Rewiring carbon metabolism in yeast for high level production of aromatic chemicals

Abstract

The production of bioactive plant compounds using microbial hosts is considered a safe, cost-competitive and scalable approach to their production. However, microbial production of some compounds like aromatic amino acid (AAA)-derived chemicals, remains an outstanding metabolic engineering challenge. Here we present the construction of a Saccharomyces cerevisiae platform strain able to produce high levels of p-coumaric acid, an AAA-derived precursor for many commercially valuable chemicals. This is achieved through engineering the AAA biosynthesis pathway, introducing a phosphoketalose-based pathway to divert glycolytic flux towards erythrose 4-phosphate formation, and optimizing carbon distribution between glycolysis and the AAA biosynthesis pathway by replacing the promoters of several important genes at key nodes between these two pathways. This results in a maximum p-coumaric acid titer of 12.5 g L^-1 and a maximum yield on glucose of 154.9 mg g^-1.

Fig. 1

Rewiring yeast central carbon metabolism for efficient production of aromatic chemicals. Introduction of a phosphoketolase (EC 4.1.2.22)-based pathway (green) to divert the glycolytic flux toward the formation of E4P. This consisted of a phosphoketolase to irreversibly cleave F6P into E4P and AcP, and a phosphotransacetylase to convert AcP into acetyl-CoA. The shikimate and AAA biosynthesis pathway were systematically investigated to relieve bottlenecks for biosynthesis of l-Phe and l-Tyr, two substrates for p-HCA production. A phenylalanine ammonia lyase and a cinnamic acid hydroxylase (PAL branch, pink) and a tyrosine ammonia lyase (TAL branch, blue) were both employed for the biosynthesis of p-HCA production. Further optimization of carbon distribution between glycolysis and the AAA biosynthesis pathway was achieved by fine-tuning glycolytic flux at phosphofructokinase and pyruvate kinase (blue triangles), via a combinatory promoter screening approach (see the main text). Glc glucose, G6P glucose-6-phosphate, F6P fructose-6-phosphate, F1,6BP fructose 1,6-biphosphate, G3P glyceraldehyde 3-phosphate, PEP phosphoenolpyruvate, Pyr pyruvate, AcP acetyl-phosphate, Ac-CoA acetyl-CoA, TCA tricarboxylic acid cycle, Ru5P ribulose 5-phosphate, R5P ribose-5-phosphate, X5P xylulose-5-phosphate, S7P sedoheptulose 7-phosphate, E4P erythrose-4-phosphate, DAHP 3-deoxy-d-arabino-2-heptulosonic acid 7-phosphate, CHA chorismic acid, PPA prephenate, l-Phe l-phenylalanine, l-Tyr l-tyrosine, l-Trp l-tryptophan, p-HCA p-coumaric acid

Fig. 2

Relieving bottlenecks in the AAA biosynthesis pathway. a Overview of yeast metabolic pathway for p-HCA biosynthesis. Both the PAL branch (pink) consisting of Arabidopsis thaliana phenylalanine ammonia lyase (AtPAL2), cinnamic acid hydroxylase (AtC4H), P450 reductase (AtATR2), and yeast native cytochrome b5 (CYB5) and the TAL branch (blue) consisting of tyrosine ammonia lyase (FjTAL) from Flavobacterium johnsoniae were constructed for p-HCA production. Overexpressed yeast endogenous genes are shown in orange, including DAHP synthase (ARO3), l-tyrosine-feedback-insensitive DAHP synthase (ARO4^K229L), pentafunctional aromatic protein (ARO1), chorismate synthase (ARO2), l-tyrosine-feedback-insensitive chorismate mutase (ARO7^G141S), prephenate dehydratase (PHA2), and aromatic aminotransferase I (ARO8). In addition, the E. coli shikimate kinase (EcaroL) was expressed, and three heterologous l-tyrosine prephenate dehydrogenase encoding genes (ZmtyrC from Zymomonas mobilis, GmPDH1 from Glycine max, and MtPDH1 from Medicago truncatula) were evaluated for enhancing l-tyrosine supply. The dashed lines indicate feedback inhibition of Aro4 and Aro7 by l-tyrosine and feedback inhibition of Aro3 by l-phenylalanine. Removal of the AAA degradation pathway was enabled by deleting the corresponding genes PDC5 and ARO10 (marked with red cross) in the engineered strains. DHQ 3-dehydroquinate, DHS 3-dehydro-shikimate, S3P shikimate-3-phosphate, SHIK shikimate, EPSP 5-enolpyruvyl-shikimate-3-phosphate, PPY phenylpyruvate, HPP para-hydroxy-phenylpyruvate, CA cinnamic acid; see Fig. 1 legend regarding abbreviations of other metabolites. p-HCA titers obtained with engineered strains derived from the PAL branch b, the TAL branch c, and the combination of both branches and the removal of the AAA degradation pathway d, respectively. ePhe represents the combination of identified upstream beneficial manipulations (shown in Fig. 2b) for phenylalanine biosynthesis. Cells were grown in defined minimal medium with six tablets of FeedBeads as the sole carbon source, and cultures were sampled after 96 h of growth for p-HCA detection. Statistical analysis was performed by using Student’s t test (one-tailed; two-sample unequal variance; *p < 0.05, **p < 0.01, ***p < 0.001). All data represent the mean of n = 3 biologically independent samples and error bars show standard deviation. The source data underlying figures b–d are provided in a Source Data file

Fig. 3

Validation of a phosphoketolase-mediated E4P generation route in the basic PAL branch strains. a Schematic overview of phosphoketolase (PHK)-based pathway for the generation of E4P for p-HCA production. The heterologous PHK pathway, consisting of a phosphoketolase from Bifidobacterium breve (BbXfpk) and a phosphotransacetylase from Clostridium kluyveri (CkPta), was introduced into the PAL branch-based (including a phenylalanine ammonia lyase and a cinnamic acid hydroxylase) strain. Ace acetate; see Fig. 1 legend regarding abbreviations of other metabolites. b Introduction of the PHK pathway in combination with feedback-insensitive DAHP synthase (ARO4^K229L), and chorismate mutase (ARO7^G141S) increases p-HCA titers via the PAL branch. c The PHK pathway outperforms the native E4P-generating route via Tkl1 for p-HCA production. d Combining the PHK pathway with deletion of the native glycerol-1-phosphatase-encoding gene GPP1 (marked with a red cross in a) enhances p-HCA production. Cells were grown in defined minimal medium with six tablets of FeedBeads as the sole carbon source, and cultures were sampled after 96 h of growth for p-HCA detection. Statistical analysis was performed by using Student’s t test (one-tailed; two-sample unequal variance; *p < 0.05, **p < 0.01, ***p < 0.001). All data represent the mean of n = 3 biologically independent samples and error bars show standard deviation. The source data underlying figures b–d are provided in a Source Data file

Fig. 4

Employment of a combinatorial strategy to increase the production of p-HCA. a Schematic overview of the metabolic pathway for p-HCA production with an improved supply of precursor E4P and dynamic control over the relevant biosynthetic genes, as indicated by triangle symbols: open triangles indicate the use of constitutive strong promoters to control gene expression, while filled triangles indicate the use of galactose-inducible promoters. See Fig. 1 legend regarding abbreviations of metabolites and Fig. 3 legend for gene details. b Integration of the PHK pathway with combined ePhe-PAL and eTyr-TAL branches leads to increased p-HCA production. eTyr refers to enhanced tyrosine biosynthesis mediated through the beneficial effect of MtPDH1. c Dynamic control of biosynthetic genes via use of the GALp-controlled expression system significantly increases p-HCA production. Cells were grown in defined minimal medium with six tablets of FeedBeads as the sole carbon source and 1% galactose as the inducer when required. Cultures were sampled after 96 h of growth for p-HCA detection. Statistical analysis was performed by using Student’s t test (one-tailed; two-sample unequal variance; *p < 0.05, **p < 0.01, ***p < 0.001). All data represent the mean of n = 3 biologically independent samples and error bars show standard deviation. The source data of figures b and c are provided in a Source Data file

Fig. 5

Optimization of carbon distribution increases p-HCA production. a Schematic illustration of carbon redistribution between glycolysis and the AAA biosynthesis pathway through a promoter library screening approach. A promoter library was created to replace the original promoters of PFK1, PFK2, and PYK1, which encode phosphofructokinase and pyruvate kinase, respectively, at key nodes between glycolysis and the AAA biosynthesis pathway. This promoter library was transformed into a yeast strain harboring the PHK pathway alongside the upregulated shikimate pathway, with the resulting strain screened by using a l-Tyr-derived pathway that indicated increased tyrosine production via the formation of the yellow pigment betaxanthin. Selected promoters exhibiting enhanced color intensity are listed in Supplementary Table 2. Furthermore, AAA degradation pathway was eliminated by deleting the corresponding genes (marked with a red cross) in the final strains. Open triangles indicate the use of constitutive strong promoters for controlling gene expression, while filled triangles indicate the use of galactose-inducible promoters. b Simultaneous optimization of the promoters of PFK1, PFK2, and PYK1 improves p-HCA production. c Removing the AAA degradation pathway further enhances p-HCA titers. Deletion of GAL80 was additionally introduced to enable the induction of GAL promoters without the addition of galactose. For shake-flask cultivation, cells were grown in defined minimal medium with six tablets of FeedBeads as the sole carbon source and 1% galactose as the inducer when required. For strains with the deletion of GAL80, no galactose was supplemented. Cultures were sampled after 96 h of growth for p-HCA detection. Statistical analysis was performed by using Student’s t test (one-tailed; two-sample unequal variance; *p < 0.05, **p < 0.01, ***p < 0.001). All data represent the mean of n = 3 biologically independent samples and error bars show standard deviation. The source data underlying figures b and c are provided in a Source Data file

Fig. 6

High-level production of p-HCA. a Fed-batch fermentation of strain QL58 and the corresponding diploid strain QL60 under glucose-limited conditions over time. p-HCA titer (filled symbols) and cell mass (open symbols) are shown for each strain. b Glucose consumption profile (filled symbols) and time course of residual glucose (open symbols) during the same fed-batch fermentation for each strain. All data represent the mean of n = 2 biologically independent samples and error bars show standard deviation. c Predicated metabolic flux distributions via flux balance analysis (FBA) in the engineered medium-level p-HCA producer QL158 (left, without carbon-rewiring strategy) and high-level p-HCA producer QL58 (right, with carbon-rewiring strategy). Based on FBA, the fluxes to the different products were represented relative to the uptake of 100 mmol of glucose. The PHK pathway, the nonnative E4P-forming biosynthetic pathway, is highlighted in green. 1,3BPG 1,3-biphosphoglycerate; see Fig. 1 legend regarding abbreviations of other metabolites. d Observed p-HCA crystals on filter membrane (left panel) and under a microscope (right panel, scale bar = 20 µm). Crystals of p-HCA were formed as a result of its high-level production and relatively low solubility in the fed-batch fermentation broth. The source data underlying figure a and b are provided in a Source Data file