A βγ Dimer Derived from G13 Transduces the Angiotensin AT1 Receptor Signal to Stimulation of Ca2+ Channels in Rat Portal Vein Myocytes*

A G protein composed of α13, β1, and γ3 subunits selectively couples the angiotensin AT1A receptors to increase cytoplasmic Ca2+ concentration ([Ca2+] i ) in rat portal vein myocytes (Macrez-Leprêtre, N., Kalkbrenner, F., Morel, J. L., Schultz, G., and Mironneau, J. (1997) J. Biol. Chem. 272, 10095–10102). We show here that Gβγ transduces the signal leading to stimulation of L-type Ca2+channels. Intracellular dialysis through the patch pipette of a carboxyl-terminal anti-βcom antibody and a peptide corresponding to the Gβγ binding region of the β-adrenergic receptor kinase 1 inhibited the stimulation of Ca2+channels and the increase in [Ca2+] i evoked by angiotensin II. The Gβγ binding peptide did not prevent the dissociation of the heterotrimeric G protein into its subunits, as it did not block activation of phospholipase C-β by Gαq in response to stimulation of α1-adrenoreceptors. Transient overexpression of the β-adrenergic receptor kinase 1 fragment and of Gα subunits also inhibited the angiotensin II-induced increase in [Ca2+] i . Both anti-α13 antibody and carboxyl-terminal α13 peptide abrogated the angiotensin II-induced stimulation of Ca2+ channels. We conclude that activation of angiotensin AT1 receptors requires all three α, β, and γ subunits of G13 for receptor-G protein interaction, whereas the transduction of the signal to L-type Ca2+ channels is mediated by Gβγ.

Specific heterotrimeric G proteins composed of different ␣, ␤, and ␥ subunits transmit signals from membrane receptors to intracellular effectors (1)(2). When a receptor is activated by an agonist, it catalyzes the exchange of GDP for GTP on the ␣ subunit of G proteins, resulting in dissociation of ␣ subunits from the ␤␥ dimers. It is now well documented that both the G␣ subunit and the G␤␥ complex are able to transmit signals to effector molecules (3)(4). After the initial observation that G␤␥ could activate K ϩ channels (5), it was found that G␤␥ can regulate certain isoforms of adenylyl cyclase (6) and phospho-lipase C-␤ (7), activate the mitogen-activated protein kinase (8) and c-Jun N-terminal kinase (9) pathways, and mediate the translocation of the ␤-adrenergic receptor kinase (␤ARK) 1 (10).
In portal vein myocytes, the G protein heterotrimer that couples angiotensin AT 1A receptors to increase [Ca 2ϩ ] i has been identified using an antisense oligonucleotide strategy. The G protein is composed of ␣ 13 , ␤ 1 , and ␥ 3 subunits, all three being required for activation of the transduction pathway (11). Angiotensin II (AII)-induced increase in [Ca 2ϩ ] i is initiated by activation of L-type Ca 2ϩ channels, producing a slow elevation of [Ca 2ϩ ] i (12) that, in turn, activates a Ca 2ϩ -induced Ca 2ϩ release from the intracellular store (13). It has been shown that AII produces Ca 2ϩ release from the intracellular store by opening of ryanodine-sensitive Ca 2ϩ release channels, as evidenced by the increase in Ca 2ϩ spark frequency (14).
The purpose of the present study was to identify which G protein subunits transduce the signal for activation of Ca 2ϩ channels after stimulation of the angiotensin AT 1 receptor in rat portal vein myocytes. A specific G␣ 13 function-blocking antibody and G␣ 13 peptide (corresponding to the last 11 amino acids of the carboxyl terminus) abrogated AII-induced stimulation of Ca 2ϩ channel current when dialyzed into the cells through the patch pipette. Intracellular infusion of specific G␤␥ binding agents, i.e. anti-␤ com antibody and ␤ARK peptide, also blocked the AII-induced stimulation of Ca 2ϩ channel current. Finally, overexpression of a ␤ARK 1 fragment and of G␤␥ scavengers, i.e. wild type of G␣ o1 and G␣ 12 subunits, largely inhibited the AII-induced increase in [Ca 2ϩ ] i . We conclude that the angiotensin AT 1 receptor uses the ␤␥ dimers of G 13 to transduce the signal leading to activation of Ca 2ϩ channels.

EXPERIMENTAL PROCEDURES
Cell Preparation-Isolated myocytes from rat portal vein were obtained by enzymatic dispersion, as described previously (15). Cells were seeded at a density of about 10 3 cells/mm 2 on glass slides imprinted with squares for localization of injected cells and maintained in shortterm primary culture in M199 medium containing 2% fetal calf serum, 2 mM glutamine, 1 mM pyruvate, 200 units/ml penicillin, and 200 g/ml streptomycin; they were kept in an incubator gassed with 95% air, 5% CO 2 at 37°C and used within 72 h.
Membrane current and [Ca 2ϩ ] i Measurements-Voltage clamp and membrane current recordings were made with a standard patch-clamp technique using a List EPC-7 patch-clamp amplifier (Darmstadt-Eberstadt, Germany). Whole-cell recordings were performed with patch pipettes having resistances of 1-3 megaohms. Membrane potential and current records were stored and analyzed using an IBM PC computer (P-clamp system, Axon Instruments, Inc., Foster City, CA). Simultane-* This work was supported by grants from Centre National de la Recherche Scientifique and Centre National des Etudes Spatiales (France) and from the Deutsche Forschungsgemeinschaft and Fonds der Chemischen Industrie (Germany). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ Supported by a fellowship from the Fondation pour la Recherche Medicale (France).
ʈ Present address: Research Laboratory of Schering AG, 13342 Berlin, Germany.
** To whom correspondence should be addressed. Tel.: 33 5 57 57 12 31; Fax: 33 5 57 57 12 26. ous measurements of intracellular Ca 2ϩ concentration were carried out in some experiments. Briefly, 60 M fura-2 was added to the pipette solution and entered cells after establishment of the whole-cell recording mode. [Ca 2ϩ ] i was estimated from the 340/380-nm fluorescence ratio using a calibration determined within cells (15). All measurements were made at 25 Ϯ 1°C.
Measurements of Cytosolic Ca 2ϩ -Cells were loaded by incubation in physiological solution containing 1 M fura-2-acetoxymethyl ester for 30 min at room temperature. These cells were washed and allowed to cleave the dye to the active fura-2 compound for at least 1 h. Fura-2 loading was usually uniform over the cytoplasm, and compartmentalization of the dye was never observed. Measurement of cytosolic Ca 2ϩ concentration was carried out by dual-wavelength fluorescence method, as described previously (15). Briefly, fura-2-loaded cells were mounted in a perfusion chamber and placed on the stage of an inverted microscope (Nikon Diaphot, Tokyo, Japan). Single cells were alternatively excited with UV light at 340 and 380 nm through a 10 ϫ oil immersion objective, and emitted fluorescent light from the Ca 2ϩ -sensitive dye was collected through a 510-nm long-pass filter with a charge-coupled device camera (Hamamatsu Photonics, Hamamatsu City, Japan). The signal was processed (Hamamatsu DVS 3000) by correcting each fluorescence image for background fluorescence and calculating 340/ 380-nm fluorescence ratios on a pixel-to-pixel basis. Averaged frames were usually collected at each wavelength every 0.5 s. [Ca 2ϩ ] i was calculated from mean ratios using a calibration for fura-2 determined in loaded cells. All measurements were made at 25 Ϯ 1°C.
Transfection-cDNA encoding ␤-adrenergic receptor kinase was cloned into expression plasmids pRK 5 (8). cDNAs encoding for G␣ o and G␣ 12 subunits were cloned into pECE (16). cDNA encoding for S65T green fluorescent protein was cloned into pcDNA 3 (CLONTECH, Palo Alto, CA). Briefly, plasmids were diluted with water from stock solutions (0.5 g/l) to final concentrations of 0.1 g/l and injected directly into the nucleus of vascular myocytes. The S65T green fluorescent protein (GFP) was included to facilitate later identification of myocytes having a successful nuclear injection and plasmid expression. Fluorescence produced by the S65T GFP was observed 3 days after injection with a confocal microscope (Bio-Rad MRC 1000, Paris, France). From injected cells (n ϭ 280), about 20% showed a detectable fluorescence signal.
Solutions-The normal physiological solution contained 130 mM NaCl, 5.6 mM KCl, 1 mM MgCl 2 , 2 mM CaCl 2 , 11 mM glucose, 10 mM HEPES, pH 7.4, with NaOH. The basic pipette solution contained 130 mM CsCl, 10 mM HEPES, pH 7.4, with CsOH. Ca 2ϩ -free external solution was prepared by omitting CaCl 2 and adding 0.5 mM EGTA. For the recording of Ca 2ϩ channel current, 5 mM BaCl 2 was substituted for CaCl 2 in the reference solution, and CsCl was used in the pipette and external solutions to block outward potassium currents. In addition, 10 mM EGTA, 5 mM Na 2 ATP, and 1 mM MgCl 2 were added to the basic pipette solution. Substances were applied to the cells by pressure ejection from a glass pipette for the period indicated on the records.
Data Analysis-Results are expressed as means Ϯ S.E. Significance was tested by means of Student's t test. P values of Ͻ0.05 were considered as significant. 1 Receptors to Stimulation of L-type Ca 2ϩ Channels-To identify the G proteins activated by the angiotensin AT 1 receptor, we used antibodies raised against the carboxyl terminus of ␣ sub-units to block interactions of G proteins with angiotensin AT 1 receptors and synthetic peptides corresponding to the carboxyl terminus of ␣ subunits to disrupt the angiotensin AT 1 receptorevoked activation of G proteins. In the continuous presence of 100 nM CGP 42112A (to block angiotensin AT 2 receptors; Ref. 17), 10 nM AII increased the Ba 2ϩ current by about 40% (Fig.  1A). The stimulatory effect of AII reached a steady state within 1-2 min and was progressively reversed within 5-10 min. When anti-␣ 12 antiserum (at 1:100 or 1:50) was added to the pipette solution for 7-10 min, the AII-induced stimulation of the Ba 2ϩ current was not significantly affected (Fig. 1B). Similarly, intracellular applications of the anti-␣ q/11 antibody (15 g/ml) for 7-8 min were without effect (Fig. 1B). In contrast, intracellular applications of anti-␣ 13 antiserum (1:100) or anti-␣ 13 purified antibody (10 g/ml) blocked the AII-induced stimulation of the Ba 2ϩ current (Fig. 1, A and B). The specificity of the anti-␣ 13 antibody is documented in Fig. 1, A and B, since intracellular application of 10 g/ml boiled anti-␣ 13 antibody (95°C for 30 min) did not alter the AII-induced stimulation of the Ba 2ϩ current. Fig. 2 illustrates the effects of synthetic peptides corresponding to the carboxyl terminus of ␣ q/11 and ␣ 13 subunits. Intracellular applications of the ␣ 13 peptide for 7-8 min inhibited the AII-induced stimulation of the Ba 2ϩ current in a concentration-dependent manner (Fig. 2, A and B). The concentration of ␣ 13 peptide producing half-maximal inhibition was estimated to be 4 ng/ml. Complete inhibition was obtained with 0.1 g/ml ␣ 13 peptide. It has to be noted that the ␣ 13 peptide by itself had no effect on the Ba 2ϩ current, as the current density, normalized by the cell capacitance, was not significantly modified (control: 12.5 Ϯ 2.5 microampere/microfarad, n ϭ 10; in the presence of 1 g/ml ␣ 13 peptide: 13.2 Ϯ 3.5 microampere/microfarad, n ϭ 8). Intracellular applications of ␣ q/11 peptide (0.1-1 g/ml) had no effect on the AII-induced stimulation of the Ba 2ϩ current (control: 46 Ϯ 5%, n ϭ 10; in the presence of 1 g/ml ␣ q/11 peptide: 44 Ϯ 4%, n ϭ 6). Taken together, these results suggest that G 13 , but not G 12 and G q/11 , functionally couples the angiotensin AT 1 receptors to stimulation of L-type Ca 2ϩ Channels.

Identification of the G Protein-coupling Angiotensin AT
G␤␥ Is Required for AII-induced Stimulation of Ca 2ϩ Channels-Anti-G␣ antibody and G␣ subunit peptide block of the response alone cannot distinguish which G protein subunit (G␣ or G␤␥) transduces the signal to Ca 2ϩ channels. To determine which G protein subunit was involved in effector activation, an anti-␤ com antibody (18), raised to the carboxyl terminus of G␤ 1 subunit, was dialyzed into the cell for 7-8 min. As shown in Fig. 3A, intracellular applications of 10 g/ml anti-␤ com antibody blocked the AII-induced stimulation of the Ba 2ϩ current. In contrast, application of the same concentration of boiled anti-␤ com antibody (95°C for 30 min) had no significant effect on the AII-induced stimulation of the Ba 2ϩ current (Fig. 3A).
In a second set of experiments, we dialyzed peptides corresponding to fragments of ␤ARK 1 (19) into the cells for 5-6 min. Carboxyl-terminal fragments of ␤ARK 1 have been used to bind G␤␥ subunits and to block activation of effectors (19 -21). Intracellular applications of peptide G (corresponding to the G␤␥ binding domain of ␤ARK 1 ) inhibited the AII-induced stimulation of the Ba 2ϩ current in a concentration-dependent manner (Fig. 3B). The concentration of peptide G producing half-maximal inhibition was estimated to be 65 nM. Complete inhibition was obtained with 1 M peptide G. In contrast, intracellular applications of 10 -100 M peptide A (corresponding to a domain of ␤ARK 1 not involved in G␤␥ binding) had no effect on the AII-induced stimulation of the Ba 2ϩ current (Fig. 3B). These results suggest that the angiotensin AT 1 receptors transduce their signal to Ca 2ϩ channels through G␤␥.
G␤␥ Transduces AII-evoked Increase in [Ca 2ϩ ] i -We have previously reported that the AII-evoked increase in [Ca 2ϩ ] i is dependent on L-type Ca 2ϩ channel activation, leading to a slow elevation of [Ca 2ϩ ] i that triggers, in turn, a subsequent Ca 2ϩ release from the intracellular store through activation of ryanodine-sensitive Ca 2ϩ release channels (13). As shown in Fig.  4A, the AII-evoked increase in [Ca 2ϩ ] i was selectively inhibited by intracellular applications of 10 g/ml anti-␤ com antibody for 7-8 min. The same concentration of boiled anti-␤ com antibody was without effect. Furthermore, intracellular dialysis of 10 M peptide G also inhibited the AII-induced increase in [Ca 2ϩ ] i , whereas 10 M peptide A was ineffective. It has to be noted that complete inhibition of the AII-induced increase in [Ca 2ϩ ] i was not obtained with increasing concentrations (100 M) of peptide G (n ϭ 8). This is in contrast to the block of the AII-induced stimulation of the Ba 2ϩ current obtained with 1 M peptide G.
Inhibition of the AII-dependent Ca 2ϩ signaling by binding of G␤␥ could be related to two distinct mechanisms. (i) The inhibitory proteins (␤ARK 1 peptide, anti-␤ com antibody) prevented the AII-induced dissociation of G␣ 13 and G␤␥ subunits, thus inhibiting the transduction process or (ii) they bind to G␤␥ after its release from G␣ and, therefore, inhibit the G␤␥-effector coupling. To distinguish between these possibilities, the cells were stimulated by 10 M norepinephrine to produce an increase in [Ca 2ϩ ] i dependent on Ca 2ϩ release from the intracellular store. This transduction pathway involves G␣ q , which activates a phospholipase C-␤, leading to InsP 3 production and the subsequent activation of InsP 3 -gated channels (15). As illustrated in Fig. 4B, intracellular applications of 10 -100 M peptide G to scavenge free G␤␥ had no effect on the norepinephrine-induced Ca 2ϩ release that was mediated by the ␣ q subunit. Moreover, intracellular applications of both peptide G (100 M) and ␣ q/11 peptide (10 g/ml) inhibited the norepinephrine-induced Ca 2ϩ release (Fig. 4B). These results show that, although G␤␥ is bound to the peptide G, G␣ q on its own can transduce and support the Ca 2ϩ response evoked by norepinephrine.
In a second set of experiments, we overexpressed a carboxylterminal fragment of ␤ARK 1 by intranuclear microinjection of expression plasmids containing cDNA inserts coding for ␤ARK 1 fragment and S65T GFP. Overexpression of the ␤ARK 1 fragment was followed by cytoplasmic detection of S65T GFP, 3 days after injection. As illustrated in Fig. 5A, the AII-induced increase in [Ca 2ϩ ] i was inhibited by about 75% in cells overexpressing the ␤ARK 1 fragment (control: 103 Ϯ 13 nM, n ϭ 20; after ␤ARK 1 fragment overexpression: 25 Ϯ 9 nM, n ϭ 12). It has to be noted that in all the cells injected with expression plasmids, the basal [Ca 2ϩ ] i level was significantly increased (control: 47 Ϯ 5 nM, n ϭ 50; after plasmid injection: 88 Ϯ 5 nM, n ϭ 56). However, in cells injected with pRK 5 plasmids alone, AII evoked a Ca 2ϩ response whose amplitude (98 Ϯ 9 nM, n ϭ 12) was not significantly different from that measured in noninjected cells (Fig. 5B).
Another strategy to confirm the role of G␤␥ in AII-induced increase in [Ca 2ϩ ] i was to increase the normal G␣:G␤␥ ratio by overexpressing G␣ subunits. Excess of GDP-bound G␣ subunits with a high affinity for G␤␥ subunits should create conditions favoring heterotrimeric formation. Injection of expression plasmids containing cDNA inserts coding for wild type (WT) G␣ o1 (which is not endogenously expressed in vascular myocytes; Ref. 22) or WT G␣ 12 inhibited the AII-induced increase in [Ca 2ϩ ] i by about 75% (Fig. 5B). In contrast, overexpression of the constitutively active Q229L G␣ 12 did not significantly affect the AII-induced increase in [Ca 2ϩ ] i . Taken together, these results are consistent with the idea that G␤␥ controls the transduction pathway, leading to stimulation of Ca 2ϩ channels and increase in [Ca 2ϩ ] i in response to activation of angiotensin AT 1 receptors. DISCUSSION Our results show that, in rat portal vein myocytes, G 13 couples angiotensin AT 1 receptors to stimulation of L-type Ca 2ϩ channels. The G␣ 13 ␤ 1 ␥ 3 heterotrimer provides the specificity for the coupling with AT 1 receptors (11), but the signal to Ca 2ϩ channel is transduced by the G␤␥ dimer. This conclusion is based on experiments using antibodies raised against the carboxyl-terminal G␣ or G␤ subunits, synthetic peptides corresponding to the carboxyl terminus of G␣ subunits or to ␤ARK 1 fragments, and overexpression of a ␤ARK 1 fragment and of G␣ subunits to disrupt the angiotensin AT 1 receptor-evoked activation of Ca 2ϩ channels.
Involvement of G␣ 13  that selectively blocks the AII-induced stimulation of Ca 2ϩ channels and the increases in [Ca 2ϩ ] i . These results support the idea that activation of Ca 2ϩ channels via G 13 is the initial step leading to an increase in [Ca 2ϩ ] i . To confirm these results, we used synthetic peptides corresponding to the carboxyl terminus of G␣ subunits. These peptides have been shown to bind to receptors and to stabilize them in a high affinity conformational state (23)(24)(25). Thus, synthetic peptides may act as competitive agonists at the receptor/G protein interface and block receptor-mediated activation of effectors (26). As carboxyl-terminal G␣ 13 peptide selectively abrogates the AII-induced stimulation of Ca 2ϩ channels whereas carboxyl-terminal G␣ q/11 peptide is ineffective, our results suggest that the activated angiotensin AT 1 receptors specifically interact with the extreme carboxyl terminus of G␣ 13 12 inhibited the AII-in-duced Ca 2ϩ response. These experiments clearly show that G␤␥ is necessary to mediate the AII-induced increase in [Ca 2ϩ ] i . The fact that the inhibitory proteins interacting either with G␤␥ or G␣ inhibit the AII-induced Ca 2ϩ signaling may indicate that they bind to the undissociated heterotrimer and prevent the dissociation of the G protein into its subunits. Another possibility is that the inhibitory proteins bind to the dissociated G protein subunits and suppress their action on the effectors. This latter possibility is supported by our observations that overexpression of G␣ o and G␣ 12 (which cannot bind to heterotrimeric G proteins) inhibited the AII-induced Ca 2ϩ responses. In addition, intracellular applications of the ␤ARK 1 peptide G did not inhibit the norepinephrine-activated Ca 2ϩ response as would be expected if this peptide prevented the G protein heterotrimer dissociation. Although G␤␥ is bound to the peptide G, G␣ q on its own can transduce and support Ca 2ϩ release from the intracellular store in response to activation of ␣ 1 -adrenoreceptors (27). Taken together, these observations indicate that these inhibitory proteins interact with the free G protein subunits and, thus, suppress their actions on the effectors.
Coupling of G proteins with Ca 2ϩ channels has been previously reported and generally promotes an inhibition of neuronal Ca 2ϩ currents (28 -30). This inhibition can result from a direct interaction between G␤␥ and the pore-forming ␣ 1 subunits of N-and P/Q-type Ca 2ϩ channels but not of L-type Ca 2ϩ channels (31). In contrast, L-type Ca 2ϩ currents in cardiac and vascular myocytes can be enhanced by phosphorylation by protein kinases A and C, probably on phosphorylation sites located on the carboxyl terminus of the ␣ 1 subunit or on the ␤ subunit of Ca 2ϩ channels (32,33). Protein kinase C has been reported to increase Ca 2ϩ channel activity through the modulation of the mean open time, a voltage-independent property, whereas protein kinase A enhances channel activity through the modification of the voltage-dependence of activation and inactivation (34). In portal vein myocytes, AII increases the Ca 2ϩ channel current at all potentials but does not shift the currentvoltage relationship. However, the relative increase in Ca 2ϩ current induced by AII is more pronounced at negative potentials, i.e. between Ϫ40 and Ϫ20 mV, than at 0 mV, 2 supporting our previous data that the AII-induced stimulation of Ca 2ϩ channels involves activation of protein kinase C (12).
G␤␥-induced Ca 2ϩ signaling may also involve various effector systems leading to second messenger production. It has been reported in several cell types that G␤␥-mediated activation of phospholipase C-␤ 2 or -␤ 3 is associated with a relatively low increase in InsP 3 (35). Moreover, G␤␥ may increase the apparent affinity of the InsP 3 -gated Ca 2ϩ channels to InsP 3 , so that Ca 2ϩ release may occur at low InsP 3 concentrations (36). Recently, it has been shown that G␤␥ is involved in the regulation of the activity of the small GTP-binding proteins Rho and Rac (37). Thus, G␤␥ may mediate the association of activated Rho and Rac to the membrane and activate, in turn, various enzymes such as phospholipase D or phosphoinositide 3-kinase (38 -40), leading to generation of various second messengers. The specificity of the G␤␥-activated coupling could depend on localized distribution of certain subsets of receptors, G proteins, and effectors within distinct membrane domains that may have access to each other (41). Our results suggest that G␤␥ activates L-type Ca 2ϩ channels by a transduction pathway that does not involve InsP 3 -gated Ca 2ϩ release channels, since the AII-induced increase in [Ca 2ϩ ] i is insensitive to both heparin (an inhibitor of the InsP 3 receptor) and anti-phosphatidylinositol antibody (13). In addition, biochemical and pharmaco- logical approaches have proposed that AII may induce phosphatidylcholine hydrolysis by phospholipases D (42) or C (12,43), leading to diacylglycerol formation. Release of diacylglycerol from phosphatidylcholine hydrolysis is a slow phenomenon (42), supporting the observation that the AII-induced stimulation of Ca 2ϩ channel current reaches a maximum value within 1-2 min. Activation of protein kinase C by diacylglycerol may be responsible for phosphorylation of L-type Ca 2ϩ channels and the subsequent increase in Ca 2ϩ channel activity (34). In vascular myocytes, openings of L-type Ca 2ϩ channels at the resting potential (Ϫ50 mV) have been previously proposed to serve as a pathway for Ca 2ϩ influx in response to receptor activation (44,45).
In conclusion, we show that in rat portal vein myocytes, activation of angiotensin AT 1 receptors requires G 13 heterotrimer for specificity and G␤␥ dimer for transducing the signal to L-type Ca 2ϩ channels and increase in [Ca 2ϩ ] i .