Native FAD was removed from chicken liver xanthine dehydrogenase (XDH) and replaced with a number of artificial flavins of different redox potential. Dithionite titration of the 2-thio-FAD- or 4-thio-FAD (high potential)-containing enzymes showed that the first center to be reduced was the flavin. With native enzyme, iron-sulfur centers are the first to be reduced. With the low potential flavin, 6-OH-FAD, the enzyme-bound flavin was the last center to be reduced in reductive titration with xanthine. These shifts in the reduction profile support the hypothesis that the distribution of reducing equivalents in multi-center oxidation-reduction enzymes of this type is determined by the relative potentials of the centers. The reaction of molecular oxygen with fully reduced 2-thio-FAD XDH or 4-thio-FAD XDH resulted in 5 electron eq being released in a fast phase and one in a slow phase. Reduction of these enzymes by xanthine was limited at a rate comparable to that for the release of urate from native XDH. Xanthine/O2 turnover with these enzymes (and native XDH) resulted in approximately 40-50% of the xanthine reducing equivalents appearing as superoxide. Steady state turnover experiments involving all modified flavin-containing enzymes, as well as native enzyme, showed that shifting the flavin potential either positive or negative relative to FAD caused a decrease in catalytic activity in the xanthine/NAD reductase reaction. In the case of the xanthine/O2 reductase activity, there is no simple obvious relationship between the activity and the redox potential of the reconstituted flavin.