Extracellular Signal-Regulated Kinase 7 (ERK7), a Novel ERK with a C-Terminal Domain That Regulates Its Activity, Its Cellular Localization, and Cell Growth

Materials.

Phorbol 12-myristate 13-acetate (PMA), myelin basic protein (MBP), peroxidase-conjugated goat anti-rabbit immunoglobulin G (IgG), and peroxidase-conjugated goat anti-mouse IgG were purchased from Sigma Chemicals (St. Louis, Mo.). Okadaic acid was purchased from LC Laboratories (Woburn, Mass.). PHAS-1 was purchased from Stratagene (La Jolla, Calif.). GST-ATF2 was purchased from New England BioLabs (Beverly, Mass.). Anti-MAP kinase antiserum (Ab283) was developed as previously described (29). Anti-ERK7 rabbit antiserum (Cocalico Biologicals, Reamstown, Pa.) was raised against a peptide representing the last 14 amino acids of the C-terminal domain of ERK7 (Research Genetics, Huntsville, Ala.). Monoclonal antibody (12CA5) against the hemagglutinin (HA) epitope was purchased from BabCo (Emeryville, Calif.). Peroxidase-conjugated 12CA5 (12CA5-PC), high-affinity rat anti-HA monoclonal antibody (3F10), and peroxidase-conjugated, affinity purified sheep anti-rat Fab Ig were purchased from Boehringer Mannheim (Indianapolis, Ind.). Affinity-purified peroxidase-conjugated goat anti-rabbit IgG, and anti-mouse IgG were purchased from Transduction Laboratories (Lexington, Ky.). Anti-active MAP kinase polyclonal antibody was purchased from Promega (Madison, Wis.). Enhanced chemiluminenscence reagents, [³⁵S]cysteine (>600 Ci/mmol), [α-³²P]dCTP (3,000 Ci/mmol), and [γ-³²P]ATP (6,000 Ci/mmol) were purchased from DuPont/NEN Research Products (Boston, Mass.). Degenerate PCR primers were obtained from Operon Technologies (Alameda, Calif.). Sequencing primers were synthesized by Integrated DNA Technologies (Coralville, Iowa) and Gibco/BRL (Grand Island, N.Y.). The pRSV–β-gal and pCMV–β-gal expression vectors were a gift from V. Sukhatme. The pSG5 MKP-1 plasmid was a gift from Lester Lau, and the control plasmid, pSG5, was purchased from Stratagene. Plasmid DNAs were prepared by CsCl-ethidium bromide gradient centrifugation as previously described (29) or by purification through columns as specified by the manufacturer (Qiagen, Chatsworth, Calif.). The unique site elimination mutagenesis kit was purchased from Pharmacia (Piscataway, N.J.).

Degenerate-primed PCR amplification.

The template for amplification was a 6-day-old rat brain cDNA library in λpCEV29 (19). Degenerate upstream and downstream oligonucleotide primers were synthesized based on regions of highly conserved amino acid sequences within MAP kinase subdomains II (VAIKKIS) [5′-CTG GAA TTC GTN GCN AT(T/A/C) AA(A/G) AA(A/G) AT-3′] and VII (TRWYRAP) [5′-GC TCT AGA TGG (A/G)GC NC(T/G) (A/G)TA CCA NC(T/G) NGT-3′], respectively. An EcoRI site (underlined) was attached to the 5′ end of the sense oligonucleotide and an XbaI site (underlined) was attached to 5′ end of the antisense oligonucleotide to allow subcloning of the PCR products into an Escherichia coli phagemid vector, pGEM-7Z(+) (Promega). The PCR buffer consisted of 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 2 mM MgCl₂, 0.2 mM deoxynucleoside triphosphates, 2 μM each primer, and 1 U of AmpliTaq DNA polymerase (Perkin-Elmer, Norwalk, Conn.) in a final volume of 50 μl. PCR was performed in a thermocycler (Perkin-Elmer) for 5 min at 97°C followed by 35 cycles of 30 s at 94°C, 60 s at 50°C, and 60 s at 72°C, and a final extension at 72°C for 5 min. The band of interest was approximately 450 bp. The PCR product was purified with the Wizard PCR prep DNA purification system as specified by the manufacturer (Promega) and subcloned into pGEM-7Z(+) with the EcoRI and XbaI restriction sites.

Screening of PCR products.

Southern blot screening was performed with an oligonucleotide probe [5′-GAG CA(C/T) CA(A/G) ACC TAC TGT CAG-3′] derived from the consensus amino acid sequence within subdomains II to VIII for rat ERK1 and ERK2, EHQTYCQ. DNA from colonies that did not hybridize to the ERK1 and ERK2 probe was sequenced (Sequenase version 2.0 kit; Amersham Life Science). The sequences were compared to the GenBank/EMBL Data Bank by using the Blastn program (4). Of the 20 colonies initially isolated, 3 were ERK1, 8 were ERK2, 1 was ERK3, 7 were BMK1/ERK5, and 1 was only moderately homologous to known ERKs.

Tissue Northern analysis.

cDNA or the PCR product was radiolabeled with ³²P by using a nick translation labeling method and used to probe a poly(A)⁺ RNA rat multiple-tissue Northern blot (Clontech, Palo, Alto, Calif.). The blot was prehybridized for 6 h in 5× SSPE (0.75 M NaCl, 50 mM NaH₂PO₄, 5 mM EDTA [pH 7.4]) with 5× Denhardt’s solution (0.5% Ficoll, 0.5% polyvinylpyrrolidone, 0.5% bovine serum albumin), 50% formamide, 0.1% sodium dodecyl sulfate [SDS], and 100 μg of sheared salmon sperm DNA per ml, hybridized overnight in the prehybridization solution at 42°C with radiolabeled probe, washed extensively with 0.1× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate)–0.1% SDS at 65°C, and exposed to X-ray film at −80°C.

Screening of cDNA library and isolation of full-length ERK7.

The ³²P-labeled ERK7 reverse transcription-PCR product was used to screen a rat testis library in λgt11 (10⁶ recombinants: Clontech 5’ Stretch). Five positive plaques were isolated after secondary screening. cDNAs were isolated and subcloned into pBluescript SK(−) vector (Stratagene), at the EcoRI site. The isolated cDNAs were evaluated by Southern blotting to verify the identity of ERK7. Nucleotide sequences of these clones were determined, and a full-length (2.0-kb) cDNA clone was identified.

Plasmid construction and preparation.

The HA-tagged mouse ERK2 (29) was cloned into pcDNA3 (Invitrogen, Carlsbad, Calif.) at the BamHI and EcoRV sites. A hemagglutinin antigen tag, YPYDVPDY, was inserted immediately following the start methionine of ERK7 by using PCR and was verified by DNA sequencing. Unique site elimination mutagenesis (14) was performed to generate ERK7 mutants (K43R T175A Y177F and T175A Y177F). The ERK2/ERK7 chimera was generated by inserting the C-terminal region of ERK7 in frame with ERK2 at a XhoI site created by unique site elimination mutagenesis and at the EcoRV site in pcDNA3. In the ERK2/ERK7 chimera, the three amino acids, GYR, at the C terminus of ERK2 were replaced with REE, from ERK7. The truncated ERK7ΔT mutant was generated by a restriction site cleavage at a XhoI site, resulting in a 356-amino-acid protein where the last three residues were replaced with the sequence alanine-cysteine-isoleucine. The mutated sequences were verified by sequencing.

Cell culture.

COS, CV-1, NIH 3T3, and PC12 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum and antibiotics (50 U of penicillin per ml and 50 μg of streptomycin per ml) at 37°C in a 95% air–5% CO₂ atmosphere. Quiescent cells were obtained by incubating cells in serum-free medium for at least 24 h. The immortalized H19-7 cells were generated from embryonic rat hippocampal cells as previously described (18).

Transient transfections.

Unless otherwise noted, approximately 8 × 10⁵ cells were seeded on 100-mm plates and incubated overnight. The medium was changed to serum-free OptiMEM (Gibco/BRL), and the cells were transfected with a total of 10 μg of plasmid DNA and 40 μl of TransIt LT-1 as specified by the manufacturer (PanVera Corp., Madison, Wis.). Of the total plasmid DNA, 10% consisted of either the pRSV–β-gal or pCMV–β-gal expression vector. β-Galactosidase activity was used to normalize for transfection efficiency between groups. Unless otherwise noted, cells were made quiescent in serum-free medium for 24 h prior to harvesting.

β-Galactosidase assay.

Cell lysate (10 μl) was diluted to a final volume of 300 μl in 0.1 M sodium phosphate buffer (pH 7.5)–0.05 M β-mercaptoethanol–1 mM MgCl₂–0.9 mg of o-nitrophenyl-β-d-galactoside per ml. The samples were incubated at 37°C until a faint yellow color developed, at which time 500 μl of 1 M Na₂CO₃ was added and the optical density at 410 nm was measured. One unit of β-galactosidase enzyme was used as a positive control.

Preparation of cell and tissue extracts.

Culture cells were washed twice with ice-cold phosphate-buffered saline and lysed with 1% Triton-based lysis buffer (TLB) (1% Triton X-100, 50 mM Tris-HCl [pH 7.5], 40 mM β-glycerophosphate, 100 mM NaCl, 50 mM NaF, 2 mM EDTA, 1 mM sodium vanadate, 1 mM phenylmethylsulfonyl fluoride, 1 μg of aprotinin per ml, 1 μg of leupeptin per ml, 20 mM p-nitrophenyl phosphate). SDS was added to a final concentration of 0.1% where noted. Pulverized, frozen rat tissue (−80°C) was lysed in TLB containing 0.1% SDS. The cell debris was removed by centrifugation (16,000 × g for 10 min at 4°C). The protein concentration in the cleared supernatant was estimated as previously described (3).

Western analysis.

Cell extracts (10 to 60 μg of protein per lane, unless otherwise noted) were resolved on an 8% acrylamide separating gel by SDS-polyacrylamide gel electrophoresis (PAGE). Proteins were transferred to a nitrocellulose membrane. Membrane blocking, washing, antibody incubation, and detection by enhanced chemiluminescence were performed as previously described (3).

Immunoprecipitation and immune complex kinase assay.

Immune complex kinase assays were performed with ectopically expressed epitope-tagged protein kinases from COS and CV-1 cells as previously described (29), except that 2 μg of each substrate was used per reaction in kinase buffer (20 mM HEPES [pH 7.4], 10 mM MgCl₂, 1 mM dithiothreitol, 0.2 mM sodium vanadate, 10 mM d-nitrophenylphosphate, 5 μCi of [γ-³²P]ATP). Proteins were separated by SDS-PAGE on a 10% acrylamide separating gel, transferred to a nitrocellulose, and subjected to autoradiography. Quantitation of substrate phosphorylation was determined by digital optical scanning on an image analyzer (AMBIS, San Diego, Calif.). The presence of epitope-tagged proteins in the immunoprecipitates was verified by Western analysis with 12CA5, 12CA5-PC, or 3F10 monoclonal antibody. Immunoprecipitation assays were performed in a similar fashion; however, the immune complexes were eluted prior to the kinase reaction and subjected to Western analysis.

Immunocytochemistry.

CV-1 cells, about 60% confluent on coverslips, were transfected with either ERK7, K43R, or ERK7 ΔT and pCMV–β-gal with TransIt LT1 reagent. Following the transfection and overnight incubation in serum-containing medium, the medium was switched to serum-free DMEM for 24 h. The cells were then fixed in 10% formalin for 15 min, washed, blocked, and incubated simultaneously with rabbit polyclonal anti-β-galactosidase antibody (5 Prime-3 Prime, Boulder, Colo.) at a 1:300 dilution and anti-HA 12CA5 monoclonal antibody (1:500). After a 1-h incubation at room temperature, the cells were washed and then incubated with Texas Red-conjugated anti-mouse and fluorescein isothiocyanate-conjugated anti-rabbit antibodies (Molecular Probes, Inc., Eugene, Oreg.). The stained cells were analyzed under a Zeiss Axioplan fluorescence microscope, and images were captured with a Photometrics PLX cooled charge-coupled device digital camera controlled by OPENLAB software (Improvision Ltd.).

GST fusion proteins.

Plasmids encoding glutathione S-transferase (GST)–c-Fos (210 to 313), GST–c-Jun (1 to 93), GST–Elk-1 (307 to 428), and GST–c-Myc were generously provided by T. Deng, E. Wattenberg, R. Treisman, and N. Hay, respectively. GST fusion proteins were prepared by using a GST purification system (Pharmacia).

BrdU incorporation in CV-1 cells.

CV-1 cells plated in 12-well dishes at approximately 6 × 10⁴ cells per well were transfected with either pcDNA3, ERK7, K43R, or ERK7 ΔT (1 μg/well) and pCMV–β-gal (0.5 μg/well) by using TransIt LT1 solution. Following the transfection and overnight incubation in serum-containing medium, the medium was switched to serum-free DMEM for 24 h. Subsequently, the cells were stimulated for 24 h with 10% fetal bovine serum (FBS) in DMEM containing BrdU (10 μM). Transfected cells were detected by an in situ β-galactosidase assay with 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) as a substrate, as previously described (29). After the β-galactosidase assay, the cells were treated with 2 N HCl for 15 min, neutralized with 0.1 M borate buffer (pH 9.0), and washed with phosphate-buffered saline. BrdU incorporation was determined by immunocytochemistry with an anti-BrdU monoclonal antibody (Oncogene Research Products, Cambridge, Mass.). A colometric detection method, consisting of the Vectastain ABC kit with Vector VIP, a peroxidase substrate kit (Vector Laboratories, Burlingame, Calif.), was used. The stained cells were analyzed under a Leitz DMIRB microscope. The β-galactosidase-positive cells with or without BrdU incorporation were scored separately, and the percentage of BrdU-positive cells was calculated.

GenBank accession number.

The GenBank accession number for the ERK7 cDNA reported here is AF078798.

Molecular cloning and tissue distribution of ERK7.

Regions of highly conserved amino acid sequences within MAP kinase subdomains II (VAIKKIS) and VIII (TRWYRAP) were used to design degenerate upstream and downstream oligonucleotide primers. By using a 6-day old rat brain cDNA library in λpCEV29 as a template, a 455 bp clone representing a novel sequence was generated by PCR. The gene encoding this sequence was eventually termed ERK7. A rat testis library was screened with the PCR product to obtain the full-length clone.

Nucleotide sequence analysis of the ERK7 clone reveals a translation start site in good agreement with the Kozak consensus sequence (28) (Fig. 1A). The open reading frame extends 1,638 bp, encoding a protein of 546 amino acids with an estimated molecular mass of 61 kDa (Fig. 1A). The deduced amino acid sequence contains a putative ATP binding site, a highly conserved lysine in subdomain II, and a threonine-glutamine-tyrosine activation sequence within subdomain VIII, indicating that the cDNA encodes for a novel member of the ERK/MAP kinase family.

Alignment of the deduced amino acid sequence of the clone with other MAP kinases containing the TEY activation motif reveals a rather unique C-terminal extension of 195 amino acids, which is considerably shorter than that of ERK5/BMK1 (Fig. 1B). We therefore named this cDNA ERK7. This C-terminal region has two PXXP proline-rich segments thought to be ligands for binding to SH3 domains (43). In addition, the C-terminal region contains a putative nuclear localization signal (PARKRGP) (38). Apart from the C-terminal region, ERK7 is approximately 40% homologous to rat ERK1 and ERK2. Overall, ERK7 has low homology to ERK5/BMK1, suggesting that it is a distinct member of the ERK/MAP kinase family.

The tissue distribution of ERK7 transcripts was determined by Northern analysis of poly(A)⁺ RNA from multiple rat tissues (Fig. 2A). A 2.0-kb ERK7 mRNA was most highly expressed in testis followed by lung tissue (Fig. 2A). Both the testis and lung contain additional bands on Northern analysis which might represent splicing intermediates or mRNA encoding for homologous proteins. Although ERK7 was originally identified by using a 6-day-old rat brain library, the presence of ERK7 in adult rat brain could be observed only following extended exposure (data not shown). To further characterize the distribution of ERK7, antiserum was raised against a peptide corresponding to the last 14 amino acids of the C-terminal domain of ERK7. Western analysis of multiple rat tissues and cell types with this antiserum indicated that ERK7 is widely distributed, although at low levels (Fig. 2B). The relative abundance of ERK7 in testes compared to other tissues evaluated by Western analysis is consistent with the results of Northern analysis. These findings suggest that ERK7 is differentially expressed between tissue types.

Ectopic ERK7 is constitutively TEY phosphorylated in COS cells.

To study ERK7 in intact cells, ERK7 was epitope tagged with the 9-amino-acid immunodominant portion of the influenza virus HA by PCR and subcloned into the mammalian expression vector, pcDNA3. In addition, four mutant forms of epitope-tagged ERK7 were generated. Arginine was used instead of highly conserved lysine-43 in subdomain II to generate the kinase-inactive mutant, K43R. The TEY activation domain was mutated either by changing the threonine residue to alanine, T175A, changing the tyrosine residue to phenylalanine, Y177F, or combining both mutations, T175A Y177F. COS cells were transiently transfected with ERK7 and the mutant forms of ERK7, and a 61-kDa protein was recognized by Western analysis with the anti-HA antibody (Fig. 3A, left). The wild-type ERK7 is seen as a double band. To determine whether the band with reduced electrophoretic mobility represented the TEY phosphorylated form of ERK7, Western analysis was performed with the anti-active MAP kinase polyclonal antibody (Fig. 3A, right). This antibody specifically recognizes the dually but not the singly phosphorylated TEY motif of ERK1 and ERK2, which corresponds to the active form of ERK (26). It also cross-reacts with ERK5/BMK1, the only other known MAP kinase with a TEY activation motif (48). Only the wild-type ERK7 had a band at about 61 kDa, recognized by the anti-active MAP kinase antibody. This band corresponded to the band with reduced electrophoretic mobility seen in the anti-HA Western blot. Since this upper band represents about half of the total ectopically expressed ERK7, Western analysis suggests that as much as 50% of ERK7 may be TEY phosphorylated and active in serum-starved transfected cells. Since the K43R mutant is not recognized by the anti-active MAP kinase polyclonal antibody (Fig. 3A, right), it is possible that kinase activity, perhaps through autophosphorylation, is required for TEY phosphorylation and activation of ERK7.

ERK7 is constitutively active.

To further explore the possibility that ERK7 has constitutive kinase activity, potential substrates were evaluated by an in vitro kinase assay. Immunoprecipitated wild-type HA-ERK7, isolated from serum-starved transfected COS cells, phosphorylated GST-c-Fos, GST-c-Myc, and MBP in in vitro kinase assays (Fig. 3B). Immunoprecipitated K43R mutant failed to phosphorylate these substrates in vitro, indicating that this activity was specific for ERK7. Unlike other members of the MAP kinase family, ERK7 did not significantly phosphorylate GST-Elk1 (39), PHAS1 (35), GST–c-Jun (24), and GST-ATF2 (45). The latter results also suggest that the phosphorylation was on c-Fos, and c-Myc but not on GST. Similar in vitro kinase assays were performed with the TEY mutants (Fig. 3C). None of these mutants had significant kinase activity. These results confirm that ERK7 is a functional kinase with discrete activity even in serum-starved cells.

To test whether the simian virus 40 transformation of COS cells contributes to the constitutive activity of ERK7, similar experiments were performed with CV-1 cells. As seen in COS cells, the wild-type ERK7 and the kinase-inactive mutant were recognized by the anti-HA antibody as a double and a single band, respectively (Fig. 4A, top). Similarly, only the electrophoretically retarded band of the wild-type ERK7 was recognized by the anti-active MAP kinase antibody, indicative of TEY phosphorylation (Fig. 4A, bottom). An in vitro kinase assay with GST–c-Fos as a substrate confirmed that immunoprecipitated ERK7 is active in serum-starved CV-1 cells (Fig. 4B). For comparison, CV-1 cells were transiently transfected with HA-ERK2. In serum-starved cells, ERK2 appeared as a single band (Fig. 4A, top). After serum stimulation, a second band with reduced electrophoretic mobility appeared (Fig. 4A, top). Western analysis with the anti-active MAP kinase antibody demonstrated that the vast majority of the ectopically expressed ERK2 in serum-starved CV-1 cells was not TEY phosphorylated and therefore was not active as a kinase (Fig. 4A, bottom). The expression of constitutively activated ERK7 probably is not limited to monkey cells. Western analysis with the anti-HA antibody of NIH 3T3 cells transfected with HA-ERK7, like COS and CV-1 cells, also revealed a second band with reduced electrophoretic mobility corresponding to the active form (data not shown). Together, these data suggest that ERK7 is constitutively active even in serum-starved cells.

Consistent with the CV-1 results, the kinase activity of ERK7 is significantly greater than that of ERK2 in serum-starved COS cells. To assess the relative kinase activities of the two enzymes, COS cells were transfected with expression vectors for either ERK7 or ERK2. The ERKs were immunoprecipitated with anti-HA antibody and then either assayed directly for kinase activity under equivalent conditions with GST–c-Fos as a substrate (13) or quantitated by immunoblotting with anti-HA antibody. Surprisingly, the level of cytomegalovirus promoter-regulated ERK7 expression in COS cells was significantly lower than that of ERK2 (Fig. 5A, bottom). Differences in transfection efficiency as assessed by β-galactosidase activity (data not shown) could not explain this disparity. Over 140 times more ERK2 than ERK7 was isolated from serum-starved cells, but both enzymes had the same kinase activity as determined by scanning densitometry (Fig. 5A, top). These results demonstrate that ERK7 has significant constitutive kinase activity relative to ERK2.

Regulation of ERK7 differs from that of ERK2.

To determine if activators of ERK2 might also activate ERK7, COS cells were transiently transfected with HA-ERK7 and HA-ERK2. Cells within each transfection group were pooled and split to equalize transfection efficiency. After serum starvation for 24 h, the cells were stimulated with several known activators of ERK2. Cellular protein immunoprecipitated with the anti-HA antibody was subjected to Western analysis with either the anti-HA or the anti-active MAP kinase antibody.

As expected, serum, epidermal growth factor (EGF), PMA, and okadaic acid dramatically increase the amount of TEY-phosphorylated HA-ERK2 in comparison to quiescent cells (Fig. 5B, bottom). However, none of these agents significantly increases the amount of TEY-phosphorylated ERK7 (Fig. 5C, bottom). The lack of additional activation of ERK7 was confirmed in an immune-complex kinase assay with GST–c-Fos as a substrate (Fig. 5C, top). Consistent with the anti-active MAP kinase Western analysis, no increase in GST–c-Fos phosphorylation was seen following stimulation with serum, EGF, PMA, or okadaic acid. In addition, activation of ERK7 was not seen following stimulation with anisomycin or H₂O₂ (data not shown). In quiescent COS cells, only a very small fraction of HA-ERK2 is TEY phosphorylated (Fig. 5B, control). In contrast, a significant portion of HA-ERK7 appears to be TEY phosphorylated (Fig. 5C, control). Since TEY phosphorylation of the MAP kinase family is necessary and sufficient for activation (5), these results indicate that regulation of ERK7 differs from that of ERK2 and are consistent with constitutive activation of ERK7.

The C-terminal domain is required for ERK7 activity.

To determine whether the unique C-terminal region of ERK7 plays a role in regulating ERK7 activity, a form of HA-ERK7 truncated at Pro-353 of the C terminus, termed ERK7ΔT, was generated. When ERK2 and ERK5 are aligned by their homologous kinase domains, the carboxy terminus of ERK7ΔT is comparable in length to the carboxy terminus of ERK2. The anti-HA antibody recognizes both the full-length and truncated ERK7 in whole-cell lysates from transfected COS cells (Fig. 6A, top), whereas under the same conditions, the anti-active MAP kinase antibody recognizes only the full-length ERK7 (Fig. 6A, bottom). However, concentration by immunoprecipitation with the anti-HA antibody allows the truncated ERK7 to be detected by the anti-active MAP kinase antibody (Fig. 6B, right). Compared to the full-length ERK7, only a small fraction of the truncated form appears to be TEY phosphorylated. Detection by the anti-active MAP kinase antibody indicates that the truncated form of ERK7 is recognized and phosphorylated by its upstream activator, although perhaps with significantly reduced efficiency. An in vitro kinase assay with GST–c-Fos as a substrate confirmed that the kinase activity of the truncated ERK7 mutant in serum-starved COS cells is significantly reduced compared to the full-length ERK7 (Fig. 6C). In addition, ERK7ΔT, like the full-length ERK7, is not activated by serum stimulation (data not shown). To determine if the C-terminal domain is sufficient for activation of ERK2, a chimeric protein consisting of ERK2 with the C-terminal domain of ERK7 was constructed. Following serum stimulation, the chimeric protein was able to phosphorylate both PHAS1 and MBP in vitro like wild-type ERK2 (Fig. 6D), but the presence of the ERK7 C-terminal region was not sufficient to activate ERK2 in serum-starved COS cells (Fig. 6E). Together, these data indicate that the C-terminal tail is required for the full constitutive activity of ERK7 in serum-starved cells but is insufficient to activate ERK2. Reactivity of the truncated ERK7 with the anti-active MAP kinase antibody is substantially reduced in comparison to that of full-length ERK7, requiring concentration by immunoprecipitation to allow detection.

MKP-1 dephosphorylates ERK7.

To explore the possibility that the C-terminal region of ERK7 contributes to the constitutive activity of ERK7 by protecting the TEY motif from dephosphorylation, cotransfections with the dual-specificity protein phosphatase, MKP-1, were performed. MKP-1 has been shown to inactivate ERK2 as well as other members of the MAP kinase family through specific dephosphorylation of the TEY activation motif (51). COS cells were cotransfected either with HA-ERK7 or HA-ERK2 and either MKP-1 or a control plasmid. When they were assayed by an electrophoretic mobility shift assay, reduced activation of both immunoprecipitated HA-ERK7 and HA-ERK2 from cells cotransfected with MKP-1 was observed compared to controls following serum stimulation (Fig. 7, top). Western analysis with the anti-active MAP kinase antibody confirmed that cotransfection with MKP-1 reduced the amount of the TEY-phosphorylated forms of both HA-ERK2 and HA-ERK7 (Fig. 7, bottom). A modest reduction in the HA-ERK7 gel shift was noted following immunoprecipitation, suggesting that ERK7 was further dephosphorylated during the immunoprecipitation process (Fig. 7, top). The effect of MKP-1 on ERK7 in serum-starved cells was identical to that in cells treated with serum (data not shown). These experiments indicate that the C-terminal region of ERK7 does not significantly protect ERK7 from TEY dephosphorylation. Therefore, it appears unlikely that the C-terminal region regulates the constitutive activity of ERK7 by protecting the activation motif against dephosphorylation.

Cellular localization of ERK7.

Another possible regulatory mechanism is that the C-terminal domain facilitates the activation of ERK7 by directing it to its activator within the cell. To explore this possibility, immunocytochemistry was performed to test whether the C-terminal domain influences the subcellular localization of ERK7. CV-1 cells were transiently transfected with either ERK7, K43R, or ERK7ΔT and pCMV–β-gal. With the anti-HA antibody, cells transfected either with ERK7 or K43R had predominantly nuclear staining (Fig. 8A and D). On the other hand, ERK7ΔT was localized outside the nucleus (Fig. 8G). Specific staining was not detected in adjacent, nontransfected cells. Staining for β-galactosidase in the transfected cells revealed a cytoplasmic pattern (Fig. 8B, E, and H). Unlike ERK1 and ERK2, which translocate to the nucleus upon serum stimulation (20, 32), ERK7 and K43R localize to the nucleus even in serum-starved cells. These findings suggest that localization of ERK7 to the nucleus is not dependent on its kinase activation state. The staining pattern of the truncated ERK7 mutant, on the other hand, suggests the C-terminal region plays a role in nuclear localization, consistent with the presence of a putative nuclear localization signal within this domain. Together, these results suggest that localization to the nucleus alone is insufficient for activation of ERK7 through TEY phosphorylation, since the majority of the kinase-inactive mutant is not recognized by the anti-active MAP kinase antibody (Fig. 3A). However, it remains possible that directed cellular localization by the C-terminal region of ERK7 is required for activation of ERK7.

Inhibition of cell growth by ERK7 requires its C-terminal domain.

The classic MAP kinases ERK1 and ERK2 have been implicated in the regulation of cell growth by growth factors (33, 46, 53). To determine whether ERK7 can also influence cell growth, we measured the effect of ectopic expression of ERK7 on DNA synthesis as monitored by BrdU incorporation in CV-1 cells. CV-1 cells were cotransfected with either pcDNA3 alone or pcDNA3 containing ERK7, K34R, or ERK7ΔT cDNA along with a β-galactosidase expression vector to identify transfected cells. No effect of ERK7 on serum-starved cells was observed (Fig. 9). However, transient transfection of wild type ERK7 suppressed DNA synthesis in serum-treated CV-1 cells by approximately 50% compared to the control (P < 0.01) (Fig. 9). Surprisingly, this reduction appears to be independent of the kinase activity of ERK7, since the kinase-inactive mutant K43R has an identical effect (P < 0.002) (Fig. 9). Since β-galactosidase and ERK7 were coexpressed in only approximately 80% of the cells that were counted, it is likely that the actual inhibition of DNA synthesis was greater than 50%. In contrast to the kinase-inactive K43R mutant, the truncated ERK7Δ mutant that is also kinase impaired did not reduce BrdU incorporation, indicating that the C-terminal domain appears to be required. These data indicate that ERK7 can function as a negative regulator of growth. Surprisingly, it appears that the C-terminal domain, rather than kinase activity, is the primary component of ERK7 required for this function.