (bottom). Specific ['z51]ICYP binding was determined a t each time point and specific binding in the presence of competitor is displayed as a percentage of specific binding in its absence. The computer-derived curves were generated from equations describing the competition between radioligand and competitor fol- lowing the law of mass action for the experimental conditions detailed as follows. Top: the conditions used to generate experimental data (W) and computer-derived curves for (-)-isoproterenol were as follows: ["51]IcYP = 45 pM, [receptor] = 2.4 pM, and [(-)-isoproter- enol] = 1 p ~ . Experimental data represent the mean k S.E. of three to five binding assays with triplicate determinations as described under "Experimental Procedures." To generate the family of curves shown, the association and dissociation rate constants were varied in half order-of-magnitude increments (top to bottom) from 2.0 X 10' to 2.0 X IO7 M" min" and 4.0 X 10" to 4.0 X 10' rnin", respectively, to maintain the experimentally derived equilibrium KD for (-)-iso- proterenol of 2.0 p~ (Table I). Clearly none of the curves describing simple competition fit the data. Bottom: the conditions used to generate experimental data (U) and computer-derived curves for (-)-propranolol were as follows: ['2'I-ICYP] = 43 PM, [receptor] = 2.0 p ~ , and [(-)-propranolol] = 1 nM. Experimental data represent the mean f SE of four binding assays with triplicate determinations as described under "Experimental Procedures." The data was fit to the equation describing the law of mass action using Marquardt's nonlinear least squares regression analysis and the resulting curve is shown. The best fit values for the association- and dissociation-rate constants of propranolol were 1.4 X lo9 M" min" and 0.7 rnin", respectively.

by guest on July 24, 2020http://w

ww

.jbc.org/D

ownloaded from

Time-dependent P-Adrenergic Binding to S49 Cells 13601

100 1 .- c E u-) 4

I- 6

60

40 6

0 5 10 15 TIME (min)

FIG. 8. Kinetics of densensitization of cAMP generation in response to (-)-isoproterenol of intact 549 cells. Triplicate 0.5- ml aliquots of wild type S49 cells (-2 X lo6 cells/ml) were incubated with 100 @M Ro 20-1724 for 10 min and then with 1 +M (-1- isoproterenol for the indicated times. Reactions were terminated by adding ice cold perchloric acid to a final concentration of 0.4 N. Samples were centrifuged in the cold for 10 min at 1520 X g. Super- natant fractions were then collected and neutralized with 1 N KHC03. After 10 min on ice, samples were centrifuged to remove the KClO, precipitate. Supernatant fractions were lyophilized, resuspended in 500 mM sodium acetate (pH 4.0) containing 2 mM isobutylmethylxan- thine, and then assayed for cAMP as described under "Experimental Procedures." Data are expressed as the percentage of maximal cAMP accumulation at 15 min and (inset) d(% CAMP)/&, the percentage of cAMP generated between each time point divided by the time interval. Data shown are the mean + SE (n = 5 ) . The SE for the last three data points are included within the symbol.

peted for ['"I]ICYP sites in a fashion that is predicted by our computer model for kinetic studies in which an unlabeled compound and radioligand compete for binding sites accord- ing to the law of mass action. In sharp contrast, the kinetics by which 1 p~ (-)-isoproterenol competes for ['251]ICYP binding sites did not match the model predictions for simple competitive interactions described by the law of mass action. The attempted simulation involved testing a wide range of values for association and dissociation rate constants for isoproterenol (consistent with its equilibrium KIJ in an at- tempt to fit observed data (Fig. 7) . No computer simulation based on actual experimental conditions and simple compet- itive interaction generated the transient high affinity binding that was observed. The data in Fig. 7 suggest that isoproter- enol initially, but transiently, binds to the receptor with high affinity.

S49 cells demonstrate an agonist-specific desensitization (28-30). To assess whether the functional response of 8- adrenergic receptors in intact S49 cells (as measured by intracellular cAMP accumulation) desensitizes over a time course similar to that of the conversion of receptor affinity for agonist, we measured the rate of desensitization under conditions identical with those used in the binding assays. As shown in the inset to Fig. 8, the isoproterenol-stimulated rate of cAMP accumulation decreased exponentially with a T,,2 of 0.65 min (r ' = 0.92).

Both the binding data (Fig. 7) and the functional data (Fig. 8) are consistent with the idea that the receptor undergoes an agonist-promoted state change resulting in decreased agonist affinity and diminished response. This idea can be expressed as follows:

A + R + A R ~ A R * + A + R * KH k KL

The following symbols are used A = agonist, AR = agonist/ receptor complex, AR* = agonist/receptor complex after the state change, R* = receptor alone after the state change, KH = the equilibrium dissociation constant of the high afffinity agonist binding, Kr, = the equilibrium dissociation constant of the low affinity agonist binding, and kt = the unimolecular rate of receptor state transition. The radioligand (an antago- nist) binds indiscriminately to both states of the receptor.

Using a computer program to solve these equations numer- ically, and providing the program with the experimentally derived rate of desensitization, we tested whether this model could predict the binding data shown in Fig. 7 for isoproter- enol. The rate of transition, kt, was set to the 95% confidence limits of the functional rate of desensitization obtained in Fig. 8 (0.78-1.33 min-'). We set the values for the remaining rate constants as indicated in the legend to Fig. 9. In contrast to the simulations based on simple mass action shown in Fig. 7, our experimental binding data could be described by the "state change" model so long as the KH was set less than 100 nM (Fig. 9). Thus the decrease in agonist binding affinity may account for desensitization in the intact cell. Moreover, an agonist-promoted state change would predict that preincuba- tion with isoproterenol will shift receptors to the low affinity form, thus increasing subsequent radioligand binding. The data in Fig. 10 show that high affinity competition of [12'1] ICYP binding by isoproterenol is lost following a 30-min isoproterenol preincubation. In contrast, preincubation of cells with propranolol decreases the rate of [ '251]ICYP binding.

80 - Ly

l o, 2 40 L O

V 1

0 1 2 3 4 5 TIME (min)

FIG. 9. Predictions of a simple "state change" model. Data points from the experiment in Fig. 7 are shown along with the predictions of a computer model incorporating a receptor state change. The model, fully described in the text, assumes that isopro- terenol binds with high affinity to the receptors, that the receptors then undergo a state change, and that after the transition the recep- tors reversibly bind isoproterenol with its equilibrium KT. The asso- ciation rate constants of isoproterenol for both forms of the receptor were set equal to the association rate of ['251]ICYP (Fig. 1). The dissociation rate constatns of isoproterenol were calculated to yield the equilibrium dissociation constant, 4 @M (Fig. 5). The concentra- tions of radioligand and agonist were set equal to those of the experiments, and the kinetic constants for the radioligand were those determined in Fig. 1. The receptor transition rate was set equal to the 95% confidence limits of the exponential decay constant calcu- lated from the d(% cAMP)/dT curve in Fig. 8: 0.78-1.33 min". The simulated curves shown are not substantially altered as long as the initial binding step has a KD less than 100 nM.

by guest on July 24, 2020http://w

ww

.jbc.org/D

ownloaded from

13602 Time-dependent P-Adrene

This result is as expected for a slowly equilibrating radioligand that competes for binding sit,es occupied by a slowly disso- ciating compound.

Competition of &Adrenergic Agonists for (””IJICYP Bind- ing Sites in I-min and 60-min Incubations with S49 Var- iants-To determine whether the differences in binding of agonists observed with wild type cells required N, compo- nent(s) necessary for activation of adenylate cyclase in S49 cells, we tested competition of (-)-isoproterenol for receptors in several S49 variants having lesions in N, (Fig. 11). In all three variants tested (cyc-, UNC, and H21a), isoproterenol showed prominent time-dependent decreases in binding affin- ity between studies conducted for 1 min and 60 min. ICso values for (-)-isoproterenol in the variants were similar to values obtained with wild type cells. Mean values for the ratio of the IC5o of (-)-isoproterenol determined at 60 and 1 min were 84,28, and 82 in cyc-, UNC, and H21a cells, respectively. In cyc- cells, the mean values for this ratio with two other agonists, (-)-epinephrine and hydroxybenzylisoproterenol,

’‘’1 (-IISOPROTERENOL I / I I /

(-)PROPRANOLOL

L CONTROL

PRENCUBATION

~

0 1 2 3 9 0 1 2 3 4 TIME (mlnl TlME (mm)

FIG. 10. Kinetic analysis of isoproterenol and propranolol competition for [1261]ICYP binding to intact wild type S49 cells. Effect of prior incubation of cells with competitors. Cells were incubated with or without 1 pM (-)-isoproterenol or 1 nM (-)- propranolol for 30 min a t 37 “C and [12,51]ICYP was then added. A portion of cells which had been incubated without drugs was also incubated with isoproterenol or propranolol together with [1251]ICYP. Data are shown as in Fig. 7 and are mean values from two experiments performed in triplicate.

l o o k “

log E(-) -1soproterenol7, M FIG. 11. Competition of (-)-isoproterenol for [‘261]ICYP and

[1261]IHYP binding sites of intact cyc-, HZla, UNC, and Kin- S49 cells: 1-min uersus 60-min incubations. Cells were incu- bated with ICYP (50-80 PM) and varying concentrations of (-1- isoproterenol for 1 min (A-A) or 60 min (.“3). Reactions were terminated as described under “Experimental Procedures.” The data shown are similar t o those obtained in three experiments, each performed in triplicate, with UNC and cyc- cells and two experiments with H21a cells and Kin- cells.

‘rgic Binding to S49 Cells

were 101 and 24, respectively. Data similar to that shown in Fig. 11 were also obtained with cyc- cells using [1251]IHYP (data not shown). Thus, the N component(s) abent or defec- tive in these variants and agonist-mediated activation of adenylate cyclase is not necessary to produce the time-de- pendent decrease in agonist affinity in wild type S49 cells.

We also tested whether CAMP-dependent protein kinase activity was required for the time-dependent decrease in re- ceptor binding affinity for agonist by testing competition of (-)-isoproterenol for [‘251]ICYP with intact Kin- 549 cells. These cells also showed a 23-fold decrease in affinity of (-)- isoproterenol in competitive binding studies conducted for 1 min and 60 min (Fig. 11). Taken together, the results in four variants thus indicate neither functional N,, CAMP itself or CAMP-dependent protein phosphorylation is required to pro- duce a time-dependent decrease in agonist binding affinity of intact S49 cells.

DISCUSSION

When receptors function in intact cells, they potentially interact with a host of membrane and intracellular constitu- ents. To fully understand receptor mechanisms and dynamics in. uiuo, therefore, it is necessry to supplement information gained from studies of membrane fragments and purified components with information obtained in experiments with intact cells. The @-adrenergic receptor has been widely studied in a variety of cell types using both radioligand binding and functional approaches. An apparent discrepancy between these approaches has been noted in many studies (11-15): the dissociation constant of agonists in radioligand binding stud- ies is several orders of magnitude higher than the concentra- tion of agonist required to produce a receptor-mediated effect. A hint of an explanation for this discrepancy was obtained in our radioligand binding studies of several years ago suggesting that agonists bind initially, but transiently, to receptors on intact S49 cells with a high affinity (16). In the current studies we confirm these findings and extent our understanding of the functional importance and molecular mechanisms of the transient high affinity binding of agonists. Other laboratories have recently reported similar findings in several cell types (14, 31). Thus transient, high affinity binding may be a common feature of agonist interactions with p-adrenergic receptors of intact cells. As shown here this interaction cannot be explained as a simple bimolecular process described by the law of mass action.

Transient, high affinity binding seems likely to be impor- tant for regulation of target cell responses, as the apparent K,,, values for several agonists approach the ICso values for agonists in the binding studies conducted for 1 min. The ICso values that we obtained at 1 min are probably only approxi- mations of the true “affinity” of the cells’ receptors at this time.s In addition different agonists might bind with different kinetics so that binding at 1 min, which we used for conven- ience in competition studies, may not be an optimal measure of the highest affinity with which a particular agonist binds. This may explain why ICso values at 1 min closely match Kact values for only certain agonists. We conclude that previously

Competition curves for agonists at 1 min are quite “shallow,” indicating pseudo-Hill coefficients 4 . These curves probably repre- sent the summation of several events occurring simultaneously. The nonequilibrium conditions of these studies proscribe the use of avail- able nonlinear curve-fitting computer programs (20,24,25) which are based on equilibrium analyses and the law of mass action. Moreover, in comparing ICso values in binding studies with Keet values in CAMP

tration (in these studies, 3-5-fold the KD of ICYP) toward the ICXJ generation, one must consider the contribution of radioligand concen-

values.

-

by guest on July 24, 2020http://w

ww

.jbc.org/D

ownloaded from

Time-dependent /?-Adrenergic Binding to S49 Cells 13603

reported KIj/Kact discrepancies (11-13) probably in part reflect determinations of KIJ a t times subsequent to productive in- teraction of receptors with other components of the adenylate cyclase complex.

We have developed a computer program to simulate com- petitive radioligand binding according to a model incorporat- ing an agonist-promoted receptor state change. When the rate at which the receptors undergo the state change is set equal to the experimentally observed rate of receptor desensitiza- tion, this model accurately predicts the time course of com- petition between unlabeled agonist and radioligand antago- nist. This model thus mathetically demonstrates that our notion about an agonist-promoted receptor state change is consistent with the data. Moreover this model quantitatively links a change in agonist binding affinity with desensitization of the receptors, as measured by receptor-mediated cAMP generation. Clearly, however, the model is incomplete, as we have not defined the reactions by which the receptors are returned to the high affinity form. Presumably the contribu- tion of these reverse reactions are insignificant when high concentrations of agonist are present during the time course that we studied. Further studies will be required to extend the applicability of this model to fully account for all aspects of desensitization.

Previous studies conducted with membrane fractions have implicated formation of a transient, high affinity ternary complex between agonist, receptor, and N, as a critical step in agonist-promoted activation of adenylate cyclase in mem- branes (1, 3, 30, 32). The current data provide evidence that agonists transiently bind with high affinity to p-adrenergic receptors on intact cells. One might suppose therefore that this transient high affinity binding to intact cells results from the formation of a ternary complex among agonist, receptor, and N,; and that the low affinity binding observed subse- quently represents receptors that are uncoupled from N,. However, we were unable to demonstrate the transient high affinity binding in membrane preparations assayed either in the presence or absence of guanine nucleotides. Thus the transient high affinity binding that we observed appears not to be identical with the transient formation of a ternary complex postulated to occur when agonists bind to membrane- bound receptors in the presence of guanine nucleotides. More- over, as discussed below, a fully intact N, is not required to observe high affinity binding of agonists in intact cells.

The molecular components involved in the transition of p- adrenergic receptors from high to low affinity binding of agonists cannot yet be totally defined. Since lower affinity binding is not thermodynamically favored, energy must be added to the system for the conversion from high to low affinity binding. Perhaps agonist binding promotes a confor- mational change in receptors which predisposes the receptors to rapid covalent modification, internalization, or sequestra- tion within the plasma membrane (33-37). Recent work has emphasized a role for agonist-promoted endocytosis of p- adrenergic receptors in cultured cells (35, 36) and frog eryth- rocytes (34, 37). Is N, involved? One might imagine that receptors must interact with additional components of the adenylate cyclase complex prior to changes in agonist affinity. However, our data indicate that the 45,000-dalton N compo- nent defective in cyc-, UNC, and H21a cells is not required for time-dependent changes in agonist affinity. Moreover this component is not required for desensitization as demonstrated by the receptor-specific decrease in the ability of membrane extracts of agonist-treated cyc- cells to functionally reconsti- tute with exogenously added N units (38) and for receptors on intact UNC and cyc- cells to undergo temperature&-

pendent changes in binding of agonists (15). Of further inter- est is the evidence that CAMP-dependent protein phospho- rylation is not required to observe either the time-dependent change in agonist affinity (Fig. 11) or down regulation of p- adrenergic receptors of S49 cells, results which contrast with the recent demonstration of CAMP-mediated phosphorylation of turkey erythrocyte @-adrenergic receptors (39).

These results also appear to provide evidence in intact cells for the possible physiological importance of “uncoupling” as a mechanism for desensitization. Several groups have studied membranes prepared from cells after exposure to agonists and have noted that receptors become “uncoupled,” Le. develop low binding affinity for agonists and agonist-specific decreases in adenylate cyclase activity (30, 33, 40). The current results, plus those of others (14, 31), indicate that one can document decreases in agonist affinity “states” of intact cells over a time course that is similar to that reported for uncoupling of membrane receptors or desensitization of cAMP generation (29) after exposure of cells to agonists. Intact target cells thus seem able to terminate their response to p-adrenergic agonists by two apparently different mechanisms: receptor uncoupling ( i . e . a rapidly occurring decrease in affinity for agonists) and down regulation, a much later occurring decrease in receptor number. Both decreases appear to be physiologically relevant means of desensitizing intact cells to catecholamines.

In summary, our results lead to the following conclusions: 1) agonists initially bind to @-adrenergic receptors on intact S49 cells with a high affinity; 2) this high affinity state of the receptor exists transiently and its transition to a low affinity state coincides with the desensitization of cAMP generation in the cells; 3) the development of the high affinity binding state does not require agonist-mediated cAMP generation, a fully intact N, protein, or CAMP-dependent protein kinase.

REFERENCES 1. Ross, E. M., and Gilman, A. G. (1980) Anrzu. Reu. Biochem. 49,

2. Rodbell, M. (1980) Nature (Lond.) 284, 17-22 3. Limbird, L. (1981) Biochem. J . 195, 1-13 4. Stadel, J. M., DeLean, A., and Lefkowitz, R. J . (1982) Adu.

5. Hildebrandt, J., and Birnbaumer, L. (1982) Horn. Cell Regul. 6,

6. Spiegel, A. M., and Downs, R. W. (1981) Endocr. Reu. 2, 275- 305

7. Johnson, G. L., Kaslow, H. R., Farfel, Z., and Bourne, H. R. (1980) Adu. Cyclic Nucleotide Res. 13, 1-33

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

9. Sternweis, P. C., Northup, J. K., Smigel, M. D., and Gillman, A.

10. Hanski, E., Sternweis, P. C., Northup, K., Dromernick, A. W.,

11. Porzig, H. (1982) Trends Pharmacol. Sci. 3, 75-78 12. Insel, P. A,, and Stoolman, L. M. (1978) Mol. Pharmacol. 1 4 ,

13. Terasaki, W. L., and Brooker, G. (1978) J. Biol. Chem. 2 5 3 ,

14. Pittman, R. N., and Molinoff, P. B. (1980) J. Cyclic Nucleotide

15. Insel, P. A., and Sanda, M. (1979) J. Biol. Chem. 254,6554-6559 16. Insel, P. A., Stoolman, L. M., and Melmon, K. L. (1977) Clin.

Res. 25,463A 17. Haga, T., Ross, E. M., Anderson, H. A., and Gilman, A. G. (1977)

Proc. Natl. Acad. Sci. U. S. A . 74, 2016-2020 18. Schleifer, L. S., Garrison, J. C., Sternweis, P. C., Northup, J. K.,

and Gilman, A. G. (1980) J. Biol. Chern. 255, 2641-2644 19. Ferguson, K. M., Northup, J. K., and Gilman, A. G. (1982) Fed,

Proc. 4 1, 1407 20. Bourne, H. R., Kaslow, D., Kaslow, H. R., Salomon, M., and

533-564

Enzymol. 53, 1-43

111-130

S. A. 77,6516-6520

G. (1981) J. Biol. Chem. 256, 11517-11526

and Gilman, A. G. (1981) J. Biol. Chem. 256, 12911-12919

549-561

5418-5425

Res. 6,421-436

by guest on July 24, 2020http://w

ww

.jbc.org/D

ownloaded from

13604 Time-dependent P-Adrenergic Binding to S49 Cells

21.

22.

23. 24.

25.

26.

27.

28.

29.

30.

31.

Licko, V. (1981) Mol. Pharrnacol. 20,435-441

(1975) Science (Wash. D. C.) 190, 896-898

(1982) J. Biol. Chem. 257, 11901-11907

Insel, P. A,, Bourne, H. R., Coffino, P., and Tomkins, G. M.

Darfler, F. J., Mahan, L. C., Koachman, A. M., and Insel, P. A.

Mahan, L. C., and Insel, P. A. (1983) Anal. Biochem., in press Munson, P. J., and Rodbard, D. (1980) Anal. Biochem. 107,220-

Burgisser, E., Hancock, A. A., Lefkowitz, R. J., and DeLean, A.

Hoyer, D., Engel, G., and Berthold, R. (1982) Naunyn-Schmie-

Ehlert, F. J., Roeske, W. R., and Yamamura, H. I. (1981) Mol.

Shear, M., Insel, P. A., Melmon, K. L., and Coffino, P. (1976) J.

Erdos, J. J., and Maguire, M. E. (1980) Mol. Pharmacol. 18,379-

Su, Y.-F., Harden, T. K., and Perkins, J. P. (1980) J. Biol. Chem.

Toews, M. L., Harden, T. K., and Perkins, J. P. (1983) Proc.

239

(1981) Mol. Pharrnacol. 19, 205-216

deberg’s Arch. Pharmacol. 318, 319-329

Pharrnacol. 19, 367-371

Biol. Chem. 251, 7572-7576

383

255,7410-7419

Natl. Acad. Sci. U. S. A . 80. 3553-3557

32. Levitski, A. (1982) Trends Pharrnacol. Sci. 3, 203-206 33. Stadel, J. M., Nambi, P., Lavin, T. N., Heald, S. L., Caron, M.

G., and Lefkowitz, R. J. (1982) J. Bid. Chem. 257, 9242-9245 34. Chuang, 0.-M., Kinnier, W. J., Farker, L., and Costa, E. (1980)

Mol. Pharrnacol. 18, 348-355 35. Harden, T. K., Cotton, C. U., Waldo, G. L., Lutton, J. K., and

Perkins, J. P. (1980) Science (Wash. D. C.) 210, 441-443 36. Staehelin, M., Simon, P., Jaeggi, K., and Wigger, N. (1983) J.

Biol. Chem. 258,3496-3502 37. Stadel, J. M., Strulovici, B., Nambi, P., Lavin, T. N., Briggs, M.

M., Caron, M. G., and Lefkowitz, R. J. (1983) J . Biol. Chem. 258,3032-3038

38. Green, D. A., and Clark, R. B. (1981) J. Bid. Chem. 256, 2105- 2108

39. Stadel, J. M., Nambi, P., Shorr, R. G. L., Sawyer, D. F., Caron, M. G., and Lefkowitz, R. J. (1983) Proc. Natl. Acad. Sci. U. S. A. 80, 3173-3177

40. Lefkowitz, R. J., Wessels, M. R., and Stadel, J. M. (1980) Curr. Top. Cell. Regul. 17, 205-228

41. Cheng, Y., and Prusoff, W. H. (1973) Biochem. Pharmacol. 22, 3099-3108

by guest on July 24, 2020http://w

ww

.jbc.org/D

ownloaded from

Time-dependent @-Adrenergic Binding to 5'49 Cells

TIME-OEPENOEM OECREASES I N B I N O I K AFFINIT! W AGONISTS FOR SUPPLEMENTAL FlGURES FOR

BETA-ADRENERGIC RECEPTORS W INTACT 549 LIMPHWA CELLS

o r i o i o $0 i o ' i o l

i

1 I -10 -9 -8 -7 -6 -5 - 4 -3

lo~C(-)ISOPROTERENOLl ( M I

13605

by guest on July 24, 2020http://w

ww

.jbc.org/D

ownloaded from

P A Insel, L C Mahan, H J Motulsky, L M Stoolman and A M Koachmanreceptors of intact S49 lymphoma cells. A mechanism of desensitization.

Time-dependent decreases in binding affinity of agonists for beta-adrenergic

1983, 258:13597-13605.J. Biol. Chem. 

  http://www.jbc.org/content/258/22/13597Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/258/22/13597.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on July 24, 2020http://w

ww

.jbc.org/D

ownloaded from