Activation of the murine type II transforming growth factor-beta receptor gene: up-regulation and function of the transcription factor Elf-3/Ert/Esx/Ese-1.

Previous studies demonstrated that differentiation of mouse embryonal carcinoma cells leads to transcriptional up-regulation of the mouse type II transforming growth factor-beta receptor (mTbetaR-II) gene. To elucidate the molecular mechanisms regulating transcription of this gene, we isolated the 5'-flanking region of the mTbetaR-II gene and characterized its expression in F9-differentiated cells. Analysis of mTbetaR-II promoter/reporter gene constructs demonstrates that two conserved Ets-binding sites play an important role in the activity of the mTbetaR-II promoter. Importantly, we present evidence that mElf-3, a member of the Ets family, plays a key role in the activation of the mTbetaR-II promoter. Northern blot analysis reveals that the steady-state levels of mTbetaR-II mRNA increase in parallel with those of mElf-3 mRNA during the differentiation of F9 embryonal carcinoma cells. We also demonstrate that mElf-3 contains one or more domains that influence its binding to DNA. Finally, we report that a single amino acid substitution in the transactivation domain of mElf-3 reduces its ability to transactivate and elevates its steady-state levels of expression. In conclusion, our data argue that mElf-3 plays a key role in the regulation of the mTbetaR-II gene, and Elf-3 itself is regulated at multiple levels.

Previous studies demonstrated that differentiation of mouse embryonal carcinoma cells leads to transcriptional up-regulation of the mouse type II transforming growth factor-␤ receptor (mT␤R-II) gene. To elucidate the molecular mechanisms regulating transcription of this gene, we isolated the 5-flanking region of the mT␤R-II gene and characterized its expression in F9differentiated cells. Analysis of mT␤R-II promoter/reporter gene constructs demonstrates that two conserved Ets-binding sites play an important role in the activity of the mT␤R-II promoter. Importantly, we present evidence that mElf-3, a member of the Ets family, plays a key role in the activation of the mT␤R-II promoter. Northern blot analysis reveals that the steadystate levels of mT␤R-II mRNA increase in parallel with those of mElf-3 mRNA during the differentiation of F9 embryonal carcinoma cells. We also demonstrate that mElf-3 contains one or more domains that influence its binding to DNA. Finally, we report that a single amino acid substitution in the transactivation domain of mElf-3 reduces its ability to transactivate and elevates its steady-state levels of expression. In conclusion, our data argue that mElf-3 plays a key role in the regulation of the mT␤R-II gene, and Elf-3 itself is regulated at multiple levels.
Transforming growth factor-␤ (TGF-␤) 1 is a family of genetically distinct polypeptides that are secreted by virtually all cells. These growth factors play important roles in cell proliferation, differentiation, and synthesis of extracellular matrix proteins (1,2). Three major TGF-␤ receptors have been identified and designated as type I (T␤R-I), type II (T␤R-II), and type III (T␤R-III) receptors. T␤R-III is a glycoprotein (ϳ280 kDa) that appears to lack intrinsic signaling activity. However, T␤R-III can present TGF-␤ to its signaling receptors, T␤R-I (ϳ55 kDa) and T␤R-II (ϳ75 kDa) (3). Appropriate cellular responses to extracellular TGF-␤ are initiated upon binding of TGF-␤ to T␤R-II, which is a constitutively active serine/threonine kinase. When bound to ligand, T␤R-II forms a heteromeric signaling complex with T␤R-I and phosphorylates T␤R-I at serine and threonine residues within its kinase domain (4). This activates a series of downstream signaling pathways (2).
Both T␤R-I and T␤R-II are required for TGF-␤-mediated growth suppression. Hence, loss of either receptor leads to TGF-␤ resistance, which often contributes to malignant progression. TGF-␤ resistance caused by defects in T␤R-II expression has been reported in numerous tumor cell lines (5)(6)(7). Moreover, several studies (8,9) have established a clear relationship between defective T␤R-II expression and malignant progression. Importantly, transfection of several different tumor cell lines with an expression vector for T␤R-II restored sensitivity of the cells to TGF-␤ and reduced their malignant behavior (5,9,10). Similarly, transfection of a T␤R-I-defective colon carcinoma cell line with a T␤R-I expression vector reversed its malignant phenotype (11). Together, these findings illustrate the importance of both T␤R-I and T␤R-II as tumor suppressor genes.
In the case of the T␤R-II gene, gross structural mutations of both T␤R-II alleles have been observed in 71-90% of colorectal tumors and in 71% of gastric cancer cell lines with microsatellite instability (6,12,13). However, in several different types of tumors, which exhibit either reduced or undetectable expression of the T␤R-II gene at the mRNA or protein level, no structural mutations were apparent within the coding region of the gene (6, 14 -16). This suggests that defects in the mechanisms regulating the transcription of the T␤R-II gene may play important roles in the aberrant expression of T␤R-II. In this regard, it has been argued that the low levels of T␤R-II mRNA expressed by A431 tumor cells are due to a point mutation located in the 5Ј-flanking region of the T␤R-II promoter (17). Additionally, it has been reported that this point mutation increases CDP/Cut transcription factor binding affinity, and overexpression of CDP/Cut reduces transcription from T␤R-II promoter (18). During the last several years, the promoter region of the human T␤R-II gene (hT␤R-II) was examined in several cell culture model systems (19 -22). Initial analysis identified three positive cis-regulatory elements within 200 bp of the major transcription start site, i.e. a CRE/ATF motif, a GC box, and two Ets-binding sites (EBS) that are separated by 3 bp (19).
Although most cells express TGF-␤ receptors, mouse embryonal carcinoma (EC) cells lack detectable TGF-␤ binding (23). After EC cells undergo differentiation, TGF-␤ binding is readily detected, and the growth of the differentiated cells becomes responsive to TGF-␤ (23). Subsequent studies demonstrated that the differentiation of mouse EC cells leads to transcriptional up-regulation of both the T␤R-II gene (20) and the gene for one of its ligands, TGF-␤2 (24). Because mouse EC cells provide an excellent model system for studying both mammalian embryogenesis (25) and the transcriptional regulation of genes that play key roles during development and cancer (26), we set out to clone and characterize the regulatory regions of the mouse T␤R-II gene (mT␤R-II). Cloning the mouse promoter would make it possible to study the transcriptional upregulation of the T␤R-II gene in a well characterized mouse model system. It would also enable us to identify and then focus on conserved cis-regulatory elements. In this study, we describe the isolation and characterization of the 5Ј-flanking region of the mT␤R-II gene. In the course of characterizing the mT␤R-II promoter, we focused our attention on the conserved EBS that are located just downstream of the major transcription start site in the human and mouse T␤R-II genes. We show that mElf-3 binds to the EBS and strongly stimulates expression of the mT␤R-II promoter in differentiated cells derived from mouse F9 EC cells. We also demonstrate that differentiation of EC cells leads to parallel increases in the expression of mRNA for mT␤R-II and mElf-3. Equally important, we demonstrate that mElf-3 itself is subject to other levels of regulation. Hence, characterization of the mechanisms that regulate mElf-3 may help identify ways to enhance the expression of the T␤R-II gene in diseased tissues that exhibit aberrantly low levels of T␤R-II.

EXPERIMENTAL PROCEDURES
Materials-Dulbecco's modified Eagle's (DME) medium was purchased from Invitrogen. Fetal bovine serum (FBS) was obtained from HyClone (Logan, UT). Unless otherwise stated all chemicals were purchased from Sigma.
Cloning the Promoter Region of the Murine Type II TGF-␤ Receptor (mT␤R-II) Gene-To assist in cloning the 5Ј end of the mT␤R-II gene, we designed a set of PCR primers based on the 5Ј end of the cDNA sequence from Suzuki et al. (see Ref. 27; GenBank TM D32072): 5Ј-CC-TCCGGGCCTCCGAGCTCC-3Ј (sense) and 5Ј-CCGTCGCTCGTCATA-GACCG-3Ј (antisense). Using these primers, Genome Systems Inc. (St. Louis, Missouri) screened a murine strain 129/O1a genomic library and identified a P1 clone that contained a 75-100-kb genomic fragment. To obtain a smaller genomic fragment, the P1 clone was digested with PstI, and the resulting fragments were ligated into Bluescript KSϩ (Stratagene) and transformed into DH5␣FЈ bacteria. Using the same set of primers used to screen the P1 genomic library, we identified a clone (ϳ4 kb) that contained the 5Ј-flanking region of the mT␤R-II gene. Sequence analysis of both strands over a region of 1021 bp indicated that it extended more than 500 bp upstream of the two major transcription start sites (see below). Sequencing was performed with an Applied Biosystems 373 DNA Sequencer (PerkinElmer Life Sciences). Alignment of the sequencing results was accomplished using version 1.1 of the Sequence Navigator program (PerkinElmer Life Sciences/Applied Biosystems). Comparisons (alignments) with other sequences were performed using programs of the Wisconsin Package version 9.1, Genetics Computer Group (GCG, Madison, WI).
Cell Culture and Differentiation-F9 EC cells were cultured on gelatinized dishes in DME medium supplemented with 10% FBS and antibiotics. Differentiation of F9 EC cells was induced by treatment with 5 M retinoic acid (RA) for the times indicated (20). HeLa and 293T cells were cultured in DME medium supplemented with 10% FBS and antibiotics. All cells were cultured at 37°C in a humidified atmosphere of 5% CO 2 in air.
Northern Blot Analysis and Primer Extension Analysis-Poly(A) ϩ RNA was isolated from F9 EC cells and F9-differentiated cells using the Invitrogen FastTrack 2.0 kit (Invitrogen) according to the manufacturer's instructions. Five g of each mRNA was fractionated on 1.2% agarose gels containing 0.22 M formaldehyde, transferred to nylon membrane by capillary transfer, and immobilized by UV cross-linking (20). Probes were labeled with [␣-32 P]dCTP by using the Prime-It II kit (Stratagene, La Jolla, CA). mElf-3 cDNA (see below) was used to generate a probe that was 864 bp in length and encoded amino acid residues 1-287. mT␤R-II cDNA was used to generate a probe that was 699 bp in length and encoded the amino acid residues 90 -322. A mT␤R-I (ALK-5) cDNA was used to generate a probe that was 657 bp in length and encoded the amino acid residues 152-370. Hybridization was performed at 68°C in the ExpressHyb TM hybridization solution (CLONTECH, Palo Alto, CA) for 1.5 h, and the blots were washed according to the manufacturer's instructions. Hybridization of GAPDH was used as a control for equal loading. Quantitation was accomplished using the PhosphorImager and ImageQuant software (Molecular Dynamics, Sunnyvale, CA). Sizes of the transcripts were estimated by comparison to Millennium RNA markers (Ambion, Austin, TX). The transcription start site of the mT␤R-II gene was examined by primer extension as described previously (28). For this purpose, we used the antisense primer mT␤R-II-PE (5Ј-GCGCGCGCGGGGGGTGTCGTCG-GTCGGTGC-3Ј (ϩ96 to ϩ67)) and mRNA isolated from F9 EC cells induced to differentiate by exposure to RA for 4 days.
Generation of Promoter/Reporter Gene Constructs-Various regions of the mT␤R-II promoter were amplified by PCR from the 4-kb genomic mT␤R-II DNA fragment, which we had subcloned into Bluescript KSϩ. Restriction sites (HindIII or BglII) were incorporated at the ends of all primers to permit directional cloning of the amplified fragments into the HindIII and BglII sites in the polycloning region of the promoterless chloramphenicol acetyltransferase (CAT) expression plasmid pBLCAT7 (29). Site-directed mutagenesis of the mT␤R-II promoter CAT construct was performed using the QuickChange TM Site-directed Mutagenesis kit (Stratagene). All promoter sequences were verified by DNA sequencing.
Construction of Expression Vectors-Three FLAG-tagged mElf-3 expression vectors were cloned from mRNA using reverse transcriptase-PCR. One contained a cDNA for full-length mElf-3 (amino acid residues 1-371); a second contained sequences coding for the AT-hook domain plus the DNA binding domain of mElf-3 (amino acids 234 -371) (ATH ϩ DBD); and the third contained sequences for the DNA binding domain of mElf-3 (amino acids 271-371) (DBD). Poly(A) ϩ RNA was extracted from F9-differentiated cells using the Invitrogen FastTrack 2.0 kit (Invitrogen, San Diego, CA). Reverse transcription was performed at 42°C for 1 h with SuperScript II TM reverse transcriptase (Invitrogen) using the antisense primer. PCR amplification reaction was performed by using Pfu polymerase (Stratagene). The cycling conditions were 94°C for 2 min followed by 35 cycles of 94°C for 15 s, 55°C for 30 s, and 70°C for 1 min. The sequences of the sense primer for full-length, ATH ϩ DBD, and DBD mElf-3 were 5Ј-CTCAAGCTTGCCACCATGGACTA-CAAGGACGACGACGATAAGATGGCTGCCACCTGTGAGATCAGC-3Ј, 5Ј-ACGAGGATCCGACTATAAGAAGGGGGAACC-3Ј, and 5Ј-AATT-GGATCCAGAGGTACTCACCTGTGGGAGTTTATCC-3Ј, respectively. The sense primer for full-length mElf-3 contained a HindIII site, a Kozac sequence, the FLAG sequence, and the 5Ј mElf-3 coding sequences. The sense primers for ATH ϩ DBD and DBD contained a BamHI site and the coding sequences. The sequence of the antisense primer was 5Ј-CGCTCTAGATTAATTCCGACTCTCTCCAACCTC-3Ј. It contained the sequence for an XbaI site followed by coding sequences for mElf-3. The PCR products were digested with restriction enzymes and directionally cloned into the polylinker region of the mammalian expression vector FNpcDNA3.0, which was kindly provided by Dr. Craig Hauser (Birnham Institute, La Jolla, CA), and described in Galang et al. (30). The resulting mElf-3 expression plasmids encoded a protein with an N-terminal FLAG tag. Site-directed mutagenesis of the mElf-3 cDNA was performed using the QuickChange TM Site-directed Mutagenesis kit (Stratagene). All sequences were verified by DNA sequencing.
Verification of mELF-3 cDNA Sequence-Our sequence of mElf-3 cDNA did not agree fully with its published sequence (31,32). To confirm the mElf-3 cDNA sequence, genomic DNA isolated from F9 EC cells using standard protocols (33) was used as a template for PCR amplification. Oligonucleotide primers were designed for the regions in question (exons 2, 4, and 7) to generate PCR products containing the exon and the intron sequences. The sequences of the primer sets were 5Ј-CCCTGAACAACCAACAGATGAC-3Ј (sense, exon 2) and 5Ј-TAGG-CTCTCTTGGAAGGACATG-3Ј (antisense, exon 4) with an expected amplification product of 852 bp, and 5Ј-CCTCCAACTCTTCTGATGAA-CTC-3Ј (sense, exon 4) and 5Ј-CAGTATTCCTTGCTCAGCTTTCTG-3Ј (antisense, exon 7) with an expected amplification product of 870 bp. PCR amplification reaction was performed using Pfu polymerase (Stratagene). PCR conditions were 94°C for 2 min followed by 35 cycles of 94°C for 15 s, 55°C for 30 s, and 70°C for 1 min. The PCR products were gel-fractionated, and the visualized band was purified using a QIAEX II gel extraction kit (Qiagen, Chatsworth, CA). The PCR products were sequenced in both directions and compared with the cDNA sequences.
Transient Transfection and CAT Reporter Gene Analysis-F9-differ-entiated cells were transfected by the calcium phosphate precipitation method (34). The pCMV-␤-gal plasmid containing the ␤-galactosidase reporter gene under the control of the CMV immediate early promoter was cotransfected to normalize for differences in transfection efficiency. F9 EC cells were plated at a density of 1 ϫ 10 5 cells per 100-mm dish in DME medium containing 10% FBS and 5 M RA. After 72 h, the cells were transfected and incubated with the DNA-calcium phosphate precipitate for 22 h and then were refed with DME medium containing 10% FBS and 5 M RA. CAT activities were determined 72 h after transfection using the method of Seed and Sheen (35) and were normalized to ␤-galactosidase activity. 293T cells were seeded at 2 ϫ 10 6 cells per 100-mm dish in DME medium containing 10% FBS. After 24 h, the cells were transfected using the calcium phosphate precipitation method (34). The cells were incubated with the precipitate for 22 h and then were refed with DME medium containing 10% FBS. CAT activities were determined 24 h after transfection and were normalized as described above. When different amounts of the expression plasmid for mElf-3 were transfected into F9-differentiated and 293T cells, the total amounts of transfected DNA were kept constant by addition of null plasmid DNA, FNpcDNA3.0, as indicated in the figure legends. HeLa cells were seeded at 7.5 ϫ 10 5 cells per 100-mm dish in DME medium containing 10% FBS. Transfections were performed as described for 293T cells. HeLa cell nuclear extracts were prepared 24 h after transfection and were used for Western blot analysis and gel mobility shift analysis as described below. All transfection experiments were repeated at least twice with different plasmid preparations. Preparation of Nuclear Extracts and Performance of Gel Mobility Shift Analysis-Nuclear extracts were prepared from HeLa cells transiently transfected with 10 g of one of the three FLAG epitope-tagged mElf-3 expression plasmids described above. Nuclear extracts were prepared using the NE-PER TM nuclear and cytoplasmic extraction kit (Pierce). Protein concentrations were determined using the Micro BCA protein assay kit (Pierce). Annealed, double-stranded oligonucleotides were labeled by a fill-in reaction using the Klenow fragment of DNA polymerase I (New England Biolabs, Beverly, MA). Nuclear extracts (12 g) were incubated for 20 min at room temperature in 20 l of binding buffer containing 50 mM Tris, pH 7.5, 50 mM NaCl, 5 mM MgCl 2 , 1 mM dithiothreitol, 1 mM EDTA, 5% glycerol, 1 g of poly(dI-dC), 100 g/ml bovine serum albumin, and ϳ20,000 cpm of probe. In competitive binding or supershift assays, unlabeled competitor DNAs or a monoclonal antibody to the FLAG epitope (M2) was preincubated with the nuclear extract for 40 min at room temperature before addition of the labeled probe. DNA-protein complexes were resolved on a 6% native polyacrylamide gel in 0.25ϫ Tris/boric/EDTA buffer as described previously (20). The gels were dried and visualized by PhosphorImager (Molecular Dynamics, Sunnyvale, CA). Oligonucleotide probes corresponding to nucleotides ϩ3 to ϩ34 of the mT␤R-II gene were as follows: wild type, 5Ј-TGGCGAGGAGTTTCCTGTTTCCCTCTCGGCGC-3Ј; mutant, 5Ј-T-GGCGAGGAGTTTGGATCCACCCTCTCGGCGC-3Ј (underline indicates mutated sequences).
Chromatin Immunoprecipitation (ChIP)-293T cells were seeded at a density of 5 ϫ 10 6 on 150-mm plates. Cells were co-transfected 24 h after seeding by the calcium phosphate precipitation method (34) with the promoter/reporter vector mT␤R-II(Ϫ108/ϩ56) and a FLAG-tagged mElf-3 expression vector containing either full-length Elf-3 or the DBD of Elf-3. After incubation with the precipitate for 20 h, the medium was changed, and ChIP analysis was initiated 4 h later. Cross-linking was performed by addition of formaldehyde to the media at a final concentration of 1% for 10 min at room temperature. The cross-linking reaction was quenched with glycine (0.125 M) for 5 min at room temperature. Cells were then washed 3 times with PBS and scraped into 1 ml of phosphate-buffered saline. Cell pellets were resuspended in Cell Lysis Buffer (5 mM PIPES, 85 mM KCl, 0.5% IGEPAL CA-630, 0.5 mM phenylmethylsulfonyl fluoride), incubated for 15 min on ice, and the nuclei pelleted. The nuclei were lysed in Nuclei Lysis Buffer (50 mM Tris-Cl, 10 mM EDTA, 1% SDS, 0.5 mM phenylmethylsulfonyl fluoride) on ice for 10 min. Shearing of the DNA was then performed using a W-225 cup horn sonifier (Misonix, Inc.), and cell debris was pelleted. The supernatant was precleared with protein G PLUS-agarose beads (pre-blocked with bovine serum albumin, Santa Cruz Biotechnology) for 1 h at 4°C. Following preclearing, the supernatant was diluted 2-fold with Dilution Buffer (0.01% SDS, 1% Triton X-100, 1.2 mM EDTA, 16.7 mM Tris-Cl, pH 8, 167 mM NaCl), and 5% was removed and save as an "Input" sample. The remaining sample was divided into 2 aliquots. One aliquot was incubated with FLAG (M2) antibody conjugated to agarose beads (Sigma), and the other aliquot was incubated with Gal4 (DBD) antibody conjugated to agarose beads (Santa Cruz Biotechnology). Both aliquot mixtures were incubated overnight at 4°C. The beads were then washed 4 times with TBS (20 mM Tris, 140 mM NaCl, pH 7.6), and complexes were eluted by two incubations with 100 l of Elution Buffer (1% SDS, 50 mM NaHCO 3 ) at 65°C for 10 min (a total of 200 l for each sample). The eluted immunoprecipitated samples and Input sample were decross-linked by addition of NaCl to a final concentration of 0.2 M and incubation at 65°C overnight. Following decross-linking, the samples were treated with 20 g of proteinase K in Proteinase K Buffer (10 mM Tris-Cl, pH 7.5, 5 mM EDTA, 0.25% SDS) for 2 h at 42°C. DNA fragments were purified using the Geneclean Turbo kit (Qbiogene). The M13-21 FOR primer (5Ј-TGTAAAACGACGGCCAG-3Ј), specific for a sequence located in the plasmid backbone, and the primer ELF3R (5Ј-ACGATGCCATTGGGATA-3Ј), specific for a sequence within the Ϫ108/ ϩ56 region of mT␤R-II, were used to amplify specifically the mT␤R-II promoter region located in the transfected promoter/reporter gene construct. PCR was performed on equivalent volumes of each sample and analyzed at multiple cycles. Results are presented for one cycle within the linear range of amplification.
Western Blot Analysis-Aliquots of nuclear extracts prepared from cells (293T or HeLa) transfected with FLAG-tagged expression vectors coding for full-length and truncated forms of mELF-3, as well as aliquots of the cell extracts from transfected cells used in the CAT assays, were immunoblotted and probed with an anti-FLAG antibody as described previously (36). Visualization was accomplished using the enhanced chemifluorescence kit (Amersham Biosciences). Quantitation was performed using the ImageQuant analysis software (Molecular Dynamics).

Isolation of the 5Ј-Flanking Region of the mT␤R-II Gene-
The 5Ј-flanking region of the mT␤R-II gene was isolated from a P1 genomic library that was screened by PCR using primers based on the 5Ј end of the published mT␤R-II cDNA sequence (27). The P1 genomic clone identified in this screen was digested with PstI, subcloned into pBlueScript II KSϩ, and rescreened with the same set of primers used to identify the P1 genomic clone. This led to the isolation of a 4-kb PstI genomic fragment that contained the 5Ј end of the mT␤R-II cDNA. Sequencing of this 4-kb genomic subclone demonstrated that it contained the 5Ј-flanking region of the mT␤R-II gene and extended 3Ј beyond the exon I/intron I border. The sequence of this region has been assigned the GenBank TM accession number AF118264. Comparison of the immediate 5Ј-flanking region of the murine and human T␤R-II genes revealed considerable homology (Fig. 1A). Previous work (19 -22) with hT␤R-II promoter/reporter gene constructs identified several cis-regulatory elements, including a CRE/ATF motif, a CCAAT box motif, a GC box, and two closely spaced EBS. The sequences and positions of these putative cis-regulatory elements are fully conserved in the mT␤R-II gene. In addition, the 5Ј-flanking region of the mT␤R-II gene, like its human counterpart, does not contain a canonical TATA box. Examination of the human T␤R-II promoter by 5Ј-RACE indicated that transcription is initiated at multiple sites over a 70-bp region (19). By using primer extension, we identified three sites of initiation for the mT␤R-II promoter (Fig. 1B). Two sites are closely spaced, and a third is ϳ37 bp downstream. One of the sites, indicated in the Fig. 1B, corresponds to the major transcription initiation site reported for the hT␤R-II gene. This site in the hT␤R-II gene was designated as the ϩ1 position (19). For the purposes of alignment, we designated the corresponding G residue in the mT␤R-II promoter region as ϩ1 in this report (Fig. 1A).
Regulatory Regions of the mT␤R-II Gene-To understand better how the T␤R-II gene is regulated in mouse F9-differentiated cells, we prepared a series of mT␤R-II promoter/reporter gene constructs (Fig. 2A). These constructs were generated with the plasmid pBLCAT7, which contains a CAT reporter gene cassette and a multiple cloning site into which different lengths of the 5Ј-flanking region of the mT␤R-II gene were inserted. The largest construct, mT␤R-II(Ϫ300/ϩ56), contains all of the regulatory regions identified in the hT␤R-II gene as well as the major transcription start sites identified for the Elf-3 and Regulation of the Type II TGF-␤ Receptor mouse gene. The mT␤R-II promoter/reporter gene constructs were transiently transfected into F9-differentiated cells (Fig.  2B). The constructs mT␤R-II(Ϫ300/ϩ56), -(Ϫ232/ϩ56), and -(Ϫ108/ϩ56) exhibited the highest levels of promoter activity. Deletion of sequences from Ϫ300 to Ϫ233 resulted in a minor increase, and further deletion of sequence from Ϫ232 to Ϫ109, which contains a putative CRE/ATF-binding site (20), resulted in a minor decrease in promoter activity. However, deletion of the sequence between ϩ2 and ϩ56, which contains the potential EBS, markedly decreased mT␤R-II promoter activity. These results suggested that the EBS are likely to play an important role in the expression of mT␤R-II gene, although other cis-regulatory elements in the mT␤R-II promoter are also likely to influence the expression of this gene.
To test the role of the EBS in the mT␤R-II promoter, the Ets-binding sites located between ϩ14 and ϩ24 were disrupted by site-directed mutagenesis. Because both Ets-binding sites in the hT␤R-II promoter were found to influence its expression (19), both sites in the construct mT␤R-II(Ϫ108/ϩ56) were mod-ified by site-directed mutagenesis. Specifically, two different mutant constructs were generated, mT␤R-II(Ϫ108/ϩ56EB-Smut1) and mT␤R-II(Ϫ108/ϩ56EBSmut2) (Fig. 3A). mT␤R-II(Ϫ108/ϩ56EBSmut1) inactivated both GGAA core element EBS but contained three GGA sequences. Because GGA sequences have been reported to provide a core-binding element for some Ets family members (37), each of the three GGA sequences in mT␤R-II(Ϫ108/ϩ56EBSmut1) were modified to create mT␤R-II(Ϫ108/ϩ56EBSmut2). When these modified constructs and their wild type counterpart were transiently transfected into F9-differentiated cells, we determined that disruption of the EBS in both mutant constructs reduced promoter activity by ϳ60% (Fig. 3B). Interestingly, a larger reduction (ϳ80%) is observed when the region between ϩ3 and ϩ56 is deleted (Fig. 2B). However, removal of this region also removes one of the two potential transcription start sites in our promoter/reporter gene constructs. In the case of hT␤R-II promoter/reporter gene constructs, removal of a comparable region reduced promoter activity by 60% (19). Hence, our studies and those conducted with hT␤R-II promoter/reporter gene constructs argue that the EBS play an important role in the function of both the human and the mouse T␤R-II promoter. Putative CRE/ATF-binding motif, CCAAT box, GC box, and Ets family transcription factor-binding motifs are boxed. The nucleotide designated ϩ1 is number 557 in the mouse genomic sequence deposited with the GenBank TM accession number AF118264. The human sequence was assembled from the published cDNA (GenBank TM M85079; number 1 is ϩ1) and promoter (GenBank TM U37070, number 1883 is Ϫ1) sequences. B, primer extension analysis to identify the transcription start sites of the mT␤R-II gene in F9-differentiated cells. Three micrograms of poly(A) ϩ RNA isolated from F9-differentiated cells was hybridized to the radiolabeled antisense primer mT␤R-II-PE. Primer extension was performed using Superscript II reverse transcriptase. The resulting cDNA product was fractionated by electrophoresis and analyzed as described in the "Experimental Procedures."

FIG. 2. Functional analysis of the mT␤R-II promoter.
A, schematic representation of mT␤R-II promoter/reporter gene constructs. Overlapping segments of the 5Ј-flanking region of the mT␤R-II promoter were generated by PCR and inserted into the multiple cloning site upstream of the CAT reporter gene in a promoterless vector, pBLCAT7. The dotted line represents deleted regions (ϩ3 to ϩ56). B, mT␤R-II promoter activity in F9-differentiated cells. Cells were transiently transfected with 15 g of each construct described in A, along with 1 g of an internal control plasmid pCMV-␤-gal. Activities of the CAT and ␤-galactosidase reporter genes were assayed and normalized as described under "Experimental Procedures." The promoter activity of each construct is calculated relative to the expression of the CAT reporter gene observed with pBLCAT7, which was set to 1. Data shown are means and S.D. for duplicate measurements from one representative transfection. This experiment was performed three times, and similar results were obtained in each case. (22) have shown that the transcription factor hELF-3 can bind to the EBS present in the hT␤R-II promoter. Moreover, recent studies (38) have shown that ectopic expression of hELF-3 increases expression of hT␤R-II in human breast cancer cells and reduces their malignant phenotype. Because we had shown previously (20) that the differentiation of EC cells leads to increases in the steadystate levels of mT␤R-II mRNA, we examined the expression of the mElf-3 gene in EC cells and their RA-induced differentiated cells. For this purpose, we initially compared mElf-3 mRNA expressed by F9 EC cells and by F9 cells that had been treated with RA for 5 days. Northern blot analysis demonstrated that the steady-state levels of mElf-3 were 7-8-fold higher in F9differentiated cells than in EC cells (data not shown). The finding that differentiation leads to significant increases in expression of mElf-3 prompted us to examine the temporal relationship between increases in mElf-3 mRNA and mT␤R-II mRNA. If, in fact, increases in mElf-3 are responsible, at least in part, for the up-regulation of the mT␤R-II gene, increases in mElf-3 mRNA should precede or parallel the increases in mT␤R-II. To address this issue, mRNA was prepared from EC cells and from EC cells that had been treated with RA for different periods up to 5 days. A Northern blot was prepared and sequentially probed for the expression of mElf-3, T␤R-II, GAPDH, and mT␤R-I transcripts. As reported by other investigators (31, 39) studying mElf-3 in other cell types, multiple mElf-3 transcripts were observed, and the most abundant transcript exhibited a size of 2.2 kb (Fig. 4). Our analysis revealed that mElf-3 mRNA increased ϳ3-fold after 24 h of exposure to RA and more than 6-fold after 3 days (Fig. 4). Moreover, analysis of the PhosphorImager data by ImageQuant demonstrated that each of the three transcripts exhibited similar increases during the 5-day period (data not shown). Interestingly, mT␤R-II mRNA increased over 6-fold during the same 5-day period, and the increases in the expression of mElf-3 and mT␤R-II correlated closely with one another, in particular during the initial stages of differentiation (the first 48 h). Hence, these findings are consistent with the possibility that the increase in mElf-3 expression during the differentiation of EC cells plays a key role in the elevated expression of the mT␤R-II gene. Surprisingly, the steady-state levels of T␤R-I mRNA did not change significantly when F9 EC cells differentiate. Both EC cells and their differentiated counterparts express similar levels of a 5.5-kb T␤R-I transcript (Fig. 4), which is the transcript size observed in various mouse tissues (40).

Increased Expression of the mT␤R-II Gene during the Differentiation of F9 EC Cells Correlates with Increased Expression of the mElf-3 Gene-Previous studies
mELF-3 Transactivates the mT␤R-II Promoter Specifically through Binding to the EBS Located between ϩ14 and ϩ24 in the mT␤R-II Promoter-To determine whether mElf-3 influences expression of the mT␤R-II gene in F9-differentiated cells, we generated an mElf-3 expression vector. For this purpose, we employed reverse transcriptase-PCR and mRNA prepared from F9-differentiated cells. The cDNA product was inserted into the multiple cloning site of the mammalian expression vector FNpcDNA 3.0 and sequenced. Previous reports (31, 32) of the coding sequence of mElf-3 differed at six amino acid residues. Sequence analysis of our cDNA agrees with one of the pub- In addition, the nucleotides that differ between the two mutant constructs are indicated by asterisks. B, F9-differentiated cells were transiently transfected with wild type (mT␤R-II(Ϫ108/ϩ56)) and two mutant promoter/ reporter gene constructs, mT␤R-II(Ϫ108/ϩ56EBSmut1) and mT␤R-II(Ϫ108/ϩ56EBSmut2) along with 1 g of plasmid CMV-␤-gal as an internal control. Promoter activity of the mutant construct is calculated relative to the wild type construct, which was assigned a relative activity of 1. Data shown are mean and S.D. for duplicate measurements from one representative transfection. This experiment was repeated, and similar results were obtained. The reduced activity of mT␤R-II(Ϫ108/ϩ56EBSmut1) relative to mT␤R-II(Ϫ108/ϩ56) was observed in multiple experiments.

Elf-3 and Regulation of the Type II TGF-␤ Receptor
lished sequences (32) at four of the amino acid residues and agrees with the other report for the two remaining differences (31) ( Table I). Each of the differences identified in our study was confirmed at the nucleotide level in four separate sequencing reactions, including the sequencing of both strands of PCR products generated from genomic DNA isolated from F9 EC cells. Currently, it is unclear whether any of the differences are due to single nucleotide polymorphisms. Our study and one of the published reports (32) utilized mouse strain 129. The mouse strain used in the other study (31) could not be determined.
The mElf-3 mammalian expression vector described above was used to test whether enhanced expression of mElf-3 would boost the activity of the mT␤R-II promoter in F9-differentiated cells. For this purpose, the cells were transiently transfected with the mT␤R-II(Ϫ108/ϩ56) promoter/reporter gene construct and increasing amounts of the mElf-3 expression vector. This resulted in a dose-dependent increase in the activity of the mT␤R-II promoter, which exceeded 20-fold (Fig. 5). In contrast, little increase was observed when either mT␤R-II(Ϫ108/ϩ2) or mT␤R-II(Ϫ108/ϩ56EBSmut1), which lack the EBS, was cotransfected with the mElf-3 expression vector. These findings indicate that overexpression of mElf-3 can increase the activity of the mT␤R-II promoter and further argues that the ability of mElf-3 to stimulate the mT␤R-II promoter occurs predominantly through the EBS located between ϩ14 and ϩ24. Furthermore, we tested the ability of Ets-2 to stimulate the promoter/reporter gene construct mT␤R-II(Ϫ108/ϩ56), because Ets-2 is expressed by EC-differentiated cells (41). However, unlike mElf-3, ectopic expression of Ets-2 in F9-differentiated cells did not stimulate the expression of the reporter gene (data not shown).
Truncated Forms of mElf-3 Inhibit mElf-3-mediated Transactivation-To better understand the action of mElf-3, we examined the effects of deleting several domains of mElf-3. In addition to the Ets domain, which is responsible for DNA binding, and the transactivation domain, which maps to amino acids 229 -259, mElf-3 contains two other domains, a pointed domain and an AT-hook domain (Fig. 6A) (31). The pointed domain exhibits sequence homology to a similar domain contained in Ets-1 (42), and the AT-hook domain is similar to a domain first identified in the high mobility group protein HMG-I(Y) (43). In this study, we examined a truncated form of mElf-3 (amino acids 234 -371) that lacks the pointed domain and the transactivation domain and a smaller form (amino acids 271-371) that also lacks the AT-hook domain. We determined that mammalian expression vectors for both truncated forms did not stimulate the activity of the mT␤R-II (Ϫ108/ϩ56) promoter/reporter construct. In fact, they reduced basal expression of the reporter gene (Fig. 6B). These findings suggested that the truncated forms of mElf-3, like the truncated forms of other Ets family members (30,44), were behaving as dominant negatives and blocking the function of an endogenous factor. This possibility was examined by transfecting F9-differenti-ated cells with an expression vector for full-length mElf-3 and increasing amounts of expression vectors for the truncated forms of mElf-3 (Fig. 6C). At 1 g of the full-length mElf-3 expression vector, the level of the reporter gene was elevated more than 12-fold. As the concentrations of the expression vectors for the truncated forms were increased, the expression of the reporter gene decreased. Moreover, the truncated form that contained both the AT-hook and the DNA binding domain reduced the expression of the reporter gene to the level observed in the absence of added full-length mElf-3.
One possible explanation for the inhibitory effects of the truncated forms of mElf-3 is that they interfere with the expression of full-length mElf-3. We examined this possibility by performing Western blot analysis of the expression of mElf-3 in the presence of the truncated form that contained both the AT-hook and the DNA binding domain. Due to low transfection efficiency of the F9-differentiated cells, we performed this study in 293T cells, which can be transfected at high efficiency. We determined that the truncated form reduces the response of the reporter gene to mElf-3 (Fig. 7A) without reducing the expression of mElf-3 in 293T cells (Fig. 7B). Taken together, our results argue that the truncated forms of mElf-3 act as dominant negatives and reduce the activity of the mT␤R-II promoter.
DNA Binding of mElf-3-There are over 30 members in the Ets family of transcription factors. Previous studies (45)(46)(47)(48)(49)(50)(51) have shown that several members of the Ets family of transcription factors contain auto-inhibitory domains that influence their ability to bind DNA in vitro. To examine whether this was also true for mElf-3, we performed gel mobility shift analysis using nuclear extracts prepared from HeLa cells transfected with expression vectors for mElf-3 or one of the two truncated forms of mElf-3 described above. The smaller truncated protein, which consists of only the DNA binding domain, formed one major and one minor DNA-protein complex (Fig.  8A), and both complexes were supershifted with the M2 monoclonal antibody that recognizes the FLAG epitope placed at the N terminus of the protein. In other studies, we determined that the intensity of both complexes is greatly diminished when a a Only the amino acid sequence was reported by Neve et al. (31). b The codon for amino acid 200 is split between exon 5 ("A") and exon 6 ("GG") (30).

FIG. 5. Effect of mElf-3 on the expression of mT␤R-II promoter/
reporter gene constructs. F9-differentiated cells were transiently transfected with 15 g of either mT␤R-II(Ϫ108/ϩ56), mT␤R-II(Ϫ108/ ϩ56EBSmut1), or mT␤R-II(Ϫ108/ϩ2) and increasing amounts of the plasmid DNA coding for mElf-3. The cells were also transfected with 1 g of pCMV-␤-gal plasmid to normalize for differences in transfection efficiencies. Total amount of DNA transfected was kept constant using the null plasmid DNA, FNpcDNA3.0. Activities of CAT and ␤-galactosidase were assayed and normalized as described under "Experimental Procedures." Data are presented as fold activation relative to cells transfected with either mT␤R-II(Ϫ108/ϩ56), mT␤R-II(Ϫ108/ϩ56EB-Smut1), or mT␤R-II(Ϫ108/ϩ2) alone. Data shown are means and S.D. for duplicate measurements from one representative transfection. This experiment was repeated twice, and similar results were obtained in each case.

Elf-3 and Regulation of the Type II TGF-␤ Receptor
100-fold excess of the unlabeled probe was added to the reaction mixture (data not shown). In contrast, these complexes were unaffected by the addition of a 100-fold excess of an unlabeled competitor that contained a scrambled EBS, and no DNA-protein complexes formed when this sequence was used as our probe (data not shown). The reason why two complexes form with the DNA binding domain of mElf-3 is unclear. How-ever, the slower migrating complex may consist of two molecules of the DNA binding domain bound to the probe. In this regard, the probe used in this analysis consists of the region ϩ3 to ϩ34 (Fig. 1A), which contains the two closely spaced Etsbinding sites.
In direct contrast to our findings with the DNA binding domain of mElf-3, we did not observe DNA binding of fulllength mElf-3 in vitro. Similarly, relatively little binding was observed in vitro by the truncated form of mElf-3 that contains both the AT-hook domain and the DNA binding domain (Fig.  8A). To ensure that mElf-3 and the longer truncated form of mElf-3 were expressed adequately, we performed Western blot analysis on the nuclear extracts that were used in the gel mobility shift analysis (Fig. 8B). Although mElf-3 was expressed at a level ϳ65% lower than expression of the truncated form containing the DNA binding domain, this is unlikely to explain the failure of full-length mElf-3 to bind DNA in vitro. This is evident in the case of the larger truncated form of mElf-3 (ATH ϩ DBD), which bound weakly to DNA and which  DBD, DBD). The cells were also transfected with 1 g of pCMV-␤-gal plasmid to correct for any differences in transfection efficiency. C, F9-differentiated cells were transfected with 15 g of the mT␤R-II(Ϫ108/ϩ56) promoter/reporter gene construct along with an expression plasmid encoding full-length mElf-3 on its own or together with increasing amounts of an expression plasmid encoding the truncated forms of mELF-3. The cells were also transfected with 1 g of pCMV-␤-gal. The total amount of DNA transfected was kept constant by the addition of FNpcDNA3.0. Data are presented as fold activation relative to cells transfected with mT␤R-II(Ϫ108/ϩ56) alone. Data shown in B and C are means and S.D. for duplicate measurements from one representative transfection. Each experiment was repeated twice, and similar results were obtained in each case.

FIG. 7. Effects of a truncated form of mElf-3 on the expression of mT␤R-II promoter/reporter gene constructs in 293T cells. A,
293T cells were co-transfected with 15 g of the mT␤R-II(Ϫ108/ϩ56) promoter/reporter construct and 3 g of the expression plasmid DNA encoding full-length mElf-3 on its own or together with 5 g of the expression plasmid encoding truncated mELF-3(ATH ϩ DBD). The cells were also transfected with 1 g of pCMV-␤-gal plasmid to correct for any differences in transfection efficiency. The total DNA transfected was kept constant by addition of FNpcDNA3.0. Data are presented as fold activation relative to the activity observed in cells transfected with mT␤R-II(Ϫ108/ϩ56) alone. Data shown are means and S.D. for duplicate measurements from one representative transfection. This experiment was repeated twice, and similar results were obtained. B, Western blot analysis of cells transfected with expression vectors encoding fulllength mElf-3 protein or the truncated form of mElf-3(ATH-DBD). Western blot analysis was performed with an anti-FLAG monoclonal antibody as described under "Experimental Procedures." The sizes (kDa) and positions of molecular weight standards are indicated on the right. The predicted molecular mass of full-length and ATH ϩ DBD is 42.2 and 16.7 kDa, respectively, plus the mass of the FLAG epitope and nuclear localization sequence added to the N terminus. Western blot analysis was repeated in a separate experiment, and similar results were obtained.
Elf-3 and Regulation of the Type II TGF-␤ Receptor was expressed at a higher level than the smaller truncated protein containing only the DBD.
Importantly, the DNA binding characteristics of mElf-3 and its truncated counterparts do not appear to be cell type-specific. Similar results were obtained with nuclear extracts prepared from 293T cells, where ectopic expression of the three proteins was similar to each other (data not shown). Furthermore, similar findings were observed using F9-differentiated cells. For this purpose, F9-differentiated cells were transfected with either the expression vector for mElf-3 or the expression vector for the truncated form of Elf-3 containing only the DBD. Although we did not observe the formation of a DNA-protein complex using nuclear extract prepared from cells transfected with full-length mElf-3, we observed a prominent DNA-protein complex using nuclear extracts prepared from cells transfected with the DNA binding domain of mElf-3 (Fig. 8C). Moreover, as in the case of HeLa cells and 293T cells, a second DNA-protein complex, which migrated more slowly, was observed when a higher concentration of the nuclear extract was tested. Hence, it appears that mElf-3, like several other members of the Ets family of transcription factors, contains one or more domains that limit its binding to DNA in vitro.
Although full-length mElf-3 does not bind to DNA in vitro, it should be able to bind in vivo, because it stimulates the expression of the mT␤R-II promoter/reporter constructs (Figs. 5-7). To test this directly, we examined whether mElf-3 could bind to the EBS in the T␤R-II promoter in the promoter/reporter gene construct T␤R-II(Ϫ108/ϩ56) using ChIP analysis. This study was performed in 293T cells, because they can be transfected at much higher efficiency than F9-differentiated cells and because mElf-3 stimulates the T␤R-II promoter in these cells. In addition, because an antibody to Elf-3 was not available, 293T cells were transfected with expression vectors that code for FLAG epitope-tagged mElf-3 (full-length) or for FLAG epitope-tagged mElf-3 DNA binding domain. For the immunoprecipitation step, two antibodies were employed, the M2 antibody, which recognizes the FLAG epitope, and a control antibody, which recognizes the DNA binding domain of the yeast transcription factor Gal4. Both antibodies were conjugated to agarose beads. ChIP analysis indicates that both mElf-3 and its DNA binding domain can bind to the T␤R-II promoter in vivo (Fig. 9). Relative to the Gal4 antibody control, we observed nearly a 15-fold enrichment of T␤R-II promoter DNA with the M2 antibody. This was true for cells transfected with either the mElf-3 expression vector or the expression vector for the mElf-3 DNA binding domain. To confirm these results, this experiment was also performed with M2 and Gal4 antibodies that were not conjugated to beads. In this case, the antibody-protein-DNA complexes were collected using protein G-agarose beads. By using this protocol, we observed a 10-fold enrichment of T␤R-II promoter DNA when the cells were transfected with the expression vector for mElf-3 and an 8-fold enrichment when the cells expressed its DNA binding domain (data not shown).

FIG. 8. Gel mobility shift analysis with the EBS of mT␤R-II.
A, gel mobility shift analysis was performed with nuclear extracts prepared from HeLa cells transfected with expression vectors for mElf-3 or truncated forms of mElf-3 (ATH ϩ DBD or DBD) as described under "Experimental Procedures." The arrow indicates the supershifted DBD complex. B, immunoblots were performed with the M2 anti-FLAG monoclonal antibody (mAb) and nuclear extracts used in A. The sizes (kDa) and positions of molecular weight standards are indicated on the left. The predicted molecular mass of full-length, ATH ϩ DBD, and DBD is 42.2, 16.7, and 12.3 kDa, respectively, plus the mass of the FLAG epitope tag and the nuclear localization signal inserted at the N terminus of the protein. The relative levels of mElf-3, mElf-3(DBD), and mElf-3(ATH ϩ DBD), which were determined as described under "Experimental Procedures," are shown below in parentheses. C, gel mobility shift analysis was performed with nuclear extracts prepared from mock-transfected F9-differentiated cells or F9-differentiated cells transfected with expression vectors for mElf-3 or truncated form of mElf-3 (DBD) as described under "Experimental Procedures." Nuclear extracts were added at 5 l (21.5 g) or 12 l (51.6 g) as indicated. The asterisk indicates the less intense DNA-protein complex observed when a higher concentration of nuclear extract from the cells transfected with the mElf-3 DBD was used. Competitor DNAs (wild type and mutant) described under "Experimental Procedures" were added at 100-fold molar excess.

FIG. 9. Binding of mElf-3 to the T␤R-II promoter in vivo.
293T cells were transfected with the promoter/reporter construct T␤R-II(Ϫ108/ϩ56) and a FLAG-tagged expression vector for full-length mElf-3 (FL) or its DNA binding domain (DBD). ChIP analysis was performed as described under "Experimental Procedures." Quantification of the PCR products was performed using Kodak Electrophoresis Documentation and Analysis System and Kodak ID Image Analysis Software.
Hence, although full-length mElf-3 does not bind to DNA in vitro, it can do so in vivo.
Transactivation Domain of mElf-3 and the Steady-state Levels of mElf-3-A critical domain required for the function of virtually all transcription factors is their transactivation domain. Previous studies (52) employed Gal4-hELF-3 fusion proteins, containing different domains of hELF-3, to localize the position of its transactivation domain. This work identified a 13-amino acid acidic core, which is able to form an ␣-helical conformation in polar solvents (52). Moreover, it was determined that changes in two specific amino acids within the acidic core (D134A,L143P) greatly reduced transactivation by the Gal4-hELF-3 fusion protein, which contained hELF-3 amino acids 129 -159. Alignment of the amino acid sequences for hELF-3 and mElf-3 demonstrated that they are highly homologous in this region and that there is perfect conservation of the 13-amino acid acidic region (Fig. 10A). This prompted us to examine whether comparable changes in mElf-3 would influence its ability to transactivate. For this purpose, we employed mElf-3 itself rather than Gal4/fusion proteins. Specifically, we used site-directed mutagenesis to prepare two modified forms of mElf-3. In one of the modified forms, aspartate at position 133 was converted to alanine (mElf-3(D133A)). In the second, leucine 142 was converted to proline (mElf-3(L142P)). Initially, we compared the transactivation activity of each mutant to the activity of mElf-3 in F9-differentiated cells (Fig. 10B). Although mElf-3(D133A) was only slightly less active in stimulating the T␤R-II(Ϫ108/ϩ56) promoter/reporter gene construct, mElf-3(L142P) was nearly 5-fold less active than unmodified mElf-3. Hence, this study provides further evidence that the transactivation domain of intact mElf-3, like its Gal4/fusion protein counterpart, appears to be localized to an acidic region of the protein.
It is possible that the weak transactivation by mElf-3(L142P) was due to its expression being lower than that of unmodified mElf-3. We examined this possibility in 293T cells, which were transiently transfected with expression vectors for mElf-3, mElf-3(D133A), or mElf-3(L142P) (Fig. 10C). As in F9-differentiated cells, mElf-3(D133A) and mElf-3(L142P) exhibited lower levels of transactivation than mElf-3. Moreover, they were expressed at higher levels than that of mElf-3 (Fig. 10D). In the case of mElf-3(L142P), its ability to transactivate was nearly 5-fold lower than that of mElf-3, even though it was expressed at a level that was 3.7-fold higher than mElf-3. This finding not only supports the argument that the acidic region of mElf-3 (residues 128 -159) is required for transactivation, it also suggests that there is an inverse relationship between the transactivation by mElf-3 and its steady-state level. The latter finding is particularly interesting given recent findings that the turnover of several transcription factors correlates inversely with their ability to activate gene expression (53)(54)(55). This intriguing possibility is discussed below. DISCUSSION Previous studies (20) demonstrated that the differentiation of EC cells leads to up-regulation of the mT␤R-II gene. In this study, we cloned and functionally characterized the 5Ј-flanking region of the mT␤R-II gene in differentiated cells derived from F9 EC cells. Analysis of the mT␤R-II promoter by transient transfection of promoter/reporter gene constructs demonstrated the importance of a cis-regulatory element located between ϩ2 and ϩ56, which maps to the EBS. In addition, we demonstrate that the transcription factor mElf-3 binds to the EBS and strongly activates the mT␤R-II gene. Importantly, we demonstrate that differentiation of EC cells leads to the upregulation of both the mT␤R-II gene and the mElf-3 gene. We also provide evidence that mElf-3 is subject to two other levels of regulation, one that affects DNA binding and one that influences its steady-state level of expression.
mElf-3 is one of over 30 members in the Ets family of transcription factors. Sequence homology within this family is primarily limited to a highly conserved 80 -90-amino acid DNA The underlined residues represent a 13-residue acidic transactivation core in human ELF-3. Amino acids 133 and 142 in the mouse sequence are underlined. B, F9-differentiated cells were transfected with 15 g of the mT␤R-II promoter/reporter gene construct mT␤R-II(Ϫ108/ϩ56) and 1 g of an expression vector for mElf-3, mElf-3(D133A), or mElf-3(L142P). The cells were also transfected with 1 g of pCMV-␤-gal plasmid to correct for any differences in transfection efficiency. The total DNA transfected was kept constant by addition of FNpcDNA3.0. Transactivation is represented as relative activity, compared with the cells transfected with mT␤R-II(Ϫ108/ϩ56) construct alone. Data shown are means and S.D. for duplicate measurements from one representative transfection. This experiment was repeated twice, and similar results were obtained in each case. C, 293T cells were transfected with 15 g of the mT␤R-II promoter/reporter gene construct mT␤R-II(Ϫ108/ϩ56) and 3 g of expression vector for mElf-3, mElf-3(D133A), or mElf-3(L142P). The cells were also transfected with 1 g of pCMV-␤-gal plasmid to correct for any differences in transfection efficiency. The total DNA transfected was kept constant by addition of FNpcDNA3.0. This experiment was repeated, and similar results were obtained. D, Western blot analysis was performed with the M2 anti-FLAG monoclonal antibody and the nuclear extracts used in C. The relative levels of mElf-3, mElf-3(D133A), and mElf-3(L142), which were determined as described under "Experimental Procedures," are shown in parentheses. This experiment was repeated, and similar results were obtained.
binding domain (the Ets domain), which interacts with a purine-rich GGA(A/T) or a GGA core sequence flanked by different nucleotides that determine the affinity and specificity of binding for a particular Ets family member (37). This family of transcription factors plays critical roles in the transcriptional regulation of genes involved in tissue development, angiogenesis, immune response, cell cycle regulation, cell proliferation, and apoptosis (56). ELF-3, also named ESE-1 (57), ESX (58), Jen (59), or ERT (22), is the prototype of a new subclass of Ets transcription factors, because it exhibits no striking homology to any particular subclass of the Ets family of transcription factors. The most obvious structural difference distinguishing mouse and human Elf-3 from other Ets family members is the presence of an AT-hook domain, which was first described in the high mobility group, non-histone chromosomal protein HMG-I(Y) (43). Until now, hELF-3 was believed to be expressed exclusively in epithelial cells (39,57,59,60). In this study, we demonstrate low expression of mElf-3 in F9 EC cells and higher levels of mElf-3 expression when EC cells are induced to differentiate into cells that exhibit the properties of parietal extra-embryonic endoderm, which is one of the first cell types to form during mammalian embryogenesis. In addition to its activation of the T␤R-II gene, hELF-3 has been reported to activate the expression of several markers of epidermal cell differentiation, including SPRR2A, transglutaminase 3, and profilaggrin (57,59).
The work described in our study provides three lines of evidence that mElf-3 plays a key role in the regulation of the mT␤R-II gene in F9-differentiated cells. First, our co-transfection studies demonstrate that mElf-3 strongly stimulates the activity of mT␤R-II promoter/reporter constructs containing functional EBS. Deletion of the region between nucleotides ϩ2 and ϩ56 or disruption of the EBS by site-directed mutagenesis eliminates the response of mT␤R-II promoter/reporter gene constructs to mElf-3. Second, the dominant negative forms of mElf-3 not only drastically reduce the stimulation in response to mElf-3, but also reduce the basal level of the mT␤R-II promoter/reporter gene constructs in F9-differentiated cells. In this regard, it should be noted that this is the first report of a dominant negative form of ELF-3. Third, and very importantly, northern blot analysis indicates that increases in the steadystate levels of mElf-3 mRNA closely parallel increases in mT␤R-II mRNA, especially during the first 48 h of differentiation. In fact, this finding suggests that increases in mElf-3 play a key role in the mechanisms responsible for the upregulation of the mT␤R-II gene. In the future, we plan to monitor the levels of mElf-3 in nuclear extracts by western blot analysis. These studies will be performed once suitable antibodies for mElf-3 become available. Interestingly, northern blot analysis also demonstrated that the steady-state levels of the mT␤R-I gene do not increase during the differentiation of F9 EC cells. If the sole role of mT␤R-I is to form a heterodimer with mT␤R-II for the purpose of mediating the effects of TGF-␤, one would expect to see parallel increases in transcripts for both receptors. Hence, mT␤R-I may play a separate role in EC cells.
Previous studies (22,31,57,58) demonstrated hELF-3 binding to DNA in vitro. Each of these studies involved hELF-3 produced by in vitro translation or recombinant hELF-3 purified from bacteria. In this study, we provide the first direct evidence that mElf-3 binds to the T␤R-II promoter in vivo (Fig.  9). In addition, we demonstrate that mElf-3 contains one or more domains that negatively influence its binding in vitro. Specifically, we did not observe in vitro DNA binding of fulllength mElf-3 produced in HeLa cells (Fig. 8A), F9-differentiated cells (Fig. 8C), or 293T cells (data not shown). In contrast, a truncated mElf-3 containing the DNA binding domain, but not the other domains, binds readily to DNA in vitro. Thus, mElf-3 joins a growing list of Ets family members that contain domains that influence their binding to DNA. To date, at least eight other members of the Ets family of transcription factors (SAP-1, Ets-1, Ets-2, Net, Elk-1, ERM, ESE-2, and PEA3) have been shown to possess binding regulatory domains (45)(46)(47)(48)(49)(50)(51). In the case of Ets-1, the autoinhibitory domains have been mapped and studied extensively (61)(62)(63)(64).
Several mechanisms could account for the regulation of mElf-3 binding to DNA. Expression of mElf-3 in mammalian cells could lead to post-translational modifications, such as phosphorylation, which modulate its binding to DNA. For example, phosphorylation of Ets-1 outside of its DNA binding domain reduces DNA binding 50-fold (62). Conversely, phosphorylation of the C-terminal transcriptional activation domain of Elk-1 enhances both DNA binding and transcriptional activation (65). Alternatively, certain cells, such as the HeLa cells and 293T cells used in our DNA binding studies, may express a cofactor that binds to mElf-3 and limits its binding to DNA in vitro. We think this is unlikely given that mElf-3 stimulates T␤R-II promoter/reporter gene constructs in F9differentiated cells (Fig. 5) and 293T cells (Figs. 7 and 10). In contrast to an inhibitory cofactor, mElf-3 binding to DNA may require cooperative binding with a positive cofactor that does not interact with mElf-3 in vitro under the conditions employed in our gel mobility shift assays. In fact, cooperative binding of transcription factors is believed to play a central role in promoting gene activation and, at the same time, preventing promiscuous activation of inactive promoters. A good case in point is Ets-1 and CBF␣2/AML1. Interaction of these two proteins counteracts the autoinhibitory domains of Ets-1 and, as a result, stimulates Ets-1 binding to DNA and enhances Ets-1 transactivation (64,66). Similarly, binding of USF-1 to PEA3 acts as a switch that modifies the autoinhibitory motifs in PEA3 to promote DNA binding activity (67). Furthermore, other studies (68 -70) have shown that the addition of monoclonal antibodies that recognize specific Ets transcription factors significantly increases Ets protein binding to their cognate DNA target. Clearly, one or more of these mechanisms could regulate mElf-3 binding to DNA, and further study of the mechanism(s) involved is warranted. Understanding the mechanisms involved is likely to shed important light on the ability of mElf-3 to regulate the T␤R-II promoter.
The work reported in this study also extends our understanding of the transactivation domain of mElf-3. Previous studies (52) identified a small acidic region of hELF-3 (amino acids 129 -159) that could strongly stimulate a viral promoter when fused to the DNA binding domain of the yeast transcription factor Gal4. In our study, we have extended this finding by focusing on mElf-3 itself and one of its likely in vivo targets. Specifically, we demonstrate that a single amino acid substitution at residue 142 (L142P) is sufficient to cause a 5-fold reduction in the ability of mElf-3 to stimulate the mT␤R-II promoter. This substitution is particularly interesting, because it has been proposed to alter the ability of this region in hELF-3 to form an ␣-helical conformation (52).
In the course of these studies, we determined that the steady-state levels of mElf-3 were consistently lower than mutant forms of mElf-3, in particular mElf-3(L142P), which exhibits significantly lower transactivation activity. Similarly, the truncated forms of mElf-3, which act as dominant negatives, also exhibited significantly higher steady-state levels of expression than mElf-3 (Fig. 8B). Currently, the mechanisms responsible have not been determined. However, it may be the result of a slower rate of mElf-3 turnover. Recent studies (54,55,71) have shown that many transcription factors turn over by a proteosome-mediated pathway and that the rate of transcription factor degradation is inversely correlated with their ability to activate transcription. Interestingly, the domains responsible for turnover of transcription factors have been shown to overlap with their transactivation domains (72), even though they are ubiquitinated at other sites. Furthermore, recent studies (55) suggest that ubiquitination of at least some transcription factors serves a dual role of stimulating both transactivation and proteosome-mediated turnover. Given these new developments, further studies regarding the connection between mElf-3 transactivation and turnover are warranted.
In conclusion, the work reported in this study argues strongly that mElf-3 plays a key role in the regulation of the mT␤R-II promoter. Our work also suggests that the up-regulation of mElf-3 is responsible, at least in part, for the upregulation of the mT␤R-II gene during the differentiation of EC cells. Furthermore, our studies suggest that mElf-3 is subject to at least two other levels of regulation. Given these findings, we suggest that determining the mechanisms used to regulate the expression, DNA binding, and turnover of ELF-3 could help identify methods to elevate the levels of T␤R-II in diseased tissues where its expression is aberrantly low.