CD28 Ligation Induces Tyrosine Phosphorylation of Pyk2 but Not Fak in Jurkat T Cells*

Protein tyrosine kinases are critical for the function of CD28 in T cells. We examined whether the tyrosine kinases Pyk2 and Fak (members of the focal adhesion kinase family) are involved in CD28 signaling. We found that ligating CD28 in Jurkat T cells rapidly increases the tyrosine phosphorylation of Pyk2 but not of Fak. Paxillin, a substrate for Pyk2 and Fak, was not tyrosine-phosphorylated after CD28 ligation. CD28-induced tyrosine phosphorylation of Pyk2 was markedly reduced in the absence of external Ca2+. Previous studies have shown that the T cell antigen receptor (TCR) induces tyrosine phosphorylation of Pyk2. In this report, the concurrent ligation of CD28 and TCR increased tyrosine phosphorylation of Pyk2; however, the extent of phosphorylation by both receptors was equivalent to the sum of that induced by each receptor alone. The Syk/Zap inhibitor piceatannol blocked CD28, and TCR induced tyrosine phosphorylation of Pyk2, suggesting that Syk/Zap is involved in Pyk2 phosphorylation. In contrast, the phosphatidylinositol 3-kinase inhibitor wortmannin blocked TCR- but not CD28-induced phosphorylation of Pyk2, suggesting that CD28 and TCR activate distinct pathways to induce tyrosine phosphorylation of Pyk2. Notably, depleting phorbol 12-myristate 13-acetate-sensitive protein kinase C did not block CD28- and CD3-induced tyrosine phosphorylation of Pyk2. These data provide evidence for the involvement of Pyk2 in the CD28 signaling cascade and suggest that neither Fak nor paxillin is involved in the signaling pathways of CD28.

The molecular mechanisms by which CD28 augments TCR function are not clear. Recent studies, however, have shown that CD28 and the TCR complex initiate distinct intracellular signals (2,3). For example, TCR-induced IL-2 production is almost completely inhibited by cyclosporin A, yet this drug has no effect on the production of IL-2 by the ligation of CD28 in conjunction with phorbol 12-myristate 13-acetate (PMA) (6 -8). These data, together with the fact that CD28 but not TCR induces the tyrosine phosphorylation of the adaptor protein Dok (9 -13), support the notion that TCR and CD28 induce biochemically distinct intracellular signals.
In the present study, we examined the involvement of Fak/ Pyk2 in the CD28 signaling pathways. We found that Pyk2, but not Fak or its substrate paxillin, is rapidly tyrosine-phosphorylated and activated after CD28 ligation. We also found that simultaneously ligating CD28 and TCR had an additive effect on tyrosine phosphorylation of Pyk2. Further analysis of CD28and TCR-induced tyrosine phosphorylation of Pyk2 showed that CD28 appears to utilize signaling pathways to phosphorylate Pyk2 distinct from those used by TCR.
Cell Activation and Preparation of Cell Lysates-Acute human T cell leukemia (Jurkat) cells, clone E6-1, were obtained from American Type Culture Collection (Bethesda, MD). Before activation, cells were washed with RPMI containing 0.01% BSA, resuspended in the same media in polyethylene tubes (Fisherbrand, Fisher). The cells were then incubated with NMG, anti-CD28 mAb, or anti-CD3 mAb, and the tubes were placed in a 37°C water bath for the indicated time. Following incubation, the cells were solubilized by adding an equal volume of 2ϫ ice-cold solubilizing buffer (10 mM Tris-HCl, pH 7.5, 1% Triton X-100, 0.5% sodium deoxycholate, 50 mM NaCl, 50 mM NaF, 2 mM Na 3 VO 4 , 0.5 unit/ml aprotinin, and 1 mM phenylmethylsulfonyl fluoride; final concentrations) and immediately placed on ice (65). The cells in the tubes were vortexed vigorously every 10 min. Following incubation, solubilized cells were moved to chilled 1.5-ml Eppendorf tubes, and insoluble material was removed by centrifugation at 21,000 ϫ g for 30 min at 4°C.
For costimulation studies, T cells in 0.01% BSA/RPMI were stimulated in suspension for 10 min at 37°C with different concentrations of anti-human CD3 mAb in the presence or absence of 3 g/ml of anti-CD28 mAb. Cell lysates were processed as described above. For stimulation in the absence of extracellular Ca 2ϩ , cells in suspension were divided into 2 aliquots (66). In 1 aliquot, the cells were washed in EMEM containing 0.01% BSA and then resuspended in the same media and stimulated with anti-CD28 mAb or anti-CD3 mAb as above. The cells in the other aliquot were washed with Ca 2ϩ -free EMEM containing 0.01% BSA and 4 mM EDTA, suspended in EMEM containing 0.01% BSA and 50 M EDTA, and then stimulated with anti-CD28 mAb or anti-CD3 mAb. Cell activation was as described above. To study the effects of inhibitors, T cells were pretreated in suspension for 30 min at 37°C with different concentrations of the PTK Syk/Zap inhibitor, piceatannol, or with the lipid kinase PI 3-kinase inhibitor wortmannin. After washing, the cells were resuspended in 0.01% BSA/RPMI and stimulated with anti-CD3 mAb or anti-CD28 mAb as above. To deplete protein kinase C (PKC), cells were incubated for 16 h at 37°C with 400 nM PMA. After washing, the cells were resuspended in 0.01% BSA/ RPMI and stimulated with anti-CD3 mAb or anti-CD28 mAb as above.
Immunoprecipitation and Immunoblotting-Cell lysates from 8 ϫ 10 6 cells were precleared by incubating for 1 h with 10 g of rabbit anti-mouse immunoglobulin that had been preincubated for 2 h with 50 l of protein A-agarose. Meanwhile, the primary antibody was added to 10 g of rabbit anti-mouse immunoglobulin that had been preincubated for 2 h at 4°C with 50 l of protein A-agarose beads. Precleared lysates were then added to the protein A-antibodies mixture and incubated for 2 h at 4°C. After incubation, the beads were pelleted by centrifugation and washed 5 times with ice-cold solubilization buffer. After the final centrifugation, the beads were resuspended in 2ϫ SDS-polyacrylamide gel electrophoresis sample buffer (final concentration, 75 mM Tris-HCl, pH 6.8, 2% SDS, 10% glycerol) and boiled for 5 min. For immunoblotting, aliquots from whole cell lysates (WCL) or immunoprecipitates were separated by SDS-polyacrylamide gel electrophoresis (10%) and electrotransferred onto polyvinylidene fluoride membranes. Immunoblotting with the horseradish peroxidase-conjugated anti-phosphotyrosine antibody PY-20 was as described previously (62,66,67). To confirm similar loading of samples, antibodies were stripped from the membranes as previously reported (66), and the proteins were reprobed with specific antibodies followed by horseradish peroxidase-coupled rabbit anti-mouse immunoglobulin (1:50,000 dilution). The signals were visualized using the LumiGLO kit according to the manufacturer's recommendations.

Induction of Pyk2 Tyrosine Phosphorylation by Aggregating
CD28 -To examine whether the PTKs Pyk2 and Fak are involved in CD28 signaling, Jurkat T cells were treated with anti-CD28 mAb, and the immunoprecipitates of Pyk2 and Fak were transferred to membranes and probed with anti-phosphotyrosine mAb. Immunoblotting of WCL with anti-phosphotyrosine mAb showed that aggregating CD28 on Jurkat T cells induces tyrosine phosphorylation of several proteins, the most prominent of which were 42-46-, 65-, 78-, and 95-kDa proteins (Fig. 1A). In contrast, NMG did not induce detectable protein tyrosine phosphorylation. CD28-induced tyrosine phosphorylation was apparent within 2 min, peaked at 5 min, and decreased slowly after that (Fig. 1B). As shown in Fig. 2A, treating the cells with 1 or 3 g/ml anti-CD28 mAb increased tyrosine phosphorylation of Pyk2. No detectable increase in the tyrosine phosphorylation of the kinase was seen in cells treated with NMG ( Fig. 2A). Time course studies showed that the phosphorylation of Pyk2 by CD28 cross-linking was first apparent within 2 min, reached a peak by 10 min, and decreased at 20 min (Fig. 2B). In contrast, incubating the cells with anti-CD28 mAb did not induce any detectable tyrosine phosphorylation of Fak (Fig. 2C). Fak tyrosine phosphorylation was FIG. 1. Induction of protein tyrosine phosphorylation by aggregating CD28. A, Jurkat T cells were treated with 3 g/ml NMG or anti-CD28 mAb for 5 min at 37°C. Proteins in cell lysates were separated by SDS-polyacrylamide gel electrophoresis, transferred to membranes, and immunoblotted with anti-phosphotyrosine mAb (␣-PY). Stim., stimulus. B, Jurkat T cells were stimulated with 3 g/ml anti-CD28 mAb for the indicated time. Tyrosine-phosphorylated proteins were detected as in A. Arrows indicate proteins that become tyrosinephosphorylated after CD28 ligation. not detected after CD28 ligation even when more cells were used in immunoprecipitation, when cells were stimulated with plate-bound anti-CD28 mAb, or when the probed membranes were exposed for extended periods to films (data not shown). The cytoskeletal protein, paxillin, becomes strongly tyrosinephosphorylated in response to the ligation of various receptors including TCR (66,68,69). However, CD28 ligation failed to induce the tyrosine phosphorylation of paxillin (Fig. 2D). These data show that Pyk2 is involved in CD28 signaling and suggest that neither Fak nor paxillin is involved in the signaling pathways of CD28.
Concurrent Ligation of CD28 and TCR Leads to Additive Tyrosine Phosphorylation of Pyk2-It is now well established that signals from TCR and CD28 cooperate in the activation of T cells. Previous studies have shown that cross-linking TCR increases the tyrosine phosphorylation of Pyk2 (52,53). Thus, we examined the effect of simultaneously aggregating CD28 and TCR on tyrosine phosphorylation of Pyk2. Immunoblotting WCL with anti-phosphotyrosine mAb showed that CD28-and CD3-induced protein tyrosine phosphorylation was different in pattern and intensity (Fig. 3A). Thus, some tyrosine phosphoproteins were common to TCR-and CD28-activated cells, but in general, the level of phosphorylation was stronger in the TCRactivated cells (Fig. 3A). These results are similar to those previously reported (11)(12)(13)24). No enhanced tyrosine phosphorylation could be detected in WCL as a result of concurrently ligating CD28 and TCR (Fig. 3A). To examine the effect of simultaneously ligating CD28 and TCR on Pyk2 tyrosine phosphorylation, Jurkat T cells were stimulated for 10 min at 37°C with different concentrations of anti-human CD3 mAb in the presence or absence of 3 g/ml anti-CD28 mAb. As shown in Fig. 3B, concurrent ligation of CD28 and TCR increased tyrosine phosphorylation of Pyk2. Interestingly, when low concentrations of anti-CD3 mAb were used, the extent of phosphorylation by both receptors was equivalent to the sum of that induced by each receptor alone, whereas at high concentrations the phosphorylation by ligating both receptors was less than the sum of that induced by each receptor alone (Fig. 3, B and  C). Similar results were obtained when the cells were stimu-lated with 0.5 instead of 3 g/ml anti-CD28 mAb (data not shown).
Optimal CD28-induced Pyk2 Tyrosine Phosphorylation Requires Ca 2ϩ -The engagement of CD28 has been shown to increase the levels of intracellular Ca 2ϩ (8); thus, we examined the role of Ca 2ϩ in CD28-induced tyrosine phosphorylation of Pyk2 by stimulating the cells with anti-CD28 in the presence or absence of external Ca 2ϩ . As shown in Fig. 4A, tyrosine phosphorylation of several proteins induced with anti-CD28 mAb or anti-CD3 mAb was profoundly reduced or completely abolished in the absence of Ca 2ϩ . As previously reported, anti-CD3 mAbinduced tyrosine phosphorylation of Pyk2 was reduced by Ca 2ϩ depletion (Fig. 4B) (52). CD28-induced Pyk2 tyrosine phosphorylation was also reduced in the absence of Ca 2ϩ (Fig. 4B). Whether the incomplete inhibition of receptor-mediated Pyk2 tyrosine phosphorylation is due to the release of Ca 2ϩ from intracellular stores or due to the tyrosine phosphorylation of Pyk2 by mechanisms that do not require Ca 2ϩ awaits further investigation. Nonetheless, these results show that optimal tyrosine phosphorylation of Pyk2 by CD28 signaling requires Ca 2ϩ .
The Syk/Zap Inhibitor Piceatannol Blocks Both TCR and CD28-induced Tyrosine Phosphorylation of Pyk2-Syk and the related PTK Zap are critical for TCR function; however, their  role in CD28 signaling is not clear. Syk has been shown to be critical for Fc⑀RI-mediated tyrosine phosphorylation of Pyk2 in mast cells (where Fc⑀RI is high affinity IgE receptor) (51). Thus, we used the specific Syk/Zap inhibitor piceatannol (70) to examine the role of these PTKs in CD28-and TCR-induced Pyk2 tyrosine phosphorylation. As shown in Fig. 5A, pretreating the cells with piceatannol reduced tyrosine phosphorylation of several proteins induced by CD28 or TCR. This is not surprising, as Syk/Zap become activated early in signaling cascades. Interestingly, piceatannol inhibited both CD28-and TCR-induced tyrosine phosphorylation of Pyk2 (Fig. 5B). Such inhibition was evident at concentrations that have been reported previously to selectively influence Syk/Zap activity (70). Furthermore, piceatannol appears to be more effective in inhibiting CD28-induced tyrosine phosphorylation of Pyk2 than inhibiting CD3-induced Pyk2 phosphorylation. Although these data suggest that Syk/Zap are involved in TCR-and CD28induced tyrosine phosphorylation of Pyk2; the possibility that piceatannol is also inhibiting PTKs other than Syk/Zap (or directly inhibiting Pyk2) cannot be ruled out. Thus, the precise role of Syk/Zap in CD28-and TCR-induced tyrosine phosphorylation of Pyk2 requires further investigation.
The PI 3-Kinase Inhibitor Wortmannin Inhibits TCR-but Not CD28-induced Tyrosine Phosphorylation of Pyk2-As the PI 3-kinase has been implicated in the signaling pathways of TCR and CD28, we examined the effect of the PI 3-kinase inhibitor wortmannin (71) on CD28-and TCR-induced tyrosine phosphorylation of Pyk2. Notably, wortmannin enhanced CD28-and TCR-induced tyrosine phosphorylation of several proteins (Fig. 6A). Similar enhancement by wortmannin of TCR-induced protein tyrosine phosphorylation has been previously reported (72). Treating the cells with wortmannin invariably decreased tyrosine phosphorylation of Pyk2 induced by aggregating TCR in a dose-dependent manner (Fig. 6B). In contrast, wortmannin had no apparent effect on CD28-induced tyrosine phosphorylation of Pyk2 (Fig. 6B). These data suggest that TCR and CD28 induce Pyk2 tyrosine phosphorylation by distinct pathways.
Depleting PMA-sensitive PKCs Does Not Block CD28-and

CD3-induced Pyk2
Tyrosine Phosphorylation-PKC is activated after CD28 and CD3 ligation (2,3,15,16); thus, we examined whether receptor-mediated tyrosine phosphorylation of Pyk2 requires PKC. Jurkat cells were treated for 16 h with PMA (400 nM) to deplete Ca 2ϩ -dependent and Ca 2ϩ -independent PKC (66,(73)(74)(75). As shown in Fig. 7A, PMA treatment of Jurkat T cells markedly reduced PKCs, as determined by blotting WCL with antibody to Ca 2ϩ -dependent PKC␣ and Ca 2ϩindependent PKC␦. Atypical PKC was not decreased, as atypical PKCs are insensitive to PMA (73)(74)(75). Interestingly, the marked decrease in PMA-sensitive PKCs did not block CD28and CD3-induced tyrosine phosphorylation of Pyk2 (Fig. 7B). Similarly, pretreating Jurkat T cells for 72 h with 10 Ϫ6 M of the PKC inhibitor chelerythrine chloride did not block receptormediated tyrosine phosphorylation (not shown). Together, the data strongly suggest that CD28 and CD3 induce tyrosine phosphorylation of Pyk2 by mechanisms that are independent of PMA-sensitive PKCs. DISCUSSION PTKs have been shown to be critical for CD28-mediated T cell activation (12,23). Here we identify the PTK Pyk2 as a potential player in CD28 signaling, as this kinase became tyrosine-phosphorylated after CD28 ligation. In contrast, the related PTK Fak was not tyrosine-phosphorylated in response to receptor aggregation. Similar tyrosine phosphorylation of Pyk2 but not of Fak has been previously reported for other receptors (47,56). Although Fak and Pyk2 are structurally related, they localize to different sites within the cell. This has been attributed to differences in the focal adhesion targeting sequence at their COOH-terminal domains (49,50). Hence, in adherent cells Fak is mainly localized to focal adhesion sites, whereas Pyk2 is mainly diffused throughout the cytoplasm (49,50). Thus, the inability of CD28 ligation to induce Fak tyrosine phosphorylation could be due to the compartmentalization of Fak to areas not involved in CD28 signaling.
Our data showed that the potent PI 3-kinase inhibitor wortmannin inhibited CD3-but not CD28-induced Pyk2 tyrosine phosphorylation. These results suggest that PI 3-kinase activation is involved in TCR-but not in CD28-induced tyrosine phosphorylation of Pyk2. Recent studies, however, have shown that some of PI 3-kinase effects are wortmannin-resistant. For example, the p85 subunit of PI 3-kinase has been reported to function as an adaptor molecule in a manner that is unrelated to the lipid or serine kinase activity of the p110 subunit (76). Thus, at this time we cannot conclusively dismiss the role of the p85 subunit of PI 3-kinase in CD28-induced tyrosine phospho-rylation of Pyk2. Regardless, the wortmannin data strongly suggest that CD28 and TCR induce tyrosine phosphorylation of Pyk2 by distinct mechanisms.
The intracellular signals generated after CD28 aggregation integrate with those initiated after TCR ligation to induce optimal T cell activation (2,3,41,77). In this report, coligating CD28 and TCR increased the tyrosine phosphorylation of Pyk2 compared with that induced by aggregating each receptor alone. The extent of tyrosine phosphorylation of Pyk2 brought about by simultaneous stimulation of both receptors was equivalent to the sum of that induced by each receptor alone; thus, the effect of stimulating both receptors is additive and not synergistic. These data suggest that the signals generated by both receptors do not integrate at the level of Pyk2 and that both receptors induce tyrosine phosphorylation of Pyk2 by distinct pathways, a possibility supported by our data with wortmannin (see above). If Pyk2 is playing a role in the costimulatory process, then Pyk2 tyrosine phosphorylated by each receptor must regulate downstream signals that eventually integrate to enhance T cell activation.
Pyk2 does not contain SH2 and SH3 binding domains; therefore, the phosphorylation of this kinase on tyrosine is critical for its interaction with the SH2 domains of other signaling molecules. Accordingly, the tyrosine phosphorylation of Pyk2 has been reported to result in its association with several SH2-containing signaling molecules, including Src kinases, and Grb2 (48,52,53,60,78). The association of Pyk2 with the Grb2⅐Sos1 complex may link Pyk2 to the small GTPase Ras signaling pathway. Recent studies also link Pyk2 to the signaling pathways of the Rho family GTPases CDC42 and Rac1 (57). These GTPases are regulated by guanine nucleotide exchange factors such as Vav and by GTPases such as Ras and have been implicated in controlling the activation of JNK (37,39,40). Indeed, there is evidence implicating Pyk2 in pathways that regulate JNK activation. For example, overexpression of Pyk2 in human embryonic kidney cells activated JNK and induced the phosphorylation of the glutathione S-transferase-c-Jun fusion protein (57). Furthermore, the expression of a catalytically inactive mutant of Pyk2 inhibited the activation of JNK induced by both UV light and by sorbitol treatment of PC-12 cells (57) and by MIP1␤ in a murine pre-B lymphoma cell line (55). In T cells, JNK activation appears to be controlled by signals triggered by CD28 ligation, as the cross-linking of TCR in the absence of CD28 ligation failed to activate JNK (41). Because Pyk2 upon phosphorylation associates with signaling proteins, and because it appears to function upstream of GTPases, the tyrosine phosphorylation of Pyk2 after CD28 ligation may be important for coupling CD28-initiated signals to GTPases and, in turn, to the JNK cascade. However, tyrosine phosphorylation of Pyk2 by itself may not be sufficient to activate JNK, as Pyk2 becomes tyrosine-phosphorylated by TCR ligation in the absence of JNK activation. Thus, additional signals initiated by CD28 ligation may also be required for JNK activation. Alternatively, CD28 and TCR ligation may cause tyrosine phosphorylation at different sites of Pyk2, and only those associated with CD28 ligation trigger pathways leading to JNK activation. FIG. 7. Depleting PMA-sensitive PKCs does not block CD28and CD3-induced Pyk2 tyrosine phosphorylation. A, Jurkat T cells were incubated for 16 h at 37°C with 400 nM PMA. To confirm PKC depletion, WCL were immunoblotted with antibody to Ca 2ϩ -dependent PKC␣, Ca 2ϩ -independent PKC␦, and atypical PKC. B, cells treated as in A were resuspended in 0.01% BSA/RPMI and stimulated with 3 g/ml anti-CD3 mAb or anti-CD28 mAb for 5 min at 37°C. Pyk2 immunoprecipitates were then blotted with anti-phosphotyrosine mAb (␣-PY, upper panel) or with anti-Pyk2 mAb (lower panel). Stim., stimulus.