A CalDAG-GEFI/Rap1/B-Raf Cassette Couples M1Muscarinic Acetylcholine Receptors to the Activation of ERK1/2*

In this study we examine signaling pathways linking the M1 subtype of muscarinic acetylcholine receptor (M1 mAChR) to activation of extracellular signal-regulated kinases (ERK) 1 and 2 in neuronal PC12D cells. We first show that activation of ERK1/2 by the M1 mAChR agonist carbachol takes place primarily via a Ras-independent pathway that depends largely upon Rap1, another small GTP-binding protein in the Ras family. Rap1 in turn activates B-Raf, an upstream activator of ERK1/2. Consistent with these results, carbachol was found to activate Rap1 more potently than Ras. Similar to other small GTP-binding proteins, activation of Rap1 requires a guanine nucleotide exchange factor (GEF) to promote its conversion from the GDP- to GTP-bound form. Using specific antibodies, we show that a recently identified Rap1 GEF,calcium- anddiacylglycerol-regulatedguanine nucleotide exchange factorI (CalDAG-GEFI), is expressed in PC12D cells and that carbachol stimulates the formation of a complex containing CalDAG-GEFI, Rap1, and activated B-Raf. Finally, we show that expression of CalDAG-GEFI antisense RNA largely blocks carbachol-stimulated activation of hemagglutinin (HA)1-tagged B-Raf and formation of the CalDAG-GEFI/Rap1/HA1-tagged B-Raf complex. Together, these data define a novel signaling pathway for M1 mAChR, where increases in Ca2+ and diacylglycerol stimulate the sequential activation of CalDAG-GEFI, Rap1, and B-Raf, resulting in the activation of MEK and ERK1/2.

of the seven-transmembrane family of receptors, which initiate intracellular signaling by activating heterogeneous G proteins. They can be classified into two functional groups as follows: subtypes M 1 , M 3 , and M 5 couple to G q/11 to activate phospholipase C (PLC)␤, and M 2 and M 4 couple to G i/o to inhibit adenylate cyclase (1)(2)(3). The M 1 subtype is widely expressed in the central nervous system, including the hippocampus and cerebral cortex, where it is thought to play a role in learning and memory (4 -7). Intracellular signaling events that are activated by M 1 mAChR include PLC␤-mediated production of diacylglycerol (DAG), which activates protein kinase C (PKC), and inositol-1,4,5-triphosphate (IP 3 ), which stimulates the release of Ca 2ϩ from the endoplasmic reticulum. This release of Ca 2ϩ is often accompanied by a sustained influx of extracellular Ca 2ϩ (3,8,9). Signaling events that occur further downstream, however, are poorly understood. One downstream event that has been reported is the activation of the mitogenactivated protein kinases (MAPK) 1 and 2, also known as extracellular signal-regulated kinases (ERK) 1 and 2 (10 -13). Activation of ERK1/2 has been suggested to be necessary for the establishment of long term potentiation of synaptic transmission (14 -16), a form of neuronal plasticity that may be the cellular basis of learning and memory (17,18). Intracellular signaling pathways linking M 1 mAChR and ERK1/2 activation in neurons, however, have not yet been elucidated.
To study mechanisms of signal transduction activated by M 1 mAChR, we have been using PC12D cells (19), a spontaneously arising variant of neuronal PC12 cells (20) that expresses M 1 and M 4 mAChRs (21). Using an M 1 mAChR-specific toxin (m1toxin) isolated from the venom of Dendroaspis angusticeps (22), we previously showed that the acetylcholine analogue carbachol induces expression of the immediate-early gene zif268 (also designated egr-1, NGF1-A, and krox-24 (23)(24)(25)) exclusively via M 1 mAChR in PC12D cells (21). We also showed that carbachol-mediated zif268 gene induction involves both the influx of extracellular Ca 2ϩ and activation of PKC (21), and that activation of ERK1/2 is also crucial for this induction (26). The signaling pathways linking influx of Ca 2ϩ and activation of PKC to the activation of ERK1/2 in PC12D cells, however, are still not defined.
In addition to this well established Ras-dependent ERK1/2 kinase pathway, the ERK1/2 kinase cascade has been proposed to be activated by the small GTP-binding protein Rap1 in cells expressing B-Raf (42)(43)(44)(45). Rap1 was first identified as a protein that antagonizes Ras-dependent activation of Raf-1 (46), but recent studies have shown that it can activate B-Raf both in vitro (47) and in vivo (42)(43)(44)(45). Activation of Rap1 and B-Raf has been proposed to be essential for the sustained activation of ERK1/2 kinase in PC12 cells by NGF (43). The ability of Rap1 to activate B-Raf is still controversial, however, since activated Rap1 does not activate ERK1/2 in some cell lines (48,49).
Several pathways for Rap1 activation have been shown to be mediated by GEFs that respond to specific intracellular signals (50). For example, increases in the intrinsic tyrosine kinase activities of growth factor receptors (e.g. TrkA, EGF receptors) activate Rap1 via the GEF C3G (51,52). The GEFs nRap GEP/PDZ-GEFI/Ra-GEF activate Rap1 upon binding to cell adhesion molecules via their PDZ domains (53)(54)(55). Increases in intracellular concentrations of cAMP activate the Rap1 GEFs Epac ((56), also designated cAMP-GEFI (57)) and cAMP-GEFII (57), which are activated by direct binding of cAMP. Increases in intracellular levels of Ca 2ϩ and DAG activate Rap1 by activating CalDAG-GEFI ( (58), structurally related to RasGRP/HCDC25L (59,60)) and CalDAG-GEFIII (which is actually a more effective GEF for Ras (61)). Among these Rap1 GEFs, CalDAG-GEFI (calcium-and diacylglycerol-regulated guanine nucleotide exchange factor I (58)), which has binding sites for both Ca 2ϩ and DAG, seemed to have just the right properties to link PLC␤-stimulated increases in Ca 2ϩ and DAG to the activation of the ERK1/2 cascade. Until now, however, there have been no studies of the function of endogenous CalDAG-GEFI.
In the present study we show that CalDAG-GEFI plays a central role in M 1 mAChR-mediated activation of the ERK1/2 cascade in PC12D cells. First, we show that the mAChR agonist carbachol activates ERK1/2 largely via a Ras-independent pathway. Second, we show that carbachol rapidly activates B-Raf, an upstream activator of ERK1/2, and that this activation depends upon the activation of Rap1. Third, we use specific antibodies to show that CalDAG-GEFI is expressed in PC12D cells and that a complex containing CalDAG-GEFI, Rap1, and activated B-Raf is formed following stimulation of M 1 mAChR. Finally, we show that carbachol-mediated activation of hemagglutinin (HA)1-tagged B-Raf and formation of the CalDAG-GEFI/Rap1/HA1-tagged B-Raf complex can be blocked by expression of CalDAG-GEFI antisense RNA. Together, these data indicate that stimulation of M 1 mAChR by carbachol causes the sequential activation of CalDAG-GEFI, Rap1, and B-Raf, leading to the activation of MEK and ERK1/2.
Plasmid Constructions-A dominant-negative (dn) Rap1 (Rap1 Asn-17 ) was obtained by PCR amplification from the mouse full-length Rap1 cDNA cloned in pT 7 T 3 D-Pac (identified in the mouse expressed sequence tag data base (GenBank TM accession number AA272379) and purchased from Genome Systems, Inc., Livermore, CA) using the mutant oligonucleotide primer GGCGTGGGGAAGAATGCTCTAACAGTTCA-GTTTGTTCAG to change the 17th amino acid from serine to asparagine with the QuikChange Site-directed TM Mutagenesis Kit (Stratagene). DNA sequencing was performed before and after mutagenesis. The Rap1 Asn-17 cDNA was excised from this vector by digestion with NotI and XhoI and subcloned in pOPRSVI/MCS LacSwitch TM II (Stratagene) between the NotI and XhoI sites to produce the isopropyl-␤-Dthiogalactoside (IPTG)-inducible expression vector (pRap1N17) containing Rap1 Asn-17 under the control of the Rous sarcoma virus-long terminal repeat promoter and the Lac operator/repressor. Full-length rat CalDAG-GEFI cDNA was obtained by reverse transcriptionpolymerase chain reaction (RT-PCR) from RNA isolated from PC12D cells and rat brain. Primers were designed based on the mouse CalDA-G-GEFI sequence (GenBank TM accession number AF081193). The forward primer was 5Ј-AGGATCAGAGGCTGAGCTGGTT-3Ј, and the backward primer was 5Ј-TCTCCAAGGCAGGAATGAGTCC-3Ј. RT-PCR product was isolated from a 1% agarose gel and cloned in pGEM-T Easy (Promega). This plasmid was then used as a template for constructing an expression vector encoding c-Myc-tagged CalDAG-GEFI. The forward primer used in this construction was 5Ј-GGTGACACTAT-AGAATACTCAAGCTATGCATCC-3Ј, and the backward primer was 5Ј-TTATCGTCGACTAAGTGGATGTCGAACA-3Ј. This PCR product was purified from a 1% agarose gel and cloned in pGEM-T Easy. CalDAG-GEFI cDNA was excised from this vector by digestion with EcoRI and SalI and subcloned between the EcoRI and SalI sites of pCMV-Tag 5 (Stratagene). The resulting plasmid was designated pCalDAG-GEF1myc. An antisense expression vector, pCalDAG-GEFI-myc-antisense, was constructed by subcloning a cDNA fragment encoding CalDAG-G-EFI between the EcoRI and BamHI sites of pCMV-Tag 5.
Cell Culture and Transfection-PC12D cells (19), a rapidly differentiating subline of rat pheochromocytoma-derived PC12 cells (20), were a gift from Prof. Mamoru Sano (Department of Biology, Faculty of Medicine, Kyoto Prefectural University of Medicine). PC12D cells were cultured in Dulbecco's modified Eagle's medium (DMEM, Nissui) supplemented with 5% fetal bovine serum, 5% horse serum, 0.16% sodium bicarbonate, 3.6 mM glutamine, 10 units/ml penicillin, and 45 ng/ml streptomycin at 37°C under 5% CO 2 as described previously (21). Cells were used in the non-differentiated state in all the experiments. Drugs were added directly to the culture medium and were present until the time when the cells were harvested. The corresponding vehicle (water, Me 2 SO, or ethanol) was added to control cells. Cells were seeded in 3.5-cm plastic dishes (Corning or Iwaki Glass Co.) at a density of 2 ϫ 10 6 cells/dish and then cultured for 1 day, allowing the cells grow to 90 -95% confluency prior to transfection. LipofectAMINE TM (Life Technologies, Inc.) was used for making stable cell lines containing the dnRap1 gene. Transfections were performed essentially as recommended by the manufacturer. 5 g of PCMVLacI repressor (Stratagene) together with 5 g of pRap1N17, and 39 l of LipofectAMINE were used for each 10 cm dish. LipofectAMINE TM 2000 Reagent (Life Technologies, Inc.) was used for other transfections (10 cm) essentially as recommended by the manufacturer. pCalDAG-GEFI-myc (1 g or as indicated), 1 g of pHA1-B1-Raf (an expression plasmid encoding hemagglutinin (HA) 1-tagged quail B-Raf, Ref. 62; a gift from Dr. Alain Eychene, Unite Mixte de Recherche 146 du CNRS, Institut Curie, Center Universitaire, Laboratoire 110, 91405 Orsay Cedex, France), 1 g of pCalDAG-GEFI-myc-antisense, 1 g of pRap1N17, or 1 g of pcDNA3, and 4 l of LipofectAMINE TM 2000 reagent were used for each dish in different combinations as indicated. Cells were kept in normal DMEM for 48 h until they were ready to be used.
Selection of Rap1 and Ras Dominant-negative Expressing Cell Lines-A PC12D subline, PC12D-37, stably expressing a dexamethasone-inducible dnRas gene (Ras Asn-17 ) (63), under the control of the mouse mammary tumor virus-long terminal repeat promoter (64), has been described previously (65). Stable cell lines that express the dnRap1 gene (Rap1 Asn-17 ) under the control of the IPTG-inducible Lac operator/repressor were obtained by transfecting PC12D-37 cells with pRap1N17 and PCMVLacI repressor using LipofectAMINE TM , followed by selection for hygromycin B-resistant colonies in DMEM containing 300 g/ml hygromycin B. (Selection was initiated 1 week after transfection.) After 2 weeks of selection, hygromycin B-resistant colonies were isolated and screened for the ability of Rap1Asn-17 (induced by exposing cells to IPTG) to block the activation of B-Raf by carbachol. Cell lines that showed complete or nearly complete inhibition of carbachol-mediated B-Raf activation were chosen for further study. In this way, the cell line PC12D-37-19 containing Ras Asn-17 under the control of a dexamethasone-inducible promoter and Rap1 Asn-17 under the control of an IPTG-inducible promoter was selected. PC12D-37-19 cells were subsequently maintained in DMEM containing 150 g/ml hygromycin B and 100 g/ml G418. Overexpression of dnRap1, dnRas, or dnRap1 and dnRas in these cells was achieved by pretreating the cells with 5 mM IPTG for 7 h, 0.5 M dexamethasone for 19 h, or by pretreating the cells with both IPTG and dexamethasone.
ERK1/2 Kinase Assays-ERK1/2 kinase assays using anti-phospho-p44/42 MAPK (Thr-202/Tyr-204) antibodies were performed essentially as recommended by the manufacturer (New England Biolabs Inc.). Briefly, cells grown to 90 -95% confluency in 3.5-cm uncoated plastic culture dishes were stimulated with various reagents for the times indicated and then lysed by adding 100 l of SDS sample buffer containing 62.5 mM Tris-Cl (pH 6.8), 2% w/v SDS, 10% glycerol, 50 mM dithiothreitol, 0.1% w/v bromphenol blue, and then immediately scraped off the plates. The resulting cell lysates were transferred to microcentrifuge tubes on ice, sonicated for 10 -15 s in a bath sonicator, and boiled for 5 min. After centrifugation for 5 min to remove cellular debris, the proteins in the samples were resolved by 12% SDS-polyacrylamide gel electrophoresis (PAGE), transferred to polyvinylidene difluoride (PVDF) membranes (Immobilon TM transfer membrane, Millipore), and probed with anti-phospho-p44/42 MAPK (Thr-202/Tyr-204) antibodies as described below. In vitro phosphorylation of the ERK1/2 substrate myelin basic protein in the presence of [␥-32 P]ATP was carried out as described previously (26). Quantification of ERK1/2 activation was performed by liquid scintillation counting.
B-Raf Kinase Assays-B-Raf kinase assays were performed as described previously (40) with some modifications. Briefly, cells grown to 90 -95% confluency in 3.5-cm uncoated plastic culture dishes were stimulated with various reagents for the times indicated and then lysed by addition of 200 l of lysis buffer containing 10 mM Tris-Cl (pH 7.4), 5 mM EDTA, 50 mM NaCl, 50 mM NaF, 0.1% bovine serum albumin, 20 units/ml aprotinin, 1 mM phenylmethanesulfonyl fluoride, 1 mM Na 3 VO 4 , and 1% Triton X-100. After 5 min centrifugation at 15,000 rpm at 4°C to remove cellular debris, 1 g of anti-B-Raf antibodies (Santa Cruz Biotechnology, for endogenous B-Raf) or 1 g of anti-HA antibodies (Roche Molecular Biochemicals, for HA1-tagged B-Raf) were added to the supernatant fractions, which were then incubated for 1 h at 4°C with rotation to provide gentle mixing. Protein A/G-agarose (10 l of resin suspension, Santa Cruz Biotechnology) was subsequently added to each sample, and the incubation was continued with rotation at 4°C for 1 h. The agarose in each sample was collected by centrifugation (2,500 rpm for 5 min) and washed twice with 200 l of lysis buffer and once with PAN buffer containing 10 mM PIPES (pH 7.0), 20 units/ml aprotinin, and 100 mM NaCl. The resin from each sample was then resuspended in 16.5 l of 1ϫ MEK buffer containing 2 g of MEK1 (Santa Cruz Biotechnology), 10 l of PAN buffer, 1 l of [␥-32 P]ATP (3000 Ci/mM), and 4 l of 5ϫ kinase buffer (5ϫ kinase buffer: 50 mM MgCl 2 , 20 units/ml aprotinin, and 10 mM PIPES (pH 7.0)). The reaction mixtures were incubated at 30°C for 30 min, then resolved by 10% SDS-PAGE, and electroblotted onto PVDF membranes. Phosphorylated MEK1 was assessed by autoradiography and quantified using the BAS2000 Bio Imaging Analyzer (Fuji Co.).
Rap1-and Ras-GTP Assays-GTP-bound forms of Rap1 and Ras were detected as described by Frank et al. (66) and Tayler and Shalloway (67), respectively. pGST-RalGDS-RBD encoding the 97 amino acids spanning the Rap binding domain (RBD) of Ral GDP dissociation stimulator (RalGDS) fused to glutathione S-transferase (GST) was a gift from Dr. Johannes Bos (Laboratory for Physiological Chemistry and Center for Biomedical Genetics, Utrecht University), and pGST-Raf-RBD encoding the 1-149 amino acids spanning the Ras-binding domain (RBD) of cRaf-1 fused to GST was a gift from Dr. David Shalloway (Section of Biochemistry, Molecular and Cell Biology, Cornell University). pGST-RalGDS-RBD and pGST-Raf-RBD were introduced into Escherichia coli (DH5␣), and the fusion proteins induced with 1 mM IPTG for 4 -5 h and purified from cell lysates using glutathione-Sepharose beads (Wako Chemicals Industries, Ltd.) as described previously (66,67). Briefly, PC12D-37-19 cells grown to 90 -95% confluency in 3.5-cm uncoated plastic culture dishes were stimulated with various reagents for the indicated times and lysed by addition of 200 l of RIPA buffer containing 10% glycerol, 1% nonylphenoxypolyethoxy ethanol (Nonidet P-40), 50 mM Tris-Cl (pH 7.5), 200 mM NaCl, 2 mM MgCl 2 , 1 mM phenylmethanesulfonyl fluoride, 2 g/ml aprotinin, 1 g/ml leupeptin, and 10 g/ml trypsin inhibitor. Lysates were clarified by centrifugation at 12,000 rpm for 10 min at 4°C. The supernatant fractions were then incubated with 5 g of GST-RalGDS-RBD or GST-Raf-RBD precoupled to glutathione-Sepharose beads for 45 min at 4°C with slight agitation. Beads were washed three times with lysis buffer prior to resolution by 12% SDS-PAGE and immunoblotting analysis with anti-Rap1 or anti-Ras antibodies (both from Transduction Laboratories) as described below.
Polyclonal Antibodies against Rat CalDAG-GEFI-A 20-amino acid oligopeptide (GCIREEEVQTVEDGVFDIHL) containing the C-terminal 18-amino acids of rat CalDAG-GEFI was used as the antigen peptide for the production of antiserum in rabbits (Biologica Co., Nagoya, Japan). Anti-CalDAG-GEFI antibodies were purified from this antiserum by affinity chromatography (SulfoLink kit, Pierce) using the antigen peptide as described previously (68).
Coimmunoprecipitations-PC12D cells were grown in DMEM and transfected with pCalDAG-GEFI-myc, pHA1-B1-Raf, pRap1N17, or pcDNA3 as described above. Cells were rinsed once with cold phosphate-buffered saline and then lysed by addition of 200 l of lysis buffer (150 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS, 50 mM Tris-Cl (pH 8.0), 2 M leupeptin, 1 mM phenylmethanesulfonyl fluoride, and 10 g/ml aprotinin). Cell lysates were incubated for 1 h at 4°C with 1 g of the anti-Rap1 antibodies, 1 g of anti-CalDAG-GEFI antibodies, 1 g of anti-c-Myc antibodies, or 1 g of anti-HA antibodies as indicated. The immune complexes were clarified by centrifugation at 2,500 rpm for 5 min, resolved by 10 or 12% SDS-PAGE, and transferred to PVDF membranes, followed by Western blot analysis as described below.
Western Blots-Brain tissues were homogenized in 1ϫ SDS sample buffer and the extracts centrifuged at 5,000 rpm to remove tissue debris. Other cell extracts were prepared as described above and mixed with the same volume of 2ϫ SDS sample buffer. After resolution by standard SDS-PAGE, proteins were electrophoretically transferred to PVDF membranes, which were then blocked overnight at 4°C with phosphate-buffered saline containing 5% powdered skim milk and 0.05% polyoxyethylene sorbitan monolaurate (Tween 20). The membranes were then exposed to affinity purified anti-CalDAG-GEFI antibodies (600 g/ml, 1:1000 dilution), anti-c-Myc antibodies (200 g/ml, 1:1000), or other antibodies as indicated in phosphate-buffered saline containing 0.5% powdered skim milk and 0.05% Tween 20 for 1 h at room temperature. The membranes were washed three times with the above buffer and incubated in buffer containing peroxidase-conjugated anti-rabbit, anti-mouse, or anti-Rat antibodies (each from Jackson Im-munoResearch Research Laboratories, Inc.; 1:2000 dilution) for 1 h at room temperature. Membranes were washed three times, and proteins were visualized by enhanced chemiluminescence (ECL kit, Amersham Pharmacia Biotech).

Carbachol Induces Sustained ERK1/2 Activation via Ras-dependent and -independent Pathways in PC12D Cells; Both
Pathways Require the Influx of Extracellular Ca 2ϩ -Our previous experiments showed that carbachol-induced zif268 gene expression in PC12D cells is blocked by the specific MEK inhibitor PD098059 (69), implying a role for ERK1/2 activation in M 1 mAChR-mediated zif268 gene expression in these cells (26). To determine whether ERK1/2 are activated following exposure to carbachol, we performed Western blot analysis of ERK1/2 using anti-phospho-ERK1/2 antibodies, which recognize only the activated forms of ERK1/2. As shown in Fig. 1, A and B, carbachol induces a rapid and sustained activation of ERK1/2 that is blocked by the mAChR antagonist atropine and by the MEK inhibitors PD098059 and U0126 (70).
Since well established pathways for activation of ERK1/2 are mediated by the small GTP-binding protein Ras (27-32), we examined whether Ras is required for the activation of ERK1/2 by carbachol in PC12D cells. For this purpose, we used the cell line PC12D-37 (65), which contains a dexamethasone-inducible dominant-negative (dn) Ras gene (Ras Asn-17 ). As reported previously, we found that ERK1/2 activation by NGF was largely blocked in cells expressing dnRas (65) but that carbachol-mediated activation was only partially blocked (Fig. 1C). These results suggest that a Ras-independent pathway plays a significant role in carbachol-mediated activation of ERK1/2 in PC12D-37 cells. We also found that carbachol-mediated activation of ERK1/2 is insensitive to pretreatment with pertussis toxin (100 ng/ml, 18 h) or wortmannin (200 nM, 10 min), 2 which inhibit G i/o and phosphatidylinositol 3-kinase (71), respectively.
Stimulation of M 1 mAChR activates PLC␤, which generates the intracellular second messengers DAG and IP 3 . DAG is an activator of PKC, and IP 3 stimulates the release of Ca 2ϩ from the endoplasmic reticulum and the subsequent influx of extracellular Ca 2ϩ (21). We therefore examined the relative contributions of PKC and Ca 2ϩ influx to ERK1/2 activation by pretreating PC12D-37 cells with the PKC inhibitor GF109203X (72) and the Ca 2ϩ chelator EGTA prior to stimulating with carbachol (Fig.  1D). These experiments show that the carbachol-mediated ERK1/2 activation is only slightly blocked by GF109203X but is largely blocked by EGTA. The same results were found in cells expressing dnRas. These data indicate that Ca 2ϩ influx is required for both Ras-dependent and -independent activation of ERK1/2 by carbachol in PC12D-37 cells and that PKC makes only a minor contribution to this activation.
Carbachol Quantification of 32 P-labeled MEK1 was performed using the BAS2000 BioImaging Analyzer. The data shown are the averages Ϯ S.E. of 3 independent experiments, performed as described in A.
Preferentially via Ras-dependent Pathways-In addition to Raf-1 (39), B-Raf (40) has also been shown to be an upstream activator of MEK, the immediate upstream activator of ERK1/2 (41). In fact, B-Raf has been proposed to be the primary activator of MEK in PC12 cells (38,73,74), where it is activated by NGF (43), and by increases in cAMP (42) or Ca 2ϩ (45). For this reason, we examined whether B-Raf is activated following exposure to carbachol in PC12D cells. In B-Raf kinase assays using MEK1 as a substrate, we found that B-Raf is rapidly activated by carbachol ( Fig. 2A, left, top row) and that this activation is blocked by atropine. 3 As reported previously (37), NGF also activates B-Raf ( Fig. 2A, right, top row).
B-Raf is well known to be activated by Ras (27)(28)(29)(30)(31)(32). More recent studies have suggested, however, that B-Raf can also be activated independently of Ras by the GTP-binding protein Rap1 (42)(43)(44)(45). To determine whether Rap1 and Ras are required for B-Raf activation by carbachol and NGF in PC12D cells, we decided to examine the effects of overexpression of dnRap1 and dnRas on B-Raf activation. For this purpose, we constructed the stable cell line PC12D-37-19, which contains an IPTG-inducible dnRap1 gene (Rap1 Asn-17 ) and a dexamethasone-inducible dnRas gene (Ras Asn-17 ).
As shown in Fig. 2A, rapid and sustained activation of B-Raf by carbachol was more potently blocked by dnRap1 than dnRas. By contrast, NGF-mediated activation was more potently blocked by dnRas than dnRap1. Activation of B-Raf by carbachol and NGF was completely blocked in cells expressing both dnRap1 and dnRas.
These results were confirmed in quantitative B-Raf assays (Fig. 2, B and C), which showed that carbachol activated B-Raf by about 8.5-fold and that this activation was blocked by about 75% in cells expressing dnRap1 and by only 25% in cells expressing dnRas (Fig. 2B). By contrast, a 9-fold increase of B-Raf activation by NGF was blocked by about 75% in cells expressing dnRas, and by only 25% in cells expressing dnRap1 (Fig.  2C). Activation of B-Raf by both carbachol and NGF was completely blocked in cells expressing dnRap1 and dnRas.
These results indicate that Rap1 is the primary mediator of carbachol-stimulated activation of B-Raf and that Ras is the primary mediator of NGF-stimulated activation in PC12D-37-19 cells. Together the Rap1-and Ras-dependent pathways seem to account for all of the B-Raf activation by carbachol or NGF.
Carbachol Preferentially Activates Rap1 and NGF Preferentially Activates Ras-To determine whether Rap1 and Ras are activated by carbachol with sufficient rapidity to function upstream of B-Raf, we examined the time courses of Rap1 and Ras activation (Fig. 3A). In these experiments, PC12D-37-19 cells were treated with carbachol or NGF, and the GTP-bound forms of Rap1 and Ras were specifically precipitated from cell extracts using RalGDS-GST-RBD and pGEX-Raf-RBD and detected in Western blots using anti-Rap1 and anti-Ras antibodies, respectively. As a control, total Rap1 or Ras was immunoprecipitated from the same volume of cells and detected using specific antibodies for each protein. These studies show that carbachol preferentially activates Rap1 and that NGF preferentially activates Ras.
The kinetics of Rap1 and Ras activation by carbachol and NGF (Fig. 3, B and C) are similar to those for B-Raf activation (Fig. 2, B and C), suggesting that the activation of Rap1 and Ras occurs with sufficient speed to participate in the activation of B-Raf and may in fact be rate-limiting for the activation of B-Raf. The preferential activation of Rap1 by carbachol and 3 E. Kumahara and D. Saffen, unpublished observations. Ras by NGF is consistent with the preferential contribution of Rap1 and Ras to B-Raf activation by carbachol and NGF, respectively.
Effects of Overexpression of dnRap1 and dnRas on Carbachol-mediated Rap1 and Ras Activation-To determine if the block of carbachol-mediated B-Raf activation by dnRap1 and dnRas is caused by inhibition of Rap1 and Ras, respectively, we examined the effects of dnRap1 and dnRas on the activation of Rap1 and Ras by carbachol.
As shown in Fig. 4A, activation of Rap1 and Ras by carbachol was blocked by overexpression of dnRap1 and dnRas, respectively. IPTG and dexamethasone, the agents used to induce the expression of dnRap1 or dnRas, did not activate Rap1 or Ras. Quantitative analysis of Rap1 and Ras activation (Fig. 4B) revealed that carbachol-mediated Rap1 activation was blocked by about 80% in cells expressing dnRap1, and that Ras activation was blocked by almost 100% in cells expressing dnRas. Activation of both Rap1 and Ras was blocked completely in cells expressing both dnRap1 and Ras. These results indicate that dnRap1 specifically inhibits the activation of Rap1 and that dnRas specifically inhibits activation of Ras in PC12D-37-19 cells.
Activation of Rap1 by Carbachol Does Not Occur via NGF or EGF Receptors-In addition to mAChRs, PC12 cells also express membrane receptors for NGF and EGF (20,75). It has been reported that EGF receptors can be transactivated by mAChR agonists via extracellular and intracellular signal pathways (76 -79). Our previous experiments have also shown that the induction of the zif268 gene by carbachol is partly blocked by the specific EGF receptor tyrosine kinase inhibitor AG1478 (80). 4 To determine whether NGF or EGF receptors are involved in carbachol-mediated activation of Rap1 and Ras, we examined the effects of pretreating the cells with the NGF receptor kinase inhibitor K252a (81) and the EGF receptor kinase inhibitor AG1478 (Fig. 5A). These experiments show that carbachol-mediated activation of Rap1 and Ras is not blocked by K252a or AG1478, although K252a blocked NGFand AG1478 blocked EGF-mediated activation of Rap1 and Ras. Pretreatment of cells with K252a and/or AG1478 had no effect on Rap1 and Ras in the absence of carbachol. Quantification of the results of three independent assays showed that carbachol increased Rap1-GTP levels by 8-fold but Ras-GTP levels only 3-fold. Carbachol-stimulated increases in 4 H. Ishii and D. Saffen, unpublished observations.

FIG. 4. Effects of dnRap1 and dnRas on the activation of Rap1
and Ras by carbachol. A, Rap1-and Ras-GTP assays. Left, PC12D-37-19 cells containing inducible dnRap1 and dnRas genes were pretreated with water, 5 mM IPTG for 7 h to induce dnRap1, 0.5 M dexamethasone for 19 h to induce dnRas, or IPTG and dexamethasone to induce dnRap1 and dnRas prior to stimulation with water (W) or 500 M carbachol (carb) for 10 min. Right, cells were stimulated with water following pretreatments for induction of dnRap1, dnRas, dnRap1, and dnRas or neither. Rap1-GTP, Ras-GTP, total Rap1, and total Ras were detected as described in the legend to Fig. 3. The results from a representative experiment are shown. B, fold increases in activated Rap1 and Ras were calculated with respect to the values obtained with the cells stimulated with water. Quantification of the results was performed using NIH Image 1.61 software. The data shown are the averages Ϯ S.E. of 3 independent experiments, performed as described in A.

FIG. 5. Effects of receptor tyrosine kinase inhibitors on the activation of Rap1 and Ras by carbachol.
A, Rap1-and Ras-GTP assays. PC12D-37-19 cells were pretreated with 0.1% Me 2 SO, 200 nM K252a, or 250 nM AG1478 for 20 min prior to stimulation with water (W), (Ϯ) 500 M carbachol (carb), (Ϯ) 10 ng/ml NGF, or (Ϯ) 10 ng/ml EGF for 10 min. Rap1-GTP, Ras-GTP, total Rap1, and total Ras were detected as described in the legend to Fig. 3. The results from a representative experiment are shown. B-E, fold increases in activated Rap1 and Ras were calculated with respect to the values obtained with the cells stimulated with water. Quantification of the results was performed using NIH Image 1.61 software. The data shown are the averages Ϯ S.E. of 3 independent experiments, performed as described in A.
Rap1-and Ras-GTP levels were not affected by K252a or AG1478 (Fig. 5B). By contrast, NGF and EGF increased Ras-GTP levels by 7-fold and Rap1-GTP levels 3-fold, but these effects were blocked completely by K252a and AG1478, respectively (Fig. 5, C and D). The inhibitors alone had no effect on Rap1 or Ras (Fig. 5E).
These results indicate that carbachol-mediated activation of Rap1 and Ras is not mediated by transactivation of the NGF or EGF receptors.
Roles of Ca 2ϩ and PKC in Carbachol-mediated Activation of Rap1 and Ras-To determine the relative contributions of Ca 2ϩ influx and PKC to Rap1 and Ras activation by carbachol in PC12D-37-19 cells, we first examined the effects of the Ca 2ϩ ionophore ionomycin and the PKC activator PMA. As shown in Fig. 6A, Rap1 was more potently activated by ionomycin than by PMA. By contrast, ionomycin and PMA activated Ras less strongly. No additive effect was observed for Rap1 activation when cells were treated with ionomycin in combination with PMA. Treatment of cells with the solvents Me 2 SO or EtOH had no effect on Rap1 or Ras. These results were confirmed by quantitative analysis of these assays (Fig. 6B), where ionomycin was shown to increase Rap1-GTP levels and Ras-GTP levels by 8-and 3-fold, respectively, and PMA to increase Rap1-GTP levels and Ras-GTP levels by 3-and 1.5-fold, respectively.
The effects of the PKC inhibitor GF109203X and Ca 2ϩ chelator EGTA are shown in Fig. 6C. Pretreatment of cells with GF109203X had little effect on carbachol-stimulated increases in Rap1-and Ras-GTP levels, but activation of Rap1 and Ras was significantly blocked by chelating extracellular Ca 2ϩ with EGTA. Again, no obvious additive effects were observed. Pretreatment of cells with GF109203X and/or EGTA had no effect on Rap1-GTP and Ras-GTP levels in the absence of carbachol (Fig. 6C). Quantification of the results of several assays of Rap1 and Ras activation showed that carbachol-mediated activation of Rap1 and Ras was blocked by over 90% when cells were treated with EGTA, but that there was no blocking effect when cells were treated with GF109203X. The inhibitors themselves did not affect activation of Rap1 or Ras (Fig. 6D).
These results indicate that Ca 2ϩ influx is necessary and sufficient for activation of Rap1 and Ras in PC12D-37-19 cells, a result that is consistent with its important role for ERK1/2 activation. PKC makes only a minor contribution to activation of Rap1 and Ras, since PMA only weakly activates Rap1 and Ras, and GF109203X does not inhibit the activation of Rap1 or Ras.

RT-PCR and Western Blot Show That CalDAG-GEFI Is Expressed in PC12D Cells-
The data presented so far establish an important role for Rap1 in carbachol-mediated activation of B-Raf in PC12D-37-19 cells and show that the activation of Rap1 is mediated primarily via Ca 2ϩ influx. The mechanism of Rap1 activation, however, remains to be defined. Like Ras, activation of Rap1 also requires a GEF to promote its conversion from the GDP-to the GTP-bound form (50). Among Rap1 GEFs identified to date, our attention was drawn to a recently discovered GEF that is expressed in brain, CalDAG-GEFI (58). The fact that CalDAG-GEFI has binding sites for Ca 2ϩ and DAG and was shown to activate Rap1 in 293T cells suggested that it might link the PLC␤-mediated elevation of Ca 2ϩ influx and PKC activation to activation of B-Raf. For these reasons, we decided to examine whether CalDAG-GEFI might play a role in carbachol-mediated activation of Rap1 and B-Raf in PC12D cells.
Our first experiment was to determine if CalDAG-GEFI is expressed in PC12D cells. Using DNA primers based on the sequence of the mouse CalDAG-GEFI (GenBank TM accession number AF081193), we were able to amplify a full-length CalDAG-GEFI from mRNA isolated from PC12D cells and from whole rat brain by RT-PCR (Fig. 7A). Sequencing analysis of the rat cDNAs revealed 99.5% homology with mouse CalDAG-GEFI. 2 We next produced antibodies for CalDAG-GEFI in rabbits using a 20-amino acid oligopeptide (GCIREEEVQT-VEDGVFDIHL) containing the last 18 amino acids of rat CalDAG-GEFI C-terminal as the antigen peptide.
After affinity purification using the antigen peptide, we used these antibodies to examine the expression of CalDAG-GEFI in PC12D-37-19 cells and brain. In untransfected PC12D-37-19 cells and in brain, single bands corresponding to CalDAG-GEFI were detected (Fig. 7B, left). By contrast, two bands were detected in PC12D-37-19 cells transfected with pCalDAG-GEFI-myc: one corresponding to endogenous CalDAG-GEFI and a second corresponding to the Myc-tagged form of the protein. The identity of the Myc-tagged CalDAG-GEFI was confirmed by probing an identically prepared membrane with anti-c-Myc antibodies (Fig. 7B, middle). The specificity of the anti-CalDAG-GEFI antibodies was confirmed by the fact that preincubation with the CalDAG-GEFI antigen peptide completely blocked the appearance of CalDAG-GEFI bands in the Western blots (Fig. 7B, right). The observation that the mobility of the brain form of CalDAG-GEFI is a little slower than that from PC12D-37-19 cells suggest that post-translational modifications of this protein might be different in different tissues.

Carbachol and Ca 2ϩ Ionophore Ionomycin Increase the Association of CalDAG-GEFI and CalDAG-GEFI-myc with Rap1 and B-Raf-Since
Rap1 is activated by carbachol and Ca 2ϩ influx in PC12D-37-19 cells, we performed coimmunoprecipitation experiments to see whether CalDAG-GEFI is affected by these stimuli. Since Rap1 has been shown to form a complex with B-Raf upon exposure to NGF (43), or following increases in cAMP (42) or Ca 2ϩ (45) in PC12 cells, we examined if this is also true for PC12D-37-19 cells. In these experiments, we used Ca 2ϩ ionophore ionomycin and PMA to make our results comparable to those of Kawasaki et al. (58), and ionomycin and the DAG analogue 1-oleoyl-2-acetyl-sn-glycerol (OAG (82)) to replicate more closely the second messengers produced by stimulating M 1 mAChR.
As shown in Fig. 8A (left), immunoprecipitation of Rap1 from PC12D-37-19 cells stimulated with carbachol, ionomycin and PMA, or ionomycin and OAG resulted in the coimmunoprecipitation of CalDAG-GEFI and B-Raf. Exposure to ionomycin alone produced coimmunoprecipitation of CalDAG-GEFI and B-Raf to an extent similar to that obtained with ionomycin and PMA or OAG. Exposure to PMA or OAG alone produced no coimmunoprecipitation. 2 In contrast to the results obtained with carbachol, there was only a weak association of CalDAG-GEFI, Rap1, and B-Raf following exposure to NGF. The same set of proteins was found to associate when CalDAG-GEFI antibodies were used for the immunoprecipitation. In PC12D-37-19 cells transfected with pCalDAG-GEFI-myc, the same results were found, except that an additional band corresponding to transfected CalDAG-GEFI-myc was also detected (Fig.  8A, right). The association of CalDAG-GEFI-Myc with Rap1 and B-Raf was also examined using anti-Rap1 or anti-c-Myc antibodies for immunoprecipitation (Fig. 8B), and similar results were obtained.
These results indicate that a complex containing CalDAG-GEFI, Rap1, and B-Raf is formed following exposure to carbachol in PC12D-37-19 cells.
B-Raf in the Immunocomplexes Is Activated-To determine if B-Raf detected in the immunocomplexes described above is activated, we performed B-Raf assays using MEK1 as a substrate. As shown in Fig. 9, A and C, B-Raf in immunoprecipitates obtained with either anti-Rap1 or anti-CalDAG-GEFI antibodies was significantly activated following exposure to carbachol, ionomycin and PMA, or ionomycin and OAG, but only weakly activated by NGF. These results were also confirmed by quantitative analysis of the B-Raf kinase assays (Fig.  9, B and D). When anti-Rap1 antibodies were used for immunoprecipitation, B-Raf was activated by 9-, 8.5-, or 8.7-fold by carbachol, ionomycin and PMA, or ionomycin and OAG, but only 2.5-fold by NGF. When anti-CalDAG-GEFI antibodies were used, B-Raf was activated to a similar extent under identical treatment. These results indicate that the complex containing CalDAG-GEFI/Rap1/B-Raf that is formed following exposure to carbachol is able to activate the downstream effector of B-Raf, MEK. (a plasmid encoding HA1-tagged B-Raf) alone or with pHA1-B1-Raf in combination with pCalDAG-GEFI-myc-antisense, pRap1N17, pcDNA3, or pCalDAG-GEFI-myc.

Expression of CalDAG-GEFI Antisense RNA Blocks Both Carbachol-mediated Activation of HA1-tagged B-Raf and
The effects of cotransfection with antisense CalDAG-GEFI RNA on HA1-tagged B-Raf activation by carbachol and NGF was first examined by performing kinase assays for HA1tagged B-Raf using MEK-1 as a substrate. As shown in Fig.  10A (top, upper row), carbachol-mediated activation of HA1tagged B-Raf in PC12D-37-19 cells was significantly blocked by expression of antisense CalDAG-GEFI RNA or dnRap1. By contrast, NGF-mediated activation was not blocked by antisense CalDAG-GEFI RNA, although it was partially blocked by expression of dnRap1 (Fig. 10B, top, upper row). No blocking effect was observed for cells transfected with the control plasmid pcDNA3 and subsequently treated with either carbachol or NGF (Fig., 10, A and B, top, upper rows). Quantification of the results of 3 independent HA1-tagged B-Raf kinase assays (Fig.  10, A and B, bottom) showed that a 9-fold activation of HA1tagged B-Raf by carbachol was blocked by about 60 and 70% when antisense CalDAG-GEFI RNA and dnRap1 were expressed, respectively. A 14-fold activation by NGF, however, was not blocked by CalDAG-GEFI antisense RNA but was blocked by about 30% by dnRap1. Cotransfection of the control plasmid pcDNA3 had no effect. These results indicate that CalDAG-GEFI is required for activation of the HA1-tagged B-Raf by carbachol but is not required for NGF-mediated activation of HA1-tagged B-Raf in PC12D-37-19 cells.
The ability of CalDAG-GEFI and Rap1 to form a complex with HA1-tagged B-Raf in PC12D-37-19 cells transfected with pHA1-B1-Raf was also investigated by immunoblotting (Fig.  10C). Expression of antisense CalDAG-GEFI RNA or dnRap1 reduced levels of CalDAG-GEFI and Rap1 detected in the pre- cipitated complex following exposure to carbachol but cotransfection with the control plasmid pcDNA3 did not reduce levels of CalDAG-GEFI and Rap1 in the complex (Fig. 10C). These results indicate that expression of antisense CalDAG-GEFI RNA inhibits the formation of the CalDAG-GEFI/Rap1/B-Raf complex associated with B-Raf activation.
The specificity of the inhibition obtained with pCalDAG-GEFI-myc-antisense was confirmed by cotransfection of cells with pCalDAG-GEFI-myc. As shown in Fig. 10D (middle), cotransfection with increasing amounts of the CalDAG-GEFImyc expression plasmid eventually overwhelmed the inhibition of HA1-tagged B-Raf activation caused by expression of CalDAG-GEFI antisense RNA. As shown in Fig. 10D (top), coexpression of pCalDAG-GEFI-myc with pCalDAG-GEFImyc-antisense at molar ratios of less than 1 (the two plasmids have the same molecular mass) resulted in complete inability to detect the Myc epitope. The Myc epitope could be detected, however, in cells transfected with both sense and antisense plasmids at molar ratios of 1 and above. These results indicate that the ability of CalDAG-GEFI antisense RNA to block activation of HA1-tagged B-Raf is correlated with its ability to block the expression of CalDAG-GEFI. DISCUSSION In the present work, we describe a novel pathway linking M 1 mAChR to the activation of ERK1/2, in which coupling is mediated by a CalDAG-GEFI/Rap1/B-Raf cassette (Fig. 11). This report provides the first evidence for the function of endogenous CalDAG-GEFI.
Activation of the ERK1/2 cascade following stimulation of M 1 mAChR has been described previously in a variety of cell lines and tissues (10 -13). Stimulation of PC12D cells with carbachol also induces a rapid and sustained activation of ERK1/2 that is mediated by mAChR and MEK (Fig. 1, A and B). How ERK1/2 is activated by mAChR in PC12D cells, however, was not un-derstood until now. Pathways for ERK1/2 activation have been most thoroughly characterized for signaling from growth factor receptors, where activation is primarily mediated by Ras (27)(28)(29)(30)(31)(32). In contrast to this pathway, activation of ERK1/2 by carbachol takes place largely by a Ras-independent pathway in PC12D cells (Fig. 1C).
PC12D cells express mRNAs for both M 1 and M 4 mAChRs, which couple to G q/11 and G i/o , respectively (21). Activation of ERK1/2 by receptors that couple to G i/o are thought to be mediated primarily through the release of ␤␥ subunits (83) and the subsequent activation of phosphatidylinositol 3-kinase (84) or nonreceptor tyrosine kinases (85). Activation of ERK1/2 by carbachol in PC12D cells, however, was found to be insensitive to pertussis toxin and wortmannin, which block the release of ␤␥ subunits from G i/o and inhibit phosphatidylinositol 3-kinase, respectively. These results argue against a role for M 4 mAChR in the activation of ERK1/2 by carbachol in PC12D cells and suggest, instead, that this activation is mediated exclusively by M 1 mAChR coupling to G q/11 . This inference is consistent with our previous findings that induction of zif268 gene expression by carbachol is exclusively mediated via M 1 mAChR in these cells (21).
Stimulation of M 1 mAChR results in the activation of PKC and influx of extracellular Ca 2ϩ (21). PKC has been proposed to play an important role in ERK1/2 activation by receptors that couple to G q/11 (86). However, cases where PKC is not required for G q/11 -coupled receptors-mediated activation of ERK1/2 have also been reported (11,87). Our results indicate that PKC makes only a minor contribution to M 1 mAChR-mediated ERK1/2 activation in PC12D cells (Fig. 1D). By contrast, Ca 2ϩ influx is required for both Ras-dependent and -independent pathways of ERK1/2 activation (Fig. 1D). This is in good agreement with previous reports indicating a requirement for Ca 2ϩ influx for ERK1/2 activation in PC12 and other cells (87)(88)(89)(90). ERK1/2 is activated exclusively by MEK (41), and MEK is activated by the upstream activators Raf-1 (39) and B-Raf (40). B-Raf is the major Raf isoform in neurons and is highly expressed in PC12 cells (38,73,74). B-Raf has been proposed to be activated by both Ras-and Rap1-dependent pathways (27)(28)(29)(30)(31)(32)(42)(43)(44)(45). However, it is still controversial whether Rap1 always activates B-Raf in cells where both are expressed (48,49). By using a derivative of PC12 cells, PC12D-37-19, containing inducible dnRap1 and dnRas genes, we found that M 1 mAChR-stimulated activation of B-Raf is primarily mediated by a Rap1-dependent pathway (Fig. 2). As reported previously for PC12 cells (37), NGF also activates B-Raf in PC12D cells, and this activation is primarily mediated by a Ras-dependent pathway (Fig. 2). The preferential contribution of Rap1 and Ras to B-Raf activation by carbachol and NGF, respectively, is consistent with the observation that carbachol activates Rap1 more strongly than Ras and that NGF activates Ras more strongly than Rap1 (Fig. 3). We confirmed the specificity of the effects of dnRap1 and dnRas on B-Raf by showing that each preferentially blocks its cognate wild-type protein (Fig. 4). Together, these results provide evidence for distinct roles for Rap1 and Ras in the activation of B-Raf by carbachol and NGF in PC12D cells.
Based upon the different abilities of Ras and Rap1 to respond to NGF, and their differential roles in NGF-mediated activation of B-Raf, we assume that Ras is required for both early and late phase of ERK1/2 activation by NGF in PC12D cells. This is consistent with the well characterized role of Ras in NGFinduced signal cascades (27)(28)(29)(30)(31)(32). In contrast, York et al. (43) reported that Ras is required only for the initial activation of ERK1/2 by NGF in PC12 cells and that the sustained activation of ERK1/2 by NGF requires Rap1. The differences between our results and those of York et al. (43) might be due to the use of different strains of PC12 cells. On the other hand, carbachol only weakly activated Ras in PC12D cells (Fig. 3), and Ras-dependent activation of B-Raf (Fig. 2) and ERK1/2 ( Fig. 1) by carbachol was also weak. These results indicate that Ras is not so important for the activation of the ERK1/2 cascade by M 1 mAChR and that signals independent of Ras are required instead. Consistent with these ideas, Rap1 was found to be the primary mediator of B-Raf activation by carbachol in PC12D cells. Activation of Rap1 by carbachol may also result in the inhibition of the activation of Raf1 by Ras, since Rap1 has been shown previously to antagonize Ras-dependent signaling (46). Therefore, Rap1 may, on one hand, mediate the Ras-independent pathway for activation of ERK1/2 by carbachol and, on the other hand, inhibit the Ras-dependent pathway of activation in PC12D cells.
When considering factors that regulate the activation of Rap1 by carbachol, it is important to determine whether EGF receptors play a role. This is because EGF receptors have been reported to be transactivated by mAChR agonists (76 -79) and because activation of EGF receptors has been shown to activate Rap1 (48). Our observation that the EGF receptor inhibitor AG1478 has little effect on the activation of Rap1 by carbachol (Fig. 5), however, argues against a major role for EGF receptors in this activation in PC12D cells.
Since we previously showed that Ca 2ϩ influx and PKC contribute differentially to ERK1/2 activation in PC12D cells, their roles for Rap1 activation were also evaluated (Fig. 6). These experiments revealed that Ca 2ϩ influx is necessary and sufficient for the activation of Rap1, a result that is consistent with its crucial contribution to ERK1/2 activation (Fig. 1D). By contrast, activation of PKC makes only a minor contribution (Fig. 6).
Given our findings that Rap1 plays a major role in M 1 mAChR-mediated signaling in PC12D cells, it is important to is blocked by expression of CalDAG-GEFI antisense RNA. Cells were lysed and anti-HA antibodies were used for immunoprecipitation (IP) as described under "Experimental Procedures." CalDAG-GEFI, Rap1, and HA1-tagged B-Raf within the precipitated complex were detected by immunoblotting with anti-CalDAG-GEFI, anti-Rap1, and anti-HA antibodies, respectively. D, cotransfection with pCalDAG-GEFI-myc restores the expression of CalDAG-GEFI-myc and activation of HA1-tagged B-Raf in cells expressing CalDAG-GEFI antisense RNA. Cells were cotransfected with pHA1-B1-Raf, (Ϯ) pCalDAG-GEFI-myc, and (Ϯ) pCalDAG-GEFI-myc-antisense (antisense) as indicated, prior to stimulation with water (W) or 500 M carbachol for 10 min. Top, CalDAG-GEFI-myc in cell lysates was detected by immunoblotting with anti-c-Myc antibodies. Middle, HA1-tagged B-Raf (HA-B-Raf) was immunoprecipitated from equal volumes of cell lysate with anti-HA antibodies, and in vitro phosphorylation of the HA1-tagged B-Raf substrate MEK1 was carried out in the presence of [␥-32 P] ATP as described under "Experimental Procedures." Bottom, total precipitated HA1-tagged B-Raf was detected using anti-HA antibodies. elucidate the mechanisms by which it is activated. Activation of Rap1 requires GEFs, and to date several GEFs for Rap1 that are activated by distinct signals have been identified (44,50). Based on the observation that activation of Rap1 requires Ca 2ϩ influx, we became particularly interested in a GEF for Rap1, CalDAG-GEFI (58), that has binding sites for both Ca 2ϩ and DAG. Since intracellular levels of Ca 2ϩ and DAG increase following stimulation of M 1 mAChR, we guessed that CalDAG-GEFI might functionally couple M 1 mAChR to the activation of ERK1/2 in PC12D cells.
To test this hypothesis, we first used RT-PCR to determine if CalDAG-GEFI mRNA is expressed in PC12D cells. As shown in Fig. 7A, an RT-PCR DNA product of equal size could be amplified from mRNA isolated from PC12D cells and brain using primers specific for CalDAG-GEFI. Using antibodies raised against a synthetic peptide containing amino acids residues from the C terminus of rat CalDAG-GEFI as the antigen peptide, we also demonstrated the presence of CalDAG-GEFI protein in PC12D cells (Fig. 7B). Interestingly, the electrophoretic mobility of the CalDAG-GEFI in PC12D cells is a little faster compared with that expressed in brain, suggesting that CalDAG-GEFI may undergo different post-translational modifications in different tissues.
We next showed in coimmunoprecipitation experiments that CalDAG-GEFI forms a complex with Rap1 and B-Raf following exposure to carbachol or calcium ionophore and the DAG analogue OAG (Fig. 8). This is the first demonstration that CalDAG-GEFI can form a complex with Rap1 and B-Raf and confirms earlier experiments showing that Rap1 and B-Raf form a complex following exposure to NGF (43), or following increases in intracellular cAMP (42) or Ca 2ϩ (45) in PC12 cells. We also showed that B-Raf in these immunocomplexes is activated (Fig. 9), providing evidence for the direct activation of B-Raf by Rap1. In contrast to these results, Zwartkruis et al. (48) found that activation of endogenous Rap1 by PMA and endothelin did not result in the activation of ERK1/2 in Rat1 cells, even though B-Raf is expressed in these cells. Thus, activation of Rap1 may not always be linked to the activation of B-Raf. Busca et al. (49) also found that expression of constitutively active Rap1 failed to activate ERK1/2 in mouse B16/F10 melanoma cells. Again, these negative results may reflect cellspecific differences in Rap1 signaling but could also be due to differences in the response of B-Raf to acute stimulation (10 min), as measured in our experiments, versus chronic stimulation (2 days after transfection), as measured in their experiments.
The original paper (58) reporting the discovery and characterization of CalDAG-GEFI suggested that it is activated by Ca 2ϩ and DAG. In PC12D cells, PMA or OAG alone did not induce the association of CalDAG-GEFI, Rap1, and B-Raf, and there was no obvious additive effect on Rap1 activation when PMA or OAG were used in combination with ionomycin, compared with stimulation by ionomycin alone (Fig. 6). These results suggest that Ca 2ϩ influx alone may be sufficient to induce the activation of CalDAG-GEFI and the formation of the complex containing CalDAG-GEFI/Rap1/B-Raf. Alternatively, endogenous levels of DAG in PC12D cells may be sufficient for CalDAG-GEFI activation in combination with Ca 2ϩ influx.
A role for endogenous CalDAG-GEFI in the activation of ERK1/2 by carbachol is suggested by the observation that antisense CalDAG-GEFI RNA partially blocked carbachol-stimulated activation of HA1-tagged B-Raf (Fig. 10A). By contrast, activation of HA1-tagged B-Raf by NGF was not affected by the expression of CalDAG-GEFI antisense RNA (Fig. 10B), showing that the inhibitory effects of the antisense RNA are specific for mAChR-mediated signaling. Formation of the complex containing CalDAG-GEFI, Rap1, and HA1-tagged B-Raf was also partially blocked by expression of CalDAG-GEFI antisense RNA following stimulation with carbachol (Fig. 10C). The levels of these reductions in HA1-tagged B-Raf activation and complex formation are comparable to those obtained with dnRap1 (Fig. 10, A and C). The specificity of the inhibition obtained with CalDAG-GEFI antisense RNA was also demonstrated by the observation that inhibition of the activation of HA1-tagged B-Raf could be reversed by cotransfecting cells with an expression plasmid encoding CalDAG-GEFI-myc (Fig. 10D).
In cells cotransfected with sense and antisense expression plasmids at low molar ratios, antisense RNA expression blocked the expression of CalDAG-GEFI-myc. This result suggests that the level of pCalDAG-GEFI antisense RNA expression in our experiments may be sufficient to block the synthesis of the endogenous CalDAG-GEFI (Fig. 10, A-C), plus a small excess of exogenously expressed CalDAG-GEFI-myc (Fig. 10D). If this interpretation is correct, the inability of the CalDAG-GEFI antisense RNA to block completely activation of HA1tagged B-Raf (Fig. 10A) suggests the presence of additional pathways for the activation of B-Raf by carbachol. These pathways are likely to include the Ras-dependent pathway for B-Raf activation, demonstrated in Fig. 2.
Recent studies have also identified pathways that activate Rap1 independently of Ca 2ϩ influx. For example, Rap1 has been reported to be activated by cAMP via the GEFs Epac ((56) also designated cAMP-GEFI (57)) and cAMP-GEFII (57), or by a PKA-dependent pathway (91). Coexistence of both cAMP-and Ca 2ϩ -regulated activation of Rap1 in PC12 cells has also been reported (92). Although we have not ruled out the involvement of a cAMP-dependent pathway in Rap1 activation, we consider a role for cAMP unlikely, since cAMP levels are not significantly elevated by Ca 2ϩ influx in PC12 cells (93,94). We also found that carbachol-mediated activation of ERK1/2 is insensitive to the PKA inhibitor H89 (10 M, 15 min), 2 indicating that this activation is not dependent upon PKA. Increases in the intrinsic tyrosine kinase activities of growth factor receptors (e.g. TrkA, EGF receptors) activate Rap1 via C3G (51,52). A C3G-dependent pathway is not likely to be important for the activation of Rap1 by carbachol in PC12D cells, however, since this activation is insensitive to inhibitors of EGF and NGF receptor tyrosine kinases (Fig. 5). Rap1GEFs like nRap GEP/ PDZ-GEFI/Ra-GEF can activate Rap1 by binding to cell adhe- sion molecules (53)(54)(55), but there is no reason to suspect a role for these GEFs in the activation of Rap1 by carbachol in PC12D cells. The number of GEFs that are known to activate Rap1 continues to grow, however, so we cannot completely rule out the involvement of additional GEFs in Rap1 activation by carbachol.
In conclusion, activation of ERK1/2 following stimulation of M 1 mAChR in PC12D cells is mediated by a complex containing CalDAG-GEFI/Rap1/B-Raf. Our working model for this cascade is depicted in Fig. 11. Since the generation of DAG and stimulation of Ca 2ϩ influx are common outcomes of the activation of G q/11 -coupled receptors, we speculate that this pathway may also serve as a mechanism for activation of ERK1/2 by other G q/11 -coupled receptors in cells expressing B-Raf. This hypothesis awaits further investigation.