Abstract
The hypothesis was tested that M2-selective antagonists partially utilize the allosteric site of muscarinic M2receptors. The interactions of the allosteric agent W84 (hexane-1,6-bis[dimethyl-3′-phthalimidopropyl-ammonium bromide]) were studied with the M2/M4-selective AF-DX 384 [(±)-5,11-dihydro-11-{[(2-{2-[(dipropylamino)methyl]-1-piperidinyl}ethyl)amino]carbonyl}-6H-pyrido(2,3-b)(1,4)-benzodiazepine-6-one], the nonselective N-methylscopolamine (NMS), and a number of other muscarinic antagonists. In isolated paced guinea pig atria, the antagonistic effect of W84 against oxotremorine- and arecaidine propargyl ester-induced negative inotropic actions reached a limiting value at higher W84 concentrations, revealing negative cooperativity (factors of cooperativity α = 311 and α = 495, respectively). The antagonistic potency of W84 in this M2receptor model (W84 binding constantK A ∼ 160 nm) was higher than at M1/M4-like receptors of rabbit vas deferens (K B ∼800 nm) and at M3 receptors of guinea pig ileum (K B ∼4,000 nm). In paced atria, combinations of W84 with muscarinic antagonists yielded more-than-additive antagonistic effects against oxotremorine in case of conventional antagonists such as NMS (α = 18) but less-than-additive effects with the M2-preferring AF-DX 384 (α = 444). In guinea pig heart homogenates, the equilibrium binding of [3H]NMS was only partially inhibited by W84 (α = 2.4), whereas [3H]AF-DX 384 binding could be suppressed completely (α = 194). The difference in cooperativity reflects that W84 inhibits [3H]NMS dissociation with a ∼40-fold higher potency (ECdiss = 900 nm) than [3H]AF-DX 384 dissociation (ECdiss = 33,300 nm). [3H]NMS dissociation also could be retarded by AF-DX 384 (ECdiss = 22,000 nm), probably via an interaction with the site used by W84. The results suggest that the binding domain of AF-DX 384 partially overlaps with the common allosteric site of the M2 receptor protein.
Conventional muscarinic antagonists are not subtype selective. The orthosteric antagonist binding site seems highly conserved among the five muscarinic receptor subtypes. The orthosteric site is thought to be located within a ligand binding pocket formed by the transmembranous helices (Wheatley et al., 1988; Trumpp-Kallmeyer et al., 1992; Wess, 1993).
M2-selective antagonists utilize different points of attachment on M2 receptors (Birdsall et al., 1989). It has been suggested that extramembranous receptor domains may be involved in the subtype-selective binding of M2 antagonists (Melchiorre et al., 1989; Pedder et al., 1991; Wess et al., 1992;Kerckhoff and Höltje, 1994).
Allosteric modulators of ligand binding to muscarinic receptors generally have a high affinity to M2 receptors (Ellis et al., 1991; Lee and El-Fakahany, 1991;Jakubı́k et al., 1995). These drugs are capable of binding to muscarinic receptors even if the orthosteric site is occupied by a ligand such as the antagonist NMS. As a consequence, the dissociation of the orthosteric ligand is altered by the allosteric modulator; commonly, ligand dissociation is retarded. When the allosteric modulator has bound to the free receptor, the association of the ligand is impaired. The effect of an allosteric modulator on the equilibrium binding of the ligand depends on the balance between the actions on ligand association and dissociation (Kostenis and Mohr, 1996). Functionally, the known allosteric modulators behave as antagonists. Results of biochemical (Pedder et al., 1991), mutagenesis (Ellis et al., 1993; Leppik et al., 1994), and chemical modification (Jakubı́k and Tuc̆ek, 1995) studies suggest that allosteric modulators bind at the entrance of the ligand binding pocket of the M2 receptor.
Because M2-selective antagonists and allosteric modulators share a similar subtype-selectivity pattern and are likely to interact with the receptor at the entrance of the ligand binding pocket, the question arises whether M2-selective antagonists may use points of attachment that are part of the common allosteric site (Ellis and Seidenberg, 1992) of muscarinic M2 receptors.
To check for such an interplay, we investigated in functional and radioligand binding experiments the interaction between the allosteric model compound W84 (Fig. 1) and the M2/M4-selective antagonist AF-DX 384 [(±)-5,11-dihydro-11-{[(2{2-[(dipropylamino)methyl]-1-piperidinyl}ethyl)amino]carbonyl}-6H-pyrido(2,3-b)(1,4)benzodiazepine-6-one] (Fig. 1). For the sake of comparison, the nonselective antagonist NMS was included with a number of structurally related and unrelated antagonists. W84 was chosen as allosteric model compound because, first, it has been shown to interact with the common allosteric site on the M2 receptor protein (Tränkle and Mohr, 1997). Second, W84 has a rather high affinity to the allosteric site in antagonist M2 receptor complexes even under organ bath conditions (Jepsen et al., 1988; Lüß and Mohr, 1992). Third, structure-activity relationships for the stabilizing effect of W84 on NMS-occupied M2 receptors have been characterized (Kostenis et al., 1994). Forth, in isolated beating guinea pig atria, combinations of W84 with atropine (Lüllmann et al., 1969) and NMS (Maaβ et al., 1995) have been shown to induce more-than-additive antimuscarinic effects.
AF-DX 384 was chosen as a representative M2-preferring antagonist (Miller et al., 1991) because it is available in a radiolabeled form, allowing for direct binding studies.
The test compounds were applied in isolated beating guinea pig left atria to study interactions under “physiological conditions” as well as in radioligand binding experiments with guinea pig cardiac membranes to gain a more direct insight into the events on the molecular level.
Materials and Methods
Experiments in Isolated Organs
Preparation of guinea pig left atria was carried out as described previously (Maaß et al., 1995). The atria were mounted in 20-ml organ baths filled with Tyrode’s solution (149.2 mm Na+, 2.7 mmK+, 1.8 mmCa2+, 1.1 mmMg2+, 145.5 mmCl−, 12.0 mmHCO3 −, 0.2 mmH2PO4 −, 5.5 mm glucose, pH 7.3), which was maintained at 32° and oxygenated with 95% O2/5% CO2. Atria were preloaded with 10 mnand electrically stimulated via platinum contact electrodes with rectangular pulses of 5-msec duration at a frequency of 3 Hz. Isometric force of contraction after an initial equilibration period of 60-min duration in the absence of any drug was set at 100%. Cumulative concentration-effect curves for the negative inotropic effect of oxotremorine were recorded with each concentration present for 10 min. The oxotremorine concentration at which the force of contraction was reduced by 50% was used as a measure of potency (EC50). After a wash-out period of 30 min, including three changes of the incubation medium after 0, 10, and 20 min, the atria were incubated with an antagonistic test compound for 60 min before the next concentration-effect curve of oxotremorine was recorded. After another washing period of 30 min to remove the agonist, a combination of the antagonistic test compound and another antagonist of interest was applied for 60 min, and the concentration-effect curve of oxotremorine was determined again. Control experiments revealed that the order in which W84 and the respective antagonists were applied did not influence the effect of the combination. To characterize the concentration dependency of the antioxotremorine effect of a test compound, the same schedule was applied except a higher concentration of antagonist was applied instead of a combination.
A set of experiments was carried out to investigate the muscarinic receptor subtype selectivity of W84; thus, the pharmacological properties of W84 were examined at prejunctional muscarinic heteroreceptors in rabbit vas deferens (M1/M4-like receptors; 4-Cl-McN-A-343 as agonist) and in guinea pig atria (M2 receptors; APE as agonist) and ileal longitudinal smooth muscle (M3 receptors; APE as agonist). Tissues were isolated from adult guinea pigs of either sex or from male New Zealand White rabbits previously killed by cervical dislocation or intravenous injection of 120 mg/kg pentobarbitone sodium, respectively. The methods used have been described in detail previously (Waelbroeck et al., 1994, 1996).
Binding Studies
Preparation of guinea pig cardiac membranes.
Cardiac membranes were prepared as described previously at an ambient temperature of 3–6° (Jepsen et al., 1988). Briefly, pieces of ventricular myocardium of guinea pig hearts were homogenized in a 0.32 m sucrose solution. The homogenate was centrifuged for 10 min at 2,500 × g (5,000 rpm in a Beckman rotor 21; Beckman Instruments, Columbia, MD). The supernatant was centrifuged for 30 min at 77,200 × g (31,500 rpm in a Beckman rotor 35). The resulting pellet was resuspended in 50 mm Tris·HCl, pH 7.4. Aliquots of 1 ml were frozen in liquid nitrogen and stored at −80°. Protein content amounted to 4.9–7.3 mg/ml membrane suspension.
Binding assays.
[3H]NMS and [3H]AF-DX 384 had specific activities of 85.1 and 90.4 Ci/mmol, respectively. Cardiac membranes at a protein concentration of 300–400 μg/ml were incubated with either 0.5 nm [3H]NMS or 2 nm[3H]AF-DX 384 in a final volume of 1.5 ml. Experiments were performed in a buffer composed of 50 mmTris·HCl, and 3 mm MgHPO4, pH 7.3, at 23°. Nonspecific 3H-radioligand binding was determined in the presence of 100 μm atropine and did not exceed 10% of [3H]NMS total binding or 25% of [3H]AF-DX 384 total binding. The binding characteristics of [3H]NMS and [3H]AF-DX 384 under control conditions were investigated in homologous competition experiments with 2 hr of incubation. The KD andB max values were 0.9 nm and 233 fmol/mg of protein for [3H]NMS binding and 11 nmand 320 fmol/mg of protein for [3H]AF-DX 384 binding. These results are in good agreement with theKD value of 8.7 nm reported by Entzeroth and Mayer (1990) in rat cardiac homogenates for [3H]AF-DX 384 binding. The effect of increasing concentrations of W84 on radioligand equilibrium binding was measured with 2 nm[3H]AF-DX 384 after 3 hr of incubation and with 0.2 nm [3H]NMS after 4 hr of incubation.
To measure drug effects on the dissociation of [3H]NMS or [3H]AF-DX 384, the respective radioligand and membranes were incubated for 120 min before the dissociation was revealed by the addition of 100 μm atropine alone or in combination with the indicated test compound. Radioligand dissociation was observed generally over 120 min. When [3H]NMS dissociation was measured in the presence of combinations of AF-DX 384 and W84, the observation period was extended to 4 hr. Membranes were separated by rapid filtration (glass fiber filters No. 6; Schleicher & Schüll, Dassel, Germany). Filters were washed twice with 5 ml of ice-cold incubation buffer, dried, and placed into scintillation vials containing 5 ml of Ready Protein (Beckman) before the filter-bound radioactivity was determined by liquid scintillation counting.
Data Analysis
Experiments with isolated organs.
Agonist concentration-effect curves were fitted by nonlinear regression analysis using the general Hill equation, and the concentration for a half-maximum effect was determined (EC50). DRs served to quantify antagonist-induced curve shifts: DR = EC50,test compound/EC50,control. The concentration dependency of the effect of a given antagonist was analyzed according to Arunlakshana and Schild (1959) or Lanzafameet al. (1996). The latter analysis is appropriate to describe antagonistic effects in terms of the ternary allosteric model of Ehlert (1988) and is based on the equation:
The expected DR (DRexp) for a combination of antagonists can be calculated as DRexp = DRant1 + DRant2 − 1, if the interaction is competitive (Lüllmann et al., 1969;Clark and Mitchelson, 1976). More-than-additive or less-than-additive effects as observed with W84 in combinations with the antagonists were analyzed according to Christopoulos and Mitchelson (1994) using the equation:
Binding data.
Experimental results were analyzed by computer-aided, nonlinear regression analysis using Prism Version 2.01 (GraphPAD Software, San Diego, CA). Curve fitting to competition data was based on the general Hill equation. Because the observed Hill coefficients in homologous competition experiments did not differ significantly from unity (partial F test, p> 0.05, data not shown), IC50 values were determined from curve fits with nH fixed to 1. KD andB max values were calculated from these IC50 values according to DeBlasi et al. (1989).
The inhibition curves for the effect of W84 on the binding of [3H]AF-DX 384 and [3H]NMS, respectively, were analyzed according to Ehlert (1988) using the equation:
Dissociation data were fitted using a monoexponential decay function that yielded the apparent rate constant of dissociation,k −1. Curve fitting to obtain concentration-effect curves for the retardation of radioligand dissociation was based on a four-parameter logistic function.
The antagonistic action of AF-DX 384 on the W84-induced retardation of [3H]NMS dissociation was analyzed according to a method described by Lazareno and Birdsall (1993). The procedure can be regarded as a condensed form of the Schild method (Arunlakshana and Schild, 1959) to analyze the action of an antagonist. The approach is based on the simultaneous analysis of the concentration-effect curve of an effector agent E determined in the absence of antagonist and of the antagonist concentration-effect curve for the attenuation of the action of a single, fixed effector concentration. Here, the observed effect was the retardation of [3H]NMS dissociation by W84 in the absence and presence of AF-DX 384. To reveal the antagonistic action of AF-DX 384, its own effect on [3H]NMS dissociation was eliminated from the analysis by normalization (for details, see Results). The W84 and AF-DX 384 curves were fitted nonlinearly using as two independent variables the concentrations of W84 and AF-DX 384, respectively. This analysis was based on the following equation (Lazareno and Birdsall, 1993) with application of SigmaPlot for Windows (version 3.03; Jandel Scientific Software, San Rafael, CA):
Drugs
[3H]NMS and [3H]AF-DX 384 were purchased from New England Nuclear-Dupont (Bad Homburg, Germany). Oxotremorine sesquifumarate, atropine sulfate, (−)-scopolamine hydrobromide, (−)-scopolamine methylbromide, and (−)-scopolamine N-butylbromide were from Sigma Chemical (München, Germany). (+)-Dexetimide hydrochloride was from ICN Biomedicals (Meckenheim, Germany). Methoctramine tetrahydrochloride was from Research Biochemicals International (Natick, MA). AF-DX 384 was generously provided by Dr. Karl Thomae GmbH (Biberach an der Riβ, Germany). W84 was synthesized by Dr. Joachim Pfeffer (University of Kiel, Germany) according to Wassermann (1970). APE and 4-Cl-McN-A-343 were synthesized by Dr. Ulrich Moser (University of Frankfurt, Germany) according to Waelbroeck et al.(1994).
Results
In paced guinea pig atria, all test compounds antagonized the negative inotropic effect of the agonist oxotremorine. The concentration-effect curves of oxotremorine (−log EC50 = 7.89 ± 0.03, mean ± standard error, 184 experiments) were shifted by the test compounds in a parallel fashion to higher concentrations; the extent of rightward shift was expressed as DR. In Fig. 2, the DRs thus obtained for various concentrations of NMS, AF-DX 384, and W84, respectively, are plotted according to Arunlakshana and Schild (1959). For NMS and AF-DX 384, the data points could be fitted by lines with a slope of unity, as to be expected for a competitive interplay with the agonist. As a measure for the affinity to the M2 receptor, the pKB values (i.e., the concentrations at log [(DR − 1] = 0), are indicated in Table1. In case of W84, the antagonistic action against oxotremorine saturated at higher concentrations. According to the allosteric model, a curvilinear Schild plot reaching a limiting value at high concentrations of antagonist indicates allosteric modulation of agonist binding in a negative cooperative fashion (Ehlert, 1988). Thus, the data analysis was based on the allosteric model (Ehlert, 1988; Lanzafame et al., 1996). At log (DR − 1) = 0, the allosteric fit yields, corresponding to a Schild analysis for a competitive interaction, an estimate of the equilibrium dissociation constant KA for the binding of W84 at the free receptor. The limiting value reached by the curved line at higher concentrations of W84 is a measure of the degree of negative cooperativity between the allosteric ligand acting at a site different from the agonist binding site and the agonist at its binding site (Lanzafame et al., 1996). Table2 indicates the binding constantKA of W84 for the free M2 receptor and the cooperativity factor α for the interaction between W84 and oxotremorine. The findings made in guinea pig atria with the muscarinic agonist arecaidine propargyl ester [−log EC50 = 8.07 ± 0.04 (mean ± standard error), 18 experiments] in the set of experiments to check for an M2 selectivity of W84 are included in Fig.2 and Table 2. In addition, with this agonist, the curvilinear fit was significantly better than a linear fit (partial F test,p < 0.05). As expected, theKA value of W84 was independent of the agonist applied (t test, p > 0.05). The cooperativity factor α (Table 2) did not differ significantly (t test, p > 0.05) for either oxotremorine or APE as the agonist.
Electrical field stimulation of the rabbit vas deferens (M1/M4-like receptors) elicited neurogenic twitch contractions that could be concentration-dependently inhibited by the M1-selective muscarinic receptor agonist 4-Cl-McN-A-343 [−log EC50 = 6.89 ± 0.05 (mean ± standard error), 12 experiments]. Isotonic contractions of guinea pig ileum (M3 receptors) were induced by cumulative addition of APE [−log EC50 = 7.67 ± 0.04 (mean ± standard error), 9 experiments]. In both tissues, W84 surmountably antagonized the responses to the agonists. There was a concentration-dependent parallel shift to the right of agonist concentration-response curves without either the basal tension or maximal effects being affected. The Schild plots were linear throughout the antagonist concentration range studied (vas deferens: 1, 3, 10, and 30 μm; ileum: 10, 30, and 100 μm), and the slopes were not significantly different from unity (t test, p > 0.05) (Table 2). Thus, W84 was an apparently simple competitive antagonist in the two preparations studied, with KB values of 759 and 4370 nm in vas deferens and ileum, respectively (Table 2). These results show that W84 exhibits a higher antimuscarinic potency at atrial M2 receptors (KA ∼160 nm, Table 2) than at the other muscarinic receptor subtypes. It possesses the following selectivity profile: M2 > M1/M4 > M3 (Table 2). This profile of W84 is similar to that of two other allosteric modulators, gallamine (Ellis et al., 1991) and alcuronium (Jakubı́k et al., 1995).
The interaction of W84 with AF-DX 384 and other antimuscarinic drugs was studied in guinea pig atria with oxotremorine as the agonist. First, the antagonistic effect of the antimuscarinic drugs alone was measured at various drug concentrations. As expected for competitive antagonists, linear Schild plots with slopes not significantly different from unity were obtained (data not shown). The respective pKB values are listed in Table 1. In the combination experiments with W84, the competitive antagonists were applied at selected concentrations that induced a rightward shift (DRant) of the oxotremorine control curve by factors of 200-1800 (Table 1). The antagonists were combined with 100 and 1000 μm W84 (except for methoctramine, which was applied only in the presence of 1000 μm W84). The DRs induced by W84 alone in the respective sets of experiments are indicated in Table 1. The overall mean ± standard error values amounted to DRW84 = 260 ± 19 at 100 μm W84 and DRW84 = 236 ± 21 at 1000 μm W84 (Table 1). The antimuscarinic effects found in the combination experiments with 1000 μm W84 are displayed in Fig.3. All combinations with W84 except that of AF-DX 384 and methoctramine elicited a higher degree of antagonism than expected for a combination of two competitive antagonists. The extent of the deviation from competitive behavior was quantified by the factor F = DRobs/DRexp(Table 1). Combined with conventional antagonists, W84 induced F values of >1, indicating more-than-additive effects (Table 1). In contrast, less-than-additive effects (F < 1) were found when W84 was combined with the cardioselective drugs AF-DX 384 and methoctramine.
The noncompetitive interactions between W84 and the various antagonists were analyzed in terms of the cooperativity model according toChristopoulos and Mitchelson (1994). Because the preceding experiments gave the values for the cooperativity of W84 with the agonist oxotremorine (α = 311, Table 2), for the affinity of W84 at the free M2 receptor (KA = 137, Table 2), and for the affinity of the respective antagonists at the free receptor (KB , Table 1), it was possible to estimate the degree of cooperativity between W84 and the antagonists by means of equation (2) (see Data Analysis). The resulting cooperativity factors α are listed in Table 1. It is obvious that the negative cooperativity of W84 with oxotremorine (α = 311, Table 2) is smaller than with AF-DX 384 (α = 444) but larger than with NMS (α = 18) (Table 1). The equilibrium dissociation constant KA for the binding of W84 to the free receptor does not depend on the type of ligand that is also present. Thus, the differences among the cooperativity factors for the various antagonists can be attributed to differences in the affinities of W84 at the respective antagonist/receptor complexes. According to the cooperativity model of Ehlert (1988), the affinity of an allosteric modulator to a receptor occupied by an orthosteric ligand is the product of α·KA . These values for NMS and AF-DX 384 are shown in Table 3. If certain conditions are met, α·KA is equivalent to the concentration at which the modulator induces half-maximal retardation of ligand dissociation (for details, see Lazareno and Birdsall, 1995). In other words, the α·KA values listed in Table 3 indicate that W84 should stabilize NMS binding to M2 receptors with a higher potency than AF-DX 384 binding (Table 3).
To test this prediction, we measured the dissociation of [3H]AF-DX 384 and [3H]NMS in guinea pig cardiac membranes under the influence of W84 (Fig. 4). Under control conditions and in the presence of the test compounds, dissociation of both radioligands proceeded monophasically. The control half-life time was 4.7 ± 0.7 min (mean ± standard error; eight experiments) for [3H]AF-DX 384 and 11.3 ± 0.7 min (eight experiments) for [3H]NMS. W84 concentration-dependently retarded the dissociation of [3H]AF-DX 384. The apparent rate constant of dissociation (k −1 = ln2/t ½) served as a measure for the rate of ligand dissociation. A plot of 100 −k −1 versus the concentration of W84 yields the concentration-response curve for the retarding action on the dissociation of [3H]AF-DX 384 (Fig. 4). The curve levels off at 86%, indicating a slightly submaximal effect of W84 on [3H]AF-DX 384 dissociation (partialF test, p < 0.05). The half-maximal effect of W84 (i.e., the inflection point of the curve) lies at EC43, diss = 33,300 nm (Table 3). In contrast, the dissociation of [3H]NMS dissociation was almost completely inhibited at appropriate concentrations of W84 with the half-maximal effect attained at ∼40-fold lower concentrations (EC50, diss = 900 nm; Fig. 4 and Table 3) compared with [3H]AF-DX 384 as the ligand. Hill slopes were not significantly different from unity (nH = 1.18 in the presence of the radioligand [3H]AF-DX 384 and = 1.15 with [3H]NMS, partial F test,p > 0.05).
The parameters KA and α can be obtained from radioligand equilibrium binding experiments (Table 3) if the allosteric modulator alters equilibrium binding of the radioligand. As shown in Fig. 4, the equilibrium binding of [3H]NMS was diminished only slightly by W84. The curve fit based on the cooperativity model of allosteric action (Ehlert, 1988) indicated that the bottom of the inhibition curve was located above the 50% level of specific [3H]NMS binding. In contrast, W84 was capable of reducing the equilibrium binding of [3H]AF-DX 384 to the zero level (Fig. 4). Because the delay of [3H]AF-DX 384 dissociation requires high concentrations of W84, the inhibition curve mainly represents the inhibitive action of W84 on [3H]AF-DX 384 association (i.e., the binding of W84 to the free M2 receptor) (KA = 215, Table 3). This value is in acceptable correspondence with the respective binding constant for W84 derived from the separate inhibition experiment with [3H]NMS (KA = 457).
At [3H]NMS-occupied M2receptors, it was determined whether AF-DX 384 is capable of acting as an allosteric modulator by its own and whether there is an interference between the actions of W84 and AF-DX 384. AF-DX 384 retarded [3H]NMS dissociation concentration-dependently (Fig. 5A). Curve fitting with a variable upper plateau did not give a better fit compared with a plateau fixed at 100% (F test, p > 0.05). The slope of the curve (nH = 0.9) was not significantly different from unity (partial F test,p > 0.05). Half-maximal retardation of [3H]NMS dissociation occurred at ECdiss = 22,000 nm (withnH = 1).
To test whether AF-DX 384 interferes with the allosteric effect of W84 on the dissociation of [3H]NMS, both compounds were applied in combination. To compensate for the retarding effect of AF-DX 384 on [3H]NMS dissociation, the effect of the respective concentrations of AF-DX 384 alone (see also Fig. 5A) was normalized to a value of k −1, normalized = 1. This procedure is equivalent to experiments previously carried out with the modulator obidoxime as allosteric “antagonist” (Tränkle and Mohr, 1997). In the presence of AF-DX 384, the allosteric delay of [3H]NMS dissociation induced by 3 μm W84 was diminished concentration-dependently (Fig. 5B). The interaction of AF-DX 384 with W84 was analyzed according to Lazareno and Birdsall (1993). Because this analysis requires the control curve for the effect of W84, the curve shown in Fig. 4 for the allosteric action of W84 alone on [3H]NMS dissociation is also included in Fig.5B. Simultaneous analysis of the effector (W84) and the inhibition data (W84 plus AF-DX 384) was carried out; the slope n of the effector curve and the Schild factor s, respectively, were each checked successively for a deviation from unity. Neither n nor s deviated from unity (p > 0.05, data not shown). The Schild factor s = 1 means that the interaction of AF-DX 384 with W84 at the [3H]NMS occupied receptor is compatible with a competitive mode of interaction. The pKB value for the inhibitive action of AF-DX 384 on the allosteric effect of W84 yielded by the analysis is pKB = 4.6 ± 0.5 (i.e.,KB = 25,000 nm). This value favorably corresponds to the EC50, diss value of 22,000 nm for the individual allosteric action of AF-DX 384 on [3H]NMS dissociation. The close correspondence suggests that both effects of AF-DX 384 (i.e., inhibition of [3H]NMS dissociation and antagonism of the effect of W84) are mediated via the same site of attachment on the [3H]NMS-occupied M2 receptor protein.
Discussion
In terms of the cooperativity model, W84 exerts at muscarinic M2 receptors a negative cooperative interaction with antagonists, which, however, is considerably smaller with NMS as a representative of the nonselective antagonists (α < 20) than with the M2/M4-selective antagonist AF-DX 384 (α > 190). In principle, this difference is observed in paced guinea pig atria as well as in radioligand binding experiments (Table 3). In other words, W84 is able to inhibit the binding of AF-DX 384 to a larger extent than the binding of NMS.
The effect of an allosteric modulator on the equilibrium binding of a ligand results from the effect of the modulator on the association and on the dissociation of the ligand. The effect on the association reflects an interaction of W84 with the unoccupied receptor. W84 inhibits the association of NMS (Jepsen et al., 1988) and AF-DX 384 (Fig. 4). The affinity of an allosteric modulator for the unoccupied receptor is independent of the ligand (KA in Table 3). In contrast, the interaction of an allosteric modulator with a ligand-occupied receptor depends on the type of ligand bound to the receptor (Lee and El-Fakahany, 1988). Ligand dissociation of NMS and AF-DX 384 is impaired by W84 with largely different activity. Obviously, the affinity of W84 at the AF-DX 384-occupied receptor is lower than that for the NMS-occupied receptor; as assessed from the radioligand dissociation experiment (ECdiss values in Table3), the affinity ratio is 37. The results of the independent radioligand equilibrium binding experiments (α·KA values in Table 3) yield an affinity ratio of 38. From the results in paced atria (α·KA values in Table 3), a very similar affinity ratio amounting to 25 is derived. The absolute binding affinity of W84 at the ligand-occupied receptors, however, is somewhat lower under the “physiological conditions” of the organ bath experiments compared with the radioligand binding experiments. This observation can be accounted for by the different incubation conditions with regard to temperature and ionic composition (Tränkleet al., 1996).
W84 also interacts in a negative cooperative fashion with the agonists oxotremorine and arecaidine propargyl ester. The extent of the negative cooperativity with oxotremorine (α = 311, Table 3) is considerably greater than that with the antagonist NMS (α = 18). Therefore, in combination experiments, the interplay is shifted by W84 in favor of the binding of NMS compared with oxotremorine, thus inducing a more-than-additive antagonistic effect. In contrast, the degree of negative cooperativity between W84 and oxotremorine is smaller than that between W84 and AF-DX 384 (α = 444, Table 3). In other words, W84 inhibits AF-DX 384 binding more than oxotremorine binding, which explains the less-than-additive effect of this combination. A less-than-additive action was also observed for the combination of W84 with the M2-selective antagonist methoctramine but not with any of the other antagonists. The latter compounds are structurally heterogeneous, but they share a nonselective behavior. This coincidence suggests a relationship between the low affinity of W84 for the AF-DX 384-occupied receptor and the special binding mode of this M2-preferring antagonist.
On the molecular level, there are two possible explanations for the low affinity of W84 at the AF-DX 384-occupied M2receptor. First, AF-DX 384 may stabilize the M2receptor in another conformation than does NMS, and the conformation induced or stabilized by AF-DX 384 may have a low affinity for W84. Second, the receptor protein may have the same conformation with bound AF-DX 384 as with bound NMS, but AF-DX 384 imposes a steric hindrance for the attachment of W84, thus allowing only part of the W84 molecule to bind. This hypothesis is illustrated in Fig.6. Because AF-DX 384 affects receptor G protein coupling with negative intrinsic activity, as does the tropate atropine (Hilf and Jakobs, 1992), it seems reasonable to assume that AF-DX 384 and the tropate NMS stabilize the receptor in the same conformation. AF-DX 384 may bind in a way that it uses part of the allosteric site. This notion is supported by the results of Kerckhoff and Höltje (1994), who used a conformational analysis and receptor modeling approach for various M2- and M1-selective derivatives of pirenzepine to assess the orientation of the compounds in the ligand binding pocket of the receptor protein. It seemed pivotal for the M2receptor binding of AF-DX 116, which is a close congener of AF-DX 384, that the side chain is directed toward the entrance of the ligand binding pocket. As noted above, the allosteric site is likely to be located in this region of the M2 receptor. According to a recent structure-activity relationship study, the W84 molecule in its entire length is involved in the binding to the allosteric site of [3H]NMS-occupied M2 receptors (Kostenis et al., 1994). Shared points of attachment between AF-DX 384 and W84 would explain why the affinity of W84 at the AF-DX 384-occupied receptor compared with the NMS-occupied receptor is reduced.
In this context, the findings of Christopoulos and Mitchelson (1994)should be mentioned; they applied the heptamethonium analogue of W84 (i.e., heptane-1,7-bis(dimethyl-3′-phthalimidopropyl) ammonium bromide) in combination with NMS and with the M1-selective antagonist pirenzepine in paced guinea pig atria. Pirenzepine has the same tricyclic backbone as AF-DX 384 and AF-DX 116 but differs in the nature of the side chain and the spatial location of the protonated nitrogen (Eberlein et al., 1989). Combinations of C7/3′-phth with pirenzepine still acted supra-additively, but the extent of supra-additivity was considerably less pronounced than that in combinations of C7/3′-phth with NMS. Remarkably, Kerckhoff and Höltje (1994) suggested that pirenzepine attaches to the M2 receptor in the same orientation as AF-DX 116 (i.e., with the nitrogen directed toward the entrance of the ligand binding pore). For pirenzepine, however, this is an unfavorable position leading to a low binding affinity at the M2 receptor. This low affinity precludes direct binding experiments with radiolabeled pirenzepine to study further the interaction with the allosteric modulator C7/3′-phth. However, having in mind the high negative cooperativity between C7/3′-phth and pirenzepine (α = 58) compared with NMS (α = 9), as reported byChristopoulos and Mitchelson (1994), it is tempting to assume that pirenzepine interferes with the binding of C7/3′-phth to the allosteric site in a similar way as proposed here for the respective analogues AF-DX 384 and W84.
As found with other M2-preferring compounds such as AF-DX 116 (Lee and El-Fakahany, 1991), AF-DX 384 is capable of stabilizing NMS receptor complexes. This effect may result from an interaction with the allosteric binding site in [3H]NMS-occupied M2receptors as illustrated in Fig. 6. This idea is compatible with the lower affinity of AF-DX 384 at the [3H]NMS-occupied receptor (ECdiss = 22,000 nm) compared with the free M2 receptor (KD = 11 nm). In the first case, NMS hinders the ring system of AF-DX 384 from binding. In the latter case, the docking place for the ring system is available, allowing for a pronounced gain in affinity. As far as [3H]NMS-occupied receptors are concerned, the binding affinity of AF-DX 384 is much lower than that of W84 (ECdiss = 22,000 versus 900 nm) yet similar to the affinity of the prototype modulator gallamine (ECdiss = 16,000 nm, Tränkle et al., 1996). The character of the interaction between W84 and AF-DX 384 at the NMS-occupied receptor is formally competitive (Fig. 5). This finding suggests that AF-DX 384 interacts with the common allosteric site of NMS-occupied M2 receptors (Ellis and Seidenberg, 1992; Tränkle and Mohr, 1997). Yet, we are aware that it is speculative to assume that the part of AF-DX 384 that is capable of attaching to the NMS-occupied receptor is identical with the part of the molecule that possibly binds to the allosteric site when AF-DX 384 attaches to the free receptor.
With regard to the interaction of methoctramine with M2 receptors, Melchiorre et al. (1989)proposed that the binding domain of this long tetraamine molecule includes the orthosteric and the allosteric recognition site of the receptor protein. This hypothesis is supported by our finding that methoctramine like AF-DX 384 is less-than-additive antimuscarinic in combination with W84 (Fig. 3). Furthermore, methoctramine is known to interact with [3H]NMS-occupied M2 receptors having a 7-fold higher affinity (EC50, diss = 3,000 nm; Tränkleet al., 1996) than AF-DX 384 (EC50, diss = 22,000 nm). Thus, methoctramine and AF-DX 384 seem to interact with M2 receptors in a similar fashion.
In conclusion, the results presented here suggest “shared points of attachment” between the muscarinic antagonist AF-DX 384 and the potent allosteric modulator W84. It is tempting to assume that M2-preferring antagonists may derive part of their receptor subtype selectivity from an interaction with the common allosteric site of the M2 receptor protein.
Acknowledgments
We gratefully acknowledge the excellent technical assistance of Micheline Neubert (Kiel, Germany) and Frauke Mörschel and Iris Witten (Bonn, Germany).
Footnotes
- Received August 6, 1997.
- Accepted October 24, 1997.
-
Send reprint requests to: Klaus Mohr, Pharmacology & Toxicology, Institute of Pharmacy, University of Bonn, An der Immenburg 4, 53121 Bonn, Germany. E-mail:k.mohr{at}uni-bonn.de
-
This work was supported by the Deutsche Forschungsgemeinschaft (K.M.) and the Fonds der Chemischen Industrie (G.L.).
Abbreviations
- NMS
- N-methylscopolamine
- APE
- arecaidine propargyl ester
- NBS
- (−)-N-butylscopolamine
- W84
- hexane-1,6-bis(dimethyl-3′-phthalimidopropyl-ammonium bromide)
- DR
- dose ratio
- 4-Cl-McN-A-343
- 4-(4-chlorophenylcarbamoyloxy)-2-butynyltrimethylammonium iodide
- The American Society for Pharmacology and Experimental Therapeutics