Re-valuation of the Inverse Docking technique to compute drugs in multi-target studies

The molecule-receptor interaction studies are limited by the possibility to analyse only one target without having the possibility to consider interactions with systemic protein systems. This is restrictive limitation taking in consideration that these molecules could represent potential hits for the treatment of several diseases. To fully develop the potentiality of the Inverse Docking (ID)1,2 technique an exhaustive analysis of the interactions between a number of ligands and a protein database, such us the Protein Data Bank (PDB), was conducted using different softwares like MOE5, AutoDock Vina3, Knime4 and Amber6 with NAMD7, in a concerted manner. We were capable to automate several different processes with a good yield with respect computer time consuming and to ameliorate time operator, at least in small-scale. The docking studies have been conducted on a multi-conformers system, as suggested by Chen et al..1 Multi-conformers have been obtained through an automatized conformational search conducted by Knime, implemented with MOE nodes. Moreover, docking analysis were conducted with scripts that culminate in an automated Autodock Vina starting, thus obtaining accurate docking scores and RMSD values. The best ligand-protein interaction will be submitted in a brief semi-automatized dynamic simulation. The protein systems involved in the present study are metalloproteinases (MMPs), namely MMP-13 and MMP-3, and recently opioid co-crystallised receptors (µ, δ, κ). To evaluate the goodness of the depicted workflow, we selected co-crystallised molecules like a blank control and a series of compounds developed and published by us and by other research groups.