Three Distinct Two-Component Systems Are Involved in Resistance to the Class I Bacteriocins, Nukacin ISK-1 and Nisin A, in Staphylococcus aureus

Staphylococcus aureus uses two-component systems (TCSs) to adapt to stressful environmental conditions. To colonize a host, S. aureus must resist bacteriocins produced by commensal bacteria. In a comprehensive analysis using individual TCS inactivation mutants, the inactivation of two TCSs, graRS and braRS, significantly increased the susceptibility to the class I bacteriocins, nukacin ISK-1 and nisin A, and inactivation of vraSR slightly increased the susceptibility to nukacin ISK-1. In addition, two ABC transporters (BraAB and VraDE) regulated by BraRS and one transporter (VraFG) regulated by GraRS were associated with resistance to nukacin ISK-1 and nisin A. We investigated the role of these three TCSs of S. aureus in co-culture with S. warneri, which produces nukacin ISK-1, and Lactococcus lactis, which produces nisin A. When co-cultured with S. warneri or L. lactis, the braRS mutant showed a significant decrease in its population compared with the wild-type, whereas the graRS and vraSR mutants showed slight decreases. Expression of vraDE was elevated significantly in S. aureus co-cultured with nisin A/nukacin ISK-1-producing strains. These results suggest that three distinct TCSs are involved in the resistance to nisin A and nukacin ISK-1. Additionally, braRS and its related transporters played a central role in S. aureus survival in co-culture with the strains producing nisin A and nukacin ISK-1.


Introduction
Many bacteria produce antibacterial agents, called bacteriocins, which interfere with other bacteria in the bacterial community [1]. Bacteriocins are peptides or proteins that are ribosomally synthesized and show antimicrobial activity, mostly against bacterial species that are closely related to the producers [2]. In gram-positive bacteria, bacteriocins are classified into two major types, classes I and II bacteriocins [3,4,5]. Class I bacteriocins (,5-kDa peptides) are called lantibiotics because they contain the unusual amino acids, lantionine and methyllanthionine, which are posttranslationally modified, whereas class II bacteriocins contain unmodified amino acids. Lantibiotics are further divided into two types (types A and B) [6]. Type A lantibiotics include two subtypes, type A(I) such as nisin A and type A(II) such as nukacin ISK-1. The mode of action of lantibiotics, especially nisin A, has been well characterized [7,8]. Nisin A exhibits pore-forming activity and the inhibition of cell wall biosynthesis. The docking molecule of nisin A is lipid II, which is a membrane component consisting of one GlcNAc-MurNAc pentapeptide subunit linked to a polyisoprenoid and is associated with peptidoglycan biosynthesis in the membrane [9]. Nukacin ISK-1 also binds to lipid II [10]. However, the mode of action of nukacin ISK-1 is not poreforming, but the inhibition of cell wall synthesis causes a bacteriostatic effect [10,11]. These bacteriocins are considered to affect other bacterial populations. In addition, some bacteriocins, such as nisin A, are used as preservatives for foods and other surfaces [1].
Staphylococcus aureus is a major pathogen in humans that can cause a variety of suppurative diseases, food poisoning, and toxic shock syndrome [12,13,14]. Furthermore, clinically isolated strains, particularly methicillin-resistant S. aureus (MRSA), exhibit multiple antibiotic resistances [15,16], resulting in serious problems with regard to therapy against S. aureus infectious diseases. Several two-component systems (TCSs) such as vraSR, agrCA, and graRS/apsRS, were recently reported as being associated with susceptibility to antibacterial agents [17,18,19,20]. A TCS, which is thought to function as a monitor and adapt to specific environmental conditions [21], is a prokaryote-specific signal transduction system that contains a sensor that encodes a sensory histidine-kinase and a regulator that encodes a cognate response regulator (RR) [22]. Recently, we and others identified one TCS, called BceRS [23], BraRS [24] or NsaRS [25], which affects susceptibility to bacitracin. In addition, this TCS is associated with resistance to nisin A [24,25,26]. Because this TCS was designated separately by three different groups, we used the name BraRS in this study because Hiron et al well characterized this TCS and the name (bacitracin resistance associated) is representative of its characteristics [24]. These findings suggest that S. aureus has several systems for resisting bacteriocins. Given that S. aureus is a commensal bacterium in the nasal cavity, skin, and intestine, this organism is faced with many other bacterial species, including other staphylococci such as S. epidermidis and S. warneri [27,28]; thus it is considered that S. aureus must resist bacteriocins to survive when it co-exists with bacteriocin-producing bacteria. Herein, we investigated the association of TCSs with susceptibility to the class I bacteriocin, nukacin ISK-1.

Bacterial Strains and Growth Conditions
The bacterial strains used in this study are listed in Tables 1 and  2. S. aureus inactivation mutants were constructed previously [23]. S. aureus and S. warneri were grown in trypticase soy broth (TSB; Becton Dickinson Microbiology Systems, Cockeysville, MD, USA) at 37uC. Escherichia coli XL-II was grown in Luria-Bertani (LB) broth at 37uC. Lactococcus lactis was grown in De Man, Rogosa, Sharpe (MRS) broth (Becton Dickinson Microbiology Systems) at 37uC. Tetracycline (TC; 5 mg/ml) or chloramphenicol (CP; 10 mg/ml) was added to S. aureus when necessary.

Evaluation of Bacteriocin Susceptibility
Two methods (the minimum inhibitory concentration [MIC] and direct methods) were used to evaluate susceptibility to bacteriocins. MICs of nisin A, nukacin ISK-1, and bacitracin were determined by micro-dilution method as described previously [29]. Nisin A [30] and nukacin ISK-1 [31] were purified as described elsewhere. MICs were determined after 10 h of incubation. Three independent experiments were performed.
In the direct method, modified from a previous method [32], 2 ml of an overnight culture of S. warneri and L. lactis were spotted on an MRS agar plate. After overnight incubation at 37uC, 5 ml of pre-warmed TSB soft agar (0.75%) containing wild-type S. aureus or the mutants at 10 6 cells/ml was poured over the TSB agar plate. Plates were incubated for 20 h at 37uC. We confirmed that the diameter of the producing colony was uniformly 7 mm among all strains. The diameter of the inhibition zones surrounding bacteriocin-producing strains was measured in three directions. Three independent experiments were performed for the direct method, and the average result of the three experiments was calculated. Statistical analysis was performed with Dunnett's method.

Effect of Nukacin ISK-1 and Nisin A on the Expression of TCSs and Transporters
A small portion (10 8 cells) of S. aureus cultured overnight was inoculated into 10 ml fresh TSB, and then grown at 37uC with shaking. When the optical density reached 0.5 at 660 nm, various concentrations of nukacin ISK-1or nisin A were added to the medium. After the appropriate incubation, bacterial cells were collected. Total RNA was extracted from the bacterial cells with a FastRNA Pro Blue kit (MP Biomedicals, Solon, OH, USA) in accordance with the manufacturer's protocol. A 1-mg aliquot of total RNA was reverse-transcribed to cDNA using a first-strand cDNA synthesis kit (Roche, Tokyo, Japan). Using cDNA as template, quantitative PCR was performed using a LightCycler system (Roche). Primers for braR, vraR, and graR (TCS), as well as for braA, vraD, and vraF (ABC transporter) were constructed and used to determine the optimal conditions for analysis of their expression, and gyrA was used as an internal control. Three independent experiments were performed, and the mean was calculated. Statistical analysis was performed with Dunnett's method. The primers are listed in Table S1.
Co-culture of S. aureus with S. warneri or L. lactis The method for the co-culture experiment is summarized schematically in Figure S1. Overnight cultures of S. aureus MM30 (MW2 harboring pCL8 [33]), S. warneri ISK-1, S. warneri ISK-1 -, L. lactis ATCC 11454, and L. lactis NZ9000 were adjusted to OD 660 = 1.0 and diluted ten-fold. Next, 100 ml of bacterial culture (S. aureus [10 7 , 10 6 , 10 5 , 10 4 cells] and S. warneri [10 7 cells], S. aureus [10 7 , 10 6 , 10 5 , 10 4 cells], and L. lactis [10 7 cells]) was mixed well. A 20-ml aliquot of the mixed culture was spotted on a 50% trypticase soy agar (TSA) plates. After overnight incubation at 37uC, the bacterial colonies growing on the agar plate were scraped and suspended in 1 ml of TSB. The appropriate dilutions were plated on TSA and TSA containing chloramphenicol (for selection of S. aureus). After 1 day, the colony-forming units (CFUs) grown on TSA and TSA containing antibiotics were determined, and we calculated the percent population of the S. aureus strain. We also extracted total RNA from scraped cells and performed cDNA synthesis using the method described above; gene expression analysis was conducted by quantitative PCR. The statistical Table 2. Bacteriocin-producing and non-producing strains used in this study.

Strain
Bacrteriocin Reference
Next, the co-culture of S. aureus TCS or ABC transporter mutants with S. warneri or L. lactis was investigated using the method described above. The concentrations of bacterial cells used in this assay were 10 7 cells/ml S. aureus mutant and 10 8 cells/ ml S. warneri or 10 6 cells/ml S. aureus mutant and 10 8 cells/ml L. lactis. For the co-culture of S. aureus with S. warneri, TSA containing chloramphenicol (wild-type S. aureus, 10 mg/ml) and tetracycline (S. aureus mutants, 10 mg/ml) were used for S. aureus selection.

Susceptibility of Bacitracin-treated S. aureus to Nukacin ISK-1 and Nisin A
To investigate whether the VraDE expression level affects susceptibility to nisin A and nukacin ISK-1, we evaluated the susceptibility of bacitracin-pretreated S. aureus to nisin A and nukacin ISK-1 using the MIC method, as described above, and the spot-on-lawn method described elsewhere [11]. Previously, we reported that bacitracin at a sub-MIC induced VraDE expression significantly [23]. S. aureus pretreated with or without bacitracin was used for both methods. S. aureus cells (10 9 /ml) were exposed to a sub-MIC (1/8 MIC: 8 mg/ml) of bacitracin (Sigma-Aldrich, Figure 1. Expression of TCSs and ABC transporters in S. aureus exposed to nukacin ISK-1. Analysis of expression levels of braR, graR, vraR, braA, vraD, and vraF were performed as described in the Materials and Methods. (A) braA, vraF, vraD, braR, graR, and vraR expression in S. aureus MW2 exposed to various concentrations of nukacin ISK-1 (5-min exposure). *, statistically significant difference from the wild-type as tested using Dunnett's method (p,0.05). (B) Time course experiment of braA, vraF, and vraD expression in S. aureus MW2 exposed to nukacin ISK-1 (4 mg/ml). (C) braA, vraF, and vraD expression in S. aureus MW2 and three mutants (braRS, graRS, and vraSR mutant) exposed to nukacin ISK-1 (4 mg/ml). *, statistically significant difference from the wild-type as tested using Dunnett's method (p,0.05). doi:10.1371/journal.pone.0069455.g001 Tokyo, Japan) for 30 min. In the spot-on-lawn method, 5 ml of nisin A (6.4 mg/ml) or nukacin ISK-1 (64 mg/ml) were spotted on a double-layered agar plate containing 8 ml of TSA soft agar with 10 6 /ml S. aureus cells as the upper layer and 10 ml of TSB agar (1.5%) as the bottom layer. After overnight incubation at 37uC, the diameter of the inhibition zone for bacterial growth was measured in three directions.
Additionally, we constructed the VraDE overexpression strain in the graRS mutant and investigated the susceptibility to nisin A and nukacin ISK-1. The gene coding vraDE was amplified with specific primers, and then the DNA fragment was cloned into pCL15, which harbored an E. coli-S. aureus shuttle vector with the Pspac promoter [34]. The expression of the cloned gene in pCL15 was significantly induced by the addition of 1 mM isopropyl b-D-1-thiogalactopyranoside (IPTG). The obtained plasmid was electroporated to S. aureus RN4220, and then the plasmid was transduced to the gra mutant using phage 80 alpha with the method described in a previous study [35]. Plasmids and primers used in the present study are listed in Table S1. Using this strain, susceptibility was investigated by the MIC method and the spoton-lawn method, as described above.

Susceptibility of TCS-inactivated Mutants to nisin A and Nukacin ISK-1
We determined the susceptibility of TCS mutants to nukacin ISK-1 using the MIC and the direct method (Table 3, Figure S2). The graRS and braRS mutants exhibited higher susceptibility to nukacin ISK-1, and the vraSR mutant exhibited higher susceptibility to nukacin ISK-1. The susceptibilities of other TCS mutants other than braRS, graRS and vraSR to nukacin ISK-1 did not change (data not shown). We also the evaluated the susceptibility of TCS mutants to nisin A and nukacin ISK-1 using the direct method ( Figure S2), and obtained similar results to those obtained using the MIC method. Furthermore, we determined the susceptibility of the mutants and their complemented strains. We found that each complemented strain could restore the respective mutation (Table 3). Also, we investigated the susceptibility of TCS mutants against nisin A, and found similar results to nukacin ISK-1 except that the susceptibility of the vraRS mutant to nisin A did not increase (Table 3, Figure S2).

Susceptibility of ABC Transporter-inactivated Mutants to nisin A and nukacin ISK-1
Previously, we demonstrated that BraRS regulates two transporters (VraDE and BraAB) for resistance against bacitracin, and that GraRS regulates one transporter (VraFG) [19,23]. We evaluated the susceptibility of these three mutants (vraDE, braAB, and vraFG mutants) to nukacin-ISK-1 (Table 3, Figure S2). The vraFG, vraDE, and braAB mutants were more susceptible to nukacin ISK-1. Additionally, we found that each complemented strain could restore the respective mutation (Table 3). Also, we investigated the susceptibilities of these mutants to nisin A, and obtained similar results to those for nukacin ISK-1 (Table 3, Figure S2).

Effect of Nukacin ISK-1 and Nisin A on the Expression of TCSs and Transporters
Given that the susceptibility to nisin A and nukacin ISK-1 was changed by inactivation of three TCSs (braRS, graRS, and vraSR) and three transporters (vraFG, vraDE, and braAB), we investigated the expressions of those TCSs and transporters upon exposure of bacterial cells to nukacin ISK-1 and nisin A.
Among the three TCSs, the expression levels of the braR and graR transcripts did not increase upon exposure to nukacin ISK-1, whereas the expression of vraR was induced ( Figure 1A). Regarding the three ABC transporters, the expression of braA and vraD in the wild-type MW2 strain was rapidly induced by the addition of nukacin ISK-1 to the medium, whereas vraF expression was not ( Figures 1A and 1B). This induction occurred after 5 min of exposure, after which the transcript levels of both transporters gradually decreased ( Figure 1B). The expressions of both braA and vraD were dose-dependent ( Figure 1A). However, the induction of vraD and braA expression by nukacin ISK-1 was not observed in the braRS mutant, although both genes showed increased expression in the graRS and vraSR mutants ( Figure 1C). In the graRS mutant, vraF expression decreased irrespective of nukacin ISK-1 addition, but in the other two mutants, the vraF expression level was not significantly different from that of the wild-type ( Figure 1C).
In addition, we investigated the expression of TCSs and transporters in the wild-type and its mutants by addition of nisin A. We obtained results similar to those of nukacin ISK-1, except that nisin A did not induce vraSR expression ( Figure S3).
Based on these results, we concluded that the expression of two ABC transporters, vraD and braA, was induced by nukacin ISK-1 and nisin A, and that this effect was mediated by one TCS, BraRS. Also, the expression of another transporter, vraF, was not induced, but vraF expression was regulated by GraRS.
Co-culture of S. aureus with S. warneri or L. lactis Co-cultures of S. aureus MM30 with S. warneri ISK-1, S. warneri ISK-1DpPI-1 (pPI-1 plasmid cured), L. lactis ATCC 11454 (nisin A-producing strain), and L. lactis NZ9000 (nisin A non-producing strain) were analyzed. When S. aureus MM30 was co-cultured with S. warneri ISK-1, which produces nukacin ISK-1, the population of the braRS mutant was significantly lower at any S. aureus/S. warneri ratio, compared to that of the wild-type (MM30) (Figure 2A). When S. aureus was co-cultured with S. warneri ISK-1DpPI-1, which does not produce nukacin ISK-1, the population of the braRS mutant was similar to that of the wild-type. We evaluated the expression of vraD and braA under co-culture conditions ( Figure 2B). The expression of both increased when S. aureus was co-cultured with S. warneri ISK-1 at various ratios. In particular, both increased gradually as the ratio of S. aureus to S. warneri decreased before spotting on the TSA plate.
S. aureus MM30 was co-cultured with L. lactis ATCC 11454, which produces nisin A, or NZ9000, which does not. Results were similar to those of nukacin ISK-1 ( Figure S4). Figure 3A shows the S. aureus population ratio when 10 6 cells of S. aureus mutants were mixed with 10 7 S. warneri ISK-1 cells. The braRS, braAB, and vraDE mutants exhibited drastically decreased population ratios compared with that of the wild-type. In addition, the graRS, vraSR, and vraFG mutants showed slight decreases compared with the wild-type.
Similar results were obtained when 10 5 cells of S. aureus mutants were mixed with 10 7 L. lactis cells ( Figure 3B). When co-cultured with the ATCC 11454 strain, the population ratios of the braRS, braAB, and vraDE mutants were significantly lowered, compared to that of the wild-type. Additionally, we investigated vraD expression in TCS mutants. The wild-type, graRS and vraSR mutants showed significantly increased vraD expression upon co-culture with bacteriocin-producing strains ( Figure 3C). However, vraD expression did not increase in the wild-type, graRS and vraSR mutants when co-cultured with bacteriocin-non-producing strains (data not shown).
From the results of co-culture assay, we found that the inactivation of one TCS (braRS) and two BraRS-regulated transporters (braAB and vraDE) caused a significant decrease in the S. aureus population when co-cultured with a nukacin ISK-1-or nisin A-producing strain. Susceptibility of Bacitracin-treated S. aureus to nisin A and nukacin ISK-1 Because vraDE expression of the S. aureus wild-type was induced by nisinA or nukacin ISK-1, we investigated whether the VraDEoverexpressing strain showed higher resistance to nisinA and nukacin ISK-1. We used S. aureus pretreated with a sub-MIC of bacitracin, which induced the expression of VraDE [23] but was not bactericidal. The wild-type strain pretreated with bacitracin showed an increased nisin A and nukacin ISK-1 MICs compared to that without bacitracin (Table 4). Also, the MICs of the graRS and vraRS mutants against nisin A and nukacin ISK-1 were increased by pretreatment with bacitracin, whereas the MIC of the braRS mutant did not change. In addition, we constructed a VraDE-overexpression strain using the pCL15 plasmid. VraDE overexpression in the graRS mutant caused decreased susceptibility to both nisin A and nukacin ISK-1 (Table 4).

Discussion
We performed a comprehensive analysis of the TCSs involved in the susceptibility to the class I bacteriocin, nukacin ISK-1, in S. aureus and identified several TCSs to be associated with nukacin ISK-1 susceptibility ( Figure S2, Table 3). Previously, BraRS and GraRS were also shown to be associated with bacitracin and nisin A resistance [23,24,25,36]. In addition, Hiron et al. demonstrated that BraRS is activated by nisin A, which induces the expression of transporters [24], as confirmed by the results in this study (Figure 1, Figure S3). Therefore, BraRS is involved in the resistance to nukacin ISK-1, nisin A and bacitracin. Bacitracin binds to Expression of braR, graR, vraR, braA, vraD, and vraF was determined using the method described in the Materials and Methods. (A) braA, vraF, vraD, braR, graR, and vraR expression in S. aureus MW2 exposed to various concentrations of nisin A (5 min exposure). *, statistically significant difference from the wild-type as tested using Dunnett's method (p,0.05). (B) Time course experiment of braA, vraF, and vraD expression in S. aureus MW2 exposed to nisin A (16 mg/ml). (C) braA, vraF, and vraD expression in S. aureus MW2 and three mutants (braRS, graRS, and vraSR mutant) exposed to nisin A (16 mg/ml). *, statistically significant difference from the wild-type as tested using Dunnett's method (p,0.05). doi:10.1371/journal.pone.0069455.t004 undecaprenyl pyrophosphate, resulting in inhibition of lipid II formation, whereas nisin A and nukacin ISK-1 bind to lipid II [9,10]. Vancomycin also binds to the D-alanine-D-alanine molecule in lipid II; however, the braRS mutant did not exhibit marked susceptibility to vancomycin [18]. Nisin A binds to the pyrophosphate moiety of lipid II, resulting in pore formation in the membrane [9]. Therefore, we propose that BraRS is associated with susceptibility to antibacterial agents related to the membraneanchoring region of lipid II. Recently, BraAB was found to act as a cofactor for BraRS but not as a direct resistance factor, suggesting that it is associated with the regulation of vraDE [24]. BraRS in S. aureus shows homology with TCSs of other gram-positive bacteria, including Enterococcus, Bacillus and Streptococcus [37,38,39,40]. Therefore, the BraRS system is widely conserved in gram-positive bacteria.
GraRS is involved in susceptibility to cationic peptides such as defensins, gentamicin, and vancomycin [41,42,43] because it regulates two factors, dlt and mprF (fmtC), both of which influence the cell surface charge of bacteria [36,42,43,44]. Inactivation of graRS causes an increase in the negative charge of the cell surface, resulting in increased attraction of the cationic peptides nisin A and nukacin ISK-1, but not bacitracin, to the cell membrane. In addition, vraFG (downstream of graRS) is associated with susceptibility to nisin A and nukacin ISK-1, but not bacitracin. Recently, Falord et al. demonstrated that VraFG did not act as a detoxification module but was associated with GraRS activation [45]. Therefore, the difference in susceptibilities between nisin A/ nukacin ISK-1 and bacitracin may be due to the charges of these peptides.
VraSR (vancomycin resistance associated sensor/regulator) was first identified as a factor responsible for vancomycin susceptibility [46]. Further investigations revealed that this TCS regulates many factors involved in cell wall biosynthesis and that it is associated with susceptibility to cell wall synthesis inhibitors including betalactams, cycloserine, teicoplanin, and bacitracin [47,48,49]. In this study, the inactivation of vraSR led to an increase in susceptibility to nukacin ISK-1, but not nisin A (Table 3 and Figure S2). Moreover, nukacin ISK-1 induced vraSR expression ( Figure 1A), whereas nisin A did not ( Figure S3). Although nisin A also exhibits an inhibitory effect on cell wall biosynthesis, we did not detect elevated vraSR expression under our conditions. Nukacin ISK-1 and nisin A are type A lantibiotics, but their subtypes (type A [I] and type A [II], respectively) and structure differ ( Figure S5). Also, their modes of action differ; nisin A exhibits a bactericidal effect by causing pore formation and the inhibition of cell wall biosynthesis [7,8], whereas nukacin ISK-1 acts as a bacteriostatic agent by inhibiting cell wall biosynthesis [11]. Therefore, we hypothesize that the different modes of action of these two bacteriocins reflect the different responses in terms of vraSR expression.
We evaluated the competition between two bacterial strains using a co-culture method. Notably, the inactivation of three TCSs, but especially graRS and braRS, caused significantly increased susceptibility to nisin A and nukacin ISK-1 by the direct and the MIC methods ( Figure S2 and Table 3); however, only one TCS, BraRS, was a major contributor to S. aureus survival in co-culture with S. warneri or L. lacti, which produces bacteriocin (Figures 2, 3 and S4). In particular, the graRS mutant showed different results between the direct and co-culture methods. We hypothesized that this difference in the graRS mutant was due to the different level of VraDE expression upon exposure of the mutant to nisin A and nukacin ISK-1. In the direct assay, S. aureus cells that expressed VraDE at a very low level were exposed to relatively high concentrations of nisin A and nukacin ISK-1. In the early period after bacteriocin exposure, VraDE expression was not sufficient for BraRS-mediated nisin A/nukacin ISK-1 resistance; thus the graRS mutation, which exhibited a more negatively charged cell surface [42,43], showed marked susceptibility to nisin A and nukacin ISK-1. Conversely, S. aureus cells in the co-culture are exposed to a low concentrations (non-lethal) of nisin A or nukacin ISK-1 during the early stage of co-culture, and so VraDE expression is induced. Upon exposure to a high concentration of nisin A or nukacin ISK-1 during further incubation, S. aureus expressed VraDE at a level sufficient for resistance. Table 4 reflects our hypothesis that pretreatment of the graRS mutant with bacitracin resulted in marked nisin A and nukacin ISK-1 resistance. Additionally, we obtained the same result when S. aureus cells were pretreated with nisin A (data not shown). Furthermore, similar results were obtained using the VraDE overexpression strain (Table 4). Therefore, the different results of the direct and co-culture methods using the graRS mutant were due to the different VraDE expression levels upon exposure of S. aureus to a high concentration of nisin A or nukacin ISK-1.
Based on our findings, we propose that BraRS and GraRS have distinct functions in terms of resistance to nisin A and nukacin ISK-1. BraRS is an intrinsic factor for such resistance. However, upon exposure of S. aureus to a relatively high level of nisin A or nukacin ISK-1, GraRS is important for resistance until significant induction of VraDE expression by BraRS occurs. Therefore, these two TCSs function coordinately in resistance to nisin A and nukacin ISK-1. Furthermore, in addition to these two TCSs, VraSR is independently activated upon inhibition of cell wall biosynthesis. Class I bacteriocins, such as nisin A and nukacin ISK-1 inhibit cell wall biosynthesis, although increased expression of VraSR was identified only in S. aureus cells exposed to nukacin ISK-1. Our results strongly indicate that S. aureus possesses three distinct class-I-bacteriocin-resistance systems.
In conclusion, we demonstrated that several TCSs and ABC transporters in S. aureus are associated with resistance to bacteriocins produced by other bacteria. Notably, S. aureus possesses multiple TCSs that resist nisin A and nukacin ISK-1 ( Figure 4). In particular, the BraRS system is specific for nisin A and nukacin ISK-1. Conversely, GraRS and VraSR confer broadspectrum resistance against cationic peptides and cell-wall synthesis inhibitors, respectively. Our findings suggest that S. aureus possesses several TCSs that facilitate its survival in complex bacterial communities. Figure S1 Method for the co-culture experiment. (TIF) Figure S2 Susceptibility of TCS-and ABC transporterinactivated mutants to nukacin ISK-1 and nisin A. The susceptibilities of S. aureus MW2 and its TCS-or ABC transportermutants to nukacin ISK-1 and nisin A were evaluated by the direct method. (A) In total, 2 ml of overnight cultures of bacteriocinproducing strains were spotted on an MRS agar plate. After overnight incubation at 37uC, pre-warmed MRS soft agar (0.75%) containing S. aureus was poured over the surface of the MRS agar plate. Plates were incubated for 20 h at 37uC. (B) The diameters of the inhibition zones surrounding the bacteriocin-producing strain were measured in three directions. Three experiments were performed independently, and the average result of the three experiments was calculated. *, statistically significant difference from the wild-type as tested using Dunnett's method (p,0.05). The error bar represents the standard deviation. (TIF) Figure S3 Expression of TCSs and ABC transporters in S. aureus exposed to nisin A.