The benzylisoquinoline alkaloid (BIA) family of tetrahydroisoquinolines (THIQs) comprises >2,500 members, including the pharmaceuticals morphine, codeine, and papaverine, as well as the antibiotics sanguinarine and chelerythrine. Agricultural cultivation can supply the demand for the BIAs that accumulate in plants, but broader access to the BIA family would facilitate additional research and commercialization. Microbial synthesis presents an attractive option due to cheap feedstock, genetic tractability, and ease of scale-up. Previously, we reported titers of the branch-point BIA (S)-reticuline of 4.6 g/L in yeast, which was achieved through leveraging the Ehrlich pathway 2-oxoacid decarboxylase Aro10 to generate the intermediate 4-hydroxyphenylacetaldehyde (4-HPAA). Here, we establish a superior route to (S)-reticuline by switching the pathway intermediate from 4-HPAA to 3,4-dihydroxyphenylacetaldehyde (3,4-dHPAA) using monoamine oxidase A (MAO). The resulting (S)-norlaudanosoline route to (S)-reticuline synthesis is more selective, resolving prior issues with off-pathway THIQs synthesized due to cascading enzyme promiscuity, and more efficient, enabling titers of 4.8 g/L (S)-reticuline while improving yields by over 40%, from 17 to 24 mg/g sucrose in fed-batch fermentations. Finally, we extend de novo (S)-reticuline synthesis to dihydrosanguinarine, achieving 635 mg/L dihydrosanguinarine and sanguinarine in fed-batch fermentation, the highest reported titer of these BIAs by a factor of 40.
Narcross et al. (Wed,) studied this question.