Coupling of the Muscarinic m2 Receptor to G Protein-activated K+ Channels via Gαz and a Receptor-Gαz Fusion Protein

G protein-activated K+ channel (GIRK), which is activated by the Gβγ subunit of heterotrimeric G proteins, and muscarinic m2 receptor (m2R) were coexpressed in Xenopus oocytes. Acetylcholine evoked a K+ current, I ACh, via the endogenous pertussis toxin (PTX)-sensitive Gi/o proteins. Activation of I ACh was accelerated by increasing the expression of m2R, suggesting a collision coupling mechanism in which one receptor catalytically activates several G proteins. Coexpression of the α subunit of the PTX-insensitive G protein Gz, Gαz, induced a slowly activating PTX-insensitive I ACh, whose activation kinetics were also compatible with the collision coupling mechanism. When GIRK was coexpressed with an m2R·Gαz fusion protein (tandem), in which the C terminus of m2R was tethered to the N terminus of Gαz, part of I ACh was still eliminated by PTX. Thus, the m2R of the tandem activates the tethered Gαz but also the nontethered Gi/o proteins. After PTX treatment, the speed of activation of the m2R·Gαz-mediated response did not depend on the expression level of m2R·Gαz and was faster than when m2R and Gαz were coexpressed as separate proteins. These results demonstrate that fusing the receptor and the Gα strengthens their coupling, support the collision-coupling mechanism between m2R and the G proteins, and suggest a noncatalytic (stoichiometric) coupling between the G protein and GIRK in this model system.

G protein-activated K ؉ channel (GIRK), which is activated by the G ␤␥ subunit of heterotrimeric G proteins, and muscarinic m2 receptor (m2R) were coexpressed in Xenopus oocytes. Acetylcholine evoked a K ؉ current, I ACh , via the endogenous pertussis toxin (PTX)-sensitive G i/o proteins. Activation of I ACh was accelerated by increasing the expression of m2R, suggesting a collision coupling mechanism in which one receptor catalytically activates several G proteins. Coexpression of the ␣ subunit of the PTX-insensitive G protein G z , G␣ z , induced a slowly activating PTX-insensitive I ACh , whose activation kinetics were also compatible with the collision coupling mechanism. When GIRK was coexpressed with an m2R⅐G␣ z fusion protein (tandem), in which the C terminus of m2R was tethered to the N terminus of G␣ z , part of I ACh was still eliminated by PTX. Thus, the m2R of the tandem activates the tethered G␣ z but also the nontethered G i/o proteins. After PTX treatment, the speed of activation of the m2R⅐G␣ z -mediated response did not depend on the expression level of m2R⅐G␣ z and was faster than when m2R and G␣ z were coexpressed as separate proteins. These results demonstrate that fusing the receptor and the G␣ strengthens their coupling, support the collision-coupling mechanism between m2R and the G proteins, and suggest a noncatalytic (stoichiometric) coupling between the G protein and GIRK in this model system.
Members of the G i/o family of heterotrimeric G proteins (G i1, G i2 , G i3 , G o1 , G o2 , and G z ) regulate numerous effectors such as adenylyl cyclase, ion channels, protein kinases, etc. (1)(2)(3)(4)(5). A great number of heptahelical G protein-coupled receptors (GPCRs) 1  2). It is widely accepted that the overall GPCR-effector coupling specificity is defined by factors such as colocalization or scaffolding of the signaling components, the presence of additional regulatory proteins such as regulators of G proteins signaling, effector/G protein specificity, etc. (6 -9). However, in the G i/o -related pathways, these factors still remain poorly understood. One proposed mechanism of ensuring a specific activation of a certain G␣ by a GPCR is the existence of a stable complex between the receptor and the G protein in the absence of agonist. The existence of such complexes is supported by several lines of data, among them coimmunoprecipitation of several GPCRs with the corresponding G␣ proteins in many cell types (6 -9). However, in other systems, a collision coupling-type mechanism (10,11) between GPCRs and the G proteins has been demonstrated. In these systems, the receptor is not a priori coupled to G␣, and the coupling takes place only after the binding of an agonist to the receptor. A receptor activated in this way shuttles between and catalytically activates several G proteins in succession.
To study activation and effector coupling of individual G␣ i/o proteins in separation from the other members of the family, it is necessary to overcome the problem that each of the G i/o proteins can be activated by almost any G i/o -interacting GPCR. Attempts to achieve specific coupling in GPCR-G␣ pairs were made by creating GPCR⅐G␣ fusion proteins (12)(13)(14). However, in G i/o -containing GPCR⅐G␣ tandems, the tethered receptor still activates "nearby" nontethered G␣ i/o proteins (15). This problem has been partly overcome (14) by utilizing the pertussis toxin (PTX) sensitivity, which is a characteristic of all G␣ i/o proteins except G␣ z (16,17). PTX catalyzes ADP-ribosylation of a cysteine near the end of the C terminus, preventing the coupling of the agonist-bound receptor to G␣, the dissociation of G␣ and G ␤␥ subunits, and the activation of their effectors (1). Changing the C-terminal cysteine to glycine or serine renders G␣ insensitive to PTX, but it can still be activated by the receptors (18,19). Thus, after PTX treatment, the GPCR of the tandem containing such a mutant G␣ i/o interacts only with the tethered G␣. However, another problem arose. In the best studied case, a tandem of the ␣2A adrenoreceptor with a PTXresistant mutant G␣ i1(C351G) , both the receptor-activated GTPase activity and the coupling to the effector (adenylyl cyclase) were substantially impaired as compared with the wildtype G␣ i (20,21). This is not surprising, given the importance of the C-terminal end of G␣ in receptor recognition and coupling (18,19) and the devastating effects of some (although not all) mutations of the C-terminal cysteine on receptor-G␣ coupling (22,23).
To avoid the use of mutant G␣ and to still be able to compare effector activation by free versus receptor-fused G␣ i/o protein, we utilized the naturally PTX-resistant G␣ z (24,25) and the Xenopus oocyte expression system. G␣ z is activated by a variety of G␣ i/o -coupled GPCRs and inhibits adenylyl cyclase like other G␣ i/o proteins (reviewed in Refs. 16 and 17), but G␣ z shows a slower GDP-GTP exchange rate and a very low GTPase activity (17). Recently, endogenous ␣-adrenergic receptors have been shown to inhibit N-type Ca 2ϩ channels and to activate the G protein-activated, inwardly rectifying K ϩ channels (GIRK) in a PTX-resistant manner in sympathetic neurons after overexpression of G␣ z (26). Both Ca 2ϩ channel inhibition and GIRK activation are mediated by a direct interaction of these channels with G ␤␥ normally released from G␣ i/o proteins (4,(27)(28)(29). The use of GIRK channels expressed in Xenopus oocytes as an assay to study the receptor-G protein-effector coupling has been widely utilized (28,30). It allows a controlled expression of different amounts of proteins under study, simply by injecting different amounts of the encoding RNAs, enabling a quantitative study of various aspects of the coupling mechanism. Furthermore, the binding and unbinding of G ␤␥ to and from the channel are fast. The rise and decay times of the GIRK current upon agonist application and washout are believed to be limited primarily by the rate of G protein activation (normally, the GDP-GTP exchange at the G␣) and by the rate of GTP hydrolysis by G␣, respectively (28,31,32). Therefore, the kinetics of GIRK currents reflect the kinetics of G protein activation and deactivation. Here, using this system, we demonstrate a collision coupling-type (10) mechanism in activation of the GIRK by m2R via PTX-sensitive G i/o proteins and via G z and a substantial improvement of the efficiency of coupling by tethering m2R and the G␣ in tandem.

EXPERIMENTAL PROCEDURES
cDNA Constructs and mRNA-Materials and enzymes for molecular biology were from Roche Molecular Biochemicals, Promega, or MBI Fermentas. The cDNA of GIRK2 (33) was a gift from H. A. Lester. cDNA of c␤ARK (34) was a gift from E. Reuveni. The coding sequences of m2R (35) were a gift from E. Peralta. G␣ z (24) and G␣ i3 (36) (gifts from M. I. Simon) were subcloned into the pGEMHJ vector (34) using a standard polymerase chain reaction procedure in which an EcoRI restriction site was created immediately before the ATG initiation codon, and another EcoRI (m2R), HindIII (G␣ z ), or BstEII (G␣ i3 ) site was created immediately after the stop codon. The pGEMHJ vector provides 5Ј-and 3Јuntranslated regions of the Xenopus ␤-globin RNA (37), ensuring a high level of protein expression in the oocytes. The m2R⅐G␣ z and m2R⅐G␣ i3 tandem cDNAs were created by ligating the m2R cDNA sequence into the EcoRI site of the corresponding G␣ cDNA. Thus, in each tandem protein, the full primary sequences of m2R and the G␣ are connected by a short, 2-amino acid linker (Glu-Phe) encoded by the EcoRI restriction site sequence GAATTC. High quality capped RNA was prepared as described (38).
Oocytes and Electrophysiology-Xenopus laevis oocytes were prepared and injected with RNAs as described (39) and incubated for 3-5 days before the experiment in the ND96 solution (96 mM NaCl, 2 mM KCl, 1 mM CaCl 2 , 1 mM MgCl 2 , 5 mM Hepes/NaOH) supplemented with 50 g/ml gentamycin and 2.5 mM sodium pyruvate. c␤ARK RNA (5 ng/oocyte) was injected 2 days after the other RNAs, 2 days before the experiment. Two-electrode voltage clamp experiments were performed as described (39). Data acquisition and analysis were done using pCLAMP software (Axon Instruments). The membrane potential was set at Ϫ80 or Ϫ40 mV in the ND96 solution, and GIRK currents were induced and measured in the high K ϩ solution (KCl, 96; NaCl, 2; CaCl 2 , 1; MgCl 2 , 1; Hepes/NaOH, 5). Exchange of solutions was performed using the BSP-4 fast perfusion system (ALA Instruments, New York). Full exchange of solution in the experimental bath (50 l volume) occurred within less than 0.5 s. The shift from ND96 to the high K ϩ solution evokes a basal current, I basal (see Fig. 1A), which flows mainly via the GIRK channels, with a minor (Ͻ0.1 A) contribution of a G protein-independent endogenous current, I native (39,40). Therefore, I basal was corrected by subtracting the average I native measured in native (not injected with RNA) oocytes. PTX treatment was done by injecting the oocytes with the activated A protomer of PTX (Alomone Labs, Jerusalem, Israel) 4 -24 h before the experiment (39).
Immunochemistry of the Expressed m2R and m2R⅐G␣ Tandems in Xenopus Oocytes-was performed essentially as described (41). Oocytes were injected with mRNAs and incubated in NDE solution containing 0.5 mCi/ml [ 35 S]methionine/cysteine (Amersham Pharmacia Biotech) for 3-4 days at 22°C. 5 oocytes were homogenized, and proteins were solubilized, immunoprecipitated with 5 l of a polyclonal antibody against m2R (Alomone Labs, Jerusalem, Israel), and electrophoresed on 7.5% polyacrylamide-SDS gel.

RESULTS AND DISCUSSION
Collision Coupling of m2 Receptor to GIRK via the Endogenous PTX-sensitive G Proteins-We used the GIRK1/GIRK2 subunit composition of the G protein-gated channels, which is abundant in the brain (see Refs. 28 and 30 for review). Expression of m2R with GIRK1 and GIRK2 in Xenopus oocytes gave rise to inwardly rectifying K ϩ currents (42, 43) that were measured using the two-electrode voltage clamp method (Fig. 1A). Exchanging the high Na ϩ solution (ND96) to a high K ϩ solution elicited a basal current (I basal ), and the addition of 10 M agonist (acetylcholine (ACh)) caused an additional current (I ACh ), which gradually inactivated over a few minutes (39,44,45). 10 M of ACh was a saturating concentration, because a maximal response was observed already at 1 M (data not shown). In each experiment, all oocytes were injected with the same amount of channel RNA. We used relatively low doses of RNA of the channel subunits (50, 80, or 100 pg/oocyte each) to ensure that the amount of the endogenous G ␤␥ subunits was sufficient to activate all the channels. Indeed, larger currents were evoked by ACh at 200 -500 pg/oocyte of GIRK1 and GIRK2 RNAs (data not shown). Increasing the dose of m2R RNA caused an increase in I ACh ; the dose-response relationship almost saturated at 500 pg/ oocyte (Fig. 1B, E). I basal was unchanged by coexpression of m2R-G␣ z Tandem Activates GIRK m2R at RNA doses between 10 and 100 pg/oocyte (40) but increased at 500 pg/oocyte (Fig. 1B, q), suggesting that at high m2R densities, the agonist-free receptor activates the G proteins. The speed of activation of I ACh (expressed as the time by which I ACh reaches 50% of peak amplitude, t 50% act ) became progressively faster with increasing expression of m2R, reaching 1.4 s at 500 pg of m2R RNA/oocyte (Fig. 1C). The simplest explanation for this phenomenon is a collision coupling-type mechanism (11,46), in which one receptor shuttles between and catalytically activates several G proteins in succession. The active state of a G␣ persists after unbinding of the receptor until GTP is hydrolyzed; only then it can bind G ␤␥ . Until then, G ␤␥ remains free ("active") too. This model predicts that the response will develop faster when more receptors are present, because each receptor has to activate less molecules of G␣, saving the shuttling and binding/unbinding time. There is an alternative theoretical possibility that each m2R is always found in a tight complex with a G i/o and each complex releases one G ␤␥ . In such a case, to explain the acceleration of activation at increased levels of m2R, one must assume that each G ␤␥ activates several channels in succession, and a channel previously activated by G ␤␥ remains open at least until and during the activation of the next one. This, however, is highly unlikely in view of the fast channel deactivation upon unbinding of G ␤␥ (31,32).
GIRK Channels Are Activated by G ␤␥ Released from G␣ z ␤␥ Heterotrimers-Expression of increasing amounts of G␣ z while keeping the levels of m2R and GIRK constant slowed down the speed of activation of I ACh (Fig. 1D). The activation appeared composed of a fast and a slow component, with the slow one becoming more prominent as the concentration of G␣ z RNA was increased. At the same time, the inactivation of I ACh became less apparent. The deactivation of the current upon washout of ACh became extremely slow, taking many minutes for completion (data not shown), in line with the slow GTPase activity of G␣ z . These features are similar to those described by Jeong and Ikeda (26) for G z -mediated activation of GIRK in sympathetic neurons, suggesting that the slow component is contributed by G ␤␥ released from G␣ z ␤␥ heterotrimers. Indeed, PTX treatment eliminated the fast component of activation, leaving the slow one intact (Figs. 1E and 2E). On the other hand, the ACh-evoked response was still mediated by G ␤␥ , because coexpression of a myristoylated C-terminal part of ␤-adrenergic receptor kinase (c␤ARK), which is a strong G ␤␥ chelator and blocks agonist-evoked GIRK currents in the oocytes (47), reduced I ACh by 85 Ϯ 6% (mean Ϯ S.E., n ϭ 3) in oocytes injected with 1 ng of G␣ z RNA. Fig. 2 summarizes the results of G␣ z coexpression experiments. Without PTX, the amplitudes of either I basal or I ACh (Fig. 2, A and B) were not changed by coexpression of G␣ z , supporting the notion that the overall amplitude of the response was limited by the number of expressed channels. Without G␣ z , PTX strongly reduced I basal and almost fully suppressed the I ACh (39), and both I basal and I ACh were dosedependently rescued by coexpression of G␣ z (Fig. 2, A and B). At 1 ng of G␣ z RNA/oocyte, I basal and I ACh were already almost fully restored. The speed of activation of I ACh became progressively slower with increasing dose of G␣ z , with t 50% act reaching 35 Ϯ 5 s at 1 ng of G␣ z RNA in the presence of PTX (Fig. 2C). At all concentrations of G␣ z , t 50% act was lower in cells untreated with PTX, although this difference was reduced from 2.9-fold at 10 pg to 1.7-fold at 1 ng of G␣ z RNA (Fig. 2C). The fast component of activation could be reasonably well fitted by a single exponent with time constant of activation () of 1-3 s in a majority of cells (Fig. 2D). This allowed to estimate its contribution to the peak amplitude of I ACh (Fig. 2E). This analysis showed that in PTX the fast component of activation was practically absent. As expected, in the absence of PTX the contribution of the fast component decreased as the expression of G␣ z increased; about 30% of total agonist-evoked current was still contributed by the endogenous G␣ i/o even at 1 ng of G␣ z RNA.
The results of Fig. 2 (A-E) demonstrate that the endogenous G␣ i/o and the expressed G␣ z compete for the available m2R, to donate G ␤␥ for channel activation. In the absence of PTX, a fast component of activation is contributed by the endogenous G␣ i/o at all doses of G␣ z RNA, even 1 ng/oocyte, and only after PTX treatment all channels can be activated solely via G ␤␥ coming from G z . PTX reveals the actual kinetics of the response evoked via G z .
These experiments also have an important general implication: PTX is often used to suppress agonist-evoked responses mediated by G i/o proteins, to evaluate in this way what part of the cellular response evoked by this agonist is mediated by PTX-insensitive G proteins. Our results call for caution in such interpretations by showing that after PTX treatment, PTXinsensitive G proteins may "take over" the function of PTXsensitive G i/o proteins and couple to a larger proportion of effector molecules than under normal conditions.
The smaller amplitude of I ACh at low levels of G␣ z expression suggests that G z protein expression is limiting; each G z may donate G ␤␥ for the activation of a limited number of channels (maybe only one). The progressive increase in t 50% act with increasing G␣ z expression in the presence of PTX (Fig. 2C) is compatible with the collision coupling mechanism: when more G␣ z are available, each receptor can activate more G proteins (hence a greater I ACh ; compare with Fig. 2B); but activation of more G proteins by each receptor simply takes more time. An alternative explanation is the presence of a diffusional barrier between G z and GIRKs (because of channel scaffolding with other G proteins such as G i , etc.), when some of the channels are poorly accessible to G ␤␥ derived from G z . Expression of m2R-G␣ z Tandem Activates GIRK more G z might eventually force the activation of the previously inaccessible population of channels, but with a slower time course, reflecting the slower diffusion of G ␤␥ to these channels. We view the collision coupling mechanism between m2R and G z as a more plausible explanation, because it is supported by an independent experiment showing acceleration of activation of I ACh by increasing the level of expression of m2R, in the presence of PTX (Fig. 2F). Importantly, the activation of GIRK by G z is much slower than by PTX-sensitive G i/o . Comparison of the group in which all the channels were activated by the endogenous G i/o (0 pg of G␣ z , no PTX) and the group in which the same amount of channels was activated by G z (1 ng of G␣ z RNA, with PTX) reveals a 22-fold difference in t 50% act (Fig. 2C). To understand the reasons for this poor speed of activation, we have created an m2R⅐G␣ z tandem and studied its interaction with GIRK. For comparison, an m2R⅐G␣ i3 tandem was also made and tested.
Stoichiometric Coupling of the m2R⅐G␣ z Tandem to GIRK-Metabolically labeled m2R, m2R⅐G␣ z , and m2R⅐G␣ i3 were immunoprecipitated from oocytes injected with 100 pg of each RNA with a polyclonal antibody against m2R. The amounts of all three proteins were approximately equal (Fig. 3A). m2R (calculated molecular mass, 51.7 kDa) is a glycosylated protein (48) with an apparent molecular mass of 65-80 kDa in mammalian cells (48,49). In the oocytes, it ran as a diffuse band of ϳ90 kDa (Fig. 3A), suggesting that it was glycosylated stronger than in mammalian cells. The apparent sizes of the tandem proteins were 116 -120 kDa, above the calculated ϳ92 kDa, suggesting that these proteins were also glycosylated. Uninjected oocytes were devoid of label.
In oocytes coexpressing the m2R⅐G␣ z tandem with GIRK, I basal was stable in the range 50 -300 pg of m2R⅐G␣ z RNA but grew at higher doses (Fig. 3B), suggesting a substantial basal activity of the receptor at high expression levels. I ACh grew dose-dependently with increasing amounts of tandem RNA (Fig. 3C). PTX reduced both I basal and I ACh at lower doses of m2R⅐G␣ z , confirming that a GPCR, even when tethered to a G␣, can still activate other nontethered G proteins (15). Above 300 pg/oocyte of tandem RNA, PTX had little effect on either I basal or I ACh (Fig. 3, B and C), suggesting that there were enough tandems to donate G ␤␥ for activation of most channels. The m2R⅐G␣ z -mediated I ACh was fully G ␤␥ -dependent, because coexpression of c␤ARK completely abolished it (data not shown). Examination of the rising phase of I ACh (Fig. 3D) revealed that activation by m2R⅐G␣ z was accelerated by the PTX treatment (Fig. 3D, trace 2, and Fig. 3F). This was true at most doses of m2R⅐G␣ z (Fig. 3E). These results reveal that the coupling of the tethered receptor to G␣ z within the tandem is better (faster) than its coupling to the nontethered G␣ i/o . Again, PTX treatment reveals the real kinetics of GIRK activation by the tandem. Fig. 3F summarizes the differences in the kinetics of GIRK activation, obtained in several experiments at the same doses of RNAs of m2R, G␣ z , or m2R⅐G␣: 100 pg/oocyte. The activation of GIRK by the m2R⅐G␣ z (in PTX) is 4-fold faster than by separately expressed m2R and G␣ z , suggesting that the slowness of coupling of m2R to GIRK via G z is primarily the result of poor receptor-G␣ z coupling. Yet, activation by the m2R⅐G␣ z tandem is still slower (p Ͻ 0.001) than by m2R via the endogenous G i/o (t 50% act , 3.6 Ϯ 0.4 s, n ϭ 16, versus 1.9 Ϯ 0.1 s, n ϭ 53). This may reflect the slow intrinsic kinetics of G␣ z activation, because a structurally similar tandem, m2R⅐G␣ i3 , activated the channel even faster than m2R via the endogenous G i/o (t 50% act , 0.8 Ϯ 0.2 s, n ϭ 7).
The most striking consequence of the physical link between m2R and G␣ z was revealed after PTX treatment, which eliminated of the contribution of the endogenous G i/o (Fig. 3E): the speed of activation of the m2R⅐G␣ z -mediated response did not depend on the expression level of the tandem. This result clearly demonstrates a stoichiometric interaction between m2R and G␣ z in the tandem protein, confirming that tethering produces an active GPCR-G␣ complex and eliminates the collision coupling. Furthermore, it confirms that collision coupling (if any) between free G ␤␥ and GIRK does not contribute to the activation kinetics. This, in turn, strengthens the conclusion reached above, that the collision coupling under normal conditions (free expressed m2R; Figs. 1C and 2F) occurs on the receptor-G protein level. The stoichiometric (noncatalytic) coupling between the released G ␤␥ and the channel may simply reflect the fast dissociation of G ␤␥ (see above). A more tantalizing possibility is that there is a high affinity preformed complex between a G protein heterotrimer and the GIRK (50,51), so that dissociation between G␣ and G ␤␥ is immediately followed by the activation of the effector, saving the time lag introduced by the diffusion. Such a scheme explains well the fast kinetics of GIRK activation in atrial myocytes and in neurons (32,52).
Conclusions-Using the Xenopus oocyte expression system, we have performed a quantitative comparison of activation of an effector (GIRK) by a GPCR (m2R) via endogenous PTXsensitive G proteins, via a coexpressed G␣ z , and via an m2R⅐G␣ z tandem (fusion) protein. The efficiency (in terms of kinetics) of coupling of the components of the signaling pathway was monitored by measuring the activation kinetics of the GIRK current. Our results demonstrate a collision coupling between m2R and the G i/o proteins (either endogenous G i/o or FIG. 3. Expression of the m2R⅐G␣ z tandem and its coupling to GIRK. A, a polyclonal antibody against m2R immunoprecipitates metabolically labeled m2R, m2R⅐G␣ i3 , and m2R⅐G␣ z tandems from oocytes injected with 100 pg of each RNA but not from uninjected oocytes. B and C, effect of the level of expression of the m2R⅐G␣ z tandem and of PTX on I basal (B) and I ACh (C). Results from a representative batch of oocyte are shown; similar results were obtained in two more batches. The bars show mean Ϯ S.E. of 3-7 oocytes. Asterisks indicate p Ͻ 0.05 or better (t test). D, representative traces of the responses to ACh in oocytes expressing the GIRK channels and either m2R (trace 1, 100 pg/oocyte) or m2R⅐G␣ z tandem (100 pg/oocyte) with (trace 3) or without (trace 2) PTX treatment. E, dependence of t 50% act of I ACh on the level of expression of m2R⅐G␣ z . Same experiment as in B and C. Similar results were obtained in two additional batches. F, comparison of t 50% act from all experiments in which 100 pg RNA/oocyte of each of the indicated constructs were injected. Each bar shows mean Ϯ S.E. of 15-53 oocytes from 5-12 batches, except for the m2R⅐G␣ i3 tandem (7 oocytes, two batches). Horizontal bars with asterisks indicate statistically significant differences (p Ͻ 0.05 or better) obtained in comparisons between selected groups (two-tailed t test). coexpressed G z ). In contrast, the relationship between the amount of activated G proteins and the activated effector is stoichiometric. The latter is in line with the existence of a preformed complex between a G protein heterotrimer and the effector. Although m2R is able to elicit a full effector response via G z , the slow kinetics of the response reveals an inherent inefficiency of this signaling pathway, compared with m2R-GIRK signaling via PTX-sensitive G i/o proteins. The use of m2R⅐G␣ reveals that the poor coupling between m2R and G␣ z and the intrinsic slow activation of G z are important ratelimiting steps in the G z pathway.