Abstract
Muscarinic M2 receptors preferentially couple with the Gi/o class of G-proteins to inhibit cAMP synthesis. However, they can also stimulate net synthesis of cAMP and inositol phosphate (IP) accumulation. We investigated in intact Chinese hamster ovary (CHO) cells expressing human M2 receptors (CHO-M2 cells) whether direct interaction of M2 receptors with Gs and Gq/11 G-proteins is responsible for the latter effects. Suppression of the Gsα subunit using RNA interference abolished stimulation of cAMP synthesis induced by 1 mM carbachol in both control and pertussis toxin-treated CHO-M2 cells but had no effect on the inhibition of forskolin-stimulated cAMP synthesis. Carbachol stimulated accumulation of IP with an EC50 of 79 μM. Removal of the Gq,G11, or both α subunits reduced this response by 78, 54, and 92%, respectively, whereas suppression of the Gsα subunit had no effect. Similar results obtained in CHO cells expressing M1 receptors that preferentially couple with Gs and Gq/11 G-proteins confirmed the efficiency of siRNA treatments. Stimulation of M2 receptors in control and pertussis toxin-treated cells by a series of full agonists with respect to inhibition of adenylyl cyclase displayed different efficacies in stimulating IP accumulation. Carbachol, acetylcholine, and oxotremorine-M [N,N,N-trimethyl-4-(2-oxo-1-pyrolidinyl)-2-butyn-1-ammonium] behaved as full agonists, furmethide (N,N,N-trimethyl-2-furanmethammonium) and methylfurmethide [(5-methyl-2-furyl)methyltrimethylammonium] were partial agonists, and oxotremorine (1-[4-(1-pyrrolidinyl)-2-butynyl]-2-pyrrolidinone) had no effect. Our results provide direct evidence of M2 receptor coupling with the α subunits of Gs and Gq/11 G-proteins and demonstrate induction of multiple receptor conformational states dependent on both the concentration and the nature of the agonist used.
Muscarinic receptors belong to a large family of membrane G-protein-coupled receptors with seven transmembrane domains. These receptors transmit various chemical signals from the extracellular space via four classes of heterotrimeric G-proteins that activate distinct intracellular signaling pathways. Muscarinic receptors have two principal binding sites, one interacting with the signaling molecule at the extracellular surface of membrane and the other interacting at the intracellular surface with G-proteins. The affinity of these two binding sites is mutually influenced by occupancy of the other binding site. Conformational change induced by extracellular ligand binding induces separation of α and βγ subunits of the G-protein complex, and both subunits can then activate their particular signaling pathways (Pierce et al., 2002).
There are five subtypes of muscarinic receptors designated as M1 to M5 receptors (Bonner, 1989; Bonner et al., 1987, 1988). Although individual subtypes preferentially interact with particular classes of G-proteins (Jones et al., 1991; Caulfield et al., 1994; Felder, 1995), specificity of their functional outcome is not absolute (Kostenis et al., 1997a,b, 2005; Milligan and Kostenis, 2006). We previously observed in Chinese hamster ovary cells (CHO), which specifically express individual subtypes of muscarinic receptors (Buckley et al., 1989), that stimulation of muscarinic M2 receptors preferentially inhibits adenylyl cyclase via Gi/o G-proteins (Michal et al., 2001). However, higher agonist concentrations lead to anomalous concentration-dependent reversal of this effect (Michal et al., 2001). Obliteration of M2 muscarinic receptor-mediated inhibition of adenylyl cyclase by pertussis toxin treatment results in revealing of strong stimulation of cAMP production. Furthermore, it was demonstrated that stimulation of porcine M2 receptors increases IP production in a concentration-dependent manner (Ashkenazi et al., 1987, Vogel et al., 1995). These and similar observations of anomalous functional outcome evoked by stimulation of various G-protein-coupled receptors (Ashkenazi et al., 1987; Peralta et al., 1987; Jones et al., 1991; Eason et al., 1992; Dittman et al., 1994; Vogel et al., 1995; Jakubík et al., 1996; Bonhaus et al., 1998; Xiao et al., 1999) may have important implications in certain pathological states and their treatment as well as in intoxication or drug overdosing.
It has been postulated that coincident activation of different signaling pathways and availability of corresponding G-proteins may explain unusual functional responses (Tuček et al., 2001, 2002). However, it has not been clearly demonstrated whether activation by the M2 receptor of nonpreferential cAMP synthesis or IP accumulation is effected through the direct activation of adenylyl cyclase or phospholipase C by the α subunit of the Gs or Gq/11 proteins, respectively, or if it is an indirect effect mediated by the βγ subunit of Gi/o consequent to activation of other signaling pathways. We investigated using RNA interference whether M2 receptor-mediated activation of alternative signaling pathways, namely stimulation of cAMP synthesis and IP accumulation, is mediated directly by the Gs and Gq/11 G-proteins, respectively. The increase of IP accumulation as well as the increase of cAMP synthesis elicited by M2 receptor stimulation appears to be largely mediated by the Gq/11 and Gs G-proteins, respectively, because knockdown of their α subunits strongly reduced or abolished these effects. Comparison of concentration-response relationships of IP accumulation and inhibition of cAMP synthesis evoked by six different muscarinic agonists indicated agonist-specific preference of signaling pathways.
Materials and Methods
Cell Cultures and Chemicals. Experiments were performed on CHO cells stably transfected with the human gene of the muscarinic M1 (CHO-M1 cells) and M2 (CHO-M2 cells) receptor subtypes kindly supplied by Professor Bonner (National Institute of Mental Health, Bethesda, MD). Cells were grown in Dulbecco's modified Eagle's medium with 10% fetal calf serum and 0.005% geneticin and used for experiments 3 to 4 days after seeding. In some experiments, pertussis toxin at a concentration of 100 ng/ml was present during the last 24 h of cultivation to inactivate Gi/o G-proteins. Chemicals were obtained from Sigma Chemical (Prague, Czech Republic) unless indicated otherwise.
Transfection. Transfection of cells was carried out by incubation with siRNA [sequences: Gsα subunit sense (5′-GGACAUCAAAAACAACCUGtt-3′) and antisense (5′-CAGGUUGUUUUUGAUGUCCtg-3′); Gqα subunit sense (5′-GAAGGUGUCUGCUUUUGAGtt-3′) and antisense (5′-CUCAAAAGCAGACACCUUCtc-3′); G11α subunit sense (5′-GAAGGUCACGACUUUUGAGtt-3′) and antisense (5′-CUCAAAAGUCGUGACCUUCtc-3′); Ambion, Huntingdon, Cambridgeshire, UK] and lipofectamine 2000 (Invitrogen, Paisley, UK) in Opti-MEM I (Invitrogen) for 6 h as recommended by the manufacturer. The concentration of siRNAs was 100 nM. The transfection mixture was then diluted using a 10 times volume of fresh Dulbecco's modified Eagle's medium, and cells were grown for additional 2 (siRNA against Gs) or 3 (siRNA against Gq and G11) days before being used for experiments. In preliminary experiments, 40 nM siRNAs did not have a full effect, and 200 nM siRNAs had no bigger effect than 100 nM siRNAs. G-protein α subunit suppression was in addition to reported functional assays confirmed using a scintillation proximity assay as described previously (Jakubík et al., 2006). In brief, membranes from control cells or cells transfected with siRNA against Gsα or Gqα plus G11α subunits were incubated to equilibrium with [35S]GTPγS (GE Healthcare, Little Chalfont, Buckinghamshire, UK) in the absence of GDP. Individual G-protein α subunits were separated using antibodies from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). In one experiment performed twice in quadruplicates, treatment with siRNA against the Gsα subunit reduced [35S]GTPγS binding by 95% to Gs but had no effect on binding to Gi/o or Gq/11 α subunits. Similarly, coincident treatment with siRNAs against Gqα and G11α subunit reduced [35S]GTPγS binding by 87% to Gq/11 but had no effect on binding to Gi/o or Gs α subunits.
Biochemical and Binding Assays. Inositol phosphate accumulation and cAMP production were assayed as described previously (Jakubík et al., 1996; Michal et al., 2001). Synthesis of cAMP was measured during 10-min incubations in medium containing 1 mM isobutylmethylxanthine. Accumulation of IP was determined in medium containing 10 mM lithium during 5- and 20-min incubations in CHO-M1 and CHO-M2 cells, respectively. The density of plasma membrane muscarinic receptors was determined as specific binding of2 nM [3H]N-methylscopolamine (Amersham Biosciences) in intact cells (Jakubík et al., 1995). The expression level was 5.46 ± 1.13 and 1.62 ± 0.12 pmol/mg protein in control M1 and M2 cells, respectively. Corresponding levels in cells transfected with siRNA against Gsα subunit were 3.24 ± 0.14 and 1.17 ± 0.11 pmol/mg protein. Cells subjected to sham treatment expressed 3.15 ± 0.24 (M1) and 1.41 ± 0.21 (M2) pmol/mg protein. All values represent the mean ± range from two independent experiments in triplicate.
Data Treatment. Curve fitting and statistical evaluation of data were done using Prism 4 (GraphPad Software Inc., San Diego, CA). Sigmoidal concentration-response curves with a slope of unity (Y = [bottom + (top – bottom)]/[1 + 10(log EC50–X)]) or variable slope (Y = [bottom + (top – bottom)]/[1 + 10(log EC50–X) × Hill slope)]) were fitted to the data as appropriate. X is the log of the agonist concentration, Y is the measured effect (cAMP synthesis or IP accumulation), and EC50 is the concentration of agonist required to produce half-maximal stimulation. A better fit was determined using an F test. Calculated parameters are expressed as means (95% confidence limits) of pooled data when two experiments on cells from independent seedings were performed or means ± S.E.M. of individual values when three or more independent experiments were done.
Results
Effect of Gsα Subunit Knockdown on cAMP Synthesis. Pertussis toxin treatment had no effect on resting cAMP level in CHO-M2 cells. However, suppression of Gsα subunits by RNA interference significantly reduced the resting level of cAMP by 26% in CHO-M2 and by 27% in CHO-M1 cells that served as a check of treatment efficiency. A decrease in the resting level of cAMP (23%) was also evident in pertussis toxin-treated CHO-M2 cells but did not reach statistical significance (Fig. 1, top row). In concert with previous experiments (Michal et al., 2001), 1 mM carbachol increased synthesis of cAMP by 2-fold in CHO-M2 cells, and pertussis toxin inactivation of Gi/o G-proteins about doubled the magnitude of this response (Fig. 1, bottom right). Knockdown of Gsα subunits inhibited 1 mM carbachol-stimulated cAMP synthesis by 93, 92, and 99% in CHO-M1, CHO-M2, and pertussis toxin-treated CHO-M2 cells, respectively (Fig. 1, bottom row). Negative siRNA had no influence on either resting or carbachol-stimulated synthesis of cAMP in both CHO-M1 and CHO-M2 and pertussis toxin-treated CHO-M2 cells.
In CHO-M2 cells, 10 μM forskolin stimulated cAMP synthesis by ∼10 times (Fig. 2, top). In line with previous findings, the concentration-response curve of cAMP synthesis inhibition by carbachol was biphasic. Low concentrations of carbachol decreased cAMP synthesis (Emax –0.89 ± 0.03% of incorporated radioactivity, IC50 0.35 μM, and Hill slope 1.67 ± 0.22; calculated disregarding concentrations higher than 10 μM) (Fig. 2, middle; Table 1). At concentrations higher than 10 μM, a reduction of inhibition became apparent. This reversal of cAMP inhibition was abolished by siRNA directed against the Gsα subunit. The concentration-response relationship after siRNA treatment fits a sigmoidal concentration-response curve with a variable slope (Emax –0.20 ± 0.01% of incorporated radioactivity, IC50 0.35 μM, and Hill slope 1.73 ± 0.34) (Table 1). Because the treatment significantly reduced both resting and forskolin-stimulated synthesis of cAMP (Fig. 2, top), the data were expressed as a percentage of control synthesis to bring them to the same scale (Fig. 2, bottom). The stimulation of cAMP synthesis induced by carbachol could be calculated as the difference between siRNA-treated and control cells. This transformation provided a curve of stimulation of cAMP synthesis that fits a sigmoidal concentration-response curve with variable slope (Emax 139.3 ± 4.7% of control, EC50 21.9 μM, and Hill slope 0.53 ± 0.07) (Table 1).
Effects Gqα and G11α Subunit Knockdown on IP Production. As shown in Fig. 3, carbachol stimulated accumulation of IP in CHO-M2 cells in a concentration-dependent manner with EC50 of 73.3 ± 9.7 μM (Table 2). Pertussis toxin treatment slightly, but not significantly, reduced Emax from 7.13 ± 0.34 to 5.89 ± 0.44% of loaded radioactivity and increased EC50 to 117.1 ± 8.9 μM. Resting values of IP production (Fig. 3, right) were slightly, but not significantly, higher in pertussis toxin-treated than in control cells (3.10 ± 0.64 and 2.54 ± 0.33% of loaded radioactivity in five experiments). Similarly, when expressed as a percentage of control (nontreated) cells in paired samples from four independent experiments, pertussis toxin treatment slightly increased resting IP production by 22.0 ± 8.3% (range from 9.6 to 46.1% of control), but this difference did not reach statistical significance. The increase of IP accumulation induced by 10 mM carbachol was completely prevented by 10 μM atropine.
Suppression of Gq and G11 α subunits in CHO-M2 cells (Fig. 4, left) significantly reduced accumulation of IP stimulated by 1 mM carbachol by 79 and 55%, respectively. Removal of both subunits had a more marked effect, as it inhibited 91% of the response. Lack of effect of negative siRNA and siRNA directed against the Gsα subunit suggested selectivity of the treatment. However, confirmation of this selectivity necessitates direct testing of the specificity of effects of the siRNAs used on various G-protein subunits. None of the siRNA treatments significantly influenced resting IP accumulation that was 1.08 ± 0.12, 0.85 ± 0.11, 0.77 ± 0.07, 0.76 ± 0.14, 1.19 ± 0.13, and 0.82 ± 0.08% of incorporated radioactivity in control cells and in Gqα,G11α,Gqα plus G11α, Gsα, and negative siRNA-transfected cells, respectively. As expected, carbachol induced in CHO-M1 cells a robust concentration-dependent increase of IP accumulation (Fig. 4, right, inset) with an EC50 of 0.86 μM and Emax amounting to 24.3% of incorporated radioactivity. Suppression of Gqα, G11α, and Gqα plus G11α reduced the responses to maximally effective 3 μM carbachol by 60, 48, and 83%, respectively. These effects are similar to those seen in CHO-M2 cells stimulated by a maximally effective carbachol concentration (1 mM). Similar to CHO-M2 cells, removal of Gq, G11, Gq plus G11, and Gs α subunits did not significantly influence resting IP formation that was 1.97 ± 0.33, 0.92 ± 0.40, 2.30 ± 0.56, and 0.88 ± 0.38 of incorporated radioactivity, respectively.
Effect of Muscarinic Agonists on IP Production. Effects of five additional muscarinic agonists that are full agonists with respect to preferential Gi/o G-protein-mediated inhibition of cAMP synthesis on IP accumulation were investigated and results are summarized in Fig. 5 and Table 2. With the exception of oxotremorine, all other tested agonists increased IP accumulation both in control (Fig. 5, left) and in pertussis toxin-treated cells (Fig. 5, right). Pertussis toxin treatment somewhat attenuated the maximal response of all effective agonists. However, the decrease in efficacy was significant only in the case of acetylcholine. Inactivation of Gi/o G-proteins by pertussis toxin significantly decreased the potency of all effective agonists with the exception of carbachol.
Discussion
Muscarinic M2 receptors preferentially couple to Gi/o G-proteins to inhibit cAMP synthesis. Nonetheless, it was shown that stimulation of M2 receptors with higher concentrations of agonists reduces inhibition of cAMP synthesis or even, depending on the level of receptor expression, increases cAMP synthesis (Michal et al., 2001) and also evokes IP accumulation (Ashkenazi et al., 1987, Vogel et al., 1995). It was not fully clarified, however, whether activation of these uncommon signaling pathways with respect to M2 receptors is triggered by their direct coupling with Gs, and Gq/11 G-proteins, respectively, as opposed to indirect effects mediated by the βγ subunits released from the Gi protein complex. In the present work, we used inactivation of specific G-proteins by their corresponding siRNA or pertussis toxin to investigate the mechanisms of coupling of M2 muscarinic receptors expressed in CHO cells to nonpreferred cellular signals. CHO-M1 cells were used as positive control for the effects of siRNA treatment, assuming that the only major difference in the two cell lines resides in the expressed receptor. Agonist stimulation of muscarinic M1 receptors was shown to directly activate IP production through Gq/11 G-proteins with high affinity (Felder et al., 1989; Gurwitz et al., 1994) and with lower affinity also directly increased cAMP synthesis through Gs G-proteins (Burford and Nahorski, 1996).
We previously found that pertussis toxin treatment in CHO-M2 cells reverts carbachol-induced inhibition of cAMP synthesis to net stimulation irrespective of the receptor expression level (Michal et al., 2001). This and similar observations on the M4 muscarinic receptor, another Gi/o G-protein-preferring receptor (Jones et al., 1991; Dittman et al., 1994), exclude a direct involvement of the Gi/o G-protein βγ subunit complex in mediating the increase in cAMP synthesis. Results of our experiments show that selective suppression of the Gsα subunit by siRNA treatment practically abolished the M2 receptor-mediated carbachol stimulation of cAMP synthesis in native as well as in pertussis toxin-treated cells, indicating that this atypical muscarinic response in CHO-M2 cells is predominantly mediated by direct interaction of the M2 receptor with the Gs G-proteins and activation of the Gsα subunit. Effectiveness of the siRNA treatment was verified by abolishment of carbachol stimulation of cAMP synthesis in CHO-M1 cells (Fig. 1, bottom left) that is mediated by the Gsα subunit (Burford and Nahorski, 1996). Selectivity of the treatment was confirmed by the lack of effect of negative siRNA on carbachol stimulation of cAMP synthesis and by the finding that M2 receptor-mediated inhibition of forskolin-activated cAMP synthesis was preserved after Gsα subunit suppression. Gsα subunit knockdown caused a significant decrease in resting cAMP synthesis in both CHO-M1 and CHO-M2 cells and in forskolin-stimulated cAMP synthesis in CHO-M2 cells. These observations demonstrate constitutive activity of Gsα G-proteins in CHO cells and are consistent with involvement of the Gsα subunit in forskolin-induced stimulation of cAMP synthesis (Yan et al., 1998). Despite the significant reduction of forskolin stimulation of cAMP synthesis, the potency and efficacy of carbachol in inhibiting cAMP synthesis was the same in control and treated cells when expressed as percentage of corresponding controls (Fig. 2; Table 1). The difference between the relative inhibition of cAMP synthesis in control and Gsα subunit knockdown CHO-M2 cells represents stimulation of cAMP synthesis. The EC50 of this calculated concentration-response curve (21.9 μM) (Fig. 2, bottom; Table 1) fits reasonably well with that estimated from increased cAMP synthesis directly measured in pertussis toxin-treated CHO-M2 cells (3.6 μM in Michal et al., 2001). Taken together, these results demonstrate that the Gsα subunit plays a predominant role in M2 receptor-induced synthesis of cAMP, and its activation explains more than 92% of the response.
Results of our experiments show that stimulation of the human M2 receptor in intact CHO cells increases IP production. This finding is in concert with the previously reported increase of IP accumulation shown in CHO cells expressing the porcine M2 receptor (Ashkenazi et al., 1987, Vogel et al., 1995). We have recently reported increased GTPγS binding to Gs and Gq G-proteins induced by the muscarinic agonists carbachol and xanomeline in membranes prepared from CHO-M2 cells (Jakubík et al., 2006). Our present results extend these observations by showing a direct interaction of the M2 receptor with Gs and G11/q G-proteins and their signaling pathways in intact cells. Reduction of carbachol-stimulated IP accumulation by suppression of Gqα or G11α subunits and additivity of these effects provides evidence for independent participation of these G-proteins in increasing accumulation of IP. Both negative siRNA and siRNA directed against the Gs G-protein had no effect on the stimulation of IP accumulation. These observations support the specificity of the treatment and demonstrate that a possible contribution of indirect stimulation of phospholipase C mediated by cAMP (Birnbaumer, 1992) does not play a major role in stimulation of IP accumulation in CHO-M2 cells. Treatment of CHO-M1 cells with siRNA for G11α, Gqα, or their combination confirmed efficacy of the treatment in targeting specific α subunits (Fig. 4, right).
Another well-established mechanism of activation of certain phospholipase C subtypes is through the βγ dimers released from various G-proteins (Birnbaumer, 1992; Camps et al., 1992; Katz et al., 1992). βγ dimers coupled not only to Gi/o but also to Gs and Gq/11 G-proteins could have participated in the IP response at the M2 receptor, because the concentration of carbachol we used was saturating for coupling of the receptor to all these G-proteins. A possible contribution of Gs G-protein βγ dimers was eliminated by the finding that prevention of receptor-Gs G-protein coupling by suppressing expression of the Gsα subunit had no effect on the IP response. However, we observed a consistent reduction of Emax for all agonists we tested in pertussis toxin-treated CHO-M2 cells, although it was statistically significant only for acetylcholine. There was also a consistent decrease of potency that was significant for all agonists with the exception of carbachol. Because pertussis toxin inactivates Gi/o and prevents dissociation of its βγ complex from the G-protein heterotrimer, our findings support a possible synergistic effect of the βγ dimer released from the Gi/o G-protein as shown by Zhu and Birnbaumer (1996).
We observed variable efficacy of agonists in stimulating IP accumulation at the M2 receptor, although the selected agonists show equal full efficacy in inhibiting cAMP synthesis. Most notable, oxotremorine did not stimulate IP production at all, although it is a full agonist with regard to reducing cAMP levels (Michal et al., 2001). Linear regression analysis of the relationship between enhancing IP accumulation and inhibition of cAMP formation resulted in a correlation coefficient of R2 = 0.0037 and a slope not significantly different from 0, indicating lack of correlation. Moreover, the potency of agonists in increasing IP production was generally ∼100-fold lower than their potency in inhibiting cAMP synthesis (Table 3). Furthermore, there was a marked discrepancy in the rank order of efficacy of agonists in stimulating IP accumulation in CHO-M2 cells (current study) and in increasing cAMP levels in pertussis toxin-treated cells (Michal et al., 2001) (Table 3). This is supported by the lack of statistical correlation between the two parameters (R2 = 0.103; slope not significantly different from 0). Potency of agonists that increased IP accumulation at M2 receptors was generally lower than that at stimulation of cAMP synthesis, with the largest difference being for carbachol (33-fold) (Michal et al., 2001) (Table 3). Some, but not all, of the reported uncommon outcomes of G-protein-coupled receptor stimulation evoked by classic agonists (Jones et al., 1991; Eason et al., 1992; Dittman et al., 1994; Vogel et al., 1995; Bonhaus et al., 1998; Xiao et al., 1999; Michal et al., 2001) or ectopic muscarinic receptor ligands (Jakubík et al., 1996, 1998, 2002, 2006) can be explained by coupling promiscuity resulting from favorable stoichiometry of receptors and various G-proteins (Tuček et al., 2002). However, all experiments presented here were run with the same CHO-M2 cells so that neither differences in receptor nor G-protein densities could be responsible for the observed variability in the efficacy or potency of various agonists in inducing different second messenger responses. Fluorescence lifetime analysis of activation of tagged β2-adrenergic receptors demonstrated the existence of multiple receptor conformations (Ghanouni et al., 2001a,b; Peleg et al., 2001; for reviews, see Kenakin, 2003; Urban et al., 2007). According to this model, differential activation of signaling pathways by different agonists depends on the affinity of distinct agonist-stabilized receptor conformations and various G-proteins. Our data are in line with this notion. Therefore, one should be cautious in extrapolating the effects of a given agonist on one cellular signaling pathway to others mediated by the same receptor.
In summary, our data provide strong evidence for direct coupling of muscarinic M2 receptors to the α subunits of Gs and G11/q, resulting in stimulation of cAMP synthesis and IP accumulation, respectively. The results also provide support for the concept of induction of multiple agonist-receptor conformations that preferentially couple to various signaling pathways and the notion that different agonists exhibit selectivity for preferential active receptor conformations.
Acknowledgments
We appreciate the invaluable help of Dr. Jan Jakubík with comments on the manuscript.
Footnotes
-
This work was supported by Project AV0Z50110509 and by Grants GACR305/05/P209, GACR305/05/0452, National Institutes of Health Grant NS25743, and MSMT CR LC554.
-
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
-
doi:10.1124/jpet.106.114314.
-
ABBREVIATIONS: CHO, Chinese hamster ovary cells; IP, inositol phosphate(s); CHO-M1, CHO-M2, Chinese hamster ovary cells expressing the M1 or M2 subtype of muscarinic receptors; GTPγS, guanosine 5′-O-(3-thio)triphosphate; ANOVA, analysis of variance.
- Received September 19, 2006.
- Accepted October 24, 2006.
- The American Society for Pharmacology and Experimental Therapeutics