Treatment of either beef heart or rat liver mitochondrial ATPase with the arginine reagent, 2,3-butanedione, resulted in enzyme inactivation. The reaction followed pseudo-first order kinetics until 90 to 95% of the enzyme had been inactivated, and prolonged incubation with butanedione resulted in complete inactivation. When the modification reaction was performed in the presence of ATP, the rate of inactivation was significantly decreased. The kinetics of inactivation indicates that the reaction of 1 molecule of reagent per active site of beef heart mitochondrial ATPase is necessary for inactivation. The loss of ATPase activity was also observed when submitochondrial particles were treated with butanedione. Studies with beef heart mitochondrial ATPase indicated that the inactivation was not due to enzyme dissociation into subunits. Kinetic studies with partially inactivated enzyme demonstrated that the Km values of ITP and of ATP in the presence of HCO3-were similar to the same constants for the control enzyme. When ATP was used as the substrate in the absence of anion activator, the partially inactivated enzyme still exhibited negative cooperativity. Inactivation was also observed when beef heart mitochondrial ATPase was treated with another arginine reagent, phenylglyoxal. The loss of ATPase activity was analyzed in terms of [14C]phenylglyoxal incorporation. From the present studies it is concluded that arginyl residues play an essential role in mitochondrial ATPase, probably at the hydrolytic site.