Microbiota assembly of specific pathogen-free neonatal mice

Abstract

Background: Neonatal mice are frequently used to model diseases that affect human infants. Microbial community composition has been shown to impact disease progression in these models. Despite this, the maturation of the early-life murine microbiome has not been well-characterized. We address this gap by characterizing the assembly of the bacterial microbiota of C57BL/6 and BALB/c litters from birth to adulthood across multiple independent litters.

Results: The fecal microbiome of young pups is simple, dominated by only a few pioneering bacterial taxa. These taxa are present at low levels in the microbiota of multiple maternal body sites, precluding a clear identification of maternal source. The pup microbiota begins diversifying after fourteen days, coinciding with the beginning of coprophagy and the consumption of solid foods. Pup stool bacterial community composition and diversity are not significantly different from dams from day 21 onwards. Short-read shotgun sequencing-based metagenomic profiling of young pups enabled the assembly of metagenome-assembled genomes for strain-level analysis of these pioneer Ligilactobacillus, Streptococcus, and Proteus species.

Conclusions: Assembly of the murine microbiome occurs over the first weeks of postnatal life and is largely complete by day 21. This detailed view of bacterial community development across multiple commonly employed mouse strains informs experimental design, allowing researchers to better target interventions before, during, or after the maturation of the bacterial microbiota. The source of pioneer bacterial strains appears heterogeneous, as the most abundant taxa identified in young pup stool were found at low levels across multiple maternal body sites, suggesting diverse routes for seeding of the murine microbiome.

Keywords: development; early-life; microbiome; microbiota; mother-infant transmission; neonatal; pioneer species; seeding.

Figure 1:. The early-life fecal microbiota begins diversifying at postnatal day 15.

(A) Taxonomic classification of stool samples collected from neonates sequenced at the 16S rRNA gene V4 region. Genera represented at greater than 5% abundance in at least one sample are shown. Each bar represents the average abundance of taxa from all samples collected from pups from the litter for a given time point; the n above each bar indicates the number of samples collected for that time point. Color families show phyla-level assignments – blue for Bacteroidetes, orange for Deferribacteres, red for Firmicutes, green for Proteobacteria, and yellow for Verrucomicrobia; grey includes phyla present at less than 5% abundance in all samples. (B) Shannon diversity calculated based on operational taxonomic unit (OTU) clustering. (C) Faith’s phylogenetic diversity was calculated based on the phylogeny of amplicon sequence variants (ASVs). (D) 16S rRNA gene copies per fecal pellet, detected by qPCR. Medians are indicated by a horizontal line. Results were compared by the Kruskal-Wallis test with Dunn’s test for multiple comparisons. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001, ns = not significant; n =12 – 57, representing samples from mice combined across four litters from two genotypes as in A within indicated age ranges.

Figure 2:. Pup fecal bacterial community structure shifts with age and shares features with dam fecal microbiota.

(A) Stool samples were clustered by litter using principal coordinate analysis (PCoA) based on theta similarity coefficients. Each point represents a single stool sample, colored according to age. Samples clustered together have a more similar community structure. (B) All stool samples were clustered using PCoA based on theta similarity coefficients. Each box represents the average of all stool samples taken at a given age for that litter, with the number indicating the postnatal day on which the samples were collected, with lines connecting subsequent times. (C) Theta similarity of samples of the indicated pup age compared to dam samples collected at the age of pup weaning. Samples from each litter were compared either to their dam or to other dams. Means are indicated by the top of the bars. Results were compared by the Kruskal-Wallis test. *** p < 0.001, **** p < 0.0001, ns = not significant; n =19-168 pup-dam pairs per group.

Figure 3:. Maternal body site microbiota samples do not cluster by site.

(A) Taxonomic classification of maternal body-site samples. Genera represented at greater than 5% abundance in at least one sample are shown. Read counts are displayed above each sample; only maternal samples with greater than 1500 reads were used for analysis. (B) Shannon diversity of maternal body-site samples. Means are indicated by the thick horizontal crossbar and error bars indicate the standard error of the mean. Results were compared by the Kruskal-Wallis test with Dunn’s test for multiple comparisons. * p < 0.05, ** p < 0.01, ns = not significant; n = 6 – 12, representing samples from mice combined across four litters from two genotypes. (C) Maternal body site samples clustered using principal coordinate analysis based on theta similarity coefficients. Numbers within points indicate the age of pups at the time of sampling. (D) Heatmap of mean relative abundance at maternal body sites of the six ASVs present at greater than 5% relative abundance in the stool of at least one pup up to P10. n = 2-20 samples.

Figure 4:. Limited strain-level diversity detected in the early pup microbiome

(A) Percent of nucleotide identity shared between strains of Ligilactobacillus murinus and Proteus mirabilis identified in seven shotgun-sequenced mouse samples and their most closely related NCBI genomes, L. murinus ASF361 and Proteus mirabilis swupm1, respectively, as determined by StrainGST. Sample names indicate mouse strain, litter, pup age in days, and sample number, separated by dots. (B) Heatmap colors indicate pairwise nucleotide identity between L. murinus strains identified in each of the seven mouse samples as determined by StrainGR. The cladogram was calculated using Euclidean distance. (C) Heatmap colors indicate pairwise nucleotide identity between P. mirabilis strains identified in each of the seven mouse samples as determined by StrainGR. The cladogram was calculated using Euclidean distance.