点击显示 收起
【关键词】 Orthosteric
Alanine substitution mutagenesis has been used to investigate residues that make up the roof and floor of the muscarinic binding pocket and regulate ligand access. We mutated the amino acids in the second extracellular loop of the M1 muscarinic acetylcholine receptor that are homologous to the cis-retinal contact residues in rhodopsin, the disulfide-bonded Cys178 and Cys98 that anchor the loop to transmembrane helix 3, the adjoining acidic residue Asp99, and the conserved aromatic residues Phe197 and Trp378 in the transmembrane domain. The effects on ligand binding, kinetics, and receptor function suggest that the second extracellular loop does not provide primary contacts for orthosteric ligands, including acetylcholine, but that it does contribute to microdomains that are important for the conformational changes that accompany receptor activation. Kinetic studies suggest that the disulfide bond between Cys98 and Cys178 may contribute to structures that regulate the access of positively charged ligands such as N-methyl scopolamine to the binding pocket. Asp99 may act as a gatekeeper residue to this channel. In contrast, the bulkier lipophilic ligand 3-quinuclidinyl benzilate may require breathing motions of the receptor to access the binding site. Trp378 is a key residue for receptor activation as well as binding, whereas Phe197 represents the floor of the N-methyl scopolamine binding pocket but does not interact with acetylcholine or 3-quinuclidinyl benzilate. Differences between the binding modes of N-methyl scopolamine, 3-quinuclidinyl benzilate, and acetylcholine have been modeled. Although the head groups of these ligands occupy overlapping volumes within the binding site, their side chains may follow significantly different directions.
Most of the amino acids that contribute to the orthosteric binding site of the muscarinic acetylcholine receptors (mAChRs) are in the transmembrane (TM) domain. Point mutations (Wess, 1996; Heitz et al., 1999), particularly alanine-scanning (Hulme et al., 2003), have been particularly useful in identifying them. The putative contacts made by the inverse agonist (-)-N-methyl scopolamine (NMS) with the TM domain of the M1 mAChR are homologous to the contacts of ground-state rhodopsin with cis-retinal (Palczewski et al., 2000; Li et al., 2004), with the exception of Trp157 in TM4 (Lu et al., 2001) (residue 4.57 using the notation of Ballesteros and Weinstein, 1995).
In rhodopsin, the second extracellular loop (ECL2) forms an antiparallel -hairpin, in which the inner strand is disulfide-bonded to the top of TM3. The disulfide bond is a conserved feature of most of the rhodopsin-related G-protein-coupled receptors, including the M1 mAChR, where Cys98 and Cys178 are the linked residues (Kurtenbach et al., 1990). Thus, ECL2 is anchored within the transmembrane domain. In rhodopsin, in addition to the conserved Cys187 (designated position 0) amino acid, side chains at positions -6, -1, +1, +2, and +4 contribute to the retinal environment, delimiting the top of the binding site (Li et al., 2004). At position -6, Glu181 may be the ancestral counter-ion for the retinal Schiff's base (Terakita et al., 2004), possibly still playing this role in the metarhodopsin-1 intermediate (Lüdeke et al., 2005). No precise homolog of this residue exists in the M1 mAChR in which Arg171 and Thr172 are inserted C-terminal to Glu170 (position -8). The other retinal contact residues lie close to the polyene chain. Their homologs in the M1 mAChR are Gln177, Cys178, Tyr179, Ile180, and Phe182.
Five of these eight positions are conserved across the vertebrate mAChR subtypes. Alanine substitution of three (Arg171, Thr172, and Phe182 human M1 mAChR) reduced NMS or ACh affinity by 3-fold or less; effects on receptor signaling were not determined (Matsui et al., 1995). In the M1 mAChR, mutation of the disulfide-bonded Cys178 and its partner Cys98 to serine abolished binding and signaling (Savarese et al., 1992) but the less disruptive alanine mutants have not been studied. In the M3 mAChR, alanine substitutions of the homologous cysteines reduced the affinity of NMS by 50- to 800-fold; of carbachol, by more than 100-fold; and the potency of the carbachol-induced phosphoinositide (PhI) response, by 20,000-fold (Zeng et al., 1999), although a very recent study of the M2 and M3 mAChRs suggests smaller effects (Huang and Ellis, 2007).
At the bottom of the binding site of rhodopsin, Trp265 (6.48) and Tyr268 (6.51) make the single largest contribution to the burial of cis-retinal surface area (Palczewski et al., 2000; Li et al., 2004). The key tryptophan residue at 6.48 has been mutated to alanine and/or phenylalanine in M1, M2, and M3 mAChRs (Wess et al., 1993; Bourdon et al., 1997; Heitz et al., 1999), but no fully coherent data set that includes both binding and functional data has been obtained for the alanine mutant. The functional role of Phe197 (5.47), homologous to Phe212, which forms the base of the binding pocket for the -ionone ring of cis-retinal, has not been studied in the mAChRs, although it was a target for nonconservative inactivating mutations in a recent random mutagenesis study on the M3 mAChR (Li et al., 2007).
Here, we report the effects on binding and signaling of alanine substitution of the residues in ECL2 of the M1 mAChR that are homologous to the retinal contact sites in rhodopsin, including the disulfide-bonded Cys178. In addition, we have obtained functional data for the mutants studied previously by binding alone (Matsui et al., 1995) and have mutated Cys98 to disrupt the disulfide bond that anchors ECL2 to the top of TM3. The results, interpreted by molecular modeling, indicate that residues in ECL2 make minimal direct contacts with the antagonists NMS and (-)-3-quinuclidinyl benzilate (QNB) or with the agonist ACh in the bound state but that the integrity of several conserved ECL2 residues, particularly Cys178, is important for ACh-induced signal transduction. Kinetic studies suggest that ECL2, assisted by Asp99 (3.26), a residue studied previously by asparagine substitution (Fraser et al., 1989), may contribute to structures that regulate the access of ligands to the transmembrane binding pocket. Phe197, at the bottom of the orthosteric binding site, is important for the binding of NMS but not QNB and ACh, suggesting that the side chains of these ligands follow different paths within the binding site. Trp378 interacts with all three ligands and has a major role in transmitting the ACh-induced signal.
Equilibrium binding assays and phosphoinositide functional assays were carried out as described previously, with small variations (Jones et al., 1995; Lu et al., 1997; Lu and Hulme, 1999).
Materials
l-[N-methyl-3H]scopolamine methyl chloride ([3H]NMS; 84 Ci/mmol), l-quinuclidinyl[phenyl -4-3H] benzilate ([3H]QNB; 42–48 Ci/mmol), and D-[myo-3H]inositol (80 Ci/mmol) were purchased from GE Healthcare (Chalfont St. Giles, Buckinghamshire, UK). Unlabeled ligands [(-)-N-methyl scopolamine bromide, acetylcholine bromide, and atropine] were from Sigma-Aldrich (Gillingham, UK). The QuikChange kit was from Stratagene (La Jolla, CA).
Mutagenesis and Expression of Receptors
In brief, residues of the rat M1 mAChR were mutated to alanine or asparagine using the QuikChange method. Mutant receptors, cloned into the pCD expression vector, and validated by dideoxy sequencing, were transiently expressed in COS-7 cells by electroporation (Bio-Rad Gene Pulser, 15 µg of DNA, 260 V, 960 µF; Bio-Rad Laboratories, Hercules, CA). Membrane preparations were made after 72 h as described previously. Where necessary, expression levels of poorly expressed mutant receptors were rescued by treatment of the cultured cells with atropine (10-6 M) for 48 h before washing and harvesting for membrane preparations or phosphoinositide turnover assays, as described previously (Lu and Hulme, 1999). The mutants W164A, Q165A, G169A, T172A, Q181A, and S184A (hM1 background) were the kind gift of Dr. Nigel Birdsall.
Equilibrium Binding and Functional Assays
Binding of [3H]NMS and [3H]QNB to membrane preparations (15–30 µg of membrane protein per milliliter) was measured at 30°C in a buffer containing 20 mM sodium-HEPES, 100 mM NaCl, and 1 mM MgCl2, pH 7.5, using an assay volume of 1 ml and an incubation time of 2 ([3H]NMS) to4h([3H]QNB). Nonspecific binding was defined with 1 µM atropine. Assays were performed in quadruplicate. The binding reaction was terminated by rapid filtration on a Brandel or TomTec cell harvester. The mean expression level of the wild-type receptor was 1.84 ± 0.52 pmol [3H]NMS binding sites/mg of protein (mean ± S.D.). The expression of [3H]QNB binding sites was up to 40% greater in direct comparisons. We have reported previously that the quaternary antagonist NMS inhibits approximately 20% of [3H]QNB binding sites, with low affinity possibly reflecting the penetration of the more lipophilic tertiary antagonist into sealed vesicles containing receptors inaccessible to [3H]NMS (Ward et al., 1999). Low-affinity sites were not seen in competition experiments with the tertiary analog scopolamine (Ward, 1999). In the present experiments, 90 ± 1% (n = 5) of the wild-type [3H]QNB binding sites showed high-affinity (pKd = 9.6 ± 0.18) inhibition by (-)-NMS, the remainder having a pKd of approximately 6. The W378A mutant also gave 87 ± 2.4% of higher-affinity sites. The higher-affinity binding constant was used in further analysis. We have not further investigated the balance of the apparent discrepancy in binding capacity in this study and cannot exclude more complex explanations.
The binding of ACh was measured by inhibition of the binding of radioligand concentrations ranging from 0.01 to 3 nM as appropriate for the mutant and radioligand combination. NMS binding to low-affinity mutants that retained significant affinity for [3H]QNB was measured similarly. ACh-stimulated total PhI turnover was assayed for 30 min at 37°C in Krebs-bicarbonate solution supplemented with 10 mM LiCl after prelabeling of the transfected COS-7 cells grown in 12-well plates with [3H]inositol (1 µCi/ml) for 48 h. The accumulation of total [3H]inositol phosphates was measured (Jones et al., 1995). Assays were performed in triplicate.
Radioligand Dissociation Assays
Membranes were labeled to equilibrium by incubation with [3H]NMS (2 x 10-9 M) or [3H]QNB (3 x 10-10 M) for 2 or 4 h, respectively. Dissociation was initiated by a 10-fold dilution into 10-5 M concentration of unlabeled NMS, and binding was assayed at specified time points.
Data Analysis
Saturation binding curves or self-competition curves for [3H]antagonists were fitted to a one-site model of binding using Sigma Plot 8.0 or 10.0 (SPSS Inc., Chicago, IL) to yield a total concentration of binding sites and an affinity constant. The results are expressed as pKd. Expression levels (RT) were normalized to simultaneously transfected wild-type controls. Inhibition curves for ACh and other ligands were fitted to the Hill equation to yield an experimental pIC50 value and a slope factor (nH).The experimental pIC50 values were corrected for the Cheng-Prusoff shift, as necessary, to yield the tabulated corrected pIC50 values. Some ACh binding curves were also fitted to a two-site model of binding providing high- and low-affinity binding constants and the associated proportions of binding sites.
In addition, a dimeric receptor model of binding was used to analyze certain experiments (Hulme, 2006). This took the form [3H]NMSbound/RT = T1/T2, where T1 = KL1 x L + KA1 x KL3 x A x L + KL1 x KL2 x L2 and T2 = 1 + 2 x KL1 x L + KL1 x KL2 x L2 + 2 x KA1 x KL3 x A x L + 2 x KA1 x A + KA1 x KA2 x A2, where [3H]NMSbound is the measured concentration of specifically bound [3H]NMS. RT is the total concentration of [3H]NMS binding sites. KL1 and KL2 represent the intrinsic affinity constants for the binding of the first and second molecule, respectively, of [3H]NMS to the protomers of the presumed receptor dimer. This was assumed to occur without cooperativity, so that KL1 and KL2 were fixed at the values determined in direct binding experiments. Fitting the above expression to the ACh inhibition curve allowed KA1 and KA2, the intrinsic affinity constants for the binding of the first and second molecules of ACh, to be estimated, as well as the affinity constant KL3 for binding of [3H]NMS to the singly ACh-occupied dimer. These values were obtained in a logarithmic form. Radioligand dissociation time courses were fitted to single (or, where appropriate, double) exponential functions yielding a dissociation rate constant, koff, the initial concentration of bound radioligand, and the level of radioligand binding after full dissociation, which was equivalent to the level of nonspecific binding.
PhI dose-response curves were fitted to a four-parameter logistic function, yielding a pEC50 (-log M) value and basal and maximum responses (Ebasal, Emax). Slope factors were close to 1.0. In these experiments, Emax was 3020 ± 430 dpm/assay (n = 22) for the wild-type rM1 receptor and 4200 ± 1020 (n = 5) for the wild-type hM1 receptor. The ratios of Emax to Ebasal were 6.4 ± 0.8 and 3.9 ± 1.3, respectively. These values are not significantly different (P > 0.11). Likewise, the pEC50 values were very similar (6.96 ± 0.06, rM1; 7.07 ± 0.16, hM1). For the combined data set, the mean ± S.E.M. ACh-independent signaling activity of the wild-type receptor was 477 ± 96 dpm, and the mean ACh-stimulated maximum activity was 2993 ± 462 dpm/assay (n = 27) after subtraction of the null-transfection background (Enull, 250 ± 23 dpm).
For the combined wild-type data set, Ebasal was correlated with Emax value over a wide range of values (correlation coefficient = 0.67, n = 27, P < 0.001). This proportionality suggests that the large variations in the responses, indicated by the high S.E.M. values, may reflect differences in the density, efficiency of transfection of the cells, and efficiency of [3H]inositol labeling of the phosphoinositide pool dependent on the state of the cells. In contrast, the pEC50 values for ACh showed little variation and were not correlated with the Emax values.
To compensate for these variations, values of Ebasal and Emax were determined relative to wild-type values obtained from contemporaneous controls. To take a specific example, for the C98A mutant, Emax values in three independent experiments were 2973, 1749, and 1859 dpm; the corresponding Emax values for the wild-type rM1 in simultaneous parallel transfections were 8031, 3111, and 4678 dpm yielding ratios of 0.37, 0.56, and 0.40, the mean of which is 0.443 ± 0.06. After the application of a correction for the irreducible background PhI signal measured in null-transfections, which was assumed to be the same for wild-type and mutant transfections, this became 0.40 ± 0.07. The correction takes the form (Emax,mutant/Emax,wild-type)cor = (Emax,mutant/Emax,wild-type)(1 + r) - r, where r is the ratio Enull/(Emax,wild-type - Enull), having a value of 0.083 for the combined rM1 and hM1 data set. A similar correction was applied to the Ebasal ratios, for which r = 0.52. The resulting values are tabulated, as a percentage of wild type, in Supplementary Table S3.
Values of the signaling efficacy of the ACh-receptor complex (eA), adjusted for variations in receptor expression and background signaling, were calculated using equations (Lu and Hulme, 1999; Hulme, 2006) that are based on the free association of receptors and G-proteins in the cell membrane (Azpiazu and Gautam, 2004). eA = ([IC50/EC50(1 - B)] - 1)/RT when the Emax value of the mutant (as a fraction of the wild-type receptor) is greater than 0.9 or eA = [Emax/(1 - Emax)]/RT when the Emax of the mutant is less than 0.9. The parameter B represents receptor-dependent basal signaling calculated as a fraction of the wild-type ACh-stimulated maximum signal and had a mean value of 0.16 for the wild-type mAChR in these experiments. The two equations yield equivalent results, but when the potency ratio IC50/EC50 is large, Emax is close to 1; conversely, when Emax is substantially less than 0.9, IC50 EC50. eA values computed for the mutants were expressed relative to the wild-type receptor.
If the signaling unit is actually a dimer in which both subunits must be occupied to produce a maximum signal, as is the case for the mGluR (Kniazeff et al., 2004), these expressions yield values that represent the geometric mean of the values for the individual subunits.
Statistical Analysis
Experiments were repeated at least 3 times. Values are tabulated as mean ± S.E.M. Statistical comparisons of affinity and rate constants for mutants and wild-type controls were carried out by one-way analysis of variance followed by Dunnett's post hoc test backed up by paired Student's t tests in marginal cases. Where values such as expression levels had been normalized to the wild-type control, two-tailed t tests were used to ascertain the level of significance of differences from the wild type.
Molecular Modeling
Construction of the M1 Model. The initial model of the TM domain of human M1 receptor was constructed by homology with the published X-ray crystal structure of bovine rhodopsin (Palczewski et al., 2000). Alignment between the M1 receptor sequence and bovine rhodopsin was based on the "classic" motifs found in each TM region, the asparagine in TM1, the aspartate in TM2, the "DRY" motif (ERY in rhodopsin) of TM3, the tryptophan in TM4, and the conserved prolines in TM5, TM6, and TM7. These alignments were used with the standard homology modeling tools in the Quanta program to construct the seven helical bundle domain of the M1 receptor. The extracellular loop regions were subsequently added using a procedure developed in-house that makes use of a combined distance geometry sampling and molecular dynamics simulation (Blaney et al., 2001). The side chains of this model were then refined using the Karplus standard rotamer library (Dunbrack and Karplus, 1993). The final model was optimized fully (500 steps of Steepest Descent followed by 5000 steps of Adopted Basis Newton Raphson) with the CHARMM force field using helical distance constraints between the ith and i + 4th residues (except proline) within a range of 1.8 to 2.5 Å to maintain the backbone hydrogen bonds of the helix bundle. A more in-depth overview of how a similar Family A aminergic receptor, the 5-hydroxytryptamine2A receptor, was built is covered in a recent publication (Blaney, 2006).
Ligand Docking Studies. Ligand molecules were built and optimized within Spartan (SPARTAN SGI, version 5.1.3; Wavefunction, Inc., Irvine, CA) using an AM1 Hamiltonian. The atom-centered charges for the ligands, used in the docking studies, were natural atomic orbital charges calculated at the Hartree-Fock level within Spartan using a 6-31G* basis set.
Ligands were docked into the receptor model manually using a variety of low-energy starting conformations. Adjustments of the receptor protein side chains were made where necessary, always ensuring that these side chains were only in allowed rotametric states (Dunbrack and Karplus, 1993). Once again, full optimization of the receptor-ligand complexes was performed using CHARMM, the only constraints used being those that maintained the hydrogen bonding pattern of the helical bundle and the charge-charge interaction between the basic nitrogen of our ligand species and the acidic head group of the Asp3.32 on TM3. This procedure allows full relaxation of both the ligand and the whole protein, which is not possible with automated docking procedures.
The interaction energies between the ligand and individual amino acid residues were calculated by energy deconvolution and represent the summed steric and electrostatic contributions from all of the atoms in the residues interacting with the ligand. The force field used was CHARMM. Before calculating the interaction energies, the complex was minimized using a Steepest Descent algorithm followed by a Newton-Raphson algorithm, the nonbonded cutoff was 15 Å, and the constant dielectric was set to 5 during both the minimization steps and energy calculation.
The pKb value of the basic nitrogen atom of QNB was calculated using the Advanced Chemistry Development pKa calculator (available at http://www.acdlabs.com).
3H-Labeled Antagonist Binding and Receptor Expression Levels. The effects of the mutations on the expression of mAChR binding sites were measured by saturation binding assays using both [3H]NMS and [3H]QNB. Full details are given in Supplementary Table S1. Although measurements were made with both antagonists, the expression levels of binding sites relative to wild type were not significantly different, with the possible exception of F182A (P = 0.03). The weighted means of the two sets of measurements are plotted in Fig. 1a.
Fig. 1. The effects of alanine substitution of residues in extracellular loop 2 and in the transmembrane domain on the level of expression and binding affinity of M1 mAChRs in COS-7 cells. Full details are given in Supplementary Tables S1 and S2. a, effects on receptor expression level, calculated from saturation curves with [3H]NMS and [3H]QNB. Bars show the change in expression level relative to contemporaneously transfected wild-type controls. Except for F197A and W378A, the values show the weighted mean ± S.E.M. of two sets of measurements, one with [3H]NMS and the other with [3H]QNB (n = 3–10; Supplementary Table S1). Hatched bars indicate mutants in which expression was rescued by atropine treatment of the cells after transfection. *, P < 0.05, **, P < 0.01 with respect to wild type. b, effects on affinity for NMS and QNB. Changes in affinity relative to wild type are displayed on a log scale and represent the mean ± S.E.M. of 3 to 10 independent measurements. *, P < 0.05, **, P < 0.01 with respect to wild type (n = 14, [3H]NMS; n = 10, [3H]QNB). c, effects on affinity for ACh. Changes in pIC50 value relative to wild type are displayed and represent the mean ± S.E.M. of three to nine independent measurements. *, P < 0.05, **, P < 0.01 with respect to wild type (n = 27).
Alanine substitutions of individual amino acid residues in ECL2 reduced M1 mAChR expression levels by a maximum of 50% relative to wild type. In contrast, it was difficult or impossible to measure specific binding of the radioligands to mutants of Cys178, Cys98, and Phe197. In the case of C178A, the best case, the maximum observed level of binding was 12 ± 3% of wild type (n = 3). This increased to 40 to 50% after treatment of the transfected cells with atropine (10-6 M) for 48 h before washing and harvesting. The expression levels of C98A, C98A/C178A, and F197A were also raised to between 15 and 30% of wild type (Fig. 1a). The ability of atropine to act as a "pharmacological chaperone" for poorly expressed mutants has been described previously (Lu and Hulme, 1999).
Alanine substitution of the ECL2 loop residues Arg171, Tyr179, Ile180, and Phe182 marginally reduced the affinity of NMS from the wild-type value of 6.6 ± 0.5 x 109/M (Fig. 1b). The maximum effect was 5-fold (R171A). The effects of R171A and F182A agree with published data for the hM1 mAChR (Matsui et al., 1995) but differ from a recent study on the M3 mAChR (Scarselli et al., 2007). These mutations did not significantly affect the affinity of QNB (2.63 ± 0.9 x 1010/M). Single and double substitution of Cys98 and Cys178 reduced the affinities of both NMS (Fig. 2a) and QNB (Fig. 1b) by 10- to 20-fold, as did the nonconservative substitution C178N. In contrast, the mutation of Asp99 (TM3), Phe197 (TM5), and Trp378 (TM6) decreased the affinity of NMS by 8-, 37-, and 370-fold, respectively, but of these, only W378A significantly reduced the affinity of QNB, by 9-fold (Fig. 1b).
Fig. 2. Representative experiments showing membrane ligand binding and whole-cell signaling properties of the wild-type M1 mAChR (), C98A (), C178A (), and C98A/C178A () mutants expressed in COS-7 cells. a, inhibition of the specific binding of [3H]NMS (1.17 x 10-11 M, WT; 2.12 x 10-10 M, mutants) by unlabeled NMS. Curves are fits to a one-site model of binding giving pKd estimates of the following: WT, 9.5; C98A, 8.3; C178A, 8.4; and C98A/C178A, 8.2. b, inhibition of the specific binding of [3H]NMS (1.07 x 10-10 M, WT; 1.9–2.5 x 10-9 M, mutants) by ACh. Full lines are fits to the Hill equation giving Cheng-Prusoff-corrected pIC50 (nH) estimates of the following: WT, 4.82 (0.71); C178A 3.29 (0.57); and C98A/C178A, 3.00 (1.04); or to a two-site model of binding giving pKd values for C98A of 5.75 (37% of total sites) and 2.28 (63% of total sites). The broken line shows the fit of the dimeric receptor model to the C98A data with pKL1 = pKL2 = 8.46, pKL3 = 8.65, pKA1 = 4.65, and pKA2 = 2.62. c, ACh-induced [3H]PhI response. Curves are fits to a four-parameter logistic function giving pEC50 values of the following: WT, 7.3; C98A, 3.75; C178A, 3.51; and C98A/C178A, 3.40. Emax (percentage of WT) values were C98A, 40%; C178A, 55%; and C98A/C178A, 30%.
Acetylcholine Binding. Consistent with previous reports, the binding of ACh to wild-type M1 mAChRs was described by a slightly flattened binding isotherm with a mean apparent affinity constant of 5.8 ± 1.3 x 104/M and a mean Hill coefficient of 0.85, slightly but significantly less than 1.0 (P < 0.01; a representative experiment is shown in Fig. 2b). Because the ACh binding curve shows little or no GTP shift under the conditions used (Lu and Hulme, 1999), it is possible that the flattening of the binding curve reflects negative cooperativity between the elements of a dimeric receptor. Analysis using a dimeric receptor model (see Materials and Methods) provided estimates of 9.57 x 104/M and 3.83 x 104/M for the mean intrinsic affinity constants of the protomers of the wild-type receptor for ACh, with a negative cooperativity of 2.5-fold. Because this is quite small and because the pK values obtained from Hill analysis are very close to the mean of the pK values from the dimeric receptor analysis, the results presented are derived from fits using the Hill equation. Full details are given in Supplementary Table S2.
Among the ECL2 residues, only mutations of Arg171 and Ile180 decreased the affinity of ACh, by 3- to 4-fold (Fig. 1c). The effect of R171A agrees with published data (Matsui et al., 1995). We did not see a significant effect of F182A on ACh affinity, although an 8-fold reduction in carbachol affinity has been reported recently for the homologous mutation in the M3 mAChR (Scarselli et al., 2007). The Hill coefficients of the ACh binding curves were unaffected. Single and double substitutions of Cys98 and Cys178 decreased the mean apparent affinity of ACh by 25- to 70-fold (Fig. 1c). Overall, the means of the Hill coefficients estimated for the C178A and C98A/C178A mutants were not significantly different from 1.0, although a degree of flattening was observed for the C178A mutant in a few experiments (Fig. 2b). However, the mean Hill coefficient for ACh binding to the C98A mutant decreased relative to the wild type to a mean value of 0.42 ± 0.17 (n = 6), significantly different from the wild-type values (P < 0.01). In favorable cases, it was possible to resolve the binding curve into a predominant low- and a residual high-affinity population (Fig. 2b). Likewise, it was possible to fit the binding curve to a dimeric receptor model (Fig. 2b). Further analysis using the dimeric receptor model (see Materials and Methods) provided mean intrinsic pKd value estimates of 2.77 ± 0.24 (C98A), 2.92 ± 0.12 (C178A), and 2.83 ± 0.11 (C98A/C178A) for the low-affinity components of the dimer and 4.0 ± 0.27 (C98A), 3.4 ± 0.24 (C178A), and 2.9 ± 0.22 (C98A/C178A) for the corresponding high-affinity components. Thus, of these mutants, it was C98A that reproducibly manifested a high-affinity component. Further experiments are needed to examine whether this mutant exhibits an enhanced GTP shift.
Mutation of Asp99 and Trp378 (the latter measured by inhibition of [3H]QNB binding) reduced ACh affinity by 10-fold, but mutation of Phe197 had no significant effect (Fig. 1c). The Hill coefficient of the ACh binding curve was increased to 1.0 by the W378A mutation.
Phosphoinositide Turnover Response. In addition to the ECL2 loop mutations reported above, we characterized the PhI response of a series of ECL2 loop mutations of residues conserved in the mAChR subtypes that had been characterized previously in the hM1 receptor subtype at the level of binding (Matsui et al., 1995). Their expression levels were confirmed by both [3H]NMS and [3H]QNB saturation assays (data not shown), and were similar to wild type, as reported previously (Matsui et al., 1995).
Both rat and human wild-type M1 mAChRs expressed in COS-7 cells elicited robust and indistinguishable PhI responses to stimulation with ACh with EC50 values of 10-7 M (Fig. 2c). Significant agonist-independent basal activity (approximately 16% of the maximum ACh-stimulated signal) was seen in most experiments relative to mock-transfected controls. This could be inhibited by atropine (data not shown) and was taken into account in the analysis of the data (see Materials and Methods).
The mutant receptors exhibited alterations in basal signaling, the potency of ACh, and the Emax of the PhI response. The full set of results is summarized in Supplementary Table S3. These parameters depend on the expression level of the mutant receptor; the potency is also proportional to the affinity constant of ACh. To compensate for these dependences, we calculated a parameter designed to estimate the efficacy of the ACh-receptor complex for initiating a PhI signal (Lu and Hulme, 1999). The results are plotted in Fig. 3, which also shows the amino acid sequence of ECL2 and the probable positioning of the outer (1) and inner (2) elements of the modeled antiparallel hairpin.
Fig. 3. Plots of mutation-induced changes in signaling efficacy against sequence position. Full details of the effects of the mutations on the PhI signaling response are given in Supplementary Table S3. ACh binding affinity and potency in eliciting a PhI response and the receptor expression level were used to calculate the signaling efficacy of the response of each mutant to ACh as described previously (see Materials and Methods). Bars represent the logs of the signaling efficacy of each mutant relative to the wild-type receptor. Where no bar is shown, the value has not been determined. a, ECL2 mutants; the arrows represent the position of the outer and inner strands of the presumed -hairpin; b, TM domain mutants.
Of the ECL2 loop mutations whose binding properties are reported above, R171A, Q177A, and I180A reduced the signaling potency of ACh by 5-fold. Of the additional residues studied, Q165A and G169A gave 3-fold, and W164A and S184A caused 15- to 30-fold reductions. None of these mutations reduced the maximum PhI response, but W164A, G169A, and S184A showed diminished ACh-independent basal signaling activity. Calculations showed that the mutation of a subset of these residues, namely Trp164, Gln177, and Ser184, decreased the ACh signaling efficacy by up to 10-fold (Fig. 3a). In the case of Q181A, the small reduction in calculated signaling efficacy relies on the reported increase in ACh affinity for this mutant (Matsui et al., 1995).
The single and double mutations of Cys98 and Cys178 all diminished the potency of the ACh-induced PhI response by 3000-fold (a representative experiment is shown in Fig. 2c). In contrast to the other ECL2 point mutants, mutations of the cysteine residues significantly reduced the maximum ACh-induced signal and the agonist-independent basal signal (Fig. 2c). The cysteine mutations all caused large, 30- to 60-fold reductions in the calculated ACh signaling efficacy (Fig. 3, a and b).
The D99A mutation reduced the potency of ACh by 100-fold but did not reduce the calculated signaling efficacy of ACh (Fig. 3b), the reduction in PhI potency being attributable to a combination of reduced ACh affinity and receptor expression (Fig. 1, a and c).
Alanine substitution of the transmembrane domain residues Phe197 and Trp378 reduced ACh potency by 30- and 300-fold. The W378A mutation also abolished agonist-independent basal signaling. Calculations indicated that because the F197A mutant showed reduced expression level (Fig. 1a), the reduction in signaling efficacy was only 4-fold (Fig. 3b). In contrast, the effect of W378A was 60-fold.
Radioligand Dissociation Rate Constants. The measurement of dissociation time courses can provide information about the transition from the bound to the free state of the radioligand. Comparison of the dissociation rate constants with the equilibrium binding constants generates additional insight into the role of particular amino acid side chains in this process.
Wild-type and mutant receptors were labeled to equilibrium with a concentration of [3H]NMS (2 x 10-9 M) or [3H]QNB (2 x 10-10 M) sufficient to give 30% or greater receptor occupancy. Dissociation was started by a 10-fold dilution into 10-5 M unlabeled NMS. The full results are given in Supplementary Table S4. A graphical summary is also used in Fig. 8.
Fig. 8. Plots of mutation-induced changes in dissociation rate constant against changes in affinity. Values are represented on a log scale. The full line is the line of equivalence. a, NMS; b, QNB.
[3H]NMS dissociated from the wild-type receptor with a monoexponential time course corresponding to a half-time of 10.5 min (a representative experiment is shown in Fig. 4a), which is in good agreement with published values (Matsui et al., 1995). The ECL2 loop mutations R171A, I180A, and F182A accelerated the dissociation rate of [3H]NMS by approximately 2.5-fold, as did C98A (5-fold) and C178A (2.4-fold) (Fig. 4a). In contrast, the E170A, Y179A, and D99A mutations did not significantly affect the dissociation rate constant of [3H]NMS.
Fig. 4. Representative experiments showing dissociation time courses of radioligands from wild-type and mutant M1 mAChRs. Unless otherwise indicated, curves are fits to single exponentials normalized to the initial level of specific binding. koff values are in units per minute. Full details are given in Supplementary Table S4. a, dissociation of [3H]NMS. koff values are WT, 0.071 (); D99A, 0.052 (); C98A, 0.19 (); and C178A, 0.14 (). b, dissociation of [3H]QNB. koff values are WT, 0.0024 (); R171A, 0.0242 (); and Y179A, 0.0113 (). c, dissociation of [3H]QNB. koff values are WT, 0.0024 (); C98A, (double-exponential fit) 0.4 (49% of total) and 0.004 (); and C178A (double-exponential fit), 0.5 (82% of total) and 0.019 ().
The dissociation of [3H]QNB from the wild-type receptor was much slower, with a mean half-time of approximately 530 min. It was also monoexponential. Y179A and R171A increased the dissociation rate by 3- and 10-fold, respectively (Fig. 4b). The C98A and C178A mutations greatly accelerated the dissociation of [3H]QNB. In the case of C178A, approximately 75% of the radioligand dissociated in 5 min (Fig. 4c), yielding a 100-fold or greater increase in the rate constant. A small fraction (<20%) of a second slower process was seen in two of three experiments. In the case of C98A, a biexponential process was regularly observed (Fig. 4c), a fast phase (approximately 50% of the total) similar to that of C178A, and a second slower phase (koff = 3.07 ± 1 x 10-3/min; t1/2 = 225 min) more like the wild type. Because of the low levels of radioligand binding and rapid dissociation kinetics of the C98A and C178A mutants, it was not possible to determine the rate constants of the fast phase of dissociation precisely. The D99A mutation did not affect the dissociation rate of [3H]QNB, but the F197A and W378A mutations accelerated it by 10- and 38-fold, respectively; the dissociation time courses were monoexponential.
Molecular Modeling. Mutations of ligand binding domain residues in TM domains 3 and 4 have been reported to cause mainly parallel reductions in the affinities of NMS, QNB, and ACh (Lu and Hulme, 1999). In contrast, some mutations in TM domains 5, 6, and 7 have differential effects on their affinities (Ward et al., 1999; Allman et al., 2000; Lu et al., 2001), as exemplified by F197A in the present study. Some of the key observations are collated in Supplementary Table S5.
An optimized rhodopsin-based homology model of the M1 mAChR has been used to dock ACh, (-)-NMS, and (-)-QNB in a manner compatible with the effects of the TM domain alanine mutations on binding affinity (see Materials and Methods). In attempting to build suitable models, we have prioritized residues whose alanine substitution caused large (10-fold or greater) reductions in binding affinity, consistent with the participation of their side chains in direct ligand interactions, and have sought arrangements of these residues that allow them to participate in a coherent binding pocket for the ligand. We have assumed that smaller (less than 10-fold) effects may indicate indirect or "second shell" effects. However, we recognize that the quantitative effects of alanine substitution mutations may be modified by factors such as entropy-enthalpy compensation, the displacement of bound water molecules, the deletion of unfavorable interactions, or the development of compensating interactions, for instance, by the substitution of a solvent molecule for the deleted side chain.
The results of the docking experiments are represented in Figs. 5 and 6. The positively charged quaternary ammonium head groups of ACh and NMS form an ion pair with the carboxylate group of Asp105 (TM3). The tertiary nitrogen on the quinuclidine ring system of QNB has a pKb value of 8.69 (ACD pKa calculator) and is protonated under physiological conditions, also allowing it to form an ionic interaction with Asp105.
Fig. 5. Molecular model of the M1 mAChR binding site docked with acetylcholine (a), S-(-)-N-methyl scopolamine (b), and R-(-)-3-quinuclidinyl benzilate (c).
Fig. 6. 2-dimensional LIGPLOTS of the modeled M1 mAChR binding site docked with acetylcholine (a), S-(-)-N-methyl scopolamine (b), and R-(-)-3-quinuclidinyl benzilate (c). Hatched semicircles indicate residues involved in hydrophobic contacts. Broken lines show hydrogen bonds; for ACh, the indicated H-bond length is 3.32 Å; for QNB, it is 2.86 Å.
This interaction is reinforced by contacts of neighboring residues, particularly Tyr106, H-bonded to the carbonyl group of ACh, contacting the scopine ring of NMS, and the quinuclidine ring and ester oxygen, as well as one of the phenyl rings of QNB. Trp378 contributes to the floor of the binding site in all three cases. Thus, the head groups of these ligands occupy overlapping spaces within the binding pocket. Furthermore, the proximal parts of the side chains all make contact with Asn382, potentially reinforced, in the case of NMS, by a hydrogen bond between the hydroxyl methyl group of the tropic acid moiety of NMS and the amide group.
In contrast, the positions of the side chain termini are different. The phenyl ring of the tropic acid side chain of NMS is modeled as projecting into the transmembrane domain past Ala193 toward the ring of Phe197. The mutation A193G has been reported to decrease NMS affinity (Allman et al., 2000). The shorter side chain of ACh does not reach as far into the transmembrane domain, instead projecting toward TM5 at a level slightly below Thr189, the mutation of which decreases ACh affinity (Allman et al., 2000). In contrast, the two phenyl rings of the bensilyl moiety of QNB are modeled in a more superficial position in the binding site, interacting with Tyr106 and Trp157, and Asn382, respectively. Such additional side chain interactions may account for the enhanced affinity of QNB with respect to NMS, diminishing the relative importance of Tyr381 (TM6), whose mutation strongly decreases the affinities of NMS and ACh but not QNB (Ward et al., 1999).
Calculations of the energy of interaction between the ligands and individual side chains in the receptor model (see Materials and Methods) are qualitatively consistent with the mutagenesis data. The energies calculated by the energy deconvolution are the summed steric and electrostatic contributions from the residues interacting with the ligand. A caveat is that alanine substitution-induced changes in ligand binding affinity are not necessarily a direct measure of side-chain interaction energies and do not address the contribution of the polypeptide backbone and the -carbon.
The plots shown in Fig. 7 compare the two data sets on a normalized scale. They emphasize the primary importance of interactions at the level of Asp105 and Tyr106 in TM3 and Trp378, Tyr381, and Asn382 in TM6 for the binding of all three ligands. By comparison with the alanine substitution data, there is apparent overemphasis of the importance of the ionic interaction with Asp105 for the binding of ACh and NMS. The calculated contributions of Trp378, Tyr381, Asn382, and Val385 to the binding of ACh and NMS follow the mutagenesis data reasonably well. In the case of QNB, the contribution of Tyr381 is predicted to be less than that of Asn382, but the experimental difference is more pronounced, because the Y381A mutation has remarkably little effect on the affinity of QNB (Ward et al., 1999), whereas for Val385, the -carbon makes the predominant contribution to the calculated energy of QNB binding, and this is preserved in the alanine mutant. In TM7, the model suggests that Tyr404 contributes to NMS and ACh but not QNB binding, in agreement with the mutagenesis data.
Fig. 7. Comparison of mutation-induced changes in binding affinity (pKd) with computed contributions of amino acid side chains to the binding energies for acetylcholine, S-(-)-N-methyl scopolamine and R-(-)-3-quinuclidinyl benzilate on a normalized scale. For plotting, each individual pKd or deconvoluted energy value was scaled by division by the sum of the corresponding values taken over the following residues: Leu102, Tyr106, Ser109, Trp157, Trp378, Tyr381, Asn382, Tyr404, and Cys407; thus, for the ith residue the values plotted were pKdi/pKd, Leu102, Tyr106... Cys407 and Ei/ELeu102, Tyr106.... Cys407.
The selective importance of Phe197 for the binding of NMS is reproduced in the energy calculations. The substantial distance between the ligands and residues in ECL2, including Cys178, is also reflected in the low calculated interaction energies, although Tyr179 is predicted to make some contribution to the binding energy of ACh and QNB but not NMS. In fact, the Y179A mutation significantly accelerated the dissociation rate of QNB (Fig. 4b) but not NMS and did not affect the equilibrium affinity constant of ACh. At the top of TM3, Cys98 and Asp99 are also not predicted to make much contribution to ligand binding energy. We have not attempted to model the more global conformational changes that might be expected to result from the mobilization of ECL2 after mutagenic disruption of the disulfide bond. These may underlie the substantial effects of these mutations on ligand binding affinities.
The role of Trp157 deserves comment. In a previous model of the M1 mAChR, the ring of Trp157 was modeled as projecting toward the outer surface of the receptor (Lu et al., 2001). To account for the important effect of the W157A mutation on the affinities of all three ligands, we have instead modeled the indole ring of Trp157 as projecting inward toward the binding cavity, where it can form a parallel stacking interaction with the phenyl ring of Tyr106. This energetically favorable conformation may help to stabilize the contributions of Tyr106 to ACh and NMS binding in addition to any direct contributions that the indole ring of Trp157 may make. In the QNB complex, the stacking interaction is relaxed to incorporate one of the two phenyl rings of QNB. This reorientation of the side chain of Trp157 enables it to make a small direct contribution to ligand binding energy, although this is still not large enough to account for the effect of the W157A mutation.
The alanine substitutions of individual ECL2 residues reduced the affinities of the antagonist NMS and the agonist ACh by less than 5-fold and did not alter the affinity of QNB. Taking published results into account (Matsui et al., 1995; Gnagey et al., 1999; Scarselli et al., 2007), 18 of the 22 non–disulfide-bonded residues of ECL2 of the mAChRs have been mutated without identifying a candidate primary contact residue for orthosteric ligands. The sequence positions that showed reduced NMS affinity are homologous to retinal contact residues in rhodopsin, but in the M1 mAChR, these amino acids may be in the "second shell," shaping the environment of the binding site without making primary ligand contacts. This view is consistent with molecular models of the binding site.
Mutations of Trp164, Gln181, and Ser184 (fully conserved in the mammalian mAChR subtypes) and Gln177 (found in M1 and M4 becoming glutamic acid in M2, M3, and M5) reduced the calculated ACh signaling efficacy by up to 10-fold. Except for Gln181, these positions were not picked up in a recent random mutagenesis and alanine substitution study of the M3 mAChR, where carbachol was used as the activating ligand (Scarselli et al., 2007). Although noticeable, these effects are not as large as the 100- to 1000-fold reductions after mutations of key residues in the TM domain (Lu et al., 2001). Several corresponding residues in rhodopsin (M1 mAChR equivalent in brackets), namely Trp175 (Trp164), Asp190 (Gln181), and Thr193 (Ser184), participate in hydrogen bond networks to the adjacent TM helices 4 and 5 (Li et al., 2004). It is plausible that the corresponding M1 mAChR residues may have a similar role in microdomains linking ECL2 to the TM domain. In contrast to a recent study on the C5a receptor (Klco et al., 2005), none of the alanine substitutions caused reproducible constitutive activity.
Single and double mutations of Cys178 and Cys98 strongly reduced receptor expression levels, underlining the importance of these highly conserved residues for receptor folding and cell-surface expression. Nevertheless, an atropine-rescue protocol enabled us to characterize their properties.
The mutations caused 10- to 20-fold reductions in the affinities of NMS and QNB, less than initially reported for the M3 receptor (Zeng et al., 1999) but comparable with a very recent study on M2 and M3 mAChRs (Huang and Ellis, 2007) and much smaller than the effect of mutating direct ligand-contact residues (Ward et al., 1999). The nonadditivity is consistent with the disulfide bond between Cys98 and Cys178. The homogeneity of the effects of C98A, C178A, and Cys178N on the affinities of NMS and QNB argues that the cysteine residues do not contribute directly to the binding of either of these antagonists. It is more likely that the reduced affinity reflects general destabilization of the transmembrane binding pocket after scission of the link between TM3 and TM5.
The mutations of Cys98 and Cys178 caused larger, 80-fold reductions in ACh affinity, as seen for primary binding site residues such as Asp105 (Lu and Hulme, 1999). However, in TM3, mutation of residues one helical turn above and below Asp105 reduced ACh affinity 30-fold, but there were no large effects of mutating neighboring residues in ECL2 and no differential effects of the C178N and C178A mutations. Therefore, the reduction in ACh affinity also probably reflects a global disturbance of the transmembrane binding pocket consequent on weakening the constraints between the tops of TM domains 3 and 5, both of which harbor residues important for the binding of ACh (Allman et al., 2000). The solvent network within the binding pocket that influences the thermodynamics of ACh binding (Hulme et al., 2006) may also be strongly perturbed. The cumulative effects may explain the 60-fold reduction in ACh signaling efficacy.
For C98A, the residual population of high-affinity ACh binding sites observed in membrane binding studies may explain the biphasic [3H]QNB dissociation curve. These sites, if present in whole cells, cannot be functional in signaling because their apparent affinity for ACh exceeds the ACh potency for the PhI response. The free Cys178 sulfhydryl may be able to form a non-natural disulfide bond (Noda et al., 1994) or promote the formation of disulfide-linked oligomers (Zeng and Wess, 1999).
The D99A mutation reduced the affinity of ACh and NMS 10-fold without affecting the affinity of QNB or changing the computed signaling efficacy of ACh. These findings are consistent with the effect of the D99N mutation (Fraser et al., 1989). They contrast sharply with the effect of mutating Asp105, which strongly reduced QNB affinity and abolished ACh signaling (Fraser et al., 1989; Lu and Hulme, 1999).
A comparison of the mutation-induced changes in the dissociation rate constants with the binding constants of NMS provided further insight (Fig. 8a). The substitution of side chains that make a favorable contribution to the free energy of the ligand in the receptor binding site should decrease the affinity constant and the free energy barrier of the reverse (dissociation) reaction to equal extents as reported for Y404A (7.39) (Matsui et al., 1995). Although the effects were small, the predicted correlation was shown by the ECL2 point mutants that exhibited reduced NMS affinity (with the possible exception of Y179A).
In contrast, for the mutations of Asp99 and the disulfide-bonded cysteine residues, the increases in the dissociation rate constant were insufficient to account for the decreases in NMS affinity. Thus, their primary effect may be to decrease the association rate constant for NMS binding. Similar effects followed alanine mutations of tryptophan residues in ECL1 and at the tops of TM3 and TM7 (Matsui et al., 1995). Structures created by the disulfide link between Cys98 and Cys178 and individual residues such as Asp99 may contribute to a channel that allows hydrophilic ligands such as NMS to access the transmembrane binding pocket. Asp99 may be a "gatekeeper" residue to the entrance channel for positively charged ligands. This may resemble the entrance channel in acetylcholine esterase (Koellner et al., 2000).
The lipophilic tertiary amine QNB gave a different picture (Fig. 8b). Alanine substitution of Asp99 did not increase its dissociation rate. However, mutation of Cys98 and Cys178 induced a fast phase of QNB dissociation. This was accelerated beyond the expectation from the decrease in the affinity constant. Mutation of Arg171 also selectively accelerated QNB dissociation. Calculated from the product of the affinity constant and the dissociation rate constant, the association rate constant of QNB for the wild-type receptor was 5.7 x 105 M-1 · s-1, approximately 10% of the value of 7.3 x 106/M-1 · s-1 for NMS, in agreement with published estimates (Waelbroeck et al., 1991). Compared with NMS, the access of the bulkier hydrophobic QNB to the wild-type receptor binding site is selectively hindered. QNB may not pass through a restrictive hydrophilic channel as readily as NMS. However, mutation of the disulfide-bonded cysteine residues, or of Arg171, increased the calculated association rate constant of QNB up to 10-fold. These mutations may destabilize the structure of the receptor, amplifying breathing motions that facilitate the access of QNB to the binding site by an alternative route. Structural destabilization, detected by accelerated dissociation of QNB, may also follow the mutation of well-packed residues such as Phe197 and Trp378 in the TM domain.
At the bottom of the binding pocket, mutagenesis and modeling suggest that Trp378 (6.48) contributes directly to the binding pocket for NMS, QNB, and ACh, although possibly with different side chain orientations. In the M3 mAChR, the mutation of tryptophan 6.48 to phenylalanine reduced NMS affinity by only 5-fold but ACh affinity by 18-fold (Wess et al., 1993). Thus, the aromatic character of Trp378 may be key for the binding of NMS but less important for the ground-state binding of ACh, as found for Tyr381 (6.51) (Ward et al., 1999). Trp378, like Tyr381, is crucial for receptor signaling, suggesting that both of the aromatic residues bracketing the kink-inducing Pro380 (6.50) may develop enhanced interactions with ACh in the activated state of the receptor.
In marked contrast, the F197A mutation reduced the affinity of NMS 40-fold without affecting QNB or ACh affinity. To achieve consistency, molecular modeling studies suggest that the side chains of NMS and QNB follow distinct vectors within the binding site (Fig. 5). The phenyl ring of the tropic acid side chain of NMS may extend deep into the transmembrane region toward TM5, whereas the bulkier benzilate group of QNB is retained at the top of the binding pocket, interacting with the upper parts of TMs 3, 4, and 6. This docking posture is consistent with the lack of effect of mutations in TM7 on QNB affinity (Lu et al., 2001). It differs from a recent ab initio modeling prediction (Peng et al., 2006). Dissimilar docking of NMS and QNB might affect not only the kinetics of formation and breakdown but also the functional properties of the receptor-ligand complexes, perhaps accounting for differences in cooperativity between NMS, QNB, and allosteric ligands (Lanzafame et al., 2006) and affecting the potential for inverse agonism (Schwartz et al., 2006). The shorter side chain of ACh also fails to penetrate far into the TM domain and is unlikely to interact strongly with TM5 in the ground state of the receptor.
Acknowledgements
We are grateful to Dr. Nigel Birdsall (National Institute for Medical Research, United Kingdom) for the gift of mutations in ECL2 and for helpful comments on the manuscript.
ABBREVIATIONS: mAChR, muscarinic acetylcholine receptor; TM, transmembrane domain; ECL extracellular loop; PhI, phosphoinositide; ACh, acetylcholine; NMS, (-)-N-methyl scopolamine; [3H]NMS, l-[N-methyl-3H]scopolamine methyl chloride; QNB, (-)-3-quinuclidinyl benzilate: [3H]QNB, l-quinuclidinyl[phenyl-4-3H] benzilate; WT, wild type.
The online version of this article (available at http://molpharm.aspetjournals.org) contains supplemental material.
【参考文献】
Allman K, Page KM, Curtis CAM, and Hulme EC (2000) Scanning mutagenesis identifies amino acid side chains in transmembrane domain 5 of the M1 muscarinic receptor that participate in binding the acetyl methyl group of acetylcholine. Mol Pharmacol 58: 175-184.[Abstract/Free Full Text]
Azpiazu I and Gautam N (2004) A fluorescence resonance energy transfer-based sensor indicates that receptor access to a G protein is unrestricted in a living mammalian cell. J Biol Chem 279: 27709-27718.[Abstract/Free Full Text]
Ballesteros JA and Weinstein H (1995) Integrated methods for the construction of three dimensional models and computational probing of structure-function relations in G-protein-coupled receptors. Methods Neurosci 25: 366-428.
Blaney FE (2006) Approaches to the molecular modelling of 7-transmembrane helical receptors, in Current Protocols in Pharmacology (Enna SJ ed), vol 9, pp 8.1-8.41, John Wiley and Sons Inc., New York.
Blaney FE, Raveglia LF, Artico M, Cavagnera S, Dartois C, Farina C, Grugni M, Gagliardi S, Luttmann MA, Martinelli M, et al. (2001) Stepwise modulation of neurokinin-3 and neurokinin-2 receptor affinity and selectivity in quinoline tachykinin receptor antagonists. J Med Chem 44: 1675-1689.
Bourdon H, Trumpp-Kallmeyer S, Schreuder H, Hoflack J, Hibert M, and Wermuth CG (1997) Modelling of the binding site of the human m1 muscarinic receptor: experimental validation and refinement. J Comput Aided Mol Des 11: 317-332.
Dunbrack RL and Karplus M (1993) Backbone-dependent rotamer library for proteins. Application to side-chain prediction. J Mol Biol 230: 543-574.
Fraser CM, Wang CD, Robinson DA, Gocayne JD, and Venter JC (1989) Site-directed mutagenesis of m1 muscarinic acetylcholine receptors: conserved aspartic acids play important roles in receptor function. Mol Pharmacol 36: 840-847.
Gnagey AL, Seidenberg M, and Ellis J (1999) Site-directed mutagenesis reveals two epitopes involved in the subtype selectivity of the allosteric interactions of gallamine at muscarinic acetylcholine receptors. Mol Pharmacol 56: 1245-1253.[Abstract/Free Full Text]
Heitz F, Holzwarth JA, Gies JP, Pruss RM, Trumpp-Kallmeyer S, Hibert M, and Guenet C (1999) Site-directed mutagenesis of the putative human muscarinic M2 receptor binding site. Eur J Pharmacol 380: 183-195.
Huang XP and Ellis J (2007) Mutational disruption of a conserved disulfide bond in muscarinic acetylcholine receptors attenuates positive homotropic cooperativity between multiple allosteric sites and has subtype-dependent effects on the affinities of muscarinic allosteric ligands. Mol Pharmacol 71: 759-768.[Abstract/Free Full Text]
Hulme EC (2006) Systematic mutagenesis of M1 muscarinic acetylcholine receptors, in G Protein-Coupled Receptors (Haga T and Takeda S eds) pp 137-178, CRC Press, Boca Raton.
Hulme EC, Lu ZL, and Bee MS (2003) Scanning mutagenesis studies of the M-1 muscarinic acetylcholine receptor. Receptors Channels 9: 215-228.
Hulme EC, Soper AK, McLain SE, and Finney JL (2006) The hydration of the neurotransmitter acetylcholine in aqueous solution. Biophys J 91: 2371-2380.
Jones PG, Curtis CA, and Hulme EC (1995) The function of a highly-conserved arginine residue in activation of the muscarinic M1 receptor. Eur J Pharmacol 288: 251-257.
Klco JM, Wiegand CB, Narzinski K, and Baranski TJ (2005) Essential role for the second extracellular loop in C5a receptor activation. Nat Struct Mol Biol 12: 320-326.
Kniazeff J, Bessis AS, Maurel D, Ansanay H, Prezeau L, and Pin JP (2004) Closed state of both binding domains of homodimeric mGlu receptors is required for full activity. Nat Struct Mol Biol 11: 706-713.
Koellner G, Kryger G, Millard CB, Silman I, Sussman JL, and Steiner T (2000) Active-site gorge and buried water molecules in crystal structures of acetylcholinesterase from Torpedo californica. J Mol Biol 296: 713-735.
Kurtenbach E, Curtis CAM, Pedder EK, AitkenA, Harris ACM, and Hulme EC (1990) Muscarinic acetylcholine receptors: peptide sequencing identifies residues involved in antagonist binding and disulfide bond formation. J Biol Chem 265: 13702-13708.[Abstract/Free Full Text]
Lanzafame AA, Sexton PM, and Christopoulos A (2006) Interaction studies of multiple binding sites on m4 muscarinic acetylcholine receptors. Mol Pharmacol 70: 736-746.[Abstract/Free Full Text]
Li B, Scarselli M, Knudsen CD, Kim SK, Jacobson KA, McMillin SM, and Wess J (2007) Rapid identification of functionally critical amino acids in a G protein-coupled receptor. Nat Methods 4: 169-174.
Li J, Edwards PC, Burghammer M, Villa C, and Schertler GF (2004) Structure of bovine rhodopsin in a trigonal crystal form. J Mol Biol 343: 1409-1438.
Lu ZL, Curtis CAM, Jones PG, Pavia J, and Hulme EC (1997) The role of the aspartate-arginine-tyrosine triad in the m1 muscarinic receptor: mutations of aspartate 122 and tyrosine 124 decrease receptor expression but do not abolish signaling. Mol Pharmacol 51: 234-241.[Abstract/Free Full Text]
Lu ZL and Hulme EC (1999) The functional topography of transmembrane domain 3 of the M1 muscarinic acetylcholine receptor, revealed by scanning mutagenesis. J Biol Chem 274: 7309-7315.[Abstract/Free Full Text]
Lu ZL, Saldanha J, and Hulme EC (2001) Transmembrane domains 4 and 7 of the M1 muscarinic acetylcholine receptor are critical for ligand binding and the receptor activation switch. J Biol Chem 276: 34098-34104.[Abstract/Free Full Text]
Lüdeke S, Beck M, Yan EC, Sakmar TP, Siebert F, and Vogel R (2005) The role of Glu181 in the photoactivation of rhodopsin. J Mol Biol 353: 345-356.
Matsui H, Lazareno S, and Birdsall NJM (1995) Probing of the location of the allosteric site on m1 muscarinic receptors by site-directed mutagenesis. Mol Pharmacol 47: 88-98.
Noda K, Saad Y, Graham RM, and Karnik SS (1994) The high affinity state of the 2-adrenergic receptor requires unique interaction between conserved and non-conserved extracellular loop cysteines. J Biol Chem 269: 6743-6752.[Abstract/Free Full Text]
Palczewski K, Kumasaka T, Hori T, Behnke CA, Motoshima H, Fox BA, Le Trong I, Teller DC, Okada T, Stenkamp RE, et al. (2000) Crystal structure of rhodopsin: A G protein-coupled receptor. Science 289: 739-745.[Abstract/Free Full Text]
Peng JY, Vaidehi N, Hall SE, and Goddard WA 3rd (2006) The predicted 3D structures of the human M1 muscarinic acetylcholine receptor with agonist or antagonist bound. ChemMedChem 1: 878-890.
Savarese TM, Wang CD, and Fraser CM (1992) Site-directed mutagenesis of the rat m1 muscarinic acetylcholine receptor: role of conserved cysteines in receptor function. J Biol Chem 267: 11439-11448.[Abstract/Free Full Text]
Scarselli M, Li B, Kim SK, and Wess J (2007) Multiple residues in the second extracellular loop are critical for M3 muscarinic acetylcholine receptor activation. J Biol Chem 282: 7385-7396.[Abstract/Free Full Text]
Schwartz TW, Frimurer TM, Holst B, Rosenkilde MM, and Elling CE (2006) Molecular mechanism of 7TM receptor activation—a global toggle switch model. Annu Rev Pharmacol Toxicol 46: 481-519.
Terakita A, Koyanagi M, Tsukamoto H, Yamashita T, Miyata T, and Shichida Y (2004) Counterion displacement in the molecular evolution of the rhodopsin family. Nat Struct Mol Biol 11: 284-289.
Waelbroeck M, Tastenoy M, Camus J, and Christophe J (1991) Binding kinetics of quinuclidinyl benzilate and methyl-quinuclidinyl benzilate enantiomers at neuronal (M1), cardiac (M2), and pancreatic (M3) muscarinic receptors. Mol Pharmacol 40: 413-420.
Ward SDC (1999) The role of residues Tyr381 to Val387 in transmembrane domain six of the rat M1 muscarinic acetylcholine receptor in agonist binding and receptor activation, Ph.D. thesis, University of London, London, UK.
Ward SDC, Curtis CAM, and Hulme EC (1999) Alanine-scanning mutagenesis of transmembrane domain 6 of the M1 muscarinic acetylcholine receptor suggests that Tyr381 plays key roles in receptor function. Mol Pharmacol 56: 1031-1041.[Abstract/Free Full Text]
Wess J (1996) Molecular biology of muscarinic acetylcholine receptors. Crit Rev Neurobiol 10: 69-99.
Wess J, Nanavati S, Vogel Z, and Maggio R (1993) Functional role of proline and tryptophan residues highly conserved among G protein-coupled receptors studied by mutational analysis of the m3 muscarinic receptor. EMBO J 12: 331-338.
Zeng FY, Soldner A, Schoneberg T, and Wess J (1999) Conserved extracellular cysteine pair in the M3 muscarinic acetylcholine receptor is essential for proper receptor cell surface localization but not G protein coupling. J Neurochem 72: 2404-2414.
Zeng FY and Wess J (1999) Identification and molecular characterisation of m3 muscarinic receptor dimers. J Biol Chem 274: 19487-19497.[Abstract/Free Full Text]
作者单位：Division of Physical Biochemistry, MRC National Institute for Medical Research, Mill Hill, London, United Kingdom (E.C.H.); and GlaxoSmithKline, New Frontiers Science Park, Harlow, Essex, United Kingdom (J.A.G., C.J.L., B.T.)