Functional and genetic markers of niche partitioning among enigmatic members of the human oral microbiome

Abstract

Introduction: Microbial residents of the human oral cavity have long been a major focus of microbiology due to their influence on host health and intriguing patterns of site specificity amidst the lack of dispersal limitation. However, the determinants of niche partitioning in this habitat are yet to be fully understood, especially among taxa that belong to recently discovered branches of microbial life.

Results: Here, we assemble metagenomes from tongue and dental plaque samples from multiple individuals and reconstruct 790 non-redundant genomes, 43 of which resolve to TM7, a member of the Candidate Phyla Radiation, forming six monophyletic clades that distinctly associate with either plaque or tongue. Both pangenomic and phylogenomic analyses group tongue-specific clades with other host-associated TM7 genomes. In contrast, plaque-specific TM7 group with environmental TM7 genomes. Besides offering deeper insights into the ecology, evolution, and mobilome of cryptic members of the oral microbiome, our study reveals an intriguing resemblance between dental plaque and non-host environments indicated by the TM7 evolution, suggesting that plaque may have served as a stepping stone for environmental microbes to adapt to host environments for some clades of microbes. Additionally, we report that prophages are widespread among oral-associated TM7, while absent from environmental TM7, suggesting that prophages may have played a role in adaptation of TM7 to the host environment.

Conclusions: Our data illuminate niche partitioning of enigmatic members of the oral cavity, including TM7, SR1, and GN02, and provide genomes for poorly characterized yet prevalent members of this biome, such as uncultivated Flavobacteriaceae.

Keywords: Candidate phyla radiation; Human oral cavity; Metagenomics; Metapangenomics; Niche partitioning; Saccharibacteria.

Fig. 1

MAGs cover most of the abundant genera of the oral microbiome as well as represent lineages absent in public genomic databases. The dendrogram in the middle of the figure organizes 227 MAGs, 1582 genomes from the HOMD, and a single archeon, which was used to root the tree, according to their phylogenomic organization based on our collection of ribosomal proteins. The bars in the innermost circular layer represent the length of each genome. The second layer shows the phylum affiliation of each genome. The third layer shows the 10 most abundant genera in our samples as estimated by KrakenUniq. The fourth layer shows the affiliation of genomes as either MAGs from our study (blue) or genomes from HOMD (gray). The outermost layer marks novel genomes of lineages that lack representation in HOMD and NCBI. The lowest taxonomic level that could be assigned using CheckM and sequence search (see “Material and methods”) is listed for each novel lineage

Fig. 2

Detection of TM7 genomes across oral metagenomes and their phylogeny. a Most TM7 populations are exclusively detected in either tongue or plaque samples in our dataset. For each of the 43 MAGs (on the x-axis), the green and blue bars represent the portion of plaque and tongue samples, respectively, in which it is detected (detection > 0.5). b Phylogenetic organization of TM7 genomes reveals niche-associated oral clades. The phylogenetic tree includes the 52 oral TM7 genomes (9 of which were previously published), as well as 5 genomes of Firmicutes that root the tree. The layers below the tree describe (top to bottom): “Oral site”—the oral site to which each of our MAGs corresponded, where blue marks tongue dorsum, green marks supragingival plaque, and turquoise marks the “cosmopolitan” TM7; “Study”—the study associated with each genome: our MAGs (purple), Espinoza et al. [41] (teal), Marcy et al. [37] (blue), He et al. [33] (red), and Cross et al. [38] (orange). A red circle appears on the dendrogram and indicates the junction that separates the majority of plaque specialists from tongue specialists, and bootstrap values appear above branches that separate major groups. † Refined versions of genomes, which we previously published [46]. ‡ Genomes from IMG that we refined in this study, but for which accession numbers for refined versions are available in Cross et al. [38]

Fig. 3

Phylogenomic analysis of human oral TM7 with all TM7 genomes on the NCBI’s GenBank shows association of plaque TM7 with environmental genomes, and tongue TM7 with TM7 from animal stool. The phylogenetic tree at the top of the figure was computed using ribosomal proteins and includes 5 Firmicutes as an outgroup. Regions of the tree that are associated with either plaque or tongue clades from Fig. 2 are marked with green or blue shaded backgrounds respectively. Bootstrap support values are shown next to branches separating major oral clades. Subclades are marked with rectangles below the branches they represent. The layers below the tree provide additional information for each genome. From top to bottom: Clade: the clade associations for finer groupings of oral genomes. Oral Site: the oral site with which the genome is associated is shown for our MAGs in accordance with Fig. 2. Source: the source of the genome, where red is human oral; brown, animal gut; cyan, dolphin oral; and black, environmental samples. Reference: the genomes from this study in blue, and genomes from Parks et al. in gray [50]. The majority of the rest of the genomes originate from various publications from the Banfield Lab at UC Berkeley. The insert at the top right of the figure shows boxplots for ANI results for genomes in each subclade against all other genomes. Data points represent the ANI score for comparisons in which the alignment coverage was at least 25%. Within-subclade comparisons appear in green, and between-subclades comparisons appear in red

Fig. 4

Detection and coverage of TM7 populations in the HMP plaque and tongue samples reveals abundant populations and niche specificity. The tree at the top of the figure and the two layers of information below it are identical to the one in Fig. 2. Barplots below the tree show the portion of plaque (green) and tongue (blue) HMP samples in which each TM7 was detected, using a detection threshold of 0.5. Boxplots at the bottom of the figure show the normalized coverages of each TM7 in plaque (green) and tongue (blue) HMP samples in which it was detected

Fig. 5

Pangenome of TM7—Accessory gene clusters include clade-specific and niche-specific markers. The dendrogram in the center of the figure organizes the 4045 gene clusters (GCs) that occurred in more than one genome according to their frequency of occurrence in the 55 TM7 genomes. The 55 inner layers correspond to the 55 genomes, where our MAGs that were associated with tongue and plaque are shown in blue and green, respectively. Previously published oral and environmental genomes are shown in black and brown, respectively. The data points in the 55 concentric layers show the presence of a GC in a given genome, and the outermost circular layer highlights groups of GCs that correspond to the core or to group-specific GCs. Genomes in this figure are ordered according to their phylogenomic organization which is shown at the top-right corner. The two top horizontal layers underneath the phylogenomic tree represent clade and oral-site associations of genomes. The next two layers display coverage statistics for each genome in the HMP oral metagenomes from tongue (blue, top) and supragingival plaque (bottom, green) samples

Fig. 6

a Type IV pilus operon is highly conserved in TM7 genomes, but missing many components in genomes of the tongue-associated clade T1. Type IV pili operons from 52 of the 55 TM7 that included pilC are aligned according to pilC (yellow). Genomes are organized according to their phylogenetic organization shown in Fig. 5. The top 10 functions identified in these operons appear with color filling, while the rest of the functions appear in gray. Contig breaks are marked with red lines for contigs that include less than 9 genes either upstream or downstream from pilC. b Some phage groups span phylogenetic clades, while other phage groups associate with specific clades. At the top of the panel, the two prophages of phage group pg08 are compared and on the bottom of the panel the two prophages of the phage group pg02 are compared. White arrows signify genes as identified by Prodigal. Homologous genes, identified as belonging to the same GC, are connected by colored areas. A function name assigned by KEGG, COG, or Pfam functional annotation source appears for genes for which it was available. On the left, the phylogenetic clade of the TM7 host of each prophage is listed next to the host genome name. The genome-wide average nucleotide identity (gANI) appears for each pair of the host genomes, where C/I stands for alignment coverage/alignment identity