In vitro antagonistic inhibitory effects of palm seed crude oils and their main constituent, lauric acid, with oxacillin in Staphylococcus aureus

Infections caused by Staphylococcus aureus are a serious global threat, and with the emergence of antibiotic resistance, even more difficult to treat. One of the possible complications in antistaphylococcal therapy represents negative interactions of antibiotics with food. In this study, the in vitro interaction between oxacillin and crude palm seed oil from Astrocaryum vulgare, Cocos nucifera, and Elaeis guineensis against nine strains of S. aureus was determined using the checkerboard method. Lauric acid was identified as a major constituent of all tested oils by gas chromatography. The results showed strong concentration dependent antagonistic interactions between palm oils and oxacillin with values of fractional inhibitory concentrations indices ranging from 4.02 to 8.56 at concentrations equal or higher than 1024 µg/mL of the tested oils. Similarly, lauric acid in combination with oxacillin produced antagonistic action with fractional inhibitory concentration indices ranging from 4.01 to 4.28 at 1024 µg/mL. These findings suggest that interference between oxacillin and palm oils and their constituents can negatively affect the treatment of staphylococcal infections in humans and other animals.

www.nature.com/scientificreports/ relatively lower concentrations in C. nucifera and A. vulgare oil (11.87 and 6.55 mg/g, respectively). The oleic acid content was the highest in all samples, ranging from 6.47 to 16.54 mg/g. In contrast, eicosenoic (0.05-0.15 mg/g) and palmitoleic acid (≤ 0.03 mg/g) were present in minor amounts. PUFA content in C. nucifera oil (5.8 mg/g) was double that of A. vulgare (2.9 mg/g) and E. guineensis (2.84 mg/g) oil. In C. nucifera oil, linoleic acid was the most abundant PUFA, similar to A. vulgare and E. guineensis oil at 4.79, 2.77, and 2.68 mg/g, respectively. The second most abundant PUFA in C. nucifera oil was α-linoleic acid (1.88 mg/g), which was also present in lower amounts in A. vulgare (0.06 mg/g) and E. guineensis (0.03 mg/g) oil. Compared to C. nucifera and A. vulgare oil (both 0.13 mg/g), arachidonic acid was the least abundant PUFA in E. guineensis oil (0.07 mg/g).
Antistaphylococcal antagonistic effect of crude oils and oxacillin. In the first step, the susceptibility of the three tested S. aureus strains to oxacillin and hydrolyzed seed oils of A. vulgare, C. nucifera, and E. guineensis was determined using broth microdilution method to evaluate the suitable starting concentrations for combined effect (MIC-minimum inhibitory concentration-values for all tested oils in non-hydrolyzed forms > 8192 µg/mL; data not shown). A. vulgare oil induced the strongest antistaphylococcal effect, with MIC values ranging from 240 to 356 µg/mL, followed by that of C. nucifera (MIC values from 241 to 512 µg/mL) and E. guineensis (MIC values from 427 to 512 µg/mL). The MIC values of oxacillin ranged from 0.72 to 56.89 µg/mL. Further, analysis of the combined effect of tested seed crude oils and oxacillin was performed by the checkerboard method. The results showed some degree of antagonism in all tested strains at certain concentrations of combinations, with indices of fractional inhibitory concentration (FICI) ranging between 8.56 and 4.02. In addition, a consistently antagonistic, concentration-dependent effect of the selected palm oils at concentration of 2048 µg/mL, and in some cases 1024 µg/mL, was seen, when combined with oxacillin against all tested S. aureus strains. All other combinations showed an indifferent relationship between tested agents, as shown in Table 2.
A strong antagonistic effect, manifested by very high FIC indices of 8.56 (clinical isolate SA1) and 8.51 (MSSA ATCC 29213), was observed for A. vulgare oil at a concentration 2048 µg/mL, when combined with oxacillin. At the same concentration, A. vulgare oil caused an adverse interaction in MRSA ATCC 43300 with a FICI of 5.78. This antagonistic effect was also observed at a concentration of 1024 µg/mL with the MSSA clinical isolate, SA1 (FICI 4.30), and the reference strain, ATCC 29213 (FICI 4.27). Indifferent relationships were displayed in all other combinations for both agents. Interestingly, a significant increase in MIC of oxacillin was observed when combined with A. vulgare oil at two lowest concentrations tested (32 and 16 µg/mL). In these cases, the MIC values rose by approximately 1/3 and 3/4, respectively, when compared to the MIC of the antibiotic alone, against MSSA ATTC 29213. In combination with oxacillin, C. nucifera and E. guineensis oils also exerted antagonistic properties in all tested strains, at a concentration of 2048 µg/mL (FICI 4.02-6.28). C. nucifera oil showed the strongest antagonistic effect for MRSA ATCC 43300 strain, followed by MRSA clinical isolate SA1 and MSSA ATCC 29213 strain (FICI values were 6.28, 6.03, and 4.09, respectively). E. guineensis oil exerted similar antagonistic activity in MSSA ATCC 29213, MRSA clinical isolate SA1, and MRSA ATCC 43300 strain, with FICI values Antagonistic growth-inhibitory effect of lauric acid with oxacillin against S. aureus. As mentioned above, GC-FID analysis revealed LA to be the predominant fatty acid in all tested oils. Based on this result, the susceptibility of nine S. aureus strains to oxacillin, LA, and a combination of oxacillin and LA was investigated using the same methodology as that used for palm oils. As shown in Table 3   www.nature.com/scientificreports/ Subsequently, FICI values were calculated and the effect of combinations was analysed. Antagonistic mode of action was observed for LA at a concentration of 1024 µg/mL, when combined with oxacillin (FICI values were 4.01-4.28) in all tested S. aureus strains; the strongest antagonistic interaction was exhibited for the MRSA ATCC 43300 strain. Other combinations of concentrations revealed indifferent relationships, except for a combination of LA at 32 µg/mL with oxacillin against the S. aureus ATCC BAA 976 strain. In this case, the MIC of oxacillin increased four times, to 128 µg/mL, and the FICI value was 4.13. In all tested bacterial strains, rise of MIC values of tested antibiotics occurred when combined with LA at certain concentrations (16-256 µg/mL), up to two times with the average increase by nearly two-thirds.
To determine whether the hydrolyzed seed oils of A. vulgare, C. nucifera, and E. guineensis and LA influence the growth of S. aureus strains, the analysis of growth curves of three S. aureus strains, namely ATCC 29213, ATCC 43300, and clinical isolate SA1, was performed. According to Fig. 1, the graphical evaluation of bacterial www.nature.com/scientificreports/ growth upon increasing concentrations of hydrolyzed oils and LA revealed decreasing growth rate and increase in generation time for all tested S. aureus strains.

Discussion
In our study, crude seed oils rich in MCFAs, hydrolyzed by porcine pancreatic lipase, and LA showed in vitro growth-inhibitory effect against reference strains and clinical isolates of S. aureus. As per the results of GC analysis, the oils used for antibacterial testing were of standard composition, as reported by other groups, who also identified LA as the major constituent of A. vulgare 48 , C. nucifera 49 , and E. guineensis oils 50 . In this study, the MIC values of LA against the tested S. aureus strains (242-256 µg/mL) were comparable to those reported by other authors. For example, Batovska et al. 42 reported an MIC values for LA, measured by the macrodilution method against S. aureus strains at ≥ 125 µg/mL. Moreover, Nitbani et al. 51 reported the antibacterial activity of LA isolated from C. nucifera oil against S. aureus. In contrast, Parsons et al. 52 determined an MIC value of 50 µg/ mL (250 µM) for LA against S. aureus, using the broth microdilution method. Similarly, Kelsey et al. 53 observed MIC values equal to 50 µg/mL of LA against three different S. aureus strains, using turbidimetry with visual evaluation and ethanol as a solvent; however, these variations could be caused by the different methodologies and S. aureus strains used. To the best of our knowledge, there is only limited information on the antibacterial properties of palm oils rich in MCFAs. Free-fatty acids are known to have antibacterial activity in contrast with those bound to triglycerides 27 . As described by various authors previously 46,52 , the antistaphylococcal effect of fatty acids is induced by disruption of bacterial cell membrane resulting in its destabilization by inducing tubule formation on the lipid bilayer and cell lysis; thus, the antimicrobial activity of oils containing fatty acids is facilitated only upon hydrolysis of triglycerides [26][27][28] . In contrast, Ubgogu 43 showed that E. guineensis oil exerted slight antibacterial effect on S. aureus, without getting hydrolyzed, using the disc diffusion technique. This result is the opposite of that observed in this study, where the oils acted as an antibacterial only after being hydrolyzed by the porcine pancreatic lipase. Rossato et al. 54 also observed no antistaphylococcal activity of unhydrolyzed A. vulgare oil. Results of in vitro screening of C. nucifera oil and other MCFA-containing fats, with lipolytic enzymes that simulated gastric conditions in piglets, showed a significant change in suppression of gut microbiota (total anaerobes and E. coli) 55 . This finding suggests that MCFA-rich oils exert their antibacterial effects after enzymatic hydrolysis by lipases synthesized in the gastrointestinal tracts of humans and other animals. Studies on the combinatory effect of MCFAs and their esters with different organic acids, inorganic compounds, and antibiotics against various bacteria, including S. aureus, can be found in literature 26,42,47,56 . However, the findings of these studies markedly vary, depending on the class of antimicrobial agent used. For example, Kitahara et al. 47 observed synergistic interaction of LA at 50 µg/mL with gentamicin (FICI values were 0.25-0.31) and imipenem (0.13-0.25) and indifferent interactions of LA (50 µg/mL) with ampicillin and oxacillin, against MRSA clinical isolates. Similarly, Hess et al. 57 observed indifferent interactions of the LA-ampicillin or vancomycin combination against S. aureus biofilms and a synergistic interaction between LA and streptomycin. We observed indifferent antistaphylococcal action of oxacillin and LA (50 µg/mL); however, increasing the amount (≥ 1024 µg/mL) of LA, produced strong antagonism in the presence of the antibiotic. The reason for the lack of antagonism in Kitahara et al. 47 and Hess et al. 57 could be the concentration of LA used in the study, which probably did not reach a sufficient value for the antagonism to be exerted. There is no information on the interaction of palm oils rich in MCFAs as their antibacterial activity was not confirmed in the study 54,58 because hydrolysis of oils was not performed 26 .
From the results presented in this study, it is not possible to infer by which mechanism the antagonism between LA, tested oils and oxacillin occurs. Nevertheless, the detected prolongation of the generation time resulting in the decrease of specific growth rate arising with increasing concentrations of LA and LA-rich tested palm oils indicates that a possible mechanism underlying the antagonistic interactions between tested compounds might lie in the halting of cell division caused by LA/A. vulgare/C. nucifera/E. guineensis oil. It can be assumed that while the LA temporarily prevents the growth of bacterial cells, oxacillin is not able to properly exert its activity. However, it is not just the membrane lipid structure that can change under the influence of exogenous lipids but the protein structure as well. Other factors including membrane strain may account for the organization of membrane proteins 59 . It can be hypothesized that under increased membrane strain, such as after the treatment of lipid bilayers with high concentrations of LA 46 , membrane protein function is altered 60 . Such a circumstance can affect the antibacterial activity of oxacillin, which strongly depends on its ability to inhibit bacterial cell wall synthesis by preferentially binding to penicillin-binding proteins (PBPs) that are located inside the bacterial cell wall 61 . Therefore, oxacillin probably becomes ineffective after change in PBP function, induced by LA. As membrane protein stability also depends on membrane energetics, LA can reduce membrane fluidity and disrupt the electron transport system, perhaps by restricting the movement of carriers within the membrane 44 . The eventual impairment of membrane electron carriers can lead to a change in the intracellular and extracellular pH, which can cause the precipitation of PBPs 62 and make them to lose the ability to interact with oxacillin. In addition, the change in extracellular pH can affect chemical structure of oxacillin as this drug is highly unstable in acidic environments 63 . Moreover, S. aureus is known to produce persisters, which are representing a fraction of the bacterial population that exhibits tolerance to antibiotics in response to various stresses 64 . According to the finding of Peyrusson et al. 65 , S. aureus in the presence of high concentration of various antibiotics including oxacillin showed a biphasic killing manner, meaning that a bulk of the bacterial population was susceptible and rapidly killed while a subpopulation with a slower killing rate was persisting for a much longer period of time, in addition showing the reversibility of the phenotype after antibiotic removal. Therefore, another possible explanation of antagonism between oxacillin and LA/LA-rich palm oils, can be in inducing persister cells of S. aureus in the presence of high concentrations of tested compounds. www.nature.com/scientificreports/ Food-drug interactions are a major threat to safe and effective oral pharmacotherapy and can result in decreased bioavailability of a drug, which predisposes the patient to treatment failure, increases the risk of adverse events, and may even precipitate toxicities 19,66 . For this reason, coadministration of a drug with specific foods is noted in medical leaflets. Generally, food intake can influence the effectiveness of an antibiotic 67 . Ingestion of food, dietary fiber, or milk reduces the bioavailability of most antibiotics, including some penicillins 68 . For example, minerals in milk and cheese create complexes with antibiotics that decrease their absorption 69 , and as seen in the case of isoxazolyl penicillins, when administered shortly before or after a meal, delayed gastric emptying and increased acidity interfere with their absorption 70 . The consumption of coconut oil and related products is currently growing among certain populations, for the claimed health benefits associated with cardiovascular disease and weight loss 71 . On the other hand, recommendations of lowering intake of saturated fatty acids and replacing them with unsaturated fatty acids exist in order to reduce risk of atherosclerosis and type-2 diabetes 72,73 . The average concentration of LA in human serum of healthy adult male and female blood donors, ages ranging from 18 to 55 years, was found to be < 10 µg/mL 74 . It has, however, been proven that higher (14.2-140 g/day) intakes of MCFAs in diet may result in higher concentrations of high density lipoprotein cholesterol than found with long-chain fatty acids, highlighting the importance of considering chain length when measuring the effect of dietary saturated fatty acids on lipid profile 75 . Moreover, in a high-carbohydrate, high-fat diet, the increases in systolic blood pressure and diastolic stiffness in the heart were inhibited in mice with diet enriched by virgin coconut oil, composed predominantly of LA, at concentrations of 200 g/kg 76 . Our results showing strong in vitro antagonistic effect of oxacillin with LA or LA-rich palm oils at concentrations ≥ 1024 µg/mL (meaning approximately ≥ 1.113 g/kg; 1 g/kg of body mass) suggest that simultaneous administration of these agents can negatively affect their pharmacological properties. The recommended intake of fats in the diet of men is estimated to be around 65 g/day (daily energy intake 2000 kcal with fats representing 30% of it) 77 . According to FAO 78 , average energy requirement of an adult female is 2410 kcal/day and of an adult men is 3100 kcal/day. Counting daily intake of fats as 30% of total energy requirement, the daily intake of fats can be estimated to be 80 g/day for women and 103 g/day for men. Studies on coconut oil supplementation in diet usually focus on the addition of the oil in range 20-50 g/day [79][80][81] . Nevertheless, there have been reports on even higher daily consumption of coconut oil reaching up to 80 g/day [82][83][84] . Therefore, the observed antagonistic action between LA-rich oils/ LA may be important especially in high-fat diets. Thus, our results showing strong in vitro antagonistic effect of oxacillin with LA or LA-rich palm oils suggest that simultaneous administration of these agents in high, but still reachable concentrations can negatively affect their pharmacological properties in the treatment of S. aureus. The risk of antagonistic interactions between oxacillin and LA-rich oils might be primarily important to systemic application, as their antibacterial effect is attributed to fatty acids unleashed from triglycerides only, therefore the topical application of LA-rich oils such as A. vulgare/C. nucifera/E. guineensis should not influence the antibacterial activity of oxacillin. But, according to Verallo-Rowell et al. 85 , 5 mL of extra virgin coconut oil applied two times a day on the affected areas that include the test sites is able to decolonise skin from S. aureus in adults with atopic dermatitis. This discrepancies between the theoretical background and practice can be debited to the lipolytic activity of skin microbiota, including staphylococci 86 , highlighting the importance of possible negative effect of LA-rich oils on topical treatment of S. aureus. Moreover, LA was previously tested in vitro at concentration of 0.24-500 μg/mL to evaluate its antibacterial properties against various bacteria causing inflammatory acne vulgaris, including S. aureus, proposing it as an promising remedy in the treatment of staphylococcal skin infections 87 . However, the mentioned tested concentrations of LA were not high enough for the antagonism with oxacillin to be exerted. Nevertheless, these hypotheses must be confirmed by further in vitro and in vivo tests and clinical trials because physiological processes can also induce changes in antibacterial activity of tested compounds.
In summary, this in vitro study revealed a concentration-dependent antagonistic effect between A. vulgare, C. nucifera, and E. guineensis oils when combined with oxacillin in higher amounts against various strains of S. aureus. The strongest antagonism was observed for A. vulgare oil, which contains the highest amount of LA. This compound was identified as the main agent responsible for antagonistic antistaphylococcal action of all oils assayed. To the best of our knowledge, this is the first study to report the antagonistic interactions between these agents. The mechanism underlying the antagonistic action of tested agents probably acts at the cellular level and is linked to the cell membranes. These findings suggest that interference between oxacillin and palm oils and their constituents can negatively affect the treatment of staphylococcal infections in humans and animals. However, these assumptions are based on in vitro tests and the negative interactions of the above-mentioned combinations should be confirmed by in vivo trials.

Methods
Chemicals and samples preparation. LA, C. nucifera and E. guineensis oils, and oxacillin sodium monohydrate were obtained from Sigma-Aldrich (Prague, CZ). A. vulgare oil was purchased from Sweet Natural Botanicals (Panama City, FL, USA). LA and A. vulgare, C. nucifera, and E. guineensis oils were dispersed in dimethyl sulfoxide, which was previously proposed for sparingly soluble substances 88 , including the LA 89 and palm oil 90 ; and emulsified using Tween 80 (Sigma-Aldrich, Prague, CZ), that is a common procedure of hydrophobic sample preparation 91 , not influencing the antibacterial properties of tested compounds when used in recommended concentrations 92,93 . Complete dissolution of LA was achieved by heating (70 °C for 10 min) it in an ultrasonic-bath. Analysed plant oils were selected based on the high content of LA as described in literature [94][95][96] . Hydrolysis of oils, to induce antimicrobial properties, was achieved by adding 5 × 10 -3 mg/mL of porcine pancreatic lipase (Sigma-Aldrich, Prague, CZ) and shaking for 1 h in a water-bath heated to 37 °C. Degree of hydrolysis of oils was dependent on the lipolytic activity of the enzyme. One unit of porcine pancreatic lipase hydrolyzes 1.0 microequivalent of fatty acid from triacetin in 1 h at pH 7.4, at 37 °C.  102 , for effectively assessing the anti-infective potential of the natural products. Antistaphylococcal effect of a combination of oxacillin/LA or oxacillin/palm oil was tested in vitro using the microdilution broth checkerboard method, as described in the Clinical Microbiology Procedures Handbook 103 . The determination of MIC of oxacillin, palm oils and LA, as well as oxacillin/LA or oxacillin/A. vulgare oil/C. nucifera oil/E. guineensis oil combinatory effect evaluation by FICI was performed in 96-well microtiter plates. For the testing of combinatory effects, eight twofold serial dilutions of oxacillin were placed in the horizontal rows of the plate and were subsequently cross-diluted vertically by eight twofold serial dilutions of the test compound (palm oil or LA), resulting in 64 different combinations of concentrations. The initial concentration for palm oils was 4096 µg/mL and for LA 2048 µg/mL; the starting concentration of antibiotic was adjusted according to the tested strain. The microtiter plate assay was performed using the automatic pipetting platform, Freedom EVO 100 equipped with a four-channel liquid handling arm (Tecan, Männedorf, CH). Cation-adjusted Müller-Hinton broth, equilibrated to a final pH of 7.6 with Trizma base (Sigma-Aldrich, Prague, CZ) was used as growth medium. Buffering the culture media was performed to ensure the stability of oxacillin that is known to decrease under the low pH conditions 63 . Inoculation of the plates was carried out using bacterial suspensions, at a final density of 5 × 10 5 CFU/mL, standardized using Densi-La-Meter II by adjusting turbidity of the microorganism suspension to 0.5 McFarland standard. Next, incubation at 37 °C for 24 h was performed. Evaluation of bacterial growth was performed spectrophotometrically using multimode reader Cytation 3 (BioTek Instruments, Winooski, VT) at 405 nm. MIC values were expressed as the lowest compound concentrations that resulted in ≥ 80% growth reduction compared to that of the agent-free growth control. The lipase added to the Müller-Hinton broth, in concentrations ranging from 0.005 to 9.77 × 10 -6 mg/mL did not affect the growth of any strain of S. aureus tested when assayed as a negative control. . According to the value of FIC, three different effects could be observed according to Odds 104 -synergy (ΣFIC ≤ 0.5), indifference (ΣFIC > 0. , and antagonism (ΣFIC > 4). All compounds and their combinations were tested in three independent experiments, each carried out in triplicate; MIC values and FICs presented in this paper are average values.

Growth rates determination.
To evaluate the influence of tested palm oils and LA concentrations on parameters of the S. aureus growth, standardised microdilution assay was used 105 . Briefly, the determination in 96-well microtiter plates where eight two-fold serial dilutions of tested compound (LA or A. vulgare/C. nucifera/E. guineensis oil emulsion cleaved by porcine pancreatic lipase as described previously) starting at concentration 4096 µg/mL in cation-adjusted Müller-Hinton broth was performed. Next, the plates were inoculated with bacterial suspensions at a final density of 5 × 10 5 CFU/mL, standardized using Densi-La-Meter II by adjusting turbidity of the microorganism suspension to 0.5 McFarland standard, same as in case of MIC and FIC determination. Subsequently, incubation at 37 °C was performed, measuring the absorbance of each well

Data availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.