Ras-dependent ERK Activation by the Human Gs-coupled Serotonin Receptors 5-HT4(b) and 5-HT7(a) *

Receptor tyrosine kinases activate mitogen-activated protein (MAP) kinases through Ras, Raf-1, and MEK. Receptor tyrosine kinases can be transactivated by G protein-coupled receptors coupling to Gi and Gq. The human G protein-coupled serotonin receptors 5-HT4(b) and 5-HT7(a) couple to Gs and elevate intracellular cAMP. Certain Gs-coupled receptors have been shown to activate MAP kinases through a protein kinase A- and Rap1-dependent pathway. We report the activation of the extracellular signal-regulated kinases (ERKs) 1 and 2 (p44 and p42 MAP kinase) through the human serotonin receptors 5-HT4(b) and 5-HT7(a) in COS-7 and human embryonic kidney HEK293 cells. In transfected HEK293 cells, 5-HT-induced activation of ERK1/2 is sensitive to H89, which indicates a role for protein kinase A. The observed activation of ERK1/2 does not require transactivation of epidermal growth factor receptors. Furthermore, 5-HT induced activation of both Ras and Rap1. Whereas the presence of Rap1GAP1 did not influence the 5-HT-mediated activation of ERK1/2, the activation of ERK1/2 was abolished in the presence of dominant negative Ras (RasN17). ERK1/2 activation was reduced in the presence of “dominant negative” Raf1 (RafS621A) and slightly reduced by dominant negative B-Raf, indicating the involvement of one or more Raf isoforms. These findings suggest that activation of ERK1/2 through the human Gs-coupled serotonin receptors 5-HT4(b) and 5-HT7(a) in HEK293 cells is dependent on Ras, but independent of Rap1.

Extracellular signals such as growth factors and hormones regulate cell proliferation. In general, growth factors activate receptor tyrosine kinases (RTKs), 1 whereas most hormones activate G protein-coupled receptors (GPCRs). Intracellular cascades such as the mitogen-activated protein (MAP) kinase, also called the extracellular signal-regulated kinase (ERK), cascade, are mainly activated as a consequence of agonist stimulation of RTKs. ERK1/2 are serine/threonine protein kinases activated upon dual phosphorylation by the MAP kinase kinase MEK (1), which is phosphorylated and activated by MAP kinase kinase kinases of the Raf family (2). In mammals, all three Raf isoforms expressed (A-Raf, B-Raf, and Raf-1) are cytosolic in resting cells and become activated upon recruitment to the plasma membrane by activated, small G proteins of the Ras family (H-Ras, K-Ras, N-Ras, and Rap1). Ras, in its active, GTP-bound conformation, may activate all isoforms of Raf. The levels of GTP-bound Ras are controlled by the actions of guanine nucleotide exchange factors and GTPase-activating proteins (GAPs). The two major Ras-specific guanine nucleotide exchange factors are Son of sevenless (SOS) (3) and Ras-GRF1 (CDC25 Mm ) (4). SOS is mainly thought to activate Ras as a consequence of RTK stimulation, whereas Ras-GRF1 activates Ras in response to Ca 2ϩ signaling and GPCR-mediated signals (5)(6)(7)(8).
Serotonin (5-hydroxytryptamine (5-HT)) mediates its diverse physiological effects through at least 14 different receptors, of which 13 are GPCRs or so-called seven transmembrane-spanning receptors (9). The 5-HT receptors, each encoded by a separate gene, have been grouped into families called 5-HT 1 to 5-HT 7 , with some of the families containing several members. The human serotonin receptors 5-HT 4 and 5-HT 7 couple to the heterotrimeric G protein G s and exist in multiple splice variants (10,11). Stimulation of these receptors leads to activation of adenylyl cyclase and a rapid increase in the formation of the intracellular second messenger cAMP. Elevated levels of cAMP have several intracellular effects, e.g. activation of cAMP-dependent protein kinase (PKA) and exchange proteins directly activated by cAMP (Epacs), guanine nucleotide exchange factors specific for Rap. Both mechanisms have been claimed to be involved in cAMP-dependent activation of Rap1 (12)(13)(14).
The main target for cAMP is PKA, which has cell typespecific effects on the ERK1/2 cascade. In several cell types, Ras-dependent ERK1/2 activation is antagonized by PKA (15)(16)(17)(18). In other cell types, PKA appears to activate the ERK1/2 cascade via Rap1 and B-Raf (12,19). In human embryonic kidney HEK293 cells, isoproterenol stimulation of the endogenous G s -coupled ␤ 2 -adrenergic receptor (␤ 2 -AR) activates Ras and Rap1, but only Rap1 is able to mediate activation of ERK1/2 by recruiting B-Raf to the membrane and thereafter activate MEK (20). However, at overexpressed levels of ␤ 2 -ARs, ERK1/2 can be activated through the G␤␥ subunits of pertussis toxin-sensitive G proteins, interpreted as a switch in coupling from G s to G i due to phosphorylation of the ␤ 2 -AR by PKA (21).
Upon stimulation of G i - (22) and G q -coupled (23) receptors, activation of MAP kinases is induced by G␤␥ subunits that mediate Ras activation through proteins such as the tyrosine kinase c-Src, Grb2, and SOS (24). Both phosphoinositide 3-kinase ␥ (25) and the calcium-sensitive kinase PYK2 (26) have been implicated in the activation of c-Src. The G␣ i and G␣ q subunits may also mediate ERK1/2 activation by binding to Rap1GAP1, a selective Rap1 GTPase-activating protein (27,28). In these cases, active Rap1 antagonizes the function of Ras, and inhibition of Rap1 by Rap1GAP1 allows Ras to signal to ERK1/2. Furthermore, in several cell types, GPCRs coupling to G i and G q have been reported to transactivate the epidermal growth factor receptor (EGFR) and thereby lead to activation of ERK1/2 (29 -31).
A G s -coupled 5-HT receptor has been implicated in ERK activation and LTP induction in Aplysia (32,33). In mouse cortical neurons, several monoamines including 5-HT, as well as forskolin, activated ERK1/2 and increased the level of GTPbound Ras, but the 5-HT receptor involved was not characterized (34). In cultured rat hippocampal neurons, 5-HT activated ERK1/2 through 5-HT 7 receptors, but the mechanism was not characterized (35). In mammalian brain, both 5-HT 4 and 5-HT 7 receptors are found in the hippocampus (36,37), and these receptors are both believed to be important in learning and memory (38). Thus, there is a need to determine whether 5-HT 4 and 5-HT 7 receptors can activate ERK, as well as the mechanism for such activation. We have previously characterized activation of adenylyl cyclase by different splice variants of the human G s -coupled 5-HT 4 (39) and 5-HT 7 (40) receptors. We now demonstrate that both the human G s -coupled serotonin receptors 5-HT 4(b) and 5-HT 7(a) , when transiently expressed in HEK293 and COS-7 cells, activate ERK1/2 through a mechanism dependent on Ras and independent of Rap1.
Cell Culture and Transfection-COS-7 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum and 50 mg/liter gentamicin at 37°C in a humidified atmosphere of 5% CO 2 in air and transfected at 60 -70% confluence with the 5-HT 4(b) or 5-HT 7(a) receptors using DEAE-dextran according to the manufacturer's protocol. HEK293 cells were cultured in UltraCULTURE™ general purpose serum-free medium with 2 mM L-glutamine and 100 units/ml penicillin-streptomycin at 37°C in a humidified atmosphere of 5% CO 2 in air and transfected at 60 -70% confluence with the indicated cDNA(s) using LipofectAMINE according to the manufacturer's protocol. When necessary, empty vector (pcDNA3.1(Ϫ); Invitrogen) was included to assure that each dish received the same amount of DNA (1.0 and 2.9 g of plasmid DNA per 35-mm and 60-mm dish, respectively). If not otherwise indicated, cells were kept in serum-free medium as described until treatment and lysis, ϳ48 h after transfection. When indicated, cells were preincubated with 20 M H89 for 25 min, 5 M PD98059 for 20 min, or 1 M PD153035 for 18 h. Cells were stimulated with agonist (10 M 5-HT or 5 nM EGF) for 5 min, if not otherwise indicated. All experiments were carried out in doublets with at least three replicates, if not otherwise indicated.
Western Blotting-Equal amounts of cell lysate proteins were separated by SDS-PAGE and electroblotted onto polyvinylidene difluoride membranes. The membranes were incubated with primary antibodies (anti-phospho-ERK1/2, 1:2,000; anti-Ras, 1:1,000; anti-Rap1, 1:500) in 5% nonfat dry milk in phosphate-buffered saline-0.05% Tween and thereafter incubated with corresponding horseradish peroxidase-conjugated secondary antibodies. The immobilized horseradish peroxidaseconjugated secondary antibodies were visualized with Bio West chemiluminescent substrate and analyzed with a UVP BioChemie system. After detection of phospho-ERK1/2, membranes were stripped by treatment with 0.5 M NaOH for 5 min at room temperature and extensive washing with distilled H 2 O and reprobed with anti-ERK1/2 antibodies (1:10,000) to confirm equal loading.
Phospho-ERK1/2 Assay-The cells were cultured in 35-mm dishes, transfected and stimulated with agonist as described, lysed in ice-cold cell lysis buffer (1% SDS, 1 mM Na 3 VO 4 , 50 mM Tris-HCl, pH 7.4 at room temperature), scraped with a rubber policeman, sheared through a 25-gauge syringe, and immediately snap frozen in liquid N 2 . The thawed cell lysates were cleared at 13,000 ϫ g at 4°C, and the protein concentrations in the supernatants were quantified using the BC assay protein quantitation kit using BSA as standard. Equal amounts of protein were prepared for separation by SDS-PAGE.
Pull-Down Experiments in HEK293 Cells-The cells were assayed as described by van Triest et al. (43). Briefly, HEK293 cells were maintained in 60-mm dishes, transfected, and stimulated as described. The cells were lysed in ice-cold pull-down lysis buffer (10% glycerol, 1% Nonidet P-40, 50 mM Tris-HCl, pH 7.5 at room temperature, 200 mM NaCl, 2 mM MgCl 2 , 1 mM phenylmethylsulfonyl fluoride, 2 g/ml aprotinin, 1 g/ml leupeptin, and 10 g/ml trypsin inhibitor). Equal amounts of cleared cell lysates were incubated at 4°C on a tumbler for 1 h with GST-Raf1-RBD or GST-RalGDS-RBD constructs precoupled to glutathione-Sepharose beads for Ras or Rap1 pull-down experiments, respectively. The beads were pelleted and washed four times with ice-cold pull-down lysis buffer. The proteins were eluted from the beads by boiling for 5 min in 1ϫ loading buffer. The protein samples were loaded and resolved on 12.5% SDS-polyacrylamide gels. Western blotting was performed as described.

ERK1/2 Activation via 5-HT 4(b) and 5-HT 7(a) Receptors in COS-7 and HEK293
Cells-Initial experiments indicated that treatment of COS-7 cells transiently transfected with G s -coupled human 5-HT 4(b) or 5-HT 7(a) receptors with 10 M 5-HT induced dual phosphorylation and thus activation of ERK1/2. Similar results were obtained with HEK293 cells. Because HEK293 cells do not express endogenous 5-HT receptors and had a lower background of ERK1/2 phosphorylation, these cells were selected for further study. Based on the potency of 5-HT to stimulate adenylyl cyclase activity through human 5-HT 4(b) (39) or 5-HT 7(a) (40) receptors and 5-HT concentration-response experiments with ERK1/2 phosphorylation (data not shown), a concentration of 10 M 5-HT was chosen for the remaining experiments. A high resolution time course showed that the maximal 5-HT-induced phosphorylation of ERK1/2 through 5-HT 4(b) receptors occurred after 5 min (Fig. 1A). Time courses with more extended periods of stimulation showed that 5-HTinduced activation of ERK1/2 through 5-HT 4(b) (data not shown) and 5-HT 7(a) (Fig. 1B) receptors peaked after ϳ5 min and then decreased. There was no second peak or sustained activation of ERK1/2.
Activation of ERK1/2 through 5-HT 4(b) and 5-HT 7(a) Receptors Is PKA-and MEK-dependent-All splice variants of the human serotonin receptors 5-HT 4 and 5-HT 7 are G s -coupled receptors that upon agonist stimulation induce a rapid increase in the formation of cAMP, leading to activation of PKA or Epac. To examine the role of PKA in the signaling pathway, HEK293 cells transiently transfected with either 5-HT 4(b) or 5-HT 7(a) were preincubated with the PKA inhibitor H89. Pretreatment with H89 led to a partial inhibition of 5-HT-induced activation of ERK1/2, as shown in Fig. 2A for the 5-HT 4(b) receptor. However, a more efficient inhibition was observed when the transiently transfected cells were maintained in serum-free Dulbecco's modified Eagle's medium for 16 -18 h before treatment with 5-HT, as shown in Fig. 2B for the 5-HT 7(a) receptor. MEK plays a central role in the classical signaling pathway from cell surface receptors to ERK1/2. Pretreatment with the selective MEK inhibitor PD98059, as described under "Experimental Procedures," blocked 5-HT-induced activation of ERK1/2 in cells transfected with 5-HT 4(b) receptors (data not shown) or 5-HT 7(a) (Fig. 2C). Taken together, the above data show that 5-HT-induced activation of ERK1/2 by human 5-HT 4(b) and 5-HT 7(a) receptors is mediated through a pathway dependent on MEK and at least partially dependent on PKA.
Activation of ERK1/2 through 5-HT 4(b) and 5-HT 7(a) Receptors Does Not Require EGFR Phosphorylation-Several studies have shown that G s -coupled receptors may transactivate RTKs (29 -31). Transiently transfected HEK293 cells were preincubated with the EGFR kinase inhibitor PD153035 to determine possible involvement of EGFR transactivation. The presence of PD153035 did not influence the 5-HT-induced activation of ERK1/2 through 5-HT 4(b) (data not shown) or 5-HT 7(a) receptors (Fig. 3A). As a positive control, the presence of PD153035 was shown to completely eliminate activation of ERK1/2 by EGF in nontransfected HEK293 cells (Fig. 3B). The above data suggest that activation of ERK1/2 through 5-HT 4(b) and 5-HT 7(a) receptors does not require transactivation of EGFRs.
5-HT-induced Activation of ERK1/2 Is Independent of Rap1-Activation of ERK1/2 by endogenous ␤ 2 -ARs in HEK293 cells has been reported to be PKA-dependent and mediated through a pathway independent of Ras. Although both Ras and Rap1 were activated by isoproterenol treatment, only Rap1 was capable of coupling to a Raf isoform and activating ERK1/2 (20). The Rap1 interfering mutant, Rap1N17, and the Rap1specific GTPase-activating protein, Rap1GAP1, were used to investigate the role of Rap1 in the activation of ERK1/2 through 5-HT 4(b) and 5-HT 7(a) receptors in co-transfection experiments. The presence of Rap1N17 (data not shown) or Rap1GAP1 (Fig. 4A) did not influence the 5-HT-induced activation of ERK1/2, although activation of Rap1 was abolished, as evidenced by pull-down of activated Rap1 (Rap1-GTP) with the RBD of RalGDS (Fig. 4B).

5-HT Induces Ras Activation through 5-HT 4(b) and 5-HT 7(a)
Receptors-GPCRs may activate the small GTPase Ras, which plays a central role in the activation of ERK1/2 through several different types of receptors. Therefore, the RBD of Raf1, which only binds activated (GTP-bound) Ras, was used in pull-down experiments to test whether 5-HT could induce activation of Ras. Such experiments documented activation of Ras by 5-HT through both 5-HT 4(b) receptors (data not shown) and 5-HT 7(a) receptors (Fig. 5A). Nontransfected HEK293 cells were treated with EGF as a positive control for Ras activation (Fig. 5A).

Activation of ERK1/2 through 5-HT 4(b) and 5-HT 7(a) Receptors Is
Ras-dependent-Dominant negative and constitutively active mutants of Ras were used to further investigate the role of Ras in the 5-HT-induced activation of ERK1/2. Activation of ERK1/2 through 5-HT 4(b) and 5-HT 7(a) receptors was highly reduced in the presence of the dominant negative variant of Ras, RasN17 (Fig. 5B). The data in Figs. 4 and 5 suggest that the 5-HT-induced activation of ERK1/2 is dependent on Ras, but independent of Rap1. When cells co-transfected with 5-HT 4(b) receptors and the constitutively active variant of Ras, RasV12, were treated with 5-HT, an additional activation of ERK1/2 was observed compared with the constitutive activation of ERK1/2 in these cells.
To determine the role of B-Raf, HEK293 cells were co-transfected with 5-HT 7(a) receptors and dominant negative (N-B-Raf) or constitutively active (CAAX-B-Raf) B-Raf. The dominant negative mutant of B-Raf corresponds to the N-terminal region of quail B-Raf (amino acids 1-443) (42). In several experiments, the presence of N-B-Raf had only a weak inhibitory effect on 5-HT-induced activation of ERK1/2 (Fig. 7A). In general, the inhibitory effect of N-B-Raf was weaker than that of RafS621A. Additional activation of ERK1/2 was induced by 5-HT in the presence of CAAX-B-Raf, as was the case with RafCAAX (Fig. 7A). Overexpression of RafS621A and N-B-Raf was verified by Western blot analysis using anti-Raf1 (Fig. 6C) and anti-B-Raf (Fig. 7B) antibodies, respectively. show representative Western blots probed as described in Fig. 1.

DISCUSSION
In this paper, we report the activation of ERK1/2 by 5-HT through the human G s -coupled serotonin receptors 5-HT 4(b) and 5-HT 7(a) transiently expressed in COS-7 and HEK293 cells. We also demonstrate that the pathway in HEK293 cells involves Ras but not Rap1. Pretreatment of transiently transfected HEK293 cells with H89 suppressed 5-HT-induced dual phosphorylation of ERK1/2, indicating a role for PKA. Intracellular signals from the large family of G proteincoupled receptors are conveyed by ␣ subunits (G␣) or ␤ and ␥ subunits (G␤␥) of heterotrimeric G proteins. Of the four main classes of heterotrimeric G proteins (G s , G i , G q , and G 12/13 ), G qand G i -coupled receptors have been shown to mediate activation of ERK1/2 through free G␤␥ subunits (44,45). G␤␥ subunits released upon stimulation of G s -coupled receptors have not been shown to mediate activation of ERK1/2, possibly because cells contain far less G s than G i or G q . However, G␣ s activates adenylyl cyclase and thereby PKA, which has been proposed to play an essential role in the ERK1/2 cascade (12,20,34). In line with this, we found that ERK1/2 activation through the G s -coupled receptors 5-HT 4(b) and 5-HT 7(a) was inhibited in cells pretreated with the PKA inhibitor H89.
In addition to activation of PKA, cAMP may lead to activation of Rap1 through the Rap1-selective exchange factors Epac1 or Epac2 (13, 14, 46). The small GTPase Rap1 is a Ras analogue, which upon isoproterenol stimulation of HEK293 cells has been reported to recruit B-Raf to the membrane and thereby activate the ERK1/2 cascade (20). We determined the role of Rap1 in the observed 5-HT-induced activation of ERK1/2 by co-transfecting HEK293 cells with either Rap1GAP1 or Rap1N17 and 5-HT receptors. Even though Rap1N17 is not a true dominant negative (47,48), it has been used to inhibit Rap1 activity in several studies (12,20,49). Therefore, we used Rap1N17 to support the data obtained with Rap1GAP1. The observed 5-HT-induced activation of ERK1/2 in HEK293 cells was not influenced by co-expression of either Rap1GAP1 or Rap1N17, although Rap1 activation was eliminated. Thus, the G s -coupled 5-HT 4(b) and 5-HT 7(a) receptors mediate activation of ERK1/2 through a Rap1-independent pathway, in contrast to what was reported for ␤ 2 -ARs (20).
The small GTPase Ras is the best characterized and serves as the main activator of the ERK1/2 cascade by linking RTKs and all isoforms of Raf. RTK-mediated activation of Ras involves a series of SH2-and SH3-dependent protein-protein interactions between tyrosine-phosphorylated receptor, the adapter proteins Shc and Grb2, and the Ras-guanine nucleotide exchange factor SOS (50,51). Certain GPCRs coupling to G i and G q mediate activation of Ras through free G␤␥ subunits, e.g. by stimulation of Shc (44,45,52). Isoproterenol stimulation of endogenous ␤ 2 -ARs in HEK293 cells was reported to activate both Rap1 and Ras, but only Rap1 was implicated in the activation of ERK1/2 (20). The mechanism of Ras activation was not addressed. Stimulation of HEK293 cells transiently transfected with 5-HT 4(b) or 5-HT 7(a) receptors also increased the amount of GTP-bound Ras, and, in addition, 5-HT-induced activation of ERK1/2 was inhibited by expression of the dominant negative mutant of Ras, RasN17. Thus, activation of ERK1/2 through G s -coupled 5-HT 4(b) and 5-HT 7(a) receptors in HEK293 cells is Ras-dependent.
Since the original manuscript was submitted for publication, two papers have been published that support the finding of Rap1-independent activation of ERK1/2 through G s -coupled receptors (53,54). Furthermore, in accordance with our finding in HEK293 cells of a Ras-dependent mechanism for activation of ERK1/2, Enserink et al. (54) report a Ras-dependent activa- tion of ERK1/2 in Chinese hamster ovary cells subsequent to elevation of intracellular cAMP.
Transactivation of RTKs has been proposed as an alternative mechanism for activation of Ras through GPCRs (55). The signal will pass the membrane three times, and activation of Ras and the ERK1/2 cascade will subsequently take place as by direct stimulation of RTKs, e.g. EGFR. Activation of ERK1/2 through 5-HT 4(b) and 5-HT 7(a) receptors was not inhibited in the presence of the EGFR kinase inhibitor PD153035, indicating that EGFR transactivation is not required.
The effects of cAMP on the ERK1/2 cascade are cell typespecific. In some cell types, PKA was found to inhibit RTKmediated activation of ERK1/2 (15)(16)(17)(18). Raf is the main target for cAMP-mediated regulation of the ERK1/2 cascade. The serine residues 43, 259, and 621 have all been suggested to be involved in the regulation of Raf1 activity (56). Phosphorylation of serine 43 does not inhibit the catalytic activity of Raf1, but it has been proposed to disrupt the normal Ras-Raf association and Raf1 activation (15). Furthermore, even though B-Raf does not have a phosphorylation site analogous to serine 43 in Raf1, elevation of intracellular cAMP levels by forskolin inhibited the EGF-dependent activation of B-Raf (57). The serine 621 residue of Raf1 has been suggested as PKA phosphorylation site. Raf1 phosphorylated at this site exhibited reduced affinity for activated Ras as well as impaired catalytic activity (58). Furthermore, mutation of serine 621 to alanine resulted in loss of catalytic activity. Thus, a serine 621 to alanine mutant, RafS621A, has been used as a "dominant negative" variant of Raf1. However, recent reports have shown that the list of residues involved in the PKA-dependent regulation of Raf1 activity may also include serine residues 233 and 338 (59,60). Based on the literature concerning cAMP-mediated regulation of Raf1, RafS621A cannot be considered a true dominant negative. Still, we found that the Ras-dependent, 5-HT-induced activation of ERK1/2 was partly inhibited by co-expression of the Raf1 interfering mutant RafS621A, implicating Raf1 in the pathway. In melanocytes, dominant negative B-Raf has been used to show that Ras and B-Raf were involved in cAMP-mediated activation of ERK1/2 (42). We found only little or no effect of co-expression of dominant negative B-Raf under conditions where RafS621A caused more pronounced reduction of ERK1/2 activation by 5-HT. Whether RafS621A is a true dominant negative or only a partial inhibitor of Raf1 activation, these data suggest that one or more isoforms of Raf are involved in the activation of ERK1/2 through 5-HT 4(b) and 5-HT 7(a) receptors in HEK293 cells.
The immediate upstream activator of ERK1/2 is MEK (61), which is involved in most pathways described for activation of ERK1/2. The presence of the MEK-specific inhibitor PD98059 inhibited 5-HT-induced activation of ERK1/2 through 5-HT 4(b) and 5-HT 7(a) receptors. These results suggest that the 5-HTinduced activation of ERK1/2 is MEK-dependent.
The high number of publications concerning activation of MAP kinases via G s -coupled receptors illustrates the complexity of this field. Several publications have shown that activation of Ras may occur after stimulation of G s -coupled receptors. However, little is known about how these receptors or agents that elevate intracellular levels of cAMP mediate activation of Ras. Src or a Src-like kinase has been suggested to be involved in the activation of ERK1/2 subsequent to stimulation of the G s -coupled A 2A -adenosine receptor (53). However, this may not account for the 5-HT-induced Ras-dependent activation of ERK1/2 in HEK293 cells because the Src-dependent mechanism was observed in Chinese hamster ovary, PC12, and NIH3T3 cells, but not in HEK293 cells. To our knowledge, a Ras-specific exchange factor activated by cAMP, analogous to Epac for Rap1, has not been published. This does not mean that a Ras-specific exchange factor directly activated by cAMP or PKA phosphorylation does not exist. Alternatively, there may be more indirect actions of PKA via known or unknown proteins that mediate activation of Ras.
The biological significance of the observed activation of ERK1/2 by 5-HT 4 and 5-HT 7 receptors is unknown. The biological effects of 5-HT are mediated by at least 14 different receptors, and there is hardly a target cell for 5-HT that expresses only one type of 5-HT receptor. Based on current knowledge about signal transduction mechanisms and ERK activation, other 5-HT receptors than 5-HT 4 and 5-HT 7 would be expected to play a more important role in ERK activation. However, examples of biological systems exist where ERK activation by G s -coupled serotonin receptors may be important. In Aplysia, activation of mitogen-activated protein kinases by serotonin is implicated in memory formation and occurs through a G scoupled 5-HT receptor (32,33). In mammalian brain, 5-HT 4 and 5-HT 7 receptors are both found in the hippocampus (36,37), and 5-HT 7 receptors were recently shown to activate ERKs in cultured hippocampal neurons (35). The importance of ERK activation for the involvement of these receptors in learning and memory (38) remains to be determined.
In conclusion, we demonstrate activation of ERK1/2 by 5-HT through the human G s -coupled 5-HT 4(b) and 5-HT 7(a) receptors in COS-7 and HEK293 cells. In HEK293 cells, the observed ERK1/2 activation was Ras-dependent but Rap1-independent and involved one or more Raf isoforms. Therefore, we suggest the following pathway for activation of ERK1/2 by 5-HT through 5-HT 4(b) and 5-HT 7(a) receptors: G s -coupled 5-HT re-ceptor33 cAMP3 PKA33 Ras3 Raf3 MEK3 ERK1/2. The unknown pathway from cAMP/PKA to Ras activation remains a challenge for future research.