Promoting cytidine biosynthesis by modulating pyrimidine metabolism and carbon metabolic regulatory networks in Bacillus subtilis

Abstract

Background

The modification of single or multiple genes via metabolic engineering can lead to the dysregulation of central metabolism and affect bacterial growth and metabolite accumulation. Meanwhile, transcription factor engineering can trigger metabolic network reprogramming at the global or systemic level, redirecting metabolic flux toward the synthetic pathways of target metabolites. In this study, we modulated pyrimidine and carbon-nitrogen metabolism in Bacillus subtilis through transcription factor engineering to promote the synthesis of cytidine, a drug intermediate.

Results

First, cytidine synthesis was enhanced by knocking out the transcriptional regulator PyrR, which increased the cytidine titer during shake flask fermentation to 0.67 g/L. Second, mutations in the transcriptional regulator catabolite control protein A (CcpA) significantly promoted cytidine synthesis, increasing the shake flask titer to 2.03 g/L. Finally, after culture in a 5 L fermenter, the cytidine titer reached 7.65 g/L, which was 3.77-fold that of shake flask fermentation. Moreover, a cytidine yield and productivity of 0.06 g/g glucose and 0.16 g/L/h, respectively, were achieved. Subsequently, the regulatory mechanisms through which PyrR and CcpA modification affect cytidine biosynthesis were explored through multi-omics analysis. Transcriptome and metabolome analysis revealed that coordinated alterations in carbon, nitrogen, nucleotide, and amino acid metabolism were essential to promote cytidine synthesis. However, the increased cytidine production in recombinant strains was attributed to the enhancement of pyrimidine metabolism, the Phosphotransferase (PTS) system, the tricarboxylic acid (TCA) cycle, the pentose phosphate (PP) pathway, and nitrogen metabolism.

Conclusions

These results indicate that PyrR knockdown can enhance pyrimidine metabolic pathway and promote cytidine synthesis. CcpA mutation can reprogram the central carbon-nitrogen metabolic network, change the metabolic flow to de novo synthesis pathway of pyrimidine nucleoside, increase the supply of cytidine synthesis precursors and promote the accumulation of cytidine. Overall, regulation of engineered carbon and nitrogen metabolic networks is essential for improving the efficiency of microbial cell factories.

Introduction

Cytidine is a valuable intermediate in the synthesis of antiviral and antineoplastic drugs and is widely used in the pharmaceutical industry and the nutraceutical and healthcare sectors [1,2,3]. In recent years, the production of cytidine using metabolically engineered or mutant strains of bacteria — such as Escherichia coli [4], Bacillus subtilis [5] and Bacillus amyloliquefaciens [6] — has emerged as a more efficient alternative to traditional chemical synthesis and nucleic acid hydrolysis techniques [7]. Metabolic engineering focuses on achieving high metabolite titers and productivity in microbial cell factories while ensuring process economy [8]. In order to achieve these goals, researchers have employed various rational approaches to modify cytidine-producing strains of bacteria. These approaches have included the blockade of catabolic pathways [9], release of metabolite feedback inhibition [6, 10], enhancement of precursor supply [10, 11], transporter engineering [5, 12], and cofactor engineering [11, 13]. However, the modification of single or multiple genes often proves inadequate for achieving the desired outcomes and product yield, since such modifications can lead to metabolic pathway disorders, resulting in an imbalance between bacterial growth and product accumulation [14]. With advancements in functional genomics, transcription factors (TFs) have been recognized as essential tools for reshaping metabolic pathways due to their unique capacity for “multifaceted regulation,” addressing the limitations associated with single-gene modifications. Meanwhile, as TFs are integral components of gene regulation, their modifications can trigger the reprogramming of metabolic networks at both the global and systemic levels, facilitating the redirection of metabolic flux toward the synthetic pathways of target products [15, 16]. Recent studies have employed TFs engineering methods to reconstruct carbon-nitrogen metabolism in various microbial strains, thereby enhancing the production of compounds such as extracellular enzymes, industrial chemicals, and heterologous proteins [17, 18]. Nevertheless, there is a notable paucity of research focused on improving pyrimidine nucleoside synthesis through TFs engineering strategies.

In B. subtilis, the de novo synthesis pathway of pyrimidine nucleosides is primarily regulated by the pyrimidine operon (pyr operon), which consists of ten genes that encode a total of eight enzymes [19]. Among these genes, eight code for six enzymes that participate in the de novo synthesis of cytidine. Notably, the pyrR gene encodes a repressor protein that regulates the expression of the pyr operon itself [19]. For example, when repressor PyrR bind to specific regulatory sequences (binding loops, BLs), the transcription of the pyr operon genes (pyrAA/AB/B/C/D/E/F/K) is inhibited. Conversely, when PyrR binds to 5-phosphoribosyl-1α-pyrophosphate (PRPP) but not to the downstream BLs, the pyr operon genes undergo normal transcription. A study conducted by Zhu et al. demonstrated that the knockout of pyrR significantly enhances the expression levels of pyr operon genes, resulting in increased uridine biosynthesis [5]. Similarly, Wang et al. reported that modifying the pyr operon can attenuate the regulatory influence of the PyrR protein [20]. CcpA is a principal TF involved in the regulation of carbon and nitrogen metabolism in B. subtilis. This TF is known to regulate approximately 300 genes, including those encoding enzymes responsible for the transport and utilization of carbon sources (bglP, ptsH, glpF, amyE, amyO, pfkA, gapA, pdhA, gntR), short-chain fatty acid (SCFA) metabolism (ackA, pta, acsA, acuABC), the PP pathway (zwf, pgl, rpe, purA), the TCA cycle (citB, citZ), branched-chain amino acid synthesis (asnB, ilv-leu operon), and nitrogen metabolism, e.g. glutamate dehydrogenase (GDH), glutamate synthase (GOGAT) and glutamine synthetase (GS) (rocG, gudB, glnA, gltAB operon), which is essential for ammonium assimilation [21,22,23,24]. Both PyrR and CcpA play crucial roles in coordinating pyrimidine nucleotide metabolism as well as carbon and nitrogen metabolism in B. subtilis.

In this study, we used the recombinant strain BSNX2 obtained by knocking out the cytidine deaminase and uridine kinase coding genes cdd and udk to block the cytidine degradation pathway as the starting strain. We investigated the influence of the regulators of pyrimidine nucleoside metabolism (PyrR) and carbon and nitrogen metabolism (CcpA) on cytidine production in B. subtilis. Our findings revealed that the cytidine titer in the pyrR knockout strain (BSNX3) increased, while the titer in the ccpA knockout strain decreased, along with reductions in sugar consumption and biomass. Consequently, a mutation library based on the D1 and D2 domains of ccpA was constructed, leading to the identification of a recombinant strain, BSNX4-DM, which exhibited a significantly increased cytidine titer. To elucidate the mechanisms by which PyrR and CcpA regulate cytidine synthesis, we performed transcriptomic and metabolomic analyses on the recombinant strains.

Materials and methods

Strains, plasmids and cultures

All strains and plasmids used in this study are listed in Table 1. E. coli DH5α was used as the intermediate cloning host for plasmid construction, and transformed clones were screened using kanamycin (50 µg/mL). B. subtilis was cultured aerobically in LB medium or on LB agar plates at 37℃, and kanamycin was added to the medium for screening. We prepared and electrotransformed B. subtilis competent cells according to the protocol described by Anagnostopoulos and Spizizen [25] and Jin et al. [26].

Table 1 Strains and plasmids used in this study

Construction and expression of recombinant plasmids

Recombinant plasmids were assembled using the method described by Ferrando et al. and Altenbuchner [27, 28]. All primers used are listed in Supplementary Table 1. The plasmid pJOE8999 was linearized using the restriction enzyme BsaI, and PyrR-sgRNA and CcpA-sgRNA sequences were introduced into the linearized plasmid to obtain the pJOE8999-PyrR-sgRNA and pJOE8999-CcpA-sgRNA recombinant plasmid. With the genome of the BSNX2 strain as a template, the upstream and downstream homology arms of the pyrR and ccpA genes were amplified using the primer pairs PyrR-U-F/PyrR-U-R, PyrR-D-F/PyrR-D-R, CcpA-U-F/CcpA-U-R, CcpA-D-F/CcpA-D-R containing SfiI restriction sites, respectively. Then, the upstream and downstream homologous arms of the pyrR and ccpA genes were ligated into a sequence via overlap extension PCR, and pJOE8999-PyrR-sgRNA and pJOE8999-CcpA-sgRNA plasmids linearized by restriction endonuclease SfiI were introduced to obtain the recombinant plasmid pJOE8999_PyrR and pJOE899_CcpA. The recombinant plasmid was transformed into the BSNX2 strain via electrotransformation, and the recombinant strains BSNX3 and BSNX4 was obtained.

The recombinant strains stored at -80℃ were inoculated on LB plates and cultured at 37℃ for 12 h. The seed medium components were as follows (g/L): glucose, 20; Yeast powder, 10; Peptone, 5; NaCl, 2.5; MgSO₄ • 7H₂O, 1; KH₂PO₄, 2; Sodium glutamate, 1.5. The shake flask fermentation medium components were as follows (g/L): glucose, 80; Peptone, 10; Yeast powder, 15; NaCl, 2.5; MgSO₄ • 7H₂O, 3; KH₂PO₄, 2.5; (NH₄)₂SO₄, 6; Sodium glutamate, 5. The pH value of the seed and fermentation medium was 7.0. Single colonies of recombinant strains on LB plates were picked and inoculated into 100 mL triangular flasks containing 10 mL of seed medium, and cultured overnight at 37 °C and 200 rpm with shaking. The seed culture was inoculated into a 500 mL flask containing 50 mL of fermentation medium and cultured at 37 °C, 200 rpm for 48 h. The fermentation broth was collected to detect the OD₆₀₀ value and cytidine titers. Based on the fermentation curve, an appropriate time point was selected to collect samples for transcriptome (logarithmic growth phase) and metabolome (logarithmic growth phase) analysis. By analyzing the changes in gene expression levels and metabolite contents in the recombinant strains, the effects of modification of transcriptional regulators PyrR and CcpA on metabolic pathways related to cytidine biosynthesis were explored.

CcpA homologous sequence alignment and construction of mutation library

In the MEME Suite (http://meme-suite.org/) website [29], the CcpA protein sequence of Bacillus subtilis 168 (NP_390852.1) was homologously compared with nine other CcpA protein sequences, including CcpA (Bacillus spizizenii) (WP_019715392.1), CcpA (Bacillus halotolerans) (WP_024122513.1), CcpA (Bacillus inaquosorum) (WP_003237907.1), CcpA (Bacillus amyloliquefaciens) (WP_289393829.1), CcpA (Bacillus atrophaeus) (WP_010789522.1), CcpA (Bacillus licheniformis) (WP_003184445.1), CcpA (Bacillus wiedmannii) (WP_001103309.1), CcpA (Bacillus cereus) (WP_001103303.1) and CcpA (Bacillus thuringiensis) (WP_001103303.1).

To obtain CcpA that promotes cytidine synthesis, two CcpA mutation libraries (D1 and D2 domains) were constructed. With the BSNX3 genome as the template, the 1.5-kb ccpA sequence was amplified using the targeted mutation primers CcpA-D1-F/CcpA-D2-R and CcpA-D2-F/CcpA-D2-R and inserted into pUC19 (pUC19-CcpA). Subsequently, linearized plasmids were obtained through PCR amplification using pUC19-CcpA as template and CcpA-F/R as primer. These plasmids were treated with polynucleotide kinase to phosphorylate the 5’ end. DpnI endonuclease was used to remove any circular DNA, and the plasmid was then transformed into DH5α cells following cyclization with T4 DNA ligase. Finally, the ccpA-targeted mutant fragment was ligated between the SfII site of the pJOE8999-CcpA-sgRNA plasmid to obtain the recombinant plasmid pJOE8999-CcpA^M. The ccpA gene mutation library was obtained following electrotransformation into the mutant BSNX3. The ccpA gene mutant in the domain D1 mutation library carries a mutation site, 43E; The D2 domains of the ccpA gene contained four mutation sites, 77 S/103 M/281S/314P. These five amino acids could be mutated into any combination of random amino acids.

High-throughput screening of CcpA gene mutation library

Mutant screening was performed using a previously described high-throughput screening method [4, 30]. First, the cultured mutants were transferred into a 96-well plate containing 200 µL seed culture medium and incubated at 37 °C, 220 rpm for 12 h. Then, an appropriate amount of culture was inoculated into a 96-well plate containing 200 µL fermentation culture medium, and continued to be cultured at 37 °C, 220 rpm for 48 h. After fermentation, the supernatant was collected via centrifugation, diluted, and transferred to a 96-well plate, and the absorbance of each well was measured at 270 nm using a microplate reader. The cytidine content (g/L) was evaluated according to the following equation: Cytidine content = (OD₂₇ – 0.019472)/1.5386. The screened mutants with the highest cytidine titer were used for shake flask fermentation and validation experiments.

Transcriptomics analysis

BSNX2, BSNX3 and BSNX4-DM strains were collected during the logarithmic growth phase of shake flask fermentation and rapidly frozen in liquid nitrogen. The samples were sent to Novogene (Beijing, China) for the extraction of total RNA, and the RNA integrity and quantity were accurately detected using an Agilent 2100 bioanalyzer. Transcriptome sequencing was performed on the Illumina HiSeq-4000 sequencing platform. The analysis of gene expression was performed using DESeq2 [31], and the screening criteria for differentially expressed genes (DEGs) were |log₂ FC| > 1.2 and p_adj < 0.05 [32]. DEGs were further analyzed based on Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment.

Metabolomics analysis

The culture in the logarithmic growth phase was centrifuged to collect the supernatant, which was quickly frozen in liquid nitrogen for metabolite extraction. Subsequently, the extraction buffer (CAN: methanol = 1:4, v/v) was added to the sample, and the supernatant was collected via centrifugation for liquid chromatography coupled with mass spectroscopy. With a 0.1% formic acid aqueous solution (A) and 0.1% formic acid acetonitrile solution (B) as the mobile phase, the samples were analyzed using a UHPLC system (LC-30 A, Shimadzu) coupled with a Waters ACQUITY Premier HSS T3 Column and a TripleTOF 6600+ mass spectrometer (SCIEX, Foster City, CA, USA). Analyst TF 1.7.1 Software was employed for data acquisition. ProteoWizard was used to convert the original data into the mzXML format [33], and peak extraction, alignment, and retention time correction were performed using the XCMS program [34]. In this study, VIP > 1.2 & p-value < 0.05 (Student’s t-test) were used as the criteria to identify differential metabolites. Further functional annotation and pathway enrichment analysis were performed using the KEGG database.

Real-time quantitative PCR (RT-qPCR) analysis of gene transcription level

The strains cultured to the logarithmic growth phase were centrifuged at 4000×g and lower temperature to collect bacterial pellets. Then, total RNA was extracted according to the procedure described in the reagent kit (OMEGA, R6950-01), and the reverse transcription kit (Vazyme, R333-01) was used to synthesize cDNA according to the manufacturer’s instructions. Subsequently, RT-qPCR was performed according to the kit instructions (Vazyme, Q711-02), and data analysis was performed using the qTOWER³G PCR system. The 16 S rRNA gene was employed as the reference gene; the RT-qPCR primers are shown in Supplementary Table 1. Data processing was performed using the 2^-∆∆Ct method [35], and at least three replicates were examined for each sample.

Fed-batch culture in a 5 L bioreactor

The strains BSNX4-DM with the best production performance were cultured in a fed-batch 5 L bioreactor (Bailun, Shanghai, China). The components of fed batch fermentation medium (g/L): glucose, 80; Yeast powder, 20; (NH₄) ₂SO₄, 5; KH₂PO₄, 1; K₂HPO₄, 5; Sodium citrate, 10; Sodium glutamate, 20; MgSO₄ • 7H₂O, 1.5; MnSO₄, 0.02; ZnSO₄, 0.02; the pH was adjusted to 7.0. The seed culture was inoculated into a 5 L bioreactor containing 1.2 L of fermentation medium with 5% inoculation amount. The aeration rate and stirrer speed were controlled at 1.0 vvm and 500 rpm, respectively, until the end of fermentation. The pH value was maintained at 7.0 by adding 50% ammonia water, and the temperature was maintained at 37℃. By adding 600 g/L of glucose, the glucose concentration in the fermentation broth was maintained at 5 g/L.

Determination of cytidine, OD₆₀₀ values, and glucose concentrations

The cytidine titer in the fermentation broth was measured using high-performance liquid chromatography (HPLC) (1260 Infinity II, Agilent) on a system equipped with an Eclipse XDB-C18 column (Agilent, USA) and a PDA detector. For HPLC, 20 mM ammonium acetate and 100% methanol (96 : 4) were used as the mobile phase at a flow rate of 1.0 mL/min at 30℃. The absorbance of the cells at 600 nm was measured using an ultraviolet spectrophotometer (Metash, Shanghai, China). The residual glucose content in the fermentation broth was detected via a biochemical sensor (SBA-40E, Shandong, China).

Statistical analysis

All data are expressed as mean ± standard deviation (n = 3). Two-sample t-test and one-way ANOVA were performed using SPSS 22.0 software. A p value of less than 0.05 was considered statistically significant, with * indicating p < 0.05 and ** indicating p < 0.01.

Results and discussion

Effects of transcriptional regulator (pyrR and ccpA) knockout on cytidine biosynthesis

To study the effects of PyrR and CcpA on cytidine biosynthesis, the recombinant strains BSNX3 and BSNX4 were constructed. Shake flask fermentation revealed that the cytidine titer of strain BSNX3 was 0.67 g/L, which was 109.38% higher than that of strain BSNX2 (0.32 g/L). Notably, the yield of BSNX3 was significantly improved, reaching 0.0083 g/g glucose, which was 107.5% higher than that of BSNX2 (Table 2). Without any significant change in cell growth (Fig. 1A, B). RT-qPCR revealed that the transcription levels of the pyr operon genes pyrAA/AB/B/C/D/F/K were significantly up-regulated in the BSNX3 strain (Fig. 1C). This indicated that the disruption of the transcriptional regulator PyrR can attenuate the transcriptional repression of the pyr operon, allowing the metabolic flux to flow toward cytidine biosynthesis, consistent with previous reports [5]. As shown in Table 2, at the end of fermentation, the cytidine titer of strain BSNX4 was 0.5 g/L, and the yield was 0.0063 g/g glucose, which was 25.37% and 24.09% lower than that of BSNX3, and the OD₆₀₀ decreased from 10.21 to 8.05. These results indicated that knocking out the transcriptional regulator CcpA affects the transport and utilization of carbon sources in B. subtilis, leading to the dysregulation of central carbon metabolism and indirectly decreasing the metabolic flux of the cytidine synthesis pathway. Similar results were also observed in a ccpA knockout strain of Bacillus licheniformis [23]. To obtain mutants that promote cytidine production after the functional modification of CcpA, a CcpA domain mutation library was constructed.

Fig. 1

Effects of transcriptional regulator (pyrR and ccpA) knockout on cytidine biosynthesis in B. subtilis. (A) Cell growth curves of different B. subtilis mutants. (B) Titers of cytidine of different B. subtilis mutants (C) The mRNA expression levels of the pyr operon gene in mutant BSNX3

Table 2 Results of shake flask fermentation of different Recombinant strains

Domain analysis and mutation library construction of global transcription factor CcpA

The genetic relationship after homology alignment of CcpA protein sequences is shown in Fig. 2A. The domain structure of the global transcription factor CcpA was analyzed using NCBI’s CD-Search tool (http://www.ncbi.nlm.nih.gov/cdd/) [36]. The results showed that the CcpA protein consists of an N-terminal DNA-binding domain (DBD) and a C-terminal dimerization effect binding domain (DEBD) (Fig. 2B). The DBD accounts for the first 60 amino acid residues of the CcpA protein structure and contains two distinct DNA-binding elements: (i) a three-helix bundle, in which helices 1 and 2 constitute the major groove-binding helix-turn-helix motif (HTH) (helix 1: residues 5–12, helix 2: residues 16–24, helix 3: residues 31–44), and (ii) a hinge helix (helix 4; residues 50–58) that is embedded in the minor groove of the central CpG step and thereby introduces a kink in the DNA structure. Owing to the conformational flexibility of residues 45–50 in the HTH module and its flexible attachment with the hinge helices, the HTH motif itself demonstrates plasticity. Thus, it can optimally dock into the major groove of DNA [22, 37]. Meanwhile, the DEBD spans amino acid residues 60 to 339 and contains two structurally similar N and C subdomains. Specifically, the N subdomain consists of a six-stranded parallel β-sheet structure and four surrounding α-helices. Meanwhile, the C subdomain consists of five β-strands and five surrounding α-helices. Notably, the DEBD can bind to the Ser46-phosphorylated form (HPr-Ser46-P) of the serine-containing phosphocarrier protein (HPr) and the Ser46-phosphorylated form (Crh-Ser46-P) of the serine-containing HPr-like protein (Crh, for the catabolite repression HPr). This interaction triggers the allosteric switch of CcpA, allowing CcpA to bind to the catabolite-responsive elements (cre) site of the cognate DNA sequence [37,38,39,40,41]. Notably, these cre sites are composed of semi-palindromic sequences with the consensus sequence “WTGNAANCGNWNNCWW (R represents G or A, W represents A or T, N represents any base)” [42]. The location of these sites in the promoter determines whether CcpA acts as a repressor or activator. For example, when the cre site is upstream of the promoter, CcpA acts as an activator, but when it is downstream of the promoter, CcpA acts as a repressor [43, 44].

Fig. 2

Analysis of the domains of CcpA. (A) The homology of B. subtilis CcpA protein sequence and other nine CcpA protein sequences were aligned in the evolutionary tree. (B) The structure of CcpA with DNA molecular binding. (C) Sequence logos of alignment results

The alignment of homologous protein sequences revealed that most of the residues within the two domains of CcpA in B. subtilis were highly conserved, although a few residues showed significant variations. Therefore, we constructed two mutation libraries for residues showing large interspecies differences in DBD and DEBD (Fig. 2C). In the DBD, one residue (43E) was mutated to construct a mutant library, D1. Additionally, four residues that showed considerable differences across homologous sequences (77 S/103M/281S/314P) were selected from the DEBD to construct a mutant library named D2 (Fig. 3C). Using high-throughput screening, the top 20 mutants showing the highest absorbance at 270 nm were identified from the mutation libraries D1 and D2 (Fig. 3A). The first five mutants in the mutant libraries D1 and D2 showing the highest absorbance at 270 nm were validated through shake flask fermentation experiments (Fig. 3B, C). D1-M11 (E43D) showed the highest cytidine titer of 0.995 g/L in mutant library D1. The cytidine titer of D2-M7 (S77A/M103L/S281A/P314K) in mutant library D2 was even higher at 1.37 g/L. Thus, the recombinant strain BSNX4-DM was obtained by integrating these five mutations (E43D /S77A/M103L/S281A/P314K) in the DBD and DEBD of CcpA. As shown in Table 2, compared with BSNX3, the cytidine titer of BSNX4-DM increased from 0.67 g/L to 2.03 g/L, the yield increased from 0.0083 g/g glucose to 0.0254 g/g glucose, the productivity increased from 0.014 g/L/h to 0.042 g/L/h, and the OD₆₀₀ increased from 10.21 to 11.23.

Fig. 3

Mutation library construction of CcpA. (A) The top 40 mutants with the highest absorbance at 270 nm were screened from the CcpA mutant library by high-throughput screening. (B) The cytidine titers of the screened mutant libraries D1 (E43D) and D2 (S77A/M103L/S281A/P314K), and the combined mutant BSNX4-DM of the two domains (E43D/S77A/M103L/S281A/P314K). (C) In the wild-type CcpA domain D1, the residue E at position 34 was mutated to D. In domain D2, the residue S at position 77 is mutated to A; the residue M at position 103 is mutated to L; the residue S at position 281 is mutated to A; and the residue P at position 314 is mutated to K

Fed-batch culture of recombinant strain BSNX4-DM

To determine the productivity of the recombinant strain BSNX4-DM, we performed fed-batch culture in a 5 L bioreactor. As shown in Fig. 4, the early stage of fermentation was dominated by growth, and the OD₆₀₀ peaked at 41.34 after 24 h. Cytidine synthesis was detected after 9 h, and the titer reached 7.65 g/L at 48 h, which was 3.77-fold that of the shake flask culture titer (2.03 g/L). A cytidine yield and productivity of 0.06 g/g glucose and 0.16 g/L/h, respectively, were achieved. During the fermentation process, the concentration of glucose was controlled around 5 g/L.

Fig. 4

Fed-batch production of cytidine using BSNX4-DM in a 5-L bioreactor. The time courses of cytidine titer, cell growth curves, and glucose concentration are presented. The maximum OD₆₀₀ value was 41.34, the cytidine titer was 7.65 g/L, and the cytidine yield was 0.06 g/g glucose

Multi-omics analysis of the effects of PyrR and CcpA modifications on cytidine biosynthesis

With the advancement of multi-omics technology, there is a growing shift towards comprehensive system-level analyses instead of isolated decomposition studies [37]. Transcriptomics explores gene expression and transcriptional regulation at the RNA level, while metabolomics extends this approach by revealing the mechanism through which metabolic changes influence biological activities and affect phenotypes [38, 39]. In recent years, studies integrating transcriptomic and metabolomic data have emerged as a key component of systems biology research, revealing the state of post-transcriptional regulation. Therefore, this combined transcriptomics and metabolomics to analyze the metabolic pathways (central carbon metabolism, nitrogen metabolism, nucleotide metabolism, and amino acid metabolism) that directly or indirectly affect cytidine biosynthesis after the modification of PyrR and CcpA in B. subtilis.

Transcriptome analysis of PyrR and CcpA-modified strains

To explore the effects of PyrR and CcpA on cytidine synthesis at the gene level, the transcriptomes of the strains BSNX2, BSNX3, and BSNX4-DM were analyzed. As shown in Fig. 5A, compared with the control strain BSNX2, 693 DEGs (up-regulated: 371; down-regulated: 322) emerged in the BSNX3 strain when pyrR was knocked out. These genes were mainly enriched in protein metabolic processes, organic nitrogen compound metabolic processes, ribosomes, cells, and the structural constituents of ribosomes (Fig. 6B). KEGG pathway enrichment analysis revealed that these DEGs were mainly enriched in ribosomes, pyrimidine metabolism, non-ribosomal peptide structures, nitrogen metabolism, and valine, leucine, and isoleucine biosynthesis (Fig. 6C). These results indicated that pyrR knockout triggers nucleotide metabolism.

Fig. 5

Transcriptome analysis of PyrR and CcpA-modified strains. (A) (B) Volcanic map of DEGs

The CcpA mutant contained 1396 DEGs (up-regulated: 685; down-regulated: 711) (Fig. 5B). GO enrichment analysis showed that the DEGs were mainly enriched in nucleoside metabolic processes, carbohydrate transport, the phosphoenolpyruvate-dependent sugar phosphotransferase system, and oxidation-reduction processes (Fig. 6C). KEGG pathway enrichment analysis showed that the DEGs were mainly enriched in carbon metabolism (PTS system, glycolysis/gluconeogenesis, pyruvate metabolism, C5-branched dibasic acid metabolism, TCA cycle), nitrogen metabolism, nucleotide metabolism (pyrimidine metabolism, purine metabolism), and amino acid metabolism (alanine, aspartate and glutamate metabolism, valine, leucine and isoleucine biosynthesis) (Fig. 6D). These data demonstrated that CcpA modification triggers pathways associated with carbon, nucleotide, nitrogen, and amino acid metabolism.

Fig. 6

Transcriptome analysis of PyrR and CcpA-modified strains. (A) (C) GO enrichment of DEGs. (B) (D) KEGG enrichment of DEGs. padj (adjusted p-value)

Metabolome analysis of PyrR and CcpA-modified strains

To detect the metabolite changes induced by the modification of the TFs PyrR and CcpA, metabolomics analysis was performed in strains BSNX2, BSNX3, and BSNX4-DM. In contrast to the control strain BSNX2, there were 798 significantly up-regulated and 436 significantly down-regulated differential metabolites (DMs) in the PyrR knockout strain BSNX3 (Fig. 7A). According to KEGG pathway enrichment, the top three pathways associated with these metabolites were nucleotide metabolism, pyrimidine metabolism, and purine metabolism (Fig. 7B). A total of 1190 significantly up-regulated and 1372 significantly down-regulated DMs were identified in the CcpA mutant strain BSNX4-DM (Fig. 7C). These DMs were mainly enriched in nucleotide metabolism (Pyrimidine metabolism, Purine metabolism), carbon metabolism (Glycolysis, Pyruvate metabolism, TCA cycle, Pantothenate and CoA biosynthesis, Pentose phosphate pathway, Propanoate metabolism) and amino acid metabolism (Glycine, serine and threonine metabolism, Arginine and proline metabolism, Biosynthesis of amino acids, Valine, leucine and isoleucine biosynthesis, Histidine metabolism,) pathways (Fig. 7D). Thus, the results of combined transcriptomics and metabolomics revealed that central carbon metabolism, nitrogen metabolism, nucleotide metabolism, and amino acid metabolism were considerably altered after PyrR and CcpA modification.

Fig. 7

Metabolome analysis of PyrR and CcpA-modified strains. (A) (C) Volcanic map of DMs. (B) (D) KEGG enrichment of DMs

Key metabolic pathways in PyrR and CcpA-modified strains revealed by multi-omics

Nucleotide metabolism

Nucleotide metabolism mainly involves a de novo synthesis pathway and a salvage synthesis pathway. The de novo synthesis pathway, which uses aspartic acid, glutamine, PRPP, and HCO₃^- as raw materials, is the main pathway generating nucleotides through a series of enzymatic reactions [7]. Specific nodes in the de novo synthesis pathway of pyrimidine nucleosides are strictly regulated by the regulatory factor PyrR via effectors such as uridine-5’-monophosphate (UMP), uridine-5’-diphosphate (UDP), UTP, and PRPP [19, 45]. As shown in Fig. 8, most genes associated with the pyrimidine metabolic pathway (pyrAA, pyrAB, pyrB, pyrC, pyrD, pyrE, pyrF, pyrK, pdp, yfkN, cmk and ndk) were significantly up-regulated in the PyrR knockout strain, and only nrdE, tmk, and carA were down-regulated, consistent with the RT-qPCR results. The genes involved in purine metabolism pathway (purA, purF, purH, purN, purN) were significantly down-regulated. Metabolomics data showed that four metabolites from the pyrimidine metabolic pathway (UMP, Orotidine-5’-phosphate (OMP), UDP, Cytidine, Cytosine) were significantly up-regulated, while three from the purine metabolic pathway (adenosine, guanosine, adenine, allantoic acid) were significantly down-regulated, in line with the transcriptomics data. These results showed that knocking out the TF PyrR can increase the expression of genes involved in the de novo synthesis pathway of pyrimidine nucleosides, thus promoting cytidine synthesis.

In the CcpA mutant, the pyrimidine metabolism-related genes pyrE, pyrF, pyrAB, pyrK, pyrD, pyrAA, pyrC, nrdE, pyrB, ndk, pdp, pdeB, and yfkN were significantly up-regulated, whereas the genes nrdF, upp, carA, thyB, tmk, and dck were significantly down-regulated. This led to an increase in the levels of OMP, UMP, UDP, cytidine-5’-diphosphate (CDP), cytidine-5’-monophosphate (CMP), cytidine, cytosine, and uracil, and a corresponding decrease in the contents of deoxyuridine (Fig. 8). Moreover, except for purB, purC, guaA, and guaB, all other genes associated with the purine metabolic pathway were significantly down-regulated, leading to a decrease in the levels of the corresponding purine nucleosides (Adenosine, Guanosine, Adenine) (Fig. 8). These results demonstrate that CcpA modification can increase the expression of pyrimidine metabolic pathway-related genes, thus promoting cytidine production. In future studies, the direction of metabolic flow could be changed by modifying CcpA to increase the expression of de novo pyrimidine nucleoside synthesis pathway genes, thus further boosting the biosynthesis of these nucleosides.

Fig. 8

Comparison of multi-omics data in cytidine biosynthesis networks. The DEGs and DMs related to nucleotide metabolism, carbon metabolism, nitrogen metabolism and amino acid metabolism in recombinant strains were analyzed (Details in Supplementary file 2). Abbreviations: G6P, glucose-6-phosphate; F6P, fructose-6-phosphate; FBP, fructose-1,6-bisphosphate; DHAP, dihydroxyacetone phosphate; GAP, glyceraldehyde-3-phosphate; 1,3-BPG, glycerate 1,3-diphosphate; 3PG, 3-phosphoglycerate; GA, gluconic acid; PEP, phosphoenolpyruvate; PYR, pyruvic acid; OAA, oxaloacetate; MAL, malic acid; SUCC, succinic acid; α-KG, α-ketoglutaric acid; ICIT, isocitrate; CIT, citrate; 6PGL, 6-phosphogluconolactone; 6PG, 6-phosphogluconic acid; RU5P, ribulose 5-phosphate; R5P, ribose 5-phosphate; PRPP, 5-phosphoribosyl-1α-pyrophosphate; IMP, inosinemonphosphate; AMP, adenosine 5’-monophosphate; GMP, guanosine monophosphate; ALA, allantoic acid; L-Glu, L-glutamate; Gln, glutamine; Cit, citrulline; Arg, arginase; L-Asp, L-asparaginase; L-Hom, L-homoserine; L-His, L-histidine; BCAA, branched-chain amino acid; CP, carbamoyl phosphate; N-CA-Asp, N-carbamoyl aspartate; DHD, dihydroorotate; OMP, orotidine 5′-phosphate; UMP, uridine monophosphate; UDP, uridine diphosphate; UTP, uridine triphosphate; CTP, cytidine triphosphate; CDP, cytidine diphosphate; CMP, cytidine monophosphate

Carbon metabolism

In B. subtilis, central carbon metabolism provides the key precursors and energy required for the de novo synthesis of cytidine [7, 10]. The PTS system, glycolysis, TCA cycle, and PP pathway are the main components of the carbon metabolism network. Studies have demonstrated that the genes related to the transport and utilization of carbon sources in B. subtilis are strictly regulated by CcpA via effectors such as fructose-1, 6-diphosphate (FBP) to achieve optimal intracellular carbon metabolism flow [21, 46]. As shown in Fig. 8, in the CcpA mutant strain, the genes related to the PTS system (sacP, bglP, ptsJ, ptsG, ptsH, gamP, malP) were all significantly up-regulated, increasing the utilization of phosphoenolpyruvate (PEP) and promoting glucose transport. Thus, the contents of glucose, glucose-6-phosphate (G6P) and fructose-6-phosphate (F6P) increased and those of PEP decreased. Among the genes related to glycolysis glcK, pgi, gapA, gapB, pdhA, pdhB, pdhC, pdhD, pckA and acsA were up-regulated, while pfkA, fbaA, pgk, pgm, eno, pyk, ackA and pta were down-regulated, leading to a decrease in dihydroxyacetone phosphate (DHAP) and acetate. Additionally, the content of acetyl-CoA dropped below detectable levels. Except for icd, the other genes associated with the TCA cycle were significantly up-regulated. As a result, the content of pyruvic acid (PYR), malic acid (MAL) and succinic acid (SUCC) decreased, and α-ketoglutarate (α-KG) and isocitric acid (ICIT) were increased, while oxaloacetate (OAA) was not detected. All key genes linked to the PP pathway (zwf, pgl, gndA, gntZ, rpe, prs, ycdF) were significantly up-regulated, and only tktA and rpiB were down-regulated. This resulted in decreased 2-phospho-glyceric acid (2PG) levels and increased gluconic acid (GA), ribulose 5-phosphate (Ru5P), xylulose 5-phosphate (Xu5P), ribose-5-phosphate (R5P) and 5-phosphoribosyl-1α-pyrophosphate (PRPP) levels. However, since metabolism is inherently post-transcriptional, changes in gene expression and metabolite content are not typically synchronized. Hence, the same substance may exhibit different trends at different omics levels.

Overall, the results indicated that the modification of the TF CcpA can alter the transcription of genes associated with the PTS system, glycolysis, TCA cycle, and PP pathway, shifting the metabolic flux toward pyrimidine metabolic pathways and increasing the expression of de novo pyrimidine nucleoside synthesis pathway genes. Thus, this alteration ultimately promotes the biosynthesis of cytidine. The overexpression of genes associated with the PP pathway (e.g. zwf, gnd, ykgB and prs) is widely used as a strategy to construct pyrimidine nucleoside-producing strains [11, 47]. However, the modification of genes related to the PTS system, glycolysis and TCA cycle (e.g. ptsG, ptsH, ptsI, glcK, gapA, pycA, pdhA, citZ, citB and ackA) has rarely been performed to enhance pyrimidine nucleoside accumulation. Therefore, these genes may serve as potential candidates for improving pyrimidine nucleoside production in the future.

Nitrogen metabolism

As a key component of biological macromolecules (nucleic acids and proteins), nitrogen also serves as an essential element for microbial growth and is essential for development and information exchange in microorganisms [48]. However, nitrogen assimilation mainly occurs via the GDH assimilation pathway and GS/GOGAT assimilation pathway, which synthesize glutamate and glutamine — the two main nitrogen donors in microbial cells [49]. In B. subtilis, GDH, GS, and GOGAT are encoded by the genes rocG/gudB, glnA and gltAB, respectively. The combination of our transcriptomic and metabolomic data showed that the nitrogen metabolism-related genes rocG, gudB, glnA, nasE, gltA and gltB genes were significantly up-regulated after CcpA modification, and the contents of L-glutamate and L-glutamine were increased (Fig. 8). Hence, CcpA modification also seemed to affect nitrogen metabolism, especially the pathways involved in the biosynthesis of glutamate and glutamine, leading to cytidine accumulation.

Amino acid metabolism

In addition to being the basic units of protein, amino acids also act as important precursors of nucleotide biosynthesis [7, 50]. In addition, the one-carbon unit (methyl, methylene, methenyl, methynyl, cresol and iminomethyl) produced via the catabolism of some amino acids serves as a key synthetic raw material for purine and pyrimidine and represents the link between amino acids and nucleotides [51]. In our study, in the CcpA mutant strain, the expression of genes associated with the L-aspartate synthesis pathway (aspB, asnB) was up-regulated, leading to an increase in the content of L-aspartate. However, arginine, histidine, and branched-chain amino acid biosynthesis pathway-related genes (argGF, hisBCD, thrBC, ilvBCDHK) were significantly down-regulated, resulting in decreased levels of the corresponding amino acids (L-Arg, L-His, L-Lle, L-Leu) (Fig. 8). These results demonstrated that the increase in cytidine production in the CcpA mutant strains could be directly related to the attenuated activity of these branched metabolic pathways.

Conclusion

This study showed that knockout of the pyrR can effectively alleviate its transcriptional repression of the pyr operon and promote cytidine synthesis. Subsequently, the ccpA mutant library was constructed, and the mutant BSNX4-DM was screened to increase the cytidine titer to 2.03 g/L in shake flask fermentation. Finally, fed-batch fermentation in a 5 L fermenter reached a cytidine titer of 7.65 g/L. Through multi-omics analysis, it was found that the expression levels of pyrimidine metabolic pathway genes and metabolites in pyrR knockout strains increased, and the expression of purine metabolic pathway genes and metabolites was down-regulated. However, in ccpA mutant, the pyrimidine metabolic pathway was enhanced, and the expression of most genes in the purine metabolic pathway was decreased. The related genes of PTS system were significantly up-regulated, and glucose transport was accelerated. The expression of genes related to TCA cycle and PP pathway was up-regulated, and the metabolic flow turned to pyrimidine nucleoside synthesis pathway. The synthesis fluxes of glutamine and aspartic acid increased, and the synthesis pathways of arginine, histidine and branched-chain amino acids decreased, which led to an increase in the supply of cytidine synthesis precursors. However, the biosynthesis of cytidine involves complex regulatory networks, and the regulatory mechanism of mutant CcpA on the global metabolic flow of B. subtilis needs to be further elucidated in future studies. Overall, these strategies help to optimizing the overall transcription level of intracellular carbon-nitrogen metabolic networks, and could be combined with systems metabolic engineering methods for constructing more efficient microbial cell factories in the future.