Overexpression of C-terminal Src kinase homologous kinase suppresses activation of Lyn tyrosine kinase required for VLA5-mediated Dami cell spreading.

The Csk homologous kinase (Chk), which is co-expressed with C-terminal Src kinase (Csk) in hematopoietic cells, negatively regulates Src family kinases in vitro with selectivity toward Lyn but not c-Src in platelets. To explore the role of Src family kinases in hematopoietic cell adhesion, we overexpressed Chk in the megakaryocytic cell line Dami and established clones exhibiting a 10-fold increase in the amount of Chk. Overexpression of Chk was found to suppress VLA5 integrin-mediated cell spreading, but not cell attachment, throughout fibronectin (FN) stimulation. Deletion and point mutagenesis analyses of Chk showed that this suppression was dependent upon both the SH3 domain, which is responsible for membrane anchoring, and kinase activity. FN-induced cell spreading accompanied a sustained increase in Lyn activity with coincidental kinetics and the activation of Lyn was also suppressed by overexpression of Chk but not a Chk mutant lacking the SH3 domain. Expression of a truncated Lyn mutant lacking the kinase domain inhibited both cell spreading and Lyn activation upon stimulation with FN. These results suggest that sustained activation of Lyn, which is regulated by membrane-anchored Chk, plays a crucial role in VLA5-mediated cell spreading but not cell attachment to a FN substrate.

Ntk, Lsk, and Batk) (11)(12)(13)(14)(15)(16)(17)(18). Like Csk, Chk has Src homology 3 (SH3) and SH2 domains and lacks the consensus tyrosine phosphorylation and myristoylation sites found in Src family kinases. Chk has been shown to phosphorylate the C-terminal negative regulatory tyrosine residue of Src family kinases (e.g. Lck, Fyn, c-Src, and Lyn) in vitro or in a yeast co-expression system, suggesting that Chk may share functional properties with Csk (13,14,19,20). However, Csk is ubiquitously expressed, whereas Chk expression is restricted to hematopoietic cells and neuronal cells in the brain. The expression of both Chk and Csk in these cell types implies either functional redundancy or specific roles for both kinases. Recent studies indicate that Chk and Csk might differentially regulate the functions of Src family kinases (18,(21)(22)(23)(24). In platelets, Chk is shown to negatively regulate Lyn but not c-Src due to the unique membrane localization of Chk, suggesting that co-expression of Chk and Csk confers specific roles for both kinases in platelet activation (20).
Cell adhesion to extracellular matrix proteins, e.g. fibronectin (FN), vitronectin, collagens, and laminin, is critical in cell growth, differentiation, and migration (25)(26)(27). Engagement of cell surface integrins triggers intracellular protein-tyrosine phosphorylation (28). There is increasing evidence that c-Src is involved in the regulation of cell adhesion. Previous studies with src Ϫ/Ϫ fibroblasts have indicated that the lack of c-Src results in a reduced rate of cell spreading on FN, although the spreading can be completed, and that expression of the SH3-SH2 domain of c-Src enhances the rate of cell spreading, suggesting that c-Src can affect cell adhesion of fibroblasts by a kinase-independent mechanism (29). In addition, Csk-overexpressing HeLa cells are reported to become spherical in cell morphology with reorganization of the vitronectin receptor (␣ v ␤ 5 integrin), suggesting a role of Csk in the regulation of integrins in HeLa cells (30). However, the involvement of Src family kinases in hematopoietic cell adhesion is still unclear.
In this study, we explored the role of Src family kinases in hematopoietic cell adhesion by means of both overexpression of Chk and expression of a truncated Lyn mutant lacking the kinase domain in the human megakaryocytic cell line Dami. We found that a sustained increase in Lyn kinase activity, which is regulated by membrane-anchored Chk, is required for VLA5-mediated cell spreading on a FN substrate, suggesting that activation of Lyn plays an important role in cell adhesion mediated by VLA5. pMKITneo vector containing human Chk-FLAG as a template. Polymerase chain reaction primers were designed with appropriate restriction sites to facilitate subsequent cloning. The resulting DNA fragments were all confirmed by DNA sequencing. The lysine to arginine mutation at position 262 in the ATP binding site of the kinase domain of Chk-FLAG was constructed previously (20). Dami cells (31) were obtained from the American Type Tissue Collection and grown in suspension in Iscove's modified Dulbecco's medium containing 10% heat-inactivated horse serum. Dami cells were transfected with the pMKITneo vector (kindly provided by Drs. K. Maruyama and T. Yamamoto) or the pM-KITneo vector containing human Chk-FLAG or each Chk mutant, and stably transfected cell clones were selected in 400 g/ml G418. To generate Lyn lacking the kinase domain (Lyn⌬K), the SR␣ promoterdriven pME18S vector encoding human p56 Lyn (kindly provided by Drs. H. Nishizumi and T. Yamamoto) (32) was digested with PstI and religated after removing the insert, resulting in the vector encoding Lyn⌬K (amino acid residues 1-298). Dami cells were cotransfected with the pME18S or the pME18S-Lyn⌬K vector, together with the pMiwhph vector (kindly provided by Dr. S. Nada) conferring hygromycin resistance, and stably transfected cells were selected in 200 g/ml hygromycin B.
Cell Spreading-Plates were coated with phosphate-buffered saline (PBS) containing 10 g/ml fibronectin (FN) or 1% bovine serum albumin for 4 h at room temperature and washed with PBS. Cells were preequilibrated in serum-free Iscove's modified Dulbecco's medium for 4 h at 37°C in bovine serum albumin-coated plates that prevent nonspecific adhesion. After washing, cells were resuspended in serum-free Iscove's modified Dulbecco's medium, seeded on FN-coated plates, and incubated at 37°C. Nonadherent cells were removed with gentle washing, and the adherent cells were observed under a phase contrast microscope. Cell spreading on FN-coated plates was quantitated by counting the number of cells showing decreased cell refractility and formation of projections around the cell periphery. For blocking, cells were reacted with an antibody against VLA4 (SG/73; Seikagaku Corp., Tokyo) (33), VLA5 (IIA1; Pharmingen) (34), or a control antibody (MOPC21; Sigma) at 20 g/ml before spreading assays.
Cell Attachment-Cells were metabolically labeled with [ 35 S]methionine (Tran 35 S-label, ICN) as described previously (35). The 35 S-labeled cells were washed with PBS and then incubated on FN-coated plates at 37°C for the indicated times. After nonadherent cells were collected, adherent cells were recovered by scraping, followed by solubilization with 2% SDS, and radioactivity was determined using a liquid scintillation counter.
Immunofluorescence-Cells were seeded on FN-coated coverslips as above and incubated at 37°C for 1 h. After fixing with 4% paraformaldehyde, cells were permeabilized with 0.1% saponin in PBS containing 3% bovine serum albumin, followed by immunostaining, using an antivinculin antibody (hVIN-1; Sigma), as described previously (36). For FACS analysis, without fixation and permeabilization cells were stained with anti-VLA4 or anti-VLA5 antibody or a control antibody, washed, and stained with secondary reagents. Viable cells were analyzed with a FACScan (Becton Dickinson).
Western Blotting and Immunoprecipitation-Cells were stimulated with FN as described above, except that 1 mM Na 3 VO 4 was included in adhesion medium during the last 60 min of preequilibration and subsequent stimulation periods. After removing nonadherent cells, cell lysates were prepared at 4°C in Triton lysis buffer (20 mM HEPES, pH 7.4, 137 mM NaCl, 5 mM EDTA, 1 mM Na 3 VO 4 , 1% Triton X-100, and protease inhibitors). Immunoprecipitations were performed with an anti-Lyn (Lyn44; Santa Cruz Biotechnology) antibody, as described elsewhere (20). Samples were subjected to SDS-polyacrylamide gel electrophoresis (37) and electroblotted onto polyvinylidene difluoride membranes (Millipore). Immunodetection was performed by enhanced chemiluminescence (Amersham Corp.) using antibodies against the FLAG epitope (M2; Eastman Kodak Co.), Chk (13G2) (20), Csk (Csk(C-20); Santa Cruz Biotechnology), Src (327; Oncogene Science), Lyn (Lyn9; Wako Chemical Co., Osaka), and phosphotyrosine (4G10; Upstate Biotechnology, Inc.) in conjunction with horseradish peroxidasecoupled F(abЈ) 2 fragments of anti-Ig (Amersham Corp.). The presence of Na 3 VO 4 in the adhesion medium had no effect on either FN-induced cell spreading or on the kinase activity of Lyn, while FN-induced tyrosine phosphorylation of cellular proteins could be detected only in the presence of Na 3 VO 4 (data not shown).
In Vitro Kinase Assay-After washing immune complexes with radioimmune precipitation buffer supplemented with 2 mM Na 3 VO 4 and with Triton lysis buffer containing 500 mM NaCl, an aliquot of Lyn immunoprecipitate was subjected to an in vitro kinase assay with 2 M [␥-32 P]ATP, and an equal aliquot was applied to a quantitative immunoblot, as described previously (20). The Lyn immunoprecipitates were incubated with acid-denatured enolase in a kinase buffer (50 mM HEPES, pH 7.4, 10 mM MnCl 2 , and 0.1% Triton X-100). After incubation at 30°C for 10 min, the reaction was terminated by addition of an equal volume of 2ϫ SDS sample buffer and boiled for 3 min. The samples were separated on SDS-polyacrylamide gel electrophoresis gels. The gels were treated with 1 N KOH at 56°C for 2 h and subjected to a BAS 2000 BioImage Analyzer (FUJIX, Tokyo).

Inhibition of FN-induced Dami Cell Spreading by Chk
Overexpression-To investigate the role of Src family kinases in hematopoietic cell adhesion, Chk was overexpressed in the human megakaryocytic cell line Dami. Transfected cell lines immunoblotted with an anti-Chk antibody showed that the expression levels of the FLAG-tagged Chk (see Fig. 3A; Chk-FLAG) (20) were about 10-fold higher than those of endogenous Chk (Fig. 1A). At least two independent sublines of each stable transfectant were used throughout the study, and results for a representative clone of each transfectant are shown.
When vector-transfected cells, which grow primarily in suspension, were seeded on a FN substrate, the cells attached to the substrate. The cell shape became irregular and the cell boundaries became difficult to visualize by phase contrast microscopy ( Fig. 1B, upper left panel), consistent with previous observations (38). Spread cells with anti-vinculin staining showed focal adhesions around the cell periphery (lower left panel). In contrast, when Chk-overexpressing cells were stimulated with FN, the size and number of spread cells were significantly reduced, and the formation of focal adhesions was inhibited (right panels), indicating that Chk overexpression leads to poor cell spreading or no morphological change.
The extent of cell spreading was quantitated by counting the number of cells that showed a flattened appearance, decreased cell refractility, and formation of projections after nonadherent cells were removed. Inhibition (Ͼ60%) of the number of cells spread was observed approximately 15 min after plating (Fig.  1C). This inhibition lasted at least 1 day (data not shown), contrasting with delayed phenotypes in fibroblasts that lack c-Src (29). No differences in cell attachment to FN-coated plates were observed between Chk-overexpressing and control cells (Fig. 1D), indicating that binding to a FN substrate is unaffected by Chk overexpression.
Dami Cell Spreading Mediated through VLA5-FACS analysis showed that Dami cells expressed the VLA4 and VLA5 integrins as major receptors for FN, and their expression levels were not affected by Chk overexpression (Fig. 2A). Cell spreading assays with blocking antibodies demonstrated that addition of an anti-VLA5 antibody, in contrast to that of an anti-VLA4 antibody, efficiently blocked cell spreading on a FN substrate (Fig. 2B). These results suggest that FN-induced cell spreading is mediated through VLA5 and that the expression level and binding activity of VLA5 to FN is unaffected by Chk overexpression.
Requirement of the SH3 Domain and Kinase Activity of Chk-To examine the structures of Chk required for inhibition of FN-induced cell spreading, various mutant forms of Chk-FLAG were prepared (Fig. 3A). Three mutants with deletions in the unique N-terminal, SH3, or SH2 domain (Chk ⌬N, Chk ⌬SH3, or Chk ⌬SH2, respectively) and a kinase-inactive mutant (Chk K262R) (20) were stably expressed in Dami cells. Each representative cell line produced a mutant protein of the predicted size, and the level of each mutant protein was comparable to or greater than that of Chk-FLAG (Fig. 3B), indicating that the expression levels in all cases are Ͼ10-fold higher than those of endogenous Chk. Fig. 3C shows that both Chk ⌬SH3and Chk K262R-expressing cells exhibited cell spreading on a FN substrate as did control cells (Vector), whereas cell spreading of Chk ⌬SH2expressing cells was significantly inhibited and Chk ⌬N-expressing cells showed significant but intermediate inhibition.
In addition, since the SH2 domain mutations in Csk are reported to destroy the kinase activity, probably due to susceptibility to denaturation upon cell lysis and immunoprecipitation (39 -41), we then examined whether the mutations in Chk affected kinase activities. In vitro kinase activities of the Chk mutants immunoprecipitated with an anti-FLAG antibody were measured using poly(Glu, Tyr) as an exogenous substrate. Wild-type Chk and all mutants except the Chk K262R mutant were found to possess nearly the same kinase activity (data not shown). These results suggest that both the SH3 domain and kinase activity of Chk are required for inhibition of FN-induced cell spreading and that the unique N-terminal domain may partly contribute to the inhibition.
Membrane Anchoring of Chk with Its SH3 Domain-Al-though Chk does not possess any known membrane anchoring motifs, about 45% of the overexpressed Chk and of the Chk ⌬SH2 mutant as well as endogenous Chk were localized to the particulate fraction (P100) which contained cellular membranes, with the remainder in the cytosolic fraction (S100) which contained the cytosolic content (Fig. 3D, upper panels). The Chk ⌬N mutation reduced the proportion of membrane localization to Ͻ30% of the mutant, which corresponds to a weak inhibition of cell spreading observed in cells transfected with this construct (Fig. 3C). However, it is important to note that the ⌬SH3 mutation completely abrogated membrane-anchoring of Chk, resulting in cytosolic localization of the Chk ⌬SH3 mutant, similar to that of the majority of Csk (Fig. 3D,  upper panel, lanes 7 and 8; upper middle panels, lanes 1-10). c-Src and Lyn were localized to the P100 fraction as expected due to their posttranslational lipid modification (lower panels). These results suggest that the SH3 domain of Chk is required for its membrane anchoring.
Effect of Chk Overexpression on Kinase Activity of Lyn-Lyn was found to be extremely abundant among the Src family kinases present in Dami cells (data not shown). To examine the effect of FN stimulation on Lyn activity, Lyn was immunoprecipitated from the Triton X-100 lysates, and in vitro kinase assays were performed with enolase. FN stimulation of control cells increased Lyn kinase activity (Fig. 4A, left panels). The level of Lyn activation, estimated to be ϳ3-fold, was sustained with a moderate increase during 45 min of stimulation. In contrast, Lyn activation was suppressed in Chk-overexpressing cells (right panels), supporting the idea that Chk negatively regulates Lyn activity in vivo. Without stimulation, the level of Lyn activity in control cells was estimated to be comparable to that observed in Chk-overexpressing cells when activity was normalized to the amount of Lyn in each sample. In addition, c-Src activity in Chk-overexpressing or control cells was unchanged upon FN stimulation (data not shown). These results suggest that Chk overexpression suppresses FN-induced Lyn activation without affecting the basal activity.
To examine whether the Chk SH3 domain was required for inhibition of FN-induced Lyn activation, the activities of Lyn were compared among the Chk mutant cells. Although the activation ratios in this experiment were slightly lower than those observed in Fig. 4A, the activity of Lyn in Chk ⌬SH3expressing cells obviously increased in response to FN stimulation, similar to the result obtained from control cells (Fig.  4B). On the other hand, the activity of Lyn in Chk ⌬SH2expressing cells was not augmented, similar to that seen in Chk-overexpressing cells. The activity of Lyn in Chk ⌬N-expressing cells was partially inhibited, which corresponds to weak phenotypes of cell spreading and membrane anchoring (Fig. 3, C and D). Collectively, these results suggest that FNinduced cell spreading involves the activation of Lyn, and the Chk SH3 domain, which promotes membrane anchoring, is required for inhibition of Lyn activation.
To further analyze kinetics of Lyn activation in control cells during FN stimulation, the levels of Lyn activity were quanti-tated. Fig. 4C shows that a slight increase in Lyn activity was detected 5 min after stimulation and a 3-fold increase was sustained after 15 min, indicating that the kinetics roughly correspond to those of cell spreading (compare with Fig. 1C).
Inhibition of Cell Spreading by Expression of Truncated Lyn-To explore whether activation of Lyn was required for FN-induced cell spreading, a truncated form of p56 Lyn lacking the kinase domain was stably expressed in Dami cells. Transfected cell lines produced mutant molecules of the predicted size of ϳ34 kDa (Fig. 5A). The expression levels were, however, varied and considerably lower than those of endogenous Lyn, suggesting that high expression of the Lyn mutant adversely affects cell growth. Nonetheless, expression of the Lyn mutant inhibited FN-induced cell spreading in a dose-dependent manner, indicative of a dominant-negative function (Fig. 5B). Most spread cells expressing the Lyn mutant exhibited immature phenotypes similar to those seen in Chk-overexpressing cells (Fig. 1B).
To ascertain whether FN-induced Lyn activation was inhibited by expression of the Lyn mutant, Lyn was immunoprecipitated and subjected to in vitro kinase assays with enolase. Fig.  5C shows that FN stimulation of vector-transfected cells gave rise to an increase in Lyn activity and a clearly detectable increase in autophosphorylation of Lyn due to the larger amounts of Lyn used than those in Fig. 4. In contrast, activation of Lyn in cells expressing the Lyn mutant was strongly inhibited, similar to that seen in Chk-overexpressing cells (Fig.  4), although the expression levels of the Lyn mutant were restricted. Without FN stimulation, Lyn activity was unaffected either with or without expression of the Lyn mutant, consistent with results seen in Chk overexpression studies. Thus, Lyn activation is required for FN-induced cell spreading. DISCUSSION In the present study, we demonstrate that Lyn plays a significant role in FN-induced cell spreading, but not cell attachment to a FN substrate, of the human megakaryocytic cell line Dami. Two lines of evidence suggest that sustained activation of Lyn is required for VLA5-mediated cell spreading of Dami cells. First, upon FN stimulation, cell spreading mediated through VLA5 accompanied a sustained increase in the kinase activity of Lyn with coincidental kinetics. Second, either overexpression of Chk or expression of a Lyn mutant lacking the kinase domain suppressed both FN-induced Lyn activation and cell spreading.
Our recent findings have shown that Chk, which localizes to membranes, selectively suppresses the kinase activity of Lyn but not c-Src in platelets (20). In this study, we overexpressed Chk in Dami cells and expected selective suppression of a Src family kinase in vivo, because Dami cells exhibit many of the morphological and biochemical characteristics of platelets and megakaryocytes (31,38). In fact, about half of the Chk protein present in Dami cells was localized to the particulate fraction via the SH3 domain (Fig. 3D) and overexpression of Chk was able to suppress Lyn but not c-Src ( Fig. 4; data not shown). Deletion of the SH3 domain caused not only cytoplasmic localization of Chk, similar to that of Csk, but also blocked inhibition of both FN-induced cell spreading and Lyn activation (Figs. 3C and 4B). The results that the kinase activity of Chk is required to inhibit Lyn activation (Fig. 3C) are likely to lead to the notion that Chk phosphorylates the C-terminal negative regulatory tyrosine residue of Lyn, as we previously demonstrated in platelets (20). These results suggest that membrane anchoring of Chk is crucial in suppression of Lyn activation in Dami cells. Indeed, we could detect a protein that specifically binds to the SH3 domain of Chk but not Csk. 2 Moreover, the expression levels of the SH2 domain-deleted mutant of Chk in every clone which we selected were greater than those of all other Chk molecules ( Fig. 3B; data not shown). This suggests that a role of the SH2 domain, if any, might be masked by excess amounts of the SH2 domain-deleted mutant. Careful evaluation of a role of the SH2 domain of Chk in suppression of Lyn activity is needed.
Previous reports have shown that Chk is present in the cytosolic, but not membrane, fraction in NIH3T3 fibroblasts transfected with Chk (18) and that ϳ80% of Chk tagged with the Src-derived membrane targeting signal (Src-Chk) is located in the cytosolic fraction in BI-141 murine T cell hybridomas transfected with Src-Chk (23). In the human immature myeloid cell line , Chk is present in the particulate fraction to an appreciable extent (our unpublished observations). It should be therefore emphasized that localization of Chk to membranes is restricted in certain cell types such as platelets/ megakaryocytic and myeloid cells and may play a role in their function. Expression of Chk K262R apparently neither reduced nor augmented the sustained increase in Lyn activity induced by FN (data not shown), consistent with the results that cells expressing Chk K262R exhibited FN-induced cell spreading to nearly the same extent as did control cells (Fig. 3C). It should be emphasized that Dami cells express endogenous Csk (Fig.  3D). These results suggest that Csk, like Chk, might be in-volved in the regulation of FN-induced Lyn activation and cell spreading.
The experiments with a Lyn mutant lacking the kinase domain demonstrated that activation of Lyn is required for cell spreading (Fig. 5). Either expression of the truncated Lyn or overexpression of Chk did not affect basal levels of Lyn activity (Figs. 4 and 5). These results suggest that the truncated Lyn or overexpressed Chk regulates Lyn activation only upon FN stimulation. Furthermore, expression of small amounts of the truncated Lyn showed a strong inhibitory effect on FN-induced Lyn activation (Fig. 5). Possibly, the ability of the truncated Lyn to be recruited to the sites of Lyn activation via its SH3 and SH2 domains may be greater than that of endogenous Lyn. This could result from freeing of the SH2 domain of Lyn as a consequence of removal of the C-terminal portion of the molecule. In addition, the inhibitory effects of the truncated Lyn on cell spreading were significant but not so drastic (Fig. 5B). As the activation of Lyn was almost completely abolished under these conditions (Fig. 5C), this finding suggests that suppression of Src family kinases might not fully explain the effect of Chk on cell spreading. Thus, these data may lead to the idea that in addition to the Src family kinases, Chk has other unidentified substrates, as has been suggested for Csk (30).
Previous studies with fibroblasts have shown that the lack of c-Src reduces a rate of FN-induced cell spreading although the cell spreading can be completed with normal flattened morphologies, and that the SH3 and SH2 domains of c-Src but not the kinase activity is sufficient to restore the rate of cell spreading (29). However, our findings show that either overexpression of Chk or expression of the truncated Lyn consisting of the SH3 and SH2 domains suppresses both cell spreading and Lyn activation throughout FN stimulation (Figs. 1C, 4, and 5, B and C), and that the level of c-Src activity is unchanged during FN stimulation (data not shown), indicating that activation of Lyn is required for Dami cell spreading. We imagine the different roles for Src family kinases in cell adhesion between fibroblasts and Dami cells.
Recent genetic analysis implicates the involvement of two Src family kinases, Hck and c-Fgr, in fibrinogen-induced cell spreading of polymorphonuclear cells mediated through ␤ 2 and ␤ 3 integrins (43). Our findings demonstrate that Lyn activation is required for FN-induced Dami cell spreading mediated by VLA5 (␣ 5 ␤ 1 integrin) (Figs. 2B, 4, and 5). These data suggest that activation of Src family kinases mediated by integrins may play a critical role in cell spreading of hematopoietic cells. It is intriguing to speculate that a class of integrin may be functionally linked to a specific member(s) of Src family kinases since expression of individual integrins varies among different lineages of hematopoietic cells (25-27, 44, 45).
On the basis of these findings, it should be emphasized that sustained activation of Lyn, which is regulated by membraneanchored Chk, is indeed a critical step in VLA5-mediated cell spreading but not cell attachment to a FN substrate. Further exploration of relevant substrates of Lyn will help us to understand the regulatory mechanism of the spreading of cells through tyrosine phosphorylation.