Targeted covalent inhibition and the use of irreversible chemical probes are important strategies in chemical biology and drug discovery. To date, the availability and reactivity of cysteine residues amenable for covalent targeting have been evaluated by proteomic and computational tools. Herein, we present a toolbox of fragments containing a 3,5‐bis(trifluoromethyl)phenyl core that was equipped with chemically diverse electrophilic warheads showing a range of reactivities. We characterized the library members for their reactivity, aqueous stability and specificity for nucleophilic amino acids. By screening this library against a set of enzymes amenable for covalent inhibition, we showed that this approach experimentally characterized the accessibility and reactivity of targeted cysteines. Interesting covalent fragment hits were obtained for all investigated cysteine‐containing enzymes.

Graphical abstract

In this study, we utilized human DNA topoisomerase IIα as a model target to outline a dynophore-based approach to catalytic inhibitor design. Based on MD simulations of a known catalytic inhibitor and the native ATP ligand analog, AMP-PNP, we derived a joint dynophore model that supplements the static structure-based-pharmacophore information with a dynamic component. Subsequently, derived pharmacophore models were employed in a virtual screening campaign of a library of natural compounds. Experimental evaluation identified flavonoid compounds with promising topoisomerase IIα catalytic inhibition and binding studies confirmed interaction with the ATPase domain. We constructed a binding model through docking and extensively investigated it with molecular dynamics MD simulations, essential dynamics, and MM-GBSA free energy calculations, thus reconnecting the new results to the initial dynophore-based screening model. We not only demonstrate a new design strategy that incorporates a dynamic component of molecular recognition, but also highlight new derivates in the established flavonoid class of topoisomerase II inhibitors.

Cholesterol is essential for development, growth, and maintenance of organisms. Mutations in cholesterol biosynthetic genes are embryonic lethal and few polymorphisms have been so far associated with pathologies in humans. Previous analyses show that lanosterol 14α-demethylase (CYP51A1) from the late part of cholesterol biosynthesis has only a few missense mutations with low minor allele frequencies and low association with pathologies in humans. The aim of this study is to evaluate the role of amino acid changes in the natural missense mutations of the hCYP51A1 protein. We searched SNP databases for existing polymorphisms of CYP51A1 and evaluated their effect on protein function. We found rare variants causing detrimental missense mutations of CYP51A1. Some missense variants were also associated with a phenotype in humans. Two missense variants have been prepared for testing enzymatic activity in vitro but failed to produce a P450 spectrum. We performed molecular modeling of three selected missense variants to evaluate the effect of the amino acid substitution on potential interaction with its substrate and the obligatory redox partner POR. We show that two of the variants, R277L and especially D152G, have possibly lower binding potential toward obligatory redox partner POR. D152G and R431H have also potentially lower affinity toward the substrate lanosterol. We evaluated the potential effect of damaging variants also using data from other in vitro CYP51A1 mutants. In conclusion, we propose to include damaging CYP51A1 variants into personalized diagnostics to improve genetic counseling for certain rare disease phenotypes.