GTK, a Src-related tyrosine kinase, induces nerve growth factor-independent neurite outgrowth in PC12 cells through activation of the Rap1 pathway. Relationship to Shb tyrosine phosphorylation and elevated levels of focal adhesion kinase.

The rat pheochromocytoma cell line PC12 is extensively used as a model for studies of neuronal cell differentiation. These cells develop a sympathetic neuron-like phenotype when cultured in the presence of nerve growth factor. The present study was performed in order to assess the role of mouse GTK (previously named BSK/IYK), a cytoplasmic tyrosine kinase belonging to the Src family, for neurite outgrowth in PC12 cells. We report that PC12 cells stably overexpressing GTK exhibit a larger fraction of cells with neurites as compared with control cells, and this response is not accompanied by an increased ERK activity. Treatment of the cells with the MEK inhibitor PD98059 did not reduce the GTK-dependent increased in neurite outgrowth. GTK expression induces a nerve growth factor-independent Rap1 activation, probably through altered CrkII signaling. We observe increased CrkII complex formation with p130(Cas), focal adhesion kinase (FAK), and Shb in PC12-GTK cells. The expression of GTK also correlates with a markedly increased content of FAK, phosphorylation of the adaptor protein Shb, and an association between these two proteins. Transient transfection of GTK-overexpressing cells with RalGDS-RBD or Rap1GAP, inhibitors of the Rap1 pathway, reduces the GTK-dependent neurite outgrowth. These data suggest that GTK participates in a signaling pathway, perhaps involving Shb, FAK and Rap1, that induces neurite outgrowth in PC12 cells.

The rat pheochromocytoma tumor cell line PC12 is commonly used to study the signaling pathways involved in neuronal cell differentiation. These cells mature into sympatheticlike neurons upon the addition of nerve growth factor (NGF) 1 (1,2). In serum-free conditions and in the absence of NGF, PC12 cells undergo programmed cell death (3). NGF induces cell cycle arrest and differentiation by binding and activating the TrkA receptor tyrosine kinase. This causes tyrosine phosphorylation of Shc (4) and fibroblast growth factor receptor substrate-2 (FRS-2), which stimulate the extracellular signalregulated kinase (ERK) pathway, phospholipase C-␥ (4), which stimulates protein kinase C and yields an increase in intracellular calcium, and phosphatidylinositol 3-kinase, which activates Akt to mediate neuronal survival (for a recent review see Ref. 5). The Ras/ERK cascade has been demonstrated to be both necessary and sufficient for NGF-induced differentiation of PC12 cells (6). Activation of ERK by growth factors can trigger either cell growth or differentiation, and a transient activation of ERK is thought to stimulate proliferation, whereas a sustained activation induces differentiation (7). Although the Ras pathway is considered to be of major importance for NGFinduced ERK activation, a recent study suggested that Ras is responsible for the initial activation of ERK, whereas the sustained activation is mediated by Rap1 (8). Rap1 (9) is a Ras family member that shares with Ras many downstream effectors, including Raf1 and RalGEFs. The involvement of Rap1 in NGF-induced ERK activation is controversial since it was recently shown that NGF fails to activate Rap1 in PC12 cells (10).
Mouse GTK (previously named BSK/IYK) (11,12) is a nonreceptor protein-tyrosine kinase belonging to the Src family of protein-tyrosine kinases. Mouse GTK is highly homologous to human FRK/RAK (13,14) and rat GTK (15). Little is known about the function of these tyrosine kinases, and they have been suggested to be members of a subgroup within the Src family with certain specific characteristics compared with the other Src family members.
We have previously investigated the importance of two tyrosine residues within the tail of GTK, namely Tyr-497 and Tyr-504, for NIH3T3 (16) and RINm5F cell proliferation (17). It was shown that expression of a kinase-active mutant that could enter the nucleus (GTK Y497F/Y504F in NIH3T3 cells and GTK Y504F and GTK Y497F/Y504F in RINm5F cells) reduced the cell proliferation rate. Furthermore, we observed that the Y504F-and Y497F/Y504F-GTK mutants increase the mRNA levels of glucagon in the RINm5F cells. These findings raise the possibility that GTK might be involved in differentiation and maturation of cells. To investigate this hypothesis we have transfected PC12 cells with GTK and studied the effects on neurite outgrowth in the absence and presence of NGF. We observe that GTK overexpression induces NGF-independent neurite outgrowth and Rap1 activation, probably through activation of the CrkII-C3G pathway. This could be the consequence of increased levels of focal adhesion kinase (FAK) and phosphorylation of the Shb adaptor protein.
Neurite Outgrowth-PC12 cells were cultured for 72 h in medium with a full serum supplement with or without the presence of 20 ng/ml NGF and/or PD98059 (a final concentration of 20 M, added 10 min prior to NGF). The percentage of cells with neurites extending two diameters of the cell body was counted every 24 h.
Western Blot Analysis on Lysates-Subconfluent cells were grown overnight in medium containing 2% FCS and 1% HS and stimulated with 100 ng/ml NGF for the indicated time points. Cells were then washed with cold phosphate-buffered saline (PBS), briefly sonicated in SDS sample buffer (containing ␤-mercaptoethanol and 2 mM PMSF), and subjected to Western blot analysis. The membranes were incubated with the indicated antibodies, and the immunoreactivity was subsequently detected by ECL.
In Vitro Kinase Assay-PC12 cells cultured to subconfluence in 10-cm dishes and maintained in 2% FCS, 1% HS overnight, were stimulated for 10 min with 100 ng/ml NGF and washed with cold PBS containing 100 M orthovanadate. Cells were lysed in 150 mM NaCl, 30 mM Tris, pH 7.5, 10 mM EDTA, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 100 M orthovanadate, 2 mM PMSF, 100 units/ml Trasylol (aprotinin), and 0.05 mM leupeptin, and the nuclei were removed by centrifugation. GTK was precipitated with GTK antibody and protein A-Sepharose, and the beads were thoroughly washed and subsequently subjected to an in vitro kinase reaction by incubation in the presence of 40 mM Hepes, pH 7.5, 10 mM MgCl 2 , 3 mM MnCl 2 , 10% glycerol, 1 mM dithiothreitol, and 7 Ci of [␥-32 P]ATP for 15 min. The reaction was stopped by addition of SDS-sample buffer. All solutions were supplemented with 100 M sodium orthovanadate. The proteins were separated by SDS-PAGE, and Western transfer was performed. The filters were either directly exposed or blotted with phosphotyrosine (4G10) antibody.
Crk-SH2 Binding Assay-The construction, expression, and purification of GST-CrkSH2 fusion protein have been described previously (20,21). PC12 cells cultured to subconfluence in 10-cm dishes were serum-starved overnight and stimulated with NGF (100 ng/ml) for 10 min and thoroughly washed with cold PBS containing 100 M orthovanadate and lysed in 300 l of lysis buffer (150 mM NaCl, 0.5% Triton X-100, 50 mM Tris, pH 7.5, 1 mM EDTA, 300 M orthovanadate, 100 units/ml aprotinin, 2 mM PMSF, 1 mM DTT, 0.05 mM leupeptin, 20 mM calpain inhibitor N-acetyl-Leu-Leu-norleucinal). Lysates were clarified at 13,000 ϫ g for 10 min at 4°C, and the supernatants were incubated on ice with GST-CrkSH2 fusion protein immobilized to glutathione-Sepharose beads for 1 h. The beads were washed with PBS ϩ 1% Triton X-100, and 50 l of SDS sample buffer was added. The proteins were separated by SDS-PAGE followed by Western blot analysis for Shb and FAK.
Rap1/Ras Activity Assay-PC12 cells cultured to subconfluence in 10-cm dishes were serum-starved overnight and stimulated with NGF (100 ng/ml) for 10 min and thoroughly washed with cold PBS containing 100 M orthovanadate and lysed in 300 l of lysis buffer (1% Nonidet P-40, 50 mM Tris, pH 7.5, 20 mM MgCl 2 , 200 mM NaCl, 10% glycerol, 100 M orthovanadate, 100 units/ml aprotinin, 2 mM PMSF, 0.05 mM leupeptin). Lysates were clarified at 13,000 ϫ g for 10 min at 4°C, and the supernatants were incubated on ice with GST-RalGDS-RBD fusion protein (see Ref. 22 for details) immobilized to glutathione-Sepharose beads for 1 h. The beads were washed with PBS ϩ 1% Triton X-100, and 50 l of SDS sample buffer was added. The proteins were separated by SDS-PAGE followed by Western blot analysis for Rap1 or Ras.
Transient Transfection-The pMT2HA-RalGDS-RBD construct was generated by polymerase chain reaction amplification of RalGDS encoding amino acids 200 -297 as a SalI/NotI fragment and subsequently subcloned into SalI/NotI-digested pMT2HA expression vector. The pMT2HA-Rap1GAP was generated by subcloning full-length Rap1GAP into the pMT2HA expression vector (according to Ref. 23). Parental PC12 and PC12-GTK cells cultured in 2-cm dishes were transfected with 0.3 g of pIRES-EGFP and 2 g of RalGDS-RBD or Rap1GAP construct using 1.8% LipofectAMINE for 3 h at 37°C in serum-free medium. As control, cells were transfected with 0.3 g of pIRES-EGFPvector and 2 g of carrier DNA. After 24 h, cells were left unstimulated or treated with 50 ng/ml NGF for 48 or 24 h, and the GFP-positive cells with neurites extending two diameters of the cell body were counted in a Zeiss fluorescence microscope.

RESULTS
Neurite Outgrowth of PC12 Cells Expressing GTK-To elucidate if GTK affects the differentiation of PC12 cells, we established PC12 cell lines stably overexpressing the wild-type GTK cDNA. Two clones (PC12 GTK-7 and GTK-10) were found to have increased intracellular levels of a 55-kDa GTK protein, compared with the control cells as assessed by Western blot analysis (Fig. 1A). A significant portion of the GTK-expressing PC12 cells displayed a flattened phenotype and extended neurites when cultured in the absence of NGF. As seen in Fig. 1B, the GTK-10 cells that express higher levels of GTK than the GTK-7 cells also show the more altered morphology, with a higher fraction of cells with neurites after 72 h in culture as well as more somal flattening compared with the GTK-7 cells.
To investigate if NGF could induce further differentiation of these cells, a time course experiment in response to NGF was performed. The cells were cultured for 24 h before treatment with 20 ng/ml NGF for another 72 h. The GTK-expressing cells responded to NGF, and the fraction of cells with neurites was significantly higher at all time points compared with the control cells (PC12 or PC12-neo, Fig. 2). The NGF-induced increment in neurite outgrowth compared with the outgrowth before NGF treatment was similar in the GTK and control cells.
GTK Kinase Activity Is Not Affected by NGF-To assess the activity of GTK in PC12 cells we performed an in vitro kinase assay using PC12-GTK10 and parental PC12 cells that were treated with NGF (100 ng/ml) for 10 min or left unstimulated. The amount of GTK present in the lysates and the total amount of tyrosine phosphorylation were determined by Western blot analysis using GTK or phosphotyrosine antibody. 32 P incorporation into a 55-kDa band was specifically detected in the immunoprecipitates from the GTK-overexpressing cells, and NGF had little impact on GTK autophosphorylation (Fig. 3). This 55-kDa band was tyrosine-phosphorylated as demonstrated by Western blot analysis with the specific phosphotyrosine antibody, 4G10. Western blot analysis of PC12-GTK10 lysate for phosphotyrosine also showed a strong 57-kDa band just above the position of GTK. Another protein of 125 kDa was also tyrosine-phosphorylated to a larger extent in the cell lysates from PC12-GTK cells compared with the control cells. Due to the proximity of GTK to IgG it was not possible to assess the GTK levels in the immunoprecipitates, but the amount GTK present in the lysate was similar in stimulated and unstimulated PC12-GTK cells. Wild-type PC12 cells express low but detectable levels of GTK.
NGF-independent Neurite Outgrowth in GTK-expressing Cells Is Not Caused by the MAPK Pathway-It is well established that the MAPK pathway is important for differentiation of PC12 cells, and therefore we wanted to assess the activation of p42 and p44 ERK in the GTK-expressing cells. Phosphorylation of p42/p44 ERK was assessed after stimulation with 100 ng/ml NGF for 0, 2, 10, and 60 min and related to their total amounts by Western blot analysis and densitometric scannings. Addition of NGF stimulates ERK phosphorylation in both GTK-overexpressing and control cells (Fig. 4A), without an elevation of the basal phosphorylation of ERK in PC12-GTK cells. Moreover, the NGF-induced activation is even somewhat delayed in these cells, with significantly decreased phosphorylation of both p42 and p44 ERK at 2 min, suggesting that the MAPK pathway is not responsible for the basal neurite outgrowth induced by GTK.
To exclude the possibility of a minor increase in ERK activation in PC12-GTK cells, we investigated if the MEK inhibitor PD98059 could influence neurite outgrowth of PC12-GTK cells. Cells were cultured as described above but in the presence or absence of PD98059 (20 M), which was added 10 min prior to the addition of NGF (20 ng/ml), and cells with neurites were counted (Fig. 4B). PD98059 inhibited the effects on neurite outgrowth induced by NGF but could not lower the fraction of PC12-GTK cells with neurites below that of the basal state. Thus, no correlation between ERK activation and neurite outgrowth in response to GTK overexpression can be found, suggesting that GTK transmits a MAPK-independent differentiation signal in PC12 cells.
GTK-expressing Cells Exhibit Elevated TrkA Phosphorylation-NGF-induced differentiation of PC12 cells is mediated by the TrkA receptor that becomes phosphorylated after NGF binding. We thus wanted to see if TrkA phosphorylation was elevated in the PC12-GTK cells. Cells were left unstimulated or treated with NGF (100 ng/ml) for 10 min followed by TrkA immunoprecipitation using a specific TrkA antibody. Western blot analysis for phosphotyrosine was performed, and the amount of phosphorylated TrkA was compared with the total amount present in lysate and immunoprecipitate. We observe a TrkA-specific band in both lysates and immunoprecipitates corresponding to 120 -145 kDa. In addition we detect two bands of 90 -100 kDa displaying TrkA immunoreactivity that showed no signs of tyrosine phosphorylation. NGF stimulated TrkA phosphorylation in the control cells. The tyrosine phos- phorylation of TrkA following NGF treatment was similar in the control and GTK-overexpressing cells, whereas the unstimulated cells from both GTK-expressing clones exhibited more tyrosine phosphorylation of TrkA compared with the control cells (Fig. 5A). Furthermore, the total amount of TrkA was lower in the PC12-GTK cells compared with the control, suggesting an attempt to down-regulate this receptor as a consequence of its constitutive activation.
When TrkA binds NGF, the tyrosine at position 490 becomes phosphorylated, and this generates a binding site for the adaptor protein Shc. We therefore performed Western blot analysis of equal amounts of crude lysate from cells stimulated with NGF for 10 or 60 min using an antiserum specifically recognizing the Tyr-490-phosphorylated form of TrkA. There is a sustained increase of Tyr-490 phosphorylation after NGF stimulation in the control cells. The GTK-expressing cells, however, show an elevated basal phosphorylation of Tyr-490 that cannot be increased further with NGF indicating that Tyr-490 is a phosphorylation site involved in GTK-mediated activation.
The Adaptor Protein Shb Is Phosphorylated in GTK-expressing Cell-We have previously shown that PC12 cells overexpressing the adaptor protein Shb exhibit enhanced NGF-induced differentiation, assessed as neurite outgrowth (24). It was found that NGF induced phosphorylation of p57 Shb in the overexpressing cells, and therefore it was suggested that Shb could be involved in the transmission of NGF-dependent differentiation signals in PC12 cells. To determine if GTK could influence Shb phosphorylation, we immunoprecipitated Shb from extracts of cells incubated in the absence or presence of NGF, and we examined its degree of tyrosine phosphorylation. High amounts of phosphorylated Shb were observed in the GTK-expressing cells (Fig. 6), whereas control cells exhibited no detectable phosphorylation despite the presence of similar amounts of Shb in all these immunoprecipitations. NGF did not exert any apparent effect on Shb phosphorylation in this experiment, which is in line with previous experiments studying Shb phosphorylation in parental PC12 cells (24). When exposing the phosphotyrosine blot for a long time, a 125-kDa band appeared in the immunoprecipitates (result not shown). To assess if this band could be focal adhesion kinase (FAK), the blot was stripped and reprobed with antibody against FAK, demonstrating the appearance of a 125-kDa band in the GTK-10 cells but not in the control cells (Fig. 6).
PC12 GTK Cells Express Elevated Levels of FAK-FAK has been postulated to play a central role in the cellular response to the extracellular matrix and for cell morphology and motility (for review see Ref. 25). Due to the GTK-dependent FAK-Shb association we decided to study FAK tyrosine phosphorylation and expression in PC12-GTK cells. The FAK protein levels in cell extracts were much increased in both GTK-7 and GTK-10 PC12 cells (Fig. 7, A and B) compared with parental PC12 cells.

Mock-transfected PC12-neo cells express similar levels of FAK as parental cells indicating that the elevated FAK levels in PC12-GTK cells are not caused by clonal selection (result not shown). Immunoprecipitation with anti-FAK antibody and
Western blot analysis for phosphotyrosine was performed, and the results show that the degree of FAK phosphorylation is increased to the same extent as the overall FAK content (Fig. 7A).
The effect of FAK on the cytoskeleton may involve the binding to p130 Cas . CrkII and C3G have been shown to associate with p130 Cas (reviewed in Ref. 26), and C3G has been identified as a guanine nucleotide exchange factor for Rap1 (for review see Ref. 27). To determine if GTK is involved in regulating this pathway, cell lysates were immunoprecipitated using anti-CrkII, anti-p130 Cas , or anti-C3G antibody, and the phosphorylation and protein complex formations were analyzed (Fig. 7,  B-D). The expression of CrkII, p130 Cas , and C3G was similar in GTK-expressing cells and control cells. Immunoprecipitation of CrkII from cell extracts and subsequent Western blot analysis for phosphotyrosine, CrkII, p130 Cas , FAK, and C3G revealed an NGF-independent phosphorylation of CrkII and an association of FAK and p130 Cas to CrkII to a higher extent in the GTKexpressing cells compared with the control cells. The GTK cells exhibited augmented p130 Cas tyrosine phosphorylation (Fig.  7C), and this might explain the increase in the association between p130 Cas and CrkII. C3G phosphorylation was similar in the GTK cells compared with the control cells (Fig. 7D). To examine if the SH2 domain of CrkII mediates an association with Shb, we performed a pull-down experiment using a GST-Crk-SH2 fusion protein. The Crk-SH2 protein was incubated with lysates of PC12-GTK and control cells, and precipitated proteins were analyzed by immunoblotting (Fig. 7E). The phosphotyrosine blot revealed stronger bands of 130 -135 and 57 kDa in the unstimulated GTK-overexpressing cells compared with the control cells. One possible explanation for the observed decrease of the 135-kDa phosphotyrosine band after NGF stimulation could be that its binding sites for the CrkII- SH2 fusion protein are blocked by an NGF-dependent association of endogenous proteins that occurs in intact cells. Shb and FAK were found to bind the SH2 domain of CrkII in a GTKdependent manner as determined by immunoblotting with specific antibodies for Shb and FAK, and the Shb band was detected at the same position as the 57-kDa phosphotyrosine band. Since it is known that FAK associates with p130 Cas directly, and since we show that Shb can bind both CrkII and FAK, it is likely that FAK associates with CrkII via Cas or Shb.
Rap1 Activation Is Increased in GTK-expressing Cells-GTPbound Rap1 associates with high selectivity and specificity to RalGDS in vitro (28); therefore, we performed a Rap1 activity assay using a GST-RalGDS-RBD fusion protein. Immunoblotting of precipitated Rap1 provides a qualitative representation of Rap1-GTP binding, corresponding to quantitative changes detected using classical GDP/GTP-binding ratio techniques (22). The relative Rap1 activation, as determined by the amount of Rap1 bound to RalGDS relative to the total amount of Rap1 in the lysate, was assessed. We also measured the amount of activated Ras bound to RalGDS-RBD by blotting for Ras. As shown in Fig. 8, Ras is activated by NGF in both PC12-GTK and control cells to a similar degree. However, there is a significant NGF-independent increase in the activation of Rap1 in GTK-expressing cells compared with parental PC12 cells. We conclude that GTK overexpression induces a constitutive increase in the amount of GTP-bound endogenous Rap1.

Transient Expression of RalGDS-RBD and Rap1GAP
Reduces Neurite Outgrowth in PC12-GTK Cells-To assess to what extent NGF-dependent activation of Rap1 in GTK-overexpressing PC12 cells is responsible for the increase in neurite FIG. 7. Activation of the CrkII signaling pathway and elevated FAK levels in PC12-GTK cells. A, subconfluent cells (parental PC12 cells, GTK-7, and GTK-10 cells) were cultured overnight in medium containing 2% FCS and 1% HS and stimulated with 100 ng/ml NGF for 10 min. Cells were immunoprecipitated (IP) with anti-FAK antibody and subjected to Western blot (WB) analysis with phosphotyrosine antibody. The membrane was then stripped and reprobed for FAK. 10% of the cell extract was subjected to Western blot analysis for total FAK. B, cells were cultured as above and stimulated with 100 ng/ml NGF for 3 min. Cells were immunoprecipitated with anti-CrkII antibody and subjected to Western blot analysis with phosphotyrosine antibody. The membrane was then stripped and reprobed for CrkII, p130 Cas , FAK and C3G (bottom). 10% of the cell extract was subjected to Western blot analysis for CrkII, p130 Cas , FAK, and C3G (top). C, cells were cultured as in A and immunoprecipitated with p130 Cas antibody. The immunoprecipitates were subjected to Western blot analysis for phosphotyrosine (4G10) and p130 Cas . Total amount of p130 Cas in lysate was assessed as above. D, cells were cultured as in A and immunoprecipitated with C3G antibody. The immunoprecipitates were subjected to Western blot analysis for phosphotyrosine and C3G. Total amount of C3G in crude lysate was assessed as above. E, control and PC12-GTK cells were cultured as above, lysed, and incubated with GST-CrkSH2 fusion protein immobilized to glutathione-Sepharose. The beads were washed, and SDS sample buffer was added. The proteins were subjected to Western blot analysis for phosphotyrosine (4G10), Shb, and FAK. outgrowth in these cells, transient transfections of RalGDS were performed. When expressed, this construct will yield a product, which contains the 97-amino acid long Rap1-binding domain (RBD), thus associating with Rap1 and blocking its interaction with physiological downstream effectors (23). Control PC12 cells and GTK-overexpressing PC12 cells were transiently transfected with the pIRES-EGFP-vector alone or together with a RalGDS-RBD construct and cultured in the absence or presence of NGF (50 ng/ml), and fluorescent cells with neurites were counted. RalGDS-RBD expression significantly decreased neurite outgrowth of PC12-GTK cells (Fig.  9A), which had been cultured for 2 days in the absence and presence of 50 ng/ml NGF, by 53 and 39%, respectively, but did not affect NGF-dependent neurite outgrowth of parental PC12 cells, indicating that Rap1 activation is at least partially responsible for the differentiation caused by GTK overexpression. The differences in neurite outgrowth between NGF-treated and untreated control-or RalGDS-RBD-transfected PC12-GTK cells was similar, 20 and 18%, respectively, suggesting that NGF-induced differentiation is not dependent on Rap1 activation in the GTK cells. This is in line with a previous study showing that Rap1 is not activated by NGF (10).
Although RalGDS-RBD can also interact with Ras, it has been reported that the RalGDS-RBD is rather specific for Rap1, with an affinity in vitro for active Rap1 about 100-fold greater than that for active Ras (28). However, to exclude the possibility that the effects of RalGDS-RBD were due to interactions with Ras, we also performed transient transfection of the Rap1specific GTPase-activating protein, Rap1GAP, together with EGFP and counted GFP-expressing cells with neurites as described above (Fig. 9B). Expression of Rap1GAP significantly decreased neurite outgrowth of PC12-GTK cells in the absence and presence of NGF but did not affect NGF-dependent neurite outgrowth of parental PC12 cells. Thus, using two different strategies for inhibiting Rap1 signaling, these experiments demonstrate a critical role for Rap1 in mediating GTK-induced neurite outgrowth.

DISCUSSION
We have previously shown that GTK plays a role in inhibiting cell proliferation in NIH3T3 (16) and RINm5F cells (17). Furthermore GTK also regulates hormone production in the insulin-producing cell line RINm5F (17), raising the possibility that this kinase plays a role for cell differentiation. The PC12 cells are commonly employed for studies on neuronal cell differentiation, and neurite outgrowth in response to v-Src expression sets a precedent for the possibility that other Src family members may operate in a similar fashion (29). In this study we show that wild-type GTK is kinase-active and has a great impact on NGF-independent neurite outgrowth when overexpressed in PC12 cells. This could either reflect the aberrant activation of new routes promoting neurite extension or the amplification of the physiological signaling pathway of endogenous GTK, which is expressed at very low levels. GTK could be one member of a family of kinases that all serve a similar role. Due to the possibility of negative regulation of GTK activity through phosphorylation of C-terminal tyrosines in PC12 cells, we also attempted to express GTK Y504F and GTK Y497F/Y504F mutants, but we failed to obtain clones that would survive and divide. We presently cannot assess to what extent there is negative regulation of GTK kinase activity in PC12 cells. However, the transfection of PC12 cells with GTK Y497F/Y504F resulted in single cells or clusters of cells with large flattened cell bodies and long branched neurites which ceased to grow and eventually died (result not shown).
Several studies have reported that overexpression or constitutive activation of proteins that activate the Ras-MAPK pathway induce spontaneous differentiation of PC12 cells or enhance the response to differentiation factors (i.e. TrkA (30), MEK1 (6), Ras (31), Raf (32), PKC⑀ (33), and Crk (34)). It has been argued that NGF-induced differentiation of PC12 cells is associated with prolonged phosphorylation and activation of ERK. We did not observe any increase in ERK activation by GTK either in the presence or absence of NGF. Moreover, the MEK1 inhibitor PD98059, a known suppressor of NGF-induced differentiation (35), did not reduce the basal neurite outgrowth of PC12 cells overexpressing GTK. These results suggest that the GTK-mediated differentiation of PC12 cells is ERK-independent, although it cannot be excluded that GTK overexpression induces a weak but prolonged ERK activation, undetectable by Western blot analysis, induced by some other factor than MEK1. Other reports of ERK-independent signals promoting neurite outgrowth in PC12 cells have been presented, i.e. the signaling through SAPK/JNK (36,37), p38 (38), Shb (24), and ␤PDGF-R (39), suggesting other pathways that control neurite outgrowth.
Rap1, a small GTPase of the Ras family (9,40), has been suggested to play a role in the sustained activation of ERK by acting through B-Raf (8,41,42), but a recent study reports that NGF fails to activate Rap1 (10) making this issue controversial. The amino acid sequences of Rap1 and Ras show about 50% identity to each other. Due to this high sequence similarity, Rap1 can bind to Ras effector molecules, however, not always activating them. Thus, it has been suggested that Rap1 functions as a Ras antagonist, suppressing Ras-dependent signaling, including the ERK pathway (43). Several distinct pathways may transduce signals toward Rap1 activation as follows: an increase in intracellular calcium, release of diacylglycerol, cAMP synthesis, and activation of C3G or some other Rap1 guanine exchange factor. We investigated Rap1 activity in PC12-GTK cells and found a significant increase in GTP-bound Rap1. We also observed a partial suppression of neurite outgrowth in GTK-overexpressing cells after transient transfection with RalGDS-RBD and Rap1GAP constructs clearly indicating that Rap1 activation is required for a significant proportion of the neurite outgrowth in PC12-GTK cells. Several pieces of evidence have been presented arguing for the CrkIIsignaling pathway as responsible for this effect. First, overexpression of Crk in PC12 cells has previously been reported to induce neurite formation (34). Second, we show an increased association of p130 Cas and FAK with CrkII in PC12-GTK cells. C3G, a specific guanine exchange factor for Rap1 (44), was also present in this complex. C3G is known to bind the N-terminal SH3 domain of CrkII and as a consequence the Crk-C3G complex is translocated to the plasma membrane via binding of the Crk SH2 domain upon stimulation (reviewed in Ref. 27). The observed binding of p130 Cas to CrkII could serve this purpose of retargeting C3G. Third, we observe a markedly increased phosphorylation of Shb and an association of Shb with the SH2 domain of CrkII. The preferred binding site for the CrkII SH2 domain contains a proline in position 3 downstream of the phosphorylated tyrosine (45). There are three candidate tyrosines in Shb with a proline in this position, namely Tyr-333, Tyr-355, and Tyr-384. If one or several of these are phosphorylated, they could serve as binding sites for CrkII. We have shown that PC12 cells overexpressing Shb (24) exhibit an increased neurite outgrowth in response to NGF, and this is at least partially dependent on Rap1 activation. 2 NGF induces phosphorylation of overexpressed Shb and an increased association between CrkII and an unknown phosphotyrosine protein of 130 -135 kDa. Tyrosine phosphorylation of Shb appears to coincide with neurite outgrowth in both the GTK-and Shboverexpressing PC12 cells, being NGF-independent in the former and NGF-dependent in the latter case. Overexpression of GTK in PC12 cells causes augmented TrkA and Shb phosphorylation that results in the association between FAK and Shb. FAK further binds p130 Cas that forms a complex with CrkII by binding via its SH2 domain. The SH3 domain of CrkII binds C3G, a guanine exchange factor for Rap1, which activates some unknown downstream effector of Rap1 that stimulates neurite outgrowth.
is the observation of an increased FAK content and its association with Shb and CrkII. Upon activation and phosphorylation in cell adhesions, FAK associates with a tyrosine kinase, i.e. Src, and this promotes the direct binding of downstream signaling proteins such as p130 Cas . Phosphorylation of p130 Cas then allows the association of CrkII to the complex via its SH2 domain (for reviews see Refs. 27 and 46). The interaction between FAK and CrkII, perhaps via p130 Cas , observed in PC12-GTK might be a consequence of the increased level of phosphorylated FAK, and this in turn may contribute, via C3G, to the increase in Rap1 activation. Shb could possibly serve a similar role as p130 Cas for CrkII-C3G retargeting to the proximity of FAK since Shb interacts with both FAK and CrkII. A recent report by Altun-Gultekin and co-workers (47) showing that v-Crk-expressing PC12 cells exhibit a flattened phenotype with broad lammelipodia and an up-regulation in the expression of FAK, similar to what we observe in GTK-overexpressing cells, supports the hypothesis that FAK is involved in neurite outgrowth of PC12 cells. However, since FAK controls cellular responses to the extracellular matrix, including adhesion, spreading, and migration, the phenotype observed in the v-Crkand GTK-overexpressing cells may be a combination of differentiation and induced spreading. The elevated FAK levels observed in PC12-GTK cells could be due to increased gene transcription or reduced FAK degradation by proteolytic proteins as a consequence of its activation and association with other proteins.
The SAPK/JNK pathway has been suggested to play a role in transducing NGF-and ERK-independent neurite outgrowth (36,37). We therefore studied SAPK/JNK phosphorylation in PC12-GTK and parental PC12 cells before and after NGF treatment by Western blot analysis using specific antisera against phosphorylated SAPK/JNK. We could not detect any increase in SAPK/JNK phosphorylation in the GTK-expressing cells making it unlikely that this pathway is involved in GTKmediated neurite outgrowth (results not shown). The observed increased phosphorylation of TrkA, Cas, and Shb in PC12-GTK cells could either result from the direct phosphorylation by GTK or as a consequence of the activation of another kinase. The strong increase in Shb phosphorylation raises the possibility that Shb is a substrate for GTK or at least an early effector for GTK signaling.
To summarize the results in the present study we suggest the following model for GTK-induced neurite outgrowth (Fig.  10). Overexpression of GTK causes augmented TrkA and Shb phosphorylation and FAK overexpression. Shb and p130 Cas associate with FAK, thus generating binding sites for the CrkII-C3G complex via the CrkII SH2 domain. This will activate Rap1 and some downstream signaling pathway, such as AF6, Nore1, Krit, or Ral (48 -51) or perhaps some unknown pathway that induces neurite outgrowth.