The Phosphatase MKP1 Is a Transcriptional Target of p53 Involved in Cell Cycle Regulation*

The tumor suppressor p53 protein suppresses cell growth by inducing cell cycle arrest or apoptosis. Despite the fact that p53-dependent p21-mediated G1 arrest induced by DNA damage is well defined, the role of p53 in the cell cycle in response to the MAKP signaling remains to be determined. Here we show that MKP1, a member of the dual specificity protein phosphatase family capable of inactivating MAPKs, is a transcriptional target of p53. MKP1 mRNA and protein levels were increased upon p53 activation in several well defined p53-regulated cell systems. p53 bound to a consensus p53 binding site located in the second intron of the MKP1 gene and transactivated MKP1 in reporter gene assays. Inhibition of phosphatase activity impaired p53-mediated G1 arrest in arrested human glioblastoma GM cells in response to growth factor stimuli. Importantly conditional expression of MKP1 prevented arrested human cancer cells from entering into the cell cycle. Thus, these results provide a novel mechanism by which p53 controls the cell cycle in response to the MAPK signaling in the absence of DNA damage and suggest that p53 may negatively control the MAKP pathway via MKP1.

The tumor suppressor p53 protein suppresses cell growth by inducing cell cycle arrest or apoptosis. Despite the fact that p53-dependent p21-mediated G 1 arrest induced by DNA damage is well defined, the role of p53 in the cell cycle in response to the MAKP signaling remains to be determined. Here we show that MKP1, a member of the dual specificity protein phosphatase family capable of inactivating MAPKs, is a transcriptional target of p53. MKP1 mRNA and protein levels were increased upon p53 activation in several well defined p53regulated cell systems. p53 bound to a consensus p53 binding site located in the second intron of the MKP1 gene and transactivated MKP1 in reporter gene assays. Inhibition of phosphatase activity impaired p53-mediated G 1 arrest in arrested human glioblastoma GM cells in response to growth factor stimuli. Importantly conditional expression of MKP1 prevented arrested human cancer cells from entering into the cell cycle. Thus, these results provide a novel mechanism by which p53 controls the cell cycle in response to the MAPK signaling in the absence of DNA damage and suggest that p53 may negatively control the MAKP pathway via MKP1.
The tumor suppressor p53 gene is the most commonly mutated gene in human cancer. The ability of p53 to induce cell cycle arrest and apoptosis plays a central role in its tumor suppression. In response to a variety of stimuli such as DNA damage, p53 becomes activated and, in turn, initiates its biological responses including cell cycle arrest, apoptosis, DNA repair, senescence, differentiation, and antiangiogenesis through transcriptional activation of specific target genes that carry p53 DNA binding sites (1)(2)(3). p53 can also promote apoptosis activity through transcriptional repression of certain genes that lack consensus binding site motifs (4). In some cell types, p53 induces apoptosis in the absence of transactivation and in the absence of new RNA and protein synthesis (5)(6)(7). Nevertheless transcriptional transactivation of p53 appears to be an important mechanism in mediating its cellular responses (8,9).
There is a growing list of transcriptional targets of p53, including p21, MDM2, Bax, Gadd45, 14-3-3, KILLER/DR5, Pigs, PERP, Noxa, and PUMA (10 -21), that lead to distinct outcomes in cells (3,22,23). The best characterized target of p53 through the transcriptional mechanism is p21 (10), a cyclin-dependent kinase inhibitor (24). Activation of p21 by p53 causes growth arrest at G 1 and G 2 /M phases of the cell cycle, and loss of p21 expression impairs p53-mediated cell cycle arrest (25)(26)(27). In addition, studies have demonstrated that p53 can transcriptionally activate Wip1, a phosphatase that negatively regulates the major growth signaling mitogen-activated protein kinase (MAPK) 1 in response to UV radiation (28,29). However, it remains unclear whether p53 can regulate the MAPK pathway in the absence of DNA damage.
The MAP kinase signaling pathway mainly consists of three subfamilies: the stress-activated protein kinase (SAPK/c-Jun NH 2 -terminal kinase (JNK)), the p38 MAPK, and the extracellular signal-related kinase (ERK). Members of the ERK family are induced by either growth factors or stresses, resulting in diverse outcomes in cells. Induction of ERK by growth factors leads to cell proliferation and growth. Both JNK and p38, on the other hand, are induced by stresses that can lead to cell differentiation and cell death. In the presence of stimuli, these MAPKs are activated through the reversible phosphorylation of both threonine and tyrosine residues of the TXY motif in the catalytic domain by upstream dual specificity kinases called MAPK kinases including MKK1/2, MKK3/6, and MKK-4, p38 MAPK, and JNK/SAPK (30 -32). Activation of MAP kinases triggers a signal cascade, leading to many changes including growth, differentiation, and apoptosis. The activities of MAPKs are negatively regulated by members of the MAPK phosphatases.
The MAPK phosphatases are a family of dual specificity protein phosphatases, which can dephosphorylate both phosphothreonine and phosphotyrosine residues, thus inactivating the activities of MAP kinases (33). This family includes MKP1, MKP2, MKP3, MKP4, MKP5, VHR, Pac1, hVH2, hVH3, Pyst1, and Pyst2 (34). MKP1, also known as CL100, 3CH134, and Erp (hereafter referred to as MKP1) (35)(36)(37), was the first member of this family to be identified as a phosphatase. MKP1 was originally identified as an oxidative damage-induced gene and an ERK-specific phosphatase (36 -38). It has now been shown that MKP1 can also inhibit JNK and p38 (39,40). Interestingly constitutive MKP1 expression blocks G 1 -specific gene expression (41), suggesting a role in cell cycle control. It has also been shown that overexpression of MKP1 inhibits UV-and cisplatininduced apoptosis (42,43). Taken together, these studies suggest a role of MKP1 in the cell cycle and chemosensitivity.
In this study, we show that MKP1 is a transcriptional target of p53. Activation of p53 induces the expression of MKP1 transcript and protein in several well defined p53-regulated cell systems. Furthermore p53 transcriptionally regulates MKP1 via binding to a p53 binding site located in the second intron of this gene. Induction of MKP1 by p53 is independent of MAP kinases. Importantly we demonstrate that inhibition of phosphatase activity impairs p53-mediated G 1 arrest in response to growth factor stimuli. Moreover we show that conditional expression of MKP1 blocks the entry of arrested cells into the cell cycle in response to mitogens. These results provide insights into the molecular basis of how p53 regulates the cell cycle in response to growth factor stimuli in the absence of DNA damage and suggest a novel mechanism by which p53 negatively controls the MAPK pathway.

MATERIALS AND METHODS
Cell Lines and Culture Conditions-The human glioblastoma cell line GM, provided by Ed Mercer (Thomas Jefferson University), was maintained in Earle's minimum essential medium as described previously (44). The human ovarian cancer cell line SKOV3 was maintained in McCoy's medium as described previously (45). The colon cancer cell line SW480 was maintained in culture as described previously (45). The human lung cancer cell line H460 was obtained from the American Type Collection (Manassas, VA) and maintained in RPMI 1640 medium as described previously (46). The human lung anaplastic carcinoma cell line Calu-6 was maintained in RPMI 1640 medium as described previously (47). These cells were supplemented with 10% fetal bovine serum and antibiotics at 37°C in a humidified atmosphere consisting of 5% CO 2 and 95% air. The Drosophila Mel-2 cells were purchased from Invitrogen and maintained in Schneider's insect medium supplemented with 10% fetal bovine serum and 2 mM glutamine plus antibiotics at 25°C.
Generation of H460 Cells Inducibly Expressing MKP1-To generate cell lines in which the human MKP1 gene is inducibly expressed, H460 cells were transfected with either pMEP4, an empty vector, or pMEP4-MKP1 using Lipofectin reagent (Invitrogen) as described previously (49). Transfected cells were selected with 250 g/ml hygromycin (Invitrogen) for 3 weeks, and individual clones were isolated as described previously (46). Clones that expressed MKP1 protein determined by Western blot analysis, as described below, were used in this study.
RNA Extraction and Northern Blot Analysis-Total cellular RNA was purified using the Trizol method (Invitrogen) according to the manufacturer's instructions. For Northern blot analysis, 10 g of total RNA was separated in a 1.5% formaldehyde-agarose gel and blotted to Hybond-N ϩ membrane (Amersham Biosciences). The blots were hybridized with the random primed radiolabeled 503-bp fragment of the MKP1 cDNA, which was generated by PCR with primer pairs of 5Ј-GAGCACATCGTGCCCAACG-3Ј and 5Ј-TGATGTCTGCCTTGTGGTT-GTC-3Ј. PCR was performed by using Pfu (Stratagene) as an enzyme and the cDNA from GM cells infected with p53 adenovirus as a template. The PCR conditions were as follows: 35 cycles of 94°C for 45 s, 64°C for 45 s, and 72°C for 3 min with one cycle of 72°C for 10 min for the final extension. Radioactive signals were analyzed by autoradiography.
cDNA Microarray Screening and Data Analysis-Total RNA from GM cells treated with or without dexamethasone (Dex) (1 M) for 10 h was purified. cRNA was prepared according to the Affymetrix GeneChip Expression technical manual followed by hybridizing to the Human Genome U95AV2 Array containing 12,625 genes. The hybridization and gene chip expression analysis was carried out at the Michigan Center for Genomic Technologies (MCGH) of Michigan Life Science Corridor at Wayne State University.
Western Blot Analysis-Whole cell lysates were prepared as described previously (49), and protein concentration was determined using the Protein Assay kit (Bio-Rad). Cell lysates (50 g) were electrophoresed through 7.5-12% denaturing polyacrylamide gels and transferred to a nitrocellulose membrane (Schleicher & Schuell, Inc.) by electroblotting. The blots were probed or reprobed with the antibodies, and bound antibody was detected using Enhanced Chemiluminescence Reagent (Amersham Biosciences) according to the manufacturer's protocol. Rabbit polyclonal anti-human MKP1 (c-19) was purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). The anti-human p53 monoclonal antibody Ab2 and the anti-human p21 WAF1 monoclonal antibody Ab1 were purchased from Calbiochem. Anti-actin antibody (AC-74) was purchased from Sigma.
Cell Cycle Studies-Subconfluent cells were maintained in medium with 0.5% FBS for GM cells and 0.1% FBS for H460 cells to arrest cells in G 0 /G 1 phase of the cell cycle. After 3 days, arrested GM cells were replaced with 10% FBS medium with or without dexamethasone (1 M) in the absence or presence of vanadate (5 M), while arrested H460 cells were cultured in medium with 10% FBS with or without CdSO 4 (5 M) for 24 h. The cells were then harvested and fixed with ice-cold 70% (v/v) ethanol for 24 h. After centrifugation at 200 ϫ g for 5 min, the cell pellet was washed with phosphate-buffered saline (pH 7.4) and resuspended in phosphate-buffered saline containing propidium iodide (50 g/ml), Triton X-100 (0.1%, v/v), and DNase-free RNase (1 g/ml). The cells were then incubated at room temperature for 1 h, and DNA content was determined by flow cytometry using a FACScan flow cytometer (BD Biosciences).
Luciferase Reporter Assays-Transfections for luciferase assays were carried out as described previously (50). In brief, SW480 cells were plated at 7.5 ϫ 10 5 /well in 6-well plates. The next day, the cells were co-transfected with 5 g of reporter constructs (pGL3-B-MKP1-L, pGL3-B-MKP1-LA, or pGL3-Basic), 600 ng of pCMV-based plasmids (pCMV-V, pCMV-p53, or pCMV-p53mt135), and 30 ng of pRLSV40 (Promega) using LipofectAMINE 2000 reagent (Invitrogen). Transfected cells were harvested 24 h later. Firefly luciferase activities were assayed using the dual luciferase reporter assay system (Promega) in a Turner TD20/20 luminometer and normalized to Renilla luciferase activity. To transfect Drosophila Mel-2 cells, the cells were co-transfected with 1 g of the MKP1-luciferase reporter constructs and 500 ng of pPacp53wt or pPacp53mt135 using FuGENE TM 6 reagent (Roche Applied Science) as described by the manufacturer. Cells were harvested after 24 h for assaying luciferase activity using the Single Luciferase Assay system (Promega). Luciferase activities were normalized to cellular protein, measured by the Bio-Rad protein assay system. The expression levels of p53 and p53mt135 in Drosophila Mel-2 cells were monitored by Western blotting.
Electrophoretic Mobility Shift Assay-Calu-6 cells infected with Ad-p53 as described previously (45) were harvested at 24 h after infection, and the nuclear extracts were prepared using a Nucbuster protein extraction kit (Novagen) according to the manufacturer's instructions.
The oligonucleotides corresponding to a potential p53 binding site in intron 2 of the MKP1 gene were as follows: 5Ј-GCTGGTCCTGCCC-AGGCAAATGGGCTTAG-3Ј and 5Ј-CTAAGCCCATTTGCCTGGGCAG-GACCAGC-3Ј. A mutant version of this site was obtained by replacing the conserved C and G with T (in lowercase) as follows: 5Ј-GCTGGTaC-TaCCCAGGCAAATGGGCTTAG-3Ј. The consensus binding site for p53 (5Ј-TACAGAACATGTCTAAGCATGCTGGGG-3Ј) was purchased from Santa Cruz Biotechnology, Inc. The oligonucleotide pairs were annealed and labeled with [␥-32 P]ATP using T4 polynucleotide kinase. Electro-phoretic mobility shift assays were carried out as described previously (50). In brief, 10 g of nuclear extract protein was incubated with ␥-32 P-labeled oligonucleotide probes (0.03 pmol) at 4°C for 20 min in a volume of 20 l of DNA binding buffer (10 mM HEPES, pH 7.6, 0.1% Nonidet P-40, 10% glycerol, 1 mM dithiothreitol, 5 mM MgCl 2 , 50 mM NaCl, and 0.5 g of poly(dI-dC) in the presence or absence of cold probe competitors. The reaction mixtures were resolved on a 5% nondenaturing polyacrylamide gel. For the supershift assays, 2 g of mouse monoclonal antibody Ab-1 (Oncogene Sciences) specific for p53 or p21 was preincubated with the nuclear extract prior to addition of the labeled probe and poly(dI-dC).

Identification of Additional p53 Target Genes Involved in
Cell Cycle Arrest in the Absence of DNA Damage-Despite a number of p53 targets being identified, the p53 signaling pathways are still not completely understood (23). We sought to identify additional transcriptional targets of p53 and characterize their roles in the p53 signaling pathways. We carried out a microarray screen to identify genes that are transcriptionally regulated by p53 in a well established GM cell system (44). The human glioblastoma cell line GM contains an endogenous mutant p53 and an exogenous Dex-inducible wild-type p53 (44). Upon addition of Dex to the medium, these cells induce the expression of p53, which, in turn, transcriptionally activates the expression of p21 (10), resulting in cell cycle arrest (10,44). Total RNA was extracted at 10 h after treatment with Dex, when p21 was induced (45), and subjected to cDNA microarray analysis on glass slides (Affymetrix), which contain 12,625 genes including expressed sequence tag sequences. RNA isolated from untreated cells was used as a control. The results were analyzed as described elsewhere (51). Using 2-fold change as a cut-off, 76 genes were induced, and 240 genes were decreased when p53 was activated (data not shown). Table I is a partial list of genes that increased expression upon p53 activation.
To confirm the accuracy of these results obtained with the microarray screen, we carried out Northern blot analysis. We first PCR-amplified cDNA fragments including MKP1 (CL100), IAP-1, and ADP-ribosylation factor-like protein 4 from the same RNA used in the cDNA array screen with Taq DNA polymerase. Each cDNA fragment was verified by DNA sequencing and then used as a probe in Northern blot analysis. Northern blot analysis demonstrated the induction of MKP1 and IAP-1 in GM cells treated with Dex for 10 h when p53 is activated (Fig. 1, A and B), thus confirming the results obtained from microarray analysis. We also confirmed that ADPribosylation factor-like protein 4 was induced (Fig. 1C, ADR). As expected, p21 was up-regulated upon p53 activation; this served as a positive control (Fig. 1D).
Induction of MKP1 in a p53-dependent Manner-We focused on MKP1 in this study for the following three reasons: 1) MKP1 was highly induced (about 47-fold induction) upon p53 activation (Table I); 2) MKP1 plays an important role in negatively regulating MAPK activity (33,34); and 3) although it is believed that there is a functional link between the p53 pathway and the MAPK pathway, the mechanism underlying such a link has not been previously clearly demonstrated.
Although Northern blot analysis demonstrated MKP1 induction by Dex in GM cells, it is possible that such induction may be due to Dex but not p53 as described recently (52)(53)(54). To determine whether p53 plays a role in MKP1 expression, we infected GM cells and the ovarian cancer cell line SKOV3 (both of which have mutant p53) with Ad-p53 and Ad-LacZ for 10 h. Total RNA was isolated and subjected to Northern blot analysis for MKP1 expression. As shown in Fig. 2, A and B, Ad-p53 infection induced MKP1 mRNA in both cell lines as compared with untreated or Ad-LacZ-infected cells in which no such induction was detected. To confirm these results, we showed that the level of MKP1 protein was induced in SKOV3 cells infected with Ad-p53 but is not evident in cells infected with control Ad-LacZ (Fig. 2C) and that this induction was not due to an unequal loading (Fig. 2C). In addition, we validated MKP1 induction by p53 in U2OS cells that contain a tamoxifeninducible p53 (55) and found that MKP1 was induced when the cells were subjected to increasing p53 expression by tamoxifen (data not shown). Taken together, our results strongly support that MKP1 is a potential new target of p53.
p53 Transcriptionally Induces MKP1 Expression-Our data demonstrate that p53 induces MKP1 expression in more than one cell system. It is well documented that p53 mediates cellular responses primarily through transcriptional activation of its targets, although the transcription-independent mechanisms also play a role (3). To determine whether p53-induced MKP1 induction is transcriptionally regulated, we treated GM cells with Dex in the presence or absence of the transcription inhibitor actinomycin D, and the expression of MKP1 was examined at different time points. Consistent with the results obtained with microarray screen and Northern blot analysis, the level of MKP1 protein was significantly increased at 8 and 12 h and decreased at 24 h, suggesting that MKP1 is a p53induced early response gene (Fig. 3A). Moreover actinomycin D inhibited induction of both p53 and MKP1 at all time points, indicating that induction of both genes occurs at the transcription level (Fig. 3A). In addition, we found that cycloheximide, a protein synthesis inhibitor, can block induction of p53, MKP1, and p21 protein (Fig. 3B). Thus, our results indicate that induction of MKP1 upon p53 activation in GM cells is regulated at the transcriptional level.
p53-dependent and MAPK-independent Induction of MKP1-It has been shown that the expression of MKP1 can be regulated by ERK and p38 or vice verse (34,40). To determine whether ERK and p38 mediate increased expression of MKP1 in GM cells, we induced p53 by Dex and examined the levels of phosphorylated (active) forms of ERK, p38, and JNK. As shown in Fig. 4, p53, p21, and MKP1 were induced between 8 and 24 h following treatment with Dex. However, neither phosphorylated ERK nor p38 was increased upon p53 activation, indicating that p53 has little effect on MAPK activities in GM cells. In addition, MAPK inhibitors such as PD98059 (39,40) had no effect on MKP1 induction in GM cells (data not shown). Thus, these results suggest that induction of MKP1 in GM cells is p53-dependent and MAPK-independent.
MKP1 Is a Direct Transcriptional Target of p53-We searched the DNA data base GenBank™ by using MKP1 cDNA sequence and identified a genomic clone containing the MKP1 gene (GenBank TM accession number NT_023169). Based on the information obtained from this clone, we found that the MKP1 gene consists of four exons separated by three introns (Fig. 5A), which is consistent with a previous report (56). When searching for p53 binding sites using the MatInspector software (57), we identified a candidate p53 binding site (GGTCCTGCCCAGGCAAATGGG) in intron 2 of the MKP1 gene (Fig. 5A).
To determine whether p53 could bind to this candidate p53 binding site, we performed electrophoretic mobility shift assays using an oligonucleotide including the p53 binding site in intron 2. Mobility shift of the MKP1 oligonucleotides incubated with the nuclear extract was observed, and the supershift was detected in the presence of p53 antibody but not p21 antibody (Fig. 5C). A similar result was obtained with the consensus binding site for p53 (58), which served as a positive control (Fig. 5C). Importantly the supershift was not ob-served with a oligonucleotide encoding a mutated p53 binding site. Moreover we found that the formation of an oligoprotein-p53 antibody complex was competed by unlabeled consensus binding site for p53 but not by a mutated MKP1 oligonucleotide (Fig. 5C). Taken together, these results strongly suggest that p53 specifically binds to the wild-type MKP1 oligonucleotides.
To further determine whether p53 could activate transcription of the MKP1 promoter in vivo, we generated the MKP1 promoter-luciferase reporter pGL3-B-MKP1-L that contains the promoter, intron 1, and intron 2 including the candidate p53 binding site. To determine the responsiveness of this construct to p53, we used the pGL3-B-MKP1-L vector, along with either empty vector pCMV-V, pCMV-wtp53, or pCMV-mtp53, to transfect the human colon cancer cell line SW480. Transfected cells were harvested after 24 h, and luciferase activity was assayed using the dual luciferase reporter assay system. As shown in Fig. 5D, luciferase activity was stimulated ϳ7-fold in cells co-transfected with the wild-type p53 vector pCMV-wtp53 as compared with cells either transfected with the empty vector pCMV-V or the mutant p53 vector pCMV-mtp53. Importantly luciferase activity did not increase when cells were co-transfected with pCMV-wtp53 and empty pGL3-basic vector. A similar induction of luciferase activity was found in transfections of the human lung cancer cell line Calu-6 with pGL3-B-MKP1-L and pCMV-wtp53 (data not shown). Finally we found that a 16-fold induction of promoter activity occurred in Drosophila Mel-2 cells transfected with p53expressing vector pPacp53 but not in cells transfected with pPacO or pPacp53mt135 (Fig. 5E). These results strongly suggest that p53 can activate transcription of the MKP1 promoter in vivo.
Since the gel shift demonstrated that p53 binds to the candidate p53 binding site in intron 2, we asked whether this binding site is responsible for p53-dependent transactivation.
To answer this question, we constructed the pGL3-B-MKP1-LA, which lacks intron 2, and then transfected it into SW480 cells. As shown in Fig. 5D, the pGL3-B-MKP1-LA was no longer responsive to p53, suggesting that this p53 binding site is required for transactivation by p53. Similar results were obtained with Calu-6 cells (data not shown). Interestingly cotransfections of the pGL3-B-MKP1-LA vector into Drosophila Mel-2 cells with pPacp53 resulted in a 5-fold increase in luciferase activity, suggesting that additional candidate p53 binding sites, possibly with low affinity, may reside in this construct and that these candidate sites may be only able to be activated by p53 in certain cell types. To further define regulation of MKP1 by p53 in the Drosophila Mel-2 system, we made two additional deletion constructs, pGL3-B-MKP1-LB and pGL3-B-MKP1-LC (Fig. 5B), which contain promoter-exon 1 and promoter-only sequence, respectively. Surprisingly transfections of pGL3-B-LB into Mel-2 cells with pPacp53 resulted in an ϳ10-fold increase in luciferase activity (Fig. 5E), whereas pGL3-B-LC was effectively inert (Fig. 5E). This suggests that the region between the promoter and intron 1 (e.g. exon 1) contributes to the p53 transactivation response in Drosophila Mel-2 cells. Both pGL3-B-MKP1-LB and -LC constructs were inactive in SW480 and Calu-6 cells (date not shown). Collectively these results strongly suggest that MKP1 is a direct transcriptional target of p53 via binding to a primary p53 binding site in intron 2 with a secondary effect at an upstream region.
The Role of Phosphatases in p53-mediated G 1 Arrest-Since activation of p53 induces cell cycle arrest or apoptosis mainly through activation of its downstream targets and since MKP1 is able to block MAPK-mediated growth signaling, we asked whether phosphatases play a role in p53-mediated growth arrest. To answer this question, we performed cell cycle analysis of the p53-inducible human glioblastoma cell line GM. GM cells were maintained in minimum essential medium with 0.5% FBS for 3 days to arrest cells in G 0 /G 1 phase of the cell cycle, and the effects of phosphatases on p53-mediated arrest were assessed in the presence or absence of the phosphatase inhibitor vanadate. Of note, vanadate has recently been shown to inhibit MKP1 activity in vivo (59). As shown in Fig. 6, the majority of cells were arrested in G 0 /G 1 following serum starvation. After replacement with normal medium containing 10% FBS, the arrested cells started to cycle. By 24 h, the S phase population had increased to 27%, and the G 2 /M populations had increased to 53% (Fig. 6). Consistent with the role of p53 in cell cycle arrest, a majority of cells remained G 0 /G 1 -arrested in cultures grown in the presence of Dex accompanying induction of p53 despite replacement with serum-containing medium. In contrast, when arrested cells were maintained in the medium with 10% FBS, Dex, and vanadate for 24 h, cell populations in S phase dramatically increased to 32% as compared with 16% in cells treated with Dex alone (Fig. 6, lower panel). These results suggest that inhibition of phosphatase activity can appreciably impair p53-mediated G 1 arrest. Importantly treatment with vanadate alone did not have any effect on the cell cycle (data not shown). Despite the fact that vanadate is not specific for MKP1, these results strongly suggest that phosphatases can play important roles in p53-mediated growth arrest in response to growth factor stimuli.
Conditional Expression of MKP1 Prevents Cells from Entering into the Cell Cycle-To further elucidate the direct role of MKP1 in growth arrest, we conditionally expressed MKP1 in the human lung cancer cell line H460 and then tested its effect on the cell cycle. We transfected the pMEP4-MKP1 (myc) expression vector, an inducible system in which the gene of interest is driven by the human metallothionein IIa promoter (39,42), into H460 cells and then selected with hygromycin B for individual clones that stably express MKP1 by Western blotting with either anti-c-Myc or anti-MKP1 antibody. The levels of MKP1 protein were controlled by adding CdSO 4 into the medium. As shown in Fig. 7A, addition of CdSO 4 to H460 MKP/clone 1 resulted in increased MKP1 expression at all time points tested. No induction was observed in a vector-transfected control clone. In addition, we isolated two more clones that show a similar induction of MKP1 as clone 1 (data not shown). By performing cell cycle analysis using flow cytometry, we showed that the G 0 /G 1 populations of MKP1 clone 1 increased from 65 to 81% before and after CdSO 4 treatment (induced MKP1), whereas the same treatment of vector clone did not affect the G 0 /G 1 populations (53 versus 52% before and  7 or lane 8). The reactions were electrophoresed on a 5% Tris borate-EDTA gel as described under "Materials and Methods." The consensus binding site for p53 served as a positive control (lanes 1-3). Unlabeled consensus binding site for p53 and a mutated MKP1 p53 binding site were used to compete for p53 binding (lane 9 and lane 10). A mutant version of the potential p53 binding site was also included (lanes 11-13). D, luciferase activity with SW480 cells. SW480 cells were co-transfected with pGL3-B-MKP1-L, pGL3-B-MKP1-LA, or empty vector pGL3-basic along with the wild-type p53-expressing vector pCMV-p53wt, the mutant p53-expressing vector pCMV-p53mt, or empty vector pCMV-V, respectively. pRLSV40 was added in transfections after CdSO 4 treatment) (Fig. 7B). Similar results were obtained in another two clones (data not shown). These results showed clearly that inducible expression of MKP1 itself can prevent arrested H460 cells from entering into the cell cycle in response to growth stimuli. Thus, our results strongly suggest that through up-regulation of MKP1, p53 may negatively regulate the cell cycle in response to growth factor stimuli. DISCUSSION p53-mediated cell cycle arrest in response to DNA damage involves activation of several p53 targets including p21, Gadd45, and 14-3-3 (2, 3, 23), but the role of p53 in regulating cell cycle progression in response to growth factor stimuli is not clearly established. In this study, we provide evidence that MKP1, the founding member of the dual specificity protein phosphatases capable of inactivating the activities of MAPKs by dephosphorylating both phosphothreonine and phosphotyrosine residues (33,34), is a direct transcriptional target of p53. Importantly overexpression of MKP1 impaired the entry of arrested cells into the cell cycle. These findings suggest that, through inactivation of the MAPK signaling via up-regulation of MKP1, p53 may play an important role in controlling entry into the cell cycle in response to growth factor stimuli in the absence of DNA damage.
MKP1 is induced by a variety of stimuli including growth factors and stressful stimuli (36 -38, 40, 59), but the mechanism of its regulation is not fully understood. It is believed that the expression of MKP1 can be regulated through multiple mechanisms involving both transcriptional and post-transcriptional pathways. For example, it has been shown that inhibition of ERK by the MEK1/2 inhibitor U0126 and p38 by SB203580 substantially inhibits MKP1 induction caused by arsenite and heat shock or H 2 O 2 , respectively (40,60). Inhibition of these two MAP kinases also abrogated lipopolysaccharide-induced MKP1 expression (54). Therefore, both p38 and ERK seem to play an important role in MKP1 expression induced by certain stimuli. We showed here that induction of MKP1 by p53 activation in GM cells is MAPK-independent because: 1) there was no detectable phosphorylated ERK and p38 induction upon p53 activation (Fig.  4), and 2) inhibition of MAPK activity by pharmacological inhibitors such as PD98059 did not abrogate MKP1 induction (data not shown).
The transcriptional regulation of the MKP1 gene is poorly understood. It has been shown that an 800-bp fragment of the human MKP1 gene exhibits promoter activity and can be activated upon serum stimulation (56). Several transcription regulatory elements were located to this region including AP1, AP2, Sp1, and cAMP-responsive element (56). In addition, it has been shown that upstream stimulatory factor, a ubiquitously expressed bHLHZip (basic region/helix-loop-helix/ leucine zipper) protein that binds to E boxes, can activate the MKP1 promoter (61). In the present study, we showed that MKP1 is induced in a p53-dependent manner in several p53- regulated cell systems (Fig. 2), and such induction occurred at the transcriptional level because actinomycin D blocked such induction (Fig. 3). It is well established that activated p53 functions as a transcription factor to activate its downstream targets by binding to p53 binding sites, commonly located in intron or promoter regions of genes (58). By examining the genomic sequence of the MKP1 gene, we found a potential p53 binding site within intron 2 (Fig. 5A). More importantly, this potential binding site can be bound and activated by p53 in vitro and in vivo (Fig. 5). Surprisingly a deletion construct lacking intron 2 (pGL3-B-MKP1-LA) still possessed luciferase activity in the Drosophila Mel-2 cells (Fig. 5E), although it did not have the same effects in human cells examined. The reason for such different results obtained with human and Drosophila cells is not clear, but it may be explained by the following possibilities. First, there may be additional p53 binding sites in this region that allow p53 to be bound and activated in Drosophila Mel-2 cells. In fact, we found a potential p53 binding site in intron 1 (GAACTTGTCATTGGCTTTGTTT) that has one base insertion between the critical positions C and G and also has two mismatches in the non-critical positions within the consensus p53 binding site (data not shown). This potential p53 binding site may be responsible for p53-dependent transcriptional activation in the Drosophila Mel-2 cells expressing a high level of introduced p53. This issue requires further investigation. Second, since the constructs lacking intron 2 did not exhibit transcriptional transactivation in the human cells tested, it is possible that this site has a low binding affinity and that activation of such a low affinity binding site by p53 may occur only at the high expression levels achieved in Drosophila Mel-2 cells. Furthermore Mel-2 cells are both Sp1 and UFS1 null, and it is not clear whether loss of Sp1/UFS in Mel-2 cells could influence the response of a secondary p53 binding site to exogenous p53. Third, we showed that the construct pGL3-B-MKP1-LB that contains the promoter and exon 1 but not intron 2 exhibited luciferase activity in Mel-2 cells, whereas pGL3-B-MKP1-LC that contains the promoter only did not (Fig. 5E). It is possible that regulatory elements located in exon 1 can also regulate the transcriptional response of a putative upstream element of p53 as observed for pGL3-B-MKP1-LB in Mel-2 cells (Fig. 5E). Our finding that pGL3-B-MKP1-LB has higher luciferase activity than pGL3-B-MKP1-LA suggests that a repressor element may occur in intron 1 (Fig. 5E). It will be of interest to test such possibilities. Taken together, our results suggest that despite the presence of p53 binding sites in the genomic sequences in many genes, different cell contexts (e.g. other transcription factors) may greatly influence binding of p53 to its different affinity binding sites. This may explain why activation of p53 induces cell cycle arrest in some cell types and causes apoptosis in other cell types.
What is the role of MKP1 in p53-mediated cellular responses? It has been shown that p53 can transcriptionally regulate a number of genes, leading to many changes including cell cycle arrest and apoptosis. Activation of some targets can substitute for the p53-mediated cellular responses, and it would be expected that MKP1 phosphatase may function in this manner. Consistent with this notion, we showed clearly that inhibition of phosphatases by vanadate impairs p53-mediated G 1 arrest (Fig. 6), which suggests that p53-mediated G 1 arrest involves activation of phosphatases. Despite the fact that vanadate is not specific for MKP1, elucidation of the role of phosphatases in p53-mediated growth arrest is both a novel and potentially important area of investigation. More importantly, we showed that conditional expression of MKP1 blocks arrested H460 entry into the cell cycle in response to growth factor stimuli (Fig. 7). This is in agreement with a previous report showing that constitutive MKP1 expression blocks G 1specific gene expression (41). However, we were unable to show the different induction of MKP1 in cells with wild-type p53 and without functional p53 after DNA damage (data not shown). This finding is not surprising considering that MKP1 plays a major role in inactivation of the MAPK signaling upon growth factor stimuli. In addition, we do not know how p53 activates MKP1 in response to growth factor stimuli, and our speculation is that under non-stressful conditions p53 may activate MKP1 to bring the MAPK signaling in check. Nevertheless our results suggest that induction of MKP1 by p53 regulates G 0 /G 1 transition in response to growth factor stimuli in the absence of DNA damage.
It has been shown that there is a functional link between the p53 pathway and the MAPK pathway. On the one hand, MAP kinases act as upstream activators to positively regulate p53 activity. For example, MAP kinases, including JNK, ERK, and p38, can phosphorylate p53 at various sites in response to a variety of stressful stimuli (62-66). Because p53 modifications including its phosphorylation are required for its activation, FIG. 7. Overexpression of MKP1 prevents cells from entering into the cell cycle. A, H460 cells were transfected with either pMEP4 or pMEP4-MKP1, and hygromycin-resistant clones were selected. The hygromycin-resistant clones were isolated and treated with CdSO 4 (5 M) for 0, 3, 6, 12, and 24 h. Induction of MKP1 was detected by Western blotting in MKP1/clone 1 but not in the vector control clone. B, both vector and MKP1expressing H460 cells were arrested by serum starvation for 3 days. Arrested cells were reinduced into the cell cycle by replacement with the media containing 10% FBS with or without CdSO 4 (5 M). After 24 h, cells were harvested for fluorescence-activated cell sorting analysis. Cells before stimulation were designated as 0 h. Data are representative of three independent experiments. phosphorylation of p53 by MAPKs leads to subsequent growth arrest or apoptosis (62)(63)(64)(65). On the other hand, p53 can transcriptionally activate the genes that are the components of the MAPK signaling pathway, leading to cell cycle arrest or apoptosis. For example, Wip1, a member of the protein phosphatase type 2C family, is a p53 target involved in growth arrest and apoptosis in response to ionizing and UV radiation (28,29). p53 has also been shown to activate the MAPK pathway via a mechanism involving up-regulation of heparin-binding epidermal growth factor-like growth factor (67,68). Induction of heparin-binding epidermal growth factor-like growth factor by p53 is believed to counteract p53-mediated growth suppression (68). Interestingly two additional genes, Cox-2 and DDR1, have been shown to be downstream targets of p53 whose expression protects cells from p53-mediated apoptosis and chemosensitivity (69,70). Therefore, p53 may regulate the MAPK pathway as a protective mechanism.
In this study, we showed that MKP1 is a direct transcriptional target of p53. Since the MAPK pathway is the major growth signaling pathway and since MKP1 is able to inactivate the MAPK-mediated growth signal, identification of MKP1 as a p53 target fits well with the idea in which the main role of p53 is to suppress cell growth by inducing cell cycle arrest or apoptosis. Thus, the identification of MKP1 as a new p53 target involved in the cell cycle regulation suggests a new mechanism by which p53 negatively controls the MAPK pathway. Interestingly while this manuscript was under preparation, Yin et al. (71) reported that Pac1, another member of the dual specificity protein phosphatase family, is a new target of p53 whose expression plays a role in oxidative damage-induced p53-mediated apoptosis. Thus, it is conceivable that p53 may negatively regulate the MAPK survival signaling pathway by inducing cell cycle arrest via MKP1 and apoptosis via Pac1, respectively.
In conclusion, we have found that MKP1 is a new p53 target and negatively regulates G 1 progression of the cell cycle in response to growth factor stimuli in the absence of DNA damage. Since induction of cell cycle arrest and apoptosis is critical for tumor suppression by p53, future experiments will determine whether MKP1 plays a role in p53mediated apoptosis. In addition, since MKP1 has been implicated in modulation of chemosensitivity (43), it would also be interesting to determine whether MKP1 is involved in p53-mediated chemosensitivity.