HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 21, 18510-18516, May 24, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/21/18510    most recent M201886200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Tarumi, T. Articles by Gailani, D. Search for Related Content PubMed PubMed Citation Articles by Tarumi, T. Articles by Gailani, D. Social Bookmarking                      What's this?

Cloning and Characterization of the Human Factor XI Gene Promoter

TRANSCRIPTION FACTOR HEPATOCYTE NUCLEAR FACTOR 4 (HNF-4) IS REQUIRED FOR HEPATOCYTE-SPECIFIC EXPRESSION OF FACTOR XI^*

Takashi Tarumi, Dmitri V. Kravtsov, Mingming Zhao, Scott M. Williams, and David Gailani^§

From the Departments of Pathology and Medicine, Vanderbilt University, Nashville, Tennessee 37232-6307 and the  Department of Microbiology, Meharry Medical College, Nashville, Tennessee 37208

Received for publication, February 25, 2002, and in revised form, March 11, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Factor XI is the zymogen of a plasma protease produced primarily in liver that is required for normal blood coagulation. We cloned ~2600 base pairs of the human factor XI gene upstream of exon one, identified transcription start sites, and conducted a functional analysis. Luciferase reporter assays demonstrate that the 381 base pairs upstream of exon one are sufficient for maximum promoter activity in HepG2 hepatocellular carcinoma cells. The removal of 19 base pairs between 381 and 363 results in a nearly complete loss of promoter activity. This region contains the sequence ACTTTG, a motif required for binding of the transcription factor hepatocyte nuclear factor 4 (HNF-4) to the promoters of several genes. Gel mobility shift assays using HepG2 or rat hepatocyte nuclear extract confirm HNF-4 binds between bp 375 and 360. Scrambling the ACTTTG motif completely abolishes promoter activity in luciferase assays. The factor XI promoter functions poorly when transfected into HeLa carcinoma cells, and gel mobility shift experiments with HeLa nuclear extracts demonstrate no HNF-4 binding to the ACTTTG sequence. When a rat HNF-4 expression construct is co-transfected into HeLa cells, factor XI promoter activity is enhanced ~10-fold. We conclude that HNF-4 is required for hepatocyte-specific expression of factor XI.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Coagulation factor XI (FXI,¹ EC 3.4.21.27) is the zymogen of a plasma serine protease (FXIa) required for normal function of the intrinsic pathway of blood coagulation (1, 2). In humans, the gene for FXI is located at the distal end of chromosome 4 (4q35.2) and contains 15 exons spread over ~25 kb of genomic DNA (3, 4) (www.ensembl.org). Plasma FXI is produced primarily in hepatocytes (5-7), a characteristic shared with other blood coagulation proteases (the vitamin K-dependent proteins prothrombin and factors VII, IX, and X). Whereas considerable information is available concerning regulation of tissue-specific expression for the vitamin K-dependent proteases (8), little data are available for the FXI gene. Furthermore, extrapolation from data obtained for vitamin K-dependent protease genes to the FXI gene would probably be difficult. The domain structure of FXI differs substantially from the vitamin K-dependent proteins (1, 2, 5, 8). Even the trypsin-like catalytic domain, although homologous to catalytic domains of other clotting enzymes, is encoded by a different number of exons with different intron-exon boundaries (8). Indeed, the analyses of protein sequence and gene organization suggest that FXI is more closely related to proteases involved in fibrinolysis and inflammation than it is to other coagulation enzymes (3, 9-11).

Recent work suggests that variations in plasma FXI levels have important physiologic and pathologic implications. For example, a large clinical study suggests that the risk of deep vein thrombosis in the 10% of individuals in a population with the highest plasma FXI levels is more than 2-fold greater than that for the remaining 90% of the population (12). Therefore, a better understanding of the processes regulating FXI gene expression is required to identify factors that predispose to thromobembolic disease. As an initial step in understanding the regulation of FXI gene expression, we cloned the 5' upstream region of the human FXI gene and identified start sites for transcription, single nucleotide polymorphisms, and an element required for expression in a hepatocyte cell line. We have determined that the transcription factor hepatocyte nuclear factor 4 (HNF-4) is required for tissue-specific expression of human FXI.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Cloning and Sequencing of the FXI Gene Promoter Region-- A 2621-base pair fragment of the 5'-flanking region and exon one of the FXI gene was amplified by nested PCR using a Human Genome Walker kit (CLONTECH Laboratories, Palo Alto, CA) and oligonucleotides complementary to overlapping sequences in the known 5'-flanking region of the gene (3). For purposes of orientation, the first base pair of the published cDNA sequence is bp +1 (Fig. 1) (5). The oligonucleotides used for nested PCR are FXI-1, 5'-GAAAAGACCTTGTTGGCTTACTTG-3' (+40 to +17), and FXI-2, 5'- CTTGCTGCAATTCTTAATAAGGGTG-3' (+20 to 5). The fragment was sequenced in both directions using a thermo-sequenase radiolabeled terminator sequencing system (Amersham Biosciences). To prepare genomic clones, the 2.6-kb product from the Gene Walker kit or human genomic DNA from normal donors was amplified by PCR using high fidelity Pfu Turbo DNA polymerase (Stratagene, La Jolla, CA). The products were ligated into pGEM-T Easy T/A cloning vector (Promega, Madison, WI) and transfected into XL-1 Blue Escherichia coli (Stratagene). DNA from individual clones were sequenced as above.

The positions of the human FXI and plasma prekallikrein (PPK, EC 3.4.21.34) genes were located in the public data base of the human genome sequence using Ensembl genome server (www.ensembl.org). The Ensembl gene designation for both FXI (also designated F11) and PPK (also designated KLKB1) is ENSG00000088962. A 12-kilobase sequence encompassing the 3' end (exons 14 and 15) of the PPK gene and the 5' end (exons 1 and 2) of the FXI gene was aligned to the sequence of the cloned FXI promoter using MacVector software, version 6.0.1 (Oxford Molecular Group, Oxford, United Kingdom).

Dideoxyfingerprinting (ddF) for Identification of Common Polymorphisms-- Two overlapping areas covering base pairs 802 to +16 were amplified from genomic DNA by PCR. The oligonucleotides used to amplify 802 to 370 are 5'-AATCACTTGAACCTGGAAGCGG-3' (upstream primer) and 5'-CTAGCTGACCTTGAATCTCAAAG-3' (downstream primer), and the oligonucleotides used for fragment 412 to +16 are 5'-GTCTCCTCCCTCCATTATATTG (upstream primer) and 5'-GAAAAGACCTTGTTGGCTTACTTG (downstream primer). PCR products were size-fractionated on 2% NuSieve-agarose gels (FMC BioProducts, Rockland, ME) and purified with a QIAEX II gel extraction kit (Qiagen, Chatsworth, CA). The PCR primers were used to perform ddF in the 5' and 3' directions as described previously (13). Oligonucleotides were labeled with [-³²P]dATP and T4 polynucleotide kinase. Each ddF reaction is 8.0 µl in volume and contains 20 mM Tris-HCl, pH 8.4, 50 mM KCl, 2 mM MgCl₂, 20 µM of each dNTP, 100 µM dideoxy-GTP, 1 pmol of ³²P-labeled primer, 0.25 units of TaqDNA polymerase, and 10 ng of PCR fragment. Reactions were performed on a PerkinElmer model 460 DNA thermal cycler using the following parameters: denaturation at 94 °C for 45 s, annealing at 60 °C for 30 s, and elongation at 72 °C for 1 min for 30 cycles. PCR reactions were mixed with 16 µl of loading buffer (98% formamide, 10 mM EDTA, bromphenol blue, and xylene cyanol) and size-fractionated at 30 watts for 3 h at room temperature on a 0.5× MDE gel (FMC BioProducts) in 0.5× Tris-boric acid-EDTA buffer. Gels were dried and exposed to x-ray film overnight. The ddF pattern obtained for the cloned promoter fragment was considered to be the wild-type pattern. The DNA base pair sequence of PCR fragments with ddF patterns differing from the wild-type pattern were determined.

RNase Protection Assay-- RNase protection assays were performed using an ribonuclease protection assay kit (RPA III, Ambion, Austin, TX). An antisense RNA probe specific for FXI mRNA was generated by in vitro transcription using T7 RNA polymerase and [³²P]UTP. A 474-bp fragment of the FXI gene (434 to +40) in pGEM-T T/A Easy vector construct was used as template. The labeled probe (8 × 104 cpm) was incubated with 20 µg of total or 0.6 µg of poly(A) human liver RNA (CLONTECH) in hybridization buffer at 37 °C for 30 min. After incubation with RNase A/RNase T1 mix (supplied by the manufacturer) for 30 min at 37 °C, RNA duplexes were separated on 6% polyacrylamide gels containing 50% urea followed by autoradiography. Hybridization of the probe to yeast tRNA served as a negative control.

5'-Rapid Amplification of cDNA Ends (5'-RACE) for Identifying Origin of Transcription-- A 5'-RACE system (Invitrogen) was used according to the manufacturer's instructions. Using total human liver RNA (2 µg) as template, first-strand cDNA synthesis was performed using a primer complimentary to sequence in FXI exon 4, 5'-GGCAGTGTTTCTGTAAC-3'). cDNA was separated from unincorporated dNTP and primer using GlassMax spin cartridges (Invitrogen). A 5'-poly(C) tail was added using dCTP and terminal deoxynucleotidyltransferase. Poly(C)-tailed cDNA was amplified directly by PCR using an anchor primer supplied with the kit and a primer complimentary to the FXI coding sequence, 5'-GTGGGATCCTCAGATGGTGAGGC-3' (+255 to +233 of the cDNA). The PCR product was used as template for a second round of PCR using nested primers, one supplied with the kit and the other complementary to FXI cDNA sequence (+207 to +185), 5'-CTTGGGTGGTAAGTGCAGACTAC-3'. The final PCR product was size-fractionated on a 1.5% agarose gel and then purified using QIAEX II resin. Purified DNA was ligated into pGEM-T Easy T/A cloning vector, and individual clones were sequenced using a primer (5'-TCCTTCAAAGCAGGTGTCCTTC-3') complementary to sequence in FXI exon 3.

Construction of Luciferase Reporter Constructs-- A 2615-bp fragment spanning base pairs 2596 to +19 of the FXI gene was amplified by PCR using Pfu Turbo polymerase from the genomic DNA of a healthy volunteer with a normal plasma FXI level. The upstream PCR primer is 5'-CGTTGAGTGGAGATCACACCA-3', and 3'-primer is 5'-AAGCTTAAGCTTCTCGAGCTTGCTGCAATTCTTAATAAGGGTG-3'. The PCR product was ligated into vector pGL3-Basic (Promega), which places the promoter upstream of the firefly luciferase cDNA. Promoter truncation mutants were prepared by cutting the vector at the 5' end of the promoter with restriction endonucleases KpnI and Mlu I and then treating with exonuclease III and S1 nuclease (Promega) for various periods of time according to the supplier's instructions. After heating to 70 °C, Klenow enzyme was used to fill in the DNA followed by ligation with T4 DNA ligase. XL-1 Blue E. coli were transformed with the ligation mixtures, and clones were selected for DNA sequencing. A 577-bp promoter fragment in which the ACTTTG sequence between base pairs 370 and 365 was changed to GACAAT (scrambled) and prepared using a Chameleon site-directed mutagenesis kit (Stratagene).

Luciferase Assays-- The human hepatoma cell line HepG2 (ATCC HB 8065) and the human epitheloid carcinoma cell line HeLa (ATCC CCL2) were from the American Type Culture Collection (Manassas, VA). HepG2 cells were grown in Eagle's minimal essential medium supplemented with 10% fetal bovine serum, 1 mM sodium pyruvate, 2 mM L-glutamine, 0.1 mM non-essential amino acids, 1.5 g/liter sodium bicarbonate, and penicillin/streptomycin. HeLa cells were grown in Dulbecco's modified eagle medium supplemented with 5% fetal bovine serum and penicillin/streptomycin. Cells were grown to 50-70% confluence in 12-well culture plates at 37 °C in a humid atmosphere containing 5% CO₂ and then transfected with 1.5 µg of FXI promoter/pGL3 constructs using SuperFect Reagent (Qiagen). For normalization of transfection efficiency, all transfections included 30 ng of pRL-CMV vector (Promega) containing the Renilla luciferase gene under the regulation of a cytomegalovirus promoter. After incubation for 48 h, cells were lysed, and luciferase activity was measured using a Dual-luciferase reporter assay system (Promega). Luminescence was measured on a Monolight 2010 Luminometer (Analytical Luminescence Laboratories, San Diego, CA).

The cDNA for rat HNF-4 (14) was a gift from Margarita Cladaras (Aristotle University of Thessaloniki, Thessaloniki, Greece) and Josephine Carew and Ken Bauer (Harvard University). The cDNA was ligated into the EcoR1 site of mammalian expression vector pJVCMV (15), which places it under the control of a cytomegalovirus promoter. 1 mg of HNF-4/pJVCMV, 1 µg of pGL3 vector with or without the 577 proximal base pairs of the FXI promoter, and 30 ng of pRL-CMV were co-transfected into HeLa cells as described above. Promoter activity was determined by the dual-luciferase method.

Preparation of Nuclear Extracts-- HepG2 or HeLa cells grown to 80% confluence were used for preparing nuclear extract by a modification of the method of Dignam et al. (16). Cells were released from flasks with trypsin/EDTA, washed with phosphate-buffered saline, and resuspended in 1 ml of hypotonic buffer (20 mM Tris-HCl, pH 7.5, 10 mM KCl, 1 mM EDTA, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 10 µg/ml leupeptin, 40 µg/ml aprotinin, 1 mM benzamidine, and 10 µg/ml pepstatin). Cell membranes were broken with a Dounce homogenizer using Pestle B. Nuclei were pelleted by centrifugation at 2,000 × g for 10 min at 4 °C and washed with hypotonic buffer. Pellets were resuspended in high salt extraction buffer (20 mM Tris-HCl, pH 7.5, 420 mM KCl, 25% glycerol, 0.2 mM EDTA, 0.5 mM MgCl₂, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 10 µg/ml leupeptin, 40 µg/ml aprotinin, 1 mM benzamidine, and 10 µg/ml pepstatin), mixed vigorously by vortexing, and placed on ice for 15 min. Suspensions underwent centrifugation at 15,000 × g for 15 min at 4 °C. Supernatants (nuclear extracts) were stored in aliquots at 80 °C. Samples were examined for degradation by SDS-PAGE. Protein concentration was determined by colorimetric assay (Bio-Rad). Rat hepatocyte nuclear extract was purchased from Geneka Biotechnologies (Montreal, Canada).

Electrophoretic Mobility Shift Assay-- Two techniques were used to label oligonucleotides. The oligonucleotides were 5' end-labeled using T4 polynucleotide kinase and [-³²P]ATP (3000 Ci/mmol) for autoradiography. The labeled product had ~0.5 × 105 cpm/ng. Alternatively, oligonucleotides were labeled with biotin for chemiluminescence using a lightshift chemiluminescent electrophoretic mobility shift assay kit (Pierce). After labeling, complementary strands were mixed together in an equimolar ratio and allowed to anneal for 1 h at 37 °C to form double-stranded probe. Gel mobility shift assays were performed by incubating 0.5 ng of labeled probe with nuclear extract (10 µg) and competing oligonucleotides in binding buffer (10 mM Tris-HCl, pH 7.5, 50 mM NaCl, 1 mM dithiothreitol, 1 mM EDTA, 5% glycerol, and 1 µg/µl poly(dI·dC) at room temperature for 30 min. Mixtures were size-fractionated on a non-denaturing 4% polyacrylamide gel followed by drying and autoradiography for ³²P-labeled probes or transfer to nitrocellulose membranes and detection by streptavidin-HRP/chemiluminescence for biotin-labeled probes. Supershift reactions were run as described above with the exception that 2 µg of polyclonal anti-HNF-4 antibody H-171, which recognizes human, mouse, and rat HNF-4 (Santa Cruz Biotechnology, Santa Cruz, CA), was added.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

DNA Sequence of the Human FXIGene Promoter-- A comparison of previously published sequences for the human FXI gene (3) and cDNA (5) revealed that 275 bp upstream from the 5' end of intron A (including exon 1) are included in the published gene sequence. We cloned a 2601-bp fragment of the human FXI gene that includes 229 base pairs of the published sequence immediately upstream of exon 1 (3) and 2372 base pairs of additional 5' sequence. The published gene sequence agrees exactly with the sequence of the cloned fragment over the 229 base pairs at the 3' end of the clone (3). After the clone was completely sequenced, the sequence became available in the public data base for the region of chromosome 4 containing the FXI gene. When the sequences are aligned, there are two single nucleotide insertions in the cloned fragment and nine single nucleotide insertions in the public data base sequence (data not shown). All insertions are located upstream of the region required for proper promoter function in HepG2 hepatocellular carcinoma cells (see below). In addition, the sequences differed at two single base pair locations (single nucleotide polymorphisms) at positions 403 and 273. A partial sequence of the promoter clone is shown in Fig. 1A.

View larger version (41K): [in this window] [in a new window]   Fig. 1.   Factor XI promoter. A, DNA sequence of the 5'-flanking region of the human FXI gene. Sequence includes 850 bases upstream of the published sequence for the FXI cDNA (5), sequence for the first exon and the first exon/intron A boundary. The bold arrowhead indicates the exon 1/intron A boundary with intron sequence in lowercase letters. Sequence from bp 233 through intron A has been reported previously (3). Start sites for transcription as determined by 5'-RACE are indicated by inverted arrows (). Underlined sequence indicates an area rich in AAAG repeats. The poly(A) sequence enclosed in the gray box starting at position 587 may be 11 (as shown), 12, or 13 adenosine residues in length. Nucleotides at 403 and 273 enclosed in gray boxes marked with an asterisk indicate the location of single nucleotide polymorphisms in West Africans (see "Results" for explanation). B, relative positions of the human PPK and FXI genes on chromosome 4. The relative positions of exons 14 and 15 of the PPK gene and exons 1 and 2 of the FXI gene on chromosome 4 are shown. The units for the scale are kilobases of DNA. The cross-hatched bar indicates the position of the 2.6 kilobases of DNA sequence described in this study.

Exon 1 of the human FXI gene is not translated. Therefore, base pair +1 in Fig. 1 was assigned to the first base pair of the published sequence for the human FXI cDNA (5). An analysis of the sequence indicates that the human FXI gene lacks CAAT and TATA boxes. There are eleven consecutive adenosine residues between bp 587 and 577 in the sequence in Fig. 1A. Subcloning of individual FXI alleles from healthy persons revealed that this poly(A) sequence may be 11, 12, or 13 residues in length. The sequence in Fig. 1A differs from the public data base at two locations ((G/T) at 403 and (C/G) at 273, respectively, indicated by asterisks in Fig. 1A). Using the ddF technique, we examined the FXI promoter in eight DNA samples from individuals of West African ancestry. These samples were chosen because prior work demonstrated significant FXI gene heterogeneity in this ethnic group (13, 17). Positions 403 and 273 appear to be sites of common single nucleotide polymorphisms in this population ((G/T) at position 403 and (C/G) at position 273). At position 403, the allele frequency for G = 0.44 and for T = 0.56, whereas at 273, C = 0.5 and G = 0.5. Interestingly, G 273 was only found in samples with T at 403, suggesting that single nucleotide polymorphisms may be in linkage disequilibrium. A larger population survey involving multiple ethnic groups is currently underway.

FXI shares similar structural organization at the protein and gene levels with the serine protease PPK. In humans, the genes for FXI and PPK have been mapped to the long arm of chromosome 4 (4q32.5) (4, 18). The recent availability of sequence in the public data base for the region of chromosome 4 containing the FXI and PPK genes allows us to determine the relationship of the genes to each other and their orientation on the chromosome. As shown in Fig. 1B, the two genes are in a head-to-tail orientation with the 5' end of the PPK gene closest to the telomere and the 3' end of the FXI gene closest to the centromere. The 5' end of the FXI promoter clone described in this paper (Fig. 1B, bp 2596, cross-hatched bar) is located 5182 bp from the 3' end of the PPK gene stop codon.

Origins of Transcription-- RNase protection assays and 5'-RACE were employed to identify start sites for transcription using human liver RNA as template (Figs. 1 and 2). Both techniques demonstrate multiple start sites for transcription between bp 30 and 64. Of 26 clones sequenced in the 5'-RACE analysis, seven represent a start site at position 47, which may correspond to the intense band on the RNase protection assay (Fig. 2). There were five clones, each for positions 49 and 64. Additional minor start sites were identified by both techniques further upstream between positions 287 and 301 (Fig. 2).

View larger version (36K): [in this window] [in a new window]   Fig. 2.   RNase A protection assay for origins of transcription for the human FXI gene. The assay was performed using a 474-bp antisense RNA probe covering bases 434 to +40 of the sequence shown in Fig. 1. Human liver poly(A) RNA was used as template. The RNase protection assay is in the lane R. A conventional dideoxy-chain termination sequencing reaction was run in lanes G, A, T, and C to serve as molecular mass standards. Numbers along the left side indicate the size of bands in bases. The marks () along the right side designate the positions of RNA standards, and the numbers accompanying the marks refer to base pairs in the sequence in Fig. 1.

Promoter Activity in HepG2 and HeLa Cells-- Promoter activity was studied by transient transfection in the human hepatocellular carcinoma cell line HepG2 using a firefly luciferase reporter system. HepG2 has been used successfully to study promoters of a number of genes expressed in liver, including the factor VII (19, 20) and factor X (21, 22) genes. More importantly for our study, HepG2 cells constitutively secrete FXI protein (23). Efficiency of transfection was monitored by co-transfection with a CMV-driven vector containing the Renilla luciferase gene. The full-length promoter in these assays contains DNA from bp 2596 to bp +19. Additional promoter constructs were prepared from this full-length sequence by exonuclease digestion from the 5' end. The activity of construct 2596 in HepG2 cells was arbitrarily assigned a value of 100% (Fig. 3). 5' truncation to bp381 was not associated with an appreciable change in activity. Construct 375 had ~60% of the activity of the full-length fragment, whereas the activity of construct 362 was similar to that of the pGL3 vector without a promoter insert (Fig. 3). Therefore, all promoter activity is lost in HepG2 cells by the removal of 19 nucleotides between bp 381 and 362. We attempted to examine the activity in normal rat hepatocytes (provided by Dr. W. Russell, Vanderbilt University); however, the activity of the full-length promoter was very low in these cells.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Luciferase reporter assay for the FXI gene promoter in HepG2 and HeLa cells. Numbers along the left side designate the 5' base pair of the promoter insert in the pGL3 basic vector. pGL3 designates the results for the vector without promoter insert. Zero percent activity was determined from untransfected cells. Data for both cell lines represent percent activity relative to the result for the 2596-construct in HepG2 cells (arbitrarily set at 100%). Each result is the mean ± S.D. of three separate transfection experiments, each run in triplicate. Results are corrected for differences in transfection efficiency as determined by co-transfection with a Renilla luciferase construct. Error bars designate the mean ± S.D. for each set of results.

The human cervical carcinoma cell line HeLa was used as a non-hepatocyte-derived control for the activity study. As shown in Fig. 3, the full-length FXI promoter and truncation constructs have little activity in this cell line when compared with HepG2 cells (~10% of the activity in HepG2). An increase in promoter activity was noted starting with construct 92 with the activity being the greatest for the pGL3 vector without promoter. This finding is probably because of cryptic promoter elements within the pGL3 vector that bind transcription factors from HeLa cells. The results clearly demonstrate a difference in promoter activity in the liver cell line compared with the squamous epithelial line and suggest that the binding of hepatocyte-specific or hepatocyte-enriched transcription factor(s) to the area between bp 381 and 362 is necessary for proper gene expression.

HNF-4 Binds to the FXIPromoter-- The sequence between bp 385 and 350 is shown at the top of Fig. 4A. This region contains the sequence ACTTTG, which has been identified in several gene promoters as a component of binding sites for the liver-enriched transcription factor HNF-4 (22, 24-27). In Table I, the sequence of the putative HNF-4 binding element in the FXI promoter is compared with corresponding sequences in the human factor VII (27-29), factor IX (25, 26, 30), and factor X (22) genes as well as the strong HNF-4 binding element APF-1 from the apolipoprotein CIII gene (31). A labeled double-stranded oligonucleotide covering this region (Fig. 4A, FXI-WT) was used in gel mobility shift assays with nuclear extracts from HepG2 cells, HeLa cells, or normal rat hepatocytes (Fig. 4B). Two bands in close proximity to each other were detected with HepG2 nuclear extract (poorly resolved in Fig. 4B) that are competed away efficiently by cold probe, whereas a single band is noted for HeLa samples. Several bands are detected when rat hepatocyte nuclear extract is used (competition experiment for the rat hepatocytes not shown). The addition of antibody specific for HNF-4 to the reactions results in the supershifting of one band in HepG2 samples and several bands in rat hepatocyte samples, demonstrating that HNF-4 is present in these nuclear extracts and that HNF-4 binds to sequence in the FXI-WT probe. In comparison, no supershift is detected with HeLa nuclear extract.

View larger version (48K): [in this window] [in a new window]   Fig. 4.   Gel mobility shift assays for the 385 to 350-region of the FXI promoter. A, oligonucleotides used for gel shift assays. Oligonucleotide FXI-WT was labeled with either 32P or biotin. All other oligonucleotides were used as unlabeled competitors. All oligonucleotides used in these experiments are double-stranded. Only the sequence corresponding to the "sense" sequence in Fig. 1 is shown. B, gel shift and supershift experiments with HepG2, HeLa, and rat hepatocyte nuclear extracts. Protein-DNA complexes between nuclear extracts (10 µg) and 32P-labeled FXI-WT probe were analyzed on 4% non-denaturing acrylamide gels. The right triangles in the Competitor row for HepG2 and HeLa cells indicate a 50-, 100-, or 200-fold excess (from left to right) of cold FXI-WT probe over labeled probe. The symbol + in the HNF-4 antibody lane indicates that 2 µg of polyclonal anti-HNF-4 IgG was added to the reaction. Arrows to the left and right of the figure indicate bands that are competed away by cold oligonucleotide or supershifted (SS) in the presence of antibody. Because some supershifted bands run close to the tops of the gels and are difficult to resolve, gels were run for prolonged periods of time, resulting in the unbound probe running off the bottom of the gel. C, competition assays using scrambled oligonucleotides. A biotin-labeled FXI-WT oligonucleotide was mixed with buffer (left lane) or 10 µg of HepG2 nuclear extract (all other lanes). Shifted bands were detected with a chemiluminescence system as described under "Experimental Procedures." The arrow to the left of the figure designates the HNF-4-containing band in the sample without competitor. Unlabeled competing oligonucleotides (listed across the top of the panel) were used in 200-fold molar excess over labeled probe. The symbol () indicates the absence of competing oligonucleotide in the reaction. WT, FXI-WT; 1-7, scrambled FXI sequence 1 through 7; 366A, sequence mimicking factor IX Leyden mutation; A, HNF-4 binding sequence APF-1 from the apolipoprotein CIII gene. D, APF-1 oligonucleotide and HepG2 nuclear extract. The HNF-4 binding sequence APF-1 from the apolipoprotein CIII promoter was labeled with 32P and mixed with either buffer (left lane) or 10 µg of HepG2 nuclear extract (all other lanes). Unlabeled arrows to the left designate bands competed away by cold oligonucleotide. The arrow labeled SS indicates the band supershifted by anti-HNF-4 antibody (far right lane). Unlabeled competing oligonucleotides are listed across the top of the panel. The symbol () indicates the absence of competing oligonucleotide in the reaction. The right triangles indicate either a 50- or 200-fold (left to right) excess of APF-1 or FXI-WT oligonucleotide. Lanes 1-7 indicate reactions containing a 200-fold excess of cold oligonucleotide FXI 1-7 (A).

View this table: [in this window] [in a new window]   Table I HNF-4 (ACTTTG) binding elements from liver-specific genes The sequence surrounding the ACTTTG motif required for HNF-4 binding to the FXI promoter is compared with the corresponding sequences in the genes for apolipoprotein CIII (APF-1) and coagulation factors VII, IX, and X. The position of scrambled sequence from oligonucleotides FXI 2-5 (Fig. 4,A) are indicated at the bottom of the table.

To further characterize this region, a series of double-stranded oligonucleotides representing base pairs 385 to 350 were prepared with scrambled sequence in five or six adjacent base pairs (Fig. 4A). These oligonucleotides were used as competitors in the gel mobility shift assay with HepG2 nuclear extract and labeled FXI-WT probe (Fig. 4C). A variant of hemophilia B referred to as factor IX Leyden may be caused by point mutations in the ACTTTG HNF-4 binding element of the factor IX gene (25, 26). An additional FXI oligonucleotide was prepared with a T  A substitution at the position corresponding to base pair 366 in the FXI promoter to mimic a factor IX Leyden-type mutation. Finally, an oligonucleotide representing the HNF-4 binding site APF-1 from the apolipoprotein CIII gene was prepared. For the competition study shown in Fig. 4C, WT-FXI oligonucleotide was labeled with biotin and detected by chemiluminescence because adequate resolution of relevant bands could not be obtained with ³²P-labeled probes. Cold FXI-WT probe as well as scrambled FXI oligonucleotides 1, 2, 6, and 7 compete successfully with labeled probe for binding to HNF-4 (Fig. 4C). As expected, APF-1 also competes successfully. In contrast, FXI oligonucleotides 3, 4, and 5 and the oligonucleotide with adenosine at 366 (Leyden substitution) failed to compete well with labeled probe. The positions of the scrambled bases in oligonucleotides 3, 4, and 5 relative to the putative HNF-4 binding site in the FXI promoter are shown at the bottom of Table I. Note that these oligonucleotides contain or flank the ACTTTG motif.

As can be seen in Fig. 4C, we had difficulty obtaining well resolved results with labeled FXI oligonucleotides and HepG2 nuclear extract in competition assays. To verify the results shown in Fig. 4C, we repeated competition assays using a labeled probe for the well characterized HNF-4 binding element APF-1 (Fig. 4D). Two clearly shifted bands are noted with this probe in the presence of HepG2 nuclear extract. The more intense lower band contains HNF-4 as demonstrated by the supershift experiment (Fig. 4D, far right lane). Both cold APF-1 probe as well as WT-FXI and FXI-1, 2, 6, and 7 effectively compete with the labeled probe for binding to nuclear extract. Consistent with the results in Fig. 4C, oligonucleotides FXI-3, 4, and 5 do not compete well indicating that the HNF-4 binding elements in these constructs have been disrupted.

The Importance of the ACTTTG Motif for Activity in Luciferase Assays-- The importance of the HNF-4 binding ACTTTG motif to promoter activity in the luciferase reporter assay was tested by replacing this sequence with a scrambled sequence GACAAT in a 577-bp promoter fragment in pGL3. As shown in Fig. 5A, the scrambled GACAAT sequence completely disrupts promoter activity in HepG2 cells. The activity of the full-length FXI promoter in HeLa cells is only ~10% of the activity in HepG2 (Fig. 3). Furthermore, gel mobility shift/supershift assays using HeLa nuclear extracts indicate that little or no HNF-4 interacts with oligonucleotides containing the FXI gene HNF-4 binding site (Fig. 4B). When a rat HNF-4 expression construct is co-transfected into HeLa cells with a FXI promoter/pGL3 construct, the activity of the FXI promoter is increased >20-fold compared with transfections in the absence of HNF-4 co-transfection (Fig. 5B). Interestingly, the co-expression of HNF-4 with the promoter construct containing the scrambled HNF-4 motif also resulted in increased promoter activity (Fig. 5B, ~9-fold over control). A similar phenomenon was reported for a factor VII promoter construct containing a mutation in the HNF-4 binding site (20). This suggests that a portion (~40-50%) of the signal increase obtained with HNF-4 overexpression is not related to the ACTTTG motif but may be related to secondary sites for HNF-4 binding within the promoter or on the pGL3 vector. Nevertheless, a significant increase in promoter activity related to the ACTTTG motif (~10-fold) occurs when HeLa cells are induced to overexpress rat HNF-4.

View larger version (10K): [in this window] [in a new window]   Fig. 5.   Effects of scrambling the ACTTTG motif, and rat HNF-4 overexpression on promoter function in luciferase reporter assays. A, HepG2 cells. HepG2 cells were transfected with 2 µg of pGL3 vector containing either a 577-bp fragment of the wild-type FXI promoter (WT), the same fragment with the ACTTTG sequence scrambled to GACAAT (SC), or no fragment (C). Zero percent activity was determined from untransfected cells. The value for the wild-type vector was arbitrarily assigned a value of 100%. Results are the mean of six separate transfections. Results are corrected for differences in transfection efficiency as determined by co-transfection with a Renilla luciferase construct. Error bars designate the standard deviation for each set of results. B, HeLa cells. HeLa cells were co-transfected with 1 µg of pGL3 vector containing either a 577-bp fragment of the wild-type FXI promoter (WT) or the scrambled sequence (SC) and with 1 µg of a CMV promoter-based expression vector containing the rat HNF-4 (HNF-4) or no cDNA (C). Numbers at the left of the panel indicate the fold increase in promoter activity compared with the cells transfected with wild-type FXI promoter without HNF-4 co-transfection, which was arbitrarily assigned a value of 1. Results are the means ± S.D. of three separate transfections and are corrected for variation in transfection efficiency as determined by co-transfection with a Renilla luciferase construct (see "Experimental Procedures"). Error bars indicate the mean ± S.D. for each set of results.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Activated FXI (FXIa) is a trypsin-like serine protease that contributes to hemostasis by activating factor IX in the presence of calcium ions (1). This appears to be an important step in sustaining thrombin generation after the initial formation of a fibrin clot. Although gene mutations responsible for congenital FXI deficiency have been well documented (32, 33), there is little information available regarding transcriptional regulation of the FXI gene or the physiologic and pathologic factors that influence plasma FXI levels. The results of a recent clinical study indicate that plasma levels of FXI may influence the risk for venous thrombotic disease. Using the registry of the Leiden Thrombophilia Study, Meijers et al. (12) determined that the age and sex-adjusted odds ratio for venous thrombosis is 2.2 in individuals with plasma FXI levels above the 90th percentile compared with those with levels below the 90th percentile. Chan et al. (34) provided evidence of a role for FXI in the thrombotic disorder affecting mice with congenital deficiency of protein C (PC), which causes death in utero or shortly after birth from disseminated coagulation. Crossing PC-deficient mice with FXI-deficient animals generated double homozygous knock-out (PC//FXI/) animals that lived for up to three months (34, 35), strongly indicating that FXI contributes to the consumptive coagulopathy in PC-deficient animals. We cloned and conducted a preliminary characterization of the promoter region of the human FXI gene as an initial step in understanding factors that control FXI mRNA transcription and plasma protein levels.

FXI shares similar domain structure and is 58% identical in amino acid sequence with PPK (5, 36). PPK is the zymogen of the kininogenase -kallikrein, which liberates the potent vasoactive nanopeptide bradykinin from kininogens (37). The genes for FXI and PPK are similarly organized, each containing 15 exons with identical intron/exon boundaries (3, 11, 18). Gene localization studies have previously mapped the two genes to the same region of the long arm of chromosome 4 (4q35.2) in humans (4, 18) and to chromosome 8 in mice (38). Taken as a whole, these findings strongly indicate that the FXI and PPK genes are the products of a duplication event involving a common ancestral gene. The recent availability of complete sequence for the area of chromosome 4 encompassing the two genes allowed us to determine their relative positions and orientations to each other. The analysis demonstrates that the two genes are indeed very close together (separated by <8 kilobases), raising the possibility of coordinated regulation. PPK and FXI are both produced primarily in liver; however, mRNA for both proteins have also been detected in the pancreas and kidney (6, 7), consistent with similar mechanisms of regulation. To date, a functional analysis of the PPK gene promoter has not been described.

Multiple start sites for transcription were identified for the human FXI gene. The gene appears to lack canonical CAAT and TATA boxes, a condition often associated with multiple start sites for transcription (39). Luciferase reporter construct experiments determined that the 381 base pairs immediately 5' of the published sequence for the FXI cDNA are sufficient for maximal promoter activity. The DNA sequence of our promoter clone differed from the public data base at only two sites, base pairs 403 and 273. An examination of the FXI promoter regions from eight individuals of West African descent suggests that these locations are the sites of common single nucleotide polymorphisms in this population. Although these polymorphisms appear to have minimal impact on promoter function in HepG2 cells (data not shown), it will be interesting to determine whether the polymorphisms are prevalent in other ethnic groups and whether there is an association among different alleles and plasma FXI levels.

The sequence between base pairs 381 and 362 is critical for promoter function in HepG2 cells, a hepatocellular carcinoma cell line that constitutively expresses FXI (23). Luciferase and gel shift experiments demonstrate that an ACTTTG sequence within this region is necessary for promoter activity and that HNF-4 binds to this sequence. Similar to other coagulation proteases, plasma FXI is produced primarily in hepatocytes (6, 7, 40). Therefore, it is not surprising that the liver-enriched transcription factor HNF-4 is important for FXI gene expression. HNF-4, originally identified as a DNA binding factor necessary for transcription of the transthyretin gene in hepatoma cells (41), contains a zinc-finger motif and is a member of the steroid hormone/thyroid hormone nuclear receptor superfamily of transcription factors (24, 42, 43). HNF-4 is generally classified as an "orphan" member of this family, because no ligand had been identified for the putative receptor (42, 43). However, Hertz and colleagues (44, 45) have presented data indicating that fatty acyl-CoA thioesters bind to the receptor domain of HNF-4 and modulate transcription factor activity. HNF-4 binds DNA as a homodimer (24, 42, 43) and is required for normal gastrulation in mouse embryos (46) and hepatocyte differentiation (47) as well as production of coagulation factors VII (27-29), IX (25, 26, 30), and X (22). The importance of HNF-4 to expression of these protease genes is attested to by the severe protein deficiencies and bleeding disorders associated with mutations in HNF-4 binding sites. A severe form of factor IX deficiency (hemophilia B) is caused by an A, C, or G for T substitution at base pair 20 or 21 in the ACTTTG sequence of the factor IX gene (25, 26). Interestingly, this condition, referred to as factor IX Leyden, is associated with a normalization of factor IX levels at puberty because an androgen response element between bp 36 and 17 can compensate for the lack of HNF-4 binding (25, 26). Two mutations in the HNF-4 binding site of the factor VII promoter have also been associated with severe factor VII deficiency and bleeding (19, 20). Thus, HNF-4 appears to be an important transcriptional regulator of multiple components of the plasma coagulation mechanism. The one essential coagulation protease gene not under regulation by HNF-4 is prothrombin, which requires transcription factor HNF-1 for tissue-specific expression (48, 49). However, HNF-4 is a transcriptional activator of the HNF-1 gene, suggesting that HNF-4 may have an indirect effect on prothrombin through the regulation of HNF-1 expression (50, 51).

HNF-4 is one of a small group of liver-enriched transcription factors, which includes HNF-1, C/EBP, and HNF-3, that governs liver-specific gene expression (52). HNF-4 controls the expression of several genes that have been implicated in vascular disease, including genes for apolipoproteins (AI, AII, AIV, B, and CIII) and coagulation factors (VII, IX, and X) (53). The finding that HNF-4 is necessary for hepatocyte-specific expression of FXI indicates that the regulation of FXI expression shares similar features with regulation of these genes. Similar to the case for FXI (12), epidemiological studies suggest that elevated levels of factors VII (54, 55), IX (56), and X (57) are associated with an increased risk for vascular thrombosis. Therefore, it is conceivable that HNF-4 may influence predisposition to thrombotic disorders through the regulation of coagulation factor gene expression. FXI mRNA has also been detected in the pancreas and kidney by Northern blot analysis (6) and in platelets by reverse transcriptase-PCR (13). In this regard, the finding that FXI expression is regulated by HNF-4 is of interest, because this transcription factor is expressed in the pancreas, kidney, and gut in addition to liver (24, 58). Therefore, HNF-4 may have a role in the regulation of FXI mRNA expression in non-hepatic tissue as well as in the liver.

	ACKNOWLEDGEMENTS

We thank Mao-Fu Sun for technical expertise and Jean McClure for graphics work and preparation of the manuscript.

	FOOTNOTES

^* This work was supported by NHLBI, National Institutes of Health Grant HL58837.The costs of publication of this article were defrayed in part by the payment of page charges. The 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/EBI Data Bank with accession number(s) AF486577.

^§ Established Investigator of the American Heart Association. To whom correspondence should be addressed: Division of Hematology/Oncology, Vanderbilt University, 777 Preston Research Bldg., 2220 Pierce Ave., Nashville, TN 37232-6307. Tel.: 615-936-1505; Fax: 615-936-3853; E-mail: dave.gailani@mcmail.vanderbilt.edu.

Published, JBC Papers in Press, March 12, 2002, DOI 10.1074/jbc.M201886200

	ABBREVIATIONS

The abbreviations used are: FXI, coagulation factor XI; HNF-4, hepatocyte nuclear factor 4; PPK, plasma prekallikrein; ddF, dideoxyfingerprinting; RACE, rapid amplification of cDNA ends; CMV, cytomegalovirus; PC, protein C; WT, wild-type.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES