Characterization and Regulation of the 5 * -Flanking Region of the Murine Endothelial Protein C Receptor Gene*

The protein C pathway plays a critical role in the negative regulation of blood coagulation. The nucleotide sequence of the murine endothelial protein C receptor (mEPCR) gene was determined for 8.8 kilobase pairs of the genomic structure and 3.4 kilobase pairs of the 5 * -flanking region. RNase protection assay revealed six major transcription start sites clustered at 2 100 to 2 109 upstream of the translation initiation site. A series of 5 * -promoter deletion fragments were fused to a luciferase reporter gene and transiently transfected into bovine aortic endothelium. Deletion of the sequence from 2 220 to 2 180 dramatically reduced luciferase expression in bovine aortic endothelial cells. This region of the murine endothelial protein C receptor gene contains one AP4 site and one SP1 site. Mutations in the core sequence of the AP4 and SP1 sites impaired both nuclear protein binding and luciferase expression. These results suggest important roles for AP4 and SP1 in the constitutive expression of mEPCR. A thrombin response element (CCCACCCC) was found to mediate the induction of mEPCR by thrombin in cell culture. Transgenic mice were developed expressing green fluorescent protein driven by the 2 350 to 2 1 or 2 1080 to 2 1 promoter. Thrombin up-regulated mEPCR and the transgene in vivo. The protein C anticoagulant pathway plays a critical role in the negative regulation Templates for riboprobe synthesis prepared by subcloning a 350-bp fragment containing the mEPCR 5 9 - flanking sequence from 2 350 to 2 1 into the Kpn I/ Xho I sites of the pBluescript II KS( 1 ) plasmid to create pBKS-mP350. [ g - 32 P]CTP (800 Ci/mmol, Amersham Pharmacia Biotech) was used with the MAXIs-cript in vitro transcription kit (Ambion) to generate riboprobes from Eco RI-linearized pBKS-mP350. Full-length, radiolabeled cRNA was purified by electrophoresis through a 5% polyacrylamide gel, followed by elution for 4 h at 37 °C in 0.5 M ammonium acetate, 1 m M EDTA, 0.2% SDS. The labeling efficiency was determined by scintillation counting, and approximately 2 3 10 5 cpm of each labeled riboprobe were used per reaction. The RNA protection assay was performed with an RPA II kit (Ambion). Briefly, the labeled riboprobes were annealed to 10–20 m g of total RNA isolated from HA or EAhy926 cells with the TRIZOL Reagent (Life Technologies, Inc.). Following annealing, unprotected single-stranded RNA digested with a mixture of RNase A (2.5 units/ml) plus RNase T1 units/ml). Two control A second round of PCR was performed using only gene-specific primers. The PCR mimic was then purified and adjusted to a concentration of 100 amol/ m l. For quantitative PCR, 10-fold dilutions of mimic were added to a series of tubes containing a fixed concentration of first-strand cDNA and a pair of gene-specific primers. 25 cycles of PCR was performed under each of the following conditions: 94 °C for 1 min, 60 °C for 1 min, and 72 °C for 2 min. The gene-specific primers for murine EPCR and GFP were derived from the published sequences (18, 22). The size of the amplified products exactly matched those predicted. The identity of the amplified fragments was further confirmed by restriction digestions with appro-priate enzymes. Murine b -actin mimic and primers were provided by the manufacturer. The mEPCR and GFP primers were as follows:

The endothelial protein C receptor (EPCR), a new member of this pathway, was identified and the cDNA sequence obtained by expression cloning (4). EPCR shares significant homology with the CD1/major histocompatibility complex class 1 family of molecules. Mature EPCR is a type 1 transmembrane glycoprotein with 235 amino acids (4,5). It is a multidomain protein with an N-terminal signal sequence, ␣1 and ␣2 domains, a transmembrane domain, and a short cytoplasmic tail (4). EPCR binds protein C and activated protein C (APC) with similar affinity (K d Ϸ30 nM) (5). A soluble form of EPCR is found in plasma (6). Soluble EPCR inhibits the ability of APC to inactivate one of its natural substrates, factor Va (7). Monoclonal antibodies that block protein C binding to EPCR reduce protein C activation rates by the thrombin-TM complex on endothelium (8). When reconstituted into phosphatidylcholine vesicles containing TM, EPCR enhances protein C activation rates in an EPCR concentration-dependent fashion (9), suggesting that the rate of protein C activation would be governed by the cellular concentration of EPCR.
The expression of EPCR, as observed by immunohistochemistry, is restricted to endothelial cells (4,10,11). Like TM, EPCR is down-regulated by tumor necrosis factor ␣ (TNF␣) on endothelium in vitro, and the down-regulation of the two receptors follows the same time course (4,12). The protein C pathway appears to be important in regulating the inflammatory response to LPS/endotoxin, as evidenced by the observations that APC can prevent lethality in a baboon sepsis model (13) and appears to be effective in treatment of septic shock in humans (14). Gene deletion of protein C in mice results in prenatal consumptive coagulopathy and the infiltration of leukocytes into the liver (15). As a receptor of protein C and/or APC, EPCR might also be involved in the inflammatory response. Recent studies have demonstrated that blocking protein C binding to EPCR exacerbates the coagulation and inflammatory responses resulting from low level Escherichia coli infusions into baboons (16). Genetic defects in EPCR have now been identified in humans and may be associated with increased risks of arterial and venous thrombosis (17). Thus, an understanding of the molecular mechanisms that control transcription of the EPCR gene may help clarify its role in the regulation of hemostasis and thrombosis.
Here we describe the initial functional characterization of the murine EPCR promoter. Among other things, we identified a thrombin response element that potently up-regulates EPCR expression in response to thrombin both in endothelial cell culture and in vivo.

EXPERIMENTAL PROCEDURES
Cell Culture-Bovine aortic endothelial cells (BAEC) were prepared as described previously (18) and maintained in Eagle's minimal essential medium containing 10% fetal calf serum (FCS). EAhy926, a hybrid cell line created by fusion of HUVEC and A549 cells (19), was kindly * These work was supported in part by Grant PO1 HL 54804 from NHLBI, National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM MO), and a 5-kilobase pair HindIII-BamHI fragment containing the 5Ј-flanking region of the mEPCR gene was inserted into the pBluescript II(ϩ). The sequence of this fragment was determined on both strands using synthetic primers employing the dideoxy terminator method (20). The newly reported sequence from Ϫ1200 to Ϫ400 has been deposited in GenBank (accession no. AF224271) and scanned for significant homology. Except for the region between Ϫ945 and Ϫ820, no significant homology was detected. The region Ϫ945 to Ϫ820 showed significant homology with a large number (100) of genes including the murine major histocompatibility locus GenBank AF1000956. The accession numbers of the 20 most similar genes follow: AF139987, AF100956, AB024501, AF133093, AC002324, AC005259, AF121351, AF125313, Y09047, AF126378, AF092505, AF111103, AC005807, AB012608, X93470, AJ001307, AF021335, L47235, AC009287, and AF109719.
RNase Protection Assay-Templates for riboprobe synthesis were prepared by subcloning a 350-bp fragment containing the mEPCR 5Јflanking sequence from Ϫ350 to Ϫ1 into the KpnI/XhoI sites of the pBluescript II KS(ϩ) plasmid to create pBKS-mP350. [␥-32 P]CTP (800 Ci/mmol, Amersham Pharmacia Biotech) was used with the MAXIscript in vitro transcription kit (Ambion) to generate riboprobes from EcoRI-linearized pBKS-mP350. Full-length, radiolabeled cRNA was purified by electrophoresis through a 5% polyacrylamide gel, followed by elution for 4 h at 37°C in 0.5 M ammonium acetate, 1 mM EDTA, 0.2% SDS. The labeling efficiency was determined by scintillation counting, and approximately 2 ϫ 10 5 cpm of each labeled riboprobe were used per reaction. The RNA protection assay was performed with an RPA II kit (Ambion). Briefly, the labeled riboprobes were annealed to 10 -20 g of total RNA isolated from HA or EAhy926 cells with the TRIZOL Reagent (Life Technologies, Inc.). Following annealing, unprotected singlestranded RNA was digested with a mixture of RNase A (2.5 units/ml) plus RNase T1 (100 units/ml). Double-stranded RNA fragments were sized on 8% denaturing polyacrylamide gels. Two control reactions were performed in which each riboprobe was annealed to yeast RNA and incubated with the RNase mixture or the reaction buffer alone. A dideoxy DNA sequencing reaction of pBKS-mP350 was used as a size marker to map the transcription start sites of the endogenous mEPCR gene. After electrophoresis, the gel was vacuum-dried and exposed to x-ray film overnight with an intensifying screen at Ϫ80°C.
Construction of Chimeric Luciferase Expression Vectors-Promoterless basic and pGL3 control (SV40 promoter) plasmids were purchased from Promega, and the p␤Gal control was from CLONTECH. An 1120-bp fragment spanning Ϫ1120 to Ϫ1 bp upstream of the translation start site of the mEPCR gene was obtained by amplification of the HB5K construct using PCR. The resulting DNA fragment was cloned into the KpnI-XhoI site of the pGL3 basic vector, generating GL3-mP1120. Constructs pGL-mP550, pGL-mP350, pGL-mP280, pGL-mP220, pGL-mP180, pGL-mP160, and pGL-mP80 were generated by ligation of the respective DNA segments generated by PCR from pGL3-mP1120 using the same strategy. Plasmid pGL-mP3340 was constructed in two steps: ligation into the HindIII and SacI sites of pBluescript II KS(ϩ) of a 2260-bp fragment from Ϫ3340 to Ϫ1080 released from HB5K with HindIII and SacI, and insertion of this 2260-bp fragment released from pBKS with KpnI and SacI between the corresponding sites in the pGL-mP1120. To construct pGL-mP350⌬280 -160, oligonucleotides 5Ј-GCCAAGCCCTTCTCCAGACAGATCCAA-3Ј and 5Ј-TTGGATCTGTCTGGAGAAGGGCTTGGC-3Ј were used to generate a DNA fragment in which Ϫ280 to Ϫ160 was deleted from the wild-type pGL-mP350. The following oligonucleotides were used to generate mutations in putative control elements: AP4, 5Ј-CCAGGCAGGAGGGCC-CAAAGATGGGAGGGGCCGAGGCG-3Ј; SP1, 5Ј-CCAGGCAGGAGGG-CCCACAGCTGGGAAAGGCCGAGGCG-3Ј; AP4 and SP1, 5Ј-CCAGG-CAGGAGGGCCCAAAGATGGGAAAGGCCGAGGCG-3Ј. These oligonucleotides were used for PCR with the wild-type pGL-mP220 plasmid as the template. For mutations of the putative thrombin response element (ThR), the QuickChange site-directed mutagenesis kit (Stratagene) was used. The sense primer was: 5Ј-CATCGAAAGCAGACGCC-CAGATCTGACTCAGCGGCGACCTAC-3Ј and the antisense primer was 5Ј-GTAGGTCGCCGCTGAGTCAGATCTGGGCGTCTGCTTTCG-ATG-3Ј.
Bold, underlined oligonucleotides represent the BglII restriction enzyme sites. These primers were used in PCR separately with both the wild-type pGL3-mP1120 and pGL3-mP350 as the templates. The PCR fragment was cloned into pGL3 that had been cut with SmaI and XhoI. All the constructs were verified by dideoxy sequencing the inserts and flanking regions with the sequenase kit (United States Biochemical Corp.).
Transient Transfection and Luciferase Assay-Transfection was carried out by the Lipofectin method (Life Technologies, Inc.). Cells were grown on 60-mm dishes to 70 -80% confluence and transfected with 10 g of various mEPCR-luciferase plasmids and the pGL3-control. To determine transfection efficiency, 5 g of plasmid p␤Gal-control (containing the SV40 promoter) was co-transfected with each test plasmid. Following transfection for 8 h at 37°C in 5% CO 2 , the medium was replaced with fresh Eagle's minimal essential medium, 10% FCS. In some experiments, the DNA-Lipofectin mix was replaced with fresh Opti-MEM I with or without bovine ␣-thrombin as indicated.
The cells were harvested 48 h after transfection. The luciferase and ␤-galactosidase activities for each transfection were measured three times with a Monolight 2001 luminometer (PharMingen, CA). Luciferase and ␤-galactosidase activities were measured with Luciferin (PharMingen) and Galactan-star (CLONTECH) according to the manufacturers' instructions. The data represent the means Ϯ S.D. of three independent experiments. Duplicate transfections were performed in each experiment.
Preparation of Nuclear Extract and Gel Mobility Shift Assay-Nuclear extracts from BAEC were prepared as described previously (21). The protein concentration of each extract was determined with the Bio-Rad protein assay. Double-stranded oligonucleotides were synthesized corresponding to Ϫ180 to Ϫ220 of the murine EPCR promoter region. These contained the wild-type sequence or mutations in the AP4, SP1, or AP4/SP1 sites.
Probes were end-labeled with T4 polynucleotide kinase and [␥-32 P]ATP (6000 Ci/mmol, Amersham Pharmacia Biotech). Nuclear protein extracts prepared from BAEC were preincubated in a 20-l reaction containing 20 mM Tris-HCl, pH 7.5, 25 mM KCl, 4 mM MgCl 2 , 0.5 mM dithiothreitol, 3 g of poly(dI-dC), and 10 g of bovine serum albumin. Competitor DNA was added as needed. After incubation at room temperature for 20 min, the 32 P-end-labeled duplex oligonucleotides (Ϸ2 ϫ 10 5 cpm) were added, and the reaction was incubated for 20 min at room temperature. Where indicated, 2 g of rabbit anti-SP1 polyclonal antibody (Santa Cruz Biotechnology) was added to the binding reaction for 30 min at room temperature prior to the addition of radioactive probe. Samples were then analyzed on a 5% polyacrylamide gel in 0.5ϫ TBE buffer (22.5 mM Tris base, 22.5 mM boric acid, and 0.5 mM EDTA). The gels were vacuum-dried and exposed to x-ray film overnight with an intensifying screen at Ϫ80°C.
DNase I Footprint Analysis-A DNA fragment of the mEPCR promoter (Ϫ280 to Ϫ1) was isolated from pBKS-mP280 with EcoRI and HindIII in the presence of calf intestinal alkaline phosphatase and was end-labeled with T4 polynucleotide kinase and [␥-32 P]ATP (6000 Ci/ mmol). After labeling, the probe DNA was completely digested with HindIII to remove the label from the 3Ј end of the fragment and to create the coding strand probe. The binding reaction (50 l) was carried out in a reaction volume containing 2 ϫ 10 4 cpm (1-2 ng) of the probe, 10 mM Tris-HCl, pH 8.0, 25 mM KCl, 2.5 mM CaCl 2 , 5 mM MgCl 2 , and 50 g of crude BAEC nuclear extract or 10 ng of purified SP1 protein (Promega). After incubation for 30 min at room temperature, the reaction mixture was digested with freshly diluted DNase I for 1 min. Digestion was stopped by adding 90 l of stop solution (200 mM NaCl, 30 mM EDTA, 1% SDS, 100 g/ml yeast tRNA). The DNA was then extracted with phenol/chloroform, precipitated with ethanol, and analyzed on an 8% denaturing polyacrylamide gel. A dideoxy DNA sequencing reaction was electrophoresed in parallel to map the location of the protected region of the mEPCR promoter. After electrophoresis, the gel was vacuum-dried and exposed to x-ray film overnight with an intensifying screen at Ϫ80°C.
Generation of Transgenic Mice-Transgenic mice were developed using the Ϫ350 to Ϫ1 or the Ϫ1080 to Ϫ1 region of the murine EPCR promoter to drive the structural gene for green fluorescent protein (GFP) to determine the in vivo activity of these regions of the EPCR promoter.
The promoter region (Ϫ350 to Ϫ1) or (Ϫ1080 to Ϫ1) of the murine EPCR gene was cloned into the pEGFP1 vector (CLONTECH), which already contains the structural gene for GFP, using KpnI and SmaI sites. The fragment containing the promoter region of mEPCR and GFP reporter gene was released by Eco47III and AflII. After purification, the DNA fragments were microinjected into the pronuclei of fertilized murine eggs by standard methods. Mice were screened by standard methods for the presence of the transgene by GFP-specific PCR and Southern blotting. Four transgenic lines were established with the Ϫ350 to Ϫ1 construct and five lines with the Ϫ1080 to Ϫ1 construct.
Transgenic animals bearing the pEGFP-350 or pEGFP-1080 construct were injected intraperitoneally with 400 g of LPS (20 mg/kg). Controls were injected intraperitoneally with saline. The tails were cut before and after LPS treatment for the preparation of total RNA. When hirudin was employed, it was dissolved in saline and injected through the tail vein (100 units/kg). 30 min after beginning the hirudin injection, LPS was injected intraperitoneally as a bolus (20 mg/kg). In some control experiments, instead of hirudin, saline was injected before the LPS treatment. 3 h after the injection of LPS or saline, the animals were sacrificed, and the heart and lungs were immediately removed for RNA preparation. The study protocol received prior approval by the Institutional Animal Care and Use Committee of the Oklahoma Medical Research Foundation.
Quantitative RT-PCR-Total RNA was isolated from murine tails or organs with the Trizol reagent. First-strand cDNA synthesis was performed with the Superscript Preamplification System as specified by the manufacturer (Life Technologies, Inc.). Before transcription, DNase I (RNase-free) was added to each RNA sample to ensure no contamination of the RNA preparation with DNA.
For the competitive quantitative amplification of GFP and mEPCR mRNA, PCR mimics were constructed using the PCR Mimic construction kit (CLONTECH) as specified by the manufacturer. Each composite primer consisted of a target gene sequence (from GFP, mEPCR, or ␤-actin) attached to one of two 20-bp sequences designed to hybridize to opposite strands of a mimic DNA fragment, the -Erb gene. A second round of PCR was performed using only gene-specific primers. The PCR mimic was then purified and adjusted to a concentration of 100 amol/l. For quantitative PCR, 10-fold dilutions of mimic were added to a series of tubes containing a fixed concentration of first-strand cDNA and a pair of gene-specific primers. 25 cycles of PCR was performed under each of the following conditions: 94°C for 1 min, 60°C for 1 min, and 72°C for 2 min.

Sequence and Structural Features of the 5Ј-Flanking
Region of the Murine EPCR Gene-1200 bp from the 5Ј-noncoding region of the gene coding mEPCR was isolated from the P1 genomic DNA library and sequenced (Fig. 1). The sequence from Ϫ400 to ϩ60 was identical to that previously reported by Liang et al. (23), while the additional sequence from Ϫ1200 to Ϫ400 was established in the present study. The bases are numbered relative to the translation initiation site. There is no canonical TATA or CCAAT box (23) as there is in the human EPCR 5Ј-flanking sequence (24). There is a putative TATA box, CTTTAAAAGCC (Ϫ152 to Ϫ142), and a weak CCAAT box, GGCCCACAGC (Ϫ220 to Ϫ211). Both of these boxes are within the optimal distance of 57-212 nucleotides upstream of the transcription start site and are credible putative elements, which are required for the RNA polymerase II promoter (25). Potential regulatory elements were identified in the 5Ј region of the murine EPCR gene including SP1, AP1, NF1, Ets, and AP4 binding sites. These sites had only one or two mismatches to the consensus sequences Identification of Transcription Initiation Sites-Having identified the genomic sequence for murine EPCR, we next attempted to define the transcription start sites of the gene using RNase protection of mRNA isolated from the murine HA cells.
A 32 P-labeled antisense riboprobe (mP350) spanning Ϫ350 to Ϫ1 upstream of the protein sequence was annealed to total RNA isolated from HA cells. Following treatment with RNase A and RNase T1, the fragments protected from RNase digestion were separated electrophoretically as described. Six major transcription start sites at Ϫ100, Ϫ102, Ϫ104, Ϫ105, Ϫ107, and Ϫ109 bp were identified by comparison to the nucleotide sequence of mEPCR initiated at the base pair representing the 3Ј end of the synthesized riboprobe (Fig. 2). No protected bands were found when human EPCR RNA (from EAhy926 cells) or yeast tRNA were used, indicating the specificity of the assay. These potential transcription start sites fall between the putative TATA box and the translation initiation codon. The centerto-center spacing between the TATA box and cap signal was 35 bp. Initiation at the furthest upstream site would result in a 110-bp 5Ј-untranslated region. Multiple start sites are consistent with this relatively broad band detected in Northern blot analysis of murine tissue RNA (data not shown).
Transient Expression of mEPCR Promoter Activity-To determine which regions of the 5Ј-flanking region of the mEPCR gene had promoter activity, we prepared chimeric constructs in which serially deleted fragments of the 5Ј-flanking sequence were inserted before a promoterless luciferase gene in the plasmid, pGL3-basic. Luciferase expression was measured after transfection of these constructs into BAEC. Constructs mP3340, mP1120, mP550, mP350, and mP280 promoted similar levels of luciferase activity, but background expression was observed in the construct mP160 and mP80 (Fig. 3). The mP180 expression was only slightly above background. Deletion of the sequence between Ϫ220 and Ϫ180 dramatically reduced luciferase expression in BAEC, suggesting the presence of a positive regulatory element(s) in this region Ϫ220 to Ϫ180.
Deletion constructs were transfected in different cell lines of both endothelial cell and non-endothelial origin. Similar results to those above were obtained in EAhy926 cells. All constructs were expressed poorly, if at all, in HeLa and 293 cells.
The AP4 Site and SP1 Site Are Required for Constitutive Expression of mEPCR in BAEC-In the promoter sequences of mEPCR, we found one AP4 site and one SP1 site between Ϫ220 and Ϫ180. It seemed likely these two elements were involved in the constitutive expression of mEPCR.
To determine whether these elements could bind nuclear proteins, we synthesized 40 bp of double-stranded oligonucleotides encompassing this sequence (Ϫ220 to Ϫ180). The core sequences in AP4 and SP1 sites were mutated (Fig. 4A). Wildtype, 32 P-end-labeled, double-stranded oligonucleotide probe was incubated with nuclear extracts from BAEC. Two DNAprotein complexes were found in the PAGE gels (Fig. 4B). The formation of these two complexes was sequence-specific since it was prevented by addition of a 50-fold molar excess of the unlabeled oligonucleotide, but not by a 50-fold molar excess of oligonucleotide containing both SP1 and AP4 mutations. The upper band was apparently SP1, because it was eliminated by a 50-fold molar excess of the AP4 mutated oligonucleotide, but not by a 50-fold excess of the SP1 mutant. The lower band was apparently AP4, because it was eliminated by a 50-fold molar excess of the SP1 mutated oligonucleotide, but not by a 50-fold excess of the AP4 mutant. Thus, both the SP1 and AP4 sites are potentially functional.
To test whether the core SP1 and AP4 sequence was required for binding of the nuclear proteins, we performed gel shift assays using different mutated oligonucleotides as probes (Fig.  4C). When using the SP1 mutant probe, the probe was unable to bind the SP1 protein. Only the AP4 DNA-protein complex could be formed. This AP4 complex can be prevented by addition of a 50-fold molar excess of cold homologous probe, or the AP4 consensus probe, but not by either the AP4 mutant or the SP1/AP4 mutant competitors. The AP4 mutant probe was unable to bind the AP4 protein and only the SP1 DNA-protein complex could be detected. The SP1 complex band was eliminated by addition of a 50-fold excess of cold homologous probe, wild-type, or SP1 consensus oligonucleotides, but not by the SP1 mutant or the SP1/AP4 mutated oligonucleotides. If both the SP1 and AP4 sites were mutated, the probe was unable to bind to BAEC nuclear proteins. These data clearly show that the SP1 and AP4 elements in the region Ϫ220 to Ϫ180 of the mEPCR promoter can bind nuclear proteins from BAEC.
A supershift assay was performed to test whether the binding of nuclear protein to the SP1 site (GGGAGGGG) of the mEPCR gene could be shifted by an SP1 antibody (Fig. 5). A single band was formed when BAEC nuclear extracts was incubated with labeled oligonucleotide probe containing the SP1 site (GGGAGGGG) of the mEPCR gene, and this band was competitively inhibited by a 50-fold molar excess of unlabeled probe and SP1 consensus oligonucleotide, but not by a 50-fold molar excess of the SP1 mutated oligonucleotide. This band was supershifted with an antibody directed against SP1.
The DNase I footprinting assay was employed to identify protected areas in the Ϫ220 to Ϫ180 region of the mEPCR promoter. When labeled probes were incubated with BAEC nuclear extracts, a protected area was identified from Ϫ200 to Ϫ214. When labeled probes were incubated with purified SP1 protein, a protected area was identified from Ϫ200 to Ϫ209 (Fig. 6). These protected areas contain an SP1 site (GG-GAGGGG) and an AP4 site (CAGCTG).
The AP4 site and SP1 site in the wild-type construct mP220 were mutated separately. When the mutant constructs were transfected into BAEC, luciferase expression in each case was reduced to 30 -40% of the wild-type level. However, if the AP4 and SP1 were both mutated, the luciferase activity of the construct was essentially eliminated (Fig. 7). These data suggest that the AP4 and SP1 elements were functional, as mutations FIG. 2. RNase protection assay to determine the transcription start sites of the mEPCR gene. A 32 P-labeled antisense riboprobe (mP350) spanning bp Ϫ350 to Ϫ1 was annealed to total RNA isolated from murine HA cells, EAhy926 human cells, and yeast tRNA. The fragments protected by RNase A and T1 digestion were separated on a denaturing polyacrylamide gel. Six major transcription start sites at Ϫ100, Ϫ102, Ϫ104, Ϫ106, Ϫ107, and Ϫ109 bp were identified by comparison to the nucleotide sequence of mEPCR. No protected bands were found when hEPCR RNA or yeast tRNA were used.

FIG. 3. mEPCR promoter activity in BAEC.
Serially deleted fragments of the 5Ј-flanking region were linked to the luciferase reporter gene. Luciferase expression was measured 48 h after transfection of these constructs into BAEC. Deletion of the sequences between Ϫ220 and Ϫ180 dramatically reduced luciferase expression in BAEC.
in the core sequences impaired both nuclear protein binding and gene expression.
Identification of a Thrombin Response Element in the Promoter Region of mEPCR-To define the region corresponding to the ThR, the serially truncated mEPCR promoter-luciferase constructs and the putative ThR mutant constructs were transiently transfected into the BAEC. These cells were subsequently treated with bovine ␣-thrombin (10 units/ml) for 6 h and luciferase activity was quantitated in the cell lysates. As shown in Fig. 8, the constructs mP1120 and mP350 retained thrombin responsiveness, increasing luciferase activity approximately 3-fold after thrombin stimulation. When the putative thrombin response element was changed from CCCACCCC to CCCAGATC in these constructs, they were no longer responsive to thrombin when transfected into BAEC. Likewise, the mP280 construct, which contains the minimal promoter required for expression in these cells, demonstrated no increased luciferase expression in response to thrombin. These observations indicate that the region spanning oligonucleotides Ϫ350 to Ϫ280 and the ThR (see Fig. 1) are involved in the transcriptional response to thrombin.
A gel shift mobility assay was employed to identify the thrombin response element of mEPCR gene capable of binding to transcription factors present in the nuclear extracts of thrombin-treated, but not in untreated, BAEC. BAEC were grown in the serum-free medium for 24 h before stimulation with thrombin. After stimulation with thrombin for 5 h, nuclear protein was prepared as described under "Experimental Procedures." A labeled double-stranded oligonucleotide that contained the thrombin response element (CCCACCCC) of mEPCR gene was incubated with nuclear extracts prepared from BAEC with or without thrombin treatment. As shown in Fig. 9, a thrombin-dependent DNA complex (TINF) was formed only when BAEC were treated with thrombin. This DNA-protein complex was eliminated by 50-fold molar excess of the unlabeled oligonucleotide, but not by a 50-fold excess of unlabeled oligonucleotide in which the CCACCC sequence was changed to CCAGAT. The DNA-protein complex could also be competed by a 50-fold excess of the sequence CCACCCACC, the sequence corresponding to the thrombin response element in the platelet-derived growth factor B chain promoter (26), suggesting that the same nuclear protein could bind to both thrombin response elements.
Taken together, these data strongly suggest that the thrombin response element (CCCACCCC) in the mEPCR promoter region is essential for thrombin induction of mEPCR transcription.
Generation of Transgenic Mice with the Construct mP350-GFP and Its Response to Endotoxin in Vivo-The experiments with BAEC transfected with the deletion constructs and ThR mutated constructs demonstrated that thrombin can up-regulate luciferase expression. To determine whether these in vitro experiments predicted in vivo responses, transgenic mice were developed using the Ϫ350 to Ϫ1 region and the Ϫ1080 to Ϫ1 region of the murine EPCR promoter to drive GFP expression. Mice were screened for the presence of the transgene by GFPspecific PCR and Southern blotting. Four transgenic animal lines were established with the Ϫ350 to Ϫ1 construct (mP350) and five transgenic lines were established with the Ϫ1080 to Ϫ1 construct (mP1080). In these lines the constitutive level of GFP and EPCR mRNA were similar (Fig. 10). 4. Gel mobility shift assay. A, the sequences of the upper strand oligonucleotides used as probes and competitors in the gel mobility shift assay. B, the 32 P-end-labeled wild-type probe (Ϫ220 to Ϫ180) was incubated with nuclear extracts from BAEC. Two DNA-protein complexes (SP1 and AP4) were found in the PAGE gel. The SP1 band can be competed and abolished by a 50-fold molar excess of cold probe and AP4 mutant oligonucleotides. The AP4 band was competed by a 50-fold molar excess of cold probe and SP1 mutant. C, gel mobility shift assays were performed using the mutated oligonucleotides as probe as indicated.
To determine the influence of LPS on the expression of the EGFP-350 and EGFP-1080 constructs in transgenic mice, animals bearing the transgenes were injected intraperitoneally with 400 g of LPS. Control mice received an injection of buffer alone. After 3 h, the mice were sacrificed and RNA was isolated from the heart and lungs. The levels of mRNA for GFP, endogenous mEPCR, and ␤-actin were measured by competitive RT-PCR as described under "Experimental Procedures." The numbers of animals, the distribution of sexes, and the transgenic line distribution are given in Table I. Fig. 10 shows representative agarose gel electrophoresis of RT-PCR products from lung tissue of the four transgenic lines. The ability of the bona fide message to compete for its "mimic" at a particular dilution of the mimic indicates the abundance of the message in the original sample. After 3 h of treatment with 400 g of LPS, the transgenic animal expressed approximately 10 times higher levels of mRNA for both GFP and endogenous mEPCR (i.e. it required approximately 10 times higher concentration of mimic to compete out the EPCR and GFP signal, whereas the amount of required to compete out ␤-actin mRNA was essentially unchanged).
To examine the potential role of thrombin in endotoxin/LPSmediated up-regulation of EPCR mRNA, hirudin, a specific thrombin inhibitor, was injected prior to administration of LPS. Hirudin diminished the LPS-mediated induction of both GFP and EPCR mRNA, resulting in levels of these mRNAs similar to those of the controls (Fig. 10). The level of GFP mRNA induction by LPS was comparable to that of EPCR in all four lines and, in all cases, hirudin treatment prior to LPS returned the GFP and EPCR level to near that observed constitutively. Similar results were obtained in the heart (data not shown). These results strongly implicate thrombin as a major mediator of endotoxin-induced up-regulation of EPCR mRNA. DISCUSSION The protein C pathway plays a critical role in the host defense against bacterial sepsis (13) and is critical to the preven-FIG. 5. Binding of nuclear proteins to mEPCR SP1 elements that supershifted by SP1 antibody. A, oligonucleotide sequences used in the binding assay; B, nuclear extracts were prepared from BAEC. Labeled double-stranded oligonucleotide probe was incubated with nuclear extracts (10 g) as described under "Experimental Procedures" in the absence or presence of 50-fold molar excess of unlabeled oligonucleotides as indicated, or 2 g of anti-SP1 antibody was added during the binding assay. Arrows indicate the positions of the SP1 band, supershifted SP1 band, and free labeled probe.
FIG. 6. DNase I footprinting analysis of the promoter region (؊220 to ؊180) of the mEPCR gene. The DNA fragment (Ϫ280 to Ϫ1) from the mEPCR promoter was end-labeled and digested with HindIII to create the coding strand probe. The end-labeled fragment was incubated with either 10 ng of purified SP1 protein or 50 g of crude BAEC nuclear extracts. The reaction mixture was then digested with DNase I for 1 min. Bound and free DNA fragments were separated by electrophoresis on an 8% denaturing polyacrylamide gel. A dideoxy DNA sequencing reaction was electrophoresed in parallel to map the location of the protected region of the mEPCR promoter gene. The film was exposed for 18 h at Ϫ80°C with an intensifying screen.  7. Mutation in the core sequence of the AP4 and SP1 sites impair luciferase expression in BAEC. The SP1 and AP4 sites in the wild-type construct mP220 were mutated separately. When the mutant constructs were transfected into BAEC, luciferase expression was reduced 30 -40%, compared with the wild type. When both the AP4 and SP1 sites were mutated, the construct expressed at levels slightly below those of mP180. tion of thrombosis (27). Recent studies have identified EPCR gene defects in patients and suggest that these may be associated with thrombotic disease (17). In addition, other recent studies have indicated that inhibiting EPCR function, like inhibiting protein C (13) or protein S (28), exacerbates the host response to E. coli in primates leading to disseminated intravascular coagulation and elevated inflammatory responses (16). In cell culture, both EPCR and TM have been shown to be down-regulated by inflammatory mediators like TNF␣, raising questions about the ability of the system to respond in the presence of a severe inflammatory challenge. Protein C (29) and activated protein C (30) are being used in clinical trials to examine their potential to treat sepsis. The clinical response to protein C supplementation in severe sepsis has appeared to be good (29); hence, if EPCR were to play any role in the protective effects of protein C or APC, it could not be down-regulated by this disease process. Consistent with this possibility, immunohistochemistry of human tissues indicates that there is a protective mechanism that prevents down-regulation in vivo. In particular, EPCR antigen was readily stained on vessels from a patient who died of respiratory distress (10). To understand the basis of this difference in in vivo and in vitro responses and to gain better insights into the regulation of EPCR under these conditions, we are establishing mouse models to examine the role of EPCR regulation in acute inflammatory diseases as well as thrombosis. To guide this process, we have extended the sequence of EPCR gene providing 800 new base pairs of 5Ј sequence and have characterized several critical regulatory elements in the gene.
Potential regulatory elements were identified in the 5Ј region of the murine EPCR gene including SP1, AP1, NF1, Ets, and AP4 binding sites. These sites had only one or two mismatches to the consensus sequences. These elements are of interest because of their frequent involvement in cell growth and development (31)(32)(33)(34). Of these, we analyzed the function of the SP1 and AP4 binding sites located between Ϫ220 and Ϫ180 and found, based on expression of fusion constructs, that both sites were required for optimal constitutive expression of mEPCR.
Constructs containing only the first 220 bp of the 5Ј region were expressed constitutively at levels comparable to the much larger constructs containing up to the first 3340 bp. In addition, considerable cell type specificity is present in the first 220 bp, since these constructs expressed well in BAEC, but poorly in HeLa cells and human kidney 293 cells. No elements were identified in this sequence that are known to impart endothelial cell specificity. Whether this sequence will direct endothelial cell-specific expression in vivo remains to be determined. FIG. 9. Characterization of the thrombin response element of mEPCR gene by gel shift mobility assay. A, oligonucleotide sequences used in the binding assay; B, nuclear extracts prepared from BAEC treated with or without bovine thrombin (10 units/ml) for 5 h. The 5Ј end-labeled mEPCR ThR oligomer was incubated with nuclear extracts (10 g) in the absence or presence of a 50-fold molar excess of unlabeled oligonucleotides as indicated in A. Gel shift mobility assay was performed as described under "Experimental Procedures." Arrows indicate the positions of the thrombin-dependent DNA complex (TINF) and free labeled probe.

FIG. 10. The LPS-mediated up-regulation of GFP and endogenous mEPCR mRNA levels is blocked by hirudin.
Mice were injected intraperitoneally with buffer (control), or buffer containing 400 g of LPS as described under "Experimental Procedures." For the LPS plus hirudin group, mice received an intravenous injection of 100 units/kg hirudin through the tail vein 30 min before the LPS injection. Three hours after LPS or saline injection, the mice were sacrificed and the organs immediately collected for RNA isolation. The levels of mRNA for GFP, mEPCR, and ␤-actin were measured by the quantitative RT-PCR described under "Experimental Procedures." The levels of competitor for the quantitative PCR are indicated above the figure. The higher the level of competitor required to block the appearance of the gene specific PCR product, the higher the level of that particular gene transcript in the tissue. Four sets of gels are shown. Each is in the following format. From top to bottom, the levels of PCR product for the transgene (labeled GFP), murine EPCR (mEPCR) and actin. Control indicates animals receiving saline injection. LPS indicates animals receiving LPS, and LPSϩhirudin represents animals pretreated with hirudin prior to LPS injection. A representative gel is shown from each of the transgenic lines examined in this study. A, transgenic line A205-3 with the Ϫ350 to Ϫ1 construct; B, line A212-1 independently derived and with the same construct; C, line B186 -1 with the Ϫ1080 to Ϫ1 construct; D, line B186-6 independently derived and with the Ϫ1080 to Ϫ1 construct.
Unfortunately, the GFP constructs used to analyze the transgenic mice failed to generate sufficient fluorescence to allow assessment of cellular specificity.
A prominent feature of EPCR gene regulation involves the thrombin response element that has a similar sequence to that found in the platelet-derived growth factor B chain promoter (26) and is upstream of the SP1 and AP4 binding sites. This element is required for thrombin stimulation of EPCR mRNA production. It is functional both in cell culture and in vivo in transgenic mice. Of particular interest, endotoxin, an agent that elicits the generation of many inflammatory cytokines such as TNF␣, up-regulates EPCR mRNA expression and the expression of the transgene. This was somewhat surprising since cell culture experiments demonstrated that EPCR mRNA and function were down-regulated almost completely within 12 h following exposure murine or human endothelial cells to TNF␣ (4,18). This up-regulation of the EPCR by endotoxin is dependent on the thrombin that is generated in response to endotoxin since both the EPCR and the transgene mRNA levels do not rise in response to endotoxin when the mice are treated with the potent and specific thrombin inhibitor, hirudin. Given the myriad of mediators that are generated in response to endotoxin (35) and the observation that gene expression levels stay constant when endotoxin is administered in the presence of hirudin, it appears that the thrombin response element is the dominant element involved in the acute inflammatory responses of mEPCR. Therefore, thrombin plays a dominant role in both protein C activation and in the regulation of EPCR, a protein intimately involved in protein C activation.
Inspection of the genomic sequences in the vicinity of the translation start site indicates that there are no canonical TATA or CCAAT boxes in the murine EPCR 5Ј-flanking region (23) as compared with the human EPCR 5Ј-flanking sequence (24), or other general RNA polymerase II promoter sequences (25). However, there is a putative TATA box, CTTTAAAAGCC (Ϫ152 to Ϫ142), and a weak CCAAT box, GGCCCACAGC (Ϫ220 to Ϫ211), located upstream of the coding region of the murine EPCR gene. Both of these boxes are centered within the optimal distance of 57-212 nucleotides upstream of the transcriptional start site and are credible putative elements required for RNA polymerase II promoter activity (25). Unlike TATA-less housekeeping genes, the 5Ј-flanking region sequence of hEPCR and mEPCR are not GC-rich.
The transcription start sites of the mEPCR gene are heterogeneous (Ϫ100 to Ϫ109). The presence of multiple transcrip-tion initiation sites may explain in part the highly varied levels of EPCR mRNA in cultured endothelial cell lines and large vessels (data not shown). Similar cases of heterogeneous transcription start sites in tissue-specific genes have been observed whether or not the gene contains a TATA box (36 -39).
These studies reveal that a thrombin response element in the EPCR gene plays a major role in the regulation of gene expression and appears to dominate the negative regulatory elements observed in vitro. To our knowledge, this is the first time that this element has been shown to function in vivo. Future experiments are aimed at determining whether this element plays a critical role in regulating EPCR functions during the host response to severe acute inflammation. Control is saline injection 3 h before organ removal. LPS is LPS injection 3 h before organ removal and LPS ϩ hirudin is hirudin infusion prior to LPS injection with organ removal 3 h after LPS injection. M, male; F, female. On the left are the designations of the different transgenic lines. Ϫ350 or Ϫ1080 indicates the number of bp from the Ϫ1 position of the promoter included in the EPCR promoter used to drive GFP expression.