Comparative genomics unravels a rich set of biosynthetic gene clusters with distinct evolutionary trajectories across fungal species farmed by termites

Abstract The use of compounds produced by hosts or symbionts for defence against antagonists has been identified in many organisms, including in fungus-farming termites (Macrotermitinae). The obligate mutualistic fungus Termitomyces plays a central role in the symbiosis through plant biomass decomposition and as the main food source for these termites. Several specialised (secondary) metabolites have been isolated from different Termitomyces species, suggesting that they may also aid in antimicrobial defence. Yet, we have a fragmented understanding of Termitomyces’ natural product repertoire. To determine the biochemical potential encoded by diverse Termitomyces species, we comparatively analysed 22 published and 17 newly generated genomes, spanning 21 of 52 described Termitomyces species and five of the 11 termite host genera. After extensive assembly and annotation optimisation, we employed fungiSMASH to detect 754 biosynthetic gene clusters (BGCs) coding for specialised metabolites. BiG-SCAPE analysis and manual curation allowed us to assign 660 of these BGCs to 61 distinct biosynthetic gene cluster families (GCFs), spanning five compound classes. Seven GCFs were shared by all 21 Termitomyces species, 21 GCFs were present in all genomes of several subsets of species, while the remaining 33 GCFs were inconsistently distributed across species. The 25 most abundant GCFs were subjected to codon-based evolutionary constraint analyses to evaluate their evolutionary histories and revealed two GCFs with consistent positive selection in the same gene across the phylogeny and seventeen genes with Termitomyces species-specific episodic positive selection. These patterns of selection indicate that millions of years of termite-fungus symbiosis have led to distinct evolutionary trajectories of biosynthetic gene clusters, ample putative chemical novelties, and uncover a vast non-random and largely unknown chemical potential of Termitomyces.