CavID: Finding the Identity Between Allosteric Cavities in Therapeutic Protein Targets and in Homologous Human Proteins

The discovery of new pharmaceutical compounds requires the use of computational approaches, since they diminish time and costs involved in this process. A successful strategy is the Structure-Based Drug Design (SBDD), in which a therapeutic target of an infectious organism has its structure explored to find or to design a molecule capable of hindering its functions, for example. However, there is a possible identity between the sequences of the allosteric site of a chosen protein target and homologous human proteins, which would represent the appearance of side effects throughout the intake of the new medicine. In this scenario, we have developed the python-based program CavID, that takes as an input a structure of a target protein and the cavities to be analyzed, as well as a multiple sequence alignment of the enzyme’s sequence with other proteins we desire to avoid overlap. In this work, we aim to apply CavID in different systems to explore its potential
in computational-aided drug discovery. To observe that, CavityPlus server was used to identify cavities on the surface of different enzyme groups and ClustalW was used to obtain a multiple sequence alignment with their sequences against human homologous. It was possible to obtain, for each target protein, detailed information on the most promising allosteric cavities that have lower identity against human proteins. This reduces the possibility of toxicity due to an inhibitor ligand being developed and used in a later treatment in humans. The druggability of the cavities, including the minimal volume of 200 A³ of them, and their pharmacophore models were also assessed and ranked. Then, it allows the CavID user to have more information about how to better explore their objects of study in the drug discovery process.