Functional Promoter Polymorphisms Govern Differential Expression of HMG-CoA Reductase Gene in Mouse Models of Essential Hypertension

3-Hydroxy-3-methylglutaryl-coenzyme A [HMG-CoA] reductase gene (Hmgcr) is a susceptibility gene for essential hypertension. Sequencing of the Hmgcr locus in genetically hypertensive BPH (blood pressure high), genetically hypotensive BPL (blood pressure low) and genetically normotensive BPN (blood pressure normal) mice yielded a number of single nucleotide polymorphisms (SNPs). BPH/BPL/BPN Hmgcr promoter-luciferase reporter constructs were generated and transfected into liver HepG2, ovarian CHO, kidney HEK-293 and neuronal N2A cells for functional characterization of the promoter SNPs. The BPH-Hmgcr promoter showed significantly less activity than the BPL-Hmgcr promoter under basal as well as nicotine/cholesterol-treated conditions. This finding was consistent with lower endogenous Hmgcr expression in liver and lower plasma cholesterol in BPH mice. Transfection experiments using 5′-promoter deletion constructs (strategically made to assess the functional significance of each promoter SNP) and computational analysis predicted lower binding affinities of transcription factors c-Fos, n-Myc and Max with the BPH-promoter as compared to the BPL-promoter. Corroboratively, the BPH promoter-luciferase reporter construct co-transfected with expression plasmids of these transcription factors displayed less pronounced augmentation of luciferase activity than the BPL construct, particularly at lower amounts of transcription factor plasmids. Electrophoretic mobility shift assays also showed diminished interactions of the BPH promoter with HepG2 nuclear proteins. Taken together, this study provides mechanistic basis for the differential Hmgcr expression in these mouse models of human essential hypertension and have implications for better understanding the role of this gene in regulation of blood pressure.


Introduction
Essential hypertension, the chief risk factor for cardiovascular and renal diseases, is often associated with and complicated by dyslipidemia [1,2]. 3-Hydroxy-3-methylglutaryl-coenzyme A [HMG-CoA] reductase gene (Hmgcr/HMGCR) is a candidate gene for hypertension; it translates to the rate-limiting enzyme in the cholesterol biosynthesis pathway and cholesterol is the precursor of glucocorticoid steroid hormones that play a profound role in blood pressure homeostasis and hypertension [3][4][5][6][7]. Consistently, the G allele of HMGCR rs17238540 (G/T) single nucleotide polymorphism [SNP] was associated with higher blood pressure [BP] and higher stroke risk in an European population of ,23,000 participants [8]. Moreover, this SNP was associated with the BP response to urinary sodium: potassium ratio [9] and response to statin (inhibitor of HMGCR enzyme) therapy in terms of total cholesterol and triglyceride lowering [10]. Two common and tightly linked HMGCR SNPs were also significantly associated with reduced efficacy of pravastatin therapy [11]. Additionally, investigations on gene expression pattern in adrenal glands of two independent, inbred, homozygous rodent models of human essential hypertension (viz. spontaneously hypertensive rat and blood pressure high [BPH] mice) revealed ,2-to 3-fold over-expression of Hmgcr in these strains as compared to their corresponding controls (viz. Wistar/ Kyoto rat and blood pressure low [BPL] mice) [12,13]. These findings suggested the possibility that an altered Hmgcr expression might be a systematic facet of hereditary hypertension in mammals, perhaps even contributing to diverse metabolic abnormalities associated with this common disorder. However, molecular basis of the differential Hmgcr expression in these animal models has not been studied. Hmgcr levels in other tissues (e.g., liver) of BPL and BPH mice also remain unknown.
The hypertensive mouse strain BPH was developed in a breeding program based solely on selection by elevated BP and it parallels human hypertension [14]. The BPH strain exhibits many of the comorbidities observed in human hypertension, such as higher heart rate, larger hearts and kidneys, higher left ventricular weight and early mortality than the hypotensive BPL strain [14]. During generation of the BPH and BPL strains, the normotensive inbred strain BPN (blood pressure normal) was derived from the unselected control population and this strain serves as a control for hypertensive BPH and hypotensive BPL mice [14].
In the present study, we sequenced the mouse Hmgcr locus (proximal promoter, 20 exons and flanking intronic regions) in the BPH, BPL and BPN strains and discovered several SNPs in promoter and coding exonic regions. Next, we assessed the quantitative impact of the promoter SNPs on Hmgcr gene expression by computational as well as experimental analyses. The results revealed that two promoter SNPs (C-874T and C-740T) altered binding affinities of several transcription factors (n-Myc, Max and c-Fos) and modulated Hmgcr expression in these mouse models of human essential hypertension.

Materials and Methods
Ethics Statement/study approval The present study was approved by the Institutional Biosafety Committee at Indian Institute of Technology Madras (IIT Madras), Chennai in June 2008.

Mouse strains and tissue samples
Liver tissue samples from 5-7 weeks old male BPH (strain BPH/2J, at inbred generation F66) and BPL (strain BPL/1J, at inbred generation F65) mice were collected in RNAlaterH solution (Ambion, USA) at the Jackson Laboratory (Bar Harbor, USA; www.jax.org) and shipped to our laboratory following institutional norms. BPH males display ,120 mm systolic BP (SBP) while BPL males display ,70 mm SBP at 4-15 weeks of age [14]. At 21 weeks of age, the SBP of BPH mice increases further to ,130 mm while that of BPL remains almost unchanged [14]. We chose 5 to 7 weeks old mice for measurement of Hmgcr gene expression levels because at that early age, BPH mice did not attain the maximal elevation of BP. Therefore, studying these mice might allow us to minimize the effects of aging-related confounding factors on Hmgcr gene expression and increase the chance of detecting pathogenic role for Hmgcr in hypertension.

Extraction of RNA and real-time PCR
Total RNA samples from liver tissues of four BPH and four BPL mice were isolated using the TRIZOL reagent (Invitrogen, USA). RNA concentrations were estimated by UV-spectrophotometry (Eppendorf Biophotometer, Germany) and the integrity of RNA molecules was assessed from the appearance of 28S and 18S bands on agarose gels.
RNA samples were subjected to total cDNA synthesis by using the   Table 1. doi:10.1371/journal.pone.0016661.g001 Table 2. Polymorphisms in the hmgcr gene in mouse models of essential hypertension. # The putative transcription factor binding sites and the binding affinity towards the promoter sequences harboring the polymorphisms were predicted using the CONSITE web tool (http://asp.ii.uib.no:8090/cgi-bin/CONSITE/consite/). doi:10.1371/journal.pone.0016661.t002 S1 for details. The relative gene expression levels were determined by calculating the 2 (2DDCt) values [17].

Sequencing of mouse Hmgcr gene for polymorphism discovery
Genomic DNA samples of BPL/1J, BPH/2J and BPN/3J mice were obtained from the Jackson laboratory (Bar Harbor). Primers were designed using the mouse Hmgcr reference sequence NM_008255.2 (from the UCSC Genome Browser) to amplify ,1 kb promoter region, each of the 20 exons as well as 50-100 bp of flanking intronic regions (Table 1 and Fig. 1). PCR was carried out using Phusion TM High-Fidelity DNA polymerase and dNTPs from New England Biolabs, USA. Agarose-gel purified PCR products served as templates for sequencing, with the exception that purified PCR products inserted into the promoterless pGL3-Basic vector (Promega, USA) were used for sequencing the promoter region. DNA samples were sequenced using ABI 3130 Genetic Analyser and BigDye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems, USA).

Construction of Hmgcr promoter-reporter plasmids
Approximately 1 kb Hmgcr promoter region was PCR-amplified from BPH/BPL/BPN genomic DNA sample using the following primers: forward, 59-CGGGGTACCTAAAGTGGGTAGGTA-TATCCG-39 and reverse, 59-CCGCTCGAGCTCACCT-CCGGATCTCAATGG-39 (with added KpnI and XhoI sites at 59 ends in forward and reverse primers respectively, shown in bold). The amplified promoter fragments were inserted between KpnI and XhoI sites in the firefly luciferase reporter vector pGL3-Basic (Promega). Resulting plasmids were named as BPH-961, BPL-961 and BPN-961, which contained 2961 bp to +43 bp region of BPH, BPL and BPN Hmgcr (numberings are with respect to the 1 st nucleotide of Exon 1 as +1). Similarly, the promoterreporter plasmids BPH-769, BPL-769 and BPN-769 (harbouring 2769 bp to +43 bp region) were generated by insertion of PCRamplified products in the pGL3-basic vector using the following primers: forward, 59-CGGGGTACCAAACGCCAGAAGCA-GAAGGTG-39 and reverse, 59-CCGCTCGAGCTCACC-TCCGGATCTCAATGG-3' (with added KpnI and XhoI sites in forward and reverse primers respectively, shown in bold). We also constructed the promoter reporter plasmids BPH-651, BPL-651 and BPN-651 (containing 2651 bp to +43 bp region) by digestion of the BPH-961, BPL-961 and BPN-961 constructs with KpnI and EcoRI (the EcoRI site is located at 2652/2647 bp position in the Hmgcr promoter), excision of ,5.4 kb fragment from gel, treatment with Mung Bean nuclease (New England Biolabs) to remove overhangs and re-circularization with T4 DNA ligase (New England Biolabs). The correct insertion/orientation and existence of SNPs in cloned DNA fragments were confirmed by sequencing of the entire inserts in several clones. The plasmids were purified on columns using an endotoxin-free plasmid DNA purification kit (Hi-Media, India) for transfection experiments.

Cell culture, transfection and reporter assay
Human hepatic cell line HepG2, Chinese hamster ovarian cell line CHO, human embryonic kidney cell line HEK-293 and mouse neuroblastoma cell line N2A were obtained from the National Center for Cell Sciences, Pune, India. Cells were cultured in Dulbecco's Modified Eagle's Medium (DMEM) with high glucose and GlutaMAX TM -I (Invitrogen), supplemented with 10% fetal bovine serum (Invitrogen), penicillin G (100 U/ml) and streptomycin sulfate (100 mg/ml) (Invitrogen) at 37uC with 6% CO 2 . HepG2 and N2A cells (grown at 50-60% confluence in 12well plates) were transfected with 1 mg/well of promoter-reporter plasmids using Lipofectamine-2000 (Invitrogen). Similarly grown CHO and HEK-293 cells were transfected with 2 mg/well of plasmids by calcium phosphate method [18]. As an internal control for transfection efficiency, cells were co-transfected with 0.5 mg/well of a b-galactosidase expression plasmid. Cells were lysed 24-30 hours after transfection for reporter assays. The luciferase assay was carried out with some modifications of a previously described method [19,20]. The beta-gal assay was carried out using ortho-nitrophenyl b-D-galactopyranoside as substrate. See Text S1 for details.The results were expressed as firefly luciferase/b-galactosidase activity.
To test the effect of nicotine on Hmgcr promoter activity, HepG2 and CHO cells transfected with promoter-reporter plasmids were treated with various doses (100 mM to 1 mM) of nicotine bitartarate (Sigma-Aldrich, USA) five hours after transfection and incubated for 16-18 hrs. Likewise, to test the effect of cholesterol, cells were grown in lipid-free DMEM medium (Hyclone-Thermo, USA) and transfected cells were treated with 0.5 mg/ml of 25-hydroxycholesterol and 12 mg/ml cholesterol (Sigma-Aldrich) for 24-30 hrs. Cells were treated with this mixture of sterols because although 25hydroxycholesterol was more potent than cholesterol in suppressing reductase activity, but it could not replace cholesterol in maintaining the cell growth [21,22]. Cells were lysed and assayed for luciferase activity as described above.
In some experiments, promoter-reporter constructs were cotransfected with various transcription factor plasmids into CHO cells: pmiw-nMyc expressing mouse n-Myc cDNA [23], pmiw-Max expressing human Max cDNA [24] and pc-Fos expressing mouse c-Fos cDNA [25]. In these co-transfection experiments, the insert-free vectors pmiwSV (in case of n-Myc/Max) and pSGI (in case of c-Fos) were used as balancing plasmids in different transfection mixtures. As a control for varying cell number within individual wells, total protein contents were measured in cell lysates using Bradford's assay reagent (Sigma-Aldrich). Luciferase activities in cell lysates were expressed as relative light units [RLU]/mg protein.

Electrophoretic mobility shift assay (EMSA)
Nuclear protein extracts from HepG2 cells were prepared using the ProteoJET cytoplasmic and nuclear protein extraction kit (Fermentas Life Sciences, USA) and stored in aliquots at 280uC until use. See Text S1 for details. For EMSA, 10 mg of nuclear protein extract was incubated with binding buffer [10 mM Tris, 50 mM KCl and 1 mM dithiothreitol at pH 7.5], 50 ng/ml poly-dI-dC and 20 fmol of biotinylated oligo for 20 min at room temperature. The reaction mixtures were resolved on 1.5 mm thick 5% non-denaturing polyacrylamide gels and transferred to nylon-66 membranes (Fluka, USA). The DNA oligomers were UV cross-linked to membrane at 312 nm for 10 min. The biotinylated probes were detected by chemiluminescence using LightShift Chemiluminescent EMSA kit (Pierce).

Data presentation and statistics
Promoter/reporter transient transfections were carried out at least three times and results were expressed as mean 6 S.E. Statistical significance was calculated by student's t-test and oneway analysis of variance (ANOVA) with Tukey-Kramer multiple comparisons post-test, as appropriate in different experiments using the InStat 3 program (GraphPad software, USA).

Discovery of polymorphisms in the mouse Hmgcr gene
Sequencing of the Hmgcr locus in BPH, BPL and BPN mice yielded several SNPs (Table 2). In the promoter region, 3 SNPs were detected: at 2874 bp (C/T), 2740 bp (C/T) and 2486 bp (T/DT). Analysis of the sequences immediately surrounding these promoter SNPs by ConSite (http://asp.ii. uib.no:8090/cgi-bin/CONSITE/consite; [26]) for identification of cis-regulatory elements revealed the presence of putative binding sites for the transcription factors c-Fos (at -881 to 2874 bp), n-Myc (at 2744 to 2739 bp) and Max ( = Mycassociated factor X; at 2747 to 2738 bp). The 2874T variant in BPL/BPN contributed to a better binding site than the 2874C variant in BPH for c-Fos (Fig. 2). Likewise, the 2740T allele in BPL contributed to better binding sites for n-Myc and Max than the 2740C allele in BPN/BPH (Fig. 2). The deletion of T at 2486 bp in BPH did not alter any transcription factor binding site.
In addition to the above-mentioned promoter variations, eight SNPs in coding exons and three SNPs in the 39-UTR were detected ( Table 2). Seven of the exonic SNPs altered amino acid residues and four of those belonged to the catalytic domain of the Hmgcr enzyme. Of note, no SNP was detected in the 59-UTR region and intronic regions flanking the exons. To test functional implication of the Hmgcr promoter SNPs, BPH/BPL/BPN promoter/luciferase reporter constructs (Fig. 3A) were transfected into HepG2, CHO, HEK-293 and N2A cells. In HepG2 cells, the BPL promoter activity was ,1.8-fold higher (p,0.01) than the BPH promoter; the BPN promoter activity was ,1.4-fold less (p,0.05) than the BPL promoter (Fig. 3B). Similarly, CHO cells showed ,1.5-fold higher (p,0.01) promoter activity in the case of BPL than BPH; the BPN promoter was ,1.3 -fold more active (p,0.05) than the BPH promoter (Fig. 3C). In HEK-293 cells, BPL and BPN promoter activities were ,1.4-fold (p,0.01) and ,2.2-fold (p,0.01) respectively higher than the BPH promoter activity (Table S1). In N2A cells, activities of the BPL and BPN promoters were ,1.9-fold (p,0.01) and ,1.4-fold (p,0.05) respectively higher than the BPH promoter (Table S1). Thus, across these cell lines, in general, the BPL promoter was more active than the BPH promoter while the BPN promoter expression was intermediate.

Endogenous Hmgcr expression in BPH and BPL mice
To study whether endogenous Hmgcr expressions differ between BPH and BPL mice in parallel to transfected promoter activities, we measured Hmgcr mRNA levels in liver tissues by real-time PCR. We chose liver tissues for this experiment because the liver is the primary site of de novo cholesterol biosynthesis and an important regulator of whole-body intermediary metabolism [27,28]. The BPL liver samples showed significantly higher (,2.6-fold when normalized to GAPDH, p = 0.002; ,3.4-fold when normalized to 18S rRNA, p = 0.022) level of Hmgcr mRNA than the BPH liver samples (Fig. 4).

Differential responses of Hmgcr promoters to cholesterols
Sterols are known to negatively regulate Hmgcr expression as well as enzyme activity [22,32,33]. Therefore, we tested the effect of cholesterols on BPH-and BPL-Hmgcr promoter activities. Both promoters displayed significant reductions in luciferase activities with respect to corresponding basal values; the extent of downregulation was more pronounced in the case of BPL than BPH (,77%, p,0.01 vs. ,53%, p,0.05 in HepG2 cells and ,49%, p,0.01 vs. ,32%, p,0.05 in CHO cells; Fig. 6).

Functional characterization of the Hmgcr promoter SNPs
To assess the contribution of each Hmgcr promoter SNP towards the differential expression of transfected BPH and BPL promoter/ reporter plasmids (Fig. 3), we undertook systematically progressive deletion of the promoter region. First, we generated the BPH-769 and BPL-769 constructs (Fig. 7A) wherein the C-874T SNP was excluded. Transfection of these constructs into HepG2 and CHO cells showed that the BPL promoter was ,25-30% (p, 0.01) more active than the BPH promoter ( Fig. 7B and 7C).
Second, we generated the BPH-651 and BPL-651 constructs (Fig. 8A) wherein the C-874T and C-740T SNPs were excluded. Transfection of these constructs into HepG2 and CHO cells displayed no significant difference in Hmgcr promoter activity between BPH and BPL strains ( Fig. 8B and 8C).
Taken together (Fig. 3, 7 and 8), while the 2486 (T/DT) SNP did not influence promoter activity, the C-874T and C-740T SNPs were functional and responsible for the higher Hmgcr promoter activity in BPL than BPH. Consistently, the BPN-769 and BPH-769 constructs [that differed only at the 2486 (T/DT) SNP position] did not display any difference in expression in HepG2 and CHO cells (data not shown). Likewise, the BPN-651 promoter-reporter construct also did not show any difference in luciferase activity as compared to BPL-651/BPH-651 construct in HepG2 and CHO cells (data not shown).

The transcription factors c-Fos, n-Myc and Max modulate Hmgcr expression
Since computational analysis of the promoter sequences around the C-874T and C-740T SNPs predicted better binding affinity of c-Fos, n-Myc and Max to the BPL promoter than the BPH promoter (Fig. 2), we tested the effect of co-transfection of expression plasmids of these transcription factors with BPL-/BPHpromoter-reporter constructs in CHO cells. The BPL-and BPHpromoters were differentially augmented by c-Fos/n-Myc/Max (Fig. 9). At the lower dose (1.0 mg/well) of the transcription factors, Figure 6. Effect of sterols on Hmgcr promoter expression. HepG2 and CHO Cells cultured in lipoprotein-deficient medium were transfected with BPL/BPH Hmgcr promoter (2961 to +43 bp)/firefly luciferase construct. Transfected cells were treated with of 25-Hydroxycholeserol (0.5 mg/ml) and of cholesterol (12 mg/ml) and incubated for 24-30 hrs. Cells were lysed and assayed for luciferase activity as well as protein concentration (for normalization). The normalized Hmgcr promoter activity in response to sterols (as percentage of control/basal) were expressed as Mean 6 S.E. Each experiment was performed in triplicate and repeated at least three times. Significant reduction in promoter activity was observed in both HepG2 and CHO cells, to a greater extent in the case of BPL than BPH. (**) and (*) indicate p,0.01 and p,0.05 respectively, with respect to the mock (without stimulation). doi:10.1371/journal.pone.0016661.g006 extents of activation of the BPL-promoter over BPH-promoter were ,1.1-, ,2.0-and ,3.7-fold for c-Fos (Fig. 9A), n-Myc (Fig. 9B) and Max (Fig. 9C) respectively. At the higher dose (2.0 mg/well of c-Fos/n-Myc/Max), although no significant difference in stimulation between BPL-and BPH-promoter was observed in the case of n-Myc, the BPL-promoter activation was still significantly more than the BPH-promoter in case of c-Fos (,1.4-fold, p,0.01; Fig. 9A) and Max (,1.6-fold, p,0.01; Fig. 9C).

Differential binding of nuclear proteins with BPL-and BPH-Hmgcr promoter domains
To test whether the BPL-and BPH-Hmgcr promoter domains harboring putative binding sites for c-Fos and n-Myc/Max interact differently with HepG2 nuclear proteins, we carried out EMSA experiments. The BPL c-Fos oligo (that contained a better binding site for c-Fos) showed altered complex formation pattern as compared to the BPH c-Fos oligo (Fig. 10A, lanes 8  and 9 vs. lanes 5 and 6). Likewise, the BPL n-Myc/Max oligo (containing a better binding site for n-Myc/Max) yielded higher amount of a specific nuclear protein-oligo complex than the BPH n-Myc/Max oligo (Fig. 10B, lanes 8 and 9 vs. lanes 5 and  6). As a positive control, we tested complex formation using consensus c-Fos oligo (Fig. 10A, lanes 2 and 3) and consensus n-Myc/Max oligo (Fig. 10B, lanes 2 and 3). As negative controls, when no nuclear extract was added to binding reactions, only labeled oligo bands were observed (Fig. 10A and 10B: lanes 1, 4  and 7).

Mouse Hmgcr genetic polymorphisms
Several SNPs were discovered in the mouse Hmgcr gene ( Table 2). Two (C-874T and C-740T) of the 3 promoter SNPs that distinguished BPL from BPH were transcriptionally active. Alignment of the orthologous mammalian sequences around these two SNP positions revealed that these promoter regions were highly conserved (Fig. 11). The C allele at 2874 bp (present in BPH) occurred in primates while the 2874T allele was found in rodents (Fig. 11A). On the other hand, the C allele at 2740 bp (present in BPH) occurred in rodents while the corresponding T allele was found in primates (Fig. 11B).
Seven of the 8 exonic SNPs altered amino acid residues while one was synonymous. Notably, two non-synonymous variations (Aspar-tate296Asparagine in the trans-membrane region and Glutamic acid455Lysine in the catalytic domain of the enzyme) distinguished BPL from BPH. Interestingly, alignment of the orthologous sequences around the non-synonymous SNPs revealed that while the Hmgcr coding sequence was extremely conserved among mammals (data not shown) amino acid variations discovered in this study were unique to these strains excepting the BPL-Aspartate296, which also occurred in rat (Fig. 11C). However, the functional implications of these polymorphisms are not yet known.
Besides the promoter and 20 exons, we also sequenced the exon-intron borders of the Hmgcr gene to probe for any alternative splicing among these strains because a number of studies reported variations in HMGCR exon/intron splicing in humans and CHO cells [34][35][36][37][38][39]. In humans, HMGCR alternative splicing has been associated with inter-individual variation in plasma low-density lipoprotein cholesterol response to statin treatment [34,35]. However, we did not find any interstrain variation in the consensus splice donor (GU) or acceptor (AG) dinucleotides in the Hmgcr introns.

Molecular basis of altered Hmgcr expression
Expression of promoter-reporter constructs in cultured cells showed that the C-874T and C-740T SNPs were functional. The 2740T allele alone contributed to ,25-30% higher promoter activity than the 2740C allele ( Fig. 7 and 8); the 2740T and 2874T alleles together resulted in ,150-180% higher expression of the BPL promoter than the BPH promoter (that contained the 2740C and 2874C alleles) (Fig. 3). Consistently, the BPN promoter (that harboured the 2740C and 2874T alleles) displayed intermediate activity (,125-135% as compared to the BPH promoter; Fig. 3).
How might the C-874T and C-740T SNPs alter the Hmgcr promoter activity? Computational analysis (by ConSite) revealed that these T alleles offered better binding sites for c-Fos, n-Myc and Max as compared to the C alleles (Fig. 2). Consistent with these computational predictions, over-expression of c-Fos, n-Myc and Max in CHO cells augmented the BPL-Hmgcr promoter activity to a greater extent than the BPH-Hmgcr promoter (Fig. 9). Interestingly, the extent of alteration of the promoter expression by c-Fos/n-Myc/Max paralleled with the differences in ConSite score (an index of binding affinity of a transcription factor protein with a DNA motif) between the motifs for BPL and BPH. For example, among these three proteins, the difference in ConSite score was the highest between BPL-Max and BPH-Max motifs (7.054-2.925<4 units) and the least between BPL-c-Fos and BPH-c-Fos motifs (3.933-2.994<1 unit); Max co-expression resulted in more dramatic differences between BPH and BPL promoter activity while the c-Fos co-expression showed the least difference in promoter activities, especially at the lower doses of the transcription factor plasmid (Fig. 9). In addition, EMSA experiments also showed enhanced complex formation of HepG2 nuclear proteins with the BPL c-Fos/n-Myc/Max oligos than the BPH c-Fos/n-Myc/Max oligos (Fig. 10) confirming roles of the promoter variants in the differential Hmgcr gene expression.
Of note, c-Fos, a basic leucine zipper (bZIP) protein and a major component of the activator protein-1 transcription factor complex, has been implicated as a regulator of cell proliferation, differentiation and transformation [40][41][42]. A recent study reported recruitment of c-Fos to the Hmgcr promoter for transcriptional regulation under acute kidney injury in mice [43]. The transcription factor n-Myc is a basic helix-loop-helix leucine zipper (bHLH-ZIP) protein that heterodimerizes with the transcription factor Max and play important roles in neuronal differentiation and cell proliferation [44][45][46][47][48]. Although n-Myc and Max have been reported to regulate transcription of many genes [49][50][51][52][53], this study, for the first time, provides evidence for regulation of the Hmgcr gene by these transcription factors.

Differential regulation of Hmgcr promoters by nicotine and cholesterols
Nicotine caused dose-dependent activation of BPL and BPH promoters in both HepG2 and CHO cells; the effect was more prominent in the case of BPL (Fig. 5). Are there functional nicotinic acetylcholine receptors (nAChRs) in these cell types to elicit nicotine-induced Hmgcr promoter stimulation? In isolated rat hepatocytes, nicotine increased intracellular calcium concentration and this phenomenon was blocked by d-tubocurarine, a nAChR antagonist [54]. A recent study also detected the presence of alpha-7 nAChRs in mouse liver by utilizing two carbon-11-labeled alpha-7 nAChR agonists [55]. On the other hand, whereas the presence of a7, a4 or b2 subunits of nAChR in CHO cells could not be detected [56], the existence of an intraovarian, non- The results are expressed as ratios of firefly luciferase activity/mg protein and are the mean 6 SE (n = 3-4 transfections for each construct). Although both BPL and BPH promoters displayed, in general, activation by each of these three transcription factors in a dose-dependent manner with respect to the control, the extents of activations were greater in case of BPL. (**) and (*) indicate p,0.01 and p,0.05 respectively with respect to the corresponding mock (i.e., without a co-transfected transcription factor). Co-transfection of Max expression plasmid resulted in more dramatic difference in promoter activities between BPL and BPH while the c-Fos expression plasmid showed the least difference in promoter activities, especially at the equimolar dose (1.0 mg/well transcription factor along with 1.0 mg/well promoter/reporter). doi:10.1371/journal.pone.0016661.g009 neuronal cholinergic system in human and rat has been reported [57,58]. In view of these reports, we speculate that the nicotinic stimulation of Hmgcr promoter in HepG2 and CHO may be mediated by some nAchR subtype(s). Further studies are required to establish the mechanism of nicotinic signal transduction in these cells.
The differential nicotinic activation of Hmgcr promoters may be mediated by altered interaction of the c-Fos/n-Myc motifs in cis with c-Fos and n-Myc in trans since these BPH-and BPL-promoter motifs have different binding affinities with these nuclear proteins ( Fig. 2 and 10). Notably, several studies have shown that nicotine regulates expression of c-Fos and Myc in different cell types [59][60][61][62]. The different extent of nicotinic stimulation of the Hmgcr promoters may also be contributed by the transcription factor cyclic AMP response element binding protein (CREB) because nicotine is known to activate the phosphorylation of CREB, which induces the expression its early target c-Fos [63,64]. Thus, our initial findings indicate hitherto un-described involvement of several transcription factors in nicotine-evoked activation of Hmgcr transcription and hence their possible roles in nicotinic modulation of cholesterol biosynthesis. However, further studies are required to confirm the contribution of these transcription factors in the upregulation of Hmgcr expression by nicotine.
Similar to nicotine, the BPL-and BPH-Hmgcr promoters responded differentially to cholesterols, albeit the effect was downregulation of gene expression; BPL-promoter showed more pronounced repression of the luciferase activity than the BPHpromoter (Fig. 6). What might be the molecular mechanism for this differential response by these promoters? Although the identities of the transcription factors involved in the sterol regulation of Hmgcr promoter still remain incompletely understood, previous studies demonstrated important roles for the sterol regulatory element (SRE-1) SRE binding proteins (SREBPs), CCAAT-binding factor/nuclear factor-Y (CBF/NF-Y) and CREB [65][66][67][68]. Since the BPL-and BPH-Hmgcr did not differ at the SRE-1 or CBF/NF-Y or CRE motifs and the differential expression under the basal conditions was mediated by c-Fos/n-Myc/Max (Fig. 2, 9, 10), the greater sterol-repression (i.e. negative feedback regulation) of the BPL-promoter may be modulated by interaction of SREBP/CBF/NF-Y/CREB with c-Fos/n-Myc/ Max. Of note, 25-hydroxycholesterol has been reported to cause calcium-dependent activation of c-Fos via the ERK1/2 signaling pathway in monocytic THP-1 cells [69]. Further studies are required to unravel the possible regulatory role of c-Fos/n-Myc/ Max in repression of Hmgcr expression by cholesterols.
Hmgcr endogenous gene expression studies: possible mechanisms for differential transcript abundance Tissue distribution pattern of the Hmgcr gene showed a high level of expression in the liver [70]. Consistently, the Hmgcr protein level was reported to be 4-to 6 -fold higher in the liver than most peripheral tissues [27]. In this study, we measured the abundance of Hmgcr transcripts in BPH and BPL liver tissues and detected as much as ,3.4-fold higher expression in BPL (Fig. 4). Intriguingly, such a large magnitude Hmgcr over-expression in BPL liver cannot be accounted for by functional promoter variants alone because the transfected BPL promoter showed only up to ,1.8-fold higher activity than the BPH promoter in HepG2 cells (Fig. 3). Other genetic and/epigenetic factors may, therefore, also contribute to the difference in the endogenous Hmgcr expression in these mouse models. However, the higher Hmgcr expression in BPL is consistent with the elevated hepatic cholesterols (,1.4-fold, p = 0.025) and plasma cholesterols (,1.3-fold, p,0.01) in these mice as compared to BPH mice (Mouse Phenome Database, Jackson Laboratory; ,www.jax.org/phenome.).
Of note, in an earlier genome-wide transcriptome profiling study in adrenal glands, BPH mice displayed ,3-fold higher Hmgcr expression than BPL mice [12]. What might be the mechanism of such directionally opposite differential Hmgcr expression between liver and adrenal glands in these mice? Given that in both these studies, BPL and BPH mice of similar age (5-7 weeks old) and same sex (male) were used, the altered Hmgcr expression might be caused/contributed by other factors. For example, Hmgcr expression in these strains might occur in a tissue-specific manner, perhaps mediated by specific transcription factors in liver versus adrenals tissues, as reported in the cases of human and rat orthologues of Hmgcr as well as other genes [19,[71][72][73][74][75][76]. Further, the concentration of cholesterol in the plasma reflects the net contribution of cholesterol synthesis, secretion, and absorption from various tissues, including the liver and adrenal glands [77]. Although extrahepatic tissues also have local cholesterol biosynthesis systems [78][79][80], the cholesterol needs of several tissues are mainly met by receptor-mediated uptake of low-density lipoprotein from the circulation [32]. Notably, in female rats, the highest rate of uptake of sterols was found in adrenal glands where only 4% of the tissue sterol content came from local synthesis [81]. In BPH mice, the supply of low-density lipoprotein cholesterol to the extrahepatic tissues including adrenals might be insufficient due to lower circulating cholesterol in this strain. Therefore, the Figure 11. Conservation of Hmgcr sequences among mammalian species. Alignment of orthologous sequences around the promoter polymorphisms C-874T (panel A), C-740T (panel B) and the amino acid variant N296D (panel C) in mouse Hmgcr was carried out using Clustal W. The promoter as well as protein sequences were found to be highly conserved among the mammals. The binding motifs for the transcription factors c-Fos (panel A) and n-Myc/Max (panel B) are shown within brackets. The BPH, BPL and BPN sequences were determined in this study (as described in the Materials and Methods section) while the other sequences were obtained from UCSC/NCBI public databases; the accession numbers for the promoter sequences are: human, NM_000859.2; chimpanzee, XM_001148324.1; rhesus monkey, XM_001104607.2; rat, NM_013134.2, and the accession numbers for the amino acids sequences are: human, NP_000850.1; chimpanzee, XP_001148324.1; rhesus monkey, XP_001104607.1; rat, NP_037266.2). The positions of the nucleotide/amino acid variations are indicated by upward arrows. doi:10.1371/journal.pone.0016661.g011 upregulation of Hmgcr expression in BPH adrenals [12] might take place to increase de novo cholesterol synthesis in order to meet the local sterol-demand.
Thus, the BPH mouse appears as a unique model where the dyslipidemia and hypertension are not inter-connected. Given that these genetically hypertensive mice were derived solely based on high BP phenotype [14], this phenomenon might be caused/ contributed by segregation of genes. However, such dissociation between severe hypertension and enhanced lipid synthesis in this model may yield important insights into factors that govern the coupling of these traits in humans. Interestingly, this is reminiscent of the recent findings on chromogranin A gene knockout mice, which displayed severe hypertension but unaltered plasma cholesterol level as compared to wild type mice [82].

Concluding remarks
The present study unfolded the quantitative impact of two Hmgcr promoter SNPs on gene expression in three mouse models of human essential hypertension. The functional implication of these SNPs for hypertension remains to be elucidated. Further investigations are also required to unravel the qualitative impact of non-synonymous SNPs (especially those in the catalytic region). Nonetheless, this is the first report on identification and molecular characterization of functional polymorphisms at the Hmgcr locus in genetically hypertensive versus genetically hypotensive mice. Future studies may establish additional genetic and molecular links of the Hmgcr gene to hypertension and associated cardiovascular disease states.

Supporting Information
Text S1 Supplementary methods.