Development of a defined medium for Corynebacterium glutamicum using urea as nitrogen source

Abstract

Corynebacterium glutamicum has been widely used for bulk and fine chemicals fermentation these years. In this study, we developed a defined medium for this bacteria based on the widely used CGXII minimal medium. We evaluated the effects of different components in CGXII on cell growth of C. glutamicum ATCC 13032 and improved the medium through single-factor experiment and central composite design (CCD). Urea, K₂HPO₄ and MgSO₄ were found to be significant factors. 7 out of the total 15 components were modified. (NH₄)₂SO₄, KH₂PO₄, and protocatechuic acid were eliminated. Amounts of urea and MgSO₄ were increased, and concentrations of biotin and glucose were reduced. The resulting R2 medium was proved to be more suitable for cell growth, plasmid amplification and protein production than the original recipe. Remarkably, cell biomass accumulation in R2 increased by 54.36% than CGXII. Transcriptome analysis revealed alteration of carbon metabolism, cation transport and energy synthesis, which might be beneficial for cell growth in R2. Considering the high nitrogen content and availability of urea, the new medium is simplified and cost effective, which holds attractive potential for future study.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13205-021-02959-6.

Introduction

With growing concerns about environmental issues and sustainable economy, production of industrially relevant chemicals utilizing microbial cell factories provides an eco-friendly alternative to current petro-based processes. As an aerobic, Gram-positive, nonsporulating, nonpathogenic bacterium with generally recognized as safe (GRAS) status, Corynebacterium glutamicum is a workhorse for industrial amino acid and vitamin production. Recent progress in synthetic and systems metabolic engineering technologies has turned C. glutamicum into a promising microbial platform for bioproduction of various value‐added products, such as aromatics, organic acids, diamines, biofuels and proteins from diverse feedstocks (Baritugo et al. 2018; Becker et al. 2018; Kogure and Inui 2018).

For microorganisms cultivation, an appropriate medium is of vital importance. To acquire comparable and repeatable data, synthetic medium is frequently used. Since it permits precise formulation with minimal complicated interactions. On the other hand, it has advantages on substrate costs and downstream product purification over complex medium most of the time. For C. glutamicum, CGXII is such a well-defined medium widely used in fundamental and applied research (Keilhauer et al. 1993). Over the years, this recipe has been adapted for different scenarios (Ko et al. 2018; Buchholz et al. 2014; Yu et al. 2015; Hoffmann and Altenbuchner 2014; Lee et al. 2015). However, few studies were focused on the effects of different compositions on cell cultivation particularly (Freier et al. 2016; Unthan et al. 2014).

Generally, medium formulation is a laborious, time-consuming, and costly process, because the optimal medium for one strain may be not applicable for others (Singh et al. 2016). As new mutants and strains are continuously being introduced, medium optimization has become an essential part of process development. With accumulation of high-throughput omics datasets and systems biology models, researchers can develop and optimize cell culture media more rationally (Savizi et al. 2019). Also, rapid development of micro-cultivation systems may enable parallel assessment of different recipes with high efficiency (Rohe et al. 2012; Morschett et al. 2017; Sundstrom and Criddle 2015; Jordan and Stettler 2014). On the other hand, conventional medium optimization strategies (e.g., single-factor experiment and statistical techniques) still have their place due to easy implementation and low equipment requirements.

In this study, effects of different compositions in CGXII on cell growth of C. glutamicum ATCC 13032 were assessed through single-factor experiment and central composite design (CCD). Accordingly, several ingredients were modified, generating an improved minimal medium R2, which supported for better cell growth than the original recipe. Then, availability of R2 was further shown in plasmid amplification and protein production. To reveal the underlying mechanisms, comparative transcriptomics was performed.

Materials and methods

Strain and media

Escherichia coli DH5α was used for cloning purpose and was propagated in Luria-Bertani (LB) medium at 37 °C under aeration. C. glutamicum ATCC 13032 was propagated in a commercial BHI medium (Hopebio, China) or CGXII medium (Keilhauer et al. 1993) at 30 °C, with aeration. BHI medium contained (per liter): 2 g glucose, 10 g tryptone, 5 g NaCl, 17.5 g brain heart infusion and 2.5 g Na₂HPO₄. CGXII medium contained (per liter): 20 g (NH₄)₂SO₄, 5 g urea, 1 g KH₂PO₄, 1 g K₂HPO₄, 0.25 g MgSO₄·7H₂O, 42 g 3-morpholinopropanesulfonic acid (MOPS), 10 mg CaCl₂, 10 mg FeSO₄·7H₂O, 10 mg MnSO₄·H₂O, 1 mg ZnSO₄·7H₂O, 0.2 mg CuSO₄, 0.02 mg NiCl₂·6H₂O, 0.2 mg biotin, 40 g glucose, and 30 mg protocatechuic acid (PCA). Solution of glucose and the media were autoclaved in separate at 115 °C for 30 min and required amounts of glucose were added into the media before use. Antibiotics (kanamycin 25 μg/mL, chloramphenicol 15 μg/mL) were added when necessary.

Plasmid construction

pEC-egfp was constructed as follows. The egfp gene (Cheng et al. 2020) was synthesized with a ribosomal binding site (AAGGAGA) added upstream of the ORF and cloned into the EcoRI-BamHI sites of pEC-XK99E (GenBank Accession no. AY219683.1). The sequence of constructed plasmid was confirmed by DNA sequencing.

Strain cultivation

For medium components assessment, a single colony of C. glutamicum ATCC 13032 was inoculated into a 300-mL shake flask containing 50 mL BHI medium and incubated for 24 h at 30 °C and 200 rpm. The preculture was inoculated into 50 mL of fresh CGXII medium (original or modified) with 1% inoculum. Cultivation was maintained at 30 °C and 200 rpm for 20 h. Optical density was measured at 4–6 h intervals.

For protein production, a single colony of C. glutamicum ATCC 13032 containing pEC-egfp was inoculated into a 300-mL shake flask containing 50 mL CGXII medium and incubated for 24 h at 30 °C and 200 rpm. The preculture was inoculated into 50 mL of specified medium with 2% inoculum. 0.25 mM IPTG was added when optical density reached about 0.8. Cultivation was maintained at 30 °C and 200 rpm for 24 h. Antibiotics were added as required.

Analytical methods

Optical density was measured at 600 nm (OD₆₀₀) with a spectrophotometer (Shimazu, Japan). Plasmids were extracted using a commercial kit (TIANGEN, China) following the manufacturer’s instructions and quantified using a Nanodrop 2000 spectrophotometer (Thermo Scientific, USA). Glucose and organic acids were quantitatively determined by high-performance liquid chromatography (HPLC) (Shimadzu, Japan) equipped with a refractive index detector (RID-10A) (Shimadzu, Japan) and an Aminex HPX-87H ion exclusion column (Bio-Rad, USA), as described previously (Yang et al. 2016). Secreted porphyrin compounds in the cell-free medium was estimated using a spectrophotometer at 405 nm (measuring Soret band) and 495 nm (measuring collective Q-bands). Green fluorescence (excitation at 488 nm and emission at 509 nm) was detected using a fluorescence spectrophotometer (Hitachi, Japan).

Transcriptomic analysis

RNA-seq was performed by Origingene (Shanghai, China) using the Illumina Hiseq 4000 platform (Illumina). Total RNAs from C. glutamicum cultivated in CGXII medium or R2 medium during exponential growth (10 h) were prepared using TRIzol reagent (Invitrogen, USA). Each sample was analyzed in triplicate. RNA was treated with RNase-free DNase I (Takara, Japan) to prevent contamination by traces of genomic DNA, and ribosomal RNAs were removed using the RiboZero rRNA removal kit (Epicenter, USA). The quality and quantity of RNAs were measured using the BioAnalyzer 2100 system (Agilent, USA), and RNA was fragmented and used as a template for PCR using random primers. Strand-specific cDNA libraries were prepared with the Illumina TruseqTM RNA sample prep Kit (Illumina). An Agilent 2100 Bioanalyzer and an ABI StepOnePlus Real-Time PCR System were used to determine the representation and quality of the library. Suitable libraries were sequenced using the Illumina Hiseq 4000 platform (Illumina). Raw sequence data were processed using Cutadapt 1.16 (https://cutadapt.readthedocs.io/en/stable/) and FastqStat.jar V1.0 (Cock et al. 2010).

The quality control of clean reads was performed using FastQC (version 0.11.4) (Andrews 2010). Hisat2 (Kim et al. 2015) was used to align the reads to the reference C. glutamicum ATCC 13032 genome sequence (NCBI: ASM284740v1). Quality control of sequencing libraries was performed using RSeQC 2.6.4 (http://rseqc.sourceforge.net/). The differential expression analysis of genes between the two groups was performed using edgeR (version: 3.24) (Robinson et al. 2010). A false discovery rate (FDR) < 0.05, and |log2 fold change (FC)|> 1 were selected as cutoff thresholds for the differentially expressed genes (DEGs) investigation. Gene ontology (GO) analysis of DEGs was performed using go_enrichment (version: 2.1.0) with a corrected P value ≤ 0.05 as a threshold (Consortium 2004).

Results

Identification of key components using single-factor experiment

To develop an effective minimal medium, the most widely used CGXII medium was referred (Keilhauer et al. 1993). The CGXII recipe comprises 15 components in total. Glucose is the main carbon source. Both urea and (NH₄)₂SO₄ serve as nitrogen source. KH₂PO₄ and K₂HPO₄ provide suitable pH while supplying phosphate. 3-Morpholinopropanesulfonic acid (MOPS) provides major buffering capacity against acidification during cell cultivation. Biotin complements the inherent auxotroph and maintains cell growth. Protocatechuic acid (PCA) acts as an iron chelator, which is supposed to be necessary to initiate cell division in C. glutamicum (Liebl et al. 1989). Other components include macro-and micro-metal elements.

To identify media components that have major impacts on cell growth, single-factor experiment was implemented. Micro-metal elements (FeSO₄·7H₂O, MnSO₄·H₂O, ZnSO₄·7H₂O, CuSO₄ and NiCl₂·6H₂O) were not included for two reasons. First, these nutrients were used in trace amounts that hardly influenced the substrate cost. Secondly, addition of metal elements was quite case-specific, especially in protein production. Since enzyme activity depends largely on metal ions, these nutrients should be determined each time. Besides, all other 10 components were assessed.

As the main carbon source, a sufficient amount of glucose should be supplied to acquire an acceptable cell growth. The original recipe contained 40 g/L glucose, which may be excess for routine batch cultivation purpose (a representative 20 h cultivation profile was shown in Fig. S1). We, thus, reduced the initial glucose to 20 g/L arbitrarily. However, it’s worth mentioning that increased dissolved oxygen would lead to a much higher biomass (e.g., when a baffled shake flask was used, the OD₆₀₀ reached about 50 at 20 h) and more glucose should be offered correspondingly.

Urea had a strong influence on cell growth. With 10 g/L urea added, OD₆₀₀ increased by 63.78% compared with the reference cultivation (Fig. 1A). However, when more urea was added, an adverse effect gradually emerged. This effect was at least partially due to the increase of culture pH resulting from urea decomposition (the final culture pH reached 9.0 when 40 g/L urea was added, compared with a pH of 6.0 in the reference cultivation). Also being a nitrogen source, (NH₄)₂SO₄ had a positive influence on cell growth. As we can see in Fig. 1B, 2 g/L (NH₄)₂SO₄ was sufficient to support cell growth and the original amount was superfluous. As a buffer agent, MOPS had a positive effect on cell growth, and we decided not to alter it (Fig. 1C). As the phosphate source, K₂HPO₄ had a positive effect on cell growth while KH₂PO₄ did not show a marked impact (Fig. 1D, E). As a result, we decided to omit KH₂PO₄ from the media in subsequent experiments. C. glutamicum is a biotin auxotrophic bacterium in which glutamate production is induced under biotin-limited conditions. During glutamate fermentation, as low as 0.5–1 μg/L biotin was used (Gutmann et al. 1992). In our results, 10 μg/L biotin was confirmed to be enough and the original amount was in excess (Fig. 1F). We determined to cut down the concentration of biotin to 10 μg/L accordingly. PCA (or other similar iron chelators) has been designated necessary to initiate cell division in C. glutamicum (Liebl et al. 1989). However, elimination of PCA did not result in a growth retardation in our study (Fig. 1G). MgSO₄ and CaCl₂ promoted cell growth within certain limits, and excessive amounts did not have further effects (Fig. 1H, I). From above results, urea, (NH₄)₂SO₄, K₂HPO₄, MgSO₄ and CaCl₂ were determined as significant input variables.

Fig. 1.

Effects of different components on cell growth of C. glutamicum ATCC 13032. A Effect of urea on cell growth of C. glutamicum ATCC 13032. B Effect of (NH₄)₂SO₄ on cell growth of C. glutamicum ATCC 13032. C Effect of MOPS on cell growth of C. glutamicum ATCC 13032. D Effect of K₂HPO₄ on cell growth of C. glutamicum ATCC 13032. E Effect of KH₂PO₄ on cell growth of C. glutamicum ATCC 13032. F Effect of biotin on cell growth of C. glutamicum ATCC 13032. G Effect of PCA on cell growth of C. glutamicum ATCC 13032. H Effect of MgSO₄ on cell growth of C. glutamicum ATCC 13032. I Effect of CaCl₂ on cell growth of C. glutamicum ATCC 13032. Strains were cultured in 300-mL shake flasks containing 50 mL modified CGXII medium, at 30 °C and 200 rpm. The initial glucose concentration was 40 g/L. Results are averages from three independent experiments, with standard deviations indicated by error bars. *Indicates original concentration. Columns marked with the same letter indicate no difference in means using the t-test (P > 0.05)

Optimization of medium composition with central composite design (CCD)

To investigate potential interactions between the previously identified medium components of influence, a full factorial experimental design with 52 experiments was applied. Variables and levels are shown in Table 1 and detailed experimental design could be found in Table S1. Other factors were set as follows: glucose 20 g/L, KH₂PO₄ 0 g/L, MOPS 42 g/L, biotin 10 μg/L, PCA 0 mg/L, trace elements the same as the original medium. By results analysis and model improvement, a canonical equation was finally obtained:

$${\text{OD}}_{600}=3.93475+2.85886X_1-1.26512X_3+0.0124737X_4-0.127847X_1^2+0.256372X_1{X}_3+0.00378728X_1{X}_4(P<0.001,R^2=0.9697).$$

Table 1.

Assigned concentrations of each compound at different levels of the CCD

Variable	Compound	Levels
		− 1	0	+ 1
X1	Urea	5	12.5	20
X2	(NH4)2SO4	1	1.5	2
X3	K2HPO4	0.5	0.75	1
X4	MgSO4	65	82.5	100
X5	CaCl2	20	35	50

Concentrations are given in g/L, except for MgSO₄ (stock solution: 100 g/L) and CaCl₂ (stock solution: 10 g/L) expressed in μL

The above model indicated that urea, K₂HPO₄ and MgSO₄ had prominent effects on cell growth with urea to be the critical factor. However, (NH₄)₂SO₄ and CaCl₂ showed little impact, which was inconsistent with the single-factor experiment. We decided not to change the concentration of CaCl₂ in following studies. From Fig. 2, it is clear that K₂HPO₄ and MgSO₄ had positive effects on cell growth in our test range but the impact of urea was parabolic. The expected maximum OD₆₀₀ was 27.79 with urea 13.64 g/L, K₂HPO₄ 1 g/L and MgSO₄ 0.2 g/L.

Fig. 2.

Contour plots of C. glutamicum ATCC 13032 cultivation. A Urea (g/L), K₂HPO₄ (g/L). B Urea (g/L), MgSO₄ (g/L). C K₂HPO₄ (g/L), MgSO₄ (g/L). The white square box represents OD₆₀₀ of cultures. The arrows represent the increase of OD₆₀₀

To test the reliability of the model, experiment was implemented with the predicted optimal parameter (generating medium R1). R1 without (NH₄)₂SO₄ (R2) was also tested. Comparison of R2 and the original CGXII recipe was shown in Table 2. As shown in Fig. 3A, the final OD₆₀₀ of R1 was slightly lower than the expected maximum (25.19 vs. 27.79), which may be due to batch-to-batch difference. Still, the final biomass increased by 54.36%, compared with the original medium. The specific growth rate of R1 was 0.32 h⁻¹, while that of the original medium was 0.19 h⁻¹. Elimination of (NH₄)₂SO₄ (R2) demonstrated no significant effect on cell growth. Residual glucose and organic acids in the cultivation broth were shown in Fig. 3B. 20 g/L glucose was proved still to be sufficient (as demonstrated in R2 with 7.52 g/L residual glucose). Lactate significantly decreased (7.98 g/L in CGXII vs. 0.2 g/L in R2) while acetate and succinate decreased moderately. Glutamate accumulation increased from 0.2 g/L in CGXII to 0.78 g/L in R2. We also found that the supernatant of R2 showed a pink color while the supernatant of CGXII was near-transparent to light yellow (inset of Fig. 3C). This color difference was consistent with the secreted porphyrin compounds of CGXII and R2, as shown in Fig. 3C. Secreted porphyrin of BHI was not determined due to the interference of the medium color.

Table 2.

Comparison of medium components between standard CGXII and R2 medium

Component	Concentration in reference medium composition	Concentration in R2 medium composition
Glucose	40 g/L	20 g/L
(NH4)2SO4	20 g/L	0 g/L
Urea	5 g/L	13.64 g/L
KH2PO4	1 g/L	0 g/L
K2HPO4	1 g/L	1 g/L
MOPS	42 g/L	42 g/L
FeSO4 * 7H2O	10 mg/L	10 mg/L
MnSO4 * H2O	10 mg/L	10 mg/L
ZnSO4 * 7H2O	1 mg/L	1 mg/L
CuSO4 * 5H2O	0.31 mg/L	0.31 mg/L
NiCl2 * 6H2O	0.02 mg/L	0.02 mg/L
CaCl2	10 mg/L	10 mg/L
MgSO4	0.13 g/L	0.2 g/L
CoCl2 * 6H2O	0.31 mg/L	0.31 mg/L
PCA	30 mg/L	0 mg/L
Biotin	0.2 mg/L	0.01 mg/L

Boldfaced components were changed in R2 medium while components in regular font remained unchanged compared with CGXII medium

Fig. 3.

Assessment of the R2 medium. A Time course of C. glutamicum ATCC 13032 cultivated in different media. Strain was cultured in a 300-mL shake flask containing 50 mL of each medium, at 30 °C and 200 rpm. B Residual glucose and organic acids in the broth of (A). C Secreted porphyrin in the broth of (A). Measurement was performed at 405 nm (measuring Soret band) and 495 nm (measuring collective Q-bands). D Plasmid production. 1 mL culture grown in various media was used for plasmid extraction using a commercial kit. E Protein production. C. glutamicum ATCC 13032 harboring pEC-egfp was grown in various media with induction. C. glutamicum ATCC 13032 harboring pEC-XK99E was used as a negative control. Results are averages from at least three independent experiments, with standard deviations indicated by error bars. *Indicates significant difference using the t-test (P < 0.05)

The availability of R2 was further assessed in terms of plasmid and protein production. For plasmid production, pEC-XK99E (Kirchner and Tauch 2003) or pXMJ19 (Jakoby et al. 1999) (two widely used medium-copy-number plasmids) was transformed into C. glutamicum ATCC 13032 and resulting strain was cultivated in CGXII, R2 or BHI media. For both plasmids, the R2 medium tended to be better than CGXII, although the difference was not significant for pEC-XK99E at 95% confidence level (P = 0.084). CGXII and BHI media showed similar capacity (P > 0.05, Fig. 3D).

For protein production, GFP was chosen as a model due to its easy detection by fluorescence of the active protein. The egfp gene (Cheng et al. 2020) was cloned into pEC-XK99E, resulting in pEC-egfp. C. glutamicum ATCC 13032 harboring pEC-egfp was cultivated in CGXII, R2 or BHI media. Fluorescence intensity was measured 12 h and 24 h after induction. Bacteria in R2 showed a stronger fluorescence signal than in CGXII at both 12 h and 24 h, which indicated more protein production (Fig. 3E). Fluorescence intensity of BHI culture was the strongest at 12 h, while it was the weakest at 24 h. This result further demonstrated the feasibility of R2 for C. glutamicum.

RNA-seq analysis

To reveal the underlying mechanisms, comparative transcriptomics was performed. Total RNAs from C. glutamicum cultivated in CGXII medium or R2 medium during exponential growth (10 h) were prepared, processed and analyzed. Data quality of raw reads and clean reads was high as reflected by the base-position error rate and mapped ratio (Table 3). To test the reliability of the experiment, correlation analysis between the CGXII group and R2 group was performed. Principal component analysis (PCA) showed that the samples between the two groups were distributed in a decentralized way, while the samples within groups were clustered (Fig. 4A). Differentially expressed genes (DEGs) revealed a total of 56 up-regulated genes and 109 down-regulated genes in the R2 group compared with the CGXII group (|log₂ FC|> 1, P < 0.05, Fig. 4B).

Table 3.

Transcriptomic profiles of C. glutamicum cultivated in CGXII and R2 medium

Sample name	Raw reads	Error%a	Q20%b	Q30%c	Clean reads	Mapped ratio (%)	Unique mapped ratio (%)
CGXII-1	32,062,428	0.0248	98.05	94.34	16,031,214	91.13	80.38
CGXII-2	25,053,266	0.025	97.99	94.2	12,526,633	94.20	84.82
CGXII-3	28,303,666	0.0255	97.74	93.69	14,151,833	90.41	75.13
R2-1	27,231,824	0.0251	97.92	94.07	13,615,912	81.78	69.75
R2-2	30,649,666	0.025	97.98	94.16	15,324,833	95.94	88.38
R2-3	25,540,028	0.0256	97.71	93.61	12,770,014	92.08	72.36

^aError: Average sequencing error rate, calculated through Qphred = − 10log10(e)

^bQ20: Proportion of bases with Qphred > 20 (Qphred = − 10log10(e))

^cQ30: Proportion of bases with Qphred > 30 (Qphred = − 10log10(e))

Fig. 4.

Overall profile of the sequenced data. A Principal component analysis. The x-axis and y-axis represent the new data set corresponding to the principal components (PCs) after dimensionality reduction, which was used to represent the difference between samples; the value in the coordinate axis label represents the percentage of variance of the corresponding PC interpretation of the population. B The volcano plot for DEGs between the two groups. Black dots show genes with no significant differences; red dots are significantly up-regulated genes; blue dots are significantly down-regulated genes

To understand the functions of the DEGs associated with medium change, sequences were searched against the (Gene ontology) GO database for enrichment analysis. This annotation resulted in three major categories: biological processes, cellular components, and molecular functions (Fig. 5). Under biological process, the most enriched GO term was ‘carbohydrate derivative metabolic process’ with 20 DEGs, followed by ‘cation transport’ and ‘polysaccharide metabolic process’. Sulfur metabolism (‘hydrogen sulfide metabolic process’, ‘cysteine metabolic process’) and nitrogen metabolism GOs (‘glutamine metabolic process’, ‘cysteine metabolic process’, ‘nitrate metabolic process’) were also enriched. Under cellular components, ‘proton-transporting ATP synthase complex’ was most enriched, which suggested an overall change of energy profile. Term ‘alpha-ketoglutarate dehydrogenase complex’ was also enriched, which was involved in metabolic flux distribution at the α-ketoglutarate node and affected glutamate accumulation. Under molecular functions, the most enriched GO term was ‘adenyl nucleotide binding’. This indicated different expression of ATP-required enzymes among the two groups. Terms involved in heme metabolism (‘heme binding’, ‘tetrapyrrole binding’) were also enriched. Overall, the new medium may have promoted cell growth through alteration of carbon metabolism, cation transport and energy synthesis.

Fig. 5.

Gene ontology (GO) enriched terms of differentially expressed genes (DEGs). The x-axis is the number of DEGs involved in each term divided by the number of total DEGs. The y-axis lists the sub-Go terms under categories of biological process, cellular component, and molecular function

Discussion

Corynebacterium glutamicum has been widely used for bulk and fine chemicals fermentation from various substrates in a green and sustainable way these years. CGXII is one of the most frequently used minimal media for this organism since reported in 1993. However, effects of components in this medium on cell cultivation have not been studied particularly till now. In this study, we developed an efficient defined medium for C. glutamicum based on the CGXII minimal medium through single-factor experiment and central composite design (CCD).

Corynebacterium glutamicum is known to tolerate high concentration of glucose. In batch fermentation, the concentration of glucose was generally high, reaching 60–140 g/L (Chen et al. 2009; Khan et al. 2005). However, for most laboratory applications (e.g., molecular operation and seed culture preparation), high amounts of glucose is not necessary. In this study, we found that 20 g/L glucose was sufficient for routine cultivation purpose in the CGXII medium.

As an organic nitrogen source, urea was demonstrated to be more preferred than the inorganic nitrogen source (NH₄)₂SO₄ here. Similar preference was also reported in several other studies when Scenedesmus bijugatus, Rhodotorula glutinis or Yarrowia lipolytica was cultivated (Arumugam et al. 2013; Brabender et al. 2018; Karakaya et al. 2012). When an insufficient amount of urea (5 g/L) was added, the impact of (NH₄)₂SO₄ was prominent. However, while the concentration of urea surpassed a threshold value (13.64 g/L), the impact of (NH₄)₂SO₄ was non-significant. This was the reason for the inconsistence in single-factor experiment and the CCD. Since urea is cheaper than ammonium sulfate on an equivalent nitrogen basis, and more readily available, a significant cost benefit may be realized. Considering only the prices of the three most variable components (glucose, CNY3900/t; (NH₄)₂SO₄, CNY1050/t; urea, CNY2700/t; based on the Chinese commodities market), the cost of medium would decrease by 60.95% when accumulating the same biomass using the R2 medium.

As the phosphate source, K₂HPO₄ had a positive effect on cell growth, while KH₂PO₄ did not show a marked impact. We speculate that maybe it was the K⁺ ion that affected cell growth actually as adjustment of these two components did not result in an obvious pH perturbation (data not shown). The magnesium ions are known to affect various enzymatic reactions and to have an impact on the stability of cell structure in bacteria (Groisman et al. 2013). Under Mg²⁺ limiting conditions, mutation occurred in corA encoding a Mg²⁺ importer to maintain cell growth through adaptive evolution (Pfeifer et al. 2017). In this study, 1.54 mM Mg²⁺ was proved to be enough and excessive amounts did not have further effects.

PCA is naturally excreted by the siderophore bacterium B. thuringiensis or B. anthracis strain Sterne under iron-restricted conditions (Williams et al. 2012). PCA (or other similar iron chelators) has been designated necessary to initiate cell division in C. glutamicum and supposed to accelerate growth of C. glutamicum when used as a co-substrate with glucose (Liebl et al. 1989; Unthan et al. 2014). However, elimination of PCA did not result in a growth retardation in our study. This discrepancy may be due to the different seed media used (BMCG medium in that study). Nutrients in the inoculum were able to compensate the PCA demand in minimal medium.

The effects of micro-metal elements (FeSO₄·7H₂O, MnSO₄·H₂O, ZnSO₄·7H₂O, CuSO₄ and NiCl₂·6H₂O) on cell cultivation were not assessed here. In some CGXII formulation, Na₂MoO₄·2H₂O and H₃BO₃ were also included (Weuster-Botz et al. 1997). These nutrients were used in trace amounts that hardly influenced the substrate cost but influenced protein production obviously. Since enzyme activity depends largely on metal ions, these nutrients should be determined each time (Savizi et al. 2019).

For plasmid amplification, as one can speculate, the amounts of plasmids produced would be positively related to cell biomass under antibiotic selection. The biomass accumulation tendency in CGXII, R2 and BHI was very similar to that displayed in Fig. 4 and not shown again. For protein production, fluorescence intensity of BHI culture was the strongest at 12 h while it was the weakest at 24 h. We suppose the rapid accumulation of GFP in this medium can be attributed to the rich complex nutrients existed, which can be readily used for protein synthesis. However, as cell biomass did not increase more after about 12 h (refer to Fig. 4), protein synthesis was also retarded. So, the fluorescence intensity at 24 h was about the same as that of 12 h.

In transcriptomics analysis, GO enrichment was performed to understand the functions of the DEGs associated with medium change. Most enriched terms could be classified as carbon metabolism, cation transport and energy synthesis. Altered carbon metabolism shaped the broth products profile (Fig. 3B), stimulated energy synthesis and changed the overall redox state. Cation transport could be influenced by expression of related transporters and the structure of cell membrane. Decrease in biotin supply was known to increase the permeability, which may affect cation transport as a result (Gutmann et al. 1992). By changing the co-nitrogen of urea and (NH₄)₂SO₄ to urea solely, genes related to urea transport and decomposition (e.g., the ure and urt operons) were supposed to be up-regulated (Beckers et al. 2004). However, expression of these genes was not significantly changed. Yet, other nitrogen metabolism related GOs (‘glutamine metabolic process’, ‘cysteine metabolic process’, ‘nitrate metabolic process’) were enriched. Perhaps, induction of related genes was already complete in CGXII, as only low concentrations make the synthesis of special carrier proteins necessary (Bendt et al. 2004).

Conclusion

In this study, an efficient defined medium R2 for C. glutamicum was developed based on the CGXII minimal medium. The optimized medium used urea as nitrogen source and showed preferable capacity in terms of cell cultivation, plasmid amplification and protein production than the original recipe. RNA-seq revealed that altered carbon metabolism, cation transport and energy synthesis may have benefited cell growth. Considering the universal availability and affordability of urea, the new medium can serve as useful reference for large-scale cultivation. To lower substrate cost further, MOPS could be replaced by CaCO₃ in future studies.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

PY and YNC performed all the laboratory experiments and drafted the paper. ADG designed the project, coordinated it, wrote and revised the manuscript. All authors read and approved the final manuscript.

Declarations

Conflict of interest statement

No conflict of interest declared.