Critical roles of c-Kit tyrosine residues 567 and 719 in stem cell factor–induced chemotaxis: contribution of src family kinase and PI3-kinase on calcium mobilization and cell migration

c-Kit is a receptor tyrosine kinase (RTK), which constitutes a type III RTK subfamily with the receptors for platelet-derived growth factor (PDGF), colony-stimulating factor 1 (CSF-1), and flt-3 ligand.1,2 The type III RTKs are characterized by an extracellular domain with 5 immunoglobulinlike domains and a cytoplasmic domain consisting of a kinase domain that is interrupted by a kinase insert. c-Kit (KIT) and its ligand stem cell factor (SCF) play an important role in hematopoiesis, melanogenesis, and gametogenesis,3 as has been clearly shown by loss of function mutations of c-kit gene. In addition, c-kit gene product has been associated with various forms of neoplasms. Activating mutants of KIT, either in the juxtamembrane domain or the catalytic domain, were identified as the cause for transformation of hematopoietic stem cells, mast cells, and gastrointestinal stromal cells.4-10 Thus, KIT/SCF has pleiotropic functions such as proliferation, survival, differentiation, and transformation. In this report, we focus on SCF/KIT-mediated cell migration, which is also a characteristic function of SCF in hematopoietic stem cells and mast cells,11-13 and has critical roles in immunity, metastasis, and development.

On ligand stimulation, KIT receptors dimerize, activate its intrinsic tyrosine kinase, and autophosphorylate. The phosphorylated KIT receptor generates binding sites for SH2 domain-containing proteins, which include proteins of the p21Ras–mitogen-activated protein kinase (MAPK) pathway,14 the p85 subunit of phosphatidylinositol 3′ kinase (PI3K),15 phospholipase C-γ₁, the Grb2 adaptor protein,16 the Src family kinases (SFKs),17 Cbl, CRKL,18p62Dok-1,19 SHP1, and SHP2.20 Those proteins are subsequently activated or phosphorylated and further transduce signaling cascades that lead to various cellular responses. However, little is known about which signaling is essential for SCF-mediated migration. Recently, a few reports indicated that Lyn or p38 MAPK plays an important role,21,22 but no comprehensive investigation has been done in which the tyrosine residue of KIT is involved in signal transduction, which is required for cell migration. In this study, we have converted all possible tyrosine (Y) residues on KIT cytoplasmic domain to phenylalanine (F) and introduced these YF substitute mutants on 293T cells or murine interleukin 3 (IL-3)–dependent BAF3 cells. We used these cell lines to elucidate signaling cascades that are important for SCF-mediated cell migration. We found the critical roles of Y567 and Y719 in KIT and defined 2 distinct pathways, SFK and PI3K, initiating from respective tyrosine residues.

Highly purified recombinant murine (rm) SCF and rmIL-3 were provided by Kirin Brewery (Tokyo, Japan). Antiphosphotyrosine antibody, a murine monoclonal antibody generated against phosphotyrosine, was generously supplied by Dr B. J. Drucker (Oregon Health Science University, Portland). Antibodies against mouse KIT, Lyn, and C-terminal Src kinase (Csk) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA) and anti-MEK1/2, antiphospho MEK1/2, anti-p38 MAPK, antiphospho p38 MAPK, anti-Akt, antiphospho Akt (pSer473), and antiphospho Src (pY416) were purchased from New England Biolabs (Beverly, MA); the anti-PI3K p85α antibody from Upstate Biotechnology (Lake Placid, NY); PI3K inhibitor, wortmannin and LY294002, p38 MAPK inhibitor, SB203850, and MEK inhibitor, U0126 from Calbiochem (San Diego, CA); BAPTA-AM and Indo 1-AM from Dojin Chemical (Kumamoto, Japan); and G418 sulfate (geneticin) from Gibco BRL (Grand Island, NY).

The mammalian expression vector pEF-BOS was donated by Dr S. Nagata (Osaka University, Osaka, Japan). A dominant-negative mutant of bovine p85 complementary DNA (cDNA), which lacks a binding site for the catalytic 110-kd subunit of PI3K, was kindly provided by Dr M. Kasuga (Kobe University, Kobe, Japan). The gene encoding murine c-kit WT cDNA was cloned into an EcoRV site of Bluescript I KS (−). BluntedHindIII-EcoRI fragment of Bluescript I KS (−) containing c-kit WT cDNA was introduced into the blunted XbaI site of pEF-BOS. To convert tyrosine residues (at codons 544, 546, 552, 567, 569, 577, 608, 645, 671, 674, 702, 719, 728, 745, 772, 821, 844, 853, 868, 878, 898, and 934) to phenylalanine in the cytoplasmic domain of KIT, point mutations were generated by overlap extension polymerase chain reaction with synthetic oligonucleotides encoding the desired amino acid substitutions using c-kit WT as a template. The digested fragments of amplified products were exchanged by the corresponding fragment of c-kit WT cDNA in pEF-BOS. The resulting plasmids were sequenced to confirm the mutations and named as Y544F, Y546F, Y552F, and so on. The cDNAs of rat WT-Csk, a kinase-defective csk[csk(−)] in which Lys222 is replaced to arginine, and a membrane-anchoring csk mutant (mcsk) possessing c-src myristoylation signal sequence, were subcloned into an expression vector, pCXN2 harboring neomycin-resistant gene.

The IL-3–dependent murine pro-B cell line, BAF3, was maintained in RPMI 1640 medium supplemented with 10% fetal calf serum (FCS) and 1 ng/mL rmIL-3. Transfection of BAF3 cells was performed by electroporation. Following gene transfer, cells were cultured in IL-3–containing medium for 24 hours and then selected in G418 at a concentration of 1 mg/mL. Multiple clones were expanded for further analysis. Human embryonic kidney 293T cell was transfected by the calcium phosphate method as described previouly.23 Two days after transfection, the cells were used for further analysis. To detect KIT proteins on the cell surface, cells were treated for 30 minutes at 4°C with biotin-conjugated rat antimurine KIT antibody (Immunotech, Marseille, France). After washing, cells were incubated for 30 minutes at 4°C with a 1:40 dilution of avidin–fluorescein isothiocyanate (FITC; Becton Dickinson, San Jose, CA). Samples were analyzed using a FACSCalibur (Becton Dickinson).

BAF3 cells transfected with c-kit constructs (BAF3/KIT) were starved from IL-3 and serum for 12 hours in RPMI 1640 medium supplemented with 0.1% FCS. Subsequently, cells were resuspended in 1 mL medium for 5 minutes at 37°C with or without 100 ng/mL SCF, washed twice with ice cold phosphate-buffered saline (PBS), and lysed with buffer containing 50 mM HEPES (pH 7.4), 10% glycerol, 150 mM NaCl, 1% Triton X-100, 1 mM EDTA, 1 mM EGTA, 50 μM ZnCl₂, 25 mM NaF, proteinase inhibitors (Complete, Boehringer Mannheim, Mannheim, Germany), 1 μM pepstatin, and 1 mM sodium orthovanadate. Cell lysates were clarified at 20 000g for 20 minutes. For immunoprecipitation, cell lysates were incubated with rabbit polyclonal antibody to murine KIT and with Protein A/G-Plus-Sepharose (Santa Cruz). The immunoprecipitates were washed 3 times with lysis buffer. Immunoprecipitates and total lysates were resuspended in sodium dodecyl sulfate (SDS) sample buffer, heated and separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE). Gels were blotted on Immobilon P membrane (Millipore, Bedford, MA) and stained with the indicated antibody. Antibody binding was detected by incubation with a horseradish peroxidase (HRP)–labeled secondary antibody followed by chemiluminescence detection (ECL-Plus, Amersham Pharmacia Biotech, Buckinghamshire, United Kingdom).

The modified Boyden chamber assay in a 24-well chemotaxis chamber (5-μm pore size, Kurabow, Osaka, Japan) was used to measure SCF-induced cell migration. The upper and lower chambers were filled with RPMI 1640 medium supplemented with 10% FCS and various concentrations of SCF. A total of 3 × 10⁵ IL-3–starved BAF3/KIT cells were plated in each of the upper chamber. Chambers were incubated for 6 hours at 37°C in a CO₂ incubator. Cells migrating into the lower chamber were counted in the flow cytometry by passing each sample at the same predetermined time and flow condition as previously described.11 

The measurements of intracellular Ca²⁺ was as done as described previously.24 In short, cells were washed twice in Hepes-buffered Krebs solution, which consisted of 124 mM NaCl, 5.4 mM KCl, 0.8 mM MgCl₂, 1.8 mM CaCl₂, 10 mM glucose, and 25 mM Hepes (pH 7.4; HBKS). Then, a total of 10⁷ cells were incubated in 1 mL HBKS containing 5 μM Indo 1-AM for 30 minutes at 37°C in the dark. The cells were subsequently washed twice with HBKS and resuspended at 2.5 × 10⁶ cells/mL. The cell suspension was plated in a continuously stirred cuvette at 37°C in a CAF-110 fluorometer (Jasco, Tokyo, Japan). Fluorescence was monitored at an excitation wavelength of 355 nm and emission wavelengths of 405 nm and 485 nm. The data are presented as the relative ratio of fluorescein excited at 405 to 485 nm.

Comparisons between other groups were made using an unpaired Student t test (2-tails).

The chemotactic property of BAF3/KIT^WT (BAF3/WT) against SCF is presented as a checkerboard diagram (Table1). Maximal response was obtained with a concentration of 30 ng/mL SCF in the lower chamber. Because Ca²⁺ is the key mediator of cell movement, we measured SCF-induced intracellular change of Ca²⁺ in BAF3/WT cells using Indo1-AM. Following SCF stimulation, BAF3/WT cells exhibited a prompt peak and subsequently sustained increase of Ca²⁺. The Ca²⁺ influx was dependent on SCF concentration within a range from 10 to 100 ng/mL. This Ca²⁺ influx peak was gradual as compared with stromal cell-derived factor–1α (SDF-1α) (Figure 1). Then, we examined the effect of the intracellular Ca²⁺ chelator BAPTA-AM on SCF-induced Ca²⁺ mobilization and cell migration. The pretreatment of cells with BAPTA-AM also inhibited SCF-induced cell migration in a dose-dependent manner (Figure2A). The pretreatment of cells with BAPTA-AM dose dependently attenuated SCF-induced intracellular Ca²⁺ mobilization as shown in Figure 2B.

As shown in Figure 3A, KIT contains 22 tyrosine (Y) residues in the cytoplasmic domain. To determine which tyrosine residue(s) is involved in Ca²⁺ mobilization and cell migration, we constructed a series of 22 single YF mutants for all these tyrosine residues. On transient transfection into 293T cells, Y567F, Y569F, and Y719F, respectively, showed 73%, 60%, and 54% reduction of Ca²⁺ mobilization when compared with WT (Figure 3B). To further examine the effects of these YF conversions in stable cell clones, we transfected these constructs into murine IL-3–dependent BAF3 cells that had no intrinsic KIT expression. The surface expressions and ligand-induced autophosphorylations of KIT were observed at the same level in BAF3 clones expressing WT or each mutant (data not shown). Similar to the results of 293T cells, Ca²⁺ influx was suppressed in BAF3 expressing Y567F (BAF3/Y567F), BAF3/Y569F and BAF3/Y719F (data not shown). SCF-induced cell migration was decreased for 85%, 70%, and 58% in BAF3/Y567F, BAF3/Y569F, and BAF3/Y719F mutants, respectively, as compared to BAF3/WT, which were similar to the reduction rate of Ca²⁺mobilization (Figure 3C).

Consistent with previous reports that indicated the close relationship between Y567 and SFK, we found that in 293T cells expressing Y567F mutant (293T/Y567F), SCF-induced Lyn activation was remarkably reduced as compared to 293T/WT (Figure4A). To determine if SFK activity is involved in SCF-stimulated Ca²⁺ mobilization and cell migration, SFK activity was inhibited specifically by expression of Csk, which is the negative regulator of SFK. We transfected WT-Csk, Csk(−) (kinase dead), or mCsk (constitutive active) into BAF3 cells and 293T cells, with c-kit WT. Csk proteins were highly expressed as shown in Figure 4B. In both cell lines, WT-Csk inhibited SCF-induced Ca²⁺ influx significantly. The suppression of Ca²⁺ influx was more prominent in mCsk transfectants. In contrast, expression of Csk(−) did not suppress Ca²⁺influx, indicating that Csk activity plays an important role in Ca²⁺ influx in the signaling cascades mediated by c-Kit (Figure 4C). We next examined the effect of WT-Csk, Csk(−), and mCsk on cell migration by using stably transfected BAF3 cells. WT-Csk inhibited cell migration and mCsk inhibited more profoundly than WT-Csk (Figure 4D). These findings suggested that SFK may regulate SCF-induced Ca²⁺ influx and cell migration.

We next explored how Csk affects SCF-mediated signal transduction, thereby leading to cell migration. As shown in Figure5A, in any Csk transfectant there was no difference in the activation of Akt. On the other hand, Csk and mCsk inhibited SCF-mediated activation of p38 MAPK and MEK1/2 (Figure 5A). SCF-mediated activations of MEK1/2 and p38 MAPK were also significantly reduced in BAF3/Y567F, where SCF-induced SFK activation was suppressed (Figure 5B). Moreover, MEK1/2 inhibitor U0126 and p38 MAPK inhibitor SB203580 suppressed cell migration dose dependently as shown in Figure6. Both U0126 and SB203580 did not affect cell viability at these concentrations. These results indicate that Y567 of KIT mediates MEK/Erk and p38 MAPK signaling through SFK and regulates SCF-induced cell migration.

We next examined the relationship between Ca²⁺mobilization, Erk, and p38 MAPK. After treatment of BAF3/WT cells with U0126 or SB203580, Ca²⁺ influx on SCF stimulation was measured. As shown in Figure 7A, p38 MAPK inhibitor SB203580 suppressed Ca²⁺ influx, but MEK inhibitor U0126 did not. Furthermore, we examined the effect of the intracellular Ca²⁺ chelator, BAPTA-AM on SCF-induced activation of MEK1/2 and p38 MAPK. As shown in Figure 7B, treatment of the cells with BAPTA-AM remarkably reduced the activation of MEK1/2, but did not suppress the activation of p38 MAPK. These data indicated that in SCF signaling, SFK-activated P38 MAPK may induce Ca²⁺ mobilization, resulting in the activation of MEK/ Erk pathway.

Y719F abolished the binding of PI3K to KIT (Figure8A). In BAF3/Y719F, Ca²⁺influx and cell migration were impaired about 50% as compared with WT. To explore the involvement of PI3K in SCF-induced cell migration, we tested specific PI3K inhibitors, wortmannin and LY294002, both of which inhibited cell migration in a concentration-dependent manner (Figure8B). Furthermore, we introduced an expression vector of dominant-negative PI3K or an empty control vector into BAF3/WT cells. Dominant-negative PI3K expression inhibited SCF-induced PI3K activity by about 85% (data not shown). BAF3/WT cell expression of dominant-negative PI3K inhibited significantly Ca²⁺ influx and cell migration in BAF3/WT cells (Figure 8C,D).

To examine the interaction between SFK and PI3K, we analyzed the migration of BAF3/WT expressing mCsk (BAF3/WT/mCsk) in the presence of the PI3K inhibitor LY294002. The migration of BAF3/WT/mCsk was reduced to about 50% of the BAF3/WT cells. When BAF3/WT/mCsk cells were treated with LY294002, the migration response was further inhibited and reduced to about 15% of BAF3/WT (Figure9). These results suggested that both SFK and PI3K cooperate for SCF-mediated cell migration.

Migration is one of the unique and important cell functions undertaken by SCF/KIT system. Much effort has been done to clarify signal transduction leading to SCF/KIT-mediated proliferation and survival, but the signaling mechanism in SCF-mediated cell migration has not been clarified yet. In this study, we determined several critical signaling molecules in SCF-induced cell migration, using a series of 22 YF mutants of KIT.

First, we found that intracellular Ca²⁺ mobilization plays a pivotal role in SCF-mediated cell migration. It has been reported that Ca²⁺ mobilization plays a critical role in cell migration of keratinocytes and neutrophils.25,26SCF-induced Ca²⁺ mobilization was not significantly impaired by most single YF substitution mutants, but it was remarkably abrogated by Y567F and moderately by Y569F and Y719F. In Y567F, Y569F, and Y719F mutants, cell motilities were also inhibited similarly, corresponding to the reduction levels of Ca²⁺ mobilization. Moreover, Ca²⁺ chelator completely abolished the migratory activity. These results indicated the importance of Ca²⁺mobilization in SCF-mediated cell migration, which are largely mediated by Y567, Y569, and Y719 of KIT.

Second, we found that Y567/SFK play an important role in SCF-mediated Ca²⁺ mobilization and cell migration. The previous study showed that Y567 of KIT is the binding site of SFK.27-29It was also shown that SFK plays an important role in PDGF-induced cell migration.30,31 Therefore, we investigated the roles of SFK on SCF/KIT-mediated Ca²⁺ mobilization and cell migration. In disagreement with previous reports,27-29 we could not find the direct binding of Lyn or Fyn with KIT after SCF stimulation in either BAF3/KIT cells or 293T/KIT cells (data not shown). However we could show the activation of SFK, Lyn, in 293T/KIT cells after SCF stimulation; this ligand-dependent activation of Lyn was remarkably reduced by the Y567F substitution, indicating that Y567 is critical for SFK activation. Although previous reports determined the importance of SFK by so-called specific SFK inhibitors PP1 or PP2, they might not be suitable for analyzing KIT signaling, because these inhibitors abolished ligand-dependent KIT autophosphorylation and kinase activity at the concentration of 10 to 30 μM in BAF3 cells (data not shown). This observation was supported by the finding that PDGF receptor activation was inhibited by PP1.32Therefore, we further investigated the role of SFK in cells overexpressing Csk. In the previous study, we showed that Csk and mCsk suppressed the activation of SFKs specifically.33,34 In this study, overexpression of Csk or mCsk inhibited SCF-induced Ca²⁺ mobilization and cell migration in BAF3/KIT cells. These results strongly suggested that SFK is a critical signaling pathway for SCF-mediated cell migration and Ca²⁺mobilization. In Y567F cells, the activations of p38 MAPK and Erk1/2 were significantly impaired. Expressions of Csk or mCsk also inhibited the activation of p38 MAPK and Erk1/2, suggesting the involvement of SFK in their activations. Furthermore, p38 MAPK inhibitor, SB203580 or MEK1/2 inhibitor, U0126, suppressed cell migration significantly. These data suggested that SFK regulates cell migration through p38 MAPK and Erk1/2. Regarding the relationship between p38 MAPK and Ca²⁺ mobilization, SB203580 impaired SCF-induced Ca²⁺ influx, but intracellular Ca²⁺ chelator BAPTA-AM did not suppress p38 MAPK activation. On the other hand, U0126 did not impair Ca²⁺ influx, but Ca²⁺ chelator suppressed MEK1/2 activation. These data indicate that p38 MAPK induces Ca²⁺ mobilization, thereby resulting in Erk1/2 activation. Previous reports showed that Ca²⁺ influx activates p38 MAPK and Erk1/2 in human neutrophils or PC12 cells.35,36Recently, it was reported that Y568 of human KIT (corresponding to murine Y567) is important of SFK activation and Shc phosphorylation resulting in Ras/Erk2 activation,37 and p38 MAPK was involved in SCF-mediated cell migration.22 In this report, for the first time, we showed that not only p38 MAPK but also Erk1/2 was involved in SCF-mediated cell migration, and that both MAPKs differentially transduce the signaling from Y567-induced SFK activation to Ca²⁺ influx and cell migration.

Finally, we showed the Y719/PI3K pathway also regulates SCF-mediated cell migration. Consistent with the previous report,15 we confirmed that Y719 was the binding site for PI3K. In PDGF-induced chemotaxis, PI3K is regarded as a major signaling pathway of Ca²⁺ influx leading to cell migration.38-40 In our study, Y719F mutant exhibited less effective inhibition of Ca²⁺ influx and cell migration than Y567F, suggesting that SFK has a more essential role than PI3K. The importance of PI3K was confirmed by the inhibition studies with chemical inhibitors and dominant negative PI3K. BecauseY719F did not show the impaired activation of p38 MAPK or Erk1/2, which were suppressed by the inhibition of SFK, PI3K may act by a different pathway from SFK, for the cell migration followed by Ca²⁺ mobilization. Furthermore, the PI3K inhibitor and Csk showed synergistic suppression of cell migration, suggesting the cooperation of SFK and PI3K in SCF-mediated migration.

In summary, this study has provided evidence that SFK and PI3K cooperatively contribute to SCF-mediated cell migration through Ca²⁺ mobilization, and the signaling of SFK is transduced sequentially from p38 MAPK, Ca²⁺ influx, and then to Erk1/2. Many kinds of growth factors can activate SFK, Erk1/2, p38 MAPK, or PI3K, but this does not always lead to the cell motility. Therefore, the coordination of this signaling by scaffolding proteins could be important for SCF-mediated chemotaxis. Future studies focused on this point will further clarify the signaling mechanism of SCF/KIT-mediated cell migration.

We thank Dr Koji Hashimoto for great help of making YF substitution mutants and their expression vectors, and we thank Noriko Kikunaga for her excellent experimental assistance.