Hepatocyte nuclear factor 1 is essential for transcription of sodium-dependent vitamin C transporter protein 1

Abstract

Transport and distribution of vitamin C is primarily regulated by the function of sodium-dependent vitamin C transporters (SVCTs). SVCT1 is expressed in the small intestine, liver, and kidney, organs that play a vital role in whole body vitamin C homeostasis. Despite the importance of this protein, little is known about regulation of the gene encoding SVCT1, Slc23a1. In this study, we present the first investigation of the transcriptional regulation of human Slc23a1, identifying transcription factors that may influence its expression. A 1,239-bp genomic DNA fragment corresponding to the 5′-flanking region of Slc23a1 was isolated from a human hepatocarcinoma cell line (HepG2) and sequenced. When cloned into a reporter gene construct, robust transcriptional activity was seen in this sequence, nearly 25-fold above the control vector. Deletion analysis of the SVCT1 reporter gene vector defined the minimal active promoter as a small 135-bp region upstream of the transcriptional start site. While several transcription factor binding sites were identified within this sequence, reporter constructs showed that basal transcription required the binding of hepatic nuclear factor 1 (HNF-1) to its cognate sequence. Furthermore, mutation of this HNF-1 binding site resulted in complete loss of luciferase expression, even in the context of the whole promoter. Additionally, small interfering RNA knockdown of both members of the HNF-1 family, HNF-1α and HNF-1β, resulted in a significant decline in SVCT1 transcription. Together, these data suggest that HNF-1α and/or HNF-1β binding is required for SVCT1 expression and may be involved in the coordinate regulation of whole body vitamin C status.

maintenance of vitamin c levels is essential to normal cell function, because it participates in a wide range of biosynthetic reactions while contributing to antioxidant defense of tissue and plasma (6, 13). Primates and certain other species have lost the ability to synthesize vitamin C de novo (47) and are at risk for developing scurvy with inadequate diet (21). Proteins that facilitate vitamin C transport across membranes govern the accumulation of vitamin C from dietary sources, maintenance of vitamin C levels in plasma, and availability of vitamin C to all tissues (46). Plasma vitamin C levels are under constant regulation, where absorption from the diet or excretion from the kidneys determines overall vitamin C tissue bioavailability (15, 21). Vitamin C transport proteins provide this “tight control” of plasma and tissue vitamin C status, though the regulatory mechanisms of vitamin C transport proteins are not well understood.

Direct uptake of the reduced form of vitamin C, ascorbic acid, is ascribed to the sodium-dependent vitamin C transport (SVCT) protein family (8, 32, 44). To date, only two functional proteins of the SVCT family have been identified (SVCT1 and SVCT2), which are the products of distinct genes, Slc23a1 and Slc23a2, respectively (42). While SVCT2 is widely expressed (32, 38, 44, 45), SVCT1 expression is limited to only a few organs, notably the epithelial cells of the liver, kidney, small intestine, and pancreas (44, 45). This tissue-selective expression of SVCT1, coupled to a high capacity for ascorbic acid transport (5, 8, 22, 45), suggests a critical role for SVCT1 in bulk intestinal absorption of vitamin C as well as renal reuptake. Thus, it is believed that the regulation of SVCT1 expression represents a critical control mechanism involved in plasma vitamin C homeostasis.

While SVCT2/Slc23a2 has an essential role in prenatal development (39) and some of the regulatory pathways for SVCT2 have been elucidated (3, 14, 35, 48, 49), there is a paucity of knowledge regarding the regulation of SVCT1. To date, no in-depth investigations focusing on the regulation of Slc23a1 transcription have been performed, and information on this subject is scarce (9, 11). However, previous studies in our laboratory (26) show that expression of SVCT1 mRNA declines in the liver of aging rats, corresponding with significantly lower vitamin C status in tissue. Because the old rats show no significant declines in SVCT2 expression, it suggests a specific role of SVCT1 transcription on maintenance of vitamin C status (26). A greater understanding of the transcriptional regulation of SVCT1 is essential for further molecular studies to elucidate the role of SVCT1 on vitamin C homeostasis.

In the present study, we examined the transcriptional regulatory elements responsible for expression of human Slc23a1. Using promoter deletion constructs, we show that a 135-bp proximal segment of the 5′-flanking region is critical for transcription of this gene. Additionally, site-directed mutagenesis of reporter gene constructs transfected into HepG2 human hepatocarcinoma cells reveal a critical role of hepatic nuclear factor (HNF-1) in Slc23a1 transcription. In addition, small interfering RNA (siRNA) knockdown of either member of the HNF-1 family, HNF-1α or HNF-1β, resulted in a significant decline in SVCT1 mRNA expression. In summary, this reliance on HNF-1 binding suggests that SVCT1-mediated vitamin C absorption and cellular distribution is influenced more by carbohydrate metabolism rather than reduction/oxidation-dependent mechanisms.

MATERIALS AND METHODS

Materials.

Cell culture materials were obtained from American Type Culture Collection (ATCC; Manassas, VA) or Sigma-Aldrich (St. Louis, MO). Restriction endonucleases and Taq polymerases were obtained from New England Biolabs (Ipswich, MA). All other chemicals were reagent grade or the highest quality available.

Cell culture.

HepG2 cells, a human hepatocellular carcinoma cell line, were obtained from ATCC. Cells were maintained in Eagle's modified minimum essential medium with the addition of 10% fetal bovine serum and 1× antibiotic antimycotic solution. Cells were cultured at 37°C and 5% CO₂ and were passaged as needed upon reaching 80–90% confluence.

Slc23a1 plasmid constructs.

A 1,239-bp fragment was generated by PCR amplification of genomic DNA from HepG2 cells using high-fidelity ExTaq polymerase (Takara Bio, Otsu, Shiga, Japan) and the full-length and reverse primer pair listed in supplemental data Table 1 (supplemental data for this article can be found online at the American Journal of Physiology-Cell Physiology website). These primers were designed against GenBank sequences for human chromosome 5 and specific for the 5′-regulatory region of Slc23a1, the gene encoding SVCT1. The PCR-amplified fragment was first subcloned into pCR-II TOPO vector (Invitrogen, Carlsbad, CA) then subcloned into the pGL3-Basic vector (Promega, Madison, WI) using the HindIII and XhoI restriction enzymes (supplemental Fig. 1). This vector was designated as the “full-length” vector, and sequence analysis confirmed the fragment's identity. This segment contains the intact transcriptional start site and 25 bases downstream of that position (supplemental Figs. 1 and 2), as identified by alignment with the human genomic SVCT1 mRNA sequence (GenBank accession no. NM_005847). The nucleotide sequence for the HepG2 Slc23a1 has been deposited in the GenBank database under GenBank accession no. (GQ279099). Transient transfection experiments show that this DNA fragment was transcriptionally active, because firefly luciferase expression was detected only in cells with the Slc23a1-containing construct, with background levels expressed in empty pGL3-Basic vector controls. Deletion constructs were synthesized using the same procedure and the primers described in supplemental Table 1. The same reverse primer was used throughout.

Mutations in transcription factor binding sites were made using the Quik-Change II Site-Directed Mutagenesis kit (Stratagene, La Jolla, CA). Primer sequences were designed that contained the putative transcription factor binding sites but incorporated two base pair mutations that would remove the consensus sequence for the identified transcription factor. Sequences of primers are listed in supplemental Table 1, with reverse primer as complements to those sequences. Annealing, elongation, and digestion steps were performed as per the manufacturer's instructions. Subcloned DNA fragments and site-directed mutations were confirmed by DNA sequencing using an ABI automated DNA sequencer (Center for Genome Research and Biocomputing, Oregon State University).

Transient transfection and luciferase assay.

For DNA transfections, HepG2 cells were grown to approximately 40–50% confluence before transient transfection on six-well plates. Transient transfection of vector DNA was performed using the Effectene reagent from Qiagen (Valencia, CA) as per the manufacturer's instructions. Briefly, 0.8 μg of firefly luciferase plasmid and 50 ng of a constitutively expressed Renilla Luciferase plasmid (pRL-CMV; Promega) were complexed with 10 μl Effectene and 6.8 μl Enhancer reagent in a final volume of 100 μl. Immediately before the addition of transfection complexes to cells, 2 ml of normal growth medium were added to the reaction. Cells remained in the presence of Effectene complexes for 18 h and were then allowed to grow in complete growth media until harvest. Transfected cells were harvested 48 h after the addition of Effectene complexes and were analyzed using the Dual Luciferase Assay Kit (Promega). Renilla luciferase values were used to normalize data for total number of cells transfected per experiment, and data were then further normalized to the indicated control plasmid as a fold change or percent variation.

siRNA knockdown.

HNF-1α, HNF-1β, SVCT1, and Negative Control siRNA were obtained commercially from Ambion (Austin, TX). Transfection of siRNAs was accomplished using Lipofectamine RNAiMax Reagent (Invitrogen) with the six-well reverse-transfection protocol described in the manufacturer's instructions, using 2 pmol total siRNA and 5 μl RNAiMax reagent per well. Cells were harvested after 72 h by trypsinization and analyzed by Western blot or quantitative RT-PCR (RT-qPCR) analysis to determine the efficiency of knockdown. For experiments requiring the transfection of siRNA and DNA, the siRNA transfection was performed 24 h before transfection of the luciferase constructs and maintained as described for DNA transfections.

Protein extract preparation and Western blot analysis.

For HNF-1 protein quantification, HepG2 cells were lysed, and nuclear and cytosolic fractions were obtained by the Nuclear Extract Kit from Active Motif (Carlsbad, CA) using the manufacturer's instructions. For SVCT1 Western blots, crude plasma membrane extracts were isolated as described by Sweet et al. (41). Protein extracts were prepared for Western blotting by addition to 2× Laemmli buffer. Soluble proteins (25 μg) were loaded on a 10% Tris·HCl gels (Bio-Rad, Hercules, CA) and separated by SDS-PAGE using a Criterion XL (Bio-Rad) apparatus at 100 V. Proteins were transferred to polyvinylidene difluoride membranes using a semi-dry electroblot transfer system. After blocking, membranes were incubated with goat anti-HNF-1α (1:500 dilution; sc-6547 Santa Cruz Biotechnology, Santa Cruz, CA), goat-anti-HNF-1β (1:200 dilution; sc-7411, Santa Cruz), goat-anti-SVCT1 (1:200 dilution, sc-9924 Santa Cruz), or mouse anti-β-Actin (1:10,000 dilution; Sigma) antibodies in PBS-Tween, washed and incubated with appropriate secondary antibody. Images were obtained after the addition of chemiluminescence detection reagents (Pierce, Rockford, IL) and exposure to autoradiography film. Relative densities of the bands were digitally quantified by using ImageJ analysis software (http://rsb.info.nih.gov/ij/).

Electrophoretic mobility shift assay.

A pair of oligonucleotides encoding the human Slc23a1 promoter sequence near the HNF-1 binding site at the −34 position was synthesized. The top strand of this sequence is as follows: 5′-GGCATTGTTCAAAGTAAATCTGTAACCAGATGCCCAGCTCCG-3′, (HNF-1 site underlined). Oligonucleotides were labeled with [γ-³²P]ATP (GE Healthcare, Piscataway, NJ) by T4 polynucleotide kinase and were separated from free ATP on Sephadex G50 (GE Healthcare) columns. Probe DNA was annealed by heating to 95°C for 5 min followed by slow cooling to room temperature. Probes were also created for competitive binding studies, with the top strand sequence as follows: HNF-1 consensus probe, 5′-TAGGTTAATAATAATTAACATTA-3′; HNF-1 binding mutant probe, 5′- GGCATTGTTCAAAGTAAGACTGTAACCAGATGCCCAGCTCCG -3′. EMSA binding assays were carried out on ice in 20-μl (total volume) reactions containing 11 mM HEPES (pH 7.9), 10% glycerol, 50 mM KCl, 2.5 mM MgCl₂, 1 mM DTT, and 0.5 μg of poly(dI-dC). This mixture was incubated at room temperature for 10 min with or without competitive inhibitors (1 μl) and with or without nuclear extract from HepG2 cells (5 μg protein). This was followed by the addition of 50,000 cpm (∼30 fmol) of the above labeled oligonucleotides and incubated at room temperature for another 20 min. Control IgG or anti-HNF-1 (pan-specific; sc-8986, Santa Cruz) antibody (2 μg) was incubated in the binding mixture for the last 10 min. The mixture was electrophoresed on a 5% polyacrylamide gel with Tris-borate-EDTA (0.5×), a running buffer. After electrophoresis, the gel was then dried, and the [³²P]-DNA pattern on the gel was analyzed by overnight exposure to autoradiography film (Amersham).

RT-qPCR.

For quantitative analysis of mRNA in cells, total RNA was isolated from 500,000 HepG2 cells using the RNAqueous RNA extraction kit (Ambion). Total RNA (2 μg) was reverse transcribed using the RETROscript kit (Ambion) and the provided polyadenylate primers. TaqMan primers specific to human HNF-1α, HNF-1β, SVCT1, and β-actin were obtained from Applied Biosystems (Foster City, CA). PCR amplification reactions were assembled using the TaqMan Gene Expression Array Master Mix per the manufacturer's instructions. Reactions were run on an MJ Research DNA Opticon 2 (Bio-Rad) for 40 cycles. Messenger RNA levels were estimated by the ΔΔC_t method (where C_t is cycle threshold) from the β-actin internal standard and are represented as changes from untreated cells.

Statistical analysis.

Data are represented as the arithmetic mean ± SE of a minimum of three separate experiments. Statistical differences between two experimental conditions were evaluated using the two-tailed Student's t-test.

RESULTS

Analysis of human Slc23a1 promoter.

The 5′-regulatory region of the human Slc23a1 gene was inserted upstream of the firefly luciferase reporter gene. To determine the cis-acting element(s) required for SVCT1 expression, constructs containing progressively smaller areas of this putative promoter region were also created (Fig. 1, left). Full-length reporter gene activity (Fig. 1, right) was approximately 25 ± 3-fold higher that the empty pGL3-Basic vector, and it remained constant in the Δ1–Δ3 constructs. In contrast, construct Δ4, containing only 160 base pairs of this regulatory sequence, exhibited a nearly 56-fold (±3.7) increase in activity over the empty vector, a significant (P < 0.01) 2-fold increase of luciferase activity over the full-length sequence. These results are suggestive of a repressive motif lying between the Δ3 and Δ4 constructs. Further deletion beyond the Δ4 fragment silenced all reporter gene activity, because luciferase production in the Δ5 and Δ6 constructs was no longer significantly higher (P > 0.9) than the empty pGL3-Basic vector. Thus, it was concluded that the minimal promoter element needed for Slc23a1 expression is contained within the Δ4 fragment which consists of the bases between −145 to +25 of the full-length Slc23a1 5′-flanking region.

Fig. 1.

Functional analysis of the human Slc23a1 promoter in HepG2 cells. Left: fragment sizes of the Slc23a1 promoter inserted in the pGL3-Basic vector construct. The full-length construct (Full) contains 1,239 base pairs of the genomic sequence stretching from the positions −1,214 to +25 bp. Right: transient transfection of HepG2 cells with the corresponding vector construct and resulting luciferase production. Data are expressed relative to the pGL3-Basic vector without DNA inserted. *Significant changes (P < 0.01) in activity.

Comparison of the Slc23a1 GenBank DNA sequence from a number of diverse species (e.g., rat, dog, mouse, chimpanzee, and human) reveals high homology in the Slc23a1 5′-flanking region proximal to the transcription start site. This highly conserved region encompasses the minimal promoter sequence, the Δ4 Slc23a1 construct, and some sequences found between the Δ3 and Δ4 constructs. Thus, these highly conserved areas may harbor regulatory elements that govern basal Slc23a1 expression. Using the MatInspector web tool from Genomatix and the Match algorithm to identify potential transcription factor binding elements (supplemental Fig. 2), multiple sites of potential regulation were noted in the Δ4 fragment. Since the Δ5 construct was not transcriptionally active, we limited the search for potential transcription factor binding sites toward the distal region of the Δ4 fragment. Interestingly, the conserved elements in this upstream region of Slc23a1 were few and included binding sites for HNF-1 at the −34 and −57 positions, a Tal-1β site at −43, as well as a TATA box at position −120 (Fig. 2A).

Fig. 2.

Hepatic nuclear factor 1 (HNF-1) binding site is essential for basal transcription in the Slc23a1 minimal and full-length promoter. A: sequence analysis of the Δ4 fragment of the Slc23a1 5′-flanking region reveals several putative regulatory elements, which are designated by name and starting position of the site relative to the transcription start (+1) on the diagram. Below is a depiction of the Slc23a1 reporter gene constructs Δ4–Δ6. B: luciferase activity of the altered transcription factor binding constructs. The unaltered vector represents the unmodified Slc23a1 Δ4 reporter vector. *Significant decline from empty vector controls (P < 0.01; n = 3). C: reporter gene constructs containing the full-length (1,239-bp) Slc23a1 sequence containing or lacking a functional HNF-1-binding site (designated HNF1-1). *Significant decline in the HNF-1 mutant vector (P < 0.01; n = 3) from empty vector control.

The relative contribution of each binding site was assessed by site-directed mutagenesis of the Δ4 reporter gene construct. These altered-binding-site vectors were transiently transfected into HepG2 cells, and the effects of specific mutations on gene expression were analyzed by measuring luciferase activity. Mutations in either the Tal-1β site or the TATA box did not diminish luciferase reporter gene expression vs. wild-type constructs; however, mutation of the first HNF-1 site (designated HNF1-1) resulted in a 50-fold (±2.4) loss of transcription (Fig. 2B). The decline in luciferase activity was so complete that this mutation essentially abrogated all basal reporter gene activity. Alterations in the second HNF-1 binding site (designated HNF1-2) resulted in a 50% (±10%) loss relative to activity observed in the unaltered Δ4 construct. Deletion of both HNF-1 binding sites or the removal of any other putative element binding in conjunction with the HNF1-1 mutation showed no significant changes in luciferase expression from the single HNF1-1 mutation (data not shown).

To further define the essential nature of the HNF1-1 site as a regulatory factor for Slc23a1 in the context of the entire promoter sequence, the HNF1-1 binding site was altered in the full-length vector. After transient transfection into HepG2 cells, expression of the HNF1-1 altered vector displayed luciferase activity that was <5% of that seen for the wild-type vector (Fig. 2C). This loss in expression was similar in magnitude to the HNF1-1 mutant constructs in the Δ4 sequence and was indistinguishable from the empty pGL3-Basic vector. From these studies, we conclude that the HNF1-1 binding site is essential for SVCT1 expression, regardless of the particular Slc23a1 promoter fragment tested. This further suggests that binding of the HNF-1 transcription factor is necessary for gene expression.

Association of HNF-1 with the Slc23a1 promoter.

To demonstrate whether members of the HNF-1 family bound to the Slc23a1 promoter, EMSAs were performed using oligonucleotides specific for the HNF1-1 site of Slc23a1. Incubation of this radiolabeled promoter sequence with HepG2 nuclear extracts resulted in the appearance of a single, strong band near the top of the gel, indicating a shift in the free probe and the formation of a protein-DNA complex (Fig. 3, lane 2). Addition of antibodies raised against HNF-1 proteins to the EMSA reaction resulted in a supershift in the probe complex (Fig. 3, lane 3), suggesting that the protein bound to the probe was HNF-1. Additionally, competition with unlabeled oligonucleotides specific for an HNF-1 consensus sequence resulted in the complete loss of the shifted band (Fig. 3, lane 4). However, when excess cold probe containing the HNF1-1 binding site mutant was added (Fig. 3, lane 5), the band remained, confirming that the altered binding site no longer bound HNF-1. These data suggest that HNF-1 binds to its respective cognate recognition element in the Slc23a1 promoter and may play an essential role in regulating human SVCT1 transcription.

Fig. 3.

HNF-1 binding to the Slc23a1 promoter. Electrophoretic mobility shift assay (EMSA) with oligonucleotides specific for the HNF1 element in the Slc23a1 reporter gene construct. Radiolabeled probes (Probe) were run alone (lane 1) or with nuclear extracts (Extract) from HepG2 cells (lanes 2–5). While nuclear extracts alone showed a shift in probe mobility (lane 2), addition of HNF-1 antibodies (HNF1ab) resulted in a supershift (lane 3), demonstrating an interaction between HNF-1 proteins and the Slc23a1 promoter. Addition of excess unlabeled consensus probe (HNF1cp) to the reaction resulted in a complete loss of signal (lane 4), but excess unlabeled probe with the binding site mutation (HNF1mut) did not (lane 5). Figure is representative of 3 individual experiments.

Expression of the HNF-1 family is required for SVCT1 transcription.

Antibodies used for EMSA analysis were pan-specific for both HNF-1 isoforms. HNF-1α is highly expressed in HepG2 cells, but RT-qPCR analysis shows that HNF-1β mRNA is also present in these cells, though at levels 8- to 9-fold lower than the α-isoform (data not shown). Because HNF-1 family members can cooperatively regulate gene promoters, HepG2 cells were transfected with either gene-specific siRNA to HNF-1α, HNF-1β, or a negative control siRNA. Results show that HNF-1α mRNA declined approximately 9-fold only in the HNF-1α siRNA-treated cells (Fig. 4A), with no appreciable decline under HNF-1β siRNA treatment. As expected, HNF-1β mRNA declined in the HNF-1β siRNA knockdown (Fig. 4B), with no observable loss in HNF-1β message when HNF-1α siRNA was used. To confirm whether this loss in gene expression was reflective of changes in HNF-1 nuclear content, nuclear extracts from siRNA transfected HepG2 cells were analyzed by Western blot. While HNF-1α siRNA transfection resulted in a undetectable HNF-1α protein in the nucleus (Fig. 4C), HNF-1β-specific siRNAs only resulted in a 30% loss in HNF-1β nuclear abundance (Fig. 4D). These results suggest that HNF-1β protein content was not completely degraded under siRNA transfection, even when longer or shorter transfection times were used (data not shown). This may indicate that HNF-1β is under a more complex translational regulation, or that is has a relatively long half-life in HepG2 cells.

Fig. 4.

Confirmation of the specificity of HNF-1 small interfering RNA (siRNA) knockdown. HepG2 cells were cotransfected with or without gene-specific or negative control siRNAs as described. mRNA in untreated cells was arbitrarily set at 1. *Significant changes (P < 0.01). A: HNF-1α mRNA levels in siRNA-transfected cells (n = 4). B: HNF-1β mRNA levels in siRNA-transfected cells (n = 3). C and D: corresponding Western blot analysis of nuclear extracts from transfected cells confirmed losses in HNF-1α (C) and HNF-1β (D) protein with only the gene-specific siRNA. Blots are representative of 3 separate experiments.

Interestingly, both HNF-1α and HNF-1β siRNA treatment resulted in a loss of SVCT1 transcription (Fig. 5A). In the HNF-1α-depleted cells, SVCT1 production was barely detectible, showing a 20-fold decline from untreated cells. Although not as robust as the HNF-1α siRNA transfection, SVCT1 mRNA expression was 5-fold lower upon HNF-1β knockdown. No change in SVCT1 mRNA expression was observed in negative control transfections or in untreated cells. These changes in SVCT1 production were then confirmed by Western blot analysis of crude plasma membrane preparations (Fig. 5B), where the levels of SVCT1 protein were approximately 63% and 55% of controls in the HNF-1α and HNF-1β knockdowns, respectively. Although the immunoreactive band shown in Fig. 5B has a lower apparent molecular mass than the predicted molecular weight for SVCT1 (∼55 kDa), it corresponds with current literature reports showing SVCT1 migrates on SDS-PAGE to approximately 35–40 kDa (35). The specific loss of SVCT1 was confirmed by the use of SVCT1-specific siRNA (Fig. 5B) which showed an approximate 80% decline in plasma membrane SVCT1 protein level.

Fig. 5.

siRNA knockdown of HNF-1 transcription factors lowers sodium-dependent vitamin C transporter 1 (SVCT) expression. HepG2 cells were transfected with gene-specific siRNAs as described. mRNA in untreated cells was arbitrarily set at 1. *Significant changes (P < 0.01). A: SVCT1 mRNA levels in siRNA-transfected cells (n = 4). B: representative SVCT1 Western blot of crude plasma membrane extracts. SVCT1 siRNAs were included as a positive control for protein knockdown. Data are representative of 3 separate experiments.

To further demonstrate the involvement of HNF family members in SVCT1 transcription, HNF-1α and HNF-1β siRNAs were cotransfected into HepG2 cells with the Δ4 fragment or the full-length Slc23a1-luciferase reporter vector (Fig. 6). As predicted by the qPCR analysis, transfection of either HNF-1α or HNF-1β siRNA resulted in a significant decline (P < 0.01) in luciferase gene expression in the Δ4 fragment (Fig. 6A). Although still statistically significant (P < 0.05), there was only a 40% decline in luciferase activity with HNF-1β knockdown compared with untransfected cells. However, in cells transfected with the full-length promoter construct, HNF-1β siRNA cotransfection resulted in large variations in luciferase activity that were no longer statistically significant (P > 0.6) from untreated cells (Fig. 6B). Again, knockdown of HNF-1α resulted in a nearly complete ablation of Slc23a1-luciferase activity, which was no longer significantly different from pGL3-Basic controls. Together, these data suggest that HNF-1α and HNF-1β regulate SVCT1 transcription, though the influence of HNF-1β may be limited by other factors binding in the entire promoter sequence.

Fig. 6.

Slc23a1 reporter gene activity is primarily modulated by HNF-1α. HepG2 cells were cotransfected with or without gene-specific siRNAs and reporter gene constructs as described. Luciferase activity is represented as fold change from activity measured in pGL3-Basic vectors. *Significant changes (P < 0.05). A: luciferase activity in Δ4 deletion construct after HNF-1 siRNA transfection (n = 3). B: luciferase activity in full-length construct after HNF-1 siRNA transfection (n = 3).

DISCUSSION

Slc23a1/SVCT1 is expressed in organs that regulate the dietary availability (5) and renal reabsorption (20) of vitamin C, and its activity contributes greatly to plasma vitamin C status. Thus, the regulation of Slc23a1 is a critical part of understanding vitamin C homeostasis. HNF-1 proteins (HNF-1α and HNF-1β) are expressed in the intestine, kidney, skin, lung, pancreas, testis, and ovary as well as the liver (4). Interestingly, these are the same tissues that express SVCT1 (37, 44). This study is the first to support this link between HNF-1 regulation and SVCT1 expression. Our initial characterization of the 5′-flanking region of human Slc23a1 revealed a critical HNF-1 binding site between positions −40 and −57 on the Slc23a1 promoter. HNF-1 proteins bound to the Slc23a1 promoter at this position, and siRNA knockdown of either HNF-1α or HNF-1β resulted in a significant loss of SVCT1 expression. The experiments demonstrated that deficiency in HNF-1 transcription factors reduced SVCT1 levels at the plasma membrane. Together, these results strongly suggest hepatic nuclear factors as principal regulatory proteins for bulk vitamin C transport through the control of SVCT1 transcription.

Regulation of HNF-1 protein levels and binding to target genes are mediated by a complex web of transcription factors and cell signaling pathways (27). HNF-1 proteins regulate the production of many secretory proteins, including insulin (25), and are necessary for the expression of many solute carriers (18, 27, 28). Our data show that Slc23a1 can be added to this growing list of genes under the functional regulation of HNF-1. However, it is difficult to draw a direct relationship between vitamin C status and HNF-1α and HNF-1β expression, because they are only two members of a complex transcription factor network. HNF-1α is under the direct control of another hepatic nuclear factor, HNF-4α (19), which is in turn influenced by insulin signaling (16) and by many protein kinases that are responsive to fasted/fed state (17). HNF-1β is linked to HNF-6 expression, where it plays critical roles in embryonic development (23) and cell differentiation (7). Declines in HNF-1α, HNF-1β, and HNF-4α activity have been linked to maturity onset diabetes of the young (MODY) (12), and impaired glucose handling, but no associations have been reported with MODY and vitamin C status. Overall, it appears that the primary regulator of SVCT1 expression is HNF-1, and this suggests a link between vitamin C transport and carbohydrate and lipid metabolism.

While HNF-1α is the primary member of the HNF-1 family expressed in the liver, HNF-1β is also detectable in many other cell types that express Slc23a1. This is not surprising, because HNF-1α and HNF-1β may form homo- or heterodimers to activate target genes (2, 34). Since the individual loss of HNF-1α or HNF-1β results in declines in SVCT1 expression, it is difficult to determine the composition of the HNF-1 dimers binding to this element. Additionally, with two adjacent HNF-1 binding sites, designated HNF1-1 and HNF1-2, present in the SVCT1 promoter, HNF-1α and/or HNF-1β dimers may bind to separate genetic sequences depending on their composition. Since the HNF1-2 site did not appear to control SVCT1 transcription to the same degree as binding to the HNF1-1, it might serve only to modulate levels of gene transcription through alternate signaling pathways governing HNF-1β rather than the abundantly produced HNF-1α. It is possible that coordinate binding by HNF-1α and HNF-1β on both of these regulatory regions may be necessary for full SVCT1 expression, for inducible regulation in times of need.

The association of HNF-1 with SVCT1 and possible control over plasma vitamin C levels suggests some interesting associations with chronic diseases. Recently, HNF-1α has been the focus of cardiovascular disease research (1), because HNF-1α controls cholesterol and lipoprotein metabolism (1) and production of inflammatory mediators (33, 50), and it may have links to the incidence of coronary artery disease (10). Although several links can be drawn between plasma vitamin C levels and cardiovascular disease, it is yet unclear what role HNF-1α plays in both processes. HNF-1α may represent a central regulatory point by which many outcomes, including whole body vitamin C status through the regulation of SVCT1, may be controlled. Furthermore, this regulation would appear to be completely separate from cellular levels of oxidative stress and antioxidant status which regulate the expression of SVCT2 (3, 24, 29, 30, 36, 40). We hypothesize that vitamin C accumulation via SVCT1 is influenced by dietary factors and glucose handling, rather than the redox nature of ascorbic acid.

Although our analysis focused on the regulation of HNF-1 in the Δ4 fragment of the Slc23a1 promoter, we also noted a possible repressor element between the Δ3 the Δ4 sequences. However, attempts at defining a repressive locus by altering transcription factor binding sites in this region failed to identify any areas of the promoter that may explain this change in luciferase activity (data not shown). Complex regulation through tertiary structure of DNA in this area, such as the formation of hairpin structures, would go undetected by a directed mutagenesis approach as employed in our study. Although computational analysis of this domain showed no thermodynamically favorable higher-order structures, further studies are warranted to determine the interaction of these elements with the core HNF-1 binding that is necessary for the regulation of this vitamin C transporter.

The absolute requirement of HNF-1 proteins for the transcriptional activation of Slc23a1 suggests that the regulation of vitamin C transport is similar to that of other carbohydrates, such as glucose, and may not be functionally related to antioxidant activity. Although future studies are needed to examine the contribution of upstream signal transduction pathways to HNF regulation of Slc23a1 transcription, there are indications for the involvement of the insulin signaling pathway in this process. However, we are aware that HepG2 cells may be an inappropriate model for the ongoing study of glucose-regulated HNF-1 transcription of SVCT1, because these cells display altered carbohydrate metabolism compared with normal liver tissue (43). Any future studies should involve in vivo monitoring of SVCT1 transport in response to alterations in HNF-1 status.

In summary, the data presented here elucidate a previously unrecognized regulatory pathway for Slc23a1 and the regulation of vitamin C uptake. Thus, like the regulation of dehydroascorbic acid uptake by the glucose transporter family (31), active ascorbic acid accumulation by SVCT1 may be influenced more by diet and insulin sensitivity than oxidative stress mechanisms.

GRANTS

This work was supported by an American Heart Association pre-doctoral fellowship (no. 0215223Z), the National Institutes of Health (2P01 AT002034-06 and 2R01 AG017141-06), and a National Institutes of Environmental Health Sciences T32 training grant (ES00240).