Reactive aldehydes might have a pivotal role in the pathogenesis of atherosclerosis by covalently modifying low-density lipoprotein (LDL). However, the identities of the aldehyde adducts that form on LDL in vivo are not yet clearly established. We previously demonstrated that the haem protein myeloperoxidase oxidizes proteins in the human artery wall. We also have shown that p-hydroxyphenylacetaldehyde (pHA), the aldehyde that forms when myeloperoxidase oxidizes L-tyrosine, covalently modifies the N(epsilon)-lysine residues of proteins. The resulting Schiff base can be quantified as N(epsilon)-[2-(p-hydroxyphenyl)ethyl]lysine (pHA-lysine) after reduction with NaCNBH(3). Here we demonstrate that pHA-lysine is a marker for LDL that has been modified by myeloperoxidase, and that water-soluble, but not lipid-soluble, antioxidants inhibit the modification of LDL protein. To determine whether myeloperoxidase-generated aldehydes might modify LDL in vivo, we used a combination of isotope-dilution GC-MS to quantify pHA-lysine in aortic tissues at various stages of lesion evolution. We also analysed LDL isolated from atherosclerotic aortic tissue. Comparison of normal and atherosclerotic aortic tissue demonstrated a significant elevation (more than 10-fold) of the reduced Schiff base adduct in fatty streaks, intermediate lesions and advanced lesions compared with normal aortic tissue. Moreover, the level of pHA-lysine in LDL recovered from atherosclerotic aortic intima was 200-fold that in plasma LDL of healthy donors. These results indicate that pHA-lysine, a specific covalent modification of LDL, is generated in human atherosclerotic vascular tissue. They also raise the possibility that reactive aldehydes generated by myeloperoxidase have a role in converting LDL into an atherogenic lipoprotein.