The F1 moiety of the rat liver mitochondrial ATP synthase/ATPase complex contains as isolated 2 mol Mg2+/mol F1, 1 mol of which is nonexchangeable and the other which is exchangeable (N. Williams, J. Hullihen, and P.L. Pedersen, (1987) Biochemistry 26, 162-169). In addition, the enzyme binds 1 mol ADP/mol F1 and 3 mol AMP.PNP, the latter of which can bind in complex formation with divalent cation and displace the Mg2+ at the exchangeable site. Thus, in terms of ligand binding sites the fully loaded rat liver F1 complex contains 3 mol MgAMP.PNP, 1 mol ADP, and 1 mol Mg2+. In this study we have used several metal ATP complexes or analogs thereof to gain further insight into the ligand binding domains of rat liver F1 and the mechanism by which it catalyzes ATP hydrolysis in soluble and membrane bound form. Studies with LaATP confirmed that MgATP is the most likely substrate for rat liver F1, and provided evidence that the enzyme may contain additional Mg2+ binding sites, undetected in previous studies of F1-ATPases, that are required for catalytic activity. Thus, F1 containing the thermodynamically stable LaATP complex in place of MgATP requires added Mg2+ to induce ATP hydrolysis. As Mg2+ cannot readily displace La2+ under these conditions there appears to be a catalytically important class of Mg2+ binding sites on rat liver F1, distinct from the nonexchangeable Mg2+ site and the sites involved in binding MgATP. Additional studies carried out with exchange inert metal-nucleotide complexes involving rhodium and the Mg2+ and Cd2+ complexes of ATP beta S and ATP alpha S imply that the rate-limiting step in the ATPase reaction pathway occurs subsequent to the P gamma-O-P beta bond cleavage steps, perhaps at the level of Mg(ADP)(Pi) hydrolysis or MgADP release. Evidence is presented that Mg2+ remains coordinated to the leaving group of the reaction, i.e., the beta phosphoryl group. Finally, in contrast to soluble F1, F1 bound to F0 in the inner mitochondrial membrane failed to discriminate between the Mg2+ complexes of the ATP beta S isomers. This indicates that a fundamental difference may exist between the catalytic or kinetic mechanism of F1 and the more physiologically intact F0F1 complex.