ABSTRACT
The integration of artificial intelligence (AI) into smart agriculture boosts production and management efficiency, facilitating sustainable agricultural development. In intensive agricultural management, adopting eco-friendly and effective pesticides is crucial to promote green agricultural practices. However, exploring new insecticides species is a difficult and time-consuming task that involves significant risks. Enhancing compound druggability in the lead discovery phase could considerably shorten the discovery cycle, accelerating insecticides research and development. The Insecticide Activity Prediction (IAPred) model, a novel classic artificial intelligence-based method for evaluating the potential insecticidal activity of unknown functional compounds, is introduced in this study. The IAPred model utilized 27 insecticide-likeness features from PaDEL descriptors and employed an ensemble of Support Vector Machine (SVM) and Random Forest (RF) algorithms using the hard-vote mechanism, achieving an accuracy rate of 86 %. Notably, the IAPred model outperforms current models by accurately predicting the efficacy of novel insecticides such as nicofluprole, overcoming the limitations inherent in existing insecticide structures. Our research presents a practical approach for discovering and optimizing novel insecticide lead compounds quickly and efficiently.
ABSTRACT
Both insect ecdysone receptors and ultraspiracle belong to the nuclear receptor family. They form a nanoscale self-assembling complex with ecdysteroids in cells, transit into the nucleus, bind with genes to initiate transcription, and perform specific biological functions to regulate the molting, metamorphosis, and growth processes of insects. Therefore, this complex is an important target for the development of eco-friendly insecticides. The diamondback moth (Plutella xylostella) is a devastating pest of cruciferous vegetable crops, wreaking havoc worldwide and causing severe economic losses, and this pest has developed resistance to most chemical insecticides. In this study, highly pure EcR and USP functional domains were obtained by constructing a prokaryotic expression system for the diamondback moth EcR and USP functional domain genes, and the differences between EcR and USP binding domain monomers and dimers were analyzed using transmission electron microscopy and zeta potential. Radioisotope experiments further confirmed that the binding affinity of PonA to the EcR/USP dimer was enhanced approximately 20-fold compared with the binding affinity to the PxGST-EcR monomer. The differences between PonA and tebufenozide in binding with EcR/USP were examined. Molecular simulations showed that the hydrogen bonding network formed by Glu307 and Arg382 on the EcR/USP dimer was a key factor in the affinity enhancement. This study provides a rapid and sensitive method for screening ecdysone agonists for ecdysone receptor studies in vitro.
