G-protein -Subunits Activate Mitogen-activated Protein Kinase via a Novel Protein Kinase C-dependent Mechanism

Mitogen-activated protein kinase (MAPK) is activated in response to both receptor tyrosine kinases and G-protein-coupled receptors. Recently, G-coupled receptors, such as the α adrenergic receptor, were shown to mediate Ras-dependent MAPK activation via a pathway requiring G-protein β subunits (G) and many of the same intermediates involved in receptor tyrosine kinase signaling. In contrast, G-coupled receptors, such as the M muscarinic acetylcholine receptor (MAChR), activate MAPK via a pathway that is Ras-independent but requires the activity of protein kinase C (PKC). Here we show that, in Chinese hamster ovary cells, the MAChR and platelet-activating factor receptor (PAFR) mediate MAPK activation via the α-subunit of the G protein. G-mediated MAPK activation was sensitive to treatment with pertussis toxin but insensitive to inhibition by a G-sequestering peptide (βARK1ct). MAChR and PAFR catalyzed G α-subunit GTP exchange, and MAPK activation could be partially rescued by a pertussis toxin-insensitive mutant of G but not by similar mutants of G. G-mediated MAPK activation was insensitive to inhibition by a dominant negative mutant of Ras (N17Ras) but was completely blocked by cellular depletion of PKC. Thus, MAChR and PAFR, which have previously been shown to couple to G, are also coupled to G to activate a novel PKC-dependent mitogenic signaling pathway.

by a variety of extracellular stimuli, including those mediated by receptor tyrosine kinases (RTKs) and G-protein-coupled receptors (GPCRs) (1)(2)(3). The mitogenic signaling pathway mediated by the epidermal growth factor RTK involves a cascade of protein-protein interactions, leading to Ras-dependent MAPK activation (4,5). Agonist binding to the epidermal growth factor RTK leads to receptor dimerization and autophosphorylation, resulting in a phosphotyrosine-dependent association with Shc. The subsequent interaction between Tyr(P)phosphorylated Shc and the Grb2 adaptor protein causes a translocation of the Grb2-SOS complex to the membrane, where SOS mediates guanine nucleotide exchange on Ras (6).
Recently, ␤␥ subunits derived from PTX-sensitive heterotrimeric G-proteins were also shown to mediate Ras-dependent MAPK activation (7)(8)(9)(10). Release of G ␤␥ promotes the tyrosine phosphorylation of Shc and its subsequent association with Grb2-SOS. Both RTK-and G ␤␥ -mediated MAPK activation are completely blocked by the expression of dominant negative mutants of mSOS1 and Ras, demonstrating that RTKs and G ␤␥ activate MAPK via a common signaling pathway involving Shc, Grb2, SOS, and Ras (7).
MAPK activation via G i -coupled receptors, such as ␣ 2A AR and the lysophosphatidic acid receptor, is sensitive to inhibition by the C-terminal fragment of ␤ARK1 (␤ARK1ct), a competitive inhibitor of G ␤␥ -mediated signals (10). However, not all GPCRs mediate MAPK activation exclusively via receptor-catalyzed release of ␤␥ subunits. For example, in COS-7 cells, MAPK activation via receptors coupled to members of the PTXinsensitive G q/11 family, such as M 1 AChR and the ␣ 1 adrenergic receptor (␣ 1 AR), is insensitive to the G ␤␥ -sequestrant ␤ARK1ct peptide (11). Instead, MAPK activation occurs predominantly via a PKC-dependent pathway. The GTP-bound ␣-subunit of the G q/11 protein activates phosphoinositide hydrolysis (12) and protein kinase C (PKC). Once activated, PKC stimulates MAPK activity via a poorly understood mechanism involving the activation of Raf kinase (13,14).
MAPK activation in CHO cells stably transfected with PAFR cDNA has been reported to be sensitive to PTX and independent of Ras (15). We have studied MAPK activation by GPCRs in COS-7 and CHO cells and find that the mechanism of M 1 AChRand PAFR-mediated MAPK activation varies between cell types. Our data demonstrate the existence of a novel PKC-dependent mitogenic signaling pathway, which is mediated by the ␣-subunit of the PTX-sensitive G o -protein and which is independent of Ras activation.

EXPERIMENTAL PROCEDURES
DNA Constructs-Hemagglutinin-tagged p44 MAPK (p44 HA-MAPK ) cDNA was provided by J. Pouysségur; the dominant negative mutant p21 N17Ras cDNA was provided by D. Altschuler and M. Ostrowski; the PTX-insensitive G o␣ (G o␣ PT) cDNA was provided by R. Taussig; the PTX-insensitive mutants of G i1␣ , G i2␣ , and G i3␣ (G i1␣ PT, G i2␣ PT, and G i3␣ PT) cDNAs were provided by S. Senogles. All PTX-insensitive G ␣subunits were created by a mutation of the C-terminal cysteine, thereby removing the site of ADP-ribosylation by PTX. PAFR cDNA was provided by R. Snyderman; M 1 AChR cDNA was provided by E. Peralta (16); G ␤ and G ␥ cDNAs were provided by M. Simon; ␣ 2A AR cDNA and * This work was supported in part by Grant HL16037 from the National Institutes of Health. 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. ‡ ‡ To whom correspondence should be addressed: Howard Hughes Medical Inst., Box 3821, Duke University Medical Center, Durham, NC 27710. 1 The abbreviations used are: MAPK, mitogen-activated protein kinase; RTK, receptor tyrosine kinase; GPCR, G-protein-coupled receptor; ␣ 1B AR cDNA were cloned in our laboratory.
Cell Culture and Transfection-COS-7 and CHO-K1 cells were maintained as described (11). Transient transfection of both cell types was performed using LipofectAMINE (Life Technologies, Inc.) as described previously (17). Cells were treated with PTX or phorbol ester (PMA) 24 h after transfection, where indicated. Assays were performed 48 h after transfection.
Measurement of MAPK Activity-Agonist-stimulated activation of p44 HA-MAPK was determined as described previously (17) using myelin basic protein (MBP) as an exogenous substrate. [ 32 P]ATP-labeled MBP was detected and quantitated after electrophoresis using a Molecular Dynamics PhosphorImager.
Immunoblotting-The expression of endogenous G ␣ subunits was assayed by immunoblotting whole cell lysates using standard methods and anti-G o␣ rabbit polyclonal antibody, anti-G q␣ rabbit polyclonal antibody, or anti-G i3␣ rabbit polyclonal antibody (Upstate Biotechnology Inc.).
Photoaffinity Labeling of Plasma Membrane G Proteins-[␣-32 P]GTP azidoanilide was made from [␣-32 P]GTP and purified by polyethyleneimine cellulose chromatography as described previously (18). Two or three days after transfection, cells were washed three times with Ham's F-12 medium and then permeabilized with streptolysin-O (10 units/5 ml ϫ 30 min). Cells were rinsed three times with medium and once with photolabel buffer (25 mM Hepes, pH 7.5, 100 mM KCl, 5 mM MgCl 2 , 5 mM CaCl 2 , 5 g/ml soybean trypsin inhibitor) and then incubated with the same buffer (500 l) containing [␣-32 P]GTP azidoanilide (Ϸ10 Ci) and 3 M GDP at 37°C for 10 min. Agonist or vehicle was added for a further 10-min incubation. Cells were rinsed twice with ice-cold photolabel buffer containing 1 mM dithiothreitol, placed on ice, and illuminated with UV light for 4 min. After exposure to UV light, cells were washed rapidly twice with unsupplemented photolabel buffer, scraped into 250 l of solubilization buffer (25 mM Hepes, pH 7.5, 100 mM NaCl, 1% Triton X-100, 0.1% SDS, 0.1% cholate), and incubated on ice for 1 h with frequent vortexing. The suspension was centrifuged at 500,000 ϫ g for 20 min. The supernatants were harvested and incubated for 1 h with 5 l of anti-␣ o antibody (Upstate Biotechnology Inc., Lake Placid, NY) and 25 l of a 50% slurry of protein A-Sepharose. Tubes were centrifuged and supernatants discarded, and pellets were washed three times with ice-cold phosphate-buffered saline and then boiled for 5 min in Laemmli sample buffer. Samples were run under reducing conditions on SDS-polyacrylamide gel electrophoresis and were subjected to autoradiography. Relative densities of the G protein bands were determined with a model GS-670 imaging densitometer and Molecular Analyst TM /PC software (Bio-Rad).

Effects of Pertussis Toxin Treatment and the G ␤␥ Sequestrant
␤ARK1ct Peptide on MAPK Activation in COS-7 and CHO Cells- Fig. 1 depicts the effects of PTX and ␤ARK1ct, an inhibitor of G ␤␥ -mediated signaling (10,19), on MAPK activation by GPCRs and overexpressed G ␤␥ subunits in COS-7 and CHO cells. In COS-7 cells, ␣ 1 AR-and M 1 AChR-mediated MAPK activation was insensitive both to PTX treatment and the expression of the ␤ARK1ct peptide (Fig. 1A). The G i -coupled ␣ 2A AR activated MAPK in a PTX-and ␤ARK1ct-sensitive manner, while MAPK activity mediated by transiently transfected G ␤␥ subunits was blocked by ␤ARK1ct. Agonist treatment of COS-7 cells transiently expressing PAFR did not result in detectable MAPK activation. These data are consistent with two distinct pathways. One, employed by ␣ 1 AR and M 1 AChR, is primarily mediated by the ␣-subunits of PTX-insensitive G proteins; the other, employed by the ␣ 2A AR, is primarily dependent upon G ␤␥ subunits derived from PTX-sensitive G proteins.
In CHO cells, three patterns emerged. First, the PTX-insensitive ␣ 1 AR-mediated signal remained PTX-insensitive, as found in COS-7 cells. Similarly, the G ␤␥ -dependent signals, mediated by either ␣ 2A AR or by transfected G ␤␥ , remained sensitive to ␤ARK1ct. In contrast, stimulation of PAFR-transfected CHO cells mediated a 5-fold increase in MAPK activity. Moreover, MAPK activation via M 1 AChR and PAFR was abolished by PTX treatment but remained insensitive to G ␤␥ sequestration by ␤ARK1ct expression (Fig. 1B). Thus, M 1 AChR can activate MAPK via two distinct pathways, one sensitive (CHO cells) and one insensitive (COS-7 cells) to PTX, while neither pathway appears to be mediated by G-protein ␤␥ subunits. Like the M 1 AChR, PAFR can mediate PTX-sensitive, ␤ARK1ct-insensitive MAPK activation in CHO cells. Interestingly, M1AChR and PAFR mediated PTX-insensitive phosphoinositide hydrolysis in both COS-7 and CHO cells (data not shown).
M 1 AChR and PAFR Are Coupled to G o in CHO Cells-To determine whether the PTX-sensitive MAPK activation pathway in CHO cells might be due to the differential expression of G-protein ␣-subunits, we compared G ␣ subunit expression between COS-7 and CHO cells. As shown in Fig. 2A, the levels of expression of G q␣ and G i3␣ were comparable between COS-7 and CHO cells. The ␣-subunits of G i1 and G i2 also showed similar expression in COS-7 cells only (data not shown). In contrast, G o␣ expression was detected only in CHO cells. The expression of G o in CHO, but not in COS-7 cells, suggested that G o␣ subunits might mediate PTX-sensitive MAPK activation in these cells.
PAFR has previously been shown to couple to G o in NCB-20 cells (20), whereas M 1 AChR, like ␣ 1 AR, is known to couple only to members of the G q/11 family (21,22). To determine whether these receptors were capable of coupling to G o in CHO cells, we measured G o␣ GTP exchange in permeabilized cell preparations. As shown in Fig. 2B, agonist stimulation of either PAFR or M 1 AChR mediated a 2-3-fold increase in the incorporation of the photoactivatable GTP analog into G o␣ in immunoprecipitates from CHO cell lysates, indicating that both receptors are

FIG. 1. Comparison of the effects of PTX treatment and ␤ARK1ct peptide on MAPK activation in COS-7 and CHO cells.
A, COS-7 cells were transiently co-transfected with p44 HA-MAPK plus the indicated cDNAs with or without the ␤ARK1ct peptide cDNA. The effects of PTX treatment (100 ng/ml, 20 h) and ␤ARK1ct peptide expression on basal and agonist-stimulated MAPK activity, assessed as phosphorylation of MBP by immunoprecipitated p44 HA-MAPK , was determined following a 5-min exposure to epinephrine (100 M), carbachol (1 mM), PAF (100 nM), or UK14304 (10 M). For G ␤␥ , cells were transfected with G ␤ and G ␥ cDNAs (stimulated) either with or without ␤ARK1ct. B, CHO cells were transiently co-transfected with p44 HA-MAPK plus the indicated cDNAs with or without the ␤ARK1ct peptide cDNA. MAPK activity was determined as described above. Data shown represent the mean Ϯ S.D. for duplicate samples from a representative experiment, which was replicated 3 times with comparable results. White column, basal; shaded column, stimulated; black column, stimulated (PTX treated); hatched column, stimulated plus ␤ARK1ct.
capable of coupling to and activating G o . The specificity of the anti-G o␣ antibody was confirmed by its inability to detect G i␣ subunits in immunoblotting assays of whole cell lysates or partially purified membrane preparations (data not shown).
G o -proteins Mediate M 1 AChR-dependent MAPK Activation in CHO Cells-PTX-insensitive mutant ␣-subunits of G i have previously been shown to rescue the PTX-mediated inhibition of adenylyl cyclase by D2 dopamine receptor (23), whereas a PTX-insensitive mutant of G o␣ rescued norepinephrine-mediated inhibition of voltage-dependent calcium current (24). To examine whether G o␣ could mediate PTX-sensitive activation of MAPK, we determined whether the M 1 AChR signal in PTXtreated CHO cells could be rescued by co-expression of a PTXinsensitive mutant of G o .
Agonist-stimulated MAPK activity was measured in CHO cells co-transfected with M 1 AChR plus the PTX-insensitive mutants of G i1␣ , G i2␣ , G i3␣ , and G o␣ (G i1␣ PT, G i2␣ PT, G i3␣ PT, and G o␣ PT, respectively). As shown in Fig. 3, MAPK activation by M 1 AChR was almost completely inhibited by PTX in control cells and in cells transfected with the PTX-insensitive mutants of G i . In contrast, M 1 AChR-mediated MAPK activation was rescued by G o␣ PT in cells treated with PTX, suggesting that the G o protein is able to mediate MAPK activation by M 1 AChR in CHO cells. G o␣ PT was unable to rescue ␣ 2A AR-mediated MAPK activation from PTX inhibition (data not shown).
G o␣ -mediated MAPK Activation Is Ras-independent and PKCdependent-MAPK activation via G ␤␥ and PTX-sensitive Gproteins has recently been shown to involve the activation of Ras (10) in addition to many of the same intermediates involved in RTK-mediated mitogenic signaling (7). To determine the involvement of Ras in G o -mediated MAPK activation in CHO cells, we assessed the effects of expression of a dominant negative mutant of Ras (25) (N17Ras) on M 1 AChR-and PAFRmediated MAPK activation. N17Ras did not affect MAPK activation by M 1 AChR and PAFR or by the G q/11 -coupled ␣ 1 AR, whereas G ␤␥ -and ␣ 2A AR-mediated signaling was significantly inhibited (Fig. 4). These data are consistent with the observation that M 1 AChR-and PAFR-mediated MAPK activation in CHO cells occurs in the absence of Ras activation (15).
To determine the role of PKC in the PTX-sensitive activation of MAPK by M 1 AChR and PAFR, we pretreated cells overnight with phorbol ester to deplete endogenous PKC. As shown in Fig. 4, MAPK activation by ␣ 1 AR, M 1 AChR, and PAFR was completely blocked by PKC depletion, whereas G ␤␥ and G icoupled ␣ 2A AR were unaffected. Thus, in CHO cells, M 1 AChR and PAFR couple to G o to activate MAPK via a signaling pathway that is independent of Ras but dependent on the activity of PKC. DISCUSSION We have characterized the mitogenic signaling pathways mediated by several G protein-coupled receptors in COS-7 and CHO cells. The data demonstrate the existence of a novel mitogenic signaling pathway mediated via the ␣-subunit of the G o protein. In CHO cells, activation of endogenous G o mediates PKC-dependent MAPK activation. Although it is not clear that PAFR and M 1 AChR activate G o under physiological conditions, we have shown that these receptors, when transiently expressed in CHO cells, activate MAPK via the ␣-subunits of G o .
The G o protein is the least well characterized of the known PTX-sensitive G proteins. G o is localized primarily to the growth cones in the mammalian brain (26) and may be involved in neuronal development and differentiation. G o is known to mediate a variety of intracellular effects, including inhibition of adenylyl cyclase (27), inhibition of voltage-dependent Ca 2ϩ channels (28,29), and stimulation of phosphoinositide hydrolysis (30). Intracellular injection of a constitutively active mu- tant of G o␣ (Q205LG o -␣) mediates a PKC-dependent resumption of the Xenopus oocyte cell cycle (31). Expression of Q205LG o -␣ stimulates mitogenesis in NIH 3T3 cells, and prolonged expression leads to cellular transformation (32).
In addition to G o , the PAF receptor has been reported to couple to the PTX-insensitive G s and the PTX-sensitive G i1 and G i2 (20). PAF mediates a variety of physiological effects, including increased expression of the c-fos and c-jun protooncogenes (33), increased neurite outgrowth in PC12 pheochromocytoma cells (34), elevation of intracellular Ca 2ϩ (35), and increased phosphoinositide hydrolysis (15). M 1 AChR has previously been shown to couple primarily to G q/11 (21,22), whereas our data demonstrate a coupling with G o to mediate mitogenic signaling. In contrast, ␣ 1 AR, which is also coupled to G q/11 in COS-7 cells, is unable to couple to G o in CHO cells, demonstrating that the interaction between M 1 AChR and G o is specific.
The activation of MAPK via a PTX-sensitive, but Ras-independent, pathway is inconsistent with the known mechanism of G i -mediated mitogenic signaling, which requires G-protein ␤␥ subunits and the activation of Shc, Grb2, SOS, and Ras (7, 10) (Fig. 5). Our data show that the PTX-sensitive activation of MAPK was insensitive to the G ␤␥ sequestering ␤ARK1ct peptide and, moreover, was specifically rescued by a PTX-insensitive mutant of G o␣ , demonstrating the direct involvement of the ␣-subunit of G o in mitogenic signaling.
It has been suggested that PKC stimulation is capable of mediating MAPK activation via direct phosphorylation of Raf (13). Consistent with this observation, G o -mediated MAPK activation was unaffected by the N17Ras dominant negative mutant and required the activity of PKC. These data corroborate the observation that PAFR, when stably expressed in CHO cells, is unable to mediate an increase in the GTP-bound form of Ras (15). Interestingly, transfection of COS-7 cells with wild-type G o␣ cDNA did not introduce a PTX-sensitive component to the M 1 AChR-mediated signal, 2 suggesting that additional downstream components, absent from COS-7 cells, may be required for G o␣ to mediate a mitogenic signal.
A model of the known mitogenic signaling pathways mediated by GPCRs (Fig. 5) shows how RTKs and G i -coupled receptors activate MAPK in a Ras-dependent manner, whereas receptors coupled to G o and G q/11 activate MAPK via a pathway that requires PKC. The mechanism by which G o activates PKC and subsequently MAPK remains unknown and is the subject of further investigation. FIG. 5. Model of G-protein-mediated mitogenic signaling. The convergent pathways of GPCR-and RTK-mediated mitogenic signaling are shown. Signals mediated by RTKs and G i -coupled receptors converge at, or before, Shc to mediate Ras-dependent MAPK activation. In contrast, receptors coupled to PTX-sensitive G o or PTX-insensitive G q/11 activate PKC which, in turn, can mediate Ras-independent MAPK activation. Dotted arrows indicate multiple or uncharacterized steps in the pathway. Jagged lines indicate lipid modifications of proteins. MEK, MAPK/extracellular regulated kinase kinase.