An in silico pipeline for enzyme-substrate modelling using arthropod P450s

The introduction of three-dimensional protein models generated using artificial intelligence (AI) to the spheres of research that rely on predicted protein structures and their interactions, such as drug development in human medicine, has been transformative. AI generated models have become an important element in providing putative functional validation of many proteins. Although important for the study of the human proteome, AI can also be applied to the study of proteins from organisms that are less well represented in the PDB database. Thanks to sequencing efforts, arthropods possess an extensive collection of protein sequences, however, these often lack functional validation or experimental structures. The plethora of online tools has made in silico structural biology more accessible than before; however, no specific guide exists allowing for effective use of these tools for scientists without an extensive background in structural biology. Here, we provide a step-by-step guide for the successful generation and interpretation of in silico cytochrome P450 models and small molecule interactions. We cover three specific examples of experimentally validated cytochrome P450s involved in nicotine and neonicotinoid metabolization in both pest and beneficial insects: CYP9Q3 from Apis mellifera, CYP6CY3 from Myzus persicae and two variants of CYP6CM1 from Bemisia tabaci. By using only publicly available tools, we provide an in silico explanation for the observed biochemical results, showcasing the pipeline's utility in augmenting laboratory-based experiments.