Cell Cycle-dependent Coupling of the Vasopressin V1a Receptor to Different G Proteins*

Arginine vasopressin (AVP) regulates biological processes by binding to G protein-coupled receptors. In Swiss 3T3 fibroblasts, expressing the V1a subtype of vasopressin receptors, AVP mobilizes calcium from intracellular stores. In proliferating cells, the AVP-induced increase in intracellular calcium concentration ([Ca2+] i ) was mediated by G proteins of the Gq family, which are insensitive to pertussis toxin (PTX) pretreatment of the cells. In quiescent cells, the AVP-induced increase in [Ca2+] i was partially PTX-sensitive, suggesting an involvement of Gi proteins. We confirmed this by photoaffinity labeling of G proteins in Swiss 3T3 cell membranes activated by AVP. In Swiss 3T3 cells arrested in the G0/G1 phase of the cell cycle, the AVP-induced increase in [Ca2+] i was also partially PTX-sensitive but was PTX-insensitive in cells arrested in other phases of the cell cycles. The blocking effect of PTX pretreatment in G0/G1 cells was mimicked by microinjection of antisense oligonucleotides suppressing the expression of the Gαi3 subunits. These results were confirmed by microinjection of antibodies directed against the C terminus of G protein α-subunits. The data presented indicate that in Swiss 3T3 fibroblasts synchronized in the G0/G1 phase of the cell cycle the V1a receptor couples to Gq/11 and Gi3 to activate the phospholipase C-β, leading to release of intracellular calcium.

The biological effects of the neuropeptide hormone arginine vasopressin (AVP) 1 are induced by binding of AVP to specific membrane receptors that are members of the GTP-binding protein (G protein)-coupled receptor (GPCR) superfamily. To date, three AVP receptor subtypes (V 1a , V 1b , and V 2 ) have been cloned and characterized according to their tissue distribution and functional properties. The V 2 receptor subtype is predom-inantly expressed in the kidney, mediating the antidiuretic effects of AVP. The activated V 2 receptor couples to the heterotrimeric G protein G s and activates adenylyl cyclases (1,2). Two V 1 -receptor subtypes have been cloned (3) and can be pharmacologically differentiated by their binding affinities to various AVP agonists and antagonists (4,5). While the expression of the V 1b receptor subtype is restricted mainly to the central nervous system (6), the V 1a receptor is expressed in different neuronal and nonneuronal tissues. In nonneuronal tissues, the V 1a receptor induces a wide range of physiological effects such as contraction of vascular smooth muscles, stimulation of hepatic glycogenolysis and cell proliferation (for a review, see Refs. 7 and 8).
By binding to GPCRs, many hormones and neurotransmitters activate various subtypes of the phospholipase C-␤ (PLC-␤) (9). These enzymes catalyze the hydrolysis of phosphatidylinositol 4,5-bisphosphate to the two second messenger molecules, diacylglycerol and inositol 1,4,5-trisphosphate. While diacylglycerol stimulates the activity of protein kinase C and Ca 2ϩ influx, inositol 1,4,5-trisphosphate binds to receptors in the endoplasmatic reticulum leading to the release of Ca 2ϩ from the intracellular stores and increases the intracellular calcium concentration ([Ca 2ϩ ] i ). The activity of most PLC-␤ isozymes is regulated by G protein ␣-subunits of the G q family (␣ q , ␣ 11 , ␣ 14 , ␣ 15/16 ). Thus, hydrolysis of phosphatidylinositol 4,5-bisphosphate and the subsequent increase in [Ca 2ϩ ] i mediated by receptors coupling to G proteins of the G q family are not inhibited by pertussis toxin (PTX). However, in some cell types, e.g. cells of the hematopoietic origin, the receptor-mediated increase in [Ca 2ϩ ] i can be blocked by pretreatment of the cells with PTX (10,11), suggesting the involvement of PTX-sensitive G proteins of the G i family in the activation of PLC-␤. Indeed the PLC-␤ 2 and -␤ 3 isoforms have been found to be activated by ␤␥ dimers released from G i proteins (11)(12)(13). Nevertheless, PLC-␤ 1 , -␤ 3 , and -␤ 4 are preferentially activated by the ␣ subunits of the G q family (14 -16).
Both V 1 receptor subtypes, V 1a and V 1b , have been shown to activate PLC-␤ leading to a transient increase in [Ca 2ϩ ] i (17,18). In most of the cell lines studied, the AVP-induced increase in [Ca 2ϩ ] i involves coupling of the V 1a receptor to the PTXinsensitive G proteins G q and G 11 and subsequent activation of PLC-␤ (19,20).
Swiss 3T3 fibroblasts, which are derived from embryonic mouse cells, endogenously express several GPCRs for neuropeptides such as bradykinin, bombesin, or AVP and, therefore, have widely been used to study intracellular signal transduction pathways induced by neuropeptides. In addition, these cells provide a useful model to study the regulation of cell growth. Swiss 3T3 cells express the V 1a receptor subtype (21). Using the technique of photoaffinity labeling of receptor-activated G␣ subunits with the GTP analog [␣-32 P]GTP azidoani-* This work was supported by grants from the Deutsche Forschungsgemeinschaft and Fonds der Chemischen Industrie. 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.
ʈ To whom correspondence should be addressed. Tel.: 49-7351-5498585; Fax: 49-7351-8398585; E-mail: Frank.Kalkbrenner@bc.boehringeringelheim.com. 1 The abbreviations used are: AVP, arginine vasopressin; aaGTP, [␣-32 P]GTP azidoanilide; DMEM, Dulbecco's modified Eagle's medium; FCS, fetal calf serum; G protein, regulatory heterotrimeric guanine nucleotide-binding protein; PLC, phospholipase C; PTX, pertussis toxin; PAGE, poyacrylamide gel electrophoresis; PBS, phosphate-buffered saline; RGS, regulator(s) of G protein signaling; GPCR, G proteincoupled receptor. lide (aaGTP), it has been shown that the AVP receptor in Swiss 3T3 cells activates G q and G 11 . No activation of G i by AVP was detected in this study (22). These results are in accordance with other studies showing that the AVP-induced activation of PLC-␤ through the V 1a receptor in various cell systems is insensitive to pretreatment of the cells with PTX (for a review, see Ref. 23). However, AVP-induced stimulation of DNA synthesis has been reported to be partially sensitive to PTX (24), suggesting that the V 1a receptor subtype principally is able to couple to PTX-sensitive G proteins in these cells. Remarkably, AVP-induced 3 H-labeled inositol phosphate formation was also inhibited by 23% by PTX pretreatment of the Swiss 3T3 cells in that study.
Herein, we studied the functional coupling of the AVP receptors in Swiss 3T3 cells to G proteins leading to the release of calcium from intracellular stores. We will provide evidence that AVP-evoked calcium release is mediated solely by the V 1a receptor subtype and that coupling of these receptors to G i3 and G q/11 is dependent on the cell cycle. 8 ]vasopressin, [1-(␤-mercapto-␤,␤-cyclopenthamethylene propionic acid),O-Me-Tyr 2 ,Arg 8 ]vasopressin, aphidicolin, and nocodazole were obtained from Sigma. Fura-2/AM and PTX were purchased from Calbiochem. 125 I-Phenylacetyl-D-Tyr(Me)-Phe-Gln-Asn-Arg-Pro-Arg-Tyr-NH 2 was from PerkinElmer Life Sciences. Phosphorothioate oligonucleotides were purchased from Eurogentec (Seraing, Belgium). The generation and the specificity of the polyclonal antibodies directed against G␣ i1-3 (AS266 and AS86, C-terminal antibodies), G␣ q/11 (AS370), and G␣ z (AS404) has been described (25). Additional affinity-purified polyclonal antibodies against G␣ q/11 (SC-392) and G␣ i3 (SC392) were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA).

Materials-[Arg
Cell Culture and Membrane Preparation-Swiss 3T3 fibroblasts were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum (FCS) in a humidified atmosphere containing 10% CO 2 and 90% air at 37°C. Pretreatment with PTX was performed by adding 100 ng/ml to the culture medium for 24 h prior to experiments. For measurement of intracellular calcium concentration ([Ca 2ϩ ] i ) and for photoaffinity labeling of G proteins, cells were cultured until they reached confluence. Before processing for experiments, cells were washed twice with PBS and incubated for 6 h either in fresh DMEM containing 10% FCS (proliferating cells) or DMEM without FCS (quiescent cells). Synchronization of cell growth in individual phases of the cell cycle was performed with exponentially growing Swiss 3T3 using the following protocols. For G 0 /G 1 phase, cells were washed with PBS and cultured in DMEM without FCS for 12 h, cultured another 16 h in fresh medium containing 10% FCS (G 0 /G 1 cells) and were detached from the culture flask with Hanks' balanced salt solution buffer (118 mM NaCl, 4.6 mM KCl, 1 mM MgCl 2 , 10 mM glucose, 5 mM EGTA, 20 mM HEPES, pH 7.2). For G 1 /S phase, aphidicolin (5 g/ml) was added to freshly changed culture medium (ϩ10% FCS), and cells were detached 20 h later with Hanks' balanced salt solution buffer. For G 2 /M phase, nocodazole (500 ng/ml) was added to freshly changed culture medium (ϩ10% FCS), and synchronized cells were detached from the culture flask 20 h later by gentle shaking. Control cells were maintained in culture medium containing 10% FCS (proliferating cells). For membrane preparation, Swiss 3T3 cells were washed twice with PBS; scraped off the tissue culture flask; resuspended in ice-cold buffer consisting of 100 mM NaCl, 0.5 mM EDTA, 50 mM KH 2 PO 4 ; and homogenized by nitrogen cavitation. Membranes were then sedimented at 100,000 ϫ g for 1 h and resuspended in 10 mM triethanolamine (pH 7.4). Protein concentration was determined with a BCA protein assay kit (Pierce), and membranes were stored at Ϫ70°C until they were used.
Flow Cytometric Analysis of Cell Synchrony-For analysis of cell cycle progression, samples of differently cultured cells were treated with ice-cold ethanol. Cell nuclei were sedimented by centrifugation, stained with propidium iodide in Ca 2ϩ -and Mg 2ϩ -free PBS supplemented with RNase (25 g/100 l), and counted according to the DNA amount using a fluorescence-activated cell sorter (Becton-Dickinson FACScan, Heidelberg, FRG). Cells with an unreplicated complement of DNA (2n) were assigned to G 0 /G 1 phase, cells with a fully replicated complement of DNA (4n) were assigned to the G 2 /M phase, and those cells containing an intermediate amount of DNA were assigned to the G 1 /S phase.
Photoaffinity Labeling of Receptor-activated G Protein ␣-Subunits-Synthesis of [␣-32 P]GTP azidoanilide and photoaffinity labeling of receptor-activated G proteins were performed as described (26). For immunoprecipitation of G proteins, photolabeled membranes were solubilized in 40 l of 2% (w/v) SDS, and 120 l of precipitation buffer (1% (w/v) Nonidet P-40, 1% (w/v) desoxycholate, 0.5% (w/v) SDS, 150 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, 0.2 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 10 mM Tris-HCl, pH 7.4) was added. Thirty l of an antiserum specific for G␣ q/11 (AS 370) or 20 l of an antiserum against all three isoforms of G␣ i (AS 266) were added, and the mixture was incubated overnight at 4°C. Protein A-Sepharose beads were added for 1 h, pelleted by centrifugation, and washed twice with 1 ml of buffer A (1% (w/v) Nonidet-P40, 0.5% (w/v) SDS, 600 mM NaCl, 50 mM Tris-HCl, pH 7.4) and twice with 1 ml of buffer B (300 mM NaCl, 10 mM EDTA, 100 mM Tris-HCl, pH 7.4). Samples were subjected to standard SDS-PAGE, and dried gels were analyzed for radioactivity with a phosphor imager (Fuji BAS 1000, Raytest, Mü nchen, FRG). Quantitative analysis of the digitized gel images was performed on a Macintosh computer using the public domain NIH Image program (developed at the U.S. National Institutes of Health and available on the Internet). The values are given as arbitrary units.
Microinjection of Oligonucleotides and Antibodies-Swiss 3T3 cells were grown on glass coverslips until they reached 50 -70% confluence in culture medium containing 10% FCS. The cells were microinjected with phosphorothioate oligonucleotides 48 h prior to measurement of cytosolic Ca 2ϩ and then serum-starved as described. Sequence specificity for the individual G protein ␣-subunits of oligonucleotides used in this study were optimized by sequence comparison and multiple alignment using MACMOLLY TETRA software (Softgene, Berlin, FRG). Approximately 10 -20 fl of a solution containing 30 -50 M oligonucleotides dissolved in 100 mM KCl was injected into the nuclei of individual cells. Microinjection was performed with a manual injection system (Eppendorf, Hamburg, FRG) using microcapillaries pulled from borosilicate glass tubes with filament (outlet diameter of 0.5 m). To facilitate localization of injected cells, coverslips were imprinted with squares, and Texas Red-conjugated bovine serum albumin (5 mg/ml) was added to the injection solution. Only cells within the square were injected, and cells on the same coverslip outside the square were used as control cells. The following oligonucleotide sequences were used: anti-␣ q11.2s , ATGGACTCCAGAGT; anti-␣ i-common , GCCGCCTTGTCC-TG; anti-␣ i1 , TGTCCTTCCACAGTCTCTTTATG; anti-␣ i2 , ATGGTCAG-CCCAGAGCCTCCGG; anti-␣ i3 , CATCTCGCCATAAACGTTTAATC; sense-␣ q/11.2 s , ACTCTGGAGTCCAT; sense-␣ i3 , GATTAAACGTTTATG-GCGAGATG; missense-␣ q/11.2 s , TGAGACCTCAGGTA; missense-␣ i3 , CTAATTTGCAAATACCGCCGCTCTAC. For microinjection of antibodies, polyclonal antisera against G␣ q/11 (AS 370), G␣ i -common (AS 86), or G␣ z (AS 404) were diluted 1:100 in injection solution (100 mM KCl, 5 mg/ml Texas Red-conjugated bovine serum albumin) and injected into the cytoplasm of individual cells 1 h prior to measurement of [Ca 2ϩ ] i .
Immunocytochemistry-Cells grown on coverslips were washed twice with PBS and fixed with acetone for 5 min at room temperature. Cells were additionally washed three times with PBS containing 3% FCS and incubated overnight at 4°C with polyclonal rabbit antisera either against G␣ q/11 (AS 370) or against all three isoforms of G␣ i (AS 266). Both antisera were used in a dilution of 1:100 in PBS containing 3% FCS. Following four washes with PBS, cells were incubated overnight at 4°C with a fluorescein isothiocyanate-labeled goat antibody against rabbit IgG (1:1000). Cells were washed four times with PBS containing 1% Triton X-100 and mounted with Moviol (Hoechst, Frankfurt, FRG). Images of the stained cells were obtained with an inverted microscope (Axiovert 100; Zeiss, Oberkochen, FRG) connected to a digital video imaging system (T.I.L.L. Photonics, Mü nchen, FRG) using monochromatic light at 495 nm and an exposure time of 1 s. Quantitative analysis of the digitized images of the control and injected cells was performed on a Macintosh computer using NIH Image. The significance of the results was determined using a U test, errors are given as S.E., and the values are given as arbitrary units.
Measurement of Cytosolic Ca 2ϩ Concentration-For measurement of cytosolic Ca 2ϩ concentration in suspension, cells were detached from tissue culture flasks, resuspended in serum-free DMEM, and loaded with 5 M fura-2/AM for 20 min at 37°C. Fura-2-loaded cells were sedimented by centrifugation at 300 ϫ g and resuspended at a concentration of 4 ϫ 10 5 cells/ml in incubation buffer consisting of 140 mM NaCl, 5 mM KCl, 0.3 mM KH 2 PO 4 , 1 mM MgCl 2 , 5 mM glucose, 1.5 mM CaCl 2 and 10 mM HEPES/NaOH, pH 7.4. Aliquots (2 ml) of the cell suspension were transferred to an acrylic cuvette and stirred continuously, and fluorescence was recorded in a luminescence spectrometer (LS 50 B; PerkinElmer Life Sciences) at 37°C with dual wavelength excitation of 340 and 380 nm and an emission wavelength of 510 nm. Before cells were stimulated with 100 nM AVP, extracellular Ca 2ϩ was chelated by the gradual addition of 20-l aliquots of Tris/EGTA (20 mM) until a stable base line was reached. For measurement of [Ca 2ϩ ] i in single cells, cells were grown on glass coverslips and loaded with 5 M fura-2/AM for 30 min. The cells were washed twice in incubation buffer, mounted on an inverted microscope (Zeiss Axiovert 100), and washed with incubation buffer supplemented with 1 mM EGTA instead of CaCl 2 . Determination of [Ca 2ϩ ] i in individual cells was performed with an imaging system (T.I.L.L. Photonics) as described (27). [Ca 2ϩ ] i values were calculated using a formula provided by Grynkiewicz et al. (28). The ⌬[Ca 2ϩ ] i was calculated as the difference of maximum [Ca 2ϩ ] i after stimulation with AVP and basal [Ca 2ϩ ] i prior to stimulation. The significance of the results was determined using a U test, and errors are given as S.E.
Radioligand Binding Assay-Binding assays on Swiss 3T3 membranes were performed in 250 l of incubation medium (1 mg/ml bovine serum albumin, 0.5 mg/ml bacitracin, 0.05 mg/ml soy bean trypsin inhibitor, pH 7.4) containing 1-300 pM of the linear AVP antagonist 125 I-phenylacetyl-D-Tyr(Me)-Phe-Gln-Asn-Arg-Pro-Arg-Tyr-NH 2 (PerkinElmer Life Sciences). The reaction was performed in duplicate and started by the addition of membranes (5-15 g/assay), and the mixture was incubated 1 h at 20°C. The reaction was stopped by adding 4 ml of ice-cold buffer (10 mM Tris-HCl, pH 7.4, 3 mM MgCl, 0.5 mM EDTA), followed by filtration through GF/C glass microfiber filters (Whatmann, Maidstone, UK) presoaked in 0.3% polyethyleneimine (Fluka, Buchs, Switzerland). Filters were washed four times with 4 ml of ice-cold buffer, and radioactivity was counted in a ␥-counter (COBRA 5002, Packard, Meriden, MA). Nonspecific binding was determined in the presence of 3 M unlabeled AVP. Binding data were analyzed by nonlinear regression using Prism software (GraphPad Software, San Diego, CA). The significance of the results was determined using a U test, and errors are given as S.D.
Immunoblot Analysis of G Proteins-Aliquots of membrane preparations of G 0 /G 1 cells and proliferating cells (50 g each) were subjected to SDS-PAGE (12% polyacrylamide), and proteins were transferred to nitrocellulose (Hybond-C Extra; Amersham Pharmacia Biotech). Nonspecific interactions were blocked overnight at 4°C with a solution of 5% nonfat dry milk in PBS, 0.1% Tween 20. Nitrocellulose membranes were incubated either with an affinity-purified polyclonal rabbit antibody against G␣ q/11 (SC-392) or G␣ i3 (SC-262) (each 1:400; Santa Cruz Biotechnology, Inc., Santa Cruz, CA) for 1 h at room temperature, followed by a 1-h incubation with horseradish peroxidase-labeled antirabbit IgG antibody from goat (1:2500; Sigma). G proteins were detected using the ECL system (Amersham Pharmacia Biotech), and chemiluminescence signals were quantified with a bioimaging system (Fuji BAS 1000, Raytest, Mü nchen, FRG).
Immunoblot Analysis of Regulators of G Protein Signaling (RGS) Proteins-Aliquots of membrane preparations of G 0 /G 1 cells and proliferating cells (100 g each) and 100 ng of recombinantly expressed RGS5 and RGS16 were subjected to SDS-PAGE (15% polyacrylamide), and proteins were transferred to nitrocellulose (Hybond-C Extra; Amersham Pharmacia Biotech). Nonspecific interactions were blocked overnight at 4°C with a solution of 3% skimmed milk in Tris-buffered saline, 0.05% Tween 20 (TTBS). Nitrocellulose membranes were incubated either with polyclonal goat antipeptide antibody against RGS1 , with polyclonal rabbit antibody against recombinantly expressed His-tagged RGS5 (1:1000), with goat polyclonal peptide antibody against RGS7 (1:1000; SC-8139, Santa Cruz Biotechnology), or with polyclonal rabbit antibody against recombinantly expressed His-tagged RGS16 (1:1000) for 2 h at room temperature, followed by a 2-h incubation with horseradish peroxidase-labeled anti-rabbit IgG antibody (1: 2000 in TTBS; Santa Cruz Biotechnology). In the case of RGS1, an anti-goat IgG antibody was used. RGS proteins were detected using the ECL system (Amersham Pharmacia Biotech).

RESULTS
In a previous publication, the G protein coupling of the AVP receptor in Swiss 3T3 cells was studied using the technique of photoaffinity labeling of receptor-activated G proteins with aaGTP and subsequent immunoprecipitation of activated G protein ␣-subunits with subtype-specific antibodies (22). In these experiments, membranes prepared from Swiss 3T3 cells grown under regular cell culture conditions were used, i.e. with change of the medium every second day, and cells were split before confluence was reached. The addition of AVP to the Swiss 3T3 membranes induced activation of both G␣ q and G␣ 11 , but no activation of G␣ i was detected (22).
We used the same technique, but compared membranes from cells grown under various cell culture conditions. The cells were grown in the same medium until they reached confluence for several days, and then the conditioned medium was replaced for 6 h by either fresh culture medium (DMEM) containing 10% FCS ("proliferating cells") or DMEM devoid of FCS ("quiescent cells"). In membranes prepared from proliferating cells as well as in membranes prepared from quiescent cells, we confirmed the AVP-induced activation of G␣ q and G␣ 11 (Fig.  1A, lanes 1 and 2, 2470 and 222 arbitrary units, respectively, and lanes 5 and 6, 4846 and 1760 arbitrary units, respectively). In membranes prepared from quiescent cells, however, incorporation of aaGTP into G␣ q and G␣ 11 was more prominent in the absence and presence of AVP. In addition, an antibody directed against all three subtypes of the G␣ i proteins (AS266) precipitated more aaGTP-labeled ␣ i -subunit in AVP-stimulated membranes than in control membranes (Fig. 1, compare lane 7 with lane 8, 643 and 220 arbitrary units, respectively), indicating the coupling of AVP receptors to G i proteins. No activation of G i proteins by AVP could be detected in proliferating cells (see in Fig. 1, lanes 3 and 4, 557 and 547 arbitrary units, respectively).
To address the question whether the AVP-induced activation of G i proteins in membranes prepared from quiescent Swiss 3T3 cells leads to a functional coupling to an effector system, the AVP-induced increase in [Ca 2ϩ ] i was measured in Swiss 3T3 cell suspensions. The AVP-induced increase of [Ca 2ϩ ] i was completely abolished by preincubation of the cells with a linear peptide antagonist specific for the V 1a receptor [1-(␤-mercapto-␤,␤-cyclopenthamethylene propionic acid),O-Me-Tyr 2 ,Arg 8 ]vasopressin (7,31), indicating that the activation of PLC-␤ and subsequent increase in [Ca 2ϩ ] i in Swiss 3T3 cells is mediated solely by the V 1a receptor subtype (data not shown). Fig. 2 shows the mean values of the amplitude of the AVP-induced increase in [Ca 2ϩ ] i (⌬[Ca 2ϩ ] i ) of 3-7 experiments with Swiss 3T3 cells cultured under the described cell culture conditions and pretreated with or without PTX (100 ng/ml for 24 h). In quiescent cells, AVP induced a significant (p Ͻ 0.05) higher Ca 2ϩ amplitude than in proliferating cells (Fig. 2, open bars). The basal calcium concentrations were not significantly different under both conditions (data not shown). In proliferating cells, the AVP-induced increase in [Ca 2ϩ ] i was not inhibited by preincubation of the cells with PTX (see Fig. 2). In contrast, To confirm that the dual coupling of the V 1a receptor to G proteins of the G q/11 and the G i subfamilies is a characteristic of the quiescent state of the cells, we analyzed the AVP-induced increase in [Ca 2ϩ ] i in various phases of the cell cycle. To clearly define the quiescent state of cell growth as G 0 /G 1 phase of the cell cycle in (in the following referred to as G 0 /G 1 cells) we used defined protocols for synchronization of proliferating, subconfluent cells. For arresting the cell growth in the G 0 /G 1 phase, subconfluent cells were serum-starved for 8 h, and then the serum-free medium was replaced by serum-containing medium. Synchrony of cell growth in the G 1 /S phase of the cell cycle was reached by preincubation of the cells with aphidicolin and in G 2 /M phase by using nocodazole. The phases of the cell cycle were analyzed by staining the DNA with propidium iodide and measurement of the DNA content by flow cytometric analysis (Fig. 3A). Control cells cultured in medium containing FCS, which was changed every second day, were found to proliferate desynchronously. Sixty-two percent of the cells were in the G 0 /G 1 phase, 26% in the S phase, and 12% in the G 2 /M phase (see Fig. 3A, a). From the cells that were serum-starved (DMEM without FCS for 12 h and subsequent incubation in medium with 10% FCS for 16 h), 92% were found in the G 0 /G 1 phase, 1% in the S phase, and 4% in the G 2 /M phase (see Fig.  3A, b). From the cells treated with aphidicolin (4 g/ml for 12 h), 75% were found in the G 0 /G 1 phase, 24% in the S phase, and 0.3% in the G 2 /M phase (see Fig. 3A, c). From the cells arrested with nocodazole (500 ng/ml for 12 h), 6% were found in the G 0 /G 1 phase, 5% in the S phase, and 88% in the G 2 /M phase (see Fig. 3A, d). Fig. 3B shows the AVP-induced increase in [Ca 2ϩ ] i (mean results of 7-10 experiments) in cell suspensions from Swiss 3T3 cells not arrested or arrested in the various phases of the cell cycle as indicated in Fig. 3A. As previously observed in confluent cells, in cells preincubated in serum-free DMEM (G 0 /G 1 cells) the AVP-induced increase in [Ca 2ϩ ] i was significantly (p Ͻ 0.05) higher than in proliferating cells. The basal calcium concentrations were the same (data not shown). Only in cells arrested by serum starvation in the G 0 /G 1 phase was the AVP-induced increase in [Ca 2ϩ ] i significantly reduced by pretreating the cells with PTX (see Fig. 3B). Therefore, the partial PTX sensitivity of the AVP-induced increase in [Ca 2ϩ ] i can be attributed to cells growth arrested in the G 0 /G 1 phase of the cell cycle and not to cells in the late G 1 /S or G 2 /M phase of the cell cycle.
To confirm the results of the PTX experiments with additional methods, we first studied whether the AVP-induced increase in [Ca 2ϩ ] i in G 0 /G 1 cells can be inhibited by microinjection of antibodies directed against the C-terminal domain of the G␣ i proteins, which prevent the interaction of the G protein with ligand-activated receptors. Cells were plated on coverslips and serum-starved for growth arrest in the G 0 /G 1 phase of the cell cycle as described. Proliferating cells served as control. Cells were microinjected with an antibody binding to the Cterminal domain of all G␣ i subunits (AS 86) or antibodies binding to the C termini of G␣ q and G␣ 11 (AS 370). Control cells, on the same coverslip, were microinjected with an antibody against G␣ z (AS 404). The G 0 /G 1 cells were compared with proliferating cells grown on separate coverslips. The AVP-induced increase in [Ca 2ϩ ] i was determined in the antibodies injected cells by using a single cell calcium-imaging system (Fig. 4). As in the previous experiments, the AVP-induced increase in [Ca 2ϩ ] i was significantly higher in G 0 /G 1 cells than in proliferating cells. Microinjection of the G␣ q/11 antibody AS 370 resulted in a partial but significant inhibition of the AVPinduced increase in [Ca 2ϩ ] i in proliferating cells as well as in G 0 /G 1 cells (see Fig. 4, dark bars). In contrast, microinjection of the G␣ i antibody AS 86 resulted in a partial inhibition of AVP-induced increase in [Ca 2ϩ ] i only in G 0 /G 1 cells, but not in proliferating cells (see Fig. 4, hatched bars). Microinjection of the G␣ z antibody AS 404 had no effect compared with noninjected cells (see Fig. 4, open bars, and data not shown).
The expression of G␣ i2 and G␣ i3 had been shown in NIH-3T3 fibroblasts (32) and in membrane preparations derived from rat fetus (33), but in both cases no G␣ i1 expression could be detected. Since the antibodies used for immunoprecipitation or for microinjection cannot discriminate between the two members of the G␣ i isoforms, G␣ i2 and G␣ i3 , we assessed which of two G␣ i isoforms are involved in the AVP-induced Ca 2ϩ release in Swiss 3T3 cells by microinjection of antisense oligonucleotides (for a review, see Ref. 34). To control whether the protein expression of G␣ i and G␣ q/11 can be inhibited by microinjection of antisense oligonucleotides into Swiss 3T3 cells, antisense oligonucleotides directed against the mRNA of G␣ q/11 (anti-␣ q/11.2s ) or all isoforms of G␣ i (anti-␣ i-common ) were used. Fortyeight h after microinjection, the G␣ q/11 or G␣ i proteins were determined semiquantitatively by indirect immunofluorescence (Fig. 5). A decrease in the expression of the respective G protein ␣-subunits was detected in proliferating cells as well as in G 0 /G 1 cells. The protein expression of the ␣ i -subunits was reduced by about 35% in proliferating cells injected with anti-␣ i-common compared with noninjected control cells (see Fig. 5A, upper panel, 127 Ϯ 9, n ϭ 12, and 82 Ϯ 5, n ϭ 8, arbitrary units, respectively, p Ͻ 0.01), whereas in G 0 /G 1 cells the ␣ i -subunit protein expression was reduced by about 50% (see Fig. 5A, lower panel, 110 Ϯ 12, n ϭ 12, and 59 Ϯ 3, n ϭ 11, respectively, p Ͻ 0.025). Injection of anti-␣ q/11.2s oligonucleotides into proliferating cells led to a reduction of the protein expression of the ␣ q/11 -subunits by about 40% compared with noninjected control cells (see Fig. 5B, upper panel, 79 Ϯ 5, n ϭ 16, and 46 Ϯ 9, n ϭ 7, arbitrary units, respectively, p Ͻ 0.005), whereas in G 0 /G 1 cells injected with anti-␣ q/11.2s oligonucleotides the ␣ q/11 -subunit expression was reduced by about 65% (see Fig. 5B, lower panel, 151 Ϯ 12, n ϭ 10, and 54 Ϯ 5, n ϭ 10, arbitrary units, respectively, p Ͻ 0.001).
Next, we determined the AVP-induced increase in [Ca 2ϩ ] i in cells microinjected with anti-␣ q/11.2s and anti-␣ i-common oligonucleotides (Fig. 6). Like microinjection of antibodies directed against G␣ q/11 , microinjection of anti-␣ q/11.2s into proliferating cells resulted in a weak inhibition of the AVP-induced increase in [Ca 2ϩ ] i (see Fig. 6A, left panel, open bars). In G 0 /G 1 cells, the effect of the anti-␣ q/11.2s oligonucleotides was more pronounced (43% inhibition of the AVP-induced increase in [Ca 2ϩ ] i ) (Fig.  6A, right panel, open bars). Preincubation with PTX had no effect in proliferating cells (see Fig. 6A, left panel, filled bars). In G 0 /G 1 cells, the AVP-induced increase in [Ca 2ϩ ] i was inhibited by about 50% upon pretreatment of the cells (see Fig. 6A, compare open to filled bars). Microinjection of anti-␣ q/11.2s in PTX-treated G 0 /G 1 cells resulted in about 75% inhibition of the AVP-induced increase in [Ca 2ϩ ] i compared with noninjected and non-PTX-treated cells (see Fig. 6A, right panel, compare the filled hatched bar to the open bar). This indicates an additive effect of G␣ q/11 and G␣ i in AVP-induced Ca 2ϩ signaling in G 0 /G 1 Swiss 3T3 cells. Neither injection of anti-␣ i-common nor preincubation with PTX had an effect in proliferating cells (see Fig. 6B, left panel). In contrast, anti-␣ i-common injection as well as PTX treatment resulted in a similar inhibition of AVP-induced increase in [Ca 2ϩ ] i in G 0 /G 1 cells (Fig. 6B, middle panel). Injection of a combination of anti-␣ q/11.2s and anti-␣ i-common caused an additive inhibition (about 80%) of the AVP-induced increase in [Ca 2ϩ ] i (see Fig. 6B, right panel) as demonstrated before with the combination of microinjection of anti-␣ q/11.2s and pretreatment of the cells with PTX (see Fig.  6A, right panel). The effect of the used antisense oligonucleotides was sequence-specific. Microinjection of oligonucleotides with sense and missense sequences corresponding to the anti- Therefore, we used this technique to address the question which of the three isoforms of G␣ i are involved in the PTXsensitive coupling of the AVP receptor to calcium release in the Swiss 3T3 cells. Specific antisense oligonucleotides directed against the mRNA of the individual isoforms of G␣ i (anti-␣ i1-3 ) were injected, and after 48 h, the AVP-induced increase in [Ca 2ϩ ] i was determined. Anti-␣ i3 , but not anti-␣ i1 or anti-␣ i2 , inhibited the AVP-induced increase in [Ca 2ϩ ] i (Fig. 7). Again, the sequence specificity of the effect of anti-␣ i3 was confirmed by microinjection of oligonucleotides with the corresponding sense or missense sequences, which showed no effects (data not shown).
Finally, we addressed the question of which mechanism may be responsible for this additional coupling of the V 1a receptor to another G protein besides G␣ q/11 . One mechanism reported to facilitate the dual coupling of a GPCR to G proteins is the altered balance in the expression of receptors or G proteins. For this reason, we studied the V 1a receptor density and the protein expression of G␣ q/11 or G␣ i3 in proliferating and G 0 /G 1 Swiss 3T3 cells. Cell membranes were prepared from both cell type and used either for determination of radioligand binding of an iodinated AVP antagonist or for immunoblotting of G␣ q/11 or G␣ i3 protein. Using the linear V 1a receptor antagonist, 125 Iphenylacetyl-D-Tyr(Me)-Phe-Gln-Asn-Arg-Pro-Arg-Tyr-NH 2 , maximal binding and affinity were determined in both membrane preparations. The K D values obtained in membranes of G 0 /G 1 cells and proliferating cells were identical. The number of binding sites, however, was reduced by 30% in membranes from G 0 /G 1 cells compared with proliferating cells (Table I). G␣ i3 protein expression was similar in proliferating or G 0 /G 1 cells. In contrast, the expression of G␣ q/11 protein was about 2-fold higher in G 0 /G 1 cells compared with proliferating cells (Fig. 8A).
The activity of G proteins is not only determined by the rate of receptor-initiated GTP exchange but also by the rate of GTP hydrolysis. Therefore, we additionally determined the expression of members of a recently discovered family of GTPaseactivating proteins termed RGS. We did not detect RGS1, -2, -4, and -7 in Swiss 3T3 cell membranes (data not shown), but expression of RGS3, -5, and -16 was verified (Fig. 8B). The expression of RGS3 was similar in G 0 /G 1 and proliferating cells. In contrast, the expression of RGS5 and RGS16 was increased in G 0 /G 1 cells compared with proliferating cells, paralleling the increased expression of G q/11 (see Fig. 8B).

DISCUSSION
Although the ability of G proteins to stimulate cell proliferation is well known, little work has been done on GPCR function during different phases of the cell cycle. A cell cycle-dependent switch of the receptor-G protein coupling from G s to G i protein, and a subsequent switch from the adenylyl cyclase to the Ca 2ϩ /protein kinase C signal pathway was reported for the calcitonin receptor in a pig kidney cell line (35). For the parathyroid hormone receptor, an additional coupling to G i proteins and a subsequent increase in Ca 2ϩ influx depending on cell cycle was shown in osteogenic sarcoma cells (36). Recently, a switch of the ␤ 2 -adrenoreceptor in coupling from G s to G i was shown to be related to phosphorylation of the receptor by pro-tein kinase A (37).

FIG. 6. AVP-induced increase in [Ca 2؉ ] i in Swiss 3T3 cells cultured under various conditions or injected with antisense oligonucleotides directed against the mRNAs of G protein
Here we describe an enhanced AVP-induced Ca 2ϩ response in quiescent Swiss 3T3 cells. A similar effect was observed by others who found in quiescent Swiss 3T3 cells an increased Ca 2ϩ mobilization and phosphorylation of proteins induced by bradykinin (38). Part of the enhanced AVP response was blocked by preincubation of the cells with PTX, suggesting an additional recruitment of a G i proteins by AVP receptors (see Fig. 2). The additional coupling of vasopressin receptors to a G i protein was confirmed by microinjection of an anti-␣ icom antiserum and anti-␣ i-common antisense oligonucleotides (see Figs. 4 and 6). In addition, AVP-induced activation of G i could be detected by the photoaffinity labeling technique in cell membranes (see Fig. 1). By microinjection of subtype-specific antisense oligonucleotides against G␣ i1 , G␣ i2 , and G␣ i3 , we identified the G protein ␣-subunit involved in the PTX-sensitive Ca 2ϩ response as G␣ i3 (see Fig. 7).
The coupling of vasopressin receptors to the G i3 was absent when the quiescent cells were stimulated for proliferation (see Figs. 1 and 2). In order to attribute the coupling of the AVP receptors to G i3 to a definite cell cycle state, we synchronized the cell proliferation in different phases of the cell cycle (see Fig. 4A). Although 62% of the proliferating cells and 75% of the aphidicolin-treated cells were found to be in the G 0 /G 1 phase, only in growth-arrested cells (92% in the G 0 /G 1 phase), an increased, partially PTX-sensitive calcium response induced by AVP was detected (see Fig. 4B). By the methodology used, we could not, however, distinguish what percentage of the cells were in the G 0 or the G 1 phase of the cell cycle. Thus, a feasible explanation for this discrepancy might be that in the growtharrested cells, most of these cells were in the G 0 phase. In contrast, the percentage of G 0 /G 1 cells in proliferating and aphidicolin-treated cells might represent mainly cells in the G 1 phase. Therefore, the PTX-sensitive AVP response might only occur in the G 0 phase.
Having established the additional coupling of AVP receptors to G i3 in quiescent Swiss 3T3 fibroblasts, we also addressed the question how this coupling might be achieved on the molecular level. The most simple explanation would be the presence of two different receptor subtypes in the quiescent Swiss 3T3 cells. Indeed, the mRNA for the V 1b receptor was found in some nonneuronal tissues (39), but binding studies, using peptide and nonpeptide ligands specific for the rat and the human V 1a receptor, demonstrated only one binding site specific for V 1a receptor ligands in Swiss 3T3 cells (21). Nevertheless, we had to exclude the possibility that in our Swiss 3T3 cells the partial PTX sensitivity of the AVP response was based on expression of a vasopressin receptor subtype other than the V 1a subtype. Our results obtained with the V 1a receptor-specific linear peptide receptor antagonist [1-(␤-mercapto-␤,␤-cyclopenthamethylene propionic acid),O-Me-Tyr 2 ,Arg 8 ]vasopressin show that only the V 1a receptor subtype is involved in the functional coupling of the AVP receptor to the Ca 2ϩ response in these cells in all phases of the cell cycle.
A second possibility would be an alteration in the receptor/G protein stochiometry. Recombinantly expressed V 1b receptor has been reported to couple solely to members of the G q family in Chinese hamster ovary cells when expressed at low levels. At higher expression levels, G i and G s were additionally recruited by V 1b receptors (40). A similar effect has been shown for some primarily G s coupled receptors (e.g. LH, V 2 , and ␤-adrenergic receptors). Dual coupling of these receptors to adenylyl cyclase and PLC-␤ also occurs with increasing numbers of expressed receptors (41). Note here that in these studies the numbers of receptors varied by a factor of 5 or more. In our hands, the number of the AVP-binding sites in G 0 /G 1 cells was reduced to  8. Expression pattern of G protein ␣-subunits and RGS proteins in G 0 /G 1 and proliferating Swiss 3T3 cells. Membrane proteins of the proliferating Swiss 3T3 cells, cells synchronized in the G 0 /G 1 phase of the cell cycle, or recombinant RGS5 and RGS16 were separated by SDS-PAGE and blotted to nitrocellulose membranes. For detection of the G protein ␣-subunits, 50 g of membrane protein were loaded per lane, and the proteins were separated in a 12% SDS gel. For detection of the RGS proteins, 100 g of membrane protein or 100 ng of recombinant protein were loaded per lane and the proteins separated in a 15% SDS gel. The bound specific antibody were detected with a second antibody coupled to horseradish peroxidase and by using a chemiluminescence detection system. A, nitrocellulose membranes were incubated with antibodies specific for G␣ q/11 or for G␣ i3 subunits as indicated at the right. B, the nitrocellulose membranes were incubated with antibodies specific for RGS3, -5, and -16 as indicated at the right. Numbers on the left indicate molecular masses of marker proteins. 70% compared with the binding sites in proliferating cells (see Table I). Although we cannot exclude this possibility, it is rather unlikely that this small decrease is responsible for the additional coupling of the V 1a receptors.
The balance of the receptor and the adjacent signal transduction molecules is also determined by the expression levels of the involved G proteins. Thus, an increased expression of G i3 protein would be another possible scenario. We could not detect a change in protein expression of G␣ i3 . On the contrary, we found a 2-fold increased expression of G␣ q/11 protein in G 0 /G 1 cells (see Fig. 8A). This increase was accompanied by a higher binding of the nonhydrolyzable GTP analog aaGTP in the absence and presence of AVP (see Fig. 1). Nevertheless, the higher expression and activation of G␣ q/11 did not result in an enhanced PTX-insensitive Ca 2ϩ -response. The PTX-insensitive ⌬[Ca 2ϩ ] i was even reduced by about 35% in quiescent cells compared with the proliferating controls (see Figs. 2 and 3). This discrepancy might be explained by the higher expression of RGS5 and RGS16 found in the quiescent cells. Both RGS proteins catalytically enhance the GTPase activity of G i as well as G q/11 proteins and, thus, negatively regulate signals mediated by these G proteins (for reviews, see Refs. [42][43][44][45]. In addition, RGS16 has been shown to regulate G q/11 -mediated signals in a receptor-selective manner (46,47). Therefore, it is feasible that the up-regulation of RGS5 and -16 is responsible for the diminished G q/11 -mediated Ca 2ϩ responses, although the expression and activation of G␣ q/11 is enhanced in quiescent cells.
Interestingly, the tumor suppressor p53, a critical regulator of growth arrest, induces the expression of RGS16 (48). Therefore, the question arises whether the up-regulation of the RGS proteins is involved in the additional coupling of the V 1a receptor to G␣ i3 in the quiescent cells. As already mentioned above and shown for RGS16 in the study of Buckbinder et al. (48), these RGS proteins negatively regulate G i -mediated pathways. We cannot estimate the impact of the up-regulation of RGS5 and -16 on G i3 -mediated signaling, since we have no PTXsensitive signal in the proliferating cells. Nevertheless, it is difficult to imagine how up-regulation of per se negative regulators of G i -mediated signals can cause an additional coupling of the V 1a receptor to G i3 .
One model of receptor/G protein interaction is based on preformed complexes of receptors, G proteins, and effectors preformed in microdomains at the plasma membrane as a possible mechanism of coupling specificity (49). A co-localization of the V 1a receptor with G␣ q/11 and G␣ i3 was described in hepatocytes (50). The AVP-induced DNA synthesis in Swiss 3T3 cells was reported to be partially sensitive to PTX (24), suggesting that in these cells the V 1a receptor couples to G proteins of the G q and the G i family. Eventually, the here described additional recruitment of G i3 by the V 1a receptor might represent an activation of such a preformed signal transduction complex in Swiss 3T3 cells. The changes in receptor numbers, which we found in G 0 /G 1 cells compared with proliferating cells, are very small, but the amount of expressed membrane proteins does not necessarily reflect the availability of these proteins in microdomains in the membranes in which the signal transduction complexes assemble. Therefore, it is possible that even small changes in the expression of endogenous receptors might have functional consequences. Therefore, one can anticipate a model that includes two functional classes of vasopressin V 1a receptors in Swiss 3T3 cells. The first is precoupled with G q/11 proteins and responsible for the AVP response in proliferating cells and partially in quiescent cells. The second class precoupled with G i3 proteins can only be recruited for the AVP response in quiescent Swiss 3T3 cells. We and others (22) were not able to detect any activation of G i proteins in proliferating cells by the sensitive method of photoaffinity labeling with aaGTP (see Fig. 1). Thus, when the Swiss 3T3 cells are stimulated to proliferate, this subpopulation of V 1a receptors is obviously desensitized on the level of receptor-G protein interaction or even removed from the cell surface. Desensitization of the V 1a receptor is accompanied by phosphorylation of the receptor molecule (51), but the functional consequence of this event remains to be elucidated. We did not study the phosphorylation state of the V 1a receptor in Swiss 3T3 cells in the various phases of the cell cycle. Therefore, we can only speculate on such a mechanism.
Taken together, our data provide evidence for a dual coupling of the V 1a receptor to PLC-␤ isoforms dependent on the proliferation state of the cells. The main axis of PLC-␤ activation is mediated by G q/11 . In the G 0 /G 1 phase of the cell cycle, which is physiologically the resting state of differentiated tissue, there is an additional coupling of the V 1a receptor to G i3 , leading to activation of PLC-␤ isozymes by ␤␥ subunits released upon activation. While AVP acts as mitogenic signal for Swiss 3T3 cells (52), activation of PLC-␤ is not sufficient for stimulation of DNA synthesis (53), and AVP-induced DNA synthesis was reported to be partially sensitive to preincubation of cells with PTX (24). Considering that many mitogenic signals were found to be dependent on PTX-sensitive G proteins, our results may indicate additional mechanisms, such as enhanced calcium release, ␤␥ (derived from G i3 )-mediated activation of MAP kinases, or phosphoinositide-3-kinase, to be involved in mitogenic signaling of the V 1a receptor.