Introduction

Bacteria often suffer from diverse external stresses both in natural and industrial environments, such as variations in pH, salinity, temperature, and osmolality (Beales 2004; Marles-Wright and Lewis 2007). Among them, low-pH fluctuations are the most frequent environmental stresses when industrial microorganisms produce acidic bio-based chemicals by fermentation (Liu et al. 2015). Acid stress usually causes a series of physiological damages in the bacterial cells; destroys the structure of biological macromolecules such as nucleic acids, proteins, and lipids; and affects the activities of enzymes involved in major metabolic pathways, thereby leading to the disturbances of intracellular substance/energy metabolism. In order to survive in an unfavorable low-pH environment, bacteria gradually develop a variety of coping strategies during the natural evolution (Kanjee and Houry 2013; Krulwich et al. 2011; Lund et al. 2014; Papadimitriou et al. 2016), such as pumping protons (H+) out of the cell, producing alkaline substances, synthesizing compatible solutes, altering intracellular redox homeostasis, changing the fatty acid composition of cell membrane, and reconstructing central metabolic pathways.

In the course of resisting acid stress, self-regulation of intracellular pH levels plays a crucial role in cell survival of microorganisms (Kobayashi et al. 2000). In neutrophilic bacteria, the intracellular pH still maintains a near-neutral state even if the external pH fluctuates within a certain range (Lund et al. 2014). Previous studies have shown that most bacteria employ multiple adaptation strategies to maintain intracellular pH homeostasis, typically by consuming excess protons or producing ammonia through several pathways, such as the glutamate decarboxylase (GAD) system, the arginine deiminase (ADI) system, urea hydrolysis, and F1F0-ATPase (Foster 2004; Kakinuma 1998; Kanjee and Houry 2013; Marles-Wright and Lewis 2007; Wu et al. 2014). In most bacteria, for example Escherichia coli, Lactobacillus, and Listeria, the GAD system plays a dominant role in resisting acid stresses, which converts a molecule of glutamate to gamma-aminobutyric acid (GABA) with the removal of an intracellular proton (Lund et al. 2014; Xu et al. 2017).

Corynebacterium glutamicum is generally regarded as a safe microorganism and is widely used for the large-scale fermentation of l-glutamate and various organic acids (Lee et al. 2016; Wieschalka et al. 2013). Interestingly, whole-genome analysis reveals that C. glutamicum has no genes with apparent homology to the described GAD and ADI systems (Heydari et al. 2014). In recent years, many studies have revealed that some physiological and biochemical processes are implicated in the defense mechanisms against low pH in C. glutamicum (Follmann et al. 2009; Liu et al. 2016; Wang et al. 2016). For example, the comprehensive analysis of pH homeostasis in C. glutamicum reveals a functional link among pH response, oxidative stress, iron homeostasis, and metabolic alternations. Mycothiol is a major thiol in most actinomycetes, and it plays an important role in protecting C. glutamicum against acid stress by scavenging intracellular reactive oxygen species (ROS). However, the molecular basis of low-pH adaptation in C. glutamicum still remains to be described in detail. In this study, an adapted C. glutamicum strain showing higher growth capacity at pH 5.6 was obtained by the adaptive laboratory evolution method. According to transcriptome analysis and functional testing, we found that multiple genes implicated in oxidative stress and sulfur assimilation were required for optimal protection against acid stress, providing new candidate functional modules for strain improvement in C. glutamicum.

Materials and methods

Strains and growth conditions

Bacterial strains and plasmids used in this study are listed in Table 1. E. coli DH5α or HST02 was used as host cells for general cloning. C. glutamicum ATCC 13032 was used as the parental strain for gene disruption and the wild-type strain for functional analysis. E. coli cells were routinely grown at 37 °C in LB medium (0.5% yeast extract, 1% tryptone, 1% NaCl), and C. glutamicum cells were cultivated at 32 °C in modified A medium (designated as the MAM medium) (7 g L−1 casamino acids, 5 g L−1 (NH4)2SO4, 2 g L−1 yeast extract, 2 g L−1 urea, 0.5 g L−1 KH2PO4, 0.5 g L−1 K2HPO4•3H2O, 0.5 g L−1 MgSO4•7H2O, 6 mg L−1 Fe2SO4•7H2O, 4.2 mg L−1 Mn2SO4•H2O, 0.2 mg L−1 biotin, 0.2 mg L−1 thiamine, 4% glucose) unless specified otherwise. When necessary, the medium was buffered with 200 mM MOPS (3-[N-morpholino]-propanesulfonic acid, pKa = 7.20), 200 mM MES (2-[N-morpholino]-ethanesulfonic acid, pKa = 5.96), or 100 mM homo-PIPES (homopiperazine-1,4-bis[2-ethanesulfonic acid], pKa = 4.55) to different desired pH values, respectively. The pH of each medium was normally adjusted as necessary using either HCl or KOH. If needed, antibiotics were added to a final concentration of 100 μg mL−1 ampicillin, 25 μg mL−1 kanamycin, or 5 μg mL−1 chloramphenicol, and 10 μM isopropyl β-d-1-thiogalactopyranoside (IPTG) was used for the induction of gene expression.

Table 1 Plasmids and strains used in this study
Full size table

Strain construction

The deletion mutants were achieved by a two-step homologous recombination method based on the temperature-sensitive plasmid pCRD206 as described previously (Okibe et al. 2011). The primers used in this study are listed in Supplemental Table S1. To avoid polar effects, markerless chromosomal in-frame deletion mutants were constructed. The ∆katA mutant was obtained as follows: the katA upstream and downstream regions were amplified with Phusion High-Fidelity DNA Polymerase using the primer pairs katA-1-For/katA-2-Rev and katA-3-For/katA-4-Rev, respectively. These two products were then fused by overlap extension polymerase chain reaction (OE-PCR) with the primer pairs katA-1-For/katA-4-Rev, ligated into the BamHI/XbaI sites of pCRD206, and directly used to transform E. coli HST02 to yield the pCRD206-∆katA plasmid. The resulting plasmid was then used to transform C. glutamicum ATCC 13032 by the electroporation method. The ∆katA mutant was obtained through a two-step selection strategy and confirmed by colony-PCR. Similar strategies were used to generate other single-deletion mutants. Moreover, the pCRD206-∆dps plasmid was used to transform C. glutamicum ∆katA mutant to generate the ∆katA ∆dps double mutant.

The detailed procedures for the construction of pXMJ19- and pXMJXU-derived recombinant plasmids were described in the Supplemental Methods. The E. coli-C. glutamicum shuttle vector pXMJXU, which was obtained by removing the lacIq region from the pXMJ19 vector (Jakoby et al. 1999), was used as the vector backbone for the improvement of acid-stress tolerance.

Adaptive laboratory evolution

The adaptive evolution experiment was performed according to the protocols described in the previous literature (Lee et al. 2013). C. glutamicum ATCC 13032 was used as the original strain for experimental evolution. The strain was grown aerobically at 32 °C in the 100-mL shake flask containing 20 mL MAM medium, and growth was monitored by measuring the optical density at 600 nm. The MAM medium was buffered with 200 MES to the desired pH values. The medium was sterilized by filtration through a 0.22-μm filter and prepared in sufficient amounts to minimize the possible influences of medium components and pH value. For the adaptive evolution experiment, the cultures were serially passaged at pH 6.0 for the first 25 days and at pH 5.8 for the subsequent 70 days. Every 24 h, the evolving cells were transferred into the indicated fresh medium to yield an initial optical density (OD600) of 0.1. After finishing the evolution experiment, aliquots of the samples were serially diluted and spread on the MAM plates with a pH of 5.8 to select a single colony of the acid-adapted strain. Dozens of colonies growing faster than the original strain were picked up from the plate and subjected to further evaluation in a liquid medium with an acidic pH of 5.8 (data not shown). Three individual colonies with clearly improved growth were chosen as representatives to test the growth ability of the adapted strain under different acid-stress conditions.

The genetic stability of the adapted strain was also examined after cultivation at normal pH for more than 100 generations. Briefly, the representative colonies of the adapted strain were inoculated into liquid MAM medium with a pH of 7.0 and incubated with shaking at 32 °C. Each cultivation process was started from an initial OD600 of 0.1 and was finished with a final OD600 of 3.0, resulting in approximately 5 generations. After the adapted strain was transferred a total of 20 times (~ 100 generations), the growth ability of the adapted strain under acid-stress conditions was tested. The growth assay suggested that the adapted strain was genetically stable for many generations (Supplemental Fig. S1).

Transcriptome analysis by RNA-seq

A single colony of C. glutamicum original (Ori-Cg) or adapted (Ada-Cg) strains was picked from the freshly streaked agar plate and inoculated into the MAM medium. After overnight cultivation (12–16 h) at 32 °C, the cells were respectively harvested, washed, and shifted to the standard CGXII minimal medium with an initial OD600 of 0.1 (Keilhauer et al. 1993). The medium was buffered to pH 5.6 with 200 mM MOPS, and every colony was cultivated in duplicate. The cultures were incubated for a further 12 h at 32 °C before RNA extraction. Total RNA of C. glutamicum cells was extracted using the RNAprep Pure Cell/Bacteria Kit (Tiangen Biotech, Beijing, China) and treated with DNase I to eliminate contaminating genomic DNA from the total RNA samples. Library construction and sequencing were performed at Novogene Bioinformatics Technology Co., Ltd. (Beijing, China). The detailed methods for transcriptome analysis are shown in the Supplemental Methods.

Quantitative reverse transcription PCR

Reverse transcription was performed by the RevertAid first-strand cDNA synthesis kit (Thermo Scientific, Waltham, MA, USA) according to the manufacturer’s protocol. For the relative qPCR analysis, the reaction mixtures were prepared using SYBR Green Realtime PCR Master Mix (Toyobo Co., Osaka, Japan). The primers used in this procedure are listed in Supplemental Table S2. The experiment was performed in triplicate and repeated three times using an Applied Biosystems 7500 fast real-time PCR system (Thermo Scientific, Waltham, MA, USA). The amplification was performed by an initial denaturation at 95 °C for 1 min, then 40 cycles of 95 °C for 15 s, 60 °C for 15 s, and 72 °C for 40 s, and a final gradient at 55–95 °C for the dissociation curve. The relative gene expression fold changes were calculated according to the delta-delta threshold cycle method after normalization using 16S rRNA as the internal control (Xu et al. 2018).

Cell growth assay

A single colony from a freshly streaked agar plate was transferred to the liquid MAM medium and incubated for 16 h at 32 °C with shaking at 200 rpm prior to growth experiments. The cultures were harvested, washed, and re-suspended in the fresh indicated medium to an initial OD600 of 0.1. Cell growth was then monitored by measuring the optical density at 600 nm at the indicated time points and shown as averages of at least three independent repeats.

Cell survival rate assay

An acid survival assay was carried out according to the previously described method with minor modifications (Zhang et al. 2007). Overnight cultures of C. glutamicum were grown to the exponential phase in MAM medium at 32 °C. For strains with IPTG-inducible constructs, 10 μM IPTG was added to the culture medium just 2 h prior to harvesting. The cells were washed with sterile 0.9% NaCl solution and re-suspended in fresh CGXII or 0.9% NaCl solution to an initial OD600 of 0.2. These solutions were buffered to pH 4.0 with 100 mM homo-PIPES, unless specified otherwise. The resulting cell suspensions were incubated at 32 °C with shaking of 100 rpm for the indicated time periods. After the low-pH treatment, ten-fold serial dilutions of each sample were spotted onto LBG agar plates (LB medium with 2% glucose) and incubated for 2~3 days before being photographed. Additionally, a 100 μL sample of the 1:10,000 dilution was spread evenly over a fresh LBG plate, and the colonies were counted after 2~3 days of incubation at 32 °C. The survival rates were calculated as the number of colony-forming units (CFUs) after acid challenge divided by the number of CFUs without acid challenge.

Transmission electron microscopy analysis

Cell morphology was examined by TEM analysis as described previously (Wu et al. 2012). Overnight cultures of C. glutamicum were pre-grown to the exponential phase in fresh MAM medium. The cells were washed with sterile 0.9% NaCl solution and re-suspended in fresh CGXII medium buffered to pH 3.8 with 100 mM homo-PIPES. After 2 h of incubation at low pH, the cells were harvested by centrifugation, mixed with 1% low-temp gelling agarose, and further cut into 1-mm3 dimensioned gel pieces. The samples were prepared according to the standard TEM procedure and observed using an HT7700 transmission electron microscope (Hitachi, Tokyo, Japan).

Measurement of cytoplasmic ROS and pH levels

Cytoplasmic ROS levels were investigated using the fluorescent probe 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) as described previously (Si et al. 2017). Cytoplasmic pH was determined using the pH-sensitive fluorescent probe 2′,7′-bis-(2-carboxyethyl)-5-6-carboxyfluorescein (BCECF) as described previously (Xu et al. 2018). The details of cytoplasmic ROS and pH measurements are shown in Supplemental Methods.

Statistical analysis

Student’s t test was used to investigate the significance of differences between two groups. Differences with P values of less than 0.01 were regarded as statistically significant and were indicated with two asterisks.

Results

Enhancement of acid tolerance in C. glutamicum by adaptive evolution

Adaptive laboratory evolution of C. glutamicum was performed in two successive stages, with the first 25 days at pH 6.0 and the subsequent 70 days at pH 5.8 (Fig. 1a). Growth assays under variable pH levels revealed that the final acid-adapted clones exhibited obviously enhanced growth compared to the original strain at pH values ranging from 5.2 to 6.0 (Fig. 1b). Growth curve analysis suggested that there was no apparent difference in growth between the original and acid-adapted strains at neutral pH (Supplemental Fig. S2). However, the adapted strain exhibited clearly accelerated growth compared to the original strain when exposed to a more acidic environment of pH 5.6.

Fig. 1
figure 1

Adaptive laboratory evolution (ALE) of C. glutamicum to improve its acid tolerance. a The workflow of the ALE experiment. C. glutamicum ATCC 13032 was used as the original strain for experimental evolution. The samples were transferred every 24 h to an initial OD600 of 0.1. b Growth differences between the original (Ori-Cg) and adapted (Ada-Cg) strains under different pH conditions. Cell growth was determined by measuring the OD600 after 20 h of incubation in a shaker at 32 °C. c Growth curves of the original and adapted strains in liquid MAM medium at pH 5.6. The data are presented as the means ± standard deviations (SD) from at least three independent experiments. The asterisk indicates a significant correlation between the original and adapted strains at **P ≤ 0.01

Full size image

The adapted strain showed the improvement of physiological characteristics under acid-stress conditions

In order to further investigate the tolerance of the original and adapted strains to low-pH stress, acid survival assays were examined using ten-fold serial dilutions and CFU counting. As shown in Fig. 2a, with the extension of the incubation period, the survival rates of both strains gradually declined to varying degrees. After 2 h of exposure to the acidic conditions, approximately 39% of the adapted strains remained viable, while the original strain had nearly zero capacity to survive such stress.

Fig. 2
figure 2

Differences in physiological properties between the original and adapted strains. a Cell survival assays of the two strains after low-pH challenge. After pH 3.8 treatment at the indicated time points, ten-fold serial dilutions were spotted onto LBG agar plates, and colonies were counted after 2~3 days of incubation. b Intracellular pH values of the original and adapted strains under various external pH conditions. c Cytoplasmic ROS levels of the original and adapted strains at neutral pH 7.0 and acidic pH 4.0. d Cell morphology before and after acid challenge was investigated using the transmission electron microscopy. Blue arrows indicate the thickness of the cell wall, and red arrows show possible sites of cell breakage. A 3-times enlarged view of the box was embedded into each image. The asterisk indicates a significant correlation between the original and adapted strains at **P ≤ 0.01

Full size image

To investigate the potential mechanisms underlying these growth differences, the intracellular pH (pHin) and ROS levels were measured by the corresponding fluorescent probes. As shown in Fig. 2b, the adapted strain exhibited higher pHin levels compared to those of the original strain under the lower external pH conditions. For example, the adapted strain maintained a stable pHin of 6.4 against an external pH of 5.5. This value was approximately 0.7 pH units higher than that of the original strain. No obvious difference of pHin was observed between these two strains in a medium with a neutral pH. As expected, Fig. 2c showed that acid stress clearly increased intracellular ROS generation. Under acid-stress conditions, the fluorescence intensity of the adapted strain was approximately 61% of the wild-type level, implying that the adapted strain developed a capacity to reduce the accumulation of intracellular ROS.

The cell wall and membrane of bacteria usually provide the first protective barrier against external stimuli (Silhavy et al. 2010). We therefore used transmission electron microscopy to examine the morphological changes of the original and adapted strains during normal conditions and following acid challenge (Fig. 2d). Under normal pH conditions, both the original and adapted strains had a capacity to maintain the integrity of cell structure, exhibiting a very similar cell wall thickness. By contrast, after 2 h of acid challenge, the cell walls of both strains became thin and their edges were rough. For the original strain, the average thickness of the cell wall was approximately 22.9 nm, and the cells displayed multiple damage sites. However, the cell wall structure of the adapted strain remained relatively intact and smooth, with an average thickness 10.9% greater than that of the original strain. The loss of cell integrity typically led to massive spillage of intracellular materials and severe cytoplasmic vacuolization. In addition, TEM analysis demonstrated that more than 80% of the adapted cells showed elongated phenotype and septum formation, indicating that the adapted strain indeed had enhanced acid tolerance.

Comparative transcriptome analysis revealed candidate genes and key pathways associated with the acid-tolerance phenotype

To further explore the potential molecular mechanisms underlying the improved growth performance of the adapted strain, differential gene expression of the adapted strain compared to the original control under acid-stress conditions was determined by transcriptome analysis. The expression patterns of 212 genes were found to be clearly altered at low pH compared with the original strain, including 108 upregulated and 104 downregulated genes (Fig. 3a). Based on KEGG pathway enrichment analysis, these genes were classified into 19 major cellular processes (Fig. 3b and Table 2). Many genes associated with the secondary metabolite biosynthesis, amino acid biosynthesis, carbon metabolism, tricarboxylic acid (TCA) cycle, oxidative phosphorylation, and microbial metabolism diversity were upregulated in the adapted strain. Furthermore, multiple important processes, including sulfur metabolism, biosynthesis of sulfur-containing amino acids, and ATP-binding cassette (ABC) transporters were downregulated in the adapted strain.

Fig. 3
figure 3

Transcriptome analysis of the original and adapted strains under acid-stress conditions. a Volcano plot of differentially expressed genes. The genes with significant differences were indicated by red (upregulation) and green dots (downregulation). The X-axis represents the log2-transformed value of the expression fold changes, and the Y-axis indicates the log10-transformed value of statistically significant differences in expression changes. b Expression profiles of genes belonging to different cellular processes according to KEGG analysis. The abscissa represents the number of genes enriched for the indicated metabolic pathways. c Relative expression levels of 16 representative genes in the adapted strain compared with those in the original strain identified by RNA-seq and qRT-PCR analyses. The data are presented as the means ± SD from three independent experiments

Full size image
Table 2 Differential gene expression analysis of RNA-seq data between the Evo-Cg and Ori-Cg strain
Full size table

Sixteen representative genes were chosen to validate the RNA-seq data by qRT-PCR analysis. As shown in Fig. 3c, the results were in good agreement with the RNA-seq analysis, despite a slight difference in fold changes. Three sulfur assimilation-related genes (cysD encoding a sulfate adenylyltransferase, ssuC encoding a sulfonate transport system permease protein, and cysK encoding a cysteine synthase) as well as two genes of ABC-type transporters (cg0589 and cg0768 encoding ATP components of Fe3+-siderophore transport system) were reduced in abundance by at least two-fold, supporting a hypothesis that the adapted strain may have impaired sulfur assimilation or external iron absorption processes. Consistent with this finding, the expression of the mcbR gene, encoding a TetR-type transcriptional inhibitor of sulfur metabolism, was approximately 2.6-fold higher in the adapted strain based on the qRT-PCR analysis. Among these upregulated genes, cg1328 and cg3292 (encoding copper chaperones) as well as cg3290 (encoding an oxidoreductase protein) were implicated in copper metabolism and trafficking (Harrison et al. 1999), while hscA (encoding a molecular chaperone of the Hsp70/DnaK family) was predicted to maintain cellular protein homeostasis (Genevaux et al. 2007). Moreover, the cydA gene (encoding a subunit of cytochrome bd respiratory oxygen reductase) was also upregulated. The enhanced expression of the respiratory bd complex contributes to the extrusion of excess intracellular protons (Giuffre et al. 2014), which was consistent with the higher pHin level displayed by the adapted strain under acid-stress conditions (Fig. 2b). Furthermore, the adapted strain displayed increased expression of ftsE (encoding a cell division ATP-binding protein) and fas (encoding a fatty acid synthase), which was in agreement with the results of TEM analysis that the adapted strain exhibited improved maintenance of cell division and membrane integrity under acid-stress condition (Fig. 2d) (Letek et al. 2008; Radmacher et al. 2005). Thus, the altered expression of these genes in the adapted strain may also be a non-negligible factor in improving growth performance.

Scavengers of intracellular ROS were required for the tolerance of C. glutamicum to low-pH stress

Previous studies have revealed that acid stress can lead to the formation of intracellular ROS in many microorganisms, including C. glutamicum (Chung et al. 2006; Liu et al. 2016; Mols et al. 2010). As the dominant component of ROS, hydrogen peroxide (H2O2) is spontaneously converted to the highly reactive hydroxyl radical (OH•) via the Fenton reaction catalyzed by free Fe2+ (Winterbourn 1995). These ROS radicals are highly reactive molecules and typically affect cell survival by damaging nucleic acids, lipids, and proteins. The RNA-seq and qRT-PCR data revealed the upregulation of genes associated with intracellular ROS scavenging, such as katA encoding a catalase and dps encoding a stress-inducible DNA-binding protein (Fig. 3c). Previous studies have demonstrated that Dps can protect DNA from oxidative damage by sequestering intracellular free Fe2+ and storing it in the form of an Fe oxyhydroxide mineral (Bellapadrona et al. 2010). Moreover, the expression of the ftn gene, encoding a major intracellular iron-storage protein, was obviously upregulated (Rivera 2017). The genes encoding the external Fe3+-siderophore transport system, including cg0589-0591, cg0768-0771, and cg0924-0928, were also inhibited in the adapted strain (Table 2). We therefore speculated that the low level of intracellular free Fe2+, largely maintained by both limited external iron uptake and enhanced intracellular iron storage, might be conducive to protecting cells against iron-mediated oxidative stress.

In this study, we selected katA, dps, and ftn as target genes to further investigate their physiological roles in acid tolerance. As shown in Fig. 4a, deletion of katA or dps significantly reduced the survival rates after low-pH challenge, whereas loss of ftn had no apparent effect on the survival of C. glutamicum under these conditions. As expected, the overexpression of katA or dps increased the percentage of viable cells to varying degrees. The double deletion of katA and dps further exacerbated the survival defect compared to the single-gene deletion, and the ∆katA ∆dps double mutant showed only 16.6 ± 3.4% survival after low-level acid challenge. Simultaneous overexpression of katA and dps increased the survival rate to 44.5 ± 3.7%, which was approximately 1.4-fold higher than that of the wild-type strain. The measurement of intracellular ROS levels revealed that they were closely related to the cell survival of different strains (Fig. 4b). The ∆katA, ∆dps, and ∆katA ∆dps strains exhibited higher levels of intracellular ROS under acid-stress conditions, which increased by 2.0-, 1.3-, and 2.5-fold compared with the wild-type strain. In accordance with the expectation, overexpression of both katA and dps led to an obvious reduction of intracellular ROS levels. Taken together, our findings suggest that ROS scavenging mediated by katA and dps plays an important role in conferring C. glutamicum cells with resistance to low-pH stress.

Fig. 4
figure 4

Effects of katA, dps, and ftn on cell survival and intracellular ROS accumulation. a Cell survival assays of the wild-type and mutant strains after low-pH challenge. b Cytoplasmic ROS levels of the wild-type and mutant strains after low-pH challenge. The data are presented as the means ± SD from three independent experiments. The asterisk indicates a significant correlation between the wild-type and mutant strains at *P ≤ 0.05 and **P ≤ 0.01

Full size image

The repression of sulfur assimilation pathway using the McbR regulator showed beneficial effects on acid-stress tolerance

The RNA-Seq data revealed that the expression of sulfur assimilation-related genes under acid-stress conditions was greatly reduced in the adapted strain compared with the original strain (Fig. 5a, Table 1). Further qRT-PCR analysis suggested that the absence of the mcbR gene, encoding a master repressor of sulfur metabolism in C. glutamicum (Rey et al. 2003), obviously increased the expression of sulfur assimilation-related genes (Supplemental Fig. S3). In addition, the expression of the mcbR gene in the original and adapted strains was also examined by qRT-PCR analysis. As shown in Fig. 5b, there was no obvious difference in transcript levels between the two strains at neutral pH conditions. However, the adapted strain exhibited clearly higher expression levels of mcbR gene compared to the original strain when challenged by low-level acid stress. Furthermore, we performed growth experiments to explore the physiological role of mcbR under different pH conditions. Deletion or overexpression of mcbR had minor effects on cell growth at neutral pH with or without the addition of cysteine (Supplemental Fig. S4). However, deletion of mcbR led to a statistically significant decrease in cell growth under low-pH stress, and overexpression of mcbR improved the tolerance of C. glutamicum to low pH (Fig. 5c). Interestingly, the presence of cysteine had a largely negative effect on cell growth of all strains at low-pH values. In the mcbR-overexpression strain, the addition of cysteine greatly decreased cell growth, supporting the opinion that the accumulation of cysteine or its metabolites is detrimental to cell survival under acid-stress condition. We therefore speculated that the observed increased expression of mcbR in the adapted strain repressed sulfur assimilation and cysteine biosynthesis, thereby alleviating the growth defects caused by cysteine accumulation under acid-stress conditions.

Fig. 5
figure 5

Roles of the McbR regulator in the response to acid stress. a Schematic diagram of the sulfur assimilation pathway in C. glutamicum. b The mRNA levels of mcbR between the original and adapted strains under neutral pH 7.0 and acidic pH 5.6 conditions. c Growth differences among the wild-type, ∆mcbR deletion, and mcbR-overexpression mutant strains under acidic pH conditions with or without 5 mM cysteine. Cell growth was determined by measuring the OD600 after 20 h of incubation

Full size image

Genetic engineering of C. glutamicum to improve acid resistance

Based on the described experimental data, katA, dps, and mcbR appeared to be promising genetic modules for strain improvement. Furthermore, we found that two copper chaperone genes, cg1328 and cg3292, were also involved in the promotion of cell survival under acid-stress conditions (Supplemental Fig. S5), which was in agreement with the interplay between acid stress and copper toxicity reported in some bacteria (Djoko et al. 2017; Palumaa 2013). Therefore, we selected four target genes (katA, dps, mcbR, and cg1328) as representatives for the construction of strains with enhanced acid tolerance. The functional module designed to confer acid tolerance was overexpressed under the control of the native promoter of C. glutamicum superoxide dismutase (sod) or glyceraldehyde-3-phosphate dehydrogenase (gap). At neutral pH, the overexpression of these potential acid-tolerance modules had slightly negative effects on cell growth during the initial 24 h of incubation, and the engineered strains ultimately showed similar growth phenotypes with the wild-type control (Supplemental Fig. S6). Under acidic conditions, these engineered strains also showed reduced cell growth during the initial incubation period, but their growth rates gradually increased with the extension of incubation (Fig. 6). As expected, the recombinant strain with a combined expression of four target genes under the control of the gap promoter reached the highest biomass yield, which was 18.4 ± 2.5% higher than that of the wild-type strain. Overall, this study provides several candidate genetic modules for the improvement of acid tolerance in C. glutamicum.

Fig. 6
figure 6

Growth assays of the wild-type and genetically engineered strains under acid-stress condition. The overnight cultures of the bacterial strains were inoculated into fresh MAM medium with an initial OD600 of 0.1. Cell growth was monitored by measuring the optical density at 600 nm at the indicated time points and shown as the averages of three independent repeats. The asterisk indicates a significant correlation between the wild-type and mutant strains at **P ≤ 0.01

Full size image

Discussion

Acid stress has previously been reported to be associated with oxidative stress in many microorganisms, including Bacillus cereus, E. coli, and C. glutamicum (Lund et al. 2014; Mols et al. 2010; Wang et al. 2016). Here, we found that the adapted strain exhibited reduced intracellular ROS formation under acid-stress condition, and overexpression of both katA and dps obviously reduced intracellular ROS accumulation and eventually elevated cell survival at acidic pH condition. Previous studies have reported that several regulators, including OxyR, RipA, and DtxR, are involved in the regulation of hydrogen peroxide-inducible antioxidant genes, (Follmann et al. 2009; Heydari et al. 2014). Studies performed in C. glutamicum revealed that the LysR-type regulator OxyR acts as a transcriptional repressor of katA, dps, ftn, and cydA (encoding a subunit of cytochrome bd oxidase) (Teramoto et al. 2013). The AraC-type regulator RipA also functions as a transcriptional repressor of katA, acn (encoding an aconitase), and other genes encoding iron-containing proteins, which is itself repressed by the global iron repressor DtxR under iron excess (Wennerhold et al. 2005). The DtxR protein acts as a dual transcriptional regulator to control iron metabolism in C. glutamicum and participates in the activation of ftn and dps (Brune et al. 2006). Interestingly, although our results indicate that the mRNA levels of katA, dps, ftn, cydA, and acn genes were greatly increased in the adapted strain compared to the original strain under acid-stress conditions, the gene expression of oxyR, ripA, and dtxR showed no obvious changes. One possible explanation for this finding is that these three regulatory genes might be affected at the genomic or protein levels rather than at the transcriptional level. The combined analysis of multi-omics data in future research may contribute to a better understanding of the mechanisms by which oxyR, ripA, and dtxR participate in the incremental expression of most identified antioxidant genes.

Sulfur is an essential macronutrient that is required for the growth of all known living organisms (Rausch and Wachter 2005). However, the accumulation of certain sulfur-containing intermediates, such as cysteine, can disrupt intracellular thiol homeostasis and cause oxidative damage by driving the Fenton reaction (Park and Imlay 2003). The transcriptional repressor McbR has been found to be a master regulator of sulfur assimilation and the biosynthesis of sulfur-containing amino acids in C. glutamicum (Rey et al. 2003). In this study, we also found that the expression levels of mcbR and sulfur assimilation-related genes were clearly altered in the adapted strain compared to those of the original strain under acid-stress condition. The impairment of the sulfur assimilation pathway exerted beneficial effects on acid tolerance of the adapted strain. In addition, multiple regulators were also reported to control the expression of genes related to sulfur metabolism, such as OxyR and CysR (Milse et al. 2014; Ruckert et al. 2008). Similarly, the expression of these regulatory genes in the adapted strain was not affected at the mRNA level.

Our transcriptome data also revealed that multiple cellular processes, such as iron transport and storage, fatty acid synthesis, copper metabolism, cytochrome bd reaction, and cell division regulation, may also play important roles in acid tolerance. The repression of Fe3+-siderophore transport system aids in the maintenance of a low intracellular free iron content and thereby alleviates the production of hydroxyl radicals via the Fenton reaction catalyzed by free Fe2+ (Andrews et al. 2003; Follmann et al. 2009; Winterbourn 1995). The upregulation of the respiratory chain complex II (sdhA/B/C/D) and the cytochrome bd oxidase complex (cydA/B) can promote the extrusion of intracellular protons (Cecchini 2003; Giuffre et al. 2014). Moreover, the activation of fatty acid synthesis (fas) and cell division process (ftsE/X) might also be conducive to acid-stress tolerance. However, the detailed mechanisms underlying the roles of these cellular processes need to be deciphered in future research.