Microbiota of the Pregnant Mouse: Characterization of the Bacterial Communities in the Oral Cavity, Lung, Intestine, and Vagina through Culture and DNA Sequencing

Abstract

Mice are frequently used as animal models for mechanistic studies of infection and obstetrical disease, yet characterization of the murine microbiota during pregnancy is lacking. The objective of this study was to characterize the microbiotas of distinct body sites of the pregnant mouse-vagina, oral cavity, intestine, and lung-that harbor microorganisms that could potentially invade the murine amniotic cavity, thus leading to adverse pregnancy outcomes. The microbiotas of these body sites were characterized through anoxic, hypoxic, and oxic culture as well as through 16S rRNA gene sequencing. With the exception of the vagina, the cultured microbiotas of each body site varied by atmosphere, with the greatest diversity in the cultured microbiota appearing under anoxic conditions. Only cultures of the vagina were comprehensively representative of the microbiota observed through direct DNA sequencing of body site samples, primarily due to the predominance of two Rodentibacter strains. Identified as Rodentibacter pneumotropicus and Rodentibacter heylii, these isolates exhibited predominance patterns similar to those of Lactobacillus crispatus and Lactobacillus iners in the human vagina. Whole-genome sequencing of these Rodentibacter strains revealed shared genomic features, including the ability to degrade glycogen, an abundant polysaccharide in the vagina. In summary, we report body site-specific microbiotas in the pregnant mouse with potential ecological parallels to those of humans. Importantly, our findings indicate that the vaginal microbiotas of pregnant mice can be readily cultured, suggesting that mock vaginal microbiotas can be tractably generated and maintained for experimental manipulation in future mechanistic studies of host vaginal-microbiome interactions. IMPORTANCE Mice are widely utilized as animal models of obstetrical complications; however, the characterization of the murine microbiota during pregnancy has been neglected. Microorganisms from the vagina, oral cavity, intestine, and lung have been found in the intra-amniotic space, where their presence threatens the progression of gestation. Here, we characterized the microbiotas of pregnant mice and established the appropriateness of culture in capturing the microbiota at each site. The high relative abundance of Rodentibacter observed in the vagina is similar to that of Lactobacillus in humans, suggesting potential ecological parallels. Importantly, we report that the vaginal microbiota of the pregnant mouse can be readily cultured under hypoxic conditions, demonstrating that mock microbial communities can be utilized to test the potential ecological parallels between microbiotas in human and murine pregnancy and to evaluate the relevance of the structure of these microbiotas for adverse pregnancy outcomes, especially intra-amniotic infection and preterm birth.

Keywords: Rodentibacter; anoxic; atmosphere; cultivation; hypoxic; microbiome; mouse model; oxic; pregnancy.

FIG 1

Study design for characterizing the microbiotas of the oral cavity, intestine, lung, and vagina of pregnant mice. Briefly, two sets of samples were collected from each body site of 11 pregnant mice. One set of samples was used for culture and the other for molecular surveys. Cultures were performed on samples from each body site, under three different atmospheric conditions on multiple medium types. Bacterial growth from each plate type was collected by plate washing with sterile PBS and then combined under each atmosphere. These samples subsequently had their DNA extracted followed by 16S rRNA gene amplification and sequencing. After classification of 16S rRNA gene sequences through DADA2, culture profiles for each body site under each atmosphere were generated as well as overall body site culture profiles after pooling of the sequence data from all three atmospheres. Samples for molecular surveys had their DNA extracted directly from the samples followed by 16S rRNA gene amplification, sequencing, and classification to generate molecular profiles. ASV, amplicon sequence variant; DADA2, divisive amplicon denoising algorithm 2; PCR, polymerase chain reaction.

FIG 2

Alpha diversity comparisons between the microbiotas cultured under different atmospheres for the oral cavity, lung, intestine, and vagina and between body sites. Bar plots indicate differences in three alpha diversity measures among anoxic, hypoxic, and oxic cultures of the oral cavity (A), lung (B), intestine (C), and vagina (D) as well as across body sites (E). For panel E, culture data from each atmosphere for each individual mouse by body site were bioinformatically pooled, and only mice with culture data from all body sites and all atmospheres (n = 5) were included in the analyses. Data points are color coded by mouse ID and are consistent across panels. Lowercase letters that are shared within each panel indicate pairwise comparisons that were not significant.

FIG 3

Comparisons of cultured microbiota from the oral cavity, lung, intestine, and vagina, controlled for atmosphere. (A and B) PCoA plots illustrating variation among cultured microbiota of the oral cavity, lung, intestine and vagina using the Jaccard dissimilarity index (A) for composition and the Bray-Curtis dissimilarity index (B) for structure. Ellipses indicate standard deviations. (C) Heatmap including ASVs with ≥1% average relative abundance within a single body site. Samples are clustered by Bray-Curtis similarities within each body site. (D) LEfSe analysis identifying taxa preferentially recovered from a particular body site. Each node represents a taxon at each taxonomic level starting with the kingdom Bacteria down through genus in the outermost nodes and are colored based on preferential recovery from a specific body site. Yellow nodes represent taxa that were not recovered preferentially from a particular body site. The diameter of each node is proportional to the relative abundance of that taxon. Phylum, class, and order (not labeled) clades are highlighted when significant for a particular atmosphere. ASV, amplicon sequence variant; LEfSe, linear discriminant analysis effect size.

FIG 4

Comparisons of sequenced microbiota and cultured microbiota from the oral cavity, lung, intestine, and vagina. (A) Heatmap showing log-transformed percent relative abundances with hierarchical clustering based on Bray-Curtis values. (B) Molecular and culture profiles were separately averaged, with dots indicating whether an ASV was detected in culture. ASVs were included if they had a ≥1% average relative abundance in the molecular profiles for one of the four body sites. ASV, amplicon sequence variant.

FIG 5

Phylogenomic and KEGG analysis of two vaginal Rodentibacter isolates. (A and B) Phylogenomic trees including the Rodentibacter isolates ASV 2 and ASV 5 and all Rodentibacter type strains (A) and all published Rodentibacter genomes. (C and D) Distribution of functional KEGG pathways enriched in the genomes of the two isolates. Phylogenomic trees were constructed by comparing 92 conserved bacterial genes as described by Na et al. (119). ASV, amplicon sequence variant; KEGG, Kyoto Encyclopedia of Genes and Genomes.