Ion Specificity of Cardiac Sarcolemmal Na+/H’ Antiporter”

In bovine cardiac sarcolemmal vesicles, an outward H+ gradient stimulated the initial rate of amiloride- sensitive uptake of “Na+, 42K+, or 86Rb+. Release of H+ from the vesicles was stimulated by extravesicular Na+, K‘, Rb’, or Li+ but not by choline or N-methylglu- camine. Uptakes of Na’ and Rb+ were half-saturated at 3 IIIM Na’ and 3 RIM Rb+, but the maximal velocity of Na+ uptake was 1.5 times that of Rb+ uptake. Na+ uptake was inhibited by extravesicular K+, Rb+, or Li+, and Rb+ uptake was inhibited by extravesicular Na‘ or Li+. Amiloride-sensitive uptake of Na’ or Rb’ in- with extravesicular and intravesicular the and

Uptakes of Na' and Rb+ were half-saturated at 3 IIIM Na' and 3 RIM Rb+, but the maximal velocity of Na+ uptake was 1.5 times that of Rb+ uptake. Na+ uptake was inhibited by extravesicular K+, Rb+, or Li+, and Rb+ uptake was inhibited by extravesicular Na' or Li+. Amiloride-sensitive uptake of Na' or Rb' increased with increase in extravesicular pH and decrease in intravesicular pH. In the absence of pH gradient, there were stimulations of Na+ uptake by intravesicular Na+ and K' and of Rb' uptake by intravesicular Rb' and Na+. Similarly, there were trans stimulations of Na+ and Rb' efflux by extravesicular alkali cations.
The data suggest the existence of a nonselective antiporter catalyzing either alkali cation/H+ exchange or alkali cation/alkali cation exchange.
Since increasing Na' caused complete inhibition of Rb+/H+ exchange, but saturating K' caused partial inhibitions of Na'/H+ exchange and Na+/Na+ exchange, the presence of a Na+-selective antiporter is also indicated.
Although both antiporters may be involved in pH homeostasis, a role of the nonselective antiporter may be in the control of Na+/K+ exchange across the cardiac sarcolemma.
The existence of alkalic cation/H+ antiporters in energytransducing membranes was postulated by Mitchell (l), and an Na'/H' antiport activity was first demonstrated by Mitchell and Moyle (2) in the mitochondrial inner membrane. Subsequently, it has been established that the plasma membranes of many cells, including the cardiac sarcolemma, also contain Na+/H' antiporters that are involved in the regulation of cytosolic pH and a variety of other cellular functions (3)(4)(5). Although the role of a plasma membrane Na'/H+ antiporter in cellular pH homeostasis must clearly depend on its ability to discriminate among the various cellular cations, the ion selectivity of the antiporter has not been examined thoroughly in all membranes in which it has been detected. Here, we present data that suggest the existence of two distinct amiloride-sensitive antiporters in the purified cardiac sarcolemmal vesicles. One is the Na+-selective Na'/H' anti- porter that transports Na+ but not K' or Rb'. The other is a nonselective alkali cation/H+ antiporter that transports all alkali cations including Na+, K', and Rb'. The physiological role of the latter may be in the regulation of Na'/K' exchange in addition to the control of intracellular pH.

EXPERIMENTAL PROCEDURES
Membrane Preparations-Cardiac sarcolemmal vesicles were prepared from fresh beef heart by the method of Jones (6) as described before (7) was filtered and washed as indicated (9). When not specified, duration of the uptake experiment was 20 s. Experiments such as those of Fig. 1 showed that 20 s was at the upper limit of the period during which the H+ gradient-dependent uptake was a linear function of time. Hence, the uptake rates presented under "Results," may be somewhat underestimated.
In some experiments (see Figs. 5, 6, and 10) the pH values of the above loading and assay media were changed by the replacement of the indicated buffers and MGA with 40 mM ACES and 20 mM MGA (pH 6.4-6.5);   (14). The initial rate of release of P, was determined calorimetrically. Materials-**Na+, *'Rb+, and @K+ were purchased from Du Pont-

Sarcolemmal
Na'/H", K+/H+, and Rb'/H+ Exchange Actiuities-Cardiac sarcolemmal vesicles were preequilibrated in acid (pH 6.0) or alkaline (pH 8.0) media and then diluted into media of varying pH, all of which contained 1 mM *'Na+. When the uptake of **Na+ by these vesicles was monitored as a function of time, the results ( Fig. la) confirmed the finding of others (9) by showing that the initial rate of **Na+ uptake in the presence of an outward H' gradient was severalfold greater than the rate of uptake either in the absence of a gradient or in the presence of an inward H+ gradient. To test the ion specificity of the system, similar experiments were done in which 22Na+ was replaced with either 86Rb+ or 42K+. The results (Fig. lb) were nearly identical for "Rb' and 42K+ and qualitatively the same as those obtained with *'Na+. Experiments similar to those of Fig. 1 were done in which the sarcolemmal vesicles were replaced with beef heart mito- Procedures"), it was established that no more than 3% of the H+ gradient-dependent uptakes of alkali cations noted in Fig. 1 could be due to mitochondrial contamination.
When experiments similar to those of Fig. la were done with vesicles that were exposed to 1 mM ouabain during the overnight loading period, the results (not shown) were nearly identical to those of Fig. la, suggesting that the observed Na' uptake was not mediated by the sodium pump.
To determine if the H+ gradient-dependent uptake of an alkali cation was accompanied by the release of H' from the vesicles, experiments of Fig. 2 were done. Vesicles that were preloaded with the acid medium (pH 6.0) were added to the alkaline (pH 8.0) medium containing acridine orange. There was a rapid quenching of acridine orange fluorescence, which is known to be due to the outward H+ gradient in such vesicles (9, 11, 12), followed by a slow increase in fluorescence, which is due to the spontaneous dissipation of the H' gradient (9) Vesicles were loaded with the acid (pH 6.0) medium and then diluted into alkaline (pH 8.0) and acid (pH 6.0) media containing the indicated concentrations of "NaCl or %RbCl. Uptake rates were measured as indicated under "Experimental Procedures." The indicated value at each Na' or Rb' concentration is the difference between uptake at pH, 8 and uptake at pH, 6. a, vesicles were loaded with the pH 6 buffer and diluted into the pH 8 buffer solution containing 1 mM **Na+ or 1 mM %Rb' and the indicated amiloride concentrations. Uptake rates were then measured as described under "Experimental Procedures." 6, **Na+ uptake rates were determined as indicated above in media containing the indicated Na' concentrations in the absence of amiioride and in the presence of indicated concentrations of amiloride. mM Na+ and 1 mM Rb', respectively, the apparent K; value of amiloride was 0.2-0.3 mM (Fig. 4a). Amiloride inhibition seemed to be competitive with respect either to Na' (Fig. 4b) or to Rb' (data not shown). In experiments similar to those of Fig. 4a, quinacrine and quinine also inhibited both exchange activities with apparent Ki values of 10 pM quinacrine and 60 pM quinine (data not shown).
Effects of pH, an d pH, on Na'/H+ and Rb+/H+ Exchange Activities-When pHi was kept constant at 6, and pH, was varied in the range of 6-9.4, both Na' uptake and Rb' uptake rates increased with increase in pH, and seemed to level off at pH, values higher than 8.3 (Fig. 5). The value of pH, for half-maximal stimulation of Rb' uptake seemed to be lower than the same value for the stimulation of Na' uptake, but the data were not good enough to establish this point (Fig. 5). In experiments of Fig. 6, pH, was kept constant (at 8 in experiments with Na' and at 8.3 in experiments with Rb'), and the effects of changes in pHi on Na' and Rb' uptake rates were studied. Both rates increased with decrease in pHi (Fig. 6). Although there was no evidence that either uptake rate had reached a maximal value at the lowest tested pHi, both exchange systems clearly seemed to approach saturation with increase in intravesicular H+ concentration (inset to Fig.  6). Note that in contrast to observations on other plasma membrane vesicles (3,34), the data of Fig. 6  of Na* or Rb+. Vesicles were loaded with the pH 6 buffer and then diluted into buffers of the indicated pH values containing either 1 mM '*Na+ or 1 mM s6Rb+. Uptake rates were measured as indicated under "Experimental Procedures." Vesicles were loaded with the pH 6 buffer and then diluted into buffers for uptake measurements containing either 1 mM *'Na+ or 1 tnM %Rb+ and the indicated concentrations of KC1 or NaCl. Uptake rates were measured as described under "Experimental Procedures." For each indicated condition, uptake was measured at pH. 8 and pH, 6; the difference between the two was taken as the value of the exchange rate.
pendent uptake of 22Na+, and extravesicular Na+ inhibited the H+ gradient-dependent uptake of "Rb+. Dixon plots of these inhibitory effects (Fig. 7) revealed K, values of 4 mM K' (Fig. 7a) and 2.5 mM Na+ (Fig. 7b), in close agreement with the Km values obtained in Fig. 3. Taken together, these data suggest that Na' and K' (or Rb') are alternate substrates that compete at the same site on an alkali cation/H+ exchanger. When the inhibitory effects of Na+ concentrations higher than those used in Fig. 7 were examined on H+ gradientdependent 86Rb' uptake, complete inhibition of "Rb' uptake was obtained with increasing Na' concentration ( Fig. 8~). In similar experiments on the effects of K' on '*Na+ uptake, however, the maximal level of inhibition of Na' uptake obtained at saturating K' concentration was less than 100% ( Fig. 8b; see also Fig. 11,below).
When Rb' was used instead of K' in experiments similar to those of Fig. 8b, the results were the same (not shown). This partial inhibition of Na'/ H' exchange by K+ or Rb' is not consistent with competition between Na' and K', or Na' and Rb', on a single carrier (see "Discussion").
In experiments similar to the above, complete inhibition of either Na'/H' or Rb'/H' exchange activity was approached by increasing Li' concentrations ( Fig. 8, a and b); however, up to 50 mM concentrations of choline and tetraethylammonium ions did not inhibit either activity (data not shown).
Alkali were varied in the range of 100-300 mM to maintain intra-and extravesicular osmolarities the same (see "Experimental Procedures"). mM concentrations of unlabeled Na', K', or Rb' as indicated. (Note that for technical reasons when the intravesicular cation was different from the labeled extracellular cation, it was necessary that the extravesicular medium also contain some of the intravesicular cation.) The results showed that uptake of "Na+ was stimulated by intravesicular Na+ or K' and that uptake of 86Rb' was stimulated by intravesicular Rb+ or Na+. The results also showed that the largest degree of trans stimulation was that of Na' uptake by intravesicular Na' (Table I).
Cardiac sarcolemmal vesicles develop a membrane potential if vesicles with a transmembrane K' gradient are exposed to valinomycin, and this potential is not significantly affected by a transmembrane Na' gradient (15,16). When the experiments of Table I on the effects of intravesicular K' on "Na+ uptake were repeated with the inclusion of 2 PM valinomycin during the course of the assay, the results were not different from those of Table I, suggesting, but not establishing, that the trans stimulatory effects of alkali cations observed in Table I are independent of the membrane potential. Experiments of Fig. 9 showed that "Na+ efflux from the vesicles was stimulated by extravesicular Na' or K' (Fig. 9a) and that 86Rb+ efflux was stimulated by extravesicular Na' or Rb+ (Fig. 9b). Here, as in the case of influx experiments, the largest trans stimulation was that of Na' efflux by extravesicular Na' (Fig. 9). Taken together, the data of Table I and Fig. 9 establish the existence of alkali cation/alkali cation exchange activities and indicate that the rate of such exchange is dependent on the nature of the alkali cation on the two sides of the membrane.
The above experiments were done in the absence of an imposed pH gradient. To examine the effect of such gradient on Na'/Na' exchange, the rate of exchange was determined at several pH, values in the range of 6-9 under two conditions: when pHi was equal to pH,, and when pHi was held constant at 6. The results (Fig. 10) showed that (a) Na+/Na+ exchange was reduced when there was a transmembrane H' gradient; and (b) either with or without a pH gradient, Na+/Na' exchange increased with increase in pH,. These results are consistent with the assumption that Na'/H' exchange and Na'/Na+ exchange occur on the same antiporter and that Na' and H' compete at both sides of the membrane.
In experiments of Fig. 11, the effects of extravesicular K' on amiloride-sensitive Na+/Na+ exchange, measured through the uptake of 'lNa+ by Na'-loaded vesicles, were studied. K' was a partial inhibitor of Na'/Na' exchange ( Fig. 11) as it was a partial inhibitor of Na+/H+ exchange (Fig. 8b). Orientation of the Sarcolemmal Vesicles-Although there is agreement that purified cardiac sarcolemmal vesicle preparations are mixtures of leaky vesicles, sealed inside-out vesicles, and sealed right side-out vesicles, there is considerable disagreement on the relative proportions of these vesicles in the various preparations (e.g. 6, 17-19). Since for the interpretation of the present data it was desirable to know the orientation of the sealed vesicles, we used the assay of (Na' + K+)-ATPase (see "Experimental Procedures") to determine the sidedness of the vesicles as others have done (6,17,18). We considered (a) the ouabain-sensitive activity in the presence of alamethicin as the total activity of all vesicles; (b) the ouabain-sensitive activity in the absence of any ionophore as that of the leaky vesicles; and (c) the digitoxigenin-sensitive activity in the presence of valinomycin and carbonylcyanide m-chlorophenylhydrazone as the sum of the activities of the leaky and the inside-out vesicles. When three preparations selected at random were assayed by this procedure, we found 38.7 + 2.6% of the vesicles to be leaky. Of the remaining sealed vesicles, 94.0 + 5.7% were right side out and the remainder inside out. In agreement with Jones (6) we conclude that the sealed vesicles are predominantly right side out. Evidently, variations among the preparative procedures lead to significant differences in the ratio of the two types of sealed vesicles.
Experiments with Other Sarcolemmal Vesicles-Limited experiments showed that the Na+/H+, Rb'/H', and alkali cation/alkali cation exchange activities described above also existed in the sarcolemmal vesicles of the canine heart obtained by the same procedure used to prepare the beef heart vesicles.

DISCUSSION
The Two Antiporters of the Cardiac Sarcolemma- Seiler et al. (9), who were the first to demonstrate an Na+/H+ exchange activity in cardiac sarcolemmal vesicles, also noted (Footnote 2 of Ref. 9) the existence of a "small amount" of K'/H' exchange in these vesicles but did not pursue their study of this activity. The data presented here characterize this K'/ H ' (or Rb'/H') exchange and its relation to the Na'/H' exchange activity. Previous studies with cardiac sarcolemmal vesicles demonstrated that an outward H' gradient stimulates Na' uptake by the vesicles (9, 20) and that this amiloridesensitive Na' uptake represents the operation of an electroneutral Na+/H+ antiporter (9). We have shown that (a) Na' uptake on this antiporter is inhibited by K' or Li+ on the cis side (Figs. 7 and 8); (b) there are also amiloride-sensitive uptakes of K+ and Rb' that are stimulated by H' on the trans side ( Figs. 1 and 4) and inhibited by Na+ or Li on the cis side (Figs. 7 and 8); (c) release of intravesicular H' is stimulated by Na+, K+, Rb', or Li', on the trans side (Fig. 2); (d) there are trans stimulations of Na+ uptake by Na+ or K+ and of Rb' uptake by Rb+ or Na+ (Table I); and (e) release of intravesicular Na+ is stimulated by Na+ or K+ on the trans side, and release of intravesicular Rb' is stimulated by Rb' or Na' on the trans side (Fig. 9). Taken together, these findings clearly indicate the existence of a nonselective antiporter that is capable of catalyzing either alkali cation/H+ exchange or alkali cation/alkali cation exchange. The same data that establish the nonselectivity of the antiporter, however, also show significant quantitative differences in the exchange activities involving Na+, K+, and Rb'. Some of these differences may be explained simply by assum-Ion Specificity of Na+/H+ Antiporter ing that alkali cations share a common binding site on the same antiporter but that within the antiport reaction cycle there are differences between one or more kinetic constants depending on the nature of the bound alkali cation (21). Two sets of our experiments, however, are difficult to reconcile with such a single antiporter: the data showing that K+ and Rb' are partial inhibitors of Na'/H' exchange and Na'/Na' exchange (Figs. 8 and 11,and "Results"), and the results indicating that Na' causes complete rather than partial inhibition of Rb'/H' exchange ( Fig. 8).* These findings can be explained if we assume the existence of two distinct antiporters in the sarcolemmal vesicles: the nonselective antiporter that transports Na+, K', and Rb' and on which the alkali cations compete; and a selective antiporter that carries Na' but not K' or Rb'. These antiporters can also account (a) for the higher maximal velocity of Na'/H+ exchange than that of Rb+/H+ exchange (Fig. 3); and (b) for the observation that in alkali cation/alkali cation exchange experiments (Table I  and Fig. 9), the largest trans stimulation is that of Na' flux by Na'. We should also note that our limited data on Li' (Figs. 2 and 8) suggest that it is a substrate for both antiporters.
Relation to Other Alkali Cation/H+ Antiporters-The plasma membrane Na'/H' antiporter that has been studied so extensively (3-5, 21) seems to have great selectivity for Na' uersus K' in most tissues. K+/H+ exchangers, however, have been identified in the brush-border membrane of the ileum (12) and suggested in red cell and bacterial plasma membranes (22,23). In mitochondria, in which the necessity of K+/H+ exchange was recognized long ago, a nonselective alkali cation/H+ antiporter that accepts either Na' or K' but functions as a K'/H' exchanger physiologically has been identified as an 82,000-kDa protein and reconstituted (24,25). Interestingly, the mitochondria also contain a second selective Na+/H+ antiporter that does not carry K+ (24,25). Garlid (26) has discussed the possibility that these antiporters may be members of the same family.
Physiological Implications-In the literature dealing with the function of Na+/H+ exchange in cardiac myocytes, it is often stated that K+ and Rb' can not substitute for Na+ (e.g. 27-29). A closer examination shows that in some cases (28, 29) the stated selectivity of the cardiac antiporter for Na' uersus K' is based on cited references on the selectivity of the antiporter in other tissues. In experiments in which the direct test of the specificity has been attempted in heart cells (Fig. 8 of Ref. 30), the data show the operation of a selective Na/ H' exchange activity but do not rule out the existence of an additional nonselective exchange mechanism that can catalyze K'/H' and K+/K+ exchanges. We consider it prudent to assume the existence of both antiporters in the reexamination of previous work and in the design of new experiments on intact cardiac myocytes.
Because the sealed sarcolemmal vesicles used here are about 90% right side out (see "Results"), the apparently equal K, and Ki values of Na+, K', and Rb' (Figs. 3 and 7) must refer to cation affinities at the extracellular side of the membrane. Considering this and the high ratio of extracellular Na' to K', it is likely that under physiological conditions, the nonspecific antiporter will also be carrying extracellular Na' in preference to K*. The situation at the intracellular side will depend on the relative affinities of Na+, K', and H' for the inward facing conformation of the nonselective antiporter. Although these affinities remain to be determined, our present ' If Na+ were also a partial inhibitor, one could still assume that Na' and K' bind to the same antiporter at different sites and that the two substrates are antagonistic but not mutually exclusive. experiments on alkali cation/alkali cation exchange (Table I  and Fig. 9) clearly show that the antiporter is not so asymmetrical that it would exclude K+ at the intracellular side. There is also the possibility that the antiporter properties in the purified sarcolemmal vesicles may be different from those in the intact cell. That the selective Na'/H' antiporter of the plasma membrane is regulated in a variety of tissues other than the heart is known (4, 5), and the physiological regulation of the nonselective mitochondrial antiporter is well established (26). In intact cardiac myocytes, there is evidence to suggest that the pH, dependence of the Na+/H' exchange may be regulated by protein kinase C (31). In this regard, we should note that our experiments on the dependence of Na+/ H+ exchange and Rb'/H' exchange on pHi (Fig. 6) do not reveal an intracellular H' modifier site that has been suggested by experiments on some intact myocyte preparations (31)(32)(33) and in other tissues (5,34). It remains to be seen if this reflects the absence or the regulated alteration of the H' modifier site in the nonselective antiporter that is predominant in our sarcolemmal preparation.
The above uncertainties notwithstanding, we believe it is reasonable to consider the possibility that without the imposition of an intracellular acid load, the nonselective antiporter may tend to catalyze Na+/K+ exchange and the coupled dissipation of the gradient that is maintained by (Na' + K+)-ATPase. This may seem wasteful at first sight. Regulation of such a futile cycle, however, is a prevalent mechanism for the efficient control of many biological systems. It is appropriate to recall the long standing evidence for the coordinated control of the "pump/leak" system in the maintenance of cell volume (35).