The Phylogenomic Diversity of Herbivore-Associated Fibrobacter spp. Is Correlated to Lignocellulose-Degrading Potential

Abstract

Members of the genus Fibrobacter are cellulose-degrading bacteria and common constituents of the gastrointestinal microbiota of herbivores. Although considerable phylogenetic diversity is observed among members of this group, few functional differences explaining the distinct ecological distributions of specific phylotypes have been described. In this study, we sequenced and performed a comparative analysis of whole genomes from 38 novel Fibrobacter strains against the type strains for the two formally described Fibrobacter species F. succinogenes strain S85 and F. intestinalis strain NR9. Significant differences in the number of genes encoding carbohydrate-active enzyme families involved in plant cell wall polysaccharide degradation were observed among Fibrobacter phylotypes. F. succinogenes genomes were consistently enriched in genes encoding carbohydrate-active enzymes compared to those of F. intestinalis strains. Moreover, genomes of F. succinogenes phylotypes that are dominant in the rumen had significantly more genes annotated to major families involved in hemicellulose degradation (e.g., CE6, GH10, and GH43) than did the genomes of F. succinogenes phylotypes typically observed in the lower gut of large hindgut-fermenting herbivores such as horses. Genes encoding a putative urease were also identified in 12 of the Fibrobacter genomes, which were primarily isolated from hindgut-fermenting hosts. Screening for growth on urea as the sole source of nitrogen provided strong evidence that the urease was active in these strains. These results represent the strongest evidence reported to date for specific functional differences contributing to the ecology of Fibrobacter spp. in the herbivore gut.IMPORTANCE The herbivore gut microbiome is incredibly diverse, and a functional understanding of this diversity is needed to more reliably manipulate this community for specific gain, such as increased production in ruminant livestock. Microbial degraders of plant cell wall polysaccharides in the herbivore gut, particularly Fibrobacter spp., are of fundamental importance to their hosts for digestion of a diet consisting primarily of recalcitrant plant fibers. Considerable phylogenetic diversity exists among members of the genus Fibrobacter, but much of this diversity remains cryptic. Here, we used comparative genomics, applied to a diverse collection of recently isolated Fibrobacter strains, to identify a robust association between carbohydrate-active enzyme gene content and the Fibrobacter phylogeny. Our results provide the strongest evidence reported to date for functional differences among Fibrobacter phylotypes associated with either the rumen or the hindgut and emphasize the general significance of carbohydrate-active enzymes in the evolution of fiber-degrading bacteria.

Keywords: Fibrobacter; carbohydrate-active enzymes; cellulose; fiber; genomics; gut microbiota; herbivores.

FIG 1

A multilocus phylogeny of Fibrobacter strains. The maximum likelihood phylogeny was constructed from the concatenated alignments of 99 essential protein genes using RAxML (68) and is midpoint rooted. Strains are annotated with circles according to their isolation source, rumen (dark green) or hindgut (dark red). Previously assigned Fibrobacter phylotypes are labeled, and their coverage of the tree is marked with black vertical bars (17). Type strains are identified with an asterisk (*). Major clades are shaded according to the following clade designations: clade A, green; clade B, blue; clade C, red; clade D, violet. The bootstrap value for clades containing more than 5 strains was 100%, unless otherwise indicated (100 replicates). The scale represents the number of substitutions per site.

FIG 2

Distribution of cellulase (A) and hemicellulase (B) CAZyme families among Fibrobacter strains. Fibrobacter strains are arranged according to the results of hierarchical clustering by relative abundance of all CAZyme family genes in their respective genomes. Fibrobacter strains are annotated with circles according to phylogenetic clade membership as clade A (green), clade B (blue), clade C (red), and clade D (violet) and according to isolation source as rumen (dark green) or hindgut (dark red). Horizontal bar plots extend along the x axis according to the number of normalized gene counts in each glycoside hydrolase (GH) family. GH families are color coded separately for cellulases (A) and hemicellulases (B).

FIG 3

Scatterplot of 2-dimensional NMDS ordination based on the relative abundances of the genes for all CAZyme families in the Fibrobacter genomes. Individual points represent a single Fibrobacter strain, colored according to the phylogenetic clade, clade A (green), clade B (blue), clade C (red), and clade D (violet). Open standard error ellipses (95% confidence interval) are plotted for each Fibrobacter clade and are colored accordingly. Vectors for CAZyme families were calculated by fitting the relative abundances of those CAZyme families in the Fibrobacter genomes to the dimensions of the NMDS, with the largest significant differences among the clades plotted with gray arrows and labeled accordingly in red.

FIG 4

Significant differences in CAZyme gene counts among Fibrobacter clades. In each plot, a single point represents a single Fibrobacter strain, colored according to phylogenetic clade, clade A (green), clade B (blue), clade C (red), and clade D (violet). Error bars, colored according to clade, represent the 95% confidence interval for the mean normalized gene count for the clade. (A) Strip plot of normalized gene counts for total CAZymes in Fibrobacter genomes by phylogenetic clade. The bracket at the top of the plot represents the results of statistical testing for an overall difference among the clades with significance indicated as follows: •, P < 0.05; *, P < 0.01; **, P < 0.001; ***, P < 0.0001 (ANOVA). Lowercase letters along the bottom of each plot represent significantly different groups based on pairwise statistical tests using comparisons between individual clades (t test, adjusted P value [P_adj] < 0.05 [Bonferroni’s correction]). (B) Strip plots of normalized gene counts for CAZyme classes. The plots are arranged horizontally by CAZyme class and labeled accordingly. Brackets represent the results of statistical testing for an overall difference among the clades for each CAZyme class, with significance indicated as follows: P_adj < 0.05 (·), P_adj < 0.01 (*), P_adj < 0.001 (**), P_adj < 0.0001 (***) (ANOVA, Bonferroni’s correction). Lowercase letters along the bottom of each plot represent significantly different groups based on pairwise statistical tests using comparisons between individual clades (t test, P_adj < 0.05 [Bonferroni’s correction]). (C) Strip plots of normalized gene counts for highly variable CAZyme families. The plots are arranged horizontally by CAZyme family and labeled accordingly. Brackets represent the results of statistical testing for an overall difference among the clades for each CAZyme family with significance indicated as follows: •, FDR < 0.05; *, FDR < 0.01; **, FDR < 0.001; ***, FDR < 0.0001 (Kruskal-Wallis rank sum test [71]). Lowercase letters along the bottom of each plot represent significantly different groups based on pairwise statistical tests using comparisons between individual clades (Wilcoxon rank sum test, P_adj < 0.05 [Bonferroni’s correction]).

FIG 5

Presence of urease in Fibrobacter strains. (A) Maximum likelihood phylogeny constructed from the concatenated alignments of 99 essential protein genes with strains labeled according to the presence of urease genes and growth on urea as a source of nitrogen. Open diamonds (◊) represent the absence of the urease genes and filled diamonds (♦) represent the presence of urease genes in the genome. Open circles (○) indicate that the strain failed to grow in media with urea as the nitrogen source and filled circles (●) indicate that the strain was able to grow in media with urea as the source of nitrogen. (B) Aligned architecture of the approximately 5.7 kb region predicted to encode urease catalytic subunits (blue) and urease accessory proteins (red) in 12 positive strains. Genes not annotated as urease components are colored gray. Genes without labels are annotated as hypothetical proteins.