Bacillus subtilis revives conventional antibiotics against Staphylococcus aureus osteomyelitis

As treatment of Staphylococcus aureus (S. aureus) osteomyelitis is often hindered by the development of antibiotic tolerance, novel antibacterial therapeutics are required. Here we found that the cell-free supernatant of Bacillus subtilis (B. subtilis CFS) killed planktonic and biofilm S. aureus, and increased S. aureus susceptibility to penicillin and gentamicin as well. Further study showed that B. subtilis CFS suppressed the expression of the genes involved in adhesive molecules (Cna and ClfA), virulence factor Hla, quorum sensing (argA, argB and RNAIII) and biofilm formation (Ica and sarA) in S. aureus. Additionally, our data showed that B. subtilis CFS changed the membrane components and increased membrane permeabilization of S. aureus. Finally, we demonstrated that B. subtilis CFS increased considerably the susceptibility of S. aureus to penicillin and effectively reduced S. aureus burdens in a mouse model of implant-associated osteomyelitis. These findings support that B. subtilis CFS may be a potential resistance-modifying agent for β-lactam antibiotics against S. aureus.


Introduction
Gram-positive Staphylococcus aureus (S. aureus) has been identified as the most common causative pathogen for osteomyelitis and other various musculoskeletal infections [1,2]. S. aureus osteomyelitis remains a significant healthcare problem in China and around the world due to high rates of recurrence and treatment failure [3,4]. Treatment of S. aureus infection in bone is complicated by its vast immune evasion, persistence mechanisms and intrinsic antibiotic resistance mechanism. S. aureus may secrete multiple virulence factors including immunomodulatory proteins, toxins and superantigens, leading to death of innate immune cells and disturbance of complement activation [5]. As the infection persists and becomes chronic, S. aureus may adhere to implanted devices, lacunae-canaliculi in cortical bone or sequestra, thereby forming biofilm phenotype [6,7]. Once a biofilm forms, S. aureus is 10 − 1,000 times more resistant to antimicrobial agents than planktonic bacteria [8] and induces phagocytosis dysfunction of macrophages [9]. Additionally, intracellular persistence of S. aureus in osteoblasts, macrophages, osteoclasts or osteocytes may induce immune cell evasion and antibiotic tolerance of S. aureus during infection [10,11]. Furthermore, S. aureus has such intrinsic mechanism for antibiotic resistance as decreasing permeability of outer membrane, activating drug efflux systems, and producing excessive β-Lactamase [12][13][14]. Zhang et al. Microb Cell Fact (2021) 20:102 Surgical debridement of necrotic bone combined with long-term administration of antibiotics is a traditional therapy to treat chronic osteomyelitis [15]. Several antibiotics are used for management of S. aureus osteomyelitis, such as vancomycin, tobramycin, daptomycin and clindamycin, but the rapid acquisition of resistance to antibiotics by S. aureus is a significant problem [16][17][18][19]. Therefore, it is urgent to find a more effective antibacterial strategy to prevent occurrence and recurrence of bone infections.
Recently, probiotics such as Bacillus subtilis (B. subtilis) has been used to prevent infection, because it is a nonpathogenic Gram-positive bacterium which can effectively maintain a beneficial microflora balance in the gastrointestinal tract of a mammalian host [20]. Accumulating evidence from animal and in vitro studies suggests that B. subtilis produces various substances, such as sufactins, iturins and fengycins, which may benefit antibacterial, anti-inflammatory and immunomodulatory applications [21,22]. Specifically, a recent report showed that the secreted substance from B. subtilis abolished colonization with S. aureus by suppressing production of the Arg-quorum-sensing signaling system [21]. In light of recent evidence implicating anti-infection and decolonization role of Bacillus lipopeptides against S. aureus, we investigated the effect of B. subtilis cell-free supernatant (B. subtilis CFS) on the growth of S. aureus in vitro and in vivo.
Here we found that B. subtilis CFS exerted a potent antimicrobial function against S. aureus and increased its susceptibility to antibiotics as well in vitro and in vivo as well. Furthermore, we demonstrated that B. subtilis CFS changed the membrane components and increased membrane permeabilization of S. aureus, which may be associated with increased susceptibility of S. aureus to antibiotics. Our data may suggest a potential application of B. subtilis CFS as an adjuvant to potentiate β-lactam antibiotics against S. aureus osteomyelitis.

Bacterial strains and culture
Staphylococcus aureus strains were isolated from the osteomyelitis subjects from Department of Orthopedics, Nanfang Hospital, Southern Medical University, using PHOENIX 100 (Becton Dickinson Microbiology System, USA). B. subtilis (CMCC-B-63,501) was obtained from China General Microbiological Culture Collection Center. Bacterial strains were cultured in TSB (Cat. LA0110, Solarbio, Beijing, China) at 37 ℃ under shaking at 200 rpm. Overnight bacterial cultures were collected by a centrifuge, and pellets washed and resuspended in phosphate-buffered saline (PBS) (Cat. C10010500BT, GIBCO, Beijing, China). The bacterial suspensions were adjusted to an optical density at 600 nm (OD 600 ) of 0.5 measured using a microplate spectrophotometer (CLAR-IOstar, BMG LABTECH, Germany), approximately equal to 1 × 10 8 colony forming unit per ml (CFU/ml).

Preparation of cell-free supernatant from B. subtilis culture and treatments
To prepare B. subtilis CFS, B. subtilis strains were cultured at 37℃ under shaking at 200 rpm overnight until the cultures reached an OD 600 of 0.4 ± 0.05. The CFS of bacterial culture was collected by centrifugation at 6000 g for 10 min, and then filtered through a 0.22 μm sterilizing-grade filter (Millipore, SLGV033RB, USA) to remove bacteria. The CFS was aliquoted and stored at − 20 ℃ until the day of experimentation.
To evaluate the effect of B. subtilis CFS on S. aureus genes expression, overnight culture of S. aureus strains was collected by a centrifuge, washed with PBS, re-suspended at 1 × 10 8 CFU/ml in TSB/PBS (1:1 v/v, control) or TSB/B. subtilis CFS (1:1 v/v) and incubated in 6-wellplate at 37℃ for 3 h. Finally, bacteria were collected for RNA extraction and analysis of genes expression.

Planktonic bacterial growth assay
To determine the antibacterial effect of B. subtilis CFS on S. aureus, the growth of planktonic S. aureus was assessed using the method as described previously [23] with some modifications. Briefly, 100 µL of S. aureus suspension (5 × 10 8 CFU/mL) from a fresh overnight culture was inoculated into 5 mL TBS/PBS (1:1 v/v, control) or TSB/B. subtilis CFS (1:1 v/v), and incubated with shaking at 200 rpm at 37 ℃. The growth of S. aureus was determined by monitoring OD 600 of the cell culture at 2, 4, 6, 8, 10, 12 and 24 h after seeding.

Biofilm formation and viability assay of biofilm S. aureus
To evaluate the effect of B. subtilis CFS on S. aureus biofilm formation, 100 µL of S. aureus (5 × 10 8 CFU/ mL) was added to 900 µL of TSB/PBS (1:1 v/v), TSB/B. subtilis CFS (1:1 v/v), TSB/PBS (1:1 v/v) with 32 µg/mL penicillin, or TSB/PBS (1:1 v/v) with 0.75 µg/mL gentamicin in each well on a 24-well plate and incubated at 37℃ for indicated time points without shaking. Next, after the medium removed, the wells were washed three times with sterile PBS. Finally, the plates were air-dried for 45 min and the adherent cells and matrix were stained with 0.1 % crystal violet solution. To quantify the biofilm production, crystal violet was extracted by incubation in solution (95 % ethanol and 0.1 % acetic acid) at room temperature for 15 min, and absorbance was measured at 600 nm in a microplate reader. SYTO9 (Cat. S34854, Invitrogen, Thermo Fisher Scientific) and propidium iodide (PI) (Cat. P346, DOJINDO, Japan) staining was performed to evaluate the effect of B. subtilis CFS on the viability of biofilm S. aureus. 100 µL of S. aureus (5 × 10 8 CFU/mL) was added to 900 µL of TSB in each well on a 12-well plate. After 24 h of static incubation at 37℃, the wells were washed three times with PBS to remove nonadherent cells and refilled with 1 mL/well of the four different sterile culture media: TSB/PBS (1:1 v/v, control), TSB/B. subtilis CFS (1:1 v/v), TSB/PBS (1:1 v/v) with 32 µg/mL penicillin, and TSB/PBS (1:1 v/v) with 0.75 µg/mL gentamicin. After 8 h incubation and washing for three times, the biofilm S. aureus were stained with 3 µM of PI and 10 µM of SYTO9 in 1 × PBS for 20 min in the dark, and visualized under a fluorescence microscope. Both live and dead bacteria were stained green, and dead ones red.

Minimum inhibitory concentration (MIC) and killing assay
The potential of synergy was evaluated via MIC evaluation and time-killing assays. MIC was determined using Epsilometer testing (E-test) following the method previously described [24,25]. Briefly, fresh overnight culture of S. aureus was collected and washed twice with PBS, and suspended and pretreated in 1 ml PBS (control) or B. subtilis CFS at 1 × 10 8 CFU/ml for 1 h. 150 µl pretreated S. aureus suspension was added and spread evenly on a Mueller-Hinton agar plate. The plate was allowed to dry for 10-15 min before applying E-test strip immobilized with predefined continuous and stable gradients of penicillin (Cat. 921,021, Liofilchem, Italy) or gentamicin (Cat. 920,090, Liofilchem, Italy). The plates were incubated at 35℃ for 24 h and the MIC value was read at the point where the ellipse intersects the E-test strip.
To monitor the response of B. subtilis CFS-pretreated S. aureus to penicillin or gentamicin, bacterial growth was continuously monitored over a time-course of 24 h (0, 2, 4, 6, 8, 10, 12, 14, 24 h). 500 µl of S. aureus suspension (1 × 10 8 CFU/ml) pretreated with PBS (control) or B. subtilis CFS was inoculated into 4.5 mL of Mueller-Hinton broth with penicillin or gentamicin at 0.5 MIC. A 200µL of sample was removed from each tube at indicated time points for measuring OD 600 .
For time-killing assay, 500 µl of S. aureus suspension (1 × 10 7 CFU/ml) pretreated with PBS (control) or B. subtilis CFS was inoculated into 4.5 mL of Mueller-Hinton broth with penicillin or gentamicin, with each drug tested at 2 × MIC and 4 × MIC. A 10 µL of sample was removed from each tube at 0, 0.5, 1, 2, 4, 6, 8, 12 and 24 h for colony count enumeration. 10 µL samples with 100-fold dilutions were plated onto Mueller-Hinton agar plates and incubated at 35℃ for 18 h. Colonies were counted and the mean CFU/mL from triplicate samples was evaluated.

RNA extraction and Quantitative real-time PCR (qRT-PCR)
Total RNA of S. aureus was extracted with a Bacterial RNA Extraction Kit (B518655-0050, Sangon Biotech, Shanghai, China) following the manufacturer's instructions. RNA purity was checked using a NanoDrop spectrophotometer (ND-1000, Nanodrop, USA). RNA was reversely transcribed using the 5× PrimeScript RT Master Mix (RR036A, Takara, Shiga, Japan) according to the manufacturer's instructions. qRT-PCR was performed using TB Green Premix Ex Taq II (RR820A, Takara, Shiga, Japan). The primers sequences are listed in Table 1. Fold change in level of chosen genes expression were determined using 2 −ΔΔCt method with gyrB as a housekeeping gene.

Transmission electron microscopy (TEM)
Staphylococcus aureus suspension (1 × 10 8 CFU) pretreated with PBS (control) or B. subtilis CFS was collected and fixed in 2.5 % Glutaric dialdehyde at 4℃ overnight. After washing, S. aureus pellets were dehydrated in a series of ethanol concentrations (50-100 %) followed by Table 1 Primers used for quantitative real-time polymerase chain reaction 100 % acetone. Samples were then embedded in Spurr resin (EM0300, Sigma-Aldrich, USA). 50 nm ultrasections were cut using an ultramicrotome (EM UC7, Leica, Germany) and stained with uranyl acetate for 10 min. After being washed with ddH 2 O, sections were stained with Reynolds lead citrate for 30 min. Finally, sections were observed on a transmission electron microscope (H-7500, Hitachi, Japan) equipped with a 16 million pixels format CCD camera and images were made at 120 kV in high contrast mode. . The pixel density of protein bands were analyzed using Image J, the relative level of PBP2a expression was normalized against GAPDH, and fold changes over control were calculated.

Implant-associated S. aureus osteomyelitis mice model
All procedures involving animals were approved by the Animal Care and Use Committee at Nanfang Hospital, Southern Medical University. 88 male C57BL/6 J mice (8-10 weeks old) were obtained from the Animal Center at Southern Medical University. Mice were housed in a facility under specific pathogen-free conditions at 24-27℃ with a 12-h light/dark cycle and had ad libitum access to food and water.
The mice model of implant-associated osteomyelitis was made as described previously with modifications [26]. In brief, prior to surgery, they were anesthetized by 125 mg/Kg tribromoethanol (Cat. T831042, Shanghai, China) via intraperitoneal injection. After being shaved and sterilized, an incision was made at the lateral side of the right hind leg and the tibiae was exposed by blunt dissection, and a uni-cortical hole was created at the proximal part of the tibia with a 29-gauge syringe needle. Next, an 8 mm stainless steel pin (0.3 mm in diameter) was inserted into the bone medullary cavity. The hole was sealed with bone wax and the wound was sutured after disinfection. By day 7 post-surgery, S. aureus (5 × 10 7 CFU/mL, 100 µL) was inoculated by intravenous injection via the tail vein. Mice were monitored twice daily for morbidity and mortality.

Infection and treatments in vivo
To determine the anti-bacterial effect of B. subtilis culture CFS in vivo, 48 mice with implant-associated S. aureus osteomyelitis were randomly divided into two groups and injected intraperitoneally with 200 µL of B. subtilis culture CFS or the same volume of PBS (control) every day from the day challenged by S. aureus. By days 3 and 14 after S. aureus inoculation, the right tibias were collected aseptically and the implanted stainless steel pin was pulled out for analysis of bacterial burden.
To evaluate the responses of B. subtilis CFS-pretreated S. aureus to penicillin in vivo, 40 mice were randomly divided into two groups and infected by S. aureus (5 × 10 7 CFU/mL, 100 µL) pretreated in 1 ml B. subtilis CFS or PBS (control) at day 7 after implantation surgery. The next day after S. aureus challenge, mice were intraperitoneally injected with penicillin (80 mg/Kg/d). All the mice were sacrificed at days 3 and 14 post-infection by cervical dislocation, the right tibias were collected and the implanted pins were removed from the bone for analysis of bacterial burden.

Antimicrobial assays in vivo
To assess bacterial burden in bone, the right tibia infected by S. aureus was dissected aseptically free from soft tissue, and homogenized in 1 ml of PBS. A 10-fold dilution of the bone homogenate was plated in TSB agar plate. Bacterial colonies were counted and calculated following plate incubation at 37℃ for 18 h. Results of bacterial burden were expressed on a log 10 scale.
To detect bacterial burden on the implant surface, pins were removed carefully from the tibia after the mice were euthanized. The pins were then sonicated in 1 ml of PBS for 5 min to obtain the biofilm bacteria. Each sample was incubated on TSB agar plates at 37℃. After 24 h incubation, the number of bacterial colonies was counted, calculated and expressed on a log 10 scale.
The survival rates were recorded within 14 days postinfection on S. aureus challenged mice. The infection rates were evaluated based on the mice with infected tibia or implant among surviving mice.

Histological analysis and immunofluorescence
To evaluate the pathological changes in bone, paraffinembedded samples were sectioned in 5-µm thickness, deparaffinized with xylene and hydrated by ethanol gradient, followed by hematoxylin and eosin (H&E) staining. Quantitative evaluation of the histopathological changes was performed using Smeltzer's scoring methods [27]. The parameters included intraosseous acute inflammation (0-4), intraosseous chronic inflammation (0-4), periosteal inflammation (0-4) and bone necrosis (0-4). A score assigned for each sample was the sum of the scores made from the above 4 parameters by two blinded observers independently.
To detect biofilm S. aureus on the implant surface, the pins implanted were removed from the tibia gently by day 14 post infection, rinsed 3 times with PBS and fixed in buffered 4 % paraformaldehyde solution for 24 h. The implants were blocked with 3 % BSA for 1 h and incubated with the rabbit polyclonal anti-S. aureus antibody (Cat. ab20920, Abcam) at 4℃ overnight. On the next day, sections were incubated with 594-conjugated secondary antibody (Cat. 712-586-153, Jackson ImmunoResearch, West Grove, PA, USA). Slides were mounted with antifade mounting medium with DAPI (Cat. S2110, Solarbio, Solarbio Life Sciences, China), and images were acquired with a fluorescence microscope (BX63, OLYMPUS, Japan).

Scanning Electron Microscopy (SEM)
Steel pins were removed from the tibias at day 14 after S. aureus infection before fixed in 2.5 % Glutaric dialdehyde at 4℃ for 16 h. After being washed and serially dehydrated in a graded series of ethanol solutions, pins were dried in a critical point dryer (HCP-2; Hitachi, Tokyo, Japan) followed by gold plasma coating (E-1010; Hitachi, Tokyo, Japan). Specimens were imaged using a scanning electron microscope (S-3000 N; Hitachi, Tokyo, Japan).

Statistical analysis
All experiments were performed for at least three times. Since the sample sizes were relatively small and the sample distributions not normally distributed, the nonparametric Mann-Whitney U test was applied to compare the differences between the two groups. For comparison of the survival time between the two groups, Gehan-Breslow-Wilcoxon test was used. For assessment of infection rate, Chi-square test was used. P < 0.05 was considered statistically significant. All statistical data were analyzed using SPSS 19.0 software.

B. subtilis CFS suppresses the growth of planktonic and biofilm S. aureus
The investigation of the effect of B. subtilis CFS on the growth of S. aureus via measuring the OD 600 of planktonic cells at indicated time points showed that B. subtilis CFS significantly suppressed the growth of planktonic S. aureus after 4 h of treatment, the inhibitory effect was as strong as that of gentamicin and continued for 24 h of treatment time (Fig. 1a). Moreover, biofilm formation in static S. aureus culture was evaluated by crystal violet staining. Results showed much faint staining in the culture of B. subtilis CFS-treated S. aureus (Fig. 1b), indicating inhibitory effect of B. subtilis CFS on S. aureus biofilm production. Next, the dissolved crystal violet was subjected to quantitative analysis. As shown in Fig. 1c, B. subtilis CFS inhibited biofilm formation during the time points examined, whereas gentamicin had limited inhibitory effect on biofilm production before 12 h of treatment. To evaluate the effect of B. subtilis CFS on biofilm S. aureus, the S. aureus biofilms were formed on plastic wells after static incubation for 24 h, followed by treatment with B. subtilis CFS, penicillin or gentamicin for 8 h. Membrane-permeable SYTO9 and membraneimpermeable PI staining was performed to evaluate the amount of biofilm S. aureus. Compared to the control group and groups treated by penicillin or gentamicin, B. subtilis CFS treatment suppressed both green-stained live biofilm S. aureus and red-stained dead ones (Fig. 1d).

B. subtilis CFS increases antibiotic susceptibility of S. aureus in vitro
To evaluate the effect of B. subtilis CFS on the response of S. aureus to antibiotics, the MIC of S. aureus pretreated with PBS or B. subtilis CFS for 1 h was detected using E-test, as shown in Fig. 2a. Quantitative analysis showed distinctly decreased MICs of penicillin and gentamicin against B. subtilis CFS pretreated-S. aureus. Specifically, the MICs of penicillin to PBS-pretreated and B. subtilis CFS-pretreated S. aureus were 32 µg/ml and 12 µg/ml, respectively (Fig. 2b). The MICs of gentamicin to PBS-pretreated and B. subtilis CFS-pretreated S. aureus were 0.75 µg/ml and 0.31 ± 0.063 µg/ml, respectively (Fig. 2c). Further study showed that B. subtilis CFS pretreatment did not suppress the growth of S. aureus in TSB, but increased the susceptibility of S. aureus to penicillin. PBS-pretreated S. aureus grew rapidly in TSB with penicillin (0.5 MIC) after 8 h of incubation, whereas the growth of B. subtilis CFS-pretreated S. aureus was substantially suppressed by penicillin after 14 h of incubation (Fig. 2d).
To further examine the effects of B. subtilis CFS on antibiotic susceptibility of S. aureus, we pretreated S. aureus with PBS or B. subtilis CFS for 1 h and then performed time-kill assay on planktonic S. aureus exposed to penicillin or gentamicin. Since this S. aureus strain was not sensitive to penicillin, we therefore made timekill curves for penicillin at 4× the MIC of antimicrobial concentration. Significantly decreased cell survival rate was observed in S. aureus pretreated with B. subtilis CFS compared to control ones after 8 h. Additionally, 99 % of S. aureus pretreated with B. subtilis CFS was killed before 24 h (Fig. 2e). Since this S. aureus strain was susceptible to gentamicin, the time-kill curves for gentamicin were made at 2× the MIC of antimicrobial concentration. Results showed that S. aureus pretreated with B. subtilis CFS had increased sensibility to gentamicin compared to control ones, 99 % of S. aureus pretreated with B. subtilis CFS was killed before 4 h (Fig. 2f ). Based on the above time-kill assay data, the minimum duration for killing 90 % (MDK 90 ) values was calculated for S. aureus exposed to penicillin or gentamicin. There was a distinct decrease in MDK 90 values of S. aureus pretreated with B. subtilis CFS than in those of control ones for both penicillin and gentamicin (Fig. 2g). Together, the above data clearly indicated that pretreatment with B. subtilis CFS led to a greater sensitivity of S. aureus to penicillin and gentamicin.

B. subtilis CFS increases membrane permeability of S. aureus
Next, we analyzed the effects of B. subtilis CFS on expression of S. aureus genes encoding adhesive molecules (Cna and ClfA) and virulence factor Hla, and genes involved in quorum sensing (argA, argB and RNAIII) and biofilm formation (Ica and sarA). Results showed that B. subtilis CFS treatment significantly down-regulated the mRNA expression of all the above genes (Fig. 3a).
Since the permeabilizing property of bacterial cell membrane is pivotal to penetration of antibiotics, we analyzed the membrane integrity of S. aureus using SYTO9-PI assay. Results showed that B. subtilis CFS disrupted the membrane of S. aureus after pretreatment for 1 h, as evidenced by the presence of PI molecules in S. aureus (Fig. 3b).
The effect of B. subtilis CFS on the membrane permeabilization of S. aureus was determined by ATP leakage assays. Results showed that B. subtilis CFS did not change the whole amount of ATP, but significantly increased the levels of extracellular ATP (0.0676 ± 0.0023 nM) (See figure on next page.) Fig. 2 B. subtilis CFS increases S. aureus susceptibility to penicillin and gentamicin. a Representative E-test images of S. aureus from penicillin and gentamicin. The minimum inhibitory concentration (MIC) was read off of the strip where the bottom portion of the ellipse intersects with the strip (see black arrows). b, c Quantitative analysis show significantly decreased MICs of penicillin and gentamicin against S. aureus pretreated by B. subtilis CFS. MIC values were measured using aliquots of S. aureus cultures from three different colonies. *P < 0.05, Mann-Whitney U test. d The growth of S. aureus pretreated with PBS or B. subtilis CFS were monitored with or without the presence of penicillin or gentamicin. Fresh overnight culture of S. aureus (5 × 10 7 CFU/mL) was pretreated with PBS or B. subtilis CFS for 1 h, and then challenged with PBS, 0.5 MIC penicillin or gentamicin. Samples were collected and OD 600 was recorded at indicated time points. N = 4/group at each time point. *P < 0.05, Mann-Whitney U test. Time-dependent killing of control S. aureus and B. subtilis CFS-pretreated S. aureus by penicillin at 4 × MIC (e) and gentamicin at 2 × MIC (f). Experiments were independently repeated for 4 times. *P < 0.05, Mann-Whitney U test. g Minimum duration for killing 90 % (MDK) measurements of control S. aureus and B. subtilis CFS-pretreated S. aureus exposed to penicillin at 4 × MIC or gentamicin at 2 × MIC. Values were determined from the quadruplicate data shown in (shown in e and f). *P < 0.05, Mann-Whitney U test compared to control (0.010 ± 0.0005 nM) (p < 0.05) (Fig. 3c), indicating that S. aureus membrane was profoundly compromised by B. subtilis CFS. Indeed, TEM analysis confirmed that B. subtilis CFS disrupted the typical semi-rigid structure of S. aureus. As can be seen in Fig. 3d, control S. aureus cells showed even cell walls, but B. subtilis CFS-treated S. aureus showed compromised cell walls, such as disruption of cell wall, displacement of cell membrane and extrusion of intracellular content.
To evaluate the effect of B. subtilis CFS on the membrane proteins of S. aureus, whole-cell and membrane proteins of S. aureus were detected using SDS-PAGE and CBB staining. As seen in Fig. 3e, compared with control, B. subtilis CFS treatment considerably changed the pattern of whole-cell protein bands in S. aureus. The levels of some proteins decreased while some new proteins appeared. Interestingly, these membrane protein bands with molecular weights of 55, 70 and 100 kDa were much weaker than those of the controls, suggesting that B. subtilis CFS has a great effect on the level of proteins in membrane. Next, we evaluated the effect of B. subtilis CFS on the mRNA expression of mecA, a gene encoding PBP2a which has a molecular weight of around 70 kDa [28]. Results showed that B. subtilis CFS substantially suppressed the mRNA expression of PBP2a (Fig. 3f ). Analyses of the protein levels of PBP2a in whole-cell lysate of S. aureus confirmed the inhibitory effect of B. subtilis CFS on PBP2a expression (Fig. 3g, h).

B. subtilis CFS reduces a hematogenous implant-associated infection in mice
To test whether B. subtilis CFS might protect against S. aureus infection in vivo, we made a mouse osteomyelitis model of hematogenous implant-associated infection. The groups of mice were infected with 5 × 10 6 CFU of S. aureus at day 7 after surgical implantation. Mice were received PBS (control group) or B. subtilis CFS injection once a day from the day challenged by S. aureus (Fig. 4a). Treatment of B. subtilis CFS improved the survival of mice challenged by S. aureus compared with control ones (Fig. 4b). The infection rate in surviving control mice increased between days 3 and 14 postinfection. In contrast, surviving mice had a significantly lower infection rate in B. subtilis CFS-treated group compared with those in control group, and the infection rate remained unchanged between days 3 and 14 post-infection (Fig. 4c). Accordingly, enumeration of bacterial burdens revealed that control mice harbored higher bacterial burdens on days 3 and 14 post-infection, while B. subtilis CFS treatment did substantially reduce bacterial burdens in the tibias and implants (Fig. 4d, e).
To detect the effect of B. subtilis CFS on growth of biofilm S. aureus and changes in bone marrow surrounding an implant, the implants and tibias were harvested on day 14. Immunofluorescence staining showed a considerable amount of S. aureus-positive staining on the implant surface in PBS-treated mice, while no obvious signals were observed on the implants in B. subtilis CFS-treated mice (Fig. 5a). SEM analysis confirmed biofilm formation rescued by B. subtilis CFS treatment (Fig. 5b). Additionally, histologic assessment using H&E staining revealed deformation of bone structure and marked abscess formation within the marrow cavity around the implant in PBS-treated control mice, with no obvious bone destruction in B. subtilis CFS-treated mice (Fig. 5c). Histological scores confirmed significantly improved bone structure in the bone of B. subtilis CFS-treated mice (Fig. 5d).

Staphylococcus aureus pretreated by B. subtilis CFS is susceptible to penicillin in vivo
To address the effect of B. subtilis CFS pretreatment on the susceptibility of S. aureus to penicillin in vivo, we examined the outcomes of penicillin treatment of mice infected by PBS-pretreated or B. subtilis CFS-pretreated S. aureus. Penicillin treatment did not extend the survival of mice infected by PBS-pretreated S. aureus but significantly prolonged the survival of mice infected by B. subtilis CFS-pretreated S. aureus (Fig. 6a). In surviving mice, penicillin significantly suppressed the infection rate in mice infected by B. subtilis CFS-pretreated S. aureus (Fig. 6b). Furthermore, enumeration of S. aureus cells in the tibias and implants by days 3 and 14 post-infection showed that the surviving mice infected by B. subtilis CFS-pretreated S. aureus had significantly decreased bacterial burdens in the infected tibias and implants (Fig. 6c,  d). Together, these data collected in vivo supported an

Discussion
S. aureus is one of the important pathogens causing various infections like osteomyelitis. It is hard to cure, in part because of the ability of S. aureus to enter into an antibiotic-tolerance state and the formation of biofilm S. aureus. The present study provided evidence for bactericidal effect of B. subtilis CFS on both planktonic and biofilm S. aureus in vitro and in vivo. We also demonstrated that B. subtilis CFS treatment increased the susceptibility of S. aureus to penicillin and gentamicin, which might have been associated with changes in membrane components and increased membrane permeability in S. aureus, respectively. Furthermore, our findings also demonstrated the sensitivity of B. subtilis CFS-pretreated S. aureus to penicillin in a mouse model of implant-associated osteomyelitis.
Several studies have reported that B. subtilis exerts an antimicrobial effect against a broad spectrum of pathogens through direct bactericidal activity or indirect enhancement of immune response, such as interrupting quorum-sensing regulatory system by production of fengycins [21], inhibiting S. aureus adhesion and biofilm formation by production of surfactins [29], and enhancing anti-microbial function of macrophage [30]. In agreement with the above reports, our study has confirmed a potent inhibitory capacity of B. subtilis CFS against both planktonic and biofilm S. aureus in vitro, which may prominently suppress expression of genes associated with S. aureus adhesion, biofilm formation, quorum-sensing and virulence. Furthermore, our data demonstrate the bactericidal effect of B. A critical finding in this study is that B. subtilis CFS increased the susceptibility of S. aureus to penicillin in vitro and in vivo. Generally, S. aureus strains are found to be resistant to almost all β-lactam antibiotics as they produce β-Lactamase that breaks down β-lactam ring or a penicillin-binding protein called PBP2a that has a low binding affinity to β-lactam antibiotics [14,31]. Our data demonstrates the inhibitory effect of B. subtilis CFS on the expression of PBP2a at both transcriptional and translational level, therefore, the increased sensitivity