The Mg2+/2H+ antiporter recently described on lutoid membrane (Z. Amalou, R. Gibrat, C. Brugidou, P. Trouslot, J. d'Auzac [1992] Plant Physiol 100 255-260) was solubilized by odylglucoside and reconstituted into soybean liposomes using the detergent dilution method. Magnesium efflux or influx experiments were used to generate a H+ influx or efflux, respectively, monitored with the fluorescent probe 9-amino-6-chloro-2-methoxyacridine. Both experiments gave saturable H+ fluxes as a function of internal or external Mg2+ concentrations with similar kinetic parameters K, and Vmax. The K, value for Mgz+ (about 2 mM) was identical to that previously found in lyophilized-resuspended lutoid (reference therein), whereas the V,, value was 14-fold higher. Since only 10% of the initial proteins were recovered in proteoliposomes, and electrophoretic patterns of the two kinds of vesicles differed significantly, it was inferred that the increase in Vmax was due essentially to an enrichment of the protein antiporter in the reconstituted fraction, owing to a selective effect of octylglucoside at both solubilization and reconstitution steps. None of the various divalent cations used could dissipate the pH gradient of control liposomes of soybean lipids, unless the divalent/H+ exchanger A23187 was added, whereas a rapid dissipation of the pH gradient was observed with reconstituted proteoliposomes from lutoid proteins, with the cation selectivity sequence ZnZ+ > CdZ+ > Mg2+ in the millimolar concentration range. The divalent ions CaZ+, BaZ+, and MnZ+ were incapable of generating a H+ efflux in reconstituted proteoliposomes, whereas both MgZ+/H+ and CaZ+/H+ exchanges were observed in lyophilized-resuspended lutoids. Therefore, the lutoid membrane seems to contain separate h4g2+/H+ and CaZ+/H+ transport systems, the latter being eliminated during the solubilization/ reconstitution of lutoid membrane proteins.