Search

Combinatorial design is an effective strategy to acquire the optimal solution in complex systems. In this study, the combined effects of pathway combination, promoters’ strength fine-tuning, copy numbers and integration locus variations caused by δ-integration were explored in Saccharomyces cerevisiae using geranylgeraniol (GGOH) production as an example. Two GGOH biosynthetic pathway branches were constructed. In branch 1, GGOH was converted from isopentenyl pyrophosphate (IPP) and farnesyl diphosphate (FPP). In branch 2, GGOH was derived directly from IPP and dimethylallyl pyrophosphate (DMAPP). Regulated by 10 combinations of 11 diverse promoters, a fusion gene BTS1-ERG20, a heterologous geranylgeranyl diphosphate synthase from Sulfolobus acidocaldarius (GGPPSsa) and an endogenous N-terminal truncated gene 3-hydroxyl-3-methylglutaryl-CoA reductase isoenzyme 1 (tHMGR), were incorporated into yeast by δ-integration, leading to a series of GGOH producing strains with yields ranging from 18.45 mg/L to 161.82 mg/L. The yield was further increased to 437.52 mg/L by optimizing the fermentation medium. Consequently, the GGOH yield reached 1315.44 mg/L in a 5-L fermenter under carbon restriction strategy. Our study not only opens large opportunities for downstream diterpenes overproductions, but also demonstrates that pathway optimization based on combinatorial design is a promising strategy to engineer microbes for overproducing natural products with complex structure.

Identification of efficient key enzymes in biosynthesis pathway and optimization of the fitness between functional modules and chassis are important for improving the production of target compounds. In this study, the taxadiene biosynthesis pathway was firstly constructed in yeast by transforming ts gene and overexpressing erg20 and thmgr. Then, the catalytic capabilities of six different geranylgeranyl diphosphate synthases (GGPPS), the key enzyme in mevalonic acid (MVA) pathway catalyzing famesyl diphosphate (FPP) to geranylgeranyl diphosphate (GGPP), were predicted using enzyme-substrate docking strategy. GGPPSs from Taxus baccata x Taxus cuspidate (GGPPSbc), Erwinia herbicola (GGPPSeh), and S. cerevisiae (GGPPSsc) which ranked 1st, 4th and 6th in docking with FPP were selected for construction. The experimental results were consistent with the computer prediction that the engineered yeast with GGPPSbc exhibited the highest production. In addition, two chassis YSG50 and W303-1A were chosen, and the titer of taxadiene reached 72.8 mg/L in chassis YSG50 with GGPPSbc. Metabolomic study revealed that the contents of tricarboxylic acid cycle (TCA) intermediates and their precursor amino acids in chassis YSG50 was lower than those in W303-1A, indicating less carbon flux was divided into TCA cycle. Furthermore, the levels of TCA intermediates in the taxadiene producing yeasts were lower than those in chassis YSG50. Thus, it may result in more carbon flux in MVA pathway in chassis YSG50, which suggested that YSG50 was more suitable for engineering the taxadiene producing yeast. These results indicated that computer-aided protein modeling directed isoenzyme selection strategy and metabolomic study could guide the rational design of terpenes biosynthetic cells.