When heat-activated F1-ATPase from chloroplasts was repeatedly exposed to Mg2+ and 2-azido-ATP, followed by separation from medium nucleotides and photolysis, a total of two sites per enzyme, both catalytic and noncatalytic, were labeled. In a coupled assay with pyruvate kinase about half the activity was lost when one site per enzyme was modified. However, increased modification resulted in no further loss of activity. In contrast, methanol-sulfite activation of the enzyme showed a loss of most of the catalytic capacity when one site per enzyme was modified. Predominant labeling of either one catalytic or one noncatalytic site caused a loss of most of the activity in either assay. An indication that the enzyme modified at one site retained some catalytic activity was verified by measurement of the [18O]Pi species formed when [gamma-18O]ATP was hydrolyzed by partially derivatized enzyme. With either catalytic or noncatalytic site modification, the distributions of [18O]Pi species formed showed that the modified enzyme had different catalytic characteristics. An interpretation is that with modification by azido nucleotides at either catalytic or noncatalytic sites, capacity for rapid catalysis is largely lost but the remaining sites retain weak modified catalytic properties.