Structure-Function Analysis of CALX1.1, a Na+-Ca2+ Exchanger fromDrosophila

Cytoplasmic Na+ and Ca2+ regulate the activity of Na+-Ca2+ exchange proteins, in addition to serving as the transported ions, and protein regions involved in these processes have been identified for the canine cardiac Na+-Ca2+ exchanger, NCX1.1. Although protein regions associated with Na+ i - and Ca2+ i -dependent regulation are highly conserved among cloned Na+-Ca2+ exchangers, it is unknown whether or not the structure-function relationships characteristic of NCX1.1 apply to any other exchangers. Therefore, we studied structure-function relationships in a Na+-Ca2+ exchanger from Drosophila, CALX1.1, which is unique among characterized members of this family of proteins in that μm levels of Ca2+ i inhibit exchange current. Wild-type and mutant CALX1.1 exchangers were expressed in Xenopus oocytes and characterized electrophysiologically using the giant excised patch technique. Mutations within the putative regulatory Ca2+ i binding site of CALX1.1, like corresponding alterations in NCX1.1, led to reduced ability (i.e. D516V and D550I) or inability (i.e. G555P) of Ca2+ i to inhibit Na+-Ca2+exchange activity. Similarly, mutations within the putative XIP region of CALX1.1, as in NCX1.1, led to two distinct phenotypes: acceleration (i.e. K306Q) and elimination (i.e. Δ310–313) of Na+ i -dependent inactivation. These results indicate that the respective regulatory roles of the Ca2+ i binding site and XIP region are conserved between CALX1.1 and NCX1.1, despite opposite responses to Ca2+ i . We extended these findings using chimeric constructs of CALX1.1 and NCX1.1 to determine whether or not functional interconversion of Ca2+ i regulatory phenotypes was feasible. With one chimera (i.e. CALX:NCX:CALX), substitution of a 193-amino acid segment, from the large intracellular loop of NCX1.1, for the corresponding 177-amino acid segment of CALX1.1 led to an exchanger that was stimulated by Ca2+ i . This result indicates that the regulatory Ca2+ i binding site of NCX1.1 retains function in a CALX1.1 parent transporter and that the substituted segment contains some of the amino acid sequence(s) required for transduction of the Ca2+ i binding signal.

The identification of novel Na ϩ -Ca 2ϩ exchange proteins has proceeded rapidly in the past 8 years. The family of Na ϩ -Ca 2ϩ exchangers includes transporters encoded by unique gene products (1)(2)(3), as well as by a variety of alternatively spliced variants (4 -7). Promoter elements underlying the tissue-specific expression of alternatively spliced exchangers have been described (8,9), and developmental changes in splice variant expression have been reported (7). Moreover, exchangers have been cloned from several species and tissue types, and the opportunity now exists for detailed comparative structurefunction studies of this family of transport molecules (10).
Regulation of Na ϩ -Ca 2ϩ exchange activity by several factors has been described, including Na ϩ i and Ca 2ϩ i , with the latter being the most thoroughly studied at the molecular level. First identified in the squid giant axon (11,12), Ca 2ϩ i -dependent regulation is apparent as a stimulation of Na ϩ -Ca 2ϩ exchange current in response to M levels of Ca 2ϩ i (13). The basis for this behavior, originally termed I 2 inactivation, is thought to involve entry of the exchanger protein into an I 2 inactive state upon removal of Ca 2ϩ i (14). A high affinity Ca 2ϩ i binding site has been identified for the canine cardiac exchanger, NCX1.1, which appears to be closely associated with the Ca 2ϩ i -dependent regulatory process. This site comprises a 138-amino acid segment of the large intracellular loop of NCX1.1; mutation of specific residues within this region can lead to reductions in both 45 Ca 2ϩ binding affinity to fusion proteins (15) and the affinity for functional Ca 2ϩ i regulation as assessed electrophysiologically (16).
The cardiac Na ϩ -Ca 2ϩ exchanger also undergoes an inactivation process in response to the application of Na ϩ i (13,17). This mechanism, termed Na ϩ i -dependent or I 1 inactivation, is analogous to ion channel gating and involves the exchanger inhibitory peptide (XIP) 1 region at the N terminus of the large cytoplasmic loop of NCX1.1 (18). This amino acid sequence, comprising residues 219 -238, was originally identified based upon primary structural similarity with calmodulin binding sites (1). Exogenous application of a peptide corresponding to this amino acid sequence (i.e. XIP) to the intracellular surface of excised membrane patches produces marked inhibition of Na ϩ -Ca 2ϩ exchange currents (19,20). More recent studies have shown that mutations within the XIP region of NCX1.1 are associated with substantial alterations in the rate and extent of I 1 inactivation (18), lending support to the notion that the XIP region of NCX1.1 is intimately involved in the mechanism of Na ϩ i -dependent inactivation and that exogenous application of XIP may mimic this process. Both of the ionic regulatory mechanisms (i.e. I 1 and I 2 ) can be eliminated by limited proteolysis of membrane patches with ␣-chymotrypsin, apparently converting NCX1.1 into a fully active, deregulated exchanger (13).
Although the role of I 1 and I 2 regulation in physiological Na ϩ -Ca 2ϩ exchange function remains poorly understood, both processes can be readily demonstrated in intact cellular preparations (21,22). Furthermore, substantial interaction between I 1 and I 2 regulation has been demonstrated in both structure-function studies and electrophysiological analyses (14,16,18). We have recently observed substantial differences in I 1 and I 2 regulation for alternatively spliced isoforms of CALX1 (23). The fact that alternative splicing has targeted these regulatory mechanisms at least suggests that they may play relevant physiological roles. However, it is unclear how I 1 and I 2 inactivation occur at the molecular level.
To date, structure-function studies of the Na ϩ -Ca 2ϩ exchanger have been restricted to NCX1.1 (16, 18, 20, 24 -26), and it is unknown whether or not these findings can be extended to other members of this family of transport proteins. Thus, we combined mutagenesis, chimeric exchanger construction, and electrophysiological measurements to investigate amino acid sequences involved in ionic regulatory mechanisms of CALX1.1, a Na ϩ -Ca 2ϩ exchanger from Drosophila melanogaster (27,28). The Drosophila protein was selected for study because it is unique among characterized exchangers in terms of its negative regulatory response to Ca 2ϩ i (29). Targets for mutagenesis of CALX1.1 were selected at or near regions analogous to the regulatory Ca 2ϩ i binding site and XIP region of NCX1.1, and we found striking parallels in terms of altered ionic regulatory properties between the two exchangers. Our results indicate that CALX1.1 and NCX1.1 employ an equivalent site for binding of regulatory Ca 2ϩ i . Other regions of the CALX1.1 cytoplasmic loop were identified that completely alleviate Ca 2ϩ i regulation. Mutations within the XIP region alter the pattern of Na ϩ i -dependent inactivation, similar to that observed for NCX1.1. Our results indicate that amino acid sequences subserving Na ϩ i -and Ca 2ϩ i -dependent regulatory processes are conserved between CALX1.1 and NCX1.1, irrespective of the fact that these exchangers are regulated by Ca 2ϩ i in opposing fashions.
Electrophysiology-Complimentary RNAs (ϳ5 ng) encoding wildtype and mutant/chimeric exchangers were injected into oocytes obtained from Xenopus laevis (29,31), and outward Na ϩ -Ca 2ϩ exchange currents were measured 3-7 days later. Na ϩ -Ca 2ϩ exchange activity was assessed using the giant excised patch technique (13) as described (29,31). Outward (i.e. reverse) exchange currents were examined exclusively because this configuration separates regulatory and transported Ca 2ϩ to opposite surfaces of the cell membrane. Borosilicate glass pipettes (N-51A, Drummond Scientific) were pulled and polished to a final diameter of ϳ20 -35 m and coated with a Parafilm:mineral oil mixture to reduce electrical noise and enhance patch stability. Oo-

RESULTS
Regulation and Deregulation of CALX1.1 Na ϩ -Ca 2ϩ Exchange Currents-In the left panel of Fig. 1, fully regulated outward Na ϩ -Ca 2ϩ exchange currents are shown from an oocyte membrane patch expressing CALX1.1. The pipette contained 8 mm Ca 2ϩ (i.e. transported Ca 2ϩ o ) and the outward current was activated by applying 100 mM Na ϩ i to the cytoplasmic surface of the patch. In the absence of Ca 2ϩ i , a large outward current was observed that decayed over seconds to a lower steady-state value. This current decay reflects Na ϩ i -dependent (i.e. I 1 ) inactivation. Upon application of 1 M Ca 2ϩ i to the intracellular side of the patch, the current was further decreased, but it recovered to the initial steady-state value upon removal of Ca 2ϩ i . This Ca 2ϩ i -dependent inhibition of outward Na ϩ -Ca 2ϩ exchange current is opposite of that observed for the cardiac exchanger NCX1.1. With the exception of CALX1 isoforms (23,29), all wild-type Na ϩ -Ca 2ϩ exchangers characterized to date are stimulated by regulatory Ca 2ϩ i . The right panels of Fig. 1 show current records from a different patch following treatment with 2 mg/ml of ␣-chymotrypsin for ϳ60 s. Following enzyme digestion, the current decay observed upon Na ϩ i application was nearly abolished, and current magnitudes were similar in the absence or presence of 1 M Ca 2ϩ i . This deregulation of CALX1.1 outward exchange FIG. 1. Regulation and deregulation of Na ؉ -Ca 2؉ exchange current for CALX1.1. In the left panel, regulated outward Na ϩ -Ca 2ϩ exchange current is shown from a membrane patch expressing CALX1.1. Current was activated by the application of 100 mM Na ϩ i in the absence of regulatory Ca 2ϩ i . A large current was observed that decays to a lower, steady-state value, reflecting the I 1 inactivation process. Following application of 1 M Ca 2ϩ i , a further decrease of current was observed (negative Ca 2ϩ i regulation), followed by a return to steady-state levels upon Ca 2ϩ i removal. Results shown on the right are from a different patch that was treated with 2 mg/ml ␣-chymotrypsin for ϳ1 min in order to deregulate CALX1.1. In this case, I 1 inactivation was nearly abolished, and outward currents were similar in the presence or absence of 1 M Ca 2ϩ i . The results shown are representative of those obtained from 16 regulated patches and 5 deregulated patches. activity by ␣-chymotrypsin was observed in five patches and is qualitatively identical to that observed with NCX1.1 or NCX2 following similar treatment (2,16).
The Putative Regulatory Ca 2ϩi Binding Site-The regulatory Ca 2ϩ i binding site of NCX1.1 has been localized to a region of its large cytoplasmic loop between amino acids 371 and 508 (10,15,16). Prominent within this region are two clusters of acidic amino acids (Fig. 2). Mutation of specific residues within these clusters of NCX1.1 produces changes in the ability of Ca 2ϩ i to regulate exchange activity and in the affinity of 45 Ca 2ϩ to bind to a fusion protein containing amino acids 371-508 (15,16). Alignment of amino acid sequences flanking these clusters from a number of cloned Na ϩ -Ca 2ϩ exchangers (Fig. 2) reveals a striking degree of similarity. Although suggestive of a conserved functional role for the region across species and subtypes of exchangers, it is unknown whether these acidic clusters play a role in Ca 2ϩ i binding for any exchanger other than NCX1.1. Therefore, we targeted residues analogous to those studied in NCX1.1 to ascertain whether or not this region also functions as the regulatory Ca 2ϩ i binding site of CALX1.1.  Fig. 3 shows results from D516V and D550I for direct comparison with NCX1.1 mutants D447V and D498I (16). Comparing the responses to 1 and 3 M regulatory Ca 2ϩ i for CALX1.1, it can be seen that with both mutant exchangers, neutralization of these conserved acidic residues leads to a reduction in the inhibitory effect of Ca 2ϩ i . Although we did not exhaustively study all mutations of CALX1.1 that are analogous to those reported to affect Ca 2ϩ idependent regulation of NCX1.1, it seems probable that this region serves an equivalent function to bind regulatory Ca 2ϩ i in both exchangers.

Mutations in the Regulatory
The Putative XIP Region-The XIP region of NCX1.1, comprising amino acids 219 -238, is located near the N terminus of its large cytoplasmic loop. Mutation of residues within this structure of NCX1.1 lead to pronounced alterations in the rate and extent of Na ϩ i -dependent (i.e. I 1 ) inactivation (18). An alignment of the corresponding XIP-like sequences from a number of cloned Na ϩ -Ca 2ϩ exchangers is presented in Fig. 4. Although less well-conserved than the acidic clusters of the regulatory Ca 2ϩ i binding site (see Fig. 2), substantial similarity is evident. For example, between NCX1.1 and CALX1.1, 13 of 20 residues are identical (65%), whereas the overall sequence identity is only ϳ50%. As for the Ca 2ϩ i binding site, the role of this region in any exchanger other than NCX1.1 is unknown. This prompted us to examine two mutations of CALX1.1 analogous to those characterized in NCX1.1. If conserved, these

FIG. 2. Acidic amino acid clusters within the putative regulatory Ca 2؉
i binding site. The regulatory Ca 2ϩ i binding site has been localized to amino acids 371-508 in the cardiac Na ϩ -Ca 2ϩ exchanger, NCX1.1. Two clusters of acidic amino acid residues are prominent within this region, and specific mutations of acidic residues lead to alterations in the affinity of Ca 2ϩ i binding and functional regulation. In this alignment, we show the analogous acidic clusters for CALX1.1 and a variety of other Na ϩ -Ca 2ϩ exchangers. Amino acid numbering for Caenorhabditis elegans and X. laevis is not shown because complete amino acid sequences were not available at the time of writing.

FIG. 3. Mutations within the regulatory Ca 2؉
i binding site of CALX1.1. Outward Na ϩ -Ca 2ϩ exchange currents are shown for CALX1.1 and three exchangers with mutations in the regulatory Ca 2ϩ i binding site. All currents were activated by applying 100 mM Na ϩ i . When present, 1, 3, or 10 M Ca 2ϩ i was present prior to and during current activation. Compared with CALX1.1, currents from G555P were insensitive to Ca 2ϩ i , whereas those from D516V and D550I exhibited a lower affinity for negative Ca 2ϩ i regulation. Current traces for G555P and D550I were filtered at 200 Hz. mutations were predicted to mirror the two phenotypic classes of XIP mutants, namely acceleration (i.e. type I) and elimination (i.e. type II) of Na ϩ i -dependent inactivation (18). Fig. 5 shows typical outward Na ϩ -Ca 2ϩ currents activated in response to a range of Na ϩ i concentrations to illustrate the I 1 inactivation process in more detail. All records were obtained in the absence of regulatory Ca 2ϩ

Mutations in the XIP Region-
i . For CALX1.1, both current magnitude and the extent of current inactivation rose with increasing [Na ϩ ] i . For K306Q (analogous to K225Q in NCX1.1), a similar response was observed, although the rate of I 1 inactivation is faster. For example, in response to a pulse of 100 mM Na ϩ i , the rate of inactivation for K306Q was 1.27 Ϯ 0.06 s Ϫ1 (n ϭ 10) as compared with 0.56 Ϯ 0.04 s Ϫ1 (n ϭ 25) for CALX1.1. Also consistent with an accelerated entry of K306Q into the I 1 inactive state as compared with CALX1.1 is the observation that the fraction of steady-state to peak outward current (F ss ) for K306Q is slightly lower than for CALX1.1 (0.21 Ϯ 0.01 versus 0.26 Ϯ 0.01 s Ϫ1 for K306Q and CALX1.1, respectively), indicative of a greater proportion of the K306Q exchanger population being in the I 1 inactive state. In contrast, for the deletion mutant ⌬310 -313 (analogous to ⌬229 -232 in NCX1.1), I 1 inactivation was apparently eliminated. These results with K306Q and ⌬310 -313 are directly analogous to those observed for NCX1.1 type I and II mutant exchangers, respectively. Apparently, the XIP-like region of CALX1.1 and the XIP region of NCX1.1 play similar functional roles in the process of Na ϩ i -dependent regulation of exchange activity.
To examine more closely the role of the XIP-like region of CALX1.1 in the process of Na ϩ i -dependent regulation, we performed paired-pulse experiments in order to directly determine the rate constant controlling recovery from the I 1 inactivation process. In all cases, outward currents were activated by 100 mM Na ϩ i in the absence of Ca 2ϩ i . Typical traces are shown in Fig. 6 for CALX1.1, K306Q, and ⌬310 -313 for recovery intervals of 16, 4, and 0.5 s between the initial Na ϩ i pulse and the second test pulse. Comparing CALX1.1 and K306Q, note that recovery from I 1 inactivation was substantially faster for the mutant exchanger. That is, following any recovery interval, the peak current of the test pulse was larger for K306Q than for CALX1.1. This is shown graphically in Fig. 6, bottom right panel, for pooled results from five patches for K306Q and eight patches for CALX1.1. On average, the recovery from I 1 inactivation occurred about twice as fast for K306Q as compared with CALX1.1. In contrast, with ⌬310 -313, outward currents were essentially the same irrespective of the test interval, suggesting that the I 1 inactivation mechanism has been eliminated in this mutant. This virtually instantaneous recovery of ⌬310 -313 was observed in four patches and is qualitatively similar to that of CALX1.1, in which, following deregulation of the exchanger by ␣-chymotrypsin-treatment, I 1 inactivation was also undetectable (Fig. 6, bottom left panel). Because ⌬310 -313 does not appear to enter the I 1 inactive state, these data are not plotted as recoveries.
Ca 2ϩ i -dependent Regulation of Chimeric Exchangers-The above results indicate that the function of the Ca 2ϩ i binding site and XIP region is conserved between CALX1.1 and NCX1.1 in that analogous mutations in either region lead to analogous changes in Ca 2ϩ i -and Na ϩ i -dependent regulation, respectively. Yet these exchangers exhibit opposite responses to regulatory Ca 2ϩ i . In an effort to determine the basis for this apparent anomaly and identify protein domains that are responsible for imparting a particular regulatory phenotype, we constructed six chimeric CALX1.1:NCX1.1 exchangers (Fig. 7, bottom  panel) and examined their responses to Ca 2ϩ i . Fig. 7 illustrates this approach. Two silent restriction sites were engineered into i . CALX1.1 exhibited the typical increase in current, and the extent of Na ϩ i -dependent (i.e. I 1 ) inactivation as Na ϩ i was increased. A similar response was observed for K306Q, although the rate and extent of current decay was greater. In contrast, the deletion mutant ⌬310 -313 did not undergo Na ϩ i -dependent inactivation.
both NCX1.1 and CALX1.1 within regions of perfect sequence conservation among all cloned Na ϩ -Ca 2ϩ exchangers to optimize our chances of obtaining functional constructs (Fig. 7,  middle panel). This allowed construction of exchanger cassettes containing 1) the first five transmembrane-spanning segments and XIP region, 2) the regulatory Ca 2ϩ i binding domain (currently defined as amino acids 371-508 in NCX1.1) (10,15,16) with adjacent sequences, and 3) the alternative splicing region and transmembrane-spanning segments 6 -11. We hypothesized that if the regulatory Ca 2ϩ i binding site and XIP region were indeed as similar in both exchangers, as our data indicate, then 1) normal Ca 2ϩ i binding and I 1 inactivation might be preserved, and 2) the approximate location of other domains more directly involved in transduction of the Ca 2ϩ i binding signal might be revealed. Fig. 8 summarizes results from a comparison of the regulatory effects of 1 M Ca 2ϩ i for NCX1.1, CALX1.1, and a chimeric exchanger designated CALX:NCX:CALX. All currents were activated by 100 mM Na ϩ i with Ca 2ϩ i present or absent, as indicated. The representative traces shown illustrate the typical stimulatory effect of Ca 2ϩ i on NCX1.1 outward current and the inhibition of CALX1.1 exchange activity. For the chimera, however, the majority of which comprises CALX1.1 sequence (see Fig. 7), Na ϩ -Ca 2ϩ exchange current was also stimulated by regulatory Ca 2ϩ i , with peak current being more prominently affected than that of steady-state. The graph in Fig. 8, bottom left, presents peak current data from 10 CALX1.1 patches and 9 CALX:NCX:CALX patches, covering a range of Ca 2ϩ i concentrations. At every concentration examined, CALX1.1 peak current was inhibited by Ca 2ϩ i , whereas that of the chimera was stimulated. Although it is clear that CALX:NCX:CALX does not behave identically to NCX1.1, it is evident that we have obtained a partial interconversion of regulatory phenotypes. Of the remaining five chimeras in this series, four did not produce measurable levels of Na ϩ -Ca 2ϩ exchange current. The common feature among this inactive group was that each had dissimilar N-and C-terminal thirds (i.e. CALX:CALX:NCX, CALX:NCX: NCX, NCX:NCX:CALX, and NCX:CALX:CALX). In contrast, the chimera NCX:CALX:NCX yielded measurable outward exchange current that retained Na ϩ i -dependent inactivation but was insensitive to Ca 2ϩ i up to 30 M (data not shown), like G555P (see Fig. 3).

DISCUSSION
In this study, we have examined ionic regulation of the Drosophila Na ϩ -Ca 2ϩ exchanger CALX1.1 using mutagenesis FIG. 6. Recovery from Na ؉ i -dependent (I 1 ) inactivation. Paired pulses of outward Na ϩ -Ca 2ϩ exchange currents were used to examine recovery from I 1 inactivation for CALX1.1 and two mutant exchangers. Currents were activated by 100 mM Na ϩ i in the absence of Ca 2ϩ i . Typical traces obtained at 16, 4, and 0.5 s recovery intervals are shown. Note that recovery of peak current of the test (i.e. second) pulse occurred more rapidly for K306Q than for CALX1.1. Pooled results from five patches for K306Q and eight patches for CALX1.1 are shown in the graph (bottom right panel). In contrast, ⌬310 -313 did not appear to enter the inactive state. This response is similar to that of deregulated CALX1.1 after treatment with ␣-chymotrypsin (ChT) (2 mg/ml) for ϳ1 min (bottom left panel). and electrophysiological techniques. Our results indicate that Ca 2ϩ i regulation in both NCX1.1 and CALX1.1 appears to employ the same regulatory Ca 2ϩ i binding site, even though these exchangers exhibit opposite patterns of Ca 2ϩ i regulation. In addition, we have examined mutations in the XIP region of CALX1.1 and obtained similar functional consequences to that observed in NCX1.1. These mutations produced the same spectrum of changes in Na ϩ i -dependent inactivation, ranging from an acceleration to complete elimination. We have not attempted to be exhaustive in our comparisons of CALX1.1 and NCX1.1, but rather to ascertain the likelihood that identified functional domains are conserved. Our data are highly supportive of this conservation. Consequently, the unique Ca 2ϩ i regulatory phenotypes between these exchangers must reside in an as yet unidentified domain(s) of the transporters. From initial results with chimeric exchangers, we have been successful in imparting a positive Ca 2ϩ i regulatory phenotype to a CALX1.1 parent transporter by interchanging a portion of its cytoplasmic loop with the corresponding region from NCX1.1. Further delineation of this region should provide insight into the transduction of the Ca 2ϩ i binding signal. Ca 2ϩ i -dependent Regulation of CALX1.1-Regulation of Na ϩ -Ca 2ϩ exchange currents by Ca 2ϩ i has been described in all exchangers studied to date, including the unique exchanger subtypes NCX1 (16), NCX2 (2), and NCX3. 2 With the exception of CALX1 splice variants, low concentrations (i.e. M range) of Ca 2ϩ i stimulate exchange current. CALX1.1 and CALX1.2 are unique in that exchange activity is decreased over the same range of Ca 2ϩ i concentrations (23,29). Although the primary structure of the regulatory Ca 2ϩ i binding site identified in NCX1.1 is well-conserved among all exchangers, there have been no studies to determine whether or not this site performs a similar function in other exchangers. Thus, we selected for study an exchanger, CALX1.1, with a Ca 2ϩ i regulatory phenotype opposite to that of NCX1.1 for two main reasons. First, if the Ca 2ϩ i binding site serves a similar functional role in both NCX1.1 and CALX1.1, we anticipated that its affinity could be predictably altered by mutagenesis. Second, we speculated that if the 138-amino acid Ca 2ϩ i binding domain of NCX1.1 contained some (or all) of the amino acid determinants directly responsible for transduction of the Ca 2ϩ i binding signal, then interchanging this region between CALX1.1 and NCX1.1 presented the greatest likelihood of observing phenotypic differences in the recipient exchanger.
Our results show that NCX1.1-analogous mutations within the regulatory Ca 2ϩ i binding site of CALX1.1 lead to parallel changes in the ability of Ca 2ϩ i to regulate exchange activity. That is, mutation of corresponding amino acid residues (see Fig. 2) reduces the affinity for Ca 2ϩ i regulation, irrespective of whether the regulatory effect is stimulatory (i.e. NCX1.1) or inhibitory (i.e. CALX1.1). This result strongly suggests that the function of the Ca 2ϩ i binding site is similar in both exchangers. Considering the striking degree of amino acid sequence similarity between the acidic clusters within the putative Ca 2ϩ i binding sites of all exchanger subtypes and species variants shown in Fig. 2, and our present results showing a conserved functional role for exchangers sharing only ϳ50% overall sequence identity and opposite Ca 2ϩ i -dependent regulatory phenotypes, it seems plausible that the regulatory Ca 2ϩ i binding site may serve the same function in other, if not all, Na ϩ -Ca 2ϩ exchangers.
Na ϩ i -dependent Regulation of CALX1.1-We have shown that the XIP region serves a similar role in both CALX1.1 and NCX1.1 because analogous mutations produce similar phenotypes in both exchangers. With K306Q, corresponding to K225Q in NCX1.1 (see Fig. 4) (18), current decay into (Fig. 5) and recovery from (Fig. 6) the I 1 inactive state was increased ϳ2-fold, mirrored by a slight increase in the extent of inactivation as indicated by a lower value of F ss , the fraction of steady-state to peak exchange current. According to the twostate model for I 1 inactivation proposed by Hilgemann et al. (17), the larger I 1 recovery rate constant of K306Q compared with CALX1.1 provides a satisfactory account for the accelerated current decay rate observed with K306Q versus CALX1.1. The observation that F ss is reduced for K306Q also indicates that entry into the I 1 inactive state has been accelerated. In contrast, the I 1 inactivation process was apparently eliminated altogether in ⌬310 -313, a result identical to that obtained with the corresponding deletion mutant in NCX1.1, ⌬229 -232 (18), and with ␣-chymotrypsin-digested (i.e. deregulated) CALX1.1 (Fig. 6). Although the degree of primary structural similarity within the XIP-like regions of the various exchangers assembled in Fig. 4 is clearly less than for the acidic clusters of the Ca 2ϩ i binding site (Fig. 2), it is sufficiently prominent to tentatively conclude that this region may also share a common function in the control of Na ϩ i -dependent inactivation of Na ϩ -Ca 2ϩ exchangers in general. This possibility is supported by the observation that both CALX1.1 (29) and NCX2 (2), like NCX1.1, are inhibited by XIP, which appears to mimic the I 1 inactivation process.
Ca 2ϩ i -insensitive and Chimeric Exchangers-Although the function of the XIP region and Ca 2ϩ i binding site appears to be conserved between CALX1.1 and NCX1.1, we have no definitive explanation for the observed differences in Ca 2ϩ i -dependent regulatory phenotypes. However, we observed that analogous mutations can render both exchangers insensitive to Ca 2ϩ i (e.g. G555P in CALX1.1 and G503P in NCX1.1) (see 3) (18,33). One possibility is that Ca 2ϩ i binds to these exchangers but the transduction process has been disabled. Supporting this notion is a report by Levitsky et al. (15) showing that a fusion protein containing the Ca 2ϩ i binding site of NCX1.1, but bearing a G503P equivalent mutation, binds 45 Ca 2ϩ i indistinguishably from the wild-type sequence. Also consistent with this hypothesis is our result with the chimeric exchanger CALX:NCX:CALX. Here, interchange of the entire Ca 2ϩ i binding site and flanking sequences of NCX1.1 with the corresponding region of CALX1.1 led to a transporter that was stimulated by regulatory Ca 2ϩ i (see Fig. 8). Thus, the Ca 2ϩ i binding function appears to have been retained, and a partial interconversion of phenotypes occurred. Although extrapolation of results from fusion and chimeric proteins to the wild-type exchanger must be made with caution, the above findings suggest that Ca 2ϩ i binding per se may be restricted to the acidic clusters within the larger Ca 2ϩ i binding domain, and that transduction of the Ca 2ϩ i binding signal may occur through sequences remote or distinct from this region. Our results with NCX:CALX: NCX neither support nor refute this notion, irrespective of the fact that the chimeric protein did acquire a novel phenotype (i.e. Ca 2ϩ i -insensitive). Although there is little evidence to indicate that Ca 2ϩ i binding and transduction domains are modular or discrete, our results are at least suggestive of this possibility. Furthermore, we can conclude that the primary structural alterations associated with these two chimeras are relatively benign with respect to the transport function and Na ϩ i -dependent regulation of the parent molecule. Complete and "symmetric" reversal of Ca 2ϩ i -dependent regulatory phenotypes between CALX1.1 and NCX1.1 may be possible through additional studies of this type.