Virtual Screening and Bioassay of Novel Protoporphyrinogen Oxidase and p-Hydroxyphenylpyruvate Dioxygenase Dual-Target Inhibitors

Abstract

Novel herbicide development is a challenge for weed control. Protoporphyrinogen oxidase (PPO) and p-hydroxyphenylpyruvate dioxygenase (HPPD) are two key enzymes involved in plant photosynthesis. The multi virtual screening protocol was adopted to design a common skeleton based on the two target enzymes, and fragment growth of the skeleton was performed. The constructed compounds were searched for structural similarity, and the accuracy of the selected compounds was further verified using the Bayesian model. Finally, eight compounds were obtained, and the binding mode with the target was studied deeply. The obtained compounds interact with the key residues of HPPD and PPO proteins similarly to commercial herbicides, and the stability of binding with proteins is also good. The activity of the screening results was determined by an enzyme activity test in vitro. The herbicidal effect of the compound was studied by phenotypic experiment. The final results showed that Z-4 and Z-7 have the potential to become new dual-target herbicides.

Keywords: Bayesian model; HPPD; PPO; dual-target herbicide; skeleton-based drug discovery.

Figure 1

The workflow of multi virtual screening.

Figure 2

Ligand interaction with the amino residues of HPPD. (A): Compound 5; (B): Compound 8; (C): Compound 10; (D): Compound 11; and (E): Compound 25.

Figure 3

Ligand interaction with the amino residues of PPO. (A): Compound 5; (B): Compound 8; (C): Compound 10; (D): Compound 11; and (E): Compound 25.

Figure 4

NBC-HPPD model. (A): PCA; (B): confusion matrix of the NBC model to distinguish the training set; (C): confusion matrix of NBC model to distinguish the test sets; and (D): ROC curve of the NBC model.

Figure 5

NBC-PPO model. (A): PCA; (B): confusion matrix of the NBC model to distinguish the training set; (C): confusion matrix of the NBC model to distinguish the test sets; and (D): ROC curve of the NBC model.

Figure 6

Molecular docking results of the obtained ligands in the active pocket of HPPD. (A): Z-1; (B): Z-2; (C): Z-3; (D): Z-4; (E): Z-5; (F): Z-6; (G): Z-7; and (H): Z-8.

Figure 7

Molecular docking results of the obtained ligands in the active pocket of PPO. (A): Z-1; (B): Z-2; (C): Z-3; (D): Z-4; (E): Z-5; (F): Z-6; (G): Z-7; and (H): Z-8.

Figure 8

MD simulation results of HPPD and ligands. (A) HPPD-RMSD of skeleton Cα atom. (B) RMSD of the heavy atom of the ligands; and (C) RMSD of the protein active pocket with 5 Å residues around the ligands.

Figure 9

MD simulation results of PPO and ligands. (A) PPO-RMSD of skeleton Cα atom. (B) RMSD of the heavy atom of the ligands; and (C) RMSD of the protein active pocket with 5 Å residues around the ligands.

Figure 10

Fluctuations of residues involved in HPPD interactions with ligands. (A) HPPD-RMSF of Cα atoms; (B) RMSF of the heavy atom of the ligands; and (C) RMSF of the protein active pocket with 5 Å residues around the ligands.

Figure 11

Fluctuations of residues involved in PPO interactions with ligands. (A) PPO-RMSF of Cα atoms; (B) RMSF of the heavy atom of the ligands; and (C) RMSF of the protein active pocket with 5 Å residues around the ligands.

Figure 12

Residue contribution degree (A) HPPD: Z-4; (B) HPPD: Z-7; (C) PPO: Z-4; and (D) PPO: Z-7.

Figure 13

Herbicidal activities of compounds Z-4, Z-7, mesotrione, and oxflurane (post-emergence, 150 g ai/ha, treated after 7 days). (A) EC. (B) AT. (C) LP.

Figure 14

Reverse synthesis routes. (A) Z-4 inverse synthetic route; (B) Z-7 inverse synthetic route.

Figure 15

Common skeleton selection.