Integration of pathway balance and protein fusion enables de novo biosynthesis of (+)-bicyclogermacrene in Escherichia coli
Chen-Yi Sun, Wen-Liang Xie, Zheng-Yu Huang, Chun-Xiu Li, Jian-He Xu

TL;DR
Scientists engineered E. coli to produce a valuable compound called (+)-bicyclogermacrene, achieving the highest production levels ever reported.
Contribution
The study reports the highest titer of (+)-bicyclogermacrene achieved through pathway balance and protein fusion in E. coli.
Findings
Genome-integrated MEP pathway increased precursor flux, raising titer from 11.3 to 50.1 mg/L.
Fusion of downstream genes and NADPH optimization boosted titer to 119 mg/L.
Engineered E. coli strain M6-36 produced 565 mg/L in a bioreactor, a 50-fold increase.
Abstract
(+)-Bicyclogermacrene and its derivatives, with promising antimicrobial, anticancer, and insecticidal properties, hold significant potential for applications in pharmaceuticals, agriculture, and industry. However, traditional extraction methods from plant essential oils are unsustainable. In this study, we achieved the de novo biosynthesis of (+)-bicyclogermacrene using a metabolically engineered Escherichia coli strain. The biosynthetic pathway of (+)-bicyclogermacrene was partitioned into upstream and downstream modules to enable precise regulation. This was accomplished through the genome-integrated overexpression of the endogenous methylerythritol phosphate pathway to ensure an adequate supply of terpenoid precursors, which pulled the titer from the initial 11.3 mg/L to 50.1 mg/L. Production was further enhanced to 96.9 mg/L by fusion of downstream key genes to facilitate precursor…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7- —http://dx.doi.org/10.13039/501100001809National Natural Science Foundation of China
- —National Key R & D Program of China
- —http://dx.doi.org/10.13039/501100012226Fundamental Research Funds for the Central Universities
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsPlant biochemistry and biosynthesis · Sesquiterpenes and Asteraceae Studies · Microbial Natural Products and Biosynthesis
Introduction
(+)-Bicyclogermacrene (C_15_H_24_) is a sesquiterpene found in the essential oils of medicinal and aromatic plants such as Syzygium levinei (Huong et al. 2023), Ageratum conyzoides (do Rosário et al. 2023), and Eugenia gracillima (Guedes et al. 2023). Studies have demonstrated its antifungal, antinociceptive, and anti-inflammatory properties, particularly when combined with germacrene D (Guedes et al. 2023; Romero et al. 2023; Valarezo et al. 2023). Recently, evidence has emerged highlighting the larvicidal potential of (+)-bicyclogermacrene-rich essential oils against malaria mosquitoes, positioning it as a sustainable alternative to conventional insecticides (Dai et al. 2023; Pham et al. 2023; Qi et al. 2023). Cramer reported a concise route started from commercially available (+)-2-carene to chemical synthesis of (+)-bicyclogermacrene in seven steps with 25% of overall yield in batches (Tran and Cramer 2014). Despite its promising applications, current methods for producing (+)-bicyclogermacrene rely on inefficient and unsustainable extraction techniques, highlighting the critical need for a green alternative strategy.
The biosynthesis of sesquiterpenes, including (+)-bicyclogermacrene, requires efficient production of isoprenoid precursors, isopentenyl pyrophosphate (IPP) and dimethylallyl pyrophosphate (DMAPP) (Niu et al. 2017; Liu et al. 2022). These precursors are canonically synthesized via two primary metabolic pathways: the methylerythritol-4-phosphate (MEP) pathway, commonly found in prokaryotes, and the mevalonate (MVA) pathway, predominant in eukaryotes, cyanobacteria and algae (Henry et al. 2015; Rinaldi et al. 2022). The interconversion of IPP and DMAPP was catalyzed by isoprenyl diphosphate isomerase (Idi). Following precursors formation, farnesyl pyrophosphate synthase (Fpps, also named IspA in Escherichia coli) condenses two IPP units with one DMAPP unit into farnesyl pyrophosphate (FPP), the immediate precursor of sesquiterpenoids, which undergoes cyclization and rearrangement reactions catalyzed by terpenoid synthases to produce diverse sesquiterpenes (Fig. 1). The isolation and functional characterization of bicyclogermacrene synthase from sources, such as Penicillium expansum (Huang et al. 2022), Jungermannia exsertifolia (Yan et al. 2022), Postia placenta (Ichinose and Kitaoka 2018), Matricaria recutita (Son et al. 2014), Handroanthus impetiginosus (Silva-Junior et al. 2018), Matricaria chamomilla var. recutita (Son et al. 2014), Citrus medica L. var. Sacrodactylis Hort (Xu et al. 2017) and Phyla dulcis (Attia et al. 2012), has laid a foundation for the biosynthesis of bicyclogermacrene. To the best of our knowledge, only two studies have reported the biosynthesis of bicyclogermacrene to date. Yan et al. achieved a titer of 107 mg/L in yeast by heterologously expressing a computationally engineered variant of JeTS4 (G91S) (Yan et al. 2022). Huang et al. employed a fusion tag strategy to improve the soluble expression of (+)-bicyclogermacrene synthase from Penicillium expansum (PeTS), which produced 188 mg/L of (+)-bicyclogermacrene (Huang et al. 2022). Nevertheless, systematic efforts toward the efficient microbial synthesis of bicyclogermacrene remain limited, highlighting considerable room for improvement in the titers achieved by engineered strains.Fig. 1. De novo construction of the biosynthetic pathway for (+)-bicyclogermacrene in Escherichia coli. The pathway was divided into two modules: upstream module encompasses conversion of glucose to terpenoid precursors via the methylerythritol phosphate pathway, downstream module involves the cyclization of the sesquiterpene precursor, farnesyl diphosphate, to form (+)-bicyclogermacrene. G-3-P: glyceraldehyde 3-phosphate; DXP: 1-Deoxy-D-xylulose 5-phosphate; IPP: isopentenyl diphosphate; DMAPP: dimethylallyl diphosphate; FPP: farnesyl diphosphate; Dxs: 1-deoxy-D-xylulose-5-phosphate synthase; Idi: isopentenyl diphosphate isomerase; IspA: farnesyl pyrophosphate synthase; PeTS: (+)-bicyclogermacrene synthase from Penicillium expansum. Metabolites are indicated in black, and gene names are indicated in red italics
In this study, we developed an engineered E. coli strain using modular strategy to regulate the synthetic pathway of (+)-bicyclogermacrene. The initial precursor supply was enhanced by integrating 6 copies of key genes expression cassettes, which involved in MEP pathway, into the genome of E. coli K-12 MG1655. The expression of IspA and PeTS in downstream module was fine-tuned via promoter engineering to coordinate the flux of precursor from the upstream module. Subsequently, the titer of (+)-bicyclogermacrene was further boosted to 96.9 mg/L through RBS engineering and fusion protein construction. The augmented supply of NADPH combined with rational design of PeTS, elevating the (+)-bicyclogermacrene titer to 162 mg/L. Finally, the process was scaled to a bioreactor, achieving a final titer of 565 mg/L, which represents the highest level reported to date. This work establishes a platform for sustainable and scalable production of (+)-bicyclogermacrene and provides a framework for the biosynthesis of other valuable sesquiterpenoids.
Materials and methods
Plasmids and strains preparation
All gene amplification were produced by PrimeSTAR® Max DNA Polymerase (TaKaRa, Japan). The PeTS gene encoding (+)-bicyclogermacrene synthase, derived from Penicillium expansum, was cloned from the lab-stored plasmid pPe4F (Huang et al. 2022). All other genes were cloned from E. coli K-12 MG1655. All plasmids, strains and primers in this study were described in Tables S1, S2 and S3, respectively. All plasmids for Gibson assemble were produced by ClonExpress Ultra One Step Cloning Kit (Vazyme, China) and further validated by Sanger sequencing (Tsingke, China). Empty vectors pRSFDuet-1, pCDFDuet-1, pACYCDuet-1 (Novagen, Germany), and lab-stored vector pRSFparaBAD, pACYCDuetTrc, pACYCDuetT5 and pCWori^+^ were selected to construct recombinant plasmids.
The 1-deoxy-D-xylulose-5-phosphate synthase (Dxs) and Idi genes, integrated as multicopy constructs on the E. coli K-12 MG1655 genome, were initially inserted into the reading frame of plasmid pACYCDuetTrc carrying promoter trc. These constructs were then transferred to donor vector pRE57I. The two additional tool plasmids, pTet-tns and pTetQcas containing transposases tnsABC and sgRNA, were saved in our laboratory. To balance the supply of precursors from the upstream IPP pool with the metabolic capacity of the downstream cyclization pathway, we modulated the expression of key enzymes (PeTS and ispA) within the downstream module. To fine-tune the promoter activity within the downstream module, we assembled a series of inducible promoters, which were cloned from our laboratory's plasmid library, while constitutive promoters were introduced via primer-based amplification.
A screening of several vector backbones, including pRSFDuet-1, pCDFDuet-1 and pACYCDuet-1, was conducted to find the suitable expression vectors for downstream applications. Additionally, a series of RBS sites were introduced to modulate the translation initiation rate of PeTS. Codon-optimized flexible linkers were further utilized to connect PeTS and ispA, enabling the formation of fusion proteins for enhanced enzymatic activity and functional synergy. To systemically optimize the metabolic flux, key endogenous E. coli enzymes involved in cofactor biosynthesis were incorporated into the downstream expression constructs to improve cofactor availability.
Subsequently, A small mutant library of PeTS was designed using the PROSS algorithm (Goldenzweig et al. 2016), and transformed into the optimized chassis cells. The resulting variants were screened to identify those that enhanced the efficiency and titer of (+)-bicyclogermacrene production in vivo.
E. coli DH5α and K-12 MG1655 were employed for plasmid construction and production.
Genome editing
The MUCICAT (multicopy chromosomal integration using CRISPR-associated transposases) tool was used to integrate key enzymes of MEP pathway and was performed according to the previous report (Zhang et al. 2020, 2021).
Culture conditions
Luria–Bertani (LB) medium (10 g/L peptone, 5 g/L yeast extract, 10 g/L NaCl) was used for competent cell preparation, plasmid propagation and inoculum preparation. Terrific broth (TB) medium (12 g/L corn steep powder, 24 g/L yeast extract, 12.54 g/L K_2_HPO_4_, 2.31 g/L KH_2_PO_4_, and 5 g/L glucose) was used for biosynthesis in shake flasks. After overnight cultivating of single colony in LB medium at 37 °C and 200 rpm, 1% (v/v) ratio of seed culture was inoculated into TB medium. Isopropylthiogalactoside (IPTG) was added to the TB medium at the final concentration of 0.05 mM when the OD_600_ reached about 0.8, then further 72 h was spent on fermenting product at 25 °C with a cover of 10% (v/v) n-dodecane to extract (+)-bicyclogermacrene. Appropriate antibiotics such as kanamycin, tetracycline, streptomycin and chloramphenicol were added into the medium at the final concentration of 50 μg/mL, 50 μg/mL, 50 μg/mL and 34 μg/mL, respectively.
Fed‑batch fermentation
Two-liquid phase fermentations were performed in a 5-L stirred-tank bioreactor. The overnight-grown seed culture was inoculated at a 5% volume ratio into the fermentation tank, which contained 2 L of culture medium supplemented with the appropriate antibiotics (12 g/L corn steep powder, 24 g/L yeast extract, 5 g/L glucose, 12.54 g/L K_2_HPO_4_, 2.31 g/L KH_2_PO_4_, 0.25 g/L MgSO_4_, 2.7 g/L NH_4_Cl). The culture was incubated at 37 °C and 200 rpm for 4–5 h, with the pH of the culture solution continuously maintained at 6.92 (adjusted with 50% sulfuric acid and 25% ammonium hydroxide). Dissolved oxygen (DO) was controlled in relation to the agitation speed, and maintained at 30%. Once the OD_600_ reached 15, the culture temperature was gradually reduced to 25 °C to induce expression, and IPTG was added to a final concentration of 0.05 mM. Concurrently, 300 mL of n-dodecane was slowly pumped into the bioreactor for two-phase fermentation. During the process, DO was closely monitored, and whenever DO increased above 40%, a glucose solution (final concentration approximately 5 g/L) was manually added to sustain cell growth until the fermentation was completed. Samples were taken every 2.5–5.0 h for OD_600_ measurement and product quantification analysis.
Analytical measurements
The (+)-bicyclogermacrene was extracted in situ by the overlaid n-dodecane layer during biphasic cultivation and subsequently analyzed by gas chromatography–mass spectrometry (GC–MS). The oven temperature was initially held at 60 °C for 2 min, then increased to 140 °C at 10 °C/min, followed by a ramp to 230 °C at 20 °C/min and finally held at 230 °C for 1 min (Huang et al. 2022).
Results
Construction of (+)-bicyclogermacrene biosynthesis pathway in E. coli through multi-copy expression of mep pathway key genes and promoter tuning
The MEP pathway is the primary route in E. coli for supplying the universal terpenoid precursors, IPP and DMAPP (Banerjee and Sharkey 2014). This pathway involves a series of enzymes, including Dxs, Dxr (1-deoxy-D-xylulose-5-phosphate reductoisomerase), IspD (4-diphosphocytidyl-2-C-methyl-D-erythritol synthase), IspE (4-diphosphocytidyl-2-C-methyl-D-erythritol kinase), and IspH ([4Fe-4S] enzyme) (Zhao et al. 2013; Wang et al. 2015), among which the overexpression of Dxs has been demonstrated to be most effective for enhancing terpenoid yields (Zu et al. 2020; Ma et al. 2021). Furthermore, although not a core MEP pathway enzyme, Idi plays a key role for balancing the IPP/DMAPP ratio by catalyzing their interconversion, thereby meeting the specific demands of downstream terpenoid synthesis (Pérez-Gil and Rodríguez-Concepcíon 2013). In this study, we constructed an upstream Dxs-RBS-Idi expression module (RBS: GAAGGAGATATACAT) under the control of a trc promoter, this cassette was then integrated as multicopy cargo into the genome of E. coli K-12 MG1655, generating a series of engineered strains with copy numbers ranging from 0 to 7, were named M0 to M7 (Fig. 2a).Fig. 2. Balancing metabolic flux between the upstream and downstream modules. a The MUCICAT tool was employed for multicopy genomic integration of the upstream module into the E. coli K-12 MG1655 genome. Agarose gel electrophoresis confirmed five recombinant strains harboring varying copy numbers. Nine promoters with different categories and strengths were introduced to regulate the downstream module. b 45 engineered strains were fermented for 72 h under two-phase fermentation condition. The bar graph represented product titer, scatter plot represented the corresponding cell density. The highest titer was marked with a red asterisk. Data were shown as mean ± standard deviation from three replicates (n = 3)
In order to effectively pull metabolic flux through the downstream, we systematically screened a panel of inducible (P1-P5) and constitutive (P6-P9) promoters with strengths ranging from weak to strong to control the expression of downstream module genes (PeTS and IspA). To this end, a library of 9 plasmids was introduced into both the wild-type and the engineered base strains, generating a series of combinatorial recombinant strains designated as MX(X=0–7)-Y(Y=1–9) (Fig. 2a) for subsequent screening.
Following 72 h of fermentation in a two-phase system with 10% n-dodecane, heterologous expression of the (+)-bicyclogermacrene synthase in the wild-type host (M0) produced only trace amounts of the product, indicating that the endogenous MEP pathway in E. coli can support moderate-scale sesquiterpene biosynthesis (Fig. 2b). Under the background of strain M0, M0-3(trc promoter), produced the highest titer of 11.3 mg/L among inducible promoters, whereas the top-performing constitutive promoter strain, M0-6 (J23119 promoter), produced only 1.5 mg/L. The increased genomic copy number of key MEP pathway enzymes raised product titer in all engineered strains, demonstrating that the upstream supply of IPP and DMAPP was the limiting step for the entire pathway flux. Notably, the highest titer of 50.1 mg/L was observed in strain M6-3. However, when the upstream copy number was further increased to 7, M7-3 showed a drop in titer to 45.5 mg/L, accompanied by a significant decrease in OD_600_ (from 24.2 to 19.8). This suggests that 6 copies provided sufficient precursor supply, exceeding this threshold imposed a metabolic burden that impaired both growth and productivity. A similar metabolic principle was observed in the biosynthesis of taxadiene, where the highest titer (approximately 1.02 g/L) was achieved when the upstream MEP pathway module was integrated as a single copy into the genome under the control of an appropriate promoter (Trc). Conversely, overexpression of the upstream module—whether through increased gene dosage or the use of excessively strong promoters—resulted in a decline in production yield (Ajikumar et al. 2010). Furthermore, in any copy number of chassis strains, the product titer achieved with inducible promoter group(P1-P5) was consistently and significantly higher than that with constitutive promoter group(P6-P9), despite no substantial differences in OD_600_ values between the two groups. We hypothesized that the continuous, unregulated expression driven by constitutive promoters may place a constant burden on the host, potentially compromise long-term pathway efficiency (Ye et al. 2016). In contrast, inducible promoters can be intermittently and tunably expressed, mitigating or reducing these issues during prolonged fermentation. Interestingly, within the inducible promoter group, we observed that the moderate-strength trc promoter consistently outperformed high-strength promoters such as tac and T5. We speculate that excessively high transcription rates may overwhelm the cellular machinery for proper protein folding and post-translational modification, leading to misfolding or aggregation, forming inclusion bodies that affect protein solubility and activity (Marschall et al. 2017). Based on these results, M6-3 was selected as the chassis strain for further research.
Evaluation of translation initiation rate and flexible linkers in downstream module
To optimize the expression of the downstream module genes (PeTS and IspA), we first attempted to clone them into vectors with varying copy numbers (reconstructed strains named as M6-Y(Y=10–11)), hypothesizing that the increased vector copy number would elevate the intracellular gene dosage and thereby enhance gene expression. However, elevated vector copy numbers imposed metabolic stress on the host cells (Singha et al. 2017), leading to a nearly half reduction in the titer of (+)-bicyclogermacrene (Fig. 3a). Sodium dodecyl sulfate—polyacrylamide gel electrophoresis (SDS-PAGE) analysis revealed that while PeTS was expressed in a soluble form across all vectors, it predominantly accumulated in inclusion bodies. In contrast, IspA was expressed in a soluble form, and its soluble protein yield increased with the vector copy number (Fig. 3b). The results suggested that disproportionate enzyme expression likely led to excessive FPP formation by IspA, which cannot be efficiently cyclized by the limited soluble PeTS. Consequently, FPP accumulated, creating a metabolic flux imbalance—a phenomenon consistent with reported pathway bottlenecks (Yang et al. 2016).Fig. 3. Systematic engineering of downstream module expression. a (+)-bicyclogermacrene production in strains harboring different plasmid backbones. b SDS-PAGE analysis of soluble PeTS and IspA expression from the corresponding backbones (S: supernatant fraction; P: precipitate fraction). Target proteins were marked with arrows (PeTS: pink arrow, 44.8 kDa; IspA: yellow arrow, 32.2 kDa). c Enhancing PeTS translation efficiency via RBS library screening. The grey circle in the vector map denoted the RBS locus. d Engineering a fusion enzyme by screening flexible linkers between PeTS and IspA to optimize catalytic efficiency. The blue beads in the vector map depicted as flexible linkers. The bar graph represented product titer, scatter plot represented the corresponding cell density. The highest titer was marked with a red asterisk. Data were shown as mean ± standard deviation from three replicates (n = 3)
The strength of ribosome binding site (RBS) serves as a key factor modulating protein expression at the translational level (Wang et al. 2023). To enhance the poor soluble expression of PeTS, we employed a series of engineered RBS sequences with low to high strengths (reconstructed strains named as M6-Y(Y = 12–20)), previously reported for Escherichia coli (https://parts.igem.org/Ribosome_Binding_Sites/Catalog). The theoretical translation initiation rate (TIR, a.u.) of these RBS variants was predicted using the RBS Calculator (https://salislab.net/software/predict_rbs_calculator) (Table S4). As shown in Fig. 3c, the titer of (+)-bicyclogermacrene at low translation initiation rate was lower than that of the control. Increasing the TIR to moderate levels slightly improved product formation, with the highest titer of 60.8 mg/L observed at a TIR of 6992 a.u.. However, SDS-PAGE analysis revealed no apparent improvement in the soluble expression of PeTS across RBS variants, as most of the target protein accumulated as inclusion bodies, with only minor soluble fractions detected (Fig. S1).
Subsequently, we proposed the construction of fusion protein by linking the domains of PeTS and IspA via an appropriate flexible linker, aiming to maintain the functional integrity of both proteins while potentially enhancing their cooperative activity (Xie et al. 2019). We introduced a series of flexible linkers into the downstream vectors (Table S5), which was transformed into the strain M6 to generate recombinant strains M6-Y(Y = 21–29)(Fig. 3d). The linker "GSGGSGGSGGSG" in strain M6-29 achieved a significant increase in (+)-bicyclogermacrene titer of 96.9 mg/L, which was 1.9-fold higher than the non-fusion control. This improvement cannot simply attribute to linker length. Given that the flexibility is predominantly depends on the Gly content and the overall linker length, the hydrophilicity is largely influenced by the proportion of Ser and Thr, the optimal linker exhibited intermediate flexibility and hydrophilicity. This balance likely creates an ideal spatial and hydrophilic microenvironment for the PeTS‑IspA fusion, promoting productive interaction and more efficient biosynthesis (Cheah et al. 2023).
Optimization of cofactor supply for enhanced metabolic flux
Since the MEP pathway employs NADPH as an electron donor in two steps (Kant et al. 2023). Overexpression of this pathway without sufficient cofactor supply can result in the accumulation of intermediates, such as 1-Deoxy-D-xylulose 5-phosphate (DXP), leading to feedback inhibition or cytotoxic effects that adversely affecting cell growth and metabolic processes. Besides, the intracellular NADPH pool primarily sustains basal cellular metabolism, strengthening the MEP pathway creates substantial additional demand on the NADPH pool, the excessive consumption of intracellular cofactor pool triggers a systemic redox imbalance, thereby redirecting central carbon metabolism flux (Wang et al. 2018; Liu et al. 2022). Therefore, optimizing the supply of NADPH is crucial for maintaining the efficiency of overexpressed MEP pathway in engineered strains. To address this, we inserted genes involved in cofactor biosynthesis into the downstream plasmid, which were co-expressed with fusion proteins in strain M6-29, generating recombinant strains designated as M6-Y(Y = 30–36). Among them, Overexpression of 6-phosphogluconate dehydrogenase (Gnd)-glucose-6-phosphate dehydrogenase (Zwf) in strain M6-30 increased the titer of (+)-bicyclogermacrene from 96.9 to 119 mg/L (Fig. 4b), whereas other constructs performed worse than M6-30. We proposed that overexpression of Gnd-Zwf redirected NADPH production via the pentose phosphate pathway (PPP pathway), facilitating redox balance and channeling carbon precursors toward sesquiterpene biosynthesis (Fig. 4a).Fig. 4. Engineering central metabolism to enhance cofactor supply for (+)-bicyclogermacrene biosynthesis. a Central metabolic pathways involved in cofactor regeneration include EMP pathway, PPP pathway, ED pathway and TCA cycle were respectively indicated with green, pink, blue, and black arrows. Interconversion between NADH and NADPH, which indirectly influences cofactor availability, was indicated by yellow arrow. G6P: Glucose-6-Phosphate; F6P: Fructose-6- Phosphate; FDP: Fructose-1,6-Bisphosphate; G3P: Glyceraldehyde-3-Phosphate; PEP: Phosphoenolpyruvate; 6PG: 6-Phospho-gluconate; Ru5P: Ribulose-5-Phosphate; KDPG: 2-keto-3-deoxy-6-phosphogluconate; DXP: 1-Deoxy-D-xylulose 5-phosphate; Zwf: glucose-6-phosphate dehydrogenase; Gnd: 6-phosphogluconate dehydrogenase; Edd: 6-phosphogluconate dehydratase; Eda: 2-keto-3-deoxy-6-phosphogluconate aldolase; Icd: Isocitrate Dehydrogenase; maeA: Malic enzyme; pntAB: pyridine nucleotide transhydrogenase; udhA: soluble pyridine nucleotide transhydrogenase.Overexpressed gene names were indicated in italics. b Impact of overexpressing cofactor-regeneration-related genes on (+)-bicyclogermacrene titer. The bar graph represented product titer, scatter plot represented the corresponding cell density. The highest titer was marked with a red asterisk. Data were shown as mean ± standard deviation from three replicates (n = 3)
Conversely, 6-phosphogluconate dehydratase (Edd) and 2-keto-3-deoxy-6-phosphogluconate aldola (Eda) constitute an alternative NADPH-enhancing strategy from the Entner-Doudoroff pathway (ED pathway), its overexpression led to the accumulation of 2-keto-3-deoxy-6-phosphogluconate (KDPG), disrupting metabolic homeostasis. Likewise, the overexpression of isocitrate dehydrogenase (Icd) and malic enzyme (maeA), key enzymes in the TCA cycle and its auxiliary routes, forces central carbon flux (e.g., from isocitrate and malate) toward α-ketoglutarate and pyruvate, thereby perturbs the natural TCA cycle flux, impairing cell growth. As key bridges connecting the intracellular NAD(H) and NADP(H) pools, overexpression of soluble pyridine nucleotide transhydrogenase (udhA) and pyridine nucleotide transhydrogenase (pntAB) may disrupt the cellular redox state in this study.
Rational design of PeTS improved the production of (+)-Bicyclogermacrene
Enzyme stability often serves as a limiting factor in the microbial production of high-value-added chemicals and commercial enzymes. Rational design provides an effective strategy approach for enhancing in vivo protein stability and folding. For instance, Li increased 1.94-fold in isoprene production by improving isoprene synthase stability through directed evolution (Li et al. 2023). In this study, we employed the Hotspot Wizard and PROSS servers to predict residues that potentially influence the stability of (+)-bicyclogermacrene synthase (Sumbalova et al. 2018; Weinstein et al. 2021). Based on the predictions, 34 amino acid residues were selected and their corresponding single-point mutants were constructed (Fig. S2). The previously engineered strain M6-30, which harbors the wild-type (+)-bicyclogermacrene synthase, was used as the control. Among all mutants, three single-site variants led to higher product titer compared to the wild-type, while other mutations either showed no significant improvement or completely abolished the activity of PeTS. Notably, the combination mutant L352F/E207P (Strain M6-36) exhibited the highest titer of 162 mg/L (Fig. 5).Fig. 5. Enhancement of (+)-bicyclogermacrene production by combining beneficial mutations. N.D. indicated not detected. The highest titer was marked with a red asterisk. Data were shown as mean ± standard deviation from three replicates (n = 3)
Scale-up production of (+)-bicyclogermacrene by strain M6-36 in a 5-L bioreactor
We performed a one-factor-at-a-time optimization of the fermentation conditions in 250 ml shake flasks, investigating four parameters: carbon source type, carbon source concentration, induction temperature, and inducer concentration. However, these modifications did not lead to a significant improvement in the product titer (Fig. S3). The following conditions were selected for bioreactor fermentation: glucose as the carbon source, induction temperature of 25 °C, and IPTG concentration of 0.05 mM. Interestingly, product titer decreased sharply when the glucose concentration exceeded 5 g/L. We hypothesized that high glucose concentrations redirected metabolic flux toward primary metabolic pathways such as glycolysis, which primarily involved in energy generation and precursor synthesis for rapid cell growth and reproduction (Law et al. 2024). This shift likely reduced the flux into the MEP pathway, limiting precursor availability for sesquiterpene synthesis and thus decreasing the product yield (Zhou et al. 2024). Subsequently, we scaled up the fermentation process to a 5-L bioreactor using a 48-h biphasic fed-batch strategy, achieving a final titer of 565 mg/L (Fig. 6). To the best of our knowledge, this represented the highest reported titer of (+)-bicyclogermacrene to date. The achievement of record-breaking titer underscores the potential for large-scale bioproduction, opening avenues for future applications in synthetic biology and industrial biotechnology.Fig. 6. Scale-up production of (+)-bicyclogermacrene via fed-batch, two-phase fermentation in a 5-L bioreactor. The culture pH was maintained at 6.9 throughout the fermentation. When the OD_600_ reached 15, the process was shifted to induce production synthesis by simultaneously adding IPTG to a final concentration of 0.05 mM, reducing temperature to 25 °C, and introducing n-dodecane for in situ product extraction. DO-stat was implied manually to maintain the residual carbon concentration below 5 g/L when DO surges above 40%
Discussion
In this study, we constructed a chassis strain for the de novo biosynthesis of (+)-bicyclogermacrene. By integrating multiple copies of key endogenous enzymes from the MEP pathway into the E. coli genome, we enhanced the precursor supply for terpenoid biosynthesis while alleviating the metabolic burden associated with multi-plasmid system. The relatively low soluble expression level of the heterologous bicyclogermacrene synthase (PeTS) was subsequently identified as a major bottleneck limiting product formation. To address this and maximize the utilization of terpenoid precursors, we systematically optimized the expression of the downstream module by adjusting plasmid copy number, promoter strength, RBS efficiency, and flexible linker screening. Furthermore, considering the high cofactor demand of terpenoid biosynthesis, we implemented complementary strategies to enhance NADPH supply, thereby supporting overall metabolic balance and pathway flux. Through final optimization in a fed-batch biphasic fermentation with glucose as the carbon source, the titer of (+)-bicyclogermacrene reached 565 mg/L, representing a 50-fold improvement over the initial strain.
Previous studies have independently or synergically expressed the MVA and MEP pathways in recombinant microorganisms to enhance precursor pools for terpenoid production, achieving the production of high-value compounds such as amorphadiene, taxadiene and α-bisabolol (Song et al. 2021; Xie et al. 2024; Zhang et al. 2025). However, plasmid-based overexpression of multiple heterologous enzymes often imposes a substantial metabolic burden on host cells and leads to plasmid instability during fermentation (Englaender et al. 2017). To overcome these limitations, we integrated multiple MEP pathway genes into the E. coli genome, which increased (+)-bicyclogermacrene titers from 11.3 to 50.1 mg/L. This result reaffirmed that endogenous levels of IPP and DMAPP in E. coli were insufficient for efficient terpenoid biosynthesis, a finding consistent with reports in other systems (e.g., Integration and modulation of heterologous MVA pathway modules into E. coli chromosome increased β-carotene titer of 51% (Ye et al. 2016).)
Interestingly, during promoter engineering to optimize downstream pathway expression, we observed that using high-copy plasmids or strong promoters reduced (+)-bicyclogermacrene titers compared to low-copy plasmids or weaker promoters. This contrasts with the typical preference for high expression levels in microbial cell factories. Similar phenomena have been reported in geraniol production (Raghavan et al. 2024), where excessive transcription of metabolic genes disrupts intracellular flux balance, leading to reduced product titers.
In E. coli, the PPP pathway is the primary source of NADPH, while other pathways, such as the TCA cycle, pyridine nucleotide transhydrogenases system, and ED pathway, play supporting roles by generating intermediates or indirectly influencing NADPH availability (Lee et al. 2011). Modulating key enzymes in these pathways is a common strategy for optimizing redox balance (Hao et al. 2020; Tang et al. 2023; Yang et al. 2024). In our work, the engineered E. coli preferentially uses the PPP pathway rather than the ED pathway to generate NADPH. The co-overexpression of Gnd and Zwf robustly enhanced the oxidative phase of the pentose phosphate pathway, directly alleviating the NADPH bottleneck limits the MEP pathway and sesquiterpene synthesis, increasing the titer from 96.9 to 119 mg/L. Overexpression of Edd and Eda may therefore lead to excessive accumulation of KDPG, which inhibits the upstream key enzyme phosphoglucomutase, reducing the cellular conversion of glucose-1-phosphate to glucose-6-phosphate and thus limiting the carbon flux into central metabolism. Moreover, ED pathway overexpression competes with glycolysis and the pentose phosphate pathway for the common substrate glucose-6-phosphate, perturbing central carbon metabolism and negatively affecting product biosynthesis. Additionally, overexpression of pntAB or udhA, which mediates NADH/NADPH transhydrogenation, might induce futile cycles and indirectly influence NADPH availability (Islam et al. 2023). Overexpression of Icd enhanced TCA cycle flux and NADPH supply but at the expense of acetyl-CoA and released CO₂, leading to carbon loss and redox stress (Masuda et al. 2017). Moreover, the massive production of its product α-ketoglutarate could inhibit key enzymes of TCA cycle like citrate synthase, through a feedback mechanism, directly inhibiting central metabolism and impairing cell growth. Overexpression of maeA consumes malate, whose replenishment requires an ATP-dependent anaplerotic reactions. Furthermore, the excess produced pyruvate may not channeled specifically into the MEP pathway, resulting in inefficient carbon use (Hua and Shimizu 1999). These results illustrated that plasmid-based overexpression of single cofactor-related genes easily to cause a chain of negative reactions in the central metabolic network, imposing significant metabolic burdens, precursor shortages or toxic intermediate accumulation.
CRISPR-based gene editing tools enable precise transcriptional and translational control of gene expression and have been widely applied in microbial engineering to optimize native pathways or introduce artificial pathways for natural product biosynthesis (Kang et al. 2016; Wang et al. 2018). In future studies, these tools could be applied to integrate downstream module genes into the genome, thereby enabling plasmid-free production of (+)-bicyclogermacrene. In enzyme engineering, we mutated surface amino acid residues of PeTS rather than those within the catalytic pocket and some led to a complete loss of enzymatic activity. While current engineering efforts on terpene synthases primarily focus on residues near the active site to manipulate product diversity or elucidate cyclization mechanisms (Edgar et al. 2017), our findings imply that surface residues may also play critical roles in catalytic activity or enzyme stability (Zhang et al. 2023), warranting further mechanistic studies on the role of surface architecture. Although this study employed an n-dodecane biphasic system for in situ extraction, downstream purification processes have not yet been developed. Future effort could prioritize the development of efficient product recovery strategies. While the modified TB medium employed in this study (with corn steep liquor replacing yeast extract and glucose replacing glycerol) achieved cost reduction to some extent, future large-scale production may involve the development and optimization of a minimal salts medium.
In summary, we successfully introduced multiple copies of key MEP pathway enzymes into the E. coli genome, enabling the efficient de novo biosynthesis of (+)-bicyclogermacrene from glucose. By fine-tuning gene expression levels, maintaining intracellular redox balance and site-directed mutation of PeTS, we alleviated metabolic stress and achieved a final (+)-bicyclogermacrene titer of 565 mg/L in bioreactor, representing the highest production reported to date. This study serves as a valuable reference for constructing sufficient precursor pools and optimizing cofactor supply for (+)-bicyclogermacrene biosynthesis. While synthetic biology researches on (+)-bicyclogermacrene are still in its early stages, continuous innovation with advanced biotechnological tools is essential to achieve further breakthroughs.
Supplementary Information
Additional file 1.
