INTRODUCTION

T cell antigen receptor (TcR)¹ engagement through antigen peptide bound to molecules encoded by the major histocompatibility complex induces the activation of multiple biochemical pathways that result in changes in gene expression, cytokine production, and ultimately T cell proliferation. However, cumulative evidence suggests that additional T cell molecules are essential for optimal T cell activation and function (Hahn et al., 1992). These so called ``accessory molecules'' have dual functions; they may regulate the adhesive interactions between T cells and antigen-presenting cells, and they may participate in generating activation signals that will synergize with those signals initiated by the TcR. The CD43 molecule (sialophorin, leukosialin, or gp115) is the predominant cell surface sialoglycoprotein expressed on B and T lymphocytes, monocytes, neutrophils, and platelets (Axelsson et al., 1985; Borche et al., 1987; Remold-O'Donnell et al., 1987). CD43 is an integral membrane glycoprotein with a cytoplasmic domain of 123 amino acids. Resting T lymphocytes express a CD43 isoform of 113-122 kDa that contains O-linked tetrasaccharides attached to serine and threonine residues. Upon activation, T lymphocytes express a 125-135-kDa form of CD43 carrying mainly O-linked hexasaccharides (Piller et al., 1988; Piller et al., 1991). The use of monoclonal antibodies (mAbs) for each CD43 isoform (Tomlinson et al., 1994; Shiota et al., 1994) suggest that specific isoforms may have specific functions in T cells. The cytoplasmic domain of CD43 from human, rat, and mouse shows high sequence homology (more than 70% identity), indicating that this protein is involved in regulating T cell function.

Earlier observations of defective CD43 expression by T lymphocytes from patients with the X chromosome-linked Wiskott-Aldrich syndrome pointed out the importance of CD43 in lymphocyte function (reviewed by Rosenstein et al. (1993)). Thereafter, it was demonstrated that anti-CD43 mAbs induced T cell proliferation (Mentzer et al., 1987; Sperling et al., 1995; Alvarado et al., 1995),² homotypic adhesion, and interleukin-2 production (Axelsson et al., 1988) and that activation of T cells via CD43 resulted in the generation of second messengers like diacylglycerol, inositol phosphates, Ca²⁺ mobilization, and protein kinase C activation (Silverman et al., 1989). Expression of human CD43 in an HLA-DR-specific murine T-cell hybridoma was shown to enhance the antigen-specific response, an effect that required the cytoplasmic domain of CD43 (Park et al., 1991). Furthermore, the cytoplasmic domain of CD43 has been reported to be phosphorylated in resting T cells and to be hyperphosphorylated upon T cell activation with PMA, an activator of protein kinase C (Piller et al., 1989). Although these data suggest a role for CD43 in T cell activation, little is known about the mechanisms involved in CD43 signaling.

In this study we show that activation of human T lymphocytes by the anti-CD43 mAb L10 induces the association of the tyrosine kinase Fyn and CD43 and that this association is mediated through the Fyn SH3 domain. We also demonstrate that T cell activation via CD43 induces phosphorylation of Fyn kinase. Together, these data suggest a role for CD43 in T cell activation through a tyrosine phosphorylation cascade.

MATERIALS AND METHODS

L10, an IgG1 mAb that recognizes CD43 (Remold-O'Donnel et al., 1986), and OKT3 (American Type Culture Collection, anti-CD3, IgG2a) were either purified from ascites on protein A-Sepharose columns or used as ascites. Rabbit anti-mouse IgG (RMIG) was generated by repeated immunization with purified mouse IgG, and anti-mouse IgG immunoglobulins were affinity-purified. The anti-Fyn antibody was generated by rabbit immunization with purified GST-Fyn SH3 fusion protein. Protein A-Sepharose and Ficoll-Hypaque were from Sigma. The GST fusion proteins expressing the SH3 domains from Fyn and Lck were generously provided by C. Rudd (Dana Farber Cancer Institute, Boston, MA). Synthetic peptides containing proline-rich sequences that are involved in p85 binding to Fyn SH3 domain were the kind gift of R. Kappeler and L. Cantley (Harvard Medical School, Boston, MA).

Jurkat cells were cultured in RPMI 1640 (Hyclone, Logan, UT) supplemented with 5% fetal calf serum (Hyclone) and 5% bovine iron supplemented calf serum (Hyclone), 2 mM L-glutamine (Sigma), 50 µg/ml penicillin, 50 µg/ml streptomycin, and 50 µM -mercaptoethanol. Peripheral blood T cells were isolated from healthy adult donors by Ficoll-Hypaque gradient centrifugation. The buffy coat was washed three times with phosphate-buffered saline and resuspended in supplemented RPMI. Adherent cells were removed by plating the cells onto 60-mm Petri dishes (4 × 10⁷ cells/plate) for at least 2 h at 37 °C in a 5% CO₂ atmosphere. Nonadherent cells were collected and loaded on a nylon column pre-equilibrated with supplemented RPMI, incubated for 45 min at 37 °C, and eluted with supplemented RPMI. The resultant purified cells were predominantly (>80%) OKT3⁺ and L10⁺ (>95%) as determined by FACS analysis.

Cells (1 × 10⁶) resuspended in 50 µl of phosphate-buffered saline containing 2% fetal calf serum and 1% sodium azide (FACS juice) were incubated with L10 mAb (ascites, 1:100) or with OKT3 mAb (1 µg/ml) for 30 min at 4 °C. Cells were then washed by centrifugation at 300 g with FACS juice and incubated with RMIG coupled to fluorescein isothiocyanate for 30 min at 4 °C and washed as above. Next, cells were resuspended in FACS juice and fixed with 1% paraformaldehyde (final concentration). Cells were analyzed with a FACSort with the CELLQUEST program (Becton and Dickinson, San José, CA).

Purified T cells (2 × 10⁷) or Jurkat cells (5 × 10⁷) were incubated in 0.5 ml of cold RPMI for 15 min at 4 °C with the following antibodies alone or in combination: L10 (1:500 dilution of ascites or 1 µg/ml of purified IgG), OKT3 (used at suboptimal concentration, 10 ng/ml (OKT3s), or at activating concentrations 1 µg/ml (OKT3o)). Cross-linking was achieved by further incubating the cells with RMIG (1 µg/ml) for 15 min at 4 °C. Cells were activated by incubating them at 37 °C for the indicated times. Cells were then lysed in 100 µl of lysis buffer (25 mM HEPES, pH 7.4, 150 mM NaCl, 1.5 mM MgCl₂, 0.2 mM EDTA, O.5% Triton X-100, 0.5 mM DTT, 20 mM -glycerophosphate, 1 mM Na₃VO₄, 5 mM NaF, 4 mM phenylmethylsulfonyl fluoride, 1 µg/ml leupeptin, 1 µg/ml aprotinin) for 30 min at 4 °C. Lysates were spun at 14,000 × g for 15 min at 4 °C, and supernatants were incubated with 10 µg of fusion protein noncovalently coupled to glutathione-Sepharose beads for 3 h at 4 °C. Beads were washed four times with cold lysis buffer, and bound proteins were resolved by SDS-PAGE and subjected to immunoblotting.

Proteins were transferred to polyvinylidine difluoride membranes (Immobilon-P; Millipore, Medford, MA), membranes were blocked with 5% nonfat milk in Tris-buffered saline (10 mM Tris at pH 7.5, 150 mM NaCl) followed by incubation with the indicated antibody diluted in Tris-buffered saline with 0.05% Tween 20 (Bio-Rad). After three washes with Tris-buffered saline/Tween, the membranes were incubated with the appropriate second antibody coupled to horseradish peroxidase (Amersham, Buckinghamshire, UK) and proteins were visualized by ECL (Amersham), following the manufacturer's instructions.

Protein A-precleared lysates from activated or nonactivated T cells (1 × 10⁷ cellular equivalents) were immunoprecipitated as described previously (Pérez et al., 1995). Briefly, lysates were incubated with the indicated antibody for 1 h at 4 °C, and immune complexes were harvested with protein A-Sepharose for 30 min on ice and then washed once with cold TNE-T (150 mM NaCl, 50 mM Tris, pH 7.5, 5 mM EDTA, 1% (w/v) Nonidet P-40), twice with TNE (150 mM NaCl, 50 mM Tris, pH 7.5, 5 mM EDTA), and once with H₂O. Immunoprecipitated proteins were subjected to SDS-PAGE and immunoblotted as described above.

RESULTS

Numerous protein-protein interactions occur during T cell activation, some of these through SH2 or SH3 domains. SH2 domains interact with proteins containing phosphotyrosines (Songyiang et al., 1993), whereas SH3 domains interact with proline-rich sequences (Yu et al., 1992). The sequence analysis of the CD43 cytoplasmic domain revealed that it lacks tyrosine residues; however, a sequence that contains several proline residues shows homology with a consensus sequence for SH3 binding domains (Fig. 1). Protein-tyrosine kinases that are members of the Src family have been shown to play an important role in signaling pathways during T cell activation; therefore, we tested whether the CD43 molecule would interact with protein-tyrosine kinases containing SH3 domains by incubating different GST-SH3 fusion proteins, noncovalently bound to glutathione-Sepharose beads, with lysates from T cells stimulated with the anti-CD43 mAb L10 alone or in combination with anti-CD3 mAb OKT3. The proteins adsorbed to the fusion protein were analyzed by anti-CD43 immunoblotting. Activation of the Jurkat T cell line with anti-CD43 mAb L10 induced a strong association of CD43 with the Fyn SH3 domain (Fig. 2, lane 1). Association of CD43 to the Lck SH3 domain was also detected, although at a very low level as compared with the association with the Fyn SH3 domain (Fig. 2, lane 3). L10 stimulation of T cells did not induce CD43 association with Src or Abl SH3 domains (data not shown). Treatment of cells with RMIG alone did not induce CD43-Fyn SH3 or CD43-Lck SH3 association (Fig. 2, lanes 2 and 4).

Co-stimulation of human peripheral T lymphocytes (Fig. 3A) or Jurkat T cells (Fig. 3B) with anti-CD43 mAb L10 and anti-CD3 mAb OKT3 abolished the interaction between CD43 and the Fyn SH3 fusion protein, whether the cells were treated with suboptimal (10 ng/ml, lane 2) or optimal concentrations (1 µg/ml, lane 4) of OKT3. Thus, upon T cell activation with the anti-CD43 mAb L10, the CD43 molecule interacts with the SH3 domain of Fyn kinase, and signaling though the TcR interferes with this association.

Following incubation of T cells with anti-CD43 mAb L10 and RMIG at 4 °C, cells were activated by incubating them at 37 °C for different time periods. As shown in Fig. 4, the association of CD43 with the Fyn SH3 domain reached maximal levels after 30 s of activation at 37 °C (compare lanes 1 and 2), remaining elevated at 1 and 2 min (lanes 3 and 4, respectively) and decreasing almost to basal levels after 5 min (lane 5). The interaction between CD43 and the Fyn SH3 domain required CD43-dependent activation of the cells; no interaction between the two molecules was detected when cells were incubated for 1 min in the presence of RMIG alone (lane 6). At time 0 there was a detectable association between CD43 and the Fyn SH3 domain, suggesting that in resting T cells a proportion of the CD43 molecules is associated with Fyn kinase and that upon activation via CD43 this association is enhanced.

To determine whether the interaction between CD43 and Fyn is mediated by the Fyn SH3 domain and the proline-rich region of CD43, T cells were stimulated as above, and the lysates were incubated with Fyn SH3 fusion protein in the presence or absence of 10 µM of synthetic peptides. The peptides used contain proline-rich sequences that are involved in p85 binding to the Fyn SH3 domain (Kappeler et al., 1994). Peptide ³⁰⁰ERQPAPALPPKPPKP³¹⁴ (Fig. 1), corresponding to the second SH3 binding domain of p85 abrogated the CD43-Fyn interaction (Fig. 5A, lane 1), whereas peptide ⁸²SPPTPKPRPPRPLP⁹⁶ (Fig. 1) had no effect on the association between CD43 and Fyn SH3 fusion protein (Fig. 5A, lane 2).

CD43-Fyn interaction also occurred in vivo; when Fyn kinase was immunoprecipitated from L10-activated T cell lysates (that had been precleared with protein A), CD43 was detected in the immune complexes (Fig. 5B, lane 3). The presence of CD43 in those immunoprecipitates was not the result of residual L10-CD43 complexes binding to protein A, since CD43 was not present when an irrelevant antibody (anti-GST) was used for immunoprecipitation (Fig. 5B, right panel). Peptide 300-314 was also able to abolish the co-immunoprecipitation of CD43 with Fyn kinase using anti-Fyn antibodies (Fig. 5B, compare lanes 2 and 3). Taken together, these results provide evidence that CD43-Fyn interactions occur in vivo and that they are mediated by the SH3 domain of Fyn kinase, probably through the proline-rich sequence of CD43 (Fig. 1).

As shown in Fig. 3, co-stimulation of T cells with anti-CD3 and anti-CD43 mAbs abolished CD43 interactions with Fyn SH3 fusion protein. To determine whether TcR signaling had an effect on the association of CD43 with endogenous Fyn kinase, precleared lysates from activated Jurkat cells were precipitated with anti-Fyn antibodies. Fyn immunoprecipitates from cells co-stimulated with anti-CD3 and anti-CD43 mAb contained very little CD43 (Fig. 6, lanes 3 and 5) as compared with Fyn immunoprecipitates from cells stimulated with anti-CD43 mAb alone (Fig. 6, lane 1). Fyn immunoprecipitates from OKT3-activated T cells did not contain CD43 (Fig. 6, lanes 2 and 4), suggesting that signals generated by cross-linking the TcR prevent CD43-Fyn association.

In order to determine whether CD43-dependent T cell activation induced tyrosine phosphorylation of Fyn kinase, total cell lysates from purified T cells activated with L10 and/or different amounts of OKT3 were analyzed by Western blot with antiphosphotyrosine mAb 4G10. Within 1 min of activation with the anti-CD43 L10 mAb, tyrosine phosphorylation of Fyn kinase was detected (Fig. 7, lane 1), reaching a maximum at 5 min (approximately 3-fold enhancement) (Fig. 7, lane 6). Lck phosphorylation was not induced by L10 stimulation (Fig. 7, lanes 1 and 6). Co-stimulation with L10 and suboptimal concentrations of OKT3 (10 ng/ml) resulted in a faster kinetics of phosphorylation; maximum levels (5-fold) of Fyn phosphorylation were observed after 1 min of activation. (Fig. 7, lane 3), decreasing dramatically after 5 min (Fig. 7, lane 8). Cellular activation with OKT3 (1 µg/ml) alone or in combination with L10 did not alter the basal level of Fyn phosphorylation induced by L10, regardless of the time of activation (Fig. 7, lanes 4, 5, 9, and 10). However, cross-linking the TcR with OKT3 (1 µg/ml) alone or in combination with L10, followed by 1 min of activation, resulted in Lck phosphorylation that decreased at 5 min, (Fig. 7, lanes 4, 5, 9, and 10). Lck phosphorylation was not observed when cells were activated with suboptimal concentrations (10 ng/ml) of OKT3 alone or in combination with L10 (lanes 2, 3, 7, and 8). As shown in Fig. 7 (bottom panel), the differences observed in Fyn phosphorylation level may not be attributed to differences in the amount of Fyn, since similar levels of Fyn were present in all lanes, as determined by blotting with anti-Fyn antibodies. Thus, activation of T cells via CD43 induces tyrosine phosphorylation of Fyn kinase but not Lck phosphorylation.

DISCUSSION

In addition to the antigen receptor, co-receptor molecules contribute to T cell activation, by increasing the avidity of the interaction with the antigen presenting cell and/or by inducing separate signal transduction events that influence the fate of the cellular response. Protein tyrosine phosphorylation by activated tyrosine kinases is important in TcR-induced signal transduction. Protein-tyrosine kinase members of the Src family as well as the Syk/Zap 70 family play an important role in tyrosine phosphorylation pathways initiated during T cell activation (Weiss and Littman, 1994; Perlmutter et al., 1993). Tyrosine phosphorylation allows multiple protein-protein interactions mediated mainly by SH2 and SH3 domains (Pawson, 1995). These interactions are required for the appropriate propagation of the signal. CD43, the most abundant membrane protein of T lymphocytes, is known to initiate signal transduction pathways that lead to Ca²⁺ mobilization and interleukin-2 production (Silverman et al., 1989; Park et al., 1991), yet the molecular events involved in the CD43 signal transduction pathway are poorly understood.

In this study we investigated the mechanisms involved in T cell activation upon CD43 cross-linking. Examination of the CD43 amino acid sequence revealed the presence of a proline-rich motif that shows homology to the SH3 binding sites in the phosphatidylinositol 3-kinase-regulatory subunit p85 (Kappeler et al., 1994). These motifs have been implicated in mediating the interaction of p85 with the SH3 domains of Lck and Fyn kinases (Prasad et al. 1993a, 1993b). We tested the possibility that T cell CD43 could interact with SH3-containing proteins. We show that activation of both purified T lymphocytes and Jurkat cells, through CD43 cross-linking with the anti-CD43 L10 mAb, leads to CD43 association with Fyn kinase. This association is mediated by the SH3 domain of Fyn, since a GST-Fyn SH3 fusion protein is able to precipitate CD43 from lysates of CD43-activated T cells. Basal levels of CD43-Fyn complexes were detected at time 0; maximal association was observed after 1 min of activation, decreasing after 5 min. Fyn immunoprecipitates from T cells activated by CD43 cross-linking contained associated CD43, suggesting that this interaction also occurs in vivo. Moreover, the specificity of this interaction was demonstrated by the ability of a synthetic peptide containing the SH3 binding sites of p85, located between amino acids 300 and 314, to inhibit binding of CD43 to the GST-Fyn SH3 fusion protein and endogenous Fyn kinase, whereas a peptide containing an SH3 binding site of p85 located between amino acids 82 and 96 had no effect. Complete inhibition of the CD43-Fyn interaction was observed at 10 µM, a peptide concentration 20 times lower than that required to inhibit Fyn SH3 binding to p85 (Kappeler et al., 1994), suggesting that CD43 contains a low affinity binding site for the SH3 domain of Fyn. This low affinity SH3 binding site in CD43 may also mediate the interaction of this molecule with other SH3-containing proteins in response to different CD43 signals provided by different ligands. Our data show that although to a lesser extent, CD43-dependent T cell activation also induces the association of CD43 with the Lck SH3 domain.

It is possible that in resting T cells a proportion of the CD43 molecules are constitutively associated with Fyn and Lck kinases. L10 immunoprecipitates from resting T cell lysates prepared at low stringent conditions contain associated Fyn (data not shown). Another anti-CD43 mAb, MEM-59, has been shown recently to co-precipitate Lck from resting T cells (Alvarado et al., 1995). At least two physiological ligands have been reported for CD43: intercellular adhesion molecule 1 and galectin-1 (Rosenstein et al., 1991; Baum et al., 1995). Further experiments are needed to test whether intercellular adhesion molecule 1 or galectin-1 is able to initiate signaling cascades through binding to CD43 and whether CD43 association with Fyn or Lck is ligand-specific.

From the data presented here it is also clear that in T cells, TcR and CD43 signals overlap, since cross-linking the CD43 molecule with the TcR inhibits the association of CD43 both with the GST-Fyn SH3 fusion protein and with the endogenous Fyn kinase. There are several potential explanations why co-stimulation of T cells through the TcR and CD43 inhibits the association of CD43 with Fyn. One possibility is that signals from CD43 and TcR synergize, accelerating the kinetics of CD43-Fyn association. In agreement with this is the fact that cross-linking CD43 with the TcR induces maximal Fyn tyrosine phosphorylation after 1 min of activation, whereas similar Fyn tyrosine phosphorylation levels are reached only after 5 min of activation when CD43 is cross-linked alone. Alternatively, activated protein kinase C by TcR cross-linking could phosphorylate the CD43 cytoplasmic region, inducing a conformational change that would make the putative SH3 binding domain of CD43 inaccessible to the Fyn SH3 domain, thus preventing Fyn interaction with CD43. The CD43 cytoplasmic region has been shown to be phosphorylated by protein kinase C following TcR cross-linking or phorbol 12-myristate 13-acetate treatment (Piller et al., 1989; Wong et al., 1990). It was recently shown that treatment of Jurkat cells with phorbol 12-myristate 13-acetate inhibited the CD28 and p85 interaction mediated by the p85 SH2 domain (Hutchcroft et al., 1995). TcR activation could also enhance Fyn affinity for the TcR-CD3 complex, making Fyn no longer available to associate with CD43. Experiments need to be addressed to testing the cross-talk between the two signaling pathways.

In the present report we also provide evidence that, upon CD43 cross-linking, Fyn is tyrosine-phosphorylated in a time-dependent manner, suggesting that CD43-Fyn interactions are a prerequisite to induce tyrosine phosphorylation of Fyn. Basal levels of Fyn tyrosine phosphorylation are observed after 1 min of activation, when maximal CD43-Fyn SH3 domain interactions occur, whereas enhanced Fyn tyrosine phosphorylation was observed at 5 min postactivation, when most of the CD43-Fyn SH3 interactions are no longer detectable. Lck phosphorylation was induced when cells were stimulated with optimal concentrations of OKT3 (1 µg/ml) for 1 min, in agreement with previously published data (Burkhardt et al., 1994). Suboptimal concentrations of OKT3 (10 ng/ml) did not induce Lck phosphorylation under our experimental conditions. Higher levels of Lck phosphorylation were observed when cells were co-stimulated with OKT3 (1 µg/ml) and L10. However, L10 stimulation alone does not induce Lck phosphorylation either at 1 min or 5 min of stimulation, suggesting that CD43 signaling pathway does not lead to Lck phosphorylation by itself but that it could interact with signals generated by the TcR. Taken together, these results suggest that CD43 cross-linking with the L10 mAb on the T cell surface induces CD43-Fyn interactions through the Fyn SH3 domain and a putative SH3 binding site in CD43 and that this interaction induces Fyn tyrosine phosphorylation, leading to signal propagation. Although most SH3 domain interactions described are not T cell activation-dependent, it is possible that conformational changes in CD43 or Fyn that result from CD43-specific T cell activation allow this interaction to occur.