Background

Natural products include terpenoids, polyketides, phenylpropanoids, alkaloids, and so on [1]. Terpenoids and their derivatives are compounds obtained from mevalonic acid with the isoprene unit (C5) as the basic structural unit in the molecular skeleton [2]. Due to their unique and diverse bioactive characteristics, terpenoids play an important role in fields and applications such as pharmaceuticals, spices, food, biofuels, cosmetics, and industrial raw materials [3,4,5,6], which have attracted the interest of many researchers. Cembratriene-ol is a monocyclic diterpenoid that exhibits good insect deterrent effects [7,8,9,10,11]. This compound is a plant natural product that is mainly found in Nicotiana plants [7], and it can be completely degraded in the natural environment without residue, making it a sustainable biopesticide ingredient [12, 13]. The titer of natural products extracted from plants cannot commonly meet the demands of their industrial application due to the lack of efficient production methods [1, 14]. The content of cembratriene-ol in plants is low (0.18% of the leaf dry weight) [10, 15], and direct extraction of cembratriene-ol from plants is difficult and expensive [15]. Moreover, cembratriene-ol is difficult and inefficiently chemically synthesised due to its complex structure [16]. The total chemical synthesis of plant natural products commonly produces significant waste, which may lead to further pollution [17].

Heterologous production pathways of natural chemicals in their natural production hosts can be transplanted or employed to construct engineered microbial cells producing these natural chemicals [1]. The biosynthesis of plant natural products with high value by engineered microbial cell factories has great prospects for their industrial production and application and provides a prospective solution for protecting plants and land and averting pollution from chemical synthesis [14, 18, 19]. Many studies have reported on heterologous synthesis of plant-based products in engineered microbial cell factories, such as cannabinoids [20], isoprenoids [21], and glycyrrhetinic acid [22]. Genetic and metabolic engineering methods have been used to optimize and modify metabolic pathways of microbial cells to efficiently produce high value chemicals [23]. Sophisticated metabolic engineering approaches can be employed to redirect the metabolic flux of microbial cells toward the formation of target products [1]. Moreover, systems metabolic engineering methods can be performed to construct an efficient microbial cell factory by generating elucidating metabolic processes and integrating these techniques with genetic engineering, synthetic biology, and systems biology [24,25,26,27].

With the continuous advances in biotechnology, an increasing number of microorganisms have been developed and modified to produce high-value chemicals, such as Escherichia coli [28, 29], Saccharomyces cerevisiae [30, 31], and Bacillus subtilis [32]. Among them, E. coli is one of the most commonly used foundational strains for the production of natural compounds, which includes many advantages (e.g. high growth rate and cell density culture, and well-established metabolic engineering tools and methods) [1, 28]. E. coli includes the natural 1-deoxy-d-xylulose-5-phosphate (DXP) pathway [1], which provides the precursor (farnesyl pyrophosphate, FPP) for synthesising cembratriene-ol from plants (Fig. 1). The overexpression of rate-limiting enzymes in the DXP pathway promotes carbon metabolic flux to the target products [1].

Fig. 1
figure 1

Synthetic pathway of cembratriene-ol in E. coli. F6P, fructose 6-phosphate; G3P, glyceraldehyde 3-phosphate; PYR, pyruvate; DXP, 1-deoxy-d-xylulose-5-phosphate; MEP, 2-C-methyl-d-erythritol-4-phosphate; CDP-ME, 4-(cytidine 5’-diphospho)-2-C-methyl-d-erythritol; CDP-MEP, 2-phospho-4-(cythidine 5’-diphospho)-2-C-methyl-d-erythritol; MEcPP, 2-C-methyl-d-erythritol 2,4-cyclodiphosphate; HMBPP, 1-hydroxy-2-methyl-2-butenyl 4-diphosphate; IPP, 4-hydroxyphenylpyruvate; DMAPP, dimethylallyl diphosphate; GPP, geranyl diphosphate; FPP, farnesyl diphosphate; GGPP, geranylgeranyl diphosphate; dxs, 1-deoxy-d-xylulose 5-phosphate synthase gene; dxr, 1-deoxy-d-xylulose 5-phosphate reductoisomerase gene; IspD, 2-C-methyl-d-erythritol 4-phosphate cytidylyltransferase gene; IspE, 4-(cytidine 5’-diphospho)-2-C-methyl-d-erythritol kinase gene; IspF, 2-C-methyl-d-erythritol 2,4-cyclodiphosphate synthase gene; IspG, (E)-4-hydroxy-3-methylbut-2-enyl-diphosphate synthase gene; IspH, 1-hydroxy-2-methyl-2-(E)-butenyl 4-diphosphate reductase; idi, isopentenyl-diphosphate Δ-isomerase gene; ispA, geranyl diphosphate synthase gene

Full size image

In this study, we overexpressed the farnesyl diphosphate synthase gene ispA, the 2-C-methyl-d-erythritol 4-phosphate cytidylyltransferase gene ispD, 2-C-methyl-d-erythritol 2,4-cyclodiphosphate synthase gene ispF, and 1-deoxy-d-xylulose-5-phosphate synthase gene dxs from E. coli, isopentenyl diphosphate isomerase gene idi from Haematococcus lacustris, geranylgeranyl pyrophosphate synthetase (CrtE) gene ggpps from Pantoea ananatis, and cembratriene-ol synthase gene cbts* from Nicotiana tabacum in recombinant E. coli and systematically regulated their expession levels to boost the titer of cembratriene-ol. Introduction of the heterologous mevalonate (MVA) pathway from Streptomyces and fermentation optimisation methods and a 15-L fermenter were also used to enhance the titer of cembratriene-ol. Moreover, the productivity, conversion yield, cell growth, residual glucose, and other parameters were determined and analysed.

Results and discussion

Construction of recombinant E. coli producing cembratriene-ol

E. coli contains the natural DXP pathway, which can synthesise terpenoid skeleton chemicals. This pathway includes seven enzyme-catalysed reaction steps and requires NADPH and ATP [1]. Farnesyl pyrophosphate (FPP) can be synthesised and accumulated via the natural DXP pathway in E. coli. Based on the initial product FPP, cembratriene-ol can be produced via heterologous expression of geranylgeranyl pyrophosphate synthase and cembratriene-ol synthase in E. coli [12]. In this study, the geranylgeranyl pyrophosphate synthase gene ggpps and cembratriene-ol synthase gene cbts were heterologously expressed to synthesize cembratriene-ol in E. coli (Additional file 1: Fig. S1/S2/S3a). As shown in Fig. 2a, recombinant E. coli Z14 was constructed, with a titer of 1.58 × 10− 2 mg L− 1. The control strain without ggpps and cbts genes could not produce cembratriene-ol. Genes ggpps and cbts were expressed in E. coli, and cembratriene-ol was successfully synthesised. The productivity and conversion yield of glucose to cembratriene-ol in Z14 were 3.3 × 10− 4 mg L− 1·h− 1 and 0.04 mg g− 1 at 48 h, respectively (Fig. 2b, Additional file 1: Fig. S4). Since the growth rate is one of significant factors for the specific productivity [33], the cell growth and growth rate were determined in this study. The maximum cell density (OD600) and specific growth rate of Z14 were 3.73 and 0.24 h− 1, which of the control was 3.65 and 0.27 h− 1, respectively (Additional file 1: Fig. S3b/c). In the previous study, the cembratriene-ol toxicity effects were not observed for E. coli at the experimental concentration (0.25%) [12].

Fig. 2
figure 2

Construction of recombinant strains producing cembratriene-ol and its efficient production via optimizing key enzyme expression level. a The titer of cembratriene-ol; b Residual glucose concentration and conversion yield of glucose to cembratriene-ol. cbts* lacked a plastid transit peptide (52 AA, 156 bp) compared with cbts; ggpps, geranylgeranyl pyrophosphate synthase gene; idi, isopentenyl-diphosphate Δ-isomerase gene; dxs, 1-deoxy-d-xylulose 5-phosphate synthase gene; ispA, geranyl diphosphate synthase gene. −, no gene overexpression; +, gene overexpression using pACYCDuet-1; 2+, gene overexpression using pCDFDuet-1; 3+, gene overexpression using pETDuet-1; 4+, gene overexpression using pRSFDuet-1.

Full size image

Effect of the cembratriene-ol synthase and its truncated variant on cembratriene-ol production

As a key enzyme in the synthesis of cembratriene-ol, cembratriene-ol synthase is of great significance for the production of cembratriene-ol in recombinant E. coli. Mischko et al. expressed one cembratriene-ol synthase lacking the plastid transit peptide to synthesize cembratriene-ol [12]. In this study, the effects of the cembratriene-ol synthase gene cbts and its truncated variant cbts* on the synthesis of cembratriene-ol in recombinant E. coli were compared, and the recombinant E. coli Z14 and E. coli Z15 were constructed (Additional file 1: Fig. S1/S5a). The cembratriene-ol titer and productivity in E. coli Z15 expressing cbts* were increased by 20-folds higher than those of E. coli Z14 (Fig. 2a). The cbts* gene was also found to be more conducive to synthesizing cembratriene-ol in recombinant E. coli than cbts. It was found that the solubility of enzyme CBTS with plastid transit peptide sequence was reduced compared with CBTS* when they were heterologously overexpressed in E. coli (Additional file 1: Fig. S6). The conversion yield of glucose to cembratriene-ol in E. coli Z15 was 2-folds higher than that of E. coli Z14, indicating that the overexpression of cbts* facilitates the conversion of glucose to cembratriene-ol. The growth of E. coli Z15 was inhibited by the expression of cbts*, and the main reason may be that the large amount of glucose was used for synthesis of cembratriene-ol instead of cell growth (Additional file 1: Fig. S5b/c).

Effect of overexpressing key genes dxs and idi in carbon metabolism flux on cembratriene-ol production

The supply of isoprene pyrophosphate (IPP) and dimethylallyl diphosphate (DMAPP) is very important for the production of cembratriene-ol [12]. The 1-deoxy-d-xylulose 5-phosphate synthase DXS (product of the dxs gene) and IPP isomerase IDI (product of the idi gene) are two key rate-controlling enzymes in the DXP pathway of E. coli. The gene idi plays a significant role in controlling the isoprenoid flux [33]. Park et al. overexpressed IDI and geranylgeranyl diphosphate synthase GPS from Archaeoglobus fulgidus to produce astaxanthin in an engineered E. coli [34]. Genes dxs and idi were co-expressed using plasmid pACYCDuet-1 in E. coli Z15 to construct recombinant E. coli Z16 (Additional file 1: Fig. S1/S7a). The titer (1.19 mg L− 1) of cembratriene-ol in E. coli Z16 was increased by 3.75-folds that of E. coli Z15 (Fig. 2a). The productivity of cembratriene-ol in E. coli Z16 was increased from 6.6 × 10− 3 mg L− 1·h− 1 in E. coli Z15 to 2.5 × 10− 2 mg L− 1·h− 1, which indicated that overexpression of dxs and idi increased the carbon metabolism flux of the DXP pathway to cembratriene-ol and further enhanced the production of cembratriene-ol in recombinant E. coli. It was also reported that genes dxs and idi were overexpressed to improve the production of heterologous terpene (e.g. cembratriene-ol) [12, 35]. The growth ability of E. coli Z16 was inhibited to a certain extent compared with E. coli Z15, suggesting that overexpression of dxs and idi increased the flux of carbon metabolism to cembratriene-ol but not cell growth (Additional file 1: Fig. S7b/c). E. coli Z16 had a higher residual glucose concentration (1.27 g L− 1) in fermentation liquid compared with that of E. coli Z15, and the conversion yield of glucose to cembratriene-ol was 5.5-folds that of E. coli Z15 (Fig. 2b). Overexpression of dxs and idi was further verified to improve the conversion efficiency of glucose to cembratriene-ol via the DXP pathway in E. coli.

To further enhance the production of cembratriene-ol, plasmid pCDFDuet-1 was used to increase the expression levels of idi and dxs, and the recombinant E. coli Z17 was constructed (Additional file 1: Fig. S1/S8a). The titer and productivity of cembratriene-ol of E. coli Z17 reached 3.03 mg L− 1 and 6.3 × 10− 2 mg L− 1·h− 1, respectively, which was increased by 2.6-folds that of E. coli Z16 (Fig. 2). It was suggested that enhancing the expression levels of idi and dxs was beneficial for the efficient production of cembratriene-ol by recombinant E. coli. Yuan et al. also found a medium-copy plasmid pDCQ108 was used to express key enzymes (e.g., idi and dxs) to result in a significant increase in production of β-carotene compared with that of the low-copy plasmid pPCB15 [33]. Cell growth of E. coli Z17 was inhibited to a certain extent compared with E. coli Z16 (Additional file 1: Fig. S8b/c), and the carbon metabolism flux in the DXP pathway to cembratriene-ol should be one of the main reasons. The residual glucose concentration in the fermentation liquid of E. coli Z17 was close to 0, and the conversion yield of glucose to cembratriene-ol reached 0.763 mg g− 1, which was 1.75-folds that of E. coli Z16. The enhanced expression levels of idi and dxs were beneficial for converting glucose to cembratriene-ol. Plasmid pRSFDuet-1 with a high copy number was used to increase the expression levels of ggpps and cbts* in E. coli Z17 for efficient synthesis of cembratriene-ol, and recombinant E. coli Z18 was constructed (Additional file 1: Fig. S8a). The titer and productivity of cembratriene-ol in E. coli Z18 was not significantly improved, and it was indicated that overexpression of ggpps and cbts* under these conditions could not yield high efficiency cembratriene-ol production.

Enhancing cembratriene-ol production via systematic regulation of key gene expression

The DXP pathway contains multiple complex regulatory elements that are difficult to systematically optimise to enhance precursor accumulation (e.g. FPP). The accumulation of sufficient precursors plays an important role in producing the target compounds. Enhancing the biosynthetic flux of precursors in the DXP pathway in E. coli is a key strategy for increasing the accumulation of target products [1]. Based on the above results, it was found that the different expression levels of key genes idi, dxs, ggpps, and cbts* have an important influence on cembratriene-ol production in E. coli. The plasmids pACYCDuet-1, pCDFDuet-1, pETDuet-1, and pRSFDuet-1 were used to regulate the expression levels of key genes idi, dxs, ggpps, and cbts*, and four recombinant E. coli Z19, E. coli Z20, E. coli Z21, and E. coli Z22 were constructed (Additional file 1: Fig. S1/S9a). The theoretical copy numbers of plasmids pACYCDuet-1, pCDFDuet-1, pETDuet-1, and pRSFDuet-1 are 10, 20, 40, and 100, respectively [36]. In this study, the actual copy numbers of empty plasmids pACYCDuet-1, pCDFDuet-1, pETDuet-1, and pRSFDuet-1 are 3.38, 4.89, 9.12, and 26.3 at the cultivation conditions, respectively (Additional file 1: Fig. S10). The titers of cembratriene-ol in E. coli Z19, E. coli Z20, E. coli Z21, and E. coli Z22 were 2.80, 4.63, 2.89, and 2.35 mg L− 1, respectively (Fig. 2a). The titer of cembratriene-ol in E. coli Z20 was the highest, exhibiting an increase by 1.53-folds compared with that of E. coli Z17. The productivity of cembratriene-ol in E. coli Z20 was 9.6 × 10− 2 mg L− 1·h− 1, which was significantly higher than that of E. coli Z17. E. coli Z20 had the highest conversion yield of glucose to cembratriene-ol (1.32 mg g− 1). It was presumed that this key enzyme expression mode (pCDFDuet-1 with idi and dxs, pRSFDuet-1 with cbts*, and pETDuet-1 with ggpps) is beneficial for converting glucose to cembratriene-ol. The systemic regulation of idi, dxs, ggpps, and cbts* expression levels inhibited cell growth compared to that of the control (Additional file 1: Fig. S9b/c). The order of gene expression levels for improving the synthesis ability of cembratriene-ol in E. coli was cbts* > ggpps > idi > dxs, suggesting that overexpression of cbts* and ggpps using plasmids with high copies was preferred for the efficient synthesis of cembratriene-ol. These results indicated that cbts* and ggpps were critical genes for the efficient synthesis of cembratriene-ol in E. coli. Wang et al. also found that geranylgeranyl diphosphate synthase was a major rate-limiting enzyme for the biosynthesis of isoprenoids in engineered E. coli [34, 36].

Enhancing cembratriene-ol production by overexpressing the farnesyl diphosphate synthase gene ispA

Farnesyl diphosphate synthase ispA synthesising GPP and FPP significantly contributes to the biosynthetic system of the terpenoid carbon skeleton [38]. In this work, the recombinant E. coli Z25 was constructed via transforming the recombinant plasmid pACYCDuet-1-ispA into E. coli Z20 (Additional file 1: Fig. S1/S11a). The titer of cembratriene-ol in E. coli Z25 was significantly improved compared to that of E. coli Z24 (Fig. 2a). The titer and productivity of cembratriene-ol in E. coli Z25 were also enhanced compared those of E. coli Z24. The overexpression of farnesyl diphosphate synthase ispA significantly enhanced the synthesis of cembratriene-ol in E. coli. Meanwhile, cembratriene-ol was also efficiently synthesised in S. cerevisiae BY-T20 with high production level of geranylgeranyl pyrophosphate (GGPP) via enhancing MVA pathway, the titer of which reached approximately 1.56 mg L− 1 [39]. The glucose consumption of E. coli Z25 was close to that of E. coli Z24, and cell growth was not inhibited compared with E. coli Z24 (Additional file 1: Fig. S11b/c). ispA overexpression has little effect on the glucose consumption and cell growth of recombinant E. coli.

Enhancing the production of cembratriene-ol via optimised fermentation

Besides engineering strategies focus on biosynthetic pathways, the microbial cell factories performance can be greatly enhanced by appropriate optimization of the fermentation conditions [40]. Fermentation processes have been optimised to further improve the production of target products in recombinant E. coli, such as medium optimisation and environmental parameter optimisation. Carbon source and nitrogen source, as important medium components, play crucial roles in the production of primary and secondary metabolites and the cell growth [41]. The production level of cembratriene-ol was further improved by adding different proportions of glucose and yeast extract (Fig. 3). When different proportions of added glucose and yeast extract were 4:2, 4:4, 4:6, 4:8, 4:10, and 4:12, the titers of cembratriene-ol in E. coli Z25 were 9.94, 21.51, 22.47, 30.37, 21.00, and 20.48 mg L− 1, respectively. When the proportion of added glucose and yeast extract was 4:8, the titer, productivity, and conversion yield of glucose to cembratriene-ol were the highest, increasing by 4.75-, 4.75-, and 2.75-folds that of the control (without addition of glucose and yeast extract), respectively. It was indicated that high nitrogen source concentration promoted production of cembratriene-ol. It was also found that the addition of nitrogen source with high concentrations was advantageous for efficient accumulation of rhamnolipids [41, 42]. There was residual glucose when glucose and yeast extract were supplemented. When the proportion of added glucose and yeast extract was 4:2, the residual glucose concentration was the highest (up to 2.99 g L− 1). High concentrations of glucose cannot be fully utilised compared to yeast extract at high concentrations. The synthesis of primary or secondary metabolites and/or the cell growth can be usually influenced by the metabolic rate of carbon sources [41]. The growth of recombinant E. coli was significantly improved by the addition of glucose and yeast extract in different proportions compared with that of the control (without addition of glucose and yeast extract). When the proportion of added glucose and yeast extract was 4:8, the cell density was the highest and the cell growth was the fastest. The maximum OD600 and the maximum specific growth rate were increased by 3.00- and 3.23-folds those of the control (without addition of glucose and yeast extract), respectively (Additional file 1: Fig. S12a/b). However, when the proportion of added glucose and yeast extract was 4:2, the cell growth was lower than those of other proportions of added glucose and yeast extract. It was also found that astaxanthin production in Xanthopyllomyces dendrorhous could be promoted by high C/N, but the cell growth could be inhibited by high glucose [40]. Meanwhile, the pH changes of fermentation broth at different C/N ratios were determined, and the pH values at 4:8, 4:10, and 4:12 (C/N ratios) were 6.55, 6.60, and 6.64 at 48 h, respectively. Nitrogen sources can promote cell growth of microorganisms. Too many nitrogen sources can make the cell row too vigorously and the pH value is too high, which is not conducive to the accumulation of metabolites. Improper C/N ratios also affect the proportion of nutrients absorbed by the cell, directly affecting the cell growth and the product synthesis.

Fig. 3
figure 3

Effects of fermentation optimization on production of cembratriene-ol. a Effects of adding glucose and yeast extract with different proportions on the production of cembratriene-ol. b Effects of adding glucose and yeast extract with different proportions on the residual glucose concentration and the conversion yield at 48 h. CK means no addition of additional glucose and yeast extract, and the glucose concentration of CK is 4 g L-1. c Effects of different fermentation time on cembratriene-ol production under the proportion of glucose:yeast extract = 4:8. d Effects of different fermentation time on the residual glucose concentration and the conversion yield. Student’s t-test was used to statistically analyze data, and statistical significance was set at p value < 0.05. Asterisks indicate significant differences compared to E. coli Z25 (***, p < 0.01; **, p < 0.05; *, p > 0.05). p < 0.05 was considered statistically significant

Full size image

The above results showed that glucose could not be completely utilised by recombinant E. coli. The concentration of residual glucose reached 1.14 g L− 1 at 48 h when the proportion of added glucose and yeast extract was 4:8. It was presumed that the extension of the fermentation period could increase the production of cembratriene-ol in E. coli Z25. In this study, the fermentation period was optimised to improve cembratriene-ol production. As shown in Fig. 3c, the cembratriene-ol production by E. coli Z25 was further enhanced during the extension of fermentation time. The titers of cembratriene-ol reached 42.13 and 53.72 mg  L− 1 at 60 and 72 h, respectively (Fig. 3c). The production intensities of cembratriene-ol also increased to 0.70 and 0.75 mg  L− 1 h− 1 at 60 and 72 h, respectively (Fig. 3d). The titer of cembratriene-ol at 72 h was 1.8-folds higher than that at 48 h and 3357.5-folds that of the initially constructed E. coli Z14. The conversion yield of glucose to cembratriene-ol reached 6.81 mg g− 1 at 72 h. It was presumed that glucose was gradually converted into cembratriene-ol during extension of fermentation time.

Introduction of the heterologous MVA pathway

E. coli only has the natural DXP pathway, which includes 7 enzymatic steps with an adenosine triphosphate (ATP) and a reducing agent (NADPH) [1]. Intermediate enzymes and their genes from DXP pathway have been identified, and terpenoid units are synthesized from DXP starting from the condensation of glyceraldehyde 3-phosphate and pyruvate in this pathway [43, 44]. Other alternative routes (e.g., from pentose) along with original DXP pathway can be utilized to improve several terpenoids production [1]. MVA pathway commonly found in archaea and eukaryotes is another known biosynthetic pathway of terpenoids [1, 45]. MVA pathway starting from condensation of 2 acetyl-CoA includes 6 enzymatic steps with 3 ATPs and 2 NADPHs [44]. The heterologous MVA pathway was commonly introduced to improve production of terpenoids in E. coli [1]. In this study, 6 genes (orfA, orfB, orfC, orfD, hmgr, and orfE) encoding enzymes from MVA pathway of Streptomyces were synthesized after codon optimization to introduce a heterologous MVA pathway in E. coli Z25, including acetyl-CoA carboxylase, acetoacetyl-CoA synthase, hydroxymethylglutaryl-CoA synthase, hydroxymethylglutaryl-CoA reductase, mevalonate kinase, and mevalonate diphosphate decarboxylase (Fig. 4a). Genes orfA, orfB, and orfC were overexpressed with pRSFDuet-1 and genes orfD, hmgr, and orfE were overexpressed with pETDuet-1 to construct E. coli Z26 (Additional file 1: Fig. S13a). Genes orfA, orfB, and hmgr were overexpressed with pRSFDuet-1 and genes orfC, orfD, and orfE were overexpressed with pETDuet-1 to construct E. coli Z27 (Additional file 1: Fig. S13a). Production of cembratriene-ol was enhanced by introduction of the heterologous MVA pathway from Streptomyces. The titer of cembratriene-ol in E. coli Z27 was significantly increased by 1.13-folds that of E. coli Z25 (p value < 0.05) at 24 h (Fig. 4b). However, the titer of cembratriene-ol in E. coli Z27 was not significantly improved compared with that of E. coli Z26. Overexpression of key enzyme genes using plasmids with high copy numbers were also preferred for production of terpenoids [1, 46]. The conversion yields of glucose to cembratriene-ol in E. coli Z26 and E. coli Z27 were lower than that of E. coli Z25, which were consistent with glucose consumption yields of these strains (Additional file 1: Fig. S13b). Meanwhile, it was also found that with the extension of fermentation time, the titers of cembratriene-ol in E. coli Z26 and E. coli Z27 were not significantly improved compared with that of E. coli Z25 (data not shown). It was presumed that the introduction of the heterologous MVA pathway from Streptomyces might cause a metabolic imbalance in the late fermentation stage. Glucose was not gradually converted into cembratriene-ol in E. coli Z26 and E. coli Z27. However, cell growth of E. coli Z26 and E. coli Z27 was inhibited compared with that of E. coli Z25 (Additional file 1: Fig. S13c/d). More byproducts were formed in E. coli Z26 and E. coli Z27 than E. coli Z25. For example, the concentrations of lactic acid in E. coli Z26 and E. coli Z27 reached 90 and 69 mg g− 1 (DCW), which were higher than that in E. coli Z25. Meanwhile, the acetic acid concentrations in E. coli Z26 and E. coli Z27 reached 3.15 and 5.15 g g− 1 (DCW), which were higher than that (2.89 g g− 1 (DCW)) in E. coli Z25 (Additional file 1: Fig. S14).

Fig. 4
figure 4

Introduction of the heterologous MVA pathway. a Schematic diagram of introduction of the heterologous MVA pathway in E. coli. b Effects of introduction of the heterologous MVA pathway on the production of cembratriene-ol at 24 h. Student’s t-test was used to statistically analyze data, and statistical significance was set at p value < 0.05. Asterisks indicate significant differences compared to E. coli Z25 (**, p < 0.05; *, p > 0.05). p < 0.05 was considered statistically significant

Full size image

Enhancing the production of cembratriene-ol via overexpressing ispD and ispF and fermentation scale-up in 15-L fermenter

It was found that the titers of cembratriene-ol in E. coli Z26 and E. coli Z27 heterologously overexpressing MVA pathway from Streptomyces were not significantly improved compared with that of the control (E. coli Z25). The enzyme encoded by dxs can condensate glyceraldehyde-3-phosphate with pyruvate to produce deoxy-d-xylulose in the DXP pathway, and deoxy-d-xylulose is converted to the C5 isoprene subunits (IPP and DMAPP) by a series of additional isoprenoid enzymes encoded by dxr, ispD, ispE, ispF, ispG, and ispH in subsequent sequential reactions [33]. Enzymatic bottlenecks genes (dxs, idi, ispD, and ispF) reported were overexpressed to increase the flux through the upstream DXP pathway [33, 47]. Therefore, key genes ispD and ispF from the DXP pathway were additionally overexpressed to enhance synthesis of intermediate metabolites to improve the accumulation of cembratriene-ol in this study. The genes ispD and ispF were overexpressed using pRSFDuet-1 and pETDuet-1 with high copy numbers in E. coli Z25 to construct the recombinant strain E. coli Z28 (Additional file 1: Table S1). The cembratriene-ol titer in E. coli Z28 reached 83.91 mg L− 1 at 72 h, which was increased by 1.56-folds that of E. coli Z25 (Fig. 5a).

Fig. 5
figure 5

Enhancing the production of cembratriene-ol via overexpressing ispD and ispF and fermentation scale-up in 15-L fermenter. a, Enhancing the production of cembratriene-ol via overexpressing ispD and ispF. −, no gene overexpression; +, gene overexpression using pACYCDuet-1; 2+, gene overexpression using pCDFDuet-1; 3+, gene overexpression using pETDuet-1; 4+, gene overexpression using pRSFDuet-1. Student’s t-test was used to statistically analyze data, and statistical significance was set at p value < 0.05. Asterisks indicate significant differences compared to E. coli Z25 (**, p < 0.05). p < 0.05 was considered statistically significant. b, Fermention scale-up in 15-L fermentor

Full size image

Through fermentation optimization in the 250 mL shaking flask, it was verified that addition of glucose and organic nitrogen source (yeast extract) was conducive to the production of cembratriene-ol and cell growth. However, the concentration of added glucose in the shaking flask was low and not used for high-density fermentation in a 15-L fermenter. In this study, a fed-batch culture with DO-stat was adopted to achieve high-density fermentation in a 15-L fermenter. The feeding medium was fed at a dissolved oxygen (DO) concentration of 45%. Among feeding media, the carbon source was glucose, and the nitrogen source was mainly provided by adding NH3 H2O (inorganic nitrogen source). In order to promote cell growth, a small amount of yeast extract and peptone was added in high concentration glucose fed to supplement a certain amount of organic nitrogen source. Feeding medium included 500 g L− 1 glucose, 7.33 g L− 1 MgSO4, 4.0 g L− 1 yeast extract, and 4.0 g L− 1 tryptone. When the cell density OD600 reached 15 at 8 h, IPTG was added to induce expression of key enzyme genes (e.g. dxs, idi, ispA, ispD, ispF, ggpps, and cbts*), which could be used to efficiently synthesize cembratriene-ol in E. coli Z28. The titers of cembratriene-ol in 15-L fermenter reached 371.2 mg L− 1 at 80 h, which was significantly enhanced compared with that of the initially constructed E. coli Z14 (Fig. 5b). The highest OD600 reached 102.3 at 80 h. The residual glucose concentrations were below 0.5 g L− 1 (Fig. 5b). Carbon substrate feeding speed was connected to the DO level. When E. coli cells were actively growing with sufficient glucose, they required a lot of oxygen and the DO level would become very low. Once glucose is depleted, the DO level started to rise. Meanwhile, the titer of cembratriene-ol in the recombinant E. coli overexpressing dxs, idi, and ggpps using different ribosomal binding sites (RBS) combinatorics reached 78.9 mg L− 1 at 120 h in a 50-L fermenter, and the dry cell weight reached 15.3 g L− 1 [12].

Conclusion

The tobacco leaf–derived cembratriene-ol was successfully synthsized in recombinant E. coli. The systematic regulation of ggpps, cbts*, dxs, and idi expression levels could promote the production of cembratriene-ol. The overexpression of the farnesyl diphosphate synthase gene ispA, ispD, and ispF was significantly for enhancement of the cembratriene-ol titer. The cembratriene-ol titer in recombinant E. coli could be importantly improved by fermentation optimisation and introduction of the heterologous MVA pathway from Streptomyces. The cembratriene-ol titer in the 15-L fermenter was the highest titer reported. Systematic optimization has great potential in facilitating high-level production of cembratriene-ol production in E. coli.

Materials and methods

Strains and plasmids

The strains constructed and used are listed in Additional file 1: Table S1. E. coli JM109 was used for cloning, and E. coli BL21(DE3) was used as the initial strain to produce cembratriene-ol. The plasmids used and constructed are listed in Additional file 1: Table S2. The primers used for gene cloning and plasmid construction are listed in Additional file 1: Table S3.

Chemical reagents and kits

The antibiotics used in this study included ampicillin (Amp), kanamycin (Kan), streptomycin (Sm), and chloramphenicol (Cm). The standard of cembratriene-ol was purchased from Boc Sciences (Shiriey, New York, United States). Yeast extract and tryptone were purchased from Thermo Fisher Scientific Ltd. (Waltham, Massachusetts, USA), glucose, isopropyl-β-d-thiogalactose (IPTG), and L-arabinose were purchased from Sangon Biotech (Shanghai) Co., Ltd.; n-hexane and ethyl acetate were purchased from Shanghai Aladdin Biochemical Technology Co., Ltd.; PCR enzymes (e.g. PrimeSTAR HS DNA polymerase), ligase solution I, and restriction endonucleases needed for molecular biology experiments were purchased from TaKaRa (Dalian, China). The kits for plasmid extraction, agarose gel DNA extraction, and DNA fragment purification were purchased from Axygen Scientific, Inc. (Union, California, USA).

Media and culture conditions

Lysogeny broth (LB) medium (1 L) included 10 g NaCl, 10 g tryptone, and 5 g yeast extract. Solid LB medium was supplemented with 2% (w/v) agar powder. 5×M9 medium (200 mL) includes 12.8 g Na2PO4·7H2O, 3.0 g KH2PO4, 0.5 g NaCl, 1.0 g NH4Cl. Initial fermentation medium (1 L) includes 200 mL 5×M9 medium, 2 mL 1 M MgSO4 solution, 0.1 mL 1 M CaCl2 solution, 20 mL 20% (w/v) glucose solution, 1 mL 1 g L− 1 thiamine solution, 1 mL 1 g L− 1 biotin solution. A single colony from solid LB medium was transferred into 50 mL liquid LB medium in a 250 mL flask, which was cultured at 37 °C and 200 rpm for 10 h. The fermentation conditions for producing cembratriene-ol in 250 mL flasks (75 mL fermentation media) were a 1% (v/v) inoculum, 30 °C, 200 rpm, and 150 µM IPTG (final concentration) and lowering the fermentation temperature to 22 °C until OD600 = 0.6. Cells were collected at 5000 ×g at 4 °C for 10 min. The fermentation medium in 15-L fermenter included 12.0 g L− 1 glucose, 13.5 g L− 1 KH2PO4, 4.0 g L− 1 (NH4)2PO4, 1.7 g L− 1 citric acid, 1.68 g L− 1 MgSO4, and 5 mL L− 1 trace element solution. Feeding medium included 500 g L− 1 glucose, 7.33 g L− 1 MgSO4, 4.0 g L− 1 yeast extract, and 4.0 g L− 1 tryptone. Trace element solution included 100 mg L− 1 FeSO4 7H2O, 52.5 mg L− 1 ZnSO4 7H2O, 30 mg L− 1 CuSO4 2H2O, 5 mg L− 1 MnSO4 4H2O, 2.3 Borax, 20 mg L− 1 CaCl2, and 1 mg L− 1 (NH4)2MoO4. Among feeding medium, the carbon source was glucose, and the nitrogen source was mainly provided by adding NH3 H2O (inorganic nitrogen source). In order to promote cell growth, a small amount of yeast extract (4.0 g L− 1) and peptone (4.0 g L− 1) was added in 500 g L− 1 glucose fed to supplement a certain amount of organic nitrogen source. Strains were first cultured in 200 mL LB medium (2 L shaking flask), 1 L of which was transformed into a 15-L fermenter when the OD600 in the shaking flask reached 0.8. The 15-L fermenter included 7 L of fermentation medium, and the pH was controlled at 7.0 by adding NH3 H2O. The initial fermentation conditions were as follows: 37 °C, 200 rpm (initial speed), dissolved oxygen (DO) = 50%, and 1.0 MPa. 150 µM IPTG (final concentration) was added when OD600 reached 15, and the culture temperature was changed to 30 °C. The feeding medium was fed at a DO concentration of 45%.

Construction of recombinant plasmids and cembratriene-ol production strains

dxs gene was amplified from the E. coli BL21 (DE3) genome using primers dxs-FW and dxs-RS, which were digested with restriction endonucleases Nde I and Xho I. It was then ligated with pCDFDuet-1 digested with the same restriction enzymes to construct the recombinant plasmid pCDFDuet-1-dxs. Isopentenyl diphosphate isomerase (GenBank No. AAC32208.1) gene idi from H. lacustris, geranylgeranyl pyrophosphate synthetase (CrtE, GenBank no. ADD79325.1) gene ggpps from P. ananatis LMG 20,103, and the cembratriene-ol synthase CBTS (GenBank no. AAS46038.1) genes cbts and cbts* without the plastid transit peptide (MSQSISPLICSHFAKFQSNIWRCNTSQLRVIHSSYASFGGRRKERVRRMNRA) from N. tabacum, acetyl-CoA carboxylase gene orfA (NCBI Reference Sequence: WP_142233022.1), acetoacetyl-CoA synthase gene orfB (GenBank no. EGE45775.1), hydroxymethylglutaryl-CoA synthase gene orfC (NCBI Reference Sequence: WP_150173950.1), hydroxymethylglutaryl-CoA reductase gene orfD (NCBI Reference Sequence: WP_150173948.1), mevalonate kinase gene hmgr (GenBank no. BAB07790.1), and mevalonate diphosphate decarboxylase gene orfF (GenBank no. BAD86801.1) from Streptomyces were synthesised by Suzhou GENEWIZ Biological Technology Co., Ltd. (Suzhou, China) after codon optimization in E. coli (Supplemental materials). Gene idi was ligated with pCDFDuet-1 to construct pCDFDuet-1-idi after digestion using Nco I and EcoR I. After digestion via Nde I and Xho I, idi was ligated with pCDFDuet-1-dxs to construct recombinant plasmid pCDFDuet-1-dxs-idi. After digestion with Nde I and Xho I, ggpps was ligated with plasmid pETDuet-1 and pRSFDuet-1 to obtain the recombinant plasmids pETDuet-1-ggpps and pRSFDuet-1-ggpps. The cbts and cbts* genes were digested with Nco I and EcoR I and ligated to plasmids pETDuet-1-ggpps and pRSFDuet-1-ggpps to obtain recombinant plasmids pETDuet-1-ggpps-cbts, pETDuet-1-ggpps-cbts*, and pRSFDuet-1-ggpps-cbts*, respectively. The digested cbts* gene was ligated with pETDuet-1 and pRSFDuet-1 to obtain the recombinant plasmids pETDuet-1-cbts* and pRSFDuet-1-cbts*. The digested gene dxs was ligated with the plasmid pACYCDuet-1 to obtain the recombinant plasmid pACYCDuet-1-dxs. Digested idi was ligated with plasmids pCDFDuet-1, pACYCDuet-1, and pACYCDuet-1-dxs to obtain the recombinant plasmids pCDFDuet-1-idi, pACYCDuet-1-idi, and pACYCDuet-1-dxs-idi. The ispA gene was cloned from E. coli BL21 (DE3) genome using primers ispA_Ecoli_FW and ispA_Ecoli_RS and ligated with pACYCDuet-1 to construct the recombinant plasmid pACYCDuet-1-ispA. Recombinant plasmids were transformed into E. coli JM109 or E. coli BL21 (DE3) to obtain recombinant E. coli Z1 − Z28.

Detection of cembratriene-ol

After fermentation, 30 mL of fermentation broth with cells was collected from each flask. All flask fermentations were performed in triplicate. Cells in each flask were lysed using a SCIENTZ-IID Ultrasonic Cell Crusher from Ningbo Scientz Biotechnology Co., Ltd. (Ningbo, China). Equal volumes of the extractant (n-hexane:ethyl acetate = 1:1) were added to the mixtures, lysed, mixed for 30 min, and centrifuged at 5000×g and 4 °C for 10 min. The supernatant was collected and dried at 30 °C using a RE100-Pro rotary evaporator (Scilogex, New Hampshire, USA). To reconstitute the dried samples, 3 mL of ethyl acetate was used, and reconstitution was determined using GC-MS. The TraceGOLD TG-5MS GC column (30 m × 0.25 mm × 0.25 µm) of Thermo Fisher Scientific Co., Ltd. (Waltham, USA) was used to determine cembratriene-ol. The column temperature was 100 °C and its hold time was 0 min; it was increased to 250 °C at a rate of 10 °C min– 1 and its hold time was 5 min. MS data of samples were recorded at 70 eV (EI) and m/z (rel. intensity in %) as total ion current. Full scan mode (m/z 35–350) was used to collect the data. The productivity of cembratriene-ol was calculated over the whole culture. The conversion yield of glucose to cembratriene-ol was calculated by dividing the cembratriene-ol titer (mg L– 1) by the glucose consumption concentration (g L– 1). Three parallel experiments were independently performed, and data are reported as mean ± standard deviation (SD).

Determination of the cell density (OD600) and the specific growth rate

Recombinant E. coli culture broth (1 mL) was used to determine the absorbance at 600 nm (OD600). The specific growth rate plots were drawn from OD measurements using the software OriginPro 8.5.1. The data processing steps are as follows: (a) Fitting of time and OD to obtain data 1; (b) Differential of OD to time to obtain data 2; (c) Data 2 were divided by data 1 to obtain specific growth rates (µ), which included 100 data and the total fermentation time was evenly divided into 100 equal parts. The obtained specific growth rates (µ) including 100 data and time divided into 100 equal parts were used to draw the specific growth rate plots using the software OriginPro 8.5.1. Three parallel experiments were independently performed, and data are reported as mean ± standard deviation (SD).

Determination of residual glucose

Fermentation broth (1 mL) was collected and centrifuged at 12,000×g at 4 °C for 3 min. The supernatant was collected, and the glucose concentration was determined using a residual sugar analyser, SGD-III (Shandong Academy of Sciences, Jinan, China). Three parallel experiments were independently performed, and data are reported as mean ± standard deviation (SD).