Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Currently used drugs that bind to G protein-coupled receptors produce their therapeutic effects either by mimicking (in the case of agonists) or blocking (in the case of antagonists) the action of the endogenous signaling molecule, by competing with it at the same primary binding site on the receptor. The effect of such drugs is to either generate continuous stimulation or chronic blockade of receptor function at all the receptor molecules. In certain therapeutic indications, it may be more desirable to have an alternative approach, that of controlling the activity of the endogenous agonist at a given receptor.

This regulation of receptor function can be accomplished if a ligand operates by an allosteric mechanism, i.e., the ligand binds to a different site on the receptor from the endogenous agonist, to modulate agonist binding and function. The simplest scheme is shown in Fig. 1 (solid lines). The two parameters that describe the action of an allosteric agent X are its affinity for the receptor (K_X) and its cooperativity () with a primary binding site ligand L, which may be the endogenous agonist. The cooperativity factor  can be greater than, less than, or, in a special case, equal to 1: this corresponds to positive, negative, or neutral cooperative interactions, respectively.

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Schematic representation of the ternary complex allosteric model (solid lines, heavy text) and a representation of the effect of competition between two allosteric agents at the allosteric site on the binding of the radioligand at the primary binding site (whole of figure). In this scheme, the radioligand L and the allosteric agents X and Y bind to the primary and allosteric sites, respectively, on the receptor R. KL, KX, and KY are the affinity constants for L, X, and Y, respectively, for binding to R, , and  are the cooperativity factors between X or Y, respectively, and L, and k1, k2, and k3 are the rate constants for L dissociating from RL, XRL, and YRL, respectively.

One example of such a potential therapeutic area is the use of an allosteric muscarinic enhancer ( > 1 with L being acetylcholine) in the early stages of Alzheimer's disease, to alleviate the cognitive deficits that are thought to be caused by the localized degeneration of cholinergic nerve terminals and the associated acetylcholine deficit (Giacobini, 1990; Ehlert et al., 1994). In fact, M₂ muscarinic receptors were the first G protein-coupled receptor at which allosterism was demonstrated, when the antagonism by gallamine of the negative inotropic effect of muscarinic agonists in the heart was investigated (Clark and Mitchelson, 1976). This work was confirmed and extended by Stockton et al. (1983). From a therapeutic viewpoint, however, it is necessary to be able to enhance acetylcholine function in a subtype-selective manner, and this has recently been achieved with brucine and related compounds (Birdsall et al., 1997, 1999; Lazareno et al., 1999).

There have been relatively few reports on the detailed characterization of the effects of allosteric ligands on agonist binding (especially kinetics) and function. Outside of the muscarinic area, studies on adenosine receptors have shown that PD81,723 enhances agonist binding and function at the adenosine A₁ receptor as well as modulating agonist kinetics at the receptor-G protein complex (Bruns and Fergus, 1990; Cohen et al., 1994; Musser et al., 1999).

With adrenergic receptors, it has been demonstrated that amilorides can act allosterically at a single allosteric site on the _2A-adrenergic receptor, and in a more complex manner at the _1A-adrenergic receptor, to inhibit antagonist binding (Nunnari et al., 1987; Leppik et al., 1998a, 2000). However, there have been no studies of agonist-amiloride interactions. In this paper we present a quantitative investigation of the allosteric interactions of amiloride analogs on agonist binding and function at the _2A-adrenergic receptor, including the effects of the amilorides on the actions of the endogenous agonists, epinephrine and norepinephrine.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. [³H]UK14304 (27 Ci/mmol) was from DuPont NEN (Hounslow, Middlesex, UK), and [³⁵S]GTPS (1050 Ci/mmol) was from Amersham International (Little Chalfont, Buckinghamshire, UK). Amiloride HCl, BZA, DMA HCl, EPA, HMA, MBA, GDP, UK14304, epinephrine, norepinephrine, and phentolamine HCl were from Sigma Chemical Co. (Poole, Dorset, UK). Tissue culture reagents were from Gibco BRL (Paisley, UK).

Cell Culture and Membrane Preparation. The CHO cell line stably expressing the human _2A-adrenergic receptor (Kurose and Lefkowitz, 1994) was generously provided by Professor Robert J. Lefkowitz (Howard Hughes Medical Institute, Duke University Medical Center, Durham, NC). The cell line was grown in -minimal essential medium supplemented with 10% newborn calf serum, 2 mM L-glutamine, 50 I.U./ml penicillin, and 50 µg/ml streptomycin, at 37°C in 5% CO₂. Membranes were prepared as described previously (Leppik et al., 1998a). Briefly, near-confluent cells were harvested in cold buffer 1 (20 mM Na-HEPES, pH 7.4, 10 mM EDTA), then homogenized, and centrifuged. The pellet was resuspended in buffer 2 (20 mM Na-HEPES, pH 7.4, 0.1 mM EDTA), recentrifuged, again resuspended in buffer 2, and then stored at 70°C. Protein concentrations were determined by the method of Bradford (1976), using bovine serum albumin as the standard.

Radioligand Binding Assays. For competition experiments, membranes (20, 25, or 30 µg of protein) were incubated with approximately 6, 0.6, or 0.3 nM [³H]UK14304, respectively, in duplicate, together with increasing concentrations of competing agent, in a final volume of 1 ml of assay buffer (20 mM Na-HEPES, pH 7.4, 100 mM NaCl, 10 mM MgCl₂), at 30°C for 120 min. Nonspecific binding was defined as the binding retained on the filter and membranes in the presence of 20 µM phentolamine. Bound and free ligand were separated by rapid filtration under vacuum through GF/B glass fiber filters (Whatman, Maidstone, Kent, UK) using a Brandell cell harvester (Semat, St. Albans, Hertfordshire, UK). The filters were washed three times with cold 20 mM sodium phosphate buffer, pH 7.4, and transferred to scintillation vials. Scintillation cocktail (Beckman, Palo Alto, CA) was added, and the filters were soaked overnight and then counted.

[³⁵S]GTPS Assays. In each tube of a set of 24 tubes, membranes (10 µg) were incubated at 30°C with GDP (1 µM), [³⁵S]GTPS (0.05 nM), a given concentration of the amiloride to be tested (where relevant), and an increasing concentration of agonist in duplicate, in a final volume of 1 ml of assay buffer. Sufficient sets of 24 tubes (normally five to seven sets) were assayed consecutively, to enable the desired concentration range of the amiloride to be tested to be covered. Where the agonist was either epinephrine or norepinephrine, sodium metabisulfite (1 mM) was also included. Sodium metabisulfite had been found in initial tests to be necessary, to minimize a decrease in the [³⁵S]GTPS bound at concentrations of epinephrine or norepinephrine over 0.1 mM (data not shown). The time course for each set of 24 tubes was started by the addition of the membrane suspension and terminated by filtration after 60 min, as described above.

Data Analysis. Data were fitted by nonlinear regression analyses, using the Grafit curve-fitting software (Erithacus Software, Staines, Middlesex, UK). This procedure allows the use of two or more independent variables (e.g., time and concentration), which was necessary for many of the analyses reported in this paper.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Effect of Amiloride Analogs on the Equilibrium Binding of [³H]UK14304. The data from competition experiments between [³H]yohimbine and UK14304 at the human _2A-adrenergic receptor, permanently expressed in a CHO-K1 cell line (Kurose and Lefkowitz, 1994; Leppik et al., 1998a), were compatible with the presence of two high-affinity GTP-sensitive agonist sites and one low-affinity GTP-insensitive agonist site, with log affinities of 10.2 ± 0.2, 8.4 ± 0.1 and 6.59 ±0.02, respectively (K_d values: 0.1, 5.8, and 260 nM, respectively; eight experiments, data not shown). The two high-affinity sites were present in approximately equal proportions and together represented 32 ± 3% of the total binding sites labeled by [³H]yohimbine. The high-affinity binding constants are in reasonable agreement with those previously reported (Neubig et al., 1988), but the proportion of high-affinity sites is somewhat lower, perhaps reflecting the higher B_max (20 ± 2 pmol/mg protein) found in the cell membranes from the transfected cells used in the current work. GTP (1 mM) converted both high-affinity states to the low-affinity state of the receptor. Equilibrium [³H]UK14304 binding more directly suggested the presence of two high-affinity agonist binding sites in approximately equal proportions, with K_d values of 0.21 ± 0.02 and 2.2 ± 0.2 nM.

Modulation by Amilorides of [³H]UK14304 Dissociation from the _2A-Adrenergic Receptor. In the dissociation assay protocol developed, the membrane suspension containing the _2A-adrenergic receptor was kept on ice, and then aliquots were added to tubes containing [³H]UK14304 in assay buffer at 30°C. After 40 min of equilibration, an aliquot of this mixture was transferred to a second tube at 30°C containing phentolamine with or without amiloride analog in assay buffer, to start the dissociation. With this protocol, the dissociation of the [³H]UK14304 from the _2A-adrenergic receptor was found to be monoexponential in the absence or presence of a given concentration of amiloride analog; a double-exponential equation did not give a significantly better fit of the data (P > .1). [³H]UK14304 alone had a dissociation rate from the _2A-adrenergic receptor at 30°C of 0.041 ± 0.003 min¹ (t_1/2 = 17 min; n = 6).

View larger version (25K): [in this window] [in a new window]   Fig. 2.   Structural formulae of amiloride and the analogs examined in this paper.

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Dissociation of [3H]UK14304 in the absence or presence of various concentrations of HMA. Aliquots of an ice-cold membrane suspension were added to [3H]UK14304 pre-equilibrated at 30°C and then, after 40 min, a portion was transferred to a second tube at 30°C containing phentolamine and various concentrations of HMA to commence the dissociation measurement. Individual data points from one experiment are shown. The data from each experiment were simultaneously fitted to the equation from the allosteric model (Fig. 1, solid lines) (eq. 10, Leppik et al., 1998), with time and HMA concentration as independent variables. The results are summarized in Table 1.

View this table: [in this window] [in a new window]   TABLE 1 Effect of amiloride analogs on [3H]UK14304 dissociation from the 2A-adrenergic receptor The experiments were performed at 30°C, as described in the legend of Fig. 3. The data from each experiment were fitted to the single-exponential equations derived in a previous study (equations 10 and 9, respectively, Leppik et al., 1998). k1, k2, and k3 are the dissociation constants for the dissociation of [3H]UK14304 from the unoccupied (k1), HMA-occupied (k2), or amiloride- or DMA-occupied (k3) receptors. The log obs and log obs are the calculated differences between the log affinity of the amiloride analog at the agonist-occupied receptor (log KX or log KY) and the log affinity at the unoccupied receptor (log KX, described in the second paragraph under Results). Values are means ± S.E. of three experiments.

View larger version (22K): [in this window] [in a new window]   Fig. 4.   Effect of either amiloride (A) or DMA (B) on the modulation by HMA of [3H]UK14304 dissociation at 30°C. The experiments were performed as described in the legend to Fig. 3. Individual data points from one experiment are shown per graph. The lines represent the simultaneous fit of the data to the equation derived from the allosteric model that allows for competition between the allosteric agents at the allosteric site (Fig. 1) (eq. 9, Leppik et al., 1998). In the fitting, time, HMA concentration and either amiloride or DMA concentration were independent variables. The arrows show the direction of the effect of amiloride (A) or DMA (B) on the modulation by HMA of the [3H]UK14304 dissociation. The parameters derived from the fits are summarized in Table 1.

Characterization of Agonist-Stimulated Binding of [³⁵S]GTPS to the _2A-Adrenergic Receptor. [³⁵S]GTPS assay conditions were established using norepinephrine. The basal (or nonspecific) level of [³⁵S]GTPS binding was defined as that level of binding in a given time period in the absence of added agonist. In all assays, norepinephrine (and epinephrine) was used in the presence of 1 mM sodium metabisulfite, to minimize oxidation of the catechol moiety. In the presence of 1 µM GDP, 0.05 nM [³⁵S]GTPS, and 0.3 mM norepinephrine, the stimulated binding increased over a 90-min time interval (data not shown). Sixty minutes was chosen as the standard incubation time. Under these assay conditions, full agonists gave over a 5-fold stimulation over basal of [³⁵S]GTPS binding to membranes containing the _2A-adrenergic receptor. There was no evidence for constitutive activity; addition of the antagonist phentolamine (1 µM) in the absence of agonists did not cause a decrease in basal level of [³⁵S]GTPS binding (data not shown).

View this table: [in this window] [in a new window]   TABLE 2 Stimulation by agonists of [35S]GTPS binding to membranes containing the 2A-adrenergic receptor Membranes were incubated at 30°C for 60 min with GDP, [35S]GTPS, sodium metabisulfite (epinephrine and norepinephrine only), and an increasing concentration of agonist in duplicate. The data from each experiment were fitted with a four-parameter logistic equation, to give the log EC50, Emax, and basal and slopes for each dose-response curve. Values are means ± S.E. of n experiments.

Effect of Amiloride Analogs on the Stimulation of [³⁵S]GTPS Binding by Agonists. All six of the amiloride analogs tested (Fig. 2) caused a concentration-dependent decrease in the potency of norepinephrine to stimulate [³⁵S]GTPS binding to CHO membranes containing the _2A-adrenergic receptor (Fig. 5; not shown). However, above a given concentration, all of the amilorides also caused a decrease in the maximal level of norepinephrine-induced [³⁵S]GTPS binding, and in some cases there was also a variation in the level of basal binding at these higher concentrations. Thus, the equation derived by Lazareno and Birdsall (eq. 37, 1995) from the allosteric model could only be applied to data derived using amiloride analog concentrations below those concentrations that caused changes in maximal and/or basal levels of [³⁵S]GTPS binding. With amiloride and DMA, a decrease in the maximal stimulation of [³⁵S]GTPS binding by norepinephrine, epinephrine, or UK14304 was only found at 3 mM amiloride analog concentration. The data obtained with concentrations of amiloride or DMA up to 1 mM were well fitted by the equation from the allosteric model (Lazareno and Birdsall, 1995), for their effect on either norepinephrine-, epinephrine-, or UK14304-stimulated [³⁵S]GTPS binding (Fig. 5 and Table 3; not shown). The data were significantly (P < .05) better fitted by the equation derived from the allosteric model (eq. 37, Lazareno and Birdsall, 1995) than by the equation for the competitive model (eq. 37, b = 0, Lazareno and Birdsall, 1995). The analyses by the allosteric model give separate estimates of both the log affinity of the amilorides at the unoccupied receptor (log K_X) and the log cooperativity between the amilorides and the agonists (log ) (Table 3), unlike a dissociation assay that gives only estimates of a single combined parameter, the log affinity of the allosteric ligand at the radioligand-occupied receptor (log K_X).

View larger version (19K): [in this window] [in a new window]   Fig. 5.   Effect of amiloride analogs on the stimulation by norepinephrine of [35S]GTPS binding to membranes containing the 2A-adrenergic receptor. In each set of 24 tubes, membranes were incubated at 30°C for 60 min with GDP, [35S]GTPS, sodium metabisulfite, a given concentration of amiloride analog, and an increasing concentration of norepinephrine in duplicate. Individual data points from one experiment are shown per graph. The data from each experiment were simultaneously fitted to the equation from the allosteric model (eq. 37, Lazareno and Birdsall, 1995), with time and amiloride analog concentration as independent variables. The results are summarized in Table 3.

View this table: [in this window] [in a new window]   TABLE 3 Amiloride analog modulation of the stimulation by agonists of [35S]GTPS binding to membranes containing the 2A-adrenergic receptor In each set of 24 tubes, membranes were incubated at 30°C for 60 min with GDP, [35S]GTPS, sodium metabisulfite (norepinephrine only), a given concentration of amiloride analog, and an increasing concentration of agonist in duplicate. The data from each experiment were simultaneously fitted by the equation derived in a previous study (eq. 37, Lazareno and Birdsall, 1995), with agonist and amiloride concentrations as independent variables, to give the log EC50, log KX, and log . The log EC50 is the log concentration of agonist needed to give half-maximal stimulation of [35S]GTPS binding in the absence of amiloride analog, log KX and log KX (the sum of the estimated values of log KX and log ) are the affinities of the amiloride analogs at the unoccupied and agonist-occupied receptor, respectively, and log  is the cooperativity factor for the interaction between the agonist and the amiloride analog. Values are means ± S.E. of n experiments.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

Previous studies have shown that the _2A-adrenergic receptor possesses a well defined allosteric site, at which amilorides can act to modulate antagonist binding at the primary binding site (Leppik et al., 1998a). This differs from the situation found for the _1A-adrenergic receptor, for which the data are compatible with two, but not one, allosteric sites at which amilorides can bind and modulate antagonist binding (Leppik et al., 2000). In this paper, we quantitate the behavior of amilorides, acting via the allosteric site on the _2A-adrenergic receptor, to modulate agonist binding and function.

Initially, competition studies were carried out between the potent, ₂-selective agonist [³H]UK14304 and the amilorides at the _2A-adrenergic receptor. The affinities calculated for the amilorides at the unoccupied receptor are in good agreement with those previously determined from competition studies versus the antagonist [³H]yohimbine (Leppik et al., 1998a). This argues that the amilorides do not have additional potent effects on G proteins that disrupt receptor-G protein interactions. These data are compatible with either competition between [³H]UK14304 and the amilorides at the agonist binding site or negative allosteric modulation ( < 0.05) by the amilorides of agonist binding via an allosteric site.

In kinetic studies, the amilorides examined caused modest concentration-dependent decreases in the dissociation rate of the [³H]UK14304 from the _2A-adrenergic receptor. The decreases were so slight that only HMA showed a large enough effect to enable the dissociation data to be fitted by the equation derived from the ternary complex allosteric model (Fig. 3). For the other amilorides, the parameters defining the allosteric interaction could only be derived by analyzing the effect on the [³H]UK14304 dissociation rate of the competition between each amiloride and HMA at the allosteric site. This was done for both amiloride and DMA, the two other amilorides successfully studied in the [³⁵S]GTPS assay, and the data obtained could be well fitted (Fig. 4) by the relevant equation (Leppik et al., 1998a). The reversal of the kinetic effects of HMA by amiloride and DMA, in a manner compatible with the allosteric model, argues against their interactions being nonspecific in nature.

The [³H]UK14304 dissociation rates from the amiloride-, DMA-, or HMA-occupied receptors were estimated to be 1.5-, 1.3-, or 2.7-fold slower, respectively, than from the unoccupied receptor (Table 1). These slight decreases are in contrast to the larger increases caused by the amiloride analogs (especially HMA) on antagonist dissociation rates (Leppik et al., 1998a). For example, the [³H]yohimbine dissociation rates from the amiloride-, DMA-, or HMA-occupied receptors were 2-, 5-, and 140-fold faster, respectively, than that from the unoccupied receptor (Leppik et al., 1998a).

From the analysis of the kinetic data, the calculated affinities of the amilorides at the [³H]UK14304-occupied receptor were found to correlate with the size of the 5-N-alkyl side chain. Amiloride, with only a 5-amino function (Fig. 2), has a log affinity of 2.38 (1/K_X: 4 mM), whereas DMA, which has a 5-dimethylamino function, has a log affinity of 3.39 (1/K_X: 400 µM), and HMA, with a hexamethylene ring on the 5-amino function, has a log affinity of 4.24 (1/K_X: 60 µM; Table 1). These values parallel the rank order of affinities of the amilorides for the unoccupied _2A-adrenergic receptor but are ~100-fold weaker. The differences between the log affinities at the [³H]UK14304-occupied and unoccupied receptors, the log cooperativities, range from 2.2 for amiloride to 2.5 for HMA (Table 1).

The values in Table 1 contrast with those found for the amilorides at antagonist-occupied _2A-adrenergic receptors (Leppik et al., 1998a). There, not only do the log affinities not correlate with the size of the 5-N-alkyl side chain but they also are found in a narrow range (2.1-2.5), whereas the observed log cooperativities vary considerably (2.4 to 4.2). In the current study, it is the log affinities at the [³H]UK14304-occupied receptor that vary (2.4-4.2), whereas the log cooperativities are approximately constant.

The kinetic data suggest that the apparent association rate constants of [³H]UK14304 for the receptor occupied by one of the amilorides (.K_L.[k₂ or k₃]) varies only over a narrow range, and is ~500-fold slower than the association rate constant of [³H]UK14304 for the unliganded receptor (K_L.k₁); i.e., (.[k₂ or k₃])/k₁ = 0.0012 to 0.0061 (Table 1). The magnitude of the decrease of the association rate constant is comparable to that found in the antagonist-amiloride analog kinetic studies [(.[k₂ or k₃])/k₁ = 0.0030-0.0088 (Leppik et al., 1998a)] and implies that the amilorides examined have similar slowing effects on the association rates of agonists or antagonists but very different effects on the dissociation of either agonists or antagonists from the ternary complex.

The [³⁵S]GTPS functional assay was chosen because it is a direct measure of the G protein activation caused by receptor-agonist interactions and is performed under the same assay conditions as used in the ligand binding studies reported above. The assay has proved its worth in functional allosteric studies with muscarinic receptors (Lazareno and Birdsall, 1995; Birdsall et al., 1999), and it has also been used recently in studies of agonist activation of recombinant ₂-adrenergic receptor subtypes (Jasper et al., 1998; Peltonen et al., 1998). In the current study, using membranes expressing a high level of the _2A-receptor, norepinephrine, epinephrine, and UK14304 were found to be full agonists, whereas guanabenz, p-aminoclonidine, and clonidine are partial agonists (Table 2). For all agonists, the slope factors were less than 1, reflecting a complex situation, possibly the interaction of the _2A-adrenergic receptor with more than one G protein.

When the effects of the amilorides on agonist function were explored, all of the amiloride analogs tested were found to cause concentration-dependent decreases in the potencies of either norepinephrine, epinephrine, or UK14304 to stimulate [³⁵S]GTPS binding. However, because high concentrations of the amilorides were found to have deleterious effects in the assay with any of the agonists, only the data obtained with amiloride, DMA, or HMA were amenable to analysis by the equation derived from the allosteric model (Lazareno and Birdsall, 1995). For these three amilorides, good fits of the data were obtained (Fig. 5 and Table 3). As predicted by the ternary complex allosteric model, the log affinities of the amilorides at the unoccupied receptor (Table 3) are in good agreement with the values obtained from competition equilibrium experiments versus [³H]UK14304 or ³H-antagonists (Leppik et al., 1998a), whereas the log affinity values at the UK14304-occupied receptor and the cooperativities are in reasonably good agreement with the values obtained from the kinetic studies (Table 1). This suggests that the magnitude of the heterotropic cooperativity between agonist and an amiloride is independent of receptor-G protein coupling, because the functional assays are carried out in the presence of 1 µM GDP, whereas the [³H]UK14304 kinetic studies are carried out at the high-affinity [³H]UK14304-receptor-G protein complex and in the absence of added guanyl nucleotides.

It is of interest to note that, with the [³⁵S]GTPS assay, the affinities of each particular amiloride at either the UK14304-occupied or norepinephrine-occupied receptors are essentially the same. A similar situation was found with three different antagonists, for which the affinities of the amilorides at the antagonist-occupied receptors were only slightly affected by which antagonist was used (Leppik et al., 1998a). It thus suggests that the binding mode of the amilorides at the agonist-occupied and antagonist-occupied _2A-adrenergic receptor are very different, and that the actual structures of the agonists or antagonists examined have only a marginal effect on that interaction.

The modulation of [³H]UK14304 dissociation by the amilorides is only compatible with the allosteric model (Fig. 1), and the data from the [³⁵S]GTPS experiments were significantly better fitted by the equation derived from the allosteric model (eq. 37, Lazareno and Birdsall, 1995) than that from a competitive model (eq. 37, b = 0, Lazareno and Birdsall, 1995). The agreement in the values of the parameters derived from the two approaches argues that the amilorides are acting via the one allosteric site to modulate both agonist dissociation and function.

The results cannot exclude the possibility that the amilorides can compete for the agonist binding site at low concentrations and bind at the allosteric site only at much higher concentrations. This is a problem intrinsic to cooperative interactions. The model (Fig. 1) only specifies that a binary complex XR is formed and does not specify to which site X is binding. Equally the formation of XRL only specifies that the two ligands bind simultaneously and does not say to which site X or L are binding. Any model with X capable of binding either to the competitive or allosteric site is formally indistinguishable from the simple allosteric model in equilibrium or kinetic studies. To determine whether the amilorides can bind to the agonist binding site would require alternative approaches.

It has previously been demonstrated that allosteric agents can modulate agonist binding and function at muscarinic receptor subtypes. Thus, gallamine was found to exhibit ~300-fold negative cooperativity on both the function and affinity of acetylcholine at M₂ muscarinic receptors (Stockton et al., 1983; Lazareno and Birdsall, 1995). Recently, compounds related to brucine have been shown to exhibit positive, neutral, or negative cooperativity on acetylcholine binding and function, depending on which brucine analog and which muscarinic receptor subtype were being studied (Birdsall et al., 1997, 1999; Lazareno et al., 1998). It has therefore been possible to discover allosteric enhancers even when the original allosteric ligands exhibited strong negative cooperativity. The finding of only 60-fold negative cooperativity between amiloride and the endogenous ligand norepinephrine at the _2A-adrenergic receptor means that it is possible that allosteric enhancers at this receptor also await discovery.

It has been demonstrated recently that the benzodiazepines, lorazepam and midazolam, can act as weak agonists of very low intrinsic activity at the three ₁-adrenergic receptor subtypes and that they increase the maximal inositol phosphate response of phenylephrine, clonidine, or epinephrine at one or more subtypes (Waugh et al., 1999). These actions are ascribed to an allosteric mechanism. However, the data from the binding and functional studies are not compatible with the predictions of the ternary complex allosteric model (Fig. 1) and suggest that a more complex mechanism is involved. Thus, the work reported here is the first clear demonstration of allosteric modulation of both agonist binding and function at an adrenergic receptor subtype.