Characterization of Muscarinic Cholinergic Receptors on Rat Pancreatic Acini by A^­[^H]Methylscopolamine Binding THEIR RELATIONSHIP WITH CALCIUM 45 EFFLUX AND AMYLASE SECRETION*

iV­[^H]MethyIscopolamine (NMS) binding, amylase sécrétion, and '*^Ca efflux from dispersed rat pan­ creatic acini were investigated in parallel, in the prés­ ence or absence of 4 muscarinic agonists and 3 mus­ carinic antagonists. Scatchard analysis of [^H]NMS saturation isotherms gave a Ko of 0.9 nM and an av­ erage binding capacity of 24,000 sites per ceU. Binding compétition curves with the antagonists atropine, dex­ etimide, and NMS gave ifx) values of 3.5, 3.5, and 0.5 nM, respectively. With the 3 full agonists oxotremo­ rine, muscarine, and carbamylcholine, the receptor population could be divided into two classes of binding sites: a minor one (15%) with high affinity (Ko = 2 0­ 35 UM ) and a major one (85%) with low affinity (Ko = 3—65 iiM). There was a receptor reserve of about 50%i with respect to carbamylcholine­stimulated amylase sécrétion. Further analysis of dose­effect curves sug­ gests that low affinity binding sites were involved in the secretory response to muscarinic stimulation. Pilo­ carpine, like muscarinic antagonists, recognized ail binding sites with the same affinity but acted as a partial agonist on amylase sécrétion and pancreatic were the incubation solution with 1 nM [^H]NMS. At the end of the incubation, 200 -^â samples were taken in duplicate and layered above 100 ^1 of di-n-butylphthalate. The acini were sedimented through the dense layer by a 15-s centrifugation in a Beckman 152 microfuge. The microfuge tube was frozen in liquid nitrogen and eut through the dense layer. The acini in the lower part were digested at 70 °C for 3 h with 1 ml of Lumasolve (Lumac, Schaesberg, The Netherlands). After adding 100 /jl of 6 M HCl to decrease chemilu- minescence, 8 ml of Aquasolve 2 (New England Nuclear, Dreieich, F.R.G.) were added, and the radioactivity was counted in a 7500 Beckman liquid scintillation spectrometer. The results were corrected for the volume of incubation médium accompanying the acini through the dense layer. This was estimated by incubating acini in the présence of 1 mM ['H]sucrose (10 mCi/mol). The volume of extracel- lular médium trapped under thèse conditions was 0.36 ± 0.08 ^1 out of a total initial volume of 200 ^1 and was not modified in the présence of muscarinic agonists and antagonists.

one récent characterization of muscarinic receptors was achieved in pancreatic acini with [^HJquinuclidinyl benzilate binding (2), despite the inconvenience of high nonspecific labeling (1) and low dissociation rate (2) observed with this radioligand. In the présent study, the binding to rat pancreatic acini of another muscarinic antagonist, Af[^H]methylscopo lamine, with much less nonspecific binding (3), was examined in the présence of 3 muscarinic agonists, of the partial agonist pilocarpine, and of 3 muscarinic antagonists. A comparison of muscarinic agonist binding with amylase release suggests that low affinity binding sites were largely if not exclusively involved in this biological effect.
Assay of PH}NMS Bindin g -In routine assays, 0.5ml aliquots of ' The abbreviations used are: Hepes, 4(2hydroxyethyl)lpipera zineethanesulfonic, acid; NMS, A^methylscopolamine; PrBCM, pro pylbenzilylcholine mustard; QNB, 3quinuclidinyl benzilate; ECso, concentration exerting halfmaximal effect; ICso, concentration ex erting halfmaximal inhibition; Lso, equilibrium concentration of free cholinergic drug at 50% saturation of its binding to receptor (3); pAa, négative logarithm to base 10 of the molar concentration of an antagonistic drug which will reduce the effect of a double dose of an agonist to that of a single dose (27); df, degree of freedom. pancreatic acini (0.4 to 0.5 mg of protein) were incubated at 37 °C in the standard incubation solution with 1 nM [^H]NMS. At the end of the incubation, 200-^â samples were taken in duplicate and layered above 100 ^1 of di-n-butylphthalate. The acini were sedimented through the dense layer by a 15-s centrifugation in a Beckman 152 microfuge. The microfuge tube was frozen in liquid nitrogen and eut through the dense layer. The acini in the lower part were digested at 70 °C for 3 h with 1 ml of Lumasolve (Lumac, Schaesberg, The Netherlands). After adding 100 /jl of 6 M HCl to decrease chemiluminescence, 8 ml of Aquasolve 2 (New England Nuclear, Dreieich, F.R.G.) were added, and the radioactivity was counted in a 7500 Beckman liquid scintillation spectrometer. The results were corrected for the volume of incubation médium accompanying the acini through the dense layer. This was estimated by incubating acini in the présence of 1 mM ['H]sucrose (10 mCi/mol). The volume of extracellular médium trapped under thèse conditions was 0.36 ± 0.08 ^1 out of a total initial volume of 200 ^1 and was not modified in the présence of muscarinic agonists and antagonists.
Nonspecific binding was defined as tracer binding in the présence of 10 MM atropine. This antagonist concentration also prevented agonist-stimulated amylase output and ''^Ca efflux without affecting basai activities. Similar inhibitory effects on the three parameters were in fact exerted by atropine tested in the 10"^ to 10"^ M concentration range (not shown).
Dissociation Kinetics-Rat pancreatic acini (5-6 mg of acinar protein) were preincubated for 1 h at 37 °C in 6 ml of standard médium in the présence of 3 nM [^H] NMS and in the présence or absence of 10 fiM atropine. Aliquots containing 60 fi\ of water (control) or atropine (allowing a final 10 iiM concentration of the unlabeled antagonist) were then added. At appropriate time intervais, duplicate 200-^1 samples were treated as described above. Results were expressed as per cent of control binding.
Scatchard Plots-Rat pancreatic acini were incubated as indicated above in the présence of increasing concentrations of [^H]NMS (0.1 to 10 nM) and in the présence or absence of 10 atropine. Spécifie binding was measured at 10, 20, 30, 60, 90, and 120 min. The saturation isotherms at each time were analyzed according to Scatchard (5).
Amylase Sécrétion-At the end of the incubation period (1 h usually), an aliquot of the suspended acini was centrifuged for 15 s in a Beckman 152 microfuge. a-Amylase activity was determined in the supernatant (6). Controls taken at the beginning of each incubation period, for determining the amylase content of the médium at time zéro, allowed appropriate corrections.
Estimation of the Degree of Receptor Reserve with Respect to Carbamylcholine Stimulation-Ariëns et al. (7) have developed a method for measuring the receptor reserve of muscarinic receptors in isolated organs that is based on the blockade of an increasing number of functional receptors with an irréversible antagonist. This method was applied to rat pancreatic acini preincubated for 15 min with various concentrations (0-100 nM) of propylbenzilylcholine mustard, an irréversible alkylating agent (8). At the end of the preincubation, the acini were centrifuged for 1 min at 200 x g and resuspended in fresh incubation médium in the absence of PrBCM.
[^H]NMS spécifie binding and amylase release with various concentrations of carbamylcholine were estimated after 60 min as previously described. The fraction q of receptors remaining free after preincubation was derived 1) frora direct binding data and 2) from a comparison of the control secretory curve of carbamylcholine with that obtained after preincubation with PrBCM. The fraction of receptors remaining free and the affinity of the stimulatory agent for receptors were derived frora the équation of Furchgott and Bursztyn (9), A q A' q Ko where A and A ' are pairs of carbamylcholine concentrations giving the same secretory response before and after PrBCM treatment, q is the fraction of free muscarinic receptors, and Ko the dissociation constant of carbamylcholine for receptors involved in sécrétion.
Study of Négative Cooperativity Interactions-Négative cooperativity interactions were tested as described by Birdsall et al. (10). Briefly, when the binding of an agonist to a receptor is characterized by flattened displacement curves and a nn < 1, négative cooperativity interactions can be tested by blocking an increasing fraction of thèse receptors. Such interactions will then be weakened with a steepening of the displacement curve and a nn tending to 1. In practice, acini were preincubated with various concentrations (0-100 nM) of PrBCM for 15 min at 37° C, washed, and resuspended in fresh incubation médium.
[^H]NMS binding was then performed in the présence of various concentrations of carbamylcholine.
'^Calcium Efflux-''^Ca was included (2.5 ixCi/m\) in the média used for preparing and washing the acini (4), i.e. for a 75-min period. When testing the muscarinic agonists, the acini were resuspended in a final unlabeled incubation médium and incubated for 5 min with the agonists. When testing the inhibitory effects of muscarinic antagonists on carbamylcholine-stimulated '"'Ca efflux, the preloaded acini were preincubated for 1 h with ''^Ca in the présence of the antagonists and then incubated for 5 min in an unlabeled incubation médium with the antagonists and 3.2 ^M carbamylcholine. This preincubation did not modify the total ''^Ca content of the acini. After the incubation, an aliquot of the suspension was centrifuged in a Beckman 152 centrifuge for 15 s. The radioactivity in the supernatant was counted. Controls taken at the beginning of each incubation period, for determining the '''Ca released into the médium at time zéro, allowed appropriate corrections. The incubation period was limited to 5 min, ••'Ca efflux remaining linear during that period, indicating no significant reuptake of the released ''^Ca.
Protein Détermination-Protein détermination was performed according to Lowry et al. (11) using bovine sérum albumin as a standard.
Analysis of the Data-Dose-effect curves of inhibition of [^H]NMS binding by muscarinic agonists were analyzed by computer fitting to a model of two classes of binding sites (12).
The goodness of fit of the two-site model was tested by comparison with a one-site model, using the method of Munson and Rodbard (13). Expected Y values were computed manually with a TI-59 using the ISIS-59 program developed by Thakur et al. (14) for a two-site model and a simple linear régression for the one-site model. Residual sums of unweighted squares of déviations of the expérimental points to the fitted curves were used for the F test.  ligand concentration in the 0.1-10 nM [^H]NMS concentration range ( Fig. 2A).

After 1 h of incubation, [^H]NMS bound with an apparent
KD of 0.9 ± 0.2 nM (n = 5) and with longer incubation periods no further decrease of the Ko value was observed (not shown), suggesting true binding equilibrium after 1 h. There was no évidence for heterogeneity among muscarinic antagonist binding sites, and the receptor concentration was 140 ± 15 fmol/ mg of protein (Fig. 2B).
To examine the reversibility of the binding reaction, acini were preincubated with 3 nM pH]NMS for 1 h at 37 °C, then incubated in the présence of 10 ^M atropine (Fig. 3) Muscarinic Antagonists-Dexetimide was at least 1000 times more potent than levitimide (the inactive stereoisomer) in displacing [^H]NMS. Compétition curves with unlabeled atropine, dexetimide, and NMS, and the derived values of the Hill coefficient indicated that thèse antagonists bound to a single class of receptors ( Fig. 4 and Table I). When correcting for tracer concentration (15), the following KD values were obtained: 3.5 nM for atropine, 3.5 nM for dexetimide, and 0.5 nM for NMS (Table I).
Thèse antagonists were also tested for their capacity to inhibit amylase sécrétion and ""^Ca efflux stimulated by 3.2 /iM carbamylcholine. In order to attain equilibrium, acini were first preincubated for 1 h in the présence of the antagonists and under conditions identical with those used for the binding assays. They were then incubated in the combined présence   (15) and are the means ± S.E. of 4 experiments.   Table I). Levitimide at 1 nM was without significant effect (Fig. 5A).

Muscarinic Agonists-Oxotremorine, muscarine, and carbamylcholine were tested for their ability to inhibit [^H]NMS binding and to stimulate amylase sécrétion during 1-h incubations. The dose-effect curves of inhibition of [^H]NMS
binding by thèse three full agonists showed that maximally effective concentrations were approximately 1000 times greater than the threshold concentrations (Fig. 4), the corresponding Hill coefficients being in the 0.6 to 0.9 range (Fig. 4 and Table II Table II. The proportion of high affinity binding sites was 14% of the total number of binding sites for oxotremorine {F = 4.82; df 2,8; p < 0.05), 15% for muscarine (F = 8.54; df 2,7; p < 0.05), and 16% for carbamylcholine {F = 7.11; df 2,12; p < 0.01). By contrast, pilocarpine recognized only one class of sites.
Dose-effect curves describing the stimulation of amylase release caused by oxotremorine, muscarine, and carbamylcholine were parallel and similar in that, as the concentration of thèse agents was increased, amylase release increased, became maximal, and then decreased. Thèse 3 agonists had the same efficacy in terms of maximal amylase sécrétion (Fig. 6A). The potency to stimulate amylase release (expressed as ECso) decreased in the following order: oxotremorine 0.2 fiM, muscarine 0.6 fiM, and carbamylcholine 1 ixM (Fig. 6A and Table  II).
The potency of the 3 full muscarinic agonists to stimulate """Ca efflux, after a 5-min incubation period, was the following when expressed as EC5o:oxotremorine 0.6 ixM, muscarine 1.5 ^tM, and carbamylcholine 2 nM (Fig. 6B and Table II).
The dose-effect curve of inhibition of [^H]NMS binding by the partial agonist pilocarpine (10) was steep as reflected by a Hill coefficient of 0.97 ( Fig. 4 and Table II) and that describing pilocarpine stimulation of amylase release was shallower than the dose-effect curves of full agonists, with a lower maximal release (Fig. GA). Finally, pilocarpine exerted only a limited effect on "^Ca efflux (Fig. 6B). Additional experiments with 10 and 100 nM pilocarpine demonstrated the capacity of this partial agonist to inhibit competitively carbamylcholine-stimulated '''Ca efflux (Fig. 7) with a Ki of 3.6 / Li M suggesting that pilocarpine and carbamylcholine were acting on the same receptors.

The Existence of a Receptor Reserve for Carbamylcholine-
The existence of a receptor's reserve for carbamylcholine was tested using the technique of Ariëns et al. (7) in an effort to relate amylase sécrétion to a given type of binding sites. Preincubating acini with 18 nM PrBCM for 15 min displaced the dose-response curve for carbamylcholine to the right without affecting the maximal response (Fig. 8A) Fig. 8B as -In q versus the concentration of PrBCM; the two séries of values yielded straight lines and were comparable within 10%. (Table III)  is the equiUbrium concentration of free cholinergic drug at 50% saturation of its binding to receptor, calculated from compétition date, according to Burgermeister et al. (3). 'The Ko for high and low affïnity sites were calculated according to Minneman et al. (12), then corrected according to Cheng and Prusoff (15). Values are expressed as percentage S of maximal increase in amylase sécrétion due to an optimal dose of agonist S",x. Basai amylase release was 3.2% of total content-60 min"\ and maximal stimulated amylase sécrétion ws 12.5% of total content-60 min~'. S, rat pancreatic acini were labeled with ^'Ca during their isolation (see under "Methods").

This shows that PrBCM acted at the receptor level (16, 17). Besides, PrBCM did not selectively block one population of sites since a PrBCM treatment did not modify the ICso and «// values of the dose-effect curves of inhibition of [^H]NMS by carbamylcholine
•"^Ca efflux was mesured after a 5-min incubation period at 37 °C in the présence of the indicated concentrations of the 4 agonists (see A). Values are expressed as a percentage E of maximal increase in ••^Ca efflux due to an optimal dose of agonist Basai '''Ca efflux was 9.8% of total initial ''^Ca content-5 min"'. Maximal *^Ca efflux was 36.0% of total initial "Ca content-5 min"'. Each value was determined in duplicate. This experiment is représentative of 3 to 5 experiments.
binding. It is recognized that, as in brain, such linear Scatchard plots obtained at equilibrium with rat acini might correspond to complex association and dissociation kinetics {e.g. a fast binding step foUowed by slow isomerization of the receptor-ligand complex (18)). The Ko value of 0.9 nM (Fig.  2 (Table I)

['H]QNB (see above and Réf. 2).
Inhibition of [^H]NMS binding by atropine and dexetimide was comparable, with a Hill coefficient close to 1, suggesting that muscarinic antagonists bound to a single class of receptors. The KD of NMS, atropine, and dexetimide were in the nanomolar range. In many tissues, the affinity of muscarinic binding sites for antagonists is high (19) in contrast to the much lower affinity for agonists (Table II and Réf. 20).
The relationship between antagonist binding and the inhibition of carbamylcholine-stimulated amylase release and "Ca efflux (Table I) suggests that [^H]NMS binding sites were involved in biological responses, since the order of potency of the antagonists for binding and for inhibiting stimulated amylase sécrétion and ''^Ca efflux was the same (Table  I).
Compétition curves of [^H]NMS with agonists developed on more than two logarithms (Fig. 4). Similar flattened agonist/[^H]antagonist compétition curves have been documented for muscarinic receptors in membranes from brain, heart, smooth muscle, and cloned neuroblastoma cells, as well as in intact smooth muscle cells from rat and guinea pig ileum and in fragments of rat brain (10, 21).
Thèse flattened compétition curves could be explained by  Effect of a pretreatment of rat pancreatic acini with propylbenzilykholine mustard on dose-effect curves of inhibition of f^HJNMS binding by carbamylcholine Acini were preincubated for 15 min with PrBCM (0-56 nM), then centrifuged, resuspended in fresh médium, and incubated in the présence of 1 nM [^H]NMS and various concentrations of carbamylcholine for 1 h (see under "Methods"). The proportions of high and low affinity binding sites were estimated according to Minneman et al. (12). This experiment is représentative of four others. Second, a subheterogeneity among muscarinic receptors might reflect bimolecular dissociation or two-step interactions with resulting receptor heterogeneity (23). However, data generated by thèse 2 models are also fitted by curves describing apparent négative cooperativity (23), so that this mechanism is also unlikely to account for the flattened displacement curves by agonists.
Third, 2 or more classes of receptors coexist, with structural différences at the level of binding sites, that show distinct affinities for agonists and a similar high affinity for antagonists. To consider the last model more quantitatively, we analyzed agonist [^H]NMS compétition curves according to a "two classes of binding sites" model (12) in order to characterize the subtyping. The data were compatible with the présence of about 15% of high affinity binding sites and about 85% of low affinity sites for oxotremorine, carbamylcholine, and muscarine corresponding to approximately 3,500 and 20,500 sites per cell, respectively, on an average. Pancreatic acini contain no more than 10% of centro-acinar and duct cells and a large majority of amylase-secreting acinar cells (24). On this ground, it cannot be excluded that the two classes of muscarinic binding sites were located on différent cell types. A minority of high affinity muscarinic binding sites and a larger proportion of low affinity muscarinic binding sites are similarly présent in brain and smooth muscle (10).
To delineate the biochemical significance of the binding data, it is of interest to note that dose-effect curves of agonist stimulation of amylase release developed within a narrow concentration range (Fig. 6), suggesting that each agonist interacted with no more than one class of muscarinic binding sites.
The EC50 of carbamylcholine for stimulation of amylase sécrétion was 1 /xM, while the [L50] of its binding to muscarinic receptors was 54 ^JM (Table II). This high efficacy of carbamylcholine suggested the présence of a receptor reserve.
To approach this problem, we first noticed that the pretreatment of acini with PrBCM was unable to modify the proportions of high and low affinity binding sites (Table III) and that the estimation of the total number of receptors blocked by PrBCM was similar when tested by [^H]NMS binding (Table III) and by the method of Furchgott and Bursztyn (9) as applied to amylase sécrétion (Fig. 8). This parallelism indicated that PrBCM could be used as an irréversible blocker of muscarinic receptors having no further action on effector mechanism(s) (16). Maximal carbamylcholine response remained possible even when 47% of the receptors were blocked by PrBCM but the efficacy of carbamylcholine started to decrease when 59% of the receptors were not available anymore (Fig. 8A). From 47 to 59% of the muscarinic receptors were, therefore, reserve receptors with respect to carbamylcholine. The "true" affinity constant KA of the carbamylcholine effect on amylase sécrétion was estimated to be 10 iiM when plotting 50% values of maximal effects (Fig. 8^1) against the log of carbamylcholine concentration (25). Besides, the Ko of carbamylcholine for receptors involved in sécrétion was estimated according to Furchgott and Bursztyn (9) (see under "Methods") using secretory curves under control conditions and after the blockade of various fractions of receptors. This Ko was found to be 8-14 nU. Since the Kp of carbamylcholine for high affinity and low affinity binding sites were, respectively, 35 nM and 65 (Table II) the sécrétion of amylase in response to carbachol correlated with the occupancy of low affinity sites, although the participation of high affinity sites cannot be excluded. Similarly, Birdsall et al. (10) and Halvorsen and Nathanson (26) have reported that smooth muscle and heart responsiveness to muscarinic agonists was best correlated with the occupancy of low affinity receptor sites.
The ECso values for pilocarpine binding and resulting amylase sécrétion were similar (Table II), suggesting that this agent reacted with a single class of binding sites among which there was no receptor reserve. This is in line with the current hypothesis that the degree of spareness is a function of the ligand bound to the receptor (17). As discussed before, be-tween 47 and 59% of the receptors were spare receptors when carbamylcholine was used; with respect to pilocarpine, there was no spareness and, furthermore, pilocarpine was unable to fully stimulate amylase sécrétion (Fig. 6A) and had hardly any effect on '"'Ca efflux (Fig. 6B). This ligand was capable, however, to inhibit competitively the carbamylcholine stimulation of ''^Ca efflux, showing, therefore, binding to genuine muscarinic receptors but with no efficient coupling to effector mechanism(s). Thus, pilocarpine met the theoretical requirements for being a partial agonist (7).