Cardenolides, which are used for the treatment of congestive heart failure, are still produced by extraction from plants. A more efficient way to produce cardenolides could be achieved by a yeast-based metabolic engineering approach [1]. However, several steps in cardenolide biosynthesis still need to be elucidated.
Malonyl-coenzymeA: 21-hydroxypregnane 21-O-Malonyltransferase (21MaT) catalyzes the formation of pregnane-21-O-malonyl hemiesters from pregnane precursors. Even though a 21MaT from D. purpurea has already been characterized [2], no 21MaT gene could be identified as yet.
Only recently, the BAHD-type malonyltransferase pMaT1, an enzyme from Arabidopsis thaliana, was found to possess 21MaT activity in vitro. Initially, the identification of an AtpMaT1 homologue in Digitalis failed. Therefore, a function-based search for malonyltransferases was performed. One sequence from the D. purpurea transcriptome [3], annotated as quercetin 3-O-glucoside-6ʹ’-O-malonyltransferase (DpQGMaT1), appeared to be a promising candidate. Primers were deduced and D. lanata mRNA was used as a template to isolate the respective D. lanata homologues. The attempt yielded a putative Dl21MaT1 sequence which was 94% identical to DpQGMaT1. Meanwhile, we identified and cloned three more promising candidate genes from D. lanata. Two recombinant Dl21MaTs were already tested for enzyme activity, using various pregnane substrates, such as 3β-O-acetyl-5β-pregnane-14β,21-diol-20-one, 5β-pregnane-21-ol,3,20-dione, and 21-hydroxyprogesterone. All substrates were accepted. However, enzyme activity was very low (1 %) compared to the recombinant AtpMaT1. Since 21MaT activity is higher in crude protein extracts from leaves of D. lanata than in A. thaliana, we cloned two more Dl21MaT genes that will now be characterized with regard to their substrate preferences.
