Sphingosine 1-Phosphate Stimulates Tyrosine Phosphorylation of Crk*

The proto-oncogene molecule c-Crk plays a role in growth factor-induced activation of Ras. Sphingosine 1-phosphate (SPP), a metabolite of cellular sphingolipids, has previously been shown to play a role in growth factor receptor signaling (Olivera, A., and Spiegel, S. (1993) Nature 365, 557–560). SPP was found to strongly induce tyrosine phosphorylation of Crk, but not Shc, in NIH-3T3 parental, insulin-like growth factor-I receptor-overexpressing and Crk-overexpressing (3T3-Crk) fibroblasts. Sphingosine, a metabolic precursor of SPP, also produced a slight increase in tyrosine phosphorylation of Crk. In contrast, other sphingolipid metabolites including ceramide did not alter Crk tyrosine phosphorylation. Furthermore, Crk enhanced SPP-induced mitogenesis, as measured by SPP-stimulated [3H]thymidine incorporation in a manner proportional to the level of Crk expression in 3T3-Crk cells. This stimulation appears to be Ras-dependent, whereas SPP stimulation of MAP kinase activity is Ras-independent. These data indicate that SPP activates a tyrosine kinase that phosphorylates Crk and that Crk is a positive effector of SPP-induced mitogenesis.

Increasing evidence suggests that the sphingolipid ceramide and its metabolites sphingosine and sphingosine 1-phosphate (SPP) 1 represent a new class of intracellular second messengers that mediate a variety of cellular functions (1)(2)(3)(4)(5). Sphingosine and SPP have been shown to induce mitogenesis in a wide range of cell types (5)(6)(7). Platelet-derived growth factor (PDGF), a potent mitogen, increases cellular levels of sphingosine and SPP (8,9). Moreover, inhibition of the PDGF-induced increase in SPP levels markedly decreased PDGF-induced cellular proliferation (8).
The downstream signaling pathways utilized by these sphingolipids have not been fully elucidated. Sphingosine and SPP are known to increase intracellular calcium (6, 10 -12) and phosphatidic acid levels (13,14) and decrease cAMP levels (6,15). SPP has also been shown to stimulate the Raf/MEK/MAP kinase signaling pathway (16). In an attempt to understand the molecular mechanisms underlying sphingosine and SPP-induced mitogenesis, we investigated the effect of these sphingolipids on intracellular signaling molecules upstream of the MAP kinase cascade, and then we examined the role of the MAP kinase signaling pathway.
It has previously been shown that activation of either Shc-or Crk-related pathways leads to activation of the MAP kinase cascade. Shc is an SH2 domain-containing protein that becomes tyrosine-phosphorylated and associates with Grb2 in response to growth factor receptor stimulation (17)(18)(19). Signaling through Shc appears to be a common pathway by which both tyrosine kinase growth factor receptors and certain G-protein-coupled receptors lead to activation of Ras (20). Crk is a noncatalytic SH2 and SH3 domain-containing adaptor molecule that shares structural homology with Grb2 and Nck (21)(22)(23). Like Grb2, Crk associates with the guanine nucleotide exchange factor mSos and a related molecule called C3G (23,24). Thus, Crk provides an alternate pathway by which growth factor receptors can signal Ras. Crk has been specifically implicated in IGF-I receptor signaling (25). Stimulation of the IGF-I receptor induces tyrosine phosphorylation of Crk (25), and the mitogenic effects of IGF-I are enhanced in Crk-overexpressing cells (26). To explore possible interactions of the sphingolipid and Crk signaling pathways, we have investigated the effects of sphingolipid metabolites on Crk and Shc signaling.
Cell Culture-Both parental and transfected NIH-3T3 mouse fibroblasts were cultured in Dulbecco's modified Eagle's medium (Biofluids, Inc., Rockville, MD) supplemented with 10% fetal bovine serum (Upstate Biotechnology Inc., Lake Placid, NY). 3T3-Crk and 3T3-Neo cells were, respectively, stably transfected with either pCXN2-CRK II, a Crk expression vector driven by the cytomegalovirus promoter and carrying neo r (kindly provided by Dr. M. Matsuda, Tokyo, Japan) or the pCXN2 vector from which the Crk cDNA had been excised, as described previously (26). Before stimulation, subconfluent cultures of cells in 100-mm dishes were switched to serum-free DMEM supplemented with 0.1% insulin-free bovine serum albumin (Intergen Co., Purchase, NY) and 20 mM HEPES (pH 7.5) for 18 h. Cells were treated with IGF-I (100 nM), SPP (10 M), sphingosine (20 M), DHS (20 M), C 2 -Cer (20 M), C 6 -Cer * This work was supported in part by National Institutes of Health Grants GM43880 and CA61774 (to S. S.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 1 The abbreviations used are: SPP, sphingosine 1-phosphate; PDGF, platelet-derived growth factor; MAP, mitogen-activated protein; MEK, mitogen-activated protein kinase kinase; IGF-I, insulin-like growth factor-I; C 2 -Cer, N-acetylsphingosine; C 6 -Cer, N-hexanoylsphingosine; DHS, DL-threo-dihydrosphingosine; DMEM, Dulbecco's modified Eagle's medium; PAGE, polyacrylamide gel electrophoresis; Erk, extracellular signal-regulated kinase; SH2 and SH3, Src homology regions 2 and 3, respectively; G-protein, guanine nucleotide-binding protein; Sph, sphingosine; DN-Ras, dominant negative Ras. NIH-3T3 cells overexpressing approximately 10-fold levels of Shc with a c-Myc tag (Shc/Myc) and control cells transfected with selectable marker vector only (pLEN) were kindly provided by Dr. Ben Margolis (University of Michigan, Ann Arbor, MI). These cell lines were cultured similarly to Crk-overexpressing NIH-3T3 cells.
Transient transfection of NIH-3T3 cells overexpressing 10-fold Crk (3T3-Crk7 cells) with the plasmid-encoding dominant negative Ras or a neo r vector alone was performed using the Dosper reagent (Boehringer Mannheim). Transfected cells were allowed to recover 24 h in complete medium before plating for thymidine incorporation assays or before serum starvation for measurement of MAP kinase activation.
Immunoprecipitations-After treatment with growth factors, cells were washed twice with ice-cold phosphate-buffered saline and harvested in a lysis buffer containing 50 mM HEPES (pH 7.4), 2 mM sodium orthovanadate (Na 3 VO 4 ) , 100 mM NaCl, 4 mM sodium pyrophosphate, 10 mM EDTA, 10 mM NaF, 1 mM phenylmethylsulfonyl fluoride, 2 g/ml leupeptin, 2 g/ml aprotinin, and 1% Triton X-100. Lysates were incubated for 1 h at 4°C then centrifuged at 10,000 ϫ g for 30 min at 4°C to remove Triton-insoluble material. Protein content of the lysates was determined by the Bio-Rad method. 600 g of protein from each dish was immunoprecipitated overnight at 4°C with either a monoclonal antibody to Crk (3 g, Transduction Laboratories, Lexington, KY), a polyclonal antibody to Shc (3 g, Transduction Laboratories), or polyclonal antiserum to C3G (10 l, kindly provided by Dr. Beatrice Knudsen, Rockefeller University, New York, NY) followed by adsorption to 50 l of 10% Protein A-Sepharose beads (Pharmacia Biotech Inc.) for 5 h at 4°C. Immunoprecipitates were washed three times with ice-cold immunoprecipitation buffer containing 10 mM Tris (pH 7.4), 150 mM NaCl, 0.2 mM sodium orthovanadate, 1 mM EDTA, 1 mM EGTA, 0.2 mM phenylmethylsulfonyl fluoride, 1% Triton X-100, and 0.5% Nonidet P-40. The entire immunoprecipitated samples were then boiled for 2 min in sample buffer containing 50 mM Tris (pH 6.7), 2% SDS, 2% ␤-mercaptoethanol, and bromphenol blue as a marker. Samples were then run on 9% SDS-polyacrylamide gel electrophoresis (SDS-PAGE) gels and transferred to nitrocellulose membranes using standard electrophoresis and electroblotting procedures.
Immunoblotting-Nitrocellulose membranes were blocked with either 3% insulin-free bovine serum albumin (for phosphotyrosine immunoblotting) or 3% nonfat dry milk (for other immunoblotting) in phosphatebuffered saline with Tween containing 10 mM sodium phosphate (pH 7.2), 140 mM NaCl, and 0.1% Tween 20. Blots were then immunolabeled overnight at 4°C for either phosphotyrosine (in antibody RC2OH 1:2500), Crk (1:1000), or Grb2 (1:1000) (all from Transduction Laboratories). Immunolabeling was detected by enhanced chemiluminescence (ECL, Amersham Life Science, Inc.) according to manufacturer conditions. Some blots were stripped and re-probed with another antibody. Blots were stripped by incubation for 1 h at 50°C in a solution containing 62.5 mM Tris-HCl (pH 6.7), 2% SDS, and 0.7% ␤-mercaptoethanol. Blots were then washed for 1 h in several changes of phosphate-buffered saline with Tween at room temperature and probed with ECL to confirm that antibodies had been completely removed. Blots were then reblocked and immunolabeled as described above. Erk1 and Erk2 in cleared cell lysates derived from NIH-3T3 cells expressing IGF-I receptors and resolved by SDS-PAGE were detected by immunoblotting with an anti-Erk1 antibody that recognizes both Erk1 and Erk2 (C16) (1:1000, Santa Cruz Biotechnology, Inc., Santa Cruz, CA). Ras expression in 3T3-Crk7 cells transfected with the neo r vector or the vector encoding dominant negative Ras was similarly detected by immunoblotting using an antibody to Ha-Ras (1:1000, Transduction Laboratories).
Thymidine Incorporation Assays-Subconfluent cell monolayers in 24-well plates were rendered quiescent in DMEM with 1% fetal bovine serum for 6 h. Cells were then incubated for 18 h in DMEM with 1% fetal bovine serum containing various concentrations of SPP. [ 3 H]Thymidine was added to a final concentration of 1 Ci/ml, and the cells were incubated for an additional hour. The cells were rinsed twice with ice-cold phosphate-buffered saline, twice with ice-cold 5% trichloroacetic acid, and twice with ice-cold 95% ethanol. The precipitates were solubilized in 0.2 ml of 1 N NaOH and neutralized with 0.2 ml of 1 N HCl, and solubilized samples were counted by liquid scintillation.
MAP Kinase Assays-MAP kinase activity using myelin basic protein as a substrate was determined from 3T3-Crk7 cells and 3T3-Crk7 cells expressing dominant negative Ras following serum starvation for 18 h and subsequent treatment with either vehicle or SPP. The cells were lysed by the addition of 0.5 ml of lysis buffer (1% Triton X-100 containing 50 mM HEPES (pH 7.9), 100 mM NaCl, 4 mM sodium pyrophosphate, 10 mM NaF, 10 mM EDTA, 2 mM Na 3 VO 4 , 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin and 10 g/ml aprotinin). Lysates were centrifuged for 10 min at 4°C, and equal amounts of protein were immunoprecipitated using 1 g of polyclonal antibody to Erk2 (C14, Santa Cruz) followed by 20 l of Protein A/G-Sepharose beads (Santa Cruz). The beads were pelleted and washed three times in lysis buffer and twice in kinase buffer (20 mM HEPES (pH 7.4), 10 mM MgCl 2 , and 1 mM dithiothreitol). The kinase assay was begun by resuspending the beads in 20 l of kinase buffer containing 50 M ATP, 2 g of myelin basic protein, and 2 Ci of [␥-32 P]ATP and incubating for 30 min at 37°C. Laemmli sample buffer was added, and samples were boiled and separated on 12% SDS-PAGE. Radioactive myelin basic protein was visualized by autoradiography and quantitated by scintillation counting.

RESULTS AND DISCUSSION
The sphingolipid ceramide is known to activate the MAP kinase pathway by stimulating phosphorylation of Raf (27). Another sphingolipid metabolite, SPP, is also known to play a role in growth factor signaling and leads to activation of the MAP kinase cascade (5,16). However, the precise biochemical mechanisms involved remain unknown. To further our understanding of these intracellular pathways, we investigated the effect of various sphingolipid metabolites on the tyrosine phosphorylation state of two signaling proteins, Shc and Crk. Shc is an SH2 domain-containing protein that is involved in activation of Ras by both growth factor receptor tyrosine kinases and G i G-protein-linked receptors (20). Tyrosine phosphorylation of Shc induces its association with the adaptor protein, Grb2. Grb2, in turn, associates with the guanine nucleotide exchange protein mSos, which leads to the activation of Ras (17)(18)(19).
NIH-3T3 cells stably transfected with the IGF-I receptor (clone NWTb3) (28) were treated for 5 min with either IGF-I, sphingosine, SPP, or dihydrosphingosine, a mitogenically inactive analog of sphingosine. In this cell line, IGF-I strongly induces tyrosine phosphorylation of Shc and thus can be used as a positive control to test for SPP-induced Shc phosphorylation. Cleared whole cell lysates were immunoprecipitated for Shc, separated by SDS-PAGE, and blotted for either phosphotyrosine or Grb2. In Fig. 1, it can be seen that IGF-I induced tyrosine phosphorylation of Shc and increased its association with Grb2. In contrast, neither sphingosine, SPP, nor DHS altered the tyrosine phosphorylation state of Shc or its association with Grb2. Identical results (i.e. no effect of sphingosine and SPP on Shc tyrosine phosphorylation) were obtained in parental NIH-3T3 cells and in a second fibroblast-derived cell line, Swiss 3T3 cells (data not shown). Thus, at least in mouse fibroblasts, these sphingolipid metabolites do not appear to mediate Ras activation via the Shc-Grb2-mSos pathway.
The SH2 and SH3 domain-containing protein Crk also plays a role in growth factor-induced activation of Ras (23-25, 29, 30). SH3 domain-mediated interactions of the Crk protein with the Ras family guanine nucleotide exchange proteins mSos and C3G represent an alternative pathway to the Shc-Grb2-mSos mechanism of Ras activation (23-25, 29, 30). To determine whether Crk was tyrosine-phosphorylated in response to sphingosine or SPP, NIH-3T3 cells were treated for 5 min with various sphingolipids. Cleared whole cell lysates were immunoprecipitated for Crk and immunoblotted for phosphotyrosine. Stimulation of NIH-3T3 cells with SPP strongly increased the tyrosine phosphorylation of Crk, as shown in Fig. 2A. This suggests that SPP activates a tyrosine kinase that specifically phosphorylates Crk. It was of interest to compare the effects of SPP with those of other biologically active sphingolipid metabolites. Sphingosine, a metabolic precursor of SPP, slightly increased the tyrosine phosphorylation of Crk, whereas DHS, a mitogenically inactive analog of sphingosine, and sphingosylphosphorylcholine, a phosphocholine ester of sphingosine (itself a potent mitogen), had no effect on Crk tyrosine phosphorylation. C 6 -Cer and C 2 -Cer, cell-permeable analogs of ceramide, have been shown to induce differentiation, apoptosis, and cell arrest in several mammalian cell lines (3,4). However, other studies have indicated that ceramide exhibits a bimodal effect on cell growth, stimulating DNA synthesis and proliferation in confluent, quiescent Swiss 3T3 fibroblasts (31, 32) but inhibiting the growth of exponentially proliferating normal fibroblasts (3,33). In contrast to SPP and sphingosine, cell-permeable ceramide analogs (C 2 -Cer and C 6 -Cer) did not alter Crk phosphorylation ( Fig. 2A). Moreover, treatment with exogenous sphingomyelinase, which hydrolyzes sphingomyelin to ceramide, also did not affect the tyrosine phosphorylation of Crk ( Fig. 2A). Fig. 2B shows a time course of the effect of SPP on tyrosine phosphorylation of Crk in NIH-3T3 cells. Cells were stimulated for various times with 10 M SPP, the optimum concentration for mitogenesis (14). Lysates were immunoprecipitated for Crk and immunoblotted for phosphotyrosine. The maximum SPPinduced tyrosine phosphorylation consistently occurred at 10 min. Crk immunoreactivity levels after 20 -30 min were not different from control (data not shown). Thus it appears that after 10 min of exposure to SPP, one or more phosphotyrosine phosphatases act to dephosphorylate Crk.
We also studied the effect of SPP on Crk phosphorylation in Crk-overexpressing cells. The 3T3-Crk7 clone has been shown to express 10-fold higher levels of Crk than NIH-3T3 parental cells (26). 3T3-Crk7 cells were treated for 5 min with either SPP or C 2 -Cer. Lysates were immunoprecipitated for Crk and immunoblotted for phosphotyrosine. As shown in Fig. 3A, SPP, but not ceramide, induced tyrosine phosphorylation of Crk in these cells. This Crk-overexpressing cell line also enabled us to study the effect of SPP on interactions of Crk with the guanine nucleotide exchange protein C3G, which is difficult to visualize in NIH-3T3 parental cells. 2 In these experiments, cleared whole cell lysates were immunoprecipitated with an antibody against C3G and immunoblotted for Crk. In Fig. 3B, it can be seen that treatment of cells with SPP, DHS, or sphingosine for 5 min had no effect on the amount of Crk protein associated with C3G. Identical results were obtained in a second Crkoverexpressing clone, 3T3-Crk9 (data not shown). IGF-I and other growth factors have also been found not to alter Crk association with C3G (data not shown). 3 Thus, alterations in the amount of Crk associated with C3G do not appear to be involved in the mechanism by which SPP and other mitogens mediate signaling through Crk.
We have previously shown that SPP, sphingosine, and ceramide induced mitogenic responses in Swiss 3T3 fibroblasts (6,31). We examined in detail the dose responses of different sphingolipid metabolites (6,8,14,31) that are mitogenic for those cells and used optimal concentrations to ensure that the different responses are not caused by differences in uptake. We have shown in the present study that, whereas SPP induced Crk phosphorylation, treatment with ceramide did not. These results suggest that the mitogenic signaling pathways utilized by SPP and ceramide differ. We have previously shown that SPP stimulates the Raf/MEK/MAP kinase pathway in Swiss 3T3 cells (16). Moreover, in myeloid HL-60 cells, ceramide activated MAP kinase activity (34) and activated a protein kinase that phosphorylates Raf1 on Thr-269, enhancing its activity toward MEK and linking ceramide signaling to the MAP kinase pathway (27). Therefore, it was of interest to examine the activation state of MAP kinases by various sphingolipid metabolites. Whereas IGF-I, SPP, and sphingosine induced a mobility shift in Erk1 and Erk2 consistent with phosphorylation and activation of these MAP kinases (Fig. 4), DHS and C 6 -Cer did not induce a mobility shift (Fig. 4 and data not  shown). Thus, in mouse fibroblasts, SPP and ceramide appear to act through distinct signaling pathways leading to Erk phosphorylation.
To further evaluate the functional effects of Crk overexpression on SPP signaling, the mitogenic effects of SPP were compared in two Crk-overexpressing (3T3-Crk) and two neomycinresistant (3T3-Neo) control clones using a [ 3 H]thymidine incorporation assay. In Fig. 5A it can be seen that SPP produced a dose-dependent increase in thymidine incorporation in both the 3T3-Neo and 3T3-Crk cell types. However, SPP-induced mitogenesis was significantly enhanced in Crk-overexpressing cells com-pared with control 3T3-Neo cells. Furthermore, SPP-induced mitogenesis was dependent on the level of Crk expression in these cell lines. That is, the highest level of SPP-induced mitogenesis was exhibited by 3T3-Crk7, which expresses 10-fold higher Crk levels than parental cells. Intermediate levels were found in 3T3-Crk9, which expresses 3-fold higher Crk levels than parental, and 3T3-Neo cells were least responsive to SPP. Interestingly, in contrast to SPP and sphingosine, which stimulate DNA synthesis by 8-and 4-fold, respectively, C 2 -ceramide and sphingomyelinase were not potent mitogens in Crk-overexpressing cells (Fig. 5B). To determine whether Shc also plays an important role in mitogenesis induced by SPP, we examined the effect of SPP on NIH-3T3 cells overexpressing Shc (Fig. 6). The SPP response in the Shc-overexpressing cell line, Shc/Myc, was similar to that seen in parental NIH-3T3 cells. Thus overexpression of Shc did not enhance the mitogenic response to SPP as did overexpression of Crk.
To determine if SPP-induced Erk1 and Erk2 activation in cells overexpressing Crk was mediated via the Ras/MAP kinase pathway, we measured MAP kinase activity and thymidine incorporation in 3T3-Crk7 cells transiently expressing dominant negative Ras and compared these results with those of 3T3-Crk7 cells transfected with vector alone. Overexpression of Crk by 10-fold (3T3-Crk7 cells) increased the basal levels of MAP kinase activity by 2.3-fold. Furthermore, in NIH-3T3 cells with neo r vector alone, MAP kinase activity was increased 4.3-fold in response to SPP, whereas 3T3-Crk7 cells exhibited only a 1.4-fold increase in MAP kinase activity after SPP stimulation. Expression of dominant negative Ras in 3T3-Crk7 cells did not affect basal MAP kinase activity nor did it block the effect of SPP on MAP kinase activity. This lack of effect is likely due to the high basal activity. The presence of endogenous Ras in 3T3-Crk7 cells and dominant negative Ras in cells transfected with the construct encoding this protein is shown in Fig.  7. These results are consistent with the hypothesis that Crk, at least in part, induces increased MAP kinase activity by a pathway other than the Ras pathway.
We considered the possibility that the increased mitogenesis in response to SPP may not have been mediated entirely by Ras in the 3T3-Crk7 cells. As shown in Fig. 8, the basal level of thymidine incorporation in 3T3-Crk7 cells expressing dominant negative Ras was drastically reduced as compared with 3T3-Crk7 cells transfected with the neo r vector. In addition, there was an attenuation of the SPP stimulation of thymidine incorporation when dominant negative Ras was expressed. Thus, although Crk mediation of SPP-induced MAP kinase activation was shown to be Ras-independent, functional Ras is essential for SPP-stimulated mitogenesis in cells expressing Crk. Results from other investigative groups indicate that Crk may function in Ras-independent as well as Ras-dependent signal transduction pathways, leading to increased mitogenesis. Early reports of downstream activators of Crk showed that Crk complexed with the Ras guanine nucleotide exchange factors mSos and C3G, suggesting that Crk may activate the Ras pathway and, thus, MAP kinase, similar to Grb2 (24,35,36). More recently, it has become clear that stimulation of Crk phosphorylation and concomitant MAP kinase activation is not solely via the Ras/Raf/MEK pathway. Expression of dominant negative Ras did not abrogate enhanced MAP kinase activity, and treatment with the MEK1 inhibitor PD98059 did not reduce the number of colonies in soft agar in NIH-3T3 cells expressing v-Crk (37,38). The finding that v-Crk, an oncogenic protein, could transform cells even in the presence of dominant negative Ras, however, may not predict the ability of c-Crk to be totally independent of the Ras pathway in mediating mitogenic or tumorigenic phenotypes. In fact, in the present report, the results indicate that c-Crk also mediates activation of MAP kinase by pathways other than the Ras pathway, but it further shows that mitogenesis is severely impaired if the Ras pathway is disrupted.
In summary, we show that the sphingolipid metabolite SPP, and to a lesser extent, sphingosine induces tyrosine phosphorylation of the signaling molecule Crk. This occurred in both parental and Crk-overexpressing NIH-3T3 cells as well as in Swiss 3T3 cells (data not shown). In contrast, ceramide, which has been implicated in cell growth, differentiation, and apoptosis, did not induce phosphorylation of Crk. Interestingly, SPP did not induce tyrosine phosphorylation of the SH2 domain-containing signaling molecule Shc. Thus, SPP appears to activate an as yet unidentified tyrosine kinase that specifically phosphorylates Crk. One potential candidate for such a kinase is c-Abl, which is known to phosphorylate Crk (39). In addition, SPP-induced mitogenesis was significantly enhanced in Crkoverexpressing cells compared with the control, and SPP-induced mitogenesis was proportional to the level of Crk expression. SPP-induced mitogenesis occurred in parallel with activation of MAP kinase, but these SPP-stimulated events were variously affected by the expression of a dominant negative Ras. Taken together, these data strongly suggest that Crk is a positive effector of SPP signaling, and furthermore, Crk sends multiple signals in response to upstream stimuli.