Epac- and Ca 2 (cid:1) -controlled Activation of Ras and Extracellular Signal-regulated Kinases by G s -coupled Receptors*

We have recently reported that two typical G s -coupled receptors, the (cid:2) 2 -adrenergic receptor and the receptor for prostaglandin E 1 , stimulate phospholipase C- (cid:3) (PLC- (cid:3) ) and increase intracellular Ca 2 (cid:1) concentration ([Ca 2 (cid:1) ] i ) in HEK-293 cells and N1E-115 neuroblastoma cells, respectively, by a pathway involving Epac1, a cAMP-activated and Rap-specific guanine nucleotide exchange factor (GEF), and the GTPase Rap2B. Here we have demonstrated that these G s -coupled receptors use this pathway to activate H-Ras and the extracellular signal-regulated kinases 1 and 2 (ERK1/2). Specifically, agonist activation of the receptors resulted in activation of H-Ras and ERK1/2. The latter action was suppressed by dominant negative H-Ras, but not Rap1A. The receptor actions were independent of protein kinase A but fully mimicked by an Epac-specific cAMP analog as well as by a constitutively active Rap2B mutant. On the other hand, a cAMP-binding-deficient Epac1 mutant, the Rap GTPase-activating proteinII, and a dominant negative Rap2B mutant suppressed receptor- and Epac-mediated activation of H-Ras and ERK1/2.

Activation of mitogen-activated protein (MAP) 1 kinases plays a prominent role in many cellular responses to a large variety of membrane receptors. Initially identified as signal transducers of growth factor receptors with intrinsic tyrosine kinase activity, activation of MAP kinases, specifically of the extracellular signal-regulated kinases 1 and 2 (ERK1/2), is now also recognized as a major signal transduction pathway of many receptors coupled to heterotrimeric G proteins. ERK activation by these second messenger-generating receptors is apparently accomplished by diverse molecular mechanisms, depending on the cell types studied as well as the receptor and the heterotrimeric G protein involved (1)(2)(3)(4)(5).
The cAMP-producing G s -coupled receptors play a rather unique role in ERK activation. These receptors inhibit ERK activation by growth factor receptors in several cell types while stimulating this cellular response in others, most notably in neuronal and endocrine cells. Several models have been proposed to explain these diverse actions of cAMP and G s -coupled receptors on ERK activation, both apparently involving in most cases the principal cAMP target, the cAMP-activated protein kinase (PKA) (6,7). In particular, PKA-dependent ERK activation in HEK-293 cells by the ␤ 2 -adrenergic receptor (␤ 2 -AR), a prototypical G s -coupled receptor, has been extensively studied. Lefkowitz and coworkers (8) reported that cAMP-activated PKA phosphorylates the ␤ 2 -AR and thereby alters the coupling specificity of the receptor from G s to G i proteins. By this G protein "switching," the ␤ 2 -AR then apparently induces ERK activation by releasing G␤␥ dimers from the pertussis toxin (PTX)-sensitive G i proteins, followed by activation of the cytosolic tyrosine kinase c-Src, the GTPase Ras, and the MAP kinase kinase kinase, Raf-1. Meanwhile, it has been reported that the switching of ␤ 2 -AR from G s to G i proteins is controlled by ␤-arrestin, which recruits the cAMP-degrading phosphodiesterase 4 to the plasma membrane and thereby alters the activity state of PKA (9). In contrast, Schmitt and Stork (10) report that ERK activation by the ␤ 2 -AR in HEK-293 cells, although also PKA-and c-Src-dependent, is PTX-insensitive and thus apparently does not require ␤ 2 -AR switching to G i proteins, a finding recently confirmed by others using a PKAinsensitive ␤ 2 -AR mutant (11). Furthermore, Schmitt and Stork (10) show that ERK activation by ␤ 2 -AR is not mediated by Ras but by the related GTPase Rap1, activating the MAP kinase kinase kinase B-Raf, but not Raf-1. Interestingly, activation of the ␤ 2 -AR also resulted in Ras activation, but this receptor action was apparently independent of cAMP and PKA (10). Recently, these authors reported that activation of Rap1 by the ␤ 2 -AR is apparently mediated by direct phosphorylation of c-Src on serine 17 by PKA, whereas activation of Ras by the ␤ 2 -AR involves G␤␥ dimers and c-Src and finally leads to phosphatidylinositol 3-kinase-dependent activation of AKT (12). Activation of Rap1 by PKA-phosphorylated c-Src obviously also accounts for the inhibition of Ras/Raf-1/ERK signaling by Rap1 by sequestering Raf-1 to specific membrane domains and thereby preventing Raf-1 activation by Ras (13,14).
Few reports indicate that cAMP-dependent ERK activation by G s -coupled receptors can also occur independently of PKA, such as the thyrotropin-induced ERK activation in FRTL5 cells (15) and the parathyroid hormone-induced ERK activation in bone cells (16). Both receptor actions apparently involve Rap1 and B-Raf (15,16). Similar data were recently reported for calcitonin-induced ERK activation in native rat kidney cells (17). Using a specific antibody, evidence was provided that the PKA-independent action of calcitonin involves Epac1 (17), a Rap-specific guanine nucleotide exchange factor (GEF) directly activated by cAMP, independent of PKA (18 -20). Very recently, an involvement of Epac in ERK activation by 5-HT 7 receptors in cultured rat hippocampal neurons and PC12 cells and by the parathyroid hormone-related protein in osteoblasts has been reported as well (21,22). On the other hand, in human melanocytes and B16 mouse melanoma cells, PKA-independent ERK activation by cAMP-elevating agents was shown to be mediated by Ras, but not Rap1 (23).
Thus, Rap1 and Ras are obviously key regulators of ERK activation by G s -coupled receptors. Activation of these small GTPases, i.e. GTP loading, is catalyzed by several distinct GEFs, comprising about 20 distinct proteins (24 -26). Some of these GEFs exhibit Rap specificity, as the above mentioned Epacs and the PDZ-GEFs, whereas others are Ras-specific, such as the SOS and GRF family members. In addition, some of these GEFs, e.g. members of the RasGRP/CalDAG-GEFs, have been reported to activate both Rap and Ras GTPases (24 -26). The latter group of GEFs also includes the recently identified phospholipase C-⑀ (PLC-⑀), which via its CDC25 domain can activate Rap1A and Ras and induce Ras-dependent ERK activation (27)(28)(29)(30). Interestingly, this PLC isoform can also be activated by Ras and Rap GTPases, apparently by binding to a separate domain, the RA2 domain (27)(28)(29)(30). We have recently shown that activation of two typical G s -coupled receptors, the ␤ 2 -AR transiently expressed in HEK-293 cells and a receptor for prostaglandin E 1 (PGE 1 ) endogenously expressed in N1E-115 neuroblastoma cells, leads to stimulation of PLC-⑀ and increases [Ca 2ϩ ] i in a cAMP-dependent manner (31). These G s -coupled receptor actions were independent of PKA but mediated by cAMP-activated Epac1 and subsequent activation of the GTPase, Rap2B (31). The aim of the present study was to analyze whether these G s -coupled receptors use this pathway, particularly Epac1 and PLC-⑀, for ERK activation. We report here that ERK activation by the ␤ 2 -AR and the PGE 1 receptor in HEK-293 cells and N1E-115 neuroblastoma cells, respectively, is mediated by cAMP-activated Epac proteins, which then lead via Rap2B and PLC-⑀ stimulation to [Ca 2ϩ ] i increase, finally resulting in activation of H-Ras as a trigger of the MAP kinase cascade.
Construction of R279K Epac1 and ⌬CDC25 PLC-⑀-The Epac1 mutant, R279K Epac1, with a point mutation in the cAMP-binding pocket (19) was generated with a QuikChange site-directed mutagenesis kit (Stratagene). PCR was carried out with 50 ng of the plasmid carrying the Epac1 cDNA (GenBank TM accession number AF103905) and complementary primers containing the desired mutation (mismatching nucleotides are in bold; forward: 5Ј-ggtgaatgatgcacccaaggcagccaccatcatcc-3Ј; reverse: 5Ј-ggatgatggtggctgccttgggtgcatcattcacc-3Ј). Subsequently, the methylated parental DNA was digested with DpnI endonuclease, and the nicked vector with the desired mutation was transformed into Escherichia coli. The PLC-⑀ deletion mutant ⌬CDC25 PLC-⑀ was constructed by replacing the KpnI/StuI fragment of wild-type PLC-⑀ cDNA by a corresponding PCR product of this region lacking the complete sequence of the CDC25 domain (amino acids 220 -523). PCR was performed with a primer pair flanking the 5Ј-terminus of cDNA encoding PLC-⑀ inclusive of the KpnI/StuI fragment (forward, 5Ј-taatacgactcactataggg-3Ј; reverse, 5Ј-ttctcgaggccttctgtgagtcctctgaaggctgcagatcgatcagctccttg-3Ј) and PLC-⑀ subcloned in pcDNA3 as template using the Expand High Fidelity PCR system (Roche Applied Science). Thereafter, the PCR fragment was ligated into the vector pGEM-T Easy by TA cloning (Promega). As PLC-⑀ contains a second StuI restriction site beyond the open reading frame, the NdeI/XbaI fragment of PLC-⑀ was subcloned into pET-23a (Novagen) to receive the pET-23a-PLC-⑀ plasmid. This construct was digested with KpnI/StuI and ligated with the appropriately cut PCR product, resulting in the fulllength cDNA of ⌬CDC25 PLC-⑀. Finally, the NdeI/XbaI fragment of pET-23a-⌬CDC25 PLC-⑀ was cloned into pcDNA3 to obtain ⌬CDC25 PLC-⑀. The correct sequences were verified by sequencing.
ERK Activation Assay-For measurement of ERK activation, HEK-293 cells and N1E-115 neuroblastoma cells were serum starved for 16 h and then incubated for the indicated periods of time at 37°C with the indicated agents, followed by cell lysis in a buffer containing 1% SDS and 10 mM Tris/HCl, pH 7.4 and five passages through a 25-gauge needle (32). Thereafter, the lysates were clarified by centrifugation, followed by determination of protein concentration and incubation in Laemmli buffer for 10 min at 95°C. After SDS-PAGE and transfer to nitrocellulose membranes, phosphorylated ERK1 and ERK2 were detected with the anti-P-ERK1/2 antibody. Densitometric analysis of the bands (P-ERK1 plus P-ERK2) was performed with ImageQuant software (Amersham Biosciences).
Activation of H-Ras and Rap2B-Serum-starved HEK-293 cells and N1E-115 neuroblastoma cells transfected with wild-type H-Ras or Rap2B were stimulated for 5 min at 37°C with the indicated agents. After cell lysis, activated H-Ras and Rap2B were extracted with glutathione S-transferase (GST)-tagged Raf1-RBD (Ras-binding domain of Raf-1) and GST-tagged RalGDS-RBD (Rap-binding domain of the Ral guanine nucleotide dissociation stimulator), respectively, each bound to glutathione-Sepharose beads, and immunoblotting with anti-H-Ras or anti-Rap2 antibodies as described before (31,33,34). Densitometric analysis of the bands was performed with ImageQuant software (Amersham Biosciences).
Immunoblot Analysis-For detection of P-ERK1/2, Epac1, Epac2 (dilution of 1:1000), c-Myc (dilution of 1:200), HA, ERK1/2, Rap2B, and H-Ras (each at a dilution of 1:500), equal (or indicated) amounts of protein were separated by SDS-PAGE on 10% acrylamide gels. After transfer to nitrocellulose membranes and a 1-h incubation with the antibodies at the dilution factors given above, the proteins were visualized by enhanced chemiluminescence.
Data Presentation-Data shown in the figures are mean Ϯ S.E. of n independent experiments, each performed in triplicate. Comparisons between means were either with the Student's paired t test or one-way analysis of variance test. A difference was regarded as significant at p Ͻ 0.05.

Characterization of ERK Activation by the ␤ 2 -AR in HEK-293
Cells-We have recently reported that the ␤ 2 -AR transiently expressed in HEK-293 cells mediates PLC-⑀ stimulation and [Ca 2ϩ ] i increase by a pathway involving the cAMP-regulated Rap-GEF Epac1 and the GTPase Rap2B (31). Here we studied whether this signaling pathway contributes to activation of MAP kinases, specifically ERK1 and ERK2, by the ␤ 2 -AR. As reported before (31), the HEK-293 cells used for the present study do not endogenously express ␤ 2 -ARs; accordingly, treatment of untransfected cells with the ␤ 2 -AR agonist, adrenaline, did not induce MAP kinase activation (data not shown). However, upon expression of the ␤ 2 -AR, agonist activation of the receptor induced a strong, but transient, phosphorylation of ERK1 and ERK2. The stimulatory effect of adrenaline reached its maximum at 2-5 min and rather rapidly disappeared thereafter (Fig. 1A). Half-maximal and maximal activation was observed at about 1 and 10 M adrenaline, respectively (Fig. 1B). Phosphorylation of ERK1/2 induced by adrenaline (10 M) was fully blocked by the ␤ 2 -AR antagonist propranolol (1 M) (see Fig. 3B), as well as by the MAP kinase kinase inhibitors PD98059 (10 M) and U0126 (10 M) (data not shown).
Previous studies on activation of ERK1/2 by the ␤ 2 -AR in HEK-293 cells indicated that this G s -coupled receptor action involves cAMP, cAMP-activated PKA, the cytosolic tyrosine kinase c-Src, eventually G i proteins and G␤␥ dimers, and either Rap1A or H-Ras as final activators of the MAP kinase cascade (8, 10 -12). Similar to these studies, we observed that ERK activation by the ␤ 2 -AR was mimicked by direct activation of adenylyl cyclase with forskolin, although its action was less rapid and more sustained than that of the receptor agonist, adrenaline ( Fig. 2A). Maximal ERK1/2 phosphorylation induced by forskolin (30 M) was observed at 15-30 min and slowly declined thereafter. Furthermore, the direct adenylyl cyclase inhibitor 2Ј,5Ј-dideoxyadenosine (10 M) suppressed the stimulatory effects of both adrenaline (n ϭ 4; p Ͻ0.001) and forskolin (not shown) on phosphorylation of ERK1/2 (Fig. 2B). However, treatment of the cells with the PKA inhibitor H-89 (10 M) did not alter ERK activation induced by the ␤ 2 -AR or forskolin (not shown) at any time point measured (Fig. 2B), indicating that the ␤ 2 -AR-induced ERK activation is cAMP-dependent, but independent of PKA. To test for a possible involvement of c-Src, the cells were treated with the c-Src inhibitor PP2 (10 M) or transfected with the kinase-deficient c-Src mutant K298M c-Src, which both interfered with epidermal growth factor signaling in these cells (35). Neither of these treatments suppressed adrenaline-induced ERK1/2 phosphorylation (Fig. 3A), although it was noted that PP2 reduced both basal and adrenaline-induced phosphorylation of ERK1/2 by some 30% (n ϭ 4; p Ͻ 0.03). Treatment of the HEK-293 cells with PTX (16 h, 100 ng/ml), to interfere with potential coupling of the ␤ 2 -AR to G i proteins, did not alter phosphorylation of ERK1/2 by adrenaline either (Fig. 3B). Furthermore, scavenging of G␤␥ dimers by expression of ␤-ARK-CT (36) did not affect ERK activation by the ␤ 2 -AR (Fig. 3B). These treatments fully (PTX) or at least partially (␤-ARK-CT) blocked ERK activation by G i -coupled sphingosine 1-phosphate receptors endogenously expressed in these cells (data not shown). Finally, we asked which type of small GTPase acts as final effector of the ␤ 2 -AR signaling cascade leading to activation of the MAP kinases. As illustrated in Fig. 3C, phosphorylation of ERK1/2 induced by adrenaline was strongly reduced by expression of dominant negative S17N H-Ras (n ϭ 5; p Ͻ 0.008), but not altered by expression of S17N Rap1A. Similar data were obtained when the effects of S17N H-Ras and S17N Rap1A were examined on ERK1/2 phosphorylation induced by forskolin (data not shown). Thus, ERK activation by the ␤ 2 -AR expressed in HEK-293 cells apparently required cAMP and H-Ras, but not PKA, G i proteins, c-Src, and Rap1A.
Involvement of Epac and Rap2B in H-Ras and ERK Activation by the ␤ 2 -AR-Besides PKA enzymes, Epac proteins, which act as specific GEFs for Rap GTPases, are directly activated by cAMP (18 -20). Thus, although Rap1A was apparently not in- volved in ␤ 2 -AR-mediated ERK activation, we studied whether these proteins may mediate cAMP-induced ERK activation. For this, we used the novel membrane-permeable Epac-specific cAMP analog, 8-pCPT-2Me-cAMP (37). Treatment of HEK-293 cells with this agent caused a time-and concentration-dependent phosphorylation of ERK1/2, similar to that observed with adrenaline or forskolin. Half-maximal and maximal effects were observed at 1 and 10 M 8-pCPT-2Me-cAMP, respectively (Fig. 4A), thus very similar to the concentrations of this agent required to activate Epac1 in vitro (38,39). Maximal phosphorylation of ERK1/2 induced by 10 M 8-pCPT-2Me-cAMP was observed at 5-10 min and returned to basal levels at 30 min (Fig. 4B). Similar to adrenaline-and forskolin-induced ERK activation, phosphorylation of ERK1/2 induced by the Epacspecific cAMP analog was strongly suppressed by expression of S17N H-Ras (n ϭ 4; p Ͻ 0.002; Fig. 4C), suggesting that all of these agents activate H-Ras. Indeed, as shown in Fig. 4D, activation of the To substantiate the involvement of Epac proteins in ␤ 2 -ARmediated ERK activation in HEK-293 cells, which endogenously express Epac1 and Epac2 (Fig. 5A), the cells were transfected with wild-type Epac1 and its mutant, R279K Epac1, which contains a single amino acid substitution in its cAMP-binding pocket and which has been shown to inhibit cAMP-induced Rap1A activation (19). Whereas overexpression of wild-type Epac1 enhanced the phosphorylation of ERK1/2 induced by the ␤ 2 -AR, the receptor response was almost fully blunted (n ϭ 4; p Ͻ 0.003) by expression of R279K Epac1 (Fig.   5A). Furthermore, expression of R279K Epac1 suppressed (n ϭ 4; p Ͻ 0.003) the activation of H-Ras induced by the agonistactivated ␤ 2 -AR (Fig. 5C). As Epac proteins act as GEFs for Rap GTPases, but not H-Ras (18 -20), we assumed that the cAMP-activated Epac primarily activates a Rap GTPase, which then induces Ras and ERK activation. In line with this assumption, expression of RapGAPII in HEK-293 cells to inhibit signaling by Rap GTPases (40) suppressed (n ϭ 3-5; p Ͻ 0.002) phosphorylation of ERK1/2 (Fig. 5B) as well as activation of H-Ras (Fig. 5C) induced by the ␤ 2 -AR.
We have recently reported that the ␤ 2 -AR induces Epac-mediated activation of the Rap GTPase Rap2B, which then leads to stimulation of PLC-⑀ (31). Therefore, we first examined whether Rap2B is also involved in H-Ras and ERK activation by the ␤ 2 -AR. As reported before (31,41) and shown in Fig. 6, A and B, activation of the ␤ 2 -AR by adrenaline (10 M) or stimulation of the cells with forskolin (30 M) or 8-pCPT-2Me-cAMP (10 M) induced GTP loading of Rap2B. Furthermore, similar to ERK and H-Ras activation by these agents (see Fig. 5), activation of Rap2B was suppressed by expression of R279K Epac1 (n ϭ 4; p Ͻ 0.003) and RapGAPII (n ϭ 5; p Ͻ 0.004) and enhanced by overexpression of Epac1 (n ϭ 4; p Ͻ 0.005) (Fig. 6, A and B). This close correlation suggested that Rap2B participates in ERK and H-Ras activation by the ␤ 2 -AR. In fact, as shown in Fig. 6C, phosphorylation of ERK1/2 induced by the ␤ 2 -AR alone or when coexpressed with Epac1 was strongly or fully suppressed by expression of dominant negative S17N Rap2B (n ϭ 3-6; p Ͻ 0.002). In addition, expression of this Rap2B mutant suppressed activation of H-Ras induced by the ␤ 2 -AR (Fig. 6D). Although these data suggested that Rap2B acts upstream of H-Ras, it had to be considered that Rap2B may act independently of H-Ras in ␤ 2 -AR-mediated ERK activation. To study this, we expressed G12V Rap2B in HEK-293 cells and examined whether this constitutively active Rap2B mutant activates H-Ras as well as ERK and whether ERK activation is dependent on H-Ras. As shown in Fig.  6D, expression of G12V Rap2B, in fact, caused activation of H-Ras as well as phosphorylation of ERK1/2. Most important, ERK1/2 phosphorylation induced by G12V Rap2B was suppressed by co-expression of S17N H-Ras. Collectively, these data suggested that H-Ras and ERK activation by the ␤ 2 -AR is mediated by cAMP-activated Epac proteins and subsequent activation of Rap2B.
PLC-⑀ and Ca 2ϩ Dependence of H-Ras and ERK Activation by the ␤ 2 -AR-We then examined whether, in addition to Epac and Rap2B, the Rap2B effector PLC-⑀ (31,35) participates in H-Ras and ERK activation by the ␤ 2 -AR. As shown in Fig. 7, overexpression of wild-type PLC-⑀ (n ϭ 5-6; p Ͻ 0.005), but not PLC-␤1 or PLC-␦1 (data not shown), increased phosphorylation of ERK1/2 as well as activation of H-Ras induced by adrenaline. The stimulatory effect of PLC-⑀ on ERK activation was blocked by co-expression of S17N H-Ras (data not shown). These data suggested that Rap2B-activated PLC-⑀ mediates H-Ras and ERK activation by the ␤ 2 -AR. As H-Ras is not involved in ␤ 2 -AR-mediated PLC-⑀ stimulation (31), we examined by which mechanism PLC-⑀ controls H-Ras and ERK activation by the ␤ 2 -AR. One possibility was that PLC-⑀ via its CDC25 domain may directly activate Ras (27). To study this, we generated a PLC-⑀ mutant lacking the CDC25 motif (⌬CDC25 PLC-⑀), expressed it in HEK-293 cells, and analyzed its effects on ERK and H-Ras activation by the ␤ 2 -AR. PLC activity measurements indicated that expression of ⌬CDC25 PLC-⑀ increased ␤ 2 -AR-mediated PLC stimulation similar to wild-type PLC-⑀ (Ref. 31; data not shown). Expression of ⌬CDC25 PLC-⑀ did not decrease but, similar to wild-type PLC-⑀, enhanced phosphorylation of ERK1/2 and activation of H-Ras induced by the ␤ 2 -AR (n ϭ 6; p Ͻ 0.003) (Fig. 7). In contrast, expression of lipase-inactive H1144L PLC-⑀ (27), which inhibited PLC and Ca 2ϩ signaling in HEK-293 cells (35), strongly reduced (n ϭ 4; p Ͻ 0.001) the adrenaline-induced ERK1/2 phosphorylation and H-Ras activation. Similar changes were observed when the PLC-⑀ enzymes were examined for their effects on phosphorylation of ERK1/2 induced by forskolin (data not shown). These data thus suggest that Rap2B-activated PLC-⑀ acts upstream of H-Ras and that H-Ras activation by the ␤ 2 -AR requires the lipase activity of PLC-⑀. In accordance, expression of either wild-type PLC-⑀ or its mutants ⌬CDC25 PLC-⑀ and H1144L PLC-⑀ did not alter activation of Rap2B by the ␤ 2 -AR (Fig. 7).
(31) and activation of protein kinase C (PKC), determined as translocation of PKC-␣ to the plasma membrane (data not shown). To study which of these two PLC-triggered cellular responses is responsible for ERK and H-Ras activation by the ␤ 2 -AR, we treated the cells with Gö 6976, an inhibitor of conventional PKC isoforms, and BAPTA/AM, a chelator of intracellular Ca 2ϩ . As illustrated in Fig. 8A, Gö 6976 (100 nM) altered neither phosphorylation of ERK1/2 nor activation of H-Ras induced by adrenaline. In contrast, chelation of intracellular Ca 2ϩ with BAPTA/AM (20 M) almost fully blunted ERK1/2 phosphorylation (n ϭ 4; p Ͻ 0.002) as well as H-Ras activation (n ϭ 5; p Ͻ 0.001) induced by the ␤ 2 -AR. Similar effects of BAPTA/AM were observed on H-Ras and ERK activation induced by the Epac-specific cAMP analog, 8-pCPT-2Me-cAMP (10 M) (Fig. 8B). These data suggested that activation of H-Ras and subsequent activation of ERK1/2 by the ␤ 2 -AR is mediated by a Ca 2ϩ -regulated GEF. Some of the Ras-GEFs, in particular the RasGRFs, are known to be regulated by Ca 2ϩ /calmodulin (24 -26). Therefore, we treated the cells with the calmodulin antagonists, calmidazolium, fluphenazine, and W7. As illustrated in Fig. 8C for HEK-293 cells treated with W7 (100 M), this calmodulin antagonist altered neither phosphorylation of ERK1/2 nor activation of H-Ras induced by adrenaline or 8-pCPT-2Me-cAMP. Similar data were obtained in cells treated with calmidazolium and fluphenazine (10 M each; data not shown). These data suggested that the Ca 2ϩ /calmodulin-activated RasGRFs may not act as triggers for H-Ras and subsequent ERK activation by the ␤ 2 -AR. A second group of putatively Ca 2ϩ -regulated Ras-GEFs is the Ras-GRP family, RasGRP1, RasGRP2, and RasGRP3, which contain Ca 2ϩ -binding motifs (24 -26). Thus, we expressed these GEFs in HEK-293 cells and measured their effects on ERK and H-Ras activation by the ␤ 2 -AR. As illustrated in Fig. 9, none of the RasGRPs affected basal unstimulated ERK and H-Ras activities. Furthermore, expression of RasGRP3 was without any effect on the ␤ 2 -AR responses. In contrast, upon expression of RasGRP1, phosphorylation of ERK1/2 as well as activation of H-Ras induced by adrenaline were strongly increased. Expression of RasGRP2 only moderately increased these cellular responses.
Epac-and Ca 2ϩ -controlled H-Ras and ERK Activation by PGE 1 in N1E-115 Neuroblastoma Cells-PGE 1 acting on an endogenously expressed prostanoid receptor in N1E-115 neuroblastoma cells has also been reported to stimulate PLC-⑀ via an Epac-and Rap2B-dependent mechanism (31). It was therefore of interest to know whether PGE 1 may induce H-Ras and ERK activation in N1E-115 neuroblastoma cells by a similar mechanism as described above for the ␤ 2 -AR expressed in HEK-293 cells. As summarized in Fig. 10, stimulation of the neuroblastoma cells with PGE 1 (1 M) induced phosphorylation of ERK1/2 and activation of H-Ras. Phosphorylation of ERK1/2 induced by PGE 1 was suppressed by expression of dominant negative S17N H-Ras, but not affected by the PKA inhibitor H-89 (10 M) (Fig. 10A), which also had no effect on H-Ras activation (Fig. 10B). Similar data on ERK and H-Ras activation were obtained when the neuroblastoma cells, which endogenously express Epac1 and Epac2 (data not shown), were stimulated with 8-pCPT-2Me-cAMP (10 M). Furthermore, activation of H-Ras, as well as phosphorylation of ERK1/2 induced by PGE 1 or 8-pCPT-2Me-cAMP, was strongly reduced by treatment of the cells with the intracellular Ca 2ϩ chelator BAPTA/AM (20 M), but not affected by the calmodulin antagonist W7 (100 M) (Fig. 10, B and C). DISCUSSION Several distinct pathways and signaling components have been described as involved in activation of the ERK1/2 MAP kinases by cAMP and cAMP-producing G s -coupled receptors. In many cases, the cAMP or G s -coupled receptor-induced ERK activation was shown to be mediated by PKA, the major cellular cAMP target (6,7). However, several recent reports demonstrate that G s -coupled receptors can also induce ERK activation in a PKA-independent manner (15)(16)(17)(21)(22)(23). Some, but not all, types of this PKA-independent ERK activation are reported to involve Epac proteins (17,(21)(22)(23), which are directly activated by cAMP, independent of PKA, and act as specific GEFs for Rap GTPases (18 -20). Furthermore, for both types of cAMP-induced ERK activation, PKA-dependent and -independent, evidence is provided for either Ras or Rap1 acting as final effectors of the MAP kinase cascade (6 -8, 10, 15-17, 23).
Here we report on a novel pathway for signaling of G scoupled receptors to ERK activation. This pathway is inde- pendent of PKA but mediated by cAMP-activated Epac proteins, which via activation of Rap2B and subsequent stimulation of PLC-⑀ induce [Ca 2ϩ ] i increase, finally resulting in activation of H-Ras as a trigger of the MAP kinase cascade. Importantly, it was demonstrated that this pathway of H-Ras and ERK activation is used by two distinct G s -coupled receptors and in two cell types, i.e. the ␤ 2 -AR heterologously expressed in HEK-293 cells and the receptor for PGE 1 endogenously expressed in N1E-115 neuroblastoma cells. Activation of H-Ras and ERK by the two G s -coupled receptors was not affected by PKA inhibition but was fully mimicked in both cell types, which endogenously express Epac1 and Epac2, by the Epac-specific cAMP analog 8-pCPT-2Me-cAMP at similar concentrations of this agent shown before to activate Epac1 in vitro (37)(38)(39). In support of these findings, receptor-mediated acti-vation of H-Ras and ERK was suppressed by expression of R279K Epac1, a cAMP-binding-deficient Epac1 mutant (19). Furthermore, in line with the known function of Epac proteins, i.e. specific GEFs for Rap GTPases (18 -20), receptor-mediated H-Ras and ERK activation was suppressed by inhibition of Rap GTPase activities by expression of RapGAPII. Finally, similar to receptor-mediated ERK activation, ERK activation induced by the Epac-specific cAMP analog was suppressed by expression of dominant negative S17N H-Ras. Activation of H-Ras and consequently of the ERK MAP kinases by the cAMPactivated Epac proteins was apparently mediated by the Rap GTPase Rap2B, resulting in stimulation of PLC-⑀ and subsequent [Ca 2ϩ ] i increase. First, similar to receptor agonists (31), direct activation of Epac proteins by 8-pCPT-2Me-cAMP induced activation of Rap2B, and this Rap2B activation was suppressed by expression of R279K Epac1 and RapGAPII. Second, receptor-mediated H-Ras and ERK activation was suppressed by expression of dominant negative S17N Rap2B, but not S17N Rap1A, whereas expression of constitutively active G12V Rap2B resulted in H-Ras and H-Ras-dependent ERK activation. Third, receptor-mediated H-Ras and ERK activation was apparently dependent on stimulation of Rap2B-activated PLC-⑀ (31,35). We very recently reported that not only G s -coupled receptors but also stimulation of HEK-293 cells and N1E-115 neuroblastoma cells with 8-pCPT-2Me-cAMP or expression of constitutively active Rap2B results in marked PLC-⑀ stimulation (41). As expression of lipase-inactive H1144L PLC-⑀ suppressed ␤ 2 -AR-mediated H-Ras and ERK activation, whereas overexpression of wild-type PLC-⑀ had the opposite effect, it can be concluded that the lipase activity of the Rap2B-activated PLC-⑀ is required for H-Ras and ERK activation. Initially, we considered that PLC-⑀ may induce H-Ras and ERK activation via its CDC25 domain, which can act as GEF for Ras and Rap GTPases (27)(28)(29). However, expression of the PLC-⑀ mutant deficient in the CDC25 domain ⌬CDC25 PLC-⑀, in contrast to the lipase-inactive mutant, enhanced the receptor-induced H-Ras and ERK activation similar to overexpression of wild-type PLC-⑀. Finally, it has been demonstrated that chelation of intracellular Ca 2ϩ , i.e. prevention of one of the major consequences of PLC-⑀ stimulation (31,35), greatly reduces or abolishes H-Ras and ERK activation induced by either the G s -coupled receptor agonists or the Epac-specific cAMP analog in HEK-293 cells and N1E-115 neuroblastoma cells. The Ca 2ϩ dependence of H-Ras and ERK activation suggested that a Ca 2ϩ -regulated Ras-GEF is involved in the G scoupled receptor actions. To corroborate this hypothesis, we overexpressed RasGRP1, RasGRP2, and RasGRP3, also termed CalDAG-GEF II, I, and III, respectively, which contain Ca 2ϩ -and diacylglycerol (DAG)-binding domains (24 -26). Interestingly, H-Ras and ERK activation induced by the ␤ 2 -AR was strongly enhanced by overexpression of RasGRP1, whereas RasGRP3 and RasGRP2 had no or only small effects. These three GEFs are known to be activated by DAG and phorbol esters (24 -26, 42-44). However, as only RasGRP1 enhanced the receptor-induced and PLC-⑀-mediated Ras activation, it is unlikely that the DAG formed by the PLC-⑀ plays a major role in Ras activation. On the other hand, data on Ca 2ϩ regulation of the Ras-GEF activity of the RasGRPs are either not available (RasGRP3) or Ca 2ϩ has even been reported to inhibit N-Ras while enhancing Rap1 activation by RasGRP2 (45). For Ras-GRP1, only Ca 2ϩ binding has been reported (42). Thus, RasGRP1 is a possible candidate for GEF mediating H-Ras activation by the G s -coupled receptors. Other candidate GEFs are the Ca 2ϩ - activated Ras-specific GEFs, GRF1 and GRF2 (24 -26). These GEFs, however, are primarily regulated by Ca 2ϩ /calmodulin (24 -26). As none of the three calmodulin antagonists studied (W7, calmidazolium, and fluphenazine) affected H-Ras and ERK activation by the G s -coupled receptors, it is rather unlikely that the Ca 2ϩ /calmodulin-regulated RasGRFs act as triggers for the Ca 2ϩ -dependent H-Ras activation presented herein. Studies are in progress to identify the specific GEF responsible for H-Ras activation, which may even be different in HEK-293 cells and N1E-115 neuroblastoma cells.
When comparing the data presented in this report on ERK activation by G s -coupled receptors and those published before for these and other G s -coupled receptors, it is obvious that there is a multiplicity of signaling pathways for these groups of receptors in different cell types or cell lines and even for a single receptor expressed in the "same" cell line, e.g. the ␤ 2 -AR expressed in HEK-293 cells (see Refs. 8 and 10 and this report) as recently pointed out (46). Different results have also been reported for the role of Epac proteins and Rap GTPases in ERK activation. For example, we have demonstrated here that the Epac-specific cAMP analog 8-pCPT-2-Me-cAMP potently and efficiently leads via Epac-activated Rap2B and PLC-⑀ to H-Ras and ERK activation in HEK-293 cells and N1E-115 neuroblastoma cells. In contrast, others report that this agent, although inducing activation of Rap1, which was also observed by us (data not shown), did not cause ERK activation when applied to Chinese hamster ovary, OVCAR3, PC12, and HEK-293T cells, which led the authors to the rather general conclusion that Epac and Rap1 are not involved in ERK activation by cAMP and G s -coupled receptors (37). The specific differences are currently unclear that account for this variability in cellular responses to this agent as well as the receptors. From our data, we conclude that G s -coupled receptors can induce activation of H-Ras and the ERK1/2 MAP kinases by an Epac-and Ca 2ϩcontrolled pathway involving Rap2B and its effector PLC-⑀ and used by two distinct receptors in two different cell types, thus adding a novel, rather well defined pathway to the apparent redundancy of signaling mechanisms from these receptors to the MAP kinases.