Prevalence of Listeria monocytogenes in ready-to-eat foods, and growth boundary modeling of the selected strains in broth as a function of temperature, salt and nisin

ABSTRACT The objectives of this study were to determine the prevalence of Listeria monocytogenes in ready-to-eat (RTE) foods, and the growth boundaries of two strains of L. monocytogenes isolated from RTE foods and one ATCC7644 strain under different temperature (4, 20, and 37°C), salt (0, 1, 2, 3, and 4% w/v), and nisin (0, 50, 100, 200, and 400 µg/mL) levels with different inoculum sizes (3, 4, 5, and 6 log10 CFU/mL). One hundred thirty-three retail RTE foods were examined and a total of 39 isolates were identified. The positive Listeria spp. and L. monocytogenes samples were 29.3% and 12.8%, respectively. Dairy and fruit-based RTE foods were L. monocytogenes negative whereas seafood-based (26.7%), meat-based (19.4%), poultry-based (16%), and vegetable-based (10%) RTE foods contained L. monocytogenes. It was shown that nisin was more effective than NaCl to avoid growth. Moreover, protective effect of NaCl on nisin activity was observed. As inoculum size increased probability of growth also increased. Suppressing the growth by using temperature, NaCl and nisin was more difficult in RTE-isolates strains (strains 120 and 137) than that of ATCC7644 strain. A mathematical model based on logistic regression could be used to determine growth limits of L. monocytogenes strains with a concordance of > 91%.

a loopful each of the ½ and full Fraser broth were streaked on both ALOA agar (Merck) and PALCAM agar (Merck), and the plates were incubated at 37°C for 24-48 h. Then, five suspected colonies each from ALOA and PALCAM agar plates were picked and subjected to morphological and biochemical tests including by Gram staining, catalase reactions, oxidase tests, carbohydrate utilization, CAMP tests, and motility at 20-25°C. Bacterial isolates were stored in Brain Heart Infusion (BHI) broth (Merck) with 30% sterile glycerol (Merck) at -80°C.

Genomic DNA extraction and PCR identification of Listeria monocytogenes
Following the biochemical tests, L. monocytogenes isolates were confirmed by polymerase-chain reaction (PCR) analysis. Bacterial cultures were grown in Tryptic Soy Broth supplemented with 0.6% of yeast extract (TSB-YE) (Merck) for both the 16S rRNA gene and listeriolysin (hlyA) gene determination. Genomic DNA extraction and PCR amplifications were used as described by Sanlibaba et al. [3] The primer pairs including forward U1 (5′-CAGCMGCCGCGGTAATWC-3′) and reverse LI1 (5′-CTCCATAAAGGTGACCCT-3′) were used to amplify a 938 bp region in the 16S rRNA gene of the Listeria genus. [21] Additionally, the primer pairs 5′-CATTAGTGGAAAGATGGAATG-3′ (forward) and 5′-GTATCCTCCAGAGTGATCGA-3′ (reverse) were used to amplify a 730 bp region of the hlyA gene of L. monocytogenes strains. [22] Escherichia coli ATCC25922 and L. monocytogenes ATCC7644 were used as negative and positive controls, respectively. Amplification conditions were slightly modified and were as follows: initial denaturation at 94°C for 3 min, followed by 35 cycles of denaturation at 95°C for 45 s, annealing at 55°C for 30 s, and extension at 72°C for 2 min, then final extension step at 72°C for 5 min, as described previously. [22] Amplified DNA was separated using 1% agarose gel electrophoresis (w/v) in 1X TBE buffer stained with ethidium bromide solution (2.5 µL/ 100 mL), and then, the gel was visualized under a UV transilluminator (SYNGENE, Biosystems UK). A 1000 bp DNA ladder (Thermo Scientific) was used as a reference marker.

Preparation of nisin solution and determination of minimum inhibitory concentration
Nisin stock solutions were prepared with Nisaplin (2.5% nisin A, Product Number: N5764, Sigma) by adding 0.020 g or 0.200 g to 10 mL of 0.02 N HCl. These solutions were kept at 4°C for a maximum of 24 h. [23] Minimum inhibitory concentration (MIC) assay was used to determine the antimicrobial activity of nisin against L. monocytogenes strains. MIC assays were carried out in 96 well microtiter plates in triplicate as described by Lynch et al. . [24] Experimental design for determination of growth rate of L. monocytogenes To determine the growth rates of 4 strains of L. monocytogenes isolated from RTE meals and an ATCC7644 strain, a microplate reader (BioTek Elisa Reader, Biotek Inc., USA) with 96 wells were used. The optical density (OD) was measured at 600 nm and growth rates were determined by using the formula below [25] : where t OD is the time (h) when OD value is equal to 0.15, that is, time to detection (TTD), λ is the lag time (h), log 10 N OD is the number of L. monocytogenes (log 10 CFU/mL) at OD = 0.15, log 10 N 0 is the initial number of L. monocytogenes (log 10 CFU/mL) and µ is the specific growth rate (h −1 ). To verify the growth rate obtained from Eq.(1), colony counts were also performed and three growth rate models were fitted to the data -see below. For the microbiological plate counts, at appropriate time intervals 1 mL of sample was added to 9 mL of sterile peptone water (0.5%) and mixed for about 15 s. Serial 10-fold dilutions were prepared in peptone water and 0.1 mL of aliquots from each dilution were spread plated in duplicate. All the experiments were repeated twice with duplicate measurements (n = 4). ÖK-BUZ GRoFiT, an Excel freeware file in which Baranyi, [26] modified Gompertz, [27] and three phase linear model [28] is inserted, was downloaded from https://mmf.gidatarim.edu.tr/en/ ok-buz-grofit and used for modeling the microbial count. [29] There are two versions of this tool. Second version of the tool is available both in Turkish (https://mmf.gidatarim.edu.tr/ok-buzgrofit) and in English. The Baranyi model in ÖK-BUZ GRoFiT is a 4-parameter alternative to the one in DMFit which has six parameters. Nevertheless, almost identical fits for these Excel tools could be obtained with the growth data having lag, exponential, and stationary phases. Two strains with the highest growth rates and the ATCC7644 strain were used for growth/no growth studies.

Experimental design for determination of growth boundaries of L. monocytogenes
A full factorial design was used to test the growth ability of L. monocytogenes in TSB with different combinations of temperature (4, 20, and 37°C), sodium chloride (0, 1, 2, 3, and 4% w/v), nisin (0, 50, 100, 200, and 400 µg/mL), and initial concentration (3, 4, 5, and 6 log 10 CFU/mL). Three replications were done for each combination which resulted in 900 data points (3 temperature levels × 5 salt levels × 5 nisin levels × 4 initial concentration levels × 3 replications). Selected temperature levels represented the refrigeration, ambient, and abuse (optimum growth) temperatures. Salt and nisin concentrations were the ones that are generally used or allowed levels (except 400 µg/mL for nisin) in foods. It should be noted that the inoculum sizes of 5 and 6 log 10 CFU/mL were higher than the expected contamination of L. monocytogenes in foods and higher growth probability with these inoculum sizes should be expected. [30] Inoculum sizes of 3 and 4 log 10 CFU/mL were also used to determine growth boundaries under more realistic conditions.

Preparation of broth and growth evaluation
Uninoculated samples were served as negative controls and broths (TSB) with no salt or nisin inoculated with 10 3 , 10 4 , 10 5 , and 10 6 CFU/mL bacteria were used as the positive control. Microbial growth was monitored weekly by measuring OD at 600 nm up to 8 weeks. Same microplate reader with 96 wells were used for growth evaluation. At higher initial inocula (10 5 and 10 6 CFU/mL), 0.1 increase in the OD values were considered as growth. [31] Growth was also verified by plate counts for some wells. At lower inocula (10 3 and 10 4 CFU/mL), growth was monitored both by checking the OD values and by plate counts.

Model development
Two models were proposed: where, p is the probability of growth in the range of 0-1 (growth = 1, no growth = 0), logit(p) = ln[p/ (1-p)] a 0 -a 4, b 0 -b 10 are the coefficients of the model to be estimated, T is the temperature (°C), NaCl is the salt concentration (% w/v), nisin is the nisin concentration (µg/mL), and IC is the initial concentration (log 10 CFU/mL) of L. monocytogenes.
SPSS (Version 22, Chicago, IL, USA) was used for logistic regression and goodness-of-fit of the models were compared by using (i) -2 · ln L with L the likelihood in its optimum, (ii) Hosmer-Lemeshow (H-L) statistic (iii) maximum rescaled R 2 statistic and (iv) percent concordant. [32] The fitted growth/no growth boundaries for p= .1, 0.5, and 0.9 were calculated by using Microsoft® Excel.
The first model was the simplest possible model with the main effects (temperature, salt, nisin, and initial concentration). The second one was more complex one with bilateral interaction terms. If the interaction terms were insignificant (P > .05), they were removed from the model and regression was repeated without those terms, that is, remaining coefficients in the model were all significant (P ≤ .05).

Prevalence of Listeria monocytogenes in RTE food samples
The identification of the 39 isolates was made according to the 16S rRNA sequence results (data not shown). The positive Listeria spp. and L. monocytogenes samples were 39/133 and 17/133, respectively. All L. monocytogenes strains showed a positive result for the presence of the virulence-associated hlyA gene. Results of the prevalence of Listeria spp. and L. monocytogenes in the positive samples are shown in Table 1. In particular, L. monocytogenes isolated from RTE foods was highlighted in 4/15 seafood-based, 6/31 meat-based, 4/25 poultry-based, 3/30 vegetable-based. However, dairy and fruitbased RTE foods were L. monocytogenes negative.

MIC results
Seventeen L. monocytogenes strains isolated in this study were tested for susceptibility to nisin using MIC assays. Nisin can inhibit L. monocytogenes; MIC ranging from 128 to 4096 μg/mL were obtained. Four L. monocytogenes strains were selected among the strains with the highest, middle and lowest MIC values (Table 2). Among these selected strains, 120 and 151 was isolated from meat-based RTE foods, 137 from vegetable-based RTE foods, and also 241 from seafood-based RTE foods. While L. monocytogenes 120 was the most resistant strain to nisin (4096 μg/mL), strains 137 and 241 exhibited the lowest MIC values of 512 μg/mL. Strain 151 also displayed low susceptibility to nisin (2048 μg/mL). However, L. monocytogenes ATCC7644 was sensitive to nisin with an observed MIC value of 128 μg/mL.

Growth rate determination
Value of N OD was determined by enumerating the number of L. monocytogenes in each well when the OD 600 value was about 0.15. Although N OD values changed with respect to strains, they were between 8.0 and 8.5 log 10 . Figure 1 shows the fit of Eq.(1) that is, how the growth rates (µ) were calculated for three strains (120, 137, and ATCC7644) and Table 3 lists the growth rates of each strain. The highest and lowest growth rates were obtained as 2.21 and 2.04 h −1 for strains 120 and 151, respectively. Plate count method was also applied for three strains (120, 137, and ATCC7644) to compare the growth rates of time-to-detection (TTD) data with colony counts. The fit of three models (modified Gompertz, Baranyi, and three-phase linear) used to describe the growth data of L. monocytogenes 137 is shown in Figure 2. All models produced reasonable fits with R 2 adj >0.97 for the other strains and closest growth rates to TTD experiments were the ones obtained from the fit of three-phase linear    (Table 4). According to the model fit, highest growth rate was 2.25 h −1 for the strain 137, followed by 2.10 h −1 for the strain 120 and 2.04 h −1 for the ATCC7644 strain. The growth rates obtained for the Baranyi and Gompertz models were given in Table 5. The information obtained from MIC study combined with the growth rate study, and the strains with higher MIC and growth rates (120, 137, and ATCC7644) were further used for growth/no growth study.

Effects of temperature, salt, nisin, and initial inoculum on the growth of L. monocytogenes
Growth responses of three strains with the inoculum levels of 10 3 and 10 6 CFU/mL are shown in Fig.3 and 4, respectively. In general, same response (growth or no growth) was observed for all three replicates; however, there were some exceptions. For example, at 4°C, without nisin and salt two out of three responses for the strain 120 were 1 (growth) and the one was 0 (no growth) meaning that growth probability was 67% (Figure 3b). For the strain 137, one out of three responses was 1 (growth) and two were 0 (no growth) indicating the probability of growth was 33% at 4°C without nisin and with 3% salt (Figure 3d). Differences between the replicates could also be seen in Figures 4b, 4d and 4f. It was observed that the temperature had a significant impact on the growth of L. monocytogenes: at 37°C, the use of high salt and nisin concentration was not enough to avoid the growth even at the lowest inoculum (10 3 CFU/mL) -results not shown. On the other hand, high nisin concentration (≥200 µg/mL) was very effective at 20°C in the absence or presence of salt when the initial inoculum     Table 7 was less than 10 6 CFU/mL for all strains in general. Refrigeration temperature (4°C) changed the results drastically: even at the lowest nisin (50 µg/mL) concentration, growth was not observed for the strains 120 and 137 (Figure 3b and Figure 3d) whereas no growth was observed for the ATCC7644 strain without nisin and salt during storage of 8 weeks (Figure 3f).

Growth boundary determination
Models proposed [Eq. (2) and (3)] were applied to the data, and coefficient estimates, their standard errors and P values for Eq. (2) and (3) are all given in Table 6 and Table 7 (3)] all main effects retained in the equation when the insignificant coefficients were removed from the model. All goodness-of-fit indices were better for the second model (Table 8 and Table 9) and therefore, Eq.(3) was further used to plot the predicted growth/no growth boundaries of L. monocytogenes. Boundaries for likely to grow probability (p= .9), equal probability of growth and no growth (p= .5), and unlikely to grow probability (p= .1) are also displayed in Figures 3 and 4.

Discussion
In this study, the prevalence of L. monocytogenes in RTE foods was as 12.8%. Previous studies reported that the percentage of positive samples of L. monocytogenes in RTE foods was 36.73% in Nigeria [33] and 33.3% in Italy. [34] On the other hand, our results showed higher prevalence of L. monocytogenes strains in RTE foods compared to some other studies conducted in different regions, such as 2.6% in Estonia, [35] 6.87% in China, [36] 1.7% in Japan, [37] 0.4% in USA, [38] 0.1% in Poland, [39] and 3.1% in Chile. [40] Previous studies in Turkey [3,41,42] documented that the prevalence of L. monocytogenes in RTE foods was in the range of 4.0-19.7%. Higher prevalence in seafood products (26.7%) than other RTE foods were observed and this result was even higher than the previous reports from Iran, [5] Greece, [43] Estonia, [44] and Nigeria. [45] The prevalence of L. monocytogenes RTE seafood products in Turkey was found as 12% by Sanlibaba et al. [3] The higher prevalence obtained in this study may indicate poor hygiene level by the food  10 0.020 0.004 <0.001HT handlers. Incidence of L. monocytogenes was slightly higher in meat-based samples (19.4%) than in poultry-based samples (16%). Available data from literature indicated that the prevalence of L. monocytogenes was between 3.6% and 27.9%. [22,39,46] Additionally, Awaisheh [22] and Sanlibaba et al. [3] found the prevalence of L. monocytogenes on RTE poultry foods at the level of 15% and 10.5%, respectively. The reason for the high prevalence of L. monocytogenes in the RTE meat and poultrybased foods may be inadequate heat treatment to eliminate of L. monocytogenes or inadequate physical separation between the raw and cooked food processing areas . [39] L. monocytogenes have been widely associated with vegetable products. [47] In this study, although L. monocytogenes was not detected in fruit-based RTE foods, the prevalence of this pathogen was 10% in vegetable-based RTE foods. Fruit-or vegetable-based RTE foods which may be considered as minimally or moderately processed foods, can be great health risk for consumers since they are consumed fresh without applying heat treatment. [48] The main reason of contamination of vegetablebased RTE foods by L. monocytogenes may be processing, packing, or poor sanitary conditions of food handlers. [8,9] It is worth noting that L. monocytogenes was not isolated from dairy-based RTE in this study. This pathogen can be controlled in dairy-based RTE by the appropriate implementation of good manufacturing practices and hazard analysis and critical control point systems. Many studies have published that the prevalence in dairy-based RTE foods ranged from 5.90 to 36%. [21,49,50] Moreover, consumption of soft cheeses named Queso Fresco caused Listeria outbreak in 2021 according to the report published by Centers for Disease Control (CDC), and 12 persons became ill, of whom 1 died. [51] Post-pasteurization contamination by L. monocytogenes is the main problem in dairy products . [2] Our results revealed that the specific growth rates of L. monocytogenes strains isolated from RTE foods obtained by the absorbance method and the plate count method were similar (Table 3 and  Table 4) which had been also observed in different studies. [52][53][54] Therefore, it may be possible to determine the specific growth rate by performing two instead of multiple plate count experiments with 2-fold dilution method [54] since the initial value of each strain was also determined by enumeration. Strain 120 was the most resistant one to nisin according to MIC results (Table 2), it had also the highest growth rate with respect to 2-fold dilution method (Table 3) and second highest growth rate according to colony count data (Table 4). Hence, it was included in the growth/no growth study. On the other hand, although strain 137 had middle susceptibility to nisin ( Table 2) it had the second highest growth rate with respect to 2-fold dilution method (Table 3) and the highest growth rate with respect to colony count data (Table 4). Therefore, it was also included in growth boundary study together with ATCC7644 strain.
Growth response data (1: growth, 0: no growth) of three strains of L. monocytogenes were monitored at different temperature, salt, nisin, and inoculum concentrations for 8 weeks (56 days), and no increase in OD values of negative controls during this period was observed indicating that no contaimination was occurred during the preparation of the liquid media. On the other hand, OD values of the positive control (without salt and nisin) increased at each temperature as expected. Growth/no growth results also indicated that the nisin was more effective than NaCl to avoid growth. Moreover, protective effect of salt on nisin activity was also observed. For example, at 20°C with 2% and 3% of salt concentrations, and with 200 µg/mL of nisin growth were observed for the strain 120 but, the responses were no growth at the same temperature and nisin concentration with 0%, 1%, and 4% salt concentrations (Figure 3a). Protective effect of salt on nisin activity against L. monocytogenes was also observed by Boziaris et al. [55] However, contradictory results on this subject were also found in the literature. For example, it was observed that the presence of salt (NaCl) improves the action of nisin against Listeria. [56,57] On the other hand, the use of 2-4% NaCl has a protective effect of nisin against L. monocytogenes. [58,59] Form a hurdle concept point of view, this phenomenon should be further investigated at a microscopic level. Initial inoculum was also an important parameter. Comparison of Figures 3 and 4 revealed this fact. Probability of growth increased as the initial inoculum increased and this was in agreement with the work of Koutsoumanis and Sofos [60] who studied the combined effect of temperature, pH, water activity, and inoculum size of L. monocytogenes. Furthermore, suppressing the growth by using temperature, NaCl, and nisin was more difficult in RTE-isolates strains (120 and 137) than that of ATCC7644 strain (Fig. 3 and 4).
While the simple model [Eq. (2)] had only five coefficients (with an insignificant main effect), the complex model [Eq.(2)] had six coefficients for the strains 120 and 137, and 8 coefficients for ATCC7644 strain after the removal of the insignificant terms and repetition of the regression ( Table 6 and Table 7). Therefore, the second model was preferable over the first one and support for this statement also comes from Table 8 and Table 9 which compared the goodness-of-fit of the models. Furthermore, HL statistics were poor for the first model with P < .05 for the strains 137 and ATCC7644. The model is rejected if P is below 0.05. [31] The concordance and maximum rescaled R 2 values were all >91% and >0.83, respectively, indicating good predictive power of Eq.(3). Growth/no growth boundaries were in general consistent with the experimental data ( Fig. 3 and 4) and it is possible to estimate the growth probability of Listeria using Eq.(3) for any combination of temperature, salt, and nisin levels within the experimental range.
The strains 120 and 137 were isolated from RTE foods, however, this study was conducted in laboratory medium (broth). Use of broth can present a quick estimation of growth boundaries nevertheless, deviations from the results could be obtained in the laboratory medium compared to a real food system. [30] Conclusion Prevalence of L. monocytogenes in RTE foods sold in Turkish market was determined in this study. It was observed that some RTE foods (dairy and fruit-based) did not contain any Listeria. On the other hand, seafood-based, meat-based, poultry-based, and vegetable-based RTE foods had L. monocytogenes showing a possible cross-contamination or poor hygiene during preparation. Growth/no growth study with different levels of temperature, NaCl, nisin, and inoculum size revealed that avoiding the growth of RTE-isolated strains were more difficult than ATCC7644 strain. Furthermore, nisin was more effective than NaCl for suppressing the growth of L. monocytogenes and protective effect of NaCl on nisin activity was observed. This issue should be taken into consideration by the researchers and food processors before the simultaneous use of salt and nisin as chemicals in RTE foods. Probabilistic model used in this study showed high concordance and it was possible to determine the growth limits of L. monocytogenes accurately.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability statement
Data are available upon request from the corresponding authors.