Cinnamaldehyde targets SarA to enhance β‐lactam antibiotic activity against methicillin‐resistant Staphylococcus aureus

Abstract Methicillin‐resistant Staphylococcus aureus (MRSA) is a current global public health problem due to its increasing resistance to the most recent antibiotic therapies. One critical approach is to develop ways to revitalize existing antibiotics. Here, we show that the phytogenic compound cinnamaldehyde (CIN) and β‐lactam antibiotic combinations can functionally synergize and resensitize clinical MRSA isolates to β‐lactam therapy and inhibit MRSA biofilm formation. Mechanistic studies indicated that the CIN potentiation effect on β‐lactams was primarily the result of inhibition of the mecA expression by targeting the staphylococcal accessory regulator sarA. CIN alone or in combination with β‐lactams decreased sarA gene expression and increased SarA protein phosphorylation that impaired SarA binding to the mecA promoter element and downregulated virulence genes such as those encoding biofilm, α‐hemolysin, and adhesin. Perturbation of SarA–mecA binding thus interfered with PBP2a biosynthesis and this decreased MRSA resistance to β‐lactams. Furthermore, CIN fully restored the anti‐MRSA activities of β‐lactam antibiotics in vivo in murine models of bacteremia and biofilm infections. Together, our results indicated that CIN acts as a β‐lactam adjuvant and can be applied as an alternative therapy to combat multidrug‐resistant MRSA infections.


INTRODUCTION
Multidrug-resistant (MDR) pathogens pose a serious global public health threat, with the rapid dissemination of methicillinresistant Staphylococcus aureus (MRSA).MRSA is highly toxigenic and can induce pore-forming and exfoliative toxins that exacerbate skin and soft tissue infections, leading to bacteremia, endocarditis, and pneumonia 1,2 .The therapeutic efficacy of most antibiotics against such bacterial infections is often compromised due to multiple resistance mechanisms [3][4][5] .For instance, the mecA gene encodes an alternative penicillinbinding protein 2a (PBP2a) that shows low-affinity to almost all β-lactam antibiotics 6 .Previous studies have shown a positive correlation between the expression of mecA and the degree of 1 State Key Laboratory for Animal Disease Control and Prevention, South China Agricultural University, Guangzhou, China. 2 Guangdong Provincial Key Laboratory of Veterinary Pharmaceutics Development and Safety Evaluation, South China Agricultural University, Guangzhou, China. 3Yantai Fushan Center for Animal Disease Control and Prevention, Yantai, China.
The staphylococcal sarA is a global regulator of MRSA virulence determinants and biofilm formation in vitro, and is also implicated in exacerbating disease severity in animal models of invasive infections [10][11][12] .Typically, sarA-mediated MRSA biofilm formation contributes to significantly elevated resistance and tolerance to antibiotics 13 .Interestingly, previous studies demonstrated that sarA regulated β-lactam antibiotic resistance in mecA-positive MRSA 7 .For instance, growth rates of sarA mutant MRSA in the presence of oxacillin were much slower than their parental strains 14 .Thus, we speculated that pharmacologic disruption of sarA expression at either the pretranscriptional or posttranscriptional level would interfere with PBP2a synthesis and resensitize MRSA to β-lactam antibiotics.In addition, the dual inhibition of both antibiotic resistance and virulence factors may prevent the resurgence of β-lactam resistance and provide additional bactericidal activity against various MRSA phenotypes.
Cinnamaldehyde (CIN) is the major component, accounting for 70%, in the essential oils of Cinnamomum cassia Presl (Lauraceae) 15 .Previous studies have demonstrated potential antimicrobial and anti-inflammatory activities of CIN [16][17][18] .For instance, addition of CIN to Escherichia coli cultures resulted in significant bacterial growth inhibition due to membrane disruption and oxidative damage 19 .In addition, CIN and substituted derivatives were previously reported to attenuate virulence in Vibrio spp.by targeting the LuxR quorum sensing response regulator and disrupting its DNA-binding activity 20 .
In the current study, we examined whether CIN could restore β-lactam antibiotic activities against MDR MRSA in vitro and in the murine models.Our experiments were successful in this respect, and we found that this potentiation effect was ascribed to the suppression of sarA expression and the increased phosphorylation level of SarA protein in the presence of CIN in combination with β-lactams.This impaired both the abilities of SarA to bind the mecA promoter and to regulate virulence-associated target genes, thus decreasing mecA-mediated MRSA resistance to β-lactam antibiotics.Our results indicate that CIN as a potential adjuvant may reestablish clinical use of older β-lactams against MRSA and provide alternatives for the treatment of MDR MRSA infections.

CIN potentiates β-lactam antibiotic activity against MRSA
Synthesis of PBP2a is a major mechanism accounting for MRSA resistance to β-lactams [21][22][23] .In the current work, we chose the 18 clinical MRSA isolates that were resistant to ampicillin (AMP) and cefotaxime (CTX) with minimum inhibitory concentrations (MICs) of 4-256 and 16-512 mg/l, respectively (Tables S1 and S2).As expected, the isogenic sarA mutant strain (JE2 ΔsarA) showed a significant decrease in β-lactam antibiotic MICs compared to the parental strain, consistent with increased susceptibility in the JE2 mecA mutant strain (JE2 ΔmecA) (Table S1).Interestingly, plasmid complementation of JE2 ΔsarA with either wild-type sarA (pALC1215) or mecA (pALC6185) restored the MIC phenotypes to levels similar to those of the isogenic parent (Table S1).
MRSA is typically associated with multiple infections and so we examined whether CIN can potentiate a β-lactam antibiotic effect against a group of 18 clinical MRSA isolates derived from human bacteremia (4 isolates), wound exudates (7 isolates), and endotracheal aspirates (7 isolates).Similar synergistic effects were observed for CIN + AMP and CIN + CTX against all MRSA isolates and these combinations generated low fractional inhibitory concentration (FIC) indices of 0.25-0.5 (Figure 1A and Table S2).Interestingly, the FIC indices of CIN combined with AMP or CTX were <0.5 for the linezolid-resistant and vancomycin-intermediate strains (Table S2).In the presence of CIN, the activities of AMP and CTX were highly potentiated, as characterized by >4-fold MIC reduction (Figure 1A,B).It is noteworthy that the JE2 ΔsarA mutant strain did not display synergism when CIN was used as a combination agent (FIC index ≥0.625).Thus, this synergistic potentiation by CIN was highly associated with sarA.As a further validation, the sarA complemented strain (JE2 ΔsarA/psarA) restored the synergistic interaction of CIN and β-lactams in the combinations (Figure 1C).
The increased activities of CIN in combination with AMP and CTX were further validated using the direct time-kill kinetic assays.Control cultures of MRSA JE2 as well as SA-70 (wound exudate) and SA-372 (bacteremia) increased by >3.0 log 10 cfu/ml after 24 h of incubation.Despite a transient growth suppression, sub-inhibitory levels of CIN, AMP, and CTX alone showed similar bacterial regrowth as the control group.In contrast, the presence of CIN + AMP and CIN + CTX combinations resulted in synergistic bactericidal effects of >3.0 log 10 (cfu/ml) reduction compared to each drug alone against all MRSA strains (Figures 1D and S1).These results suggest that CIN may be a potent combination partner for boosting the activity of β-lactam antibiotics against MRSA infections.

CIN and β-lactam antibiotic combinations inhibit sarA-mediated MRSA biofilm formation
SarA is a key regulator of MRSA biofilm formation both in vitro and in vivo 24 .We therefore investigated whether CIN potentiation is conserved beyond planktonic cells and examined whether these combinations would alter MRSA biofilm formation capacity and composition.Interestingly, when combined with AMP or CTX, CIN resulted in significant decrease in JE2 biofilm formation compared to each drug alone (p < 0.0001; Figure 2A).These results were further confirmed by looser and thinner biofilm structures for the combination treatments in scanning electron microscopy (SEM) photomicrographs (Figure 2B).Combination treatments also even destroyed mature MRSA biofilms and resulted in similar significant reduction in viable bacterial count and metabolic activity detected by 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay compared to each monotherapy (Figure 2C,D).
MRSA biofilms consist of exopolysaccharide (EPS), polysaccharide intercellular adhesin (PIA), proteins, and extracellular DNA (eDNA) in variable proportions.Thus, we examined whether any of these individual biofilm components were preferentially targeted with CIN addition.The production of EPS in the combination groups ranged from 4.72 to 24.2 mg/l, with lower average EPS (13.39 mg/l) compared to the single-drug groups (51.77 mg/l, p < 0.005; Figure 2E).Similar results were found for PIA production between single and combination groups (Figure 2F).In addition, the CIN and β-lactam combinations significantly activated the expression of the icaR gene, a negative regulator of the icaADBC operon that represses PIA-dependent biofilm production (p < 0.05; Figure 2G).In the presence of CIN, the amounts of eDNA were also markedly reduced compared to each drug alone and their respective controls (p < 0.05; Figure 2H).This is further confirmed by the higher level of extracellular nuclease production in DNase agar plates (Figure 2I).Overall, the combination of CIN with either AMP or CTX effectively inhibited MRSA biofilm formation by suppressing the biosynthesis of extracellular matrix components, including EPS, PIA, and eDNA.

CIN and β-lactam antibiotic combinations suppress MRSA virulence and inhibit infection progression
An essential aspect of MRSA pathogenicity and necessary for the progression of bloodstream and skin infections involves evasion of killing by the innate host defense peptides (HDPs), such as thrombin-induced platelet microbicidal protein (tPMP) and cathelicidin LL-37 25 .We thus exposed MRSA cells to sub-MIC CIN in the presence of tPMP or LL-37, and found a clear concentration-dependent decrease in bacterial survival for strain JE2 (Figure 3A,C).In particular, the combination of CIN and βlactams led to increased killing by tPMP or LL-37 compared to the single-drug groups (Figure 3B,D).Invasive MRSA must attach host cells and extracellular matrix ligands such as fibronectin to facilitate bacterial adhesion and internalization 26 .These abilities are crucial for establishing MRSA bacteremia.Interestingly, the presence of CIN + AMP and CIN + CTX combinations resulted in dramatic and significant (p < 0.05) decrease in MRSA binding to fibronectin (18.8%-21.8% vs. 31.2%-36.5%)and endothelial cells (5.33%-6.23%vs. 20.9%-24.4%)compared to single antibiotic treatments (Figure 3E,F).Moreover, the combined treatments showed synergistic inhibition of MRSA hemolysin activity (Figure 3G).
Consistent with these sarA-related virulence phenotypes, the relative steady-state mRNA expression of agr RNAIII, a key downstream regulator of sarA, and other virulenceassociated genes (fnbA and hla) were significantly inhibited after exposure to CIN alone or in combination with β-lactams (Figure 3H-J).In particular, exposure to sub-MIC levels of AMP or CTX resulted in increased fnbA expression compared to controls.In contrast, this effect was completely abolished when administered together with CIN (Figure 3I).These results suggest that CIN most likely alters the regulatory functions of sarA, consequently resulting in the inhibition of MRSA infection progression.
In order to obtain more insights into the molecular mechanism underlying CIN activity and CIN-induced mRNA expression level changes, we performed a transcriptome analysis of strain JE2 following CIN exposure.We identified 446 differentially expressed genes (DEGs) between the CIN treatment and control groups, including 156 upregulated and 290 downregulated genes (Figure 4A and Table S3).Gene Ontology (GO) analysis of these DEGs provided major functional annotations for biological processes (small-molecule metabolic and organic substance biosynthetic processes), cellular components (intracellular part), and molecular functions (cation and nucleotide binding) (Figure 4B).Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis revealed that the markedly downregulated genes were associated with S. aureus infection and two-component system pathways (Figure 4C).In particular, the genes encoding microbial surface components recognizing adhesive matrix molecules (MSCRAMM), such as clfA (clumping factor A), spa (staphylococcal protein A), fib, and sdrC (serineaspartate repeat protein C), were dramatically downregulated (Figure 4D).These reductions were associated with decreased MRSA adhesion to endothelial cells in the presence of sub-MIC levels of CIN.Consistently, the expression of the staphylokinase (sak) and IgG-binding protein (sbi) genes was reduced in the CIN treatment groups compared to the controls (Figures 4D and S2).We further verified the transcriptome results using quantitative real-time-PCR (qRT-PCR), and their expression profiles were highly similar (Data not shown).These results suggest that CIN suppresses the secretion of virulence-associated enzymes in MRSA and prevents production of components that compromise the host immune defense.
CIN exposure alters SarA phosphorylation and binding to the mecA promoter that contributes to potentiation of anti-MRSA activity for β-lactam antibiotics Given that sarA is an important positive regulator of virulence factor expression, we thus speculated that the regulation of virulence genes by CIN is achieved through targeting SarA.Indeed, there is precedent for these types of CIN effects as has been demonstrated for LuxR 20 .As expected, CIN was able to decrease the expression of sarA in a concentration-dependent manner (Figure S3) as well as directly interact with the SarA protein (Figure 5A).In particular, the arginine Arg A: 90 (R90) basic residue is a key target DNA site for SarA protein family members 27 .The in silico molecular docking analysis indicated that CIN could interact with the DNA-binding active sites of SarA protein with a binding energy of −4.1 kcal/mol, and form a single Hydrogen-bonding interaction with the R90 residue (Figure 5A).This interaction may alter the shape of SarA binding pockets when they bind to CIN.Addition of CIN + AMP and CIN + CTX combinations to MRSA cell cultures resulted in significant reduction of both sarA and mecA expression compared to either drug alone (p < 0.0005; Figure 5B,C).Interestingly, the mecA suppression was accompanied by sarA suppression.This suggests that sarA may positively regulate mecA expression by direct interaction of SarA protein at the mecA promoter.We examined this possibility via electrophoresis mobility gel shift assays (EMSAs) using mecA promoter element oligonucleotides and purified SarA protein.The presence of the SarA protein generated clear band-shifts, indicating direct binding of SarA to the mecA promoter fragment.In particular, sub-MIC levels of CIN remarkably reduced the binding of the mecA promoter to various concentrations of SarA protein (Figure 5D).
As a global regulator, SarA played an important role in target gene regulation at the posttranscriptional level.We therefore detected the phosphorylation level of SarA treated with CIN.In line with a previous study 28 , SarA was mainly phosphorylated by serine/threonine kinase Stk1 due to the inability to autophosphorylate in vitro (Figure S4A).As shown in Figure 5E,F, treatments with CIN increased the phosphorylation levels of SarA in the presence of Stk1 in a concentration-dependent manner.A slight upregulation in stk1 expression was observed after treatment with CIN (Figure 5G).Compared to monotherapy, the combinations of CIN and AMP/CTX demonstrated higher phosphorylation levels of SarA (p < 0.05; Figure S4B,C).It has been reported that phosphorylation of SarA weakens its DNA-binding capacity 28,29 .Thus, the effects of CIN alone or in combination with AMP/CTX, as demonstrated, are most likely due to a promotion of SarA phosphorylation that influences the differential binding of SarA to the mecA promoter, subsequently interfering with mecA expression.Together, these results indicated that treatments with CIN alone or in combination with β-lactam antibiotics decreased sarA-mediated virulence and mecA-mediated resistance in MRSA by suppressing sarA expression and facilitating SarA phosphorylation.

CIN enhances the efficacy of β-lactam antibiotics in the murine models of MRSA bacteremia and subcutaneous catheter-related biofilm infection
Therapeutic applications of antibiotics and plant-natural compound combinations have been proven to be beneficial, as outlined in numerous clinical studies 30,31 .However, novel combination therapies must first be evaluated for toxicity.To assess the potential treatment toxicity of CIN used as an antibiotic adjuvant, we analyzed hemolysis, histopathologic, and cytotoxicity of CIN alone or in combination with β-lactams.The addition of CIN to blood cells at a high concentration of 12,500 mg/l did not result in any significant hemolytic activity, either alone or in combination with β-lactams.Similarly, no visible cytotoxicity was observed in either HEK-293 or Vero cells for these treatments (Figure S5).Furthermore, histologic evaluation of kidney tissues retrieved from the monotherapy and combination therapy groups revealed no significant nephrotoxicity compared to the controls (Figure S6).These data indicated that neither CIN nor the combination therapies showed toxicity, suggesting a good safety profile of CIN as a potential synergist to enhance β-lactam antibiotic efficacy.
MRSA is among the leading causes of biofilm colonization and bloodstream infections such as bacteremia and sepsis.Given the promising potentiation effect by CIN in vitro, we therefore investigated the potential of CIN to restore β-lactam antibiotic efficacy in the murine bacteremia and catheterrelated biofilm infection models (Figure 6A).In the bacteremia model, mice infected with strains JE2 and MW2 did not respond to CIN or the β-lactams AMP and CTX monotherapies, and MRSA densities in target tissues were similar to those in the respective untreated controls (Figures 6B and S7).Importantly, the CIN + AMP and CIN + CTX combination treatments resulted in highly significant (p < 0.0001) reduction of MRSA densities in kidneys (>4.1 log 10 [cfu/g]) and spleens ( > 1.7 log 10 [cfu/g]) of infected mice compared to each monotherapy (Figure 6B).In the murine catheter-related biofilm infection model, both bioluminescence intensity and MRSA burden in catheters in the untreated controls increased considerably at 3 days after biofilm colonization (Figure 6C).At the end of therapy, despite the lower MRSA density, CIN, AMP, or CTX monotherapy did not reduce the MRSA biofilm burden compared to the initial colonization (~10 6 cfu/catheter; Figure 6D).Interestingly, the CIN + AMP and CIN + CTX combination therapies resulted in >2.2 log 10 cfu reduction of MRSA in catheters compared to the monotherapy groups.Similarly, the bioluminescence from mice receiving the combination treatments with CIN + AMP and CIN + CTX showed a significant decrease compared to that from each monotherapy group (Figure 6C,D).

DISCUSSION
Clinical treatment options for MDR MRSA infections are limited to last-line antibiotics such as vancomycin and daptomycin.This problem has been exacerbated by the recent emergences of vancomycin-intermediate and -resistant S. aureus (VISA and VRSA) 32 .Thus, effective therapeutic strategies are urgently needed against such resistant pathogens.A feasible approach to overcome bacterial resistance is to develop bioactive adjuvants to revitalize existing antibiotics.A well-known paradigm of this strategy is the coadministration of β-lactam antibiotics (e.g., amoxicillin) alongside β-lactamase inhibitors (e.g., clavulanic acid) that helps to neutralize β-lactamases 33 .Other studies also reported diverse adjuvants targeting bacterial MDR mechanisms to increase the efficacies of existing antibiotics 30,34 .For example, the active compounds pterostilbene and metformin could effectively restore the activities of carbapenem and tetracycline against MDR bacteria, respectively 35,36 .These approaches were similar to our current work demonstrating a potentiation effect from CIN and β-lactam antibiotic combinations.
The importance of global regulatory locus sarA in β-lactam antibiotic resistance has been reported previously in vitro and in an experimental endocarditis infection model 7 .The present study further investigated whether the synergistic effect of CIN and β-lactams was associated with the sarA regulon.We demonstrated that the sarA mutation rendered MRSA JE2 more susceptible to AMP and CTX killing compared to the parental strain.Complementation of the sarA mutant strain with mecA restored β-lactam resistance.We also found significant synergy for this combination treatment against all clinical MRSA strains, except for the sarA knockout mutant.These results suggest that sarA is a promising target for the development of antibiotic adjuvants against MDR MRSA infections.In addition, the present study showed that CIN could inhibit sarA expression in a concentration-dependent manner in the range of 78.12-312.50mg/l (Figure S3).In particular, the level of sarA inhibition by the high concentration of CIN at 312.50 mg/l was comparable to that observed with 156.25 mg/l CIN when used in combination with AMP or CTX (Figures 5B and S3).We thus speculated that the disorganization of the cell wall by AMP or CTX may increase MRSA exposure to CIN 37 .This speculation was subsequently corroborated by our preliminary analysis using high-performance liquid chromatography (HPLC), which demonstrated higher CIN accumulation in MRSA cells in the presence of AMP or CTX (data not shown).More importantly, a significant reduction in sarA expression was observed for all CIN-based combination therapies compared to each monotherapy.These results indicated that sarA inhibition by CIN + AMP and CIN + CTX combinations may play an important role in triggering this potentiation effect.Besides offering mechanistic insights into transcription regulation, our studies revealed that CIN promoted the phosphorylation level of SarA and interfered with SarA binding to the mecA promoter, resulting in the inhibition of PBP2a expression encoded by mecA (Figure 7).This process thereby re-sensitized MRSA to β-lactam therapies.
MRSA is recognized as a common pathogen in both nosocomial and community settings, often infecting soft tissues and forming tenacious biofilms on these surfaces 38 .Biofilm formation enables further MRSA colonization of host tissues and provides an antibiotic-impenetrable structure that also acts to exclude host immune factors 39 .Previous studies demonstrated that sarA was a positive regulator of MRSA biofilm formation because deletion of the sarA gene resulted in a loose biofilm architecture 7,40,41 .This phenotype was partly due to lack of the positive regulation of the icaADBC operon by sarA resulting in decreased levels of poly-N-acetylβ-(1-6)-glucosamine that functions as the PIA 42 .Furthermore, biofilm-associated protein (Bap) expression was also activated by sarA via an agr-independent mechanism 43 .Consistently, our results indicated that CIN alone or in combination with β-lactams effectively reduced the levels of EPS, eDNA, and PIA.In parallel, the combinations of CIN and β-lactams reduced the steady-state levels of sarA mRNA, resulting in increased expression of the icaR repressor of the icaADBC operon.Numerous studies suggest that sarA is a global transcriptional regulator of virulence genes including fnb, hla, cap5, and agr, and plays a key role in MRSA infection establishment 43,44 .In particular, MRSA capability to circumvent elimination by locally secreted HDP is a necessary step for bacteremia 45 .Furthermore, sarA is a regulator of adhesion and internalization in cases of invasive MRSA infections 46,47 .Previous studies demonstrated that SarA protein positively modulated the production of fibrinogen binding protein (FnBP) by binding to the fnbA promoter fragments, thus facilitating adhesions to extracellular matrix ligands (fibronectin and fibrinogen) and to endothelial cells and platelet surface proteins 24,47 .However, there is an inconsistent result that AMP or CTX induced fnbA expression but repressed sarA expression in this study (Figures 3I and 5B).This result may be attributed in part to the complexity of the sarA virulence regulon, and repression of sarA may have pleiotropic effects that are difficult to link to a specific virulence regulatory pathway for fnbA.Thus, fnbA may be regulated by multiple upstream factors in addition to sarA, and this is consistent with previous findings [48][49][50] .Importantly, CIN in combination with AMP or CTX decreased the expression of these pathogenic virulence genes including fnbA and sarA and resulted in consistent phenotypic observations.In addition, CINtreated MRSA cells showed higher susceptibility to the cationic HDPs (tPMP and LL-37) and β-lactam killing in vitro, reduced fibronectin adherence and minimal damage to endothelial cells, unstable biofilm, and decreased α-hemolysin levels.These findings indicated that CIN at sub-MIC concentrations clearly downregulated virulence factor expression that enhanced the efficacy of β-lactam antibiotics against MRSA.
Our results also add to the growing body of research that helps explain improved clinical outcomes with the use of multifunctional natural products 51 .However, we did observe an unsatisfactory result from the AMP and CIN combination therapy in blood versus AMP monotherapy (Figure 6B).In addition to MRSA enrichments in the kidney and spleen, this diminished bacterial response to AMP-based combination therapy in blood was most likely a result of the shorter half-life and lower area under the concentration-time curve in plasma for AMP compared to CTX [52][53][54] .Nevertheless, all combination therapies achieved uniform significant reduction of MRSA densities in both the kidney and spleen compared to each monotherapy in the murine bacteremia model.Repurposing CIN from its current use as a food flavoring agent to a SarA inhibitor dramatically interfered with MRSA biofilm formation.This is especially important since biofilm-associated infections are typically resistant to antibiotic therapy.As a result, CIN-treated MRSA biofilms became susceptible to β-lactam antibiotic therapies in a murine model of catheter-related biofilm infection.Similarly, a previous study has also reported that ΔsarA MRSA strains are highly susceptible to vancomycin therapy in a rabbit endocarditis model 7 .In addition, our work contributes a context for the transformation of in vitro studies into in vivo treatment outcomes.CIN-mediated suppression of MRSA virulence factors occurs predominantly by targeting SarA but is not bactericidal activity when used alone; rather, CIN functions by promoting SarA phosphorylation and perturbing SarA regulation of mecA and synergizes with βlactams to kill MRSA.The unique aspect of this approach is the targeting of a global regulator of MRSA and in general, virulence determinants that affect numerous infection-related processes, and could thus be used in the development of alternative treatment strategies for MDR MRSA infections.
In summary, we demonstrated that β-lactam antibiotic activity was increased by the presence of CIN against MRSA and biofilm-associated infections both in vitro and in vivo.The combinations of CIN and β-lactams inhibited sarA gene expression and increased SarA protein phosphorylation.This impaired SarA-mecA binding capacity and thus contributed to CIN potentiation effects on β-lactam antibiotics.Together, our data suggest that CIN represents a promising β-lactam antibiotic adjuvant to combat MDR MRSA infections and inspires us to find more effective candidates for use as potential antibiotic adjuvants.

Chemicals and reagents
CIN (>95% purity) was purchased from Yuanye Biotechnology Company Ltd. and dissolved in dimethyl sulfoxide.AMP and CTX powders were purchased from Sigma.

Bacterial strains and growth medium
MRSA strains used in this study are listed in Tables S1 and S2.Strain JE2 is a plasmid-cured derivative from the communityassociated MRSA (CA-MRSA) LAC (USA300) and its isogenic mecA mutant strains were obtained from the Nebraska Transposon Mutant Library (NTML) 7,55 .A sarA gene knockout in JE2 (JE2 ΔsarA) was generated by transduction of the sarA::kan mutation from strain ALC2543 with phage 85 7 .Strain JE2 ΔsarA was complemented by transformation with plasmids pALC1215 (JE2 ΔsarA/psarA) and pALC6185 (JE2 ΔsarA/pmecA) as described previously 7,24,56 .The JE2 and its isogenic mutant and complemented strains were used for subsequent mechanistic studies.CA-MRSA strain MW2 (USA400) was used for in vivo experiments.An additional 18 clinical MRSA isolates for the checkerboard assays were derived from human bacteremia, wound exudates, and endotracheal aspirate (Table S2).All strains were routinely grown, subcultured, and quantified using tryptic soy broth (TSB; Becton Dickinson) or on TSB agar plates.Brain-heart infusion broth (BHI) was used for biofilm assays.

In vitro susceptibility testing and checkerboard test
The MICs of CIN, AMP, and CTX against MRSA strains were determined using the Clinical and Laboratory Standards Institute (CLSI) microdilution method 57 .All MIC determinations were performed in duplicate with three biological replicates.The median MIC of replicate assays was reported.Chequerboard tests were performed to assess interactions between CIN and β-lactam antibiotics in standard 96-well plates as reported previously 58 .An FIC index of ≤ 0.5 indicates a synergistic effect 59 .

In vitro time-kill curves
Time-kill experiments were performed to evaluate the synergistic activity of CIN and β-lactam antibiotic combinations as previously described 30 .Briefly, we used an initial inoculum of ~10 5.5 cfu/ml of logarithmic-phase MRSA cells in the presence of sub-MICs of CIN alone or in combination with AMP or CTX.Serial dilutions of MRSA cultures obtained at 0, 3, 6, 9, and 24 h were taken for bacterial density measurement using the plate count technique.Synergistic activity was considered as a reduction of ≥2 log 10 (cfu/ml) at 24 h by the combination compared to each drug alone 58 .

Biofilm formation, viability, and microscopic analysis
Exponential MRSA cells were adjusted to an OD 600 of 0.5 and diluted 1:100 into BHI broth supplemented with 0.5% glucose 26 .The suspensions were transferred to six-well plates containing CIN (312.50 mg/l) alone or in combination with AMP (16 mg/l) or CTX (32 mg/l) and incubated for 72 h at 37°C.After incubation, the plates were then gently washed with phosphate-buffered saline (PBS), air-dried, and stained using 0.1% crystal violet.The adhering dye was dissolved in 30% acetic acid and absorption was measured at OD 600 to quantify biofilm formation.For the mature MRSA biofilm cultured for 72 h, the capacity of CIN (156.25 mg/l), AMP (8 mg/l), and CTX (16 mg/l) alone or in combination to kill biofilm-embedded bacteria was determined by viable bacterial count and MTT metabolic assays 60,61 .Biofilm formation was confirmed by SEM as previously described 18,62 .In brief, pieces of plate material were fixed with 2.5% glutaraldehyde and 0.1% OsO 4 for 1 h, chemically dried in a graded ethanol series (30%-100%), and sputter-coated with gold for SEM to analyze adherence and biofilm formation.

EPS, eDNA, and nuclease production
MRSA biofilms were treated with CIN (156.25 mg/l), AMP (4 mg/l), and CTX (8 mg/l) alone or in combination.After treatment, biofilms were rinsed, removed, and dispersed by sonication at 20 kHz for 5 s.The suspension was collected for EPS determinations and quantified using the anthrone-sulfuric acid colorimetric assay 26,63 .Release of eDNA was determined from a 72 h biofilm treated with CIN (156.25 mg/l), AMP (8 mg/l), and CTX (16 mg/l) alone or in combination using a microplate fluorescence assay with Hoechst 33258 dye 26,64 .Extraction and purification of biofilm eDNA were performed as previously described 64 .The eDNA was quantified using an EnSight fluorescence plate reader (Perkin Elmer) at Ex 350 /Em 460 .Extracellular nuclease activities of MRSA after exposure to CIN (78.12 mg/l), AMP (4 mg/l), and CTX (8 mg/l) alone or in combination, were determined using the DNase test agar 26 .

PIA assay
Colony morphology on Congo red agar (CRA) was used as a surrogate marker of PIA production as described previously 65 .To assess the effect of CIN on slime production as characterized by PIA, a 10 µl aliquot of bacterial solution (OD 600 = 1.2) was inoculated on CRA plates containing 0.25-0.5 mg/l AMP and 1-2 mg/l CTX alone or in combination with CIN (312.50 mg/l), and the plates were observed after incubation for 24 h at 37°C 66 .

In vitro susceptibility to HDPs
Our previous study of in vitro HDP susceptibilities to MRSA and in vivo treatment outcomes involved the use of tPMP and cathelicidin LL-37 26 .We therefore used these peptides in this study, and the tPMP preparation and bioactivity assays were performed as detailed elsewhere 67 .Cathelicidin LL-37 was purchased from Eurogentec.MRSA cells were treated with/without CIN (78.12-312.50mg/l), AMP (8 mg/l), and CTX (16 mg/l) alone or in combination for 2 h.After the pretreatment, cells were gently washed three times with PBS.In vitro HDP susceptibilities were assessed by adding tPMP (1 mg/l equivalent) to 10 4 cfu/ml MRSA cells and LL-37 (5 mg/l) to 10 5 cfu/ml MRSA cells.The peptide concentrations were selected based on the pilot studies identifying their levels that did not rapidly kill MRSA cells over 2 h incubation period 26 .The results were expressed as the percentage (±SD) of the initial bacterial inoculum that survived under HDP exposure.

Measurement of adherence to fibronectin and endothelial cells
To measure the effect of CIN on the ability of MRSA to adhere to fibronectin, six-well plates were coated with 50 mg/l purified human fibronectin (Sigma) overnight at 4°C and washed with PBS.Plates were then treated with 3% bovine serum albumin to prevent nonspecific adhesion 25,26 .Logarithmic-phase MRSA cells were pretreated with/without CIN (156.25 mg/l), AMP (8 mg/l), and CTX (16 mg/l) alone or in combination for 2 h, and then added at a final inoculum of 5 × 10 3 cfu/ml to fibronectin-coated plates.Unbound MRSA cells were removed by gently washing three times with PBS, and melted TSA was added to each well and allowed to solidify at 37°C.The level of adherence was quantified by determining the percentage of the initial bacterial inoculum bound to fibronectin.
Human aortic endothelial cells (HAECs) were prepared and cultured as previously described 68,69 .After pretreatment with CIN (39.06 mg/l) alone or in combination with AMP (0.25 mg/l) or CTX (0.5 mg/l), MRSA cells (5 × 10 5 cfu/ml) were added to confluent endothelial cell monolayers in six-well plates (MOI = 1:1) and incubated for 2 h under static conditions 26 .Cells were washed with Hanks balanced salt solution to remove the unbound bacteria, followed by permeabilization with 1.0% Triton X-100, after which bacterial counts were directly determined by plating the appropriate serial dilutions on TSA.Adherence was calculated as the percentage of the initial bacterial inoculum bound to HAEC cells.

Hemolysis quantification assay
The hemolytic activity of MRSA strains after exposure to CIN alone or in combination with β-lactam antibiotics was determined as previously described 70,71 .Briefly, culture supernatants lacking bacterial cells were collected from control and drug-treated (39 mg/l of CIN, 0.25 mg/l of AMP, and 0.5 mg/l of CTX alone or in combination) MRSA strains by centrifugation at 6000 rpm for 10 min.Hemolytic activity was determined by incubating the cell-free culture supernatants with an equal volume of defibrillated sheep blood cells (8% in PBS) for 20 min at 37°C.After incubation, the absorbance of released hemoglobin in the supernatant was determined at OD 543 to obtain the hemolysis rate (HR) using the following equation 70 : HR(%) = OD treated /OD control × 100%.

Transcriptomic analysis
To determine the gene expression profile of MRSA in the presence of CIN, MRSA JE2 cultures were treated with or without CIN (156.25 mg/l) for 4 h.Total RNA was extracted using an EZNA bacterial RNA kit (Omega Biotek) and pairedend sequenced with a read length of 2 × 150 bp on an Illumina NovaSeq.6000 system (Novogene).Raw sequencing reads were filtered and mapped against the S. aureus NCTC 8325 reference genome (NC_007795.1).A Cuffdiff analysis was performed based on FPKM (Fragments per Kilobase of transcript per Million mapped reads) to identify DEGs between the control and the treatment group.Criteria for selecting DEGs were fold change (FC) values > 2 and p < 0.05.Three biological replications were conducted for each treatment.

qRT-PCR analysis
Exponential phases of MRSA cells were pretreated with CIN (78.12-312.50mg/l) alone or in combination with AMP (8 mg/l) or CTX (16 mg/l).Total RNA was extracted from MRSA cells using the RNeasy extraction kit (Qiagen) as per the manufacturer's instructions.A HiScript III 1st Strand cDNA Synthesis Kit (Vazyme) was used for cDNA synthesis.Gene expression was quantified by qRT-PCR using Taq Pro Universal SYBR qPCR Master Mix (Vazyme).Three biological replicates each containing three technical replicates were carried out for each gene.Amplification of sarA, mecA, icaR, fnbA, RNAШ, hla, and stk1 genes was performed as previously described in detail, using gene-specific primers to verify transcriptome results 30 .The relative abundance of gyrB mRNA was used as an internal control.

Molecular docking analysis
To evaluate the binding interactions of CIN with the SarA protein, AutoDock v. 4.2.6 (https://autodock.scripps.edu/)was used for in silico docking analysis.The initial X-ray crystal structure of the SarA protein was obtained from the Protein Data Bank (https://www.rcsb.org/)with PDB code 2FRH.The CIN chemical structure (C 9 H 8 O; PubChem ID: 637511) was obtained from the NCBI PubChem database (https://pubchem.ncbi.nlm.nih.gov/).The outputs from AutoDock and all modeling studies as well as images were rendered with PyMol v. 1.5.0.3 (https://pymol.org/2/).

In vitro SarA phosphorylation assay
To determine the phosphorylation-level changes of SarA treated with CIN alone (78.12-312.50mg/l) or in combination with AMP (8 mg/l) or CTX (16 mg/l), the expression and purification of SarA and Stk1 proteins were performed as described previously in detail 72 .In brief, plasmid pET28a was transformed into E. coli BL21 (DE3) for the fusion protein expression, and His-tagged SarA and Stk1 proteins were purified using an NGC chromatography system (BioRad).In vitro phosphorylation of purified SarA protein was performed as previously described, with some modifications 28,73 .In brief, 2 μg of SarA was added to kinase buffer together with 1 mM ATP and 1 µg of Stk1 in the presence or absence of CIN (0-312.50mg/l) for 20 min at 37°C.The reaction was terminated with addition of SDS loading buffer and heated at 95°C for 5 min.Subsequently, samples were separated on polyacrylamide gels supplemented with Phos-Tag acrylamide.Protein mobility shift and intensity were quantified with ImageJ software (https://imagej.net/).

The impact of CIN on SarA-mecA binding
EMSAs were used to investigate whether SarA directly regulates mecA expression and the effect of CIN on SarA-mecA promoter binding.Purified SarA protein and the 200 bp DNA fragment containing the mecA promoter region were incubated in the presence and absence of 156.25 mg/l CIN for 20 min at room temperature.The gel was stained with SYBR green using a commercially available EMSA kit (Invitrogen) and visualized by UV epi-illumination at 302 nm.

In vivo murine MRSA bacteremia model
Six-week-old specific-pathogen-free ICR mice (25-27 g) were obtained from Guangdong Medical Lab Animal Center (Guangzhou, China).The Animal Research Committee (IACUC) of the South China Agricultural University approved the animal studies (#2021C086).In the murine bacteremia model, mice were infected via the tail vein with a 0.25 ml MRSA bacterial suspension delivering ∼10 5.5 cfu/mouse 26 .At 24 h postinfection, mice were randomized to receive (i) control treatment; (ii) CIN at 20 mg/kg intravenously (i.v.) once daily (qd); (iii) AMP at 40 mg/ kg orally (p.o.) three times daily (tid); (iv) CTX at 100 mg/kg subcutaneously (s.c.) twice daily (bid); and (v) a combination of CIN with CTX or AMP.These doses were chosen to mimic pharmacokinetic values similar to those achieved by the clinical dosing of humans (i.e., 0.5 g AMP p.o. tid, and 1.0 g CTX i.v.bid) 53,54,74,75 .Treatments lasted for 3 days using six to eight mice per group.After killing the mice, the target tissues (blood, kidney, and spleen) were removed, weighed, and quantitatively cultured for CFU determinations.MRSA burdens in target tissues were determined from each treatment, and expressed as the mean (±SD) of log 10 (cfu/g or cfu/ml).

In vivo murine model of subcutaneous catheter-related MRSA biofilm infection
An in vivo bioluminescence imaging system (IVIS) was used for noninvasive monitoring of the therapeutic efficacy of CIN alone or in combination with AMP or CTX in a murine catheter-related MRSA biofilm model 76 .Mice were infected by subcutaneously implanting a precolonized Teflon catheter segment (1 cm) inoculated with the lux bioluminescent MRSA strain Xen30 (Perkin Elmer) at ~10 6 cfu/catheter 77 .At 1 day postimplantation, mice received either no therapy (control), CIN at 20 mg/kg i.v.qd, AMP at 80 mg/kg p.o. tid, and CTX at 100 mg/kg s.c.bid or their combinations.Mice were imaged after 3 days of implantation for bioluminescent signals (BLS) using the IVIS Lumina imaging system (PerkinElmer), which was presented as radiance (photons/s/cm 2 /steradian) using a pseudocolor scale (red, most intense; blue, least intense) 76,78 .After killing the mice, the catheters were quantitatively cultured using the standard assays and reported as cfu/catheter 79 .

Safety assessment
The effect of CIN alone or in combination with β-lactam antibiotics on the hemolytic activity was determined as previously described 36 .In brief, a 2% sheep red blood cell suspension was incubated with CIN (1560-12,500 mg/l), AMP (1280 mg/l), and CTX (1280 mg/l) alone or in combination at 37°C for 1 h.Triton X-100 (0.2%) and saline were used as positive and negative controls, respectively.The hemolytic activity (HA) was calculated according to the following equations: HA = 1−(OD sample − OD negative )/(OD positive −OD negative ) × 100%.In vitro cytotoxicity of CIN alone or in combination with β-lactam antibiotics against human embryonic kidney (HEK-293) cells and Vero cells (2 × 10 5 cells/well) was determined using the CCK-8 assay 30 .For in vivo toxicity evaluation, mice receiving CIN (20 mg/kg) or CTX (100 mg/kg) monotherapy and combination therapy were sacrificed after 3 days of treatment, and their kidneys were harvested for hematoxylin and eosin staining using a standard staining protocol 80 .Mice treated with saline of the same volume served as controls.

Statistical analyses
Statistical analysis was performed using GraphPad Prism 8.0 software.All data were presented as mean ± SD.Student's t test was used for statistical comparisons between groups.Differences with p ≤ 0.05 were considered statistically significant.

Figure 3 .
Figure 3. Impact of CIN and β-lactam antibiotic combination on MRSA virulence and pathogenicity.(A-D) In vitro killing of MRSA JE2 by the host defense peptides thrombin-induced platelet microbicidal protein (tPMP) (A, B) and cathelicidin LL-37 (C, D) in the presence of CIN alone or in combination with AMP (8 mg/l) or CTX (16 mg/l).CIN level for combination treatments with AMP or CTX in panels (B) and (D) was 78.12 mg/l.(E, F) Adherence of MRSA to immobilized fibronectin (E) and EC (F) in the presence of CIN, AMP, CTX alone, or their combination.(G) Effects of CIN alone or in combination with AMP (0.25 mg/l) or CTX (0.5 mg/l) on hemolysin production of MRSA JE2.(H-J) Relative gene expression levels of agr RNAIII (H), fnbA (I), and hla (J) in the presence of sub-MIC concentrations of CIN, AMP, CTX alone, or in combination.Data represent the mean ± SD of three biological replicates, and statistical significances were determined using an unpaired Student's t test.EC, endothelial cells.

Figure 4 .
Figure 4. Transcriptome analysis of MRSA JE2 after exposure to sub-MIC of CIN.(A) Volcano plot of the differentially expressed genes (DEGs) in MRSA JE2 treated with 156.25 mg/l CIN for 4 h.The x and y axes represent the FCs of expression and corresponding statistically significant degree, respectively.(B) Gene Ontology (GO) annotation classification analysis of downregulated DEGs in the biological process, cellular component, and molecular functions.(C) Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis showing the 10 most significant biological pathways enriched in the downregulated DEGs.(D) Differential expression of selected downregulated genes associated with S. aureus infection, the citrate cycle, and a two-component system.Data are presented as the means of three biological replicates.FC, fold change.

Figure 5 .
Figure 5. Synergistic mechanisms of CIN with β-lactam antibiotics against MRSA.(A) Molecular docking analysis of interaction patterns of CIN to SarA protein.CIN interacts with Arg A: 90 (R90) in the DNA-binding active site of SarA.(B, C) Relative expression levels of sarA (B) and mecA (C) in MRSA JE2 after exposure to CIN (156.25 mg/l) alone or in combination with AMP (8 mg/l) or CTX (16 mg/l).(D) Effects of CIN exposure on SarA-mecA promoter binding.(E, F) SarA phosphorylation levels in the presence of CIN (0-312.50mg/l).SarA: 2 μg; Stk: 1 µg; ATP: 1 mM.(G) Relative expression levels of stk1 after exposure to CIN (0-312.50mg/l).Data represent the mean ± SD of three biological replicates, and statistical significances were determined using an unpaired Student's t test.

Figure 6 .
Figure 6.CIN enhances the efficacy of β-lactam antibiotics in two murine infection models.(A) Experimental protocols for the murine models of bacteremia and subcutaneous catheter-related biofilm infection due to MRSA.(B) Densities of MRSA JE2 strain in the kidney, spleen, and blood after a 3-day β-lactam antibiotic monotherapy or in combination with CIN in the murine bacteremia model.CIN at 20 mg/kg, intravenously (i.v.), once daily (qd); AMP at 40 mg/kg, orally (p.o.), three times daily (tid); CTX at 100 mg/kg, subcutaneously (s.c.), twice daily (bid).Each dot represents one mouse and the horizontal lines indicate the means of results from groups of six to eight mice.(C) Monitoring of therapeutic efficacy of AMP (80 mg/kg, p.o., tid), CTX (100 mg/kg, s.c., bid), or CIN (20 mg/kg, i.v., qd) monotherapy or combination therapy in a murine subcutaneous catheter-related MRSA biofilm model using in vivo bioluminescent imaging.(D) Viable bacterial counts of MRSA attached on the catheters after monotherapy or combination therapy.Statistical significances were determined using an unpaired Student's t test.IVIS, in vivo bioluminescence imaging system.