(Received for publication, December 4, 1996)
From the Division of Basic Sciences and Program in
Molecular Signal Transduction, National Jewish Center for
Immunology and Respiratory Medicine, Denver, Colorado 80206 and the
Department of Pharmacology, University of Colorado Medical
School, Denver, Colorado 80206
Mitogen-activated protein kinases (MAPKs) are components of sequential kinase cascades that are activated in response to a variety of extracellular signals. Members of the MAPK family include the extracellular response kinases (ERKs or p42/44MAPK), the c-Jun amino-terminal kinases (JNKs), and the p38/Hog 1 protein kinases. MAPKs are phosphorylated and activated by MAPK kinases (MKKs or MEKs), which in turn are phosphorylated and activated by MKK/MEK kinases (Raf and MKKK/MEKKs). We have isolated two cDNAs encoding splice variants of a novel MEK kinase, MEKK4. The MEKK4 mRNA is widely expressed in mouse tissues and encodes for a protein of approximately 180 kDa. The MEKK4 carboxyl-terminal catalytic domain is approximately 55% homologous to the catalytic domains of MEKKs 1, 2, and 3. The amino-terminal region of MEKK4 has little sequence homology to the previously cloned MEKK proteins. MEKK4 specifically activates the JNK pathway but not ERKs or p38, distinguishing it from MEKKs 1, 2 and 3, which are capable of activating the ERK pathway. MEKK4 is localized in a perinuclear, vesicular compartment similar to the Golgi. MEKK4 binds to Cdc42 and Rac; kinase-inactive mutants of MEKK4 block Cdc42/Rac stimulation of the JNK pathway. MEKK4 has a putative pleckstrin homology domain and a proline-rich motif, suggesting specific regulatory functions different from those of the previously characterized MEKKs.
Eukaryotic cells have developed specific signal transduction pathways for response to and integration of extracellular stimuli. Mitogen-activated protein kinases (MAPKs)1 represent a family of kinases that respond to diverse stimuli and are composed of sequential protein kinase cascades (1, 2). MAPKs are activated through phosphorylation of a specific threonine and tyrosine by dual specificity MAPK kinases referred to as MKKs or MEKs. MKK/MEKs are phosphorylated and activated by MKK/MEK kinases (MKKK/MEKKs). Homologous kinases in several sequential protein kinase cascades have been identified in yeast and mammalian cells, indicating conserved MAPK modules for signal transduction in eukaryotes (3, 4). There are three well defined MAPK pathways: ERK1/ERK2, also referred to as p42/p44 MAPKs (5, 6); the p38/HOG1 kinases (7-9); and the c-Jun NH2-terminal kinases/stress-activated protein kinases (JNK/SAPKs) (10-13). In addition, ERK3, which shares 50% sequence homology with ERK1/ERK2 (5), and MAPK5 (14) have been cloned but not characterized in terms of regulation and substrate recognition.
Activation of growth factor receptor tyrosine kinases, heterotrimeric
G-protein coupled receptors and specific cytokine receptors activate
the ERKs (1, 15). The p38 protein kinases (p38 and p38) are
activated by exposure of cells to lipopolysaccharide, proinflammatory
cytokines, and cellular stresses such as osmotic imbalance (7-9,
16-18). JNKs include three different isoforms with 2-4 splice
variants of each (10-13). JNKs are activated by diverse stimuli,
including cellular stresses such as UV and protein synthesis
inhibitors, proinflammatory cytokines, and G-protein coupled and
tyrosine kinase growth factor receptors (2, 10-13, 19-21).
Each MAPK group is phosphorylated and activated by specific MKKs/MEKs.
MEK1 and MEK2 activate ERKs (22-24), MKK3 and MKK6 activate p38
kinases (25-28), and MKK4 activates JNKs (26, 29). MEK5 is presumed to
activate MAPK5, but this has not been demonstrated biochemically (14,
30). The activators of MKK/MEKs comprise a group of diverse kinases
allowing integration of upstream inputs to regulate MAPK pathways. Raf
phosphorylates and activates MEK1/MEK2 (31, 32). MEKKs 1, 2, and 3 (33-35) and tumor progression locus 2 (Tpl-2) (36) activate both the
ERKs and JNKs, and immunoprecipitates of these kinases have been shown
to phosphorylate and activate MEK1/MEK2 and MKK4. Germinal center
kinase (GCK) (37), the mixed lineage kinases (MLK) MLK3/SPRK, DLK/MUK
(38-40), and TGF--activated protein kinase (TAK1) (41) show
selectivity for activation of the JNK pathway, and immunoprecipitates
of these kinases from transiently transfected cells stimulate the
phosphorylation of MKK4 in vitro. The multiple kinases that
can apparently function as MKKK/MEKKs for the JNK pathway indicates
that many different extracellular stimuli will regulate this pathway,
emphasizing its importance in the control of cell function. In this
report, we present the cloning of a novel MEK kinase, MEKK4. The
properties of MEKK4 differ from those of MEKKs 1, 2, and 3 in that it
is highly selective for the activation of the MKK4/JNK pathway. MEKK4 has interesting structural motifs suggesting regulatory mechanisms for
the control of MEKK4 that are not encoded in some of the previously characterized MEKKs.
The degenerate primers 5-GA(A
or G)(C or T)TIATGGCIGTIAA(A or G)CA-3
(sense) and 5
-TTIGCICC(T or
C)TTIAT(A or G)TCIC(G or T)(A or G)TG-3
(antisense) were used in a
polymerase chain reaction (PCR) using first strand cDNA generated
from polyadenylated RNA prepared from NIH 3T3 cells as template.
Taq DNA polymerase (Boehringer Mannheim) was used in a PCR
of 30 cycles (1 min, 94 °C; 2 min, 37 °C; 3 min, 72 °C),
followed by a 10-min cycle at 72 °C. A band of approximately 300 bp
was recovered from the PCR mixture, and the products were cloned into
pGEM-T (Promega). The PCR cDNA products were sequenced and compared
with the MEKKs 1, 2, and 3 sequences. A unique cDNA sequence of 262 bp, having significant homology to the catalytic domains of MEKKs 1, 2, and 3, was identified and used to screen cDNA libraries from mouse
brain, liver, and NIH 3T3 cells (Stratagene). The
phage libraries
were plated, and DNA from plaques were transferred to Hybond-N filters
(Amersham Corp.) followed by UV cross-linking of DNA to the filters.
Filters were prehybridized for 2 h and then hybridized overnight
in 0.5 M Na2H2PO4, pH
7.2, 10% bovine serum albumin, 1 mM EDTA, 7% SDS at
68 °C. Filters were washed 2 times at 42 °C with 2 × SSC,
once with 1 × SSC, and once with 0.5 × SSC containing 0.1%
SDS (1 × 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 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. To isolate the 5
end
of the gene, we utilized the rapid amplification of cDNA ends
(RACE) procedure using the mouse brain Marathon-Ready cDNA kit
(Clontech). The first PCR was performed using the Vent exo(-) DNA
polymerase (New England Biolabs) and the sense anchor primer supplied
with the kit in combination with the MEKK4-specific antisense primer
5
-AGTCCAACATGAATGAGCACTGTGCAT-3
. The PCR reaction involved 25 cycles
(1.5 min, 99 °C; 1 min, 68 °C; 2 min, 74 °C) in a final volume
of 50 µl. A second PCR was performed using 5 µl of the reaction
mixture from the first PCR as template with the nested anchor primer
supplied with the kit in combination with the nested MEKK4-specific
antisense primer 5
-GCTCCGTTGTTCTCAGAGTTGCTCGAA-3
. The second PCR
generated a 650-bp fragment that was subcloned into pGEM-T and 21 clones sequenced. The identity of the RACE product was confirmed by PCR
using the sense primer 5
-AAAATCTAGACCTGCGGCGGGCTAGAGGCGGAGG-3
that encodes the extreme 5
end of the RACE product and the antisense primer 5
-GCTCCCGTAGTTAACTTTGAAGGTGA-3
that is based on the cDNA encoded in the original clone that was isolated from the brain cDNA
library. Vent exo(-) DNA polymerase was used in a PCR consisting of 30 cycles (1.5 min, 99 °C; 1.5 min, 66 °C; 3 min, 74 °C) in a
final volume of 50 µl with first strand cDNA generated from mRNA isolated from NIH 3T3 cells as template. A second PCR was performed using 2 µl of the first reaction as a template and the nested MEKK4-specific antisense primer
5
-CTGGAATCGATTTTTTTGGCAAAGACC-3
that is also encoded in the
original brain cDNA clone. PCR conditions and the sense primer were
the same as in the first PCR reaction. The resulting 645-bp PCR product
was purified, cloned into pGEM-T, and five cDNAs from each of two
separate PCRs were sequenced, and all confirmed the 5
sequence of
MEKK4 that was obtained through the RACE procedure.
Poly(A)+ RNA (2 µg)
from eight different mouse tissues were separated by denaturing
formaldehyde, 1.2% agarose gel electrophoresis, transferred to a
charge-modified nylon membrane by Northern blotting, and fixed by UV
irradiation (mouse multiple Northern blot, Clontech). The membrane was
hybridized with either a 300-bp cDNA fragment derived from the
catalytic domain of MEKK4 (recognizes both splice forms of MEKK4) or a
-actin control probe. The probes were randomly primed and labeled
with [
32P]dCTP (Prime it II, Stratagene), and
hybridization was performed as described for screening of cDNA
libraries.
All cDNAs encoding MEKK4 used
in transient transfection assays were subcloned into the mammalian
expression plasmid pCMV5. Where indicated, MEKK4 and
MEKK4
kin
were epitope-tagged using a PCR at their carboxyl
terminus with the hemagglutinin (HA)-tag sequence YPYDVPDYA. The
insertion of a carboxyl-terminal epitope tag was performed using the
sense oligonucleotide 5
-TCTAAGCAGGGGCCCATAGAAGCTATC-3
, which encodes an ApaI restriction site encoded in MEKK4, and the
antisense oligonucleotide 5
-CTCTCTAGAGGTACCTCATTAAGCATAATCTGGAACATCATATGGATACTCTTCATCTGTGCAAACCTTGAC-3
, which encoded an XbaI restriction site, two sets of
termination codons, the HA-tag sequence, and the MEKK4 sequence. These
primers were used in a PCR using Vent exo(-) DNA polymerase with
MEKK4
or MEKK4
kin
(see below) as templates. The
PCR reaction involved 30 cycles (1 min, 99 °C; 2 min, 68 °C; 3 min, 74 °C). The PCR products were purified and digested with
ApaI, blunt ended, and then digested with XbaI
and cloned into pCMV5. This gives a catalytic domain of MEKK4 with an
initiation methionine at position 1301. The sequences were confirmed by
DNA sequencing.
Kinase inactive variants of MEKK4 were generated by changing lysine at
position 1361 to a methionine using the Altered Sites II in
vitro Mutagenesis Systems (Promega) with the mutagenisis oligonucleotide 5-ACAGGGGAGCTGATGGCCATGATGGAGATTCGATTTCAGCCTAAC-3
. The mutation was confirmed by DNA sequencing.
The MEKK4 catalytic domain used in polyhistidine tagged or GST-fusion
proteins was generated by PCR using the sense primer 5-AAGCTTGGATCCGAATTCAGGAGAAAGAATATCATCGGCCAA-3
, which encodes EcoRI and BamHI restriction sites, followed by
MEKK4 sequence starting at amino acid 1302 in combination with the T7
primer encoded in pBluescript II SK-. MEKK4
in
pBluescript II SK- was used as template for a PCR of 30 cycles (1 min, 92 °C; 2 min, 58 °C; 3 min, 72 °C) using
Taq DNA polymerase (Boehringer Mannheim). The PCR product
was digested with restriction endonucleases directed toward sites
encoded in the sense primer and in the multiple cloning site of
pBluescript SK II and cloned into pRSET (BamHI and
XhoI; New England Biolabs) or pGEX (EcoRI and
XhoI; Pharmacia Biotech In.). Growth, induction, and
purification was performed according to the manufacturer
instructions.
Wild type and kinase inactive (K1361M), non-epitope tagged, catalytic domains of MEKK4 were constructed using the same procedure as for the fusion protein constructs except that cDNAs were cloned into pCMV5.
The plasmids encoding HA-tagged JNK1 (pSR) and JNK2 (pCMV5) have
been described previously (10, 11). Rac (Q61L) and Cdcd42 (Q61L) were
in pCMV5.
HEK293 cells were maintained in Dulbecco's modified Eagle's medium that contained 10% bovine calf serum, 100 units/ml penicillin, and 100 µg/ml streptomycin at 37 °C in 5% CO2, 95% air. COS cells were maintained under the same conditions except that 5% bovine calf and 5% newborn calf serum was used. Cells were grown and transfected in 100-mm tissue culture dishes and transfected using the calcium phosphate (HEK293) or DEAE-dextran (COS) methods (42).
Antibody ProductionPeptides corresponding to the COOH-terminal sequences of MEKK4 (CLESDPKIRWTASQLLD) and p38 (CFVPPPLDQEEMES) were conjugated to KLH and used to immunize rabbits. Antisera were characterized for specificity by immunoblotting of lysates prepared from appropriately transfected HEK293 cells.
Assay of JNK ActivityJNK activity was measured by
immunoprecipitation of endogenous JNK1 and JNK2 using specific
antibodies (Santa Cruz Biotechnology), transiently overexpressed
HA-tagged JNK1 and JNK2 using a monoclonal antibody directed against
the hemagglutinin epitope (the 12CA5 antibody from Berkeley Antibody
Co.), or by a solid phase assay using glutathione
S-transferase (GST)-c-Jun (1-79) coupled to
glutathione-Sepharose-4B. Transfected cells were lysed in 0.5% Nonidet
P-40, 1% Triton X-100, 20 mM Tris-HCl, 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 µg/ml aprotinin, and 5 µg/ml leupeptin. Nuclei were removed by centrifugation at 15,000 × g for 10 min, and the supernatants were used for
immunoprecipitation of JNKs. For immunoprecipitations, 400 µg of
protein was mixed with the appropriate antibody (1:100 dilution for
endogenous JNK1 and JNK2 antibodies and 1:250 dilution for the 12CA5
antibody) in a final volume of 400 µl of lysis buffer and rotated at
4 °C for 1 h. Immune complexes were captured by adding 15 µl
of a 1:1 slurry of protein A-Sepharose (Sigma). The
mixture was rotated at 4 °C for an additional hr and washed 2 times
in lysis buffer and once in kinase buffer (20 mM Hepes, pH
7.5, 10 mM MgCl2, 20 mM
-glycerophosphate, 10 mM p-nitrophenyl
phosphate, 1 mM dithiothreitol, 50 µM sodium
vanadate). Beads were suspended in 40 µl of kinase buffer containing
10 µCi of [
32P]ATP and 2 µg of GST-c-Jun as
substrate. For solid phase JNK assays, 400 µg of protein was mixed
with 10 µl of a 1:1 slurry of GST-c-Jun-(1-79)-Sepharose (3-5 µg
of GST-c-Jun-(1-79)). The mixture was rotated at 4 °C for 1 h
and washed as described above, and kinase reactions were performed by
adding 10 µCi of [
32P]ATP. Kinase reactions were
performed at 30 °C for 20 min, and reactions were terminated by
addition of Laemmli sample buffer. Samples were boiled, and
phosphorylated proteins were resolved on SDS, 10% polyacrylamide gels
and visualized by autoradiography.
In
vitro activation of JNK was determined essentially as described
(33). HEK293 cells were transfected with the expression vector pCMV5 or
pCMV5 containing HA-tagged wild type or kinase inactive MEKK4. Cells
were lysed, and
MEKK4 was immunoprecipitated using the 12CA5
antibody as described for the JNK assay above. In addition,
recombinant, bacterially expressed, and purified polyhistidine-tagged
MEKK4 was used in an in vitro reconstituted coupled
kinase assay, where immunoprecipitates or recombinant
MEKK4 were
mixed with recombinant, bacterially expressed and purified, wild type
or kinase inactive (K116M) MKK4 in combination with wild type or kinase
inactive (K55M) JNK in the presence of 50 µM ATP (40 µl
final volume) and incubated at 30 °C for 20 min. GST-c-Jun-(1-79)-Sepharose beads (3-5 µg) were then added, and samples were rotated at 4 °C for 20 min. Beads were washed,
resuspended in 40 µl JNK kinase buffer containing 10 µCi of
[
32P]ATP, and incubated for 20 min at 30 °C. Direct
phosphorylation of MKK4 was demonstrated by incubating recombinant
purified GST-
MEKK4 and kinase inactive (K116M) MKK4 together with 10 µCi of [
32P]ATP in 40 µl of kinase buffer (20 mM Pipes, pH 7.0, 10 mM MnCl2, 20 µg/ml aprotinin) for 20 min at 30 °C. Reaction mixtures were added
to Laemmli sample buffer and boiled, proteins were resolved on SDS,
10% polyacrylamide gels, and phosphorylated proteins were visualized
by autoradiography.
HEK293 cells transiently transfected with empty vector (pCMV5), the indicated MEKK4 constructs, or treated with EGF (cells starved in 0.1% bovine serum albumin overnight and then activated with 30 ng/ml of EGF for 10 min) were lysed, and ERK1 and ERK2 were immunoprecipitated using specific antibodies (1:125 dilution; Santa Cruz Biotechnology) as described above for immunoprecipitation of JNK. Kinase reactions were performed as described (43) using the EGF662-681 receptor peptide as a substrate. The amounts of 32P incorporated into EGF662-681 receptor peptide was quantitated by spotting samples on P81 phosphocellulose paper followed by scintillation counting.
Assay of p38 Kinase ActivityHEK293 cells transiently transfected with empty vector (pCMV5), the indicated MEKK4 constructs, or cells treated with sorbitol (0.4 M, 20 min) were lysed, and p38 was immunoprecipitated from cell lysates (400 µg) as described for the JNK assay using rabbit antiserum (1:250 dilution) raised against the COOH-terminal peptide sequence of p38 (described above). In vitro kinase assays were carried out as described for the JNK assay with the exception that recombinant ATF-2 was used as a substrate. Proteins were separated through SDS, 10% polyacrylamide gel electrophoresis, and phosphorylated ATF-2 was visualized by autoradiography.
Binding of Rac and Cdc42 to MEKK4In vitro
binding was performed as described (44). COS cells were transiently
transfected with MEKK4 and lysed in extraction buffer (10 mM Tris-HCl, pH 7.4, 1% Triton X-100, 5 mM
EDTA, 50 mM NaCl, 50 mM NaF, 20 µg/ml
aprotinin, 1 mM phenylmethylsulfonyl fluoride, 2 mM Na3VO4) and 200 µg of lysates
incubated with 3-5 µg of GST, GST-Rac, or GST-Cdc42 fusion proteins
(coupled to Sepharose beads). The GST-fusion proteins were preloaded
with either GDP or GTP-
-S through incubation with 1 mM
GDP or GTP-
-S in PBS at 30 °C for 20 min followed by addition of
12 mM MgCl2 to terminate loading reactions.
Binding was performed for 2 h at 4 °C in a final volume of 0.5 ml. Beads were then washed 3 times in extraction buffer and resuspended
in sample buffer, and proteins were separated by SDS, 10% PAGE. MEKK4
proteins were visualized through Western blotting using a primary
rabbit anti-MEKK4 antibody (described above) and horseradish
peroxidase-linked protein A followed by chemiluminescence (DuPont
NEN).
T47D human breast carcinoma cells or HEK293 cells on glass coverslips were fixed in 3% paraformaldehyde in PBS. Cells were permeabilized in 0.2% Triton X-100 in PBS and incubated for 10 min in Dulbecco's modified Eagle's medium, 10% calf serum. The rabbit antibody raised against the MEKK4 COOH terminus was incubated with the cells for 1 h, and cells were then washed extensively in PBS. A Cy3 donkey anti-rabbit antibody (Jackson Immunological Laboratories) was then used as a secondary antibody for visualization. The MEKK4 anti-COOH-terminal peptide antibody has been extensively characterized and shown not to recognize other MEKK proteins. Characterization of the subcellular localization of the other MEKK proteins is being described elsewhere.2 Bodipy-ceramide bound to bovine serum albumin was incubated with T47D cells for 15 min prior to fixation for specific staining of the Golgi (45).
Degenerate primers were used in a
PCR with cDNA, synthesized from RNA isolated from NIH 3T3 cells, as
a template. Of 185 PCR products that were sequenced, 155 encoded MEKK1,
15 encoded MEKK2, and 5 encoded a novel MEKK-like sequence. Using this
novel MEKK cDNA fragment, cDNA libraries from mouse brain,
liver, and NIH 3T3 cells were screened. A 5.1-kb cDNA was isolated
from the brain library that contained a polyadenylated 3 tail and
encoded a 1536-amino acid open reading frame. No 5
in-frame stop sites were apparent, indicating that the full coding sequence was not within
the 5.1-kb cDNA. The RACE procedure was used to isolate the
remainder of the cDNA clone with a 5
in-frame stop codon. The
identity of the 5
end was confirmed by PCR with oligonucleotide primers that were designed so that the sense primer started at the
extreme 5
end of the RACE product and the antisense primer started in
the original clone obtained from cDNA library screens and hence
should cover the junction between the original clone and the RACE
product. PCR, using these primers, and cDNA, prepared from NIH 3T3
cells, gave a cDNA fragment that when sequenced confirmed the
junction between the original clone and the RACE product. The isolated
cDNA encoding MEKK4 is 5426 base pairs, which encodes a 1597-amino
acid protein with a deduced molecular mass of 180 kDa (Fig.
1A).
An alternative splice form was also isolated from the brain library
with a deletion of 52 amino acids between residues 1162 and 1213 in the
reported sequence (Fig. 1A). The alternatively spliced
mRNAs were confirmed by reverse transcriptase-PCR using oligonucleotides flanking the splice region and cDNA prepared from
mRNA from PC12 cells. The cDNAs for both splice forms could be
detected; the larger variant expressing the additional 52 residues has
been designated MEKK4 and the smaller alternatively spliced form
without the 52 amino acids as MEKK4
. Northern blot analysis revealed
an mRNA for MEKK4 of approximately 6 kb that is expressed in
several different mouse tissues (Fig. 2).
The COOH-terminal moiety of MEKK4 contains the 11 consensus subdomains for the catalytic domain of protein kinases (46) and has approximately 55% amino acid homology with MEKKs 1, 2, and 3 (Fig. 1B). The NH2-terminal region of MEKK4 shares little or no homology with MEKKs 1, 2, or 3, but several sequences within this moiety suggests specific regulatory functions (Fig. 1A). Near the NH2 terminus is a proline-rich region suggesting a possible interaction with SH3 domain encoded proteins (47-49). A predicted pleckstrin homology (PH) domain (50, 51) is also encoded in the MEKK4 NH2-terminal region. In addition, a modified consensus Cdc42/Rac interactive binding (CRIB) domain (52) is encoded in the MEKK4 sequence just upstream of the catalytic domain.
MEKK4 Selectively Activates the MKK4/JNK PathwayCharacterization of MEKK4-mediated regulation of the JNK
pathway was performed by transient transfection of HEK293 cells. Endogenous JNK1 and JNK2 and transfected HA-tagged JNK1 and JNK2 were
characterized for their regulation in response to expression of
full-length and the truncated catalytic domains of MEKK4 using immunoprecipitation of JNKs followed by an in vitro kinase
assay using GST-c-Jun as a substrate (Fig. 3).
Expression of the catalytic domain of MEKK4 (MEKK4) resulted in the
activation of the endogenous JNK1 and JNK2 proteins. The mutant kinase
inactive
MEKK4 (
MEKK4 kin
) did not increase JNK
activity. To verify the regulation of JNK1 and JNK2 by
MEKK4,
HA-tagged forms of JNK1 and JNK2 were coexpressed with
MEKK4 and
selectively immunoprecipitated from HEK293 cell lysates. Similar to the
regulation of total endogenous JNK activity, measured by the binding to
GST-c-Jun beads or immunoprecipitation with specific JNK antibodies,
the HA-tagged JNKs were activated in cells expressing
MEKK4 (Fig.
3). Expression of either full-length MEKK4 splice variant did not
significantly activate JNK. The transfected HA-tagged JNKs show a very
modest stimulation relative to endogenous JNK proteins when the
full-length MEKK4 is expressed, probably related to the amplification
of the signal resulting from overexpression of the kinases. Using
ultraviolet irradiation as a control for stimulating JNK activity, it
is clear that
MEKK4 is a constitutively active mutant and that the
full-length MEKK4 does not show significant activity when transiently
overexpressed in HEK293 cells. This finding suggests that full-length
MEKK4 is most likely regulated by several upstream effectors.
To verify that MEKK4 directly regulated the MKK4/JNK pathway, the
sequential kinase cascade was reconstituted in vitro (Fig. 4). Immunoprecipitates of HA-tagged MEKK4 from HEK293
cell lysates or bacterially expressed recombinant
MEKK4 were mixed
with combinations of wild type and kinase inactive mutants of MKK4 and
JNK. JNK activation was assayed by capture of JNKs onto GST-c-Jun
beads, followed by a kinase assay using [
32P]ATP.
MEKK4 expressed in HEK293 cells or recombinantly in
Escherichia coli required functional MKK4 and JNK proteins
for stimulation of GST-c-Jun phosphorylation (Fig. 4A).
Recombinant
MEKK4 also directly phosphorylated kinase inactive MKK4
(Fig. 4B). Thus, MEKK4 phosphorylates and activates MKK4,
which regulates JNK activity both in vitro and in
vivo. Expression of
MEKK4 in HEK293 cells neither activates the
ERK1/ERK2 pathway nor the p38 pathway (Fig. 5,
A and B). The selective MEKK4-induced activation
of the JNK pathway distinguishes it from MEKKs 1, 2, and 3, which are
able to activate the ERK pathway when overexpressed in different cell types (33-35).
MEKK4 Interacts with Cdc42 and Rac
Cdc42 and Rac are
GTP-binding proteins of the Rho superfamily; they are well
characterized for their ability to regulate cytoskeletal functions,
including the formation of filopodia and lamellopodia (53). Cdc42 and
Rac have also been shown to regulate pathways leading to the activation
of JNKs (54-56). Recently, a CRIB motif was proposed having the
sequence ISXP(X2-4)FXH(X2)HVG (52). A
related sequence in MEKK4 is
CDTKSDNVM (identical or
conserved residues underlined within residues 1311-1324, Fig.
1A). Fig. 6A shows that MEKK4
binds GST-Cdc42 and GST-Rac but not GST alone. The binding of
MEKK4
to Cdc42 was significantly GTP-dependent. In contrast, the
binding of
MEKK4 to GST-Rac was significant when GDP was bound. We
have found that GST-Rac has a lower specific activity for GDP/GTP
binding relative to Cdc42 (not shown), suggesting not all the GST-Rac
is functional. This result may be in part due to altered properties of
Rac when fused to GST. Fig. 6B shows that a kinase inactive
mutant of MEKK4
inhibits activated Rac and Cdc42 stimulation of JNK
activity. Combined results indicate that
MEKK4 interacts with Cdc42
and probably Rac in a GTP-dependent manner and activates
the JNK pathway. We have been unable to express enough recombinant
MEKK4 protein in E. coli with sufficient native folding
and specific activity to perform similar binding studies as was done
with COS cell-expressed
MEKK4. Hence, the possibility exists that
the interaction of Cdc42/Rac with MEKK4 involves additional proteins.
Further analysis of the interactions of MEKK4 and Cdc42/Rac will
require expression of high specific activity
MEKK4 using Sf9/baculovirus or other systems for protein purification.
MEKK4 Is Localized in a Perinuclear Vesicular Structure
Immunostaining of MEKK4 demonstrates that it is
localized in perinuclear vesicular-like structures (Fig.
7). Staining with ceramide, a marker that stains Golgi
and Golgi-derived vesicles shows an overlap in staining with MEKK4.
Thus, MEKK4 appears associated with Golgi-associated vesicles.
Interestingly, Cdc42 was recently shown to associate with Golgi
vesicles (57). Thus MEKK4 and Cdc42 are in a common location where
interaction and regulation might occur.
A number of MKKK/MEKK-like kinases have been postulated to
regulate the JNK pathway. These include the MEKKs, GCK, MLK3/SPRK, DLK/MUK, PAKs, TAK1, and Tpl-2 (33-41, 58-60). Immunoprecipitates of
these kinases (with the exception of MEKK 3 and PAKs) from transiently
transfected cells have all been shown capable of phosphorylating and
activating MKK4. This suggests that either many different kinases at
the level of MKK4 are capable of regulating MKK4 and the JNK pathway or
that the transient transfection and in vitro kinase analysis
may not maintain the fidelity of regulation observed in oligomeric
assemblies of these kinase modules. To sort out these potential
multiple regulatory inputs versus "cross talk" of
normally parallel pathways, the analysis of the endogenous proteins and
their regulation in addition to gene inactivation studies will probably
be required. It is also probable that additional MKK genes are yet to
be identified. It should be noted that only MEKK4 of all these
kinases has been expressed as a recombinant protein in E. coli and shown to directly phosphorylate and activate MKK4.
Recombinant forms of the other kinases will have to be used to define
if they are capable of directly phosphorylating and activating MKK4. In
this regard, it has been suggested that PAKs are upstream of MEKKs in
the activation of the JNK pathway (58-60). Our unpublished
observations suggest that PAKs are poor activators of the JNK pathway
and do not recognize MEKKs or MKK4 as substrates, suggesting their
actions may be indirect.
MEKK4 binds to Cdc42/Rac, and a kinase inactive mutant of MEKK4
inhibits the ability of GTPase deficient, activated, Cdc42 and Rac to
stimulate the JNK pathway. Thus, MEKK4 is a strong candidate for being
a Cdc42/Rac-regulated MEKK. Testing of this hypothesis will require the
generation of antibodies that efficiently immunoprecipitate the
full-length MEKK4 protein from cell lysates. These antibodies are
currently being generated using fusion proteins as antigens. The fact
that MEKK4 has a potential CRIB-like domain and that GST-Cdc42 binds
MEKK4 encoding this sequence in a GTP-dependent fashion
suggests MEKK4 is in the JNK pathway regulated by Cdc42 and Rac.
MLK3/SPRK was recently demonstrated to be regulated by Cdc42 and Rac and to selectively activate the JNK pathway similar to MEKK4 (38, 39). For MLK3, six of the eight consensus CRIB domain residues are conserved, whereas five of the eight conserved residues are found in MEKK4. The other signature sequences for MLK3 and MEKK4 are different. MLK3 has a leucine zipper-like sequence suggesting it may form homo- or hetero-dimers with other proteins, MEKK4 has a putative pleckstrin homology domain, and both MLK3 and MEKK4 have proline rich sequences that may be involved in SH3 domain or other protein-protein interactions. This suggests that MLK3 and MEKK4 are regulated by different upstream inputs and potentially different extracellular stimuli even though they may both interact with Cdc42/Rac GTP-binding proteins.
Our analysis of the four MEKK proteins begins to suggest one regulatory scenario for having multiple MKKKs regulating the JNK pathway. Subcellular localization studies using antibodies to the COOH terminus of MEKKs 1, 2, 3, and 4 indicate they are localized in different places in the cell.2 MEKK4 is shown to be in a perinuclear, Golgi-like localization, whereas MEKK1 is in a post-Golgi vesicle-like compartment. Differential subcellular localization suggests different regulation of the MEKKs by different extracellular stimuli and intracellular signal transduction proteins. If this is combined with localized or differential regulation of JNKs, then each MEKK could indeed have very specific regulatory functions that are not obvious by transient transfection analysis. We are currently pursuing activation, subcellular localization, and the potential redistribution of the MEKKs and other MKKKs to define their respective roles in the control of cell function and responses to extracellular stimuli.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) U85607[GenBank], for MEKK4, and U85608[GenBank], for MEKK4
.