Pathway engineering of Escherichia coli for one-step fermentative production of L-theanine from sugars and ethylamine

Abstract

L-theanine is the most abundant free amino acid in tea that offers various favorable physiological and pharmacological effects. Bacterial enzyme of γ-glutamylmethylamide synthetase (GMAS) can catalyze the synthesis of theanine from glutamate, ethylamine and ATP, but the manufacturing cost is uncompetitive due to the expensive substrates and complex processes. In this study, we described pathway engineering of wild-type Escherichia coli for one-step fermentative production of theanine from sugars and ethylamine. First, the synthetic pathway of theanine was conducted by heterologous introduction of a novel GMAS from Paracoccus aminovorans. A xylose-induced T7 RNA polymerase-P _T7 promoter system was used to enhance and control gmas gene expression. Next, the precursor glutamate pool was increased by overexpression of native citrate synthase and introduction of glutamate dehydrogenase from Corynebacterium glutamicum. Then, in order to push more carbon flux towards theanine synthesis, the tricarboxylic acid cycle was interrupted and pyruvate carboxylase from C. glutamicum was introduced as a bypath supplying oxaloacetate from pyruvate. Finally, an energy-conserving phosphoenolpyruvate carboxykinase from Mannheimia succiniciproducens was introduced to increase ATP yield for theanine synthesis. After optimizing the addition time and concentration of ethylamine hydrochloride in the fed-batch fermentation, the recombinant strain TH11 produced 70.6 ​g/L theanine in a 5-L bioreactor with a yield and productivity of 0.42 ​g/g glucose and 2.72 ​g/L/h, respectively. To our knowledge, this is the first report regarding the pathway engineering of E. coli for fermentative production of theanine. The high production capacity of recombinant strain, combined with the easy processes, will hold attractive industrial application potential for the future.

Keywords: Escherichia coli; Fermentative production; L-theanine; Pathway engineering; γ-Glutamylmethylamide synthetase.

Fig. 1

Overall metabolic engineering strategies for theanine overproduction in E.coli. ppc: pyruvate carboxylase gene; gltA: citrate synthase gene; sucCD: succinyl-CoA synthetase gene; pckA: phosphoenolpyruvate carboxykinase gene from M. succiniciproducens; cgl0689: pyruvate carboxylase gene from C. glutamicum GDK-9; cgl2079: glutamate dehydrogenase gene from C. glutamicum GDK-9; gmas: γ-glutamylmethylamide synthase gene from P. aminovorans.

Fig. 2

Selection of GMAS for the fermentative production of theanine. a Specific activities of GMAS_Mu, GMAS_Me, GMAS_Mm and GMAS_Pa expressed in E. coli BL21(DE3). b Effect of integrating different P_trc-gmas gene in E. coli W3110 on theanine titer and biomass. Each experiment was repeated three times, and data are represented as the means of three replicates and bars represent the standard deviations. ∗, P ​< ​0.05 and ∗∗, P ​< ​0.01 indicate the significance level between the two engineered strain. e.g., GMAS-Mu v GMAS-Me, GMAS-Mm v GMAS-Me, GMAS-Pa v GMAS-Mm, TH2 v TH1, TH3 v TH2, TH4-1 v TH3.

Fig. 3

Optimization of the expression levels of gmas_Pa for the fermentative production of theanine. a Optimization of the promoter and copy numbers of gmas_Pa. b Optimization of xylose inducer concentration. TH4-1: one copy of P_trc-gmas_Pa; TH4-2: two copies of P_trc-gmas_Pa; TH4-3: three copies of P_trc-gmas_Pa; TH5-1: one copy of P_T7-gmas_Pa; TH5-2: two copies of P_T7-gmas_Pa. Each experiment was repeated three times, and data are represented as the means of three replicates and bars represent the standard deviations. ∗, P ​< ​0.05 and ∗∗, P ​< ​0.01 indicate the significance level between the two engineered strain. e.g., TH4-2 v TH4-1, TH4-3 v TH4-2, TH5-1 v TH4-3, TH5-2 v TH5-1.

Fig. 4

Effect of overexpressing the native citrase synthase (TH6) and glutamate dehydrogenase from C. glutamicum GDK-9 (TH7) on theanine titer and biomass (a), and the formation of acetate and formate (b). Each experiment was repeated three times, and data are represented as the means of three replicates and bars represent the standard deviations. ∗, P ​< ​0.05 and ∗∗, P ​< ​0.01 indicate the significance level between the two engineered strain. e.g., TH6 v TH5-1, TH7 v TH6.

Fig. 5

Effect of knocking out succinyl-CoA synthetase (TH8) and overexpressing pyruvate carboxylase from C. glutamicum GDK-9 (TH9) on theanine titer and biomass (a), and the formation of acetate and formate (b). Each experiment was repeated three times, and data are represented as the means of three replicates and bars represent the standard deviations. ∗, P ​< ​0.05 and ∗∗, P ​< ​0.01 indicate the significance level between the two engineered strain. e.g., TH8 v TH7, TH9 v TH8.

Fig. 6

Effect of using phosphoenolpyruvate carboxykinase from M. succiniciproducens instead of phosphoenolpyruvate carboxylase (TH10) and complementing native phosphoenolpyruvate carboxylase (TH11) on theanine titer and biomass (a), and the formation of acetate and formate (b). Each experiment was repeated three times, and data are represented as the means of three replicates and bars represent the standard deviations. ∗, P ​< ​0.05 and ∗∗, P ​< ​0.01 indicate the significance level between the two engineered strain. e.g., TH10 v TH9, TH11 v TH10.

Fig. 7

Effect of optimizing the addition time and concentration of ethylamine hydrochloride in a 5-L bioreactor on theanine titer (a), biomass (b), and glucose consumption (c). Each experiment was repeated three times, and data are represented as the means of three replicates and bars represent the standard deviations.

Fig. 8

Fed-batch production of theanine in a 5-L bioreactor. Ethylamine hydrochloride (200 ​g/L) was automatically fed at a rate of 30 ​mL/h from 6 ​h to 26 ​h.