Introduction

Guanosine is used to synthesize flavoring additives used in the food industry and nucleoside-related drugs produced by the pharmaceutical industry (Wang et al. 2014). Bacillus subtilis and B. amyloliquefaciens have low purine nucleoside phosphorylase activities that favor guanosine accumulation, and they are important species for the rational design of high-yield industrial strains (Kröger et al. 2013). Various molecular engineering strategies have been used to improve their purine productivity, including the inactivation of the branch pathway genes purA, guaC, punA, and deoD (Asahara et al. 2010); the overexpression of prs (PRPP synthetase) and purF (PRPP amidotransferase) to relieve product-feedback regulation (Zakataeva et al. 2012); and the regulation optimization of the purine promoter to increase the transcription of the purine operon (Lobanov et al. 2011).

The quinol oxidase branch contains of the B. subtilis respiratory chain contains three oxidases: cytochrome aa 3 oxidase (qox), cytochrome bd oxidase (cyd), and a putative Yth oxidase (yth). Under aerobic conditions, qox is the primary terminal oxidase used to support growth, cyd plays a minor role during rapid growth, and the deletion of cyd increases the efficiency of energy generation (Li et al. 2006).

B. amyloliquefaciens XH7 is a guanosine-overproducing strain used for industrial production, and its complete genome sequence has been completed (Yang et al. 2011). In this work, to improve guanosine production in B. amyloliquefaciens XH7, the 5′-UTR of the purine operon containing the guanine-sensing riboswitch was deleted, and the native purine operon promoter (Pw) was individually replaced by strong promoters. In addition, the efficiency of energy generation was optimized. The combined modulation of product synthesis metabolism and energy metabolism provides a new strategy to improve strain performance and guanosine production.

Materials and methods

Bacterial strains

The bacterial strains used in this work are listed in Supplementary Table 1.

Escherichia coli DH5α and E. coli TG1 were used for plasmid construction. E. coli JM110 (dam−, dcm−) was used to prepare unmethylated plasmids used for transformation. B. amyloliquefaciens XH7 is an industrial guanosine-producing strain and was screened by traditional mutagenesis.

Media and growth conditions

Recombinant plasmids were grown aerobically at 37 °C with shaking at 200 rpm in lysogeny broth (LB) supplemented with appropriate antibiotics as necessary, including ampicillin (50 μg/ml), kanamycin (25 μg/ml), erythromycin (200 μg/ml for E. coli or 5 μg/ml for B. amyloliquefaciens), as necessary.

The fermentative production of guanosine was performed as previously described (Wang et al. 2014). B. amyloliquefaciens XH7 and its mutants were cultured in 20 ml seed medium until the OD600 reached 7–8. Approx. 10% (v/v) inoculum was added aseptically to a 500 ml shake-flask containing 30 ml fermentation medium. The seed medium was at pH 7 and was composed of 25 g glucose/l, 10 g yeast extract/l, 10 g peptone/l, 5 g NaCl/l, and 25 g adenine/l. The fermentation medium was pH 6.8 and contained 100 g glucose/l, 10 g dry powdered yeast/l, 5 g yeast extract/l, 10 g (NH4)2SO4/l, 2 g KH2PO4/l, 3 g MgSO4·7H2O/l, 15 g monosodium glutamate/l, 20 g CaCO3/l and 5 mg MnSO4/l. Glucose and CaCO3 were sterilized separately.

Plasmid construction

Plasmids used in this study are listed in Supplementary Table 1. Primers used in their construction are listed in Supplementary Table 2.

The expression plasmid pBEPw-bgaB was constructed as follows (Supplementary Fig. 1). Using B. amyloliquefaciens XH7 genomic DNA as template, the Pw promoter fragment was amplified using the primers pair pur-w-F/R. The amplified fragment was then digested with EcoRI and speI, and inserted into pBE-rbs to obtain the expression plasmid pBEPw-bgaB.

The integrative plasmids pKS2-T1 and pKS2-T2 were constructed to obtain 5′-UTR of deletion mutants of the purine operon (Supplementary Fig. 2). The primer pairs T1-p1/T1-p2 and T1-p3/T1-p4 were used to amplify homologous fragments of the 5′-UTR of the purine operon from the B. amyloliquefaciens XH7 genome. The homologous fragments were digested with KpnI and SpeI and were inserted into the corresponding sites of pKS2, thus yielding pKS2-T1. The pKS2-T2 was constructed by following the same procedure.

To replace the purine promoter with strong promoters (P41, P r2 and P43), the integrative plasmids pKS2-P41, pKS2-P r2 , and pKS2-P43 were constructed as follows (Supplementary Fig. 3). The promoter P41 fragment, and the upstream and downstream homologous fragments of the purine promoter were amplified using the primer pairs P41-p3/P4, P41-p1/P2, and P41-p5/P6, respectively. The fragments were mixed and subjected to a second round of PCR amplification using the primer pair P41-p1/P6 to obtain the p41 promoter replacement cassette. Subsequently, the cassette was phosphorylated and inserted into the pKS2 plasmid, yielding the plasmid pKS2-P41. The plasmids pKS2-P r2 and pKS2-P43 were constructed by following the same procedure.

For the cyd gene in-frame deletion, the upstream and downstream fragments of the cyd gene were amplified using the primer pairs cyd-p1/P2 and cyd-p3/P4. The upstream and downstream fragments were fused by PCR amplification using the primer pair cyd-p1/P4. The fused fragment was inserted into the corresponding sites of pKS2, yielding the plasmid pKS2-cyd with a truncated cyd gene (Supplementary Fig. 4).

Strain construction

Plasmids were transformed into B. amyloliquefaciens by electroporation as previously described (Liao et al. 2015). The XH7purET1 and XH7purET2 recombinants were identified by PCR using the primer pairs T1-p1/T1-p4 and T2-p1/T2-p4, respectively. The Mutants XH7purE::P41, XH7purE::P r2 , and XH7purE::P43 were identified by PCR using primers P41-t-F/R, P r2 -t-F/R, and P43-t-F/R. The cyd deletion mutants were confirmed by PCR using primer pair cyd-p1/cyd-p4.

Biomass analysis

Cell dry weights were measured using 1 ml cell suspensions that were harvested by centrifugation at 10,000×g for 1 min at room temperature. Afterwards, the cell pellets were washed with phosphate buffer (pH 7) and dried at 110 °C until no change in sample weight was observed. Measurements were performed in triplicate.

Assay of guanosine production

The guanosine concentration in samples was measured by HPLC using a Phenomenex Gemini 5 μm C18 110 Å (250 × 4.6 mm) column with detection at 254 nm. Approx. 10% (v/v) methanol was used as the mobile phase at 1 ml/min. 10 μl of sample was injected. Data are reported as the averages of three independent experiments.

Quantitative real-time PCR

RNA isolation and RT-PCR were performed as described previously (Liao et al. 2015). The RT-PCR results were analyzed using the comparative CT (2 −△△CT) method (Livak and Schmittgen 2001). The gapA gene was used as the internal reference gene.

β-Galactosidase (β-Gal) activity assay

β-Gal activity was measured as previously described (Liao et al. 2015). Error bars indicate the standard errors of the mean values.

Analysis of intracellular metabolites

The detection of intracellular purine intermediates was performed as previously described (Peifer et al. 2012). Samples were collected during the fermentation stages of growth and mixed with an equal volume of 0.5 M perchloric acid at 55 °C. The extracted samples were diluted to the appropriate concentrations with perchloric acid (0.5 M) and measured by HPLC using the same column as mentioned above but with the mobile phase of solvent A (methanol) and solvent B (0.5% KH2PO4/H3PO4 buffer, pH 3.6). The following gradient was used: 100% B for 0–2 min, 20% A and 80% B for 2–10 min, and 100% B for 10–15 min. Injections were 10 μl.

Organic acid analysis was performed as previously described (Asakura et al. 2007). Supernatants were diluted to appropriate concentrations with mobile phase buffer (20 mM KH2PO4/H3PO4, pH 2.9) and measured by HPLC as above with detection at 210 nm.

Results

Genetic modification of the purine operon to improve guanosine production

To improve guanosine production, the purine operon in B. amyloliquefaciens XH7 was genetically modified. The 5′-UTR of the purine operon contains two functional domains: namely, the G-Box-conserved domain (three-stem P1 through P3) and the expression platform (tandem stem-loop structure, for antitermination) (Supplementary Fig. 5A). We deleted either both the terminator stem and the expression platform (XH7PurET1: +38 to +202) or a portion of the expression platform (XH7PurET2: +100 to +186) to disrupt the function of the riboswitch. Next, we used high-strength promoters (P41, P r2 , and P43) (Supplementary Fig. 5B) to replace the native Pw promoter and terminate the PurR-mediated transcriptional repression and guanine-mediated transcriptional attenuation, thus producing the mutant strains XH7purE::P41, XH7purE::P r2 , and XH7purE::P43.

To investigate the effects of these genetic modifications, the guanosine production of these mutant strains was compared (Table 1). Of these strains, the XH7purE::P41 strain gave the highest guanosine yield (16.25 g/l) relative to B. amyloliquefaciens XH7 (13.26 g/l). The guanosine production of XH7purET2 was improved by 18%. Conversely, the guanosine production of the mutant strains XH7purE::P r2 and XH7purE::P43 decreased by 66 and 87%, respectively. In addition, the cell mass of the XH7purE::P r2 and XH7purE::P43 strains also dramatically decreased (Table 1).

Table 1 Guanosine yield and cell mass of purine operon mutants
Full size table

Relative transcription level and intracellular metabolite target analysis of the purine pathway

The transcription of key genes (purE, purF, and purD) in the purine operon was measured in the mutants during fermentation (Fig. 1). The relative transcription of purE, purF, and purD in the mutants XH7purET1, XH7purET2, and XH7purE::P41 was significantly upregulated relative to B. amyloliquefaciens XH7. However, the relative transcription of purE, purF, and purD in the mutants XH7purE::P r2 and XH7purE::P43 was remarkably decreased. These results revealed that the high transcription of the purine operon consequently enhanced guanosine production (Fig. 1; Table 1).

Fig. 1
figure 1

Relative transcription level of purine operon genes (purE, purF, and purD) in purine operon mutants during fermentation. W, XH7; T2, XH7purET2; T1, XH7purET1; P41, XH7purE::P41; P r2 , XH7purE::P r2 ; P43, XH7purE::P43. Data are averages of three independent experiments, and the error bars represent the standard errors of the means

Full size image

Purine pathway intermediates, IMP, guanosine monophosphate (GMP), hypoxanthine (HX) and inosine, were measured in these mutants to investigate the deregulation of the purine operon for guanosine biosynthesis. The concentrations of IMP and inosine in the mutants XH7purET2 and XH7purE::P41 increased relative to B. amyloliquefaciens XH7. Nevertheless, the GMP pool size in the mutants XH7purET2 and XH7purE::P41 significantly decreased (Fig. 2). This result suggests that enhancing the generation of IMP precursors and reducing GMP by-products could contribute to improving the guanosine biosynthetic flux. During the later stage of fermentation (after 48 h), the IMP concentration in the mutants XH7purET2 and XH7purE::P41 significantly increased compared with B. amyloliquefaciens XH7, which favored guanosine production. The results are consistent with the trends of guanosine production after 48 h (Fig. 3). However, the concentration of the main intermediates IMP and GMP in mutants XH7purE::P r2 and XH7purE::P43 significantly decreased compared with the wild-type strain B. amyloliquefaciens XH7 during the early fermentation stage (24 h) and returned to normal after 36 h. The results demonstrate that the replaced promoter was repressed in the early stage of fermentation and was subsequently activated.

Fig. 2
figure 2

Concentration of purine intermediates in purine operon mutants during fermentation. IMP, inosine monophosphate; GMP, guanosine monophosphate; and HX, hypoxanthine. W, XH7; T2, XH7purET2; T1, XH7purET1; P41, XH7purE::P41; P r2 , XH7purE::P r2 ; P43, XH7purE::P43. Data are averages of three independent experiments, and the error bars represented standard deviation

Full size image
Fig. 3
figure 3

Time cours of guanosine production and cell mass of purine operon mutants in shake flask fermentation. Data are averages of three independent experiments, and the error bars represented standard deviation

Full size image

Combined modification of cytochrome bd to improve guanosine production

To improve guanosine production by modifying energy generation in the respiratory chain, a series of cyd deletion mutants (XH7△cyd, XH7purET2△cyd, and XH7purE::P41△cyd) were constructed. As shown in Table 1, the cyd mutants had higher guanosine yields than B. amyloliquefaciens XH7. The XH7purE::P41△cyd stain produced the highest level of guanosine (19 g/l), 41% higher than B. amyloliquefaciens XH7. The cell mass of cyd deletion mutants showed a moderate increase compared with B. amyloliquefaciens XH7 (Table 1; Fig. 3), indicating that the deletion of cyd did not exert an influence on cell growth.

The concentration of organic acids (pyruvate, oxoglutarate, acetate, and acetoin) produced during carbon metabolism (glycolysis, TCA cycle) were also measured during fermentation (Fig. 4). The concentration of pyruvate (especially at 24 h) and acetoin in the cyd deletion mutants were both reduced. The accumulation of oxoglutarate was slightly decreased in these mutants compared with B. amyloliquefaciens XH7. The concentration of acetate in the cyd mutants was the same as in B. amyloliquefaciens XH7 and was significantly decreased at 24 h. The concentration of pyruvate also decreased, indicating that TCA cycle flux was repressed in the cyd deletion mutants.

Fig. 4
figure 4

Time course of organic acids in TCA cycle of cyd mutants. Wild, XH7; P41-cyd, XH7purE::P41△cyd; T2-cyd, XH7purET2△cyd; cyd, XH7△cyd. Data are averages of three independent experiments, and the error bars represented standard deviation

Full size image

Discussion

The regulation of the purine operon at the transcriptional level is crucial for guanosine production. Our results demonstrate that deletion of the 5′-UTR regions of the purine operon alters the structure and regulation of the guanine-sensing riboswitch. Additionally, the relative transcription levels of purine operon genes were significantly increased, which could contribute to guanosine production. Similar results were reported previously (Shi et al. 2014). The results indicate that strong promoters could be inserted into the upstream region of the purine operon to regulate the double-negative control involving the repressor protein PurR and transcriptional attenuation in B. amyloliquefaciens.

The low-coupling respiratory chain branch (cytochrome bd oxidase) was disrupted to increase the efficiency of respiratory energy generation. The cyd deletion mutants showed a slightly higher biomass than the wild-type strain. The concentration of acetate decreased during rapid growth (24 h) and this was beneficial for cell growth (Eiteman and Altman 2006). In the present study, the different degrees to which the levels of by-products (pyruvate, oxoglutarate, and acetoin) were reduced in the TCA cycle of the cyd deletion mutants indicated that TCA cycle flux decreased, which is consistent with a previous report (Zamboni and Sauer 2003). Therefore, the absence of the less efficient cytochrome bd oxidase may result in efficient energy generation and a low maintenance energy, which were beneficial for guanosine synthesis.

Conclusion

Guanosine production in B. amyloliquefaciens XH7 was improved by genetically modifying the purine operon. The increased activity of the purine promoter in B. amyloliquefaciens facilitated the biosynthesis of the purine nucleotide. The energy metabolism was then modified by deleting the less efficient cytochrome oxidase bd in the respiratory chain to further improve guanosine production. Increasing respiratory energy generation exerted a synergistic effect on increasing guanosine production. Therefore, the combined regulation of purine pathway, central carbon metabolism, ATP synthesis, and NADPH (an important co-factor) could be used as a modification strategy to further improve the performance of purine producing strains.