Inside-out erythrocyte membranes attached to polycationic beads manifested glutathione disulfide (GSSG)-stimulated ATPase activity. A Lineweaver-Burk plot of the ATPase activity as a function of GSSG concentration was biphasic and gave apparent Km values of 0.13 mM and 2.0 mM. These kinetics are similar to those reported for the ATP-requiring GSSG-transport systems in human erythrocytes and for the GSSG-stimulated ATPase activity in the plasma membranes of rat hepatocytes. Erythrocyte membranes that were depleted of extrinsic proteins were solubilized in 0.5% Triton X-100. Affinity chromatography on S-hexylglutathione-Sepharose 6B, with elution by a linear gradient of S-hexyl-glutathione, resulted in the resolution of two peaks of enzyme activity. One enzyme, which was eluted at approximately 0.5 mM S-hexylglutathione, had a high affinity for GSSG (apparent Km of 150 microM) and for ATP (80 microM). The other enzyme, which was eluted at approximately 1 mM S-hexylglutathione, had a low affinity for GSSG (apparent Km of 2.0 mM) and ATP (140 microM). GSSG-independent Mg2+-ATPase, Ca2+-dependent Mg2+-ATPase and Na+, K+-dependent Mg2+-ATPase were undetectable in the fractions. Addition of Ca2+, ouabain, or vanadate neither activated nor inhibited the activities, further indicating that the enzymes are distinguishable from ion-pumping ATPases. The enzymes required GSSG for activation; reduced glutathione (GSH) was ineffective. The ATPase activity of the high-Km enzyme was inhibited by addition of p-chloromercuribenzoate, N-ethylmaleimide, and iodoacetamide and was activated by treatment with dithiothreitol, whereas the ATPase activity of the low-Km enzyme was not modified by these thiol reagents. The properties of the enzymes are similar to those of ATP-dependent GSSG-transport systems in human erythrocytes, suggesting that these ATPases may function in the active transport of GSSG.