Abstract. An important group of antimalarial drugs consists of the endoperoxide sesquiterpene lactone artemisinin and its semi-synthetically prepared derivatives. Because of the development of resistance against other malarial drugs, the demand for artemisinin has rapidly increased during the past decade. However, the supply of artemisinin is troublesome as neither total nor semi-synthesis are economically feasible and the only plant species known to produce artemisinin, Artemisia annua L., contains only low amounts of this compound. For a directed strategy to improve artemisinin production, one of the first requirements is to have insight into the regulation of its biosynthesis. Surprisingly, until some years ago hardly anything was known about the biosynthesis of artemisinin, particularly about the early enzymatic steps leading to (dihydro)artemisinic acid. To elucidate this important missing part in the pathway, we have analysed the terpenoids in A. annua leaves and the presence of sesquiterpene synthases, the enzyme class that should theoretically catalyse the first step in the formation of artemisinin. This led to the discovery of amorpha-4,11-diene synthase, which indeed catalyses the first step in artemisinin biosynthesis and for which the corresponding gene was cloned shortly afterwards. A subsequent search for oxidized derivatives of amorpha-4,11-diene -which could be postulated to be intermediates en route to artemisinin -revealed the presence of artemisinic alcohol, dihydroartemisinic alcohol, artemisinic aldehyde, dihydroartemisinic aldehyde and dihydroartemisinic acid. We also demonstrated the presence of the biosynthetic enzymes amorpha-4,11-diene hydroxylase, artemisinic alcohol dehydrogenase, artemisinic aldehyde reductase and dihydroartemisinic aldehyde dehydrogenase in A. annua. From these results, we conclude that the early steps in artemisinin biosynthesis involve: cyclization of farnesyl diphosphate by amorpha-4,11-diene synthase to amorpha-4,11-diene, hydroxylation of amorpha-4,11-diene to artemisinic alcohol, oxidation of artemisinic alcohol to artemisinic aldehyde, reduction of the C11-C13 double bond yielding dihydroartemisinic aldehyde, and oxidation of dihydroartemisinic aldehyde to dihydroartemisinic acid. During this biochemical research we have also managed to develop a trichome isolation protocol which has not only accelerated our biochemical work but also enables us to clone -in a very directed way -the genes that encode the enzymes responsible for artemisinin production. This should in the next few years result in several approaches to boost artemisinin production.
276H. BOUWMEESTER ET AL.