Characterization of the Promoter of Human Leukocyte-specific Transcript 1

The gene for the human leukocyte-specific transcript 1 (LST1) encodes a small protein that modulates immune responses and cellular morphogenesis. The LST1 transcripts are expressed at high levels in dendritic cells. Because of the complex splicing pattern, use of alternative 5′-untranslated exons, and a biologically interesting pattern of expression of LST1 mRNA, we studied the human LST1 gene promoter and regulatory elements. We identified an additional upstream 5′-untranslated exon in U937 monocytic cells. Transient transfection studies demonstrated that the combination of regions from −1363 to −621 with −112 to −54, relative to the translation start codon, produced the highest level of transcripts from among the various constructs tested, but the pattern of transcripts produced was only a subset of those produced from the endogenous gene. DNase I footprinting analysis and electrophoretic mobility shift assays showed that oligonucleotide probes corresponding to three regions, −1171 to −1142 (BI), −1136 to −1111 (BII), and −783 to −751 (BIV), bound proteins in U937 nuclear extracts. Competition and supershift electrophoretic mobility shift assay did not identify any known transcription factors responsible for BII probe binding. These studies suggest that a novel DNA-binding site and interaction of multiple regulatory elements may be involved in mediating the expression of the various forms of LST1 mRNA.

The leukocyte-specific transcript 1 (LST1) 1 gene, the human homologue of the mouse B144 transcript, is located within 15 kilobases upstream of the tumor necrosis factor cluster in the major histocompatibility complex class IV region, a region that contains a number of genes that may play a role in various aspects of stress, inflammation, and immune responses (1)(2)(3)(4)(5)(6)(7)(8).
The human LST1 mRNA is most actively transcribed in monocytes and dendritic cells, and its mRNA levels can be enhanced by interferon (IFN)-␥, suggesting a role of LST1 in the immune response (9 -12). A recent study showed that this gene product had an inhibitory effect on lymphocyte proliferation (13). Transient transfection studies in our laboratory showed that the LST1 gene product induced extensive thin cytoplasmic extensions in a wide variety of cells in vitro and, in particular, was strongly expressed in dendritic cells in vivo and in vitro. 2 Due to alternative splicing, the LST1 gene in human and mouse encodes multiple transcripts, each 800 nucleotides or less in length. Previous studies on the human LST1 gene showed that this gene consisted of at least eight exons and four introns, including four different alternative 5Ј-UT exons (termed 1A, 1B, 1C, and 1D, respectively) and four additional exons spanning 2.7 kilobases. The extensive alternative splicing of LST1 mRNA results in high structural diversity of the putative encoded polypeptides (9). The DNA sequences between these 5Ј-UT exons do not consistently show sequence features resembling promoters.
There are certain unusual features of the relationship between the human LST1 mRNA and the gene sequence. For example, one alternative intron of the LST1 gene begins with the dinucleotide GA rather than the common GT or the occasional GC. Interestingly, each of the four published 5Ј-UT exons is between 100 and 150 nucleotides in length and lies closely downstream of the sequence CCCAG. Furthermore, the genomic sequence in the vicinity of the 5Ј-UT exons contains at least two other stretches of sequences about the same length as the published 5Ј-UT exons in which the nucleotide stretches begin with a CCCAG and end with an apparently conventional 5Ј-end splice site sequence. Therefore, it is possible that one or more additional 5Ј-UT exons may exist in the 5Ј-flanking region of the human LST1 gene.
We carried out the present study to address the following questions. (a) Does the human LST1 gene have an additional 5Ј-UT exon, which serves as a common initial exon, or does transcription apparently begin separately upstream of each of the closely expressed alternative 5Ј-UT exons? (b) Where is the promoter in the 5Ј-flanking region of the human LST1 gene? (c) Which cis-acting elements are involved in the human LST1 gene expression? (d) Can we separate elements responsible for production of subsets of transcription initiation sites and splicing events in the LST1 gene? We have identified an additional 5Ј-UT exon upstream of the published exon 1A in U937 cells, a human monocytic cell line, indicating that there are at least five alternative 5Ј-UT first exons (termed 1a, 1b, 1c, 1d, and 1e) that are spliced directly to exon 2 containing the ATG translation start codon. The exon 1b-exon 2 splicing pattern was the most abundant transcript of the human LST1 gene in U937 cells and the major transcription initiation site was located upstream of exon 1b. We have also characterized the human LST1 promoter region and regulatory elements. These studies suggest that a novel DNA-binding site and interaction of multiple regulatory elements may be involved in the regulation of the human LST1 gene expression.

EXPERIMENTAL PROCEDURES
Rapid Amplification of cDNA Ends (5Ј-RACE)-The total RNA of U937 cells (6 ϫ 10 6 ) was isolated by using TRIZOL reagent (Life Technologies, Inc.), and digested with 2 l (10 units/l) of RNase-free DNase I (Roche Molecular Biochemicals) at 37°C for 30 min. 2 g of the total RNA from U937 cells were reverse transcribed into first-strand cDNA using ThermoScript II Reverse Transcriptase (Life Technologies, Inc.) and the reverse primer GSP-RTN (see Table I) designed from the exon 2 region of the human LST1 cDNA. After degrading the original RNA template with RNase H, the first-strand cDNA was extended from its 3Ј-end using terminal transferase (Roche Molecular Biochemicals) to form a poly(A) tail. The poly(A)-tailed cDNA was then used as a template in 30 cycles of polymerase chain reaction (PCR) at 94°C for 1 min, 55°C for 1 min, and 72°C for 1 min with the antisense primer GSP3 and the sense primer QT and QO (14). The products of PCR amplifications were subjected to a second PCR amplification with the primer GSP42N and the primer QI for another 30 cycles. The PCR products of the second amplification were separated by gel electrophoresis in a 1.2% agarose gel. Purified PCR fragments were cloned into the TA vector (Invitrogen) and sequenced. All enzymes were purchased from New England BioLabs unless otherwise indicated.
Promoter-Reporter Plasmid Constructs-The different genomic fragments of the 5Ј-flanking region of the human LST1 gene were amplified by PCR using primers designed with SacI/NcoI restriction sites, and then inserted into the SacI-NcoI sites upstream of the luciferase gene in the pGL3-Basic vector (Promega) to generate various LST1 reporter constructs. The pGL2423Ap construct (Ϫ1363 to Ϫ621 plus Ϫ112 to Ϫ54) was generated by digesting the pGL2423 construct (Ϫ1363 to Ϫ54) with ApaI to delete the fragment spanning from Ϫ621 to Ϫ112, relative to the translation start codon. The pGL24SA construct (Ϫ1363 to Ϫ954 plus Ϫ112 to Ϫ54) was generated by digesting the pGL2423 CCCAGTTTCCAGACCCTAGTCAGTATATCTGGCTCT a Position numbers are relative to translation start codon of the LST1 gene. b Underlined sense sequences are SacI sites, underlined antisense sequences are NacI sites. c QT, QI, and QO primers are specially used for 5Ј-RACE. d Lowcase letters indicate the mutant nucleotides in oligonucleotide probes. e Luc1 and Luc2 primers designed are based on the Luciferase gene in pGL3-basic report vector are specially used for RT-PCR of transfected U937 cells. Luc2 is 7 bp downstream of GSP42N in the pGL2423 construct.
construct with ApaI-StyI to delete the fragment (Ϫ954 to Ϫ112) in the 5Ј-flanking region. The pGL24SX34 construct was generated by replacing the luciferase gene in the pGL2423 construct with the entire downstream region (Ϫ51 to ϩ1368) that included the middle of exon 2 and the end of exon 5, and putting an insert of 87 nucleotides amplified from the luciferase gene as a marker in a position Ϫ54 bp upstream of the translation start ATG codon. The mutant pGL2423mAp construct was generated by overlap extension using PCR (15,16). Two partially overlapping fragments were amplified using either USP2S and GSPmII or USPII and GSP42N as primers (see Table I), and these fragments were combined and used as template DNA in the second PCR reaction with primers USP2S and GSP42N to generate an 802-bp PCR product. The final PCR product was digested with NcoI-SacI and subcloned into the pGL3-Basic vector. The orientation and sequence of the inserts were verified by sequencing through the insert-vector junctions. The plasmid DNA was purified according to the Plasmid Maxi Kit (QIAGEN) manual.
Transient Transfection and Luciferase Assay-The U937 cell line was purchased from the American Type Culture Collection (Rockville, MD). Media and reagents for cell culture were purchased from Life Technologies, Inc. unless otherwise indicated. The U937 cell line was maintained in RPMI 1640 medium containing L-glutamine, 10% fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin at 37°C under 5% CO 2 . Cells in each well of 24-well plates were transfected with 1.5 g of promoter-reporter constructs and 4 l of the FuGene 6 transfection reagent (Roche Molecular Biochemicals) according to the manufacturer's instructions. After 48 h of transfection, the cells were washed, lysed, and centrifuged. The luciferase activity in the nuclear extracts was measured with a commercial luciferase assay system (Promega) according to the manufacturer's instructions and luminescence was determined with an LB9510 Lumat illuminometer (Berthold, Germeny). All cell lysate concentrations were normalized to recovered protein concentrations that were measured by the Bio-Rad protein assay based on the method of Bradford (17). The luciferase activities of the constructs are shown as a value relative to that of the pGL3-Control vector that contains the SV40 promoter (Promega), which was arbitrarily set to 100%. Each transfection was performed in duplicate and repeated three times or more. The plasmid pCMV-␤-gal (CLONTECH) was used as a control for transfection efficiency. In addition, different cell lines, such as K562, Jurkat, 293T, and HeLa (from our laboratory storage) were also used in the transfection experiments with certain reporter constructs. For the IFN-␥ stimulation experiment, U937 cells transfected with the pGL2423Ap construct were treated with 50, 100, 200, or 500 units/ml IFN-␥ (Life Technologies, Inc.) for 6, 24, and 48 h prior to performing the luciferase assay. S1 Nuclease Protection Assay-1 g of the 24SA probe (405 bp) recovered by digesting the pGL24SA construct (Ϫ1363 to Ϫ1001 plus Ϫ112 to Ϫ54) with SacI-NcoI was labeled by filling in the 5Ј-end overhangs with [␣-32 P]dCTP and Klenow polymerase. The unincorporated [␣-32 P]dCTP was removed by using Sephadex G-50 columns (Roche Molecular Biochemicals). Hybridization of 1 ng of the labeled 24SA probe with 5 g of the total RNA from U937 cells was performed in 10 l of hybridization buffer (80% deionized formamide, 0.4 M NaCl, 1 mM EDTA, 50 mM Pipes, pH 6.4) and incubated at 42°C overnight. 300 units of S1 nuclease were added to digest the unbound single strand DNA at 37°C for 30 min. The digested DNA was recovered by ethanol precipitation and analyzed on a denaturing 6% polyacrylamide gel containing 8 M urea. The gels were dried and exposed to x-ray film. The 24SA probe was sequenced according to the manual of the T7 sequencing kit (Amersham Pharmacia Biotech).
RT-PCR-After U937 cells were transfected with promoter report constructs (e.g. pGL2423, pGL54212, pGL4214, pGL24SX34, and pGL3C8S) for 48 h, the total RNA of U937 cells (6 ϫ 10 6 ) was isolated by using TRIZOL reagent (Life Technologies, Inc.) and digested with 2 l (10 units/l) of RNase free DNase I (Roche Molecular Biochemicals) at 37°C for 30 min to remove traces of contaminating DNA. 2 g of total RNA from these cells were reverse transcribed into first strand cDNA with ThermoScript II Reverse Transcriptase (Life Technologies, Inc., Life Technologies, Inc.). 2 l of the first strand cDNA were used in 50 l of PCR with the sense primers for different forms of the 5Ј-UT exon 1 (such as USP2 for exon 1a, USP1S for exon 1b, USP7 for exon 1c, USP8 for exon 1d, and USP10 for exon 1e) and the antisense primers (Luc1 for vector or GSP-RTN for exon 2). The second round of PCR reaction was performed with different internal primers (USP2S, USP1, USP7, USP81, and USP10S) and the primer (Luc2) from vector and the primer (GSP42N) from exon 2. The DNase-treated RNA without reverse transcriptase was used as a control for the PCR reactions. The PCR reaction was performed for 30 cycles at 94°C for 1 min, 55°C for 1 min, 72°C for 1 min. PCR products were electrophoresed in a 2.0% agarose gel and subcloned into the TA vector (Invitrogen) for sequencing.
Nuclear Extract Preparation-The nuclear extracts were prepared according to the method of Lee et al. (18). Briefly, after U937 cells were washed with a phosphate buffer, the pellets were resuspended in one packed cell volume of buffer A (10 mM Hepes, pH 8.0, 1.5 mM MgCl 2 , 10 mM KCl, 1 mM DTT, 0.5 mM phenylmethylsulfonyl fluoride) and placed on ice for 15 min before disruption with a hypodermic syringe with a narrow-gauge needle attached. After centrifugation, the crude nuclear pellet was suspended in 0.7 volume of buffer C (20 mM Hepes pH 8.0, 1.5 mM MgCl 2 , 25% glycerol, 420 mM NaCl, 0.2 mM EDTA, pH 8.0, 1 mM DTT, 0.5 mM phenylmethylsulfonyl fluoride) and incubated on ice by stirring for 30 min. The nuclear debris was centrifuged and the supernatant was dialyzed for 1.5 h in 500 ml of buffer D (20 mM Hepes, pH 8.0, 20% glycerol, 100 mM KCl, 0.2 mM EDTA, pH 8.0, 1 mM DTT, 0.5 mM phenylmethylsulfonyl fluoride). After centrifugation of the extracts, the protein concentration of the supernatant was determined with a protein assay kit (Bio-Rad) and stored at Ϫ80°C until use.
DNase I Footprinting Analysis-The two segments spanning from Ϫ1364 to Ϫ954 and from Ϫ1001 to Ϫ621 plus Ϫ113 to Ϫ55, which were generated by digesting the pGL24SA and the pGL14266Ap constructs with SacI-NcoI, were incubated with nuclear extracts and subjected to DNase I footprinting analysis. The 3Ј-end of the segment was labeled by filling in 5Ј-overhangs with [␣-32 P]dCTP and Klenow polymerase. The unincorporated [␣-32 P]dCTP was removed by using Sephadex G-50 columns. Binding reactions were performed in a total volume of 50 l containing 50 mM Tris-HCl, pH 8.0, 20% glycerol, 100 mM KCl, 12.5 mM MgCl 2 , 1 mM EDTA, 1 mM DTT. 20 -40 g of nuclear extracts were incubated with 1 ng of radiolabeled probes at 4°C for 20 min. CaCl 2 was added to a final concentration of 2.5 mM, and 0.3 units of freshly diluted DNase I (Promega) were subsequently added to digest the probe at room temperature for 1 min. Digestion was stopped by adding an equal volume of stop buffer (200 mM NaCl, 30 mM EDTA, 1% SDS, 100 g/ml yeast RNA). The reaction mixtures were extracted with phenol/chloroform/isoamyl alcohol (25:24:1) followed by ethanol precipitation. The digested probes were analyzed by electrophoresis on a 6% urea-polyacrylamide gel. The two probes were sequenced with the T7 sequencing kit.
Electrophoretic Mobility Shift Assay (EMSA)-Oligonucleotides for mobility shift analyses were end-labeled with [␥-32 P]ATP and T4 polynucleotide kinase. The reactions were stopped by adding 80 l of TE buffer and heated at 70°C for 10 min. The unincorporated [␥-32 P]ATP was removed by using Sephadex G-25 columns. 10 ng of radiolabeled oligonucleotides were added to 8 g of the U937 cell nuclear extracts in a final volume of 25 l containing 1 g of nonspecific competitor poly(dI-dC) (Amersham Pharmacia Biotech), 50 mM Tris-HCl, pH 8.0, 50 mM KCl, 1 mM EDTA, 20% glycerol, and 1 mM DTT. The reaction mixtures were incubated at room temperature for 20 min. The DNA-protein complexes were separated from the free DNA by electrophoresis through a 6% nondenaturing polyacrylamide gel in 0.25 ϫ TBE buffer. The gels were dried and exposed to x-ray film. Competition experiments included a 100-fold excess of cold unlabeled DNA oligonucleotides, while supershift analysis included the addition of 0.8 -1.0 g of anti-PU.1 and anti-SRY antibodies (Santa Cruz Biotechnology, Biotech, Inc.) to the mixture and incubation for an additional 20 min. Oligonucleotides mutated at certain DNA-binding sites, such as putative PU.1, SOX5, and C/EBP␤ sites (see Table I), were also used in EMSA.

An Additional 5Ј-UT Alternative Exon 1 of the Human LST1
Gene Was Identified in the U937 Cells-Using RNA from U937 cells that strongly express the LST1 gene, we performed RT-PCR with primers (see Table I) designed to be unique for upstream sequences of the potential first exons and identified an additional exon 1 (termed exon 1a) that resides upstream of the published exon 1A (equivalent to the exon 1b in present study) of the LST1 gene (Fig. 1). Therefore, there are five alternative first exons (termed 1a, 1b, 1c, 1d, and 1e) in the 5Ј-flanking region of the human LST1 gene (Fig. 2). To determine the 5Ј-ends of the various mRNA forms, we performed 5Ј-RACE with RNA from U937 cells. The sequencing results from 36 cDNA clones of 5Ј-RACE products demonstrated that 23 clones (64%) represented the exon 1b-exon 2 splicing pattern (Fig. 3); two clones (5.6%) represented the exon 1c-exon 2 splicing pattern; one clone (2.8%) represented the exon 1e-exon 2 splicing pattern; and 11 clones (27.8%) represented unspliced RNA containing exon 1e, the intron between exon 1e and exon 2. These findings indicate that the transcription initiation site in exon 1b of the LST1 gene is preferentially utilized, and the exon 1b-exon 2 splice form is the most abundant transcript of the human LST1 mRNA.
The 5Ј-Flanking Region of the Human LST1 Gene Has Several Regions That Contribute to the Strong Promoter Activity-To characterize the LST1 gene promoter, a series of different reporter plasmid constructs that contained various 5Јflanking regions of the LST1 gene were generated in the luciferase reporter plasmid pGL3-Basic. We first prepared several deleted constructs that contained the 5Ј-flanking regions of each possible transcript of this gene (Fig. 4A). The levels of relative luciferase activities from extracts of transiently transfected U937 cells showed that the pGL2423 construct (Ϫ1363 to Ϫ54) that contained all 5Ј-UT alternative exons of the LST1 gene including exons 1a, 1b, 1c, 1d, and 1e had substantially stronger promoter activity than those constructs with longer or shorter 5Ј-flanking regions of the LST1 gene. The promoter activity of the pGL14266 construct (Ϫ1001 to Ϫ54) with deletion of the exon 1a-exon 1b genomic region (Ϫ1363 to Ϫ1001) dropped by a large amount compared with that of the pGL2423 construct, which suggests that the region from exon 1a to exon 1b (Ϫ1363 to Ϫ1001) is important for the LST1 promoter activity. However, as shown in Fig. 4B, the pGL24ApSt construct that contained the region from exon 1a to exon 1b (Ϫ1363 to Ϫ954) alone did not demonstrate strong promoter activity. Based on the pGL2423 construct, we further narrowed down the region within exon 1a-exon 2 and found that (a) the apparent promoter activity of the pGL12S10N construct (Ϫ1277 to Ϫ268) with deletion of the exon 1e-exon 2 region (Ϫ267 to Ϫ54) was reduced about 50% compared with that of the pGL12STN (Ϫ1277 to Ϫ54); (b) on comparison of the pGL24ApSt (Ϫ1363 to Ϫ954) with the pGL2S11N construct (Ϫ1363 to Ϫ579), the promoter activity of the latter construct that contained the exon 1b-exon 1c region was about 60% higher than that of the pGL24ApSt, indicating that the exon 1b-exon 1c region is also important; (c) comparison of the pGL12S10N (Ϫ1277 to Ϫ268) with the pGL12S11N (Ϫ1277 to Ϫ579) showed that the promoter activity of the latter construct with the deletion of the exon 1c-exon 1e region (Ϫ578 to Ϫ268) was relatively higher than that of the pGL12S10N (Fig. 4B). This result may indicate that the region from exon 1c to upstream of exon 1e contains some silencer elements.
The above deleted constructs did not contain the upstream region of exon 2 that included the splice site at the beginning of exon 2, and this may be a reason why these constructs (e.g. pGL12S10N, pGL12S11N, pGL2S11N etc.) produced relatively lower levels of luciferase mRNA. Therefore, beginning with the pGL2423 construct, we generated the deleted pGL2423Ap construct (Ϫ1363 to Ϫ621 plus Ϫ112 to Ϫ54) that retained the sequences from the region upstream of exon 2. As we expected, the result demonstrated that the pGL2423Ap construct pro-duced more RNA than the other constructs (Fig. 4C). Further deletion of this pGL2423Ap construct remarkably reduced its promoter activity as seen with constructs pGL12SNAp (Ϫ1277 to Ϫ621 plus Ϫ112 to Ϫ54), pGL14266Ap (Ϫ1001 to Ϫ54 plus Ϫ112 to Ϫ54), pGL24W13 (Ϫ1363 to Ϫ811 plus Ϫ112 to Ϫ54), and pGL24SA (Ϫ1363 to Ϫ954 plus Ϫ112 to Ϫ54). Furthermore, the promoter activities of two constructs that contained the exon 1a-exon 1b region (e.g. pGL24ApSt) and exon 1b-exon 1c region (e.g. pGL14266Ap), respectively, were significantly decreased (Fig. 4C). We found that the region from the end of exon 1b to the beginning of exon 1c (Ϫ1001 to Ϫ621) had an important role in the LST1 gene promoter activity. For instance, the promoter activity of the pGL24SA that deleted this region was reduced from about 70 to 10% compared with the pGL2423Ap (Fig. 4C). Thus, taken together, the data indicate that the combinations of these three regions (Ϫ1363 to Ϫ954, Ϫ954 to Ϫ621, and Ϫ112 to Ϫ54) are necessary for maximal LST1 transcript production.
To address the potential competitive effect of the other promoter regions, we constructed two other constructs, pGL14266luc Ϫ (Ϫ1001 to Ϫ54) that derived from the pGL14266 construct by deleting the luciferase gene, and pGL74226luc Ϫ (Ϫ571 to Ϫ54) that derived from the pGL74226 construct by deleting the luciferase gene, to perform co-transfection experiments with the pGL2423Ap construct. The results showed that both these constructs containing some of the alternative promoter regions could reduce the levels of relative luciferase activity of the pGL2423Ap construct from about 69 to 20% (for pGL14266luc Ϫ construct) or to 28% (for pGL74226luc Ϫ construct), indicating that these regions may compete partially for common transcription factors.
The 5Ј-UT Exon-splicing Pattern of the Human LST1 Gene in Transfected U937 Cells-We transfected U937 cells with certain constructs to identify the splicing pattern in U937 cells. We first transfected U937 cells with the pGL2423 construct that contained the regions encoding the 5Ј-UT alternative ex-

FIG. 3. 5-RACE analysis of the 5-UT alternative exons of the human LST1 gene in U937 cells.
The 5Ј-RACE experiment was performed using a pair of primers, QI and GSP42N, as described under "Experimental Procedures." The 5Ј-RACE products were subcloned and sequenced. 23 out of 36 cDNA products with exon 1b-exon 2 splicing are denoted with arrows. The thick arrows indicate that more than one 5Ј-RACE product has the same upstream end. The nucleotide sequence of transcript of exon 1b-exon 2 splicing is shown on the top. Numbers are relative to the translation start codon within exon 2. The triangle represents the major transcription start site by the majority of 5Ј-RACE products (8 clones). The star represents the major transcription start site identified by an S1 nuclease protection assay. The sequence complementary to the antisense primer GSP42N is underlined. The arrow represents the junction of exon 1b and exon 2.  (1a, 1b, 1c, 1d, and 1e) and amplified the resulting cDNA to identify the splicing pattern in the RNA transcribed from this construct. The results showed that exon 1a-exon 2 and exon 1b-exon 2 splicing patterns were identified and verified by sequencing (Fig. 5A) when we used the primers from 5Ј-UT exons in the LST1 gene and luciferase gene in the pGL3-Basic vector, while exon 1c-exon 2, exon 1d-exon 2, and exon 1e-exon 2 demonstrated unspliced products that were confirmed by sequencing. However, when we used the primers from 5Ј-UT exons and the endogenous exon 2, each of five 5Ј-UT alternative exons from the endogenous gene was able to splice to exon 2 in the same RNA preparations or RNA from untransfected U937 cells. PCR amplification directly from the RNA sample without reverse transcription excluded DNA contamination in the RNA preparation. We had essentially similar results when we used the pGL4214 (Ϫ1472 to Ϫ54) and the pGL54212 (Ϫ2401 to Ϫ54) constructs in which the upstream region of the 5Ј-UT region were longer than in the pGL2423 construct (data not shown). We also transfected U937 cells with the pGL24SX34 construct (Ϫ51 to ϩ1368) that included sequences from the middle of exon 2 to the end of exon 5 containing the stop codon of the LST1 gene. Even with this construct there was no substantial change in the 5Ј-UT exon-splicing pattern, and, in particular, no spliced form containing exon 1e-exon 2 was detectable. Also, the RT-PCR result of U937 cells transfected with the pGL3C8S construct in which the region from upstream of exon 1e to exon 2 was inserted into the pGL3-Control vector containing an upstream SV40 promoter did not show the exon 1e-exon 2 splicing (Fig. 5B). These data indicated that the exon 1a or 1b-exon 2 splicing patterns in the transfected U937 cells were the upstream splices that could be produced from the transiently transfected constructs, while exon 1c-exon 2, exon 1d-exon 2, and exon 1e-exon 2 splicing occurred in the same cells for the endogenous gene transcript.
A Major Transcription Initiation Site Is Located in the Region Upstream of Exon 1b in the Human LST1 Gene-On analysis of 23 cDNA clones of transcripts with exon 1b-exon 2 splicing generated by 5Ј-RACE, we found that the majority of the transcripts began about 199 nucleotides upstream of the ATG although the lengths of transcripts of the LST1 gene were different (Fig. 3). To further examine the major transcription start site within upstream regions of exon 1b, we used the pGL24SA construct spanning the upstream region of exon 1b as the protection probe in an S1 nuclease protection assay. The results showed that the major transcription initiation site of the LST1 gene was located 205 nucleotides upstream of the ATG translation start codon (Fig. 6).
Identification of DNA-binding Sites Within the Human LST1 Promoter Region-To identify transcription factors involved in the regulation of LST1 promoter activities, DNase I footprinting analysis and EMSA were used to locate the protein binding sites within the LST1 promoter region. We used the insert (Ϫ1363 to Ϫ954 plus Ϫ112 to Ϫ54) of the pGL24SA construct and the insert (Ϫ1001 to Ϫ621 plus Ϫ112 to Ϫ54) of the pGL14266 construct as probes for DNase I footprinting analysis. The results demonstrated that several regions bound to  Table I) which represent the transcripts of exon 1a-exon 2, exon 1b-exon 2, exon 1c-exon 2, exon 1d-exon 2, and exon 1e-exon 2 in transfected U937 cells. results represent the mean Ϯ S.E. for three independent transfection experiments with duplicate samples. A, the promoter activities of construct containing 2.4 kilobases of the 5Ј-flanking region of the LST1 gene (e.g. pGL54212, Ϫ2401 to Ϫ54) and constructs with the 5Ј-end serial deletions. B, the promoter activities of constructs with serial deletions from both sides based on the pGL2423 construct (Ϫ1363 to Ϫ54) that shows significant promoter activity. C, the promoter activities of constructs with serial deletions from both sides but containing part of the region upstream of exon 2 (Ϫ112 to Ϫ54). The 5Ј-UT exons are filled with different patterns and the coding region is filled with dark. Enzyme sites used for generating certain constructs are shown on the top. Numbers are nucleotides relative to the translation start codon. nuclear extracts of U937 cells (Fig. 7, A and B).
We performed gel shift analysis with five double-stranded DNA candidate probes (see Table I) termed BI, BII, BIII, BIV,  and BV corresponding to DNase-protected regions (I, II, and III, located within exon 1a-exon 1b region; and IV and V, located within exon 1b-exon 1c region). The EMSA results demonstrated that BI, BII, and BIV bound to nuclear extracts of U937 cells and produced distinct binding patterns, and BI and BII showed stronger shift bands (Fig. 8, A and B). The competition EMSA with nonradioactive competitors (e.g. unlabeled BI, II, and IV) showed that these cold competitors could significantly inhibit the binding of radioactive probes to nuclear extracts (Fig. 8, A, B, and D). By searching the data bases of TRANSFAC and TESS for putative transcription factors, we found some putative DNA-binding sites within these probe regions, for example, C/EBP␤ and GATA sites in BI probe, PU.1 and SOX5 sites in BII probe, and an SP1 site in BIV probe. However, EMSA competition assays with the cold consensus C/EBP␤ or PU.1 oligonucleotides could not inhibit the BI or the BII probe binding to U937 nuclear extracts (Fig. 8, A  and B). The protein-DNA complexes of the labeled PU.1 and C/EBP␤ consensus probes with the U937 nuclear extracts had a different electrophoretic motility from that of the BI and the BII probes (Fig. 8C). Attempts at supershift analyses with either anti-PU.1 rabbit antibody or anti-SRY goat antibody that can cross-react with SOX5 protein provided evidence that PU.1 and SOX5 transcription factors were not involved in DNA-protein binding (Fig. 8E). EMSA with mutated probes at the transcription factor-binding sites of PU.1 and SOX5 showed that the binding patterns were similar to that of the original probes (Fig. 9A). These findings suggest that a novel DNAbinding site in the promoter region may be associated with human LST1 gene regulation. Since the two 1mBII and 2mBII probes, which separated the putative PU.1-and SOX5-binding sites, did not demonstrate significant gel shifts (Fig. 9B), we assumed that the nucleotides between these two binding sites may be related to DNA-protein binding.
In order to verify this DNA-protein binding site, we performed EMSA by introducing a "CC" deletion mutation (6mBII) in the BII probe sequence that lies between the PU.1-and SOX5-binding sites. The result showed that this mutant probe could not effectively bind to the U937 nuclear extracts (Fig.  9A). When the probe BI was divided into two parts (2mBI and 3mBI) or narrowed down (4mBI) from both ends of this probe, we found that both the divided probes and the narrowed probe demonstrated a significant gel shift as the original BI did (Fig.  9C), and the competition EMSA with these cold mutant probes (2mBI, 3mBI, and 4mBI) could inhibit these bindings (data not shown). The other mutant probes (5mBI and 6mBI) did not effect the formation of the DNA-protein complexes (Fig. 9D). These findings indicate that multiple regulatory elements may be involved in LST1 gene expression. Also, we found the cold consensus SP1 oligonucleotide could inhibit BIV probe binding to U937 nuclear extracts but the cold consensus PU.1 oligonucleotide could not (Fig. 8E).
The Novel DNA-binding Site in the LST1 Gene Promoter Region Is Necessary for Gene Expression-To determine the functional significance of the DNA-binding site found in our EMSA result, we made the mutant construct termed pGL2423mAp based on the pGL2423Ap construct which changed the nucleotides at positions Ϫ1126 through Ϫ1123 (5Ј-ACTTCCTCTCCTAACAATGCTGGGG-3Ј) from "CCTA" between the PU.1 and SOX5 sites to "AGCT" within the BII probe sequence. The promoter activity of this mutant construct was significantly reduced compared with the pGL2423Ap control in U937, K562, and Jurkat cells although the levels of the pro-moter activities were apparently different in these different cell lines (Fig. 10, A-C). These data confirm that a novel DNAbinding site in the human LST1 promoter region plays an important role in regulation of this gene expression. DISCUSSION Based on the observation that each of the four known 5Ј-UT exons of the human LST1 gene lies closely downstream of the sequence CCCAG, we anticipated that other stretches of sequences that are preceded by CCCAG and followed by a plausible GT containing 5Ј-end splice site may also serve as 5Ј-UT exons. The template regions for the initial exons are close to one another and are not all preceded by obvious transcription factor-binding sites, so it seemed unlikely that a conventional promoter was located separately immediately upstream of each exon. We therefore also tried to identify an upstream common exon by RT-PCR with primers from exon 2 that contains the translation start codon and from regions within each of these putative exons. This uncovered an additional 5Ј-UT alternative exon (termed exon 1a) about 180 nucleotides upstream of the FIG. 6. S1 nuclease protection assay for the major transcription start site of the LST1 predominant transcript in U937 cells. The S1 nuclease protection assay was performed as described under "Experimental Procedures." The radiolabeled 24SA probe spanning the exon 1b-exon 2 splicing region (Ϫ1363 to Ϫ1001 plus Ϫ112 to Ϫ54) was hybridized to the total RNA from U937 cells and digested with S1 nuclease. The protected fragment was analyzed by electrophoresis on a 6% polyacrylamide sequencing gel. The 24SA probe was sequenced and shown on the left. The protected band is indicated with a star representing the transcription start site. A schematic representation of the 24SA probe used for S1 nuclease protection assay is shown on the top.
previously published exon 1A of the human LST1 gene (Figs. 1-2). However, our 5Ј-RACE analysis using RNA from U937 cells did not show a common initial exon preceding the 5Ј-UT alternative exons. We sequenced 36 cDNA clones of our 5Ј-RACE products and found that the majority of cDNA clones (64%) was transcripts from exon 1b to exon 2 (Fig. 3). The frequency of exon 1b-exon 2 splicing representation in the isolated cDNA clones indicates that the transcription initiation site in exon 1b of the LST1 gene is preferentially used in U937 cells. We also analyzed the EST data base of LST1 cDNAs focused on the 5Ј-end region and found that four cDNA from an adult brain showed exon 1b-exon 2 splicing. One cDNA from U937 cells stimulated with IFN-␥ showed exon 1c-exon 2 splicing. 7 of 8 cDNA from fetal liver, spleen, heart, and placenta showed exon 1e-exon 2 splicing, but one cDNA from a postnatal brain contained the intron between exons 1e and exon 2. Combining the EST data with our 5Ј-RACE results, there is a suggestion that the 5Ј-UT splicing pattern of the human LST1 gene may have tissue specificity, that is, the exon 1b-exon 2 splicing or exon 1e-exon 2 unspliced form appears to be abundant in U937 cells or adult tissues, while exon 1e-exon 2 splicing is more easily found in the fetal tissues, such as fetal liver, spleen, placenta and heart, etc. Also, de Baey's (9) results of RT-PCR showed that the major LST1 transcripts initiated from exon 1A (equivalent to our exon 1b) were detected in lymphocytes (e.g. PBMC, CD4 ϩ T cell clones, CD8 ϩ T cell clones, and B lymphoblastoid cell line LG2), in monocytic cell lines (e.g. U937 and Mono Mac 6), and also in many human tissues (e.g. lung, tonsil, thymus, placenta, kidney, spleen, and liver). However, a high level of LST1 mRNA was detected only in macrophage and monocytic cell lines (3,6). The previous RT-PCR data from our group showed that the expression pattern of the LST1 gene in U937 cells was similar to that found in human monocytic and dentridic cells derived from human blood (data not shown). Thus, we undertook the present study on the LST1 gene promoter in U937 cells and speculated that the region from the end of exon 1a to upstream of exon 1b probably contained the major LST1 promoter. However, our transient transfection studies did not fully support this prediction. The pGL24ApSt construct containing only this region (Ϫ1363 to Ϫ954) did not show a high level of promoter activity (Fig. 4B). We found that the region from the end of exon 1b to upstream of exon 1c (Ϫ1001 to Ϫ621) was very important for LST1 promoter activity although this region alone did not show ap- parent promoter activity (Fig. 4C). However, the pGL2423Ap construct combining three genomic regions (Ϫ1363 to Ϫ1001, Ϫ1001 to Ϫ621, and Ϫ112 to Ϫ54) together formed an active promoter. In addition, transfection experiments with some reporter constructs (e.g. pGL54212, pGL4214, pGL2423, pGL14266, and pGL2423Ap) in different cell lines, such as K562, Jurkat, 293T, and HeLa cells, showed much weaker promoter activities in these cells than in U937 cells (data not shown).
Holzinger's data showed that the human LST1 gene expression was enhanced by stimulation with 200 units/ml IFN-␥ for   (6), or the regulation of LST1 gene expression by IFN-␥ may be due to post-transcriptional mechanisms.
In addition to a number of alternative splice forms that occur within the coding region of the LST1 gene in human monocytes, there are at least five 5Ј-UT alternative exons spliced to exon 2 in U937 cells (9,13). To understand the splicing patterns of the 5Ј-UT alternative exons in the transfected cells or if the 5Ј-UT exon splicing would influence LST1 gene expression, we used an RT-PCR approach to analyze the cDNA in U937 cells transfected with our reporter constructs. Only exon 1a-exon 2 and exon 1b-exon 2 splicing that occurred in the transcripts from the different constructs tested were found in transfected U937 cells (Fig. 5A). We did not identify any exon 1e-exon 2 splicing in the mRNA derived from constructs transiently transfected into U937 cells even when the constructs were designed so that transcription started at approximately the 5Ј end of exon 1e (Fig. 5B).
Analysis of the 2.4-kilobase nucleotide sequence of the 5Јflanking region by searching the data bases of transcription factors showed that this region contained a number of putative regulatory cis-acting elements (Fig. 2). For example, BII and BI probe sequences contain a putative PU.1 and C/EBP␤ site, respectively. PU.1 is a myeloid-specific transcription factor that has a role in hematopoietic cell differentiation, proliferation, and apoptosis (19 -26). C/EBP␤ is expressed in myelomonocytic cells and its binding sites have been detected within regulatory and lipopolysaccharide-responsive elements of genes expressed in monocytes and macrophages (27)(28)(29)(30). We focused on the BII and BI probes and tried to identify which transcription factors might be involved in the regulation of the LST1 gene expression. However, the EMSA with mutant probes at the putative PU.1, SOX5, C/EBP␤, and GATA-binding sites did not show any evidence that these DNA-binding sites were related to these factors (Figs. 8A and 9, A and D).
EMSA with radiolabeled or cold consensus C/EBP␤ and PU.1 probe, or the supershift EMSA with anti-PU.1 and SRY antibodies did not show that these transcription factors were responsible for their observed gel shifts (Fig. 8, A-C). The putative C/EBP␤ site is predicted only by searching TRANSFAC but not by TESS, and C/EBP␤ is unlikely to be the factor binding to the BI probe. Evidence from another study (9) also showed that C/EBP was not involved in LST1 gene expression. The PU.1 site (GGAA) is located at the extreme 5Ј-end of the BII probe and it needs additional bases on its 5Ј side to form an appropriate protein-binding motif (31). When we narrowed down the BII probe (5Ј-ACTTCCTCTCCTAACAATGCTGGGG-3Ј) from both sides, we found the mutant 5mBII probe (5Ј-TTCCTCTCCTAACAATGCTG-3Ј) also significantly bound to U937 nuclear extracts (Fig. 9A). However, when we divided the BII probe into two part sequences (such as 1mBII and 2mBII) and performed EMSA, neither of these probes bound to U937 nuclear extracts (Fig. 9B). Taken together, these data indicate that there may be a novel DNA-binding site present between the putative PU.1 and SOX5 sites within the BII probe sequence. Furthermore, EMSA with 6mBII probe and transfection study with the pGL2423mAp construct that was mutated in the putative DNA-binding site confirmed that this binding site plays an important role in DNA-protein binding and LST1 promoter activity (Figs. 9A and 10).
In summary, we have identified an additional alternative exon 1 (termed exon 1a) residing upstream of the published exon 1A. The exon 1a-exon 1c genomic region (Ϫ1363 to Ϫ621) combined with sequences upstream of exon 2 (Ϫ112 to Ϫ54) of the human LST1 gene had the highest promoter activity. The exon 1b-exon 2 splice form was the dominant transcript of the human LST1 gene in U937 cells, and the major transcription initiation site was located at 205 base pairs position upstream of exon 1b. Transient transfection studies in U937 cells did not produce transcripts with the unusual splicing pattern seen in a portion of the endogenous transcripts. A novel DNA-binding site in the human LST1 promoter region and the interaction of multiple transcription factors may play an important role in regulating the LST1 gene expression. man's laboratory for helpful discussions, and Aumugham Raghunathan for electronic manuscript assistance.