Muscarinic Agonist Binding and Phospholipid Turnover in Brain

The ability of muscarinic cholinergic agonists to interact with muscarinic receptors in nerve ending preparations and elicit an increased labeling of phosphatidate and phosphatidylinositol from 32Pi has been investigated. Two groups of brain muscarinic agonists are distinguished. Addition of acetylcholine, carbamylcholine, methacholine, or muscarine resulted in a 2-fold stimulation of phosphatidate and phosphatidylinositol labeling, while bethanechol, pilocarpine, arecoline, and oxotremorine were less effective. Simultaneous addition of two agonists from the more effective group did not result in any further increase in stimulated labeling, while the addition of agonists from the less effective group antagonized the stimulatory effect of carbamylcholine. All of the agonists could completely displace binding of [3H]quinuclidinyl benzilate, a muscarinic antagonist. The displacement of the labeled antagonist by the more effective agonists was more complex than that predicted from a simple mass action isotherm and was compatible with the interaction of the agonists with high and low affinity forms of the receptor. Conversely, the displacement data from less effective agonists did not deviate markedly from those predicted for interaction of the agonists with a single affinity form of the receptor. Dose-response curves for stimulated phosphatidate labeling obtained in the presence of acetylcholine, carbamylcholine, and methacholine were predominantly correlated with occupation of the low affinity form of the muscarinic receptor. These results suggest that the enhancement of phosphatidate and phosphatidylinositol turnover in brain is caused by agonist-mediated conformational changes in the muscarinic receptor and that the ability of an agonist to induce this conversion may be predicted by its differential binding to the high and low affinity forms of the receptor.

The ability of muscarinic cholinergic agonists to interact with muscarinic receptors in nerve ending preparations and elicit an increased labeling of phosphatidate and phosphatidylinositol from 32Pi has been investigated. Two groups of brain muscarinic agonists are distinguished. Addition of acetylcholine, carbamylcholine, methacholine, or muscarine resulted in a 2-fold stimulation of phosphatidate and phosphatidylinositol labeling, while bethanechol, pilocarpine, arecoline, and oxotremorine were less effective. Simultaneous addition of two agonists from the more effective group did not result in any further increase in stimulated labeling, while the addition of agonists from the less effective group antagonized the stimulatory effect of carbamylcholine. All of the agonists could completely displace binding of [3H]quinuclidinyl benzilate, a muscarinic antagonist. The displacement of the labeled antagonist by the more effective agonists was more complex than that predicted from a simple mass action isotherm and was compatible with the interaction of the agonists with high and low affinity forms of the receptor. Conversely, the displacement data from less effective agonists did not deviate markedly from those predicted for interaction of the agonists with a single affinity form of the receptor. Dose-response curves for stimulated phosphatidate labeling obtained in the presence of acetylcholine, carbamylcholine, and methacholine were predominantly correlated with occupation of the low affinity form of the muscarinic receptor. These results suggest that the enhancement of phosphatidate and phosphatidylinositol turnover in brain is caused by agonist-mediated conformational changes in the muscarinic receptor and that the ability of an agonist to induce this conversion may be predicted by its differential binding to the high and low affinity forms of the receptor.
Biochemical correlates of muscarinic receptor activation in a variety of tissues include an increased synthesis of cGMP (l), a decreased synthesis of CAMP (2), an increase in Ca2+ fluxes (3), a decreased K' conductance (4), and an increased turnover of two quantitatively minor phospholipids, PhA' and PhI (5-7). Of these correlates, only stimulated phospholipid turnover (usually determined as an increased incorporation of 32P, into PhA and PhI) consistently accompanies muscar- inic receptor activation. While the precise physiological significance of enhanced phospholipid turnover remains unresolved (8), the biochemical mechanism appears to involve an initial phosphodiesteratic breakdown of PhI or its phosphorylated derivatives, phosphatidylinositol 4-phosphate or phosphatidylinositol 4,5-bisphosphate, to yield diacylglycerol, which in turn is rephosphorylated in the presence of ["PI ATP to yield labeled PhA and PhI (9)(10)(11)(12).
Known physiological consequences of muscarinic receptor activation, such as contraction or secretion, typically require only a partial receptor occupancy for maximal effect.
In contrast, concentrations of agonists approaching those required for full receptor occupancy are necessary to evince a maximal increase in the turnover of PhA and PhI (13). This observation has prompted the hypothesis that enhanced phospholipid turnover serves in a biochemical amplification step that follows receptor-ligand interaction. However, the relationship between binding of the agonist to the receptor and the initiation of the biochemical response is unclear. For example, while the occupancy hypothesis predicts that the concentration-dependence curves for enhanced phospholipid turnover should be displaced to the right (ie. higher concentration range) of the physiological curve, this is not always the case (14, 15). Similarly, the addition of partial muscarinic agonists does not invariably result in the predicted smaller increase in phospholipid turnover (16,17). In addition, while it has been proposed that muscarinic receptors exist in both high and low affinity forms (18), it has yet to be clearly established whether one or both forms of the receptor are coupled to stimulated phospholipid turnover.
In this study, we have examined the relationship between the binding of agonists to brain muscarinic receptors and the initiation of a metabolic response, enhanced PhA and PhI labeling from 32Pi in a nerve ending fraction obtained from guinea pig cerebral cortex.
Phospholipid labeling from 32Pi in the absence and presence of cholinergic agonists was carried out by incubating aliquots of NE1.1 fraction or chromaffin cells a t 37 "C in Buffer A containing 20-50 pCi of 32P, in a total volume of 0.5 ml. Incubations were terminated after 30 min for NE,,, fractions and 60 min for chromaffin cells, and 7358 by guest on October 15, 2017 http://www.jbc.org/ Downloaded from the lipids were extracted, separated, and quantitated as previously described (22). Under these conditions, the amount of 3zP, incorporated into PhA and PhI was proportional to the amount of protein present, and the stimulation (per cent control) of both lipids was independent of protein concentration. Data for stimulated PhA or PhI labeling as a function of agonist concentration was evaluated for evidence of a one-or two-site interaction from the following equations.
where f = fractional increase in labeling of PhA or PhI at a concentration [A] of agonist, Fmax = maximum increase in labeling, and K = affinity constant of the receptor for stimulated PhA or PhI turnover. For a two-site fit, Equation 1 is modified as follows.
where F H and F L are proportions of high and low affinity sites with affinities KH and KL, respectively, assuming Fmax = FH + FL.
[3H]QNB binding to NE,., fractions was carried out as previously described (23). Briefly, approximately 60 pg of NE,., protein were incubated in 2 ml of Buffer A a t 25 "C, in the presence of 0.65 nM L-[3H]QNB. This temperature assures minimal dissociation of the QNB-receptor complex. Reactions were terminated after 60 min by addition of 3 ml of Buffer A and rapid filtration with a manifold filtration apparatus. Filters were washed twice with 6 ml of chilled Buffer A and radioactivity was determined after the addition of 10 ml of ACS and allowing the vials to stand overnight. Under these conditions >95% of 13HJQNB was free in solution. Nonspecific binding (<2% of total) was that radioactivity bound to the filter in the presence of 1.0 pM atropine. For displacement studies, agonists were added in varying concentrations prior to the addition of [3H]QNB. The counting efficiency for 3H was 40%.
Analysis of [3H]QNB binding to NE,., fractions and its competition by muscarinic agonists was performed by the method of McKinney and Coyle (24)

Muscarinic Agonist Binding and
Phospholipid Turnover analysis (18). In addition, plots of occupancy of the receptor by agonists as a function of [A] were visually examined for deviations from predicted one-or two-site models. In some experiments the data were fitted to a three-site model, incorporating the terms for displacement of [3HJQNB from a "super high" affinity site by a muscarinic agonist (25). Only in a limited number of experiments was there a statistically significant improvement in fit, so that this data treatment was not routinely performed. Protein was determined by the method of Geiger and Bessman (26).

Stimulation by Muscarinic Agonists of PhA and PhZ Labeling from "Pi in NE,.1 Fractions and Dissociated Adrenal
Chromaffin Cells-Muscarinic agonists (1.0 mM) differed considerably in their ability to enhance PhA and PhI turnover in the fraction. Addition of carbamylcholine,.acetylcholine, methacholine, or muscarine (Table I, Group A) resulted in a larger stimulation of PhA and PhI labeling than agonists in Group B, i.e. bethanechol, pilocarpine, arecoline, or oxotremorine. This pattern of agonist efficacy was also observed for NE1.l fractions obtained from rat cerebral cortex and other areas of guinea pig brain (hippocampus, inferior and superior colliculi, and pons/medulla, not shown). However, in bovine adrenal chromaffin cells, oxotremorine and arecoline were relatively more efficacious, although still less effective than carbamylcholine in stimulation of PhI labeling, while in this tissue pilocarpine was the least efficacious agonist.
Group B Agonists: Dose-Response and Effect on Carbamylcholine-stimulated PhA and PhZ Labeling-To establish that optimum concentrations of agonists in Group B had been employed, dose-response curves for stimulated PhA and PhI labeling were constructed (Fig. 1). Stimulated PhA and PhI labeling was optimum a t approximately 1.0 mM concentrations of oxotremorine, arecoline, bethanechol, and pilocarpine. Agonists in Group B antagonized the stimulatory effect of carbamylcholine addition on phospholipid labeling. The rank order potency so obtained was oxotremorine > arecoline = pilocarpine > bethanechol. Conversely, the simultaneous addition of agonists in Group A did not elicit an increase in labeling of either PhA or PhI greater than that produced by each agonist alone (Table 11).  of muscarinic agonists. The displacement of I3H]QNB binding was then analyzed for evidence of one-site or two-site interaction (see under "Materials and Methods"). The necessary criteria for an agonist to fit a two-site model included a statistical difference between the sum of the squared residuals for one-and two-site models, nearest neighbor residual analysis, and visual inspection of receptor occupancy curves. Agonists which did not meet these requirements were considered to interact with a single affinity form of the receptor. Results shown are the best fits (one-or two-site) for each muscarinic agonist, and the number of separate experiments (n) in which this pattern was observed. Values for the high and low affinity forms of the receptor (KH and KL) and percentage high affinity sites are expressed as mean f S.E. For agonists in Group B, a single affinity constant has been calculated. pmol (n = 8). All agonists could completely displace [3H]QNB bound to the NE1., fraction, the most potent being oxotremorine, while the least effective was bethanechol. Plots of receptor occupancy by the agonist as a function of agonist concentration revealed that Group A agonists showed a systematic deviation from the curve predicted for the interaction of an agonist with a single affinity site. The displacement data could, however, be readily fitted to a two-site model in which the muscarinic receptor was proposed to exist in both high and low affinity forms (Fig. 2). In contrast, the displacement data for agonists in Group B could readily be fitted to a one-site model, and there was no significant improvement in fit for a two-site model (Fig. 3). Although the sum of the squared residuals for a two-site model was invariably smaller for agonists in both groups, only for Group A agonists were statistically significant differences in fit between one-site and two-site models seen consistently. A summary of the computer-derived data obtained for KH, KL, BH, BL, and percentage high affinity sites obtained from eight separate NE1,, fractions is shown in Table 111. These values are in close agreement with published studies on rat brain (18,27).
Approximately 30-40% of the sites present in guinea pig cerebral cortex NE,., fractions are in the high affinity form. For Group A agonists the ratio KJKH was greatest for acetylcholine and least for muscarine.
Dose-Response Curves for Stimulation of PhA Labeling by Agonists in Group A-To determine whether the high, low, or both affinity forms of the muscarinic receptor were coupled to enhanced phospholipid turnover, dose-response curves were constructed for stimulated PhA labeling. This parameter

Muscarinic Agonist Binding and Phospholipid
Turnover was selected as the preferred measure of receptor activation, since it represents an early plasma membrane event, unlike PhI resynthesis, which occurs subsequently at the endoplasmic reticulum (5). In addition, enhanced PhA labeling is of sufficient magnitude to permit accurate detection at low agonist concentrations, in contrast to stimulated inositol lipid breakdown which is often 25% or less for full agonists at optimum concentrations (14,17,28,29). Carbamylcholine stimulation of PhA labeling was detectable at 1.0 p~ and was maximal at 10 mM (Fig. 4). While this data could be dissociated into high and low affinity components using computer fits, only 12% of the inferred muscarinic receptors were in the high affinity state. The constant for the low affinity state of the muscarinic receptor (109 p~) as determined from enhanced PhA labeling was in the range of that obtained from the binding data (77 p~) .
Similarly, dose-response curves for PhA labeling obtained with methacholine and acetylcholine indicated that both of these agonists also interact predominantly with a single low affinity form of the receptor (Fig. 5 ) .
The values for affinity constants so obtained were 30 p~ and 20 p~, respectively, and are in the range of values obtained for the low affinity form of the muscarinic receptor derived from binding studies (20 p~ and 6 p~, respectively).

DISCUSSION
The present study indicates that muscarinic receptors coupled to enhanced phospholipid turnover in nerve ending preparations from cerebral cortex possess a distinctive profile with regard to agonist efficacy. The ability of Group B agonists to antagonize the stimulatory effect of carbamylcholine indicates that while both groups of agonists bind to the same population of brain muscarinic receptors, Group B agonists are less efficacious and may thus be considered "partial" agonists. In contrast, when muscarinic receptors on dissociated adrenal chromaffin cells are examined, arecoline and oxotremorine are relatively more efficacious agonists than they are in the brain. The categorization of agonists, therefore, appears to be highly dependent on the source of receptors.
Evidence in favor of a tissue-specific profile of muscarinic agonist efficacy for stimulated phospholipid turnover and/or the attendant physiological response has been obtained from other sources. Thus, bethanechol addition to the superior cervical ganglion elicits increases in PhI turnover and membrane depolarization comparable to other full agonists in this preparation such as carbamylcholine or muscarine, while the physiological responses (depolarization) due to pilocarpine and oxotremorine addition are characteristic of partial agonism (4,30). In contrast, the stimulation of amylase release from pancreatic preparations and the accompanying PhI breakdown evoked by the addition of pilocarpine are comparable in magnitude to those produced by acetylcholine or carbamylcholine (31,32). While these tissue differences might be interpreted to reflect the presence of different receptor molecules, present evidence favors the existence of a single receptor protein in varying molecular environments (33).
The pattern of agonist efficacy in brain is related to the characteristics of agonist displacement of [3H]QNB binding in nerve ending preparations. Agonists that were most effective in enhancing PhA and PhI turnover displaced [3H]QNB in a complex manner, with binding curves deviating markedly from a simple mass action isotherm (Fig. 2). This binding behavior of the agonists can be explained by postulating the presence of two noninteracting forms of the muscarinic receptor with large (30-to 150-fold) differences in KL and KH, as has been observed in rat brain preparations (18). In contrast, less effective agonists displaced [3H]QNB in a simpler manner, with minor deviation from a single mass action isotherm and relatively small differences in values of KL and KH. These results are consistent with a model for muscarinic receptor activation proposed by Birdsall and Hulme (34) in which agonists, unlike antagonists, are assumed to induce a conformational change in the receptor, and differences in agonist efficacy arise as a result of conformational restraints imposed by the coupling state of the receptor. From this, it can be predicted that the efficacy of an agonist is a function of KL/ KH. In the present study this prediction is met; full agonists with regard to phospholipid turnover also possess large values for KL/KH, while the partial agonists, bethanechol, pilocarpine, arecoline, and oxotremorine, display lower KL/KH ratios. The fact that muscarine (KL/KH = 28) was as effective as acetylcholine (KL/KH = 148) indicates a requirement for a minimum value of KL/KH for full agonist status, below which a reduced response is seen. While full agonists typically possess a high KL/KH ratio, the low affinity form of the brain muscarinic receptor rather than the high affinity form appears in general to be coupled to phospholipid turnover. Although computer fits of carbamylcholine-stimulated PhA labeling could be dissociated into high and low affinity components, almost 90% of the receptors coupled were in the L-form. Similarly, dose-response curves for acetylcholine-and methacholine-stimulated PhA labeling also revealed the sole involvement of a low affinity form of the receptor. Other studies have also indicated the importance of the L-form of the receptor, e.g. cGMP stimulation in neuroblastoma cells (l), inhibition of adenylate cyclase in heart (2), smooth muscle contraction in the ileum (35), and stimulation of PhI turnover by methacholine in pancreas (17). In contrast, enhancement of PhI turnover in guinea pig ileum correlates with the occupancy of both Land H-forms of the receptor (13), and stimulation of Ca2+ influx into smooth muscle reflects the activation of the H-form (36). While little information exists as to the molecular nature of these high and low affinity forms of the receptor, recent photoaffinity probe studies suggest that the muscarinic receptor exists in oligomeric forms, with the L-form being a dimer of molecular weight 86,000 Da, while the H-form of the receptor is a 160,000-Da tetramer (37).
The observation that the L-form of the muscarinic receptor is preferentially coupled to phospholipid turnover in brain is in complete accord with the model proposed by Birdsall and Hulme (34). The present results also indicate that stimulated phospholipid turnover in brain is tightly coupled to the OCcupancy of the muscarinic receptor, with little or no "receptor reserve,'' in agreement with observations made on enhanced PhI turnover in the ileum (13) and pancreas (17) but in contrast to recent results for methacholine stimulation of PhI turnover in rat parotid (14).
Michell and Kirk (38) have suggested that this apparent discrepancy may arise from measurement of enhanced PhI turnover by means of labeling experiments, rather than by inositol lipid breakdown. HOWever, results from the same laboratory have indicated a good correlation between the concentrations of agonists required to elicit a maximal increase in inositol phospholipid breakdown and increased PhI labeling (28).
In conclusion, the results from this study suggest that the profile of agonist efficacy for stimulated PhA and PhI turnover is a reflection of the relative abilities of muscarinic agonists to induce a conformational change in the brain receptor. The low affinity form of the receptor is that coupled to PhA and PhI turnover, and full receptor occupancy is required for a maximum effect.