Mitochondrial creatine kinase was first proposed to act as a functional component in respiratory control in 1966 (Bessman, S. P., and Fonyo, A. (1966) Biochem. Biophys. Res. Commun. 22, 597-602). Since that time, evidence has accumulated to support the theory of a creatine-phosphorylcreatine shuttle mechanism involved in supplying energy for aerobic muscle contraction (Bessman, S. P., and Geiger, P. J. (1981) Science 211, 448-452). To demonstrate directly the interaction between mitochondrial oxidative phosphorylation and that of creatine phosphate synthesis, we have studied the labeling of adenine nucleotides and creatine phosphate with [33P]H3PO4 or [gamma-32P]ATP over a range of adenine nucleotide concentrations incubated with rabbit cardiac and rat skeletal muscle mitochondria. An apparent direct mitochondrial ATP contribution to creatine phosphate synthesis was observed that varied inversely with the total adenine nucleotide present in the reaction system. This reaction of de novo synthesized ATP with creatine phosphokinase prior to equilibration with the total ATP pool was observed regardless of the entry point of electrons from oxidizable substrate into the electron transport chain. This special relation was not observed for added yeast hexokinase in forming glucose 6-phosphate. Mitochondria could not synthesize creatine phosphate in the presence of atractyloside, thus underscoring the requirement for adenine nucleotide translocase-linked transport of ATP prior to reaction with the bound creatine phosphokinase. These studies show that there is coupling or compartmentation of ATP synthesis and transport with creatine phosphate formation in heart and skeletal muscle mitochondria.