INTRODUCTION

One of the earliest biochemical events seen in T lymphocytes triggered through the T cell antigen receptor complex is the enhanced phosphorylation of a number of cellular proteins on tyrosine residues (1, 2). Inhibition of this event by pharmacological agents prevents T cell activation as measured by both functional readouts and biochemical assays (3, 4). It has become evident that several protein tyrosine kinases (PTKs)¹ and the CD45 protein tyrosine phosphatase (PTPase) play crucial roles (reviewed in Refs. 5-7), and that the T cell antigen receptor-induced cascade of transient tyrosine phosphorylation events depends on a dynamic interplay between these and, presumably, many additional PTKs and PTPases. In addition to CD45 (8-10), only two other PTPases have been implicated in T cell activation, namely SHP1 (11) and SHP2 (12).

The low molecular weight PTPases LMPTP-A and LMPTP-B constitute a class of PTPases with limited sequence homology to the other PTPases (13-16). Nevertheless, these enzymes are highly specific for PTyr (14). Chemical modifications and mutagenesis experiments have shown that their catalytic mechanism involves a cysteine residue, Cys-12, which participates in phosphoenzyme intermediate formation (15), as in other PTPases. The recent crystallization of LMPTP (16) showed that the catalytic center is quite similar to that of "classical" PTPases, with Cys-12 residing in the bottom of the catalytic pocket.

The physiological functions of LMPTP are unknown. Overexpression of LMPTP in cells transformed by PTK oncogenes leads to decreased proliferation and the ability to form colonies in soft agar (17). Thus, a potential physiological function of the LMPTPs is to control normal cell growth by interacting directly or indirectly with the PTK signaling network. Reportedly, LMPTP can interact directly with the platelet-derived growth factor receptor PTK in NIH3T3 fibroblasts (18).

We have examined the expression of LMPTP in T cells and have begun to investigate its potential role in T cell activation. We show that LMPTP is present in T cells, is tyrosine phosphorylated mainly at Tyr-131 presumably by Lck or Fyn, and becomes rapidly dephosphorylated on receptor ligation. As mutation or thiophosphorylation of Tyr-131 (and Tyr-132 to a lesser degree) affects the catalytic activity of LMPTP, it seems that LMPTP is regulated during T cell activation.

MATERIALS AND METHODS

The 4G10 anti-PTyr mAb and purified Lck were from Upstate Biotechnology Inc. (Lake Placid, NY). The 12CA5 mAb, which recognizes the hemagglutinin (HA) epitope, was from Boehringer Mannheim. The OKT3 hybridoma was from the American Type Culture Collection, and the mAb was used as ascites. The pcDNA-3 vector was purchased from Invitrogen (La Jolla, CA). The pRc/CMV-csk expression plasmid has been described before (19). Zap was in pEF/HA (20), and Lck was in pEF (21) without tag.

Jurkat T leukemia cells, and their Lck-negative variant JCaM1, the CD45-negative Jurkat J45.01, and other leukemic cell lines were kept at logarithmic growth in RPMI 1640 medium supplemented with 10% heat-inactivated fetal calf serum, L-glutamine, and antibiotics. Peripheral blood mononuclear cells (>80% T cells) were isolated from the blood of healthy volunteers by gradient centrifugation on Lymphoprep (Nycomed, Oslo, Norway). Jurkat T cells were stimulated with 5 µg/ml anti-CD3 mAb OKT3 in RPMI at 37 °C. The blood T cells were activated with 10 µg/ml phytohemagglutinin for 72 h.

Messenger RNA was prepared using standard protocols and was reverse transcribed using avian myeloblastosis virus reverse transcriptase using the manufacturer's protocol (Life Technologies, Inc.). For assessment of isoform expression, two isoform-specific 5-oligonucleotide primers were synthesized, termed primer A (5-A ACT CCG GGG TAT GAG ATA GGG AA-3) and primer B (5-T GCT GTT TCT GAC TGG AAC GTG GG-3). These, together with a 3-primer (5-ATA ACC ACT CGA GTG GTC AGT GGG CCT TCT CCA AGA A-3), were used in a PCR reaction and gave rise to the expected 343-base pair fragment from the LMPTP cDNA.

The full-length cDNA for human LMPTP was obtained from Jurkat mRNA by RT-PCR using the 5-primer (5-ATA ACC ACT CGA GTG GAA GAT GGC GGA ACA GGC TAC G-3) corresponding to the beginning of the open reading frame and the same 3-primer as above. The resulting 477-base pair product was inserted into the pCRII^TM cloning vector (Invitrogen) and sequenced. Subsequently, it was subcloned into the pcDNA-3 eukaryotic expression vector with an added N-terminal HA epitope tag. It was also subcloned into the pGEX-3T prokaryotic expression vector (Pharmacia Biotech Inc.) and the encoded glutathione S-transferase (GST) fusion protein expressed and purified as recommended by the manufacturer and then used for immunization of two rabbits.

The codon for Cys-12 was changed into a codon for serine (C12S), and the codons for Tyr-131 and Tyr-132 were changed into codons for phenylalanine (Y131F, Y132F, and Y131F/Y132F mutations) in the pcDNA-3/lmptp-B plasmid using the Transformer^TM site-directed mutagenesis kit as recommended by the manufacturer (Clontech, Palo Alto, CA) and verified by sequencing. The inserts were also subcloned into the pGEX-3T prokaryotic expression vector, and the proteins were expressed and purified.

COS-1 cells were transfected with 10 µg of plasmid by lipofection as before (21-23). 48 h after transfection, the cells were lysed in 20 mM Tris/HCl, pH 7.5, 150 mM NaCl, 1% Nonidet P-40, 5 mM EDTA, 10 µg/ml aprotinin and leupeptin, 100 µg/ml soybean trypsin inhibitor, 1 mM phenylmethylsulfonyl fluoride, and 1 mM Na₃VO₄. For PTPase assays, Na₃VO₄ was omitted.

Immunoprecipitation was carried out as described before (12, 21-23). Proteins were resolved by SDS-polyacrylamide gel electrophoresis, transferred onto nitrocellulose filters, and immunoblotted with specific antisera or mAbs. The blots were developed by the enhanced chemiluminescence technique (ECL; Amersham Corp.).

The catalytic activity of LMPTP was measured as described (24). The reaction was for 30 min at 37 °C in 100 µl of 50 mM sodium citrate, pH 5.5, with 5 mM p-nitrophenyl phosphate as a substrate. The production of p-nitrophenol was measured colorimetrically as A at 410 nm.

The phosphorylation of LMPTP was carried out as described (24). Briefly, 5 µg of GST-LMPTP protein was incubated in 40 µl of 25 mM Hepes, pH 7.5, containing 100 mM NaCl, 10 mM MgCl₂, 10 mM MnCl₂, 0.1% Nonidet P-40, 10% glycerol, and 50 µM ATPS without or with 5 units of purified Lck for 3 h at 30 °C. 4-µl aliquots containing 0.5 µg of GST-LMPTP were then directly assayed for PTPase activity by addition of 100 µl of PTPase assay buffer as above.

RESULTS AND DISCUSSION

The two identified splice variants, termed A and B, of LMPTP differ in their use of an exon that encodes for amino acids 40-73 (25). To survey T cells and leukemic cell lines for the expression of these two isoforms, we used two isoform-specific 5-oligonucleotide primers and a common 3-primer in a RT-PCR assay starting with 0.5 µg/sample total RNA isolated from a variety of cell lines and freshly isolated blood mononuclear cells (>80% T cells). Jurkat T cells were found to express very low amounts of the B isoform only, whereas the CD45-negative Jurkat line J45.01 expressed both isoforms at relatively higher levels. Normal T lymphocytes expressed only the B isoform at levels that were higher than in Jurkat cells, but not as high as in J45.01 cells. Expression did not change during activation of the cells with phytohemagglutinin for 3 days. The myelomonocytic leukemia cell line HL-60 expressed both isoforms, whereas the erythroleukemia cell line K562 contained only isoform B message. Thus, LMPTP seems to be expressed in all tested cell lines, but the alternative splicing mechanism seems to be independently and variably used. Importantly, normal T lymphocytes seem to express exclusively the B isoform. Therefore, we decided to initially concentrate on this isoform.

To obtain the full-length cDNA of LMPTP-B, we synthesized oligonucleotide primers corresponding to both ends of the open reading frame of human LMPTP and used them in a RT-PCR as above. The resulting 477-base pair PCR product was first cloned into the pCRII^TM vector and sequenced. The obtained nucleotide sequence was 100% identical to the published sequence of the B isoform of LMPTP from human red blood cells (25). This cDNA was subsequently cloned into the pcDNA-3 eukaryotic expression vector, which adds a hemagglutinin epitope tag recognized by the 12CA5 mAb to the N terminus of the insert. In transiently transfected COS cells, this construct resulted in the appearance of a ~19-kDa protein that was both immunoblotted and immunoprecipitated with the 12CA5 mAb. The pcDNA-3/lmptp-B construct also gave a positive result in the RT-PCR with the B-specific primer but not with the A-specific primer. A catalytically inactive C12S mutant of LMPTP (LMPTP-C12S) was generated by site-directed mutagenesis. The full-length cDNA for LMPTP-B was also cloned into the prokaryotic expression vector pGEX, and the recombinant fusion protein was expressed, purified, and used for immunization of two rabbits. The resulting antiserum both immunoblotted and immunoprecipitated an 18-kDa protein in T cells (Fig. 1A) and other human or murine hematopoietic cell lines. It also reacted with a slightly slower migrating protein in COS cells transfected with the pcDNA-3/lmptp-B construct having the HA tag (not shown).

LMPTP-B immunoprecipitated from unstimulated Jurkat T cells was found to react with anti-PTyr mAbs (Fig. 1B). On treatment of the cells with 5 µg/ml of anti-CD3 mAb, the PTyr content of LMPTP decreased rapidly (Fig. 1B). A decrease was detected within 15 s (Fig. 1C) and continued during the first minute of stimulation. The amount of LMPTP in the immunoprecipitates did not change (Fig. 1C, lower panel). In a reverse experiment, anti-PTyr immunoprecipitates from resting Jurkat cells contained easily detected amounts of LMPTP, which declined on activation of the cells (not shown).

Tyrosine phosphorylation of LMPTP-B was also achieved in COS cells by co-expression of the Lck (Fig. 2) or Fyn (not shown) PTKs, but not by Csk, Zap (Fig. 2), or Jak2 (not shown) PTKs. Recombinant Lck also readily phosphorylated GST-LMPTP in the presence of [-³²P]ATP in vitro. (not shown).

As LMPTP-B contains only 5 tyrosine residues, and 2 of them are exposed on the surface of the molecule (16) and are preceded by two acidic residues at positions 3 and 2, as preferred by many PTKs, we decided to test whether 1 of these was the modified tyrosine residue. The Y131F, Y132F, and double Y131F/Y132F mutants were generated by site-directed mutagenesis, expressed with Lck in COS cells, and analyzed for PTyr content. As seen in Fig. 3, these experiments showed that Tyr-131 is the primary site of phosphorylation, and that Tyr-132 may also be phosphorylated, but to a much lower stoichiometry.

Tyrosine phosphorylation of LMPTP-B or its Tyr  Phe mutants by Lck in COS cells.

To begin to examine the effects of tyrosine phosphorylation of LMPTP on its catalytic activity, we generated the recombinant GST fusion proteins of the wild-type enzyme and its Y131F, Y132F, and double Y131F/Y132F mutants. After determination of protein concentration and verification of purity by SDS-polyacrylamide gel electrophoresis and Coomassie Blue staining, equivalent amounts of these proteins were subjected to PTPase assays using p-nitrophenyl phosphate as a substrate. As seen in Fig. 4A, wild-type LMPTP and the Y132F mutant were active, whereas the activity of the Y131F mutant was much reduced, and the double mutant was essentially inactive. This result was obtained in several independent experiments, although the activity of the Y132F mutant was more reduced in some experiments.

Catalytic activity of LMPTP-B or its Tyr  Phe mutants.

To exclude the possibility that the mutant LMPTPs are less active due to instability in the Escherichia coli system, we also immunoprecipitated wild-type LMPTP and its Tyr  Phe mutants from transfected COS cells and measured their activity (Fig. 4B). The results were very similar to those seen with the recombinant protein. Taken together, these results indicate that Tyr-131 is important for the activity of LMPTP. This conclusion is supported by the location of Tyr-131 next to the catalytic cleft in LMPTP (16).

A recent paper by Rigacci and co-workers (24) reported that LMPTP is tyrosine phosphorylated in v-Src-transformed NIH3T3 cells, and that Src was able to phosphorylate LMPTP in vitro. In addition, these authors showed that tyrosine thiophosphorylation of LMPTP by v-Src in the presence of ATPS caused activation of the PTPase, similar to the effect of tyrosine thiophosphorylation of CD45 by Csk (19). Rapid autodephosphorylation precluded the use of regular ATP in these assays (24).

To test whether phosphorylation of LMPTP at Tyr-131 or Tyr-132 is responsible for this change in enzymatic activity, we incubated the recombinant LMPTP proteins with purified Lck and 50 µM ATPS as described by Rigacci and co-workers (24). Subsequent PTPase assays revealed that Lck activated the wild-type LMPTP about 2-fold (Fig. 5). Lck also caused a small increase in the activity of the Y131F mutant and a large increase in the activity of the Y132F mutant. In contrast, the double mutant remained inactive. Thus, it seems that phosphorylation of either Tyr-131 or Tyr-132 can activate LMPTP, but that phosphorylation at Tyr-131 is more potent (perhaps due to higher stoichiometry of phosphorylation). Similar results were obtained in two additional independent experiments, and neither ATPS nor Lck alone contained any PTPase activity.

We report that T cells contain the B isoform of LMPTP, which is tyrosine-phosphorylated in resting cells but rapidly loses its PTyr on stimulation of the cells. We identify Tyr-131 as the major phosphorylation site and Tyr-132 as a minor site and the Src family PTKs Lck and Fyn as enzymes capable of phosphorylating these sites in vivo and in vitro. Both Tyr-131 and Tyr-132 are located next to the catalytic pocket of LMPTP, and especially, Tyr-131 seems to be important for the activity of LMPTP. Phosphorylation of Tyr-131 or Tyr-132, particularly the former, caused an increase in the activity of LMPTP. The location of these residues suggests that they may also affect interaction with cellular substrates. Taken together, our observations indicate that LMPTP-B may by inducibly tyrosine dephosphorylated and inactivated following T cell antigen receptor triggering in T cells. Thus, this PTPase may promote T cell activation by reducing the dephosphorylation of its substrate(s) on T cell antigen receptor triggering.