Oxidative reactions that modify proteins have been implicated in the pathogenesis of aging and disease (1 ). It has been difficult to identify the physiologically relevant pathways, however, because the reactive intermediates are short-lived. We attempt to determine which oxidative pathways damage proteins in vivo by first identifying stable end products of potential pathways through in vitro experiments. We then analyze normal and diseased tissues for those compounds. For example, two stable isomers of p -tyrosine—ortho -tyrosine and meta -tyrosine—appear after hydroxyl radical modifies protein-bound phenylalanine residues (2 –4 ). In contrast, o ,o′ -dityrosine forms when hydroxyl radical crosslinks tyrosine residues. o ,o′ -Dityrosine also appears when free or protein-bound tyrosine is attacked by tyrosyl radical (5 ), which is produced from tyrosine and H2 O2 by the heme enzyme myeloperoxidase (6 ,7 ). Tyrosyl radical does not generate ortho -tyrosine and meta -tyrosine, however (2 –5 ). Another oxidant, hypochlorous acid (HOCl), produces 3-chlorotyrosine when it reacts with tyrosine (8 ,9 ). HOCl is generated only by myeloperoxidase, which requires H2 O2 and Cl− to perform the reaction. Thus, determining relative levels of ortho -tyrosine, meta -tyrosine, o ,o′ -dityrosine, and 3-chlorotyrosine can indicate which pathway might have inflicted protein damage in vivo in a particular tissue. These amino acid products are useful markers because they are stable to acid hydrolysis, an essential analytical step.