Functional and physical interaction of protein-tyrosine kinases Fyn and Csk in the T-cell signaling system.

The Src-like protein-tyrosine kinase Fyn is associated with T-cell antigen receptor. Transient expression of actively mutated Fyn, having Phe-528 instead of Tyr-528 or Thr-338 instead of Ile-338, in Jurkat T-cells stimulated the serum response element (SRE), 12-O-tetradecanoyl-phorbol-13-acetate response element, cyclic AMP response element, and c-fos promoter. The stimulation of SRE was particularly prominent not only with active Fyn but also with normal (wild-type) Fyn. SRE was also stimulated by both normal and active Lck. Furthermore, normal and active Fyn stimulated transcription from the IL-2 gene promoter when transfected cells were stimulated by concanavalin A plus 12-O-tetradecanoylphorbol-13-acetate. Under the same conditions, Lck did not stimulate IL-2 promoter unless it was activated by mutation. Interestingly, a mutant Fyn, which has deletions within the SH2 region and so is able to transform chicken embryo fibroblasts, did not stimulate either the c-fos or IL-2 promoter, suggesting the importance of this region in T-cell signaling. Csk, which phosphorylates tyrosine residues in the negative regulatory sites of Src family kinases, down-regulated Fyn- and Lck-mediated stimulation of the serum response element and Fyn-mediated enhancement of IL-2 promoter activity. These data suggest that Fyn and Lck, whose activities are regulated by Csk, are involved in different phases of T-cell activation.

li To whom correspondence should be addressed.
Crk (11). Recent evidence also shows that Csk is able to rescue the c-Src-induced lethal phenotype of Saccharomyces pombe, presumably by regulating kinase activity of the c-Src protein (12). Csk is expressed in lymphoid tissues and in neonatal brain (lo), which is rich in Src family kinases (1,2,13). Therefore, Csk may negatively regulate the functions of these Srclike kinases in lymphoid and neural cells.
The response of T-cells to antigen bound to the antigen receptor (T-cell receptor (TCR)')-major histocompatibility complex consists of a series of cellular events that result in expression of a number of genes including the c-fos and IL-2 genes (14). There are reports that activation of protein-tyrosine kinases following TCR stimulation is required and precede these intracellular events (15). In most cases TCR consists of an LIP heterodimer and CD3 subunits that are y, 6, E, and 5 (and/or chains. However, none of these molecules possesses intrinsic protein-tyrosine kinase activity. Recent evidence suggests that Src-like protein-tyrosine kinases couple the TCR for the signaling events (15). In particular, the Fyn protein, which is expressed in mature T-cells, physically associates with the CD3 complex (16). The importance of Fyn in TCR-mediated signal transduction is further suggested by the following studies. By use of a transgenic mice system, thymocytes expressing a high level of the fyn transgene were shown to be hyperstimulatable (17). Studies using fyn-mice revealed that TCR-mediated signaling was defective in mature single positive (CD4+CD8-or CD4-CD8+) thymocytes, although the signaling was still present in immature thymocytes and peripheral T-cells (18,20). The latter data suggest that fyn is critically important in TCRmediated signaling, at least in a developmentally restricted subpopulation of thymocytes. In addition, our unpublished data suggest that IL-2 production is elevated in T-cell hybridomas expressing high levels of wild-type and actively mutated

EXPERIMENTAL PROCEDURES Cell Culture-Human
Jurkat T-cells, donated from T. Taniguchi (Osaka University, Japan) and Class 11-restricted mouse helper T-cell hybridoma HBC21.7.31 and its derivatives2 were maintained in RPMI 1640 medium supplemented with 10% fetal calf serum. The monkey The abbreviations used are: TCR, T-cell receptor; CAT, chloramphenicol acetyltransferase; C o d , concanavalin A , TPA, 12-0-tetradecanoylphorbol- 13 (22). Relative kinase activity was determined by measuring the level of in oitro autophosphorylation of each Fyn immunoprecipitated from CV-1 cells that had been transfected with the expression plasmid containing corresponding fin cDNA. Tyrosine to p h e n y l a l a~e mutation at amino acid positions 528 and 531 is indicated as Y528F and Y531F, respectively, and isoleucine to threonine muta- exon 7B (221, which was synthesized by reverse transcription of mRNA from Jurkat T-cetls followed by polymerase chain reaction (23). The primers used for polymerase chain reaction were 5'-CCACCCGcGc-GccTAC-3', corresponding to the sense sequence at the 3' end of exon 6 and the antisense sequence at the 5' end of exon 8, respectively. These primers contained Hind111 sites, which were created by introducing point mutations (shown above in boldface letters) in the authentic sequence of fin cDNA. The boundaries of the exons were deduced from those of the c-src gene 124). The amplified DNA fragments for exon 7 were cleaved by Hind111 and cloned in pUC119. The nu~~eotide sequence of exon 7B was determined by the standard dideoxy chain termination method. The deduced amino acid sequence of exon 7B was compared with that of exon 7A (Fig. 1B).
The T-fin cDNA as well as B-fin, t-yjn, and f-14fj.n cDNAs were excised from pSRD vector (21) and recloned into pME18SM at the MluI site. The DNAfragments of the Zck cDNA (25) obtained from T. Shimotohno (National Cancer Center, Tokyo), the lyn cDNA (13), and the csk cDNA (26) were converted to the and c-fos CAT (28) were from K. Kaibuchi (Kobe University, Japan). The ZL-2 CAT plasmid (29) was from T. Taniguchi (Osaka University). ppact-@gal plasmid, in which the p-actin promoter was placed upstream of the p-galactosidase gene, was from S, Ishii (RIKEN, Japan). Dansfection Assay for CAT-Effector plasmid DNAs were transfected into Jurkat T-cells together with the reporter plasmid and ppact-@gal by the DEAE-dextran method as described (29). The transfected cells were incubated with or without ConA (30 pg/ml) and TPA (50 ng/ml) for 8 h before cell harvest. Cells were then harvested at 48 h after transfection, washed, and lysed as described (30). The lysates (80 pg of pwtein) were forwarded to CAT assay as described (30,31). CAT activities of the reaction products were quantitated with a Fujix Bio-Image Analyzer BAS 2000 (Fuji Film, Tokyo). The transfection efficiency was monitored by measuring the @-galactosidase activity of the cell lysates.

X~~s e -~t i v e
Fyn-The cDNA of active T-fin (T-FynF, Fig. 1) that encodes thymus-type Fyn (22) with Phe-528 instead of Tyr-528 was cloned into the expression plasmid pMEl8SM containing a chimeric promoter (SRcr) of the SV40 and human T

I~~h o t~p~c virus4 promoters (20). Reporter plasmid c-fos CAT (28) was cotransfected with the T-fjmF expression plasmids into Jurkat T-cells, which were chosen because of the similarity of their response to TCR stimulation to that of normal resting T-cells (34). Transfection of T-finF significantly
increased the c-fos promoter activity (Table I)

TABLE I Stimulation of fos and IL-2 promoters by Fyn
(Fos), SRECAT (SRE), TRECAT (TRE), CRECAT (CRE), or IL-SCAT (5 pg) as described (28,29). The lysates (80 pg of protein) of the transfectants Various Fyn cDNA constructs (5 pg) in pME18SM vector were transfected into Jurkat T-cells (2 x 109 together with reporter plasmids, c-fosCAT were subjected to CAT assay as described (30,31). The transfectants of IL-2CAT were treated with ConA (30 pg/ml) and TPA (50 ng/ml) for 8 h ( C o d + TPA) before cell harvest. Relative CAT activitiees of Fyn transfectants in comparison with mock (vector) transfectants are presented.  Table I). These data suggested that the c-fos promoter was activated by T-FynF primarily through SRE. In contrast, Jurkat cells transfected with T-fynF showed little activation of CAT activity of ZL-2 CAT over the cells cotransfected with the vector plasmid (Table I). This result indicates that elevated tyrosine kinase activity alone is not sufficient for induction of the ZL-2 gene. However, when T-cells cotransfected with ZL-2 CAT and the T-fynF constructs were stimulated simultaneously by TPA and a mitogen such as ConA or phytohemagglutinin, elevated kinase activity of T-FynF markedly stimulated the ZL-2 promoter (Table I). In this system, the phorbol ester TPA provided a second signal that is required for TCR-mediated mitogenesis. Next, we cross-linked the CD3 molecule with anti-C D~E antibodies on the surface of the Jurkat T-cells that had been transfected with the ZL-2 CAT and T-fynF constructs. On TPA treatment, CD3 cross-linking induced significant transactivation of the ZL-2 promoter (Fig. 3). Treatment with TPA and anti-CD~E without exogenous T-fynF had little effect. Thus we concluded that increased kinase activity of Fyn augments transcription of the ZL2 gene upon T-cell activation.
Effects of Normal Fyn, Lck, and Lyn on SRE and ZL-2 Promoter-Previous reports show that active forms of other Src family members, namely F505 p56" (36) and v-Src (371, enhance T-cell responsiveness. As ~5 6 ' "~ is abundantly expressed in T-cells and can contribute to T-cell signaling (3-5, 38), it is important to clarify whether Fyn and Lck regulate the signaling through similar or different mechanisms. Then we examined the effects of normal Fyn and normal Lck on SRE and the ZL-2 promoter. For comparison, the effect of the lyn gene product, which is not expressed in normal T-cells, was also tested. Two forms of normal fyn cDNAs, T-fyn and B-fyn ( Fig. 1) (22), transfected into Jurkat cells stimulated SRECAT and caused stimulation of the ZL-2 promoter only when the transfectants were stimulated with TPA plus ConA (Table I). No significant difference was observed between T-fyn and B-fyn. SRE was also stimulated by normal and active F505Lck (Table 11). Furthermore, as c-fos CAT was transactivated by Lck to a similar extent as by Fyn (Fig. 2 B ) , we tentatively concluded that the c-fos gene is also regulated by Lck most likely through SRE on T-cell activation. Active Lyn (F508Lyn) also stimulated SRE, while normal Lyn had only a slight effect (Table 11). These findings were substantiated by the observation that introduction of increasing amounts of the fyn and lck constructs, but not the lyn construct, stimulated SRE dose dependently (Fig. 4A). In contrast, we found that neither normal Lck nor Lyn activated the ZL-2 promoter in the presence or absence of ConA plus TPA stimulation (Table 11). When the amounts of the expression constructs were increased, the level of transactivation  lanes 1 and 5), TRECAT (lanes 2 and 6), CRECAT (lanes 3  and 7), and SRECAT (lanes 4 and 8). The lysates (80 pg of protein) of the transfectants were subjected to CAT assay. B, the expression plasmids pME18SM (5 pdreaction) containing cDNAs for normal T-Fyn of the ZL2 promoter by normal Lck or Lyn kinase remained low, while that by normal Fyn increased with an increase in the amount offyn construct transfected (Fig. 4B). From these data, we conclude that Fyn is a key protein-tyrosine kinase involved in the process of T-cell activation. It should be noted, however, that both active F505Lck (LckF) and F508Lyn (LynF) together with ConA plus TPA stimulated the ZL-2 promoter (Table II) (5 pg) were transfected into Jurkat T-cells (2 x lo6) together with reporter plasmids, SRECAT (SRE), or IL-2CAT (5 pg) as described in Table I.
The lysates (80 pg of protein) of the transfectants were subjected to CAT assay, and the enhancer/promoter activities were shown as described in dissect further the molecular mechanisms of T-cell responsiveness by Fyn, we examined the effects of Fyn mutants on SRE and ZL-2 promoters. Two mutants of B-fyn, f-14 and t-lfyns, which we isolated before (Fig. l), transform chicken embryo fibroblasts while normal fun does not (21). The products of these mutant fyn genes were shown to have elevated proteintyrosine kinase activities in vitro ((21) summarized in Fig. 1). The mechanisms of activation seem to be different because the nature of the mutations in the genes are different from each other. The f-14 mutant cames a point mutation resulting in Thr-338 instead of Ile-338 in the kinase domain. The t-1 mutant carries deletions of 78 base pairs in the sequence encoding the Src homology-2 (SH2) region. The product of f-14fyn stimulated the CAT activity of SRECAT, and the stimulation was greater than that by normal Fyn. On the other hand, the t-lfyn did not transactivate SRE (Table I). The possibility that the expression level of t-lfyn was much lower than that of normal fun or f-14fun in transfected cells is unlikely for the following reasons. First all the B-fyn constructs, the wild type, f-14, and t-1, directed almost the same level of Fyn expression when transfected into CV-1 cells (Fig. 5A) in which the level of endogenous Fyn is low. Second, Jurkat cells transfected with the SH2 deletion mutant t-1 expressed a significant amount of the exogenous Fyn protein as determined by immune complex kinase assay (Fig. 5B). Third, even when the increased amounts of the t-lfyn construct were cotransfected, the CAT activity of SRECAT remained at low levels (Fig. 4A). Thus, the SH2 region seems important in Fyn-mediated transactivation of SRE. Similarly the CAT activity of ZL-2 CAT was stimulated by f-14 Fyn but not by t-1Fyn (Table I) A4 and 841, f-14fyn (lanes A3 and B3 ), and t-lfyn (lanes A2 and

B 2 ) or no insert (lanes A1 and B l ) were transfected into CV-1 cells ( A )
and Jurkat T-cells ( B ) . The transfectants were harvested 48 h after transfection and lysed with RIPA buffer, and the lysates were subjected to Western blot analysis ( A ) or to immune complex kinase assay ( B ) using anti-Fyn antibodies. The position of Fyn is indicated on the right. Only a slight difference of migration between t-1Fyn and normal or f-14Fyn is acknowledged in B, since the kinase reaction products were analyzed in SDS-polyacrylamide gel containing a relatively hgher concentration (12%) of polyacrylamide.
plus WA. These findings suggest that Fyn kinase enhances signaling into the nucleus and further that the SH2 domain of Fyn is involved in this enhancement. Functional and Physical Interactions of Fyn with Csk in TCR-mediated Signaling-To determine the effects of Csk on transactivation of c-fos and ZL-2 promoters by Src-like kinases, we inserted csk cDNA (26) in the correct and reverse orienta-  and 4). normal Lck (lanes 7 and 8), activated T-FynF (lanes 5 and 6), or activated LckF (lanes 9 and IO) and SRECAT ( A ) or IL-2 CAT (B) reporter plasmid (3 pg). When the IL-2 CAT construct was used as a reporter plasmid, the transfectants were treated with C o d and TPA as described in Table I. The pMECsk (lane I ) or pMEKsc (lane 2 ) plasmid (3 pg) was transfected with the reporter plasmids (3 pg) and pME18SM vector plasmid without fin cDNA (3 pg). At least two independent experiments gave essentially identical resuits. It should be noted that the data with antisense csk cDNA in this experiment were consistent with those given in Tables I and I1 except that the level of stimulation of SRE by LckF was relatively lower than that by Lck. Probably this could be attributed to the poorer quality of the LckF plasmid that we prepared for this particular experiment. tions under the SRa promoter and transfected the resulting constructs into Jurkat cells together with the reporter CAT constructs and the fyn or Zck expression construct. Expression of sense csk cDNA suppressed normal Fyn-or Lck-mediated stimulation of SRE, while expression of antisense csk cDNA had virtually no effect on SRE stimulation by Lck and Fyn (Fig.   6A 1. As expected, SRE stimulation by the active forms of these kinases with tyrosine to phenylalanine mutations at the carboxyl-terminal phosphorylation sites (T-FynF and LckF) was barely affected by Csk. Transactivation of the ZL-2 promoter by normal Fyn was also suppressed by Csk (Fig. 6B). Again T-FynF-mediated transactivation of the promoter was little affected by Csk. These results suggested that Csk is involved in TCR-mediated signaling by regulating the kinase activities of Fyn and Lck to stimulate the c-fos and/or ZL-2 promoters. Then the possibility that Csk physically interacts with Fyn was ex- amined using the T-cell hybridomas HBC21.7.31 in which normal or f-l4fyn was exogenously introduced.2 From lysates of both normal fyn and f-14fyn overexpressed cells, we could coimmunoprecipitate Fyn with Csk using anti-Csk antibodies, which was demonstrated by in vitro kinase reaction (Fig. 7).
The amount of Fyn co-immunoprecipitated with Csk was apparently dependent on the level of Fyn expression (lanes 4 and

) .
More phosphorylation of f-14Fyn as compared with normal Fyn was detected in the Csk immunoprecipitates (lunes 5 and 6). This may suggest that f-14Fyn that is supposed to be highly autophosphorylated at Ty~-420 in vivo forms a stabler complex with Csk than normal Fyn. The associated Csk would phosphorylate Fyn in vitro. It is also possible that both forms of Fyn equally associate with Fyn and the levels of phosphorylation merely reflect the in vitro kinase activities of f-14Fyn and normal Fyn. Our experimental data do not discriminate these two possibilities. It is important to examine whether phosphorylation in vitro (lanes 5 and 6) occurred on Tyr-420 or on Tyr-531 in order to further analyze the mechanism of association. Nonetheless our data at least demonstrate physical association of Fyn with Csk, which suggests that Csk is involved in TCRmediated signaling.

DISCUSSION
Using a transient expression system we have shown that Fyn kinase activated by a point mutation, Tyr-528 Phe, confers hyperresponsiveness on T-cells. The T-cell responsiveness was determined by measuring stimulation of the promoter activities of both the c-fos and ZL-2 genes. Our evidence suggests that the c-fos promoter is stimulated by T-FynF through enhancement of SRE and TRE within the promoter ( Fig. 2A, Table I).
The ZL-2 promoter activity was augmented by T-FynF only when the cells received mitogenic stimulation that mimics TCR-mediated activation (Table I). This suggests that the elevated kinase activity alone is not sufficient to deliver the signal toward the nucleus. It is likely that T-cell specific signaling through Fyn may require its dimerization, as the signaling of the receptor-type protein-tyrosine kinase occurs via dimerization of the receptor (39). Since Fyn interacts with TCR complex, ligand-dependent clustering of TCR would result in Fyn dimerization. Similar, though less, stimulation of these promoters was observed by overexpressing normal Fyn (Table  I). In addition, kinase-active Fyn having a deletion in the SH2 region could hardly stimulate SRE or affect the promoter activity of the ZL-2 gene in response to activating signal ( Table I,   see below). From these data we have concluded that Fyn plays an important role in delivering signal from TCR to nucleus. In addition we would like to point out that the experimental system employed in this study is useful to analyze the molecular basis of Fyn function in T-cell activation. Namely, it is possible to coexpress possible regulatory proteins such as protein tyrosine kinase Csk and protein tyrosine phosphatases together with Fyn in Jurkat cells and examine the activities of c-fos and ZL-2 promoters in T-cell signaling (see below).
The suggestion of functional involvement of Fyn in TCR signal transduction is provided by previous studies. First, a rapid but transient increase in Fyn kinase activity followed TCR stimulation stimulation. Finally TCR-mediated signaling was defective in mature thymocytes of Fyn-deficient mice, although the signaling was unaffected in immature thymocytes and peripheral T-cells (18,19). The conclusion of this manuscript is consistent with these observations.
Previous studies have shown that another Src-like kinase Lck that is primarily expressed in T-cells is also important in the process of T-cell activation. For example, a failure to express functional Lck was unable to activate the TCR-regulated pathway of the protein tyrosine kinase (42). Consistently SRE and the c-fos promoter were stimulated by Lck as well as T-Fyn (Table 11, Fig. 2). However, we also demonstrated that normal Lck, unlike normal Fyn, was unable to stimulate transcription of the ZL-2 gene upon T-cell activation ( Table 11, Fig. 4B), suggesting that Lck and Fyn play roles in different phases of signaling. We propose that Fyn functions more proximal to the IL-2 gene than Lck.
The experimental data in this study also show that at least part of the SH2 region of Fyn is important in Fyn-mediated signaling in T-cells (Table I, Fig. 4A). Many lines of evidence show that SH2-containing proteins are important in signal transduction (4345). These proteins include a phosphatidylinositol-specific phospholipase C-y, the ras GTPase-activating protein, the p85 subunit o f phosphatidylinositol 3-kinase, and Grb-YSem5, which have been demonstrated to be phosphorylated a t tyrosine residues and/or associated with receptor protein-tyrosine kinases upon ligand stimulation. In a preliminary study, we found that the CD3 [ chain was associated with the Fyn protein when these proteins were expressed together in CV-1 cells (data not shown). Others also showed physical inter- -containing signaling molecules such as phospholipase C-y, GTPase-activating protein, and phosphatidylinositol 3-kinase, as insulin receptor-associated p185, which becomes phospho-rylated at multiple tyrosine residues by the insulin receptor upon insulin stimulation, interacts with phosphatidylinositol 3-kinase (51). It should be noted that ZAP-70 protein-tyrosine kinase, which was identified as ["associated kinase, contains two SH2 motives and is rapidly tyrosine-phosphorylated aRer stimulation of T-cells (52, 53). In addition, our present data suggest that at least some fraction of Csk is associated with TCR. Csk might be recruited by phosphorylated 5 to complete negative regulation of Fyn and/or Lck kinases, resulting in down-regulation of signaling. The kinase activity of the Src family is regulated by phosphorylation and dephosphorylation of a tyrosine residue that corresponds to QT-527 of c-Src and is conserved among the family members (1, 2). Phosphorylation of this regulatory tyrosine is catalyzed by a cytoplasmic protein-tyrosine kinase Csk (9). Regulation of the Src kinase activity by Csk has been demonstrated in vitro and in yeast (10, 12, 26). Although the kinase activity of Fyn has been shown to be down-regulated in vitro by Csk (lo), regulation of Fyn kinase by Csk in intact cells has not been demonstrated. Our present data clearly show that Csk suppresses activities of Fyn and Lck to stimulate the c-fos and ZL-2 promoters in T-cells (Fig. 6), which suggests that Csk is involved in TCR-mediated signaling. Furthermore we demonstrated co-immunoprecipitation of Csk and Fyn from the lysates of Fyn-overexpressed cells by means of in vitro kinase assay (Fig. 7). The efficiency of co-immunoprecipitation was low, and preliminary data indicated that no more than 1% of Fyn was physically associated in unstimulated cells. The molecular basis of this interaction remains to be elucidated. It should be noted, however, that no association of Csk with Src was observed in fibroblasts (54). Nevertheless recent studies in this laboratory suggested physical interaction of Lyn and Fyn with Csk when both kinases were overexpressed in CV-1 monkey kidney cells.3 Further studies are needed to explain the apparent discrepancy. Both Fyn and Lck are associated with the antigen-major histocompatibility complex receptor system (3, 4, 6). Moreover Csk is expressed ubiquitously but is expressed a t highest levels in certain tissues that include thymus (10) and has been isolated from the cellular membrane fraction (9). These data together suggest that Csk functions in the TCR-CD3-mediated signaling pathway.