Introduction
This project aims to explore the efficient production of sulforaphane using Saccharomyces cerevisiae, achieving promising results. The project primarily utilizes synthetic biology tools to modify S. cerevisiae through homologous recombination, constructing a pathway in yeast for the de novo synthesis of sulforaphane from methionine. To achieve these objectives, we constructed 18 single-cassette plasmids: 16 gene cassettes required for the complete pathway and 2 modified P450 enzyme genes. Additionally, we combined these 18 plasmids into 6 co-expression plasmids, which were then introduced into S. cerevisiae via homologous recombination to express the desired proteins, enabling the de novo synthesis of sulforaphane from methionine.
Section1-Plasmid amplification and reconstruction.
We successfully constructed a single-cassette plasmid encompassing the entire metabolic pathway. The gene sequences were initially synthesized by the International Gene Synthesis Consortium (IGSC). Using PCR amplification, we obtained the plasmid backbone, along with gene cassettes, promoters, and terminators equipped with homologous arms. These components were assembled through seamless cloning via the In-Fusion method, facilitating homologous recombination.
The resulting recombinant plasmids were transformed into Escherichia coli strains Trans-5α and Trans-T1 for amplification. The engineered strains were cultured in LB medium supplemented with ampicillin to enable antibiotic selection. Colony PCR was performed using both universal and verification primers to screen for positive clones. Plasmids from colonies yielding positive PCR results were subsequently sequenced by IGSC to confirm the accuracy of the plasmid constructs.
BCAT4 catalyzes the transamination step in the methionine chain elongation pathway, serving as a crucial enzyme in the biosynthesis of methionine-derived glucosinolates. We successfully amplified the ADH1p-BCAT4-CYC1t construct (approximately 1700 bp) using PCR, and the observed fragment lengths matched the expected theoretical size of 1751 bp, as calculated based on primer design spanning from the promoter to the terminator. This result confirms the successful construction and acquisition of the desired plasmids.
MAM1 specifically catalyzes the first and second condensation reactions in the methionine carbon chain elongation pathway, thereby determining the side chain length of aliphatic glucosinolates. The PCR-amplified fragment CDC19p-MAM1-ADH1t (approximately 2500 bp) was found to be consistent with the theoretical length of 2853 bp, as predicted from the primer locations spanning from the promoter to the terminator. This result confirms the successful construction and acquisition of the plasmid containing the MAM1 gene.
LSU1 plays a crucial role in catalyzing the isomerization between 2-isopropylmalate and 3-isopropylmalate. The PCR-amplified fragment TEF1p-IPMI-LSU1-ADH1t (approximately 2000 bp) was found to match the theoretical length of 2265 bp, as estimated from the designed primer locations spanning from the promoter to the terminator. This result confirms the successful construction and acquisition of the LSU1 plasmid.
SSU2 catalyzes the isomerization between 2-isopropylmalate and 3-isopropylmalate, potentially via the formation of 2-isopropylmaleate, and functions redundantly with LEUD2 in the methionine chain elongation pathway for aliphatic glucosinolate formation. The PCR-amplified fragment ADH1p-IPMI-SSU2-PYK1t (approximately 1500 bp) was found to be consistent with the theoretical length of 1697 bp, as estimated from the designed primer locations spanning from the promoter to the terminator. This result confirms the successful construction and acquisition of the SSU2 plasmid.
IPMDH1 catalyzes the conversion of alkylmalate into homoketo acid. The PCR-amplified fragment GPDp-IPMDH1-CYC1t (approximately 2000 bp) was found to match the theoretical length of 2164 bp, as estimated from the designed primer locations spanning from the promoter to the terminator. This result confirms the successful construction and acquisition of the IPMDH1 plasmid.
BCAT3 catalyzes the transamination of homoketo acid to produce dihomomethionine. The PCR-amplified fragment ENO2p-BCAT3-HXT7t (approximately 2000 bp) was found to be consistent with the theoretical length of 2259 bp, as estimated from the designed primer locations spanning from the promoter to the terminator. This result confirms the successful construction and acquisition of the BCAT3 plasmid.
In this experiment, the enzyme specifically catalyzes the conversion of dihomomethionine (DHM). The PCR-amplified fragment CDC19p-CYP79F1-PYK1t (approximately 3000 bp) was found to match the theoretical length of 3144 bp, as estimated from the designed primer locations spanning from the promoter to the terminator. This result confirms the successful construction and acquisition of the CYP79F1 plasmid.
CYP83A1 metabolizes (E)-aldoximes derived from chain-elongated methionine, such as dihomomethionine. The PCR-amplified fragment PGI1p- CYP83A1-HXT7t (approximately 2000 bp) was found to match the theoretical length of 2464 bp, as estimated from the designed primer locations spanning from the promoter to the terminator. This result confirms the successful construction and acquisition of the CYP83A1 plasmid.
GGP1 participates in the biosynthesis of glucosinolates by hydrolyzing the γ-glutamyl bond in glutathione (GSH) conjugates, producing Cys-Gly conjugates. The PCR-amplified fragment ADH1p-GGP1-PYK1t (approximately 1500 bp) was found to match the theoretical length of 1550 bp, as estimated from the designed primer locations spanning from the promoter to the terminator. This result confirms the successful construction and acquisition of the GGP1 plasmid.
SUR1 plays a crucial role in glucosinolate biosynthesis by catalyzing the conversion of S-(alkylacetohydroximoyl)-L-cysteine to thiohydroximate. The PCR-amplified fragment HXT7p-SUR1-HXT7t (approximately 2000 bp) was found to match the theoretical length of 2185 bp, as estimated from the designed primer locations spanning from the promoter to the terminator. This result confirms the successful construction and acquisition of the SUR1 plasmid.
UGT74B1 plays a crucial role in glucosinolate biosynthesis by participating specifically in the glycosylation step. It is specific for the thiohydroximate functional group and does not glycosylate the carboxylate or hydroxyl groups. The PCR-amplified fragment PGK1p-UGT74B1-ADH1t (approximately 2500 bp) was found to match the theoretical length of 2532 bp, as estimated from the designed primer locations spanning from the promoter to the terminator. This result confirms the successful construction and acquisition of the UGT74B1 plasmid.
SOT18 exhibits a preference for long-chain desulfo-glucosinolates derived from methionine. The PCR-amplified fragment PGI1p-SOT18-HXT7t (approximately 1800 bp) was found to be consistent with the theoretical length of 1954 bp, as estimated from the designed primer locations spanning from the promoter to the terminator. This result confirms the successful construction and acquisition of the SOT18 plasmid.
FMO GS-OX1 encodes a flavin monooxygenase that catalyzes the conversion of methylthioalkyl glucosinolates into methylsulfinylalkyl glucosinolates. The PCR-amplified fragment GPDp-FMO-PYK1t (approximately 2000 bp) was found to match the theoretical length of 2464 bp, as estimated from the designed primer locations spanning from the promoter to the terminator. This result confirms the successful construction and acquisition of the FMO GS-OX1 plasmid.
Myrosinase 1 is a key enzyme involved in the degradation of glucosinolates. In this experiment, it primarily catalyzes the hydrolysis of Glucoraphanin to produce Sulforaphane, a compound known for its anticancer, antibacterial, and antioxidant properties. The PCR-amplified fragment PGI1p-Myrosinase(TGG)-PYK1t (approximately 2000 bp) was found to match the theoretical length of 2499 bp, as estimated from the designed primer locations spanning from the promoter to the terminator. This result confirms the successful construction and acquisition of the Myrosinase 1 plasmid.
Cytochrome b5 primarily functions as an auxiliary electron transfer protein, facilitating the targeted metabolic reactions. The PCR-amplified fragment GPDp-CYB5-PYK1t (approximately 1500 bp) was found to match the theoretical length of 1518 bp, as estimated from the designed primer locations spanning from the promoter to the terminator. This result confirms the successful construction and acquisition of the cytochrome b5 plasmid.
The enzyme ATR2 facilitates electron transfer not only to cytochrome P450 but also interacts with cytochrome b5, thereby assisting its function. Additionally, ATR2 can reduce a range of substrates, including cytochrome c, ferricyanide, and dichlorophenol indophenol, demonstrating its broad substrate reduction capability. The PCR-amplified fragment ADH1p-ATR2-TDH2t (approximately 3000 bp) was found to match the theoretical length of 3066 bp, as estimated from the designed primer locations spanning from the promoter to the terminator. This result confirms the successful construction and acquisition of the ATR2 plasmid.
In this experiment, the enzyme specifically catalyzes the conversion of dihomomethionine (DHM). CDC19p- CYP79F1(truncated)-PYK1t (3000bp+) from PCR are identical to the theoretical lengths of 3028bp estimated by the designed primer locations (promoter to terminator), which could demonstrate that these plasmids had successfully been obtained.
CYP83A1 metabolizes (E)-aldoximes derived from chain-elongated methionine, such as dihomomethionine. PGI1p- CYP83A1(truncated)-HXT7t (2000bp+) from PCR are identical to the theoretical lengths of 2322bp estimated by the designed primer locations (promoter to terminator), which could demonstrate that these plasmids had successfully been obtained.
The results above could demonstrate that these plasmids are correctly constructed, allowing the combination of these single-cassette plasmids into co-expression plasmids.
Possesses glutathione S-transferase activity toward 1-chloro-2,4-dinitrobenzene (CDNB) and benzyl isothiocyanate (BITC), and glutathione peroxidase activity toward cumene hydroperoxide and linoleic acid-13-hydroperoxide. GPDp-GSTF9-ADH1t（1500+）from PCR are identical to the theoretical lengths of 1627bp estimated by the designed primer locations (promoter to terminator), which could demonstrate that these plasmids had successfully been obtained.
The results above could demonstrate that these plasmids are correctly constructed, allowing the combination of these single-cassette plasmids into co-expression plasmids.
The pntA gene encodes a component of the membrane-bound transhydrogenase complex, which facilitates the transhydrogenation between NADH and NADP. This reaction is coupled to respiration and ATP hydrolysis, functioning as a proton pump across the membrane. The PCR-amplified fragment ADH1p-pntA-TDH2t (approximately 2400 bp) was found to match the theoretical length of 2430 bp, as estimated from the designed primer locations spanning from the promoter to the terminator. This result confirms the successful construction and acquisition of the pntA plasmid.
The pntB gene encodes a component of the membrane-bound transhydrogenase complex, which facilitates the transhydrogenation between NADH and NADP. This reaction is coupled to respiration and ATP hydrolysis, functioning as a proton pump across the membrane. The PCR-amplified fragment ADH1p-pntB-TDH2t (approximately 2200 bp) was found to match the theoretical length of 2286 bp, as estimated from the designed primer locations spanning from the promoter to the terminator. This result confirms the successful construction and acquisition of the pntB plasmid.
Section2- Construction of Co-expression Plasmids and Genome Integration
After constructing the single-cassette plasmid, to facilitate the integration of gene cassettes into the Saccharomyces cerevisiae genome, we first combined each gene into five co-expression plasmids.
BCAT4 encodes an aminotransferase and MAM1 encodes a monooxygenase. The process involves the transamination of methionine by BCAT4 and subsequent coupling with acetyl-CoA to produce 2-Alkylmalate, thus facilitating the effective synthesis of intermediates in sulforaphane biosynthesis. ADH1p-BCAT4-CYC1t-CDC19p-MAM1-ADH1t (4000+) from PCR are identical to the theoretical lengths of 4512bp estimated by the designed primer locations (promoter to terminator), which could demonstrate that these plasmids had successfully been obtained.
We constructed a catalytic element utilizing the IPMI, IPMDH1, and BCAT3 genes, capable of catalyzing the isomerization of 2-alkylmalate, oxidative decarboxylation, and transamination, synthesizing the intermediate dihomomethionine within this pathway. TEF1p-LSU-ADH1t-ADH1p-IPMI2-PYK1t-ENO2p-BCAT3-HXT7t-GPDp-IPMDH1-CYC1t (8000+) from PCR are identical to the theoretical lengths of 8294bp estimated by the designed primer locations (promoter to terminator), which could demonstrate that these plasmids had successfully been obtained.
We constructed a catalytic element utilizing the CYP79F1 and CYP83A1 genes, catalyzing the conversion of dihomomethionine (DHM) to aldoxime, followed by the oxidation of aldoxime to an aci-nitro compound, thereby achieving the oxidation of DHM.CDC19p-CYP79F1-PYK1t-PGI1p-CYP83A1-HXT7t (5000+) from PCR are identical to the theoretical lengths of 5491bp estimated by the designed primer locations (promoter to terminator), which could demonstrate that these plasmids had successfully been obtained.
We constructed a catalytic element using the GGP1, GSTF9, SUR1, and UGT genes. The GGP1 enzyme hydrolyzes the γ-glutamyl peptide bond of glutathione (GSH) conjugates, producing S-alkyl thiohydroxime. SUR1, acting as a CS lyase, converts S-alkyl thiohydroxime to thiohydroxime, while UGT is involved in glycosylation reactions to synthesize desulfo-glucosinolate. ADH1p-GGP1-PYK1t-GPDp-GSTF9-ADH1t-HXT7p-SUR1-HXT7t-ADH1t-UGT-PGK1p (7000+) from PCR are identical to the theoretical lengths of 7894bp estimated by the designed primer locations (promoter to terminator), which could demonstrate that these plasmids had successfully been obtained.
We constructed a catalytic element using the SOT18, ATR2, and FMO genes, which accomplished the sulfation and sulfoxidation of desulfo-glucosinolate, leading to the synthesis of glucoraphanin, the precursor compound for sulforaphane. PGI1p-SOT18-HXT7t-ADH1p-ATR2-TDH2t-GPDp-FMO-PYK1t (7000+) from PCR are identical to the theoretical lengths of 7400bp estimated by the designed primer locations (promoter to terminator), which could demonstrate that these plasmids had successfully been obtained.
After constructing the co-expression plasmids, we need to electroporate the target gene cassettes into Saccharomyces cerevisiae. To prevent plasmid loss, we utilized the yeast's native homologous recombination system to integrate the genes at the corresponding loci in the genome.
First, we used PCR technology to design and amplify the gene cassettes of all co-expression plasmids, obtaining the corresponding gene cassettes with integration sites. These cassettes were then electroporated into the Saccharomyces cerevisiae genome via homologous recombination for expression validation. Below is an introduction to the cassettes:
- G1: BCAT4 -MAM1.
- G2: LSU -IPMI2 -BCAT3 -IPMDH1.
- G3: CYP79F1 -CYP83A1 -SOT18 -ATR2.
- G4: GGP1 -GSTF9 -SUR1 -UGT - FMO
- G5:Myrosinase.
The following are the PCR amplification results for each fragment：
The design strategy and validation results for the above complex segments are as follows (using Ⅴ as a validation reference example:
All gene fragment bands were consistent with the theoretical sizes, indicating successful amplification of these cassettes.
Next, we designed a unique electroporation protocol: first, we prepared competent Saccharomyces cerevisiae cells, and then used electroporation to introduce the cassette with genomic homologous arms into the yeast cells. The yeast utilized homologous recombination to assemble the gene cassettes, successfully constructing five strains, defined as I, II, III, IV, and V.
Below are the PCR verification results for the genomes of these strains:
All bands were consistent with the theoretical sizes, indicating that the gene cassettes with genomic homologous arms have been successfully transformed.
Section3-P450 enzyme engineering
What work we have done:
- Further validation of the expression and function of the fusion protein.
- Specific in vitro characterization of the fusion protein and pntAB
Work has not finished yet:
- Further validation of the expression and function of the fusion protein.
- Specific in vitro characterization of the fusion protein and pntAB
We first used computational modification methods to predict the improvement of the stability and expression efficiency of the two P450 enzymes. This was achieved by modifying the hydrophobicity and hydrophilicity of the substrate channel and evaluating through molecular dynamics simulations.
1.Results of computer-aided enzyme structure optimization
The specific semi-rational sequence modifications are as follows:
Next, we constructed the fusion protein. To enhance reaction efficiency, we introduced a flexible linker peptide between the CYP79F1 and CYP83A1 genes, reducing the spatial distance between them to form a fusion protein with improved functionality, aiming to increase catalytic efficiency. The introduction of the linker peptide optimized the interaction between the proteins, thereby improving enzymatic reaction efficiency.
Computational analysis revealed that the stability of the modified CYP79F1 increased by 31.22%, while the stability of the modified CYP83A1 increased by 29.46%.
For detailed results, please visit the model👈2.Fusion protein construction and subcellular localization modification results
We added a mitochondrial targeting sequence (MLS) at the N-terminus. To better modify its electron transfer environment, we targeted the two P450 enzymes involved in the pathway to the mitochondrial inner membrane.
After constructing the fusion protein, we integrated it into the Saccharomyces cerevisiae genome through homologous recombination and added EGFP to verify its successful targeting to the mitochondrial inner membrane. And we validated the fusion protein through SDS-PAGE gel electrophoresis.
3. Cofactor supply
However, since one of the cofactors for P450 enzymes is NADPH, which lacks the necessary cofactors and electron transfer-associated proteins in the mitochondria, we plan to further modify the enzyme system: introducing the cyb5 and pntAB proteins into the mitochondrial inner membrane using the same system, allowing the fusion protein to be localized and expressed on the mitochondrial inner membrane.
MLS-CYP79F1(truncated,Δter)-CYP83A1(truncated)-pntAB-CYB5（8000+）from PCR are identical to the theoretical lengths of 9978bp estimated by the designed primer locations (promoter to terminator), which could demonstrate that these plasmids had successfully been obtained.
We constructed the aforementioned P450s modification system into the yeast plasmid vector pRS425 and transformed it into the engineered strains producing GRA and SFN for fermentation experiments.
Section4-Fermentation and production testing
The detection of final and intermediate products during the fermentation process for sulforaphane production is relatively complex and involves numerous challenges. According to our investigation, the fermentation process for sulforaphane faces issues such as the instability of intermediate products, a lack of established quantitative detection methods, low product concentrations, and making detection difficult. To address this, we contacted relevant companies to analyze the lyophilized and concentrated samples using LC-MS or HPLC techniques.
After the construction of the strains, we proceeded with fermentation experiments in YPD medium at a constant temperature of 30°C,220 rpm for 96 h. Samples were taken, and cells were disrupted using ultrasonic treatment. The supernatant was collected, freeze-dried, and concentrated before being sent to a testing company for analysis. Using techniques such as LC-MS and HPLC, we achieved groundbreaking high yields, marking the first successful de novo synthesis of sulforaphane in Saccharomyces cerevisiae, reaching the highest yields reported to date.
We present the LC-MS raw data spectrum for GRA, one of the most important precursors and final products of the pathway, along with the final yield data from the 96 h fermentation.
We conducted large-scale fermentation of sulforaphane using a 3L bioreactor. With the assistance of our partner company, we processed the product through freeze-drying, concentration, ethyl acetate extraction, and purification steps, achieving a purity of over 95%. We also performed cellular experiments (approved by the safety committee) to validate the inhibitory effect of our produced sulforaphane on PANC-1 cells.
Summary and Outlook
In fact, the construction, optimization, fermentation, and eventual industrialization of such a complex metabolic pathway is an extremely challenging process. We are thrilled to have completed this "impossible mission" in less than five months, a feat made possible through everyone's collective effort. Although our results have not yet reached the anticipated industrial standards, we have successfully completed the most critical first step!
Looking ahead, we will continue to explore, optimize, and increase yields in pursuit of our target goals. It's worth noting that many naturally bioactive isothiocyanates share similar synthesis pathways, meaning that our engineered strains have the potential to evolve into a platform capable of efficiently synthesizing natural isothiocyanate compounds in the future.
