It was shown recently that ATP present at near saturating concentrations did not prevent binding and hydrolysis of submicromolar concentration of trinitrophenyl adenosine triphosphate (Tnp-ATP) by F-1-ATPase [Murataliev, M. B. & Boyer, P. O. (1994) J. Biol. Chem. 269, 15431-15439]. To explore F-1-ATPase binding sites that bind Tnp-ATP a new photoreactive analog of ATP, 2-azido-trinitrophenyl adenosine triphosphate (2-N-3-Tnp-ATP) has been synthesized and used for photoaffinity labeling of mitochondrial F-1-ATPase. The analog shares many properties of the parent non-azido Tnp-ATP as shown from spectral characteristics, binding with F-1-ATPase, and kinetic and inhibition studies, 500 mu M ATP does not prevent binding and hydrolysis of low concentrations of 2-N-3-Tnp-ATP by F-1-ATPase. Photoirradiation of the enzyme-analog complex formed under such conditions results in the labeling of the catalytic-site peptide. This shows that in the presence of near saturating ATP, Tnp-ATP can enter the catalytic cycle and inhibit ATP hydrolysis by initial binding at a third catalytic site. The results give strong evidence that only two catalytic sites need to have bound substrate for near maximal turnover rate, and that three catalytic sites of F-1-ATPase participate equally in catalysis. When F-1-ATPase binds substoichiometric 2-N-3-Tnp-ATP in the presence of Mg2+, illumination of the inactive complex formed results in the covalent labeling of a catalytic site. This shows that F-1-ATPase forms similar inactive complexes when ADP or Tnp-ADP is bound at a catalytic site in the presence of Mg2+. Exposure of the nucleotide-depleted F-1-ATPase to 20 mu M 2-N-3-Tnp-ATP followed by a short incubation with excess of Tnp-ATP results in binding, and, upon illumination, in a covalent labeling of a non-catalytic-sire peptide.