The reaction mechanism of the molybdoenzyme xanthine oxidase has been further investigated by 13 C and 17 O ENDOR of molybdenum(V) species and by kinetic studies of exchange of oxygen isotopes. Three EPR signalgiving species were studied: (i) Very Rapid, a transient intermediate in substrate turnover, (ii) Inhibited, the product of an inhibitory side reaction with aldehyde substrates, and (iii) Alloxanthine, a species formed by reaction of reduced enzyme with the inhibitor, alloxanthine. The Very Rapid signal was developed either with [8-13 C]xanthine or with 2-oxo-6-methylpurine using enzyme equilibrated with [ 17 O]H2O. The Inhibited signal was developed with 2 H 13 C 2 HO and the Alloxanthine signal by using [ 17 O]H2O. Estimates of Mo−C distances were made, from the anisotropic components of the 13 C-couplings, by corrected dipolar coupling calculations and by back-calculation from assumed possible structures. Estimated distances in the Inhibited and Very Rapid species were about 1.9 and less than 2.4 Å, respectively. A Mo−C bond in the Inhibited species is very strongly suggested, presumably associated with side-on bonding to molybdenum of the carbonyl of the aldehyde substrate. For the Very Rapid species, a Mo−C bond is highly likely. Coupling from a strongly coupled 17 O, not in the form of an oxo group, and no coupling from other oxygens was detected in the Very Rapid species. No coupled oxygens were detected in the Alloxanthine species. That the coupled oxygen of the Very Rapid species is the one that appears in the product uric acid molecule was confirmed by new kinetic data. It is concluded that this oxygen of the Very Rapid species does not, as frequently assumed, originate from the oxo group of the oxidized enzyme. A new turnover mechanism is proposed, not involving direct participation of the oxo ligand group, and based on that of Coucouvanis et al. [