The amyloid beta-peptide (A beta)-associated free radical oxidative stress model for neuronal death in Alzheimer's disease (AD) brain predicts that neuronal protein oxidation is a consequence of A beta-associated free radicals [8]. In this study we have used both in vitro and in vivo models of beta-amyloid (A beta) toxicity to detect free radical induced oxidative stress by the measure of protein carbonyl levels. These model systems employed cultured hippocampal neurons exposed to exogenous synthetic A beta(1-42) and transgenic Caenorhabditis elegans (C. elegans) animals expressing A beta(1-42). We also investigated the importance of the A beta(1-42) Met35 residue for free radical formation in peptide solution and for peptide-induced protein oxidation and neuronal toxicity in these model systems. A beta(1-42) in solution yielded an EPR spectrum, suggesting that free radicals are associated with this peptide; however, neither the reverse [A beta(42-1)] nor methionine-substituted peptide [A beta(1-42)Met35Nlc] gave significant EPR spectra, suggesting the importance of the methionine residue in free radical formation. A beta(1-42) addition to cultured hippocampal neurons led to both neurotoxicity (30.1% cell death, p < 0.001) and increased protein oxidation (158% of controls, p < 0.001). and both of those effects were not observed with reverse or Met35Nle substituted peptides. C. elegans transgenic animals expressing human A beta(1-42) also had significantly increased in vivo protein carbonyls (176% of control animals, p < 0.001), consistent with our model. In contrast, transgenic animals with a Met35cys substitution in A beta(1-42) showed no increased protein carbonyls in vivo, in support of the hypothesis that methionine is important in A beta-associated free radical oxidative stress. These results are discussed with reference to the A beta-associated free radical oxidative stress model of neurotoxicity in AD brain.