Treatment of xanthine oxidase with ferricenium at high pH gives rise to an EPR signal not previously seen with this enzyme. The signal is apparently isotropic at 9 GHz with a gavg of approximately 2 and once generated is stable to pH 6.0, so long as the sample is kept in the dark. Treatment of the signal-giving species with hydroxyurea results in complete loss of the signal, indicating that the signal is radical-based. Pretreatment of the enzyme with iodoacetate has no effect on signal formation with ferricenium. The ferricenium-generated EPR signal shows proton hyperfine coupling that is not lost upon exchange into D2O and bears considerable resemblance to the tyrosyl radical of the photosynthetic reaction center and other systems. These observations lead us to interpret the new ferricenium-generated EPR signal of xanthine oxidase as arising from a tyrosyl radical, the result of one-electron oxidation of a protein tyrosinate residue. Kinetic parameters for the reductive half-reaction of ferricenium-treated xanthine oxidase with xanthine were determined by stopped-flow spectrophotometry; kred and KDxanthine (15 s-1 and 12 microM, respectively) were essentially unchanged. Addition of 2-hydroxy-6-methylpurine (in the presence of 2 mg/mL catalase and superoxide dismutase) generated the "very rapid" MoV EPR signal while preserving the ferricenium-derived EPR signal, providing a further indication that the modified enzyme remains fully functional and the presence of the tyrosyl radical does not impact turnover by the enzyme. Coupling of the two signals was not evident, nor was coupling to the two 2Fe-2S centers or the flavin semiquinone evident. The implications of covalent modifications of proteins mediated by ferricenium are discussed.