Aims: When subjected to dynamic temperatures surpassing the expected maximum growth temperature, Escherichia coli K12 MG1655 shows disturbed growth curves. These irregular population dynamics were explained by considering two subpopulations, i.e. a thermoresistant and a thermosensitive one (Van Derlinden et al. 2010a). In this paper, the influence of the initial cell concentration on the subpopulations’ dynamics is evaluated.
Methods and Results: Experiments were performed in a bioreactor with the temperature increasing from 42 to 65·2°C (1 and 4°C h−1) with varying initial cell concentrations [6, 12 and 18 ln(CFU ml−1)]. When started from the highest cell concentration, the population was characterized by a higher overall maximum growth temperature and a higher inactivation temperature. For all experimental set-ups, resistant cells were still growing at the final temperature of 65·2°C.
Conclusions: The initial cell concentration had no effect on temperature resistance. The increase in temperature resistance of the sensitive subpopulation was because of the change of the physiological state to the stationary phase.
Significance and Impact of the Study: A higher initial cell concentration leads to higher heat stress adaptation when cultures reach a maximum cell concentration. The observed growth at a temperature of 65·2°C is very important for food safety and the temperature treatment of micro-organisms.
The influence of temperature and temperature history on microbial behaviour has been the subject of several studies, with relevance to several disciplines. In industrial biotechnology, fermentations with thermoresistant micro-organisms performed at higher temperatures can result in higher reaction rates and better productivity (see, e.g. Banat et al. 1998; Ke et al. 2008). Acquired thermotolerance is of special interest for the food industry, as adapted pathogens and spoilage micro-organisms can resist the thermal processes presumed to be lethal (see, e.g. Tsuchido et al. 1982; Stephens et al. 1994; Juneja and Marks 2003; Manãs et al. 2003; Hassani et al. 2005, 2006; Cebrián et al. 2008 and Valdramidis et al. 2006). In environmental sciences, micro-organisms are also thought to be killed at higher temperatures during sanitizing in anaerobic digestion and composting (see, e.g. Albihn and Vinneras 2007; Lang and Smith 2008; Sahlstrom et al. 2008).
When micro-organisms are subjected to unfavourable conditions, this causes a reduced specific growth rate at mild stress levels or results in a reduced viability of the majority of the population when stress levels are severe (Nikolaev 2004). The microbial cell can also adapt to mild stress by altering its metabolism. As a consequence, cells are able to survive or even grow at conditions presumed to be lethal. For heat stress adaptation, this can imply that the cells obtain a higher maximum growth temperature Tmax (Van Derlinden et al. 2010a).
The individual cells of a microbial isogenic population also display phenotypic diversity to enhance their chances of surviving stress conditions (Sumner and Avery 2002; Aertsen and Michiels 2005). For this reason, a population can be divided in subpopulations depending on the degree in which they possess certain characteristics. In considering the temperature resistance of a population, Van Derlinden et al. (2009, 2010a) postulated that the population of Escherichia coli K12 consists of a temperature sensitive and a temperature-resistant subpopulation when grown in brain heart infusion broth (BHI) at super-optimal temperatures. This hypothesis was based on experimental results in both static and dynamic temperature conditions, where a sequence of growth, inactivation and re-growth was observed, with the growth and inactivation assigned to the sensitive subpopulation and the re-growth to the resistant subpopulation. Moreover, these researchers observed a heat stress adaptation of the resistant cell population, i.e. growth up to 60°C.
Van Derlinden et al. (2009) studied the influence of the initial cell concentration on the stress adaptation in static temperature conditions and noticed that, for high initial cell concentrations, the population heterogeneity is suppressed. This was observed by the growth curves being approximated by a smooth sigmoid pattern, in contrast to the disturbed growth curve obtained at low initial cell concentrations. Possible explanations were that, when starting at higher cell concentrations, (i) the concentration of extracellular induction component produced by the resistant subpopulation was sufficiently high to protect the sensitive subpopulation, (ii) the concentration of extracellular heat-shock proteins was higher and protected the sensitive subpopulation, or (iii) the cells were already close to the stationary growth phase, resulting in higher stress tolerance.
The objective of this paper was to further elucidate the conditions likely to lead to adaptive responses. The influence of the initial cell concentration on the heat stress adaptation of E. coli was evaluated in dynamic experiments using two different heating rates. These experiments could give an answer to the following question: at higher initial cell concentrations, is the sensitive subpopulation growing at higher temperatures in the dynamic temperature situation, as can be expected from results in the static experiments?
Materials and methods
A stationary culture of E. coli K12 MG1655 was stored at −80°C in BHI (Oxoid, Basingstoke, UK), supplemented with 25% glycerol (Acros Organics, Geel, Belgium).
The inoculum was prepared as described in Van Derlinden et al. (2009, 2010a,b).
Experiments were performed in a bioreactor (BioFlo 3000; New Brunswick Scientific, Edison, NJ, USA) filled with 3·5l BHI. The dynamic temperature profile, pH (7·55), agitation speed (400 rev min−1) and aeration rate (2 l min−1) were monitored and controlled as in Van Derlinden et al. (2010a). At regular time-points, a sample was taken and the cell density was measured by plate counting after serial dilutions [full details are given by Van Derlinden et al. (2010a)]. Some bioreactor samples were analysed for their glucose content with an enzymatic UV test using hexokinase (DiaSys Diagnostic Systems GmbH, Holzheim, Germany).
In a series of dynamic bioreactor experiments, different initial cell concentrations of E. coli were subjected to selected temperature profiles. A short constant phase at 42°C (0·8 h) was followed by a linear temperature increase to 65·2°C at a slope of 1 or 4°C h−1. When 65·2°C was reached, this temperature was maintained. For several of these experiments, replicates were performed.
The resulting cell concentration measurements of the dynamic experiments are depicted in Fig. 1. Depending on the initial cell concentration and the slope of the temperature profile, the evolution of the microbial population shows three, four or five subsequent phases, i.e. a first exponential growth phase, a first (pseudo-)stationary phase, an inactivation phase, a second growth phase and a second stationary phase. The first growth and stationary phase are highly reproducible, while the second growth and stationary phase vary significantly.
In the experiments at higher initial cell concentrations [12 and 18 ln(CFU ml−1) at 1°C h−1], cells grew until a maximum cell concentration Nmax was reached (see Fig. 1a). After a stationary phase, inactivation occurred. In this case, the inactivation started later than for the lower initial cell concentration of 6 ln(CFU ml−1). The inactivation phase is followed by a second stationary phase with a final cell concentration, which is found to be around 12–16 ln(CFU ml−1). When starting at a cell concentration of 6 ln(CFU ml−1), the inactivation phase is followed by a second growth phase before reaching the same final cell concentration around 12 and 16 ln(CFU ml−1).
Figure 1b shows, for experiments with lower initial cell concentrations [6 and 12 ln(CFU ml−1)] and a high heating rate (4°C h−1), that the cell concentration decreases immediately after reaching a maximum value. After inactivation, a second growth phase is visible. The cell count is still increasing at the final temperature of the experiment, i.e. 65·2°C. In the experiment with an initial cell concentration of 18 ln(CFU ml−1), after a first growth phase, the culture reaches a short stationary phase followed by inactivation.
When subjected to dynamic temperature changes near the maximum growth temperature (Tmax), E. coli K12 MG1655 shows disturbed growth curves, which were previously explained by considering two subpopulations, i.e. a temperature resistant and a thermosensitive one (Van Derlinden et al. 2010a). In this paper, the dynamics of E. coli K12 MG1655 at super-optimal and lethal temperatures were studied to reveal the influence of the initial cell concentration on growth and cell death. All experimental data are shown in Fig. 1.
For the selected temperature profiles and initial cell concentrations, a sequence of growth, inactivation and regrowth is observed. In accordance with the subpopulations hypothesis, as proposed in Van Derlinden et al. (2009, 2010a), this can be explained as follows: the first growth phase, the first (pseudo-)stationary phase and the inactivation phase are assigned to the sensitive subpopulation of E. coli. The second growth and stationary phase are attributed to the resistant subpopulation (Van Derlinden et al. 2010a). In the following parts, the influence of the initial cell concentration, for a higher and a lower heating rate, is discussed for the sensitive and resistant subpopulations, respectively.
Influence of the initial cell concentration on the dynamics of the sensitive subpopulation
The dynamics of the sensitive subpopulation is summarized by a sequence of growth and inactivation. For the different initial cell concentrations, grouped per heating rate, the cell density in the initial growth phase increased approximately at the same rate, i.e. parallel growth curves are observed (Fig. 1). The inoculum size apparently has no effect on the specific growth rate of the sensitive subpopulation.
Reaching the maximum cell concentration accounts for the transition from the growth phase to the stationary phase (Baranyi and Roberts 1994). The maximum cell concentration for E. coli cultures grown in bioreactor experiments under dynamic temperature conditions in BHI, because of depletion of nutrients, is typically around 22·5 ln(CFU ml−1) (Van Derlinden et al. 2008b). When the maximum temperature (Tmax) is attained before Nmax, the exponential phase ends earlier. As can be seen in Fig. 1, the higher initial cell concentrations result in a stationary phase for the sensitive subpopulation. Figure 1 depicts a stationary phase at initial cell concentrations of 12 and 18 ln(CFU ml−1) and a heating rate of 1°C h−1, and the initial cell concentration of 18 ln(CFU ml−1) and a heating rate of 4°C h−1. The Nmax for an initial inoculum level of 12 ln(CFU ml−1) is lower than the aforementioned concentration of 22·5 ln(CFU ml−1) (Fig. 1). Possibly, this can be explained by a combined stress effect of glucose depletion and higher temperatures. Whether the maximum cell number supported by the medium is reached is determined by the concentration of carbon source, i.e. glucose. This is confirmed by measuring the glucose concentration, which dropped significantly to a constant level for all experiments starting from 18 ln(CFU ml−1) (data not shown).
However, for experiments with low initial cell concentrations, i.e. 6 ln(CFU ml−1) at 1°C h−1 and 6 and 12 ln(CFU ml−1) at 4°C h−1, the maximum cell concentration was not reached, i.e. glucose was not depleted (data not shown). For these conditions, growth of the sensitive subpopulation would stop at the maximum growth temperature of the sensitive subpopulation, Tmax,S.
In the experiments with a lower heating rate (1°C h−1), Tmax,S for an inoculum size of 12 ln(CFU ml−1) was, based on visual inspection, undoubtedly higher than the maximum growth temperature at an inoculum size of 6 ln(CFU ml−1). Thus, the population has a higher thermoresistance. This higher thermoresistance of the stationary phase bacteria is most likely because of a large number of physiological changes and of the production of heat-shock proteins (Rees et al. 1995; Kaur et al. 1998; Nyström 2004; Diaz-Acosta et al. 2006). A higher Tmax,S is not observed for the experiment with an initial cell density of 18 ln(CFU ml−1), because the maximum cell concentration was reached before the maximum growth temperature was attained. The experiments with heating rate of 1°C h−1 showed a higher Tmax,S, and consequently a higher stress resistance, when the cells were in their stationary phase. For experiments with 4°C h−1, Tmax,S seems to be the same for all three initial cell concentrations.
The thermoresistance of the stationary phase cells of the sensitive subpopulation was also reflected by the temperature of inactivation (Tinact). This temperature was the same as the maximum growth temperature for the lower cell concentrations in Fig. 1b. At the higher cell concentrations, when Nmax is reached and a stationary phase precedes the inactivation, the inactivation is delayed (Fig. 1a,b).
The higher Tmax,S and higher Tinact, and thus a higher thermoresistance, is in agreement with the findings of Van Derlinden et al. (2009) for static temperature experiments. Skandamis et al. (2007) studied the effect of the inoculum size of E. coli on several other stress factors, such as pH and aw, and also concluded that the higher initial cell concentration results in a higher stress resistance. In our study, the pH is controlled and kept constant at 7·55. Gnanou Besse et al. (2006) found that the effect of inoculum size on the growth of Listeria monocytogenes was dependent on the physiological state of the cells. Our results lead to the same conclusion for E. coli, where the higher inoculum concentration has an indirect influence on the heat resistance of the sensitive subpopulation, because at higher initial cell concentrations the stationary phase is reached. Stationary phase cells possess a greater temperature resistance.
All experiments, regardless of the initial cell concentrations, exhibited an inactivation phase. The slope of the inactivation curve, reflecting the inactivation rate kmax(T), seems independent of the initial cell concentration (Fig. 1). At higher initial cell concentrations, i.e. when cells are in the stationary phase, Fig. 1 displays a steeper inactivation curve as inactivation is started at a higher temperature. This is in agreement with the Weibullian model of Corradini and Peleg (2007).
In summary, the inoculum level has no effect on the population dynamics of the sensitive subpopulation, except for experiments with a high initial cell concentration and low heating rate where the cells reach a maximum cell concentration. Stationary phase cells undergo physiological and morphological changes, which increase their general stress resistance.
Influence of the initial cell concentration on the evolution of the resistant subpopulation
In experiments with an initial cell concentration of 6 ln(CFU ml−1), the curves display a second growth phase (Fig. 1). This second growth phase is in agreement with the hypothesis of Van Derlinden et al. (2009), i.e. the E. coli population is composed of a thermosensitive and a temperature resistant population with the second growth phase being assigned to the temperature resistant subpopulation. The high variability of the dynamics of the resistant subpopulation is most likely related to the experimental protocol, i.e. the cells undergo a supplementary stress because of the large temperature decrease during sampling and dilution (65°C to room temperature). Despite this high variability, some general conclusions can be drawn.
Van Derlinden et al. (2010a) performed experiments at a heating rate of 1°C h−1 and obtained a maximum growth temperature of 60°C for the heat resistant subpopulation. In our research, the E. coli cultures, subjected to a heating rate of 4°C h−1, still grow at the final temperature of the profile, i.e. 65·2°C (Fig. 1b, n(0) = 6 ln(CFU ml−1)). This second growth phase is most likely not observed for the heating rate of 1°C h−1, as the resistant subpopulation has already reached its maximum cell concentration before 65·2°C was attained.
As the cells were still growing at 65·2°C after more than 20 h, this temperature is not expected to inhibit growth. The maximum growth temperature of the resistant cell population must be equal to or >65·2°C and is thus at least 18·5°C higher than the assumed maximum growth temperature of 46–47°C from dynamic experiments (Van Derlinden et al. 2008b).
For experiments with a high initial cell concentration, e.g. for an initial cell concentration of 18 ln(CFU ml−1), a second growth phase was not observed (Fig. 1). In this case, the inactivation phase of the sensitive subpopulation was immediately followed by the stationary phase of the resistant subpopulation, i.e. the resistant subpopulation was already in the stationary phase when it started to dominate the total population dynamics.
In general, the resistant E. coli population reached a stationary phase with a final cell concentration of between 12 and 15 ln(CFU ml−1) (Fig. 1). This final cell concentration seems independent of the initial cell concentration. Compared to the maximum cell count normally obtained in bioreactor experiments (i.e. 22·5–23·0 ln(CFU ml−1) (see, e.g. Van Derlinden et al. 2008b), the maximum population level of the resistant subpopulation was rather low. In the static test tube experiments at 45–46°C (published in Van Derlinden et al. 2008a,b; Van Derlinden et al. 2010b), the secondary growth phase, i.e. the resistant subpopulation, reaches a cell count of c. 18 ln(CFU ml−1), which is the same as that obtained under nonstress temperature conditions in test tubes, e.g. at the optimal growth temperature of 40°C. Consequently, it can be expected that the lower Nmax observed in the present experiments is either because of a shortage in one or more nutrients, or of the higher temperature.
In Van Derlinden et al. (2010a), a ratio of resistant cells with respect to the initial cell population was determined for a dynamic experiment with n(0) = 6 ln(CFU ml−1) and a heating rate of 1°C h−1. For this experimental set-up, it was determined that c. 7 of 10 000 inoculated cells are characterized by an increased heat resistance. It is, however, yet to be determined whether this value is an intrinsic property of the population or is rather induced by the environmental conditions. Therefore, a series of dynamic experiments with different set-ups (i.e. different n(0) values, initial and final temperatures and heating rates) need to be performed to further identify the relationship between the resistant population fraction, the population and the environmental conditions. [A more extensive discussion on this ratio can be found in Van Derlinden et al. (2010a,b)].
For all experiments, the resistant subpopulation grew until a final cell concentration of between 12 and 15 ln(CFU ml−1) was attained or is still growing at the final temperature of 65·2°C. This can be a major problem in mass catering, as in hospitals, where, according to food safety regulations, food is stored at temperatures above 63°C and is regarded to be safe (Roberts 1990).
In further research, intermediate initial cell concentrations and different initial and final temperatures and heating rates will be implemented to clarify the conditions of overall heat stress adaptation measured by an increase in the maximum growth temperature.
This research is supported in part by projects OT/09/025 and EF/05/006 (OPTEC – Optimization in Engineering) of the Research Council of the Katholieke Universiteit Leuven, project KP/09/005 (SCORES4CHEM) of the Industrial Research Fund and by the Belgian Program on Interuniversity Poles of Attraction, initiated by the Belgian Federal Science Policy Office. I. Cornet is supported by the Artesis University College of Antwerp. J. Van Impe holds the chair safety Engineering sponsored by the Belgian chemistry and life sciences Federation essenscia.