De novo production of protoberberine and benzophenanthridine alkaloids through metabolic engineering of yeast

Abstract

Protoberberine alkaloids and benzophenanthridine alkaloids (BZDAs) are subgroups of benzylisoquinoline alkaloids (BIAs), which represent a diverse class of plant-specialized natural metabolites with many pharmacological properties. Microbial biosynthesis has been allowed for accessibility and scalable production of high-value BIAs. Here, we engineer Saccharomyces cerevisiae to de novo produce a series of protoberberines and BZDAs, including palmatine, berberine, chelerythrine, sanguinarine and chelirubine. An ER compartmentalization strategy is developed to improve vacuole protein berberine bridge enzyme (BBE) activity, resulting in >200% increase on the production of the key intermediate (S)-scoulerine. Another promiscuous vacuole protein dihydrobenzophenanthridine oxidase (DBOX) has been identified to catalyze two-electron oxidation on various tetrahydroprotoberberines at N7-C8 position and dihydrobenzophenanthridine alkaloids. Furthermore, cytosolically expressed DBOX can alleviate the limitation on BBE. This study highlights the potential of microbial cell factories for the biosynthesis of a diverse group of BIAs through engineering of heterologous plant enzymes.

Fig. 1. Schematic presentation of reconstructing the biosynthetic pathway for de novo production of protoberberines and BZDAs in yeast.

To enable the reconstruction of the complete pathway more accessible, it is divided into eight modules as different colors highlighted. Module I (pink) initiates from the condensation of dopamine and 4-HPAA to generate the first BIA compound (S)-norcoclaurine ((S)-NOR), which was sequentially converted into the core intermediates (S)-reticuline ((S)-RET) and (S)-scoulerine ((S)-SCO) in module II (blue) and module III (green). Module IV (sky blue), module V (gray), module VI (yellow) and module VIII (orange), diverged from the branch point intermediate (S)-SCO, aim to produce palmatine, berberine, sanguinarine, and chelerythrine, respectively. Module VII (rose) focuses on extending the pathway from sanguinarine to biosynthesize Chelirubine.

Fig. 2. Optimizing (S)-NOR production in yeast.

a Biosynthetic pathway from glucose to generate (S)-NOR in yeast. Color scheme of pathway: green, overexpression of endogenous genes or mutants; orange, introduction of heterologous genes; purple arrows, plant-based tyrosine pathway; black dashed arrow, multiple steps; red, gene deletion. EcAROL, shikimate kinase; MtPDH1, prephenate dehydrogenase; AtPAT, prephenate aminotransferase; MtncADH, noncanonical arogenate dehydrogenase; PsAAAD*, aromatic amino acid decarboxylase mutant. PPP, pentose phosphate pathway; E4P, erythrose 4-phosphate; PEP, phosphoenolpyruvate; DAHP, 3-deoxy-D-arabino-heptulosonate 7-phosphate; SA, shikimic acid; S3P, shikimate 3-phosphate; EPSP, 5-enolpyruvylshikimate-3-phosphate; CHA, chorismate; PPA, prephenate; HPP, 4-hydroxyphenylpyruvate; 4-HPAA, 4-hydroxyphenylacetaldehyde; Tyrosol, 2-(4-Hydroxyphenyl) ethanol; 4-HPAC, 4-Hydroxyphenylacetic acid. (S)-NOR titers and final OD₆₀₀ in engineered strains with b the removal of potential aldehyde reductases or dehydrogenases; c increasing more copies of CjNCS∆35; d introducing aromatic amino acid synthase and e plant-based tyrosine pathway. At least three biologically independent colonies were grown for 72 h in 20 mL minimal media with 20 g/L glucose. Augmenting the flux indicates overexpression of yeast native ARO1, ARO2, ARO3, and expression of EcAROL, MtPDH1, ARO4^K229L and ARO7^G141S. (S)-NOR pathway indicates the introduction of optimal CYP76AD5, DODC and CjNCS∆35. Significance was calculated using two-tailed t-test. Data are presented as mean ± standard deviations (n = 3 or 4 biologically independent samples). Source data are provided as a Source Data file.

Fig. 3. Extending the biosynthetic pathway from (S)-NOR to produce (S)-RET.

a Schematic presentation of module II. P450 NMCH catalyzes the rate-limiting step of hydroxylation of N-methylcoclaurine. CPR could shuttle two electrons to P450 for its catalyzation. Optimal coupling of alternative P450 and CPR enables improved activity. (S)-RET titers and final OD₆₀₀ in engineered strains carrying b Various combinations of NMCH and CPR, and c Multiple copies of rate-limiting enzymes. Significance was calculated using two-tailed t-test. Data are presented as mean ± standard deviations (n = 4 biologically independent samples). Source data are provided as a Source Data file.

Fig. 4. Fine-tuning BBE to improve (S)-SCO production.

a Schematic illustration of module III regarding fine-tuning BBE to increase (S)-SCO production. Green diamond indicates wild type CyBBE, transporting to vacuole; red diamond indicates Golgi-retained GOTS_CyBBE, retaining in Golgi; purple diamond indicates ER-targeted CyBBE_ERTS, retrograding to ER. b Fluorescence colocalization of CyBBE and its variants, (S)-SCO titers and final OD₆₀₀ in engineered strains, in which CyBBE was tailored to target various organelles. Scale bar represents 5 µm. c (S)-SCO titers and final OD₆₀₀ obtained from the engineered strains, in which ER was expanded by OPI1 deletion, and ER-localized mPRDX4 was expressed to alleviate the stress induced by H₂O₂ production. Significance was calculated using two-tailed t-test. Data are presented as mean ± standard deviations (n = 3 or 4 biologically independent samples). Source data are provided as a Source Data file.

Fig. 5. Extending the pathway to biosynthesize palmatine in module IV and berberine in module V.

a Schematic presentation of module IV and module V. b (S)-THP and (S)-canadine titers in strains XJ0839 and XJ0834, respectively. c 36h-heating of fermentation broth led to increased palmatine and berberine titers. Significance was calculated using two-tailed t-test. Data are presented as mean ± standard deviations (n = 3 biologically independent samples). Source data are provided as a Source Data file.

Fig. 6. Optimizing the biosynthesis of BZDAs in yeast.

a Schematic representation of strategies for chelerythrine (CHE) and sanguinarine (SAN) production. MCH5, plasma-bounded riboflavin transporter; BsRibc, FAD synthase from B. subtilis. Orange circle indicates FAD; purple hexagon indicates ER-targeted CyBBE_ERTS; green and dark green hexagon indicate vacuole-localized McDBOX2 and cytosolically expressed McDBOX2∆29, respectively. b Metabolite titers and final OD₆₀₀ in engineered strains with McDBOX2 expression. c Time course of CHE titer and OD₆₀₀ in engineered strains expressing McDBOX2∆29, McDBOX2 or no DBOX candidate. The solid line represents CHE production; the dashed line represents OD₆₀₀. Purple indicates the strain XJ07411 without the expression of DBOX candidate; light green indicates the strain XJ0747 (XJ07411 + McDBOX2) expressing wild type McDBOX2; dark green indicates the strain XJ07472 (XJ07411 + McDBOX2Δ29) expressing McDBOX2Δ29. d Fluorescence analysis of McDBOX2∆29 and McDBOX2 localization in yeast. Scale bar represents 5 µm. e (S)-RET, CHE titers and OD₆₀₀ in engineered strains XJ07411, XJ0747 (XJ07411 + McDBOX2) and XJ07472 (XJ07411 + McDBOX2Δ29). f Increased CHE production by improving FAD availability and OPI1 deletion-mediated ER expansion. Significance was calculated using two-tailed t-test. Data are presented as mean ± standard deviations (n  =  3 or 4 biologically independent samples). Source data are provided as a Source Data file.