1325 Journal o f Food Protection, Vol. 77, No. 8, 2014, Pages 1325-1335 doi:10.4315/0362-028X.JFP-13-393 C o p yrig h t © , Internatio nal A sso cia tio n fo r Food Protection

Antimicrobial Efficacy of Emulsified Essential Oil Components against Weak Acid-Adapted Spoilage Yeasts in Clear and Cloudy Apple Juice MYRIAM LOEFFLER, SOPHIA BEISER, SARISA SURIYARAK, MONIKA GIBIS,

a n d

JOCHEN WEISS*

Department o f Food Physics and Meal Science, Institute o f Food Science and Biotechnology, University o f Hohenheim, Garbenstrasse 21125, 70599 Stuttgart, Germany MS 13-393: Received 22 September 2013/Accepted 14 March 2014

ABSTRACT The antimicrobial activity of oil-in-water emulsions containing dual combinations of the essential oil components cinnamaldehyde, periUaldehyde, and citral was examined against two acid-resistant yeast strains (Zygosaccharomyces bailii) in beverage systems composed of diluted clear or cloudy apple juice and in a Sabouraud dextrose broth model. Antimicrobial properties of an encapsulated oil-in-water emulsion and of essential oil components dissolved in 10% dimethyl sulfoxide were compared using plate counts and turbidity measurements. Growth curves were modulated to qualitatively assess differences in antimicrobial efficacy. The impact of the presence of a beverage emulsion without essential oils (unloaded; 5% oil and 1% modified starch, pH 3.0) on the antimicrobial efficacy also was investigated. Dual combinations of essential oil components were sufficient to completely inhibit and/or kill yeast cells in diluted apple juice and Sabouraud dextrose broth systems at very low concentrations (100 to 200 pg/ml). However, the combination of perillaldehyde and citral had the weakest antimicrobial effect; a concentration of 400 pg/ml was necessary to prevent yeast growth in beverages, and up to 800 pg/ml was required in-systems to which an unloaded emulsion had been added. The antimicrobial activity of essential oil components did not differ in diluted clear and cloudy apple juices and was not affected by being added in emulsified form or dissolved in dimethyl sulfoxide as long as there was no unloaded emulsion also present. These results indicate that formulations of essential oil combinations encapsulated together in emulsions are highly effective for inhibiting and/or killing microorganisms in real beverage systems.

The susceptibility of a particular food product to spoilage depends on a variety of factors, including intrinsic parameters and manufacturing conditions (11). Because sweet nonalcoholic soft drinks contain ingredients such as golden syrup, organic fruit acids, and essential oils (e.g. citral), these drinks are intrinsically resistant to many food spoilage organisms. Low pH and a high concentration of solubilized C 0 2 in carbonated beverages contribute to that resistance. Typical spoilage organisms of soft drinks that can tolerate these conditions are lactic acid bacteria and especially yeasts in the Davenport group, such as Zygo­ saccharomyces bailii (2). These yeasts are hard to control because they can develop strong resistance to weak acidic preservatives, more precisely to benzoic and sorbic acids, possibly due to enzymatic systems that can degrade or convert the organic acids (25). Because of this resistance and consumer demands for milder preservation techniques, essential oils, which are secondary plant metabolites, might be an excellent alternative to traditional preservatives. Essential oils can be isolated by steam distillation from herbs and spices and are commonly applied to food products to add a particular flavor, thereby enhancing product * Author for correspondence. Tel: + 4 9 711 459 24415; Fax: + 4 9 711 459 24446; E-mail: [email protected].

quality. Essential oils also can prolong shelf life and improve food safety because of the presence of highly active antimicrobial compounds, such as phenols (e.g., eugenol and thymol) and/or aldehydes (e.g., cinnamaldehyde and perillaldehyde). For example, vanillin inhibits the growth of spoilage yeasts such as Z. bailii (8). The mechanism of the antimicrobial action of these oils is associated with the functional groups. The cell membranes of microorganisms have been presumed to be the main target of essential oil components (22). Increased permeability or disruption of these membranes leads to a loss of cell contents, and the reduction of the membrane potential causes a collapse of the proton motive force, followed by a depletion of the ATP pool (27). Combinations of essential oil components may have synergistic effects if they simultaneously attack different targets of a microbial cell. Zhou et al. (30), who evaluated how cinnamaldehyde, thymol, and carvacrol influenced the growth of Salmonella Typhimurium, found that the combination of these essential oil components was much more effective for inhibiting bacterial growth than were the single components. However, some natural antimicrobials are only effective as preservatives in foods at concentrations exceeding the acceptable flavor thresholds (7). This adverse effect on flavor may be attributed to food ingredient interactions or nonhomogenous distribution of

1326

LOEFFLER ET AL.

essential oils in the food matrices (29). Therefore, higher concentrations of antimicrobial compounds are needed to ensure food safety. Essential oils cannot simply be added to aqueous systems because of their hydrophobicity. Even at low concentrations, these oils separate rapidly from any aqueous phase, which causes antimicrobials to be distributed nonhomogenously throughout the product. Consequently, the frequency of collision of antimicrobials with target microorganisms is substantially decreased. Encapsulation of essential oils may reduce interactions with interfering food ingredients such as proteins and ensure homogenous distribution throughout the product. Oil-in-water emulsions could function as carriers, thereby preventing a decrease in the antimicrobial activity of the essential oils. In addition to macroemulsions, nanoemulsions are potential carrier sys­ tems because of their high kinetic stability, which is due to their small droplet size (diameter < 1 0 0 nm). McClements et al. (16) suggested that the bioavailability of encapsulated hydrophobic compounds is lower in conventional macro­ emulsions than in nanoemulsions, which was explained by the higher surface-to-volume ratio of smaller droplets (1,19, 28). Initial studies carried out in our laboratories revealed that oil-in-water emulsions loaded with cinnamaldehyde, perillaldehyde, and citral had high antimicrobial activity against spoilage yeasts (Candida albicans, Zygosaccharomyces bisporus, and Z. bailii) in beverage model systems. The antimicrobial efficacy of loaded emulsions could be increased by combining these essential oil components, which would result in synergistic interactions. We hypothesized that the interaction of essential oil components with apple juice ingredients and the presence of an unloaded emulsion would reduce the antimicrobial activity of these essential oils. The main focus of this study was to determine the influence of a real beverage system composed of diluted clear or cloudy apple juice on the antimicrobial activity of essential oil components. To ensure the applicability of this work, e.g., in soft drinks, the impact of an unloaded emulsion on antimicrobial activity was investigated. The activity in beverages of encapsulated antimicrobials was compared with the activity of antimicrobials solubilized in dimethyl sulfoxide (DMSO) to assess the effect of encapsulation. The antimicrobial effect of essential oils on yeast growth in diluted beverage systems also was quantitatively and qualitatively analyzed using a growth model. M A TER IA LS A N D M E TH O D S Materials. Essential oil components, cinnamaldehyde (>93% ), citral (>95% ), perillaldehyde (>92% ), nonionic Tween 20 (polyoxyethylene (20)-sorbitan-monolaureate), and DMSO (>99.9% ) were obtained from Sigma Aldrich (Steinheim, Germany). Sabouraud dextrose broth (SDB), Sabouraud dextrose 4% agar (SDA), and peptone water were obtained from Carl Roth GmbH (Karlsruhe, Germany). Miglyol 812 (caprylic-capric triglyceride) oil was supplied by Sasol Germany GmbH (Hamburg, Germany). Calcium chloride dihydrate, magnesium chloride hexahydrate, disodium DL-malate, and DL-malic acid were purchased from Merck (Darmstadt, Germany). An unloaded

J. Food Prot., Vol. 77, No. 8

emulsion (5% oil and 1% modified starch, pH 3.0) was obtained from PepsiCo (Valhalla, NY), and two strains of Z. bailii (ATCC 60484 and 906 Pepsi) were obtained from the Pepsico R&D Culture Collection. The cultures were stored at —75°C in SDB with 20% glycerol. Apples (Jonagold variety from Italy), the enzymes pectinase (Pectinex ultra SP-L, Novo Nordisk, Mainz, Germany) and amylase (BAN 480 L, Novozymes, Bagsvaerd, Denmark), bentonite (Puranit, Begerow, Langenlonsheim, Ger­ many), silica sol (Bevasil 30, Begerow), and gelatine (finely granulated; Begerow) were donated by the Department of Plant Foodstuff Technology (University of Hohenheim, Stuttgart, Germany). Distilled water was used for the preparation of all solutions, and tap water was used for the apple juice manufacturing process. Preparation of malic buffer solution. Buffer was prepared with distilled water of hardness ca. 40 mg/ml, according to instructions from Pepsico. Water hardness was achieved by mixing distilled water with 1 ml/liter dissolved magnesium chloride hexahydrate (stock solution of 3.9 g/100 ml of distilled water) and calcium chloride dihydrate (stock solution of 2.7 g/100 ml of distilled water). The water was titrated to a pH of 3.0 with 1 M malic acid and then titrated to a final pH of 3.4 with 0.5 M sodium malate. Application of essential oil components. The combinations cinnamaldehyde-citral (ratio 3:1), cinnamaldehyde-perillaldehyde (ratio 1:1), and perillaldehyde-citral (ratio 3:1) were also directly applied to the microbiological test systems. Essential oil com­ ponent combinations at 47.5 mg/ml were dissolved in an aqueous solution containing malic buffer (pH 3.4) and 10% DMSO. Preparation and characterization of emulsions. Three oilin-water emulsions (5% oil phase with Miglyol 812 and essential oil components, 1% [wt/wt] Tween 20, and malic buffer; pH 3.4) were prepared with different combinations of essential oil components: cinnamaldehyde-citral (ratio 3:1), cinnamaldehydeperillaldehyde (ratio 1:1), and perillaldehyde-citral (ratio 3:1). A lipid blend was prepared by mixing 5% Miglyol with 95% essential oil components (equal to 47.5 mg/ml). For example, the oil phase containing a 1:1 mixture of essential oil components hence contained a mixture of 2.375 g of component 1, 2.375 g of component 2, and 0.25 g of Miglyol. A premix of 95% surfactant solution and 5% lipid blend was prepared by mixing for 2 min at maximum speed (24,000 rpm) with a high shear blender (IKA Werke GmbH, Staufen, Germany). The premix was then passed through a microfluidizer (MilOP, Microfluidics, Westwood, MA) three times at a homogenization pressure of 10,000 lb/in2 to form fine emulsions. The emulsions were adjusted to pH 3.4 by addition of HC1 or NaOH. Droplet size was measured using a dynamic light scattering technique (Zetasizer Nano ZS, Malvern Instruments, Malvern, UK). Emulsion samples were diluted to 1:100 with malic buffer solution (pH 3.4) to avoid multiple scattering effects. Emulsion stability was investigated by measuring the droplet diameter repeatedly during 2 weeks of storage at room temperature. Formation of cream layers indicates destabilization of emulsions, but no cream layers were observed in any of the test tubes. All measurements were repeated three times on two different samples. Apple juice manufacturing. Clear and cloudy apple juice was produced at the pilot plant of the University of Hohenheim. Apples were washed with tap water to remove any adhering

J. Food Prot., Vol. 77, No. 8

REPLACEMENT OF TRADITIONAL PRESERVATIVES IN FRUIT JUICES

1327

substances. After milling, the apples were pressed with a connected pack press (TPZ 7, Buchner Unipektin AG, Niederweningen, Switzerland). To obtain clear apple juice, the juice was treated with pectinases and amylases (each 25 ml/500 liters), which hydrolyzed pectin and amylose. The fining agents bentonite (400 g/500 liters), silica sol (500 ml/500 liters), and gelatine (120 g/500 liters) were added to remove proteins and polyphenols. The apple juice with the fining agents was left overnight to precipitate the suspended solids. The next day, particles were removed by decanting, and the juice was pasteurized at 89°C and poured into sterilized bottles. To obtain cloudy apple juice, the juice was centrifuged (Separator, Westfalia, Oelde, Germany) to remove the coarse pulp particles, pasteurized at 89°C, and poured into sterilized bottles. Both types of apple juice were stored at 4°C until used.

Antimicrobial activity assays. The clear apple juice and essential oil components emulsified or dissolved in DMSO were filter sterilized using 0.20-pm-pore-size polyethersulfone mem­ brane filters (Pradusc 25, Whatman Inc., Kent, UK) before all antimicrobial assays. Refreshed yeast cultures on SDA were incubated in a shaking incubator (Thermo Scientific Heratherm, Karlsruhe, Germany) for 48 h in SDB (25°C, 150 rpm) before use. MICs in model medium were determined with a microtiter assay performed on 96-well plates with 300 pi per well (Falcon, BD Biosciences GmbH, Heidelberg, Germany). To reach the desired antimicrobial concentrations in the wells, 120 pi of an emulsion loaded with essential oil components (stock emulsions of essential oil components at 47.5 mg/ml) had to be diluted with buffer (pH 3.4) and was then dispensed into the 96-well microtiter plates, and then 120 pi of inoculated double-concentrated yeast suspension (Z. bailii ATCC 60484 or Z. bailii 906 Pepsi, approximately 1 x 105 CFU/ml) was added. Plates were covered with breathable seals to prevent evaporation of essential oil components and incubated with shaking at 150 rpm and 25°C for 7 days. The optical density (OD) was measured at 630 nm every 24 h with a synergy HT Multi-Mode Microplate Reader (BioTek Instruments, Bad Friedrichshall, Germany). The MIC was defined as the concentration sufficient to prevent relative increases in OD (AOD = OD, — OD,=0) over a 7-day period, where OD, is the OD measured each 24 h and OD,=0 is the OD at time 0. The antimicrobial efficacy of different combinations of essential oil components applied as pure compounds, as dilutions in DMSO, or as emulsions was investigated against Z. bailii ATCC 60484 and Z. bailii 906 Pepsi in the macrobroth assay. Activity was tested in different media (SDB and diluted clear and cloudy apple juice) and in the absence or presence of an emulsion without essential ods (i.e., unloaded) to simulate the presence of a flavor emulsion in a beverage system. The pH of all systems was adjusted to pH 3.4. Various 1.2-ml tubes (Fisher Scientific GmbH, Schwerte, Germany) were filled with 2% medium, buffer (pH 3.4) solution, or essential oil components (diluted in 10% DMSO or encapsulated in emulsions) in various concentrations (0 to 800 pg/ml depending on the essential oil components used) and, when required, 0.15% (wt/wt) unloaded emulsion, so that a headspace of 10% remained. The blends were inoculated with Z. bailii, which had been previously diluted in peptone water to 102 CFU/ml. Samples were incubated at 25°C, and yeast growth was determined by plate counts on SDA on days 0, 1, 3, 5, 7, 12, 15, 20, and 27. Except for investigations of cloudy apple juice (which plugged the spiral plater), cell levels higher than 102 CFU/ml were evaluated by spiral plating (automatic spiral plater, Don Whitley Scientific, Shipley, UK). Undiluted samples were spread plated on SDA plates that were then incubated for 48 h at 25°C. Colonies

Time (day)

10910 h v ^ ] = A'V">'mL + e r p [ K tc - t ] “ l + exp[k t j l

FIGURE 1. Logistic function as a model o f microbial growth (macrobroth assay; plate counts). d N asym is the asymptote o f the growth curve, k is the growth rate, k is the position o f the inflection point, No is the initial number o f microorganisms, and N is the number o f microorganisms at time t (CFU per milliliter). were counted with an automated plate counter (Acolyte, Synbiosis, Cambridge, UK).

Modeling of yeast growth. A model was needed to qualitatively and quantitatively assess the differences in the effect of antimicrobials in their emulsified and dissolved form, in the presence or absence of an unloaded emulsion, and in clear and cloudy juice and broth. Experimental data on inhibition of yeast growth in the presence of a sublethal concentration of antimicro­ bials were obtained with the macrobroth assays, and the cell levels derived from plate counts were then fitted using the model described by Dai et al. (6). Growth curves determined from plate counts (detection limit of viable cells, 102 CFU/ml) were fitted with the following equation: log

'm

_N0

-AN,asym

1 + exp [k(tc —r)]

1 + exp [ktc

( 1)

where AN^ym is the asymptote of the growth curve, k is the growth rate, and tc is the position of the inflection point. The model can be used for sigmoid growth behavior starting at the point of origin (Fig. 1). The parameter A/Vasym was defined as the concentration of microorganisms at the stationary phase and is an indicator of the inhibition of microbial growth. The growth rate k and the growth rate at the inflection point (A/Vasym x k)/4 represent the maximum rate of microbial growth. The inflection point tc can be used to describe the inhibition “ length,” i.e., the period for which growth is inhibited by the presence of antimicrobials.

RESULTS AND DISCUSSION Characterization of emulsions. Oil-in-water emul­ sions can be used as carrier systems for colors, flavors, antioxidants, or antimicrobials. In this study, oil-in-water emulsions were prepared by high pressure homogenization (10,000 lb/in2, three passes) to function as carrier systems for the essential oil components cinnamaldehyde, perillal-

1328

LOEFFLER ET AL.

J. Food Prot., Vol. 77, No. 8

cinnamaldehyde-citral, 3:1 perillaldehyde-citral, and 1:1 cinnamaldehyde-perillaldehyde) were measured during 2 weeks of storage at room temperature as an indicator of emulsion stability. The particle sizes of loaded emulsions were 50 to 80 nm (Fig. 2), and emulsions remained stable during the 2 weeks. No visible cream layers were formed. Antimicrobial activity assays. Microtiter assays allow the monitoring of microbial growth behavior and the determination of the MICs needed to totally inhibit microbial growth by recording changes in the OD. Table 1 gives a summary of the MICs determined for the dual-loaded emulsions cinnamaldehyde-citral (3:1), cinnamaldehyde-perillaldehyde (1:1), and perillaldehydecitral (3:1) against Z. bailii ATCC 60484 and Z. bailii 906 Pepsi, which were subsequently used as starting concen­ trations in the antimicrobial activity tests via plate counts in the beverage systems. The highest antimicrobial activity was achieved with emulsions that contained cinnamalde­ hyde as one component. Both strains of Z. bailii were either inhibited or inactivated by emulsions with concen­ trations of 200 to 250 pg/ml. The high antimicrobial activity of cinnamaldehyde, which had been previously reported (14, 18), may be the result of its relatively high solubility in water compared with many other essential oil components (21). Results from microtiter assays revealed differences in the growth behavior of Z. bailii ATCC 60484 and Z. bailii 906 Pepsi in the presence of essential oil components; the 906 Pepsi strain was more sensitive than the ATCC 60484 strain for all examined antimicrobial combinations (Fig. 3). However, use of microtiter assays to determine the antimicrobial activity of encapsulated essential oils in beverage systems is limited when samples have high turbidity, which is especially a problem for testing cloudy apple juice (particle sizes of 0.5 to 2 pm and very few large cell fragments). Thus, the antimicrobial impact of essential oils was studied with macrobroth assays, which are not affected by sample turbidity. The antimicrobial activity of the combinations of cinnamaldehyde, citral, and perillaldehyde used against Z. bailii were examined with macrobroth assays in SDB and in “ diluted” beverage systems (2% clear or cloudy apple juice plus malic buffer). Because of their water insolubility, the essential oil components were added in two different ways: encapsulated in nanoemulsions or dissolved in DMSO. Test

ca/pa (1:1) ca/cit (3:1) pa/cit(3:1)

FIGURE 2. Z-average droplet diameter o f oil-in-water emulsions with 5% oil phase (5% Miglyol and 95% essential oil components) and 95% water phase (1% [wtlwt] Tween 20 and malic buffer, pH 3.4). The oil phase was loaded with combinations o f cinnamaldehyde (ca)-perillaldehyde (pa) (ratio, 1:1), ca-citral (cit) (ratio, 3:1), and pa-cit (ratio, 3:1). Emulsions were stored for 14 days at room temperature.

dehyde, and citral to inhibit growth of spoilage yeasts in fruit-based soft drinks. Emulsions could be loaded with essential oil components up to 95% referred to the oil phase (5%, wt/wt) without causing rapid destabilization of the emulsion due to Ostwald ripening, i.e., the growth of large droplets at the expense of smaller ones (5). Destabilization caused by Ostwald ripening was prevented by using Miglyol as the carrier oil. Miglyol 812 does not have antimicrobial activity within its normally used concentration ranges in emulsions (approximate oil phase concentration of 5 to 10%). However, Miglyol used at very high concentra­ tions may have an impact on particle size and thus also on the release or partitioning behavior of such compounds as essential oil components. Nevertheless, the Miglyol con­ centrations used in this study were very low (stock emulsions used were 95% aqueous phase and 5% oil phase; the oil phase was 95% essential oil and 5% Miglyol), and therefore no impact on the antimicrobial activity was expected. The mean diameters of droplets in emulsions con­ taining various blends of essential oil components (3:1

TABLE 1. MICs o f dual emulsions o f essential oil components (EOCs) cinnamaldehyde, perillaldehyde, and citral against Zygosaccharomyces bailii 906 Pepsi and ATCC 60484a MICs against Z. bailii (ng/ml)* EOC1

EOC2

Mixing ratio (EOCl:EOC2)

906 Pepsi

ATCC 60484

Cinnamaldehyde Perillaldehyde Cinnamaldehyde

Citral Citral Perillaldehyde

3:1 3:1 1:1

200 800 250

200 >1,200 250

a Emulsion composition was 5% oil phase (5% Miglyol and 95% EOCs) and 95% water phase (malic buffer and 1% [wt/wt] Tween 20, pH 3.4) at different mixing ratios. h MICs were determined against Z. bailii inoculated at 102 to 103 CFU/ml, using a microtiter assay at OD630. Samples were incubated at 25°C and examined for 7 days.

J. Food Prot., Vol. 77, No. 8

REPLACEMENT OF TRADITIONAL PRESERVATIVES IN FRUIT JUICES

1329

FIGURE 3. Growth behavior o f Zygosaccharomyces bailii 906 Pepsi (A panels) and Z. bailii ATCC 60484 (B panels) in the presence o f different concentrations o f oilin-water emulsions: 5% oil phase (5% Miglyol and 95% essential oil components) and 95% water phase (1% [wtlwt] Tween 20 and malic buffer; pH 3.4). The oil phase was loaded with dual essential oil compo­ nent combinations: (1A, IB) cinnamaldehyde (caf-citral (cit) (ratio, 3:1), (2A, 2B) perillaldehyde (paj-cit (ratio, 3:1), or (3A, 3B) ca-pa (ratio, 1:1). Growth was assessed by optical density (OD) measurements (630 nm) every 24 h.

TABLE 2. Growth parameters o f Zygosaccharomyces bailii 906 Pepsi and ATCC 60484 in model medium and apple juice without essential oil components as determined with a macrobroth assaya Medium

ANasym (log CFU/ml)

k (log CFU/ml/day)

AWasym x A/4 ([log CFU/ml]2/day)

h (days)

R2

3.80 3.88 4.62

16.72 18.88 19.53

15.89 18.30 22.54

1.00 1.00 1.00

0.98 0.98 0.99

3.95 4.00 3.98

18.10 13.59 21.08

17.60 13.58 20.95

1.00 1.00 1.00

0.98 0.96 1.00

Z. bailii 906 Pepsi Clear apple juice Cloudy apple juice Sabouraud dextrose broth Z. bailii ATCC 60484 Clear apple juice Cloudy apple juice Sabouraud dextrose broth

a Growth parameters are from the model equation (Fig. 1), where A/Vasym is the asymptote of the growth curve, k is the growth rate, tc is the position of the inflection point, and R2 is the coefficient of determination. Yeast inoculation levels were approximately 102 CFU/ml; final levels after 27 days of storage at 25°C were approximately 106-CFU/ml.

1330

LOEFFLER ET AL.

J. Food Prot., Vol. 77, No. 8

TABLE 3. Effect o f essential oil components emulsified or dissolved in DMSO in model medium and apple juice on the growth o f Z. bailii 906 Pepsi and ATCC 60484 as determined with a macrobroth assaya Medium

Application system*

Cinnamaldehyde-perillaldehyde (1:1)

Cinnamaldehyde-citral (3:1)

Perillaldehyde-citral (3:1)

62.5 pg/ml

125 pg/ml

50.0 pg/ml

100 pg/ml

200 pg/ml

400 pg/ml

+ + +

—

+ + +

—

+ + +

— +

+

—

+

—

+

—

+ + +

—

+ + +

—

+ + +

+

+

—

+

—

+

—

+ —

—

-

+ + +

— -

+

—

+

+

—

—

—

+

—

+ + +

—

+ + +

—

+ + +

+

+

—

+

+

+

+ + +

—

+ + +

—

+ + +

+

+

—

+

—

+

+

+ + +

—

+ -

— -

— -

+

—

+ + +

+

+

—

+

—

+

—

Z. bailii 906 Pepsi Clear apple juice

Cloudy apple juice

Sabouraud dextrose broth

Emulsified DMSO Emulsified + unloaded emulsion DMSO + unloaded emulsion Emulsified DMSO Emulsified + unloaded emulsion DMSO + unloaded emulsion Emulsified DMSO Emulsified + unloaded emulsion DMSO + unloaded emulsion

Z. bailii ATCC 60484 Clear apple juice

Cloudy apple juice

Sabouraud dextrose broth

Emulsified DMSO Emulsified + unloaded emulsion DMSO + unloaded emulsion Emulsified DMSO Emulsified + unloaded emulsion DMSO + unloaded emulsion Emulsified DMSO Emulsified + unloaded emulsion DMSO + unloaded emulsion

a Essential oil components were applied at one-half or one-quarter of the MIC. MICs were 250 pg/ml for cinnamaldehyde-perillaldehyde, 200 pg/ml for cinnamaldehyde-citral, and 800 pg/ml for perillaldehyde-citral. + , growth; —, no growth. * Components were emulsified or dissolved in DMSO and added to the model medium (broth) or to diluted clear or cloudy apple juice with or without 0.15% (wt/wt) unloaded emulsion.

m edia (juice or broth) were adjusted to pH 3.4 and then inoculated w ith the Z. bailii strains at 102 CFU /m l because beverages are generally contam inated w ith such yeast strains at low levels. Sam ples w ere stored at 20°C (approxim ately room tem perature) before m icrobiological analysis to sim ulate typical storage conditions o f apple juice in a household. Plate counts w ere then m odulated, and the grow th param eters w ere determ ined w ith equation 1. In the first step, the im pact o f pH and sam ple com position w ithout essential oil com ponents (control sam ples) w as determ ined. G row th o f both strains o f Z.

bailii was not inhibited at pH 3.4, and final yeast levels of s 106 C FU /m l were reached in the control sam ples after 3 days o f storage at 25°C. T able 2 gives the grow th param eters o f both Z. bailii strains when no essential oil com ponents were present in the test media. N either inhibition o f Z. bailii grow th by com pounds in the apple juice nor substrate lim itation occurred, and the inflection point tc was always 1 (independent o f sam ple com position and yeast strain). T he Z. bailii levels rem ained at approxim ately 106 C FU /m l after 27 days in sam ples w ith either SDB or apple juice. N evertheless, som e differences

J. Food Prot., Vol. 77, No. 8

REPLACEMENT OF TRADITIONAL PRESERVATIVES IN FRUIT JUICES

1331

FIGURE 4. Growth behavior o f Zygosaccharomyces bailii ATCC 60484 in the presence o f 50 pg/ml (sublethal concentra­ tion, corresponds to one-quarter o f the MIC) blended cinnamaldehyde (ca) and citral (cit) (ratio, 3:1) during 27 days at 25 °C. Systems with 2% clear apple juice (A), 2% cloudy apple juice (B), or SDB (C) and with 0.15% (wtlwt) unloaded emulsion (A , Mi were inoculated with 102 CFU/ml yeast levels. Ca-cit was added at 3:1 either within an oil-in-water emulsion (O , A ) o f 5% oil phase (5% Miglyol plus 95% essential oil components) and 95% water phase (1% [wtlwt] Tween 20 and malic buffer; pH 3.4) or dissolved in DMSO (V, ■ ). Growth was assessed by plate counts (two samples measured in duplicate).

C —• — —O— —▼— —A— —*—

a:

control ca/cit (3:1) emulsified ca/cit (3:1) dissolved in DMSO ca/cit (3:1) emulsified + unloaded emulsion ca/cit (3:1) dissolved in DMSO + unloaded emulsion initial decrease o f viable cells below detection limit

regarding the growth rate k were observed. The higher k values for microbial growth in broth than in juice could be explained by differences in the matrix structure (e.g., presence of particles up to 2 pm in the cloudy juice). Initially, the impact on yeast growth of combinations of essential oil components at concentrations equal to their previously determined MICs was evaluated. These concen­ trations inhibited and/or inactivated both strains of Z. bailii in all beverage systems. After 1 day of sample storage at 25°C, the germination of both yeast strains was below the detection level, indicating inactivation. Similar results were reported by Kisko and Roller (13), who examined the antimicrobial activity of the essential oil components carvacrol and p-cymene against Escherichia coli in apple juices. Antimicrobial concentrations of 1.25 mM inactivated 104 CFU/ml levels of E. coli in less than 1 day. Thus, essential oil components can be very effective within a short time. Concentrations of essential oil components at half of the previously determined MIC (Table 1) had almost the same effect on both strains of Z. bailii as did the MIC. Independent of medium composition and the method of essential oil component application, no microbial growth occurred after treatment with 100 pg/ml cinnamaldehydecitral (3:1; MIC, 200 pg/ml) or 125 pg/ml cinnamaldehydeperillaldehyde (1:1; MIC, 250 pg/ml). In contrast, 400 pg/ml perillaldehyde-citral (3:1; MIC, 800 to >1,200 pg/ml) were

not sufficient to completely inactivate Z. bailii ATCC 60484 when an unloaded emulsion (0.15% wt/wt) was present in the beverage model system. However, in systems containing 2% apple juice to which no unloaded emulsion had been added, Z. bailii ATCC 60484, which was the more resistant strain, was also completely inactivated at half of the MIC. These results differed from those obtained with the microtiter assay, where the yeasts were apparently not inactivated with any of the combinations of essential oil components at half of the MIC. The difficulties associated with comparing results from different methods is well known. The MIC or minimal lethal concentration (MLC) of antimicrobials depend on such factors as temperature, type of organism, and inoculum level (15). Unlike the macrobroth assay, it is not possible to differentiate between dead and viable cells with the microtiter assay (OD measurement). In this study, slight differences in inoculum levels may have affected the results of the two assays (4,20). The 96-well plates of the microtiter assay were shaken during storage, which increased the available oxygen and therefore may have provided a more hospitable environment for yeast growth, making these microorganisms less susceptible to the antimicrobials. The lowest concentrations of combinations of essential oil components examined with the macrobroth assay were one-quarter of the previously ascertained MICs (Table 1), i.e., 50 pg/ml cinnamaldehyde-citral (3:1), 200 ppm of perillaldehyde-citral (3:1), and 62.5 pg/ml cinnamaldehyde-

1332

LOEFFLER ET AL.

FIGURE 5. Growth behavior of Zygosaccharomyces bailii ATCC 60484 in the presence of 62.5 pglml (sublethal concen­ tration, corresponds to one-quarter of the MIC) of blended cinnamaldehyde (ca) and perillaldehyde (pa) during 27 days at 25 °C. Systems with 2% clear apple juice (A), cloudy apple juice (B), or SDB (C) and with 0.15% (wtlwt) unloaded emulsion (A, ■) were inoculated with 102 CFUIml yeast levels. Ca-pa was added at 1:1 either within an oil-in-water emulsion (O, A) of 5% oil phase (5% Miglyol plus 95% essential oil components) and 95% water phase (1% [wtlwt] Tween 20 and malic buffer; pH 3.4) or dissolved in DMSO (▼, ■). Growth was assessed by plate counts (two samples measured in duplicate).

J. Food Prot., Vol. 77, No. 8

A

B

C —• — control —o — ca/pa (1:1) emulsified r ca/pa (1:1) dissolved in DMSO —a — ca/pa (1:1) emulsified + unloaded emulsion -■ ca/pa (1:1) dissolved in DMSO + unloaded emulsion

a:

perillaldehyde (1:1). Table 3 gives a summary of the effect of essential oil components in all media (with or without unloaded emulsion) on the growth of both yeast strains in the macrobroth assay; the component concentrations (one-half or one-quarter of the MIC) were derived from the MICs initially determined with the microtiter assay. In most cases, onequarter of the MIC of the essential oil components did not completely prevent but only inhibited yeast growth during 27 days of incubation. In most cases a prolonged lag phase was observed, but final populations were similar, as shown in Figure 4 for cinnamaldehyde-citral (3:1) and Figure 5 for cinnamaldehyde-perillaldehyde (1:1) against Z. bailii ATCC 60484. The number of viable cells decreased initially below the detection limit of 102 CFU/ml in some beverage systems, but the cells recovered and grew back possibly because of adaption to the media. Because the growth model applied can describe only sigmoidal microbial growth, it could not be used in these cases. Z. bailii 906 Pepsi and Z. bailii ATCC 60484 behaved differently when exposed to 50 pg/ml cinnamaldehyde-citral (3:1) or to 62.5 pg/ml cinnamalde­ hyde-perillaldehyde (1:1). Z. bailii 906 Pepsi was easier to inhibit and thus considered more sensitive. Inactivation of this yeast occurred in SDB and in all beverage systems except those containing the emulsified essential oil components and an unloaded emulsion. Because of initial inactivation and later recovery of Z. bailii 906 Pepsi in the apple juice systems, most of the data could not be fitted with the model. Thus, the

in itial d ecrease o f v iab le c ells below d e te c tio n lim it

growth parameters apply only to Z. bailii ATCC 60484. However, the model is suitable for other strains and isolates, as determined by Dai et al. (6). Table 4 shows the results of the model equation (Figs. 4 and 5) for cinnamaldehyde-citral (3:1) and cinnamaldehyde-perillaldehyde (1:1) according to the val­ ues plotted. In most cases, addition of essential oils had no effect on the parameter A/Vasym, the level of microorganisms at the stationary phase. In contrast, the growth rates (k and A/Vasym x k/4) of Z. bailii ATCC 60484 were affected by the essential oil components; growth was markedly slowed in their presence. Increases in the tc values in response to the essential oil components were observed, indicating a retardation of microbial growth. The degree of inhibition did not differ greatly across the beverage systems. No inhibition of yeast growth was only observed in samples containing the second unloaded emulsion in addition to the antimicrobial emulsion. Low tc values were calculated for these systems (Table 4). Use of flavor emulsions are very common in the beverage industry. These emulsions are loaded with citrus oils and chemicals such as alcohols, aldehydes, ketones, and esters to create a characteristic aroma and flavor profile (9). Weighting agents are added to increase the specific gravity of the oil phase and, therefore, play a part in contributing to the stabilization of flavor emulsions. Modified starches and gum arabic are typically used as emulsifiers (10). Gum arabic was the

J. Food Prot., Vol. 77, No. 8

1333

REPLACEMENT OF TRADITIONAL PRESERVATIVES IN FRUIT JUICES

TABLE 4. Growth parameters o f Z. bailii ATCC 60484 in model medium and apple juice with cinnamaldehyde plus perillaldehyde

or citrala Medium

Application system*

AA/aSym (log CFU/ml)

A/Vasym x k/4 k (log CFU/ml/day) ([log CFU/mlf/day)

tc (days)

R2

Cinnamaldehyde-perillaldehyde (1:1) (62.5 pg/ml) Clear apple juice

Cloudy apple juice

Sabouraud dextrose broth

Emulsified DMSO Emulsified + unloaded emulsion DMSO + unloaded emulsion Emulsified DMSO Emulsified + unloaded emulsion DMSO + unloaded emulsion Emulsified DMSO Emulsified + unloaded emulsion DMSO + unloaded emulsion

4.05 3.71 3.67

0.88 1.09 1.57

0.89 1.01 1.44

8.03 8.40 2.70

0.99 1.00 0.98

3.79

0.40

0.38

7.06

0.95

3.87 3.75 3.69

0.82 0.60 1.52

0.80 0.56 1.40

7.04 4.54 2.77

1.00 0.82 0.98

3.69

0.55

0.50

5.82

0.91

4.49 3.74 3.82

0.20 0.57 1.62

0.22 0.54 1.54

13.5 4.65 2.75

0.87 0.81 0.98

3.46

7.71

6.67

6.83

0.95

Emulsified DMSO Emulsified + unloaded emulsion DMSO + unloaded emulsion Emulsified DMSO Emulsified + unloaded emulsion DMSO + unloaded emulsion Emulsified DMSO Emulsified + unloaded emulsion DMSO + unloaded emulsion

4.32 NG 3.79

0.32 NG 2.98

0.34 NG 2.82

14.30 NG 4.49

0.99 NG 0.99

4.26

.0.26

0.28

9.63

0.92

3.49 3.14 3.99

7.74 2.81 1.41

6.75 2.20 1.40

2.92 14.82 2.49

0.88 0.96 0.98

3.92

0.53

0.52

4.20

0.93

3.49 NG 4.00

2.05 NG 1.74

1.79 NG 1.74

3.51 NG 3.68

0.89 NG 1.00

4.34

0.37

0.40

12.10

0.90

Cinnamaldehyde-citral (3:1) (50 pg/ml) Clear apple juice

Cloudy apple juice

Sabouraud dextrose broth

a Growth parameters are from the model equation (Fig. 1), where A/Vasym is the asymptote o f the growth curve, k is the growth rate, tc is the position of the inflection point, and R2 is the adjusted coefficient of determination (P < 0.001). NG, no growth. * Essential oil components were emulsified or dissolved in DMSO and added to the model medium (broth) or diluted clear or cloudy apple juice with or without 0.15% (wt/wt) unloaded emulsion.

stabilizer in the unloaded emulsion used in this study and was added to the beverage systems to simulate the presence of a flavor emulsion. The antimicrobial activity of dissolved and emulsified essential oil components was influenced by the addition of the unloaded emulsion (Fig. 4) possibly because of oil exchange between the loaded and unloaded emulsions directly through the aqueous phase or via surfactant micelles through the aqueous phase ( 17). The same phenomenon was reported when essential oils were applied in food products with high fat contents ( 23). In summary, the antimicrobial activity of essential oil

components generally decreased in the presence of an unloaded emulsion. Differences in the growth behavior of Z. bailii in apple juice and SDB were observed. The presence of pulp in juices can affect the activity of essential oils. Clear apple juice is manufactured by the separation of clouding agents, such as proteins and polyphenols, through enzymatic and centrifugal clarification. Tserennadmid et al. ( 26) found that particles in cloudy apple juice (approximately 0.5 to 2 pm) can decrease the antimicrobial activity of essential oils. Essential oil components may adhere to particles and sink to

1334

LOEFFLER ET AL.

the bottom of the container, which reduces their antimicro­ bial effectiveness. Tserennadmid et al. (26) also found lower MICs in SDB than in apple juice. The same tendency was noted in our study. The antimicrobial activity was influenced by the type of medium, possibly because of interactions between the functional groups of the essential oil components and the juice ingredients. However, differences between clear and cloudy apple juice were not observed, which might be due to the low concentration of apple juice (2%) in the samples. These results suggest that cinnamaldehyde alone could be used as a natural preservative to control growth of yeast strains resistant to weak acids. However, combinations of antimicrobials that simultaneously attack different targets in a microbial cell can have strong synergistic effects, particularly when their modes of action are also different. In this case, concentrations of the components in combination required to cause total growth inhibition or inactivation were significantly lower than those needed if a single component would have been used. Bevilacqua et al. (3) reported improved control of germination of Alicyclobacillus spores in apple juice during a period of 7 days when a mixture of 40 ppm of cinnamalde­ hyde and 40 ppm of eugenol (a phenolic compound like citral) was added to the product. The effects of the combinations used in that study also were synergistic. In the present study, low concentrations of emulsions (one-quarter of the MIC) loaded with combinations of the essential oil components cinnamaldehyde, perillaldehyde, and citral inhibited and at sufficient concentrations inactivated the spoilage yeast Z. bailii. Encapsulation ensures a homogenous distribution of hydrophobic essential oil components within the food matrix. Very little impact of variations in the beverage matrix on the antimicrobial activity of these compounds was observed. Therefore, addition of emulsified essential oil components is a promising alternative to the use of synthetic preservatives, such as benzoic acid or sorbic acid, in beverages suited for the application of essential oils, which should not contain high amounts of lipids or proteins that could, e.g., protect microorganisms from the antimicrobials or interact with the essential oil components, lowering their antimicrobial efficacy (12, 24). Some essential oils, such as citrus and cinnamon oils, have already been used in beverages as flavoring agents, and the acidic pH of many beverages promotes antimicrobial activity of essential oils. Because of the increased hydrophobicity, these oils dissolve in the lipid phase of the cell membrane (1). However, analyses of the impact of essential oil components on the organoleptic properties of the final product are needed before these components can be used; a negative impact at the low concentrations needed is not expected. The combination of essential oil components with traditional preservatives may even lead to lower MICs and therefore should also be investigated. ACKNOWLEDGMENT We thank PepsiCo for their financial support, which enabled this study.

REFERENCES 1.

Acosta, E. 2009. Bioavailability of nanoparticles in nutrient and nutraceutical delivery. Curr. Opin. Colloid Interface Sci. 14:3-15.

J. Food Prot., Vol. 77, No. 8

2.

3.

4.

5. 6.

7. 8.

9. 10. 11.

12.

13. 14.

15.

16.

17.

18. 19.

20.

21.

22.

23.

24.

25.

Back, W. 1993. Mikrobiologie der Fruchtsaft- und Erfrischungsgetranke. In H. H. Dittrich (ed.), Mikrobiologie der lebensmittel. Getraenke Behr’s Verlag, Hamburg, Germany. Bevilacqua, A., M. R. Corbo, and M. Sinigaglia. 2010. Combining eugenol and cinnamaldehyde to control growth of Alicyclobacillus acidoterrestris. Food Control 21:172-177. Borobio, M. V., A. Pascual, M. Carmen Dominguez, and E. J. Perea. 1986. Effect of medium, pH, and inoculum size on activity of ceftizoxime and Sch-34343 against anaerobic bacteria. Antimicrob. Agents Chemother. 30:626-627. Capek, I. 2004. Degradation of kinetically-stable o/w emulsions. Adv. Colloid Interface Sci. 107:125-155. Dai, Y., L. A. McLandsborough, J. Weiss, and M. Peleg. 2010. Concentration and application order effects of sodium benzoate and eugenol mixtures on the growth inhibition of Saccharomyces cerevisiae and Zygosaccharomyces bailii. J. Food Saf. 75:482-488. Dillon, V. M., and R. G. Board (ed.). 1994. Natural antimicrobial systems and food preservation. CAB International, Wallingford, UK. Fitzgerald, D. J., M. Stratford, and A. Narbad. 2003. Analysis of the inhibition of food spoilage yeasts by vanillin. Int. J. Food Microbiol. 86:113-122. Friberg, S., K. Larsson, and J. Sjoblom (ed.). 2004. Food emulsions. Marcel Dekker, New York. Given, P. S., Jr. 2009. Encapsulation of flavors in emulsions for beverages. Curr. Opin. Colloid Interface Sci. 14:43^17. Gram, L., L. Ravn, M. Rasch, J. B. Bruhn, A. B. Christensen, and M. Givskov. 2002. Food spoilage— interactions between food spoilage bacteria. Int. J. Food Microbiol. 78:79-97. Gutierrez, J., C. Barry-Ryan, and P. Bourke. 2008. The antimicrobial efficacy of plant essential oil combinations and interactions with food ingredients. Int. J. Food Microbiol. 124:91-97. Kisko, G., and S. Roller. 2005. Carvacrol and p-cymene inactivate Escherichia coli 0157:H7 in apple juice. BMC Microbiol. 5:36. Kurita, N., M. Miyaji, and R. Kurane. 1979. Antifungal activity and molecular orbital energies of aldehyde compounds from oils of higher plants. Agric. Biol. Chem. 43:2365-2371. Lambert, R. J. W. 2000. Susceptibility testing: inoculum size dependency of inhibition using the Colworth MIC technique. J. Appl. Microbiol. 89:275-279. McClements, D. J., E. A. Decker, and J. Weiss. 2007. Emulsionbased delivery systems for lipophilic bioactive components. J. Food Sci. 72:R109-R124. McClements, D. J., S. R. Dungan, J. B. German, and J. E. Kinsella. 1992. Oil exchange between oil-in-water emulsion droplets stabilised with a non-ionic surfactant. Food Hydrocolloids 6:415^122. Moleyar, V., and P. Narasimham. 1992. Antibacterial activity of essential oil components. Int. J. Food Microbiol. 16:337-342. Qian, C., and D. J. McClements. 2011. Formation of nanoemulsions stabilized by model food-grade emulsifiers using high-pressure homogenization: factors affecting particle size. Food Hydrocoil. 25: 1000-1008. Riha, W. E., and M. Solberg. 1975. Clostridium perfringens inhibition by sodium nitrite as a function of pH, inoculum size and heat. J. Food Sci. 40:439-442. Samperio, C„ R. Boyer, W. N. Eigel, K. W. Holland, J. S. McKinney, S. F. O ’Keefe, R. Smith, and J. E. Marcy. 2010. Enhancement of plant essential oils’ aqueous solubility and stability using alpha and beta cyclodextrin. J. Agric. Food Chem. 58:12950-12956. Sikkema, J., J. A. M. De Bont, and B. Poolman. 1994. Interactions of cyclic hydrocarbons with biological membranes. J. Biol. Chem. 269: 8022-8028. Singh, A., R. K. Singh, A. K. Bhunia, and N. Singh. 2003. Efficacy of plant essential oils as antimicrobial agents against Listeria monocytogenes in hotdogs. Food Sci. Technol. 36:787-794. Smith-Palmer, A., J. Stewart, and L. Fyfe. 2001. The potential application of plant essential oils as natural food preservatives in soft cheese. Food Microbiol. 18:463^170. Thomas, D. S., and R. R. Davenport. 1985. Zygosaccharomyces bailii—a profile of characteristics and spoilage activities. Food Microbiol. 2:157-169.

J. Food Prot., Vol. 77, No. 8

26.

27.

28.

REPLACEMENT OF TRADITIONAL PRESERVATIVES IN FRUIT JUICES

Tserennadmid, R., M. Tako, L. Galgoczy, T. Papp, M. Pesti, C. Vagvolgyi, K. Almassy, and J. Krisch. 2011. Anti yeast activities of some essential oils in growth medium, fruit juices and milk. Int. J. Food Microbiol. 144:480-486. Ultee, A., M. H. J. Bennik, and R. Moezelaar. 2002. The phenolic hydroxyl group of carvacrol is essential for action against the food-borne pathogen Bacillus cereus. Appl. Environ. Microbiol. 68:1561-1568. Wang, X., Y. Jiang, Y.-W. Wang, M.-T. Huang, C.-T. Ho, and Q. Huang. 2008. Enhancing anti-inflammation activity of curcumin through o/w nanoemulsions. Food Chem. 108:419-424.

29.

30.

1335

Weiss, J., S. Gaysinsky, M. Davidson, and J. McClements. 2009. Nanostructured encapsulation systems: food antimicrobials. In G. Barbosa-Canovas, A. Mortimer, D. Lineback, W. Spiess, K. Buckle, and P. Colonna (ed.), Global issues in food science and technology. Academic Press, San Diego, CA. Zhou, F., B. Ji, H. Zhang, H. U. I. Jiang, Z. Yang, J. Li, and W. Yan. 2007. The antibacterial effect of cinnamaldehyde, thymol, carvacrol and their combinations against the foodborne pathogen Salmonella Typhimurium. / . Food Saf. 27:124133.

Copyright of Journal of Food Protection is the property of Allen Press Publishing Services Inc. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.