Rapid Chagas Disease Drug Target Discovery Using Directed Evolution in Drug-Sensitive Yeast

Abstract

Recent advances in cell-based, high-throughput phenotypic screening have identified new chemical compounds that are active against eukaryotic pathogens. A challenge to their future development lies in identifying these compounds' molecular targets and binding modes. In particular, subsequent structure-based chemical optimization and target-based screening require a detailed understanding of the binding event. Here, we use directed evolution and whole-genome sequencing of a drug-sensitive S. cerevisiae strain to identify the yeast ortholog of TcCyp51, lanosterol-14-alpha-demethylase (TcCyp51), as the target of MMV001239, a benzamide compound with activity against Trypanosoma cruzi, the etiological agent of Chagas disease. We show that parasites treated with MMV0001239 phenocopy parasites treated with another TcCyp51 inhibitor, posaconazole, accumulating both lanosterol and eburicol. Direct drug-protein binding of MMV0001239 was confirmed through spectrophotometric binding assays and X-ray crystallography, revealing a binding site shared with other antitrypanosomal compounds that target Cyp51. These studies provide a new probe chemotype for TcCyp51 inhibition.

Fig. 1. MMMV001239 activity against yeast and T. cruzi

A. MMV001239 dose-response curve for S. cerevisiae. Data points represent mean measurements taken from three independent 18 hour, 8-point dose response experiments. Error bars represent the standard error. The calculated average IC₅₀ concentration for the ABC₁₆-Monster strain (○) is 8.1 µM ± 1.2. For wild-type S. cerevisiae (●), no IC₅₀ value could be calculated because complete growth inhibition could not be achieved. B. MMV001239 activity against T. cruzi intracellular amastigotes. C2C12 myocytes were infected with T. cruzi trypomastigotes at a 15:1 parasite-to-host-cell ratio and treated with MMV001239 in varied concentrations. Compound activity was assessed after 72 h of treatment by determining the number of amastigotes per total host cells, normalizing to vehicle control and to positive control (uninfected cells).

Fig. 2. Characterization of MMV001239-resistant lineages obtained through directed evolution

A) After multiple rounds of selections, the IC₅₀ value of each mutant lineage was determined. Average IC₅₀ values were calculated from three independent experiments done in triplicate. B and C) Cells of the parental strain as well as the four resistant lineages were streaked out on a YPD plate (B) or a plate containing 10 µM MMV001239 (C) and incubated for 3 days at 30°C. ABC₁₆, parental strain (GM); R1, lineage 1; R2, lineage 2; R4, lineage 4; R11, lineage 11.

Fig. 3. ERG11 and the sterol biosynthesis pathway

A. Sequence similarity of CYP51 proteins (pfam00067) is shown through a profile Hidden Markov Model (pHMM). The highly conserved amino acid T318 that is mutated in lineage R2 is marked by an asterisk. B) CYP51 model (PDB ID 4K0F). Residues shown as pink and yellow solid surfaces (T318 and V154, respectively) underwent directed-evolution-induced changes. Parts of the protein have been removed or otherwise modified to facilitate visualization. C). S. cerevisiae ergosterol biosynthesis pathway and associated drug classes. Figure adapted and modified from.

Fig. 4. GC–MS sterol-profile analysis of intracellular T. cruzi amastigotes

C2C12 myoblast cultures infected with T. cruzi show parasite-specific lipids corresponding to chromatographic peaks labeled as the following: (a) cholesterol originating from host cells, m/z = 458, tR = 17.3–17.7 min; (b) ergosterol, m/z = 468, tR = 19.3–19.8 min; (c) ergosta-7,24-diene-3-β-ol (episterol), m/z = 470, tR = 21.3–21.8 min; (d) ergosta-8,24-diene-3-β-ol (fecosterol), m/z = 470, tR = 21.7–22 min; (e) lanosterol, m/z 498, tR = 22.7–23 min; (f) 4-methylepisterol, m/z = 484, tR = 24.4–24.9 min; (g) eburicol, m/z = 512, tR = 24.8–25.1 min. DMSO (vehicle) and benznidazole (a reference drug for Chagas disease) were used as negative controls. Posaconazole, a potent Cyp51 inhibitor, was used as a positive control. Inhibition of T. cruzi Cyp51 by posaconazole and MMV001239 is demonstrated by the accumulation of lanosterol (e) and eburicol (g), and the decline of fecosterol (d).

Fig. 5. Determining binding potency with absorbance difference spectra

A) To estimate the relative affinity of MMV001239 binding to TcCyp51, spectra were recorded across a range of compound concentrations. The MMV001239 spectra were compared to those of fluconazole and posaconazole, canonical weak and strong TcCyp51 inhibitors, respectively. B) Absorbance values at 390 were subtracted from those at 420 and plotted against inhibitor concentration. Relative potency can be determined from the maximum difference and the concentration at which it is reached.

Fig. 6. Small-molecule binding to ScERG11p

A) MMV001239 was docked into CYP51 from S. cerevisiae (PDB ID 4K0F). Residues that underwent directed-evolution-induced changes are shown as pink and yellow solid surfaces. Parts of the protein have been removed or otherwise modified to facilitate visualization. B) The MMV001239 docked pose into ScERG11p matched the low-resolution crystallographic density. The electron density (in green mesh) was visualized at an isovalue of 0.04. Most of the density was removed to facilitate visualization. C) The crystallographic pose of fluconazole, taken from the 4WMZ structure, shown superimposed on 4K0F for reference. D) The crystallographic pose of lanosterol, taken from 4LXJ, similarly superimposed on 4K0F. Like MMV001239, these two well-characterized inhibitors also contain a nitrogenous heterocycle that interacts with the iron atom at the center of the heme group.