Identification of c-Jun NH2-terminal Protein Kinase (JNK)-activating Kinase 2 as an Activator of JNK but Not p38*

c-Jun NH2-terminal protein kinase (JNK), a distant member of the mitogen-activated protein (MAP) kinase family, regulates gene expression in response to various extracellular stimuli. JNK is activated by JNK-activating kinase 1 (JNKK1), a dual specificity protein kinase that phosphorylates JNK on threonine 183 and tyrosine 185 residues. Here we show that JNKK2, a novel member of the MAP kinase kinase family, was phosphorylated and activated by MEKK1, a MAP kinase kinase kinase in the JNK signaling cascade. JNKK2 activity was also stimulated by constitutively active forms of Rac and Cdc42Hs, members of the Rho small GTP-binding protein family. Unlike JNKK1 that activates both JNK and p38 MAP kinases, JNKK2 stimulated only JNK. Transient transfection assays demonstrated that JNKK2 potentiated the stimulation of c-Jun transcriptional activity by MEKK1. The existence of multiple JNK-activating kinases may contribute to the specificity of the JNK signaling cascade.

A challenge in understanding the mammalian MAP kinase cascade is how signaling specificity is achieved. Despite the MAP kinase modules, cross-talk still exists. The cross-talk may allow the cells to coordinate the activity of different signaling cascades to produce a specific physiological response. However, it also makes the signaling cascades prone to lack of specificity. Other mechanisms are needed to ensure the specificity of each MAP kinase cascade, including subcellular localization, specific associated proteins, high enzymatic specificity, and selective responsiveness to extracellular stimuli.
The JNK cascade is activated by various stimuli such as growth factors, cytokines, tumor promoters, protein synthesis inhibitors, ultraviolet (UV) light irradiation, and oncogenes (1). JNK in turn stimulates the activity of several transcription factors including c-Jun, ATF-2, Elk, and Sap-1 (6 -8, 31-33). It is not completely understood how the specificity of the JNK cascade is maintained. One plausible mechanism is through the existence of multiple JNKKs that have high substrate specificity and respond to distinct upstream signals. To test this hypothesis, we have isolated a novel JNK-activating kinase, JNKK2, which is a close homologue of JNKK1. In contrast to JNKK1, which is selectively expressed in skeletal muscle and brain, JNKK2 is expressed in many tissues examined. Unlike JNKK1 which activates both JNK and p38, JNKK2 stimulated only JNK. Furthermore, JNKK2 did not respond to several extracellular stimuli that activate JNKK1. These data suggest that JNKK2 is a specific JNK activator and may contribute to the specificity of the JNK cascade.
cDNA Cloning-A human JNKK1 cDNA (20) was used as a probe to isolate JNKK1 homologues by screening a ZAPII HeLa cDNA library (Stratagene) at low stringency. Thirty positive clones were obtained after screening 2 ϫ 10 6 phage. These clones were grouped and sequenced on both strands by the dideoxy chain termination method using Sequenase Version II (U. S. Biochemical Corp.). Nucleotide sequence comparison of clones 4 and 23 with the GenBank TM data base reveals five sets of nucleotide sequences (GenBank TM accession numbers AA194047, AA019720, AA194193, AA258025, and AA252650) that are nearly identical. These clones were partially sequenced by the Washington University-Merck EST project. The clones AA194047 and AA019720 were requested and completely sequenced.
cDNA Constructs-The JNKK2 expression vector was constructed by inserting a polymerase chain reaction-generated NcoI-BglII fragment encoding JNKK2 into NcoI and BglII sites of pSR␣3-hemagglutinin (HA) vector (20). To construct pGEX-KG-JNKK2, the BglII site of the NcoI-BglII fragment of JNKK2 was blunted, and the fragment was inserted between the NcoI site and blunted HindII site of pGEX-KG vector. A Chameleon™ mutagenesis kit (Stratagene) was used to re-* This work was supported by National Institutes of Health Grant CA73740, American Heart Association Scientist Development Grant 9630261N, American Cancer Society Grant IRG66-37, and the start-up fund from the Department of Pathology, University of Alabama at Birmingham (to A. L.). 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 /EBI Data Bank with accession number(s) AF006689.
Transfections, Immunoprecipitation, and Protein Kinase Assay-HeLa cells were transiently transfected with expression vectors and harvested as described (20). HA-tagged or M2-Flag-tagged protein kinases were immunoprecipitated with specific antibodies for 3 h at 4°C. The activity of the immune complex was assayed at 30°C for 30 min in 30 l of kinase buffer (20) in the presence of 10 M [␥-32 P]ATP (10 Ci/mmol) with appropriate substrates, as indicated in the figure legends. The proteins were resolved by 13% SDS-polyacrylamide gel electrophoresis, followed by autoradiography. The phosphorylated proteins were quantitated by a phosphorImager.

RESULTS
Isolation of JNKK2 cDNA-Sequence analysis revealed that the JNKK2 clone encodes a complete open reading frame which appears to be a novel MAP kinase kinase (Fig. 1). The new MAP kinase kinase is closest to hep, a Drosophila JNK-activating kinase (36) (56.2% identity), followed by human JNKK1 (20) (42.4% identity), based on the comparison of all known MAP kinase kinases (PILE-UP program, Wisconsin Genetics Computer Group). Northern blot analysis revealed that unlike JNKK1, which is expressed mainly in skeletal muscle and brain, JNKK-2 is widely expressed in many tissues, with the highest level of expression in skeletal muscle (data not shown), suggesting that the two JNK activators may have different roles in different cells.
JNKK2 Can Be Activated by MEKK1, Rac, and Cdc42Hs-We and others have shown that MEKK1 acts as a MAP kinase kinase kinase which directly phosphorylates and activates JNKK1 (20,23,24). In addition, Rac and Cdc42Hs, members of the Rho small GTP-binding protein family, are both strong activators of JNKK1 and JNK (34,37,38). The effects of expression of MEKK1 and Rac/Cdc42Hs on JNKK2 activity were examined in transient transfection assays. Coexpression of the active forms of MEKK1, Rac1 and -2, and Cdc42Hs strongly activated JNKK2 (Fig. 2A). The active form of RhoA did not activate JNKK2 ( Fig. 2A) but was able to stimulate ERK activity (data not shown). MEKK1 also directly phosphorylated JNKK2 but not the JNKK2 (S272A/T276A) mutant, in which the putative activating phosphorylation residues Ser-272 and Thr-276 were replaced with alanines (Fig.  2B). These results demonstrate that JNKK2 is a JNK-activating kinase which acts downstream of MEKK1 and Rac/ Cdc42Hs in the JNK signaling pathway.
We also examined the regulation of JNKK2 by UV and the protein synthesis inhibitor anisomycin, both of which are strong activators of JNKK1 (Fig. 2C, top panel). UV and anisomycin only weakly stimulated JNKK2 activity (Fig. 2C, middle panel), even though JNK activity was stimulated manyfold in the same transfection assays (Fig. 2C, bottom panel). It is likely that JNKK2 may respond to upstream signals other than UV and anisomycin.
JNKK2 Is a Specific JNK-activating Kinase-MEKK1-activated HA-JNKK2 was isolated from transfected HeLa cells, and its substrate specificity was determined by protein kinase assays. JNKK2 efficiently phosphorylated GST-JNK1, but not GST-p38 or histidine-ERK2 (Fig. 3A). However, GST-p38 and histidine-ERK2 can be phosphorylated by their specific upstream kinases, MKK6 and MEK1(EE), respectively (Fig. 3A). The phosphorylation of JNK by JNKK2 is specific, since JNKK2 did not phosphorylate mutants JNK1 (APY) or JNK1 (APF), in which the activating phosphorylation residues Thr-183 and Tyr-185 were replaced by alanine and phenylalanine (Fig. 3B). Phosphorylation of JNK by JNKK2 in vitro led to JNK activation, as measured in a coupled kinase assay (Fig.  3C).
The effect of JNKK2 on JNK in vivo was examined in transient transfection assays. In HeLa cells, cotransfection of JNKK2 stimulated M2 Flag-tagged JNK1 (M2-JNK1) activity (Fig. 3D). The stimulation by JNKK2 was further potentiated by a suboptimal amount of cotransfected MEKK1 (Fig. 3D). Under the same conditions, JNKK2 or JNKK2 plus MEKK1 did not stimulate M2-p38 activity (Fig. 3D). These data demonstrate that JNKK2 is a specific activator of JNK in vivo.
JNKK2 Potentiates the Stimulation of c-Jun Transcriptional Activity by MEKK1-To determine the effect of JNKK2 on c-Jun transcriptional activity, HeLa cells were cotransfected with the active form of MEKK1 and GAL4-c-Jun fusion protein (20). MEKK1 stimulated GAL4-c-Jun activity 18-fold, as measured by a luciferase reporter gene driven by a GAL4 responsive promoter (Fig. 4). Coexpression of wild-type JNKK2 potentiated the stimulatory effect of MEKK1 on GAL4-c-Jun, resulting in 28-fold activation (Fig. 4). Neither MEKK1 alone nor MEKK1 plus JNKK2 stimulated the activity of GAL4-c-Jun Ala-63/73, in which both Ser-63 and Ser-73 have been replaced with alanines (Fig. 4). These results demonstrate that JNKK2 participates in stimulation of c-Jun transcription activity.

DISCUSSION
The specificity of the MAP kinase cascade is determined, in part, by enzymatic specificity of MAP kinase kinases in the MAP kinase module (2). JNKK1 was shown to be an efficient MAP kinase kinase of JNK in vitro (20 -22) and is required for JNK activation in vivo (39,40). However, JNKK1 is also involved in regulation of p38 (20,22). In contrast to JNKK1, JNKK2 is a highly selective kinase that phosphorylates and activates JNK but not p38 or ERK2 (Fig. 3A). The high substrate specificity of JNKK2 should contribute to the specificity of the JNK cascade.
The specificity of the JNK cascade may also be achieved by limiting the responsiveness of JNKKs to a certain subset of upstream stimuli. It has been shown that various extracellular stimuli and proto-oncogenes stimulate JNK activity, some of which do so through activating JNKK1. JNKK2 can be significantly activated by MEKK1, Rac, and Cdc42Hs, which are activators of JNKK1 ( Fig. 2A). However, JNKK2 was only weakly activated by UV and anisomycin (Fig. 2C), both of which strongly activate JNKK1 (Fig. 2C). Osmotic stress stimulates JNKK1, but only weakly stimulated JNKK2 activity (data not shown). The differential response of JNKK1 and JNKK2 to extracellular stimuli suggests that signaling leading to JNK activation may diverge upstream of JNK at the level of JNKKs. It is possible that JNKK2 may mediate the activation of JNK by extracellular stimuli that stimulate only JNK but not p38. Identification of extracellular stimuli that specifically activate JNKK2 should provide insights into the specificity of the JNK signaling pathway. The transfected kinases were immunoprecipitated, and their activities were determined by kinase assays with GST-JNK1, GST-p38, and histidine-ERK2 as substrates, as indicated. B, MEKK1-activated HA-JNKK2 was isolated as described in A, and its activity was measured by kinase assays with GST-JNK1 wild-type (WT), GST-JNK1 (APY), and GST-JNK1 (APF) mutants as substrates. C, MEKK1-activated HA-JNKK2 was isolated as described in A, and its activity was measured in a coupled kinase assay with GST-c-Jun as a substrate. D, HeLa cells were transfected with expression vectors encoding M2-JNK1 and M2-p38 (1 g each), HA-JNKK2 (2 and 3 g), and MEKK (20 ng), as indicated. M2-JNK1 or M2-p38 were immunoprecipitated with anti-M2-Flag antibody, and their activities were measured by kinase assays with GST-c-Jun or GST-ATF2 as substrates, respectively. Phosphorylated GST-c-Jun shows a slower migration in the SDS gel, as indicated. . Luciferase activity was determined as described (20). The averages of three experiments are shown. Luciferase activity expressed by cells transfected with pSR␣ was given an arbitrary value of 1.