Cloning and characterization of MEK6, a novel member of the mitogen-activated protein kinase kinase cascade.

Mitogen-activated protein kinases are members of a conserved cascade of kinases involved in many signal transduction pathways. They stimulate phosphorylation of transcription factors in response to extracellular signals such as growth factors, cytokines, ultraviolet light, and stress-inducing agents. A novel mitogen-activated protein kinase kinase, MEK6, was cloned and characterized. The complete MEK6 cDNA was isolated by polymerase chain reaction. It encodes a 334-amino acid protein with 82% identity to MKK3. MEK6 is highly expressed in skeletal muscle like many other members of this family, but in contrast to MKK3 its expression in leukocytes is very low. MEK6 is a member of the p38 kinase cascade and efficiently phosphorylates p38 but not c-Jun N-terminal kinase (JNK) and extracellular signal-regulated kinase (ERK) family members in direct kinase assays. Coupled kinase assays demonstrated that MEK6 induces phosphorylation of ATF2 by p38 but does not phosphorylate ATF2 directly. MEK6 is strongly activated by UV, anisomycin, and osmotic shock but not by phorbol esters, nerve growth factor, and epidermal growth factor. This separates MEK6 from the ERK subgroup of protein kinases. MEK6 is only a poor substrate for MEKK, a mitogen-activated protein kinase kinase kinase that efficiently phosphorylates the related family member JNKK.

consists of several modules. Most of the mammalian MAPK modules have been identified by analogy with yeast protein kinase cascades. We are following the nomenclature of Seger and Krebs (10) and name the individual levels of kinases regardless of their pathway MAPK, MAPKK, and MAPKKK. The extracellular signal-regulated kinase (ERK) subgroup of MAPKs has been examined in detail (11)(12)(13)(14)(15)(16). The c-Jun Nterminal kinase (JNK) subgroup (also known as stress-activated protein kinase, or SAPK) was first described as a kinase cascade leading to the phosphorylation of Jun (17)(18)(19). More recently the p38 MAPK pathway has been identified in analogy to the yeast HOG1 pathway (20,21).
We were interested in identifying novel members of the MAPKK family that use p38 as substrate. So far only two of these kinases have been identified in humans: MKK3 and JNKK (MKK4, SEK). JNKK is a substrate for MEKK1 (22)(23)(24) and phosphorylates and activates JNK1, JNK2, and p38 in vivo and in vitro (23,25). In contrast, MKK3 phosphorylates and activates exclusively p38 (25); however, the upstream MAP-KKK for MKK3 is unknown, since MEKK1 is not an activator of p38 in vivo (22). In an attempt to find novel members of the p38 MAP kinase cascade we cloned and characterized a new human MAPKK, which we named MEK6. MEK6 differs from its closest homologue, MKK3, most significantly in the N-terminal and C-terminal amino acid regions. MEK6 efficiently phosphorylates p38 but not JNK or ERK. MEK6 is strongly activated by UV, anisomycin, and osmotic shock.

EXPERIMENTAL PROCEDURES
cDNA Cloning-The Expressed Sequence Tags (EST) subdivision of the National Center for Biotechnology Information (NCBI) GenBank TM databank was searched with the tblastn program and the human MKK3 amino acid sequence as query using the BLAST e-mail server. The 223-bp EST sequence F00521 displayed the highest similarity score. A reverse PCR primer (5Ј-CACATCTTCACTTGACCGAGAGCA-3Ј) directed against this sequence was designed with the help of the program Oligo, version 4.0 (National Biosciences, Inc.). Poly(A) ϩ RNA was prepared from unstimulated Jurkat T cells using the Micro-Fast Track Kit (Invitrogen). One g of this RNA was used to generate an adaptor-ligated cDNA library that can be used for 5Ј and 3Ј rapid amplification of cDNA ends (Marathon cDNA Amplification Kit, Clontech). The adaptor-specific primer from the kit and the gene specific reverse primer were used to PCR-amplify the 5Ј portion of MEK6. PCR amplification was performed with a combination of Taq and Pwo polymerases (Expand Long Template PCR system, Boehringer Mannheim) in the presence of TaqStart antibody (Clontech). This mixture is designed to produce high yield of long PCR fragments and to provide proofreading function. All PCR amplifications were carried out in 0.2-ml Perkin-Elmer thin-wall MicroAmp tubes and a Perkin-Elmer model 2400 or 9600 thermocycler. The resulting 0.8-kb PCR fragment was ligated into pGEM-T (Promega) and sequenced (dye terminator cycle sequencing) with an ABI 373 Automated Sequencer. The sequence information from the 5Ј end of the partial MEK6 cDNA was used to design a forward PCR primer (5Ј-TTGTGCTCCCCTCCCCCATCAAAG-GAA-3Ј) for 3Ј rapid amplification of cDNA ends. The gene-specific forward primer and the adaptor-specific primer were used to PCRamplify the complete MEK6 cDNA from an adaptor-ligated MOLT-4 cDNA library. This library was generated using 1 g of MOLT-4 * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EMBL Data Bank with accession number(s) U49732.
‡ To whom correspondence should be addressed: Signal Pharmaceuticals Inc., 5555 Oberlin Dr., San Diego, CA 92121. Tel.: 619-558-7500; Fax: 619-558-7513; E-mail: bstein@signalpharm.com. 1 The abbreviations used are: MAPK, mitogen-activated protein kinase; ERK, extracellular signal-regulated kinase; MAPKK, MAP kinase kinase; MAPKKK, MAP kinase kinase kinase; JNK, c-Jun N-terminal kinase; JNKK, JNK kinase; MEKK, MEK kinase; EST, Expressed Sequence Tags; NCBI, National Center for Biotechnology Information; HA, hemagglutinin; PCR, polymerase chain reaction; GST, glutathione S-transferase; kb, kilobase pair(s); bp base pair(s); NGF, nerve growth factor; EGF, epidermal growth factor. poly(A) ϩ RNA (Clontech) and the Marathon cDNA Amplification Kit (Clontech). The 1.6-kb PCR fragment was ligated into pGEM-T (Promega), and three clones were sequenced several times on both strands with an ABI 373 automated sequencer. The BLAST program was used to search the NCBI GenBank TM data base for related cDNAs. The Bestfit program from the Wisconsin Genetics Computer Group (Madison, WI) was used for calculating the amino acid identities between MKK3 and MEK6. The MacVector program (Kodak-IBI) was used for aligning the amino acids of MKK3 and MEK6.
Human p38 (GenBank TM accession number U10871) was cloned by PCR amplification of a Jurkat cDNA library with primers against the 5Ј end (5Ј-CCAACCATGGCTCAGGAGAG-3Ј) and 3Ј end (5Ј-CGGTACCT-TCAGGACTCCATCT-3Ј) of the published human p38 sequence. Each strand of the PCR fragment was sequenced several times with an ABI 373 automated sequencer.
Plasmids-3ϫHA-MEK6-SR␣3 was constructed by replacing serine in position 2 of MEK6 with alanine, adding sequence encoding three copies of a 10-amino acid hemagglutinin (HA) epitope to the N terminus of MEK6 and ligating the resulting cDNA into SR␣3. 3ϫHA-JNKK-SR␣3 was constructed by adding sequence encoding three copies of the HA epitope to the N terminus of mouse ⌬JNKK initiated at amino acid 35 (22) and ligating the resulting cDNA into SR␣3. Both plasmids expressed proteins of the expected size as verified with the TnT SP6 Coupled Reticulocyte Lysate System (Promega). GST-MEK6 was constructed by ligating a 1.3-kb DNA fragment encoding amino acid 1 through the stop codon of MEK6 with a serine to alanine substitution of amino acid 2 into pGEX-KG (26). Similarly, GST-p38 and GST-JNK2 were constructed by ligating the respective cDNA fragments encoding amino acid 1 through the stop codon into pGEX-KG. The following plasmids have been described previously: CMV5-MEKK (23), His-ERK1(K52R) (27), GST-c-Jun(1-79) (17), GST-ATF2 (28).
Northern Blot Analysis-Northern blots were performed using 2 g of poly(A) ϩ RNA isolated from 16 different adult human tissues, fractionated by denaturing formaldehyde 1.2% agarose gel electrophoresis, and transferred onto a charge-modified nylon membrane (Clontech). The blots were hybridized to a MKK3 probe (700-bp MKK3 cDNA fragment) or MEK6 probe (870-bp MEK6 cDNA fragment) using Ex-pressHyb (Clontech) according to the manufacturer's instructions. Both probes were prepared by random prime labeling (Prime It II, Stratagene) of the cDNA with [␣-32 P]dCTP (DuPont NEN). For control purposes the blots were also hybridized to a radiolabeled ␤-actin probe.
Transient Transfection and Extract Preparation-HeLa and COS cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 500 mg/liter L-glutamine, and antibiotics. HeLa cells were transfected by calcium phosphate-mediated DNA precipitation (29). COS cells were transfected by the DEAEdextran method (30). Twenty-four hours later, cells were stimulated for 45 min unless otherwise indicated and then solubilized in lysis buffer as described (18). The protein concentration of lysates was determined by Bradford assay (31).
Protein Expression and Purification-Expression of bacterial GST fusion proteins and purification by affinity chromatography on GSH-Sepharose 4B beads (Pharmacia Biotech Inc.) was performed as described previously (32).
Protein Kinase Assays-Direct and coupled kinase assays were performed as described previously (18,22) with minor modifications. For in vitro kinase assays 0.5 g of recombinant kinase and 1 g of recombinant substrate were used, and for in vivo kinase assays 30 g of cell lysate was immunoprecipitated for 2 h with the anti-HA antibody 12CA5 (Boehringer Mannheim) and then incubated with 1 g of recombinant substrate. The concentration of [␥-32 P]ATP was 50 nM. Phosphorylated proteins were separated by SDS-polyacrylamide gel electrophoresis on 10% gels and then subjected to autoradiography. Incorporation of [ 32 P]phosphate was quantitated with a PhosphorImager and ImageQuant software (Molecular Dynamics, Inc.).

RESULTS
Isolation of MEK6 cDNA-We performed BLAST homology searches of the EST subdivision of the NCBI GenBank TM data bank with the tblastn program to identify EST sequences that encode peptides related to human MKK3. We identified a 223-bp EST fragment with the accession number F00521 that was related to MKK3 at the amino acid level but showed significant differences at the nucleotide level. A reverse PCR primer was used to amplify the 5Ј portion of the potential new gene from an adaptor-ligated Jurkat cDNA library. A popula-tion of 0.8-kb fragments was produced and subcloned into pGEM-T. Sequencing revealed several identical clones with long open reading frames preceded by about 250 bp of sequence with stop codons in all three reading frames. The sequence information from this 5Ј-untranslated region was used to design a forward PCR primer for 3Ј rapid amplification of cDNA ends. We amplified a 1.6-kb cDNA fragment from an adaptorligated MOLT-4 cDNA library. Sequence analysis of three independent clones revealed that all three clones were identical, which minimizes the possibility of introduction of point muta- tions during the PCR amplification procedure. A GenBank TM BLAST search revealed no similar sequences. Since the most recently published MAPKK was called MEK5 (33, 34), we named this new gene MEK6. The 1.6-kb cDNA encodes a potential protein of 334 amino acids with a calculated molecular weight of 37,500. MEK6 has 82% amino acid identity with its closest homologue, MKK3. All relevant kinase subdomains are conserved. The most divergent regions are the N-terminal region with an additional 18 amino acids and the C-terminal region (Fig. 1).
The expression of human MKK3 and MEK6 was examined by Northern blot analysis of RNA isolated from various human tissues. MKK3 is widely expressed in many adult human tissues with highest levels in skeletal muscle and leukocytes ( Fig.  2A). In contrast, MEK6 is predominantly expressed in skeletal muscle and at lower levels in heart and pancreas (Fig. 2B). All tissue samples expressed similar levels of ␤-actin mRNA (data not shown).
Substrate Specificity of MEK6 -To characterize the kinase activity of MEK6, we subcloned the cDNA into a bacterial GST fusion protein expression vector. We investigated the substrate specificity of MEK6 in an in vitro kinase assay with bacterially expressed MAPKK substrates (JNK2, p38, and ERK1(K52R)). Autophosphorylation of MEK6 (Fig. 3A) was very low compared with that of MKK3 (25). JNK2 was found to autophosphorylate in contrast to p38 and ERK1(K52R), which did not. MEK6 efficiently phosphorylated p38 but none of the other substrates (Fig. 3A, compare lanes 1-4 with lanes 5-8). To determine whether phosphorylation of p38 is an activating event we analyzed the phosphorylation of recombinant ATF2 (a substrate for p38) in a coupled in vitro kinase assay. MEK6 did not cause increased phosphorylation of Jun either directly or in combination with JNK2 (Fig. 3B, lanes 1-3). ATF2, however, was strongly phosphorylated by p38 that had been activated by MEK6 (Fig. 3B, lane 5). ATF2 was not directly phosphorylated by MEK6. These data establish that MEK6 is a functional MAPKK in vitro and that p38 is a substrate for MEK6.
MEK6 Is Activated by Anisomycin and UV-Next, we examined whether in vivo activated MEK6 can phosphorylate and activate p38 in an immune complex kinase assay. HeLa cells were transiently transfected with an expression vector encoding epitope-tagged MEK6 (3ϫHA-MEK6). After 24 h, 3ϫHA-  MEK6 was isolated by immunoprecipitation with an anti-HA monoclonal antibody (12CA5). In an initial experiment we investigated the time course of MEK6 activation by anisomycin and UV treatment of transfected cells. MEK6 activation by anisomycin as measured by its ability to phosphorylate p38 was observed as early as 10 min after treatment (Fig. 4). The activation was transient and peaked at 40 min after treatment. In contrast, activation by UV was delayed by about 10 -15 min and reached a plateau between 60 and 120 min (Fig. 4). Analysis of the UV dose response of MEK6 in HeLa cells revealed that doses up to 120 J/m 2 yielded increasing activity of MEK6 (Fig. 5).
To determine whether the increase in p38 phosphorylation by activated MEK6 augments p38 kinase activity, a coupled immune complex kinase assay was performed. Epitope-tagged MEK6 was isolated from anisomycin-treated HeLa cells and subjected to two subsequent kinase reactions using recombinant p38, ATF2, and GST alone. In support of our in vitro results, anisomycin treatment caused increased phosphorylation of ATF2 only when MEK6 and p38 were present (Fig. 6, compare lanes 5 and 6 with lanes 7 and 8). Similar results have been found with MEK6 activated by UV treatment of cells (data not shown). No inducible phosphorylation of p38 or ATF2 was observed in HeLa cells transfected with the empty expression vector SR␣3 (Fig. 6, compare lanes 5 and 6 with lanes 13 and  14). This clearly indicates that the inducible phosphorylation of ATF2 depends on a kinase cascade comprised of MEK6 and p38. Interestingly, p38 also phosphorylated weakly a protein with a mobility slightly faster than ATF2 (indicated by an asterisk in Fig. 6). This phosphorylation event was slightly augmented by anisomycin in the presence of MEK6 (Fig. 6,  compare lanes 3 and 4 with lanes 11 and 12).
MEK6 Is Activated by Stress-inducing Agents-MAPK cascades in mammalian cells respond to a variety of extracellular stimuli. To investigate the pattern of MEK6 regulation, cells were transiently transfected with expression vector for 3ϫHA-MEK6 and treated with various stimulators of the MAPK pathway. In HeLa cells, strongest inducers of MEK6 were UV, anisomycin, and NaCl followed by weak induction with interleukin-1␤ (Fig. 7A). NGF and EGF, two strong inducers of the ERK pathway, did not activate MEK6, although we noted the inducible phosphorylation of two lower molecular weight bands (see "Discussion"). Similar experiments were performed in COS cells, demonstrating a strong induction of MEK6 by UV and to a lesser extent by anisomycin (Fig. 7B). To exclude the possibility that changes of MEK6 kinase activity are caused by different levels of expression of MEK6 in response to treatment of cells with stimulators we performed Western blot analysis. MEK6 was present at equal levels in all cell lysates as determined by Western blot analysis with an anti-HA antibody (data not shown).
MEK6 Is Not a Physiological Substrate for MEKK-MEKK has been described as a MAPKKK leading to the phosphorylation and activation of JNKK (22)(23)(24). In an initial experiment with 1000 ng of cotransfected expression vector for MEKK we observed stimulation of MEK6 activity in COS cells but not in HeLa cells (Fig. 7, A and B). This prompted us to examine more carefully whether MEKK is able to activate MEK6. COS cells were transfected with increasing amounts of expression vector encoding MEKK in the presence of a constant amount of expression vector encoding epitope-tagged MEK6 or JNKK (Fig.  8). As assayed by JNK2 phosphorylation we observed strong JNKK activation in cells transfected with as little as 125 ng of the MEKK expression vector. Comparable amounts of MEK6 activation, however, were not observed until 1000 ng of the MEKK expression vector were cotransfected. These data suggest that MEKK is not a physiological activator of the MEK6 and p38 kinase cascade.

DISCUSSION
In this report, we describe the cloning and characterization of a novel member of the MAPKK family of dual specificity protein kinases, which we named MEK6. MEK6 was first identified in a BLAST homology search of the EST subdivision of NCBI GenBank TM as a 223-bp partial cDNA fragment. The full-length cDNA was obtained by PCR amplification of Jurkat and MOLT-4 cDNA libraries with gene-specific primers. We minimized the chance of introducing errors in the amplified sequence by using a DNA polymerase mixture with proofreading function. Further, we sequenced three independent clones that were identical. The cDNA has a long open reading frame that is preceded by about 250 bp of sequence with stop codons in all three reading frames. The size determination of in vitro translated protein indicates that translation starts at the identified start codon of MEK6 cDNA. Further, Northern blot analysis revealed that the size of MEK6 mRNA corresponds to the size of MEK6 cDNA. This suggests that we cloned the fulllength cDNA of MEK6. MEK6 has significant homology on the amino acid level to MKK3 (82% amino acid identity) but is not significantly related to MKK3 at the DNA level. This clearly demonstrates that MEK6 is encoded by a novel gene. MEK6 differs from MKK3 most significantly in the C terminus and N FIG. 6. Coupled kinase assay in HeLa cells. HeLa cells were transiently transfected with epitope-tagged MEK6 (lanes 1-8) or the empty expression vector SR␣3 (lanes 9 and 16) and treated for 45 min with anisomycin (An., 50 ng/ml) or left untreated (ctrl) as indicated. Cell lysates were prepared and used in a coupled immune complex kinase assay as described under "Experimental Procedures" and the legend to terminus, which has an additional 18 amino acids. All relevant kinase subdomains, the ATP acceptor site, and phosphorylation sites are conserved.
Many members of the MAPK cascades are expressed at high levels in skeletal muscle (18,25,33), which prompted us to compare the expression of MEK6 and its closest homologue MKK3 in 16 different adult human tissues. We observed very high expression of MEK6 mRNA in skeletal muscle, consistent with the high levels of other MAPK family members, but surprising was the absence of MEK6 mRNA in spleen, thymus, prostate, ovary, small intestine, colon, and leukocytes. Analysis of the same blot showed that MKK3 is widely expressed, and all 16 tissues expressed equal amounts of ␤-actin mRNA. We assume that MEK6 is expressed in these tissues at levels that are below the detection limit of Northern blot analysis. The fact that we isolated MEK6 cDNA from human T cells implies that activated T cells can express MEK6. Interestingly, some of the tissues expressed an MEK6-related mRNA of about 4.2 kb. This band was not observed when we used a MEK6-specific probe that was directed against the 3Ј end of MEK6 cDNA.
Similarities between MEK6 and MKK3 prompted us to investigate whether MEK6 is able to utilize p38 as substrate. In vitro MEK6 efficiently phosphorylated p38 but not ERK and JNK, although in parallel experiments the phosphorylation of JNK by JNKK was observed (data not shown). This indicates that MEK6, like MKK3, has substrate selectivity for the p38 subgroup of MAPK. Activation of p38 requires phosphorylation of Thr 180 and Tyr 182 (20). We subjected p38 that has been phosphorylated by MEK6 to a second kinase reaction with ATF2 as substrate. MEK6 induced phosphorylation of ATF2 by p38 but did not directly phosphorylate ATF2. These experiments were carried out with MEK6 prepared in bacteria as GST fusion protein and with epitope-tagged MEK6 isolated from HeLa cells after stimulation with anisomycin or UV. These results demonstrate that MEK6 specifically phosphorylates p38, resulting in its activation.
In some of the coupled kinase assays we also observed extra bands migrating slower and faster than ATF2. Some of these bands represent degradation products of ATF2 and phosphorylated forms of p38 and MEK6, as indicated in the figures. Additionally, we observed weak phosphorylation of a protein with a migration slightly faster than ATF2. This protein was not observed in in vitro kinase assays and therefore is most likely a contamination of the immunoprecipitation.
MEK6 is strongly activated by stress-inducing and DNAdamaging agents, anisomycin, UV, and also osmotic shock. Phorbol esters, NGF, and EGF, strong stimulators of the ERK pathway in the same cell lines analyzed (11,12,14), did not stimulate MEK6. Similarly, cycloheximide, a stimulator of p54 kinase (JNK2) (35) and of the ERK pathway (36), did not significantly activate MEK6. Anisomycin and cycloheximide are known to be effective protein synthesis inhibitors, yet the described effects on the MAPK pathways occur at concentrations that only marginally affect protein synthesis. Interestingly, we noted in our in vivo kinase assays with lysates prepared from HeLa cells but not from COS cells two bands of variable intensity that were stimulated by NGF and EGF. These bands most likely represent contaminants in the immunoprecipitate that became phosphorylated by ERK family members.
A time course experiment revealed that the induction of MEK6 by UV lagged behind the anisomycin induction by about 10 -15 min. Further, in contrast to the transient activation by anisomycin, UV-induced MEK6 stayed active for at least 120 min. The slight reduction of MEK6 activity at the 90-min time point was not observed in other experiments. This time course experiment suggests that different pathways are used for anisomycin versus UV stimulation of MEK6. The time course correlates well with the activation of JNK1 by UV (18). Additionally we tested different doses of UV. Consistent with the dose-response curves of JNK1 activation (18), we observed an increase in MEK6 stimulation with UV doses up to 120 J/m 2 .
MAPKK are activated by phosphorylation through MAP-KKK. MEKK (37), the upstream activator for JNKK, activates the JNK pathway much more efficiently than it does the ERK pathway (23). Overexpression of MEKK activates MEK1 and MEK2, but this does not cause activation of ERK2 (38). p38 activity was only weakly potentiated by cotransfected MEKK vector (in the presence or absence of the JNKK vector) in either COS or HeLa cells (22). In vitro, however, extracts from MEKK-transfected cells activated p38 as efficiently as they activated JNK1 (22). Therefore, at physiological concentrations JNKK seems to be the only substrate for MEKK.
Studies in yeast have shown the existence of a scaffolding protein (STE5) that independently associates with all members of the MAPKKK, MAPKK, and MAPK module (39,40). Although the precise function of STE5 is still open to speculation, researchers propose that mammalian cells also have a protein(s) that act(s) as scaffold for the MAP kinase module. Additionally, it is possible that this protein acts as a kinase itself and phosphorylates other MAP kinases. The existence of different MAP kinase modules held together by scaffolding protein(s) may contribute to signal and tissue specificity and may restrict the cross-talk between MAP kinase cascades as illustrated with the MEKK example. This prompted us to study whether MEKK can activate MEK6 in COS cells. We performed a titration experiment cotransfecting increasing amounts of the expression vector for MEKK with a fixed amount of the expression vector for epitopetagged MEK6 or JNKK, respectively. JNKK activity as assayed in vitro by phosphorylation of JNK2 was already strongly stimulated with 125 ng of MEKK, while similar levels of stimulation of MEK6 as assayed in vitro by phosphorylation of p38 needed 1000 ng of MEKK. This suggests that MEK6 is not a physiological substrate for MEKK, although we cannot exclude the possibility that in vivo scaffolding proteins are involved in bringing together a signaling complex containing MEKK, MEK6, and p38. A schematic presentation of the MAPK cascades based on our information and published information is presented in Fig. 9.