A COMPUTER AIDED EXCURSUS INTO MOLECULAR BIOSYSTEMS

The application of computational methods to study biological and biochemical processes is a growing field in both industrial and academical environment. They can be used to explore condition that would be extremely expensive or flat out impossible to reproduce in a laboratory; to filter out molecules that would be unlikely to react with the designed target, thus reducing the price needed to produce a new drug; to understand experimental data by looking at the behaviour of molecules and proteins at an atomistic level. In this PhD thesis, molecular modeling, molecular dynamics and different enhanced sampling methods are employed to address different aspects of relevant biosystems. • Chapter 1 is a theoretical introduction to the principles of molecular dynamics and the different approaches employed in the following chapters. It also explains the strengths and weaknesses of each method and how are they relevant in their respective fields of application. • Chapter 2 focuses on the interaction between Aspergillomarasmine-A and the zinc ion of a Metallo--lactamases. Metallo--lactamases are a relevant target in fighting bacterial resistance due to their ability to inactivate -lactam rings that are the core fragment in penicillin like antibiotics. To study this process we had to develop a method that was able to overcome the limitations of the force fields used to simulate the protein and the ligand, which were not able to correctly predict the coordination geometry. We could generate different conformations using well tempered metadynamics and filter them in order to obtain, after a quantomechanical minimization the correct structures. • In Chapter 3 six nucleosides obtained by modifying the phosphate sidechain of guanosine triphosphate and their ability to interfere with the microtubule dynamics is analyzed. This study is carried out by calculating the binding free energy of the six nucleosides with the - of tubuline. We also investigated which aminoacids are more important in the process by calculating the binding free energy difference upon mutation for each of the 36 aminoacids in the binding site. • In Chapter 4 is described the conformational study of a small peptide containing a non natural aminoacid. We simulated the two possible enantiomers and both the possible rotamers of one of the two enantiomer with well tempered metadynamics to study the ability of the new aminoacid to act as a mimetic of a -turn. To better understand the small peptides behaviour we also ran simulations where the NMR data were considered. • In Chapter 5 the protein protein interactions between two monomer of the universal protein YeaZ which is composed by five identical monomer were investigated. We started by running a Solvent Accessible Surface Area analysis to identify which residues were located in the protein protein interaction surface. We then proceded to calculate which residues were important to the binding by performing a computational alanine scanning which allowed us to identify a small peptide that, in our simulations, proved to be able to bind in the interaction site. We then tried to improve the stability of the peptide by stiffening the helix. • The final chapter describes the work carried on in collaboration with Prof. Birgit Schiøtt during my staying at Aarhus University. Here we employed accelerated molecular dynamics to simulate the binding of glutamine to gamma-glutamyltransferase. We ran 90 simulations which resulted in three binding events. By analyzing these three simulations we shed light on the residues that are more important during the binding process, recognized four general configurations that are common in the successful simulations and elucidated the importance of a loop known as lid-loop in the recognition of the substrate.