Phylogenetic systematics of Butyrivibrio and Pseudobutyrivibrio genomes illustrate vast taxonomic diversity, open genomes and an abundance of carbohydrate-active enzyme family isoforms

Abstract

Butyrivibrio and Pseudobutyrivibrio dominate in anaerobic gastrointestinal microbiomes, particularly the rumen, where they play a key role in harvesting dietary energy. Within these genera, five rumen species have been classified ( Butyrivibrio fibrisolvens , Butyrivibrio hungatei , Butyrivibrio proteoclasticus , Pseudobutyrivibrio ruminis and Pseudobutyrivibrio xylanivorans ) and more recently an additional Butyrivibrio sp. group was added. Given the recent increase in available genomes, we re-investigated the phylogenetic systematics and evolution of Butyrivibrio and Pseudobutyrivibrio . Across 71 genomes, we show using 16S rDNA and 40 gene marker phylogenetic trees that the current six species designations ( P. ruminis , P. xylanivorans , B. fibrisolvens , Butyrivibrio sp., B. hungatei and B. proteclasticus) are found. However, pangenome analysis showed vast genomic variation and a high abundance of accessory genes (91.50–99.34 %), compared with core genes (0.66–8.50 %), within these six taxonomic groups, suggesting incorrectly assigned taxonomy. Subsequent pangenome accessory genomes under varying core gene cut-offs (%) and average nucleotide identity (ANI) analysis suggest the existence of 42 species within 32 genera. Pangenome analysis of those that still group within B. fibrisolvens , B. hungatei and P. ruminis , based on revised ANI phylogeny, also showed possession of very open genomes, illustrating the diversity that exists even within these groups. All strains of both Butyrivibrio and Pseudobutyrivibrio also shared a broad range of clusters of orthologous genes (COGs) (870), indicating recent evolution from a common ancestor. We also demonstrate that the carbohydrate-active enzymes (CAZymes) predominantly belong to glycosyl hydrolase (GH)2, 3, 5, 13 and 43, with numerous within family isoforms apparent, likely facilitating metabolic plasticity and resilience under dietary perturbations. This study provides a major advancement in our functional and evolutionary understanding of these important anaerobic bacteria.

Keywords: Butyrivibrio; Pseudobutyrivibrio; evolution; pangenome; rumen; taxonomy.

Fig. 1.

Chronological identification and classification of Butyrivibrio and Pseudobutyrivibrio .

Fig. 2.

Functional annotation of the 71 Butyrivibrio and Pseudobutyrivibrio genomes used in this study. Gene functionality is sorted by colour, as indicated in the key. Annotation was performed using EggNOG [44] (http://eggnogdb.embl.de/#/app/emapper).

Fig. 3.

ANI comparison of 71 strains of Butyrivibrio and Pseudobutyrivibrio using pyani.py (https://github.com/widdowquinn/pyani/tree/version_0_2) with the MUMmer alignment option (a) (cells in the heat map that are coloured red have >95 % sequence similarity, whilst blue cells have <95 % similarity, and as nucleotide identity reaches 95 % the cells are coloured white). Alignment coverage of all strains using pyani.py with the MUMmer alignment option (b) (cells in the heat map that are coloured red have >50 % coverage, whilst blue cells have <50 % similarity, and as nucleotide identity reaches 95 % the cells are coloured white).

Fig. 4.

Effect of core definition percentage (a) and gene similarity homology (b) cut-off percentage on the pangenome composition of current Butyrivibrio and Pseudobutyrivibrio taxa as denoted by coloured lines: green being B. fibrisolvens (BF), purple being B. hungatei (BH), blue being B. proteoclasticus (BP), orange being Butyrivibrio sp. (Bsp), red being P. ruminis (PR) and grey being P. xylanivorans (PX).

Fig. 5.

Functional annotations of the core and accessory genomes of the ANI-defined B. fibrisolvens , B. hungatei and P. ruminis groups.

Fig. 6.

UpSet plot showing orthologous gene cluster intersections across the two genera, Butyrivibrio (B) and Pseudobutyrivibrio (P). Intersections are denoted by the corresponding bar.

Fig. 7.

Phylogenetic tree showing the relatedness of all GH family 2 genes found in all 71 strains used in this analysis. presence of a black square on the outermost layer indicates that the gene was found to be present in the Shi et al. metatranscriptome dataset [55]. The tree is rooted using a β-galactosidase large subunit sequence from L. acidophilus NCFM, which is coloured in black. Tree scale indicates number of substitutions per site.

Fig. 8.

Phylogenetic tree showing the relatedness of all GH family 3 genes found in all 71 strains used in this analysis. The presence of a black square on the outermost layer indicates that the gene was found to be present in the Shi et al. metatranscriptome dataset [55]. The tree is rooted using a β-N-acetylhexosaminidase sequence from L. acidophilus NCTC13720, which is coloured in black.

Fig. 9.

Phylogenetic tree showing the relatedness of all GH family 5 genes found in all 71 strains used in this analysis. The presence of a black square on the outermost layer indicates that the gene was found to be present in the Shi et al. metatranscriptome dataset [55]. The tree is rooted using a β-glucosidase sequence from Lactobacillus mucosae LM1, which is coloured in black.

Fig. 10.

Phylogenetic tree showing the relatedness of all GH family 13 genes found in all 71 strains used in this analysis. The presence of a black square on the outermost layer indicates that the gene was found to be present in the Shi et al. metatranscriptome dataset [55]. The tree is rooted using a sucrose phosphorylase sequence from L. acidophilus NCFM, which is coloured in black.

Fig. 11.

Phylogenetic tree showing the relatedness of all GH family 43 genes found in all 71 strains used in this analysis. The presence of a black square on the outermost layer indicates that the gene was found to be present in the Shi et al. metatranscriptome dataset [55]. The tree is rooted using a β-xylosidase sequence from Lactobacillus mucosae LM1, which is coloured in black.