Pronounced conformational changes following agonist activation of the M(3) muscarinic acetylcholine receptor.

The conformational changes that convert G protein-coupled receptors (GPCRs) activated by diffusible ligands from their resting into their active states are not well understood at present. To address this issue, we used the M(3) muscarinic acetylcholine receptor, a prototypical class A GPCR, as a model system, employing a recently developed disulfide cross-linking strategy that allows the formation of disulfide bonds using Cys-substituted mutant M(3) muscarinic receptors present in their native membrane environment. In the present study, we generated and analyzed 30 double Cys mutant M(3) receptors, all of which contained one Cys substitution within the C-terminal portion of transmembrane domain (TM) VII (Val-541 to Ser-546) and another one within the C-terminal segment of TM I (Val-88 to Phe-92). Following their transient expression in COS-7 cells, all mutant receptors were initially characterized in radioligand binding and second messenger assays (carbachol-induced stimulation of phosphatidylinositol hydrolysis). This analysis showed that all 30 double Cys mutant M(3) receptors were able to bind muscarinic ligands with high affinity and retained the ability to stimulate G proteins with high efficacy. In situ disulfide cross-linking experiments revealed that the muscarinic agonist, carbachol, promoted the formation of cross-links between specific Cys pairs. The observed pattern of disulfide cross-links, together with receptor modeling studies, strongly suggested that M(3) receptor activation induces a major rotational movement of the C-terminal portion of TM VII and increases the proximity of the cytoplasmic ends of TM I and VII. These findings should be of relevance for other family A GPCRs.

The superfamily of G protein-coupled receptors (GPCRs) 1 represents the largest group of cell surface receptors found in nature (1)(2)(3). A structural hallmark of all GPCRs is the presence of a bundle of seven transmembrane helices (TM I-VII) that are connected by alternating intracellular and extracellular loops (4 -6) (Fig. 1). The structural elements determining ligand binding and G protein recognition have been studied in considerable detail, at least for some members of the GPCR superfamily (5)(6)(7)(8). In contrast, the conformational changes that activating ligands induce in their target receptors are still not well understood at present. The currently available evidence suggests that GPCR activation opens a cleft on the intracellular side of the receptor that promotes the recognition and activation of specific G protein heterotrimers (4,6,9,10,34).
At present, bovine rhodopsin (in its inactive state) is the only GPCR for which high resolution structural information is currently available (11). Most GPCRs share a considerable degree of structural homology with bovine rhodopsin (12) and are therefore also referred to as rhodopsin-like or family A GPCRs. However, whereas the endogenous ligand of rhodopsin, 11-cisretinal, is covalently bound to the receptor protein, all other family A GPCRs known to date are activated by diffusible ligands. The possibility therefore exists that the precise structural mechanisms involved in receptor activation may not be identical between rhodopsin and other class A GPCRs.
During the past decade, considerable progress has been made in elucidating the light-induced conformational changes in bovine rhodopsin (9,10,13). Biophysical and biochemical studies suggest that rhodopsin activation triggers a reorientation of the cytoplasmic end of TM VI and changes in the relative disposition of TM VI and III, along with smaller movements involving several other TM helices (10, 14 -16). Considerable evidence indicates that a similar movement (reorientation of the cytoplasmic end of TM VI versus that of TM III) occurs in other GPCRs, including the ␤ 2 -adrenergic receptor (17)(18)(19)(20)(21). More specifically, site-directed spin labeling studies (15) suggested that rhodopsin activation involves a rigid body movement of the cytoplasmic end of TM VI (away from the C terminus of TM III) that is accompanied by a rotational movement of ϳ30°(clockwise as viewed from the cytoplasm).
The pioneering biophysical and biochemical studies carried out with bovine rhodopsin have led to important new insights into the structural mechanisms involved in rhodopsin activation. However, the vast majority of these studies were carried out with mutant versions of rhodopsin in the solution state (receptor proteins were solubilized in dodecyl maltoside micelles), and some data suggest that the structural and dynamic properties of rhodopsin present in solution may not be identical * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. with those found in native disk membranes (10). Thus, the development of techniques that would allow the monitoring of agonist-induced conformational changes in GPCRs present in their native membrane environment would be highly desirable. To address this issue, we recently described a novel in situ disulfide cross-linking strategy that allows the formation of disulfide bonds using Cys-substituted mutant M 3 muscarinic acetylcholine receptors present in their native membrane environment (22,23). The M 3 muscarinic receptor is a prototypical class A GPCR that preferentially interacts with G proteins of the G q family (24).
Agonist binding to the M 3 muscarinic receptor and most other class A GPCRs involves, among other sites of contact, several key residues present within the exofacial portion of TM VII (24,25). Moreover, the endofacial segment of TM VII contains the highly conserved NPXXY motif (corresponding to Asn-539 to Tyr-543 in the rat M 3 receptor sequence) (Fig. 1), which may provide a point of flexibility for agonist-induced structural changes. We therefore tested the hypothesis that diffusible ligands may induce conformational changes within the cytoplasmic segment of TM VII (the region located Cterminal of the NPXXY motif), using our previously developed in situ disulfide cross-linking strategy (22,23).
Since the C-terminal segment of TM VII is predicted to be located in the vicinity of the C-terminal portion of TM I (11), we generated 30 double Cys mutant M 3 muscarinic receptors, all of which contained one Cys substitution within the C-terminal portion of TM VII (Val-541 7.51 -Ser-546 7.56 ) and another Cys substitution within the C-terminal segment of TM I (Val-88 1.53 -Phe-92 1.57 ) (the superscripts indicate amino acid positions according to the nomenclature proposed by Ballesteros and Weinstein (36)). All Cys mutations were introduced into a modified version of the rat M 3 muscarinic receptor (M3Ј(3C)-Xa) (22,26) that lacked most native Cys residues and contained a factor Xa cleavage site within the third intracellular loop (i3 loop) (Fig. 1).
Disulfide cross-linking experiments, carried out in the absence or the presence of a muscarinic agonist (carbachol), led to the identification of three double Cys mutant M 3 muscarinic receptors (V88C 1.53 /Y543C 7.53 , A91C 1.56 /L545C 7.55 , and A91C 1.56 /S546C 7.56 ) that showed agonist-promoted disulfide bond formation. The observed cross-linking pattern, in combination with a newly established three-dimensional model of the rat M 3 muscarinic receptor, strongly suggested that receptor activation leads to a major rotational movement of the Cterminal portion of TM VII and increases the proximity of the C-terminal ends of TM I and VII. Given the high degree of structural homology found among all class A GPCRs, our findings should be of broad general relevance. Generation of Cys-substituted Mutant M 3 Muscarinic Receptor Constructs-All Cys substitutions were introduced into a pCD-based expression plasmid coding for a modified version of the rat M 3 muscarinic receptor, previously referred to as M3Ј(3C)-Xa receptor (Fig. 1). The generation of the M3Ј(3C)-Xa expression plasmid has been described previously (26). The M3Ј(3C)-Xa receptor contains an N-terminal he-magglutinin epitope tag and lacks all five potential N-terminal Nglycosylation sites and most endogenous Cys residues, except for Cys 140 , Cys 220 , and Cys 532 . Importantly, the central portion of the i3 loop (Ala 274 -Lys 469 ) was replaced by two factor Xa cleavage sites. Cys residues were reintroduced into the M3Ј(3C)-Xa construct at positions Val-88 1.53 -Phe-92 1.57 and Val-541 7.51 -Ser-546 7.56 , by using the QuikChange TM site-directed mutagenesis kit (Stratagene), according to the manufacturer's instructions. Double Cys mutant receptors were obtained by subcloning a 1.7-kb BglII-NdeI fragment derived from the mutant M3Ј(3C)-Xa constructs containing single Cys substitutions at positions Val-541 to Val-546 into the M3Ј(3C)-Xa constructs containing single Cys substitutions at positions Val-88 to Phe-92. The identity of all mutant constructs was verified by DNA sequencing.

Materials-Copper
Expression of Receptor Constructs in Mammalian Cells-All mutant M 3 muscarinic receptors were transiently expressed in COS-7 cells. Cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 2 mM L-glutamine, 100 units/ml penicillin, and 100 g/ml streptomycin at 37°C in a humidified 5% CO 2 incubator. Approximately 24 h prior to transfections, 1 ϫ 10 6 cells were seeded into 100-mm dishes. Cells were transfected with 4 g of receptor plasmid DNA/dish using the Lipofectamine TM Plus kit (Invitrogen), according to the manufacturer's recommendations. In order to increase muscarinic receptor expression levels, 1 M atropine was routinely added to the incubation medium for the last 24 h of culture.
Preparations of Membranes from Transfected COS-7 Cells-Transfected cells were harvested ϳ48 h after transfections. To ensure complete removal of atropine that was present in the incubation medium during the last 24 h of culture, cells were washed twice (10 min each wash) with 10 ml of ice-cold phosphate-buffered saline (pH 7.4). Subsequently, 2 ml of ice-cold buffer A (25 mM sodium phosphate and 5 mM MgCl 2 , pH 7.4) was added to each 100-mm dish, followed by a 15-min incubation at 4°C. Cells were then scraped off of the plates and homogenized using a Polytron tissue homogenizer (setting 5; 20 s), followed by a 15-min centrifugation at 20,000 ϫ g at 4°C. The membrane pellets were then resuspended in buffer A (1 ml/100-mm dish), rehomogenized, frozen on dry ice, and stored at Ϫ70°C until needed. Protein concentrations were measured using the Micro BCA protein assay reagent kit with bovine serum albumin as a standard.
Radioligand Binding Studies-Radioligand binding assays were carried out using membrane homogenates prepared from transfected COS-7 cells essentially as described previously (22). In brief, all incubations were carried out in 1 ml of buffer A (ϳ10 -20 g of membrane protein/tube) for 2 h at room temperature (22°C). In saturation binding assays, six different concentrations (ranging from 20 to 3,000 pM) of [ 3 H]NMS were used. In competition binding assays, a fixed concentration of [ 3 H]NMS (500 pM) was employed in the presence of 10 different concentrations (13 pM to 10 M) of the cold competitor, carbachol, a muscarinic agonist. Reactions were terminated by rapid filtration over GF/C Brandel filters followed by three washes (ϳ4 ml each) with icecold distilled water. In all assays, nonspecific binding was defined as the binding remaining in the presence of 1 M atropine. The amount of bound radioactivity was determined by liquid scintillation spectrometry. Binding data were analyzed using the nonlinear curve-fitting program Prism 3.0 (GraphPad).
Agonist-induced Stimulation of Phosphatidylinositol Hydrolysis-The ability of the muscarinic agonist, carbachol, to stimulate increases in intracellular inositol monophosphate (IP 1 ) levels was determined using transfected COS-7 cells grown in 6-well plates, as described previously (26). After labeling of cells for 20 -24 h with myo-[ 3 H]inositol (3 Ci/ml), cells were incubated in the presence of 10 mM LiCl for 1 h at 37°C with increasing concentrations of carbachol. The IP 1 fraction was isolated and quantitated as described (26). Carbachol concentrationresponse curves were analyzed using the nonlinear curve-fitting program Prism 3.0 (GraphPad).
Oxidation, Solubilization, and Factor Xa Treatment of Mutant M 3 Muscarinic Receptors-Membrane preparations obtained from transfected COS-7 cells were thawed at room temperature and rehomogenized as described under "Preparations of Membranes from Transfected COS-7 Cells." Membranes from one 100-mm dish (ϳ1 mg of protein present in a 1-ml volume) were incubated in microcentrifuge tubes with end-over-end rotation (30 rpm; 10 min at room temperature) with the oxidizing agent, Cu(II)-phenanthroline (2.5 M), either in the absence or the presence of different concentrations of the muscarinic agonist, carbachol, or the antagonist, atropine. Reactions were terminated by the addition of EDTA and N-ethylmaleimide (10 mM each), followed by a 10-min incubation on ice.
To obtain membrane lysates (22), samples were then centrifuged at 8,000 ϫ g for 10 min at 4°C. The resulting membrane pellets were incubated with 250 l of 0.2% digitonin in phosphate-buffered saline (pH 7.4) for 20 min on ice (to remove peripheral membrane proteins). Following another centrifugation step (8,000 ϫ g for 10 min at 4°C), membrane pellets were incubated with 1.2% digitonin in buffer B (50 mM Tris-HCl, pH 8, 100 mM NaCl, and 1 mM CaCl 2 ) for 90 -120 min at 4°C with end-over-end rotation (30 rpm). After another centrifugation step (same conditions as above), the supernatants (membrane lysates containing solubilized mutant M 3 muscarinic receptors) were transferred to fresh microcentrifuge tubes. Membrane lysates (ϳ15 g of protein) were then incubated with factor Xa protease (final concentration, 0.1 g/l) at room temperature for 16 -20 h (final volume, 50 l). The reactions were then terminated by incubation for 30 min at room temperature with a mammalian protease inhibitor mixture (1:25 dilution; Sigma). Samples were then used directly for SDS-PAGE or stored at Ϫ70°C until use.
Western Blot Analysis-SDS-PAGE was performed essentially as described (22). Samples were incubated for 30 min at 37°C with Laemmli loading buffer (nonreducing conditions) and then loaded onto 10 -20% Tris-glycine polyacrylamide gels, which were run at 125 V in the presence of 0.1% SDS. Western blotting studies were carried out essentially as described by Ward et al. (22), using the anti-M 3 antibody directed against the C-terminal 18 amino acids of the M 3 receptor protein (27). Receptor proteins were visualized by using ECL detection reagents and autoradiography. The intensities of immunoreactive bands were quantitated by scanning densitometry using the program ImageQuant TL (Amersham Biosciences).
A Three-dimensional Model of the Rat M 3 Muscarinic Receptor-A three-dimensional model of the TM core of the rat M 3 muscarinic receptor (including the various loop regions and helix 8) was built using homology modeling based on the high resolution (2.8 Å) x-ray structure of the inactive state of bovine rhodopsin (11). All calculations were performed on a Silicon Graphics Octane work station (300-MHz MIPS R12000 (IP30) processor) using the SYBYL 6.9 program (Sybyl Molecular Modeling System, version 6.9; Tripos Inc., St. Louis, MO). For the conformational refinement of the initial M 3 muscarinic receptor model, the optimized structures were used as the starting point for subsequent molecular dynamics studies. Overall, the M 3 receptor model showed high structural similarity with that of the rhodopsin template, especially within the regions endowed with secondary structure (for details, see Supplemental Data).

Generation of 30 Double Cys Mutant M 3 Muscarinic
Receptors-This study was designed to monitor agonist-induced conformational changes in the M 3 muscarinic receptor with the receptor being present in its native membrane environment. Our major goal was to detect potential activity-dependent structural changes occurring at the cytoplasmic end of TM VII. Toward this aim, we used an in situ disulfide cross-linking strategy to monitor the positions of six consecutive amino acids located at the C terminus of TM VII, relative to a string of residues located at the C terminus of TM I. Altogether, we generated 30 double Cys mutant receptors, all of which contained one Cys substitution within the C-terminal segment of TM I (Val-88 1.53 -Phe-92 1.57 ) and another Cys substitution within the C-terminal portion of TM VII (Val-541 7.51 -Ser-546 7.56 ) (Fig. 1). All Cys mutations were introduced into a modified version of the rat M 3 muscarinic receptor (M3Ј(3C)-Xa) (26) that lacked most native Cys residues and contained a factor Xa cleavage site within the i3 loop ( Fig. 1).

Transient Expression of Mutant M 3 Muscarinic Receptors and Radioligand
Binding Studies-All 30 double Cys mutant M 3 receptor constructs, along with the M3Ј(3C)-Xa "background" receptor, were transiently expressed in COS-7 cells and initially examined for their ability to bind the muscarinic radioligand, [ 3 H]NMS. To increase receptor expression levels, transfected cells were incubated with atropine (1 M) for the last 24 h of culture. We previously demonstrated that this strategy leads to a pronounced increase in the density of the M3Ј(3C)-Xa receptor and all Cys-substituted mutant receptors derived from this construct (22,23).
Saturation binding studies with the muscarinic antagonist, tors differed from this value by less than 2-fold (Table I). Interestingly, several mutant receptors yielded B max values that were significantly higher than that observed with the M3Ј(3C)-Xa construct ( Table I) Competition binding studies with the muscarinic agonist, carbachol, showed that most of the 30 double Cys mutant receptors displayed carbachol binding affinities that were similar to or even higher than that found for the M3Ј(3C)-Xa construct (K i ϭ 15.2 Ϯ 1.7 M; Table I). Interestingly, most mutant receptors containing the F92C point mutation exhibited carbachol affinities that were increased by ϳ3-10-fold, as compared with the M3Ј(3C)-Xa "base mutant" (Table I). In contrast, only two of the analyzed double Cys mutant receptors, I89C 1.54 /S546C 7.56 and V90C 1.55 /A544C 7.54 , showed a clear reduction (by ϳ3-10-fold) in carbachol binding affinities (Table I).
G Protein-coupling Properties of Double Cys Mutant M 3 Muscarinic Receptors-To examine whether the 30 double Cys mutant receptors were still able to couple to G proteins, we next studied their ability to mediate carbachol-induced increases in inositol monophosphate (IP 1 ) production (phosphatidylinositol hydrolysis). As shown in Table II, all analyzed double Cys mutant receptors retained the ability to stimulate phosphatidylinositol hydrolysis with high efficacy (E max , expressed as -fold increase in IP 1 production above basal levels). In general, the apparent increase in E max values displayed by several of the double Cys mutant receptors, as compared with the M3Ј(3C)-Xa construct, correlated well with reduced basal IP 1 levels displayed by these mutant receptors (Table II). This observation suggests that the observed increases in E max values observed for several of the double Cys mutant receptors probably do not reflect a "true" gain in G protein coupling efficacy.
Most of the analyzed double Cys mutant receptors showed reduced carbachol potencies (increased EC 50 values), as compared with the M3Ј(3C)-Xa construct (EC 50 ϭ 0.025 Ϯ 0.004 M; Table II), indicative of impaired receptor/G protein coupling efficiency. This observation is consistent with the concept that residues located within the C-terminal segment of TM VII play an important role in receptor/G protein coupling (28,29).
Taken together, the ligand binding studies indicated that all 30 double Cys mutant receptors were able to bind muscarinic ligands with high affinity, indicating that all mutant receptors were properly folded. Moreover, all double Cys mutant receptors retained the ability to activate G proteins, although the efficiency of receptor/G protein coupling was reduced in most cases.
Disulfide Cross-linking Studies-The results of the [ 3 H]NMS binding studies indicated that all 30 double Cys mutant M 3 receptors, similar to the M3Ј(3C)-Xa construct, were properly expressed, at moderate to high levels, in transiently transfected COS-7 cells. Consistent with these results, Western blotting studies using a polyclonal antibody directed against the C-terminal portion of the rat M 3 muscarinic receptor (referred to as "anti-M 3 antibody") showed that all double Cys mutant

muscarinic receptors
The indicated mutant M 3 muscarinic receptors were transiently expressed in COS-7 cells. All double Cys mutant receptors were derived from the M3Ј(3C)-Xa construct. B max and K D values for ͓ 3 H͔NMS were determined from saturation binding experiments using membrane homogenates prepared from transfected COS-7 cells. Carbachol binding affinities (K i ) were determined in ͓ 3 H͔NMS competition binding assays (n H ϭ Hill coefficient). Carbachol binding data were corrected for the Cheng-Prusoff shift. Binding data were analyzed using the nonlinear curve-fitting program Prism 3.0 (GraphPad). Data are given as means Ϯ S.E. from 2-5 independent experiments, each performed in duplicate. receptors could be easily detected in membrane lysates prepared from transfected COS-7 cells (data not shown) in a fashion similar to the M3Ј(3C)-Xa receptor (22,23).
To probe the potential proximity of the Cys pairs present in the 30 double Cys mutant receptors, we examined the ability of the 30 Cys pairs to form intramolecular disulfide bonds. Disulfide cross-linking studies were performed with mutant receptors being present in their natural membrane environment (in situ), using membrane preparations obtained from transfected COS-7 cells (22,23). To facilitate the formation of disulfide bonds, all reactions were carried out in the presence of a low concentration (2.5 M; 10-min incubation at room temperature) of the mild oxidizing agent, Cu(II)-phenanthroline. Subsequently, receptors were solubilized, digested to completion with factor Xa, and subjected to SDS-PAGE and Western blotting (reducing and nonreducing conditions). In this system, the appearance of a ϳ38-kDa receptor band under nonreducing conditions that is absent under reducing conditions is indicative of successful disulfide cross-linking. We previously demonstrated, using surface biotinylation (22) and other biochemical techniques (23), that this ϳ38-kDa immunoreactive species corresponds to properly folded cell surface receptors. To study activity-dependent changes in disulfide cross-linking patterns, cross-linking studies were carried out in the absence and in the presence of the muscarinic agonist, carbachol.
The results of the cross-linking studies obtained with the 30 double Cys mutant M 3 receptors are summarized in Fig. 2. Under basal conditions (no carbachol added to the incubation medium), only the V88C 1.53 /Y543C 7.53 receptor yielded a robust 38-kDa signal (under nonreducing conditions) (Fig. 2), indicat-ing that Val-88 (located on TM I) and Tyr-543 (located on TM VII) lie adjacent to each other in the three-dimensional structure of the M 3 muscarinic receptor. The intensity of the 38-kDa band further increased when cross-linking reactions were car-  Fig. 1 for the structure of the mutant receptors) were incubated with the oxidizing agent, Cu(II)-Phen (2.5 M), for 10 min at room temperature, either in the absence or the presence or the muscarinic agonist, carbachol (1 mM). Receptors were then solubilized and digested to completion with factor Xa, as indicated under "Experimental Procedures." Samples containing equal amounts of protein (ϳ5 g) were then run, under nonreducing conditions, on 10 -20% Tris-glycine polyacrylamide gels, followed by Western blotting analysis using the anti-M3 antibody. Note that three of the 30 investigated double Cys mutant receptors (V88C/Y543C, A91C/L545C, and A91C/S546C) displayed agonist-dependent crosslinking, as indicated by the appearance of a clearly visible ϳ38-kDa full-length receptor band. Data shown here are representative of three independent experiments. ried out in the presence of the muscarinic agonist, carbachol (1 mM). Strikingly, in the presence of carbachol, two additional double Cys mutant receptors, A91C 1.56 /L545C 7.55 and A91C 1.56 /S546C 7.56 , also showed pronounced disulfide crosslinking, as indicated by the appearance of a prominent 38-kDa receptor species (Fig. 2). These 38-kDa bands were not observed when Western blotting studies were carried out under reducing conditions (data not shown), indicating that they were not caused by incomplete digestion by factor Xa.

FIG. 2. Agonist-induced disulfide cross-linking in mutant M 3 muscarinic receptors studied by Western blot analysis. Thirty M3Ј(3C)-Xa-based double Cys mutant receptors (see
Agonist Dependence of Disulfide Cross-linking Signals-To examine the agonist dependence of the disulfide cross-linking signals in more detail, we carried out additional cross-linking experiments in which the V88C 1.53 /Y543C 7.53 , A91C 1.56 / L545C 7.55 , and A91C 1.56 /S546C 7.56 receptors were oxidized (in situ) in the presence of different carbachol concentrations (0.01-10 mM). In all three mutant receptors, carbachol induced concentration-dependent increases in the intensity of the 38-kDa cross-linking signal (nonreducing conditions) (Fig. 3A). As expected, no cross-linking was observed with the M3Ј(3C)-Xa "background" receptor, which served as a negative control.
Under the same experimental conditions, the muscarinic antagonist, atropine (0.1 and 1 M), had no significant effect on the weak cross-linking signals observed in the absence of carbachol (Fig. 4).
The results of three independent carbachol cross-linking experiments carried out with the V88C/Y543C, A91C/L545C, and A91C/S546C mutant receptors are summarized in Fig. 3B, based on the quantification of the intensity of the 38-kDa receptor species. To obtain a quantitative measure of agonistinduced disulfide cross-linking, ratios were formed between band intensities determined in the presence versus the absence of carbachol. As shown in Fig. 3B, the relative increases in agonist-induced disulfide cross-linking were most pronounced for the A91C/L545C and A91C/S546C receptors.
Disulfide Bonds Form Intramolecularly Rather than Intermolecularly-Like many other GPCRs (30,31), M 3 muscarinic receptors have been shown to form dimers or oligomers (32). To exclude the possibility that the agonist-dependent formation of disulfide bonds observed with the V88C/Y543C, A91C/L545C, and A91C/S546C constructs was due to the formation of intermolecular cross-links (involving Cys residues located on different receptor molecules), we carried out an additional set of experiments. We first generated five M3Ј(3C)-Xa-based mutant receptors containing the V88C, A91C, Y543C, L545C, and S546C single Cys substitutions. Radioligand binding studies showed that these mutant receptors were properly expressed in transiently transfected COS-7 cells (data not shown). We then cotransfected COS-7 cells with the following pairs of single Cys mutant receptor constructs: V88C ϩ Y543C, A91C ϩ L545C, and A91C ϩ S546C. In the absence of Cu(II)-phenanthroline and factor Xa treatment, Western blotting studies using membrane lysates prepared from these cotransfected COS-7 cells led to the appearance of prominent 38-kDa immunoreactive bands in all three cases (Fig. 5), indicative of proper protein expression. However, despite the presence of carbachol (1 mM), these bands were no longer observed following oxidation of membrane homogenates with Cu(II)-phenanthroline (2.5 M) and factor Xa treatment (Fig. 5). On the other hand, under the same experimental conditions, the V88C/Y543C, A91C/L545C, and A91C/S546C double Cys constructs (which served as positive controls) gave pronounced cross-linking signals (Fig. 5), as expected (Figs. 2 and  3). Taken together, these data strongly support the concept that carbachol promotes the formation of intramolecular rather than intermolecular disulfide bonds in the V88C/Y543C, A91C/L545C, and A91C/S546C mutant receptors.
A Three-dimensional Model of the Rat M 3 Muscarinic Receptor-To facilitate the interpretation of the disulfide cross-linking data, we built a three-dimensional model of the TM core of the rat M 3 muscarinic receptor (including the various loop regions and helix 8) using homology modeling based on the high resolution x-ray structure of the inactive state of bovine rhodopsin (11,33). Overall, the calculated M 3 receptor model (resting state of the receptor) showed high structural similarity with the rhodopsin template, especially within the TM helical bundle (for details, see Supplemental Data). In bovine rhodopsin, hydrogen bonding networks and Van der Waals contacts link the different TM helices, stabilizing the ground state structure of the receptor (11,33). Analogously, the TM bundle of the inactive state of the rat M 3 receptor is also predicted to be stabilized by many interhelical hydrogen bonds and hydrophobic interactions that involve residues that are highly conserved in family A GPCRs (including bovine rhodopsin) as well as residues that are highly conserved only within the muscarinic receptor family (M 1 -M 5 ; for more detailed information, see Supplemental Data).
The predicted locations within the three-dimensional model of the rat M 3 muscarinic receptor of the residues involved in agonist-dependent disulfide cross-linking are indicated in the legend to Fig. 6. As described in detail above, carbachol promoted the formation of disulfide cross-links between Cys residues present at positions 88 1.53 /543 7.53 , 91 1.56 /545 7.55 , and 91 1.56 /546 7.56 . A side view of the TM receptor core shows that positions 88 and 543 are located at about the same level within TM I and TM VII, respectively (Fig. 6A). The same is true for positions 91 (TM I) and 545/546 (TM VII). The estimated distances between the C␣ atoms in the three residue pairs are (in Å): 9.61 (Val-88/Tyr-543), 15.08 (Ala-91/Leu-545), and 15.52 (Ala-91/Cys-546), respectively. Fig. 6B, which shows a cytoplasmic view of the TM receptor core of the rat M 3 receptor (with the primary focus on TM I and VII), indicates that residues 88 and 543 directly face each other at the TM I/TM VII interface. In striking contrast, residues 545 and 546 face away from TM I, including residue 91. This observation, together with the relatively large distances between the C␣ atoms in the 91/545, and 91/546 residue pairs, strongly suggests that the C-terminal portion of TM VII undergoes a major conformational change following M 3 receptor activation. The potential structural nature of these changes are discussed in detail below.

DISCUSSION
Many biochemical and biophysical studies of bovine rhodopsin have revealed the structural changes involved in lightinduced activation of rhodopsin in considerable detail (10,13). In contrast, little is known about the agonist-induced conformational changes that occur in GPCRs activated by diffusible ligands (6,34). In addition, almost all rhodopsin studies were carried out with mutant versions of rhodopsin in the solution state (10). Thus, studies that allow the monitoring of agonistinduced conformational changes in GPCRs present in their native membrane environment should be highly informative.
To address these issues, we recently developed a novel disulfide cross-linking strategy that allows the formation of disul- fide bonds using Cys-substituted mutant M 3 muscarinic receptors present in their native membrane environment (in situ) (22,23). Using this approach, we recently demonstrated that M 3 receptor activation leads to a conformational change that moves the cytoplasmic end of TM VI closer to that of TM V, consistent with the results of fluorescence spectroscopic studies of the ␤ 2 -adrenergic receptor (35).
Biochemical and biophysical analysis of bovine rhodopsin suggests that the C-terminal segment of TM VII undergoes significant light-induced conformational changes. Site-directed spin labeling studies showed that Tyr-306 7.53 underwent a light-induced mobility change, which was interpreted as an increase in the polarity of the environment of Tyr 306 (37). Abdulaev and Ridge (38) demonstrated that a monoclonal antibody direct against amino acids 304 7.51 -311 7.58 of bovine rhodopsin did not recognize the inactive state of rhodopsin but efficiently bound to light-activated rhodopsin, suggesting that rhodopsin activation is associated with pronounced conformational changes at the cytoplasmic end of TM VII. However, the precise molecular nature of these activity-dependent structural changes remains unknown.
Thus, one major goal of the present study was to investigate whether activation of GPCRs by diffusible ligands also leads to conformational changes involving the C-terminal portion of TM VII. To address this question, we employed a previously developed in situ disulfide cross-linking strategy (22,23). Moreover, we speculated that this analysis would yield more specific information about the molecular nature of the agonist-induced conformational changes occurring in this region.
Specifically, we generated 30 double Cys mutant M 3 muscarinic receptors, all of which contained one Cys substitution within the C-terminal portion of TM VII (Val-541 7.51 -Ser-546 7.56 ) and another Cys substitution with the C-terminal segment of TM I (Val-88 1.53 -Phe-92 1.57 ). The x-ray structure of bovine rhodopsin and our newly generated M 3 muscarinic receptor model suggest that these receptor segments lie adjacent to each other in the three-dimensional structure of class A GPCRs.
In the inactive state of the M 3 receptor (in the absence of carbachol), only the V88C/Y543C double Cys mutant receptor showed a robust disulfide cross-linking signal (Fig. 2). This observation suggests that Val-88 1.53 and Tyr-543 7.53 directly face each other at the TM I/TM VII interface, consistent with the newly developed three-dimensional model of the M 3 muscarinic receptor (Fig. 6B). This model suggests that the distance between the C␣ atoms of Val-88 and Tyr-543 is about 9.6 Å. Although the distance between the C␣ carbons of two Cys residues engaged in a disulfide bond usually ranges from about 3.8 to 6.8 Å (9, 39), disulfide cross-linking studies carried out with Cys-substituted versions of a bacterial chemoreceptor of known structure indicate that disulfide bonds can readily form when the two C␣ carbons are less than ϳ12 Å apart (40).
Interestingly, carbachol activation of the V88C/Y543C double Cys receptor led to a concentration-dependent increase in the intensity of the disulfide cross-linking signal (increase in intensity of the 38-kDa band following digestion with factor Xa; Fig. 3), suggesting that Val-88 1.53 and Tyr-543 7.53 move closer to each other following M 3 receptor activation. Mutational analysis of many different class A GPCRs suggests that the conserved Tyr 7.53 residue plays a key role in receptor activation (28,29,(41)(42)(43). In the inactive state of the receptor, Tyr 7.53 is predicted to interact with a highly conserved Phe residue (position 7.60) located in helix 8 (11,41,42). A recent study provided evidence that this interaction is disrupted following light-induced activation of rhodopsin (42). Our disulfide crosslinking data therefore suggest that the loss of this interaction may be due to the activity-dependent change in the position of Tyr 7.53 (Tyr-543 in the M 3 muscarinic receptor).
Strikingly, following carbachol treatment, two additional double Cys mutant M 3 receptors, A91C/L545C and A91C/ S546C, also showed pronounced disulfide cross-linking (Fig. 2). In both cases, as seen with the V88C/Y543C receptor, the intensity of the cross-linking signals increased with increasing agonist concentrations (Fig. 3), indicating that the number of agonist-occupied receptors correlates well with the extent of disulfide cross-linking.
In contrast to carbachol, the muscarinic antagonist, atropine, did not stimulate disulfide cross-link formation in the case of the V88C/Y543C, A91C/L545C, and A91C/S546C mutant receptors (Fig. 4), consistent with the concept that the observed conformational changes are agonist-specific. Atropine, like all other classical muscarinic antagonists that have been tested so far, is able to act as an inverse agonist, being able to suppress constitutive signaling by wild-type or mutant muscarinic receptors (43)(44)(45). At present, it remains unclear why atropine failed to reduce the weak cross-linking signals observed with the V88C/Y543C, A91C/L545C, and A91C/ S546C mutant receptors in the absence of carbachol.
In the inactive state of the M 3 muscarinic receptor, the C␣ atoms of Ala-91 1.56 and Leu-545 7.55 are predicted to be about 15.1 Å apart, as suggested by the M 3 muscarinic receptor model. A similar distance (15.5 Å) lies between the C␣ atoms of Ala-91 1.56 and Ser-546 7.56 . Fig. 6B also shows that Leu-545 and Ser-546, located at the C terminus of TM VII, face the lipid bilayer and TM VI, respectively, pointing away from Ala-91 (TM I). The ability of the muscarinic agonist, carbachol, to promote the formation of disulfide cross-links in the A91C/ L545C and A91C/S546C mutant receptors suggests that activation of M 3 receptors in their native membrane environment triggers major conformational changes at the cytoplasmic end of TM VII. A side view of the M 3 muscarinic receptor (Fig. 6A) shows that Leu-545 and Ser-546 are located at approximately the same level within the TM receptor core as Ala-91. The most straightforward explanation for the disulfide cross-linking patterns observed with the A91C/L545C and A91C/S546C mutant receptors therefore is that carbachol activation leads to a rotational movement of the C-terminal end of TM VII (clockwise as viewed from the cytoplasm) that brings residues 545 and 546 within cross-linking distance of position 91. One might argue that such a rotational movement would move Tyr-543 further away from Val-88, thus reducing the likelihood of disulfide cross-link formation between positions V88C and Y543C. One possible explanation for the observation that carbachol promoted rather than reduced disulfide cross-linking between these two residues is that the proposed rotational movement of the cytoplasmic end of TM VII follows an initial lateral movement of TM VII that moves Y543C closer to V88C.
TM VII, in contrast to TM I, contains several amino acids (e.g. Tyr-529 7.39 and Tyr-533 7.43 in the M 3 muscarinic receptor) that are involved in the binding of muscarinic agonists (24,25). Moreover, the endofacial segment of TM VII contains the highly conserved NPXXY motif (corresponding to Asn-539 7.49 -Tyr-543 7.53 in the rat M 3 receptor) (Fig. 1) that is known to be critically involved in receptor activation. The rhodopsin x-ray structure indicates that the Pro residue contained with the conserved NPXXY motif induces a pronounced kink in TM VII (11). A similar kink can be observed in the three-dimensional model of the M 3 muscarinic receptor. Since proline kinks in TM helices are considered points of flexibility within otherwise rigid helical rods (12,46), it is possible that the agonist-induced structural changes observed within the C-terminal segment of TM VII depend on the presence of the conserved TM VII Pro 7.50 residue.
In contrast to the conclusions based on our in situ disulfide cross-linking experiments, a site-directed spin labeling study using double Cys rhodopsin mutants in solution suggested that light-induced activation of rhodopsin increases the distance between position 65 1.60 (located at the C terminus of TM I) and 306 7.53 (located at the C terminus of TM VII) or 310 7.57 (located between TM VII and helix 8) (47). Based on this observation, the authors proposed that rhodopsin activation results in a movement of the cytoplasmic portion of TM VII away from TM I (10,47). There are several possible explanations for the discrepant conclusions drawn by Altenbach et al. (47) and those drawn from the results of the present study, suggesting that the cytoplasmic ends of TM I and VII move closer to each other following M 3 receptor activation. For example, the possibility exists that the discrepant conclusions drawn from the two studies are caused by the different experimental procedures used. Whereas the spectroscopic techniques used for studying conformational changes in rhodopsin measure the average of multiple possible receptor conformations (10), the disulfide cross-linking approach used in the present study allows the detection of agonist-induced receptor conformations that may be relatively short lived by "trapping" these conformations via disulfide bond formation (40,48). Another possibility is that the structural mechanisms involved in receptor activation are not identical between rhodopsin and class A GPCRs activated by diffusible ligands. Finally, it is also possible that the differences observed between the two studies are caused by the fact that the disulfide cross-linking experiments were carried out with receptors being present in their native membrane environment, whereas the rhodopsin studies were performed with mutant rhodopsin molecules in the solution state. In the present study, the activity-dependent formation of disulfide crosslinks was restricted to a rather small number of amino acids (Cys pairs). It is therefore unlikely that the observed disulfide cross-linking pattern was caused by a generalized increase in the mobility of TM VII. In any case, our findings emphasize the need to explore activity-dependent dynamic changes in receptor conformation by different (complementary) experimental techniques, preferably with receptors being present in their native membrane environment.
The high resolution x-ray structure of bovine rhodopsin indicates that TM VII is followed by another helical segment, referred to as helix 8, that lies nearly perpendicular to TM VII (11). Several lines of evidence suggest that GPCR activation leads to a structural rearrangement of helix 8 (37, 41, 49 -54). Moreover, studies with different class A GPCRs including rhodopsin suggest that helix 8 contains amino acids that are important for G protein recognition and activation (8,(55)(56)(57). One possible scenario therefore is that the activity-dependent structural changes observed at the C terminus of TM VII are propagated to the adjacent helix 8 region, allowing helix 8 to contribute to productive receptor/G protein interactions. Fig. 7 shows an intracellular view of the cytoplasmic surface of the M 3 muscarinic receptor, highlighting the structural changes that accompany M 3 receptor activation, as predicted by the results of our in situ disulfide cross-linking studies. We previously presented data suggesting that agonist activation moves the cytoplasmic end of TM VI closer to that of TM V (22). In addition, a recent disulfide cross-linking study using double Cys mutant M 3 receptors containing Cys substitutions at the intracellular ends of TM III and TM VI suggested that the cytoplasmic end of TM VI undergoes an agonist-dependent clockwise rotational movement, 2 consistent with previous bio-physical and biochemical studies carried out with rhodopsin (15) and the ␤ 2 -adrenergic receptor (17,18,35). The results of the present study support the concept that the cytoplasmic end of TM VII also undergoes an activity-dependent rotational movement and, at the same time, moves closer to the cytoplasmic end of TM I. It remains unclear at present whether all of these conformational changes occur in a concerted fashion or whether a primary structural change (e.g. the reorientation of TM VI) triggers secondary changes in the structure and orientation of other TM helices. In any case, our in situ disulfide cross-linking data are consistent with the concept that agonist activation of the M 3 muscarinic receptor opens a cleft on the intracellular receptor surface that increases the accessibility of various residues located at the cytoplasmic ends of different TM helices including TM III, VI, and VII (Fig. 7). These activitydependent changes are predicted to promote receptor binding to heterotrimeric G proteins, ultimately triggering productive receptor/G protein coupling.
In sum, our findings provide novel insights into the activitydependent conformational changes occurring in a GPCR activated by a diffusible ligand. Our results indicate that the applied in situ disulfide cross-linking strategy, combined with molecular modeling studies, represents a powerful approach for studying activity-dependent changes in GPCRs present in their native membrane environment. Systematic application of this strategy involving the entire TM receptor core should eventually lead to a refined model of the dynamic structural changes that convert a nonrhodopsin GPCR from its resting into its activated state. Given the high structural homology found among class A GPCRs, our results should also be relevant for other members of this receptor superfamily. The results of the present study suggest that receptor activation involves a conformational change that moves the cytoplasmic end of TM VII moves closer to that of TM I. This movement is predicted to be accompanied by a clockwise rotational movement of the cytoplasmic end of TM VII. In a previous disulfide cross-linking study (22), we showed that agonist activation of the M 3 muscarinic receptor allows the cytoplasmic end of TM VI to move closer to that of TM V. This movement is predicted to be accompanied by a clockwise rotational movement of the cytoplasmic end of TM VI, 2 consistent with similar findings in bovine rhodopsin (15) and the ␤ 2 -adrenergic receptor (17,18,35). These agonist-induced structural changes are likely to expose previously inaccessible receptor residues or surfaces (e.g. on TM III, VI, and VII) to G protein heterotrimers, ultimately triggering productive receptor/G protein coupling.