Introduction

Application of Hazard Analysis of Critical Control Point (HACCP) is a common hygiene control practice in the Korean instant food industry; food-poisoning control depends on this system (1). However, for instant food storage refrigeration and the limitation of storage period are applied. Consequently, the addition of preservatives is restricted, and alternative methods are required to prevent food decay and development of foodborne pathogens. Because the materials added to prolong shelf life should not remain in the food, more studies on the use of volatile antibacterial materials for food preservation and the prevention of microorganism development are required (2).

Many of the known volatile compounds with antibacterial effect are found in herbs. For example, yeast contamination of food can be prevented by garlic, onion, oregano, thyme (3), tea tree (4), or summer savory (5); an antimold effect can be achieved by adding thyme, garlic, or onion (6-8); and thyme (9), horseradish (10), oregano, coriander, and basil have antibacterial effects (11). The bactericidal effect of allyl isothiocyanate on Listeria monocytogenes has also been investigated (12).

Callicarpa japonica Thunb. belongs to the Verbenaceae family and is indigenous to Korea, Japan, China, and Taiwan. The water extracts of its leaves have hemostatic and antibacterial functions, and the extracts of its leaves, stems, and roots are a traditional cure for extravasation, intestinal bleeding, uterine bleeding, lung infection, and tonsillitis (13) and are rich in 5,6,7-trimethoxyflavone, which has antiviral (14-16) and insecticidal (17) effects. Kim (18) reported that C. japonica complex found in East Asia shows diverse morphological variations. Kobaisy et al. (19) found that the volatile compounds of C. japonica leaves are spathulenol (18.1%), germacrene B (13.0%), and biocyclogermacrene (11.0%) and that these noticeably differ from the volatiles of Callicarpa americana.

However, few studies have been conducted on the antibacterial effects of the essential oil of C. japonica leaves. Therefore, we undertook the present study to extract the essential oil of C. japonica leaves under different conditions, to identify the volatiles present, and to conduct tests of their antibacterial activities, with a view toward increasing the shelf life of instant foods.

Materials and Methods

Microorganisms and Cultures. Six different foodborne bacterial species were used. Bacillus cereus (ATCC 11778) and Salmonella typhimurium (ATCC 14028) strains were grown at 30 C in nutrient broth or nutrient agar (Oxoid Ltd., Basingstoke, Hampshire, U.K.). Escherichia coli O157:H7 (ATCC 43894) and Staphylococcus aureus (ATCC 25923) were grown at 37 C and Listeria monocytogenes (ATCC 19111) was grown at 30 C in tryptic soy broth or tryptic soy agar (Difco, Detroit, MI). Vibrio parahaemolyticus (ATCC 33844) strain was grown at 37 C in tryptic soy broth or tryptic soy agar supplemented with 3% (w/v) NaCl. The bacteria were grown for 24 h in sterilized broth medium. A portion of each culture (0.1 mL) was transferred into new broth medium (9.9 mL) and grown for 18 h for the antibacterial experiments.

Extraction of Volatile Components. The leaves of C. japonica were collected from the Jeonju Arboretum (Jeonju, Korea) in September 2001 and washed and stored at -20 C. Extracts of the volatile compounds in the leaves were obtained by simultaneous steam distillation and solvent extraction (SDE) using the "improved" Likens-Nickerson unit (20). After 50 mL of the extracting solvent (redistilled diethyl ether) had been circulated through the apparatus at 36 C, 100 g of the leaves was ground in a Waring blender (Waring, New Hartford, CT) and added to 1000 mL of distilled water in a round-bottom flask. The mixture was heated at 100 C under different conditions, as follows. The extraction conditions used for C. japonica leaves were varied by using SDE times of 0.5, 1.0, 1.5, and 2.0 h at pH 6.0. The pH was then adjusted to 4.0, 5.0, 6.0 (control) and 7.0 and extracted for 1.5 h. Anhydrous sodium sulfate (~15 g) was added to remove water, and the ether mixture was cooled to -20 C for 12 h. The mix was then evaporated to 1 mL using nitrogen flow. Ten microliters of 1-pentanol (n-amyl alcohol) was then added to the extract as an internal gas chromatography (GC) standard. The concentrates obtained were tested for antibacterial activity and their volatile components analyzed.

Analysis and Identification of Volatile Constituents. GC (GC-17A V3) and GC-MS (QP5050, Shimadzu Co., Kyoto, Japan) were used for the analysis. Both used a Supelcowax 10 fused silica capillary column (60 m × 0.25 mm; 0.25 m film thickness). Helium was used as the carrier gas at a flow rate of 1 mL/min. The GC oven temperature was maintained at 50 C for 5 min, then increased to 230 C at a rate of 2 C/min, and then held for 10 min. The temperature of the injector was 250 C and that of the FID detector, 260 C. The GC split ratio used was 1:60, and 0.5 L of extract was injected for each run. The mass spectra ranged from m/e 28 to 400, and the ionizing voltage was 70 eV. Detected components were identified by comparing the spectra obtained with a mass spectrum library (Wiley NBS 139) and by comparing GC retention indices versus known standards.

Antibacterial Activities of SDE Extracts. C. japonica leaf extract (543.1 mg from 1 kg of fresh leaves) was obtained after completely evaporating the diethyl ether under nitrogen atmosphere. Ten percent (v/v) Tween 80 (Showa Chemical Co., Ltd., Tokyo, Japan) in water was then added up to a volume of 0.5 mL. The extract was then sterilized by passing it through a membrane filter (0.2 m) (21, 22). To determine its antibacterial activity, 5.82 mL of each medium was supplemented with 0.12 mL of the sterilized extract or with 10% Tween 80 as a control and then inoculated with 0.06 mL (10⁵-10⁶ cfu/mL) of each bacterial strain. The concentration of the dispersion was adjusted to an extract strength of 2, 4, 8, or 10% (v/v) with 10% Tween 80 to produce a total mixture volume of 6.00 mL. Aliquots of these cultures (0.3 mL) were dispensed into Bioscreen C (Labsystem, Helsinki, Finland) (an apparatus for automeasurement of optical density) wells and incubated as described for the respective bacterial strains. Optical density (600 nm) was measured every 12 h for 3 days against the Tween 80 control.

Antibacterial Effect of Extract Component. Thirty-two compounds identified in the volatile extract of C. japonica leaves by GC were tested for antibacterial activity. All compounds tested had a purity of >95% except for 2,4-heptadienal (90%). Each volatile compound (the content in C. japonica) was dissolved in 0.5 mL of 10% Tween 80 and then added to medium (2% v/v). Antibacterial activity was measured at 600 nm by using Bioscreen C every 24 h for 72 h. Activities were compared against 10% Tween 80 control.

Results and Discussion

Antibacterial Effect of Leaves Extracted by SDE for Different Times. The antibacterial activities of the C. japonica leaf extracts obtained by SDE for 0.5, 1.0, 1.5, and 2.0 h at pH 6.0 are shown in Table 1. The growth of B. cereus was slightly inhibited in the presence of the 0.5 h extract, whereas the antibacterial activities of the 1.5 and 2.0 h extracts were similar. Also, the growth inhibitory effects on S. typhimurium and other strains were lowest for the 0.5 h extract and increased with extraction time. Seo et al. (23) found that the initial water extract of hydrolyzed mustard leaves had a weak antibacterial effect, but after 12 h of incubation, the effect increased considerably and reached maximum activity after 24 h, after which it remained constant, which is in line with the results of the present study in which the antibacterial activity increased up to 1.5 h and remained constant thereafter. Therefore, in the latter experiments of the present study, an extraction time of 1.5 h was selected. Moreover, a prolonged heat treatment can break down the effective volatile components (24).

Changes in the Compositions of the Volatile Components with Extraction Time. It has been reported that the volatile components of extracts obtained by SDE may differ with the extraction time (25) and that this may affect the antibacterial activity (23). The volatile components of C. japonica leaf extracts are shown in Table 2. It was found that as the extraction time increased, the number of volatile components also increased; for example, 33, 47, 52, and 58 components were identified at SDE times of 0.5, 1.0, 1.5, and 2.0 h, respectively. Furthermore, the peak area of the total volatile components increased versus the internal standard as extraction time increased; that is, 79.3, 216.4, 543.1, and 706.7 mg/kg were recorded for extraction times of 0.5, 1.0, 1.5, and 2.0 h, respectively.

The main volatile components (78.85-80.48%) of C. japonica leaf extract were identified as -caryophyllene, 1-octen-3-ol, 2-hexenal, germacrene B, and aromadendrenepoxide II, accounting for 35.55-51.57, 9.70-19.55, 5.86-13.30, 5.21-8.71, and 4.01-6.42% of the total peak area, respectively. For an SDE time of 0.5 h, the peak area of the low molecular weight volatiles, that is, 2-hexenal (MW 98.15) and 1-octen-3-ol (MW 128.22), were 13.30 and 19.55%, respectively, but these reduced to 5.86 and 9.70% for an SDE time of 2.0 h. On the other hand, the peak areas of the higher molecular weight compounds, -caryophyllene (MW 204.36) and germacrene B (MW 204.36), at an SDE time of 0.5 h were 35.55 and 5.21%, respectively, but these increased to 51.47 and 8.17% after an SDE time of 2.0 h. Au-Yeung et al. (25) showed in an efficacy test using a Likens-Nickerson extractor that the extraction efficiencies of highly volatile compounds such as allyl isothiocyanate reduced with increasing extraction time, whereas those of nonvolatiles, such as methylpyrazine, increased.

Using the GC-MS, 58 volatile compounds were identified.

Effect of Extraction pH on Antibacterial Activity. The antibacterial activities of the extracts prepared at pH 4.0, 5.0, 6.0 (control), and 7.0 are presented in Table 3. Growth inhibitory effect on B. cereus and S. typhimurium tended to be lower at neutral pH, and at pH <6, the activity did not differ. The growth inhibitory effect on V. parahaemolyticus reduced with increasing pH, but activity against other bacterial species was found to be unaffected by extraction pH.

Effect of Extraction pH on Extract Volatile Composition. Investigators have shown that for the SDE method volatile components are affected by the salt content of the dispersion media (26) and pH (27, 28). In this study, the amount of volatile compounds was also affected by the medium pH (Table 4); for example, 53, 50, 52, and 52 compounds were identified at pH values of 4.0, 5.0, 6.0, and 7.0, respectively. The volatile content at pH 6.0 (543.1 mg/kg) was higher than that at pH 4.0 (291.9 mg/kg), 5.0 (311.6 mg/kg), or 7.0 (484.4 mg/kg).

The peak areas of -caryophyllene and germacrene B from C. japonica leaves were 38.66 and 2.95% of total GC peak area, respectively, at pH 4.0, and these increased to 41.31-48.57 and 5.24-5.76% at pH 5.0-7.0. However, the peak area of 2-hexenal was 9.15-9.87% at pH 4.0-6.0 but decreased to 6.99% at pH 7.0. No significant change in the peak area of 1-octen-3-ol (12.47-16.25%) occurred with pH. The above four compounds and aromadendrenepoxide II were found to be the main volatile compounds in the C. japonica leaf extract dispersion at all pH values and represented 67.23-79.00% of the total volatile compound content.

In a previous study, it was found that salt (NaCl) can enhance the steam pressure and increase the number of volatile compounds (26). Choi et al. (29) found that the volatile compound yield in Capsella bursa-pastiris extracts produced by SDE was maximal at pH 7.0 and that at pH 3.0 high levels of pentadecane, octanol, and indole were present. However, in general, minor volatile compounds such as hexanol were unaffected by pH. Nevertheless, this trend between volatile component yields and pH effect (29) concurs with the results of the present study.

Effect of SDE Extract Concentration on Antibacterial Activity. Different concentrations of 1.5 h SDE (pH 6.0) extracts (2, 4, 8, and 10% v/v) were tested for antibacterial activity (Figure 1).