Measurement of catalysis by MnSOD using direct observation of the UV absorbance of superoxide allows determination of steady-state catalytic constants. Stabilizing superoxide in aprotic solvents such as dimethyl sulfoxide permits the use of stopped-flow spectrophotometry, although significant information is lost in the 2- to 4-msec mixing time; generating superoxide by pulse radiolysis requires no mixing time. Studies show that kcat/Km for the decay of superoxide catalyzed by MnSOD proceeds at diffusion control. Investigations using solvent hydrogen isotope effects and enhancement of catalysis by exogenous proton donors show that kcat near 10(4) sec-1 contains a significant contribution from proton transfer steps. The active site of MnSOD is dominated by a hydrogen bond network comprising the manganese-bound aqueous ligand, the side chains of four residues (Gln-143, Tyr-34, His-30, and Tyr-166 from an adjacent subunit), as well as other water molecules. Interrupting this hydrogen bond network by conservative replacement of residues 30, 34, and 166 causes a 10- to 40-fold decrease in maximal velocity, interpreted as an effect on proton transport to the active site, with smaller effects on kcat/Km. Replacement of Gln-143 causes a much greater decrease in catalytic activity, by two to three orders of magnitude, and causes significant changes to the redox potential as well. During catalysis, MnSOD is inhibited by a peroxide complex of the metal in the active site, different from the inhibition of FeSOD and Cu,ZnSOD by Fenton chemistry. Site-specific mutagenesis of active-site residues alters the extent of product inhibition of MnSOD as well, indicating that this is not only a property of the metal. The replacement of Trp-161 with phenylalanine results in a variant that is completely blocked in catalysis by product inhibition.