Analysis of protein-protein interaction based pharmacophores as a screening tool for small molecule protein-protein interaction inhibitiors

Pharmacophore modelling is a well-known structure based drug design method allowing fast virtual screening of huge libraries of (commercial available) compounds. The construction of a pharmacophore model based on the protein-protein interface is a prerequisite for success. Here we present an analysis of the amino acid composition of the protein-protein interfaces with respect to the amino acids mimicked by small molecule protein-protein interaction inhibitors (SMPPII) representing the pharmacophore features. A dataset was screened and a series of properties (interacting surface area and the Van der Waals, electrostatic and hydrogen bonding contributing to the total interaction energy) were calculated for all amino acids involved in protein-protein interaction. The obtained energy values of the mimicked amino acids were compared. Rules could be deduced allowing a more precise amino acid selection to be used in the construction of a pharmacophore model. From the superposition of the protein-protein structure and the protein-SMPPII we could deduct rules for the pharmacophore transcript of the interacting entities. This analysis shows that the pharmacopore is based on a sub-population of the amino acids present in the total protein-protein interface. The amino acids, used as pharmacophore features, have lower interaction energy, are hydrophobic and interact via a large surface area. In conclusion; structure based discovery (SMPPII as well as small molecules) can be facilitated by using valid pharmacophores derived from the protein-protein interfaces.status: publishe