COMPUTATIONAL APPROACHES TO STUDY METALLOPROTEINS AND ZINC-CONTAINING SYSTEMS

Computational representation of metal-containing systems is a considerable challenge that derives from the complex nature of the metal coordination bond. Herein we tested computational methods with increasing level of accuracy to handle zinc containing systems, starting from simplistic approaches and moving to more complex descriptions. Fragment-based molecular docking was applied for drug design of novel inhibitors of HDAC, a metallo-enzyme containing a catalytic zinc ion within the binding pocket. Although metal ion interactions were treated with classic approaches, some promising compounds have been identified. Then docking, MD simulations and MM-GBSA free energy calculations were applied to study the interaction between AChE and a new metallo-organic reactivator. The MD simulation was carried out using the 12-6-4 non-bonded model, which was able to maintain the integrity of the ligand and its correct octahedral geometry through all the trajectories. A similar workflow was applied also for drug discovery of new inhibitors of CN1, a di-peptidase containing two catalytic zinc ions within the binding pocket. Small molecules able to coordinate the metal ions were found by docking of different libraries, then the hit compounds underwent to MD simulations using the cationic dummy atoms approach, where we parametrized de novo the five-coordinated zinc ion. As a last step, the free binding energy was computed using the MM-GBSA method. Finally, a multiscale approach for the binding mode prediction of new metallo-b-lactamase IMP-1 inhibitors was developed. Docking and MD simulation using a restrained non-bonded model were applied, followed by a QM/MM refinement step, which was able to recover the correct coordination geometry and to fix the errors introduced by the MD simulation. Different modelling techniques applied to representative case studies are reported, considering increasing level of accuracy and illustrating the advantages and weakness of each approach.