Phosphorylation of rap1GAP in Vivo and by CAMP-dependent Kinase and the Cell Cycle ~ 3 4 " ~ " ~ Kinase in Vitro*

raplGAP is a GTPase activating protein that specif-ically stimulates the GTP hydrolytic rate of the ras- related protein p2lraP'. raplGAP undergoes post- translational modification that causes a substantial change in its mobility on sodium dodecyl sulfate-poly-acrylamide gels. At least part of this modification is due to the phosphorylation. Expression of a raplGAP cDNA in insect cells labeled with "Pi resulted in high level incorporation of radioactivity into serine residues of the expressed protein. Purified raplGAP was phosphorylated in vitro by CAMP-dependent kinase and the cell cycle ~34"~"' kinase. The molar ratio of incorporated phosphate/raplGAP was approximately 3 by CAMP-dependent kinase and 2 by ~ 3 4 ~ " ' . The sites of phosphorylation by both kinases were localized to a 100-residue segment contained in the carboxyl-termi-nal region of the predicted primary structure of raplGAP. Highly favorable recognition sequences for the two kinases are contained within this fragment and are proposed as the sites of phosphorylation. Treat- ment of SK-MEL-3 cells with dibutyryl CAMP pro- moted phosphorylation of raplGAP in vivo. Based on the results of comparative phosphopeptide mapping the sites of phosphorylation in vivo and in vitro are iden- tical.

* This work was partly funded by Hoffmann-La Roche and National Cancer Institute Grant CA51992-02 (to F. M.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisentent:' in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The complete lack of structural identity between GAPS specific for the different GTP-binding proteins suggests they function in disparate signaling systems in the cell. The identification of proteins, enzymes, or second messengers that associate with or modify GAPS will help to assemble the signaling networks in which the various GTP-binding proteins operate. For example, the finding that rasGAP is phosphorylated by growth factor receptors (8,9) supports the proposal that p21' " itself is coupled to growth factor receptor signaling (10). It is conceivable that raplGAP is also coupled to cell signaling events which have already been extensively characterized. As a first step toward understanding the signaling systems affecting raplGAP we have investigated the nature of a previously reported post-translational modification of the protein (6). This modification, manifested as a multiple banding pattern on SDS-polyacrylamide gels, was detected with both the recombinantly expressed protein and lysates from mammalian tissues and cell lines (6).
Since phosphorylation is known to affect the mobility of proteins on SDS-gels, and the raplGAP amino acid sequence contains strong consensus sites for phosphorylation, we sought to determine whether raplGAP was a phosphoprotein. Here we report that raplGAP is substantially phosphorylated on serine residues in uiuo and that this phosphorylation is enhanced by reagents that activate CAMP-dependent kinase. We also demonstrate that both CAMP-dependent kinase and the cell cycle ~34'~'' kinase phosphorylate raplGAP quantitatively in vitro at sites proximal to each other in the polypeptide chain.
Expression and Purification of Recombinant raplGAP-Insect Sf9 cells were infected with recombinant baculovirus carrying the HuB10-A raplGAP cDNA as previously described (6). Approximately 100 ml of cell suspension was centrifuged and the pelleted cells lysed by freeze-thawing in 10 ml of 20 mM Tris, pH 8.0, 1 mM dithiothreitol, 1 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, leupeptin, and pepstatin at 1 pg/ml. Following centrifugation at 100,000 X g for 1 h the supernatant was adjusted to pH 6.5 and chromatographed on S-Sepharose (Pharmacia LKB Biotechnology Inc.) equilibrated in 50 mM sodium phosphate, pH 6.5, containing 1 mM dithiothreitol, 1 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, leupeptin, and pepstatin at 1 pg/ml. The column was eluted with a 100-ml gradient of 0-300 mM NaCl, at a flow rate of 20 ml/h. Peak fractions containing raplGAP activity were chromatographed on a 500 ml S-300 Sephacryl column (Pharmacia LKB) equilibrated in 25 mM Tris, pH 8.0, containing 1 mM dithiothreitol, 1 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, leupeptin, and pepstatin at 1 pg/ml. Peak fractions were pooled and concentrated to approximately 0.5 mg/ml total protein. These preparations were judged to be greater than 95% pure raplGAP by SDS-PAGE analysis. For the RG12 raplGAP mutant the protein was purified directly from the Sf9 cell lysate by immunoaffinity chromatogaphy using an antibody specific to a 9-mer artificial epitope expressed at the carboxyl terminus of the protein as described previously (6). This preparation of RG12 was greater than 95% pure and exhibited specific GAP activity equivalent to the fulllength raplGAP.
Metabolic Labeling of Cells-Sixty-mm dishes of Sf9 cells expressing raplGAP cDNA were starved for either phosphate or methionine for 60 min a t 23 h post-infection and then incubated with either 1 mCi of "Pi or 0.5 mCi of ["'SJmethionine for 2 h at 25 "C. Cells were lysed in 0.5 ml/dish 1% Nonidet P-40 in 25 mM Tris, pH 8.0, containing 1 mM dithiothreitol, 1 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, leupeptin, and pepstatin at 1 pg/ml. For labeling SK-MEL-3 cells with "Pi, cells were labeled at confluency in 60-mm dishes essentially as described above for Sf9 cells. For treatment with dibutyryl CAMP, cells were incubated for 2 h with 2 mM dibutyryl CAMP beginning at the time of addition of the "Pi.
Phosphorylation of raplGAP in Vitro-Kinase reactions were carried out in a 100-pl final volume containing 25 mM Tris, pH 7.5, 10 mM MgC12, 1 mM dithiothreitol, 50 mM [y-3ZP]ATP (10,000-20.000 cpm/pmol), and 5 pg of purified raplGAP. Reactions were started by the addition of either 10 ng of CAMP-PK or 2 ng of purified ~34'~'' kinase. For estimating the level of phosphorylation, aliquots of 5 p1 were removed from the reaction and diluted into 4 ml of Tris, pH 7.5, and immediately filtered through nitrocellulose filter disks at the indicated times. Radioactivity was determined by scintillation counting of the dried filters. The molar amounts of incorporated phosphate were estimated from the specific activity of the [y-"PIATP determined by scintillation counting of a known volume of [y-'>P]ATP. All quantitations presented were determined from the means of duplicate filtrations. When phosphorylated raplGAP was to be applied to SDS-polyacrylamide gels the reactions were stopped by addition of SDS-gel sample buffer.
Phosphoamino Acid Analysis-Phosphorylated raplGAP was immunoprecipitated from lysates of 3ZPi-labeled Sf9 cells overexpressing raplGAP. Immunoprecipitates were resolved by SDS-polyacrylamide gel electrophoresis and then electroblotted to a PVDF filter membrane. The radiolabeled raplGAP was excised from the blot and then hydrolyzed by vapor phase contact with 6 N HCI a t 110 "C for 2 h essentially as described by Hildebrandt and Fried (11). The amino acids were eluted from the membrane with distilled water and standard phosphoamino acids (5 nmol each) were added to the sample prior to chromatography. Phosphoamino acids were resolved by thin layer chromatography using the ammonia/ethanol solvent system described by Munoz and Marshall (12). Standard amino acids were visualized by staining with ninhydrin and "Pi-labeled residues were detected by autoradiography.
Electrophoretic and Immunologic Procedures-For the separation and Western blotting of intact raplGAP, SDS-polyacrylamide gel electrophoresis was performed in 8% gels essentially as described by Laemmli (13). Digestions of raplGAP were resolved on 16% SDSpolyacrylamide gels using the Tricine gel system described by Chagger and VonJagow (14). Western blotting for detection of raplGAP with antisera was performed using nitrocellulose filters and 20 mM Tris, 192 mM glycine, 20% methanol. Electroblotting for amino acid sequencing of resolved peptides was performed using PVDF filters and 10 mM CAPS buffer, pH 11.0, containing 10% methanol as described by Matsudaira (15). Antibodies specific to raplGAP were raised in rabbits against the purified recombinant protein as previously described (6). The sera was affinity-purified against immobilized purified raplGAP and diluted to a stock concentration of 0.25 mg/ml. For Western blotting, the purified antibody was used a t a dilution of 1/10,000 and for immunoprecipitations, 5 pl were added to 500 pl of cell lysate and incubated a t 4 "C overnight. Immunocomplexes were recovered using protein A-Sepharose. Western blots were developed using goat anti-rabbit IgG coupled to horseradish peroxidase (Bio-Rad, Richmond, CA) and visualized using the ECL system (Amersham, UK).
Proteolvtic and Chemical Dieestion of radGAP-For the dieestion of r a p l G i P by CNBr, 25 pg ofihe puriked'raplGAP was phosphorylated in uitro and then precipitated on ice with trichloroacetic acid (20% final volume). The protein was redissolved in 70% formic acid containing 30 mg/ml CNBr and digested overnight a t room temperature. The sample was evaporated, redissolved in SDS-gel sample buffer, and applied to the gel. For enzymatic digestions, the procedure described by Cleveland (16) was followed. Five pg of the phosphorylated raplGAP was boiled for 2 min in the presence of 100 mM Tris, pH 6.8, 0.5% SDS, 10 mM dithiothreitol, 10% glycerol, and the indicated proteases were added to a final concentration of 25 pg/ml. Digestions were carried overnight a t room temperature, and the samples were boiled in SDS-gel sample buffer prior to electrophoresis.
Other Procedures-Protein concentration of raplGAP was determined by amino acid analysis on a Beckman System 6300 high performance analyzer. Amino acid sequencing was performed on an Applied Biosystems model 470A gas phase automatic sequenator.

RESULTS
To determine whether raplGAP was phosphorylated in vivo we expressed raplGAP cDNA in insect Sf9 cells using the . We next carried out phosphoamino acid analysis of the 32P-labeled raplGAP. Following immunoprecipitation of raplGAP from the Sf9 cell lysate, the protein was electroblotted from an SDS-gel and the immobilized protein was hydrolyzed with HC1. A chromatogram of the hydrolysate revealed that most radioactivity comigrated with the phosphoserine standard, while a lower level of phosphothreonine was also detected (Fig. 1B). A trace amount of radioactivity was also detected near the top of the chromatogram but migrated slightly faster than the phosphotyrosine standard.
The deduced amino acid sequence of raplGAP contains some of the most commonly identified recognition sequences for phosphorylation by CAMP-dependent protein kinase (reviewed in Refs. 17 and 18). To test whether raplGAP was a substrate for this kinase we incubated purified recombinant raplGAP with the catalytic subunit of CAMP-dependent kinase (CAMP-PK) in the presence of [r-"P]ATP. raplGAP was rapidly phosphorylated and maximally incorporated approximately 3 mol of phosphate/mol of protein (Fig. 2). The incorporation of phosphate was also reflected in a shift in mobility of raplGAP on SDS-polyacrylamide gels (Fig. 2,  inset). To localize the site of phoshorylation by CAMP-PK the radiolabeled protein was digested with CNBr and the fragments resolved by gel electrophoresis. A single 13.5-kDa digestion product contained the majority of the radioactivity (Fig. 3A). Amino acid sequencing of this peptide yielded the sequence shown underlined in Fig. 3B. This sequence immediately follows methionine residue 445 in the raplGAP sequence and the next methionine COOH-terminal to this is found at position 545. The predicted molecular mass of this peptide is comparable to that determined for the CNBr fragment on SDS-polyacrylamide gels. Consistent with the phosphorylation of this region by CAMP-PK is the identification of several consensus recognition motifs for this kinase. In particular, the most commonly identified motif, R-R/K-X-S/ T, is represented twice and the R-X-S/T motif once within the phosphorylated raplGAP fragment.
The raplGAP sequence also contains consensus sites for the cell cycle ~3 4~~'~ kinase (17,18). When incubated with purified ~3 4~~' ' , raplGAP again rapidly incorporated phosphate. The maximal incorporation reached a plateau at approximately 2 mol of phosphate/mol of raplGAP (Fig. 4).  Phosphorylation was carried out in vitro as described under "Experimental Procedures." ~34'~'' kinase was added at time zero, and phosphorylation was assayed at the indicated times. A&r 90 min the sample was divided and incubated further with either CAMP-PK or a second addition of ~3 4~'~ kinase.

200
-ND T f P C z G* C C E 4 5 .  4). This demonstrates that the sites of phosphorylation for these two kinases are distinct. To compare the sites of phosphorylation we phosphorylated raplGAP using ~3 4~'~ or CAMP-PK and digested the radiolabeled proteins with CNBr. The protein phosphorylated by p34'd'2 generated a 13.5-kDa radiolabeled fragment identical in mobility to that produced from the raplGAP phosphorylated by CAMP-PK (Fig. 5). Thus, the sites of phosphorylation by ~3 4 '~~' are also within the carboxyl-terminal region located between methionine residues 445 and 545. Within this region there are several S/T residues immediately followed by prolines; a requirement for phosphorylation by ~34"~'' (17,18). We further compared the sites of phosphorylation by carrying out limited enzymatic digestions of raplGAP phosphorylated by the two kinases. This treatment resulted in the generation of a set of overlapping radiolabeled peptides that were resolved by SDS-polyacrylamide gel electrophoresis (Fig. 5). Nearly identical peptide maps were generated from the raplGAP phosphorylated by ~3 4 '~~~ or CAMP-PK. Again, this suggests that although the sites of phosphorylation by these two kinases are distinct they are likely very near each other in the polypeptide chain. The only detectable difference in the comparative peptide maps was an approximate 4-kDa peptide produced with chymotryptic digestion of the CAMP-PK-labeled raplGAP but not with the protein labeled by ~3 4~" (Fig. 5). However, phosphorylation of this residue may reflect secondary, artifactual labeling in the in vitro system as this labeled 4-kDa fragment was not detected in chymotryptic digests of .. To confirm that the major sites of phosphorylation were localized to the carboxyl-terminal region of raplGAP, a deletion mutant lacking the carboxyl-terminal one-third of the protein was generated. This mutant, designated RG12, was produced in the Sf9 cell, purified to near homogeneity, and found to stimulate the GTPase activity of p2lTaP' with a specific activity nearly identical to that of full-length raplGAP.' Significant phosphorylation of the RG12 mutant could not be achieved in vitro with either CAMP-PK or ~34'~'' under conditions that resulted in the rapid phosphorylation of full-length raplGAP (Fig. 6). These results suggest that the domain of raplGAP which serves as a substrate for kinases is distinct from the domain responsible for GTPase activation of p2lrnp1. Moreover, the kinase substrate domain did not influence the activity of the catalytic domain because RG12 did not differ in activity from full-length raplGAP and we could not detect a change in the GAP activity of fulllength raplGAP following its phosphorylation i n vitro (data not shown).
That the phosphorylation of raplGAP altered its mobility on SDS-polyacrylamide gels suggested that phosphorylation may also be responsible for the multiple banding pattern observed for raplGAP extracted from recombinant and native sources (6). We attempted to alter the banding pattern of raplGAP by incubating intact cells with dibutyryl CAMP, a reagent known to activate CAMP-PK. The human melanoma cell line, SK-MEL-3, contains relatively high levels of raplGAP mRNA (6) and was therefore used for this analysis. Lysates from the SK-MEL-3 cell contained three resolvable forms of raplGAP as detected by Western blotting (Fig. 7A). When the cells were incubated with dibutyryl CAMP, the fastest migrating form of the raplGAP was diminished and the second band became more intense relative to the control (Fig. 7A). Essentially identical results were obtained using isobutylmethylxanthine, an inhibitor of phosphodiesterase which promotes accumulation of cAMP in the cell. SK-MEL-3 cells were also metabolically labeled with 32Pi, incubated with or without dibutyryl CAMP, and the raplGAP immunoprecipitated from the cell lysates. These results show that both the stimulated and unstimulated SK-MEL-3 cells contain substantial levels of phosphorylated raplGAP (Fig. 7B).  (-) or incubated with 2 mM dibutyryl cAMP for 2 h (+) and lysates subjected to SDS-PAGE and immunoblotting with raplGAP-specific antisera. B, raplGAP was immunoprecipitated from lysates of SK-MEL-3 cells that were metabolically labeled with "Pi and untreated (-) or incubated with dibutyryl cAMP (+). The raplGAP phosphorylated in vivo or in vitro with the indicated kinases was subjected to SDS-PAGE before (first 5 lams) or after digestion with chymotrypsin (last 5 lanes). An autoradiogram is presented. Values a t right are molecular weights of standard proteins X lo-'.
ations in the recovery of the immunoprecipitates, a dibutyryl CAMP-induced increase in phosphorylation was not readily observed by this analysis. However, on Western blotting of the immunoprecipitates a band shift identical to that seen in Fig. 7A was observed indicating a qualitative response to dibutyryl cAMP (data not shown). Nevertheless, the phosphoproteins, radiolabeled either in vivo or in vitro, generated the same set of polypeptides on digestion with chymotrypsin (Fig. 7B). These results demonstrate that raplGAP is phosphorylated in vivo at sites that are proximal or identical to those phosphorylated in vitro by CAMP-PK and ~3 4~~~~. One exception is the 4-kDa phosphopeptide obtained from digestion of raplGAP phosphorylated by CAMP-PK in uitro but not in vivo. This is the case even for cells stimulated with dibutyryl CAMP. As noted above, this suggests that the site of phosphorylation contained in this 4-kDa peptide may not be relevant to phosphorylation as it occurs in the cell. In this experiment we also digested raplGAP that was phosphorylated both by CAMP-PK and ~34"~". No new digestion products were noted in comparison to those produced from raplGAP phosphorylated by either kinase alone (Fig. 7B).

DISCUSSION
When expressed in recombinant systems raplGAP undergoes a post-translational modification that affects its mobility on SDS-polyacrylamide gels (6). Multiple forms of raplGAP, differing in apparent molecular mass, were also consistently identified in lysates from various mammalian tissues and cell lines (6). Here we have shown that these multiple forms of raplGAP are at least in part due to phosphorylation of the protein on serine residues. Although the exact stoichiometry of phosphorylation in vivo was not determined, the relative intensity of the multiple forms of raplGAP detected on Western blots suggests that a substantial percentage of the protein exists in the phosphorylated state in the cell. The multiple banding pattern is apparently indicative of phosphorylation, because the relative intensities of the bands were altered following treatment of the cells with an activator of CAMP-PK. Moreover, this shift in mobility of raplGAP on SDS-gels was reproduced in vitro by phosphorylation of the purified protein with CAMP-PK.
From the in uitro phosphorylation data it is clear that raplGAP incorporated several stoichiometric equivalents of phosphate. The molar ratio of phosphate to raplGAP ranged from 2 to 3 using CAMP-PK and was as high as 5 in some experiments where both CAMP-PK and ~3 4 '~' were employed. The maximum level to which raplGAP could be phosphorylated varied to some degree across preparations and was likely dependent upon the relative level of phosphorylation pre-existing in the purified raplGAP. During the purification of the recombinant protein used in these studies we biased the yields for the lowest molecular mass forms of raplGAP in an attempt to enrich for the nonphosphorylated forms. However, slower migrating forms, presumably phosphorylated, were still evident in the final preparations. Moreover, the shift in mobility observed on SDS-gels following phosphorylation of raplGAP in vitro did not occur until an approximate molar ratio of 2 was achieved. Therefore, even the fastest migrating forms of raplGAP may nevertheless be substantially phosphorylated. In any event, it is clear that raplGAP undergoes hyperphosphorylation in vitro and based on comparative phosphopeptide mapping this is also likely the case for phosphorylation in vivo.
The carboxyl-terminal region of raplGAP contains the sites of phosphorylation by CAMP-PK and ~3 4~' ' .
These phosphorylation sites were localized to an approximate 13.5-kDa fragment generated by digestion with CNBr. This segment of the raplGAP primary structure, bordered by methionine residues 445 and 545, contains 24 threonine and serine residues. From our data it cannot be determined which of these residues are phosphorylated by CAMP-PK. However, 3 of these serines are situated with basic residues at the -2 and/or -3 position with respect to the putative phosphate acceptor: a requirement for recognition by CAMP-PK (17,18). Moreover, two of these, Ser4" and Ser4", comply with the most typical motif for recognition by CAMP-PK (17,18). Phosphorylation in vitro by ~34'~'' also occurs within this same serine/threoninerich segment of raplGAP. A primary recognition sequence for ~34'~"' has not been well defined, but in all cases there is a requirement for a proline residue immediately carboxyl-terminal to the Ser/Thr phosphate acceptor (17,18). Most of the reported sites for ~34'~'', such as those identified in Tantigen (19), p60""" (20), the rab 1A and 4 proteins (21), nucleolin (22), and nuclear lamin proteins (23) also contain basic residues 1 residue carboxyl-terminal to the essential proline. The phosphorylated CNBr fragment of raplGAP contains 4 Ser/Thr-Pro pairs, and one of these, Ser483, is followed by a basic residue in this context. In fact, the raplGAP S4=PTR motif is identical to that conserved in the amino-terminal tail of all lamin proteins (23) and a 17-mer peptide containing this SPTR motif inhibits phosphorylation of lamin and H1 histone by ~34'~'' (23).
The proposed sites of phosphorylation are also consistent with the peptide mapping results. When digested with three different proteases, nearly identical sets of phosphopeptides were generated from raplGAP phosphorylated by either p3PdC2 or CAMP-PK. The observation that most of the peptides produced contained both a ~34"~"' site and a CAMP-PK site suggests that extensive primary structure does not lie between the phospho-acceptor sites for the two kinases. In addition, the inability to phosphorylate the carboxyl-terminal truncation mutant, RG12, indicates that the sites of phosphorylation are clustered within the same general region of raplGAP primary structure.
The significance of phosphorylation of raplGAP by either CAMP-PK or ~34'~"' is not yet understood. A significant enhancement of raplGAP activity toward p2lraP' was not observed when raplGAP was phosphorylated in vitro. However, it is more difficult to appreciate how phosphorylation might affect raplGAP activity in vivo. Phosphorylation of raplGAP in vivo may influence its interaction with other proteins or second messengers, which in turn could have a profound affect on its activity toward p21"P' . Alternatively, the association of raplGAP with a particular subcellular component may be influenced by its state of phosphorylation. This has been proposed for a number of regulatory proteins including p2lraP' itself (24). raplGAP is detected both in membrane and cytosolic fractions prepared from a variety of tissue and cell lysates (6) and may therefore shuttle between these compartments in response to phosphorylation. Whatever the consequences of phosphorylation it is intriguing that raplGAP contains closely juxtaposed sites for CAMP-PK and ~34'~'' kinase and that these two kinases typically have opposing effects on cell growth (24). However, a recent study indicated that CAMP-PK and ~34'~'' may even associate with each other suggesting possible interdependent activities (25). It is clear from this study that raplGAP is significantly phosphorylated on serine residues both in vivo and in vitro.
The major sites of phosphorylation are contained between methionine residues 445 and 545 in the complete 663-amino acid primary structure of raplGAP. Interestingly, this region of raplGAP also contains a duplicated amino acid sequence resulting from an alternative mRNA splicing event (6). Two distinct cDNAs coding for raplGAP were found to differ by the inclusion of an additional 78 base pairs inserted in the 3' coding region of one of the clones (6). These additional base pairs code for a 26-amino acid sequence highly homologous to the sequence immediately following it and contained between residues 478-505 in the raplGAP used in the present study (see Fig. 2). The repeated region present in the splicing variant encodes an additional putative CAMP-PK site, RRGS, but the ~3 4 '~'~ consensus site is not duplicated. Recombinant proteins expressed from either cDNA exhibit approximately equal specific activities towardp2lrnp' (6). This again suggests that the carboxyl-terminal portion of raplGAP is not directly involved in the stimulation of p2lrnp' but rather may serve as a regulatory domain interfacing raplGAP to the activation of serine kinases in the cell. The cellular function of p21rap' is not understood. The protein has been shown to bind with high affinity to rasGAP (26, 27), and it has been suggested that this competition may account for its ability to revert the ras-transformed phenotype in 3T3 cells (28). The finding that raplGAP is abundant in some myeloid cell types and is diminished following their differentiation suggests that the p2lrnp1-raplGAP system may participate in the control of cell growth (6). This is consistent with our results showing that raplGAP is a good substrate for the cell cycle ~3 4 '~"~ kinase. The phosphorylation of raplGAP by CAMP-PK is intriguing considering that ~2 1 "~' itself is also a good substrate for this kinase. It would make sense that interacting proteins such as p2lTnP' and raplGAP respond to the same second messenger systems. The identification of raplGAP as a substrate for serinebhreonine kinases and the identification of those kinases acting on it in vivo will help bring together the various components functioning in this signaling network.