Effects of natural antimicrobials with modified atmosphere packaging on the growth kinetics of Listeria monocytogenes in ravioli at various temperatures

Abstract The objective of this study was to investigate the antimicrobial effects of cultured sugar/vinegar (CSV) blend and nisin to control the risk of Listeria monocytogenes in ready to cook (RTC) ravioli. Ravioli dough was prepared with 0.1, 0.3, 0.5, 1% CSV blend and 0.1, 0.2, and 0.3% nisin. Inoculated spinach or artichoke raviolis with 2.0 ± 0.5 log cfu/g of L. monocytogenes were packed aerobically or using modified atmosphere packaging (MAP), and then stored at 4, 10, 17, and 24 °C for 60 days. Growth kinetic parameters of the observed data fit well to the Baranyi equation. Ravioli with spinach filling materials yielded a higher risk than that with artichoke. L. monocytogenes was able to survive in ravioli with artichoke, but did not grow. The addition of 1% CSV blend or 0.3% nisin in spinach ravioli with the combination of MAP effectively controlled the growth of L. monocytogenes at the temperature below 10 °C. The organoleptic quality of spinach ravioli was not also affected by the application of 1% CSV blend. Therefore, the CSV blend can be recommended to improve the microbial safety and quality of natural RTC ravioli at retail market. Practical applications The risk of ravioli was affected by the filling materials of ravioli at retail market. Addition of 1% cultured sugar/vinegar blend in dough substantially contributes to the extension of shelf‐life of MAP spinach raviolis. classification and regression tree analysis results indicate that refrigeration temperature is the main control factor to affect lag time and growth rate, while packaging method is critical for maximum population density.

foods (Organic Trade Association, 2009). Thus, to emphasize the safety of used ingredients, producers are seeking clean-label alternative ingredients with less "chemical sounding" names. Cultured sugar and vinegar blend (CSV blend, Purac verdad NV55), a natural alternative to lactic acid salt compound, is generally regarded as safe and can be added to food products during processing. A natural blend of CSV consists of fermentation with specially curated cultures, such as sugars, organic acids, peptides, and aromas. The reagent is designed to have a high level of Listeria control (Purac, 2013). Glass and Sindelar (2010) and Sullivan et al. (2012) reported that the CSV blend inhibited the growth of L. monocytogenes in meat products. Recently, antimicrobial effect of 3% CSV blend on behavior of Campylobacter jejuni and Salmonella Typhimurium in chicken breast was also reported (Park, Hong, & Yoon, 2014).
In the present study, we investigated the antimicrobial effect of CSV blend on the growth control of L. monocytogenes in raviolis filled with spinach or artichoke and compared to that of nisin. Combined effects of modified atmosphere packaging (MAP) and storage temperature as hurdle techniques at retail market were also evaluated.

| M AT ER I AL S A N D M E TH ODS
2.1 | Bacterial culture for inoculation study L. monocytogenes strains (ATCC 19111, ATCC 19115, and ATCC 15313) were purchased from the Korean Culture Center of Microorganisms (KCCM, Seoul, Korea) and maintained at 280 8C with tryptic soy broth (TSB, Difco, Sparks, MD) with 0.6% yeast extract (TSBYE) (Oxoid, Basingstoke, United Kingdom) in beads. For each experiment, each stock culture was thawed and inoculated in 10 ml of TSBYE, followed by incubation at 37 8C for 24 hr. Viable cell counts of L. monocytogenes ranged between 8.5 and 9.0 log cfu/ml after incubation. An aliquot of 1 ml of the initial culture in stationary phase was transferred into 9 ml of 0.85% NaCl, which was serially diluted. A mixture of three L. monocytogenes strains was prepared before inoculation into the ravioli samples.
Spinach or artichoke base materials were provided from Giloy ravioli manufacturer (Pulmuone Co., CA). The CSV blend was uniformly added to the spinach or artichoke ravioli dough at the concentration of 0.1, 0.3, 0.5, or 1%, while nisin was added to only spinach ravioli dough at the concentrations of 0.1, 0.2, or 0.3%. Control was prepared without antimicrobial agents. The dough samples were formed into rectangle ravioli shape (8.5 6 0.5 g) and then filled with 18 g of spinach or artichoke base materials. The prepared raviolis were then heated at 121 8C for 10 min to remove any background microflora. After cooling, the raviolis were transferred into petri dishes (9 3 2 cm). The surface of each ravioli was uniformly inoculated with 0.1 ml of the diluted cultures of L. monocytogenes mixture using a sterile repeater pipette to reach the target inoculation levels (2.0 6 0.5 log cfu/g). The inoculated spinach raviolis were packaged aerobically or in modified atmosphere with gas mixture (0% O 2 , 0.3% CO 2 , and N 2 99.7%) (Composite Deoxidation Desiccant, Lipmen, Incheonm, Korea). The inoculated artichoke raviolis were also packed under MAP. Gas-tight wrapper type bags filled with a deoxidizer (PPEC Eumsung Fresh Noodle Co., Ltd., Eumsung, Korea) were used for MAP. The packaged samples were separately stored at 4, 10, 17, and 24 8C for 60 days.

| Enumeration of L. monocytogenes
At the specific time intervals during storage, each sample was homogenized (Bag-Mixer 400, Interscience, Paris, France) in 90 ml of 0.85% NaCl (vol/vol) 2 min. One milliliter of the homogenized sample was diluted with 9 ml of 0.85% NaCl (vol/vol) and 0.1 ml aliquots of two dilutions of each sample was spiral plated (Automatic spiral plater, Interscience, Saint Nom, France) in duplicate on PALCAM agar (Oxoid, Basingstoke, United Kingdom) for L. monocytogenes and then incubated aerobically at 36 8C for 48 hr. The colonies on the duplicate plates of each sample were counted by an automated colony counter (Scan 300, Interscience, Saint Nom, France). The results were expressed as log cfu/g and each experiment was repeated twice.

| Growth kinetics of primary modeling for L. monocytogenes in ravioli
Growth representing viable cell counts (log cfu/g) of L. monocytogenes as a function of time was iteratively fit to the Baranyi equation using the DM Fit 3.5 curve-fitting program (Institute of Food Research, Norwich, United Kingdom). The equation used was as follows (Baranyi & Roberts, 1994): ln 11 e l max A 21 where y is the logarithm of the cell numbers (log cfu/g), y 0 is the initial cell number, y max is the final cell number, A is the time variable, l max is the specific growth rate (SGR; log per day), q 0 is the physiological state of the inoculum; t lag is the lag time (LT); and t is the sampling time. The goodness of fit of the data was evaluated based on the coefficient of determination (R 2 ). Three parameters, namely, LT, SGR, and maximum population density (MPD) were calculated from the equations described by Baranyi and Roberts (1994) and used for evaluation criteria.
2.5 | Secondary modeling for the effect of temperature on the growth kinetics of L. monocytogenes Response surface equations as a function of temperature and the concentration of the CSV blend were developed for LT, SGR, and MPD of L. monocytogenes in spinach ravioli with MAP by multiple regressions using the SAS (V 9.3) General Linear Models Procedure: ln y5 a 0 1 a 1 A1 a 2 B1a 3 A3B1 a 4 A3A1 a 5 B3B1E (2) where ln y is the natural logarithm of the modeled growth parameters (LT, SGR, and MPD), A is the temperature, B is the concentration of the CSV blend, a 0 -a 5 are regression coefficients, and e is the random error.

| Sensory evaluation
A triangle test was performed to determine whether a significant flavor change occurred in spinach ravioli with 1% CSV blend. The minimum number of panelist needed for the test was determined from the table of significant test for triangle test (Kim, Kim, Sung, & Lee, 1993). To be >95% certain (b 5 0.05 and a 5 0.05) that no more than 50% (p d 5 50%) of consumers would be able to detect a difference if CSV the blend was applied to spinach ravioli, the required number of panelists is n 5 19. In the present study, 40 trained panelists from the department of research and development at Pulmuone Co., Ltd. participated in sensory evaluation. Spinach ravioli with 1% CSV blend were compared with two controls without antimicrobial treatment. Three samples were randomly numbered with 3digit codes and placed in a random order. Two of three samples were identical and the other was different. The panelists were asked to state which product they believed was the odd one, which was different based on appearance, smell, taste, etc. Testing was conducted in individual booths in the Pulmuone R&D center. The panelists were equally spaced throughout the room and instructed not to speak to one another during the test.

| Statistical analysis
Three growth kinetics parameters for L. monocytogenes-namely, SGR, LT, and MPD were used to analyze the data. ANOVA model was applied to analyze the effect of multiple treatments for each of the factors: temperature (4, 10, 17, 24 8C) and concentration of antimicrobials (CVS blend: 0.1, 0.3, 0.5, or 1%, nisin: 0.1, 0.2, or 0.3%). For each ANOVA, complete randomized block design was used by setting the other factor combinations as a block using statistical analysis system SAS V 9.3 (SAS Institute Inc., Cary, NC). R package (version 3.1.2) was also used for the classification and regression tree (CART) (Breiman, Friedman, Stone, & Olshen, 1984) analysis. The CART analysis was applied to visualize the suggested hierarchy of variables with respect to the three measured parameters (SGR, LT, and MPD). The CART analysis generates a tree-like structure by the set of decision points yielding partitions of the original group. The more important variables to affect the response, the higher the position on the tree they have as decision points.

| RE SULT S A ND D I SCUSSION
3.1 | Comparison of the growth kinetics of L. monocytogenes by the kind of ravioli, antimicrobial agent, and packaging The growth curves of L. monocytogenes in both spinach and artichoke ravioli for 60 days well fitted to the Baranyi model. The mean values of the growth kinetics of L. monocytogenes were compared according to the kind of raviolis (spinach versus artichoke), antimicrobial agents (CSV blend versus nisin), and packaging (modified atmosphere versus aerobic) (Table 1). There was a higher growth potential of L. monocytogenes in spinach ravioli as compared to artichoke ravioli. Although no significant differences were observed in the SGR and LT values, the MPD values were differed significantly between spinach (5.420 log cfu/g) and artichoke raviolis (4.569 log cfu/g). These results indicate that the different filling in ravioli affects the level of stationary phase of L. monocytogenes and ravioli with spinach filling has a higher risk than that with artichoke filling. L. monocytogenes in artichoke was able to survive, but did not grow, with remaining counts at the initial concentration (around 4.5 log cfu/g) during the storage period (Sanz, Gim enez, & Olarte, 2003). In previous research, extracts of artichoke exhibited antimicrobial activity against bacteria species, yeasts, and molds (Emanuel, Adrian, Sultana, & Svetlana, 2011;Ionescu et al., 2013;Zhu, Zhang, & Lo, 2004). 7.7% contamination rate of L. monocytogenes in spinach was reported, whereas the pathogen was not isolated from artichokes (Cordano & Jacquet, 2009). Other reports also emphasized the hazard of spinach as a food ingredient for humans. Specifically, Pingulkar, Kamat, and Bongirwar (2001) and Yolanda Moreno et al. (2012) reported that the higher percentage of L. monocytogenes positive was detected in spinach samples.
L. monocytogenes prevalence under such environment was attributed to the cross-contamination and ability of growth even at refrigeration, as well as to ambient temperature of processed foods (Beuchat, 1996;Wilks, Michels, & Keevil, 2006). The population of L. monocytogenes increased from 2.4 to 8.8 log cfu/ml in autoclaved spinach powder cultures at 30 8C, indicating that spinach products provide a good environment for the growth of L. monocytogenes (Babic, Watada, & Buta, 1997).
In addition, antimicrobial effects of 0.1% CSV blend and 0.3% nisin on the control of L. monocytogenes growth in spinach raviolis were compared. No significant differences in the LT, SGR, and MPD values were observed between 0.1% CSV blend and 0.3% nisin. Despite this comparable activity, the CSV blend has its advantage of allowing "organic" label. Nisin, which is derived from milk bacteria Streptococcus lactis spp., has an effect on controlling a wide range of gram-positive organisms, including: Listeria spp., Bacillus spp., and Clostridium spp., and their spores (Bruno & Montville, 1993). Although the use of bacteriocins is a novel approach in eliminating or controlling L. monocytogenes in food, the development of resistant strains remains the main concern and limits this use as a biopreservative (Kaur et al., 2011). In addition, nisin is not currently approved as an allowable ingredient in organic food and is expensive to use. Therefore, if two reagents are considered as comparable antimicrobial activity, food industry could benefit more by adapting the CSV blend in ravioli products, rather than nisin. The CSV blend is a  Table 1).
Moreover, the LT in spinach ravioli with MAP was significantly extended up to 15.5 days, as compared to 2.5 days in aerobic packaged spinach ravioli. MPD was also lower in spinach ravioli with MAP (5.68 log cfu/g) than with aerobic packaging (7.51 log cfu/g). The combined treatment of low oxygen and high concentration of CO 2 or N 2 , can provide adequate suppression of the growth of L. monocytogenes (Kostaki et al., 2009). According to Whitley, Muir, and Waites (2000), N 2 MAP in the absence of O 2 increases the LT of L. monocytogenes in cheese up to 3 weeks and retards the growth of L. monocytogenes.

| Modeling of L. monocytogenes growth in spinach ravioli as a function of the CSV blend and storage temperature
To predict the growth restriction of L. monocytogenes by adding the CSV blend in spinach ravioli with MAP, the primary growth models were generated at various temperatures. The impacts of storage temperature and the level of the CSV blend on SGR, LT, and MPD are summarized in  LT 5 lag time (day); SGR 5 specific growth rate (log/day); MPD 5 maximum population density (log cfu/g). a Modified packaged ravioli with CSV blend at 0, 0.1, 0.3, 0.5, and 1% was compared. b CSV blend and nisin at 0.1 and 0.3% was compared in spinach ravioli. c Only spinach ravioli treated CSV blend was compared. *Significant differences between the mean values of growth kinetics were determined by t test (p < .05).
function of amount. For example, Glass and Sindelar (2010), Sullivan (2012), and Schrader (2010 also reported that the 3% CSV blend reduced L. monocytogenes growth on meat products. In addition, the survival of other pathogens, such as C. jejuni and S. Typhimurium, on precooked chicken breasts was also controlled with 3% CSV blend (Park et al., 2014). The observed concentration-discrepancy regarding the effectiveness of the antimicrobial could be explained by the difference in food products and food formulations during the course of experiment. The CSV blend used in the present study (up to 1%) in dough was lower than the recommended levels (3-3.5%) by the PURAC and other studies. There were significant interactions between the concentrations of the CSV blend and storage temperatures in spinach ravioli with MAP in the present study. The combined treatment of the CSV blend in dough with MAP packaging was effective at pathogen reduction, without significantly affecting the quality, and demonstrated its potential as a novel method to increase the microbial safety in ravioli, possibly other products, such as pasta and noodle, etc. Of all available options for ensuring safety, maintenance of low temperature throughout this food chain is likely to be the most effective.
The response surface models were also developed to describe the effects of the CSV blend concentration and storage temperature on SGR, LT, and MPD (Table 3) In the present study, the growth of L. monocytogenes was not observed in MAP packed spinach ravioli with 1% CSV blend at 4 8C (Table 2). Therefore, these conditions were excluded from the development of the secondary growth model in the present study. Increased amounts of CSV blend resulted in a significant growth control of L. monocytogenes. The additional hurdle provided by low temperature enhanced the effect of the CSV blend as an antimicrobial agent. Conclusively, the growth of L. monocytogenes in spinach ravioli with MAP was safely controlled by the addition of 0.3 CSV blend at the  LT 5 lag time (day); SGR 5 specific growth rate (log/day); MPD 5 maximum population density (log cfu/g). ac Means (n 5 4) 6 SD in the same column with different superscripts are significantly different by Duncan's multiple range test at p < .05. AD Means (n 5 4) 6 SD in the same row with different superscripts are significantly different by Duncan's multiple range test at p < .05. | 5 of 8 refrigeration temperature. In summary, the models in the matrix of conditions described in the present study can be used as a tool to estimate the impact of CSV blend and temperature on the growth of L. monocytogenes in retail ravioli (Table 3).

| Determination of important control factor for the growth of L. monocytogenes by the CART analysis
We used the CART analysis to discriminate important variables (concentration of antimicrobials, packaging method, and storage temperature) in determining the growth kinetics of L. monocytogenes in spinach ravioli.   packaging was 7.64 log/g, which is 1.4 times higher than MAP. This result indicates that packaging method is the main control factor to affect MPD. In L. monocytogenes in spinach raviolis with MAP, the samples treated with above 0.1% CSV blend showed the smallest MPD value (4.34 log/g), which was 38% lower than that of raviolis without CSV blend (7.05 log/g). In the case of spinach raviolis with aerobic packaging, 0.5% concentration of CSV blend was the critical level. The MPD value of L. monocytogenes in spinach raviolis with less than 0.5% CSV blend with aerobic packaging was 1.3 times (8.50 log/g) higher than that of the CSV blend concentration equal to or above 0.5% (6.36 log/g). It should be noted that the threshold concentration for the MAP partition was between 0.1 and 0.3%, while that for the aerobic packaging ranged between 0.3 and 0.5%. These results indicate that low concentration of antimicrobial agent can be effective to control L. monocytogenes in spinach ravioli with MAP. Although the degree of importance varied, temperature was the most important factor, followed by packaging method and concentration of the CSV blend to control the growth of L. monocytogenes in spinach ravioli.
Consistently with the results of previous studies, our CART analysis results further emphasized that the storage at the refrigeration temperature is a critical factor in controlling L. monocytogenes growth in food products. The combined treatment of the CSV blend with MAP was effective in increasing L. monocytogenes reduction without significantly affecting the quality, and demonstrated its potential as a novel method to increase the microbial safety in ready-to cook refrigerated ravioli.

| Sensory evaluation
In the triangle test, 16 of 40 (45%) panelists correctly chose spinach sample treated with the 1% CSV blend that was different from the other two samples. Thus, it can be concluded with 95% confidence that not more than 50% of the population would be able to detect a difference in color, smell, or taste of spinach ravioli with 1% CSV blend.
Our results also indicated that the product was well accepted by the panelists. Other studies reported that many conventional antibacterial interventions (e.g., irradiation) can result in undesirable alterations to the appearance, taste and smell of food (Gecgel, 2013) and make those foods less desirable to the consumer. However, our results showed that the organoleptic quality of spinach ravioli was not affected by application of 1% CSV blend and that no differences in taste, color, or appearance were detectable.

| CON CLU S I ON S
The results of this study confirm the combined efficacy of the CSV blend as a clean label/natural antimicrobial and other huddle factors, such as low temperature and MAP, on decreasing the risk of