Pyrethroids in an AlphaFold2 Model of the Insect Sodium Channel.

Abstract

Simple SummaryPyrethroids are a major class of insecticides for controlling arthropod pests and human disease transmitting vectors. Resistance to pyrethroid insecticides presents a serious obstacle to effective pest control. One major mechanism of pyrethroid resistance is caused by mutations in the voltage-gated sodium channel, the target of pyrethroid toxic action. Previous mutational analyses, coupled with homology modeling, predicted two pyrethroid receptor sites on insect sodium channels. Mutations that are located in or close to these receptor sites likely confer resistance by directly reducing pyrethroid binding affinity. However, many mutations appear to be far from the receptor sites, and how these mutations confer pyrethroid resistance remains unknown. In this study, we employed the AlphaFold2 neural network to generate new models of mosquito and cockroach sodium channels. We then used computational methods to dock pyrethroids in the open and closed states of sodium channels to understand the atomic mechanisms of interactions between pyrethroids and sodium channels. Our analysis suggests that mutations residing beyond the receptor sites alter the action of pyrethroids by allosterically affecting the geometry of the receptor sites. The information gained from this study could be valuable for a structure-based design of novel pyrethroids.Pyrethroid insecticides stabilize the open state of insect sodium channels. Previous mutational, electrophysiological, and computational analyses led to the development of homology models predicting two pyrethroid receptor sites, PyR1 and PyR2. Many of the naturally occurring sodium channel mutations, which confer knockdown resistance (kdr) to pyrethroids, are located within or close to these receptor sites, indicating that these mutations impair pyrethroid binding. However, the mechanism of the state-dependent action of pyrethroids and the mechanisms by which kdr mutations beyond the receptor sites confer resistance remain unclear. Recent advances in protein structure prediction using the AlphaFold2 (AF2) neural network allowed us to generate a new model of the mosquito sodium channel AaNav1-1, with the activated voltage-sensing domains (VSMs) and the presumably inactivated pore domain (PM). We further employed Monte Carlo energy minimizations to open PM and deactivate VSM-I and VSM-II to generate additional models. The docking of a Type II pyrethroid deltamethrin in the models predicted its interactions with many known pyrethroid-sensing residues in the PyR1 and PyR2 sites and revealed ligand-channel interactions that stabilized the open PM and activated VSMs. Our study confirms the predicted two pyrethroid receptor sites, explains the state-dependent action of pyrethroids, and proposes the mechanisms of the allosteric effects of various kdr mutations on pyrethroid action. The AF2-based models may assist in the structure-based design of new insecticides.