Conformational changes that occur during M3 muscarinic acetylcholine receptor activation probed by the use of an in situ disulfide cross-linking strategy.

The structural changes involved in ligand-dependent activation of G protein-coupled receptors are not well understood at present. To address this issue, we developed an in situ disulfide cross-linking strategy using the rat M(3) muscarinic receptor, a prototypical G(q)-coupled receptor, as a model system. It is known that a tyrosine residue (Tyr(254)) located at the C terminus of transmembrane domain (TM) V and several primarily hydrophobic amino acids present within the cytoplasmic portion of TM VI play key roles in determining the G protein coupling selectivity of the M(3) receptor subtype. To examine whether M3 receptor activation involves changes in the relative orientations of these functionally critical residues, pairs of cysteine residues were substituted into a modified version of the M(3) receptor that contained a factor Xa cleavage site within the third intracellular loop and lacked most endogenous cysteine residues. All analyzed mutant receptors contained a Y254C point mutation and a second cysteine substitution within the segment Lys(484)-Ser(493) at the intracellular end of TM VI. Following their transient expression in COS-7 cells, mutant receptors present in their native membrane environment (in situ) were subjected to mild oxidizing conditions, either in the absence or in the presence of the muscarinic agonist, carbachol. The successful formation of disulfide cross-links was monitored by studying changes in the electrophoretic mobility of oxidized, factor Xa-treated receptors on SDS gels. The observed cross-linking patterns indicated that M(3) receptor activation leads to structural changes that allow the cytoplasmic ends of TM V and TM VI to move closer to each other and that also appear to involve a major change in secondary structure at the cytoplasmic end of TM VI. This is the first study employing an in situ disulfide cross-linking strategy to examine agonist-dependent dynamic structural changes in a G protein-coupled receptor.

G protein-coupled receptors (GPCRs) 1 constitute the largest class of signaling molecules in the mammalian genome (1-3).
All GPCRs are predicted to share a conserved molecular architecture consisting of a bundle of seven ␣-helically arranged transmembrane domains (TM I-VII) linked by alternating intracellular and extracellular loops (Fig. 1). Despite the remarkable structural diversity of the ligands that exert their physiological functions via interaction with specific classes of GPCRs, all GPCRs are thought to share a conserved mechanism of activation. Several lines of evidence indicate that the binding of ligands to the extracellular side of the receptor leads to changes in the arrangement of distinct TM helices, which are then propagated to the intracellular surface of the receptor, thus enabling the receptor to recognize and activate specific classes of heterotrimeric G proteins (4 -7).
Accumulating evidence suggests that GPCR activation may involve a change in the relative disposition of TM III and VI (4 -11). Elegant site-directed spin labeling studies (9) carried out with the photoreceptor, rhodopsin, indicated that rhodopsin activation may involve a rigid body movement of the cytoplasmic end of TM VI, away from the C terminus of TM III. In agreement with this concept, cross-linking of the cytoplasmic ends of TM III and TM VI, either via disulfide bonds in rhodopsin (9) or via metal ion bridges in rhodopsin and other GPCRs (10,11), prevented receptor activation. However, despite these recent findings, our knowledge of the molecular nature of the structural changes involved in GPCR activation is still very incomplete.
During the past decade, we have used the M 3 muscarinic acetylcholine receptor, a prototypical member of the superfamily of class I GPCRs (rhodopsin family), as a model system to delineate the structural elements involved in determining the selectivity of receptor/G protein coupling selectivity (reviewed in Ref. 12). Like the M 1 and M 5 muscarinic receptors, the M 3 receptor subtype is preferentially coupled to G proteins of the G q family, activation of which leads to the breakdown of phosphatidylinositol (PI) lipids (12). Mutational analysis of the rat M 3 muscarinic receptor indicated that a limited number of amino acids located within the second (i2) and third intracellular (i3) loops and the adjacent cytoplasmic ends of TM III, V, and VI largely determine the G q coupling preference of this receptor subtype (13)(14)(15)(16). These findings are in agreement with a large number of studies using other classes of GPCRs (17,18). Both loss-and gain-of-function mutagenesis studies revealed that a series of primarily hydrophobic amino acids located at the TM V/i3 loop and TM VI/i3 loop junctions are of particular importance for proper recognition of G q proteins by the M 3 muscarinic receptor. These residues include an aromatic residue located at the C terminus of TM V (Tyr 254 ) and several mostly hydrophobic residues present at the cytoplasmic end of TM VI (Ala 488 , Ala 489 , Leu 492 , and Ser 493 ) ( Fig. 1; Refs. [13][14][15][16]. Mutational analysis of the M 5 muscarinic receptor has yielded similar results (19,20).
To gain insight into the structural mechanisms involved in M 3 receptor activation, we wanted to investigate whether these functionally important residues change their relative positions on the cytoplasmic surface of the receptor protein following agonist-dependent receptor activation. To address this issue, we decided to employ Cys substitution mutagenesis followed by disulfide cross-linking of Cys residues that are adjacent to each other in the three-dimensional structure of the receptor. This approach has been used successfully in the past to monitor dynamic changes in a number of different membrane proteins including various bacterial chemoreceptors (21)(22)(23)(24) and, more recently, bovine rhodopsin (9,(25)(26)(27)(28)(29).
To examine the potential proximity of the cytoplasmic ends of TM III and TM VI, we recently described a cross-linking protocol involving the use of a series of Cys-substituted mutant M 3 muscarinic receptors (30). The observed disulfide crosslinking patterns suggested that the cytoplasmic surface of the M 3 receptor protein is highly dynamic. In this previous study (30), mutant M 3 receptors were oxidized following solubilization and factor Xa digestion (a factor Xa site was present in the central portion of the i3 loop of all analyzed mutant receptors). Thus, the possibility cannot be completely excluded that the observed disulfide cross-linking "promiscuity" may have been caused, at least partially, by an increase in conformational flexibility of the mutant receptor proteins resulting from the solubilization and protease digestion steps.
Thus, to facilitate the proper interpretation of disulfide cross-linking data, another major goal of the present study also was to establish a disulfide cross-linking protocol that would allow the formation of disulfide bonds using Cys-substituted mutant M 3 receptors present in their native membrane environment (in situ). Specifically, we generated a series of 10 mutant M 3 muscarinic receptors, all of which contained a Y254C point mutation at the bottom of TM V and a second Cys substitution within the C-terminal segment of TM VI (Lys 484 to Ser 493 ) (Fig. 1). All Cys substitutions were introduced into a mutant version of the M 3 muscarinic receptor, referred to as M3Ј(3C)-Xa (30), in which the central portion of the i3 loop was replaced with a factor Xa cleavage site and which, among other modifications, lacked most endogenous Cys residues (see Fig. 1 for details).
Following the transient expression of the 10 double Cys mutant M 3 receptors in COS-7 cells, we used a novel in situ cross-linking protocol to induce the formation of disulfide bonds between adjacent Cys residues, using Cu(II)-(1,10-phenanthroline) 3 (hereafter Cu (II)-phenanthroline) as a mild oxidizing agent. Cross-linking studies were carried out in the absence or in the presence of the muscarinic agonist, carbachol. The successful formation of disulfide cross-links was monitored by studying changes in the electrophoretic mobility of oxidized, factor Xa-treated receptors on SDS gels.
Using this strategy, we identified several double Cys mutant receptors that displayed agonist-dependent disulfide crosslinking. The observed pattern of disulfide bonds indicates that agonist-induced M 3 receptor activation involves a structural change that allows the cytoplasmic ends of TM V and VI to move closer to each other. Moreover, our data suggest that the cytoplasmic end of TM VI undergoes an activity-dependent increase in conformational flexibility, perhaps due to changes in local secondary structure. The in situ disulfide cross-linking strategy described here should be highly useful to probe the structural mechanisms involved in GPCR activation in a systematic fashion.  (20 Ci/mmol) was obtained from American Radiolabeled Chemicals. Sulfosuccinimidyl-2-(biotinamido)ethyl-1,3-dithiopropionate (EZ-Link TM Sulfo-NHS-SS-Biotin), ImmunoPure ® immobilized streptavidin, and the Micro BCA protein assay reagent kit were purchased from Pierce. The protease factor Xa was from Roche Molecular Biochemicals. Precast Novex® Tris-glycine polyacrylamide gels and SeeBlue®-Plus 2 prestained molecular mass standards were obtained from Invitrogen. Hybond TM ECL TM nitrocellulose membranes, anti-rabbit IgG antibodies conjugated to horseradish peroxidase, ECL TM detection reagents, and Hyperfilm TM ECL TM chemiluminescence film were from Amersham Pharmacia Biotech. Laemmli loading buffer was from Bio-Rad. All other reagents used were of the highest grade commercially available.
Construction of Mutant M 3 Muscarinic Receptors-All mutations were introduced into a pCD-based expression plasmid coding for the M3Ј(3C)-Xa mutant M 3 muscarinic receptor (Fig. 1). The construction of the M3Ј(3C)-Xa expression construct has been described previously (30). As shown in Fig. 1, the M3Ј(3C)-Xa receptor contained an Nterminal hemagglutinin tag and lacked all five potential N-terminal N-glycosylation sites and most endogenous Cys residues (except for Cys 140 , Cys 220 , and Cys 532 ). In addition, the central portion of the i3 loop (Ala 274 -Lys 469 ) was replaced by two factor Xa cleavage sites (in error we previously reported (30) that the M3Ј(3C)-Xa construct contains three rather than two factor Xa recognition sites). Cys residues were reintroduced into the M3Ј(3C)-Xa construct at positions Tyr 254 and Lys 484 -Ser 493 , by using the QuikChange TM site-directed mutagenesis kit (Stratagene) according to the manufacturer's instructions. Double Cys mutant receptors, in the M3Ј(3C)-Xa background, 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 Lys 484 to Ser 493 into the M3Ј(3C)-Xa construct containing the Y254C point mutation. The identity of all mutant constructs was verified by DNA sequencing.
Transient Expression of Receptor Constructs-All mutant M 3 muscarinic receptors were transiently expressed in COS-7 cells. The cells were grown 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 transfection, 1 ϫ 10 6 cells were seeded into 100-mm dishes. Cells were transfected with 4 g of DNA/100-mm dish using the LipofectAMINE TM Plus kit (Life Technologies), as per the manufacturer's instructions. If required, 1 M atropine was added to the incubation medium 24 h after transfection to increase the expression levels of the different mutant M 3 muscarinic receptors.
Preparation of Membrane Homogenates-Receptor-expressing COS-7 cells were harvested 48 h after transfection to obtain membrane preparations. Briefly, the culture medium was removed, and cells were washed twice with 10 ml of phosphate-buffered saline (PBS; pH 7.4) with the wash being left on the plates for a 10-min period to remove atropine. After these wash steps, muscarinic binding buffer (25 mM sodium phosphate, 5 mM MgCl 2 , pH 7.4) was added to the cells (1 ml/100-mm dish), and cells were placed at 4°C for 15 min. The cells were then scraped off the plates with a Teflon cell scraper, transferred to a centrifuge tube, and homogenized using a Polytron tissue homogenizer (speed 5, 20 s), followed by a 15-min centrifugation step (20,000 ϫ g; 4°C). The membrane pellets (particulate fractions) were resuspended in muscarinic binding buffer (1 ml/100-mm dish) and homogenized as before. Membrane homogenates were then aliquoted and stored at Ϫ70°C until needed. The protein concentration of membrane preparations was determined by using the Micro BCA protein assay reagent kit according to the manufacturer's instructions.
Radioligand Binding Assays-Radioligand binding assays were performed using membrane homogenates prepared from transfected COS-7 cells essentially as described (31). All binding assays were carried out in 25 mM sodium phosphate buffer (pH 7.4), supplemented with 5 mM MgCl 2 , in a final volume of 1 ml (ϳ20 g of membrane protein/ tube). Incubations were carried out for 2 h at 22°C (room temperature). In saturation binding assays, six different concentrations of the radioligand, [ 3 H]NMS, ranging from 20 to 5,000 pM, were tested. In competition binding assays, incubations were carried out with a fixed concentration of [ 3 H]NMS (500 pM) in the presence of 10 different concentrations of the competing ligand (carbachol). In all assays, nonspecific binding was defined as binding remaining in the presence of 1 M atropine. Saturation and competition binding curves were analyzed as described (32), by fitting data to a one-site model of ligand binding and the Hill equation using the program SigmaPlot 4.01 (SPSS Inc.).
PI Assays-Carbachol-induced increases in intracellular inositol monophosphate levels were determined using transfected COS-7 cells grown in six-well plates as described in detail previously (30). Following labeling of cells for 20 -24 h with myo-[ 3 H]inositol (3 Ci/ml), cells were incubated for 1 h at 37°C in the presence of 10 mM LiCl and increasing concentrations of carbachol. The inositol monophosphate fraction was isolated and quantitated as described (30). The dose-response curves were fitted to a four-parameter logistic function using the program SigmaPlot 4.01 (SPSS Inc.).
Disulfide Cross-linking with Cu(II)-Phenanthroline-Membrane preparations were thawed at room temperature and rehomogenized as described above (see "Preparation of Membrane Homogenates"). Membranes from one 100-mm dish (ϳ1 mg of protein; 1-ml aliquots) were incubated in microcentrifuge tubes with 2.5 M Cu(II)-phenanthroline, either in the presence of muscarinic agonists (1 mM carbachol, 1 mM oxotremorine, or 1 mM acetylcholine), in the presence of the muscarinic antagonist, atropine (1 M), or in the absence of ligands, for 10 min at room temperature with end over end rotation (30 rpm). Reactions were terminated by the addition of EDTA and N-ethylmaleimide (10 mM each), followed by a 10-min incubation on ice. Membrane extracts were then generated as described below.
Preparation of Membrane Extracts-Membrane extracts (lysates) were prepared using a protocol similar to that described by Zeng et al. (30). Membrane homogenates that had been subjected to different treatments (see above) were centrifuged at 8,000 ϫ g for 10 min at 4°C. The pellets were then incubated with 250 l of 0.2% digitonin in PBS (pH 7.4) for 20 min on ice. Following centrifugation at 2,000 ϫ g for 10 min at 4°C, the pellets were incubated with 1.2% digitonin in PBS (pH 7.4) or in factor Xa digestion buffer (50 mM Tris-HCl, pH 8, 100 mM NaCl, 1 mM CaCl 2 ), as required, for 90 -120 min at 4°C with end over end rotation (30 rpm). After centrifugation of samples at 20,000 ϫ g for 30 min at 4°C, the supernatants (membrane extracts) were transferred to fresh microcentrifuge tubes and stored at Ϫ70°C until use. Protein concentrations were determined by using the Micro BCA protein assay reagent kit according to the manufacturer's instructions.
Factor Xa Digestion of Mutant M 3 Muscarinic Receptors-Membrane extracts were incubated in factor Xa digestion buffer (50 mM Tris-HCl, pH 8, 100 mM NaCl, 1 mM CaCl 2 ) with factor Xa protease (final concentration, 0.1 g/l) at room temperature for 16 -20 h. Reactions were terminated by the addition of mammalian protease inhibitor mixture (Sigma; final concentration, 10%), followed by a 30-min incubation at room temperature. Samples were then placed at Ϫ70°C for at least 3 h prior to SDS-PAGE and immunoblotting.
[ 35 S]GTP␥S Binding Assays-[ 35 S]GTP␥S binding assays were performed using a procedure similar to that described by Lazareno et al. (33). COS-7 cells grown in 100-mm dishes were co-transfected with 3 g of receptor DNA or pCD vector (control) and 2 g of a pcDNAI-based plasmid directing the expression of wild-type mouse G␣ q (34). Approximately 48 h after transfections, membranes were prepared as described above and resuspended in a buffer containing 20 mM HEPES, 100 mM NaCl, and 10 mM MgCl 2 , pH 7.4 (buffer A). After brief homogenization with a Polytron homogenizer (20 s), membranes (about 60 g of protein) were incubated in buffer A (see above) containing 0.1 nM [ 35 S]GTP␥S, 0.1 M GDP, and 10 g/ml saponin, in the presence or absence of the agonist, carbachol (1 mM). Incubations were carried out for 60 min at room temperature in a 1-ml volume. Binding reactions were terminated by rapid filtration on presoaked GF/B filters followed by several washes with cold phosphate buffer (pH 7.4). The bound radioactivity was determined by liquid scintillation spectrometry.
Urea Treatment of Cell Membranes-To inactivate G proteins coupled to heterologously expressed mutant M 3 muscarinic receptors, we treated membranes prepared from receptor-expressing COS-7 cells with a high concentration of urea (35,36), using a procedure similar to that described by Lim and Neubig (36). Briefly, membranes were prepared from transfected COS-7 cells grown in a 100-mm dish as described above and centrifuged at 8,000 ϫ g (4°C) for 30 min. The pellets were resuspended in 1 ml of binding buffer (25 mM sodium phosphate buffer, pH 7.4, supplemented with 5 mM MgCl 2 ) and incubated on ice for 30 min, either in the presence or in the absence of 5 M urea. The samples were then recentrifuged (same conditions as above), and pellets were resuspended in 1 ml of binding buffer (urea-free) followed by a 30-min incubation on ice. Washed membranes were centrifuged (same conditions as above), and the final pellets were resuspended in 1 ml of the appropriate buffer to be used for GTP␥S binding or disulfide crosslinking studies.
Biotinylation of Cell Surface Receptors-Cell surface proteins were biotinylated using a protocol similar to that described by Chen et al. (37). About 48 h after transfection, COS-7 cells were washed twice with PBS as described above (preparation of membrane homogenates). Cells were then incubated for 15 min (room temperature) with EZ-Link TM Sulfo-NHS-SS-Biotin (1.5 mg in 1 ml of PBS; pH 8.0). After this initial incubation step, a new aliquot of EZ-Link TM Sulfo-NHS-SS-Biotin (1.5 mg in 1 ml of PBS; pH 8.0) was added, followed by another 15-min incubation at room temperature. The biotinylation solution was then removed, and 1 ml of 0.1 M glycine (in PBS; pH 8.0) was added to each dish, followed by a 20-min incubation at 4°C. Cells were then scraped off the plates, homogenized, and centrifuged as described above (see "Preparation of Membrane Homogenates"). Membrane pellets were resuspended in PBS (pH 7.4; 1 ml/100-mm dish), homogenized as described above, and stored at Ϫ70°C until use (see below).
Isolation of Biotinylated Cell Surface Proteins-Membrane extracts were prepared from surface-biotinylated COS-cells as described above (preparation of membrane extracts), followed by incubation of membrane extracts with 2% SDS for 30 min at 37°C. Subsequently, PBS (pH 7.4) containing 1% bovine serum albumin was added to the samples so that the SDS concentration was reduced to ϳ0.05%. In the next step, 30 mg of ImmunoPure ® immobilized streptavidin was added to the samples, which were then incubated at 4°C with end over end rotation (30 rpm) for 15-20 h. Subsequently, samples were centrifuged at 1,000 ϫ g at 4°C for 5 min, and the supernatants were discarded. The agarose pellets were then washed with 0.1% digitonin in PBS (pH 7.4). This centrifugation/wash step was repeated three times (the last wash step was carried out in the absence of digitonin). The agarose pellets were then incubated with Laemmli sample buffer (38) containing 50 mM dithiothreitol for 30 min at 37°C (reducing conditions), followed by SDS-polyacrylamide gel electrophoresis (PAGE) and Western blotting (see below).
Western Blotting Analysis-SDS-PAGE was performed essentially as described by Laemmli (38). Samples were incubated with Laemmli loading buffer for 30 min at 37°C (nonreducing conditions) and then loaded onto a Tris-glycine polyacrylamide gel, which was run at 125 V in the presence of 0.1% SDS. If required, 50 mM dithiothreitol was added to the samples prior to the initial incubation step (30 min at 37°C) to break up existing disulfide bonds (reducing conditions). Gels were blotted onto nitrocellulose membranes using the procedure described by Burnette (39). Membranes were then blocked with 5% fat free milk in PBS (pH 7.4) containing 0.05% Tween 20 (PBS-T) (16 h at 4°C). Following a wash with PBS-T, the membranes were incubated with 1 g/ml anti-M3 antibody (40) in PBS-T for 1 h at room temperature. The blots were then washed three times with PBS-T (10 min each wash) and incubated for 1 h at room temperature with a 1:3,000 dilution of anti-rabbit IgG antibody conjugated to horseradish peroxidase in PBS-T. After three 15-min washes with PBS-T, proteins were visualized by using ECL TM detection reagents and exposure to film. The intensities of immunoreactive bands were quantitated by scanning densitometry using the program NIH Image (National Institutes of Health).

RESULTS
A modified version of the rat M 3 muscarinic receptor (M3Ј(3C)-Xa) that lacked all potential N-terminal N-glycosylation sites and most native Cys residues and contained a factor Xa cleavage site within the i3 loop was used as a background for Cys-scanning mutagenesis studies ( Fig. 1; Ref. 30). Although the M3Ј(3C)-Xa construct is expressed at significantly lower levels than the wild-type M 3 receptor, it retains the ability to bind muscarinic ligands with high affinity and to interact with G proteins efficiently, in a fashion similar to the wild-type receptor (30).
To investigate activity-dependent changes in the relative positions of residues located at the cytoplasmic ends of TM V and VI in the M 3 muscarinic receptor, we reintroduced pairs of Cys residues into the M3Ј(3C)-Xa construct, thus generating 10 double Cys mutant receptors. All mutant receptors contained a Y254C point mutation at the bottom of TM V, N-terminal of the factor Xa cleavage site, and a second Cys substitution within the segment Lys 484 -Ser 493 at the cytoplasmic end of TM VI, C-terminal of the factor Xa cleavage site (Fig. 1).
Transient Expression of Mutant M 3 Muscarinic Receptors in COS-7 Cells (Atropine "Rescue")-All mutant receptor constructs were transiently expressed in COS-7 cells and initially examined for their ability to bind the muscarinic radioligand, [ 3

H]NMS. Incubation of membranes prepared from transfected COS-7 cells with a saturating concentration of [ 3 H]NMS (3 nM)
indicated that all double Cys mutant receptors, similar to the M3Ј(3C)-Xa receptor from which they were derived (30), were expressed at relatively low levels (Յ0.5 pmol/mg protein). To facilitate the development of a sensitive in situ disulfide crosslinking protocol, we considered it advantageous to devise a strategy that would lead to increased mutant receptor expression levels. Interestingly, recent studies have shown that treatment of transfected cells with antagonist ligands can lead to significant increases in the expression levels of certain mutant GPCRs (41,42). To examine the potential usefulness of this approach to boost the expression levels of the mutant M 3 muscarinic receptors analyzed in this study, we incubated transfected COS-7 cells with the muscarinic antagonist, atropine (1 M), for the last 24 h of culture. During the harvesting of cells, special care was taken to wash out atropine bound to cell membranes (for details, see "Experimental Procedures"). Strikingly, atropine treatment of transfected cells led to pronounced increases (Ն10-fold) in the expression levels of all analyzed mutant muscarinic receptors, as determined in [ 3 H]NMS saturation binding studies (shown for the M3Ј(3C)-Xa construct and a representative Cys double mutant, Y254C/A489C, in Fig.  2A). As indicated in Table I Table I).
To investigate whether the 10 double Cys mutant receptors were able to couple to G proteins, we next studied their ability to mediate carbachol-dependent increases in inositol phosphate production (PI hydrolysis). Like all other experiments described below, the PI assays were carried out with cells that had been grown in the presence of atropine (1 M) in order to boost receptor expression levels. As indicated in Table I, all FIG. 1. Structure of mutant M 3 muscarinic acetylcholine receptors analyzed in the present study. All Cys substitutions were introduced into the M3Ј(3C)-Xa mutant rat M 3 muscarinic receptor background (30). To generate the M3Ј(3C)-Xa construct, the M 3 receptor was modified as follows (30). The five potential N-glycosylation sites present in the N-terminal receptor segment (Asn 6 , Asn 15 , Asn 41 , Asn 48 , and Asn 52 ) were replaced with glutamine residues (shown circled), and the central portion of the i3 loop (Ala 274 -Lys 469 ) was replaced with two factor Xa cleavage sites ((IEGR) 2 ; double underlined). Most remaining endogenous cysteine residues (Cys 111 , Cys 516 , Cys 519 , Cys 542 , Cys 546 , Cys 560 , and Cys 562 ) were substituted with serine or alanine (filled squares) except for Cys 140 , Cys 220 , and Cys 532 , which proved to be essential for proper receptor expression and function (30). Ten M3Ј(3C)-Xa-based double Cys mutant receptors were constructed, all of which contained a Y254C point mutation at the intracellular end of TM V (boldface type), and a second Cys substitution within the segment Lys 484 -Ser 493 at the intracellular end of TM VI (boldface type and underlined). All receptor constructs contained an N-terminal hemagglutinin epitope tag. To detect the different mutant M 3 muscarinic receptors via Western blotting, we used a rabbit polyclonal antibody (anti-M3) directed against the indicated C-terminal receptor sequence (40). The numbers refer to amino acid positions in the rat M 3 muscarinic receptor sequence (58). mutant receptors retained the ability to stimulate the breakdown of PI lipids in an agonist-dependent fashion. The Y254C/ E485C and Y254C/Q490C receptors showed carbachol EC 50 values that were not significantly different from the corresponding value obtained with the M3Ј(3C)-Xa construct (ϳ70 nM). On the other hand, the remaining seven double Cys mutant receptors displayed significantly reduced carbachol EC 50 values (Table I). The most pronounced reduction in agonist potency (ϳ350-fold, as compared with the M3Ј(3C)-Xa construct) was observed with the Y254C/K487C mutant receptor ( Table I). Most of the analyzed mutant receptors also showed a significant reduction in basal inositol phosphate levels, as compared with the M3Ј(3C)-Xa receptor (set at 100% in Table I). In general, the extent of reduction in basal activity correlated well with the degree of reduction in carbachol potency displayed by the individual double Cys mutant receptors. It is likely that the reduced functional activity of most mutant receptors is due to the mutational modification of residues predicted to be critical for efficient G protein coupling (13)(14)(15)(16) and, probably to a minor extent, to the somewhat reduced B max levels displayed by some of the mutant receptors. However, as shown in Table I, all double Cys mutant receptors were able to stimulate the breakdown of PI lipids with high efficacy (E max , expressed as -fold increase in inositol monophosphate production above basal levels). The apparent increase in E max values displayed by many of the double Cys mutant receptors correlated very well with their reduced degree of basal activity ( Table I), suggesting that the observed increases in E max values probably do not reflect a "true" gain in G protein coupling efficacy.
Western Blot Analysis of Double Cys Mutant M 3 Muscarinic Receptors-We next examined the expression of the M3Ј(3C)-Xa construct and the 10 double Cys mutants receptors via Western blotting (reducing conditions) using a polyclonal antibody (anti-M3; Fig. 1) directed against the C terminus of the M 3 receptor protein (40). As shown in Fig. 2B for the M3Ј(3C)-Xa construct and a representative double Cys mutant receptor (Y254C/A489C), the anti-M3 antibody detected multiple immunoreactive bands, ranging in size from about 30 to 60 kDa. Interestingly, atropine (1 M) treatment of transfected cells for the last 24 h of culture resulted in the appearance of a pronounced band at ϳ38 kDa, which was barely visible in samples prepared from cells grown in the absence of atropine (Fig. 2B). Since the increase in the intensity of this band correlated well with the atropine-induced increase in the number of detectable [ 3 H]NMS binding sites ( Fig. 2A), we speculated that the ϳ38-kDa receptor species might correspond to properly folded cell surface receptors. It is likely that the additional immunoreactive species observed in Fig. 2B resulted from partial receptor degradation or complex formation with other proteins. Fig. 2B also shows that incubation of transfected COS-7 cells with the biotinylation reagent, Sulfo-NHS-SS-Biotin (2.5 M), had no effect on the observed protein expression patterns (see below).
Western Procedures." Subsequently, membrane extracts were prepared and processed in two different ways. B, membrane extracts were directly subjected to SDS-PAGE and Western blotting (reducing conditions), using the anti-M3 antibody. Note that atropine treatment resulted in a pronounced increase in the intensity of the 38-kDa receptor species, independent of whether or not cells were incubated with the biotinylation reagent. C, membrane extracts were first incubated with immobilized streptavidin to capture biotinylated cell surface proteins, followed by SDS-PAGE and Western blotting (reducing conditions), using the anti-M3 antibody. Note that the 38-kDa immunoreactive species represents the major biotinylated (cell surface) M 3 receptor species when transfected cells were grown in the presence of atropine (1 M). All samples were run on 12% Tris-glycine polyacrylamide gels (see "Experimental Procedures" for details). Protein molecular mass standards (in kDa) are indicated to the left.
background staining was seen in samples prepared from cells that had not been exposed to the biotinylation reagent (Fig.  2C). On the other hand, following biotinylation of cell surface proteins, the anti-M3 antibody detected multiple immunoreactive bands ranging in size from about 30 to 60 kDa. Analogous to the findings described in the previous paragraph, a prominent ϳ38-kDa band was observed with biotinylated samples prepared from cells grown in the presence of atropine (1 M). These observations, together with the results of the [ 3 H]NMS binding studies, strongly suggest that the 38-kDa species corresponds to properly folded cell surface receptors. Therefore, all disulfide cross-linking experiments described in the following were carried out with samples prepared from cells grown in the presence of atropine.
Agonist-promoted Disulfide Cross-linking of Mutant M 3 Muscarinic Receptors-To probe the potential proximity of Cys residues present on the intracellular surface of the 10 double Cys mutant receptors, we examined the ability of these Cys residues to form intramolecular disulfide bonds. To be able to perform such studies with receptors in their native conformations, we developed an experimental protocol that allowed the formation of disulfide cross-links with the mutant receptors being present in their natural membrane environment (in situ). The protocol that we employed is summarized in Scheme 1. Initially, membrane preparations derived from COS-7 cells individually expressing the M3Ј(3C)-X construct or the different double Cys mutant receptors were incubated with a low concentration (2.5 M; 10 min at room temperature) of the mild oxidizing agent, Cu(II)-phenanthroline, either in the absence of ligands or in the presence of the antagonist, atropine (1 M), or the agonist, carbachol (1 mM). Receptors were then 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 38-kDa receptor band under nonreducing conditions that is absent under reducing conditions is indicative of successful disulfide cross-linking.
The results of a representative cross-linking experiment carried out according to Scheme 1 are shown in Fig. 3. In this experiment, samples were processed for Western blotting studies under nonreducing conditions to ensure the integrity of disulfide bonds. Under basal conditions (incubation of samples with Cu(II)-phenanthroline with or without atropine (1 M)), only the Y254C/A489C double Cys mutant receptor yielded a 38-kDa signal that was clearly higher than background (this can be seen more clearly in Figs. 5 and 7). In contrast, when samples were oxidized in the presence of agonist (carbachol, 1 mM), the Y254C/A489C construct as well as three additional double Cys mutant receptors (Y254C/Q490C, Y254C/T491C, and Y254C/L492C) yielded a prominent 38-kDa band, indicative of agonist-promoted disulfide cross-linking (Fig. 3). These 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. 4 summarizes the results of three independent crosslinking experiments, based on the quantification of the intensity of the 38-kDa receptor band (nonreducing conditions) via scanning densitometry (NIH Image). To obtain a quantitative measure of agonist-induced disulfide cross-linking, the ratios of band intensities determined in the presence of agonist versus antagonist were calculated. In most cases, consistent with the representative Western blot shown in Fig. 3, the presence of agonist did not result in the appearance of a specific 38-kDa receptor signal. On the other hand, in the case of the Y254C/ A489C, Y254C/Q490C, Y254C/T491C, and Y254C/L492C mutant receptors, agonist treatment led to significant increases (by ϳ100 -200%) in the intensity of the 38-kDa receptor species SCHEME 1. Experimental procedure used for the induction (in situ) and detection of disulfide cross-links in mutant M3 muscarinic receptors.  (Fig. 4).
To investigate the agonist dependence of disulfide crosslinking in more detail, we carried out additional cross-linking experiments in which the Y254C/A489C, Y254C/Q490C, Y254C/T491C, and Y254C/L492C mutant receptors were oxidized in situ in the presence of different carbachol concentrations (0.001-1 mM). In all cases, increasing concentrations of carbachol led to concentration-dependent increases in the intensity of the 38-kDa receptor species (nonreducing conditions; all samples were processed and digested with factor Xa according to Scheme 1; Fig. 5). Cross-linking usually reached a plateau at around 10 -100 M carbachol (Fig. 5), in good agree-ment with the carbachol binding data indicating that the majority of receptors become saturated in this concentration range (Table I).
To examine whether carbachol-promoted disulfide crosslinking could be mimicked by other muscarinic agonists, the Y254C/A489C, Y254C/Q490C, Y254C/T491C, and Y254C/ L492C mutant receptors were also oxidized in the presence of oxotremorine or acetylcholine (1 mM each). In agreement with the findings obtained with carbachol, both oxotremorine and acetylcholine mediated a pronounced increase in disulfide cross-linking in each of the four mutant receptors (Fig. 6).
Agonist-promoted Disulfide Cross-linking Is Not Due to Receptor Oligomerization-Previous studies have shown that M 3 muscarinic receptors (43), like many other GPCRs (44), can form receptor dimers or oligomers. To facilitate the proper interpretation of the above disulfide cross-linking data, we therefore had to exclude the possibility that disulfide bonds were formed by intermolecular cross-links (between two or more different receptor molecules). In fact, processing of samples according to Scheme 1 is predicted to lead to the appearance of a 38-kDa band independent of whether disulfide bonds are formed intramolecularly or intermolecularly. To address this issue, we coexpressed an M3Ј(3C)-Xa-based mutant receptor containing the Y254C single point mutation with M3Ј(3C)-Xa-derived mutant receptor constructs containing the A489C, Q490C, T491C, or L492C single Cys substitutions. These mutant receptors were expressed at levels similar to the 10 double Cys mutant receptors, as determined in [ 3 H]NMS binding studies (data not shown). As shown in Fig. 7, oxidation and processing of samples prepared from co-transfected COS-7 cells, according to Scheme 1 cells failed to reveal significant intermolecular disulfide cross-linking (lack of 38-kDa receptor species). In contrast, the Y254C/A489C, Y254C/Q490C, Y254C/ T491C, and Y254C/L492C mutant receptors, which were included as positive controls in this set of experiments, showed the expected agonist-dependent increase in the intensity of the 38-kDa receptor species, indicative of successful disulfide crosslinking. These observations strongly suggest that the agonist promotes the formation of intramolecular rather than intermolecular disulfide bonds in these mutant receptors.
Agonist-promoted Disulfide Cross-linking Is Not Due to Agonist-induced G Protein Release-Since cross-linking studies were done with native membranes, we needed to exclude the possibility that the agonist dependence of disulfide cross-linking observed with some of the investigated mutant receptors was due the precoupling of receptors to heterotrimeric G proteins. For example, one can imagine a scenario in which receptor/G protein precoupling prevents the formation of disulfide bonds between the cytoplasmic ends of TM V and VI or interferes with the access of the oxidizing agent to the intracellular receptor surface in the inactive state of the receptor. In this case, cross-linking of adjacent cysteine residues exposed on the cytoplasmic side of the receptor would occur only after agonistdependent dissociation of the activated G protein from the receptor. To address this issue, disulfide cross-linking studies were also carried out with receptor-containing membranes that had been pretreated with high concentration (5 M) of the chaotropic agent, urea. Previous studies have shown that this strategy leads to the almost complete inactivation or removal of heterotrimeric G proteins, while leaving uncoupled receptors fully functional (35,36).
To confirm that urea (5 M) pretreatment (30 min on ice) of receptor-containing cell membranes was indeed able to disrupt receptor/G protein coupling, we examined the ability of the agonist, carbachol (1 mM), to stimulate [ 35 S]GTP␥S binding to receptor-containing COS-7 cell membranes. Preliminary studies had shown that cotransfection of receptor DNA with a plasmid coding for wild-type G␣ q resulted in more robust and reproducible [ 35 S]GTP␥S binding activity, as compared with cells transfected with receptor DNA alone (data not shown). All subsequent [ 35 S]GTP␥S binding assays were therefore performed with membranes prepared from cotransfected cells. To assess receptor-mediated [ 35 S]GTP␥S binding activity, membrane samples were incubated with the muscarinic agonist, carbachol (1 mM), for 60 min at room temperature, and carbacholdependent increases in [ 35 S]GTP␥S binding were determined by employing a standard filtration binding assay. As shown in Strikingly, after pretreatment of M3Ј(3C)-Xa-expressing membranes with urea, carbachol-mediated increases in [ 35 S]GTP␥S binding were greatly reduced (by ϳ75%; Fig. 8A), in good agreement with previous studies using a similar approach to uncouple heterologously expressed 5-HT 2C receptors (35). Moreover, urea-treated samples showed a drastic reduction (by ϳ80%) of basal [ 35 S]GTP␥S binding activity (untreated membranes: 18,973 Ϯ 317 cpm; urea-treated membranes: 4,129 Ϯ 84 cpm), consistent with urea-mediated inactivation or removal of membrane-attached G proteins.
Finally, we examined the effect of urea treatment on the ability of the Y254C/A489C, Y254C/Q490C, Y254C/T491C, and Y254C/L492C mutant receptors to form agonist-dependent disulfide cross-links in the presence of the oxidizing agent, Cu(II)-phenanthroline (2.5 M; 10 min). As shown in Fig. 8B, preincubation of membranes prepared from transfected COS-7 cells with urea (5 M) had essentially no effect on carbachol (1 mM)-dependent increases in disulfide cross-linking observed with each of the four mutant receptors (all samples were processed and digested with factor Xa according to Scheme 1). Taken together, these data strongly suggest that the observed agonist-induced changes in disulfide cross-linking patterns reflect conformational changes intrinsic to the receptor proteins rather than being caused by receptor/G protein precoupling.  1 M), or the agonists, oxotremorine (Oxo; 1 mM) or acetylcholine (ACh; 1 mM). Membrane preparations were the processed according to Scheme 1 (see "Experimental Procedures" for details). Samples containing equal amounts of protein (5 g) were then run on 10 -20% Tris-glycine polyacrylamide gels, followed by Western blotting analysis using the anti-M3 antibody (nonreducing conditions). Note that the agonists oxotremorine and acetylcholine, similar to carbachol, promoted disulfide cross-linking in the Y254C/A489C, Y254C/Q490C, Y254C/T491C, and Y254C/L492C mutant receptors. The 38-kDa receptor bands, which correspond to functional cell surface receptors (see "Results" for details), are shown. brane samples with carbachol (0.1 mM) alone, in the absence of oxidizing agent (data not shown). These observations suggest that the high agonist concentrations required to achieve a significant degree of disulfide cross-linking led to receptor desensitization, precluding the assessment of the effect of disulfide cross-linking on receptor function. DISCUSSION In the present study, we have used a site-directed disulfide cross-linking strategy to monitor agonist-induced conformational changes that occur during the activation of the M 3 muscarinic receptor. Disulfide cross-linking approaches have been used previously to investigate light-induced structural changes in the photoreceptor, rhodopsin (9,(25)(26)(27)(28)(29). However, this is the first study demonstrating the usefulness of a disulfide mapping strategy to investigate activity-dependent structural changes in a hormone-activated GPCR.
In the case of rhodopsin, disulfide cross-linking studies have been carried out with wild-type and mutant photoreceptor proteins following detergent solubilization and additional purification steps (9,(25)(26)(27)(28)(29). Similarly, to map intramolecular contact sites in the ground state of the M 3 muscarinic receptor, we recently used a strategy that involved the formation of disulfide bonds in mutant receptors that had been solubilized and treated with factor Xa protease (30). This latter approach may lead to increased flexibility of the analyzed receptor proteins and may also increase the likelihood of partial receptor unfolding, thus complicating the proper interpretation of experimental results. To overcome these potential difficulties, we developed a new experimental protocol that allows the formation of intramolecular disulfide bonds in Cys-substituted mutant M 3 receptors present in their native membrane environment (in situ). To our knowledge, this is the first study applying an in situ disulfide cross-linking strategy to the analysis of liganddependent conformational changes in a GPCR.
Specifically, we generated a series of 10 double Cys mutant M 3 receptors by introducing pairs of Cys residues into the M3Ј(3C)-Xa base mutant, which lacks most endogenous Cys residues and harbors a factor Xa cleavage site within the i3 loop (Fig. 1). All mutant receptors contained the Y254C point mutation at the C terminus of TM V and a second Cys substitution within the C-terminal portion of TM VI (positions 484 -493). In an attempt to boost mutant receptor expression levels, transfected cells were incubated with the muscarinic antagonist, atropine (1 M), for the last day of culture (41). This strategy led to pronounced increases (Ն10-fold) in the B max levels of all examined mutant receptors, as determined in [ 3 H]NMS binding studies. Preliminary studies 2 showed that incubation of transfected COS-7 cells with the membrane-impermeable atropine derivative, N-methylatropine, led to similar results, suggesting that atropine (or N-methylatropine) treatment increases the expression levels of the examined mutant M 3 receptors by binding to and stabilizing receptors present on the cell surface.
Radioligand binding studies showed that all double Cys mutant receptors were able to bind the muscarinic antagonist, [ 3 H]NMS, and the muscarinic agonist, carbachol, with high affinities. In addition, PI assays demonstrated that all mutant receptors retained the ability to activate G proteins, suggesting that all mutant receptor proteins were folded properly. However, in the PI assays, most double Cys mutant receptors showed a significant reduction in carbachol potency and basal activity, as compared with the M3Ј(3C)-Xa construct (Table I). This finding was not unexpected, since all mutant receptors lacked the functionally critical tyrosine residue at the bottom of TM V (Tyr 254 ) and contained additional point mutations within the C-terminal portion of TM VI, a region known to be intimately involved in G protein recognition and activation (12,17,18,(45)(46)(47).
When the 10 double Cys mutant receptors were subjected to the in situ disulfide cross-linking protocol outlined in Scheme 1, only the Y254C/A489C mutant receptor showed significant cross-linking in the absence of agonist (Figs. 5 and 7). However, the intensity of the cross-linking signal (38-kDa receptor species detectable after factor Xa digestion; nonreducing conditions) significantly increased when the Y254C/A489C mutant receptor was oxidized in the presence of the agonist, carbachol Y254C/Q490C, Y254C/T491C, and Y254C/L492C, which showed little or no specific cross-linking under basal conditions, also displayed agonist-dependent disulfide cross-linking (Figs. 3-7). In all four cases, the extent of disulfide bond formation correlated well with the fraction of receptors occupied by carbachol, as indicated by cross-linking studies carried out in the presence of increasing agonist concentrations (Fig. 5). In addition, control experiments showed that carbachol induced the formation of intramolecular disulfide bonds rather than promoting cross-linking between different receptor molecules (intermolecular disulfide bonds) (Fig. 7). In another set of control experiments, cell membranes prepared from transfected COS-7 cells were pretreated with urea (5 M) prior to disulfide cross-linking (Fig. 8). Consistent with previous reports (35,36), urea treatment greatly reduced the activity of heterotrimeric G proteins (Fig. 8A), probably due to denaturation of heterotrimeric G proteins and other nonintegral membrane proteins (36). In contrast, urea treatment had essentially no effect on the extent of agonist-dependent disulfide cross-linking displayed by the Y254C/A489C, Y254C/Q490C, Y254C/T491C, and Y254C/L492C mutant receptors (Fig. 8B). These findings strongly suggest that the observed agonist-induced changes in disulfide cross-linking patterns reflect conformational changes intrinsic to the receptor proteins rather than being caused by receptor/G protein precoupling. Taken together, the cross-linking data therefore support the concept that M 3 receptor activation is associated with a structural change that moves the cytoplasmic ends of TM V (containing Tyr 254 ) and VI (containing Ala 489 -Leu 492 ) closer to each other, thus facilitating crosslinking between cysteine residues present at these positions.
Very recently, Ghanouni et al. (48) carried out fluorescence spectroscopic studies using a modified version of the ␤ 2 -adrenergic receptor that contained a single fluorophore at position Cys 265 at the i3 loop/TM VI junction (corresponding to Ile 483 in the rat M 3 muscarinic receptor) and a single fluorescence quencher at position 224 at the bottom of TM V (corresponding to Lys 255 in the rat M 3 muscarinic receptor). In agreement with the results of our in situ disulfide cross-linking studies, spectroscopic analysis of the detergent-solubilized purified mutant receptor suggested that ␤ 2 -adrenergic receptor activation also involves a movement that places the cytoplasmic end of TM VI closer to that of TM V (48). Since these two independent approaches (Ref. 48 and present study) using different receptor subtypes arrived at a very similar conclusion, the observed structural change is likely to be conserved among all members of class I GPCRs and to be functionally relevant.
Consistent with a large body of mutagenesis data and theoretical models of GPCR structure (12,17,50), the high resolution crystal structure of the inactive state of bovine rhodopsin (49) predicts that Ala 489 , Gln 490 , Thr 491 , and Leu 492 are located within the C-terminal portion of the TM VI ␣-helix (Figs. 1 and  9). The high resolution rhodopsin structure also predicts that Tyr 254 (TM V) is located in relatively close proximity to Ala 489 (TM VI; predicted distance between ␣ carbons is ϳ11 Å). Consistent with this proposed structural arrangement, the Y254C/ A489C mutant receptor showed some degree of disulfide crosslinking in the inactive state of the M 3 receptor (see discussion above). In the ground state of the receptor, Gln 490 , Thr 491 , and Leu 492 are predicted to face away from Tyr 254 to other regions of the receptor or toward the lipid bilayer (Fig. 9), explaining why no specific disulfide cross-links were detectable with the Y254C/Q490C, Y254C/T491C, and Y254C/L492C mutant receptors in the absence of agonist or in the presence of the antagonist, atropine. However, as discussed above, these three mutant receptors, as well as the Y254C/A489C construct, underwent disulfide cross-linking in the presence of carbachol.
The use of two additional muscarinic agonists, oxotremorine and acetylcholine, resulted in disulfide cross-linking patterns similar to those observed with carbachol (Fig. 6). Based on the strict agonist dependence of these effects, the involved structural changes are likely to be relevant for the receptor activation process.
Based on the predicted positions of the cross-linked residues in the inactive state of the M3 receptor, the cross-linking patterns observed in the presence of muscarinic agonists are consistent with a model in which agonist activation of the M 3 muscarinic receptor leads to significant alterations in the secondary structure of the C-terminal segment of TM VI, involving perhaps a partial unraveling of this portion of the TM VI helix. Another possibility is that M 3 receptor activation causes pronounced increases in the rotational flexibility of the cytoplasmic end of TM VI, involving perhaps major backbone fluctuations. Consistent with this hypothesis, fluorescence spectroscopic studies (51) using a modified version of the human ␤ 2 -adrenergic receptor carrying a fluorescence reporter molecule at position Cys 265 at the i3 loop/TM VI junction also suggested that receptor activation leads to pronounced increases in the conformational flexibility of the region around Cys 265 . It is therefore likely that an increase in the conformational flexibility of this receptor domain is a prerequisite for the proper recognition and activation of G protein heterotrimers (6,7,51).
Interestingly, the results of several biophysical studies carried out with mutant versions of bovine rhodopsin (9) and mutant ␤ 2 -adrenergic receptors (48,52,54) suggest that receptor activation involves an outward and/or rotational movement of TM VI. For example, site-directed spin labeling experiments suggest that light-dependent activation of rhodopsin involves an "outward" movement of the cytoplasmic portion of TM VI, accompanied by a clockwise rotation of ϳ30 o , as viewed from FIG. 9. Predicted location of Tyr 254 , Ala 489 , Gln 490 , Thr 491 , and Leu 492 on the intracellular surface of the M 3 muscarinic receptor. The depicted structure was modeled according to the high resolution crystal structure of bovine rhodopsin (49), which corresponds to the inactive state of rhodopsin. The rhodopsin structure was modified such that Val 227 , Thr 251 , Arg 252 , Met 253 , and Val 254 in rhodopsin were replaced with the homologous M 3 receptor residues (Tyr 254 , Ala 489 , Gln 490 , Thr 491 , and Leu 492 , respectively). The intracellular receptor surface is viewed from the cytoplasm. Whereas Tyr 254 is predicted to be located at the C terminus of TM V (5), Ala 489 , Gln 490 , Thr 491 , and Leu 492 are thought to be contained within the cytoplasmic end of TM VI (6) (inactive state of the M 3 receptor). Note that Ala 489 , Gln 490 , Thr 491 , and Leu 492 project to different receptor domains or the lipid phase in the inactive state of the receptor. the cell interior (9). A similar "rigid body movement" is thought to occur during activation of the ␤ 2 -adrenergic receptor, as suggested by the use of fluorescence spectroscopic techniques (48,52,54) and the analysis of constitutively active mutant receptors using the "substituted Cys accessibility method" (53,55,56). In addition, based on fluorescence spectroscopic studies with mutant ␤ 2 -adrenergic receptors, Jensen et al. (54) recently proposed that ␤ 2 -adrenergic receptor activation may involve a relative "straightening" of helix VI around a conserved proline kink in the middle of TM VI (49,50), thus pushing the Cterminal segment of TM VI away from the receptor core deeper into the lipid bilayer.
The spectroscopic techniques used to study activity-dependent changes in rhodopsin (9) and the ␤ 2 -adrenergic receptor (48,51,52,54) measure the average of multiple possible receptor conformations. In contrast, 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 (21,57). It is therefore likely that the activity-dependent conformational changes detected by our disulfide cross-linking strategy occur in addition to a more pronounced "outward" rigid body movement of TM VI. One possibility is that the increased conformational flexibility of the cytoplasmic end of TM VI detected in the present study is required for the proper "docking" of G proteins to the intracellular surface of the receptor.
The high resolution x-ray structure of rhodopsin (56) suggests that at least two residues located at the cytoplasmic end of TM VI (corresponding to Glu 485 and Ala 489 in the M 3 muscarinic receptor) can interact with a highly conserved arginine residue (part of the conserved (E/D)R(Y/W) motif; corresponding to Arg 165 in the M 3 receptor) present at the C terminus of TM III. Several lines of evidence suggest that these contacts are required for keeping GPCRs of the rhodopsin family in an inactive state (see Ref. 56 and references therein). It is therefore likely that the agonist-dependent structural changes in the M 3 receptor protein detected in the present study destabilize these interactions, thus "freeing" the conserved arginine residue for interactions with heterotrimeric G proteins.
In conclusion, we have developed a novel disulfide crosslinking strategy that allows the monitoring of activity-dependent changes in the M 3 muscarinic receptor, using Cys-substituted mutant receptors present in their native membrane environment. Systematic application of this strategy to other receptor regions should eventually lead to a more comprehensive picture of the structural changes that convert a GPCR from its resting into its active state. Given the difficulty in overexpressing and purifying GPCRs in quantities sufficient for biophysical studies, the approach described here should be highly useful to study the structure and activity-dependent changes in other classes of GPCRs.