The Epithelium-specific ETS Protein EHF/ESE-3 Is aContext-dependent Transcriptional RepressorDownstream of MAPK Signaling Cascades*

Received for publication, December 4, 2000, and in revised form, February 23, 2001Published, JBC Papers in Press, March 19, 2001, DOI 10.1074/jbc.M010930200

Antonio Tugores,a Jennifer Le, Irina Sorokina,b A. J. Snijders,c Mabel Duyao,d

P. Sanjeeva Reddy,e Leone Carlee,f Mathew Ronshaugen,g Arcady Mushegian,h Tim Watanaskul,i

Sunny Chu,d Alan Buckler,d Spencer Emtage,b and Mary Kay McCormickj

From Axys Pharmaceuticals, Inc., South San Francisco, CA 94080 and the gDepartment of Biology, University ofCalifornia at San Diego, La Jolla, California 92093

Exon trapping and cDNA selection procedures wereused to search for novel genes at human chromosome11p13, a region previously associated with loss of het-erozygosity in epithelial carcinomas. Using these ap-proaches, we found the ESE-2 and ESE-3 genes, codingfor ETS domain-containing transcription factors.These genes lie in close proximity to the catalase genewithin a ;200-kilobase genomic interval. ESE-3 mRNAis widely expressed in human tissues with high epithe-lial content, and immunohistochemical analysis with anewly generated monoclonal antibody revealed thatESE-3 is a nuclear protein expressed exclusively indifferentiated epithelial cells and that it is absent inthe epithelial carcinomas tested. In transient transfec-tions, ESE-3 behaves as a repressor of the Ras- or phor-bol ester-induced transcriptional activation of asubset of promoters that contain ETS and AP-1 bindingsites. ESE-3-mediated repression is sequence- and con-text-dependent and depends both on the presence ofhigh affinity ESE-3 binding sites in combination withAP-1 cis-elements and the arrangement of these siteswithin a given promoter. We propose that ESE-3 mightbe an important determinant in the control of epithelialdifferentiation, as a modulator of the nuclear responseto mitogen-activated protein kinase signaling cascades.

Intracellular signaling through mitogen-activated protein(MAP)1 kinases is a universal mechanism controlling multipleaspects of cell division, migration, and differentiation (1).Among the known nuclear effectors downstream of MAP kinasesignaling cascades are members of the ETS family of transcrip-tion factors, a group of proteins that share a conserved DNAbinding domain, known as the ETS domain, and that are onlyfound in metazoans. ETS transcription factors have been im-plicated as positive and negative regulators in signal transduc-tion pathways that control a number of cellular processes,including differentiation, proliferation, cellular migration orinvasion, and inflammation, and some of these factors areindeed direct targets of MAP kinases (2–5).

The biological role of ETS proteins in the integration ofsignals from MAP kinases has been best characterized in Dro-sophila melanogaster. Genetic and biochemical studies haveled to the identification of two ETS proteins, Pointed and An-terior Open (Aop)/Yan, as pivotal components in the nuclearintegration of MAPK signaling cascades (6–10). The activity ofPointed P2, a product of the pointed gene (11), and a putativeortholog of the vertebrate ETS-1 and ETS-2 proteins (12, 3), isdependent upon direct phosphorylation by MAP kinases. Onthe other hand, Aop/Yan has been described as a repressor ofPointed-responsive genes (6–10) and is also a downstreamtarget of MAP kinase cascades. No mammalian epithelial or-tholog of Aop/Yan has been identified yet. It has been proposedthat phosphorylation of Aop/Yan by MAP kinases results in itsinactivation and nuclear export, thus alleviating repressionand allowing the transcriptional activation of Pointed targetgenes (6). The balance between the activity of these two tran-scription factors appears to modulate the specificity of thenuclear response to MAPK signaling cascades.

ETS proteins have also been shown to interact with eachother and with other structurally unrelated transcription fac-tors (see Refs. 3, 4, 13, and 14 and references therein), therebyproviding additional complexity to their role as transcriptionalregulators. In particular, the interaction of ETS proteins withJun has significant relevance, since ETS binding sites appearin the vicinity of AP-1 cis-elements in the promoters of a num-ber of genes involved in cell proliferation and migration, includ-ing those of extracellular proteases. This functional interaction

* This work was supported by Boehringer Ingelheim Pharma KG,and Axys Pharmaceuticals Inc. (formerly Sequana Therapeutics, Inc.).The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby marked“advertisement” in accordance with 18 U.S.C. Section 1734 solely toindicate this fact.

The nucleotide sequence(s) reported in this paper has been submittedto the GenBankTM/EBI Data Bank with accession number(s) AF212848.

a To whom correspondence should be addressed: Dept. of Biology, 0349,University of California, San Diego, 9500 Gilman Dr., La Jolla, CA 92093.Tel.: 858-822-0461; Fax: 858-822-0460; E-mail: atugores@yahoo.com.

b Present address: Structural Genomics, Inc., 10505 Roselle St., SanDiego, CA 92121.

c Present address: Highwire Press, 1454 Page Mill Rd., Palo Alto, CA94304-1124.

d Present address: Ardais Corp., 128 Spring St., Lexington, MA 02421.e Present address: Maxim Pharmaceuticals, 6650 Nancy Ridge Dr.,

San Diego, CA 92121.f Present address: Division of Molecular Genetics/H5, The Nether-

lands Cancer Institute, Plesmanlaan 121, 1066 CX Amsterdam, TheNetherlands.

h Present address: Stowers Institute for Medical Research, 1000 E.50th St., Kansas City, MO 64110.

i Present address: Biocept Inc., 2151 Las Palmas Dr., Suite C, Carls-bad, CA 92009.

j Present address: GenProbe Inc., 10210 Genetic Center Dr., SanDiego, CA 92121.

1 The abbreviations used are: MAP, mitogen-activated protein;MAPK, MAP kinase; AP-1, activator protein 1; BAC, bacterial artificialchromosome; LOH, loss of heterozygosity; mAb, monoclonal antibody;ORE, oncogene response element; PCR, polymerase chain reaction;PMA, phorbol myristate acetate; YAC, yeast artificial chromosome;MMP, matrix metalloproteinase; ORF, open reading frame; kb, kilo-base; contig, group of overlapping clones.

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 276, No. 23, Issue of June 8, pp. 20397–20406, 2001© 2001 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

This paper is available on line at http://www.jbc.org 20397

by guest on March 19, 2020

http://ww

w.jbc.org/

Dow

nloaded from

results in the transcriptional activation of these genes in re-sponse to both extracellular signal-regulated kinase and JunN-terminal kinase/stress-activated protein kinase MAP kinasesignaling cascades (3, 15–17). Paradoxically, similar regulatoryelements are present in the promoters of protease inhibitorsand genes expressed during epithelial cell differentiation (18–23), raising the question of how specificity is achieved at thenuclear level. It is plausible that members of the ETS familymay be important elements in the generation of target genespecificity. In this regard, transcription factors belonging to aspecific group within the ETS family of proteins, such as ESE-1/ESX/JEN/ERT/ELF-3, ELF-5/ESE-2, and more recentlyEHF/ESE-3, have been implicated in the control of the tran-scription of epithelial differentiation genes (21, 24–32). Thepresence of various members with diverse binding specificitiesand the distinct expression patterns within the ESE groupcould be important for the selective control of the expression ofepithelial target genes by these ETS transcription factors.

Elevated levels of ETS proteins have been also associatedwith cellular transformation (30, 33). Cancer is the result of amultistep process that is commonly associated with the loss offunction of genes required for the maintenance of nonprolifera-tive, differentiated cell fates and whose loss would facilitatetumor survival and/or metastasis. One of the mechanisms thatmay lead to the haploinsufficiency of such regulatory genes isthe deletion of genomic regions containing these genes as aresult of the increased genomic instability associated with thetransformation process, a mechanism known as allelic loss ofheterozygosity (LOH) (34).

We have performed an extensive search for genes at humanchromosome 11p13, in a region that has been previously asso-ciated with allelic loss of heterozygosity in epithelial carcino-mas (see Refs. 34–37 and references therein). In particular,LOH at this region is associated with increased metastaticpotential, thus suggesting that this region contains genes re-quired for the control of cell migration (36). As a result of thissearch, we have found the genes coding for two epithelium-specific ETS transcription factors, which we initially identifiedas I and J in our study (38) and have been later reported asELF5/ESE-2 and EHF/ESE-3, respectively (26–28, 31). Theanalysis of EHF/ESE-3 revealed that the expression of thistranscription factor is restricted to differentiated epithelialcells, and remarkably, it is absent in epithelial carcinomasoriginated from the same tissues where this factor is expressednormally. EHF/ESE-3 is a transcriptional repressor of a subsetof ETS/AP-1-responsive genes, including the interstitial colla-genase gene (matrix metalloproteinase 1; MMP-1) promoter.We suggest that EHF/ESE-3 might be an important regulatoryfactor in the maintenance of a differentiated phenotype inepithelial cells by modulating the nuclear response to MAPkinase signaling cascades.

EXPERIMENTAL PROCEDURES

Cloning of Genes in Chromosome 11p13—To generate a physical mapfor the 11p13 region, we began with searches of publicly availableinformation sources for yeast artificial chromosome (YAC) clones thatbest represented the region. Queries of the Whitehead Institute Ge-nome Center’s physical map data base yielded multiple CEPH mega-YAC clones that were detected by genetic markers from the region.Each clone was analyzed by STS content mapping with all availablemarkers by pulsed field gel electrophoresis and blotting and by fluores-cent in situ hybridization to metaphase chromosomes. This extensiveanalysis led to the establishment of a minimal tiling path of clones thatappeared to be nonchimeric, stable, and free of deletions. The resulting;4-Mb YAC contig encompasses an 8–9-centimorgan interval definedby D11S935 and D11S1776. To increase the map’s resolution and utilityfor gene identification experiments, the minimal set of YACs was thenused to identify plasmid-based genomic clones corresponding to theregion. DNA samples from YACs purified by pulsed field gel electro-

phoresis were radiolabeled and used to screen high density filter gridsof a human total genomic DNA BAC library (Research Genetics, Hunts-ville, AL), as well as a human chromosome 11-specific cosmid library(39). A subset of the detected clones could subsequently be ordered andoriented by STS content mapping. Overlap relationships between theseclones were defined using DIRVISH, and gaps in the contigs were closedby end walking.

BAC clones covering the region were subjected to exon trapping (40)and cDNA selection rounds using a commercial kit (CLONTECH, PaloAlto, CA) to search for coding sequences. Full-length cDNAs corre-sponding to exon-trapped products or selected cDNAs were cloned,sequenced, and assigned letters as they were identified. cDNA clonescontaining the full-length J ORF were obtained from a human prostatepDR2 cDNA library (CLONTECH).

Computer-aided Sequence Analysis—Partitioning of protein se-quences into globular and nonglobular segments was performed usingthe local compositional complexity measures implemented in the SEGalgorithm run with the following optimized parameters: window length,45; trigger complexity, 3.4; extension complexity, 3.8 (41). Protein database searches were performed using the PSI-BLAST program (42).

Phylogenetic analysis was performed on a continuous 127-amino aciddata set derived by concatenating the conserved ETS and SAM/SPMdomains. Positions containing gaps within these domains were deletedsitewise. Hypotheses of phylogenetic relationships were constructed usingneighbor joining, a distance-based method, and the maximum likelihoodmethods. Support for neighbor joining topologies was evaluated usingbootstrap analysis with 1000 replicates, and support for maximum like-lihood tree interference utilized quartet puzzling reliability values from10,000 puzzling steps (analogous to bootstrap replicates) (43). The dis-tance matrix in the neighbor joining tree construction was calculated bytwo methods: the simple proportional difference and Lake’s paralineardistance because of its robustness of interference on data sets havingsignificantly unequal rate effects (44). The calculation of the distancematrices, neighbor joining interference of tree topology, and the subse-quent bootstrap analysis were implemented in DAMBE (45). The neigh-bor joining tree derived from the maximum likelihood distance matrixwas used to compute the parameters for the models of substitution andrate heterogeneity. A model of rate heterogeneity with one invariant andeight g-distributed rate categories was used.

DNA Clones—A cDNA clone containing the full-length J (EHF/ESE-3b) ORF, was subjected to PCR to construct different clones containinga FLAG epitope either at the 59 (clone C7132) or at the 39 (clone F5)ends, that were subcloned into pRC/CMV (Invitrogen, Carlsbad, CA). Togenerate a full-length GST fusion bacterial expression construct (clonepGEX B4), a PCR fragment containing the full-length ORF was sub-cloned into the PGEX-20T vector, a derivative of PGEX-2T (AmershamPharmacia Biotech). A similar approach was used to generate a His-tagged full-length construct in the pProEx-HTB vector (Life Technolo-gies, Inc.). An ESE-1a/ESX expression vector was constructed by per-forming PCR on an expressed sequence tag containing the full-lengthESE1a/ESX ORF (expressed sequence tag 78004; GenBankTM accessionnumber AA 3667460), also inserting a FLAG epitope at the 59-end, andsubcloned into pRC/CMV. The murine ETS-2 and the oncogenic Ras61L expression vectors, the human collagenase (MMP-1) (21200 to163), the (Py)2 D56 dE Fos, (PyEts)2 D56 dE Fos, and the (Py 1 6F)2 D56dE Fos luciferase reporter constructs have been described previously(46–48). To make the TIMP-1 D56 dE Fos luciferase reporter, the HT-1sequence (see electrophoretic mobility shift assay probes below) wassubcloned upstream of the D56 dE Fos minimal promoter (48) drivingthe luciferase reporter gene. A 2635 to 15 human matrilysin promoterfragment (49) was obtained through PCR amplification of total humangenomic DNA and subcloned into a high copy number derivative ofp19Luc (50).

Cell Culture, Transfections, and Luciferase Assays—The colon HCT15, COLO 205, prostate LNCaP FGC 10, and lung NCI H-1299 adeno-carcinoma cell lines (ATCC, Manassas, VA) were cultured in RPMI1640, supplemented with 10% heat-inactivated fetal bovine serum, 10mM L-glutamine, 100 IU/ml penicillin, and 100 mg/ml streptomycin (LifeTechnologies). COS-7 cells (ATCC) were cultured in Dulbecco’s modifiedEagle’s medium (Life Technologies) supplemented as above. Phorbol12-myristate 13-acetate (PMA) was purchased from Sigma. Cells werekept at 37 °C, in a humidified atmosphere containing 5% CO2, 95% air.

Plasmids for transfections were isolated by alkaline lysis, RNase Atreatment, and polyethyleneglycol precipitation (51), followed by a sin-gle centrifugation through a CsCl gradient. HCT-15 cells were trans-fected by using LipofectAMINE (Life Technologies) as recommended bythe manufacturer. After 5 h, the transfection mixture was removed, andRPMI 1640 containing 0.5% fetal bovine serum was added to the cells.

ESE-3 Is a Transcriptional Repressor20398

by guest on March 19, 2020

http://ww

w.jbc.org/

Dow

nloaded from

After 24 h, cells were treated with 10 ng/ml PMA for 12 h, and theluciferase activity of the cell extracts was analyzed in a 3010 Luminom-eter (Analytical Bioluminiscence, Ann Arbor, MI) using a commerciallyavailable reagent kit (Analytical Bioluminiscence). For the Ras 61Lco-transfections, cell lysates were collected 24 h after the addition offresh medium containing 0.5% fetal bovine serum. Relative luciferaseunits were normalized according to the protein concentration of theextracts. The transfection experiments shown were done in triplicateand repeated at least three different times.

Northern Analysis—A full-length J cDNA fragment was radiolabeledwith 32P using the Prime-It kit as recommended by the manufacturer(Stratagene, La Jolla, CA). Membranes containing polyadenylated hu-man tissue mRNAs (CLONTECH) or 20 mg of total RNA extracted fromepithelial cell lines were hybridized to the 32P-labeled DNA probe byusing RapidHyb (Amersham Pharmacia Biotech) as recommended bythe manufacturer. Filters were washed up to 0.13 SSC, 0.1% SDS at65 °C (13 SSC: 0.15 M NaCl, 15 mM sodium citrate), analyzed in aStorm PhosphorImager using ImageQuant software (Molecular Dy-namics, Inc., Sunnyvale, CA), and digital images were finally processedin Adobe Photoshop software (Adobe Systems, San Jose, CA) to elabo-rate the figures.

Recombinant Protein Production—Bacterial expression constructswere transformed into the Escherichia coli strain BL21(DE3), and cul-tures at A600 5 0.5–0.6 were induced with 0.5 mM isopropyl-b-thioga-lactoside, for 4 h at 37 °C. His tag full-length ESE-3 was purified usingthe HiszBind resin and buffer kit, following the manufacturer’s protocol(Novagen, Madison, WI). GST-full-length ESE-3 was purified on gluta-thione-Sepharose 4B resin according to the manufacturer’s protocol(Amersham Pharmacia Biotech).

Generation of Monoclonal Antibodies—Harlan Sprague-Dawley ratsimmunized with denatured purified His-tagged ESE-3 were tested forESE-3-reactive antisera using Western blots. Rat 362, whose serumshowed the highest immunoreactivity, was sacrificed, and spleen cellswere fused to the SP2 mouse myeloma (ATCC) using standard proce-dures (52). Positive cloned hybridomas were selected by enzyme-linkedimmunosorbent assay and were further tested for specificity by per-forming sequential immunoprecipitations and Western blots (52) withthe monoclonal antibodies (mAbs) and the anti-FLAG antibody (Sigma)on cell extracts from COS7 cells transfected with EHF/ESE-3b, ESE1a/ESX, and ETS-2 and by using immunocytochemical analysis on theprostate cell line LNCaP FGC 10. The mAb 5A5.5 was shown to detectEHF/ESE-3b with high specificity and did not show any cross-reactivitywith the closest human homolog ESX/ESE-1. The epitope recognized bythe 5A5.5 mAb is not included in either conserved domain (ETS orSAM) of the EHF/ESE-3b protein (not shown).

Immunohistochemistry—Neutral buffered formalin-fixed, paraffin-embedded human tissues were sectioned (6 mm), baked on Superfrostglass slides (Fisher) at 65 °C for 2 h, rehydrated, and treated withantigen unmasking solution as recommended by the manufacturer(DAKO, Carpinteria, CA). Immunohistochemistry was carried out byusing either the mAb 5A5.5-containing culture supernatant or the ratmyeloma SP2 culture supernatant as a control, followed by incubationwith a biotinylated mouse anti Rat mAb (DAKO) and a streptavidin-horseradish peroxidase conjugate (DAKO). Detection was performedwith the Immunopure Metal Enhanced DAB Substrate Kit (Pierce), andafter the detection reaction was completed (10 min to 2 h), slides weresubjected to a light hematoxilin stain (Harris hematoxilin 1:10 in waterfor 20 s), rinsed in tap water, dehydrated, and mounted with DPX(Fluka, Buchs, Switzerland). Digital images were captured with a Po-laroid DMC1 digital camera (Polaroid, Cambridge, MA) and importedinto Adobe Photoshop software (Adobe Systems) to elaborate thefigures.

Electrophoretic Mobility Shift Assays—Nuclear protein extracts wereprepared, and electrophoretic mobility shift assays were performedessentially as described (53). After electrophoresis, gels were dried andanalyzed as described above for Northern blots. The sequences for theelectrophoretic mobility shift assay probes containing the ETS sites on

the ESE-3 promoter and their variants are shown in Fig. 5D. The fullsequence of the ETS-AP-1 tandem on the human TIMP-1 gene promoter(HT1) is TGGGTGGATGAGTAATGCATCCAGGAAGCCTGGAGGCC-TGTGGTTT (sites are underlined) (18), and the ESE1a/ESX bindingsite at the neu/erb2 gene (ESX) was as described (30).

RESULTS

Isolation of ETS-related Genes at Human Chromosome11p13—As part of a genomic mapping and sequencing study(38), we used a combination of exon trapping and cDNA selec-tion to isolate novel genes at 11p13 within a genomic intervalcentered on the catalase gene and flanked proximally by CD44and distally by CD59. Two of the genes identified, initiallytermed I and J (38), potentially coded for ETS domain proteins.Because of the role of ETS domain transcription factors in theregulation of multiple processes including cell division, migra-tion, and transformation (2–5, 33) and the known association ofthis region to LOH in several adenocarcinomas (see Refs.34–37 and references therein), these genes were investigatedin more detail. During the course of this study, I has beenreported as ELF5 (31) and as ESE-2 (26), and J has beenreported as EHF and ESE-3b (27, 28). For simplicity, we willrefer to I as ESE-2, and to J as ESE-3.

ESE-2 and ESE-3 are transcribed in opposite directions(“head-to-head”) and are located ;110 Kb apart, with ESE-3being more proximal (Fig. 1). Their close proximity and se-quence similarity suggest that these two genes arose as a resultof ancestral gene duplication. The 39-end of the catalase gene islocated ;7 kb from the 39-end of ESE-2. The microsatellitemarkers that define the ESE-3 locus are D11S2008, situated;2.5 kilobase pairs upstream of exon A, and D11S907, on thesecond intron and at 0.3 kilobase pairs from exon C. Oursequence differs slightly from those reported previously, and itis available from GenBankTM under accession number AF212848.

Expression of the ESE-3-specific mRNAs—Northern blotanalysis showed two major ESE-3 mRNA transcripts of 5.6 and1.3 Kb that presumably arise from alternative utilization of twodifferent consensus polyadenylation signals. An additionaltranscript of 4.6 kb could also be detected. ESE-3 mRNAs werestrongly expressed in a variety of tissues with significant epi-thelial content including trachea, colon, pancreas, prostate,and lung and were absent or below the level of detection innonepithelial tissues (Fig. 2A), showing a wider distributionthan that previously reported for EHF or ESE-3 (27, 28). Fromthe tissues analyzed, ESE-3 mRNAs appeared to be morewidely expressed in tissues with high epithelial content thanESE-I/ELF5/ESE-2 (not shown).

Expression analysis in epithelial cell lines showed that theprostate carcinoma cell line LNCaP FGC 10 expressed highlevels of ESE-3 mRNAs, followed by the colon adenocarcinomacell lines COLO 205 and HCT 15. ESE-3 mRNA transcriptswere absent or below the level of detection in the lung adeno-carcinoma cell line NCI H-1299 (Fig. 2B).

ESE-3 Is a Nuclear Protein Expressed in Differentiated Epi-thelial Cells—To evaluate the expression of the ESE-3 protein,a mAb was generated (see “Experimental Procedures” for de-tails). Immunoprecipitation and Western blot analyses showed

FIG. 1. Physical map of the ESE-3genomic region. Microsatellite markersare listed across the top. The lines imme-diately below the markers depict YACclones. Below the YAC clones are BACand cosmid clones. Genes are shown atthe bottom, and the arrows indicate thedirection of transcription.

ESE-3 Is a Transcriptional Repressor 20399

by guest on March 19, 2020

http://ww

w.jbc.org/

Dow

nloaded from

that the ESE-3 protein migrated as a Mr 5 33,000 singlepolypeptide in both LNCaP FGC 10 and ESE-3-transfectedCOS-7 cell protein extracts, and it was absent in mock-trans-fected COS 7 cell extracts (not shown). Immunohistochemistryanalysis on several human tissues revealed that ESE-3 expres-sion was restricted to the nuclei of highly differentiated epithe-lial cells, particularly those with secretory functions (Fig. 3).Expression was highest in the serous and mucous glands, andsecretory ducts of the airways (Fig. 3, A and B), the secretoryepithelium of the seminal vesicle (Fig. 3C), the salivary glands(Fig. 3D), and the mucus-producing cells in the crypts ofLieberkuhn in the colon (Fig. 3, E and F). Moderately highlevels of expression were also observed in the basal cells andto a lesser extent in the secretory epithelium of the prostate(Fig. 3G).

In the squamous nonkeratinized epithelium of the esopha-gus, the expression of ESE-3 was evident in the differentiatinglayers, while the basal cells, undergoing active proliferation,were negative (Fig. 3H). Expression of ESE-3 could be alsodetected in the columnar epithelial lining of the airways (Fig.3A), and it was absent or below the level of detection in thelining epithelium of the colon (Fig. 3E).

ESE-3 Expression Is Lost during Epithelial Carcinogenesis—Previous reports have shown an association between oncogen-esis and elevated levels of expression of ETS proteins (see Refs.3, 30, 33, and references therein). However, ESE-3 expressionwas shown to be restricted to differentiated epithelial cells andabsent in proliferating cells. This was clearly evident in theexpression pattern observed in the squamous epithelium ofthe esophagus (Fig. 3H). Additionally, expression of ESE-3 inthe prostate was highest in the basal cells, a cell type that islost in malignant prostate carcinomas (54). Consistently, weobserved loss of ESE-3 positive cells in malignant prostatecarcinomas (not shown).

Based these observations, ESE-3 immunoreactivity wasmonitored in other epithelial carcinomas. Using the ESE-3-specific mAb, we analyzed protein expression in samples ofbladder adenocarcinoma (Fig. 4A), oral epithelium carcinoma(Fig. 4B), and breast ductal carcinoma (C and D). As shown,ESE-3 protein expression was restricted to the normal epithe-lium and was clearly excluded from the transformed epithe-lium. Both normal and transformed epithelium were evident inall samples, allowing us to exclude sample quality or processingas an explanation for the loss of immunoreactivity.

The ESE-3 Protein Binds Selectively to a Small Subset ofETS Binding Sites—Since a number of transcription factorsregulate their own transcription, we examined the ESE-3 pro-moter for putative ETS binding sites. Two potential ETS bind-ing sequences were found: one at 2240 and another at 2730

nucleotides from the major transcription start site.2 Both Histag (Fig. 5A, lane 1) and GST-full-length ESE-3 (Fig. 5A, lane 3)could bind the 2730 site with high affinity, as evidenced by acompetition profile using cold oligonucleotide as competitor(not shown). The mobility of these nucleoprotein complexes wasaffected by the addition of an anti-ESE-3 polyclonal antiserum(a362; see “Experimental Procedures” for details) (Fig. 5A,lanes 2 and 4).

A number of natural known ETS binding sites were com-pared with the 2730 site and with several sequences derivedfrom the 2730 site by single nucleotide substitutions. Full-length GST-ESE-3 displayed different relative affinities for thesites tested (Fig. 5B), suggesting the consensus A . G:C .G:A:G:G:A:A:G:T, as a high affinity binding site for ESE-3(Fig. 5D).

To assess the binding of native ESE-3 protein, nuclear ex-tracts from the LNCaP FGC 10 prostate epithelial adenocarci-noma were tested for binding to the 2730 m2 probe (see below)and showed a number of nucleoprotein complexes (Fig. 5C, lane1). The addition of the anti-ESE-3 polyclonal serum to thebinding reactions selectively altered the migration of two nu-cleoprotein complexes (Fig. 5C, lane 2). The migration profile ofthese DNA-protein complexes was similar to that of purifiedHis-tagged ESE-3 (Fig. 5A, lane 1), indicating that both bandsarose from a single polypeptide. The slower migrating com-plexes, which were not recognized by the anti ESE-3 serum,were shown to bind specifically to the 2730 m2 sequence, andthey did not appear when the 2730 m1 oligonucleotide wasused as a probe in the binding reactions (not shown). Thus,these complexes may result from the binding of other ETSproteins present in the extracts. The band near the top of thegel only appeared in the lanes that contained preimmune se-rum and not in those containing antiserum or no serum at all(not shown), indicating that it is an artifact caused by thepreimmune serum.

ESE-3 Is a Context-dependent Transcriptional Repressor—The polyoma virus enhancer ETS site contains a very similarsequence to the one that we experimentally determined as ahigh affinity binding site for ESE-3 and is contained within aregulatory cis module termed the oncogene response element(ORE). Transcription from the ORE is synergistically stimu-lated through the cooperation of ETS transcription factors andAP-1 in response to Ras signaling (48). To analyze the tran-scriptional role of ESE-3 in the ORE-mediated transcription,HCT-15 colon adenocarcinoma cells were transfected with anESE-3 expression vector and the (Py)2 D56 dE Fos luciferasereporter, either in the presence or the absence of an oncogenic

2 A. Tugores, unpublished observations.

FIG. 2. Expression of the ESE-3mRNA. The expression of ESE-3 mRNAswas analyzed by Northern blot hybridiza-tion on human tissues (A) and the epithe-lial cell lines COLO 205, LNCaP FGC 10,HCT-15, and NCI-H1299 (B). The migra-tion of the molecular size standards isshown on the left. The arrows indicate thethree major mRNA transcripts observed.

ESE-3 Is a Transcriptional Repressor20400

by guest on March 19, 2020

http://ww

w.jbc.org/

Dow

nloaded from

FIG. 3. ESE-3 is a nuclear protein expressed in differentiated epithelial cells. Human tissues were subjected to immunohistochemicalanalysis by using the ESE-3-specific monoclonal antibody 5A5.5. Brown staining reveals immunoreactivity, and blue staining corresponds tohematoxilin nuclear stain. A, bronchial columnar lining epithelium (top) and secretory ducts; B, bronchial secretory ducts and mucous and serousglands; C, seminal vesicle; D, salivary glands; E, colon lining and crypts of Lieberkuhn; F, crypts of Lieberkuhn (transversal section); G, prostategland; H, esophagus stratified epithelium. Bar, 100 mm.

ESE-3 Is a Transcriptional Repressor 20401

by guest on March 19, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Ras (Ras 61L) expression construct. The (Py)2 D56 dE Fosluciferase reporter contains two ORE modules upstream of aminimal Fos promoter driving the luciferase gene (48). HCT-15cells were used in this study due to their low, albeit detectable,endogenous levels of ESE-3 mRNA expression (see Fig. 2B). Asshown (Fig. 6A), co-expression of ESE-3 repressed the onco-genic Ras-mediated transcriptional activation of the polyomavirus ORE-driven reporter. Similar results were obtained whena constitutively active membrane-targeted form of the nucleo-tide exchange factor SOS (55) was co-transfected instead of theoncogenic Ras mutant (not shown).

The human interstitial collagenase (MMP-1) promoter hasbeen also shown to be a Ras-responsive promoter by virtue ofthe presence of ETS and AP-1 cis-elements, which act syner-gistically to activate Ras-dependent transcription of the gene(15, 16). Similarly to what was observed previously with thesynthetic ORE reporter, ESE-3 was able to repress the 21200to 163 collagenase promoter response to Ras (Fig. 6B). In bothcases, expression of ETS-2 resulted in enhanced expression ofthe reporter in these cells (Fig. 6, A and B).

Similar results were obtained when the tumor promoterPMA was used as an activator of the transcription of thereporter constructs. Phorbol esters activate protein kinase C,which in turn is able to phosphorylate Raf leading to activationof the MAPK cascade (56). ESE-3 repressed, in a dose-depend-ent manner, the PMA-induced transcriptional activation ofboth reporters (Fig. 6, C and D). Co-transfection of ETS-2 inthis setting resulted neither in an increase nor in a decrease ofluciferase activity (not shown).

To determine the specificity of the transcriptional repressionmediated by ESE-3, other ETS/AP-1-responsive promoterswere analyzed. Expression of ESE-3 did not repress transcrip-tion stimulated by the tandem ETS/AP-1 from the humanTIMP-1 gene promoter cloned upstream of the minimal D56 dEFos promoter (Fig. 6E). This is consistent with the low affinityof ESE-3 for the ETS cis-element present in the TIMP-1 pro-moter (see Fig. 5).

The epithelium-specific metalloprotease matrilysin (MMP-7)also contains ETS and AP-1 cis-elements in its promoter, andits expression, like that of MMP-1, is also induced by PMA,tumor necrosis factor-a, interleukin-1, and epidermal growthfactor (16, 49). The activity of matrilysin has been associatedwith normal epithelial cell migration, although it is also ex-pressed at high levels in nonmigrating glandular epithelialcells (57), which also express ESE-3 (this study). In agreementwith their tissue co-localization, we found that ESE-3 wasunable to repress transcription from the matrilysin promoter(Fig. 6F) despite the presence of a high affinity binding site forESE-3, suggesting that repression might be context-dependent.

The context-dependent repression by ESE-3 is further con-firmed by deletion of one tandem of ETS/AP-1 sites or bymutation of the AP-1 cis-elements in the context of the OREreporter. Because it is impossible to eliminate the AP-1 site inthe polyoma ORE without affecting the ETS site, since theyoverlap, we used the (Py 1 6F)2 reporter containing the ETS/AP-1 tandem of the polyoma ORE, where the ETS and AP-1binding sites are separated by 6 nucleotides (48). As shown,ESE-3 was still able to repress the PMA-mediated trans-activation of the (Py 1 6F)2 reporter (Fig. 6G). Although(Py 1 6F)2 reporter is, on average, 10–20-fold less active thanthe parental(Py)2 D56 dE Fos luciferase reporter, we feel that,qualitatively, they both reflect similar results. It is also possi-ble that ESE-3 repressed less efficiently in this setting. In anycase, ESE-3 was unable to repress a reporter that contained asingle copy of the Py 1 6F module (Fig. 6H). Mutation of theAP-1 cis-elements in this context, represented by the (PyEts)2D56 dE Fos luciferase reporter (48), showed that repressionwas dependent on the presence of functional ESE-3 bindingsites in combination with AP-1 cis-elements. In fact, ESE-3then behaved as an activator at low concentrations (Fig. 6I),although this activity was very moderate in comparison withthe related polypeptide ESE-1/ESX (Fig. 6J). On the otherhand, synthetic promoters where the ETS binding sites wereeliminated in this context, leaving multimers of the AP-1 cis-

FIG. 4. ESE-3 immunoreactivity is lost during carcinogenesis. Sections from bladder (A) and oral epithelial (B) adenocarcinomas andbreast ductal carcinoma (C and D) containing normal tissue adjacent to the carcinoma cells were subjected to immunohistochemical analysis byusing the ESE-3-specific monoclonal antibody 5A5.5. The black arrows indicate the normal ESE-3-expressing epithelium, while the red arrowsshow the transformed epithelium, which does not show ESE-3 immunoreactivity. Bar, 100 mm in A and B and 50 mm in C and D.

ESE-3 Is a Transcriptional Repressor20402

by guest on March 19, 2020

http://ww

w.jbc.org/

Dow

nloaded from

elements (33AP-1, and 63AP-1; see Ref. 48), were not re-pressed by ESE-3 (not shown), indicating that DNA binding tocognate cis-elements is essential to mediate this activity.

Previous reports indicate that overexpression of ETS-2 maylead to transcriptional repression (58). In our experiments,co-transfection of ETS-2 did not lead to a decrease of transcrip-tion in any of the cases analyzed (not shown). Western analysisalso revealed that both the ETS-2 and the ESE-3 expressionvectors were expressed at comparable levels (not shown), andtherefore we excluded the possibility that the repression ob-served upon co-transfection with ESE-3 is due to nonspecificrepression.

DISCUSSION

The genes coding for ELF5/ESE-2 and EHF/ESE-3 werefound during an extensive search for genes localized at humanchromosome 11p13 (38), in a region that has been previouslyassociated with LOH in a subset of epithelial carcinomas (seeRefs. 34–37 and references therein). In the case of some lungcarcinomas analyzed, LOH at this locus has been associatedwith a higher patient mortality rate, a direct measurement ofthe tumors’ metastatic potential (36), thus suggesting that thisregion may contain genes required for the control of cellularmigration. Genetic mapping of this LOH has been mostly as-sociated with the Catalase locus (35–37), which we found inclose proximity to ESE-2, and ESE-3.

The expression of ESE-3 is widespread in tissues with highepithelial content. We show that ESE-3 mRNA has a much

wider expression profile than previously reported (27, 28), andcontains an additional 1.3-Kb mRNA species. We did not iso-late the short splice variant reported by others, indicating thatit was underrepresented in the library used. Using a newlygenerated specific monoclonal antibody, we show that theESE-3 protein is specifically expressed in differentiated epithe-lial cells. This is most evident in the stratified squamous epi-thelium of the esophagus, where ESE-3 is absent in the basalproliferating stem cells, and appears as soon as they differen-tiate. Importantly, ESE-3 expression is lost in a number ofepithelial carcinomas, including bladder, oral squamous,breast ductal (Fig. 4), and prostate epithelial carcinomas (notshown). Certainly, it can be argued that the decreased expres-sion of ESE-3 is not an essential pathogenic event and thatother oncogenic mechanisms lead to the loss of ESE-3 expres-sion. Indeed, although our data are suggestive altogether, wedo not have direct evidence correlating the loss of ESE-3 ex-pression with cancer progression. Nonetheless, the restrictedassociation of ESE-3 with differentiated epithelial cells indi-cates that, at least, could be a marker for malignancy, since thecell fate associated with ESE-3 expression appears to be nolonger determined. Others have reported variable EHF/ESE-3mRNA expression in several epithelial carcinomas analyzed(27), although increased mRNA expression does not necessarilyhave to correlate with elevated protein levels. This seems to bethe case of the prostate LNCaP FGC 10 adenocarcinoma cellline, in which the ESE-3 mRNA levels are the highest among

FIG. 5. DNA binding specificity ofthe ESE-3 protein. A, electrophoreticmobility shift assays were used to monitorthe binding of E. coli expressed His-tagged full-length ESE-3 (His-ESE-3;lanes 1 and 2) or GST-full-length ESE-3(GST-ESE-3; lanes 3 and 4) fusion pro-teins (10–100 fmol) to a 32P-labeled dou-ble-stranded oligonucleotide (40 fmol; 3nM final concentration) containing theETS consensus binding site present at po-sition 2730 on the ESE-3 gene promoter.Specificity of the nucleoprotein complexeswas determined by using the 362 rat an-tiserum against ESE-3 (362 aS; lanes 2and 4). B, binding of E. coli expressed andpurified GST-ESE-3 to double-strandedoligonucleotides containing the following:the ETS consensus sites at 2240 (240;lane 1) and at 2730 (730; lanes 2 and 6)on the ESE-3 gene promoter; a variant ofthe 2730 site (730 m1; lane 3); the ETSand AP-1 sites on the human TIMP-1gene promoter (HT-1; lane 4); the ESE1a/ESX binding site at the human NEU gene(ESX; lane 5); and the variants of the 730sequence (730 m2 (lane 7), 730 m3 (lane8), 730 m4 (lane 9), and 730 m5 (lane 10)).C, detection of ESE-3 binding activity inthe prostate LNCaP FGC 10 epithelialcell line. Twenty mg of nuclear proteinswere incubated with the 730 m2 32P-la-beled oligonucleotide in the presence ofthe 362 rat anti ESE-3 antiserum (362aS, lane 2), preimmune serum (362 PI;lane 1), or a 50-fold molar excess of coldcompetitor oligonucleotide (lane 3). Thearrows show the nucleoprotein complexesthat are reactive to the anti-ESE-3 anti-serum. D, sequences of the probes used,relative binding affinities, and proposedconsensus.

ESE-3 Is a Transcriptional Repressor 20403

by guest on March 19, 2020

http://ww

w.jbc.org/

Dow

nloaded from

the cell lines tested, but the protein is barely detectable both inWestern blots (not shown) and in mobility shift assays (Fig. 5).More tumor samples will have to be analyzed to clarify thisissue.

Functionally, ESE-3 is a transcriptional repressor of a subsetof genes that contain ETS and AP-1 cis-elements within theirpromoters and have been shown to be downstream targets thatare positively regulated by MAP kinase signaling cascades.ESE-3-mediated repression of these promoters is dependent onthe presence of high affinity binding sites, which we havedefined in good agreement with those described by others (28).The restricted binding affinity of this repressor to a limitednumber of ETS cis-elements, combined with specific promoterrequirements, indicates that ESE-3 is only a repressor for aspecific subset of ETS/AP-1-responsive genes. These includethe interstitial collagenase gene (MMP-1) promoter, also ob-served by others (27), and a dimer of the oncogene responseelement of the polyoma virus enhancer, both well characterizedRas-responsive genes. Alternatively, ESE-3 may also modulatethe transcription of genes downstream of other pathways thatare activated via phorbol esters and that are only secondarilylinked to the main Ras pathway.

The restricted expression pattern of ESE-3 in the squamousstratified epithelium of the esophagus may provide insightsinto the physiological role of this repressor. Stratified epitheliahave served as a paradigm for the study of epithelial differen-tiation (59). In this dynamic system, a population of self-renew-ing basal stem cells give rise to committed, nonproliferatingprecursors that undergo progressive differentiation as they arepushed toward the lumen, where they finally delaminate (60).The molecular mechanisms that govern the transition from aproliferating basal cell to a committed nondividing precursorare of great importance for deciphering the etiology of cancer. Afailure to adopt the committed fate by the basal cell may lead tothe development of adenomas or carcinomas, giving rise tomost solid tumors.

Both autocrine and paracrine factors are known to maintainthe proliferative potential of the basal cells. Some of thesefactors, such as transforming growth factor-a, keratinocytegrowth factor, and fibroblast growth factors, are ligands fortyrosine kinase receptors and thus trigger the activation ofsignaling cascades involving MAP kinases. These signalingcascades are reiteratively used throughout the differentiationof these cells, since they activate the expression of differentia-tion markers in the upper layers of the epidermis (59). Thisimplies that additional factors are required to modulate appro-priate response (e.g. proliferation versus differentiation) toMAP kinase signaling cascades.

As mentioned earlier, transcription factors of the ETS pro-tein family are good candidates as modulators of the nuclearresponse to MAP kinase signaling cascades. Among these, fac-tors belonging to the ESE subgroup seem to play an importantrole in the activation of epithelium-specific genes (21, 24–29).Indeed, a number of genes whose expression is associated withepithelial cell differentiation contain ETS cis-elements withintheir promoters, thereby supporting this view (19–23). Both

HCT 15 cells were also transfected with increasing amounts of theESE-3 (C–I) or ESE-1/ESX expression construct (J) (0.1 mg, hatchedbar, and 0.5 mg, gray bar) as indicated, together with the (Py)2 D56 dEFos luciferase (C), human intersticial collagenase (MMP-1) (D), TIMP-1D56 dE Fos luciferase (E), matrilysin (F), (Py 1 6F)2 D56 dE Fosluciferase (G), Py 1 6F D56 dE Fos luciferase (H), and (PyEts)2 D56 dEFos luciferase (I and J) luciferase reporter constructs and treated withPMA where indicated. Luciferase activity was measured from cell ly-sates, and values were normalized to the protein concentration of theextracts. Experiments shown are representative of three independentexperiments, all of them performed in triplicate. Error bars, S.D.

FIG. 6. ESE-3 represses the Ras- and the phorbol ester-depend-ent trans-activation of ETS/AP-1-responsive genes. HCT-15 colonadenocarcinoma cells were transfected with either (Py)2 D56 dE Fosluciferase (A) or the human collagenase Luc (B) luciferase reporterconstructs (1 mg/30 mm dish) and with eucaryotic expression constructsencoding murine ETS-2 (hatched bar) (0.5 mg/30-mm dish) or humanESE-3 (clone pRC/C7132; 0.5 mg/30-mm dish) (gray bar) in the presenceof a Ras 61L expression construct (0.5 mg/30-mm dish) where indicated.

ESE-3 Is a Transcriptional Repressor20404

by guest on March 19, 2020

http://ww

w.jbc.org/

Dow

nloaded from

ESE-1 and ESE-2 are expressed at later stages of keratino-cyte differentiation (21, 26), indicating that they are unlikelyto play a role in the determination of the early differentiatedspinous precursor in stratified squamous epithelia. Thus, itappears that, among the ESE family members described todate, ESE-3 is expressed the earliest in the differentiatedprecursors. Based on the functional analysis and the expres-sion profile for this gene, as presented herein, we hypothesizethat ESE-3 could play a role in the determination of the earlydifferentiated fate in stratified squamous epithelia. SinceESE-3-mediated repression may only occur at a specific set ofMAP kinase target promoters, while allowing the transcrip-tion of others, we conclude that ESE-3 might be an importantfactor in the nuclear specific response to MAP kinase signal-ing cascades.

The molecular mechanisms underlying the differentiation ofsecretory epithelial cells is less known, although recent reportsindicate that epidermal growth factor receptor signaling is alsoinvolved in the determination of goblet cells in the airways andsubsequent mucus production (61). The epithelium-specificETS factors ESE-1/ESX/JEN/ERT/ELF3 and ELF5/ESE-2 andmore recently also EHF/ESE-3 have also been implicated in theselective transcriptional activation of genes specific to glandu-lar epithelia (26, 28). Further work will be needed to under-stand how the interplay between these genes orchestrates geneexpression in secretory epithelia and their role (if any) in theinduction of secretory fates.

The role of ESE-3 as a repressor resembles that described forAop/Yan, an ETS protein in D. melanogaster (6). Additionally,ESE-3 and Aop/Yan not only share a similar structure but alsorecognize the same DNA sequences, since the polyomavirusenhancer in the context of the interstitial collagenase genepromoter has been used to demonstrate repression by Aop/Yan(62). However, ESE-3 does not have any consensus extracellu-lar signal-regulated kinase phosphorylation sites; nor is itphosphorylated by extracellular signal-regulated kinase or Jun

N-terminal kinase 1 in vitro,3 suggesting that, unlike Aop/Yan,ESE-3 may not be regulated post-translationally by direct MAPkinase phosphorylation.

The ESE family has been shown to contain two distinguish-able domains, namely the ETS DNA binding domain at thecarboxyl terminus and an additional domain that has beenrecognized in some ETS-domain proteins as the A domain, orthe Pointed domain (after the Drosophila ETS protein Pointed;reviewed in Ref. 3), located at the amino terminus. Recentdetailed analysis of conserved motifs and comparison of three-dimensional structures established the similarity of the A do-main to the SAM/SPM (sterile a motif/scm-polyhomeotic motif)domains, also involved in protein-protein interactions (3, 63).In the case of the ETS proteins belonging to the TEL subgroup,this domain has been shown to mediate homotypic interactions(14, 64), which are required for repression (65–67), but we havenot observed such a role in the case of ESE-3.3

The semiautonomous nature of the SAM/SPM domain andits apparent rapid and diverse evolution (68) impose a chal-lenge in defining the phylogenetic relationship between ESE-3and related ETS proteins in other species. In particular, theETS domain sequence in Drosophila with highest similarity tothat of ESE-3 is found in E-74 (GenBankTM accession numberAAA28493), which does not contain a SAM/SPM domain. It isfor this reason that we do not favor the ELF nomenclature(E-74-like factor) for this group. Phylogenetic analysis of theETS family members that contain ETS and SAM/SPM domainsusing both domains in the analysis clearly indicates that Aop/Yan and the TEL proteins are members of a monophyleticgroup (66, 28) also containing the ESE and the PDEF/ETS 98Bfamilies. This clade of ETS proteins clusters to the exclusion ofa second monophyletic group containing Drosophila Pointed P2and its vertebrate orthologs c-ETS-1 and c-ETS-2 (Fig. 7).

Functionally, TEL-1, TEL-2b, ESE-1, ESE-2, and EHF/

3 P. S. Reddy, unpublished results.

FIG. 7. Evolutionary relationship among ETS transcription factors structurally related to ESE-3. Phylogenetic analysis on ETS familymembers that contain both ETS and SAM/SPM domains was performed, including both domains as described under “Experimental Procedures.”A star phylogeny for the maximum likelihood tree is shown. All three methods of phylogenetic interference used (see “Experimental Procedures”for details) recovered the same topology. The numbers close to the internal nodes represent the quartet puzzling support values for the branchingrelationships. The human sequences used were as follows: ESE-1a/ESX/JEN/ERT/ELF3 (AAB65824), ESE-2a/ELF5 (AAD22960.1), TEL-1 (NP001978), TEL-2b (AAD43251), c-ETS-1 p54 (P 27577), c-ETS-2 (P 15036), FLI-1 (Q 01543), human GABP a (Q 06546), and the human epithelialprostate-specific PDEF (AAC 95296). Our ESE-3/EHF cDNA sequence data are available (as ESE-J) from GenBankTM under accession numberAF 212848. The tree also includes all the Drosophila ETS proteins that contain both the ETS and SAM domains: Aop/Yan/Pok (Q 01842), PointedP2 (P 51023), ETS 21C (AAF51484.1), ELG/ETS 97D (Q 04688), and ETS 98B (AAF 56746).

ESE-3 Is a Transcriptional Repressor 20405

by guest on March 19, 2020

http://ww

w.jbc.org/

Dow

nloaded from

ESE-3 may all behave as repressors (Refs. 14, 26–28, 32, and65–67 and this study). It is tempting to speculate that theregulatory role of Aop/Yan may have diversified in vertebratesand could be represented by these genes. This regulatory rolemay have acquired additional functions during evolution, sincesome of these proteins also behave as activators, although todate, it has not been reported whether Aop/Yan is only arepressor. Since TEL-1 appears to be only expressed in meso-derm, where it appears to have an important role during he-matopoiesis and angiogenesis (69), it seems plausible that theESE subgroup could represent the function of Aop/Yan in ver-tebrate epithelia. Furthermore, we hypothesize that the func-tion of these genes could be partially redundant, as evidencedby the fact that deletion of ESE-3 coding sequences in mice didnot result in any clearly apparent phenotype.4

As Aop/Yan, transcription factors belonging to the ESE sub-group might be also involved in the election of alternative cellfates, such as the choice between proliferation and migrationversus differentiation, in response to complex environmentalcues. We propose that the role of these proteins has beenconserved throughout evolution as a nuclear switch necessaryfor the modulation of MAPK signaling cascades in epithelialcells.

Acknowledgments—The contributions of the Physical Mapping, Se-quencing, and Bioinformatics groups at Sequana were fundamental tothis work. We thank Javier Fraga for advice and Michael Karin,William McGinnis, John K. Westwick, Jose Javier Garcıa Ramırez, andDavid A. Brenner for valuable suggestions on the manuscript. J. Fraga,Craig Hauser, and M. Karin kindly provided materials used in thisstudy.

REFERENCES

1. Widmann, C., Gibson, S., Jarpe, M. B., and Johnson, G. L. (1999) Phys. Rev. 79,143–180

2. Treisman, R. (1996) Curr. Opin. Cell Biol. 8, 205–2153. Graves, B. J., and Petersen, J. M. (1998) Adv. Cancer Res. 75, 1–554. Wasylyk, B., Hagman, J., and Gutierrez-Hartman, A. (1998) Trends Biochem.

Sci. 23, 213–2165. Tan, P. B. O., and Kim, S. K. (1999) Trends Genet. 15, 145–1496. Wassarman, D. A., Therrien, M., and Rubin, G. M. (1995) Curr. Opin. Genet.

Dev. 5, 44–507. Rogge, R., Green, P. J., Urano, J., Horn-Saban, S., Mlodzik, M., Shilo, B-Z.,

Hartenstein, V., and Banerjee, U. (1995) Development 121, 3947–39588. Gabay, L., Scholz, H., Golembo, M., Klaes, A., Shilo, B-Z., and Klambt, C.

(1996) Development 122, 3355–33629. Riesgo-Escovar, J. R., and Hafen, E. (1997) Genes Dev. 11, 1717–1727

10. Hacohen, N., Kramer, S., Sutherland, D., Hiromi, Y., and Krasnow, M. A.(1998) Cell 92, 253–263

11. Scholz, H., Deatrick, J., Klaes, A., and Klambt, C. (1993) Genetics 135,455–468

12. Albagli, O., Klaes, A., Ferreira, E., Leprince, D., and Klambt, C. (1996) Mech.Dev. 59, 29–40

13. Sieweke, M. H., Tekotte, H., Frampton, J., and Graf, T. (1996) Cell 85, 49–6014. Kwiatkowski, B. A., Bastian, L. S., Bauer, T. R., Tsai, S., Zielinska-

Kwiatkowska, A. G., and Hickstein, D. D. (1998) J. Biol. Chem. 273,17525–17530

15. Crawford, H. C., and Matrisian, L. M. (1996) Enzyme Protein 49, 20–3716. Borden, P., and Heller, R. A. (1997) Crit. Rev. Eukaryotic Gene Expression 7,

159–17817. D’Orazio, D., Besser, D., Marksitzer, R., Kunz, C., Hume, D. A., Kiefer, B., and

Nagamine, D. (1997) Gene (Amst.) 201, 179–18718. Logan, S. K., Garabedian, M. J., Campbell, C. E., and Werb, Z. (1996) J. Biol.

Chem. 271, 774–78219. Zhang, M., Maass, N., Magit, D., and Sager, R. (1997) Cell Growth Differ. 8,

179–18620. Seth, A., Robinson, L., Panayiotakis, A., Thompson, D. M., Hodge, D. R.,

Zhang, X. K., Watson, D. K., and Papas, T. S. (1994) Oncogene 9, 469–47721. Andreoli, J. M., Jang, S-I., Chung, E., Coticchia, C. M., Steinert, P. M., and

Markova, N. G. (1997) Nucleic Acids Res. 25, 4287–429522. Sark, M. W., Fischer, D. F., de Meijer, E., van de Putte, P., and Backendorf, C.

(1998) J. Biol. Chem. 273, 24683–2469223. Fischer, D., Sark, M., Lehtola, M., Gibbs, S., van de Putte, P., and Backendorf,

C. (1999) Genomics 55, 88–99

24. Oettgen, P., Alani, R. M., Barcinsky, M. A., Brown, L., Akbaraly, Y., Boltax, J.,Kunsch, C., Munger, K., and Libermann, T. A. (1997) Mol. Cell. Biol. 17,4419–4433

25. Choi, S-G., Yi, Y., Kim, Y-S., Kato, M., Chang, J., Chung, H-W., Hahm, K-B.,Yang, H-K., Rhee, H. H., Bang, Y-J., and Kim, S-J. (1998) J. Biol. Chem.273, 110–117

26. Oettgen, P., Kas, K., Dube, A., Gu, X., Grall, F., Thamrongsak, U., Akbarali,Y., Finger, E., Boltax, J., Endress, G., Munger, K., Kunsch, C., andLibermann, T. A. (1999) J. Biol. Chem. 274, 29439–29452

27. Kleinbaum, L., Duggan, C., Ferreira, E., Coffey, G., Buttice, G., and Burton, F.(1999) Biochem. Biophys. Res. Commun. 264, 119–126

28. Kas, K., Finger, E., Grall, F., Gu, X., Akbarali, Y., Boltax, J., Weiss, A.,Oettgen, P., Kapeller, R., and Libermann, T. A. (2000) J. Biol. Chem. 275,2986–2998

29. Chang, J., Lee, C., Hahm, K. B., Yi, Y., Choi, S-G., and Kim, S-J. (2000)Oncogene 19, 151–154

30. Chang, C. H., Scott, G. K., Kuo, W-L., Xiong, X., Suzdaltseva, Y., Park, J. W.,Sayre, P., Erny, K., Collins, C., Gray, J. W., and Benz, C. C. (1997) Oncogene14, 1617–1622

31. Zhou, J., Ng, A., Tymms, M. J., Jermiin, L. S., Seth A. K., Thomas, R. S., andKola, I. (1998) Oncogene 17, 2719–2732

32. Brembeck, F., Opitz, O., Libermann, T., and Rustgi, A. (2000) Oncogene 19,1941–1949

33. Dittmer, J., and Nordheim, A. (1998) Biochim. Biophys. Acta 1377, F1–F1134. Weinberg, R. A. (1991) Science 254, 1138–114635. Reeve, A. E., Sin, S. A., Raizis, A. M., and Feinberg, A. P. (1989) Mol. Cell. Biol.

9, 1799–180336. Fong, K. M., Zimmerman, P. V., and Smith, P. J. (1994) Genes Chromosomes

Cancer 10, 183–18937. Shipman, R., Schraml, P., Colombi, M., and Ludwig, C. U. (1998) Int. J. Oncol.

12, 107–11138. Brooks-Wilson, A. R., Buckler, A., Cardon, L., Carey, A. H., Galvin, M., Miller,

A., and North, M. (July 11, 2000) U. S. Patent 6,087,48539. Evans, G. A., and Lewis, K. A. (1989) Proc. Natl. Acad. Sci. U. S. A. 83,

7821–782540. Buckler, A., Chang, D. D., Graw, S. L., Brook, J. D., Haber, D. A., Sharp, P. A.,

and Housman, D. E. (1991) Proc. Natl. Acad. Sci. U. S. A. 88, 4005–400941. Wootton, J. C., and Federhen, S. (1996) Methods Enzymol. 266, 554–57142. Altschul, S. F., Madden, T. L., Schaffer, A. A., Zhang, J., Zhang, Z., Miller, W.,

and Lippman, D. J. (1997) Nucleic Acids Res. 25, 3389–340243. Strimmer, K., and von Haesler, A. (1996) Mol. Biol. Evol. 13, 964–96944. Lake, J. A. (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 1455–145945. Xia, X. (2000) DAMBE: Data Analysis in Molecular Biology and Evolution,

Department of Ecology and Biodiversity, University of Hong Kong46. Pankov, R., Umezawa, A., Maki, R., Der, C. J., Hauser, C. A., and Oshima,

R. G. (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 873–87747. Angel, P, Baumann, I., Stein, B., Delius, H., Rahmsdorf, H. J., and Herrlich, P.

(1987) Mol. Cell. Biol. 7, 2256–226648. Galang, C. K., Der, C. J., and Hauser, C. A. (1994) Oncogene 10, 2913–292149. Gaire, M., Magbanua, Z., Mc Connel, S., McNeil, L., Lovett, D. H., and

Matrisian, L. M. (1994) J. Biol. Chem. 269, 2032–204050. van Zonneveld, A-J., Curriden, S. A., and Loskutoff, D. J. (1988) Proc. Natl.

Acad. Sci. U. S. A. 85, 5525–552951. Hattori, M., and Sakaki, Y. (1986) Anal. Biochem. 152, 232–23852. Harlow, E., and Lane, D. (1988) Antibodies: A Laboratory Manual, Cold Spring

Harbor Laboratory Press, Cold Spring Harbor, NY53. Tugores, A., Alonso, M. A., Sanchez-Madrid, F., and Landazuri, M. O. (1992)

J. Immunol. 269, 30789–3079754. Yang, X. J., Lecksell, K., Gaudin, P., and Epstein, J. I. (1999) Am. J. Surg.

Pathol. 23, 147–15255. Aronheim, A., Engelberg, D., Li, N., al-Alawi, N., Schlessinger, J., and Karin,

M. (1994) Cell 78, 949–96156. Bruder, J. T., Heidecker, G., and Rapp, U. R. (1992) Genes Dev. 6, 545–55657. Saarialho-Kere, U. K., Crouch, E. C., and Parks, W. C. (1995) J. Invest.

Dermatol. 105, 190–19658. Foos, G., Garcia-Ramirez, J. J., Galang, C. K., and Hauser, C. A. (1998) J. Biol.

Chem. 273, 18871–1888059. Fuchs, E. (1990) J. Cell Biol. 111, 2807–281460. Marques-Pereira, J. P., and C. P. Leblond, C. P. (1965) Am. J. Anat. 117, 73–9061. Takeyama, K., Dabbagh, K., Lee, H-M., Agustı, C., Lausier, J. A., Ueki, I. F.,

Grattan, K. M., and Nadel, J. A. (1999) Proc. Natl. Acad. Sci. U. S. A. 96,3081–3086

62. Treier, M., Bohman, D., and Mlodzik, M. (1995) Cell 83, 753–76063. Stapleton, D., Balan, I., Pawson, T., and Sicheri, F. (1999) Nat. Struct. Biol. 6,

44–4964. Jousset, C., Carron, C., Boureux, A., Quang, C. T., Oury, C., Dusanter-Fourt,

I., Charon, M., Levin, J., Bernard, O., and Ghysdael, J. (1997) EMBO J. 16,69–82

65. Lopez, R., Carron, C., Oury, C., Gardellin, P., Bernard, O., and Ghysdael, J.(1999) J. Biol. Chem. 274, 30132–30138

66. Poirel, H., Lopez, R., Lacronique, V., Della Valle, V., Mauchauffe, M., Berger,R., Ghysdael, J., Bernard, O. (2000) Oncogene 19, 4802–4806

67. Gu, X., Shin, B-H., Akbarali, Y., Weiss, A., Boltax, J., Oettgen, P., andLibermann, T. A. (2001) J. Biol. Chem. 276, 9421–9436

68. Shao, X., and Grishin, N. (2000) Nucleic Acids Res. 28, 2643–265069. Wang, L. C., Kuo, F., Fujiwara, Y., Gilliland, D. G., Golub, T. R., and Orkin,

S. H. (1997) EMBO J. 16, 4374–43834 L. Carlee, and M. Duyao, unpublished results.

ESE-3 Is a Transcriptional Repressor20406

by guest on March 19, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Chu, Alan Buckler, Spencer Emtage and Mary Kay McCormickReddy, Leone Carlée, Mathew Ronshaugen, Arcady Mushegian, Tim Watanaskul, Sunny Antonio Tugores, Jennifer Le, Irina Sorokina, A. J. Snijders, Mabel Duyao, P. Sanjeeva

Transcriptional Repressor Downstream of MAPK Signaling CascadesThe Epithelium-specific ETS Protein EHF/ESE-3 Is a Context-dependent

doi: 10.1074/jbc.M010930200 originally published online March 19, 20012001, 276:20397-20406.J. Biol. Chem. 

  10.1074/jbc.M010930200Access the most updated version of this article at doi:

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/276/23/20397.full.html#ref-list-1

This article cites 66 references, 28 of which can be accessed free at

by guest on March 19, 2020

http://ww

w.jbc.org/

Dow

nloaded from