Molecular cloning and characterization of a novel dual specificity phosphatase, LMW-DSP2, that lacks the cdc25 homology domain.

A novel dual specificity phosphatase (DSP) designated LMW-DSP2 was cloned with a combination of reverse transcription-polymerase chain reaction and cDNA library screening strategies. The LMW-DSP2 open reading frame of 194 amino acids contained a single DSP catalytic domain but lacked the cdc25 homology domain, which is conserved in most known DSPs. Northern blot and reverse transcription-polymerase chain reaction analyses revealed that LMW-DSP2 was specifically expressed in testis. Recombinant LMW-DSP2 protein exhibited phosphatase activity toward an artificial low molecular weight substrate para-nitrophenyl phosphate, and the activity was inhibited completely by sodium orthovanadate but not sodium fluoride, pyrophosphate, and okadaic acid. The substitution of critical amino acid residues, aspartic acid and cysteine, resulted in a dramatic reduction of phosphatase activity. Transient transfection of LMW-DSP2 in COS7 cells resulted in the expression of a 21-kDa protein, and the phosphatase was shown to be distributed in both the cytosol and the nucleus. LMW-DSP2 dephosphorylated and deactivated p38, to a higher extent, and stress-activated protein kinase (SAPK)/c-Jun N-terminal kinase (JNK), but not extracellular signal-regulated kinase 1/2 mitogen-activated protein kinases, in transfected COS7 cells and in vitro. Interestingly, mutation in a conserved docking motif of p38 and SAPK/JNK as well as in a cluster of aspartic acids of LMW-DSP2 did not affect the deactivation of the mitogen-activated protein kinases by LMW-DSP2. Furthermore, the binding between LMW-DSP2 and p38 and SAPK/JNK was also not disrupted by such mutations. Among the DSPs lacking the cdc25 homology domain, LMW-DSP2 is the first one that dephosphorylates and deactivates p38 and SAPK/JNK.

A novel dual specificity phosphatase (DSP) designated LMW-DSP2 was cloned with a combination of reverse transcription-polymerase chain reaction and cDNA library screening strategies. The LMW-DSP2 open reading frame of 194 amino acids contained a single DSP catalytic domain but lacked the cdc25 homology domain, which is conserved in most known DSPs. Northern blot and reverse transcription-polymerase chain reaction analyses revealed that LMW-DSP2 was specifically expressed in testis. Recombinant LMW-DSP2 protein exhibited phosphatase activity toward an artificial low molecular weight substrate paranitrophenyl phosphate, and the activity was inhibited completely by sodium orthovanadate but not sodium fluoride, pyrophosphate, and okadaic acid. The substitution of critical amino acid residues, aspartic acid and cysteine, resulted in a dramatic reduction of phosphatase activity. Transient transfection of LMW-DSP2 in COS7 cells resulted in the expression of a 21-kDa protein, and the phosphatase was shown to be distributed in both the cytosol and the nucleus. LMW-DSP2 dephosphorylated and deactivated p38, to a higher extent, and stress-activated protein kinase (SAPK)/c-Jun N-terminal kinase (JNK), but not extracellular signal-regulated kinase 1/2 mitogen-activated protein kinases, in transfected COS7 cells and in vitro. Interestingly, mutation in a conserved docking motif of p38 and SAPK/JNK as well as in a cluster of aspartic acids of LMW-DSP2 did not affect the deactivation of the mitogenactivated protein kinases by LMW-DSP2. Furthermore, the binding between LMW-DSP2 and p38 and SAPK/JNK was also not disrupted by such mutations. Among the DSPs lacking the cdc25 homology domain, LMW-DSP2 is the first one that dephosphorylates and deactivates p38 and SAPK/JNK. Extracellular signal-regulated kinase (ERK), 1 stress-activated protein kinase (SAPK)/c-Jun N-terminal kinase (JNK) and p38/RK/ CSBP (p38) are distinct classes of mitogen-activated protein (MAP) kinases. ERK is activated mainly by a variety of growth factors and phorbol esters and is associated with cellular proliferation and differentiation. SAPK/JNK and p38 are activated by extracellular stresses such as UV irradiation, osmotic stress, and inflammatory cytokines but are poorly activated by growth factors and phorbol esters. Activation of these protein kinases leads to a variety of responses such as gene expression, cell proliferation, differentiation, cell cycle arrest, apoptosis, early development, etc., depending on the cell type (1)(2)(3)(4)(5)(6)(7)(8).
As a part of the studies on DSP diversity and cellular function, we have been trying to clone novel members of the DSP family. A combination of RT-PCR and cDNA library screening revealed a novel clone designated LMW-DSP2 containing a single catalytic domain but lacking the cdc25 homology domain. Despite the absence of a putative common docking site for MAP kinases, which is normally located in the cdc25 homology domain, LMW-DSP2 dephosphorylated and deactivated p38, to a higher extent, and SAPK/JNK, but not ERK1/2, in transfected COS7 cells and in vitro. Further analyses using various mutants suggested that LMW-DSP2 was a member of a distinct class of DSPs and that the mode of its dephosphorylation and deactivation action toward MAP kinases was different from that of other DSPs reported.
PCR Amplification and cDNA Cloning of Mouse LMW-DSP2-Degenerate oligonucleotide sense and antisense primers were based on the consensus sequences for two conserved amino acid stretches within the catalytic domains of low molecular weight type DSPs: ITH(V/I)(L/V/ I)NA and IXPNXGF. 2 Random-primed cDNA (up to 100 ng) from * 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.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AF237619.
Cell Culture, Transfection, Cell Lysis, and Western Blotting-COS7 cells were inoculated at a density of 2 ϫ 10 5 cells/6-cm dish in Dulbecco's modified Eagle's medium containing 10% fetal calf serum and transfected with expression plasmids by the modified calcium phosphate precipitation method (37). Prior to stimulation, cells were serumstarved for at least 18 h. The transfected cells were lysed as described (34) and subjected to SDS-polyacrylamide gel electrophoresis (38) followed by blotting onto nitrocellulose membranes (Hybond C ϩ , Amersham Pharmacia Biotech). The membranes were probed with the indicated antibodies and visualized with an enhanced chemiluminescence (ECL) detection kit (Amersham Pharmacia Biotech).
MAP Kinase Assay-The MAP kinase assay was done by using a nonradioactive p44/42 MAP kinase assay kit, SAPK/JNK assay kit, and p38 MAP kinase kit (Cell Signaling Technology). All the procedures were essentially based on the manufacturer instructions. Briefly, COS7 cells were co-transfected with HA-p44 ERK1, HA-p38, or HA-p54 SAPK␤ and varying amounts of Myc-LMW-DSP2. After stimulation, MAP kinases were immunoprecipitated and then subjected to an in vitro kinase assay. Phosphorylated Elk-1, c-Jun, and ATF-2 were detected with specific antibodies.
Subcellular Fractionation and Cell Staining-COS7 cells were transfected with Myc-LMW-DSP2 and fractionated essentially as described (39). For cell staining, transfected COS7 cells were fixed with 3.7% formaldehyde in PBS for 10 min at room temperature. After PBS washing, the cells were permeabilized in 0.5% Triton X-100 in PBS and washed with PBS three times. The cells were successively incubated with an anti-Myc antibody and anti-mouse IgG secondary antibody in 3% bovine serum albumin in PBS. After three washes with PBS and two washes with Milli-Q water, coverslips were mounted. Fluorescence was viewed with an Olympus fluorescence microscope.
Northern Blot and RT-PCR Analyses-Total RNA was prepared from the indicated tissues of male mice except for the mammary gland from lactating female mice with ISOGEN reagent (Nippon Gene, Tokyo, Japan). Poly(A) ϩ RNA was isolated using OligotexTM-dT30 super (TaKaRa). Digoxigenin-conjugated dUTP-labeled LMW-DSP2 riboprobe was generated by T7 RNA polymerase (Roche) transcription of the linearized LMW-DSP2 open reading frame. All procedures were based on manufacturer protocols.
RT , C88S (C/S) and D57A (D/A) mutants were expressed in Escherichia coli and purified on glutathione-Sepharose. Indicated amounts of the proteins were assayed for phosphatase activity. The reaction was terminated by the addition of 1 N NaOH, and the absorbance at 405 nm was determined. B, 4 g of GST-LMW-DSP2 wild type was incubated with para-nitrophenyl phosphate in the absence or presence of 1 mM sodium orthovanadate (vanadate), 10 mM sodium fluoride (NaF), 10 mM sodium pyrophosphate (pyrophosphatase), or 300 nM okadaic acid, and processed as above. The values are shown as means Ϯ S.E. of three independent assays. family, but it lacks the cdc25 homology domain and accordingly has a low molecular weight (12), suggesting that a distinct class of the DSP family is present. To look for other members of this low molecular weight-type of DSP family, we searched the data bases and found some sequences of various species that matched the criteria. Alignment of the amino acid sequences of a catalytic domain revealed relatively well conserved amino acid stretches. 2 Degenerate primer sets (see "Experimental Procedures") were then designed, and RT-PCR amplification was performed using poly(A) ϩ RNA prepared from various mouse tissues as a template. PCR clone number 2 (LMW-DSP2) exhibited sequence similarities but was not identical to any previously known DSPs or protein-tyrosine phosphatases (PTPs). Using the LMW-DSP2 PCR-generated cDNA fragment, we isolated a full-length cDNA clone from a mouse testis cDNA library and characterized it by sequencing. Cells were serum-starved and stimulated with epidermal growth factor (50 nM) for ERK1 activation or anisomycin (10 g/ml) for p38 and SAPK/JNK activation. Cells were lysed, and the respective MAP kinases were immunoprecipitated with an anti-HA antibody followed by immunoblotting with anti-phospho MAP kinase antibodies (top panels). The same membranes were reprobed with an anti-HA antibody (middle panels). The expression of Myc-tagged LMW-DSP2 was assessed by immunoblotting (bottom panels). B, the intensity of MAP kinase phosphorylation in A was densitometrically quantified. The phosphorylation level of MAP kinase in the mock transfectant was set as 100%. The values are shown as means Ϯ S.E. of three independent experiments. C, COS7 cells were transiently transfected with 2 g of HA-ERK1, HA-p38, or HA-SAPK/JNK together with the indicated amounts of Myc-LMW-DSP2 wild type and processed as above. MAP kinases were immunoprecipitated and subjected to an in vitro MAP kinase assay. The expression of Myc-LMW-DSP2 in total cell lysate (TCL) was assessed by immunoblotting. D, the intensity of MAP kinase activity in C was densitometrically quantitated. MAP kinase activity in the mock transfectant was set as 100%. The values are shown as means Ϯ S.E. of three independent experiments. E and F, the same experiments as described for C and D were performed except that MKP-4 or hVH5 was used instead of LMW-DSP2.

FIG. 5. In vitro dephosphorylation of p38 and SAPK/JNK by recombinant GST-LMW-DSP2.
A, COS7 cells were transfected with expression plasmids for HA-p38 or HA-SAPK/JNK and stimulated with anisomycin (10 g/ml) for 30 min after serum starvation. HA-p38 and HA-SAPK/JNK were immunoprecipitated, washed with lysis buffer, and then subjected to an in vitro dephosphorylation assay. After termination of the incubation, proteins were separated by SDSpolyacrylamide gel electrophoresis and analyzed with anti-phospho-p38 or antiphospho-SAPK/JNK antibody. The same blots were reprobed with an anti-HA antibody after stripping. WT, wild type; C/S, cysteine 88 to serine mutant. B, COS7 cells were transfected with HA-p38 or HA-SAPK/JNK and processed as above. The blots were successively incubated with anti-phospho-tyrosine (top panels), anti-phospho-threonine (middle panels), and anti-HA antibodies (bottom panels).

FIG. 4. continued A Novel Dual Specificity Phosphatase, LMW-DSP2 27579
The cDNA and deduced amino acid sequences of LMW-DSP2 are shown in Fig. 1A. The open reading frame encoded a protein of 184 amino acids with a predicted molecular mass of ϳ21 kDa. The deduced amino acid sequence of LMW-DSP2 had nearly matched the extended active site sequence motif DX26(V/L)X(V/I)HCXAG(I/V)SRSXT(I/V)XXAY(L/I)M (where X is any amino acid) conserved in DSPs (Fig. 1A) but lacked the cdc25 homology domain, which is present in all known DSPs except for human VHR (Fig. 1C). Alignment of the DSP catalytic domain of LMW-DSP2 exhibited 33-41% identity to that of other known DSPs (Fig. 1, B and C). The two amino acids of LMW-DSP2, Asp-57 and Cys-88, are likely to participate in the catalytic mechanism of DSP activity (Fig. 1A).
Catalytic Activity of LMW-DSP2 in Vitro-To examine whether LMW-DSP2 has phosphatase activity, LMW-DSP2 wild-type GST fusion proteins were assayed for enzymatic activity against the well known artificial phosphatase substrate para-nitrophenyl phosphate. GST-LMW-DSP2 wild-type hydrolyzed para-nitrophenyl phosphate in a dose-dependent fash-ion ( Fig. 2A). The substitution of cysteine 88 to serine (LMW-DSP2 C/S) resulted in a complete loss of catalytic activity. Furthermore, substitution of aspartic acid 57 to alanine (LMW-DSP2 D/A) also resulted in a dramatic reduction in phosphatase activity, but slight activity was retained. The catalytic activity of LMW-DSP2 was strongly inhibited by a potent tyrosine phosphatase inhibitor, sodium orthovanadate at the concentration of 1 mM, but not sodium fluoride, pyrophosphate, or okadaic acid (Fig. 2B).
Expression and Subcellular Localization of LMW-DSP2 in COS7 Cells-Wild-type as well as catalytically inactive mutants of Myc-LMW-DSP2 were constructed and transiently transfected into COS7 cells, and then the expression was assessed by Western blotting. LMW-DSP2 was detected as a 21-kDa protein band (Fig. 3A), which was well consistent with the predicted molecular mass. Indirect immunofluorescence cell staining showed that Myc-LMW-DSP2 localized in both the cytosol and the nucleus in COS7 cells and was enriched especially in the perinuclear regions (Fig. 3B). This subcellular localization of the phosphatase was confirmed further by biochemical cell fractionation, which revealed that LMW-DSP2 was distributed in both the cytosol, to a higher extent, and the nucleus (Fig. 3C).
Dephosphorylation and Deactivation of p38 and SAPK/JNK by LMW-DSP2-Individual DSPs have their own substrate specificity for MAP kinases (32). To examine the substrate specificity of LMW-DSP2, we tested the activity of LMW-DSP2 against MAP kinases in cultured cells. COS7 cells were co- transfected with each of the HA-tagged MAP kinases and each of the Myc-LMW-DSP2 wild-type, C88S, or D57A mutants. After stimulation of the cells by the appropriate agonists to activate MAP kinases (50 nM epidermal growth factor for 20 min for MAPK/ERK and 10 g/ml anisomycin for 30 min for p38 and p54 SAPK/JNK), HA-tagged MAP kinases were immunoprecipitated and subjected to immunoblot analysis using anti-phospho (activated) MAP kinases. LMW-DSP2 wild type dephosphorylated p38 and SAPK/JNK but not MAPK/ERK (Fig. 4A, top panels). On the contrary, the catalytically inactive C88S and D57A mutants exhibited no dephosphorylation activity toward the MAP kinases. It was confirmed that nearly the same amounts of MAP kinases were precipitated in each lane (Fig. 4A, middle panels). Comparable amounts of Myc-LMW-DSP2 were expressed in the transfected COS7 cells (Fig.  4A, bottom panels). Densitometric quantitation of the data clearly showed that LMW-DSP2 dephosphorylated p38 more strongly than SAPK/JNK (Fig. 4B).
To establish whether LMW-DSP2 specifically targets p38 and SAPK/JNK, MAP kinase activity was assayed. To obtain a clear impression of the relative effectiveness of LMW-DSP2 to deactivate each MAP kinase, COS7 cells were transfected with a varied amount of LMW-DSP2 plasmid (0 -2 g). After stimulation, the cells were lysed, each HA-tagged MAP kinase was immunoprecipitated, and the activity was assayed in vitro. As clearly shown in Fig. 4, C and D, LMW-DSP2 selectively and strongly deactivated p38, and more than 80% of the MAP activity was lost when the highest amount of plasmid was introduced. SAPK/JNK was also deactivated, to a lesser extent, and ϳ50% of the MAP kinase activity was lost upon the highest expression of LMW-DSP2. The effectiveness of LMW-DSP2 on the activity of p38 and SAPK/JNK was greater than that on the phosphorylation level of the MAP kinases (Fig. 4, A and B), but the relative specificity of LMW-DSP2 for the MAP kinases was similar. On the contrary, ERK1 was not deactivated by LMW-DSP2 within the amounts of plasmid examined. These experiments were controlled by MKP-4 (29, 30) and hVH5 (24,25), where MAPK/ERK and p38 and SAPK/JNK were deactivated, respectively, in a dose-dependent manner of the plasmid introduced (Fig. 4, E and F).
We then examined whether p38 and SAPK/JNK were directly dephosphorylated by LMW-DSP2. COS7 cells, which had been transfected with HA-p38 or HA-SAPK/JNK, were stimulated with anisomycin, and then phosphorylated HA-p38 or HA-SAPK/JNK was immunoprecipitated. Each immune complex was incubated with GST-LMW-DSP2 wild type or the C88S mutant, and the phosphorylation levels of p38 and SAPK/ JNK were analyzed by immunoblotting with anti-phospho-p38 and anti-phospho-SAPK/JNK, respectively. The results clearly showed that both p38 and SAPK/JNK were direct targets of LMW-DSP2 (Fig. 5A). Phosphorylated HA-MAPK/ERK could not be dephosphorylated by incubation with GST-LMW-DSP2 (data not shown).
Next, we studied whether LMW-DSP2 was really a dual specificity phosphatase that could dephosphorylate both Tyr and Thr residues. Phosphorylated p38 and SAPK/JNK were incubated with GST-LMW-DSP2 and immunoblotted first with an anti-phospho Tyr antibody. The same blots were reprobed FIG. 8. Northern blot and RT-PCR analyses of LMW-DSP2. A, 20 g of total RNAs prepared from indicated tissues of male mice except for mammary gland were separated and blotted onto a nylon mem-brane. The membrane was hybridized with a digoxigenin-labeled RNA probe of LMW-DSP2. Ethidium bromide staining of 28S and 18S ribosomal RNAs is shown for the loading control. B and C, the RT-PCR amplification was carried out using a specific primer set for LMW-DSP2 (B) and JNK1, JNK2, JNK3, p38␣, p38␤, p38␥, and p38␦ (C), with cDNA templates prepared from DNase I-digested poly(A) ϩ RNAs of the indicated mouse tissues. Aliquots of the PCR product were run on a 1.0% agarose gel. GAPDH was also analyzed as a control.
with an anti-phospho Thr antibody followed by incubation with an anti-HA antibody for normalization. As shown in Fig. 5B, both phosphorylated Tyr and Thr residues on p38 were dephosphorylated efficiently by LMW-DSP2 in vitro. Dephosphorylation of Tyr on SAPK/JNK was also obvious, but that of Thr was indistinguishable as compared with the control experiment. In vitro dephosphorylation experiments also demonstrated that dephosphorylation of p38 by LMW-DSP2 was greater than that of SAPK/JNK. Thus, it was confirmed that LMW-DSP2 is actually a member of the DSP family.
A Conserved Docking Motif in p38 and SAPK/JNK Is Not Essential for Deactivation by and Binding with LMW-DSP2-Recently, Nishida and co-workers (33) have reported that a conserved docking motif in MAP kinases is essential for the binding and biological functions of their substrates, activators, and regulators and that such a docking motif also existed in the substrates, activators, and regulators including MAP kinase phosphatases/DSPs. To test whether such a mechanism is also true to the deactivation of p38 and SAPK/JNK by LMW-DSP2, we constructed docking site mutants of p38 and SAPK/JNK, the aspartic acid stretches of which were substituted with asparagines (referred to as p38 DN and SAPK/JNK DN, respectively). Although LMW-DSP2 does not contain the cdc25 homology domain that has been shown to be involved in interaction with MAP kinases through a cluster of basic amino acids, two repeated arginine residues are located at its C terminus. The two arginines in LMW-DSP2 were also substituted with methionines and are referred to as LMW-DSP2 RRMM. As shown in Fig. 6, mutation in the common docking sites of p38 and SAPK/JNK revealed apparently no effect on the deactivating efficiency of LMW-DSP2 against the MAP kinases. Mutation in the cluster of arginines of LMW-DSP2 also resulted in no reduction in the phosphatase activity of LMW-DSP2 toward the MAP kinases. Furthermore, co-expression of MAP kinase DN mutants with the LMW-DSP2 RRMM mutant also resulted in deactivation of the MAP kinases. It was also revealed that such mutations in LMW-DSP2 and p38 and SAPK/JNK had no effect on the mutual binding capacity of the molecules (Fig. 7).
Distribution of LMW-DSP2 mRNA-In Northern blot analysis, a 1.0-kilobase mRNA species for LMW-DSP2 was detected specifically in testis but was scarcely in other tissues examined (Fig. 8A). The transcript corresponded to nearly the same size of the full-length cDNA obtained. RT-PCR using DNase I-digested RNA preparations also revealed that LMW-DSP2 was specifically expressed in testis (Fig. 8B).
In the above experiments, LMW-DSP2 was shown to dephosphorylate p38 and SAPK/JNK in cultured cells and in vitro (Figs. 4 -6). To relate the physiological significance of the LMW-DSP2, the expression of p38 and SAPK/JNK was examined by RT-PCR analysis. As shown in Fig. 8C, all the reported members of p38 and SAPK/JNK except for p38␤ were shown to be actually expressed in testis, and all of them were detected in brain. JNK3 and p38␤ were not shown to be expressed in liver. DISCUSSION We have isolated a cDNA clone encoding a novel dual specificity phosphatase, LMW-DSP2. Unlike most of other DSPs/ MKPs hitherto identified, LMW-DSP2 was unique in that it lacked the cdc25 homology domain that is conserved in the N-terminal region of DSPs (32) and accordingly had a smaller molecular mass. VHR and recently cloned TMDP did not have the cdc25 homology domain as well (12,40). Furthermore, we recently have cloned three additional members of this family by RT-PCR strategy, 3 and putative DSPs with a low molecular mass of various species have been registered in the GenBank and Swissprot data bases. It is now conceivable that a distinct subfamily of DSPs is actually present.
Although LMW-DSP2 does not contain the cdc25 homology domain that functions in other DSPs as a regulatory and essential domain in dephosphorylation toward MAP kinases (33,42,43), it could dephosphorylate and deactivate p38 and SAPK/JNK MAP kinases in cultured cells (Fig. 4) and in vitro (Fig. 5). Further experiments with a variety of mutants of the phosphatase and MAP kinases revealed that static interaction between LMW-DSP2 and p38 and SAPK/JNK might not be required for dephosphorylation and deactivation of the MAP kinases by LMW-DSP2 in cultured cells (Fig. 6) and that specific binding between LMW-DSP2 and MAP kinases was not disrupted by mutating putative common docking sites (Fig. 7). It has also been reported that VHR lacking the cdc25 homology domain could dephosphorylate activated MAPK/ERK kinase in cultured cells and in vitro (44). This and our present study suggest the presence of a distinct dephosphorylation mechanism of MAP kinases by DSPs with a low molecular weight. Elucidation of this putative novel dephosphorylation mechanism is currently in progress using a series of mutants in our laboratory.
LMW-DSP2 was shown here to localize evenly in the cytosol and the nucleus (Fig. 3). This subcellular localization of LMW-DSP2 is very similar to that of MKP-5 (31). The subcellular localization of LMW-DSP2 was unchanged after stimulation with anisomycin and NaCl (data not shown). p38 and SAPK/ JNK also localized in both the cytosol and nucleus, and their subcellular localizations were not dramatically affected by stress stimulation, whereas MAPK/ERK translocated into the nucleus upon stimulation (data not shown). The correspondence of the subcellular localization between LMW-DSP2 and its target substrates, p38 and SAPK/JNK, might reflect some physiological relevance.
LMW-DSP2 was shown to be specifically expressed in testis (Fig. 8, A and B), and p38 and SAPK/JNK were also shown to be expressed in testis (Fig. 8C), suggesting that p38 and SAPK/ JNK are physiological substrates of LMW-DSP2. However, we cannot exclude the possibility that LMW-DSP2 might target other unknown molecules specifically expressed in testis. Substrate-trapping mutants of LMW-DSP2 (D57A and C88S) might be useful in the identification of the other putative substrates.
Several protein phosphatases are reported to be expressed specifically in the testis. Serine/threonine protein phosphatase PP12 is abundant in the rat testis and localized in the nuclei of late spermatocytes and early spermatids (45). The PTP Typ was also shown to be expressed specifically in testicular germ cells (46). We also observed that cytosolic PTP20 is expressed abundantly in mouse and rat testis, 4 and we showed that one of the splice variants of PTP36 (PTP36-B) is specifically expressed in testis (11). One of the DSPs, TMDP, was also shown to be predominantly expressed in the testis and skeletal muscle (40). Moreover, we have observed that three novel DSP clones are also specifically expressed in testis. 4 Testis-specific or predominant expression of many protein phosphatases might suggest involvement in a testis-specific cellular event, particularly in spermatogenesis.
In summary, we have cloned and characterized a novel DSP with low molecular weight designated LMW-DSP2. LMW-DSP2 does not seem to belong to known subfamilies of DSPs with respect to structural features and substrate specificity. LMW-DSP2 is specific for p38 and SAPK/JNK, but the dephosphorylation and deactivation mechanism does not require static interaction between the phosphatase and MAP kinases. Thus, LMW-DSP2 may be a novel member of the DSP family acting on the MAP kinase family.