Fibroblast Transformation by Fps/Fes Tyrosine Kinases Requires Ras, Rac, and Cdc42 and Induces Extracellular Signal-regulated and c-Jun N-terminal Kinase Activation*

The small GTP-binding proteins Ras, Rac, and Cdc42 link protein-tyrosine kinases with mitogen-activated protein kinase (MAPK) signaling cascades. Ras controls the activation of extracellular signal-regulated kinases (ERKs), while Rac and Cdc42 regulate the c-Jun N-terminal kinases (JNKs). In this study, we investigated whether small G protein/MAPK cascades contribute to signal transduction by transforming variants of c-Fes, a nonreceptor tyrosine kinase implicated in cytokine signaling and myeloid differentiation. First, we investigated the effects of dominant-negative small G proteins on Rat-2 fibroblast transformation by a retroviral homolog of c-Fes (v-Fps) and by c-Fes activated via N-terminal addition of the v-Src myristylation signal (Myr-Fes). We observed that dominant-negative Ras, Rac, and Cdc42 inhibited v-Fps- and Myr-Fes-induced growth of Rat-2 cells in soft agar, indicating that activation of these small GTP-binding proteins is required for fibroblast transformation by Fps/Fes tyrosine kinases. To determine whether MAPK pathways are activated downstream of these small G proteins, we measured ERK and JNK activity in the v-Fps- and Myr-Fes-transformed Rat-2 cells. Both ERK and JNK activities were elevated in the transformed cells, suggesting that these pathways are involved in cellular transformation. Dominant-negative mutants of Ras (but not Rac or Cdc42) specifically inhibited ERK activation by v-Fps and Myr-Fes, demonstrating that ERK activation occurs exclusively downstream of Ras. All three dominant-negative small G proteins inhibited JNK activation by v-Fps and Myr-Fes, indicating that JNK activation by these tyrosine kinases requires both Ras and Rho family GTPases. These data demonstrate that multiple small G protein/MAPK cascades are involved in downstream signal transduction by Fps/Fes tyrosine kinases.

Members of the Ras superfamily of small GTP-binding proteins regulate diverse cellular functions, including cell proliferation, differentiation, and organization of the actin cytoskeleton (1)(2)(3)(4). One key function of these small GTP-binding proteins is to link upstream protein-tyrosine kinases with mi-togen-activated protein kinase (MAPK) 1 pathways. Perhaps the best characterized MAPK pathway involves Raf/MEK/ERK activation downstream of Ras (5)(6)(7). Raf binds to the GTPbound, active form of Ras at the plasma membrane and becomes activated. Active Raf phosphorylates and activates the dual specificity MAPK kinases MEK1 and MEK2 which in turn activate the MAPKs ERK1 and ERK2. Activated ERKs can then translocate to the nucleus and phosphorylate a number of transcription factors. In this way, the Ras/ERK pathway provides a critical link between tyrosine kinase signal transduction and transcriptional events in the nucleus.
In addition to the Ras/ERK cascade, recent work has revealed parallel small G protein/MAPK pathways to the nucleus. Small GTPases of the Rho family including Rac and Cdc42 (but not Rho itself) are linked to a kinase cascade leading to the activation of the c-Jun N-terminal kinase (JNK), an ERK-related serine kinase that specifically phosphorylates and activates the c-Jun transcription factor (8 -10). This cascade is also linked to a related kinase known as p38. Both p38 and JNK are activated in response to cytokines as well as cellular stress. Rac and Cdc42 are connected to JNK via intermediate MEKK (analogous to Raf) and SEK (analogous to MEK) activities. The Rac/Cdc42/JNK cascade also includes additional kinases upstream of MEKK, such as p21-activated kinase 65, which may represent the direct targets of active Rac and Cdc42 (7,8). Thus in addition to their roles in the regulation of cytoskeletal architecture (3,4,11), Rac and Cdc42 can control transcriptional events in the nucleus as well.
Activation of small GTPase/MAPK pathways is a common feature of both normal and oncogenic protein-tyrosine kinase signal transduction (12). The biochemical connection between tyrosine kinases and the activation of small GTPases downstream often involves recruitment and activation of guanine nucleotide exchange factors. In the case of Ras, autophosphorylation of activated tyrosine kinases can create binding sites for the SH2 domains of adaptor proteins such as Shc and Grb-2, which serve as intermediates in the recruitment of the SOS guanine nucleotide exchange factor to the plasma membrane, focal adhesions, or other sites of tyrosine kinase signaling (13,14). The translocation of SOS to these sites allows for Ras activation and subsequent stimulation of the MAPK pathway downstream. In the case of Rho family GTPases, tyrosine phosphorylation of the guanine nucleotide exchange factor may be essential for activation. For example, recent studies have shown that tyrosine phosphorylation stimulates the exchange activity of Vav, leading to activation of the JNK pathway (15,16). In addition, products of the phosphatidylinositol 3Ј-kinase pathway have recently been shown to regulate the exchange activity of Vav and SOS toward Rho family GTPases (17,18).
The Fps/Fes family of cytoplasmic protein-tyrosine kinases are expressed primarily in myeloid hematopoietic cells and the vascular endothelium (19 -23). Human c-Fes has been linked to multiple cytokine receptors (24 -28) and may play a direct role in myeloid differentiation and angiogenesis (22,29,30). However, the critical downstream signaling pathways regulating the biological activities of Fes have not been fully characterized. Previous studies have identified Bcr, Ras GAP, and Shc as substrates for c-Fes and its transforming viral homolog, v-Fps (31)(32)(33)(34)(35). All of these proteins have been linked to the regulation of the Ras and Rho families of small GTPases. For example, Bcr has a C-terminal GAP domain for Rac and Cdc42 (36) as well as a central guanine nucleotide exchange region with homology to Dbl (37,38). Previous work from our laboratory has shown that tyrosine phosphorylation of Bcr in v-Fps-transformed fibroblasts induces its association with Grb-2/SOS, suggesting that Bcr may couple Fps/Fes tyrosine kinases to both the Ras and Rho signaling pathways downstream (34). In this study, we provide evidence that small G proteins of both the Ras and Rho families are required for Fps/Fes-induced fibroblast transformation and that small G protein activation is coupled to the ERK and JNK pathways downstream. These data provide strong evidence that activation of multiple small G protein/ MAPK cascades is critical to Fps/Fes signal transduction.

Construction of Plasmid Vectors-
The full-length cDNAs for N17Ras, V12N17Rac, and N17Cdc42 containing a C-terminal Myc epitope tag were the generous gift of Dr. Alan Hall, MRC, London (39). These cDNAs were subcloned into the retroviral vector pSR␣MSVtkneo (40) and into the mammalian expression vector pCDNA3 (Invitrogen). Addition of the N-terminal myristylation signal sequence of v-Src to the N-terminal region of c-Fes to create the transforming oncoprotein Myr-Fes was accomplished using a polymerase chain reaction-based approach and will be described in detail elsewhere. 2 Construction of a full-length human c-fes cDNA with a C-terminal FLAG epitope tag is described elsewhere (42). A similar polymerase chain reaction-based procedure was used to add the FLAG coding sequence to the C-terminal region of v-Fps and Myr-Fes. The coding sequence of the v-src oncogene was amplified by polymerase chain reaction from a Rous sarcoma virus template obtained from the ATCC and subcloned into the pSR␣MSVtkneo and pCDNA3 expression vectors described above.
Production of Recombinant Retroviruses and Stably Transfected Rat-2 Cells-Retroviral stocks were prepared by transient transfection of 293T cells (43) with the pSR␣MSVtkneo constructs together with an ecotropic packaging vector (40). Briefly, 293T cells were grown to 50 -80% confluence in 10-cm 2 dishes and transfected using a calcium phosphate method as described previously (42). Culture medium containing the recombinant retroviruses was harvested from 2 to 5 days posttransfection. Viral supernatants were pooled and stored in frozen aliquots at Ϫ80°C. Negative control retroviruses were prepared using the parent vector alone.
Rat-2 cells were infected with recombinant retroviruses in the presence of polybrene as described (44). G418 was added to a final concentration of 0.8 mg/ml 48 h later, and the cells were grown under G418 selection for 2-4 weeks. The expression of N17Ras, V12N17Rac, and N17Cdc42 was confirmed by immunoprecipitation followed by immunoblotting with the anti-Myc monoclonal antibody, 9E10 (39). Stable expression of c-Fes was confirmed by immunoprecipitation and immunoblotting with the anti-FLAG monoclonal antibody, M2 (Kodak Scientific Imaging Systems).
Transformation Assay-Rat-2 cells (5 ϫ 10 4 ) stably expressing N17Ras, V12N17Rac, N17Cdc42, or the neo marker alone were plated in 6-well dishes. Cells were infected with v-Src, v-Fps, or Myr-Fes retroviruses 24 h later as described above and incubated for 2 addi-tional days. The cells were then trypsinized, and 1 ϫ 10 4 cells were replated in triplicate in Dulbecco's modified Eagle's medium containing 0.3% agar and 20% fetal calf serum. Transformed colonies were counted after 2 weeks.
To establish Rat-2 cell lines stably expressing each of the tyrosine kinases, single transformed colonies were isolated from the soft agar plates and grown in Dulbecco's modified Eagle's medium with 5% fetal calf serum. The expression of v-Src was confirmed by immunoprecipitation and immunoblotting with an anti-Src antibody (N-16, Santa Cruz Biotechnology); Myr-Fes and v-Fps expression were confirmed with the M2 anti-FLAG antibody.
In Vitro JNK Assay-Rat-2 or 293T cells were incubated overnight in serum-free medium prior to harvesting. Clarified cell lysates were prepared as described previously (35). For some experiments, a solid phase JNK assay was employed (45). Cell lysates (75 g of protein) were incubated with 2 g of an immobilized GST-Jun fusion protein (Jun amino acids 1-72) in a final volume of 300 l of HBB buffer (20 mM HEPES, pH 7.7, 50 mM NaCl, 0.1 mM EDTA, 2.5 mM MgCl 2 , and 0.05% Triton X-100) for 3 h at 4°C. The GST-Jun fusion protein was then washed with four 1.0-ml aliquots of HBB buffer and incubated with 35 l of kinase buffer (20 mM HEPES, pH 7.5, 20 mM ␤-glycerophosphate, 10 mM MgCl 2 , 40 M ATP, 1 mM dithiothreitol, and 50 M Na 3 VO 4 ) containing 5 Ci of [␥-32 P]ATP (3000 Ci/mmol, NEN Life Science Products) for 20 min at 30°C. Phosphorylated proteins were separated by SDS-PAGE and visualized by storage phosphorimaging (Molecular Dynamics PhosphorImager). Alternatively, JNK activity was assessed by immune complex kinase assay. Clarified cell lysates (75 g) were incubated with 0.8 g of JNK1 antibody (C-17, Santa Cruz Biotechnology) in HBB buffer at 4°C for 3 h. Immune complexes were precipitated with protein G-Sepharose, washed with HBB buffer, and divided into two aliquots. One aliquot was tested for JNK activity using the in vitro kinase assay conditions described above for the solid phase assay. The second aliquot was immunoblotted with the JNK antibody to verify equivalent levels of the JNK protein in each reaction.
In Vitro ERK Assay-Clarified cell lysates (75 g) were incubated with MAPK lysis buffer (40 mM Tris-HCl, pH 7.4, 120 mM NaCl, 5 mM EDTA, 2 mM EGTA, 1 mM phenylmethylsulfonyl fluoride, 0.5% Triton X-100, 10 mM NaF, 2 mM Na 3 VO 4 , 10 mM sodium pyrophosphate, 10 g/ml leupeptin, and 5 g/ml aprotinin) and immunoprecipitated with ERK1 and ERK2 antibodies (0.4 g each; Santa Cruz Biotechnology). Immune complexes were precipitated with protein G-Sepharose, washed three times with 1.0 ml of MAPK lysis buffer and twice with MAP kinase buffer (40 mM Tris-HCl, pH 7.4, 20 mM MgCl 2 , 2 mM MnCl 2 , 25 M ATP, 0.5 mM dithiothreitol), and divided into two aliquots. One aliquot was assayed for ERK activity in 40 l of MAP kinase buffer containing 2 g of myelin basic protein (MBP) and 5 Ci of [␥-32 P]ATP. Phosphorylation reactions were incubated for 20 min at 30°C, and labeled MBP was separated by SDS-PAGE and visualized by storage phosphorimaging. The second aliquot was immunoblotted with the ERK1 antibody to verify equivalent amounts of ERK protein in each immunoprecipitate.

Dominant-Negative Mutants of the Small GTP-binding Proteins Ras, Rac, and Cdc42 Block Rat-2 Transformation by Fps/ Fes
Tyrosine Kinases-In this study, we used fibroblast transformation as a model system to study small G protein/MAPK signaling by v-Fps and Myr-Fes, two activated forms of the Fps/Fes tyrosine kinase family. v-Fps is a potent transforming homolog of c-Fes found in the chicken sarcoma virus, FSV (46,47). Myr-Fes is a form of human c-Fes that is activated by N-terminal addition of the v-Src myristylation signal sequence, which results in membrane targeting (22). To investigate whether Ras, Rac, and Cdc42 are involved in fibroblast transformation by fps/fes oncogenes, we inhibited the function of these small G proteins in Rat-2 cells with dominant-inhibitory mutants in which the Ser residue at position 17 was changed to Asn (N17 mutants). These mutants bind preferentially to GDP and act as dominant inhibitors of endogenous small G protein function presumably by blocking their access to exchange factors (39,48,49). Populations of Rat-2 fibroblasts stably expressing each of these mutants were infected with v-Fps and Myr-Fes retroviruses and assayed for soft agar colony formation as a measure of transformation. As a control, we also examined transformation by v-Src in these cell lines, because previous work has shown that dominant-negative mutants of Ras and Rac interfere with transformation by this oncogenic tyrosine kinase (10). As shown in Fig. 1A, we observed that v-Src, v-Fps, and Myr-Fes consistently produced fewer transformed colonies from cells expressing the dominant-negative small G proteins as compared with control cells. Control immunoblots demonstrate equivalent expression of v-Src, v-Fps, or Myr-Fes in all four cell lines (Fig. 1B). These results indicate that fibroblast transformation by oncogenically activated Fps/ Fes tyrosine kinases as well as v-Src requires the activation of Ras and Rho family GTPases.
To verify that the dominant-negative small GTPases did not induce a general suppression of cellular proliferation, the growth rates of the Rat-2 cell populations stably expressing each of the mutants were compared with that of the control cells expressing only the drug-selection marker. As shown in Fig. 2A, Rat-2 cells stably expressing dominant-negative Ras, Rac, and Cdc42 grew at the same rate and to the same saturation density as the vector control cells. Expression of the dominant-negative small G proteins was verified by immunoprecipitation followed by immunoblotting (Fig. 2B).
As an alternative experimental approach to investigate the effect of dominant-negative small G proteins on Fps/Fes-mediated transformation, soft agar colony-forming assays were per-formed on Rat-2 fibroblasts following co-infection with recombinant retroviruses carrying the transforming oncogenes and the dominant-negative GTPase mutants. As shown in Fig. 3, this approach also led to strong suppression of transforming activity. These results are in good agreement with those observed using the Rat-2 cell populations stably expressing the dominant-negative mutants (Fig. 1). Taken together, these data support the conclusion that Ras, Rac, and Cdc42 play essential roles in transformation signaling downstream from oncogenic Fps/Fes tyrosine kinases as well as v-Src.
Activation of ERK by v-Fps and Myr-Fes-Activation of many tyrosine kinase families is linked to stimulation of ERK activity (see Introduction). To determine whether the ERK pathway is activated by Fps/Fes kinases, we measured ERK activity in v-Fps-and Myr-Fes-transformed Rat-2 cell lines derived from single transformed soft-agar colonies. We also examined ERK activation in a Rat-2 cell line stably expressing c-Fes and in control G418-resistant Rat-2 cells. Endogenous ERK1 and ERK2 were immunoprecipitated from clarified cell lysates and an in vitro kinase assay was conducted with MBP and [␥-32 P]ATP as co-substrates. As shown in Fig. 4, ERK activity was increased 6 -7-fold in the v-Fps-and Myr-Fes-transformed cells. In contrast, stable expression of c-Fes, which is nontransforming, resulted in only a 2-fold increase in ERK activity. These results suggest that transformation by activated forms of Fps/Fes kinases may require sustained ERK activation. In contrast, no activation of ERK was observed in the v-Src transformants, suggesting that maintenance of the v-Src-transformed phenotype does not require constitutive MAPK activation. and Cdc42 (f). Following G418 selection, 4 ϫ 10 4 cells from each drug-resistant population were plated in each of five 60-mm culture plates. Viable cells present per plate were determined on each of the days indicated by trypan blue exclusion. These selected cell populations were used for the transformation assays used shown in Fig. 1. B, the expression of N17Ras (dnRas), V12N17Rac (dnRac), or N17Cdc42 (dnCdc42) in the selected Rat-2 cell populations from A was detected by immunoprecipitation followed by immunoblotting with the anti-Myctag monoclonal antibody, 9E10. Each of the small G protein mutants carries a C-terminal Myc epitope tag (39). Cells expressing only the neo selection marker were included in this experiment as a negative control (vector).
The surprising result that transformation by v-Src was not associated with constitutive ERK activation led us to investigate the effects of v-Src and Fps/Fes kinases on this pathway under transient expression conditions. For these experiments, we employed the human embryonic kidney cell line, 293T (43), and the in vitro ERK assay described above. As shown in Fig.  4, both v-Src and v-Fps stimulated ERK activity in this system. However, both c-Fes and Myr-Fes were unable to stimulate ERK activity in 293T cells. This lack of activation may reflect the tight regulation of both c-Fes and Myr-Fes tyrosine kinase activity in this cell line, despite the high level of protein expression (42). 2 A nearly identical pattern of ERK activation was also observed in Rat-2 cells transiently expressing v-Fps, Myr-Fes, c-Fes, and v-Src, validating the use of 293T cells as a model system for the investigation of small G protein/MAPK signaling following transient expression of these tyrosine kinases (data not shown).

Activation of JNK by v-Fps and Myr-Fes-Recent studies have shown that JNK is activated downstream of Rac and
Cdc42 (see Introduction). To determine whether the JNK pathway is activated by v-Fps and Myr-Fes, JNK activity was measured in the v-Src-, v-Fps-, and Myr-Fes-transformed Rat-2 cells. Endogenous JNK was precipitated from cell lysates with an immobilized GST-Jun fusion protein and kinase activity of substrate-bound JNK was assayed by addition of [␥-32 P]ATP. As shown in Fig. 5, JNK activity was elevated in the v-Src-, v-Fps-, and Myr-Fes-transformed Rat-2 cells, suggesting that constitutive JNK activation may contribute to transformation by these oncogenes. In contrast, no increase in JNK activity was observed in nontransformed Rat-2 cells stably expressing c-Fes. Thus, JNK activation correlates with the transforming activity of these tyrosine kinase oncogenes.
Endogenous JNK activity was also measured in 293T cells transiently expressing v-Src, v-Fps, Myr-Fes, and c-Fes. As shown in Fig. 5, expression of v-Src and v-Fps readily stimulated JNK activity in 293T cells. In contrast to Rat-2 cells, JNK activity was only modestly increased by Myr-Fes in 293T cells.
Similar differences were observed for Myr-Fes-induced activation of ERK in Rat-2 versus 293T cells (Fig. 4), and are consistent with the hypothesis that cells responding with activation of  these pathways are selected for during transformation. c-Fes also slightly activated JNK over background in 293T cells, suggesting that JNK may be more sensitive than ERK to the low level of Fes tyrosine kinase activity resulting from the overexpression of Fes in this cell line (42).

Dominant-Negative Ras Selectively Blocks ERK Activation by v-Fps and Myr-Fes-Given that ERKs are often activated
downstream of Ras, we tested whether dominant-negative Ras selectively blocks ERK activation in response to transforming signals from v-Fps and Myr-Fes. Populations of Rat-2 cells stably expressing dominant-negative mutants of Ras, Rac, and Cdc42 (as well as vector control cells) were infected with v-Fps and Myr-Fes retroviruses. One week later, endogenous ERK activity was immunoprecipitated from cell lysates and analyzed by in vitro kinase assay using MBP as a substrate. As shown in Fig. 6, dominant-negative Ras dramatically inhibited ERK activation by both v-Fps and Myr-Fes. In contrast, dominant-negative Rac and Cdc42 showed no effect on ERK activation. These results clearly indicate that constitutive ERK activation in Myr-Fes and v-Fps transformed cells occurs downstream of Ras.
The effects of dominant-negative Ras, Rac, and Cdc42 on ERK activation by v-Src and v-Fps were also examined in 293T cells. Myr-Fes was not tested because it does not activate ERK in this cell line (Fig. 4). v-Src and v-Fps were transiently expressed either alone or together with the dominant-negative small G proteins, and endogenous ERK activity was analyzed by in vitro kinase assay. As shown in Fig. 7, dominant-negative Ras inhibited ERK activation by both v-Src and v-Fps in 293T cells, while the other mutant GTPases were without effect. Expression of the mutant small G proteins and the tyrosine kinase oncogenes was verified by immunoblotting (data not shown). Similarly, co-infection of Rat-2 cells with recombinant retroviruses carrying the transforming tyrosine kinase oncogenes and dominant-negative Ras also blocked ERK activation under transient expression conditions (data not shown). These results indicate that transient ERK activation by v-Src and v-Fps is also regulated by Ras.

Dominant-Negative Ras, Rac, and Cdc42 Block JNK Activation by v-Fps and Myr-Fes-Recent studies have shown that
Rac and Cdc42 specifically regulate the JNK pathway (8 -10, 18). In some cases, Rac and Cdc42 may be regulated by Ras upstream (49,50). To determine the contribution of each of these small G proteins to JNK activation by Fps/Fes kinases, populations of Rat-2 cells expressing dominant-negative mutants of these small G-proteins as well as vector control cells were infected with v-Fps, Myr-Fes, and v-Src retroviruses. One week later, endogenous JNK activation was analyzed by immune complex kinase assay with the GST-Jun substrate as described under "Materials and Methods." As shown in Fig. 8, activation of JNK by v-Fps and Myr-Fes was inhibited by all three dominant-negative small G protein mutants. These data suggest that both Rac and Cdc42 contribute to JNK activation by Fps/Fes tyrosine kinases, and that Ras activation is also involved in this pathway. Mutant Rac and Cdc42 also strongly suppressed JNK activation in v-Src-transformed Rat-2 cells (Fig. 8). However, dominant-negative Ras did not interfere with JNK activation by v-Src, suggesting that Ras is not required for activation of the Rac/Cdc42/JNK pathway in this case. Control immunoblots demonstrated an equivalent amount of JNK in each kinase assay (Fig. 8, bottom). Equivalent expression of each oncogenic tyrosine kinase was also verified by immunoblotting (data not shown). DISCUSSION In this study, we investigated the role of small GTPases of the Ras and Rho families and their cognate MAPK pathways in signal transduction by transforming variants of Fps/Fes tyro-

FIG. 6. Dominant-negative Ras blocks ERK activation by v-Fps and Myr-Fes in Rat-2 cells.
Populations of Rat-2 cells expressing dominant-negative mutants of Ras (dnRas), Rac (dnRac), and Cdc42 (dnCdc42) as well as the neo selection marker alone (vector) were infected with v-Fps and Myr-Fes retroviruses for 4 h as described under "Materials and Methods." One week later, endogenous ERK activity was analyzed using the in vitro kinase assay described under "Materials and Methods." Phosphorylated MBP (arrowhead) was separated by SDS-PAGE and visualized by storage phosphor technology (top). Experiments were performed in triplicate and the bar graphs (center) show the mean fold increase relative to the uninfected vector control cells Ϯ S.E. Control immunoblots verified similar amounts of ERK (arrowhead) in each immunoprecipitate (bottom).

FIG. 7. Dominant-negative Ras blocks ERK activation by v-Src and v-Fps in 293T cells.
Human 293T cells were co-transfected with v-Src or v-Fps and dominant-negative mutants of Ras (dnRas), Rac (dnRac), Cdc42 (dnCdc42) or empty vector (vector). As a negative control, cells were transfected with the empty parent vector alone. Endogenous ERK activity was analyzed using the in vitro kinase assay described under "Materials and Methods." Phosphorylated MBP (arrowhead) was separated by SDS-PAGE and visualized by storage phosphor technology (top). Experiments were performed in triplicate and the bar graphs show the mean fold increase relative to the vector control Ϯ S.E. Control immunoblots verified similar amounts of ERK (arrowhead) in each immunoprecipitate (bottom). sine kinases. Using dominant-negative mutants of Ras, Rac, and Cdc42, we observed that activation of all three of these small GTPases is a necessary downstream event in Fps/Fesmediated fibroblast transformation. Consistent with this requirement for small G protein activation in Fps/Fes-mediated signaling is our observation that both the ERK and JNK signaling pathways were constitutively activated in the transformed cells. Taken together, our data implicate multiple small G protein/MAPK kinase cascades in mitogenic signal transduction by Fps/Fes tyrosine kinases. These results are consistent with several previous reports that regulatory proteins for Ras and Rho family GTPases are substrates for this nonreceptor tyrosine kinase family (31)(32)(33)(34)(35).
Experiments with dominant-negative mutants of Ras and Rho family GTPases presented here provide insight into the mechanism of ERK and JNK activation by Fps/Fes tyrosine kinases. ERK activation by Fps/Fes in both the transformed fibroblasts and a transient expression system (human 293T cells) was potently inhibited by dominant-negative Ras but not by dominant-negative Rac or Cdc42. This result provides strong evidence that Ras is the upstream regulator of the ERK pathway in both systems. In contrast to the Ras/ERK pathway, dominant-negative Ras, Rac, and Cdc42 all inhibited JNK activation in Fps/Fes-transformed fibroblasts, indicating overlap of Ras, Rac, and Cdc42 function in the regulation of JNK. These data suggest that Rac and Cdc42 are at least partially responsible for the activation of JNK in Fps/Fes-transformed cells, and are in agreement with earlier reports that these small G proteins are important regulators of JNK activity in vivo (8 -10). The ability of dominant-negative Ras to suppress JNK activation is consistent with previous work showing that Ras may lie upstream of Rac and Cdc42 (49 -52). Alternatively, p120 Ras GAP may provide a link between Ras activation and JNK (53).
Data presented here suggest that activation of the Rac/ Cdc42/JNK pathway contributes to fibroblast transformation by v-src, v-fps, and other tyrosine kinase oncogenes. Consistent with this hypothesis are the previous findings that Bcr-Abl, the oncogenic protein-tyrosine kinase associated with chronic myelogenous leukemia (54), preferentially induces JNK activation in fibroblasts and hematopoietic cells transformed by this oncoprotein (55). JNK-mediated activation of the c-Jun transcription factor may be an important aspect of transformation by both tyrosine kinase oncogenes as well as Ras. For example, fibroblast transformation by Bcr-Abl is blocked by dominantnegative mutants of c-Jun (55), and Ras is unable to transform fibroblasts in which Jun has been inactivated by gene targeting (56). However, in addition to the activation of JNK, Rac and Cdc42 also regulate the organization of the actin cytoskeleton (3,4,11) and Rac-induced changes in cytoskeletal architecture correlate with its ability to induce cellular transformation (57). Another alternative pathway downstream of Rho family GTPases may involve JNK-independent activation of nuclear factor-B (58). Therefore, dominant-negative mutants of Rac and Cdc42 may inhibit transformation by Fps/Fes, v-Src, and other tyrosine kinases by blocking multiple intracellular signaling events in addition to interfering with JNK activation.
Comparison of small G protein/MAPK signaling by v-src and fps/fes oncogenes revealed some important differences. While both v-Src and v-Fps both potently activated ERK in the 293T cell transient expression system in a Ras-dependent manner (Figs. 4 and 7), constitutive ERK activation was not observed in the v-Src-transformed Rat-2 cells (Fig. 4). This finding suggests that although ERK activation may be an essential early event in v-Src-mediated transformation, maintenance of the transformed phenotype does not require continuous activation of this kinase pathway. Our results agree with other recent work showing that rodent fibroblasts transformed by v-Src as well as mutationally activated c-Ha-Ras do not demonstrate constitutive ERK activation (59). Unlike v-Src, transformation by both v-Fps and Myr-Fes does correlate with constitutive ERK activation, suggesting that this pathway is important to the maintenance of transformation by Fps/Fes tyrosine kinases. In contrast to the ERK pathway, constitutive activation of JNK was observed with v-src and both fps/fes oncogenes, suggesting that JNK activation may play a general role in tyrosine kinasemediated transformation.
A second major difference between v-Src-and Fps/Fes-mediated transformation may involve the signaling pathway leading from the tyrosine kinase to JNK activation downstream. While dominant-negative Ras, Rac, and Cdc42 all inhibited JNK activation by both v-Fps and Myr-Fes, only dominantnegative Rac and Cdc42 blocked JNK activation by v-Src; N17 Ras was without effect (Fig. 8). This finding suggests that the major pathway from v-Src to Rho family GTPases does not require Ras. One possible mechanism linking v-Src with Rho family small G protein activation may involve direct phosphorylation of a guanine-nucleotide exchange factor. A precedent for this idea is provided by the recent observation that Vav guanine nucleotide exchange activity toward Rac and Cdc42 requires tyrosine phosphorylation by the Src family tyrosine kinase, Lck (15,16). Although Vav is restricted in its expression to hematopoietic cells (60), a related protein may exist in other cell types that is a direct target for v-Src.
Although our results show a strong correlation of Ras/ERK signaling and Fps/Fes-mediated fibroblast transformation, activation of this pathway in other cell types may contribute to Fps/Fes-mediated differentiation. Several previous studies have implicated Fps/Fes kinases in the regulation of hematopoietic differentiation. For example, expression of v-Fps is suf- ficient to induce differentiation of chicken myeloid stem cells into macrophages in the absence of colony-stimulating factor-1 (61). Similarly, introduction of c-Fes into the leukemia cell line K-562 induces growth suppression and terminal myeloid differentiation (29). Interestingly, recent work with K-562 cells shows that activation of the ERK pathway may be sufficient to induce terminal differentiation as well. In one study, constitutively active MEK mutants were shown to be sufficient to promote megakaryocytic differentiation in this cell line (62). These results suggest that MEK may be an important downstream effector for Fes-induced differentiation in a hematopoietic cell context.
Other work points to the Ras/Raf/ERK pathway as a general regulator of hematopoietic differentiation. Induction of monocytic differentiation of human myeloid leukemia cells has been correlated with activation of Raf-1 and ERK activities (63). Furthermore, introduction of a constitutively activated mutant of Ras induced monocytic differentiation of the human monoblastic cell line, U-937 (64), while dominant-negative Ras was shown to block phorbol-ester-induced differentiation of HL-60 cells into macrophages (65). These data strongly implicate the Ras/ERK pathway in some hematopoietic differentiation events, and identify a possible downstream signaling pathway for Fps/Fes-mediated myeloid differentiation. Although these studies suggest that the Ras/ERK pathway may contribute to myeloid differentiation, the possible role of Rac/Cdc42/JNK signaling in this process is less clear. Signal transduction by several hematopoietic cytokines has recently been linked to the activation of JNK (41). Interestingly, two of these cytokines (erythropoietin and interleukin-3) have been previously shown to activate c-Fes (24,27). Thus, Fes may provide one connection between the cytokine receptor and activation of JNK downstream. Whether or not JNK contributes to the differentiation response will require further investigation.