Cell wall proteins with sialidase activity are involved in carbohydrate assimilation, adhesion to mucosal surfaces, and biofilm formation. Gardnerella spp. inhabit the human vaginal microbiome and encode up to three sialidase enzymes, two of which are suspected to be cell wall associated. Here, we demonstrate that the gene encoding extracellular sialidase NanH3 is found almost exclusively in Gardnerella piotii and the closely related species Gardnerella genome sp. 3, and its presence correlates with a sialidase-positive phenotype in a collection of 112 Gardnerella isolates. The nanH3 gene sequence includes a homopolymeric repeat of cytosines that varies in length within cell populations, indicating that this gene is subject to slipped-strand mispairing, a mechanism of phase variation in bacteria. Variation in the length of the homopolymer sequence results in production of either the full-length sialidase protein or truncated peptides lacking the sialidase domain due to introduction of reading-frame shifts and premature stop codons. Phase variation in NanH3 may be involved in immune evasion or modulation of adhesion to host epithelial cells and formation of biofilms characteristic of the vaginal dysbiosis known as bacterial vaginosis.

Objective: Microbiological and virulence characterization of 2 moderately hemolytic Brachyspira strains. Materials and methods: Clinical isolates were obtained from diarrheic (3603-F2) and healthy (G79) pigs. Phenotypic characterization included assessment of hemolytic activity on blood agar and biochemical profiling. Genotyping was performed by sequencing the nicotinamide adenine dinucleotide oxidase (nox) gene, whole genome sequencing, and comparison to relevant Brachyspira. Pig inoculation included 4 treatment groups in 2 challenge experiments: negative control (sterile broth media; n = 12), positive control (Brachyspira hampsonii genomovar 2 strain 30446; n = 18), and 3603-F2 (n = 12) or G79 (n = 12). Fecal scoring and rectal swabbing for culture were performed daily. Animals were euthanized following onset of mucohemorrhagic diarrhea or between 21 and 28 days post inoculation (dpi). Gross and microscopic pathology were assessed. Terminal colon samples were used to characterize post-infection mucosal ion secretion. Results: Both strains were moderately hemolytic. Whole genome and nox sequencing identified 3603-F2 as Brachyspira murdochii and G79 as a novel strain. Both challenge trials revealed intestinal colonization, but no mucohemorrhagic diarrhea. Sporadic watery diarrhea was induced by 3603-F2 associated with a pattern of microscopic lesions similar to pigs with swine dysentery (positive controls). No diarrhea was observed in G79 inoculated pigs, but microscopic lesions were more severe than in controls. Both strains induced greater colonic anion secretory potential than negative controls 21 dpi. Implications: Allegedly avirulent Brachyspira species most closely related to B murdochii can be associated with subclinical colitis and may be a concern for grow-finish pigs.

Gardnerella vaginalis was first described in 1953, and subsequently identified as the causative agent of a cluster of vaginal symptoms currently known as vaginosis. Research has so far failed to confirm whether and by which mechanism G. vaginalis initiates vaginosis, with, consequently, poor diagnostics and treatment outcomes. Recent molecular analyses of protein-coding genes demonstrate that the taxon G. vaginalis consists of at least four distinct species. This development may represent a critical turning point in clarifying ecological interactions and virulence factors contributing to symptoms and/or sequelae of vaginosis.

Increased abundance of Gardnerella vaginalis and sialidase activity in vaginal fluid is associated with bacterial vaginosis (BV), a common but poorly understood clinical entity associated with poor reproductive health outcomes. Since most women are colonized with G. vaginalis, its status as a normal member of the vaginal microbiota or pathogen causing BV remains controversial, and numerous classification schemes have been described. Since 2005, sequencing of the chaperonin-60 universal target (cpn60 UT) has distinguished four subgroups in isolate collections, clone libraries and deep sequencing datasets. To clarify potential clinical and diagnostic significance of cpn60 subgroups, we undertook phenotypic and molecular characterization of 112 G. vaginalis isolates from three continents. A total of 36 subgroup A, 33 B, 35 C and 8 D isolates were identified through phylogenetic analysis of cpn60 sequences as corresponding to four "clades" identified in a recently published study, based on sequencing 473 genes across 17 isolates. cpn60 subgroups were compared with other previously described molecular methods for classification of Gardnerella subgroups, including amplified ribosomal DNA restriction analysis (ARDRA) and real-time PCR assays designed to quantify subgroups in vaginal samples. Although two ARDRA patterns were observed in isolates, each was observed in three cpn60 subgroups (A/B/D and B/C/D). Real-time PCR assays corroborated cpn60 subgroups overall, but 13 isolates from subgroups A, B and D were negative in all assays. A putative sialidase gene was detected in all subgroup B, C and D isolates, but only in a single subgroup A isolate. In contrast, sialidase activity was observed in all subgroup B isolates, 3 (9%) subgroup C isolates and no subgroup A or D isolates. These observations suggest distinct roles for G. vaginalis subgroups in BV pathogenesis. We conclude that cpn60 UT sequencing is a robust approach for defining G. vaginalis subgroups within the vaginal microbiome.