MKP-7, a Novel Mitogen-activated Protein Kinase Phosphatase, Functions as a Shuttle Protein*

Mitogen-activated protein kinase (MAPK) phosphatases (MKPs) negatively regulate MAPK activity. In the present study, we have identified a novel MKP, designated MKP-7, and mapped it to human chromosome 12p12. MKP-7 possesses a long C-terminal stretch containing both a nuclear export signal and a nuclear localization signal, in addition to the rhodanese-like domain and the dual specificity phosphatase catalytic domain, both of which are conserved among MKP family members. When expressed in mammalian cells MKP-7 protein was localized exclusively in the cytoplasm, but this localization became exclusively nuclear following leptomycin B treatment or introduction of a mutation in the nuclear export signal. These findings indicate that MKP-7 is the first identified leptomycin B-sensitive shuttle MKP. Forced expression of MKP-7 suppressed activation of MAPKs in COS-7 cells in the order of selectivity, JNK ≫ p38 > ERK. Furthermore, a mutant form MKP-7 functioned as a dominant negative particularly against the dephosphorylation of JNK, suggesting that MKP-7 works as a JNK-specific phosphatase in vivo. Co-immunoprecipitation experiments and histological analysis suggested that MKP-7 determines the localization of MAPKs in the cytoplasm.

Mitogen-activated protein kinase (MAPK) phosphatases (MKPs) negatively regulate MAPK activity. In the present study, we have identified a novel MKP, designated MKP-7, and mapped it to human chromosome 12p12. MKP-7 possesses a long C-terminal stretch containing both a nuclear export signal and a nuclear localization signal, in addition to the rhodanese-like domain and the dual specificity phosphatase catalytic domain, both of which are conserved among MKP family members. When expressed in mammalian cells MKP-7 protein was localized exclusively in the cytoplasm, but this localization became exclusively nuclear following leptomycin B treatment or introduction of a mutation in the nuclear export signal. These findings indicate that MKP-7 is the first identified leptomycin B-sensitive shuttle MKP. Forced expression of MKP-7 suppressed activation of MAPKs in COS-7 cells in the order of selectivity, JNK > > p38 > ERK. Furthermore, a mutant form MKP-7 functioned as a dominant negative particularly against the dephosphorylation of JNK, suggesting that MKP-7 works as a JNK-specific phosphatase in vivo.

Co-immunoprecipitation experiments and histological analysis suggested that MKP-7 determines the localization of MAPKs in the cytoplasm.
The activation of the mitogen-activated protein kinase (MAPK) 1 cascade plays a key role in transducing various extracellular signals to the nucleus to induce responses such as gene expression, cell proliferation, differentiation, cell cycle arrest, and apoptosis (1,2). MAPKs consist of three major subfamilies, extracellular signal kinases (MAPK/ERK), stressactivated kinase/c-Jun N-terminal kinases, and homologues of the budding yeast HOG1 protein (p38). For full activation of these MAPKs, phosphorylation of both threonine and tyrosine residues found in TXY motifs is required by dual specificity kinases, known as MAPK kinases. Thus dephosphorylation of the TXY motif is critical for negative regulation of MAPK activity (3).
MKPs are potential negative regulators of MAPK cascades and as such are assumed to be involved in carcinogenesis by regulating cell proliferation and apoptosis. As a result, all human MKP genes have been mapped. Among them, MKP-2 and MKP-3 are mapped to a gene locus encoding tumor suppressors for prostate and pancreatic cancer, respectively (20,21). MKP-X and MKP-5 are mapped to 3p21 (21) and to 1q41 (13), respectively, where frequent deletions are reported in a number of different tumors.
Activation/phosphorylation of MAPKs leads to their nuclear translocation and phosphorylation of certain DNA-binding proteins that contribute to transcriptional regulation. The mechanism of nuclear-cytoplasmic transport of MAPKs is, however, not clear. Recent reports indicate that several different proteins contain an intrinsic nuclear export signal (NES) motif mediating their subcellular localization and nuclear-cytoplasmic shuttling through association with the export receptor, CRM1/exportin 1 (22)(23)(24). Among them, MEK1, one of MAPK kinases (MAPKKs) has been well characterized (25,26). Its NES motif was shown to function as an anchor protein of ERK in the cytoplasm when cells are unstimulated, thereby suppressing cell transformation. Leptomycin B (LMB), a specific inhibitor of nuclear export that blocks binding between the NES and CRM1 (27)(28)(29), caused nuclear accumulation of MEK1. Substitutions of crucial leucines in the NES motif with alanines caused nuclear accumulation of MEK1 and ERK (30).
By screening an EST library, we identified a human cDNA clone encoding a novel member of the MKP family, MKP-7.
Interestingly MKP-7 contains predicted functional motifs such as a nuclear export signal (NES) and nuclear localization signals (NLSs), suggesting that it functions as a shuttle protein and a MAPK phosphatase. In this report, the substrate specificity, subcellular localization, and regulation of MKP-7 are presented and discussed.

EXPERIMENTAL PROCEDURES
Identification of a Novel MKP cDNA-By using the amino acid sequence of human MKP-4, we screened an expressed sequence tag data base, dbEST, and identified a novel MKP. A human clone (GenBank TM accession number AI274662, IMAGE clone ID 1986459) and a mouse clone (GenBank TM accession number AA879894, IMAGE clone ID 1230637) had high sequence homology to human MKP-4. The human and mouse clones were obtained from Research Genetics, Inc. (Huntsville, AL), and their nucleotide sequences were determined using the dideoxynucleotide chain termination method on a 373A DNA sequencer (Applied Biosystems, Foster City, CA), with a Dynamic dye terminator cycle sequencing kit (Amersham Pharmacia Biotech). We performed 5Јand 3Ј-RACE using cDNA derived from Jurkat cells as template and primers based on the human clone. Nucleotide sequencing analysis showed that the deduced amino acid sequences of the PCR product are identical to those of clone AI 274662 with the presence of some polymorphism in 5Ј-and 3Ј-UTR regions (data not shown). The obtained PCR fragment had an ORF of 1995 base pairs. The mouse clone lacked the first ATG; therefore, 5Ј-and 3Ј-RACE using cerebellar mRNA was performed, resulting in a PCR fragment containing an ORF of 1980 base pairs. All RACE methods were performed with a SMART TM RACE cDNA amplification kit (CLONTECH, Palo Alto, CA) according to the manufacturer's protocol.
Northern Blot Analysis-Total RNAs from various tissues of 6-weekold male mice were isolated by acid guanidinium thiocyanate extraction (31). The RNAs were fractionated on a 1.5% formaldehyde-agarose gel and transferred to nitrocellulose membranes (Schleicher & Schuell). The membranes were hybridized with a 32 P-labeled 1.8-kb insert of mouse MKP-7 cDNA, which contains the full-length ORF. Hybridization was performed at 42°C in 50% formamide, 0.65 M NaCl, 5 mM EDTA, 1ϫ Denhardt's solution, 10% dextran sulfate, 0.1 M PIPES, pH 6.8, 0.1% SDS, and 100 g/ml denatured salmon sperm DNA. The membranes were washed twice with 2ϫ SSC containing 0.1% SDS at room temperature for 5 min, followed by sequential washes with 0.5ϫ SSC containing 0.1% SDS and 0.2ϫ SSC containing 0.1% SDS at 50°C, each for 15 min. The filters were exposed to an x-ray film using an intensifying screen at Ϫ80°C. MKP-7 Expression Plasmids-To construct pEGFP-MKP-7, the coding region of human MKP-7 cDNA was amplified by PCR to introduce a BglII site on the 5Ј end and a SalI site on the 3Ј end using Platinum Pfx DNA polymerase (Life Technologies, Inc.) and ligated to BglII and SalI-digested pEGFP-C2 vector (CLONTECH) in frame with the EGFP-coding sequence. To construct pFLAG-MKP-7, the same region was amplified by PCR to introduce a NotI site on the 5Ј end and a SalI site on the 3Ј end using Platinum Pfx DNA polymerase and ligated to NotI-and SalI-digested pFLAG-CMV2 vector (Sigma) in frame with the FLAG epitope sequence. Several constructs encoding MKP-7 mutant proteins, including DA (D213A), CS (C244S), delC1 (residues 1-290), delC2 (residues 1-370), delC3 (residues 1-604), LA (L380A, L383A, and L385A), delR (residues 162-665), and delR-LA (residues 162-665 of LA), were constructed by PCR and subcloned into pFLAG-CMV2 (Sigma). The final PCR products were cloned into pGEM-T Easy (CLONTECH) and sequenced. No substitution was found except for the targeted mutation.
Cell Culture and Transient Transfection-HeLa and COS-7 cells were maintained in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum at 37°C under 5% CO 2 . Cells were co-transfected with pFLAG-MKP-7 (wild type or mutant constructs) together with SR␣-HA-ERK2, SR␣-HA-JNK1, or pMT3-HA-p38␣. For transient assays, cells were transfected using Fugene-6 (Roche Molecular Biochemicals) according to the manufacturer's recommendation. Twenty four hours after transfection, cells were maintained with or without serum for 18 h and then stimulated with either 5 ng/ml PMA for 10 min for ERK2 activation or with 0.4 M sorbitol for 30 min for JNK1 and p38␣ activation.
Leptomycin B Treatment-Twenty four hours after transfection, HeLa cells were maintained without serum for 18 h and then treated with 5 nM LMB (provided by Dr. M. Yoshida) for the indicated periods.
Co-immunoprecipitation-Transfected COS-7 cells were lysed on a plate (300 l/60-mm plate) in co-IP buffer (50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 2 mM EDTA, 10% glycerol, 0.5% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, and 10 g/ml aprotinin). The cell lysate was clarified by centrifugation, and protein concentrations were measured as above. The supernatant (500 g) was incubated with 2 g of mouse anti-FLAG M2 antibody and 15 l of protein G-Sepharose 4 fast flow (Amersham Pharmacia Biotech), which had been equilibrated with the co-IP buffer in a 500-l tube. After 1 h rotation at 4°C, the beads were washed 5 times with 500 l of the co-IP buffer. The immunoprecipitates were resuspended in 40 l of 1 ϫ Laemmli's SDS sample buffer, boiled for 2 min, separated by SDSpolyacrylamide gel electrophoresis on 10% gels, and transferred to a nitrocellulose membrane (Amersham Pharmacia Biotech). FLAG-or HA-tagged proteins were detected by the respective antibodies using ECL reagents.
Cell Staining-HeLa cells on coverslips coated with Vitrogen 100 (Collagen Biochemical, Palo Alto, CA) were transfected with wild type or mutant forms of pFLAG-MKP-7. Transfected cells were fixed in PBS containing 3.7% formaldehyde for 10 min and then permeabilized with PBS containing 0.5% Triton X-100 for 5 min. After incubation in PBS containing 3% BSA (PBS-B) for 2 h, cells were incubated with anti-FLAG M2 antibody or anti-FLAG polyclonal antibody (provided from Dr. K. Yamashita) to detect FLAG-tagged proteins and an anti-HA (12C5) antibody to detect HA-tagged proteins in PBS-B overnight at 4°C. After three washes with PBS, cells were incubated for 20 min at 37°C with Cy3-conjugated goat anti-mouse IgG ϩ IgM (H ϩ L) antibody (Chemicon International), fluorescent isothiocyanate-conjugated goat anti-mouse IgG (H ϩ L) (Kirkegaard & Perry Laboratories, Inc.), or AlexaFluor 546-conjugated goat anti-rabbit IgG (H ϩ L) (highly crossabsorbed) (Molecular Probes, Eugene, OR) in PBS-B. After three washes with PBS, coverslips were mounted with PBS containing 90% glycerol. Fluorescence signals were visualized by a fluorescence microscope.

Isolation of a Human and Mouse MKP-7-To search for novel
MKPs, we screened a human dbEST using the amino acid sequence of human MKP-4 as a probe. This clone (GenBank TM accession number AI274662, IMAGE clone ID 1986459) also showed high sequence homology to human MKP-4. The nucleotide sequence of this clone was determined, and then 5Ј-RACE was performed to identify the first methionine codon and sequences of the 5Ј-UTR region. The full-length clone was obtained by reverse transcriptase-PCR using mRNA from Jurkat cells. The nucleotide sequences were verified by comparing three independent clones (Fig. 1). The open reading frame (ORF) of this cDNA was predicted to encode 665 amino acids. A dual specificity phosphatase (DSP) catalytic site motif, VXVH-CXAGXSRSXTXXXAYXM, which is essential for phosphatase activity, was found in this clone (32). A MAPK-docking motif composed of a kinase-interacting motif (33) at residues 51-65 and ␦-like domain (34 -36) at residues 161-169 were also present. Since this ORF contains these two essential sequences, we designated the clone a new member of MKP gene family (Fig.  1). To date, 6 species among 12 DSPs have been designated MKP-1 to MKP-6 based on their structural similarity and substrate specificity toward MAPKs. Following this nomenclature, we named our novel DSP MKP-7 (Fig. 1).
MKP-7 exhibits several other predicted functional motifs. Two bipartite NLS motifs (37) were located at amino acid residues 296 -313 (NLS1) and 610 -627 (NLS2). One leucinerich NES motif was located at amino acid residues 376 -385. The presence of a NES as well as an NLS suggested that MKP-7 acts as a shuttle protein. PEST sequences, which are thought to be involved in rapid degradation through ubiquitinmediated proteolysis (38), were found at residues 332-353 and 441-462 residues. The presence of these motifs suggested that a C-terminal stretch of MKP-7 is important for localization and stability of the protein.
To obtain the mouse homologue, we screened mouse dbESTs and found one clone (GenBank TM accession number AA879894, IMAGE clone ID 1230637) that had high sequence homology to human MKP-7 but lacked the 5Ј-half of the ORF. We obtained a full-length cDNA clone by RACE using mouse cerebellum RNA as a template. The nucleotide sequence was verified using three independent reverse transcriptase-PCR fragments covering the entire ORF of mouse MKP-7. In Fig. 2A the deduced amino acid of mouse MKP-7 (lower line) is aligned with human MKP-7 (upper line). The identity between mouse and human MKP-7 is 90.4% at the amino acid level. MKP-7 appears to be composed of three domains. Domain 1 is a rhodanese-like domain that contains two CH2 domains and a kinase-interacting motif. Domain 2 is a DSP catalytic domain that contains ␦-like domain and catalytic motif, and domain 3 is a long C-terminal stretch. Domains 1 and 2 are highly conserved between human and mouse MKP-7s.
The alignment of domain structures of MKP-7 with those of other MKPs is shown in Fig. 2B. MKP-7 has similar domain structures and the highest sequence similarity to hVH5. The similarities between human MKP-7 and hVH5 are 76.1, 59.6, and 30.5% in the DSP domain, rhodanese-like domain, and C-terminal stretch, respectively.
Chromosomal Location of the MKP-7 Gene-The location of the MKP-7 gene on human chromosomes was determined by the identification of MKP-7 cDNA between two sequencetagged site markers, G24001 and G41293, in Homo sapiens 12p BAC RP11-253I19 (GenBank TM accession number AC007619.22). This gene was localized to human chromosome 12p12 as shown in Fig. 3A. By comparing the nucleotide sequences of the MKP-7 genomic clone with the cDNA, the MKP-7 gene was shown to be composed of at least seven exons (Fig. 3B). The catalytic core and C-terminal stretch were each encoded on a single exon, whereas the rhodaneselike domain was contained on three exons. Since the 5Ј-UTR sequence did not match any sequences in the NCBI data base or others, we did not identify the promoter region of MKP-7.
Theoretically, DSPs have tumor suppressor activity. The Recurrent Chromosome Aberrations in Cancer Data base searcher (www.cgap.nci.nih.gov/Chromosomes/Recurrent Aberrations) showed that chromosome 12p12 is a region where deletions often occur in acute lymphoblastic leukemia, acute and chronic myeloid leukemia, and myelodysplastic syndrome, suggesting that tumor suppressor genes for leukemia lie within this region.
Tissue Distribution of MKP-7 mRNA-The expression pattern of MKP-7 mRNA in mouse tissues was examined by Northern blot analysis using mouse MKP-7 cDNA containing the entire ORF as a probe. As shown in Fig. 4, two mRNA species of 4.1 kb as a major transcript and 2.1 kb as a minor transcript were detected. The 4.1-kb transcript was abundantly expressed in the brain, kidney, intestine, and testis but expressed at low levels in the thymus, spleen, and bone marrow. The 2.1-kb transcript was detected only in the testis.
In order to analyze specificity further, we used two types of catalytically inactive proteins with mutations in conserved residues, MKP-7-CS (C244S) and MKP-7-DA (D213A). Both mutant proteins significantly enhanced phosphorylation of HA-JNK1 but had little effect on activation of HA-ERK2 and HA-p38␣ (Fig. 5B, lanes 10 and 11), indicating that both mutant proteins function as dominant negatives toward JNK. Under the same conditions, MKP-5 inactivated JNK1 and p38␣ more strongly than ERK2, as reported (Fig. 5, A-C, lane 12, and Refs. 11 and 12). Therefore, MKP-7 blocked activation of MAP kinases in the order of selectivity, JNK1 Ͼ Ͼ p38␣ Ͼ ERK2.
In Vivo Interaction between MKP-7 and MAPKs-Since MKP-7 inactivated JNK1 and p38␣ in vivo, we asked whether MKP-7 binds MAPKs in vivo. We tested an in vivo interaction between MKP-7 and MAPKs by co-immunoprecipitation experiments to determine whether MKP-7 has a binding preference among MAPKs and, if a direct interaction occurs, whether MAPKs must be phosphorylated for that interaction.
FLAG-MKP-7 and HA-MAPKs (either HA-ERK2, HA-JNK1, or HA-p38␣) were co-expressed in COS-7 cells (Fig. 6A). As expected, HA-JNK1 was co-immunoprecipitated with FLAG-MKP7 (Fig. 6A, lanes 10 -13); however, stimulation did not affect this interaction (Fig. 6A, compare lanes 12 and 13). The interaction was observed even under culture conditions lacking starvation or stimuli (Fig. 6A, lane 10). When we expressed FLAG-MKP-7DA, an inactive mutant, interaction of FLAG-MKP-7DA and HA-JNK1 was similar to that of FLAG-MKP-7 and HA-JNK1 (data not shown), suggesting that MKP-7 binds the dephosphorylated as well as the phosphorylated protein. It should be noted that MKP-7 binds not only JNK1 but also ERK2 and p38␣. Under standard conditions, we did not observe any binding preference of MKP-7 toward a specific MAPK. Under these conditions, the MAPK binding specificity of MKP-5 (Fig. 6B) and MKP-2 (data not shown) was confirmed as already reported (11,39), which excludes a possibility that the binding of MKP-7 and MAPKs is due to be artificial by overexpression. Taken together, we conclude that MKP-7 interacts with ERK2 and p38␣ as well as JNK1 with similar preference in vivo and that such interaction does not depend on the phosphorylation state of MAPKs.
An  6 -9, and B, lanes 1 and 2), SR␣-HA-JNK1 (A, lanes 10 -13, and B, lanes 3 and 4), or pMT3-HA-p38␣ (A, lanes 14 -17, and B, lanes 5 and 6) in 60-mm dishes. Twenty four hours after transfection, the cells were maintained with or without serum for 18 h and then stimulated with either 5 ng/ml PMA for 10 min (ERK2 activation) or 0.4 M sorbitol for 30 min (JNK1 and p38␣ activation). An immunoprecipitation (IP)-Western was done by using anti-FLAG M2 antibody for immunoprecipitation and blotted with anti-HA antibody. The expression levels of FLAG-MKPs (FLAG-MKP-7 and FLAG-MKP-5) and HA-MAPKs (HA-ERK2, HA-JNK1 and HA-p38␣) were assessed by immunoblot using anti-FLAG or anti-HA antibody. Data are representative of three independent experiments. ined (Fig. 6), we further tested the effect of MKP-7-inactive mutants CS and DA on the phosphorylation state of MAPKs under unstimulated conditions (Fig. 7). Expression levels of either FLAG-MKP-7 proteins or HA-MAPKs were similar in each lane (data not shown). A dominant negative effect against dephosphorylation of JNK was observed even in unstimulated cells (Fig. 7B, lanes 5 and 7). A similar effect was observed for ERK (Fig. 7A, lanes 5 and 7) but not for p38␣ (Fig. 7C, lanes 5  and 7).
Next we examined those effects under conditions lacking starvation or stimulation (Fig. 7, D-F). Accumulation of phosphorylated forms of ERK2 and JNK1 was evident (Fig. 7, D and  E, lanes 5 and 7) but that of p38␣ was not (Fig. 7F, lanes 5 and  7). These results suggest that MKP-7 may block ERK and JNK phosphorylation/activation when cells are unstimulated.
MKP-7 Is Localized in the Cytoplasm-In order to understand the function of MKP-7, we investigated the subcellular localization of MKP-7. The localization of EGFP-MKP-7 in HeLa cells is shown in Fig. 8A. The control EGFP protein was distributed evenly in transfected cells, whereas EGFP-MKP-7 was specifically localized in the cytoplasm (Fig. 8A, a and c). In a separate experiment, we used FLAG-tagged MKP-7 to ensure that localization of MKP-7 to the cytoplasm was not an artifactual result of the green fluorescent protein domain being fused to the phosphatase (Fig. 8A, e). We examined the subcellular distribution of FLAG-MKP-7 in several cell lines, including COS-7, NIH3T3, and 293 cells. FLAG-MKP-7 was localized exclusively in the cytoplasm of all these cell lines (data not shown).

MKP-7 Is an LMB-sensitive Shuttle
Protein-To determine whether the predicted NES is functional, we examined the effect of LMB on distribution of FLAG-MKP-7 when the cells were starved (Fig. 8B). Without LMB treatment, FLAG-MKP-7 was localized exclusively in the cytoplasm, but it accumulated in the nucleus in a manner proportional to incubation time with LMB. By 120 min of LMB treatment, FLAG-MKP-7 had exclusively accumulated in the nucleus. Similar results were obtained in cells without starvation (data not shown). These data suggested that MKP-7 shuttles between the nucleus and the cytoplasm and that nuclear export of MKP-7 is LMB-sensitive.
Analysis of the Sequences Required for Shuttling-To determine the importance of C-terminal stretch for nuclear transport and export, we analyzed localization of the following three deletion mutant proteins: FLAG-MKP-7-delC1, FLAG-MKP-7-delC2, and FLAG-MKP-7-delC3 (Fig. 9A). In sharp contrast to the wild type protein, FLAG-MKP-7-delC1 lost its specific localization and was evenly distributed in the cell. It also lacked sensitivity to LMB (Fig. 9), strongly suggesting that the cytoplasmic localization and LMB sensitivity of MKP-7 is determined by the C-terminal stretch. FLAG-MKP-7-delC2 localized mainly in the nucleus and its localization was not affected by LMB, whereas FLAG-MKP-7-delC3 was localized in the cytoplasm and this localization was sensitive to LMB. These data show that the region (residues 291-370) containing NLS1 functions for nuclear import, and the region containing the NES functions for LMB-sensitive nuclear export. NLS2 appears not to be critical since the localization of FLAG-MKP-7delC3 is the same as that of the wild type protein.
To verify the importance of the NES motif, FLAG-MKP-7-LA, which has crucial three leucines substituted with alanines, was expressed in HeLa cells. This mutant protein was com-pletely accumulated in the nucleus. These results were observed in other cell lines such as COS-7, NIH3T3, and 293 cells (data not shown).
We also investigated the involvement of the rhodanese-like domain in subcellular localization. FLAG-MKP-7-delR and FLAG-MKP-7-delRLA did not translocate to the nucleus even with LMB treatment. These results also support the idea that MKP-7 is a shuttle protein between the nucleus and the cytoplasm. NLS1, in collaboration with rhodanese-like domain, seems function to allow nuclear import, and the NES in the C-terminal stretch is critical for the nuclear export.
MKP-7 Determines MAPK Localization-The observation that MKP-7 was localized exclusively in the cytoplasm (Fig. 8) and that it co-immunoprecipitated with MAPK ( Fig. 6) led us to analyze effect of MKP-7 on localization of MAPKs. Without stimulation, HA-ERK2 is localized in the cytoplasm; however, HA-JNK1 and HA-p38␣ are distributed evenly in the nucleus as well as the cytoplasm as reported (Fig. 10A) (40 -45). However, following co-transfection with FLAG-MKP7, HA-JNK1 or HA-p38␣ became accumulated in the cytoplasm. Localization of co-expressed FLAG-MKP-7 was similar to that of HA-ERK2, HA-JNK1, and HA-p38␣.  (g and h). B, after transfection of pFLAG-MKP7, HeLa cells were maintained without serum for 18 h and then exposed to 5 nM LMB for the indicated periods. FLAG-MKP-7 was detected by immunofluorescence using an anti-FLAG M2 antibody with Cy-3 conjugated goat anti-mouse secondary antibody. appears to be regulated by both an NLS and an NES located in the C-terminal stretch of the protein.
Since MKP is a potential negative regulator of the MAPK cascade, it could play a role in carcinogenesis by regulating cell proliferation and apoptosis. To further our understanding of the relationship of MKPs to diseases, six human MKP genes have been already mapped. Among them, hVH2/MKP-2 and MKP-4 map to suppressor gene loci for prostate cancer and pancreatic cancer, respectively. PYST2/MKP-X maps to 3p21, where frequent deletions are found in several different tumors. We have mapped MKP-7 to 12p12, where deletions often occur in acute lymphoblastic leukemia, acute and chronic myeloid leukemia, and myelodysplastic syndrome. Since MKP-7 was identified as a phosphatase specific for JNK, it could also function as a tumor suppressor in cancers through negative regulation of the JNK pathway. Whether the MKP-7 gene is deleted or mutated in such tumors is currently under investigation, although levels of MKP-7 gene expression were very low in hematopoietic and lymphoid cells.
In Northern blots a mouse MKP-7 probe detected 4.1-and 1.8-kb mRNAs as ubiquitous and testis-specific transcripts, respectively. An interesting feature of MKP-7 expression is that the level of the 4.1-kb mRNA is very low in some tissues, such as thymus, spleen, and bone marrow. It is possible that expression is down-regulated in hematopoietic or proliferating cells. Recently we reported that testis-and skeletal musclespecific DSP TMDP is abundantly expressed in the testis (46), and MKP-5 is expressed as a shorter transcript in the testis (13). It is interesting that both testis-specific transcripts were expressed specifically during meiosis in the testis. The structure and properties of the testis-specific MKP-7 transcript are being investigated and compared with those of TMDP and the testis-specific MKP-5 transcripts. MKP-7 is likely to function as a JNK phosphatase. MKP-7 was more effective toward phosphorylated and activated JNK1 than ERK2 and p38␣ (JNK1 Ͼ Ͼ p38␣ Ͼ ERK2). Also the finding that inactive mutants of MKP-7 worked as strong dominant negatives against dephosphorylation of JNK supports the idea that MKP-7 functions as a JNK phosphatase in vivo. The substrate specificity of MKP-7 toward MAPKs was similar to that of hVH-5/M3/6 (JNK Ϸ p38 Ͼ Ͼ ERK) (47) and MKP-5 (JNK Ϸ p38 Ͼ Ͼ ERK) (11,12) but very different from that of PAC-1 (ERK ϭ p38 Ͼ JNK) (48 -50), MKP-2 (ERK ϭ JNK Ͼ p38) (50), MKP-3 (ERK Ͼ Ͼ JNK ϭ p38) (9,47), and MKP-4 (ERK Ͼ p38 ϭ JNK) (10). MKP-7 and hVH-5/M3/6 have high sequence homology, similar domain structures, and similar substrate specificities toward MAPKs. Compared with the wild type protein, FLAG-MKP-7delC2 showed higher activity toward p38␣, although its activity toward JNK1 and ERK2 was unchanged (data not shown). These results suggest that 1) the high specificity to JNK1 depends on conserved sequences between MKP-7 and hVH-5/M3/6, which include the rhodaneselike domain and catalytic domain, and 2) the C-terminal stretch of MKP-7 may interfere with its recognition of p38␣ as a substrate.
It is unclear why, despite its high specificity toward JNK1, MKP-7 can bind ERK2 and p38␣ as well as JNK1 with similar affinity. This observation suggests that binding is necessary but not sufficient for determination of substrate specificity. To address this issue, experiments either substituting the catalytic domain of MKP-7 with the corresponding domain from other MKPs or mutating the catalytic domain of MKP-7 will be required. It is of note that MKP-7-inactive mutants increased levels of phosphorylated HA-ERK2 as well as HA-JNK1, when cells are unstimulated (Fig. 7). MKP-7 may play a role as a gatekeeper for ERK as well as JNK by setting a high threshold for stimulation.
To our knowledge, MKP-7 is the first identified shuttle MKP. By substitution experiments and LMB treatment, we showed that the NES in the C-terminal stretch is functional. It is of interest that the NES motif of MKP-7 (LXXXLXXLXL) is identical to that of MEK1 (25,30). For nuclear import, the NLS1 region in addition to the rhodanese domain was identified as functional. The rhodanese-like domain may be involved in conformational changes of MKP-7, but the details remain to be clarified. An important question is the function and role of MKP-7 as a shuttle molecule. Based on the results shown in Fig. 10 that MKP-7 trapped MAPKs in the cytoplasm, we propose two models. One model is that MKP-7 translocates into the nucleus and interacts with activated MAPKs and then dephosphorylates and transports them back to the cytoplasm. Another is that MKP-7 remains in the cytoplasm to anchor and dephosphorylate MAPKs. To distinguish between these two models, we analyzed the activity of FLAG-MKP-7 LA. This mutant was localized in the nucleus (Fig. 9) and showed activity toward MAPKs similar to that of the wild type protein (data not shown), supporting the former model. Future experiments focusing on MKP-7 will address the question how localizations and activity of MAPKs are regulated.