Phosphorylation of the Ras-GRF1 exchange factor at Ser916/898 reveals activation of Ras signaling in the cerebral cortex.

The Ras-GRF1 exchange factor, which is regulated by increases in intracellular calcium and the release of G beta gamma subunits from heterotrimeric G proteins, plays a critical role in the activation of neuronal Ras. Activation of G protein-coupled receptors stimulates an increase in the phosphorylation of Ras-GRF1 at certain serine residues. The first of these sites to be identified, Ser(916) in the mouse sequence (equivalent to Ser(898) in the rat sequence), is required for full activation of the Ras exchange factor activity of Ras-GRF1 by muscarinic receptors. We demonstrate here that Ras-GRF1 is highly expressed in rat brain compared with the Sos exchange factor and that there is an increase in incorporation of (32)P into Ser(898) of brain Ras-GRF1 following activation of protein kinase A. Phosphorylation of Ras-GRF1 at Ser(916) is also required for maximal induction of Ras-dependent neurite outgrowth in PC12 cells. A novel antibody (termed 2152) that selectively recognizes Ras-GRF1 when it is phosphorylated at Ser(916/898) confirmed the regulated phosphorylation of Ras-GRF1 by Western blotting in both model systems of transfected COS-7 and PC12 cells and also of the endogenous protein in rat forebrain slices. Indirect confocal immunofluorescence of transfected PC12 cells using antibody 2152 demonstrated reactivity only under conditions in which Ras-GRF1 was phosphorylated at Ser(916/898). Confocal immunofluorescence of cortical slices of rat brain revealed widespread and selective phosphorylation of Ras-GRF1 at Ser(898). In the prefrontal cortex, there was striking phosphorylation of Ras-GRF1 in the dendritic tree, supporting a role for Ras activation and signal transduction in neurotransmission in this area.

The Ras GTPases are timed molecular switches that cycle between GDP-and GTP-bound forms to control pathways of cellular growth and differentiation (1). In addition to the well established roles for Ras in the proliferation of normal and malignant cells (2), there is increasing evidence that Ras also regulates critical functions in terminally differentiated cells such as neurons (3). The GTPase cycle is controlled by guanine nucleotide exchange factors (GEFs) 1 and GTPase-activating proteins (GAPs) (4), with the balance between the effective GEF and GAP activities determining the activation state of Ras because it is the GTP-bound form that activates downstream effector pathways. The principal control in many instances may be the activation process through the activity or subcellular localization of the GEF, as there is increasing evidence that these Ras activators are highly regulated (5). In other cases, such as type 1 neurofibromatosis, where there is loss of the Ras-GAP activity of the protein neurofibromin (6), physiological control is shown to require the deactivation step. Indeed, oncogenic mutations in Ras that block the GAP-catalyzed deactivation of Ras from the GTP-to GDP-bound states (7), emphasize the importance of this switch mechanism.
The Ras-GRF1 exchange factor (8), which is also termed CDC25 Mm (9,10), is highly expressed in neurons of the central nervous system (11)(12)(13). Ras-GRF1 is expressed predominantly at synapses (14), a neuronal subcellular distribution it shares with the Ras-GAP termed SynGap (15). This colocalization suggests that the proteins may reciprocally regulate Ras at synapses. Interestingly, Ras (16) and the Ras effector mitogenactivated protein kinase (17,18) have been implicated in the control of synaptic plasticity, which is the proposed cellular corollary of memory storage, whereas mice that lack Ras-GRF1 have defects in memory (19,20). Thus, Ras-GRF1 may participate in the regulation of synaptic plasticity in the central nervous system.
Ras-GRF1 couples heterotrimeric G proteins (21)(22)(23)(24) and calcium signals (8,25) to the activation of Ras. The activation of Ras-GRF1 by G protein-coupled receptors is closely associated with an increase in its phosphorylation at certain serine residues (24), with the first of these to be identified being Ser 916 (in the mouse sequence, equivalent to Ser 898 in the rat sequence) (26). Phosphorylation of Ser 916/898 , which is an in vivo and in vitro substrate for protein kinase A (PKA), is necessary for full activation of the Ras-GEF activity of Ras-GRF1 (26). The current study confirms that Ser 916/898 is a physiologically relevant site of regulated phosphorylation in the endogenous Ras-GRF1 exchange factor that is expressed in rat forebrain. Localization of this regulatory phosphorylation event to the apical dendrites of prefrontal pyramidal cells suggests that Ras signaling is activated at these loci.
Metabolic Labeling and Digestion of Ras-GRF1-Forebrain slices cut from postnatal day 15 rat brains were labeled in six-well tissue culture plates in 1 ml/well phosphate-free Dulbecco's modified Eagle's medium (Invitrogen) with 2 mCi of [ 32 P]orthophosphate (ICN, Costa Mesa, CA) for 2 h in a 37°C incubator that was supplied with 95% O 2 and 5% CO 2 . The slices were lysed, and Ras-GRF1 was immunoprecipitated using antibody sc-224 (Santa Cruz Biotechnology, Santa Cruz, CA) as described (23). Immunoprecipitated labeled Ras-GRF1 was processed for Western blotting and digestion with cyanogen bromide as described (26).
Development of Polyclonal Antibody 2152 against Ras-GRF1 Ser(P) 916 -Antibodies were developed in rabbits immunized with a 13-residue phosphopeptide that represents the region flanking Ser 916 of mouse Ras-GRF1. This sequence is highly conserved (with only three conservative substitutions) in rat Ras-GRF1 (8). The specificity of the antibodies was tested, and affinity-purified antibody 2152 was characterized as a selective agent for recognition of both mouse and rat Ras-GRF1 only when phosphorylated at the relevant serine (see below). Western blots were developed by enhanced chemiluminescent detection (32).
Forebrain slices of postnatal day 15 rat brain were fixed in 4% paraformaldehyde in phosphate-buffered saline and incubated in 50% sucrose/phosphate-buffered saline overnight at 4°C. The slices were then permeabilized for 10 min with methanol that had been precooled to Ϫ20°C; rinsed with phosphate-buffered saline; and blocked with phosphate-buffered saline containing 2% bovine serum albumen, 5% goat serum, and 0.25% Triton X-100 for 4 h at room temperature. Indirect confocal immunofluorescence was carried out using the following primary polyclonal antibodies: antibody 2152 (anti-Ras-GRF1 Ser(P) 916/898 ; 1:200 dilution); antibody sc-224 (anti-Ras-GRF1; 1:200 dilution); or, as a negative control, antibody sc-863 (anti-Ras-GRF1 antibody that is competent for Western blotting, but does not recognize the protein in immunoprecipitation or immunofluorescence protocols; 1:200 dilution). The secondary antibody used was Oregon Green-coupled anti-rabbit immunoglobulin (1:150 dilution). Pictures were taken with an Olympus Fluoview laser scanning confocal microscope using a ϫ20 objective.

Ras-GRF1 Is Highly Expressed in Rat
Brain-It is now clear that there are multiple GEFs that can serve to activate Ras proteins, including those in the Sos, Ras-GRF1/CDC25 Mm , Rasguanyl nucleotide-releasing protein, and other families (5). Much work has been performed on the ubiquitous Sos exchange factors that activate Ras in response to stimulation of tyrosine kinase-mediated signals (28,35,36). In contrast, Ras-GRF1 is expressed predominantly in the neurons of the central nervous system (11)(12)(13) and activates Ras in response to G proteincoupled and calcium signals (8,(21)(22)(23)(24). To examine the relative expression levels of Sos and Ras-GRF1 in rat brain, a quantitative Western blot protocol was developed. Ras-GRF1 and Sos1 were expressed in COS-7 cells with identical triple-HA1 epitope tags at their N termini. The lysates from these transfections were standardized by Western blotting with anti-HA1 monoclonal antibody 12CA5. The standardized HA1 3 -Ras-GRF1 and HA1 3 -Sos1 lysates were then used to establish the relative sensitivities of two polyclonal antibodies directed against the Ras-GRF1 and Sos proteins (Fig. 1). The polyclonal FIG. 1. Ras-GRF1 is expressed at a higher level compared with Sos in rat forebrain slices. Lysates of COS-7 cells expressing HA1 3 -Ras-GRF1 and HA1 3 -Sos1 were subjected to Western blotting with the anti-HA1 monoclonal antibody (left panels) and standardized by dilution such that equivalent levels of the expressed proteins were present. These amounts were then run in parallel with lysates of rat brain slices (right panels) and subjected to Western blotting with polyclonal antibodies to either Ras-GRF1 (upper right panel) or Sos (lower right panel). In all cases, a graded scale of four 3-fold dilutions of the lysates was loaded to ensure that the signals could be measured in a range where the signals were not saturated. Data shown are representative of four independent experiments and were collected digitally using a Fuji Film LAS-1000 Plus imaging system. Quantification was performed with ImageGauge Version 3.3 software. Ras-GRF1 proteins were immunoprecipitated and digested with cyanogen bromide, and the fragments were separated by SDS-PAGE using a Tricine cathode buffer to sharpen resolution of the phosphopeptides (26). The results shown are autoradiographs of a region of the dried gels, with molecular mass markers (in kilodaltons) shown to the left. Exposure times were overnight for the COS-7 cell-derived samples and 2 weeks with intensifying screens for the brain-derived samples. The arrow indicates a phosphopeptide fragment that was stimulated by the activation of PKA and represents phosphorylation of Ser 898 (26). The result is representative of three independent experiments. antibodies detected significantly more Ras-GRF1 than Sos in lysates of rat brain forebrain, even though the sensitivity of the anti-Sos antibody is greater. Correcting for the sensitivity of the antibodies, we found that there was 11-fold more Ras-GRF1 than Sos in rat forebrain slices.
Ras-GRF1 Is Phosphorylated at Ser 898 in Forskolin-treated Slices of Rat Brain-Ras-GRF1 is regulated by a complex of mechanisms that include regulated phosphorylation of serine residues (24). Phosphorylation of Ser 916 , which is an in vivo substrate for PKA, has been shown to be necessary in model systems for full activation of the Ras-GEF activity of Ras-GRF1 by muscarinic receptors (26). To test whether phosphorylation of this residue also occurs in endogenous Ras-GRF1, we metabolically labeled rat forebrain slices with [ 32 P]orthophosphate, stimulated the slices with forskolin to activate PKA, and immunoprecipitated Ras-GRF1. Ras-GRF1 was then cleaved at methionine residues using cyanogen bromide, and the fragments were separated by peptide PAGE (Fig. 2). Phosphoryla-tion of mouse Ras-GRF1 at Ser 916 produces a 32 P-labeled fragment with an apparent mobility of ϳ6.5 kDa (26). The positions of the methionines that flank this phosphorylation site are conserved between mouse and rat. Transfection of COS-7 cells with rat Ras-GRF1 demonstrated that the equivalent phosphorylation (at Ser 898 ) was revealed in a similarly sized cyanogen bromide cleavage product. Stimulation of forebrain slices of rat brain with forskolin induced the appearance of 32 P in the equivalent fragment of endogenous Ras-GRF1, confirming that Ser 898 is a site of regulated phosphorylation in the brain.
Phosphorylation of Ras-GRF1 at Ser 916 Promotes Ras-dependent Outgrowth of Neurites in PC12 Cells-We have previously shown that phosphorylation of Ras-GRF1 at Ser 916 is necessary for maximal activation of Ras-GEF activity in a biochemical assay with recombinant Ras substrate (26). Because we have shown here that this phosphorylation event occurred in the endogenous exchange factor in the brain, we investigated whether this phosphorylation plays a critical role in Ras activation in a neuronal context. We used the well established model of neurite extension from PC12 cells in response to Ras activation (37). The results show that expression of either the C-terminal half of Ras-GRF1 (Ras-GRF1⌬N) or a Myc-tagged H-Ras protein alone did not induce neurite outgrowth (Fig. 3). It has previously been demonstrated that wildtype Ras proteins (those without a constitutively activating, oncogenic mutation) are not sufficient to induce neurite outgrowth (37), and this result further suggests that the Ras-GRF1⌬N protein is not able to activate the endogenous Ras protein in PC12 cells to induce neurite extension. Because H-Ras is the preferred in situ substrate for Ras-GRF1 (29), we assayed for neurite extension in cotransfections of Ras-GRF1⌬N and wild-type H-Ras and found robust outgrowth. Deletion of the Ras-GRF1 construct to leave the C-terminal third (Ras-GRF1⌬900), although removing the Ras exchange motif sequence that is common to Ras-GEFs (5), did not reduce the H-Ras-dependent induction of neurite outgrowth. Further deletion to leave the C-terminal quarter of Ras-GRF1 (Ras-GRF1⌬976) did, however, significantly reduce neurite extension. This deletion leaves intact the CDC25 Ras-GEF domain that is necessary for biochemical activity (38) and so reveals a stimulatory function within residues 900 -975 of Ras-GRF1. The same reduction in neurite extension was produced by point mutation of the Ser 916 phosphorylation site to alanine (Fig. 3).
Development of an Antibody That Selectively Recognizes Ras-GRF1 Ser(P) 916/898 -The appearance of the 32 P-labeled fragment of Ras-GRF1 following cyanogen bromide cleavage pro- vides an assay for the regulated phosphorylation of Ser 916/898 . This assay is time-consuming and provides poor sensitivity, however. To develop a more sensitive and powerful assay approach, antibodies that selectively recognize the Ser(P) 916/898 form of Ras-GRF1 were developed. To confirm the selectivity of antibody 2152, it was tested against recombinant Ras-GRF1 proteins that were subjected to phosphorylation by PKA (Fig.  4). The data show that antibody 2152 exhibited great selectivity for recognition of Ras-GRF1 only when it was phosphorylated, with detectable signal produced from as little as 10 ng of Ras-GRF1 Ser(P) 916 , whereas unphosphorylated Ras-GRF1 was only detected when microgram quantities were present. Note that PKA treatment of the S916A mutant protein did not produce any reactivity to antibody 2152.
Regulated Phosphorylation of Ras-GRF1 at Ser 916/898 -To test whether antibody 2152 recognizes both mouse and rat Ras-GRF1 expressed in mammalian cells, it was tested against whole cell lysates of COS-7 cells that had been transfected to express Ras-GRF1 and then treated with various agonists (Fig.  5). The data show that antibody 2152 reacted with a single band that was present only in the cells transfected to express Ras-GRF1 and that this band had the same mobility as that recognized by an independent anti-Ras-GRF1 antibody (sc-863). Furthermore, it is clear that the recognition of the mouse and rat Ras-GRF1 proteins was similar and that all recognition was absolutely dependent on phosphorylation at Ser 916/898 , as no reactivity occurred in cells expressing Ras-GRF1(S916A). Both forskolin and serum (the latter to a lesser extent, which is clearer when the Western blots were given prolonged exposure) stimulated an increase in the phosphorylation of Ras-GRF1 at Ser 916/898 . Interestingly, treatment of the cells with thapsigargin, which increases cytosolic calcium levels (39), failed to increase phosphorylation of Ras-GRF1 at Ser 916/898 above the basal level. This result supports previous observations on transfected fibroblasts that increases in calcium do not stimulate the phosphorylation of Ras-GRF1 (40), although it has been reported that Ras-GRF1 is an in vitro substrate for calmodulin-dependent kinase II (14).
Multiple Agonists Stimulate the Phosphorylation of Ras-GRF1 at Ser 916/898 in PC12 Cells-To investigate the phosphorylation of Ras-GRF1 in a neuronal context, we coexpressed Ras-GRF1 with muscarinic receptors in PC12 cells. Western blotting of PC12 cell lysates (Fig. 6) demonstrated that activation of muscarinic receptors by carbachol, stimulation of PKA by forskolin, and activation of endogenous Trk receptors by nerve growth factor (NGF) all increased the phosphorylation of Ras-GRF1 at Ser 916 . To extend these results, either HA1 3 -Ras-GRF1 or its S916A mutant was coexpressed in PC12 cells with muscarinic receptors, and the cells were stimulated and then fixed and processed for indirect confocal immunofluorescence (Fig. 7). It is clear that antibody 2152 reactivity was again dependent on phosphorylation at Ser 916 , as no reactivity was seen in untransfected cells, in transfected cells that were not stimulated, or in stimulated cells that were transfected with the Ras-GRF1(S916A) mutant. In agreement with the Western blot results, carbachol, forskolin, and NGF all induced the phosphorylation of Ras-GRF1 at Ser 916 .
Regulated Phosphorylation of Ras-GRF1 at Ser 898 in Rat Forebrain Slices-To investigate whether endogenous Ras-GRF1 is regulated by phosphorylation in a manner similar to that defined in the model systems, antibody 2152 was used to assess phosphorylation of Ras-GRF1 in rat forebrain slices. It was difficult to detect any reactivity to antibody 2152 in lysates of brain slices by Western blotting, perhaps suggesting that only a subset of Ras-GRF1 proteins (perhaps, in particular, neurons) is phosphorylated at Ser 898 . When Ras-GRF1 was first immunoprecipitated from the lysate using an antibody directed against the C terminus (sc-224), the regulated phosphorylation at Ser 898 became apparent by Western blotting (Fig. 8). There was little reactivity in Ras-GRF1 from untreated slices, but there were clear increases following stimulation with carbachol, forskolin, or NGF. Thus, just as in the expression systems constructed in PC12 cells, endogenous Ras-GRF1 is phosphorylated at Ser 898 following activation of muscarinic or Trk receptors or PKA.
To investigate the distribution of Ras-GRF1 Ser(P) 898 in the cortex, brain slices were fixed and processed for indirect confocal immunofluorescence with antibodies 2152 (anti-Ras-GRF1 Ser(P) 916/898 ) and sc-224 (anti-Ras-GRF1). Prefrontal cortex includes two classes of neurons, GABAergic interneurons and glutamate-secreting pyramidal cells. The latter class is easily recognizable by its distinctive morphology, including a prominent single apical dendrite that branches in the upper layers. In these immunolocalization studies, we observed expression of Ras-GRF1 throughout pyramidal neurons, including the cell FIG. 7. Regulated phosphorylation of Ras-GRF1 at Ser 916 in PC12 cells as revealed by indirect confocal immunofluorescence. PC12 cells were cotransfected with pKH3Ras-GRF1 (left panels) or pKH3Ras-GRF1(S916A) (right panels) plus expression vectors for human muscarinic receptor subtypes 1 and 2. The cells were deprived of serum overnight and then stimulated with 100 M carbachol, with 100 M IBMX plus 20 M forskolin, or with 10 ng/ml NGF. The cells were fixed and processed for indirect confocal immunofluorescence using red detection for the HA1 epitope tags at the N termini of the Ras-GRF1 constructs and green detection for antibody 2152 (raised against Ras-GRF1 Ser(P) 916 (Ras-GRF1phospho916)). The fluorescence images are presented overlaid with a phase-contrast image of the cells that reveals the presence of untransfected control cells. Note that green fluorescence due to antibody 2152 reactivity was found only when wild-type Ras-GRF1 (Ras-GRF1,wt) was present (thus colocalized with red fluorescence and appears yellow) and only in cells that had been stimulated with agonists that induce Ser 916 phosphorylation. Data shown are representative of five independent experiments. Ras-GRF1,916A, Ras-GRF1(S916A). bodies ( Fig. 9), whereas phosphorylated Ras-GRF1 was strikingly localized to the apical dendrites of pyramidal cells (Fig. 9). DISCUSSION This study identifies Ser 916/898 of the Ras-GRF1 exchange factor as a target for regulated in vivo phosphorylation in both transfected model systems of COS-7 and PC12 cells and endogenously in rat brain slices. The role of Ras-GRF1 in central neurons is intriguing (41,42), but confusing, as the phenotypes of mice deficient in Ras-GRF1 are not in complete agreement. Brambilla et al. (19) described Ras-GRF1-deficient mice that were impaired in amygdala-dependent learning and memory tests such as emotional conditioning of memory consolidation. The authors stated that there were no deficits in the Morris water maze test, suggesting that hippocampal function was normal, but did find larger field excitatory postsynaptic potentials upon electrophysiological recording from both the amygdala and hippocampus (19). Subsequent work has detailed the hyperexcitability of hippocampal neurons from these mice (43). Giese et al. (20) described smaller-than-control mice with defects in hippocampus-dependent learning and memory as revealed in contextual discrimination, social transmission of food preference, and Morris water maze tests. A third group of Ras-GRF1-deficient mice, studied by Itier et al. (44), were also found to be small, with evidence that this characteristic is due to a decrease in postnatal insulin-like growth factor I production that is secondary to a reduction in pituitary growth hormone levels. Because pituitary growth hormone is under the control of hypothalamic neurotransmission and Ras-GRF1 is normally expressed in the postnatal hypothalamus, the authors suggested that hypothalamic function may be reduced in their Ras-GRF1-deficient mice, but did not report results from any other tests of neuronal function (44). Nevertheless, there is strong support for the concept that Ras-GRF1 is highly expressed in certain central neurons and has the potential to play critical roles in processes (such as memory) to which Ras signaling has previously been linked (3,45,46). In this regard, it is interesting to note that Ras-GRF1 is regulated by cAMP/ PKA phosphorylation (26,47), signaling elements that are thought to play a key role in synaptic plasticity and hence memory storage in the brain (17). The identification of activated Ras-GRF1 in the apical dendrites of prefrontal pyramidal neurons in this study is completely consistent with its postulated role as an integrator of neuronal signal transduction (41,42).
One possibility that might have resolved the differences in the phenotypes of mice that are deficient in Ras-GRF1 would be variable compensation from the highly homologous Ras-GRF2 protein (48). Recently, however, mice deficient in Ras-GRF2 have been described to have neither an apparent phenotype nor any compensatory changes in the expression of Ras-GRF1 (49). Conversely, Ras-GRF1-deficient mice have no alteration in Ras-GRF2 expression. Furthermore, results from mice doubly deficient in both Ras-GRF1 and Ras-GRF2 were also reported. Remarkably, there is no further phenotypic effect beyond that produced by loss of Ras-GRF1 alone, suggesting that there is no overlap of function between these highly similar GEFs (49). In this regard, it is important to note that the Ser 916/898 phosphorylation site, although conserved in Ras-GRF1 proteins from rodents and humans, is absent in Ras-GRF2. Thus, it is plausible that this regulatory mechanism is part of the unique function of Ras-GRF1 that cannot be replaced by Ras-GRF2.
Phosphorylation of Ras-GRF1 at Ser 916 has previously been shown to be essential for maximal activation of Ras-GRF1 in a Rat forebrain slices were stimulated with 100 M carbachol, with 100 M IBMX plus 20 M forskolin, or with 10 ng/ml NGF for 20 min at room temperature. Lysates were prepared, with 2.5% reserved for direct analysis. Ras-GRF1 was immunoprecipitated from 97.5% of the lysate using antibody sc-224; the immunoprecipitates were then split, with 10% subjected to Western blotting for total Ras-GRF1 using antibody sc-863 (left panel) and 90% subjected to Western blotting with antibody 2152 (raised against Ras-GRF1 Ser(P) 898 (GRF1(phospho898)); right panel). Negative control immunoprecipitations ((-ve) i.p.) were performed with rabbit anti-mouse immunoglobulin (Jackson Laboratories). Data shown are representative of four independent experiments.
FIG. 9. Regulated phosphorylation of endogenous Ras-GRF1 at Ser 898 in rat cerebral cortex as revealed by indirect confocal immunofluorescence. Forebrain slices of postnatal day 15 rat brains were fixed and processed for indirect confocal immunofluorescence. The primary antibodies used were as follows: antibody 2152, raised against Ras-GRF1 Ser(P) 898 (A, the yellow arrows indicate apical dendrites of pyramidal cells in the medial prefrontal cerebral cortex, and the inset shows a lower magnification view); antibody sc-224, raised against total Ras-GRF1 (B1, the yellow arrows indicate apical dendrites of pyramidal cells in the medial prefrontal cerebral cortex, and the red arrows indicate cell bodies); and antibody sc-863, which recognizes Ras-GRF1 only in Western blot protocols, but not in immunofluorescence and so was used as a staining control (B2). biochemical assay with recombinant Ras substrate (26). In the current study, we have shown that there is a regulatory function for residues 900 -975 of Ras-GRF1 to increase Ras activation in the PC12 system and that this function is also lost by point mutation of the Ser 916 phosphorylation site. Although Ras-GRF1 is phosphorylated at multiple sites in response to agonist stimulation, Ser 916 is the only detectable site of phosphorylation in this region (26). These results therefore support the significance of Ras-GRF1 phosphorylation at Ser 916/898 both as a marker for increased Ras activation within a neuronal context and in the physiological control of neuronal Ras activation.
Muscarinic stimulation was previously demonstrated to increase the phosphorylation and activity of Ras-GRF1 in mouse brain explants (23); thus, stimulation of rat brain Ras-GRF1 Ser 898 phosphorylation by carbachol treatment was expected. Activation of muscarinic receptor signaling in the cortex is also correlated with memory function in humans (50), with broad pharmacological significance (51). Whether Ras-GRF1 phosphorylation and activation of Ras in response to muscarinic agonists are a physiological part of this process remains to be demonstrated.
The Ser 916/898 context fits a consensus PKA substrate sequence (52) and does indeed serve as a PKA site both in vitro and in vivo in model systems (26). Phosphorylation at this site was not, however, recorded in in vitro experiments with PKA and recombinant Ras-GRF1 by Baouz et al. (47). The basis for this difference is unclear, but the current study provides further in vivo support for this PKA-dependent phosphorylation through demonstration of stimulation of Ser 916/898 phosphorylation by forskolin treatment. More surprising was the robust ability of NGF to stimulate Ras-GRF1 phosphorylation at Ser 916/898 in both the transfected PC12 model and the endogenous protein in rat brain slices. Previously, Ras-GRF1 has generally been found to be regulated independently from growth factor/tyrosine kinase pathways (21,23), although Ras-GRF1 was reported to associate with the NGF-activated TrkA receptor (53). The current result suggests that the interaction of the TrkA and Ras-GRF1 signaling pathways may be functionally important in those neurons that express both proteins.
In addition to Ras-GEF activity that is stimulated by a G protein ␤␥ subunit-dependent pathway and is associated with increased serine phosphorylation (23), Ras-GRF1 has also been demonstrated to exhibit Rac-GEF activity that can be induced by the coexpression of G protein ␤␥ subunits and is associated with increased tyrosine phosphorylation (54). Both of these pathways can be induced by the same agonist, lysophosphatidic acid, at least in NIH-3T3 fibroblast model systems (24,55). Tyrosine phosphorylation of Ras-GRF1 that is induced by v-Src increases Rac-GEF activity, but does not increase Ras-GEF activity (56). In contrast, ACK1-dependent tyrosine phosphorylation of Ras-GRF1 increases Ras-GEF activity, but does not induce Rac-GEF activity (57). These results suggest that there may be multiple sites of tyrosine phosphorylation on Ras-GRF1 with distinct effects. Because ACK1 can couple to both growth factor pathways and the CDC42 small GTPase, it is possible that ACK1 may participate in the interactions between the NGF/TrkA pathway and Ras-GRF1 or between CDC42 and Ras-GRF1 (58,59). In view of the close parallels between the control of the Ras-GEF and Rac-GEF activities of Ras-GRF1, it is possible that phosphorylation of Ser 916/898 may also participate in the control of Rac exchange activity, but this remains unclear.
Ras-GRF1 provides an example for other GEFs that are also subject to complex regulation through phosphorylation (60, 61), allosteric interactions (62)(63)(64), and subcellular redistribution (65). The activation of Ras-GRF1 is also complicated by the potential for dimerization and for effects that may be mediated through the activation of Raf, the Ras effector, even in the absence of apparent further activation of Ras (66). It is likely that these complexities reflect the underlying function of Ras-GRF1 and other GEFs to serve as key integrators and controllers of signaling pathways with activities beyond the simple control of GTP binding to their substrates. For example, an emerging model suggests that the GEFs provide a scaffold or framework function that can select the downstream targets of their substrate GTPases (67,68). Thus, it is essential to consider the cellular and subcellular context within which the GEF acts to activate Ras signal transduction. In this regard, the selectivity and sensitivity of anti-Ras-GRF1 Ser(P) 916/898 antibody 2152 will provide a powerful approach toward the elucidation of Ras activation pathways in central neurons.