Deep Learning-Based Structure Modeling of the Treponema pallidum Proteome: Insights into Pathogenesis and Syphilis...

Treponema pallidum ssp. pallidum , the causative agent of syphilis, has a small proteome and encompasses numerous strains. Knowledge gaps remain in understanding the molecular mechanisms of pathogenesis of this bacterium, as well as the structure and function of the full complement of proteins encoded by T. pallidum . Here, an AI-based structure-to-function modeling workflow was used to investigate the complement of proteins encoded by T. pallidum . High-confidence structure models were generated for 976 T. pallidum proteins, covering 99% of the proteome. Analysis of the generated models using the protein structure comparison server DALI enabled high-confidence, structure-based functional annotation of 877 T. pallidum proteins, including 240 of the 323 proteins of unknown function encoded by this pathogen. Additionally, 63 putative pathogenesis related proteins (PPRPs) and seven treponemal proteins with previously uncharacterized similarity to outer membrane proteins (OMPs) from Gram-negative bacteria were identified. A workflow for B cell epitope (BCE) prediction identified 1133 surface-exposed, host-facing potential epitopes in known and predicted T. pallidum OMPs, of which 92 were prioritized based on bioinformatic analyses, biophysical properties, amino acid sequence conservation, and previous protein expression data. This work provides insight into T. pallidum pathogenesis through structure modeling-based functional annotation, including characterization of proteins of unknown function. This study also informs syphilis vaccine design by identifying new potential T. pallidum OMPs, as well as host-facing regions of T. pallidum OMPs that have conserved amino acid sequences in globally circulating strains.

This study presents the first AI-based global structure modeling-to-function analysis of the proteome of Treponema pallidum , the bacterium that causes syphilis. Structure-based functional predictions of previously uncharacterized proteins, including proteins potentially involved in virulence, provide novel insight into mechanisms of pathogenesis. The work also informs syphilis vaccine development by the identification and structural characterization of new candidate vaccine proteins in globally circulating strains of T. pallidum .