Skip to main content

Efficient production of guanosine in Escherichia coli by combinatorial metabolic engineering



Guanosine is a purine nucleoside that is widely used as a raw material for food additives and pharmaceutical products. Microbial fermentation is the main production method of guanosine. However, the guanosine-producing strains possess multiple metabolic pathway interactions and complex regulatory mechanisms. The lack of strains with efficiently producing-guanosine greatly limited industrial application.


We attempted to efficiently produce guanosine in Escherichia coli using systematic metabolic engineering. First, we overexpressed the purine synthesis pathway from Bacillus subtilis and the prs gene, and deleted three genes involved in guanosine catabolism to increase guanosine accumulation. Subsequently, we attenuated purA expression and eliminated feedback and transcription dual inhibition. Then, we modified the metabolic flux of the glycolysis and Entner-Doudoroff (ED) pathways and performed redox cofactors rebalancing. Finally, transporter engineering and enhancing the guanosine synthesis pathway further increased the guanosine titre to 134.9 mg/L. After 72 h of the fed-batch fermentation in shake-flask, the guanosine titre achieved 289.8 mg/L.


Our results reveal that the guanosine synthesis pathway was successfully optimized by combinatorial metabolic engineering, which could be applicable to the efficient synthesis of other nucleoside products.


Guanosine and its nucleotide derivatives play essential physiological roles in nucleic acid synthesis, and possess antioxidant activity, neurotrophic and neuroprotective effects [1,2,3]. Additionally, guanosine has extensive applications as a crucial precursor for some medications used in the treatment of viral infections and tumors as well as food additives [4,5,6]. Guanosine production is mainly based on chemical synthesis, RNA enzymatic hydrolysis and microbial fermentation. Nevertheless, the first two approaches have the drawbacks of producing a large number of byproducts, complex separation and purification procedures, and high cost. Currently, microbial fermentation is the most commonly used approach for guanosine production. With the increasing demand for guanosine, a cost-effective method for the production of guanosine is needed. However, the lack of strains with high guanosine synthesis leads to low efficiency of guanosine production and high costs, hindering industrial application.

Guanosine is synthesized though the Embden–Meyerhof–Parnas (EMP, or glycolysis) pathway, ED pathway, pentose phosphate pathway (PPP), and purine synthesis pathway (Fig. 1). First, the guanosine biosynthetic pathway in vivo begins with the formation of IMP and GMP. Subsequently, GMP is further converted to guanosine by phosphatase or 5’-nucleotidase. The de novo synthesis of guanosine includes 13 steps from 5′-phosphoribosyl pyrophosphate (PRPP), and precursors include bicarbonate, glycine, aspartate, glutamine, ribose-5′-phosphate (R5P), and some cofactors. In the salvage pathway, phosphoribosyltransferases can catalyze nucleobases to generate nucleotides. In B. subtilis, a gene cluster, purEKB-purC(orf)QLF-purMNH(J)-purD, constitutes the purine operon (Bspur) that contains three overlapping coding units and a single gene [7, 8]. Additionally, purA, guaA, and guaB are scattered across the genome as single genes. The dual regulation of the Bspur operon is mediated by transcription initiation and termination [9]. The pur operon exists as single genes or small operons across the chromosome in E. coli, and its expression is affected by the PurR repressor and the addition of guanine or hypoxanthine [10].

Fig. 1
figure 1

An overview of engineering strategies to increase guanosine production in E. coli. Blue and green arrows indicate overexpression and attenuation of the target genes. The red X indicates deletion of the corresponding gene. Dashed arrows indicate repression by regulatory protein. Bs indicates Bacillus subtilis. Ba indicates Bacillus amyloliquefaciens. Abbreviations: pfkA, 6-phosphofructokinase I; BsglpX, fructose-1,6-bisphosphatase II; edd, phosphogluconate dehydratase; eda, KHG/KDPG aldolase; Baprs, ribose-phosphate diphosphokinase; BspurF, glutamine PRPP amidotransferase; BspurD, phosphoribosylglycinamide synthetase; BspurN, THFA-dependent phosphoribosylglycinamide transformylases; BspurQLS, phosphoribosylformyl-glycinamidine synthetases I, II, and III; BspurM, phosphoribosylaminoimidazole synthetase; BspurEK, phosphoribosylaminoimidazole carboxylases I and II; BspurC, phosphoribosylaminoimidazolesuccinocarboxamide synthetase; BsPurB, adenylosuccinate lyase; BspurH, phosphoribosylaminoimidazole carboxamide formyltransferase and IMP cyclohydrolase; purR, DNA-binding transcriptional repressor; purA, adenylosuccinate synthase; guaA, GMP synthase; guaB, IMP dehydrogenase; guaC, GMP reductase; deoD, purine nucleoside phosphorylase; ppnP, nucleoside phosphorylase; gsk, inosine/guanosine kinase; nupG, nucleoside: H+ symporter; nepI, purine ribonucleoside exporter

Native guanosine synthesis in bacteria is relatively complex and regulated by the EMP, ED and PPP, several branch and degradation pathways, feedback inhibition, etc. Guanosine anabolism requires a large quantity of purine precursors and considerable energy, and accumulating high levels of guanosine is difficult. Traditionally, chemical or physical mutagenesis methods have been used to generate strains with high guanosine yields. However, it is increasingly difficult to further increase the titre of guanosine based on existing guanosine-producing strains. Microbial cell factories constructed by system metabolic engineering provide an option for increasing the guanosine production. Moreover, such cell factories have been used to boost the production of uridine, inosine and guanosine [11,12,13]. Multiple engineering strategies including increasing the supply of precursor, disrupting the branch and degradation pathways, deleting the repressor PurR, overexpressing nucleoside efflux transporter PbuE, constructing an unnecessary protein-reduced chassis, and building genome-scale metabolic network models, have been applied to promote the synthesis of purine nucleosides [11, 14,15,16]. E. coli was applied to synthesize up to 7.5 g/L inosine by overexpression of the desensitized prs and purF genes and knockout of the purA, purR, deoD, and purF genes [17, 18]. Deletion of the gsk gene encoding inosine/guanosine kinase led to increased guanosine accumulation [19]. It’s reported that the guanosine production in engineered E. coli and B. subtilis was 120 mg/L and 115.2 mg/L in shake flask, respectively [13, 19]. Overall, previous studies focused on several genetic modifications of synthesis pathways of purine nucleoside. Employing a systems metabolic engineering strategy to construct engineered strains is necessary for the effective production of guanosine.

As a model microorganism, E. coli is a promising chassis strain for guanosine synthesis. Multiple gene editing tools and synthetic biology and metabolic engineering strategies are favourable for constructing cell factories [20]. CRISPR/Cas9-mediated homologous recombination is a precise gene editing method that can achieve stable targeted gene knockout or integration [21]. To enhance the guanosine titre, we adopted a systematic metabolic engineering strategy to modify the chassis strain E. coli MG1655. First, overexpression of the purine operon Bspur and the prs gene was carried out to enhance the flux into guanosine synthesis. Then, the guanosine degradation pathway was blocked to reduce guanosine consumption. Subsequently, branch pathway for adenosine synthesis, feedback inhibition from repressor PurR, EMP and ED pathways were modified to drive more carbon flux into guanosine synthesis. Next, redox cofactors rebalancing, transporter engineering and enhancing the guanosine synthesis pathway were used to promote guanosine accumulation. Finally, the best strain MQ39 produced 289.8 mg/L guanosine in fed-batch fermentation in shake flasks after 72 h. The constructed cell factory of guanosine production in this study can also be served as chassis strains to synthesize other valuable purine nucleosides like adenosine and inosine.

Materials and methods

Strains, plasmids, and cultivation conditions

E. coli DH5α was used for vector construction, and E. coli MG1655 ΔlacI was used as the starting chassis strain to perform genomic editing. Plasmids #1 and #2 for the CRISPR/Cas9 genome editing were kindly supplied by Prof. Huo and Dr. Huang of Beijing Institute of Technology [21]. Plasmid #1 consists of a p15A replication origin, a kanamycin resistance gene KanR, the sucrose-sensitive sacB gene, a CAS9 expression system induced by L-arabinose and an isopropyl-β-D-thiogalactopyranoside (IPTG) inducible λ-Red (Gam, Beta, and Exo) recombination system. Plasmid #2 includes a temperature-sensitive pSC101 replication origin, the promoter PBAD-N20-gRNA scaffold expressing cassette for Cas9 binding, a donor DNA-generation system serving as an editing template and the ampicillin resistance gene AmpR. The sacB gene of plasmid #1 and the pSC101 replication origin of plasmid #2 were used for plasmid curing. The strains and plasmids used in this study are listed in Table 1. E. coli DH5α and MG1655 ΔlacI were cultured in Luria–Bertani (LB) medium at 37 °C. Ampicillin (100 µg/mL) and kanamycin (50 µg/mL) were added to the cultures when needed. IPTG, L-arabinose, glucose, and sucrose were added at concentrations of 1 mM, 20 mM, 20 g/L and 20 g/L, respectively.

Table 1 Strains and plasmids used in this study

Construction of plasmids and recombinant strains

gRNA expression plasmids were constructed based on plasmid #2 and used for gene deletion and genomic integration. The purified two homologous arms (~ 500 bp) and the inserted expression cassette were fused by overlap extension PCR to obtain donor DNA. Subsequently, the two fragments from the plasmid #2 backbone and donor DNA were ligated with the ClonExpress II One Step Cloning Kit (Vazyme, Nanjing, China) to form the specific gRNA plasmid #2.

The recombinant strains were obtained according to a previous protocol [21]. Briefly, the specific gRNA plasmid #2 was introduced into the MG1655ΔlacI strain harbouring plasmid #1for gene knockout and chromosomal integration. Colonies grown on the plates were verified by colony PCR, and the correct colonies were confirmed by DNA sequencing. Both plasmids #1 and #2 could be cured by cultivation on LB solid plate containing 2% sucrose at 37 °C for 24 h when the obtained strain was not required to further genetic modifications. The primers are listed in Supplementary Table S1. The integration expression cassettes and a flowchart of gene editing for generating the engineered strains are shown in Supplementary Figs. S1 and S2.

Fed-batch fermentation

For guanosine fermentation, fresh single colonies of engineered strains were inoculated in 10 mL of liquid LBG medium overnight at 37 °C and 220 rpm. Then, the seed cultures of 3 mL were transferred into 27 mL of LBG (LB plus 2% glucose) medium in a 500-mL shake flask and cultured for 72 h at 37 °C and 220 rpm. Fed-batch fermentation was almost the same as the batch method. During the fed-batch process, the pH was kept at approximately 7.0 by adding ammonia with a microinjector, and phenol red was used as the pH indicator. Glucose solution (30%) was added for fed-batch culture when the glucose is depleted.

Substrate and product analysis

Optical density of the engineered strains was measured by the ultraviolet spectrophotometer at 600 nm. The glucose concentration was analysed by the 3,5-dinitrosalicylic acid (DNS) method. Cells were collected and centrifuged at 12,000 rpm for 10 min. The resultant supernatants were filtered through 0.22 μm syringe filters for the next analysis. The guanosine and hypoxanthine concentrations were quantified by HPLC (Agilent 1200) using a RD-C18 5 μm column (Zhongpu scientific, China) using the mobile phase methanol/water mixture (15:85 v/v) with a flow rate of 1 mL/min. The detection temperature and wavelength were set to 25 °C and 254 nm, respectively.

Statistical analysis

All experiments were performed with four independent cultures. Statistical significance was analysed by a two-tailed Student’s t-test. p value < 0.05 (*) was defined as statistical difference; p value < 0.01 (**) was defined as significant; p value < 0.001 (***) was defined as highly significant; ns was defined as no significant difference.


Increasing the metabolic flux and precursor PRPP supply

To increase the metabolic flux of guanosine synthesis, we sought to introduce the purine synthesis pathway of B. subtilis 168 into chromosome of E. coli. However, the complete purine synthesis pathway Bspur operon contains 11 genes (~ 12 kb) without purR (Fig. 2a and b). It is quite difficult for chromosomal integration expression for large DNA fragments by the CRISPR/Cas9-mediated homologous recombination. With increasing length of the integrated fragment, the recombination efficiency significantly decreases, which is not conducive to obtaining positive clones. Therefore, the Bspur operon was divided into four parts, each 1000–5000 bp. These parts were controlled by the Ptrc promoter and different synthetic terminators to generate expression cassettes (Fig. 2b). Four expression cassettes were further flanked by upstream and downstream homologous arms to form the donor DNA. Meanwhile, we deleted the lacI gene to eliminate regulation of the LacI repressor on the Ptrc promoter, and the resultant strain MG1655 ΔlacI was used as the starting chassis. After integration of the donor DNA into different loci located on the chromosome, the engineered E. coli strain MQ4 overexpressing the Bspur operon was obtained. The results showed that MG1655 ΔlacI strain did not produce inosine, while the MQ4 strain achieved the highest inosine production at 7.3 mg/L after 24 h of fermentation, indicating that Bspur operon integration is beneficial for increasing the metabolic flux of purine synthesis in E. coli (Fig. 2d). In addition, MG1655 ΔlacI and MQ4 strains did not produce guanosine (data not shown). MG1655 ΔlacI and MQ4 accumulated 8.1 mg/L and 10.9 mg/L of hypoxanthine after 24 h of fermentation, respectively, and the amount of hypoxanthine still showed an increasing trend. The final OD600 of MQ4 was very similar to that of MG1655 ΔlacI.

Fig. 2
figure 2

Overexpressing the purine synthesis operon from B. subtilis. (a) The native metabolic pathway of purine biosynthesis in E. coli; (b) Structural gene diagram of the purine operon in B. subtilis; (c) Integrated expression cassettes of the purine operon; (d) Production of hypoxanthine and inosine by overexpressing the purine synthesis operon. All data represent the mean ± s.d. (n = 4 biologically independent samples). Error bars were analysed by Student’s t test (two-sample, two-tailed; ** p value < 0.01, *** p value < 0.001)

PRPP is a crucial precursor for purine and pyrimidine nucleoside synthesis in vivo. PRPP is generated from R5P and ATP, which is catalysed by PRPP synthase (Figs. 1 and 3a). PRPP synthase is regulated by feedback inhibition via ADP, and mutated PRPP synthase can relieve nucleotide inhibition [17]. The prs gene from E. coli, the mutated EcprsD128A gene and the Baprs gene from Bacillus amyloliquefaciens were integrated into the aslA-glmZ locus of MQ4 (Fig. 3b), respectively. Guanosine and adenosine were not detected by HPLC in MQ5, MQ6, and MQ7. The production of inosine in MQ5, MQ6, and MQ7 was 7.8 mg/L, 8.8 mg/L, and 10.6 mg/L, respectively, which increased by 6.8%, 20.5%, and 45.2% compared to the MQ4 strain. In addition, MQ5, MQ6 and MQ7 strains accumulated 11.6 mg/L, 13.9 mg/L, and 16.0 mg/L of hypoxanthine, respectively (Fig. 3c). These results implied that although the flux of inosine synthesis was further enhanced by the overexpression of the prs gene, the guanosine concentration was possibly too low to be monitored by HPLC. In the de novo synthesis pathway of purine nucleosides, IMP requires a one-step reaction to generate inosine, while the conversion of IMP to guanosine and adenosine requires a three-step reaction. We speculate that IMP may be used to preferentially synthesize inosine, and the metabolic flux of the inosine synthesis pathway is higher than that of guanosine. In addition, guanosine is possibly degraded in E. coli.

Fig. 3
figure 3

Overexpressing the prs gene for guanosine production. (a) Schematic of the PRPP synthesis pathway in E. coli; (b) Integrated expression cassettes of the prs gene; (c) Production of hypoxanthine and inosine by overexpressing the prs gene. All data represent the mean ± s.d. (n = 4 biologically independent samples). Error bars were analysed by Student’s t test (two-sample, two-tailed; ** p value < 0.01, *** p value < 0.001)

Blocking the degradation pathway

Guanosine can be degraded by multiple cellular enzymes, including inosine/guanosine kinase encoded by gsk gene, two guanosine phosphorylases encoded by deoD and ppnP genes, and three nucleoside hydrolases encoded by rihA, rihB, and rihC genes [19, 22] (Fig. 4a). Notably, phosphorylase DeoD can also degrade inosine and adenosine into hypoxanthine and adenine, respectively. These enzymes were selected to be blocked in E. coli. The results showed that MQ9 (deoD deletion) produced 13.1 mg/L guanosine after 72 h of fermentation (Fig. 4b). After further knockout of ppnP and gsk, the titre of guanosine in MQ14 and MQ15 reached 16.2 and 23.2 mg/L, respectively. However, the production of guanosine in the MQ16 (rihA deletion), MQ17 (rihB deletion) and MQ18 (rihC deletion) strains was obviously decreased, indicating that the deletion of rihA, rihB and rihC is unfavourable for guanosine accumulation. Subsequently, MQ19 (triple deletion of deoD, ppnP and gsk) accumulated 41.5 mg/L guanosine after 72 h fermentation, an increase of 78.9% compared with MQ15 (Fig. 4b). These results suggested that disrupting the degradation pathway is beneficial for the accumulation of guanosine in E. coli.

Fig. 4
figure 4

Disruption of guanosine catabolism-related genes. (a) The related pathways of guanosine catabolism in E. coli. (b) Production of guanosine by deleting guanosine catabolism-related genes. All data represent the mean ± s.d. (n = 4 biologically independent samples). Error bars were analysed by Student’s t test (two-sample, two-tailed; *** p value < 0.001)

Down regulating purA expression and removing feedback and transcription dual inhibition

The conversion of IMP to adenosine was catalysed by adenylosuccinate synthetase (ADSS), which is a competitive pathway for guanosine production (Fig. 5a). Given that adenosine is important for cell growth, the adenosine synthesis pathway was attenuated to direct more carbon flux towards guanosine synthesis. The ssrA degradation tag (DAS + 4) was added to the C-terminus of ADSS encoded by purA gene to induce cytoplasmic degradation of ADSS [23]. Additionally, the native purA was replaced with BspurAP242N to lower the flux towards adenosine branch from IMP. The guanosine titres of MQ20 (purA disruption), MQ21 (BspurAP242N) and MQ23 (purA-ssrA) were 56.3, 57.7 and 50.0 mg/L, respectively (Fig. 5c). However, the final OD600 of the MQ20 and MQ21 strains was 1.9, which was significantly reduced by 70.3% compared with the MQ19 strain. The biomass of MQ23 strain is similar to that of MQ19 strain, and the guanosine production increased by 20.4%. Considering that OD600 is an important factor for strain engineering and fermentation, the MQ23 strain was selected for subsequent genetic modification. These results suggested that the purA gene is indispensable for cell growth and that its deletion can lead to severe growth deficiency.

The PurR repressor can inhibit the transcription of the pur operon and the prs gene (Fig. 5b). We sought to delete the purR gene to remove the repression regulation and enhance the intracellular concentration of PRPP. The results showed that MQ24 (purR deletion) accumulated 62.4 mg/L guanosine, an increase of 24.8% compared with MQ23 (Fig. 5c).

Fig. 5
figure 5

Adjusting the metabolic flux from adenosine synthesis and the feedback inhibition of PurR. The intrinsic pathway of adenosine synthesis (a) and the feedback inhibition mediated by PurR (b); (c) Production of guanosine in the engineered strains. All data represent the mean ± s.d. (n = 4 biologically independent samples). Error bars were analysed by Student’s t test (two-sample, two-tailed; ** p value < 0.01, *** p value < 0.001, ns represents no significant difference)

Redistributing the metabolic flux of the EMP and ED pathways and balancing redox cofactors

Glucose-6-phosphate (G6P) is an important intermediate of guanosine production as well as a component of the EMP and ED pathways. The biosynthetic pathway of purine competes for metabolic flux with the glycolysis and ED pathways (Fig. 6a). Nevertheless, the disruption of glycolysis and the ED pathway may lead to metabolic imbalance and sever growth defect. The carbon flux between the EMP and purine synthesis pathways should be efficiently distributed to achieve higher guanosine production. In glycolysis, the Pfk enzyme catalyzes the conversion of fructose-6-phosphate (F6P) to fructose-1,6-diphosphate (FBP). Pfk has two isoenzyme forms, and 6-phosphofructokinase I (Pfk-I, encoded by pfkA) possesses more than 90% of Pfk enzymatic activity. Fbpase, including Fbpase I and Fbpase II, can hydrolyse FBP to F6P (Fig. 6a). Compared with Fbpase I (encoded by fbp), Fbpase II, encoded by glpX, exhibits reduced sensitivity towards feedback inhibition of G6P [24, 25]. In the ED pathway, phosphogluconate dehydratase encoded by edd and 2-keto-3-deoxygluconate-6-phosphate aldolase encoded by eda convert 6-phosphogluconate (6-PG) into pyruvate and glyceraldehyde-3-phosphate.

Fig. 6
figure 6

Redistributing the metabolic flux of EMP and ED and redox cofactor rebalancing for the accumulation of guanosine. (a) Schematic of the EMP and ED pathways; (b) Schematic of redox cofactor rebalancing; (c) Increased guanosine production by downregulating the metabolic flux of EMP and ED and redox cofactor rebalancing. All data represent the mean ± s.d. (n = 4 biologically independent samples). Error bars were analysed by Student’s t test (two-sample, two-tailed; ns represents no significant difference)

To enhance flux into the PPP towards the production of guanosine, we deleted the pfkA gene and overexpressed glpX and BsglpX using the weak synthetic promoter PJ23116 to adjust the carbon flux of the glycolysis and ED pathways. The guanosine titres of MQ25 (pfkA deletion), MQ26 (pfkA deletion plus glpX overexpression) and MQ27 (pfkA deletion plus BsglpX overexpression) reached 75.6, 78.6 and 90.7 mg/L, respectively (Fig. 6c). Compared to the MQ24 strain, the production of guanosine was increased by 21.2%, 26.0% and 45.4%, respectively. The ED pathway was also interrupted by double knockout of the edd and eda genes. The resultant MQ28 strain accumulated 100.9 mg/L guanosine, an increase of 11.2% compared with MQ27 (Fig. 6c). Engineering the EMP and ED pathways displayed a synergistic effect for guanosine accumulation. These results proved that attenuating the flux of the EMP and ED pathways is a useful strategy for enhancing guanosine synthesis.

Although adjusting glycolysis and the ED pathway can improve guanosine production, it might result in cofactor imbalance. The inactivation of pfkA, edd, and eda and overexpression of glpX can redirect more carbon flux into the PPP, which can lead to excess NADPH and NADH deficiency [26]. To restore cofactor balance, soluble pyridine nucleotide transhydrogenase SthA (UdhA), encoded by sthA, was introduced into the genome to enhance the conversion of NADPH to NADH, and the membrane-bound transhydrogenase encoded by the pntAB gene was deleted to decrease the conversion of NADH to NADPH (Fig. 6b). The guanosine titres of MQ29 (pntAB deletion), MQ30 (sthA overexpression) and MQ31 (pntAB deletion and sthA overexpression) were 88.3, 82.3 and 96.3 mg/L, respectively (Fig. 6c). Redox cofactor rebalancing failed to further improve the production of guanosine, which suggested that the NADH and NADPH levels in MQ28 might be suitable for guanosine synthesis.

Transporter engineering for guanosine accumulation

Excessive accumulation of nucleosides leads to metabolic burden on the cell and causes product-mediated feedback inhibition. Previous studies illustrated that the NupG protein of E. coli can transport extracellular nucleosides into cells, and knockout of the nupG and nupC genes can increase the uridine and cytidine titres [11, 27, 28]. Meanwhile, overexpression of the nucleoside efflux transporters PbuE and NepI notably increased inosine secretion [29]. Transporters are essential for the hyperproduction of target products. To investigate the effects of these transporters on guanosine synthesis, we deleted the nupG gene in MQ28 to obtain strain MQ32. The inactivation of the nupG gene led to a slight increase in the guanosine titre (Fig. 7). Subsequently, overexpression of nepI and BspbuE in MQ32 generated the MQ33 and MQ34 strains, respectively. The guanosine titre of MQ33 was 123.6 mg/L, which was 16.4% higher than that obtained from MQ32. Blockage and overexpression of nucleoside transporters can prevent the absorption of exogenous guanosine and facilitate the extracellular outflow of guanosine [30, 31]. The results demonstrated that engineering transporters can enhance the performance of chassis strains by facilitating efflux and decreasing the accumulation of the desired intracellular products.

Fig. 7
figure 7

Engineering transporters for guanosine production. All data represent the mean ± s.d. (n = 4 biologically independent samples). Error bars were analysed by Student’s t test (two-sample, two-tailed; *p value < 0.05)

Strengthening the guanosine synthesis pathway

The reduction reaction from GMP to IMP is catalyzed by GMP reductase encoded by guaC gene, and overexpression of guaA and guaB genes is beneficial to guanosine synthesis. To favor guanosine production, we integrated the guaAB gene into ykgH-betA locus and guaC locus to obtain strains MQ38 and MQ39. These two strains accumulated 127.0 mg/L and 134.9 mg/L guanosine, respectively (Fig. 8). These results illustrated that enhancing the synthesis pathway of guanosine is conducive to guanosine accumulation.

Fig. 8
figure 8

The effect of engineering the guanosine synthesis pathway. (a) Schematic of guanosine synthesis pathway; (b) Effects of strengthening the guanosine synthesis pathway on guanosine accumulation. All data represent the mean ± s.d. (n = 4 biologically independent samples). Error bars were analysed by Student’s t test (two-sample, two-tailed; *p value < 0.05)

Guanosine production in shake flasks

The strain MQ39 was grown in shake flasks to evaluate the performance of guanosine production under fed-batch fermentation. The cells grew rapidly from the initiation of fermentation, and the OD600 reached 11.7 at 72 h (Fig. 9). The production of guanosine gradually increased in the whole fermentation process, representing a cell growth-independent production trend. The guanosine titre reached 289.8 mg/L with a yield of 9.89 mg/g glucose after fermentation for 72 h.

Fig. 9
figure 9

Assessment of guanosine production using the strain MQ39. Fed-batch fermentation was performed in shake flasks. Four biological replicates were performed, and the error bars indicate the standard deviation


Guanosine and other purine nucleosides are crucial for various cellular physiological processes and have broad applications in the fields of antitumor/antiviral medication as well as food additives. Nevertheless, microbial production of guanosine is still a large challenge because of multiple metabolic pathway interactions and complex regulatory mechanisms. Here, we successfully constructed a microbial cell factory for efficient guanosine production by combinational metabolic engineering strategies. The engineered strains can produce 289.8 mg/L guanosine after 72 h of fermentation.

Low metabolic flux and insufficient precursor supply are the main limiting factors for synthesizing target products in microbial hosts [32]. Overexpression of homologous or heterologous metabolic pathways is an effective method to enhance the metabolic flux for target product production [33]. Previous methods have mainly focused on plasmid expression. The advantage of plasmid expression is that more copies of the target genes possibly lead to high expression levels. However, it is necessary to add amino acids or antibiotics to prevent plasmid loss, and the continuous addition of antibiotics may increase the cost of fermentation. In addition, a large recombinant plasmid may bring about a metabolic burden on the strain and result in growth delay, which might be unsuitable for constructing pyrimidine-producing strains [11]. Plasmid loss is also easily observed in fed-bath cultivation, which results in ineffective consumption of carbon sources and reduced product titre [26]. Integration of the desired metabolic pathway into the genome is a stable overexpression strategy for multiple genes [34]. Chromosomal integration of multiple genes possesses several advantages, such as achieving stable gene expression, avoiding vector capacity limitations and plasmid incompatibility. For example, the pyrimidine biosynthetic operon from B. subtilis was introduced into the yghX locus of E. coli with the help of two exogenous protospacer and protospacer-adjacent motifs and two corresponding gRNA plasmids [11]. We divided the Bspur operon into four parts and integrated them into different neutral sites using the CRISPR/Cas9-mediated homologous recombination method. The introduced Bspur operon was used to enhance the enzymatic activities of purine nucleotide synthesis pathway. However, we could not detect the accumulation of guanosine in MQ4 (Fig. 2). In addition, the reaction of R5P to PRPP is considered a rate-limiting step for purine and pyrimidine nucleoside synthesis. prs overexpression also did not lead to detectable guanosine accumulation. Unexpectedly, the titre of inosine was significantly increased by overexpression of the pur operon and prs, indicating that high purine metabolic flux is directed towards inosine synthesis in cells (Fig. 3). Subsequently, the accumulation of guanosine was observed by disrupting degradation-related genes (∆deoD, ∆ppnP and ∆gsk), and MQ19 obtained 41.5 mg/L guanosine (Fig. 4). These results indicated the necessity of blocking product catabolism pathways for the target products, which was in accord with previous reports [11, 35]. Notably, because guanosine synthesis involves multiple precursors, enhancing the supply of other precursors, such as glutamine and aspartate, should be taken into consideration in the future.

Metabolic flux redistribution is an effective strategy to enhance chemical synthesis. In this study, downregulating the flux of adenosine synthesis and the EMP and ED pathways drove more metabolic flux into guanosine biosynthesis, which illustrated a synergistic effect on guanosine production (Figs. 5 and 6). Optimization of central carbon metabolism can enhance the availability of precursors, thereby improving the titre, productivity and yield of the biosynthetic target compounds and enhancing the performance of engineered strains. This strategy has been used to enhance the titre of other chemicals, including nucleosides, terpenoids and fatty acid derivatives, amino acids, organic acids and natural products [15, 36,37,38,39,40]. For example, the titre of inosine was significantly improved by overexpressing the key enzyme Zwf and blocking two essential backflow nodes including the purine synthesis pathway towards the PPP, and from the PPP to the glycolysis [15]. A previous study illustrated that deleting the pfkA, edd and eda genes can result in an approximately 11-fold increase in riboflavin titre in shake flasks [41]. The fbp gene overexpression also led to a shift in metabolic flux from the glycolysis to the PPP and an increase in riboflavin titre [26]. Alternatively, we can fine-tune metabolic pathway through promoter replacement with different strengths, inducible promoters and quorum sensing system to dynamically regulate key metabolic nodes to increase guanosine production. Further engineering should be focused on balancing biomass and products by controlling the carbon flux of glucose into the glycolysis.


In summary, guanosine production in E. coli engineered strains was progressively increased by combinatorial metabolic engineering strategies, which included overexpression of the Bspur operon and prs gene, blockage of the guanosine degradation pathway, downregulation of purA expression, elimination of feedback and transcription dual inhibition, redirection of the metabolic flux of the EMP and ED pathways, redox cofactor rebalancing and engineering of transporters and strengthening the guanosine synthesis. Ultimately, the final optimized strain MQ39 produced 289.8 mg/L guanosine. Combinatorial metabolic engineering strategies would be beneficial to further engineer E. coli to act as an excellent chassis strain for industrial guanosine production in the future.

Data availability

No datasets were generated or analysed during the current study.







pentose phosphate pathway


Clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated protein 9 (Cas9)


5′-phosphoribosyl pyrophosphate




polymerase chain reaction


High-performance liquid chromatography




3,5-dinitrosalicylic acid










adenylosuccinate synthetase


  1. Gudkov SV, Shtarkman IN, Smirnova VS, Chernikov AV, Bruskov VI. Guanosine and inosine display antioxidant activity, protect DNA in vitro from oxidative damage induced by reactive oxygen species, and serve as radioprotectors in mice. Radiat Res. 2006;165:538–45.

    Article  CAS  PubMed  Google Scholar 

  2. Lanznaster D, Dal-Cim T, Piermartiri TC, Tasca CI. Guanosine: a neuromodulator with therapeutic potential in Brain disorders. Aging Dis. 2016;7:657–79.

    Article  PubMed  PubMed Central  Google Scholar 

  3. Kappock TJ, Ealick SE, Stubbe J. Modular evolution of the purine biosynthetic pathway. Curr Opin Chem Biol. 2000;4:567–72.

    Article  CAS  PubMed  Google Scholar 

  4. Elfiky AA. Novel guanosine derivatives against Zika virus polymerase in silico. J Med Virol. 2020;92:11–6.

    Article  CAS  PubMed  Google Scholar 

  5. De Clercq E, Andrei G, Snoeck R, De Bolle L, Naesens L, Degrève B, et al. Acyclic/carbocyclic guanosine analogues as anti-herpesvirus agents. Nucleos Nucleot Nucl. 2001;20:271–85.

    Article  Google Scholar 

  6. Chen Z, Ku TC, Seley-Radtke KL. Thiophene-expanded guanosine analogues of gemcitabine. Bioorg Med Chem Lett. 2015;25:4274–6.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  7. Ebbole DJ, Zalkin H. Cloning and characterization of a 12-gene cluster from Bacillus subtilis encoding nine enzymes for de novo purine nucleotide synthesis. J Biol Chem. 1987;262:8274–87.

    Article  CAS  PubMed  Google Scholar 

  8. Ebbole DJ, Zalkin H. Bacillus subtilis pur operon expression and regulation. J Bacteriol. 1989;171:2136–41.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  9. Johansen LE, Nygaard P, Lassen C, Agersø Y, Saxild HH. Definition of a second Bacillus subtilis pur regulon comprising the pur and xpt-pbux operons plus pbuG, nupG (yxjA), and pbuE (ydhL). J Bacteriol. 2003;185:5200–9.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  10. Rolfes RJ, Zalkin H. Escherichia coli gene purR encoding a repressor protein for purine nucleotide synthesis. Cloning, nucleotide sequence, and interaction with the purF operator. J Biol Chem. 1988;263:19653–61.

    Article  CAS  PubMed  Google Scholar 

  11. Wu H, Li Y, Ma Q, Li Q, Jia Z, Yang B, et al. Metabolic engineering of Escherichia coli for high-yield uridine production. Metab Eng. 2018;49:248–56.

    Article  CAS  PubMed  Google Scholar 

  12. Liao Y, Ye Y, Wang B, Pan L. Optimization of the purine operon and energy generation in Bacillus amyloliquefaciens for guanosine production. Biotechnol Lett. 2017;39:1675–82.

    Article  CAS  PubMed  Google Scholar 

  13. Li Y, Zhu X, Zhang X, Fu J, Wang Z, Chen T, et al. Characterization of genome-reduced Bacillus subtilis strains and their application for the production of guanosine and thymidine. Microb Cell Fact. 2016;15:94.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  14. Asahara T, Mori Y, Zakataeva NP, Livshits VA, Yoshida K, Matsuno K. Accumulation of gene-targeted Bacillus subtilis mutations that enhance fermentative inosine production. Appl Microbiol Biotechnol. 2010;87:2195–207.

    Article  CAS  PubMed  Google Scholar 

  15. Deng A, Qiu Q, Sun Q, Chen Z, Wang J, Zhang Y, et al. In silico-guided metabolic engineering of Bacillus subtilis for efficient biosynthesis of purine nucleosides by blocking the key backflow nodes. Biotechnol Biofuels Bioprod. 2022;15:82.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  16. Li P, Jing R, Zhou M, Jia, Pei, Li Z, Liu G, Wang Z, et al. Whole-cell biosynthesis of cytarabine by an unnecessary protein-reduced Escherichia coli that coexpresses purine and uracil phosphorylase. Biotechnol Bioengin. 2022;119:1768–80.

    Article  Google Scholar 

  17. Shimaoka M, Takenaka Y, Kurahashi O, Kawasaki H, Matsui H. Effect of amplification of desensitized purF and prs on inosine accumulation in Escherichia coli. J Biosci Bioeng. 2007;103:255–61.

  18. Liu M, Fu Y, Gao W, Xian M, Zhao G. Highly efficient biosynthesis of hypoxanthine in Escherichia coli and transcriptome-based analysis of the purine metabolism. ACS Synth Biol. 2020;9:525–35.

    Article  CAS  PubMed  Google Scholar 

  19. Matsui H, Shimaoka M, Takenaka Y, Kawasaki H, Kurahashi O. Gsk disruption leads to guanosine accumulation in Escherichia coli. Biosci Biotechnol Biochem. 2001;65:1230–5.

    Article  CAS  PubMed  Google Scholar 

  20. Yang D, Prabowo CPS, Eun H, Park SY, Cho IJ, Jiao S, et al. Escherichia coli as a platform microbial host for systems metabolic engineering. Essays Biochem. 2021;65:225–46.

    Article  CAS  PubMed  Google Scholar 

  21. Huang C, Guo L, Wang J, Wang N, Huo YX. Efficient long fragment editing technique enables large-scale and scarless bacterial genome engineering. Appl Microbiol Biotechnol. 2020;104:7943–56.

    Article  CAS  PubMed  Google Scholar 

  22. Petersen C, Møller LB. The RihA, RihB, and RihC ribonucleoside hydrolases of Escherichia coli: substrate specificity, gene expression, and regulation. J Biol Chem. 2001;276:884–94.

    Article  CAS  PubMed  Google Scholar 

  23. McGinness KEBT, Sauer RT. Engineering controllable protein degradation. Mol Cell. 2006;22:701–7.

    Article  CAS  PubMed  Google Scholar 

  24. Kim YM, Cho HS, Jung GY, Park JM. Engineering the pentose phosphate pathway to improve hydrogen yield in recombinant Escherichia coli. Biotechnol Bioeng. 2011;108:2941–6.

    Article  CAS  PubMed  Google Scholar 

  25. Sun X, Peng Z, Li C, Zheng Y, Cheng Y, Zong J, et al. Combinatorial metabolic engineering and tolerance evolving of Escherichia coli for high production of 2’-fucosyllactose. Bioresour Technol. 2023;372:128667.

    Article  CAS  PubMed  Google Scholar 

  26. Liu S, Hu W, Wang Z, Chen T. Rational engineering of Escherichia coli for high-level production of riboflavin. J Agric Food Chem. 2021;69:12241–9.

  27. Patching SG, Baldwin SA, Baldwin AD, Young JD, Gallagher MP, Henderson PJ, et al. The nucleoside transport proteins, NupC and NupG, from Escherichia coli: specific structural motifs necessary for the binding of ligands. Org Biomol Chem. 2005;3:462–70.

    Article  CAS  PubMed  Google Scholar 

  28. Ma R, Fang H, Liu H, Pan L, Wang H, Zhang H. Overexpression of uracil permease and nucleoside transporter from Bacillus amyloliquefaciens improves cytidine production in Escherichia coli. Biotechnol Lett. 2021;43:1211–9.

    Article  CAS  PubMed  Google Scholar 

  29. Sheremet AS, Gronskiy SV, Akhmadyshin RA, Novikova AE, Livshits VA, Shakulov RS, et al. Enhancement of extracellular purine nucleoside accumulation by Bacillus strains through genetic modifications of genes involved in nucleoside export. J Ind Microbiol Biot. 2011;38:65–70.

    Article  CAS  Google Scholar 

  30. Liu Y, Pan X, Zhang H, Zhao Z, Teng Z, Rao Z. Combinatorial protein engineering and transporter engineering for efficient synthesis of L-Carnosine in Escherichia coli. Bioresour Technol. 2023;387:129628.

    Article  CAS  PubMed  Google Scholar 

  31. Chen Y, Wang J, Wang M, Han A, Zhao X, Wang W, et al. Engineering the metabolism and morphology of the filamentous fungus Trichoderma reesei for efficient L-malic acid production. Bioresour Technol. 2023;387:129629.

    Article  CAS  PubMed  Google Scholar 

  32. Jung J, Lim JH, Kim SY, Im D-K, Seok JY, Lee S-JV, et al. Precise precursor rebalancing for isoprenoids production by fine control of gapA expression in Escherichia coli. Metab Eng. 2016;38:401–8.

    Article  CAS  PubMed  Google Scholar 

  33. Kong X, Wu Y, Yu W, Liu Y, Li J, Du G, et al. Efficient synthesis of limonene in Saccharomyces cerevisiae using combinatorial metabolic engineering strategies. J Agric Food Chem. 2023;71:7752–64.

    Article  CAS  PubMed  Google Scholar 

  34. Goormans AR, Snoeck N, Decadt H, Vermeulen K, Peters G, Coussement P, et al. Comprehensive study on Escherichia coli genomic expression: does position really matter? Metab Eng. 2020;62:10–9.

    Article  CAS  PubMed  Google Scholar 

  35. Yang K, Li Z. Multistep construction of metabolically engineered Escherichia coli for enhanced cytidine biosynthesis. Biochem Eng J. 2020;154:107433.

    Article  Google Scholar 

  36. Li B, Zhang B, Wang P, Cai X, Chen YY, Yang YF, et al. Rerouting fluxes of the central carbon metabolism and relieving mechanism-based inactivation of l-Aspartate-α-decarboxylase for fermentative production of β-alanine in Escherichia coli. ACS Synth Biol. 2022;11:1908–18.

    Article  CAS  PubMed  Google Scholar 

  37. Wu J, Zhang X, Zhou J, Dong M. Efficient biosynthesis of (2S)-pinocembrin from d-glucose by integrating engineering central metabolic pathways with a pH-shift control strategy. Bioresour Technol. 2016;218:999–1007.

    Article  CAS  PubMed  Google Scholar 

  38. Wu J, Wang Z, Duan X, Zhou P, Liu P, Pang Z, et al. Construction of artificial micro-aerobic metabolism for energy- and carbon-efficient synthesis of medium chain fatty acids in Escherichia coli. Metab Eng. 2019;53:1–13.

    Article  CAS  PubMed  Google Scholar 

  39. Wu Y, Yan P, Li Y, Liu X, Wang Z, Chen T, et al. Enhancing β-carotene production in Escherichia coli by perturbing central carbon metabolism and improving the NADPH supply. Front Bioeng Biotechnol. 2020;8:585.

    Article  PubMed  PubMed Central  Google Scholar 

  40. Liu H, Song R, Liang Y, Zhang T, Deng L, Wang F, et al. Genetic manipulation of Escherichia coli central carbon metabolism for efficient production of fumaric acid. Bioresour Technol. 2018;270:96–102.

    Article  CAS  PubMed  Google Scholar 

  41. Liu S, Kang P, Cui Z, Wang Z, Chen T. Increased riboflavin production by knockout of 6-phosphofructokinase I and blocking the Entner-Doudoroff pathway in Escherichia coli. Biotechnol Lett. 2016;38:1307–14.

    Article  CAS  PubMed  Google Scholar 

Download references


This research was funded by the National Natural Science Foundation of China (31901071).

Author information

Authors and Affiliations



Kun Zhang: Conceptualization, Investigation, Writing – original draft, Funding acquisition. Mengxing Qin: Investigation, Writing – original draft. Yu Hou: Investigation. Wenwen Zhang: Investigation. Zhenyu Wang: Resources, Supervision, Writing – review & editing. Hailei Wang: Conceptualization, Resources, Supervision, Writing – review & editing.

Corresponding author

Correspondence to Hailei Wang.

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Zhang, K., Qin, M., Hou, Y. et al. Efficient production of guanosine in Escherichia coli by combinatorial metabolic engineering. Microb Cell Fact 23, 182 (2024).

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: