The formation of beta-amyloid (Abeta) deposits in the brain is likely to be a seminal step in the development of Alzheimer's disease. Recent studies support the hypothesis that Abeta soluble oligomers are toxic to cells and have potent effects on memory and learning. Inhibiting the early stages of Abeta aggregation could, therefore, provide a novel approach to treating the underlying cause of AD. We have designed a retro-inverso peptide (RI-OR2, H(2)N-r<--G<--k<--l<--v<--f<--f<--G<--r-Ac), based on a previously described inhibitor of Abeta oligomer formation (OR2, H(2)N-R-G-K-L-V-F-F-G-R-NH(2)). Unlike OR2, RI-OR2 was highly stable to proteolysis and completely resisted breakdown in human plasma and brain extracts. RI-OR2 blocked the formation of Abeta oligomers and fibrils from extensively deseeded preparations of Abeta(1-40) and Abeta(1-42), as assessed by thioflavin T binding, an immunoassay method for Abeta oligomers, SDS-PAGE separation of stable oligomers, and atomic force microscopy, and was more effective against Abeta(1-42) than Abeta(1-40). In surface plasmon resonance experiments, RI-OR2 was shown to bind to immobilized Abeta(1-42) monomers and fibrils, with an apparent K(d) of 9-12 muM, and also acted as an inhibitor of Abeta(1-42) fibril extension. In two different cell toxicity assays, RI-OR2 significantly reversed the toxicity of Abeta(1-42) toward cultured SH-SY5Y neuroblastoma cells. Thus, RI-OR2 represents a strong candidate for further development as a novel treatment for Alzheimer's disease.