Discovery of Multitarget Inhibitors against Insect Chitinolytic Enzymes via Machine Learning-Based Virtual Screening

Authors：Ding, Yi; Chen, Sizhe; Liu, Huan; Liu, Tian^*; Yang, Qing

Source：Journal of Agricultural and Food Chemistry, 2023, 71(23): 8769-8777.

DOI：10.1021/acs.jafc.3c00633

• machine learning
• virtual screening
• chitinase
• insecticide
• natural product
• inhibitor

Multitarget inhibitors of insect chitinolytic enzymesare promisingsources of green insecticides. Machine learning (ML) is an emergingvirtual screening method that can accelerate drug discovery and reducecosts. Taking advantage of the data from our previous high-throughputscreening work, we established a strategy integrating ML and moleculardocking to screen a large natural product library (17 600 compounds)to identify novel multitarget inhibitors of four chitinolytic enzymesfrom the insect Ostrinia furnacalis (OfChtI, OfChtII, OfChi-h, and OfHex1). 3,5-Di-O-caffeoylquinicacid and gamma-mangostin were identified as inhibitors of all ofthese enzymes with K (i) values at the mu Mlevel. Moreover, they showed significant biological activities againstlepidopteran pests. Transcriptomic analysis of compound-treated insectssuggested the physiological relationship between these compounds andchitinolytic enzymes. This study highlights the potential of ML forinsecticide discovery and provides novel and low-cost scaffolds ofmultitarget inhibitors against insect chitinolytic enzymes as potentialpesticide leads.