Utilizing Whole <i>Fusobacterium</i> Genomes To Identify, Correct, and Characterize Potential Virulence Protein Families

Abstract

<i>Fusobacterium</i> spp. are Gram-negative, anaerobic, opportunistic pathogens involved in multiple diseases, including a link between the oral pathogen <i>Fusobacterium nucleatum</i> and the progression and severity of colorectal cancer. The identification and characterization of virulence factors in the genus <i>Fusobacterium</i> has been greatly hindered by a lack of properly assembled and annotated genomes. Using newly completed genomes from nine strains and seven species of <i>Fusobacterium</i>, we report the identification and corrected annotation of verified and potential virulence factors from the type 5 secreted autotransporter, FadA, and MORN2 protein families, with a focus on the genetically tractable strain <i>F. nucleatum</i> subsp. <i>nucleatum</i> ATCC 23726 and type strain <i>F. nucleatum</i> subsp. <i>nucleatum</i> ATCC 25586. Within the autotransporters, we used sequence similarity networks to identify protein subsets and show a clear differentiation between the prediction of outer membrane adhesins, serine proteases, and proteins with unknown function. These data have identified unique subsets of type 5a autotransporters, which are key proteins associated with virulence in <i>F. nucleatum</i> However, we coupled our bioinformatic data with bacterial binding assays to show that a predicted weakly invasive strain of <i>F. necrophorum</i> that lacks a Fap2 autotransporter adhesin strongly binds human colonocytes. These analyses confirm a gap in our understanding of how autotransporters, MORN2 domain proteins, and FadA adhesins contribute to host interactions and invasion. In summary, we identify candidate virulence genes in <i>Fusobacterium</i>, and caution that experimental validation of host-microbe interactions should complement bioinformatic predictions to increase our understanding of virulence protein contributions in <i>Fusobacterium</i> infections and disease.<b>IMPORTANCE</b><i>Fusobacterium</i> spp. are emerging pathogens that contribute to mammalian and human diseases, including colorectal cancer. Despite a validated connection with disease, few proteins have been characterized that define a direct molecular mechanism for <i>Fusobacterium</i> pathogenesis. We report a comprehensive examination of virulence-associated protein families in multiple <i>Fusobacterium</i> species and show that complete genomes facilitate the correction and identification of multiple, large type 5a secreted autotransporter genes in previously misannotated or fragmented genomes. In addition, we use protein sequence similarity networks and human cell interaction experiments to show that previously predicted noninvasive strains can indeed bind to and potentially invade human cells and that this could be due to the expansion of specific virulence proteins that drive <i>Fusobacterium</i> infections and disease.