Molecular Cloning and Characterization of a Novel Dual Specificity Phosphatase, MKP-5*

A group of dual specificity protein phosphatases neg-atively regulates members of the mitogen-activated protein kinase (MAPK) superfamily, which consists of three major subfamilies, MAPK/extracellular signal-regulated kinase (ERK), stress-activated protein kinase (SAPK)/c-Jun N-terminal kinase (JNK), and p38. Nine members of this group of dual specificity phosphatases have previously been cloned. They show distinct substrate speci-ficities for MAPKs, different tissue distribution and subcellular localization, and different modes of inducibility of their expression by extracellular stimuli. Here we have cloned and characterized a novel dual specificity phosphatase, which we have designated MKP-5. MKP-5 is a protein of 482 amino acids with a calculated molecular mass of 52.6 kDa and consists of 150 N-terminal amino acids of unknown function, two Cdc25 homology 2 regions in the middle, and a C-terminal catalytic domain. MKP-5 binds to p38 and SAPK/JNK, but not to MAPK/ERK, and inactivates p38 and SAPK/JNK, but not MAPK/ERK. p38 is a preferred substrate. The subcellular localization of MKP-5 is unique; it is present evenly in both the cytoplasm and the nucleus. MKP-5 mRNA is widely expressed in various tissues and organs, and its expression in cultured cells is elevated by stress stimuli. These results suggest that MKP-5 is a novel type of dual specificity phosphatase specific for p38 and SAPK/JNK. Members of the mitogen-activated protein kinase (MAPK) was transformed into Escherichia coli. Positive clones were picked up, and mutagenesis was verified by sequencing. Full-length catalytically inactive p38 was cloned into pBridge Gal4 activation vec- tor (CLONTECH). A human liver cDNA library for two-hybrid screening was purchased from CLONTECH. 5 9 -Rapid Amplification of cDNA Ends— A SuperScript human fetal brain cDNA library (Life Technologies, Inc.) was used as a template. The vector sequence (5 9 -caccaaacagctatgacc-3 9 ) and a gene-specific

The magnitude and duration of activation of the MAPK superfamily are properly regulated and have a great effect on determination of the fates and responses of cells. In the case of PC12 cells, nerve growth factor treatment causes sustained activation of MAPK/ERK and p38 and drives the cells into neuronal differentiation, whereas epidermal growth factor treatment causes transient activation of MAPK/ERK and p38 and induces proliferation of the cells (7, 9 -13). MAPK/ERK, SAPK/JNK, and p38 are regulated by dual phosphorylation and dephosphorylation within the motifs TEY, TPY, and TGY, respectively, by several upstream dual specificity kinases (MAPK kinases) and several types of protein phosphatases.
Here we report the cDNA cloning and characterization of a novel member of the dual specificity phosphatase family. We named it MKP-5. We isolated MKP-5 from a human liver cDNA library by yeast two-hybrid screening using p38 as a bait. MKP-5 shares structural features with other members of the dual specificity phosphatase family, including a Cdc25-like domain and a C-terminal phosphatase motif. MKP-5 has a high basal activity in vitro and inactivates both p38 and SAPK/JNK, but not MAPK/ERK. MKP-5 specifically binds to p38 and SAPK/JNK in the yeast two-hybrid system. Intracellular localization of MKP-5 differs from that of other dual specificity phosphatases. The expression of MKP-5 mRNA is elevated by stress stimuli. These results show that MKP-5 is a novel dual specificity phosphatase specific for p38 and SAPK/JNK.

EXPERIMENTAL PROCEDURES
Two-hybrid Screening-Yeast two-hybrid screening of the Gal4 system was performed with catalytically inactive p38 (human p38␣) as bait. Catalytically inactive p38, in which Lys-53 was replaced by Met, was constructed by PCR-based mutagenesis. The primers used were 5Ј-cggggttacgtgtggcagtcgacaagctctccagacc-3Ј and 5Ј-ggtctggagagcttgtcgactgccacacgtaacccc-3Ј. Full-length p38 in pCR blunt vector (Invitrogen) was used as a template. PCR was performed using Pfu polymerase (Stratagene). A DpnI restriction enzyme (Stratagene)-treated PCR product was transformed into Escherichia coli. Positive clones were picked up, and mutagenesis was verified by sequencing. Full-length catalytically inactive p38 was cloned into pBridge Gal4 activation vector (CLONTECH). A human liver cDNA library for two-hybrid screening was purchased from CLONTECH.
5Ј-Rapid Amplification of cDNA Ends-A SuperScript human fetal brain cDNA library (Life Technologies, Inc.) was used as a template. The vector sequence (5Ј-caccaaacagctatgacc-3Ј) and a gene-specific primer (5Ј-ccacactggtgagcttcctc-3Ј) were used for the first PCR, and the SP6 primer and a gene-specific primer (5Ј-gtcacacaaccgtctccacg-3Ј) were used for nested PCR. Advantage cDNA polymerase mix (CLONTECH) was used in PCR, and PCR products were subcloned into TOPO TA cloning vector (Invitrogen). Full-length MKP-5 was obtained by the PCR method using the SuperScript human fetal brain library as a template.
Bacterial Expression and Purification of Recombinant MKP-5-For bacterial expression, the open reading frame of MKP-5 was amplified by PCR and subcloned into pGEX6P3 (Amersham Pharmacia Biotech). E. coli cells transformed with pGEX-MKP-5 were grown overnight to saturation in 10 ml of LB medium containing 50 g/ml ampicillin. The cells were grown in 3 liters of LB medium containing 50 g/ml ampicillin at 37°C to reach A 600 nm ϭ 0.6. One hour after the temperature shift to 25°C, isopropyl-␤-D-thiogalactopyranoside was added to a final concentration of 400 M, and cells were cultured for 9 h. Purification of GST-MKP-5 was performed by the method described previously (33). The yield of GST-MKP-5 was ϳ18 mg from 3 liters culture.
Phosphatase Assay-The catalytic activity of GST-MKP-5 fusion protein was measured at 37°C using p-nitrophenyl phosphate (pNPP) (Sigma) as a substrate. Reactions were performed for 15 min in 200 l of 50 mM imidazole (pH 7.5) containing 10 mM dithiothreitol, 20 mM pNPP, and the indicated amounts of GST-MKP-5 fusion protein. The reaction was stopped by the addition of 0.1 N NaOH, and the pNPP hydrolyzed was measured by absorbance at 405 nm with a microplate reader (Life Technologies, Inc.).
Cell Cultures-NIH3T3, KB, and HeLa cells were cultured in Dulbecco's modified Eagle's medium containing 10% calf serum. COS-7 cells were cultured in Dulbecco's modified Eagle's medium containing 10% fetal calf serum. These cells were maintained in 5% CO 2 at 37°C.
Construction of the Catalytically Inactive Mutant of MKP-5-The catalytically inactive mutant of MKP-5, in which Cys-408 was replaced by Ser, was constructed by PCR-based mutagenesis. The primers used were 5Ј-aggggcttctcatccacagccaggctggggtgtc-3Ј and 5Ј-gacaccccagcctgggtgtggatgagaagcccct-3Ј. Full-length MKP-5 in TOPO TA cloning vector was used as a template.
Phosphoamino Acid Analysis-Phosphorylated p38 was prepared by incubating GST-p38 (20 g) with His-MKK6 (10 g) in kinase reaction buffer for 1 h at 37°C. After precipitation with GSH beads, the sample was divided equally into four aliquots and then incubated with GST-MKP-5 (5 g each) for the indicated times. The reaction was stopped by the addition of Laemmli sample buffer. GST-p38 was extracted from SDS-polyacrylamide gel after electrophoresis. One-sixtieth of each sam-ple was spotted on a thin-layer cellulose plate. Extraction of GST-p38 from SDS-polyacrylamide gel and phosphoamino acid analysis were performed as described (35). Phosphoamino acids were detected using a BAS-2500 imager (Fuji Film).
Cell Staining-NIH3T3 and COS-7 cells were transfected with Myctagged MKP-5 or Myc-tagged CL100. After 24 h, the cells were fixed with 3.7% formaldehyde in PBS for 10 min at room temperature. After three washes with PBS, the cells were permeabilized in 0.5% Triton X-100 in PBS and washed with PBS three times. The cells were then incubated with anti-Myc antibody (9E10) in 3% bovine serum albumin in PBS for 16 h at 4°C, washed three times with PBS, and incubated with anti-mouse IgG secondary antibody in 3% bovine serum albumin in PBS for 1 h at 37°C. After three washes with PBS and two washes with Milli-Q water, coverslips were mounted with Mowiol. Fluorescence was viewed with a Zeiss fluorescence microscope.
Northern Blot Analysis-Digoxigenin-conjugated, dUTP-labeled MKP-5 riboprobe was generated by T7 RNA polymerase (Takara) transcription of the linearized MKP-5 open reading frame. Northern analysis was performed using a multiple-tissue Northern blot (CLON-TECH) according to the manufacturer's protocol.
Quantification of mRNA by Reverse Transcription-PCR-Total RNA was extracted from cells in a 100-mm dish/stimulus using an RNeasy minikit (QIAGEN Inc.) according to the manufacturer's protocol. Total RNA was reverse-transcribed into DNA by the random primer method with Rivetra Ace reverse transcriptase (Toyobo). PCR was performed using 5Ј-atgaccaaatgcagcaag-3Ј and 5Ј-ggagctggagggagttgtcac-3Ј as 5Јand 3Ј-primers, respectively. [␣-32 P]dCTP-incorporated reverse transcription-PCR products were detected by autoradiography (Eastman Kodak x-ray film).

RESULTS
Isolation of MKP-5 cDNA-We screened the human liver cDNA library by yeast two-hybrid screening with catalytically inactive p38 as bait. Seven independent positive clones were obtained. Among these, one clone, containing a 1436-base pair cDNA fragment, shares the highest similarity with dual specificity phosphatases. By performing the 5Ј-rapid amplification of cDNA ends/PCR method in the human fetal brain cDNA library, we obtained a 2061-base pair cDNA fragment. The cDNA contains an open reading frame encoding a protein of 482 amino acids with a calculated molecular mass of 52.6 kDa (Fig.  1A). An independently obtained sequence from a human testis cDNA library showed the same amino acid sequence of the open reading frame as this brain cDNA sequence (data not shown). Sequence comparison revealed that it is similar to all known dual specificity phosphatases (Fig. 1D). We then designated it MKP-5. MKP-5 contains the active-site sequence LLIHC-QAGVSRSATIVIAYLM (residues 404 -424) at the C-terminal portion (Fig. 1C), which corresponds to, but is slightly different from, the consensus sequence VXVHCXXGXSRSXTXXXAYLM (where X is any amino acid), deduced from known dual specificity phosphatases (36). The two amino acids Cys-408 and Ser-415 in this motif, together with Asp-377, are likely to participate in the catalytic mechanism of dual specificity phosphatase activity (37,38). The middle portion of MKP-5 contains two regions of amino acid similarity to the Cdc25 phosphatase (Fig. 1B). These two regions are referred to as Cdc25 homology 2 (CH2) domains and are present in all known dual specificity phosphatases (39,40). Here we tentatively call this portion containing CH2 domains the "Cdc25-like domain." Recently, it was reported that an N-terminal portion, containing this Cdc25-like domain, of MKP-3/Pyst1/rVH6 and M3/6 (hVH5) determines the binding specificity and substrate specificity (41). The amino acid sequence identity of the Cdc25-like domain between MKP-5 and known dual specificity phosphatases is 20 -34% (Fig. 1, B and D) and that of the phosphatase catalytic domain is 37-50% (Fig. 1, C and D). MKP-5 has an N-terminal stretch of 150 amino acids (Fig. 1, A and D) that is absent from the other dual specificity phosphatases. This Nterminal region has no homology with other proteins, and no distinct structural motif can be found. were used. Full-length MKP-5 interacted strongly with p38 and SAPK/JNK and very weakly with MAPK/ERK ( Fig. 2A). MKP-5 interacted slightly more strongly with p38 than with SAPK/JNK. It was reported that the Cdc25-like domain of MKP-3/Pyst1/rVH6 interacted with MAPK/ERK and that the corresponding region of M3/6 (hVH5) interacted with p38 (41). We tested the binding ability and specificity between the corresponding region of MKP-5 and MAPKs. We subcloned a 396base pair fragment of MKP-5 that corresponded to the Cdc25like domain (residues 159 -290) into the pB42AD plasmid (coding the LexA transactivation domain) and performed the two-hybrid assay. The Cdc25-like domain of MKP-5 interacted specifically with p38 (Fig. 2B), although the binding observed was very weak compared with full-length MKP-5.
Catalytic Activity of MKP-5 in Vitro-To determine whether MKP-5 has a phosphatase activity, recombinant MKP-5 (as a GST fusion protein) was expressed in E. coli, purified, and assayed for enzymatic activity against pNPP, a well known phosphatase substrate. GST-MKP-5 protein hydrolyzed pNPP in a dose-dependent manner (Fig. 3), and sodium vanadate, a potent inhibitor of tyrosine phosphatase, strongly inhibited the catalytic activity of MKP-5 against pNPP (Fig. 3). Thus, it was confirmed that MKP-5 is indeed a member of the phosphatase family. It has previously been reported that the binding of MAPK/ERK to MKP-3/Pyst1/rVH6, Pyst2, and MKP-4/Pyst3 enhances the activities of these phosphatases (31,42). We then assayed the activity of GST-MKP-5 in the presence of increasing concentrations of recombinant MAPKs. Incubation of GST-MKP-5 with MAPKs stimulated moderately the phosphatase activity of MKP-5 for pNPP in vitro. p38 and SAPK/JNK (10 g each) enhanced the MKP-5 (1 g) activity up to 1.6-fold and MAPK/ERK (10 g) ϳ1.2-fold (data not shown). Increasing the amount of MAPKs (up to 16 g) did not induce further enhancement of the activity (data not shown).
MKP-5 Inactivates p38 and SAPK/JNK-Each dual specificity phosphatase has its own substrate specificity for MAPKs (30,31,43,44). To identify the substrate specificity of MKP-5, we tested the activity of MKP-5 toward various MAPKs in cells. NIH3T3 cells were cotransfected with each of the epitopetagged MAPKs and increasing amounts of a plasmid encoding MKP-5. After stimulation of the cells by appropriate agonists to activate MAPKs (0.5 M NaCl for 20 min for p38 and SAPK/JNK and 50 ng/ml TPA for 15 min for MAPK/ERK), epitope-tagged MAPKs were immunoprecipitated and assayed for their kinase activity. MKP-5 inactivated p38 and SAPK/JNK, but not MAPK/ERK, in a dose-dependent manner (Fig. 4A, first, third, and fifth panels). It was confirmed that nearly the same amounts of MAPKs were precipitated in each lane (Fig. 4A, second, fourth, and sixth panels). Quantification of the data showed clearly that MKP-5 inactivates p38 more strongly than SAPK/JNK (Fig. 4B). It is known that when a Cys residue in the catalytic active site (VXVHCXXGXSRSXTXXXAYLM) is replaced by Ser, dual specificity phosphatases are converted to catalytically inactive forms. We tested whether MKP-5 was made catalytically inactive by such a mutation. As shown in Fig. 4C, the mutant form of MKP-5 (C408S) could not inactivate either p38 or SAPK/JNK. We then examined whether p38 and SAPK/JNK are directly inactivated by MKP-5. Phosphorylated p38 was prepared by incubating GST-p38 with His-MKK6 in the presence of [␥-32 P]ATP, and phosphorylated Myc-tagged SAPK/JNK was isolated by immunoprecipitation from Myc-tagged SAPK/JNKtransfected cells that had been exposed to 0.5 M NaCl osmotic shock. Each sample was incubated with GST-MKP-5 in vitro in the presence or absence of vanadate, and the phosphorylation states of SAPK/JNK and p38 were investigated by immunoblotting with anti-phospho-SAPK/JNK antibody and autoradiography, respectively. The results clearly show that both p38 and SAPK/JNK are direct targets of MKP-5 (Fig. 5A). Phosphorylated Myc-tagged MAPK/ERK could also be dephosphorylated by incubation with GST-MKP-5, but was a poorer substrate than p38 (data not shown).
Next, we examined whether MKP-5 was really a dual specificity phosphatase that could dephosphorylate both Thr and Tyr residues. The phosphoamino acid analysis showed that when phosphorylated p38 was incubated with MKP-5 in vitro, both phosphorylated Thr and Tyr residues were dephosphorylated in a time-dependent manner. This result clearly indicates that MKP-5 is indeed a member of the dual specificity phosphatase family.
Subcellular Localization of MKP-5-Each dual specificity phosphatase shows distinct subcellular localization (19, 21, 24, 28, 29, 31, 32). We expressed Myc-tagged MKP-5 or Myc-tagged CL100 in COS-7 and NIH3T3 cells. Indirect immunofluorescence with anti-Myc antibody showed that Myc-MKP-5 localized in both the cytoplasm and the nucleus in both NIH3T3 and COS-7 cells, whereas Myc-CL100 localized in the nucleus, as previously shown (Fig. 6). There was no marked difference in the staining intensity of Myc-MKP-5 between the cytoplasm and the nucleus. Essentially the same result was obtained using green fluorescent protein-MKP-5 (data not shown).
Distribution of MKP-5 mRNA-In Northern blot analysis, a 4.6-kilobase mRNA species for MKP-5 was detected in heart, lung, liver, skeletal muscle, and kidney, but was scarcely detected in brain, spleen, or testis (Fig. 7).
Elevation of MKP-5 mRNA Expression by Stress Stimuli-Many dual specificity phosphatases are rapidly induced after stress stimuli or serum stimulation (18, 23-26, 29, 30, 40, 43). By the reverse transcription-PCR method, we examined whether or not MKP-5 mRNA was induced by various stimuli. In KB and HeLa cells, the phorbol ester TPA, which is a good agonist for MAPK/ERK, did not induce a marked elevation of MKP-5 mRNA (Fig. 8). In contrast, anisomycin and osmotic FIG. 4. MKP-5 inhibits p38 and SAPK activation. A, an expression plasmid of Myc-MAPK, Myc-SAPK, or HA-p38 (0.7 g for a 35-mm dish) was transfected into NIH3T3 cells with the indicated -fold amounts of SR␣-MKP-5. Plasmid concentrations were maintained constant using SR␣ empty plasmids. After 24 h, the cells were stimulated by 0.5 M NaCl for 20 min (for p38 and SAPK) or by 50 ng/ml TPA for 15 min after incubation in serum-free medium for 15 h (for MAPK). Immune complex kinase assays were then performed using their respective substrates: myelin basic protein for MAPK, GST-cJun-(1-79) for SAPK, and His-ATF2 for p38. Phosphorylation of substrates was detected by autoradiography (Activity). The amounts of Myc-MAPK, Myc-SAPK, and HA-p38 in each immunoprecipitate were determined by Western blotting (second, fourth, and sixth panels). It was verified that MKP-5 protein expression was comparable between coexpressions with each MAPK in a different series of experiments using Myc-MKP-5, which gave essentially the same results as shown here. B, the intensity of substrate phosphorylation in A was quantified. C, an expression plasmid of HA-SAPK or HA-p38 (0.7 g for a 35-mm dish) was transfected into NIH3T3 cells with wild-type SR␣-Myc-MKP-5 (myc-MKP-5 wt) or mutant SR␣-Myc-MKP-5, in which Cys-408 was replaced by Ser (myc-MKP-5 cs) (1.4 g). Plasmid concentrations were maintained constant using SR␣ empty plasmids. After 24 h, the cells were stimulated by 0.5 M NaCl for 20 min. Immune complex kinase assays were then performed using their respective substrates: GST-cJun-(1-79) for SAPK and His-ATF2 for p38. Phosphorylation of substrates was detected by autoradiography (Activity). The amounts of HA-SAPK and HA-p38 in each immunoprecipitate were determined by Western blotting (anti-HA antibody for HA-p38 and anti-JNK2 antibody (Santa Cruz Biotechnology) for HA-SAPK) (second and fifth panels). The expression of wild-type Myc-MKP-5 and mutant Myc-MKP-5 was detected by Western blotting (third and sixth panels). stress (NaCl), which are potent agonists for p38, did induce a marked elevation of MKP-5 mRNA. Tumor necrosis factor-␣ significantly increased the MKP-5 mRNA level in KB cells, but not in HeLa cells. UV irradiation, which potently activates SAPK/JNK, did not induce an elevation of the mRNA (Fig. 8). DISCUSSION We have isolated a cDNA clone encoding a novel dual specificity phosphatase, MKP-5, by yeast two-hybrid screening using p38 as bait. All of the amino acids previously shown to be important for the catalytic activity of dual specificity phosphatases are conserved in MKP-5. MKP-5 also has two CH2 domains (the Cdc25-like domain) in its middle portion that are also conserved in dual specificity phosphatases (39,40). The previously described dual specificity phosphatases, except hVH5 (M3/6), consist of only two regions, the Cdc25-like domain and the phosphatase catalytic domain (Fig. 1D). hVH5 (M3/6) has a long C-terminal stretch whose function is unknown (26,27). MKP-5 is unique in that it has an additional long N-terminal region, consisting of ϳ150 amino acids, that is absent from other dual specificity phosphatases (Fig. 1D). This region has no homology with other proteins, and so the role of this N-terminal region is unknown at present. The amino acid sequence identity of the Cdc25-like domain between MKP-5 and other dual specificity phosphatases is 20 -34%. hVH5 shows 34% identity. The catalytic domain of MKP-5 shows 37-50% identity to that of other dual specificity phosphatases. From the standpoint of the amino acid sequence, MKP-3, Pyst2, and MKP-4 form one group, and CL100, PAC1, and MKP-2 form another. The identities among MKP-3, Pyst2, and MKP-4 are 42-59% in the Cdc25-like domain and 74 -86% in the catalytic domain, and the identities among CL100, PAC1, and MKP-2 are 34 -54% in Cdc25-like domain and 73-81% in the catalytic domain. hVH5 (M3/6) and hVH3/B23 share much lower similarities with either of the two groups and with each other. Because MKP-5 does not share high similarity with either of the two groups, hVH5 (M3/6) or hVH3/B23, MKP-5 may be a distinct member of the dual specificity phosphatase family.
Bacterially expressed GST-MKP-5 possessed a high phosphatase activity for pNPP, which seems relatively strong. For example, under almost the same reaction conditions, incubation with 40 g of GST-MKP-4/Pyst3 for 60 min (32) is equivalent to incubation with 3 g of GST-MKP-5 for 15 min.
It has previously been reported that the Cdc25-like domains of MKP-3/Pyst1/rVH6 and hVH5 (M3/6) determine their binding specificity and substrate specificity for MAPKs (39). Using the semiquantitative yeast two-hybrid interaction assay, we tested whether MKP-5 could bind to MAPKs and whether the Cdc25-like domain was responsible for the binding. Full-length MKP-5 could bind to p38 and SAPK/JNK, but only very weakly to MAPK/ERK. The affinity for p38 appeared stronger than that for SAPK/JNK. The Cdc25-like domain of MKP-5 interacted specifically with p38, although the affinity seemed very low. Thereby, the proper binding of MKP-5 to p38 may require not only the Cdc25-like domain, but also other regions. That the binding between MKP-5 and SAPK/JNK is slightly weaker than that between MKP-5 and p38 may be accounted for by the fact that the Cdc25-like domain of MKP-5 has no affinity for SAPK/JNK.
The binding of MAPK/ERK to MKP-3/Pyst1/rVH6, Pyst2, and MKP-4/Pyst3 has been reported to enhance the activities of these phosphatases up to ϳ20-fold (31,42). In contrast, incubation with p38 or SAPK/JNK enhanced the catalytic activity of MKP-5 only 1.6-fold. There may be two kinds of dual specificity phosphatases. Members of one subfamily, which comprises MKP-3/Pyst1/rVH6, Pyst2, and MKP-4/Pyst3, may be activated by binding to their physiological substrates, and members of the other subfamily may exist as constitutively active forms. In fact, MKP-5 seems to have a higher basal activity than MKP-3/Pyst1/rVH6, Pyst2, and MKP-4/Pyst3 (see above). It is possible that the N-terminal portion of MKP-5 may serve as regulatory sites, i.e. binding sites for inhibitors and/or regulators.
Previous studies showed that CL100/MKP-1, PAC1, MKP-2/ hVH2/TYP-1, and hVH3/B23 localize in the nucleus, whereas MKP-3/Pyst1/rVH6, Pyst2, and MKP-4/Pyst3 localize in the cytoplasm (19, 21, 24, 28, 29, 31, 32). MKP-5 is shown here to localize evenly in the cytoplasm and the nucleus. This subcel- The level of phosphorylated SAPK was examined by immunoblotting with anti-phospho-SAPK/JNK antibody. Phosphorylated p38 was prepared by incubating GST-p38 with MKK6 in the presence of [␥-32 P]ATP. After precipitation with GSH beads, 0.1 g of phosphorylated GST-p38 was incubated at 37°C with 0.1 g of GST-MKP-5 for the indicated times with or without 5 mM vanadate in the same buffer solution (10 l). The level of phosphate incorporation of p38 was determined by autoradiography. B, MKP-5 dephosphorylates both Thr and Tyr residues of p38 in vitro. Phosphorylated p38 was incubated with MKP-5 for the indicated times, and phosphoamino acid analysis was performed. p-Ser, p-Thr, and p-Tyr indicate phosphorylated Ser, Thr, and Tyr, respectively. lular localization of MKP-5 is different from that of previously described dual specificity phosphatases. The subcellular localization of MKP-5 is unchanged after stress stimulation (data not shown). p38 and SAPK/JNK, targets of MKP-5, also localize in both the cytoplasm and the nucleus, and their localization does not dramatically change after stress stimulation. In con-trast, MAPK/ERK, which is present in the cytoplasm before stimulation, is translocated to the nucleus upon stimulation. The correspondence of the subcellular localization between MKP-5 and its targets, p38 and SAPK/JNK, might have some physiological relevance or may reflect binding between them.
The expression of some dual specificity phosphatases is known to be induced after growth factor stimulation and/or stress stimulation (18, 23-26, 29, 30, 40, 43). The MKP-5 mRNA was elevated significantly by anisomycin treatment, osmotic stress, and tumor necrosis factor-␣ treatment, which are known to activate p38 strongly. TPA, which is a good agonist for MAPK/ERK, did not induce MKP-5 mRNA; so there may be a correlation between MKP-5 mRNA inducibility and p38 activation in extracellular stimuli. However, UV irradiation, which is able to activate p38, did not increase the mRNA level. The mechanism of elevated expression of MKP-5 mRNA by stress stimuli should be elucidated in future studies.
In summary, we have cloned and characterized a novel dual specificity phosphatase, designated MKP-5. MKP-5 does not seem to belong to known subfamilies of dual specificity phosphatases in terms of structural features. MKP-5 is specific for p38 and SAPK/JNK, and its subcellular localization is unique in that it is present in both the cytoplasm and the nucleus. Thus, MKP-5 may be a novel member of the dual specificity phosphatase family acting on the MAPK family molecules. KB and HeLa cells were grown to subconfluence (ϳ70% confluent) and stimulated by TPA (50 ng/ml for 20 min), UVC (60 J/m 2 ), anisomycin (5 g/ml for 20 min), NaCl (0.5 M for 20 min), or tumor necrosis factor-␣ (TNF-␣; 50 ng/ml). After 2 h, the cells were lysed, and reverse transcription-PCR was performed to evaluate the relative level of MKP-5 mRNA expression as described under "Experimental Procedures." Glyceraldehyde-3-phosphate dehydrogenase (G3PDH) was also analyzed as a control.