Calcium Transport Driven by a Proton Gradient in Inverted Membrane Vesicles of Escherichia coZi *

Calcium transport into inverted vesicles of Escherichia coli was observed to occur without an exogenous energy source when an artificial proton gradient was used. The orientation of the proton gradient was acid inside and alkaline outside. Either phosphate or oxalate was necessary for transport, as was found for respiratory-driven or ATP-driven uptake (Tsuchiya, T., and Rosen, B. P. (1975) J. Biol. Chem. 250, 7687-7692). Phosphate accumulation was found to occur in conjunction with calcium accumulation. Calcium transport driven by an artificial proton gradient was stimulated by dicyclohexylcarbodiimide, an inhibitor of the Mg’+ATPase (EC 3.6.1.3). Valinomycin, which catalyzes electrogenic potassium movement, stimulated calcium accumulation, while nigericin, which catalyzes electroneutral exchange of potassium and protons, inhibited both artificial proton gradient-driven transport and respiratory-driven transport. Other properties of the proton gradient-driven system and the previously reported energy-linked calcium transport system are similar, indicating that calcium is transported by the same carrier whether energy is supplied through an artificial proton gradient or an energized membrane state. These results suggest the existence of a calcium/proton antiport.

, and such vesicles are believed to be predominantly right side out, that is, they retain the same orientation of the cytoplasmic membrane as found in whole cells (21). In contrast, an inward flow of protons was found in vesicles prepared by lysis of cells in a French press during NADH oxidation or ATP hydrolysis (22). These results suggest that an artificial proton gradient alkaline outside and acid inside should be used with inverted vesicles in order to drive calcium uptake.
We now report that the establishment of such a proton 962 Proton-coupled Calcium Transport gradient can directly cause the uptake of calcium in inverted vesicles of E. coli. The uptake is inhibited by ionophores which dissipate the proton gradient but not by ionophores which dissipate only the membrane potential. These observations suggest that calcium transport occurs via a calcium/proton antiport.

MATERIALS AND METHODS
Growth of Cells--Escherichia coli strain 7 (23) cultures were grown to midexponential phase in a basal salts medium (24) supplemented with 68 mM glycerol as a carbon source.
Chemicals-'5CaC1, ( Transport Assays-Transport assays were performed as described previously (15) in a buffer consisting of 10 rnM Tris-HCl, 10 rnM potassium phosphate, and 0.14 M KCl, with modifications depending on the assay. The experiments described in Fig. 5B were performed as described previously (15). In assays using an artificial proton gradient the pH of the assay and wash buffer was 8.5 instead of 8.0, except in the experiments described in Fig. 2B, where the pH of the assay and wash buffer was varied as described in the text. In the experiments described in Figs. 1 to 5A, the reaction was initiated by a 10. to 20.fold dilution of the vesicle suspension into transport assay buffer containing %aCl, rather than by addition of "CaCl,.
In control experiments vesicles were added to the transport assay buffer 10 min prior to the initiation of the assay by the addition of '%aCl,.
The concentration of '%aCI, was 0.5 mM in all assays, except that 0.05 mM YZaCI, was used when oxalate was present in the assay. The concentration of vesicles in the assays was 0.1 to 0.24 mg/ml of membrane protein.
Protein Assays-Protein assays were performed according to a micro modification of the method of Lowry et al. (25).

RESULTS
Calcium Transport Driven by Proton Gradient-Vesicles were washed and resuspended in an unbuffered medium as described under "Materials and Methods" and then transferred to buffered media at various pH values in the presence of 'YJaCl. As shown in Fig. 1, calcium transport occurred only when the outer pH was alkaline (pH 8.5). Neutral (pH 7.3) and acidic (pH 6.5) buffers were not effective. In a control experiment vesicles were allowed to remain in alkaline buffer for 10 min prior to the addition of calcium in order to allow the pH gradient to dissipate, and no calcium uptake was observed. Thus, inward flow of calcium was found concomitant with the establishment of a proton gradient. Presumably there would be an outward flow of protons, although this is not directly measured.
When the magnitude of the pH gradient was decreased, either by keeping the outer pH at 8.5 and raising the inner pH ( Fig. 2A) or by keeping the inner pH at 5.6 and lowering the outer pH (Fig. 2B), the magnitude of calcium uptake was likewise decreased. As the outer pH was increased, the Control level of calcium radioactivity associated with the vesicles increased, perhaps because of the lower solubility product of calcium phosphate at high pH. For that reason our standard assay utilized pH 8.5 for the outer pH. It is not clear why little accumulation was observed at pH 8.0, although the formation of a transient calcium phosphate precipitate may be necessary to observe the reaction. Although the level of calcium within the vesicles decreases with time as the pH gradient dissipates, we cannot be certain of the concentration of free calcium ion within the vesicles and consequently cannot be certain that accumulation against a gradient has occurred. However, the fact that the level of '%a*+ inside of the vesicles just after the change in pH is higher than that found after 30 min in the presence of D-1aCtat.e (15) suggests that concentration against a gradient has occurred.
Anion Requirement for Proton Gradient-driven Calcium Transport-A requirement for phosphate or oxalate in the transport of calcium has been found in mitochondria (26), sarcoplasmic reticulum (27), and inverted vesicles of E. coli (14,15). This requirement has been interpreted as a mechanism for trapping calcium inside as an insoluble precipitate or, by action as a counterion.
As shown in Fig. 3, proton gradient-driven calcium uptake also requires phosphate or oxalate. With oxalate as the anion no efflux of calcium occurred after the pH gradient dissipates (Fig. 3B), while efflux was observed when phosphate was used (Fig. 3A). However, precipitation with phosphate must occur to some extent since the level of calcium within the vesicles never returns to the control level. As stated above, calcium may form a transient complex with phosphate even at early times. Since the wlubility constant of calcium oxalate is much lower than calcium phosphate, the external concentration of calcium had to be reduced in assays using oxalate, accounting for the difference in the absolute level of calcium transport between assays with phosphate and those with oxalate. Neither acetate nor arsenate were effective anions.
We have previously reported the accumulation of '*P, during the accumulation of calcium (15). Since phosphate can be replaced by oxalate, it is unlikely that the transport system is a calcium/phosphate symport. Phosphate was also accumulated concomitant with calcium accumulation in the proton gradi- ent-driven system (Fig. 4). With NADH-driven uptake, phosphate appeared to enter more slowly than calcium, with a final Ca*+/P, ratio of about 1.5, suggesting the formation of Ca,(PO,), (15). The final ratio in the experiment described in Fig. 4 is also about 1. 5 5A) or by respiration (Fig. 5B).
Since these compounds can also affect the electron transport chain, it was not possible to determine if they acted directly on the calcium transport system. As shown in Table  I Fro 5. Effect of ionophores on calcium transport. Vesicles were prepared and assayed for transport as described under "Methods," so that the concentration of potassium was approximately 0.14 M both inside the vesicles and in the assay medium. Vesicles were incubated. for 10 min in the presence of 2 @g/ml of valinomycin (Val., A-A), 2 &ml of nigericin (Nig., W---D), or 0.2% ethanol (04) before dilution into assay buffer containing those compounds at the same concentrations. Valinomycin and nigericin were prepared in ethanol. A, calcium transport driven by an artificial proton gradient using 0.23 mg/ml of membrane protein, final concentration; B, calcium transport in the presence of 20 mM n-lactate using 0.24 mg/ml of membrane protein. Membrane vesicles were prepared and assayed for calcium transport as described under "Methods." Each compound was added to both the vesicles and the assay mixture 10 min prior to the start of the assay. Divalent cations were used at 0.1 mM final concentrations. KCN  driven by ATP hydrolysis. Similarly, cyanide stimulated proton gradient-driven calcium uptake (Table I) without increasing the background level. Thus, the stimulatory effect of cyanide appears to be directly on the transport system.
Thiocyanate is a permeant anion which, like valinomycin in the presence of potassium, can dissipate a membrane potential without dissipating a proton gradient. Thiocyanate had no inhibitory effect and may have been slightly stimulatory (Table I) as was found for valinomycin. These results suggest that the chemical gradient of protons is the portion of the protonmotive force which derives calcium uptake, and that the presence of a membrane potential may be to some extent inhibitory in this artificial system and possibly in the natural energy-coupling system as well.
As shown in Table II, pCMB inhibitied calcium uptake, while 2-mercaptoethanol reversed the effect of pCMB. These results are similar to those found for energy-linked calcium Membrane vesicles were prepared and assayed for calcium transport as described under "Methods," except that 2-mercaptoethanol was omitted from all buffers. The concentrations of &MB and 2-mercaptoethanol were 0.2 and 2 mM, respectively. The two reagents were added to both membranes and assay medium 10 min prior to the start of the assay. The uptake values given are those found just before efflux of calcium. uptake (15). Thus, it appears likely that the calcium transport system studied here is the same as that reported previously (15).

DISCUSSION
Active transport of small molecules in E. coli has been shown to occur by two different mechanisms, one utilizing directly phosphate bond energy and the other utilizing the "high energy state of the membrane" (34,35). The calcium transport system falls into the second category for several reasons: (a) it is present in membrane vesicles (14), while systems of the first type are not; (b) it can utilize the energy of oxidation by the respiratory chain even in the absence of the Mg*+ATPase (33), while phosphate bond-linked systems can use respiration only when the MgS+ATPase is functional; and (c) it can utilize ATP only when the Mg*+ATPase is functional (15), while systems of the other type use the energy of ATP by some other, undefined mechanism.
The most likely form in which the "high energy membrane state" occurs is a protonmotive force, consisting of a membrane potential and a chemical gradient of protons, both derived from the electrogenic extrusion of protons by the electron transport chain or by the Mg*+ATPase (2,3). Mitchell has postulated that transport systems can use either or both of these component parts of the protonmotive force to drive transport systems of the second type (3). In this proposal proton/substrate symports utilize protons directly, where proton movement through the symport could be driven by a chemical gradient of protons or a membrane potential, or both. Cations may also be extruded from whole cells by cation/proton antiports, as has been shown for sodium (36). If such cations are extruded from whole cells, then they should be accumulated by inverted vesicles. However, since the vesicles are inverted, the sign of the electrochemical gradient of protons must be the reverse of that found in whole cells, that is, protons must be translocated into inverted vesicles. Our results demonstrate that a gradient of protons, acid inside, causes the uptake of calcium into vesicles. Moreover, the uptake is dependent on the chemical gradient of protons, since it is inhibited by nigericin, but does not depend on the membrane potential directly, since valinomycin and thiocyanate do not inhibit.
The rate at which the proton gradient dissipates must be related to the uptake of calcium, As the gradient dissipates, Proton-coupled-Calcium Transport calcium (and phosphate) efflux occurs (Fig. 4). Compared with the transport of galactosides driven by a proton gradient in whole cells (9)