The structural diversity of terpenoids is limited by the isoprene rule which states that all primary terpene synthase products derive from methyl-branched building blocks with five carbon atoms. With this study we discover a broad spectrum of novel terpenoids with eleven carbon atoms as byproducts of bacterial 2-methylisoborneol or 2-methylenebornane synthases. Both enzymes use 2-methyl-GPP as substrate, which is synthesized from GPP by the action of a methyltransferase. We used E. coli strains that heterologously produce different C11-terpene synthases together with the GPP methyltransferase and the mevalonate pathway enzymes. With this de novo approach, 35 different C11-terpenes could be produced. In addition to eleven known compounds, it was possible to detect 24 novel C11-terpenes which have not yet been described as terpene synthase products. Four of them, 3,4-dimethylcumene, 2-methylborneol and the two diastereomers of 2-methylcitronellol could be identified. Furthermore, we showed that an E. coli strain expressing the GPP-methyltransferase can produce the C16-terpene 6-methylfarnesol which indicates the condensation of 2-methyl-GPP and IPP to 6-methyl-FPP by the E. coli FPP-synthase. Our study demonstrates the broad range of unusual terpenes accessible by expression of GPP-methyltransferases and C11-terpene synthases in E. coli and provides an extended mechanism for C11-terpene synthases.

Terpenes constitute the largest class of natural products with more than 70 000 compounds. Many different terpenes find applications in the flavor and fragrance industry or can be used as fine chemicals or drugs. In some bacteria, noncanonical terpenes with 11 carbon atoms are synthesized via a GPP-C2-methyltransferase and the subsequent conversion of 2-methyl-GPP by certain terpene synthases into mainly 2-methylisoborneol and 2methylenebornane. Many other C 11 -terpenes were reported as side products, but they are synthesized only in minor amounts by the bacterial C 11 -terpene biosynthesis pathway. To enable biotechnological synthesis of these largely unexplored natural products, we changed the product selectivity of the 2methylenebornane synthase from Pseudomonas f luorescens by a semirational protein engineering approach. Active site amino acids with impact on the product selectivity were identified and variants with completely altered product spectra could be identified and characterized. The gathered data provide new insights into the structure−function relationship for C 11 -terpene synthases and demonstrate the production of formerly inaccessible noncanonical terpenes.