Biochimica et Biophysica Acta 1843 (2014) 875–884

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta

j ourna l homepage: www.e lsev ie r .com/ locate /bbamcr

Map2k4δ — Identification and functional characterization of a novelMap2k4 splice variant

Wiebke Haeusgen, Leif Tueffers, Thomas Herdegen, Vicki Waetzig ⁎Institute of Experimental and Clinical Pharmacology, University Hospital Schleswig-Holstein, Campus Kiel, Hospitalstrasse 4, 24105 Kiel, Germany

Abbreviations: AKT, protein kinase B; ATF-2, activatingbromo-2-deoxyuridine; DVD, domain of versatile dockingcellular signal-regulated kinase; Fas-L, Fas ligand; ID, immcipitation; JNK, c-Jun N-terminal kinase; MAPK, mitoMAP2K/MKK, mitogen-activated protein kinase kinase; Mtein kinase kinase kinase; Mdm-2, mouse double minute 2onic fibroblasts; MEK, mitogen-activated protein kinaseprotein kinase; p21, p21WAF1/CIP1; PI3K, phosphatidylinosSP600125; WB, Western blot⁎ Corresponding author. Tel.: +49 431 597 3511; fax: +

E-mail address: vicki.waetzig@pharmakologie.uni-kiel

0167-4889/$ – see front matter © 2014 Elsevier B.V. All rhttp://dx.doi.org/10.1016/j.bbamcr.2014.01.028

a b s t r a c t

a r t i c l e i n f o

Article history:Received 5 June 2013Received in revised form 22 January 2014Accepted 24 January 2014Available online 31 January 2014

Keywords:Map2k4JNKp38ERKMKK4Cell-cycle control

Mitogen-activated protein kinase kinase 4 (Map2k4) is a dual specificity serin/threonine protein kinase that isunique among all MAP2Ks in activating two different subfamilies of mitogen-activated protein kinases, the c-Jun N-terminal kinases (JNKs) and p38 kinases. Map2k4 is essential during embryogenesis and involved in a va-riety of physiological and pathological processes. However, studies on its role in cancer development revealedpartially conflicting data. In the present study, we report the identification of a novel splice variant of Map2k4,Map2k4δ, with an additional exon in front of the substrate binding D-domain. Map2k4δ is expressed togetherwith Map2k4 in various tissues from rat, mouse and human. In PC12 cells, both splice variants control cellcycle progression and basal apoptosis by using different signaling pathways. If expression and activation ofMap2k4 andMap2k4δ are at a certain, cell type-specific equilibrium, an appropriate cell growth is ensured. Over-expression of one kinase disrupts the intricate balance and either results in a highly proliferative or pro-apoptoticphenotype, partially reflecting the discrepancies in the literature on Map2k4 and its role in tumor development.Our findings contribute to the understanding of previous studies and point out thatMap2k4 has not always a def-inite function, but rather triggers a cellular reaction in concert with other modulators.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

The mitogen-activated proteinkinase (MAPK) pathways initiate awide spectrum of cellular responses. Starting with basal cellular pro-cesses like gene transcription or protein synthesis to highly concerteddevelopmental programs like proliferation, differentiation or apoptosis,their functions are not only versatile but can also be contradictory [1–3].All enzymes of the three-tiered MAPK signaling cascade are highlyconserved, mostly ubiquitously expressed and overlapping in theirsubstrates. Therefore, the specificity of their cellular effects has to besecured by expression, cellular localization and interactions with theirtargets or scaffolds. The last few years have shown that individual splicevariants of mitogen-activated protein kinases or their activators cancause distinct and sometimes conflicting cellular responses [4–9].

transcription factor 2; BrdU, 5-; E, embryonic day; ERK, extra-unodepletion; IP, immunopre-gen-activated protein kinase;AP3K, mitogen-activated pro-homolog; MEF, mouse embry-/extracellular signal-regulateditol-3-kinase; SB, SB203580; SP,

49 431 597 3522..de (V. Waetzig).

ights reserved.

The mitogen-activated protein kinase kinase 4 (Map2k4, MKK4) isunique among all known MAP2Ks in its ability to phosphorylate twodifferent MAPKs, p38 and JNK (c-Jun N-terminal kinase), respectively.TheMap2k4 gene encodes for a dual specificity serin/thereonine proteinkinase, 399 (human; 397 formouse and rat) amino acids long and high-ly conserved throughout evolution [10,11]. The tertiary structure of theprotein consists of three functional domains: (i) the kinase domain,where Map2k4 is activated by the majority of MAP3Ks (mitogen-activated protein kinase kinase kinases) by dual phosphorylation inthe S-I-A-K-T motif, (ii) a domain for versatile docking (DVD) whereupstream kinases and scaffold proteins bind to Map2k4 and (iii) aD-domain type of docking site for binding of p38 and JNK. Indeed, stud-ies withMap2k4−/−mice showed no obvious changes in p38 activationbut dramatic defects in JNK activation [12].

Map2k4 is involved in a variety of physiological and pathologicalprocesses. During embryogenesis, Map2k4 plays essential roles inhepatogenesis and brain development [13,14]. Therefore, mice with atargeted deletion of Map2k4 die between E11.5 and E13.5 [13]. More-over, Map2k4 is relevant in embryonic stem cell differentiation, theimmune system and the heart [15–17]. Regarding its chromosomallocus next to the tumor suppressor p53 on chromosome 17, Map2k4was intensively studied in tumor initiation. Indeed, studies of a widespectrum of primary cancers demonstrated loss-of-function mutationsin Map2k4 especially in lung and pancreatic tumors [18–22]. Further-more, endogenous expression of Map2k4 could reduce metastasis ofprostate and ovarian tumors in mice [23,24]. In contrast, an equal

876 W. Haeusgen et al. / Biochimica et Biophysica Acta 1843 (2014) 875–884

number of studies revealed a pro-oncogenic or metastasis promotingrole for Map2k4 [25–28]. For example, skin-specific Map2k4-deficientmice are resistant to carcinogen-induced tumorigenesis and injectionof Map2k4−/− pancreatic cells into mice reduced lung metastases ofthe primary pancreatic tumor [29]. However, as recently pointed out itis neither possible nor necessary to define a clear-cut role for Map2k4as a suppressor or promoter of tumorigenesis, but more important tounderstand Map2k4-dependent signaling in tumor cells [30].

In the present study, we identified a novel Map2k4 splice variant,Map2k4δ, expressed in various tissues of rat, mouse and human and al-ways together with Map2k4. We investigated the functional relevanceof Map2k4 and Map2k4δ in regulation of MAPK signaling in PC12cells, a well studied cellular model system. The results reveal distinctsignaling and functional properties of bothMap2k4 splice variants lead-ing to partially opposing cellular reactions. Our observations indicatethat the expression and activation of Map2k4 and Map2k4δ have to beat equilibrium to control cell growth and basal apoptosis. If unbalanced,proliferation or cell death predominate and lead to a phenotype that iseither suggestive of a pro-oncogene or a tumor-suppressor. Thus, ourfindings partially explain the discrepancies between the existing studiesand also emphasize that a single molecule does not always have a defi-nite role with a definite cellular effect, but rather fulfills a function inconcert with other modulators.

2. Materials and methods

2.1. Cell culture

Rat pheochromocytoma PC12 cells (DSMZ; Braunschweig, Germany)were cultured on collagen-coated plates in RPMI 1640 (PAA; Cölbe,Germany) supplemented with 5% fetal bovine serum (Biochrom; Berlin,Germany) and 10% horse serum (Biochrom; Berlin, Germany) at 37 °Cand 5% CO2. Transfected cell clones were kept in the presence of500 μg/ml G418 (Biochrom; Berlin, Germany) to maintain selection.For cell cycle analyses, cultures were transferred to serum-free mediumfor 24 h to stop proliferation. Cell growth was restarted by returning thecells to normal growth medium supplemented with fetal bovine- andhorse serum, resulting in a synchronized cell cycle [31].

2.2. Plasmids and transfection

cDNA of Map2k4 (GenBank™ Accession Number ID: AY880882.1)and Map2k4δ (GenBank™ Accession Number ID: KC795688) fromPC12 cells were inserted in the expression vector pEGFP-C3 (Clontech;Saint-Germain-en-Laye, France) as a BamHI/HindIII fragment. Transfec-tion of the Map2k4-pEGFP, Map2k4δ-pEGFP constructs or the pEGFPvector was performed using the Turbofect™ transfection system(Thermo Fisher Scientific; St. Leon-Rot, Germany) according to themanufacturer's instructions. For comparability, stable cell clones withsimilar expression levels of the endogenous Map2k4/δ protein and theMap2k4-pEGFP, Map2k4δ-pEGFP constructs or the pEGFP vector werechosen. The experiments were performed in at least 3 different cellclones of each splice variant. Microscopy, cell counts and Westernblots with five independent stable PC12 clones revealed the sameresults.

2.3. Cell counting

Electronic cell counting analyses were performed by CASY® TT cellcounter (Roche Applied Sciences; Mannheim, Germany). An aliquot ofthe cell suspension was mixed with CASYton® buffer and cell numberand viabilitywere determined according tomanufacturer's instructions.The CASY®TT cell counter can determine the complete cell number, theratio of cells with damaged cell membrane and themean volume of onecell.

2.4. 5-Bromo-2-deoxyuridine (BrdU) incorporation assay

Cell proliferation was measured using a BrdU incorporation assay(Roche Applied Sciences; Mannheim, Germany) according to themanufacturer's instructions. Cells grown in a 96-well-plate were incu-bated with BrdU for 4 h and fixed. After incubation with an anti-BrdU-POD-antibody and the subsequent substrate reaction, absorbance wasmeasured using an ELISA reader (Bio-Rad; München, Germany).

2.5. Whole cell extracts

If not indicated otherwise, chemicals were purchased from Sigma-Aldrich (Steinheim, Germany) or Carl Roth (Karlsruhe, Germany).Whole cell lysates were generated from subconfluent cells. Beforeharvesting, cells were washed with phosphate-buffered saline. Forwhole cell extracts, cells were resuspended in denaturing lysis buffer(20 mM Tris (pH 7.4), 2% SDS, 1% phosphatase inhibitor, and 1%protease inhibitor), incubated at 95 °C for 5 min, briefly sonicated, andcentrifuged to remove insoluble material (15,000 ×g for 15 min).

2.6. Western blot

Twenty micrograms of total cellular proteins was separated on10–15% SDS-polyacrylamide gels and transferred to poly-vinylidenedifluoride transfer membranes (Millipore, Schwalbach, Germany). Themembranes were blocked with 4% non-fat dry milk and incubated withthe primary antibodies according to the manufacturer's recommenda-tions. After three washing steps with TBST, the membranes were incu-bated with the HRP-conjugated secondary antibody for 30 min. AllWestern blotswere developed using the ECL chemiluminescence systemand Hyperfilm ECL (GE Healthcare; München, Germany). Between thestainings with phosphospecific antibodies and total kinase or total tran-scription factor antibodies, blots were stripped in 2% SDS, 62.5 mM Trisand 100 mM 2-mercaptoethanol for 30 min at 50 °C, washed withTBST and blocked again. All measurements of dual-phosphorylated ki-nase levels were normalized by hybridization with antibodies againsttotal kinase protein. To normalize for the protein content of each laneand to confirm equal loading, all membranes were finally stained withPonceau S. Antibodies against the following targets were purchasedfrom the indicated sources: mouse IgG (GE Healthcare; München,Germany); phospho-JNK and phospho-p38 (Promega; Mannheim,Germany); β-Actin (Sigma-Aldrich; Steinheim, Germany); AKT,Caspase-8, Caspase-9, cleaved Caspase-3, cleaved Caspase-8, cleavedCaspase-9, c-Myc, ERK, JNK, Map2k7, MEK1/2, p53, phospho-p53,phospho-AKT, phospho-ATF-2, phospho-c-Jun, phospho-c-Myc,phospho-ERK, phospho-Mdm-2, phospho-MEK1/2, phospho-Map2k4, phospho-Map2k7 and rabbit IgG (New England Biolabs;Frankfurt, Germany); CyclinD1, JNK1 and Mdm-2 (BD Pharmingen;Heidelberg, Germany); α-Tubulin, ATF-2, Caspase-3, Fas, Fas-L, JNK2,Map2k4, p21/WAF1 and p38 (Santa Cruz; Heidelberg, Germany).

2.7. RT-PCR

Total RNAwas isolated fromPC12 cells and tissues usingRNeasy PlusKit (Qiagen; Hilden Germany) according tomanufacturer's instructions.After quantification and quality validation, the remaining RNA wasaliquoted and stored at −80 °C. The reverse transcriptase reactionwas performed using the RevertAid™ H Minus First Strand cDNASynthesis Kit (Thermo Fisher Scientific; St. Leon-Rot, Germany) forRT-PCR according to the manufacturer's instructions with 0.5 μg ofRNA for each sample. The subsequent PCR was carried out with theDreamTaq™ PCR reagent system (Thermo Fisher Scientific; St. Leon-Rot, Germany) by using 0.4–0.6 μl of cDNA, 0.2 mM of each dNTP,0.2 μM of each primer, and 2.5 units of Taq DNA polymerase. Fornormalization, the histone H2A variant H2A.z was used. The absenceof contamination was checked by negative control assays, in which

877W. Haeusgen et al. / Biochimica et Biophysica Acta 1843 (2014) 875–884

RNA in the RT reaction or cDNA in the PCR were replaced by diethylpyrocarbonate-treated water or genomic DNA. Amplification was per-formed with the following primer sets: Gapdh_rat: sense 5′-AACGACCCCTTCATTGAC-3′, and antisense 5′-TCCACGACATACTCAGCAC-3′;Gapdh_mouse/human: sense 5′-TCCATGACAACTTTGGTATCGTGG-3′,and antisense 5′-GACGCCTGCTTCACCACCTTCT-3′; Map2k4_rat: sense5′ AGGGTAAACGCAAAGCCCT-3′, and antisense 5′-CGTATTCTTTTAACTGCCATTATCTG-3′; Map2k4_ mouse/human: sense 5′-AGGGTAAACGCAAAGCACTG-3′, and antisense 5′-TCAC TACTCCGCATTACTACATCC-3′;Map2k4δ_rat: sense 5′-GCATGCAGGGCTTTCAGATA-3′, and antisense5′-ACGAGCTGTCCACTGATGCCGA-3′; Map2k4δ_mouse/human: sense5′-GCATGCAGGGCTTTCAGATA-3′, and antisense 5′-TCACTACTCCGCATTACTACATCC-3′. The specificity of the primer pairswas verified by com-parative alignment using the BLAST database (NCBI, National Institutesof Health, Bethesda, MD).

2.8. Immunoprecipitation and immunodepletion

For immunoprecipitation (IP) and immunodepletion (ID) native celllysates were prepared from subconfluent cells. Before harvesting, cellswere washed with phosphate-buffered saline. For whole cell extracts,cells were resuspended in native lysis buffer (20 mM Tris (pH 7.6),250mMNaCl, 3mMEDTA, 3mMEGTA, 0.5%Non-Idet P40, 1% phospha-tase inhibitor, 1% β-glycerolphosphate and 1% protease inhibitor), incu-bated on ice for 30 min and centrifuged to remove insoluble material(15,000 ×g for 15 min). One hundred fifty μg of total proteins was usedfor IP and ID with the ImmunoCruz™ reagents (Santa Cruz; Heidelberg,Germany). All steps were performed at 4 °C if not mentioned other-wise. The samples were precleared for 30 min with the appropriatepreclearing Matrix. The antibody-IP matrix complex was formed byincubation of 1 μg primary antibody with 50 μl of the appropriate IPmatrix in 500 μl of PBS for 1 h. The IP antibody-IP matrix complex waswashed twice with cold PBS, the pre-cleared sample was mixed withthe antibody-IP matrix complex and incubated overnight. The negativecontrol was prepared by incubation of the antibody-IP matrix complexwith native lysis buffer. After centrifugation, the supernatants werecollected for Western blots as ID samples, while the pellets containingimmobilized proteins with IP matrix complexes were washedthree times with PBS. The immobilized proteins were dissociatedwith 5× reducing electrophoresis buffer in 20 μl of ultra-pure water at95 °C for 3 min. The supernatants from the dissociation were carefullyloaded onto 12% SDS-gels whereas the matrix was removed bycentrifugation.

2.9. Statistical analyses

All data were obtained from at least 4–5 independent sets of exper-iments. Statistical analyses were performed with the GraphPrism soft-ware (www.graphpad.com) using one-way analyses of variance withrepeated measurements.

3. Results

3.1. Identification and cloning of a novel Map2k4 splice variant

While amplifying the Map2k4mRNA from rat phaeochromocytomaPC12 cells, we identified an additional PCR product. Cloning and se-quence analyses of both cDNAs revealed two different splice variantsof the transcript. Apart from the expected Map2k4 transcript (11exons; 1191 bp), we amplified an apparent splice variant of Map2k4(12 exons; 1224 bp) including an additional exon (nt 103–135)(Fig. 1A). This exon contains canonical splice sites and is located be-tween exon 1 and exon 2. It encodes for an 11 amino acid fragment atthe N-terminus of Map2k4 (Fig. 1B). BLAST-like Alignment Tool (BLAT,UCSC genome browser) examinations of mouse and human genome

assemblies revealed that it is totally conserved within the Map2k4genes of all three species. We named the novel gene product Map2k4δ.

To confirm the appearance of the novelMap2k4 splice variant in dif-ferent tissues from rat, mouse and human, we designed splice variantspecific primers. Total RNAwas isolated from brain, lung, heart, pancre-as, kidney, liver and gut sections and transcribed into cDNA. As shown inFig. 1C, both splice variants of Map2k4 could be detected in all tissuesfrom all species, indicating a ubiquitous expression and splicing ofMap2k4.

3.2. Characterization of Map2k4- and Map2k4δ-transfected PC12 cells

As the expression analysis in different tissues indicates a generalcoexistence of both splice variants, we concluded that the balancebetween Map2k4 and Map2k4δ might be important for regular signaltransduction. Therefore, we chose a cell model where both splicevariants could be expressed together, but at different equilibrium condi-tions. For thus, both cDNAs were cloned into themammalian pEGFP-C3expression vector and stably transfected Map2k4- and Map2k4δ-PC12cell lines were generated. We monitored individual clones of both celllines with regard to expression levels and activity of exogenousMap2k4 and Map2k4δ. For further analyses we chose cell clones withsimilar protein levels.

Microscopy revealed no morphological differences (data notshown). Only a strong increase in cell debris was detected in theMap2k4-transfected cultures. To investigate the growth behavior ofvector-, Map2k4- and Map2k4δ-PC12 cells, electronic cell counts werecarried out. For this purpose, 0.3 × 106 cells were grown on collagen-coated 6-well plates for 5 days. Transfection with Map2k4 decreasedthe number of viable PC12 cells by 35%, while transfection with thenovelMap2k4δ increased cell counts by 40% compared to vector trans-fection (Fig. 2A).

Subsequently, stable cell clones were screened for protein and acti-vation levels of the direct Map2k4 substrates, JNK and p38, and theircounterpart Map2k7 by Western blotting. Transfection of PC12 cellswith Map2k4-pEGFP or Map2k4δ-pEGFP increased the overall proteinlevels as well as the phosphorylation of Map2k4 compared to vectorcontrols, while the intracellular protein amount of Map2k7 did notchange (Fig. 2B). However, phosphorylation of Map2k7 was reducedin Map2k4δ-transfected cells and further decreased by transfectionwith Map2k4 compared to vector-transfected PC12 cells. All three celllines showed similar levels of JNK1/2 proteins. Activation of JNK1/2, es-pecially the 54 kDa splice variants, was slightly elevated by transfectionwith Map2k4 and Map2k4δ. Looking at the protein amounts of singleJNK isoforms, transfection with Map2k4 and especially withMap2k4δ increased JNK2-p54 levels compared to control cells,whereas JNK1 was unaffected. For the second direct Map2k4 sub-strate p38, no changes could be observed in the amount and phos-phorylation status (Fig. 2B).

Furthermore, transfection with Map2k4 and Map2k4δ affectedactivity and protein levels of p38 and JNK targets. Protein amountsand phosphorylation of the transcription factor and proto-oncogenec-Jun were almost unaffected by transfection with Map2k4 comparedto vector-transfected cells but decreased following transfection withMap2k4δ (Fig. 2B). Interestingly, transfection with Map2k4δ enhancedphosphorylation of the activating transcription factor 2 (ATF-2) com-pared to vector- andMap2k4-PC12 cells. The amount of phosphorylatedc-Myc was strongly decreased by Map2k4.

To examine isoform-specific JNK activity, we performed immuno-depletion assays, which showed that transfection with Map2k4 in-creased phosphorylation of JNK2, while overexpression of Map2k4δreduced it (Fig. 2C). The phosphorylation of JNK1, however, was onlyelevated in Map2k4δ-transfected cells (Fig. 2C).

In summary, exogenous Map2k4 and Map2k4δ changed the activityof their counterpart Map2k7 and their substrate JNK and thereby affect

Fig. 1. Identification of a novelMap2k4 splice variant. (A) Schematic presentation of the exon/intron organization ofMap2k4 andMap2k4δ.Map2k4δ is produced by alternative splicing ofexon 2. The nucleotide sequence of exon 2 is shown under the scheme. (B) Sequence alignment of Map2k4δ (rat), Map2k4 (rat), Map2k4 (mouse) and Map2k4 (human). Differences be-tween the species are highlighted in gray. (C) Detection ofMap2k4 andMap2k4δ in different tissues from rat, mouse and human. Total RNAwas isolated from tissues and transcribed intocDNA. Both splice variants were amplified with specific primer pairs.

878 W. Haeusgen et al. / Biochimica et Biophysica Acta 1843 (2014) 875–884

phosphorylation patterns of transcription factors in a splice-variant-specific manner.

3.3. Map2k4 and Map2k4δ oppositely influence cell proliferation

Next we wanted to know if the different quantity of viable cells wasdue to changes in proliferation or cell death. We performed 5-bromo-deoyuridine (BrdU) assays. The transfection with Map2k4 reducedBrdU incorporation by 20% compared to vector transfected cells, whileMap2k4δ increased BrdU incorporation by 18% (Fig. 3A). These findings

indicate that an increased expression ofMap2k4 slows cell proliferation,whereas elevated levels of Map2k4δ enhance it.

As p21WAF1/Cip1 (p21), CyclinD1 and p53 are important regulators ofcell cycle progression and targets of both, p38 and JNK signaling, we an-alyzed their protein levels in the three different transfected PC12 celllines. Indeed,Map2k4-PC12 cells had increased amounts of p21 proteinand a decreased signal of CyclinD1 compared to vector- and Map2k4δ-transfected PC12 cells (Fig. 3B). In contrast, Map2k4δ transfection en-hanced the expression of CyclinD1, p21 levels were not affected. Proteinlevel of the tumor suppressor p53was increased by transfection of bothMap2k4 variants. However, the expression of the E3 ubiquitin-protein

Fig. 2. Characterization of Map2k4- and Map2kδ-PC12 cells. (A) Effect of Map2k4 andMap2k4δ on cell growth. Electronical cell counts of vector-, Map2k4- and Map2k4δ-transfected PC12 cells under normal growth conditions. Vector = 100%. ***/###, p b

0.001 for vector- vs. Map2k4- and Map2k4δ-transfected cells/transfection with Map2k4δvs. Map2k4. The data represents five independent experiments with ternary ascertain-ment. (B) Effect on activity and protein levels ofMAP2Ks,MAPKs and transcription factors.Immunoblots of total lysates (20 μg) of vector-,Map2k4- and Map2k4δ-transfected PC12cells were performed using antibodies against phospho-active forms of MAP2Ks, MAPKsand transcription factors. Each blot was re-probed with antibodies that recognize totalproteins. (C) Effect on activity levels of JNK isoforms. Single JNK isoforms wereimmunodepleted (ID) in native protein extracts (150 μg) from vector-, Map2k4- andMap2k4δ-transfected PC12 cells. The remaining extracts were first immunoblotted withan antibody against phosphorylated JNK1/2, then stripped and incubated with anantibody against JNK1/2. Western blots are representative for four independentexperiments.

Fig. 3. Effect ofMap2k4 andMap2k4δ on cell proliferation and basal apoptosis. (A) Effect ofMap2k4 and Map2k4δ on proliferation of PC12 cells. BrdU incorporation assay of vector-,Map2k4- and Map2k4δ-transfected cells. The cell lines were kept under the same condi-tions. The data shown represent five separate experiments with ternary ascertainment.***/###, p b 0.001 for vector- vs. Map2k4 and Map2k4δ-transfected cells/transfectionwith Map2k4δ vs. Map2k4. (B) Protein levels of the cell cycle regulators p21, CyclinD1,p53 and Mdm-2. Whole cell extracts (20 μg) from vector-, Map2k4- and Map2k4δ-transfected PC12 cells were immunoblotted against p21, CyclinD1, p53, Mdm-2 and re-probed with β-Actin antibody. (C) Immunoprecipitations (IP) of native protein extractsfrom unstimulated protein extracts of vector-, Map2k4- and Map2k4δ-transfected PC12cells. ∅, mock, without IP antibody; C, control, lysate. Native protein lysates wereimmunoprecipitatedwith an antibody raised against p53 and immunoblottedwith an an-tibody against Mdm-2. (D) Protein levels of apoptosis regulators. Whole cell extracts wereimmunoblotted against Fas, Fas-L and (cleaved) Caspase-3, -8, -9. Western blots are rep-resentative for four independent experiments.

879W. Haeusgen et al. / Biochimica et Biophysica Acta 1843 (2014) 875–884

ligase Mdm-2, which negatively regulates p53, was only increasedin Map2k4δ-PC12 cells. Immunoprecipitations showed that complexesof p53 and Mdm-2 were slightly reduced following Map2k4δ-transfection and nearly undetectable following overexpression ofMap2k4 (Fig. 3C).

Another reason for the observed changes in cell viability could bevariations in basal apoptosis. The death receptor Fas is inversely affectedby transfection with the twoMap2k4 splice variants (Fig. 3D). WhereasMap2k4 caused a reduction of Fas protein level, transfection withMap2k4δ slightly increased it. The amount of Fas ligand (Fas-L) was un-affected by Map2k4-overexpression. Surprisingly, effector Caspase-3was activated by transfection with Map2k4δ and especially Map2k4, asshown by increased cleavage. To analyze if Caspase-3 activation is trig-gered by extrinsic or intrinsic apoptosis, we monitored cleavage of

initiator Caspases-8 and -9. While, Map2k4-PC12 cells showed a higheractivation of Caspase-9, Map2k4δ-transfected cells had higher levels ofcleaved Capsase-8 (Fig. 3D).

These findings indicate that the cell number is regulated via prolifer-ation and apoptosis: (i) transfection of Map2k4 slowed cell cycle pro-gression and increased intrinsic apoptosis, resulting in a decreasednumber of viable cells; (ii) the novel splice variant Map2k4δ enhancedproliferation and therebymasked the elevated extrinsic apoptosis, lead-ing to elevated cell numbers.

880 W. Haeusgen et al. / Biochimica et Biophysica Acta 1843 (2014) 875–884

3.4. Regulation of p38 and JNK signal transduction by Map2k4and Map2k4δ

Our results show, that both Map2k4 splice variants have differenteffects on kinases, transcription factors and regulators of cell cycle andapoptosis. As Map2k4 is an activator of two individual signalingpathways, p38 and JNK, we investigatedMap2k4-splice-variant-specificregulation via both mitogen-activated protein kinases.

For defining the role of JNKs and p38 inMap2k4 andMap2k4δ signaltransmission, experiments with the specific kinase inhibitors SB203580(SB) or SP600125 (SP)were carried out. Inhibition of JNK reduced phos-phorylation and protein levels of c-Jun in all three cell lines (Fig. 4). Incontrast, the regulation of the activity of ATF-2, p53 and c-Myc by p38and JNK is more complex. In control cells, inhibition of p38 or JNK in-creased activity levels of ATF-2 and reduced phosphorylation of p53.SP additionally elevated p53protein and c-Myc activity levels. However,overexpression of Map2k4 reversed the effects of SB and SP on phos-phorylation of ATF-2. Both compounds decreased the activity levels ofATF-2. Inhibition of JNK also increased phosphorylated p53 and SB de-creased expression of p53. InMap2k4δ-transfected PC12 cells inhibitionof p38 and JNK only slightly reduced phosphorylation of p53. Interest-ingly, SB strongly increased the activity level of c-Myc and both inhibi-tors slightly increased ATF-2 phosphorylation.

Protein levels of the cell cycle regulator p21WAF/Cip1 were slightly in-creased by SB and SP in vector transfected PC12 cells but decreased inMap2k4 and Map2k4δ cells (Fig. 4). CyclinD1, however, was clearly re-duced by SB and even more by SP in vector controls. Map2k4-transfected cells only had low CyclinD1 levels, and inhibition of p38and JNK further decreased it. In Map2k4δ-PC12 cells CyclinD1 proteinlevels were only attenuated after JNK inhibition.

Fig. 4. Effect of p38 and JNK inhibition on phosphorylation and protein levels of their tar-gets. Vector-, Map2k4- and Map2k4δ-transfected cells were unstimulated (control; C) orincubated for 24 h with SB203580 (SB; 5 μM) or SP600125 (SP; 2 μM). Western blotsfrom whole cell extracts (20 μg) were first incubated with antibodies against phospho-proteins and re-probed with antibodies against total protein. Western blots are represen-tative for five independent experiments.

Those findings indicate that bothMap2k4 splice variants affect theircellular targets via distinct signal ways by p38 and JNK.

3.5. Effects of Map2k4 and Map2k4δ on related signaling pathways

The extracellular signal-regulated kinases (ERKs) 1/2 form the thirdgroup ofmammalianMAPKs and are also implicated in the regulation ofproliferation and cell death. We wanted to find out, if up-regulation ofMap2k4 signaling has an impact on activation of ERK1/2 and its activa-tor MEK1/2. Western blots of unstimulated whole cell extracts showedthat overexpression ofMap2k4 slightly reduced phosphorylation of ERK,whereasMap2k4δ strongly enhanced it (Fig. 5).

Apart from MAPK signaling, the PI3K/AKT kinase pathway is an im-portant regulator of survival, proliferation and differentiation in PC12cells that could interfere with the observed effects of exogenousMap2k4 and Map2k4δ. Protein levels of AKT were unaffected byMap2k4 andMap2k4δ (Fig. 5). Interestingly, activation of AKTwas clearlyincreased in Map2k4δ-PC12 cells under normal growth conditions.

Taken together,Map2k4 andMap2k4δ not only affect cellular signal-ing by phosphorylating their direct substrates, p38 and JNK, but also bycrosstalk with ERK and AKT in PC12 cells. Both kinases are stronglyactivated in Map2k4δ-PC12 but attenuated byMap2k4.

3.6. Effects of Map2k4 and Map2k4δ on signalosome compositions

Efficient signal transduction is ensured by the formation of proteincomplexes, so called signalosomes, built of kinases and their substrates.As experiments with kinase inhibitors clearly showed differencesbetween signal pathways of Map2k4- and Map2k4δ-PC12 cells, weanalyzed, if the observed changes are based on altered protein-proteininteractions. Immunoprecipitations demonstrated that the compositionof PC12 cell signalosomes was substantially affected by exogenous ex-pression of Map2k4 and Map2k4δ, respectively. Complexes of Map2k4and JNK1 were reduced in Map2k4 cells and further diminished byMap2k4δ overexpression (Fig. 6A). Map2k4:JNK2 signalosomes werealso reduced inMap2k4δ-PC12 cells but increased inMap2k4 cells. Inter-actions between Map2k4 and its second direct substrate p38 werestrongly increased in Map2k4δ-transfected PC12 cells. This suggeststhat Map2k4 prefers JNK2 whereas Map2k4δ favors p38 for interactionand activation. For the second activator of JNK, Map2k7, which showeddecreased activity levels following Map2k4- and Map2k4δ-transfection(Fig. 2B), also a diminished complexation could be detected comparedto control cells (Fig. 6B).

Fig. 5. Effects of Map2k4 and Map2k4δ on ERK1/2- and AKT-signaling. Immunoblots oftotal lysates (20 μg) of unstimulated vector-, Map2k4- and Map2k4δ-transfected PC12cells were performed using antibodies against phospho-active forms (p-MEK1/2, p-ERK1/2, p-AKT). Each blot was re-probed with antibodies that recognize total proteins.Western blots are representative for three independent experiments.

Fig. 6. Effects ofMap2k4 andMap2k4δ on protein–protein interactions. Immunoprecipita-tions (IP) of native protein extracts from unstimulated protein extracts of vector (V)-,Map2k4 (4)- and Map2k4δ (4δ)-transfected PC12 cells. ∅, mock, without IP antibody; C,control, lysate. Blots were re-probed with antibodies which recognize theimmunoprecipitated proteins. (A) Native protein lysates were immunoprecipitated withan antibody raised against Map2k4 and immunoblotted with antibodies against JNK1,JNK2 and p38. (B) Native protein lysates were immunoprecipitated with an antibodyraised against JNK1/2 and immunoblotted with an antibody against Map2k7. (C) Nativeprotein lysates were immunoprecipitated with an antibody raised against p38 or JNKand immunoblottedwith antibodies against p53, ATF-2 and c-Myc.Western blots are rep-resentative for four independent experiments.

881W. Haeusgen et al. / Biochimica et Biophysica Acta 1843 (2014) 875–884

Moreover, binding of JNK and p38 to their downstream substrateswas altered by the twoMap2k4 splice variants. ATF-2 was preferentiallybound to p38 in Map2k4- andMap2k4δ-PC12 cells, whereas JNK:ATF-2complexes were primarily found in control cells (Fig. 6C). Associationof p38 and p53 was almost diminished in Map2k4-cells compared tovector- and Map2k4δ-transfected PC12 cells. The activity of c-Myc wasalmost abolished by overexpression of Map2k4 (Fig. 2C). Indeed, bind-ing of c-Myc to p38 or JNK could hardly be detected in Map2k4-PC12cells. InMap2k4δ cells, the amount of p38:c-Myc complexes was nearlythe same as in control cells, but JNK:c-Myc interactions were decreasedas well, possibly explaining the reduced c-Myc phosphorylation com-pared to controls.

Togetherwith our previous results the followingmodels forMap2k4and Map2k4δ signaling can be suggested. Map2k4 preferentially acti-vates JNK2, leading to c-Jun activation and enhanced p21 protein levelscombined with a reduction of CyclinD1 and Fas. This signaling pathway

leads to reduced cell proliferation and increased basal intrinsic apopto-sis. In contrast, Map2k4δ increasingly binds to p38, which can contrib-ute to phosphorylation of ATF-2 and increased CyclinD1 and Fasprotein levels. Thereby, Map2k4δ enhances proliferation and sensitizesthe cells to extrinsic apoptosis.

4. Discussion

4.1. Map2k4δ — a novel splice variant of Map2k4

This study presents a novel splice variant ofMap2k4, calledMap2k4δand the cellular functions it modulates in concert with Map2k4 in PC12cells. So far, only one gene product of mammalian Map2k4 genes hasbeen described.Map2k4δ contains an additional exon of 30 bp. Genomeanalyses revealed that this exon is conserved in rat, mouse and human.Herewe demonstrate thatMap2k4δ is expressed in brains, hearts, lungs,pancreas, livers, testis and guts of all these species. Thus, both tran-scripts co-occur and, as Map2k4, Map2k4δ seems to be ubiquitouslyexpressed. Since the alternative spliced exon is located directly infront of the region encoding the D-domain, where p38 and JNK bindto Map2k4 [12], a functional relevance was assumable.

4.2. Map2k4 and Map2k4δ as regulators of proliferation and apoptosis

Both Map2k4 splice variants in naïve PC12 cells affect the growthbehavior of PC12 cells. Whereas Map2k4-overexpression decreasedthe number of viable PC12 cells under basal conditions,Map2k4δ trans-fection increased it. Those changes in cell numbers can be caused byaltered proliferation or cell death. BrdU-assays showed that elevatedexpression ofMap2k4 slowed downproliferation,whereasMap2k4δ en-hanced it. Indeed, Map2k4 mediated an increase in p21 protein levelswhereas CyclinD1 levels were down-regulated compared to vectortransfection, which reduces proliferation of PC12 cells [5,6,32].Our data shows, that up-regulation of p21 in Map2k4 cells isJNK-dependent. JNKs control p21 expression via p53 [33]. Furthermore,increased levels of p21 were shown to repress CyclinD1 expres-sion [34,35]. Transfection with Map2k4δ, however, strongly increasedCyclinD1 expression compared to control PC12 cells and did not affectthe p21 level. Here, an obvious enhancement of ERK and AKT phosphor-ylationwas found, which could contribute to cell cycle progression [36].Thus, depending onwhich splice variant predominates, cellular prolifer-ation is attenuated or elevated. Since cells do not even have to changethe expression of either Map2k4 variant, but just have to modify theiractivities by targeting scaffold proteins, activators, phosphatases ordegrading enzymes, keeping Map2k4 and Map2k4δ at just the rightequilibrium enables them to swiftly adapt to different surrounding con-ditions. This concept of balancing two different Map2k4s might alsohelp to understand the partly conflicting data that exist on the role ofMap2k4. For example, expression of dominant-negativeMap2k4 in em-bryonic stem cells reduced their proliferation rate [37]. Skin-specificknock-out of Map2k4 reduced carcinogen-induced tumorigenesis inmice [29]. In contrast, Map2k4 is a tumor suppressor in lung adenocar-cinomas and strongly reduced the number of overt metastasis in amouse xenograft model [25]. Of course, even if those contradictory re-sults were obtained from the same model system, heterogeneity ofcells and experimental conditions are a critical factor. Nevertheless,our results from the two Map2k4 splice variants could also account forthe inconsistencies in Map2k4 research so far. As known for variantsof other MAPK, e.g. JNK or p38, Map2k4 should not be considered as asingle, working enzyme that has a definite role, but rather as a pool ofMap2k4 splice forms that act in concert with other signaling moleculesto fulfill certain functions.

Looking at apoptosis regulators, interpretation of our results is not asclear. Expression of dominant-negativeMap2k4 in embryonic stem cellsreduced constitutive apoptosis [37], but Map2k4 knock-out studiesshowed no effect on cell death or even a pro-apoptotic one [28,29,38].

882 W. Haeusgen et al. / Biochimica et Biophysica Acta 1843 (2014) 875–884

In Map2k4-PC12 cells, reduced levels of death receptor Fas and in-creased amounts of cleaved Caspase-9 and -3 are detectable. Apparent-ly, overexpression ofMap2k4 promotes basal apoptosis via the intrinsicway. In contrast, Map2k4δ-PC12 cells have increased levels of Fas,cleaved Caspase-8 and -3, indicating extrinsic apoptosis. Therefore,both Map2k4 variants promote PC12 cell apoptosis under normal cellgrowth conditions, which is partially in line with previous findings[37]. In Map2k4δ cells, however, enhanced proliferation masks celldeath, slowed cell division of Map2k4 cells further amplifies the de-creased cell number.

4.3. Signal transduction by Map2k4 and Map2k4δ

Map2k4 is unique among all MAP2Ks in its ability to phosphorylatetwo different MAPKs, p38 and JNKs. Nevertheless, some studies re-vealed that theMap2k4 activity towards p38 is less pronounced than to-wards JNKs. For example, knockout of Map2k4 in MEFs or pancreaticcancer cells causes a decrease in phospho-JNK levels but not inphospho-p38 levels under basal or stress conditions [25,39]. Indeed,overexpression of the Map2k4 splice variants only enhanced JNK1/2phosphorylation, whereas phophorylated p38 was unaffected. Immu-noprecipitations show, that Map2k4:JNK2 complexes were mainlyfound inMap2k4-transfected PC12 cells. In contrast,Map2k4δ increasedthe formation of Map2k4:p38 complexes, while binding of Map2k4 toJNKs was less detectable. This raises the question why the amount ofphosphorylated p38 is unaffected compared to control cells. One possi-ble explanation is the increase of activated ATF-2 in Map2k4δ-transfected cells. ATF-2 can also be phosphorylated via ERK bymitogen-ic signals and via JNKand p38during stress signaling [40–44]. Therefore,the enhanced activity of ERK1/2 could account for the increasedphospho-ATF-2 levels in Map2k4δ-cells, which, in turn, limits p38 ki-nase activity through a negative feedback including transcriptional acti-vation of MAPK phosphatases [45]. Still, ATF-2 can be associated withp38 after transfection withMap2k4 and Map2k4δ. The enhanced phos-phorylation of JNK in Map2k4δ-transfected cells seems to be triggeredbyMap2k7, as the activity of Map2k7was increased as well as complexformation with JNK compared toMap2k4-PC12 cells. Concurrently, theactivity level of Map2k7 was decreased in Map2k4-transfected PC12cells, which indicates that here the increase of activated JNK1/2 wasdue to Map2k4. Likewise, our results suggest that the differential

Fig. 7. Cellular signaling of Map2k4 and Map2k4δ. Schematic model of Map2k4- and Map2k4phosphorylation of p38. Both kinases have opposed effects on AKT- and ERK1/2-signaling, soan appropriate cell growth.

activation of Map2k4 and Map2k7 in the transfected cell lines leads tothe phosphorylation of specific JNK isoforms: in Map2k4-transfectedcells, JNK2 was activated, while JNK1 phosphorylation predominatesin Map2k4δ-transfected cells.

Surprisingly, phosphorylation of c-Jun was unaffected by Map2k4-transfection and even reduced followingMap2k4δ-transfection. Studieson the distribution of Map2k4, JNKs and c-Jun in primary neurons haveshown that Map2k4 and JNKs are mainly located in the cytoplasm,whereas c-Jun and Map2k7 are mainly located in the nucleus underbasal conditions [46], so that JNKs would be refrained from phosphory-lating c-Jun. This, in turn, could explain the enhanced ERK1/2 activity,since c-Jun-mediated gene transcription can be critical to attenuateERK activation [47].

Experiments with kinase inhibitors and immunoprecipitations haveshown that both splice variants alter binding and phosphorylation oftranscription factors by p38 and JNK. It is known, that p38 stabilizesp53, whereas JNK trigger its ubiquitination [48,49]. Map2k4δ increasescomplexes of p53 and p38which leads to higher amounts of p53 protein.Concurrently, binding of p53 toMdm-2 ismore frequently. This ubiquitinE3-ligase of p53 is a target not only of p38 and JNK, but also of AKTwhichshows enhanced phosphorylation in Map2k4δ-transfected cells [50].Thus,Map2k4 is a positive regulator of tumor suppressor p53 but overex-pression of Map2k4δ indirectly stimulates its degradation.

Interestingly, exogenous Map2k4 nearly abolished phosphorylationand protein levels of the c-Myc. The proto-oncogene c-Myc promotescell cycle progression, e.g. by repression of p21 [51]. In fact, reductionof c-Myc is accompanied by high levels of p21 in Map2k4-transfectedcells.

In conclusion, in the present study we have identified a novel splicevariant of Map2k4. Map2k4δ is a result of alternative splicing on exon 2of the Map2k4 gene and it is expressed in various mammalian tissues.In contrast to the anti-proliferative effect of overexpressed Map2k4,Map2k4δ drives cell proliferation of PC12 cells. Interestingly, dependingon which Map2k4 splice variant predominates cellular reactions areconsiderably different. Whereas Map2k4 promotes the Map2k4-JNK-c-Jun signal pathway, Map2k4δ-PC12 cells show a preference forMap2k4-p38-ATF-2 signaling. Placing our results in context of previousfunctional studies, Map2k4 alone tightly controls cell proliferationand would rather be anti-oncogenic. If Map2k4δ signaling prevailsMap2k4, cellular effects of pro-oncogenic ERK1/2 and AKT are promot-ed (Fig. 7).

δ-mediated signaling. Map2k4 preferentially activates JNK, whereas Map2k4δ augmentsthat a balanced expression and activation of both Map2k4 splice variants is important for

883W. Haeusgen et al. / Biochimica et Biophysica Acta 1843 (2014) 875–884

Ethical standards

The authors declare that all experiments performed complywith thecurrent laws of Germany.

Conflict of interest

The authors declare that they have no conflict of interest.

Acknowledgements

The authors thank Dr. Heinz-Hermann Hugo for human tissuesand Irina Naujoks, Elke Schröder and Angela Schulz for their excellenttechnical assistance. This work was supported by the DeutscheForschungsgemeinschaft Grant SFB415, the STREP STRESSPROTECTwithin the European 6th FP and the intramural grant of the Faculty ofMedicine Christian-Albrecht's-University Kiel.

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.bbamcr.2014.01.028.

References

[1] M. Raman, W. Chen, M.H. Cobb, Differential regulation and properties of MAPKs,Oncogene 26 (2007) 3100–3112.

[2] R. Seger, E.G. Krebs, The MAPK signaling cascade, FASEB J. 9 (1995) 726–735.[3] C. Widmann, S. Gibson, M.B. Jarpe, G.L. Johnson, Mitogen-activated protein kinase:

conservation of a three-kinase module from yeast to human, Physiol. Rev. 79(1999) 143–180.

[4] J. Cui, S.Y. Han, C. Wang, W. Su, L. Harshyne, M. Holgado-Madruga, A.J. Wong, c-JunNH(2)-terminal kinase 2alpha2 promotes the tumorigenicity of human glioblastomacells, Cancer Res. 66 (2006) 10024–10031.

[5] W. Haeusgen, T. Herdegen, V. Waetzig, Specific regulation of JNK signalling by thenovel rat MKK7gamma1 isoform, Cell. Signal. 22 (2010) 1761–1772.

[6] W. Haeusgen, T. Herdegen, V. Waetzig, MKK7gamma1 reverses nerve growth factorsignals: proliferation and cell death instead of neuritogenesis and protection, Cell.Signal. 23 (2011) 1281–1290.

[7] R.T. Nitta, C.A. Del Vecchio, A.H. Chu, S.S. Mitra, A.K. Godwin, A.J. Wong, The role ofthe c-Jun N-terminal kinase 2-alpha-isoform in non-small cell lung carcinomatumorigenesis, Oncogene 30 (2011) 234–244.

[8] Y.D. Shaul, G. Gibor, A. Plotnikov, R. Seger, Specific phosphorylation and activation ofERK1c by MEK1b: a unique route in the ERK cascade, Genes Dev. 23 (2009)1779–1790.

[9] C.L. Winchester, H. Ohzeki, D.A. Vouyiouklis, R. Thompson, J.M. Penninger, K.Yamagami, J.D. Norrie, R. Hunter, J.A. Pratt, B.J. Morris, Converging evidence thatsequence variations in the novel candidate gene MAP2K7 (MKK7) are functionallyassociated with schizophrenia, Hum. Mol. Genet. 21 (2012) 4910–4921.

[10] W. Haeusgen, T. Herdegen, V. Waetzig, The bottleneck of JNK signaling: molecularand functional characteristics of MKK4 and MKK7, Eur. J. Cell Biol. 90 (2011)536–544.

[11] A. Lin, A. Minden, H. Martinetto, F.X. Claret, C. Lange-Carter, F. Mercurio, G.L.Johnson, M. Karin, Identification of a dual specificity kinase that activates the Junkinases and p38-Mpk2, Science 268 (1995) 286–290.

[12] D.T. Ho, A.J. Bardwell, M. Abdollahi, L. Bardwell, A docking site in MKK4 mediateshigh affinity binding to JNK MAPKs and competes with similar docking sites inJNK substrates, J. Biol. Chem. 278 (2003) 32662–32672.

[13] S. Ganiatsas, L. Kwee, Y. Fujiwara, A. Perkins, T. Ikeda, M.A. Labow, L.I. Zon, SEK1 de-ficiency reveals mitogen-activated protein kinase cascade crossregulation and leadsto abnormal hepatogenesis, Proc. Natl. Acad. Sci. U. S. A. 95 (1998) 6881–6886.

[14] J.K. Lee, W.S. Hwang, Y.D. Lee, P.L. Han, Dynamic expression of SEK1 suggests mul-tiple roles of the gene during embryogenesis and in adult brain of mice, Brain Res.Mol. Brain Res. 66 (1999) 133–140.

[15] G. Choukroun, R. Hajjar, S. Fry, F. del Monte, S. Haq, J.L. Guerrero, M. Picard, A.Rosenzweig, T. Force, Regulation of cardiac hypertrophy in vivo by thestress-activated protein kinases/c-Jun NH(2)-terminal kinases, J. Clin. Invest. 104(1999) 391–398.

[16] H. Nishina, K.D. Fischer, L. Radvanyi, A. Shahinian, R. Hakem, E.A. Rubie, A. Bernstein,T.W. Mak, J.R. Woodgett, J.M. Penninger, Stress-signalling kinase Sek1 protects thy-mocytes from apoptosis mediated by CD95 and CD3, Nature 385 (1997) 350–353.

[17] W. Swat, K. Fujikawa, S. Ganiatsas, D. Yang, R.J. Xavier, N.L. Harris, L. Davidson, R.Ferrini, R.J. Davis, M.A. Labow, R.A. Flavell, L.I. Zon, F.W. Alt, SEK1/MKK4 is requiredfor maintenance of a normal peripheral lymphoid compartment but not for lympho-cyte development, Immunity 8 (1998) 625–634.

[18] H.L. Kim, D.J. Vander Griend, X. Yang, D.A. Benson, Z. Dubauskas, B.A. Yoshida, M.A.Chekmareva, Y. Ichikawa, M.H. Sokoloff, P. Zhan, T. Karrison, A. Lin, W.M. Stadler, T.Ichikawa, M.A. Rubin, C.W. Rinker-Schaeffer, Mitogen-activated protein kinase

kinase 4 metastasis suppressor gene expression is inversely related to histologicalpattern in advancing human prostatic cancers, Cancer Res. 61 (2001) 2833–2837.

[19] K. Nakayama, N. Nakayama, B. Davidson, H. Katabuchi, R.J. Kurman, V.E. Velculescu,M. Shih Ie, T.L. Wang, Homozygous deletion of MKK4 in ovarian serous carcinoma,Cancer Biol. Ther. 5 (2006) 630–634.

[20] G.H. Su, J.J. Song, E.A. Repasky, M. Schutte, S.E. Kern, Mutation rate ofMAP2K4/MKK4 in breast carcinoma, Hum. Mutat. 19 (2002) 81.

[21] D.H. Teng, W.L. Perry III, J.K. Hogan, M. Baumgard, R. Bell, S. Berry, T. Davis, D. Frank,C. Frye, T. Hattier, R. Hu, S. Jammulapati, T. Janecki, A. Leavitt, J.T. Mitchell, R. Pero, D.Sexton, M. Schroeder, P.H. Su, B. Swedlund, J.M. Kyriakis, J. Avruch, P. Bartel, A.K.Wong, S.V. Tavtigian, et al., Human mitogen-activated protein kinase kinase 4 as acandidate tumor suppressor, Cancer Res. 57 (1997) 4177–4182.

[22] W. Xin, K.J. Yun, F. Ricci, M. Zahurak, W. Qiu, G.H. Su, C.J. Yeo, R.H. Hruban, S.E. Kern,C.A. Iacobuzio-Donahue, MAP2K4/MKK4 expression in pancreatic cancer: geneticvalidation of immunohistochemistry and relationship to disease course, Clin. CancerRes. 10 (2004) 8516–8520.

[23] S.D. Yamada, J.A. Hickson, Y. Hrobowski, D.J. Vander Griend, D. Benson, A. Montag, T.Karrison, D. Huo, J. Rutgers, S. Adams, C.W. Rinker-Schaeffer, Mitogen-activated pro-tein kinase kinase 4 (MKK4) acts as a metastasis suppressor gene in human ovariancarcinoma, Cancer Res. 62 (2002) 6717–6723.

[24] B.A. Yoshida, Z. Dubauskas, M.A. Chekmareva, T.R. Christiano, W.M. Stadler, C.W.Rinker-Schaeffer, Mitogen-activated protein kinase kinase 4/stress-activatedprotein/Erk kinase 1 (MKK4/SEK1), a prostate cancer metastasis suppressor geneencoded by human chromosome 17, Cancer Res. 59 (1999) 5483–5487.

[25] S.C. Cunningham, E. Gallmeier, T. Hucl, D.A. Dezentje, E.S. Calhoun, G. Falco, K.Abdelmohsen, M. Gorospe, S.E. Kern, Targeted deletion of MKK4 in cancer cells: adetrimental phenotype manifests as decreased experimental metastasis andsuggests a counterweight to the evolution of tumor-suppressor loss, Cancer Res.66 (2006) 5560–5564.

[26] T.S. Khatlani, M. Wislez, M. Sun, H. Srinivas, K. Iwanaga, L. Ma, A.E. Hanna, D. Liu, L.Girard, Y.H. Kim, J.R. Pollack, J.D. Minna, Wistuba II, J.M. Kurie, c-Jun N-terminalkinase is activated in non-small-cell lung cancer and promotes neoplastic transfor-mation in human bronchial epithelial cells, Oncogene 26 (2007) 2658–2666.

[27] H.Y. Lee, S.H. Oh, Y.A. Suh, J.H. Baek, V. Papadimitrakopoulou, S. Huang, W.K.Hong, Response of non-small cell lung cancer cells to the inhibitors of phos-phatidylinositol 3-kinase/Akt- and MAPK kinase 4/c-Jun NH2-terminal kinasepathways: an effective therapeutic strategy for lung cancer, Clin. Cancer Res.11 (2005) 6065–6074.

[28] L. Wang, Y. Pan, J.L. Dai, Evidence of MKK4 pro-oncogenic activity in breast andpancreatic tumors, Oncogene 23 (2004) 5978–5985.

[29] K.G. Finegan, C. Tournier, The mitogen-activated protein kinase kinase 4 has apro-oncogenic role in skin cancer, Cancer Res. 70 (2010) 5797–5806.

[30] S.C. Cunningham, E. Gallmeier, S.E. Kern, MKK4 as oncogene or tumor suppressor: incancer and senescence, the story's getting old, Aging (Albany NY) 2 (2010)752–753.

[31] B.B. Rudkin, P. Lazarovici, B.Z. Levi, Y. Abe, K. Fujita, G. Guroff, Cell cycle-specific ac-tion of nerve growth factor in PC12 cells: differentiation without proliferation,EMBO J. 8 (1989) 3319–3325.

[32] J.A. Erhardt, R.N. Pittman, Ectopic p21(WAF1) expression induces differentiation-specific cell cycle changes in PC12 cells characteristic of nerve growth factor treat-ment, J. Biol. Chem. 273 (1998) 23517–23523.

[33] W.S. el-Deiry, T. Tokino, V.E. Velculescu, D.B. Levy, R. Parsons, J.M. Trent, D. Lin, W.E.Mercer, K.W. Kinzler, B. Vogelstein, WAF1, a potential mediator of p53 tumorsuppression, Cell 75 (1993) 817–825.

[34] E. Stepniak, R. Ricci, R. Eferl, G. Sumara, I. Sumara, M. Rath, L. Hui, E.F. Wagner,c-Jun/AP-1 controls liver regeneration by repressing p53/p21 and p38 MAPKactivity, Genes Dev. 20 (2006) 2306–2314.

[35] H. Wu, M. Wade, L. Krall, J. Grisham, Y. Xiong, T. Van Dyke, Targeted in vivo expres-sion of the cyclin-dependent kinase inhibitor p21 halts hepatocyte cell-cycle pro-gression, postnatal liver development and regeneration, Genes Dev. 10 (1996)245–260.

[36] J.C. Chambard, R. Lefloch, J. Pouyssegur, P. Lenormand, ERK implication in cell cycleregulation, Biochim. Biophys. Acta 1773 (2007) 1299–1310.

[37] M. Cazillis, A.F. Bringuier, D. Delautier, M. Buisine, D. Bernuau, C. Gespach, A. Groyer,Disruption of MKK4 signaling reveals its tumor-suppressor role in embryonic stemcells, Oncogene 23 (2004) 4735–4744.

[38] D. Xia, H. Srinivas, Y.H. Ahn, G. Sethi, X. Sheng, W.K. Yung, Q. Xia, P.J. Chiao, H. Kim,P.H. Brown, I.I. Wistuba, B.B. Aggarwal, J.M. Kurie, Mitogen-activated protein kinasekinase-4 promotes cell survival by decreasing PTEN expression through anNF kappaB-dependent pathway, J. Biol. Chem. 282 (2007) 3507–3519.

[39] T. Wada, K. Nakagawa, T. Watanabe, G. Nishitai, J. Seo, H. Kishimoto, D. Kitagawa, T.Sasaki, J.M. Penninger, H. Nishina, T. Katada, Impaired synergistic activation ofstress-activated protein kinase SAPK/JNK in mouse embryonic stem cells lackingSEK1/MKK4: different contribution of SEK2/MKK7 isoforms to the synergistic acti-vation, J. Biol. Chem. 276 (2001) 30892–30897.

[40] B. Derijard, J. Raingeaud, T. Barrett, I.H. Wu, J. Han, R.J. Ulevitch, R.J. Davis, Indepen-dent human MAP-kinase signal transduction pathways defined by MEK and MKKisoforms, Science 267 (1995) 682–685.

[41] S. Gupta, D. Campbell, B. Derijard, R.J. Davis, Transcription factor ATF2 regulation bythe JNK signal transduction pathway, Science 267 (1995) 389–393.

[42] J.M. Kyriakis, P. Banerjee, E. Nikolakaki, T. Dai, E.A. Rubie, M.F. Ahmad, J. Avruch, J.R.Woodgett, The stress-activated protein kinase subfamily of c-Jun kinases, Nature369 (1994) 156–160.

[43] D.M. Ouwens, N.D. de Ruiter, G.C. van der Zon, A.P. Carter, J. Schouten, C. van derBurgt, K. Kooistra, J.L. Bos, J.A. Maassen, H. van Dam, Growth factors can activateATF2 via a two-step mechanism: phosphorylation of Thr71 through the Ras-

884 W. Haeusgen et al. / Biochimica et Biophysica Acta 1843 (2014) 875–884

MEK-ERK pathway and of Thr69 through RalGDS-Src-p38, EMBO J. 21 (2002)3782–3793.

[44] J. Raingeaud, S. Gupta, J.S. Rogers, M. Dickens, J. Han, R.J. Ulevitch, R.J. Davis,Pro-inflammatory cytokines and environmental stress cause p38 mitogen-activated protein kinase activation by dual phosphorylation on tyrosine and threo-nine, J. Biol. Chem. 270 (1995) 7420–7426.

[45] W. Breitwieser, S. Lyons, A.M. Flenniken, G. Ashton, G. Bruder, M. Willington, G.Lacaud, V. Kouskoff, N. Jones, Feedback regulation of p38 activity via ATF2 is essen-tial for survival of embryonic liver cells, Genes Dev. 21 (2007) 2069–2082.

[46] E.T. Coffey, V. Hongisto, M. Dickens, R.J. Davis, M.J. Courtney, Dual roles for c-JunN-terminal kinase in developmental and stress responses in cerebellar granule neu-rons, J. Neurosci. 20 (2000) 7602–7613.

[47] Y.H. Shen, J. Godlewski, J. Zhu, P. Sathyanarayana, V. Leaner, M.J. Birrer, A. Rana, G.Tzivion, Cross-talk between JNK/SAPK and ERK/MAPK pathways: sustained

activation of JNK blocks ERK activation by mitogenic factors, J. Biol. Chem. 278(2003) 26715–26721.

[48] D.V. Bulavin, S. Saito, M.C. Hollander, K. Sakaguchi, C.W. Anderson, E. Appella, A.J.Fornace Jr., Phosphorylation of human p53 by p38 kinase coordinates N-terminalphosphorylation and apoptosis in response to UV radiation, EMBO J. 18 (1999)6845–6854.

[49] S.Y. Fuchs, V. Adler, M.R. Pincus, Z. Ronai, MEKK1/JNK signaling stabilizes and acti-vates p53, Proc. Natl. Acad. Sci. U. S. A. 95 (1998) 10541–10546.

[50] L.D. Mayo, D.B. Donner, A phosphatidylinositol 3-kinase/Akt pathway promotestranslocation of Mdm2 from the cytoplasm to the nucleus, Proc. Natl. Acad. Sci.U. S. A. 98 (2001) 11598–11603.

[51] R.A. Steinman, B. Hoffman, A. Iro, C. Guillouf, D.A. Liebermann, M.E. el-Houseini, In-duction of p21 (WAF-1/CIP1) during differentiation, Oncogene 9 (1994)3389–3396.