Computational Study of Solvent Effects on the Stability of Native Structures in Proteins

The commonly accepted protein folding mechanism hinges upon the idea that all the information required for a protein to reach the native conformation are encoded in its primary sequence. However, an equally important role is played by the interaction with the solvent, as recent both experimental and computational studies have clarified. The present thesis will build upon these study to investigate the stability of few representative proteins in the presence of solvents of different nature, ranging from highly polar, as water, to highly non-polar, as ciclohexane. This task will be pursued using a two fold approach. In the first approach, we will collaborate with a group at the University of Kyoto to compare the solvation free energy differences for proteins folding into different solvents, as compared with a number of in-house generated decoys. Within the framework we will develop a tool capable to build several decoys with a wide variety of α-helix and β-sheet contents and three-dimensional misfolded structures. In the second approach, we will use thermodynamic integration and molecular dynamics simulations to compute the same free energy difference with the aim of disentangling the entropic and the enthalpic contributions. This will be carried out at the level of single amino acids both in water and in cyclohexane. The aim of this second part will be to back up with a detailed numerical calculations the results of the approximate calculations carried out in the first part.