cDNA Cloning and Expression of the Cardiac Na+/Ca2+ Exchanger from Mozambique Tilapia (Oreochromis mossambicus) Reveal a Teleost Membrane Transporter with Mammalian Temperature Dependence*

The complete cDNA sequence of the tilapia cardiac Na+/Ca2+ exchanger (NCX-TL1.0) was determined. The 3.1-kb transcript encodes a protein 957 amino acids in length, with a predicted signal peptide cleaved at residue 31 and two potential N-glycosylation sites in the extracellular N terminus. Hydropathy analysis and sequence comparison predicted a mature protein with nine transmembrane-spanning segments, consistent with the structural topologies of other known mammalian and teleost NCX isoforms. Overall sequence comparison shows high identity to both trout NCX-TR1.0 (∼81%) and mammalian NCX1.1 (∼73%), and phylogenetic analyses confirmed its identity as a member of the NCX1 gene family, expressing exons A, C, D, and F in the alternative splice site. Sequence identity is even higher in the α-repeats, the exchanger inhibitory peptide (XIP) site, and Ca2+-binding domains, which is reflected in the functional and regulatory properties of tilapia NCX-TL1.0. When NCX-TL1.0 was expressed in Xenopus oocytes and the currents were measured in giant excised patches, they displayed both positive regulation by Ca2+ and Na+-dependent inactivation in a manner similar to trout NCX-TR1.0. However, tilapia NCX-TL1.0 exhibited a relatively high sensitivity to temperature compared with trout NCX-TR1.0. Whereas trout NCX-TR1.0 currents displayed activation energies of ∼7 kJ/mol, tilapia NCX-TL1.0 currents showed mammal-like temperature dependence, with peak and steady-state current activation energies of 53 ± 9 and 67 ± 21 kJ/mol, respectively. Using comparative sequence analysis, we highlighted 10 residue positions in the N-terminal domain of the NCX that, in combination, may confer exchanger temperature dependence through subtle changes in protein flexibility. Tilapia NCX-TL1.0 represents the first non-mammalian NCX to exhibit a mammalian temperature dependence phenotype and will prove to be a useful model in defining the interplay between molecular flexibility and stability in NCX function.

The Na ϩ /Ca 2ϩ exchanger (NCX) 1 is a polytopic membrane protein that belongs to the cation/Ca 2ϩ antiporter superfamily of Ca 2ϩ transporters. Catalyzing the electrogenic exchange of three Na ϩ ions for one Ca 2ϩ ion, the NCX plays a key role in regulating cytosolic Ca 2ϩ concentrations in many different cell types. The NCX is present in a wide array of species, with high overall sequence identity and conservation of ion transport and regulatory components (1,2). However, knowledge of the molecular function of the NCX is based primarily on examination of mammalian isoforms and is therefore restrictive from a phylogenetic perspective. Examination of non-mammalian isoforms has previously provided valuable insights into NCX function (3)(4)(5). The NCX gene family consists of three cloned members present in vertebrates, NCX1 (6), NCX2 (7), and NCX3 (8). Phylogenetic analyses suggest that these exchanger isoforms arose from separate gene duplication events occurring before the emergence of vertebrates, as evidenced by the presence of all three NCX genes in fish species (2). In addition, a putative fourth member of the NCX family present only in fish species was derived from genomic data, but has unknown function and expression patterns at this time (2).
The relative expression of the NCX among tissues varies greatly and is generally correlative with the need for a high degree of transmembrane Ca 2ϩ flux regulated in a precise manner. Whereas NCX2 and NCX3 are found exclusively in the brain and skeletal muscle, expression of NCX1 is virtually ubiquitous, with the highest levels present in excitable tissues (9). In addition, differential expression of exons from the alternative splice site yields an array of NCX splice variants that are expressed in a tissue-specific manner (9,10). The highest exchange activity is in the heart, where the role of cardiac NCX1.1 in excitation-contraction coupling has been studied extensively (11)(12)(13). NCX1.1 serves as the prime mechanism of Ca 2ϩ extrusion from the cardiomyocyte and is therefore an important contributor to relaxation (14,15). It has been suggested that reverse mode exchange can contribute to cardiac contraction via Ca 2ϩ influx (16 -18), a role that is increased in both the failing heart and neonate heart (19 -21). In addition, NCX1 has been implicated in several cardiac pathophysiologies, including arrhythmogenesis (22)(23)(24) and cellular damage associated with ischemia/reperfusion injury (25,26). Elucidating the molecular mechanisms behind NCX function is crucial to understanding and manipulating its role in the heart.
Previously, we used temperature as a probe to gain insight into the molecular function of the NCX (3,27). Proteins involved in Ca 2ϩ active transport and regulation in the mammalian heart, such as NCX1.1, are highly temperature-dependent (3,28). Homologous proteins in cold-adapted species have evolved differently to maintain adequate cardiac function under conditions that are cardioplegic to mammals. Using both native (29) and cloned (3) proteins, it has been shown that the temperature dependence of NCX1 from cold-and warm-adapted species varies substantially. Comparison of peak outward exchange currents from cloned trout NCX-TR1.0 and canine NCX1.1 gave Q 10 values (-fold change in activity/10°C change in temperature) of 1.1 and 2.4, respectively (3). Furthermore, these disparities in temperature dependence are not due to differences in inactivation kinetics or NCX regulatory properties. At the amino acid level, these isoforms exhibit high overall identity (ϳ73%), which is significantly higher in the ␣-repeat region (ϳ92%), the exchanger inhibitory peptide (XIP) site (ϳ85%), and regulatory Ca 2ϩ -binding domains (95%). Construction of trout and canine NCX chimeras revealed that the region responsible for the differential temperature dependence between isoforms is within the N-terminal transmembrane domain (27). This region composes approximately one-third of the protein, including the first five transmembrane segments up to the end of the XIP site. Further attempts to define specific areas or substitutions responsible for the temperature dependence of the exchanger have proven difficult and given equivocal results. 2 Despite the high overall identity between trout NCX-TR1.0 and canine NCX1.1, the species are separated by ϳ450 million years of evolution (30). To make a more accurate comparison between NCX proteins adapted to high and low temperatures, examination of more closely related NCX isoforms may reduce the number of substitution permutations required to confer the temperature dependence phenotype. Teleosts successfully populate habitats over a large temperature range, and as such, their proteins have evolved to function accordingly. Tilapias (Oreochromis sp.) are euryhaline teleosts that inhabit environments ranging from freshwater to seawater of high salinity and can withstand temperatures as high as 42°C (31,32). Due to their tolerance of relatively extreme environments, tilapias have gained important roles as model organisms in the laboratory (31,33,34) and as a prominent fish in aquaculture (32,35). In this study, we have chosen the Mozambique tilapia (Oreochromis mossambicus) as a model system because the cardiac NCX from this species has likely adapted to function at temperatures similar to those in the mammalian heart, but lethal to trout. We report here the cDNA cloning and characterization of tilapia NCX-TL1.0 and demonstrate, for the first time, a non-mammalian NCX with mammal-like temperature dependence of Na ϩ /Ca 2ϩ exchange.

EXPERIMENTAL PROCEDURES
Animals-Cultured tilapias (O. mossambicus; weighing 40ϳ50 g) were maintained at the Institute of Zoology (Academia Sinica, Taipei, Taiwan) in freshwater (local tap water) at 27°C under a 14-h light/10-h dark photoperiod.
RNA Extraction-Tilapias were killed with a sharp blow to the head. An appropriate amount of heart tissue (ϳ0.5 g) was collected and homogenized in 4 ml of solution containing 4 M guanidine thiocyanate, 1.25 M sodium citrate, 35% N-lauroylsarcosine sodium salt, and 0.1 M 2-mercaptoethanol. Tissue homogenates were mixed with 0.4 ml of 2 M sodium acetate, 0.8 ml of chloroform/isopropyl alcohol (49:1), and 4 ml of phenol and shaken thoroughly. After centrifugation at 13,000 rpm for 30 min at 4°C, the supernatants were mixed with an equal volume of isopropyl alcohol. Pellets were precipitated by another centrifugation at 13,000 rpm for 30 min at 4°C, washed with 70% alcohol, and stored at Ϫ20°C prior to use. The amount and quality of the total RNA were determined by measuring the absorbance at 260 and 280 nm with a Hitachi U-2000 spectrophotometer and by running RNA-denatured gels.
cDNA Cloning and Sequencing-For cloning and sequencing, mRNA was purified from the total RNA of three tilapia hearts using an Oligotex mRNA kit (Qiagen Inc., Hilden, Germany). The cDNA used for all subsequent PCR cloning was prepared by 5Ј-and 3Ј-rapid amplification of cDNA ends (RACE) using a SMART TM RACE cDNA amplification kit (Clontech). A partial cDNA fragment was amplified by PCR using forward (5Ј-AARCARAARCAYCCNGA-3Ј) and reverse (5Ј-ACRAAYTGYTCNCK-CCA-3Ј) degenerate primers based on conserved domains within the intracellular loop of canine NCX1.1 (GenBank TM accession number M57523). The forward primer corresponds to a region just downstream of the XIP site (nucleotides 1018 -1035), whereas the reverse primer matches a region just before transmembrane segment (TMS)-6 (nucleotides 2248 -2265). For PCR amplification, 2 l of cDNA was used as template in a 50-l final reaction volume containing 0.25 mM dNTP, 2.5 units of Ex Taq TM DNA polymerase (TaKaRa, Shiga, Japan), and 0.2 M each primer. For PCR, the reaction conditions were as follows: initial denaturation at 95°C for 3 min, followed by 30 cycles of denaturation at 95°C for 30 s, annealing at 50°C for 30s, extension at 72°C for 1 min, and a final extension at 72°C for 10 min. PCR products were purified and subcloned into the pGEM-T Easy vector (Promega Corp., Madison, WI). Nucleotide sequencing was determined with an ABI 377 Automated DNA Sequencer. From these partial sequences, specific 5Ј-and 3Ј-RACE primers were designed and used for subsequent PCR steps. PCR products were subcloned into the pGEM-T Easy vector and sequenced. The full-length tilapia NCX-TL1.0 clone was covered by three overlapping fragments, each ϳ1 kb in length. These three fragments were ligated together to produce the full-length clone.
Expression of Tilapia NCX-TL1.0 in Xenopus Oocytes-Full-length tilapia NCX-TL1.0 was subcloned from the original sequencing pGEM-T Easy vector into the expression vector pBSTA. To do this, BglII sites were introduced at the start and stop codons of tilapia NCX-TL1.0 in the pGEM-T Easy vector using the QuikChange TM sitedirected mutagenesis kit (Stratagene, La Jolla, CA). The pBSTA vector contains the 5Ј-and 3Ј-untranslated regions of Xenopus ␤-globin, which, in the final construct, replaced the endogenous tilapia untranslated regions and flanked the coding region. The pBSTA vector containing full-length tilapia NCX-TL1.0 was linearized with SacII, and cRNA was synthesized in vitro using the T7 mMessage mMachine kit (Ambion Inc., Austin, TX). The cRNA amount was assessed spectroscopically, whereas cRNA purity was determined using RNA-containing 1% agarose gel. Oocytes were prepared as described previously (36) and injected with 46 nl of cRNA diluted to 0.5 ng/nl. Tilapia NCX-TL1.0 exchange activity was assessed 3-7 days after injection using the giant excised patch technique (see below).
Electrophysiology-Using the giant excised patch technique, we measured outward Na ϩ /Ca 2ϩ exchange currents as described previously (3,27). Briefly, oocytes with the vitellin layer removed were placed in a solution containing 100 mM KOH, 100 mM MES, 20 mM HEPES, 5 mM EGTA, and 5 mM MgCl 2 , and pH was adjusted to 7.0 at room temperature (ϳ22-23°C) with MES. Borosilicate glass pipettes were polished (inner diameter of ϳ20 -30 m) and coated with a Parafilm/mineral oil mixture. Suction was used to form gigaohm seals, and membrane patches were excised by moving the pipette tip.  (37) was used to adjust Mg 2ϩ and Ca 2ϩ to yield free concentrations of 1 mM and 0, 1, or 10 M, respectively. Outward Na ϩ /Ca 2ϩ exchange currents were activated by switching from Li ϩ -to Na ϩ -based bath solution and were recorded at 30, 14, and 7°C. To account for current rundown within a single patch, multiple current traces at all three temperatures were measured. Hardware and software from Axon Instruments, Inc. were used for data acquisition and analysis.
Current Trace Analyses and Statistics-Inactivation kinetics and temperature dependence parameters for the tilapia NCX-TL1.0 currents were calculated as described previously (3). The inactivation rate constant () was obtained by fitting current-time traces to a single exponential. The energy of activation (E a ) and temperature coefficient (Q 10 ) were calculated as indices of temperature dependence as described by Marshall et al. (27). All values are displayed as the means Ϯ S.E. Statistical significance of the results was determined by mean comparison using Tukey's test and one-way analysis of variance performed with Microcal Origin and GraphPad software. Unless indicated otherwise, a p value Ͻ0.05 was considered significantly different.
Sequence Alignments and Phylogenetic Analysis-Sequence and phylogenetic analyses were performed as described previously (2). Sequence data were acquired from either the NCBI Non-redundant Protein Database or derived from whole genome sequences. The following sequences were obtained from the NCBI Protein Database: human NCX1.1 (GenBank TM gi:10863913), human NCX2.1 (gi:10720116), human NCX3.3 (gi:22087483), rainbow trout NCX-TR1.0 (gi:6273849), and green spotted pufferfish NCX (gi:47219419). NCX1, NCX2, NCX3, and NCX4 from Fugu rubripes (Japanese pufferfish) and Danio rerio (zebrafish) were derived from whole genome sequences (2). Acquired NCX sequences were aligned with ClustalX (Version 1.83) (38) using the default parameters. The alignment was then imported into GeneDoc (Version 2.6.002) (39) for manual editing and creation of figures. Neighbor-joining trees were generated using ClustalX, followed by tree evaluation with bootstrap re-sampling (1000 times). The program TreeView (Version 1.6.6) (40) was used to examine and display the resulting phylogram and was rooted with invertebrate CALX (gi:2266953) from fruit fly. The variable alternative splice site of the NCX was not included in subsequent phylogenetic analyses due to the high potential for homoplasy in this region.

RESULTS
Tilapia NCX-TL1.0 cDNA and Deduced Protein Topology-A 3106-bp cDNA clone was isolated from the heart of Mozambique tilapia (GenBank TM accession number AY283779) using a combination of RT-PCR and 5Ј-and 3Ј-RACE. An open reading frame of 2871 bp is initiated by a methionine at nucleotide 59 associated with a partial Kozak initiation site, ACCATGA (with the start codon in italics). This open reading frame encodes a protein of 957 amino acids (designated NCX-TL1.0) with a deduced molecular mass of 107 kDa. The 174-bp 3Јuntranslated region (excluding the stop codon) does not contain a polyadenylation signal.
The deduced amino acid sequence for tilapia NCX-TL1.0 is shown in Fig. 1. The only other teleost NCX to be cloned and characterized, trout NCX-TR1.0 (GenBank TM accession number AF175313), is aligned for comparison of the sequences. The topology of tilapia NCX-TL1.0 is similar to that of trout NCX-TR1.0 (41) and canine NCX1.1 (6) based on hydropathy analysis (42) and prediction of secondary structure using PSI-PRED (43) (data not shown). For tilapia NCX-TL1.0, an N-terminal signal peptide with a cleavage site between residues 31 and 32 was predicted both through homology to experimentally determined NCX signal peptides (44,45) and by the prediction program SignalP (Version 3.0) (46). In the N-terminal region of mature tilapia NCX-TL1.0 (between the signal cleavage site and TMS1), there are two potential N-linked glycosylation sites at positions 10 and 15 (aspargines). If glycosylation of tilapia NCX-TL1.0 and mammalian NCX1 is similar, then Asn 10 in NCX-TL1.0 would be glycosylated; however, this requires experimental conformation. In addition, the cysteines at positions 19 and 25 are homologous to positions 14 and 20 in mammalian NCX1, which are thought to be involved in a disulfide bond with Cys 792 (47). This latter cysteine is located in the extracellular loop between TMS6 and TMS7 and corresponds to Cys 780 in tilapia NCX-TL1.0.
Overall sequence comparison of full-length tilapia NCX-TL1.0 and trout NCX-TR1.0 shows ϳ81% identity at the amino acid level. As expected, identity is especially high in the ␣-repeats (ϳ95%), the XIP site (90%), and Ca 2ϩ -binding domains (100%). Phylogenetic analysis confirmed NCX-TL1.0 as a member of the NCX1 gene family, expressing exons A, C, D, and F in the alternative splice site. The phylogram in Fig. 2 shows the evolutionary relationship of all known NCX isoforms from fish.
Functional Expression of Tilapia NCX-TL1.0 and Exchange Currents-To confirm that tilapia NCX-TL1.0 is functional, we expressed the exchanger in Xenopus oocytes and measured outward Na ϩ /Ca 2ϩ exchange currents in giant excised patches (Fig. 3). With the pipette (extracellular surface) containing 8 mM Ca 2ϩ , currents were activated by switching from a 100 mM Li ϩ to a 100 mM Na ϩ solution applied to the intracellular surface of the excised patch. Tilapia NCX-TL1.0 outward exchange currents showed characteristics similar to those of both trout and mammalian NCX1 isoforms. As shown in Fig. 3A, tilapia NCX-TL1.0 demonstrated positive regulation by Ca 2ϩ i .

At 1 M Ca 2ϩ
i , the outward current increased to a peak value and then decayed slowly in a time-dependent manner indicative of Na ϩ -dependent inactivation (48). Increased regulatory Ca 2ϩ i (10 M) eliminated Na ϩ -dependent inactivation, thereby increasing steady-state current (Fig. 3B). The current-voltage relationship for tilapia NCX-TL1.0 recorded in the presence of 100 mM Na ϩ i and 1 M Ca 2ϩ i is shown in Fig. 3C. To obtain the current-voltage plot, a series of 10-mV voltage steps from a holding potential of 0 mV were applied from Ϫ100 to ϩ100 mV for 10-ms intervals. This voltage clamp protocol was applied in both the presence and absence of Na ϩ to allow for leak subtraction. The current-voltage plot obtained for tilapia NCX-TL1.0 was similar to that for trout NCX-TR1.0.
Arrhenius plots of peak and steady-state Na ϩ /Ca 2ϩ exchange currents of tilapia NCX-TL1.0 and trout NCX-TR1.0 are shown in Fig. 4B. Within the same patch, currents at 7 and 14°C were normalized to that at 30°C. The E a was obtained from the slope of a linear regression fit of the logarithms of normalized currents (Fig. 4C). Previously, we used the E a as an indication of NCX temperature dependence (3,27). The E a values were 53 Ϯ 9 and 67 Ϯ 21 kJ/mol for the tilapia NCX-TL1.0 peak and steady-state currents, respectively. Compared with trout NCX-TR1.0 (peak and steady-state E a ϭ 7.0 Ϯ 2.0 and 6.0 Ϯ 0.1 kJ/mol, respectively), tilapia NCX-TL1.0 exchange activity was highly temperature-dependent. This phenomenon was corroborated by comparison of tilapia NCX-TL1.0 and trout NCX-TR1.0 Q 10 values (Fig. 4D). The Q 10 values for the peak and steady-state currents are 2. The reason for the increase in the value for tilapia NCX-TL1.0 at 7°C is not known, but may be a result of poor fitting due to small absolute currents. Examination of Sequence Differences in the N-terminal Domain- Fig. 5 shows a multiple sequence alignment of the N terminus, previously shown to be the area solely responsible for NCX temperature dependence (27). The alignment contains all known fish species NCX isoforms and uses human NCX isoforms as mammalian representatives. A sequence comparison based on common temperature phenotypes highlights 10 residue positions (Pro 70 , Val 87 , Asn 163 , Leu 186 , Thr 196 , Ala 210 , Val 230 , Leu 238 , Leu 249 , and Phe 259 ) numbered from the tilapia start codon that may confer NCX temperature sensitivity. All residues except Asn 163 (Thr in mammals) and Leu 249 (Phe in mammals) are common between mammalian NCX1.1 and tilapia NCX-TL1.0, but are different in trout NCX-TR1.0. is compared with that of trout NCX-TR1.0 (lower sequence), with dots representing identity between species. Exchanger topology is represented schematically above the sequence alignment. The NCX is modeled to have nine TMSs (gray cylinders) and a signal peptide (white cylinder) that is cleaved during biosynthesis. The ␣-repeat regions (shaded dark gray) show intramolecular sequence similarity and are important for ion translocation. Also shown are important regions of the large intracellular loop: the XIP site, regulatory Ca 2ϩ -binding domains, and an alternative splice site (boxed). Asterisks denote potential N-linked glycosylation sites in the N-terminal region of tilapia NCX-TL1.0.

DISCUSSION
With the underlying motivation of finding a teleost NCX capable of functioning at mammalian core temperatures, we have cloned, expressed, and characterized an NCX from tilapia heart. Hydropathy and secondary structure analyses of tilapia NCX-TL1.0 yielded results similar to those obtained with other NCX1 isoforms. Currently, the NCX is modeled to have nine TMSs, five in the N-terminal domain and four in the C-terminal domain (49,50). These TMSs are important for ion translocation (51) and are separated by a large intracellular loop that confers NCX regulatory properties and alternative splicing. In addition, known NCX1 post-translational modifications, including signal peptide cleavage, N-linked glycosylation, and intramolecular disulfide bond formation, are predicted to be conserved in tilapia NCX-TL1.0. Thus, the overall topology of tilapia NCX-TL1.0 is unlikely to be different from that proposed for NCX1.
Sequence and evolutionary analyses support the assignment of NCX-TL1.0 as a member of the NCX1 subfamily. Fish species genomes have been recently shown to possess up to four separate NCX genes, whereas other vertebrate species have only three NCX genes (2). However, the degree and distribution of NCX gene expression in fish species are largely unknown, with the exception of the trout NCX1 ortholog (NCX-TR1.0), which is known to be expressed in cardiac tissue (41). Phylogenetic analysis places tilapia NCX-TL1.0 as a member of the teleost NCX1 family, with its most closely related isoform being trout NCX-TR1.0. Sequence comparison at the amino acid level shows overall identities of ϳ81% to trout NCX-TR1.0 and ϳ73% to mammalian NCX1. Conversely, comparison of tilapia NCX-TL1.0 protein sequence with other NCX paralog (NCX2, NCX3, and NCX4) sequences yields lower identities in the range of 60 -65%. Consistent with the NCX subfamily as a whole, the tilapia NCX-TL1.0 sequence displays the highest divergence in the N-terminal region, but is highly conserved in the TMSs and regulatory regions (1,2).
Tissue-specific NCX splice variants are created through differential expression of six small exons (A-F) located in the C-terminal region of the large intracellular loop (9,10,52). Exons A and B are mutually exclusive and are used in conjunction with the cassette exons C-F to produce alternatively spliced NCX isoforms (10). In mammals, the cardiac-specific NCX isoform is designated NCX1.1 and uses exons A, C, and D-F. Interestingly, the tilapia NCX-TL1.0 cloned in this study expresses exons A, C, D, and F, the same combination present in trout NCX-TR1.0 cloned from heart tissue (41). Both cardiac-specific NCX isoforms from fish species appear to lack exon E, which is only five amino acids in length. The functional significance of the alternative splice site remains unclear, as it is not essential for ion transport (53). Potential roles of the region, including modulation of NCX current through protein kinase A sensitivity (54,55), Ca 2ϩ -dependent activation (55,56), and Na ϩ inactivation (56,57), have been suggested. However, the specific physiological relevance of exon E is not known, and its absence in the unique splice variant combination of exons A, C, D, and F is thus far restricted to expression in fish hearts (thus named NCX1.0). Based on phylogenetic analyses and examination of exon expression patterns in the alternative splice site, it is clear that we have expressed a cardiac-specific NCX1 from tilapia.
The high sequence identity of tilapia NCX-TL1.0 to trout NCX-TR1.0 in known regulatory regions is reflected in its Na ϩ /Ca 2ϩ exchange current. Tilapia NCX-TL1.0 is positively regulated by Ca 2ϩ , making it similar to trout NCX-TR1.0 (41), mammalian NCX1 (58), and squid NCX-SQ1 (4), but opposite from fruit fly CALX, which is characterized by decreased current in response to increased cytoplasmic Ca 2ϩ (5). At 30°C, the values for trout NCX-TR1.0 and tilapia NCX-TL1.0 (ϳ0.15) are indicative of slower inactivation rates compared with those reported previously for mammalian cloned ( ϳ 0.25) (3) and native ( ϳ 0.23) (60) NCX1.1 outward currents. Na ϩ -dependent inactivation of NCX1 is attributed in part to the highly conserved XIP site, which is composed of 20 residues and is located in the first portion of the intracellular loop just past TMS5 (59). Common to tilapia NCX-TL1.0 and trout NCX-TR1.0 are methionine and arginine at positions 9 and 18, respectively (numbering starting at the XIP site). Mammalian NCX1.1 has valine and glutamine at these respective positions. To our knowledge, these residue positions have not undergone mutational analysis and may be directly or indirectly involved in altering NCX inactivation rates. The physiological significance of fish NCX isoforms having slower Na ϩ -dependent inactivation is not known. We have thus confirmed that tilapia NCX-TL1.0 is functional when expressed in oocytes and that it is regulated in a manner similar to that of trout NCX-TR1.0.
Mammalian cardiac NCX1 activity measured in cardiac myocytes (60, 61), sarcolemmal vesicles (29), and Xenopus oocytes (3) has been shown to be highly temperature-dependent, with E a values in the range of 48 -66 kJ/mol. In contrast, trout NCX-TR1.0 currents exhibit an E a value of ϳ7 kJ/mol when expressed in Xenopus oocytes (3). The relative insensitivity of trout NCX-TR1.0 to temperature compared with mammalian NCX1.1 has been shown to be an intrinsic property of the protein rather than dependent on lipid environment (29). In this study, we measured the temperature dependence of cardiac NCX-TL1.0, cloned from the warm-adapted fish species Mozambique tilapia. This teleost is tolerant of environmental temperatures greater than mammalian body temperature (up to 42°C) and has an optimal growth temperature (ϳ30°C) that is lethal to trout. Conversely, the lower temperature limit of tilapia (ϳ14°C) coincides with the optimal temperature for trout survival. The relatively high E a value (ϳ60 kJ/mol) we recorded for tilapia NCX-TL1.0 in giant excised patches is consistent with the environmental temperature range of tilapias and is a typical value for a warm-adapted species. To our knowledge, this is the first NCX from non-mammalian species shown to display mammal-like temperature dependence. The only other non-mammalian NCX temperature dependence reported is for the frog NCX, which has intermediate tempera-ture sensitivity, with E a ϳ 25 kJ/mol (62).
Elucidating the specific molecular determinants of NCX temperature dependence has proven a difficult undertaking. Previous chimeric studies using mammalian NCX1.1 and trout NCX-TR1.0 have shown that the region responsible for the differential temperature dependence between these isoforms is attributable to the N-terminal hydrophobic domain (27). This includes the XIP region and makes up only one-quarter of the mature protein, consisting of ϳ240 residues. There are 50 substitutions between trout NCX-TR1.0 and tilapia NCX-TL1.0 in this region, with over half of these substitutions residing in the highly variable N terminus before the putative TMS1. The rest of the hydrophobic region (TMS1 to the end of the XIP site) is highly conserved, with only 22 substitutions. It is generally accepted that proteins adapted to cold temperatures, such as trout NCX-TR1.0, gain catalytic efficiency through increased flexibility in structural moieties involved in i dependence of peak and steady-state outward currents. Currents were normalized to the largest peak or steady-state current attained within the same patch. In the absence of regulatory Ca 2ϩ i , peak and steady-state currents were minimal and not well defined. Therefore, the maximal current in the trace was taken as the peak value, and the current measured at the end of the 32-s trace was taken as the steadystate value. *, statistically significant differences (p Ͻ 0.05) between mean values with respect to that in the presence of 10 M Ca 2ϩ i ; **, significant difference between mean values found at 0 and 1 M regulatory Ca 2ϩ i . C, leak-subtracted current-voltage relationship obtained in the presence of 100 mM Na ϩ i and 1 M Ca 2ϩ i . This relationship was obtained using 10-ms voltage steps from Ϫ100 to 100 mV in 10-mV increments from a holding potential of 0 mV. i at the temperatures indicated. B, Arrhenius plots of the peak and steady-state (SS) currents for tilapia NCX-TL1.0 (circles) and trout NCX-TR1.0 (squares). For tilapia NCX-TL1.0, data were pooled from seven patches, whereas for trout NCX-TR1.0, the points are as previously measured. Within the same patch, the currents at 7 and 14°C were normalized to that at 30°C, which was the highest temperature studied. C, E a values obtained from the slope of a linear regression fit of the logarithms of normalized peak (white bars) and steadystate (black bars) currents. *, statistically significant differences (p Ͻ 0.05) between the mean peak and steady-state current E a values for trout NCX-TR1.0 and tilapia NCX-TL1.0. D, Q 10 values for the peak (white bars) and steady-state (black bars) currents for the 7-30°C interval. *, statistically significant differences (p Ͻ 0.001) between the mean peak and steady-state current Q 10 values for trout NCX-TR1.0 and tilapia NCX-TL1.0. E, temperature dependence of the current decay rates () for trout NCX-TR1.0 (white bars) and tilapia NCX-TL1.0 (black bars). Analysis of variance and mean comparison using Tukey's test did not reveal significant differences at the temperatures indicated. catalysis (Refs. 63 and 64; for review, see Refs. [65][66][67]. The opposite also holds true in that adaptation to high temperatures is postulated to involve increased rigidity for thermal stability. Unfortunately, this relationship between flexibility and stability does not allow prediction of single amino acid effects on protein temperature dependence. There is no absolute consensus as to which residues modify protein flexibility, suggesting that evolution has utilized numerous strategies to FIG. 5. Multiple sequence alignment of NCX N-terminal domain. A, multiple sequence alignment of the NCX N-terminal domain from representative species. Sequences were aligned with ClustalX (Version 1.83) using the default parameters and imported into GeneDoc (Version 2.6.002) for manual editing. The sequences of human NCX1.1 (ncx1 mammalian), human NCX2.1 (ncx2 mammalian), human NCX3.3 (ncx3 mammalian), rainbow trout NCX-TR1.0 (ncx1 trout), and green spotted pufferfish NCX (ncx4 gr puffer) were obtained from the NCBI Nonredundant Protein Database. NCX1, NCX2, NCX3, and NCX4 from F. rubripes (ncx1-4 fugu) and D. rerio (ncx1-4 danio) were derived previously from whole genome sequences (2). Due to almost complete identity in this region among mammals, human NCX sequences were used as representatives. Residues thought to compose the signal peptide are shown in italics, and asparagine residues that are part of a potential glycosylation site (NX(S/T)X, where X is not Pro) are shown in boldface. Residue positions that potentially confer NCX temperature dependence are boxed and numbered. ncx1 tilapia, tilapia NCX-TL1.0. B, approximate locations of the boxed residues in a topological model of the NCX N-terminal domain. SP, signal peptide; XIP, exchanger inhibitory peptide. confer function over a range of temperatures. In addition, it is not known if membrane proteins (which may make up ϳ30% of the proteome of an organism) employ the same adaptations to temperature as do soluble proteins. The vast majority of functional data examining protein temperature adaptation are derived from soluble proteins (64, 68 -70) and, in the case of comparative proteomic studies (71)(72)(73), fail to differentially treat soluble and membrane proteins in their analysis. These caveats advocate the study of membrane protein temperature dependence on a case-by-case basis, preferably using orthologs that are highly conserved. Tilapia NCX-TL1.0 is potentially a very good model to study NCX temperature dependence because it is evolutionarily close to trout NCX-TR1.0, but displays a mammalian temperature dependence phenotype.
A simple sequence analysis comparing amino acids that are common to tilapia NCX-TL1.0 and mammalian NCX1.1 but are different in trout NCX-TR1.0 yielded 10 residue positions in the N-terminal domain that may alter exchanger flexibility (Fig. 5). These highlighted substitutions are spread throughout the N terminus and are not localized to a single area in the primary sequence; however, because little of the NCX tertiary structure is known, it is possible that some of these residues are in close proximity in the tertiary structure. All residues except Asn 163 and Leu 249 (numbering is from the tilapia NCX-TL1.0 start codon) are identical between the tilapia and mammalian NCX, but are different in trout NCX-TR1.0. Tilapia Asn 163 (Asp in trout and Thr in mammals) is predicted to be in the re-entrant loop of the ␣ 1 -repeat, spanning TMS2 and TMS3. The TMS of the ␣ 1 -repeat is known to be important for ion translocation and is highly conserved, whereas the reentrant loop shows more variability among NCX isoforms (2) and has a less defined role in ion exchange. Recent mutational analysis at this position showed no effect on ion translocation or intracellular Na ϩ affinity, arguing against the role of Asn 163 in the ion translocation mechanism (74). Tilapia Leu 249 (Met in trout and Phe in mammals) is located in TMS5 and is implicated in sensitivity to the NCX inhibitor SEA0400 (75). The remaining tilapia NCX-TL1.0 residues (Pro 70 , Val 87 , Leu 186 , Thr 196 , Ala 210 , Val 230 , Leu 238 , and Phe 259 ) are identical to NCX1.1. Phe 259 has been shown to modify Na ϩ -dependent inactivation (59), but the remaining residues have not undergone mutational analysis and have unknown roles in NCX function. The most interesting of these positions is tilapia Pro 70 , located in the N terminus just upstream of TMS1. In all vertebrate species, proline is located at this position, except in trout NCX-TR1.0, which has threonine. This substitution introduces a potential N-linked glycosylation site (consensus NX(S/T)X, where X is any residue except Pro) in trout NCX-TR1.0 that is not found in any other vertebrate NCX isoform. Glycosylation in mammalian NCX1 occurs ϳ23 residues upstream from this position and has been experimentally determined not to modify exchange based on the parameters measured to date (45). However, the effect of glycosylation on NCX temperature dependence is not known. The addition of a glycosylation site in trout NCX-TR1.0 close to TMS1 raises the intriguing possibility of glycosylation being involved in the temperature dependence of the NCX.
In general, the high variability among NCX orthologs and paralogs at these positions is consistent with the fact that none of the residues has been shown to be important for NCX ion translocation. It is likely that the NCX contains two types of amino acid residues: one group that has been conserved throughout evolution and that is critical for the core function of ion translocation and regulation and another group that is not essential for NCX function but rather confers the subtle "tweaking" of the protein to function optimally under its respective environmental conditions. NCX temperature dependence is therefore dictated by a series of substitutions in positions that may or may not be common among homologs, indicating that each isoform has evolved independent mechanisms to adapt to temperature. Furthermore, these substitutions most likely modify the flexibility of NCX core regions in an allosteric manner that allows function at different temperatures.
In summary, we have cloned and characterized NCX-TL1.0 from tilapia heart and shown that it displays mammal-like temperature dependence. Tilapia NCX-TL1.0 is a potentially good model to study the temperature dependence of the NCX and should lead to a better understanding of NCX function in general.