First published August 15, 2001; 10.1152/ajprenal.00092.2001.—Na-phosphate (Pi) cotransporters in the apical membrane of renal proximal tubular cells play a major role in the maintenance of Pi homeostasis. Although two such cotransporters, Npt1 and Npt2, have been identified, little is known about the function and regulation of Npt1. We cloned and characterized the murine (Npt1) and human (NPT1) genes, isolated the 5′-flanking region ofNpt1, and analyzed its promoter activity. Npt1 is ∼29 kb with 12 exons, whereas NPT1 is ∼49 kb with one additional exon. The Npt1 promoter has a TATA-like box but no CAAT box, and the transcription start site was identified by primer extension and 5′-rapid amplification of cDNA ends. Transfection of opossum kidney cells with Npt1 promoter-reporter gene constructs demonstrated significant activity in a 570-bp fragment that was completely inhibited by cotransfection with the transcription factor, hepatocyte nuclear factor (HNF)-3β. Deletion of 200 bp from the 3′-end of the 570-bp fragment abrogated its promoter activity. In addition, promoter activity of a 4.5-kb fragment, but not the 570-bp fragment, was stimulated fourfold by cotransfection with HNF-1α. Other well-characterized cis-acting elements were identified in theNpt1 promoter. We suggest that Npt1 expression is transcriptionally regulated and provide a basis for the investigation of Npt1 function by targeted mutagenesis.
- brush-border membrane
- hepatocyte nuclear factor
there is considerable evidence to suggest that the rate-limiting step in renal phosphate (Pi) reabsorption is mediated by Na-Pi cotransporters that reside in the brush-border membrane of proximal tubular cells where the bulk of filtered Pi is reabsorbed (19). Two classes of renal Na-Pi cotransporters, type I (Npt1) (31) and type II (Npt2) (16), have been identified by expression cloning and localized to the brush-border membrane of proximal tubular cells (2, 7). Studies in our laboratory demonstrated that Npt2 is by far the most abundant Na-Pi cotransporter in mouse kidney (29) and that disruption of the Npt2 gene in mice results in increased urinary Pi excretion, an 85% loss in brush-border membrane Na-Pi cotransport, and significant hypophosphatemia (1). Moreover, mice homozygous for the disrupted Npt2 gene fail to respond to Pideprivation with an adaptive increase in brush-border membrane Na-Pi cotransport (12) and to parathyroid hormone (PTH) with a decrease in transport (34). These findings underscore the importance of Npt2 in the maintenance of Pi homeostasis and indicate that it is the target for regulation by dietary Pi and PTH, major regulators of renal Pi reabsorption.
In contrast to Npt2, the precise function of Npt1 is not clear. Npt1 accounts for ∼13% of total Na-Pi cotransporter mRNAs in mouse kidney (29) and is also expressed in liver (15, 31) and brain (17). Stable transfection of MDCK and LLC-PK1 cells with Npt1 cDNA results in increased cellular uptake of Pi (25), and electrophysiological studies in cRNA-injected oocytes demonstrated that Npt1 mediates electrogenic Na-dependent transport of Pi(4). However, Npt1 also induces a Cl−conductance that is inhibited by Cl− channel antagonists and organic anions (4) and mediates the transport of anionic drugs, such as benzylpenicillin (33). On the basis of these findings, it was suggested that Npt1 not only functions as a Na-Pi cotransporter but also serves as a channel for Cl− transport and the excretion of anionic xenobiotics. In addition, more recent studies have suggested that Npt1 may function as a modulator of intrinsic Pi transport, rather than as a Na-Pi cotransporter per se (3).
Consistent with the differences in Npt1- and Npt2-mediated transport function are the differences in their pattern of regulation. RenalNpt1 gene expression is not modulated by either dietary Pi intake (12, 29, 30) or PTH (34). In addition, renal abundance of Npt1 mRNA and protein is not upregulated by Npt2 gene disruption (12, 34). Of interest are the findings in rat hepatocyte cultures that Npt1 mRNA expression is increased by insulin in the presence of glucose and decreased by glucagon and cAMP (15). In addition, in intact rats, Npt1 mRNA expression in liver and kidney is decreased by fasting and increased by streptozotocin-induced diabetes (15). Moreover, studies in rat hepatoma cells demonstrated that insulin regulation of Npt1 mRNA abundance is mediated through the phosphatidyl 3-kinase/p70 ribosomal S6 kinase pathways (32). However, the mechanism whereby these signaling pathways increase Npt1 gene expression remains to be determined.
As a first step toward understanding the precise physiological role of Npt1 and its regulation, we cloned and characterized the murine (Npt1) and human (NPT1) genes, isolated the 5′-flanking region of Npt1, and analyzed its promoter activity. We demonstrate the presence of well-characterized cis-acting elements in the Npt1 promoter and provide evidence for regulation of Npt1 promoter activity by hepatocyte nuclear factor (HNF)-1α and HNF-3β, hepatocyte transcription factors that regulate gene expression in a variety of tissues. Our data suggest thatNpt1 gene expression is regulated at the transcriptional level and provide a basis for the investigation of Npt1 gene function by targeted mutagenesis.
MATERIALS AND METHODS
Southern blot analysis.
Genomic DNA, isolated from mouse (129/SvJ strain) liver and human peripheral blood leukocytes (26), was digested with restriction endonucleases (5 U/μg). The digests were resolved on 0.8% agarose gels and transferred to nitrocellulose membranes [BA-(S)85, Schleicher & Schuell, Xymotech Biosystems, Montreal, Quebec, Canada] (26). Full-length murine and humanNpt1/NPT1 cDNAs (gift of Dr. M. R. Hughes, Georgetown University, Washington, DC) (5, 6) were radiolabeled with [α-32P]dCTP (3,000 Ci/mmol; ICN Biomedicals, Irvine, CA), and hybridization was performed as previously described (10). The blots were washed twice with 2× standard saline citrate (SSC)/0.1% SDS at room temperature for 5 min each and exposed to a Kodak Biomax MR film (Kodak) at −70°C.
PCR amplification and DNA sequencing.
The mouse and human Npt1/NPT1 genes were cloned and characterized using Long Amplification-PCR (LA-PCR), and the primers are listed in Table 1. Primer sequences were derived from the Npt1/NPT1 (NaPi-1)cDNA sequences (GenBank accession nos. X77241 and X71355). LA-PCR was achieved with Elongase (Life Technologies) or Expand Long Template enzyme mix (Roche Diagnostics, Laval, Quebec, Canada) as follows: initial denaturation step at 94°C, 40 cycles of amplification, and a final elongation step at 68°C for 20 min. The template-primer annealing temperatures were generally 2°C below T m (melting temperature, Table 1). PCR products <8 kb were purified and subcloned into pCR2.1 using the TOPO TA cloning kit (Invitrogen, Carlsbad, CA). Genomic Npt1and NPT1 subclones were sequenced by an automated dye termination method at the Sheldon Biotechnology Center (McGill University) or at the DNA Synthesis and Sequencing Facility of the Université Laval (Quebec, Quebec, Canada).
Genomic NPT1 clones.
To characterize the 3′-end of the NPT1 gene, we did a blast search in GenBank (nr database), using as probe a segment of DNA that encompasses the NPT1 promoter region (GenBank accession no. D83236). Two presequenced BAC clones containing a large fragment of the human gene (RPCI-11–317E16.TJ and HS3223A2E06) were identified, purchased from Research Genetics (Huntsville, AL), and subjected to PCR screening using primer pairs derived from the human NaPi-1 (NPT1) cDNA sequence. DNA from positive BAC clones was purified and sequenced to identify intron/exon boundaries by methods described above.
Isolation of Npt1 promoter region.
The 5′-flanking region of the Npt1 gene was isolated by screening a 129/SvJ mouse genomic BAC library with a 205-bp Npt1cDNA fragment derived from 5′-rapid amplification of cDNA ends (5′-RACE; see 5′-RACE and primer extension). One positive clone was identified by Southern blotting and confirmed by PCR, using primers F-61/R25 (Table 1). To establish a restriction map of the 5′-region of Npt1 gene, the BAC clone was digested with several restriction enzymes for Southern analysis and hybridized with the 32P-labeled 205-bp 5′-RACE cDNA fragment. TheNpt1 BAC clone was digested with KpnI andHindIII, and the DNA fragments were shotgun subcloned into pBluescript KS(−) digested with the same enzymes. The resulting recombinant clones were screened by PCR, and a clone containing a 4.8-kb fragment of the Npt1 promoter region (designated pKS-Npt1-4800) was purified and sequenced as described above.
5′-RACE and primer extension.
5′-RACE was performed according to the method of Ranasinghe and Hobbs (Elsevier Trends Journals Technical Tips online,http://tto.biomednet. com) with minor modifications. Total RNA (15 μg), extracted from mouse kidney using Trizol reagent (Life Technologies), was reverse transcribed with Superscript II RT (Life Technologies) using a Npt1-specific antisense primer R255 (Table 1). The cDNA-RNA hybrid was treated with RNase A, phenol-chloroform extracted, and ligated in pBluescript KS(−) cut withEcoRV. PCR was performed using the pBluescript-cDNA-RNA complex as template with T7 primer and Npt1-specific antisense primer R25 (Table 1). PCR products were subcloned into pCR2.1 and sequenced. Primer extension was accomplished with a Primer Extension System (Promega, Madison, WI) as described by the supplier, using two 32P-labeled Npt1-specific antisense primers (R25 and R70; see Table 1) located in exons 2 and3, respectively.
Preparation of Npt1 reporter plasmids.
Six different constructs were generated using the promoterless luciferase reporter plasmid pGL3-basic (Promega) by PCR-based strategies using the 4.8-kb Npt1 promoter fragment as template. PCR was performed with high-fidelity DNA polymerases (ELONGASE or Pwo). A 570-bp fragment, which includes exon 1and a short sequence of 5′-region, was PCR amplified with primers MP7 and MP8 (Table 2). The PCR product was digested with MluI and BglII and introduced, in a sense orientation, into the same sites of pGL3-basic, generating the reporter plasmid Npt1-570. The primer pair MP5/MP11 (Table2) was used to PCR amplify a 4,500-bp fragment, spanning exon 1 and the 5′-genomic region. The PCR product was digested withMluI and HindIII, whose sites were engineered by the primers, subcloned in a sense orientation upstream of the luciferase gene in pGL3-basic, and designated Npt1-4500. The same 4,500-bp fragment was inserted into pGL3-basic, in an antisense orientation, yielding Npt1–4500R. Three additionalNpt1 promoter constructs, Npt1–200,Npt1-570Δ200, andNpt1–4500Δ200, were generated by PCR, using, respectively, primer sets MP23/MP11, MP7/MP24, and MP5/MP24 (Table 2), and inserted in the reporter plasmid pGL3-basic. The reporter constructs Npt1-570Δ200 andNpt1-4500Δ200 were lacking 200 bp at the 3′-end ofNpt1-570 and Npt1–4500, respectively.
Cell culture, transient transfection, and reporter assays.
Opossum kidney (OK) proximal tubule cell line OK/E13 (obtained from Dr. J. Cole, University of Missouri, Columbia, MO), mouse distal convoluted tubule (MDCT) cell line (28), hepatoma cell line HepG2 [obtained from Dr. C. Deal, Ste. Justine Hospital Research Centre, Montreal, Quebec, Canada (20)], COS-1, and HEK-293 cells were maintained in DMEM supplemented with 10% serum (5% bovine serum + 5% fetal calf serum) at 37°C in a humidified atmosphere of 95% air-5% CO2. At 80% confluence, cells were cotransfected with 0.4 μg of the Npt1 promoter-luciferase reporter plasmid and pCMV-βgal plasmid, an internal standard for transfection efficiency, using LipofectAMINE 2000 (Life Technologies). In experiments involving HNF-1α and HNF-3β [kindly provided by Drs. C. Goodyer (McGill University, Montreal, Quebec, Canada), E. Holthuizen (University Medical Center, Utrecht, The Netherlands), J. Crabtree (Stanford University, Stanford, CA), and R. H. Costa (University of Illinois, Chicago, IL)], cells were cotransfected withNpt1 promoter-reporter gene constructs, HNF-1α or HNF-3β constructs, and pCMV-βgal. Positive and negative controls were carried out using, respectively, pGL3-control (containing the luciferase coding region under the control of the SV40 early promoter) and pGL3-basic (promoterless reporter). For enzymatic assays, transfected cells grown on 24-well plates for 24 h (48 h for MDCT) were lysed with buffer provided with the luciferase reporter assay kit (Roche Diagnostics). β-Galactosidase activity was measured using the Galacto-Star kit (Tropix, Bedford, MA) according to the manufacturer's protocol. An EG & G Berthold Luminometer (Fisher Scientific, Montreal, Quebec, Canada) was used for all enzymatic assays, and experiments in quadruplicate wells were repeated at least three times.
Organization of the Npt1 and NPT1 genes.
Southern blot analysis of murine (Fig. 1) and human (data not shown) genomic DNA, using the corresponding full-length cDNAs as probes (5, 6), was used to estimate the size and complexity of the Npt1 and NPT1genes. Based on digests with at least seven restriction enzymes, or combinations thereof, an approximate size of 23 kb was estimated for the Npt1 and NPT1 genes. This estimate is below that determined by genomic cloning (see below), indicating that large regions of the genes were not detected with the probes used. Blots washed at both high and low stringency gave similar results, suggesting that both the Npt1 and NPT1 genes are single-copy genes.
The Npt1 and NPT1 genes were cloned using a long-range PCR approach. The organization of the Npt1 gene, the PCR-amplified genomic fragments used to generate its structure, and a partial restriction map of the gene are shown in Fig.2. The mouse gene is ∼29 kb and consists of 12 exons, with the ATG translation start site in exon 2 and the translation termination site in exon 11 (Fig.3). The human gene is larger (∼49 kb) and has one additional exon, the translation initiation site is inexon 2, and the translation termination site is inexon 12 (Fig. 3). Despite differences between the mouse and human genes, the amino acids and codon usage at intron/exon boundaries are remarkably conserved (Table 3). Moreover, nucleotide sequences at the intron/exon boundaries conform to the GT/AG rule for donor and acceptor splice sites (Table 3). Although several genomic databases were searched using NPT1 cDNA sequences downstream from nucleotide 1,726 (GenBank accession no.D28532), we were unable to identify the 3′-acceptor site forintron 13. Rather, the genome search revealed a high homology with several cloning vectors, suggesting that the 3′-NPT1 cDNA sequences in this region may not be part of theNPT1 gene.
Transcription initiation site and features of the 5′-flanking region of the Npt1 gene.
Primer extension and 5′-RACE were performed to determine the Npt1transcription start site. Products of 118 and 159 bp were detected using antisense primers R25 and R70, respectively (Fig.4 A). This located the transcription start site to position −95 relative to the first nucleotide of the translation initiation site (designated1). To confirm these findings, 5′-RACE was performed with antisense primer R25. The sequence of the resulting product depicted in Fig. 4 B is consistent with the primer extension data.
The nucleotide sequence of the 4.8-kb Npt1 promoter fragment generated from a mouse genomic BAC clone (Fig. 2) was deposited in GenBank (accession no. AF361762). The region was 50% G + C, and a TATA-like box was identified 37 bp upstream of the transcription start site. However, a classical CAAT-box was not found in close proximity to the transcription initiation site.
The 4.8-kb Npt1 promoter fragment contains a number of potential cis-acting elements recognized by well-characterized transcription factors (Table 4). These include associated protein (AP) 1, AP2, AP4, glucocorticoid response element (GRE), HNF-1α, HNF-3β, and thyroid hormone receptor-α1 binding sequence (T3R), and suggest that Npt1 gene expression may be highly regulated at the transcriptional level.
Transcriptional activity of the Npt1 promoter.
To characterize the transcriptional activity of the 5′-flanking region of the Npt1 gene, Npt1 promoter-luciferase reporter constructs were transfected into a variety of cell lines. OK cells, derived from the renal proximal tubule of opossum kidney, were previously used to examine NPT1 (27) andNpt2 promoter activity (11). MDCT cells were derived from mouse distal tubule (21) and shown to expressNpt1 mRNA in addition to type III Na-Picotransporter mRNAs (28). HepG2 cells, derived from a human hepatoma, were also studied because Npt1 expression was documented in rat hepatoma cells (32) and in rat (15, 32) and mouse (33) hepatocytes. In OK cells, a 570-bp fragment (Npt1–570) that includesexon 1 was able to induce the highest luciferase activity compared with constructs containing longer (Npt1–4500) and shorter (Npt1–200) 5′-fragments (Fig.5 A). These data suggest that negative regulatory elements are present upstream ofNpt1–570 and that the 200-bp 5′-fragment is not sufficient to support transactivation. Npt1–570 was also able to drive expression of the luciferase reporter gene in MDCT and HepG2 cells, albeit at a lower level than in OK cells (Fig.5 A), but not in COS-1 cells (Fig. 5 A) or HEK-293 cells (data not shown), suggesting tissue specificity ofNpt1 promoter activity. Npt1–200 was not able to drive transcriptional activity in MDCT, HepG2, or COS-1 cells, consistent with data in OK cells (Fig. 5 A). Deletion of the 200 bp from Npt1–4500 and Npt1–570significantly inhibited promoter activity in OK cells (Fig.5 B), indicating that this region is necessary for transactivation.
Cotransfection of OK cells with transcription factor HNF-1α led to an increase in Npt1–4500R promoter activity but had no significant effect on Npt1–570 promoter activity (Fig.5 C), consistent with the presence and absence of HNF-1α consensus sequence in the respective promoter fragments (Table 4). Of interest, cotransfection with HNF-1α had no significant effect onNpt1–4500 promoter activity (data not shown), suggesting a directional effect of promoter regulation. Cotransfection with the transcription factor HNF-3β significantly inhibitedNpt1–4500 (data not shown), Npt1–4500R, and Npt1–570 promoter activity (Fig. 5 C), consistent with the presence of HNF-3β consensus sequences in all three 5′-fragments (Table 4).
The type I Na-Pi cotransporter, Npt1/NPT1, is a 465/467-amino acid protein that resides in the brush-border membrane of renal proximal tubular cells and is also expressed in liver and brain (19). Electrophysiological studies demonstrated that Npt1 mediates not only the Na-dependent transport of Pi but also induces a Cl− conductance that is inhibited by Cl− channel blockers and organic anions (4). As a first step to elucidate the physiological function of Npt1/NPT1, we cloned and characterized the mouse and human genes and demonstrated that, although they differ in size and exon number, both genes exhibit similar amino acids and codon usage at intron/exon boundaries. We also characterized the 5′-flanking region of the mouse gene and show that hepatic nuclear transcription factors HNF-1α and HNF-3β, respectively, stimulate and inhibit Npt1 promoter activity. These findings and the identification of well-characterized cis-acting elements in the Npt1 promoter suggest that Npt1gene expression is subject to transcriptional regulation.
In contrast to Npt1, there is little evidence to suggest that Npt2 gene expression is regulated at the transcriptional level (11, 18), despite the identification of several cis-acting elements in the promoter region (13). Rather, the regulation of Npt2-mediated Na-Pi cotransport occurs primarily at the posttranscriptional level (18). Functional studies (19), as well as targeted disruption of theNpt2 gene in mice (1), indicate that Npt2 is the target for regulation of renal Na-Pi cotransport by dietary phosphate and PTH, major determinants of renal phosphate reabsorption. In this regard, it is of interest in that neither dietary phosphate (12, 30), PTH (34), nor knockout of the Npt2 gene (12) has an effect on renalNpt1 gene expression.
It is well known that thyroid hormone and glucocorticoids are important regulators of renal phosphate handling (14) and that both hormones elicit their cellular effects via receptors that bind to specific DNA response elements in the 5′-region of target genes and thereby modulate gene transcription (9). Thus the demonstration of thyroid hormone and glucocorticoid response elements in the promoter region of the Npt1 gene suggests that Npt1 may be a target for this regulation. Consistent with this notion are the findings of a previous study showing that neither T3nor dexamethasone had an effect on Npt2-promoter luciferase reporter gene expression, suggesting that other targets may be involved (11).
Several well-characterized cis-acting elements have also been identified in a 1.4-kb promoter fragment of the NPT1 gene (27), demonstrating species conservation of core promoter elements. However, a comparison of Npt1 and NPT1promoter sequences exhibited little homology (data not shown). The physiological significance of these consensus sequences in theNpt1/NPT1 genes requires further investigation. In the present study, we examined the effect of two transcription factors, HNF-1α and HNF-3β, for which recognition sequences were identified in the Npt1 promoter, and demonstrate that both have the ability to modulate Npt1 gene transcription in vitro.
Our findings that HNF-1α stimulates and HNF-3β inhibitsNpt1 gene transcription may be relevant to the regulation of renal Pi handling in vivo. These transcription factors play an important role in the regulation of a variety of genes and are expressed in polarized epithelia of liver, digestive tract, pancreas, and kidney (22). HNF-1α expression in the kidney is confined to the proximal tubule (22), the segment of the nephron where the bulk of filtered Pi is reabsorbed (19) and where Npt1 mRNA (8) and protein (2) have been localized.
Recent studies demonstrated that inactivation of theHNF-1α gene in mice results in a renal Fanconi syndrome, which is characterized by severe urinary wasting of Pi, glucose, and amino acids (23). Moreover, patients with maturity-onset diabetes type 3 (MODY3), which results from mutations in the HNF-1α gene, also exhibit defects in the renal reabsorption of Pi, glucose, and amino acids (24). Clearly, further studies are necessary to determine whether the renal phosphate leak in HNF-1α-deficient mice and patients with MODY3 can be attributed to decreased renalNpt1/NPT1 gene expression. In this regard, it is of interest that decreased renal expression of SGLT2, a proximal tubular Na-glucose cotransporter that plays a key role in glucose reabsorption, can account for the renal glucose leak in HNF-1α-deficient mice (23). Moreover, HNF-1α has a direct effect onSGLT2 gene transcription, as demonstrated in cells cotransfected with SGLT2 promoter-reporter gene constructs and HNF-1α (23). These findings provide a molecular basis for the increase in urinary excretion of glucose in the mutant mouse strain and suggest that similar mechanisms may account for the renal Pi leak in the murine and human disorders. Finally, HNF-1α may also play a role in the regulation of Npt1/NPT1gene expression in the liver. Previous studies have shown that Npt1 mRNA abundance is modulated by changes in glucose metabolism in the liver (15, 32) and that mutations in HNF-1αgene in mice and humans lead to severe disturbances in glucose homeostasis (22).
In summary, we have cloned and characterized the mouse and human genes encoding the type I Na-Pi cotransporter, Npt1/NPT1,and demonstrate that hepatic nuclear transcription factors HNF-1α and HNF-3β, respectively, stimulate and inhibitNpt1 promoter activity. In addition, we identified several well-characterized cis-acting elements in the Npt1 promoter. We suggest that Npt1 expression is transcriptionally regulated and provide a basis for the investigation of Npt1function by targeted mutagenesis.
We thank Drs. N. Zhao for initiating the promoter studies, D. Leclerc for advice on genome databases, and Y. Sabbagh for computer support.
This work was supported by Canadian Institutes for Health Research Grant GR-13297 to H. S. Tenenhouse.
Address for reprint requests and other correspondence: H. S. Tenenhouse, Montreal Children's Hospital Research Institute, Rm. 222, 4060 Ste. Catherine St. West, Montreal, Quebec, Canada H3Z 2Z3 (E-mail:).
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.
First published August 15, 2001;10.1152/ajprenal.00092.2001
- Copyright © 2001 the American Physiological Society