Molecular Cloning of a Sixth Member of the K+-dependent Na+/Ca2+ Exchanger Gene Family, NCKX6*

Bioinformatic and molecular cloning tools were used to identify and isolate cDNA clones from mouse and human tissues that encode the sixth member of the K+-dependent Na+/Ca2+ exchanger family, NCKX6. The mouse NCKX6 protein is 585 amino acids long and shares about 62% sequence similarity with previously identified exchangers in the α-repeat regions but has little primary sequence similarity outside these regions. NCKX6 transcripts of 4 kb are abundantly expressed in all tissues examined and are thus more broadly distributed than previously described NC(K)X family members. Two alternatively spliced products of this novel gene were identified that encode proteins of different length. The short isoform differs from the full-length isoform at the C-terminal hydrophobic domain as a result of a shift in the reading frame caused by the deletion of two exons. Both NCKX6 isoforms were expressed in HEK-293 cells. Functional analysis by digital imaging of fura-2 loaded transfected HEK-293 cells demonstrated that the short isoform exhibited K+-dependent Na+/Ca2+ exchange activity whereas the full-length isoform did not. The latter was retained within the endoplasmic reticulum, whereas the short isoform was present at the plasma membrane in transfected cells. Immunofluorescence studies examining NCKX6 expression in native tissue using an NCKX6-specific antibody showed intense labeling of the cardiac sarcolemmal membrane. The discovery of NCKX6 therefore reveals a novel member of the Na+/Ca2+ exchanger superfamily whose ubiquitous expression in all tissues suggests an important role for K+-dependent Na+/Ca2+ exchange in maintaining cellular Ca2+ homeostasis in diverse tissues and cell types.

Calcium signaling plays an important role in mediating numerous cellular processes in virtually all types of animal cells (1). In order to maintain long term cellular Ca 2ϩ homeostasis, Ca 2ϩ entering the cytoplasm to initiate a signal must be removed once the signal has been terminated. Plasma membrane Na ϩ /Ca 2ϩ exchangers are a major pathway for transmembrane Ca 2ϩ efflux and have been extensively investigated in a wide range of tissues, particularly in the heart and brain (2,3). Detailed functional and molecular studies have revealed the existence of two families of Na ϩ /Ca 2ϩ exchanger genes encoding proteins that share sequence similarity in two internally homologous domains known as ␣-repeats (4). One family, K ϩindependent Na ϩ /Ca 2ϩ exchangers (NCX), 1 is thought to catalyze the electrogenic exchange of either 3 or 4 Na ϩ for 1 Ca 2ϩ (2,(5)(6)(7). The NCX family is composed of three distinct gene products: NCX1 (SLC8A1) (8), NCX2 (SLC8A2) (9) and NCX3 (SLC8A3) (10). NCX1 is expressed at high levels in cardiac muscle, brain, and kidney and is also present to a lesser extent in many other tissues (11,12). NCX2 and NCX3, in contrast, are expressed primarily in only two tissues, brain and skeletal muscle (9,10). All three exchangers share about 70% overall amino acid identity that rises to more than 80% within the predicted transmembrane segments (TMS) (10). NCX1 and NCX3, but not NCX2, undergo tissue-specific alternative splicing, which occurs in a small region of the large intracellular loop of the exchanger (12). The second family, K ϩ -dependent Na ϩ /Ca 2ϩ exchangers (NCKX), catalyzes the electrogenic countertransport of 4 Na ϩ for 1 Ca 2ϩ and 1 K ϩ (13,14). NCKX exchangers differ from NCX proteins in their absolute requirement for K ϩ , lower Ca 2ϩ transport rates, and primary amino acid sequence outside the ␣-repeats (2). NCKX1 (SLC24A1) was initially cloned from bovine rod photoreceptors where it was believed to play a central and unique role in the mammalian phototransduction pathway because its ionic stoichiometry allowed the maintenance of Ca 2ϩ extrusion despite the unusual ionic environment of the vertebrate photoreceptors (15,16). Evidence from functional measurements, however, suggested the existence of K ϩ -dependent Na ϩ /Ca 2ϩ exchange processes in tissues other than eye, for instance, brain synaptic plasma membrane (17) and platelets (18). Those observations led to the search for other members of the K ϩ -dependent Na ϩ /Ca 2ϩ exchanger gene family.
Consequently, NCKX2 (SLC24A2) was cloned, first from rat brain (19) and then from chick and human cone photoreceptors (20). NCKX3 (SLC24A3) and NCKX4 (SLC24A4) were cloned and characterized recently (21,22), and NCKX5 has also been cloned (GenBank TM accession number AB085629) (23). Northern blot analysis was employed to compare the distribution pattern of individual NCKX exchangers among different tissues (24). NCKX1 expression is restricted to a 6-kb transcript present only in eye. The NCKX2 transcript of 11 kb is expressed at high levels only in brain. NCKX3 transcripts of 5 kb, and NCKX4 transcripts of 10 and 4.5 kb, are broadly expressed with particular abundance in various brain regions, aorta, lung, and intestine. The tissue-specific expression patterns of these known NCKX members may reflect the different Ca 2ϩ handling properties of different tissues or cells. Expansion of the NCKX family suggests a wider role for K ϩ -dependent Na ϩ / Ca 2ϩ exchange in maintaining cellular Ca 2ϩ homeostasis than previously anticipated. The functional significance of K ϩ -dependent Na ϩ /Ca 2ϩ exchange has been demonstrated recently (25) in axon terminals of the rat neurohypophysis where more than 60% of Ca 2ϩ efflux is mediated by Na ϩ /Ca 2ϩ exchange activity, of which in excess of 90% is due to NCKX.
Analysis of NCKX mRNA expression using probes encoding the C-terminal hydrophobic portion conserved between NCKX paralogs has revealed evidence for further, as yet uncharacterized, mammalian members of the Na ϩ /Ca 2ϩ exchanger family, i.e. a 3.5-kb band in stomach (26). In this study, starting from a bioinformatics-based search of the GenBank TM data base, we have identified a novel member of the Na ϩ /Ca 2ϩ exchanger gene superfamily, NCKX6, that has various unique properties distinguishing it from previously identified members. We demonstrate that expression of NCKX6 transcripts is ubiquitous in all tissues examined and that HEK-293 cells expressing NCKX6 indeed display K ϩ -dependent Na ϩ /Ca 2ϩ exchange activity.

EXPERIMENTAL PROCEDURES
All molecular procedures were performed according to standard protocols (27,28) or the directions of reagent manufacturers, unless noted otherwise. Common chemical reagents were obtained from Fisher, Sigma, or BDH and were of analytical grade or better, unless indicated otherwise. Fluorescent dye terminator cycle sequencing was done at the University of Calgary Core DNA Service facility. Nucleic acid sequence, protein amino acid sequence, and phylogenetic analyses were done in MacVector (Accelrys Inc., Madison, WI). BLAST (29) searches of the sequence data bases were run at the National Center for Biotechnology Information (NCBI) web site (Bethesda) or at the Ensembl Genome Server of the European Bioinformatics Institute (Hinxton, UK).
Identification and Cloning of the Mouse and Human NCKX6 cDNAs-The draft human genome sequence protein data base was searched with the conserved amino acids of the rat NCKX2 ␣-repeats (19) using the BLASTP program (29). 23 hits were found in seven proteins: six of them belonged to previously cloned NCX/NCKX members, whereas one hypothetic protein FLJ22233 (GenBank TM accession number NP_079235) was novel. Multiple sequence alignment demonstrated that the FLJ22233 protein aligned with the C-terminal half of known human NCX/NCKX proteins including the ␣-repeat domain (data not shown).
The FLJ22233 protein sequence was then used to screen the NCBI data bases more extensively. This analysis resulted in the identification of a novel mouse cDNA (GenBank™ accession number AF261233) encoding a homologous protein. The deduced novel mouse protein, which had not been characterized functionally (30), showed a pattern of hydrophobicity analogous to other known mammalian NCKX molecules. In addition, AF261233 had a stop codon upstream from the first possible translational start site (GGCCCATGG) which is similar to the Kozak consensus initiation sequence (31). Two pairs of primers based on the AF261233 sequence were used to amplify the coding region in two parts by RT-PCR using Superscript II Reverse Transcriptase followed by Expand High Fidelity PCR System (Invitrogen and Roche Applied Science, respectively): 1) the 5Ј end, A1 (CAT CCG GCT AGA GGA AGA CTC) to A4 (GAT GAC CAG CTG CAG ACA GTT G); 2) the 3Ј end, A3 (GAG ACC ACT GTC CAG ATC CTG) to A2 (AGG ACC TTC TCA CCC TGC AGA). Total RNA samples isolated from mouse brain, kidney, heart, lung, and thymus were used. Amplification with A1 and A4 produced a single band of 1201 bp, and amplification using A3 and A2 resulted in two fragments of 1005 and 856 bp. Full-length coding fragments were generated by combining the A14 fragments with either the long or short A32 fragments, amplifying with A1 and A2 primers, and cloning into the EcoRV site of pBluescript II SK(Ϫ) (Stratagene, La Jolla, CA). The resulting "long" and "short" clones were sequenced entirely. The longer clone, encoding the full-length NCKX6 protein, was identical to the novel mouse homolog previously identified in Gen-Bank TM (accession number AF261233). While this manuscript was being prepared, the sequence data for the short clone was deposited in GenBank TM with accession number BC043689 by the Mammalian Gene Collection group at the National Institutes of Health.
To compare the relative expression level between the full-length and the short transcripts of NCKX6 in various mouse tissues, a pair of primers was designed based on sequences flanking exons 13 and 14: E12F (CCT GGG GAA ACA GCA TTG GAG) and E15R (CAG TAA TCC GTC TGG CTC CAG). Amplification by RT-PCR using E12F and E15R was carried out for 15, 20, 25, or 30 cycles. To investigate if alternative splicing occurred in the central cytoplasmic loop, as observed in NCX1 and NCX3 (12), NCKX1 (26), NCKX2 (19), and NCKX4 (22), a pair of primers was designed based on sequences encoding the transmembrane segments flanking the large cytoplasmic loop of NCKX6: E6F (TTC ACT GCA CTC TAT CTT GGC) and E11R (ACA GGA AGG AGA CCG CCA AT). The PCR products were then analyzed in 5% polyacrylamide gels, purified, and sequenced.
The complete mouse NCKX6 coding sequence was used to perform a BLAST search of the human genome sequence, and a region sharing about 80% sequence similarity to the second exon in the mouse NCKX6coding sequence was identified. BLAST searches of human EST data bases using 96 nucleotides from this region revealed six sequences (GenBank™ accession numbers BQ229748, BM560555, BI918586, BI753221, H46908, and H20230) with various lengths and 100% identity to the query nucleotides. Two sequence contigs, assembled from these six clones and the original partial human cDNA (AK025886), had distinct 5Ј end sequences extending to 84 nucleotides upstream from the start codon but were otherwise identical. Human NCKX6 was amplified by PCR from Marathon-Ready human thymus cDNA (Clontech, Palo Alto, CA) using High Fidelity Taq polymerase (Roche Applied Science), and the following two primers designed based on the common human sequence: 1) starting from the start codon, H1 (ATG GCC GGC AGA AGG CTG AAT C), and 2) just downstream from the C terminus of the deduced coding region, H2 (AGT GAG GCC ACA GCA CT AAG). A single PCR product was obtained, subsequently cloned into pBluescript II SK (Ϫ), and sequenced. During preparation of this manuscript, a clone identical in sequence to the human NCKX6 cDNA we isolated, and corresponding to the full-length mouse NCKX6 clone, was deposited in GenBank TM with accession number AX537505.
Northern Blot Analysis-Tissue distribution of the NCKX6 transcripts in rodent tissues was studied on Northern blots of total RNA samples using a digoxigenin-UTP-labeled antisense riboprobe from the full-length mouse NCKX6 cDNA sequence, according to instructions of the manufacturer (Roche Applied Science) as described previously (32). Briefly, 10 g of total RNA isolated by guanidine isothiocyanate extraction and CsCl centrifugation from various rat or mouse tissues was separated on 1% formaldehyde-agarose gels and transferred to nylon membranes by capillary diffusion overnight. The UV cross-linked membranes were then hybridized at high stringency using the labeled riboprobe.
Antibody Preparation-Affinity-purified antibody to mouse NCKX6 was prepared at Affinity Bioreagents, Inc. (Golden, CO), by immunizing rabbits with a carrier-conjugated synthetic peptide corresponding to the sequence VDPDKDDRNWKRPLN (amino acids residues 345-359; located in the large cytoplasmic loop of the mouse NCKX6 protein), followed by purification using the immunogenic peptide coupled to a solid support.
Expression in HEK-293 Cells-Expression constructs encoding the different NCKX6 proteins were obtained by digesting corresponding constructs in pBluescript II SK (Ϫ) with convenient unique restriction endonucleases and ligating into the mammalian expression vector pcDNA3.1(ϩ) or -(Ϫ) (Invitrogen). In order to detect protein expression, a FLAG epitope (amino acids DYKDDDDK) was inserted at the predicted extracellular sequence between Arg-70 and Asn-71 of the mouse full-length and short NCKX6 proteins using two chimeric primers and two external primers that flanked convenient unique restriction enzyme digestion sites in a two-step process essentially as described previously (19).
Transient expression of proteins encoded by Qiagen-purified cDNA constructs in HEK-293 cells was performed using a standard calciumphosphate precipitation protocol with BES buffer essentially as described previously (19). Two days following transfection, postnuclear extracts were prepared by solubilizing transfected cells for 20 min in ice-cold lysis buffer (1% Triton X-100, 0.5% deoxycholate, 0.14 M NaCl, 10 mM EDTA, 25 mM Tris-Cl, pH 7.4, 100 units/ml aprotinin, 0.1 mM phenylmethylsulfonyl fluoride (PMSF)) followed by centrifugation for 5 min at 16,000 ϫ g. Protein concentration was determined by Bradford assay using the reagent from Bio-Rad, with bovine ␥-globulin as a standard. Immunoblotting was performed as described previously (19,33), using either anti-peptide antibody or M2 anti-FLAG monoclonal antibody (Sigma), and detected using SuperSignal Plus ECL reagents (Pierce).
Analysis of NCKX6 Function by Ca 2ϩ Imaging-Ca 2ϩ transport into transfected HEK-293 cells was measured by fura-2 fluorescence ratio digital imaging essentially as described previously (19,21). In brief, 2 days after transfection, HEK-293 cells on poly-D-lysine-precoated coverslips were loaded with 5 M fura-2 AM (Molecular Probes, Eugene, OR) and mounted in a perfusion chamber on a microscope stage. The ratio of fura-2 fluorescence with excitation at 340 or 380 nm was followed with time and captured using an intensified CCD camera (ICCD200) and the Image Master System (Photon Technology International, Lawrenceville, NJ). Cells were initially perfused with Na ϩcontaining K ϩ -free solution (145 mM NaCl, 10 mM D-glucose, 0.1 mM CaCl 2 , 10 mM HEPES-trimethylamine, pH 7.4) for 2 min, followed by alternating changes to solutions in which the NaCl was replaced with either 140 mM LiCl and 5 mM KCl or 145 mM LiCl.
Indirect Immunofluorescence-Subcellular location of the FLAG epitope was determined using immunofluorescence essentially as described previously (34) with some modifications. In brief, HEK-293 cells transfected with FLAG-tagged mouse full-length NCKX6 clone, FLAGtagged mouse short NCKX6 clone, FLAG-tagged rat NCKX2 (as a positive control) (19), or vector alone (as a negative control) were grown on glass coverslips that had been pre-coated with 1 mg/ml poly-D-lysine (Sigma). Transfected HEK-293 cells were rinsed in PBSCM (phosphatebuffered saline supplemented with 0.1 mM CaCl 2 and 1 mM MgCl 2 ), fixed in 4% paraformaldehyde in PBSCM, and blocked with 0.2% fish gelatin/PBSCM for 30 min. The cells were incubated with the monoclonal anti-FLAG antibody (1:500) in PBSCM containing 0.2% fish gelatin for 1 h at room temperature, rinsed three times for 5 min each in PBSCM, and then stained by an FITC-conjugated anti-mouse secondary antibody (1:500) in 0.2% fish gelatin/PBSCM for 30 min. Double labeling experiments were performed as follows. HEK-293 cells were first fixed in 4% paraformaldehyde and then permeabilized with 0.2% Triton X-100 in PBSCM for 5 min. After staining for the FLAG epitope, cells were stained with the rabbit anti-SERCA antibody N1 (1:400) to visualize the endoplasmic reticulum (ER), followed by a rhodamineconjugated anti-rabbit secondary antibody.
Mitochondria were stained using MitoTracker Red essentially according to the instructions of the manufacturer (Molecular Probes Inc., Eugene, OR) with some modifications. Briefly, transfected HEK-293 cells on glass coverslips were washed three times with 10 ml of Dulbecco's modified Eagle's medium and then incubated for 30 min at 37°C with 10 ml of pre-warmed (37°C) growth medium containing 100 nM MitoTracker Red. Cells were washed again with 10 ml of Dulbecco's modified Eagle's medium and treated for 15 min at 37°C with 10 ml of pre-warmed growth medium containing freshly prepared 3.7% formaldehyde. After rinsing twice with 10 ml of PBSCM, the cells were permeabilized and stained for the FLAG epitope as described above.
Golgi were identified using pEGFP-g67 (C113) (generous gift from Dr. D. J. Fujita, University of Calgary), which is an expression construct that encodes a fusion protein between EGFP and the C-terminal 113 amino acids of a recently identified Golgi protein Golgin-67 (35). This construct was co-transfected into HEK-293 cells with the FLAGtagged mouse full-length NCKX6 clone. The EGFP fusion protein was visualized using the FITC channel of the immunofluorescence microscope. The cells were then permeabilized and incubated with monoclonal anti-FLAG antibody followed by rhodamine-conjugated antimouse antibody as described above.
In all cases, coverslips were mounted in an anti-fade solution containing 4,6-diamidino-2-phenylindole (Vector Laboratories, Burlingame, CA). Immunofluorescence microscopy was performed using standard epifluorescent optics on a Zeiss Axioscop II through a Fluar 63ϫ objective. Images were captured using a Spot digital camera (Diagnostic Instruments, Inc., Sterling Heights, MI) and processed with Adobe Photoshop 6.0 (Adobe Systems Inc., San Jose, CA).
Detection of Endogenous NCKX6 Protein in Rat Ventricular Myocytes-A sarcolemmal membrane fraction was prepared from adult rat ventricles following a modified protocol as described previously (36). All steps were performed at 4°C. In brief, the ventricles were rinsed in phosphate-buffered saline, minced by scissors, and homogenized in mannitol buffer (250 mM mannitol, 70 mM Tris, pH 7.4, 100 units/ml aprotinin, and 0.1 mM PMSF) in a volume equal to three times the weight of starting material, with three 5-s bursts using a Polytron homogenizer T-3000 (Brinkmann Instruments) at a dial setting of 13 K.
The homogenate was centrifuged at 14,000 ϫ g (max) for 20 min, and the pellet was re-homogenized with the Polytron three times for 30 s, each burst separated by an interval of 30 s. After centrifugation again as described above, the supernatant was kept and the pellet was resuspended, homogenized, and centrifuged once again as described above. The collected supernatants were combined and centrifuged at 48,000 ϫ g (max) for 30 min. The pellet was resuspended in ϳ0.55 ml of mannitol buffer per 1 g of original weighed ventricle, layered over 0.64 M sucrose solution containing 20 mM imidazole HCl, pH 7.4, 100 units/ml aprotinin, and 0.1 mM PMSF, and centrifuged for 90 min at 161,000 ϫ g (max). The turbid fraction at the interface between the sucrose and mannitol layers was collected, diluted with 3 volumes of mannitol buffer, and sedimented at 161,000 ϫ g (max) for 30 min. The pellet was resuspended at 200 l per ventricle in 160 mM NaCl, 20 mM MOPS/Tris, pH 7.4, 100 units/ml aprotinin, and 0.1 mM PMSF. Immunoblotting was performed as described above using NCKX6 polyclonal anti-peptide antibody or anti-peptide antibody preincubated with competing peptide for 30 min at room temperature (1 l of antibody, 1 g of peptide).
Single rat ventricular myocytes (kind gift from Dr. Henry Duff, University of Calgary) were freshly isolated by using a modified Langendorff procedure as described previously (37). Myocytes were rinsed in PBSCM twice, fixed in 4% paraformaldehyde in PBSCM, and permeabilized with 0.2% Triton X-100 for 10 min followed by blocking with 0.2% fish gelatin/PBSCM for 30 min. The myocytes were incubated with NCKX6 anti-peptide polyclonal antibody (1:200) or peptide-absorbed antibody in 0.2% fish gelatin/PBSCM for 1 h, rinsed three times for 5 min each in PBSCM, and then stained by rhodamine-conjugated antirabbit second antibody (1:400) for 30 min. Visualization and processing of images were performed as described above.

Identification and Isolation of Mouse and Human cDNAs
Encoding a Novel Na ϩ /Ca 2ϩ Exchanger-The Na ϩ /Ca 2ϩ exchanger superfamily members are defined by the presence of two conserved ␣-repeats and a similar pattern of membranespanning segments (3). As part of our interest in molecular diversity of Na ϩ /Ca 2ϩ exchangers, BLAST searches of the NCBI data bases were performed using the conserved amino acid sequence of the ␣-repeat regions of the rat brain NCKX2 (19). Initially, a partial human homolog cDNA sequence and then a corresponding full-length mouse homolog cDNA sequence were identified (GenBank™ accession numbers, AK025886 and AF261233, respectively). By combining PCR cloning and bioinformatics analysis, we were able to isolate full-length mouse and human cDNA clones encoding a novel member of the Na ϩ /Ca 2ϩ exchanger superfamily (Fig. 1). Comparison of the deduced amino acid sequence of this novel gene product with the other family members (Fig. 2) revealed a moderate level of similarity in the hydrophobic segments, especially the putative ␣-repeat regions, and a closer phylogenetic relation to the NCKX branch of the family than to the NCX branch. Therefore, this novel gene encodes a new member of the Na ϩ /Ca 2ϩ exchanger superfamily. Recent NCBI data base submissions (GenBank TM accession numbers AB085629, AK089225, XM_230584, and XM_208771) have identified another NCKX family member, designated "NCKX5" (23), which is more closely related to the existing published four members NCKX1-NCKX4 (15,19,21,22) than the gene described here, which is therefore referred to as NCKX6.
Initial analysis of NCKX6 transcripts from various mouse tissues by PCR (Fig. 3) using primers that spanned various regions of the coding sequence indicated a single species, except for two PCR products of different size corresponding to the C-terminal end of the encoded protein. Further analysis of the transcripts by sequencing and RT-PCR using exon-specific primer pairs (Fig. 3C) suggested alternative splicing with inclusion or exclusion of a 149-nucleotide sequence corresponding to exons 13 and 14. The relative intensities of the two species observed by PCR amplification with different reaction cycles suggested that transcripts lacking exons 13 and 14, designated as the short species in this study, were present at a significant level that was lower than that of the full-length species (Fig.  3C). Consistent with such a relative abundance, no short clones could be isolated by PCR from human thymus and heart cDNA. No other alternatively spliced products were observed in either the region encoding the N-terminal half or the large cytoplasmic loop of NCKX6 (Fig. 3, B and D).
The full-length NCKX6 cDNA sequence encodes a protein of 585 (mouse) or 584 (human) amino acids (Fig. 1). Mouse and human NCKX6 share an overall amino acid identity of 83% that rises to more than 92% within the predicted TMSs. Hydropathy analysis (Fig. 4) suggests 13 hydrophobic transmembrane spans resulting in a protein with a topology analogous to the current models of both NCX1 (38) and NCKX2 (34) in the N-terminal half. Seven transmembrane segments are proposed in the C-terminal half, which results in the second ␣-repeat having a membrane orientation consistent with current models. In addition, both a potential cleavage site for signal peptidase and two consensus glycosylation sites are present on the extracellular loop between M0 and M1. The location of several putative phosphorylation sites in the large intracellular loop is also illustrated in Fig. 1.
Mapping the Mouse and Human NCKX6 Genes-BLAST searches of the mouse and human genomic sequences with the corresponding mouse and human NCKX6 gene coding se- FIG. 1. Comparison of mouse and human NCKX6 amino acid sequences. A, the deduced amino acid sequences of mouse and human NCKX6 clones have been aligned using the ClustalW algorithm in MacVector. Similar or identical amino acids are boxed, with identities highlighted in boldface with a dark gray background. Dashes indicate gaps introduced to maximize the alignment. Hydrophobic, putative transmembrane spanning regions identified by hydropathy analysis using the GES hydrophobicity scale (58) are highlighted and labeled M0 to M12. The position of the inserted FLAG epitope, consensus sites for signal peptidase cleavage (SPase?), glycosylation (CHO), and for phosphorylation via protein kinase A (PKA) are also indicated. Asn-442, Gly-459, Leu-463, and Val-528, which align with the corresponding residues in NCKX2 shown to be extracellular (41), are labeled with asterisks. B, amino acid sequence alignment of the C terminus of the proteins encoded by the mouse full-length (Mu-L) and short (Mu-S) NCKX6 clones. The site of divergence that arises from the alternative splicing process is indicated. quences revealed their chromosomal locations at 5.11 (mouse) and 12q24 (human), respectively, and the intron-exon structure of the gene. The coding region of the human and mouse NCKX6 genes consists of 15 exons spread over 19.6 kb (in mouse) with exon 1 separated by more than 6 kb from exons 2-15, which are more closely spaced. Similar to NCKX3 and shown to be extracellular (41). C, phylogenetic tree of the Na ϩ /Ca 2ϩ exchanger superfamily. The two main branches, NCX, NCKX, of the Na ϩ /Ca 2ϩ exchanger gene family are clearly defined. The numbers above each branch of the tree correspond to the p values, which are a measure of proportional differences between sequences. Calculation was performed using ClustalW multiple sequence alignment with MacVector software and confirmed by bootstrapping.
NCKX4 gene arrangement, NCKX6 does not have the unusually long first coding exon found as a conserved feature in the NCX1, NCX3, NCKX1, and NCKX2 genes (21,22). However, the exon boundaries in NCKX6 do not match those of NCKX3 and NCKX4.
Tissue Distribution of NCKX6 Gene Expression-Northern blotting analysis of NCKX6 distribution using total RNA isolated from various rat and mouse tissues is shown in Fig. 5. The major transcript is about 4 kb in length and was found to be abundantly and widely expressed in various tissues examined. This pattern of distribution clearly distinguishes NCKX6 from the restricted distribution of NCX2 (9), NCX3 (10), NCKX1 (15), and NCKX2 (19) and is similar to, but broader than, the widely expressed NCX1 (8), NCKX3 (21), and NCKX4 (22).
Functional Analysis of NCKX6 -The ability of the heterologously expressed NCKX6 protein to operate as a Na ϩ /Ca 2ϩ exchanger was examined by fluorescence ratio imaging of fura-2-loaded transfected HEK-293 cells. Ca 2ϩ influx with or without extracellular K ϩ was measured when cells were perfused with Li ϩ to reverse the Na ϩ gradient across the plasma membrane, thus employing the reverse mode of the exchanger.
Positive control cells transfected with wild-type NCKX2 demonstrated a strong K ϩ -dependent Na ϩ /Ca 2ϩ exchange activity, measured as a large increase in fura-2 fluorescence ratio only in the presence of extracellular K ϩ , whereas cells transfected with vector alone showed no significant change in fluorescence ratio (19). In cells transfected with either the mouse NCKX6 full-length clone (Fig. 6) or the human full-length clone (data not shown), Na ϩ -to-Li ϩ or Na ϩ -to-Li ϩ ϩ K ϩ solution switches did not elicit any significant increase in the fura-2 fluorescence ratio. In cells transfected with the mouse NCKX6 alternatively spliced clone missing exon 13 and 14, however, the Na ϩ -to-Li ϩ ϩ K ϩ perfusion switch, but not the Na ϩ -to-Li ϩ solution switch, caused a significant increase in fura-2 fluorescence, which rapidly returned to the basal level upon switching back to 145 mM Na ϩ -containing solution (Fig. 6). The pattern of change and the strong dependence on K ϩ for function is highly comparable with the NCKX2 exchanger, demonstrating the short NCKX6 isoform indeed operates as a K ϩ -dependent Na ϩ / Ca 2ϩ exchanger.
Subcellular Localization of Mouse Full-length and Short NCKX6 Proteins-Because the full-length NCKX6 protein did  13 and 14). The location of the named primers is illustrated (arrows), and the PCR products isolated are shown by lines below the transcript. A14 (primers A1 and A4); Loop (primers E6F and E11R); AS (primers E12F and E15R); A32 (primers A3 and A2). B, sample of total RNA from each of the indicated mouse tissues was reverse-transcribed and amplified by PCR using two pairs of primers, A1/A4 and A3/A2. The position of size markers (-HindIII and X-174 RF-HaeIII digests) are indicated to the left. For further details, see "Experimental Procedures." Ϫ, negative control (kidney sample processed in the absence of the reverse transcriptase enzyme); Th, thymus; Ht, heart; Kd, kidney; Br, brain; Lu, lung. C, RT-PCR as above was performed using a pair of primers, E12F and E15R, for detection of the full-length transcripts or those with exons 13 and 14 deleted (short). Amplification was stopped after 15 (not shown), 20 (not shown), 25, or 30 cycles. For further details, see "Experimental Procedures." D, RT-PCR as above was performed using a pair of primers, E6F and E11R, for detection of alternative splicing in the large cytoplasmic loop. For further details, see "Experimental Procedures." not exhibit Na ϩ /Ca 2ϩ exchange activity in the HEK-293 cell assay, we investigated the protein expression level and subcellular targeting of the mouse full-length NCKX6 protein. A FLAG epitope was introduced into the extracellular domain of NCKX6 between TMS0 and TMS1, and the resulting molecules were expressed in HEK-293 cells. Immunoblot analysis using the M2 anti-FLAG antibody to test a postnuclear lysate from transfected cells showed both the full-length and the short mouse NCKX6 proteins were detected as bands of about 55 kDa, whereas the positive control, FLAG-tagged NCKX2, was detected around 76 kDa (Fig. 7). To verify that the bands detected by anti-FLAG antibody corresponded to the mouse full-length and short NCKX6 proteins, the same blot was stripped and reprobed with the NCKX6-specific anti-peptide antibody, which resulted in detection of two identical bands for both full-length and short NCKX6 proteins but not for the FLAG-tagged NCKX2. The predicted sizes for the full-length and short NCKX6 proteins, based solely on amino acid sequence, are 64 and 60 kDa, respectively. The observed bands of 55 kDa suggested both proteins underwent post-translational modification(s), i.e. signal peptide cleavage and/or glycosylation (data not shown), as documented previously for NCX1 (39).
Surface delivery of the expressed NCKX6 molecules was assessed in HEK-293 cells by immunofluorescence. As shown in Fig. 8A, fixed but unpermeabilized cells expressing the FIG. 4. Hydropathy profile and model for the transmembrane topology of mouse NCKX6. Hydropathy profiles (generated using the GES hydrophobicity scale (58) with a window of 19 amino acids) for the mouse full-length NCKX6 protein (A) and the mouse short NCKX6 protein (C), respectively. The position of the predicted transmembrane segments are indicated with bars labeled M0 to M12. The corresponding topological models (B and D) are illustrated with cylinders indicating the predicted transmembrane segments. The schematic models also indicate the putative glycosylation sites (CHO), and signal peptidase cleavage sites (SigPase?), as well as the ␣-repeats (boxes).

FIG. 5. Northern blot analysis of NCKX6 tissue distribution.
Ten-g samples of total RNA isolated from the indicated rat (A) or mouse (B) tissues were separated on 1% formaldehyde-agarose gels, transferred to nylon membranes, and hybridized at high stringency using a riboprobe from mouse NCKX6. The position of RNA size markers is shown in kb on the left. Th, thymus; Ao, aorta; Ht, heart; Lu, lung; Te, testis; Mu, mucosal layer of small intestine; SI, small intestine; LI, large intestine; St, stomach; Kd, kidney; Cc, cerebral cortex; Cb, cerebellum; Bs, brainstem; Br, brain; and SM, skeletal muscle.
FIG. 6. Functional analysis of mouse NCKX6 using Ca 2؉ imaging. The averaged fura-2 intensity ratio for the indicated number of cells is shown for background-subtracted fluorescence determined by Ca 2ϩ imaging with excitation at 340 or 380 nm. Error bars indicate the mean Ϯ S.E. HEK-293 cells on coverslips were transfected with cDNA constructs encoding positive control rat NCKX2 (NCKX2), the mouse full-length NCKX6 clone (Mu-L), the mouse short NCKX6 clone (Mu-S), or vector alone (Neg). Cells were then loaded with fura-2, mounted in a perfusion chamber on the microscope, and observed by fluorescent digital imaging. An initial perfusion of 2 min with Na ϩ -containing K ϩ -free solution (Na, 0K) was followed by alternating changes to Li ϩcontaining solutions supplemented with (Li, 5K) or without (Li, 0K) 5 mM K ϩ , as indicated.
FLAG-tagged mouse short NCKX6 protein showed clear staining at the cell surface, present as punctate clusters, similar to the surface-staining pattern observed for the positive control FLAG-tagged NCKX2. Cells expressing the FLAG-tagged mouse full-length NCKX6 protein, on the other hand, showed no significant fluorescent staining under non-permeabilized conditions. Cells permeabilized with Triton X-100 demonstrated that the mouse full-length NCKX6 protein was predominantly distributed in the perinuclear region and overlapped well with the endogenous SERCA pump, suggesting an ER localization (Fig. 8B). We also employed markers for Golgi complex and mitochondria, and we confirmed that the mouse full-length NCKX6 protein was not associated with these compartments to any appreciable extent (Fig. 8B).
Detection and Localization of Endogenous NCKX6 Protein at the Plasma Membrane of Rat Ventricular Myocytes-Immunoblots performed on a cardiac sarcolemmal membrane fraction using the NCKX6 anti-peptide polyclonal antibody showed two bands of about 64 and 55 kDa, both of which disappeared when the antibody was first incubated with blocking peptides (Fig.  9A). Distribution of NCKX6 immunoreactivity in freshly isolated ventricular myocytes was studied after permeabilization with 0.2% Triton X-100. Immunofluorescence experiments using the polyclonal NCKX6 anti-peptide antibody showed intense labeling of the cardiac sarcolemmal membrane, visible as a peripheral fluorescent staining mid-cell, and as a regular punctate pattern at the base of the cell where the membrane surface was flattened by contact with the coverslip (Fig. 9B). These data demonstrate that in native cells, the NCKX6 protein could reach the plasma membrane. No significant fluorescence was observed when the polyclonal anti-NCKX6 antibody was first incubated with competing peptide before treatment of the myocytes (Fig. 9B).

DISCUSSION
In this study, we have identified a novel Na ϩ /Ca 2ϩ exchanger molecule, NCKX6, which is ubiquitously expressed in various tissues examined and shares a moderate degree of sequence similarity in the ␣-repeat regions with previously cloned NCX and NCKX exchangers. Phylogenetic analysis and multiple sequence alignments suggested that NCKX6 has a closer relationship with NCKX molecules than with NCX members. The observed requirement of K ϩ for Na ϩ /Ca 2ϩ exchange function, as tested by calcium imaging, demonstrated that NCKX6 is a K ϩ -dependent Na ϩ /Ca 2ϩ exchanger.
Hydropathy analysis suggested that the full-length NCKX6 protein possesses a topology in the N-terminal domain very similar to known NCX and NCKX members: a signal peptide, M0, a glycosylated extracellular loop, and a cluster of five TMSs followed by a large intracellular loop. Unlike NCX and NCKX members, whose C-terminal halves consist of six hydrophobic segments predicted by hydropathy plots, NCKX6 contains seven predicted hydrophobic spans in the C-terminal half. Experimental tests of the initial topology for NCX1 (38) and NCKX2 (34,40) have provided evidence for a revised model where the number and orientation of transmembrane segments in the C-terminal hydrophobic region are altered. In the human cone NCKX2 exchanger, amino acids Trp-519, Gly-536, Leu-540, and Leu-603 were recently identified in positions close to the extracellular surface of the plasma membrane (41). The topology model represented in Fig. 4 is consistent with the extracellular location of the corresponding amino acids in the mouse NCKX6 protein (Asn-442, Gly-459, Leu-463, and Val-528) (Fig. 1A and Fig. 2B). The proposed topology of NCKX6, along with current models for NCX-and NCKX-type exchangers, place the ␣-repeat regions on opposite faces of the membrane. Therefore, it seems that NCX1, NCKX2, and NCKX6 exchangers have similar conserved structural elements formed FIG. 8. Subcellular localization of mouse NCKX6 alternatively spliced isoforms in transfected HEK-293 cells. A, HEK-293 cells were transfected with either the FLAG-tagged mouse full-length NCKX6 clone (FLAG-Mu-L), the FLAG-tagged mouse short NCKX6 clone (FLAG-Mu-S), the positive control FLAG-tagged rat NCKX2 (FLAG-NCKX2), or the negative control vector alone (not shown) and then subjected to immunofluorescence analysis without membrane permeabilization using the M2 anti-FLAG monoclonal antibody followed by FITC-conjugated anti-mouse antibody. B, HEK-293 cells transfected with FLAG-Mu-L were fixed, permeabilized, and double-stained using monoclonal anti-FLAG antibody followed by FITC-conjugated antimouse antibody, and subsequently polyclonal antibody to the endogenous ER marker SERCA (anti-SERCA) followed by rhodamine-conjugated anti-rabbit antibody (top row). In the middle row, HEK-293 cells co-transfected with FLAG-Mu-L and the Golgi marker pEGFP-G67(C113) expression construct were fixed, permeabilized, and labeled using monoclonal anti-FLAG antibody followed by rhodamine-conjugated anti-mouse antibody. The EGFP fluorescence was visualized by the FITC channel. In the bottom row, HEK-293 cells expressing FLAG-Mu-L were first incubated with the mitochondrial staining dye, Mito-Tracker Red, and then fixed, permeabilized, and labeled with monoclonal anti-FLAG antibody followed by FITC-conjugated anti-mouse antibody. Coverslips were mounted on slides in an anti-fade medium supplemented with 4,6-diamidino-2-phenylindole for nuclei staining (blue). by the ␣-repeat regions, surrounded by a different overall transmembrane structure.
Even though NCKX6 is clearly capable of functioning as a K ϩ -dependent Na ϩ /Ca 2ϩ exchanger, its amino acid sequence is surprisingly divergent from the other NCX and NCKX family members, especially within the two ␣-repeats, as shown in Fig.  2. For example, within the ␣1-repeat, the conserved sequences G(S/G)SAPE and GSA(A/V)FN are replaced with GNGAPD and TTVVAG in NCKX6. Previously studies in both NCX1 (42) and NCKX2 (43) had demonstrated that among other residues, the second Ser of G(S/G)SAPE and the Ser of GSA(A/V)FN were essential to function. Similarly, in the ␣2-repeat, the conserved sequences GTS(I/V)PD and GSN contain essential Thr and Ser residues, respectively, but are replaced with GNSIGD and GGI in NCKX6. The lack of conservation of these key residues suggests they are not essential to ion binding and transport function but may instead play a critical role in providing an appropriate chemical microenvironment for ion binding. Possibly the changes in NCKX6 may tune this exchanger to operate with ions other than, or in addition to, Na ϩ , K ϩ , and Ca 2ϩ . These ideas will need to be explored experimentally.
Alternative splicing is a functionally important mechanism for generating diversity of proteins in general and is often used in genes encoding membrane transporters (44) and channels (45). In the Na ϩ /Ca 2ϩ exchanger superfamily, a number of splice variants of NCX1 have been extensively documented that result in changes in a region of the large intracellular loop of the exchanger (11,12,46). Experimental evidence demonstrates that the alternatively spliced protein products have significant functional and regulatory differences (47,48). That the different splice variants are expressed in a tissue-specific pattern suggests properties of each splice variant have evolved to fit different ionic environments or particular requirements of different tissues or cell types (3). There is also evidence for alternative splicing in NCX3 (12), NCKX1 (26), NCKX2 (19), and NCKX4 (22), all at sites encoding segments of the large intracellular loop. In all these cases the alternative splicing is of the cassette-type, and the normal protein reading frame is maintained downstream from the splice site. In contrast, the alternative splicing in NCKX6 occurs in the middle of the second hydrophobic cluster and results in a shift in the reading frame disrupting the last three predicted transmembrane segments, which are replaced by a unique long intracellular Cterminal tail (Fig. 4D). The consequence is that the short clone, lacking exons 13 and 14, encodes a protein in which the second ␣-repeat is disrupted. Analysis of RNA from several tissues indicated that, although expressed at a lower level than the full-length transcript, mRNA encoding the transcript lacking exons 13 and 14 is still reasonably abundant. The existence of similar single ␣-repeat exchanger proteins has been reported previously. Studies from our laboratory have demonstrated an abundant and unusual circular transcript of the NCX1 gene that encodes a truncated protein, containing only the first ␣-repeat, which is nevertheless functional when expressed in HEK-293 cells (49). Other groups have also reported similar truncated, functional NCX1 proteins (50 -52). The mechanism for how such single ␣-repeat proteins including NCKX6 act as exchangers is not clear but may involve oligomeric association of exchanger proteins (34,53).
Functional analysis of the two NCKX6 proteins by calcium imaging revealed that the mouse short NCKX6 protein clearly demonstrated K ϩ -dependent Na ϩ /Ca 2ϩ exchange activity in transfected HEK-293 cells. Neither the mouse nor human fulllength NCKX6 proteins, however, produced Na ϩ /Ca 2ϩ exchange activity in the same system. Immunofluorescence studies revealed that the full-length NCKX6 clone was expressed well in HEK-293 cells, but the protein product was retained primarily in the endoplasmic reticulum. All previously cloned Na ϩ /Ca 2ϩ exchangers mediate transport at the plasma membrane. Immunostaining studies in ventricular myocytes demonstrated that the endogenous NCKX6 protein was clearly located at the cell surface of these native cells (Fig. 9B). Thus, we propose that the full-length NCKX6 isoform, with two complete ␣-repeats, functions at the plasma membrane in native cells. It is not clear why the full-length NCKX6 protein is retained in the ER when expressed in HEK-293 cells. It is possible that an essential subunit(s), scaffolding interactions, or post-translational processing is lacking in this heterologous system. Interestingly, we detect NCKX6 as a doublet of bands on Western blots of cardiac sarcolemmal membrane, at 64 and 55 kDa (Fig. 9A), whereas only bands at 55 kDa are present in transfected HEK-293 cells (Fig. 7), possibly indicative of altered or defective processing.
It is noteworthy that full-length bovine NCKX1 is also functionally silent when expressed heterologously, similar to what FIG. 9. Analysis of native NCKX6 protein expression in rat ventricular tissue. A, immunoblot to test endogenous NCKX6 protein expression using the NCKX6 anti-peptide polyclonal antibody with or without preincubation with competing peptide. Five-g sample of rat ventricular sarcolemma membranes was loaded in each lane. Molecular weight markers are shown on the left. B, freshly isolated rat ventricular myocytes were fixed, permeabilized with 0.2% Triton X-100, and subsequently stained with polyclonal anti-NCKX6 antibody followed by rhodamine-conjugated anti-rabbit antibody. In the control experiment, the anti-NCKX6 antibody was mixed with competing peptide before the staining procedure. The images were taken at focal planes near the middle of the myocyte or near the base of the myocyte where it attached to the coverslip as indicated. DIC, differential interference contrast image.
we observe for NCKX6 (54). Transfected HEK-293 cells expressing bovine or dolphin NCKX1 molecules containing a region near the N-terminal end of the large cytoplasmic loop did not display Ca 2ϩ transport, whereas deletion of this region from either clone resulted in a functional exchanger, although protein expression from all these clones was similar. In bovine photoreceptors, where transcripts encoding the full-length NCKX1 are in the majority, the molecule is clearly functional at the native cell membrane. Similar findings had been reported during characterization of other alternatively spliced membrane channels and transporters. For example, a novel voltage-dependent and Ca 2ϩ -activated K ϩ channel ␣ subunit splice variant containing a 33-amino acid splice insert in the S1 transmembrane domain was retained in the ER, preventing the channel subunit from producing detectable surface labeling or ionic current at the plasma membrane when expressed in HEK-293 cells (55).
In conclusion, we have described the cloning and characterization of a novel, structurally divergent, and distantly related Na ϩ /Ca 2ϩ exchanger gene family member, NCKX6. Ubiquitous expression of NCKX6 in various tissues suggests a key role for this molecule in regulating intracellular Ca 2ϩ homeostasis in mammalian cells and tissues. As discussed previously (56) and demonstrated in this study, advances in genomic biology over the past years have revolutionized the Na ϩ /Ca 2ϩ exchanger field in revealing a complexity of related genes and alternatively spliced products expressed in many different organisms. The precise, unique physiological role for NCKX6 will no doubt require the application of mouse genetics and other selective molecular tools (57).