Structural organization of the human NBC1 gene: kNBC1 is transcribed from an alternative promoter in intron 3

Gene 251 (2000) 109–122 www.elsevier.com/locate/gene

Structural organization of the human NBC1 gene: kNBC1 is transcribed from an alternative promoter in intron 3 Natalia Abuladze 1, Mark Song 1, Alexander Pushkin, Debra Newman, Ivan Lee, Susan Nicholas, Ira Kurtz * Division of Nephrology, UCLA School of Medicine, 10833 Le Conte Avenue, Rm 7-155 Factor Building, Los Angeles, CA 90095-1689, USA Received 12 November 1999; received in revised form 13 April 2000; accepted 28 April 2000 Received by K. Gardiner

Abstract Several electrogenic sodium bicarbonate cotransporters have been cloned from different human organs. In the renal proximal tubule, the electrogenic sodium bicarbonate cotransporter kNBC1 (1035 aa) mediates the majority of basolateral sodium bicarbonate absorption. In pancreatic ducts, the electrogenic sodium bicarbonate cotransporter pNBC1 (1079 aa) mediates basolateral sodium bicarbonate influx. hNBC1 (hhNBC ), cloned from human heart, is identical to pNBC1 at the amino acid level. We have demonstrated that kNBC1 and pNBC1 are highly homologous proteins that have different N-termini. In kNBC1, 41 amino acids replace the initial 85 amino acids of pNBC1. Whether these proteins are coded by one or more genes is unknown. In order to determine the genetic basis for these transcripts, we first characterized the genomic organization of the NBC1 gene (SLC4A4). NBC1 spans approximately 450 kilobases containing 26 exons that are flanked by typical splice donor and acceptor sequences at the intron–exon boundaries. Exon 1 is specific for the pNBC1 transcript. The first alternative exon of the hNBC1 transcript, containing the 5∞-untranslated region, is derived from the last 43 nucleotides of intron 1 in the NBC1 gene coupled to exon 2. kNBC1 is transcribed from an alternative promoter in intron 3. In the first alternative exon of kNBC1, the last 313 nucleotides of intron 3 are coupled to exon 4, which is common to pNBC1 and hNBC1. The major transcription initiation site in kNBC1 is located 192 nucleotides upstream from the translation initiation codon. A minor start site is located 182 nucleotides upstream from the translation initiation codon. Structural analysis of the proximal kNBC1 promoter revealed an atypical TATA sequence (−33) and several potentially important transcription factor binding sites. Functional studies showed that the 5∞-flanking region of the alternative kNBC1 promoter (−159 to+43) is sufficient for promoter activity. This work is the first demonstration that the three N-terminal transcripts of the human electrogenic sodium bicarbonate cotransporter NBC1 are encoded by the SLC4A4 gene. Furthermore, knowledge of the genomic organization and alternative promoter usage in the NBC1 gene provides a molecular basis for understanding disorders involving electrogenic sodium bicarbonate cotransporters and facilitates the elucidation of transcriptional control of NBC1 expression. © 2000 Elsevier Science B.V. All rights reserved. Keywords: Bicarbonate; Kidney; Pancreas; Sodium; Transport

1. Introduction The Na(HCO ) cotransporters (NBCs) are a family 3n of integral membrane proteins found in most mammaAbbreviations: hhNBC, human heart sodium bicarbonate cotransporter; hNBC1, heart sodium bicarbonate cotransporter 1; kNBC1, kidney sodium bicarbonate cotransporter 1; NBC1, sodium bicarbonate cotransporter 1; NBC3, sodium bicarbonate cotransporter 3; pNBC1, pancreas sodium bicarbonate cotransporter 1; UTR, untranslated region. * Corresponding author. Tel.: +1-310-206-6741; fax: +1-310-825-6309. E-mail address: [email protected] (I. Kurtz) 1 These authors contributed equally to this work.

lian cells that mediate electroneutral and electrogenic sodium bicarbonate cotransport (Burnham et al., 1997; Romero et al., 1997; Abuladze et al., 1998a; Ishibashi et al., 1998; Amlal et al., 1999; Choi et al., 1999; Pushkin et al., 1999a,b; Kwon et al., 2000). Members of the NBC protein family mediate the transepithelial transport of sodium and bicarbonate in several tissues, and contribute to intracellular pH (pHi) regulation. The electrogenic sodium bicarbonate cotransporter kNBC1 is expressed in the renal proximal tubule, where it mediates the majority of basolateral bicarbonate absorption (Burnham et al., 1997; Romero et al., 1997; Abuladze et al., 1998b; Schmitt et al., 1999). The kNBC1 transcript

0378-1119/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S0 3 7 8 -1 1 1 9 ( 0 0 ) 0 0 20 4 - 3

110

N. Abuladze et al. / Gene 251 (2000) 109–122

(~7.6 kilobases) has also been detected in corneal endothelial cells ( Usui et al., 1999) and duodenum (GenBank EST clone AF141934). A homologous protein, pNBC1, is highly expressed in pancreas, where it mediates electrogenic sodium bicarbonate cotransport in pancreatic ducts and acinar cells (Muallem and Loessberg, 1990; Ishiguro et al., 1996; Abuladze et al., 1998a). pNBC1 is expressed in several tissues, including brain, spinal cord, kidney, colon, thyroid, and prostate (Abuladze et al., 1998a). Although the C-terminal 994 amino acids of pNBC1 and kNBC1 are identical, pNBC1 has a unique N-terminus of 85 amino acids, which replaces the first 41 amino acids in kNBC1 (Abuladze et al., 1998a). More recently, an electrogenic sodium bicarbonate cotransporter transcript hNBC1 (hhNBC ) has been cloned from human heart (Choi et al., 1999). hNBC1 is identical to pNBC1 at the amino acid level, and differs only in its 5∞-untranslated region. The genetic basis for these differences is currently unknown. Differences in the N-terminus protein sequence of kNBC1 and pNBC1 (hNBC1) may contribute to tissuespecific regulation, targeting, and functional properties. The N-terminus of pNBC1 has unique consensus phosphorylation sites for protein kinase A, protein kinase C, and casein kinase II, lacking in kNBC1 (Abuladze et al., 1998a). Furthermore, the use of alternative promoters could contribute to tissue-specific transcriptional regulation. Although the amino acid sequences of pNBC1 and hNBC1 are identical, the 5∞-UTR of these transcripts differs. This is of interest given that the 5∞-UTR can function as a tissue-specific translational enhancer and in regulating the transcriptional activity of several genes. In order to determine the molecular basis for the N-terminal differences among the three highly homologous human electrogenic sodium bicarbonate cotransporter transcripts, we have characterized the genomic organization of the NBC1 gene and determined the genetic basis of kNBC1, pNBC1 and hNBC1. The results demonstrate that all three transcripts are encoded by the NBC1 (SLC4A4) gene. Furthermore, functional studies showed that kNBC1 is transcribed from an alternative promoter in intron 3.

random primed and labeled 95 bp synthetic oligonucleotide common to both kNBC1 (271–365), pNBC1 (371– 465) and hNBC1 (298–392). 2.2. Screening of the human BAC clone genomic library and characterization of BAC clones A BAC human genomic library (Clontech, Palo Alto, CA) was screened using a PCR approach. Eight BAC clones spanning the entire coding region of kNBC1, pNBC1 and hNBC1 were isolated and mapped by direct sequencing, restriction endonuclease cleavage, and Southern blot analysis (GenBank Accession Nos: AF089725, AF231466S1 and AF231466S2). Various [32P]ATP random primed synthetic oligonucleotide probes corresponding to various regions of the kNBC1, pNBC1 and hNBC1 transcripts were used. Nucleotide sequences were determined bidirectionally by automated sequencing (ABI 310 Perkin Elmer, Foster City, CA) using Taq Polymerase (Ampli-Taq FS, Perkin-Elmer, Foster City, CA). Sequence assembly and analysis was carried out using GeneWorks software (Oxford Molecular Group, Oxford, UK ). 2.3. Cloning and characterization of the kNBC1 promoter Genomic DNA from BAC clone 15800 was purified and digested with BamHI, HindIII, and EcoRI. Southern analysis was performed using a 94 bp [32P]ATP random primed and labeled synthetic specific kNBC1 oligonucleotide probe (175–268), and a [32P]ATP random primed and labeled 95 bp synthetic specific pNBC1 oligonucleotide probe (118–212). The pNBC1 probe was common to the hNBC1 sequence (45–139). BAC clone 15800 contained the translation initiation site of pNBC1, hNBC1, and kNBC1. An ~8 kilobase EcoRI fragment from BAC clone 15800, which contained the promoter and the first alternative exon of kNBC1, was subcloned into PCR-ScriptSK(+) (Stratagene, La Jolla, CA). The kNBC1 promoter sequence was submitted to GenBank (Accession No. AF089725). 2.4. 5∞-RACE assay

2. Methods 2.1. Southern blot analysis of total human genomic DNA Total human genomic DNA (Novagen, Madison, WI ) was digested with HindIII, NcoI, PstI, EcoRI, and BamHI at 37°C for 2 h, and a Southern blot analysis was performed. In brief, the samples were separated on a 0.8% agarose gel, transferred to a nylon membrane by diffusion using a Turboblotter (Schleicher and Schuell, Keene, NH ), and probed with a [32P]ATP

5∞-RACE PCR was performed using Marathon-ready human kidney cDNA (Clontech, Palo Alto, CA). The following PCR primers specific for the kNBC1 transcript were used: 5∞-CCATCCTAATACGACTCACTATAGGGC-3∞ (sense), 5∞-CCTGCTAAGCCTCCCAAATCC-3∞ (antisense). A nested pair of primers was then utilized: 5∞-CGATCACTATAGGGCTCGAGCG-3∞ (sense), 5∞GCCCTTCTCTATTCCTTTGGTTCTG-3∞ (antisense). A total of three separate 5∞-RACE PCR reactions were performed. A total of 18 fragments were subcloned into

N. Abuladze et al. / Gene 251 (2000) 109–122

PCR-ScriptSK(+) (Stratagene, La Jolla, CA) and characterized by direct sequencing. 2.5. S1 nuclease protection assay S1 nuclease protection assays were performed using the S1-Assay Protection Kit (Ambion, Austin, TX ) with the following modifications. A specific kNBC1 probe synthetic 98 bp oligonucleotide, which corresponded to a position 165–262 nucleotides prior to the kNBC1 translation initiation codon, was used with human kidney, heart, and pancreas total RNA, and tRNA. In brief, the probe was initially dephosphorylated using alkaline phosphatase, 5∞end-labelled with [c-32P] ATP using T4 polynucleotide kinase, and purified by ethanol precipitation. The 5∞ [32P] end-labeled oligonucleotide (2×105 cpm) was mixed with 25 mg of RNA. The mixture was incubated in 10 ml of S1 nuclease hybridization buffer (80% deionized formamide/100 mM sodium citrate, pH 6.4/300 mM sodium acetate, pH 6.4/1 mM EDTA) at 65°C for 10 min, then at 30°C for 20 h. The samples were then digested with 100 units of S1 nuclease for 1 h at 37°C in 200 ml of the S1 digestion buffer containing 5 mg of glycogen. The protected fragments were purified by ethanol precipitation, resuspended in the gel loading buffer, denatured at 90°C and analyzed on an 8% polyacrylamide sequencing gel. The dried gel was exposed to X-ray film to visualize the protected bands. The size of the protected fragments was determined by comparison with a DNA sequencing ladder (Amersham, Piscataway, NJ ), and a 10 bp ladder (Life Technologies, Rockville, MD). In all cases, sizes were confirmed with a minimum of three different experiments and gels.

111

DNA sequencing ladder (Amersham, Piskataway, NJ ), and a 25 bp ladder (Life Technologies, Rockville, MD). In all cases, sizes were confirmed with a minimum of three different experiments and gels. 2.7. Plasmid construction and transient gene expression assays Chimeric kNBC1-luciferase gene constructs were created by PCR from the ~8 kilobase EcoRI fragment derived from BAC clone 15800 and sublcloned into pGL3 basic vector (Promega, Madison, WI ). This vector contains KpnI and XhoI restriction sites. A KpnI restriction site was added to the 5∞-end of the sense primers and an XhoI restriction site was added to the 5∞-end of the antisense primers for cloning into the pGL3 vector. For cloning construct pGL3-606-R in the reverse orientation into the pGL3 vector, an XhoI restriction site was added to the 5∞-end of the sense primer, and a KpnI restriction site was added to the 5∞-end of the antisense primer. The chimeric kNBC1-luciferase gene constructs used in the transfection experiments in the transformed murine S1 cells are shown in Table 1. Plasmid purification was performed using a Wizard Minipreps DNA purification system (Promega, Madison, WI ), followed by phenol/ chloroform extraction and ethanol precipitation. A Dual-Luciferase reporter assay was used (Promega, Madison, WI ). Assays were performed in a dual autoinjector luminometer ( Turner Designs, Sunnyvale, CA). The experiments were performed in quadruplicate and the results reported as mean±S.D. 2.8. Cell culture and transfection procedure

2.6. Primer extension assay Confirmation of the transcriptional start site of kNBC1 was obtained by primer extension analysis. The following kNBC1 specific antisense primer was used: 5∞-GCCCTTCTCTATTCCTTTGGTTCTGAAC-3∞. Ten picomoles of primer were labeled with 0.03 mCi of [c-32P]ATP in a 10 ml reaction using 10 U of T4 polynucleotide kinase and purified with phenol and chloroform extraction. The specific radioactivity was 1×106 cpm/pmol of primer. Total RNA (50 mg) from human kidney, heart, and pancreas, and tRNA was combined with 0.5 pmol of primer and precipitated with ethanol. The mixture was dissolved in 10 ml of hybridization buffer, heated to 65°C for 10 min, and then cooled slowly to 21°C for 10 min. The reaction was reversetranscribed at 42°C for 1 h with AMV reverse transcriptase, denatured at 90°C, and analyzed on an 8% polyacrylamide sequencing gel. The dried gel was exposed to X-ray film to visualize the bands. The size of the fragments was determined by comparison with a

Murine SV40 transformed S1 proximal tubular cells (kindly provided by Dr G. Nagami), derived from the proximal tubule of the Tg(SV40E )Bri7 mouse, have been previously characterized and were grown as described to ~50–60% confluence in six well culture plates before transfection ( Kaunitz et al., 1993). One microgram of each chimeric promoter construct and 0.04 mg of the Rinella luciferase pRL-TK control vector were mixed with Lipofectamine (Gibco-BRL, Gaithersburg, MD) in serum-free media. The mixture was added to the cells for 8 h for transient transfection with the various plasmid constructs. The assays were performed 24 h following the initiation of transfection. In separate experiments, the promoter constructs pGL3-2595 (insert size 2595 nucleotides), pGL3-606 (insert size 606 nucleotides), and pGL3-69 (insert size 69 nucleotides) were transfected with equimolar amounts of plasmid. The results were not significantly different from those obtained when 1 mg of the promoter constructs was used.

112

N. Abuladze et al. / Gene 251 (2000) 109–122

Table 1 Chimeric kNBC1-luciferase gene constructs used in the transfection experiments in the transformed murine S1 cells Name

Position

5∞ primer

3∞ primer

pGL3-501 pGL3-1002 pGl3-1534 pGL3-2010 pGL3-2595 pGL3-2113 pPGL3-1613 pGL3-606 pGL3-368 pGL3- 202 pGL3-157 pGL3-96 pGL3-69 pGL3-424 pGL3-260 pGL3-123 pGL3-606-R

−2552 −2052 −2552 −1551 −2552 −1019 −2552 −543 −2552 +43 −2070 +43 −1570 +43 −563 +43 −325 +43 −159 +43 −114 +43 −53 +43 −26 +43 −563 −140 −563 −304 −563 −441 −563 +43

CCGGTACCTTGTACAGTCTTATCAGGAGTAG CCGGTACCTTGTACAGTCTTATCAGGAGTAG CCGGTACCTTGTACAGTCTTATCAGGAGTAG CCGGTACCTTGTACAGTCTTATCAGGAGTAG CCGGTACCTTGTACAGTCTTATCAGGAGTAG CGGTACCGGAGACAAGAAGGCAACAC CGGTACCCCAATCTTCCTTTATGGATG CCGGTACCCTGCTTTTCTTTCTCTCTCTC GCGGTACCGTGTACAATATTTTACTTGGGC CGGTACCGACCTGGGGGAAATTCTAG GGGTACCGGGGGATCTCAGGATTG CGGTACCGTTCAAGCTGGCTTTTG GCGGTACCTTGGATGAGTCATAAGTGGAG CCGGTACCCTGCTTTTCTTTCTCTCTCTC CCGGTACCCTGCTTTTCTTTCTCTCTCTC CCGGTACCCTGCTTTTCTTTCTCTCTCTC CCTCGAGCTGCTTTTCTTTCTCTCTC

CTCGAGGTGTTGCCTTCTTGTCTCC CCTCGAGCATCCATAAAGGAAGATTGG CTCGAGCAACATGGTGAAACCCTG CCTCGAGGAGAGAGAGAAAGAAAAGCAG CCTCGAGCTTTAGCCCTCAGATGAGC CCTCGAGCTTTAGCCCTCAGATGAG CCTCGAGCTTTAGCCCTCAGATGAG CCTCGAGCTTTAGCCCTCAGATGAGC CCTCGAGCTTTAGCCCTCAGATGAG CCTCGAGCTTTAGCCCTCAGATGAG CCTCGAGCTTTAGCCCTCAGATGAG CCTCGAGCTTTAGCCCTCAGATGAG CCTCGAGCTTTAGCCCTCAGATGAG CCTCGAGCCTCTGTCCAATGCCTAAAG CCCTCGAGGCCCAAGTAAAATATTGTACAC CCTCGAGCTAGAATTTCCCCCAGGTC ATGGTACCCTTTAGCCCTCAGATGAGC

3. Results and discussion 3.1. Structure of the human sodium bicarbonate cotransporter NBC1 gene To determine whether kNBC1, pNBC1 and hNBC1 are transcribed from a single gene, total human genomic DNA was digested with HindIII, NcoI, PstI, EcoRI, and BamHI and probed with a [32P] 95 bp synthetic oligonucleotide common to both kNBC1 (271–365), pNBC1 (371–465), and hNBC1 (298–392). As shown in Fig. 1, a single band was detected in each lane, suggesting that a single gene encodes all three transcripts. BAC clone 15800 was initially detected with the following kNBC1 specific primers: 5∞-CAAACTGGAGGAGCGACGGAAG-3∞ and 5∞-TCTCAAGAGATTCTCTCCTTTTCCTTC-3∞. Direct sequencing showed that this clone contained the translational start site of kNBC1, pNBC1 and hNBC1. Fig. 2 shows a Southern blot analysis of purified genomic DNA from BAC clone 15800 probed with a 94 bp synthetic oligonucleotide specific for human kNBC1 sequence (175–268), and a 95 bp synthetic oligonucleotide specific for the human pNBC1 sequence (118–212). The pNBC1 probe was common to the hNBC1 sequence (45–139). In separate mapping studies of BAC clone 15800, the two probes were found to be separated by >50 kilobases. The genomic organization of the human NBC1 (SLC4A4) gene was determined by sequencing genomic DNA inserts from eight overlapping BAC clones that spanned the published pNBC1, kNBC1, and hNBC1 sequences (Fig. 3a). The gene spans ~450 kilobases and contains 26 exons. All intron–exon splice junctions follow the GT–AG rule and conform to established exon boundary sequences. The translational start site of the longest of the three N-terminal variants, pNBC1, was

Fig. 1. Southern blot of total human genomic DNA. Total human genomic DNA was digested with HindIII, NcoI, PstI, EcoRI, and BamHI and probed with a [32P]ATP labeled 95 bp synthetic oligonucleotide common to both kNBC1 (271–365), pNBC1 (371–465), and hNBC1 (298–392). The position of the size standards (kilobases) was derived from HindIII-digested l phage DNA.

N. Abuladze et al. / Gene 251 (2000) 109–122

113

the 3∞-untranslated sequence of each gene (Fig. 5). Interestingly, the intron–exon boundaries are more similar in the proposed transmembranous regions of pNBC1 and NBC3 proteins than in the N- and C-terminus cytoplasmic domains. Three of seven intron–exon boundaries in the N-terminal cytoplasmic regions of pNBC1 and NBC3 are identical. In the transmembranous region of pNBC1 and NBC3, 10 of 15 intron– exon boundaries are identical. No identical intron–exon boundaries were found in the C-terminal cytoplasmic regions of pNBC1 and NBC3. Furthermore, of the 13 identical intron–exon boundaries, five interrupted an amino acid, which is an arginine at three sites. The conservation of intron–exon boundaries, particularly in the transmembranous region, suggests gene duplication and divergence as the origin of the two isoforms. Interestingly, the amino acid similarity between pNBC1 and NBC3 in the N-terminus and transmembrane domains mirrored the intron–exon identity. Specifically, the N-terminal cytoplasmic domain of each protein was 33% identical, whereas the transmembrane domain showed 59% identity. A similar correlation between intron–exon and amino acid identity has been found among members of the NHE family ( Kandasamy and Orlowski, 1996). 3.2. Transcriptional start site and nucleotide sequence of the 5∞-flanking region of the kNBC1 transcript Fig. 2. Southern blot of genomic DNA from BAC clone 15800. Genomic DNA was purified from BAC clone 15800 and digested with BamHI, HindIII, and EcoRI. Southern analysis was performed using an (a) 94 bp [32P]ATP labeled synthetic specific kNBC1 oligonucleotide probe (175–268), and (b) a [32P]ATP random primed and labeled 95 bp synthetic specific pNBC1 oligonucleotide probe (118–212). The pNBC1 probe was common to the hNBC1 sequence (45–139). BAC clone 15800 was shown to contain the translation initiation site of both pNBC1 and kNBC1. An ~8 kilobase EcoRI fragment from BAC clone 15800 containing the 5∞-UTR of kNBC1 was subcloned into PCR-ScriptSK(+) (Stratagene, La Jolla, CA). The position of the size standards (kilobases) was derived from HindIII-digested l phage DNA.

mapped to exon 2 ( Fig. 3b). Exon 25 contains the TGA stop codon. The protein-coding exons range in size from 69 to 272 nucleotides, with the exception of exon 26, which contains a large 3∞-untranslated region of 4293 nucleotides ( Fig. 4). As a general rule, members of gene families have been shown to have a similar genomic organization. The results of the present study indicate that this also appears to be true for members of the NBC gene family, which have been characterized thus far. We have recently determined the genomic organization of the NBC3 (SLC4A7) gene (Pushkin et al., 1999a). A comparison of the exons in the coding regions of the human NBC1 and NBC3 genes indicates that the structure of both genes is similar, including the large 3∞ exon containing

To determine the mechanisms responsible for regulation of kNBC1 at a transcriptional level, the transcription initiation site of kNBC1 was characterized using S1 nuclease protection and primer extension analysis, and 5∞-RACE. S1 nuclease protection studies ( Fig. 6a) were performed using total RNA isolated from human kidney, heart, and pancreas. A specific kNBC1 probe synthetic 98 bp oligonucleotide, which corresponded to a position 165–262 nucleotides prior to the kNBC1 translation initiation codon, was used. Using kidney total RNA, a major protected fragment corresponding to a position 192 nucleotides upstream from the translation start site (size of protected fragment: 27 nucleotides) and a minor protected fragment at a position 182 nucleotides upstream from the start site were detected (size of protected fragment: 17 nucleotides). No fragments were detected with human heart and pancreas total RNA. Similar results were obtained in the primer extension studies ( Fig. 6b). A major band of 140 nucleotides and a minor band of 130 nucleotides were detected using total kidney RNA. No bands were detected with heart or pancreas RNA. These results also correspond to a major transcription initiation site 192 nucleotides upstream from the translation start site and a minor transcription initiation site, 182 nucleotides upstream from the translation start site. The 5∞-RACE experiments ( Fig. 6c) confirmed these findings. The PCR products

Fig. 3. Organization of the human NBC1 (SLC4A4) gene. (a) Overlapping BAC clones encoding the human NBC1 gene. Eight BAC clones spanning the entire coding region of kNBC1, pNBC1, and hNBC1 were isolated and mapped by direct sequencing, restriction endonuclease cleavage, and Southern blot analysis using [32P]ATP random primed synthetic oligonucleotide probes corresponding to various regions of the pNBC1 and kNBC1 cDNA. The alignment of pNBC1 cDNA (the longest of the three transcripts) and each BAC clone is depicted. The arrows represent the position of introns in pNBC1 cDNA in relationship to each BAC clone. (b) Position of introns within the pNBC1 amino acid sequence. The exon sequence surrounding each intron is shown above the amino acid sequence.

114 N. Abuladze et al. / Gene 251 (2000) 109–122

N. Abuladze et al. / Gene 251 (2000) 109–122

115

Fig. 4. Intron–exon boundaries of the NBC1 (SLC4A4) gene. Nucleotide sequences at the intron–exon boundaries are shown with exon sequences in upper-case letters and intron sequences in lower-case letters. Exons are numbered from left to right. The length of each exon determined by sequence analysis is shown.

116

N. Abuladze et al. / Gene 251 (2000) 109–122

Fig. 5. Intron–exon boundaries in human pNBC1 and NBC3. The amino acid sequences of pNBC1 and NBC3 were aligned using GeneWorks software (Oxford Molecular Group, Oxford, UK ). The locations of the intron–exon boundaries are shown by arrows.

clustered at a position 192 nucleotides upstream from the kNBC1 translation initiation codon, with a minor start site located 182 nucleotides upstream. The data confirm that the major transcription start site is at 43 nucleotides upstream from the 5∞-terminus of the published human kNBC1 sequence (Burnham et al., 1997). Thus, we have defined the residue at position 192 nucleotides upstream from the translation initiation as the major transcriptional start point in human kidney. The presence of second minor transcriptional start site 10 nucleotides upstream of the major site suggests that transcriptional initiation may be staggered. As shown in Fig. 7, the transcription and translation initiation sites of kNBC1 are in intron 3 of the NBC1 gene. In the first alternative exon of kNBC1, the last 313 nucleotides of intron 3 are coupled to exon 4, which is common to pNBC1 and hNBC1. The existence of multiple transcriptional initiation sites is not uncommon for GC-rich promoters, which lack TATA boxes. Furthermore, multiple start sites have

also been found in promoters, which have atypical TATA boxes ( Kandasamy and Orlowski, 1996). An additional feature of the kNBC1 major start site is the lack of a typical cap signal motif, NCA (G/C/T ) 0 (Bucher, 1990). In some promoters, the use of a T as the initiation site is also found. Although the kNBC1 major initiation site lacks this canonical cap sequence, it is estimated that this motif is absent in 40% of promoters, suggesting that it is not essential for promoter function. 3.3. Functional promoter activity of the 5∞-flanking region of the kNBC1 transcript To determine whether the 5∞-flanking region of the kNBC1 transcript contains a functional promoter, chimeric kNBC1-luciferase gene constructs were created by PCR from the ~8 kilobase EcoRI fragment derived from BAC clone 15800 (Fig. 2), and subcloned into the pGL3 promoterless vector (Promega, Madison, WI ).

N. Abuladze et al. / Gene 251 (2000) 109–122

117

Fig. 6. Determination of the transcription initiation site of the kNBC1 transcript. (a) S1 nuclease protection analysis. A synthetic [c-32P]-labeled oligonucleotide 98 nucleotides in length, corresponding to a position 165–262 nucleotides prior to the kNBC1 translation initiation codon, was hybridized with 25 mg of total human kidney. S1 nuclease protection analysis was performed as described in Section 2. The protected fragments were analyzed on an 8% denaturing polyacrylamide sequencing gel. M, size markers; K, human kidney total RNA; H, human heart total RNA; P, human pancreas total RNA; t, tRNA; GATC, sequencing ladder. (b) Primer extension analysis: The following kNBC1 specific antisense primer was used: 5∞-GCCCTTCTCTATTCCTTTGGTTCTGAAC-3∞. M, size markers; K – human kidney total RNA; H, human heart total RNA; P, human pancreas total RNA; t, tRNA; GATC, sequencing ladder. (c) 5∞-RACE PCR performed using Marathon-ready human kidney cDNA (Clontech, Palo Alto, CA) with primers specific for the kNBC1 transcript. A total of three separate 5∞-RACE PCR reactions were performed. Eighteen fragments (depicted by diamonds) were subcloned into PCR-ScriptSK(+) (Stratagene, La Jolla, CA), and characterized by direct sequencing. The 5∞-RACE experiments generated products that were clustered around the start site corresponding to both the major and minor protected fragments in the S1 nuclease studies.

Fig. 7. Transcription initiation site of kNBC1 in intron 3. The nucleotides shown in bold in exon 4 are common to pNBC1 and hNBC1.

118

N. Abuladze et al. / Gene 251 (2000) 109–122

The PCR primers used to generate the various chimeric constructs are shown in Table 1. Murine SV40 transformed S1 proximal tubular cells were transfected with the chimeric kNBC1-luciferase gene contructs and the Rinella luciferase pRL-TK control vector, which served as an internal control for monitoring the transfection efficiency. The promoter activity of the chimeric luciferase constructs was assayed using a dual luciferase reporter assay. As shown in Fig. 8a, pGL3-2595 exhibited an ~10-fold increase in luciferase activity compared with the pGL3 vector. This indicated that the 5∞-flanking region of the kNBC1 transcript contains a functional promoter. Successive truncation of the 5∞-terminus revealed that the promoter activity was minimal within nucleotides −2552 to −543 relative to the transcription start site. Furthermore, the activity of pGL3-606 containing nucleotides −563 to +43 relative to the transcriptional start site was not distinguishable from the activity of the pGL3-2595 (−2552 to +43). The reverse orientation of pGL3-606 (pGL3-606-R) did not demonstrate any promoter activity (Fig. 8b). To determine the minimum promoter, pGL3-606 was further truncated. As shown in Fig. 8b, the 5∞-flanking region of the kNBC1 promoter (−159 to +43) is sufficient for promoter activity. 3.4. Analysis of the kNBC1 promoter transcription factor binding sites Fig. 9 shows the nucleotide sequence of the 5∞-flanking region of the kNBC1transcript. The proximal promoter (−952 to +43) of the kNBC1 transcript was analyzed using the MatInspector program, which is continuously updated from the TRANSFAC database. There is an atypical TATA sequence (CTTATAATT ) at position −33. The position of the TATA box corresponds well with the accepted location of basal transcription factor binding, usually at 25–30 nucleotides upstream from the demonstrated transcription start site. There are three Oct1-like sites at positions −68, −712, and −812 (consensus sequence ATGCAAAT ); a GC box that binds the Sp-1 protein at position −125 consensus sequence (G/T )GGGCGG(G/A)(G/A) (C/T ); an NF-Y site at −196 (consensus CCAAT ); two C/EBP-like sites at positions −253, and −480 (consensus TKNNGYAAK ); an AP-1 site at position –279 (consensus sequence TGASTMA); three Nkx2.5 sites at positions −286, −296, and −329 (consensus sequence TNAAGTG); two HNF-3 sites at positions −356 and −637 (consensus AWTRTTKRYTY ); a canonical estrogen-responsive element (ER) half site at position −407 (GGTCA) (Dana et al., 1994); a GATA site at −412 [consensus sequence (A/T )GATA(A/G)] at position −412; and two E box sites at positions −651, and −796. Analysis of the kNBC1 promoter reveals binding sites for multiple potentially important transcription factors.

The atypical TATA-box at position −33 contains core and flanking residues that differ from the canonical sequence [G/CTAT AAA(A/T ) (G/A)] at two residues O (Bucher, 1990). This DNA element most likely constitutes the functional site for localization of basal transcription factors in the promoter. SP1 sites are often found within 200 nucleotides upstream of the transcription start site and can increase the rate of transcription. The kNBC1 promoter differs from the promoter of a number of TATA-less housekeeping genes that have multiple SP-1 binding sites, and are highly (G+C )-rich. CCAAT boxes are typically located 50–200 nucleotides upstream from the transcription start site and are recognized by the NF-Y transcription factor. The atypical CCAAT-box at position −194 in the kNBC1 promoter differs from canonical sequence in the +1 position. The substitution of an A for a G in this position is observed in 12% of genes examined (Bucher, 1990). The NF-Y binding site in the kNBC1 promoter may play a role in significantly increasing the efficiency of transcription, as has been demonstrated in many mammalian promoters (Bucher, 1990). The binding sites for transcription factors OCT-1, AP-1, SP-1, GATA, C/EBP, NF-kB, and Nkx2.5, may be relevant to basal as well as inducible transcriptional regulation of the kNBC1 promoter. The results obtained using truncated promoter constructs suggest that the SP-1 and Oct-1 binding sites immediately upstream from the TATA-box play a role in increasing the efficiency of transcription. C/EBP proteins have been detected in the kidney and are involved in the expression of phosphoenolpyruvate carboxykinase and gene expression in response to cAMP (Liu and Curthoys, 1996). The HNF-3 motif belongs to a family of transcription factors that share homology with the winged helix DNA-binding domain. Members of this family of transcription factors have been implicated in cell-fate determination, in organogenesis, and in celltype-specific gene expression. Two E-boxes were identified, which could be binding sites for basic helix–loop– helix (bHLH ) proteins. E-boxes have been identified in the kidney NKCC2 promoter and are important in gene expression in muscle, neurons, lymphocytes, and pancreas (Igarashi et al., 1996). 3.5. Mapping of hNBC1 hNBC1 has an identical amino acid sequence to pNBC1 but differs at the 5∞-UTR. BAC clone sequencing revealed that the first alternative exon of the hNBC1 transcript is derived from the last 43 nucleotides of intron 1 coupled to exon 2 of pNBC1 (Fig. 10). Sequence analysis ( TESS program, http://dot. imgen.bcm.tmc.edu: 9331/gene-finder/gf.html ) of the genomic region 1000 nucleotides upstream to the putative transcription start site of hNBC1 did not reveal a potential promoter, suggesting that hNBC1 is derived

N. Abuladze et al. / Gene 251 (2000) 109–122

119

Fig. 8. (a) Functional activity of the kNBC1 promoter. Chimeric kNBC1-luciferase gene constructs were created by PCR from the ~8 kilobase EcoRI fragment derived from BAC clone 15800 and sublcloned into pGL3 basic vector (Promega, Madison, WI ). Murine SV40 transformed S1 proximal tubular cells were transfected with 1 mg of each chimeric promoter construct and 0.04 mg of the Rinella luciferase pRL-TK control vector using Lipofectamine (Gibco-BRL, Gaithersburg, MD) in serum-free media. Assays were performed in a dual auto-injector luminometer ( Turner Designs, Sunnyvale, CA). Experiments were performed in quadruplicate and the results reported as mean±S.D. Successive truncation of the 5∞-terminus revealed that the promoter activity was minimal within nucleotides −2552 to −543 relative to the transcription start site. Furthermore, the activity of pGL3-606 containing nucleotides −563 to +43 relative to the transcriptional start site was not distinguishable from the activity of the pGL3-2595 (−2552 to +43). The activity of pGL3-2595, pGL3-2113, pGL3-1613, and pGL3-606 was significantly greater than the activity of the pGL3 vector, P<0.05 (Dunnett’s t-test). (b) Determination of the minimal kNBC1 promoter. To determine the minimum promoter, pGL3-606 was further truncated. The 5∞-flanking region of the kNBC1 promoter (−159 to+43) is sufficient for promoter activity. The activity of pGL3-96, pGL3-69, pGL3-424, pGL3-260, pGL3-123, and pGL3-606-R was significantly less than the activity of pGL3-606, P<0.05 (Dunnett’s t-test).

120

N. Abuladze et al. / Gene 251 (2000) 109–122

Fig. 9. Potential transcription factor binding sites within the kNBC1 promoter. Numbering of nucleotides begins with +1 at the transcription start site. The sequence was analyzed for known transcription factors using the MatInspector and the TRANSFAC database. Multiple consensus sequences for known transcription factors that are potentially of functional significance are underlined (see text for details).

Fig. 10. Mapping of hNBC1in the NBC1 gene. BAC clone sequencing revealed that the first alternative exon of the hNBC1 transcript is derived from the last 43 nucleotides of intron 1 coupled to exon 2 of pNBC1. Sequence analysis of the genomic region upstream to the putative transcription start site of hNBC1 did not reveal a potential promoter, suggesting that hNBC1 is derived from alternative splicing of a cryptic splice site in intron 1.

from alternative splicing of a cryptic splice site in intron 1. The 5∞-UTR of several genes can function as a tissuespecific translational enhancer and in regulating transcriptional activity. Analysis of the EST database has revealed several clones containing the 5∞-UTR of pNBC1 but failed to detect any clones containing the hNBC1 5∞-UTR, suggesting that the tissue distribution of hNBC1 may be more restricted. 3.6. N-terminal variants of NBC1 and other members of the bicarbonate transporter superfamily NBC transporters belong to the bicarbonate transporter superfamily (BTS), which also includes the Cl−/

HCO− exchanger gene family consisting of AE1, AE2, 3 and AE3 (Alper, 1991). The variations among the kNBC1, pNBC1, and hNBC1 transcripts result from differences in their N-termini. All three anion exchanger genes also encode multiple N-terminal transcripts. The kidney-specific AE1 transcript (kAE1) expressed on the basolateral membrane of type A intercalated cells in the renal collecting duct is transcribed from an alternative promoter in the third intron of the erythrocyte transcription unit (Alper, 1991). mRNAs encoding three N-terminal AE2 variants (AE2a, AE2b, and AE2c) were also shown to be potentially transcribed from alternate promoters ( Wang et al., 1996). The alternative promoter for AE2b is located in intron 2 of the AE2 gene, whereas

N. Abuladze et al. / Gene 251 (2000) 109–122

the promoter for AE2c is located in intron 5. The AE3 gene encodes brain (bAE3) and cardiac (cAE3) variants, and the cardiac variant is thought to be transcribed from an alternative promoter in intron 6 (Linn et al., 1992). The data suggest that it is likely that new additional members of the BTS family that are characterized will also have multiple N-terminal variants. 3.7. Role of mutations in NBC1 in mediating proximal renal tubular acidosis and ocular abnormalities Igarashi et al. (1994, 1999) reported two unrelated patients with homozygous mutations in the kNBC1 transcript. The first patient had a homozygous R298S mutation, and the second patient had a R510H mutation. Both mutations decreased the functional activity of the transporter to ~55% of the wild-type activity when assayed in a eukaryotic expression system. Clinically, the patients were mentally retarded with short stature and had severe proximal renal tubular acidosis, systemic acidemia and hypokalemia. Ocular abnormalities were detected including glaucoma, cataracts, and band keratopathy. No measurements of pancreatic fluid bicarbonate concentration/pH were made, although the serum amylase was elevated, suggesting a possible pancreatic phenotype. Head CT scans revealed bilateral calcifications of the basal ganglia. Thyroid abnormalities were also documented. These patients clinically resemble two brothers previously characterized by Winsnes et al. (1979). Given the findings in the present study that the kNBC1, pNBC1, and hNBC1 transcripts are derived from the NBC1 (SLC4A4) gene, the location of the mutations described by Igarashi et al. would be expected to affect all three transcripts. Since basolateral proximal tubular bicarbonate transport is mediated predominantly by kNBC1, inactivating mutations in the kNBC1 transcript can account for the renal phenotype in this disorder. The ocular abnormalities can also be ascribed to inactivating mutations in the NBC1 gene, although it is less clear which cell types are affected and which N-terminal variants are involved. Electrogenic sodium bicarbonate cotransport has been previously functionally characterized in corneal endothelial cells (Jentsch et al., 1984), retinal astrocytes and Muller cells (Newman, 1999), and retinal pigment epithelium ( Kenyon et al., 1997). Both kNBC1 and pNBC1 transcripts are present in corneal endothelial cells ( Usui et al., 1999); however, it is unknown as to which protein mediates electrogenic sodium bicarbonate cotransport in this cell type. It is expected that additional patients with transcript specific mutations will clarify some of these issues. We initially demonstrated that pNBC1 is expressed in thyroid, which may account for the thyroid abnormalities in these patients (Abuladze et al., 1998a). However, the function of electrogenic sodium bicarbonate cotransport in thyroid has not yet been studied.

121

Finally, although electrogenic sodium bicarbonate cotransport has been detected in specific cell types in the brain (Rose and Ransom, 1998), the mechanisms by which inactivating mutations in NBC1 result in basal ganglia calcification and mental retardation remain to be determined. Of interest, NBC1 transcripts are expressed in several additional tissues including prostate, colon, stomach, and heart. Whether more subtle abnormalities in the function of these organs are present in patients with NBC1 mutations will require further study.

Acknowledgements This work was supported by NIH grant DK46976, the Iris and B. Gerald Cantor Foundation, the Max Factor Family Foundation, the Verna Harrah Foundation, the Richard and Hinda Rosenthal Foundation, and the Fredericka Taubitz Foundation. N.A. is supported by a training grant from the National Kidney Foundation of Southern California J891002.

References Abuladze, N., Lee, I., Newman, D., Hwang, J., Boorer, K., Pushkin, A., Kurtz, I., 1998a. Molecular cloning, chromosomal localization, tissue distribution, and functional expression of the human pancreatic sodium bicarbonate cotransporter. J. Biol. Chem. 273, 17689–17695. Abuladze, N., Lee, I., Newman, D., Hwang, J., Pushkin, A., Kurtz, I., 1998b. Axial heterogeneity of sodium-bicarbonate cotransporter expression in the rabbit proximal tubule. Am. J. Physiol. 43, 628–633. Alper, S.L., 1991. The band 3-related anion exchanger (AE) gene family. Annu. Rev. Physiol. 53, 549–564. Amlal, H., Burnham, C.E., Soleimani, M., 1999. Characterization of Na+/HCO− cotransporter isoform NBC-3. Am. J. Physiol. 276, 3 F903–F913. Bucher, P., 1990. Weight matrix descriptions of four eukaryotic RNA polymerase II promoter elements derived from 502 unrelated promoter sequences. J. Mol. Biol. 212, 563–578. Burnham, C.E., Amlal, H., Wang, Z., Shull, G.E., Soleimani, M., 1997. Cloning and functional expression of a human kidney Na+: HCO− cotransporter. J. Biol. Chem. 72, 19111–19114. 3 Choi, I., Romero, M.F., Khandoudi, N., Bril, A., Boron, W.F., 1999. Cloning and characterization of a human electrogenic Na+–HCO− cotransporter isoform (hhNBC ). Am. J. Physiol. 276, 3 C576–C584. Dana, S.L., Hoener, P.A., Wheeler, D.A., Lawrence, C.B., McDonnell, D.P., 1994. Novel estrogen response elements identified by genetic selection in yeast are differentially responsive to estrogens and antiestrogens in mammalian cells. Mol. Endocrinol. 8, 1193–1207. Igarashi, T., Ishii, T., Watanabe, K., Hayakawa, H., Horio, K., Sone, Y., Ohga, K., 1994. Persistent isolated proximal renal tubular acidosis — a systemic disease with a distinct clinical entity. Ped. Nephrol. 8, 70–71. Igarashi, P., Whyte, D.A., Li, K., Nagami, G.T., 1996. Cloning and kidney cell-specific activity of the promoter of the murine renal Na–K–C1 cotransporter gene. J. Biol. Chem. 271, 9666–9674. Igarashi, T., Inatomi, J., Sekine, T., Cha, S.H., Kanai, Y., Kunimi, M., Tsukamoto, K., Satoh, H., Shimadzu, M., Tozawa, F., Mori,

122

N. Abuladze et al. / Gene 251 (2000) 109–122

T., Shiobara, M., Seki, G., Endou, H., 1999. Mutations in SLC4A4 cause permanent isolated proximal renal tubular acidosis with ocular abnormalities. Nat. Genet. 23, 264–266. Ishibashi, K., Sasaki, S., Marumo, F., 1998. Molecular cloning of a new sodium bicarbonate cotransporter cDNA from human retina. Biochem. Biophys. Res. Commun. 246, 535–538. Ishiguro, H., Steward, M.C., Lindsay, A.R., Case, R.M., 1996. Accumulation of intracellular HCO− by Na+–HCO− cotransport in 3 3 interlobular ducts from guinea-pig pancreas. J. Physiol. 495, 169–178. Jentsch, T.J., Keller, S.K., Koch, M., Wiederholt, M., 1984. Evidence for coupled transport of bicarbonate and sodium in cultured bovine corneal endothelial cells. J. Membr. Biol. 81, 189–204. Kandasamy, R.A., Orlowski, J., 1996. Genomic organization and glucocorticoid transcriptional activation of the rat Na+/H+ exchanger NHE3 gene. J. Biol. Chem. 271, 10551–10559. Kaunitz, J.D., Cummins, V.P., Mishler, D., Nagami, G.T., 1993. Inhibition of gentamicin uptake into cultured mouse proximal tubule epithelial cells by -lysine. J. Clin. Pharmacol. 33, 63–69. Kenyon, E., Maminishkis, A., Joseph, D.P., Miller, S.S., 1997. Apical and basolateral membrane mechanisms that regulate pHi in bovine retinal pigment epithelium. Am. J. Physiol. 273, C456–C472. Kwon, T.-H., Pushkin, A., Abuladze, N., Nielsen, S., Kurtz, I., 2000. Immunoelectron microscopic localization of NBC3 sodium bicarbonate cotransporter in rat kidney. Am. J. Physiol. 278, F327–F336. Linn, S.C., Kudrycki, K.E., Shull, G.E., 1992. The predicted translation product of a cardiac AE3 mRNA contains an N terminus distinct from that of the brain AE3 Cl−/HCO− exchanger. Cloning 3 of a cardiac AE3 cDNA, organization of the AE3 gene and identification of an alternative transcription initiation site. J. Biol. Chem. 267, 7927–7935. Liu, X., Curthoys, N.P., 1996. cAMP activation of phosphoenolpyr-

uvate carboxykinase transcription in renal LLC-PK1-F+ cells. Am. J. Physiol. 271, F347–F355. Muallem, S., Loessberg, P.A., 1990. Intracellular pH-regulatory mechanisms in pancreatic acinar cells. I. Characterization of H+ and HCO− transporters. J. Biol. Chem. 265, 12806–12812. 3 Newman, E.A., 1999. Sodium-bicarbonate cotransport in retinal astrocytes and Muller cells of the rat. Glia 26, 302–308. Pushkin, A., Abuladze, N., Lee, I., Newman, D., Hwang, J., Kurtz, I., 1999a. Cloning, tissue distribution, genomic organization, and functional characterization of NBC3, a new member of the sodium bicarbonate cotransporter family. J. Biol. Chem. 274, 16569–16575. Pushkin, A., Yip, K.P., Clark, I., Abuladze, N., Kwon, T.H., Tsuruoka, S., Schwartz, G.J., Nielsen, S., Kurtz, I., 1999b. NBC3 expression in the rabbit collecting duct: colocalization with the vacuolar H+-ATPase. Am. J. Physiol. 277, F974–F981. Romero, M.F., Hediger, M.A., Boulpaep, E.L., Boron, W.F., 1997. Expression cloning and characterization of a renal electrogenic Na+/HCO− cotransporter. Nature 38, 409–413. 3 Rose, C.R., Ransom, B.R., 1998. pH regulation in mammalian glia. In: Kaila, K., Ransom, B.R. ( Eds.), pH and Brain Function. Wiley, New York, pp. 253–275. Schmitt, B.M.D., Biemesderfer, D., Romero, M.F., Boulpaep, E.L., Boron, W.F., 1999. Immunolocalization of the electrogenic Na+– HCO− cotransporter in mammalian and amphibian kidney. Am. 3 J. Physiol. 276, F27–F38. Usui, T., Seki, G., Amano, S., Oshika, T., Miyata, K., Kunimi, M., Taniguchi, S., Uwatoko, S., Fujita, T., Araie, M., 1999. Functional and molecular evidence for Na+–HCO− cotransporter in human 3 endothelial cells. Pflug. Arch. Eur. J. Physiol. 438, 458–462. Wang, Z., Schultheis, P.J., Shull, G.E., 1996. Three N-terminal variants of the AE2 Cl−/HCO− exchanger are encoded by mRNAs tran3 scribed from alternative promoters. J. Biol. Chem. 271, 7835–7843. Winsnes, A., Monn, E., Stokke, O., Feyling, T., 1979. Congenital persistent proximal type renal tubular acidosis in two brothers. Acta Paed. Scand. 68, 861–868.