Predicting functions of putative fungal sesquiterpene synthase genes based on multiomics data analysis

Abstract

Sesquiterpenes (STs) are secondary metabolites, which mediate biotic interactions between different organisms. Predicting the species-specific ST repertoires can contribute to deciphering the language of communication between organisms of the same or different species. High biochemical plasticity and catalytic promiscuity of sesquiterpene synthases (STSs), however, challenge the homology-based prediction of the STS functions. Using integrated analyses of genomic, transcriptomic, volatilomic, and metabolomic data, we predict product profiles for 116 out of 146 putative STS genes identified in the genomes of 30 fungal species from different trophic groups. Our prediction method is based on the observation that STSs encoded by genes closely related phylogenetically are likely to share the initial enzymatic reactions of the ST biosynthesis pathways and, therefore, produce STs via the same reaction route. The classification by reaction routes allows to assign STs known to be emitted by a particular species to the putative STS genes from this species. Gene expression information helps to further specify these ST-to-STS assignments. Validation of the computational predictions of the STS functions using both in silico and experimental approaches shows that integrated multiomic analyses are able to correctly link cyclic STs of non-cadalane type to genes. In the process of the experimental validation, we characterized catalytic properties of several putative STS genes from the mycorrhizal fungus Laccaria bicolor. We show that the STSs encoded by the L.bicolor mycorrhiza-induced genes emit either nerolidol or α-cuprenene and α-cuparene, and discuss the possible roles of these STs in the mycorrhiza formation.

Keywords: Functional prediction; Fungi; Multiomics; Mycorrhiza; Sesquiterpene synthase genes; Sesquiterpenes.