Signaling from G Protein-coupled Receptors to ERK5/Big MAPK 1 Involves G a q and G a 12/13 Families of Heterotrimeric G Proteins EVIDENCE FOR THE EXISTENCE OF A NOVEL Ras AND Rho-INDEPENDENT PATHWAY*

,

expression of GPCRs with chimeric G protein ␣ subunits to begin delineating the biochemical route linking GPCRs to ERK5. We show that the G␣ q and G␣ 12/13 families of heterotrimeric G proteins ␣ subunits, but not the G␣ i , G␣ s , or ␤␥ subunits, are able to regulate ERK5 activity. Furthermore, we obtained evidence that the stimulation of the ERK5 cascade by GPCRs involves a novel pathway, which is distinct from those regulated by Ras and Rho GTPases.

EXPERIMENTAL PROCEDURES
Expression Plasmids-PAR1, kindly provided by Dr. L.F. Brass, was subcloned into the pCEFL vector as an EcoRI fragment. DNA encoding a G 13 /G i chimera, in which 5 amino acids at the C terminus of G␣ 13 were replaced by the corresponding sequence of G␣ i2, was prepared by polymerase chain reaction amplification using pcDNA3 HA-G␣ 13 (27) as a template, and the resulting DNA was subcloned into the pCEFL HA vector (24) as a BglII-EcoRI fragment. Sequences of mutagenic oligonucleotides will be made available upon request. Plasmids expressing epitope-tagged ERK5, MAPK, and JNK, pCEFL HA-ERK5, pcDNA3 HA-MAPK, and pcDNA3 HA-JNK, respectively, as well as expression plasmids for constitutively activated forms of Ras, Rho, Rac1, Cdc42, G␣ q , G␣ i2 , G␣ s , G␣ 12 , G␣ 13 , ␤ and ␥ subunits of G proteins, dominant negative mutants of Ras and Rho, the CRIB (Cdc42, Rac interactive binding) domain of PAK (PAK-N), m1 and m2 muscarinic receptors, a Gal4 fusion protein containing the transactivating domain of MEF2C and dominant negative and active mutants of MEK5, MEK5AA, and MEK5DD, respectively, were described previously (21,24,27,28). A DNA plasmid encoding a G q /G i chimeric protein, in which 5 amino acids at the C terminus of G␣ q were replaced by the corresponding sequence of G␣ i2 , was a gift from Dr. B. Conklin (29). Reporter plasmids that express the chloramphenicol acetyltransferase (CAT) gene under the control of the mutant form of the serum response element (SRE) from the c-fos promoter, lacking the ternary complex factor binding site (SREmutL) as well as an expression vector for the C3 toxin were kindly provided by Dr. R. Treisman (30).
Cell Culture and Transfection-COS-7 cells were maintained in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) supplemented with 10% fetal bovine serum. Cells were transfected using Lipo-fectAMINE Plus TM reagent (Life Technologies, Inc.) according to the manufacturer's protocol. In each experiment, the total amount of DNA was adjusted to 3-10 g/plate with a plasmid for green fluorescent protein.
Kinase Assays-The ERK5 kinase activity in cells transfected with an expression plasmid for HA-ERK5 was measured as described previously (24), using 3 g of GST⅐MEF2C fusion protein containing the transactivating domain of MEF2C as a substrate. MAPK and JNK activities in cells transfected with an epitope-tagged MAPK (HA-ERK2, referred in here as HA-MAPK) or JNK (HA-JNK) were determined as described previously (21), using myelin basic protein (Sigma) or bacterially expressed GST⅐ATF2(96) fusion protein as a substrate, respectively. The expression level of HA-ERK5, HA-MAPK, and HA-JNK in lysates from transfected cells was assessed by Western blot analysis after SDS-polyacrylamide gel electrophoresis with the specific antibody against HA (HA.11; Berkeley Antibody Company).
Reporter Gene Assays-The transactivating activity of MEF2C and the SRE activity were determined as described previously (24,27). Briefly, for MEF2C, COS-7 cells plated in a 24-well plate were transfected with different expression plasmids together with 2 ng of pCDNAIII-Gal4-MEF2C, a plasmid expressing a Gal4 fusion protein containing the transactivation domain of MEF2C (amino acids 161-350) as well as 50 ng of pGal4-Luc and 10 ng of pRL-null (Promega). To measure the SRE activity, COS-7 cells were transfected with the indicated plasmids together with 0.1 g of pCDNAIII-␤-galactosidase, a plasmid expressing the enzyme ␤-galactosidase, and 0.1 g of pSRE-mutL, the reporter plasmid expressing a CAT gene under the control of the mutant SRE lacking a ternary complex factor binding site. After transfection, cells were cultured for ϳ24 h in serum-free Dulbecco's modified Eagle's medium, then stimulated with the indicated ligands for an additional 6 h, and lysed using reporter lysis buffer (Promega). Luciferase activities in cell extracts were determined using a dual luciferase assay system (Promega). CAT activity was assayed in the cell extracts by incubation at 37°C for 1 h in the presence of 0.25 Ci of [ 14 C]chloramphenicol (100 mCi/mmol) (ICN) and 200 g/ml butyryl-CoA (Sigma) in 0.25 M Tris-HCl, pH 7.4. Labeled butyrylated products were extracted using a mixture of xylenes and 2,6,10,14-tetramethylpentadecane (ratio 1:2), and radioactivity was counted. ␤ϪGalactosid-ase activity present in each sample was assayed by a colorimetric method and was used to normalize for transfection efficiency.

RESULTS AND DISCUSSION
Coupling Specificity of G Protein-linked Receptors Stimulating ERK5 Kinase Activity-ERK5 has been recently found to participate in the regulation of the c-jun promoter by transforming GPCRs (24). To begin exploring the nature of the pathway linking these cell surface receptors to ERK5, we first investigated which classes of GPCRs are able to stimulate ERK5 kinase activity. For these experiments, COS-7 cells were transiently transfected with expression plasmids for a HAtagged form of ERK5, and its kinase activity was measured by an in vitro kinase assay using MEF2C fused to GST as a substrate. As shown in Fig. 1A, stimulation by carbachol, a cholinergic agonist, of transfected G␣ q -coupled m1 muscarinic receptors (31) potently activated ERK5. Similarly, stimulation with a tyrosine kinase receptor agonist, epidermal growth factor, also enhanced ERK5 kinase activity, as recently reported (32,33). However, the stimulation of m2 muscarinic receptors, which are typical G i -coupled receptors (34), had no effect on ERK5 activity, although it potently activated MAPK, which served as an internal control (8). ERK5 activity was also stimulated by exposure to thrombin, which acts on endogenously expressed GPCRs, and this effect was slightly enhanced by overexpression of its cognate receptors, PAR1 (Fig. 1A). Kinetics of ERK5 activation mediated by m1 and thrombin receptors were very similar, and responses were evident within 5 min after agonist addition and reached a maximal level around 10 min (Fig. 1B). As m1 and m2 muscarinic receptors couple to G␣ q and G␣ i types of heterotrimeric G proteins, respectively, and thrombin receptors can stimulate both the G␣ q and G␣ i as well as the G␣ 12/13 families of G proteins (35)(36)(37)(38), these findings suggest that receptors coupled to G␣ q and, possibly G␣ 12/13 , may harbor the ability to transduce a signal to ERK5, whereas G i -coupled receptors do not.
The MEK5-ERK5 Kinase Cascade Is Involved in the Activation of MEF2C Transcriptional Activity Mediated by G Proteincoupled Receptors-MEF2C is a physiological substrate for ERK5 (39). Thus, we next asked whether the ability to enhance the in vitro phosphorylating activity of ERK5 by GPCRs resulted in enhanced transcriptional activity of MEF2 proteins in vivo. For these experiments, we fused the transactivation domain of MEF2C to the DNA binding domain of Gal4 and tested the ability to induce the expression from a pGal4-Luc reporter plasmid, as described previously (24). As shown in Fig. 2A, expression from the Gal4-driven luciferase reporter was induced by the stimulation of m1 and thrombin receptors, but not m2 receptors, which is consistent with their abilities to stimulate ERK5 kinase activity. Furthermore, transfection of a DNA plasmid for MEK5AA, which acts as a dominant negative mutant of MEK5 (24), completely blocked the increased transcriptional activity of MEF2C elicited by thrombin and partially inhibited m1 mediated-transcriptional activation (Fig. 2B). As a control, the activation of SRE mediated by m1 and thrombin receptors was unaffected by co-expression of MEK5AA (Fig.  2B). These findings suggested that the MEK5-ERK5 kinase pathway is functionally activated by GPCRs and that this kinase cascade is involved in the activation of MEF2C by m1 and thrombin receptors.
Activated Forms of G␣ 12 and G␣ 13 Stimulate ERK5-To investigate which classes of G proteins mediate ERK5 activation induced by GPCRs, we examined the effects of activated forms of G␣ subunits as well as overexpression of ␤␥ subunits of heterotrimeric G proteins on ERK5 kinase activity. As shown in Fig. 3A, expression of G␣ 12 QL and G␣ 13 QL could induce ERK5 activation, whereas ERK5 kinase activity was not al-tered by the expression of G␣ q QL, G␣ i2 QL, G␣ s QL, and the ␤␥ subunit, thus suggesting the involvement of the G␣ 12/13 family of G proteins in GPCR-mediated signaling to ERK5. Regarding G␣ i , these observations are in line with the failure of m2 muscarinic receptors, which transduce signals through G␣ i and ␤␥ subunits, to activate ERK5. In addition, ERK5 activation mediated by m1 and thrombin receptors was found to be insensitive to pertussis toxin, which ADP ribosylates and inactivates G␣ i and G␣ o (data not shown), together indicating that pertussis toxin-sensitive G proteins cannot signal to ERK5 and that they do not mediate the activation of ERK5 by thrombin and m1 receptors.
The Use of Chimeric G Protein ␣ Subunits Reveals That the G␣ q and G␣ 12/13 Families of Heterotrimeric G Proteins Can Signal to ERK5-Stimulation of m1 receptors, that couple to G␣ q , can activate ERK5. However, expression of an activated form of G␣ q could not enhance the ERK5 kinase activity. Interestingly, this is highly reminiscent to that observed for the MAPK pathway (8,40). Indeed, several lines of evidence suggest the ability of G␣ q to stimulate the MAPK pathway, but expression of activated G␣ q could not induce MAPK activation (8,40), and it even prevents the further stimulation of MAPK by a variety of stimuli (41). This discrepancy may be ascribed to the desensitization of downstream signaling pathways, such as that two days after the transfection of expression plasmids the activity of MAPK and/or ERK5 may no longer be demonstrable.
To solve this problem, we took advantage of the finding that a G q /G i chimera, where a C-terminal region of G␣ q is replaced by the corresponding region of G␣ i , can be stimulated by G i -coupled receptors and is able to transmit G q -mediated signaling pathways (29). Thus, upon coexpression of this G q /G i chimera together with a G i -coupled receptor, such as m2 receptors, on and off G q -mediated signaling can now be controlled by agonist addition. Moreover, the treatment with pertussis toxin can now make it possible to specifically activate G q -mediated signals, because G␣ i is inhibited by pertussis toxin and no longer has the ability to stimulate downstream signaling pathways. The potential usefulness of this system prompted us to design also a similar G 13 /G i chimera in which the C-terminal 5 amino acids of G␣ 13 were replaced by the corresponding sequences of G␣ i . Expression of this protein was confirmed by Western blot analysis (data not shown). The functional activity of these chimeras was assessed by examining their ability to stimulate the transcriptional activation of an SRE-containing reporter plasmid, as both G␣ q and G␣ 12/13 classes of G proteins are known to activate a SRE, but not the stimulation of G␣ i (42). As previously reported, the stimulation of m2 by carbachol did not activate the SRE (42), although transcription from the SRE was potently increased by m1 receptor stimulation (Fig. 3B). However, the exposure to carbachol of cells expressing both m2 receptors and either a G q /G i chimera or a G 13 /G i chimera significantly induced SRE activation, whereas the expression of either a G q /G i chimera or a G 13 /G i chimera alone did not affect the SRE activity (Fig. 3B). These results indicated that these G q /G i and G 13 /G i chimeras can be stimulated by m2 receptors and are capable of transmitting G q -and G 13 -mediated signaling pathways, respectively.
We then used this system to investigate whether G␣ q and G␣ 13 have the ability to stimulate ERK5. As shown in Fig. 3C, stimulation with carbachol induced a limited activation of ERK5 in cells expressing m2 receptors. However, in cells expressing either the G q /G i or G 13 /G i chimera, m2 receptor stimulation significantly activated ERK5 (Fig. 3C). These results indicated that both the G␣ q and G␣ 12/13 classes of heterotrimeric G proteins are able to signal to ERK5 and thus might participate in ERK5 activation by GPCRs.

FIG. 1. Differential activation of MAPK and ERK5 by G protein-coupled receptors exhibiting distinct coupling specificity.
COS-7 cells were transfected with expression plasmids for HA-ERK5 or HA-MAPK, together with plasmids expressing GFP, m1, or PAR1 receptors, as indicated, and stimulated with vehicle (c), 100 M carbachol (Cch), 5 units/ml thrombin (Thr), or 100 ng/ml epidermal growth factor (EGF) for 10 min (A) or for the indicated time (B). Kinase reactions were performed using anti-HA immunoprecipitates from the corresponding cellular lysates. Labeled substrates are indicated. Data shown are from a representative experiment for each assay, which was repeated three to five times with similar results. Western blot (WB) analysis was performed with anti-HA antibodies using total cellular lysates. Data represent the mean Ϯ S.E. of three to five independent experiments expressed as fold increase with respect to unstimulated cells (control).

Signaling from G Protein-coupled Receptors to ERK5 Does Not Involve Ras and Rho
GTPases-G i -and G q -coupled receptors can stimulate the Ras-MAPK pathway effectively (6). However, we found that G i -coupled m2 receptors fail to stimulate ERK5, suggesting that the pathway linking GPCRs to ERK5 is different from that which communicates these receptors to MAPK. Indeed, we observed that activated Ras causes only a very limited increase in the enzymatic activity of ERK5, although it potently stimulates MAPK (Fig. 4). On the other hand, activation of the JNK pathway is believed to be mediated by Rac and Cdc42, two members of the Rho family of GTPases (21). However, whereas expression of activated Rac and Cdc42 strongly enhanced the kinase activity of JNK, these GTPases failed to stimulate ERK5 (Fig. 4). Similarly, expression of an activated form of Rho, which stimulates the SRE-driven reporter plasmid potently (30), also failed to enhance the enzymatic activity of ERK5. Furthermore, although some minor variations in the activity of ERK5 can be observed upon expression of these GTPases, an activated form of MEK5, MEK5DD, consistently induced ERK5 activation under these experimen-tal conditions (Fig. 4). Thus, activation of Ras, Rho, Rac, and Cdc42 may not be sufficient to stimulate the ERK5 pathway. However, it is still possible that these small GTPases are required for GPCR-mediated ERK5 activation. To address this possibility, we used the expression of dominant interfering molecules for each of these GTPases. As shown in Fig. 5A, the activation of ERK5 mediated by m1 and thrombin receptors was not affected by the expression of a dominant negative mutant of Ras, RasN17, although this inhibitory molecule effectively inhibited MAPK activation when induced by m1 stimulation (Fig. 5B) and by thrombin (data not shown), but not by phorbol esters, as previously reported (43,44). Thus, together these data suggest that Ras is unlikely to play a prominent role in ERK5 activation by GPCRs. For Rac and Cdc42, we used the overexpression of a molecule containing the CRIB domain of PAK fused to GST, which can specifically bind the GTP-bound forms of Rac and Cdc42 thereby inhibiting these GTPases (28,45,46). Indeed, expression of the CRIB domain of PAK (PAK-N) significantly inhibited JNK activation evoked by the expression of activated forms of Rac and Cdc42 and by m1 FIG. 2. Stimulation of G proteincoupled receptors enhances the transcriptional activity of MEF2C through the MEK5-ERK5 signaling pathways. A, COS-7 cells were cotransfected with pcDNA3 Gal4-MEF2C, pGal4-Luc, and pRL-null, together with expression vectors for GFP and m1, m2, and PAR1 receptors, as indicated. Cells were then exposed to vehicle (c), 100 M carbachol (Cch), and 5 units/ml thrombin (Thr) for 6 h and processed as described in "Experimental Procedures." The data represent luciferase activity normalized by the luciferase activity from Renilla reniformis present in each cellular lysate, expressed as fold induction with respect to control cells, and are the mean Ϯ S.E. of triplicate samples from a typical experiment. Similar results were obtained in three separate experiments. B, COS-7 cells were cotransfected with pSREmutL (SRE) or with pcDNA3 Gal4-MEF2C and pGal4-Luc (MEF2C), together with the expression plasmid for m1 (left panel) or PAR1 (right panel) receptors and with increasing amount of expression plasmid for MEK5AA, as indicated, and stimulated with vehicle, 100 M carbachol, and 5 units/ml thrombin for 6 h. Cells were processed as described under "Experimental Procedures." The data represent CAT (for SRE) and luciferase activities (for MEF2C) expressed as percentage relative to those observed in cells that did not include MEK5AA and are the mean Ϯ S.E. of triplicate samples from a typical experiment. Similar results were obtained in three independent experiments. stimulation but had only a limited effect on the activation of JNK by anisomycin, which served as a control for specificity (Fig. 5C). However, PAK-N did not affect the abilities of m1 and thrombin receptors to activate ERK5 (Fig. 5A). Together, these results indicated that the small GTPases Ras, Rac, and Cdc42 are not involved in the signaling from GPCRs to ERK5.
Interestingly, both G␣ q and G␣ 12/13 classes of G proteins, but not G␣ i , have been shown to activate Rho-dependent signaling pathways (42), and recent studies suggested that Rho-specific exchange factors such as p115-RhoGEF and PDZ-RhoGEF could be directly activated by the G␣ 12/13 family of G proteins (27,47). Together, these findings suggested the possibility that Rho may participate in signaling to ERK5. However, an activated form of Rho did not enhance the kinase activity of ERK5 (see above, Fig. 4) and that of any other member of the MAPK superfamily in this cellular setting. Nonetheless, these obser-FIG. 3. Signaling to ERK5 through G␣ q and G␣ 12/13 families of heterotrimeric G proteins. A, effects of activated mutants of G proteins ␣ subunits and ␤␥ subunits on ERK5 kinase activity. COS-7 cells were transfected with an expression plasmid for HA-ERK5 together with plasmids expressing GFP, G␣ q QL, G␣ i QL, G␣ s QL, G␣ 12 QL, G␣ 13 QL, or ␤ 1 ␥ 2 subunits. Kinase reactions were performed using anti-HA immunoprecipitates from the corresponding cellular lysates. Data represent the mean Ϯ S.E. of three independent experiments, expressed as fold increase with respect to control cells. B, COS-7 cells were cotransfected with pSREmutL and pCMV-␤-galactosidase plasmid DNAs as well as with the indicated expression vectors and stimulated with vehicle (Ϫ) or 100 M carbachol (ϩ) for 6 h. Cells were processed as described under "Experimental Procedures." The data represent CAT activity normalized by the ␤-galactosidase activity present in each cellular lysate, expressed as fold induction with respect to control cells, and are the mean Ϯ S.E. of triplicate samples from a typical experiment. Nearly identical results were obtained in three additional experiments. C, COS-7 cells were transfected with an expression plasmid for the HA-ERK5 together with the indicated expression vectors and stimulated with vehicle (Ϫ) or 100 M carbachol (ϩ) for 10 min. Kinase reactions were carried out using anti-HA immunoprecipitates from the corresponding cellular lysates. Data represent the mean Ϯ S.E. of three independent experiments, expressed as fold increase with respect to control cells. , or HA-JNK (C) as well as with expression vectors carrying cDNAs for GFP, m1, and PAR1 receptors or the activated mutant of Rac1 (RacQL) and Cdc42 (Cdc42QL), together with plasmids encoding inhibitory molecules (RasN17, RhoN19, and PAK-N). Cells were stimulated with or without 100 M carbachol (Cch), 5 units/ml thrombin (Thr), 100 ng/ml 12-Otetradecanoylphorbol-13-acetate (TPA), or 10 g/ml anisomycin as indicated. Kinase reactions were performed using anti-HA immunoprecipitates from the corresponding cellular lysates. Labeled substrates are indicated. Data shown are from a representative experiment for each assay. Western blot (WB) analysis was performed with anti-HA antibodies using total cell lysates (HA-ERK5 and HA-MAPK) or anti-HA immunoprecipitates (HA-JNK). D, COS-7 cells were cotransfected with pSREmutL and pCMV-␤-galactosidase plasmid DNAs as well as with expression vectors carrying cDNAs for GFP, m1 and PAR1 receptors, and the activated mutant of Cdc42 (Cdc42QL), with or without expression plasmids for C3 toxin, as indicated. The next day, the cells were stimulated with vehicle, 100 M carbachol (Cch) or 5 units/ml thrombin (Thr) for 6 h. Cells were processed as described under "Experimental Procedures." The data represent CAT activity normalized by the ␤-galactosidase activity present in each cellular lysate, expressed as fold induction with respect to control cells, and are the mean Ϯ S.E. of triplicate samples from a typical experiment. Similar results were obtained in three separate experiments. E, COS-7 cells were transfected with expression vectors carrying DNA for HA-ERK5 and GFP, m1 or PAR1 receptors, with or without expression plasmid for C3 toxin, and stimulated with vehicle, 100 M carbachol (Cch), 5 units/ml thrombin (Thr), or 100 ng/ml epidermal growth factor (EGF) for 10 min. Kinase reactions were performed in anti-HA immunoprecipitates from the corresponding cellular lysates. Data represent the mean Ϯ S.E. of three independent experiments, expressed as fold increase with respect to unstimulated cells. Autoradiograms correspond to representative experiments. Western blot (WB) analysis was performed in the corresponding cellular lysates and immunodetected with the antibody to HA. vations cannot rule out the possibility that Rho stimulates certain MAPKs, which might not be revealed by the expression of its activated mutants for reasons such as those described for G␣ q . Thus, to explore further the possibility that G q -and G 12/13 -coupled receptors uses Rho to stimulate ERK5, we used as a more definitive approach the expression of Clostridium botulinum C3 exoenzyme, which specifically ADP ribosylates Rho thus preventing its activation (48). As shown in Fig. 5D, a C3 toxin inhibited the activation of SRE, a typical Rho-dependent response (30), by m1 and thrombin receptors, although the Cdc42-mediated activation of SRE was unaffected and served as a control. However, this toxin did not change the abilities of m1 and thrombin receptors to activate ERK5 (Fig. 5E), strongly suggesting that the activation of ERK5 by GPCRs is independent of Rho.
In conclusion, the present study demonstrates that the ERK5 pathway can be functionally activated by the stimulation of GPCRs depending on their coupling specificity and that the G q and G 12/13 families of heterotrimeric G proteins can mediate this effect. As ERK5 regulates the activity of a growing number of nuclear transcription factors, these findings may help explain the distinct ability of G q -and G 12/13 -coupled receptors to promote the expression of growth related genes. Furthermore, our observations raise the possibility of the existence of a novel signaling pathway whereby GPCRs enhance the activity of ERK5. Although the precise nature of this biochemical route is still unknown, we provide evidence that the pathway linking GPCRs to the MEK5-ERK5 kinase cascade is distinct from those utilized by these cell surface receptors to stimulate MAPK, JNK, and Rho GTPases.