Role of Corynebacterium glutamicum sprA Encoding a Serine Protease in glxR-Mediated Global Gene Regulation

The global regulator glxR of Corynebacterium glutamicum is involved in many cellular activities. Considering its role, the GlxR protein likely interacts with other proteins to obtain, maintain, and control its activity. To isolate proteins interacting with GlxR, we used a two-hybrid system with GlxR as the bait. Subsequently, the partner, a subtilisin-like serine protease, was isolated from a C. glutamicum genomic library. Unlike glxR, which showed constitutive expression, the expression of sprA, encoding a serine protease, was maximal in the log phase. Purified His6-SprA protein underwent self-proteolysis and proteolyzed purified GlxR. The proteolytic action of SprA on GlxR was not observed in the presence of cyclic adenosine monophosphate, which modulates GlxR activity. The C. glutamicum sprA deletion mutant (ΔsprA) and sprA-overexpressing (P180-sprA) strains showed reduced growth. The activity of isocitrate dehydrogenase (a tricarboxylic acid cycle enzyme) in these strains decreased to 30–50% of that in the wild-type strain. In the P180-sprA strain, proteins involved in diverse cellular functions such as energy and carbon metabolism (NCgl2809), nitrogen metabolism (NCgl0049), methylation reactions (NCgl0719), and peptidoglycan biosynthesis (NCgl1267), as well as stress, starvation, and survival (NCgl0938) were affected and showed decreased transcription. Taken together, these data suggest that SprA, as a serine protease, performs a novel regulatory role not only in glxR-mediated gene expression but also in other areas of cell physiology. In addition, the tight control of SprA and GlxR availability may indicate their importance in global gene regulation.


Introduction
Corynebacterium glutamicum is a non-pathogenic, Gram-positive organism that belongs to the order Actinomycetales. Based on its metabolic capability, this organism has been widely used in the industrial production of nucleotides and amino acids [1]. Therefore, the catabolic and anabolic pathways associated with amino acid metabolism have attracted scientific attention and have been extensively analyzed. Moreover, the availability of genome sequence data [2][3][4] has significantly increased the understanding of gene expression and regulation on a global scale [5].
The global regulator glxR of C. glutamicum was initially identified through its capability to repress glyoxylate bypass genes [6]. The bypass, mediated by isocitrate lyase (ICL) and malate synthase (MS), is not required by cells grown with glucose as the sole carbon source, and therefore the genes mediating the bypass are repressed by glxR. However, when cells are provided with two-carbon compounds, such as acetate, the bypass plays a critical role because it conserves carbon by bypassing the CO 2 -generating steps of the tricarboxylic acid cycle. ICL (encoded by aceA) catalyzes the conversion of the tricarboxylic acid cycle intermediate isocitrate into glyoxylate and succinate, and MS (encoded by aceB) catalyzes the subsequent condensation of glyoxylate with acetyl-coenzyme A to produce malate. Furthermore, transcriptional regulators such as RamA and RamB are also involved in the regulation of aceA and aceB expression [7][8][9].
GlxR is a cyclic adenosine monophosphate (cAMP)-binding transcriptional regulator [10] that belongs to the cAMP receptor protein/fumarate and nitrate reduction regulator (CRP/FNR) family of proteins and shows homology to the global regulator CRP from Escherichia coli [6]. In addition to its role in acetate metabolism, glxR is involved in many other cellular activities, such as carbon metabolism [6,[11][12][13][14], energy metabolism [15], lipid metabolism [16], anaerobic nitrate respiration [17], inorganic phosphate uptake [18], and cell resuscitation [19]. Further, Kohl et al [20] analyzed the whole genome using in silico tools in combination with in vitro analysis and predicted that the GlxR regulon comprises many genes (96 genes in 53 transcription units) presumed to be involved in diverse cellular activities such as carbohydrate metabolism, aromatic compound degradation, glutamate uptake and nitrogen assimilation, fatty acid biosynthesis, stress response, and resuscitation. The global function of GlxR has recently been confirmed and expanded by Jungwirth et al [21] and Toyoda et al [22], who analyzed GlxR-bound target DNA sequences isolated from cells via immunoprecipitation. A total of 14% of C. glutamicum genes may be under the direct transcriptional control of GlxR [23].
Despite the importance of GlxR as a global regulator, many questions regarding its mechanism and the control of its action remain largely unanswered, partly owing to the essential role of glxR in cell physiology. For example, in earlier studies, glxR deletion mutants could not be isolated [6,13] or grew too slowly for study when isolated [24]. However, Park et al [25] were able to isolate a glxR deletion mutant; they demonstrated that the mutant strain had a severe growth defect and that glxR repressed glyoxylate bypass genes. Incidentally, GlxR functions as a dual regulator that can be a repressor or an activator depending on its target genes. The position of its consensus binding site TGTGA-N 2 -TA-N 2 -TCACA [20] relative to the transcriptional start site also supports its dual role. In all the cases studied, the in vitro binding of GlxR to its target binding sites depended on the presence of cAMP. Unlike the findings in E. coli, intracellular cAMP levels in C. glutamicum are reported to be elevated during growth on glucose and low during growth on acetate [6].
Although only a small degree of information is available about the regulation of global regulatory proteins, it is clear they must also be subject to a regulatory process. The autoregulation of transcription occurs for many regulatory genes [5]. Activity regulation with effector molecules is also commonly seen with many regulatory proteins [10]. In addition, proteolytic control of regulators to maintain their appropriate intracellular concentrations is also seen in the case of some proteins [26]. Considering the importance of global regulators, several regulatory mechanisms are thought to be present for these genes and proteins. This type of regulation may also occur for GlxR in order to ensure its optimum performance under the physiological conditions of the cell. GlxR is very likely to interact with other proteins to obtain, maintain, or control its activity. For this reason, we aimed to isolate proteins interacting with GlxR by implementing a two-hybrid screening system. We used GlxR as bait and screened the C. glutamicum genome. Herein, we report the identification of a subtilisin-like serine protease encoded by sprA as an interacting partner for GlxR. Given the physiological and biochemical data, we propose a role for sprA in glxR-mediated global gene expression.

DNA and RNA Analyses
Plasmids were introduced into C. glutamicum cells through electroporation [30]. Total RNA isolation for C. glutamicum was performed using the FastPrep24 system (MP Biomedicals). Polymerase chain reaction (PCR) was performed as previously described [6]. cDNA conversion and real-time quantitative PCR (RT-qPCR) were also performed as previously described [29]. CFX96 Real-Time PCR Detection System (Bio-Rad) was used for gene expression analysis. Standard curves, expression normalization, and standard error values were obtained using CFX Manager software ver. 1.5 (Bio-Rad), which employs the DDC t method, and16S rRNA was used for normalization. RT-qPCR products were verified by melting curve and peak analyses.

Two-hybrid System
A C. glutamicum genomic library was prepared as previously described [29]. The BacterioMatch II Two-Hybrid system (Agilent Technology) was used according to manufacturer instructions. Briefly, the two plasmids, pBT and pTRG, containing the ''bait'' and ''target'' genes, respectively, were used to transform E. coli simultaneously. Protein-protein interactions were screened based on the expression of his3 and aadA, which confer histidine prototrophy (His + ) and streptomycin resistance (Str + ), respectively. For screening, 50 ng of each pSL500 (i.e., pBT-glxR) and target library DNA were introduced into reporter cells and spread onto selective media (His 2 and Str + ). Colonies were isolated and the plasmids in the growing cells were analyzed.

Protein Purification and Proteolytic Assays
MBP-GlxR was purified as previously described [6]. His 6 -SprA was overexpressed, isolated, and refolded on a HisTrap FF column (GE Healthcare) as described in the Recombinant Protein Purification Handbook (GE Healthcare). Briefly, we grew E. coli BL21-CodonPlus (DE3)-RIL cells harboring plasmid pSL510 in 300 mL Luria-Bertani broth to an optical density of 0.5 at 600 nm, added isopropylthio-b-galactoside to a final concentration of 0.3 mM, and harvested the cells after an additional 2 h of incubation. Cells were resuspended in 10 mL buffer (20 mM Tris-HCl, 100 mM NaCl, pH 8.0) and disrupted by sonication. Inclusion bodies were collected via centrifugation at 9,500 6 g for 1 h, washed twice with buffer (2% Triton X-100, 10 mM Tris-HCl, 100 mM NaCl, 2 M urea, pH 8.0), and incubated at room temperature for 4 h in 10 mL denaturation buffer (8 M urea, 20 mM Tris-HCl, 100 mM NaCl, 10 mM imidazole, pH 8.0) for denaturation. Solubilized proteins were recovered by centrifugation at 9,500 6 g for 1 h, followed by passage of the supernatant through a syringe filter (0.45 mm). Proteins were loaded onto a Ni + column (1 mL HisTrap His-FF, GE Healthcare), which was subsequently washed with 10 mL denaturation buffer at a flow rate of 0.1 mL min 21 . Refolding of His 6 -SprA was induced by slowly (0.1 mL min 21 ) replacing the denaturation buffer with 20 mL refolding buffer (20 mM Tris-HCl, 100 mM NaCl, 10 mM imidazole, 5 mM dithiothreitol, pH 8.0) through the application of a linear gradient. The proteins were then eluted with 5 mL elution buffer (20 mM Tris-HCl, 100 mM NaCl, 1 M imidazole, pH 8.0) at a flow rate of 1 mL min 21 . Proteins were concentrated via ultrafiltration (Amicon Ultra Centrifugal Filter Devices, Millipore), and the buffer was replaced with assay buffer (40 mM potassium phosphate, pH 8.0).
Proteolytic assays were conducted in a total volume of 25 mL at 30uC by incubating proteins for 0-24 h. The assay buffer was composed of 40 mM potassium phosphate (pH 8.0). Subsequently, the assay mixtures were analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). The proteins in the assay mixtures were as follows (pmol): 200 MBP-GlxR, 400 GlxR, 630 His 6 -SprA, and 111 bovine serum albumin (BSA).
Cell lysates of C. glutamicum cells were prepared as described previously [6] and enzymatic activities of MS, ICL, and isocitrate dehydrogenase (ICDH) were determined as previously described [35].

Isolation of Proteins Interacting with GlxR
To isolate protein(s) interacting with GlxR, we used a twohybrid system with pSL500, which expresses GlxR, as the bait vector. Protein-protein interactions were screened based on the expression of his3 and aadA, which confer His + and Str + , respectively. Introduction of the C. glutamicum genomic library into E. coli reporter cells harboring the bait vector allowed us to isolate five target clones exhibiting the His + and Str + phenotypes. Subsequently, the plasmids were isolated, and the DNA fragments in the ''target'' vector were sequenced. Two of the clones contained gene fragments that expressed peptide sequences, which turned out to be subtilisin-like serine protease (NCgl0550) and Xaa-proline peptidase (NCgl1430; Table 1). To verify the interaction of the proteins with GlxR, we cloned the full-length open reading frames of NCgl0550 and NCgl1430 into the target vector, introduced them into reporter cells carrying the glxRcarrying bait vector, and monitored their growth on selection media. Only the cells carrying the NCgl0550 gene showed growth on the medium (Figure 1). To quantify the protein-protein interaction, we measured the transcriptional level of the reporter gene his3 by RT-qPCR and determined it to be 11.3% compared with that of the positive control cells, which was set at 100%. The value observed in cells harboring empty vectors was 3.4%. Considering that the screening conditions may not have been physiologically ideal for inducing protein interactions (see Discussion), we tentatively concluded that GlxR specifically interacts with the NCgl0550-encoded protein. We then analyzed the gene further to see if the protein product interacted with GlxR and whether this interaction had physiological significance.

Analysis of NCgl0550
The open reading frame of NCgl0550 was 1,245 bps long and encoded a 43,143-Da protein composed of 414 amino acids. The NCgl0550-encoded protein had conserved features found in subtilisin-like serine proteases (InterProScan program, http:// www.ebi.ac.uk/Tools/pfa/iprscan), such as hydrophobic domains in the N-terminus and near the C-terminus and a catalytic triad characterized by the conserved amino acids Asp, His, and Ser [36,37]. The catalytic triad within conserved sequences D90, H121, and S332 is common among bacterial serine proteases of the subtilisin family. In addition, the encoded protein exhibited typical subtilisin family domain architecture, which is most common among subtilisin-like serine proteases and is associated with a proposed biological function of proteolysis [38]. In addition, a putative signal peptide cleavage site (between the 23rd and 24th amino acids) was also found in the N-terminal region of the protein, suggesting that the protein could be secreted or membrane anchored. The encoded protein showed approximately 35% identity with the mycosins (encoded by mycP1-5), which are subtilisin-like serine proteases, of closely related Mycobacterium tuberculosis H37Rv, whereas the five serine proteases of the organism have been reported to show 36-47% identity with each other [39]. The encoded protein also showed 32% identity with the subtilisin-like serine protease (YaB, encoded by ale) of Bacillus subtilis [40]. Homologies with other subtilisin-like serine proteases were generally low (data not shown). Based on the homology with serine proteases, NCgl0550 was designated as sprA (serine protease A).
To study the role of sprA in glxR-mediated regulation, we first used RT-qPCR to analyze the expression profile of sprA during growth. The expression of glxR showed only marginal differences, whereas that of the sprA gene was maximal in the log phase and gradually decreased as the cells entered the stationary phase, suggesting that it plays a role in the active growth phase (Figure 2).
To determine the importance of sprA, we constructed a C. glutamicum sprA deletion mutant (DsprA) and a sprA-overexpressing strain (P 180 -sprA) and monitored their growth properties on minimal media. Internal deletion of a 494-bp fragment in sprA was verified by PCR (data not shown). The identity of the deleted gene was verified by complementing the DsprA strain with the plasmid pSL535 (i.e., pMT1-sprA), which harbors the sprA gene with its own promoter; the complemented strain showed wild typelike growth ( Figure 3A). Promoter P 180 is known to overexpress the fused gene irrespective of the growth phase [31]. Overexpression of sprA (approximately 15-fold relative to that of the wild-type strain) was confirmed by RT-qPCR. As shown in Figure 3A, when grown on glucose minimal media, the DsprA and P 180 -sprA strains exhibited reduced growth in the log phase relative to that of the wild-type strain, requiring longer incubation time to reach the stationary phase. In accordance with the expression data (see Figure 2), deletion of the gene did not significantly affect cell growth in the stationary phase. The differences in growth were further minimized when the cells were grown on acetate minimal medium ( Figure 3B). Collectively, these data suggest a role for sprA in central metabolism during growth.

In vitro Proteolysis of GlxR by SprA
Because SprA was isolated based on its capability to interact with GlxR, we aimed to investigate protein-protein interactions in a purified system. The overexpressed MBP-GlxR fusion protein was soluble and could be purified using a conventional approach without difficulty. However, the overexpressed His 6 -SprA fusion protein was insoluble due to the formation of inclusion bodies.  Therefore, we used a different purification approach, in which His 6 -SprA fusion protein was denatured with urea, bound to the column, and renatured on the column by slowly removing urea before elution. Once the protein was purified, it remained soluble ( Figure 4A). However, during the purification process, partial degradation of His 6 -SprA was observed. Subsequently, we tested the stability of the purified His 6 -SprA by incubating it at 4uC or 30uC for 14 h. As shown in Figure 4B, the 48.2-kDa His 6 -SprA fusion protein underwent self-proteolysis during incubation at 4uC: SDS-PAGE revealed that fractions of purified SprA were digested to 27-to 35-kDa proteins (see Figure 4B, bracket 1). When the incubation was performed at 30uC, even smaller fragments were formed (see Figure 4B, bracket 2), indicating nearly complete digestion of His 6 -SprA. Although serine proteases are commonly synthesized as zymogens, which are converted to active enzymes via proteolytic cleavage at a particular peptide bond, no major identifiable protein band was detected in His 6 -SprA. In contrast, the 67.5-kDa MBP-GlxR fusion protein remained stable during purification and incubation; apparent self-proteolysis was not noted on performing SDS-PAGE ( Figure 5, lanes 1 and 2) after a 24-h incubation period at 4uC or 30uC. After observing the proteolytic activity of His 6 -SprA, we tested it against MBP-GlxR.
To minimize activity loss, we did not remove the protein fusion partners, such as MBP and His 6 , from each protein, eliminating additional purification steps. Interestingly, when the MBP-GlxR and His 6 -SprA fusion proteins were incubated together, the protein band corresponding to MBP-GlxR, as well as the band corresponding to SprA ( Figure 5, lane 6), disappeared completely. However, a new 45-kDa protein band was detected by SDS-PAGE and remained stable. The M r of the protein ( Figure 6A, lane 1) was nearly identical to that of purified MBP (see Figure 6A, lane 2). To elucidate its nature, we excised the band from the gel for quadrupole time-of-flight (Q-TOF) analysis. The amino acid composition of the band indicated that the protein product was MBP, suggesting that SprA is specific for its GlxR target. We performed additional analyses by incubating His 6 -SprA with BSA. As shown in Figure 6B, proteolysis of BSA was not observed, whereas self-proteolysis of His 6 -SprA was evident. Although we observed digestion of GlxR by SprA, the reaction was slow, taking a maximum of 24 h. This result could be due to the poor activity of SprA, which was purified using an on-column refolding procedure. The digestion of GlxR by SprA was only observable with fresh protein preparations. In addition, the fusion partners MBP and His 6 may have hindered protein-protein interaction. Nevertheless, these data indicate that, although unstable, SprA acts on the globular form of GlxR and digests it with specificity. Furthermore, self-proteolysis of SprA may indicate the importance of controlling the availability of the protein in cells for GlxR and other proteins.

Effects of cAMP on the Proteolysis of GlxR by SprA
Intracellular cAMP levels in C. glutamicum are elevated during growth on glucose, especially in the early log phase, and low during growth on acetate. The DNA-binding activity of GlxR is also modulated by cAMP. Once SprA was found to specifically proteolyze GlxR, the effects of cAMP on the proteolysis were investigated, using native GlxR by removing MBP from MBP-GlxR. As shown in Figure 7 (lanes 1 and 2), purified GlxR remained stable during the incubation period at 4uC or 30uC. When GlxR was incubated in the presence of SprA at 30uC (see

Effects of sprA on the Expression of aceA, aceB, and icd
The genes aceA and aceB, which encode ICL and MS, respectively, are repressed by glxR in cells supplied with glucose as the sole carbon source. Therefore, one can speculate that C. glutamicum P 180 -sprA cells overexpressing the sprA gene may have depleted intracellular GlxR owing to proteolysis by SprA and may show derepression of aceB and aceA genes, even in cells supplied with glucose as the sole carbon source. Conversely, C. glutamicum cells grown on glucose have elevated levels of intracellular cAMP [6]. Therefore, as shown in Figure 7, SprA may be unable to proteolyze GlxR due to high intracellular cAMP levels in glucosegrown cells, resulting in the repression of aceA and aceB. Furthermore, the C. glutamicum DglxR strain, which has no intracellular GlxR, is known to have a severe growth defect [24,25], whereas the C. glutamicum P 180 -sprA strain, which is considered equivalent to the DglxR strain, showed reduced but reasonable growth on glucose minimal medium (see Figure 3 A), suggesting that the intracellular GlxR might be intact. To determine whether the in vitro data of the present study agree with the reported observations for whole cells, we measured the ICL and MS activities from P 180 -sprA cells supplied with glucose as the carbon source. As shown in Table 2, the ICL and MS activities in P 180 -sprA cells were nearly comparable to those in the wild-type strain. These observations suggest that GlxR is almost intact in the cells, supporting the growth of the strain (see Figure 3A). Conversely, when acetate was supplied as the carbon source, as in the wild-type strain, derepression of glyoxylate bypass enzymes was observed in both the DsprA and P 180 -sprA strains (see Table 2). The observation that the level of derepression in the DsprA strain was comparable to that in the wild-type and P 180 -sprA strains may indicate that the amount of intracellular SprA is not important under those growth conditions. The results obtained with both DsprA and P 180 -sprA strains on glucose or acetate as the carbon source are in accordance with the in vitro data and suggest that the in vivo state of GlxR, the activity of which responds to cAMP, may determine its fate as the substrate for the SprA protein (see Discussion).
Next, we measured the activity of isocitrate dehydrogenase (ICDH, encoded by icd), which is a key enzyme for the tricarboxylic acid cycle and converts isocitrate to a-ketoglutarate and CO 2 with the concomitant release of reduced nicotinamide adenine dinucleotide phosphate. In C. glutamicum, the enzyme is constitutively formed independent of the growth substrate [41,42]. Kim et al [6] have observed increased ICDH activity (twofold higher) in glucose-grown C. glutamicum cells harboring multiple copies of glxR, suggesting a positive role of glxR in icd expression, whereas no such effect was observed in acetate-grown cells. As shown in Table 2, the ICDH activity observed in the P 180 -sprA strain was only 32% of that of the wild-type strain, suggesting that the stimulatory effect of glxR on icd was abolished in the P 180 -sprA strain. In accordance with the finding by Kim et al [6], the enzyme activities of ICDH were found to be unaffected in acetate-grown cells regardless of the presence of the P 180 -sprA plasmid or DsprA mutation (see Table 2). The decrease in ICDH activity in glucosegrown cells was speculated to be due to the transcriptional control of icd because activity was unaffected in acetate-grown cells. Although we do not know the cause of this result, we also observed a decrease in ICDH activity in DsprA cells grown with glucose as the sole carbon source. The activity was 50% of that in the wildtype strain (see Table 2). Decreases in ICDH activity in glucosegrown DsprA cells suggest the involvement of additional regulatory proteins (see Discussion). We hypothesized that the decrease in ICDH activity in P 180 -sprA and DsprA mutant cells contributed to the slower growth of the cells on glucose minimal medium (see Figure 3A). In support of this hypothesis, growth differences in the cells grown on acetate minimal medium were marginal (see Figure 3B). Similarly, the attenuation of ICDH activity decreases biomass yield and increases lysine production in lysine-producing C. glutamicum cells [43]. Collectively, these data suggest that sprA has unknown regulatory roles and that its involvement in glxRmediated regulation may be more complex than anticipated.

Identification of Proteins under the Control of sprA
After investigating the role of sprA in general cell physiology (growth differences are shown in Figure 3, and differences in ICDH activity between P 180 -sprA and DsprA strains are in Table 2), we assumed that additional proteins might be affected in P 180 -sprA cells. To identify proteins with altered expression, we used 2D-PAGE analysis and compared the protein profiles of P 180 -sprA cells with those of wild-type cells. We chose P 180 -sprA cells because they exhibited a more severe growth defect than DsprA cells. Our analysis revealed the following five protein spots (Figure 8): pyruvate kinase-like protein (NCgl2809), succinate-semialdehyde dehydrogenase (NCgl0049), S-adenosyl-L-homocysteine hydrolase (NCgl0719), D-alanyl-alanine synthetase A (NCgl1267), and exopolyphosphatase (NCgl0938). These proteins showed decreased spot density in the P 180 -sprA strain. Although we found a pyruvate kinase-like protein, pyruvate kinase is a well-known enzyme that catalyzes the last step in glycolysis, forming ATP from ADP by converting phosphoenolpyruvate to pyruvate [44]. Succinate-semialdehyde dehydrogenase activity has been suggested to play important roles in nitrogen metabolism in E. coli [45]. S-Adenosyl-L-homocysteine hydrolase is an important regulator of Sadenosylmethionine-dependent biological methylation reactions and catalyzes the reversible breakdown of S-adenosyl-L-homocysteine to adenosine and homocysteine [46]. D-Alanyl-D-alanine ligase (or synthetase) is a cytoplasmic enzyme that provides the Dalanyl-D-alanine dipeptide substrate in the initial cytoplasmic phase of peptidoglycan biosynthesis [47]. Exopolyphosphatase catalyzes the hydrolysis of exopolyphosphate, which plays a role in cell stress, starvation, survival, and virulence [48]. Next, to determine whether the decreased intensity in protein spot was due to a decrease in the transcription of the corresponding genes or increased proteolysis by the SprA, we measured the amount of corresponding mRNA using RT-qPCR. The amount of mRNA in P 180 -sprA cells decreased to approximately 10-30% of that in the wild-type strain (Figure 9), showing that the decrease in spot intensity was caused by decreased transcription of the corresponding genes. Reduced transcription of the genes was also observed in DsprA cells (Figure 9), suggesting that the action of SprA was conveyed via transcriptional regulators, such as GlxR. Although genes in the GlxR regulon are mostly known from studies involving in silico genome analysis and in vivo and in vitro DNAbinding assays [20,21,23], the genes that we identified in the above analysis have not been listed among those in the GlxR regulon. However, the known functions of these proteins, such as those in energy and carbon metabolism (NCgl2809), nitrogen metabolism (NCgl0049), methylation reactions (NCgl0719), peptidoglycan biosynthesis (NCgl1267), as well as stress, starvation, and survival (NCgl0938) are associated with the functional category in which glxR plays a role. All of the data obtained suggest that SprA performs an important role not only in glxR-mediated gene regulation but also in other areas of cell physiology.

Discussion
In this study, we used a two-hybrid system to isolate the subtilisin-like serine protease SprA as a partner that interacts with GlxR. In the two-hybrid system, the bait (GlxR) and target (SprA) proteins are expressed as fusion proteins linked to the lcI repressor and RNA polymerase a subunit, respectively. Interactions between the bait and target proteins stabilizes the interaction of RNA polymerase with its weak promoter, resulting in the expression of aadA and his3 reporter genes, which confer Str+ and His+, respectively. The in vivo GlxR-SprA interaction was quantified using RT-qPCR, which measured his3 transcription, and determined to be low relative to that of the positive control. This decrease could be due to the proteolysis of GlxR with the interacting SprA, decreasing the transcription of the reporter genes. This assumption is highly supported by the observed GlxR proteolysis by SprA in a purified system. Self-proteolysis of SprA may have also contributed to the decrease in the reporter gene transcription. In addition, the binding of GlxR to SprA as a proteolytic substrate may not be strong enough to stabilize interactions between RNA polymerase and its promoter to allow transcription.
The finding that GlxR interacts with a protease is surprising because the common role of proteases is protein quality control, such as the degradation of damaged or unfolded proteins [49]. Subtilisin-like serine proteases are also generally secreted extracellularly to scavenge nutrients, although intracellular roles are also commonly found. Proteases are known to conduct critical regulatory functions via the proteolysis of regulators, enzymes, and other proteins [26,50]. Regulators that show either functional or structural similarity or both to GlxR also exist. For example, the global transcription factor FNR, which plays a role in anaerobiosis in E. coli, is controlled by proteolysis [51,52] as well as the oxidative stress response regulator Spx in low-GC Gram-positive bacteria [53]. The global regulator Mlc, which is a transcriptional repressor of sugar-metabolizing enzymes and uptake systems in E. coli, is also regulated by proteolysis [54]. A protease system is also involved in the cellular turnover of FixK2, which is a CRP-like transcription factor that controls the endosymbiotic lifestyle of Bradyrhizobium japonicum [55]. In C. glutamicum, the Clp complex and FtsH protease are involved in nitrogen control through action on GlnK [56,57]. In addition, C. glutamicum FtsH protease has been implicated in the regulation of energy and carbon metabolism as well as in amino acid biosynthesis [58]. Therefore, it is not surprising to find SprA proteolyzing the global regulator GlxR with specificity. A serine protease MycP1 from M. tuberculosis, a close relative of C. glutamicum, also performs a novel role with defined substrate specificity [59]. Considering the constitutive transcription of glxR (see Figure 2), one can speculate the involvement of additional regulatory mechanisms, such as proteolysis, for GlxR.
In general, serine proteases show broad substrate specificity. Although many function as general proteases for unfolded substrates, some cleave only one substrate. In addition, many prokaryotic serine proteases show domain architectural diversity and thus function in an organism-specific manner [38]. Several observations suggest that sprA performs a novel role in C. glutamicum: (a) Both deletion and overexpression of the gene resulted in a retarded growth phenotype, indicating a global role for the gene. (b) The transcription of sprA showed an atypical temporal pattern, suggesting a specialized role for the gene.
Typically, most subtilisin genes are expressed at low levels during the log growth phase [60]. M. tuberculosis myc genes are also expressed constitutively [39]. (c) SprA was not homologous to other subtilisin-like serine proteases. (d) Complete self-proteolysis of SprA is unique and suggests that tight control of SprA availability is important owing to its regulatory role in cell physiology. (e) Although five myc genes are present in closely related M. tuberculosis, only one homologous gene (sprA) is found in C. glutamicum, suggesting a novel role.
Although we initially isolated SprA as a protein that specifically interacted with GlxR, its regulatory role now appears to be much more complex than we expected. First, the P 180 -sprA strain showed no lethal phenotype, indicating that GlxR was still active in the cell but reduced cell growth, which suggests that other cellular functions were affected. In addition, the proteins identified with 2D-PAGE are not known to be included in the GlxR regulon, suggesting that putative GlxR-binding motifs are absent in the regulatory region of the identified genes. Furthermore, the activity of ICDH was severely affected in DsprA and P 180 -sprA cells. Icd, encoding ICDH, showed no glxR-binding motif in its promoter and regulatory regions and is excluded from the GlxR regulon. However, this gene is still regulated by glxR, as noted in our study and by Kim et al [6], suggesting unknown regulation in central carbon metabolism. In E. coli, the icd gene has two different promoters and is controlled by a complex regulatory system involving the global regulatory proteins ArcA, Fnr, and Cra [61,62]. The expression of icd in E. coli is controlled by carbon and oxygen availability; its expression decreased during anaerobic growth and increased in the presence of poor carbon sources, such   as acetate. In this type of control, ArcA and Fnr act negatively, and Cra exerts positive control. Analogous to this, our results suggest a complex regulatory mechanism for the C. glutamicum icd gene as well.
The activity of GlxR is modulated by cAMP. In all the cases studied, the in vitro binding of GlxR to its target binding sites is strictly dependent on the presence of cAMP. Unlike the levels in E. coli, intracellular cAMP levels in C. glutamicum are reported to be elevated during growth on glucose and low during growth on acetate [6]. Therefore, cells grown on glucose may have GlxR in its holo form; that is, GlxR may exist as a cAMP-bound protein.
However, as shown in Figure 7, SprA may not act on the holo form of GlxR. We observed proteolysis of GlxR by SprA only in the absence of cAMP, suggesting that only the apo form of GlxR could serve as the substrate for SprA. The in vitro binding of GlxR to its target DNA requires the addition of cAMP to the assay mixture, probably to convert the GlxR into its cAMP-bound holo form, which is considered active and protected from proteolysis by SprA, thereby performing its function by repressing glyoxylate bypass genes. As a result, ICL and MS activities are minimal in glucose-grown DsprA and P 180 -sprA strains, as shown in Table 2. The reasonable growth of the P 180 -sprA strain on glucose minimal medium (see Figure 3A) suggests that intracellular GlxR is mostly intact and supports the above-mentioned hypothesis. Conversely, Figure 8. Identification of protein affected in sprA-overexpressing C. glutamicum (P 180 -sprA). 2D-PAGE was performed as described in the Materials and Methods. A total of 150 mg of protein was loaded onto each gel. The identity of each protein spot was determined by electrospray ionization mass spectroscopy. doi:10.1371/journal.pone.0093587.g008 in acetate-grown cells, GlxR probably exists in its apo form; that is, it does not bind to cAMP owing to low intracellular cAMP levels.
If cAMP is absent in the cells, the DNA-binding activity of GlxR remains minimal, thereby derepressing the glyoxylate bypass genes. Under this condition, the overexpression of sprA (P 180 -sprA) or deletion of the gene (DsprA) has no significant effect on the expression of glyoxylate bypass genes, as shown in Table 2. In wild-type cells, residual intracellular GlxR in its apo form can be further proteolyzed by SprA and cleared from the cell, preventing residual GlxR from performing any regulatory role (see below). Thus, our hypothesis that only the apo form of GlxR, which is not bound to cAMP, serves as the substrate for SprA is logical, because otherwise, uncontrolled degradation of GlxR by SprA would occur and disturb the cell physiology by depleting GlxR.
We can also speculate about why proteolysis is used to regulate GlxR, which has DNA-binding activity that is already controlled by intracellular cAMP levels. Any protein subject to regulated synthesis will also be degraded as part of a regulatory network [26], and this arrangement may be particularly important for proteins, such as glxR, that show autoregulation [19] because the basal levels of the protein can be kept even lower if the protein is subject to degradation. In the case of global regulators, small changes in synthesis may inappropriately switch on genes and disturb cell physiology. One of the roles of SprA may be to keep the intracellular levels of GlxR below working concentrations, which may be why GlxR is the target for a serine protease rather than an ATP-dependent AAA+ protease that requires a more complex regulatory mechanism. Alternatively, the removal of GlxR may be important when it is no longer needed. GlxR is well expressed during growth (see Figure 2), but as conditions fluctuate, GlxR, which once served a useful function, likely requires clearance from the cell at times. The primary goal of SprA may be to remove GlxR when it is no longer needed. As mentioned above, cAMP probably plays a key role in GlxR-SprA interaction and the stimulation of proteolysis. The constitutive expression of glxR, temporal expression of the sprA gene (see Figure 2), and selfproteolysis of SprA to remove itself after use (see Figure 4) support this hypothesis.
In conclusion, we isolated SprA as an interacting partner for GlxR. SprA preferentially functions in the log phase and mediates the proteolysis of GlxR with specificity. In addition, sprA appears to play additional regulatory roles in general cell physiology. The self-proteolytic activity of SprA suggests that the immediate and timely removal of SprA is also important to ensure precise regulatory control.