The ADP analogue NAP(3)-2N(3)ADP is able to bind to one or two high-affinity sites on mitochondrial F-1-ATPase, depending on the nucleotide content of the F-1 preparation. In both cases studied (enzyme with three bound nucleotides and enzyme with four bound nucleotides), the binding is accompanied by the exchange of one tightly-bound adenine nucleotide and nearly complete inhibition of the ATPase activity upon UV illumination. In both cases the ADP-analogue binds at a high-affinity catalytic site, replacing a bound nucleotide. The apparent K-D value for the exchange equals 25-30 mu M, but the newly bound ligand does not dissociate. With F-1 containing 3 bound nucleotides NAP(3)-2N(3)ADP is able to bind to a second high-affinity site as well. This binding induces already in the absence of illumination 45% inhibition of the ATPase activity. The additionally bound molecule does not exchange within a short period of turnover with Mg-ATP. Therefore it has to be bound at a slowly exchangeable non-catalytic site, with a regulatory influence on the activity of the enzyme. Binding of NAP(3)-2N(3)ADP to this non-catalytic site is influenced by the presence of Mg2+ or EDTA: tight binding requires Mg2+ and in the absence of Mg2+ and presence of EDTA the ligand is removed from this site relatively easily, just like ADP. The presence of EDTA instead of Mg2+ lowers the measured affinity of this site for NAP(3)-2N(3)ADP with a factor 5. Kinetic measurements after an incubation of F-1 with NAP(3)-2N(3)ADP show a decrease of the V-max with ATP as substrate, without effect on the two measured K-m values. With ITP as substrate, however, incubation of F-1 with NAP(3)-2N(3)ADP results in an increase of the K-m values, without effect on the V-max. Comparison of our data with the literature shows that this non-catalytic site is not the site responsible for hysteretic inhibition by ADP. We conclude that this latter form of inhibition is observed when ADP or a suitable analogue is bound at the first (potentially catalytic) beta-site, in disagreement with the conclusions of Jault and Allison (J. Biol. Chem. 269 (1994) 319-325).