Trypanosoma cruzi CYP51 inhibitor derived from a Mycobacterium tuberculosis screen hit

Abstract

Background: The two front-line drugs for chronic Trypanosoma cruzi infections are limited by adverse side-effects and declining efficacy. One potential new target for Chagas' disease chemotherapy is sterol 14alpha-demethylase (CYP51), a cytochrome P450 enzyme involved in biosynthesis of membrane sterols.

Methodology/principal finding: In a screening effort targeting Mycobacterium tuberculosis CYP51 (CYP51(Mt)), we previously identified the N-[4-pyridyl]-formamide moiety as a building block capable of delivering a variety of chemotypes into the CYP51 active site. In that work, the binding modes of several second generation compounds carrying this scaffold were determined by high-resolution co-crystal structures with CYP51(Mt). Subsequent assays against the CYP51 orthologue in T. cruzi, CYP51(Tc), demonstrated that two of the compounds tested in the earlier effort bound tightly to this enzyme. Both were tested in vitro for inhibitory effects against T. cruzi and the related protozoan parasite Trypanosoma brucei, the causative agent of African sleeping sickness. One of the compounds had potent, selective anti-T. cruzi activity in infected mouse macrophages. Cure of treated host cells was confirmed by prolonged incubation in the absence of the inhibiting compound. Discrimination between T. cruzi and T. brucei CYP51 by the inhibitor was largely based on the variability (phenylalanine versus isoleucine) of a single residue at a critical position in the active site.

Conclusions/significance: CYP51(Mt)-based crystal structure analysis revealed that the functional groups of the two tightly bound compounds are likely to occupy different spaces in the CYP51 active site, suggesting the possibility of combining the beneficial features of both inhibitors in a third generation of compounds to achieve more potent and selective inhibition of CYP51(Tc).

Figure 1. Spectral characteristics of CYP51_Tc (A) and CYP51_Tb (B).

Main panel shows the absolute protein spectrum, while insert shows CO-bound reduced difference spectrum.

Figure 2. Chemical structures and binding affinities of compounds.

¹Compounds are identified by the numbers in the ChemDiv, Inc., product library catalog.

Figure 3. Overall view of compound binding in the active site.

Compounds 11 (A), 9 (B), and 8 (C) (highlighted in pink) bound in the active site of CYP51_Mt are shown looking in from the active site opening. For clarity only one conformation of each compound is shown. Protein is represented by the semitransparent accessible surface (gray). The ordered BC-loop obstructs the view in the CYP51_Mt-8 complex in (C). The invariable elements of the CYP51 active site, Y76, H259 (yellow), and heme (green), are in a stick mode. Water molecules are shown as red spheres. Oxygen atoms are red, nitrogen blue, sulfur yellow. Images were generated using the VMD program .

Figure 4. Stereo view of compounds in the active site.

Compounds 11 (A), 9 (B), and 8 (C) are shown surrounded by the CYP51 active site residues. The fragments of the electron density 2Fo-Fc map (gray mesh) are cut at 1.2 σ. Different conformers in (A) and (B) are highlighted in pink and cyan. Images were generated using PYMOL .

Figure 5. Spectroscopic binding of compounds.

(A) Type II spectral responses of CYP51_Tc to increasing concentrations of compound 10. The concentration dependence of compound 10, fluconazole (B), and compound 8 (C) binding were deduced from the difference absorption changes obtained from the titration of CYP51_Tc with increasing concentrations of the inhibitor. The concentration dependence of fluconazole (D) was deduced from the difference absorption changes obtained from the titration of CYP51_Tb.

Figure 6. Inhibition of T. cruzi intracellular amastigotes by compound 10.

T. cruzi intracellular multiplication was evaluated at 52 hr of incubation at several concentrations of the inhibitor by determining the number of parasites/cell. Intracellular parasites were counted per one hundred cells to estimate a mean number of parasites per cell. Approximation of concentration dependence of mean P/cell±SD data with a smooth curve highlights the 50% drop in parasite count at ∼1 nM compound 10. SD did not exceed 14% of the mean.