Unit of Food Safety Research (SECALIM), National Institute of Agronomic Research – Veterinary School of Nantes, BP 40706, F-44307 Nantes Cedex 3, France. E-mail: firstname.lastname@example.org
Aims: To investigate potential resuscitation of Listeria monocytogenes and Salmonella Typhimurium after high hydrostatic pressure treatments.
Methods and Results: Pressure treatments were applied at room temperature for 10 min on bacterial suspensions in buffers at pH 7 and 5·6. Total bacterial inactivation (8 log10 CFU ml−1 of bacterial reduction) obtained by conventional plating was achieved regarding both micro-organisms. Treatments at 400 MPa in pH 5·6 and 600 MPa in pH 7 for L. monocytogenes and at 350 MPa in pH 5·6 and 400 MPa in pH 7 for S. Typhimurium were required respectively. A ‘direct viable count’ method detected some viable cells in the apparently totally inactivated population. Resuscitation was observed for the two micro-organisms during storage (at 4 and 20°C) after almost all treatments. In the S. Typhimurium population, 600 MPa, 10 min, was considered as the treatment achieving total destruction because no resuscitation was observed under these storage conditions.
Conclusions: We suggest a delay before performing counts in treated samples in order to avoid the under-evaluation of surviving cells.
Significance and Impact of the Study: The resuscitation of pathogen bacteria after physical treatments like high hydrostatic pressure has to be considered from the food safety point of view. Further studies should be performed in food products to study this resuscitation phenomenon.
For many years, conventional thermal treatment has constituted the method most widely employed to ensure the microbiological safety of foods. Recent trends in processing have aimed to produce more healthy, nutritious and convenient foods. High hydrostatic pressure treatment is currently considered as an attractive, nonthermal process to preserve foods (Knorr 1993). Previous studies showed that treatment at a pressure of 500–700 MPa readily kills vegetative bacterial, yeast and mould cells, while bacterial spores are more resistant (Hayakawa et al. 1994; Mackey et al. 1994; Wuytack et al. 1998). The resistance of micro-organisms to high pressure is variable and a reduction in microbial loads is directly related to the level of hydrostatic pressure applied (Arroyo et al. 1997). The industrial equipment used to preserve foods is routinely designed to generate at least 400 MPa of pressure every cycle (Zimmerman and Bergman 1993). For any preservation treatment, one of the primary considerations is its ability to eradicate pathogenic micro-organisms and thus ensure product safety. Unlike many other aspects of high-pressure research, relatively few studies have been carried out in this area. In microbiology, the effectiveness of preservative treatments is usually assessed by the difference between bacterial cell counts before and after treatment. This method has also been used to inactivate Listeria monocytogenes and Salmonella Typhimurium using high hydrostatic pressure (Winckel et al. 1997; Ritz et al. 1998). Raffali et al. (1994) mentioned the possibility of resuscitation after treatment, but some doubts still remained in their study because of the statistical significance of sampling method employed and the use of selective media. In addition, Ritz et al. (2001) demonstrated heterogeneity in the damage inflicted in a pressurized L. monocytogenes population and proposed that a small part of the bacterial population might be sublethally damaged by the pressure applied. The aim of the present study was to determine whether high hydrostatic pressure treatments displayed a real ability to eradicate these two pathogenic micro-organisms. This investigation involved staining procedures and storage tests to determine whether a degree of resuscitation was possible as a result of sublethal stress.
Materials and methods
Bacterial culture preparation
Listeria monocytogenes ScottA and S. Typhimurium ATCC 13311 were obtained from Institut Pasteur (Paris, France). Two successive cultures were performed in brain–heart infusion (BHI; Biokar, Beauvais, France) for 18 h at 37°C. Under these conditions, the initial cell count was around 8·2 log10 CFU ml−1, with SE of 0·2.
Preparation of samples
Pressurized samples were composed of 1 ml of subculture diluted in 9 ml of buffer. Two buffers were used: a phosphate buffer (pH 7) composed of Na2HPO4 (0·2 mol l−1) and NaH2PO4 (0·2 mol l−1) and a citrate buffer (pH 5·6) composed of Na2HPO4 (0·2 mol l−1) and C6H8O7 (0·1 mol l−1) (Merck, Darmstadt, Germany). The samples were placed in a sterile ministomacher bag (AES, Combourg, France). Experiments under each treatment condition were performed in triplicate.
The stomacher bags were sealed after the elimination of any air and placed in a hydrostatic pressure vessel (3 l) in a high-pressure apparatus (Alstom, Nantes, France). High-pressure levels were generated using water pressurized by a hydrostatic pump. The pressure levels required for total inactivation had been defined during previous studies (Ritz et al. 1998, 2000). Pressurized bacterial populations were defined as being totally inactivated when no culture was recorded in the whole sample analysed on plate count agar (PCA). The pressure treatments required to achieve the total inactivation of L. monocytogenes were 400 MPa in pH 5·6 and 600 MPa in pH 7, while for S. Typhimurium they were 350 MPa in pH 5·6 and 400 MPa in pH 7 respectively. These pressure levels were applied in this study. In order to minimize variations between samples, the rates of pressure increase and decrease were strictly equivalent. For all treatments, the kinetic parameters were 3 MPa s−1 for the pressurization rate and <1 s for depressurization. The temperature at which treatments were performed was fixed at 20°C. The temperature during treatment was checked and any variations (1·5°C, 100 MPa−1) were compensated for automatically during the holding time (10 min).
The heat treatment used as a standard to compare the viability of the high pressure method was 121°C for 15 min.
Count of surviving cells
The numbers of cells were determined before and after treatment. A sample of cells subjected to each treatment was counted 2 h after high-pressure treatment and then were stored at 4°C until analysis. Cells were counted by plating 0·1 ml volumes twice on PCA (Biokar), which were then incubated for at least 48 h at 37°C. When only a few cells had survived, the entire 10 ml of cell suspension was divided between two plates and covered with PCA. When no colony could be observed in the entire sample, it was considered that the population had been totally inactivated.
After high-pressure treatments, half of the pressurized samples were incubated at 4°C, while others were stored at 20°C. During this period, population cell counts were performed daily. When no resuscitation was observed after 7 days of storage at 4°C, the remaining samples were incubated at 20°C. The maximum storage period was 42 days.
Direct viable counts
Before and after treatments, cells were collected under aseptic conditions and viable cells counted using the direct viable count (DVC) method. To count the elongated cells of S. Typhimurium and L. monocytogenes, a modified version of the technique described by Kogure et al. (1979) was used. The antibiotics used were nalidixic acid for S. Typhimurium and ciprofloxacin for L. monocytogenes, the aim being to inhibit cell cleavage. One millilitre of the cell population was centrifuged at 12 000 g for 5 min at 4°C. The pellet was suspended in 9 ml of BHI (Merck) containing 5 μg ml−1 of nalidixic acid for S. Typhimurium and 1·2 μg ml−1 of ciprofloxacin for L. monocytogenes. After 6 h of incubation, the suspensions were diluted, filtered over a 0·2-μm pore size polycarbonate filter and stained for 5 min with 5 μg ml−1 of 4′,6-diamino-2-phénylindole (DAPI) solution (Interchim, Montluçon, France). Finally, the filter was air dried and placed on a slide in nonfluorescent immersion oil and a cover-slip was added before observation under an epifluorescence microscope (Olympus BX 40; Olympus, Rungis, France), equipped with a BW2-RFL-T3100-W light source. Counting was performed randomly and in triplicate on the basis of 20 microscope fields per filter. A typical count was between 10 and 100 cells per field. Only cells at least double the length of cells in fresh culture were counted as being viable.
All experiments were performed in triplicate. The data were the mean values of experimental findings. Statistical analysis was performed using a one-way analysis of variance according to the general linear model procedure, with least-square mean values effects to determine significant differences between treatments. The multiple range test was applied to determine which mean values were significantly different according to Fisher's ‘least significant differences’. Significant differences in inactivation were determined with 5% level of significance (P < 0·05) using Student's t-test. Statistical analysis was performed using Statgraphics plus, version 2.1 software (Statistical Graphics Corp., Princeton, NJ, USA).
The residual viability of pressurized bacterial populations was studied by the direct counting of viable cells. A DVC count was performed with respect to inactivated S. Typhimurium and L. monocytogenes populations under both pH conditions. The results obtained for cell populations suspended in both pH 7 or 5·6 buffers were quite similar [no survival observed from the plating count, whereas a high concentration (around 4 log10 ml−1) of viable cells was recorded by DVC counts]; for this reason, only the results obtained under pH 7 are presented herewith. Figures 1 and 2 show the comparison between total microscopic counts and elongated cell counts (viable cells) in untreated, pressurized and heat-treated (121°C, 15 min) populations. The total microscopic cell count remained almost constant (8 ± 0·2 log10 ml−1) in all experiments. The concentration of elongated cells in untreated populations was also constant, at around 7 log10 ml−1 (data not shown). No elongated cells (viable cell) were observed in heat-treated populations. Unlike the apparent inactivation of pressurized populations observed in different culture media, the results obtained using the DVC method demonstrated a concentration of elongated cells (considered as viable cells) just after pressure treatment that was equal to 4·12 ± 0·1 log10 ml−1 for S. Typhimurium and 4·15 ± 0·17log10 ml−1 for L. monocytogenes. In S. Typhimurium, concentration (400 MPa, 20°C, 10 min) was equal to 4·44 ± 0·34 log10 ml−1 after 1 day of storage and rose to reach 7·29 ± 0·3 log10 ml−1 after 7 days of storage. In the same pressurized population stored at 4°C, the elongated cell count remained statistically constant, at around 4·3 (±0·3) log10 ml−1 (data not shown).
Resuscitation capacity of S. Typhimurium cells after inactivation treatment
These results encouraged us to study the resuscitation capacities of treated cell populations. This approach was used to estimate the proportions of totally destroyed and damaged cells in the treated population. The results showed that S. Typhimurium populations (pressurized at 350 and 400 MPa and stored at 20°C after treatment), were capable of resuscitation (Fig. 3); this was observed only after 1 day of storage and reached 2·3 ± 0·11 and 4·3 ± 0·27 log10 CFU ml−1 with the treatments at 350 MPa in pH 5·6 and 400 MPa in pH 7 respectively. Further, the results showed that after this first step of resuscitation and during a 7-day storage period at 20°C, the population increased up to a stationary phase concentration. As for pressurized samples (350 and 400 MPa) stored at 4°C, resuscitation was observed after 2 days at low concentrations (<2 log10 CFU ml−1) in buffers at both pH values. This concentration remained statistically constant for c. 20 days, which was followed by a decline phase. In order to compare these results with standard bacterial behaviour, two populations (around 102 log10CFU ml−1 of Listeria and Salmonella) were subjected to the same conditions of buffer and temperature, but no decline phase was observed, even after 42 days at 4°C (data not shown). Whatever the pH of treatment and the storage temperature, no resuscitation was observed in the population treated at 600 MPa for 10 min.
Resuscitation capacity of L. monocytogenes cells after inactivation treatment
In the case of L. monocytogenes, treatments (at 400 MPa in pH 5·6 and 600 MPa in pH 7) achieved the maximum inactivation (Fig. 4). Storage at 20°C of the pressurized samples led to resuscitation in the two populations, which appeared after 4 days of storage. The bacterial concentrations of the stationary phase reached after resuscitation were statistically different (at a 95% confidence level) according to the pH of buffers 7·54 ± 0·22 log10 CFU ml−1 under pH 7 and 4·61 ± 0·38 log10 CFU ml−1 under pH 5·6 respectively. No resuscitation was observed in pressurized samples stored at 4°C, even after centrifugation and suspension of the bacterial population in BHI. However, growth of an unpressurized L. monocytogenes population was achieved under the same experimental conditions (data not shown).
Experimental simulation of a weak link in the chill chain showed that if after 7 days of storage at 4°C without resuscitation the pressurized samples were stored at 20°C, a resuscitation of populations was observed under all the treatment conditions. In the case of a weak link in the chill chain simulation, the final bacterial concentrations obtained after resuscitation were again statistically higher (at 95% confidence level) under pH 7 (7·2 ± 0·18) than under pH 5·6 (4·54 ± 0·34).
The effect of pressure on bacterial inactivation has been well documented (Arroyo et al. 1997; Simpson and Gilmour 1997). Studies on the pressure resistance of micro-organisms have usually demonstrated stronger resistance of Gram-positive than of Gram-negative bacteria (Shigehisa et al. 1991). The present experiments using L. monocytogenes and S. Typhimurium produced similar results, which also agreed with a previous review concerning the pressure resistance of various Salmonella and Listeria strains, which demonstrated bacterial reductions of more than 6 log10 CFU ml−1 (Cheftel 1995). However, during our study, the different pH of the buffers used could lead to a difference in treatment efficiency of up to 5 log10 CFU ml−1. These results were in line with the effects of low pH values on the pressure resistance of S. Typhimurium, described by Ritz et al. (1998), who reported a significant interaction between the pH of the buffer and the level of pressure treatment. In all the experiments reported, the number of surviving micro-organisms was determined by counting the bacterial suspension on media before and after treatments (Patterson et al. 1995; Simpson and Gilmour 1997). It is also widely recognized that classic culture techniques may underestimate the actual number of viable bacteria, especially when cells have been damaged by physical treatments. In addition, the physiological damage suffered by L. monocytogenes after high hydrostatic pressure has recently been documented by Ritz et al. (2001), who showed that cell damage is not homogenous, suggesting that resistant cells may exist in the pressurized bacterial population. On the other hand, Raffali et al. (1994) mentioned the possibility of the resuscitation of a Listeria innocua population inoculated in a dairy cream and treated with high hydrostatic pressure. This type of resuscitation could be compared with stress, because the experiment demonstrated a difference between the count on selective media (Oxford; Merck, Darmstadt, Germany) and in tryptone soy yeast extract (TSYE) broth. In addition, the count was performed by analysing 0·1 ml of dairy cream (in triplicate) taken from a total sample of 30 g dairy cream. Thus these resuscitation results may have been markedly influenced by the sampling procedure. All these questions help to explain the aim of the present study, which was to determine whether high hydrostatic pressure treatments could eradicate these two micro-organisms or if a degree of resuscitation might be induced by the existence of sublethal damage.
Few techniques are currently being developed in the field of viability assessment. The method described by Kogure et al. (1979) is useful in demonstrating that dormant cells can retain viability even though they cannot be cultured on conventional media. Using the DVC method, antibiotics are added to samples and after a period of incubation in fresh BHI, elongated cells can readily be observed and counted as viable cells (Roszak et al. 1984). Our results showed that even if the pressurized population appeared to be totally inactivated, a concentration of viable cells (c. 104 bacterial cells per ml) could be recorded just after pressure treatment. The same degree of heterogeneity had been already described in the case of cells in a viable but non-culturable (VBNC) state after nutrient limitation (Federighi et al. 1998; Besnard et al. 2002) or following disinfectant treatment (Desmonts et al. 1990) but this method had never been used to characterize a physical treatment.
As many previous results suggested that some less damaged cells might be present in a treated population, we investigated the resuscitation capacities of these pressurized populations. During these experiments, the entire sample (10 ml) was counted and resuscitation was studied with respect to cells which remained in the treatment buffer under all storage conditions. The results of this study showed that classic culture techniques performed just after treatments overestimated treatment efficacy. Indeed, the results showed that sublethal stress affecting the cells prevented the counting of viable cells in classic culture media for a period of 2–4 days. After that period (except in the case of S. Typhimurium treated at 600 MPa at both pH levels, data not shown), storage at 20°C gave rise to resuscitation in all pressurized populations. From a food safety point of view, these results thus reveal an under-evaluation of the residual viability of pressurized-pathogenic micro-organisms, so that a delay should be allowed after treatment in order to count the number of surviving cells. Our experiments also made it possible to show that some cells could recover their culturable state (after a period which depended on the strain under study). However, as described by Ravel et al. (1995), we cannot exclude the possibility that the counts observed were not solely because of resuscitation but also resulted from a combination of resuscitated cells and cell growth, once they had recovered their ability to divide. In the case of S. Typhimurium, subjected to high pressure treatment and stored at 4°C (a temperature where no growth has been recorded in untreated populations under the same conditions), one in 106 bacteria retained a capacity for resuscitation. Surprisingly, in L. monocytogenes, no resuscitation was observed at low temperatures, even though this bacteria is usually able to grow under these conditions (Tienungoon et al. 2000). These observations thus suggested that the cell targets and consequently the damage inflicted on Gram-positive and Gram-negative bacteria might differ. This idea had already been put forward by Shigehisa et al. (1991), who described a graduation of pressure sensitivity between Gram-positive and Gram-negative cells. Some authors, such as Jyothirmayi et al. (1998) mentioned that BHI should be the best method to achieve the recovery of growth capacity in stressed cells. However, during our experiments, this resuscitation procedure did not improve the resuscitation rate in the pressurized population.
The principal advantage of high hydrostatic pressure processing is that it preserves the initial food product (vitamins and flavour). However, with respect to bacterial destruction, the results obtained during this present study remind us that the possibility of bacterial resuscitation needs to be borne in mind. Further investigations will be of value to determine whether the results observed in these buffers could be extrapolated to food products. In this case, the possibility of alternative treatments to pasteurization should be considered and an efficient chill chain imposed after treatment.