Combination Therapy Using Low-Concentration Oxacillin with Palmitic Acid and Span85 to Control Clinical Methicillin-Resistant Staphylococcus aureus

The overuse of antibiotics has led to the emergence of multidrug-resistant bacteria, such as methicillin-resistant Staphylococcus aureus (MRSA). MRSA is difficult to kill with a single antibiotic because it has evolved to be resistant to various antibiotics by increasing the PBP2a (mecA) expression level, building up biofilm, introducing SCCmec for multidrug resistance, and changing its membrane properties. Therefore, to overcome antibiotic resistance and decrease possible genetic mutations that can lead to the acquisition of higher antibiotic resistance, drug combination therapy was applied based on previous results indicating that MRSA shows increased susceptibility to free fatty acids and surfactants. The optimal ratio of three components and the synergistic effects of possible combinations were investigated. The combinations were directly applied to clinically isolated strains, and the combination containing 15 μg/mL of oxacillin was able to control SCCmec type III and IV isolates having an oxacillin minimum inhibitory concentration (MIC) up to 1024 μg/mL; moreover, the combination with a slightly increased oxacillin concentration was able to kill SCCmec type II. Phospholipid analysis revealed that clinical strains with higher resistance contained a high portion of 12-methyltetradecanoic acid (anteiso-C15:0) and 14-methylhexadecanoic acid (anteiso-C17:0), although individual strains showed different patterns. In summary, we showed that combinatorial therapy with a low concentration of oxacillin controlled different laboratory and highly diversified clinical MRSA strains.


Introduction
Over decades, the overuse of antibiotics has brought about multidrug-resistant bacteria [1]. Among these, methicillin-resistant Staphylococcus aureus (MRSA) is difficult to treat in communities and healthcare facilities owing to its quick spread and multidrug resistance [2]. To be more specific, they biofilm formation of community-associated MRSA (CA-MRSA) LAC and its ∆agr mutant, which has a higher oxacillin minimum inhibitory concentration (MIC) (>200 µg/mL) [9]. Concomitant treatment with palmitic acid and oxacillin led to a dramatic increase in the efficacy of oxacillin. Similarly, span85, which is mainly used in medicines, cosmetics, textiles, paints, and petroleum as an emulsifier, thickener, anti-rust agent, and biodegradable surfactant based on a natural fatty acid (oleic acid) and sugar alcohol (sorbitol), eliminated the biofilm of the MRSA strains and decreased the MIC of oxacillin on MRSA. However, the effect of palmitic acid and span85 was investigated only at fixed concentrations, and the combinatorial effect was not investigated for possible applications. Considering that soaps contain more than 10% palmitic acid and that span85 is used at a concentration of 0.5-5% in drugs and cosmetics, these compounds could be used to control resistant bacteria causing several skin issues.
In an initial analysis, the effects of different concentrations of palmitic acid and span85 were tested for dose-dependency; furthermore, ∆agr mutant, which is more resistant than the LAC MRSA strain, was used to evaluate the antibacterial activity of palmitic acid and span85. The results showed that palmitic acid at concentrations lower than 1 mg/mL exerted no significant effect on the growth of the ∆agr mutant cells ( Figure 1A). However, it started to inhibit the cell growth when the concentration was over 1 mg/mL. Span85 also inhibited the growth of the ∆agr mutant cells at concentrations over 0.1% (v/v) ( Figure 1B). Collectively, the MICs for each antibacterial compound are calculated (oxacillin = 256 µg/mL, palmitic acid = 2 mg/mL and span85 = 2% (v/v) ≈ 19.14 mg/mL) Antibiotics 2020, 9, x FOR PEER REVIEW 3 of 11 which has a higher oxacillin minimum inhibitory concentration (MIC) ( >200 μg/mL) [9]. Concomitant treatment with palmitic acid and oxacillin led to a dramatic increase in the efficacy of oxacillin. Similarly, span85, which is mainly used in medicines, cosmetics, textiles, paints, and petroleum as an emulsifier, thickener, anti-rust agent, and biodegradable surfactant based on a natural fatty acid (oleic acid) and sugar alcohol (sorbitol), eliminated the biofilm of the MRSA strains and decreased the MIC of oxacillin on MRSA. However, the effect of palmitic acid and span85 was investigated only at fixed concentrations, and the combinatorial effect was not investigated for possible applications.
Considering that soaps contain more than 10% palmitic acid and that span85 is used at a concentration of 0.5-5% in drugs and cosmetics, these compounds could be used to control resistant bacteria causing several skin issues. In an initial analysis, the effects of different concentrations of palmitic acid and span85 were tested for dose-dependency; furthermore, Δagr mutant, which is more resistant than the LAC MRSA strain, was used to evaluate the antibacterial activity of palmitic acid and span85. The results showed that palmitic acid at concentrations lower than 1 mg/mL exerted no significant effect on the growth of the Δagr mutant cells ( Figure 1A). However, it started to inhibit the cell growth when the concentration was over 1 mg/mL. Span85 also inhibited the growth of the Δagr mutant cells at concentrations over 0.1% (v/v) ( Figure 1B). Collectively, the MICs for each antibacterial compound are calculated (oxacillin = 256 μg/mL, palmitic acid = 2 mg/mL and span85 = 2% (v/v) ≈ 19.14 mg/mL)

Response Surface Methodology Analysis to Study the Effect of the Interaction of Different Antibacterial Agents
The advantage of combination therapies is the reduction in the antibiotic concentration used, as multiple activities can better attenuate or evade the antibiotic-resistance mechanisms of pathogenic bacteria. Response surface methodology analysis using the Box-Behnken design was introduced to set up the optimal concentration of three antibacterial agents to effectively eliminate MRSA [21][22][23]. Using concentrations higher than the MIC of each compound is meaningless; thus, the MIC50 of each agent was selected for the Box-Behnken design using Minitab 18 software to analyze the interaction and examine the desired response. The three significant variables, namely oxacillin, palmitic acid, and span85, were investigated with the values shown in Table 1 based on the diagonal sampling method [24,25].

Response Surface Methodology Analysis to Study the Effect of the Interaction of Different Antibacterial Agents
The advantage of combination therapies is the reduction in the antibiotic concentration used, as multiple activities can better attenuate or evade the antibiotic-resistance mechanisms of pathogenic bacteria. Response surface methodology analysis using the Box-Behnken design was introduced to set up the optimal concentration of three antibacterial agents to effectively eliminate MRSA [21][22][23]. Using concentrations higher than the MIC of each compound is meaningless; thus, the MIC 50 of each agent was selected for the Box-Behnken design using Minitab 18 software to analyze the interaction and examine the desired response. The three significant variables, namely oxacillin, palmitic acid, and span85, were investigated with the values shown in Table 1 based on the diagonal sampling method [24,25].
To monitor the effect of the oxacillin concentration, we selected three values with different concentrations of oxacillin. The range of the oxacillin concentration changed from 0 to 100 µg/mL. The experimental design and results are shown in Table 2. The regression equation obtained after analysis of variance gave the response (optical density, 595 nm) as a function of three significant variables. To Antibiotics 2020, 9, 682 4 of 11 obtain a polynomial equation, a quadratic model was conducted to fit the data by least squares, and all terms, regardless of their significance, were included in the following equation (1): where X 1 : oxacillin, X 2 : palmitic acid, and X 3 : span85. Oxacillin (X 1 , µg/mL) 0 50 100 Palmitic acid (X 2 , mg/mL) 0 0.75 1.5 Span85 (X 3 , %) 0 0.25 0.5 Table 2. Box-Behnken experimental design matrix with experimental values of the cell growth of the ∆agr strain. A surface plot for oxacillin, palmitic acid, and span85 is shown in Figure 2. Analysis of variance of the selected response showed a p value of <0.05, which indicated that the designed model was appropriate (Table 3). Surface plots showed that the concentrations of oxacillin and palmitic acid were important for the bactericidal effect ( 50 0.75 0.25 0.07 ± 0.00 A surface plot for oxacillin, palmitic acid, and span85 is shown in Figure 2. Analysis of variance of the selected response showed a p value of <0.05, which indicated that the designed model was appropriate (Table 3). Surface plots showed that the concentrations of oxacillin and palmitic acid were important for the bactericidal effect (

Effect of Combined Therapy on the ∆agr Strain and Clinically Isolated Strains
Once the concentrations of the three significant variables (oxacillin, palmitic acid, and span85) were set with the ∆agr mutant by response surface methodology analysis and response optimizer, we treated clinically isolated strains with the different combination therapies (Table 4). Low to high concentrations of oxacillin were considered. The focus was on the low concentrations, as antibiotic overuse has led to the emergence of multidrug resistance in MRSA. Thus, we examined whether the selected conditions were effective even for clinical isolates. Except for MRSA6230 and MRSA14459, the other MRSA strains were HA-MRSA, having SCCmec type II and type III, but all clinical strains were found to have an oxacillin MIC of over 128 µg/mL ( Figure 3A). The MIC for each strain is listed in Table 5. Multilocus sequence typing and spa typing, which were determined in the previous study, are also included in the table [26,27]. In addition, biofilm formation was compared between the different oxacillin concentrations ( Figure 3B). Depending on the degree of their antibiotic resistance, the clinical isolates were classified into high-resistance (MRSA8471 and MRSA9291), intermediate-resistance (MRSA2065, MRSA6288, MRSA7557, MRSA12779, MRSA14278, and MRSA14459), and sensitive (MRSA6230 and MRSA7875) groups.
typing, which were determined in the previous study, are also included in the table [26,27]. In addition, biofilm formation was compared between the different oxacillin concentrations ( Figure 3B). Depending on the degree of their antibiotic resistance, the clinical isolates were classified into highresistance (MRSA8471 and MRSA9291), intermediate-resistance (MRSA2065, MRSA6288 IV  20  t008  8  2065  MRSA  III  1024  t037  239  6230  MRSA  IV  128  t324  72  6288  MRSA  III  1024  t037  239  7557  MRSA  II  1024  t9353  5  7875  MRSA  IV  128  t664  72  8471  MRSA  II  1024  t9353  5  9291  MRSA  II  1024  t601  5  12779  MRSA  II  1024  t2460  5  14278  MRSA  II  1024  t9353  5  14459  MRSA  IV  1024  t324  72 As a control experiment, we found that all three conditions were sufficient to kill the ∆agr mutant, which mimics the characteristics of HA-MRSA with attenuated virulence. Additionally, the growth of MRSA2065, MRSA6230, MRSA6288, MRSA7875, and MRSA14459 was totally inhibited by all three combinations ( Figure 4A). All clinical SCCmec type III and IV strains have a much higher MIC of oxacillin of more than 128 µg/mL. This makes it difficult to kill MRSA with a low concentration of oxacillin. However, our combination therapy sets were able to kill five different strains, even with 15 µg/mL of oxacillin. Though the combination therapy was set up with the ∆agr mutant strain, the results showed that it was still enough to kill the clinical SCCmec type III and IV strains. However, the oxacillin concentration used was not enough to eliminate MRSA7557, MRSA8471, MRSA9291, MRSA12779, and MRSA14278, which are found to be SCCmec type II strains, which exhibit multidrug resistance with an oxacillin MIC level of about 1024 µg/mL. Killing assays were carried out once again after the confirmation of the oxacillin MIC; the results confirmed that even the SCCmec type II clinical isolates could be killed with 256 µg/mL of oxacillin ( Figure 4B). The clinical isolates contained a high ratio of odd anteiso-fatty acids in the membrane (data not shown), and the biofilm was thicker in the clinical isolates [28]. The individual treatment of the clinical strains with palmitic acid and span85 showed a similar pattern to the ∆agr mutant (Figure 1), but with different values. However, the combination of the three components decreased the amount of each component. even the SCCmec type II clinical isolates could be killed with 256 μg/mL of oxacillin ( Figure 4B). The clinical isolates contained a high ratio of odd anteiso-fatty acids in the membrane (data not shown), and the biofilm was thicker in the clinical isolates [28]. The individual treatment of the clinical strains with palmitic acid and span85 showed a similar pattern to the Δagr mutant (Figure 1), but with different values. However, the combination of the three components decreased the amount of each component.

Characterization of Clinically Isolated strains with Phospholipid Fatty Acid (PLFA) Analysis
MRSA strains with higher antibiotic resistance tend to have extraordinary features, such as compositional changes in membrane lipid, biofilm formation, persistent cell formation, stable membrane integrity for membrane microdomain assembly for optimal oligomerization of PBP2a, high mecA expression, and increased cell surface hydrophobicity [29,30]. To elucidate the reason for the different effects of our combinations, we performed phospholipid fatty acid (PLFA) analysis because fatty acid composition in the cytoplasmic membrane can affect the antibiotic resistance of pathogenic bacteria [9,31]. PLFA analysis showed that a major portion of phospholipids in the clinical MRSA strains contained abundant 12-methyltetradecanoic acid (anteiso-C15:0) and 14methylhexadecanoic acid (anteiso-C17:0) instead of hexadecanoic acid (C16:0) ( Figure 5). It is known that methyl branching modifies the thermotropic behavior and enhances the fluidity of lipid bilayers. It reduces lipid condensation, decreases the bilayer thickness, and lowers chain ordering and formation of kinks at the branching point [32]. Highly resistant SCCmec type II strains appeared to

Characterization of Clinically Isolated strains with Phospholipid Fatty Acid (PLFA) Analysis
MRSA strains with higher antibiotic resistance tend to have extraordinary features, such as compositional changes in membrane lipid, biofilm formation, persistent cell formation, stable membrane integrity for membrane microdomain assembly for optimal oligomerization of PBP2a, high mecA expression, and increased cell surface hydrophobicity [29,30]. To elucidate the reason for the different effects of our combinations, we performed phospholipid fatty acid (PLFA) analysis because fatty acid composition in the cytoplasmic membrane can affect the antibiotic resistance of pathogenic bacteria [9,31]. PLFA analysis showed that a major portion of phospholipids in the clinical MRSA Antibiotics 2020, 9, 682 8 of 11 strains contained abundant 12-methyltetradecanoic acid (anteiso-C15:0) and 14-methylhexadecanoic acid (anteiso-C17:0) instead of hexadecanoic acid (C16:0) ( Figure 5). It is known that methyl branching modifies the thermotropic behavior and enhances the fluidity of lipid bilayers. It reduces lipid condensation, decreases the bilayer thickness, and lowers chain ordering and formation of kinks at the branching point [32]. Highly resistant SCCmec type II strains appeared to show different PLFA compositions, except for MRSA7557, showing a lower amount of 12-methyltetradecanoic acid (anteiso-C15:0) than type III and IV strains and a relatively higher amount of 13-methyltetradecanoic acid (iso-C15:0) and hexadecanoic acid (C16:0), although it was difficult to link this to the increased resistance of SCCmec type II strains. Although SCCmec type II strains showed different results than those of type III and IV, these results in clinical strains showed the potential of combination therapy by decreasing oxacillin concentration with the same antibiotic activity.

Bacterial Strains, Media, and Culture Conditions
For cell preparation, the Δagr mutant [33] was cultured in tryptic soybean broth (TSB) agar and/or liquid broth. For pre-culture, a single colony of the strain from a TSB agar plate was used to inoculate 5 mL of TSB medium. Next, 1% (v/v) of the cell culture suspension was inoculated in a 96well plate for the antibiotic resistance test, and the cells were cultivated overnight in an incubator at 37 °C without shaking unless stated otherwise.

Antibacterial Agents
Oxacillin, palmitic acid, and span85 were purchased from Sigma-Aldrich (St. Louis, MO, USA). Stock solutions of these agents were prepared at various concentrations in sterile dimethyl sulfoxide.

Analysis of Cell Growth and Biofilm Formation
Cell growth was measured at 595 nm using a 96-well microplate reader (TECAN, Männedorf, Switzerland). Biofilm formation was analyzed by crystal violet staining according to a previously described protocol [9]. Briefly, the supernatant was aspirated. The biofilm was then fixed with methanol, air-dried, and stained with 200 μL of 0.2% crystal violet solution for 5 min. Next, the crystal violet solution was removed, and the biofilm was washed with distilled water and air-dried. Finally, the optical density of the biofilm was detected at 595 nm using a 96-well microplate reader.

Bacterial Strains, Media, and Culture Conditions
For cell preparation, the ∆agr mutant [33] was cultured in tryptic soybean broth (TSB) agar and/or liquid broth. For pre-culture, a single colony of the strain from a TSB agar plate was used to inoculate 5 mL of TSB medium. Next, 1% (v/v) of the cell culture suspension was inoculated in a 96-well plate for the antibiotic resistance test, and the cells were cultivated overnight in an incubator at 37 • C without shaking unless stated otherwise.

Antibacterial Agents
Oxacillin, palmitic acid, and span85 were purchased from Sigma-Aldrich (St. Louis, MO, USA). Stock solutions of these agents were prepared at various concentrations in sterile dimethyl sulfoxide.

Analysis of Cell Growth and Biofilm Formation
Cell growth was measured at 595 nm using a 96-well microplate reader (TECAN, Männedorf, Switzerland). Biofilm formation was analyzed by crystal violet staining according to a previously Antibiotics 2020, 9, 682 9 of 11 described protocol [9]. Briefly, the supernatant was aspirated. The biofilm was then fixed with methanol, air-dried, and stained with 200 µL of 0.2% crystal violet solution for 5 min. Next, the crystal violet solution was removed, and the biofilm was washed with distilled water and air-dried. Finally, the optical density of the biofilm was detected at 595 nm using a 96-well microplate reader.

Response Surface Methodology Analysis
After selecting the optimal concentrations for oxacillin, palmitic acid, and span85, combination therapies were optimized using Minitab software 18 (Minitab, State College, PA and SPSS, IBM Corp. 2011, Version 18, Armonk, NY, USA) through the Box-Behnken design and response surface methodology analysis. Experiments were conducted in triplicate, and the cell growth of MRSA was determined. Coefficients were determined using the experimental values using the full quadratic model f(x, y, z) = (x, y, z) = ax 2 + by 2 + cz 2 + dxy + eyz + fxz + gx + hy + iz + j, (a,b,c 0). Using surface plots, the relationships between the variables were investigated and validated.

PLFA Analysis
Briefly, 10 mL of the liquid culture was cultivated in TSB with 1% inoculum in an incubator at 37 • C with shaking at 200 rpm. Samples were collected at 8 and 16 h. Next, the samples were centrifuged at 3500 rpm for 20 min, and total fatty acids were extracted with 0.15 M citric acid buffer/chloroform/methanol (7:7.5:5, v/v/v) and incubated in an incubator at 37 • C with shaking at 200 rpm for 2 h. The chloroform phase was collected, and the chloroform was slowly evaporated under compressed N 2 to avoid oxidation. The sample was loaded into a sialic acid column and then serially eluted with 5 mL each of chloroform, acetone, and methanol. The methanol phase was collected for PLFA analysis. Next, 1 mL of toluene was added to the sample, which was subjected to mild alkaline trans-methylation with 1 mL of KOH/MeOH at 37 • C for 15 min, followed by cooling to room temperature. A 2 mL aliquot of 4:1 n-hexane/chloroform was added, and the sample was then neutralized with 1 mL of 1 M acetic acid. Subsequently, 2 mL of Milli Q water was added, and the phases were separated by centrifugation. The upper hexane layer was removed, and this step was repeated with fresh 2 mL aliquots of 4:1 n-hexane/chloroform. The combined hexane fractions were concentrated under compressed N 2 , and the fatty acids were re-solubilized with chloroform and analyzed.

Conclusions
In this study, we examined the effect of oxacillin combined with the fatty acid palmitic acid and the surfactant span85 on clinical strains, due to their antibacterial characteristics. To discover the optimal condition, we used the Box-Behnken design and response surface methodology analysis. We then proposed several conditions that were optimal to kill highly resistant clinical strains with very low concentrations of oxacillin. In addition, we showed that it is possible to kill more resistant strains, such as SCCmec type II strains, by increasing the oxacillin concentration.
To elucidate the reasons for the high resistance of clinical strains, PLFA analysis was conducted, and the results revealed different patterns of membrane fatty acid composition: more resistant strains contained a higher ratio of odd-chain fatty acids, such as 12-methyltetradecanoic acid (anteiso-C15:0) and 14-methylhexadecanoic acid (anteiso-C17:0). Although they may not be directly linked to the higher resistance of clinical samples and the effectiveness of a simple combination in killing all the strains, the different PLFA patterns appeared to be responsible for the higher resistance, based on the interaction between fatty acids and surfactants, which affected the membranes.
Our results showed that by combining oxacillin with palmitic acid and span85, the same level of antibacterial effects could be achieved with a lower concentration of oxacillin, thereby reducing the possibility of the strain acquiring drug resistance.
In conclusion, our data suggested a possible recycling strategy of safe antibiotics at present, in which their efficacy against resistant bacteria is increased via combined use with effective molecules.