Role of the Liver-Enriched Transcription Factor Hepatocyte Nuclear ...

MOLECULAR AND CELLULAR BIOLOGY, May 1996, p. 1936–1945 0270-7306/96/$04.0010 Copyright q 1996, American Society for Microbiology

Vol. 16, No. 5

Role of the Liver-Enriched Transcription Factor Hepatocyte Nuclear Factor 1 in Transcriptional Regulation of the Factor VIII Gene L. K. MCGLYNN, C. R. MUELLER, M. BEGBIE, C. R. NOTLEY,

AND

D. LILLICRAP*

Departments of Pathology and Biochemistry, Queen’s University, Kingston, Ontario, Canada K7L 3N6 Received 3 May 1995/Returned for modification 13 June 1995/Accepted 1 February 1996

Coagulation factor VIII is an essential cofactor required for normal hemostatic function. A deficiency in factor VIII results in the bleeding disorder hemophilia A. Despite the fact that the factor VIII gene was cloned a decade ago, the mechanisms which control its transcription remain unresolved. In our studies, we have characterized 12 protein binding sites within the factor VIII promoter by DNase I protection assays performed with rat liver nuclear extracts. Three of these elements (sites 1 to 3) are situated within the 5* untranslated region of the gene, while three other sites (sites 4 to 6) lie within the first 100 bp upstream of the transcriptional start site. We have identified an additional site (site 7) ;300 bp upstream from site 6, as well as a cluster of five sites in a 250-bp region which terminates ;1 kb from the transcriptional start site. Seven of these binding sites (sites 2, 3, 4, 6, 7, 9, and 10) bind members of the C/EBP family of transcription factors. DBP also binds to five of these sites (sites 3, 4, 6, 7, and 9). Utilizing transient transfection studies in HepG2 cells, we have shown that deletion of the factor VIII promoter sequences distal to nucleotide 244 results in a significant but small increase in promoter activity. The activity of each of the various 5* deletion constructs is significantly enhanced by cotransfection of C/EBPa and D-site-binding protein expression plasmids, while cotransfection of both C/EBPa and C/EBPb plasmids resulted in a further enhancement of transactivation. These studies also provide evidence of a repressor element located between nucleotides 2740 and 21002. Since the minimal promoter sequence (244 to 1148) maintains the transcriptional activity of the full-length promoter sequence, we proceeded to identify additional factors binding to sites 1 to 4. Competition studies revealed that a ubiquitous transcription factor, NF-Y, binds to site 4, while the liver-enriched transcription factor hepatocyte nuclear factor I (HNF-1) binds to site 1. Mutation analysis of the minimal promoter demonstrated that HNF-1 is critical for activating transcription of the factor VIII gene in vitro. Our results also suggest that the multiple upstream elements that we have identified may act as a backup regulatory region in the event of disruption of the HNF-1 element in the 5* untranslated region. factor VIII may be attributed to mRNA accumulation and inefficiencies in the processing and secretion of the factor VIII protein (8, 9, 21). Thus, factor VIII synthesis may be regulated at the posttranscriptional level as well as at the transcriptional level. The mutations responsible for the hemophilia A phenotype consist of a variety of single base substitutions, deletions, and insertions that have been found throughout the gene (54). To data, no mutations have been identified in the factor VIII promoter. In addition to cloning the factor VIII coding region in their initial studies, Gitschier et al. (14) sequenced 1.2 kb of the 59 flanking region of the gene. They also identified the transcriptional start site by RNase protection to be 170 bp 59 of the initiation methionine codon. There is a GATAAA sequence 30 bp 59 of the transcriptional start site which is thought to represent an alternative TATA box, but other details concerning the cis-acting sequences which regulate transcription of the gene remain unknown. In this study, we demonstrate that members of the basic leucine zipper (b/ZIP) family of proteins, the CCAAT/enhancer-binding proteins (C/EBPa and C/EBPb) and D-site-binding protein (DBP), bind to several cis-acting elements in the factor VIII promoter and are important for transactivating the expression of the factor VIII gene. In addition, we show evidence of a possible repressor element approximately 1 kb 59 of the gene. Furthermore, our studies show that hepatocyte nuclear factor 1 (HNF-1) binding to its cognate sequence element in

Factor VIII is a large plasma glycoprotein that circulates in a noncovalent complex with von Willebrand factor (20). When factor VIII is released from von Willebrand factor and activated by thrombin, factor VIIIa participates as a cofactor in the activation of factor X by factor IXa. Deficiency of factor VIII results in hemophilia A, an X-linked bleeding disorder which affects 1 in 5,000 males. The gene for factor VIII was cloned and characterized in 1984 (14, 53). It spans 186 kb on the long arm of the X chromosome at position Xq28. The gene encodes a mature mRNA of 9 kb which is expressed in a number of different tissues (58). Wion et al. (58) have shown, using RNase protection studies, that factor VIII is expressed in whole liver and isolated hepatocytes as well as in the spleen and in lymph nodes. Factor VIII mRNA was also found at low levels in the pancreas, kidney, and muscle. The clinical evidence that supports the liver as a major site of factor VIII synthesis is the finding that liver transplantation in hemophiliacs corrects their bleeding disorder (3, 24). In normal individuals, factor VIII circulates in the plasma in trace amounts (;100 to 200 ng/ml). The biosynthesis and secretion of factor VIII have been extensively studied (22, 42, 55, 59). It has been suggested that the low circulating levels of * Corresponding author. Mailing address: Department of Pathology, Richardson Laboratory, Queen’s University, Kingston, Ontario, Canada K7L 3N6. Phone: (613) 545-2823. Fax: (613) 545-2907. Electronic mail address: [email protected]. 1936

VOL. 16, 1996

ROLE OF HNF-1 IN REGULATION OF FACTOR VIII GENE

the 59 untranslated region (UTR) plays an important role in sustaining the transcriptional activity of this promoter. MATERIALS AND METHODS Factor VIII promoter constructs. Various 59 deletion constructs of the factor VIII promoter sequence were constructed by using PCR. To generate factor VIII promoter fragments, PCR was performed with the sense primers (mutated nucleotides are in boldface and underlined) F8S-1002 (59-CTCACCATGGCTA CATTCTGA-39), F8S-739 (59-GTGGGAGTGGAACTACGT-39), F8S-211 (59GGCTAGGAATAGGAGCACAA-39), F8S-44 (59-GGCTGCTTCCCACTGA TAAA-39), F8Psite 7 (59-TAAAACAATAGTTGCCTAACCTC-39), F8Psite 4 mut 215 (59-CCACTGATAAAAAGGAAGCATTCCTATCGGTTACT-39), F8Psite 3 mut 117 (59-TTACTGCTTAGTGCTGAGCACTTCCAGT-39), F8Psite 4 mut 117 (59-CCACTGATAAAAAGGAAGCAATCCTATCGGTTA CTGCTTAGTGCTGAGCACTTCCAGT-39), F8Psite 2 mut (59-ATTGG GACTTTTCATTAAAGCAGTATAT-39), and F8Psite 3 mut INR (59-TTA GAGCTTAGTGCTGAGCA-39) with the antisense primer F8AS1148 (59-CT GGAGAAGCAAAAGGTTAAT-39), F8AS125 (59-CCACTGGATGTGCT CAGCACT-39), or F8Psite 1 HNF-1-mut (59-CTGGAGAAGCAAAAGGTTG GTACCTCTCT-39). Amplifications were in 13 PCR buffer [67 mM Tris-HCl (pH 8.8), 16.6 mM (NH4)2SO4, 6.7 mM MgCl2, 100 mg of bovine serum albumin (BSA) per ml, 14.4 mM b-mercaptoethanol] with 200 mM each deoxynucleoside triphosphate, 5 pmol of each primer, 1 mg of genomic DNA, and 2.5 U of Taq DNA polymerase (Gibco/BRL). The reaction mixture was denatured for 5 min at 988C prior to 30 cycles of amplification (denaturation at 948C for 1 min, annealing at 608C for 1 min [488C for mutant oligonucleotides], and extension at 728C for 3 min). The PCR products were then purified from a 1% low-meltingpoint agarose gel and cloned into a TA vector, pCRII (Invitrogen). Dideoxy sequencing confirmed the factor VIII promoter sequence and determined the orientations of the promoter fragments for further manipulation. The fragments were subsequently excised with XhoI and KpnI and subcloned into pUC19 (Gibco/BRL) cut with SalI and KpnI. The promoter fragments were then cloned into the KpnI-HindIII sites of pGL2-Luc Basic (Promega). DNase I protection assays. Probes for DNase I protection were restriction fragments of the 1.2-kb factor VIII promoter sequence. The 1.2-kb PCR product was cloned into a pT7 blue vector (Novagen). Two fragments, an EcoRI-KpnI fragment corresponding to the proximal promoter (2277 to 1146) and an EcoRI-HindIII fragment corresponding to the distal promoter (21002 to 2278), were excised from the pT7 blue vector and subcloned into pUC19 to facilitate their 39 end labeling. The probes were prepared by digestion of the proximal and distal promoter fragments with EcoRI or NcoI (distal) followed by end labeling with [a-32P]dATP and Klenow polymerase. The appropriate fragment was purified from low-gelling-temperature agarose after digestion with KpnI (proximal), HindIII (distal), or EcoRV (distal). DNase I protection assays were carried out as described by Lichtsteiner et al. (25) with rat liver nuclear extracts prepared by the method of Gorski et al. (15), with modification as described by Maire et al. (29). A Maxam-and-Gilbert sequencing ladder was used to determine the locations of the protected sites. Gel mobility shift assays. Oligonucleotides which corresponded to the DNase I-protected regions in footprint analysis were synthesized. The double-stranded oligonucleotides were end labeled with [a-32P]dATP and Klenow polymerase. The labeled oligonucleotides were extracted once with an equal volume of chloroform and separated from free nucleotide on a Sephadex G-50 column equilibrated with TEN50 (10 mM Tris HCl [pH 7.0], 1 mM EDTA, 50 mM NaCl). The binding reaction was performed in a final volume of 20 ml of 13 band shift buffer (25 mM N-2-hydroxyethylpiperazine-N9-2-ethanesulfonic acid [HEPES; pH 7.6], 5 mM MgCl2, 34 mM KCl) containing 1 ng of labeled probe, 1 mg of poly(dI-dC) (Sigma), 10 mg of BSA (Gibco/BRL), and 1 mg of nuclear proteins. After 15 min on ice, the mixture was loaded on a 6% nondenaturing polyacrylamide gel in 0.253 Tris-borate-EDTA and electrophoresed for 2 to 2.5 h at 150 V at 48C. The gel was dried and autoradiographed. For competition gel mobility experiments, a 10-, 50-, or 100-fold molar excess of the appropriate unlabeled competitor oligonucleotide was added to the binding reaction prior to the addition of nuclear proteins. For supershift assays, 1 ml of the appropriate antiserum was added 10 min after incubation of the labeled probe with nuclear proteins. The incubation period was extended for an additional 10 min on ice. The anti-NF-YA antibody was a monoclonal antibody, and the anti-NF-YB antibody was a polyclonal antibody; both were kind gifts from Christophe Benoist (Laboratoire de Ge´ne ´tique Mole ´culaire des Eucaryotes du CNRS et Unite´ 184 de l’INSERM, Strasbourg, France). Recombinant C/EBPa and DBP were from the laboratory stock of Chris Mueller (Department of Biochemistry, Queen’s University). Transient transfection of HepG2 cells. HepG2 cells were grown at 378C as a monolayer on 100-mm-diameter plates supplemented with 5% CO2 in minimal essential medium (Gibco) containing 10% fetal bovine serum, 50 U of penicillin G per ml, and 50 mg of streptomycin per ml. Confluent plates of HepG2 cells were split 1:4 the day before transfection. DNA was transfected into the HepG2 cells by the calcium phosphate coprecipitation method as described by Minty and Kedes (33). For each plate, a total of 20 mg of DNA consisting of 5 mg of luciferase reporter plasmid, 3.6 mg of b-galactosidase expression vector pCH110 (Pharmacia), and 1 mg of either C/EBPa (pMSV-C/EBP), C/EBPb (pMEX-

1937

CRP2), or DBP (pCMV-DBP) expression plasmid was transfected. The total DNA was adjusted to 20 mg with calf thymus DNA. The cells were harvested 48 h after removal of the calcium phosphate coprecipitate, and the cell extracts were prepared by using 13 cell lysis buffer (Fisher/Promega) and centrifugation to remove cell debris. Fifty microliters of the cell lysate was added to the luciferin substrate buffer. Luciferase activity was measured in a Bio Orbit 1250 Luminometer. Values for luciferase activity were corrected for each transfection with the corresponding result for b-galactosidase activity.

RESULTS PCR and DNA sequencing of the factor VIII promoter. Following convention, the transcriptional start site is referred to as nucleotide (nt) 11. A region of 1,150 bp of the factor VIII gene corresponding to nt 1148 to 21002 was amplified by PCR from genomic DNA, cloned, and sequenced. The sequencing data confirmed the presence of an extra cytosine nucleotide (underlined) at 116 in the untranslated region (GCACATC instead of GCAATC) missing from the original published sequence (14, 53). This sequence change has been documented previously by one other study (57). Twelve DNase I-protected sites occur in the 5* flanking region of the factor VIII promoter. To examine the extent and location of nuclear protein interactions, DNase I footprint analysis of the factor VIII promoter was performed with rat liver nuclear extracts. The results show that with the factor VIII proximal promoter fragment (nt 2277 to 1148) shown in Fig. 1a and b, three regions of protein-DNA interactions occur in the 59 UTR (sites 1 to 3). Site 1 is located between nt 1118 and 1148 and has 10-of-13-nt homology to HNF-1 consensus sequence (5). Site 2 is located between nt 168 and 185 and has some similarity to the consensus sequence 59-ATTNNG CAAT-39 of the C/EBP family of transcription factors (30). A large protected region which extends from nt 123 to the GATAAA sequence at nt 230 was revealed (Fig. 1b). We hypothesize that this is a composite of two contiguous sites, given the relatively large size of the protected area. Evidence for delineation of the two sites is provided by footprint analysis of the 39-labeled coding strand, which shows a separation between site 3 (11 to 123) and site 4 (21 to 230) (data not shown). Factor VIII promoter sites (F8 sites) 3 and 4 exhibit sequence similarity to the C/EBP consensus sequence. The fifth site is located between nt 245 and 262, and site 6 is located between nt 279 and 2109. The factors binding to these sites are still unknown. The remaining six sites are found in the distal region of the factor VIII promoter. The seventh site (nt 2401 to 2418) shown in Fig. 1c lies an additional 300 nt upstream of site 6 and also shows sequence similarity to the C/EBP consensus sequence. The five remaining sites are clustered in a 250-bp region which terminates ;1 kb from the transcriptional start site: site 8 (nt 2705 to 2722) and site 9 (nt 2800 to 2822) (Fig. 1d), site 10 (nt 2852 to 2869) and site 11 (nt 2889 to 2913) (Fig. 1e), and site 12 (nt 2946 to 2963) (Fig. 1f). These five sites have no apparent homology to known consensus sequences for transcription factor binding sites. The locations of these binding sites in the 59 flanking sequence of the factor VIII gene are summarized in Fig. 1g. The liver-enriched members of the b/ZIP family of transcription factors, C/EBP and DBP, have common binding sites in the factor VIII promoter. To study the proteins binding to each of these sites, double-stranded oligonucleotides which correspond to the 12 sites identified by DNase I footprint analysis were synthesized. Protein binding was assessed through gel mobility shift assays using rat liver nuclear extracts as a source of liver-enriched transcription factors. The banding patterns shown in Fig. 2a to c by F8 sites 2, 3, 4, 7, 9, and 10

1938

MCGLYNN ET AL.

MOL. CELL. BIOL.

FIG. 1. DNase I footprinting of the promoter region of the human factor VIII gene, using rat liver nuclear extracts. DNase I protection analysis with a 425-bp factor VIII proximal promoter fragment (2277 to 1148) is shown in panels a and b, with an extended electrophoresis shown in panel a. The noncoding strand is shown here. The arrow in panel b indicates the location of the transcriptional start site, with the direction of transcription toward the top. The number above each lane indicates the amount of liver nuclear extract used in the protection assay. The boundaries of the protected regions are marked by brackets indicating the relative position of each site with respect to the transcription initiation site. DNase I footprint analyses of a 724-bp (21002 to 2278) and a 303-bp (2845 to 2542) factor VIII distal promoter fragment are shown in panels c and d, respectively. The coding strands are shown in these instances. DNase I footprint analysis of a 156-bp distal fragment (21002 to 2846) is shown in panel e, with a shorter electrophoresis shown in panel f. The noncoding strand is shown here. The sequences corresponding to the DNase I footprints are shown in panel g (underlined and in boldface). Double-stranded oligonucleotides corresponding to these sites were synthesized for gel mobility shift assays.

show size heterogeneity of the protein-DNA complexes. These banding patterns, in turn, show a distinct similarity to that of the D site of the albumin promoter that has been reported to bind C/EBPa, C/EBPb, and DBP (29, 34). The differential

migration of the protein-DNA complexes observed is due to the b/ZIP structure of the C/EBP isoforms, which allows them to form homo- and heterodimers. We were able to confirm that C/EBPa can bind as a homodimer to at least six of the factor

VOL. 16, 1996


1939

FIG. 2. Gel mobility shift analysis of factor VIII oligonucleotides. (a to c) Rat liver nuclear extracts were incubated with oligonucleotide probes corresponding to the DNase I-protected areas of the human factor VIII promoter (sites 1 to 12). The DNA-protein complexes (BOUND) and the free oligonucleotides (FREE) are marked. The double-stranded oligonucleotides were also analyzed for the ability to bind recombinant C/EBPa (rC/EBP) (d) and recombinant DBP (rDBP) (e). An oligonucleotide corresponding to the D site of the albumin promoter is shown in lanes D in panels d and e.

VIII sites, as a single protein-DNA complex is obtained with a recombinant C/EBP protein in lanes 3, 4, 6, 7, 9, and 10 of Fig. 2d. Furthermore, in competition studies, the D site of the albumin promoter (a well-characterized C/EBP binding site) was able to compete for the smear of complexes at site 2 (data not shown), suggesting that additional factors, including C/EBP, may also bind to this site. Supershift analysis with antibodies against various members of the C/EBP family confirms that both the C/EBPa and C/EBPb isoforms bind to sites 3, 4, and 7 (data not shown; supershift studies were not performed on sites 6, 9, and 10). In addition, C/EBP and DBP have been shown to bind to common sites in the factor VIII promoter (Fig. 2e), a finding that has been documented previously for a number of other promoters of liver-specific genes (17, 23, 29, 34, 40, 52). Transactivation of the factor VIII promoter by members of the C/EBP family and the PAR family. Since hepatocytes represent one of the cellular sites of synthesis of factor VIII, the human hepatocellular carcinoma cell line HepG2 was chosen as the cell line in which transcription of this gene might most appropriately be studied. Five constructs with progressive 59 deletions from the 1.2-kb factor VIII promoter sequence were assessed for the ability to direct transcription of a luciferase reporter gene in HepG2 cells. All of the constructs include 148 nt of the 59 UTR of factor VIII. All transfection results have been reported with respect to the activity of the full-length construct. Analysis of a series of deletion constructs (Fig. 3a) indicates that removal of sequences distal to nt 244 results in a significant but small increase in promoter activity (at most ;2-fold). The construct containing sites 1 to 4 (pFVIII(244/1148)-Luc) maintains the full activity of the 1.2-kb promoter sequence, indicating that these four sites play a major role in regulating transcription of this locus. The presence of multiple C/EBP and DBP sites in the prox-

imal promoter indicates that members of the C/EBP and PAR domain family of transcription factors likely play a role in the activation of this gene. To test this hypothesis, HepG2 cells (which express only ;10% of the normal levels of C/EBPa and DBP [13, 35]) were cotransfected with expression plasmids for C/EBPa, C/EBPb, and DBP (Fig. 3b). In these studies, cotransfection of the 59 deletion constructs with both C/EBPa and DBP resulted in strong transactivation of the factor VIII promoter with an overall additive effect. Similarly, cotransfection of the two C/EBP expression vectors, either alone or in combination, resulted in strong transactivation of the factor VIII promoter, particularly when both C/EBP isoforms were present. In addition, these experiments showed a significant enhancement of transactivation with the combined cotransfection of C/EBPa/DBP (;3-fold, P , 0.005) or C/EBPa/C/EBPb (;10-fold, P , 0.001) when the region between nt 2740 and 21002 was deleted, suggesting the presence of a repressor element within this region. The liver-enriched transcription factor HNF-1 and a ubiquitous factor, NF-Y, bind to distinct sites in the minimal factor VIII promoter. Upon scanning of the protein binding sites in the factor VIII promoter for similarity to consensus sequences for transcription factors implicated in the regulation of other liver-specific genes, it became apparent that a sequence within site 1 had significant homology to the HNF-1 consensus sequence 59-GTTAATNATTAAC-39 (5). HNF-1 is a member of the POU domain family of transcription factors. The two isoforms of HNF-1, HNF-1a and HNF-1b, are able to form homo- and heterodimers with each other and other related proteins (32). Competition studies with an HNF-1 oligonucleotide (Fig. 4a) demonstrated that a 10-fold molar excess of the HNF-1 competitor prevents the formation of complex 1, whereas a 100-fold excess of the competitor is required to compete for complex 2. The nature of the two complexes is not known since the expression patterns of HNF-1a and HNF-1b

1940

MCGLYNN ET AL.

MOL. CELL. BIOL.

VOL. 16, 1996


1941

FIG. 3. Effects of 59 deletions on the transcriptional activity of the factor VIII promoter. (a) Promoter activity was determined by the luciferase activity of cell lysates from HepG2 cells transiently transfected with the pFVIII-Luc deletion constructs. LUC is the luciferase plasmid containing no promoter sequence. pFVIII(21002)-Luc is the reference plasmid containing the full 1.2-kb factor VIII promoter sequence. Transcriptional activity obtained with this construct is assigned a value of 1, with all other values shown relative to this. The results shown represent the means 6 1 standard deviation for four separate experiments. Luciferase activity was normalized for transfection efficiency by cotransfection with a b-galactosidase expression plasmid. The numbers below the graph indicate the extents of the 59 deletions. (b) Various deletions (del) of the pFVIII-Luc constructs (5 mg) were cotransfected with 1 mg of either pMSV-C/EBP (C/EBPa), pMEX-CRP2 (C/EBPb), or pCMV-DBP (DBP) alone or in combination. The means 6 1 standard deviation for three separate experiments are shown. Cotransfection of the promoterless luciferase vector with C/EBPa, C/EBPb, and DBP shows negligible activity. A schematic diagram of the factor VIII promoter showing the locations of the 12 DNase I-protected sites is presented below the graph.

differ. Although HNF-1a and HNF-1b mRNA has been found in the liver, the HNF-1b protein is absent (32). We were also interested in identifying the distinct proteinDNA complex above the smear in band shift analysis of F8 site 4. Analysis of this sequence showed that it contained an element which was similar to the C site of the albumin promoter.

FIG. 4. Competition gel shift analysis and supershift analysis of proteins binding to F8 sites 1 and 4. A double-stranded oligonucleotide corresponding to F8 site 1 was used as a probe in panel a. A competition assay was carried out in the presence of 10-, 50-, and 100-fold molar excesses of unlabeled HNF-1 oligonucleotide (52) prior to the addition of 1 mg of rat liver nuclear extracts. The DNA-protein complexes (BOUND) and the free oligonucleotides (FREE) are marked. The arrows on the right indicate the complexes (1 and 2) which disappear from left to right. In panel b, a double-stranded oligonucleotide corresponding to F8 site 4 was used as a probe. A competition assay was carried out in the presence of 10-, 50-, and 100-fold molar excesses of unlabeled albumin (Alb) site C (15). The arrow on the right side indicates the complex which disappears from left to right. In panel c, a supershift assay was carried out with rat liver nuclear extracts and either wild-type F8 site 4 (F8 site 4) or mutant F8 site 4 (F8 mut S4). No antibodies (0), a rabbit polyclonal antibody to the B subunit of NF-Y (aNF-YBa), a monoclonal antibody to the A subunit of NF-Y(a-NF-YA), or an affinity-purified polyclonal antibody to the B subunit of NF-Y(a-NF-YBb) was added to normal band shift assays. The DNA-protein-antibody complexes (Supershift) and the DNA-protein complexes (BOUND) are marked.

It has been previously reported that the CCAAT-binding factor NF-Y binds to this site (25). The C site of the albumin promoter contains a CCAAT sequence preceded by GGAA, which is virtually identical to the site 4 sequence 59-GGAAGCAAT-39. Previous studies of the albumin promoter have determined that NF-Y is the major transcription factor binding to site C, whereas C/EBP binds only weakly to this site (60). When we used site C of the albumin promoter as a competitor (Fig. 4b), the prominent gel-retarded band at site 2 disappeared, as did the heterogeneous complexes which have been previously attributed to be C/EBP binding complexes. Both polyclonal and monoclonal antibodies to rat NF-Y supershifted the distinct F8 site 4 complex (Fig. 4c). With the antiNF-YB polyclonal antibodies, the supershifted complex has been retarded in the wells of the gel or is more apparent by the lack of formation of the complex which is present in a normal band shift. With the anti-NF-YA monoclonal antibody, the supershifted complex appears as a doublet; this may be due to the fact that in rats, two isoforms of the NF-Y A subunit, produced by the initiation of transcription at different start sites, have been found (56). Mutations which disrupt HNF-1 binding, but not C/EBP binding, to the factor VIII minimal promoter abolish transcriptional activity. We extended our study of the transcription factors binding to sites 1 to 4 by modifying important nucleotides within the consensus sequences for HNF-1 and C/EBP by site-directed mutagenesis. The nucleotide substitutions which were made within F8 site 1 were identical to those which had been previously reported to inhibit HNF-1 binding in studies with the a2,6-sialyltransferase gene promoter (52). It is well known that C/EBP binds to a notoriously loose consensus (30). When we compared many of the reported C/EBP sequences, we noticed that they have in common a highly conserved adenine residue (underlined) within at least one of the GCAAT half sites. We thus decided to substitute the adenine residue at nucleotide position 117 within site 3 and at position 215 within site 4 to thymine (Fig. 5a). It was hoped that this transversion mutation would disrupt the binding of C/EBP to both site 3 and site 4. Additional evidence which supported our choice of mutation was that one of the naturally occurring mutations at site 1 of the factor IX promoter, which gives rise to the hemophilia B Leyden phenotype, was a transition at this residue of an A to G (nt 113), which disrupts C/EBP binding (1). In band shift studies, we have confirmed that the 117 mutation disrupts the binding of C/EBP, whether the source is rat liver nuclear extract or recombinant C/EBPa (Fig. 5b). The binding of DBP to this site is also reduced. The 215 mutation prevents the binding of C/EBP and DBP but not NF-Y (Fig. 5b). To confirm that NF-Y still binds to the site 4 mutant oligonucleotide, polyclonal and monoclonal antibodies to the A and B subunits of NF-Y were shown to supershift a proteinDNA complex which has previously been attributed as NF-Y binding to the wild-type site 4 oligonucleotide (Fig. 4c). The site 2 C/EBP binding mutant possesses four transversion mutations at positions 175, 179, 181, and 182. As expected, this

1942

MCGLYNN ET AL.

MOL. CELL. BIOL.

FIG. 5. Effects of mutations on the activity of the factor VIII minimal promoter. (a) Nucleotide sequence of the minimal promoter pFVIII(244)-Luc construct. The arrow indicates the transcriptional start site as well as the boundary which separates site 3 from site 4. Asterisks indicate the nucleotides modified by site-directed mutagenesis. The mutations in site 3 and site 4 are transversion mutations, A to T at nucleotides 215 and 117 (within the GCAAT sequence). The nucleotide changes in site 1 consist of transition and transversion mutations which have been previously described to disrupt HNF-1 binding (52). (b) Double-stranded oligonucleotides corresponding to the wild-type F8 site 3 (WT), mutant F8 site 3 (M3), wild-type F8 site 4 (WT), and mutant F8 site 4 (M4) were analyzed in gel mobility shift assays for the ability to bind rat liver nuclear extracts (Liver), recombinant C/EBPa (rC/EBP), and recombinant DBP (rDBP). These mutations were substituted in the factor VIII minimal promoter. The schematic diagrams in panel c illustrate the effects of the various promoter mutations on the binding of the transcription factors C/EBP, DBP, NF-Y, and HNF-1. HepG2 cells were transfected with 5 mg of a pFVIII-Luc reporter plasmid under the control of the full-length [pFVIII(21002/1148] promoter, wild-type minimal promoter [pFVIII(244/1148], and mutant forms of the factor VIII promoter sequence. The pFVIII(21002/1148)-Luc construct is the reference plasmid. The transcriptional activity obtained with this construct has been assigned a value of 1, with all other values shown relative to this. The first column represents the basal activity 6 1 standard deviation from four separate experiments with the factor VIII promoter constructs alone. The second column represents the fold increase in activity 6 1 standard deviation with cotransfection of C/EBPa and C/EBPb.

VOL. 16, 1996


1943

FIG. 6. Schematic diagram of the factor VIII promoter illustrating the 12 cis-acting DNA sequences and their coordinates within 1.2 kb of the 59 flanking sequence of the factor VIII gene. Nucleotide numbering is relative to the transcriptional start site at position 11. Some of the trans-acting factors which bind to these sites are indicated above the sites. Site 1 (1118 to 1148) binds HNF-1; site 2 (168 to 185) binds C/EBP (heterodimeric forms); site 3 (11 to 123) binds C/EBP and DBP; site 4 (230 to 21) binds C/EBP, NF-Y, and DBP; site 7 (2418 to 2401) binds C/EBP and DBP; site 9 (2822 to 2800) binds C/EBP and DBP; site 10 (2869 to 2852) binds C/EBP; site 5 (262 to 245), site 6 (2109 to 279), site 8 (2722 to 2705), site 11 (2913 to 2889), and site 12 (2963 to 2946) bind proteins that remain uncharacterized.

sequence has been shown to disrupt C/EBP binding in a supershift assay using liver nuclear extracts (data not shown). Figure 5c summarizes the functional effects of the factor VIII promoter mutations on the binding of HNF-1, C/EBP, DBP, and NF-Y to sites 1 to 4. These studies include cotransfection of both C/EBPa and C/EBPb expression plasmids but do not include cotransfection of HNF-1, since HepG2 cells have been shown to express normal levels of HNF-1 (1). Disruption of C/EBP binding to sites 2 to 4 did not result in any significant change in basal promoter activity (Fig. 5c). However, the transactivation potential of C/EBPa and C/EBPb was reduced by 40 to 60% (P , 0.025) for all of these mutant promoters. Furthermore, transcription initiation appears to be unaffected when the GATAAA sequence is removed [pFVIII(25/1148)mutS3-Luc], suggesting that an initiator (Inr) element at the transcriptional start site may be important in this construct. Mutational analysis of nt 22 and 21 in the putative Inr element [pFVIII(25/1148)mutInr/S3-Luc] does not interfere with basal promoter activity; however, the ability of C/EBPa and C/EBPb to transactivate the mutant Inr construct is significantly reduced by an additional ;60% (P , 0.005). It is only upon mutation of the HNF-1 binding site [pFVIII(25/1148)mutS1/S3-Luc] that basal promoter activity is significantly reduced by ;75% (P , 0.005) and that the transactivation potential of C/EBPa and C/EBPb is decreased by 95% (P , 0.005). These results suggest that HNF-1 binding to site 1 plays a critical role in the context of the activity of the minimal factor VIII promoter sequence. However, our results with a promoter construct including only sites 3 to 12 [pFVIII(21002/125)-Luc] also show that in the absence of HNF-1 binding, the multiple upstream sites maintain the activity attainable by the full-length factor VIII promoter. This finding suggests that the multiple upstream sites may play a compensatory role should HNF-1 binding be prohibited. DISCUSSION The 1.2-kb 59 flanking region of the factor VIII gene was cloned in 1984 (14); in the same study, the transcriptional start site was mapped 170 bp upstream of the initiation methionine codon. In our initial experiments, we have shown that the regulation of factor VIII gene expression involves a complex array of transcription factors that mediate their function through both positive and negative cis-acting regulatory elements. We have identified 12 cis-acting binding sites in the factor VIII promoter by DNase I protection analysis (Fig. 6). There is a cluster of five distal elements within 250 bp that extends to ;1 kb from the transcriptional start site. Within this region between nt 2740 and 21002, there appears to be a negative regulatory element since the activity of the factor VIII

promoter is increased upon removal of this region. Three proximal elements lie within 100 bp 59 of the transcriptional start site, while an additional three elements lie downstream of the CAP site. The structure of the factor VIII promoter suggests a number of parallels to the regulatory elements of the factor IX gene (6, 40, 46). In both cases, sites for DNA-binding proteins overlap the transcriptional start site. It has been postulated that in the factor IX promoter, these sites may be important in the stabilization of the preinitiation complex, since each of these sites in isolation has been reported to mediate only weak direct transcriptional activation of factor IX (41). In addition, the sites immediately downstream of the CAP site bind isoforms of C/EBP in both promoters. A naturally occurring mutation at position 113 of the factor IX gene results in the reduced binding of C/EBP to this site and produces a hemophilia B Leyden phenotype (6, 41). C/EBPa binding to sites 1, 4, and 5 of the factor IX promoter mediates transcriptional activation, and the coexpression of C/EBPa with DBP is thought to play a role in the phenotypic recovery of patients with the hemophilia B Leyden phenotype following the onset of puberty (41). Similarly, there are at least seven C/EBP binding sites in the factor VIII promoter, five of which also bind DBP. Therefore, both of these b/ZIP proteins have been shown to be important factors that mediate the transcriptional regulation of these genes in vitro. An additional common feature of these two genes concerns the occurrence of several critical protein binding sites within their 59 UTRs. In considering all of these similarities, one can suggest that factor VIII and factor IX (and perhaps other coagulation factors) have evolved some common mechanisms for their transcriptional regulation, thus facilitating their coordinate expression. A fundamental difference between the factor VIII and factor IX genes is their patterns of tissue expression. Factor IX is expressed only in the liver, whereas factor VIII is expressed at moderate levels in the spleen, kidney, and lymph nodes, in addition to the liver (58). Although there are a number of liver-specific factors which regulate factor VIII expression, other ubiquitous and tissue-specific factors likely play a role in its expression in tissues other than the liver. The results from our studies suggest that HNF-1 may be one such factor that appears to be involved in determining the tissue-specific expression of factor VIII, while the relative abundance of the two isoforms of this protein, HNF-1a and HNF-1b, may regulate the magnitude of this response. HNF-1a and HNF-1b have a high degree of homology in their dimerization domains and homeodomains, which allows them to bind to similar sequences, but HNF-1b lacks an activation domain (32). HNF-1a is expressed in the same tissues as factor VIII (liver, spleen, kidney, and intestine), whereas HNF-1b has a more

1944

MCGLYNN ET AL.

ubiquitous expression pattern. In addition, mammalian tissues contain different ratios of HNF-1a and HNF-1b, and although HNF-1b does not block the transcriptional activity of HNF-1a (32), exclusive expression of HNF-1b has been implicated in the repression of a number of hepatocyte-specific genes (32). In light of this evidence, we suggest that factor VIII is likely to be expressed in the same cells in which HNF-1a is synthesized, with its expression being enhanced in the liver because of the apparent absence of HNF-1b protein in this tissue (32) and the presence of other liver-enriched factors that bind to this promoter. A recent publication by Figueiredo and Brownlee (11) presents evidence indicating that the ubiquitous transcription factor NF-kB binds adjacent to the downstream C/EBP site (site 2). Our results show that a mutation within the HNF-1 binding site significantly reduces the C/EBP responsiveness of the minimal promoter sequence. This finding suggests the possibility of a functional interaction between C/EBP and HNF-1. Furthermore, there is previously published evidence of a physical and functional association between C/EBP and NF-kB (44, 51), suggesting that further study is required to elucidate the interactive role of these three factors in the 59 UTR of the factor VIII gene. The presence of many cis-acting sequences establishes binding of transcription factors at a location where they can interact with the transcription initiation complex that must assemble in order for the polymerase to initiate transcription. As an alternative to the TATA box, the GATAAA sequence at position 230 of the factor VIII gene is thought to be important for assembly of the preinitiation complex. We have shown that deletion of the alternative TATA sequence in factor VIII does not disrupt transcription in vitro [Fig. 5c, promoter construct pFVIII(25/1148)mutS3-Luc]. Similarly, Figueiredo and Brownlee have shown that upon mutation analysis of the GATAAA sequence, no significant changes in promoter activity are found (11). There is a putative Inr element between nt 24 and 14 in the factor VIII gene promoter which has sequence homology (7 of 8 nt) to the terminal deoxynucleotidyltransferase Inr element (59-PyAPyTCPyPyPy-39) (50), suggesting that this sequence may play a role in localizing the preinitiation complex at this locus. However, in this report, we have demonstrated that mutations within the Inr element do not result in a complete down-regulation of transcription but may reduce the frequency of initiation or allow for a variety of transcription initiation sites (48). A number of examples in which mutations in the protein binding sites of promoters have had clinical consequences have been documented. It has been found that promoter mutations in factor IX (4, 7, 12, 19, 28, 39, 45, 47, 49), protein C (2), haptoglobin (16), and b-globin (10, 31, 36–38) lead to disruption of the binding of transcription factors, with a resulting clinically significant reduction of gene expression. Such a lesion might be expected to affect the level of expression of the factor VIII protein and give rise to a clinical phenotype. In 1982, when the hemophilia B Leyden phenotype was first described, the additional discovery of variable factor VIII expression in rare patients was thought to represent a factor VIII form of the Leyden phenotype (4). These patients exhibited a mild form of hemophilia A with levels of factor VIII between 10 and 20% of normal, which gradually increased to 40 to 160% of normal (27, 43, 57). In a previous study of 127 unrelated patients with hemophilia A, no changes were reported in the promoter sequence extending 200 nt 59 of the transcriptional start site (18). In addition, the full 1.2-kb factor VIII promoter has been sequenced in three other unrelated patients with a variable hemophilia A phenotype (26, 57). However, no

MOL. CELL. BIOL.

regulatory mutation causing hemophilia A has been found. In our studies, we generated mutations at critical residues for HNF-1 and C/EBP binding in what was thought to be a minimal promoter. Our results suggest that even though HNF-1 binding is critical in maintaining the activity of the factor VIII promoter in the absence of upstream elements, it is not necessary in the native promoter context, as the multiple upstream sites seem to compensate for its loss. We postulate from these results that our inability to detect the down-regulation of factor VIII promoter activity in vitro parallels our inability to detect a disease-causing mutation which would give rise to a clinical phenotype. The role of the multiple upstream sites can be hypothesized as a safeguard or backup regulatory system in the instance of a potentially harmful mutation occurring in this very narrow region of the downstream HNF-1 binding site. ACKNOWLEDGMENTS We acknowledge David Picketts for helpful discussions concerning this work. This study was supported by a grant from Bayer Canada and the Canadian Red Cross Society. L.K.M. and M.B. are recipients of Canadian MRC studentships. C.R.M. is a Research Scientist of the Canadian NCI, and D.L. is a Career Investigator of the Heart and Stroke Foundation of Ontario. REFERENCES 1. Bach, I., M. G. Mattei, S. Cereghini, and M. Yaniv. 1991. Two members of an HNF-1 homeoprotein family are expressed in human liver. Nucleic Acids Res. 19:3553–3559. 2. Berg, L.-P., D. White, V. V. Kakkar, and D. N. Cooper. 1993. Functional characterization of a promoter mutation in the protein C gene. Thromb. Haemostasis 69:999a. 3. Bontempo, F. A., J. H. Lewis, T. J. Gorenc, J. A. Spero, M. V. Ragni, J. P. Scott, and T. E. Starzl. 1987. Liver transplantation in hemophilia A. Blood 69:1721–1724. 4. Briet, E., R. M. Bertina, N. H. van Tilburg, and J. J. Veltkamp. 1982. Hemophilia B Leyden: a sex-linked hereditary disorder that improves after puberty. N. Engl. J. Med. 306:788–790. 5. Courtois, G., S. Baumhueter, and G. R. Crabtree. 1988. Purified hepatocyte nuclear factor 1 interacts with a family of hepatocyte specific promoters. Proc. Natl. Acad. Sci. USA 85:7937–7941. 6. Crossley, M., and G. G. Brownlee. 1990. Disruption of a C/EBP binding site in the factor IX promoter is associated with hemophilia B. Nature (London) 345:444–446. 7. Crossley, M., P. R. Winship, D. E. G. Austen, C. R. Rizza, and G. G. Brownlee. 1990. A less severe form of haemophilia B Leyden. Nucleic Acids Res. 18:4633. 8. Dorner, A. J., D. G. Bole, and R. J. Kaufman. 1987. The relationship of N-linked glycosylation and heavy chain-binding protein association with the secretion of glycoproteins. J. Cell. Biol. 105:2665–2674. 9. Dorner, A. J., L. C. Wasley, and R. J. Kaufman. 1990. Protein dissociation from GRP78 and secretion are blocked by depletion of cellular ATP levels. Proc. Natl. Acad. Sci. USA 87:7429–7432. 10. Elion, J., P. E. Berg, C. Lapoumeroulie, G. Trabuchet, M. Mettelman, R. Krishnamoorthy, A. W. Schechter, and D. Labie. 1992. DNA sequence variation in a negative control region 59 to the b-globin gene correlates with the phenotypic expression of the bS mutation. Blood 79:787–792. 11. Figueiredo, M. S., and G. G. Brownlee. 1995. cis-acting elements and transcription factors involved in the promoter activity of the human factor VIII gene. J. Biol. Chem. 270:11828–11838. 12. Freedenburg, D. L., and B. Black. 1991. Altered developmental control of the factor IX gene: a new T to A mutation at position 16 of the factor IX gene resulting in hemophilia B Leyden. Thromb. Haemostasis 65:964a. 13. Friedman, A. D., W. H. Landschultz, and S. L. McKnight. 1989. CCAAT/ enhancer binding protein activates the promoter of the serum albumin gene in cultured hepatoma cells. Genes Dev. 3:1314–1322. 14. Gitschier, J., W. I. Wood, T. M. Goralka, K. L. Wion, E. Y. Chen, D. H. Eaton, G. A. Vehar, D. J. Capon, and R. M. Lawn. 1984. Characterization of the human factor VIII gene. Nature (London) 312:326–330. 15. Gorski, K., M. Carneiro, and U. Schibler. 1986. Tissue-specific in vitro transcription from the mouse albumin promoter. Cell 47:767–776. 16. Grant, D. J., and N. Maeda. 1993. A base substitution in the promoter associated with the human haptoglobin 2-1 modified phenotype decreases transcriptional activity and responsiveness to interleukin-6 in human hepatoma cells. Am. J. Hum. Genet. 52:974–980. 17. Gregori, C., A. Kahn, and A.-L. Pichard. 1993. Competition between tran-

VOL. 16, 1996

18.

19.

20. 21.

22. 23. 24. 25. 26. 27. 28.

29. 30. 31.

32.

33. 34. 35. 36.

37.

38. 39. 40.


scription factors HNF-1 and HNF-3, and alternative cell-specific activation by DBP and C/EBP contribute to the regulation of the liver-specific aldolase B promoter. Nucleic Acids Res. 21:897–903. Higuchi, M., C. Wong, L. Kochhan, K. Olek, S. Aronis, C. K. Kasper, H. H. Kazazian, Jr., and S. E. Antonarakis. 1990. Characterization of mutations in the factor VIII gene by direct sequencing of amplified genomic DNA. Genomics 6:65–71. Hirosawa, S., J. B. Fahner, J.-P. Salier, C.-T. Wu, E. W. Lovrien, and K. Kurachi. 1990. Structural and functional basis of the developmental regulation of human coagulation factor IX gene: factor IX Leyden. Proc. Natl. Acad. Sci. USA 87:4421–4425. Hoyer, L. W. 1981. The FVIII complex: structure and function. Blood 58:1– 13. Kaufman, R. J., L. C. Wasley, M. V. Davies, R. J. Wise, D. I. Israel, and A. J. Dorner. 1989. Effect of von Willebrand factor coexpression on the synthesis and secretion of factor VIII in Chinese hamster ovary cells. Mol. Cell. Biol. 9:1233–1242. Kaufman, R. J., L. C. Wasley, and A. J. Dorner. 1988. Synthesis, processing, and secretion of recombinant human factor VIII expressed in mammalian cells. J. Biol. Chem. 263:6352–6362. Lee, Y.-H., J. A. Alberta, F. J. Gonzalez, and D. J. Waxman. 1994. Multiple, functional DBP sites on the promoter of the cholesterol 7a-hydroxylase P450 gene CYP7. J. Biol. Chem. 269:14681–14689. Lewis, J. H., F. A. Bontempo, J. A. Sperio, M. V. Ragni, and T. E. Starzl. 1985. Liver transplantation in a hemophiliac. N. Engl. J. Med. 312:1189– 1190. Lichtsteiner, S., J. Wuarin, and U. Schibler. 1987. The interplay of DNA binding proteins on the promoter of the mouse albumin gene. Cell 51:963– 973. Lillicrap, D., et al. Unpublished data. Ljung, R., and I. M. Nilsson. 1982. Hemophilia B Leyden and a similar variant of hemophilia A. N. Engl. J. Med. 307:897–898. Ludwig, M., R. Schwaab, A. Eigel, J. Horst, H. Egli, H. H. Brackmann, and K. Olek. 1989. Identification of a single nucleotide C to T transition and five different deletions in patients with severe hemophilia B. Am. J. Hum. Genet. 45:115–122. Maire, P., J. Wuarin, and U. Schibler. 1989. The role of cis-acting promoter elements in tissue-specific albumin gene expression. Science 244:343–345. McKnight, S. L. 1992. CCAAT/enhancer binding protein, p. 771–792. In S. L. McKnight (ed.), Transcriptional regulation. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. Meloni, A., M. C. Rosatelli, V. Faa, R. Sardu, L. Saba, S. Murru, G. V. Sciarratta, M. Baldi, N. Tannoia, A. Vittucci, and A. Cao. 1992. Promoter mutations producing mild beta-thalassaemia in the Italian population. Br. J. Haematol. 80:222–226. Mendel, D. B., L. P. Hansen, M. K. Graves, P. B. Conley, and G. R. Crabtree. 1991. HNF-1a and HNF-1b (vHNF-1) share dimerization and homeo domains, but not activation domains, and form heterodimers in vitro. Genes Dev. 5:1042–1056. Minty, A., and L. Kedes. 1986. Upstream regions of the human cardiac actin genes that modulate its transcription in muscle cells: presence of an evolutionarily conserved repeated motif. Mol. Cell. Biol. 6:2125–2136. Mueller, C. R. 1992. The down-regulation of albumin transcription during regeneration is due to the loss of HNF-1 and D-site transcription factors. DNA Cell Biol. 11:559–566. Mueller, C. R., et al. Unpublished data. Murru, S., M. C. Pischedda, A. Cao, M. C. Rosatelli, M. Pirastu, G. V. Sciarratta, L. Manca, D. Gablisai, and C. Toffoli. 1993. A promoter mutation in the b-globin gene (2101 C3T) has an age-related expression pattern. Blood 81:2818–2819. Orkin, S. H., S. E. Antonarakis, and H. H. Kazazian, Jr. 1984. Base substitution at position 288 in a beta thalassemia globin gene: further evidence for the role of distal promoter element ACACCC. J. Biol. Chem. 259:8679– 8681. Orkin, S. H., J. P. Sexton, T. C. Cheng, S. C. Goff, P. J. V. Giardina, J. I. Lee, and H. H. Kazazian, Jr. 1983. ATA box transcription mutation in betathalassemia. Nucleic Acids Res. 11:4727–4733. Picketts, D. J., C. D’Souza, P. J. Bridge, and D. Lillicrap. 1992. An A to T transversion at position 25 of the factor IX promoter results in hemophilia B. Genomics 12:161–163. Picketts, D. J., D. P. Lillicrap, and C. R. Mueller. 1993. Synergy between

41.

42. 43. 44. 45.

46. 47. 48. 49. 50. 51. 52. 53.

54.

55.

56.

57. 58. 59.

60.

1945

transcription factors DBP and C/EBP compensates for a hemophilia B Leyden factor IX mutation. Nat. Genet. 3:175–179. Picketts, D. J., C. R. Mueller, and D. P. Lillicrap. 1995. Transcriptional control of the factor IX gene: analysis of five cis-acting elements and the deleterious effects of naturally occurring hemophilia B Leyden mutations. Blood 84:2992–3000. Pittman, D. D., and R. J. Kaufman. 1989. Structure-function relationships of factor VIII elucidated through recombinant DNA technology. Thromb. Haemostasis 61:161–165. Ratnoff, O. D., and J. H. Lewis. 1975. Heckathorn’s disease: variable functional deficiency of antihemophilic factor (factor VIII). Blood 46:161–173. Ray, A., M. Hannink, and B. K. Ray. 1995. Concerted participation of NF-kappa B and C/EBP heteromer in lipopolysaccharide induction of serum amyloid A gene expression in liver. J. Biol. Chem. 270:7365–7374. Reijnen, M. J., K. Peerlinck, D. Maasdam, R. M. Bertina, and P. H. Reitsma. 1993. Hemophilia B Leyden: substitution of thymine for guanine at position 221 results in a disruption of a HNF-4 binding site in the factor IX promoter. Blood 82:151–158. Reijnen, M. J., F. M. Sladek, R. M. Bertina, and P. H. Reitsma. 1992. Disruption of a C/EBP binding site for hepatic nuclear factor results in hemophilia B Leyden. Proc. Natl. Acad. Sci. USA 89:6300–6303. Reitsma, P. H., R. M. Bertina, J. K. Ploos van Amstel, A. Riemans, and E. Briet. 1988. The putative factor IX gene promoter in hemophilia B Leyden. Blood 72:1074–1076. Roeder, R. G. 1991. The complexities of eukaryotic transcription initiation: regulation of pre-initiation complex assembly. Trends Biochem. Sci. 16:402– 408. Royle, G., N. S. Van De Water, E. Berry, P. A. Ockelford, and P. J. Browett. 1991. Haemophilia B Leyden arising de novo by point mutation in the putative factor IX promoter region. Br. J. Haematol. 77:191–194. Smale, S. T., and D. Baltimore. 1989. The ‘‘initiator’’ as a transcription control element. Cell 57:103–113. Stein, B., and A. S. Baldwin, Jr. 1993. Distinct mechanisms for regulation of the interleukin-8 gene involve synergism and cooperativity between C/EBP and NF-KB. Mol. Cell. Biol. 13:7191–7198. Svensson, E. C., P. B. Conley, and J. C. Paulson. 1992. Regulated expression of a2,6-sialyltransferase by the liver-enriched transcription factors HNF-1, DBP and LAP. J. Biol. Chem. 267:3466–3472. Toole, J. J., J. L. Knopf, J. M. Wozney, L. A. Sultzman, J. L. Buecker, D. D. Pittman, R. J. Kaufman, E. Brown, C. Shoemaker, E. C. Orr, G. W. Amphlett, W. B. Foster, M. L. Coe, G. J. Knutson, D. N. Fass, and R. M. Hewick. 1984. Molecular cloning of a cDNA encoding human antihaemophilic factor. Nature (London) 312:342–346. Tuddenham, E. G. D., R. Schwaab, J. Seehafer, D. S. Millar, J. Gitschier, M. Higuchi, S. Bidichandani, J. M. Connor, L. W. Hoyer, A. Yoshioka, I. R. Peake, K. Olek, H. H. Kazazian, J.-M. Lavergne, F. Giannelli, S. E. Antonarakis, and D. N. Cooper. 1994. Haemophilia A: database of nucleotide substitutions, deletions, insertions and rearrangements of the factor VIII gene, second edition. Nucleic Acids Res. 22:4851–4868. Vehar, G. A., B. Keyt, D. Eaton, H. Rodriguez, D. P. O’Brien, F. Rotblat, H. Oppermann, R. Keck, W. I. Wood, R. N. Harkins, E. G. D. Tuddenham, R. M. Lawn, and D. J. Capon. 1984. Structure of human factor VIII. Nature (London) 312:337–342. Vuorio, T., S. N. Maity, and B. de Crombrugghe. 1990. Purification and molecular cloning of the A chain of a rat heterodimeric CCAAT-binding protein: sequence identity with the yeast HAP3 transcription factor. J. Biol. Chem. 265:22480–22486. Wieland, K., L.-P. Berg, V. V. Kakkar, D. N. Cooper, and U. Martinowitz. 1990. Molecular genetic analysis of a novel form of haemophilia A characterized by the variable expression of factor VIII. Thromb. Res. 59:871–877. Wion, K. L., D. Kelly, J. A. Summerfield, E. G. D. Tuddenham, and R. M. Lawn. 1985. Distribution of factor VIII mRNA and antigen in human liver and other tissues. Nature (London) 317:726–728. Wood, W. I., D. J. Capon, C. C. Simonsen, D. L. Eaton, J. Gitschier, B. Keyt, P. H. Seeburg, D. H. Smith, P. Hollingshead, K. L. Wion, E. Delwart, E. G. D. Tuddenham, G. A. Vehar, and R. M. Lawn. 1984. Expression of active human factor VIII from recombinant DNA clones. Nature (London) 312:330–337. Wuarin, J., C. Mueller, and U. Schibler. 1990. A ubiquitous CCAAT factor is required for efficient in vitro transcription from the mouse albumin promoter. J. Mol. Biol. 214:865–874.