Coexpression of ligand-gated P2X and G protein-coupled P2Y receptors in smooth muscle. Preferential activation of P2Y receptors coupled to phospholipase C (PLC)-beta1 via Galphaq/11 and to PLC-beta3 via Gbetagammai3.

P2 receptor subtypes and their signaling mechanisms were characterized in dispersed smooth muscle cells. UTP and ATP stimulated inositol 1,4,5-triphosphate formation, Ca2+ release, and contraction that were abolished by U-73122 and guanosine 5'-O-(3-thio)diphosphate, and partly inhibited (50-60%) by pertussis toxin (PTX). ATP analogs (adenosine 5'-(alpha, beta-methylene)triphosphate, adenosine 5'-(beta, gamma-methylene)triphosphate, and 2-methylthio-ATP) stimulated Ca2+ influx and contraction that were abolished by nifedipine and in Ca2+-free medium. Micromolar concentrations of ATP stimulated both Ca2+ influx and Ca2+ release. ATP and UTP activated Gq/11 and Gi3 in gastric and aortic smooth muscle and heart membranes, Gq/11 and Gi1 and/or Gi2 in liver membranes, and Go and Gi1-3 in brain membranes. Phosphoinositide hydrolysis stimulated by ATP and UTP was mediated concurrently by Galphaq/11-dependent activation of phospholipase (PL) C-beta1 and Gbetagammai3-dependent activation of PLC-beta3. Phosphoinositide hydrolysis was partially inhibited by PTX or by antibodies to Galphaq/11, Gbeta, PLC-beta1, or PLC-beta3, and completely inhibited by the following combinations (PLC-beta1 and PLC-beta3 antibodies; Galphaq/11 and Gbeta antibodies; PLC-beta1 and Gbeta antibodies; PTX with either PLC-beta1 or Galphaq/11 antibody). The pattern of responses implied that P2Y2 receptors in visceral, and probably vascular, smooth muscle are coupled to PLC-beta1 via Galphaq/11 and to PLC-beta3 via Gbetagammai3. These receptors co-exist with ligand-gated P2X1 receptors activated by ATP analogs and high levels of ATP.

P 2 receptors have been classified recently into two classes comprising ligand-gated cationic channels or P 2X receptors and G protein-coupled P 2Y receptors (1, 2); P 2U and P 2T receptors have been subsumed into the P 2Y class of receptors. The term P 2 recognizes the fact that purine and pyrimidine nucleotides can act as preferential ligands of various receptor subtypes (2). Up to seven P 2X receptor subtypes (3)(4)(5)(6)(7)(8)(9) and eight P 2Y receptor subtypes (10 -16) have been cloned from mammalian and avian species. Fuller understanding of the functions subserved by discrete receptor subtypes is hampered by the organization of native P 2X receptors into homopolymers or heteropolymers (5) and by the co-existence of P 2X and P 2Y receptors on the same cell (17). Earlier classifications based on agonist potency profiles had been confounded by the paucity of selective antagonists and radioligands (2), and by the rapid degradation of some nucleotides, mainly ATP and 2-methylthio-ATP, by ecto-nucleotidases (18), and the interconversion of adenine and uridine nucleotides by ecto-nucleoside diphosphokinases (19,20). P 2X1 is the main P 2X receptor subtype expressed in visceral and vascular smooth muscle (21), whereas P 2X2 and P 2X3 are the main receptor subtypes expressed in peripheral sensory ganglia (8,(21)(22)(23). Both P 2X1 and P 2X3 receptors have high affinity for ATP and AMP-PCP 1 and are rapidly desensitized (23,24). P 2X2 , P 2X4 , and P 2X6 receptors are the predominant receptor subtypes expressed in the adult brain where they are present in various heteromeric combinations; these receptor subtypes exhibit lower affinity for ATP, are insensitive to AMP-PCP, and are not readily desensitized (7,8,22,23). Their insensitivity to AMP-PCP restricts the usefulness of this analog as a radioligand for all but the P 2X1 and P 2X3 receptor subtypes (24). P 2Y receptors exhibit variable affinity for purine and pyrimidine nucleotides. P 2Y1 are purinoceptors and are adenine nucleotide-specific (10,13), whereas P 2Y2 receptors (P 2U in earlier classifications) have equal affinity for adenine and uridine nucleotide triphosphates (UTP Ն ATP) (11,19). P 2Y3 , P 2Y4 , and P 2Y6 are pyrimidinoceptors: P 2Y4 is UTP selective whereas P 2Y3 and P 2Y6 are UDP selective (14,19). The functional status of P 2Y5 which has low homology to other P 2Y receptors has not been resolved (16,25), while the P 2Y7 receptor has now been identified as the leukotriene B 4 receptor (26). P 2Y receptors are variously coupled to pertussis toxin-sensitive and -insensitive G proteins which activate or inhibit various effector enzymes including phospholipase C-␤ (PLC-␤) (15,16,(27)(28)(29)(30), phospholipase D (31,32), phospholipase A 2 (33), and adenylyl cyclase (28,30,34).
ATP, UTP, and AMP-PCP can mobilize Ca 2ϩ and elicit contractile responses in vascular and visceral smooth muscle suggesting that both P 2X and P 2Y receptors are present (15,17,(35)(36)(37). Their co-existence raises the question as to which receptor subtype mediates preferentially the action of the endogenous ligand, ATP. In the present study, we have used a series of purine and pyrimdine agonists to characterize P 2 receptors in dispersed gastric smooth muscle cells and identify the sig-naling pathways to which they are coupled. Comparative studies characterized the coupling of P 2Y receptors to G proteins in vascular smooth muscle, heart, liver, and brain. P 2X1 and P 2Y2 receptors were shown to co-exist on gastric smooth muscle cells and to mediate Ca 2ϩ mobilization and muscle contraction via three distinct pathways. UTP and nanomolar concentrations of ATP activated exclusively P 2Y2 receptors, whereas micromolar concentrations of ATP activated additionally P 2X1 receptors. The pattern suggests that contraction induced by purine and pyrimidine nucleotides may be preferentially mediated by G protein-coupled receptors.

Dispersion of Gastric Smooth Muscle Cells-Smooth muscle cells
were isolated from the circular muscle layer of rabbit stomach by sequential enzymatic digestion, filtration, and centrifugation as described previously (38 -40). The cells were resuspended in enzyme-free medium consisting of 120 mM NaCl, 4 mM KCl, 2.6 mM KH 2 PO 4 , 2 mM CaCl 2 , 0.6 mM MgCl 2 , 25 mM HEPES, 14 mM glucose, and 2.1% Eagle's essential amino acid mixture. The muscle cells were harvested by filtration through 500-m Nitex mesh and centrifuged twice at 350 ϫ g for 10 min.
In some experiments, the muscle cells were reversibly permeabilized using the Trans.Port reagent (Life Technologies, Inc.) as described previously (40). The cells were washed in Ca 2ϩ -and Mg 2ϩ -free HEPES medium and re-suspended in a medium containing 10 mM NaCl, 140 mM KCl, 2.4 mM MgCl 2 , and 10 mM HEPES. Trans.Port reagent (15 l/ml) was added with or without GDP␤S (10 M) and the mixture incubated at 31°C for 20 min. Permeabilization was terminated by addition of Stop solution (30 l/ml) and the cell suspension centrifuged for 15 min at 350 ϫ g. The cells were resuspended in control HEPES medium containing 0.1% bovine serum albumin and incubated at 31°C for 1 h. The resealed cells were shown to exclude trypan blue and respond to contractile agonists and depolarizing concentrations of KCl (20 mM) but not to 2 mM CaCl 2 or inositol 1,4,5-trisphosphate (1 M) (40). The effectiveness of GDP␤S was tested by measuring its ability to abolish the contractile response to the contractile agonist, cholecystokinin octapeptide (40).

Identification of G Protein Subtypes and PLC-␤ Isozymes in Gastric Smooth
Muscle by Western Blot-The expression of G proteins and PLC-␤ isozymes was determined by Western blot analysis as described previously (41)(42)(43). Homogenates prepared from dispersed gastric muscle cells were solubilized on ice for 1 h in 20 mM Tris (pH 8.0), 1 mM dithiothreitol, 100 mM NaCl, and 0.5% sodium cholate. The suspension was centrifuged at 13,000 ϫ g for 5 min. Solubilized proteins were resolved by SDS-polyacrylamide gel electrophoresis and electrophoretically transferred to nitrocellulose membranes. The blots were incubated for 12 h at 4°C with subtype-specific G protein or PLC-␤ antibodies, and then for 1 h with secondary antibody conjugated with horseradish peroxidase. The bands were identified by enhanced chemiluminescence.
Selective Protection of P 2 Receptors-A technique of selective receptor protection previously used to determine the co-existence and function of various G protein-coupled receptors (44 -48) was used to determine the presence and function of P 2 receptor subtypes. The technique involves protection of one receptor subtype with selective agonists or antagonists followed by inactivation of all unprotected receptors with a low concentration of N-ethylmaleimide (5 M). Freshly dispersed muscle cells were incubated with one agonist (AMP-PCP, AMP-CPP, UTP, or ATP) at 31°C for 2 min followed by addition of 5 M N-ethylmaleimide for 20 min. The cells were centrifuged twice at 150 ϫ g for 10 min and resuspended in control HEPES medium for 60 min to ensure complete re-sensitization. The contractile response of cells treated in this fashion was compared with the response of untreated cells. As previously shown (44 -48), muscle cells incubated with N-ethylmaleimide without protective agent did not contract in response to receptor-linked agonists, but they responded fully upon addition of agents that bypass receptors (e.g. ionomycin, KCl, and forskolin), implying that post-receptor mechanisms were intact.
Measurement of Contraction in Dispersed Muscle Cells-Contraction of dispersed muscle cells was measured by scanning micrometry as described previously (38 -40). The length of 50 muscle cells treated with one concentration of a contractile agent was measured by scanning micrometry and compared with the length of 50 untreated muscle cells. All measurements were done in the presence of adenosine A 1 and A 2 antagonists (1 M DPCPX and 0.1 M CGS-15943, respectively) (47).
Time course measurements were done at intervals ranging from 5 s to 5 min. As with other agonists, peak contraction was measured at 30 s and the response used to construct concentration-response curves. Contraction was expressed as the mean decrease in cell length from control in micrometers or as the percent decrease in cell length (range of control cell length in various experiments 96 Ϯ 4 to 103 Ϯ 5 m).
Measurement of Cytosolic Free Ca 2ϩ in Dispersed Muscle Cells-Cytosolic free Ca 2ϩ ([Ca 2ϩ ] i ) was measured by fluorescence in suspensions of muscle cells loaded with the fluorescent Ca 2ϩ dye, fura 2, as described previously (40,45). Autofluorescence of unloaded cells was determined in each suspension and subtracted from the fluorescence of fura 2-loaded cells. Measurements were done in the presence of adenosine A 1 and A 2 antagonists. Ca 2ϩ levels were calculated under basal conditions and upon addition of agonist from the ratios of observed, minimal and maximal fluorescence (49).
Inositol 1,4,5-Trisphosphate (IP 3 ) Radioreceptor Assay-IP 3 was measured in dispersed muscle cells by a radioreceptor assay which utilizes 3 H-labeled D-myo-IP 3 and bovine brain microsomes as described previously (41,42). Agonists were added for 30 s in the presence of adenosine A 1 and A 2 antagonists to 1 ml of muscle cell suspension (10 6 cells/ml) and the reaction terminated with an equal volume of ice-cold 10% perchloric acid. The supernatant was extracted and IP 3 content in the aqueous phase was measured. The results were expressed as picomoles of IP 3 /10 6 cells.
Assay of PLC-␤ Activity in Plasma Membranes-PLC-␤ activity was determined in plasma membranes by a modification of the method of Uhing et al. (50) as described previously (43,51). The membranes were isolated from dispersed muscle cells labeled with myo-[ 3 H]inositol. PLC-␤ assay was initiated by addition of 0.4 mg of membrane protein to 25 mM Tris-HCl (pH 7.5), 0.5 mM EGTA, 10 mM MgCl 2 , 300 nM free Ca 2ϩ , 1 M GTP␥S, 5 mM phosphocreatine, and 50 units/ml creatine phosphokinase in a total volume of 0.4 ml. After incubation at 31°C for 60 s, the reaction was terminated with 0.6 ml of 25% trichloroacetic acid (w/v). The supernatant was extracted four times with 2 ml of diethyl ether and the amount of labeled inositol phosphates in the aqueous phase was counted. All measurements were done in the presence of adenosine A 1 and A 2 antagonists. PLC-␤ activity was expressed as counts/min/mg protein/min.
Identification of Receptor-activated G Proteins-G proteins selectively activated by P 2 receptor agonists in muscle cell membranes were identified by the method of Okomoto et al. (52) as described previously (41,42,48). Muscle cell homogenates were centrifuged at 27,000 ϫ g for 15 min, and the crude membranes solubilized for 60 min at 4°C in 20 mM HEPES medium (pH 7.4) containing 2 mM EDTA, 240 mM NaCl, and 1% CHAPS. The membranes were diluted 10-fold and incubated at 37°C with 60 nM [ 35 S]GTP␥S in a medium containing 10 mM HEPES (pH 7.4), 100 M EDTA, and 10 mM MgCl 2 . The reaction was stopped with 10 volumes of 100 mM Tris-HCl medium (pH 8.0) containing 10 mM MgCl 2 , 100 mM NaCl, and 20 M GTP, and the mixture incubated for 2 h on ice in wells precoated with specific G protein antibodies. The wells were washed three times with phosphate buffer solution containing 0.05% Tween 20 and the radioactivity from each well was counted. Coating with G protein antibodies (1:1000) was done after the wells were first coated with anti-rabbit IgG (1:1000) for 2 h on ice. The measurements were done in the presence of adenosine A 1 and A 2 antagonists.
In separate experiments on rabbit aortic smooth muscle, heart, liver, and whole brain, membranes were obtained by homogenization of these tissues without prior cell isolation. The homogenates were treated as described above for gastric smooth muscle cells.
Data Analysis-Results were expressed as means Ϯ S.E. of n separate experiments and evaluated statistically using Student's t test for paired or unpaired data. Concentration-response curves were analyzed using the P.fit 6.0 program.

Contraction and Ca 2ϩ Mobilization in Dispersed Smooth
Muscle Cells by Purine and Pyrimidine Nucleotides-Exposure of muscle cells to 1 M UTP or ATP caused immediate contraction that was virtually linear during the first 10 s and attained a peak in 30 s followed by a decline to lower levels (Fig. 1A). The biphasic time course was identical to that observed with other contractile agonists (38,53). The peak response at 30 s was used to construct concentration-response curves. Prolonged exposure of muscle cells to purine or pyrimidine agonists resulted in time-dependent desensitization that was more rapid with P 2X receptor agonists (e.g. AMP-PCP) than with P 2Y receptor agonists (e.g., UTP) (Fig. 1B). With either type of agonist, however, there was minimal desensitization (Ͻ2% of control response) during the initial 30-s period when peak response was measured.
ATP, UTP, AMP-PCP, and AMP-CPP increased cytosolic Ca 2ϩ ([Ca 2ϩ ] i ) in dispersed smooth muscle cells by 1-fold at 10 nM and by 3-fold at 10 M (Table II). The increase induced by 2-methylthio-ATP was also concentration-dependent but significantly lower (Table II).
Contraction and the increase in [Ca 2ϩ ] i induced by AMP-PCP, AMP-CPP, and 2-methylthio-ATP were abolished by nifedipine (1 M) and in Ca 2ϩ -free medium but were not affected by pretreatment of the cells for 1 h with 400 ng/ml PTX, or for 10 min with the PLC-␤ inhibitor, U-73122 (1 M); insertion of GDP␤S into transiently permeabilized muscle cells had no effect (Tables I and II). The pattern of response suggested that contraction and Ca 2ϩ mobilization induced by ATP analogs with high affinity for P 2X receptors was mediated by Ca 2ϩ influx via dihydropyridine-sensitive Ca 2ϩ channels.
In contrast, contraction and the increase in [Ca 2ϩ ] i induced by 10 nM or 10 M UTP, and by 10 nM ATP were not affected by nifedipine or Ca 2ϩ -free medium but were abolished by GDP␤S or U-73122 (Tables I and II). Pertussis toxin partly inhibited contraction (46 Ϯ 5 to 50 Ϯ 7%) and the increase in [Ca 2ϩ ] i (49 Ϯ 6 to 68 Ϯ 7%) induced by 10 nM and 10 M UTP and by 10 nM ATP. The pattern suggested that contraction and Ca 2ϩ mobilization induced by UTP, which has high affinity for P 2Y2 receptors, and by low concentrations of ATP were mediated by IP 3 -dependent Ca 2ϩ release resulting from activation of PLC-␤ via both PTX-sensitive and -insensitive G proteins.
Contraction induced by 10 M ATP was not affected by GDP␤S, PTX, U-73122, nifedipine, and Ca 2ϩ -free medium (Table I), while the increase in [Ca 2ϩ ] i was only slightly inhibited (13 Ϯ 5 to 28 Ϯ 6%) (Table II). However, a combination of U-73122 or GDP␤S with either nifedipine or Ca 2ϩ -free medium abolished the contraction and the increase in [Ca 2ϩ ] i (Table I). Thus, contraction and the increase in [Ca 2ϩ ] i induced by high concentrations of ATP could be independently mediated by Ca 2ϩ influx and Ca 2ϩ release and appears to reflect activation of both P 2X and P 2Y receptors.
The extent of ATP-induced contraction mediated by P 2X receptors was evaluated at different concentrations of ATP in the presence of 1 M U-73122. The concentration-response curve was shifted to the right by U-73122 (Fig. 2), and the EC 50 for ATP acting via P 2X receptors was 0.61 Ϯ 0.07 M (compared with 33 Ϯ 9 nM in the absence of U-73122). In Ca 2ϩ -free medium, the contractile response to all concentrations of ATP was abolished by U-73122 (Fig. 2).
The effect of a combination of PLC-␤1 antibody and G ␤ antibody was additive, eliciting complete inhibition of PLC-␤ ac-

FIG. 4. Expression of G proteins and PLC-␤ isozymes in gastric smooth muscle.
Homogenates were prepared from dispersed gastric circular muscle cells and solubilized with sodium cholate in Tris buffer. Proteins were resolved by SDS-polyacrylamide gel electrophoresis, electrophoretically transferred to nitrocellulose membranes, and probed with specific antibodies to the PLC-␤ isozymes and ␣-subunits of various G proteins and then with anti-rabbit IgG conjugated to horseradish peroxidase. The proteins were identified by enhanced chemiluminescence. tivity (91 Ϯ 4%), whereas the effect of a combination of PLC-␤3 antibody and G ␤ antibody was not additive (62 Ϯ 6% inhibition by the combination versus 59 Ϯ 8 and 56 Ϯ 6% inhibition for either antibody alone). Identical results were obtained for UTPstimulated PLC-␤ activity which was inhibited 90 Ϯ 6% by the combination of PLC-␤1 and G ␤ antibodies and 53 Ϯ 7% by a combination of PLC-␤3 and G ␤ antibodies (Fig. 5).
The results implied that phosphoinositide (PI) hydrolysis induced by ATP and UTP was mediated by G␣ q/11 -dependent activation of PLC-␤1, and by PTX-sensitive, G ␤␥ -dependent activation of PLC-␤3. The pattern is consistent with PTXsensitive and -insensitive stimulation of IP 3 formation in dispersed muscle cells by ATP and UTP (Fig. 3).

Identification of G Proteins Coupled to P 2Y Receptors-The
PTX-sensitive and -insensitive G protein(s) activated by ATP and UTP in gastric smooth muscle were identified by a technique that did not involve functional blockade with antibodies. Solubilized muscle cell membranes were incubated with [ 35 S]GTP␥S (60 nM) with or without ATP or UTP and added to wells precoated with different G␣ antibodies; an increase in the binding of [ 35 S]GTP␥S⅐G␣ complexes to a specific G␣ antibody reflected activation of the corresponding G protein. Addition of ATP (10 M) caused a time-dependent increase in the binding of [ 35 S]GTP␥S to G␣ q/11 and G␣ i3 antibodies (Fig. 7), but not to G␣ s G␣ i1-2 or G␣ o antibodies (Table III). An identical pattern was observed with UTP which stimulated the binding of [ 35 S]GTP␥S to G␣ q/11 and G␣ i3 antibodies but not to G␣ s , G␣ i1-2 , or G␣ o antibodies (Table III). Pretreatment of muscle cells with 400 ng/ml PTX for 1 h before membrane isolation abolished the ATP-and UTP-stimulated increase in steadystate binding of [ 35 S]GTP␥S to G␣ i3 antibody, but not to G␣ q/11 antibody (Table III). Peptide I (KNNKECGLY), comprising the G protein sequence against which the G␣ i3 antibody was raised, inhibited selectively ATP-and UTP-stimulated activation of G␣ i3 (Table Aliquots were added to wells precoated with G␣ i3 or G␣ q/11 antibody for 2 h and bound radioactivity measured (counts/min/mg of protein). ATP caused a significant increase in binding of [ 35 S]GTP␥S⅐G␣ complexes to wells precoated with G␣ i3 antibody (A) or G␣ q/11 antibody (B) but not to wells precoated with G␣ i1-2 , G␣ s , or G␣ o antibody (see Table III). Values are mean Ϯ S.E. of four experiments. III). Conversely, peptide II (QLNLKEYNLV), comprising the G protein sequence against which G␣ q/11 antibody was raised, inhibited selectively ATP-and UTP-stimulated activation of G␣ q/11 (Table III). Peptides I and II inhibited both activation of G proteins by GTP␥S as well as the increase in activation induced by ATP or UTP. Peptides I and II were used at a concentration (1 M) previously shown to abolish activation of G␣ i3 and G␣ q/11 , respectively (48). AMP-PCP, AMP-CPP, and 2-methylthio-ATP did not cause activation of G␣ q/11 , G␣ i1-2 , G␣ i3 , G␣ s , and G␣ o (Table III).
To determine whether P 2Y receptors were invariably coupled to the same G proteins, similar measurements were done on solubilized membranes from rabbit aortic smooth muscle, heart, liver, and whole brain (Figs. 8 and 9). The results obtained in heart and vascular smooth muscle membranes were identical to those obtained in visceral smooth muscle membranes: UTP and ATP activated G q/11 and G i3 but not G i1 , G i2 , G o , or G s (Fig. 8). In liver, ATP and UTP activated G q/11 and G i1 and/or G i2 , but did not activate G i3 , G o , or G s (Fig. 9). In contrast to vascular and visceral smooth muscle, heart, and liver, ATP and UTP activated predominantly G o , as well as G i3 and G i1 and/or G i2 in brain membranes, but did not activate G q/11 or G s (Fig. 9). The extent of activation of specific G proteins by ATP and UTP was similar in all tissues except brain where activation of all inhibitory G proteins by ATP was more pronounced, suggesting interaction of ATP with P 2Y2 receptors and ATP-preferring P 2Y1 receptors.

Identification of P 2Y and P 2X Receptors in Smooth Muscle Cells by Selective Receptor Protection-
The pattern of PI hydrolysis, IP 3 formation, Ca 2ϩ mobilization and contraction suggested that UTP and ATP interacted with a common P 2Y receptor coupled to PTX-sensitive and -insensitive G proteins, whereas AMP-PCP, AMP-CPP, and 2-methylthio-ATP interacted with a distinct ligand-gated P 2X receptor; the latter was also activated by high concentrations of ATP. This notion was corroborated by selective receptor protection to enrich the muscle cells with one receptor subtype. After selective receptor protection, muscle cells were incubated for 60 min in control medium to allow complete resensitization of the cells (see "Experimental Procedures").
Receptor protection with 10 nM AMP-PCP preserved completely the contractile response to 10 nM AMP-PCP (15 Ϯ 3% decrease in cell length) and AMP-CPP (14 Ϯ 2%) (see Fig. 2 and Table I for comparison with responses to untreated muscle cells), but not the responses to ATP or UTP. An identical pattern was obtained by receptor protection with 10 nM AMP-CPP. In contrast, receptor protection with 10 nM UTP preserved completely the responses to 10 nM UTP (15 Ϯ 1%) and ATP (13 Ϯ 2%), but not the responses to AMP-PCP and AMP-CPP. An identical pattern was obtained by receptor protection with 10 nM ATP. Receptor protection with a high concentration of ATP (10 M) preserved completely the responses to UTP and ATP as well as the responses to AMP-PCP and AMP-CPP, FIG. 8. Binding of ATP-and UTP-stimulated GTP␥S⅐G␣ complexes to G␣ protein antibodies in solubilized membranes from aortic muscle and heart. Membranes isolated from rabbit aortic smooth muscle and heart were solubilized and incubated with [ 35 S]GTP␥S in the presence or absence of 10 M ATP or UTP for 20 min. Aliquots were added to wells precoated with various G␣ antibodies for 2 h and bound radioactivity measured (counts/min/mg of protein). ATP and UTP caused significant (p Ͻ 0.01) increases in binding of [ 35 S]GTP␥S⅐G␣ complexes to G␣ i3 antibody and G␣ q/11 antibody but not to G␣ i1-2 , G␣ s , or G␣ o antibody. Values are mean Ϯ S.E. of five experiments. **, significant increase above control GTP␥S binding, p Ͻ 0.01. implying that at this concentration ATP interacted with P 2X and P 2Y receptors on muscle cells.
After selective protection of P 2Y receptors with UTP, the contractile response to a high concentration of ATP (10 M) could be abolished by U-73122 (control response: 28 Ϯ 1% decrease in cell length; with U-73122: 3 Ϯ 2%) implying that it was exclusively mediated by PI hydrolysis. Following desensitization of ligand-gated P 2X receptors by preincubation of muscle cells for 30 min with 10 M AMP-PCP, the contractile response to 10 M ATP was virtually abolished by GDP␤S and U-73122 (control, 30 Ϯ 2% decrease in cell length; GDP␤S, 3 Ϯ 1%; U-73122, 4 Ϯ 2%) and partly inhibited by PTX (control, 30 Ϯ 2%; PTX, 13 Ϯ 2%). Thus, after selective desensitization of P 2X receptors or selective protection of P 2Y receptors, the response to a high concentration of ATP (10 M) reflected exclusively activation of G protein-coupled pathways. DISCUSSION This study demonstrates the co-existence of ligand-gated P 2X and G protein-coupled P 2Y receptors on freshly dispersed gas-tric smooth muscle cells and suggests that ATP activates preferentially P 2Y receptors to elicit Ca 2ϩ mobilization and muscle contraction. The P 2Y receptors selectively activated by UTP and ATP were coupled to PLC-␤1 via G␣ q/11 and to PLC-␤3 via G␤␥ i3 . Concurrent activation of the two effector enzymes resulted in PTX-sensitive and -insensitive IP 3 formation and IP 3 -dependent Ca 2ϩ release from sarcoplasmic stores. The high affinity for UTP and ATP suggested that these were P 2Y2 receptors: UTP-selective P 2Y4 receptors and UDP-selective P 2Y6 receptors coupled to PLC-␤ can be expressed in smooth muscle but they exhibit low or minimal affinity for ATP (11,19). The P 2X receptors selectively activated by AMP-PCP and AMP-CPP-mediated Ca 2ϩ influx via dihydropyridine-sensitive, voltage-gated Ca 2ϩ channels; the Ca 2ϩ channels were activated by depolarization of the plasma membrane that resulted from the opening of ligand-gated cation P 2X receptor/channels (22,23). Their presence on smooth muscle (which predominantly expresses P 2X1 receptors (21)), and their activation by AMP-CPP (which selectively interacts with P 2X receptors on smooth muscle (36)), and by AMP-PCP (which interacts with P 2X1 and P 2X3 receptors (23,24,36)), suggested that the receptors were of the P 2X1 subtype. Although the activity profile in visceral smooth muscle (AMP-PCP ϭ AMP-CPP Ͼ 2-methylthio-ATP Ͼ ATP) resembled that seen in vascular smooth muscle (17), it differed from the activity profile determined in patch-clamp studies of the cloned human and rat P 2X1 receptors where ATP and 2-methylthio-ATP were more potent than AMP-PCP (4,5,23). It seems unlikely that the difference reflected degradation of ATP or 2-methylthio-ATP by ecto-nucleotidases, since the measurements of response, particularly those of [Ca 2ϩ ] i , in both visceral and vascular smooth muscle were virtually instantaneous (Ͻ2 s), and the ratio of medium to cell volume (5000:1) was very high. It is possible that the rabbit P 2X1 receptor is different from the human and rat homologs or that its conformation or extent of polymerization is different when it is expressed in smooth muscle.
The evidence for the co-existence and subtype of P 2Y and P 2X receptors in smooth muscle may be summarized as follows. First, ATP and UTP stimulated IP 3 formation, Ca 2ϩ release, and contraction in dispersed smooth muscle cells. The responses to UTP and low nanomolar concentrations of ATP were abolished by GDP␤S and the PLC-␤ inhibitor, U-73122, and partly inhibited by PTX, implying the participation of PTXsensitive and -insensitive G proteins in IP 3 formation and IP 3 -dependent Ca 2ϩ release. In contrast, ATP analogs with high affinity for P 2X receptors did not stimulate IP 3 formation; contraction and the increase in [Ca 2ϩ ] i were abolished by nifedipine and in Ca 2ϩ -free medium implying that they were mediated by Ca 2ϩ influx. Higher micromolar concentrations of ATP stimulated both Ca 2ϩ influx and IP 3 -dependent Ca 2ϩ release. A similar pattern was observed by Pacaud et al. (17) for the [Ca 2ϩ ] i response in single aortic smooth muscle cells.
Second, the interaction of ATP analogs with P 2X receptors, and UTP and ATP with P 2Y receptors was corroborated in experiments using smooth muscle cells enriched with one receptor subtype. The validity of this approach was previously established for several agonists (44 -48). Selective protection of P 2Y2 receptors with either UTP or ATP preserved the contractile response to both UTP and ATP, whereas selective protection of P 2X1 receptors with either AMP-PCP or AMP-CPP preserved the response to both analogs and to high concentrations of ATP.
Third, ATP and UTP activated both G q/11 and G i3 in muscle membranes; activation of G i3 was suppressed by pretreatment of muscle cells with PTX. PI hydrolysis stimulated by ATP and UTP in plasma membranes was mediated concurrently by the FIG. 9. Binding of ATP-and UTP-stimulated GTP␥S⅐G␣ complexes to G␣ protein antibodies in solubilized membranes from liver and brain. Membranes isolated from rabbit liver and whole brain were solubilized and incubated with [ 35 S]GTP␥S in the presence or absence of 10 M ATP or UTP for 20 min. Aliquots were added to wells precoated with various G␣ antibodies for 2 h and bound radioactivity measured (counts/min/mg of protein). ATP and UTP caused significant (p Ͻ 0.01) increases in binding of [ 35 S]GTP␥S⅐G␣ complexes to G␣ q/11 and G␣ i1-2 antibodies in liver membranes, and to G␣ o , G␣ i1-2 , and G i3 antibodies in brain membranes. Values are mean Ϯ S.E. of five experiments. **, significant increase above control GTP␥S binding, p Ͻ 0.01. The increase induced by ATP in brain membranes was significantly greater than that induced by UTP, p Ͻ 0.02.