Abstract P-40: Computational Evaluation of Binding of Heme-Focused Library to Cholesterol-Metabolizing P450 Enzymes from Mycobacterium tuberculosis

Abstract

Background: Tuberculosis (TB), caused by pathogenic bacteria Mycobacterium tuberculosis (MTb), is known to be among top-10 causes of death worldwide. In 2019, WHO mentioned drug-resistance with special focus on drug-resistant TB as one of the top-10 threats to the global health. Hence, there is need for novel antimycobacterial drugs. Methods: Recently, structures of several P450 enzymes from MTb CYP125 and CYP142), capable to conversion of cholest-4-en-3-one to its 25-26-dihydroxy product, have been determined. They are also known to be important for bacterial growth in vitro. Current work is focused on structure-based drug design (SBDD) targeting these enzymes. Using available structures and open data, we analyze the ability of existing computational algorithms to predict binding of fragment-like heterocyclic compounds from experimentally tested library to the enzyme. Results: We find that apo structures of all three enzymes have the most promising capabilities in terms of both overall discriminating ability (ROC AUC) and top-10% enrichment. Conclusion: Obtained results establish a blueprint for both structure-based and fragment-based drug design of novel therapeutics against Mycobacterium tuberculosis.

Keywords

• Mycobacterium tuberculosis
• structure-based drug design