Functional and phylogenetic analyses of camel rumen microbiota associated with different lignocellulosic substrates

Abstract

Rumen microbiota facilitates nutrition through digestion of recalcitrant lignocellulosic substrates into energy-accessible nutrients and essential metabolites. Despite the high similarity in rumen microbiome structure, there might be distinct functional capabilities that enable different ruminant species to thrive on various lignocellulosic substrates as feed. Here, we applied genome-centric metagenomics to explore phylogenetic diversity, lignocellulose-degrading potential and fermentation metabolism of biofilm-forming microbiota colonizing 11 different plant substrates in the camel rumen. Diversity analysis revealed significant variations in the community of rumen microbiota colonizing different substrates in accordance with their varied physicochemical properties. Metagenome reconstruction recovered genome sequences of 590 bacterial isolates and one archaeal lineage belonging to 20 microbial phyla. A comparison to publicly available reference genomes and rumen metagenome-assembled genomes revealed that most isolates belonged to new species with no well-characterized representatives. We found that certain low abundant taxa, including members of Verrucomicrobiota, Planctomycetota and Fibrobacterota, possessed a disproportionately large number of carbohydrate active enzymes per Mb of genome, implying their high metabolic potential to contribute to the rumen function. In conclusion, we provided a detailed picture of the diversity and functional significance of rumen microbiota colonizing feeds of varying lignocellulose composition in the camel rumen. A detailed analysis of 591 metagenome-assembled genomes revealed a network of interconnected microbiota and highlighted the key roles of certain taxonomic clades in rumen function, including those with minimal genomes (e.g., Patescibacteria). The existence of a diverse array of gene clusters encoding for secondary metabolites unveiled the specific functions of these biomolecules in shaping community structure of rumen microbiota.

Fig. 1. Forage physicochemical properties and the diversity of their attached microbiota during incubation in the camel rumen.

Forages degradability was evaluated by measuring (A) neutral detergent fiber (NDF), (B) acid detergent fiber (ADF) and (C) acid detergent lignin (ADL) concentrations during incubation in the rumen. Concentrations are presented based on percentage of dry matter. Data points are mean and error bars represent standard errors. Beta diversity of camel rumen microbiota attached to the forages was explored by searching reads against ~1.1 M lineage-specific marker genes using MetaPhlAn. Principal coordinate analysis (PCoA) plot shows community diversity of rumen microbiota attached to 11 different forages incubated in the camel rumen based on (D) weighted UniFrac and (E) Bray_Curtis dissimilarity matrices. Samples are grouped based on feeds and each feed is represented by a unique shape and color.

Fig. 2. Taxonomic classification of RUIs and their sequence comparison with previously reported rumen cultured isolates and metagenome-assembled genomes (MAGs).

A The tree shows phylogenetic relationship of 591 RUIs taxonomically classified at the phylum level. Taxonomies were assigned based on comparison with ~40,000 reference genomes from Genome Taxonomy Database (GTDB) using GTDB-Tk. Average nucleotide identities (ANIs) along with aligned fractions (AFs) of the RUIs against rumen cultured isolates (n = 408) sequenced by Hungate1000 genome project (B), rumen metagenome-assembled genome (MAGs) from Scottish cattle (n = 4941) (C), Taleshi cattle (n = 538) (D) and camel (n = 65) (E). An AF of 65% was considered as cutoff for correct species delineation.

Fig. 3. Beta diversity analysis of the rumen uncultured isolates (RUIs) reconstituted from metagenome-assembled contigs.

A PCoA plot showing significant separation of forages in a two-dimensional space based on community diversity (weighted UniFrac dissimilarity matrix) of RUIs reconstituted from metagenomes of forage-attached microbiota. Each feed is represented by a unique shape and color. Significant differences in community compositions were tested using permutation multivariate analysis of variance (PERMANOVA). B The homogeneity of group dispersion between forages was also tested using permutational multivariate analysis of dispersion (PERMDISP). C RUIs abundance profiles (phylum level) were correlated with forage degradation indices (i.e., NDF, ADF and ADL) using redundancy analysis (RDA). The first two axis explaining the highest variations are displayed. RDA biplot was created using PAST software. D Significant correlations of genome bin abundances (grouped at taxonomic level of order) with forage physicochemical properties. Correlations with r > 0.38 or r < −0.38 and P-value < 0.01 were considered statistically significant.

Fig. 4. The profile of carbohydrate active enzymes (CAZymes) predicted in the RUIs and their sequence comparison with public databases.

Heatmap shows the abundance profile of CAZymes in RUIs classified at the phylum level (A). The sequences of the predicted CAZymes were searched against non-redundant (nr) protein database (B) and dbCAN database (C). GHs, glycoside hydrolases; GTs, glycosyl transferases; PLs, polysaccharide lyases; CEs, carbohydrate esterases; AAs, auxiliary activities; and CBs, carbohydrate binding modules. The boxes represent the median (center line) and interquartile ranges, red triangle represents mean and whiskers extend to one and a half times of the interquartile range.

Fig. 5. Metabolic characterization of RUIs.

RUIs were characterized with respect to their abilities for major polysaccharides degradation, sugar utilization and fermentation. RUIs were grouped into class level taxonomic clades based on taxonomies inferred by GTDB-Tk. Metabolic pathways were reconstructed based on the presence of key marker genes of each metabolic pathway using KOfam annotations. The details of Hidden Marcov Model (HMM) profiles searched are presented in Supplementary Data 3.