Molecular cloning of mitogen-activated protein/ERK kinase kinases (MEKK) 2 and 3. Regulation of sequential phosphorylation pathways involving mitogen-activated protein kinase and c-Jun kinase.

Mitogen-activated protein/ERK kinase kinases (MEKKs) phosphorylate and activate protein kinases which in turn phosphorylate and activate the p42/44 mitogen-activated protein kinase (MAPK), c-Jun/stress-activated protein kinases (JNKs), and p38/Hog1 kinase. We have isolated the cDNAs for two novel mammalian MEKKs (MEKK 2 and 3). MEKK 2 and 3 encode proteins of 69.7 and 71 kDa, respectively. The kinase domains encoded in the COOH-terminal moiety are 94% conserved; the NH-terminal moieties are approximately 65% homologous, suggesting this region may encode sequences conferring differential regulation of the two kinases. Expression of MEKK 2 or 3 in HEK293 cells results in activation of p42/44 and JNK but not of p38/Hog1 kinase. Immunoprecipitated MEKK 2 phosphorylated the MAP kinase kinases, MEK 1, and JNK kinase. Titration of MEKK 2 and 3 expression in transfection assays indicated that MEKK 2 preferentially activated JNK while MEKK 3 preferentially activated p42/44. These findings define a family of MEKK proteins capable of regulating sequential protein kinase pathways involving MAPK members.

Mitogen-activated protein/ERK kinase kinases (MEKKs) phosphorylate and activate protein kinases which in turn phosphorylate and activate the p42/44 mitogen-activated protein kinase (MAPK), c-Jun/stressactivated protein kinases (JNKs), and p38/Hog1 kinase. We have isolated the cDNAs for two novel mammalian MEKKs (MEKK 2 and 3). MEKK 2 and 3 encode proteins of 69.7 and 71 kDa, respectively. The kinase domains encoded in the COOH-terminal moiety are 94% conserved; the NH 2 -terminal moieties are approximately 65% homologous, suggesting this region may encode sequences conferring differential regulation of the two kinases. Expression of MEKK 2 or 3 in HEK293 cells results in activation of p42/44 MAPK and JNK but not of p38/Hog1 kinase. Immunoprecipitated MEKK 2 phosphorylated the MAP kinase kinases, MEK 1, and JNK kinase. Titration of MEKK 2 and 3 expression in transfection assays indicated that MEKK 2 preferentially activated JNK while MEKK 3 preferentially activated p42/ 44 MAPK . These findings define a family of MEKK proteins capable of regulating sequential protein kinase pathways involving MAPK members.
A variety of extracellular signals including growth factors, hormones, cytokines, antigens, and stresses such as heat shock and osmotic imbalance activate members of the mitogen-activated protein kinase (MAPK) 1 family (1)(2)(3)(4)(5). MAPKs are char-acterized as serine/threonine-protein kinases activated by dual phosphorylation on both a tyrosine and a threonine (6). The MAPK family includes p42/44 MAPK (also referred to as ERK2 and -1) (7), the c-Jun kinases (JNKs which are also referred to as stress-activated protein kinases) (8), and p38, the osmotic imbalance responsive kinase similar to the yeast Hog1 enzyme (5). The regulation of different MAPKs including p42/44 MAPK , JNKs, and p38 involves sequential protein kinase pathways whose upstream activators include the MEKs (MAPK/ERK kinases) and the JNK kinases (also referred to as SEKs or stress/ERK kinases) (9 -13). MEKs and JNK kinases (JNKKs) phosphorylate specific MAPK family members on both a tyrosine and threonine resulting in MAPK activation.
Raf-1 and B-Raf are serine/threonine-protein kinases that selectively phosphorylate and activate MEK 1 and MEK 2 (14 -17). Recently, we isolated the cDNA for a novel serine/ threonine-protein kinase referred to as MEK kinase (MEKK 1) that phosphorylates and activates MEK 1 and 2 and JNKKs (13, 18 -20). The catalytic domain of MEKK 1 is homologous to the kinase domains of the Ste11 and Byr2 serine/threonineprotein kinases, involved in the control of mating in Saccharomyces cerevisiae and Schizosaccharomyces pombe, respectively (21)(22)(23). In this report, we have isolated and expressed the cDNAs for two new MEKK proteins. MEKK 2 and 3, when expressed in cells, are similar to MEKK 1 in that they are capable of regulating MEK and JNKK activities. Thus, there exists a family of MEKKs controlling sequential protein kinase systems involving MAPK members in addition to the Raf family of protein kinases.

Isolation of MEKK 2 and 3 cDNAs-
The degenerate primers GA(A/ G)(C/T)TIATGGCIGTIAA(A/G)CA (sense) and TTIGCICC(T/C)TTIAT-(A/G)TCIC(G/T)(A/G)TG (antisense) were used in a polymerase chain reaction (PCR) using first strand cDNA generated from polyadenylated RNA prepared from NIH 3T3 cells. The PCR reaction involved 30 cycles (1 min, 94°C; 2 min, 52°C; 3 min, 72°C) followed by a 10-min cycle at 72°C. A band of approximately 300 base pairs was recovered from the PCR mixture, and the products were cloned into pGEM-T (Promega). The PCR cDNA products were sequenced and compared to the MEKK 1 sequence (19). A unique cDNA sequence of 322 base pairs having significant homology to MEKK 1 cDNA was identified and used to screen an oligo(dT)-primed mouse brain cDNA library (Stratagene). The phage library was plated, and DNA from plaques were transferred to Hybond N filters (Amersham) followed by UV-cross-linking of DNA to the filters. Filters were prehybridized for 2 h and then hybridized overnight in 0.5 M Na 2 H 2 PO 4 (pH 7.2), 10% bovine serum albumin, 1 mM EDTA, 7% SDS at 68°C. Filters were washed twice at 42°C with 2 ϫ SSC, once with 1 ϫ SSC, and once with 0.5 ϫ SSC containing 0.1% SDS (1 ϫ SSC is 0.15 M NaCl, 0.015 M sodium citrate, pH 7.0). Positive hybridizing clones were purified and sequenced. To resolve GC-rich * This work was supported in part by National Institutes of Health Grants DK 37871, GM 30324, CA 58157, and DK 48845. 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.
regions, cDNAs were subcloned into M13 vectors (New England Biolabs), and single strand DNA was sequenced. In all cases, both strands of DNA were sequenced. Plasmid Expression of MEKK 2 and 3-The proteins for MEKK 2 and 3 were epitope-tagged at their NH 2 terminus with the hemagglutinin (HA)-tag sequence GYPYDVPDYAS (24) using a PCR strategy as described previously (25). For inserting the NH 2 -terminal epitope tag in MEKK 2 and 3, sense oligonucleotides were synthesized having a methionine codon (ATG), 33 bases coding for the GYPYDVPDYAS epitopetag sequences, and 20 bases of MEKK 2 or 3 sequence starting at codon 2. For MEKK 2, the sense oligonucleotide was ATGGGGTACCCGTA-CGACGTGCCGGACTACGCTTCCGATGATCAGCAAGCTTTGAA. The sense oligonucleotide for MEKK 3 was ATGGGGTACCCGTACGACG-TGCCGGACTACGCTTCCGATGAACAAGAGGCATTAGA. The antisense oligonucleotides for MEKK 2 and 3 were AGACTTAGATCTCAG-GTCTTC encoding a BglII site for MEKK 2 and GATTCTGACGTC-ACTCTGCCT encoding an AatII site for MEKK 3. The PCR reactions were performed for 30 cycles using MEKK 2 or MEKK 3 cDNAs as template. The PCR products were purified, and a second PCR reaction was performed using the first PCR product as template, the MEKK 2 or 3 antisense oligonucleotide described above and the common sense oligonucleotide encoding a XbaI restriction site, a consensus Kozak initiation site and 17 bases overlapping with the initiation methionine, and HA-tag sequence (TGACGTTCTAGAGCCACCATGGGGTAC-CCGTACGA). The resulting PCR products were digested with XbaI and BglII for MEKK 2 and XbaI and AatII for MEKK 3 and ligated in-frame into the appropriate MEKK 2 or 3 cDNA. The sequences were confirmed by DNA sequencing, and the cDNAs were inserted into the expression plasmid pCMV5. HEK293 cells were transfected with pCMV5 expression plasmids using LipofectAMINE (Life Technologies, Inc.) and assayed 48 h later. The 12CA5 monoclonal antibody (Berkeley Antibody Co.) was used for recognition of the HA epitope tag encoded in expressed MEKK 2 and 3.
Antibody Production-Peptides corresponding to COOH-terminal sequences of MEKK 3 (CEARQRPSAEELLTHHFAQ) and p38 (CFVPP-PLDQEEMES) were conjugated to keyhole limpet hemocyanin and used to immunize rabbits. Antisera were characterized for specificity by immunoblotting of lysates prepared from appropriately transfected HEK293 cells.
Assay of JNK Activity-JNK activity was measured using GST (glutathione S-transferase)-c-Jun  coupled to glutathione-Sepharose 4B (26). Cells transfected with MEKK 2 or 3 and control transfected cells were lysed in 0.5% Nonidet P-40, 20 mM Tris⅐HCl, pH 7.6, 0.25 M NaCl, 3 mM EDTA, 3 mM EGTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 2 mM sodium vanadate, 20 g/ml aprotinin, and 5 g/ml leupeptin. Nuclei were removed by centrifugation at 15,000 ϫ g for 10 min, and the supernatants (25 g of protein) were mixed with 10 l of a slurry of GST-c-Jun (1-79) -Sepharose (3-5 g of GST-c-Jun  ). The mixture was rotated at 4°C for 1 h, washed twice in lysis buffer and once in kinase buffer (20 mM Hepes, pH 7.5, 10 mM MgCl 2 , 20 mM ␤-glycerophosphate, 10 mM p-nitrophenyl phosphate, 1 mM dithiothreitol, 50 M sodium vanadate). Beads were suspended in 40 l of kinase assay buffer containing 10 Ci of [␥-32 P]ATP and incubated at 30°C for 20 min. Reaction mixtures were added to Laemmli sample buffer, boiled, and phosphorylated proteins were resolved on SDS-10% polyacrylamide gels. When JNK activity was assayed following fractionation by Mono Q ion exchange chromatography, 50 l of each fraction was incubated with the GST-c-Jun  beads.
To demonstrate MEKK activation of JNKK activity, the in vitro kinase reactions were performed with different combinations of recombinant wild type or kinase-inactive JNKK (lysine 116 mutated to methionine) and wild type or kinase-inactive JNK. Kinase-inactive JNK was made by mutating the active site lysine 55 to methionine (provided by Dr. Matt Jarpe). Incubations were for 30 min at 30°C in the presence of 50 M ATP. GST-c-Jun (1-79) -Sepharose beads were then added, and the mixture was rotated at 4°C for 30 min. The beads were washed, suspended in 40 l of c-Jun kinase assay buffer containing 20 Ci of [␥-32 P]ATP, and incubated for 15 min at 30°C. Reaction mixtures were added to Laemmli sample buffer, boiled, and phosphorylated proteins were resolved on SDS-10% polyacrylamide gels.
Assay of p38 Kinase Activity-Sorbitol-treated (0.4 M, 20 min) or control HEK293 cells were lysed in the same buffer as that used for assay of MEKK 2 and 3. Supernatants (200 g of protein) were used for immunoprecipitation of p38 using rabbit antiserum raised against the COOH-terminal peptide sequence of p38 (5). Immunoprecipitates were washed once in lysis buffer, once in assay buffer (25 mM Hepes, pH 7.4, 25 mM ␤-glycerophosphate, 25 mM NaCl 2 , 2 mM dithiothreitol, 0.1 mM sodium vanadate), resuspended, and used in an in vitro kinase assay with a recombinant NH 2 -terminal fragment of ATF 2 (20 -50 ng) as substrate and 20 Ci of [␥-32 P] ATP (31). For analysis of p38 kinase activity from Mono Q FPLC fractions, 20-l aliquots were mixed with kinase buffer containing 20 -50 ng of recombinant ATF 2 and 10 Ci of [␥-32 P]ATP (29,31). Reactions were quenched in Laemmli sample buffer, boiled, and proteins were resolved using SDS-10% polyacrylamide gels.

RESULTS
Cloning of MEKK 2 and 3-Degenerate primers were used in polymerase chain reactions with cDNA synthesized from RNA isolated from NIH3T3 cells (see "Materials and Methods"). Approximately 80% of the polymerase chain reaction products when sequenced encoded MEKK 1 (18). 15-20% of the reaction products encoded a novel cDNA sequence having homology to MEKK 1. The unique cDNA fragment was used to screen a mouse brain cDNA library which resulted in the isolation of two unique MEKK cDNAs. Fig. 1 shows the DNA sequence and deduced amino acid sequence for MEKK 2 and 3. MEKK 2 encodes a 619-amino acid protein having a mass of 69.7 kDa. MEKK 3 encodes a 626-amino acid protein having a mass of 71 kDa. The two proteins share a common structure with the kinase catalytic domain encoded in the COOH-terminal moiety. The amino-terminal moiety does not encode any definable domain such as a SH2 or SH3 domain sequence.
The 5Ј ends of both MEKK 2 and 3 are highly G/C-rich making DNA sequencing difficult. To verify the presence of stop codons in all three possible reading frames 5Ј to the predicted start site methionine, the MEKK 2 and 3 cDNAs were inserted in pRSET A, B, and C (Invitrogen) and expressed in Escherichia coli (not shown). Each construct gave a truncated RSET fragment confirming that the MEKK 2 and 3 cDNAs encoded 5Ј stop sites and that the isolated cDNAs encode fulllength proteins.
Alignment of the deduced amino acid sequences demonstrated significant homology between the two proteins ( Fig.  2A). Overall, the two proteins are approximately 77% homologous. The catalytic domain is encoded in the COOH-terminal moiety of both MEKK 2 and 3. The first consensus kinase domain (32) comprising the catalytic site of MEKK 2 and 3 begins at residues 361 and 367, respectively. The COOH-terminal catalytic domains of MEKK 2 and 3 are approximately 94% conserved, whereas the NH 2 -terminal moieties are only 65% conserved in amino acid sequence. These findings indicate that the primary sequences of MEKK 2 and 3 diverge significantly in the NH 2 -terminal half of the proteins. The conservation in sequence of the catalytic domains suggests they may recognize an overlapping set of substrates. The divergent NH 2 termini would be consistent with this region encoding se- FIG. 1. DNA and deduced amino acid sequences for MEKK 2 and 3. A, MEKK 2; B, MEKK 3. In-frame stop codons 5Ј to the predicted start methionine are underlined. quences for the differential regulation of the two proteins. Fig. 2B shows the alignment of MEKK 1, 2, and 3 catalytic domains. The 11 conserved subdomains comprising the protein kinase catalytic domain are designated by Roman numerals. The COOH terminus of MEKK 1 encoding the catalytic domain is only 50% homologous to the corresponding regions of MEKK 2 and 3. Thus, the catalytic domains of MEKK 2 and 3 are very similar to each other but significantly divergent from MEKK 1. As shown below, MEKK 1, 2, and 3 can all stimulate JNK and p42/44 MAPK activities in transfected cells. The significance of the sequence differences in the catalytic domains of MEKK 1, 2, and 3 is presently unclear. MEKK 2 and 3 Activate c-Jun Kinase and p42/44 MAPK Activity-Transient expression of MEKK 2 and 3 resulted in the stimulation of c-Jun kinase (JNK) activity (Fig. 3A). The JNK activity also eluted early from a Mono Q column using a linear sodium chloride elution gradient (Fig. 3B). Immunoblotting (not shown) demonstrated that this activity corresponded to the JNK/stress-activated protein kinase (1,8).

FIG. 2. Comparison of amino acid sequences for MEKK 2 and 3.
A, amino acids not having homology are boxed in the alignment of MEKK 2 and 3. B, alignment of the catalytic domains for MEKK 1, 2, and 3. Roman numerals indicate the 11 conserved regions within the protein kinase catalytic domain, with the most highly conserved residues underlined (32). Lowercase letters represent nonconserved amino acids in one or more of the MEKK sequences. and 3 were expressed at similar levels in HEK293 cells when 2 g of plasmid DNA was used per transfection (not shown). To determine whether MEKK 2 and 3 demonstrated selectivity in activating the JNK and p42/44 MAPK pathways, plasmid DNAs were titrated over a range of concentrations in the transfections. Fig. 5 shows that MEKK 2 has a greater selectivity for stimulation of the JNK pathway. In contrast, MEKK 3 had a greater selectivity for activating p42/44 MAPK relative to JNK. Thus, even though the kinase domains are approximately 94% conserved, MEKK 2 and 3 differ in their selectivity for regulation of the JNK and p42/44 MAPK pathways. This was particularly evident for MEKK 3 at low plasmid concentrations where the p42/44 MAPK pathway was preferentially activated. MEKK 2 Phosphorylates Both MEK 1 and JNK Kinase in Vitro-HEK293 cells expressing MEKK 2 and 3 were lysed, and the recombinant MEKK proteins were immunoprecipitated using the 12CA5 antibody recognizing the HA epitopetag. The immunoprecipitates were then used for in vitro kinase assays with recombinant purified MEK 1 and JNK kinase (JNKK) as substrates (Fig. 6A). MEKK 2 clearly phosphorylates both MEK 1 and JNKK consistent with its ability to activate JNK and p42/44 MAPK in HEK293 cells. Fig. 6B shows that the MEKK 2-catalyzed phosphorylation of recombinant JNKK resulted in the enhancement of JNKK activity. Thus, JNKK is a MEKK 2 substrate whose activity is stimulated both in vitro and in vivo by MEKK 2. We were unable to demonstrate the ability of MEKK 3 to phosphorylate MEK 1, MEK 2 (not shown), or JNKK in vitro using a variety of immunoprecipitation procedures. Although MEKK 3 was efficiently immunoprecipitated, as determined by Western blot analysis, it did not show measurable kinase activity toward MEK 1 or JNKK or show detectable autophosphorylation. This contrasted dramatically with the ability of MEKK 3 to activate both JNK and p42/44 MAPK in cells (Figs. 3-5). MEKK 3 protein was clearly immunoprecipitated using the 12CA5 antibody in these experiments, and a rabbit antisera raised against a keyhole limpet hemocyanin-conjugated peptide encoding the last 15 amino acids of MEKK 3 recognized the intact immunoprecipitated protein indicating that it was not degraded. The failure of immunoprecipitated MEKK 3 to phosphorylate recombinant MEK 1 or JNKK suggests one of three possibilities: (i) MEKK 3 is denatured but not degraded during immunoprecipitation, (ii) MEKK 3 requires an additional protein or co-factor for its activity in vitro that is lost during immunoprecipitation, (iii) the relevant substrate for MEKK 3 in cells is neither MEK 1 or 2 nor JNKK. At present, it is not clear which of these possibilities is responsible for the failure to detect MEKK 3 activity in vitro. We demonstrated that a mutant MEKK 3 having lysine 391 mutated to methionine, rendering it kinase-inactive (19,27,32), did not stimulate JNK or p42/44 MAPK activity when expressed in HEK293 cells (not shown). This finding indicated that the functional kinase activity of MEKK 3 was required for the in vivo regulation of JNK and p42/44 MAPK . MEKK 2 and 3 Do Not Regulate p38 Activity in HEK293 Cells-The p38 kinase is activated by hyperosmotic conditions (5) and recognizes the transcription factor ATF 2 as an in vitro substrate (33). Incubation of HEK293 cells with sorbitol activated p38 kinase (Fig. 7A). Immunoprecipitation and in vitro kinase assay of p38 from MEKK 2 and 3 transfected HEK293 cells indicated that neither MEKK 2 nor MEKK 3 stimulated p38 kinase activity (not shown). Mono Q FPLC fractionation of lysates from MEKK 2 or 3 transfected HEK293 cells confirmed that p38 kinase activity was similar to that from control transfected cells (Fig. 7B). ATF 2 is also a substrate for JNK (31). Fractions 2-8 from cells transfected with MEKK 2 or 3, that contain immunoreactive JNK, have increased kinase activity toward ATF 2. This is a predicted result based on the ability of both MEKK 2 and 3 to stimulate JNK activity in HEK293 cells. Expression of MEKK 2 and 3 also activated additional ATF 2 phosphorylating activities resolved by Mono Q fractionation. These activities are seen to elute in fractions 9 -12 and 13-18 for lysates from both MEKK 2 and 3 expressing cells. These activities do not correspond by immunoblotting to JNK, p42/ 44 MAPK , p38, or MEKK 2 or 3 and represent novel kinase activities capable of phosphorylating recombinant ATF 2 that are regulated by both MEKK 2 and 3.

DISCUSSION
The cloning and characterization of MEKK 2 and 3 define a family of MEKK proteins. MEKK 1, 2, and 3 are all capable of regulating both p42/44 MAPK and JNK activities. MEKK 1 and 2 appear to preferentially regulate the JNK pathway, whereas MEKK 3 shows a preference for activation of the p42/44 MAPK pathway in vivo. Cumulatively, our current and previous results (19) indicate that the different MEKKs when transiently expressed do not display a high selectivity for the p42/44 MAPK or JNK regulatory pathways. At more modest levels of expression, a greater degree of selectivity is observed for MEKK regulation of sequential protein kinase pathways. In MEKK 1-inducible clones of NIH3T3 (34) and Swiss 3T3 (35) cells, JNK is preferentially activated relative to p42/44 MAPK . MEKK 1, 2, and 3 do not measurably activate the p38 kinase pathway in these cell types.
Based on the biochemical characterization of MEKK proteins, it is evident that their activities are quite distinct from those of Raf-1 and B-Raf kinases. The Raf kinases selectively regulate MEK 1 and 2 and do not recognize the JNKK proteins (1,14,20,(33)(34)(35). Thus, Raf proteins which evolved in metazoan organisms appear to be highly selective for the regulation of p42/44 MAPK pathways (1,8). At present, no distinct pathways or substrates other than MEK and p42/44 MAPK have been defined for Raf kinases, although it is probable that additional MEK-like kinases will be identified in the future that serve as Raf substrates. MEKK proteins, in contrast, are capable of regulating both JNK and p42/44 MAPK pathways.
The ability of MEKKs to regulate multiple sequential protein kinase pathways in the cell suggests that a different mechanism exists for their regulation compared to the Raf kinases. The simplest prediction would be that MEKK proteins are selectively organized in "signalsome" complexes much like that for the mating pathway in S. cerevisiae. In this pathway, the protein kinases Ste11 (MEKK-like), Ste7 (MEK-like), and Fus3 (MAPK-like) are held together in a high affinity complex by Ste5 (36 -38). Ste5 functions as a scaffold to keep these proteins in an organized complex. Expression of gain-of-function Ste11 mutants can result in overcoming a threshold where Ste11 crosses over to activate another sequential protein kinase pathway such as that involved in morphogenesis (21)(22)(23). No Ste5 equivalent has yet been reported for mammalian cells and MEKKs. If such scaffold-like proteins do exist in mammalian cells and their expression is limiting, it would explain the ability of transiently expressed MEKKs to regulate both JNK and p42/44 MAPK pathways. It will obviously be necessary to define the organization of potential MEKK signalsome complexes in the cell and the constituent kinases in each complex. Forty-eight h post-transfection, cells were lysed, and MEKK 2 and 3 were immunoprecipitated using the 12CA5 monoclonal antibody. Immunoprecipitates were then assayed for kinase activity using recombinant kinase-inactive MEK 1 or JNKK as substrate. The JNKK only lane shows the low level of autophosphorylation of recombinant JNKK. Results are representative of 5-6 experiments for both MEKK 2 and 3. B, MEKK 2 phosphorylation of JNKK stimulates JNK kinase activity. MEKK 2 immunoprecipitates were incubated with the indicated combinations of wild-type or kinase-inactive JNKK and JNK as described under "Materials and Methods. Gst-c-Jun  phosphorylation by JNK was used as a measure for activation of the JNKK/JNK pathway.
Finally, the MEKK/JNK pathways can be activated by a diverse set of stimuli. These include cytokines such as TNF and IL-1 (39), low molecular weight GTP-binding regulatory proteins including Ras, Rac, and Cdc42 (40,41), high intracellular calcium (42), and stresses such as ultraviolet light, heat shock, osmotic imbalance, sphingomyelinase activity, protein synthesis inhibitors, etc. (1,8,13,14,20). Based on this array of stimuli capable of activating JNK, it is likely that several independent pathways converge to regulate the JNK sequential protein kinase pathway. It is possible that MEKK 1, 2, and 3 may all regulate the JNK pathway and each functions to respond to different upstream inputs. Alternatively, it is possible that MEKK 1, 2 and 3 are not only capable of regulating the JNK pathway but also other sequential protein kinase pathways as well. Such a mechanism would allow co-ordinate regulation of both the JNK pathway and additional parallel sequential protein kinase pathways in the cell. In support of this hypothesis, we have found that MEKK 1 also selectively regulates a protein kinase pathway leading to the phosphorylation and transactivation of c-Myc that is independent of JNK and c-Jun (35). The magnitude of a specific MEKK activation in response to a stimulus could selectively regulate different pathways such as those for JNK and c-Myc kinase activity by requiring different thresholds of MEKK activity to be obtained for stimulation of each pathway. The thresholds for MEKK regulation of these pathways might be regulated in part by the relative abundance or cellular localization of specific signalsome complexes. The cloning of MEKK 2 and 3 allows us to now address these potential regulatory mechanisms for the three MEKK proteins using both genetic and biochemical approaches.