Ca2+ exchanger 3, expressed in fetal brain

Gene 348 (2005) 143 – 155 www.elsevier.com/locate/gene

Molecular cloning and characterization of two novel truncated isoforms of human Na+/Ca2+ exchanger 3, expressed in fetal brainB Rose-Marie Lindgrena, Jian Zhaob, Susanne Hellera, Hanna Berglindb, Monica Niste´rb,T a Department of Genetics and Pathology, Uppsala University, Rudbeck Laboratory, SE-751 85 Uppsala, Sweden Department of Oncology-Pathology, Karolinska Institutet, CCK R8:05, Karolinska University Hospital Solna, SE-171 76 Stockholm, Sweden

b

Received 28 June 2004; received in revised form 24 November 2004; accepted 6 January 2005 Received by E. Boncinelli

Abstract The human gene encoding the Na+/Ca2+ exchanger family member 3 (NCX3) undergoes extensive alternative splicing, with four variants previously identified. In this study, we report two novel alternative transcripts encoding two N-terminally truncated NCX3 proteins specifically expressed in human fetal brain. The identified transcripts, designated NCX3-tN.1 and NCX3-tN.2, are approximately 2.8 kb and 2.9 kb, respectively. The open reading frames (ORFs) are predicted to encode separately a 284 and a 298 amino acid (aa) polypeptide. Sequence analysis and bioinformatics reveal that NCX3-tN.1 and NCX3-tN.2 are the result of alternative splicing of the NCX3 gene. They have their own potential start codons and unique 5V untranslated regions (UTRs) that are different from those of the known NCX3 variants. The variants include a part of intron 2 of the original gene organization as their first exon (exon baQ) at the 5V end of the novel transcripts. NCX3-tN.2 consists of six exons including exon baQ and exons 4, 6, 7, 8 and 9 of NCX3, while NCX3-tN.1 lacks exon 4, but is otherwise similar to NCX3tN.2. Expression studies show that both variants can be translated into protein and NCX3-tN.1 seems more efficiently translated. Based on their structural features, NCX3-tN.1 and NCX3-tN.2 proteins are potentially involved in regulation of Na+/Ca2+ homeostasis. D 2005 Elsevier B.V. All rights reserved. Keywords: Alternative splicing; CNS; NCX3; SLC8A3

1. Introduction Intracellular calcium is an important messenger of many physiological functions and multiple pathways regulate its Abbreviations: aa, amino acid(s); bp, base pair(s); cDNA, DNA complementary to RNA; DD-PCR, differential display-polymerase chain reaction; dNTP, deoxyribonucleoside triphosphate; EST, expressed sequence tag; EtdBr, ethidium bromide; G3PDH, glyceraldehyde 3phosphate dehydrogenase; kb, kilobase(s) or 1000 bp; MTN, Multiple Tissue Northern (Blot); NCX, Na+/Ca2+ exchanger; nr, non-redundant; ORF, open reading frame; RACE, rapid amplification of cDNA ends; SDS, sodium dodecyl sulfate; SLC8A1, 2 and 3, solute carrier family 8 (sodium– calcium exchanger) genes, members 1, 2 and 3; SSC, 0.15 M NaCl/0.015 M Na3 citrate pH 7.6; TM, trans-membrane domain; u, unit(s); UTR, untranslated region(s); XIP, exchanger inhibitory protein. B The cDNA sequences described in this work have been deposited to GenBank under the accession numbers AJ745101 and AJ745102. T Corresponding author. Tel.: +46 8 517 70 309; fax: +46 8 321 047. E-mail address: [email protected] (M. Niste´r). 0378-1119/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.gene.2005.01.003

concentration. The sodium–calcium exchanger (NCX) plays an important role in the maintenance of the intracellular Ca2+ homeostasis and is present in the plasma membrane of most cells. The primary function of NCX is to transport Ca2+ in exchange for Na+, in an inward or outward direction across the plasma membrane. The direction of transport depends on the prevailing Na+ and Ca2+ electrochemical gradients, with an exchange of three Na+ ions for one Ca2+ ion (for review, see Philipson and Nicoll, 2000; Quednau et al., 2004). Three mammalian members of the exchanger family have been identified, NCX1 (GenBank accession no. NM_ 021097), NCX2 (GenBank accession no. XM_375633) and NCX3 (GenBank accession no. AF508982). Although all three members arise from separate genes on different chromosomes, they share high amino acid identity (about 70%), especially in the hydrophobic regions (Nicoll et al., 1990, 1996b; Li et al., 1994). NCX1 was first cloned from canine cardiac sarcolemma (Nicoll et al., 1990) and later the

144

R.-M. Lindgren et al. / Gene 348 (2005) 143–155

corresponding human gene, SLC8A1 was mapped to chromosome 2p22.1 (Kraev et al., 1996). NCX1 is bestcharacterized and nearly ubiquitously expressed, but with a high level in heart, brain and kidney (Philipson et al., 1996). NCX2 and NCX3 seem to have similar properties as NCX1 but their expression appears to be restricted to brain and skeletal muscle (Li et al., 1994; Nicoll et al., 1996b). The NCX2 gene, SLC8A2, is located on chromosome 19q13.2 (Kikuno et al., 1999), while the SLC8A3 gene, encoding NCX3, maps on chromosome 14q24.2 (Gabellini et al., 2002). The mature exchanger protein consists of an extracellular N-terminal signal sequence and two sets of hydrophobic domains, separated by a large central hydrophilic cytoplasmic loop and followed by an intracellular C-terminal region. The hydrophobic domains comprise five trans-membrane domains (TMs) in the N-terminal region, TM1–TM5, and four TMs in the C-terminal region, TM6–TM9 (Nicoll et al., 1999; for review, see Nicoll et al., 2002). The large intracellular hydrophilic loop is involved in regulation of the exchanger, but is not essential for transport (Nicoll et al., 1999). The 60 amino acids (aa) long and similar h-1 and h-2 repeats are two conserved regions in the loop, with no suggested regulatory function (Philipson and Nicoll, 2000). The endogenous exchanger inhibitory peptide (XIP) is a 20 aa long region, located near the TM5, in the intracellular loop (Li et al., 1991). The XIP might have an auto regulatory function and may also be involved in Na+dependent inactivation of the exchanger (Li et al., 1991; Matsuoka et al., 1997). Further, in the N- and C-terminal halves of all the three NCX members, there are highly conserved internal repeat sequences, the a-1 and a-2 repeats (Nicoll et al., 1996a; Schwarz and Benzer, 1997). These repeats are located in between the TM2 and TM3 (a-1) and the TM7 and TM8 (a2), formed by gene duplication during evolution and considered to be involved in ion binding and translocation of the exchanger (Nicoll et al., 1996a; Schwarz and Benzer, 1997). The function of the exchanger is highly sensitive to mutations in the a-repeat regions (Nicoll et al., 1996a). Alternative splicing is common among the NCX genes. The NCX1 mRNA is alternatively spliced and at least 12 distinct variants have been described in various species and tissues (Furman et al., 1993; Kofuji et al., 1994; Lee et al., 1994; Reilly and Lattanzi, 1996; Quednau et al., 1997; Van Eylen et al., 2001). In addition, three alternative tissuespecific promoters have been reported (Lee et al., 1994; Barnes et al., 1997; Nicholas et al., 1998). Recently, several alternative transcripts of NCX3 have been characterized in different tissues, mainly generated by alternative splicing of exons encoding the large cytoplasmic loop (Gabellini et al., 2002). The NCX3 and the NCX1 genes have comparable arrangements of exons, for instance, both contain a large exon 2 and have a similar number of nucleotides between corresponding exons (Kraev et al., 1996; Gabellini et al.,

2002). The numbering of NCX1 and NCX3 exons differ, however, since the NCX3 gene includes only nine exons (numbered 1 to 9) (Gabellini et al., 2002) compared to the 12 exons of NCX1. Exons 2–5 of the NCX3 are homologous to the corresponding exons of NCX1, but exons 6–9 of NCX3 correspond to exons 9–12 of NCX1 (Kraev et al., 1996). The predicted NCX3 protein is coded by eight exons (2 to 9), with and without either exon 3 or 4. The start codon is predicted to be located in the 5V end of exon 2 and the stop in exon 9 (Gabellini et al., 2002). In spite of the NCX3 gene lacking the commonly alternatively spliced exons 6–8 of NCX1 (Kraev et al., 1996; Gabellini et al., 2002), alternative splicing of NCX3 does occur. Four alternatively expressed variants of NCX3 have been described, in both human (Gabellini et al., 2002) and rat (Quednau et al., 1997) (Fig. 1). The variants NCX3.1, NCX3.2 (found in human brain and neuroblastoma) and NCX3.3 (found in human skeletal muscle) are derived from alternative splicing in the region of exons 3–5 (Quednau et al., 1997; Gabellini et al., 2002). NCX3.4, found in human skeletal muscle, is a truncated isoform produced by skipping of exons 3–5, leading to a frame shift and generating a short NCX3 variant with a deletion of the C-terminal portion of the h-2 repeat domain and the whole hydrophobic C-terminal half of the protein (Gabellini et al., 2002). Since it has been reported that the C-terminal part of the protein is not crucial for ion transport (Li and Lytton, 1999; Van Eylen et al., 2001), NCX3.4 should be functional. In this study, we report two novel alternative transcripts encoding two N-terminally truncated NCX3 proteins from human fetal brain, designated as NCX3-tN.1 and NCX3tN.2. Genomic sequence analysis revealed that they result from alternative splicing of the NCX3 gene localized on chromosome 14. Sequence analysis showed that NCX3-tN.1 and NCX3-tN.2 have their own start codons and unique 5V untranslated regions (UTRs) that are different from the 5V UTRs of the currently known NCX3 variants. The new variants both start from a novel exon, designated as exon baQ, located in the intron 2 of the previously described gene organization. The expression of the two truncated mRNAs is likely to be induced from an alternative promoter of the NCX3 gene and is brain-specific in examined fetal tissues. NCX3-tN.2 consists of six exons including exon baQ, exon 4 and exons 6–9, while NCX3-tN.1 lacks exon 4, but is otherwise similar to NCX3-tN.2.

2. Materials and methods 2.1. Differential display-PCR—isolation of 47-1 cDNA fragment from human glioma cells A 281 base pair (bp) long cDNA (DNA complementary to RNA) fragment (47-1) was isolated as differentially expressed in the two clonal human malignant glioma cell lines, U-343 MG and U-343 MGa Cl 2:6 (Nister et al.,

R.-M. Lindgren et al. / Gene 348 (2005) 143–155

145

Exons 1

2

3 4

67 8

9

1

2

4

67 8

9

1

2

4 5 67 8

9

1

2

NCX3.1 NCX3.2 NCX3.3 6

NCX3.4 a

67 8

9

67 8

9

NCX3-tN.1

10kb

a

4

NCX3-tN.2

Fig. 1. The exon organization of different NCX3 isoforms. The four known splice variants, found in brain (NCX3.1, GenBank accession no. NM_033262, and NCX3.2, GenBank accession no. NM_182932) and skeletal muscle (NCX3.3, GenBank accession no. NM_183002, and NCX3.4, GenBank accession no. NM_182933) and NCX3-tN.1 and NCX3-tN.2, presented in this work are shown. The baQ indicates the previously not described exon baQ.

1986). Synthesis of cDNA and differential display-PCR (DD-PCR) was performed as described in the manufacturer manual (RNAimagek Kit 1, GenHunter Corporation, Nashville, TN). Total RNA from the two cell lines was used as template, prepared according to the acid-phenol guanidinium method (Chomczynski and Sacchi, 1987). The DD-PCRs were run for 40 cycles on a Perkin Elmer 2400 and generated 3V UTR PCR products were separated on denaturing polyacrylamide gels (6%). PCR products of interest were reamplified, cloned into pGEM or pGEM-T easy vectors (Promega, Madison, WI) and sequenced as described below. The cloned fragment (47-1) corresponded to an expressed sequence tag (EST), Morton Fetal Cochlea Homo Sapiens cDNA clone (GenBank accession no. AW022249) in the dbEST database. Part of this 654 bp long EST showed complete homology to the 281 bp long 47-1 sequence. The sequence also had homology to a part of the human chromosome 14 DNA sequence (GenBank accession no. AL135747), in the NCBI (National Center for Biotechnology Information) non-redundant (nr) database. Primers for the subsequent RACE (rapid amplification of cDNA ends)

PCR analysis were designed from the 47-1 sequence and from the corresponding EST (Table 1). 2.2. Marathon RACE-PCR, agarose gel electrophoresis, cDNA fragment purification and cloning The Human Fetal Brain BDk Marathon-Ready cDNA (BD Biosciences, Clontech, Palo Alto, CA) was used as template in RACE-PCR. The AdvantageR 2 PCR Kit (BD Biosciences, Clontech, Palo Alto, CA) was used in the 5V RACE reactions with the gene specific primers RML2 and RML3 at 0.2 AM (Table 1) (all primers were purchased from DNA Technology, Aarhus, Denmark). The reactions were carried out according to the BDk Marathon-Ready cDNA User Manual in the GeneAmp PCR System 2400 (Perkin Elmer) by incubation at 94 8C for 30 s, followed by five cycles of 94 8C for 5 s and 72 8C for 4 min, five cycles of 94 8C for 5 s and 70 8C for 4 min, 25 cycles of 94 8C for 5 s and 68 8C for 4 min, and finally 7 min of extension at 72 8C. Reaction products were separated on a 1.2% agarose gel; bands of interest were excised and the PCR fragments were recovered with QIAEX II Gel Extraction Kit or QIAquick

Table 1 Primers used for RACE-PCR and detection of NCX3 variants Primer

Sequence (5V–3V)

Used ina

Directionb

RML2 RML3 RML6 RML13 RML20 RML22 RML40 RML42 RML51 RML52 RML54 RML68

GGCTAAAGATCCGTGGGCAGTTTCGTG CCAACCCAACATTTCCACTGGCAACTG AGGAAGCAAGAGCGTGGTGAAGCTGAC GGCGGGTCTTCCTGCAGAAGGA GCTGGTGATCTGGGACCTTCTACC GAAGATGCCTGCATGGTCATCATCC CAATGATGACTTCTAGTTTGGGG AAGAAGAGGAGGCCAAGAGG AGCCAGGAGGGTGGATAAGT CTGCTGAAGAAACGGAAGCT TTAGAACCCCTTGATGTAGCAATA CTGCACCATTGGTCTCAAAG

5V RACE PCR, exon 9 5V RACE PCR, exon 9 Southern blot, probe PCR, exon 9 PCR, exon 2 PCR, exon 2 Nested PCR (outer), exon 6 Expression cloning, sequencing, exon 6 PCR/Nested PCR (outer), exon baQ Nested PCR (inner), exon baQ PCR, exon 9c Expression cloning, sequencing, exon 8

R R R R F R R F F F R R

a b c

See also Fig. 5 for location of the different primers. Reverse (R) or forward (F). Sequence from Gabellini et al. (2002).

146

R.-M. Lindgren et al. / Gene 348 (2005) 143–155

Gel Extraction Kit (Qiagen) and eluted with Tris–HCl, 10 mM, pH 8.5. The purified PCR fragments were either ligated into a pGEMR-T Vector (pGEMR-T Vector System II, Promega, Madison, WI) for amplification and sequencing or directly sequenced. 2.3. Sequencing The sequencing reactions were prepared as described in the ABI PRISMR Big Dyek Terminator Cycle Sequencing Ready Reaction Kit protocol (Applied Biosystems). Universal M13 forward and M13 reverse vector sequencing primers and gene specific primers were used. When sequences were obtained, new primers were designed from them and used for extending the sequence. Results were analyzed using the ABI PRISM 377-96 Collection software, Sequence Analysisk, version 3.4 (Applied Biosystems). Sequences were aligned and analyzed using the ABI PRISM Sequence Navigatork software, version 1.0.1, with ABI PRISM Facturak version 1.2.Or6 (Applied Biosystems) and the Sequencherk, version 4.1.2 (Gene Codes Corporation, Ann Arbor, MI). 2.4. Bioinformatics Homology analyses of the cloned sequences, including the alignment of obtained sequences, the 47-1 sequence, the Morton Fetal Cochlea sequence (GenBank accession no. AW022249) and EST database sequences, were performed using the BLAST search of NCBI website (http:// www.ncbi.nlm.nih.gov/blast). The human genomic DNA sequences on chromosome 14 (GenBank accession nos. AL135747 and AL160191) and the human NCX3 gene (SLC8A3) sequence (GenBank accession no. AF508982) (Gabellini et al., 2002) were used to analyze exon–intron junction sequences using programs from EMBL-EBI (European Bioinformatics Institute, http://www.ebi.ac.uk/). Open reading frames (ORF) were identified by the ORF Finder program (http://www.ncbi.nlm.nih.gov/gorf/ gorf.html). Protein similarity searches were performed using the BLAST program (http://www.ncbi.nlm.nih.gov/ BLAST/). SMART (Simple Modular Architecture Research Tools, http://smart.embl-heidelberg.de/) was used to predict protein sequence domains of the deduced protein sequences. The program predicts signal peptide sequences and putative trans-membrane regions. Multiple sequence alignments were generated with Clustal W (1.82) and shaded using BOXSHADE 3.21 (http://www.ch.embnet.org/). Promoter predictions were performed using the NIX (Nucleotide Identification of unknown sequences) program, including the GrailEXP (Grail Experimental Gene Discovery Suite) analysis program at the UK HGMP Resource Centre website (http://www.hgmp.mrc.ac.uk/Bioinformatics/). Transcription factor binding sites were searched for by the MatInspector tool at the Genomatix website (http://www. genomatix.de/index.html).

2.5. PCR for identification of NCX3 cDNA variants The 50 Al reactions contained GeneAmpR 10 PCR Buffer (Applied Biosystems Inc.), 10 mM of each dNTP (Amersham Biosciences, Uppsala, Sweden), 0.25 U AmpliTaq DNA Polymerase (Applied Biosystems Inc.), 10 AM of each primer (Table 1), 8% dimethylsulfoxide and 1.5 Al template. The template was either the cloned and purified PCR fragments (Section 2.2), as controls, or the Human Fetal Brain BDk Marathon-Ready cDNA (BD Biosciences, Clontech, Palo Alto, CA). All reactions were conducted in the GeneAmp PCR System 2400 (Perkin Elmer), started by 2 min at 94 8C, followed by 25 cycles of 94 8C for 30 s, 58 8C for 1 min and 72 8C for 30 s and completed by 7 min at 72 8C for the primers RML51 or RML52 combined with RML40. When primer RML51 was combined with RML54, the cycles were 30, annealing temperature 60 8C and extension time 2.5 min; otherwise, the same as above. Where necessary, PCR products were amplified again, the same way, with similar primers or with one new inner primer, RML52 combined with RML40 (semi-nested PCR) (Table 1). The reaction products were electrophoresed and bands were cut out from the gel, purified and sequenced (Section 2.3). 2.6. Multiple tissue Northern (MTNk) Blots The pre-made hybridization-ready Multiple Tissue Northern (MTNk) Blots (BD Biosciences, Clontech, Palo Alto, CA) used were the Human Fetal MTNk Blot II, including RNA from fetal brain, lung, liver and kidney and the Human MTNk Blot, including RNA from adult heart, whole brain, placenta, lung, liver, skeletal muscle, kidney and pancreas. Purified, cloned and linear PCR fragments, amplified from the Human Fetal Brain Marathon-Readyk cDNA, the purified plasmids containing NCX3 cDNA clones (Section 2.2) or 47-1 (Section 2.1), were used as probes. Probes for exon 1 and exon 2 were designed from the NCX3 gene sequence (GenBank accession no. AF508982), exon 1 (440 bp) at position +69 to +511 and exon 2 (804 bp) at position +853 to +1657. The exon baQ probe covered almost the complete sequence of the newly identified exon baQ from position +1 to +209 (Section 2.5). The 47-1 probe reached from position +2626 to the end of the transcript, all within the predicted exon 9 of the NCX3 gene (GenBank accession no. AF508982). The probes were labeled with [a-32P]dCTP using the Megaprime DNA labelling System (Amersham Biosciences, Buckinghamshire, UK) according to the manufacturer’s protocol. Membranes were hybridized, as described in the manufacturer’s manual, in 5 ml ExpressHybk Hybridization Solution (BD Biosciences, Clontech, Palo Alto, CA) containing denatured probes. Washing was in 2SSC (standard saline citrate, 0.3 M NaCl, 0.03 M sodium citrate), 0.05% SDS (sodium dodecyl sulfate) at RT, followed by 0.1SSC, 0.1% SDS at 50 8C. BIOMAX MR or X-OMAT

R.-M. Lindgren et al. / Gene 348 (2005) 143–155

3. Results 3.1. Characterization of exon baQ and identification of a novel cDNA variant The adaptor-ligated cDNA (Human Fetal Brain BDk Marathon-Ready cDNA) was amplified in the 5V direction with the primers RML2 and RML3, both located in the end of exon 9 of NCX3 and in the 47-1 fragment and about 160 bp away from each other. Two 5V RACE reactions generated two products of about 3 kb and 5 kb (Fig. 2), which were cloned and sequenced. 3VRACE reactions with primers located in the 47-1 fragment resulted in around 300 bp long products, which were not further analyzed (data not shown). The complete sequences of the longest cloned fragments were obtained. The sequences of three clones were identical in their 3V ends, but not in the 5V ends, indicating the presence of different variants. The 2622 bp sequence of the plasmid 10

2.8. Protein expression analysis The DNA fragments with putative open reading frames (ORFs) of the two truncated isoforms NCX3-tN.1 and NCX3-tN.2 were amplified by PCR using the corresponding ligated pGEM-T vectors as templates. The AdvantageR 2 PCR Kit (BD Biosciences) was used in the reactions with the gene specific primers RML51 and NCX3-orf-end (GAACCCCTTGATGTAGCAATAGGC). The reactions were run on a Mastercycle personal (Eppendorf) by incubation at 94 8C for 1 min, followed by 35 cycles of 94 8C for 30 s, 55 8C for 40 s and 72 8C for 2.5 min and finally 7 min of extension at 72 8C. NCX3.2 with full-length ORF was amplified from human fetal brain cDNA using a pair of primers: NCX3-ex2-F1 (forward, TCTCTGGCCTATCAGGAGGA) and NCX3-orf-end. The reaction products were separated on a 1% agarose gel and the bands of correct size were excised and purified with the QIAquick Gel Extraction Kit (Qiagen). The PCR fragments were cloned into the pcDNA3.1/V5-His-TOPO vector (pcDNA3.1/V5-Hisn TOPOR TA Expression Kit, Invitrogen). All constructs were verified by sequencing. For cell culture and transfection, Cos7 cells were grown in DMEM medium containing 10% FBS in 6-well plates. Transient transfections were performed with Lipofectamine

→ →

1 kb DNA Ladder

The Human Fetal MTCk Panel, containing PCR-ready first-strand cDNA from skeletal muscle, brain (ages 21–30 weeks), lung, liver, kidney, heart, thymus and spleen (BD Biosciences, Clontech, Palo Alto, CA), was used as template for PCR. The NCX3-tN.1 and NCX3-tN.2 mRNAs were amplified with the forward primer RML51 in exon baQ and the reverse primer RML54 in exon 9 (Table 1), giving a 1034 bp fragment for NCX3-tN.1 and a 1137 bp for NCX3tN.2. The exon 2 of NCX3 was amplified with the primers RML20 and RML22, both located in exon 2 (Table 1), generating a product of 1112 bp. G3PDH (glyceraldehyde 3phosphate dehydrogenase) primers were used as controls. The 20 Al reactions contained Taq DNA polymerase, with buffer containing 15 mM MgCl2 (both Roche Diagnostics Scandinavia), 4 mM of each dNTP (Invitrogen) and 10 AM of each primer (Table 1). PCR was performed in an Eppendorf PCR machine under the following conditions: 1 min at 94 8C, followed by 38 cycles of denaturation for 20 s at 94 8C, annealing for 20 s at 58 8C (for RML51 and RML54) or at 55 8C (for RML20 and RML22), and extension for 1 min at 72 8C, then 72 8C for 7 min. PCR products were analyzed by 2% agarose gel electrophoresis.

5'-RACE, RML3

2.7. Tissue expression profile of NCX3-tN.1 and NCX3-tN.2 analyzed by RT-PCR

Plus (Invitrogen) with 2 Ag of plasmids with NCX3-tN.1, NCX3-tN.2 and NCX3.2, respectively. Empty vector was used as a negative control. Cells were harvested 24 h after transfection, and polypeptides were separated by 12% SDSPAGE gels followed by Western blotting with the monoclonal antibody against the V5 tag in fusion proteins (Invitrogen). Western blot signals were detected with the enhanced chemiluminescent (ECL) kit (Amersham).

5'-RACE, RML2

AR films (Kodak) were developed after 24 h to 14 days of exposure at 70 8C. Probes were stripped from the blot by washing for 20 min in heated sterile dH2O containing 0.5% SDS.

147

kb

-5.0 -3.0 -2.0 -1.5 -1.0

Fig. 2. RACE PCR using the Human Fetal Brain Marathon-Ready cDNA as template. The PCR products were analyzed by gel electrophoresis in 1.2% agarose with 0.5TBE (Tris–base, boric acid, EDTA) and EtdBr. In the 5V RACE reactions, primers RML2 and RML3 (Table 1) were used. The 1 kb DNA ladder (Promega) was used as a size marker. Southern blot analysis of the 5V RACE products, using RML6 (Table 1) as a probe, indicated that the desired amplified cDNA fragments were present (data not shown). Thus, the two 5V RACE products of lanes 1 and 2 (arrows) were excised and cloned. The cDNAs described in this work correspond in size to the lower of the two products.

148

R.-M. Lindgren et al. / Gene 348 (2005) 143–155 Exon "a" CGAAAACCTAGGGAATGACAGCTAGAGGCATCCAGACGATAACTGGCAGCCAGGAGGGTGGATAAGTCAAAGGAAGGGGTCAAGGAAAGAGGGGAAGGAAAGGGAACCATCACTTGCTGA 120

4 GCCTGCTGCCTGTGCTTCCTCATGTCACCCGCACGACAACCCAATGTGAATGTTATCATCTCCAGGTAACTGCTGAAGAAACGGAAGCTCAAAGAGGAAAACCATAAGGGTTAAAATAGT 240

6 AGATGAGGAGGAATACGAAAGGCAAGAGAATTTCTTCATTGCCCTTGGTGAACCGAAATGGATGGAACGTGGAATATCAGATGTGACAGACAGGAAGCTGACTATGGAAGAAGAGGAGGC 360 2) M E R G I S D V T D R K L T 1)M E E E E A

7 CAAGAGGATAGCAGAGATGGGAAAGCCAGTATTGGGTGAACACCCCAAACTAGAAGTCATCATTGAAGAGTCCTATGAGTTCAAGACTACGGTGGACAAACTGATCAAGAAGACAAACCT 480 K R I A E M G K P V L G E H P K L E V I I E E S Y E F K T T V D K L I K K T N L

8 GGCCTTGGTTGTGGGGACCCATTCCTGGAGGGACCAGTTCATGGAGGCCATCACCGTCAGTGCAGCAGGGGATGAGGATGAGGATGAATCCGGGGAGGAGAGGCTGCCCTCCTGCTTTGA 600 A L V V G T H S W R D Q F M E A I T V S A A G D E D E D E S G E E R L P S C F D CTACGTCATGCACTTCCTGACTGTCTTCTGGAAGGTGCTGTTTGCCTGTGTGCCCCCCACAGAGTACTGCCACGGCTGGGCCTGCTTCGCCGTCTCCATCCTCATCATTGGCATGCTCAC 720 Y V M H F L T V F W K V L F A C V P P T E Y C H G W A C F A V S I L I I G M L T

9

CGCCATCATTGGGGACCTGGCCTCGCACTTCGGCTGCACCATTGGTCTCAAAGATTCAGTCACAGCTGTTGTTTTCGTGGCATTTGGCACCTCTGTCCCAGATACGTTTGCCAGCAAAGC 840 A I I G D L A S H F G C T I G L K D S V T A V V F V A F G T S V P D T F A S K A TGCTGCCCTCCAGGATGTATATGCAGACGCCTCCATTGGCAACGTGACGGGCAGCAACGCCGTCAATGTCTTCCTGGGCATCGGCCTGGCCTGGTCCGTGGCCGCCATCTACTGGGCTCT 960 A A L Q D V Y A D A S I G N V T G S N A V N V F L G I G L A W S V A A I Y W A L GCAGGGACAGGAGTTCCACGTGTCGGCCGGCACACTGGCCTTCTCCGTCACCCTCTTCACCATCTTTGCATTTGTCTGCATCAGCGTGCTCTTGTACCGAAGGCGGCCGCACCTGGGAGG 1080 Q G Q E F H V S A G T L A F S V T L F T I F A F V C I S V L L Y R R R P H L G G GGAGCTTGGTGGCCCCCGTGGCTGCAAGCTCGCCACAACATGGCTCTTTGTGAGCCTGTGGCTCCTCTACATACTCTTTGCCACACTAGAGGCCTATTGCTACATCAAGGGGTTCTAAGC 1200 E L G G P R G C K L A T T W L F V S L W L L Y I L F A T L E A Y C Y I K G F * CACACAACAGAGCCTCCAGCAGGGCAGGCCTAGGACTTCTCCTAAGAGAAGGGCACTTCCCCACCAGTGATCTCTCCCGACTGCACTGCCCTGGAGAGGCAGCATCAGGACCTAAGCCCC 1320 AGGAACTTCACCCAACTTAGGCCCTGGCAATTAACTGAAAGGGCAAAGTCTTAATCAATCAAACAATGGAGGAATCACCGACTTTACACAGTATTTAATTGAATACAAACAAGCAACAGC 1440 AACAAATCCACCTCCACCCCATCTCCCCCTCATATCCCTGACCCAAAGCAAAGGTCAGAGCCTTTCGCCTCCTTCTATTCCATCTTTTGATTATTCCTTTGCCTCTCATTTCTTTGGAAG 1560 CAGGGTTTCTCCTCTCTGCCCAATTCCATATGTCCCTATTATCTCACTCAGCTGACAAGACGTGAAAATGAGTCACATTCATGTGGCTGGGGTGGGGTTCTTTTTTCATTGTAATCATTA 1680 TTGTGGTTGCTTTCGTTTTGCCATTAGGTTTTGCTTATTATTTTGTTTTGTCTTTTTTTTCTGAAGTGGGTGAAAAAGGTGCCACAAAGGAATTCCAGGTCCGAGCCACCAGAGAGAAAC 1800 ATGAATTTTTAGACACATGCTCTCCTGCCACCTCTTGGCTCCATCAAGATCCAGTTCCCCATCTCACTGTTTTCTCTGAGTTCTTGGGAGGAGTGATGGTGTTGGGGTAGAAATAAGCTC 1920 ACTCACCCACGCAGGGTACTAAAGATCTTACAGGAGCTTCAACTGGAGCAGGAGGAGCTTTTTATGCTTATGTTGAATCAAGTCAGATACAAAAAGCAATTGTCCCTCTTTGCCCAAGCC 2040 TTTCCAATTCTGTGTGTCTTGTTGTGTCAGTGTCCACTTGTGTATCCTTCTGCAGGAAGACCCGCCAAATAGAAGAGATGGGACAAAAATAGGAATGGTGTGTGACGACAAAGGGCTACT 2160 GGAAGAACAAAAGGGATACAGGCCTTCTTGATTATCTTTGGCTTTGTACCTGAGGCAGGAGAGAAGAGATGTCCAACCAGTGAGATCTTTAAGAGAAAAGTTTGTATTTTAAATGTCAAT 2280 GTGCCTGAGAAATGTCAGCTTCACCACGCTCTTGCTTCCTAATGCTCTATACAAAGAGGGCTGACTATATTTCTTGAAGTGGTGTAAAAACTTAGAGATTTTATAAGAGAACCAGGGGCT 2400 CCCTTCACCTCTCCTGGTCCCTCAGGTCACATATGAAAGCATTTTTACAAGATAGGAACTGGAATTCCTCATTTCTCCCATGTTCCTGCTTGTTCTTAAACTTCATGAAGCTATTTTTCC 2520 AGTCTATGGGGTAGTTCTTGCTCCAGTAAGAGGAATCTTAGTTGTCATAATCCCTTGGAGCCTGGGTTTTTGGAGAAAGAGATCTCCGTGCCCTACAGACCTTTTCTCAACGAATGTGGG 2640 AAGGACCTGGCTTTAAAACACGCGCACAAACACACAAATAACCAGGCATAAGATGTCATCGCGAAACTGAACACGGATCTTTAGGC

2726

Fig. 3. Nucleotide and predicted amino acid sequences of NCX3-tN.2. The NCX3-tN.1 splice variant (plasmid 10) has the same predicted sequence except that exon 4 is missing. Exon baQ in the 5V end is marked with a square and is followed by exon 4 (in bold type). The numbered arrows mark starting-points for the predicted exons 4, 6, 7, 8 and 9. Potential translational start sites are underlined for (1) NCX3-tN.1 and (2) NCX3-tN.2. The predicted stop codon is marked with *.

was the most interesting one, containing an ORF of 843 nucleotides (Fig. 3). This ORF was predicted to code for a protein of 285 aa. In the C-terminal part, it was found to be identical to the exons 6–9 of the human NCX3 gene (SLC8A3) (GenBank accession no. AF508982) (Gabellini et al., 2002). In the 5V end, it contained a previously

undescribed exon, here named exon baQ. New RACE reactions were performed, with primers in the 5V end of the exon baQ, to make sure that the true 5V end was reached. These reactions generated no clear products (data not shown), indicating that the actual 5V end of the transcript had been reached. The 216 bases long exon baQ showed

Table 2 Intron–exon boundaries of two N-terminally truncated variants of human NCX3 Exons

Exon size (bp)

a 4 6 7 8 9

216 104 125 100 276 2068

Intron site

5V splice site

3V splice site

Intron site

. . .tccctacag . . .ttgttgcag . . .tcttaacag . . .tgccttcag . . .cttttccag

CGAAAACCT. . . GAAAACCAT. . . ATGTGACAG. . . ACTACGGTG. . . CAGGGGATG. . . ATACGTTTG. . .

. . .CTCAAAGAG . . .GAATATCAG . . .GAGTTCAAG . . .TCAGTGCAG . . .CTGTCCCAG

gtaagagat. . . gtgtgagat. . . gtcaggcaa. . . gtgagaagt. . . gtgagagtg. . .

The information is based on a comparison of the cDNA sequences of NCX3-tN.1 and NCX3-tN.2 with human SLC8A3 gene DNA sequence (GenBank accession no. AF508982). NCX3-tN.2 contains exon 4, but NCX3-tN.1 does not.

0.5

1.0 -

-

149

RML51 and RML54

kb

1 kb DNA Ladder

kb

b

RML52 and RML40

100 bp DNA Ladder

a

RML51 and RML40

R.-M. Lindgren et al. / Gene 348 (2005) 143–155

1.5

←NCX3-tN.2 ←NCX3-tN.1 ←NCX3-tN.2

0.5

←NCX3-tN.2 ←NCX3-tN.1

-

←NCX3-tN.1 0.1

-

Fig. 4. PCR analysis with Human Fetal Brain Marathon-Ready cDNA as template identifying two different truncated NCX3 variants. (a) PCR was first conducted with primers RML51, located in exon baQ and RML40, located in exon 6. The product from this PCR was used as template in a semi-nested PCR, also shown in (a), with RML52 located in exon 6. Primers are indicated above the lanes. The expected lengths of the generated PCR products are, for primer RML51 combined with RML40, 276 bp (NCX3-tN.1, without exon 4) and 379 bp (NCX3-tN.2 with exon 4) and for primer RML52 combined with primer RML40, 136 and 239 bp, respectively. The 100 bp DNA Ladder (New England BioLabs) was used as a size marker. In (b), the primers used were RML51, located in exon baQ together with RML54, located downstream of the stop codon in exon 9. These products include the whole expected ORFs, with expected sizes of 1034 bp (NCX3-tN.1, excluding exon 4) and 1137 bp (NCX3-tN.2, including exon 4). The 1 kb DNA Ladder (New England BioLabs) was used. The PCR products were analyzed by gel electrophoresis in 1.5% (a) and 1.2% (b) agarose, with 0.5TBE and EtdBr. All the PCR fragments were cut from the gel and their identity, including presence or not of exon 4, was confirmed by sequencing.

homology to the NCX3 gene sequence (Gabellini et al., 2002), but not to any known mRNA sequences. The exon baQ is located in the almost 103 kb long intron 2 (location 109159–109374 in the gene, GenBank accession no. AF508982). In comparison to the known transcripts of the human NCX3 gene, the truncated transcripts isolated here lacked a part of the polyadenylation sequence of the

Exons 1

10kb

estimated 3V end. If these 202 missing nucleotides are added to the predicted plasmid 10 transcript, a longer transcript of 2824 bp will be the result. The intron–exon boundaries of the putative exon baQ show similarity to the most common sequences described for intron–exon boundaries in eucaryotic genes (Mount, 1982), and also to the boundaries of the exons already

2

a

3 4 5 67 8 9

→RML20 ←RML22

NCX3-tN.2

a

4

6

7

8

3'UTR

5'UTR

NCX3-tN.1

9

a

6

7

5'UTR ←RML40 →RML51 →RML52 →RML42

8

9 3'UTR ←RML54 ←RML13 ←RML68 ←RML6 ←RML2 ←RML3

Fig. 5. Organization of the NCX3 gene and the N-terminally truncated variant forms NCX3-tN.1 and NCX3-tN.2. The location of the novel exon baQ is marked and included in the exon/intron organization of the gene in the top panel. The two N-terminally truncated mRNAs produced in the human fetal brain are shown below. The predicted ORF regions are shown in gray. Arrows mark the locations of the different PCR primers used.

150

R.-M. Lindgren et al. / Gene 348 (2005) 143–155

known in the NCX3 gene. The intron following exon baQ starts with GT and ends with AG, which is according to the consensus (Table 2). Two potential translational start sites (ATG) are marked in Fig. 3 and, since there are two truncated transcripts (Section

a

2

3.3), these two could use separate start sites, or both the same. The start signal marked with (1) appears to be the better candidate since it has an A in the 3 position, followed by a C and also a G in the +4 position. Sequences that enhance initiation of translation have been described

signal peptide

NCX3-tN.1 NCX3.2 NCX3-tN.2

1 ---------------------------------------------------------------------1 MAWLRLQPLTSAFLHFGLVTFVLFLNGLRAEAGGSGDVPSTGQNNESCSGSSDCKEGVILPIWYPENPSL 1 ----------------------------------------------------------------------

NCX3-tN.1 NCX3.2 NCX3-tN.2

1 ---------------------------------------------------------------------71 GDKIARVIVYFVALIYMFLGVSIIADRFMASIEVITSQEREVTIKKPNGETSTTTIRVWNETVSNLTLMA 1 ----------------------------------------------------------------------

NCX3-tN.1 NCX3.2 NCX3-tN.2

1 ---------------------------------------------------------------------141 LGSSAPEILLSLIEVCGHGFIAGDLGPSTIVGSAAFNMFIIIGICVYVIPDGETRKIKHLRVFFITAAWS 1 ----------------------------------------------------------------------

NCX3-tN.1 NCX3.2 NCX3-tN.2

1 ---------------------------------------------------------------------211 IFAYIWLYMILAVFSPGVVQVWEGLLTLFFFPVCVLLAWVADKRLLFYKYMHKKYRTDKHRGIIIETEGD 1 ----------------------------------------------------------------------

NCX3-tN.1 NCX3.2 NCX3-tN.2

1 ---------------------------------------------------------------------281 HPKGIEMDGKMMNSHFLDGNLVPLEGKEVDESRREMIRILKDLKQKHPEKDLDQLVEMANYYALSHQQKS 1 ----------------------------------------------------------------------

NCX3-tN.1 NCX3.2 NCX3-tN.2

1 ---------------------------------------------------------------------351 RAFYRIQATRMMTGAGNILKKHAAEQAKKASSMSEVHTDEPEDFISKVFFDPCSYQCLENCGAVLLTVVR 1 ----------------------------------------------------------------------

NCX3-tN.1 NCX3.2 NCX3-tN.2

1 ---------------------------------------------------------------------421 KGGDMSKTMYVDYKTEDGSANAGADYEFTEGTVVLKPGETQKEFSVGIIDDDIFEEDEHFFVRLSNVRIE 1 ----------------------------------------------------------------------

NCX3-tN.1 NCX3.2 NCX3-tN.2

1 ---------------------------------------------------------------------491 EEQPEEGMPPAIFNSLPLPRAVLASPCVATVTILDDDHAGIFTFECDTIHVSESIGVMEVKVLRTSGARG 1 ----------------------------------------------------------------------

NCX3-tN.1 NCX3.2 NCX3-tN.2

β-2 repeat 4 6 1 ---------------------------------------------------------------------561 TVIVPFRTVEGTAKGGGEDFEDTYGELEFKNDETVKTIRVKIVDEEEYERQENFFIALGEPKWM ER GI SD 1 ---------------------------------------------------------------M ER GI SD 7

TM1

TM2

α-1 repeat

TM4

TM3

TM5

XIP

β-1 repeat

NCX3-tN.1 NCX3.2 NCX3-tN.2

1 ------- ME EEE AK R IA E M GK P V LG EH PK LE V I I EE SY E FK TT VD K LI KK TN L ALVV G T HS WRD QF ME AI 631 VT DRKLT ME EEE AK R IA E M GK P V LG EH PK LE V I I EE SY E FK TT VD K LI KK TN L ALVV G T HS WRD QF ME AI 8 VT DRKLT ME EEE AK R IA E M GK P V LG EH PK LE V I I EE SY E FK TT VD K LI KK TN L ALVV G T HS WRD QF ME AI 8

NCX3-tN.1 NCX3.2 NCX3-tN.2

64 TV SAAGD ED EDE SG E ER L P SC F D YV MH FL TV F W K VL FA C VP PT EY C HG WA CF A VSIL I I GM LTA II GD LA 701 TV SAAGD ED EDE SG E ER L P SC F D YV MH FL TV F W K VL FA C VP PT EY C HG WA CF A VSIL I I GM LTA II GD LA 78 TV SAAGD ED EDE SG E ER L P SC F D YV MH FL TV F W K VL FA C VP PT EY C HG WA CF A VSIL I I GM LTA II GD LA α-2 repeat 9

NCX3-tN.1 NCX3.2 NCX3-tN.2

134 SH FGCTI GL KDS VT A VV F V AF G T SV PD TF AS K A A AL QD V YA DA SI G NV TG SN A VNVF L G IG LAW SV AA IY 771 SH FGCTI GL KDS VT A VV F V AF G T SV PD TF AS K A A AL QD V YA DA SI G NV TG SN A VNVF L G IG LAW SV AA IY 148 SH FGCTI GL KDS VT A VV F V AF G T SV PD TF AS K A A AL QD V YA DA SI G NV TG SN A VNVF L G IG LAW SV AA IY

NCX3-tN.1 NCX3.2 NCX3-tN.2

204 WA LQGQE FH VSA GT L AF S V TL F T IF AF VC IS V L L YR RR P HL GG EL G GP RG CK L ATTW L F VS LWL LY IL FA 841 WA LQGQE FH VSA GT L AF S V TL F T IF AF VC IS V L L YR RR P HL GG EL G GP RG CK L ATTW L F VS LWL LY IL FA 218 WA LQGQE FH VSA GT L AF S V TL F T IF AF VC IS V L L YR RR P HL GG EL G GP RG CK L ATTW L F VS LWL LY IL FA

NCX3-tN.1 N CX3.2 N CX3-tN.2

274 TL EAYCY IK GF 911 TL EAYCY IK GF 288 TL EAYCY IK GF

TM6

TM7

TM8

TM9

Fig. 6. (a) The amino acid sequence of the NCX3.2 (Gabellini et al., 2002) isoform is aligned with the sequences of the novel truncated variants NCX3-tN.1 and NCX3-tN.2, shown shaded. The positions of the predicted signal peptide, the putative trans-membrane (TM) regions and the endogenous exchanger inhibitory peptide (XIP) are marked. The h-1 and h-2 repeat domains and the regions corresponding to the a-1 and a-2 repeats are boxed. Numbering refers to the alignment positions. The arrows with numbers mark starting-points for the predicted exons 2, 4, 6, 7, 8 and 9. (b) The topographical structure of the bfulllengthQ NCX3 and the NCX3-tN.1 and NCX3-tN.2 as predicted by bioinformatics and following (Gabellini et al., 2002). AltSplice: commonly differentially spliced region in NCXs, Out: extracellular, In: intracellular, Ca: calcium-regulated region.

R.-M. Lindgren et al. / Gene 348 (2005) 143–155

151

b α−1

N

Out

1

2

3

4

5

6

7

8

9

NCX3 XIP

In α−2

β−1

β−2

Ca

C

AltSplice

Out

6

7

8

9

NCX3-tN.1/2 In α−2

C

Fig. 6 (continued).

and there is usually A or G at 3, often followed by two Cs (A/GCCAUG). Efficient initiation has also been shown to be promoted by a G following the initiation codon at position +4 (Kozak, 2002). We also noticed that the full-length of NCX3 uses a translation start signal with a G in the +4 position (CGTGTAUGG). Comparing this with both truncated NCX3 isoforms presented in this manuscript, it is worth to mention that all the three isoforms have a G in the +4 position, but are different in the 3 positions. 3.2. Identification of cDNA variants containing exon baQ To further confirm the presence of an NCX3 variant starting with exon baQ, different PCRs were performed. The primers were chosen so that the 5V primers were located in the exon baQ sequence and the 3V primers in exon 6 (Fig. 4a) or in the vicinity of the stop codon in exon 9 (Fig. 4b) (Table 1). By PCR, followed by sequencing of PCR products, we found two variants present in the human fetal brain cDNA. The first variant, of about 2.8 kb, designated as NCX3-tN.1, included exon baQ, followed by exons 6, 7, 8 and 9, while the second variant (NCX3-tN.2), of 2.9 kb, included also exon 4 (Fig. 4a and b). The novel variants are designated as NCX3-tN.1 and NCX3-tN.2 because the putative proteins correspond to NCX3 truncated in its amino terminal part. The truncated forms were verified by a number of different PCRs. In addition to those presented in Fig. 4, other primer combinations were used, e.g. RML51 combined with RML2 or RML13, both located in exon 9 (Table 1). The presence

of two differentially sized transcripts was thus confirmed (data not shown). 3.3. Predicted structural features of the new NCX3 variants Fig. 5 shows the organization of the gene, with the two alternative transcripts, NCX3-tN.1 and NCX3-tN.2, both containing exon baQ, in comparison to the described organization of introns and exons in NCX3.2 (Gabellini et al., 2002). The expected structural features of the putative proteins encoded by the truncated transcripts were defined by bioinformatic analysis (Fig. 6a). The aa sequence of NCX3.2 was aligned to the predicted sequences of NCX3-tN.1 and NCX3-tN.2 by the programs Clustal W (1.82) and BOXSHADE. The positions of the predicted signal peptide and the putative TM regions of NCX3.2 were marked as in Gabellini et al. (2002). The topological structure of the two truncated transcripts was compared to the structure of NCX3 (Fig. 6b) and the two variants are predicted to consist of four trans-membrane regions (TM6– TM9) while NCX3.2 has all the nine trans-membrane motifs (TM1–TM9) (Gabellini et al., 2002). Worth to be noted is also that the h-repeat domains (Philipson and Nicoll, 2000; Gabellini et al., 2002) were missing in the truncated transcripts. Furthermore, the truncated variants are expected to include only the a-2 repeat, but not the a-1 (Schwarz and Benzer, 1997), and to lack the endogenous XIP region (Li et al., 1991).

4.4 -

Kidney

Liver

Lung

Brain

Kidney

Liver

Lung

Brain

Kidney

Liver

Brain

Liver

kb

Lung

Brain

a

Lung

R.-M. Lindgren et al. / Gene 348 (2005) 143–155

Kidney

152

7.5

⇐

Negative control

47-1

Spleen

Thymus

Heart

Liver

kb

Lung

b

Exon "a"

Brain

Exon 2

Skeletal muscle

Exon 1

Kidney

-

100 bp DNA Ladder

2.4

⇐

←

1.0 0.5 -

1.0 0.5 -

1.0 0.5 -

Fig. 7. Expression of NCX3 variants in fetal tissues. (a) The pre made hybridization-ready Multiple Tissue Northern (MTNk) Blot, Human Fetal MTNk Blot II was used. The blot was hybridized with radioactive-labeled probes specific for the first and second exons of NCX3 (exon 1 and exon 2, respectively), the new transcript (exon baQ) and for part of exon 9 (47-1). The various tissues on the blot are indicated. The different hybridizations were exposed at 70 8C for 14 days (exon 1 and exon 2) and 4 days (exon baQ and 47-1). Probes were stripped from the membrane after development and the same membrane was used for all four hybridizations. The 3 kb transcript corresponding to the truncated NCX3 variants found in fetal brain are marked with o, the longer bfull-lengthQ NCX3 transcript with and the unknown transcript in fetal liver with p. In (b), the tissue expression patterns of NCX3-tN.1 and NCX3-tN.2 mRNAs were studied by PCR, using the Human Fetal MTCk Panel as template. Products were analyzed by gel electrophoresis in 2% agarose. PCR of NCX3-tN.1 and NCX3-tN.2 mRNAs, with primers RML51 and RML54 (top), of NCX3 with exon 2 specific primers (middle) and of G3PDH as control (bottom) are shown. Negative control is with dH2O instead of cDNA template. Expected size of the NCX3-tN.1 PCR product is 1034 bp, and of NCX3-tN.2 1137 bp. The exon 2 PCR product is 1112 bp and the G3PDH product 1000 bp. The 100 bp DNA Ladder (Invitrogen) was used.

Empty vector

NCX3.2/V5-His

kDa

NCX3-tN.2/V5-His

The Human Fetal MTN Blot II and the Human MTN Blot were probed with exon 1, exon 2, exon baQ and 47-1 cDNA probes (Fig. 7a). The 47-1 fragment corresponds to exon 9 of the exchanger. In fetal brain, but not in adult brain (data not shown), the probes for exon baQ and 47-1 recognized a transcript of about 3 kb (Fig. 7a), which corresponds to the estimated size of the truncated NCX3 variants (NCX3-tN.2, 2.9 kb with exon 4 and NCX3-tN.1, 2.8 kb without exon 4). Exon 1 and exon 2 are not included in the new variants and, as expected, the c3 kb transcript could not be visualized with the corresponding probes. The c3 kb transcript was not detected in any of the adult tissues analyzed, i.e. brain (whole), heart, placenta, lung, liver, skeletal muscle, kidney and pancreas (data not shown). In fetal brain, exon 1, exon 2 and 47-1 probes all hybridized to a longer transcript, of about 6.0 kb. This corresponds to the previously described bfull-lengthQ isoforms of the NCX3 (including most of the exons). Exon baQ is, as shown above, not expected to be included in these longer variants and consistent with this, the exon baQ probe did not recognize the 6.0 kb transcript (Fig. 7a). Among the adult tissues analyzed by Northern blot, the 6.0 kb transcript was clearly visible only in skeletal muscle, not evident in brain (data not shown). In fetal liver, a transcript of about 2.4 kb could be seen, recognized only by the exon 1 probe (Fig. 7a). The identity of this transcript was not further addressed. In summary, an exon baQ, containing transcript, approximately 3 kb in size, was visualized in fetal brain, but not in other fetal or adult tissues analyzed. It should be noted that fetal muscle was not included in the analysis. Moreover, it is not known whether the signal represents one or both truncated transcripts (NCX3-tN.1 and NCX3-tN.2), as the two variants are too similar in size to be discriminated in the Northern blot analysis. To further characterize the expression pattern of NCX3tN.1 and NCX3-tN.2 in fetal tissues, a multiple tissue cDNA panel was used to amplify mRNA by RT-PCR strategy, with primers RML51 and RML54 (Table 1). The expression patterns of NCX3-tN.1 (1034 bp) and NCX3tN.2 (1137 bp) in fetal tissues are shown in Fig. 7b (top), with exon 2 of NCX3 (1112 bp) (middle) and G3PDH (1000 bp) (bottom) as controls. Among fetal tissues, the two PCR products representing NCX3-tN.1 and NCX3-tN.2 mRNAs were only detected in brain (Fig. 7b, top). They were not expressed in skeletal muscle or in heart. When using the sensitive PCR method, the exon 2 that represents bfull-lengthQ NCX3 isoforms was detected in skeletal muscle, lung, liver and heart in addition to its strong signal in brain (Fig. 7b, middle). Based on the Northern blots and RT-PCR results, we concluded that NCX3-tN.1 and NCX3tN.2 transcripts are brain-specific among analyzed fetal tissues.

NCX3-tN.1/V5-His

3.4. Tissue distribution of NCX3-tN.1 and NCX3-tN.2 transcripts

153

Molecular size marker

R.-M. Lindgren et al. / Gene 348 (2005) 143–155

30-

220 120 80 60 50 40

-

20

Fig. 8. Fused expression of NCX3-tN.1 and NCX3-tN.2 in Cos7 cells. DNA fragments with the ORFs of NCX3-tN.1, NCX3-tN.2 and NCX3.2 were cloned into expression vector pcDNA3.1/V5-His. Resulting plasmids were transfected into Cos7 cells and expression products were examined 24 h after transfection by Western blotting using a monoclonal antibody against V5 tag at the C-terminal end of the fusion proteins.

3.5. Protein expression analysis To determine whether the two truncated transcripts, NCX3-tN.1 and NCX3-tN.2, could produce polypeptides, the putative ORFs of the splicing variants were cloned into the expression vector pcDNA3.1/V5-His-TOPO. Expected inserts were confirmed by sequencing. The expression vectors were transfected into Cos7 cells and products were analyzed by Western blotting with a monoclonal antibody against V5 tag at the C-terminal part of the fusion proteins. The empty vector and NCX3.2-pcDNA3.1/V5-His were used as negative and positive control, respectively. The predicted molecular weights of the fusion proteins for NCX3-tN.1/pcDNA3.1/V5 and NCX3-tN.2/pcDNA3.1/V5His are 36 and 37.6 kDa, respectively. As shown in Fig. 8, it is clear that both the two truncated transcripts, NCX3-tN.1 and NCX3-tN.2, were able to produce polypeptides of 30– 40 kDa, but translation of NCX3-tN.1 was, apparently, more efficient than NCX3-tN.2.

4. Discussion In the present study, we describe the cloning and characterization of two novel alternative NCX3 transcripts from human fetal brain cDNA, encoding two truncated NCX3 variants. The truncated transcripts were designated as NCX3-tN.1 and NCX3-tN.2, because the putative encoded

154

R.-M. Lindgren et al. / Gene 348 (2005) 143–155

proteins (284 and 298 aa, respectively) correspond to NCX3 truncated in its N-terminal portion. The two variants are about a third of the bfull-lengthQ NCX3, represented by the NCX3.2, also expressed in brain (Gabellini et al., 2002). The truncated forms lack the putative N-terminal signal sequence, the following five TMs, and most of the large central cytoplasmic loop. As a consequence, they are composed of a small C-terminal portion of the large cytosolic loop, not including any h-repeat domains, and the C-terminal set of hydrophobic domains with four putative TMs and the a-2 repeat. Bioinformatic analysis revealed that these two novel transcripts likely result from the use of an alternative gene promoter of the NCX3 gene. Initiation of transcription of NCX3-tN.1 and NCX3-tN.2 occurs in intron 2 of the NCX3 gene, with a novel identical initial exon, here named exon baQ. Although, if the complete exon baQ has been found, remains to be seen. The size of the transcripts found by Northern blot compared to the size of the RACE PCR products implies that the 5V end of the transcript has been reached. Further, during this work, a full-length cDNA sequence (GenBank accession no. AK122728), corresponding to NCX3-tN.2 in this work, but with a 24 bp longer exon baQ at the 5V end has recently been released (Ota et al., 2004). NCX3-tN.1 consists of five exons including exon baQ and exons 6–9 of the original gene organization, while NCX3-tN.2 also includes exon 4. The exon 4 codes for the C-terminal end of the h-2 repeat, but the expected translational start site of NCX3-tN.2 is located just C-terminal of the corresponding genomic sequence (cf. Figs. 3, 5 and 6). Cloning of the two cDNAs into expression vectors and transient transfection into Cos7 cells confirmed that both could be translated into proteins. NCX3-tN.2 is identical to NCX3-tN.1 except that it is expected to be 14 aa longer at its N-terminal. Their deduced amino acid sequences are 100% identical to the C-terminal portion of NCX3. The differing 14 aa could have an unknown but potentially important function, since they are located in the 110 aa variable region in the intracellular loop, where alternative splicing often takes place in NCX genes (Kofuji et al., 1994; Quednau et al., 1997). Although all three members of the NCX family and their alternative variants generally contain two sets of hydrophobic domains with several TM fragments separated by the large intracellular loop, at least two C-terminally truncated variants have been identified (NCX1.33 and NCX3.4) (Fig. 1 and not shown). These two have a similar topology consisting of only five N-terminal TM fragments and a shortened hydrophilic loop (Van Eylen et al., 2001; Gabellini et al., 2002). However, the truncated isoforms of NCX3 lacking the N-terminal hydrophobic domain and about 2/3 of the large cytoplasmic loop as revealed in this study (Fig. 1) are not previously described for any of the three mammalian members of the NCX family. Two regions of hydrophobic domains, in the N- and Cterminal halves, respectively, of the NCX contain an arepeat, that plays an important role in the ion binding and

transport (Nicoll et al., 1996a; Schwarz and Benzer, 1997). Several studies have shown that the truncated isoform of NCX1, lacking the hydrophobic domain of the C-terminal half, retain the capacity to transport ions (Gabellini et al., 1995; Nicoll et al., 1996a; Li and Lytton, 1999; Van Eylen et al., 2001). Moreover, active truncated isoforms could form dimers, because of the homology between the two a-repeat sequences (Gabellini et al., 1996). A similar topological structure as in the C-terminally truncated exchanger is deduced for the N-terminally truncated forms, implying that NCX3-tN.1 and NCX3-tN.2 are likely to be functionally active (Fig. 6b). Further, the two novel variants described here lack about 2/3 of the large intracellular loop, including the h-1 and the h-2 repeat. The h-repeats are considered to be involved in functional regulation of the exchanger (Nicoll et al., 1999), implying that the regulation of these two new variants could be different or abolished. One result of the differential splicing is that NCX3-tN.1 and NCX3-tN.2 transcripts include their own individual potential translational start sites and unique 5V UTRs that are different from those of the known NCX3 variants. Protein expression studies showed that NCX3-tN.1 was more efficiently translated into protein than NCX3-tN.2 (Fig. 8). This finding is consistent with the prediction from mRNA sequence that (1) is a better translational start site than (2) (cf. Fig. 3 and Kozak, 2002). The distribution of NCX3-specific transcripts is highly tissue specific (brain and skeletal muscle) (Nicoll et al., 1996b). In rat brain tissue, cell-specific expression of the three NCX subtypes has been seen in most areas of the brain (Thurneysen et al., 2002). In hippocampus, the expression of NCX3 was restricted to a small subpopulation of neuronal cells and NCX3 was not at all detected in glia cells, while NCX2 expression was reported to be specific for various types of glia cells (Thurneysen et al., 2002). Also the NCX3-tN.1 and NCX3-tN.2 described here seem to be subject to developmental regulation in the brain. The explanation for this might be found in their unique 5V UTRs. By the NIX program, including the GrailEXP, a promoter was predicted upstream of the exon baQ, with a good quality of prediction and the score=77 (data not shown). The predicted promoter was approximately 370 bp. A TATA box was identified and Clustal W alignment of human, mouse and rat sequences of NCX3 gene showed a high degree of conservation in this region. Thus, the bioinformatics analyses indicate that the truncated transcripts have their own promoter, although the function of this promoter needs to be proven experimentally. Alternative splicing affecting gene regulation has been described for a number of trans-membrane proteins, e.g. variable regions of NCX1 and NCX3 in neonatal compared to adult brain and skeletal muscle (in rat tissues) (Quednau et al., 1997). In conclusion, we have cloned and characterized two new N-terminally truncated NCX3 transcript variants that are identical to the C-terminal portion of NCX3, and encode a

R.-M. Lindgren et al. / Gene 348 (2005) 143–155

small part of the large cytoplasmic loop, four C-terminal TMs and in addition contain a novel exon at the 5V end (exon baQ). The NCX3-tN.2 is predicted to be 14 aa longer than NCX3-tN.1 at its N-terminus. Among different fetal tissues, the truncated variants are specifically expressed in the brain. Further studies will be necessary to determine the cellular localization and function of these novel proteins, as well as to determine how their tissue specific expression is regulated.

Acknowledgements We thank Mrs. Marianne Kastemar for excellent technical assistance and Dr. Erik Bongcam-Rudloff and Dr. Lucia Cavelier for valuable bioinformatic advice. This work was supported by grants from the Swedish Cancer Foundation (MN), the Cancer Society in Stockholm (MN), Karolinska Institutet (MN) and the Swedish Society of Medicine (SH).

References Barnes, K.V., Cheng, G., Dawson, M.M., Menick, D.R., 1997. Cloning of cardiac, kidney, and brain promoters of the feline ncx1 gene. J. Biol. Chem. 272, 11510 – 11517. Chomczynski, P., Sacchi, N., 1987. Single-step method of RNA isolation by acid guanidinium thiocyanate–phenol–chloroform extraction. Anal. Biochem. 162, 156 – 159. Furman, I., Cook, O., Kasir, J., Rahamimoff, H., 1993. Cloning of two isoforms of the rat brain Na(+)–Ca2+ exchanger gene and their functional expression in HeLa cells. FEBS Lett. 319, 105 – 109. Gabellini, N., Iwata, T., Carafoli, E., 1995. An alternative splicing site modifies the carboxyl-terminal trans-membrane domains of the Na+/ Ca2+ exchanger. J. Biol. Chem. 270, 6917 – 6924. Gabellini, N., Zatti, A., Rispoli, G., Navangione, A., Carafoli, E., 1996. Expression of an active Na+/Ca2+ exchanger isoform lacking the six C-terminal transmembrane segments. Eur. J. Biochem. 239, 897 – 904. Gabellini, N., Bortoluzzi, S., Danieli, G.A., Carafoli, E., 2002. The human SLC8A3 gene and the tissue-specific Na+/Ca2+ exchanger 3 isoforms. Gene 298, 1 – 7. Kikuno, R., et al., 1999. Prediction of the coding sequences of unidentified human genes: XIV. The complete sequences of 100 new cDNA clones from brain which code for large proteins in vitro. DNA Res. 6, 197 – 205. Kofuji, P., Lederer, W.J., Schulze, D.H., 1994. Mutually exclusive and cassette exons underlie alternatively spliced isoforms of the Na/Ca exchanger. J. Biol. Chem. 269, 5145 – 5149. Kozak, M., 2002. Pushing the limits of the scanning mechanism for initiation of translation. Gene 299, 1 – 34. Kraev, A., Chumakov, I., Carafoli, E., 1996. The organization of the human gene NCX1 encoding the sodium–calcium exchanger. Genomics 37, 105 – 112.

155

Lee, S.L., Yu, A.S., Lytton, J., 1994. Tissue-specific expression of Na(+)– Ca2+ exchanger isoforms. J. Biol. Chem. 269, 14849 – 14852. Li, X.F., Lytton, J., 1999. A circularized sodium–calcium exchanger exon 2 transcript. J. Biol. Chem. 274, 8153 – 8160. Li, Z., et al., 1991. Identification of a peptide inhibitor of the cardiac sarcolemmal Na(+)–Ca2+ exchanger. J. Biol. Chem. 266, 1014 – 1020. Li, Z., et al., 1994. Cloning of the NCX2 isoform of the plasma membrane Na(+)–Ca2+ exchanger. J. Biol. Chem. 269, 17434 – 17439. Matsuoka, S., Nicoll, D.A., He, Z., Philipson, K.D., 1997. Regulation of cardiac Na(+)–Ca2+ exchanger by the endogenous XIP region. J. Gen. Physiol. 109, 273 – 286. Mount, S.M., 1982. A catalogue of splice junction sequences. Nucleic Acids Res. 10, 459 – 472. Nicholas, S.B., Yang, W., Lee, S.L., Zhu, H., Philipson, K.D., Lytton, J., 1998. Alternative promoters and cardiac muscle cell-specific expression of the Na+/Ca2+ exchanger gene. Am. J. Physiol. 274, H217 – H232. Nicoll, D.A., Longoni, S., Philipson, K.D., 1990. Molecular cloning and functional expression of the cardiac sarcolemmal Na(+)–Ca2+ exchanger. Science 250, 562 – 565. Nicoll, D.A., Hryshko, L.V., Matsuoka, S., Frank, J.S., Philipson, K.D., 1996a. Mutation of amino acid residues in the putative transmembrane segments of the cardiac sarcolemmal Na+–Ca2+ exchanger. J. Biol. Chem. 271, 13385 – 13391. Nicoll, D.A., Quednau, B.D., Qui, Z., Xia, Y.R., Lusis, A.J., Philipson, K.D., 1996b. Cloning of a third mammalian Na+–Ca2+ exchanger, NCX3. J. Biol. Chem. 271, 24914 – 24921. Nicoll, D.A., Ottolia, M., Lu, L., Lu, Y., Philipson, K.D., 1999. A new topological model of the cardiac sarcolemmal Na+–Ca2+ exchanger. J. Biol. Chem. 274, 910 – 917. Nicoll, D.A., Ottolia, M., Philipson, K.D., 2002. Toward a topological model of the NCX1 exchanger. Ann. N.Y. Acad. Sci. 976, 11 – 18. Nister, M., Heldin, C.H., Westermark, B., 1986. Clonal variation in the production of a platelet-derived growth factor-like protein and expression of corresponding receptors in a human malignant glioma. Cancer Res. 46, 332 – 340. Ota, T., et al., 2004. Complete sequencing and characterization of 21,243 full-length human cDNAs. Nat. Genet. 36, 40 – 45. Philipson, K.D., Nicoll, D.A., 2000. Sodium–calcium exchange: a molecular perspective. Annu. Rev. Physiol. 62, 111 – 133. Philipson, K.D., Nicoll, D.A., Matsuoka, S., Hryshko, L.V., Levitsky, D.O., Weiss, J.N., 1996. Molecular regulation of the Na(+)–Ca2+ exchanger. Ann. N.Y. Acad. Sci. 779, 20 – 28. Quednau, B.D., Nicoll, D.A., Philipson, K.D., 1997. Tissue specificity and alternative splicing of the Na+/Ca2+ exchanger isoforms NCX1, NCX2, and NCX3 in rat. Am. J. Physiol. 272, C1250 – C1261. Quednau, B.D., Nicoll, D.A., Philipson, K.D., 2004. The sodium/calcium exchanger family-SLC8. Pflugers Arch. 447, 543 – 548. Reilly, R.F., Lattanzi, D., 1996. Identification of a novel alternatively spliced isoform of the Na(+)–Ca2+ exchanger (NACA8) in heart. Ann. N.Y. Acad. Sci. 779, 129 – 131. Schwarz, E.M., Benzer, S., 1997. Calx, a Na–Ca exchanger gene of Drosophila melanogaster. Proc. Natl. Acad. Sci. U. S. A. 94, 10249 – 10254. Thurneysen, T., Nicoll, D.A., Philipson, K.D., Porzig, H., 2002. Sodium/ calcium exchanger subtypes NCX1, NCX2 and NCX3 show cellspecific expression in rat hippocampus cultures. Brain Res. Mol. Brain Res. 107, 145 – 156. Van Eylen, F., Kamagate, A., Herchuelz, A., 2001. A new Na/Ca exchanger splicing pattern identified in situ leads to a functionally active 70 kDa NH(2)-terminal protein. Cell Calcium 30, 191 – 198.