GATA-6 Activates Transcription of Thyroid Transcription Factor-1*

Thyroid transcription factor-1 (TTF-1) is expressed in respiratory epithelial cells, where it regulates the transcription of target genes expressed in a cell-selective manner. GATA-5 and -6, members of the zinc finger family of transcription factors, are also expressed in various cell types within in the developing lung. In the present work, GATA-6 mRNA was detected in adult mouse lung, purified mouse type II epithelial cells, and differentiated mouse pulmonary adenocarcinoma cells (MLE-15 cells), being co-expressed with TTF-1 mRNA. In order to test whether GATA factors regulated TTF-1 gene transcription, GATA-5 and -6 expression vectors were co-transfected with TTF-1 luciferase expression vector. GATA-6, but not GATA-5, markedly activated TTF-1 gene transcription in HeLa cells. EMSA and supershift analysis with GATA-6 antiserum demonstrated that GATA-6 in MLE-15 cell nuclear extracts bound to an element located 96–101 base pairs from major start of TTF-1 gene transcription. Site directed mutagenesis of the GATA element in the TTF-1 promoter region inhibited transactivation by GATA-6 in HeLa cells. GATA-6 is co-expressed with TTF-1 in the respiratory epitheliumin vivo and respiratory epithelial cells in vitro. GATA-6 strongly enhanced activity of the human TTF-1 gene promoter in vitro. These findings support the concept that GATA-6 may play an important role in lung cell differentiation and gene expression, at least in part by altering the expression of TTF-1 and its potential targets.

and in lung (surfactant proteins A, B, and C (SP-A, SP-B, and SP-C) and Clara cell secretory protein (CCSP)) (8 -12). TTF-1 also plays a critical role in lung morphogenesis. Deletion of TTF-1 in transgenic mice causes severe thyroid and lung hypoplasia (13). Thus, TTF-1 modulates both lung morphogenesis and respiratory epithelial cell gene expression in the developing and postnatal lung.
TTF-1 is expressed in the progenitor cells of the lung buds early in rat, mouse, and human development. TTF-1 is encoded by a single gene in humans located on chromosome 14q13 (1). While the elements governing TTF-1 expression have not been defined with certainty, three distinct transcriptional start sites within the proximal promoter region and the expression of alternatively spliced TTF-1 cDNAs have been described in the rat and human TTF-1 genes (14 -18). The 5Ј regulatory regions of the human and rat TTF-1 genes have been studied in vitro (15,19). TTF-1 transcription was enhanced by HNF-3␤, which bound to two distinct sites located at bp Ϫ135 to Ϫ124 and Ϫ14 to Ϫ3 relative to the major start of transcription in the human gene, suggesting that TTF-1 regulation may be influenced by other transcription factors expressed in the foregut endoderm (19).
The GATA family of transcription factors includes a family of zinc finger domain-containing nuclear proteins that includes at least six distinct proteins that are expressed in a variety of tissues. GATA family members, particularly GATA-1, -2, and -3, were initially identified as important in the control of gene regulation during hematopoiesis (20). Gene targeting of GATA-1, -2, and -3 demonstrated their important roles in the development of the hematopoietic system, being critical for gene expression and development of erythroid cell lineages in the developing mouse embryo (21)(22)(23)(24). More recently, several of the GATA family members were identified in nonhematopoietic sites including the heart, gastrointestinal, and genitourinary tracts. The finding that GATA-4, -5, and -6 are expressed in the foregut or cardiac regions of the developing embryo suggested their potential role in organogenesis and gene expression in the heart and lung (25)(26)(27). GATA-5 and -6 are expressed in developing mouse lung, GATA-6 being detected primarily in epithelial cell respiratory tubules of the fetal lung, in a pattern similar to that previously described for TTF-1. The finding that, in the heart, Nkx2.5 (structurally closely related to TTF-1) and GATA-4 interact in the regulation of cardiac specific gene expression, suggested that GATA family members and TTF-1 might interact in respiratory epithelial cells (28).
The present work was undertaken to determine whether GATA-5 or -6 was expressed in respiratory epithelial cells in vivo and in vitro and whether they might play a role in the regulation of TTF-1 gene expression, thereby influencing TTF-1 transcriptional targets. GATA-6, but not GATA-5, was expressed at high levels in the mouse pulmonary adenocarcinoma cells in vitro and in freshly isolated rat type II epithelial cells. GATA-6 activated a regulatory region of the mouse TTF-1 gene in HeLa cells and in human pulmonary adenocarcinoma cells (H441 cells).

EXPERIMENTAL PROCEDURES
RNA Isolation and Northern Blot Analysis-Cell lines were maintained in standard tissue culture conditions prior to isolation of RNA. Rat type II epithelial cells were isolated after protease digestion of adult rat lung as described previously (29). Total RNA was isolated by an adapted method of Chirgwin et al. (30). Cells were lysed directly on the tissue culture plate using 4 M guanidine thiocyanate, 0.5% N-lauroylsarcosine, 25 mM sodium citrate, and 0.1 M ␤-mercaptoethanol. Tissue was homogenized in the same buffer. Phase lock gels (5 Prime 3 3 Prime, Inc., Boulder, CO) were used to prepare RNA. RNA was quantitated by absorbance at 260 nm. Fifteen g of total RNA were loaded onto a 0.8% formaldehyde gel and transferred to Maximum strength Nytran (Schleicher & Schuell) and bound to the filter by UV crosslinking in a Stratalinker (Stratagene, La Jolla, CA). Hybridization was done in 50% formamide, 5ϫ SSPE, 2.5ϫ Denhardt's solution, 1% SDS, and 200 g/ml salmon sperm DNA at 42°C.
Identification of Genomic Clones-A 129SvJ mouse liver genomic DNA library (Lambda DASH II, Stratagene) was provided by Marcia M. Shull, University of Cincinnati. Approximately 3.4 genome equivalents were plated and lifted onto Hybond N (Amersham Pharmacia Biotech). This was screened with a 1.3-kb rat TTF-1 cDNA provided by Dr. R. Di Lauro (Naples, Italy). The probe was labeled by random primer labeling and hybridized at a concentration of 10 6 dpm/ml in 6ϫ SSC, 5ϫ Denhardt's, 0.1% SDS, and 0.1 mg/ml salmon sperm DNA at 55°C. The final wash conditions were 0.5ϫ SSC, 0.1% SDS at 55°C for 30 min. Four clones contained the TTF-1 gene, the largest clone containing 1.7 kb of 5Ј-flanking sequence. Sequence of the promoter region was determined by using the dideoxy sequencing reaction product (Sequenase, U.S. Biochemical Corp.). Restriction fragments were used to compare the maps of the clone with mouse genomic DNA.
TTF-1 Luciferase Plasmid Constructions and Mutagenesis-Primers were used to amplify the desired region of the TTF-1 gene. The 3Ј-end of all constructs included all of the 5Ј-untranslated region up to, but not including, the ATG and had a HindIII site inserted in to aid in cloning. The DNA fragment was subcloned into pGL2-Basic (Promega, Inc., Madison, WI).
To make mTTF-1-142 luciferase, primers ϩ200 to ϩ185 and Ϫ168 to Ϫ149 were used on a mouse genomic clone. The polymerase chain reaction product was digested with HindIII and cloned into pGL2-Basic (Promega). To make mTTF-1-110 luciferase primers, ϩ200 to ϩ185 and Ϫ110 to Ϫ87 were used. To make mTTF-95, primers ϩ200 to ϩ185 and Ϫ95 to Ϫ72 were used. To make mTTF-1-87 luciferase primers, ϩ200 to ϩ185 and Ϫ87 to Ϫ69 were used. To make a construct mutant for GATA binding, primers ϩ200 to ϩ185 and Ϫ110 to Ϫ87 mutant primer and in a separate reaction primers GL1 and Ϫ87 to Ϫ110 mutant primer were used on a mTTF-1-142 template. Products were gel-purified and the combined in a polymerase chain reaction with primers ϩ200 to ϩ185 and GL1. This product was digested with HindIII, cloned into pGL2-Basic. The resultant plasmid was called mTTF-1-142 GATA mutant luciferase. The veracity of mutations was confirmed by both restriction analysis and nucleotide sequencing of double-stranded templates.
Cell Transfection, Luciferase, and ␤-Galactosidase Assays-Cell transfection was performed by the calcium phosphate coprecipitation method except that glycerol shock was not used (31). Transfections were done in six-well plates with 0.5 pmol of reporter luciferase construct, 0.2 pmol of transactivator (or empty vector), and 0.1 pmol of CMV-␤galactosidase to normalize for transfection efficiency. CMV-rTTF-1 was a gift of Dr. R. Di Lauro (Stazione Zoologica, A. Dohrn, Naples, Italy), and CMV-GATA-5 and CMV-GATA-6 were a gift of Dr. J. Molkentin (Division of Molecular Cardiology, Children's Hospital Medical Center, Cincinnati, OH). Fresh control medium was added approximately 20 h after transfection. Cells were harvested in Reporter Lysis Buffer (Promega, Madison, WI). The lysates were prepared and aliquots assayed for ␤-galactosidase and luciferase activity with a luminometer (Analytical Luminescence Laboratory, San Diego, CA). To correct for variations in transfection efficiency, luciferase activity was normalized to ␤-galactosidase activity. The normalized luciferase activity of the mTTF pro-moter was divided by the normalized luciferase activity of pGL2-Basic in the presence of the appropriate transactivating plasmid to determine relative luciferase activity.
Preparation of Nuclear Extracts-Cell lines were maintained in standard tissue culture conditions prior to making the nuclear extracts. Cells were cooled to 4°C, washed in phosphate-buffered saline, and scraped off of the plates. Cells were spun at 3000 rpm at 4°C for 5 min and then washed again twice with phosphate-buffered saline before lysing. Cells were then resuspended in three packed cell volumes of lysis buffer consisting of 10 mM HEPES, pH 7.9, 10 mM KCl, 0.1 mM EDTA, 1.5 mM MgCl 2 , and 0.2% Nonidet P-40, 100 M dithiothreitol, and 50 M phenylmethylsulfonyl fluoride. Cells were then vortexed at medium speed for 10 s and incubated on ice for 5 min. Nuclei were then pelleted at 3000 rpm and 4°C for 5 min. The pellet fraction contained the nuclei. One packed volume of extract buffer consisting of 20 mM HEPES, pH 7.9, 420 mM NaCl, 0.1 mM EDTA, 1.5 mM MgCl 2 , 25% glycerol, 100 M dithiothreitol, and 50 M phenylmethylsulfonyl fluoride was added to the nuclei and vortexed at medium speed every minute for 10 min. The nuclear extracts were then centrifuged at 15,000 rpm for 5 min, aliquoted into chilled tubes, quick frozen in liquid nitrogen, and stored at Ϫ80°C. Protein concentration was determined by the method of Lowry (32).
Annealing of Synthetic Oligonucleotides-Oligonucleotides were annealed at a concentration of 250 mM in 250 mM Tris (pH 7.8) at 95°C for 15 min and then cooled slowly to room temperature overnight and stored at Ϫ20°C. The annealed oligonucleotides were diluted to 5 mM in the same buffer. Probes were labeled using [␥-32 P]ATP and T4 polynucleotide kinase. End-labeled probes were purified from unincorporated [␥-32 P]ATP by an Amersham Pharmacia Biotech Nick column eluted in a 400-l volume. Unlabeled probe was placed on a Nick column at a 10-fold greater concentration and also eluted in 400 l. Oligonucleotides used in shifts consisted of nucleotides Ϫ110 to Ϫ87 and their complement Ϫ87 to Ϫ110, and the GATA mutant consisting of the nucleotide Ϫ110 to Ϫ87 region (and complement containing the mutation TTATCT 3 TGAATT). Oligonucleotides were generated from the published sequence of the histidine decarboxylase promoter (33) consisting of the sequence TACTGCTGATAAGGAAA and complement TTTCCTTATCAGCAGTA.
Electrophoretic Mobility Gel Shift (EMSA) Assay-The procedure for EMSA was adapted from methods previously described (34). Briefly, 5 g of extract and 10 l of unlabeled oligonucleotide competitor DNA (125 nM) were incubated in the same volume of buffer consisting of 50 mM HEPES, pH 7.9, 100 mM KCl, 4 mM EDTA, 200 ng/ml poly(dI-dC), 4 mM dithiothreitol, 20 mM MgCl 2 , 40% glycerol, and 2 mM phenylmethylsulfonyl fluoride for 10 min on ice. One l of 12.5 nM labeled probe was added, and the mixture was incubated for another 10 min on ice. Supershifts were performed by preincubating anti-GATA-6 antiserum or control rabbit serum with the labeled oligonucleotide under similar assay conditions. GATA-6 rabbit polyclonal antiserum was kindly provided by Dr. Kenneth Walsh (Tufts University, Boston). Bound and free probe were resolved by nondenaturing polyacrylamide gel electrophoresis. Recombinant rat TTF-1 homeodomain and supershift with rabbit anti-TTF-1 were performed as described previously (9). Five percent gels (acrylamide/bisacrylamide, 29:1, 0.5ϫ TBE, 2.5% glycerol, 1.5 mm thick) were run in 0.5ϫ TBE running buffer at constant current (30 mA) for approximately 85 min. Gels were blotted to Whatman 3 MM paper, dried under vacuum, and exposed to x-ray film from 4 h to overnight at room temperature.

GATA-6 and TTF-1 Are Co-expressed in Respiratory
Epithelial Cells-Northern blot analysis of mRNA obtained from various respiratory epithelial cells and nonpulmonary cells was used to assess expression of GATA-5, GATA-6, and TTF-1 mRNAs. GATA-5 and -6 mRNAs were readily identified in RNA isolated from whole mouse lung, migrating as 8.0-and 3.2-kb species, respectively (Fig. 1). GATA-6, but not GATA-5, mRNAs were detected at high concentrations in freshly isolated rat type II epithelial cells, previously shown to express TTF-1 and varying amounts of TTF-1 target genes including surfactant protein A, B, and C (14,19). GATA-6 and TTF-1 were also co-expressed in various SV40 large T immortalized clonal mouse lung epithelial cell lines, including MLE-15 and MLE-F6 cells; however, GATA-6 was not readily detected in H441 human pulmonary adenocarcinoma cells by Northern blot analysis. GATA-6 mRNA was detected in 3T3 cells and in SV40 large T antigen immortalized fetal mouse lung mesenchymal cells (mFLM cells). Neither GATA-6 nor TTF-1 mRNA was detected in HeLa cells (data not shown). GATA-5 mRNA was not detected in H441, MLE pulmonary adenocarcinoma cells or in type II cells. Thus, GATA-6 and TTF-1 were coexpressed in various differentiated pulmonary adenocarcinoma cells and in freshly isolated rat type II cells, consistent with the finding that both mRNAs were detected in subsets of respiratory epithelial cells in the fetal mouse lung (26,27). In contrast, GATA-5 mRNA was not readily identified in the immortalized respiratory epithelial cells or in type II cells, consistent with its localization in nonepithelial cells in the developing mouse lung as assessed by in situ hybridization (26,27). The co-localization of GATA-6 and TTF-1 with TTF-1 targets including SP-A, SP-B, and SP-C in various respiratory epithelial cells is consistent with the possibility that GATA-6 may influence transcription of TTF-1 or its target genes.
5Ј Regulatory Region of the Mouse TTF-1 Gene Contains a GATA Regulatory Element-Inspection of the 5Ј region of the mouse TTF-1 gene revealed the nucleotide sequence TTATCT, located at Ϫ101 to Ϫ96 relative to the major transcription start site, consistent with its role as a potential GATA cis-acting element. Previous work from this laboratory identified a large DNase protected site by in situ footprint analysis with nuclear extracts from MLE-15 cells in this region (14). The nucleotide sequence of this region of the TTF-1 gene is highly conserved in human rat and mouse TTF-1 genes. The potential GATA site was located near the element TGTTT, previously demonstrated to bind HNF-3␤, activating TTF-1 gene promoter activity (19). The potential GATA site was also located close to a potential TTF-1 site (CAAG), located 5 bp from the GATA site. To test whether GATA family members might regulate TTF-1 promoter activity, the 5Ј region of the mouse TTF-1 gene was ligated to the luciferase genes to produce mTTF-1-142. This region of the TTF-1 gene contains 142 bp 5Ј of the primary transcription start site, as well as 5Ј-untranslated sequence derived from the major TTF-1 transcript (Fig. 2). This 5Јuntranslated region encodes two alternative transcripts and includes a small potential intron (15).
Transient co-transfection experiments were performed in which mTTF-1-142 was co-transfected with expression vectors encoding mouse GATA-5 (CMV-GATA-5) and GATA-6 (CMV-GATA-6) into HeLa cells. mTTF-1-142 was markedly activated by co-transfection with GATA-6 but not GATA-5 (Fig. 3) Likewise, cotransfection of GATA-6 activated the mTTF-1-142 luciferase construct after transfection into H441 adenocarcinoma cells from 17 Ϯ 2.3 to 28 Ϯ 3.6 luciferase units (mean Ϯ S.E., p Ͻ 0.01). Deletion mutations were made producing vectors at Ϫ110, Ϫ95, and Ϫ87 bp relative to the start of transcription of the TTF-1 gene. In HeLa cells, significant stimulation of TTF-1 promoter activity by GATA-6 was observed with 5Ј sequences from Ϫ110 bp; however, maximal basal and stimulated activities were observed when a larger region (Ϫ142 bp) was included. The nucleotide sequences between Ϫ110 and Ϫ95 contained a region that encodes a consensus GATA binding site TTATCT (Fig. 4). Mutation of the GATA site in mTTF-1-142 (TTATCT to TGAATT) completely inhibited transactivation by GATA-6 after co-transfection in HeLa cells (Fig. 4). Thus, a GATA-6 response element was located at Ϫ101 to Ϫ96 in the murine TTF-1 promoter. In contrast to the findings in HeLa cells, transfection of MLE-15 cells with CMV-GATA-6 did not further transactivate the wild type TTF-1-142 construct, perhaps related to the high levels of exogenous GATA-6 expression in this cell line. Luciferase activity of the mTTF-1-142 (TGAATT mutant) construct was similar to the wild type construct in MLE-15 cells (data not shown).
GATA-6 Binds to the mTTF-1 Gene-Electromobility shift assays were performed using nuclear extracts from MLE-15, MLE-F6, 3T3, and H441 and HeLa cells (Fig. 5). A gel shift was readily detected with an oligonucleotide Ϫ110 to Ϫ87 containing the GATA site with nuclear extracts from MLE-15, MLE-F6, and 3T3 cells consistent with the detection of GATA-6 mRNA as determined by Northern blot analysis. This DNA complex was lacking in HeLa and H441 cells, also consistent with the lack of GATA-6 mRNA in these cells. The protein-DNA complex was inhibited by paucity or absence of the unlabeled self-oligonucleotide. The same oligonucleotide containing the mutated GATA site (Ϫ110 to Ϫ87) (TTATCT 3 TGAATT) did not compete for binding to MLE-15 cell nuclear extracts (Fig. 5,  lane 3). An oligonucleotide containing the GATA-6 site identified in the histidine decarboxylase gene (33) bound to MLE-15 nuclear extracts. Binding of the histidine carboxylase GATA-6 site was competed by self oligonucleotide and by the oligonucleotide containing the GATA-6 site from the TTF-1 gene. Formation of the complex with the histidine decarboxylase GATA-6 site was not inhibited by the oligonucleotide containing the mutant GATA site (Fig. 6), further demonstrating the specificity of the interaction.
GATA-6 Enhances TTF-1 Transcription Independently of TTF-1 or HNF-3␤-A canonical TTF-1 binding site, CAAG, is located in close apposition to the GATA-6 binding site in the murine TTF-1 gene. Since GATA-4 and Nkx2.5 are known to interact in a synergistic manner to activate myosin heavy chain gene expression in cardiac cells (28), we tested whether GATA-6 or TTF-1 might interact synergistically on the TTF-1 promoter. Expression vectors encoding rat TTF-1 (CMV-TTF-1) and GATA-6 (CMV-GATA-6) were co-transfected with the parental TTF-1-luciferase reporter construct (TTF-1-142) into HeLa cells. TTF-1 did not enhance, but slightly inhibited, GATA-6-dependent activity (Fig. 7). Thus, GATA-6 stimulation of TTF-1 transcription neither requires nor is synergized by TTF-1. Consistent with this observation, recombinant TTF-1 homeodomain did not bind to the CAAG element in the GATA site-containing oligonucleotide from the TTF-1 gene. Likewise, MLE-15 nuclear extracts did not bind to the potential TTF-1 site as assessed by supershift analysis (data not shown). While HNF-3␤ enhanced activity of the TTF-1-142 construct and HeLa cells, co-transfection of CMV-GATA-6 with an expression vector expressing rat HNF-3␤ (CMV-HNF-3␤) did not further influence the effects of GATA-6 on TTF-1 promoter activity. DISCUSSION TTF-1 and GATA-6 mRNAs were co-expressed in freshly isolated type II epithelial cells and immortalized respiratory epithelial adenocarcinoma cells in vitro. GATA-6, but not GATA-5, activated transcriptional activity of the murine TTF-1 promoter construct, GATA-6 binding to a cis-active element identified at Ϫ101 to Ϫ96 from the start of the major start site of TTF-1 gene transcription. GATA-6 activated transcription of TTF-1 independently of TTF-1 and HNF-3␤. The present findings support the concept that GATA-6 may influence lung differentiation and gene expression by modulating TTF-1 gene transcription, thereby potentially influencing the regulation of TTF-1 transcriptional targets.
Previous studies demonstrated that TTF-1 plays a critical FIG. 2. Nucleotide sequence of the 5 region of the mouse gene. Nucleotide sequence was determined as described under "Experimental Procedures." The downward arrows represent the 5Ј-end of the mTTF-1 luciferase constructs. The upward arrow represents the 3Ј-end, which stops just short of the ATG codon. The 5Ј-end of the fragments are located at Ϫ142, Ϫ110, Ϫ95, and Ϫ87 bp upstream of the major start site. The boldface asterisk denotes the major transcriptional start site, and the other two asterisks denote the other minor start sites as previously published (14,15). The previously identified HNF-3␤ and GATA-6 binding sites and the TTF-1 consensus site are illustrated. The alternatively spliced intron in the 5Ј-untranslated region is underlined. The mutant base pair changes for transfections and EMSA are shown above the GATA binding site. Transfections were normalized to ␤-galactosidase activity. Luciferase activity of the TTF-1 promoter construct was divided by the luciferase activity from the promoterless luciferase vector (pGL2-Basic) transfected with the appropriate transactivator. Values are mean Ϯ S.E. n ϭ 3, p Ͻ 0.0005 as assessed by unpaired t test versus pCDNA1.1.

FIG. 4.
A site of GATA-6 activation is located between ؊142 and ؊95. Deletion analysis was used to identify the site of GATA-6 activation on the mTTF-1 promoter. GATA-6-dependent activity was lost in deletions between Ϫ110 and Ϫ95 in the mTTF-1 promoter; however, maximal GATA-6-dependent activity was observed with the larger construct Ϫ142. A consensus site (TTACTC) for GATA binding was located at Ϫ96 to Ϫ101 in this region. This site was mutated to TGAATT and tested for GATA-6 transactivation in HeLa cells. Relative luciferase values were determined as described under "Experimental Procedures." The deletion and mutant constructs are further described under "Experimental Procedures" and in Fig. 1. Values represent mean Ϯ S.E. of three separate experiments. * and **, p Ͻ 0.001 and p Ͻ 0.005, respectively, versus pCDNA1.1 control. role in pulmonary morphogenesis and in the regulation of a number of gene products involved in pulmonary function and host defense. TTF-1 is expressed in the developing lung buds of the rat, mouse, human, and chick, the level of expression decreasing in the perinatal and postnatal period of lung development (2,4,5,35). TTF-1 is increasingly restricted in the postnatal lung, being detected primarily in type II epithelial cells and in subsets of bronchiolar epithelial cells (4). Thus, the concentration of TTF-1 is likely to influence both respiratory cell differentiation and the expression of target genes, including surfactant proteins A, B, and C and CCSP (8 -11), proteins that play important roles in host defense and surfactant function. TTF-1 binds to and activates the promoters of each of these genes, DNA binding and transcriptional activity of TTF-1 being further modified by phosphorylation and oxidationreduction (36 -38).
Co-localization of GATA-6 and TTF-1 in developing and mature respiratory epithelial cells and the finding that TTF-1 transcription is activated by GATA-6 are consistent with the potential interaction of the two gene families during lung development. A similar relationship was demonstrated between TTF-1 and HNF-3␤, the latter a member of the forkhead family of transcription factors. HNF-3␤ and TTF-1 are also co-expressed in respiratory epithelial cells of the embryonic and postnatal lung. Recent studies demonstrated that TTF-1 gene transcription was enhanced by HNF-3␤, a process mediated by binding of HNF-3␤ to a cis-active element located near the presently described GATA-6 binding site in the TTF-1 gene (19). Thus, distinct members of the HNF-3, GATA, and Nkx families may act in combination on target genes in the foregut endoderm. The phylogenetic conservation of such a pathway, wherein members of these same families of transcription fac-tors interact, is supported by recent findings in Caenorhabditis elegans. Forkhead homologue pha-4, GATA homologue elt-2, and Nkx homologue ceh-22 interact in a transcriptional pathway mediating pharyngeal and gut development in the worm (39). In the present study, GATA-6 activation of the TTF-1 gene transcription did not require expression of TTF-1 or HNF-3␤, neither being expressed in HeLa cells (14). 2 TTF-1, HNF-3␤, and GATA-6 are expressed in a similar distribution in respiratory epithelial cells in the fetal lung in vivo and in MLE-15 cells in vitro. Expression of TTF-1, HNF-3␤, and GATA-6 in fetal lung overlaps with that of the SP-A, SP-B, and SP-C genes. Recent findings from this laboratory demonstrated that GATA-6 activated transcriptional activity of SP-A and SP-C genes by binding to cis-active elements located in the 5Ј regulatory regions of each of the genes. 2 The present findings that GATA-6 activates TTF-1, taken together with observations that GATA-6 directly influences expression of the surfactant protein genes by binding to their promoters support a potential regulatory role for GATA-6 in surfactant protein expression.
The finding that GATA-6, but not GATA-5, activated TTF-1 gene transcription is consistent with the similar distribution of GATA-6 and TTF-1 in subsets of developing respiratory epithelial cells in the mouse lung in vivo (5,26,27). GATA-5 mRNA was detected in mesenchymal but not epithelial cells of the developing lung. The finding that GATA-6, but not GATA-5, activated TTF-1 gene expression was somewhat surprising, given that both proteins bind to similar consensus elements in target genes. Of interest, the selectivity observed for GATA-6 in activation of the TTF-1 gene was not observed in the studies FIG. 5. A, EMSA of the GATA-6 binding site. EMSA analysis was performed using an oligonucleotide located Ϫ110 to Ϫ87 in the mTTF-1 gene using nuclear extracts from MLE-15 (lanes 1-3), HeLa (lanes 4 -6), 3T3 (lanes 7-9), H441 (lanes 10 -12), and MLE-F6 (lane 13) cells. The arrow indicates the GATA-6-dependent complex. Lanes with no competitor are labeled TE. The complex was inhibited by a 100-fold excess of the same unlabeled oligonucleotide and labeled self. The same oligonucleotide bearing a mutant GATA site (TTACTC 3 TGAATT), changed at base pairs Ϫ100 to Ϫ97, did not compete for binding. These lanes are labeled mut. B, supershift analysis of the mTTF-1 GATA-6 binding site. EMSA analysis was performed using mTTF-1 gene sequences from Ϫ110 to Ϫ87 and MLE-15 cell nuclear extracts. Lane 1 represents a GATA-6-specific complex formed by the wild type oligonucleotide identified by the arrow. This complex was supershifted by GATA-6 antiserum (lane 2), but not by preimmune serum (lane 3). Sera were diluted 1:100. The formation of a GATA-6specific protein complex was markedly inhibited by introduction of a mutation into the GATA binding site (lanes 4 -6).
with the SP-C promoter, wherein both GATA-5 and GATA-6 were active in enhancing SP-C gene transcription. 2 Thus, the effects of GATA-6 on TTF-1 gene transcription are probably determined by mechanisms distinct from that involved in activation of surfactant protein C gene transcription by GATA-6.
GATA-6 bound to a site in a consensus element located near the start of transcription of the TTF-1 gene. The cis-active element and flanking regions are highly conserved in the TTF-1 gene sequences available for comparison (human, mouse, and rat), supporting the potential importance of this region in gene regulation (14,15,18). The DNA binding site identified in the murine TTF-1 gene is consistent with binding sites for many other GATA family members. Therefore, the selectivity of the response to GATA-6 is likely to be dependent upon the sites of expression of GATA-6 as distinct from other known GATA family members. Alternatively, interactions between GATA-6 and other transcription proteins that are expressed in a cell-selective manner are required. Other transcription factors are known to form complexes with GATA factors to activate target gene transcription. These include FOG (friend of GATA) and GATA-1 (41); Lmo2 and GATA-1 (42); and EKLF (erythroid kruppel-like factor), Sp1, and GATA-1 (43). GATA-6 is also expressed in the gastrointestinal and genitourinary tract, as well as in the heart. Known GATA-6 target genes include histidine decarboxylase, lactase, vitellogenin II, and cardiac m2 muscarinic acetylcholine gene (33,40,44,45). While GATA-6 activated TTF-1 transcription in HeLa and H441 cells, GATA-6 did not activate the promoter in MLE-15 cells, perhaps related to the high concentrations of GATA-6 present in this cell line or to unique combinations of transcription factors present in the MLE cell line.
The stimulatory effects of GATA-6 on the TTF-1 promoter in HeLa cells did not require TTF-1. The finding that the coexpression of TTF-1 and GATA-6 did not further enhance TTF-1 transcription in HeLa demonstrated that GATA-6 activ-ity was not dependent upon interactions with this region of the TTF-1 gene and TTF-1 itself. This finding is distinct from recent studies in which Nkx2.5 (a TTF-1 family member) and GATA-4 enhanced myosin heavy chain gene expression in the heart (28). Co-transfection with TTF-1 modestly inhibited transcriptional activity of the mTTF-1 construct in the presence of GATA-6 in HeLa cells; however, the significance of this observation is unclear. Activation of the TTF-1 promoter by GATA-6 was not influenced by co-transfection with HNF-3␤, which is known to bind to two sites in the region Ϫ135 to Ϫ3 in the human TTF-1 gene. Thus, GATA-6 binds and activates the cis-acting element in the TTF-1 gene independently of either TTF-1 or HNF-3␤.
The finding that GATA-6 activates the TTF-1 promoter adds to our understanding of the potential interactions among various families of transcription factors in the developing respiratory epithelium. Present data support a model in which the transcription factor(s) bind to and activate target genes, including the surfactant proteins and CCSP, but may also interact at the level of regulation of the transcription factors themselves, HNF-3␤ and GATA-6 modulating TTF-1 gene expression. The co-expression of HNF-3␤, GATA-6, and TTF-1 in the developing lung is consistent with the concept that respiratory epithelial differentiation and gene expression may be modulated by interactions among these distinct families of transcription factors (9).