cAMP-dependent Protein Kinase Induces cAMP-response Element-binding Protein Phosphorylation via an Intracellular Calcium Release/ERK-dependent Pathway in Striatal Neurons*

Activation of the cAMP-dependent protein kinase A (PKA) pathway may induce cAMP-response element-binding protein (CREB) phosphorylation either directly or via cross-talk mechanisms with other signal transduction pathways. In this study, we have investigated in striatal primary cultures the mechanism by which activation of the cAMP/PKA-dependent pathway leads to CREB phosphorylation via the extracellular signal-regulated kinase (ERK)-dependent pathway. We have found that PKA-induced CREB phosphorylation and CREB-dependent transcription are mediated by calcium (Ca2+) release from intracellular stores and are blocked by inhibitors of the protein kinase C and ERK pathways. This mechanism appears to be mediated by the small G-protein Rap1, whose activation appears to be primed by PKA-induced Ca2+ release but not further induced by direct or indirect PKA- or protein kinase C-dependent phosphorylation. These results suggest that, in striatal neurons, intracellular Ca2+ release, Rap1, and ERK pathway play a crucial role in the PKA-induced CREB phosphorylation and CREB-dependent transcription.

The dopaminergic striatal system is the main target of the antipsychotic agents used in the treatment of schizophrenia (1) and of the psychostimulant drugs cocaine and amphetamines (2). The prolonged administration of these drugs increases the synaptic availability of dopamine and induces many dopaminedependent adaptive responses culminating in the transcription of striatal cAMP-response element (CRE) 1 -dependent genes, such as the immediate early gene c-fos and the neuropeptides dynorphin, substance P, and enkephalins.
The ERK pathway plays a pivotal role in stimulus-dependent gene regulation in the central nervous system, because pharmacological manipulations of the ERK pathway functionality affect the synaptic plasticity mechanisms supposed to underlie learning and memory (13). The cascade responsible of the ERK pathway activation requires the stimulus-dependent recruitment of the small G protein Ras, which in turn activates the Raf and MEK kinases. Although Ca 2ϩ -dependent activation of Ras has been demonstrated in several experimental models (14), in the central nervous system an alternative route to ERK activation via the cAMP/PKA-or Ca 2ϩ -dependent pathway has been recently characterized. This route implies the participation of the small G protein Ras-related Rap1 and the downstream kinase B-Raf (15,16).
The importance of the Rap1/B-Raf system in the cross-talk between the cAMP/PKA-and ERK-dependent pathways has received further attention by the recent identification of two novel families of Rap1GEF activated by direct binding to Ca 2ϩ and cAMP (17). However, these results were obtained mainly in platelets and PC12 cells, rather than in primary neural cells, which might display distinct signal transduction pathways. * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Furthermore, the exact mechanism by which the cAMP/PKAdependent pathway induces ERK activation and the role, if any, of Rap1 in CREB phosphorylation and CRE-dependent transcription have not yet been investigated.
In an attempt to address these questions, in this study we have investigated the mechanisms involved in cAMP/PKA-dependent ERK activation, CREB phosphorylation, and CREdependent gene transcription in striatal neurons.
Cell Culture-Primary cultures were obtained as previously described (18). Briefly, striata of 16 -17-day-old Harlan Sprague-Dawley rat embryos were dissected and incubated with papain. Tissue fragments were mechanically dissociated, and the cells were plated in 60-mm diameter poly-L-lysine-coated (10 g/ml) dishes and grown in 1:1 minimal essential medium/F-12 medium containing 2 mM glutamine, 2.5% fetal bovine serum, and 2.5% horse serum. About 48 h after plating, 10 M cytosine-C arabinoside was added to the cultures to prevent the growth of nonneuronal cells. The cells were grown for 6 days in vitro for intracellular Ca 2ϩ measurement experiments (see below) or 12 days in vitro for all other experiments. The total percentage of glial cells was below 5%, as assessed by glial fibrillary acid protein staining and counterstaining with 1% cresyl violet. Unless otherwise indicated, cells were rinsed twice with Krebs buffer and then left for 20 min in Krebs buffer before stimulation with the indicated agents.
In the experiments with BAPTA-AM, the cells were preincubated for the indicated time with BAPTA-AM and then incubated with stimulating agents in Krebs buffer. Experiments in which the role of excitatory amino acids was studied were performed in the presence of 1 M tetrodoxin to block action potentials and excitatory amino acid release.
Intracellular Ca 2ϩ Measurement-The intracellular Ca 2ϩ concentration was measured with the fluorescent Ca 2ϩ indicator fura-2/AM. Striatal cells, prepared as described above, were plated at low density into 24-well multiwell plates containing 13-mm glass coverslips previously coated with 10 g/ml poly-L-lysine. After 6 days in vitro, cells were washed twice with Krebs buffer and then incubated in the same buffer containing 5 M fura-2/AM at 37°C in a humidified incubator with 5% CO 2 for 30 min, followed by a postincubation period in the same buffer for 30 min. The dye-loaded cells were epi-illuminated with light from a 75-watt xenon lamp filtered through 340-and 380-nm interference filters, and the fluorescence emitted at 510 nm was revealed by a photon-counting photomultiplier. The 340/380-nm fluorescence ratio, averaged over a period of 2 s, was measured for 10 min in the Krebs buffer, and this value was taken as basal reference value. The stimulating agents were added as 10ϫ concentrated solution, and then changes in the 340/380-nm ratio were monitored for the time indicated in Fig. 9. For each experimental condition, fluorescence responses were taken in at least three fields on each coverslip.
Pull-down Assay for the Determination of Rap1 Activation-The pulldown assay for the determination of Rap1 activation was carried out as previously reported (19). Briefly, the GST fusion protein of the minimal Rap1-binding domain of ralGDS (ral-RBD) was induced in Escherichia coli (strain BL21DE3) by isopropyl-1-thio-␤-D-galactopyranoside, and bacterial lysates were prepared as described (19). The GST fusion protein was immobilized by incubating the bacterial lysate for 1 h at 4°C with glutathione-Sepharose preequilibrated in pull-down buffer (50 mM Tris buffer, pH 7.5, 200 mM NaCl, 2 mM MgCl 2 , 1 mM phenylmethylsulfonyl fluoride, 10% glycerol, 1% Nonidet P-40, 2 g/ml aprotinin, 1 g/ml leupeptin). Sepharose beads were washed three times to remove excess GST fusion protein. After stimulation with agonists for 10 min, the cells were rinsed twice with stimulation buffer and then chilled immediately in pull-down buffer, and then cell lysate was cleared by centrifugation at 10,000 ϫ g for 10 min at 4°C. Approximately 5% of the total cell lysate was analyzed to normalize for the total amount of Rap1. The remaining supernatant (about 500 g of proteins) was incubated with the GST-Sepharose beads (50 l of a 1:1 slurry containing about 10 g of immobilized GST fusion protein) for 2 h to allow for the association of activated Rap1 with the effector-GST fusion protein. Samples were washed twice with pull-down buffer, resuspended in Laemmli sample buffer and applied to SDS-polyacrylamide gels. The Rap1 band was then visualized using specific antibodies in a 1:1000 dilution by immunoblot analysis (see below).
Immunoblot Analysis-Stimulated cells were rinsed twice with icecold phosphate-buffered saline and then lysed in 400 l of ice-cold lysis buffer (10 mM Tris-HCl, pH 7.5, 140 mM NaCl, 1% Nonidet P-40, 1 mM orthovanadate, 1 mM phenylmethylsulfonyl fluoride, 2 g/ml aprotinin, 2 g/ml pepstatin, and 2 g/ml leupeptin). Cell lysate were centrifuged at 16,000 ϫ g for 5 min at 4°C, and the supernatant was incubated for 15 h at 4°C with primary antibodies and then with 15 l of protein A-agarose for 1 h. Immunoprecipitates were washed three times with lysis buffer and resuspended either in SDS-sample buffer for immunoblot experiments or in kinase buffer for immunocomplex protein kinase assay (see below). Western blot analysis was carried out either on 20 g of nuclear extracts (for phosphorylated CREB, phosphorylation stateindependent CREB, and p90 RSK ), 20 g of total cell lysate (for phosphorylated ERK and phosphorylation state-independent ERK), or immunoprecipitates for PYK2. Proteins resolved by 12.5% (PCREB, CREB, ERK, and PERK) or 7% (PYK2 and p90 RSK ) SDS-polyacrylamide gel electrophoresis were transferred to nitrocellulose membranes by tank blotting overnight at 4°C. The membranes were rinsed in Tris-buffered saline (TBS) (composition 10 mM Tris-HCl, pH 8.0, 150 mM NaCl, and 0.05% Tween 20), incubated for 1 h in TBS containing 4% nonfat dry milk (TBSM), and then incubated overnight at 4°C with primary antibodies in TBSM. After three 15-min washes in TBS, the membranes were incubated in TBSM for 1 h at room temperature with horseradish peroxidase-conjugated goat polyclonal anti-rabbit IgG for rabbit polyclonal primary antibodies, or with horseradish peroxidase-conjugated goat polyclonal anti-mouse IgG for mouse monoclonal primary antibodies. The chemiluminescent signals were detected using the Super signal kit, and the films were scanned using an Agfa T1200 scanner with Photolook software. The band intensity was quantified with NIH Image software. The same protocol was used for the detection of tyrosinephosphorylated PYK2 using the mouse monoclonal antibody specific to phosphotyrosine PY20, except that 1% bovine serum albumin was used instead of 4% nonfat dry milk.
Protein Kinase Assay-Cells were preincubated with protein kinase inhibitors as reported in the figure legends and then stimulated with the indicated agents. In vitro PKA and PKC activities were performed as previously reported (18,20).
Northern Blot Analysis-Striatal cultures were treated with appropriate stimuli for the indicated time and rinsed twice with phosphatebuffered saline. The RNA extraction reagent Trizol was added directly to each culture dish. Total RNA was extracted with a 1:10 volume of chloroform, and the upper aqueous phase was precipitated overnight at Ϫ20°C with a 1:1 volume of isopropyl alcohol. Total RNA was pelleted by centrifugation at 20,000 ϫ g for 15 min at 4°C, washed twice with ethanol, and quantified by UV spectrophotometry at 260 nm. Total RNA (10 g) was size-resolved on a 1.1% denaturing agarose gel containing 1 M paraformaldehyde in 20 mM MOPS, pH 7.0, 5 mM sodium acetate, 1 mM EDTA and capillary-blotted overnight in 10ϫ SSC onto a positively charged Hybond nylon membrane. The membranes were UV-autocross-linked and then prehybridized for 1 h at 65°C in 0.5 M sodium-phosphate buffer, pH 7.3, 7% SDS, 1 mM EDTA. Hybridization was carried out for 16 h at 65°C in 50 ml of prehybridization solution to which about 1 ϫ 10 6 cpm/ml of a full-length cDNA c-fos or 18 S ribosomal gene (internal loading control) cDNA random primed 32 Plabeled probes were added. Membranes were washed three times for 2 min in 300 ml of prewarmed (65°C) washing solution (50 mM Na 2 HPO 4 -NaH 2 PO 4 , pH 7.4, 1% SDS) and once in 100 ml of the same solution on a shaking platform for 30 min at 65°C. The radioactivity associated with the membranes was visualized and quantified with a Cyclone storage phosphor system (Packard Instrument Co.).
Statistical Analysis-Results are expressed as mean Ϯ S.E., and the significance of differences among treatments was evaluated using Student's t test for paired data or analysis of variance by Dunnett's test.

Activation of the cAMP/PKA Pathway Induces CREB Phosphorylation in Striatal Neurons-
The striatum is the brain area with the most abundant dopaminergic innervation, and striatal neurons in primary cultures express all of the dopaminergic receptor subtypes. We first sought to determine which receptor subtypes and signal transduction pathways are involved in CREB phosphorylation induced by the neurotransmitter dopamine in cultured striatal neurons. Dopaminergic regulation of Ser 133 CREB phosphorylation was investigated by using selective agonists and antagonists specific for the D 1 -like and D 2 -like dopaminergic receptor subtypes. Dopamine-induced increase in Ser 133 CREB phosphorylation (2.1 Ϯ 0.2-fold basal level, n ϭ 4) was mimicked by the D 1 -like receptor agonist SKF38393, not affected by the D 2 -like receptor antagonist sulpiride, and blocked by the D 1 -like antagonist SCH 23390 (Fig. 1A). To confirm the involvement of cAMP and to rule out any possible G protein-dependent activation of cAMP-independent signal transduction pathways, CREB phosphorylation was stimulated by agents capable of increasing cAMP formation downstream of dopamine receptors. The membrane-permeable cAMP analogue 8-CPT (200 M) and the direct activator of adenylylcyclase forskolin (20 M) induced a significant increase (3.2 Ϯ 0.5-fold basal level, n ϭ 4) of CREB phosphorylation, similar to that induced by 100 M dopamine.
The recent characterization of a family of cAMP-binding proteins that directly activates the Ras-like Rap1 protein (21), raises the possibility that the forskolin-induced CREB phosphorylation might be mediated by a cAMP-dependent but PKA-independent mechanism; this mechanism seems unlikely under our conditions, since forskolin-induced CREB phosphorylation was blocked by the PKA inhibitors KT5720 and H89 (Fig. 1B).
PKA-induced ERK Activation and CREB Phosphorylation Require PKC Activity-In different experimental models (16,22), the activation of the cAMP/PKA-dependent pathway leads to activation of the ERK pathway. To assess whether this cross-talk occurs in cultured striatal neurons, we performed a set of experiments using a phosphospecific antibody directed against the dually phosphorylated ERK isoforms ERK1 and ERK2. We found that DA and forskolin stimulated predominantly the phosphorylation of the ERK p42 isoform ERK2 (FSK: 4.5 Ϯ 0.8-fold over basal level, n ϭ 4; DA: 3.2 Ϯ 0.5-fold over basal level, n ϭ 4), and this effect was PKA-dependent, being reduced by the PKA inhibitor H89 (Fig. 2). PKA-induced activation of the ERK pathway was an intermediate step in PKA-induced CREB phosphorylation, since the ERK kinase (MEK) inhibitor PD98059 strongly blocked CREB phosphorylation induced by forskolin and dopamine (Fig. 3). In previous in vitro kinase assays, we found that, under our experimental conditions, the effect of PD98059 was specific for the ERK pathway and did not interfere with the activity of other kinases such as PKA and CaMKs (data not shown).
Since in the hippocampus the ERK pathway could be activated by PKC (22), we investigated the role of this kinase in cAMP/PKA-induced CREB phosphorylation. The PKC inhibitor bisindoleylmaleimide I strongly reduced DA-and FSKinduced ERK2 and CREB phosphorylation (Fig. 4A); the same results were obtained using two other specific PKC inhibitors, chelerythrine chloride and Gö6976 (Fig. 4B). In vitro kinase assays demonstrated that, under our experimental conditions, PKC inhibitors blocked PKC activity but did not interfere with nuclear PKA activity (Fig. 4C). To further confirm a causal relationship between PKC activation and ERK2/CREB phosphorylation, we investigated the functional consequence of short and long term treatment of striatal cells with the direct PKC activator PMA on ERK and CREB phosphorylation. When cells were stimulated for 10 min with 1 M PMA to activate PKC, the MEK inhibitor PD98059 blocked both PMA-induced ERK2 and CREB phosphorylation (Fig. 5A). Accordingly, when cells were treated for 18 h with 10 nM PMA to desensitize PKC, forskolin-induced ERK2 and CREB phosphorylation were significantly reduced, as compared with untreated cells (Fig. 5B). Taken together, these experiments strongly suggest that PKA causes CREB phosphorylation via a PKC/ERK-dependent pathway.
PKA-induced CREB Phosphorylation and CRE-dependent Gene Transcription Require Intracellular Ca 2ϩ -Previous studies have shown that neuronal Ca 2ϩ represents a link between the ERK-and PKA-dependent pathways (16). These observations prompted us to investigate the Ca 2ϩ requirement in PKA-induced CREB phosphorylation mediated by ERK in striatal neurons; a set of experiments was performed using the cell-permeable intracellular Ca 2ϩ chelator BAPTA-AM. When neurons were preincubated with BAPTA-AM (50 M) for 30 min, both ERK2 (Fig. 6A) and CREB phosphorylation (Fig. 6B) induced by forskolin were significantly reduced.
Having established the requirement of Ca 2ϩ in PKA-mediated CREB phosphorylation, we then analyzed the possible involvement of L-type Ca 2ϩ channels and NMDA receptors in this mechanism. Both L-type Ca 2ϩ channels and NMDA receptors have been previously shown to be involved in CREB phosphorylation and CRE-dependent gene expression (7). Preincubation of striatal cells with the competitive NMDA antagonist APV or the noncompetitive NMDA antagonist MK-801 did not interfere with forskolin-induced CREB phosphorylation (Fig.  7A). This finding clearly demonstrates that in striatal neurons NMDA receptors are not responsible for the intracellular Ca 2ϩ elevation required in forskolin-induced CREB phosphorylation.
In a subset of striatal neurons, PKA-induced phosphorylation of the L-type voltage-sensitive Ca 2ϩ channel increases the peak current through these channels, thereby augmenting the amount of Ca 2ϩ entering the cell (23). We therefore investigated whether extracellular Ca 2ϩ was required for PKA-induced CREB phosphorylation in striatal neurons. To our surprise, we found that neither the extracellular Ca 2ϩ chelator EGTA nor the specific L-type voltage-dependent Ca 2ϩ channel blocker nifedipine interfered with CREB phosphorylation induced by forskolin (Fig. 7B). To test whether an increase in intracellular Ca 2ϩ could induce CREB phosphorylation, we analyzed the effect of the SERCA pump inhibitor thapsigargin. As expected from the previous experiments, thapsigargin induced an increase (2.8 Ϯ 0.3-fold basal level, n ϭ 4) of CREB phosphorylation (Fig. 7B). Interestingly, we found that neither nifedipine nor MK-801 interfered with forskolin-induced ERK2 phosphorylation (Fig. 7C), strongly suggesting, although without proving definitively, that ERK2 activation mediates PKAdependent CREB phosphorylation.
To investigate whether intracellular Ca 2ϩ was also required for the cAMP/PKA-induced gene transcription, we measured the mRNA levels of the immediate early gene c-fos, which contains in its promoter a CRE sequence. Treatment of striatal neurons for 30 min with 20 M forskolin induced a 2.3 Ϯ 0.3-fold increase (n ϭ 4) of c-fos mRNA over basal levels; when striatal neurons were preincubated for 30 min with BAPTA-AM (50 M) to chelate intracellular Ca 2ϩ or with the MEK inhibitor PD98059 to block the ERK pathway, forskolininduced increase of c-fos mRNA was prevented (Fig. 8).
PKA Activation Releases Intracellular Ca 2ϩ -Our experiments with BAPTA-AM and thapsigargin indicate that PKA activation might induce the release of intracellular Ca 2ϩ from Ca 2ϩ store organelles in striatal neurons. To test this hypothesis, intracellular Ca 2ϩ measurements in striatal neurons were performed using the fluorescent Ca 2ϩ probe fura-2/AM. The Ca 2ϩ ionophore A23187 (5 M), which causes a large Ca 2ϩ entry, was first tested as a positive control (Fig. 9, open squares, n ϭ 4). Application of forskolin (20 M) to fura-2/AMloaded neurons caused a significant long lasting (about 30 s) and slowly decreasing rise in intracellular Ca 2ϩ levels (Fig. 9,  full circles, n ϭ 4). The amplitude, but not the shape, of the forskolin-induced Ca 2ϩ response was greatly reduced by preincubating the cells with the PKA inhibitor H89 (20 M; Fig. 9, open triangles, n ϭ 4). Finally, to study the role of extracellular Ca 2ϩ , fura-2/AM-loaded neurons were preincubated with the chelator of extracellular Ca 2ϩ EGTA (5 mM) and then stimulated with forskolin. In the absence of extracellular Ca 2ϩ , for-skolin still elicited a significant sharp, very short lasting (about 7 s) and fast decreasing rise in intracellular Ca 2ϩ levels (Fig. 9, open circles). Taken together, these data indicate that, in addition to playing a role in regulating voltage-dependent Ca 2ϩ channel function, PKA may also increase intracellular Ca 2ϩ by releasing Ca 2ϩ from intracellular stores.
PKA Stimulation Induces Rap1 and PYK2 Activation-In several experimental models, the family of small G proteins plays a key role in Ca 2ϩ -induced activation of ERK. Recently, it has been reported that Rap1, a Ras-like small G protein that may activate the ERK pathway by interacting with the kinase B-Raf, is activated by membrane depolarization in a PKAdependent manner (16) and by PKA via the ERK pathway (24). To test the possibility that Rap1 participates in the signaling pathway responsible for PKA-induced CREB phosphorylation, we decided to investigate the activation of Rap1 by means of a pull-down assay using the GST-ralRBD fusion protein. Stimulation of striatal cells with 20 M forskolin for 10 min induced a large increase in the amount of Rap1-GTP (Fig. 10A). We then investigated the effects of both kinase inhibitors and of agents capable of interfering with intracellular Ca 2ϩ levels. Forskolin-induced Rap1 activation was blocked by the PKA inhibitor H89, but not by the specific PKC inhibitor bisindoleylmaleimide I. Intracellular Ca 2ϩ release appears to be necessary and sufficient for Rap1 activation, because Rap1 activation was not observed in the presence of BAPTA-AM and was mimicked by elevation of intracellular Ca 2ϩ induced by thapsigargin (Fig.  10A). These data indicate that forskolin mainly stimulates Rap1 activation in a PKA/Ca 2ϩ -dependent but PKC-independent manner, suggesting that the role of PKC in PKA-dependent CREB phosphorylation is likely to be downstream of Rap1.
Finally, since in the central nervous system the concomitant elevation of intracellular Ca 2ϩ and activation of PKC leads to tyrosine phosphorylation and activation of PYK2 (25), we investigated whether PYK2 was activated by PKA under our experimental conditions. As expected, PKA activation induced an increase of tyrosine phosphorylation of PYK2 immunoprecipitated from total cell lysate of striatal neurons (Fig. 10B).

DISCUSSION
This study describes a novel mechanism by which activation of the cAMP/PKA-dependent pathway may stimulate CREB phosphorylation in striatal neurons (Fig. 11). Activation of PKA, induced by dopamine or forskolin, releases Ca 2ϩ from the intracellular stores, which in turn activates the Rap1⅐B-Raf complex whose main target is the phosphorylation of ERK. Finally, the activation of ERK induces CREB phosphorylation, probably via the translocation of the ERK⅐RSK complex into the nucleus. The prevailing model dictates that an increase of cAMP induces CREB phosphorylation via a mechanism in which the PKA catalytic subunit translocates into the nucleus and directly phosphorylates CREB; in this mechanism, the localization of PKA may contribute to enhance cAMP-dependent CREB phosphorylation and CRE-dependent gene expression (20). However, several recent studies have demonstrated that, in addition to this direct PKA-dependent pathway, other signaling pathways, such as the ERK pathway (10), can induce CREB phosphorylation in a PKA-dependent manner. The participation of the ERK pathway appears to be very important for PKA-dependent transcription, because several neurotransmitter receptor systems (22) as well as depolarization-induced Ca 2ϩ influx (16) have been shown to activate the ERK pathway via a PKA-mediated mechanism. This mechanism appears to have an important functional role, since a previous study in Aplysia has shown that long term facilitation involves PKA-dependent activation of ERKs (26). In agreement with these data, we have demonstrated that in striatal neurons the novel mechanism involving PKA-induced intracellular Ca 2ϩ release participates in ERK and CREB phosphorylation and ultimately regulates the transcription of the immediate early gene c-fos.
Previous studies in various experimental models have shown that the functionality of intracellular Ca 2ϩ store organelles could be modulated by PKA-dependent phosphorylation in vivo and in vitro. The neuronal type-I IP3 receptor, immunopurified from rat cerebellar membranes, can be phosphorylated by PKA in vitro (27), and activation of PKA induces phosphorylation of the type-I IP3 receptor in cerebellar slices (27) and in cultured cells (28,29). Notably, a recent paper has described that, in cardiac tissue, PKA-dependent phosphorylation of ryanodine receptor/Ca 2ϩ release channel regulates the channel open prob-ability (30). These studies suggest that PKA-dependent phosphorylation could modulate the functionality of intracellular Ca 2ϩ store organelles by increasing their sensitivity to physiological agonists or toward external Ca 2ϩ . Given the importance of the cross-talk between PKA-and Ca 2ϩ -dependent pathways, the precise mechanism of PKA-mediated intracellular Ca 2ϩ release will certainly become a focus of future studies in neuronal cell physiology.
In PC12 cells and primary cortical neurons, ERK activation induced by Ca 2ϩ entry through voltage-sensitive channels or by the release of Ca 2ϩ from intracellular stores is mediated by the small G protein Ras (14). Several factors, such as the route and magnitude of Ca 2ϩ influx or the subcellular compartmentalization of protein kinases in the cell type under investigation, contribute to determine which signaling pathway upstream of ERK is activated. Several recent studies performed in various cell types have demonstrated that Ca 2ϩ , in addition to Ras, may also activate ERK via another route involving the Ras . Preincubation with these agents did not modify basal Rap1-GTP levels (not shown). Total cell lysate (0.5 mg) was processed for the pull-down assay of Rap1 using the GST-ral-RBD protein as described under "Experimental Procedures." Western blot analysis was performed on the pulled down Rap1-GTP (upper panel) and on total Rap1 (lower panel) in a 5% (25-g) extract of total cell lysate as a loading control (bottom panel). Data are reported as Rap1-GTP immunoreactivity, and the average -fold increase (mean Ϯ S.E.) over basal level is reported in the bar graph (n ϭ 4). *, p Ͻ 0.05 compared with forskolin. B, cells were stimulated for 10 min with forskolin (FSK, 20 M), and then PYK2 was immunoprecipitated from total cell lysate (1 mg). The immunoprecipitate was analyzed by Western blot with antiphosphotyrosine (WB P-Tyr, upper panel) or anti-PYK2 (WB PYK2, lower panel) antibodies as described under "Experimental Procedures." Data are reported as Tyr(P) immunoreactivity (IR), and the average -fold increase (mean Ϯ S.E.) over basal level is reported in the bar graph (n ϭ 4). *, p Ͻ 0.05 compared with basal level.
homologue Rap1 and its effector Raf-like kinase B-Raf. The Rap1⅐B-Raf system participates in PKA-mediated ERK activation induced by growth factors in PC12 cells (14,31,32) and by ␤ 2 adrenergic receptor stimulation in HEK 293 cells (24). In agreement with these reports, we propose that in striatal neurons the Rap1⅐B-Raf system is an intermediate step in the activation of the ERK pathway triggered by PKA-induced intracellular Ca 2ϩ release. This model is in contrast with what was found by other authors in PC12 cells (16), in which depolarization-induced Ca 2ϩ entry activates the ERK pathway via a two-step mechanism involving first PKA and then Rap1. In a previous study (33) and in experiments not reported here, we found that, in cultured striatal neurons, Ca 2ϩ entry induced by a Ca 2ϩ ionophore or by depolarization inhibits cAMP formation, probably interacting with the Ca 2ϩ -inhibitable type V adenylyl cyclase isoform, which is abundantly expressed in these neurons (18).
The complexity of Rap1 activation mechanisms in neuronal cells has been emphasized by the characterization of new families of Rap guanine nucleotide exchange factors activated by Ca 2ϩ and diacylglycerol (34) or by cAMP and Ca 2ϩ (17,21). On the other hand, our findings demonstrate that, in agreement with another report (35), Rap1 activation is induced by PKAdependent release of intracellular Ca 2ϩ (35). The differences in the mechanism of Rap1 activation are probably due to the differential expression of adenylyl cyclase isoforms and small G protein guanine nucleotide exchange factors in the cell types under investigation. Moreover, since Rap1 activation is not triggered by a PKA-or PKC-dependent phosphorylation, two mechanisms may be considered. First, Rap1 could be directly activated by the novel Rap1GEFs, which in turn are directly activated by Ca 2ϩ . Second, Rap1 could be activated indirectly via a multiple step mechanism involving the Ca 2ϩ -dependent stimulation of PYK2, which then interacts with Rap1GEF C3G (36). In agreement with this latter hypothesis, in the present study we have found that activation of the cAMP/PKA-dependent pathway leads to activation of the tyrosine kinase PYK2 (25). This is consistent with the observation that, in striatal brain slices, stimulation of D 2 receptors induces tyrosine phosphorylation of PYK2 via a phospholipase C-mediated intracellular Ca 2ϩ release (37). So far, it is still unclear whether in striatum activation of PYK2 is implicated in CREB phosphorylation, because D 2 -dependent CREB phosphorylation was not blocked by the broad spectrum tyrosine kinase inhibitor genistein (37).
Although the experiments presented in this study show that in striatal neurons PKC inhibitors block PKA-induced CREB phosphorylation, no conclusion can be drawn about the mechanism by which the cAMP/PKA-dependent pathway activates PKC and at what level PKC modulates CREB phosphorylation. Having excluded a direct role of PKC in Rap1 activation in our model, it is reasonable to assume that PKC acts downstream of Rap1 and upstream of B-Raf. This hypothesis is supported by studies showing that PKC may play a facilitator role on the small G protein-dependent ERK pathway activation by enhancing Raf-1 or B-Raf functionality (38).
In conclusion, we have found that intracellular Ca 2ϩ release induced by PKA leads to CREB phosphorylation via a PKCand ERK-dependent mechanism, most likely involving Rap1 activation. Given the complexity of cross-talk between the cAMP/PKA and the ERK pathways, the plethora of signaling pathways converging on ERK, and the wide array of transcription factors activated by these signaling pathways, this and other studies (16,23) raise fundamental questions concerning the role of Rap1 and the functional effects of its activation at the level of transcription factor activation and CRE-dependent gene expression.