A Bioinformatics Tool for Efficient Retrieval of High-Confidence Terpene Synthases (TPS) and Application to the Identification of TPS in Coffea and Quillaja

Terpenoids are a class of compounds that are found in all living organisms. In plants, some terpenoids are part of primary metabolism, but most terpenes found in plants are classified as specialized metabolites, encoded by terpene synthases (TPS). It is not obvious how to assign the putative product of a given TPS using bioinformatics tools. Phylogenetic analyses easily assign TPS into families; however members of the same TPS family can synthetize more than one terpenoid-and, in many biotechnological applications, researchers are more interested in the product of a given TPS rather than its phylogenetic profile. Automated protein annotation can be used to classify TPS based on their products, despite the family they belong to. Here, we implement an automated bioinformatics method, search_TPS, to identify TPS proteins that synthesize mono, sesqui and diterpenes in Angiosperms. We verified the applicability of the method by classifying wet lab validated TPS and applying it to find TPS proteins in Coffea arabica, C. canephora, C. eugenioides, and Quillaja saponaria. Search_TPS is a computational tool based on PERL scripts that carries out a series of HMMER searches against a curated database of TPS profile hidden Markov models. The tool is freely available at https://github.com/liliane-sntn/TPS .