Overcoming Bottlenecks for Metabolic Engineering of Sesquiterpene Production in Tomato Fruits

Abstract

Terpenoids are a large and diverse class of plant metabolites that also includes volatile mono- and sesquiterpenes which are involved in biotic interactions of plants. Due to the limited natural availability of these terpenes and the tight regulation of their biosynthesis, there is strong interest to introduce or enhance their production in crop plants by metabolic engineering for agricultural, pharmaceutical and industrial applications. While engineering of monoterpenes has been quite successful, expression of sesquiterpene synthases in engineered plants frequently resulted in production of only minor amounts of sesquiterpenes. To identify bottlenecks for sesquiterpene engineering in plants, we have used two nearly identical terpene synthases, snapdragon (Antirrhinum majus) nerolidol/linalool synthase-1 and -2 (AmNES/LIS-1/-2), that are localized in the cytosol and plastids, respectively. Since these two bifunctional terpene synthases have very similar catalytic properties with geranyl diphosphate (GPP) and farnesyl diphosphate (FPP), their expression in target tissues allows indirect determination of the availability of these substrates in both subcellular compartments. Both terpene synthases were expressed under control of the ripening specific PG promoter in tomato fruits, which are characterized by a highly active terpenoid metabolism providing precursors for carotenoid biosynthesis. As AmNES/LIS-2 fruits produced the monoterpene linalool, AmNES/LIS-1 fruits were found to exclusively produce the sesquiterpene nerolidol. While nerolidol emission in AmNES/LIS-1 fruits was 60- to 584-fold lower compared to linalool emission in AmNES/LIS-2 fruits, accumulation of nerolidol-glucosides in AmNES/LIS-1 fruits was 4- to 14-fold lower than that of linalool-glucosides in AmNES/LIS-2 fruits. These results suggest that only a relatively small pool of FPP is available for sesquiterpene formation in the cytosol. To potentially overcome limitations in sesquiterpene production, we transiently co-expressed the key pathway-enzymes hydroxymethylglutaryl-CoA reductase (HMGR) and 1-deoxy-D-xylulose 5-phosphate synthase (DXS), as well as the regulator isopentenyl phosphate kinase (IPK). While HMGR and IPK expression increased metabolic flux toward nerolidol formation 5.7- and 2.9-fold, respectively, DXS expression only resulted in a 2.5-fold increase.