Detection of Molecular Interaction Field Similarities for the Rational Drug Design of Multi-Functional Inhibitors

Challenge: Binding promiscuity plays a major role in medicine as promiscuous drug interactions may lead to undesirable cross-reactivity effects. This represents a considerable challenge for the pharmaceutical industry. Drugs act by modulating the function of target proteins. However, additional off-site non-target proteins may also be affected due to similarities between binding sites. This unintentional effect may lead to the serendipitous discovery of new applications for a specific drug, but it can equally result in unwanted safety concerns.

Solution: The detection of such similarities early in the drug discovery process may prevent deleterious side effects. To address this issue, the researcher proposed to develop novel software for the detection of 3D atomic similarities in the binding sites of drugs. The software, called IsoMIF, fills a gap in the current computational methods used for drug design since it allows the detection of similarities across protein families at the level of molecular interactions. IsoMIF uses X-ray, NMR or homology modeling 3D structure of a protein as input to calculate combined molecular interactions termed Molecular Interaction Field (MIF) for proteins and detect MIF similarities between pairs of MIFs. IsoMIF has been validated against experimental data for the detection of similarities across protein families. IsoMIF can detect binding-site similarities in cases where neither sequence similarities nor fold similarities can predict binding cross reactivity.

Achievements/Impacts: Rather than limiting the focus on the specific relative position of atoms in the surface of binding-sites, a more global vision of the combined total interactions is considered critical for selective rational drug design. IsoMIF is proving to be a powerful tool for drug design with the ability to predict binding cross reactivity in drug development and drug repurposing for new targets via MIF similarities. It can be also used to analyse and predict protein function from structure and exploit MIF similarities to target two and potentially more proteins at once within the concept of poly-pharmacology.

Principal Investigator:

Rafael Najmanovich
University of Sherbrooke

Completed Project
$ 300,000 / 2 years
Supported by CQDM through:
• AstraZeneca
• Boehringer Ingelheim
• Eli Lilly
• Merck
• Pfizer