Bergenin is the main bioactive component of the traditional Chinese medicine Ardisia japonica, which has multiple biological activities, including antitussive, anti-inflammatory, anti-diabetic and anti-cancer effects. Bergenin is mainly extracted from the rhizomes or stems of various plants, but its biosynthetic pathway remains unclear.
The collaborative team of Jungui Dai from the Chinese Academy of Medical Sciences & Peking Union Medical College and Lin Yang from Minzu University of China published an article entitled "Unravelling and reconstructing the biosynthetic pathway of bergenin" in Nature Communications on April 26, 2024, providing a basis for the sustainable supply of bergenin.
Structurally, bergenin is the lactone product of 4-O-methyl gallic acid 2-C-β-D-glycoside (4-OMGA-Glc). Existing 14C-glucose incorporation experiment has shown that Gallic Acid (GA) is the glucosyl acceptor in the biosynthesis of bergenin, so it may be produced by the regio- and/or stereoselective C-glycosylation and O-methylation of GA. The authors speculated three routes of the catalytic process by 4-O-methyltransferase (4-OMT) and 2-C-glycosyltransferase (2-CGT).
Fig. 1 Proposed biosynthetic pathway from GA to bergenin in A. japonica. (Yan, et al., 2024)
Bergenin can be detected throughout the plant of A. japonica, with different contents in leaves, stems, and rhizomes. The authors sequenced the transcriptome of the whole plant. First, 85 candidate Glycosyltransferase genes were screened based on the transcriptomes of roots, stems, and leaves. Analysis of these genes with CGTs in plants showed that two candidate glycosyltransferase genes, AjCGT1 and AjCGT2, were clustered in the same branch as typical plant CGTs. The expression levels of AjCGT1 and AjCGT2 were closely related to the content of bergenin. Therefore, these two genes were selected as candidate genes.
Then, the authors studied GA (2) and 4-OMGA (3) as substrates, respectively. The results of in vivo and in vitro experiments showed that both enzymes could catalyze the glycosylation of both substrates at position 2. Kinetic experiments showed that AjCGT1 and AjCGT2 had higher catalytic activity on 3 than 2, indicating that 3 was more likely to be a natural substrate, so glycosylation may occur after methylation. At the same time, the conversion rate of AjCGT1 to 3 was higher than that of AjCGT2, so AjCGT1 was selected as the target 2-CGT for further biochemical characterization and engineering strain construction.
The authors screened 10 genes with high expression, and only 6 genes (AjOMT2−7) were positively correlated with the expression of AjCGT1 and AjCGT2 genes. AjOMT2 and AjOMT3 clustered into the clade of caffeoyl-CoA OMTs and tended to modify the adjacent hydroxyl groups on the aromatic ring of the substrate.
Then, the authors used 2 and Norbergenin (6) as substrates to verify the activities of the two enzymes in vivo and in vitro. The results showed that AjOMT2 showed higher methylation activity for 2 and 6, and the conversion rates were similar. The kinetic properties showed that AjOMT2 had a significantly lower Km value for substrate 2. Therefore, the natural substrate of AjOMT2 may be 2 rather than 6, indicating that 4-O-methylation may occur before 2-C-glycosylation.
So far, the authors have identified two key enzymes (AjOMT2 and AjCGT1) in the biosynthesis of bergenin, and their synthesis pathway is the inferred route 1.
The authors attempted to use Escherichia coli as a chassis to achieve the de novo biosynthesis of 4-OMGA-Glc (4), the direct precursor of bergenin. The authors divided the Synthesis pathway into two modules, namely the GA production module and the methylation and glycosylation post-modification module.
Fig. 2 The designed de novo artificial biosynthesis of bergenin from glucose in E. coli. (Yan, et al., 2024)
The mutant PobA** obtained by engineering the p-hydroxybenzoate hydroxylase (PobA) of Pseudomonas aeruginosa can efficiently catalyze 4-hydroxybenzoic acid (4-HBA) to produce GA. PobA** and chorismate lyase (UbiC) were co-expressed in E. coli to achieve the construction of a high-yield GA strain. AjOMT2 and AjCGT1 were co-expressed in E. coli, and the production of 4-OMGA-Glc was finally detected by feeding the exogenous substrate GA.
The genes of the two modules were constructed on plasmids with different copy numbers for combined expression, resulting in six different strains. The experimental results showed that strain A3 (pCDF-PobA**-UbiC + pETDuet-AjOMT2-AjCGT1) had the highest 4-OMGA-Glc production. The pH of the final fermentation broth was adjusted to 0.1, and 4 can be completely converted to 1. The de novo biosynthesis of bergenin in E. coli was achieved. However, a large amount of GA accumulation was detected in the fermentation broth of the optimal strain A3, and the authors believe that the efficiency of the OMT-CGT module should be further optimized.
First, the structure of AjOMT2 and AjCGT1 was optimized. The authors used molecular docking to perform alanine scanning on the amino acid residues near the substrate binding pocket of AjOMT2, and the results showed that the Y203A mutant could improve the catalytic activity. Subsequently, the Y203 residue was subjected to site-directed saturation mutation, and the Y203T mutant showed the highest activity. Since alanine scanning of AjCGT1 did not obtain mutants with improved catalytic activity, the authors used HotSpot Wizard to predict hotspot mutagenesis residues that may change enzyme activity in AjCGT1. Among these mutants, AjCGT1-A332F showed the highest catalytic activity.
AjOMT2-Y203T (AjOMT2*) and AjCGT1-A332F (AjCGT1*) mutants were co-expressed in engineered bacteria (M1-M3). However, the catalytic activity of the engineered bacteria containing pET-AjOMT2*-AjCGT1* (M3) was significantly lower than that of the strain containing wild-type AjOMT2-AjCGT1 (E1). The conversion rate of the M2 strain carrying pET-AjOMT2-AjCGT1* was higher than that of the wild-type AjOMT2-AjCGT1 strain. Considering the codon preference of E. coli, codon optimization was performed on AjOMT2, AjCGT1, AjOMT2* and AjCGT1*. A total of 5 strains Y1-Y5 were designed. The results showed that strain Y4 containing the codon-optimized mutants AjOMT2*opt and AjCGT1opt had the highest yield. Therefore, the Y4 strain carrying pET-AjOMT2*opt-AjCGT1opt was used to construct the next generation of engineered bacteria.
In the biosynthetic pathway of 4, donors (SAM and UDP-GLc) play a vital role. The authors tried to enhance the supply of donors required in the catalytic process to increase the catalytic efficiency of the OMT-CGT module. The authors overexpressed the key enzymes involved in the supply of these two donors and established different versions of strains. The fermentation results showed that strain D4 (which only enhanced the recycling process of SAM) had the highest yield.
Finally, a fed-batch experiment was carried out with D4 as the fermentation strain, and 4 reached the maximum titer after 60 h of fermentation. Then the pH was adjusted to 0.1, and 4 was completely esterified to bergenin. Due to the low content of by-products during the fermentation process, the final product can be easily separated and purified.
Thus, the authors reconstructed the artificial biosynthetic pathway of bergenin and increased its yield to grams per liter, showing great potential to solve the problem of its resource scarcity.
Fig. 3 De novo biosynthesis of 4-OMGA-Glc. (Yan, et al., 2024)
In this study, the key enzymes AjCGTs and AjOMTs in the biosynthesis of bergenin were functionally identified, and its biosynthetic pathway was fully resolved. AjOMT2 and AjOMT3 performed highly regioselective 4-O-methylation of GA, while AjCGT1 and AjCGT2 acted as C-glycosyltransferases of 4-OMGA, showing strict regio- and stereo-specificity to form 4-OMGA-Glc.
The de novo biosynthetic pathway of 4-OMGA-Glc, the direct precursor of bergenin, was reconstructed in E. coli, including the GA production module, the OMT-CGT module, and the donor supply module, which increased the yield of 4-OMGA-Glc. This work demonstrates the potential of bacterial fermentation systems to increase the yield of rare Glycosides through metabolic engineering, and establishes a biochemical synthesis method for bergenin, laying the foundation for the commercial production of bergenin.
Reference
