Calcium Transport and Monovalent Cation and Proton Fluxes in Sarcoplasmic Reticulum Vesicles*

ATP-dependent Ca2+ uptake by rabbit skeletal muscle sarcoplasmic reticulum vesicles has been studied in the presence and absence of artificially generated pH gradients and membrane potentials. H+ and K+ diffu- sion potentials were generated via the H+ and K,Na channels of sarcoplasmic reticulum by transfer of ves- icles from low to high pH, or from high to low K’. Membrane potentials were measured using the voltage-sensitive fluorescent dye 3,3’-dipentyl-t,2’-oxacarbo- cyanine. The initial rate of Ca2+ uptake was found to be increased in the presence of a pH gradient and mem- brane potential (negative inside). In turn, the rates of decay of K’- or H’-induced membrane potentials were accelerated during Ca2+ transport, suggesting that active Ca2+ uptake stimulated the release of K+ and H+ from the vesicles. The ratio of K’ (or H’) release to Ca2+ transport was near two. Release of K’ did not appear to be directly catalyzed by the Ca2’-ATPase. Evidence against a directly coupled ATP-mediated 2 K+-Ca2+ or K+-Ca2+ exchange re- action was that (i) similar results were obtained when K+ was substituted by Na’ or by organic cations which could rapidly permeate through the channel of K+,Na+-permeable vesicles and (ii) Ca2+ transport did not result in an equivalent release of 86Rb+ or “Na+ from K+,Na+-impermeable vesicles. These studies are in support of an electrogenic Ca2+ transport system in sarcoplasmic reticulum. The

ATP-dependent Ca2+ uptake by rabbit skeletal muscle sarcoplasmic reticulum vesicles has been studied in the presence and absence of artificially generated pH gradients and membrane potentials. H+ and K+ diffusion potentials were generated via the H+ and K,Na channels of sarcoplasmic reticulum by transfer of vesicles from low to high pH, or from high to low K'.
Membrane potentials were measured using the voltagesensitive fluorescent dye 3,3'-dipentyl-t,2'-oxacarbocyanine. The initial rate of Ca2+ uptake was found to be increased in the presence of a pH gradient and membrane potential (negative inside). In turn, the rates of decay of K'or H'-induced membrane potentials were accelerated during Ca2+ transport, suggesting that active Ca2+ uptake stimulated the release of K+ and H+ from the vesicles. The ratio of K' (or H') release to Ca2+ transport was near two.
Release of K' did not appear to be directly catalyzed by the Ca2'-ATPase. Evidence against a directly coupled ATP-mediated 2 K+-Ca2+ or K+-Ca2+ exchange reaction was that (i) similar results were obtained when K+ was substituted by Na' or by organic cations which could rapidly permeate through the channel of K+,Na+permeable vesicles and (ii) Ca2+ transport did not result in an equivalent release of 86Rb+ or "Na+ from K+,Na+impermeable vesicles.
These studies are in support of an electrogenic Ca2+ transport system in sarcoplasmic reticulum. The results further suggest that during Ca2+ transport development of a membrane potential (positive inside) is likely nullified by the countermovement of the permeant cations K+, Na', and H'.
Isolated sarcoplasmic reticulum (SR)' vesicles rapidly sequester e a 2 + against a concentration gradient upon energization with ATP through the action of a membrane-bound, Mg"-dependent, Ca"-stimulated ATPase (see reviews by Tada et Hasselbach, 1978;de Meis andVianna, 1979, Inesi, 1979). The Ca"-ATPase is a major membrane component of skeletal muscle SR accounting for up to 90% of the total protein (Meissner, 1975). Recently, evidence has been presented that the Ca"'-ATPase, when incorporated into phospholipid bilayer vesicles (Zimniak and Racker, 1978), transfers positive charges into vesicles during Ca2' transport.
* The work was supported by Research Grant AM 18687 from the United States Public Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Gerhard Meissner
From the Deoarfments of Biochemistm and Phvsiolom, School of Medicine, University of North Carolina, Chapel Hill, The phospholipid bilayer model system offers the advantage of being relatively impermeable to ions. Study of the electrogenic nature of eaz' transport in native SR vesicles is more difficult, because these membranes are permeable to ions in a complex manner. SR vesicles are relatively impermeable to CaZ+, Mg2+, and larger ions such as gluconate-, choline', or Tris' (Meissner and McKinley, 1976;Nagasaki and Kasai, 1980). About two-thirds of the vesicles contain a K,Na channel making them highly permeable to K', Rb', and Na'. The remaining one-third of the vesicles lack the K,Na channel and are therefore relatively impermeable to these cations (Mc-Kinley and Meissner, 1978). K',Na'-impermeable vesicles are thought to arise as a consequence of a limited number of K,Na channels in SR. Both K+,Na'-permeable and -impermeable vesicles are highly permeable to protons and chloride suggesting that these 2 ions can pass the sarcoplasmic reticulum membrane by a pathway separate from that of the K,Na channel (McKinley and Meissner, 1978;Meissner and Young, 1980). In this study we have taken into consideration the particular permeability properties of native SR vesicles to investigate the effect of ea'+ transport on monovalent cation and proton fluxes. The data suggest that active uptake of positively charged Ca" stimulates the release of an equivalent amount of charge including K+, Na+, or H'. With the possible exception of H+, movement of these ions does not appear to be directly catalyzed by the ea'+-ATPase. Some of these results have appeared elsewhere in a preliminary form (Meissner, 1979).

MATERIALS AND METHODS
Reagents-The fluorescent dye 3,3"dipentyl-2,2"oxacarbocyanine was the generous gift of Dr. Alan S. Waggoner (Amherst College, Amherst, Mass.). Gluconic acid (technical grade, Eastman, Rochester, N. Y.) was treated with charcoal before use to prepare the salt Tris/ gluconate. Analytical grade reagents were used otherwise.
Sarcoplasmic Reticulum Vesicles-"Intermediate" density rabbit skeletal muscle sarcoplasmic reticulum vesicles used in this study have been characterized previously (Meissner, 1975). Unless otherwise indicated, vesicles (0.5 to 1.0 mg of sarcoplasmic reticulum protein/ml) were incubated for 6 to 15 h a t 0°C in 480 mosM K/ gluconate or Tris/Pipes medium at pH 7. Vesicles were sedimented by centrifugation for 30 min at 35,000 rpm in a Beckman 42.1 rotor, resuspended in a small volume (15 to 20 mg of protein/ml) of the above media, incubated for another 2 to 3 h a t O"C, and then stored a t -65°C for later use.
Membrane Potential Measurements-Membrane potentials were generated by gradients of permeant ions between the intravesicular cavity and the medium into which the vesicles were diluted. Membrane potentials (negative inside) were detected by the use of the fluorescent dye 3,3'-dipentyl-2,2'-oxacarbocyanine iodide (diO-Cd3)) (Sims et al., 1974) as previously described (McKinley and Meissner, 1978). Formation of positive membrane potentials and pH gradients did not affect significantly the fluorescence emission of diO-Cd3) under the conditions used in the present study (Meissner and Young,

636
This is an Open Access article under the CC BY license. 1980). The polarity of membrane potentials are reported according to standard convention, that is, reference (ground) is extravesicular. Unless otherwise indicated, fluorescence assays were carried out at 15°C under stirring in a Farrand model 801 Fluorometer. Excitation was at 470 nm and emission was recorded at 495 nm. Slits used resulted in a half-band width of 2.5 nm. Vesicle concentrations (approximately 15 pg of protein/ml) were used which produced negligible perturbation of the fluorescence emission during dilution with incubation medium.
Isotope Flux Measurements-Radioisotope efflux rates from sarcoplasmic reticulum vesicles were measured as previously described (Meissner and McKinley, 1976). Briefly, vesicles were incubated for 6 to 15 h at 0°C in the presence of a radioactive compound. The vesicles were then diluted 400-fold into an unlabeled release medium with rapid mixing. Efflux of the radioactive compounds was monitored at various time intervals by placing 1.0-ml aliquots on 0.45 pm HAWP Millipore filters followed by rapid rinsing with unlabeled medium. The time required to execute filtration and rinsing was 15 to 20 s and was taken into account. The radioactivity retained on filters was counted in a liquid scintillation system.
Ca2' Transport Measurements-Ca'+ uptake was initiated by the addition of SR vesicles (20 to 50 pg of protein) to 1 ml of assay medium containing 50 pM 45Ca2+ (0.01 pCi/ml), 1 mM M e , and 0.5 mM ATP. In order to stop "%a2' uptake and 45Ca'+ efflux, 10 pl of 1 M LaCL was added (Meissner and McKinley, 1976;Chiesi and Inesi, 1979). Only minimal "Ca*+ uptake was observed when La.'+ was added to the assay media prior to the vesicles. Duplicate aliquots (0.4 m l ) were placed on 0.45 pm Millipore filters and '%a2+ not taken up by the vesicles was removed by rinsing the filters with assay medium containing 10 mM La'+ but no ATP. Radioactivity retained on fdters was counted in a liquid scintillation system (Meissner and McKinley, 1976).

Effect of Ca2+
Uptake on K" a n d €€+-induced Membrane Potentials-Membrane potential measurements with the fluorescent dye diO-C5-(3) are an effective means of distinguishing SR vesicles which contain the K,Na channel (type I) and those that do not (type 11). As previously shown (McKinley and Meissner, 1978), negative membrane potentials could be generated in K',Na+-permeable (type I) vesicles by dilution of K/gluconate-filed vesicles into Tris/gluconate media ( or presence (B, C) or 0.5 pM valinomycin (Val). Dilution media contained 1.5 pM diO-C~-(3), 1 mM Mg", 50 pM Ca2+, no or 0.5 mM ATP, and 10 mM Pipes, 20 m M Tris, pH 7.0. The K' of the vesicle medium served to establish an initial 60-fold K' gradient throughout the experiments. Dotted lines represent the initial slopes of fluorescence rebound curves and are extrapolated back to the time of vesicle addition. In A, only type I vesicles are expected to be polarized by dilution; in B, only type I1 vesicles; in C, both type I + I1 vesicles; in D, none of the vesicles. 1A). The absence of a potential in K',Na'-impermeable (type 11) vesicles was due to the similarly low permeability of K' and Tris'. On the other hand, negative potentials were exclusively formed in type 11 vesicles by transferring the entire SR vesicle population from K/gluconate medium to Na/gluconate medium containing the K' selective ionophore valinomycin (Fig. 1B). Under these conditions, no membrane potential was formed in K',Na'-permeable (type I) vesicles. Because of the presence of the K,Na channel, these vesicles rapidly exchanged all of their K' for Na+ within 1 to 2 s, the experimental limit of detection. Use of valinomycin in Tris/gluconate dilution medium enabled development of a membrane potential in all vesicles. Accordingly, the trace of Fig. 1C exhibited a deflection which was nearly the sum of the signals seen for type I (Fig. L 4 ) and type I1 (Fig. 1B) vesicles.
K',Na'-permeable and -impermeable vesicles are permeable to H' (Meissner and Young, 1980). Permeability of vesicles to both K' and H' resulted in rapid K'-H' exchange until K' and H' gradients of equal size and direction were formed. The extent of K' for H' exchange was minimized in the experiments of Fig. 1 with the use of the impermeable nonbuffering gluconate anion (Meissner and Young, 1980). External cation concentrations were considered to be essentially unaffected by K'-H' exchange since the dilution media were an infinite bath compared to the total volume of the vesicles.
Fluorescence emission of diO-C,-(3) reached a minimum within 1 to 2 s after the addition of the vesicles to the three polarizing media (Figs. 1, A to C). The signal returned within 10 to 20 min to that observed for nonpolarized vesicles (Fig.  ID). This gradual collapse was likely due to inward movement of Tris' (Fig. 1, A and C) or Na' (Fig. lB), and the eventual dissipation of the K' (and H') gradients. The control nonpolarized signal was obtained by diluting Tris/gluconate-fdled vesicles into Tris/gluconate medium (Fig. 1D). A similar trace was obtained when K/gluconate-filled vesicles were diluted into K/gluconate medium (not shown).
The ability of SR vesicles to generate and maintain membrane potentials in the presence of an ion gradient has been used to study cation movements across SR membranes which actively transported Ca"'. Fluorescence signals indicated that the developed membrane potentials returned faster to baseline when media contained 0.5 mM ATP, 50 p~ Ca", and 1 mM Mg" (Fig. 1, A to C), i e . when SR vesicles actively transported Ca" (see below). The effect of ATP (in the presence of Ca" and Mg") on shortening the lifetime of membrane potentials in type I (K -+ Tris) or type I1 (K + Na + Val) vesicles was more pronounced at 21°C than at 3°C (Table I).
Addition of ATP or Ca'+ uptake per se were likely not responsible for the enhanced breakdown of the fluorescence signals. Russell et al. (1979) and  have found that the fluorescence of oxacarbocyanine dyes responds to binding of ATP and Ca2+ to the membranes. However, experimental conditions used by these investigators were different from the present study. We observed a normal slow rebound of fluorescence signals in type I + I1 vesicles when Ca2' was omitted from the media or ATP was substituted by the nonhydrolyzable ATP analog adenylyl-(/?,y-methylene)diphosphonate (AMP. PCP) ( Table 11). In the absence of an initial potential where Tris/gluconate-filled vesicles were diluted into Tris/gluconate medium, fluorescence signals were not significantly changed by ATP (Fig. 1D). Addition of 10 mM Ca2+ to the assay media (in the absence of ATP) also showed no effect (not shown). In another control, we tested the effect of internal Ca2' on diO-C5-(3) fluorescence. Vesicles were passively loaded with Ca" by preincubation in K/gluconate medium in the presence of 1 mM or 5 mM Ca". Rapid

Effect of Ca2+ transport on lifetime of cation-induced membrane potentials
SR vesicles filled with 480 mosM K,Na or monoethanolarnine (EtOH.NHn)/gluconate at pH 7 were diluted 20-fold into isoosmolal Tris/gluconate or Na/gluconate media a t pH 7. The Na/gluconate dilution medium contained 0.5 p~ valinomycin to render type I1 vesicles permeable to K' . All dilution media contained 1.5 p~ di0-CS-(3), 1 mM Mg", 50 pM ca2+, 20 mOSM Tris/Pipes, and no or 0.5 mM ATP. tlIP was the time in which the fluorescence signal, extrapolated to time of vesicle addition (cf. Fig. l), returned to half its final value. Type of vesicles expected to be polarized by dilution is indicated.

c5-(3)
K' diffusion potentials were created in type I + I1 vesicles by diluting vesicles 50-fold from a 480 m o m K/gluconate medium at pH 7 into a solution of 460 mOSM Tris/gluconate and 20 mOSM Tris/Pipes at pH 7 (A). Type I1 vesicles were rendered permeable to K' by the addition of 0.5 p~ valinomycin. The extravesicular K' from the sample served to define the size of the K' gradient (50-fold). H' diffusion potentials were established by diluting vesicles present in Tris/Pipes at pH 6.1 into Tris/Pipes media at pH 7.6 (B). Dilution media contained 1.5 p~ diO-C~-(3), and where indicated 1 mM Mg2' , 50 PM Can+, 0.5 m~ ethyleneglycol-bis-(&xninoethyl ether)-N,N" tetraacetic acid (EGTA), 0.5 mM ATP, or 0.5 mM adenylyl-(P,ymethylene)-diphosphonate (AMP. PCP). Fluorescence decreases were obtained by back-extrapolation to the time of vesicle addition (cf. Figs. 1 and 2). tlrL was the time in which the extrapolated fluorescence signal returned to half its final value. and slow transient fluorescence signals similar to those seen in Fig. 1 were observed in the presence and absence of ATP, respectively. Further, light-scattering effects could not account for the observed rapid changes in fluorescence intensity.
No or only minimal changes in light intensity at 495 nm (X,,, = 470 nm) were recorded during Ca2' uptake when the dye was omitted.
Our explanation for the Cap' uptake-induced enhancement of potential breakdown is an increase in the rate of K' (and H' ) release during Ca" transport. It appeared likely that intrinsic K' and H' permeability was sufficient to mediate increased cation efflux since addition of increasing amounts of valinomycin and the H+-carrier carbonyl cyanide p-trifluoromethoxyphenyl hydrazone did not cause a more rapid collapse of membrane potentials during Ca" transport (not shown).
On the other hand, because of the K' permeability of the vesicles, the above experiments could not reveal whether K' release was catalyzed by an obligate exchange process directly coupled with Ca2+ translocation or whether active uptake of Ca2' was compensated in charge by release of K' by an independent mechanism such as the K,Na channel of SR. To establish the mechanism of K' release, we studied the effect of Ca2' transport on membrane potentials created by permeable ions other than K' . We expected that an obligate exchange process would be fairly specific with regard to the number and kind of ions transported during Ca2+ transport, while an independent mechanism would only require that the membrane is permeable in some way to the released ions.
Membrane potentials were generated in type I SR by transferring Na/ or monoethanolamine/gluconate-filed vesicles to Tris/gluconate medium. Both cations were previously found to pass rapidly through the K,Na channel (McKinley and Meissner, 1978). Uptake of Ca2' increased the rate at which fluorescence signals returned to the base-line (Table I). These results would suggest that membrane potentials formed by a K', Na' or monoethanolamine' gradient were similarly rapidly nullified during active Ca2' transport.
Negative H' diffusion potentials were established as previously described (Meissner and Young, 1980) by diluting SR vesicles from a Tris/Pipes medium at pH 6.2 into a Tris/Pipes medium at pH 7.6 ( Fig. 2). In these experiments, Pipes was used as the impermeable anion to increase the internal buffering capacity of the vesicles, and thus the lifetime of the membrane potentials. As above (Fig. 1, Table II), addition of 50 PM Ca2', 0.5 mM ATP, and 1 mM Mg2+ enhanced the rate with which the fluorescence signal returned to base-line. This result indicated that active uptake of Ca" accelerated the collapse of H'-induced membrane potentials as well. Together, the similar behavior of membrane potentials formed by K', Na', monoethanolamine, or H' gradients was in support of independent but complementary movement of K' rather than movement directly catalyzed by the Ca"-ATPase.
Effects ofR+-or €€+-induced Membrane Potentials on CaZ+ Uptake-If the Ca2' transport system of SR is electrogenic, one might expect that the initial rate of Ca" uptake is retarded by a positive potential and stimulated by a negative potential. To test this hypothesis, Ca" uptake rates were measured in the presence of H'and K'-induced membrane potentials. H'-induced membrane potentials were generated in Tris/ Pipes buffer (cfi Fig. 2). In control experiments, in the absence of a pH gradient and potential, at pH 6.2 or 7.6, similar initial rates of Ca2+ uptake were observed in spite of the fact that Ca2' uptake capacity of vesicles was twice as great at pH 6.2 (Fig. 3). Establishment of a negative membrane potential (theoretical Nernst potential AV = -75 mV) by transferring vesicles from pH 6.2 to 7.6, nearly doubled the initial rate of Ca2' uptake. On the other hand, initial Ca2+ uptake was slightly reduced when a positive membrane potential was formed by diluting vesicles from pH 7.6 to 6.2. Thus, the Ca2+ transport system of SR was affected by a H+-induced diffusion potential, a pH gradient, or both.
Membrane potentials were also generated in all SR vesicles using K' gradients in the presence of the K'-ionophore, valinomycin. In standardization experiments in the absence of a potential and gradient Ca2' uptake was faster in 480 mOSM K/gluconate medium than in Tris/gluconate medium containing K/gluconate at a concentration of 48 mom (K -K, Tris + Tris in Fig. 4). The increased rate suggested an activating effect of K' , in agreement with earlier reports (Shigekawa and Pearl, 1976;Duggan and Jacob, 1977;Ribeiro and Vianna, 1978). When vesicles were diluted from high (480 mosM K/ gluconate, pH 7) to low K' concentration (48 mOSM K/gluconate + 432 mOSM Tris/gluconate, pH 7), a negative potential (theoretical Nernst potential AV = -55 mV) and a 10-fold H' gradient (pHi = 6) were formed. Under these conditions, a maximal initial rate of Ca2+ uptake was observed that was 1.5fold greater than the rate in K/gluconate medium (K' -Tris' and K+ -K+ in Fig. 4). Therefore, Ca2' uptake was stimulated, despite a decreased K+ concentration in the dilution medium. Establishment of 10-fold K' and H+ gradients of opposite direction (AV = +55 mV, pHi = 8) had a retarding effect on initial Ca2+ uptake (Tris' + K+). Ca2+ uptake was therefore slowed down, despite an increased K' concentration in the dilution medium (Tris' -+ K' and Tris' -Tris' in Fig.   4). At later times, membrane potentials collapsed (cfi Fig. 1) and similar amounts of Ca2+ were accumulated by vesicles transferred to K' or Tris' media. These results showed that in the presence of a membrane potential, Ca2+ uptake was accelerated and slowed down as predicted by the electrogenic model. Simultaneous formation of a pH gradient prevented, however, ruling out an alternative explanation, that stimulation and retardation in Ca2+ uptake was attributable to a pH gradient rather than a membrane potential. Ca'+ uptake by SR vesicles in the presence and absence of K+-induced membrane potentials. SR vesicles present in 480 mom K/ or Tris/gluconate medium at pH 7 were diluted into isoosmolal K/ or Tris/gluconate media. Tris/gluconate incubation and dilution media contained K/gluconate at a final concentration of 48 mom so that a 10-fold K+ gradient was formed by dilution (0, 0). Dilution media contained 1 mM Mg2+, 50 p~ 45Ca2+, 0.5 mM ATP, 0.5 p~ valinomycin, and 10 m~ Pipes, 20 m~ Tris, pH 7. Ca2+ transport was initiated at 15OC by 100-fold dilution of vesicles and stopped at the indicated time by the addition of 10 mM La3' (cf. "Materials and Methods").

ION GRADIENT ([IN]/[OUT])
FIG. 5. Dependence of fluorescence emission of diO-C5-(3) on size of ion gradients in the absence of Ca2+ transport. Size of K' gradients was varied by replacing in the vesicle medium part of the K/gluconate with Tris/gluconate so that the sum of the two salts remained at 480 mosM. Composition of dilution media was kept constant so that final Tris/gluconate and K/gluconate concentrations were 432 and 48 m o a , respectively. Type I1 vesicles were rendered permeable to K' by the addition of 0.5 p~ valinomycin to the dilution media. Similarly, size of H+ gradients was varied with the use of vesicles that were present in Tris/Pipes media of different pH (pH 6.2 to pH 7.6) and were diluted into Tris/Pipes media of constant final pH (pH 7.6). Fluorescence decreases extrapolated back to the time of vesicle addition are indicated.
that Ca2+ uptake enhances the rate of potential collapse. Since, as noted above, the addition of valinomycin or the proton carrier carbonyl cyanidep-trifluoromethoxyphenyl hydrazone did not promote faster potential breakdown, it appeared that intra-and extravesicular K+ and H' were at electrochemical equilibrium at all times during Ca2+ transport. The likely explanation for Ca2' uptake-induced enhancement of potential breakdown was therefore the induction of faster  and Meissner, 1978), so that the actual amounts of K' or H' released potentials were established as described in the legend of Fig. 2. by the vesicles could be calculated. These were found to be 2.5 pl/mg Amounts of Ca" taken up by the vesicles in the presence of ATP of protein for vesicles present in K/gluconate medium and 2.1 pl/mg were determined as described in the legends of Figs. 3 and 4. K' or of protein for vesicles present in Tris/Pipes medium at pH 6.2. AK' ,,t H' release from the vesicles, in the absence and presence of ATP, and AH'out are the difference in the amount of K' and H' that were was estimated as follows: From the rebound portion of fluorescence released by the vesicles 3, 5, and 10 s after vesicle dilution in the signals (cf Figs. 1 and 2) and the two calibration curves in Fig. 5 Figs. 1 and 2). In the absence of Ca2+ uptake, the slow breakdown in potential was probably due to the inward movement of Tris', a slowly permeating cation (McKinley and Meissner, 1978), and the subsequent outward diffusion of K+ or HZ. The uptake of each Caz+ stimulated the release of about 2 K' or 2 H' more than when no Ca2+ uptake occurred (Table 111). Ratios near 2 were also calculated for K' release/Ca" transport of type I (measured in the absence of valinomycin, Fig. lA) and type 11 (assayed in Na gluconate media containing valinomycin, Fig. 1B) vesicles (not shown).
It was assumed that type I and I1 vesicles were equally capable of Ca2+ transport ( Fig. 1) and had the same average size. Transport ratios of type I or I1 vesicles could therefore be estimated from total vesicle space and the total amount of Ca2' uptake. Thus, Ca2+ uptake appeared to stimulate the release of an equivalent amount of charge in the form of K' and H+, resulting in the faster breakdown in potential.
Effect of ea2' Transport on "Rb+ a n d "Na+ Efflux from K+,Na'-impermeable Vesicles-About one-third of sarcoplasmic reticulum vesicles (type 11) are relatively impermeable to Rb', K', or Na', permitting a direct test of a 2 K+-Ca2+ or K+-Ca"-coupled transport system in SR. Cation efflux rates from type I1 vesicles were determined during or in the absence of Ca2+ transport by Millipore filtration in the presence of the highly permeable type I vesicles (McKinley and Meissner, 1978). '"RbCl (or "NaCl) was equilibrated at a concentration of 10 mM across vesicle membranes by prolonged incubation. To assess nonspecific permeability changes during Ca" transport, incubation media contained in addition the organic cation [''Hlcholine', at a concentration of 190 mM. Sarcoplasmic reticulum vesicles, composed of type I and 11, were then diluted into the unlabeled incubation medium containing 50 PM Ca2', 1 mM Mg2+, without ATP (control), or with 0.5 [3H]choline C1, 5 mM Pipes, 10 mM Tris, pH 7, were diluted 400-fold into an unlabeled medium of the same composition at 15'C, except that the dilution medium contained 1 mM Mg'+ and 50 p~ Ca2', and 0.5 mM ATP where indicated. "Rb+, "Na+, and ['H]choline' efflux from the vesicles was determined by Millipore filtration as described under "Materials and Methods." Rb' has been used rather than K' since there is no convenient radioisotope of K'. mM ATP. Efflux of the radioactive compounds was monitored at various times at 15°C by Millipore fitration. Fig. 6 shows that the apparent volumes occupied by "Rb' and "Na' were about one-third that of [3H]choline+. This was consistent with two-thirds of the vesicles (type I) being highly permeable to Rb' (or K') and Na', but not to ['HH]choline+, and one-third (type 11) being relatively impermeable to Rb', Na', and choline+. In the presence of ATP, similar =Rb+ and "Na' as well as ['HH]choline' efflux rates were observed (Fig.  6). Addition of ATP resulted in the uptake of about 55 nmol of Ca2+/mg of protein within 30 s (Table IV). It appeared unlikely that minimal stimulation of *6Rb' or 22Na+ efflux by ATP in K+,Na'-impermeable vesicles was due to low ca2+ uptake. Fluorescence measurements suggested that K',Na'permeable and -impermeable vesicles can accumulate Ca" with a similar rate (Fig. 1). Further, the amount of ca2+ accumulated within 30 s did not seem to be greatly affected by the impermeability of type I1 vesicles to Rb', since similar amounts of Ca2+ were taken up at this time by vesicles in the presence or absence of valinomycin (not shown). It was therefore concluded that about one-third of the Ca2' (18 nmol/mg of protein) was taken up by type 1 1 vesicles. Accordingly, Rb' and Na' efflux from K+,Na'-impermeable vesicles was increased by less khan 0.1 Rb' or Na'/Ca2' taken up (Table   IV). Moreover, as indicated by the enhanced [3H]choline' efflux rates during Ca" transport, these small increases in Rb' and Na' fluxes were likely due to nonspecific membrane permeability increases during Ca' uptake. In agreement with such an interpretation were the following three observations. First, increase in Rb+ and Na+ concentrations to 200 mM did not affect appreciably isotope efflux rates in the presence or absence of Ca" transport. At 200 mM, amounts of Rb+ or Na' were released which approached those seen for ["H]choline+ at a concentration of 190 mM (Table IV). Second, replacement of the permeable chloride anion (McKinley and Meissner, 1978) by the impermeable gluconate anion resulted in a greater stimulation of Rb' and Na' release from actively transporting vesicles (Table IV). Third, addition of 5 mM oxalate which increased Ca'+ uptake 2-to 3-fold at 15"C, had little effect on isotope release rates (not shown).
In summary, isotope efflux and fluorescence experiments suggested that Rb' and Na+ efflux from K+, Na'-impermeable and -permeable vesicles was differently affected by the active uptake of Ca". An initial Rb' or Na' release to Ca2' transport ratio of less than 0.1 was found in K',Na'-impermeable (type 11) vesicles. In contrast, a transport ratio near 2 was estimated TABLE IV CaZ+ transport and "Rb' and '=Nu+ efflux from K',Na'impermeable vesicles SR vesicles were preincubated in the media (all in miuimolar) in the presence of the indicated rahoisotope (cf. "Materials and Methods"). Vesicles were diluted 200-fold at 15°C into unlabeled medium of the same composition but containing 1 mM Mg" and 50 pM Can+ (as the c1or gluconate-salts), no or 0.5 mM Tris/ATP, and 10 mM Pipes, 20 mM Tris, pH 7. Apparent isotope spaces and amounts of Ca2' accumulated by type I + I1 vesicles 30 s after addition of vesicles to dilution media were determined as described under "Materials and Methods." AM',,,,, was calculated from the apparent isotope spaces, in the presence and absence of ATP, and represents the amount of radioactively labeled cation that was released by type I1 (""Rb, "'Na) or type I + I1 ([:1H]choline) vesicles in the presence of Ca2+ transport minus the amount when no Ca2+ uptake occurred. ARb',, (or ANa+,,t)/Ca2+,, was calculated assuming that K',Na+impermeable vesicles accumulated one-third of total Ca2+ uptake. AM+"", and Ca"," are given on a molar basis. Data are the average of three determinations. S.E. = +15%. for type I vesicles as well as type I1 vesicles, if the latter were made permeable to K' with valinomycin.

DISCUSSION
Membrane potential and ion flux experiments reported here suggest that ATP-mediated Ca2' uptake by sarcoplasmic reticulum may be compensated in charge by the counter movement of the permeable cations K+, Na', and/or H' . Compensating movement of these biologically relevant cations during Ca2+ transport would seem to prevent the development of a membrane potential, positive inside, which could impede further Ca2+ uptake. Transport of K' and Na+ appeared to be mediated by the K,Na channel of SR rather than by an obligate exchange process directly coupled with Ca" translocation.
In isolated SR vesicles, either intrinsically K'-permeable (type I) or rendered permeable to K' by valinomycin (type II), a sufficient amount of K' was released outward to neutralize the charge transported by Ca2+ into the vesicles. Na' or monoethanolamine' could be substituted for K', supporting the notion of a complementary movement of ions not directly catalyzed by the Ca2'-ATPase. Furthermore, during Ca" transport, Rb+ or Na+ efflux from K+,Na'-impermeable vesicles was only minimally stimulated when these contained the permeable ion C1F. Thus, it was unlikely that the observed ATP-mediated 2 K+-Ca2+ exchange reaction occurred via the Ca2+-ATPase.
The lack of SR vesicles impermeable to H' prevented a direct test of a putative 2 H'-Ca" coupled transport system. Indirect studies revealed an increase in the initial rate of Ca" uptake when a proton gradient and a negative membrane potential were induced by H' or by dilution of vesicles from high to low K' media. Since decreased K' lowered Ca2+ uptake in the absence of a membrane potential, it was conceivable that enhanced Ca2+ uptake may have been due to the negative potential rather than to the ions present in the media. The ability of vesicles to increase Ca2+ uptake in response to a membrane potential was compatible with the presence of a fully electrogenic Ca2+ transport system or an electrogenic H+-Ca2'-ATPase. Because of the simultaneous formation of pH gradients the present experiments did not rule out, however, a directly coupled electroneutral2 H+-Car+ exchange reaction.
The existence of an electrogenic Ca2' transport system in sarcoplasmic reticulum was suggested in earlier studies with native SR vesicles (Dupont, 1979;Akerman and Wolff, 1979) or a purified Ca"-ATPase incorporated in phospholipid bilayer vesicles (Zimniak and Racker, 1978). The rate of Ca2+ uptake was found to be affected by a membrane potential. Use of the dyes 8-anilino-1-napthalenesulfonic acid (Zimniak and Racker, 1978) or diS-C:3-(5) (Dupont, 1979) suggested the formation of a membrane potential, positive inside, of 50 to 60 mV during Ca2' transport. Because of the ineffectiveness of the dye di0-Cs-(3) as an indicator for positive membrane potentials under our experimental conditions (Meissner and Young, 1980) we did not measure positive potentials in the present investigation. A possible involvement of H' or the existence of an obligatory Ke-Ca2+ exchange reaction was not ruled out in the earlier studies. In this connection it may also be noted that interpretation of the earlier experiments with native vesicles (Dupont, 1979;Akerman and Wolff, 1979) is complicated by the failure to consider the presence of vesicles that were permeable to H' and differed in their K',Na+ permeability. Permeability properties of the phospholipid bilayer vesicles used by Zimniak and Racker (1978) were not directly studied. Our studies do not support the suggestion of forms a H' gradient during Ca'+ uptake and that a transmembrane proton gradient may provide the motive force.
The high permeation rates of H+, K' , and Na+ allow for their rapid redistribution during Ca2+ transport. Carvalho and Leo (1967) demonstrated that the active uptake of Ca2' by SR initiates the release of an equivalent amount of K+, H', and/or Mg2' from vesicles within 15 to 30 min (the time required to sediment the vesicles). It seems unlikely that Mg2+ is released in an obligatory manner during Ca2+ uptake as discussed by Kanazawa et al. (1971), Froehlich and Taylor (1976. Use of the fluorescence probe chlortetracycline (Nagasaki and Kasai, 1980) and membrane potential measurements with Mg/gluconate-filed vesicles2 have indicated that Mg2+, like Ca2' or Mn2+ (Meissner and McKinley, 1976), is a relatively impermeant cation. An obligatory Ca2+ for Mg2+ exchange would therefore appear to be incompatible with a K+ or H+ release to Ca2+ transport ratio of about two. Also, Ueno and Sekine (1978) found that Ca2' uptake could far exceed the internal Mg'+ concentration of the vesicles.
We envision an electrogenic Ca2+ transport system of SR as part of a membrane which is permeable to various biologically relevant ions, as summarized in Model A presented in Fig. 7. The cation permeability of sarcoplasmic reticulum in vivo is likely reflected by vesicles permeable to H', K' , and Na' rather than by H'-or K',Na'-impermeable vesicles. Studies with sonicated (McKinley and Meissner, 1978) and reconstituted (Young et al., 1981) vesicles have led us to suggest that SR contains a limited number of randomly distributed K,Na channels, approximately 50/pm2. Because of this low number of channels, fragmentation of the reticulum structure during homogenization yields two vesicle fractions, one of which lacks the K,Na channel. That no H+-and C1"impermeable vesicles (Meissner and McKinley, 1976;Meissner and Young, 1980) are observed would suggest a higher density of H' or C1channels. As indicated by the present study, both H+ and K,Na channels are capable of promoting sufficiently rapid countermovement of H' and K' to maintain these ions at electrochemical equilibrium during Ca2+ transport. In this regard it is of interest that a voltage-gated, K+-conducting channel has been found to display single channel conductance fluctuations between an "open" and "closed" state when SR vesicles were fused with planar bilayers (Miller, 1978). Whether the translocation of 2 Ca2+ at the expense of 1 ATP (Hasselbach and Makinose, 1963;Weber et al., 1966) by an electrogenic Ca2+-ATPase occurs independently of other ions (Model A, Fig. 7) or is coupled to the outward movement of n H+ (n = 1 to 3) or inward movement of phosphate anions (formed during ATP hydrolysis) (Model B, Fig. 7), could not be resolved in the present study since SR is permeable to H' and phosphate anions. Ca2' transport appears to be not directly coupled to C1-since Ca2+ uptake was observed in essentially chloride-free media.
We propose that a physiological function of H' and K,Na channels is to act as a complementary part of the Ca2+ transport system by allowing rapid H+ and K+ movement to counter Ca2+ fluxes during muscle relaxation. Ca2+ uptake could generate a momentary positive membrane potential causing the efflux of H' and/or K' , the exact number of which would be determined by the proton-buffering capacity and permeant ion concentrations. A consequence of a monovalent cation and proton permeant membrane, together with the similar ionic composition of SR and myofibrillar spaces (Somlyo et al., 1977), is that SR may not undergo significant potential changes during muscle relaxation, even though it appears to contain an electrogenic Ca2+ transport system.