The biphasic response of muscarinic cholinergic receptors in cultured heart cells to agonists. Effects on receptor number and affinity in intact cells and homogenates.

A biphasic time course of the agonist-mediated loss of muscarinic cholinergic receptors has been demonstrated in cultured chick embryo heart cells by radioligand binding studies using the muscarinic antagonist [3H]quinuclidinyl benzilate ([3H]QNB). This agonist-mediated receptor loss was associated with decreased affinity of the receptor for agonist as judged by competitive binding of the agonist carbamylcholine with [3H]QNB to cell homogenates (Galper, J. B., and Smith, T. W. (1980) J. Biol. Chem. 255, 9571-9579). In the current studies the concentration dependence of agonist-mediated receptor loss was also found to be biphasic. The apparent shift of affinity following brief (15 min) agonist exposure coincided with the agonist-mediated loss of a subclass of high affinity receptors with an IC50 for carbamylcholine inhibition of [3H]QNB binding of 3.9 x 10(-7) M. Those receptors remaining constituted a subclass of low affinity receptors with IC50 = 8.2 x 10(-5) M. The data further suggest that an apparent decrease in agonist affinity after guanine nucleotide exposure represents conversion of high affinity receptors to a similar low affinity state, IC50 = 8.6 x 10(-5) M. The rapid loss of [3H]QNB binding sites in the presence of agonist did not require interaction of agonist with intact cells, but also occurred if cells were homogenized and then subjected to a brief (15 min) exposure to agonist. The slow loss over 3 h of [3H]QNB binding sites could only be demonstrated in intact cells incubated with agonist prior to homogenization. To probe further the later phase of agonist-mediated receptor loss, we developed new assay methods for determining muscarinic antagonist binding to intact cells. In control cells, binding of the hydrophobic antagonist [3H]QNB was quite similar in extent to binding of the more hydrophilic antagonist [3H]methylscopolamine ([3H]MS), with Kd values of 0.11 and 0.47 nM, respectively. Kinetic analysis of the binding of these two ligands was performed to determine whether they might distinguish between two states of the receptor. Both ligands bound to the receptor by a two step mechanism consistent with the formation of a low affinity complex followed by conversion to a high affinity complex. However, the ratio of reverse to forward rate constants of the second step of [3H]MS binding was roughly 100-fold greater than that for the more hydrophobic ligand [3H]QNB. Comparison of the time course of agonist-induced receptor loss as measured by binding of [3H]MS or [3H]QNB was consistent with muscarinic agonist mediation of a stepwise alteration in receptor configuration from a form that bound both [3H]MS and [3H]QNB to a form that bound only [3H]QNB and thence to a form that bound neither [3H]MS nor [3H]QNB. The relationship of such a sequential mechanism to agonist-induced changes in the relationship of the receptor to the cell membrane and agonist-induced endocytosis of the receptor is discussed.

A biphasic time course of the agonist-mediated loss of muscarinic cholinergic receptors has been demonstrated in cultured chick embryo heart cells by radioligand binding studies using the muscarinic antagonist [3H]quinuclidinyl benzilate ([3HlQNB). This agonistmediated receptor loss was associated with decreased affinity of the receptor for agonist as judged by competitive binding of the agonist carbamylcholine with [3H]QNB to cell homogenates (Galper, J. B., and Smith, T. W. (1980) J. Biol. Chem. 255,[9571][9572][9573][9574][9575][9576][9577][9578][9579]. In the current studies the concentration dependence of agonist-mediated receptor loss was also found to be biphasic. The apparent shift of affinity following brief (15 min) agonist exposure coincided with the agonist-mediated loss of a subclass of high affinity receptors with an ICso for carbamylcholine inhibition of [3H]QNB binding of 3.9 X lo-' M. Those receptors remaining constituted a subclass of low affinity receptors with ICs0 = 8.2 X M. The data further suggest that an apparent decrease in agonist affinity after guanine nucleotide exposure represents conversion of high affinity receptors to a similar low affinity state, ICso = 8.6 X

M.
The rapid loss of [3H]QNB binding sites in the presence of agonist did not require interaction of agonist with intact cells, but also occurred if cells were homogenized and then subjected to a brief (15 min) exposure to agonist. The slow loss over 3 h of 13H]QNB binding sites could only be demonstrated in intact cells incubated with agonist prior to homogenization.
To probe further the later phase of agonist-mediated recpetor loss, we developed new assay methods for determining muscarinic antagonist binding to intact cells. In control cells, binding of the hydrophobic antagonist C3H]QNB was quite similar in extent to binding of the more hydrophilic antagonist [3H]methylscopolamine ([3H]MS), with Kd values of 0.11 and 0.47 n~, respectively. Kinetic analysis of the binding of these two ligands was performed to determine whether they might distinguish between two states of the receptor. Both ligands bound to the receptor by a two step mechanism consistent with the formation of a low affinity complex followed by conversion to a high affinity complex. However, the ratio of reverse to forward rate constants of the second step of C3H]MS binding was * This study was supported by research Grants HL-22775 and HL-18003 from the National Heart, Lung and Blood Institute, National Institutes of Health and American Heart Association Grant 79825. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. + Recipient of Young Investigator Award, National Heart, Lung, and Blood Institute, National Institutes of Health. roughly 100-fold greater than that for the more hydrophobic ligand 13H]QNB. Comparison of the time course of agonist-induced receptor loss as measured by binding of 13H]MS or [3H]QNB was consistent with muscarinic agonist mediation of a stepwise alteration in receptor configuration from a form that bound both [3H] MS and 13H]QNB to a form that bound only [3H]QNB and thence to a form that bound neither [3H]MS nor [3H]QNB. The relationship of such a sequential mechanism to angonist-induced changes in the relationship of the receptor to the cell membrane and agonist-induced endocytosis of the receptor is discussed.
The ability of hormones and neurotransmitters to regulate the number and/or affinity of cell surface receptors has been demonstrated in excitable tissues, endocrine glands, and other tissues subject to hormonal control (1). Guanine nucleotides have also been shown to modulate the binding of agonists to several classes of receptors including the P-adrenergic receptor, the glucagon receptor, and the muscarinic cholinergic receptor (2)(3)(4).
Recently, experiments using the potent muscarinic antagonist [3H]QNB' for the identification of muscarinic binding sites have demonstrated that prior exposure of neuroblastoma cells (5) or cultures of embryonic chick heart cells (6) to muscarinic cholinergic agonists decreased the number of muscarinic binding sites by greater than 70%. In cultured heart cells the time course of loss of binding sites during agonist exposure was biphasic with an early rapid loss (1 min) of 26% of binding sites followed by a 20-min lag phase and the gradual loss of another 43% of receptors over 3 h (7). The rapid loss of 26% of receptor sites was accompanied by a 6-fold decrease in apparent affinity of the remaining receptors for agonist. When homogenates of rat heart (8,9) or chick embryo heart cell cultures (7) were exposed to guanine nucleotides, affinity of muscarinic receptors for agonist decreased while receptor number remained unchanged. When homogenates of chick embryo heart cell cultures which had been exposed briefly to agonist were homogenized and incubated with guanine nucleotides, affinity remained decreased while the apparent number of receptor sites recovered to control levels (7).
The slow loss of 43% of muscarinic receptors during a 3-h exposure to agonist was only reversible after a 12-h incubation in the absence of agonist, and required protein synthesis (7). Inhibitors of microtubule function inhibited 43% of total ago-nist-induced receptor loss. These data suggested that the subclass of 43% of receptors that disappear slowly during agonist exposure of cultured heart cells might be subject to endocytosis and degradation.
Agonist-induced endocytosis of cell surface receptors has been demonstrated for several types of hormone binding sites (1). In one of the most extensively studied cases, epidermal growth factor interaction with the human fibroblast has been shown to involve endocytosis of the receptor-growth factor complex with probable subsequent lysosomal degradation of both the hormone and receptor within the cell (10).
The experiments described here address two major aspects of the interaction between muscarinic agonists and receptors in cultured heart cells. First, we extend previous studies of the early rapid phase of agonist binding. The findings presented support the hypothesis that the rapid loss of 26% of receptors after brief exposure to agonist is due to persistent binding of agonist to a class of high affinity receptor sites, and that guanine nucleotides mediate recovery of these sites by facilitating release of persistently bound agonist during guanine nucleotide-mediated conversion of the high affinity receptor to a low affinity form. A second set of experiments in which we developed an assay for the binding of muscarinic antagonists to intact heart cells explores further the mechanism by which prolonged agonist exposure mediates loss of receptors. These experiments are divided into two parts. First, we compared the binding of the muscarinic antagonists ['HIQNB and ['HIMS to the intact cell. Kinetic studies demonstrated that although both antagonists bound to the receptor via a sequential two step mechanism, the antagonist-high affinity receptor complex formed with the more hydrophobic ["HIQNB demonstrated marked kinetic differences from the ['HIMS-high affinity receptor complex. In the second set of experiments we compared the binding of ["HIQNB and ['HIMS to cells which had been exposed for various times to high concentrations of agonist prior to antagonist binding. Our findings support the view that properties of binding of these radioligands to intact cells are capable of differentiating between two states of the receptor: the control state (before agonist binding) which bound both ["HIMS and ["HJQNB and a state present during the late phase of agonist exposure which bound only ['HIQNB and which may be an intermediate state in the process of receptor endocytosis. Thus, alterations in the muscarinic receptor subsequent to binding of agonist involve distinct and separable initial and late events that can be distinguished by analysis of specific radioligand binding data.
Heart Cell Cultures-Heart cell cultures were prepared by a modification of the method of DeHaan (11) as described (7). Embryos were removed from embryonated Leghorn chicken eggs on day 10 in trypsin in Cay'-Mg"-free Hanks' balanced salt solution at 37 "C for OUO. Hearts were removed, minced, and incubated with 0.25% (w/v) 8 min. The trypsin solution was removed and diluted into medium "199 containing 50% heat-inactivated horse serum a t room temperature. After successive trypsinizations, suspensions of trypsinized cells were sedimented a t 1000 rpm in a desk top centrifuge, resuspended in growth medium, and incubated in a 100-mm Petri dish (Falcon, Oxnard, CA) for 45 min a t 37 "C in a humidified atmosphere of 5% CO2/95% air. During this incubation, nearly 95% of the fibroblasts in the suspension adhered to the dish. The heart cells then were plated at a density of 1.3 X lo5 cells/cm' on collagen-coated 100-mm Petri dishes or a t a density of 2.0 X lo5 cells/cm' in 16-mm multiwell dishes (Costar, Cambridge, MA) containing 24 wells/unit. On the third day of incubation, the medium was changed. Unless otherwise indicated, cells were used for experiments on culture day 4.
Measurement of [:' H]QNB Binding to Homogenates-The assay procedure was a modification of the method of Yamamura and Snyder (12). After three rinses with ice-cold "199 HEPES, cells were harvested in a small volume of HEPES-buffered medium "199. After freezing at -70 "C and thawing twice, the cells were homogenized in a glass on glass homogenizer and allowed to warm to 22 "C prior to assay.
The final assay mixture was 0.5 ml of HEPES-buffered "199 containing l-["H]QNB (43 Ci/mol, I nM in ["HIQNB unless otherwise specified), drugs as indicated in the figure legends, and 0.5 ml of homogenate. At the appropriate time (1 h a t 22 "C unless otherwise stated), 5 ml of wash medium (120 mM NaC1, 5.4 mM KCI, 0.8 mM MgSO,, 1.8 mM CaC12,50 mM HEPES, 1.0 mM NaH2P04, adjusted to pH 7.4 with NaOH) at 22 "C was added to terminate the incubation.
The reaction mixture was passed through a Whatman glass fiber (GF/C) filter, the assay tube washed 3 times, and filters dried and assayed for radioactivity in a liquid scintillation counter with 10 ml of Instagel (Packard). Protein was determined by the procedure of Lowry et al. (13) after precipitation of the homogenate protein with trichloroacetic acid, using bovine serum albumin as standard.
Specific binding is defined as binding inhibited by saturating concentrations (0.1 n m ) of oxotremorine. ['HIQNB binding to homogenates was 90% specific by this criterion. All data were corrected for nonspecific binding in the presence of 0.1 mM oxotremorine. A typical assay tube containing 0.2 mg of protein/0.5 ml of homogenate suspension bound 900 to 1000 cpm of rH]QNB at a counting efficiency of 33%.
Measurement of ["H]MS Binding to Intact Cells-Cells prepared as described in 16-mm multiwell plates were fed with fresh medium containing [U-'4C]leucine, 0.01 mCi/ml, on the third culture day and incubation continued for 24 h. To study the binding of [,"HIM& 0.5 ml of HEPES-buffered "199, and the indicated concentration of ["HIMS were added to each well and incubated at 37 "C for 1 h.
Wells were rinsed rapidly 3 times with 3 ml of ice-cold wash medium and 1 ml of 0.5 N NaOH added to each well to solubilize cell proteins. Aliquots, 0.8 ml , were taken from each well, neutralized with 0.5 ml of 1 M Tris, pH 7.4, and assayed for ,'H and 14C with 10 cc of Instagel in a liquid scintillation counter. Protein content was determined for one well in four by the method of Lowry et al. (13) and the mean protein content per 1000 14C-counts calculated. This factor permitted calculation of protein content of cells in each well. All [:' HI MS binding was normalized to specific binding per mg of cell protein.
In principle, the four rate constants can be evaluated simultane-O U S~' using Equation 1. However, more reproducible estimates were obtained if k -I and h-2 were f i s t determined from dissociation data dissociation of bound ligand is and then used as fured values in Equation 1. The time course of BIIBo = ae -k -d + (1a ) e -k l', (2)  The rapid loss of 26% of sites during the first minute of agonist exposure was followed by the gradual loss of an additional 43% of the total receptor population over 3 h. The experiments described here were designed to determine whether the rapid loss of receptors involved the disappearance of a subclass of high affinity receptors while the slower loss of receptors involved the disappearance of a subclass of low affinity receptors.
The results of an experiment to determine the effect of a brief (15 min) exposure of cells to various concentrations of the agonist carbamylcholine on the binding of ['HIQNB to homogenates of these cells are shown in Fig. 1. These data demonstrate a 23% maximal decrease in ["HIQNB binding sites with a half-maximal effect at M carbamylcholine. In the same experiment, a 3-h exposure of cells to various concentrations of agonist resulted in a maximal decrease of 70% in ["HIQNB binding to homogenates with a half-maximal effect at 1.7 X lo-" M carbamylcholine. The data in Fig. 1 show that at low concentrations of agonist (less than M ) , the loss of high affinity binding sites was complete within 15 min and no further loss could be detected following a 3-h exposure to agonist. If the subclass of receptors lost during a 15-min exposure to agonist is subtracted from the total receptor loss at 3 h, the data further indicate that during the 2 h and 45 min following rapid loss of receptors, a second subclass of 46% of receptor sites is lost with a half-maximal effect of agonist at 3.0 X M carbamylcholine. Analysis of the data for the effect of a 3-h exposure to various concentrations of agonist on ['HIQNB binding (Fig. lA (Fig. 1B). These data are consistent with the view that both the time course and concentration dependence of agonist-induced loss of ["HIQNB binding sites are biphasic, involving at least two separate classes of receptor sites.
We have shown that the relative effectiveness of a muscarinic agonist in mediating a decrease in r3H]QNB binding sites was proportional both to its ability to compete with ['HIQNB for receptor binding and to the relative pharmacologic potency of that agonist (7). These findings indicated that agonist modulation of receptor number involves specific binding of agonist to muscarinic receptors. Hence, the most obvious interpretation of the data in Fig. 1 is that brief agonist exposure results in loss of high affinity receptors while prolonged agonist exposure results in loss of lower affinity receptors. The Effects of Agonist and Guanine Nucleotides on High Affinity Muscarinic Receptors-More direct evidence for the presence of two subclasses of muscarinic receptors and the rapid loss of high affinity sites following brief agonist exposure was obtained from replicate measurements of the affinity of the muscarinic receptor for carbamylcholine. Affinity was determined by the ability of various concentrations of agonist to compete with [3H]QNB for binding to cell homogenates. In homogenates of cells not subjected to agonist exposure prior to homogenization, such measurements of affinity were biphasic. The shoulder seen at low concentrations of carbamylcholine ( Fig. 2, closed circles) suggests the presence of a subclass of high affinity receptors that reaches a maximum value of approximately 37 fmol/mg of protein. Computer analysis of these data by nonlinear regression for a two receptor class model was consistent with a subset of high affinity recpetors (RH), 34 k 3.5 (S.D., n = 25) fmol/mg of protein (26% of total receptors), with an ICX, for carbamylcholine displacement of ['HIQNB at 3.9 f 1.0 X and a second subset of low affinity receptors, 95 3.6 fmol/mg of protein, This curve was nearly identical with that derived for R,. from the control curve in Fig. 2 by subtracting the extrapolation of the high affinity portion of the curve from total control binding ( Fig. 2, open squares). The simplest explanation of these data is that brief exposure of cells to agonist results in loss of the ability of labeled antagonist to bind to high affinity sites. With only R,, remaining, the affinity of the remaining receptors for agonist would be that of RL. The data suggest that agonist has little, if any, direct effect on the affinity of RI,.
Further evidence for the presence of high affinity receptors and their relationship to RL can be demonstrated by the effects of guanine nucleotides on RH. Exposure of cell homogenates to M Gpp(NH)p also eliminated the shoulder seen in the control curve (Fig. 2) and gave a close fit to a curve describing a single class of low affinity receptors with ICso of 8.6 f 1.2 X M (Fig. 2, open circles). However, exposure to Gpp(NH)p had no effect on total receptor number. This value for 1C.m is similar to the IC50 of 8.2 X M for those receptors remaining in homogenates of cells exposed for 15 min to lo-'' M carbamylcholine. Similar results were obtained when homogenates were incubated with GTP (data not shown).
The shift of the ICrfi to a value similar to that for RI, following treatment of homogenates with Gpp(NH)p and the absence of any effect of Gpp(NH)p on total receptor number supports the view that Gpp(NH)p mediates the conversion of RH to RL. The data do not support a significant direct effect of guanine nucleotides on the affinity of RI, sites. Hence, both guanine nucleotides and brief agonist exposure appear to affect only the high affinity subclass of receptors. Any direct effect of these agents on the low affinity receptor is small and of doubtful significance.

The Role of the Intact Cell in the Biphasic Response of the
Muscarinic Receptor to Agonist Binding-In experiments described previously, intact cells were exposed to agonist and the effect on the number of [,'H]QNB binding sites and affinity of receptors for agonist were measured in homogenates of these cells ( Fig. 2; Refs. 6 and 7). To determine whether the effects of both short term and more prolonged agonist exposure on receptor number required an intact cell for their expression, cell-free homogenates were exposed to muscarinic agonists. The experiment summarized in Fig. 3 demonstrates the effect of a brief (15 min) exposure of cell homogenates to various carbamylcholine concentrations, followed by removal of agonist by repeated centrifugations, on both the total binding of ['HIQNB and on receptor affinity for agonist. Under these conditions 26% of receptor sites were lost and the apparent ICSo for carbamylcholine competition with ["HIQNB binding increased from 2.0 X to 7.1 X M. Although the shoulder at low carbamylcholine concentrations seen in Fig. 2 is not as apparent in the control curve of Fig. 3 (open circles), the shift in affinity following exposure to carbamylcholine is consistent with loss of high affinity receptors. These data indicates that both the agonist-induced rapid loss of high affinity receptors and the effect of guanine nucleotides on high Pellets were resuspended in a small volume of M199-HEPES at approximately 3 mg/ml and incubated at room temperature in 1 nM ['HIQNB and the indicated concentrations of carbamylcholine. Each point is the mean of three replicate determinations. The solid lines are drawn by eye. 0, control homogenates not exposed to carbamylcholine; 0, homogenates pre-exposed to 10 .' M carbamylcholine for affinity receptors (Fig. 2) do not require the intact cell for their expression.
Previously published results suggest that the slow phase of agonist-induced receptor loss involves microtubule function and endocytosis of the ceU surface receptor (7). If this view is correct, the cell surface and its interaction with microtubules should be intact in order to mediate the slow response to agonist exposure. When cell homogenates were incubated for 3 h with M carbamylcholine, no further loss of ["HIQNB binding sites could be detected beyond that noted after a 15min exposure to agonist (Fig. 3, open squares). Hence, the slow phase of receptor loss does not take place in a broken cell preparation.
Binding of Muscarinic Antagonists to Intact Cultured Heart Cells-The studies to be described next were designed to elucidate the mechanism by which prolonged exposure to agonist results in a loss of receptor sites. Because the slow phase of agonist-induced receptor loss requires an intact cell membrane and does not take place in cell homogenates (Fig.  3), studies of the effects of prolonged agonist exposure on receptor number must be carried out in intact cells. Furthermore, the physiologic effects of muscarinic agonists on the heart including changes in K+ permeability (17) and the rate and force of contraction (18) require the interaction of agonist with receptors on the surface of the intact cell. Receptor number measured in heart cell homogenates may not reflect the subset of receptors in the intact cell available for agonist binding. For these reasons, we characterized the binding of muscarinic antagonists to the intact cell. We compared the binding of ["HIMS and ['HIQNB to intact cells under control conditions and following exposure to muscarinic agonists. Although at pH 7.4 both ['HIMS and ['HIQNB are predominantly positively charged, we determined that ['HIQNB is markedly more hydrophobic than ["HIMS. A comparison of the partition of ["HIQNB and [?H]MS between an aqueous phase consisting of the wash solution described under "Experimental Procedures" for ligand binding studies (pH 7.4, physiologic ionic strength) and an organic phase consisting of ether or chloroform revealed that the ether:aqueous phase partition coefficient for ['HIQNB was 10 times greater than that for r3H]MS and the chlorofonmaqueous partition coefficient for ['HIQNB was 30 times greater than the corresponding values for ['HIMS. Our previous data (7) having suggested that loss of receptors during prolonged agonist exposure may involve an endocytotic process, we postulated that this process might involve as an initial step an agonist-induced alteration in the configuration of the receptor within the cell membrane.
We considered the possibility that the more hydrophobic r3H] QNB might diffuse into or through the intact cell membrane and bind to receptors that were not available for interaction with agonist at the cell surface while the less hydrophobic ['HIMS might bind only to receptors on the cell surface. Hence, differences in binding of these two antagonists to the receptor in the control state might offer a probe for the study of changes in the configuration and/or localization of receptors in the cell membrane following agonist exposure. Fig. 4 ( The concentrations at which carbamylcholine and oxotremorine inhibited 50% of binding of 1 nM ['HIQNB to intact cells are significantly higher than those reported previously for competition with 1 nM ['HIQNB in cell homogenates (6, 7 ) .

An experiment comparing the binding of ['HJMS and ['HI QNB to intact cells after a 2-h incubation with various concentrations of either ['H]MS or ["HIQNB is shown in
For example, in homogenates IC50 was 3 x compared to 3 X W 4 M in intact cells. One explanation for these findings is that in intact cells, during the measurement of affinity of the receptor for agonist by competition of carbamylcholine with "H-antagonist, carbamylcholine is simultaneously decreasing the affinity of the receptor for further agonist binding. Thus, apparent IC& might be increased compared to measurements in homogenates. Alternatively, the presence of endogenous guanine nucleotides in intact cells could mediate the conversion of RH to RL in a manner similar to the effect of exogenous guanine nucleotides on homogenates (Fig. 2,  open circles). Fig. 4 indicate that [3H]QNB has a 4-fold higher affinity for the receptor than ['HIMS. In order to determine whether these differences in affinity might reflect important differences in the kinetics of binding due to preferential binding of each ligand to a different state of the receptor, a comparison of the kinetics of binding of ['HJQNB and r3H]MS was carried out. Data presented below indicate that the binding of both ligands involves sequential formation of a low affinity agonist-receptor complex followed by the conversion of this complex to a high affinity form. Although the rate constants for formation of the low affinity complex are quite similar for ["HIMS and [3H]QNB, the probability of rearrangement of the ['H)MS low affinity complex to a high affinity form is much less favorable than that for ['HIQNB. These data suggest that a form of the receptor does exist which preferentially binds ['HIQNB.

Kinetic Analysis of ['HIMS and r3H]QNB Binding to Intact Cultured Heart Cells-Data from
As shown in Fig. 6, at a concentration of 2 nM binding of each ligand proceeded without a lag and reached 50% of the equilibrium value at 4.5 min for ["HIMS and 2 min for ["HI QNB, reflecting a significantly lower rate of binding for the less hydrophobic r3H]MS.
Data from an experiment in which initial rates of binding are plotted according to the equation for a simple bimolecular process are shown in Fig. 7 . Both ['HIQNB and [3H]MS gave biphasic curves. The simplest mechanisms that would yield the data shown in Fig. 7 would be parallel reactions of the ligands with two binding sites having different kinetic properties or a two stage sequential reaction, e.g. binding followed by a rearrangement yielding a more stable complex. The fraction of total sites reacting in the slow phase can be approximated by extrapolating tbe linear portion of the plots in Fig. 7 to t = 0. Inspection of kinetic plots such as those in Fig. 7 indicates that the fraction of receptors reacting in the rapid phase of the reaction increases from less than 10% to greater than 50% as the concentration of QNB or methylscopolamine increases (data not shown). This finding is inconsistent with the parallel reaction mechanism, in which a constant fraction of receptors should react rapidly. It is, however, consistent with a sequential reaction mechanism of the  3. Furthermore, we have previously demonstrated that the binding of ["HIQNB (Q) can be demonstrated to proceed with the formation of a rapidly reversible complex QR * which forms quickly, reaches a maximum concentration by 2 to 3 min, and then disappears.
A slowly reversible complex QR appears more slowly and continues to increase until nearly 90% of total QNB bound is in a slowly reversible form (19). The data, then, are consistent with a sequential reaction mechanism where k l and k:! are forward rate constants for the initial and late phase of the reaction, respectively, and k-l and h-, are the reverse rate constants for these two phases. When ligand is present in excess the data in Fig. 7 can be represented by the sum of two exponentials, as indicated in Equation 1 under "Experimental Procedures." When In [B, -Bt)/B,,] is plotted as a function of time (Fig. 7), the limiting slopes of the rapid and slow phases of the binding curves correspond respectively to X2 and As of Equation 1. Bt represents total ligand bound at time t, the sum of the rapidly and slowly reversible complexes QR * plus QR; Be, equals the value of Bt at equilibrium. The data in Fig. 7 were fit to Equation 1 by a nonlinear least squares analysis (see "Experimental Procedures") and gave the fit demonstrated by the coincidence of the experimental and computer-fitted data in Fig. 7 with a kl at 37 "C for binding of ["HIQNB of 0.21 X 10' M-hin" and a k , for binding of FHIMS of 0.13 X 10' M"min-'. The value of k l for ["HIQNB is in close agreement with the value of 0.19 X lo' "'rnin-' obtained for the binding of ["HIQNB to cell homogenates at 23 "C (6) and confirms the very rapid nature of the initial phase of binding. The rate constant k I for binding of ["HIMS was only slightly less than that for ["HIQNB, indicating that initial binding of ["HIMS and ["HIQNB to the intact cell is quite similar. However, the forward rate constants estimated from Equation 1 for the postulated second phase of antagonist binding (k2) were 0.54 min" for ["HIQNB and 0.04 min" for VHIMS, consistent with a markedly slower rearrangement of the initial ['HIMS-receptor complex during the second phase of the reaction (Table I).
The dissociation of ["HIQNB-and ['HIMS-receptor complexes was also biphasic; dissociation rat,e constants were estimated from the slopes in   However, the reverse rate constants for the second, more slowly reversible phase of the reaction ( L 2 of Equation 2) were estimated to be 0.006 min" and 0.045 min" for [3H]QNB and ['HIMS, respectively, a nearly 8-fold more rapid reverse reaction rate for ["HIMS. Hence, ['HIMS not only has a slower rate of transition for the second forward phase of the reaction than ['HIQNB, but also a more rapid rate of reversal for this reaction, L r , than ["HIQNB (Table I).
In determining Kd from k-2/h2 for the second phase of the reaction only a range of values could be obtained because of the propagated errors in the parameter estimates for h-2.
Hence, the Kd estimate for ["HIMS varied from 0.5 to 1.7 while that for ['HH]QNB varied from 0.004 to 0.023. These data are consistent with a roughly 100-fold decrease in the effectiveness with which the ['HJMS-receptor complex is converted from QR* to QR (Equation 3) compared to the analogous ['HIQNB transition. One possible interpretation of these data is that the second phase of the antagonist binding reaction involves a rearrangement of the ligand-receptor complex within the membrane. If this were the case, the more hydrophobic properties of [3H]QNB might facilitate its interaction with a more hydrophobic domain outside the active site of the receptor and within the surrounding membrane, and account for the markedly higher efficiency with which the ["HIQNB-receptor complex is converted to this form.
The range of values for the apparent K d for the binding of ['HIQNB and E3H]MS derived from kinetic parameters are compared to the K,, for equilibrium binding for C:'H]MS and ['HIQNB in Table I. The discrepancy between the kinetically derived values of Kd and those derived by equilibrium binding studies reflect the uncertainties in the determination of h2 and especially of k-*. Delineation of the slow phase of dissociation involves the measurement of a small change in the small amount of residual binding seen a t late times, while the slow phase of formation of the complex involves the measurement of small increments in binding at times when binding is approaching equilibrium. Hence, both of these measurements are subject to more uncertainty than the remaining kinetic data.
Agonist-mediated Changes in the Binding of ['HJQNB a n d f H ] M S to Intact CetZs-Previously reported data support the hypothesis that the slow phase of agonist-induced receptor loss involves endocytosis of cell surface receptors (7). If this were the case, agonist-induced receptor loss might involve an agonist-mediated configurational change in the receptor followed by movement of the receptor into or across the cell membrane. If ["HIMS and ['HIQNB bind with markedly different affinities to a form of the receptor altered as a result of agonist interaction, we reasoned that these two ligands might distinguish between different configurations assumed by the receptor during prolonged agonist exposures. We therefore compared the binding of ["HIMS and ["HIQNB to intact cells following prolonged agonist exposure. Data from an experiment in which ['HIQNB and ["HIMS binding were compared in cells following a 6-h exposure to various concentrations of carbamylcholine is shown in Fig. 9. No difference could be detected between the binding of the two ligands. Sixty-eight per cent of radioiigand binding sites were lost with a half-maximal effect at 3 X M carbamylcholine. A similar IC, for carbamylcholine-induced receptor loss was noted previously for cell homogenates (7).
An experiment comparing the time course of agonist-induced receptor loss in intact cells as measured by the binding of ['HIMS or VHIQNB is shown in Fig. 1OA. The binding of ['HIMS to intact cells (closed circles) was unaffected by a 10to 12-min prior exposure to muscarinic agonists. However, following this lag period nearly 80% of ['HIMS binding sites were lost over the next 6 h of agonist exposure. Half of these receptors were lost during the fiist 30 min of agonist exposure following the lag phase (Fig. 1OA). These binding sites recovered during a 12-h incubation in fresh medium. Recovery was inhibited by cycloheximide, 2 ,ug/ml (data not shown), and hence was dependent on synthesis of protein. The time course of agonist-mediated receptor loss as measured by the binding of [3H]QNB (open circles) was significantly slower than that measured with r3H]MS. Following a 35-37 min lag period, 76% of [3H]QNB binding sites were lost over the next 5% h (Fig. 1OA). Half of the receptor sites were lost approximately 107 min after the end of the lag phase. Hence, although a 6-h exposure to carbamylcholine caused a nearly identical decrease in total receptor number as measured by either ["HI MS or C3H]QNB binding (Fig. 9), the loss of [3H]QNB binding sites takes place following a longer lag period and at a slower rate than does the loss of 13H]MS binding sites.
Although these data might be explained by an agonistinduced decrease in E"H]MS binding which preceeded a decrease in [3H]QNB binding, the findings could also be due to an agonist-mediated decrease in affinity of these ligands for the receptor with a greater effect on [3H]MS affinity than that of ["HIQNB. An experiment comparing the binding of [3H] QNB or [3H]MS to intact cells following a 3-h agonist exposure, however, disclosed no significant effect of agonist exposure on Kd for either r3H]MS or [3H]QNB ccimpared to that in control cells (20).
Since the loss of the ability of the receptor to bind [3H]MS precedes the loss of binding sites for ['HIQNB, these data suggest the possibility of an agonist-mediated sequential conversion of the receptor to different forms. Such a conversion might be represented by a series of consecutive irreversible first order reactions:  Table I). The reactions can be considered irreversible during the time course of these studies since agonist-mediated loss of the ability to bind ["HI MS or 13H]QNB was not reversed for up to 3 h following removal of agonist and recovered fully only 9 to 12 h after agonist removal (data not shown). Three differential equations describe the time course of changes of A, B, and C according to classical precursor-product considerations (21): The sequential reactions in Equation 4 should not be confused with those described in Equation 3 for the two step binding of ["HIQNB and [3H]MS to the receptor. The sequential reactions in Equation 4 apply here only to cells that are first exposed to agonist for various times before binding of r3H]MS or [3H]QNB is measured.
The predicted concentrations of A, B, and C at any time may be derived from the data in Fig. 1OA. Computer analysis of these data by nonlinear regression for two sequential irreversible reactions described by Equations 5a-c demonstrated that KI and KII were approximately equal. Consequently, the data were fit to a model in which Kr was set equal to ~I I .

The
solutions of Equations 6a-c when hl = hII are

C = A o -A -B (6c)
where An is the number of receptor sites at time 0.
A fit of the data in Fig. 10B to Equations 6a-c gave An = 157 f 4 fmol/mg of protein ( n = lo), h = 0.0117 f 0.0006 rnin", and total receptor number remaining at 6 h = 30 f 4 fmol/mg or about 20% of receptor sites. The equivalence of kI and KII suggests either that loss of both [3H]MS and 13H]QNB binding sites are part of a continuous time-dependent process rather than two independent processes, or that Kr and KIr describe the rates of two independent processes which happen to be similar in rate. In any event, these data show that after a 10-to 12- 6 h (open squares). These data strongly suggest that ['HIMS and [3H]QNB binding data are capable of delineating two distinct states of the receptor.

DISCUSSION
The studies described here comprise two parts. First, we demonstrate the presence of two receptor subtypes and provide evidence consistent with the hypothesis that almost immediately upon agonist exposure an agonist-high affinity receptor complex is formed which renders a subclass of high affinity receptors unavailable for subsequent binding of ["HI QNB. The second set of studies comprises two parts. First, we demonstrate that I3H]QNB and ['HIMS bind to the receptor by a sequential process with formation of a rapidly reversible low affinity antagonist receptor complex followed by rearrangement of this complex to a slowly reversible form. The more hydrophobic [3H]QNB formed the high affinity complex 100-fold more readily than did [3H]methylscopolamine. Hence, the receptor was capable of assuming a form which preferentially bound I"H]QNB. In the second set of experiments, we compared the binding of ['HIMS and ['HIQNB to receptors which had been exposed for various times to high concentrations of muscarinic agonists. In these studies, prior exposure to agonist resulted in conversion of muscarinic receptors from a form which bound both ["HIMS and ['HIQNB to a form which bound only ["HIQNB. This set of studies indicates that ['HIQNB and E3H]MS recognize different forms assumed by the receptor during agonist-induced receptor down-regulation and support the hypothesis that this process involves conversion of the receptor to an altered form which may be an intermediate in the process of receptor endocytosis. In the discussion that follows, the relationship of these immediate and late effects of agonist-receptor interaction to the ability of the cultured heart cell to respond to muscarinic stimulation will be considered.
The initial response of the muscarinic receptor population to agonist is the rapid loss of a fraction of ["HIQNB binding sites (26%) which corresponds to a subclass of high affinity receptors with an ICs0 for carbamylcholine inhibition of ['HI QNB binding of 3.9 X lo" M (Fig. 2). This loss of receptor sites did not require an intact cell for its expression and could be demonstrated in homogenates from intact cells which had been exposed briefly to agonist prior to homogenization (Fig.  2) or in cell-free homogenates exposed to agonist (Fig. 3) following homogenization. Our studies indicate that the loss of RH was complete as early as 1 to 15 min following exposure of cells to concentrations of carbamylcholine less than M (Fig. 1). The affinity for agonist of those receptors remaining in homogenates of agonist-treated cells was similar to that for RL (Fig. 2 ) . A plausible interpretation of these data is that

Control of Muscarinic Receptors in Cultured Heart
Cells 10353 rapid receptor loss involves persistent binding of agonist to RH following brief exposure to agonist, leaving the high affinity receptor occupied and hence unavailable for "H-antagonist binding. Under these conditions RH would appear to be lost, leaving only RL. Alternatively, brief agonist exposure could lead to assumption of a receptor configuration inaccessible to 'H-antagonist. Incubation of cell homogenates with Gpp(NH)p also caused the disappearance of RH, but without a decrease in total ["HI QNB binding sites (Fig. 2). These findings suggested that guanine nucleotides mediate the conversion of RH to RL while having only a small direct effect, if any, on RL. We have previously shown (7) that incubation of homogenates of cells which had been briefly exposed to agonist with guanine nucleotides resulted in the reappearance of the subset of 26% of receptor sites lost during brief agonist exposure. These receptors reappeared in a low affinity form. These findings suggest that persistently bound agonist was released during the guanine nucleotide-mediated conversion of RH to RL.
The data in Fig. 10 revealed that the time course of agonistinduced receptor loss measured by ["HIMS binding in intact cells is quite similar to the time course of the slow phase of receptor loss assayed in homogenates of cells exposed to agonist prior to homogenization (7). In both intact cells and homogenates a 15-min lag period was followed by a slow receptor loss that was half-maximal at 30 min. However, unlike studies of cell homogenates, neither ['HIMS nor ["HI QNB binding to intact cells demonstrated the rapid loss of a subclass of 26% of receptors preceding the lag phase. However, the assay of ['HIMS and/or ['HIQNB binding to the intact cell differed from that in the homogenate.
Following brief exposure to agonists, cells were washed and then incubated for 1 h with "H-antagonist. During this incubation cells continued to contract a t 140 k 5 (S.D., n = 20) beats/min, indicating that normal energy metabolism was maintained. We have demonstrated that exogenously added guanine nucleotides mediate the recovery of ['HIQNB binding sites in homogenates of cells which had been exposed briefly to agonist prior to homogenization. During binding of "-antagonist to the intact cell, physiologic levels of guanine nucleotides could mediate recovery of ['HIQNB or ['HIMS binding sites. If the rapid loss of "-antagonist binding sites were due to persistent binding of agonists to high affinity receptors, then in the intact cell endogenous GTP might mediate the release of bound agonist. Such a mechanism might play a role in the recovery of the receptor to a state available for a subsequent cycle of agonist binding.
The presence of muscarinic receptors of high and low affinity has been demonstrated by Birdsall et al. (22) in studies of direct binding of "H-agonists to synaptosomal preparations from rat cerebral cortex in which 25-3055 of receptors from that source existed in a high affinity form. Ehlert et al. (23,24) have studied the binding of a recently available tritiumlabeled muscarinic agonist, [cis-"H]methyldioxalane, to rat brain. Although levels of nonspecific binding were quite high, their data also suggested the presence of two classes of receptors of differing affinity. Several groups have demonstrated that exposure of cell homogenates to guanine nucleotides results in a decrease in the apparent affinity of the muscarinic receptor for agonist, consistent with conversion of RH to a low affinity form (7-9, 23).
The early and late events resulting from muscarinic agonistreceptor interaction may be divided into an initial rapid formation of a complex capable of stimulating a characteristic physiologic response, followed by a slow decrease in receptor number and/or affinity of receptor for agonist that would regulate the ability of subsequent agonist exposure to elicit a physiologic response. Such a dual role of agonist has been suggested for the insulin receptor, /3-adrenergic receptor, luteinizing hormone receptor, and others (1). Recent evidence presented by Stadel et al. (25) suggests that a long-lived agonist-high affinity receptor complex might be formed between the P-adrenergic agonist and receptor in turkey and frog erythrocyte membranes. They demonstrated that the addition of guanine nucleotide resulted in the dissociation of agonist from the receptor with concomitant activation of adenylate cyclase. DeLean et al. (26), using computer modeling for the binding of the P-adrenergic agonist ['H]hydroxybenzylisoproterenol to frog erythrocytes, demonstrated a good correlation of the data with a model in which a ternary complex is formed among agonist, recpetor, and a guanine nucleotide regulatory protein. The interaction of muscarinic agonist, high affinity receptors, and guanine nucleotides described in our studies suggests the following model, which contains interesting parallels with the scheme presented by DeLean (26) for the ,&adrenergic receptor. RHN + A+RHNA + G+(RHNG) + A+RL + (NG) (7) where N is the guanine nucleotide regulatory protein; A is the agonist; and G is a guanine nucleotide (GTP under physiological circumstances). The evidence for the existence of the species in parentheses is thus far indirect. In this scheme, in the absence of G, agonists binds persistently to R H N rendering these sites unavailable for binding of ['HIQNB. In the presence of G, the complex RHNA would interact with G with release of agonist and regeneration of the receptor in a low affinity form which would now be available for binding of ['HI QNB. Since a guanine nucleotide regulatory protein must be present in order for the receptor to interact with G (27), the absence of any significant effect of G on R1. suggests that N is associated with the high affinity state of the receptor. Hence, conversion of RH to RL in the presence of guanine nucleotides may be associated with the release of the guanine nucleotide regulatory protein. Conversely, regeneration of RH from Rl, may be associated with binding of a guanine nucleotide regulatory protein to the low affinity form of the receptor. Such an association might mediate conversion of RI, to RH: R L + N-RHN (8) Whether G is capable of interacting with RHN in the absence of A cannot be determined from these studies, since all assays of the effect of G on affinity of receptor for agonist were carried out in the presence of agonist.
The negative inotropic effect of muscarinic agonists in canine (28) and rabbit heart (29) has been shown to be due a t least in part to a GTP-and Na'-dependent inhibition of both basal and /?-adrenergic agonist-stimulated adenylate cyclase activity. An agonist-high affinity receptor complex, RHNA, might constitute an "activated" form of the receptor. One might speculate that in the presence of G, such a complex could mediate an inhibition of adenylate cyclase, perhaps through release of a species such as NG.
The second aspect of the studies reported here deals with the mechanisms by which prolonged agonist exposure decreases the number of receptors available for labeled antagonist binding. Unlike the rapid effects of agonist on the high affinity receptor associated with receptor activation, the slower effect of prolonged agonist exposure on receptor number may represent an agonist-mediated modulation of the ability of subsequent agonist binding to elicit a physiologic response, and as such would constitute an important biological control mechanism.
Muscarinic cholinergic stimulation of the heart causes decreases in the rate and force of contraction, and these changes are accompanied by an increase in K' permeability (16) and a decrease in the movement of Ca2+ into the cell via the Ca+ slow channel (18,30). We have demonstrated that prior exposure of cultured heart cells to agonist results in the loss of the ability of muscarinic agonist to decrease beating rate and to increase K' permeability as measured by altered efflux of 42K+ from the cells (20). The time course of this agonistmediated loss of physiologic response to muscarinic agonists corresponds closely to the agonist-mediated loss of [3H]MS binding sites shown in Fig. 10, with a 15-min lag period followed by loss of half the physiologic response after a 30min agonist exposure. Recently, Halvorsen and Nathanson (31) reported that agonist exposure of 6 h duration markedly decreased the ability of muscarinic agonists to mediate a decrease in beating rate in intact embryonic chick hearts 8 days in ouo. This effect was associated with the loss of high affinity receptors in heart homogenates. The close coupling between the loss of [''HIMS binding sites and the loss of physiologic response demonstrates that a modest change in receptor number is related to a comparable change in physiologic response. Such sensitivity of physiologic response to agonist exposure suggests that agonist-mediated receptor loss could be a sensitive mechanism for modulating the level of responsiveness of the heart to muscarinic stimuli.
Our data relating loss of muscarinic receptors during preexposure to various agonist concentrations ( Fig. 1) suggested that the population of receptors lost slowly during agonist exposure corresponded to a subclass of low affinity receptors. The concentration of carbamylcholine required to mediate a half-maximal loss of receptors over 3 h (Fig. 1) was 1.7 X M compared to a concentration of 1 X M for a halfmaximal decrease in rapidly lost receptors. Furthermore, since we have shown that brief agonist exposure results in the loss of high affinity receptors and leaves only low affinity receptors with an for carbamylcholine inhibition of ["HIQNB binding of 3.0 X M (Fig. 2), more prolonged exposure to agonist must involve the loss of a low affinity subset of receptors. Our previous studies demonstrated that a large fraction of agonistmediated receptor loss (46% of total receptor sites) was inhibited by agents that interfered with microtubule function (7), suggesting that disappearance of this subclass of receptors involved endocytosis. In the studies reported here, the slow loss of receptors could only be demonstrated in intact cells or in homogenates of intact cells that had been exposed to agonist prior to homogenization (Fig. 3). This was also consistent with the view that the slow phase of agonist-induced receptor loss involves endocytosis of receptors and hence requires functions which take place in the intact cell.
To pursue this issue further, we postulated that following agonist exposure, but prior to irreversible degradation of the receptor, receptors might assume an altered configuration involving a more intimate association with membrane lipids. Such a change in configuration might prevent subsequent agonist binding or interfere with the ability of agonists to mediate a physiologic response. Although a functional agonistreceptor complex might not form in cells pre-exposed to agonist, we further postulated that a muscarinic ligand which was sufficiently hydrophobic might bind to such an altered form of the receptor. Hence, comparison of the binding of radiolabeled ligands with markedly differing hydrophobicities might provide a means for studying the agonist-mediated transition in the state of the receptor.
Since ["HIMS contains one less phenyl group than ["HI QNB as well as an ether linkage, [:'H]MS should be less capable of interacting with the lipid bilayer. The relative solubility of substances in the lipid bilayer of the cell membrane may be approximated by comparison of partition coef-ficients between an aqueous phase at physiologic pH and ionic strength and immiscible organic solvents such as ether and chloroform. We have demonstrated the relative preference of ["HIQNB over ["HIMS for the organic phase in terms of partition with an aqueous phase. By these criteria, one might expect ["HIQNB to be substantially more soluble (and permeant) in the lipid bilayer of the cell membrane than ["]MS.
In addition to the difference in hydrophobicity, there are significant steric differences between ["HIMS and ["HIQNB. [3H]QNB has an extra phenyl group, while ["HIMS has two extra methyl groups on its bridged nitrogen (Fig.  11). We carried out two sets of studies to determine whether these differences in structure and hydrophobicity between ['HIQNB and ['HIMS could be used to study different states of the receptor. First, we compared the kinetics of binding of these two ligands to intact cultured heart cells. Second, we compared the binding of ["HIQNB and ['HIMS to intact cells which had been subject to prior exposure to high concentrations of agonist.
Studies comparing the kinetics of binding [:'H]QNB and ["HIMS to intact cells indicated that the binding of both ligands appears to be biphasic, with initial formation of a low affinity complex followed by a conversion to a relatively high affinity state. Burgisser et al. (32) have reported data for the binding of (&) '"I-hydroxybenzylpindolol to frog erythrocyte membranes which indicate that biphasic kinetics of binding and dissociation are due to the use of a racemic mixture of the optically active radiolabeled antagonists. However, both the ['HIQNB and ['HIMS used in these and our previous studies (7) are in the levo form. Hence, the biphasic association rates we observed for muscarinic antagonists require another explanation.
The kinetic rate constants for the initial binding step Q + kl R QR* yield similar values of kl and k-1 for ['HIQNB and k-I ["HIMS and hence similar dissociation constants (Table I). Acetylcholine, the physiologic muscarinic agonist, and carbamylcholine, the agonist used in these studies, are both positively charged and contain a carbonyl moiety in ester linkage to a choline moiety (Fig. 11). Given the proper juxtaposition, both the nitrogen of the choline moiety and the carbonyl appear to be necessary for agonist activity (33). The similarity of the relative positions of the carbonyl and nitrogen in ["HI QNB and ["HIMS might explain the similarity in their relative ability to recognize the active site of the receptor and hence to form the initial complex QR*. However, our kinetic data (Table I) indicate that the ['HIQNB-receptor complex ( Q R * ) initially formed undergoes transition to the state designated QR in Equation 3 at a substantially more rapid rate than is the case for ['HIMS. At the same time, the back reaction from QR to QR* is much slower for ["HIQNB than for ["HIMS, resulting in a k-z/k2 ratio roughly 100-fold more favorable to QR formation for [3H]QNB. If the steric requirements for formations of QR favored a less bulky ligand, then r3H]MS might bind more readily. Since this is not the case, our tentative interpretation is that differences in hydrophobicity between the two antagonist radioligands are responsible for the differing kinetics of ["HIMS and ['HIQNB binding to intact cells.
Having established that a form of the receptor existed which could bind ['HIQNB more readily than ['HIMS, we compared the binding of r3H]MS and r3H]QNB to cells which had been subject to prior exposure to high concentrations of agonists for various times. Prolonged exposure to agonist mediated conversion of the receptor from a form (designated A , Equation 4) which binds ["HIMS and ['HIQNB to a form (designated B, Equation 4) which binds [3H]QNB but which does not bind r3H]MS, suggesting that the configuration of the receptor and/or relationship of the receptor to the plasmalemma is altered. We observed a close temporal correlation between loss of [3H]MS binding sites in intact cells and loss of physiologic response to agonist (20). This finding is consistent with the hypothesis that accessibility of the receptor in the intact cell to [3H]MS is correlated with accessibility of the receptor to agonist. Alternatively, r3H]MS may bind preferentially to a functioning state of the receptor. Hence, the receptors in form B do not appear to mediate a physiologic response (20). More prolonged agonist exposure resulted in conversion to a form C, presumably a degraded form of the receptor that binds neither r3H]MS nor ['HIQNB. Hence, form B of the receptor may constitute an intermediate state of the receptor formed during the process of endocytosis. It should be noted that the experiments described cannot distinguish between differences in r3H]QNB and ['HIMS binding to an altered configuration of the receptor within the membrane ( i e . plasmalemma) and differences in access to receptor sites that have been endocytosed but not yet degraded.
It is intriguing to speculate that the agonist-mediated change in the receptor described as A+ B (Equation 4) from a form A which binds both ['HIQNB and ['HIMS to a form B which binds only r3H]QNB might represent a transition similar to that described in the process QR * QR (Equation

3) in which both [3H]QNB and r3H]MS bind with equal
affinity to the complex QR *, but the transition to a rearranged form of the receptor-antagonist complex QR is markedly more favorable for r3H]QNB. Both the transition designated A-B and QR *=QR involve rearrangement to a state which binds r3H]MS less well. Both could represent rearrangement to a more hydrophobic domain and both transitions are induced by ligand binding. Although endocytosis of receptors will only take place in the intact cell, it has not been determined whether the conversion of A+B requires an intact cell. Our previous studies with cell homogenates (6) and recent work with a solubilized form of the receptor (34) has demonstrated that the transition QR*=QR does take place in a cell-free system.
Because in most receptor studies the binding of antagonists is only very slowly reversible, the effect of prior exposure to antagonists on subsequent antagonist binding has not been kl kz E 2 determined. Hence, it is not known whether antagonist binding can modulate receptors number in a manner similar to the effect of agonist exposure. It seems possible that the second step in ligand binding could constitute an antagonist-mediated rearrangement similar to the change in the state of the receptor that occurs during prolonged agonist exposure. Hence, one might speculate further that exposure to any specific ligand, antagonist or agonist, is capable of mediating a change in receptor conformation and that the state QR in Equation 3 is similar to the state B in Equation 4.
In conclusion, we have used radioligand binding techniques to demonstrate a response of the muscarinic receptor to occupancy by agonist that is biphasic with respect to both time and agonist concentration. Subsets of receptors were further delineated by kinetic analysis of binding of ["HIQNB and r3H]MS to intact, beating heart cells. Our findings provide further evidence for a dual role of the muscarinic agonistreceptor interaction as mediator of immediate cellular biochemical and physiological responses and also as modulator of responsiveness of the cell to subsequent or continued agonist exposure.