Elucidation of the sesquiterpene lactone biosynthetic pathway in feverfew (Tanacetum parthenium)

Parthenolide is the major bioactive compound of feverfew and has anti-inflammatory and anti-cancer activity. Chapter 1gives an overview of the history and current status of research on parthenolide in feverfew. As a promising anti-cancer drug, parthenolide has attracted a lot of attention from medical institutes and companies. A search with ‘parthenolide’ in Google patents yields more than 2000 hits on extraction of parthenolide or its use in treating cancer or other diseases. However, information on the parthenolide biosynthetic pathway is scarce. Elucidation of the full pathway to parthenolide would open up new opportunities for production of this compound in heterologous, more efficient production platforms. To elucidate the biosynthetic pathway of parthenolide, knowledge on the tissue(s) in which parthenolide is produced and stored is important. In Chapter 2, parthenolide was found to highly accumulate particularly in floral trichomes, suggesting that this is also the preferred site of biosynthesis. These floral trichomes were subsequently used to isolate germacrene A synthase (TpGAS), the gene encoding the first dedicated step in parthenolide biosynthesis, using a degenerate primer PCR approach. The transcript level of TpGASwas indeed highest in glandular trichomes. The high expression of TpGASin glandular trichomes which also contain the highest concentration of parthenolide, supports the assumption that glandular trichomes are the organ where parthenolide biosynthesis and accumulation occur. During my work on Chapter 2, a Canadian group reported a germacrene A oxidase (GAO) from lettuce. As the 454 cDNA library of feverfew trichomes was not available yet, we decided to use a 454 cDNA library of chicory (which also produces costunolide) to continue screening candidate genes involved in the next step of the parthenolide biosynthetic pathway, costunolide synthase (COS). In Chapter 3, four P450s (belonging to the CYP71 group) were selected from the chicory cDNA library for functional characterisation in yeast. One of them, named CYP71BL3, was found to be costunolide synthase, and can catalyse the oxidation of germacra-1(10),4,11(13)-trien-12-oic acid to yield costunolide. The biosynthetic pathway of costunolide was reconstituted in Nicotiana benthamiana by transient expression (agro-infiltration) ofTpGAS, CiGAO(which we also identified in the chicory library) and CiCOS, whichresulted in costunolide production of up to 60 ng.g-1FW. In addition, two new compounds were formed that were identified as costunolide-glutathione and costunolide-cysteine conjugates. When the 454 sequences of the feverfew trichome library became available, we continued to identify additional genes involved in the biosynthetic pathway of parthenolide. In Chapter 4, the parthenolide biosynthetic pathway was elucidated by isolating all the structural genes from feverfew, TpGAS, TpGAO, TpCOS and TpPTS. Moreover, the whole pathway was reconstituted in N. benthamiana, through transient expression. In the agro-infiltrated plants, parthenolide as well as a number of conjugates (with cysteine and glutathione) were produced. In an anti-cancer bioassay, these relatively polar conjugates were highly active against colon cancer cells, with only slightly lower activity than free parthenolide. Finally, also a gene encoding a costunolide and parthenolide 3β-hydroxylase was identified, which could potentially be used in biotechnological applications to produce hydroxylated parthenolide. The conjugation and hydroxylation of parthenolide open up new options to improve the water solubility of parthenolide and therefore its potential as a drug. Besides genes involved in the biosynthetic pathway of parthenolide, we also identified two other P450 genes that can utilize costunolide as substrate. In Chapter 5, Tp8879 is identified. Tp8879 can cyclise the monocyclic germacranolide sesquiterpene lactone costunolide to form the bicyclic guaianolide sesquiterpene lactone kauniolide, and is hence called kauniolide synthase. The biosynthetic pathway of kauniolide was reconstituted in N. benthamiana, through transient expression. This thesis combines a series of existing and new technologies for gene discovery – transcriptomics and metabolomics - as well as optimisation of plant metabolic engineering – using transient expression in N. benthamiana- and reports on novel combinatorial biochemistry occurring in metabolic engineering of heterologous plant hosts, resulting in novel sesquiterpene lactone derivatives with the potential to be new drug leads. The use of transient expression and metabolomics for unexpected product identification are technologies that will be of great value to others working in the field of metabolic engineering. The strategies for identification and characterization of candidate genes, the strategies and tools for metabolic engineering and the possibilities to further improve pathway metabolic engineering are discussed in Chapter 6.