Topology of a Functionally Important Region of the Cardiac Na+/Ca2+ Exchanger*

The cardiac Na+/Ca2+ exchanger, NCX1, has been modeled to consist of 11 transmembrane segments and a large cytoplasmic loop (loop f). Cysteine mutagenesis and sulfhydryl modification experiments demonstrate that the loop connecting transmembrane segments 1 and 2 (loop b) is located on the cytoplasmic side of the membrane, as previously modeled. A mutation in loop b, asparagine 101 to cysteine (N101C), renders the exchanger insensitive to regulation by cytoplasmic Na+ and Ca2+. Nearby mutations at residue threonine 103 (T103C or T103V) increase the apparent affinity of the exchanger for cytoplasmic Na+ and also produce a significant Li+ transport capacity. The evidence suggests that the region at the interface of cytoplasmic loop b and transmembrane segment 2 is important in Na+ transport and also in secondary regulation. Thus, this region may form part of the link between the ion translocation pathway formed by the transmembrane segments and regulatory sites that have previously been localized to loop f.

The Na ϩ /Ca 2ϩ exchanger transports one Ca 2ϩ across the cell membrane in exchange for three Na ϩ and is important in maintaining Ca 2ϩ homeostasis in many cell types, including muscle, nerve, and epithelium. In cardiac muscle, Na ϩ /Ca 2ϩ exchange has been well characterized as the primary mechanism for extrusion of the Ca 2ϩ that enters the cell to initiate contraction. Na ϩ /Ca 2ϩ exchange is therefore essential to cardiac relaxation, and an understanding of the function and regulation of this protein is essential to understanding the basis of cardiac excitation-contraction coupling.
Na ϩ /Ca 2ϩ exchange activity can be readily observed using inside-out giant excised patches in which "reverse" exchange is initiated by applying Na ϩ to the intracellular surface. This generates an outward current as intracellular Na ϩ is exchanged for extracellular Ca 2ϩ . This protocol, and other studies, have demonstrated that the exchanger is subject to modulation by several cytoplasmic factors, including Na ϩ and Ca 2ϩ .
Secondary Na ϩ regulation is demonstrated by rapidly raising the intracellular Na ϩ concentration to activate outward Na ϩ /Ca 2ϩ exchange current. Current peaks and then declines over several seconds to a reduced steady-state level. The extent and rate of this partial inactivation are dependent on the concentration of intracellular Na ϩ . This regulatory process has been referred to as Na ϩ -dependent inactivation (1,2).
Regulation of the Na ϩ /Ca 2ϩ exchanger by intracellular Ca 2ϩ has been observed in several preparations. For example, squid axon and mammalian cardiac Na ϩ /Ca 2ϩ exchangers have an absolute requirement for intracellular Ca 2ϩ . The stimulatory effect of intracellular Ca 2ϩ is most easily demonstrated by measuring reverse Na ϩ /Ca 2ϩ exchange. Under these conditions, the nontransported regulatory Ca 2ϩ is on the opposite side of the membrane from the transported extracellular Ca 2ϩ . When NCX1 is expressed in oocytes, a reduction in the intracellular Ca 2ϩ markedly decreases subsequent Ca 2ϩ influx via reverse exchange (3). These and other regulatory effects can be abolished by partial proteolysis with chymotrypsin (2).
Based on hydropathy analysis, the exchanger is modelled to consist of 11 transmembrane segments with a large intracellular loop (loop f) between transmembrane segments 5 and 6 ( Fig.  1). The extracellular location of the amino terminus (4) and the intracellular location of loop f (6, 7) have been experimentally verified. Also, loop e, between transmembrane segments 4 and 5, appears to be extracellular, based on the glycosylation of an exchanger truncated at that site (5). The transmembrane segments are essential for ion transport, whereas loop f is involved in regulation (6).
Through mutagenesis studies, it has been determined that the endogenous XIP 1 region in loop f (8) is involved in Na ϩ -dependent inactivation (9) and that another portion of loop f is involved in the binding of regulatory Ca 2ϩ ( Fig. 1) (3,10). The two regulatory processes, however, are not independent. Binding of regulatory Ca 2ϩ decreases Na ϩ -dependent inactivation (11), and mutations in the endogenous XIP region alter the kinetics of Ca 2ϩ regulation (9).
Other mutagenesis studies have implicated the involvement of two highly conserved and internally homologous regions (the ␣-repeats) in ion transport (12). However, the molecular details of ion transport are unknown, as are the mechanisms by which loop f interacts with the transmembrane segments to effect regulation.
We used substituted cysteine mutagenesis and sulfhydryl modification reagents to determine the membrane topology and explore the function of residues modelled to be at the interface of loop b and transmembrane segment 2. Wild-type NCX1 mutants were expressed in Xenopus oocytes, and Na ϩ /Ca 2ϩ exchange activity was measured as 45 Ca 2ϩ uptake into intact oocytes or as Na ϩ i -activated outward current in giant patches excised from oocyte membranes. We show that this region is located on the cytoplasmic side of the membrane and may form part of the link between transport and regulation.

EXPERIMENTAL PROCEDURES
Mutations in the wild-type exchanger were generated using the Sculptor in vitro mutagenesis kit (Amersham Corp.). cRNA was synthesized with mMessage mMachine (Ambion) and injected into Xenopus laevis oocytes. Na ϩ /Ca 2ϩ exchange activities were measured after incubation of the oocytes for 3-6 days at 17-19°C. To directly measure 45 Ca 2ϩ fluxes, oocytes were first loaded with Na ϩ by incubation in ice-cold Barth's solution containing 20 M nystatin, 0.2% dimethyl sulfoxide for 10 min. Oocytes were then washed four times in 4 ml of 90 mM NaCl, 5 mM HEPES, pH 7.5, and placed into 45 Ca 2ϩ uptake medium containing either NaCl (90 mM) (to measure nonspecific Ca 2ϩ uptake) or KCl (90 mM) (to measure Na ϩ gradient-dependent Ca 2ϩ uptake) and 10 M CaCl 2 , 6.4 mM 4-morpholinepropanesulfonic acid, pH 7.4, and 6 Ci/ml 45 CaCl 2 for 10 min, at which time Ca 2ϩ uptake was still in a linear phase. Unincorporated 45 Ca 2ϩ was then removed by washing the oocytes four times in 10 ml of ice-cold 90 mM KCl, 1 mM EGTA. Oocytes were individually placed into scintillation vials and dissolved in 0.25 ml of 1 M NaOH at 100°C for 20 min. Scintillation fluid (10 ml) was added, and the amount of 45 Ca 2ϩ uptake was determined by liquid scintillation counting. The small nonspecific Ca 2ϩ uptake was subtracted from the Na ϩ gradient-dependent Ca 2ϩ uptake as a measure of Na ϩ /Ca 2ϩ exchange activity. We have previously used this technique extensively to quantitate Na ϩ /Ca 2ϩ exchange activity (6,12,13). Where indicated, oocytes were preloaded with Li ϩ , instead of Na ϩ .
To investigate the effects of extracellular sulfhydryl reagents NEM, MTSES, and MTSET, oocytes were preincubated with reagents (10 mM) for 10 min and then washed four times. To determine the effects of intracellular sulfhydryl reagents, 100 mM MTSES or MTSET was injected into oocytes 10 min before measuring 45 Ca 2ϩ uptake to produce an estimated cytoplasmic concentration of 10 mM (assuming an intracellular oocyte volume of 0.5 l). MTS reagents were purchased from Toronto Research Chemicals.
Data summaries are presented as an average of n experiments Ϯ S.E. Results from 6 -8 oocytes were averaged for each experiment measuring 45 Ca 2ϩ flux. Fig. 1 shows a model for the Na ϩ /Ca 2ϩ exchange protein, based primarily on hydropathy analysis. Loop b consists of amino acids 63-101 and is modeled to be on the cytoplasmic side of the membrane. To determine the orientation of loop b with respect to the membrane, we used substituted cysteine mutagenesis. Specifically, residues modeled to be near the surface of the membrane were mutated to cysteine, and the mutants were exposed to sulfhydryl modifying reagents that form covalent bonds with accessible cysteines. We then examined the effects on exchange activity of membrane-permeable reagents and of impermeable reagents applied either extracellularly or intracellularly.

RESULTS
The wild-type and mutated canine NCX1 exchangers were expressed in Xenopus oocytes. Amino acids Ser-100, Asn-101, Leu-102, Thr-103, and Ala-106 were each individually mutated to cysteine. The mutant T103V (12) was also examined in this study. Fig. 2 shows the activity of each of the mutants relative to the wild-type exchanger. Except for mutant N101C, which had a wild-type level of activity, all of the mutants displayed approximately half the activity of the wild-type exchanger. This may reflect either a reduction in the level of protein expression or a direct effect of the mutation on exchange activity.
The membrane-permeable reagent NEM had no significant effect on the wild-type exchanger but inhibited each of the cysteine mutants except A106C (Fig. 3). T103C was inhibited least. T103C and A106C were modeled to be within the membrane, and may be less accessible, even to membrane-permeable sulfhydryl reagents, than the other cysteine mutants.
The effects of intracellular or extracellular application of MTSET or MTSES are illustrated in Fig. 4. MTS reagents have been used extensively in sulfhydryl modification experiments (14 -16). These relatively large, charged molecules (MTSET is cationic, and MTSES is anionic) are membrane-impermeable. The MTS reagents were either applied extracellularly, by incubating oocytes in medium containing 10 mM reagent, or applied intracellularly, by injecting 100 mM MTS solution into the oocyte for an estimated cytoplasmic concentration of 10 mM. The wild-type exchanger is relatively insensitive to either MTSET or MTSES, although intracellular MTSET caused a modest stimulation (Fig. 4). Extracellular applications of MTSET or MTSES had no substantial effects on any of the cysteine mutants. On the other hand, intracellular application of MTSET or MTSES inhibited FIG. 1. Model of NCX1. Loop f contains regions important in Na ϩ -dependent inactivation (endogenous XIP region) and Ca 2ϩ -regulation (Ca 2ϩ binding site). The ␣-repeats in the transmembrane segments appear to be important in ion translocation. The indicated residues located at the interface of loop b and transmembrane segment 2 were mutated in this study. mutants S100C and L102C. Mutant T103C may also be modestly inhibited by intracellular MTSES. The results verify that the loop b-transmembrane segment 2 interface is located at the cytoplasmic surface of the membrane. The results also suggest that the region is functionally significant because the activity of mutants was inhibited by sulfhydryl modifications.
Further investigation showed that two mutations in this region had important effects on specific aspects of Na/Ca exchange function. Mutant N101C showed altered sensitivity to intracellular Na ϩ and Ca 2ϩ , which suggests that loop b is involved in secondary regulation by these factors. Fig. 5A illustrates the regulatory effects of intracellular Na ϩ and Ca 2ϩ on wild-type Na ϩ /Ca 2ϩ exchange current. Current is activated by application of intracellular Na ϩ ; the slow decay is due to the secondary inhibitory effect of Na i ϩ (2). When intracellular Ca 2ϩ is removed, the current is further inhibited, demonstrating the Ca 2ϩ regulatory effect. These regulatory effects are also characteristic of cardiac Na ϩ /Ca 2ϩ exchange activity measured in situ (1). Both of these effects are abolished following partial proteolysis of the exchanger by chymotrypsin. Fig. 5B shows that N101C has altered regulatory properties. There was no decay in current amplitude after current was activated by raising Na i ϩ , and when Ca i 2ϩ was lowered to 0, there was only a small reduction in current amplitude. In a series of experiments, the wild-type Na ϩ /Ca 2ϩ exchange current decayed to 43 Ϯ 4% of peak after Na ϩ -dependent inactivation (n ϭ 5), but decay of N101C current was insignificant (n ϭ 8). Removal of regulatory Ca 2ϩ reduced the wild-type current by 55 Ϯ 5% even after Na ϩ -dependent inactivation, but it reduced the N101C current by only 24 Ϯ 4%. N101C appeared to function like the wild-type exchanger after deregulation by chymotrypsin. In preliminary experiments, exchangers S100C and L102C displayed normal regulatory properties. Although mutant N101C resembles the chymotrypsintreated wild-type exchanger in terms of a lack of sensitivity to regulatory ions, chymotrypsin digestion still stimulates the current of N101C. This stimulation was small in the experiment shown in Fig. 5B, but in other experiments, the stimulation was more substantial (97 Ϯ 14%; n ϭ 5). Interestingly, chymotrypsin also abolished NEM inhibition of N101C, as shown in Fig. 6. In the experiment shown in Fig. 6A, outward current activated by raising cytoplasmic Na ϩ was inhibited by NEM applied in the cytoplasmic solution, and this inhibitory effect was irreversible. The results indicate that NEM modified the cysteine introduced at residue 101 in such a way as to interfere with exchange, consistent with the data shown above (Fig. 3). When the exchanger was partially proteolyzed by cytoplasmic chymotrypsin and then exposed to NEM, no inhibitory effect was seen (Fig. 6B).
The preceding results indicate that mutation N101C alters the response of the Na ϩ /Ca 2ϩ exchanger to regulatory ions. We also find that mutations in putative transmembrane segment 2, near loop b, affect ion translocation properties. Mutation of threonine 103 modifies the dependence of the Na ϩ /Ca 2ϩ exchanger on cytoplasmic monovalent cations. Fig. 7 shows the dependence of outward Na ϩ /Ca 2ϩ exchange current on cytoplasmic Na ϩ concentration. The excised patches were first exposed to chymotrypsin to remove complicating regulatory effects in these experiments. The wild-type Na i ϩ dependence was sigmoidal, with half-maximal activation at 16 mM Na i ϩ . The Na i ϩ dependence for mutant T103V was shifted to the left, with half-maximal activation at lower Na i ϩ (8 mM) and with a decreased Hill coefficient. Mutant T103V thus has a higher apparent affinity for cytoplasmic Na ϩ and an apparent decreased cooperativity of Na i ϩ -binding. The observed increase in apparent affinity for cytoplasmic Na ϩ raised the possibility that mutant T103V might also have an altered selectivity for monovalent cations. We tested this FIG. 2. Relative activity of mutant exchangers. 45 Ca 2ϩ uptake into oocytes expressing each of the mutant Na ϩ /Ca 2ϩ exchangers was normalized to 45 Ca 2ϩ uptake into oocytes expressing the wild-type exchanger (n ϭ 6 -37).

FIG. 3. Effect of NEM on wild-type and cysteine mutant exchangers.
Oocytes expressing exchangers were incubated in 10 mM NEM for 10 min before 45 Ca 2ϩ uptake was measured. 45 Ca 2ϩ uptake was normalized to uptake in the absence of NEM pretreatment (n ϭ 2-14).

FIG. 4. Effects of intracellular or extracellular application of MTS reagents on wild-type and mutant Na ؉ /Ca 2؉ exchangers.
Oocytes expressing wild-type or mutant exchangers were pretreated by intracellular (MTSET i or MTSES i ) or extracellular (MTSET o or MTSES o ) application of sulfhydryl reagents before 45 Ca 2ϩ uptake was measured. 45 Ca 2ϩ uptake was normalized to uptake in the absence of sulfhydryl reagents for each mutant (n ϭ 5-17). possibility using 45 Ca 2ϩ flux measurements. Oocytes were loaded with Li ϩ in place of Na ϩ , and 45 Ca 2ϩ uptake was measured. The T103V mutant is capable of a greater Li ϩ -dependent 45 Ca 2ϩ uptake than the wild-type exchanger (Fig. 8). Similar results were seen after loading with K ϩ or Cs ϩ (data not shown). Mutant T103V therefore appears to have a reduced selectivity for Na ϩ over other monovalent cations and may be capable of translocating other monovalent cations besides Na ϩ in exchange for Ca 2ϩ .
An alternative possibility is that T103V could support a higher level of Ca 2ϩ /Ca 2ϩ exchange than does the wild-type exchanger, so that increased extracellular 45 Ca 2ϩ would be taken up in exchange for endogenous intracellular Ca 2ϩ , irrespective of the internal monovalent cation. In this case, there would be no associated membrane current, because Ca 2ϩ /Ca 2ϩ exchange is not electrogenic. However, in voltage clamp experiments (Fig. 9), Li ϩ activates a significant, concentration-dependent, outward current in oocytes expressing T103V. The wild-type exchanger shows no change in current when Li i ϩ is applied. Neither Na ϩ nor Li ϩ activated significant outward current in control water-injected oocytes (data not shown). The ability of K ϩ to activate currents was not tested, although Cs ϩ did not appear to do so (Fig. 9). Ca 2ϩ /Ca 2ϩ exchange may have been the prominent mode of 45 Ca 2ϩ uptake for T103V when oocytes were loaded with these monovalent cations (see "Discussion").
A summary of the dependence of outward membrane current on [Na ϩ ] i or [Li ϩ ] i for mutant T103V is shown in Fig. 10. The estimated K1 ⁄2 for Na ϩ is 8 mM, and for Li ϩ , it is 100 mM. Thus, in addition to an apparent increase in affinity for Na i ϩ (Fig. 7), T103V displays a significant Li i ϩ -dependent current. Mutant T103C has similar properties (data not shown).

DISCUSSION
In this study, we have investigated a previously unexamined region of the Na ϩ /Ca 2ϩ exchange protein. Mutations in loop b and the cytoplasmic end of transmembrane segment 2 verify that this region is located at the cytoplasmic face and suggest that it is involved in ion translocation and regulation by cytoplasmic factors.
The transmembrane segments of the Na ϩ /Ca 2ϩ exchange protein have been assigned based on hydropathy analysis of the amino acid sequence (13). Previous results support the extra-cellular assignment of the amino terminus (4) and loop e, between transmembrane segments 4 and 5 (5), and the intracellular assignment of loop f, between transmembrane segments 5 and 6 (6, 7). In this study, mutants S100C and L102C, located in loop b, were sensitive only to intracellular application of membrane-impermeable reagents MTSET and MTSES, thus assigning these residues to the intracellular surface. These results extend our understanding of the structure of the exchanger by assigning loop b, between transmembrane segments 2 and 3, to the inside of the cell.
These results were dependent upon the fact that none of the sulfhydryl modifying reagents we used inhibited the wild-type exchanger. It appears that the endogenous cysteines are inaccessible to modifying reagents or that the accessible cysteines are not intimately involved in exchanger function. An alternative, though unlikely, interpretation is that the mutations in loop b induce a conformational change that exposes a native cysteine to cytoplasmic MTS reagent. We are in the process of constructing a cysteine-less mutant to investigate this possibility.
There are two interesting points in the effect of the membrane-permeable reagent NEM on the loop b mutants. First, the effect of NEM was reduced as the mutations were introduced further into transmembrane segment 2 (Fig. 3). This may be an indication that the mutated residues with reduced NEM sensitivity are in the transmembrane segment as modeled. Second, mutants N101C and T103C are sensitive to inhibition by NEM but not by the MTS reagents. Thus, these two residues are solvent-accessible but are more restricted to chemical reaction than residues Ser-100 or Leu-102. There may perhaps be local hydrophobic or steric constraints preventing reaction with the MTS reagents.
Secondary regulation by cytoplasmic Na ϩ and Ca 2ϩ has previously been localized to functional sites in loop f, but no other regions of the exchanger were implicated in the regulatory process. Mutations in the endogenous XIP region (Fig. 1) can remove Na ϩ -dependent inactivation but not Ca 2ϩ stimulation (9), whereas deletion of amino acids 562-685 has the opposite effect (6), removing Ca 2ϩ stimulation but not Na ϩ -dependent inactivation. A functionally important binding site for regulatory Ca 2ϩ has been identified at amino acids 371-508 (10). Some mutations in loop f alter both Na ϩ inhibition and Ca 2ϩ stimulation, indicating that the two processes interact (9).
Loop f must exert regulatory effects on transport by influencing the transmembrane segments that catalyze ion translocation. We show here that a mutation in loop b (N101C) FIG. 8. 45 Ca 2؉ uptake into oocytes expressing wild-type or T103V exchangers after loading with Na ؉ or Li ؉ . 45 Ca 2ϩ uptake was measured after loading oocytes with Na ϩ or Li ϩ (n ϭ 6 -7 oocytes) as described under "Experimental Procedures" and was normalized to the average 45 Ca 2ϩ uptake in Na ϩ -loaded oocytes. interferes with secondary regulation by Na ϩ and Ca 2ϩ . Two explanations are considered. First, the residue at position 101 may interact with loop f. Mutating residue Asn-101 to cysteine may then alter interaction between loops b and f and block regulation. Second, the mutation may alter the protein conformation so that loop f is no longer able to appropriately influence, and regulate, the transmembrane segments. We are unable to distinguish these possibilities. Nevertheless, in both cases, the interface of loop b and transmembrane segment 2 is identified as a key site in the coupling of regulation to translocation.
We explored the possibility that the cysteine introduced in N101C interrupts regulatory mechanisms by forming an internal disulfide bond with a native cysteine. We found that the reducing agent dithiothreitol does not restore regulation to the mutant as would be expected under these circumstances (data not shown). Thus, it does not appear that disulfide bond formation is the cause of loss of regulation in N101C.
Partial proteolysis by cytoplasmic chymotrypsin, which removes secondary regulatory effects of intracellular Na ϩ and Ca 2ϩ in the wild-type Na ϩ /Ca 2ϩ exchanger, also prevents the inhibitory effect of the sulfhydryl reagent NEM on N101C. Proteolysis may remove secondary regulation in the wild-type exchanger by disrupting the interaction between regulatory sites in loop f and the translocation pathway. It is thought that the major site for proteolysis of the exchanger is within loop f (17). The fact that proteolysis removes the NEM effect on mutant N101C suggests a functional relationship between loops b and f. Again, the relationship could be either a direct or an indirect interaction.
The cytoplasmic end of transmembrane segment 2 appears to play a role in the translocation of Na ϩ ions. Mutant T103V shows an increased apparent affinity for cytoplasmic Na ϩ and also seems to support Li ϩ /Ca 2ϩ exchange. T103C has identical characteristics, but mutant N101C does not show any Li ϩactivated current (data not shown). The Li ϩ -activated current of mutant T103V is relatively small. At 100 mM activating ion, the Li ϩ -activated exchange current is about 20% of Na ϩ -activated exchange current. Because mutant T103V has a higher affinity for Na ϩ than the wild-type exchanger, the possibility arises that contaminating Na ϩ in the cytoplasmic Li ϩ solution could activate significant current in the mutant but not the wild-type exchanger. This is unlikely because we used 99.95% pure LiOH, and 2-3 mM contaminating Na ϩ would have to be present in the 100 mM Li ϩ solution to produce a contamination artifact.
The 45 Ca 2ϩ flux measurements show that mutant T103V transports Ca 2ϩ ions across the membrane without requiring a Na ϩ gradient (Fig. 8), and the Li ϩ -activated outward current indicates that an electrogenic process accompanies this Ca 2ϩ transport (Fig. 9). This suggests that mutant T103V can directly catalyze Li ϩ /Ca 2ϩ exchange. Other Na ϩ transporters, such as the Na ϩ /H ϩ exchanger (18), the Na ϩ /K ϩ ATPase (19), and the mitochondrial Na ϩ /Ca 2ϩ exchanger (20), can substitute Li ϩ for Na ϩ to varying degrees. However, this is the first report of a plasma membrane Na ϩ /Ca 2ϩ exchanger capable of Li ϩ transport.
The relative level of Ca 2ϩ -Ca 2ϩ exchange may also be higher for mutant T103V than for the wild-type exchanger. When 45 Ca 2ϩ uptake was measured in the presence of a Li ϩ gradient (Fig. 8), 45 Ca 2ϩ flux is due to a combination of Li ϩ /Ca 2ϩ exchange and Ca 2ϩ -Ca 2ϩ exchange (external 45 Ca 2ϩ exchanging for internal endogenous Ca 2ϩ ). On the other hand, Li ϩ -induced currents (Fig. 9) are due only to Li ϩ /Ca 2ϩ exchange (Ca 2ϩ -Ca 2ϩ exchange is not electrogenic). For mutant exchanger T103V, the 45 Ca 2ϩ flux into Li ϩ -loaded oocytes was about 80% of that into Na ϩ -loaded oocytes, but the Li ϩ -induced current was only about 20% of the Na ϩ -induced current. Thus, Ca 2ϩ -Ca 2ϩ exchange activity of T103V appeared to be much more robust than that of the wild-type exchanger.
In parallel experiments (not shown; n ϭ 2), Li ϩ -loaded oocytes were diluted into 45 Ca 2ϩ uptake medium containing K ϩ (Li ϩ gradient present) or Li ϩ (no Li ϩ gradient). 45 Ca 2ϩ uptake levels for the wild-type exchanger were 33 and 27%, respectively, of that obtained in the presence of a Na ϩ gradient. For mutant T103V, these values were 107 and 173%. These results are consistent with the idea that the level of Ca 2ϩ -Ca 2ϩ exchange is greater in mutant T103V than in the wild-type exchanger.
The data in Fig. 10, showing the relative transport activity of Na ϩ /Ca 2ϩ exchange and Li ϩ /Ca 2ϩ exchange for T103V, were fit well with a simple model. Using a sequential Na ϩ /Ca 2ϩ exchange cycle to represent the wild-type exchanger, we added low-affinity Li ϩ binding and Li ϩ /Ca 2ϩ exchange. In this model, the best fit is produced when Li ϩ binds with a K1 ⁄2 of 100 mM and the Li ϩ translocation rate is 40% of the Na ϩ translocation rate. It is possible that residue Thr-103 is part of a Na ϩ binding site and that the mutation increases the affinity of the site for Na ϩ and also for Li ϩ . It is striking to note the proximity of the residue to ␣-repeat 1 (Fig. 1), which begins at alanine 106. The ␣-repeats are highly conserved among different isoforms of the Na ϩ /Ca 2ϩ exchanger and play an important role in function (12).
The functional map of the Na ϩ /Ca 2ϩ exchange protein is augmented by the data presented here. Loop b has been localized to the cytoplasmic side of the membrane and assigned a potential role in secondary regulation of the exchanger by cytoplasmic factors. The cytoplasmic end of transmembrane segment 2 may have an important functional role in the transport of Na ϩ . The loop b-transmembrane segment 2 interface could be an important structural link between ion translocation and regulation.