Unique roles of vaginal Megasphaera phylotypes in reproductive health

Abstract

The composition of the human vaginal microbiome has been extensively studied and is known to influence reproductive health. However, the functional roles of individual taxa and their contributions to negative health outcomes have yet to be well characterized. Here, we examine two vaginal bacterial taxa grouped within the genus Megasphaera that have been previously associated with bacterial vaginosis (BV) and pregnancy complications. Phylogenetic analyses support the classification of these taxa as two distinct species. These two phylotypes, Megasphaera phylotype 1 (MP1) and Megasphaera phylotype 2 (MP2), differ in genomic structure and metabolic potential, suggestive of differential roles within the vaginal environment. Further, these vaginal taxa show evidence of genome reduction and changes in DNA base composition, which may be common features of host dependence and/or adaptation to the vaginal environment. In a cohort of 3870 women, we observed that MP1 has a stronger positive association with bacterial vaginosis whereas MP2 was positively associated with trichomoniasis. MP1, in contrast to MP2 and other common BV-associated organisms, was not significantly excluded in pregnancy. In a cohort of 52 pregnant women, MP1 was both present and transcriptionally active in 75.4 % of vaginal samples. Conversely, MP2 was largely absent in the pregnant cohort. This study provides insight into the evolutionary history, genomic potential and predicted functional role of two clinically relevant vaginal microbial taxa.

Keywords: BV; Megasphaera; trichomoniasis; vaginal microbiome.

Fig. 1.

Maximum-likelihood phylogenetic tree of the class negativicutes. A total of 145 orthologous genes from 110 genomes assigned to the class Negativicutes were included in this analysis. Clostridium botulinum A strain Hall was designated as the outgroup. This maximum-likelihood phylogenetic tree was generated using 100 bootstrap replicates. Bootstrap values as present at nodes of the tree. Families within the tree highlighted in different colours: Sporomusaceae: yellow, Selenomonadaceae: blue, Acidaminococcaceae: green, Veillonellaceae: purple. MP1 and MP2 genomes are outlined with dotted lines and labelled.

Fig. 2.

Maximum-likelihood phylogenetic tree of 16S ribosomal RNA gene. A maximum-likelihood tree with 1000 bootstraps was generated from full-length 16S ribosomal RNA gene sequences (nucleotide). Numbers at nodes are indicative of bootstrap support of that node placement. Dialister micraerophilus was selected as the outgroup and is a human oral isolate also classified in the family Veillonellaceae. Remaining isolates are coloured by their site of isolation: blue – mammalian gut, green – human oral, purple – human vaginal.

Fig. 3.

Distinctive GC composition, codon preference and genomic structure between vaginal Megasphaera phylotypes. Differences in both (a) whole genome and (b) protein-coding GC composition are shown. (c) Codon preference is distinct between MP1 and MP2 genomes based on differences in GC composition at specific codon positions and in the overall coding GC composition (MP1: blue, MP2: orange). (d) Synteny is conserved within phylotype but lost between MP1 and MP2 genomes. Synteny plots demonstrate structural alignment of genomic content at the amino acid level. Colour designates similarity at the amino acid level. The left panel shows the strong conservation of genomic synteny and protein identity (red colour) between two MP1 genomes. The right panel show massive genome rearrangement and loss of amino acid sequence conservation between a MP1 and a MP2 genome.

Fig. 4.

Vaginal Megasphaera phylotypes associated with increased alpha diversity. Alpha diversity of vaginal microbiome profiles was calculated using the inverse Simpson’s index. Distribution of inverse Simpson’s index for each group is shown. Boxes show median and interquartile ranges, with whiskers denoting maximum and minimum values. Outliers are shown as dots. Significance was determined using a two-tailed Student’s t-test. Four different groups are shown, samples containing neither MP1 or MP2 (n=1901, red), samples containing MP1 only (n=845, yellow), samples containing both MP1 and MP2 (n=182, green) and samples containing only MP2 (n=163, blue). Taxa were determined to be present in a sample if they comprised greater than or equal to 0.1 % of the sample. Samples with MP1 only, MP2 only and both phylotypes all exhibit increased alpha diveristy, with MP2 only samples being the most highly diverse. All comparisons were found to be highly signifcant (P<0.01) with the exception of MP2 only and both phylotypes.

Fig. 5.

Megasphaera phylotype 1 (MP1) not significantly excluded in pregnancy. Results were generated from a cohort of 779 pregnant women case matched 1 : 1 with non-pregnant controls (N=1558). The distribution of the proportional abundance of taxa of interest is shown in both pregnant (red) and non-pregnant (blue) cohorts. Boxes show median and interquartile ranges, with whiskers denoting maximum and minimum values. Outliers are shown as dots. Significance was determined using a Mann–Whitney U test with FDR correction. Lactobacillus iners is shown to be significantly more prevalent in the pregnant cohort (q=1.20E-6). Lactobacillus crispatus cluster, Lactobacillus jensenii, Lachnospiraceae BVAB1, Megasphaera phylotype 1 (MP1) and Ureaplasma cluster 23 are not significantly different between the two cohorts (q=0.19, 0.23, 0.43, 0.56, 0.26, respectively). Megasphaera phylotype 2 (MP2), Prevotella cluster 2, Atopobium vaginae and Gardnerella vaginalis are significantly lower in the pregnant cohort (q=5.82×10⁻³, 6.09E-8, 7.95E-8, 2.82E-7, respectively).

Fig. 6.

Relationship between Megasphaera 16S read abundance and transcript abundance in paired datasets. Results were generated from samples collected from a cohort of pregnant women that participated in the MOMS-PI study. Samples were processed for whole metagenome microbiomics and transcriptomics. Percent of total transcripts attributed to the taxon of interest is shown on the y-axis. Percent of total whole metagenome sequencing reads attributed to the taxon of interest are shown on the x-axis. Each dot represents an individual sample. The relationship between WMGS and WMTS representation of Megasphaera phylotype 1 (MP1) is shown. Data points representing samples from women who went on to deliver full term are shaded blue, while data points representing samples from women who went on to deliver preterm are shaded red. The dotted line extending across the graph diagonally represent the expected 1 : 1 relationship of WMGS and WMTS- based abundance measures.