The pathogenesis of Alzheimer's disease (AD) involves a key event which changes the morphology of amyloid-beta 42 (A beta(42)) peptide from its soluble monomeric form into the fibrillated aggregates in the brain. Aluminum ion, Al(III), is known to act as a pathological chaperone of the A beta(42) in this process; curcumin, a natural phenolic compound, is considered capable of binding Al(III) and A beta(42); nevertheless, little is known about the combined action of curcumin and Al(III) on the A beta(42) fibrillation and neurotoxicity. Here, combinations of circular dichroism spectroscopy, thioflavin T fluorescence, atomic force microscopy, Bradford and MTT assays, it is demonstrated that although Al(III) can promote the A beta(42) fibrillation dose-dependently, leading to the high neurotoxicity to PC12 cells, curcumin can inhibit the events. Besides, we found that curcumin is able not only to inhibit the formation of Al(III)-induced A beta(42) fibrillation, but also to form the Al(III)-curcumin complexes which in turn can remold the preformed, mature, ordered A beta(42) fibrils into the low toxic amorphous aggregates. These findings suggest that curcumin could block the binding of Al(III) with A beta(42) and form the Al(III)-curcumin complexes, so as to inhibit the Al(III)-induced A beta(42) fibrillation and neurotoxicity. The Al(III)-curcumin complexes are worth potentially developing as a therapy agent against the neurodegenerative disorders in the future. (C) 2012 Elsevier B.V. All rights reserved,