The Minimal Essential Core of a Cysteine-based Protein-tyrosine Phosphatase Revealed by a Novel 16-kDa VH1-like Phosphatase, VHZ*

The smallest active protein-tyrosine phosphatase yet (only 16 kDa) is described here and given the name VHZ for VH1-like member Z because it belongs to the group of small Vaccinia virus VH1-related dual specific phosphatases exemplified by VHR, VHX, and VHY. Human VHZ is remarkably well conserved through evolution as it has species orthologs in frogs, fish, fly, and Archaea. The gene for VHZ, which we designate as DUSP25, is located on human chromosome 1q23.1 and consists of only two coding exons. VHZ is broadly expressed in tissues and cells, including resting blood lymphocytes, Jurkat T cells, HL-60, and RAMOS. In transfected cells, VHZ was located in the cytosol and in other cells also in the nucleoli. Endogenous VHZ showed a similar but more granular distribution. We show that VHZ is an active phosphatase and analyze its structure by computer modeling, which shows that in comparison with the 185-amino acid residue VHR, the 150-residue VHZ is a shortened version of VHR and contains the minimal set of secondary structure elements conserved in all known phosphatases from this class. The surface charge distribution of VHZ differs from that of VHR and is therefore unlikely to dephosphorylate mitogen-activated protein kinases. The remarkably high degree of conservation of VHZ through evolution may indicate a role in some ancient and fundamental physiological process.

The VH1-like phosphatases (1) are members of the cysteinebased protein-tyrosine phosphatase (PTP) 1 family and contain the extended consensus signature motif DX 27-30 HCX 2 GX 2 R(S/ T/A)X 5 A(Y/F)LM or slight variations thereof. The crystal structures of the catalytic domains of the VH1-like enzymes VHR (2), mitogen-activated protein kinase phosphatase-3 (MKP-3) (3), KAP (4), PTEN (5), and MTMR2 (6) show that they have the same topology and catalytic machinery as other members of the Class I cysteine-based PTP family exemplified by PTP1B (7) and RPTP␣ (8). However, whereas the catalytic cleft of most VH1-like phosphatases is only 6 Å deep to allow access by the shorter phosphoamino acid side chains of phosphoserine (Ser(P)) and phosphothreonine (Thr(P)) to the catalytic Cys at the bottom of the pocket, the deeper (9 Å) pocket of classical phosphatases like PTP1B only permits dephosphorylation of phosphotyrosine. On the other hand, the catalytic pockets of PTEN (5) and MTMR2 (6) are considerably wider to accommodate their unique physiological substrate, the lipid-bound inositol ring with phosphate at the D position.
VH1-like phosphatases are found in all main classes of organisms from bacteria to plants, yeast, insects, worms, and mammals. The human and mouse genomes each encode at least 61 VH1-like enzymes 2 , 11 of which are collectively referred to as the MKPs. In addition to a catalytic VH1-like domain, these MKPs contain a non-catalytic CDC25 homology (CH2) domain for docking with mitogen-activated protein kinases (10), which subsequently are dephosphorylated by the catalytic domain. Another group of 19 small VH1-like phosphatases lack the CH2 domain (in both mice and humans), 2 but some of the phosphatases still dephosphorylate the phosphotyrosine and phosphothreonine residues in the activation loop of mitogen-activated protein kinases in vitro and in cells. In vitro, all of these 30 phosphatases as well as many other VH1-like enzymes dephosphorylate peptides with phosphate on any of three hydroxyl amino acids (serine, threonine, and tyrosine), and they are therefore often referred to as dual specific protein phosphatases (DSPs) (10). The CDC25 phosphatases also show dual specific activity against threonine-and tyrosine-phosphorylated cyclin-dependent protein kinases (11), but they have a different structure and topology (12) and do not share sequence similarity with DSPs apart from the CX 5 R motif common to all Cys-based phosphatases. Thus, the three CDC25s are not members of the VH1-like family (as defined above) but apparently became dual specific phosphatases as the result of convergent evolution. In addition, several VH1-like phosphatases dephosphorylate Tyr(P), Ser(P), and Thr(P) in vitro but have non-protein substrates in vivo. For example, phosphatase and tensin homology deleted on chromosome 10, the related phosphatases TPTE and TPIP, and the myotubularins dephosphorylate the Asp-3 phosphate of phosphatidylinositol (13), * This work was supported by a fellowship from the Spanish Ministries of Education and Culture and by National Institutes of Health Grants AI35603, AI48032, AI53585, AI55741, and CA96949 (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  whereas PIR1 (phosphatase that interacts with RNA/RNP complex 1) (14) and mRNA capping enzyme (RNGTT) dephosphorylate the 5Ј end of RNA transcripts during the mRNA capping process (14,15).
Although much of the early work on DSPs focused on the biochemical and structural aspects of these enzymes, more recent work has begun to unravel the physiological functions of DSPs particularly in invertebrates. For example, the puckered (16) and slingshot (17) genes in Drosophila melanogaster and LIP-1 (18) in Caenorhabditis elegans have non-redundant functions revealed by gene deletion. Our understanding of the physiological roles of DSPs in mammals is much more rudimentary. The two reported DSP knock-out mice, MKP-1 Ϫ/Ϫ (19) and VHX Ϫ/Ϫ (20) animals, had no apparent phenotype. However, new functions in cell lines and new substrates have recently been found for DSPs. For example, the cytoskeletal protein ADF/cofilin was identified as the substrate for the slingshot DSPs in the fruit fly D. melanogaster (17), whereas the hYVH1 (DUSP12) DSP was shown to dephosphorylate glucokinase, an enzyme involved in glucose metabolism (21). Another small DSP, laforin (EMP2A), associates via a glycogen-binding domain (22) with glycogen, an interaction that is important for cellular health because mutations in laforin cause a severe human disease, progressive myoclonus epilepsy or Lafora disease (23).
A first step toward understanding the physiological roles and possible redundancies within the DSP group of PTPs is the elucidation of the full repertoire of these enzymes expressed in cells and the characterization of their expression patterns and biochemical properties. Here we begin the characterization of a novel VH1-like phosphatase, which we call VHZ for VH1-like member Z.

EXPERIMENTAL PROCEDURES
Antibodies and Reagents-The 9E10 hybridoma producing the monoclonal antibody that recognizes the c-Myc epitope tag was from the American Type Culture Collection (Manassas, VA). Polyclonal rabbit antisera were raised against the recombinant GST fusion protein of full-length VHZ protein and were used at 1:1,000 dilution.

FIG. 1. Amino acid sequence and Northern blots for VHZ.
A, deduced amino acid sequence of VHZ compared with that of VHX, VHR, and CDC14. Amino acid residues that are identical to those in human VHZ are shaded in gray. B, amino acid sequence comparison between human VHZ and presumed species orthologs from the puffer fish (T. rubripes), frog (X. tropicalis), and the Archaea T. kodakaraensis. Amino acid residues that are identical to those in human VHZ are shaded in gray. The right column of panels shows Nomarski-differential contrast images of the same cells. E, confocal microscopy of 293T cells stained with the anti-VHZ antibody followed by fluorescein isothiocyanate-conjugated anti-rabbit IgG. The second and fourth panels are Nomarski-differential contrast images of the same cells. The white arrows point to staining of the nucleolus. F, confocal microscopy of HeLa cells triple stained with the anti-VHZ antibody followed by fluorescein isothiocyanate-conjugated anti-rabbit IgG (green), anti-tubulin (red), and with 4Ј6-diamidino-2phenylindole for DNA (blue). The fourth column of panels shows merges of the three colors.
Expression Plasmids for VHZ-The cDNA for VHZ was cloned into the 5Ј EcoRI to 3Ј XhoI sites of the pcDNA3.1myc-his vector leaving the epitope tags in the C terminus of the insert.
p-Nitrophenyl Phosphate (pNPP) Dephosphorylation Assay-The hydrolysis of pNPP was assayed at 30°C for 30 min with 20 g of enzyme in 100 l of 0.1 M bis-Tris, pH 6.0, 150 mM NaCl, 1 mM dithiothreitol. The reaction was initiated by the addition of various concentrations of pNPP (ranging from 0.2 to 10 K m ) to the reaction mixtures and terminated by the addition of 200 l of Biomol reagent. Nonenzymatic hydrolysis of the substrate was corrected by measuring the control without the addition of enzyme. After a 30-min incubation, the absorbance at 620 nm was determined using a PowerWaveX340 microplate spectrophotometer (Bio-Tek Instruments, Inc.) and converted to a molar amount of product using a standard curve. The K m and K cat values were determined by fitting the data to the Michaelis-Menten equations using nonlinear regression and the software program GraphPad Prism® (version 4.0).
Cells and Transfections-293T, COS-1, HeLa, and other cell lines 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. Peripheral blood lymphocytes (ϳ80% T cells) were obtained from venous blood from healthy donors (Red Cross Blood Bank, San Diego, CA); Ficoll gradient centrifugation was performed, and monocytes were removed by adherence to plastic at 37°C for 2 h. 293T and COS-1 cells were transfected with a total of 5-10 g of DNA by lipofection. Empty vector was added to control samples to make a constant amount of DNA in each sample. Cells were used for experiments 24 -48 h after transfection.
Immunoprecipitation, SDS-PAGE, and Immunoblotting-These procedures were performed as before (24 -26). Briefly, cells were lysed in 20 mM Tris/HCl, pH 7.5, 150 mM NaCl, 5 mM EDTA containing 1% Nonidet P-40, 1 mM Na 3 VO 4 , 10 g/ml aprotinin and leupeptin, 100 g/ml soybean trypsin inhibitor, and 1 mM phenylmethylsulfonyl fluoride and clarified by centrifugation at 15,000 rpm for 20 min. The clarified lysates were preadsorbed on protein G-Sepharose and then incubated with antibody for 2 h; this was followed by the addition of protein G-Sepharose beads. Immune complexes were washed three times in lysis buffer: once in lysis buffer with 0.5 M NaCl, again in lysis buffer, and then suspended in SDS sample buffer. Proteins resolved by SDS-PAGE were transferred electrophoretically to nitrocellulose filter and were then immunoblotted with 1:1,000 anti-VHZ or optimal dilutions of monoclonal antibodies followed by anti-rabbit or anti-mouse Ig-peroxidase, and the blots were developed by the enhanced chemiluminescence technique (ECL kit, Amersham Biosciences) according to the manufacturer's instructions.
Indirect Immunofluorescence and Confocal Microscopy-Immunofluorescence staining of transfected or endogenous VHZ was performed as described previously (25)(26)(27)(28). 293T cells were washed in phosphatebuffered saline, fixed in 3.7% formaldehyde, permeabilized with 0.1% saponin in phosphate-buffered saline, and blocked in 2.5% normal goat serum, 0.1% saponin in phosphate-buffered saline for 30 min at room temperature. Primary and secondary antibody were diluted in the same buffer and incubated with the cells for 1 h each at room temperature. After washing three times with phosphate-buffered saline, the cells were mounted onto glass slides and viewed under a confocal laser scanning microscopy MRC-1024 (Bio-Rad). DNA was stained with TO-PRO-3 iodide (Molecular Probes) at 1:1,000. Nomarski-differential interference contrast images were also taken.
Bioinformatics-A VHZ model was built using a comparative modeling approach based on the structure of the human MAPK phosphatase (Protein Data Bank code 1mkp) and the phosphatase domain of the human cell cycle protein Cdc14 (Protein Data Bank code 1ohe) based on a -fold prediction metaserver alignment (29). Modeling was done by a SWISS-MODEL (30); the same package was used for model visualization. The final model was optimized by the SCRWL program (31). The tree in Fig. 3C was obtained by the STRUCLA server (32) based on several measures of protein structural similarity. Topology diagrams were obtained with the TOPS algorithm (33).

RESULTS AND DISCUSSION
During our attempts to map the entire human set of PTP genes (the "PTPome"), 2 we uncovered a number of uncharacterized cDNAs in the public data bases, including a 150-amino acid open reading frame termed "hypothetical protein FLJ20442" (NCBI number gi12654609) with the canonical PTP consensus sequence HC-X 2 -G-X 2 -R. In the new open reading frame, this sequence, HCALGFGR, is unusual in having a phenylalanine after the conserved glycine, and in this respect it resembles VHR (34), which has a tyrosine at the corresponding position. BLAST searches revealed that the new open reading frame had the highest degree of homology (67% identity to residues 89 -147) with hypothetical protein COS41.7 from the sea squirt Ciona intestinalis and the next highest degree of homology with the CDC14 cell cycle phosphatases from human and many other species. Significant homology (36% identity with residues 55-147) was also seen with PTP9Q22, an uncharacterized, CDC14-like brain-expressed PTP of unknown physiological function. Because the new open reading frame is related to VHR and belongs to the group of small (Ͻ30 kDa) dual specific phosphatases, including the "founding member" Vaccinia virus VH1, and the human VHR, VHX (35), and VHY, 3 we decided to name the new member "VHZ" for "VH1-like member Z." This name also signifies the placement of VHZ last in the list of genomic nomenclature for PTPs, which starts with the larger receptor-like enzymes (e.g. receptor type A, B, C, etc.) and then tends to go toward smaller and smaller enzymes, such as LMW-DSP21 (DUSP21) (36) and VHX (DUSP22) (35). We have given the gene that encodes VHZ the genomic designation DUSP25. 2 Apparent species orthologs of VHZ (Fig. 1B) were found in a variety of organisms from mammals (e.g. mouse) to amphibians (Xenopus tropicalis; 68.5% identity, 82.2% similarity), fish (Takifugu rubripes; 64.4% identity, 72.6% similarity), insects (D. melanogaster), and even a possible species ortholog in the Archaea Thermococcus kodakaraensis (NCBI numbers gi:18147126; dbj BAB83049.1), which is 30.3% identical and 49.7% similar over 145 residues). These sequences indicate that VHZ is a well conserved and ancient gene.
The human gene that encodes VHZ (DUSP25, NCBI locus identification number 54935) is located at 1q23.1 and consists of a short non-coding exon 1 followed by two coding exons, both of which are represented by 6 -20 expressed sequence tags. To further ensure that this is a real expressed gene, we performed Northern blot analysis with the full-length VHZ as a probe and detected a 1-kb mRNA in all examined tissues ( Fig. 2A). Highest signals were seen in spleen, prostate, colon, adrenal gland, mammary gland, thyroid, and trachea, whereas the lowest signals were seen in uterus and bladder.
Next we cloned the VHZ cDNA into the pcDNA3.1myc-his vector and expressed it in COS cells (Fig. 2B, lane 3). We also created the C95S and R101M mutants, which were expressed at similar levels as the wild-type enzyme (lanes 4 and 5). All three proteins migrated at ϳ18 kDa on SDS gels as expected from the predicted molecular mass of 16 kDa with an additional 2 kDa from the Myc-His tag (15 amino acids). For comparison, a similarly tagged VHY 3 was expressed at comparable levels.
To be able to detect the endogenous VHZ protein, we generated an anti-VHZ antiserum by immunizing rabbits with a GST fusion protein containing full-length VHZ. The resulting antisera reacted very well with the expressed tagged VHZ (Fig.  2C, lane 2) and with a protein of 16 kDa in freshly isolated blood lymphocytes, Jurkat T cells, HL-60 myelomonocytic cells, RAMOS B cells, and the tumor cell lines M45 and MT29.
Confocal microscopy of 293T cells transfected with VHZ and stained with the monoclonal antibody against its epitope tag showed that VHZ had a range of subcellular locations (Fig. 2D). In most cells, VHZ was cytoplasmic excluding the nucleus (top panels), whereas in some cells it was found also in the nucleus (second and third row of panels). A minority of cells had VHZ only in the nucleus (bottom panels). Interestingly, in every cell where VHZ was found in the nucleus, the nucleoli were intensely stained. Endogenous VHZ (Fig. 2E) was also primarily located in the cytosol but in a more granular manner with a faint staining of the nucleolus only in some cells (Fig. 2E, white  arrows). To determine whether this somewhat variable distribution pattern of VHZ had any correlation to the cell cycle as the 50% homology with CDC14 might suggest, we double stained HeLa cells for endogenous VHZ and DNA and focused on cells in various stages of mitosis (Fig. 2F). In interphase cells, VHZ showed a granular distribution throughout the cytosol. The only change in the location of VHZ during mitosis was its entry into the vicinity of the condensed chromosomes upon breakdown of the nuclear envelope before cytokinesis. In this location, VHZ was more diffusely distributed but not significantly enriched. Staining for tubulin to visualize the mitotic spindle showed that VHZ was in the same region but not organized on the microtubules.
Having established that a 16-kDa VHZ protein exists in cells, we turned to ask if this remarkably small DSP protein is a catalytically active phosphatase. The smallest experimentally verified VH1-like phosphatase so far is the 185-amino acid residue VHR (34), which has been crystallized (2) and found to comprise a catalytic domain with only a few extra residues in its N and C termini. VHZ contains only 150 amino acid residues and is therefore smaller than the catalytic domain of VHR.
A GST fusion protein containing full-length VHZ as well as its C95M and R101M mutant were expressed in Escherichia coli and purified. A standard phosphatase assay with p-nitrophenyl phosphate showed that VHZ readily dephosphorylated this substrate with a K m of 1.498 mM and a K cat of 0.0094 s Ϫ1 (Table I). Although the K m is similar to the values for the small DSPs VH1, VHX, and VHY (Table I), the catalysis rate is slow. Nevertheless, VHZ is an active phosphatase.
To better understand how the structural elements and catalytic core of a phosphatase of only 150 amino acid residues might be organized, we generated a computer model of VHZ. This model (Fig. 3A) showed that VHZ consists of the same core of five parallel ␤-sheets flanked by one and three ␣-helices as all other Class I Cys-based PTPs. The main difference between the VHZ and VHR structure seems to be the shortening of several loops between strands and helices as well as significant shortening of the longest helix in VHR. Also the predicted surface topology and surface charge distribution of VHZ (Fig.  3B) differs from those of VHR and VHX (35). VHZ has two acidic protrusions (one on each side of the catalytic pocket), but it lacks a patch of basic charge adjacent to the catalytic pocket found in VHR. Because this patch has been suggested to be important for docking of phospho-Erk with VHR (37), our model suggests that VHZ is unlikely to target this substrate. We want to mention that an alternative model of VHZ, where one of the helices (␣2 in Fig. 3A) is replaced by a ␤-strand (making the core of the protein similar to that of the human MAPK phosphatase) (Protein Data Bank code 1mpk), is also possible and could not be ruled out based on model quality considerations.
A more comprehensive comparison of the predicted structure of VHZ with nine other phosphatases (Fig. 3C) showed that although the central core, which is composed of four parallel ␤-strands (black upside-down triangles) flanked by one and three ␣-helices on each side (black circles), is found in all Cys-based PTPs, many additional elements have been added to this core structure during evolution. These added elements are shown in Fig. 3C as triangles for ␤-sheets and circles for ␣-helices, each color coded according to placement in the overall structure. For example, the classical PTPs (PTP1B, SHP2, and LAR) contain eight ␤-sheets with the four additional ones being both antiparallel (triangles) and parallel (upside-down triangles) to the common four parallel ␤-sheets. They also have additional flanking short ␤-strands and ␣-helices. These features contribute to the deeper catalytic pocket of these classical PTPs, making them strictly tyrosine-specific. In contrast, VHR and VHZ consist essentially only of the common core and thus are close to the likely three-dimensional structure of an archetypal protophosphatase. One could envision an evolutionary sequence of events in which an ancestral Cys-based hydrolase consisting of the minimal essential core of four or five parallel ␤-strands flanked by only one or two ␣-helices on each side was used as a starting point for building a first DSP-like phosphatase, which gave rise to more specialized protein phosphatases and eventually tyrosine-specific PTPs through the addition of more elements. This scenario is supported by the existence of DSP-like enzymes in all kingdoms of life, including Archaea and prokaryotes, whereas "classical" PTPs are found primarily in higher eukaryotes. Interestingly, VHZ appears to be a phylogenetically well conserved enzyme with a probable ortholog even in the Archaea T. kodakaraensis. Thus, it is possible that VHZ represents an ancient stage in the evolution of tyrosine phosphatases. Its presence in the human genome and wide expression in tissues and cell types may indicate that VHZ has retained some important and presumably basic function in cell physiology. On the amino acid sequence level, the closest relatives of VHZ are the CDC14 phosphatases, which are critical regulators of cell division (9). However, the ϳ50% sequence identity with CDC14 is too low for any prediction of VHZ function. This topic will have to be addressed experimentally in the future.