Virtual screening of natural products against 5-enolpyruvylshikimate-3-phosphate synthase using the Anagreen herbicide-like natural compound library

Abstract

The shikimate pathway enzyme 5-enolpyruvylshikimate-3-phosphate synthase (EPSPS) catalyzes a reaction involved in the production of amino acids essential for plant growth and survival. EPSPS is the main target of glyphosate, a broad-spectrum herbicide that acts as a competitive inhibitor concerning phosphoenolpyruvate (PEP), which is the natural substrate of EPSPS. In the present study, we introduce a natural compound library, named Anagreen, which is a compendium of herbicide-like compounds obtained from different natural product databases. Herein, we combined the structure- and ligand-based virtual screening strategies to explore Anagreen against EPSPS using the structure of glyphosate complexed with a T102I/P106S mutant of EPSPS from Eleusine indica (EiEPSPS) as a starting point. First, ligand-based pharmacophore screening was performed to select compounds with a similar pharmacophore to glyphosate. Then, structure-based pharmacophore modeling was applied to build a model which represents the molecular features of glyphosate. Then, consensus docking was performed to rank the best poses of the natural compounds against the PEP binding site, and then molecular dynamics simulations were performed to analyze the stability of EPSPS complexed with the selected ligands. Finally, we have investigated the binding affinity of the complexes using free energy calculations. The selected hit compound, namely AG332841, showed a stable conformation and binding affinity to the EPSPS structure and showed no structural similarity to the already known weed EPSPS inhibitors. Our computational study aims to clarify the inhibition of the mutant EiEPSPS, which is resistant to glyphosate, and identify new potential herbicides from natural products.

Fig. 1. Per-residue energy decomposition analysis of main residues with energy decomposition value ≤ or ≥0.5 kcal mol⁻¹ of (A) glyphosate and (B) AG332841 complexed with EiEPSPS_wt and EiEPSPS_mut structures.

Fig. 2. RMSD plots of the heavy atoms in the EiEPSPS_wt– and EiEPSPS_mut–ligand complexes (orange) and the ligands (maroon) were analyzed over 150 ns of MD simulation. (A) Complex of EiEPSPS_wt and AG332841; (B) complex of EiEPSPS_mut and AG332841; (C) complex of EiEPSPS_wt and AG351308; (D) complex of EiEPSPS_mut and AG351308.

Fig. 3. (A) Regions of EiEPSPS with fluctuation values ≥3.0 Å. RMSF plot, in Å, of the backbones atoms in the (B) EiEPSPS_wt– and (C) EiEPSPS_mut–ligand complexes analyzed over the last 50 ns of MD simulation. Panels (D) and (E) show the fluctuations of the main binding pocket residues in EiEPSPS_wt– and EiEPSPS_mut–ligand complexes, respectively.

Fig. 4. The 2D diagram of the (A) EiEPSPS_wt and (B) EiEPSPS_mut complexed with AG332841 shows the average distances of the main intermolecular interactions with occupancy value ≥50% obtained over the MD simulation.

Fig. 5. The electrostatic potential maps of EiEPSPS_wt and EiEPSPS_mut enzymes and AG332841 and AG351308 ligands.

Fig. 6. Computational workflow applied in the virtual screening strategy.

Fig. 7. Construction of the Anagreen herbicide-like natural compound library.