The M3 muscarinic acetylcholine receptor expressed in HEK-293 cells signals to phospholipase D via G12 but not Gq-type G proteins: regulators of G proteins as tools to dissect pertussis toxin-resistant G proteins in receptor-effector coupling.

The M(3) muscarinic acetylcholine receptor (mAChR) expressed in HEK-293 cells couples to G(q) and G(12) proteins and stimulates phospholipase C (PLC) and phospholipase D (PLD) in a pertussis toxin-insensitive manner. To determine the type of G protein mediating M(3) mAChR-PLD coupling in comparison to M(3) mAChR-PLC coupling, we expressed various Galpha proteins and regulators of the G protein signaling (RGS), which act as GTPase-activating proteins for G(q)- or G(12)-type G proteins. PLD stimulation by the M(3) mAChR was enhanced by the overexpression of Galpha(12) and Galpha(13), whereas the overexpression of Galpha(q) strongly increased PLC activity without affecting PLD activity. Expression of the RGS homology domain of Lsc, which acts specifically on Galpha(12) and Galpha(13), blunted the M(3) mAChR-induced PLD stimulation without affecting PLC stimulation. On the other hand, overexpression of RGS4, which acts on Galpha(q)- but not Galpha(12)-type G proteins, suppressed the M(3) mAChR-induced PLC stimulation without altering PLD stimulation. We conclude that the M(3) mAChR in HEK-293 cells apparently signals to PLD via G(12)- but not G(q)-type G proteins and that G protein subtype-selective RGS proteins can be used as powerful tools to dissect the pertussis toxin-resistant G proteins and their role in receptor-effector coupling.

Stimulation of phosphatidylcholine-specific phospholipase D (PLD), 1 leading to the formation of the putative second messenger phosphatidic acid, has been described in a wide range of cell types in response to stimulation of a large variety of different membrane receptors (1)(2)(3)(4)(5). Particularly, numerous receptors coupled to heterotrimeric G proteins have been shown to cause PLD stimulation. Interestingly, almost every G pro-tein-coupled receptor (GPCR) known to stimulate phosphoinositide-specific phospholipase C (PLC) also stimulates PLD activity. This concomitant activation of the two phospholipases may be caused at several levels of signal transduction by these GPCRs.
First, stimulation of PLD activity may be secondary to PLC stimulation, and the cellular signals generated by the PLC reaction specifically increase in the cytosolic Ca 2ϩ concentration and activation of protein kinase C isoforms. In fact, evidence has been provided for some GPCRs in different cell types that PLD stimulation is apparently a consequence of the primary PLC stimulation (1)(2)(3).
Second, the same receptor-activated G protein may stimulate both PLC and PLD. There is ample evidence that one type of G protein can regulate different effectors (6,7). Two distinct G protein subtypes mediate GPCR-PLC coupling, the pertussis toxin (PTX)-insensitive G q -type G proteins and the PTX-sensitive G i -type G proteins (2,8). Studies on stimulation of PLD activity by chemoattractants in neutrophils suggest that in these cells GPCRs couple to PLD and PLC via the same G i -type G proteins (4,9). On the other hand, PLC-independent coupling of GPCRs to PLD via G q -type G proteins has not yet been reported. However, it has recently been reported that the expression of constitutively active G q can strongly increase PLD activity (10), suggesting that this possibility should be considered.
Third, the receptor may couple to PLC and PLD by activating two distinct types of heterotrimeric G proteins. For example, GPCR-PLC coupling may be mediated by G q proteins and GPCR-PLD coupling may be mediated by G i proteins or vice versa. Other potential candidates for mediating specific GPCR-PLD coupling are the PTX-insensitive G 12 -type G proteins G 12 and G 13 , which are not directly involved in GPCR-PLC coupling (2,11,12). Expression of constitutively active G␣ 12 and G␣ 13 has been shown to strongly increase the activity of a coexpressed PLD1 enzyme like the constitutively active G␣ q (10).
The M 3 muscarinic acetylcholine receptor (mAChR) stably expressed in HEK-293 cells is coupled to both PLD and PLC (13), and stimulation of either phospholipase by this GPCR is resistant to the treatment of the cells with PTX (14), which indicates that G i -type G proteins are not involved in coupling M 3 mAChR to either PLC or PLD. Previous studies furthermore demonstrate that the M 3 mAChR-induced PLD stimulation is not affected by protein kinase C inhibition or downregulation or by chelation of intracellular Ca 2ϩ (15,16), suggesting that it is not secondary to PLC stimulation. Thus, because the M 3 mAChR coupled to G q -type G proteins in these cells (14), we had to consider that the receptor coupled to PLD and PLC via the same type of G proteins, i.e. G q proteins, or that two distinct PTX-insensitive G proteins mediate GPCR coupling to the two phospholipases.
To resolve this question, we took advantage of the regulators of G protein signaling (RGS). These are GTPase-activating proteins (GAPs) that accelerate GTP hydrolysis by G␣ proteins and thereby attenuate signaling via heterotrimeric G proteins (17)(18)(19). More than 20 different RGS proteins have been identified, of which RGS4 has received the most extensive biochemical characterization (20 -24). RGS4 exhibits GAP activity for G␣ i -and G␣ q -type G proteins but not for G␣ 12 -type proteins (21). Furthermore, structural data suggest that an interaction of either G␣ 12 or G␣ 13 with RGS4 is unlikely (23). Recently, an NH 2 -terminal RGS homology domain has been identified in the Rho-specific guanine nucleotide exchange factor (GEF), p115 RhoGEF, its murine homolog Lsc, and some closely related RhoGEFs (25). This RGS homology domain exhibited GAP activity for G␣ 12 and G␣ 13 but not for other G␣ proteins. Thus, the use of the RGS homology domain of these RhoGEFs inactivating G 12 -type G proteins in comparison to RGS4 inactivating G q -proteins should help to clarify whether G q -and/or G 12type G proteins mediate M 3 mAChR-PLD coupling in HEK-293 cells. Using this approach in combination with the overexpression of relevant G␣ proteins, strong evidence is provided that G 12 but not G q proteins mediate M 3 mAChR-PLD coupling.

EXPERIMENTAL PROCEDURES
Expression Plasmids-The pCis vectors carrying cDNAs for wildtype G␣ q , G␣ 12 , G␣ 13 , and the constitutively active G␣ q mutant, G␣ q R183C, were described previously (26,27). FLAG-tagged RGS4 in a pCMV-based expression vector and the anti-RGS4 antibody were kind gifts of Dr. J. H. Kehrl. cDNA encoding Lsc was donated by Dr. R. Kay. For the expression of a Myc-tagged variant of Lsc containing the amino terminus with the RGS homology domain but lacking the Dbl and pleckstrin homology domains (Lsc-RGS, amino acids 1-283), the corresponding cDNA fragment was subcloned into pCMV3-Tag3 expression vector (Stratagene).
Cell Culture and Transient Transfection-HEK-293 cells stably expressing the M 3 mAChR were cultured as reported previously (15). For experiments, cells were grown to about 50% confluence on 145-mm culture dishes and transfected with the indicated amounts of either plasmid DNA or empty expression vectors using the calcium phosphate precipitation method. As studied by the expression of the green fluorescent protein, transfection efficiency ranged from 50 to 70%. All assays were performed 48 h after transfection. Expression of the proteins was verified by the immunoblotting of cell lysates with specific antibodies. The antibodies against G␣ q , G␣ 12 , G␣ 13 , and His 6 were from Santa Cruz, the anti-Myc antibody was from Roche Molecular Biochemicals, and the anti-FLAG antibody was from Stratagene.
Construction of and Infection with Recombinant Adenoviruses-For construction of the recombinant adenoviruses encoding human RGS4 or mouse Lsc-RGS, the Adeasy system kindly provided by Dr. B. Vogelstein was used (28). The constructs were transfected into HEK-293 cells using LipofectAMINE, and recombinant viruses were amplified in several steps of this cell line. After purification over a CsCl 2 gradient, high titer viral stocks of 3-4 ϫ 10 8 plaque-forming units/l were obtained. Infection of subconfluent monolayers of HEK-293 cells stably expressing the M 3 mAChR was performed at a multiplicity of infection of 30 for 24 h before labeling. Before the PLC and PLD assays, infected cells were visualized and quantified by fluorescence microscopy as the constructs additionally expressed green fluorescent protein under the control of an independent CMV promoter. For control, an adenovirus encoding bacterial ␤-galactosidase LacZ, a kind gift of Dr. T. Eschenhagen, was used.
Photoaffinity Labeling of G Proteins with [␥-32 P]GTP-azidoanilide-Synthesis and purification of [␥-32 P]GTP-azidoanilide were performed as described previously (29). G proteins in the membranes of HEK-293 cells (100 -200 g of protein) were photoaffinity-labeled as described (29) with the following modifications: the reaction mixture (50 l) contained 60 mM HEPES, pH 7.4, 60 mM NaCl, 10 mM MgCl 2 , 0.2 mM EDTA, and 2 Ci of [␥-32 P]GTP-azidoanilide. Incubation was for 5 min at 30°C in the absence and presence of 1 mM carbachol. Solubilization of irradiated membranes and immunoprecipitation of G␣ q and G␣ 12 were exactly as described previously (30) using the antibodies AS 370 (5 l/tube) and AS 233 (10 l/tube), respectively (kindly provided by Dr. S. Offermanns). Immunoprecipitates were eluted from protein A-Sepharose with sample buffer and resolved by SDS-polyacrylamide gel electrophoresis (9% acrylamide plus 6 M urea). Gels were stained with Coomassie Brilliant Blue and dried. Labeled proteins were visualized by phosphoimaging (PhosphorImager, Molecular Dynamics) (29).
Assays of PLD and PLC Activities-For measurement of PLD and PLC activities, the transfected cells were replated 24 h after transfection on 145-mm culture dishes. Cellular phospholipids were labeled by incubation for 20 -24 h with [ 3 H]oleic acid (2 Ci/ml) and myo-[ 3 H]inositol (1 Ci/ml) in inositol-free growth medium. Thereafter, the cells were detached from the dishes, washed once with Hanks' balanced salt solution (118 mM NaCl, 5 mM KCl, 1 mM CaCl 2 , 1 mM MgCl 2 , and 5 mM D-glucose buffered at pH 7.4 with 15 mM HEPES), supplemented with 10 mM LiCl, and resuspended at a density of 1 ϫ 10 7 cells/ml. Then phospholipase activities were measured for 30 min at 37°C in a total volume of 200 l containing 100 l of the cell suspension (1 ϫ 10 6 cells), 1.75% ethanol, and the indicated stimulatory agents. Alternatively, the transfected cells were split to poly-L-lysine-coated 35-mm dishes, and after labeling PLD and PLC activities were determined in adherent cells (15). Stop of the enzyme reactions and analysis of [ 3 H]inositol phosphates and labeled phospholipids including the specific PLD product [ 3 H]phosphatidylethanol were carried out as described previously (15). Protein levels were measured by the Bradford method in separate culture dishes. The formation of [ 3 H]phosphatidylethanol was expressed as the percentage of total labeled phospholipids. The formation of [ 3 H]inositol phosphates was given as counts/min per 10 6 cells or per mg of protein. Similar results were obtained whether the cells were in suspension or were adherent.
Data Presentation-Data shown are means Ϯ S.D. from one representative experiment performed in triplicate and repeated as indicated or means Ϯ S.E. with n providing the number of independent experiments. Concentration response curves were analyzed using iterative nonlinear regression analysis (GraphPAD Prism, GraphPAD Software for Science, San Diego, CA).

Effects of Various G␣ Proteins on PLD and PLC Activities-
Agonist activation of the M 3 mAChR stably expressed in HEK-293 cells results in rapid and strong stimulation of PLD and PLC activities that are insensitive to PTX treatment (13)(14)(15). As a first approach to determine the type of G proteins mediating M 3 mAChR-PLD coupling in comparison to M 3 mAChR-PLC coupling, HEK-293 cells were transiently transfected with expression vectors for wild-type G␣ 12 , G␣ 13 , and G␣ q as well as the constitutively active G␣ q R183C. In cells overexpressing either G␣ 12 or G␣ 13 , the stimulation of PLD activity by carbachol (1 mM) was increased by 60 -100% compared with vectortransfected control cells (Fig. 1A), whereas the overexpression of G␣ q did not alter PLD activity (Fig. 1B). In contrast to M 3 mAChR-mediated PLD stimulation, basal PLD activity and PLD stimulation by 100 nM phorbol 12-myristate 13-acetate (PMA) were not affected by the overexpression of G␣ 12 or G␣ 13 (Fig. 2). On the other hand, basal and carbachol-stimulated PLC activities were not altered by the overexpression of G␣ 12 or G␣ 13 (Fig. 3A) but were enhanced by 3-4-fold in cells overexpressing G␣ q (Fig. 3B). Expression of the constitutively active G␣ q R183C caused an even larger increase in basal PLC activity, which was not further enhanced by carbachol (Fig.  4A). However, despite this marked PLC stimulation, neither basal nor carbachol-stimulated PLD activities were altered by the expression of G␣ q R183C (Fig. 4B).
To investigate whether the M 3 mAChR is able to activate G qand G 12 -type G proteins, the incorporation of the photoreactive GTP analog [␥-32 P]GTP-azidoanilide into G␣ q and G␣ 12 proteins overexpressed in HEK-293 cells was studied. In line with previous findings (14), the addition of carbachol (1 mM) strongly increased the incorporation of [␥-32 P]GTP-azidoanilide into G␣ q proteins (Fig. 5A). Binding of the GTP analog to G␣ 12 was also, but less efficiently, enhanced by carbachol (Fig. 5B). Thus, the M 3 mAChR stably expressed in HEK-293 cells can activate both G q -and G 12 -type G proteins.
Effects of RGS Proteins on M 3 mAChR-induced PLD and PLC Stimulation-To determine the endogenous G protein subtype involved in the coupling of M 3 mAChR to the stimulation of PLD and PLC, we made use of the two RGS proteins, RGS4 and the RGS homology domain of Lsc (Lsc-RGS), which act as GAPs for G␣ q and G␣ 12 family members, respectively (17)(18)(19)(20)(21)(22)(23)(24)(25). As shown in Fig. 6A, M 3 mAChR-stimulated PLD activity was strongly reduced (by about 50%) in cells transiently expressing Lsc-RGS, whereas the expression of RGS4 was without effect. On the other hand, the expression of RGS4 markedly reduced (by about 50%) 1 mM carbachol-stimulated PLC activity, whereas the expression of Lsc-RGS was without effect (Fig. 6B). In contrast to M 3 mAChR-mediated PLD stimulation, expression of Lsc-RGS did not alter 100 nM PMAinduced PLD stimulation (Fig. 7).
The extent of inhibition of carbachol-stimulated PLD and PLC activities observed upon the transient expression of Lsc-RGS and RGS4, respectively, roughly reflects the transfection efficiency. However, a small effect of Lsc-RGS and RGS4 on PLC and PLD activities, respectively, may have escaped detection. Therefore, the expression of RGS proteins was increased by a second approach, i.e. infection of cells with recombinant adenoviruses encoding Lsc-RGS or RGS4. As judged by the expression of green fluorescent protein, the efficiency of adenoviral gene transfer into HEK-293 cells was Ͼ95% (data not shown). The expression of RGS4 by adenoviral infection was without any effect on M 3 mAChR-stimulated PLD activity (Fig.  8A). Neither the maximal extent nor the concentration dependence of carbachol-induced PLD stimulation was altered in RGS4 expression compared with control cells infected with an adenovirus encoding LacZ. In contrast, the carbachol-induced PLD stimulation was blunted (by about 80%) by the expression of Lsc-RGS. On the other hand, adenoviral expression of Lsc-RGS did not affect the stimulation of PLC activity at any carbachol concentration that had been examined, whereas PLC stimulation was strongly reduced (by about 70%) by the expression of RGS4 (Fig. 8B). DISCUSSION A large variety of GPCRs that stimulate phosphoinositidehydrolyzing PLC also activates PLD. This concomitant stimulation of the two phospholipases may be a sequential reaction in that the receptor initially leads to PLC stimulation followed by the increase in PLD activity that is caused by intracellular Ca 2ϩ and/or activated protein kinase C, consequences of the PLC reaction that are in fact reported for some GPCRs (1-3). Alternatively, PLC and PLD enzymes are independently regulated by the receptor either by activating one heterotrimeric G protein that then leads to the stimulation of both PLC and PLD or two distinct G proteins, one responsible for PLC stimulation and the other responsible for mediating PLD stimulation. The M 3 mAChR expressed in HEK-293 cells is a prototypical example of a GPCR that stimulates both PLC and PLD (13)(14)(15). Previous studies indicate that PLD stimulation is apparently  independent of changes in intracellular Ca 2ϩ and protein kinase C activation (15,16), suggesting that it is not a consequence of the PLC stimulation. Thus, two other possibilities were considered for mediating M 3 mAChR-PLD coupling. As the M 3 mAChR couples to G q -type G proteins in HEK-293 cells and stimulation of either phospholipase is resistant to PTX (14), we first studied whether G q proteins may mediate M 3 mAChR-PLD coupling. However, as shown here, the basal PLD activity and its stimulation by the M 3 mAChR were not affected by the overexpression of G␣ q (wild type or constitutively active), which on the other hand strongly increased PLC activity. Furthermore, overexpression (transient or by infection with a recombinant adenovirus) of RGS4, which is known to act as a GAP for G q proteins (17)(18)(19)(20)(21)(22)(23)(24), suppressed the M 3 mAChR-mediated PLC stimulation but did not affect PLD stimulation by this GPCR. These results are in line with previously published data on the regulation of PLC stimulation by G␣ q and RGS4 (2,12,(17)(18)(19)(31)(32)(33) and strongly corroborate the idea that PLD stimulation by the M 3 mAChR in HEK-293 cells is in fact not a consequence of the concomitant PLC stimulation.
As the M 3 mAChR couples to PLD but apparently not by activating the PLC-stimulatory G q proteins, we examined whether the PTX-resistant G 12 -type G proteins may act as specific transducers for PLD stimulation. The overexpression of G␣ 12 or G␣ 13 enhanced PLD stimulation by the M 3 mAChR by up to 2-fold. On the other hand, the expression of Lsc-RGS, the murine homolog of the RGS domain of p115 RhoGEF that acts as a specific GAP for G␣ 12 and G␣ 13 (25), suppressed the M 3 mAChR-mediated PLD stimulation. In contrast, the basal and M 3 mAChR-stimulated PLC activities were not altered by the overexpression of G␣ 12 or G␣ 13 or by the expression of Lsc-RGS. These results are in line with our knowledge that G 12 -type G proteins do not directly influence the activities of PLC enzymes (2,11,12) and indicate that the expression of the PLD-inhibitory Lsc-RGS does not unspecifically inhibit the M 3 mAChR. As the expression of G␣ 12 , G␣ 13 , or Lsc-RGS did not alter phorbol ester-stimulated PLD activity, it can be concluded that these proteins, such as G␣ 12 and G␣ 13 , specifically enhance or interfere, such as Lsc-RGS, with M 3 mAChR-PLD coupling. Thus, together with the finding that the M 3 mAChR induced the incorporation of [␥-32 P]GTP-azidoanilide into G 12 proteins, these results strongly suggest that G 12 -type G proteins specifically mediate the coupling of the M 3 mAChR to the PLD signaling pathway in HEK-293 cells.
Recently, the G protein specificities for coupling of the angiotensin II type 1 receptor to PLD and PLC in vascular smooth muscle cells were reported (34,35). Stimulation of PLD activity by this GPCR was inhibited (by 50%) by antibodies against G␣ 12 but not G␣ q/11 . However, both types of antibodies suppressed in an additive manner angiotensin II stimulation of PLC-␤1. Thus, in contrast to the specific function of G 12 -type G proteins in M 3 mAChR-PLD coupling in HEK-293 cells, G 12 proteins are apparently involved in GPCR-induced stimulation of both phospholipases, PLD and PLC, in vascular smooth muscle cells.
As direct activation of PLD enzymes by heterotrimeric G proteins including G 12 -type G proteins has not been observed (1)(2)(3)(4)(5), these G proteins most probably activate intermediate signaling pathways and proteins, finally leading to PLD stimulation. The best characterized intermediates involved in receptor-induced PLD stimulation are small GTPases of the ADP-ribosylation factor and Rho families as well as protein kinase C isoenzymes (1)(2)(3)(4)(5). As G 12 -type G proteins do not stimulate PLC (11,12), PLD stimulation by these G proteins most probably is not caused by protein kinase C activation. Many of the diverse cellular responses observed upon expression of constitutively active G␣ 12 or G␣ 13 including PLD1 stimulation, actin stress fiber formation, neurite retraction, activation of Na ϩ -H ϩ -exchanger, and serum response elementdependent gene transcription were inhibited by dominant- negative Rho GTPases (10, 36 -40), suggesting that Rho GTPases are major targets of the action of G 12 -type G proteins. However, from the finding that Rho GTPases are involved in a cellular response, it cannot be concluded that G 12 -type G proteins mediate the receptor action as Rho-dependent cellular actions as well as direct Rho activation can also be induced by activated G␣ q proteins (37, 40 -43). Also, in HEK-293 cells, the expression of mutationally activated G␣ 12 and G␣ q family members caused strong and comparable stimulation of serum response factor-mediated gene transcription that was sensitive to Clostridium botulinum C3 transferase (data not shown), indicating the involvement and probably activation of Rho by either type of heterotrimeric G protein. Thus, although G q proteins can cause Rho activation, the Rho-dependent PLD stimulation by the M 3 mAChR in HEK-293 cells (44) is apparently independent of this reaction but mediated by G 12 -type G proteins. In combination with the finding that the M 3 mAChR activates PLC and protein kinase C (data not shown) but stimulates PLD that is apparently independent of this reaction, these data suggest that the M 3 mAChR-PLD coupling occurs in a highly organized subcellular compartment.
PLD stimulation by the M 3 mAChR-expressed HEK-293 cells is dependent on Rho GTPases apparently acting via Rho kinase (44,45). Thus, from the data discussed above and those data presented in this study, it seems reasonable to assume that the M 3 mAChR activates endogenous G 12 -type G proteins, which then through an unidentified GEF activate Rho and subsequently PLD. In addition to and apparently independent of Rho proteins, ADP-ribosylation factor GTPases are required for M 3 mAChR-induced PLD stimulation in HEK-293 cells (46,47). As PLD stimulation was not affected by the inactivation of G q proteins (by expression of RGS4), it has to be concluded that the activation of ADP-ribosylation factor GTPases is probably also mediated by G 12 -type G proteins. Experiments are in progress to study whether both G 12 -type G proteins, G 12 and G 13 , are involved in M 3 mAChR-induced PLD stimulation and may even exhibit selectivity for coupling the receptor to the activation of Rho and ADP-ribosylation factor GTPases.
In summary, our data indicate that distinct heterotrimeric G proteins couple the M 3 mAChR to the stimulation of PLD and PLC in HEK-293 cells. Whereas G q -type G proteins mediate M 3 mAChR-PLC coupling, the stimulation of PLD activity is apparently mediated specifically by G 12 -type G proteins. Moreover, we demonstrate that G protein subtype-specific RGS proteins can be used as powerful tools to dissect the PTX-resistant G protein families, G q and G 12 , and their role in receptor-effector coupling.