PMC

(A) The 16S rRNA sequences were taken from a previously published alignment created using the NAST aligner []. A maximum-likelihood tree was generated using parameters estimated with ModelTest 3.7 and PAUP* (version 4.0b10). Terminal branch lengths are not drawn to scale.
(B) The average percent amino acid sequence identities were calculated using ClustalW alignments for the 530 sets of seven-way orthologs that include the five intestinal Bacteroidetes genomes, P. gingivalis, and Cy. hutchinsonii. B. thetaiotaomicron was used as a reference.

(A) CTn-mediated duplication of B. vulgatus CPS loci. Homologous gene pairs in the two duplicated regions are linked with fine gray lines, underscoring the high level of synteny. Genes constituting CPS loci 1 and 2 are highlighted in red, with the first and last genes numbered. Green denotes essential component genes of CTns. Blue brackets indicate two subregions that share 100% nucleotide sequence identity. The asterisk (*) indicates three open reading frames encoding two conserved hypothetical proteins and a hypothetical protein, suggesting an insertion that occurred after the duplication event.
(B) Locations of putative glycosyltransferase xenologs and inserted phage genes in CPS loci of the sequenced gut Bacteroidetes. Color code: integrases (green), UpxY transcriptional regulator homologs (black), putative xenologs (primarily glycosyltransferases, red), phage genes (blue), and remaining genes (gray). See for functional annotations.

The number of genes assigned to each GO term from each genome is shown. Significant enrichment is denoted by pink (p < 0.05) or red (p < 0.001), whereas depletion is indicated by light blue (p < 0.05) or dark blue (p < 0.001), as calculated by a binomial comparison followed by Benjamini-Hochberg false-discovery rate correction (see ).
(A) Genes assigned to GO terms related to core metabolic functions are enriched in the subset of common gut-associated Bacteroidetes orthologs shared with non-gut Bacteroidetes (seven-way comparison; abbreviated 7w), compared to the reference set of 1,416 orthologs common to the five sequenced gut Bacteroidetes genomes (5w), suggesting that all Bacteroidetes have inherited a core metabolome from their common ancestor. The set of orthologs that is not shared with non-gut–associated Bacteroidetes (five-way unique; 5wU) is enriched, relative to all orthologs (5w), for genes in three classes—amino acid biosynthesis; membrane transport; and two-component signal transduction systems—suggesting that these genes were important in the process of adaptation to the gut and/or other habitats by the common ancestor of gut Bacteroidetes.
(B) Various GO terms related to environmental sensing, gene regulation, and carbohydrate degradation are enriched in gut Bacteroidetes relative to Cy. hutchinsonii (Ch). A similar pattern is observed relative to P. gingivalis (Pg) (unpublished data). Note that these same classes of genes are depleted in the subset of shared gut Bacteroidetes orthologs ([A] 5w) relative to the full B. thetaiotaomicron (Bt) genome ([A] Bt-G). Thus, these classes of genes, though enriched in all gut Bacteroidetes, are widely divergent between them. Other classes of genes vary between species: B. distasonis (Bd) and B. vulgatus (Bv) show an expanded repertoire of proteases, whereas B. thetaiotaomicron lacks genes involved in synthesis of cobalamin. 6w refers to the orthologs shared by the five sequenced gut Bacteroidetes genomes (Bt, Bv, Bd, plus two B. fragilis strains [NCTC 9343 (BfN) and YCH 46 (BfY)] and Pg.

(A) Genes involved in core metabolic processes are enriched among non-laterally transferred genes identified by a phylogenetic approach (see ). The proportion of genes identified as not laterally transferred in each genome (light blue), as well as assigned to the GO terms “Primary metabolism” (yellow) and “Protein biosynthesis” (red), are shown. Significant increases (enrichment) relative to each whole genome are shown by an upward-pointing arrowhead, and decreases (depletion) by a downward-pointing arrowhead, whereas the corresponding probability, determined by a binomial test, is denoted by asterisks: a single asterisk (*) indicates p < 0.05; double asterisks (**) indicate p < 0.01; and triple asterisks (***) indicate p < 0.001.
(B) Laterally transferred genes are enriched among genes assigned to the GO term “DNA methylation” (e.g., restriction-modification systems) (red), relative to each complete genome (light blue). Glycosyltransferases (yellow) and genes located within CPS loci (green) are also enriched within the set of transferred genes. Significance was determined and denoted as in (A).
(C) B. distasonis (light blue) possesses a significantly larger proportion of laterally transferred genes than the other Bacteroidetes, as shown by significant increases in the proportion of genes in each category of our analysis (“LGT in,” laterally transferred into the genome; “Novel,” no homologs identified from other species; “LGT direction unresolved,” laterally transferred but direction unknown; “LGT out,” laterally transferred out of the genome; and “Unresolved,” lateral transfer uncertain; see for detailed explanations of categories and http://rd.plos.org/pbio.0050156.a for a complete list of genes in each category). Significant changes, denoted as in (A), were determined by a binomial test, using the average proportion within all other genomes used in the analysis as the reference. Other strains are B. vulgatus (red), B. thetaiotaomicron (yellow), B. fragilis NCTC 9343 (green), B. fragilis YCH 46 (purple), and P. gingivalis (orange).
(D) A prominent laterally transferred locus within B. distasonis contains a ten-gene hydrogenase complex, likely allowing B. distasonis to use hydrogen as a terminal electron acceptor in anaerobic respiration. Genes transferred into B. distasonis are colored red, whereas genes whose phylogeny could not be resolved are shown in yellow. Letters indicate functional components of the hydrogenase complex: L, large subunit; M, maturation or accessory factor; and S, small subunit.

(A) Cladogram generated from all fully sequenced Bacteroidetes. Branches that are unique to each species are color-coded as indicated. The homologous RagA/RagB proteins from P. gingivalis were selected as an arbitrary root (dashed branches). Dashed lines surrounding the tree indicate (1) a clade that is dominated by B. thetaiotaomicron SusC/SusD pairs (39/45 pairs, red dashes) and (2) a clade that is poorly represented in B. thetaiotaomicron (7/34 pairs, black dashes). Colored hash marks surrounding the cladogram represent the linkage of two other protein families, which show syntenic organization within related B. thetaiotaomicron SusC/SusD-containing loci: NHL repeat–containing proteins (light blue) and a group of conserved hypothetical lipidated proteins (light green). These protein families are not represented in the other sequenced Bacteroidetes, occur only adjacent to SusC/SusD pairs, and have no predicted functions. See http://rd.plos.org/pbio.0050156.a for locus tags for each taxon, branch bootstrap values, and lists of SusC/SusD-linked genes.
(B) An example of a recently amplified polysaccharide utilization locus in which the synteny of three flanking SusC/SusD genes has been maintained. The locations of the four SusC/SusD pairs encoded within these amplified clusters are indicated on the cladogram shown in (A) by asterisks (*). The locus schematic is arranged so that groups of related proteins (mutual best BLAST hits) are aligned vertically within the yellow box. The functions of amplified genes are indicated by numbers over each vertical column and, where applicable, are color coded to correspond to (A): 1, conserved hypothetical lipidated protein; 2, SusD paralog; 3, SusC paralog, 4, NHLrepeat–containing protein; and 5, glutaminase A (note that in three clusters, this gene has been partially deleted). Gray-colored genes downstream of each amplified cluster encode hypothetical proteins or predicted enzymatic activities (e.g., dehydrogenase, sulfatase, and glycoside hydrolase) that are unique to each cluster. A xenolog that has been inserted in one gene cluster is indicated in red; other genes are black. Dashed lines connecting gene clusters show linkage only, and do not correspond to actual genomic distance.
(C) An example of a recently duplicated locus from B. distasonis that includes duplicated regulatory genes. Syntenic regions are aligned as in (B) and include a single sulfatase (1, dark green), a SusD paralog (2, light purple), SusC paralog (3, dark purple), an anti-σ factor (4, light orange), and an ECF-σ factor (5, dark orange). Two other downstream sulfatase genes (gray) are also included in one cluster. The locations of the two SusC/SusD pairs encoded within these clusters are indicated on the cladogram shown in (A) by black arrows.