Improving Olfactory Receptor Structure Modeling via Hybrid Methods

Understanding the structure of Olfactory Receptors (ORs) is pivotal in deciphering the molecular complexities of smell, a sense crucial for health, and survival, and holding immense therapeutic potential. However, the scarcity of detailed experimental data on ORs has hindered progress, demanding inventive approaches. This study explores various structure prediction methods tailored to ORs based on their phylogenetic and structural characteristics, using OR51E2 as a reference. We employed a hybrid homology modeling approach, utilizing AlphaFold structures, yielding models with lower RMSD (1.019 {degrees}A across pruned/significant pairs and 2.33 {degrees}A over-all) and enhanced structural metrics compared to standalone AlphaFold (RMSD - 2.5 {degrees}A) predictions. Our pipeline successfully replicated experimental findings for OR51E2 and was applied to homologous ORs: OR51E1, OR51D1, and OR51G2. Various tools were also used to predict potential binding sites for each receptor. Molecular dynamics simulations validated the stability of these OR models in a lipid bilayer environment, with biophysical analyses revealing that AlphaFold models exhibit relatively less ideal behavior compared to the Hybrid Model. Our study presents a targeted approach to investigate and generate optimum OR structures for further conformational analyses.