The free-energy landscape of the Alzheimer beta-amyloid peptide Abeta(12-36) in a 40% (v/v) 2,2,2-trifluoroethanol (TFE)/water solution was determined by using multicanonical molecular dynamics simulations. Simulations using this enhanced conformational sampling technique were initiated from a random unfolded polypeptide conformation. Our simulations reliably folded the peptide to the experimental NMR structure, which consists of two linked helices. The shape of the free energy landscape for folding was found to be strongly dependent on temperature: Above 325 K, the overall shape was funnel-like, with the bottom of the funnel coinciding exactly with the NMR structure. Below 325 K, on the other hand, the landscape became increasingly rugged, with the emergence of new conformational clusters connected by low free-energy pathways. Finally, our simulations reveal that water and TFE solvate the polypeptide in different ways: The hydrogen bond formation between TFE and Abeta was enhanced with decreasing temperature, while that between water and Abeta was depressed.

We developed a new molecular dynamics simulation method for protein-ligand binding free energy calculation in an explicit water model. This method consists of three steps: (1) generation of a compound dissociation path starting from a stable protein-compound complex structure, (2) calculation of the free energy surface along the dissociation path, and (3) calculation of the free energy surface of a small area around the protein-compound complex structure, which is a free energy minimum. The protein-compound binding free energy is estimated from the information obtained by the above three steps. This method was applied to a small system, a 18-crown-6 ether with its ligand ion, and a realistic system consisting of a target protein with its inhibitor. This approximation worked well; the protein-inhibitor dissociation was successfully performed, and the binding free energies were calculated.