Sodium taurocholate cotransporting polypeptide (Ntcp) is the major uptake system for conjugated bile acids. Deletions of hepatocyte nuclear factor (HNF)-1α and retinoid X receptor-α:retinoic acid receptor-α binding sites in the mouse 5′-flanking region corresponding to putatively central regulatory elements of rat Ntcp do not significantly reduce promoter activity. We hypothesized that HNF-4α, which is increasingly recognized as a central regulator of hepatocyte function, may directly transactivate mouse (mNtcp). A 1.1-kb 5′-upstream region including the mouse Ntcp promoter was cloned and compared with the rat promoter. In contrast to a moderate 3.5-fold activation of mNtcp by HNF-1α, HNF-4α cotransfection led to a robust 20-fold activation. Deletion analysis of mouse and rat Ntcp promoters mapped a conserved HNF-4α consensus site at −345/−326 and −335/−316 bp, respectively. p-475bpmNtcpLUC is not transactivated by HNF-1α but shows a 50-fold enhanced activity upon cotransfection with HNF-4α. Gel mobility shift assays demonstrated a complex of the HNF-4α-element formed with liver nuclear extracts that was blocked by an HNF-4α specific antibody. HNF-4α binding was confirmed by chromatin immunoprecipitation. Using Hepa 1–6 cells, HNF-4α-knockdown resulted in a significant 95% reduction in NTCP mRNA. In conclusion, mouse Ntcp is regulated by HNF-4α via a conserved distal cis-element independently of HNF-1α.
- hepatocyte-enriched transcription factors
- nuclear hormone receptors
- transcriptional coactivator
- gene regulation
- bile acid transport
the coordinated expression of a large number of genes is required to maintain essential functions of the adult liver in metabolic homeostasis including bile formation. Physiologically, bile acids and other organic anions are taken up from portal blood and secreted into bile by ATP-dependent, export pumps thereby generating an osmotic gradient. The Na+/taurocholate cotransporter (Ntcp; SLC10A1 in humans, Slc10a1 in rodents), a 56-kDa basolaterally localized membrane transporter, probably represents the main hepatocellular sodium-dependent uptake system for conjugated bile acids from sinusoidal blood in human and rodent liver (2, 12, 13).
Regulation of Ntcp gene expression primarily occurs at the transcriptional level (6, 20, 28). Hepatocyte nuclear factor-1α (HNF-1α, TCF1) and retinoid X receptor (RXR-α, NR2B1):retinoic acid receptor (RAR-α, NR1B1) heterodimer complex have so far been identified to control rat Ntcp gene transactivation (19, 28). Decreased binding activity at these two regulatory elements in vivo occurs by either induction of inflammatory cytokines or retention of bile acids and leads to downregulation of Ntcp gene expression (3, 7, 28). In accordance with the concept of HNF-1α as a major regulator of the Ntcp gene, HNF-1α knockout mice (Tcf1−/−) exhibit a decreased Ntcp expression to less than 10% of wild-type controls (24).
In recent studies, HNF-4α (NR2A1) has become increasingly recognized as a central regulator of hepatocyte differentiation and function (29). HNF-4α belongs to the nuclear hormone receptor family of transcription factors and binds DNA as a homodimer (25, 29). Recently, HNF-4α has been found to contribute to a large fraction of the liver transcriptome (21). In this study, with the use of combined chromatin-immunoprecipitation with a promoter microarray, HNF-4α has been found to bind over 40% of the active promoters in hepatocyte tissue (those occupied by RNA polymerase II, and thus likely to be transcribed).
In adult hepatocyte-specific conditional HNF-4α knockout mice (H4LivKO) using the Cre-loxP system, expression of a large number of genes whose gene products are essential for adult liver function is disrupted (15, 16, 18, 26, 27). These mice with HNF-4α-null hepatocytes present with high serum bile acid levels and defects in bile acid transport including decreased expression of the Ntcp (Slc10a1) gene (15).
HNF-4α and HNF-1α occupy each other's promoter as an example of a multicomponent regulatory loop in hepatocytes (11, 21). However, HNF-4α occupies significantly more genes in the pancreatic β-cell and the hepatocyte than does HNF-1α (11). This finding does not support the theory that HNF-4α necessarily mediates its effects on gene expression (including Ntcp) solely through the indirect regulation of HNF-1α (11). Further information will be particularly useful in the interpretation of changes in Ntcp expression in HNF-4α and HNF-1α gene knockout mouse models.
Deletions of HNF-1α and RXR-α:RAR-α binding sites in the mouse 5′-flanking region corresponding to putatively central regulatory elements within the minimal rat Ntcp promoter did not significantly reduce promoter activity. This fact further questions the importance of these transactivators for rodent Ntcp gene expression (9, 17). On the basis of these data, we hypothesized that HNF-4α may directly transactivate mouse and rat Ntcp through a thus far unknown element. We therefore characterize the role of HNF-4α in controlling rodent Ntcp gene transactivation.
MATERIALS AND METHODS
Cloning of mouse Ntcp promoter constructs.
The mouse Ntcp 5′-untranslated region including the promoter was cloned as previously described (9) and is published in GenBank (accession no. AF190698). The numbering of the mouse promoter is deduced from the rat transcription start site (GenBank L76612) as determined by Karpen et al. (19) by alignment of the mouse sequence (AF190698, with the transcription start site starting at position 975). Further numbering of the mouse promoter fragments was made relative to this site by being designated as +1. Mouse Ntcp promoter fragments p-974bpmNtcp luciferase (LUC) and p-475bpmNtcpLUC in pXP2Δ2 vector were constructed as previously described (9). Additional deletion constructs were generated by PCR and cloned into the KpnI and BglII sites of pXP2Δ2. The following constructs were generated: p-350bpmNtcpLUC, p-325bpmNtcpLUC, p-175bpmNtcpLUC, p-120bpmNtcpLUC, and p-69bpmNtcpLUC.
Cloning of rat Ntcp promoter constructs.
The rat Ntcp promoter was subcloned from pDB1.5, GenBank L76612 (19) into pGL3basic with MluI and BglII restriction sites generating a construct spanning nt −1,230 to +46 of the rat (rNtcp) promoter region. Further deletion constructs were generated by PCR and cloned into the MluI and BglII sites of pGL3basic, thereby generating the following constructs: p-146bprNtcpLUC, p-307bprNtcpLUC, p-368bprNtcpLUC, p-469bprNtcpLUC, p-620bprNtcpLUC, p-770bprNtcpLUC, p-930bprNtcpLUC, and p-1230bprNtcpLUC. The sequence of the oligonucleotides used for PCR is available from the authors on request.
Cell culture and transient transfections.
Hepa 1–6 and HepG2 hepatoma cells (1 × 105 cells/well) cells were obtained from American Tissue Culture Collection (ATCC, Manassas, VA) and grown in DMEM or DMEM/F12 (Gibco BRL, Grand Island, NY), respectively, and supplemented with 10% fetal calf serum and penicillin/streptomycin. Cells were subcultured weekly by using trypsinization followed by 1:20 dilution and plating in 100-mm Petri dishes. Transfections of Ntcp promoter constructs in HepG2 cells were performed utilizing either Fugene6 (Roche, Indianapolis, IN) according to manufacturer's instructions or standard calcium phosphate DNA coprecipitation techniques. In cotransfection experiments, 1.8 μg/ml of each mNtcpLUC pXP2Δ2 construct was combined with 0.1–0.75 μg/ml of HNF-4α- or HA-tagged peroxisome-proliferator-activated receptor-γ coactivator-1-α (PGC-1-α) expression plasmids (Dr. A. Kralli, Scripps Institute, La Jolla, CA). Cointroduced pCMV-β-Gal expression vector served as an internal control for transfection efficiency. Promoter activities are given as the means ± SD of triplicate transfections.
Electrophoretic mobility shift assays.
Preparation of nuclear extracts was performed as previously described (8). Protein (10 μg) was incubated on ice for 30 min with 2 × 104 cycles/min 32P-end-labeled oligonucleotide probe (5′-TTAGATGAGGAAGGCAAAGGCAGAAA-3′ corresponding to −352/−327 in the mNtcp promoter). For supershift assays, nuclear extracts were preincubated for 30 min on ice with 5 μg of a polyclonal HNF-4α antibody (sc-8987, Santa Cruz Biotechnology, Santa Cruz, CA) before addition of labeled oligonucleotides. For competition assays, a 100-fold molar excess of unlabeled oligonucleotides was coincubated with the labeled probe. Separation of protein-DNA complexes was obtained by electrophoresis through a nondenaturing 6% polyacrylamide gel and quantified by phosphorimaging.
Chromatin immunoprecipitation assay.
Chromatin immunoprecipitation (ChIP) assays were performed with the Chromatin Immunoprecipitation Assay Kit from Upstate Cell Signaling Solutions (Lake Placid, NY) according to their protocol. Briefly, mouse Hepa 1–6 cells were grown to 80–90% confluency in 10-cm dishes and crosslinked with 1% formaldehyde in tissue culture medium for 10 min at 37°C. Cell samples were washed twice in ice-cold PBS containing protease inhibitors (1 mM PMSF, 1 μg/ml aprotinin, and 1 μg/ml pepstatin A). Cells were pelleted and lysed in SDS lysis buffer (Upstate Biotechnology). Samples were sonicated to shear DNA to a size between 200 and 1,000 bp. Subsequently, an aliquot of sheared DNA was analyzed by agarose gel electrophoresis to confirm the correct size of sheared DNA length for immunoprecipitation. The chromatin samples were incubated overnight at 4°C with HNF-4α (Santa Cruz, sc-8987) antibody. As positive and negative controls the recommended anti-acetyl histone H4 polyclonal antibody (Upstate, no. 06-866) and anti-cytochrome C antibody (Santa Cruz Biotechnology, no. sc-8385), respectively, were used, and another reaction without antibody served as a second negative control. Immune complexes were precipitated with salmon sperm DNA-bovine serum albumin-Sepharose beads. DNA was prepared by treatment with DNase- and RNase-free proteinase K, extraction with phenol and chloroform, and ethanol precipitation. PCR was performed with primers flanking the predicted HNF-4α-binding site (forward: 5′-ACAAAGCAAGGTCTCAGAGGAGGAC-3′, reverse: 5′- GCTTCTACCCCAGTCCAGAAAACTC-3′).
Mutations in the HNF-4α site of the mouse Ntcp promoter were carried out with QuickChange site-directed mutagenesis kit (Stratagene, La Jolla, CA) according to manufacturer's directions. Primers used for generating the desired mutants are listed in Table 1. Mutated DNA sequences were verified by sequencing using ABI Prism machine.
Transfection of Hepa 1–6 cells with control and HNF-4α siRNAs and analysis of message levels by qPCR.
Hepa 1–6 cells were plated at a density of 5 × 105 cells/well in six-well plates. Twenty-four hours later, the cells were transfected with control or mHNF-4α small interfering (si)RNAs (siGENOME SMART pool; Dharmacon, Lafayette, CO) at a final concentration of 50 nM. The SMART pool against mouse HNF-4α contained a mix of four duplexes of siRNAs with the following sense sequence: duplex 1: 5′-GAAGGAAGCUGUCCAAAAUUU-3′; duplex 2: 5′-AGAGGUCCAUGGUGUUUAAUU-3′; duplex 3: 5′-UGUCGUUACUGCAGGCUUAUU-3′; duplex 4: 5′-CUAACACGAUGCCCUCUCUCAUU-3′. Briefly, 5 μl control or HNF-4α of siRNA in 95 μl of OPTI-MEM and 5 μl of Trans-IT TKO (Mirus Bio, Madison, WI) in 95 μl of OPTI-MEM are incubated for 5 min, after which they are mixed and further incubated for 20 min at room temperature. The mixture was added to cells in six-well plates after they had been washed once with OPTI-MEM. After a 24-h incubation, the medium was replaced with DMEM, and the incubation continued for additional 24 h. The wells were then processed for total RNA isolation using TRIZOL reagent according to manufacturer's instructions. A sample (1 μg) of total RNA was reverse transcribed with Invitrogen (Carlsbad, CA) Superscript First Strand Synthesis system for RT-PCR according to instructions. For qPCR 50 ng reverse transcribed cDNA was used per well using Quantitect SybrGreen PCR kit (Qiagen, Valencia, CA) in a Bio-Rad (Redmond, CA) Mini Opticon 3 with the following cycle parameters: 1) one cycle at 95°C for 15 min, 2) 40 cycles each consisting of denaturation at 95°C for 15 s, annealing at 56°C for 30 s, and extension at 72°C for 30 s with a final extension cycle at 72°C for 10 min. After each cycle, the wells were read for emitted fluorescence and, at the end of completion of PCR, a melting curve analysis was included in the program. Primers were selected using Primer Express (Applied Biosystems, Norwalk, CT) to amplify the 80–90-bp region by using cDNA sequences for mouse HNF-4α, Ntcp, and 36B4 in the NCBI database and are listed in Table 1. For all primers, melting curve analysis was conducted to verify that the primers resulted in a single peak of fluorescence with no primer dimers. Ct values of each of the genes obtained were normalized by subtraction of Ct values for 36B4 (a constitutive ribosomal RNA gene), where Ct is cycle threshold. Relative expression levels were obtained with the comparative method as described in ABI Reference Manual using the formula 2−ΔΔCt, keeping the message level obtained in the control siRNA-treated cells at 100%.
Western blot analysis of HNF-4α and HNF-1α protein expression after control and HNF-4α siRNA treatment.
Hepa 1–6 cells were treated with control siRNA or HNF-4α siRNA as described above. Cell lysates were prepared by resuspension of cell pellet in mammalian protein extraction reagent (Pierce Chemical, Rockford, IL) with protease inhibitor cocktail (Sigma, St. Louis, MO), incubation on ice for 15 min followed by centrifugation at 14, 000 g for 15 min at 4°C. Western blotting of cell lysates and analysis of HNF-1α and HNF-4α protein levels were carried out according to procedures previously described (1). Equal amount of protein (184 μg) was loaded from the differentially treated samples. Polyclonal antibodies to HNF-1α (no. sc-8986, H-205) and HNF-4α (no. sc6556, C-19) were obtained from Santa Cruz Biotechnology and used at a 1:2,000 dilution. Protein levels for HNF-1α and HNF-4α as seen in the Western blots were quantitated using Image J (from the National Institutes of Health web site). As positive controls for HNF-4α, nuclear extract from Huh-7 cells transfected with an expression plasmid for rat HNF-4α cDNA was used.
Transactivation of mouse Ntcp promoter by liver-enriched transcription factors.
The mouse Ntcp promoter activity has been shown to be increased 3.5-fold upon cotransfection of an HNF-1α expression plasmid in luciferase assays (9). To investigate the potential of HNF-4α to transactivate the mouse Ntcp promoter, an HNF-4α expression plasmid was cotransfected into HepG2 cells together with the longest mNtcp promoter construct p-974bpmNtcpLUC. To test for a linear dependency of HNF-4α-induced gene transcription with the transfected amount of the HNF-4α expression plasmid, overexpression of HNF-4α reporter gene activity was increased in a concentration-dependent fashion between 0.1 and 0.5 μg/ml plasmid DNA (Fig. 1A). In subsequent experiments, a concentration of 0.25 μg HNF-4α expression plasmid per milliliter of medium was used for cotransfection experiments. Strikingly, promoter activity was upregulated ∼20-fold compared with basal levels (Fig. 1B).
This suggests the presence of an HNF-4α cis-acting element in the 5′-untranslated region of the mNtcp promoter. Interestingly, the minimal p-475bpmNtcpLUC promoter construct, which lacks any identified HNF-1α binding site and is not transactivated by HNF-1α (9), shows a 50-fold enhanced activity upon cotransfection with HNF-4α compared with basal levels (Fig. 1B). These data strongly argue for a direct HNF-4α-based effect on reporter gene expression independently of HNF-1α.
Basal mouse Ntcp promoter activity.
Transfection studies with mouse Ntcp promoter reporter constructs were performed in HepG2 cells. A series of promoter deletion constructs extending from −974 bp to −69 bp (p-974bpmNtcpLUC, p-475bpmNtcpLUC, p-350bpmNtcpLUC, p-325bpmNtcpLUC, p-175bpmNtcpLUC, and p-69bpmNtcpLUC) or the empty reporter vector pXP2Δ2LUC were transfected into HepG2 cells to determine basal promoter activities (Fig. 1C). The minimal p-69bpmNtcpLUC construct exhibited background luciferase activity, and the activity of the p-325bpmNtcpLUC promoter fragment was ∼40% compared with the full-length p-974bpmNtcp promoter. Addition of a further 25 nt resulting in p-350bpmNtcpLUC increased basal promoter activity roughly fivefold (Fig. 1C). These results indicate the presence of a regulatory element responsible for constitutive Ntcp gene expression within promoter sequence −350 to −325 bp.
Mapping the HNF-4α cis-acting element in the mouse Ntcp promoter.
To identify the location of the putative HNF-4α-binding site in the mouse Ntcp promoter, deletion constructs of the p-974bpmNtcpLUC construct were generated as described in materials and methods and cotransfected in HepG2 cells with an HNF-4α-expressing plasmid. HNF-4α conferred high reporter gene activity in constructs containing a minimal promoter sequence of 350 bp, ranging between 20- and 24-fold compared with basal activity of the full-length promoter (Fig. 1C). In contrast, the truncated promoter beginning at −325 bp exhibits a significantly reduced activation accounting for only around 25% of the p-350bpmNtcpLUC activity. Another drop in LUC activity was observed between −175 and −69 bp.
HNF-4α binds the mouse Ntcp promoter.
To confirm that the putative HNF-4α binding site truly serves as a cis-acting element, DNA and chromatin binding studies were performed. First, a complex formation was detected after incubation of radiolabeled oligo mNtcp −352−327 encompassing the newly identified site with mouse liver nuclear extract. Complex formation could be blocked using an HNF-4α-specific antibody (Fig. 2A).
Further corroboration came from ChIP assays (Fig. 2B) performed on DNA from Hepa 1–6 cells. Chromatin DNA was precipitated using either anti-HNF-4α antibody, anti-acetyl-histone 4 (H4) antibody (positive control), or a cytochrome C antibody (negative control) before incubation with protein A Sepharose. Subsequent to purification of the precipitated DNA, the sequence encompassing the HNF-4α binding site was amplified using specific primers. First, it was confirmed that the primer pairs gave a specific product with purified input DNA (data not shown). HNF-4α antibody but not the cytochrome C antibody (as a negative control) brought down chromatin-associated DNA, which when used in PCR amplified the 226-bp fragment containing the HNF-4α response element. Again, this experiment confirms that HNF-4α occupies the mouse Ntcp promoter. Respective controls verified the specificity of the signal (Fig. 2B).
Mutation of the HNF-4α binding site in mouse Ntcp promoter results in decreased reporter gene activity.
To further characterize the functional role of HNF-4α on mNtcp gene expression, we studied the effects of an HNF-4α binding site mutation on Ntcp promoter transactivation. Mutation of the HNF-4α binding site led to a 42% decrease in activity (11.8 ± 3.3 in wild-type compared with 6.9 ± 0.5 normalized luciferase activity in the mutated promoter-transfected cells; 2nd bar vs. 3rd bar) (Fig. 3). PGC-1-α has previously been shown to serve as a coactivator for HNF-4α for a number of genes using promoter studies (22, 30). Consistent with this notion, when we cotransfected PGC-1-α with the mouse Ntcp promoter into hepatoma cells, we observed a 50% increase in activity compared with cells that were not cotransfected, suggesting that PGC-1-α further potentiated the effect of HNF-4α on the mouse Ntcp promoter (Fig. 4, 1st bar vs. 2nd bar; 17.8 ± 0.9 in PGC-1-α cotransfected cells vs. 11.8 ± 3.3 normalized luciferase activity in controls). Transfection of PGC-1-α alone with the promoter did not result in significant change in activity.
HNF-4α siRNA treatment results in downregulation of Ntcp message.
To see whether HNF-4α plays a significant role of transcription of Ntcp mRNA in vivo, mouse-derived Hepa 1–6 cells were treated with siGENOME SMARTpool siRNAs for mouse HNF-4α and a control siRNA obtained from Dharmacon as described in materials and methods. SMARTpool consists of a mixture of siRNAs for different parts of the message, which has been shown to result in uniform downregulation of message levels for many proteins compared with treatment with a single siRNA species. As would be predicted, transfection of Hepa 1–6 cells with HNF-4α siRNA led to significant downregulation of HNF-4α mRNA (4.3 ± 0.6% of control siRNA, P < 0.005, Fig. 4) as measured by real-time qPCR. Consistent with the data on the mNtcp promoter studies shown above, Ntcp message levels were also significantly downregulated in HNF-4α siRNA-treated cells compared with control siRNA treatment (4.6 ± 1.2% controls; P < 0.001). We wanted to see whether HNF-4α siRNA treatment also led to downregulation of HNF-4α and HNF-1α proteins by Western blot analysis using specific antibodies to HNF-4α and HNF-1α. Data in Fig. 5 show that HNF-4α siRNA treatment resulted in decreased HNF-4α (0.4% of control) and HNF-1α (26% of control) proteins consistent with the real-time PCR data. Taken together with the data on the mNtcp promoter, these observations show that HNF-4α plays a significant role in transcriptional regulation of the Ntcp gene. Thus our studies also confirm the fact that liver-specific HNF-4α-null mice exhibited significant downregulation of Ntcp mRNA in their livers (15).
Comparison of HNF-4α transactivation of mouse and rat Ntcp promoters.
To investigate whether HNF-4α mediates transactivation of rodent Ntcp promoters in general, several deletion constructs of the rat Ntcp promoter were also tested in the luciferase gene reporter assay by using HNF-4α cotransfection experiments. Similar to the mouse Ntcp promoter, the promoter fragment encompassing ∼350 bp (p-368bprNtcpLUC) displayed a fourfold higher reporter gene activity than the slightly shorter p-307bprNtcpLUC truncation (Fig. 6A).
In general, activities of the rat Ntcp promoter mirrored those of the highly homologous mouse sequence. Analysis of the mouse sequence between −350 and −325 bp revealed a putative HNF-4α binding site, which is conserved in the rat promoter sequence (Fig. 6B).
NTCP/Ntcp is a 56-kDa polypeptide localized exclusively at the basolateral membrane of the hepatocyte and mediates Na-dependent uptake of conjugated bile acids. Previous work from our laboratory and those of others have implicated HNF-1α, RAR-α:RXR-α, CCAAT/enhancer binding protein (C/EBP-β), and glucocorticoid receptor (GR) in activation of the rat/human genes (5, 19). We recently performed preliminary analysis of the mouse 1.1-kb Ntcp promoter and found that an upstream HNF-1α site is involved in regulation by HNF-1α (9). HNF-4α (NR2A1) is a member of the nuclear receptor superfamily that is highly conserved and expressed at high levels in liver, kidney, intestine, and pancreas in mammals and homologous structures in vertebrates. It can be activated in the absence of ligands, but some studies have suggested that fatty acyl CoA-thioesters can act as agonists or antagonists depending on fatty acid chain length and degree of saturation (23). A large number of genes was shown to be targets of HNF-4α in liver and pancreas based on a recent study using ChIP-based promoter analysis (21). Most of the target genes were also positive for ChIP using an antibody to RNA polymerase II, suggesting that HNF-4α is involved in initiation.
HNF-4α is a transcription factor controlling a majority of hepatocyte-specific genes. We wanted to investigate whether it is involved in direct regulation of NTCP/Ntcp although it can also indirectly regulate via HNF-1α. To map the potential HNF-4α response element in mouse Ntcp promoter, we initially created nested deletion constructs from −69 to −974 bp upstream of the start site and found that significant basal activity is obtained in constructs containing at least −350 bp (Fig. 1C). To differentiate between HNF-1α and HNF-4α transactivation, we examined the transactivation after cotransfection with HNF-1α and HNF-4α encoding plasmids. Although there was significant transactivation by HNF-4α in reporter constructs containing at least −475 bp, most activation was obtained in constructs less than −475 bp of the 5′-untranslated region where HNF-1α-induced transactivation is almost fully abolished (Fig. 1B). A further deletion construct that differed only by 25 bp between −325 bp and −350 bp was then generated to finely map the binding site as shown in Fig. 1C. The p-325 construct had modest activation, whereas further addition of 25 bp led to sharp increase in HNF-4α transactivation capacity.
Furthermore, we established that the murine response element is bound by HNF-4α using EMSA and ChIP assays. The gel shift experiment demonstrates that the HNF-4α-oligonucleotide complex can be blocked by a specific antibody to HNF-4α (Fig. 2A). Albeit there is no supershift visible upon HNF-4α antibody addition, the disappearance of the complex is a clear signal for specificity together with the specific competition experiment resulting in loss of labeled complex. To assess direct promoter occupation at the chromatin level, we performed ChIP assays with the mouse Ntcp promoter and a specific HNF-4α antibody and demonstrated that the region of the promoter containing −345/−326 is indeed occupied by HNF-4α (Fig. 2B).
Additional studies were conducted to functionally prove that HNF-4α directly transactivates Ntcp promoter. Mutation of the putative binding site in mutNtcp promoter exhibited a significant loss of transactivation by HNF-4α (Fig. 3). The reason for the inability to completely abrogate the transactivation by HNF-4α in the mutant construct may be due to the presence of yet another HNF-4α binding site in the mutated construct. However, on the basis of TRANSFAC and MATINSPECTOR sequence analyses, no other potential HNF-4α site can be identified within the respective region of the mouse Ntcp promoter. Further proof that the transactivation by HNF-4α is direct comes from the fact that the transactivation was potentiated by cotransfection with PGC-1-α (Fig. 3), as shown for a number of HNF-4α-responsive genes.
To demonstrate that HNF-4α siRNA had any effect on endogenous Ntcp levels, we transfected Hepa 1–6 cells with a SMARTPool siRNA for HNF-4α and measured Ntcp mRNA levels in addition to protein and mRNA levels for HNF-4α and HNF-1α. The results from this analysis shown in Figs. 4 and 5 indicate that there was significant decrease in the mRNA levels for HNF-4α and, more importantly, as expected from our hypothesis, for Ntcp. Since the HNF-1α promoter contains a binding site for HNF-4α (21), it is quite to be expected that HNF-1α protein levels dropped to 26% in HNF-4α siRNA-treated cells.
To study whether the homologous rat Ntcp promoter also possessed an HNF-4α binding site, nested deletion constructs were created and the transactivation by HNF-4α was tested. As shown in Fig. 6A, HNF-4α cotransfection experiments had a comparable effect on rat and mouse promoters when their promoters contained >−300 bp upstream of the start site. Comparison of the sequences in this region (−345/−326, Fig. 6B) between rat and mouse showed the presence of a conserved binding site in both species. The rat sequence had only two changes compared with the mouse sequence, suggesting high conservation of the binding site. Very recent studies from our group further corroborate the notion that HNF-4α also regulates Ntcp expression in rats (4). In rats, Ntcp was upregulated 1.5-fold after 48 h of food deprivation concomitant with a twofold enhanced DNA binding activity of HNF-4α.
On the basis of these data, we propose that HNF-4α is a direct transactivator for rodent Ntcp (Fig. 7). Our findings are further supported by a liver-specific knockout of HNF-4α using the Cre-lox system achieved by Hayhurst et al. (15). An analysis of a large number of genes and their expression patterns in null mice carried out in this study corroborated the data from the genomic analysis of Odom et al. (21) in that there were global alterations in expression of genes involved in multiple metabolic pathways. More specifically was the fact that the levels of Ntcp message were significantly decreased in the livers of the knockout mice compared with controls, and their sera exhibited higher bile acid levels due to their decreased uptake from portal circulation. Ntcp mRNA in HNF-1α knockout mice is also reduced by more than 90% (24). However, the facts that HNF-1α also regulates HNF-4α expression (14) and that mutation of the HNF-1α binding site in the mNtcp promoter does not result in a significant loss in promoter activity (7) strengthen the hypothesis of HNF-4α acting as the central regulator of murine Ntcp expression.
Although mouse, rat, and human Ntcp/NTCP genes exhibit a similar pattern of downregulation upon inflammatory stimuli and cholestasis (10), the underlying regulatory mechanisms are distinct. In contrast to our study, a recent study by Jung et al. (17) where they compared the regulatory profiles of rat, mouse, and human ntcp/NTCP genes concluded that the rat but not the human or mouse promoter is activated by HNF-4α. However, their study showed that all three promoters were inhibited by HNF-3α-β, whereas the human and the mouse but not the rat promoter was activated modestly by C/EBP-β. The same group also showed in a separate study that GR activates human NTCP promoter, which is potentiated by PGC-1-α and suppressed by bile acids in a short heterodimer partner-dependent manner (5). We attribute the failure of Jung et al. to see transactivation by HNF-4α in the mouse promoter to their use of very short promoter constructs for all three promoters (up to −131 bp upstream of start site). Clearly our studies using promoter constructs including sequences upstream of −300 bp showed significant stimulation by HNF-4α, which was further strengthened by EMSA, mutational analysis, and siRNA knockdown studies.
In summary, we have shown that rodent Ntcp promoters and possibly also the human counterpart are significantly transactivated directly by HNF-4α. Our studies are consistent with the recent liver-specific knockout studies in which the null mice showed elevated serum bile acid levels. Our data also further confirm the genomic scan analysis of Odom et al. (21), who employed ChIP assay, that HNF-4α is a major transactivator of multiple liver genes.
This work was supported by the Deutsche Forschungsgemeinschaft grant SFB542 TP C1 (to A. Geier), GE 1219/1-1 (to A. Geier), DI 729/3-1 (to C. G. Dietrich), and National Institutes of Health grant HD20632 (to F. J. Suchy and M. Ananthanarayanan).
The authors thank Petra Schmitz for excellent technical assistance, Dr. Yanfeng Li for help with siRNA experiments, as well as Juliane Luescher-Firzlaff and Bernhard Luescher, Institute of Biochemistry, Aachen University, for continuous support regarding the ChIP assay.
↵* A. Geier and I. Martin contributed equally to this work.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society