INTRODUCTION

The mitogen-activated protein kinase (MAPK)¹ cascade that consists of three protein kinases, MAPK, MAPK kinase (MAPKK), and MAPKK kinase, is conserved in many eukaryotic signal transduction pathways (1, 2, 3, 4). In budding yeast Saccharomyces cerevisiae several MAPK pathways have been identified, and these pathways are thought to function independently in distinct phenomena (5, 6). Recently, several subgroups of MAPK have been reported in vertebrate cells, including stress-activated protein kinase (SAPK)/c-Jun NH₂-terminal kinase (JNK) (7, 8) and p38 (also known as HOG1, MPK2, or CSBP) (9, 10, 11) in addition to classical MAPKs. Classical MAPKs are stimulated by growth factors and tumor promoters, whereas SAPK/JNK and p38 are activated in response to environmental stresses and cytokines (1, 2, 3, 4, 7, 8, 9, 10, 11, 12, 13). SAPK/JNK and p38 partially overlap in activating stimuli and their substrates. For example, ATF-2 is phosphorylated and activated in its transcriptional activity by both SAPK/JNK and p38 (14, 15, 16, 17), whereas MAPKAP kinase-2 can be activated by p38 but not by SAPK/JNK (10, 18). SAPK/JNK and p38 can complement a yeast osmosensitive hog1 mutant (9, 19), suggesting that both SAPK/JNK and p38 can be activated by PBS2, a HOG1 activator that is a member of the yeast MAPKK superfamily molecules.

Several members of vertebrate MAPKK superfamily molecules can activate SAPK/JNK and p38, such as MKK3 (20), MKK4/SEK1/JNKK (20, 21, 22), and MAPKK6/MKK6/MEK6 (17, 23, 24, 25). MKK3 and MAPKK6 can act as a specific activator for p38, and MKK4/SEK1 may function as an activator for SAPK/JNK, although it can also activate p38 in vitro (13, 20, 21, 22, 23, 24, 25, 26). However, it is unclear which of these MAPKK family molecules functions in vivo in each signaling pathway, as purification or identification of a MAPKK family molecule in each pathway has not been reported.

To identify an in vivo activator for p38 in an osmosensing pathway, we fractionated extracts obtained from mouse lymphoma L5178Y cells exposed to hyperosmolar media, purified a major activator for p38, and identified it as MAPKK6. Furthermore, we searched for a new mammalian activator for p38 by making use of a yeast osmoregulation MAPK cascade, and isolated MKK3b, an alternatively spliced form of MKK3. By using several antibodies, we showed that MKK3/MKK3b and MAPKK6 are expressed differently in various types of cells and that MAPKK6 is activated strongly in KB cells by tumor necrosis factor- (TNF-), H₂O₂ and hyperosmolarity and thus may act as a major activator for p38 in these pathways.

EXPERIMENTAL PROCEDURES

His-tagged wild type p38, His-tagged kinase-negative (KN-) MPK2, and ATF-2 were expressed in Escherichia coli and purified as described previously (25, 26). GST-p38 was prepared by using the expression vector pGEX-2T (Pharmacia Biotech Inc.) and purified by affinity chromatography on glutathione-Sepharose 4B (Pharmacia). An anti-MKK3 polyclonal antibody was purchased from Santa Cruz Biotechnology Inc. This antibody was used for immunoblotting at 1:100 dilution (1 µg of IgG/ml). Anti-MAPKK6 polyclonal antibodies were raised in both rabbits and mice by immunizing them with His-tagged MAPKK6 (25). These antisera were used for immunoprecipitation (3 µl of antiserum/200 µl of cell extracts) and immunoblotting (at 1:500 dilution). An anti-MAPKK6 monoclonal antibody (3F2) was produced by immunizing mice with His-tagged MAPKK6 (25) and used for immunoblotting (1 µg of IgG/ml).

L5178Y cells were cultured in RPMI 1640 containing 10% fetal calf serum. KB cells were cultured in Dulbecco's modified Eagle's medium containing 10% calf serum. COS7 cells were cultured and transiently transfected using LipofectAMINE (Life Technologies Inc.) as described previously (25).

An aliquot (7 µl) of the fractions was incubated with 100 µM ATP and 20 mM MgCl₂ in the presence or absence of 1 µg of His-tagged p38 (final volume 10 µl) at 30 °C. After 30 min, 5 µl of 0.5 mg/ml recombinant ATF-2 and 1 µCi of [-³²P]ATP were added and incubated for another 20 min at 20 °C. Reactions were stopped by the addition of Laemmli's sample buffer and boiling. One unit of p38 activator is defined as 1 nmol of [³²P]phosphate incorporated into recombinant ATF-2 per min in the above reaction. To estimate p38 activator activity in crude extracts, 1 µg of GST-p38 was incubated in the extracts in the presence of 50 µM ATP and 20 mM MgCl₂ for 30 min at 30 °C. Then, 40 µl of 1:1 slurry of glutathione-Sepharose 4B beads was added and rotated at 4 °C for 1 h. The beads were washed three times with a solution containing 20 mM Tris-Cl, pH 7.5, 500 mM NaCl, 2 mM dithiothreitol, and 0.05% Tween 20 and washed once with buffer A consisting of 20 mM Tris, pH 7.5, 2 mM EGTA, 25 mM 2-glycerophosphate, 2 mM dithiothreitol, and 1 mM vanadate. The beads were incubated with 2.5 µg of ATF-2, 50 µM [-³²P]ATP (1 µCi), and 20 mM MgCl₂ in a final volume of 15 µl. The reactions were terminated after 20 min at 20 °C by the addition of Laemmli's sample buffer and boiling.

A 7-liter culture of L5178Y cells at 7.4 × 10⁵ cells/ml was stimulated by 0.7 M NaCl for 30 min at 37 °C, washed twice in ice-cold Hepes-buffered saline, collected by centrifugation, quick-frozen in liquid nitrogen, and stored at 80 °C. About 2.0 × 10¹⁰ frozen L5178Y cells (20-liter culture of L5178Y cells) were the starting material for one-cycle purification. Cytosolic extracts were prepared as described (26) and loaded onto a 150-ml Q-Sepharose Fast Flow column (Pharmacia) equilibrated with buffer A. The column was washed with 300 ml of buffer A, and unadsorbed fractions were pooled and used as peak 1 fraction. Proteins that bound to the column were eluted with 500 ml of buffer A containing 0.1 M NaCl to yield peak 2 fraction. The peak 1 fraction was loaded onto a 30-ml blue-Sepharose CL-6B (Pharmacia) column equilibrated with buffer A. The column was washed with 100 ml of buffer A, and proteins were eluted with 120 ml of buffer A containing 1.2 M NaCl. 10 ml of buffer A containing 5 M NaCl was added to the eluted fraction (final NaCl concentration was adjusted to ~1.2 M NaCl) and loaded onto a 3 × 5-ml HiTrap phenyl-Sepharose HP (Pharmacia) column. After washing the column, proteins were eluted with 225 ml of decreasing linear gradient of NaCl in buffer B (buffer A + 0.01% Brij35) and eluted further with 30 ml of buffer B. The active fractions were pooled and concentrated to 2.5 ml by Centriprep 30 (Amicon). Two-ml concentrated phenyl-Sepharose HP pools were loaded onto a HiLoad 16/60 Superdex 200 gel filtration column (Pharmacia) equilibrated in buffer B supplemented with 0.1 M NaCl. The flow rate was 0.8 ml/min, and 0.8-ml fractions were collected. 5.6 ml of active fractions were pooled, diluted to 12 ml with buffer B, and applied to a HiTrap heparin (1 ml, Pharmacia) equilibrated with buffer B. The column was washed with 5 ml of buffer B and developed with an 18-ml NaCl gradient. The peak 2 fraction of Q-Sepharose chromatography was subjected to sequential chromatography in the same manner as above except for omitting blue-Sepharose column chromatography.

Concentrated proteins from three-cycle purification steps described above were resolved by SDS-polyacrylamide gel electrophoresis and transferred to a polyvinylidene difluoride membrane. The band was reduced and S-carboxymethylated in situ and digested with Achromobacter protease I (27). Peptides released from the membrane were fractionated by reverse-phase high performance liquid chromatography on a Wakosil-II 5C18 AR column and subjected to amino acid sequence analysis with a Shimadzu gas-phase sequenator model PSQ-10.

The DNA fragment of human p38 was cloned into a yeast integrating plasmid. Yeast strain TM334 (MAT ura3 leu2 trp1 his3 lys2 pbs2::HIS3) (28) was transformed by this p38 plasmid to yield TM334[YIplac128-p38]. The HeLa and Jurkat cDNA libraries were constructed in the yeast expression vector pNV7 (29). The murine BAF-B03 library was described previously (30). TM334[YIplac128-p38] was transformed with BAF-B03, HeLa, and Jurkat libraries, and approximately 3 × 10⁵ clones were screened, respectively. Isolated murine and human MKK3b clones could suppress the pbs2 defect in the presence of p38. The detailed procedures for this method will be described elsewhere.² The sequences of all plasmids were determined by dye terminator sequencing with an Applied Biosystems model 373A machine.

RESULTS AND DISCUSSION

To purify p38 activator, we utilized a mouse T lymphoma cell line, L5178Y, which grows rapidly and has a high p38 activating activity when exposed to hyperosmolar media. L5178Y cells were stimulated by 0.7 M NaCl, and extracts obtained from these cells were subjected to chromatography on Q-Sepharose Fast Flow (Fig. 1A). The p38 activating activity was measured by increased phosphorylation of ATF-2 in the presence of recombinant p38. Two peaks of p38 activating activity were observed; one was eluted in the flow-through and the other in 0.1 M NaCl fractions, termed peak 1 and peak 2, respectively. The ATF-2 phosphorylating activities that were eluted in the 0.2-0.3 M NaCl fractions were attributed mostly to activation of endogenous MAPK and SAPK/JNK, and the ATF-2 phosphorylating activity that was eluted in the 0.35 M NaCl was due to activation of endogenous p38 (data not shown). No significant p38 activating activity was detected in the 0.2-0.35 M NaCl fractions even when we adopted another method of detecting p38 activating activity which uses GST-p38 as described under ``Experimental Procedures.'' The p38 activating activity in the combined fractions of peak 1 and peak 2 was stimulated maximally when the cells were exposed to 0.4-0.8 M NaCl (Fig. 1B), and the activation was time-dependent (Fig. 1C).

Characterization of p38 activating activity from L5178Y cells. Panel A, L5178Y cells (250 ml of culture) were exposed to 0.7 M NaCl for 30 min. Soluble extracts obtained from these cells were subjected to Q-Sepharose chromatography (10 ml), and each fraction was assayed for p38 activating activity by measuring ATF-2 phosphorylating activity in the absence (p38) or presence (+p38) of His-tagged p38. Phosphorylation of ATF-2 was detected by autoradiography (upper panels) or quantified by Fujix BAS2000 (lower panel). Panels B and C, L5178Y cells were exposed to indicated concentration of NaCl for 30 min (panel B) or exposed to 0.7 M NaCl for the indicated times (panel C). Cell extracts were obtained and mixed with a 0.5 volume of Q-Sepharose beads equilibrated with buffer A (see ``Experimental Procedures'') containing 0.1 M NaCl. The unadsorbed fractions were assayed for p38 activating activity by measuring ATF-2 phosphorylating activity in the absence (open circles) or presence (closed circles) of His-tagged p38 as described under ``Experimental Procedures.''

To purify p38 activator further, the peak 1 and peak 2 fractions from Q-Sepharose chromatography were subjected separately to sequential chromatography on blue-Sepharose (for only peak 1), phenyl-Sepharose HP, Superdex 200, and heparin-Sepharose (Table I and Fig. 2). Both peak 1 and peak 2 activities gave similar elution profiles on each column chromatography (Fig. 2, A and B). The column fractions from the heparin chromatography (for peak 1, see Fig. 2A) were analyzed by SDS-polyacrylamide gel electrophoresis and silver staining (Fig. 2C). Three proteins, with apparent molecular masses of 40, 38, and 36 kDa, were detected in the peak fractions of p38 activating activity (indicated by arrowheads in Fig. 2C). Since 36-kDa protein was eluted coincidentally with p38 activating activity, it was subjected to protein microsequencing as described under ``Experimental Procedures.'' Seven polypeptide sequences derived from 36-kDa protein were determined. As shown in Fig. 3, five sequences of the peptides perfectly matched the sequence of human MAPKK6 (25). The other two peptide sequences (AP-6 and AP-7) had high similarities to MAPKK6, and the sequence of AP-6 matched completely MKK6c (23). MKK6c is a murine MAPKK6 reported previously which has only 237 amino acids and lacks the kinase subdomains I and II. The 40- and 38-kDa proteins were also subjected to microsequencing were found to be hnRNP-E2 protein and isocitrate dehydrogenase (NAD), respectively. These results indicate that the p38 activator activity in peak 1 of the Q-Sepharose chromatography resides in 36-kDa protein that is a murine homolog of MAPKK6. Thus, there is a murine MAPKK6 whose molecular size is similar to that of human MAPKK6, in addition to MKK6c, a truncated form of a murine homolog of MAPKK6. In these purification steps described above, major SAPK/JNK activating activities were separated from the p38 activating activities,³ indicating that direct activators for p38 are different from those for SAPK/JNK and endogenous MAPKK6 activates only p38.

Table I. Purification of p38 activators from stimulated L5178Y cells Step Volume Total protein Activity Specific activity Yield Purification ml mg units units/mg % fold Peak 1   Total extract 150 930  Q-Sepharose 500 175 233 1.33 100 1  Blue-Sepharose 150 110 198 1.8 85 1.4  Phenyl-Sepharose 45 8.3 126 15.1 54 11.4  Centriprep 30 2.5  Superdex 200 5.6 0.16 53.6 335.6 23 252.3  Heparin-Sepharose 2.5 0.018 42.5 2360 18 1774 Peak 2   Total extract 150 930  Q-Sepharose 500 92 191 1.79 100 1  Phenyl-Sepharose 45 7.9 122 13.4 64 7.5  Centriprep 30 2.5  Superdex 200 5.6 0.10 40.1 346 21 193  Heparin-Sepharose 2.5 0.007 32.5 3843 17 2147

To identify other possible activators for p38, we used another approach, the complementation of a yeast pbs2 mutant by BAF-B03, HeLa, and Jurkat cDNA libraries in a p38-dependent manner. We isolated murine and human clones, which were termed mMKK3b and hMKK3b, respectively. These clones have an open reading frame whose predicted amino acid sequence is almost identical to MKK3 except for the NH₂-terminal region. Their cDNA sequences lack the first in-frame termination codon, which is present in the 5-untranslated region of the MKK3 cDNA and have another in-frame initiation codon upstream of the initiation codon of MKK3 (data not shown). These clones, therefore, may be an alternatively spliced form of MKK3 and thus are termed MKK3b. The deduced amino acid sequences of murine and human MKK3b cDNAs are shown in Fig. 4A. MKK3b contains 347 amino acids, and MKK3 contains 318. 

To examine whether MKK3b has kinase activity, COS7 cells were transiently transfected with an expression vector encoding an epitope tagged MKK3b (HA-MKK3b). After 24 h, cells were exposed to hyperosmolarity, and HA-MKK3b was immunoprecipitated with anti-HA monoclonal antibody (12CA5) and examined for kinase activity toward kinase-negative MPK2 (a Xenopus homolog of p38). The immunoprecipitated HA-MKK3b had kinase activity, and its activity was increased after osmotic shock (Fig. 4B). This HA-MKK3b did not catalyze phosphorylation of recombinant SAPK/JNK markedly (data not shown), indicating that MKK3b is also a specific activator for p38, like MKK3 and MAPKK6. HA-MKK3 and HA-MAPKK6 were also expressed as above, and their activity was examined. As shown in Fig. 4B, HA-MAPKK6 had the highest activity and MKK3 the lowest activity among three kinases.

To characterize endogenous MAPKK6 further, we produced anti-MAPKK6 antibodies by using a bacterially expressed recombinant MAPKK6 as an antigen. Mouse monoclonal anti-MAPKK6 antibody reacted strongly with HA-MAPKK6, which was overexpressed in COS7 cells, and not with HA-MKK3 or HA-MKK3b (Fig. 5A, MAPKK6). Rabbit and mouse polyclonal antibodies against MAPKK6 showed the same reactivity (data not shown). Rabbit polyclonal antibody raised against the COOH-terminal peptide of MKK3 (Santa Cruz Biotechnology Inc.) reacted strongly with HA-MKK3 and HA-MKK3b and faintly with HA-MAPKK6 (Fig. 5A, MKK3). In total L5178Y extracts, the anti-MAPKK6 antibody reacted with a 36-kDa band, and the anti-MKK3 antibody reacted with 37- and 35-kDa bands, which were clearly different from 36-kDa MAPKK6 in their migration on SDS-polyacrylamide gel electrophoresis (Fig. 5B). When exposed further, the 36-kDa band was detected very weakly by anti-MKK3 antibody (data not shown). These data have suggested that 35-, 36-, and 37-kDa bands correspond to murine MKK3, MAPKK6, and MKK3b, respectively. The 36-kDa band that reacted with anti-MAPKK6 antibodies was expressed highly in HeLa cells, KB cells, and porcine brain, but not in PC12 cells, 3Y1 cells, and Mv1Lu cells (Fig. 5C, and data not shown). The 35-kDa and/or 37-kDa band(s), which reacted with anti-MKK3 antibody, could be detected in all cells examined (Fig. 5C).

After chromatography on Q-Sepharose (see Fig. 1A), the peak 1 and peak 2 p38 activating activities were subjected separately to chromatography on phenyl-Sepharose HP (Fig. 2, A and B), and eluted proteins were analyzed by Immunoblotting with anti-MAPKK6 antibody (Fig. 6A, peak 1 and peak 2). Rather surprisingly, a large amount of MAPKK6 was eluted in the low salt concentration fractions (Fig. 6A, fractions 29-34 for peak 1; fractions 30-34 for peak 2) where no significant p38 activating activity was observed, and a small amount of MAPKK6 was coeluted in the peak fractions of p38 activating activity (Fig. 6A, fractions 20-23 for peak 1; fractions 20-22 for peak 2). This result indicated that a small portion of MAPKK6 was activated in response to hyperosmolarity, although hyperosmolarity was one of the strongest stimuli among various cellular stresses and cytokines to activate MAPKK6 (see below). MKK3 (35 kDa) and MKK3b (37 kDa), which reacted with anti-MKK3 antibody, were eluted in the flow-through fractions and in fractions 1-10, separate from the p38 activating activity on phenyl-Sepharose HP chromatography. In the final heparin chromatography fractions from peak 2 (see Fig. 2B, Heparin), MAPKK6 was coeluted with the peak fractions of p38 activating activity (Fig. 6B). These results have suggested that the p38 activating activity in peak 2 also resides in activated MAPKK6. It may be that activated MAPKK6 was resolved into two peaks (peak 1 and peak 2) on Q-Sepharose chromatography possibly because of the existence of at least two kinds of activated forms due to difference in the phosphorylation state, like MAPKK1 (31, 32).

To characterize further the p38 activating activity in extracts obtained from stimulated L5178Y cells, MAPKK6 was immunodepleted from the extracts using rabbit polyclonal antibody raised against recombinant MAPKK6, and p38 activating activity remaining in the supernatants was measured using GST-p38 as described under ``Experimental Procedures.'' The p38 activating activity in the supernatant was decreased to less than 10% of the original value by the addition of increasing amounts of anti-MAPKK6 antibody (Fig. 7A). The immunoblotting of the supernatant with anti-MAPKK6 antibody and anti-MKK3 antibody revealed that MAPKK6 was almost completely removed by this immunodepletion procedure, whereas MKK3 and MKK3b were not removed at all (Fig. 7B). This again indicated that MAPKK6 is a major activator for p38 in an osmosensing pathway in L5178Y cells, consistent with the observation that two major peaks of the p38 activating activity in Q-Sepharose chromatography were accounted for by activated MAPKK6.

MAPKK6

MKK3

It has been reported that MAPKK6, when overexpressed in cells, can be activated by UV irradiation, osmotic shock, and anisomycin (17, 24, 25). To identify extracellular stimuli that can induce activation of endogenous MAPKK6 and p38, we followed their activity in response to a variety of stimuli in KB cells that expressed MAPKK6 as well as MKK3/MKK3b (Fig. 5C). Immune complex protein kinase assays revealed that MAPKK6 (Fig. 8, lower panels) was activated strongly by TNF- (Fig. 8A), H₂O₂ (Fig. 8B), and okadaic acid (Fig. 8C), moderately by cycloheximide (Fig. 8E), and very weakly by epidermal growth factor (Fig. 8D). The magnitude and the time course of changes in the kinase activity of MAPKK6 correlated with those of p38 (Fig. 8, upper panels). Then, we performed the immunodepletion experiment with anti-MAPKK6 antibody in the extracts obtained from the KB cells that had been stimulated by osmotic shock or H₂O₂. In both cases, the p38 activating activity in the stimulated cell extracts was reduced to less than 10% of the original level by immunodepletion of MAPKK6 protein (data not shown and Fig. 8B inset), indicating that MAPKK6 is a major activator for p38 in osmotic shocked and H₂O₂-stimulated pathways in KB cells. Meier et al. (33) reported recently that a major activator peak for p38 in Mono S chromatography (termed SAPKK-3 fraction) accounted for >95% of the p38 activator activity in KB cells stimulated by interleukin-1, UV, anisomycin, and sorbitol. It is likely that the entity of this active fractions is MAPKK6.

In this study, we have shown that MAPKK6 as well as p38 is strongly activated by TNF- in KB cells, but the other study did not observe activation of MAPKK6 by TNF- in COS cells and HeLa cells (24). This might be caused by the difference of cell lines used. Cycloheximide, which has been reported to be a stimulator of SAPK/JNK (8), caused a modest increase of MAPKK6 and p38 activity in KB cells. The MAPKK6 activation induced by cycloheximide was slow compared with that induced by other stimuli such as TNF-. The oxidative agent H₂O₂ has also been shown to induce the strong activation of MAPKK6 and p38 in this study. A previous study has shown that H₂O₂ activates strongly classical MAPK, and activated MAPK plays a critical role in cell survival following oxidant injury (34). We might hypothesize that p38 is involved in cell death induced by oxidative stress because p38 has been implicated in nerve growth factor withdrawal-induced cell death in PC12 cells (35). It has been reported recently, however, that MAPKAP kinase-2, a target of p38, may be a major component of the signal transduction pathway which triggers the adaptive response to oxidative stress (36). Further studies will be needed to reveal the roles of the MAPKK6/p38 cascade and classical MAPK cascade in oxidative stress. The function of the MAPKK6/p38 pathway in TNF- signal transduction is also to be clarified.

In this study, we have identified a major activator for p38 as MAPKK6 in L5178Y cells and KB cells stimulated by osmotic shock. We isolated a novel p38 activator, termed MKK3b, which is an alternatively spliced form of MKK3. MKK3, MKK3b, and MAPKK6 can act as a specific activator for p38, but they differ in the expression pattern and the magnitude of kinase activity, suggesting that they might have different functions. Furthermore, the MAPKK6/p38 kinase cascade has been shown to be activated strongly by TNF- and H₂O₂. It could be speculated that MAPKK6, when expressed, may act as a major activator for p38 in various signaling pathways.