Single mutations in BraRS confer high resistance against nisin A in Staphylococcus aureus

Abstract Nisin A is a lantibiotic produced by Lactococcus lactis that is widely used as a food preservative. In Staphylococcus aureus, the BraRS two‐component system (TCS) senses nisin A and regulates the expression of the ABC transporter VraDE, which is responsible for nisin A resistance. In this study, we exposed S. aureus to a sub‐minimum inhibition concentration of nisin A and obtained three spontaneous mutants that were highly resistant to this lantibiotic, designated as SAN (S. aureus nisin resistant) 1, SAN8, and SAN87. In the wild‐type S. aureus strain, VraDE expression was induced by nisin A. In contrast, SAN8 and SAN87 showed constitutively high VraDE expression, even in the absence of nisin A, while SAN1 showed higher BraRS expression, which resulted in high VraDE expression in the presence of nisin A. We identified a single mutation in the promoter region of braXRS in SAN1, whereas SAN8 and SAN87 had single mutations in braR and braS, respectively. Interestingly, even the unphosphorylated form of the mutant BraR protein induced VraDE expression. These results indicate that conformational changes in BraS or BraR resulting from the point mutations may result in the constitutive expression of VraDE, allowing S. aureus to adapt to high concentrations of nisin A.

. The primary mode of action of nisin A involves its binding to lipid II to inhibit cell wall biosynthesis and promote pore formation in bacterial membranes (Bierbaum & Sahl, 2009).
We and other researchers previously reported on the association between the BraRS two-component system (TCS) and nisin A resistance (Hiron, Falord, Valle, Débarbouillé, & Msadek, 2011;Kawada-Matsuo, Yoshida, et al., 2013). In addition, BraDE has also been shown to be involved in nisin A sensing and signaling through BraRS (Hiron et al., 2011). Finally, the phosphorylated BraR protein induces the expression of the ABC transporter VraDE, an intrinsic factor for nisin A resistance. BraRS is also associated with resistance to bacitracin and nukacin ISK-1, which act upon undecaprenol pyrophosphate and lipid II, respectively (Bierbaum & Sahl, 2009;Islam, et al., 2012). However, because S. aureus MW2 showed a relatively low minimum inhibition concentration (MIC) for nisin A (MIC: 512 µg/ ml), high concentrations of this lantibiotic have antibacterial activity against S. aureus (Hiron et al., 2011;Kawada-Matsuo, Yoshida, et al., 2013). Many studies have investigated lantibiotics such as nisin A for their clinical use and as food additives (Breukink & de Kruijff, 2006;Field, Cotter, Hill, & Ross, 2015;Gharsallaoui et al., 2016;Shin et al., 2016). To determine whether the application of nisin A could select for a mutant with decreased susceptibility to nisin A, we attempted to isolate such mutants by exposing S. aureus cells to a sub-MIC of nisin A. As a result, we obtained several strains exhibiting a decreased susceptibility to nisin A. We also identified several point mutations in the braXRS region resulting in high VraDE expression.
These results indicate that endogenous mutations conferring high levels of nisin A resistance in S. aureus can arise through exposure of cells to a sub-MIC of nisin A.

| Bacterial strains and growth conditions
The bacterial strains used in this study are listed in Table 1 and Luria Bertani (LB) broth, respectively. Tetracycline (5 µg/ml) and chloramphenicol (5 µg/ml) were used to select for S. aureus, and ampicillin (100 µg/ml) was used to select for E. coli when necessary.

| Isolation of spontaneous mutants by nisin A exposure
The S. aureus strain MW2 was used to isolate spontaneous mutants that were highly resistant to nisin A using a microdilution method.
The MW2 strain was cultured overnight, and an aliquot (containing 10 5 cells) was inoculated into 100 µl of TSB containing various concentrations of nisin A (Sigma-Aldrich; twofold dilutions: 16,384 to 16 µg/ml) and incubated at 37°C overnight. Next, using the bacterial cells that grew in the 1/2 MIC of nisin A, the same experiment was repeated two additional times. After the last subculture, the bacterial cells grown in the 1/2 MIC of nisin A were appropriately diluted and plated on TSA. After an overnight incubation, seven colonies were randomly picked and replated on TSA. Subsequently, the MICs of nisin A were determined for the seven strains. This experiment was performed three times independently (experiments 1, 2, and 3).
The expression of VraD (MW2620) was investigated in the strains exhibiting increased MICs for nisin A compared to the wild-type strain. The S. aureus strains were cultured overnight, and an aliquot (containing 10 8 cells) was inoculated into 5 ml of fresh TSB and grown at 37°C with shaking. When the optical density at 660 nm reached 0.5, nisin A (64 µg/ml) was added to the bacterial culture. After a 15 min of incubation, the bacterial cells were collected and total RNA was extracted using a FastRNA Pro Blue Kit (MP Biomedicals, Solon, OH, USA) according to the manufacturer's protocol. Next, 1 µg of total RNA was reverse-transcribed to cDNA using a first-strand cDNA synthesis kit (Roche, Tokyo, Japan). Using the cDNA as template, quantitative PCR was performed using a LightCycler system (Roche, Tokyo, Japan). Primers were designed to amplify MW2620 (vraD), and gyrB was used as an internal control.
The primers used in this assay are listed in Table 2. Finally, the strains exhibiting increased MICs and an increased expression of MW2620 in the absence of nisin A were selected for further analysis.
In addition, primers were designed to amplify the vraD promoter region (Table 2). To prepare chromosomal DNA from the mutant strains, the cells from 1 ml of overnight cultures were collected. The cells were resuspended in 100 µl of 10 mM Tris-HCl (pH 6.8) containing 10 µg lysostaphin (Sigma-Aldrich) and incubated at 37°C for 20 min followed by an incubation at 95°C for 15 min. After centrifugation, the cell lysates were used as template DNA for PCR. PCR was performed using the Takara Ex Taq system, and the amplicons were purified using a QIAquick kit (Qiagen, Hilden, Germany). The nucleotide sequences of each DNA fragment were determined using specific primers, the sequences of which are listed in Table 2.

TA B L E 1 Strains used in this study
Construction of the plasmid for reporter assay using β-galactosidase MW2546-5′ phosphate RT primer: 5′-ccccatttgtattgc-3′ Amplification of DNA fragments used in gel shift assay Construction of the plasmid for recombinant protein

| Inactivation of braRS in the mutant and its complementation
The method used to inactivate braRS with the thermosensitive plasmid pCL52.1 was described previously (Kawada-Matsuo, Yoshida, et al., 2013). For gene complementation, the isopropyl-β-d-thiogalactopyranoside (IPTG)-inducible vector pCL15 was used. A DNA fragment for complementation was PCR-amplified using chromosomal DNA from the wild-type or mutant strains described above. The DNA fragment was cloned into pCL15 and transformed into E. coli XL-II competent cells. The obtained plasmid was electroporated into S. aureus RN4220 and was subsequently transduced into the mutant strain using the phage 80α.

| Analysis of the vraDE and braXRS promoter activities in nisin-resistant mutants using a reporter system
Before the reporter assay, we identified the promoter region of braXRS using the rapid amplification of cDNA ends (RACE) method.
RACE was performed using a 5′-Full RACE Core Set (Takara Bio Inc., Ohtsu, Japan) according to the manufacturer's protocol, and the primers used in this assay are listed in Table 2. To analyze the braXRS and vraDE promoter activities, the respective promoter regions were fused to the β-galactosidase gene using a PCR method.
Briefly, the promoter regions and the β-galactosidase gene were PCR-amplified such that the downstream primer of the promoter region and the upstream primer of the β-galactosidase gene contained ten overlapping nucleotides to allow the two PCR fragments to be joined together. After the first PCR, the two resulting PCR fragments were mixed and heated at 95°C, after which they were cooled to 37°C. Next, a second PCR was performed to amplify the fused fragment using primers listed in Table 2. The fragment was cloned into pLI50, a shuttle vector for E. coli and S. aureus, and the resulting plasmid was electroporated into S. aureus RN4220. Next, the plasmid was transduced into several S. aureus strains using the method described above. β-Galactosidase assays were performed with a SensoLyte ONPG β-Galactosidase Assay Kit (ANASPEC, CA, USA).

| Expression of braR and vraD
Quantitative PCR and immunoblotting were performed to assess the expression of braR/BraR and vraD/VraD. The S. aureus strains were cultured overnight, and aliquots (containing 10 8 cells) were inoculated into 5 ml of fresh TSB and then grown at 37°C with shaking. When the optical density reached 0.5 at 660 nm, nisin A (64 µg/ml) was added to the bacterial culture. After incubating for 15 min (for quantitative PCR) and 2 hr (for immunoblotting), the bacterial cells were collected. For quantitative PCR, RNA extraction, cDNA synthesis, and PCR were performed as described above. For immunoblotting, antiserum against VraD was obtained by immunizing mice with the recombinant protein as described previously . Briefly, histidine-tagged recombinant VraD (rVraD) was constructed for the immunization. The DNA fragment encoding VraD was amplified with the specific primers listed in Table 2

| Electrophoretic mobility shift assay
For the electrophoretic mobility shift assay (EMSA), 6× histidinetagged recombinant BraR (rBraR) was utilized. A DNA fragment encoding BraR was amplified with the specific primers listed in Table 1 and was subsequently cloned into pQE30 (Qiagen). The plasmid was then transformed into E. coli M15 (pREP4), and the recombinant protein was purified according to the manufacturer's instructions. Purified protein was phosphorylated with a method described elsewhere (Gao, Gusa, Scott, & Churchward, 2005). The rBraR protein was incubated for 2 hr at room temperature in 50 mM Tris-HCl (pH 8.0), 10 mM MgCl2, 3 mM dithiothreitol, and 32 mM acetyl phosphate. To assess the binding of rBraR to the region upstream of vraDE, an EMSA was performed as described previously (Mazda et al., 2012). A DNA fragment encompassing the region upstream of vraDE and a fragment lacking the binding region were amplified with the specific primers listed in

| Isolation of S. aureus strains with high levels of nisin A resistance and VraD expression
To obtain S. aureus MW2 mutants with high nisin A resistance, cells were exposed to increasing nisin A concentrations (1st, 256 µg/ml; 2nd, 1,024 µg/ml; and 3rd, 2,048 µg/ml). All 21

| DNA sequence of the braXRS, braAB, and vraD regions
The DNA sequences of the braRS, braAB, and vraD regions in the SAN1, SAN8, and SAN87 strains were determined. In the SAN1 strain, only one mutation was observed in the promoter region of braXRS ( Figure 2).
In the SAN8 strain, one mutation in the braR region was observed that

| Susceptibility of strains to various antibacterial agents
We evaluated the MICs of various antibacterial agents against MW2, the nisin A-resistant mutants, their braRS-inactivated mutants, and the braRS-complemented strains (

| Expression of VraD in the mutants
We investigated the expression of VraD by immunoblotting and quantitative PCR (Figures 1 and 3). We observed similar protein and mRNA expression patterns in both experiments. The wild-type MW2 strain showed inducible expression by nisin A. In contrast,

VraD expression was very low in the SAN1 strain in the absence of nisin A, while VraD expression increased in the presence of nisin A,
showing higher expression than that observed in the wild-type strain ( Figure 3). However, when braRS was inactivated in the SAN1 strain,

VraD expression was not increased in the presence of nisin A, while
in the braRS-complemented strain, the VraD expression was similar with that observed in the SAN1 strain. In the SAN8 and SAN87 strains, VraD expression was higher in the absence and presence of nisin A than in the wild-type strain with no nisin A added. When braRS was inactivated in the SAN8 and SAN87 strains, VraD expression in these strains was absent. In the complemented strains, VraD expression was recovered and showed similar expression levels as the SAN8 and SAN87 strains.
We next assessed the expression of braR in these strains and observed that only the SAN1 strain showed high braR expression compared to the wild-type strain (Figure 4a), with the SAN8 and SAN87 strains showing similar expression as the wild-type strain (data not shown).

| braXRS and vraDE promoter activities
Since a point mutation in the SAN1 strain was observed in the braXRS promoter region, and the expression of braR was observed to be

TA B L E 3 Susceptibility of
Staphylococcus aureus mutants to various antibacterial agents increased compared to that in the wild-type strain by quantitative PCR (Figure 4a), we hypothesized that the braXRS promoter activity was increased in the SAN1 strain. We investigated the braXRS promoter activity in the wild-type and SAN1 strains, and the activity was higher in the SAN1 strain than in the wild-type strain (Figure 4b).
We also investigated the vraDE promoter activity in the SAN1, SAN8, and SAN87 strains. The results were similar to those observed in the quantitative PCR and immunoblotting analyses. The SAN8 and SAN87 strains exhibited higher vraDE activity than the wild-type strain in the absence and presence of nisin A, while the SAN1 strain showed higher activity only in the presence of nisin A ( Figure 5).

| Effect of the BraR mutation (in the SAN8 strain) on vraD expression and nisin A susceptibility
Using braRS from MW2 or SAN8, we complemented the braRS-inactivated MW2 and SAN8 strains. We observed that complementation using braRS from the SAN8 (MM2147) strain but not MW2 (MM2145) resulted in almost identical vraD expression levels as was observed in the SAN8 strain (Figure 6a). In addition, when braRS from the  (Figure 6b). Furthermore, we assessed the MIC of nisin A in these strains and observed that the MM2195 and MM2197 mutants showed increased nisin A MICs, the same as that observed for the SAN8 strain, while for the MM2194 and MM2196 strains, the introduction of braR from the MW2 wild-type strain did not increase the MIC of nisin A (Table 4). We also transduced braR from the SAN8 strain into RN4220, a methicillin-susceptible strain. An RN4220 strain harboring braR from the SAN8 strain showed an increase in the MIC of nisin A, whereas an RN4220 strain harboring braR from the MW2 strain showed no alteration in the MIC (Table 4).

| Binding of the wild-type and mutated BraR proteins to the upstream region of vraDE
Electrophoretic mobility shift assay showed the binding of phosphorylated MW2-rBraR and SAN8-rBraR (phosphorylated and nonphosphorylated) with the upstream region of vraDE. Figure 7a shows

| D ISCUSS I ON
In this study, we isolated three spontaneous mutants (SAN1, SAN8, and SAN87) exhibiting high levels of nisin A resistance and VraDE expression, all of which possessed single mutations in the braXRS

region. Gene inactivation and complementation experiments clearly
demonstrated that the point mutation in braXRS was directly associated with the high resistance of the mutants to nisin A. Two mutants, SAN8 and SAN87, showed constitutively high VraDE expression, even in the absence of nisin A. In contrast, the SAN1 strain showed low VraDE expression in the absence of nisin A but higher expression in the presence of nisin A than the wild-type strain. In previous reports, ApsRS, a TCS in S. aureus, was also associated with nisin A susceptibility (Kawada-Matsuo, Yoshida, et al., 2013). ApsRS regulates the expression of dlt and mprF to suppress the negative charge of the bacterial cell surface (Meehl, Herbert, Götz, & Cheung, 2007;Sakoulas et al., 2002). However, mutations in apsRS were not detected in the mutants isolated in this study.
In the SAN1 strain, only one point mutation was observed between the −35 and −10 box in the braXRS promoter region. We observed an increase in braRS expression in the SAN1 strain compared to that detected in the wild-type strain (Figure 4a). The reporter assay also revealed that the braXRS promoter activity in the SAN1 strain was 10 times higher than that observed in the wild-type strain (Figure 4b). Based on these results, we speculated that high amount of BraRS in the SAN1 strain increased the level of phosphorylated BraR by the addition of nisin A, which resulted in a higher induction of VraDE in response to nisin A than in the wild-type strain, although we did not quantify the level of phosphorylated BraR.
In the SAN8 strain, a BraR mutation at amino acid position 96 (aspartic acid to valine) (BraR M ) caused the expression of VraDE to be constitutively increased. When the BraR M allele was introduced into the braRS-inactivated MW2 strain (MM2195), this strain showed a similar MIC for nisin A to that observed F I G U R E 4 Expression of braR and the braXRS promoter activity in the MW2 and SAN1 strains. The expression of braR mRNA in the MW2 and SAN1 strains was evaluated by quantitative PCR (a) as described in the Section 2. The promoter activity of braXRS was evaluated using a β-galactosidase reporter system (b), as described in the Section 2. *p < 0.05, as determined by t test F I G U R E 5 Activity of the vraDE promoter in the mutants. The vraDE promoter activity was evaluated using a β-galactosidase reporter system as described in the Section 2. The plasmid for the reporter assay was constructed by fusing the vraDE promoter region with the gene encoding β-galactosidase. Next, the plasmid was transduced into various strains, and β-galactosidase activity was evaluated. *p < 0.05, as determined by Dunnett's post hoc tests compared to untreated MW2 for the SAN8 strain (Table 3). In addition, inactivation of braS alone in the SAN8 strain (MM2116) did not decrease the MIC of nisin A (data not shown). Based on these results, we considered that unphosphorylated BraR M could enhance the expression of VraDE. In the reporter assay, the vraDE promoter deleted of the BraR-binding region (MM2263) had no activity (Figure 8). These results suggested that BraR M bound to the same binding region upstream of vraDE as the native BraR. The EMSA assay also showed the binding affinity of BraR M to the upstream vraDE region and that its affinity was higher than that of the native BraR. Khosa, Hoeppner, Gohlke, Schmitt, and Smits (2016) reported the structure of NsrR from Streptococcus agalactiae, which showed homology with BraR from S. aureus. NsrRK is a TCS responsible for the expression of nsr and nsrFP, which are involved in nisin resistance (Khosa, AlKhatib, & Smits, 2013). Figure 9a shows an amino acid sequence alignment of response regulators reported to be as- Hill, 1999). According to the structural analysis of NsrR (Khosa et al., 2013), we showed the active site aspartate residue, the two switch residues, and the dimer interface regions of S. aureus BraR and other response regulators that were shown to be associated with nisin A susceptibility ( Figure 9a). The BraR M mutation site at position 96 is an aspartic acid residue (black triangle) that is adjacent to a phenylalanine residue, which is a conserved amino acid residue involved in a switch residue (dashed arrow). In a structural analysis of NsrR and ArcA, the phosphorylation of an aspartic acid causes a conformational change in two amino acid residues called switch residues (shown in the box in Figure 9a; Khosa et al., 2013;Toro-Roman, Mack, & Stock, 2005).
This conformational change induces the response regulator to form a dimer. In addition, four amino acid residues (three boxes and a dashed box in Figure 9a) After electrophoresis, DNA bands were detected as described in the Section 2 F I G U R E 8 The vraDE promoter activity in the braS-inactivated mutants. The vraDE promoter activity was evaluated using a β-galactosidase reporter system as described in the Section 2. The plasmid for the reporter assay was constructed by fusing the wild-type vraD promoter region or the vraD promoter region with the BraR-binding site deleted with the gene encoding β-galactosidase. Next, the plasmid was transduced into various strains, and β-galactosidase activity was evaluated. *p < 0.05, as determined by Dunnett's post hoc tests compared to untreated MW2 Since the braR gene in strain SAN87 had no mutations, we believe that BraR is phosphorylated even in the absence of nisin A in the SAN87 strain. Figure 9b shows an amino acid sequence alignment of sensor proteins that were reported to be associated with nisin Although they did not investigate the expression of VraDE, the mutation of BraS may allow a conformational change to occur that mimics the phosphorylated BraS protein without the need for nisin A stimulation.
F I G U R E 9 Protein alignments of BraR with other proteins and amino acid sequence of BraS. Protein alignment of BraR with other response regulators exhibiting homology with BraR (a). The active site aspartate residue (arrow), the two switch residues (dashed arrows), and the dimer interface regions (shown in the box and the dashed box) are shown. The triangle represents the mutation site in the mutant. Protein sequence of BraS (b). The dashed underline, double underline, and underline represent the region for the membrane-spanning region, histidine kinase domain, and ATPase domain region, respectively. The active site histidine residue (the arrow), the mutation site in the SAN87 strain (black triangle), and the mutation sites reported previously (white triangle) are shown In conclusion, we obtained three spontaneous S. aureus MW2 mutants with high levels of nisin A resistance by exposing cells to a sub-MIC of nisin A. Interestingly, the mutants harbored single point mutations in the braRS region that induced constitutive expression of the target gene without the need for environmental stimuli. Our findings also provide new insights into the key amino acids of BraRS required for nisin A resistance in S. aureus.

ACK N OWLED G EM ENT
This study was supported in part by Grants-in-Aid for Young Scientists (B) (Grant No: 15K11017, 18K09553) from the Ministry of Education, Culture, Sports, Sciences, and Technology of Japan.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no competing interests.

E TH I C S S TATEM ENT
None required.

DATA ACCE SS I B I LIT Y
All data are provided in full in the results section of this paper.