Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

The mammalian Na⁺/Ca²⁺ exchanger forms a multigene family comprising three isoforms (NCX1, NCX2, and NCX3) that shares ~70% identity in the overall amino acid sequences (Nicoll et al., 1990; Li et al., 1994; Nicoll et al., 1996b). NCX1 is highly expressed in heart and brain and at lower levels in many other tissues, whereas expression of the other isoforms is limited mainly to brain. In heart, NCX1 plays the primary role in the extrusion of cytosolic Ca²⁺ from myocytes during each heart beat, whereas the contribution of NCX isoforms in Ca²⁺ handling in other tissues still remains to be precisely defined (Blaustein and Lederer, 1999).

Recent topological studies (Nicoll et al., 1999; Iwamoto et al., 1999a) have suggested that the mature NCX1 protein consists of nine transmembrane segments (TMs) and a large intracellular loop between putative TM5 and TM6, with the N and C termini localized on the extracellular and cytoplasmic sides, respectively (see Fig. 1A). In all known members of the Na⁺/Ca²⁺ exchanger family, there exist highly conserved internal repeat sequences designated the -1 and -2 repeats in the N- and C-terminal halves of the exchanger molecule, respectively (Schwarz and Benzer, 1997). Similar intragenic repeat structures were found in the Na⁺/Ca²⁺/K⁺ exchanger (Reiländer et al., 1992; Tsoi et al., 1998) and the recently identified Mg²⁺/H⁺ exchanger (Shaul et al., 1999). In NCX1, the -1 repeat comprises most of TM2-TM3 and their connecting loop, whereas the -2 repeat comprises putative TM7 and its C-terminal segment (see Fig. 1A). Our recent study has provided evidence that the putative loop connecting TM2-TM3 forms a re-entrant membrane loop with both ends facing the extracellular side, whereas the -2 repeat, except for TM7, forms a domain mostly accessible from the cytoplasm (Iwamoto et al., 1999a). Mutational analyses have further suggested that these putative TMs and loops in the -1 and -2 repeats are important in the interaction of the exchanger with transport substrates, a nonselective cation inhibitor (Ni²⁺), and an activator (Li⁺) (Nicoll et al., 1996a; Doering et al., 1998; Iwamoto et al., 1999b).

View larger version (42K): [in this window] [in a new window]   Fig. 1.   A topological model of NCX1 (A) and chimeric constructs between NCX1 and NCX3 (B). A, a nine-TM model of NCX1 based on the recent topology analyses (Nicoll et al., 1999; Iwamoto et al., 1999a). The amino acid residues of NCX1 whose mutation alters the KB-R7943 sensitivity are shown. B, two series of chimeras (N1 and N3) were constructed by substituting segments of NCX1 with homologous segments of NCX3, and vice versa. In each series, chimeras are named based on the amino acid numbers for respective isoforms. Broken lines show the restriction enzyme cut sites and the black boxes indicate the substituted segments. The linear model of exchanger is indicated at the top.

Iwamoto et al. (1996b) and Watano et al. (1996) have recently reported synthesis and characterization of a potent inhibitor of the cardiac Na⁺/Ca²⁺ exchanger, KB-R7943. This compound is highly selective for the Na⁺/Ca²⁺ exchanger, because at up to 10 µM, it exerted little influence, not only on many other ion transporters but also on several cardiac action potential parameters, diastolic [Ca²⁺]_i, and spontaneous beating in cardiomyocytes (Iwamoto et al., 1996b). It is effective in inhibiting all three NCX isoforms (Iwamoto and Shigekawa, 1998). It is now being widely used to study the physiological and pathological roles of the exchanger at the cellular and organ levels. For example, the compound has been shown to efficiently protect the ischemia-associated reperfusion injury of heart (Iwamoto et al., 1996b; Nakamura et al., 1998; Ladilov et al., 1999; Mukai et al., 2000; Satoh et al., 2000), brain (Schröder et al., 1999; Breder et al., 2000), and kidney (Kuro et al., 1999), suggesting its therapeutic potential in this type of cell damage. More recently, however, this compound has been reported to inhibit neuronal nicotinic acetylcholine receptors (Pintado et al., 2000) and N-methyl-D-aspartate receptor channels (Sobolevsky and Khodorov, 1999), although evidence has been presented that it does not inhibit the latter directly (Hoyt et al., 1998). An interesting feature of the drug is that it inhibits the Ca²⁺ influx mode (reverse mode) of Na⁺/Ca²⁺ exchange much more effectively than the forward mode (Iwamoto et al., 1996b; Watano et al., 1996; Satoh et al., 2000), although the reason for such difference is currently unclear. Another interesting feature is that it is 3-fold more effective on NCX3 than NCX1 or NCX2 (Iwamoto and Shigekawa, 1998). Currently, however, little is known about the molecular mechanism of action for this drug and location of its receptor site in the exchanger molecule.

In this study, to obtain insights into the mechanism of action of KB-R7943 as well as into the mechanism of Na⁺/Ca²⁺ exchange, we constructed a series of chimeric exchangers between NCX1 and NCX3 and searched for the structural domain(s) of the exchanger involved in the determination of sensitivity to KB-R7943. We identified several amino acid residues within the -2 repeat of the exchanger whose mutations alter apparent affinity of the drug.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Cell Cultures. CCL39 cells and NCX transfectants were maintained in Dulbecco's modified Eagle's medium supplemented with 7.5% heat-inactivated fetal calf serum, 50 U/ml penicillin, and 50 µg/ml streptomycin.

Construction and Stable Expression of Wild-Type, Chimeric, and Mutant Exchangers. We used major splice variants of cDNAs for NCXs and NCKX that are prevalent in dog heart (NCX1.1) and rat brain (NCX2.1, NCX3.3, and NCKX2) (Quednau et al., 1997; Tsoi et al., 1998). NCX cDNAs were cloned into SacII and HindIII restriction sites in pCRII (Invitrogen, San Diego, CA) or in the mammalian expression vector pKCRH as described previously (Iwamoto et al., 1996a; Iwamoto and Shigekawa, 1998). For isolation of cDNA of NCKX2, total RNA was extracted from rat brain with a TRIzol reagent (Life Technologies, Rockville, MD) and first-strand cDNAs were synthesized using an oligo(dT) primer and SuperScript preamplification system (Life Technologies). cDNAs were amplified by polymerase chain reaction using pairs of sense and antisense primers, 5'-ATATGAATTCACCACCATATCCACCAGAAGATCC-3' and 5'-ATATGGTACCGTTTAAGATATGGCTTTTCCACTA-3' [nucleotides 24- 1 and 2011-2034 of NCKX2 (Tsoi et al., 1998)] containing exogenous EcoRI and KpnI restriction sites, respectively (underlined). Amplified cDNA insert was cloned into EcoRI and KpnI restriction sites in the mammalian expression vector pcDNA3.1 (Invitrogen).

Assay of ⁴⁵Ca²⁺ Uptake. Na⁺_i-dependent ⁴⁵Ca²⁺ uptake into cells expressing the wild-type or mutated NCXs were assayed as described in detail previously (Iwamoto and Shigekawa, 1998). Briefly, confluent transfectants in 24-well dishes were loaded with Na⁺ by the incubation at 37°C for 30 min in 0.5 ml of balanced salt solution (BSS) (10 mM HEPES/Tris, pH 7.4, 146 mM NaCl, 4 mM KCl, 2 mM MgCl₂, 0.1 mM CaCl₂, 10 mM glucose, and 0.1% bovine serum albumin) containing 1 mM ouabain and 10 µM monensin. ⁴⁵Ca²⁺ uptake was then initiated by switching the medium to Na⁺-free BSS (replacing NaCl with equimolar choline chloride) or to normal BSS, both of which contained 0.1 mM ⁴⁵CaCl₂ (1.5 µCi/ml) and 1 mM ouabain. After a 30-s incubation, ⁴⁵Ca²⁺ uptake was terminated by washing cells four times with an ice-cold solution containing 10 mM HEPES/Tris, pH 7.4, 120 mM choline chloride, and 10 mM LaCl₃. Cells were then solubilized with 0.1 N NaOH, and aliquots were taken for determination of radioactivity and protein. When present, KB-R7943 was included in medium at concentrations of up to 100 µM two min before the start of ⁴⁵Ca²⁺ uptake. Assay of Na⁺_i/K⁺_o-dependent ⁴⁵Ca²⁺ uptake into cells expressing the wild-type NCKX2 was performed under the same condition as that for NCX.

Measurement of Whole-Cell Outward Current. Outward exchange currents from NCX1 transfectants were measured using the whole-cell, patch-clamp technique as described previously (Uehara et al., 1997). Bath solution contained 150 mM LiCl (replacing NaCl), 1 mM MgCl₂, 0 or 1 mM CaCl₂, 20 µM ouabain, 2 µM nicardipine, 5 µM ryanodine, 0 to 10 µM KB-R7943, and 5 mM HEPES, pH 7.2, whereas pipette solution contained 20 or 100 mM NaOH, 20 mM CsOH, 1.1 mM MgCl₂, 20 mM tetraethylammonium chloride, 2 mM MgATP, 2 mM creatine phosphate, 19.8 mM CaCl₂, 50 mM EGTA, 0 to 30 µM KB-R7943, and 50 mM HEPES, pH 7.2. The ionized Ca²⁺ concentration in pipette solution was calculated to be 0.16 µM. The outward current was activated by switching bath solution from one without CaCl₂ to one with CaCl₂. Intracellular application of KB-R7943 was performed by perfusion of pipette solution by the method described by Soejima and Noma (1984). All experiments were performed at about 35°C and the holding and test potentials were 40 mV. All data were acquired and analyzed by the pCLAMP software (Axon Instruments, Foster City, CA).

Statistical Analysis. Data are expressed as means ± S.E. of three independent determinations. Differences for multiple comparisons were analyzed by unpaired t test or one-way analysis of variance followed by the Dunnett's test. Values of p < 0.05 were considered statistically significant.

Materials. Chinese hamster lung fibroblasts (CCL39) were purchased from American Type Culture Collection (Manassas, VA). KB-R7943 was synthesized by New Drug Research Laboratories, Kanebo (Osaka, Japan). ⁴⁵CaCl₂ was purchased from Amersham Pharmacia Biotech (Buckinghamshire, UK). All other chemicals were of the highest grade available.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Chimera Analysis of Effect of KB-R7943. KB-R7943 inhibited the initial rate of Na⁺_i-dependent ⁴⁵Ca²⁺ uptake into NCX1-, NCX2-, or NCX3-expressing cells with IC₅₀ values of 4.3 ± 0.3, 4.7 ± 0.5, or 1.4 ± 0.2 µM (n = 3), respectively (see Fig. 5), confirming our previous report (Iwamoto and Shigekawa, 1998) that NCX3 was about 3-fold more sensitive to KB-R7943 than other isoforms. Taking advantage of the fact that NCX1 and NCX3 exhibit a high sequence homology, we constructed chimeric exchangers between these isoforms to identify region(s) of the exchanger that is important for the interaction with KB-R7943. We constructed two series of chimeras in which one or two segments from NCX3 were transferred into NCX1 in exchange for the homologous segment(s) in the latter (N1 chimeras), and vice versa (N3 chimeras) (Fig. 1B). All these chimeras had exchange activities similar to that of the wild-type NCX1 or NCX3 (see the legend to Fig. 2).

View larger version (24K): [in this window] [in a new window]   Fig. 2.   Effects of KB-R7943 on Na+i-dependent 45Ca2+ uptake into cells expressing N1 (A) and N3 (B) chimeric exchangers. The initial rate of 45Ca2+ uptake was measured in the presence or absence of 3 µM KB-R7943 as described under Experimental Procedures. The uptake rates of cells expressing these chimeras were 5 to 9 nmol/mg/30 s in the absence of KB-R7943, which were similar to those in cells expressing the wild-type NCX1 or NCX3 (10 and 7 nmol/mg/30 s). Data are presented as percentage of the values obtained in the absence of drug. Data are presented as mean ± S.E. of three independent experiments. *p < 0.05 versus NCX1 (in A); *p < 0.05 versus NCX3 (in B).

Identification of Residues Influencing KB-R7943 Sensitivity. Fig. 3A shows amino acid sequences of the -2 repeat regions from NCX isoforms and NCKX2 (Nicoll et al., 1990; Li et al., 1994; Nicoll et al., 1996b; Tsoi et al., 1998). In these regions of NCX1 and NCX3, there are only four residues unique to each isoform. To identify the residues that influence KB-R7943 sensitivity, the unique residues were exchanged between NCX1 and NCX3 or mutated to other amino acids. We found that single mutations Leu808-to-Phe and Thr823-to-Leu in NCX1 and their reciprocal mutations in NCX3 did not produce any change in sensitivity to inhibition by 3 µM KB-R7943 (Fig. 3, B and C). The Val820-to-Ala mutation in NCX1 was also silent, but its reciprocal mutation in NCX3 (N3-A809V) significantly decreased KB-R7943 sensitivity. On the other hand, the Gln826-to-Val mutation in NCX1 increased KB-R7943 sensitivity of the exchanger to a level comparable with that seen in the wild-type NCX3, whereas its reciprocal mutation in NCX3 (N3-V815Q) was silent. In contrast, simultaneous mutations of Val820 to Ala and Gln826 to Val in NCX1 and their reciprocal mutations in NCX3 mostly reproduced the drug sensitivities seen in the parental NCX3 and NCX1, respectively (Fig. 3, B and C). We found that the Val820-to-Gly mutation in NCX1 and the Ala809-to-Ile mutation in NCX3, but not their reciprocal mutations in the other isoforms, produced alterations in drug sensitivity almost comparable with those seen between the parental exchangers. Other mutations, such as Gln826 to Glu or Arg in NCX1 and their reciprocal mutations in NCX3, were silent. Therefore, residues at positions 820 and 826 of NCX1 are mostly responsible for the differential drug response between NCX1 and NCX3.

View larger version (33K): [in this window] [in a new window]   Fig. 3.   Amino acid alignment of -2 repeat regions in NCX isoforms and NCKX2 (A) and effects of KB-R7943 on Na+i-dependent 45Ca2+ uptake into cells expressing NCX variants mutated in the -2 repeat (B and C). A, conserved amino acids among exchangers are indicated with dots and alignment gaps with dashes. Boxes show the amino acid positions of NCX1 and NCX3 whose mutation alters the KB-R7943 sensitivity. B and C, we constructed mutants in which the amino acid residues in the -2 repeat unique to each NCX isoform were exchanged between NCX1 and NCX3 or mutated to other residues. The initial rates of 45Ca2+ uptake into cells expressing NCX1 mutants (B) or NCX3 mutants (C) were measured as described in Fig. 2. The uptake rates in these mutants were 3 to 9 nmol/mg/30 s in the absence of KB-R7943. Data are presented as percentage of the values obtained in the absence of drug. Data are presented as means ± S.E. of three independent experiments. *p < 0.05 versus NCX1 (in B); versus NCX3 (in C).

View larger version (33K): [in this window] [in a new window]   Fig. 4.   Effect of KB-R7943 on activities of NCX1 mutants carrying single cysteine substitution in the -2 repeat. Initial rates of Na+i-dependent 45Ca2+ uptake were measured in the presence or absence of 10 µM KB-R7943; otherwise, details as in Fig. 2. Data are presented as mean ± S.E. (n = 3). *p < .05 versus wild-type NCX1.

View larger version (29K): [in this window] [in a new window]   Fig. 5.   Dose-response curves for the effects of KB-R7943 on Na+i-dependent 45Ca2+ uptake into cells expressing wild-type NCX isoforms, several important NCX1 mutants, and NCKX2. The initial rates of 45Ca2+ uptake were measured in the presence or absence of indicated concentrations of KB-R7943. Data are means ± S.E. of three independent experiments.

View larger version (29K): [in this window] [in a new window]   Fig. 6.   Effects of multiple amino acid substitutions at Gly833 of NCX1 on dose-responses for KB-R7943. The initial rates of Na+i-dependent 45Ca2+ uptake into cells expressing wild-type NCX1 and substitution mutants were measured in the presence or absence of indicated concentrations of KB-R7943; otherwise, details as in Fig. 2. Data are presented as means ± S.E. of three independent experiments.

View this table: [in this window] [in a new window]   TABLE 1 Kinetic parameters as determined from double-reciprocal plot analyses of Ca2+-dependences of the rate of Na+i-dependent Ca2+ uptake into cells expressing NCX1, NCX3, or NCX1 mutants in the presence or absence of KB-R7943. The initial rates of 45Ca2+ uptake were measured in the presence of 0.067 to 2 mM CaCl2 and at 0 to 10 µM KB-R7943 and analyzed as shown previously (Iwamoto and Shigekawa, 1998). Values are presented as means ± S.E. (n = 3).

Sidedness of Effect of KB-R7943. Fig. 7, A and B, shows the results from electrophysiological measurements in which sidedness of action of KB-R7943 was examined. We measured whole-cell outward exchange current evoked by the extracellular application of 1 mM Ca²⁺ in cells expressing the wild-type NCX1 preincubated with or without external 0.8 µM KB-R7943 or perfused internally with or without 30 µM KB-R7943. These outward currents were never observed in nontransfected CCL39 cells (Uehara et al., 1997). When applied externally for 6 min, 0.8 µM KB-R7943 significantly reduced the current; its initial peak was reduced to 39 ± 2.3% (n = 3) of the control (Fig. 7A). When cells were washed with drug-free solution, the current was recovered to the control level. In contrast, KB-R7943 produced no significant inhibition of the outward current when it was applied intracellularly for 15 min (Fig. 7B). These results suggest that KB-R7943 inhibits the exchanger from the extracellular side. The IC₅₀ value of the wild-type NCX1 for external KB-R7943 obtained from these electrophysiological measurements was 0.89 ± 0.23 µM (n = 4), which is significantly lower than that from ⁴⁵Ca²⁺ uptake measurements (Fig. 5). A similarly low IC₅₀ value was reported by Watano et al. (1996), who used electrophysiological methods. The reason for such difference in the IC₅₀ is not clear, but could be caused by differences in the conditions used, such as the preincubation time.

View larger version (20K): [in this window] [in a new window]   Fig. 7.   Effect of externally or internally applied KB-R7943 on the whole-cell exchange current from cells expressing the wild-type NCX1. A, the outward current was evoked successively in the same cell by external application of 1 mM Ca2+; the cell was placed in drug-free bath solution and then in bath solution containing 0.8 µM KB-R7943 for 6 min, followed by washing for 6 min with drug-free solution. B, a cell was perfused internally with drug-free pipette solution and then with 30 µM KB-R7943-containing solution for 15 min by the method of Soejima and Noma (1984). C, the outward current was evoked by 1 mM external Ca2+ for 20 s. At intervals of 30 s, the current was successively evoked by external application of 1 mM Ca2+ plus 10 µM KB-R7943. After washing for 5 min, the current was evoked again by 1 mM external Ca2+. Similar results were obtained from three to five independent experiments.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

In this study, taking advantage of our previous observation that KB-R7943 is 3-fold more effective in inhibiting NCX3 than NCX1 or NCX2 (Iwamoto and Shigekawa, 1998), we initially employed a chimera strategy to identify structural domain(s) of the exchanger responsible for the difference in the drug sensitivity of isoforms. Analyses with N1 and N3 chimeras, in which homologous segments were transferred individually or in combination from NCX3 into NCX1 and vice versa, revealed that a NheI-MluI segment corresponding to amino acids 788-829 of NCX1 or 777-818 of NCX3 was fully responsible for the difference in drug sensitivity between the two isoforms (Fig. 2). This segment contains a large fraction of the -2 repeat that is highly conserved in all homologs of the Na⁺/Ca²⁺ exchanger family (Schwarz and Benzer, 1997).

By exchanging four different amino acid residues of the -2 repeat between NCX1 and NCX3 (Fig. 3A) and testing the resulting mutants for altered drug sensitivity, we found that NCX1 mutants N1-Q826V and N1-V820A/Q826V and NCX3 mutants N3-A809V and N3-A809V/V815Q were able to reproduce ~70 to 85% of the difference in KB-R7943 response observed between the parental exchangers (Fig. 3, B and C). This suggests that Val820 and Gln826 of NCX1 and the corresponding Ala809 and Val815 of NCX3 play predominant roles in generating the differential drug responses of exchangers. On the other hand, although other mutants N1-V820G and N3-A809I exhibited similar alterations in drug sensitivity, N1-V820A, N1-V820I, N3-A809G, and N3-V815Q-A were silent. Because these mutations were silent and because the effective mutations produced only a small (3-fold) change in drug sensitivity, it seems likely that Val820 and Gln-826 of NCX1 (and the corresponding residues in NCX3) do not interact directly with KB-R7943 with their mutations allosterically influencing binding of the drug to its receptor.

To identify other residues that possibly interact with KB-R7943 but cannot be detected in the chimera study, we screened cysteine-substituted NCX1 mutants of most residues in the -2 repeat and about half of residues in the -1 repeat for altered drug sensitivity. We found that N1-G833C alone exhibited a much-reduced affinity for the drug compared with the wild-type NCX1 (Fig. 4). We made multiple amino acid substitutions at this position with relatively small residues except Lys (Fig. 6). We found that Thr produced the largest effect on the IC₅₀ value, increasing it by much more than 30-fold. Lys was almost as effective as Cys in reducing drug sensitivity, although Asp did not produce any effect. Interpretation of these data is not simple, because single factors, such as the residue size, do not seem to be able to explain them. The observed difference between Lys and Asp may suggest the involvement of electrostatic interaction and may be consistent with our previous report that the positive charge on the isothiourea moiety of KB-R7943 is important for the high inhibitory potency (Iwamoto et al., 1996b). The marked reduction in KB-R7943 sensitivity caused by mutation at Gly833 suggests that this residue may participate in the formation of the drug receptor.

Fig. 7 has provided strong pieces of evidence that KB-R7943 inhibits Na⁺/Ca²⁺ exchanger only from the external side in intact CCL39 cells. We confirmed the same side-dependent effect using the Xenopus laevis oocyte expression system, in which up to 100 µM KB-R7943 was injected into oocytes or added extracellularly to examine its effect on Na⁺_i-dependent ⁴⁵Ca²⁺ uptake (S. Kita, T. Iwamoto, and M. Shigekawa, unpublished observations). Such absence of effect of intracellular KB-R7943 was also observed in cardiomyocytes by Watano et al. (1996) (see under Discussion in this reference). However, we also suggested previously that the drug may inhibit the exchanger from the cytoplasmic side, because Na⁺_i-dependent ⁴⁵Ca²⁺ uptake into cardiac sarcolemmal vesicles with predominantly (~70%) inside-out orientation was inhibited almost completely by KB-R7943 (Iwamoto et al., 1996b). Consistent with the latter finding, the exchange current of NCX1 expressed in X. laevis oocytes, as measured using the giant excised patch technique, was reported to be inhibited by KB-R7943 applied to the cytoplasmic surface (Elias et al., 2000). Why the exchanger is inhibited from the external side in intact cells is unclear at present. One possibility could be that intimate interactions between the membrane and submembrane cytoskeleton existing in intact cells are disrupted in isolated membrane or excised patch preparations, which might result in alteration of the cytoplasmic surface of the exchanger, thereby exposing a binding site for the drug. Another possibility is that the intracellular drug concentration may not reach a level sufficient to inhibit the exchanger, possibly because of binding or sequestration of the drug by cytoplasmic proteins. Because we perfused the cell interior with high concentrations of drug, we feel that the latter possibility is unlikely.

From substituted cysteine scanning analysis of NCX1 topology using the inhibition of exchange activity by thiol-modifying probes as the marker for accessibility (Karlin and Akabas, 1998), we found that cysteine at position 833 was accessible to internally but not externally applied 2-trimethylammonioethylmethane thiosulfonate, suggesting that Gly833 is exposed on the cytoplasmic side of the membrane (see Iwamoto et al., 1999a and also Fig. 1A). Val820 and Gln826 also seem to be located on the cytoplasmic side, because cysteines incorporated at positions 818, 821, 825, and 827 of NCX1 were accessible to internal 2-trimethylammonioethylmethane thiosulfonate (Iwamoto et al., 1999a). In our more recent study, we have found that D825C and N842C are also accessible to external smaller probes, transitional metal ions and 2-aminoethylmethane thiosulfonate, respectively (Iwamoto et al., 2000). It seems, therefore, that the C-terminal region of -2 repeat is part of a reentrant membrane loop with both ends facing the cytoplasmic side. We speculate that the segment containing Gly833 undergoes alternate exposure to cytoplasmic or extracellular medium via the conformation change of the exchanger during Na⁺/Ca²⁺ exchange. On the basis of a large reduction of drug sensitivity in N1-G833C or N1-G833T (Fig. 6) and the above topological consideration, we hypothesize that the segment containing Gly833 forms part of the drug receptor site or at least is near it. At present, however, we cannot rule out the possibility that mutations at Gly833 located on the cytoplasmic side may allosterically alter the conformation of the external KB-R7943 binding site, thereby reducing its affinity for the drug.

NCX isoforms exhibit difference in other pharmacological properties; divalent cation Ni²⁺ inhibits exchange activity of NCX1 or NCX2 10-fold more potently than that of NCX3, whereas stimulation of exchange by monovalent cation Li⁺ is significantly greater in NCX2 and NCX3 than in NCX1 (Iwamoto and Shigekawa, 1998). Ni²⁺ and Mg²⁺ inhibit Na⁺/Ca²⁺ exchange competitively with extracellular Ca²⁺ for the transport site (Kimura, 1996; Iwamoto and Shigekawa, 1998). Li⁺, on the other hand, seems to accelerate the exchange by increasing the V_max (Iwamoto and Shigekawa, 1998). In our previous study, in which we used the same chimera strategy as that used here, we showed that residues corresponding to Asn125 and Thr127 in the -1 repeat and Val820 in the -2 repeat of NCX1 are involved in the generation of most of the difference in sensitivity to Ni²⁺ between NCX isoforms (Iwamoto et al., 1999b). In the same study, we also showed that residues corresponding to Val820 and Gln826 of NCX1 are responsible for the differential sensitivity of NCX isoforms to Li⁺. Because Val820 and Gln826 of NCX1 were shown here to alter KB-R7943 affinity, it is intriguing to note that double substitution at positions 820 and 826 of NCX1 with corresponding residues from NCX3 (i.e., N1-V820A/Q826V) is able to generate simultaneous changes in responses to different modifiers.

As discussed briefly above, our recent topological study of NCX1 has provided evidence suggesting that the loop connecting the putative TM2 and TM3 in the -1 repeat forms a re-entrant membrane loop, with both ends facing the extracellular side, whereas the C-terminal portion of the -2 repeat also forms a reentrant membrane loop domain largely accessible from the cytoplasm (Iwamoto et al., 1999a, 2000) (see Fig. 1A). We have recently observed that site-directed mutations of highly conserved aspartic acids localized in the loop domains of  repeats (Asp130 in the -1 repeat and Asp825 and Asp829 in the -2 repeat) cause up to 6-fold reduction in the apparent affinity of NCX1 for the transport substrate extracellular Ca²⁺ (Iwamoto et al., 2000). Therefore, the important residues whose mutations alter apparent affinities of the exchanger for inhibitors (Ni²⁺ and KB-R7943), an activator (Li⁺), and the transport substrate are all located in the putative loop domains of the -1 and -2 repeats. It seems, then, that the putative loop regions of  repeats may form parts of the ion translocation pathway of the exchanger. KB-R7943 seems to block this pathway by interacting with a receptor site in the -2 repeat loop domain.