Previously, we found that the ability of Csk to inhibit T-cell activation required the presence of its SH3 domain and that, through this region, Csk interacted at a high stoichiometry with the proline-enriched protein tyrosine phosphatase PEP (8
). Herein, we demonstrated that, like Csk, PEP was a strong negative regulator of antigen receptor–mediated protein tyrosine phosphorylation and lymphokine secretion in T cells. This effect required the phosphatase activity of PEP, and correlated with the capacity of PEP to bind to Csk. Reconstitution experiments in Cos-1 cells provided further support for the idea that PEP and Csk cooperate to inhibit TCR signaling, and that this synergism is dependent on their association. Finally, studies using a substrate-trapping mutant of PEP suggested that PEP inhibited TCR signaling by dephosphorylating the Src- related enzymes themselves and at least one of their targets, the protein tyrosine kinase Zap-70.
These observations suggested that the SH3 domain of Csk is needed for its inhibitory function in T cells because it mediates binding to PEP (or a PEP-related phosphatase; see below). Although it is formally possible that this domain also interacts with other proteins contributing to its activity, it should be mentioned that we and others have failed to identify additional ligands for the Csk SH3 region (9 and references therein). Taking into consideration the similar impact of Csk and PEP on TCR signaling, the high stoichiometry and specificity of their association, and the structural requirements for their respective inhibitory effects, it appears likely that the ability of Csk and PEP to inhibit TCR signaling requires their physical association. Through its dual ability to inactivate Src-related enzymes by phosphorylating their COOH-terminal tyrosine and to dephosphorylate critical TCR-regulated substrates, the Csk-PEP complex constitutes a potent mechanism to inactivate T cells. The purpose of this kinase–phosphatase complex is clearly different from that of CaMKIV-PP2A, Erk2-MKP3, and Lck-CD45, in which the associated phosphatase directly regulates the activity of the kinase (11
We found that enhanced PEP expression in BI-141 T cells caused a diminution in tyrosine phosphorylation of most TCR-regulated substrates, including Zap-70 and the ITAMs of ζ. Given the complex regulatory interactions between Src family kinases, ITAMs, and Zap-70, it is difficult to determine whether PEP acted by dephosphorylating and inactivating the ITAMs, Zap-70, and/or the positive regulatory site of Src family kinases. To help resolve this issue, we attempted to identify the immediate target(s) of PEP using a substrate-trapping variant of PEP (D195A PEP). In transiently transfected Cos-1 cells, this mutant was found to associate with Zap-70 and FynT. A similar association could be seen with Lck (data not shown). By contrast, D195A PEP did not interact with ζ. Hence, these findings suggested that PEP inhibited TCR signaling by regulating Zap-70 and Src family kinases, and that the reduction in ITAM phosphorylation in PEP-overexpressing cells was secondary to inactivation of Src-related enzymes. Of course, we cannot exclude the possibility that PEP also dephosphorylated more downstream targets in the TCR signaling cascade.
It is unlikely that the impact of PEP in BI-141 cells was the nonspecific consequence of overexpression of an irrelevant PTP. Firstly, the biochemical and biological effects of PEP were seen with moderate degrees (three- to fourfold) of overexpression. Secondly, and most importantly, they were not noted in cells containing ΔP1 PEP, a PEP variant that was defective in the ability to associate with Csk, but that remained catalytically active. On this basis, we feel confident that the effects of PEP in these cells were genuine, and that they reflected a physiological role of PEP in TCR signaling. The inability of phosphatase-inactive versions of PEP (C227S and R233M PEP) to behave as dominant-negative mutants and enhance TCR signaling in BI-141 cells may be viewed as a contradiction of this model. However, it should be remembered that ~5% of Csk is associated with endogenous PEP molecules. As indicated in Fig. C, there remains a large pool of presumably unoccupied Csk molecules than can accommodate additional PEP polypeptides. Consequently, it is not surprising that expression of the phosphatase-inactive mutants had no effect on TCR signaling, as they would not be expected to displace the endogenous PEP from Csk.
In addition to PEP, other PTPs have been implicated in the control of T-cell activation. The transmembrane PTP CD45 was shown to be a positive regulator of TCR signaling, as a consequence of its ability to dephosphorylate the negative regulatory site of Src-like kinases (41
). SHP-1 and SHP-2, two PTPs bearing tandem SH2 domains at their NH2
termini, have also been linked to this process. There is evidence that SHP-1 has an inhibitory impact on T-cell activation (43
), and that it can be activated by tyrosine phosphorylated Zap-70 (45
) or, in some T-cell subsets, inhibitory receptors such as KIRs and LAIR-1 (46
). In comparison, the function of SHP-2 is less certain. Whereas it has been proposed to be a negative regulator of TCR signaling based on its association with the inhibitory receptor CTLA-4 (48
), recent studies with dominant-negative SHP-2 mutants have suggested an opposite role (49
). Finally, other phosphatases such as PAC-1 and HePTP have been found to inhibit T-cell activation by acting on more downstream events of TCR signaling, such as MAP kinase activation (50
). Thus, in a manner analogous to protein tyrosine kinases, several PTPs appear to be involved in regulating protein tyrosine phosphorylation in T cells. Even though certain PTPs, especially SHP-1, share substrates with PEP, these enzymes are most probably regulated by distinct mechanisms. As a result, they are likely to be involved in different aspects of T-cell inactivation. Additional studies, including determination of the mechanism by which Csk-PEP is recruited to the site of Src family kinase activation, should help to substantiate this notion.
is ubiquitous and PEP is also expressed in hemopoietic cells other than T cells, the Csk-PEP complex may provide a similar function in cell types such as B cells and macrophages. Furthermore, it is noteworthy that Csk also associates via its SH3 domain with PTP-PEST (16
), a PEP-related PTP contained in all cell lineages (20
). Since the SH3 region of Csk seems to be critical for its function in nonhemopoietic cells (52
), the inhibitory kinase may also cooperate with PTP-PEST in these cells. In keeping with this notion, it was reported that expression of the catalytic domain of PTP-PEST was able to suppress the lethal effect of Src in yeast cells, seemingly by dephosphorylating its substrates (55
). Hence, a complex between Csk and a PTP could have a generalized role in the regulation of Src family kinases in vivo.
Because of the central role played by Src family kinases in cell signaling, significant interest has been directed towards understanding their regulation. Earlier studies showed that Csk-mediated phosphorylation of the COOH-terminal tyrosine of Src-like enzymes decreases their catalytic activity, by allowing an intramolecular interaction between the phosphorylated COOH terminus and the SH2 domain (5
). However, as revealed by recent structural data (1
), this phosphorylation is not sufficient for full inactivation of Src-related enzymes. Other mechanisms, including a documented intramolecular association between their SH3 domain and linker region, are involved in additional inhibition. The data presented in this report raise the possibility of a novel facet in this regulation, by suggesting that a complex between Csk and a PTP is also critical to repress signaling by Src family kinases in vivo.