On the Sidedness of Membrane Phosphorylation by Pi and ATP Synthesis during Reversal of the Ca’+ Pump of Sarcoplasmic Reticulum Vesicles*

The membrane sidedness of Pi interaction in reactions which characterize reversal of the Caz+ pump of sarcoplasmic reticulum vesicles isolated from rabbit skeletal mu’scle was investigated. Vesicles previously loaded with calcium [;*P]phosphate were incubated with 0.1 mM ADP and different concentrations of nonradioactive PI. Alternatively, vesicles loaded with nonradioactive calcium phos- phate were incubated in a medium containing ‘*Pi. The rates of Ca*+ efflux and ATP synthesis were significantly activated only when P, was included in the assay medium. P,

Fragmented sarcoplasmic reticulum vesicles SRV' isolated from skeletal muscle actively take up calcium from the medium at the expense of ATP hydrolysis. Recently it has been shown that the calcium pump can be reversed. Evidence has been presented that the very same ATPase involved in calcium transport is able to promote the synthesis of ATP using the transmembrane calcium gradient as a source of energy (l-6). When vesicles previously loaded with CaZ+ are incubated in a medium containing EGTA, a calcium chelating agent, a steep concentration gradient is formed across the vesicle membrane, but calcium leaves the vesicles at a very low rate due to the low permeability of the membrane to Ca*+ (1, 3). If ADP and P, are added to the assay medium, the rate of calcium efflux sharply increases (3,7). Coupled with the increment of the efflux, the membrane is phosphorylated by P, (6,(8)(9)(10)(11)(12)(13), and ATP is synthesized (5,12,13 EGTA, ethylene glycol bis(&aminoethyl ether)-NJ'-tetraacetic acid.
'*Pi, and calcium, calcium phosphate is accumulated by the vesicles (14) and a calcium concentration gradient is built up until a steady state is reached in which a slow Ca*+ efflux is balanced by an ATP-driven influx. When this condition is reached a steady rate of exchange between '*P, and ATP is observed (4, [15][16][17]. In previous papers (9,(14)(15)(16) evidence has been presented that this exchange is the result of the two reactions shown above, operating simultaneously forward (ATP hydrolysis) and backward (ATP synthesis from ADP and "'P,).
The aim of this paper was to investigate whether 32P, reacts at the inner or outer surface of the vesicle membrane for the formation of E-P, and whether newly formed ATP is released inside the vesicles or outside them.  (14). Once this precipitate is formed, the free calcium concentration inside the vesicles should remain constant as the ionic species in equilibrium with the calcium phosphate crystals flow out of the vesicles. For each experiment, the SRV were simultaneously loaded under two conditions: one with 'sCaCI and nonradioactive P,, and the other with nonradioactive CaCl, and "P,. The ratio of Ca:P, accumulated by the vesicles in different SRV preparations tested varied between 1.4 and 2.0. In some experiments (Fig. 6), the vesicles were loaded with calcium oxalate. For these experiments, 5 mM ammonium oxalate was used instead of 10 mM P, in the loading medium.
Oxalate has been shown to be more effective than P, in increasing the calcium uptake by sarcoplasmic reticulum vesicles; a smaller concentration of oxalate is required in the assay medium to increase maximally the rate of calcium uptake and the calcium storage capacity of the vesicles (1, 21-23). This is probably due to a lower solubility of calcium oxalate. Thus, oxalate anions would be more efficient than P, in decreasing the Calf concentration inside the vesicles and in attenuating the inhibition promoted by high Ca *+ inside the vesicles (15,16,(24)(25)(26) In order to measure the '"Ca or '*P, which leaks out of the vesicles during the manipulation following the loading procedure, an aliquot of the efflux mixture was filtered 10 s after addition of the vesicles. This was taken as zero incubation time and the amount of '%a or "P, found in this initial determination was subtracted from the subsequent determinations. The concentrations of '"Ca and szP, found in the first filtrate varied between 25 NM and 35 PM. If the vesicles were not previously washed with ice-cold Mg ATP as described above, the concentrations of %a and '*P, found in this first filtrate were much higher (100 to 250 PM). ATP Synthesis-This was assayed on aliquots of the Ca*+ or P, efflux media described above, either by measuring the formation of [y-SZP]ATP from szP,, the excess of '*P, being extracted from the assay medium as phosphomolybdate with isobutyl alcohol-benzene (15) Phosphorylation of SRV by I',-For measuring E-P formation, the reaction was started by the addition of loaded vesicles to the assay media described above for synthesis or exchange. For 1 ml of the assay medium, 2 ml of ice-cold perchloric acid (125 mM) containing 2 mM P, were used to stop the reaction.
The resulting suspension was centrifuged in the cold at 5000 x g for 8 min. The protein pellet was washed five times with 4-ml aliquots of the ice-cold perchloric acid solution.
After the washings, the pellet was resuspended in 0.3 ml of a solution containing 0.1 N NaOH, 2% Na2COJ, and 1 mM orthophosphate, and dissolved by heating in boiling water for 30 min. An aliquot of this was counted in a liquid scintillation counter, and another was used for protein determination by the method of Lowry et al. (27). In order to measure the unspecific binding of ,*P, (Table  I, B and C)  controls were performed in which the vesicles were previously denaturated by adding the perchloric acid solution before addition to 1 ml of the assay medium.
The values found varied between 0.35 and 0.55 pmol of "'P,/g of protein.
The data of Table I are not corrected  for this  unspecific binding of a*P,.

RESULTS
ATP Synthesis-When vesicles loaded with calcium phosphate were incubated in a medium containing EGTA but no ADP or P,, a steady efflux of Caa5 and szP, were measured (Fig.  1, A and B). This represents the passive efflux, and its rate was determined by the concentrations of these two ionic species inside the vesicles. The addition of ADP to the medium promoted a slight increment in the efflux of both ?a and "P,, which was more pronounced the longer the incubation time, this efflux was accompanied by the synthesis of a small amount of ATP (Fig. 1C). All the ATP synthesized was labeled with radioactive phosphate. If nonradioactive P, was included in the medium along with the ADP, both '%a and "P, effluxes sharply increased. This increment of efflux was coupled to the synthesis of a larger amount of ATP, which could be measured by the enzymatic method. However, only a very small fraction of the ATP synthesized was labeled with radioactive phosphate (Fig. 1C). This indicates that only the phosphate which reacted with the membrane at the outer surface of the vesicles was involved in the reversal of the calcium pump. The synthesis of ATP observed when ADP alone was added was probably promoted by the 32P, which left the vesicles by passive efflux and could then react with the outer surface of the membrane. This conclusion was substantiated by the experiments described in Figs. 2 to 4.
The data of Fig. 2A show that the rate of release of calcium and Pr was progressively enhanced when the PI concentration of the medium was raised. In the absence of added PI, the [T-~*P]ATP synthesized was equal to the ATP measured enzymatically (Fig. 2B). Whereas the total ATP formed increased in parallel with the effluxes, the [-y-32P]ATP formed decreased as the 32Pi leaving the vesicles was diluted by increasing amounts of nonradioactive P, in the medium.
The apparent K, for P, (Fig. 3)  The passive efflux obtained in the nonradioactive phosphate and incubated in media containing increasing concentrations of 9,. The [y-szP]ATP formed was equal to the amount of ATP measured by the enzymatic method.
When the SRV were loaded, the ratio of '5Ca:82P1 accumulated by the vesicles varied in different preparations between 1.4 and 2.0. Since the same ratio as that of Ca:P, accumulated was maintained in a given vesicle preparation for the initial rates of '%a and szP, effluxes, the effluxes appear to be controlled by the internal concentrations of these two species. The increment of s*P, efflux promoted by the addition of ADP and PI to the medium was probably not an active process like that of Ca ZI+ , but rather a passive efflux. The rapid solubilization of the calcium phosphate crystals inside the vesicles which was promoted by the rapid draining of Cal+ from the vesicles should result in a sharp increase of the P, concentration inside the vesicles, and hence an increment of the passive efflux of P,. The conclusion that the P, efflux was not involved in the ATP synthesis was substantiated by the finding that an increment of the Ca'+ efflux coupled with ATP synthesis was also found in vesicles loaded with calcium oxalate (l-4) (Fig. 6). Makinose and Hasselbach (5) reported that for the increment of Cal+ efflux promoted by ADP and P,, the release of The ADP concentration was 0.5 mM and that of P, 10 mM. Other additions to the assay medium and experimental condition were as described under "Materials and Methods," using SRV loaded with calcium oxalate. The reaction was started by the addition of vesicles and arrested after different incubation intervals at 30" by removal of vesicles with Millipore filters. Left, 0---0, calcium efflux measured in the absence of ADP and P, (passive efflux); O---O, with ADP and P,; and A-A, with ADP, P,, and 12 c(g of X-537 A/ml. Right, ATP synthesis measured enzymatically in media containing ADP and P, (0) or with ADP, P,, and 12 pg of X-537 A/ml (A).