Activation of ZAP-70 through specific dephosphorylation at the inhibitory Tyr-292 by the low molecular weight phosphotyrosine phosphatase (LMPTP).

The ZAP-70 protein-tyrosine kinase plays a central role in signaling from the T cell antigen receptor. Recruitment and activation of ZAP-70 are transient and are terminated by phosphorylation of negative regulatory tyrosine residues and dephosphorylation of positively acting sites. We report that the low molecular weight protein-tyrosine phosphatase (LMPTP) specifically dephosphorylates the negative regulatory Tyr-292 of ZAP-70, thereby counteracting inactivation of ZAP-70. Expression of low levels of LMPTP resulted in increased ZAP-70 phosphorylation, presumably at the activating Tyr-493 and other sites, increased kinase activity, and augmented downstream signaling to the mitogen-activated protein kinase pathway. The ZAP-70 Y292F mutant was not affected by LMPTP. Our results indicate that LMPTP, like CD45, dephosphorylates a negative regulatory tyrosine site in a protein-tyrosine kinase and thereby strengthens T cell receptor signaling.

The ZAP-70 protein-tyrosine kinase plays a central role in signaling from the T cell antigen receptor. Recruitment and activation of ZAP-70 are transient and are terminated by phosphorylation of negative regulatory tyrosine residues and dephosphorylation of positively acting sites. We report that the low molecular weight protein-tyrosine phosphatase (LMPTP) specifically dephosphorylates the negative regulatory Tyr-292 of ZAP-70, thereby counteracting inactivation of ZAP-70. Expression of low levels of LMPTP resulted in increased ZAP-70 phosphorylation, presumably at the activating Tyr-493 and other sites, increased kinase activity, and augmented downstream signaling to the mitogen-activated protein kinase pathway. The ZAP-70 Y292F mutant was not affected by LMPTP. Our results indicate that LMPTP, like CD45, dephosphorylates a negative regulatory tyrosine site in a protein-tyrosine kinase and thereby strengthens T cell receptor signaling.
The molecular mechanisms of signal transduction and T cell activation have been intensely studied during the past few years. It has become evident that several protein-tyrosine kinases (PTKs) 1 and protein-tyrosine phosphatases (PTPases) play crucial roles (1)(2)(3)(4)(5)(6). Thus, one of the earliest biochemical events seen in T lymphocytes triggered through the TCR complex is the enhanced phosphorylation of a number of cellular proteins on tyrosine residues (7,8). Inhibition of this phosphorylation by pharmacological agents prevents T cell activation as measured by both functional readouts and biochemical assays (9,10). TCR signaling is regulated by several PTPases, most of which oppose the PTKs by dephosphorylating them or their substrates (6,(11)(12)(13). A striking exception is CD45, which plays a predominantly positive role (14 -17) by dephosphorylating and activating the Src family PTKs Lck (18 -21) and Fyn (22). In addition, CD45 may have other regulatory effects on Src family PTKs (23), TCR- (24), or the ZAP-70 kinase (25).
The low molecular weight PTPase (LMPTP) is a unique enzyme with limited sequence homology to the other PTPases (26,27). LMPTP (also known as ACP1) was originally classified as an ubiquitously expressed acid phosphatase (28), but was later found to be highly specific for phosphotyrosine (29). Its crystal structure (30) confirmed that LMPTP uses a cysteinebased catalytic mechanism similar to other PTPases. The physiological function of LMPTP is unknown. A high level overexpression of LMPTP in cells transformed by PTK oncogenes leads to decreased proliferation and ability to form colonies in soft agar (26,31). It has also been suggested that LMPTP can interact directly with the platelet-derived growth factor receptor (32) and other receptor PTKs. We have recently reported that LMPTP is expressed in T cells and is transiently phosphorylated on tyrosine (33). LMPTP also differed from 13 other tested T cell-expressed PTPases in that it augmented TCRinduced reporter gene activation (34).
Here we report that LMPTP, like CD45, plays a positive role in TCR signaling by preferentially dephosphorylating a negative regulatory site, namely Tyr-292 of ZAP-70. This leads to a severalfold increase in the tyrosine phosphorylation of the kinase at its positive regulatory sites and enhanced kinase activity. In contrast, LMPTP did not affect the Src family PTKs Lck or Fyn or the phosphorylation of the TCR-or Itk/Emt by these kinases. Based on our results we suggest that LMPTP plays a positive role in TCR signaling by counteracting the negative regulation of ZAP-70.

MATERIALS AND METHODS
Reagents and Plasmids-The 4G10 anti-phosphotyrosine mAb was from Upstate Biotechnology Inc. (Lake Placid, NY). The 12CA5 mAb, which recognizes the hemagglutinin epitope, was from Roche Molecular Biochemicals, and the 9E10 mAb, which reacts with the Myc epitope, was from ATCC. Anti-ZAP-70 mAb was from Zymed Laboratories Inc. (San Francisco, CA). Anti-Itk/Emt and the Itk/Emt expression plasmids were kind gifts from T. Kawakami. The cDNA for ZAP-70 (kindly provided by C. Rudd and A. Chan) and the ZAP-70-FF and ZAP-70-Y292F mutants (from A. Chan) were all cloned into the pEF/HA vector (35). The GST-Vav fusion protein (amino acids 161-191) was generously provided by A. Altman. The expression plasmids for LMPTP, VHR, HePTP, and TCPTP have been described (34). GST fusion proteins were generated using the pEGST vector (36). All point mutants were verified by sequencing.
Cells and Transfections-Jurkat T leukemia cells, the Lck-negative variant JCaM1, the ZAP-70-negative P116, and COS cells were kept at logarithmic growth in RPMI 1640 supplemented with 10% heat-inactivated fetal calf serum, L-glutamine, and antibiotics. Jurkat T cells were stimulated with 5 g/ml of the anti-CD3⑀ mAb, OKT3, in RPMI at 37°C. COS-1 cells were transfected with 0.1-5 g of plasmid by lipofection. Jurkat cells were transfected with a total of 5-10 g of DNA by electroporation at 20 ϫ 10 6 per sample at 960 microfarads and 260 V. Cells were used 48 h after transfection.
Tryptic Peptide Mapping-These experiments were performed as described previously (12,39). Briefly, a piece of nitrocellulose filter containing the radiolabeled protein was blocked with polyvinyl pyrrolidone-360 and digested with three additions of 10 g of tosylsulfonyl phenylalanyl chloromethyl ketone-treated trypsin. The released phosphopeptides were separated by electrophoresis on cellulose thin layer plates at 1,000 V in pH 1.9 buffer for 27 min followed by ascending chromatography in n-butyl alcohol/pyridine/acetic acid/water (75:50:15: 60) and exposed to film.

LMPTP Causes Hyperphosphorylation of ZAP-70 -
We recently reported that expression of LMPTP caused an augmentation of TCR-induced NFAT/AP-1 reporter gene activation (34). This effect of LMPTP was unique among PTPases in that 13 other tested intracellular T cell-expressed PTPases either inhibited or had no effect on the reporter gene. To understand this effect and its possible physiological significance, we studied upstream TCR-induced signaling events. The most TCRproximal parameter found to be affected by LMPTP was the tyrosine phosphorylation of ZAP-70, a PTK that associates with the TCR-chain of the receptor complex after TCR-phosphorylation by Src family PTKs. Subsequently, ZAP-70 is activated by Lck-mediated phosphorylation at Tyr-493 (40,41) in subdomain VIII, a site corresponding to the activating auto-or transphosphorylation site in most protein kinases.
When Jurkat T cells were co-transfected with low levels of LMPTP together with a HA-epitope-tagged ZAP-70 in the pEFneo vector, ZAP-70 was found to be hyperphosphorylated up to 5-fold both in resting cells and in cells treated with anti-CD3 mAb for 5 min (Fig. 1a, upper panel). The amount of ZAP-70 was similar in each sample (lower panel). In contrast to this increase in the in vivo tyrosine phosphorylation of ZAP-70, the associated TCR-was unaffected. Expression of ZAP-70 together with LMPTP in COS-1 cells resulted in a similar increase in phosphotyrosine content of ZAP-70 (Fig. 1b, upper  panel). As in Fig. 1a, the amount of ZAP-70 was similar in all samples (lower panel). Furthermore, catalytically inactive LMPTP-C12S had the opposite effect (lane 6), while another small PTPase, VHR, potently dephosphorylated ZAP-70 (lane 3). Finally, LMPTP did not cause any augmentation of the Lck-dependent phosphorylation of Itk/Emt (42) (Fig. 1c, upper  panel). Thus, the positive effect of LMPTP on ZAP-70 seems to be relatively specific to this PTPase-PTK pair. Importantly, the effect was seen at physiological levels of LMPTP (0.1-0.3 g of plasmid DNA); at higher concentrations LMPTP became less stimulatory and then inhibitory, presumably due to nonspecific effects.
LMPTP Activates ZAP-70 -To test whether the LMPTPinduced hyperphosphorylation of ZAP-70 depends on the activation loop tyrosine 493, we expressed the Y492F/Y493F mutant of ZAP-70 (here termed ZAP-70-FF) in COS cells alone or together with LMPTP, or LMPTP-C12S as a control. Antiphosphotyrosine immunoblots of the immunoprecipitated ZAP-70 molecules demonstrate that this mutant of ZAP-70 was not inducibly phosphorylated on tyrosine by a co-expressed LMPTP, while the wild-type ZAP-70 was (Fig. 2a, upper panel). Anti-HA blots of the same immunoprecipitates showed a comparable level of ZAP-70 in each sample (lower panel). Thus, the Since phosphorylation of ZAP-70 at Tyr-493 has been shown to cause its enzymatic activation, we measured the ability of LMPTP to enhance the kinase activity of ZAP-70 against a Vav-derived peptide containing Tyr-174, which has been shown to be an excellent substrate for the related Syk kinase (43). Indeed, ZAP-70 immunoprecipitated from COS cells was capable of phosphorylating this substrate, and co-expression of LMPTP caused a marked increase in this phosphorylation (Fig.  2b). In contrast, LMPTP-C12S did not activate ZAP-70, and ZAP-70-FF remained inactive. A co-expressed Lck also caused activation of ZAP-70, as expected (40,41), and LMPTP further augmented this effect (Fig. 2b, lanes 8 -11). Thus, LMPTP activates ZAP-70 to the same extent as Lck, and the two enzymes have an additive effect on ZAP-70. Taken together these results suggest that LMPTP causes hyperphosphorylation and activation of ZAP-70 in a manner that requires the activating Tyr-493 site.
LMPTP Dephosphorylates Tyr-292 of ZAP-70 -The simplest explanation for the effect of LMPTP on ZAP-70 would be the specific dephosphorylation of a negative regulatory site, analogous to the effect of CD45 on Lck. To test the notion that ZAP-70 might be a direct substrate for LMPTP, we expressed a substrate-trapping (44) LMPTP-D129A mutant in T cells and assessed the ability of this mutant to co-immunoprecipitate with ZAP-70. As shown in Fig. 3a, the substrate-trapping mutant LMPTP bound to ZAP-70 at both low and high salt conditions, while ZAP-70 was detectable in wild-type LMPTP immunoprecipitates performed only in very low salt buffer. Since high salt conditions prevent nonspecific or weak protein-protein interactions, these results suggest that LMPTP can bind ZAP-70 through its catalytic cleft.
To directly test whether LMPTP can interact with ZAP-70 and dephosphorylate a negative regulatory site, we analyzed the effects of LMPTP on ZAP-70 phosphorylation by tryptic peptide mapping. Recombinant ZAP-70 was incubated with [␥-32 P]ATP to allow the kinase to autophosphorylate, including at the regulatory Tyr-292 residue (40,45). Subsequently, phospho-ZAP-70 was incubated with different nanomolar concentrations of recombinant LMPTP. As seen in Fig. 3b, 20 nM (assuming 100% purity and activity) of LMPTP caused a selective decrease in a spot in comparison with other peptides on the maps. The spot contained 1,857 cpm in the control sample, 1,252 cpm in the presence of 20 nM LMPTP, and 633 cpm in the presence of 60 nM LMPTP. This spot was absent in maps of an immunoprecipitated Y292F mutant ZAP-70 allowed to autophosphorylate in the presence of [␥-32 P]ATP, but was present in parallel maps of wild-type ZAP-70 (Fig. 3c). These experiments suggest that LMPTP preferentially dephosphorylates ZAP-70 at Tyr-292. In agreement with this notion, the substrate-trapping LMPTP-D129A mutant precipitated considerably more wild-type ZAP-70 than Y292F mutant from pervanadate-treated cells (not shown).
LMPTP Has No Effect on the ZAP-70 Y292F Mutant-Next, we expressed wild-type or Y292F-mutated ZAP-70 together with LMPTP in the ZAP-70-negative P116 clone of Jurkat. The cells were stimulated for a few minutes with anti-CD3 mAb and then analyzed for ZAP-70 tyrosine phosphorylation by immu-  1, 4, 7, and 10), wild-type LMPTP (lanes 2, 5, 8, and 11), or the substrate-trapping LMPTP-D129A (lanes 3, 6, 9, and 12). Cell lysis and immunoprecipitation were carried out in the indicated NaCl concentrations. b, tryptic peptide maps of recombinant ZAP-70 treated with the indicated concentrations of LMPTP for 15 min at 30°C. The arrow points at the spot containing phospho-Tyr-292. c, tryptic peptide maps of wild-type ZAP-70 or ZAP-70-Tyr-292 immunoprecipitated with the anti-HA tag mAb from pervanadate-treated Jurkat cells and then allowed to autophosphorylate in vitro. Note that the spot corresponding to Tyr-292 is missing in the Y292F mutant. noprecipitation and immunoblotting. As in previous experiments, LMPTP augmented the tyrosine phosphorylation of wild-type ZAP-70 (Fig. 4), although the expression levels were lower in these cells. In contrast, the Y292F mutant ZAP-70 was equally phosphorylated in the presence of LMPTP as in its absence. Quantitation of the bands indicated that Y292F mutant ZAP-70 contained 2.3 times as much phosphotyrosine as wild-type ZAP-70 in the absence of LMPTP. In the presence of LMPTP, the phosphotyrosine content of wild-type ZAP-70 was increased by 2-fold, while the Y292F mutant was unchanged. Thus, LMPTP can augment the tyrosine phosphorylation of ZAP-70 to a level that approaches that of Y292F mutated ZAP-70, suggesting that a substantial portion of phospho-Tyr-292 was dephosphorylated by LMPTP in the cells.
LMPTP Also Augments Signals Downstream of ZAP-70 -The positive effect of LMPTP on the phosphorylation and activity of ZAP-70 was also reflected in the downstream activation of the Erk2 mitogen-activated protein kinase (Fig. 5a). As specificity controls, we expressed two other PTPases normally found in T cells, HePTP and TCPTP. As expected (12,37), the former inhibited Erk2 activation, while TCPTP had little effect. These results have been obtained in several independent experiments, and they fit well with our previous observation (34) that LMPTP augmented TCR-induced NFAT/AP-1 reporter gene induction.
The ability of LMPTP to augment TCR-induced activation of Erk2 was also observed in JCaM1 cells co-transfected with Y505F-mutated Lck, which cannot be activated by dephospho-rylation. These experiments clearly demonstrated that the positive effect of LMPTP was independent of the phosphorylation of Lck at Tyr-505 (Fig. 5b). Thus, it is clear that LMPTP does not augment TCR signaling by activating Lck. Instead, the direct dephosphorylation of Tyr-292 of ZAP-70 is by itself sufficient to explain the effect of LMPTP. DISCUSSION We were initially puzzled by the finding that LMPTP, a phosphatase, caused increased phosphorylation of ZAP-70. However, the effect was observed repeatedly, and we decided to consider three different possibilities: (i) LMPTP may dephosphorylate and activate Lck to increase its phosphorylation of ZAP-70; (ii) another PTPase (e.g. SHP-1) is functionally inactivated by LMPTP-mediated dephosphorylation and subsequently reduces its dephosphorylation of ZAP-70; (iii) LMPTP may dephosphorylate a negative regulatory site on ZAP-70 leading to an allosteric activating effect or a reduced associa- tion with a negative regulator, such as c-Cbl (46), or another PTPase. Our data are consistent with the third model, although more complex or additional effects are difficult to exclude.
Our experimental results do not support the possibility that LMPTP augments ZAP-70 phosphorylation by activating Src family PTKs. First, we have been unable to detect any effects of LMPTP on Lck or Fyn. Second, LMPTP augmented the TCRinduced activation of the mitogen-activated protein kinase pathway in JCaM1 cells co-expressing the Y505F-mutated Lck, which cannot be activated through dephosphorylation of Tyr-505. Third, the tyrosine phosphorylation of TCR-was unaltered in ZAP-70 immunoprecipitates. Fourth, Lck-mediated tyrosine phosphorylation of Itk/Emt in co-transfected COS cells was not augmented by LMPTP, but was counteracted by LMPTP. Fifth, although LMPTP was able to cause some dephosphorylation of Lck in vitro, the dephosphorylation was slow and showed no preference for Tyr-505 over Tyr-394 (not shown).
The possibility that LMPTP reduces the action of another PTPase is more difficult to exclude. Only two PTPases have been suggested to dephosphorylate ZAP-70, namely SHP-1 (47,48) and PEP (49). Of these two, only SHP-1 is known to be tyrosine phosphorylated (50), but this phosphorylation does not affect the catalytic activity of the enzyme. We recently reported (48) that SHP-1 can indeed dephosphorylate ZAP-70, but that this required activation of SHP-1 by removal of the SH2 domains. Deletion of the C-terminal tyrosine phosphorylation sites in SHP-1 had no effect. Thus, a potential dephosphorylation of SHP-1 by LMPTP would not suffice as a mechanism for the effect of LMPTP that we report.
The model we favor is that LMPTP directly dephosphorylates ZAP-70 at Tyr-292. As mutation of Tyr-292 causes a significant increase (45) in the total tyrosine phosphorylation of ZAP-70 (despite the lack of one tyrosine), we propose that the removal of phosphate from Tyr-292 by LMPTP results in an increase of ZAP-70 phosphorylation at other sites (including Tyr-493) to an extent that exceeds the loss of phosphate at Tyr-292. This notion is supported by the lack of effects of LMPTP on Y292F-mutated ZAP-70 (Fig. 4).
The physiological relevance of LMPTP-mediated Tyr-292 dephosphorylation is difficult to address conclusively. The absence of phosphate at this site in resting T cells clearly indicates that a cellular PTPase with this specificity must exist in resting T cells. In our hands, none of the nearly 20 different PTPases we have tested (11-13, 34, 37, 38) has a positive effect on TCR signaling, except LMPTP and CD45. A physiological relevance is also suggested by the low level of LMPTP required for optimal stimulation of ZAP-70 and the membrane-proximal subcellular location of LMPTP. Finally, the opposite effect of the catalytically inactive LMPTP-C12S may be the result of a "dominant-negative" competition with endogenous LMPTP.
Based on our findings, we propose that LMPTP, like CD45, has a positive role in TCR signaling. By dephosphorylating Tyr-292 of ZAP-70, LMPTP counteracts the negative regulation of this kinase through the action of c-Cbl or c-Cbl-associated molecules. Since LMPTP itself can be regulated by tyrosine phosphorylation in T cells, it is likely that this role of LMPTP is modulated during T cell activation. Thus, LMPTP may be an important regulator of the threshold and duration of signaling through ZAP-70.