To investigate the role of sialic acid metabolism in bacterial vaginosis, we explored both the genetic and phenotypic diversity among clinical isolates of each bacterial species: Gardnerella spp, Prevotella bivia and Atopobium vaginae. We have collected vaginal clinical samples, created an isolate collection of 88 Gardnerella spp., 65 P. bivia and seven A. vaginae, and developed a growth protocol on solid surfaces. This growth assay allows both reliable monoculture growth of all our isolates as well as community experiments with co-cultures of isolates, so that we can test community-level interactions.
We first assessed genotypic and phenotypic diversity among our Gardnerella spp. and P. bivia isolates. We sequenced a conserved gene to evaluate genetic relationships among isolates, and we targeted three different sialidase genes, related to sialic acid metabolism. We further performed biochemical assays to quantify bacterial sialidase enzyme activity. For A. vaginae, we could not find any sialidase genes, and only one isolate displayed a very low sialidase activity. For P. bivia, the genetic diversity was relatively low and all isolates possessed a putative sialidase gene, but the actually expressed sialidase activity varied greatly between isolates. For Gardnerella spp. we found great genotypic variation in terms of whole-genome relatedness and of patterns of presence/absence of sialidase genes. When crossing the genetic data with the actual sialidase activity production, we conclude that the presence of the nanH3 sialidase gene was predictive of expressed sialidase activity in Gardnerella spp, and we observed a very large sialidase phenotypic diversity among isolates, both qualitatively and quantitatively. No pattern of genotypic nor phenotypic diversity, however, was linked to a clinical diagnosis of BV.
We then evaluated the sialidase production patterns at the community-level. For this we performed bacterial co-cultures, so that we could evaluate the direction and strength of intra- and interspecific interactions. The experimental system we implemented enabled us to measure expressed sialidase activity for 82 co-cultures formed by a selection of Gardnerella spp., P. bivia and A. vaginae isolates, in a spatially structured environment. All isolates were categorized based on their expressed sialidase activity in monoculture (strong-, low-, or non-producer), and we assessed intra and inter-specific interactions through combinations of two or three isolates of different taxa and sialidase categories. Our results support the hypothesis that both intra- and interspecific interactions likely play a key role in sialic acid-related bacterial functions. Indeed, sialidase activity in cocultures can strongly diverge from expectations based on monoculture data with numerous cases suggesting the presence of synergistic interactions amongst bacterial vaginosis-associated bacteria. Synergistic interactions where also found for interspecific co-cultures and were strongest in co-cultures of Gardnerella spp. and A. vaginae isolates with co-cultures expressing substantially higher sialidase activity levels than the monocultures of both partner isolates added together.
The first scientific article of this project (corresponding to WP1 and 2) is currently in the final phase of preparation and another one will follow (for WP4). The results have further been presented at the institute seminars, to our international scientific collaborators (mainly in South Africa) and to the medical collaborators at the hospital, and at the international conference of the French Society of Microbiology.
