the human paracellin-1 gene (hpcln-1): renal epithelial ... · the 7.5-kb 5-ﬂank-ing sequence...

The human paracellin-1 gene (hPCLN-1): renal epithelial cell-specificexpression and regulation

Edna Efrati,1,* Julia Arsentiev-Rozenfeld,1,* and Israel Zelikovic1,2

1Laboratory of Developmental Nephrology, Faculty of Medicine, Technion-Israel Institute of Technology,and 2 Pediatric Nephrology Unit, Department of Nephrology, Rambam Medical Center, Haifa, Israel

Submitted 20 January 2004; accepted in final form 2 September 2004

Efrati, Edna, Julia Arsentiev-Rozenfeld, and Israel Zelikovic.The human paracellin-1 gene (hPCLN-1): renal epithelial cell-specificexpression and regulation. Am J Physiol Renal Physiol 288: F272–F283, 2005. First published September 7, 2004; doi:10.1152/ajprenal.00021.2004.—Tubular reabsorption of Mg2� is mediated bythe tight junction protein paracellin-1, which is encoded by the genePCLN-1 (CLDN16) and exclusively expressed in the kidney. TubularMg2� reclamation is modulated by many hormones and factors. Theaim of this study was to define regulatory elements essential for renaltubular cell-specific expression of human PCLN-1 (hPCLN-1) and toexplore the effect of Mg2� transport modulators on the paracellin-1gene promoter. Endogenous paracellin-1 mRNA and protein weredetected in renal cell lines opossom kidney (OK), HEK293, andMDCT, but not in the fibroblast cell line NIH3T3. A 7.5-kb hPCLN-15�-flanking DNA sequence along with seven 5�-deletion productswere cloned into luciferase reporter vectors and transiently transfectedinto the renal and nonrenal cells. The highest levels of luciferaseactivity resulted from transfection of a 5�-flanking 2.5-kb fragment(pJ2M). This activity was maximal in OK cells, was orientationdependent, and was absent in NIH3T3 cells. Mg2� deprivation sig-nificantly increased pJ2M-driven activity in transfected OK cells,whereas Mg2� load decreased it compared with conditions of normalMg2�. Deletion analysis along with electrophoretic mobility-shiftassay demonstrated that OK cells contain nuclear proteins, which binda 70-bp region between �1633 and �1703 of major functionalsignificance. Deleting this 70-bp segment, which contains a singleperoxisome proliferator-response element (PPRE), or mutating thePPRE, caused a 60% reduction in luciferase activity. Stimulating the70-bp sequence with 1,25(OH)2 vitamin D decreased luciferase ac-tivity by 52%. This effect of 1,25(OH)2 vitamin D was abolished inthe absence of PPRE or in the presence of mutated PPRE. Weconclude that the PPRE within this 70-bp DNA region may play a keyrole in the cell-specific and regulatory activity of the hPCLN-1promoter. Ambient Mg2� concentration and 1,25(OH)2 vitamin Dmay modulate paracellular, paracellin-1-mediated, Mg2� transport atthe transcriptional level. 1,25(OH)2 vitamin D exerts its activity on thehPCLN-1 promoter likely via the PPRE site.

magnesium; renal tubule; transcription; promoter; gene expression; generegulation; peroxisome proliferator response element; 1,25(OH)2 vitamin D

MAGNESIUM IS THE MOST ABUNDANT divalent cation in the intra-cellular fluid. It plays a critical role in a wide variety ofmetabolic and cellular processes, including cellular energystorage, DNA/RNA processing, ion transport, membrane sta-bilization, and nerve conduction (33). Abnormalities in Mg2�

homeostasis are relatively common in clinical practice and maylead to neuromuscular disturbances, central nervous system

manifestations, and cardiovascular dysfunction (16, 30). Inmammals, the kidney is the principal organ responsible forMg2� balance (16, 30). Normally, �95% of the filtered Mg2�

is reabsorbed by the renal tubule. Ten to fifteen percent of thefiltered Mg2� is reabsorbed in the proximal tubule and 10% inthe distal tubule. The major site of Mg2� reabsorption is thethick ascending limb of the loop of Henle (TAL), where60–70% of the filtered load is reclaimed (16, 30). Mg2�

transport in this tubule segment occurs primarily throughparacellular conductance driven by the lumen positive electri-cal potential (30). While renal Mg2� handling has been thor-oughly investigated at the tubular and cellular levels (8, 16,30), the molecular mechanisms of tubular Mg2� reabsorptionare poorly understood.

Recently, Simon et al. (38) using positional cloning, haveidentified a human gene, hPCLN-1 (also known as CLDN16,NCBI accession no. NM-006580), mutations in which causefamilial hypomagnesemia-hypercalciuria syndrome. hPCLN-1consists of five exons and resides on chromosome 3q27. Thegene encodes a protein, paracellin-1, which is composed of 305amino acids (38). Northern blot analysis of human tissues hasshown that the 3.5-kb PCLN-1 mRNA transcript is expressedexclusively in the kidney (38). RT-PCR analysis of mRNAfrom nephron segments of the rabbit (38) and rat (42) hasdemonstrated that PCLN-1 is expressed in the TAL and thedistal convoluted tubule (DCT). The paracellin-1 protein,which is located in the paracellular tight junctions of the TALand DCT, is a member of the claudin family of tight junctionproteins (27) and appears to mediate resorption of both Mg2�

and Ca2� (38).In the kidney, the specialized reabsorptive and/or secretory

function of each tubule segment depends upon its structuralarrangement and upon the specific pattern of gene expressionin each tubular cell type. The promoters of several transporterand channel genes including aquaporin (28), the Na�-phos-phate cotransporter (36), the Na�-K�-Cl� cotransporter (39),and chloride channels (40) as well as the promoters of nephrin(46) and cadherin (13, 43) genes have been cloned and shownto direct kidney-specific expression in vitro and/or in trans-genic mice. Several transcription factors including myc-asso-ciated zinc finger proteins and Kruppel-like factor (41), hepa-tocyte nuclear factor-3 (39), and hepatocyte nuclear factor-1�(2), were found to be involved in kidney-specific expression ofthe ClC-K1 chloride channel, thiazide-sensitive Na-Cl cotrans-porter, and cadherin genes, respectively. However, very little isknown about the regulatory elements responsible for cell-

* E. Efrati and Julia Arsentiev-Rozenfeld contributred equally to this work.Address for reprint requests and other correspondence: I. Zelikovic, Pedi-

atric Nephrology, Rambam Medical Ctr. 8 Ha’Aliyah St., PO Box. 9602, Haifa31096, Israel (E-mail: [email protected]).

The costs of publication of this article were defrayed in part by the paymentof page charges. The article must therefore be hereby marked “advertisement”in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Am J Physiol Renal Physiol 288: F272–F283, 2005.First published September 7, 2004; doi:10.1152/ajprenal.00021.2004.

0363-6127/05 $8.00 Copyright © 2005 the American Physiological Society http://www.ajprenal.orgF272

by 10.220.33.6 on April 30, 2017

http://ajprenal.physiology.org/D

ownloaded from

http://ajprenal.physiology.org/

specific expression of transporter and channel genes in thekidney. As yet, the promoter region of the human paracellin-1gene has not been characterized, and the molecular mecha-nisms for renal epithelial-specific activity of this gene have notbeen investigated.

Many factors are known to modulate Mg2� reabsorption invarious nephron segments. These factors include hormonessuch as insulin, 1,25(OH)2 vitamin D, aldosterone, and para-thyroid hormone as well as nonhormonal factors such as Mg2�

restriction and load and acid-base changes (8, 30). Most ofthese factors influence both transcellular magnesium transportin the DCT as well as paracellular transport of this cation in theTAL (8, 30). However, very little is known about the molecularmechanisms of this modulation and whether it occurs at theprotein or DNA/RNA level. The magnesium restriction (30)-and 1,25(OH)2 vitamin D (32)-induced increase in Mg2�

uptake by renal MDCT cells was diminished by pretreatmentof cells with actinomycin D, suggesting that this stimulationoccurs through transcriptional activation. Nevertheless, themolecular mechanisms whereby hormones and other factorsmodulate Mg2� transport across the paracellular pathway inthe renal tubule are unknown.

The purpose of this study was to define cis-acting promoterregulatory elements and to examine trans-acting factors essen-tial for renal tubular epithelial-specific expression of hPCLN-1.We also explored the effect of modulators of Mg2� transporton the hPCLN-1 gene promoter. We show that a 70-bp 5�-flanking region of the paracellin-1 gene may determine renalepithelial-specific expression of this gene. We demonstrate anincrease in hPCLN-1 promoter activity in response to Mg2�

depletion and a decrease in response to Mg2� load. In addition,we show that 1,25(OH)2 vitamin D may modulate Mg2�

transport at the transcriptional level, probably via the peroxi-some proliferator response element (PPRE) contained withinthe 70-bp region.

MATERIALS AND METHODS

Opossum kidney (OK) cells (provided by Dr. J Green, Technion,Haifa, Israel), human embryonic kidney (HEK293) cells (provided byDr. K. Skorecki, Technion, Haifa, Israel), mouse distal convolutedtubule (MDCT) cells (provided by Dr. P. Friedman, University ofPittsburgh, Pittsburgh, PA), and mouse embryonic fibroblast(NIH3T3) cells were grown and maintained in DMEM/F-12 supple-mented with 10% fetal calf serum, 2 mM glutamine, 50 IU/mlpenicillin, and 50 �g/ml streptomycin at 37°C in a humidified atmo-sphere of 95% air-5% CO2.

RT-PCR followed by Southern blot analysis. Total RNA isolatedwith Tri-reagent (MRB, Cincinnati, OH) from the cell lines mentionedabove was reverse-transcribed using OmniscriptRT (Qiagen, Hilden,Germany) with random hexamers (Promega, Madison, WI). PCR wasperformed using HotStarTaq DNA polymerase (Qiagen) with two setsof primers (Table 1). The first (F1,R1) complementary to exon 1 ofhPCLN-1 and the second (F2,R2) complementary to a region betweenexons 3 and 5 of PCLN-1 (spanning two introns) highly homologousamong human, rat, and mouse (mPCLN-1). The resultant DNA wasseparated on 1% agarose gels and transferred to nylon membranes(Osmonics, Minnetonka, MN). The membranes were probed withPCLN-1 cDNA probes generated by PCR using primers F1,R1 forhPCLN-1 and primers F3,R2 (Table 1) from exon 5 for mPCLN-1,with human genomic DNA as a template. A similar procedure wascarried out with actin primers (F4,R4) as control. Probes were 32Plabeled by random priming (Biological Industries, Beit Ha’Emek,Israel) using [�-32P]dCTP (DuPont-New England Nuclear, Boston, MA).

Western blot analysis. Protein extracts from HEK293, OK, andNIH3T3 cells were prepared using standard protocols (35). Proteinswere separated on 10% SDS polyacrylamide gels, transferred tonitrocellulose membranes (Schleicher and Schuell, Dassel, Germany),and probed with anti-human paracellin-1 antibody (Santa Cruz Bio-technology, Santa Cruz, CA) using chemiluminescense (BiologicalIndustries). To verify antibody specificity, a peptide competition assaywas carried out with 100-fold excess of paracellin-derived peptide.

Cloning of hPCLN-1 5�-flanking DNA. Using the published se-quence of human paracellin-1 cDNA (NCBI accession no. NM-006580), we localized the hPCLN-1 gene on chromosome 3, drafts ofwhich have recently been deposited in NCBI (accession no. NT-00962). A DNA fragment of 7.5 kb in the 5�-flanking sequence region,to but not including the PCLN-1 translation start site (Fig. 1), wassynthesized using PCR with human genomic DNA as a template andprimers F5/R5 (Table 1). To minimize the possibility of PCR-gener-ated mutations, the High-Fidelity Long-Range PCR system (Roche,Mannheim, Germany) was used. The fragment was TA cloned intopCR-XL-TOPO vector (Invitrogen, Carlsbad, CA) and sequence ver-ified. Computer analysis (Wisconsin Package version 8.0, GeneticComputer Group) of the 5�-flanking region of hPCLN-1 from thegene’s translation start site was carried out, and putative transcriptionfactor binding sites were identified.

Generation of promoter/reporter constructs. The 7.5-kb 5�-flank-ing sequence described above was subcloned into pGL3-basic vector(Promega) upstream of a luciferase reporter gene (and named pJ12)(Fig. 1). In addition, a set of seven deletion fragments decreasing insize from the 5�-end of the 7,514-bp genomic fragment were producedand cloned into pGL3-basic vector. Five inserts were prepared usingPCR-based strategies with genomic DNA as template with primersF6, F7, F8, F9, F10, and R5 (Table 1). The additional two wereprepared by removing specific 5�-segments from cloned hPCLN-1sequences in pGL3-basic, using restriction enzymes. All eight con-structs were sequenced to verify the orientation and integrity of theinserts. As illustrated in Fig. 1, the 5�-ends of the deletion fragmentswere positioned at �4687 (pJ2L), �3986 (pJ2/3.9), �3317 (pJ2/3.3),�2554 (pJ2M), �1982 (pJ2/1.9), �1458 (pJ2.1/4), and �733 (pJ2.7),respectively. All fragments extended to the ATG site, but did not

Table 1. Primers complementary to various segmentsof the human PCLN-1 gene used in PCR

F1 ctcagcccttgcactgacctgR1 agactgacacccgccacttaagtgF2 ctggaggtgagcacaaaatgR2 ctctggtgtctacagcatacF3 gttggacctgagagaaactatcF4 tgacggggtcacccacactgtgcccatctaR4 ctagaagcattgcggtggacgatggagggF5 gaaactatcaacagagtaaacagR5 cttctgattggaggctggttgF6 gttctagcagacacattgcctacF7 ctagcatggcctttcagaacF8 gatgcactggtgaacgaaagF9 ctcccagatcagcagagtaaatcF10 cagaccatgattaaccactgF12 ttatc tatgaaattt cccatgtaat ttttgcagacF13 cgatc gctaatgtgtacagtaatca ccccttctccF14 ttgaggaata tgttcaaaga ccctcagtgg atgtcF15 agaaacgaaggacag tactggatcctacacacactF16 atacgcaaca gtaggtcaga tcaccgaggc agctaF17 ttaag tgactaaggg gtgggtagtg catacaatgtF18 ggaaatgctg gccaaagggatgattcatat cccatF19 ctatacgcaacagtatagcagatcaccgaggcR12 aaagcatctgtaatgtacctcgcagatagttataaR13 tcagaacagtgttcaattta ccactgatga attgtR14 ctgcctcggtgatctgctatactgttgcgtatag

See text for details.

F273EXPRESSION AND REGULATION OF THE HUMAN PARACELLIN-1 GENE

AJP-Renal Physiol • VOL 288 • FEBRUARY 2005 • www.ajprenal.org

by 10.220.33.6 on April 30, 2017


ownloaded from


include it. All eight constructs were cloned into the vector in senseorientation, and the 2.5-kb fragment in pJ2M was also cloned inantisense orientation.

A reporter plasmid containing a 586-bp nested deletion (fromposition �1772 to �1186) within this 2.5-kb DNA region (termedpJ2Mdel) was generated using PCR with pJ2M as template andprimers F12/R12 (Table 1) complementary to the gap ends. FollowingDpnI treatment, the PCR-generated pGL3-basic construct was ligated,and the sequence was verified. Empty pGL3 vector containing noinsert was used as a negative control. pGL3-control containing theSV40 enhancer/promoter was used as a positive control. Promoteractivity was estimated from the ratio of hPCLN-1-driven luciferaseactivity in pGL3-basic vector to promoterless pGL3-basic vector.

For reporter gene activity driven by distal hPCLN-1 promoterfragments, the DNA region from position �1738 to �1493 as well asa series of five deletion fragments 5�-truncated in increments of 35 bpwere PCR generated with human genomic DNA and primers F13,F14, F15, F16, F17, F18, with R13. All six fragments were cloned intothe pGL3-promoter vector, which carries an SV40 minimal promoterupstream to the luciferase reporter gene. In these experiments, pro-moter activity was determined from the ratio of hPCLN-1 promoter-driven luciferase activity in the pGL3- promoter vector to that in theempty pGL3-promoter vector.

Point mutations in the 2.5-kb insert-containing plasmid, pJ2M, andthe 210-bp insert-containing plasmid, pJ210 (see RESULTS), weregenerated using PCR with the mutated primers F19 and R14 (Table 1).

In all transfection experiments, plasmid pCH110 (Pharmacia, Upp-sala, Sweden), containing the LacZ gene driven by the CMV pro-moter, was used to normalize for transfection efficiency. DNA fortransfections was purified using Nucleobond AX (Macherey Nagel,Duren, Germany).

Transient transfections and reporter gene assays. OK, HEK293,and MDCT cells were plated (5 � 104/dish) in 24-well dishes inserum-containing medium. Cotransfections were performed usingFugene 6 (1.2 �l/well, Roche) with 0.3 �g reporter plasmid and 0.3�g pCH110. NIH3T3 cells (4 � 105/dish) were seeded in 6-wellplates and incubated overnight at 37°C. Cells were transfected 24 hlater using Polyfect (10 �l/well, Qiagen) with 0.75 �g reporterplasmid and 0.75 �g pCH110. All plates were incubated at 37°C for48 h. For enzymatic assays, cells were washed with PBS (150 mMNaCl, 15 mM sodium phosphate, pH 7.3) and lysed by incubating in200 �l/well M-Per (Pierce, Cheshire, UK) for 5 min at 37°C. Lysedcells were centrifuged, and the supernatant was aliquoted (50 �l/well)into 96-well plates. Fifty microliters of Luciferase Assay Reagent(Promega) were automatically added, and the light intensity of thereaction was immediately read in a luminometer (Lucy, Anthos,Austria) for a period of 10 s. Luciferase activity was normalized to

�-galactosidase activity, which was measured in identical cell lysates.One hundred sixty microliters of ONPG substrate (Sigma, St. Louis,MO) were added to 30 �l of cell lysate in each well of a 96-well plate.The reaction mixture was incubated for 30 min at 37°C, or untilyellow color developed. �-Galactosidase measurements were per-formed by a luminometer with a 405-nm filter. Measurements ofluciferase and �-galactosidase were performed in duplicate.

In some experiments, 48 h after transfection, the medium wasreplaced with fresh medium containing 0 (low), 0.7 (normal), or 1.5mM (high) Mg2� (Biological Industries). In other experiments, thefresh medium contained 1,25(OH)2 vitamin D (from a stock solutionof 10�4 M in ethanol, Sigma) at a final concentration of 5 � 10�7 M.Control experiments were carried out with 0.5% ethanol. Cells wereexposed to experimental media for 24 h.

EMSA. Nuclear extracts were prepared from cells using the methodof Dignam (9). Briefly, confluent cells were grown on 100-mm plates,washed in 3 ml of PBS (150 mM NaCl, 15 mM sodium phosphate, pH7.3) supplemented with protease inhibitor mix (Complete, Roche),scraped, and pelleted. The pellet was resuspended in ice-cold lysisbuffer (in mM: 10 HEPES, pH 7.9, 1.5 MgCl2, 420 NaCl, 0.2 EDTA,and 1 DTT as well as protease inhibitor mix) and incubated at 4°C for20 min. Following centrifugation, the supernatant, containing thenuclear extract, was diluted 1:2 with (in mM) 20 HEPES (pH 7.9), 100KCl, 0.2 EDTA, and 1 DTT as well as 20% glycerol and proteaseinhibitors. Protein concentration was measured using Bradford re-agent (Sigma) at 595 nm with bovine serum albumin as the standard.

Double-stranded oligonucleotides corresponding to promoter se-quences of interest were end-labeled with [-32P]ATP (DuPont-NewEngland Nuclear) using T4 polynucleotide kinase (NEB, Beverly,MA). Binding reactions contained (in mM) 10 Tris �HCl (pH 8.0), 250KCl, 0.5 EDTA, and 0.2 DTT as well as 0.1% Triton-X 100, 12.5%glycerol (vol/vol), 1 �g poly-dIdC-labeled probe (5 � 104 counts/min), and 50 �g of nuclear extracts. In some reactions, a 50-foldmolar excess of unlabeled double-stranded oligonucleotide was addedfor specific competition. Following a 30-min incubation period on ice,complexes were resolved on 4% nondenaturing polyacrylamide gelsin 1� TBE buffer. The gels were dried and autoradiographed.

RESULTS

Cloning the 5�-flanking region of hPCLN-1. A DNA frag-ment stretching over 7.5 kb in the 5�-flanking region ofhPCLN-1, reaching 9 bp from the gene’s translation start site,was isolated. Computer analysis of this fragment disclosed avariety of putative transcription factor binding sites, amongthem recognition sites for several transcription factors known

Fig. 1. DNA constructs of the human paracellin-1 gene(hPCLN-1) promoter in pGL3-basic vectors. Promoter frag-ments of decreasing size from the 5�-end were prepared byPCR with human genomic DNA as a template and clonedupstream to the luciferase reporter gene in pGL3-basicvectors.

F274 EXPRESSION AND REGULATION OF THE HUMAN PARACELLIN-1 GENE


by 10.220.33.6 on April 30, 2017


ownloaded from


to be tissue restricted and expressed in the kidney. Theseincluded hepatocyte nuclear factor 5 (HNF-5), GATA factors,and a peroxisone proliferator-activated receptor (PPAR) bind-ing site. The hPCLN-1 promoter also contained binding sitesfor transcription factors involved in signal transduction, suchas activator protein (AP)-1 and AP-3 as well as a putativeTATA box.

The 5�-flanking regions of PCLN-1 in mouse and ratgenomic DNA were located using NCBI-BLAST with therespective cDNA sequences. Comparison of the 5�-flankingregion of hPCLN-1 to that in mouse (accession no. NW-000107, region: 23187252 . . . 23287252) and rat (accessionno. AC106700) revealed very little overall sequence identitywithin 2.4 kb 5� to the translation site. However, the 51-bpregion from position �392 to �341, relative to the hPCLN-1translation start site, was 79 and 81% identical to the mouseand rat PCLN-1 5�-flanking region sequences, respectively.

Comparison of the hPCLN-1 5�-flanking region to that of avariety of genes encoding various ion channels and transport-ers, known to be specifically expressed in the kidney, such asthe voltage-gated Cl� channels ClC-K1 and ClC-K2, aqua-porin-2, the K� channel ROMK, Tamn-Horsfall protein, andthe thiazide-sensitive NaCl cotransporter, did not reveal areasof sequence homology.

Kidney cell-specific expression of hPCLN-1. RT-PCR fol-lowed by Southern blot analysis with nested probes (seeMATERIALS AND METHODS) detected endogenous PCLN-1 mRNAin HEK293 and OK cells using an hPCLN-1 probe and inHEK293 and MDCT cells using an mPCLN-1 probe (Fig. 2A).Paracellin-1 mRNA was not detected in NIH3T3 cells.

Western blot analysis with paracellin-1-specific antibodydetected paracellin-1 protein expression in HEK293 and OKbut not in NIH3T3 cells (Fig. 2B). Addition of 100-fold molarexcess of paracellin-1 peptide to the antibody, 24 h beforeincubation with the blot, competed out the paracellin-1 band(data not shown).

Kidney cell-specific activity of the hPCLN-1 promoter. Toverify that the 5�-flanking region of PCLN-1 contained afunctional cell-specific promoter, reporter gene assays wereconducted. The 7.5-kb hPCLN-1 5� flanking region, as well asthree 5�-deletion products, were cloned upstream to a lucif-erase reporter gene in promoterless pGL3-basic vectors insense orientation.

The resulting plasmids (Fig. 1), designated pJ12 (7,514-bpinsert), pJ2L (4,687-bp insert), pJ2M (2,554-bp insert) andpJ2.7 (733-bp insert), were transfected into the PCLN-1-ex-

Fig. 2. Cell-specific expression of paracellin-1 mRNA (A) and protein (B) inrenal cell lines. A: Southern blot analysis of mRNA amplified by RT-PCRusing primers complementary to exon 1 of hPCLN-1 (I), or primers from aregion of high homology among human, rat, and mouse PCLN-1 (mPCLN-1),spanning 2 introns between exons 3 and 5 (hence, nothing is visible in 570-bpregion in lane with genomic DNA as a template (II) or actin primers (III). Twohundred-bp transcripts are observed in HEK293 and opossum kidney (OK)cells, and 570-bp transcripts are visible in HEK293 and MDCT cells. B:Western blot analysis of proteins from renal cell lines using anti-paracellin-1antibody. Probing the blot with 2 �g/ml anti-paracellin-1 IgG displayed a bandcorresponding to a 33-kDa molecular weight marker in HEK293 and OK, but notin NIH3T3, cells.

Fig. 3. Deletion analysis of the hPCLN-1promoter region in renal cell lines. Cellswere transfected with 0.3 �g (0.75 �g forNIH3T3) pGL3-basic or 0.3 �g (0.75 �g forNIH3T3) pGL3-basic containing hPCLN-1promoter fragments of decreasing size fromthe 5�-end (see Fig. 1). Luciferase activitywas measured after 48 h. To control fortransfection efficiency, cells were cotrans-fected with 0.3 �g (0.75 �g for NIH3T3)pCH110, and luciferase activity was normal-ized to �-galactosidase activity. Activity ofreporter plasmids is expressed as %pGL3-basic vector activity. Values are means SEof 3–5 independent experiments, each per-formed in quadruplicate. OK, opossum kid-ney; MDCT, mouse distal convoluted duct.



by 10.220.33.6 on April 30, 2017


ownloaded from


pressing cell lines HEK293, OK, and MDCT. Reporter geneactivities were compared with those in transfected NIH3T3cells, which do not express PCLN-1. As shown in Fig. 3, allthree PCLN-1-expressing renal cell lines displayed a similarpattern of luciferase activity when transfected with pJ12, pJ2L,pJ2M, and pJ2.7. Luciferase activity was maximally inducedby pJ2M (2.5 kb) and gradually diminished, as the insert sizewas either increased or decreased. Specifically, the lowestluciferase activity resulted from transfection of the 7.5-kbinsert-containing plasmid (pJ12). Truncation of the 5�-flankingregion from �7,514 to �4,687 bp (pJ2L) did not significantlyincrease luciferase activity. A further decrease in size from�4,687 to �2554 bp (pJ2M) caused a significant increase inluciferase activity in all three renal cell lines. When the 7.5-kbfragment in the luciferase vector was further shortened to 0.733kb (pJ2.7), luciferase activity in all four renal cell lines wasmarkedly reduced. Transfection of each of the four constructsinto the control cell line NIH3T3 showed negligible inductionof luciferase activity.

Taken together, these results indicated that the 2.5-kb hP-CLN-1 promoter fragment cloned in pJ2M contained an activepromoter. The lack of stimulation in NIH3T3 cells suggestedthat the activity of the hPCLN-1 promoter was kidney cellspecific. The results also suggested that kidney-specific expres-sion of the hPCLN-1 promoter was due, at least in part, totissue-specific transcriptional regulation.

Although the induction pattern was similar in all threePCLN-1-expressing renal cell lines, its magnitude was verydifferent between cells. The highest levels of induction ap-peared in OK cells and the lowest in MDCT cells. HEK293cells displayed intermediate levels of induction. Based on these

findings, OK cells were selected as experimental cells in ournext set of experiments.

Deletion analysis of the hPCLN-1 promoter. To furtherexplore the 5�-flanking region of hPCLN-1, a second set ofdeletion constructs was tested (Fig. 1), two between pJ2L andpJ2M (designated pJ2/3.9 and pJ2/3.3) and two between pJ2Mand pJ2.7 (designated pJ2/1.9 and pJ/1.4). All eight constructswere transiently transfected into OK and NIH3T3 cells. Asevident from Fig. 4, the highest level of luciferase activityresulted from transfection of the 2.5-kb insert-containing plas-mid (pJ2M). Increasing the size of this fragment by 763 bp(pJ2/3.3) caused a 50% reduction in luciferase activity, whichremained unchanged when the 2.5-kb fragment was lengthenedby 1,432 bp (pJ2/3.9). When the 2.5-kb fragment was length-ened by 2,133 (pJ2L) or 4,960 bp (pJ12), luciferase activitygradually diminished by 60 and 90%, respectively, comparedwith pJ2M. Decreasing the size of the 2.5-kb fragment by 572(pJ2/1.9) and 1,096 bp (pJ2/1.4) brought about a 30 and 77%reduction in luciferase activity, respectively. A further trunca-tion of the insert from 1,458 to 733 bp left luciferase activity at23% that of pJ2M. No significant differences in luciferaseactivity between the various fragments tested were demon-strated in control NIH3T3 cells (data not shown).

To further explore the 2.5-kb 5�-flanking region ofhPCLN-1, which displayed the highest promoter activity, anorientation study was performed in OK cells. As shown in Fig.5, when the 2.5-kb promoter fragment in pGL3 vector wasreversed (pJ235M), luciferase activity was 80% lower thanactivity induced by the same sequence in sense orientation(pJ2M).

Fig. 4. Deletion analysis of the hPCLN-1promoter region in OK cells. Cells weretransfected with 0.3 �g pGL3-basic or 0.3�g pGL3-basic containing hPCLN-1 pro-moter fragments of decreasing size from the5�-end (see Fig. 1). Luciferase activity wasmeasured after 48 h. To control for transfec-tion efficiency, cells were cotransfected with0.3 �g pCH110, and luciferase activity wasnormalized to �-galactosidase activity. Ac-tivity of reporter plasmids is expressed as%pGL3-basic vector activity. Values aremeans SE of 3–5 independent experi-ments, each performed in quadruplicate.

Fig. 5. Orientation analysis of the hPCLN-1 promoter re-gion in OK cells. Cells were transfected with pGL3-basic or0.3 �g pGL3-basic containing the 2.5-kb 5�-flanking regionof hPCLN-1 cloned in sense (pJ2M) and antisense(pJ235M) orientations. Luciferase activity was measuredafter 48 h. To control for transfection efficiency, cells werecotransfected with 0.3 �g pCH110, and luciferase activitywas normalized to �-galactosidase activity. Activity ofreporter plasmids is expressed as %pGL3-basic vector ac-tivity. Values are means SE of 3–5 independent experi-ments, each performed in quadruplicate.



by 10.220.33.6 on April 30, 2017


ownloaded from


Taken together, these results suggested that a positive reg-ulatory element, most likely located between positions �1982and �1458, and a negative regulatory element likely posi-tioned between positions �3317 and �2554, are involved inhPCLN-1 promoter activity in renal cells.

Binding of nuclear proteins to the hPCLN-1 promoter. Toinvestigate the location and nature of the presumed positiveregulatory element between positions �1982 and �1458,binding of nuclear proteins to this region was examined. The525-bp DNA fragment was divided into 15 sequential double-stranded oligonucleotides, which were 5� end-labeled with 32Pand incubated with nuclear proteins. DNA-protein complexeswere resolved from unbound DNA by nondenaturing gel elec-trophoresis. Binding patterns were compared between nuclearextracts from OK cells and those from control NIH3T3 cells.As shown in Fig. 6, incubation of seven sequential DNAfragments located in the proximal part of the 525-bp sequence(between positions �1770 and �1491) with nuclear extractsfrom OK cells produced several retarded bands (� in Fig. 6)that were absent when the DNA was incubated with nuclearextracts from NIH3T3 cells. Binding was specific, since addi-tion of a 50-fold molar excess of unlabeled oligonucleotideabolished DNA-protein complexes. DNA binding was absentin lanes without the nuclear extract. When oligonucleotidesoriginating from the distal part of the 500-bp region (fromposition �1982 to �1771) were incubated with nuclear pro-teins from OK and NIH3T3 cells, the protein-DNA bindingpattern was similar in the two cell lines (data not shown).

Taken together, these results suggested that OK cells, therenal cell line OK, but not nonrenal NIH3T3 cells, containnuclear proteins that bind specifically to the hPCLN-1 pro-moter in the 280-bp region between positions �1770 and�1491.

Deletion analysis of the DNA region containing the pro-posed positive regulatory element of hPCLN-1. To investigatethe functional importance of the 280-bp hPCLN-1 promoterregion implicated in transcriptional activation and nuclearprotein binding, a deletion of the region between positions�1772 and �1186 within the pJ2M reporter vector was cre-ated. When the resulting plasmid, designated pJ2Mdel, was

transfected into OK cells, luciferase activity was reduced by60% relative to pJ2M (Fig. 7A).

To further define the positive regulatory element within the280-bp promoter sequence of functional importance, deletionanalysis of this segment was carried out. The hPCLN-1 pro-moter sequence from positon �1738 to �1493 was 5�-trun-cated in increments of 35 bp. These progressively shorter DNAfragments were cloned upstream to a luciferase reporter gene inSV40 minimal promoter-containing vectors, pGL3-promoter,and transfected into OK cells. Figure 7B illustrates that lucif-erase activity was highest with the construct containing thehPCLN-1 promoter sequence from position �1703 to �1493.Truncation of the promoter to position �1668 caused littlechange in reporter activity. However, when the promoter wastruncated to position �1633, reporter activity diminished by50%. Two more sequential 35-bp deletions caused only minorreductions in luciferase activity.

Taken together, these experiments suggest that the 70-bpDNA sequence between positions �1703 and �1633, whichbinds nuclear proteins from renal cells, contains a positiveregulatory element involved in transcriptional activation. Com-puter analysis of this segment, revealed a single PPRE atposition �1655, which is known to bind the transcriptionfactor PPAR (11, 15, 20).

Most PPREs described so far consist of a direct repeat of twohexamer half-sites separated by several nucleotides (14). Thefirst half-site is highly conserved (AGGTCA) whereas thesecond is not. The sequence of the PPRE identified in thehPCLN-1 promoter contains one half-site, which is a perfectconsensus sequence, but no clearly recognizable second half-site. This sequence, however, has been shown to enable tran-scription factor binding and activation (3).

Several in vivo (37, 44) and in vitro (7, 30) studies havedemonstrated that Mg2� restriction increases Mg2� transportin the renal tubule whereas Mg2� load decreases it (22, 31). Toexplore the molecular mechanisms of this modulation, weexamined the effect of changes in ambient Mg2� concentrationon the activity of the hPCLN-1 promoter. For this purpose, OKcells were transiently transfected with the 2.5-kb hPCLN-1promoter fragment (pJ2M) driving maximal luciferase activity.

Fig. 6. Binding of nuclear proteins to thehPCLN-1 promoter. EMSAs were performedusing nuclear extracts (50 �g) from OK andNIH3T3 cells. Nuclear extracts were incu-bated with 32P-labeled oligonucleotides(oligo) containing 7 sequential 35-bp nucle-otides (fragments 7–14) stretching from po-sitions �1770 to �1491. Assays were per-formed in the absence (�) or presence (�)of 50-fold molar excess of unlabeled oligo-nucleotide. �, Specific retarded bands ob-served by the addition of nuclear extractsfrom OK cells. Data for fragments 1–6 and8 are not shown.



by 10.220.33.6 on April 30, 2017


ownloaded from


Following transfection, the cells were exposed for 24 h tomedia containing 0 (low), 0.7 (normal), or 1.5 mM (high)Mg2�. As demonstrated in Fig. 8, Mg2� restriction caused a52% increase, and Mg2� load a 72% decrease, in luciferase

activity in OK cells compared with conditions of normal Mg2�

concentration. These findings suggest that ambient Mg2� con-centration may modulate Mg2� transport at the transcriptionallevel.

Effect of 1,25(OH)2 vitamin D on the PPRE-containing DNAregion in the hPCLN-1 promoter. Since the vitamin D receptor(VDR) and PPAR both belong to the nuclear receptor super-family of transcription factors and thus share several biologicalcharacteristics (6, 15, 26), we examined the effect of1,25(OH)2 vitamin D on the hPCLN-1 promoter sequencecontaining the PPRE half-site. We first examined the action of1,25(OH)2 vitamin D on the 2.5-kb hPCLN-1 promoter frag-ment (pJ2M). For this purpose, OK cells, transiently trans-fected with this fragment were exposed to 5 � 10�7 M1,25(OH)2 vitamin D for 16–20 h, which caused promoteractivity to decrease by 56% (Fig. 9A). Experimental conditionswere selected based on preliminary experiments with varying1,25(OH)2 vitamin D concentrations (10�6 to 10�7M) (datanot shown).

We next examined whether the effect of 1,25(OH)2 vitaminD on the promoter was PPRE dependent. For this purpose, OKcells transfected with pGL3-promoter vector containing a210-bp hPCLN-1 promoter fragment (pJ210) corresponding to

Fig. 7. Deletion analysis of the DNA region containing the proposed positive regulatory element of hPCLN-1 in OK cells. A: cells were transfected with 0.3�g pGL3-basic or 0.3 �g pGL3-basic containing the 2.5-kb promoter fragment (pJ2M) or the same fragment with a deletion of the region between positions�1772 and �1186 (pJ2Mdel). Luciferase activity was measured after 48 h. To control for transfection efficiency, cells were cotransfected with 0.3 �g pCH110,and luciferase activity was normalized to �-galactosidase activity. Activity of reporter plasmids containing hPCLN-1 promoter fragments is expressed as%pGL3-basic vector activity. Values are means SE of 3–5 independent experiments, each performed in quadruplicate. B: deletion analysis of the proposedpositive regulatory element-containing region of hPCLN-1 in OK cells. Cells were transfected with 0.3 �g pGL3 SV40 minimal promoter vector or 0.3 �gpGL3-SV40 containing 6 hPCLN-1 promoter fragments decreasing in size from the 5�-end, between positions �1738 and �1493. Luciferase activity wasmeasured after 48 h. To control for transfection efficiency, cells were cotransfected with 0.3 �g pCH110, and luciferase activity was normalized to�-galactosidase activity. Activity of reporter plasmids containing hPCLN-1 promoter fragments is expressed as %activity of SV40 minimal promoter-containingreporter vector (pGL3-SV40). Values are means SE of 3–5 independent experiments, each performed in quadruplicate.

Fig. 8. Effect of ambient Mg2� concentration on the hPCLN-1 promoter inOK cells. Cells were transfected with 0.3 �g pGL3-basic or 0.3 �g pGL3-basiccontaining the 2.5-kb 5�-flanking region of hPCLN-1 (pJ2M). Cells wereexposed to 0 (low), 0.7 (normal), or 1.5 mM (high) Mg2� for 24 h. Subse-quently, luciferase activity was measured. To control for transfection effi-ciency, cells were cotransfected with 0.3 �g pCH110, and luciferase activitywas normalized to �-galactosidase activity. Activity of reporter plasmids isexpressed as %pGL3-basic vector activity. Values are means SE of 3–5independent experiments, each performed in quadruplicate.



by 10.220.33.6 on April 30, 2017


ownloaded from


the hPCLN-1 promoter sequence between positions �1703 and�1493 harboring the PPRE site, or a 70-bp 5� deletion product(pJ140) of this same promoter sequence, between positions�1633 and �1493, lacking the PPRE, were exposed to1,25(OH)2 vitamin D under similar conditions. 1,25(OH)2

vitamin D treatment reduced the activity of the PPRE-contain-ing promoter fragment by 52% compared with control non-1,25(OH)2 vitamin D-treated cells (Fig. 9B). This effect of1,25(OH)2 vitamin D was markedly diminished in cells trans-fected with the PPRE-lacking fragment. Moreover, the actualdifference in promoter activity between the PPRE-containingand PPRE-lacking fragments decreased fivefold following1,25(OH)2 vitamin D treatment (Fig. 9B).

To further establish the role of PPRE in hPCLN-1 promoteractivity, three bases of the PPRE half-site were point-mutatedin both pJ2M and pJ210. OK cells transfected with theseconstructs were exposed to 1,25(OH)2 vitamin D as above. Asshown in Fig. 10A, the decrease in pJ2M activity following1,25(OH)2 vitamin D treatment was greatly diminished in cellstransfected with the mutated promoter. When a similar exper-iment was carried out with mutated pJ210 (Fig. 10B), the1,25(OH)2 vitamin D-induced reduction in luciferase activityseen in cells transfected with wild-type, PPRE-containingpJ210 was completely abolished once the PPRE site wasmutated.

These experiments suggest that 1,25(OH)2 vitamin D mod-ulates hPCLN-1 activity by a mechanism that appears toinvolve the PPRE half-site in the gene promoter.

DISCUSSION

In this study, we describe the transcriptional analysis of thepromoter of the human paracellin-1 gene. We demonstrate thata 70-bp region between positions �1633 and �1703 may playa key role in the activity of the hPCLN-1 promoter (Figs. 6 and7). Furthermore, we provide evidence that an interplay betweena positive regulatory element located within this 70-bp region,and a more distally located negative regulatory element on the5�-flanking region of the paracellin-1 gene, may determinerenal cell-specific expression of this gene (Fig. 4). In addition,we demonstrate that ambient Mg2� concentration affectshPCLN-1 promoter activity (Fig. 8). Finally, we show that1,25(OH)2 vitamin D modulates hPCLN-1 promoter activityvia this 70-bp, PPRE-containing, hPCLN-1 promoter fragment(Figs. 9 and 10), thereby suggesting that paracellin-1 geneexpression is subject to PPAR/PPRE-mediated transcriptionalregulation by this hormone.

Reabsorption of solutes across the tubular epithelial layerdepends on two separate routes: a transcellular pathway and aparacellular pathway (1, 47). The paracellular pathway isregulated by tight junctions, which form a barrier to thediffusion of solutes across epithelial cells and function as aboundary between the apical and basolateral membranes main-taining epithelial cell polarity (21). Paracellin-1, which appearsto regulate the paracellular transport of Mg2� in the TAL, isthe first tight junction protein reported to be involved in ionresorption (38). Paracellin-1, or claudin 16, (38) belongs to the

Fig. 9. Effect of 1,25(OH)2 vitamin D (VitD) on the hPCLN-1 promoter in OK cells. A: effect of 1,25(OH)2 vitamin D on the 5�-flanking 2.5-kb promoterfragment. Cells were transfected with 0.3 �g pGL3-basic or 0.3 �g pGL3-basic containing the 2.5-kb 5�-flanking region of hPCLN-1 (pJ2M). Cells were exposedto 5 � 10�7 M 1,25(OH)2 vitamin D (prepared from a stock solution of 10�4 M in ethanol) or to 0.5% ethanol (control) for 16–20 h. Subsequently, luciferaseactivity was measured. Activity of reporter plasmids is expressed as %pGL3-basic vector activity. B: effect of 1,25(OH)2 vitamin D on the 210-bp peroxisomeproliferator response element (PPRE)-containing DNA region. Cells were transfected with 0.3 �g pGL3-SV40 minimal promoter vector or with pGL3-SV40containing the 210-bp hPCLN-1 promoter fragment corresponding to the promoter sequence between positions �1703 and �1493, harboring the PPRE site, orpGL3-SV40 containing a 140-bp hPCLN-1 promoter fragment, corresponding to the region between positions �1633 and �1493, lacking the PPRE site. Cellswere exposed to 1,25(OH)2 vitamin D or ethanol as in A, and luciferase activity was measured. Activity of reporter plasmids is expressed as %pGL3-SV40activity. In both A and B, to control for transfection efficiency, cells were cotransfected with 0.3 �g pCH110, and luciferase activity was normalized to�-galactosidase activity. Values are means SE of 3–5 independent experiments, each performed in quadruplicate.



by 10.220.33.6 on April 30, 2017


ownloaded from


claudin family of proteins that participate in the formation oftight junction strands in various tissues (27) and are thought tohave a major role in regulating the magnitude and nature ofparacellular permeability (1). The disease mutations found inparacellin-1 in familial hypomagnesemia-hypercalciuria syn-drome result in an increased resistance to the electrical seal oftight junctions, thereby decreasing epithelial ionic permeabilityand selectively impeding magnesium and calcium reabsorption(38, 47). Recently, it has been reported that deletion of theparacellin-1 gene is responsible for renal tubular dysplasia incattle (29). This finding suggests that paracellin-1, in additionto its function in ion resorption, may play an important role inthe normal development and organization of the renal tubule.

hPCLN-1 mRNA was found to be expressed exclusively inthe kidney (38, 42). Although cis-acting regulatory elementsinvolved in regulating gene transcription may be dispersedthroughout the gene locus, often the proximal promoter regioncontains elements sufficient for high levels of tissue-specificgene transcription. These elements may function as bindingsites for tissue-restricted proteins that arbitrate transcriptionalactivation in expressing cells. In an effort to identify cis-actingelements involved in regulating kidney-specific expression ofhPCLN-1, we focused our study on the 5�-flanking region ofthis gene. We isolated a 7.5-kb genomic sequence correspond-ing to the 5�-flanking region of hPCLN-1 and, using reportergene assays, showed that the proximal 2.5-kb region containedcis-acting, positive and negative, regulatory elements that playa role in renal epithelial-specific expression of hPCLN-1. This

promoter displayed high activity in PCLN-1-expressing renalcell lines but not in the fibroblast cell line NIH3T3. In therabbit (38) and rat (42) kidney, PCLN-1 expression has beenshown to be restricted to the TAL and the DCT. Several renalcell lines were used in this study. These included HEK293 andMDCT cells, as well as OK cells of proximal tubular origin.These renal cells were shown to express paracellin-1 mRNA(Fig. 2A) and protein (Fig. 2B) and to display hPCLN-1promoter-driven luciferase activity (Fig. 3) as opposed tomouse embryonic fibroblast cells (NIH3T3), which demon-strated no expression/activity, indicating that the PCLN-1 pro-moter was kidney specific.

In our study, OK cells displayed the highest level ofhPCLN-1 promoter-driven reporter gene activity (Fig. 3).Hence, these cells served as the recipient cells in most of ourtransfection experiments. It is not entirely clear why OK cellsof proximal tubular origin express paracellin-1, which wasfound only in the TAL and the DCT of the rabbit and the rat.Several considerations could provide an explanation for thisfinding. First, the nephron segment-specific expression of theparacellin-1 gene has been examined in the kidney of these tworodents only. It is possible that the expression of the paracel-lin-1 gene in other animals and species, including humans,extends to the tight junction of more proximal nephron seg-ments. Second, although several properties of OK cells areconsistent with a proximal tubular site of origin, this cell linewas originally derived from the whole kidney and may possesscharacteristics of more distal regions of the nephron (10, 19).

Fig. 10. Effect of 1,25(OH)2 vitamin D on the hPCLN-1 promoter in OK cells. A: effect of 1,25(OH)2 vitamin D on the 2.5-kb 5�-flanking region. Cells weretransfected with 0.3 �g pGL3-basic or 0.3 �g pGL3-basic containing the 2.5-kb 5�-flanking region of hPCLN-1 (pJ2M-wild-type) or 0.3 �g pJ2M harboring apoint-mutated PPRE site (pJ2M-mutated). Cells were exposed to 5 � 10�7 M 1,25(OH)2 vitamin D (prepared from a stock solution of 10�4 M in ethanol) orto 0.5% ethanol (control) for 16–20 h. Subsequently, luciferase activity was measured. Reporter gene activity is expressed as %pGL3-basic activity. B: effectof 1,25(OH)2 vitamin D on the 210-bp PPRE-containing DNA region. Cells were transfected with 0.3 �g pGL3-SV40 minimal promoter vector or 0.3 �gpGL3-SV40 containing the 210-bp hPCLN-1 promoter fragment with a wild-type PPRE site (pJ210-wild-type) or 0.3 �g pJ210 harboring a point-mutated PPREsite (pJ210-mutated). Cells were exposed to 1,25(OH)2 vitamin D or ethanol as in A, and luciferase activity was measured. Activity of reporter plasmids isexpressed as %pGL3-SV40 vector activity. In both A and B, to control for transfection efficiency, cells were cotransfected with 0.3 �g pCH110, and luciferaseactivity was normalized to �-galactosidase activity. Values are means SE of 3–5 independent experiments, each performed in quadruplicate.



by 10.220.33.6 on April 30, 2017


ownloaded from


Finally, in the immature rat, Mg2� reabsorption occurs pre-dominantly via the paracellular pathway in the proximal tubulerather than in the loop of Henle (23, 30). It is possible that theOK cell line used in our study contains cells with characteris-tics of the immature proximal tubule, including expression andcell-specific activity of genes, such as PCLN-1, participating inMg2� reabsorption in the immature proximal tubule tightjunction. Taken together, we believe that the expression ofparacellin-1 mRNA and protein in OK cells, coupled with themarked increase in reporter gene activity driven by thehPCLN-1 promoter in this cell line, has made OK cells suitableas experimental cells in our study.

Deletion analysis of reporter gene constructs containinghPCLN-1 promoter DNA suggested that hPCLN-1 transcrip-tion may be regulated by a proximal positive regulatory ele-ment positioned between �1982 and �1458 and a moredistally located negative regulatory element between positions�3317 and �2554 (Fig. 4). EMSA studies of the presumedpositive regulatory element-bearing region indicated that OKcells contain nuclear proteins that bind specifically to thisfunctionally important region of the hPCLN-1 promoter (Fig.6). NIH3T3 nuclear proteins did not bind to this DNA region.These proteins are potentially involved in tissue-specific ex-pression of hPCLN-1. When this protein-binding DNA regionwas deleted from the 2.5-kb PCLN-1 promoter, a 60% reduc-tion in promoter activity occurred. Detailed analysis revealed a70-bp sequence within this DNA region, which seems to beresponsible, at least in part, for modulating kidney-specifichPCLN-1 transcription (Fig. 7). Computer analysis of thissegment disclosed a single PPAR binding site (PPRE).

Tubular Mg2� reclamation is modulated by a variety ofhormonal and nonhormonal factors (8, 16, 30). Mg2� restric-tion and Mg2� load influence both transcellular Mg2� transportin the DCT and paracellular Mg2� flux in the TAL (7, 22, 31,37, 44). Micropuncture experiments in the rat nephron havedemonstrated that a low-Mg2� diet leads to urinary retention ofMg2� due to increased Mg2� reabsorption in the loop of Henle(37, 44). Culturing MDCT cells in Mg2�-free medium in-creased their Mg2� transport rate (7, 30). Pretreatment ofMDCT cells with actinomycin D, an inhibitor of transcription,resulted in a significant decrease in this adaptive response,suggesting that the adaptive regulation of Mg2� depletion mayinvolve gene transcription (30). As opposed to the effect ofMg2� deprivation, Mg2� infusion (22) and acute elevation ofMg2� concentration at the contraluminal membrane of theTAL (31) in rats inhibited Mg2� resorption in this tubulesegment. The molecular mechanisms underlying the adaptiveresponse of tubular Mg2� transport to changes in Mg2� levelshave not been investigated. In this study, we show that ambientMg2� concentration affects hPCLN-1 promoter activity in OKcells. Specifically, Mg2� restriction increases hPCLN-1 pro-moter activity, whereas Mg2� load reduces the activity of thispromoter. These findings are in concert with the modulatoryeffects of Mg2� restriction and load on Mg2� transport ob-served at the tubular and cellular levels (7, 22, 31, 37, 44) anddemonstrate, for the first time, that changes in Mg2� availabil-ity may influence Mg2� transport at the transcriptional level.The exact molecular mechanisms whereby this Mg2� level-induced effect is achieved remain unknown. The possibleinvolvement of the cell membrane-bound Ca2�/Mg2�-sensingreceptor, or the potential role of a hypothetical Mg2� response

element residing on the hPCLN-1 promoter in the Mg2�-induced effect on transcription of this gene, remains to beexplored.

Paracellular Mg2� transport in the TAL is known to beregulated by several hormones. These include parathyroidhormone, arginine vasopressin, aldosterone, insulin, and1,25(OH)2 vitamin D, among others (30). Using microperfu-sion studies of isolated mouse TAL segments, it has beenshown that most of these hormonal responses are mediated bychanging the transepithelial voltage or by altering the perme-ability of the paracellular pathway (24, 30, 45). However, theexact molecular mechanisms of this hormone-induced effect onMg2� transport have not been investigated. 1,25(OH)2 vitaminD has been shown to increase Mg2� uptake by MDCT cells(32). The effect of this hormone on paracellular Mg2� trans-port has not been explored. Several hormones, including1,25(OH)2 vitamin D (6), exert many of the biological actionsby receptor-mediated effects on gene transcription. In thisstudy, we show that 1,25(OH)2 vitamin D decreases hPCLN-1promoter-driven luciferase activity (Figs. 9 and 10). OK cellsare known to harbor vitamin D receptors (18) and, as discussedabove, appear to express the paracellin-1 gene as well aspossess the machinery necessary for its activity. Hence, thiscell line may serve as an excellent model with which to explorethe effect of this hormone on the paracellin-1 gene. Takentogether, our findings provide the first direct evidence thatparacellular, paracellin-1-mediated Mg2� transport may beregulated at the transcriptional level by 1,25(OH)2 vitamin D.However, the physiological significance of the transcription-inhibiting effect of 1,25(OH)2 vitamin D shown in our studyand its role in overall magnesium transport in the renal tubuleremain to be clarified.

The effect of 1,25(OH)2 vitamin D on the hPCLN-1 pro-moter was demonstrated in our study to involve the PPREhalf-site located between positions �1633 and �1703 withinthis promoter (Figs. 9 and 10). The PPARs comprise a group oftranscription factors that belong to the nuclear receptor super-family to which the VDR, the thyroid hormone receptor, andthe all-trans-retinoic acid receptor also belong (11, 15, 20).PPARs are ligand-regulated transcription factors that controlgene expression by binding to specific response elements(PPREs) within promoters, having formed heterodimers withthe retinoid X receptor (11, 15, 20). Three PPAR isoforms,PPAR�, PPAR�, and PPAR, have been identified (11, 15)and are expressed in several tissues including the kidney (11,48). PPARs participate in a variety of biological processescommon to various cell types, including lipid metabolism,glucose homeostasis, inflammation, cell proliferation and dif-ferentiation, apoptosis, and early development (4, 11). How-ever, despite their abundant expression in various segments ofthe renal tubule (11, 48), very little is known about specifictubular transport processes that are controlled by the PPARs.Noteworthy is a recent study demonstrating enhanced renaltubular cell surface expression of the epithelial sodium channelin response to PPAR activation (12).

In our study, we focused on 1,25(OH)2 vitamin D, known tomodulate Mg2� transport on one hand (30, 32) and to interactwith the PPAR system on the other. The DNA binding site ofthe VDR is 46% homologous to the DNA binding site ofPPAR, and both receptors are known to heterodimerize withthe retinoid X receptor before binding to their specific DNA



by 10.220.33.6 on April 30, 2017


ownloaded from


motifs (6, 15, 26). However, recent studies suggest that there isconsiderable flexibility in the binding sites recognized by theVDR, including the existence of some single half-sites (17).Furthermore, the outcome of the interaction between the VDRand its binding site may be an increase (5) or a decrease (25)in gene transcription. Of note is a recent study demonstratingthat the VDR represses transcriptional activity of PPAR� inCOS1 cells (34). In our study, we provide evidence for PPRE-dependent transcriptional regulation of hPCLN-1 by 1,25(OH)2

vitamin D, and we show here, for the first time, that thePPAR/PPRE axis may play a specific role in the hormonalregulation of Mg2� transport in the renal tubule.

In conclusion, an interplay between a positive regulatoryelement and a more distally located negative regulatory ele-ment on the 5�-flanking region of the paracellin-1 gene maydetermine renal epithelial-specific activity of this gene. The70-bp, PPRE-containing, DNA region between positions�1633 and �1703 may play a key role in the activity of thePCLN-1 promoter. Ambient Mg2� concentration and1,25(OH)2 vitamin D may modulate paracellular, paracellin-1-driven, Mg2� transport at the transcriptional level. 1,25(OH)2

vitamin D exerts its activity on the hPCLN-1 promoter likelyvia the PPRE site.

Future studies utilizing transgenic animals may verifywhether the activity of this 70-bp DNA fragment, examined inthe cell culture model, demonstrates tubular epithelial specifityin the intact organism, where the transgene is exposed to amore physiologically relevant environment (i.e., hormones,proteins, etc.). Once established, it is possible that these trans-genic animals will serve as an excellent model with which tostudy the transcriptional effects of a variety of physiological aswell as pathophysiological factors on the paracellin-1 gene,and thus on tubular Mg2� reabsorption.

ACKNOWLEDGMENTS

We thank Dr. Karl Skorecki, Dr. Maty Tzukerman, and Dr. Sara Selig forhelpful suggestions and Ayal Kelmachter and Adva Hermoni for technicalassistance. We thank Judi Fichman and Hagar Shafrir for expert secretarialassistance.

GRANTS

This work was supported by the Ruth and Allen Ziegler Fund for PediatricResearch and the Kronovet Fund for Medical Research, Technion-IsraelInstitute of Technology.

REFERENCES

1. Anderson JM. Molecular structure of tight junctions and their role inepithelial transport. News Physiol Sci 16: 126–130, 2001.

2. Bai Y, Pontoglio M, Hiesberger T, Sinclair AM, and Igarashi P.Regulation of kidney-specific Ksp-cadherin gene promoter by hepatocytenuclear factor-1�. Am J Physiol Renal Physiol 283: F839–F851, 2002.

3. Barrero MJ, Camarrero N, Marrero PF, and Haro D. Control ofhuman carnitine palmitoyltransferase II gene transcription by peroxisomeproliferator-activated receptor through a partially conserved peroxisomeproliferator-responsive element. Biochem J 369: 721–729, 2003.

4. Berger J and Moller DE. The mechanisms of action of PPARs. Annu RevMed 53: 409–435, 2002.

5. Breen EC, Van Wijen AJ, Lian JB, Stein GS, and Stein JL. In vivooccupancy of vitamin D responsive element in the osteocalcin genesupports vitamin D-dependent transcriptional upregulation in intact cells.Proc Natl Acad Sci USA 91: 12902–12906, 1994.

6. Brown AJ, Dusso A, and Slatopolsky E. Vitamin D. Am J Physiol RenalPhysiol 277: F157–F175, 1999.

7. Dai LJ, Raymond L, Friedman PA, and Quamme GA. Mechanisms ofamiloride stimulation of Mg2� uptake in immortalized mouse distal

convoluted tubule cells. Am J Physiol Renal Physiol 272: F249–F256,1997.

8. Dai LJ, Ritchie G, Kerstan D, Kang HS, Cole DE, and Quamme GA.Magnesium transport in the renal distal convoluted tubule. Physiol Rev 81:51–84, 2001.

9. Dignam JD, Lebovitz RM, and Roeder RG. Accurate transcriptioninitiation by RNA polymerase II in a soluble extract from isolatedmammalian nuclei. Nucleic Acids Res 11: 1475–1489, 1983.

10. Gomes P, Xu J, Serrao P, Doria S, Jose PA, and Soares-da-Silva P.Expression and function of sodium transporters in two opossum kidneycell clonal sublines. Am J Physiol Renal Physiol 283: F73–F85, 2002.

11. Guan Y and Breyer MD. Peroxisome proliferator-activated receptors(PPARs): novel therapeutic targets in renal disease. Kidney Int 60: 14–30,2001.

12. Hong g Lockhart A, Davis B, Rahmoune H, Baker S, Ye L, ThompsonP, Shou Y, O’Shaughnessy K, Ronco P, and Brown J. PPAR activa-tion enhances cell surface ENac� via up-regulation of SGK1 in humancollecting duct cells. FASEB J 17: 1966–1968, 2003.

13. Igarashi P, Cooduvalli S, Shashikant R, Thomson RB, Whyte DA,Liu-Chen S, Ruddle FH, and Aronson PS. Ksp-cadherin gene promoter.II. Kidney-specific activity in transgenic mice. Am J Physiol Renal Physiol277: F599–F610, 1999.

14. Ijpenberg A, Jeannin E, Wahli W, and Desvergne B. Polarity andspecific sequence requirements of peroxisome proliferator-activated recep-tor (PPAR)/retinoid X receptor heterodimer binding to DNA. A functionalanalysis of the malic enzyme gene PPAR response element. J Biol Chem272: 20108–20117, 1997.

15. Issemann I and Green S. Activation of a member of the steroid hormonereceptor superfamily by peroxisome proliferators. Nature 347: 645–650,1990.

16. Kelepouris E and Agus ZS. Hypomagnesemia: renal magnesium han-dling. Semin Nephrol 18: 58–73, 1998.

17. Koszewski NJ, Kiessling S, and Malluche H. Isolation of genomic DNAsequences that bind vitamin D receptor complexes. Biochem Biophys ResCommun 283: 188–194, 2001.

18. Koszewski NJ and Rowan A. Heterodimer requirement for gene regula-tion by vitamin D in variant OK cells. Steroids 68: 307–314, 2003.

19. Koyama H, Goodpasture C, Miller MM, Teplitz RL, and Riggs AD.Establishment and characterization of a cell line from the Americanopossum (Didelphys virginiana). In Vitro 14: 239–246, 1978.

20. Laudet V, Hanni C, Coll J, Catzeflis F, and Stehelin D. Evolution of thenuclear receptor gene superfamily. EMBO J 11: 1003–1013, 1992.

21. Le Grimellec C, Friedlander G, el Yandouzi EH, Zlatkine P, andGiocondi MC. Membrane fluidity and transport properties in epithelia.Kidney Int 42: 825–836, 1992.

22. Le Grimellec C, Roinel N, and Morel F. Simultaneous Mg, Ca, P, K, Naand Cl analysis in rat tubular fluid. II. During acute Mg plasma loading.Pflugers Arch 340: 197–210, 1973.

23. Lelievre-Pegorier M, Merlet-Benichou C, Roinel N, and de RouffignacC. Developmental pattern of water and electrolyte transport in rat super-ficial nephrons. Am J Physiol Renal Fluid Electrolyte Physiol 245: F15–F21, 1983.

24. Mandon B, Siga E, Chabardes D, Firsov D, Roinel N, and De Rouffig-nac C. Insulin stimulates Na�, Cl�, Ca2�, and Mg2� transports in TAL ofmouse nephron: cross-potentiation with AVP. Am J Physiol Renal FluidElectrolyte Physiol 265: F361–F369, 1993.

25. Meszaros JG, Karin NJ, Akanbi K, and Farach-Carson MC. Down-regulation of L-type Ca2� channel transcript levels by 1,25-dihydroxyvi-tamin-D3. J Biol Chem 271: 32981–32985, 1996.

26. Misra P, Owuor ED, Li W, Yu S, Qi C, Meyer K, Zhu YJ, Rao MS,Kong AN, and Reddy JK. Phosphorylation of transcriptional coactivatorperoxisome proliferator-activated receptor (PPAR)-binding protein (PBP).Stimulation of transcriptional regulation by mitogen-activated proteinkinase. J Biol Chem 277: 48745–48754, 2002.

27. Morita K, Furuse M, Fujimoto K, and Tsukita SH. Claudin multigenefamily encoding four-transmembrane domain protein components of tightjunction strands. Proc Natl Acad Sci USA 96: 511–516, 1999.

28. Nelson RD, Stricklett P, Gustafson C, Stevens A, Ausiello D, Brown D,and Kohan DE. Expression of an AQP2 Cre recombinase transgene inkidney and male reproductive system of transgenic mice. Am J PhysiolCell Physiol 275: C216–C226, 1998.

29. Ohba Y, Kitagawa H, Kitoh K, Sasaki Y, Takami M, Shinkai Y, andKunieda T. A deletion of the paracellin-1 gene is responsible for renaltubular dysplasia in cattle. Genomics 68: 229–236, 2000.



by 10.220.33.6 on April 30, 2017


ownloaded from


30. Quamme GA. Renal magnesium handling: new insights in understandingold problems. Kideny Int 52: 1180–1195, 1997.

31. Quamme GA and Dirks JH. Intraluminal and contraluminal magnesiumon magnesium and calcium transfer in the rat nephron. Am J Physiol RenalFluid Electrolyte Physiol 238: F187–F198, 1980.

32. Ritchie G, Kerstan D, Dai LJ, Kang HS, Canaff L, Hendy GN, andQuamme GA. 1,25(OH)2D3 stimulates Mg2� uptake into MDCT cells:modulation by extracellular Ca2� and Mg2�. Am J Physiol Renal Physiol280: F868–F878, 2001.

33. Romani A and Scarpa A. Regulation of cell magnesium. Arch BiochemBiophys 298: 1–12, 1992.

34. Sakuma T, Miyamoto T, Jiang W, Kakizawa T, Nishio S, Suzuki S,Takeda T, Oiwa A, and Hashizume K. Inhibition of peroxisome prolif-erator-activated receptor � signaling by vitamin D receptor. BiochemBiophys Res Commun 312: 513–519, 2003.

35. Sambrook J, Fritsh EF, and Maniatis T. Molecular Cloning: A Labo-ratory Manual (2nd ed.). Cold Spring Harbor, NY: Cold Spring HarborLaboratory, 1989.

36. Shachaf C, Skorecki KL, and Tzukerman M. Role of AP2 consensussites in regulation of rat Npt2 (sodium-phosphate cotransporter) promoter.Am J Physiol Renal Physiol 278: F406–F416, 2000.

37. Shafik IM and Quamme GA. Early adaptation of renal magnesiumreabsorption in response to magnesium restriction. Am J Physiol RenalFluid Electrolyte Physiol 257: F974–F977, 1989.

38. Simon DB, Lu Y, Chaote KA, Velazquez H, Al-Sabban E, Praga M,Casari G, Bettinelli A, Colussi G, Rodriguez-Soriano J, McCredie D,Milford D, Sanjad S, and Lifton RP. Paracellin-1, a renal tight junctionprotein required for paracellular Mg2� resorption. Science 285: 103–106,1999.

39. Taniyama Y, Sato K, Sugawara A, Urono A, Ikeda Y, Kudo M, Ito S,and Takeuchi K. Renal tubule-specific transcription and chromosomallocalization of rat thiazide-sensitive Na-Cl cotransporter gene. J BiolChem 276: 26260–26268, 2001.

40. Uchida S, Rai T, Yatsushige H, Matsumura Y, Kawasaki M, Sasaki S,and Marumo F. Isolation and characterization of kidney cell-specificClC-K1 chloride channel gene promoter. Am J Physiol Renal Physiol 274:F602–F610, 1998.

41. Uchida S, Tanaka Y, Ito H, Saitoh-Ohara F, Inazawa J, YokoyamaKK, Sasaki S, Marumo F. Transcriptional regulation of the CLC-K1promoter by myc-associated zinc finger protein and kidney-enrichedKruppel-like factor, a novel zinc finger repressor. Mol Cell Biol 20:7319–7331, 2000.

42. Weber S, Schlingmann KP, Peters M, Nejsum LN, Nielsen S, Engel H,Grzeschik KH, Seyberth HW, Grone HJ, Nusing R, and Konrad M.Primary gene structure and expression studies of rodent paracellin-1. J AmSoc Nephrol 12: 2664–2667, 2001.

43. Whyte DA, Li C, Thomson RB, Nix SL, Zanjani R, Karp SL, AronsonPS, and Igarashi P. Ksp-cadherin gene promoter. I. Characterization andrenal epithelial cell-specific activity. Am J Physiol Renal Physiol 277:F587–F598, 1999.

44. Wittner M, Jounier S, Deschenes G, de Rouffignac C, and Di StefanoA. Cellular adaptation of the mouse cortical thick ascending limb ofHenle’s loop (CTAL) to dietary magnesium restriction: enhanced trans-epithelial Mg2� and Ca2� transport. Pflugers Arch 439: 765–771, 2000.

45. Wittner M, Mandon B, Roinel N, de Rouffignac C, and Di Stefano A.Hormonal stimulation of Ca2� and Mg2� transport in the cortical thickascending limb of Henle’s loop of the mouse: evidence for a change in theparacellular pathway permeability. Pflugers Arch 423: 387–396, 1993.

46. Wong MA, Cui S, and Quaggin SE. Identification and characterizationof a glomerular-specific promoter from the human nephrin gene. Am JPhysiol Renal Physiol 279: F1027–F1032, 2000.

47. Wong V and Goodenough DA. Paracellular channels (Abstract). Science285: 52, 1999.

48. Yang T, Michele DE, Park J, Smart AM, Lin Z, Brosius 3rd FC,Schnermann JB, and Briggs JP. Expression of peroxisomal proliferator-activated receptors and retinoid X receptors in the kidney. Am J PhysiolRenal Physiol 277: F966–F973, 1999.



by 10.220.33.6 on April 30, 2017


ownloaded from


the human paracellin-1 gene (hpcln-1): renal epithelial ... · the 7.5-kb 5-ﬂank-ing sequence...

Documents