Unveiling lignocellulolytic potential: a genomic exploration of bacterial lineages within the termite gut

Abstract

Background: The microbial landscape within termite guts varies across termite families. The gut microbiota of lower termites (LT) is dominated by cellulolytic flagellates that sequester wood particles in their digestive vacuoles, whereas in the flagellate-free higher termites (HT), cellulolytic activity has been attributed to fiber-associated bacteria. However, little is known about the role of individual lineages in fiber digestion, particularly in LT.

Results: We investigated the lignocellulolytic potential of 2223 metagenome-assembled genomes (MAGs) recovered from the gut metagenomes of 51 termite species. In the flagellate-dependent LT, cellulolytic enzymes are restricted to MAGs of Bacteroidota (Dysgonomonadaceae, Tannerellaceae, Bacteroidaceae, Azobacteroidaceae) and Spirochaetota (Breznakiellaceae) and reflect a specialization on cellodextrins, whereas their hemicellulolytic arsenal features activities on xylans and diverse heteropolymers. By contrast, the MAGs derived from flagellate-free HT possess a comprehensive arsenal of exo- and endoglucanases that resembles that of termite gut flagellates, underlining that Fibrobacterota and Spirochaetota occupy the cellulolytic niche that became vacant after the loss of the flagellates. Furthermore, we detected directly or indirectly oxygen-dependent enzymes that oxidize cellulose or modify lignin in MAGs of Pseudomonadota (Burkholderiales, Pseudomonadales) and Actinomycetota (Actinomycetales, Mycobacteriales), representing lineages located at the hindgut wall.

Conclusions: The results of this study refine our concept of symbiotic digestion of lignocellulose in termite guts, emphasizing the differential roles of specific bacterial lineages in both flagellate-dependent and flagellate-independent breakdown of cellulose and hemicelluloses, as well as a so far unappreciated role of oxygen in the depolymerization of plant fiber and lignin in the microoxic periphery during gut passage in HT. Video Abstract.

Keywords: CAZymes; Cellulase; Functional genomics; Lignin; Lignocellulose degradation; Termite microbiota.

Fig. 1

Distribution of MAGs from major bacterial phyla among metagenomes from different host groups. The ordinate shows the number of MAGs in each phylum as a heatmap that is based on the total number of MAGs (2,223) in the dataset. The abscissa shows the number of metagenomes from which the MAGs were recovered, summarized for each host group, as a heatmap based on the total number of metagenomes (51). Relative abundance values indicate the proportion of reads in a metagenome that mapped against the corresponding MAGs, expressed as averages per termite family (lower termites) or subfamily (higher termites, Termitidae). The termite species contained in each host group, the relative abundances of individual MAGs, and their classification down to genus level are given in Tables S1 and S2

Fig. 2

The distribution of carbohydrate-active enzymes with lignocellulolytic activities (cellulase, hemicellulase, auxiliary) and cellulose-binding modules (CBM) among the MAGs of major bacterial families recovered from termite gut metagenomes. The heatmap indicates the mean gene abundance in selected CAZyme families for the MAGs from the respective bacterial family. CAZYme families were clustered using a hierarchical clustering algorithm; values were scaled from 0 to 1 to facilitate visualization. An interactive spreadsheet with details for all MAGs is given in Table S4; the complete dataset is given in Table S5

Fig. 3

Ordination analysis of CAZyme composition in 2,204 MAGs of termite-associated bacteria. The UMAP analysis (stress = 0.12) used the entire dataset of CAZymes (256,212 genes across 534 CAZyme families; see Table S5) to visualize differences in the general content of CAZymes among phyla. The biplot displays MAGs color-coded according to their respective phyla (A) and according to the termite (sub)families from which they were recovered (B). The theoretical distances were calculated based on the KNN score, and the statistics are based on ANOSIM (bacterial phyla R = 0.34; termite (sub)families R = 0.04; p < 0.001) and ADONIS (bacterial phyla R² = 0.24; termite (sub)families R² = 0.03; p < 0.001) tests

Fig. 4

Gene density (GD) of cellulases (A) and hemicellulases (B) in the termite-associated MAGs of different bacterial phyla, and the relationships between the respective GDs and the relative abundance (RA) of the MAGs in the metagenomes of lower termites and higher termites (C, D). The number of (hemi)cellulolytic genes of a given MAG is normalized by its total number of predicted genes and expressed as percentages. Statistical tests were performed both globally and pairwise, against the mean values (dashed line). Significance values: ns = non-significant; * ≤ 0.05; ** ≤ 0.01; *** ≤ 0.001; **** ≤ 0.0001. Linear regressions are shown for all phyla but a positive correlation (Spearman ϱ > 0.1 and p < 0.05) was present only for LT-associated MAGs of Bacteroidota and HT-associated MAGs of Fibrobacterota, Bacteroidota, Spirochaetota, Bacillota and Actinomycetota

Fig. 5

Abundance and proportion of selected CAZymes from GH families with cellulolytic (A) and hemicellulolytic (B) functions in the MAGs of selected bacterial families with a high fibrolytic potential. Each bar shows the total number of CAZymes classified as cellulase or hemicellulase that are encoded by a particular MAG (gray and black color indicates the origin from lower or higher termites) and the proportion of genes in the respective GH families (visualized with different colors). Only MAGs that encode more than one cellulase or hemicellulase were included. The families GH5 and GH30 appear twice but refer to different subfamilies classified as cellulases or hemicellulases (see Table S3 for details)

Fig. 6

Principal component analysis (PCA) of the abundance of cellulases (A) and hemicellulases (B) in the MAGs from selected bacterial families with a high fibrolytic potential (see Fig. 5). The biplots show MAGs color-coded either according to bacterial phylum (left) or according to their respective host group (lower or higher termite; right)

Fig. 7

Distribution of gene densities in CAZymes with lignin-modifying activities in MAGs from different bacterial phyla. The insets break down selected phyla to the order level and show the proportion in percentages of genes from different enzyme families: POL (phenol-oxidizing laccases, AA1), LMP (lignin-modifying peroxidases, AA2), and AAO (aryl alcohol oxidases, AA3_2)