Combinatorial optimization engineering for the high-level production of shikimic acid in Escherichia coli

Abstract

Shikimic acid is a crucial precursor for the production of indole derivatives and chiral drugs, and the biomanufacturing offers reduced cost and simplified process compared to traditional plant extraction. Current research on industrial-scale production of shikimic acid achieves notable progress, yet the limitations associated with high production cost, complex operation process, and excessive byproduct formation still persist. To tackle these difficulties, this study aimed at establishing a high-efficient shikimic acid-producer through systematic modification and fermentation optimization in Escherichia coli. First, the route for the synthesis of shikimic acid was divided into three modules (i.e., sugar uptake, bioproduct synthesis, and metabolite catabolism), of which the disruption of glucose uptake system, reinforcement of bioproduct synthesis, and block of its catabolism rewired metabolic fluxes and provided the possibility for efficient synthesis. Next, series of systematic engineering involving in deleting pyruvate bypass, reinforcing the pentose phosphate (PP) pathway, and adjusting promoter strength promoted precursor supply and avoided metabolic overflow. The best strain EC022 generated 5.27 g/L of shikimic acid, highlighting the significance of precursor availability. Then, fermentation conditions including substrate concentration, inoculation size, cultivation temperature, induction time, and C/N ratio were comprehensively optimized, which boosted the bioproduction to 8.70 g/L in the optimized conditions. Finally, a final titer of 61.87 g/L shikimic acid was achieved in a 7 L bioreactor. This study developed an efficient shikimic acid-producer with industrial potential through metabolic engineering and bioprocess optimization.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12934-026-03016-8.

Introduction

Shikimic acid, as crucial intermediate in the aromatic amino acid pathway, not only serves as raw material for the production of indole derivatives and chiral pharmaceuticals, but also is the precursor for antiviral drug Tamiflu [1]. It is also used in the food and cosmetics fields due to unique biological activities and safety profiles [2]. Shikimic acid is traditionally obtained by plant extraction or chemical synthesis, whereas these processes encounter scarce raw materials, complex reaction processes, stringent reaction conditions, high production costs, and potential environmental pollution [3, 4]. As an alternative, the biomanufacturing exhibits the superiority in simple process, high titer, and environmental sustainability, thereby garnering extensive attentions during the past decades [5].

Microbial metabolic networks consist of complex, interconnected pathways involving multiple steps, competing reactions, and regulatory nodes, and alteration of the specific pathway may cause global metabolic disruption [6]. Therefore, efficient synthesis of the desirable bioproduct without impairing cell growth poses major challenge. Shikimic acid synthesis generally involves the glycolysis and PP pathways [7], of which glucose is converted into glucose-6-phosphate, and then it individually enters the glycolysis (EMP) and PP pathways for generating phosphoenolpyruvate (PEP) and erythrose-4-phosphate (E4P). Sequential transformation of 3-dehydroquinate synthase, 3-dehydroquinate dehydratase, and shikimic acid dehydrogenase forms the desired product [8, 9]. Effective production requires for rewiring central metabolism by using diverse approaches to maximize target metabolite accumulation while maintaining robust growth. Until date, the metabolic engineering strategies involve three aspects: (i) precursor synthesis (conversion of glucose in PEP and E4P); (ii) product synthesis (condensation to shikimic acid); (iii) product degradation (shikimic acid conversion to aromatic amino acids) [10–12]. These modules are interconnected and collectively determine synthesis efficiency, and therefore a comprehensive genetic engineering approach should be performed for enhancing overall productivity of shikimic acid.

So far, the biorefineries on shikimic acid synthesis center on Corynebacterium glutamicum and Escherichia coli though engineering approaches differ significantly. Shikimic acid degradation is primarily regulated by single shikimic acid kinase in C. glutamicum, and the pathway features numerous native bypasses, necessitating multi-pronged metabolic interventions [13]. Strategies such as RBS optimization to balance expression of shikimic acid pathway genes in the translational level, along with the deletion of 3-dehydroshikimate dehydrase (qsuB), shikimic acid dehydrase (qsuD) and dihydroxyacetone phosphate phosphatase (hdpA) enhance metabolic flux toward shikimic acid [14, 15]. Further improvement was achieved by inactivating the phosphoenolpyruvate: carbohydrate phosphotransferase system (PTS) system and overexpressing inositol transporter (IolT1) and glucokinase (glk) to boost glucose uptake and PEP availability [16]. These combined interventions enabled to synthesize the highest titer of shikimic acid (141 g/L) [17]. However, knockout of shikimic acid kinase creates an auxotrophic requirement that complicates fermentation process, and suboptimal nutrient supplementation negatively impacts the titer. In addition, limited genetic tools and complex fermentation process pose other challenges for industrial scaling-up [18].

In contrast, E. coli, as a promising model microorganism, possesses fewer native shikimic acid bypasses, a well-annotated genome, and highly developed genetic tools, making it more amenable to high-level engineering. Its fermentation does not require cell harvesting or secondary cultivation, reducing process complexity and production cost [19]. Common strategies consists of replacement of the PTS with the non-PTS glucose uptake system [20], reinforcement of the bioproduct pathway [21], and deletion of the competing pathways [22]. For example, a titer of 101 g/L shikimic acid was achieved by replacing the PTS with the non-PTS system, reinforcing biosynthetic pathway, deleting bioproduct catabolic pathway, and using tyrosine-dependent transcriptional regulator, pyruvate kinase, and shikimic acid transporter to reduce precursors consumption [23]. Building upon this foundational strategy, a recent study boosted its production to 126.4 g/L by serine hydroxymethyltransferase and hyperosmolar response protein for enhancing metabolic flux and bioproduct tolerance [24]. Although remarkable achievements have been made, genetic engineering frequently involves complex genetic manipulations that may not be readily adaptable to streamlined industrial process.

To address metabolic bottlenecks and develop novel strategies for high-titer production, this study aims to construct high-performance strains through a modular approach that couples metabolic engineering with fermentation optimization, thereby offering an alternative framework balancing product titer, process simplicity, and scalability. First, various modifications were evaluated to assess their impact on production and identify optimal chassis for subsequent engineering. Then, the non-oxidative glycolysis (NOG) pathway was attempted for investigating its potential for enhancing cellular growth and shikimic acid synthesis. Finally, comprehensive optimization of EMP and PP pathways, shikimic acid synthesis modules, and fermentation parameters was performed, followed by scaling-up validation in a bioreactor. A central insight from this integrated approach is the successful reconciliation of the inherent trade-off between biomass formation and product synthesis. Strategic resource allocation was achieved through dynamic fermentation control and regulated expression systems, which temporally separated robust cellular growth from optimized production phases. This work establishes novel engineering concepts and fermentation strategies for high-level production of shikimic acid and other multi-precursor biochemicals in microbial systems (Fig. 1).

Fig. 1.

Microbial engineering for shikimic acid synthesis. Shikimic acid is synthesized from PEP and E4P through series of enzymatic reactions. To enhance shikimic acid synthesis, several strategies were conducted: (i) replacing PTS system by introducing a Zymomonas mobilis-derived non-phosphorylative glucose uptake system to enhance PEP supply; (ii) blocking pyruvate bypass for reducing byproduct formation and enhance precursor availability; (iii) enhancing the PP and shikimic acid pathways for promoting its synthesis by enhancing metabolic flux through these interconnected pathways; and (iv) optimizing fermentation conditions for altering substrate utilization. Genes: glf, glucose-stimulated protein; glk, glucokinase; aroG, 3-deoxy-d-arabino-heptulosonate-7-phosphate synthase; aroB, 3-dehydroquinate synthase; aroD, 3-dehydroquinate dehydratase; aroE, shikimic acid dehydrogenase; ldhA, lactate dehydrogenase; poxB, pyruvate dehydrogenase; adhE, alcohol dehydrogenase; pta, phosphate acetyltransferase; tktA, transketolase; talB, transaldolase; zwf, glucose-6-phosphate dehydrogenase. G6P glucose-6-phosphate, F6P fructose-6-phosphate, G3P glyceraldehyde-3-phosphate, Ru5P ribulose-5-phosphate; Xu5P xylulose-3-phosphate, R5P ribose-5-phosphate, S7P sedobertulose-7-phosphate, DAHP 3-deoxy-d-arabino-heptulosonate-7-phosphate, DHQ 3-dehydroquinate, DHS 3-dehydroshikimic acid

Materials and methods

Strains and plasmids

E. coli DH5α was employed in plasmids construction and propagation, and E. coli MG1655 was used as the host strain for the synthesis of shikimic acid. All strains and plasmids used in this study are summarized in Table S1 and S2, respectively. The medium-copy number plasmids pTrcHisA and pBAD33 served as the backbone for constructing recombinant plasmids, which was controlled by isopropyl-β-d-thiogalactoside (IPTG)- and L-arabinose-induced promoter P_trc and P_araBAD, respectively. The primers and recombinant plasmids were synthesized and sequenced by Sangon Biotech (Shanghai, China) and Tsingke Biotech (Beijing, China), respectively. The DNA fragment amplification was conducted by KOD One PCR Master Mix (Toyobo, Osaka, Japan), Phanta Flash Master Mix (Vazyme, Nanjing, China), and 2 × Taq Master Mix (Cwbio, Taizhou, China), and restriction endonucleases were obtained from ThermoFisher (Waltham, MA, USA).

Culture condition

The engineered strains were commonly cultivated at 37 ℃, 220 rpm. The Luria-Bertani (LB) medium (containing 10 g/L NaCl, 10 g/L tryptone, and 5 g/L yeast extract) was used for strain cultivation during DNA manipulation, supplemented with appropriate concentrations of the antibiotics, including 100 µg/mL ampicillin, 50 µg/mL kanamycin, 33 µg/mL chloramphenicol, or 25 µg/mL spectinomycin. Auxotrophic growth and fermentation production was performed in M9 minimal and organic culture (OC) medium, respectively. In particular, the M9 medium consisted of 3 g/L KH₂PO₄, 15.11 g/L Na₂HPO₄·12H₂O, 0.5 g/L NaCl, 1 g/L NH₄Cl, 0.25 g/L MgSO₄·7H₂O, 10 g/L glucose, and 1 mL/L trace metal solution (2.4 g/L Fe₃Cl·12H₂O, 0.3 g/L CoCl₂·6H₂O, 0.15 g/L CuCl₂, 0.15 g/L ZuCl₂·4H₂O, 0.3 g/L NaMnO₄, 0.075 g/L H₃BO₃, and 0.5 g/L MnCl₂·4H₂O dissolved in 0.1 M HCl); the OC medium contained 20 g/L glucose, 10 g/L yeast extract, 4 g/L (NH₄)₂SO₄, 7.5 g/L K₂HPO₄·3H₂O, 2.1 g/L citric acid monohydrate, 0.3 g/L ferric ammonium citrate, 0.24 g/L MgSO₄, 0.2 g/L l-tryptophan, 0.2 g/L l-tyrosine, 0.2 g/L l-phenylalanine, 0.01 g/L biotin, and 1 mL/L trace metal solution (0.025 g/L CuSO₄·5H₂O, 0.0029 g/L ZnCl₂·7H₂O, 0.0037 g/L (NH₄)₈Mo₇O₂₄, 0.0247 g/L H₃BO₃, and 0.0158 g/L MnCl₂·4H₂O dissolved in 0.1 M HCl).

Construction of recombinant plasmids

All DNA manipulations, such as polymerase chain reaction, DNA restriction, Gibson assembly, and ligation, were performed following standard procedures by a thermal cycler (Baiheng, Hangzhou, China). Then, these plasmids were introduced into the host for generating engineered strains by the electroporation (Eppendorf, Hamburg, Germany). Main primers used in this study are listed in Table S3. Heterologous genes including glucose kinase and glucose facilitator protein from Zymomonas mobilis, and phosphoketolase from Bifidobacterium adolescentis were codon-optimized and synthesized by Genewiz (Suzhou, China). Additionally, the CRISPR-Cas9 technique was used for knocking out aroK, aroL, ptsG, ptsH, ptsI, pykA, and pykF genes. In specific, plasmid pCas9 were introduced and endonuclease expression was induced, and the cultures were incubated at 30 ℃. Then, the pTargetF and donor DNA were transformed for verifying.

Microbial tolerance test to shikimic acid

Seed culture of the wild-type was inoculated into LB medium supplemented with varying concentrations of shikimic acid (0–10 g/L), and incubated at 37 ℃ for 12 h. Subsequently, an OD₆₀₀ was measured every 2 h over a 12 h period to obtain growth curve.

Shikimic acid synthesis in shaking flasks

Recombinant strains were routinely cultured in LB medium. The shaking flask fermentation procedure was as follows: recombinant strains stored at ‒80 ℃ were streaked onto solid LB medium for the activation. A single colony was then inoculated into 50 mL of LB medium and cultured overnight at 37 ℃. Subsequently, the pre-culture was transferred to either 50 mL of M9 or OC medium with an inoculation volume of 10% (v/v). When an OD₆₀₀ reached 0.6, gene expression was induced by adding IPTG and l-arabinose, and the temperature was adjusted to 30 ℃. The samples were collected for measuring an OD₆₀₀, glucose consumption, and shikimic acid concentration.

High-level production of shikimic acid in a 7 L bioreactor

Engineered strain was inoculated into 30 mL of LB medium and cultured at 37 ℃ for 10–12 h to obtain the primary seeding culture. This primary seed was then transferred at a 1% (v/v) inoculum into 50 mL of LB medium and cultured for another 10–12 h to generate the secondary seeding culture. Subsequently, it was transferred at a 15% (v/v) inoculum into a 7 L bioreactor (Dibier, Shanghai, China) harboring an initial volume of 1.6 L. The following supplements were added: 20 mL 25 g/L L-phenylalanine, 20 mL of 25 g/L L-tyrosine, and 20 mL of 25 g/L L-tryptophan. Additionally, L-arabinose and IPTG were added under specific conditions. The initial glucose concentration was set at 20 g/L. During the fermentation, glucose was continuously supplemented to maintain concentration, resulting in a final working volume of 2 L. The aeration rate was maintained at 3 vvm, and the dissolved oxygen (DO) level was controlled at 30%. The pH was controlled by supplementing with ammonia solution. Samples were collected at 4 h intervals to measure an OD₆₀₀, glucose consumption, and shikimic acid concentration.

Analytical methods

Cell growth was determined by measuring the optical density at 600 nm (OD₆₀₀) was measured using a UV755B UV-Vis spectrophotometer (Youke, Shanghai, China). Glucose and shikimic acid were analyzed by using high-performance liquid chromatography (HPLC) (Waters, Shanghai, China). An Aminex 87 H organic acid column was employed for compound separation and differential detection. The chromatographic conditions were as follows: column temperature of 60 ℃, mobile phase consisting of 2.5 mM H₂SO₄, flow rate of 0.6 mL/min, detector temperature of 40 ℃, injection volume of 10 µL. As for the quantification of shikimic acid, the collected samples were centrifuged and supernatant was acidified. Then, the solution was filtered into sample vials by using a 0.22 μm PCDF filter membrane.

Statistical analysis

Each experiment was conducted at least twice, with a minimum of two independent biological replicates. The error bars were calculated by the function STDEV in Microsoft Excel.

Results

Tolerance test of shikimic acid

To evaluate the effect of shikimic acid on cell growth, growth kinetics was assessed by using different concentrations of shikimic acid (0, 1, 2, 5, and 10 g/L). As depicted in Fig. 2, an OD₆₀₀ reached 4.90 after 12 h in the absence of shikimic acid, while growth rate was declined by the increased shikimic acid concentration. It is worth noting that at a concentration of 2 g/L, its specific growth rate (µ) decreased from 1.12 to 1.08 compared with a concentration of 0 g/L, with a reduction of 3.57%. In particular, its growth was completely repressed upon a higher concentration (5 g/L). Hence, these results indicate that an efficient shikimic acid-producing strain should tackle inhibitory effect, and it is necessary for reinforcing acid tolerance and optimizing fermentation conditions.

Fig. 2.

Microbial tolerance to shikimic acid. Seed culture of the wild-type was inoculated into 50 ml LB medium supplemented with varying shikimic acid concentrations for culturing 12 h at a proportion of 5% (v/v). The OD₆₀₀ was recorded for generating growth curves every 2 h

De novo synthesis of shikimic acid

To construct the chassis with high potential for shikimic acid synthesis, microbial metabolic network was reprogrammed. Shikimic acid is primarily regulated by shikimic acid kinase (aroK and aroL), which is converted into aromatic amino acids (i.e., l-tyrosine, l-phenylalanine, and l-tryptophan) (Fig. 3a) [25]. On this condition, this characteristic negatively affects intracellular availability. Therefore, these genes were knocked out for reinforcing its accumulation in the wild-type, generating auxotroph EC01 (MG1655 ΔaroK ΔaroL). As expected, this strain showed the limited growth with an OD₆₀₀ of 0.15 when cultured for 12 h without amino acid supplementation, with its µ of 0.01 h^{− 1}. In contrast, the supplementation with aromatic amino acids significantly improved its growth, reaching an OD₆₀₀ and µ of 0.63 and 0.05 h^{− 1}, respectively. It represented an improvement of 320.00% in OD₆₀₀ and 400.00% in µ compared to the control (Fig. 3b). This growth restoration was attributed to the inability of strain EC01 to synthesize l-tyrosine, l-phenylalanine, and l-tryptophan unless these amino acids were exogenously provided. Apart from bioproduct catabolism pathway, glucose importation was another factor for hindering high-level synthesis. The PTS that is mainly controlled by phosphate carrier protein IIBC (ptsG), phosphate carrier protein HPr (ptsH), and phosphate carrier protein EI (ptsI) consumes PEP for glucose uptake and cellular growth. Since PEP also serves as key precursor for shikimic acid synthesis, engineering intracellular PTS availability is necessary to increase precursors availability (Fig. 3c) [26]. To compare the differential effects on cellular growth and shikimic acid synthesis between partial PTS disruption combined with shikimic acid pathway knockout versus complete PTS disruption without shikimic acid pathway modification, strains EC02 (MG1655 ΔaroK ΔaroL ΔptsH ΔptsI) and EC03 (MG1655 ΔptsG ΔptsH ΔptsI) were constructed. When cultured in M9 medium, both strains exhibited severely impaired growth, with strain EC02 reaching an OD₆₀₀ of only 0.17 with its µ of 0.01 h^{− 1} after 12 h, while strain EC03 showed similar growth deficiency with an OD₆₀₀ of 0.16 with µ of 0.01 h^{− 1} under the same conditions (Fig. 3d). These results indicate that neither strain EC02 nor EC03 could effectively utilize glucose for the growth. In summary, endogenous PTS disruption in conjunction with bioproduct catabolism knockout significantly compromised glucose uptake, necessitating subsequent implementation of complementary strategies to restore glucose utilization. Taken together, the modifications to both upstream and downstream segments of the bioproduct pathway impair cellular growth, likely due to the disrupted synthesis of essential metabolites. This growth deficiency underscores the need for complementary engineering strategies to restore and optimize growth performance.

Fig. 3.

Reconstructing microbial metabolism for synthesizing shikimic acid. a Eliminating the downstream pathway of shikimic acid. b Growth curves of strain EC01; The engineered strains were transferred into M9 medium at 37 ℃ for 12 h under two nutritional conditions: one supplemented with aromatic amino acids and the other without this supplementation. A final concentration of 1.21 mM L-tyrosine, L-phenylalanine, and L-tryptophan was added. c Modification and optimization of glucose transportation system. d Growth curves of strains EC02 and EC03; The engineered strains were transferred in M9 medium supplemented for the cultivation at 37 ℃ for 12 h. e Gene expression associated with synthetic pathway for enhancing the synthesis. f Fermentation profiles of strains E. coli MG1655, EC04, EC05, and EC06. Abbreviations: L-Phe, l-phenylalanine; L-Trp, L-tryptophan; L-Tyr, L-tyrosine

To enhance synthetic capacity and identify suitable chassis for subsequent experiments, key genes involved shikimic acid synthesis were systematically engineered. The shikimic acid synthesis is controlled by 3-deoxy-d-arabino-heptulosonate-7-phosphate synthase (AroG), 3-dehydroquinate synthase (AroB), 3-dehydroquinate dehydratase (AroD), and shikimic acid dehydrogenase (AroE), of which the former is subject to feedback inhibition by regulator TyrR, a repressor of l-tyrosine and l-tryptophan synthesis (Fig. 3e) [27]. For such that, series of strategies was conducted for enhancing metabolic flux of the synthesis of shikimic acid: (i) The variant D146N of AroG was selected for alleviating the TyrR-mediated feedback inhibition [28]; (ii) The start codon of AroB was altered from AAG to ATG to enhance its expression; (iii) Fusion proteins of AroG^fbr‒AroB^opt and AroD‒AroE was designed for facilitating metabolic turnover and decreasing intermediate formation. Additionally, glucose utilization was restored by using a glucose facilitator from Z. mobilis, and this method avoided PEP consumption [29]. Subsequently, plasmids pT1 (pTrcHisA-P_trc-aroG^fbr-GGGS-aroB^opt-.

P_trc-aroD-GGGS-aroE) and pB1 (pBAD33-P_araBAD-glk-glf) were introduced into the strain EC01, EC02, and EC03 for yielding strain EC04 (EC01 harboring pT1), EC05 (EC02 harboring pT1 and pB1), and EC06 (EC03 harboring pT1 and pB1), respectively. Compared with the wild-type, strains EC04, EC05, and EC06 achieved bioproduct titer of 1.06, 0.94, and 2.09 g/L, corresponding to 53.00-, 47.00-, and 104.50-fold increases, respectively (Fig. 3f). This significant improvement confirms the effectiveness of PTS disruption, non-PTS glucose uptake, and pathway reinforcement. Notably, whereas direct elimination of the downstream pathway impaired cell physiology, introduction of the non-PTS system enhanced PEP accumulation and precursor conservation, leading to a 97.13% higher titer in strain EC06 than in that in strain EC04. Additionally, strain EC04 and strain EC05 benefited from aromatic amino acid degradation, while their auxotrophy for these compounds restricted both biomass accumulation and final titer. In contrast, the superior performance of strain EC06 reflected synergistic effect of pathway enhancement via overexpression of aroG^fbr, aroB^opt, aroD, and aroE, circumvention of TyrR-mediated feedback inhibition by AroG^fbr, and efficient intermediates channeling through a fusion protein. Collectively, PTS disruption, introduction of the non-PTS glucose uptake system, and pathway reinforcement redirected carbon flux toward shikimic acid synthesis, although further optimization was needed.

Modifying the NOG pathway for forming the growth-dependent pathway

As mentioned above, we have confirmed the feasibility of glucose into shikimic acid, whereas its titer was relatively low. Therefore, we speculated that the unbalanced metabolic flow between cell growth and product synthesis might be one of essential factor for the reduced bioproduction. On the one hand, the conversion of pyruvate into acetyl-CoA will release 1 molecule of CO₂, and carbon loss reduces the yield. On the other hand, the introduction of the NOG pathway directly forms 3 molecules of acetyl-CoA from glucose. This process does not involve carbon release, and theoretically generates more energy by channeling into TCA cycle. Therefore, we attempted to assess the effect of the NOG pathway on shikimic acid synthesis.

To evaluate potential application of the NOG pathway in shikimic acid synthesis, its effects on cell growth and bioproduct synthesis were investigated by modifying the EMP pathway. In the EMP pathway, pyruvate kinases (pykA and pykF) convert pyruvate to acetyl-CoA, which subsequently enters the TCA cycle to support the growth (Fig. 4a) [30]. To determine whether E. coli could utilize the NOG pathway for the growth, these genes were knocked out, generating strains EC07 (EC03 ΔpykA) and EC08 (EC03 ΔpykA ΔpykF). Then, plasmid pB1 was introduced for assessing the impact of impaired EMP pathway on cell growth, yielding strains EC09 (EC07 harboring pB1) and EC010 (EC08 harboring pB1). It could be seen that incomplete glycolysis pathway leads to a decrease in metabolic flux, thereby reducing growth rate (Fig. 4b). Subsequently, plasmid pB2 was constructed by incorporating phosphoketolase, generating strains EC011 (EC07 harboring pB2) and EC012 (EC08 harboring pB2). As shown in Fig. 4c, the resulting strain EC011 and EC012 achieved an OD₆₀₀ of 3.72 and 3.33 with its µ of 0.31 h^{− 1} and 0.28 h^{− 1}, representing a 197.60% and 246.88% increase compared to strain EC09 and EC010, respectively (Fig. 4c). Growth profiles demonstrated that down-regulation of the EMP pathway severely impaired cellular growth, whereas subsequent introduction of the NOG pathway partially restored the growth, indicating its potential for decoupling shikimic acid synthesis from microbial growth.

Fig. 4.

Synergetic optimization of the NOG and glycolysis pathways for shikimic acid synthesis. a The modification strategy of introducing the NOG pathway. b Growth curves of strains EC09 and EC010. c Growth curves of strains EC011 and EC012. d Schematic diagram of optimized recombinant plasmids. e Bioproduct titer. f Growth curves; These strains were inoculated in OC medium at 30 °C for 120 h, and their OD₆₀₀ was measured at an interval of 12 h. ^**p < 0.01, ^*p < 0.1, ns p > 0.01

Having established the functionality of NOG pathway in supporting the growth, its influence on shikimic acid synthesis was subsequently examined. Strains EC013 (EC07 harboring pT1 and pB2) and EC014 (EC08 harboring pT1 and pB2) were constructed, whereas the final titer merely reached 0.62 g/L and 0.42 g/L, with an OD₆₀₀ of 16.28 and 17.18, respectively (Figs. 4e and f). These results indicated that the EMP pathway in strain EC013 remained partially functional, allowing improved cell growth and higher bioproduct formation. To further investigate whether the impaired EMP pathway was responsible for the decreased titer, strain EC015 (EC03 harboring pT1 and pB2) was constructed, which produced 1.29 g/L of shikimic acid with an OD₆₀₀ of 13.64. Although its growth was lower than that of strains EC013 and EC014, the titer exceeded both, yet remained below that of strain EC06. This suggested that direct introduction of the NOG pathway caused metabolic imbalance, thereby limiting shikimic acid synthesis. Consequently, applying the NOG pathway for shikimic acid synthesis required further metabolic engineering. To evaluate the feasibility of implementing the NOG pathway for shikimic acid synthesis, the inducible promoters were replaced with growth-dependent promoters, yielding plasmids pT2 and pB3 (Fig. 4d). For comprehensive analysis of NOG pathway effects on shikimic acid synthesis, these plasmids were introduced into strains EC03 and EC08, generating strains EC016 (EC03 harboring pT2 and pB3) and EC017 (EC08 harboring pT2 and pB3). Notably, strain EC016 and EC017 achieved an OD₆₀₀ of 23.18 and 15.22 with the titers of 3.85 g/L and 0.82 g/L, respectively. Compared to strain EC015, strain EC016 exhibited 69.94% higher growth rate and 198.45% increased production, while strain EC017 showed 6.96% lower growth rate but 95.23% higher titer than strain EC014 (Figs. 4e and f). These results confirmed that dynamic control of the specific genes enabled effective application of the NOG pathway for shikimic acid synthesis, while how to coordinate carbon flux at glucose-6-phosphate node was crucial for shikimic acid synthesis. Together, direct introduction of the NOG pathway is insufficient to enhance shikimic acid synthesis, while its targeted implementation via systematic engineering and dynamic regulation represents a promising strategy.

Combinatorial metabolic engineering for enhancing shikimic acid synthesis

Given that the introduction of the NOG pathway failed to enhance shikimic acid synthesis while multi-branched pathways reduced precursor availability, strain EC06 was selected for the following experiments. Next, a comprehensive approach was performed for determining metabolic bottlenecks for the high-level synthesis. On the basis of strain E. coli ΔpoxB ΔldhA ΔadhE Δpta, the strain EC018 (MG1655 ΔpoxB ΔldhA ΔadhE Δpta ΔptsG ΔptsH ΔptsI) was generated for minimizing pyruvate consumption and maximizing intracellular PEP availability (Fig. 5a).

Fig. 5.

Enhancement of precursor pathway in the synthesis of shikimic acid. a The disruption of pyruvate catabolic pathways. b Genes overexpression of PP pathway for enhancing metabolic flux and promoting bioproduct synthesis. c Schematic diagram of recombinant plasmids. d Bioproduct titer in strains EC019‒EC022; Plasmids were designed to modify the target gene on the plasmid, which was subsequently introduced into strain EC018 to construct the corresponding strains. The resulting recombinant strains were cultured in OC medium at 30 °C for 120 h under shaking flasks conditions. ^*p < 0.1

First, as key precursor for shikimic acid synthesis, increasing E4P supply represents an effective strategy. Generally, the E4P is generated by transketolase and transaldolase in the PP pathway (Fig. 5b) [21, 23]. To enhance intracellular E4P supply, these genes were amplified for yielding plasmid pB4. Subsequently, introduction of pT1 and pB4 yielded strain EC019, which achieved a titer of 2.75 g/L in shaking flasks ‒ a 31.58% increase over strain EC06 (Fig. 5d). It indicated that blocking competing pathways and reinforcing PP pathway effectively reduced metabolic overflow and improved precursor supply, thereby enhancing the flux toward shikimic acid synthesis.

Secondly, in the consideration of positive function of tktA and talB genes on the synthesis, PP pathway was further optimized for enhancing E4P supply. Entry of glucose-6-phosphate into the PP pathway is mediated by glucose-6-phosphate dehydrogenase, a reaction that also generates NADPH (Fig. 5b) [31, 32]. To increase NADPH availability and strengthen metabolic flux, it was magnified for resulting in strain EC020 (EC018 harboring pT1 and pB5), achieving a titer of 3.04 g/L, marking an increase of 10.54% over strain EC019 (Fig. 5d). This suggested that glucose-6-phosphate dehydrogenase expression enhanced synthesis rate and underscored the importance of enhancing key precursors. Taken together, optimization of precursor-supplying pathways such as the PP pathway remains a promising strategy for enhancing shikimic acid production.

Finally, gene expression directly affects protein abundance and product synthesis efficiency [33]. To improve their expression, those genes related to the bioproduct synthesis were modified by replacing the linker with the co-expression strategy, yielding strain EC021 (EC018 harboring pT3 and pB5) (Fig. 5c). As depicted in Fig. 5d, a 32.57% increase in the titer relative to strain EC020, while an OD₆₀₀ decreased from 21.44 to 20.77 (a 3.13% decrease), which demonstrates the superiority of the co-expression over fusion protein strategies for enhancing the production in plasmid-based systems. It also indicated that the usage of fusion protein might affect the enzyme’s catalytic activity and decrease reaction rate. On this condition, we speculated whether optimizing the PP pathway would be beneficial for controlling competing pathways and achieving the most optimal metabolic distribution. In pB5, zwf gene was regulated by the strong promoter P_T5, whereas the other genes were controlled by the inducible promoter P_araBAD. Although P_T5 driven higher expression, its overexpression might lead to imbalanced metabolic distribution and consequently reduced the accumulation. Therefore, zwf gene was controlled by the P_araBAD and resulted in plasmid pB6, which was subsequently evaluated for its ability to improve shikimic acid production. These plasmids were introduced into strain EC018 for yielding strain EC022 (EC018 harboring pT3 and pB6). Strain EC022 produced 5.27 g/L shikimic acid ‒ a 73.36% increase over strain EC020 ‒ with an OD₆₀₀ of 23.71, representing a 33.20% increase relative to strain EC020 (Table S4 and Fig. 5d). These incremental improvements demonstrate that modulating the expression of tktA, talB, and zwf via promoters of varying strengths effectively balances NADPH regeneration and carbon flux partitioning, underscoring the importance of modular optimization.

Optimizing fermentation conditions for reinforcing the synthesis of shikimic acid

To identify the factors influencing bioproduct synthesis and determine the optimal fermentation conditions, a comprehensive optimization of multiple parameters was conducted in the best strain EC022. Here, series of factors including substrate concentration, inoculation size, cultivation temperature, C/N ratio, and induction time were systematically evaluated.

First, the effect of the medium on the synthesis was tested. Cultured in OC medium, a titer of 1.25 g/L was completed (OD₆₀₀: 11.60), whereas only 0.15 g/L (OD₆₀₀: 3.94) was produced in M9 medium (Fig. 6a). These results indicate that yeast extract, as an organic nitrogen source, effectively enhances cell growth and bioproduct synthesis, suggesting its essential role in achieving high titers. Next, to investigate the influence of organic nitrogen content on the titer, further modifications were evaluated in OC medium: (i) increasing yeast extract concentration to 20 g/L (Mode II), and (ii) supplementing with 5 g/L peptone (Mode III). Compared to the unmodified OC medium (Mode I, 3.95 g/L, OD₆₀₀: 24.40), Mode II reduced the titer to 3.65 g/L with an OD₆₀₀ of 25.18, while Mode III increased its production to 4.81 g/L accompanied by an elevated OD₆₀₀ of 26.34 (Fig. 6b). These results showed that excessive yeast extract promoted cell growth but was unfavorable for shikimic acid synthesis, likely because it strengthened competing pathways and increased byproduct formation. In contrast, supplementing a moderate amount of yeast extract with peptone provided richer nutrients, supporting the growth and product accumulation. In the meantime, the impact of glucose feeding strategies was also examined: the fed-batch fermentation yielded 6.10 g/L shikimic acid with an OD₆₀₀ of 19.76, while batch culture resulted in 5.31 g/L with an OD₆₀₀ of 24.06. Although growth rate increased by 21.76% under batch culture compared to the fed-batch fermentation, bioproduct titer was decreased by 12.95% (Fig. 6c). These results suggested that directly supplying high concentrations of glucose, while stimulating growth might promote byproduct accumulation and reduce shikimic acid synthesis. In contrast, the fed-batch feeding allowed better control of metabolic flux toward the desired product. For subsequent fermentations, the medium might be optimized by supplementing an appropriate amount of yeast extract with peptone to provide balanced nutrients, thereby improving both cell growth and bioproduct synthesis. Additionally, a fed-batch feeding strategy should be maintained to avoid negative effects associated with high glucose concentrations. Taken together, nutritional components should be maintained within an optimal range to maximize the production.

Fig. 6.

The optimization of fermentation conditions on the synthesis. a Yeast extract. b C/N ratio. c Fermentation mode. d Temperature. e Inoculation proportion. f The pre-induction. g Induction time. h Fermentation profiles of the unoptimized (left) and optimized (right) conditions. The initial glucose was 20 g/L, and then 10 g/L of glucose was fed for reaching a final concentration of 50 g/L at an interval of 24 h; The dotted line indicates the added glucose

Then, the growth-associated parameters were subsequently optimized. On the one hand, the effect of temperature on product synthesis was evaluated. Given that 30 °C and 37 °C represent the typical cultivation temperatures for product synthesis and cell growth, respectively. Therefore, these two conditions were selected to assess the growth-production trade-off. As depicted in Fig. 6d, the cultivation at 30 °C resulted in a titer of 9.92 g/L with an OD₆₀₀ of 25.21, compared to only 2.16 g/L at 37 °C with an OD₆₀₀ of 32.33. Although the growth rate increased by 28.24% at 37 °C relative to 30 °C, the titer decreased by 77.93%. This suggested that this strain preferentially directed carbon toward biomass formation at high temperature, while low temperature was more favorable for the synthesis of shikimic acid. On the other hand, the evaluation of inoculation sizes revealed that a 10% (v/v) proportion yielded a higher shikimic acid concentration (reaching 6.75 g/L). Nevertheless, increasing the inoculation to 20% (v/v) and 40% (v/v) reduced the titer to 6.31 g/L and 5.01 g/L, respectively. Compared to a 10% (v/v) proportion, these conditions reduced the production by 6.51% and 25.78%, respectively (Fig. 6e). Taken together, shikimic acid titer requires precise control of both the temperature and biomass concentration within optimal ranges to maximize the titer.

Finally, considering gene expression was controlled by inducible promoters, temperature and induction time significantly affect synthetic performance. Early induction during the pre-log phase increases metabolic burden due to resource competition, while late induction reduces protein expression and product synthesis. To determine optimal induction time, it was first applied at an OD₆₀₀ of 0.6 during seed culture. Compared with non-induced seed cultures, an OD₆₀₀ increased from 22.04 to 24.04 (a 9.07% increase), yet the titer decreased from 5.37 g/L to 3.91 g/L (a 27.50% decrease) (Fig. 6f). Further tests involving induction at 0, 3, 6, and 9 h revealed that the induction at 6 h resulted in an OD₆₀₀ of 27.64 but only yielded 2.45 g/L shikimic acid, whereas induction at 3 h achieved 5.76 g/L shikimic acid with an OD₆₀₀ of 21.1 ‒ similar to the 5.98 g/L obtained with induction at 0 h (OD₆₀₀: 19.32, representing a 9.32% difference from 3 h induction) (Fig. 6g). Thus, the induction during seed culture was omitted, and the induction at 3–4 h was selected to avoid premature metabolic burden. Taken together, optimal bioproduct titer under inducible promoter regulation requires precise induction time. Premature induction increases metabolic burden, whereas delayed induction fails to maximize promoter activity, both leading to the reduced titer.

Based on these results, an integrated fermentation strategy was implemented: supplementation of OC medium with 5 g/L peptone, induction at the mid-growth stage, temperature shift from 37 °C to 30 °C at 3 h, and fed-batch feeding. On this condition, strain EC022 was subsequently cultured under the unoptimized and optimized conditions for direct comparison. After a total of 120 h fermentation period, the unoptimized conditions resulted in 6.88 g/L of shikimic acid, with its OD₆₀₀, yield, and productivity of 24.78, 0.22 g/g, and 0.06 g/L/h, respectively. In contrast, under the optimized conditions, a final bioproduct concentration reached 8.70 g/L, with its OD₆₀₀, yield, and productivity of 20.87, 0.24 g/g, and 0.073 g/L/h, respectively. Although growth rate decreased by 15.78%, its production increased by 26.45%, indicating effective redirection of carbon flux toward shikimic acid (Fig. 6h). Therefore, combined nutritional enhancement and dynamic bioprocess control led to a significant improvement in the production, demonstrating the effectiveness of systematic fermentation optimization in metabolic engineering applications.

High-level production of shikimic acid in a 7 L bioreactor

To evaluate the industrial-scale production potential of strain EC022, shikimic acid synthesis was scaled up in a 7 L bioreactor. The fermentation conditions consisted of OC medium, inducer supplementation and temperature reduction to 30 ℃ at 4 h, followed by intermittent glucose feeding (10–20 g/L every 4 h from 24 h) (Mode I). After 72 h of fermentation, this process achieved a titer of 25.8 g/L, with its yield, productivity, and OD₆₀₀ of 0.13 g/g, 0.36 g/L/h, and 30.52, respectively (Fig. 7a). Bioproduct titer showed an increase compared to shaking flasks, yet remained suboptimal, highlighting the need for further optimization to improve bioprocess efficiency.

Fig. 7.

Shikimic acid production using different fermentation modes in a 7 L bioreactor. a Mode (I) b Mode (II) c Mode III; The dotted line indicates the temperature changes; strain EC022 was cultivated in a 7 L bioreactor using the previously described fermentation strategy. The temperature was shifted from 37 ℃ to 30 ℃ at 8 h post-inoculation. After 24 h of fermentation, the glucose monitoring interval was reduced from every 4 h to every 2 h, and glucose feeding was adjusted based on real-time residual glucose concentration in the medium

To improve cell growth and substrate utilization, we implemented a temperature-shift strategy wherein the culture was initially incubated at 37 ℃ for 8 h to promote biomass accumulation, then shifted to 30 ℃ with inducer addition, and finally returned to 37 ℃ upon growth deceleration over a 72 h fermentation period (Mode II). This approach increased an OD₆₀₀ by 65.86% to 50.62 and glucose consumption by 14.67% to 232.58 g/L relative to Mode I; however, the titer remained relatively low at 27.91 g/L, with a yield and productivity of 0.12 g/g and 0.39 g/L/h, respectively (Fig. 7b). These results indicated that this approach significantly enhanced biomass formation and glucose uptake, whereas it did not proportionally improve the production, suggesting a potential decoupling between cell growth and product synthesis under the applied conditions.

To maximize shikimic acid production by balancing growth and synthesis phases, the initial 37 ℃ phase was extended to 8 h for biomass accumulation and then maintained 30 ℃ throughout the remainder of the 72 h fermentation (Mode III). Notably, this strategy resulted in a titer of 61.87 g/L ‒ a 121.68% increase over Mode I ‒ along with a glucose consumption of 294.62 g/L (26.67% higher than Mode I), an OD₆₀₀, yield, and productivity of 49.15, 0.21 g/g, and 0.91 g/L/h, respectively (Fig. 7c). However, it could be found that an OD₆₀₀ was decreased, probably attributed to acid repression. Although tolerance test indicated that directly adding a higher concentration of shikimic acid would inhibit cell growth, through fermentation, microbial tolerance could be enhanced. Nevertheless, a higher bioproduct concentrations remained one of key factors limiting efficient growth. Therefore, it is important for improving its tolerance to sshikimic acid in the following experiments. Despite a higher titer was achieved under Mode III, the overall product yield was relatively low, suggesting that a considerable fraction of consumed glucose was diverted toward biomass formation, energy metabolism, or the accumulation of potential by-products. Overall, these results demonstrated that efficient shikimic-producer was established by combining metabolic engineering strategies and bioprocess optimization (Fig. 8). Taken together, the implementation of Mode III strategy, building upon Modes I and II, successfully reconciled cell growth and bioproduct synthesis, resulting in a substantial improvement in shikimic acid titer and overall process efficiency (Table 1).

Fig. 8.

Schematic diagram of the strategies on shikimic acid improvement. Synthetic efficiency of engineered strains was continuously improved with the application of various modification strategies

Table 1.

Shikimic acid production using different fermentation modes in a 7 L bioreactor

Mode	Consumed glucose (g/L)	OD600	Titer (g/L)	Yield (g/g)	Productivity (g/L/h)
I	198.46 ± 0.13	30.52 ± 0.57	25.80 ± 0.79	0.13 ± 0.00	0.36 ± 0.00
II	232.58 ± 0.04	50.62 ± 0.48	27.91 ± 0.10	0.12 ± 0.00	0.39 ± 0.00
III	294.62 ± 0.31	49.15 ± 1.48	61.87 ± 0.47	0.21 ± 0.01	0.91 ± 0.00

Conclusions and discussion

Significant progress has made in designing high-performance shikimic acid-producing strains during the past decades, and such efforts greatly promoted industrial biotechnology. To identify key determinants of shikimic acid synthesis and define future directions, strain E. coli MG1655 was selected as the chassis for constructing high-titer producer in this study. First, the modifications to bioproduct catabolism and implementation of a non-PTS system were performed to evaluate their individual and combined effects on the synthesis. Then, the NOG pathway was attempted, while multiple metabolic bypass led to metabolic diversion and inefficient synthesis. Subsequently, the deletion of pyruvate bypass and reinforcement of PP pathway increased precursor availability, and promoter engineering optimized genes expression related to shikimic acid pathway, collectively improving its efficiency. Finally, the best strain EC022 achieved a final titer of 61.87 g/L in a 7 L bioreactor through systematic optimization of fermentation parameters. The stepwise engineering approach illustrates how targeted genetic modifications, when combined with dynamic process control, can resolve classic trade-off between biomass formation and bioproduct synthesis, establishing a scalable framework for microbial production of shikimic acid and other value-added biochemicals.

The implementation of the NOG pathway established promising concept and methodology for developing high-performance biorefineries with a high yield. Metabolic engineering strategies offer excellent opportunity for enhancing carbon flux toward target bioproducts and improving production efficiency. As an emerging alternative to the EMP pathway, the NOG pathway rearranges fructose-6-phosphate metabolism for yielding 3 molecules of acetyl-phosphate that are subsequently converted to acetyl-CoA [34]. Theoretically, the increased acetyl-CoA supply enhances ATP generation and improve biomass formation. Owing to its metabolic flexibility, the NOG pathway has applied for synthesizing biofuels, terpenoids, and other acetyl-CoA-derived compounds. For example, its introduction elevated acetyl-CoA and acetyl-phosphate pools, thereby increasing the supply of direct precursor for acetone synthesis and improving acetone yield. Similarly, in the production of citramalate, the combination of the NOG pathway with acetate kinase overexpression enabled more efficient conversion of acetyl-phosphate to acetyl-CoA, enhancing precursor availability. Compared to the control, this strategy resulted in a 35.0% increase in titer and a 33.3% improvement in yield [35, 36]. Although acetyl-CoA is not a direct precursor of shikimic acid, its elevated pools may improve energy metabolism and biomass formation, and the NOG-mediated growth reduced PEP consumption, suggesting potential synergistic benefits. In this study, strain EC03 was selected as the chassis for preventing excessive genetic perturbations that could confound the assessment of NOG pathway effects on the growth and production. This allowed systematic evaluation of NOG pathway effects on both cell growth and pathway efficiency, enabling an analysis of its industrial applicability for shikimic acid production. First, downstream pathways of the EMP pathway were modified and the NOG pathway was introduced to evaluate growth effects. Then, the NOG pathway was coupled with the shikimic acid synthesis module to analyze synthesis performance. Finally, inducible promoters were replaced with the growth-dependent promoters for dynamic control. Unfortunately, the strain EC017 only achieved a titer of 0.82 g/L, with its OD₆₀₀ of 15.22. While the NOG pathway sustains strain growth with compromised central carbon metabolism, its expression reduces the capacity. This decline is attributed to the increased branching at fructose-6-phosphate node, which intensifies precursor competition between biomass formation and product synthesis. Furthermore, direct disruption of the primary pathway impairs cellular physiology, and these perturbations synergistically reduce the metabolic rate. Although this pathway alone was insufficient to enhance product synthesis, its exploration served as a proof-of concept for sustaining cell growth under glycolytic impairment and revealed metabolic constraints, including increased flux branching and cofactor imbalance. Notably, the partial restoration achieved with growth-dependent promoters underscores the importance of balancing pathway activity with cellular physiology. These findings indicate that dynamic control is more effective in enabling the NOG pathway to function and in further improving product yield. Overall, this study establishes a conceptual framework for applying the NOG pathway to shikimic acid synthesis and offers a promising strategy for developing advanced microbial cell factories.

Enhancing the supply of synthetic precursors represents a key strategy for improving bioproduct titers. Generally, the PEP serves as crucial precursor for shikimic acid synthesis and cell growth, and the engineering on the EMP provides alternative approaches to enhance PEP availability except for the non-PTS system. Current strategies for enhancing PEP accumulation include the overexpression of phosphoenolpyruvate synthase (ppsA) to improve pyruvate-to-PEP conversion efficiency [23], or knocking out of dihydroxyacetone kinase (dhaL) to reduce PEP degradation [24]. In contrast to these direct interventions, this study adopted an alternative strategy focusing on minimizing pyruvate consumption to indirectly increase PEP supply, based on severe disruption of EMP pathway severely compromises cellular growth, which was achieved through targeted knockout of pyruvate-consuming bypass pathways [31, 32]. The targeted gene knockouts reduced non-essential pyruvate consumption, increased PEP supply, and enhanced pyruvate flux toward the TCA cycle, thereby boosting acetyl-CoA generation and improving provision of both precursors and energy for cell growth and product synthesis. Beyond the PEP, E4P availability constitutes another major bottleneck in shikimic acid production. Previous studies have shown that the reinforced tktA and talB genes enhances E4P supply by strengthening PP pathway [23, 24]. To further improve E4P generation, the PP pathway was systematically investigated. Glucose-6-phosphate entry into the PP pathway is primarily regulated by glucose-6-phosphate dehydrogenase, a step that also generates NADPH. Since the conversion of 3-dehydroquinate to shikimic acid requires NADPH, cofactor availability ‒ including NADPH supply ‒ may also limit the synthesis of shikimic acid [37, 38]. Consequently, zwf gene was overexpressed and led to a titer of 3.04 g/L. Subsequent optimization of shikimic acid pathway and zwf gene expression to balance metabolic fluxes, achieving a titer of 5.27 g/L in strain EC022 ‒ a 73.36% increase over strain EC020. These results demonstrate that coordinated engineering of pyruvate bypass, PP pathway, and shikimic acid synthesis modules significantly enhances production, underscoring the importance of precursor-oriented optimization in microbial metabolic engineering. The targeted re-routing of pyruvate flux conserved PEP without imposing a growth penalty ‒ a contrast to strategies involving disruptive inhibition of central EMP pathway. Furthermore, coordinated overexpression of tktA, talB, and zwf genes balanced NADPH regeneration, with promoter fine-tuning preventing redox stress. Through modular optimization from strain EC020 to EC022, a 73.36% increase in titer was achieved while growth characteristics were maintained, effectively separating biomass formation from synthesis phases. This phased strategy, which prioritizes establishing robust growth before enhancing product-specific pathways, offers a generalizable framework for high-level microbial biosynthesis.

Systematic optimization of fermentation conditions offers crucial direction for industrial production, as environmental factors significantly influence microbial performance and bioproduct synthesis [39]. Current fermentation optimization strategies include glycerol supplementation as a co-substrate alongside glucose, pH control, and DO adjustment to improve shikimic acid synthesis [40]. To systematically investigate how fermentation parameters influence its production, a comprehensive evaluation was conducted by varying medium composition, induction time, inoculum density, and temperature, which enabled identification of optimal conditions for shikimic acid accumulation while providing insights for developing novel engineering strategies based on fermentation parameter effects. Analysis revealed that cellular growth critically affects production efficiency: insufficient growth fails to establish adequate biocatalytic capacity, while excessive growth promotes byproduct formation and reduces shikimic acid synthesis rates. Therefore, maintaining intermediate growth range proved essential for maximizing the titer. Further experiments demonstrated that strain EC022 preferentially directs metabolism toward growth at high temperature and shikimic acid production at low temperature. To leverage this physiological differentiation, a temperature-shift strategy was implemented: the cultures were initiated at 37 ℃ for rapid biomass accumulation, then switched to 30 ℃ after 3–4 h to enhance shikimic acid synthesis. This approach established robust cellular capacity while minimizing potential feedback inhibition, achieving 8.70 g/L shikimic acid. Although this study identified 30 ℃ as conducive for shikimic acid synthesis and 37 ℃ as optimal for cell growth, further optimization using finer temperature gradients is needed to identify the exact thermal threshold for decoupling biomass accumulation from product formation in large-scale bioreactors. Scaling-up to a bioreactor with extended thermophilic growth phase and dynamic feeding further improved titer to 61.87 g/L, with its yield and productivity reaching 0.21 g/g and 0.91 g/L/h, respectively. Fermentation optimization revealed a critical balance between biomass accumulation and metabolic specialization for synthesis. Product efficiency depends not on maximal growth rate but on an optimal physiological window that establishes sufficient biocatalytic capacity without triggering overflow metabolism. A temperature-shift strategy was implemented to temporally separate these competing objectives: a 37 ℃ phase for efficient cellular platform construction, followed by a shift to 30 ℃ to redirect metabolism toward shikimic acid synthesis, thereby minimizing resource competition that limits titer in constant-temperature cultures. An extended thermophilic phase under controlled conditions further optimized biomass, while dynamic feeding matched glucose supply to the shifted metabolic demand at 30 ℃. The substantial improvements in the titer, yield, and productivity demonstrate that growth-production decoupling via temperature control is a scalable process. Using temperature as a ubiquitous environmental cue to dynamically steer metabolism offers a practical, easily implementable alternative to complex genetic circuits for flux control in industrial bioprocessing. In summary, systematic fermentation optimization through phased temperature regimes effectively resolves the classic tension between growth and production. The “grow then produce” paradigm ‒ first building a robust cellular platform and then redirecting metabolism toward the target pathway ‒ enables concurrent achievement of high titers and process efficiency. This approach provides a generalizable framework for microbial synthesis of shikimic acid and other metabolites whose precursor supply is tightly coupled to central growth metabolism.

While previous studies have established promising microbial platform for shikimic acid synthesis, the limitations in current engineered strains necessitate novel engineering strategies to advance industrial implementation [23, 24]. This study addresses key limitations through integrated metabolic engineering and systematic fermentation analysis, revealing the competition between growth and production. Unlike previous studies reliant on complex genetic interventions, this work proposed a stepwise modular strategy integrating precursor pathways engineering, heterologous NOG pathway deployment, and temperature-shift fermentation strategy. This integrated approach is operationally straightforward and confers marked advantages in process robustness and scalability. Thus, the study provides a complementary engineering framework that balances high productivity with practical industrial applicability. Nevertheless, the overall bioproduct yield still remains below the theoretical maximum because carbon is diverted toward biomass formation and by-product accumulation, and the fermentation protocol requires further optimization to enhance metabolic turnover. To overcome these constraints, future efforts should employ omics-based approaches to elucidate flux distribution across competing pathways, and implement dynamic pathway regulation coupled with tolerance engineering to mitigate product inhibition and improve bioproduction efficiency. Such advances will be instrumental in translating the present framework into efficient industrial processes.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Author contributions

CY, YL, HF and JW participated in the design of this study. CY, YL and ML were involved in fermentations. CY, YL and JW analyzed the raw data. CY, YL, HF and JW participated in the analysis of the results. CY, YL and JW wrote the manuscript. All authors have read and approved the final manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (22178133).

Data availability

All data generated or analyzed during this study are included in this published article and its additional files.

Declarations

Ethics approval and consent to participate

Not applicable.

Competing interests

The authors declare no competing interests.