Improved protein glycosylation enabled heterologous biosynthesis of monoterpenoid indole alkaloids and their unnatural derivatives in yeast

Abstract

With over 3,000 reported structures, monoterpenoid indole alkaloids (MIAs) constitute one of the largest alkaloid groups in nature, including the clinically important anticancer drug vinblastine and its semi-synthetic derivatives from Catharanthus roseus (Madagascar’s periwinkle). With the elucidation of the complete 28-step biosynthesis for anhydrovinblastine, it is possible to investigate the heterologous production of vinblastine and other medicinal MIAs. In this study, we successfully expressed the flavoenzyme O-acetylstemmadenine oxidase in Saccharomyces cerevisiae (baker’s yeast) by signal peptide modification, which is a vinblastine biosynthetic gene that has not been functionally expressed in this system. We also report the simultaneous genomic integration of ∼18 kb MIA biosynthetic gene cassettes as single copies by CRISPR-Cas9 in baker’s yeast, which enabled the biosynthesis of vinblastine precursors catharanthine and tabersonine from the feedstocks secologanin and tryptamine. We further demonstrated the biosynthesis of fluorinated and hydroxylated catharanthine and tabersonine derivatives using our yeasts, which showed that the MIA biosynthesis accommodates unnatural substrates, and the system can be further explored to produce other complex MIAs.With over 3,000 members, monoterpenoid indole alkaloids (MIA) are one of the largest and most diverse alkaloids in nature including many human medicines, such as chemotherapeutics vinblastine from Catharanthus roseus (Madagascar’s periwinkle) and camptothecin from Camptotheca accuminata (happy tree), and antiarrhythmic ajmaline from Rauwolfia serpentina (Indian snakeroot).1 Recent studies have elucidated the complete 28-step biosynthetic pathway for anhydrovinblastine in C. roseus, which involves diverting a primary monoterpene geranyl pyrophosphate into the biosynthesis of secologanin via the iridoid pathway (9 steps), genesis of the first MIA strictosidine that is the universal precursor to almost all MIAs (2 steps), conversion of strictosidine to iboga type MIA catharanthine and aspidosperma type tabersonine (9 steps), decorating tabersonine to vindoline (7 steps), and the final step that couples vindoline and catharanthine to make anhydrovinblastine (Fig. 1). 2-12 These studies not only revealed the remarkable complexity of MIA formations but also enabled the exploration in heterologous production of bioactive MIAs and intermediates that are usually found in low quantities in their natural sources. Notably, strictosidine and a related corynanthe type MIA ajmalicine have been produced de novo in Saccharomyces cerevisiae (baker’s yeast), 13,14 while vindoline has been produced in baker’s yeast from tabersonine feedstock. 3,15,16 For strictosidine production in yeast, the challenges lie in the generally low monoterpene biosynthesis output and the intermediates consumption by yeast native metabolism.13,14,17 While studies did not report rapid MIA consumption by yeast, vindoline yields were improved by optimizing the stoichiometry of cytochrome P450 monooxygenase (CYP), CYP redox partner CYP reductase (CPR), and other factors related with CYP activities such as endoplasmic reticulum (ER) homeostasis and NADPH co-factor regeneration that are commonly exploited.15,16 In this study, we constructed yeast strains containing the remaining vinblastine biosynthetic segment and produced catharanthine and tabersonine by feeding precursors, secologanin and tryptamine, as well as their unnatural derivatives by feeding substituted tryptamine.