The Electric Charge Stoichiometry of Respiration-dependent Ca2+ Uptake by Mitochondria*

The number of positive electric charges transported with each calcium ion into energized rat liver mito- chondria was determined with the aid of continuous electrode measurements of oxygen consumption, H’ ejection, Ca2+ uptake, K’ movements, as well as color- imetric measurements of phosphate transport. It was found that in the presence of N-ethylmaleimide Ca2+ is transported into respiring rat liver mitochondria in such a way that 2 Ca2+ are accumulated and 4 H’ are ejected per pair of electrons passing through each en-ergy-conserving site of the respiratory chain. Since no detectable inward movement of phosphate occurred during Ca2+ uptake under these circumstances, exit of 2 H’ provided complete charge compensation for in- ward movement of 1 calcium ion, which thus carries two positive electric charges. No N-ethylmaleimide-in- sensitive calcium-phosphate symport into the mito- chondria was observed. taken the of is limited

was driven by an outward directed K+ diffusion potential in the presence of N-ethylmaleimide, again no influx of phosphate was found to take place; uptake of each Ca2' ion was fully charge compensated by the efflux of 2 K'. Thus, each calcium ion carries two positive charges when it enters mitochondria energized by a K+ gradient.
N-ethyhnaleimide-insensitive symport of the type CaZ4' -HP042-, as postulated by Moyle and Mitchell ((1977) FEBS L&t. 77, 136-140), in which only one positive charge is carried per calcium ion transported, could not be detected under any circumstances tested.
It is now widely agreed that Ca" ions are transported into respiring mitochondria in an average stoichiometric ratio approaching 2 Ca2+ per pair of electrons passing through each energy-conserving site of the respiratory chain (l-9). The available evidence has also strongly supported the conclusion that calcium enters mitochondria electrophoretically in a uniport process in which each calcium ion carries two positive electric charges, as shown in Fig. lA (7,8,(10)(11)(12)(13)(14)(15)(16)(17)(18). Entry of * This study was supported by grants from the National Institutes of Health (GM05919) and the National Science Foundation (PCM 75-21923) and a contract from the National Cancer Institute (NOl-CP-45610). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. phosphate has been generally assumed to occur via the well known phosphate-hydroxide antiporter. However, reports have recently appeared with the interpretation that inward calcium transport by respiring mitochondria has an absolute requirement for phosphate or some other potential proton-carrying anion, such as / 3-hydroxybutyrate (9, 19-22). From such observations Moyle and Mitchell have proposed (19) that calcium is carried into respiring mitochondria in an obligatory (Ca&4'-HP042-or Cazc-monocarboxylate-symport process, promoted by a specific carrier, in which only one positive electric charge is carried per calcium ion transported inward (Fig. 1B). The evidence for an electrophoretic Ca2+ uniporter consists of 1) osmotic swelling experiments in nonrespiring mitochondria in which any one of several permeant charge-delocalized anions (SCN-and NOs-, etc) may enter nonrespiring mitochondria with Ca2+ (11-13); 2) stoichiometric charge compensation of the entry of Ca2+ by efflux of 2 H+ generated by electron transport (7,8,15); and 3) stoichiometric charge compensation of entry of Ca2+ by efflux of 2 K+ (10, 17, 23, 24). It is well known that any one of a number of different anions may accompany uptake of large amounts of Ca2+, but it is required that the anion can potentially carry H' into respiring mitochondria. Such anions, which include phosphate, acetate and other anions of weak lipophilic acids, and HC03-in the form of dissolved CO2 (12, 13, 25), pass into respiring mitochondria either by a separate and specific carrier, such as the phosphate--hydroxide-antiporter (26-29), or by simple unmediated diffusion, as in the case of lipophilic weak acids or C02, or via an ionophore, such as tributyltin (30) 1. Pathways of Ca'+ transport into respiration-energized mitochondria. A shows electrophoretic uniport of Ca2+ into mitochondria in response to the negative inside membrane potential generated by respiration; each calcium ion carries two positive charges. Phosphate is carried inward in response to the alkaline inside gradient, in exchange for hydroxide, on the well known phosphate-hydroxide exchanger. This model is stoichiometrically consistent with much evidence that at least 3 and probably 4 H' are ejected per pair of electrons per energy-conserving site. B shows the alternative calcium transport model of Moyle and Mitchell (19)) in which calcium enters in obligatory symport with phosphate on a specific carrier, in such a way that each calcium ion carries one positive charge. For inward transport of 2 Ca*' per pair of electrons per site, a H+ ejection stoichiometry of 2 per site is required. chiometric approaches in several laboratories that at least 3 and probably as many as 4 H' are ejected (or charges separated) per pair of electrons per site (6-8, 15, 32-48).
In this communication we describe experiments in which the charge stoichiometry of calcium transport by rat liver mitochondria is re-examined with simultaneous electrode recordings of oxygen consumption, Ca2' uptake, H+ ejection, and K' movements, moreover, measurements of phosphate movements between the mitochondria and the medium were carried out simultaneously in order to assess the possible participation of a (CaJ'+-HP04'-symport process. The data presented unequivocally confii that the calcium ion is transported into respiring mitochondria by an electrophoretic process in which each Ca2+ ion carries two positive charges and demonstrate further that no obligatory inward phosphate movement is required. Moreover, it is shown that observations which have been taken to support an obligatory (Ca#'-HP04'-symport are in fact readily explained by electrophoretie uniport of Ca2+ and the independent operation of the well documented N-ethylmaleimide-sensitive phosphate-H* symporter. No evidence was found for occurrence of an Nethyhnaleimide-insensitive calcium-phosphate symport process.

EXPERIMENTAL PROCEDURES
Mitochondria were isolated from the livers of large male Charles River CD rats fasted overnight. The homogenate was prepared in unbuffered 250 nm sucrose (Mallinckrodt), the mitochondrial fraction was washed three times, and then suspended in cold 250 mM sucrose at 100 mg of protein/ml; the pH of this suspension was 6.9. Absence of N-Ethylmaleimide-insensitive Phosphate Uptake during Respiration-coupled Uptake of Car+-In preceding papers from this laboratory (7,8) it was shown that in the presence of N-ethylmaleimide or mersalyl, potent inhibitors of H+-H2P04-symport (cf Ref. 29), each Ca2+ enters respiring rat liver or heart mitochondria in electroneutral exchange with exactly 2 H+ under conditions in which 'no other permeant anion was available, indicating that each calcium ion must carry two positive charges into the mitochondria.
Nevertheless, it appeared necessary to establish by direct measurement whether phosphate of the medium accompanies the influx of Ca" in such experiments in order to test the question whether an N-ethylmaleimide-insensitive inward symport of (Ca2)4+-HP042-, such as that postulated by Moyle and Mitchell (9,19,20), takes place. Such an experiment is shown in Fig. 2. Rat liver mitochondria were allowed to become deenergized by preincubation in an aerobic system containing rotenone to prevent oxidation of NAD-linked substrates, oligomycin, Ca2+, and N-ethylmaleimide.
(The addition of acetazolamide to inhibit utilization of dissolved CO2 as a source of anion (22, 25) was found to be unnecessary when C02-free reagents were employed.) After 5-min preincubation, at which time a total of about 150 nmol of phosphate (equivalent to 43 pM) had appeared in the medium, derived by efflux from the de-energized mitochondria (40)(41)(42)(43)(44)(45)50), succinate was added to initiate oxygen uptake. Concomitant with the electron flow so induced, uptake of Ca2' and ejection of H' took place at nearly linear initial rates, as indicated by the traces from the Ca2+ and H' electrodes. 'The initial rate of H' ejection (218 ng-ions X mm-' X mg-') was almost exactly twice the initial rate of Ca2+ uptake (115 ng-ions x min-' x mg-*), as reported earlier (7,8). The observed H'/Ca" ratio of 218/115 = 1.96 showed that entry of two positive charges with each calcium ion was charge compensated by the exit of 2 Hf. Moreover, from the initial rates of Ca2+ uptake, oxygen consumption, and H+ ejection, the Ca"/site uptake ratio was determined to be 2.1 and the H+/site ejection ratio was 4.0, also in full confiiation of earlier results (7,8). However, the crucial point of the experiment in Fig. 2 is that under these conditions there was absolutely no uptake of phosphate from the medium, as determined by a series of direct measurements, even though ample phosphate was available to accompany influx of calcium if the hypothetical N-ethyhnaleimide-insensitive (Ca$+-HP04'-symport process proposed by Moyle and Mitchell (9,19,20) had taken place. This experiment clearly eliminates the latter type of process as making any significant contribution to respiration-dependent Ca2+ uptake. In conditions such as those shown in Fig. 2, Ca2+ uptake, H' ejection, and oxygen consumption are stoichiometrically related only in the early initial reaction period from which the rate data were obtained. In the absence of proton-carrying permeant anions the further uptake of Ca2+ can result in an elevated matrix pH and consequent respiratory inhibition and membrane binding of Ca'+, a condition designated as State 6 (51-53). Further uptake of Ca2+ can occur only if a source of protons, such as proton-carrying anions, can enter the mitochondria to bring the matrix pH back to the normal levels characteristic of State 4 (31). This conclusion, already drawn some years ago (5,13,(51)(52)(53), is supported by an additional experiment in Fig. 2, which shows that the addition of phydroxybutyrate, after maximum Ca2+ uptake and H+ extrusion had occurred in the presence of N-ethyhnaleimide, evokes further Ca" accumulation.
Moreover, the H+ trace in Fig. 2 shows that inward transport of HC occurred on addition of P-hydroxybutyrate since the medium pH rose very rapidly before further uptake of Ca2+ occurred. P-Hydroxybutyrate is known to enter the matrix either as the lipophilic free acid or on the H'-keto acid anion-symporter (54); because of the presence of rotenone, P-hydroxybutyrate is not oxidized by the mitochondria.
No uptake of phosphate occurred during this period of P-hydroxybutyrate-induced Ca2+ uptake. The stimulation of Ca*' uptake by /3-hydroxybutyrate under these conditions is not specific; similar stimulation at about the same rate is also given by anions of other weak lipophilic acids (cf Ref. 13).
From Fig. 2 it may, therefore, be concluded 1) that 2 calcium ions are taken up per pair of electrons passing each energyconserving site; 2) that 4 H' are ejected per energy-transducing site; 3) that exit of 2 H+ provides near exact electrical compensation for the entry of each calcium ion, which thus carries two positive electric charges inward; and 4) that phosphate does not enter the mitochondria with Ca2+ in an Nethylmaleimide-insensitive pathway when N-ethylmaleimide is present at a level just sufficient to block H2P04--H+ symport activity.
Ca2+ Uptake when Phosphate Is Available--When a proton-donating anion is readily available for inward transport during the entire course of Ca*' uptake in experiments such as that in Fig. 2, the results are quite different. The experiment in Fig. 3 demonstrates that in the absence of N-ethylmaleimide initiation of respiration with succinate is accompanied by uptake of Ca*+ and ejection of H', but in this case there is very rapid uptake of phosphate from the medium. Respiration-dependent Ca2' uptake was more extensive than in Fig.  2 since the simultaneous uptake of H+ and H2P04-on the phosphate-H+ symporter provided a supply of H+ to replace the matrix H+ ejected in exchange for Ca2+, thus delaying the onset of matrix alkalinization and the State 6 condition that would ultimately occur in the absence of phosphate. However, since only a limited amount of phosphate was available in the medium in the experiment of Fig. 3, the amount of Ca*' that could be taken up was necessarily limited. However, after phosphate-supported ?a*+ uptake had declined to a very low rate, the addition of the proton-carrying anion P-hydroxybutyrate immediately yielded further Ca*+ uptake. Additional uptake of Ca*+ under these precise conditions also takes place upon addition of phosphate, acetate, or other permeant proton-donating anions (experiments not shown; see Ref. 13). As in the experiment described in Fig. 2, the second period of Ca2+ uptake in Fig. 3 was preceded by alkalinization of the medium following addition of /?-hydroxybutyrate, again demonstrating that the latter carries protons into the matrix, Effect of N-ethyhnaleimide on Ca2+ uptake with ethylmaleimide.
The test system was the-same as in Fig. 2 with the ascorbate + TMPD as oxidizable substrates.
The conditions were the exception that N-ethylmaleimide was omitted from the medium.  In the experiment in Fig. 3, the absence of N-ethylmaleimide and the presence of phosphate resulted in a lowered ratio of H+ ejected/Ca'+ accumulated in the first period of Ca2+ uptake, to about 1.2, in the range (0.8 to 1.2) already observed in many earlier reports from a number of laboratories (1,(55)(56)(57)(58). Moreover, the H'/site ratio was also lowered, to about 2.4, as has been shown before (8). However, as has been discussed in extenso elsewhere (40)(41)(42), the lowering of the observed H+/Ca2' ratio is simply due to H+ influx via the phosphate--H+ symporter. For this reason the observed number of H+ ejected per pair of electrons per energy-transducing site when Ca2+ and phosphate are taken up together cannot be taken as a measure of the intrinsic H+/site ratio of electron transport nor of the charge stoichiometry of Ca2+ transport. The experiments in Figs. 2 and 3 thus show that the data on Ca2+ uptake and H+ ejection in the absence or presence of phosphate can be easily explained in terms of the already well documented individual Ca2+ and phosphate--H+ transport systems and do not require postulation of a hypothetical (Ca2)4+-HP042-symport carrier (9,19,20); indeed, these experiments show that no significant N-ethyhnaleimide-insensitive influx of phosphate with Ca2+ occurs (19).
Another experiment (Fig. 4) further emphasizes the independence of the Ca2+ and phosphate transport systems. Mitochondria were preincubated in the absence of N-ethylmaleimide. During this period some endogenous phosphate escaped into the medium (42, 50). On addition of Ca*+, uptake of Ca2+ and phosphate proceeded almost to the limit set by the available phosphate.
At this point ruthenium red, an inhibitor of Ca2+ influx (59, 60), was added to stop further Cazc accumulation.
Additional phosphate was then added to the medium. As expected, no further Ca2+ uptake ensued, despite the presence of added phosphate; nevertheless, over 100 nmol of phosphate were very rapidly taken up by the mitochondria.
Coinciding with the phosphate accumulation was the disappearance of approximately the same quantity of H' from the medium. Thus, phosphate uptake can occur in response to an alkaline inside pH gradient, independent of Ca2+ transport. This observation parallels that in Fig. 3, which showed that Ca2+ uptake can occur in response to a negative inside membrane potential, independent of phosphate transport.
Apparent Inhibition of Ca2' Transport N-Ethylmaleimide-The traces in Figs. 2 and 3 indicate that the initial rate of Ca2' inilux supported by succinate oxidation was decreased in the presence of N-ethyhnaleimide, as has been reported by others under similar conditions (21, 61, 62). However, this observation cannot be taken as evidence that Ca2+ uptake has an absolute requirement for phosphate or that Ca2+ uniport is inhibited by N-ethyhnaleimide since there is a very simple explanation for this effect. In the presence of N-ethyhnaleimide or mersalyl, electron flow from succinate to oxygen is depressed (cf Refs. 42 and 63); consequently, the rate of Ca2+ uptake coupled to succinate oxidation must necessarily decrease in proportion to the respiratory inhibition. That the Ca" carrier is not inhibited by N-ethyhnaleimide is shown by an experiment in which succinate was replaced as electron donor by the ascorbate + TMPD' system in the presence of rotenone + antimycin A (Fig. 5). In this case the initial rates of neither electron flow nor CazC uptake were inhibited by Nethylmaleimide amount required for inhibition of phosphate transport. Nethyhnaleimide did limit the net amount of Ca2+ accumulated to approximately 40 ng-ions/mg, the same limit observed with succinate as respiratory substrate, since phosphate influx is limiting under these conditions. The addition of N-ethyhnaleimide also affected the H'/ Ca2+ stoichiometry in the predicted manner. The traces in Fig. 5 show that H' ejection occurred at a rate of 64 ng-ions x mh-' x mg-' when the uptake of phosphate was allowed to proceed and at 168 ng-ions min-' mg-' when phosphate trahsport was blocked with N-ethylmaleimide.
The observed H+/ Ca2+ rate ratio was thus 0.64 in the absence and 1.66 in the presence of N-ethylmaleimide.
The latter value agrees closely with the predicted value of H+/Ca2+ = 1.67 for the operation of an electrogenic Ca2+ uniport when ascorbate + TMPD is the respiratory substrate (see Ref. 46). Particularly significant was the observation in Fig. 5 that phosphate was not taken up with Ca2+ when N-ethyhnaleimide was present. This type of experiment excludes an alternative explanation of data such as those in Figs. 2 to 4 that would theoretically be possible ifan obligatory (Ca2)4+-HP042symport were a major pathway of inward Ca2' transport. With succinate as respiratory substrate the possibility could exist that phosphate might enter the mitochondria via the hypothetical N-ethylmaleimide-insensitive (Ca#+-HP04'-symport pathway as proposed by Moyle and Mitchell (19) and then exit via succinate-phosphate exchange on the dicarboxylate carrier (65), thus constituting a phosphate cycle. Such cycling of phosphate, without a net accumulation, could conceivably account for the observation (Fig. 2) that there is no net phosphate uptake as Ca2+ is accumulated. However, since ascorbate or TMPD (reduced form) or both together fail to release phosphate from rat liver mitochondria,2 such hypothetical phosphate cycling can be excluded.

Explanation of Reported Experiments
Showing Failure of Ca2+ Uptake in the Absence of Phosphate Transport-There is another type of experiment that has been cited (21, 22) as evidence that Ca2+ entry cannot take place in the absence of the transport of phosphate or some other proton-donating anion. In one such experiment in which Ca2' movements were followed spectrophotometrically with Arsenazo III (21), Ca2+ was added to respiring rat liver mitochondria supplemented with succinate as substrate in the presence of rotenone. In the absence of an inhibitor of phosphate transport the Ca2+ was rapidly accumulated, but when the inhibitor was present the Ca2+ remained in the medium. This experiment was taken to indicate that Ca2+ uptake cannot take place without uptake of phosphate, contrary to the conclusions drawn from the experiments described in Fig. 2. However, the conditions employed in that experiment do not warrant this conclusion since the participation of endogenous Ca2+ in the total Ca*' movements in such experiments was not taken into consideration. Endogenous Ca2+ (and phosphate) are known (66) to leak rapidly from mitochondria during de-energizing conditions, such as the anaerobiosis, that usually exist in concentrated unstirred suspensions of isolated rat liver mitochondria.
When an aliquot of such a mitochondrial suspension is added to a medium containing a respiratory substrate, rapid uptake of the previously released Ca2+ and phosphate would be expected to occur as respiration is instituted. It would also be expected from the results shown in Fig. 2 that the presence of N-ethyhnaleimide would not interfere with such uptake of Ca2+ so long as the level of endogenous Ca2+ is less than approximately 40 ng-ions/mg of protein, the apparent maximum capacity of rat liver mitochondria for Ca2+ uptake in the absence of a permeant protein-carrying anion in the medium.
However, the uptake of Ca"+ beyond a total of about 40 ngions/mg of protein will be inhibited by N-ethylmaleimide since no further uptake of Ca'+ can occur after the limiting alkalinization of the matrix produced by the first period of Ca2+ accumulation is reached unless phosphate (or some other proton-carrying anion) can enter with the Ca". Experiments similar to those reported in ref. 21 are shown in Fig. 6. In these experiments rat liver mitochondria were added to a medium containing rotenone, Mg'+, oligomycin, and phosphate. Trace I of Fig. 6 shows that within a few minutes in the absence of electron flow, a total of 13 ng-ions of endogenous Ca2'/mg of protein was released from the mitochondria.
Subsequent addition of succinate caused rapid uptake of the released Ca2', as well as an additional 5 ng-ions of Ca'+/mg that was initially present in the medium. An addition of 37 ng-ions of Ca2+/mg of protein was then made, which was immediately and completely accumulated by the phosphate-supplemented mitochondria, as expected. The total Ca2+ taken up by the mitochondria was thus 18 + 37 = 55 ng-ions of Ca2+/mg of protein.
Trace II of Fig. 6 shows the effect of N-ethyhnaleimide on Ca2+ uptake under the same conditions.
The addition of mitochondria to the medium again was accompanied by release of endogenous Ca2'. On addition of succinate the available 18 ng-ions of Ca2+/mg was immediately taken up, despite the presence of N-ethylmaleimide, since the capacity of rat liver mitochondria to accumulate Ca2' in exchange with ejected H+ was not exceeded. However, when a subsequent addition of 37 ng-ions of Ca2'/mg was then made to the Nethylmaleimide-treated mitochondria, only 12 of the 37 ngions of Ca2+ were taken un. as can be expected, since this addition of Ca2', together with the reuptake of the endogenous Ca2+, represents the upper limit of Ca2+ uptake when Ca2+ entry is compensated solely by respiration-coupled H' extrusion. That these mitochondria are still capable of Ca2+ uptake, providing a proton-donating anion is available, is shown by Trace III. In this experiment, identical with that in Trace II, but with 10 mM acetate added, the second addition of Ca2+ was taken up completely in the presence of N-ethylmaleimide, which does not inhibit the entry of undissociated free acetic acid into the matrix where it can deliver protons to neutralize the alkalinity generated by proton extrusion coupled to electron transport.
Charge Stoichiometry of Ca'+ Uptake Driven by a K+ Gradient-One of the most important experiments for determination of the number of positive electric charges transferred per calcium ion transported is to couple Ca2' uptake to K+ efflux down an outward directed K* gradient in respirationinhibited mitochondria, as originally reported by Rossi et al.  added to allow the release of mitochondrial K' from the rotenone-treated mitochondria down its gradient into the low K' medium. Since valinomycin-mediated K' transport is electrogenic, with one positive charge transferred per potassium ion (67), an outward directed K' diffusion potential is established across the mitochondrial membrane, which can drive the inward transport of other cations. Azzone and co-workers (10, 23,24), and more recently Akerman (17), have reported that valinomycin-induced effIux of mitochondrial K+ down a gradient of K' drives the uptake of Ca2+ in the stoichiometric ratio of 2 K' per Ca2', thus indicating that two positive electric charges are transferred per calcium ion accumulated. However, simultaneous measurements of phosphate movements during such experiments were not reported, leaving open the possibility that the postulated N-ethylmaleimide-insensitive (Ca2)4+-HP042-symport may have taken place. In the experiments in Fig. 7 Ca2' uptake in respiration-inhibited mitochondria was coupled to K+ effIux; measurements were made of K' eMux, Ca2+ uptake, H+ movements, and movements of inorganic phosphate. Fig. 7 shows that when valinomycin was added to mitochondria suspended in a medium low in K+ and containing rotenone, antimycin, oligomycin, and N-ethyhnaleimide, the ratio of the net K' efflux and net Ca" uptake 30 s after addition of valinomycin was K+/Ca2' = 2.08, very close to the expected value 2.0, in confirmation of the earlier reports (10,17,23,24). Simultaneous measurement with the pH electrode showed that only a very small uptake of H+ occurred. Most important, during this period of K+ eftlux and Ca2+ influx no phosphate was taken up by the mitochondria, as is shown by the direct measurements recorded in Fig. 7. Therefore, as in the case of respiration-dependent Ca2' uptake in exchange for H' (Fig. 2), Ca2+ uptake driven by an outward directed K' gradient also proceeds without concurrent uptake of phosphate when N-ethylmaleimide is present. In both instances close to two positive charges were extruded for every calcium ion accumulated, but in neither case did inorganic phosphate enter the mitochondria, thus excluding the participation of the hypothetical N-ethylmaleimide-insensitive (Ca2)4+-HP042-symport process. In a comparable experiment carried out in the absence of N-ethyhnaleimide, the observed K+ efflux/Ca*' influx ratio was again close to the expected value of 2.0. Again, no uptake of phosphate was observed under these conditions, as expected, since a pH gradient was not generated. These observations not only exclude a (Ca#+-HP04*-symport process (19) but also provide further evidence that the well documented transport systems for Ca*' uptake and for phosphate-H+ symport function independently of each other under a variety of experimental conditions and can account for all types of observations on Ca'+ uptake that have been reported.

DISCUSSION
The data collected in this paper, taken together with evidence reported from other laboratories (7, 8, 10-18, 23, 24) demonstrate conclusively that calcium is transported into respiring mitochondria by an electrophoretic uniport process in which each calcium ion carries two positive electric charges. The alternative proposal of Moyle and Mitchell (9,19,20) that calcium is transported by an obligatory (Ca#+-HP04'or Ca*+-monocarboxylate-symport process with inward transport of only one positive charge per calcium ion has been excluded by direct measurement of the transmembrane movements of phosphate accompanying uptake of Ca*+ in the absence of other permeant anions. No inward movement of phosphate occurred as each calcium ion was transported inward in exchange for either 2 H', when electron transport provided the driving force, or for 2 K+, when an outward directed K+ diffusion potential was the driving force in valinomycin-treated respiration-inhibited mitochondria. In these experiments the concentration of phosphate in the medium was determined at frequent intervals over the entire course of Ca*+ uptake, which lasted less than 1 min. Although Moyle and Mitchell (19) reported experiments in which extramitochondrial phosphate was determined, only two measurements were reported, the fist made at an unspecified time prior to Ca*+ uptake and the second 5 min after Ca2+ uptake was initiated. It appears probable that during this period, very long in relationship to the actual time required for the Ca2+ uptake, the phosphate permeability of the mitochondria could have increased, perhaps due to swelling, so that some Nethylmaleimide-insensitive phosphate uptake occurred as a secondary consequence after the limiting amount of Ca'+ uptake had already taken place. Since we have consistently observed a total lack of phosphate uptake within the exact time period in which Ca*' is accumulated, as in Fig. 2, we conclude that there is no significant phosphate influx obligatorily coupled to Ca*+ uptake in a symport process insensitive to N-ethylmaleimide, such as that proposed by Moyle and Mitchell (19). When no proton-donating per-meant anion is available, the extent of Ca*' uptake by respiring mitochondria is limited by the increase of matrix pH that accompanies the uptake of Ca'+ in the exchange for matrix H+. The limit of Ca2+ uptake is attained when matrix pH rises sufficiently to inhibit electron transport and/or to increase membrane permeability.
Factors that tend to prevent a large rise in matrix pH will permit respiration and Ca2+ uptake to continue; those that oppose this action will tend to limit the extent of Ca*+ accumulation. The rise in matrix pH that accompanies Ca*+ uptake and ultimately limits it can be prevented by 1) the influx of H+ coupled to influx of phosphate on the H+-H2P04-symporter; 2) passage of undissociated lipophilic free acid forms of such anions as acetate and /?-hydroxybutyrate (13,36); or by 3) the exchange of H' or hydroxide ions for other ions as catalyzed by exogenous ionophores, such as tributyltin (30). When such pathways of H' access into the matrix are unavailable, the extent of Ca2+ uptake is necessarily limited by the buffering power of the matrix solutes, furnished largely by matrix phosphate, adenine nucleotides, and also matrix and/or inner membrane proteins.
Direct evidence that the increased extent of Ca2+ uptake by respiring mitochondria produced by phosphate is actually due to the Hf carried inward by symport with the phosphate is given by the experiments shown in Figs. 2,3, and 6. When the extent of Ca*' uptake was limited due to inhibition of phosphate transport (Fig. 2), further Ca*' uptake could still be evoked by the addition of /3-hydroxybutyrate, which is transported into mitochondria in symport with protons (or in exchange for hydroxide ions) (54). Addition of P-hydroxybutyrate resulted in an immediate rise in medium pH due to the loss of H' carried into the alkaline matrix with P-hydroxybutyrate. Only after substantial influx of H+ was there onset of further Ca*' accumulation.
Such a rapid uptake of H+ by Ca*'-loaded mitochondria was also observed (Fig. 3) after addition of acetate or phosphate (in the absence of N-ethylmaleimide).
In the absence of H' furnished by proton-donating anions in the medium a maximum of about 40 ng-ions of Ca" were observed to be accumulated per mg of rat liver mitochondrial protein under the conditions used (see Figs. 2,5,and 6). Although this value is in the range reported by other investigators (36, 58), several groups have suggested that the maximum capacity for accumulation of Ca*+ is rather less (22, 61, 62); at least one report (21) has indicated a total lack of Ca*' uptake under these conditions. However, as is shown here, these discrepancies can be attributed to differences in experimental conditions that affect the point at which the elevation of matrix pH limits Ca*+ accumulation.
Thus, as shown in Fig.  6, the reuptake of endogenous Ca*+ that had earlier leaked from the mitochondria during the de-energizing preincubation can account for the apparent inability of respiring mitochondria under some conditions to take up added Ca"' in the absence of added phosphate. This factor was revealed by a simple change in experimental design, i.e. by initiating Ca"' influx by addition of the respiratory substrate rather than by addition of CaZf (Fig. 6).
Another reason for some variability of values reported for the maximum extent of Ca"' accumulation in the absence of H'-carrying anions is that mitochondria prepared by different procedures or incubated in media of differing composition may differ in their endogenous buffering capacities due to differing levels of intramitochondrial phosphate, adenine nucleotides, and pyridine nucleotides or possibly differences in intramitochondrial Mg*+. Preliminary experiments indicate that the inorganic phosphate content of freshly isolated mitochondria can vary considerably depending on the pH and composition of the isolation medium. Although it is now certain from evidence provided in this and other reports that Ca*+ enters energized mitochondria in an electrophoretic uniport process in which each calcium ion carries two positive charges, there is much evidence that the efflux of Ca2+ from energized liver mitochondria takes place via a different transport process (16, 6%71), differently sensitive to inhibitors, possibly in exchange with H+. In mitochondria from heart and other excitable tissues Ca*+ leaves in exchange for Na+ (72,73). In an earlier communication from this laboratory (74) it was shown that the Ca*' influx and Ca"+ efflux systems of rat liver and heart mitochondria are independently regulated, the efflux system being most active when the mitochondrial pyridine nucleotides are in the oxidized steady state and least active when they are in the reduced state. For this reason the steady state distribution of Ca2+ between the medium and mitochondrial matrix may be deter-mined by the relative kinetics of Ca*+ influx and efflux, as determined by specific regulatory effecters, and thus may not always be in thermodynamic equilibrium with the membrane potential. These considerations provide an explanation for the reported variations in the number of electric charges carried by calcium ions as calculated from the Nernst equation, on the assumption that calcium distribution across the membrane is in thermodynamic equilibrium with the membrane potential.
Disequilibrium evidently prevails in experiments yielding less than two positive charges per Ca2+ (cf Ref. 36).
Finally, it may be pointed out here that the postulation of a Ca'+-anion-carrier, with transport of one electric charge per calcium atom, was in part an ad hoc necessity to accommodate at one and the same time the fact that 2 calcium ions are accumulated per 2e-passing through each energy conserving site of the respiratory chain together with the no longer tenable conclusion of Mitchell (31) that only 2 H+ ions are ejected per 2e-per site during electron transport.