Exposure to Anacardiaceae Volatile Oils and Their Constituents Induces Lipid Peroxidation within Food-Borne Bacteria Cells

The chemical composition of the volatile oils from five Anacardiaceae species and their activities against Gram positive and negative bacteria were assessed. The peroxidative damage within bacterial cell membranes was determined through the breakdown product malondialdehyde (MDA). The major constituents in Anacardium humile leaves oil were (E)-caryophyllene (31.0%) and α-pinene (22.0%), and in Anacardium occidentale oil they were (E)-caryophyllene (15.4%) and germacrene-D (11.5%). Volatile oil from Astronium fraxinifolium leaves were dominated by (E)-β-ocimene (44.1%) and α-terpinolene (15.2%), whilst the oil from Myracrodruon urundeuva contained an abundance of δ-3-carene (78.8%). However, Schinus terebinthifolius leaves oil collected in March and July presented different chemical compositions. The oils from all species, except the one from A. occidentale, exhibited varying levels of antibacterial activity against Staphylococcus aureus, Bacillus cereus and Escherichia coli. Oil extracted in July from S. terebinthifolius was more active against all bacterial strains than the corresponding oil extracted in March. The high antibacterial activity of the M. urundeuva oil could be ascribed to its high δ-3-carene content. The amounts of MDA generated within bacterial cells indicate that the volatile oils induce lipid peroxidation. The results suggest that one putative mechanism of antibacterial action of these volatile oils is pro-oxidant damage within bacterial cell membrane explaining in part their preservative properties.

resultant antibacterial activity of these oils against Gram-positive and Gram-negative bacteria. We also report on the oxidative effects of these oils on bacterial cells by quantifying the extent of lipid peroxidation.
The oils extracted from the leaves of M. urundeuva revealed the monoterpene δ-3-carene at 78.8% concentration. This monoterpene was also the major component in oils from plants collected in two other states in Brazil, Maranhão (78.1%) and Tocantins (56.3%) [25].
The chemical composition of S. terebinthifolius leaves oil in both sampling periods (March and July) confirmed the seasonal variation previously observed for this species [26]. The oil extracted in March presented a high concentration of myrcene (15.4%) and (E)-caryophyllene (14.7%) whilst in July, this components represented only 0.8% and 2.7%, respectively of the total oil. Germacrene-D content dropped from 21.0% in July to 8.8% in March, whereas the monoterpene α-phellandrene was undetectable in the oils collected in March, but rose to 18.2% in July. In addition, the oils extracted in July contained 15.5% of oxygenated sesquiterpenes compared with only 5.8% in the oils extracted in March.
The essential oils demonstrate a high level of variability in terms of yield and composition and this has been attributed to the interactions between factors such the geographic origin, edaphic and climate features, genetic variability and phenological phase of the plants [27][28][29].

Antibacterial Activity of the Essential Oils
Antibacterial activities of the essential oils were assessed by the agar disc diffusion and the microdilution methods against S. aureus, B. cereus and E. coli (Tables 2 and 3). The oils of all tested species, except the one from A. occidentale, exhibited varying levels of antibacterial activity against Gram-positive and Gram-negative bacteria. In general, the inhibition zones were higher for Gram-positive bacteria ( Table 2). Tassou and Nychas [6] also observed a greater effect on Gram-positive organisms for the oil from P. lentiscus (Anacardiaceae). Similar results were observed by Shimizu et al. [30] for the oils from fruits of Lithraea molleoides.
A relationship between the inhibition zones diameters (Table 2) and the MIC values (Table 3) could not be found. Due to the inherent characteristics of the method, the diffusion coefficients of the different constituents found in essential oils can influence the results. The capacity for each oil constituent to migrate within the agar can markedly influence the size of the inhibition zone. Although this method was standardized by CLSI, it was nevertheless developed for the analysis of conventional antimicrobial agents such as antibiotics. Such drugs are mostly hydrophilic in nature and consequently diffuse more easily in agar, as opposed to essential oils that are volatile, insoluble in water, viscous and possess a complex chemical composition. In addition, for the microdilution method, the oils were dissolved in the broth with the surfactant Tween 80, which may modulate their biological availability and, consequently, their antibacterial activity.  Extracts of S. terebinthifolius oils prepared in July were more active against all bacterial strains than the oils extracted from plants collected in March (Tables 2 and 3). The difference in antimicrobial activity is most probably due to the differing chemical composition (Table 1). Therefore, the seasonal variation influences the pharmacological properties of S. terebinthifolius oil and this should be carefully regulated when considering its application as an antimicrobial agent.
The region where the plants are grown may also influence the chemical composition of the oils produced. For instance, Erazo et al. [8] identified the major component of S. polygamus oil collected in Chile as β-pinene and this oil proved active against both Gram-positive and Gram-negative bacteria. However, for the oils extracted from plants collected in Argentina, the major constituents were limonene and α-phellandrene and this oil was active against Gram-positive B. cereus.
Oils extracted from S. terebinthifolius leaves collected in July showed elevated oxygenated sesquiterpene concentrations (15.5%) compared with the oils extracted in the month of March (5.8%). The relationship between the oxygenated compounds in the constituents with the higher antimicrobial activity of essential oils has been previously reported. For example, Ultee et al. [13] demonstrated that the presence of the phenolic hydroxyl group in carvacrol is essential for its activity against B. cereus. Therefore, the higher concentration of oxygenated components observed within oils extracted in July may be related to their greater antibacterial activity. Consequently, A. fraxinifolium oil, which also presented a higher concentration of oxygenated compounds, demonstrated high antibacterial activity.
However, oxygenated compounds were absent from M. urundeuva oil, but its antibacterial activity was comparable to that of the other oils investigated. The oil of this species was characterized by a high concentration of δ-3-carene (78.8%). Some studies do report the effect of this constituent on the tissues and cells of the respiratory system such as broncho-constriction and mucosal irritation [31][32][33], and toxicity to alveolar macrophages [34]. However, to our knowledge, studies on the antibacterial action of δ-3-carene remain unreported.
From the data presented in Table 3, it is apparent that δ-3-carene is more active itself than the total volatile oils extract. This compound is a hydrocarbon monoterpene ( Figure 1) and its hydrophobicity (calculated logp = 2.9) enables it to readily partition within the lipids of the bacterial cell membrane, thereby disrupting the structures and rendering them more permeable [1,13]. In conclusion, this compound appears to be a promising lead molecule for analogue synthesis, optimization of which may produce useful new compounds with antibiotic activity.

Lipid Peroxidation within Bacterial Cells
The results presented in Table 4 indicate that essential oils promote an increase in the malondialdehyde (MDA) content within bacterial cells, signaling the process of lipid peroxidation. The highest levels of peroxidation were observed for S. terebinthifolius (July) and M. urundeuva oils (MDA-TBA 2 = 2.14 to 3.34 nmol mg −1 protein) whereas the lowest effect was seen for the monoterpene δ-3-carene. In biological membranes, lipid peroxidation is normally associated with oxidative stress and free radical attack and, therefore, generates a complex variety of products, many of which are reactive electrophiles. For instance, malondialdehyde (MDA), a bifunctional tautomer of (E)-3-hydroxyacrylaldehyde, remains the most widely studied among such products [35]. It is formed by successive β-scission cascades of peroxidized polyunsaturated fatty acids and is commonly measured by reaction with thiobarbituric acid (TBA) to produce a colored chromogen. This highly reactive bis-aldehyde is toxic and interacts with DNA and proteins and it has potentially mutagenic properties [36].
Peroxidation damage of membrane lipids caused by synthetic antimicrobial agents has been described in the literature [37,38]. However, to the best of our knowledge, the effect of volatile oils on lipid peroxidation within bacterial membranes has not been previously reported. Considering the large number of different groups of chemical compounds present in essential oils, it is most likely that their antibacterial activity is not attributable to one specific mechanism and that there are several critical targets in the cell [1,11,39]. Among such targets, the cell membrane is particularly affected because the essential oil components can enter between the fatty acyl chains leading to membrane disruption [40]. This effect was observed to Bacillus subtilis treated with Fortunella crassifolia essential oil, the cell walls and membranes were partially disintegrated, causing the outflow of cytoplasm [41]. Consequently, our results indicate that the peroxidation of lipid is an important event contributing to the mechanism of antibacterial action of essential oils and deserves more widespread attention.

Plant Material
Aerial parts of A. humile Engl., A. occidentale L., A. fraxinifolium Schott ex Spreng., M. urundeuva Allemão, and Schinus terebinthifolius Raddi were collected in Minas Gerais state, Brazil. The materials were identified, herborized and voucher specimens were deposited in the VIC Herbarium of the Plant Biology Department, Federal University of Viçosa (Registration Numbers: 31,600, 36,629, 15,709, 20,815 and 30,839). Samples were collected in March 2008. The species S. terebinthifolius were collected in the summer (March) and in the winter (July).

Essential Oil Extraction
Leaves were collected in triplicate and in a completely randomized way amongst specimens of the populations investigated. Each sample (100 g) was freshly chopped and subjected to three h hydrodistillation in a Clevenger-type apparatus. The resulting oils were separated from the aqueous phase, weighed and the reported yields were calculated with respect to dry matter mass. Oils were stored under a nitrogen atmosphere and maintained at −4 °C, until they were analyzed by gas chromatography and mass spectrometry and before use within the bioassays reported herein.

Chemical Analysis of the Essential Oil Extraction-GC-FID and GC-MS
GC analyses were carried out with a GC-17A Series instrument (Shimadzu, Japan) equipped with a flame ionization detector (FID). Chromatographic conditions were as follows: fused silica capillary column (30 m × 0.22 mm i.d.) with a DB-5 bonded phase (0.25 μm film thickness); carrier gas, N 2 at a flow rate of 1.8 mL min −1 ; injector temperature 220 °C, detector temperature 240 °C; column temperature was programmed to start at 55 °C (isothermal for 2 min), with an increase of 3 °C min −1 up to 240 °C, isothermal at 240 °C for 15 min; injection of 1.0 μL (1% w/v in dichloromethane); split ratio 1:10; column pressure of 115 kPa. The analyses were carried out in triplicate and the abundance of each compound was expressed as a relative percentage of the total area of the chromatograms.
The GC-MS unit (model GCMS-QP5050A, from Shimadzu) was equipped with a DB-5 fused silica column (30 m × 0.22 mm i.d., film thickness 0.25 µm) and interfaced with an ion trap detector. Transfer line temperature, 240 °C; ion trap temperature, 220 °C; carrier gas, He at a flow rate of 1.8 mL min −1 ; injector temperature 220 °C, detector temperature 240 °C; column temperature was programmed to start at 55 °C (isothermal for 2 min), with an increase of 3 °C min −1 up to 240 °C, isothermal at 240 °C for 15 min; injection of 1.0 µL (1% w/v in dichloromethane); split ratio 1:10; column pressure of 100 kPa; ionization energy, 70 eV; scan range, 29-450 u; scan time, 1 s. The identity of each component was assigned by comparison of their retention indexes (RI), relative to a linear alkane standards series (C 8 -C 27 ) and also by comparison of its mass spectrum with either reference data from the equipment database (Wiley 7) and from literature sources [42].
The agar disc diffusion method was employed to determine the antimicrobial activity of the essential oils, as previously described [44]. Briefly, a suspension of the tested microorganism (2 × 10 8 CFU mL −1 ), previously activated twice at intervals of 24 h, was spread on Petri plates with Mueller Hinton agar. Filter paper discs (6 mm diameter) were individually impregnated with 5 µL of the essential oils and placed on the inoculated plates. The plates were incubated for 48 h at 37 °C in the cases of S. aureus and E. coli and at 32 °C for B. cereus. The diameters of the inhibition zones were measured using a paquimeter and expressed in millimeters. The antibiotic Chloramphenicol (30 μg) and sterile water were included in this experiment as positive and negative controls. Each test was performed in triplicate and repeated three times. The results were analyzed by ANOVA and Scott-Knott's multiple-range tests at p ≤ 0.05 by using the software GENES (Genetics and Statistical Analysis [45]. A broth microdilution method was used to determine the minimum inhibitory concentration (MIC) for the oil under test [46]. Overnight broth cultures of each strain were prepared in Brain Heart Infusion Broth (Himedia) and the final concentration in each well was adjusted to 2 × 10 5 CFU/mL following inoculation. The concentration of each inoculum was confirmed by viable count on Plate Count Agar (Himedia). The broth was supplemented with Tween 80 (Merck, Germany) at a concentration of 0.1% (v/v) in order to enhance essential oils solubility. Preliminary control experiments showed that Tween 80 alone did not affect the growth the microorganisms tested. A serial doubling dilution of each essential oil was prepared in a 96-well microtiter plate over the range 0.039 to 5.0 g L −1 . Positive growth control was prepared in the broth supplemented with 0.1% (v/v) Tween 80. The plates were incubated aerobically for 24 h at 37 °C for S. aureus, E. coli and lastly at 32 °C for B. cereus. The bacterial growth was monitored by turbidity by assaying in a UV spectrophotometer (625 nm) and the MIC was defined as the lowest concentration that reduced the bacterial growth.

Determinations of Lipid Peroxidation Levels
Tubes with 10 mL of Brain Heart Infusion Broth (Himedia), supplemented with 0.1% (v/v) Tween 80 (Merck, Germany), were inoculated with 2 × 10 5 CFU/mL of bacterial strains. These conditions were used for the control group. Within treatments, the broth was also supplemented with the MIC for each essential oil. After 24 h, the bacterial cells were collected by centrifugation at 10,000 g for 10 min. They were washed with 2 mL of 50 mM potassium phosphate buffer at pH 7.5 before being re-suspended in 2.0 mL of this buffer, and the lipid peroxidation was detected by assaying for malondialdehyde (MDA), which reacts with thiobarbituric acid (TBA) [37]. The MDA levels were expressed relative to the concentration of bacterial proteins. Proteins content was determined by the Bradford assay (1976) [47].

Conclusions
In conclusion, the essential oils from A. fraxinifolium, M. urundeuva and Schinus terebinthifolius leaves, collected in Minas Gerais, Brazil, have promising activity as an organic alternative to commonly used disinfectants and preservatives against both Gram-positive and Gram-negative food-borne bacteria. The volatile oil from M. urundeuva is also a potential novel source of the monoterpene δ-3-carene. Finally, pro-oxidant damages to the cell membrane appear to be associated with some of the essential oils described herein and may therefore play an important role and so far unrecognized role in the mechanism of antibacterial action of these natural compounds.