Toothbrush microbiomes feature a meeting ground for human oral and environmental microbiota

Abstract

Background: While indoor microbiomes impact our health and well-being, much remains unknown about taxonomic and functional transitions that occur in human-derived microbial communities once they are transferred away from human hosts. Toothbrushes are a model to investigate the potential response of oral-derived microbiota to conditions of the built environment. Here, we characterize metagenomes of toothbrushes from 34 subjects to define the toothbrush microbiome and resistome and possible influential factors.

Results: Toothbrush microbiomes often comprised a dominant subset of human oral taxa and less abundant or site-specific environmental strains. Although toothbrushes contained lower taxonomic diversity than oral-associated counterparts (determined by comparison with the Human Microbiome Project), they had relatively broader antimicrobial resistance gene (ARG) profiles. Toothbrush resistomes were enriched with a variety of ARGs, notably those conferring multidrug efflux and putative resistance to triclosan, which were primarily attributable to versatile environmental taxa. Toothbrush microbial communities and resistomes correlated with a variety of factors linked to personal health, dental hygiene, and bathroom features.

Conclusions: Selective pressures in the built environment may shape the dynamic mixture of human (primarily oral-associated) and environmental microbiota that encounter each other on toothbrushes. Harboring a microbial diversity and resistome distinct from human-associated counterparts suggests toothbrushes could potentially serve as a reservoir that may enable the transfer of ARGs. Video abstract.

Keywords: Antimicrobial resistance; Built environment; Metagenomics; Oral microbiome; Toothbrush.

Fig. 1

Toothbrush microbiota taxonomic diversity reflects a subset of the human oral microbiota with minor influence from other body sites and the built environment. a Frequency-abundance of toothbrush microbiota; color intensity corresponds to taxon frequency of detection in HMP-II [20] oral microbiome samples. b Frequency of detection of conserved toothbrush-associated taxa (i.e., those in at least 50% toothbrush samples) across different sample types, including toothbrushes, indoor dust [3, 4], tap water [21, 22], shower head biofilms [unpublished], and various body sites from the human microbiome project (HMP-II) [20]. Heatmap intensity reflects frequency of taxon detection in the respective sample types. Sidebar color corresponds to taxon phylum and an asterisk indicates member of the “core toothbrush microbiota” (i.e., in at least 75% samples). c PCoA displaying species-level beta-diversity across microbiota of toothbrushes, relevant environmental samples, and human microbiomes. Color/shape correspond to sample type. d Alpha-diversity (Shannon index) for each sample type

Fig. 2

Toothbrush-associated microbiota are derived from a mix of prominent human and environmental taxa. a Distributions of putative origins of the toothbrush microbiota (genus-level) predicted by SourceTracker [23]. b PCoA displaying genus-level beta-diversity across toothbrush microbiota; color intensity indicates the cumulative fraction of taxa putatively derived from the human microbiome per sample. c Average relative abundances of highly abundant genera (i.e., > 2% average relative abundance) in all toothbrush metagenomes and those with putatively less than or greater than 50% microbiota derived from the human microbiome (i.e., primarily human-derived vs. non-niche-specific, respectively)

Fig. 3

Although toothbrushes contained a less diverse microbiota and fewer commonly occurring antibiotic resistance genes (ARGs) than the oral microbiome, they had a more diverse resistome. a K-means clustering of all HMP-II oral sample taxonomic profiles (n = 1259; gray points) was used to generate centroids equal in size to the number of toothbrush samples (n = 34). Samples closest to the centroids (orange points) were selected for resistome analysis. b Shannon indices for overall taxonomic and ARG profiles for the subset oral samples (orange) and on toothbrushes (red). Mann-Whitney test p values for differences by sample type are displayed. c Overlap of conserved (i.e., detected in at least 50% respective samples) microbial species and antibiotic ARGs

Fig. 4

Enrichments in the toothbrush resistome correlate with the environmental-derived fraction of toothbrush microbial communities. a Left: Frequency of ARGs detected in oral (orange) and toothbrush (red) samples for ARGs present in > 10% of all samples (n = 68). Vertical dotted lines distinguish conserved ARGs (i.e., detected in at least 50% respective samples). An asterisk indicates enrichment by sample type (Bonferroni q < 0.05). Right: Log-transformed normalized RPKM counts for marker hits to each ARG protein family. Crossbar indicates median value for samples with > 0 counts. Axis tick colors correspond to ARG drug class. b PCoA displaying beta-diversity of toothbrush microbiomes based on normalized RPKMs of ARGs; color intensity indicates the cumulative fraction of taxa putatively derived from the human microbiome per sample (see Fig. 2a). c Log-transformed normalized RPKMs (average ± standard error) of ARGs that were enriched on toothbrushes (indicated in a) for toothbrush metagenomes containing less than 50% (gray) or greater than 50% (orange) microbiota putatively derived from the human microbiome

Fig. 5

Metagenome-assembled genomes containing ARGs were most often linked to environment-associated taxa. Columns correspond to ARGs detected among bins (n = 29 ARGs); bottom side color corresponds to drug class, and asterisk indicates that the ARG had been enriched by toothbrush or oral-associated metagenome sample type (Bonferroni q < 0.05). Rows correspond to bins with at least one ARG detected (i.e., 28/110 bins); toothbrush sample ID, bin number, and the predicted bacterial species are displayed. Oral-associated or not niche-specific categories were assigned based on whether the bacterial species (or genus) was more or less abundant on toothbrushes that contained primarily oral-derived microbiota, as predicted by SourceTracker [23]

Fig. 6

Summary statistics for subject population (n = 34). Participant information regarding oral hygiene, health and activity, and bathroom attributes. These metadata were selected from all survey responses (n = 20/40) for statistical analysis with toothbrush microbiome data, as they had relatively well-balanced responses (i.e., at least 20% for more than one categorical answer in the respective question)

Fig. 7

Trends for possible dental hygiene associations with diversity of toothbrush microbiome taxonomic (top row) and ARG (bottom row) profiles. Comparisons yielding Mann-Whitney test p ≤ 0.05 or p ≤ 0.1 are displayed as “**” or “*,” respectively

Fig. 8

Significant associations between metadata and relative abundances of bacterial species (a) or RPKMs of ARGs (b) in toothbrush microbiomes based on Hierarchical All-against-All association (HAllA) testing. All metadata presented in Fig. 4 (n = 20 variables) were included in the HAllA analyses and those with significant association (q < 0.1) to at least one taxon/ARG are presented in the respective figure panels (i.e., 12/20 and 7/20 variables associated with particular taxa and ARGs, respectively). Significant clusters of related features (q < 0.1) are outlined and ranked based on hierarchy of similarity scores. Grid color intensity corresponds to normalized mutual information similarity metric. Sidebar colors in b correspond to the ARG drug class