The ability of sarA to limit protease production plays a key role in the pathogenesis of Staphylococcus aureus osteomyelitis irrespective of the functional status of agr

ABSTRACT

We demonstrate that mutation of the staphylococcal accessory regulator A (sarA) in the USA300 strain LAC limits virulence in a murine osteomyelitis model to a greater extent than mutation of the accessory gene regulator (agr) and that it does so irrespective of the functional status of agr. Protease production was decreased in the agr mutant but increased in sarA and sarA/agr mutants to a degree that limited biofilm formation. Extracellular protein A (eSpa) and full-length extracellular nuclease (Nuc1) were absent in the conditioned medium (CM) from sarA and sarA/agr mutants, and their abundance was restored in both mutants by eliminating protease production. Cytotoxicity of CM for osteoblasts and osteoclasts was also reduced in both mutants. Cytotoxicity was restored in a protease-deficient sarA mutant but not in the protease-deficient sarA/agr mutant. Reduced cytotoxicity was correlated with the reduced abundance of full-length α-toxin, LukF, and LukS in sarA and sarA/agr mutants. The abundance of these toxins in their full-length form was increased in the protease-deficient sarA mutant by comparison to LAC, demonstrating that mutation of sarA increases the production of these toxins but increased protease production limits their abundance in full-length and presumably functional forms. Most importantly, eliminating protease production enhanced the virulence of sarA and sarA/agr mutants, but had no impact in the agr mutant. We conclude that a key factor in the attenuation of LAC sarA and sarA/agr mutants in osteomyelitis is the increased production of extracellular proteases and its impact on virulence factors that contribute to biofilm formation and cytotoxicity.

IMPORTANCE

The persistent emergence of antibiotic-resistant strains has rekindled interest in anti-virulence strategies to combat S. aureus infections. Numerous reports describe anti-virulence strategies focusing on key regulatory elements that globally influence virulence factor production, the two most commonly targeted being the accessory gene regulator (agr) and the staphylococcal accessory regulator A (sarA). We demonstrate that mutation of sarA limits virulence to a greater extent than mutation of agr and that this can be attributed to increased protease production in both sarA and sarA/agr mutants. This illustrates the critical role of sarA in protease-mediated post-translational regulation in S. aureus. It also suggests that an inhibitor of sarA would be more effective than an inhibitor of agr in overcoming the therapeutic recalcitrance of osteomyelitis and that such an inhibitor would remain effective even in the context of agr mutants known to arise in vivo during the transition from acute to chronic infection.

INTRODUCTION

Staphylococcus aureus employs a complex and highly integrated regulatory system to control the production of its arsenal of virulence factors (1, 2). Two of the best studied regulatory loci at the center of this system are the accessory gene regulator (agr) and the staphylococcal accessory regulator (sarA). The agr system is a quorum-sensing system that modulates the production of AgrA and a regulatory RNA designated RNAIII, the accumulation of which defines a transition from the production of surface-associated virulence factors to the production of extracellular toxins and enzymes. This transition is evident in vitro as cultures go from the exponential to stationary growth phases (2). The effector molecule of the sarA system is a 15-kDa protein (SarA) that modulates the transcription of many S. aureus genes, including agr itself, with SarA being required for maximum expression of agr (1, 3).

Mutation of sarA results in increased protease production and decreased virulence in murine models of sepsis and osteomyelitis, and eliminating the ability of sarA mutants to produce extracellular proteases restores virulence to a statistically significant extent (4–9). Mutation of agr has the opposite effect on protease production but also limits virulence in these models (9, 10). This suggests that the impact of mutating agr on virulence is mediated at the level of protein production, while that of sarA is mediated at the level of protein degradation. It also suggests that the impact of mutating sarA on virulence is independent of agr. Nevertheless, the relative impact of agr and sarA on the pathogenesis of osteomyelitis remains unknown, particularly given that sarA is required for maximum expression of agr (1, 3, 11, 12). However, we are unaware of any studies that have made direct comparisons necessary to definitively assess the relative impact of the agr-dependent and agr-independent pathways of sarA-mediated regulation in the pathogenesis of osteomyelitis.

It is important to address this issue because sarA and agr are the two regulatory loci that have been evaluated most extensively as targets for prophylactic and therapeutic intervention in S. aureus infection (12–19). Spontaneous agr mutants arise in vivo and ultimately become the dominant subpopulation, and it has been proposed that this promotes the transition from acute to chronic infection (18–22). This suggests that inhibitors of agr might have the unintended consequence of promoting this transition. If the impact of sarA is agr-dependent, then inhibitors of sarA might have the same effect, while if the impact of sarA is agr-independent, this presumably would not be the case. Additionally, an effective inhibitor of sarA could retain its efficacy even in the context of these agr mutants.

In this report, we addressed the relative impact of sarA and agr on the pathogenesis of osteomyelitis by generating isogenic sarA, agr, and sarA/agr mutants in the USA300 strain LAC and assessing their relative virulence in a murine osteomyelitis model. To assess the role of extracellular proteases, we also evaluated derivatives of LAC and each regulatory mutant unable to produce the extracellular proteases aureolysin (Aur), staphopain A (ScpA), serine protease A (V8 protease, SspA), cysteine protease B (staphopain B, SspB), and the serine protease-like proteases (SplA-F). These 10 proteases were chosen because they are the primary S. aureus extracellular proteases and because eliminating their production was previously shown to impact virulence in LAC and its isogenic sarA mutant (4–7, 23).

RESULTS

Relative impact of sarA and agr on the pathogenesis of osteomyelitis

We used an established murine osteomyelitis model (4, 5) to make direct comparisons between mice infected with LAC and its isogenic sarA and agr mutants. We did not observe a statistically significant difference in the number of bacteria isolated from the femurs of mice infected with the sarA or agr mutants, but a downward trend was observed in mice infected with the sarA mutant (P = 0.7918), while an upward trend was observed in mice infected with the agr mutant (P = 0.2213) (Fig. 1A).

Fig 1.

Impact of sarA and agr on virulence in a murine osteomyelitis model. Virulence in a murine osteomyelitis was assessed based on bacterial burdens in the femur (A), reactive new bone formation (B), cortical bone destruction (C), and cumulative osteomyelitis score (D). Numbers indicate P values based on one-way ANOVA comparing the results observed with LAC to those observed with isogenic sarA and agr mutants.

As assessed by quantitative micro-computed tomography, reactive new bone formation (NBF) was significantly reduced in mice infected with the sarA mutant (P = 0.0177) by comparison to mice infected with LAC (Fig. 1B). There was also a reduction in the NBF in mice infected with the agr mutant, but it was not statistically significant (P = 0.6853). Although the differences were not statistically significant, mice infected with sarA and agr mutants exhibited less cortical bone destruction (CBD), and like NBF, the decrease was greater in mice infected with the sarA mutant (P = 0.0792) than mice infected with the agr mutant (P = 0.7067) (Fig. 1C). Combining the results from all of these parameters into a cumulative osteomyelitis score (4, 5), it was revealed that mice infected with the sarA mutant exhibited a significant reduction in virulence (P = 0.0010), while mice infected with the agr mutant did not (P = 0.2741) (Fig. 1D). Based on these results and the possibility that the differences we observed might be related to protease production, we generated a sarA/agr mutant and derivatives of LAC and its regulatory mutants with null mutations in the genes and operons encoding aureolysin, ScpA (staphopain A), SspA, SspB (staphopain B), and the serine protease-like proteases SplA-F (4–9, 24).

Impact of sarA and agr on protease activity and biofilm formation

There was a significant reduction in the overall protease activity in all protease-deficient strains and a comparable reduction in the isogenic agr mutant (Fig. 2A, inset). There was a significant increase in protease activity in the sarA and sarA/agr mutants, although the increase was greater in the sarA mutant (Fig. 2A). However, the increase in the sarA/agr mutant was sufficient to limit biofilm formation to a comparable degree, and in both mutants, biofilm formation was restored by eliminating protease production (Fig. 2B). Mutation of agr did not have a significant impact on biofilm formation, irrespective of protease production (Fig. 2).

Fig 2.

Impact of sarA and agr on protease production and biofilm formation. Protease production (A) and biofilm formation (B) were assessed in LAC, isogenic derivatives with mutations in sarA (S), agr (A), sarA and agr (SA), and protease-deficient derivatives of LAC and its isogenic sarA, agr, and sarA/agr mutants that do not produce aureolysin (Aur), staphopain A (ScpA), serine protease A (V8 protease, SspA), cysteine protease B (staphopain B, SspB), or the serine protease-like proteases (SplA-F). (P, SP, AP, and SAP, respectively). Asterisks indicate a significant increase in protease production (A) or decrease in biofilm formation, (B) as determined by one-way ANOVA with Dunnett’s correction for comparison to the results observed with LAC. The difference in protease production between the sarA and sarA/agr mutants was also significant (P < 0.0001), while the difference in biofilm formation was not. The inset in panel A shows data from the same experiment after excluding the sarA and sarA/agr mutants. Asterisks indicate a significant decrease in protease production, relative to LAC.

Impact of sarA and agr on the extracellular proteome

SDS-PAGE analysis demonstrated the absence of high-molecular weight proteins (>40 kDa) in conditioned media from sarA, agr, and sarA/agr mutants (Fig. S1). These changes were correlated with reduced cytotoxicity for osteoclasts and osteoblasts. Specifically, CM from LAC, its protease-deficient derivative, and the protease-deficient sarA mutant were cytotoxic for both osteoblasts and osteoclasts (Fig. 3), while cytotoxicity was reduced in CM from all mutants in which the abundance of high-molecular weight proteins was limited (Fig. S1).

Fig 3.

Impact of sarA and agr on cytotoxicity as a function of protease activity. RAW and MC3T3 cells were used for surrogates for osteoclasts and osteoblasts, respectively. Cytotoxicity was assessed using CM from LAC; isogenic derivatives with mutations in sarA (S), agr (A), and sarA and agr (SA); and protease-deficient derivatives of all four strains (P, SP, AP, and SAP, respectively). Asterisks indicate a statistically significant difference by comparison to the results observed with LAC. Results are reported as the fluorescence intensity, with increased fluorescence reflecting increased cell viability.

Impact of sarA and agr on virulence factors implicated in cytotoxicity

By comparison to CM from LAC, α-toxin was diminished or absent in all mutants that exhibited reduced cytotoxicity (Fig. 4). The abundance of α-toxin was restored in CM from the protease-deficient sarA mutant and even increased by comparison to that in LAC and its protease-deficient derivative. This demonstrates that mutation of sarA results in the increased production of α-toxin and that its abundance is limited in the sarA mutant owing to increased protease production. Full-length α-toxin was not detectable in CM from the agr mutant and its protease-deficient derivative, but it was detected in CM from the sarA/agr mutant and its protease-deficient derivative (Fig. 4). CM from these strains was not cytotoxic (Fig. 3), suggesting that the amount of α-toxins in CM from sarA/agr mutants is below a critical threshold for our cytotoxicity assay.

Fig 4.

sarA limits the production of specific virulence factors via protease-dependent and -independent mechanisms. Western blots were done with CM from overnight cultures of LAC; isogenic derivatives with mutations in sarA (S), agr (A), and sarA and agr (SA); and protease-deficient derivatives of all of these strains (P, SP, AP, and SAP, respectively) using primary antibodies for protein A (Spa), α-toxin (Hla), leukocidin S (LukS), leukocidin F (LukF), and nuclease 1 (Nuc1). Nuc1 is proteolytically processed from its full-length form (NucB) into a truncated form (NucA).

LukS and LukF, the two components of the Panton–Valentine leukocidin (PVL), were not readily detectable by Western blot of CM from LAC or its protease-deficient derivative (Fig. 4), but CM from both strains was cytotoxic (Fig. 3). In contrast, LukF and LukS were detected in CM from the sarA mutant despite its reduced cytotoxicity. However, CM from the sarA mutant contained truncated forms of both proteins, and in the case of LukF, this was the only form detected (Fig. 4). Only full-length forms of LukF and LukS were detected in CM from the protease-deficient sarA mutant, which was cytotoxic (Fig. 3). The fact that LukF was only present in a truncated form in CM from the sarA mutant suggests that this could be a limiting factor in PVL-associated cytotoxicity.

Whether in truncated or full-length forms, the abundance of LukF and LukS was also increased in CM from the sarA and protease-deficient sarA mutants, thus demonstrating that mutation of sarA also results in the increased production of LukF and LukS (Fig. 4). In fact, mutation of sarA resulted in a much greater increase in the abundance of LukF and LukS than mutation of sarS (Fig. S2), which was previously implicated as a key repressor of leukocidin production in S. aureus (25).

Impact of sarA and agr on virulence factors implicated in biofilm formation

The extracellular nuclease Nuc1 was also more abundant in CM from the sarA and protease-deficient sarA mutants (Fig. 4). However, in CM from the sarA mutant, all of Nuc1 was in the smaller NucA form, while in CM from the protease-deficient sarA mutant, it was in the larger NucB form. Both NucA and NucB are enzymatically active (26), suggesting that this difference may not be phenotypically apparent. However, this pattern was fully replicated in the sarA/agr mutant and its protease-deficient derivative (Fig. 4), thus demonstrating that the impact of sarA on Nuc1 production is not dependent on the functional status of agr.

Staphylococcal protein A (Spa) is produced in both surface-anchored and extracellular forms, and both forms contribute to biofilm formation and function as activators of osteoclasts to a degree associated with increased bone destruction in osteomyelitis (27, 28). The experiments we report focused on protein abundance in CM and thus were limited to extracellular Spa (eSpa). Unlike LukF, LukS, and Nuc1, all of which were absent in CM from the LAC agr mutant and its protease-deficient derivative, the abundance of eSpa was increased in CM from the agr mutant and its protease-deficient derivative (Fig. 4). In contrast, eSpa was absent in CM from the sarA and sarA/agr mutants, and its abundance was restored to levels comparable to those of LAC in their protease-deficient derivatives (Fig. 4).

Impact of protease production on the attenuation of sarA and agr mutants

To assess in vivo relevance of these results, we used our osteomyelitis model to evaluate the virulence of LAC; its sarA, agr, and sarA/agr mutants; and derivatives of all four strains unable to produce extracellular proteases. Representative µCT images illustrate extensive NBF and CBD in mice infected with LAC and its protease-deficient derivative (Fig. S3). They also suggest that, by comparison to sarA and sarA/agr mutants, the mutation of agr has little effect on either of these pathological parameters, irrespective of the ability to produce extracellular proteases. In contrast, both were limited in mice infected with the isogenic sarA and sarA/agr mutants, and both were enhanced in isogenic protease-deficient derivatives (Fig. S3).

These differences were confirmed by quantitative µCT analysis. Specifically, by comparison to mice infected with LAC, a reduction in CBD was observed in mice infected with the sarA mutant, although as assessed by one-way ANOVA, this difference was not statistically significant (P = 0.0503). Nevertheless, by comparison to mice infected with the sarA mutant, this was reversed to a statistically significant extent in mice infected with the protease-deficient sarA mutant (P = 0.0163) (Fig. 5). The only significant difference observed in mice infected with the sarA/agr mutant by comparison to mice infected with LAC was decreased NBF (P = 0.0243), and this was also reversed by eliminating the production of extracellular proteases (P = 0.0066). NBF was also reduced in mice infected with the sarA mutant by comparison to mice infected with LAC, but the difference was not statistically significant (P = 0.6429). NBF and CBD are both indicators of overall virulence in osteomyelitis, but this suggests a disconnect between NBF and CBD, which may reflect a spatial difference in osteoblast and osteoclast activity (29) that is driven by the functional status of sarA and agr relative to each other.

Fig 5.

Quantitative analysis of infected femurs. Bacterial burdens in the femur, cortical bone destruction, new bone formation, and cumulative osteomyelitis score were determined for each mouse infected with LAC; its sarA, agr, and sarA/agr mutants; and protease-deficient derivatives of all four strains. Double asterisks indicate a statistically significant decrease (P $\le$ 0.0439) by comparison to mice infected with LAC, as determined by one-way ANOVA. The reduction in cortical bone destruction in mice infected with the sarA mutant was not statistically significant (P = 0.0503). A single asterisk indicates a statistically significant increase in the protease-deficient derivative by comparison to its isogenic regulatory mutant, as determined by unpaired t-test.

As in the preliminary experiment that provided the foundation for these expanded in vivo studies, no significant differences were observed in bacterial burdens in the femur, but downward trends were evident in mice infected with the sarA and sarA/agr mutants but not in mice infected with the agr mutant (Fig. 5). Moreover, when all of these virulence indicators were combined into a cumulative osteomyelitis score that also allows us to include fractured bones (4, 5), a statistically significant reduction in virulence was found in the sarA (P = 0.0021) and sarA/agr mutants (P = 0.0009), but not in the agr mutant. The impact of eliminating proteases was significant in the sarA/agr mutant (P = 0.0006), but not in the sarA mutant, although an upward trend was observed in the protease-deficient sarA mutant (P = 0.1139). In contrast, there were no significant differences between mice infected with agr and protease-deficient agr mutants, as assessed by the overall osteomyelitis score (P = 0.5682) or any of the individual indicators of virulence in our osteomyelitis model (Fig. 5).

Finally, bacterial burdens in the femur, cortical bone destruction, new bone formation, and overall osteomyelitis score were all increased in mice infected with the protease-deficient derivative of LAC itself, although none of these differences were statistically significant (Fig. 5). There was also definitive evidence of systemic dissemination from the bone to multiple soft tissues, which was limited to mice infected with the protease-deficient derivative of LAC (Fig. S4).

DISCUSSION

The persistent problem of acquired antibiotic resistance in S. aureus has rekindled interest in the development of anti-virulence therapeutic strategies. Given the arsenal of S. aureus virulence factors, much of this effort has focused on key regulatory loci that control the abundance of multiple virulence factors. The staphylococcal accessory regulator (sarA) and accessory gene regulator (agr) are the two regulatory loci that have received the most interest in this regard (12–18). The interest in agr is based on its role in increasing production of S. aureus exotoxins and the role of these toxins in acute infections (30), while the interest in sarA is based on its role in biofilm formation and chronic infections (31). The fact that mutation of sarA limits biofilm formation while mutation of agr has the opposite effect largely defines this distinction (32, 33), although the impact of agr on biofilm formation has been called into question particularly under in vivo conditions (34). However, agr-defective variants of S. aureus are often isolated from patients, and it has been proposed that such mutants promote the transition between acute and chronic forms of S. aureus infection, including osteomyelitis (18–22). In contrast, we are unaware of any reports of sarA mutants or strains exhibiting altered protease activity being isolated from patients suffering from orthopedic infection or any other form of S. aureus infection.

Our clinical focus on biofilm-associated orthopedic infections accounts for our interest in sarA as a prophylactic and therapeutic target. Indeed, our previous studies confirmed that mutating sarA has a greater impact on protease production, biofilm formation, and virulence in osteomyelitis than mutation of any other S. aureus regulatory locus we have examined (4–9, 35–41). This is not to say that mutation of other regulatory loci does not have a significant effect on these phenotypes, one example being saePQRS (sae), but rather that mutation of these other loci has a reduced effect by comparison to sarA. However, our previous comparative studies were limited to S. aureus mutants that exhibit increased protease production, and as confirmed here, this does not include agr.

This accounts for our focus on sarA and agr in this report, and we believe the results we present confirm that mutation of sarA limits virulence in our osteomyelitis model to a greater extent than mutation of agr. As with sae (7), this is not to suggest that mutation of agr does not limit virulence in our osteomyelitis model. Indeed, we observed downward trends in mice infected with the LAC agr mutant, particularly with respect to CBD. However, the limited virulence observed in mice infected with an LAC agr mutant by comparison to mice infected with LAC was not statistically significant in either of our in vivo experiments. Rather, the collective results we report support the conclusion that mutation of sarA has a greater impact on the virulence of S. aureus in osteomyelitis than mutation of agr and that this is due to the increased production of extracellular proteases to a degree that is evident even in an sarA/agr mutant. This is important in the context of anti-virulence targets, in that our results suggest that an effective inhibitor of sarA would have a greater therapeutic effect than an inhibitor of agr, at least in the context of osteomyelitis, and that an sarA inhibitor would retain its efficacy even against the spontaneous agr mutants known to arise in vivo during the transition between acute and chronic S. aureus infections (18–22). Such inhibitors could be used along with conventional antibiotics as adjunct therapy following traumatic injury to the bone, perioperatively in elective orthopedic procedures including total joint arthroplasty or, given the frequent need for targeted local delivery of therapeutic agents following debridement and revision of infected bone and orthopedic implants, incorporated into local antibiotic delivery matrices (42, 43).

Our previous studies established that the increased production of proteases in sarA mutants occurs at a transcriptional level (6), and here we confirm that mutation of sarA also results in the increased production of specific virulence factors implicated in biofilm formation and cytotoxicity. This includes α-toxin, LukF, LukS, and Nuc1, all of which were present in increased amounts in CM from the protease-deficient sarA mutant by comparison to CM from LAC and its protease-deficient derivative. With the exception of α-toxin, which was undetectable in the sarA mutant, these proteins were also present in increased amounts in CM from the sarA mutant itself. In fact, the abundance of LukF and LukS was increased in CM from an LAC sarA mutant even above the levels observed in an LAC sarS mutant (Fig. S2), which was previously reported to play a primary role in limiting the production of S. aureus leukocidins (25). This demonstrates that sarA plays a critical role in limiting the production of α-toxin, LukF, and LukS and suggests that it does so in an agr-independent manner. However, we believe this would be an oversimplification. Specifically, these proteins were absent in CM from the LAC agr and isogenic sarA/agr mutants, thus demonstrating that the increased production of α-toxin, LukF, and LukS in the sarA mutant is dependent on the functional status of agr.

The α-toxin phenotype of the sarA/agr mutant provides a potential explanation that is independent of agr to the extent that it is not dependent on the influence of sarA on agr expression. Specifically, the α-toxin phenotype of the sarA/agr mutant was intermediate between that of the isogenic protease-deficient sarA mutant and the parent strain. We hypothesize that the abundance of α-toxin in the sarA/agr mutant may reflect increased transcription of the corresponding gene (hla) owing to the sarA mutation, but a limited capacity to produce α-toxin owing to the agr mutation and the absence of RNAIII, which is required for translation of hla mRNA (44). In this scenario, the observation that this intermediate phenotype was not apparent with LukF or LukS could reflect the importance of the balance between these two factors, thus further illustrating the complex and interactive nature of sarA and agr. Nevertheless, as suggested by the reduced cytotoxicity of the sarA mutant, the increased production of these proteins in an sarA mutant may be phenotypically irrelevant owing to the increased production of extracellular proteases and its impact on the abundance of full-length and presumably functional toxins.

It is noteworthy that mutation of agr limited the abundance of these same toxins, and this was correlated with reduced cytotoxicity for osteoblasts and osteoclasts. However, it was not correlated with reduced biofilm formation. In contrast, mutation of sarA limited both of these phenotypes to a degree that could be correlated with reduced virulence. Given the multifactorial nature of osteomyelitis and the potential importance of both of these in vitro phenotypes, this likely contributes to the reduced virulence of sarA mutants, irrespective of the functional status of agr. The increased production of extracellular proteases in both sarA and sarA/agr mutants, together with the fact that eliminating protease production enhanced the virulence of both mutants, is consistent with the conclusion that extracellular proteases play an important role in this regard.

This is consistent with the eSpa and Nuc1 phenotypes of sarA and sarA/agr mutants, both of which have been implicated in biofilm formation (26, 45). In fact, the observation that the phenotype of sarA mutants was precisely replicated in the isogenic sarA/agr mutant with both of these proteins further illustrates the importance of the agr-independent pathway of sarA-mediated regulation. In fact, mutation of agr resulted in an apparent increase in the abundance of eSpa as expected based on the current agr regulatory paradigm (2), and this was reversed in the protease-deficient sarA and sarA/agr mutants. The virtual absence of eSpa in CM from the sarA and sarA/agr mutants due to protease-mediated degradation may be particularly relevant, in that Spa contributes to osteomyelitis-associated phenotypes and the pathogenesis of osteomyelitis owing to its role in biofilm formation, osteoclastogenesis, and cortical bone destruction (27, 28).

Finally, S. aureus extracellular proteases are proven virulence factors that contribute to nutrient acquisition, tissue invasion, and avoiding host defenses (46–50). The results we present do not contradict this conclusion but rather demonstrate that the increased production of proteases contributes to the attenuation of sarA and sarA/agr mutants. To the extent that the inability to produce proteases enhances virulence in both mutants, our results are consistent with a scenario in which it is important for S. aureus to produce proteases for multiple reasons including balancing its virulence factor repertoire, but equally important that the production of these proteases is kept in check. Failing that, the increased production of extracellular proteases observed in sarA mutants limits the availability of multiple virulence factors that contribute to osteomyelitis-associated phenotypes, including biofilm formation and cytotoxicity for osteoblasts and osteoclasts, as evidenced by their impact on the virulence factors examined in this report. In fact, our results suggest that even in an agr mutant, the impact of extracellular proteases extends beyond their impact on these virulence factors to include other elements of the S. aureus proteome. The need to fine-tune protease production is further reflected in the number of S. aureus regulatory loci that have been implicated in this regard (6, 35). Thus, the results presented here confirm the critical importance of sarA in limiting protease production to levels that benefit S. aureus without compromising its own virulence factor repertoire and demonstrate that sarA is important in this regard irrespective of the functional status of agr.

MATERIALS AND METHODS

Bacterial strains and growth conditions

Mutants were generated in LAC by phage-mediated transduction, as previously described (35, 37, 39). Bacterial strains were recovered from frozen stock cultures and grown overnight (16 hr). Cultures of each strain were then standardized to an optical density (OD₅₆₀) of 10.0. CM was prepared from each standardized culture by removing bacterial cells by centrifugation, followed by filter sterilization with 0.20-micron filters.

Phenotypic assays

Total protease activity was assessed as previously described (4, 5) using CM from standardized bacterial cultures. Assays were done using the EnzChek Gelatinase/Collagenase Assay Kit (Thermo Fisher Scientific, Cat. #E12055). Biofilm assays were performed as previously described using plasma-coated microtiter plates and tryptic soy broth (TSB) supplemented with salt and glucose (32, 33). Cytotoxicity for osteoblasts and osteoclasts was assessed as previously described (4, 5) using RAW 264.7 and MC3T3-E1 cells as surrogates for osteoclasts and osteoblasts, respectively. All phenotypic assays were done with at least two biological replicates, each of which included at least three experimental replicates.

SDS-PAGE and Western blot analysis

Protein electrophoresis of CM samples was done using 4%–12% gradient Bolt Bis-Tris Plus gels (Thermo Fisher Scientific). Gels were stained with SimplyBlue SafeStain (Thermo-Fischer Scientific) and imaged with a Bio-Rad ChemiDoc MP imaging system (Bio-Rad Laboratories) or used for Western blots. Western blots for extracellular protein A (eSpa), Nuc1, LukS, LukF, and α-toxin were done using commercially available antibodies (Sigma and Toxin Technologies, abCAM, and United States Biological) as previously described (4, 5).

Murine osteomyelitis model

For in vivo experiments, each strain was grown overnight (16 hours) with constant shaking at 37°C in TSB without antibiotic selection. Bacterial cells were harvested by centrifugation, washed three times with sterile PBS, and resuspended in PBS at a density of 5 × 10⁸ colony-forming units (CFUs) per mL. Cell density and strain identity were confirmed by plating serial dilutions on TSA with and without antibiotic selection. Mice were infected with 1 × 10⁶ CFUs as previously described (4, 5). After 14 days, mice were humanely euthanized, and the femurs frozen at −80°C. Frozen femurs were used for μCT analysis, followed immediately by homogenization for determination of bacterial burdens. Soft tissues were also harvested immediately after euthanasia and homogenized to determine bacterial burdens in individual tissues. The identity of isolates obtained from bone and soft tissues was confirmed by plating on TSA with antibiotic selection for the appropriate mutations and by polymerase chain reaction (PCR) analysis to confirm the absence of cna (forward: CAAGCAGTTATTACACCAGACGG, reverse: CACCTTTTACAGTACCTTCAATACC) and the presence of lukS (forward: AATTGCATTGCTTTTGCTATCC, reverse: ATTTTGAACCATTACCTCCACC). Colony counts were logarithmically transformed for statistical analysis. Samples with no bacterial burden were assigned a colony count of 1 to allow inclusion of results from all experimental animals.

Microcomputed tomography (μCT)

Femurs were scanned with a Skyscan 1275 Microtomograph (Bruker) at 40 kV (100 uA) using an isotropic voxel size of 6.8 um. After scanning, images were reconstructed using Skyscan Nrecon software and then processed with Skyscan CT-analyzer software as previously described (4–7). A semi-automated protocol (global thresholding 90–255, round closing in 3D with pixel size 4, round opening in 3D with pixel size 1, round closing in 3D with pixel size 8, and round dilation in 3D pixel size 3) was used to generate preliminary regions of interest (ROIs) of cortical bone. Every 20th automated ROI image was loaded onto the sample to act as a base for manual ROI adjustment to ensure inclusion of only cortical bone within the ROI. Cortical bone volume was then calculated within the ROIs with a threshold of 70–255, and cortical bone destruction was determined by subtracting the cortical bone volume from the average amount of cortical bone in a sham surgical bone. New bone ROIs were generated using the subtractive volume function on the cortical ROIs, calculating bone volume with a threshold of 45–135, and then subtracting the corresponding average amount of bone present in sham surgical bones to account for any trabecular bone inclusion as well as normal post-surgical new bone formation. Example uCT images from randomly selected bones from mice in each experimental group were generated in Skyscan CTvox software.

Statistical analysis

Statistical analysis of in vitro and in vivo results was done by one-way ANOVA with Dunnett’s correction to allow comparison of the results from all experimental groups relative to the results observed with LAC. Unpaired t-tests were used to individually assess the impact of extracellular proteases in LAC and its regulatory mutants. Error bars indicate standard error of the mean. All statistical analyses were done using GraphPad Prism software (version 10.0.0 for Windows, GraphPad Software).

DATA AVAILABILITY

The authors confirm that the data supporting the findings of this study are available within the article [and/or] its supplementary materials.

ETHICS APPROVAL

All animal experiments were approved by the Institutional Animal Care and Use Committee of the University of Arkansas for Medical Sciences under protocol number 4124 and were performed in compliance with NIH guidelines, the Animal Welfare Act, and United States Federal Law.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/iai.00473-24.

Fig. S1. iai.00473-24-s0001.tif.

Supplemental Figure 1.

Fig. S2. iai.00473-24-s0002.tif.

Supplemental Figure 2.

Fig. S3. iai.00473-24-s0003.tif.

Supplemental Figure 3.

Fig. S4. iai.00473-24-s0004.tif.

Supplemental Figure 4.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.