Overview

Cancer is a major health concern globally, and chemotherapy, as one of the primary treatment approaches, encounters significant limitations such as severe side effects and reduced efficacy. ^[1]Sulforaphane, an isothiocyanate compound found in cruciferous vegetables, particularly in broccoli seeds, has demonstrated potent anticancer activity with minimal adverse effects. ^[2]However, current sulforaphane production relies on extraction from broccoli seeds, which contain less than 2% sulforaphane. This extraction process requires large-scale broccoli cultivation, is dependent on the plant's growth cycle, and results in land overuse, high production costs, and unsustainable practices.

Our objective is to engineer S.cerevisiae for the de novo biosynthesis of sulforaphane, optimizing its production through computational modeling and synthetic biology techniques. By reducing production costs and land use, we aim to offer a sustainable, scalable alternative to traditional extraction methods, contributing to both human health and global climate change mitigation efforts.

Introduction and pathway selection

Cruciferous vegetable intake has been significantly associated with a reduced risk of various cancers, such as prostate cancer and colon cancer ^[3]. This protective effect is primarily attributed to isothiocyanates found in broccoli, especially sulforaphane. In broccoli, sulforaphane exists in its precursor form—glucosinolate—which can be hydrolyzed by gut microbiota into bioactive sulforaphane ^[4]. Notably, broccoli seeds have the highest content of this glucosinolate, approximately 20 mg/g of fresh weight ^[5].

Recent clinical studies have confirmed that oral intake of sulforaphane is safe ^[6], with significant health benefits, including reducing obesity, decreasing hepatic glucose production, and improving blood glucose control in type II diabetes patients ^[7]. Due to these considerable health benefits, sulforaphane has gained widespread attention. However, current processes for extracting sulforaphane from cruciferous vegetables remain costly and are limited by the plant growth cycle ^[8], requiring substantial land resources.

Microbial cell factories have made significant progress in producing natural active compounds, such as successfully producing artemisinin in yeast ^{[9, 10]}, as well as flavonoids ^[11] and opioids ^[12]. With their efficiency, controllability, scalability, and environmental friendliness, microbial cell factories have become an ideal platform for synthesizing high-value plant compounds ^[13].

To date, there have been few reports on microbial synthesis of glucosinolates, including sulforaphane, and engineering research in this field is still in its early stages.

In this project, we plan to integrate the entire sulforaphane biosynthetic pathway into the genome of S.cerevisiae S288c for the first time. By optimizing the synthetic pathway and the P450 enzyme, we aim to achieve efficient sulforaphane synthesis in yeast. This work will not only advance research on sulforaphane biosynthesis but may also provide new solutions for its industrial production.

Pathway Construction

Sulforaphane is an aliphatic glucosinolate derived from methionine. Through the analysis of the A.thaliana genome, the natural synthesis process of sulforaphane has been revealed. This synthesis can be divided into three main stages ^[14]:

First, the methionine side chain is extended by inserting two methylene groups, generating a precursor amino acid derivative—dihomomethionine ^{[13, 15, 16]}.

Second, dihomomethionine undergoes a series of metabolic rearrangements to form the core structure of glucosinolates. This process involves complex enzymatic reactions that build the basic framework of the glucosinolate ^[17-20].

In the third step, the side chain of the initially formed glucosinolate is further modified to ultimately produce biologically active sulforaphane (as shown in Figure 1) ^[21].

In optimizing the synthetic pathway, we focused specifically on five key genes: FMO-GSOX, SOT18, IPMDH1, IPMI-SSU, among others. We selected specific sources of these genes or their better-performing mutants to improve sulforaphane production efficiency. Through the optimization and screening of these genes, our goal is to construct an efficient biosynthetic system for high-yield sulforaphane production.

In our study, we also introduced the ATR2 gene from A.thaliana and the CYb5 gene from S.cerevisiae strain S288c to assist in P450 enzyme expression. The ATR2 gene plays a critical role in regulating plant metabolism, while the CYb5 gene enhances the function of the P450 enzyme, which is crucial for improving sulforaphane synthesis efficiency ^[22].

Additionally, different species exhibit codon usage preferences ^[23]. For instance, the codon usage pattern in A.thaliana, a plant, differs significantly from that in S.cerevisiae, a fungus ^[24]. To enhance the transcription and translation efficiency of A.thaliana genes in S.cerevisiae, we performed codon optimization on all genes derived from A.thaliana. This optimization process aimed to adjust the codon sequences in A.thaliana genes to better align with the codon preferences of S.cerevisiae, thus improving protein synthesis efficiency and stability.

Furthermore , for the selection of the biosynthetic pathway for sulforaphane, we plan to utilize computational techniques for prediction to propose new pathways that differ from the natural synthesis route. This may provide new directions for future research. The design of biosynthetic pathways is typically based on the modification of natural routes and the design of new pathways ^[25]. Building on the natural pathway, we aim to explore novel synthetic routes.

We engineered a reverse biosynthesis approach to simulate the synthesis pathway of sulforaphane. This method involves two key steps: pathway prediction and pathway screening.

First, during the pathway prediction phase, we will leverage various biological databases to collect all possible synthetic pathways. We will integrate information from literature and databases, covering known enzymes and intermediates to construct potential biosynthetic pathways ^[26].

Second, in the pathway screening phase, we will conduct multi-criteria assessments of all predicted potential pathways to evaluate their experimental feasibility. This step involves a comprehensive evaluation of each pathway, including enzyme availability, metabolic efficiency, and the potential generation of by-products. Through this process, we aim to identify the most promising new pathways, providing a reliable theoretical basis for experiments.

We aim to develop new sulforaphane synthesis pathways through these computational simulations and database analyses, providing a generalized scheme for metabolic simulation. This research will not only aid in understanding the synthesis mechanism of sulforaphane but may also promote new advancements in the field of synthetic biology.

For detailed designs and results, please refer to:Metabolic simulation

Chassis selection

Due to the selection of synthesis pathways in this study that include numerous eukaryotic genes derived from A.thaliana, yeast, as a eukaryote, possesses post-translational modification mechanisms similar to those in plants, such as glycosylation and phosphorylation. These post-translational modifications play a crucial role in ensuring the correct expression and functionality of foreign plant gene products, which is especially vital for the synthesis of certain plant secondary metabolites.

Additionally, S.cerevisiae was chosen as the chassis microorganism based on its efficiency and plasticity in genetic manipulation. The S288c of S.cerevisiae, being one of the most commonly used experimental strains, exhibits a stable genetic background and high efficiency in expressing foreign proteins, making it an ideal engineering chassis for the expression of heterologous plant genes and the construction of complex metabolic pathways ^[27]. The widespread use of S.cerevisiae not only simplifies genetic operations but also enhances our yields and efficiency in the synthesis of secondary metabolites, ultimately providing a solid foundation for subsequent industrial biomanufacturing.

Pathway construction and genomic integration

To construct a complete biosynthetic pathway for sulforaphane, we first designed and constructed 18 single-cassette plasmids in Escherichia coli DH5α. These plasmids primarily include 16 genes encoding key enzymes of the synthesis pathway, along with two P450 enzyme genes that have been enzymatically modified, both of which are crucial for sulforaphane synthesis. We utilized the pRS series of shuttle plasmids as vector carriers to ensure stable gene expression in E. coli.

To achieve the construction of this complex metabolic pathway, we employed In-Fusion homologous recombination technology, which allows efficient assembly and integration of DNA fragments in E. coli ^[28]. The application of this technique enabled us to precisely construct the desired complete synthesis pathway in E. coli, laying the groundwork for subsequent expression in S.cerevisiae.

Considering the potential instability of plasmid fermentation in S.cerevisiae ^[29], we plan to merge the 18 single-cassette plasmids into 6 co-expression plasmids (as illustrated). This integration strategy not only simplifies genetic operations in S.cerevisiae but also enhances plasmid stability and expression efficiency. Through this approach, we hope to achieve efficient homologous recombination integration in S.cerevisiae, optimizing the sulforaphane synthesis pathway and increasing product yield.

To achieve the de novo synthesis of sulforaphane from methionine, we will integrate the required expression frames into the S.cerevisiae genome at the Delta high-copy integration site using homologous recombination technology [30]. This ensures efficient expression of the desired proteins in S.cerevisiae, facilitating the biosynthesis of sulforaphane.

By integrating different genes into the yeast genome, we can construct engineered yeast strains capable of producing intermediates in the sulforaphane biosynthetic pathway. The accumulation of these intermediates will help us evaluate the overall efficiency of the synthesis pathway and identify rate-limiting steps. This information will not only assist in optimizing the current biosynthetic process but also provide reference data for future improvements and expansions.

• G1:BCATT4-MAM1
• G2:LSU -IPMI2 -BCAT3 -IPMDH1
• G3:CYP79F1-CYP83A1 -SOT18 -ATR2
• G4:GGP1-GSTF9 -SUR1 -UGT-FMO
• G5:Myrosinase

P450 enzymes engineering

P450 enzymes are members of a heme-thiolate superfamily widely present in various organisms, primarily responsible for catalyzing mixed-function monooxygenase reactions ^[31]. Despite their crucial roles in many biological processes, the expression efficiency of P450 in heterologous systems is often low, limiting their application in biosynthetic pathways ^[32]. Particularly for sulforaphane synthesis, previous studies have indicated that CYP79F1 and CYP83A1 in the pathway are rate-limiting steps, significantly affecting overall synthesis efficiency. To address this, we conducted systematic modifications of these P450 enzymes to enhance their expression efficiency and catalytic activity in S.cerevisiae. Our modification plan includes the following aspects:

Computer-Aided Rational Protein Design

Optimization of Substrate Channel Hydrophobicity Through computer-aided rational design methods, we optimized the hydrophobicity of the substrate channel of the P450 enzymes to improve substrate affinity. This process involved structural adjustments of the substrate channel and optimization of amino acid residues. We used molecular docking techniques to simulate the binding of the optimized enzyme to the substrate, combined with molecular dynamics simulations to evaluate the stability and reaction kinetics of the enzyme-substrate complex.

For the modification of CYP79F1, we introduced amino acids with high affinity for dihomomethionine but low affinity for aldoxime (such as isoleucine and leucine) to enhance substrate binding specificity and efficiency. Conversely, for CYP83A1, we optimized polar amino acids (such as glutamine and asparagine) with high affinity for aldoxime to improve interactions with the substrate.

For detailed design plans and results, please refer to the model page.

Construction of fusion protein

Inclusion of Flexible Linkers To enhance reaction efficiency, we introduced a flexible linker (GGAGGG) between the CYP79F1 and CYP83A1 genes. This flexible linker consists of rotatable amino acid residues (such as glycine and alanine), providing greater conformational freedom and allowing sufficient space for independent folding and functional expression of the two fused protein domains. Compared to other types of linkers, this design reduces interference between domains and prevents unnecessary restrictions or contacts in three-dimensional space, helping to maintain the original activity of each functional domain. The inclusion of this linker shortens the spatial distance between the two enzymes, thereby fully preserving the functionality of the fusion protein and resulting in a fusion protein with enhanced performance. This structural optimization aids in improving interactions between the enzymes, ultimately increasing catalytic efficiency. We aim to validate the functionality of the fusion protein through in vitro enzyme activity assays.

Modification of Subcellular Localization To improve the catalytic efficiency of the P450 enzymes, we undertook a bold approach to alter their electron transfer environment. We aimed to localize the two P450 enzymes involved in the pathway to the mitochondrial inner membrane for potentially better results. To achieve this, we truncated the original endoplasmic reticulum anchoring sequence and signal peptide and added a mitochondrial localization sequence (MLS) to the N-terminus of the fusion protein. This modification is intended to enhance the overall catalytic efficiency by optimizing the enzymes' environment

Cofactor Supply

Supplementation of NADPH and Electron Transfer Auxiliary Proteins The catalytic reactions of P450 enzymes require NADPH as a cofactor; however, mitochondria lack appropriate NADPH and electron transfer auxiliary proteins. To overcome this limitation, we plan to further modify the enzyme system.

Specifically, we will introduce two proteins, CYB5 (cytochrome b5) and PntAB (NADH phosphatase), into the mitochondrial inner membrane. The PntAB sourced from Escherichia coli provides the necessary reducing power for biosynthesis by converting NADH into NADPH . NADPH is an essential reducing agent in fatty acid synthesis, nucleic acid synthesis, and antioxidant reactions. Under aerobic conditions, PntAB is one of the primary pathways for NADPH production, contributing approximately 35-45% of NADPH in glucose media. This is critical for the metabolic balance and growth of the cells.

CYB5 (cytochrome b5) provides the necessary reducing power in this reaction, supporting a redox-dependent mechanism. It acts as a reductive component, supplying electrons during the synthesis of very long-chain alkanes (VLC alkanes) ^[22] .

This electron transfer is critical for the proper functioning of various enzymatic reactions involved in VLC alkane biosynthesis, facilitating the necessary reductions that drive the conversion of substrates into these complex molecules. By enhancing the electron transfer process, CYTB5 plays a vital role in improving the overall efficiency and yield of the biosynthetic pathway.

We constructed the aforementioned P450 modification system within the yeast plasmid vector pRS425 and transformed it into the engineered strains producing GRA and SFN for fermentation experiments.

This introduction can comprehensively supplement the cofactors and electron transfer auxiliary proteins that are lacking in the mitochondria. This modification will help ensure that the fusion protein can be effectively localized and function on the mitochondrial inner membrane.