Comparison of Desiccation Tolerance Among <I>Listeria</I>

104

Journal of Food Protection, Vol. 78, No. 1, 2015, Pages 104–110 doi:10.4315/0362-028X.JFP-14-249 Copyright G, International Association for Food Protection

Comparison of Desiccation Tolerance among Listeria monocytogenes, Escherichia coli O157:H7, Salmonella enterica, and Cronobacter sakazakii in Powdered Infant Formula

SHIGENOBU KOSEKI,1,2* NOBUTAKA NAKAMURA,2 AND TAKEO SHIINA2 Downloaded from http://meridian.allenpress.com/jfp/article-pdf/78/1/104/1688456/0362-028x_jfp-14-249.pdf by guest on 28 September 2021 1Research Faculty of Agriculture, Hokkaido University, Kita 9 Nishi 9, Kita-ku, Sapporo 060-8589, Japan; and 2National Food Research Institute, 2-1-12 Kannondai, Tsukuba, Ibaraki 305-8642, Japan

MS 14-249: Received 29 May 2014/Accepted 23 September 2014

ABSTRACT Bacterial pathogens such as Listeria monocytogenes, Escherichia coli O157:H7, Salmonella enterica, and Cronobacter sakazakii have demonstrated long-term survival in/on dry or low–water activity (aw) foods. However, there have been few comparative studies on the desiccation tolerance among these bacterial pathogens separately in a same food matrix. In the present study, the survival kinetics of the four bacterial pathogens separately inoculated onto powdered infant formula as a model low-aw food was compared during storage at 5, 22, and 35uC. No significant differences in the survival kinetics between E. coli O157:H7 and L. monocytogenes were observed. Salmonella showed significantly higher desiccation tolerance than these pathogens, and C. sakazakii demonstrated significantly higher desiccation tolerance than all other three bacteria studied. Thus, the desiccation tolerance was represented as C. sakazakii . Salmonella . E. coli O157:H7 ~ L. monocytogenes. The survival kinetics of each bacterium was mathematically analyzed, and the observed kinetics was successfully described using the Weibull model. To evaluate the variability of the inactivation kinetics of the tested bacterial pathogens, the Monte Carlo simulation was performed using assumed probability distribution of the estimated fitted parameters. The simulation results showed that the storage temperature significantly influenced survival of each bacterium under the dry environment, where the bacterial inactivation became faster with increasing storage temperature. Furthermore, the fitted rate and shape parameters of the Weibull model were successfully modelled as a function of temperature. The numerical simulation of the bacterial inactivation was realized using the functions of the parameters under arbitrary fluctuating temperature conditions.

Outbreaks of foodborne illness associated with the these bacterial pathogens under the same dry environments consumption of dry or low–water activity (aw) foods and and/or food matrix (17, 20). A quantitative and mathemat- food ingredients have been documented (4). A large ical analysis of the differences and/or similarities in the majority of these outbreaks were caused by Salmonella (4, desiccation tolerance among bacterial pathogens would play 34). Outbreaks of salmonellosis were caused by several an important role in a comparative quantitative microbial different Salmonella serotypes including Enteritidis, Typhi- risk assessment of dry or low-aw foods. murium, Saintpaul, etc. (7, 8, 13, 16, 18, 19, 22, 23, 38). The objective of the present study was to clarify the Other bacterial pathogens, including Cronobacter sakazakii differences and/or similarities in the desiccation tolerance (6, 15, 17) and verotoxigenic Escherichia coli, such as E. of C. sakazakii, Salmonella, E. coli O157:H7, and L. coli O157:H7 (27, 28, 30) and E. coli O104 on dry seeds of monocytogenes. We investigated the desiccation tolerance sprouts (2, 5), have also been associated with these of these four pathogenic bacteria separately during 1-year outbreaks. Listeriosis associated with the consumption of storage using powdered infant formula as a model dry low-aw food containing Listeria monocytogenes has not food matrix. To clarify the differences and/or similari- been documented. However, L. monocytogenes has been ties in the desiccation tolerance among these bacterial detected in several types of dry foods (19, 20). These pathogens, the effect of storage temperature on the bacterial pathogens can survive under dry or low-aw survival characteristics was also examined. In addition, environments for long periods of time in or on various the survival kinetics of each bacterium was mathemati- foods and food contact surfaces (4, 19, 20, 33). cally analyzed and modeled as a function of time and However, there have been few reports directly com- temperature. paring the characteristics of desiccation tolerance among MATERIALS AND METHODS

* Author for correspondence. Present address: Research Faculty of Agriculture, Bacterial strains. We used generally available bacterial Hokkaido University, Kita 9 Nishi 9, Kita-ku, Sapporo 060-8589, Japan. Tel strains from American Type Culture collection (ATCC) as follows. and Fax: z81-11-706-2552; E-mail: [email protected]. Six strains of L. monocytogenes (ATCC 19111, ATCC 19117, J. Food Prot., Vol. 78, No. 1 PATHOGEN SURVIVAL UNDER DRY ENVIRONMENT 105

ATCC 19118, ATCC 13932, ATCC 15313, and ATCC 35152), (Corning Incorporated Life Sciences, Tewksbury, MA) at 5, 22, six strains of E.coli O157:H7 (ATCC 35150, ATCC 43889, and 35uC for 12 months. ATCC 43895, ATCC 51657, ATCC 700378, and ATCC BAA- 460), and four strains of C. sakazakii (ATCC 12868, ATCC Sampling and enumeration of bacteria. The survival of the 29004, ATCC 29544, and ATCC 51329) were used. The S. microorganisms at 5, 22, and 35uC was determined using samples of enterica used in the present study represented a mixture of the contaminated formula after storing for 0.5, 1.5, 4, 6, 9, and Salmonella Enteritidis and Salmonella Typhimurium. Two strains 12 months. Duplicate 10-g samples were taken from the storage of Salmonella Enteritidis (ATCC BAA-708 and ATCC 4931) and bottle at each sampling time to confirm the internal uniformity. Each three strains of Salmonella Typhimurium (ATCC 29057, ATCC 10-g sample of powdered infant formula was combined with 90 ml 29629, and ATCC 29630) were used. All strains were maintained of 0.1% peptone water in a 400-ml stomacher bag (PYXON-20, at 285uC in brain heart infusion broth (Merck, Darmstadt, Elmex, Tokyo, Japan) and pummeled for 2 min using a stomacher- Germany) containing 10% glycerol. A sterile disposable plastic type blender (model CE-97, ILU Instrument, Barcelona, Spain). The loop was used to transfer the frozen bacterial cultures by undiluted sample solution (0.25 ml) was plated on the surface in scratching the surface of the frozen culture into 10 ml of brain quadruplicate, and the samples, serially diluted in 0.1% buffered heart infusion in a glass tube. The cultures were incubated without peptone water (0.1 ml), was plated in duplicate onto tryptic soy agar Downloaded from http://meridian.allenpress.com/jfp/article-pdf/78/1/104/1688456/0362-028x_jfp-14-249.pdf by guest on 28 September 2021 agitation at 35uC for 24 h, and transferred using loop inocula at (TSA) plates (Merck). The plates were subsequently incubated at three successive 24-h intervals to obtain a more homogeneous and 37uC for 48 h, and the number of colonies was enumerated. Thus, stable cell population. Grown cells were collected by centrifuga- the detection limit of the bacteria was 10 CFU/g (1 log CFU/g). tion (3000 | g, 15 min at 20uC), and the resulting pellet was Typical three colonies on each plate were picked up and then washed by 0.1% peptone water twice and subsequently confirmed the presence of each pathogen. The presence of E. coli resuspended in 5 ml of 0.1% peptone water. To generate a single O157:H7 was confirmed using a latex agglutination test (E.coli sample of each pathogen comprising every strain, equal volumes O157 Singlepath, Merck). The presence of Salmonella was of the cell suspensions from four, five, or six strains of each confirmed using a latex agglutination test (Salmonella Singlepath, pathogen were combined to achieve approximately equal Merck). The presence of L. monocytogenes was confirmed using the populations of each strain. Singlepath Listeria diagnosis kit (Merck). The presence of C. sakazakii was confirmed after streaking the colony onto Chromo- Inoculation of powdered infant formula. Commercial Cult Enterobacter sakazakii agar (Merck) plates, and the typical powdered infant formula (Sukoyaka; Bean Stalk Snow Co., green colonies forming on the plates were observed. Ltd., Sapporo, Japan) was used as a model food substrate in this To confirm the presence of pathogens that did not occur at study. The powdered infant formula contained proteins (12.3%, detectable levels in the plate count experiments, the stomaching wt/wt), fats (27.8%, wt/wt), carbohydrates (54.9%, wt/wt), mixture was incubated at 37uC for 24 h. The incubated mixture several vitamins (A, B1, B2, B6, B12, C, D, E, and K), and was streaked onto TSA plates and inoculated into tryptic soy broth mineral salts (Ca, Fe, K, Mg, Mg, Cu, and Zn). One kilogram of (5 ml), followed by additional incubation at 37uC for 24 h. The powdered infant formula was placed onto a sterile stainless steel presence or absence of colonies on each plate and/or presence or tray (60 by 30 by 10 cm) under a biological safety hood. A absence of the changes in optical density at 595 nm (.0.05) was 9 cocktail of approximately 3.5 | 10 CFU of concentrated culture evaluated as positive or negative. of each bacterium in 1.0 ml of 0.1% peptone was added by 100 10-ml droplets to the dry infant formula and then thoroughly Data analysis. Triplicate independent experiments were mixed manually. The inoculated infant formula was placed into a conducted to examine variability between the trials. As previously sterile plastic bag and vigorously shaken for 5 min. The mixed described, duplicate 10-g samples for each sampling time were powdered infant formula was again placed onto the sterile taken for confirming the internal uniformity. The colony-counting stainless steel tray and maintained in a biological safety hood (40 data of duplicate samples for each sampling interval in each to 50% relative humidity) for 12 h to ensure dryness (14, 15). The bacterium were transformed to log CFU per gram and the duplicate procedure described above was separately conducted with each samples were averaged to represent the powdered infant formula bacterium and replicated three times for each bacterium to data in each sampling time. Consequently, triplicate data in each confirm reproducibility of the experiment. sampling interval, for each bacterium, at each storage temperature There was no microbial contamination in the tested infant were obtained. The data set of each replicate was independently formula prior to pathogen inoculation and throughout the storage analyzed for further statistical modeling. period, as confirmed in the negative control samples (without The Weibull model, one of the most frequently used survival inoculation). Although theoretical final concentration of the models (29, 31), was used as a representative model. This model is 6 inoculated sample was expected around 3.5 | 10 CFU/g, the appropriately represented using the following formula: resulting final concentration of bacterial cells in the dry powder NtðÞ ranged between 1.0 | 105 and 2.5 | 105 CFU/g. This might be log StðÞ~ log ~{btn ð1Þ due to either a loss of cells during preparation and/or an initial N0 reduction during preparation. The aw of the contaminated powder where S(t) is the momentary (instantaneous) survival ratio, i.e., (ca. 5 g, due to size of the sample holder) stored at each N(t)/N0, and N(t) and N0 are the momentary and initial counts, temperature was determined once at each sampling interval for respectively. The initial counts (N0) were assumed to be the state each bacterium using a water activity meter (Novasina aw after drying the samples under a biological safety hood for 12 h. CENTER; Novasina, Pfa¨ffikon, Switzerland). The aw of the The b and n are parameters representing survival rate and curvature powdered infant formula, measured immediately after inoculation of kinetics, respectively. The survival ratio data for each with the bacterial suspension, remained below 0.28. Each of the temperature condition were fitted with the Weibull function 1-kg powder samples, separately contaminated with one described by equation 1 using nonlinear least-squares regression pathogenic bacterium, was divided into three 330-g samples with R statistical software (Version 2.15.0 for Mac OS X; http:// and stored in tightly closed 1,000-ml sterile Pyrex bottles www.r-project.org). 106 KOSEKI ET AL. J. Food Prot., Vol. 78, No. 1

FIGURE 1. Survival kinetics of L. monocytogenes (A), E. coli O157:H7 (B), Salmonella (C), and C. sakazakii (D) at 5uC(N), 22uC(n), and 35uC(&) during a 1-year period. The curves were represented by the best-fitted Weibull model. The results are presented as the means ¡ SD (n ~ 3). The values with data points below detection limit (,1 log CFU/g) represent the ratio of positive samples versus tested samples by enrichment. The shaded regions represent 95% prediction intervals obtained from the Monte Carlo simulation. Downloaded from http://meridian.allenpress.com/jfp/article-pdf/78/1/104/1688456/0362-028x_jfp-14-249.pdf by guest on 28 September 2021

The estimated parameters b and n of each experimental the momentary temperature, at a time that corresponds to the condition in each bacterium were compared with each other by momentary survival ratio. Thus, when the inactivation pattern using Tukey’s honest significant difference method to determine follows equation 1 as a model, the momentary isotemperature statistically significant differences (P # 0.05) among the bacteria survival rate, d log S(t)/dt, at a given temperature is types. All statistical procedures were conducted by using R d log S(t) n{1 statistical software. ~{bTðÞ|n|t ð4Þ dt Monte Carlo Simulation for the inactivation kinetics. To According to equation (1), the time t*, corresponding to the estimate the variability of the pathogen inactivation kinetics, the momentary logarithmic survival ratio log S(t)is Monte Carlo simulation was performed with 10,000 iterations 1 log StðÞ n using @Risk version 5.7 (Palisade Corp., Ithaca, NY). The rate t~ ð5Þ bTðÞ parameter b and shape parameter n were assumed to follow a normal distribution of the mean value and standard deviation (SD) Combining equations 4 and 5, using and the temperature as a derived from replicate experiments of each condition as follows: function of time, i.e., temperature ~ T(t) yielded the following the survival rate equation: bii&Normal½ EðÞbii ,SDðÞbii ð2Þ n{1 d log S(t) log S(t) n n &Normal½EnðÞ, SDðÞ n ð3Þ ~{bT½(t) |n|{ ð6Þ ii ii ii dt b(T) where i and j represent the bacteria type (C. sakazakii, S. enterica, This model is an ordinary differential equation that can be solved E. coli O157:H7, and L. monocytogenes) and temperature (5, 22, numerically by the fourth-order Runge-Kutta method using R 35uC), respectively. Equation 2 represents a normal distribution software with a ‘‘deSolve’’ package. with expected value E(bij) and SD(bij) values obtained from the replicated experiments. Likewise, equation 3 means a normal RESULTS distribution with expected values E(nij) and SD(nij) obtained from the replicated experiments. We simulated the inactivation kinetics There was no significant change in the aw of the described by Weibull model (equation 1) for each bacterium using samples throughout the storage period regardless of the the distribution of the parameters obtained from the simulation storage temperature and kind of bacteria (0.28 ¡ 0.03, n ~ mentioned above. 216, comprising four pathogens, three temperatures, six sampling intervals, and triplicate trials). The inactivation Numerical simulation of bacterial inactivation during storage. The temperature (T) dependency of ‘‘rate parameter’’ b kinetics in all conditions was apparently nonlinear with and ‘‘shape parameter’’ n were described using fitted exponential tailing (Fig. 1). The inactivation kinetics of the tested type functions b(T) and n(T), respectively (25, 26, 31). bacteria was satisfactory and accurately described using the We assume that under fluctuating temperature conditions, Weibull model as shown in the Table 1. The decreases of the momentary inactivation rate is the rate that corresponds to the viable bacterial numbers were more rapid at 35uC than at J. Food Prot., Vol. 78, No. 1 PATHOGEN SURVIVAL UNDER DRY ENVIRONMENT 107

22 and 5uC. However, survivors remained during storage at 5uC for 1 year, regardless of the kind of bacteria. Comparing 0.01 0.02 0.01 0.02 C

u the survival kinetics among the tested bacteria, C. sakazakii ¡ ¡ ¡ ¡ showed a significantly (P # 0.05) lower survival rate parameter (b) than those of the other three bacteria, except at 35uC storage. This result indicates that C. sakazakii is highly resistant to desiccation stress compared with the other pathogenic bacteria tested in the present study. 0.01 0.97 0.02 0.96 0.02 0.97 0.02 0.96 2 C35 u The Monte Carlo simulation was conducted using the ¡ ¡ ¡ ¡ estimated mean parameter values and SDs for each bacterium

SD of triplicate experiments. and temperature combination. The 2.5th and 97.5th percen-

¡ tiles obtained by simulation are shown in Figure 1 as shaded regions. Considering the variability of the inactivation Downloaded from http://meridian.allenpress.com/jfp/article-pdf/78/1/104/1688456/0362-028x_jfp-14-249.pdf by guest on 28 September 2021 0.01 0.98 0.01 0.95 0.02 0.96 0.02 0.93 kinetics, the effect of temperature on bacterial inactivation C22 u ¡ ¡ ¡ ¡ was apparent, regardless of the bacteria type. The relationship between the estimates of the parameters (b and n) and the temperature is shown in Figure 2. While the rate parameter b was described as an exponential function of temperature (Table 2), the shape parameter n did 0.05 0.99 0.04 0.97 0.03 0.98 0.05 0.94 C5 u not show any common trend among the bacteria. If we ¡ ¡ ¡ ¡ describe the temperature dependency of the parameter n, a a logarithm function could be fitted with relatively high

The values presented here are mean accuracy for each bacterium except for the case of E. coli n. O157:H7 which showed constant values (n ~ 0.165) 0.04 0.28 0.04 0.20 0.04 0.17 0.06 0.25 and

C35 regardless of the temperature (Table 2). u b ¡ ¡ ¡ ¡ Using the obtained functions and/or values for these parameters, numerical simulation was conducted for arbitrary

) and goodness-of-fit fluctuating temperature conditions. The simulation was con- n ducted for E. coli O157:H7 as a model case, as the shape and parameter n of E. coli O157:H7 showed small variations b 0.01 0.28 0.04 0.18 0.03 0.16 0.04 0.26 C22

u representing almost constant values regardless of the tempera- ¡ ¡ ¡ ¡ ture conditions. We used the rate parameter b(T) function for E. coli O157:H7 and set the shape parameter n as a constant value of 0.165 for the numerical simulation (Table 2). As shown in Figure 3, the numerical simulation was conducted using the 0.26 0.03 0.26 0.28 0.23 0.16 0.22 0.06 function of the parameters under an arbitrary temperature history C5 u ¡ ¡ ¡ ¡ as an example of warehouse for a commercial product.

) fitted survival parameters ( DISCUSSION n bt In the present study, we examined the desiccation {

~ tolerance for four types of pathogenic bacteria in powdered 0.12 2.81 0.18 3.34 0.18 3.63 Þ 0.16 2.41 0 C35

u infant formula. Notably, in this study, we compared the N ¡ ¡ ¡ ¡ bnR =

N desiccation tolerance among the pathogenic bacteria in the ð same food matrix. The results suggest that there are log

n differences and/or similarities in the desiccation tolerance bt among the pathogenic bacteria. C. sakazakii showed the {

~ highest tolerance for desiccation, regardless of the storage 0.02 1.71 0.09 2.11 0.07 2.43 0.05 1.29 Þ 0 C22

u temperature (Fig. 1). Salmonella showed significantly ¡ ¡ ¡ ¡ N 5 = higher resistance to desiccation than E. coli O157:H7 and N

ð 1.38 1.08 0.72 L. monocytogenes (P # 0.05) in terms of the inactivation rate parameter. Although these results are consistent with previous studies, there are only a few studies on the quantitative direct comparison of desiccation tolerance among pathogenic bacteria (17, 20). Weibullian (log Several previous studies have examined the survival of

O157:H7 1.35 C. sakazakii in infant formula (3, 10, 13). Although Kandhai et al. (17) conducted a quantitative analysis of the survival The data set of each replicate was independently fitted to the Weibull model and estimated the parameters TABLE 1. Salmonella a L. monocytogenes E. coli C. sakazakii of C. sakazakii, the estimated rate parameter values obtained 108 KOSEKI ET AL. J. Food Prot., Vol. 78, No. 1

kinetics was represented by the most tolerant strain. Thus, it was expected that one or more of the other three strains would have higher desiccation tolerance. The present study did not evaluate the bacterial survival using D-value. The inactivation kinetics of the bacterial pathogens examined in the present study showed nonlinear kinetics on a semilog plot. Because D-values only work appropriately on log-linear inactivation kinetics, the D- values calculated from nonlinear kinetics would lead to the over or under estimation of the inactivation effect (29–31). Although as an alternative indicator, the first 1-log reduction has been proposed (24), this idea is just an attempt to simplify a nonlinear kinetics. Thus, to respect the whole Downloaded from http://meridian.allenpress.com/jfp/article-pdf/78/1/104/1688456/0362-028x_jfp-14-249.pdf by guest on 28 September 2021 inactivation kinetics, both the rate and shape parameters of the Weibull model were evaluated in the present study. Three representative pathogenic bacteria (E. coli O157:H7, Salmonella, and L. monocytogenes) showed survival under low aw conditions, such as aw ~ 0.28, for long periods of time under chilled to room temperatures. While the low-aw food did not show the risk of pathogenic bacterial growth, the low-aw food might be a vehicle for pathogenic bacteria due to its desiccation tolerance. Indeed, there have been documented the outbreaks of foodborne diseases associated with consumption of low-aw foods (4, 5, 33). It is extremely difficult to efficiently inactivate or sterilize the pathogenic bacteria in a dried low-aw food, due to technical limitations (1, 21, 25, 26, 35). Physical inactivation techniques, such as heating and/or pressurizing, show a slight effect on dry or low-aw matrices, as the existence of free water is a key factor for microbial inactivation. Although gaseous atmosphere such as ozone, chlorine dioxide, and propylene oxide were reported its inactivation effect, the bacterial inactivation effect under dry condition was limited (36). Once a dry or low-aw food is contaminated with pathogenic bacteria, it is difficult to sufficiently reduce the number of those bacteria in the food. Indeed, irradiation is effective in microbial inactivation in and on low-aw foods. However, a special facility is needed for irradiation and the treatment cost must be higher than FIGURE 2. Temperature dependency of the Weibullian survival those of conventional procedure. It might not be easy to parameters b (A) and n (B) of L. monocytogenes (%), E. coli adopt irradiation technique in low-a foods processing. O157:H7 (e), Salmonella (n), and C. sakazakii (#), respectively. w The results are presented as the means ¡ SD of triplicate trials. Thus, care should be taken for the handling of low-aw foods Values with different letters within each temperature represent to avoid cross contamination. significant difference (P # 0.05). The variability of the bacterial inactivation kinetics was represented as 95% prediction intervals (Fig. 1) using the using the Weibull model were slightly faster than those of Monte Carlo simulation based on the probabilistic distribu- the present study. These differences might reflect the tion of the parameters. The effect of storage temperature on sampling intervals. Kandhai et al. (17) measured samples the survival of each bacterium was significant in the tested in short intervals, such as 2 weeks, and the observation conditions, even if the variability of bacterial inactivation duration was shorter (150 days) than that of the present was considered. There were few overlaps of the 95% study (365 days); thus, the inactivation curve might be prediction interval region among the temperatures, suggest- different from the results of the present study, which used ing that the storage temperature significantly influenced longer sampling intervals and a longer observation period. bacterial survival behavior in low-aw foods. It has been Furthermore, the difference between the strains used might suggested that low temperature conditions in which the cell also influence the estimated parameters. Indeed, in the activity is suppressed might be suitable for pathogenic present study, we used a cocktail of four different strains of bacteria under desiccation stress (3, 17). Adaptation to C. sakazakii, which included the strain ATCC 29544 as desiccation stress might stimulate the potential of survival reported by Kandhai et al. (17). Due to existence of one under more stressful conditions, such as suboptimal tolerant strain more than others in a cocktail, the survival temperature, as a survival strategy (11, 32, 37). J. Food Prot., Vol. 78, No. 1 PATHOGEN SURVIVAL UNDER DRY ENVIRONMENT 109

TABLE 2. Relationship between temperature and the Weibullian-fitted parameters Rate parameter b Shape parameter n Pathogen (5uC , Ta , 35uC) (5uC , T , 35uC)

Salmonella b ~ 1:164 expðÞ 0:023 | T n ~ 0:186 z 0:321 | log(T) R2 ~ 0.92 R2 ~ 0.94 L. monocytogenes b ~ 0:903 expðÞ 0:038 | T n ~ 0:352 0:111 | log(T) R2 ~ 0.99 R2 ~ 0.86 E. coli O157:H7 b ~ 1.555 exp(0.033 | T) n ~ 0:165 + 0:005 R2 ~ 0.99 C. sakazakii b ~ 0.569 exp(0.040 | T) n ~ 20.107 z 0.250 | log(T) R2 ~ 0.99 R2 ~ 0.94 a T, temperature. Downloaded from http://meridian.allenpress.com/jfp/article-pdf/78/1/104/1688456/0362-028x_jfp-14-249.pdf by guest on 28 September 2021

The estimated parameters of the Weibull model and the characterize the relationship between desiccation stress and functions derived from the parameters facilitated the the occurrence of viable but nonculturable bacteria in future simulation of the pathogen survival in powdered infant studies. formula under a given temperature history. For several In conclusion, we quantitatively clarified the inactiva- organisms, the power n(T) in equation 1 was practically tion kinetics of four different pathogenic bacteria under constant or could be assigned a fixed numerical value with temperatures ranging from 5 to 35uC during 1 year. The only a minor effect on the model’s fit (9, 24, 30). We will desiccation tolerance was represented as C. sakazakii . assume that this condition is true and for the following Salmonella . E. coli O157:H7 ~ L. monocytogenes. The analyses, we used the model with n(T) ~ n. Although it is relationship between the estimated parameters of the necessary to validate the model accuracy under various Weibull model and the temperature indicate the potential temperatures, the approach and calculation procedure prediction of the changes in the number of pathogenic proposed in the present study would contribute to an bacteria under arbitrary temperate conditions. exposure assessment in a quantitative microbial risk assessment to estimate the changes in the number of ACKNOWLEDGMENT bacteria over time. Thus, the validation study should be This work was supported through a grant from the Information Center conducted in future studies. of Particle Technology, Japan (2011–2012). The results presented in the present study show the viable bacterial counts on nonselective agar plates (TSA), REFERENCES suggesting that the cells had potential for colony forming on 1. Archer, J., E. T. Jervis, J. Bird, and J. E. Gaze. 1998. Heat resistance agar plates. However, the decreasing number of viable cells of Salmonella weltevreden in low-moisture environments. J. Food might not reflect the number of actual live cells. The state of Prot. 61:969–973. 2. Aurass, P., R. Prager, and A. Flieger. 2011. EHEC/EAEC O104:H4 the bacterial cells under desiccation stress might actually strain linked with the 2011 German outbreak of haemolytic uremic reflect viable but nonculturable bacteria strains (2, 23, 33, syndrome enters into the viable but non-culturable state in response to 34). Although the present study used nonselective medium various stresses and resuscitates upon stress relief. Environ. (TSA) to determine the bacterial number, which possibly Microbiol. 13:3139–3148. took into account the injured cells recovery, we should 3. Beuchat, L. R., H. Kim, J. B. Gurtler, L.-C. Lin, J.-H. Ryu, and G. M. Richards. 2009. Cronobacter sakazakii in foods and factors affecting its survival, growth, and inactivation. Int. J. Food Microbiol. 136: 204–213. 4. Beuchat, L. R., E. Komitopoulou, H. Beckers, R. P. Betts, F. Bourdichon, S. Fanning, H. M. Joosten, and B. H. Ter Kuile. 2013. Low-water activity foods: increased concern as vehicles of foodborne pathogens. J. Food Prot. 76:150–172. 5. Beutin, L., and A. Martin. 2012. Outbreak of shiga toxin–producing Escherichia coli (STEC) O104:H4 infection in Germany causes a paradigm shift with regard to human pathogenicity of STEC strains. J. Food Prot. 75:408–418. 6. Bowen, A. B., and C. R. Braden. 2006. Invasive Enterobacter sakazakii disease in infants. Emerg. Infect. Dis. 12:1185–1189. 7. Centers for Disease Control and Prevention. 2007. Multistate outbreak of Salmonella serotype Tennessee infections associated with peanut butter—United States, 2006–2007. Morb. Mortal. Wkly. Rep. 56:521–524. FIGURE 3. An example of the predictive simulation of E. coli 8. Centers for Disease Control and Prevention. 2009. Multistate O157:H7 survival during storage under fluctuating temperature outbreak of Salmonella infections associated with peanut butter and conditions using a combination of equation 6 and the functions of peanut butter–containing products—United States, 2008–2009. parameters in Table 2. Morb. Mortal. Wkly. Rep. 58:85–90. 110 KOSEKI ET AL. J. Food Prot., Vol. 78, No. 1

9. Corradini, M. G., and M. Peleg. 2004. Demonstration of the 24. Mafart, P., O. Couvert, S. Gaillard, and I. Leguerinel. 2002. On applicability of the Weibull-log-logistic survival model to the calculating sterility in thermal preservation methods: application of isothermal and nonisothermal inactivation of Escherichia coli K-12 the Weibull frequency distribution model. Int. J. Food Microbiol. 72: MG1655. J. Food Prot. 67:2617–2621. 107–113. 10. Edelson-Mammel, S. G., M. K. Porteous, and R. L. Buchanan. 2005. 25. Mattick, K. L., F. Jorgensen, P. Wang, J. Pound, M. Vandeven, L. Survival of Enterobacter sakazakii in a dehydrated powdered infant Ward, J. Legan, H. Lappin-Scott, and T. Humphrey. 2001. Effect of formula. J. Food Prot. 68:1900–1902. challenge temperature and solute type on heat tolerance of Salmonella 11. Elmnasser, N., M. Ritz-Bricaud, S. Guillou, F. Leroi, N. Orange, A. serovars at low water activity. Appl. Environ. Microbiol. 67:4128–4136. Bakhrouf, and M. Federighi. 2006. Adaptative response of Listeria 26. Mattick, K. L., J. D. Legan, T. J. Humphrey, and M. Peleg. 2001. monocytogenes to osmotic and chill stress: consequences on food Calculating Salmonella inactivation in nonisothermal heat treatments safety. Rev. Med. Vet. Toulouse 157:92–101. from isothermal nonlinear survival curves. J. Food Prot. 64:606–613. 12. Gill, O. N., P. N. Sockett, C. L. Bartlett, M. S. Vaile, B. Rowe, R. J. 27. Miller, B. D., C. E. Rigdon, J. Ball, J. M. Rounds, R. F. Klos, B. M. Gilbert, C. Dulake, H. C. Murrell, and S. Salmaso. 1983. Outbreak of Brennan, K. D. Arends, P. Kennelly, C. Hedberg, and K. E. Smith. Salmonella napoli infection caused by contaminated chocolate bars. 2012. Use of traceback methods to confirm the source of a multistate Lancet 321:574–577. Escherichia coli O157:H7 outbreak due to in-shell hazelnuts. J. Food

13. Gurtler, J. B., and L. R. Beuchat. 2007. Survival of Enterobacter Prot. 75:320–327. Downloaded from http://meridian.allenpress.com/jfp/article-pdf/78/1/104/1688456/0362-028x_jfp-14-249.pdf by guest on 28 September 2021 sakazakii in powdered infant formula as affected by composition, 28. Neil, K. P., G. Biggerstaff, J. K. MacDonald, E. Trees, C. Medus, water activity, and temperature. J. Food Prot. 70:1579–1586. K. A. Musser, S. G. Stroika, D. Zink, and M. J. Sotir. 2012. A novel 14. Gurtler, J. B., J. L. Kornacki, and L. R. Beuchat. 2005. Enterobacter vehicle for transmission of Escherichia coli O157:H7 to humans: sakazakii: a coliform of increased concern to infant health. Int. J. multistate outbreak of E. coli O157:H7 infections associated with Food Microbiol. 104:1–34. consumption of ready-to-bake commercial prepackaged cookie 15. Isaacs, S., J. Aramini, B. Ciebin, J. A. Farrar, R. Ahmed, D. dough–United States, 2009. Clin. Infect. Dis. 54:511–518. 29. Peleg, M., and M. D. Normand. 2004. Calculating microbial survival Middleton, A. U. Chandran, L. J. Harris, M. Howes, E. Chan, A. S. parameters and predicting survival curves from non-isothermal Pichette, K. Campbell, A. Gupta, L. Y. Lior, M. Pearce, C. Clark, F. inactivation data. Crit. Rev. Food Sci. 44:409–418. Rodgers, F. Jamieson, I. Brophy, A. Ellis, for the Salmonella 30. Peleg, M., and C. M. Penchina. 2000. Modeling microbial survival Enteritidis PT30 Outbreak Investigation Working Group. 2005. An during exposure to a lethal agent with varying intensity. Crit. Rev. international outbreak of Salmonellosis associated with raw almonds Food Sci. 40:159–172. contaminated with a rare phage type of Salmonella Enteritidis. J. 31. Peleg, M., C. M. Penchina, and M. B. Cole. 2001. Estimation of the Food Prot. 68:191–198. survival curve of Listeria monocytogenes during non-isothermal heat 16. Iversen, C., and S. Forsythe. 2003. Risk profile of Enterobacter treatments. Food Res. Int. 34:383–388. sakazakii, an emergent pathogen associated with infant milk formula. 32. Pin, C., T. Hansen, M. Mun˜oz-Cuevas, R. de Jonge, J. T. Trends Food Sci. Technol. 14:443–454. Rosenkrantz, C. Lo¨fstro¨m, H. Aarts, and J. E. Olsen. 2012. The 17. Kandhai, M. C., M. W. Reij, M. van Schothorst, L. G. M. Gorris, and transcriptional heat shock response of Salmonella Typhimurium M. H. Zwietering. 2010. Inactivation rates of Cronobacter spp. and shows hysteresis and heated cells show increased resistance to heat selected other bacterial strains in powdered infant formulae stored at and acid stress. PLoS One 7:e51196. different temperatures. J. Food Prot. 73:839–848. 33. Podolak, R., E. Enache, W. Stone, D. G. Black, and P. H. Elliott. 18. Kapperud, G., S. Gustavsen, I. Hellesnes, A. H. Hansen, J. Lassen, J. 2010. Sources and risk factors for contamination, survival, persis- Hirn, M. Jahkola, M. A. Montenegro, and R. Helmuth. 1990. tence, and heat resistance of Salmonella in low-moisture foods. J. Outbreak of Salmonella typhimurium infection traced to contaminat- Food Prot. 73:1919–1936. ed chocolate and caused by a strain lacking the 60-megadalton 34. Rowan, N. J. 2004. Viable but non-culturable forms of food and virulence plasmid. J. Clin. Microbiol. 28:2597–2601. waterborne bacteria: quo vadis? Trends Food Sci. Technol. 15:462– 19. Kenney, S. J., and L. R. Beuchat. 2004. Survival, growth, and thermal 467. resistance of Listeria monocytogenes in products containing peanut 35. Shachar, D., and S. Yaron. 2006. Heat tolerance of Salmonella and chocolate. J. Food Prot. 67:2205–2211. enterica serovars Agona, Enteritidis, and Typhimurium in peanut 20. Kimber, M. A., H. Kaur, L. Wang, M. D. Danyluk, and L. J. Harris. butter. J. Food Prot. 69:2687–2691. 2012. Survival of Salmonella, Escherichia coli O157:H7, and Listeria 36. Spotts Whitney, E. A., M. E. Beatty, T. H. Taylor, R. Weyant, J. monocytogenes on inoculated almonds and pistachios stored at 219, Sobel, M. J. Arduino, and D. A. Ashford. 2003. Inactivation of 4, and 24uC. J. Food Prot. 75:1394–1403. Bacillus anthracis Spores. Emerg. Infect. Dis. 9:623–627. 21. Koseki, S., and K. Yamamoto. 2007. Water activity of bacterial 37. van Bokhorst-van de Veen, H., T. Abee, M. Tempelaars, P. A. Bron, suspension media unable to account for the baroprotective effect of M. Kleerebezem, and M. L. Marco. 2011. Short- and long-term solute concentration on the inactivation of Listeria monocytogenes by adaptation to ethanol stress and its cross-protective consequences high hydrostatic pressure. Int. J. Food Microbiol. 115:43–47. in Lactobacillus plantarum. Appl. Environ. Microbiol. 77:5247– 22. Lehmacher, A., J. Bockemu¨hl, and S. Aleksic. 1995. Nationwide 5256. outbreak of human salmonellosis in Germany due to contaminated 38. Werber, D., J. Dreesman, F. Feil, U. van Treeck, G. Fell, S. paprika and paprika-powdered potato chips. Epidemiol. Infect. 115: Ethelberg, A. M. Hauri, P. Roggentin, R. Prager, I. S. T. Fisher, S. C. 501–511. Behnke, E. Bartelt, E. Weise, A. Ellis, A. Siitonen, Y. Andersson, H. 23. Lleo, M. M., B. Bonato, M. C. Tafi, C. Signoretto, M. Boaretti, and P. Tscha¨pe, M. H. Kramer, and A. Ammon. 2005. International outbreak Canepari. 2001. Resuscitation rate in different enterococcal species in of Salmonella Oranienburg due to German chocolate. BMC Infect. the viable but non-culturable state. J. Appl. Microbiol. 91:1095–1102. Dis. 5:7.