AbstractNatural products from plants are useful as lead compounds in drug discovery. Plant benzylisoquinoline alkaloids (BIAs) exhibit various pharmaceutical activities. Although unidentified BIAs are expected to be of medicinal value, sufficient quantities of such BIAs, for biological assays, are sometimes difficult to obtain due to their low content in natural sources. Here, we showed that high productivity of BIAs in engineered Escherichia coli could be exploited for drug discovery. First, we improved upon the previous microbial production system producing (S)-reticuline, an important BIA intermediate, to obtain yields of around 160 mg/L, which was 4-fold higher than those of the previously reported highest production system. Subsequently, we synthesised non-natural BIAs (O-sulphated (S)-reticulines) by introducing human sulphotransferases into the improved (S)-reticuline production system. Analysis of human primary cells treated with these BIAs demonstrated that they affected a biomarker expression in a manner different from that by the parent compound (S)-reticuline, suggesting that simple side-chain modification altered the characteristic traits of BIA. These results indicated that highly productive microbial systems might facilitate the production of scarce or novel BIAs and enable subsequent evaluation of their biological activities. The system developed here could be applied to other rare natural products and might contribute to the drug-discovery process as a next-generation strategy.