Computational structural genomics unravels common folds and predicted functions in the secretome of fungal phytopathogen Magnaporthe oryzae

Abstract

Magnaporthe oryzae relies on a diverse collection of secreted effector proteins to reprogram the host metabolic and immune responses for the pathogens benefit. Characterization of the effectors is thus critical for understanding the biology and host infection mechanisms of this phytopathogen. In rapid, divergent effector evolution, structural information has the potential to illuminate the unknown aspects of effectors that sequence analyses alone cannot reveal. It has recently become feasible to reliably predict the protein structures without depending on homologous templates. In this study, we tested structure modeling on 1854 secreted proteins from M. oryzae and evaluated success and obstacles involved in effector structure prediction. With sensitive homology search and structure-based clustering, we defined both distantly related homologous groups and structurally related analogous groups. With this dataset, we propose sequence-unrelated, structurally similar effectors are a common theme in M. oryzae and possibly in other phytopathogens. We incorporated the predicted models for structure-based annotations, molecular docking and evolutionary analyses to demonstrate how the predicted structures can deepen our understanding of effector biology. We also provide new experimentally testable structure-derived hypotheses of effector functions. Collectively, we propose that computational structural genomic approaches can now be an integral part of studying effector biology and provide valuable resources that were inaccessible before the advent of reliable, machine learning-based structure prediction.