Subtype Differences in Pre-Coupling of Muscarinic Acetylcholine Receptors

Based on the kinetics of interaction between a receptor and G-protein, a myriad of possibilities may result. Two extreme cases are represented by: 1/Collision coupling, where an agonist binds to the free receptor and then the agonist-receptor complex “collides” with the free G-protein. 2/Pre-coupling, where stable receptor/G-protein complexes exist in the absence of agonist. Pre-coupling plays an important role in the kinetics of signal transduction. Odd-numbered muscarinic acetylcholine receptors preferentially couple to Gq/11, while even-numbered receptors prefer coupling to Gi/o. We analyzed the coupling status of the various subtypes of muscarinic receptors with preferential and non-preferential G-proteins. The magnitude of receptor-G-protein coupling was determined by the proportion of receptors existing in the agonist high-affinity binding conformation. Antibodies directed against the C-terminus of the α-subunits of the individual G-proteins were used to interfere with receptor-G-protein coupling. Effects of mutations and expression level on receptor-G-protein coupling were also investigated. Tested agonists displayed biphasic competition curves with the antagonist [3H]-N-methylscopolamine. Antibodies directed against the C-terminus of the α-subunits of the preferential G-protein decreased the proportion of high-affinity sites, and mutations at the receptor-G-protein interface abolished agonist high-affinity binding. In contrast, mutations that prevent receptor activation had no effect. Expression level of preferential G-proteins had no effect on pre-coupling to non-preferential G-proteins. Our data show that all subtypes of muscarinic receptors pre-couple with their preferential classes of G-proteins, but only M1 and M3 receptors also pre-couple with non-preferential Gi/o G-proteins. Pre-coupling is not dependent on agonist efficacy nor on receptor activation. The ultimate mode of coupling is therefore dictated by a combination of the receptor subtype and the class of G-protein.


Introduction
G-protein coupled receptors (GPCR) represent the largest family of receptors, with more than 900 encoding genes [1]. They process and transduce a multitude of signals elicited by hormones, neurotransmitter and odorants and are thus involved in a very wide array of physiological and pathological processes. This makes this class of receptors a major pharmacological target for drug development [2].
Agonist-stimulated GPCRs in turn activate heterotrimeric GTP-binding proteins (G-proteins) that activate various signaling pathways. Two distinctive types of interaction between a receptor and G-protein exist: collision coupling and pre-coupling. In the former case, an agonist binds to the free receptor, activates it and then the receptor with bound agonist ''collides'' with free Gprotein and activates it. In the latter case, stable receptor-Gprotein complexes exist in the absence of agonist, agonist binds to this complex, induces change in the receptor conformation that leads to G-protein activation and dissociation of the complex [3]. It should, however, be noted that the distinction between collision coupling and pre-coupling is rather a matter of kinetics of receptor-G-protein interaction, activation state and receptor to Gprotein stoichiometry [4]. Additional modes of interaction intermediate between pure collision coupling and pre-coupling, like transient receptor to G-protein complexing (''dynamic scaffolding''), have been observed [5].
There is accumulating evidence for both collision coupling and pre-coupling of GPCRs. Interestingly, coimmunoprecipitation studies showed pre-coupling of a 2A -adrenergic receptors [6] with G i/o G-proteins and b 2 -adrenergic receptors with G s/olf G-proteins [7]. In contrast, rapid collision coupling of G-proteins with a 2Aadrenergic receptors has been demonstrated in resonance energy transfer studies [8] and with b 2 -adrenergic receptors in living cell imaging studies [9]. Overall, current data on GPCR coupling suggest that the mode of receptor to G-protein coupling may differ depending on the receptor type, cell type and membrane composition [3,10]. Thus, understanding the dynamic behavior of GPCR systems including receptor-G-protein coupling is important in discovery and development of more organ-specific drugs.
Muscarinic acetylcholine receptors are GPCRs present at synapses of the central and peripheral nervous systems but also exist in non-innervated cells and tissues. There are five subtypes of muscarinic receptors encoded by distinct genes without splicing variants [11]. Development of selective ligands for muscarinic receptors thus represents an enormous challenge due to their omnipresence, with only a few types of tissues being endowed by a single or predominant subtype of these receptors. So far very little is known about the nature of coupling of muscarinic receptors to G-proteins [12]. We have demonstrated that the M 2 receptor can directly activate all three classes of G-proteins [13], and that it probably pre-couple to G i/o but not to G s/olf G-proteins [14]. To further clarify the mechanisms of muscarinic receptor subtypes signaling we analyzed the mode of coupling of M 1 through M 4 muscarinic receptors with G i/o , G s/olf and G q/11 G-proteins in membranes from Chinese hamster ovary cells expressing individual receptor subtypes. We show that while M 1 and M 3 receptors pre-couple both with their preferential G q/11 and non-preferential G i/o G-proteins, M 2 and M 4 receptors pre-couple only to preferential G i/o G-proteins.

Stimulation of [ 35 S]GTPcS binding to G i/o , G s/olf and G q/11 G-proteins
Membranes from CHO cells containing from 1.4 to 2.5 fmol of M 1 through M 4 muscarinic receptors per mg of protein were exposed to carbachol in concentrations ranging from 0.1 mM to 1 mM and binding of [ 35 S]GTPcS to G-protein classes was determined using a scintillation proximity assay (SPA) (Fig. 1). Carbachol stimulated [ 35 S]GTPcS binding to all three major classes of G-proteins via all four receptors, with highest potency (EC 50 about 1 mM) and efficacy (more than 3-fold increase over basal) for preferential G-proteins (G q/11 for M 1 and M 3 and G i/o for M 2 and M 4 receptors) ( Table 1). The potency of carbachol in stimulating [ 35 S]GTPcS binding to non-preferential G-proteins was 2-(M 3 G s / olf ) to 10-fold (M 2 G s/olf ) lower than to preferential G-proteins.

Competition of carbachol with [ 3 H]NMS binding at M 1 through M 4 receptors
Binding of the tritiated antagonist N-metylscopolamine ([ 3 H]NMS) in the presence of agonist carbachol concentrations ranging from 10 nM to 10 mM (Fig. 2) was best described by competition for two sites (Eq. 3) at all four receptor subtypes. The equilibrium inhibition constant (K I ) of carbachol was similar among receptor subtypes, both for high and low affinity sites ( Table 2). At M 1 and M 3 receptors that preferentially couple to G q/11 G-proteins carbachol displayed more low affinity binding sites than at M 2 and M 4 receptors that preferentially couple to G i/o G-proteins. In some cases preincubation of membranes with antibodies directed against the C-termini of a-subunits of individual classes of G-proteins led to an increase in the proportion of low affinity sites. The proportion of low affinity sites was increased by anti-G i/o and anti-G q/11 antibodies at M 1 and M 3 receptors but only by anti-G i/o antibodies at M 2 and M 4 receptors. The anti-G s/olf antibody did not change the proportion of low affinity sites at any receptor subtype. None of the antibodies affected K I of either the low or high affinity sites.

Competition of agonists with [ 3 H]NMS binding at M 1 and M 2 receptors
All tested agonists at M 1 receptors (carbachol, furmethide, oxotremorine, and pilocarpine) bound to two binding sites (Fig. 3 full circles). Although they bound with different affinities they recognized the same proportion of low-affinity sites (Table 3). Anti-G i/o and anti-G q/11 antibodies increased the proportion of low affinity sites to a comparable extent for all tested agonists (Fig. 3, open circles and open diamonds). The anti-G s/olf antibody did not change the proportion of the low-affinity binding sites for any of the agonists tested (Fig. 3, open squares). None of the antibodies affected K I values.
Similarly, all tested agonists bound to two binding sites at M 2 receptors (Fig. 4, full circles). As in the case of the M 1 receptor they bound with different affinities but they recognized the same proportion of low-affinity sites (Table 4). Similar to carbachol, the proportion of low-affinity sites was lower at M 2 than at M 1 receptors for all tested agonists (Table 4 vs. Table 3) and only the anti-G i/o antibody increased the proportion of low affinity sites (Fig. 4, open circles). The anti-G s/olf and anti G q/11 antibodies did not change either the proportion of low-affinity binding sites or K I s for any of the agonists tested (Fig. 4, open squares and open diamonds).

Effects of mutations of M 1 receptors that affect receptor activation
To further investigate the role of receptor activation in receptor-G-protein pre-coupling we prepared cell lines expressing mutant M 1 receptor with mutations known to interfere with receptor signaling. Mutation of aspartate 71 in the middle of the second transmembrane domain to asparagine (D71N) has been shown to abolish receptor activation [15]. Mutation of aspartate 122 in the conserved E/DRY-motif at the intracellular edge of the third transmembrane domain to asparagine (D122N) has been shown to reduce the potency of muscarinic agonists [16]. Opsin arginine in the conserved E/DRY-motif at the intracellular edge of the third transmembrane domain of has been shown to directly interact with the C-terminal cysteine of the a-subunit of G-protein [16]. At M 1 muscarinic receptors mutation of corresponding arginine 123 asparagine (R123N) blocks activation of G-proteins [17]. The appropriate control CHO cell line expressing the wild-type receptor was also generated using the same expression vector. Expression levels of receptor mutants    Table 5).
On the newly prepared cell line expressing M 1 wt receptors carbachol displayed binding to two binding sites in competition with [ 3 H]NMS with the same proportion of low affinity sites and similar affinities (Fig. 6, full circles) as in Fig. 2. While mutation D71N did not change the binding parameters of carbachol, mutation D122N brought about an increase in low affinity sites and mutation R123N completely abolished high-affinity binding ( Fig. 6 and Table 6).  Table 3. doi:10.1371/journal.pone.0027732.g003   Table 7 vs. Table 1) and decreased its efficacy more than 5-times. The efficacy of carbachol in stimulation of [ 35 S]GTPcS binding to G s/olf or G q/11 G-proteins was unchanged while its potency increased about 3-times in both cases.
Based on competition binding of agonists and [ 3 H]NMS ( Fig. 8; Table 8), attenuation of G i/o expression led to an increase in the proportion of low-affinity sites for all tested agonists (see controls in Table 4 and Table 8) without change in K I values. The anti-G i/o antibody further increased the proportion of low-affinity sites in G i/o G-proteins-depleted membranes only for the full agonists carbachol and furmethide. In contrast to control M 2 -CHO cells, the proportion of low-affinity sites of the partial agonists oxotremorine and pilocarpine in G i/o G-proteins-depleted mem-   branes was not changed by the anti-G i/o antibody. Similar to untreated M 2 -CHO cells, the anti-G s/olf and anti-G q/11 antibodies had no effect on either the proportion of low affinity sites or K I values in membranes with attenuated expression of G i/o Gproteins.

Discussion
Binding of an agonist to a G-protein-coupled receptor (GPCR) results in transforming the receptor to an active state that facilitates guanosine diphosphate (GDP) dissociation from the asubunit of interacting heterotrimeric G-proteins and its exchange for guanosine trisphosphate (GTP) [18]. In principle there are many possible ways for receptor-G-protein interactions to take effect, with two extreme possibilities. In one scenario receptors and G-proteins diffuse freely within the plasma membrane, agonist binds to the free receptor that then randomly ''collides'' with Gproteins and activates them. Alternatively, receptors and Gproteins form stable complexes regardless of the receptor activation state and agonist binding, the agonist binds to this complex and induces conformational changes in the receptor protein that leads to G-protein activation and dissipation of the receptor-G-protein complex. We will refer to the former situation as ''collision coupling'' and the latter one as ''pre-coupling'' [3]. It should, however, be noted that even if receptors are partially precoupled to G-proteins, an agonist can also bind to free receptors and then ''collide'' with G-protein. Also the distinction between collision coupling and pre-coupling is rather a matter of kinetics of   receptor-G-protein interaction and activation and receptor to Gprotein stoichiometry [4]. Thus a myriad of possible ways for interaction among receptor, G-protein and agonist exist, e.g., transient receptor to G-protein complexing (''dynamic scaffolding'') [5]. Receptor G-protein pre-coupling plays an important role in signaling. It may accelerate kinetics of signal transduction. If receptor G-protein complexes pre-exist, instantaneous activation of G-protein takes place upon agonist binding to the receptor [4]. As shown repeatedly [10,13,19,20] and also in Fig. 1, muscarinic acetylcholine receptors couple with all 3 major classes of G-proteins (G i/o , G s/olf and G q/11 ). Our recent data show that at M 2 receptors the agonist carbachol slows down the association of GDP with G i/o but not G s/olf G-proteins [14]. This finding may evidence the pre-existence of a receptor/G-protein complex prior to carbachol binding. Data thus suggest that muscarinic M 2 receptors pre-couple to G i/o but not to G s/olf G-proteins. Alternatively, M 2 receptors may precouple with G s/olf but carbachol has no effect on GDP association. To exclude this possibility we analyzed in detail pre-coupling of all 3 major classes of G-proteins with M 2 receptors and compared it with precoupling at other G i/o preferring (M 4 ) and G q/11 preferring (M 1 and M 3 ) muscarinic receptors.
At all receptor subtypes carbachol displays a ''two site'' binding curve with high affinity binding in the nanomolar range and low affinity binding in the micromolar range (Fig. 2, Table 2). According to the ternary complex model of GPCRs [21] agonists bind with high affinity to the receptor-G-protein complex and with low-affinity to receptors uncoupled from G-proteins. The interface of interaction between the receptor and G-protein consists of the intracellular edge of the third, fifth and sixth transmembrane domains and adjacent parts of the third and fourth intracellular loops of the receptor and the C-terminus of the G-protein asubunit [16,22]. Antibodies directed against the C-terminus of Gprotein should prevent receptor-G-protein interaction (or break existing receptor G-protein complex) and lower the affinity of the receptor for agonists. Indeed, IgG antibodies directed against the C-terminus of the G i/o class of G-proteins increased the fraction of low-affinity sites at all receptor subtypes including the G i/o nonpreferring M 1 and M 3 receptors. Similarly, IgG antibodies directed against the C-terminus of G q/11 class of G-proteins increased the fraction of low-affinity sites only at their preferring M 1 and M 3 receptors, but IgG antibodies directed against the Cterminus of G s/olf class of G-proteins had no effect. Antibodies only changed the proportion of low-affinity sites without an effect on receptor affinity. Our data also show that all receptors precouple with their preferential G-proteins (M 1 and M 3 with G q/11 and M 2 and M 4 with G i/o ) and that M 1 and M 3 receptors also precouple with non-preferential G i/o G-proteins. In contrast, precoupling of G s/olf G-proteins was not detected at any subtype of muscarinic receptors. In other words, the interaction between receptor and G s/olf is so short-lived that cannot be detected by antibodies. This is in agreement with our kinetic measurements at G s/olf and M 2 receptors [14]. We tested binding of four structurally different agonists (carbachol, furmethide, oxotremorine and pilocarpine) that also differ in potency and efficacy in activating muscarinic receptors [14] and binding kinetics [23]. Importantly, all tested agonists recognize the same proportion of low-affinity binding sites (Figs 3  and 4; Tables 3 and 4). It is very unlikely that these agonists induce the same proportion of transient high-affinity states (in collision coupling, dynamic scaffolding or a similar scenario). Rather receptor G-protein complexes preexist prior to agonists binding and their proportion is given by stoichiometry of receptors and Gproteins. Moreover, the antibodies have the same effect on binding of all agonists, further excluding the role of agonists in the formation of receptor-G-protein complexes.
To further investigate the role of receptor activation in receptor-G-protein pre-coupling we prepared cell lines expressing mutant M 1 receptor with mutations known to interfere with receptor signaling. Mutation of aspartate 71 in the middle of the second transmembrane domain to asparagine (D71N) has been shown to prevent activation of the M 1 receptor [15]. This residue is neither part of the agonist binding site nor the receptor-G-protein interface. It is supposed that the D71N mutation disrupts intramolecular hydrogen bond network and prevents the receptor from gaining an active conformation. Its effect on suppressing receptor activation is confirmed in Fig. 5 (upper right). Although D71N receptors cannot be activated by carbachol they still display both high-and low-affinity sites for carbachol with the same proportion as control wild-type M 1 receptors (Fig. 6, full and open circles). Thus, an active conformation of the receptor is not a prerequisite for receptor-G protein pre-coupling. These data are in perfect fit with a report by Quin et al. [24] published during completion of this manuscript that M 3 receptors form inactive complexes with G q G-proteins.
In contrast, mutations that interfere with receptor signaling by being directed against the receptor G-protein interface do affect pre-coupling. Mutation of aspartate 122 in the conserved E/DRYmotif at the intracellular edge of the third transmembrane domain to asparagine (D122N) has been shown to reduce the potency of muscarinic agonists [13]. Measurements of the association rate of GTPcS shows that at D122N receptors GTPcS binding under basal conditions (in the absence of agonist) is accelerated and that carbachol has smaller effect on GTPcS association rate than at wild-type M 1 receptors (Fig. 5, lower left). Meanwhile, the proportion of low-affinity sites for carbachol is increased in D122N receptors in comparison with control (Fig. 6). Most likely, increased basal activity of D122N results in more activated Gproteins and thus more uncoupled receptors in membrane preparations. Crystal structure of complex of opsin and Cterminus of G-protein a-subunit revealed that arginine in the conserved E/DRY-motif at the intracellular edge of the third transmembrane domain interacts directly with the a-subunit Cterminal cysteine [16]. At M 1 muscarinic receptors mutation of arginine 123 to asparagine (R123N) blocks activation of Gproteins (Fig. 5, lower right). In accordance with the ternary complex model of GPCRs [21], R123N receptors (uncoupled from G-proteins) display only low-affinity for carbachol (Fig. 6,  triangles).
It is worth noting that carbachol can activate all three classes of G-proteins at both M 1 and M 2 receptors (Fig. 1) and M 1 receptors pre-couple both to preferential G q/11 and non-preferential G i/o Gproteins. In contrast, M 2 receptors pre-couple only to preferential G i/o G-proteins. G i/o are the major class of G-proteins in membranes from CHO cells, representing almost half of all Gproteins. To exclude the possibility that M 2 receptors do not precouple with G q/11 G-proteins due to competition with preferential G i/o G-proteins, we attenuated the expression of G i/o a-subunits by siRNA to one quarter, making G i/o G-proteins the least abundant class in CHO membranes. Such reduction in expression of G i/o G-proteins diminishes the efficacy of carbachol in activating these preferential G-proteins to a level lower than at any of non-preferential G-proteins (Fig. 7). It also reduced its potency (Table 7 vs. Table 1). On the other hand, the potency of carbachol to stimulate GTPcS binding increases at non-preferential G s/olf and G q/11 G-proteins, demonstrating competition among G-proteins for M 2 receptors [25]. In concert, the proportion of low-affinity sites increases and the effect of the anti-G i/o antibody is reduced (Fig. 8, cf. Table 4 and Table 8). Again, these findings indicate the presence of a lower proportion of high-affinity receptor/G-protein complexes. However, the anti-G q/11 antibody has no effect on the proportion of low-affinity sites even after such reduction in the expression of G i/o G-proteins. This suggests that the lack of pre-coupling of G q/11 G-proteins with M 2 receptors is not due to competition with G i/o G-proteins.
In summary, we show that muscarinic receptors pre-couple with their preferential class of G-proteins in the absence of an agonist. In contrast to the M 1 and M 3 receptors that pre-couple both with preferential G q/11 and non-preferential G i/o G-proteins, the M 2 and M 4 receptors pre-couple only with their preferential G i/o G-  Table 5. doi:10.1371/journal.pone.0027732.g007 proteins. Lack of pre-coupling of the M 2 and M 4 receptors to G q/11 G-proteins is not due to competition with preferential G i/o G-proteins. None of the four subtypes of muscarinic receptors precouples to G s/olf G-proteins. Thus, the mode of coupling of a given subtype of muscarinic receptors is governed by a combination of the receptor subtype and the class of G-protein. Advanced instrumental methods like fluorescence resonance energy transfer (FRET) between receptor and G-protein [7] and plasmon surface resonance [5] were developed to monitor kinetics of receptor Gprotein interactions. Although these methods give better picture of receptor G-protein interaction, our simple method, that can only detect pre-coupling, does not require recombinant systems like FRET-based methods nor reconstituted systems like plasmon surface resonance methods and can be easily applied ex vivo, e.g. to tissues of experimental animals.   Table 6. doi:10.1371/journal.pone.0027732.g008 Table 8. Effects of IgG antibodies directed against a-subunits of individual subtypes of G-proteins on binding parameters of muscarinic agonists in membranes of M 2 CHO cells with reduced expression of G i/o G-proteins by siRNA.

Cell culture and membrane preparation
Chinese hamster ovary cells stably transfected with the human M 1 to M 4 muscarinic receptor genes (CHO cells) were kindly donated by Prof. T.I.Bonner (National Institutes of Health, Bethesda, MD). Cell cultures and crude membranes were prepared as described previously [18]. Briefly, cells were grown to confluency in 75 cm 2 flasks in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. Two million of cells were subcultured to 100 mm Petri dishes. Medium was supplemented with 5 mM butyrate for the last 24 hours of cultivation to increase receptor expression. Cells were detached by mild trypsinization on day 5 after subculture. Detached cells were washed twice in 50 ml of phosphate-buffered saline and 3 min centrifugation at 250 x g. Washed cells were suspended in 20 ml of ice-cold incubation medium (100 mM NaCl, 20 mM Na-HEPES, 10 mM MgCl 2 ; pH = 7.4) supplemented with 10 mM EDTA and homogenized on ice by two 30 sec strokes using Polytron homogenizer (Ultra-Turrax; Janke & Kunkel GmbH & Co. KG, IKA-Labortechnik, Staufen, Germany) with a 30-sec pause between strokes. Cell homogenates were centrifuged for 30 min at 30,000 x g. Supernatants were discarded, pellets resuspended in fresh incubation medium and centrifuged again. Resulting membrane pellets were kept at 220uC until assayed within 10 weeks at a maximum.

Preparation of new stable cell lines
New stable CHO cell lines expressing wild-type M 1 and mutant M 1 receptors have been prepared. Coding sequence of wild-type human M 1 muscarinic receptor (in expression vector pcDNA ver. 3.2, cDNA resource center, University of Missouri-Rolla, MO, USA) was mutated by PCR and mismatch primers using Qiagene QuickChange kit. Mutations were verified by sequencing of complete receptor coding sequence. Then CHO-K1 cells were transfected with either original M 1 -pcDNA or mutated plasmid using Lipofectamine 2000 (Lipofectamie 10 ml/ml, DNA 0.5 mg/ml). After 48 hours geneticine was added to cultivation medium to final concentration of 800 mg/ml. After selection, the concentration of geneticine was lowered to 50 mg/ ml and maintained during cultivation.

Equilibrium radioligand binding experiments
All radioligand binding experiments were optimized and carried out as described earlier [20]. Briefly, membranes were incubated in 96-well plates at 30 o C in the incubation medium described above that was supplemented with freshly prepared dithiothreitol at a final concentration of 1 mM.

Scintillation proximity assay
In case of scintillation proximity assay, incubation with [ 35 S]GTPcS as described above was terminated by membrane solubilization by the addition of 20 ml of 10% Nonidet P-40. After 20 min 10 ml of individual primary antibodies against C-termini of G-protein a-subunits were added and incubation was continued for 1 h. The final dilution was 1:500 in case of anti-G i/o -a and anti-G s/olf -a antibodies and 1:1000 in case of the anti-G q/11 -a antibody. One batch of anti-rabbit IgG-coated scintillation beads was diluted in 20 ml of incubation medium and 50 ml of the suspension was added to each well for 3 h. Then plates were spun for 15 min at 1,000 x g and counted using the scintillation proximity assay protocol in a Wallac Microbeta scintillation counter.

Data analysis
In general binding data were analyzed as described previously [20]. Data were preprocessed by Open Office version 3.2 (www. openoffice.org) and subsequently analyzed by Grace version 5.1 (plazma-gate.weizman.ac.il) and statistic package R version 2.13 (www.r-project.org) on Scientific Linux version 6 distribution of GNU/Linux.
The following equations were fitted to data: Saturation of radioligand binding y, binding of radioligand at free concentration of radioligand x; B MAX , maximum binding capacity; K D , equilibrium dissociation constant.

Concentration-response
y~1z(E MAX -1)=(1z(EC 50 =x) nH ) ð2Þ y, radioactivity in the presence of agonist at concentration x normalized to radioactivity in the absence of agonist; E MAX , maximal increase by agonist; EC 50 , concentration of agonist producing 50% of maximal effect; nH, Hill coefficient.

Interference of agonist with [ 3 H]NMS
y~(100-f low ) Ã (1-x=(IC 50high zx))zf low Ã(1-x=(IC 50low zx)) ð3Þ y, binding of radioligand at a concentration of displacer x normalized to binding in the absence of displacer; f low , percentage of low affinity sites; IC 50high , concentration causing 50% decrease in binding to high affinity sites; IC 50low , concentration causing 50% decrease in binding to low affinity sites. Equilibrium dissociation constant of displacer (K I ) was calculated according to Cheng and Prusoff [26]. Rate of association y, binding of radioligand at a time x; B eq , equilibrium binding; k obs , observed rate of association.