Bovine insulin was reacted at pH 4.0 with submolar amounts of hypochlorite. At least one molecule of insulin was modified per two molecules of hypochlorite added, as estimated by HPLC of native and modified insulin. About 5% of the hypochlorite-modified insulin reacted with dinitrophenylhydrazine (DNPH), a reagent which specifically labels carbonyl groups. The major DNPH-labeled product was isolated from the native insulin on reverse-phase HPLC, using trifluoroacetic acid/water/acetonitrile gradients. The UV spectrum of the major peak on the HPLC diode-array detector was representative of DNPH adducts, with lambda max = 365 nm. Several methods, including total amino acid analysis, tryptic digestion, and collision-induced dissociation-electrospray MS, indicate that the major carbonyl in the DNPH-labeled product was on the amino-terminal phenylalanine of the insulin B-chain. Amino acid analysis indicated that tyrosine was also degraded by hypochlorite, but we could not detect a carbonyl group formed at tyrosine. These findings suggest that the terminal amino groups of proteins are highly vulnerable to carbonyl formation during hypochlorite attack. The use of relatively low amounts of active oxygen species (such as hypochlorite), followed by chromatographic isolation of the protein labeled with a carbonyl-specific reagent, can be a useful approach to the study of reactive sites on proteins.