Ancient Genomes From Bronze Age Remains Reveal Deep Diversity and Recent Adaptive Episodes for Human Oral Pathobionts

Abstract

Ancient microbial genomes can illuminate pathobiont evolution across millenia, with teeth providing a rich substrate. However, the characterization of prehistoric oral pathobiont diversity is limited. In Europe, only preagricultural genomes have been subject to phylogenetic analysis, with none compared to more recent archaeological periods. Here, we report well-preserved microbiomes from two 4,000-year-old teeth from an Irish limestone cave. These contained bacteria implicated in periodontitis, as well as Streptococcus mutans, the major cause of caries and rare in the ancient genomic record. Despite deriving from the same individual, these teeth produced divergent Tannerella forsythia genomes, indicating higher levels of strain diversity in prehistoric populations. We find evidence of microbiome dysbiosis, with a disproportionate quantity of S. mutans sequences relative to other oral streptococci. This high abundance allowed for metagenomic assembly, resulting in its first reported ancient genome. Phylogenetic analysis indicates major postmedieval population expansions for both species, highlighting the inordinate impact of recent dietary changes. In T. forsythia, this expansion is associated with the replacement of older lineages, possibly reflecting a genome-wide selective sweep. Accordingly, we see dramatic changes in T. forsythia's virulence repertoire across this period. S. mutans shows a contrasting pattern, with deeply divergent lineages persisting in modern populations. This may be due to its highly recombining nature, allowing for maintenance of diversity through selective episodes. Nonetheless, an explosion in recent coalescences and significantly shorter branch lengths separating bacteriocin-carrying strains indicate major changes in S. mutans demography and function coinciding with sugar popularization during the industrial period.

Keywords: Streptococcus mutans; Tannerella forsythia; ancient pathogen genomics; microbial evolution; oral microbiome.

Fig. 1.

Oral microbiome retrieval from the Early Bronze Age at Killuragh Cave. a) A schematic of sampling for this study. Two teeth were sampled—KGH1 (left mandibular second molar) and KGH2 (right mandibular first molar). An aliquot was taken from each root (A and B, respectively) and each crown (E and F). b) Normalized relative abundance of 5 common oral pathobionts for ancient teeth, calculus and modern microbiome samples (Human Microbiome Project Consortium 2012; Lloyd-Price et al. 2017) and a lab control from our sequencing (see also supplementary table S3, Supplementary Material online). Relative abundance is normalized to the highest abundance sample for each taxon (S. mutans: 10.6%, KGH2-B; Tr. denticola: 4.5%, CS05; T. forsythia: 30.6%. CS36; P. gingivalis: 13.3%, CS32; F. nucleatum: 8.1%, SRR062298): color is scaled within species. Ancient samples are labeled and colored by publication (Skoglund et al. 2014; Warinner et al. 2014; Jones et al. 2015; Schiffels et al. 2016; Philips et al. 2017; Mann et al. 2018; Willmann et al. 2018; Velsko et al. 2019; Brunel et al. 2020; Cassidy et al. 2020; Fagernäs et al. 2020; Jacobson et al. 2020; Neukamm et al. 2020; Seguin-Orlando et al. 2021; Dulias et al. 2022).

Fig. 2.

Tannerella forsythia phylogenetic structure and virulence profiles. a) Maximum likelihood (ML) tree of T. forsythia sequences, using a Neanderthal sequence as an outgroup, alongside a matrix showing presence (black) or absence (white) of virulence genes examined by Philips et al. (2020), with sample IDs and symbols aligned with their positioning on the ML tree. Colors and shapes reflect time period and continental region, respectively. Grayed-out sample IDs were excluded from the main tree due to low coverage and are positioned with respect to time period. Trees with these additional 3 medieval and 2 industrial samples are shown in supplementary figs. S13 to 15, Supplementary Material online supporting the temporal structure seen here. Virulence categories are shown along the x axis in different colors. b) Bayesian skyline plot from BEAST showing inferred population size over the last 1,500 years, illustrating an exponential increase starting approximately 1,400 CE. c) Sampling locations of the genomes included (Europe and North Africa inset).

Fig. 3.

Streptococcus mutans phylogeny and mutacin profiles. Phylogeny inferred using MCC tree based on median values from independent BEAST runs of 500 million chains each. The deepest clades with >90% posterior support are colored (seven total). KGH2-B (black) and three modern genomes (gray) do not reliably group with any others. Eight nodes had median values lower than their direct descendent nodes, resulting in negative edge lengths, which were set to zero for visualization purposes only. A presence/absence matrix of 7 mutacins investigated in Watanabe et al. (2021) is aligned to tree tips. Black indicates mutacin positive. Samples lacking any mutacins are highlighted in gray and tend to have long branches. Underneath the phylogeny, median node heights from the BEAST tree are plotted in 250 year bins, showing a sharp increase in coalescences in the last 750 years. Eras of major dietary change are highlighted: cereal domestication in the Near East (Allaby et al. 2017); sugarcane spread from centers of domestication (Grivet et al. 2004) and the popularization of sugar in the 18th century.