Contribution of hla Regulation by SaeR to Staphylococcus aureus USA300 Pathogenesis

The SaeRS two-component system in Staphylococcus aureus is critical for regulation of many virulence genes, including hla, which encodes alpha-toxin. However, the impact of regulation of alpha-toxin by Sae on S. aureus pathogenesis has not been directly addressed. Here, we mutated the SaeR-binding sequences in the hla regulatory region and determined the contribution of this mutation to hla expression and pathogenesis in strain USA300 JE2.

ABSTRACT

The SaeRS two-component system in Staphylococcus aureus is critical for regulation of many virulence genes, including hla, which encodes alpha-toxin. However, the impact of regulation of alpha-toxin by Sae on S. aureus pathogenesis has not been directly addressed. Here, we mutated the SaeR-binding sequences in the hla regulatory region and determined the contribution of this mutation to hla expression and pathogenesis in strain USA300 JE2. Western blot analyses revealed drastic reduction of alpha-toxin levels in the culture supernatants of SaeR-binding mutant in contrast to the marked alpha-toxin production in the wild type. The SaeR-binding mutation had no significant effect on alpha-toxin regulation by Agr, MgrA, and CcpA. In animal studies, we found that the SaeR-binding mutation did not contribute to USA300 JE2 pathogenesis using a rat infective endocarditis model. However, in a rat skin and soft tissue infection model, the abscesses on rats infected with the mutant were significantly smaller than the abscesses on those infected with the wild type but similar to the abscesses on those infected with a saeR mutant. These studies indicated that there is a direct effect of hla regulation by SaeR on pathogenesis but that the effect depends on the animal model used.

INTRODUCTION

Staphylococcus aureus is an important human pathogen capable of causing diseases ranging from superficial to life-threatening infections (1). The organism is endowed with a large repertoire of secretory and cell surface-associated virulence factors that enable it to infect virtually all human tissues (2). Depending on the environmental cues and infection stages, S. aureus is able to coordinate the expression of virulence factors through an orchestrated network of regulatory elements, including DNA-binding proteins (3, 4), regulatory RNAs (5, 6) and two-component systems (TCSs) (7, 8). It is believed that successful infections by S. aureus depend on appropriate expression of virulence factors. However, virulence gene regulation in S. aureus is very complex, as many of the regulators affect a large number of target genes and the regulons often overlap. In addition, many regulators affect one another. As such, it is very difficult to demonstrate the role of a specific regulatory pathway in staphylococcal pathogenesis.

The Sae system is one of several TCSs in S. aureus and plays an important role in regulating the transcription of many virulence genes under both in vivo and in vitro conditions (9–12). The sae operon consists of the saeP, saeQ, saeR, and saeS genes. The saeS and saeR genes encode a histidine kinase sensor and a response regulator, respectively (13). Upon sensing external stimuli such as human neutrophil peptide-1, SaeS autophosphorylates a conserved His residue from which the phosphate group is transferred to a conserved Asp residue of the response regulator (13, 14). The phosphorylated SaeR then regulates expression of its target genes, usually by binding to their promoter regions (10, 15). The role of the SaePQ complex is to activate the phosphatase activity of SaeS in order to fine-tune the TCS signal by dephosphorylating the response regulator (16).

Sae is one of the major regulators that have been shown to regulate the expression of alpha-toxin, a key virulence factor, by binding to the consensus SaeR-binding site upstream of the hla promoter (10, 15). The role of alpha-toxin in S. aureus pathogenesis has been extensively demonstrated by comparing isogenic hla mutants with their wild-type strains using various animal models (11, 17–20). However, like many staphylococcal global regulators, Sae also regulates many other virulence genes (8, 9, 13). Conversely, many virulence genes, including hla, can also be regulated by multiple regulators (5, 10, 11, 21).

Because SaeR activates hla transcription through a well-defined SaeR-binding site, we envision that this specific regulatory pathway could be used to assess its impact on bacterial pathogenesis experimentally using appropriate animal models. To this end, we mutated the SaeR-binding sequences upstream of the hla promoter and demonstrated that regulation of hla expression by Sae contributed to pathogenesis in strain USA300 JE2 in a rat skin and soft tissue infection (SSTI) model but not in an infective endocarditis (IE) model.

RESULTS

Construction of SaeR-binding mutant.

Sae has been shown to be a major regulator for hla transcription (11, 13, 22). It activates hla expression through its response regulator, SaeR, by binding to a direct repeat consensus sequence located 22 bp upstream of the −35 region of the hla promoter (15). As such, disruption of the SaeR-binding site should specifically abolish SaeR activation of hla transcription, allowing us to assess the contribution of this specific regulation to virulence. Accordingly, we mutated this SaeR-binding site in the JE2 chromosome by allelic replacement (Fig. 1A). To ensure that this mutation, which we named the sbm mutation, would abolish SaeR binding, we employed the competitive electrophoretic mobility shift assay (EMSA). However, our initial experiment using purified SaeR-His₆ was not successful in EMSAs (not shown). It was shown previously that phosphorylation of SaeR protein by the sensor kinase SaeS is essential for its DNA-binding activity (15, 23). We therefore attempted to purify SaeS-His₆ for in vitro phosphorylation of SaeR. The latter approach also failed, as expressing SaeS-His₆ or its N-terminal half in the Escherichia coli BL21 Star(DE3) strain resulted in the presence of recombinant proteins in inclusion bodies. To circumvent this outcome, we coexpressed SaeS-His₆ and SaeR-His₆ in the E. coli Rosetta 2(DE3) strain in the hope that phosphorylation of SaeR by SaeS could occur in vivo. This is likely, as a low level of SaeR phosphorylation is sufficient to activate the hla gene (23). After cell lysis, we found that almost all of the SaeS-His₆ protein remained in inclusion bodies whereas about 10% to 20% of the SaeR-His₆ was soluble. The soluble fraction of SaeR-His₆ was then purified by the use of HisPur Cobalt resin and used for the EMSA. A shift in EMSA was detected when a 1.8 μM concentration of the cobalt-purified SaeR-His₆ (presumably phosphorylated by SaeS-His₆) was used (Fig. 1B), but no shift was detected with SaeR-His₆ protein that was not coexpressed with SaeS-His₆ even at 4 μM (data not shown). The amount of SaeR required for shifting the DNA fragment was comparable to that previously reported (15). We found that the shifted band detected at a 2.7 μM concentration of SaeR could be competed away with 250-fold excess of the unlabeled wild-type 44-mer but not with the 44-mer containing the mutated sequence (Fig. 1B), suggesting that SaeR binding to the consensus SaeR-binding site is abolished by the sbm mutation. These results were further confirmed by using the 44-mer containing the sbm mutation as a probe in a repeated EMSA in which we detected no SaeR binding to the sbm sequence (Fig. 1C). In contrast, the probe containing the wild-type sequence was readily bound by SaeR.

FIG 1.

Construction of the sbm SaeR-binding mutation. (A) The promoter region containing the SaeR-binding consensus is in bold. The nucleotide changes in the sbm mutant are in red. (B) EMSA was performed using 40 fmol DIG-labeled wild-type 44-mer with increasing amounts of SaeR-His₆. Competition was done with 250-fold excess unlabeled sbm mutant 44-mer (mt) or with unlabeled wild-type 44-mer (wt). (C) EMSA was performed using 40 fmol DIG-labeled wild-type 44-mer or sbm 44-mer with increasing amounts of SaeR-His₆.

Mutation of SaeR-binding site on a hla promoter drastically reduces alpha-toxin production.

To assess the effect of abrogation of SaeR binding to the hla promoter region on Hla production, we compared the levels of Hla production by the wild type and the sbm mutant by Western blotting. A hla::bursa mutant and a saeR::bursa mutant were included as controls. In addition, as the sbm mutant cannot be complemented by plasmid, we also included a revertant mutant (CYLA308) to rule out the possibility that inadvertent secondary mutations might have occurred during the mutant construction. The Western analysis (Fig. 2) results revealed a drastic reduction of alpha-toxin production by the sbm mutant compared to the wild type. The revertant had a level comparable to that seen with the wild type, whereas the hla::bursa mutant had an undetectable level. These results suggest that SaeR is a primary activator of Hla production in strain JE2. In addition, the results presented in Fig. 2 revealed that the sbm mutant had produced an amount of alpha-toxin similar to that produced by the saeR::bursa mutant, indicating that SaeR controls Hla through binding to the SaeR-binding site to exert its effect at the transcriptional level.

FIG 2.

Western blot analysis of alpha-toxin production. Supernatants of 6-h cultures were analyzed by SDS-PAGE and blotted against anti-alpha-toxin antibody.

SaeR-binding mutation predominately affects sae regulation of hla.

Alpha-toxin is regulated by multiple regulators. Other than SaeR, hla transcription has been shown to be directly activated by MgrA and CcpA (24, 25). In addition, RNAIII of the agr system has been shown to affect hla through Rot and MgrA (26, 27). To determine whether AgrA or MgrA contributes to the reduction of alpha-toxin production in the sbm mutant, we first compared the levels of hla transcription in agr, mgrA, saeR, and sbm mutants by reverse transcription-quantitative PCR (qRT-PCR) on RNA samples isolated from early-stationary-phase cultures, as hla expression peaks during this growth phase. The results presented in Fig. 3A showed that the levels of transcription of hla in the sbm mutant and the saeR::bursa mutant were similarly reduced by 101-fold whereas 1.8-fold reduction was found in the mgrA::cat single mutant and 2.4-fold reduction in the agrA::bursa single mutant. The small effects of agrA and mgrA mutations are consistent with the previous reports although the effect of MgrA on hla transcription has previously been shown to be strain dependent (24, 27–29). These results indicate that SaeR is the main transcriptional regulator of hla by binding to the SaeR-binding site and that MgrA and AgrA have a minor effect. We then compared the hla transcripts of the sbm-agrA::bursa and the sbm-mgrA::cat double mutants to those of the sbm single mutant. We found a significant reduction (P = 0.0406) in the sbm-agrA::bursa double mutant compared to the sbm mutant, suggesting that the sbm mutation does not affect the effect of AgrA on hla. However, we found no significant difference between the sbm mutant and the sbm-mgrA::cat double mutant. It is likely that our method was not sensitive enough to detect the small effect of the mgrA mutation in the sbm mutant background. To further test whether the sbm mutation could affect MgrA binding, we determined whether the sbm mutation affected MgrA binding using the 44-mer oligonucleotide containing either the wild-type SaeR-binding site or the sbm mutation. As shown in Fig. 4, MgrA bound similarly to both of the 44-mer oligonucleotides, suggesting that the sbm mutation does not affect MgrA binding to the hla promoter region.

FIG 3.

Effect of the sbm SaeR-binding mutation on hla regulation by (A) Agr or MgrA (B) CcpA. Mean relative hla expression levels determined by qRT-PCR on early-stationary-phase cultures are presented with standard errors. Data represent results from three independent experiments. Statistical significance was analyzed by Student's t test. *, P < 0.05; ns, not significant.

FIG 4.

DNA binding of MgrA to hla regulatory region. EMSA was performed using 40 fmol DIG-labeled wild-type (wt) 44-mer or sbm 44-mer with increasing amounts of MgrA-His₆.

Next, we tested whether the sbm mutation affected regulation of hla by CcpA by the same strategy as described above using growth medium with and without glucose, as CcpA is involved in glucose-mediated catabolite repression. However, it has been reported that the presence of mild acid due to glucose fermentation during bacterial growth has a broad effect on expression of virulence genes, including hla (30, 61). To avoid the pH effect, we employed buffered medium using 50 mM HEPES as described previously by Seidl et al. (32), who showed that a pH drop resulting from glucose fermentation could be effectively buffered by adding 50 mM HEPES (pH 7.5) to the growth media without affecting the bacterial growth kinetics (32). Using this buffered medium, we assessed hla transcripts by qRT-PCR in the absence or presence of 10 mM glucose. The results presented in Fig. 3B showed that the ccpA::tet mutation decreased hla expression by 2.4-fold in the absence of glucose whereas the sbm mutation and the sbm-ccpA double mutation reduced hla expression by 247-fold and 294-fold, respectively. However, the difference between the sbm mutant and sbm-ccpA double mutant was not statistically significant. No CcpA effect was detected in the presence of glucose, and both the sbm single mutant and the sbm-ccpA::tet double mutant showed an approximately 100-fold reduction. Although our results did not rule out the possibility that the sbm mutation affects CcpA regulation of hla, we contend that this possibility is unlikely because the binding sites of CcpA and SaeR in the hla regulatory region do not overlap—one is located downstream of the promoter, and the other is located upstream (15, 25). Furthermore, the sbm mutation has a much stronger effect than that of CcpA on hla. As indicated above, the small additive effect may not be detected by our method.

It should be noted here that an earlier report showed that CcpA had a slight positive effect on hla in the absence of glucose at the early stationary phase but that the presence of glucose shifted the effect from the early stationary phase to the stationary phase (32). In the experiments whose results are presented in Fig. 3B, the bacterial cells were harvested at the early stationary phase. Thus, our findings indicating that the effect of CcpA on hla was detected in the absence but not in the presence of glucose are consistent with those reported previously by Seidl et al. (32).

SaeR regulation of hla impacts S. aureus USA300 JE2 pathogenesis in a rat model of SSTI.

Both Hla and Sae have independently been shown to contribute to USA300 pathogenesis in mouse and rabbit soft tissue infection models (11, 17, 31, 33). Our in vitro results described above showed that a SaeR-binding mutation had the same strong effect on Hla production and hla transcription as the saeR::bursa mutation, suggesting that Sae is a primary regulator of alpha-toxin in JE2. To examine whether activation of Hla by SaeR DNA binding has an impact on JE2 pathogenesis, we employed two different animal models of infection in rats, the IE and SSTI models, as described in Materials and Methods.

For the IE model, a group of 5 rats were surgically implanted with catheters to induce heart valve damage in the left ventricle before infection with either wild-type JE2 or the sbm mutant via tail vein was performed. After day 3 postinfection, rats were sacrificed and bacteria in the heart valves and heart tissues were examined. No statistically significant difference between the results seen with JE2 and the sbm mutant was found (results not shown), suggesting that regulation of hla by SaeR had no impact on S. aureus pathogenesis in this model under our experimental conditions.

For the SSTI model, a group of 10 rats were infected subcutaneously in the flanks with JE2, the sbm mutant, or the saeR::bursa mutant for 12 days. No obvious dermonecrosis was observed throughout the 12-day infection period. The absence of dermonecrosis, which is often observed in mouse models, could have been a consequence of the use of rats, rather than mice, in this study. The sizes of abscesses were measured daily. As shown in Fig. 5, the abscesses reached maximal size at day 2 postinfection for the wild type and the sbm mutant, whereas the abscesses seen with the sarR::bursa mutant continued to reduce in size, having begun to do so on day 1. During the 12-day period, the abscesses on rats infected with the sbm mutant were significantly smaller from day 1 to 7 than the abscesses on rats infected with the wild type. Similarly, significantly smaller abscesses were found in the saeR::bursa mutants than in the wild type (day 1 to 8). Although consistently smaller abscesses were found in the saeR::bursa mutant than in the sbm mutant, the differences were not statistically significant. At day 12 postinfection, rats were sacrificed and bacterial loads of abscesses were determined. The bacterial counts of rats infected with the sbm mutant were not significantly lower than the counts measured for those infected with the wild type but were significantly higher (P = 0.0044) than the counts measured for those infected with the saeR::bursa mutant (Fig. 6A). To assess the bacterial loads during infection, an additional group of 10 rats were infected and sacrificed at day 2 postinfection, when the abscesses infected with the wild type reached the peak volume. No statistically significant differences among the wild type, the sbm mutant, and the saeR::bursa mutant were found (Fig. 6B). Taken together, these results suggest that the regulation of hla by SaeR had a significant impact on S. aureus pathogenesis in the rat SSTI model.

FIG 5.

Contribution of hla regulation by SaeR to pathogenesis of USA300 SSTI in rats. Mean values of abscess sizes (measured daily) are presented with standard errors. *, P < 0.05 (versus wild-type JE2 using two-way ANOVA and Bonferroni’s post hoc comparisons).

FIG 6.

Bacterial loads of SSTI postinfection at day 12 (A) and day 2 (B). Results are expressed as mean values with standard errors. Statistical significance was analyzed by Student's t test. ns, not significant.

DISCUSSION

S. aureus virulence factors are regulated by a large number of regulators, many of which have pleiotropic effects. Studies have shown that most virulence factors are affected by multiple regulators (2, 8, 34). Globally, the regulatory pathways form a complicated network of virulence regulation in which examples of cross talk among the pathways are common. Thus, the conventional knockout approach to study the role of a regulator in pathogenesis can address only the combined regulon effects and not the specific regulatory effects of individual virulence factors. In this context, we chose the SaeR regulation of hla to test a novel approach to study the impact of virulence regulation on pathogenesis. Our results show that SaeR-hla regulation had a major impact on staphylococcal pathogenesis in a rat model of SSTI but not in a rat model of IE.

Our approach was to interrupt the specific regulation of hla by SaeR so that we could test the effect of this regulation on S. aureus pathogenesis in an appropriate animal model. The rationale was based on a previous report showing that SaeR affects hla transcription through a well-defined DNA-binding site (15). To demonstrate that SaeR indeed affects hla expression through interaction with the DNA-binding site, we first constructed a mutation in which 8 of the 10 nucleotides in the DNA binding consensus sequence were altered. We showed that the binding site mutant had an effect on Hla protein and hla transcription comparable to that exerted by the saeR::bursa mutant (Fig. 2; see also Fig. 3), indicating that the DNA binding pathway is likely the only pathway by which SaeR activates hla. The mutation was further verified by phenotypic analysis of the revertant.

In addition to SaeR, Hla production has been reported to be upregulated by CcpA, MgrA, and Agr (21, 24, 25). By analyzing sbm double mutants with each of these regulators, we showed that the sbm mutation did not affect the effect of AgrA regulation on hla. However, we were unable to conclusively determine whether the sbm mutation affects hla regulation by MgrA or CcpA. In the case of MgrA, we found its binding to the hla promoter region was not altered by the sbm mutation, suggesting that MgrA regulation of hla is not affected. As for CcpA, it is unlikely that the sbm mutation affects CcpA regulation, as CcpA-binding sites and SaeR-binding sites are separated by the hla promoter. On the basis of these analyses and the finding that the sbm mutation has a very strong effect on hla transcription, we conclude that regulation of hla by these regulators has little or no impact on the effect of hla by the sbm mutation.

It should be mentioned here that SarZ has also been shown to activate hla transcription by direct binding to the hla regulatory region in a heavily mutagenized strain, RN4220 (35). However, subsequent gene profiling study using strain Newman found no effect of hla transcription (36). In addition, SarA has been shown to activate hla by direct binding to an AT-rich motif in the hla regulatory region (37). However, it was later shown that the SarA effect on hla transcription was strain dependent and that it occurred largely at the posttranslational level through the activity of proteases (38). We therefore did not include SarZ and SarA in our study.

A number of regulators, including SarT, SarS, KdpE, and CodY, have been shown to repress hla transcription, likely by direct DNA binding (39–42). The binding sites of these regulators in the hla regulatory region have not been defined. Since the sbm mutant and the saeR mutant both strongly reduced hla expression and did so at about the same level, it is unlikely that the SaeR-binding mutation has a significant effect on the regulation by these negative regulators. Thus, these regulators were not included in our study.

In this study, we used the USA300 JE2 strain to demonstrate the effect of hla regulation by SaeR on the pathogenesis of S. aureus. USA300 strains produce a high level of toxins correlating with their high virulence, which has been associated with high levels of agr activity in these strains (43–45). Since RNAIII of the agr system has been shown to affect Hla at the translational level by interacting with sequences in the untranslated region of hla mRNA (21), our initial plan was to include RNAIII regulation of Hla production in our study. However, we found alteration of the RNAIII interaction site did not significantly affect Hla production by Western blot analyses (data not shown), suggesting that translational regulation of Hla by RNAIII may not be critical for Hla production in JE2. These results also suggest that regulation of hla by agr is mostly mediated by hla transcription indirectly through intermediary regulators such as Rot or MgrA (26, 27). However, we found agr deletion had a much lower effect than sae deletion in vitro (Fig. 3), suggesting that Hla is primarily regulated by SaeR in JE2.

Staphylococcal alpha-toxin is one of most extensively studied virulence factors. The role of alpha-toxin in invasive and superficial infections has been studied in different animal models (17, 18, 20, 31, 33, 46). Here we tested the effect of hla regulation by SaeR in two rat infection models. Previously, we showed that hla expression increased in vivo in the same rat IE model that we employed here (29). In addition, an increased level of hla expression in murine heart tissue has been reported previously (47). However, to our surprise, our results indicated that SaeR regulation of hla was not critical in a rat IE model. Although we did not test the hla deletion in this model, our results were consistent with a study showing that hla mutations did not impact the bacterial load in a rabbit IE model (48). On the other hand, the positive role of alpha-toxin in murine SSTI infection has been well established (17, 49). It is also known that USA300 is the major lineage causing skin infection in which Hla and other toxins are highly expressed compared to other clinical lineages (50). Our finding that SaeR regulation of hla had an impact on skin abscess sizes implies that Sae regulation of hla transcription plays a role in rat SSTI. This further underscores the previously reported findings indicating that sae genes are highly expressed in vivo (29, 51, 52). Although we showed that regulation of hla by SaeR was important for abscess size, our results showed no difference between the sbm mutant and the wild type in bacterial load. As a comparison, the saeR::bursa insertion mutant did show a significantly lower bacterial load than the wild-type strain and the sbm mutant at day 12 postinfection. This difference may be attributable to the fact that SaeR is a major regulator for many other virulence factors (53, 54). Furthermore, as the size of an abscess is directly proportional to the level of the host immune response, the differences between the saeR::bursa mutant and the sbm mutant also suggest that alpha-toxin is involved in promoting host immune response whereas other SaeR regulated virulence factors may be more extensively involved in bacterial growth and persistence in the abscesses.

MATERIALS AND METHODS

Bacterial and culture conditions.

Strains and plasmids used in this study are listed in Table 1. NE1354, NE1622, and NE1532 were obtained from the Nebraska transposon library collection (55) distributed by BEI Resources (https://www.beiresources.org/) through the Network of Antimicrobial Resistance in S. aureus (NARSA) program. Culture conditions were essentially as described previously (56). For testing the effect of CcpA on hla expression by qRT-PCR, S. aureus strains were grown in 50 mM HEPES (pH 7.5)-buffered tryptic soy broth (TSB) with no glucose or with 10 mM glucose as previously described (32). Antibiotics were added to the culture medium when necessary at final concentrations of 10 μg/ml for erythromycin, 10 μg/ml for chloramphenicol, 3 μg/ml for tetracycline, 100 μg/ml for ampicillin, and 30 μg/ml for kanamycin.

TABLE 1.

Strain and plasmids used in this study

Strain or plasmid	Description	Source or reference
Strains
S. aureus JE2	USA300 strain	NARSA
S. aureus CYLA207	JE2 SaeR-binding mutant (sbm) on hla promoter	This study
S. aureus CYLA270	JE2 ΔmgrA::cat	This study
S. aureus CYL1040	Becker ΔmgrA::cat	28
S. aureus NE1354	USA300FPR3757 hla::bursa	NARSA
S. aureus NE1532	USA300FPR3757 agrA::bursa	NARSA
S. aureus NE1622	USA300FPR3757 saeR::bursa	NARSA
S. aureus CYLA393	JE2 hla::bursa	This study
S. aureus CYLA498	JE2 saeR::bursa	This study
S. aureus CYLA537	JE2 sbm agrA::bursa	This study
S. aureus CYLA544	JE2 sbm ΔmgrA::cat	This study
S. aureus CYLA567	JE2 agrA::bursa	This study
S. aureus CYLA538	JE2 sbm ΔccpA::tet(L)	This study
S. aureus CYLA588	JE2 ΔccpA::tet	This study
S. aureus KS30	NM143 ΔccpA::tet(L)	32
S. aureus CYLA308	JE2 sbm revertant	This study
S. aureus RN4220	Restriction-negative laboratory strain	J. Iandolo
E. coli XL1-Blue	Host strain used for cloning	Stratagene
E. coli Rosetta 2(DE3)	Host strain for protein purification	Novagen
E. coli BL21Star(DE3)	Host strain for protein purification	Invitrogen
E. coli BL21(DE3)	Host strain for protein purification	Novagen
Plasmids
pJB38	Allelic replacement vector	57
pET28a-SaeR	For SaeR purification	15
pMCSG19-SaeS	For in vivo phosphorylation of SaeR	16
pET15b-MgrA	For MgrA purification	This study

Strain construction.

Primers used in this study are listed in Table 2. To construct a SaeR-binding mutation (referred to here as the sbm mutation), the consensus SaeR-binding sequence, GTTAAN₆GTTAA, in the wild type was mutated to GCTCCN₆TCCGC by overlapping PCR using primers Hla1F, Hla2R, Hla3F, and Hla4R. The product of the overlapping PCR was cloned to pJB38 vector (57) at the EcoRI and SalI restriction sites. Allelic replacement was performed as previously described (57, 58). The resultant mutant was named CYLA207. A DNA fragment containing the sbm mutation in the CYLA207 chromosome was amplified by PCR using HlaseqF and HlaseqR, and the mutation was confirmed by sequencing the amplified PCR fragment. In addition, the sbm mutant was made to revert to the wild type by allelic replacement using pJB38 vector carrying the wild-type DNA fragment, which was amplified by Hla1F and Hal4R. The revertant, CYLA308, was confirmed by PCR and sequencing as described above. CYLA270 (JE2 ΔmgrA::cat) and CYLA544 (JE2 sbm ΔmgrA::cat) were constructed by the use of phage 52A from CYL1040. CYLA393 (JE2 hla::bursa), CYLA498 (JE2 saeR::bursa), CYLA567 (JE2 agrA::bursa), and CYLA537 (JE2 sbm agrA::bursa) were transduced from the corresponding transposon mutants by the use of phage 52A. CYLA538 (JE2 sbm ΔccpA::tet) and CYLA588 (JE2 ΔccpA::tet) were transduced from KS30 by the use of phage 80α.

TABLE 2.

Primers and probes used in this study

Primer	Sequence (5′–3′)a
Hla1F	GATGATATCGAATTCGTAGAGCCTAAGACATTGATTTATTATG
Hla2R	CTATTAAATAAAAAGCGGATATATAGGAGCTAGTGTAAATTAATAAATAGAAATAGAG
Hla3F	TTTATTAATTTACACTAGCTCCTATATATCCGCTTTTTATTTAATAGTTAATTAATTG
Hla4R	GATGATATCGTCGACTATAAACTCTATATTGACCAGCAATG
HlaseqF	AACATAATTAATACCCTTTTTCTC
HlaseqR	AGTCAAAGTTGAATAAATTCTTTATG
HlaS-RT	GAACCCGGTATATGGCAATCA
HlaR-RT	GAGAACTTGCTTTGTTAGGATCAAG
SGhu2	ATCCAAAACTCACTTGCTAAAGG
SGhu3	ACCAGCTTTGAATGCTGGAAC
Wt-oligoF	ATTAATTTACACTAGTTAATATATAGTTAATTTTTATTTAATAG
Wt-oligoR	CTATTAAATAAAAATTAACTATATATTAACTAGTGTAAATTAAT
Mt-oligoF	ATTAATTTACACTAGCTCCTATATATCCGCTTTTTATTTAATAG
Mt-oligoR	CTATTAAATAAAAAGCGGATATATAGGAGCTAGTGTAAATTAAT

^aUnderlined characters indicate restriction sites.

Western blotting.

Culture was grown overnight and was then diluted to optical density at 600 nm (OD₆₀₀) of 0.05 in TSB and grown at 37°C with shaking at 225 rpm for 6 h. The culture was standardized by adjusting the optical density of each strain before harvesting the supernatants. The relative amounts of alpha-toxin in the supernatant were determined by Western blot using polyclonal anti-alpha-toxin antibody (Sigma-Aldrich, St. Louis, MO) and horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG secondary antibody (Sigma-Aldrich). The blot was developed using a ChemiDoc XRS+ system (Bio-Rad Laboratories, Hercules, CA) and chemiluminescent HRP substrate (Millipore, Billerica, MA).

RNA isolation and reverse transcription-quantitative PCR.

RNA samples were isolated from early-stationary-phase cultures (OD₆₀₀, ∼3.00) using RNAzol-RT (Molecular Research Center, Inc., Cincinnati, OH) as described previously (59). One-step RT-quantitative PCR was performed using one-step TB Green PrimeScript RT-PCR kit II master mix (TaKaRa Bio Inc., Kusatsu, Japan), a QuantStudio 6 Flex real-time PCR system (Applied Biosystems, Foster City, CA), and HlaRT-F and HlaRT-R primers (Table 2) for the alpha-toxin-encoding hla gene. The hu gene, amplified by primers SGhu2 and SGhu3 (Table 2), was used for normalization. The relative expression levels of hla were analyzed by the comparative threshold cycle (ΔΔC_T) method as previously described (60).

Protein expression and purification and electrophoretic mobility shift assay (EMSA).

E. coli Rosetta 2(DE3)pLysS containing pET28a-SaeR and pMCSG19-SaeS was used to coexpress SaeR-His₆ and SaeS-His₆ as described previously by Sun et al. (15). MgrA-His₆ was expressed using pET15b vector (Novagen) and E. coli BL21(DE3)pLysS. The His-tagged proteins were purified by the use of HisPur Cobalt resin (Thermo Fisher Scientific, Rockford, IL) per the guidelines from the manufacturer. The purified proteins were used in EMSAs. To construct the 44-mer oligonucleotide probe harboring the consensus SaeR-binding sequence, Wt-oligoF and Wt-oligoR (Table 2) were first annealed and then labeled at the 3′ end with digoxigenin (DIG)-11-ddUTP per the guidelines from the manufacturer (Roche Applied Sciences, Indianapolis, IN). The 44-mer containing the sbm mutation was constructed by annealing Mt-oligoF and Mt-oligoR (Table 2) and was labeled with digoxigenin as described above. Conditions for EMSA were essentially as described previously (15). The DIG-labeled DNA was detected by the use of a chemiluminescent detection system using reagents purchased from Roche Applied Sciences.

Animal infection models.

The protocols for animal experiments were reviewed and approved by the institutional animal care and use committee at the University of Arkansas for Medical Sciences (UAMS). Sprague-Dawley rats (Charles River Laboratories, Willington, MA) were housed in Institutional Animal Care and Use Committee-accredited animal care facilities at the UAMS. For the SSTI model, 10 rats (5 females and 5 males) were anesthetized with isoflurane and inoculated subcutaneously with 10⁷ CFU S. aureus in 0.1 ml DBPS (Dulbecco's phosphate-buffered saline [HyClone Laboratories Inc., South Logan, UT]) into both sides of the shaved flanks. The size of the abscess was measured with a Vernier caliper. For bacterial counts, each whole abscess from the euthanized animals was aseptically removed from the infection site and homogenized either manually using a 1.5-ml-pellet pestle (Nalge Nunc International, Naperville, IL) or by the use of a Polytron homogenizer (Kinematica AG, Littau/Lucerne, Switzerland). The homogenates were diluted and plated on tryptic soy agar (TSA) followed by overnight incubation at 37°C. The IE model experiments were performed as described previously (29). Briefly, polyethylene tubing (Braintree Laboratories Inc., Braintree, MA) was positioned in the left ventricle of 5 rats followed by infection with 5 × 10³ CFU S. aureus in 0.5 ml DPBS via tail vein 48 h postcatheterization. Rats were euthanized 3 days postinfection, and cardiac vegetations were aseptically removed from the heart tissue and homogenized by the use of a 1.5-ml-pellet pestle. The homogenates were diluted and plated on TSA as described for the skin infection model.

Statistics.

Results of comparisons of means of abscess sizes were analyzed by two-way analysis of variance (ANOVA) and Bonferroni’s posttest (Fig. 5). Paired Student's t tests were used to compare quantitative RT-PCR data (Fig. 3) and bacterial loads (Fig. 6). Analyses were performed using GraphPad Prism version 6.05 (San Diego, CA).