Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Muscarinic receptors are widely distributed throughout the body, mediate a variety of functions, and consequently represent potential pharmacological targets in many disorders and diseases (Hulme et al., 1990; Caulfield, 1993; Wess, 1996). Five subtypes of muscarinic receptors are known to exist in mammals (Caulfield and Birdsall, 1998). Currently, there are markedly more subtype-selective muscarinic antagonists available than agonists. There has been active search for potent muscarinic agonists with selectivity for the M₁ receptor subtype with the aim of pharmacological application of such agonists to compensate for acetylcholine deficiency in Alzheimer's disease because the M₁ receptor is important for learning and memory (McKinney and Coyle, 1991). One of the promising compounds is 3-[3-hexyloxy-1,2,5-thiadiazo-4-yl]-1,2,5,6-tetrahydro-1-methylpyridine (xanomeline) (Bodick et al., 1997; Veroff et al., 1998; Shannon et al., 2000). Xanomeline behaves as an agonist with a high potency, and is functionally selective for the M₁ subtype of muscarinic receptors (Ward et al., 1995; Bymaster et al., 1998; Ward et al., 1998; Wood et al., 1999). It has been shown recently in our laboratory that a significant part of xanomeline binding to the M₁ receptors is wash-resistant, resulting in persistent activation of receptors even after extensive washing (Christopoulos et al., 1998).

A hypothetical interpretation of xanomeline binding has been proposed (Christopoulos et al., 1998, 1999, 2000) whereby xanomeline possibly interacts reversibly with the classical binding site on the receptor, but also binds to a secondary site, "exosite", in a wash-resistant manner. The objective of the present study was to carefully test the hypothesis that wash-resistant binding of xanomeline does not take place at the primary binding site on the M₁ muscarinic acetylcholine receptor. Experiments were designed so that one of the two components of xanomeline binding (either the reversible or the wash-resistant one) was measured predominantly at a time. Additional experiments were performed to exclude experimental artifacts that might contribute to xanomeline wash-resistant binding.

The data we obtained confirm the concept that xanomeline displays two substantially different modes of binding to the M₁ muscarinic receptors: 1) reversible binding conforming to an interaction at the receptor primary binding site; and 2) wash-resistant, very stable binding that takes place somewhere else. The latter mode of binding is accompanied by marked allosteric modulation of the interaction of competitive ligands with the receptor.

We had previously speculated that wash-resistant binding of xanomeline to the M₁ muscarinic receptor might take place at an exosite on the receptor, similar to the mode of binding of salmeterol to ₂-adrenergic receptors (Coleman et al., 1996). The present data, however, do not conclusively support this mode of xanomeline binding.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Reagents

[³H]NMS was from PerkinElmer Life Sciences (Boston, MA). Atropine was from Sigma-Aldrich (St. Louis, MO). Xanomeline tartrate was a kind gift from Eli Lilly & Co. (Indianapolis, IN).

Cells and Cell Membranes

Experiments were performed on whole cells or membranes of Chinese hamster ovary (CHO) cells stably transfected with the human gene for muscarinic M₁ receptors (kindly provided by Dr. M. Brann, University of Vermont Medical School, Burlington, VT). Cells were grown in plastic dishes in Dulbecco's modified Eagle's medium with 10% calf serum and 0.005% Geneticin. They were harvested by mild trypsinization 7 days after subculturing, washed twice through centrifugation (3 min at 300g), and resuspended in HEPES medium (110 mM NaCl, 5.3 mM KCl, 1.8 mM CaCl₂, 1 mM MgCl₂, 25 mM glucose, 20 mM HEPES, and 58 mM sucrose, with pH adjusted to 7.4 and osmolarity to 340 mOsM). Membranes were prepared by homogenization with a Polytron PT3000 homogenizer (Kinematica, Littau-Luzern, Switzerland) and centrifugation for 20 min at 30,000g. Pellets were kept frozen at 20°C.

Experimental Design of Binding Experiments

Two main types of experiments were performed to study xanomeline-reversible and wash-resistant binding. In one type of experiment, xanomeline was present during incubation with [³H]NMS. In the first type of binding experiments, xanomeline was present during incubation simultaneously with [³H]NMS. In these experiments, changes in [³H]NMS binding induced by xanomeline were due both to xanomeline, which associated with receptors in a reversible manner, and to xanomeline, which associated in the wash-resistant manner. In the second type of experiments, membranes had been first pretreated with xanomeline, after which they were washed and incubated with [³H]NMS without xanomeline. In these experiments, changes in [³H]NMS binding induced by xanomeline were only due to xanomeline, which was attached to the receptors in the wash-resistant manner.

Pretreatment of Cells or Membranes with Xanomeline

In experiments designed to measure the wash-resistant binding of xanomeline, cells or membranes were diluted in the HEPES medium to a final concentration of 3 to 10 million cells or corresponding membranes per milliliter and incubated with xanomeline at 37°C for 1 h (except measurements of xanomeline association where incubation time varied). Pretreatment was terminated with 10-fold dilution of sample with HEPES medium, and brief centrifugation (cells, 1 min at 1000g; membranes, 10 min at 37,000g). Excess xanomeline was removed by centrifugation and resuspension in 10 times the incubation volume of HEPES medium. After incubation at room temperature for 30 min. Samples were centrifuged and supernatant was discarded. Dilution, incubation, and centrifugation were repeated three times.

Radioligand Binding

Measurements of radioligand binding were performed essentially as described (Jakubík et al., 1995). Cells or membranes corresponding to 300,000 to 1,000,000 cells were incubated at 37°C in a final incubation volume of 0.8 ml. The incubation medium corresponded to the HEPES medium described above, with added ligands as indicated for specific experiments. Atropine (5 µM) was used to determine nonspecific binding of [³H]NMS. The incubation was terminated by filtration through GF/C glass fiber filters (Whatman, Maidstone, UK) in a cell harvester (Brandel, Inc., Gaithersburg, MA), and the radioactivity retained on the filters was measured by liquid scintillation spectrometry.

Saturation Binding Experiments

Membranes were preincubated for 1 h in the absence or presence of xanomeline at concentrations of up to 1 mM. Then membranes were centrifuged and washed three times with fresh HEPES medium. After that, membranes were incubated for 1 h with [³H]NMS (36 pM-4 nM). In another type of experiment, untreated membranes were simultaneously incubated with both [³H]NMS and xanomeline. Equations 1 and 2 under "Treatment of Data" were fitted to the resulting data sets.

Competition-Type Experiments

Membranes were preincubated for 30 min in the absence or presence of xanomeline in concentrations ranging from 1 pM to 3.16 mM. Then the membranes were centrifuged and washed with HEPES medium three times and incubated for 1 h with 125, 250, 500, or 1000 pM [³H]NMS. Alternatively, untreated membranes were incubated with [³H]NMS in the simultaneous presence of increasing concentrations of xanomeline (1 pM-3.16 mM). Equation 3 under Treatment of Data was fitted to the resulting data sets.

Kinetic Experiments

Three kinds of kinetic experiments were performed: 1) kinetics of xanomeline wash-resistant binding; 2) kinetics of xanomeline association to the receptor in the presence of atropine; and 3) kinetics of [³H]NMS association to and dissociation from the receptor in the simultaneous presence of xanomeline, and [³H]NMS association to and dissociation from membranes pretreated with xanomeline.

To measure the rate of the wash-resistant association of xanomeline to the receptors, intact cells were incubated with xanomeline at concentrations up to 0.1 mM for up to 30 min. Then cells were briefly centrifuged and extensively washed three times with HEPES medium. After that, cells were incubated for 1 h with [³H]NMS and the loss of [³H]NMS caused by xanomeline binding was taken as a measure of the wash-resistant association of xanomeline. To measure the dissociation rate of the wash-resistant bound xanomeline, membranes were incubated for 30 min with 0.1 mM xanomeline, centrifuged, and washed three times with HEPES medium. After that, the wash-resistant bound xanomeline was allowed to dissociate in a 50-fold excess of HEPES medium for up to 24 h. After dissociation, membranes were again extensively washed three times with HEPES medium and incubated with [³H]NMS. The binding of [³H]NMS increased in parallel with the dissociation of the wash-resistant bound xanomeline and served as the indicator of its rate.

To measure the effects of the wash-resistant bound xanomeline on [³H]NMS association, membranes were preincubated for 1 h in the absence or presence of xanomeline at concentrations of up to 10 µM. Then they were centrifuged and washed three times with HEPES medium. After that, 125 pM [³H]NMS was added and its association was followed for up to 30 min. Alternatively, [³H]NMS association to untreated membranes was studied in the simultaneous presence of xanomeline. Equations 4 and 5 under "Treatment of Data" were fitted to the resulting data sets. To measure the effects of xanomeline on [³H]NMS dissociation, membranes were preincubated in the absence or presence of xanomeline at concentrations of up to 10 µM. Then they were centrifuged and washed three times with HEPES medium. After that, membranes were labeled with 125 pM [³H]NMS for 1 h. The dissociation of [³H]NMS was started by addition of 1 µM atropine. Alternatively, membranes were prelabeled with [³H]NMS for 1 h, and [³H]NMS dissociation was initiated by atropine added either alone or simultaneously with xanomeline. Equations 6 and 7 under "Treatment of Data" were fitted to the resulting data sets.

To measure the effects of atropine on the association rate of xanomeline wash-resistant binding, membranes were preincubated with atropine for 30 min and then xanomeline was added and allowed to associate for 5 to 30 min. After that, membranes were centrifuged and extensively washed with HEPES medium three times, and incubated for 1 h with [³H]NMS. The rate of the wash-resistant association of xanomeline was determined according to the rate at which the binding of [³H]NMS diminished dependent on the duration of the preincubation with xanomeline.

Treatment of Data

Equations for saturation, association, and dissociation built in GraphPad Prism, version 3.00 (GraphPad Software, San Diego, CA) were fitted to the data. The equations were as follows.

Saturation Binding Experiments. After subtraction of nonspecific binding eqs. 1 and 2 were fitted to the data:

(1)

(2)

where Y is [³H]NMS binding at concentration X of free [³H]NMS, K_d is the equilibrium dissociation constant, and B_MAX is the number of binding sites. Data from the better fit were transformed and plotted as Scatchard plots.

Competition-Type Experiments. After subtraction of nonspecific binding and normalization (to express the binding of [³H]NMS in the presence of xanomeline as percentage of the binding in the absence of xanomeline), eq. 3 was fitted to the data:

(3)

where Y is [³H]NMS binding at logarithm of concentration of xanomeline X, I_MAX is maximal inhibition of binding, and IC₅₀ is the concentration of xanomeline that results in 50% of maximal inhibition.

Kinetic Experiments. Association experiments. After subtraction of nonspecific binding, eqs. 4 and 5 were fitted to the data:

(4)

(5)

where Y is [³H]NMS binding at time X, B_eq is equilibrium binding, and K_obs is the observed association rate.

Dissociation experiments. After subtraction of nonspecific binding eqs. 6 and 7 were fitted to the data:

(6)

(7)

where Y is [³H]NMS binding at time X, Plateau is the wash-resistant component of binding, Span is the reversible component of binding, and K_off is the dissociation rate constant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

The characteristics of xanomeline binding to M₁ muscarinic receptors were determined indirectly according to changes in [³H]NMS binding. Previous reports from this laboratory (Christopoulos et al., 1998, 1999) have shown that a fraction of xanomeline binding is wash-resistant. Therefore, two kinds of experiments were performed to distinguish between the mode of xanomeline reversible and wash-resistant binding. In one type of experiment xanomeline was present during incubation with [³H]NMS and in the other membranes were first pretreated with xanomeline then washed and incubated with only [³H]NMS. Although in the former type of experiment the changes in [³H]NMS binding are due to both xanomeline wash-resistant and reversible binding, the changes in [³H]NMS binding in the latter type of experiment are only due to xanomeline wash-resistant binding, allowing us to distinguish between these two modes of xanomeline binding to M₁ muscarinic receptors. In preliminary experiments, we tested whether the observed xanomeline-persistent binding is an experimental artifact due to the inefficiency of the washing protocol, or is the result of locking residual xanomeline into a high-affinity agonist-binding receptor conformation. Accordingly, membranes were treated with xanomeline, washed, and the supernatant from the last wash was tested for its ability to change [³H]NMS binding. The results of these experiments were negative (data not shown). Furthermore, wash-resistant binding of xanomeline to membranes was found to be unaffected by the addition of a stable GTP analog, GppNHp (data not shown).

Kinetics of Xanomeline Wash-Resistant Binding. Whole cells rather than membranes were used in these experiments to shorten centrifugation times and expedite the washing procedure as much as possible. To measure the association rate of xanomeline-persistent binding to M₁ receptors, intact CHO cells expressing the receptor were incubated with xanomeline at concentrations ranging from 10^5.5 to 10⁴ M for up to 30 min. Then cells were briefly centrifuged and extensively washed three times. After that cells were incubated for 1 h with [³H]NMS and residual [³H]NMS binding was used to measure xanomeline wash-resistant binding. The association of xanomeline was fast (Fig. 1). Thus, xanomeline obliterated from 17 (3.2 µM) to 62% (100 µM) of the binding sites in a wash-resistant manner, even when cells were spun down and washed immediately after xanomeline addition. Fitting eq. 4 to the data in Fig. 1 resulted in an observed association rate (K_obs) of 0.22, 0.27, 0.32, and 0.32 min¹ at 3.2, 10, 32, and 100 µM xanomeline, respectively. Note that the slope of the relationship between the observed rate of association (K_obs) of xanomeline wash-resistant binding and xanomeline concentration is markedly shallower than what would be expected for a simple bimolecular interaction (Fig. 1B). Moreover, there was no increase in value of the observed rate of association when xanomeline concentration was increased from 32 to 100 µM (Fig. 1B).

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Kinetics of xanomeline association to its wash-resistant site. A, CHO cells expressing M1 receptors were incubated with xanomeline at concentrations of 105.5 M (), 105 M (), 104.5 M (), and 104 M () for the time indicated on the abscissa. Then they were briefly centrifuged and washed three times with HEPES medium and incubated with [3H]NMS for 1 h in the absence of xanomeline. [3H]NMS binding is expressed as percentage of [3H]NMS binding to the cells not pretreated with xanomeline but sham-treated in parallel. Data are means ± S.E.M. of three to four experiments performed in triplicate. B, Kobs values for xanomeline were obtained by fitting eq. 4 to the data in A. The dashed line represents theoretical values of Kobs based on Kobs for 105.5 M xanomeline and assumption of a simple bimolecular interaction.

View larger version (11K): [in this window] [in a new window]   Fig. 2.   Kinetics of xanomeline dissociation from its wash-resistant site. Membranes were incubated for 30 min with 10 µM xanomeline then washed three times with HEPES medium and suspended in a 50-fold excess of HEPES medium to allow dissociation. Samples were then incubated with [3H]NMS for 1 h. [3H]NMS binding is expressed as percentage of binding to cells not treated with xanomeline. Data are means ± S.E.M. of three experiments performed in triplicate.

Saturation Binding Experiments. In saturation experiments (Fig. 3), increasing concentrations of [³H]NMS were added to membranes in the continued presence of various concentrations of xanomeline (Fig. 3A) or to membranes that had been pretreated with xanomeline at various concentrations for 30 min, followed by removal of xanomeline (Fig. 3B). In either case, incubation of membranes with [³H]NMS was for 1 h. The results of nonlinear regression analysis of the data in Fig. 3 are summarized in Table 1. Equations describing binding to either one (eq. 1) or two (eq. 2) binding sites were fitted to the data. [³H]NMS binding was adequately described by a single homogeneous population of sites (eq. 1) in all cases except in the presence of 0.1 µM xanomeline, where eq. 2 gave a significantly better fit than eq. 1 (F = 17.47, P < 0.01). Under both experimental conditions, xanomeline decreased maximum binding capacity (B_MAX) and increased equilibrium dissociation constant (K_d; diminished affinity) of receptors for [³H]NMS. Higher concentrations of xanomeline were required in the preincubation/washing paradigm to produce effects similar to those in continued presence of xanomeline (Fig. 3, A and B). Moreover, effect of preincubation/washing with xanomeline on K_d approached a maximal limit (Fig. 3B; Table 1). This phenomenon was not observed in the simultaneous presence of xanomeline and [³H]NMS (Fig. 3A).

View larger version (19K): [in this window] [in a new window]   Fig. 3.   [3H]NMS saturation binding. [3H]NMS binding to membranes was carried out in the absence or presence of various concentrations of xanomeline (A) or membranes that were pretreated with various concentrations of xanomeline for 30 min, and excess of xanomeline removed by washing (B). The membranes were incubated for 1 h with [3H]NMS at concentrations ranging from 36 pM to 4 nM. Data are means of three to six experiments performed in triplicate. Results of nonlinear regression analysis have been summarized in Table 1.

Competition-Type Experiments. In competition-type experiments (Fig. 4), binding of 125, 250, 500, and 1000 pM [³H]NMS was measured either in the presence of increasing concentrations of xanomeline, or after the membranes were pretreated with increasing concentrations of xanomeline for 30 min and washed. In both cases membranes were incubated for 1 h with [³H]NMS. Potency of xanomeline to inhibit [³H]NMS binding when continuously present was approximately 2 orders of magnitude higher than the inhibitory potency of the persistent component of xanomeline binding. Equation 3 under "Treatment of Data" was fitted to the resulting data sets. Results of nonlinear regression analysis are summarized in Table 2. Although increasing [³H]NMS concentration caused a significant increase in xanomeline IC₅₀, this shift was slightly smaller than what would be expected in competitive interaction. Thus, increasing [³H]NMS concentration from 125 to 1000 pM resulted in a change in xanomeline IC₅₀ of 2.24-fold (preincubation/washing) and 2.34-fold (simultaneous presence), which is slightly but significantly lower than a 2.77-fold shift predicted by the Cheng-Prusoff equation (Cheng and Prusoff, 1973) for competitive interaction. Interestingly, although the simultaneous presence of xanomeline caused complete inhibition of [³H]NMS binding, the inhibition of [³H]NMS binding produced by pretreatment with xanomeline was incomplete (92% for 500 and 1000 pM [³H]NMS) (Fig. 4).

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Effects of changing [3H]NMS concentration on the apparent potency of xanomeline to inhibit [3H]NMS binding. The binding of 125 pM (squares), 250 pM (triangles), 500 pM (circles), and 1000 pM (diamonds) [3H]NMS was measured in the presence (closed symbols) of increasing concentrations of xanomeline, or after the membranes have been pretreated (open symbols) with increasing concentrations of xanomeline, followed by washing. Data are means ± S.E.M. of three to four experiments performed in triplicate.

Effects of Xanomeline on Kinetics of [³H]NMS Binding. To measure the effects of xanomeline on [³H]NMS association, membranes were preincubated in the absence or presence of xanomeline at concentrations up to 10 µM. Then membranes were centrifuged and washed three times with buffer. After that, 125 pM [³H]NMS was added and its association was followed for up to 30 min (Fig. 5B). For membranes not pretreated with xanomeline, xanomeline was added simultaneously with [³H]NMS (Fig. 5A). Note that concentrations of xanomeline used in Fig. 5A are lower than those used in Fig. 5B, so that the ratio of the applied concentrations to IC₅₀ values of xanomeline under the two experimental conditions is the same. Also, note that the concentrations of xanomeline (0.1 and 1 µM) used simultaneously with [³H]NMS are expected to exert minimal wash-resistant interactions. Equations 4 and 5 under Treatment of Data were fitted to the resulting data sets. Association of [³H]NMS to receptors in the absence of xanomeline was well described by eq. 4 as an association with a single homogeneous population of binding sites. According to two-way analysis of variance, both membrane preincubation with xanomeline and xanomeline presence during incubation of membranes with [³H]NMS significantly slowed the rate of [³H]NMS association. When present simultaneously with [³H]NMS, xanomeline at concentration of 0.1 µM drug (i.e., at a concentration corresponding to its IC₅₀ under these conditions) slowed the rate of [³H]NMS association by 2.8-fold. This is consistent with competition between xanomeline and [³H]NMS for the receptor. In contrast, the wash-resistant bound xanomeline (at 10 µM, i.e., a concentration corresponding to the IC₅₀ of the wash-resistant component) slowed the rate of [³H]NMS association by only 16%. This is in spite of decreasing [³H]NMS equilibrium binding by 45%. Taken together, this is clearly inconsistent with a competitive mode of inhibition of [³H]NMS binding by prebound xanomeline.

View larger version (20K): [in this window] [in a new window]   Fig. 5.   Effects of xanomeline on [3H]NMS association. Measurement of the rate of [3H]NMS association to membranes was carried out either in the presence of various concentrations of xanomeline (A) or on membranes that had been pretreated with various concentrations of xanomeline, followed by washing (B). The concentration of [3H]NMS was 250 pM. Data are means ± S.E.M. of three to six experiments performed in triplicate.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Effects of xanomeline on [3H]NMS dissociation. Measurement of the rate of [3H]NMS dissociation from the membranes was carried out either in the presence of various concentrations of xanomeline (A) or on membranes that had been pretreated with various concentrations of xanomeline and washed (B). Membranes were incubated with 250 pM [3H]NMS for 1 h and dissociation was initiated (time 0) by addition of a mixture of 1 µM atropine and xanomeline (A) or atropine alone (B). Incubation was terminated at times indicated on abscissa. Ordinate, percentage of binding before the start of dissociation. In B, preincubation of membranes with 1 or 10 µM xanomeline followed by washing resulted in a decrease of [3H]NMS binding by 1 and 48%, respectively. Data are means ± S.E.M. of three experiments performed in triplicate.

Effects of Atropine on Xanomeline Wash-Resistant Binding. To measure the effects of occupancy of the classical primary receptor binding site on the association rate of xanomeline wash-resistant binding, membranes were preincubated with receptor-saturating and -supersaturating concentrations of atropine for 30 min then xanomeline was added and allowed to associate for up to 30 min. After that, membranes were centrifuged, extensively washed three times with buffer, and incubated for 1 h with [³H]NMS. The results are summarized in Fig. 7. Atropine (1 µM) slowed the association of both 10 and 100 µM xanomeline with its wash-resistant binding site, but did not prevent it. Increasing the concentration of atropine over 1 µM did not cause further slowing of xanomeline association. Thus, xanomeline is able to gain access to its wash-resistant binding site, even if the primary receptor-binding site is completely blocked.

View larger version (24K): [in this window] [in a new window]   Fig. 7.   Effect of atropine on the rate of xanomeline association with its wash-resistant site. Membranes were preincubated for 30 min with atropine at the indicated concentrations. Then membranes were exposed to 10 µM (A) or 100 µM (B) xanomeline (in the continuing presence of atropine) for the times indicated on the abscissa. After extensive washing to remove both atropine and xanomeline, membranes were incubated for 1 h with [3H]NMS. Abscissa, duration of preincubation with xanomeline. Ordinate, [3H]NMS binding to membranes pretreated with atropine and xanomeline, expressed as percentage of binding to membranes pretreated with atropine alone. Data are means ± S.E.M. of three experiments performed in triplicate.

View larger version (14K): [in this window] [in a new window]   Fig. 8.   Concentration dependence of the wash-resistant binding of xanomeline on the presence of atropine. Membranes were preincubated for 30 min without or with 1, 10, or 100 µM atropine. Then xanomeline was added to the membranes at different concentrations and preincubation continued for additional 10 min. After extensive washing to remove both atropine and reversibly bound xanomeline, membranes were incubated for 1 h with [3H]NMS. Data are means ± S.E.M. of three experiments performed in triplicate.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Previous work from this laboratory unfolded unique features of the interaction of xanomeline with the M₁ subtype of acetylcholine muscarinic receptors. Although the majority of xanomeline binding to this receptor is readily reversible, a distinct binding component is wash-resistant (Christopoulos et al., 1998). The wash-resistant binding of xanomeline results in persistent activation of the receptor in an atropine-sensitive manner. Previous data were tentatively interpreted in terms of an interaction of xanomeline with a hypothetical secondary binding site on the receptor (an exosite), from which xanomeline interacts with the classical binding site to cause long-lasting receptor activation. In the present work we investigated the properties of the persistent binding of xanomeline in more depth. The main results of our experiments may be briefly summarized as follows: Wash-resistant binding of xanomeline is fast and occurs with an apparent affinity in the micromolar range. Its dissociation is extremely slow, with a half-life of over 30 h. Inhibition of binding of the classical competitive antagonist NMS by persistently bound xanomeline was accompanied by a slight slowing down of the rate of NMS association and marked enhancement of its rate of dissociation. There was also an apparent decrease in the number of NMS binding sites in saturation experiments. Atropine bound to the orthosteric binding site only slightly hindered the wash-resistant association of xanomeline with the receptor, and this effect reached a limit regardless of further increasing atropine concentrations.

Taken together, our data provide strong support to the notion that xanomeline-persistent binding does not take place at the classical primary binding site on the M₁ muscarinic receptor. More importantly, this mode of binding results in allosteric modulation of interaction of competitive ligands at the receptor primary binding domain.

Control experiments were designed to eliminate the possibility that residual, unwashed xanomeline contributes to the observed persistent binding. In these experiments the supernatant from the last wash failed to affect [³H]NMS binding, indicating that xanomeline wash-resistant binding is not due to insufficient removal of free xanomeline. Furthermore, wash-resistant binding of xanomeline to membranes was not affected by the addition of the stable GTP analog GppNHp, indicating that it is not due to a locking of xanomeline by receptors in a high-affinity agonist binding state. However, we cannot at present exclude the possibility that xanomeline partitioned in the cell membrane contributes to its observed binding profile.

The rate of formation of the wash-resistant xanomeline binding to M₁ muscarinic receptors was quite fast, reaching a steady state in 10 min. The half-life of this type of xanomeline binding exceeded 24 h. Although persistent binding of xanomeline has features of near-irreversible binding, data in Fig. 1A indicate that it is not truly irreversible because it reaches a concentration-dependent plateau instead of progressing indefinitely with time. Additional insight is needed but is difficult to obtain in the absence of commercially available radiolabeled xanomeline.

Several pieces of evidence support our previously proposed hypothetical interpretation whereby xanomeline-persistent binding involves its interaction with a domain(s) different from the classical receptor binding site. This is clearly revealed by the fact that xanomeline was capable of exerting wash-resistant binding even when the classical binding site was completely obliterated by atropine, albeit at a slightly slower rate and lower potency. Interestingly, both the rate and the magnitude of xanomeline-persistent binding to the M₁ receptor remained constant when the concentration of atropine was increased from 1 µM (receptor saturating concentration) to 100 µM (suprasaturating concentration).

The observation that the wash-resistant binding of xanomeline was minimally affected by a complete blockade of the classical binding site demonstrates that an initial interaction of xanomeline with the classical binding site is not a prerequisite for its persistent binding somewhere else. It should be noted, however, that it was previously shown that atropine blocks activation of the receptor induced by the wash-resistant prebound xanomeline (Christopoulos, 2001). Two explanations seem plausible to reconcile these observations. First, persistently bound xanomeline might activate the receptor by an allosteric action (Jakubík et al., 1996). Atropine, being an inverse agonist (negative antagonist) stabilizes the receptor in an inactive conformation to suppress its activation by xanomeline. Second, as previously proposed, persistently bound xanomeline might also interact with the receptor classical site in a direct but not as-yet understood way (Christopoulos et al., 1999), which would make receptor activation sensitive to atropine.

There is previous evidence of avid interaction of certain 2-adrenoreceptor agonists with an exosite on the receptor (Clark et al., 1996; Coleman et al., 1996). Although our present findings strongly suggest that persistent binding of xanomeline to the M₁ muscarinic receptor does not take place at the conventional primary binding site on the receptor, certain anomalies in the characteristics of xanomeline binding described herein are at odds with the involvement of a saturable secondary binding site that exhibits simple bimolecular mass-action kinetics. Most importantly, there is marked discrepancy between the apparent affinity of xanomeline wash-resistant binding derived from kinetic (Figs. 1 and 2) and equilibrium (Figs. 3 and 4) experiments. The fast association and slow dissociation in Figs. 1 and 2 implicate nanomolar affinity (ranging between 0.7 and 15 nM) for xanomeline wash-resistant binding. On the other hand, apparent affinity of xanomeline-persistent binding obtained in equilibrium experiments in Figs. 3 to 5 is in the micromolar range. It is difficult at the present time to offer a highly satisfactory explanation for such discrepancy, particularly because a spontaneous time-dependent conversion of xanomeline-persistent binding from an initial high-affinity to a late lower affinity state is not thermodynamically feasible. As a matter of speculation, however, one plausible interpretation would implicate negative allosteric modulation of xanomeline-persistent binding by xanomeline binding to the orthosteric binding site. Due to different kinetics of the two modes of xanomeline binding, one would anticipate that the magnitude of such allosteric modulation changes over the time. If xanomeline wash-resistant binding is faster than its reversible interaction at the orthosteric binding site then this mode of binding should exhibit initial high affinity in absence of the dampening effects of xanomeline bound at the primary site. Upon approaching equilibrium, however, a higher proportion of xanomeline binds to the orthosteric site and allosterically decreases affinity of wash-resistant binding of the drug. Another related anomalous observation is the nonlinear relationship between xanomeline concentration and the observed rate of its persistent binding (Fig. 1B). A simple bimolecular mass-action mechanism of binding to a defined saturable site should result in a linear relationship. Moreover, persistently bound xanomeline brings about a loss of the apparent number of NMS binding sites, instead of the anticipated progressive decrease in ligand affinity without influencing its maximal binding.

Taken together, these observations suggest quite complex mechanisms, and perhaps stoichiometry, of formation of xanomeline-persistent binding to the M₁ muscarinic receptor. Future experiments using radiolabeled xanomeline will be of paramount importance toward understanding these mechanisms.

Binding of lower concentrations of xanomeline is completely reversible in nature. In contrast to the wash-resistant component of xanomeline binding, several pieces of evidence suggest that its reversible binding is competitive with ligands that interact with the receptor classical binding site. One salient feature of our present data and the proposed interpretation is that reversible and persistent interactions of xanomeline (taking place at low and high concentrations, respectively) influence binding of competitive receptor ligands differently. It is to be noted, that high concentrations of xanomeline will exert both modes of binding to the M₁ muscarinic receptor, unless the reversible component of xanomeline binding is eliminated or blocked. In vivo, this could be attained by administering a high dose of xanomeline, followed by a waiting period that allows for metabolism of free xanomeline. Alternatively, xanomeline would bind primarily in a wash-resistant manner in presence of drugs that interact competitively with the classical binding site on the M₁ muscarinic receptor [e.g., anticholinergics, antidepressants, or antipsychotics (Richelson, 1999)].