Abstract
Insect odorant receptors (iORs) are seven-transmembrane-domain (7TM) ion channels crucial for insect survival, playing key roles in insect behaviour such as foraging, pollination, social interaction, and prey recognition. Unlike G-protein coupled receptors (GPCRs), iORs are "inverted" in the membrane and function independently of accessory G proteins. iORs form complexes with a highly conserved co-receptor (Orco), allowing calcium ions to enter the cell upon ligand binding and are promising targets for pest control strategies. However, due to their structural complexity, experimental determination of iOR structures remains a challenge. This scarcity of experimental data underscores the urgent need for reliable computational modeling approaches to understand iOR structure and function. With the rise of the deep learning (DL) approach AlphaFold, currently in its third version, we explored whether detailed template-based modeling (TBM), using the few available experimental insect OR structures, is still required, instead of a quick AI-generated AlphaFold 3 (AF3) model. For six OR sequences from three insect orders, we compared TBM models with standard AF3 models, and AF3 models with lipid molecules, creating an artificial membrane. Our evaluation of functional mutagenesis data supports TBM models for iORs rather than AF3 models, for ligand discovery for insect pest-specific control using structure-based approaches.