Endospores and other lysis-resistant bacteria comprise a widely shared core community within the human microbiota

Abstract

Endospore-formers in the human microbiota are well adapted for host-to-host transmission, and an emerging consensus points to their role in determining health and disease states in the gut. The human gut, more than any other environment, encourages the maintenance of endospore formation, with recent culture-based work suggesting that over 50% of genera in the microbiome carry genes attributed to this trait. However, there has been limited work on the ecological role of endospores and other stress-resistant cellular states in the human gut. In fact, there is no data to indicate whether organisms with the genetic potential to form endospores actually form endospores in situ and how sporulation varies across individuals and over time. Here we applied a culture-independent protocol to enrich for endospores and other stress-resistant cells in human feces to identify variation in these states across people and within an individual over time. We see that cells with resistant states are more likely than those without to be shared among multiple individuals, which suggests that these resistant states are particularly adapted for cross-host dissemination. Furthermore, we use untargeted fecal metabolomics in 24 individuals and within a person over time to show that these organisms respond to shared environmental signals, and in particular, dietary fatty acids, that likely mediate colonization of recently disturbed human guts.

Fig. 1

Resistant fraction sequencing of human fecal bacteria. a Overview of resistant cell enrichment and 16S rDNA sequencing protocol. Resistant fraction samples are treated with a series of physical, enzymatic, and chemical lysis steps to deplete vegetative cells. DNA from bulk community and resistant fraction samples are extracted via a mechanical lysis protocol, and 16S rDNA libraries prepared. Communities are analyzed to determine the change in abundance of each OTU in the resistant fraction relative to the bulk community. (right) Phase contrast images of bulk community and resistant fraction—phase-bright cells are endospores. Endospores stain green when heat fixed with malachite green, vegetative cells appear red from safranin counter stain. b Representative results of 16S rDNA profile for bulk community and endospore-enriched samples. Reads from each OTU are summed across 24 individuals to give a meta-bulk and meta-endospore community. Phylogenetic classes in black text increase with resistant fraction; gray text classes decrease with resistant fraction. c Distribution of resistant fraction proportion across phyla aggregated across individuals filtered to remove OTUs with single counts in a sample (for visualization purposes). Colors represent phyla. OTUs with a resistant fraction proportion of 0 are absent from the resistant fraction; OTUs with a resistant fraction proportion of 1 are absent from the bulk community and only found in the resistant fraction

Fig. 2

Resistant fraction OTUs are more shared across individuals than bulk community OTUs. a Alpha-diversity metrics measured for the bulk community (x-axis) and resistant fraction enrichments (y-axis). P-values are for the test of differences between alpha-diversity metric distributions using paired Wilcoxon rank-sum test. b Distribution of Jaccard distance between resistant fraction and bulk communities, within the resistant fraction, and within the bulk communities. c Multidimensional scaling on the Jensen-Shannon divergence of all resistant fraction and bulk community samples. Black dots represent resistant fraction communities, gray dots represent bulk community samples. d Comparison of the number of rOTU sequence variants in only one person and in multiple people to nOTU sequence variants in only one person and in multiple people

Fig. 3

Taxa show heterogeneous patterns of resistant cell fractions across individuals. Phylogenetic placement of the fraction of resistant organisms for taxa present within at least eight individuals estimated by the ratio of counts scaled by qPCR-estimates of biomass in the resistant fractions and bulk communities. Tree branch colors represent the degree to which a taxonomic group was enriched (higher relative abundance) in the resistant fraction with pink branches never enriched and blue and green branches enriched at least once across samples. Classes within each phylum are shown with a colored bar. Arrows indicate OTUs showing the maximum (black) and minimum (gray) within-OTU variability in enrichment scores. (Color figure online)

Fig. 4

Common signals govern resistant state exit and growth in the GI tract. a Boxplot of the distribution of correlation distances (pairwise Euclidean distance between the Spearman correlation vectors for two OTUs) between rOTUs and nOTUs, within rOTUs (enriched), and within nOTUs (unenriched). b Abundance of OTUs in the resistant fraction as a function of bile acid exposure for nine phylogenetically distant OTUs

Fig. 5

Resistant OTUs show disproportionate turnover in diverse contexts. a Overview of approach for identifying resistant-cell forming OTUs in 16S rDNA sequencing datasets. rOTU database sequences are matched to sequences in other datasets, and then patterns within those datasets among the identified rOTUs are determined. b Fraction of rOTUs present during microbial colonization of an infant gut annotated with major diet and health perturbations. rOTUs encompass both putative endospore-forming organisms and those not known to form endospores, but which possess a resistant state (Actinobacteria and non-endospore-forming Firmicutes). c Fraction of rOTUs present as a function of C. difficile infection status (fCDI = first time C. difficile diagnosis, rCDI = at least 3 episodes of C. difficile infection following initial treatment). d Fraction of rOTUs and all other OTUs (non-resistant OTUs) transferred from donors to recipients by fecal microbiota transplant. e Time series of rOTUs (top) and all other (non-resistant) OTUs (bottom) from a human male infected with Salmonella, with OTUs significantly more abundant pre-infection (dark gray) and significantly more abundant post-infection (light gray)