Decoding and reprogramming of the biosynthetic networks of mushroom-derived bioactive type II ganoderic acids in yeast

Abstract

Mushroom's specialized secondary metabolites possess important pharmacological activities, but their biosynthetic pathway elucidation is extremely challenging, not to mention reprogramming of their biosynthetic networks to target metabolites. By taking Ganoderma lucidum, a famous traditional medicinal mushroom, as a lead example, here we decoded the biosynthetic networks of type II ganoderic acids (TIIGAs), a group of its main bioactive metabolites by studying the coordinated gene expression in G. lucidum, identifying endogenous or heterologous enzymes capable of C22 hydroxylation, configuration conversion of C3 hydroxyl group, and acetylation on C3, C15 and C22 hydroxyl groups. Notably, we revealed the catalytic mechanism of the C22 hydroxylase CYP512W6, and an unexpected bifunctional acetyltransferase GlAT that is required to transfer acetyl groups to C15 and C22. Using a fluorescence-guided integration method, we achieved efficient biosynthesis of significant TIIGAs applicable to industrial fermentation. After introducing all the identified enzymes to baker's yeast, we observed that biosynthesis of downstream TIIGAs was severely impeded, and dredged the metabolic block by temporally regulating the expression of acetyltransferases. By reprogramming of the biosynthetic networks of TIIGAs, we were able to produce over 30 TIIGAs, exhibiting 1-4 orders of magnitude higher titers or efficiencies than those from farmed mushrooms. The work enables the access to valuable TIIGAs, facilitates their widespread application, and sheds light on research of other mushroom products.

Fig. 1. Analysis of the biosynthetic network of TIIGAs in G. lucidum.

a A series of unknown post-modification enzymes and their possible working sequences. Early steps in the biosynthesis of TIIGAs lead to the generation of the precursor lanosterol via the mevalonate (MVA) pathway (highlighted in light pink). Two previously characterized CYPs from G. lucidum facilitate the conversion of lanosterol to HLDOA, and to GA-Y and GA-Jb (highlighted in light yellow). A series of complex post-modifications, including C3 configuration conversion (highlighted in red arrow), C22 hydroxylation (highlighted in yellow arrow), and C3, C15, C22 acetylation (highlighted in blue arrow), are necessary for the formation of the end TIIGAs (highlighted in light blue). OSC, oxidosqualene cyclase. b Transcriptomic analysis of G. lucidum for identifying enzyme candidates. Co-expression analysis of G. lucidum RNA samples under dark and blue light exposure (Supplementary Table S3) using cyp5150l8 and cyp512w2 as bait genes. Linear regression analysis is employed to screen genes with PCC > 0.6 relative to the baits. PCC values are averages calculated using two baits. Enzymes highlighted in red refer to the baits used for the analysis, enzymes highlighted in black refer to the functional enzymes associated with unknown steps in biosynthesis of TIIGAs, and enzymes highlighted in gray indicate enzymes in the MVA pathway. For the remaining candidates, see Supplementary Table S4. The heatmap displays the Z-score calculated from log₂-normalized expression, with three replicates for each sample. The heatmap is drawn using https://www.chiplot.online/.

Fig. 2. Functional characterization of C3 oxidase and reductase.

a HPLC analysis of fermentation extracts of strains SC62-CK-r-CYP512W2-r (CK-r-CYP512W2-r), SC62-AfuSDR-r-CYP512W2-r (Afu-r-W2-r), and SC62-CsSDR-r-CYP512W2-r (Cs-r-W2-r). b HPLC analysis of fermentation extracts of engineered S. cerevisiae strains BJ5464-CK-r (BJ-CK-r), BJ5464-AKR1C2-r (BJ-AKR1C2-r), BJ5464-AKR1C4-r (BJ-AKR1C4-r), and BJ5464-FusC1-r (BJ-FusC1-r) after incubation with GA-TR for 48 h. c HPLC analysis of the fermentation extracts of strains SC62-CsSDR-AKR1C4-r (Cs-C4-r), SC62-CK-r (CK-r), SC62-CsSDR-AKR1C4-CYP512A2-r (Cs-C4-A2-r), and SC62-CYP512A2-r (A2-r). d Production of GA-Jb, GA-TR, and GA-Ja by strains Afu-C2, Afu-C4, Afu-Fu, Cs-C2, Cs-C4, and Cs-Fu after 120 h fermentation. All data represent the mean of three independent samples, and the error bars indicate the standard deviation.

Fig. 3. Discovery of CYP512W6 as a C22 hydroxylase.

a HPLC analysis of the fermentation extracts of strains SC62-CK-r-CsSDR-AKR1C4-CYP512W2-r (CK-r-Cs-C4-W2-r), SC62-CYP512W6-r-CsSDR-AKR1C4-CYP512W2-r (W6-r-Cs-C4-W2-r). b HPLC analysis of the fermentation extracts of strains SC62-CYP512W6-r-CYP512W2-r (W6-r-W2-r) and SC62-CK-r-W2-r (CK-r-W2-r). c HPLC analyses of the fermentation extracts of strains SC62-CK-r, and SC62-CYP512W6-r (SC62-W6-r). d HPLC analysis of the in vitro enzymatic extracts by incubating CYP512W6-containing microsomes (prepared from strain CYP512W6-r-iGLCPR-r) with GA-Jb. CK indicates microsomes prepared from the control strain CK-r-iGLCPR-r. e The optimized structures and transition states with their respective energy barriers for C22 hydroxylation of spin 2 and spin 4. f Eight residues contributed more than 1.4 kcal/mol in binding affinity calculated by MM/GBSA. g The residues with binding energies more than 1.4 kcal/mol and their conservativeness. Hydrogen bond and sulfur–hydrogen bond are shown as red dots. Residues with high variability are selected for experimental verification (circled by red rectangle, including conservative R366). The conservation is calculated by the ConSurf website. h Production of 3β-TLTOA and GA-Jb by strains SC62-CYP512W6-r-CYP512W2-r (WT), SC62-V105L-r-CYP512W2-r (V105L), SC62-L108A-r-CYP512W2-r (L108A), SC62-T212A-r-CYP512W2-r (T212A), SC62-M365A-r-CYP512W2-r (M365A), and SC62-R366A-r-CYP512W2-r (R366A) after 120 h of fermentation.

Fig. 4. Functional characterization of GlAT.

a HPLC analysis of the fermentation extracts of strains SC62-GlAT-r-CYP512W2-r (GlAT-r-W2-r) and SC62-CK-r-CYP512W2-r (CK-r-W2-r). b HPLC analysis of the fermentation extracts of strains SC62-GlAT-r-CsSDR-AKR1C4-CYP512W2-r (GlAT-r-Cs-C4-W2-r) and SC62-CK-r-CsSDR-AKR1C4-CYP512W2-r (CK-r-Cs-C4-W2-r). c HPLC analysis of the in vitro enzymatic extracts by incubating the GlAT-containing microsomes (prepared from the strain YL-T3-GlAT-r) with different substrates. CK indicates the microsomes prepared from the control strain YL-T3-CK-r.

Fig. 5. Discovery of GlAT as a bifunctional acetyltransferase on C15 and C22.

a The extracts from the in vitro enzymatic assay of GlAT-containing microsomes with GA-T2 (left), and the enlarged view as highlighted in the black dotted line frame in the middle (right). CK indicates microsomes prepared from the control strain YL-T3-CK-r. b GA-T2 is docked into the T-tunnel of GlAT (shown by surface) at the C22 pose (slate) and C15 pose (cyan), respectively. c The percentage and represented structure at PRSs for the C15 and C22 poses, respectively. H305 and Acyl-CoA are represented by green sticks. GA-T2 in the C15 pose and C22 pose is represented by cyan and slate sticks, respectively. d The concentration (indicated as mg/L or mAU·s) of different substrates (indicated as light blue column) and products after 3 h incubation of strains BJ5464-GlAT-r (WT), BJ5464-A211L-r (A211L), BJ5464-A211F-r (A211F), BJ5464-G208L-r (G208L), BJ5464-R218L-r (R218L), BJ5464-S79L-r (S79L), and BJ5464-H305A-r (H305A) with respective substrates.

Fig. 6. Functional characterization of BsAT.

a HPLC analysis of the fermentation extracts of strains SC62-BsAT-r-CYP512W2-r (BsAT-r-W2-r), SC62-CK-r-CYP52W2-r (CK-r-W2-r). b HPLC analysis of the fermentation extracts of strains SC62-BsAT-r-CsSDR-AKR1C4-CYP512W2-r (BsAT-r-Cs-C4-W2-r) and SC62-CK-r-CsSDR-AKR1C4-CYP512W2-r (CK-r-Cs-C4-W2-r). c LC-MS analysis of the in vitro enzymatic reaction extracts by incubating BsAT-containing crude enzyme (prepared from the strain BL21-BsAT-r) with different substrates. CK indicates the crude enzyme prepared from the control strain BL21-CK-r.

Fig. 7. Integration of key enzyme expression cassettes leads to efficient production of significant TIIGAs.

a The principle of integration. RFP and key enzyme expression cassettes, flanked by 500 bp homologous recombination arms, are used as donors for integration at the EGFP locus of strain SC62. Recombinant clones with high RFP but no GFP fluorescent signal are collected by FACS. Test tube fermentation is performed to screen strains with high production of TIIGAs. Then, the CRISPR plasmid is recycled by culturing in 5-FOA containing medium. The resultant strain is used for integration of other enzyme expression cassettes. For this round of integration, a synonymous mutant EGFP is designed in the donor to avoid its own homologous recombination with EGFP locus of SC62. b–d Production of GA-Jb (b), GA-Ja (c), and TLTOA (d) by selected clones after 120 h fermentation. The colonies with the highest TIIGA production are indicated by the red arrows.

Fig. 8. Reprogramming of the biosynthetic network to target TIIGAs.

a Biosynthesis of target TIIGAs can be achieved by tunable expression of key enzymes either via changing the concentration of antibiotics (green), integration of key enzyme expression cassettes without antibiotics (blue), or inducible expression of acetyltransferases (yellow). For speculating the chemical structures corresponding to peaks 20, 21, 28–33, see Supplementary Figs. S84–S86 and Note S2. b Production of TIIGAs by constitutive and inducible expression of acetyltransferases, and by integration of key enzyme expression cassettes at different loci (H1, H2, EGFP and RFP). Cs+C4, CsSDR and AKR1C4; Con, constitutive; Ind, inducible expression. All data represent the mean of three independent samples, and the error bars indicate the standard deviation.