Lignocellulose degradation at the holobiont level: teamwork in a keystone soil invertebrate

Abstract

Background: Woodlice are recognized as keystone species in terrestrial ecosystems due to their role in the decomposition of organic matter. Thus, they contribute to lignocellulose degradation and nutrient cycling in the environment together with other macroarthropods. Lignocellulose is the main component of plants and is composed of cellulose, lignin and hemicellulose. Its digestion requires the action of multiple Carbohydrate-Active enZymes (called CAZymes), typically acting together as a cocktail with complementary, synergistic activities and modes of action. Some invertebrates express a few endogenous lignocellulose-degrading enzymes but in most species, an efficient degradation and digestion of lignocellulose can only be achieved through mutualistic associations with endosymbionts. Similar to termites, it has been suspected that several bacterial symbionts may be involved in lignocellulose degradation in terrestrial isopods, by completing the CAZyme repertoire of their hosts.

Results: To test this hypothesis, host transcriptomic and microbiome shotgun metagenomic datasets were obtained and investigated from the pill bug Armadillidium vulgare. Many genes of bacterial and archaeal origin coding for CAZymes were identified in the metagenomes of several host tissues and the gut content of specimens from both laboratory lineages and a natural population of A. vulgare. Some of them may be involved in the degradation of cellulose, hemicellulose, and lignin. Reconstructing a lignocellulose-degrading microbial community based on the prokaryotic taxa contributing relevant CAZymes revealed two taxonomically distinct but functionally redundant microbial communities depending on host origin. In parallel, endogenous CAZymes were identified from the transcriptome of the host and their expression in digestive tissues was demonstrated by RT-qPCR, demonstrating a complementary enzyme repertoire for lignocellulose degradation from both the host and the microbiome in A. vulgare.

Conclusions: Our results provide new insights into the role of the microbiome in the evolution of terrestrial isopods and their adaptive radiation in terrestrial habitats.

Keywords: CAZyme; Holobiont; Host–symbiont interactions; Isopods; Microbiome; RT-qPCR; Transcriptome.

Fig. 1

Model for lignocellulose degradation in the A. vulgare holobiont. Diagrams represent the CAZy families contributed by the host (red) and the microbiome (blue). (I) Lignin would be partially degraded to release cellulose and hemicellulose. (II) Cellulose would be degraded by the action of endoglucanases and β-glucosidases. A high number of β-glucosidases and mechanical fragmentation by A. vulgare could compensate for the lack of exoglucanases. (III) The A. vulgare holobiont could degrade most types of hemicellulose due to the high diversity of Debranching enzymes* (CE1, CE3, CE4, CE5, CE6, CE7, CE12, GH3, GH4, GH43, GH51), Endo-hemicellulases* (GH5, GH8, GH9, GH10, GH11, GH16, GH30, GH43, GH51, GH53, GH74, GH113, GH128, GH134), and Exo-hemicellulases* (GH1, GH2, GH3, GH4, GH5, GH27, GH29, GH30, GH31, GH35, GH36, GH39, GH42, GH43, GH51, GH57, GH116, GH120)

Fig. 2

Prediction of enzymatic functions (EC number) of debranching enzymes (DE), endo-hemicellulases (Endo), exo-hemicellulases (Exo), cellulases, and lignin modifying enzymes (LMEs) identified in the metagenomes of specimens from the field and the laboratory and in the host transcriptome. Relative abundance (in %) for a given predicted enzymatic function was calculated by dividing the identified counts for a given enzyme by the total counts identified in a metagenome or in the transcriptome

Fig. 3

Quantitative RT-PCR analysis of the expression of representative host lignocellulose-degrading CAZymes in caeca (C), gut content (GC), hindgut (HG), and non-digestive tissues (T). Transcripts with the highest RPKM value were chosen to represent each family of interest. Expression of each gene was normalized based on the expression of Ribosomal Protein L8 (RbL8) and Elongation Factor 2 (EF2) as reference genes. Different letters indicate statistically significant differences (p < 0.05) after Kruskal–Wallis rank sum test

Fig. 4

Relative abundance of prokaryotic taxa contributing lignocellulose-degrading CAZymes depending on a host origin, gender, and Wolbachia infection status, b host origin alone, and c for several genes consistently present in isopods of both field and laboratory origin. See Table 3 for a detailed annotation of these genes

Fig. 5

Lignocellulose-degrading enzymes and their associated microbial community in (a) isopods from the laboratory and (b) isopods from a natural population. c The microbial taxa contributing lignocellulose-binding modules in both field and laboratory specimens