Inhibition of T Cell Antigen Receptor Signaling by VHR-related MKPX (VHX), a New Dual Specificity Phosphatase Related to VH1 Related (VHR)*

A cDNA encoding a novel, human, dual-specific protein phosphatase was identified in the Incyte data base. The open reading frame predicted a protein of 184 amino acids related to theVaccinia virus VH1 and human VH1-related (VHR) phosphatases. Expression VHR-related MKPX (VHX) was highest in thymus, but also detectable in monocytes and lymphocytes. A VHX-specific antiserum detected a protein with an apparent molecular mass of 19 kDa in many cells, including T lymphocytes and monocytes. VHX expression was not induced by T cell activation, but decreased somewhat at later time points. In vitro, VHX dephosphorylated the Erk2 mitogen-activated protein kinase with faster kinetics than did VHR, which is thought to be specific for Erk1 and 2. When expressed in Jurkat T cells, VHX had the capacity to suppress T cell antigen receptor-induced activation of Erk2 and of an NFAT/AP-1 luciferase reporter, but not an NF-κB reporter. Thus, VHX is a new member of the VH1/VHR group of small dual-specific phosphatases that act in mitogen-activated protein kinase signaling pathways.

Phosphate is removed from phosphoproteins by two unrelated classes of protein phosphatases, the serine/threoninespecific phosphatases (PP1, PP2, etc.) and the protein tyrosine phosphatases (PTPases; reviewed in 1,2). The latter group uses a cysteine-based catalytic mechanism (3,4) shared with a broader family of hydrolases, including phosphatases specific for phospholipids (5) and RNA (6). Phosphotyrosine specificity is generally achieved by the depth of the catalytic pocket (5). An exception from this rule is the PTPase subfamily of dual-specific phosphatases (DSPs), 1 which readily dephosphorylates both phosphotyrosine and an adjacent phosphothreonine residue (7)(8)(9). Most DSPs are specific for mitogen-activated protein kinases Erk, Jnk, and p38, which are activated by dual phosphorylation of a tyrosine and threonine residue in the consensus motif Thr-X-Tyr. Accordingly, these DSPs are often referred to as the MAP kinase phosphatases or MKPs.
The first DSP to be cloned was the VH1 protein from Vaccinia virus (10). A related enzyme was subsequently found in mammalian cells and termed VHR, for VH 1-related (11). Both VH1 and VHR differ from other DSPs in that they are much smaller, only 19 and 21 kDa, respectively. VH1 has been reported to dephosphorylate both MAP kinases and Stat1 (12), while VHR appears to be specific for Erk and Jnk (13,14). However, the physiological function of VHR has remained somewhat unclear as it seems to be less efficient than many other MAP kinase-specific DSPs.
We report the identification and initial characterization of a new human gene that encodes a DSP that is much more closely related to VH1 and VHR than to other DSPs. During our work, the nucleotide sequence and predicted open reading frame of this enzyme was deposited by others in GenBank TM under the name MKPX. However, this name is already occupied (15) and cannot be used as such. Instead, we honor this name by proposing the acronym VHX for "VHR-related MKPX."

MATERIALS AND METHODS
Antibodies-The 12CA5 anti-hemagglutinin mAb was from Roche Molecular Biochemicals (Indianapolis, IN). The anti-phosphotyrosine mAb 4G10 was from Upstate Biotechnology, Inc., (Lake Placid, NY) and the OKT3 hybridoma that produces the anti-CD3⑀ mAb was from the American Type Culture Collection (Manassas, VA). The mAb against CD28 was from PharMingen (San Diego, CA). The polyclonal antiextracellular signal-regulated kinase 2 (Erk2) antibody was from Santa Cruz Biotechnology Inc.
Generation of Antisera-The cDNA for VHX was subcloned into the pGEX-4T1 vector and transformed into Escherichia coli strain XL-10. Expression of the GST-VHX protein was induced with 1 mM isopropyl-1-thio-␤-D-galactopyranoside, and the protein purified over glutathione-Sepharose TM 4B (Amersham Biosciences, Inc.) and used for immunization of two rabbits (HTI Inc., Ramona, CA). The resulting antiserum was purified by affinity chromatography on immobilized VHX (without GST).
PTPase Assay and Erk2 Dephosphorylation-The catalytic activity of VHX was measured as described (16). The reaction was for 30 min at 37°C in 200 l of 50 mM Tris/HCl, pH 6.8, 5 mM dithiothreitol with 10 mM p-nitrophenyl phosphate as a substrate. The production of p-nitrophenol was measured colorimetrically as absorbance at 405 nm.
Purified His 6 -Erk2 (17) was incubated with active Mek (Upstate Biotechnology Inc., Lake Placid, NY) in 20 mM MOPS, pH 7.2, 25 mM MgCl 2 , 1 mM ATP, 1 mM dithiothreitol, 5 mM EGTA at 30°C for 30 min. The reaction was stopped by dilution into ice-cold buffer without MgCl 2 and ATP, but with 5 mM EDTA. Subsequently, 200-ng aliquots of * This work was supported by a fellowship from the Spanish Ministries of Education and Culture and by Grants AI35603, AI40552, AI41481, and AI48032 from the National Institutes of Health (to T. M.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. phospho-Erk2 were mixed with VHX (cleaved from GST-VHX by thrombin) and incubated for 1-30 min at 37°C. Reactions were terminated by addition of SDS sample buffer, and the dephosphorylation of Erk2 analyzed by anti-phosphotyrosine immunoblotting.
RT-PCR-Amplification of a 293-bp fragment of VHX cDNA was done with the 5Ј-upstream primer 5Ј-GAG CTG CCT TGT ACA CTG CCT GGC CGG GG-3Ј and the 3Ј-downstream primer 5Ј-GGC CCA GAA CTT CAG AAT TCC TGG AGC GGC C-3Ј from the Human MTC TM Panel II (CLONTECH, Palo Alto, CA) and Human Blood Fractions MTC TM Panel (CLONTECH) with Titanium TM Taq polymerase (CLONTECH) as recommended by the manufacturer. The manufacturer's primers for G3PDH were used as recommended.
Plasmids and Site-directed Mutagenesis-The cDNA for VHX was also subcloned into the pEF/HA vector (18), which adds a hemagglutinin (HA) tag to the N terminus of the insert. To generate a catalytically inactive mutant of VHX, the codon for Cys-88 was changed into a codon for serine in the pGEX-VHX plasmid using the Transformer TM sitedirected mutagenesis kit as recommended by the manufacturer (CLONTECH, Palo Alto, CA). The mutation was verified by nucleotide sequencing, and the cDNA subcloned into pEF/HA.
Cells and Transfections-Jurkat T leukemia and murine thymoma cells were kept at logarithmic growth in RPMI 1640 medium supplemented with 10% fetal calf serum, 2 mM L-glutamine, 1 mM sodium pyruvate, nonessential amino acids, and 100 units/ml each penicillin G and streptomycin. These cells were transiently transfected with a total of 5-15 g of DNA by electroporation at 950 F and 240 V (14). Empty vector was added to control samples to make a constant amount of DNA in each sample. Cells were used for experiments 24 h after transfection. Peripheral blood lymphocytes (ϳ80% T cells) were obtained from venous blood from healthy donors (Red Cross Blood Bank, San Diego, CA), by Ficoll gradient centrifugation and removal of monocytes by adherence to plastic at 37°C for 2 h. The adherent cells were also recovered.
Immunoblotting-Proteins resolved by SDS-PAGE were transferred electrophoretically to nitrocellulose filter, which were immunoblotted as before (19 -22) with optimal dilutions of mAbs, followed by antimouse-Ig-peroxidase, and the blots developed by the enhanced chemiluminescence technique (ECL kit, Amersham Biosciences, Inc.) according to the manufacturer's instructions.

Identification of a Novel Dual Specificity Phosphatase,
VHX-To identify novel protein phosphatases, the Incyte data base was screened for open reading frames encoding proteins with homology to PTPases. One such cDNA encoded a 184amino acid-residue polypeptide that had the canonical PTPase signature motif, His-Cys-X 5 -Arg, at positions 87-94. In comparison with other PTPases, the new open reading frame had highest similarity to human VHR and the Vaccinia virus VH1 (Fig. 1A). Very recently, the same open-reading frame was deposited in GenBank TM under the (unpublished) name MKPX (accession number AF165519, gi9294744). However, since this name is already occupied by another unrelated DSP, MKP-X (15), also known as Pyst2 (15,24), B59 (25), or DUSP7 (26), we propose the name VHX for VHR-related MKPX for the new protein. This name honors the prior GenBank TM deposition by Gu and colleagues, but circumvents the confusing duplication of names. The name VHX also does not imply a function, which currently is unknown.
Expression of VHX in Tissues and Cells-To determine in which tissues and cell types VHX is expressed, we used a RT-PCR approach with gene-specific primers. An amplification product of the appropriate size (293 bp) was best obtained from Jurkat cells and thymus, while weaker bands were seen in spleen, prostate, testis, and peripheral blood lymphocytes (Fig.  1B). The band from Jurkat was sequenced to verify that the correct cDNA fragment had been obtained. As a control, we amplified glyceraldehyde-3-phosphate dehydrogenase. Because of the lymphoid bias of the result, we repeated the RT-PCR with a lymphoid cell cDNA library panel (Fig. 1B, right hand panels), which showed that monocytes, B lymphocytes, as well as CD4-positive and CD8-positive T cells express VHX. Interestingly, activated cells contained less VHX relative to the glyceraldehyde-3-phosphate dehydrogenase control than resting cells (compare lanes 3-7 with 8 -11).
Detection of Endogenous VHX Protein-A GST fusion protein containing the entire open reading frame of VHX was generated, purified, and used for immunization of two rabbits. The resulting antisera reacted well with HA-tagged VHX expressed in Jurkat or COS cells ( Fig. 2A). In contrast, the prebleed serum from the same rabbit did not recognize this protein. As a control, the HA-tagged VHX was readily detected by the anti-HA mAb (Fig. 2B). Immunoblotting of a number of cell lysates from human and murine leukemic cells lines detected an endogenous protein of the expected size, 19 kDa (Fig. 2C) level of VHX protein (Fig. 2D) despite relatively abundant VHX mRNA (Fig. 1B, lane 11).
T Cell Activation Does Not Induce VHX Expression-Since most DSPs are inducible genes encoded by immediate early genes, we wanted to determine whether this was also true for VHX. Normal blood T cells were treated with 5 g/ml of the activating anti-CD3⑀ mAb OKT3 plus the anti-CD28 mAb for 0 -48 h and then immunoblotted with the anti-VHX antibody. The amount of VHX immunoreactivity was not induced in these experiments, but rather showed a decline at the 24 and 48 h time points (Fig. 2E), as also indicated by the lower presence of VHX mRNA in activated cells (Fig. 1B, lanes 8 -11). Thus, VHR is constitutively expressed in resting T cells and is not induced.
Topology of the Substrate Binding Surface of VHX-To gain some initial insight into the biology of VHX, we first generated a computer model of the protein by comparative modeling using the crystal structure of VHR (27) as a template. The model showed that the amino acid chain of VHX readily adopts the same mixed ␣-helical and ␤-sheet tertiary structure as VHR with Cys-88 in the bottom of the catalytic pocket. A GRASP representation of the model (Fig. 3A), showed that VHX presents a largely basic surface surrounding the catalytic center. In VHR the catalytic cleft is in a similarly basic groove, but bordered by an acidic ridge. In VHX, this ridge is missing. The Vaccinia virus VH1 also lacks the acidic ridge, and it shares a second basic patch with VHX. Based on this comparison, all three enzymes are likely to prefer substrates with acidic resi-dues around the phosphotyrosine target residue, but VHX is likely to display a substrate preference more similar to that of VH1 than VHR.
Catalytic Activity of the VHX Protein-To formally demonstrate that VHX is a catalytically active phosphatase, we first measured its capacity to dephosphorylate the general PTPase substrate p-nitrophenyl phosphate. This substrate was readily converted into the yellow p-nitrophenol by GST-VHX, but not by control GST, in a dose-dependent manner (Fig. 3A) and with a pH optimum of 7.0 (not shown). Thus, VHX is a catalytically active phosphatase.
VHX belongs to the subgroup of dual-specific PTPases, many of which are specific for dually tyrosine and threonine phosphorylated MAP kinase family members. To directly test if VHX has phosphatase activity against the dual-phosphorylated Erk2 protein, we incubated purified Erk2 with GST-Mek and then treated the resulting phospho-Erk with GST-VHX. As shown in Fig. 3C, 200 ng of VHX dephosphorylated 100 ng of Erk2 nearly completely within a few minutes at 37°C. The same amount of VHR also dephosphorylated Erk2, but at a considerably slower rate. In contrast, the low molecular weight PTPase (LMPTP), an unrelated enzyme of similar size, did not dephosphorylate Erk2 (lanes 8 and 9). Control blots for Erk2 amount (Fig. 3C, lower panel) also revealed that VHX caused a rapid shift of the apparent M r of Erk2 from 44 to 42 kDa, while VHR did not. This M r shift is caused by loss of phosphothreonine. To study this phenomenon better, we performed similar experiments that also included a shorter time point (Fig. 3, D  and E), which showed that VHX caused appearance of the faster migrating Erk2 protein within 30 s and a complete conversion by 5 min. Again, VHR did not convert Erk2 to the lower M r form.

Active VHX Inhibits MAP Kinase and Anti-CD3 Plus Anti-CD28-induced Activation of a NFAT/AP-1 Reporter Gene
Taken from the Interleukin-2 Gene-Since endogenous VHX was present in T lymphocytes, we decided to test VHX for biological activity in T cell antigen receptor signaling pathways. For these experiments, we first co-expressed VHX with a Myc-tagged Erk2 and measured its activation in response to anti-CD3⑀ mAbs. As shown in Fig. 4A, Erk2 activation was severely reduced by VHX, but not by another PTPase (PRL-2), which was expressed at higher levels than VHX (lower panel). At the shown 5-min time point, the inhibition was 42%, and in time-course experiments the inhibition was similar (40 -60%) at all time points (1, 2, 3, 5, and 10 min).
Next, we co-expressed VHX with a sensitive luciferase reporter gene, in which luciferase transcription is under the control of a tandem NFAT/AP-1 element taken from the interleukin-2 gene promoter. This reporter responds to T cell antigen receptor ligation with anti-CD3⑀ mAbs alone or together with co-ligation of the CD28 co-stimulatory molecule (14). VHX inhibited the response in a dose-dependent manner with 50% inhibition at 0.5-1 g of plasmid (Fig. 4B). The expression of VHX protein correlated with DNA dose (insert). In contrast to the effect of catalytically active VHX, expression of the catalytically inactive VHX-C88S mutant did not inhibit reporter gene activation (Fig. 4C), suggesting that the catalytic activity of VHX was required for inhibition. The C88S mutant protein was expressed at the same level as the wild-type VHX (insert).
VHX Does Not Affect Anti-CD3 Plus Anti-CD28-induced Activation of a NF-B-We next tested VHX for effects in a parallel signaling pathway emanating from the T cell antigen receptor and the CD28 co-receptor, namely the activation of NF-B-driven transcription. In these experiments, VHX failed to have any significant effects over the same 0.5-10-g range as used in the NFAT/AP-1 experiments (Fig. 4D). Only at the highest concentration (10 g) was there a slight reduction in the reporter gene response, which is unlikely of any significance. We conclude that VHX does not affect signaling pathways required for NF-B activation in T cells.
Subcellular Location of VHX-To determine where in the cell VHX is located, we first transfected Jurkat T cells with the VHX expression plasmid and stained them with a fluorescein isothiocyanate-conjugated anti-HA mAb. When these cells were viewed under a confocal microscope, it was clear that most of the fluorescence was in the cytoplasm of the cells (Fig. 5, B and C). Cells transfected with empty vector did not display any fluorescence at all (Fig. 5A). To verify that this subcellular location was also true for endogenous VHX, we stained blood T lymphocytes with preimmune serum (Fig. 5D) or the affinitypurified anti-VHX plus TRITC-conjugated anti-rabbit IgG (Fig.  5, E and F). With this staining protocol, the immunofluores- cence was also cytoplasmic. We conclude that VHX is not nuclear like many DSPs, but cytoplasmic. DISCUSSION In this paper, we describe and begin the characterization of a novel DSP, VHX, which is related to Vaccinia VH1 and the human VHR. Although, the three enzymes are quite similar in size and share amino acid similarity, they also differ in many significant respects. First, VHX has a longer C-terminal tail, in which there is a possible tyrosine phosphorylation site. Second, VHX has a different expression profile than VHR. VHX is expressed best in the thymus, which is lowest in VHR (14). Neither enzyme is induced, but VHX decreases at 24 and 48 h of T cell activation, while VHR remains unchanged (14). Third, VHX differs from VHR in that it is much more efficient toward dually phosphorylated Erk2. In contrast to VHR, which dephosphorylates phosphotyrosine much faster than phosphothreonine (28), VHX appears to readily dephosphorylate Thr-183 in Erk2, as well as Tyr-185. Thus, the apparent M r of Erk2, as seen by anti-phosphotyrosine immunoblotting, shifts from 44 to 42 kDa within the first minutes of co-incubation before it disappears. VHR, on the other hand, maintains Erk2 as a 44-kDa band that gradually weakens. Thus, VHX is a bona fide DSP, while VHR has evolved toward a more tyrosine-selective catalytic activity.
The group of small DSPs defined by VHR and VHX differs from the MKPs that constitute the majority of DSPs, in that they lack obvious non-catalytic targeting or regulatory motifs. This explains their much smaller molecular mass, only 21 and 19 kDa, respectively, compared with over 40 kDa for most MKPs. Nevertheless, two recent papers (14,29) suggested that the Erk1 and Erk2 kinases are physiological targets for VHR. We also found that activation of Jnk in intact T cells was diminished by co-expression of VHR, while the catalytically inactive VHR-C124S had little effect (14). In contrast to Erk and Jnk, activation of p38 kinase was not affected by VHR (14,29). It remains unclear, however, if VHR dephosphorylates MAP kinases in cells that normally express many other, and more efficient, MKPs. In this report, we compared the new enzyme VHX with VHR and found that VHX is considerably more efficient as a Erk phosphatase and also removes phosphate from Thr-183 as well as from Tyr-185. Transfected VHX also inhibits activation of Erk2 in Jurkat T cells. Nevertheless, the physiological substrate(s) for VHX in intact cells remains to be established. Our results with reporter genes suggest that VHX can dephosphorylate a protein that is involved in T cell antigen receptor-mediated activation of gene transcription driven by a NFAT/AP-1 module, but not in pathways to NF-B driven transcription. This result is also compatible with Erk2 being a target, but since all three types of MAP kinases (Erk, Jnk, and p38) are activated by the T cell antigen receptor and participate in activation of components of AP-1, it is premature to conclude that Erk2 is the only substrate. This is under further investigation in our laboratory.