Transcription of the Interleukin-1 Receptor-related T1 Gene Is Initiated at Different Promoters in Mast Cells and Fibroblasts*

The delayed early serum response gene T1 encodes glycoproteins of the immunoglobulin superfamily with significant sequence similarity to the type 1 interleu- kin-1 receptor. The T1 gene is transcribed in fibroblasts into an abundant 2.7-kilobase (kb) and a rare 5-kb mRNA in response to proliferation-inducing stimuli. It gives predominantly rise to the longer transcript in the bone marrow of adult mice and in cultured mast cells. Alternative 3 (cid:42) processing is responsible for the two mRNA forms. The short transcript encodes a secreted protein with marked similarity to the extracellular domain of the interleukin-1 receptor, whereas the long mRNA is translated into a protein with an additional putative transmembrane and an intracellular domain. Here we demonstrate that T1 transcription in mast cells and fibroblasts initiates at two different start sites which are 10.5 kb apart. The alternative first exons are both spliced to exon 2 which contains the translation start site. Northern blot analysis and primer extension experiments revealed that promoter usage is strictly cell type-specific. T1 transcription in mast cells is initiated exclusively at the distal promoter, whereas in fi- broblasts both the short and the long T1 mRNA start at the proximal promoter. Two GATA-1 elements were identified in the 5 (cid:42) -flanking region of the mast cell-spe-cific distal exon 1. The T1 gene, also designated ST2 system (Promega) to enrich for mRNA. Five or ten micrograms of total RNA and 2.5 (cid:109) g of poly(A) mRNA were denatured by glyoxylation, separated on 1% agarose gels (25), transferred onto nylon membranes (GeneScreen Plus; DuPont-NEN) and UV cross-linked (UV stratalinker 1800; Stratagene). Hybridization was performed either with DNA frag- ments labeled with [ (cid:97) - 32 P]dCTP (3000 Ci/mmol; Amersham Corp.) according to Feinberg and Vogelstein (26, 27) in 50% formamide, 2 (cid:51) SSC (0.3 M NaCl, 34.2 m M sodium citrate, pH 7.0), 5 (cid:51) Denhardt’s, 1% SDS, 10 m M Tris-HCl (pH 7.5), and 100 (cid:109) g/ml denatured carrier DNA or with digoxigenin (DIG)-UTP-labeled RNA fragments (26, 28) in 50% form-amide, 5 (cid:51) SSC, 0.1% N -lauroylsarcosine, 0.02% SDS, 20 m M maleic acid, pH 7.5, 30 m M NaCl, and 2% blocking reagent. DIG-labeled RNA was produced using the DIG RNA labeling kit (Boehringer Mannheim) and the plasmid pMV7TORF (4), linearized with Sal I, as the template. The resulting 1-kb RNA spans the whole small T1 open reading frame. DIG-labeled RNA was detected using an anti-DIG antibody coupled to alkaline phosphatase followed by the enzymatic reaction with the sub- strate Lumigen™ PPD (DIG luminescent detection kit; Boehringer Mannheim). The radioactively labeled probes were the following: a 1.15-kb Hin dIII/ Bam HI fragment from the 17 spanning exon Bam HI/ Hin dIII fragment plasmid, proximal exon

The delayed early serum response gene T1 encodes glycoproteins of the immunoglobulin superfamily with significant sequence similarity to the type 1 interleukin-1 receptor. The T1 gene is transcribed in fibroblasts into an abundant 2.7-kilobase (kb) and a rare 5-kb mRNA in response to proliferation-inducing stimuli. It gives predominantly rise to the longer transcript in the bone marrow of adult mice and in cultured mast cells. Alternative 3 processing is responsible for the two mRNA forms. The short transcript encodes a secreted protein with marked similarity to the extracellular domain of the interleukin-1 receptor, whereas the long mRNA is translated into a protein with an additional putative transmembrane and an intracellular domain.
Here we demonstrate that T1 transcription in mast cells and fibroblasts initiates at two different start sites which are 10.5 kb apart. The alternative first exons are both spliced to exon 2 which contains the translation start site. Northern blot analysis and primer extension experiments revealed that promoter usage is strictly cell type-specific. T1 transcription in mast cells is initiated exclusively at the distal promoter, whereas in fibroblasts both the short and the long T1 mRNA start at the proximal promoter. Two GATA-1 elements were identified in the 5-flanking region of the mast cell-specific distal exon 1.
The T1 gene, also designated ST2 or DER4, was independently isolated as a Ha-ras oncogene-responsive gene (1) and as a gene inducible by growth factors in murine fibroblasts (2,3). Growth factor-mediated T1 gene induction requires ongoing protein synthesis, thus defining T1 as a delayed early serum responsive gene (3,4). Upon stimulation of fibroblasts with proliferation-inducing agents such as serum, platelet-derived growth factor, fibroblast growth factor, 12-O-tetradecanoylphorbol-13-acetate, lysophosphatidic acid, or upon oncogene expression T1 is transcribed into an abundant 2.7-kb 1 and a rare 5-kb mRNA (4). 2 The shorter transcript encodes a primary translation product of 337 amino acids with an NH 2 -terminal leader peptide sequence (6). The mature small T1 protein is a secreted, heavily N-glycosylated protein (7,8) of the immunoglobulin superfamily with marked sequence similarity to the carcinoembryonic antigen (6) and the extracellular portion of the interleukin 1 (IL-1) receptors (2). The rare 5-kb T1 mRNA arises by alternative 3Ј processing of the primary transcript and encodes a 567-amino acid protein. The corresponding cDNA (designated ST2L) was cloned from serum stimulated BALB/c-3T3 cells (9). The sequences of the two T1 proteins are identical at the amino terminus and diverge in nine amino acids in front of the COOH-terminal end of the small protein.
The extension of the large T1 protein consists of a putative transmembrane and an intracellular domain with significant homology to the type 1 IL-1 receptor (9).
The in vivo expression patterns of the two T1 transcripts differ drastically. The short mRNA has been found in experimental Ha-ras-induced murine mammary tumors, in the developing mammary gland 3-4 weeks after birth (10) as well as in embryonic skin, retina, and bone (11). Expression of the long transcript is restricted to distinct cells of the hematopoietic organs (embryonic liver, spleen, bone marrow) (11) and to the lung (12) throughout ontogenesis. Despite their similarity to the IL-1 receptor, the T1 proteins neither bind IL-1␣ nor IL-1␤ (11), suggesting that the large protein is a novel orphan receptor predominantly expressed in a subset of hematopoietic cells.
Growth factor-induced T1 gene expression in fibroblasts is mediated by AP-1 (4, 13), a homodimeric or heterodimeric protein complex formed by transcription factors of the Jun and Fos families (14). A 12-O-tetradecanoylphorbol-13-acetate-responsive element located within the T1 enhancer at position Ϫ3.6 kb is essential for gene induction (13). Furthermore, expression of the immediate early transcription factors c-Fos and FosB resulted in the accumulation of T1 mRNA (4). Likewise, the rat T1 homologue fit-1 was identified as a c-Fos-responsive gene in fibroblasts (15).
Sequence analysis of cDNA clones obtained from T1 mRNA in hematopoietic cells and fibroblasts revealed different transcription initiation sites in these two cell types. Here we demonstrate that the T1 gene is transcribed from two promoters which are 10.5 kb apart. Promoter usage is strictly tissuespecific, the distal and proximal promoters being used exclusively in hematopoietic cells and fibroblasts, respectively.

EXPERIMENTAL PROCEDURES
Cell Culture-NFR-2 is a cell clone of NIH 3T3 cells transfected with the Ha-ras (EJ) oncogene (16). NIH 3T3 and NFR-2 cells were grown in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) supplemented with 10% fetal calf serum (FCS), penicillin (100 units/ml), and streptomycin (100 g/ml). To arrest cell growth, the cells were incubated for 24 h in Dulbecco's modified Eagle's medium containing 0.5% FCS. This medium was replaced with fresh Dulbecco's modified Eagle's medium containing 10% FCS to stimulate entry into the cell cycle.
* This work was supported by Swiss National Science Foundation Grant 31.32182.91 (to R. K.) and by the Swiss Cancer League. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EMBL Data Bank with accession number(s) U38789.
Sequencing and Primer Extension-The 0.6-kb EcoRI fragment of plasmid pETH was subcloned into pBluescript KS ϩ (Stratagene), and the nucleotide sequence was determined by the dideoxynucleoside triphosphate-chain termination reaction (29).
For primer extension analysis, 1-2 ϫ 10 5 cpm of end-labeled oligonucleotides TG1p (GGGCTGGAGAGAATAAGCTCAGAGCCGTGAG-GGC) or TG2 (GCTCTCTGAGGTAGGGTCCAGAAGAGAAATCAC) (purified on 7 M urea, 13% polyacrylamide gels (30) prior to labeling) were mixed with 5-30 g of total RNA, ethanol-precipitated, and resuspended in 30 l of 50% formamide, 400 mM NaCl, 10 mM PIPES (pH 6.4), and 1 mM EDTA. Annealing was performed by incubation at 80°C for 3 min followed by slow cooling to room temperature. The nucleic acids were precipitated with isopropanol and primer extension reactions were carried out in 20 l of 50 mM Tris-HCl (pH 8.3), 75 mM KCl, 3 mM MgCl 2 , 5 mM dithiothreitol, 1 g of actinomycin D, 125 M of each deoxynucleoside triphosphate, 20 units of RNase Inhibitor (Boehringer Mannheim), and 300 units of RNase H Ϫ Moloney murine leukemia virus reverse transcriptase (SuperScript™; Life Technologies, Inc.) at 39°C for 1 h. For dideoxy chain termination primer extension reactions (29), the conditions used were identical to the above with the following changes: the concentration of each of the three driver dNTPs (not in competition with the ddNTP) was 10 M, the concentration of the dNTP which was competed for by the corresponding ddNTP was 6 M, ddATP was 25 M, ddCTP was 20 M, ddGTP was 10 M, and ddTTP was 30 M. The reactions were stopped by the addition of EDTA to a final concentration of 25 mM, and the RNA was degraded with 5 g of RNase A at 37°C for 30 min. After phenol/chloroform extraction of the mixture and subsequent ethanol precipitation the primer extension products were analyzed on 7 M urea, 9% polyacrylamide sequencing gels (30).

Localization of Mast Cell-specific T1 Gene Sequences-
The sequences of cDNA clones derived from mast cells and fibroblasts diverge upstream of the 5Ј end of exon 2 (data not shown). The whole genomic T1 DNA has been sequenced up to the SacI site at position Ϫ3.987 (13,31). No sequence corresponding to the 5Ј end of the mast cell-specific T1 cDNA was found within this region. Therefore the T1 transcription start site in mast cells must be located further upstream.
In order to identify this start site, we first determined the genomic organization upstream of the end point of the published sequence (Fig. 1A). DNA of plasmid pETH (Fig. 1A) as well as genomic DNA of NIH 3T3 cells and of primary mast cells were subjected to Southern blot analysis to verify that the organization of the recombinant DNA corresponds to the one of the genomic DNA. The restriction fragment patterns of the different DNAs, obtained by hybridization with the radioactively labeled 0.85-kb SacI/SphI fragment (Fig. 1B), were identical, indicating that no DNA rearrangement had occurred during cloning. This result was confirmed using the 0.6-kb EcoRI fragment as a probe (data not shown).
Next we localized the mast cell-specific sequence by Southern blot hybridization utilizing as a probe the oligonucleotide TG1d which represents part of the extra sequence present at the 5Ј end of mast cell but not of fibroblast T1 mRNA. The hybridization pattern indicated that the novel sequence is limited to the 0.6-kb EcoRI fragment of the plasmid pETH located approximately 10.5 kb upstream of the previously described transcription start site ( Fig. 2A). Sequence analysis of this restriction fragment revealed that the entire extra 5Ј sequence of the mast cell-specific T1 transcript is colinear with genomic DNA, thus representing an alternative first exon within this 0.6-kb EcoRI fragment (Fig. 2B). In order to accurately map the transcription initiation site, we performed primer extension analysis (Fig. 2C). A major and a minor extension product were obtained. Reactions in the presence of dideoxynucleoside triphosphates allowed the exact positioning of the transcription start site. The 5Ј ends determined by primer extension and by cDNA cloning coincide precisely.
Cell Type-specific Expression of the Two Alternative First Exons-The T1 gene is transcribed into a major 2.7-kb and a minor 5-kb mRNA in fibroblasts and almost exclusively into the longer 5-kb mRNA in mast cells. Initiation of transcription at the distal and proximal exon 1 could either be cell typespecific or specific for the long and the short transcript as has been suggested for the fit-1 gene, the rat homologue of T1 (12).
Northern blot analyses of RNA extracted from primary and established mast cells, fibroblasts, and adherent primary bone marrow cells were performed to address this question (Fig. 3). The probes used for hybridization recognized either the T1 open reading frame, the exon 2, or the proximal or distal exon 1. As expected, serum-stimulated fibroblasts mainly expressed the short 2.7-kb mRNA, whereas the 5-kb transcript accumulated predominantly in mast cells and, to a much lower extent, in primary adherent bone marrow cells.
The T1 gene has previously been identified as a delayed early serum response gene based on the observation that ongoing protein synthesis is a prerequisite for its stimulation. However, in certain cell lines such as NIH 3T3 fibroblasts, T1 induction is possible in the presence of the protein synthesis inhibitors cycloheximide and anisomycin (32). 3 We even observed that anisomycin on its own led to a very substantial accumulation of T1 mRNA (data not shown). We found particularly high levels of the long 5-kb T1 transcript in the Ha-ras (EJ)-transformed NIH 3T3 cell line NFR-2 (33). Therefore, analysis of T1 transcription initiation in this cell line allowed us to distinguish between transcript-specific versus tissue-specific promoter usage. The Northern blot analyses shown in Fig. 3 demonstrate that all T1 transcripts contain sequences downstream of the exon 1/exon 2 boundary (top two panels), whereas hybridization to the distal and proximal exon 1-specific probes occurred only with RNA derived from mast cells and fibroblasts, respectively, independent of the transcript size. Hybridization with the probe which recognizes the proximal exon 1 is rather inefficient probably due to the small size of this exon (Fig. 3, third panel). We therefore analyzed poly(A)-enriched RNA to detect the 5-kb T1 transcript in fibroblasts (Fig. 3, right panel).
Primer extension experiments were performed to confirm these results at the level of transcription initiation. The primers used in these experiments as well as the expected elongation products are depicted in Fig. 4A. The oligonucleotide TG2, which is complementary to the 5Ј region of exon 2, was extended on RNA derived from NFR-2 cells (Fig. 4B, lane 1), NIH  (Fig. 1A) were digested with the restriction endonucleases indicated above the lines and subjected to Southern blot hybridization with the oligonucleotide TG1d. This probe is complementary to the 3Ј part of the distal exon 1 (Fig. 4A). The lengths of the hybridizing restriction fragments are indicated. B, BamHI; E, EcoRI; H, HindIII; S, SacI. B, the nucleotide sequence of the 0.6-kb EcoRI fragment at position Ϫ10.5 kb. The highlighted 57 nucleotides represent the extra sequence identified at the 5Ј end of cDNAs derived from mast cells but not fibroblasts. The two putative TATA boxes for the major and minor transcription start sites are underlined, and putative binding sites for the GATA transcription factors and SP1 are in brackets and double underlined, respectively. C, mapping of the 5Ј end by primer extension. The primer TG2 (Fig. 4A)  3T3 cells (Fig. 4B, lane 2), and mast cells (Fig. 4B, lane 3). The extension product indicative for transcriptional initiation at the proximal promoter was obtained with RNA from the two fibroblast cell lines. In contrast, primer extension on mast cell-derived RNA gave rise to the two extension products characteristic for initiation at the distal promoter (Fig. 4B). Thus, transcription initiation seems to be strictly cell type-specific. This implies that synthesis of the 5-kb T1 mRNA starts at the distal (Fig. 4B, lane 3) as well as the proximal (Fig. 4B, lane 1) promoter. However, one could argue that the long T1 transcript in fibroblasts is initiated at the distal promoter but that the corresponding extension product could not be detected due to  Fig. 4B. The amount of loaded RNA from fibroblasts (lanes 1, 2, and 4) and mast cells (lanes 3 and 4)  were used as templates. The position for the 75-bp extension product, which is characteristic for transcription start at the proximal promoter, is indicated. its low abundance. To disprove this argument, we performed a primer extension experiment with a mixture of RNA derived from serum-stimulated NIH 3T3 fibroblasts and from mast cells. Northern blot analysis demonstrated that the ratio of the 2.7-kb to the 5-kb T1 transcript in this mixture is approximately the same as in anisomycin-treated NFR-2 cells (Fig. 4C,  compare lanes 1 and 4). We obtained extension products indicative of transcription initiation at the proximal and distal promoter (Fig. 4B, lane 4). Even a mixture with a 4-fold reduced amount of mast cell RNA gave rise to easily detectable primer extension products derived from the distal exon 1 (data not shown). Hence, if the long T1 mRNA in NFR-2 cells were initiated at the distal promoter, we should have detected it by primer extension (Fig. 4B, lane 1). Therefore, we conclude that synthesis of the large majority of both T1 transcripts in NFR-2 fibroblasts is started at the proximal exon 1 and that transcript size is independent of the promoter used.
Furthermore, primer extension experiments with the oligonucleotide TG1p, which anneals to the proximal exon 1, yielded a product with RNA from fibroblasts but not from mast cells (Fig. 4D). This confirms that T1 gene transcription in mast cells is exclusively initiated at the distal promoter.

DISCUSSION
In this report we have demonstrated that the synthesis of T1 mRNA starts at different promoters in fibroblasts and in mast cells. The two alternative first exons are 10.5 kb apart, and both are spliced to exon 2 where translation starts. Alternative 3Ј processing gives rise to a short 2.7-kb and a long 5-kb mRNA that encode a secreted and a putative trans-membrane protein, respectively. In fibroblasts the short transcript predominates, but small amounts of the long mRNA are always observed. Under certain conditions, such as treatment with anisomycin, substantial amounts of the long T1 mRNA are synthesized in fibroblasts, particularly in ras-transformed NIH 3T3 cells. This finding allowed us to demonstrate that all T1 transcripts are initiated at the proximal promoter in fibroblasts, irrespective of their processing mode (Fig. 4, B and C). This is in line with the observation that the two T1 transcripts accumulate in parallel under all inducing conditions in fibroblasts and thus are apparently controlled by the same transcriptional regulatory elements. The large T1 mRNA predominates in mast cells, but substantial amounts of the 2.7-kb T1 transcript could be observed under certain conditions such as treatment with Ca 2ϩ ionophores (data not shown). The synthesis of both T1 mRNAs is initiated at the distal promoter in mast cells.
Expression of fit-1, the rat homologous gene of T1, is initiated at two tissue-specific promoters as well, the transcription start sites being 14 kb apart (12). The model has been put forward that transcription initiation is tightly coupled to 3Ј processing. Accordingly, whenever transcription starts at the distal promoter, the long mRNA is produced, whereas initiation at the proximal promoter instructs the transcription machinery to terminate in exon 8, giving rise to the short transcript. However, this model is rendered unlikely by our finding that substantial amounts of the long T1 mRNA are observed in fibroblasts under some conditions and that these transcripts are exclusively initiated at the proximal promoter. Rather, we suggest that transcription initiation and 3Ј processing (poly(A) selection) are two independent tissue-specific events.
Sequence analysis of the genomic DNA around the mast cell-specific first exon revealed the presence of two GATA elements and one putative SP1 binding site (Fig. 2B). The GATA consensus motif (T/A)GATA(A/G) is the recognition site of the GATA transcription factor family of which six members have so far been described. The patterns of both sequence preferences for DNA binding and expression in tissues as well as cell lines are overlapping, but distinct for these zinc finger proteins. Thus, the GATA transcription factors might exert differential gene regulation by distinct tissue distribution, selective binding to DNA target sequences or by different interactions with other nuclear proteins. Among the GATA factors, expression of GATA-1 (34,35), GATA-2 and GATA-3 (35) was detected in several mouse and rat mast cell lines but not in mouse 3T3 fibroblasts. These three GATA proteins are reported to regulate the promoter of the carboxypeptidase A gene in mast cells (35). Hence, we consider the possibility that GATA factors direct T1 transcription in mast cells to the distal promoter. It is also striking that two of the four important GATA elements in the murine ␤-globin promoter (36) are at almost identical relative positions to those in the T1 gene (positions Ϫ210 and Ϫ75 in the ␤-globin and positions Ϫ210 and Ϫ76 in the T1 promoter). A further indication that the GATA elements might be important for mast cell-specific T1 transcription is the finding of a putative SP1 site adjacent to them. Cooperation of SP1 and GATA transcription factors has been observed for several genes (37)(38)(39). Promoter studies have been initiated to test whether these GATA elements and the SP1 binding site are indeed instrumental for mast cell-specific T1 gene expression and whether they work in concert with the T1 enhancer. This AP-1 binding regulatory element is centrally located between the two alternative first exons at position Ϫ3.6 kb, and it is essential for the induction of the proximal T1 promoter (4,13). Moreover, Gong et al. have demonstrated that the transcription of the ⑀-globin promoter requires both promoter bound GATA-1 and enhancer bound AP-1/NF-E2, whereas GATA-1 acts solely as a mediator of the enhancer effect on transcription (5).
In summary, the T1 gene is transcribed from two promoters which are 10.5 kb apart. Their usage is strictly cell typespecific, as initiation at the distal and the proximal promoter exclusively occurs in mast cells and fibroblasts, respectively. Further characterization of the distal promoter containing the two GATA elements and one SP1 binding site will help to identify the transcription factors that orchestrate tissue-specific 5Ј initiation. Differential expression of the putative T1 receptor protein and its soluble form is achieved by two independent tissue-specific events: transcription initiation at two promoters and alternative 3Ј processing. T1 is therefore a particularly useful tool for gene regulation studies to reveal the mechanisms responsible for the tissue-specific expression of a transmembrane versus a secreted form of a protein.