Improving Effector Protein Prediction in Phytoplasmas Through Structural Analysis of Signal Peptide Cleavage

Phytoplasmas are highly destructive phloem-restricted pathogens, acting as obligate plant parasites transmitted by sap-feeding insect vectors. They infect over 1000 plant species, including critical crops, leading to severe agricultural losses globally. Evolving from Gram-positive bacteria, phytoplasmas underwent extreme genome reduction, resulting in some of the smallest known bacterial genomes. Despite their minimal genetic content, they effectively manipulate host and vector cellular processes through effector proteins. These virulence factors are thought to be secreted via signal peptide (SP)-dependent cleavage by signal peptidase I (SPase I). Since phytoplasmas remain unculturable in vitro, identification of these effectors relies heavily on in silico SP and cleavage site (CS) prediction methods, which often produce unreliable results, leading to misidentified effector candidates. In this study, to improve prediction accuracy, we applied a structural modeling approach that complements sequence-based methods by assessing SPs through 3D modeling of SP-SPase I hetero-oligomer complexes. We analyzed reference virulence proteins (RVPs) with experimentally validated SPs, identifying potential errors in their annotated CSs. Through structural characterization, we classified phytoplasma SPase Is as eukaryotic ER-type-a rare trait in bacteria-and modeled SP-SPase I hetero-oligomers using ColabFold. Our findings reveal structural determinants governing cleavable SP binding to SPase I, enabling more accurate SP/CS predictions. This work underscores the unique molecular adaptations of phytoplasmas and provides insights for targeting their effector secretion mechanisms in disease control.