Treatment of malic enzyme with arginine-specific reagents phenylglyoxal or 2,3-butanedione results in pseudo-first-order loss of oxidative decarboxylase activity. In-activation by phenylglyoxal is completely prevented by saturating concentrations of NADP+, Mn2+, and substrate analog hydroxymalonate. Double log plots of pseudo-first-order rate constant versus concentration yield straight lines with identical slopes of unity for both reagents, suggesting that reaction of one molecule of reagent per active site is associated with activity loss. In parallel experiments, complete inactivation is accompanied by the incorporation of four [14C]phenylglyoxal molecules, and the loss of two arginyl residues per enzyme subunit, as determined by the colorimetric method of Yamasaki et al. (R. B. Yamasaki, D. A. Shimer, and R. E. Feeney (1981) Anal. Biochem. 14, 220-226). These results confirm a 2:1 ratio for the reaction between phenylglyoxal and arginine (K. Takahashi (1968) J. Biol. Chem. 243, 6171-6179) and yield a stoichiometry of two arginine residues reacted per subunit for complete inactivation, of which one is essential for enzyme activity as determined by the statistical method of Tsou (C. L. Tsou (1962) Acta Biochim. Biophys. Sinica 2, 203-211) and the Ray and Koshland analysis (W. J. Ray and D. E. Koshland (1961) J. Biol. Chem. 236, 1973-1979). Amino acid analysis of butanedione-modified enzyme also shows loss of arginyl residues, without significant decrease in other amino acids. Modification by phenylglyoxal does not significantly affect the affinity of this enzyme for NADPH. Binding of L-malate and its dicarboxylic acid analogs oxalate and tartronate is abolished upon modification, as is binding of the monocarboxylic acid alpha-hydroxybutyrate. The latter result indicates binding of the C-1 carboxyl group of the substrate to an arginyl residue on the enzyme.