Predictive engineering of class I terpene synthases using experimental and computational approaches

Terpenoids are a highly diverse group of natural products with considerable industrial interest. Increasingly, engineered microbes are used for the production of terpenoids to replace natural extracts and chemical synthesis. Terpene synthases (TSs) show a high level of functional plasticity and are responsible for the vast structural diversity observed in natural terpenoids. Their relatively inert active sites guide intrinsically reactive linear carbocation intermediates along one of many cyclisation paths via exertion of subtle steric and electrostatic control. Due to the absence of a strong protein interaction with these intermediates, there is a remarkable lack of sequence-function relationship within the TS family, making product-outcome predictions from sequences alone challenging. This, in combination with the fact that many TSs produce multiple products from a single substrate hampers the design and use of TSs in the biomanufacturing of terpenoids. This review highlights recent advances in genome mining, computational modelling, high-throughput screening, and machine-learning that will allow more predictive engineering of these fascinating enzymes in the near future.