Habituation of Salmonella spp. at Reduced Water Activity and Its Effect on Heat Tolerance

Abstract

The effect of habituation at reduced water activity (a_w) on heat tolerance of Salmonella spp. was investigated. Stationary-phase cells were exposed to a_w 0.95 in broths containing glucose-fructose, sodium chloride, or glycerol at 21°C for up to a week prior to heat challenge at 54°C. In addition, the effects of different a_ws and heat challenge temperatures were investigated. Habituation at a_w 0.95 resulted in increased heat tolerance at 54°C with all solutes tested. The extent of the increase and the optimal habituation time depended on the solute used. Exposure to broths containing glucose-fructose (a_w 0.95) for 12 h resulted in maximal heat tolerance, with more than a fourfold increase in D₅₄ values. Cells held for more than 72 h in these conditions, however, became as heat sensitive as nonhabituated populations. Habituation in the presence of sodium chloride or glycerol gave rise to less pronounced but still significant increases in heat tolerance at 54°C, and a shorter incubation time was required to maximize tolerance. The increase in heat tolerance following habituation in broths containing glucose-fructose (a_w 0.95) was RpoS independent. The presence of chloramphenicol or rifampin during habituation and inactivation did not affect the extent of heat tolerance achieved, suggesting that de novo protein synthesis was probably not necessary. These data highlight the importance of cell prehistory prior to heat inactivation and may have implications for food manufacturers using low-a_w ingredients.

The genus Salmonella includes important international food-borne pathogens and has been reported in approximately 30,000 cases of food-borne illness each year in England and Wales since 1990, with a decline in the number of cases over the last 2 years (4). Reported cases are likely to represent only a fraction of the true numbers, and a recent study by the Public Health Laboratory Service (PHLS) in England has indicated that infectious intestinal disease is often underreported (43). This study indicated that, for every case reported to PHLS Communicable Disease Surveillance Centre, there are 3.2 cases of Salmonella infection in the community. Salmonella enterica serovar Typhimurium definitive type (DT) 104 is the second most prevalent serotype isolated from human cases after S. enterica serovar Enteritidis phage type (PT) 4 (PHLS Communicable Disease Surveillance Centre, unpublished data for 1990–94; 16). DT104 is of particular concern due to the severity of human disease caused, its multiple-drug resistance, and its extensive animal reservoirs (42), and PT4 is clearly of interest due to the large numbers of cases associated with it. These and other Salmonella serotypes have caused food poisoning outbreaks associated with low-water-activity (a_w) foods (3, 15, 19, 27, 39). The processing of certain foods containing low-a_w ingredients is known to involve heat treatments, and it is in such situations that a knowledge of the effect of prior exposure to low a_w on the heat tolerance of Salmonella spp. is required.

Measured heat tolerance is known to depend on the conditions during heat inactivation and cell recovery (35, 40). The prehistory of Salmonella spp. prior to challenge at high temperature has also proved critical to subsequent survival data. Factors such as growth temperature and age of culture (12), broth type (17), pH of incubation conditions (17, 23), and whether cells are washed (17) are all important. One such factor that may also be important is the incubation of bacteria in conditions where water is limited prior to heat challenge.

Although it is well known that a_w during heat inactivation has a profound effect on heat tolerance (40), little is known about the effect of exposure to non-growth-permitting low a_w (e.g., <0.96 for Salmonella spp.) prior to heat challenge. Early work showed that the heat tolerance of Escherichia coli increased when cells were incubated in the presence of 50% (wt/vol) sucrose (14). In Listeria spp., habituation in broth with added NaCl (9% [wt/vol], corresponding to a_w 0.94) led to increased heat tolerance in low-a_w broths, and this was also demonstrated in a food system (24). It has also been shown that air-dried Salmonella cells, in which a_w is lowered without the use of solutes, become more heat tolerant (28). Thus, Salmonella spp. contaminating a low-a_w food ingredient (e.g., spices) may be more heat tolerant than expected.

In this study, the effect of habituation of Salmonella cells at low a_w on their ability to survive a subsequent combination of lethal heat and low a_w was investigated. Habituation over an extended time period was chosen to simulate real situations, such as the contamination of a food ingredient prior to heat processing. This study examines habituation at low a_w prior to heat challenge in a systematic manner, investigating the effect of habituation time and solute. This is believed to be the first publication to report that the time taken for Salmonella to reach optimal habituation at a defined a_w varies with respect to solute and to demonstrate that this habituation is likely to be independent of de novo protein synthesis and RpoS expression.

MATERIALS AND METHODS

Salmonella strains and preparation of cultures.

DT104 strain 30 (44), PT4 strain LA5, and PT4 strain EAV54 were used in this study. Strain 30 was isolated from cattle feces. LA5 is from a natural chicken infection (1). Strain EAV54 is an otherwise isogenic rpoS mutant of LA5 (1), kindly provided by M. J. Woodward, Veterinary Laboratory Agency, Weybridge, United Kingdom. Salmonella strains were recovered from storage at −20°C on Protect beads (Mast Diagnostics). A bead was streaked onto 5% horse blood agar and incubated at 37°C for 24 h. Stationary-phase cultures were prepared by inoculation of 9 ml of tryptone soya broth (TSB; Oxoid, Hampshire, United Kingdom) and incubation at 37°C. After 3 h, 1 μl of the initial broth was transferred into 9 ml of TSB before incubation at 37°C for 15 h by which time the cultures had reached early stationary phase (22).

Preparation of reduced-a_w broths.

AnalaR grades of glucose, fructose, NaCl, and glycerol (BDH, Leicestershire, United Kingdom) were used as humectants to produce reduced-a_w TSB. TSB base was supplemented with glucose-fructose (in equal portions), NaCl, or glycerol, and deionized water was added such that the final a_w would be lower than the required value. The broths were steamed for 30 min to avoid caramelization. The pH of the broths was adjusted to pH 6.5 ± 0.2, using HCl and NaOH, with pH measured with a pH meter (PHM93; Radiometer, Copenhagen, Denmark). The broths were then adjusted using TSB (also adjusted to pH 6.5 using HCl and NaOH) to give the required a_w values (0.91, 0.93, 0.95, 0.97, and 0.99) ± 0.003. An a_w of 0.95 is equivalent to approximately 32% (wt/vol) glucose-fructose, 8% (wt/vol) NaCl, or 20% (vol/vol) glycerol. The a_w of the broths was measured using an Aqualab CX-3T (Labcell, Hampshire, United Kingdom) water activity meter at 25°C. The water activity meter works on the “dew point” principle, involving the detection of condensation on a mirror during cooling-heating cycles. An aliquot of each batch of broth was incubated at 37°C to ensure that no viable microorganisms remained.

Measurement of heat tolerance.

One hundred and fifty microliters of a stationary-phase Salmonella culture was inoculated into 15 ml of TSB, with or without added humectant. This gave an initial cell density of approximately 10⁷ CFU ml⁻¹. These broths were then incubated under the conditions described below prior to heat challenge. The heat challenge was performed by injecting cultures into a submerged heating coil (8) held at either 54 or 60°C (± 0.1°C) measured with a mercury thermometer (Zeal, London, United Kingdom). At predetermined time intervals, 0.2 ml of the culture was ejected from the coil and an immediate 10-fold dilution was made in 1.8 ml of maximum recovery diluent to reduce the broth temperature and minimize further cell death. Further 10-fold dilutions were made in maximum recovery diluent as appropriate. Viable counts were performed using the method of Miles and Misra (31) with recovery on horse blood agar and incubation for 48 h at 37°C.

Habituation of DT104 at low a_w.

An a_w of 0.95 was chosen for the majority of habituation studies, since that was the lowest a_w at which habituation can proceed for 100 or more h without significant cell death or growth (29). Cultures of DT104 strain 30 were prepared as above and incubated at 21°C in TSB at a_w 0.95 for up to 168 h prior to heat challenge at 54 or 60°C. For TSB with glucose-fructose, the pH values were measured during the habituation. To investigate other a_ws, DT104 strain 30 was inoculated into TSB containing glucose-fructose at a_w 0.91, 0.93, 0.95, and 0.97, and it was habituated for 48 h at 21°C prior to heat challenge at 54°C in these media. Control cultures were heat challenged at the appropriate a_w without habituation or habituated for 48 h in broth with no added solute (a_w 0.99).

Involvement of protein synthesis during habituation of DT104 at a_w 0.95.

Cultures were prepared as above using DT104 strain 30 and were heat challenged at 54°C in triplicate after habituation for 48 h at 21°C in TSB containing glucose-fructose at a_w 0.95, in the presence of 100 μg of chloramphenicol per ml or 15 μg of rifampin per ml to inhibit protein synthesis. This strain of DT104 had previously been found to be sensitive to both antibiotics (data not shown). The control cultures were habituated in TSB at a_w 0.95 in the absence of antibiotic.

Habituation of PT4 at a_w 0.95.

PT4 strain LA5 was inoculated into TSB containing glucose-fructose at a_w 0.95. The cells were heat challenged at 54°C before and after habituation for 48 h at a_w 0.95 (glucose-fructose) at 21°C.

Involvement of RpoS expression during habituation of PT4 at a_w 0.95.

The PT4 rpoS mutant strain EAV54 was inoculated into TSB at a_w 0.95 (glucose-fructose) and heat challenged at 54°C before and after habituation for 48 h in this medium at 21°C.

Data analysis.

For each habituation time and solute combination, heat challenge was carried out at least in triplicate and for some as many as six times. Values for decimal reduction time (D values [26; time required to kill one log concentration of bacteria]) were calculated to summarize much of the data. Data analysis was performed with Microsoft Excel 97. Statistical significance was calculated using a t test on two samples, assuming equal variance.

RESULTS

Habituation of DT104 at low a_w prior to heat challenge at 54 or 60°C.

Habituation at a_w 0.95 prior to heat challenge resulted in increased heat tolerance at 54°C. The extent of this increase depended on the humectant used and the incubation time. After habituation of cells in TSB containing glucose-fructose (a_w 0.95), maximal heat tolerance (>4-fold increase in D₅₄ [the time necessary to kill one log concentration of bacteria at 54°C]) was observed after approximately 12 h, although a significant increase was seen after only 30 min (P = 0.002) (Fig. 1). The pH of cultures exposed to TSB containing glucose-fructose (a_w 0.95) did not decrease until after 100 h (data not shown), suggesting that the increased heat tolerance was not a result of acid-induced cross-protection to heat. When sodium chloride was the humectant, maximal heat tolerance was observed after approximately 24 h, and the increase in heat tolerance (∼2-fold increase in D₅₄) was significantly lower than with glucose-fructose (P = 0.003). When glycerol was the humectant, the measured increase in heat tolerance was similar to that seen with NaCl but occurred after only 30 min (Fig. 1). After maximal heat tolerance had been achieved, further incubation in low-a_w media caused a decline in tolerance for all humectants examined (Fig. 1). In contrast to the data obtained at 54°C, habituation in TSB containing glucose-fructose (a_w 0.95) had no significant effect on measured death rates at 60°C (Fig. 1).

FIG. 1.

Inactivation rates for DT104 strain 30 at 54°C (open symbols) or 60°C (closed symbols), following habituation at a_w 0.95 for different durations prior to heat challenge. Humectants used to reduce a_w were glucose-fructose (triangles), NaCl (squares), and glycerol (circles).

Media in the a_w range 0.91 through 0.99 were examined to see whether culture under these conditions brought about similar increases in heat tolerance compared to that seen following habituation at a_w 0.95 (Fig. 1). Cells were heat challenged in the medium in which they were habituated. These experiments revealed an interesting relationship between a_w and heat tolerance. When cells were challenged in a medium without prior habituation at reduced a_w, heat tolerance was greatest at a_w 0.91 and lowest at a_w 0.95 while an intermediate level of tolerance was observed at a_w 0.93 and 0.97 (Fig. 2A). When cells were habituated in the low-a_w media for 48 h prior to heat challenge, a different pattern emerged. Salmonella cells habituated at a_w 0.91, 0.93, 0.95, and 0.97 exhibited similar heat tolerance, but control cultures incubated for 48 h at a_w 0.99 prior to heat challenge were far more heat sensitive (Fig. 2B). The cells that were originally most heat sensitive (i.e., those at a_w 0.95) therefore showed the greatest increase in heat tolerance.

FIG. 2.

The rate of inactivation of DT104 strain 30 at 54°C, following habituation at a_w 0.91 (closed square), 0.93 (open square), 0.95 (closed circle), 0.97 (open circle), or 0.99 (closed triangle), achieved using glucose-fructose for 0 h (A) or 48 h (B).

Protein synthesis and low-a_w-induced heat tolerance at 54°C in DT104.

Inhibition of protein synthesis using either chloramphenicol or rifampin had no significant effect on the induction of heat tolerance by habituation in TSB containing glucose-fructose (a_w 0.95) when the challenge temperature was 54°C (Fig. 3), despite being present throughout habituation and inactivation.

FIG. 3.

The rate of inactivation of DT104 strain 30 at 54°C, following habituation at a_w 0.95 (glucose-fructose) for 48 h in the presence of protein synthesis inhibitors, chloramphenicol (open circles), and rifampin (open triangles). Controls were habituated with no protein synthesis inhibitors (closed squares) or with no habituation (open squares).

Habituation of PT4 at a_w 0.95 prior to heat challenge at 54°C and involvement of RpoS expression.

Habituation of PT4 strain LA5 at a_w 0.95 results in a clear increase in heat tolerance at 54°C; thus the effect of low-a_w habituation is not limited to DT104 strain 30 (Fig. 4). RpoS expression was not required for habituation to occur, since the rpoS mutant, EAV54, showed significantly enhanced heat tolerance after 48 h of incubation at a_w 0.95 (P = 0.017) (Fig. 4). This enhanced heat tolerance was similar to the increase seen in LA5, even though rpoS mutants are inherently less heat tolerant (30).

FIG. 4.

The rate of inactivation of PT4 strain LA5 (squares) and rpoS mutant EAV54 (triangles) at 54°C, following habituation for 48 h at a_w 0.95 (glucose-fructose [closed symbols]) or with no habituation (open symbols).

DISCUSSION

It is well documented that transient conditioning to a sublethal stress (habituation) can induce tolerance to a more extreme stress and that habituation to one type of stress may cross-protect to others. In this investigation, we have demonstrated that habituation to low a_w at 21°C can increase the heat tolerance in DT104 and PT4 and that the extent of heat protection afforded is dependent on the humectant used, habituation time, and challenge temperature. The effects of the three humectants were very different at the same a_w; thus it would appear that the cells are not reacting to a_w per se. Following habituation for only 30 min in TSB containing glucose-fructose (a_w 0.95), a significant increase in heat tolerance at 54°C was observed, and after 12 h, a >4-fold increase in the D₅₄ value was seen. In contrast, habituation with NaCl or glycerol at this a_w had less effect on heat tolerance (with D₅₄ values increasing only twofold). The increase in heat tolerance when glycerol is the humectant is rapid but very short lived. Habituation with NaCl caused a slower rise in heat tolerance than either glucose-fructose or glycerol, with tolerance being maximized after approximately 24 h. In all cases, after maximal heat tolerance had been achieved, further incubation at low a_w caused a decline in D₅₄ values.

It has previously been reported that at the same a_w, the heat tolerance of Salmonella spp. differs according to the solute used (17). In our study, it appeared that the extent of heat tolerance at 54°C resulting from habituation at low a_w is also affected by the solute. Glycerol is rapidly permeable (18), entering the cell very quickly, and accordingly the optimal habituation time is very short. Glycerol does not cause plasmolysis of the cell to the extent that sugars and NaCl can, and this could be a reason for its transient protective effect against heat (18). Sodium chloride is ionic and consequently disruptive to cellular activities (11, 18). The decline in heat tolerance after prolonged incubation at low a_w may relate to energy expenditure required in maintaining homeostasis (11, 25, 32). We have excluded the possibility that cells in broth may enter log phase after a certain time and thus become more sensitive to heat, since there is no significant increase in cell numbers in the broths until 168 h, when the decrease in heat resistance has already occurred.

In the current study we have grown Salmonella spp. in a standard microbiological broth (a_w 0.99) and then habituated the stationary-phase cells at non-growth-permitting low a_w. Some very early investigations indicated that heat resistance of Escherichia coli increased when that bacterium was incubated even for a short time in 50% sucrose, but after 7 h a return to normal resistance was observed (14). This pattern is consistent with our findings, although the time taken to return to normal resistance is shorter than predicted from our data, possibly due to different effects exerted by different sugars. In Listeria spp., habituation in broth with added NaCl (9% [wt/vol], corresponding to an a_w of 0.94) led to increased survival in low-a_w broths at 60°C, and this was also demonstrated in a food system (24). This is a similar pattern to that seen in this study with Salmonella spp. at 54°C. Cells dried for 48 h showed increased resistance, but no further increase was seen with longer periods of dehydration (28). Therefore, the habituation effect observed in this study may apply to air-dried cells in addition to cells in a liquid where available water is limited by the addition of a solute.

Other studies have examined the heat tolerance of bacteria following prior growth at less extreme levels of reduced a_w (0.96 through 0.98). Even at these growth-permitting a_ws (reduced using various solutes), there was usually some effect on heat tolerance (7, 9, 17), and when an increase was not observed, this could be explained in terms of the experimental protocol.

The lack of effect of habituation on heat tolerance at 60°C in this study may reflect different targets for cell death at the higher temperature compared with 54°C. Previous work demonstrated that neither adaptation to high sugar concentrations or growth at an elevated temperature resulted in increased heat tolerance at 65 or 58°C, respectively (10, 45).

During habituation and heat challenge at a_ws 0.91 through 0.99, cells cultured at a_w 0.95 developed from being the most heat-sensitive population to among the most heat-resistant population. It has been reported that solutes can cause a decrease in heat resistance when present at relatively low levels, whereas at high concentration they may afford considerable protection to heat (33). Baird-Parker et al. (5) reported a decrease in heat resistance for certain Salmonella strains at about a_w 0.94, but below this, the heat resistance increased. These findings were reflected in this study in nonhabituated controls at a_ws 0.91 through 0.97 (Fig. 2). After habituation, the cultures at a_w 0.91 through 0.97 all exhibited similar heat tolerance.

The experiments with the rpoS mutant indicated that RpoS expression is not required for the observed increase in heat tolerance following habituation in glucose-fructose. Despite evidence that proteins produced at low a_w can protect against heat challenge (6), habituation occurred in the presence of chloramphenicol and rifampin. Thus, RpoS expression and de novo protein synthesis are unlikely to have a major role in low-a_w-induced heat tolerance, unless the required proteins can be synthesized in the presence of ribosome inhibitors (as has been reported with certain other proteins [13, 36]). It may be that existing cell proteins can modify their structure and function when environmental conditions change, as demonstrated for certain heat shock proteins (37, 38). The solutes used in this study are also likely to produce substantial osmotic stress, leading to the accumulation of compatible solutes. Such accumulation may be independent of protein synthesis and can result in increased heat tolerance (20). Finally, targets for heat inactivation may be directly protected by the solutes, possibly in a temperature-dependent manner.

This work has possible important implications for the design of experimental protocols and for food manufacturing and processing. In earlier studies, a low-a_w habituation stage was built with the same duration for each solute (34). The new results presented here indicate that an optimal habituation time for one solute may give no benefit for another and that the habituation step should be tailored to the solute type. In addition, exposure of Salmonella spp. to low a_w could reduce the effectiveness of any subsequent heat processing. This will be of particular concern to food manufacturers whose processing conditions involve a heat treatment step and, in particular, when ingredients that are dry or have a high solute concentration are to be processed.