Transgenic Mice with −6A Haplotype of the Human Angiotensinogen Gene Have Increased Blood Pressure Compared with −6G Haplotype*

Hypertension is a serious risk factor for cardiovascular disease, and the angiotensinogen (AGT) gene locus is associated with human essential hypertension. The human AGT (hAGT) gene has an A/G polymorphism at −6, and the −6A allele is associated with increased blood pressure. However, transgenic mice containing 1.2 kb of the promoter with −6A of the hAGT gene show neither increased plasma AGT level nor increased blood pressure compared with −6G. We have found that the hAGT gene has three additional SNPs (A/G at −1670, C/G at −1562, and T/G at −1561). Variants −1670A, −1562C, and −1561T almost always occur with −6A, and variants −1670G, −1562G, and −1561G almost always occur with −6G. Therefore, the hAGT gene may be subdivided into either −6A or −6G haplotypes. We show that these polymorphisms affect the binding of HNF-1α and glucocorticoid receptor to the promoter, and a reporter construct containing a 1.8-kb hAGT gene promoter with −6A haplotype has 4-fold increased glucocorticoid-induced promoter activity as compared with −6G haplotype. In order to understand the physiological significance of these haplotypes in an in vivo situation, we have generated double transgenic mice containing either the −6A or −6G haplotype of the hAGT gene and the human renin gene. Our ChIP assay shows that HNF-1α and glucocorticoid receptor have stronger affinity for the chromatin obtained from the liver of transgenic mice containing −6A haplotype. Our studies also show that transgenic mice containing −6A haplotype have increased plasma AGT level and increased blood pressure as compared with −6G haplotype. Our studies explain the molecular mechanism involved in association of the −6A allele of the hAGT gene with hypertension.

The renin-angiotensin system plays an important role in the regulation of blood pressure. The octapeptide angiotensin-II is one of the most active vasopressor agents and is ob-tained from angiotensinogen (AGT) 2 by the combined proteolytic action of renin and angiotensin-converting enzyme. AGT is primarily synthesized in the liver and to a lesser extent in the fat, kidney, brain, heart, adrenal, and vascular walls (1). The plasma concentration of AGT is close to the Michaelis constant of the enzymatic reaction between renin and AGT (2). For this reason, a rise in plasma AGT levels can lead to a parallel increase in the formation of angiotensin-II that may ultimately result in hypertension. Injection of recombinant AGT increases plasma AGT level and blood pressure in rats (3). Overexpression of the AGT gene increases blood pressure in transgenic animals (4), and AGT gene knock-out mice have reduced blood pressure (5). Kim et al. (6) have introduced up to four copies of the AGT gene in mice with each copy of the gene, resulting in a successive increase in blood pressure. These results directly demonstrated that small increases in plasma AGT level could increase blood pressure in a gene dose-dependent manner.
The AGT gene is associated with essential hypertension in Caucasians (7), Japanese (8), and Asian Indian subjects (9). Although several investigators have confirmed these initial association studies, some studies have refuted this claim (reviewed in Ref. 10). However, a number of large studies and meta-analysis have confirmed this association (10 -12). This gene may also be involved in cardiac hypertrophy (13), coronary atherosclerosis (14), and microangiopathy-related cerebral damage (15,16). Initially, it was shown that variant M235T is associated with essential hypertension and increased plasma AGT level in hypertensive patients. However, because amino acid 235 is located far away from the renin cleavage site, this polymorphism does not explain the molecular mechanism involved in increased plasma AGT level. hAGT gene also has an A/G polymorphism at the Ϫ6-position. Inoue et al. (17) have shown that molecular variants 235T and Ϫ6A are in linkage disequilibrium, and reporter constructs containing the human AGT gene promoter with Ϫ6A have increased promoter activity compared with Ϫ6G. This observation suggests that increased plasma AGT levels by 235T in hypertensive patients may actually be due to in-* This work was supported, in whole or in part, by National Institutes of creased transcriptional activity of the human AGT gene by Ϫ6A allele. Two large genome-wide analyses have been performed recently by the Wellcome Trust consortium (52) and the Framingham Heart study consortium (18) to identify genes involved in human hypertension using the Affymetrix 500K SNP platform. However, this platform has neither M235T nor Ϫ6A/G polymorphism and therefore could not evaluate the role of this SNP in hypertension.
In order to prove the role of Ϫ6A allele in hypertension, Sigmund's group used a knock-in approach at the hypoxanthine-guanine phosphoribosyltransferase locus and generated transgenic mice containing either Ϫ6A or Ϫ6G in the 13.5-kb hAGT gene (containing 1.2 kb of the promoter and all of the exons and introns) (19). However, analysis of these transgenic mice showed that variation at nucleotide Ϫ6 has no effect either on the blood pressure or on the tissue-specific expression of the hAGT gene. This suggested that polymorphism at the Ϫ6-position may only be a marker and not a functional polymorphism. We have found that the hAGT gene has three additional SNPs (A/G at Ϫ1670, C/G at Ϫ1562, and T/G at Ϫ1561), and variants Ϫ1670A, Ϫ1562C, and Ϫ1561T almost always occur with Ϫ6A, and variants Ϫ1670G, Ϫ1562G, and Ϫ1561G almost always occur with Ϫ6G. Thus, the hAGT gene may be subdivided into Ϫ6A and Ϫ6G haplotypes. We have also shown that an oligonucleotide containing Ϫ1670A has stronger affinity for hepatocyte nuclear factor-1␣ (HNF-1␣) as compared with Ϫ1670G, and an oligonucleotide containing Ϫ1562C and Ϫ1561T has stronger homology with glucocorticoid-responsive element (GRE) as compared with Ϫ1562G and Ϫ1561G. In order to understand the role of these polymorphic sites on transcriptional regulation of the hAGT gene, we have synthesized reporter constructs containing a 1.8-kb promoter with either Ϫ6A or Ϫ6G haplotype and used them for transient transfection in human liver and kidney cells. We have shown that the reporter construct with Ϫ6A haplotype has 1.4 -1.6-fold increased basal promoter activity and 4-fold increased glucocorticoid-induced promoter activity as compared with the reporter construct containing Ϫ6G haplotype. These experiments have suggested that the Ϫ6A allele of the hAGT gene may be associated with increased plasma AGT level and with increased blood pressure due to increased binding of HNF-1␣ and glucocorticoid receptor (GR) to the hAGT gene promoter containing variants Ϫ1670A, Ϫ1562C, and Ϫ1561T that always occur with Ϫ6A. In order to understand the role of these SNPs in the expression of the hAGT gene in a tissue-specific manner and in the regulation of blood pressure in an in vivo situation, we have recombineered 180-kb-long BAC DNA (containing 116 kb of the 5Ј-flanking region, all five exons and four introns, and 54 kb of the 3Ј-flanking region of the hAGT gene) that contains Ϫ6G haplotype. We exchanged the 1.8-kb hAGT gene promoter in this BAC with the 1.8-kb fragment obtained from the genomic DNA from a hypertensive subject containing Ϫ6A haplotype. We used these BAC DNAs and produced transgenic mice containing either Ϫ6A or Ϫ6G haplotype of the hAGT gene. Because mouse renin is not able to cleave hAGT, we have made double transgenic mice containing human renin gene and either Ϫ6A or Ϫ6G haplotype of the hAGT gene to understand the physiological role of these haplotypes in an in vivo situation. We show here that double transgenic mice containing Ϫ6A allele of the hAGT gene have increased plasma AGT level and increased blood pressure as compared with transgenic animals containing Ϫ6G allele.

MATERIALS AND METHODS
Cell Culture-Human hepatoma cells (HepG2) and human kidney cells (HK-2) were grown as a monolayer in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum, 100 units/ml penicillin, and 100 g/ml streptomycin in an atmosphere of 5% CO 2 at 37°C. Cells were grown to 80% confluence in 6-well plates for transient transfections.
Plasmid Construction-The reporter construct pHAGT2.6luc (or pHAGT2572luc), and its deletion constructs pHAGT1748luc, pHAGT1657luc, pHAGT1624luc, pHAGT1604luc, pHAGT1544luc, pHAGT1459luc, and pHAGT1223luc were synthesized by PCR amplification of the human AGT gene using TTAGCTAGCAGCACATCTTAC-ATGGTGGC, TTAGCTAGCTAGCCACAGATAAGCT,  TTAGCTAGCAGGCTTTAAAGAGCCGA, TTAGCTAGC-GCACAAGGTTGT, TTAGCTAGCCAGTGCTAGATC-CTT, TTAGCTAGCCCACTTGCTGATCCTCA, TTAGCT-AGCAGGCAGGAGCACTTTGA, and TTAGCTAGCCCA-GACAAGTGATT as the respective forward primers with the  NheI restriction site and TCGAAGCTTCACAGCTCAGTT-ACATCT as the reverse primer with the HindIII restriction site. The amplified fragments contained the nucleotide sequence Ϫ2581 to ϩ98, Ϫ1748 to ϩ98, Ϫ1651 to ϩ98, Ϫ1618 to ϩ98, Ϫ1598 to ϩ98, Ϫ1541 to ϩ98, Ϫ1457 to ϩ98, and Ϫ1223 to ϩ98, respectively. The PCR-amplified product was subcloned in the pGL3 basic vector that lacks eukaryotic promoter and enhancer sequences (Promega, Madison, WI). The mutant plasmid pHAGT1618lucmutC/EBP and pHAGT1618lucmutGRE were obtained by site-specific mutagenesis using the Stratagene site-directed mutagenesis kit. The nucleotide sequence of oligonucleotides used for mutation of the CCAAT/enhancer-binding protein (C/EBP) site was CAAGGTTGTTATTTCATCAG, and the sequence for the mutation of GRE was TCCAGGTCCACTTGTG-CTAC (mutated nucleotides are shown in boldface italic type).
Transient Transfections-Transfections in HepG2 and HK-2 cell cultures in 6-well plates were carried out using the Lipofectamine reagent (Polyfect, Qiagen, Valencia, CA) using the manufacturer's protocol. Briefly, 250 ng of reporter constructs and 50 ng of RSV ␤-gal was used in each experiment. RSV-GR was co-transfected to increase the expression of glucocorticoid receptor (20). After 4 h of transfection, the medium was changed to the serum-free medium, and after 24 h of transfection, cells were treated for an additional 24 h with dexamethasone (100 nM). The procedures for cell extract preparation and luciferase or ␤-gal activity were carried out as described previously (21).
Gel Mobility Shift Assay-The probes for the electrophoretic mobility shift assay (EMSA) were chemically synthesized, annealed, and radiolabeled at the 5Ј-ends by polynucle-otide kinase using [␥-32 P]ATP. The EMSA protocol was followed as described previously (22). Nuclear extracts for gel mobility shift assays were prepared by modification of a method described previously (23). Antibodies against C/EBP and HNF-1␣ were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA).
Chromatin Immunoprecipitation (ChIP) Assays-The ChIP assay was performed using the ChIP assay kit from Upstate Biotechnology, Inc. (Lake Placid, NY). The protocol was followed as described previously (22). Finally, the immunoprecipitated DNA (1 l) and the input DNA (1 l) were subjected to 35 cycles of PCR amplification using CAGGCACAGTGGAAACT as a forward and ACTGCAG-TAACAAGTC as a reverse primer. This amplified 203-bp product from the upstream region of the human AGT gene promoter (Ϫ1546 to Ϫ1748) spanning HNF-1␣, C/EBP, and GRE was analyzed on 2% Metaphore agarose gel.
Preparation and Modification of hAGT BAC-Human BAC containing the AGT gene (RP11-505D24) was obtained from Invitrogen. BAC was grown in low salt LB medium (5 g/liter NaCl, 10 g/liter tryptone, and 5 g/liter yeast extract) plus 12.5 g/ml chloramphenicol at 37°C, and BAC DNA was purified by using the Nucleobond BAC Maxiprep kit (BD Biosciences). The promoter sequence in the wild type AGT BAC contained Ϫ6G haplotype and was modified by recombineering using the galK two-step procedure in SW102 cells (24). SW102 cells were grown at 32°C. First, the 1.8 kb of the promoter region of the hAGT gene was deleted by inserting the galK cassette. The following PCR primers were used for this purpose (AGT homology is underlined): AAAGACCAGGATCTTTGTTTT-GTTCCCTGACATATGCTGAGCACCAGGAACCTGTTG-ACAATTAATCATCGGCA and AGCTGAGGGGGCCCCC-GGCTTACCTTCTGCTGTAGTACCCAGAACAACGGTC-AGCACTGTCCTGCTCCTT. Next, the galK cassette was replaced with a 1.8-kb promoter region from the DNA of a hypertensive subject containing the Ϫ6A haplotype of the hAGT gene. The genomic DNA from the hypertensive subject was amplified by using AAAGACCAGGATCTTTGTTTTG-TTCCCTGACATATGCTGAGCACCAGGAA as a forward primer and AGCTGAGGGGGCCCCCGGCTTACCTTCTG-CTGTAGTACCCAGAACAACGG as a reverse primer to produce a 1.8-kb region of the promoter. Wild type and modified BAC DNAs were analyzed by fingerprinting after treatment with SpeI. The modified region of the AGT BAC was finally confirmed by direct sequencing. The nucleotide sequences of the promoter, all of the exons, and the 3Ј-flanking region were sequenced to confirm that they differ only in the promoter region and contained Ϫ6A or Ϫ6G haplotypes, respectively.
Generation of Transgenic Mice-The hAGT BAC plasmids were used for microinjection to generate transgenic mice using standard procedure. The presence of the hAGT gene was confirmed by PCR amplification of the tail DNAs of transgenic animals using human AGT gene-specific primers. The forward and reverse primers for the amplification of the hAGT gene were CAGCAGTGAAACTCTGC and TTCAGT-CATCACCGTGC, respectively, to produce a 342-bp amplification product. Chimeras were bred to C57BL/6 mice and then successfully backcrossed to C57BL/6 for at least seven generations. Copy number of the AGT gene in transgenic and C57 mice was determined by quantitative PCR using human and mouse AGT-specific primers, and animals with single copy genes were selected for future experiments. Initially, four founders from each injection were selected, but after gene copy analysis, one line was established from each group for further work. Nucleotide sequence analysis of the hAGT gene promoter from the tails confirmed that transgenic mice with Ϫ6A haplotype have variants Ϫ6A, Ϫ20A, Ϫ217A, Ϫ532T, Ϫ793A, Ϫ1074T, Ϫ1178G, Ϫ1561T, Ϫ1562C, and Ϫ1670A, whereas transgenic mice with Ϫ6G haplotype have variants Ϫ6G, Ϫ20A, Ϫ217G, Ϫ532C, Ϫ793G, Ϫ1074G, Ϫ1178A, Ϫ1561G, Ϫ1562G, and Ϫ1670G. All experiments involving animals were approved by the institutional review board of the New York Medical College.
Generation of Double Transgenic Mice-Female transgenic mice containing either Ϫ6A or Ϫ6G haplotype of the hAGT gene were crossed with male transgenic mice containing the human renin gene (PAC-hRen) developed by Sinn et al. (25). The genetic background of the PAC-hRen mice was originally B6SJL (C57BL/6J ϫ SJL/JF2), but they have since been backcrossed for at least five generations with C57BL/6J prior to breeding with hAGT mice. The double transgenic mice were genotyped for the human AGT and renin genes by PCR amplification of the DNA isolated from their tails. The primers for the amplification of the hAGT gene were the same as used in the analysis of single transgenic mice. The forward and reverse primers for the amplification of the hRen gene were CTCTTCGATGCTTCGGATTC and TGGCAGAG-TAGGGTGTTCCT, respectively, to produce a 250-bp product.
Quantitative RT-PCR of hAGT and Renin mRNA-Liver and kidney from 8-week-old male transgenic mice containing either Ϫ6A or Ϫ6G haplotype of the hAGT gene and C57 control were harvested following CO 2 asphyxiation and stored in Allprotect tissue reagent (Qiagen). RNA was isolated using the RNeasy Plus minikit (Qiagen). One microgram of RNA was reverse-transcribed into cDNA using the Revert Aid first strand cDNA synthesis kit (Fermentas). Quantitative real-time RT-PCR was performed using Brilliant II SYBR Green QPCR master mix (Stratagene) and a Stratagene thermocycler (Mx3005P, Agilent Technologies). Primers for mouse and human AGT (catalog nos. PPM04219A and PPH01807A), mouse and human renin (catalog nos. PPM0748A and PPH07193A), and mouse GAPDH (catalog no. PPM02946A) genes were purchased from SuperArray Bioscience Corp. Following a 95°C incubation for 10 min, 40 cycles of PCR (95°C for 30 s, 60°C for 30 s) were then performed using 1 l of cDNA, 50 nM PCR primers, and 12.5 l of SYBR Green PCR Master Mix in 25-l reactions. Threshold cycles for three replicate reactions were determined using MxPro-Mx3005P software (version 4.10), and relative transcript abundance was calculated following normalization with human GAPDH.
Western Blotting-Blood samples were collected by cardiac puncture from mice immediately after CO 2 asphyxiation. Approximately 250 l of plasma was collected from each mouse by centrifuging 500 l of blood sample at 12,000 rpm for 15 min at 4°C. Liver protein extracts were prepared from the tissue stored in AllProtect reagent (Qiagen) using a protein isolation kit (Qiagen). Liver protein extracts (25 g) and equal volumes of plasma samples from transgenic mice were fractionated by SDS-PAGE (10% polyacrylamide) in duplicate and transferred to Immobilon-P transfer membrane (Millipore Corp.) to detect hAGT protein. Membranes were blocked in 10% nonfat dry milk (Bio-Rad), and one membrane was immunoblotted with a 1:1500 dilution of commercially available monoclonal antibodies for hAGT (catalog no. H00000183-M01, Abnova Corp.), whereas the other was incubated with a 1:10,000 dilution of mouse albumin antibody (catalog no. A90134P, Bethyl Laboratories). Immune complexes were detected by HRP-conjugated anti-mouse IgG (catalog no. I1904-25C, U.S. Biologicals) using the SuperSignal West Pico chemiluminescence assay (Pierce) according to the manufacturer's protocol. Densitometric analysis of protein bands was performed by Quantity One quantitation software from Bio-Rad.
Immunohistochemistry-Mice were perfused with normal saline. Liver was excised in phosphate-buffered saline (PBS) and fixed in 4% paraformaldehyde-PBS for 3 h at room temperature, kept overnight at 4°C in 30% sucrose solution, and then embedded in OCT compound (Sakura Finetek, Torrence, CA). Croystat sections (8-m thickness) were taken on slides and treated with blocking solution (5% goat serum in PBS with 0.2% Triton X-100) for 1 h at room temperature. Sections were then incubated overnight at 4°C with mouse anti-human primary antibody for AGT (Abnova), followed by washes with PBS (four times for 10 min each) and incubation (2 h) with secondary Cy3-conjugated goat anti-mouse antibody (Abnova) at room temperature. Slides were finally washed with PBS (four times for 10 min each) and mounted. Immunofluorescence was visualized using a Zeiss Axioplan-2 fluorescent microscope. Images were captured and analyzed using AxioVision 2 multichannel image processing software (Zeiss, Göttingen, Germany).
In Vivo ChIP Assay-The ChIP assay was performed using the ChIP assay kit from Upstate Biotechnology, Inc. Mice were perfused with normal saline. Liver was excised and washed in PBS, minced into smaller pieces, fixed with 1% formaldehyde for 20 min at room temperature, and washed with chilled PBS followed by lysis. The DNA was fragmented by sonication, and 10 l of the chromatin solution was saved as input. A 5-g amount of anti-GR or HNF-1␣ or RNApol antibody or rabbit immunoglobulin G was added to tubes containing 900 l of sonicated chromatin solution, and the mixture was incubated overnight at 4°C. The antibody complexes were captured with protein A-agarose beads and subjected to serial washes (as described in the manufacturer's protocol). The chromatin fraction was extracted with SDS buffer and reverse cross-linked at 65°C for 4 -6 h. The DNA was then purified using a Qiagen miniprep column. The immunoprecipitated DNA (1 l) and the input DNA (1 l) were subjected to 35 cycles of PCR amplification using Ϫ314AGTfor (CTCAGGCTGTCACACACCTA) as a forward and Ϫ6AGTrev (TCTTCCCCTGGCCGGGTCACGAT) as a reverse primer when GR or RNA polymerase antibodies were used for immunoprecipitation. This amplified 317-bp fragment spanning the GR and RNA polymerase binding site (located between Ϫ314 and ϩ3) of the human AGT gene promoter. We used Ϫ1757AGTfor (CAGGCACAGTGGAAAC-TCTCC) as a forward primer and Ϫ1554AGTrev (AGTAAC-AAGTCCACCTGGAC) as a reverse primer when HNF-1␣ antibody was used. This amplified the 205-bp fragment containing the HNF-1␣ binding region of the human AGT gene promoter. The PCR-amplified products were analyzed on 2% agarose gel.
Plasma Angiotensin II Level-Plasma angiotensin-II levels were determined by ELISA. Equal amounts of plasma from control (non-transgenic) and double transgenic mice containing either Ϫ6A or Ϫ6G haplotype of the hAGT gene were passed through phenylsilylsilica extraction columns (ALPCO Diagnostics, Salem, NH). The reversed phase extraction method was applied to extract the plasma as per the manufacturer's protocol. A peptide enzyme immunoassay for Ang II was performed in extracted plasma samples using an enzyme immunoassay kit (Peninsula Laboratories) as per the manufacturer's protocol. Plasma angiotensin-II levels were calculated as described by the manufacturer using a standard curve.
Blood Pressure Measurement-All mice were fed with standard mice chow and had access to water ad libitum. Blood pressure (BP) was measured in the conscious state by telemetry. A radiotelemetric system from Data Science International (St. Paul, MN) was used for this procedure. Briefly, mice were anesthetized with ketamine and xylazine (90 and 10 mg/kg, respectively), and the left carotid artery was isolated. The tip of the telemetric catheter (model TA11PA-C10) was then inserted into the carotid artery and advanced into the aortic arch, with the telemetric device main body positioned into a subcutaneous pocket into the right flank. After 1 week of recovery from the surgical procedure, BP readings were recorded every 10 min using a Data-Science International instrument as described previously (26). Mean BP values were calculated for every hour from the values taken over 4 days. Animals were then treated with angiotensin receptor blocker, losartan (Sigma-Aldrich) (30 mg/kg/day), in drinking water. The blood pressure was then measured 24 h after losartan treatment.
Statistical Analysis-The unpaired t test was used to compare relative luciferase activities of reporter constructs in transient transfection experiments. All experiments were conducted in triplicate in three independent transfections. Data are expressed as means Ϯ S.E. from each group containing six mice. Differences between group means were determined by a two-factor analysis of variance, followed by a Newman-Keuls post hoc analysis (NCSS 2007); p Ͻ 0.05 was considered sig-nificant (NCSS LLC, Kaysville, UT). Statistically significant results are marked by an asterisk (p Ͻ 0.05).

RESULTS
Variant Ϫ6A of the hAGT Gene Is Associated with Hypertension in Caucasian Subjects-In order to confirm a previous observation that Ϫ6A allele is associated with hypertension in Caucasian subjects, we analyzed genomic DNA from 293 Caucasian subjects with normal blood pressure and 354 Caucasian subjects with hypertension (mean age 59 Ϯ 10 years, equal number of male and female subjects) for A/G polymorphism at Ϫ6 of the hAGT gene. All of these subjects were recruited from the outpatient department of Westchester Medical Center (Valhalla, NY). All case and control subjects gave informed consent before participating in the research. The research protocol was approved by the institutional review board at New York Medical College. All cases were diagnosed as having primary hypertension, and patients with secondary hypertension, diabetes mellitus, or ischemic heart disease were excluded. The criteria for hypertension were defined as a systolic BP of Ͼ140 mm Hg, a diastolic BP of Ͼ90 mm Hg, or under antihypertensive therapy. Blood pressure was measured twice with the subject seated with a 5-min interval. The normotensives (with systolic BP/diastolic BP of Ͻ140/90 mm Hg) without a history of hypertension and without diabetes mellitus were recruited from the same population and matched for sex and age. All participants completed a standard questionnaire on personal medical history and family history of hypertension. Genomic DNA was isolated from peripheral blood samples, and Ϫ6A/G polymorphism of the hAGT gene was analyzed by PCR amplification of DNA followed by BstNI restriction analysis as described (9). All patients and control subjects were in Hardy-Weinberg equilibrium. The frequency of the Ϫ6A allele in hypertensive patients was 0.51 as compared with 0.40 in the normotensive population, which is highly significant (p ϭ Ͻ.001 and OR ϭ 2.2, confidence interval 1.49 -3.31). Statistical analysis based on genotype also suggested a highly significant role of the Ϫ6A allele in hypertension in this cohort (p Ͻ 0.001).

Upstream Region of the hAGT Gene Promoter Increases
Glucocorticoid-induced Promoter Activity-Because previous studies have shown that glucocorticoid treatment increases expression of the AGT gene in liver (20,27), we were interested in identifying glucocorticoid-responsive elements in the promoter of the hAGT gene. We therefore synthesized reporter constructs that contained either 1253 bp (pHAGT1.2-luc) or 2572 bp (pHAGT2.6-luc) of the hAGT gene promoter attached to the luciferase gene. These reporter constructs were then transiently transfected in HepG2 cells along with RSV-GR (an expression vector containing rat GR coding sequence), and promoter activity was analyzed after dexamethasone treatment as described previously (27)(28)(29). Results of this experiment showed that whereas the reporter construct with the 1.2-kb promoter gave only 2.3-fold glucocorticoid-induced promoter activity, the reporter construct with the 2.6-kb promoter gave 24-fold increased promoter activity upon transient transfection in human liver cells (Fig. 1A). Similarly, the reporter construct with the 1.2-kb promoter gave only 4.5fold glucocorticoid-induced promoter activity, whereas the reporter construct with the 2.6-kb promoter gave 37-fold glucocorticoid-induced promoter activity upon transient transfection in human kidney cells (Fig. 1B). This experiment suggested that nucleotide sequence located between 1.3 and 2.6 kb of the promoter is responsible for glucocorticoid-induced promoter activity of the hAGT gene.
Deletion Constructs Show That Glucocorticoid-induced Promoter Activity Is Located between Ϫ1541 and Ϫ1748-In order to identify the promoter region responsible for glucocorticoid-induced promoter activity of the hAGT gene, we synthesized deletion constructs using pHAGT2.6-luc. These constructs contained 1748, 1651, 1618, 1598, 1541, and 1457 nucleotides from the transcriptional initiation sites attached to the luciferase gene (supplemental Fig. 1A). These deletion constructs were then used for transient transfection in HepG2 cells in the absence and presence of dexamethasone. Results of this experiment suggested that nucleotide sequence located between 1748 and 1541 is mainly responsible for the glucocorticoid-induced promoter activity (supplemental Fig. 1B).
Nucleotide Sequence Located between Ϫ1541 and Ϫ1748 Contains Binding Sites for Multiple Liver-enriched Transcription Factors-In order to identify nucleotide sequence that may be involved in glucocorticoid-induced promoter activity, we determined the nucleotide sequence of the hAGT gene located between Ϫ1748 and Ϫ1541. The nucleotide sequence of this region of the promoter revealed homology with binding sites of multiple liver-enriched transcription factors, such as HNF-3, HNF-1, and C/EBP, as well as GRE and SP1 (supplemental Fig. 2A). Binding of multiple transcription factors in this region of the promoter may be necessary for glucocorticoid-induced expression of the hAGT gene because previous studies have suggested that interaction of multiple transcription factors is crucial for glucocorticoid-induced promoter activity of a gene (30 -32).
Nucleotide Sequence Located between Ϫ1541 and Ϫ1748 of the hAGT Gene Contains Three Additional SNP Sites-Previous studies have shown that the hAGT gene has eight SNPs in the 1.2-kb promoter (located at Ϫ6, Ϫ20, Ϫ217, Ϫ532, Ϫ776, Ϫ793, Ϫ1074, and Ϫ1178) that are involved in transcriptional regulation of this gene. We were therefore interested in examining whether nucleotide sequence located between Ϫ1748 and Ϫ1541 of the hAGT gene promoter contains additional SNPs that may modulate its transcriptional activity. We therefore determined nucleotide sequence of this region of the promoter from five normotensive and five hypertensive subjects by amplification and direct sequencing of this region of the promoter. Results of this experiment showed that this region of the promoter has three new SNPs (A/G at Ϫ1670, C/G at Ϫ1562, and T/G at Ϫ1561) (supplemental Fig. 2B). The position of these SNPs is also shown by an asterisk at the top of each nucleoside (supplemental Fig. 2A).
Variant Ϫ6A Almost Always Occurs with Variants Ϫ1561T, Ϫ1562C, and Ϫ1670A-Because variant Ϫ6A of the hAGT gene is associated with increased plasma AGT level and increased blood pressure, we were interested in finding whether Ϫ6A is in linkage disequilibrium with three new SNPs identified in the upstream region of the hAGT gene promoter. We therefore sequenced DNA from 100 subjects in these regions of the promoter. Results of this analysis suggested that variant Ϫ6A always occurred with variants Ϫ1561T, Ϫ1562C, and Ϫ1670A, and variant Ϫ6G always occurred with Ϫ1561G, Ϫ1562G, and Ϫ1670G (only four samples with Ϫ6A had Ϫ1561G and three samples had Ϫ6G and Ϫ1561T). Therefore, the hAGT gene may be subdivided into two haplotypes: Ϫ6A haplotype containing variants Ϫ6A, Ϫ1561T, Ϫ1562C, and Ϫ1670A; and Ϫ6G haplotype containing variants Ϫ6G, Ϫ1561G, Ϫ1562G, and Ϫ1670G.
Nucleotide Sequence of the hAGT Gene Containing Ϫ1670A Binds Strongly to Liver-enriched Transcription Factor HNF-1␣ as Compared with Ϫ1670G-Because an A/G polymorphic site is located between Ϫ1748 and Ϫ1651 of the hAGT gene promoter and deletion of this region affects glucocorticoidinduced promoter activity, we were interested in finding whether nucleotide sequence around this polymorphic site binds to a transcription factor. Transfac analysis suggested that indeed nucleotide sequence containing Ϫ1670A has stronger homology with the consensus HNF-1 binding site as compared with Ϫ1670G ( Fig. 2A). In order to confirm that HNF-1 binds to this region of the promoter, we performed a ChIP assay using HepG2 cells. Results of this assay show that HNF-1␣ binds to this region of the promoter (Fig. 2B). In order to analyze whether variants Ϫ1670A and Ϫ1670G bind differentially to HNF-1␣, we performed an EMSA using radiolabeled oligonucleotides containing Ϫ1670A in the presence of HepG2 nuclear extract. Results of this experiment (Fig. 2C) showed that the complex formed by radiolabeled oligonucleotide containing consensus HNF-1 site (lane 1)   DECEMBER 24, 2010 • VOLUME 285 • NUMBER 52

JOURNAL OF BIOLOGICAL CHEMISTRY 41177
Nucleotide Sequence Located between Ϫ1624 and Ϫ1604 Contains a C/EBP Binding Site That Plays an Important Role in Glucocorticoid-induced Promoter Activity-Our transient transfection assay using different deletion constructs showed that removal of nucleotide sequence between Ϫ1618 and Ϫ1598 reduces glucocorticoid-induced promoter activity (supplemental Fig. 1). Nucleotide sequence located in this region of the promoter has strong homology with the C/EBP site (Fig. 3A) (core binding sequence is underlined). Our ChIP assay showed that C/EBP␤ binds to this region of the promoter (Fig. 3B). In order to further confirm that this sequence binds to C/EBP, we performed an EMSA (Fig. 3C). A complex obtained with radiolabeled oligonucleotide containing the consensus C/EBP site in the presence of HepG2 extract (lane 1) was removed in the presence of a 20-fold excess of cold self-oligonucleotide (lane 2) and a 50-or 100-fold excess of cold oligonucleotide containing nucleotide sequence located between Ϫ1606 and Ϫ1614 (labeled CEBP 1606) (lanes 3 and 4) but only slightly reduced in the presence of mutated CEBP1606 (lanes 5 and 6); this complex was supershifted in the presence of C/EBP␤ antibody (lane 7) but not in the presence of preimmune serum (lane 8). In another experiment, radiolabeled oligonucleotide CEBP1606 formed a complex with HepG2 extract (Fig. 3D, lane 1), which was removed in the presence of a 50-or 100-fold excess of cold self-oligonucleotide (lanes 2 and 3) but only slightly reduced in the presence of mutated CEBP 1606 (lanes 4 and 5); this complex was not removed or supershifted in the presence of preimmune serum (lane 6) but was supershifted in the presence of C/EBP␤ antibody (lane 7) and removed in the presence of cold oligonucleotide containing consensus C/EBP site (lane 8). Results of these experiments suggested that nucleotide sequence located between Ϫ1606 and Ϫ1614 binds to the C/EBP family of transcription factors. In order to confirm the role of this C/EBP binding site in glucocorticoid-induced promoter activity of the hAGT gene, we performed site-specific mutagenesis in the reporter construct pHAGT1618-luc to mutate the C/EBP site and performed a transient transfection assay. Results of this experiment show that whereas wild type construct had 15-fold dexamethasoneinduced promoter activity, reporter construct containing mutation of the C/EBP site located between Ϫ1606 and Ϫ1614 had only 5-fold dexamethasone-induced promoter activity (Fig. 3E). These experiments confirmed that the C/EBP site located around Ϫ1610 is necessary for glucocorticoid-induced promoter activity of the hAGT gene.
Nucleotide Sequence Containing SNPs at Ϫ1561 and Ϫ1562 Has Homology with GRE-We were next interested in determining whether nucleotide sequence of the hAGT gene promoter containing SNPs at Ϫ1562 and Ϫ1561 binds to a transcription factor that may modulate glucocorticoid-induced promoter activity. Nucleotide sequence containing Ϫ1562C and Ϫ1561T (that is in linkage disequilibrium with Ϫ6A) (Fig.   FIGURE 3. A, sequence homology between the consensus C/EBP binding site and Ϫ1600 region of the hAGT gene promoter. Line 1, consensus C/EBP binding site; line 2, nucleotide sequence located between Ϫ1596 and Ϫ1613 of the hAGT gene; line 3, oligonucleotide sequence containing mutated C/EBP binding site. B, ChIP assay shows that C/EBP binds to the Ϫ1600 region of the hAGT gene promoter in HepG2 cells. The ChIP assay was performed as described in the legend to Fig. 6B except that C/EBP␤ antibody was used for immunoprecipitation. Shown are PCR product obtained in the presence of C/EBP␤ antibody (lane 1), PCR product in the presence of preimmune serum (lane 2), PCR product obtained in the presence of C/EBP␤ antibody from another region of the promoter that does not contain C/EBP site (lane 3), and PCR product obtained from the genomic DNA without immunoprecipitation (positive control) (lane 4). C, EMSA confirms that nucleotide sequence located between Ϫ1596 and Ϫ1613 of the hAGT gene promoter competes with the C/EBP binding site. EMSA was performed using radiolabeled oligonucleotide containing the consensus C/EBP binding site in the presence of HepG2 extract and competed with an oligonucleotide containing either wild type or mutated C/EBP site located in the Ϫ1596 to Ϫ1613 region of the hAGT gene promoter. Lane 1, complex obtained in the presence of radiolabeled oligonucleotide containing consensus C/EBP site; lane 2, competition with a 20-fold excess of cold C/EBP consensus oligonucleotide; lanes 3-6, competition in the presence of a 50-and 100-fold excess of an oligonucleotide containing either wild type or mutant C/EBP sequence from the hAGT gene; lane 7, reaction in the presence of C/EBP␤ antibody; lane 8, reaction in the presence of preimmune serum (PIS). D, EMSA confirms that nucleotide sequence located between Ϫ1596 and Ϫ1613 of the hAGT gene promoter binds to C/EBP␤. EMSA was performed using radiolabeled oligonucleotide containing the C/EBP site located between Ϫ1596 and Ϫ1613 of the hAGT gene promoter in the presence of HepG2 extract. Lane 1, complex obtained in the presence of radiolabeled oligonucleotide alone; lanes 2-5, competition in the presence of a 50-and 100fold excess of wild type and mutant cold C/EBP oligonucleotides; lane 6, reaction in the presence of preimmune serum; lane 7, reaction in the presence of C/EBP␤ antibody; lane 8, reaction in the presence of a 50-fold excess of cold consensus C/EBP oligonucleotide. E, mutation of C/EBP site located between Ϫ1596 and Ϫ1613 of the hAGT gene promoter reduces glucocorticoid-induced promoter activity. Reporter constructs containing 1618 bp of the hAGT gene promoter with either wild type or mutated C/EBP site (located between Ϫ1596 and Ϫ1613) were transiently transfected in HepG2 cells, and glucocorticoid-induced promoter activity was analyzed as described above. The empty box shows glucocorticoid-induced promoter activity of the wild type construct, and the filled box shows glucocorticoid-induced promoter activity of the mutated construct. Error bars, S.E.
4A, lane 2) has stronger homology with consensus GRE (Fig.  4A, lane 1) as compared with the nucleotide sequence containing Ϫ1562G and Ϫ1561G (that is in linkage disequilibrium with Ϫ6G) (Fig. 4A, lane 3). We next performed a ChIP assay using GR antibody and HepG2 cells. Results of this experiment confirmed that indeed GR binds to this region of the promoter (Fig. 4B). Because variants Ϫ1562C and Ϫ1561T had stronger homology with GRE, we wanted to examine their role in GR-induced promoter activity. We therefore mutated Ϫ1562C to Ϫ1562G and Ϫ1561T to Ϫ1561G by sitespecific mutagenesis in the reporter construct containing pHAGT1618-luc and performed transient transfection in HepG2 cells. Results of this experiment showed that mutation of Ϫ1562C and Ϫ1561T reduced glucocorticoid-induced promoter activity from 13-fold to 8-fold (Fig. 4C).
Reporter Construct Containing Haplotype Ϫ6A of hAGT Gene Has Increased Basal and Glucocorticoid-induced Promoter Activity as Compared with Haplotype Ϫ6G upon Transient Transfection in Human Liver and Kidney Cells-In order to understand the role of Ϫ6A and Ϫ6G alleles in the context of the 1.8-kb promoter of the hAGT gene upon transcriptional regulation, we synthesized reporter constructs using genomic DNA from a hypertensive subject (that contained Ϫ6A allele) and a normotensive subject (that contained Ϫ6G allele). Because previous studies have shown that the hAGT gene contains eight polymorphic sites in the 1.2-kb region of the promoter, we determined the nucleotide sequence of the entire region of the promoter present in these constructs. The positions of different SNPs in the 1.8-kb promoter of reporter construct containing Ϫ6A haplotype and Ϫ6G haplotype are shown in Fig.  5A. These reporter constructs were then transiently transfected in human liver cells (HepG2) and human kidney (HK-2) cells to determine their basal and glucocorticoid-induced promoter activity. Basal and glucocorticoid-induced promoter activity of reporter construct pHAGT1748-luc containing Ϫ6A and Ϫ6G haplotypes in HepG2 cells is shown in Fig. 5, B and C, and pro-moter activity in HK-2 cells is shown in Fig. 5, D and E. Results of these experiments showed that the reporter construct containing 1.8 kb of the hAGT gene promoter with Ϫ6A haplotype has 40 -50% increased basal promoter activity and 4 -6-fold increased glucocorticoid-induced promoter activity upon transient transfection in human liver and kidney cells as compared with Ϫ6G haplotype.
Angiotensinogen mRNA Is Increased in the Liver and Kidney of Transgenic Mice Containing Ϫ6A Haplotype of the hAGT Gene as Compared with the Ϫ6G Haplotype-In order to understand the role of Ϫ6A and Ϫ6G haplotypes in the expression of the hAGT gene in an in vivo situation, we generated double transgenic mice containing the human renin gene and either Ϫ6A or Ϫ6G haplotype of the hAGT gene as described under "Materials and Methods." We started with a 180-kb BAC containing Ϫ6G haplotype of the hAGT gene (supplemental Fig. 3). This BAC was recombineered to produce another BAC with Ϫ6A haplotype of the hAGT gene. These BACs were then used to generate transgenic mice containing either Ϫ6A or Ϫ6G haplotype of the hAGT gene. These mice were then crossed with transgenic mice containing the human renin gene to obtain double transgenic mice. Because liver and kidney are the most important sites for the expression of AGT gene, we examined the effect of Ϫ6A and Ϫ6G haplotypes on hAGT mRNA level in these tissues of transgenic mice by quantitative RT-PCR. Quantitation of RNA level in Ϫ6A haplotype was performed using RNA level in Ϫ6G haplotype as 1. Results of these experiments show that the AGT mRNA level in the liver of transgenic mice containing Ϫ6A haplotype is increased about 2.5-fold, and in the kidney, it was increased 2.1-fold as compared with the transgenic mice containing Ϫ6G haplotype (Fig. 6A). We also confirmed the overexpression of hAGT protein in the liver of Ϫ6A haplotype compared with Ϫ6G haplotype by immunohistochemistry. Results of this experiment showed that the hAGT gene is expressed more abundantly in the liver from a transgenic animal  DECEMBER 24, 2010 • VOLUME 285 • NUMBER 52

JOURNAL OF BIOLOGICAL CHEMISTRY 41179
containing Ϫ6A haplotype as compared with Ϫ6G haplotype (Fig. 6B). Finally, we examined the AGT protein level in the liver of transgenic mice by Western blot analysis. A typical Western blot is shown in Fig. 6C, and quantitation of AGT protein level normalized to the ␤-actin is shown in Fig. 6D. hAGT mRNA level was increased about 2.5-fold in the liver of transgenic mice containing Ϫ6A haplotype as compared with Ϫ6G haplotype, and AGT protein level was increased about 1.6-fold.
Chromatin Immunoprecipitation Assay Shows That HNF-1␣, GR, and RNA Polymerase Have Higher Affinity for the Chromatin of Liver from Transgenic Mice Containing Ϫ6A Haplotype of the hAGT Gene as Compared with Ϫ6G Haplotype-Because (a) our studies described here have shown that nucleotide se-quence of hAGT gene promoter containing Ϫ1670A has stronger homology with the HNF-1 consensus binding site as compared with Ϫ1670G, (b) our studies here have shown that nucleotide sequence containing Ϫ1562C and Ϫ1561T has stronger homology with consensus GRE as compared with Ϫ1562G and Ϫ1561G, and (c) previous studies have shown that nucleotide sequence containing Ϫ217A has stronger affinity for GR as compared with Ϫ217G (29), we wanted to confirm the effect of these polymorphisms on the binding of their cognate transcription factors using chromatin from the liver of double transgenic mice containing either the Ϫ6A or Ϫ6G haplotype of the hAGT gene. Results of our in vivo ChIP assay showed that HNF-1␣ binds more strongly to the nucle- otide sequence around the Ϫ1670 region in the chromatin from the liver of transgenic mice containing Ϫ6A haplotype of the hAGT gene as compared with the Ϫ6G haplotype (Fig.  7a, compare lanes 1 and 2). Similarly, GR binds more strongly to the Ϫ1550 and Ϫ220 regions of the hAGT gene promoter in the chromatin of the liver from transgenic mice containing Ϫ6A haplotype as compared with Ϫ6G haplotype (Fig.  7, b and c). GR also binds to the Ϫ220 region of the promoter because we have previously shown that the hAGT gene has an A/G polymorphism at the Ϫ217-position, and GR shows stronger affinity for an oligonucleotide containing Ϫ217A as compared with Ϫ217G (29). Finally, we performed a ChIP assay using antibody against RNA polymerase to examine whether it binds more strongly to the hAGT gene promoter in the chromatin of liver containing Ϫ6A haplotype as compared with Ϫ6G haplotype. Results of this experiment showed that indeed RNA polymerase binds more strongly to the hAGT gene promoter in the chromatin of liver of transgenic mice containing Ϫ6A haplotype as compared with Ϫ6G haplotype (Fig. 7d). Taken together, results of this experiment confirmed that RNA polymerase, HNF-1␣, and GR have stronger affinity for the hAGT gene promoter in the chromatin obtained from the liver of transgenic mice containing Ϫ6A haplotype as compared with Ϫ6G haplotype.  . ChIP assay shows that GR, HNF-1␣, and RNA polymerase bind strongly to the human AGT promoter from liver chromatin of transgenic mice containing ؊6A haplotype as compared with ؊6G haplotype. A ChIP assay was performed to examine the binding of HNF-1␣, GR, and RNA polymerase to the liver chromatin in double transgenic mice containing Ϫ6A and Ϫ6G haplotypes of the hAGT gene. Immunoprecipitated DNA from the two haplotypes in the presence of antibodies against HNF-1␣ (a), GR (b and c), and RNA polymerase (d) was used to amplify nucleotide sequence containing their respective binding sites as described under "Materials and Methods." In all panels (a-d), lanes 1 and 2, PCR-amplified product obtained from the ChIP of Ϫ6A and Ϫ6G haplotypes, respectively; lanes 3 and 4, PCR-amplified product using input genomic DNA from Ϫ6A and Ϫ6G haplotypes; lanes 5 and 6, PCR-amplified product in the presence of rabbit IgG for Ϫ6A and Ϫ6G haplotype, respectively. Chip assays were performed in triplicate.

Plasma Angiotensinogen Level Is Increased in Transgenic
Mice Containing Ϫ6A Haplotype of the hAGT Gene as Compared with the Ϫ6G Haplotype-We next determined if (a) hAGT protein in transgenic mice is correctly processed in the liver and secreted into plasma and (b) plasma AGT level is increased in transgenic mice containing Ϫ6A haplotype of the hAGT gene as compared with the Ϫ6G haplotype. Equal volumes of plasma from control C57 mice and male transgenic animals containing either Ϫ6A or Ϫ6G haplotype of the hAGT gene were analyzed by a Western blot assay along with human plasma as described (19) (Fig. 8A). Quantitation of AGT level was performed after normalization with mouse albumin (Fig. 8B). Results of this experiment showed that AGT level is increased by about 40% in the plasma of transgenic mice containing Ϫ6G haplotype as compared with the non-transgenic control animals. The plasma AGT level was increased by about 70% in transgenic mice containing Ϫ6A haplotype as compared with Ϫ6G haplotype (p ϭ Ͻ 0.05). Electrophoretic mobility of hAGT protein in transgenic mice was similar to that of AGT in human plasma.
Plasma Angiotensin-II Level Is Increased in Male Transgenic Mice Containing Ϫ6A Haplotype of the hAGT Gene as Compared with the Ϫ6G Haplotype-We next examined the effect of Ϫ6A and Ϫ6G haplotypes of the hAGT gene on plasma angiotensin-II level in double transgenic mice. Plasma was collected from control (non-transgenic), and male double transgenic mice containing either Ϫ6A or Ϫ6G haplotype of the hAGT gene, and plasma angiotensin-II levels were analyzed by ELISA. Results of this experiment (supplemental Fig.  4) showed that the plasma angiotensin-II level in transgenic mice containing Ϫ6G haplotype was increased by about 50% as compared with the non-transgenic littermate. On the other hand, plasma angiotensin-II level in transgenic mice containing Ϫ6A haplotype was increased by 100% as compared with the non-transgenic littermate (p ϭ Ͻ0.05). Thus, plasma angiotensin-II level was increased by about 50% in double transgenic mice containing Ϫ6A haplotype of the hAGT gene as compared with Ϫ6G haplotype. Plasma angiotensin-II levels in single transgenic mice containing either the hAGT or hRen gene were the same as in control C57 mice (data not shown).

Mouse and Human Renin Level Is Decreased in Double Transgenic Mice Containing hRen and either Ϫ6A or Ϫ6G
Haplotype of the hAGT Gene-Because it is known that an increased amount of AGT down-regulates renin expression, we were interested in examining the effect of Ϫ6A and Ϫ6G haplotypes of the hAGT gene on mouse and human renin gene expression in double transgenic mice. For this purpose, we analyzed the mRNA level of endogenous mouse renin and human renin gene in the kidney of non-transgenic and transgenic mice by quantitative RT-PCR. Results of this experiment showed that mouse renin mRNA level in transgenic mice containing Ϫ6G haplotype of the hAGT gene was about 40% less as compared with control C57 animals. On the other hand, renin mRNA level in transgenic mice containing Ϫ6A haplotype was 60% less as compared with control C57 animals (supplemental Fig. 5A). Human renin mRNA level in transgenic mice containing Ϫ6G haplotype of the hAGT gene was about 20% less and the mRNA level in Ϫ6A haplotype was 40% less as compared with control PAC-hRen animals (supplemental Fig. 5B). This suggested that although mouse as well as human renin gene expression is down-regulated by overexpression of the hAGT gene in transgenic mice, this effect is more pronounced in transgenic mice containing Ϫ6A haplotype as compared with the Ϫ6G haplotype.
Blood Pressure Is Increased in Transgenic Mice Containing Ϫ6A Haplotype of the hAGT Gene as Compared with Ϫ6G Haplotype-Finally, we measured the effect of Ϫ6A and Ϫ6G haplotypes of the hAGT gene on blood pressure in double transgenic conscious male mice for 24 h using telemetry over a period of 2 weeks. Control C57 (non-transgenic) and double transgenic mice containing either Ϫ6A or Ϫ6G haplotype of the hAGT gene and hRen gene (3-month-old male mice) were surgically instrumented with telemetric probes. After 1 week of recovery from the surgical procedure, BP readings were recorded every 10 min using a Data-Science International instrument. Systolic blood pressure was measured over a period of 4 days, and mean values for daytime (6 a.m. to 6 p.m.) and nighttime (6 p.m. to 6 a.m.) as well as the average for the whole day (24 h) were calculated. A graph showing the mean BP values over a 24-h period of the Ϫ6A and Ϫ6G haplotypes of transgenic mice (number of animals ϭ 6, days of recording ϭ 4) is shown in Fig. 9A. Mean BP of control non-transgenic mice was 114 mm Hg (Con), mean BP of transgenic mice containing Ϫ6G haplotype was 127 mm Hg, and mean BP of transgenic mice containing Ϫ6A haplotype was 138 mm Hg (n ϭ 6) (p Ͻ 0.05). Taken together, the results of this experiment showed that blood pressure of transgenic mice containing Ϫ6A haplotype was increased by 11 mm Hg as compared with transgenic mice containing Ϫ6G haplotype of the hAGT gene. Blood pressure of single transgenic mice containing either the hAGT or hRen gene was the same as that of control C57 mice. Because our experiments so far suggested that hAGT gene expression is increased in transgenic mice containing Ϫ6A haplotype that leads to increased plasma angiotensin-II level and increased blood pressure as compared with transgenic animals containing Ϫ6G haplotype, we next examined whether angiotensin receptor blockers reduce blood pressure in these animals. We used losartan as an AT1 receptor blocker and determined the blood pressure by the tail cuff method. Results from this experiment show that losartan reduces the systolic blood pressure in double transgenic mice containing both Ϫ6A and Ϫ6G haplotypes of the hAGT gene to or below the non-transgenic base-line level (Fig. 9B).

DISCUSSION
Although Ϫ6A allele of the hAGT gene is associated with increased plasma AGT level and increased blood pressure in humans, the molecular mechanism involved in this process is not clear. Previous in vitro and in vivo studies have shown that glucocorticoids play an important role in the expression of the AGT gene in the liver (33) and are the greatest inducers of its expression in adipocytes (34). Glucocorticoid treatment increases blood pressure in humans (35), and increased cortisol level is a marker for human hypertension (36). We have found that a reporter construct containing 1.8 kb of the hAGT gene promoter has about 10-fold increased glucocorticoid-induced promoter activity on transient transfection in human liver and kidney cells as compared with the reporter construct containing 1.2 kb of the promoter. Surprisingly, the hAGT gene has three new SNPs in this region of the promoter (A/G at Ϫ1670, C/G at Ϫ1562, and T/G at Ϫ1561). Variants Ϫ1670A, Ϫ1562C, and Ϫ1561T almost always occur with Ϫ6A; and variants Ϫ1670G, Ϫ1562G, and Ϫ1561G almost always occur with Ϫ6G. Therefore, the hAGT gene may be subdivided into Ϫ6A and Ϫ6G haplotypes. We have shown that a reporter construct containing 1.8 kb of the hAGT gene promoter with Ϫ6A haplotype has 50 -60% increased basal promoter activity and 4 -6-fold increased glucocorticoidinduced promoter activity upon transient transfection in human liver and kidney cells as compared with a reporter construct with Ϫ6G haplotype.
We have also shown that an oligonucleotide containing Ϫ1670A has stronger affinity for HNF-1␣ as compared with the same oligonucleotide containing Ϫ1670G, and deletion of the HNF-1␣ binding site reduces basal and glucocorticoid induced promoter activity upon transient transfection in liver cells. Although HNF-1 was originally discovered as a liverspecific transcription factor, recent studies have suggested that it is critical for diverse metabolic functions in pancreatic islets, liver, intestine, and kidneys (37). An intact and occupied HNF-1 site plays a crucial role in basal and IL-6-induced expression of the phosphoenolpyruvate carboxykinase gene in the kidney (38,39). Therefore, variant Ϫ1670A (which always occurs with Ϫ6A) may increase the expression of the hAGT gene in the liver as well as the kidney of human subjects containing Ϫ6A haplotype by increased binding of HNF-1␣ to nucleotide sequence containing Ϫ1670A.
The upstream region of the hAGT gene promoter has two polymorphic sites, T/G at Ϫ1561 and C/G at Ϫ1562, and nucleotide sequence containing Ϫ1562C and Ϫ1561T has stronger homology with consensus GRE as compared with the nucleotide sequence containing Ϫ1562G and Ϫ1561G. Our transient transfections show that reporter construct containing Ϫ1562C and Ϫ1561T has increased glucocorticoid-induced promoter activity as compared with the reporter construct containing Ϫ1562G and Ϫ1561G. We have previously shown that the hAGT gene has an A/G polymorphic site at Ϫ217, and reporter construct containing 1.2 kb of the promoter with Ϫ217A has increased glucocorticoid induced promoter activity as compared with Ϫ217G (29). Because transient transfection of reporter construct containing 1.8 kb of the promoter has 10-fold increased GR-induced promoter activity as compared with the reporter construct containing the 1.2-kb promoter, it seems that upstream GRE (located around Ϫ1560) is required for full GR-induced promoter ac- FIGURE 9. A, blood pressure is increased in double transgenic mice containing Ϫ6A haplotype of the hAGT gene as compared with Ϫ6G haplotype. The mean of 24 h blood pressure of control (CON) and double transgenic mice containing Ϫ6G or Ϫ6A haplotype of the hAGT gene was measured by telemetry for 4 days. Each bar represents mean from six male animals taken over 4 days. B, blood pressure in transgenic mice is reduced by losartan treatment. Shown is the mean of 24 h blood pressure of double transgenic mice containing Ϫ6G or Ϫ6A haplotypes in the presence (ϩ) and absence (Ϫ) of losartan. Error bars, S.E.
tivity. Previous studies have suggested that GREs may exist in the upstream region of the promoter and act as an enhancer. For example, a GRE is located (a) 6 kb upstream from the transcriptional initiation site in mouse HNF-4␣1 promoter (40), (b) 11 kb upstream from the rat tyrosine aminotransferase promoter (41), and (c) 1.4 kb from the rat phosphoenolpyruvate carboxykinase gene promoter (30,38). Our observation that deletion of the HNF-1 binding site around Ϫ1670 reduces GR-induced promoter activity suggests that an intact HNF-1 site is also required for glucocorticoid-induced expression of the hAGT gene. This observation is in accordance with previous studies suggesting the role of HNF-1 in GRinduced promoter activity of the HNF-4␣1 gene (42). In addition, the HNF-1 binding site is required for glucocorticoidinduced transcription of the insulin-like growth factor binding protein-1 gene (43) and the phosphoenolpyruvate carboxykinase gene in the kidney (39).
Another finding of this study is that deletion of a 20-bp region of the promoter located between Ϫ1600 and Ϫ1620 drastically reduces glucocorticoid-induced promoter activity. This region of the promoter binds to the C/EBP family of transcription factors. Site-specific mutagenesis of this C/EBP binding site reduces GR-induced promoter activity and shows that this C/EBP site is important for glucocorticoid-induced expression of the hAGT gene. C/EBP family of transcription factors play an important role in liver-and adipose-specific gene expression, two tissues in which the AGT gene is abundantly expressed (44,45). Glucocorticoids also induce the expression of C/EBP␤ and C/EBP␦ in hepatoma cells (46), primary culture of hepatocytes (47), and adipocytes (48). In addition, GR physically interacts with C/EBP␤ and increases expression of a gene that contains the C/EBP binding site (32). C/EBP binding sites have been shown to be essential for GR-induced promoter activity of mouse HNF-4␣1 (40) and arginase (32) genes.
In order to unambiguously prove the role of Ϫ6A allele in the expression of the hAGT gene in a tissue-specific manner and in the regulation of blood pressure in an in vivo situation, we have made transgenic mice containing either Ϫ6A or Ϫ6G haplotype of the hAGT gene present in 180-kb-long BAC DNA (containing 116 kb of the 5Ј-flanking region, all five exons and four introns, and 54 kb of the 3Ј-flanking region of the hAGT gene). The reason behind using BAC DNA was that because it contains 116 kb of the 5Ј-flanking region, it will have all of the cis-acting DNA elements required for tissue-and hormone-specific expression of the hAGT gene. Because there is a strict species specificity in the renin cleavage site and mouse renin is not able to cleave hAGT, we also made double transgenic mice containing the human renin gene and either the Ϫ6A or Ϫ6G haplotype of the hAGT gene. We show here that hAGT mRNA level is increased about 2.5-fold in the liver of male double transgenic mice containing Ϫ6A haplotype as compared with Ϫ6G haplotype. In order to confirm that an increase in mRNA level correspondingly increases its protein level, we also analyzed AGT protein level by Western analysis. Our results indicate that AGT protein level is also increased in the liver of transgenic mice containing Ϫ6A haplotype by about 1.6-fold as com-pared with the liver of transgenic mice containing Ϫ6G haplotype.
Because our EMSA experiment has shown that an oligonucleotide containing Ϫ1670A has stronger affinity with HNF-1␣ as compared with the same oligonucleotide containing Ϫ1670G, we performed a ChIP assay to confirm that HNF-1␣ binds strongly to the chromatin obtained from the liver of transgenic mice containing Ϫ6A haplotype as compared with Ϫ6G haplotype. Results of our in vivo ChIP assay have shown that indeed HNF-1␣ binds more strongly to the hAGT gene promoter in the chromatin obtained from the liver of transgenic mice containing Ϫ6A haplotype of the hAGT gene as compared with Ϫ6G haplotype. Our experiments suggest that transcription factors present in mouse cells correctly recognize the changes in cis-acting DNA elements present in the chromatin of two different haplotypes of a human gene and differentially regulate their transcription.
Because nucleotide sequence of the hAGT gene promoter containing Ϫ1562C and Ϫ1561T has stronger homology with consensus GRE as compared with Ϫ1562G and Ϫ1561G, we performed a ChIP assay to analyze the binding of GR in liver chromatin of transgenic mice containing either Ϫ6A or Ϫ6G haplotype of the hAGT gene. Our in vivo ChIP assay has shown that GR has stronger affinity for the chromatin obtained from the liver of transgenic animals containing Ϫ6A haplotype as compared with Ϫ6G haplotype. Glucocorticoid treatment increases blood pressure in humans (35), and increased cortisol level is a marker for human hypertension (36). Previous in vitro and in vivo studies have shown that glucocorticoids play an important role in the expression of the AGT gene in the liver (49) and are the greatest inducer of its expression in adipocytes (34). Overexpression of the AGT gene in adipocytes increases plasma AGT level and blood pressure in transgenic mice (50). AGT levels are increased in visceral adipose tissue as compared with the subcutaneous adipose tissue of human subjects and may be responsible for increased blood pressure in obese human subjects (51). Our experiments suggest that an increase in cortisol level (such as in stress and obesity) may increase the expression of the AGT gene in subjects containing Ϫ6A haplotype as compared with Ϫ6G haplotype. This may lead to increased tissue/plasma AGT level and increased blood pressure in subjects containing Ϫ6A haplotype of the AGT gene.
Because plasma AGT level is increased in hypertensive subjects containing the Ϫ6A allele of the hAGT gene as compared with the Ϫ6G allele (7), it was important to examine the plasma hAGT level in transgenic mice containing Ϫ6A and Ϫ6G haplotypes of the hAGT gene. We show that plasma AGT level is increased about 1.7-fold in transgenic mice containing the Ϫ6A haplotype of the hAGT gene as compared with the Ϫ6G haplotype. This increase in plasma AGT level may be responsible for increased blood pressure in transgenic mice containing Ϫ6A haplotype because previous studies have shown that plasma AGT level is directly associated with blood pressure. Kim et al. have shown that plasma AGT concentration is increased 1.45-fold between two-copy and fourcopy AGT-containing transgenic mice, which results in an increase in blood pressure of about 8 mm Hg per copy num-ber (6). Klett et al. (3) have also shown that administration of 1.2 mg/kg recombinant AGT to rats increases plasma AGT level by 1.4-fold within 30 min and increases blood pressure by 19 mm Hg. An increase in the plasma AGT level concurrently increased the plasma angiotensin-II level. Thus, plasma angiotensin-II level was increased by about 50% in transgenic mice containing Ϫ6A haplotype as compared with Ϫ6G haplotype. We have also shown that, as expected, mouse and human renin gene expression is down-regulated in the kidney of double transgenic mice as compared with the control animals. Even with this down-regulation of the renin gene expression, transgenic animals with Ϫ6A haplotype had increased blood pressure as compared with Ϫ6G haplotype.
We show here that systolic blood pressure is increased by 11 mm Hg in male double transgenic mice containing hRen gene and Ϫ6A haplotype of the hAGT gene as compared with Ϫ6G haplotype. Our results for blood pressure of transgenic mice containing Ϫ6A and Ϫ6G haplotypes of the hAGT gene differ from those of Cvetkovic et al. (19), who developed transgenic mice by inserting the hAGT gene containing either Ϫ6A or Ϫ6G in the context of 1.2 kb of the promoter at the hypoxanthine-guanine phosphoribosyltransferase locus. Probably three new polymorphisms identified in the upstream region of the promoter of hAGT gene that are present in our transgenic mice are responsible for this difference. The fact that this increase is due to an increased level of angiotensin-II is shown by our finding that angiotensin receptor blocker (losartan) reduced the blood pressure in transgenic mice.
In conclusion, we have shown that the hAGT gene promoter contains three new polymorphic sites at Ϫ1670, Ϫ1562, and Ϫ1561. Moreover, variants Ϫ1670A, Ϫ1562C, and Ϫ1561T almost always occur with Ϫ6A, and variants Ϫ1670G, Ϫ1562G, and Ϫ1561G almost always occur with Ϫ6G. Therefore, the hAGT gene may be subdivided into Ϫ6A and Ϫ6G haplotypes. We have shown that (a) HNF-1␣ and GR bind strongly to the promoter containing Ϫ6A haplotype of the hAGT gene, and (b) reporter construct containing 1.8 kb of the hAGT gene promoter with Ϫ6A haplotype has 50 -60% increased basal and 4 -6-fold increased glucocorticoidinduced promoter activity upon transient transfection in human liver and kidney cells as compared with Ϫ6G haplotype. We have generated double transgenic mice containing human renin and either Ϫ6A or Ϫ6G haplotype of the hAGT gene. Transgenic mice containing Ϫ6A haplotype have increased plasma AGT and angiotensin-II level and increased hAGT mRNA level in the liver as compared with transgenic mice containing Ϫ6G haplotype. Our in vivo ChIP assay shows that HNF-1␣, GR, and RNA polymerase bind strongly to the chromatin obtained from the liver of transgenic mice containing Ϫ6A haplotype of the hAGT gene as compared with Ϫ6G haplotype. We also show that (a) transgenic mice containing Ϫ6A haplotype have 11 mm Hg increased systolic blood pressure as compared with the transgenic mice containing Ϫ6G haplotype, and (b) angiotensin receptor blocker reduces blood pressure in these transgenic mice. Our experiments for the first time explain the molecular mechanism involved in association of Ϫ6A allele of the hAGT gene with increased blood pressure in human subjects and provide a new experimental animal model that is very similar to human hypertension.