We isolated a diverse set of novel and clinically significant anaerobes from the human vagina using conventional approaches with systematic molecular identification. Several previously "uncultivated" bacteria are amenable to conventional cultivation.

A major challenge for microbiome studies is maintaining an even and accurate DNA extraction in the presence of samples with a wide range of bacterial content. Here we compare five DNA extraction methods using replicate stool samples that were diluted to create high and low biomass samples. Our results indicate greater variation in microbiome composition between high and low biomass samples than variation between methods. Many of the extraction methods had reduced yield from low biomass samples; however, our adapted plate column-based extraction method was evenly efficient and captured the largest number of high-quality reads. Based on these results, we have identified a DNA extraction method that ensures adequate yield in metagenomic microbiome studies that have samples with a broad range of bacterial content.

Bacterial vaginosis (BV) is a vaginal dysbiotic condition linked to negative gynecological and reproductive sequelae. Flagellated bacteria have been identified in women with BV, including Mobiluncus spp. and BV-associated bacterium-1 (BVAB1), an uncultivated, putatively flagellated species. The host response to flagellin mediated through toll-like receptor-5 (TLR5) has not been explored in BV. Using independent discovery and validation cohorts, we examined the hypothesis that TLR5 deficiency–defined by a dominant negative stop codon polymorphism rs5744168–is associated with an increased risk for BV and increased colonization with flagellated bacteria associated with BV (BVAB1, Mobiluncus curtisii, Mobiluncus mulieris). TLR5 deficiency was not associated with BV status and TLR5 deficient women had decreased colonization with BVAB1 in both cohorts. We stimulated HEK-hTLR5-overexpressing, NF-kB reporter cells with whole, heat-killed M. mulieris or M. curtisii, and with partially purified flagellin from these species; as BVAB1 is uncultivated, we used cervicovaginal lavage (CVL) supernatant from women colonized with BVAB1 for stimulation. While heat-killed M. mulieris and CVL from women colonized with BVAB1 stimulate a TLR5-mediated response, heat-killed M. curtisii did not. In contrast, partially purified flagellin from both Mobiluncus species stimulated a TLR5-mediated response in vitro. We observed no correlation between vaginal IL-8 and flagellated BVAB concentrations among TLR5 sufficient women. Inter-species variation in accessibility of flagellin recognition domains may be responsible for these observations as reflected in the potentially novel flagellin products encoded by Mobiluncus species versus those encoded by BVAB1.