1.1 Utilization of Red Algae as a Novel Carbon Source

To avoid the drawbacks of using fast-carbon glucose for terpenoid production, our project innovatively introduces red algae as a novel alternative carbon source. The core of this project is to endow S. cerevisiae with red algal polysaccharide degradation capability through heterologous expression of key enzyme components. Specifically, we introduced and overexpressed the agarase gene AqAga (BBa_258028F8) from Aquimarina agarilytica and the neoagarobiose hydrolase gene agaNash (BBa_257L75AN) from Cellvibrio in S. cerevisiae, combined with the S. cerevisiae α-mating factor secretion peptide gene (BBa_258RYIFY) to guide extracellular secretion of the enzymes. AqAga specifically hydrolyzes β-1,4 glycosidic bonds in agar, initially degrading red algal polysaccharides into neoagarobiose; agaNash further hydrolyzes neoagarobiose into galactose, providing an accessible carbon source for S. cerevisiae.

1.2 Establishing the Rh1 Synthesis Pathway in Saccharomyces cerevisiae

As a high-value tetracyclic triterpenoid, rare ginsenoside Rh1 has limited natural sources and inefficient traditional production methods. Our project achieved efficient microbial synthesis of Rh1 through genetic engineering by heterologously integrating key enzyme components for Rh1 synthesis into the S. cerevisiae chassis. We introduced five exogenous genes into the S. cerevisiae genome: the dammarenediol synthase gene PgDDS (BBa_255LOGT6), protopanaxadiol synthase gene CYP716A47 (BBa_25RY2RMX), cytochrome P450 reductase gene PgCPR1, protopanaxatriol synthase gene CYP716A53v2 (BBa_25NSM6TW), and glycosyltransferase gene UGTPg100 (BBa_255ROLW2). Experimental results showed that the engineered strain could produce up to 141.78 mg/L of Rh1 after 144 hours of fermentation under optimal conditions, providing an efficient and green microbial cell factory solution for industrial production of rare ginsenosides.

2.1 Open-Source FBA Computational Software Tool

We have developed a software tool for Flux Balance Analysis (FBA) that integrates an agent-based Q&A system and operational assistance. Currently, most FBA tools still require programming by users and lack intuitive visualization and graphical interaction capabilities. Visualization is often limited to results presentation, whereas the computational process itself remains code-dependent. Under the traditional research paradigm, conducting metabolic network analysis demands proficiency in programming languages such as MATLAB or Python. To address this, our software aims to provide a low-threshold, flexible, and user-friendly platform with graphical interaction support for metabolic modeling and analysis, tailored for synthetic biology research teams. The platform utilizes COBRApy for core FBA computations and offers a web-based user interface for visualization. It also incorporates an agent system with a dedicated FBA knowledge base, allowing users to interact with the software using natural language.

The complete workflow and implementation details are thoroughly described in the Software section.

2.2 Protein Molecular Docking Methods

This study achieved efficient screening of target enzymes through protein structure modeling and molecular docking techniques. During the project, we discovered that red algae cannot be directly utilized by yeast cells and require the synergistic action of two specific enzymes. We established a computational screening pipeline: first predicting protein structures of candidate enzymes from amino acid sequences using AlphaFold2, and then performing molecular docking with the Vina force field to evaluate the interaction efficiency of different enzyme combinations, thereby identifying the optimal enzyme system. Experimental validation of the model predictions partially confirmed the reliability of the docking results, indicating that the model is accurate and practically useful.

The full methodology is elaborated in Model 3.

2.3 Promoter Combination Prediction Tool

We developed a promoter combination prediction model based on ODE kinetic analysis. In the biosynthetic pathway of ginsenoside Rh1 in red algae, the yield of squalene directly determines the synthesis efficiency of Rh1. This process involves multiple coupled factors, including promoter regulation, enzyme expression levels, energy allocation, and cell growth. With theoretically thousands of possible promoter combinations, conventional experimental screening poses significant challenges. Our model effectively addresses the need for efficient selection of optimal promoter combinations. After expansion, the model demonstrated significantly improved predictive performance, achieving an overall R² of 0.8384 and a K-S test p-value of 0.7864. Predictions for specific combinations such as “PGal1+PGPD” were consistent with experimental rankings, demonstrating high reliability.

The complete modeling process and validation results are provided in Model 4.

3.1 空缺

正文

3.2 空缺

正文