Genomes of Gardnerella Strains Reveal an Abundance of Prophages within the Bladder Microbiome

Abstract

Bacterial surveys of the vaginal and bladder human microbiota have revealed an abundance of many similar bacterial taxa. As the bladder was once thought to be sterile, the complex interactions between microbes within the bladder have yet to be characterized. To initiate this process, we have begun sequencing isolates, including the clinically relevant genus Gardnerella. Herein, we present the genomic sequences of four Gardnerella strains isolated from the bladders of women with symptoms of urgency urinary incontinence; these are the first Gardnerella genomes produced from this niche. Congruent to genomic characterization of Gardnerella isolates from the reproductive tract, isolates from the bladder reveal a large pangenome, as well as evidence of high frequency horizontal gene transfer. Prophage gene sequences were found to be abundant amongst the strains isolated from the bladder, as well as amongst publicly available Gardnerella genomes from the vagina and endometrium, motivating an in depth examination of these sequences. Amongst the 39 Gardnerella strains examined here, there were more than 400 annotated prophage gene sequences that we could cluster into 95 homologous groups; 49 of these groups were unique to a single strain. While many of these prophages exhibited no sequence similarity to any lytic phage genome, estimation of the rate of phage acquisition suggests both vertical and horizontal acquisition. Furthermore, bioinformatic evidence indicates that prophage acquisition is ongoing within both vaginal and bladder Gardnerella populations. The abundance of prophage sequences within the strains examined here suggests that phages could play an important role in the species’ evolutionary history and in its interactions within the complex communities found in the female urinary and reproductive tracts.

Focusing the diversity of Gardnerella vaginalis through the lens of ecotypes

“…It is important to emphasize that the phylogenetic relationships of G. vaginalis inferred in this work are consistent with the findings of Ahmed et al. () based on a smaller set of 17 genomes and only looking at gene content in a model‐free comparison, as well as a recent study that included four G. vaginalis isolates from the bladder genome (Malki et al., ). Ahmed et al.…”

“…This difference is particularly important given that strains JCP8481A and JCP8481B lack sialidases and when clades 3A and 3B(3 and 4 inAhmed et al, 2012) are compared, no significant differences in the content of sialidases can be found.Our analysis of 35 isolates of G. vaginalis supports the existence of three major ecotypes based on the phylogenetic structure of their core and accessory genes and the cohesiveness in functional gene composition within ecotypes. These ecotypes likely evolved in response to different selective pressures imposed by differences among microbial communities and hosts, although these patterns might have arisen through neutral turnover of genes in strains evolving as segregated populations.It is important to emphasize that the phylogenetic relationships of G. vaginalis inferred in this work are consistent with the findings ofAhmed et al (2012) based on a smaller set of 17 genomes and only looking at gene content in a model-free comparison, as well as a recent study that included four G. vaginalis isolates from the bladder genome(Malki et al, 2016). Ahmed et al remarked that the genomic diversity among strains was great enough to warrant designation as four separate species.…”

Metagenomic analysis with strain-level resolution reveals fine-scale variation in the human pregnancy microbiome

“…However, healthy vs. "pathogenic" strains did not 10 map to specific clades. Consistent with previous comparative genomics reports of G. vaginalis (Ahmed et al 2012;Malki et al 2016), the G. vaginalis genomes recovered here show high degrees of re-arrangements with many unique islands. Although multiple G. vaginalis strains were isolated from the same subject previously (Ahmed et al 2012), the co-occurring genomes were not compared, and the importance of co-occurring strains was not addressed.…”

“…http://dx.doi.org/10.1101/266700 doi: bioRxiv preprint first posted online Feb. 18, 2018; based on conserved chaperonin cpn60 genes (Schellenberg et al 2016); four genotypes were proposed with molecular genotyping (Santiago et al 2011); and four clades were suggested from phylogenies reconstructed from core-gene sets (Malki et al 2016). Phylogenomic trees reconstructed from full-genome alignments, as well as from core SNPs, based on 34 publicly available genomes along with the genomes recovered here suggested a fifth divergent clade 5 within the G. vaginalis tree (Fig.…”

Metagenomic analysis with strain-level resolution reveals fine-scale variation in the human pregnancy microbiome

“…Although G. vaginalis has been associated with bacterial vaginosis and with increased risk of preterm birth (Schwebke et al 2014;DiGiulio et al 2015;Callahan et al 2017), it is also commonly found in vaginal CST IV communities of healthy women (Romero et al 2014). G. vaginalis "genotypes" have been proposed previously: Four main types were identified based on ∼473 concatenated core-gene alignments (Ahmed et al 2012); three groups were suggested based on conserved chaperonin cpn60 genes (Schellenberg et al 2016); four genotypes were proposed with molecular genotyping (Santiago et al 2011); and four clades were suggested from phylogenies reconstructed from core-gene sets (Malki et al 2016). Phylogenomic trees reconstructed from full-genome alignments, as well as from core SNPs, based on 34 publicly available genomes along with the genomes recovered here suggested a fifth divergent clade within the G. vaginalis tree.…”