Alterations in Subcellular Localization of p38 MAPK Potentiates Endothelin-stimulated COX-2 Expression in Glomerular Mesangial Cells*

Endothelin-1 (ET-1) is a potent vasoconstrictor peptide with mitogenic actions linked to activation of tyrosine kinase signaling pathways. ET-1 induces cyclooxygenase-2 (COX-2), an enzyme that converts arachidonic acid to pro-inflammatory eicosanoids. Activation of each of the three major mitogen-activated protein kinase (MAPK) pathways, ERK1/2, JNK/SAPK, and p38 MAPK (p38), have been shown to enhance the expression of COX-2. Negative regulation of MAPK may occur via a family of dual specificity phosphatases referred to as mitogen-activated protein kinase phosphatases (MKP). The goal of this work was to test the hypothesis that wild type MKP-1 regulates the expression of ET-1-induced COX-2 expression by inhibiting the activation of p38 in cultured glomerular mesangial cells (GMC). An adenovirus expressing both wild type and a catalytically inactive mutant of MKP-1 (MKP-1/CS) were constructed to study ET-1-regulated MAPK signaling and COX-2 expression in cultured GMC. ET-1 stimulated the phosphorylation of ERK and p38α MAPK and induced the expression of COX-2. Expression of COX-2 was partially blocked by U0126, a MEK inhibitor, and SB 203580, a p38 MAPK inhibitor. Adenoviral expression of MKP-1/CS augmented basal and ET-1-induced phosphorylation of p38α MAPK with less pronounced effects on ERK1/2 phosphorylation. Ectopic expression of wild type MKP-1 blocked the phosphorylation of p38α MAPK by ET-1 but increased the phosphorylation of p38γ MAPK. Co-precipitation studies demonstrated association of MKP-1 with p38α MAPK and ERK1/2. Immunofluorescent image analysis demonstrated trapping of phospho-p38 MAPK in the cytoplasm by MKP-1/CS/green fluorescent protein. ET-1-stimulated expression of COX-2 was increased in MKP-1/CS versus LacZ or green fluorescent protein-infected control cells. These results indicate that MKP-1 demonstrates a relative selectivity for p38α MAPK versus p38γ MAPK in GMC and is likely to indirectly regulate the expression of COX-2.

Selective inhibition of cyclooxygenase-2 (COX-2) 1 is a therapeutic strategy for limiting the production of pro-inflammatory eicosanoids (1). Interestingly, the use of selective COX-2 inhibitors is associated with an increased risk of acute renal failure, although the exact mechanisms are unknown (2). Endothelin-1 (ET-1) is a 21-amino acid peptide best known for its potent vasoconstrictor properties with probable contributions to renal inflammatory disease like glomerulonephritis (3,4). Additionally, ET-1 induces the expression of COX-2 in rat mesangial cells (RMC) (5). Investigations into possible signaling mechanisms involved in the regulation of COX-2 expression indicate that in 3T3 fibroblasts and human umbilical vein endothelial cells, activation of ERK1/2 by adenoviral delivery of constitutively active MEK1 induces the expression of COX-2. Additionally, human mesangial cells infected with adenovirus expressing constitutively active mitogen-activated protein kinase kinase-3/6 and constitutively active mitogen-activated protein kinase kinase-7 demonstrated that the p38 MAPK and JNK/ SAPK, respectively, induce the COX-2 enzyme (6). Significantly, a major mechanism of action attributed to activation of p38 MAPK appears to be stabilization of COX-2 and other mRNAs (7,8).
p38 MAPK belongs to the canonical MAPK pathway and is expressed as four isoforms (␣, ␤, ␥, and ␦) (9,10). Several pharmacological antagonists (including SB 203580) are potent inhibitors of the ␣ and ␤ isoforms (11). Activation of p38 MAPK occurs as a result of tyrosine and threonine phosphorylation within the TGY motif (9). A downstream target of p38 MAPK includes MAPKAP-2, that when phosphorylated by p38 MAPK, translocates to the cytoplasm and phosphorylates HSP27 (12,13). Inactivation of p38 MAPK results from the coordinated removal of phosphates from tyrosine, threonine, or both residues (14). Although serine/threonine phosphatases such as PP2A members (15) or protein tyrosine phosphatases such as PTP-SL or STEP are capable of inactivating MAPK (16), it is also clear that a family of dual-specificity phosphatases (DSP), the mitogen-activated protein kinase phosphatases (MKPs), are emerging as likely candidates for the negative regulation of MAPKs in vivo (14).
Although transcriptional regulation appears to be a primary mechanism of DSP expression (21), subcellular localization of DSPs also may contribute to their regulation. For example, MKP-1 is a nuclear localized enzyme (22), whereas MKP-3 is retained in the cytoplasm (23). In addition, the phosphatase activity of MKP-3 is stimulated by binding of ERK to its Nterminal domain (24). Similarly, the catalytic activity of MKP-1 is increased by C-terminal binding of JNK1, ERK1/2, and p38␣ MAPK (25). These results indicate that substrate specificity is tightly linked with protein binding domains, which can have a significant impact on the catalytic activity of the DSP. The DSPs MKP-3 and M3/6 appear to demonstrate selectivity toward ERK and p38 MAPK, respectively (26), whereas MKP-1 has been reported to inactivate target kinases including ERK, SAPK/JNK, and p38 MAPK (27)(28)(29). The reasons for these observations are unclear but may partially rely on the cell type and experimental stimulus used to activate the respective MAPK pathways.
Previously, we found that ET-1 induced the expression of MKP-1 in a time frame compatible with a model for negative feedback control of ERK1/2 activation (30). However, those studies failed to address the potential role of other MAPK pathways in regulating the expression of COX-2. Therefore, the present work utilizes adenoviral delivery of wild type and a catalytically inactive, "substrate trapping" mutant of MKP-1 to determine their potential role in regulating the MAPKs, ERK1/2 and p38, and COX-2 expression in glomerular mesangial cells in response to ET-1 treatment (31). Additionally, studies were performed with inhibitors of MEK and p38 MAPK to determine the importance of these MAPK pathways for ET-1stimulated expression of COX-2. Our findings provide initial evidence for a substrate specificity of MKP-1 for p38␣ MAPK but not p38␥ MAPK and demonstrate that alteration in the localization of active p38 MAPK by the catalytically inactive form of MKP-1 can result in enhanced expression of COX-2. Based on these results and those of earlier studies, we have proposed a model for DSP regulation of p38-dependent COX-2 regulation in glomerular mesangial cells.
Cell Culture and Transfection-Primary RMCs were cultured from isolated glomeruli from male Sprague-Dawley rats as described previously (32). Mesangial cells were cultured in RPMI 1640 containing HEPES (25 mM) and supplemented with 17% fetal bovine serum, 100 units/ml penicillin, 100 g/ml streptomycin, 5 g/ml each insulin and transferrin, and 5 ng/ml selenium at 37°C in an atmosphere of 95% air/5% CO 2 . For some experiments, a transformed human mesangial cell line was used and cultured in RPMI 1640 containing HEPES (25 mM) and supplemented with 10% fetal bovine serum and antibiotics. Human embryonic kidney (HEK) 293 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin and incubated under the same conditions as RMC. Cells were typically cultured in 60or 100-mm Petri dishes and serum-starved for 48 h prior to the addition of stimulus.
Antibodies and Plasmid Vectors-Rabbit polyclonal phospho-MAPK, phospho-p38 antibodies, and phospho-HSP27 were from Cell Signaling Technology (Beverly, MA). Mouse anti-phospho p38 MAPK antibodies were from Sigma. Goat polyclonal antibodies against COX-2 (N20) and p38 (C20) were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Antibodies against p38␥ MAPK were from Upstate Biotechnology (Lake Placid, NY). Rabbit polyclonal antibodies to MKP-1 were produced by immunizing rabbits with a synthetic peptide derived from the C terminus of the protein (21). Mouse monoclonal antibodies against the FLAG epitope were from Sigma and were used for immunoprecipitation of p38-FLAG. Mouse monoclonal antibodies recognizing the hemagglutinin epitope (HA) (Santa Cruz Biotechnology, Inc.) were used for immunoprecipitation of HA-ERK. Rabbit polyclonal antibodies recognizing HSP27 were from Stressgen (Victoria, British Columbia, Canada). Goat anti-rabbit IgG or anti-mouse IgG (Sigma) conjugated to TRITC were used as the secondary antibody for detection of p38 and p38-HA, respectively.
MKP-1 was mutated to the catalytically inactive form as follows. The cDNA encoding wild type MKP-1 was mutated using the Transformer site-directed mutagenesis kit (Clontech). In position 1006, thymidine was changed to adenine resulting in the substitution of serine for  cysteine. The mutation was confirmed by manual sequencing using Sequenase version 2.0 (United States Biochemical, Cleveland, OH). The primers used for mutagenesis were as follows: catalytic domain, 5Ј-GTGTTTGTCCCACAGCCAGGCAGGC-3Ј; NotI site introduced for selection, 5Ј-CCGAGCTCCGAGGCCGCCTGCTG-3Ј.
The cDNA encoding wild type MKP-1 and MKP-1/CS was digested with KpnI and PvuII and ligated into pEGFP(N3) vector (Clonetech) to create the 67-kDa fusion protein. HA-ERK was a kind gift from Dr. Joseph Barbieri (Medical College of Wisconsin, Milwaukee, WI), and HA-p38 was a kind gift from Dr. Steven Keyse (Biomedical Research Centre, Dundee, Scotland).

Construction of Adenoviral Recombinant MKP-1/CS and Infection of Primary Rat Mesangial Cells-The recombinant adenoviral vectors
AdvCMVMKP-1/CS and AdvCMVMKP-1 were constructed according to a method described previously (30). Briefly, cDNA encoding MKP-1/CS or MKP-1 was digested with KpnI and XhoI and cloned into pCA3 shuttle vector for homologous recombination in HEK 293 cells. Linearized plasmid was cotransfected with ClaI-digested Ad-dl327 using HEK 293 cells. Individual plaques resulting from virion formation were isolated and amplified in HEK 293 cells, and viral DNA was isolated (33) and analyzed for identification of recombinant viruses by HindIII DNA restriction analysis. High titer preparations (1 ϫ 10 10 to 1 ϫ 10 11 pfu/ml) were made using HEK 293 cells, and resultant viral DNA was purified by CsCl density gradient centrifugation (34). Purified viral DNA was dialyzed and stored at Ϫ70°C in 10 mM Tris-HCl, pH 7.4, 1 mM MgCl 2 , and 10% glycerol until used. Virus titer was determined by plaque assay using HEK 293 cells (34).
In Vitro Transcription/Translation and Immunoprecipitation-Enhanced GFP and wild type MKP-1/GFP were subcloned into pcDNA 3.1 (ϩ). ERK2 was subcloned into pBKII. 4.5 g of each plasmid was transcribed and translated using the Promega TNT kit. Equal volumes of [ 35 S]methionine-labeled proteins were mixed and immunoprecipitated with anti-FLAG or anti-HA antibody followed by electrophoretic separation on a 10% SDS-polyacrylamide gel, and labeled proteins were detected by autoradiography.
Metabolic Labeling, Cell Lysis, Immunoprecipitation, and Immunoblot Analysis-Cell labeling was performed essentially as described (35). Metabolic labeling of RMC was performed by incubating cells in phosphate-free RPMI 1640 supplemented with 0.5 mCi/ml [ 32 P]orthophosphate for 3 h at 37°C. After stimulation with ET-1 for the appropriate time, the labeling solution was removed and discarded. The cells were lysed as described previously (30). Briefly, cells were washed twice with ice-cold phosphate-buffered saline (PBS) after stimulation. Cells were lysed for 20 min at 4°C in 500 l of Triton lysis buffer containing the following (in mM): HEPES (50), pH 7.5, NaCl (150), MgCl 2 (1.5), EGTA (1), 1% Triton X-100, 10% glycerol, protease inhibitor mixture, phenylmethylsulfonyl fluoride (1), sodium orthovanadate (0.2), and sodium fluoride (10). Cell lysates were collected in microcentrifuge tubes and cleared by centrifugation at 15,000 ϫ g for 15 min at 4°C. Supernatants were then resolved by SDS-PAGE for Western blot analysis or subjected to immunoprecipitation.
Cells lysates, normalized for protein, were immunoprecipitated with antibodies against MKP-1 for 1.5 h at 4°C with constant rotation. Protein A-Sepharose was added, and the lysates were incubated for an FIG. 6. Adenoviral-delivered MKP-1 preferentially inactivates p38 MAPK. RMC and HMC were infected with adenovirus encoding GFP (AdvGFP) or MKP-1 (AdvMKP-1). The cells were serum-starved for 48 h and stimulated with 100 nM ET-1 for the indicated times. The resultant cell lysates were normalized for protein and resolved on 10% SDS-polyacrylamide gel. After transfer to PVDF the membranes were probed with anti-phospho-ERK (A) or anti-phospho-p38 antibodies (B). additional 1 h. After washing the immunoprecipitates three times in Triton lysis buffer, 2ϫ Laemmli sample buffer was added, and the samples were boiled for 5 min, subjected to SDS-PAGE, and transferred to a polyvinylidene difluoride (PVDF) membrane (Millipore, Bedford, MA). After transfer of proteins, the PVDF membrane was blocked with Tris-buffered saline (TBS) containing the following: 20 mM Tris-HCl, pH 7.8, 150 mM NaCl, and 3% bovine serum albumin (United States Biochemical, Cleveland, OH) for 2 h at room temperature. Primary antibodies were added, and the membranes were incubated overnight at 4°C with constant rotation. The following day the membranes were washed five times with TBS containing 0.1% Tween 20 followed by two washes with TBS prior to the addition of secondary antibodies. Secondary antibodies were diluted in TBS containing 5% non-fat dry milk and incubated with the membrane for 1 h at room temperature. The membranes were again washed five times with TBS containing 0.1% Tween 20 and twice with TBS prior to development of immunoreactive signal by chemiluminescence.
Immunofluorescence and Digital Imaging-Human mesangial cells were plated to 60-mm dishes prior to transfection. The cells were transfected the following day with a total of 4 g of cDNA using LipofectAMINE. The following day the cells were plated to 4-well glass chamber slides. The next day the cells were washed twice with PBS and fixed in 2% paraformaldehyde in PBS, pH 7.4, for 15 min at room temperature. The cells were washed three times with PBS and permeabilized with 0.2% Triton X-100 in PBS for 10 min at room temperature. The cells were washed once for 5 min with PBS containing 0.2% Tween 20 and incubated overnight at 4°C with primary antibodies in PBS containing 0.1% Tween 20 and 5% normal goat serum. The following day the primary antibodies were removed, and the cells were washed three times over 15 min with PBS containing 0.2% Tween 20. The cells were then incubated for 1 h with secondary antibody in PBS containing 5% normal goat serum and 0.1% Tween 20. The cells were washed three times over 15 min with PBS containing 0.2% Tween 20 and once for 5 min with PBS. Nuclei were stained with DAPI (50 ng/ml) in PBS for 5 min. The cells were washed once with PBS, a coverslip was mounted, fluorescence was detected, and images were taken using a Nikon Eclipse microscope (model E600) with an RT Slider SPOT CCD camera and Spot Advanced imaging software.

Differential Regulation of COX-2 Expression in Mesangial
Cells Transduced with Wild Type or Catalytically Inactive MKP-1-ET-1-induced expression of COX-2 was determined in RMC infected with GFP, wild type, or catalytically-inactive MKP-1 (CS). COX-2 protein was increased after 4 h of treatment with ET-1 when compared with serum-restricted control cells (Fig. 1). Adenoviral delivery of wild type MKP-1 resulted in diminished basal and ET-1-induced expression of COX-2 whereas MKP-1/CS-infected cells had enhanced basal and ET-1-induced expression of COX-2. Notably, ET-1 stimulated an increase in adenoviral delivered MKP-1 at 4 h when compared with control. The reason for this increase is unclear at present, although evidence from other laboratories indicates that the turnover of MKP-1 may be slowed in an ERK-dependent manner by phosphorylation of C-terminal serine residues of MKP-1 (36). There was no apparent change in the expression of the catalytically inactive MKP-1 mutant, suggesting that stabilization of this mutant with an activated (i.e. phosphorylated) substrate may protect it from degradation.
Constitutively active members of the MEKs have been shown previously (6) to drive the expression of COX-2. Thus, experiments were conducted to determine whether inhibition of MEK1/2, the upstream kinase for ERK1/2, and/or inhibition of p38 MAPK could account for the observed ET-1-induced increase in COX-2 expression. The MEK1/2 inhibitor, U0126, completely inhibited the phosphorylation of ERK1/2 in response to ET-1 in RMC at 5 and 10 M (Fig. 2, top panel) but had no effect on the phosphorylation of p38 MAPK (Fig. 2, middle panel). Furthermore, U0126 had no effect on the phosphorylation of HSP27, an indirect measure of p38 MAPK activity (Fig. 2, bottom panel) (9). To support this, we found that SB 203580, a direct inhibitor of p38 MAPK (11), blocked ET-1stimulated phosphorylation of HSP27 at 10 M with little additional effect at 25 M (Fig. 2, bottom panel). SB 203580 had no effect on the phosphorylation of ERK1/2 at either concentration tested (Fig. 2, top panel).
Once the concentrations of inhibitors to achieve specificity were determined, their effects on ET-1-stimulated expression of COX-2 were examined. Expression of COX-2 in RMC pretreated with either U0126 or SB 203580 alone was reduced when compared with control cells (Fig. 3A). Combined treatment of RMC with U0126 and SB 203580 resulted in COX-2 expression that was only slightly less than either inhibitor used alone (Fig. 3B). Taken together, these data suggest that ET-1 stimulates the expression of COX-2 in an ERK1/2 and p38-dependent manner in RMC.
Temporal Activation of ERK and p38 MAPK by ET-1 in Rat Mesangial Cells-Mutation of cysteine 258 to serine within the active site of MKP-1 was shown previously (35) to stabilize its association with one of its proposed substrates, phospho-ERK. Thus, we examined ET-1-stimulated RMC labeled with [ 32 P]orthophosphate to determine whether antibodies against MKP-1 might co-precipitate phosphorylated protein(s). Cell lysates from RMC infected with an adenovirus encoding MKP-1/CS co-precipitated a phosphoprotein using anti-MKP-1 antibodies (Fig. 4). Phosphorylation of this protein was enhanced by the addition of ET-1 in a time-dependent manner. The phosphoprotein detected was in the molecular weight range of the MAPKs, ERK1/ERK2 and p38 MAPK.
To identify the MAPK that co-immunoprecipitated with catalytically inactive MKP-1, we first investigated the effects of the cysteine-to-serine mutant on ET-1-induced phosphorylation of ERK1/2 and p38 MAPK. We have demonstrated previously (30) that recombinant adenovirus can deliver a gene of interest to 100% of primary RMC. Therefore, experiments were performed to determine whether MKP-1/CS could exert a dominant interfering effect on ERK1/2 and p38 MAPK signaling cascades in RMC and human mesangial cells (HMC). Phosphorylation of ERK1/2 and p38 MAPK were followed at several time points after the addition of ET-1.
Parallel dishes of RMC and HMC were infected with an adenovirus encoding LacZ or GFP, as reporter genes for in- fection. RMC and HMC stimulated with ET-1 displayed biphasic kinetics of ERK1/2 phosphorylation (Fig. 5A, upper and lower panels, respectively). ERK1/2 phosphorylation was maximal at 5 min, returned to basal by 15 min, and was again phosphorylated within 4 h of ET-1 treatment in RMC and within 3 h of treatment in HMC. Adenoviral delivery of MKP-1/CS resulted in elevated basal phosphorylation of ERK1/2 with marginal enhancement of ERK1/2 phosphorylation in response to ET-1. In contrast, adenoviral delivery of MKP-1/CS greatly accentuated the basal and ET-1-stimulated phosphorylation of p38 MAPK in RMC and HMC (Fig.  5B, upper and lower panels, respectively). These results sug-gest that MKP-1/CS may stably interact with p38 MAPK to maintain its phosphorylated state.
Adenoviral-delivered MKP-1 Attenuates ET-1-stimulated Phosphorylation of ERK1/2 and p38 MAPK-Because adenoviral delivery of MKP-1/CS resulted in a greater accentuation of basal and ET-1 stimulated phosphorylation of p38 MAPK than ERK1/2, we hypothesized that adenoviral delivery of the wild type form of MKP-1 would result in greater inhibition of basal and ET-1-stimulated phosphorylation of p38 MAPK. Fig. 6A demonstrates a partial inhibition of ERK1/2 phosphorylation stimulated by ET-1 in HMC (lower panel) but not RMC (upper panel) infected with an adenovirus encoding wild type MKP-1 when compared with cells infected with GFP. However, ET-1stimulated phosphorylation of p38 MAPK was markedly reduced in the same cell lines transduced with MKP-1 when compared with cells infected with GFP (Fig. 6B, upper and  lower panels). These data support a role for MKP-1 in regulating both p38 MAPK and ERK1/2 in glomerular mesangial cells of both species.
MKP-1/CS Traps phospho-p38 in the Cytoplasm-Because our data indicated that MKP-1/CS enhances the basal and ET-1-induced phosphorylation of p38 MAPK, we examined the localization of phospho-p38 MAPK to identify potential intracellular sites of MKP-1 and p38 MAPK interactions. A monoclonal antibody raised against the dually phosphorylated (i.e. active) p38 MAPK was used. Mesangial cells co-transfected with the GFP vector and p38-HA (Fig. 7, VECTOR) were compared with cells co-transfected with p38-HA and MKP-1/GFP (Fig. 7, WT) or p38-HA and MKP-1/CS/GFP (Fig. 7, CS). As illustrated in Fig. 7 (WT column), a mesangial cell co-expressing the GFP fusion of wild type MKP-1 and p38-HA resulted in no detectable phospho-p38 immunoreactivity. In contrast, the two cells negative for MKP-1/GFP expression (i.e. lack of fluorescence in GFP row; see Fig. 7, WT column) were positive for phospho-p38 immunoreactivity. In striking contrast, cells pos- itive for MKP-1/CS/GFP expression (Fig. 7, CS column) were also positive for phospho-p38 MAPK that was restricted to the cytoplasm. Thus, these studies would indicate for the first time that the substrate trapping MKP-1 mutant (MKP-1/CS) effectively retains phospho-p38 MAPK in the cytoplasm. Because nuclear/cytoplasmic shuttling is important for proper MAPK function, the restriction of the active form of p38 MAPK to the cytoplasm by MKP-1/CS may provide a unique tool to dissect nuclear and cytoplasmic p38-dependent signaling.
Adenoviral Delivery of Wild Type MKP-1 Results in Phosphorylation of p38␥ MAPK-In an effort to support the observation that the monoclonal antibody used to detect phospho-p38 in Fig. 7 was specific, this same antibody was used to detect ET-1 induced phosphorylation of p38 MAPK by Western blot in RMC infected with GFP, wild type, or catalytically inactive MKP-1. As was shown previously with the polyclonal antibody (see Fig.  5B and Fig. 6B), the anti-active p38 MAPK monoclonal antibody also detected ET-1-stimulated phosphorylation of p38 MAPK in the GPF-infected cells (Fig. 8A). The phosphorylation of p38 MAPK induced by ET-1 was abolished in cells infected with wild type MKP-1, whereas cells infected with MKP-1/CS displayed an increased basal phosphorylation of p38 MAPK that was further enhanced by the addition of ET-1. Notably, monoclonal anti-active p38 MAPK also detected a higher molecular mass phosphorylated protein in those cells infected with the wild type phosphatase (Fig. 8A, top panel). As the p38␥ isoform migrates with a molecular mass of ϳ43 kDa (10), the same lysates were probed with isoform-specific antibodies directed against p38␥ MAPK and p38␣ MAPK. Both isoforms were detected in the mesangial cell lysates (Fig. 8A, middle panels). Summarized data from three independent experiments demonstrate a consistent finding that expression of wild type MKP-1 greatly attenuates ET-1-stimulated p38␣ MAPK but enhances p38␥ MAPK phosphorylation (Fig. 8B). These data provide initial evidence for the in vivo specificity of wild type MKP-1 for p38␣ but not p38␥ MAPK. Additionally, the observation that overexpression of the wild type form of MKP-1 results in activation of p38␥ suggests a complex intra-MAPK signaling network.
p38 MAPK Co-precipitates with Wild Type MKP-1 in Vitro and in Vivo-To investigate a possible direct interaction of p38␣ MAPK or ERK with wild type MKP-1, cDNAs encoding MKP-1, MKP-1/GFP, ERK2, and FLAG-tagged p38␣ MAPK (p38-FLAG) were in vitro transcribed and translated in rabbit reticulocyte lysates. All proteins were efficiently translated, and antibodies against the FLAG epitope efficiently immunoprecipitated p38-FLAG and MKP-1/GFP. Antibodies directed against ERK2 efficiently immunoprecipitated ERK2, but a lower amount of MKP-1 was co-precipitated (Fig. 9).
Given these results, experiments were conducted in transiently transfected HEK 293 cells to determine whether p38␣ MAPK associates with MKP-1. Experiments were performed in HEK 293 cells co-transfected with MKP-1, MKP-1/GFP, and MKP-1/CS/GFP, in combination with either p38-FLAG or HA-ERK1. Wild type MKP-1 (39 kDa), MKP-1/GFP (67 kDa), and MKP-1/CS/GFP (67 kDa) were efficiently expressed in cells co-transfected with p38-FLAG (Fig. 10A, lanes 1, 3, and 5). Interestingly, only the GFP fusions of MKP-1 and MKP-1/CS were detected in cells co-transfected with HA-ERK1 (Fig. 10A,  compare lanes 4 and 6 with lane 2). As was observed with the in vitro experiment MKP-1/GFP and MKP-1/CS/GFP co-precipitated with p38-FLAG from cell lysates (Fig. 10B). Additionally, the native 39-kDa form of MKP-1 also co-precipitated with p38-FLAG (Fig. 10B). In contrast, little MKP-1/GFP or MKP-1/CS/GFP was detected in cell lysates immunoprecipitated with anti-HA antibodies (Fig. 10C). Taken together, these data support a direct and specific interaction of MKP-1 with the p38␣ MAPK isoform. DISCUSSION The findings of this study suggest that the p38 MAPK pathway acts in concert with the ERK1/2 pathway to stimulate the expression of COX-2. Specifically, we demonstrated an interaction between wild type MKP-1 and p38␣ MAPK consistent with the hypothesis that MKP-1, in addition to dephosphorylating ERK1/2, also may act as an in vivo regulator of p38␣ MAPK activity. Novel aspects of this study include the observations that: 1) catalytically inactive MKP-1 traps phospho-p38 MAPK in the cytoplasm, preventing any potential nuclear shuttling of p38 MAPK, and 2) adenoviral-delivery of wild type MKP-1 abolishes basal and ET-1-stimulated phosphorylation of p38␣ MAPK but stimulates the activation of p38␥ MAPK. Thus, we propose a model of ET-1 signaling in glomerular mesangial cells that is dependent on ERK1/2, p38 MAPK, and the DSP MKP-1. Importantly, trapping of active p38 MAPK in the cytoplasm may result in the potentiation of COX-2 expression. Significantly, these data reflect the importance of identifying the subcellular location of signaling complexes, using tools like dominant interfering mutants.
ET-1 mediates both vasoconstrictor and mitogenic actions in a number of cell types. In mesangial cells, the mitogenic actions of ET-1 can be attributed to primarily a Ras-dependent activation of the Raf-MEK-ERK signaling module (37). Our previous work has showed that the DSP, MKP-1, is induced by serum and ET-1 in mesangial cells (21,30). Furthermore, the induction of MKP-1 appears to be primarily dependent on activation of ERK1/2 (38). Interestingly, MKP-1 may be an in vivo substrate for ERK1/2 protecting the phosphatase from degradation via the proteasome pathway (36,39). This has led our laboratory and others to postulate that MKP-1 may act as a negative feedback enzyme for the regulation of MAPKs (30,40).
In addition to a biphasic activation of Ras by ET-1, we and others (41,42) have demonstrated a biphasic time course of ERK1/2 phosphorylation. In the present study, we observed that ET-1 treatment resulted in the biphasic phosphorylation of ERK1/2 in mesangial cells, which peaks at 5 min, returns to basal conditions by 30 min, and subsequently returns to a phosphorylated state within 4 h of ET-1 treatment. Interestingly, expression of MKP-1/CS extended the kinetics of ERK1/2 phosphorylation with the resultant loss of phosphorylation at 4 h. Previously, we suggested that because the kinetics of MKP-1 induction appeared to follow the time course of inactivation of ERK1/2, MKP-1 may act as a negative regulator of ERK1/2 in RMC (30). In the present study, investigations using a catalytically inactive form of MKP-1 (MKP-1/CS) demonstrated that anti-MKP-1 antibodies co-precipitated a phosphoprotein from RMC infected with an adenovirus encoding MKP-1/CS. Further, in vitro and in vivo co-precipitation and colocalization experiments indicated that phospho-p38 MAPK may be a prime candidate for interacting with MKP-1/CS. Additional studies including co-immunoprecipitation and mass spectrometry peptide mapping will be needed to support this hypothesis.
The p38 MAPK pathway is involved in the regulation of COX-2 expression in response to cytokines (43) and may be important for controlling expression of COX-2 in response to low salt conditions (44). Indeed, we observed that ET-1 induced a brief pulse of p38 MAPK phosphorylation in mesangial cells, and the addition of the p38 MAPK inhibitor, SB 203580, attenuated ET-1-stimulated COX-2 expression. Combined treatment of RMC with the MEK1/2 inhibitor U0126, which reduces ERK1/2 phosphorylation, and SB 203580 resulted in a marked reduction of ET-1-stimulated expression of COX-2. Moreover, in those cells infected with MKP-1/CS, there was a dramatic enhancement of basal and ET-1-stimulated p38 MAPK phosphorylation that was accompanied by enhanced basal and ET-1-stimulated expression of COX-2. Taken together, these results suggest that both the p38 MAPK pathway and the ERK1/2 pathway contribute to the expression of COX-2 in RMC. In concordance with our results, Guan et al. (43) have demonstrated a p38 MAPK dependence on COX-2 expression in RMC stimulated with interleukin-1␤. Additionally, MAP-KAP-2 stimulated by p38 MAPK has been implicated as a mechanism for stabilizing COX-2 mRNA in HeLa cells (7). Thus, an emerging idea imparts a general p38 MAPK and ERK dependence on the expression of COX-2. However, we cannot rule out the contributions of other pathways including NF-B (45) and NFAT (46) toward COX-2 expression or the possible dependence of these pathways on the upstream activation of p38 MAPK or ERK1/2 activation.
Our data suggest that adenoviral delivery of MKP-1/CS enhances phosphorylation of p38 MAPK whereas wild type MKP-1 results in diminished phosphorylation of p38␣ MAPK. Several reports indicate that wild type MKP-1 can inactivate ERK1/2, p38 MAPK, and SAPK/JNK (28,29,37), inferring that MKP-1 may be a general inhibitor of MAPK family members with specificity conferred by the stimulus. In contrast, the substrate trapping mutant of MKP-1 clearly augmented the phosphorylation of p38 MAPK more than ERK1/2 in our study, suggesting that MKP-1 displays a relative selectivity for p38␣ MAPK compared with ERK1/2 in glomerular mesangial cells. Evidence supporting this conclusion are as follows. First, wild FIG. 11. Model for regulation of COX-2 by MKP-1 and p38 MAPK in RMC. A, ET-1 stimulates the phosphorylation of ERK and p38 MAPK (1). ET-1 induces the expression of COX-2 in quiescent cells in an ERK1/2 and p38 MAPKdependent manner (2). MKP-1 directly interacts with p38␣ MAPK as a probable mechanism for inactivating the dually phosphorylated MAPK (3). Adenoviral delivered MKP-1/CS traps the phosphorylated p38␣ MAPK in the cytoplasm, preventing any potential nuclear shuttling of the enzyme. B, delivery of adenoviral MKP-1 (ADVMKP-1) results in decreased COX-2 expression (1), attenuation of ET-1 stimulated p38␣ MAPK phosphorylation (2), and activation of p38␥ MAPK (3). type MKP-1 was co-precipitated with p38-FLAG in our in vitro transcription/translation experiments. Our data suggest that not only is MKP-1 specific for inactivating p38␣ MAPK but that it also results in activation of p38␥ MAPK by an as yet unidentified mechanism. This result may suggest a complex intra-MAPK regulatory pathway that has yet to be fully examined. This finding of p38␣ MAPK and MKP-1 interaction is supported by two separate laboratories utilizing recombinant forms of MKP-1 and p38 MAPK. Specifically, p38␣ MAPK, but not ␥ or ␦ isoforms, were shown to directly interact with the wild type form of MKP-1, increasing its catalytic activity (25). Additionally, Tanoue et al. (47) have also described a similar interaction of MKP-1 with JNK/SAPK and p38 MAPK and proposed a classification scheme for the MKPs based on a consensus arginine-rich MAPK docking site (RRRAK/R). Second, MKP-1 and MKP-1/GFP were co-precipitated with p38-FLAG MAPK to a greater extent than with ERK from transiently transfected HEK 293 cells, a finding supported by other studies (48). These combined observations support a role for MKP-1 as a specific negative regulator of p38␣ MAPK but not p38␥ MAPK. However, we cannot rule out completely the contribution of other DSP as a detailed study of their expression has not yet been performed. Additional studies using knockout technology like short interfering RNA will be needed to provide the critical link among COX-2 expression, p38 MAPK activity, and MKP-1 expression.
Based on our past (30) and present work and the work of others (12,49) we propose a model for the role of MKP-1 in regulating the expression of COX-2 in mesangial cells (Fig. 11). Mitogenic agents like ET-1 rapidly induce the phosphorylation of both ERK1/2 and p38 MAPK. An additional effect is the induction of MKP-1 in a time frame that overlaps with their return to basal levels of phosphorylation. That we observe an increase in both basal and ET-1-stimulated phosphorylation of p38 MAPK concurrent with an increased basal and ET-1-stimulated expression of COX-2 in mesangial cells transduced with MKP-1/CS (Fig. 11A) suggests that a major role for wild type MKP-1 may be to limit the expression of genes that rely on the p38 MAPK pathway. Indeed, its possible that the expression of MKP-1 relates to the subcellular localization of p38 MAPK, although future studies will be necessary to validate this hypothesis. Additionally, our model implicates a role for MKP-1 in regulating the expression of COX-2 (Fig. 11B). Notably, the up-regulation of MKP-1 by the use of adenovirus in our studies reduced the basal and ET-1-stimulated phosphorylation of p38␣ MAPK. This implicates p38␣ MAPK as an important regulator of COX-2 expression. Conversely, under these same conditions, we found the phosphorylation state of p38␥ MAPK was increased (Fig. 11B). Whether this increase is a result of some compensatory mechanism or suggests a complex intra-MAPK signaling network remains unclear.
The coordinated expression of genes is of obvious importance during development, and dysregulation of gene expression may be involved in disease progression following inflammation and tumor formation. Mitogen-activated protein kinases are uniquely positioned to play an integral part in the gene regulatory process. Normally contained within the cytosolic compartment complexed with regulatory molecules, extracellular stimuli induce an activation accompanied by nuclear translocation of these enzymes (37). Phosphorylation of transcription factors by MAPKs results in transcriptional activation of immediate early genes like c-jun, c-fos, and even the DSP, mkp-1 (21,50). Thus, as is evidenced by the work presented here, DSP may play a key role in regulating the expression of p38 MAPKdependent genes such as COX-2. Importantly, this work suggests that mutations resulting in expression of endogenous catalytically inactive dual-specificity phosphatases (STYX) (51) could significantly contribute to altered MAPK signaling and ultimately affect gene expression.