An investigation of the thermal inactivation of Staphylococcus aureus and the potential for increased thermotolerance as a result of chilled storage


Declan J. Bolton, Foods Safety Department, Teagasc – The National Food Centre, Ashtown, Dublin 15, Ireland (e-mail:


Aims:  The aims of this study were; (i) to provide thermal inactivation data for Staphylococcus aureus; (ii) to examine the kinetics, including decimal reduction times (D-value) and rate constants (k), that describe the thermal inactivation of Staph. aureus and to compare two different methods of calculating D-values and (iii) to determine whether or not chilled storage would toughen these microorganisms resulting in increased thermotolerance.

Methods and Results:  Isolates of Staph. aureus recovered from domestic refrigerators were grown in shaken culture for 8 h at 37°C, recovered and washed by centrifugation and combined to form a cocktail of five strains. Samples from this cocktail were (a) heat treated at 50, 55 and 60°C or (b) held under simulated domestic refrigeration conditions for 72 h and then heat treated as above. The numbers of Staph. aureus in heat treated and chill held, heat treated samples were enumerated by direct selective plating onto Baird Parker Agar (BPA) and recovery plating on Tryptone Soya Agar (TSA) subsequently overlaid with BPA. D-values were obtained using two different methods both of which may be used when the thermal inactivation follows first order kinetics. In the first method D-values are obtained by plotting the Log10 of the surviving cells against time and using the equation D = −1/slope. The second method uses the rate constant (k) which is obtained from the slope of a plot of ln N/N0vs time and D is obtained using the equation D = 2·303 k−1. D50, D55 and D60 values ranged from 94·3 to 127·9 min, 13 to 21·7 min and 4·8 to 6·5 min. Prechilling did not enhance thermal resistance. The method of calculation did not affect the D-values obtained because the thermal inactivation of Staph. aureus in this study followed first order kinetics with r2 values of 0·91–0·99.

Conclusions:  The thermal inactivation of Staph. aureus in tryptone soya broth (TSB) follows first order kinetics and in general chilling of these bacteria does not increase the resistance to thermal destruction during subsequent thermal processes such as cooking.

Significance and Impact of the Study:  This study provides much needed data on the thermal resistance of Staph. aureus and validates chilling as a food storage activity which does not cause toughening of the microorganisms to subsequent cooking. However, the data generated strongly suggests that Staph. aureus is more thermotolerant than Listeria monocytogenes and should be used as the target microorganism in designing mild thermal treatments for food, in which case the current recommendations for pasteurization (70°C for 2 min, minimum) should be revised.


Although not capable of growth at temperatures below 6·5°C (Halpin-Dohnalek and Marth 1989a), Staphylococcus aureus are the most prevalent pathogenic bacteria found in domestic refrigerators (Enriquez et al. 1997; Kennedy et al. 2005). This might be attributable to the fact that an estimated 50% of the human population carry this pathogen on the skin and in the nasal passage (Arbuthnott et al. 1990) presenting numerous opportunities for direct and indirect transfer of this pathogen into refrigerators during routine domestic food preparation and storage activities. Although long recognized as an opportunistic infective agent in wounds, etc. (Halpin-Dohnalek and Marth 1989a; Jablonski and Bohach 1997), Staph. aureus infection now poses even more significant challenges associated with the emergence and dissemination of multiple antibiotic resistant Staph. aureus strains (MRSA), particularly in relation to postsurgical and hospital acquired infections. In terms of food safety, Staph. aureus is most important as the source of a range of heat-stable enterotoxins (SEs) produced in contaminated foods. Ingestion of foods containing SEs leads to very rapid development of staphylococcal food poisoning, involving sudden and violent gastroenteritis. The severity of the symptoms and the frequency of occurrence of staphylococcal food poisoning means that it is more costly in terms of medical expenses and lost working days than all the other food-associated pathogens combined (Halpin-Dohnalek and Marth 1989b; Jablonski and Bohach 1997). It is fortunate that Staph. aureus does not compete well in mixed populations and is usually not a problem in unheated foods. However, mild pasteurization treatments, designed to destroy heat sensitive bacterial pathogens such as Salmonella or Escherichia coli O157:H7, may reduce the competing microflora allowing the more heat resistant Staph. aureus to survive, multiply, and produce toxin in treated foods.

In an expanding range of bacteria, it is now clear that sub-lethal treatments may in fact ‘stress harden’ surviving cells, increasing their ability to resist subsequent challenge with the same or different stresses. Thus habituation to adverse but sub-lethal temperatures may enhance the ability of surviving cells to resist more extreme adverse temperatures, acid, or osmotic conditions by the production and action of stress proteins (Perl et al. 1998; Sheridan and McDowell 1998; Miller et al. 2000; Lee 2003).

While such stress adaptations have been demonstrated in many species including Arthrobacter globiformis, E. coli, Bacillus subtilis, Listeria monocytogenes, Pseudomonas putid, Salmonella enteritidis, and Lactococcus lactis subsp. lactis, (Tang and Jackson 1979; Van Bogelen and Neidhardt 1990; Jones et al. 1987; Willimsky et al. 1992; Jones and Inouye 1994; Lottering and Streips 1995; Panoff et al. 1995; Bayles et al. 1996; Berger et al. 1996; Gumley and Inniss 1996; Jeffereys et al. 1998; Bryan et al. 1999), the study of such phenomena in Staph. aureus has yet to be fully investigated. Recent studies within our research group suggest that Staph. aureus occurs frequently in domestic refrigerators and that such domestic units may frequently operate at the margin between slow growth/bacteriostatic conditions, i.e. sub-lethal stress conditions (Flynn et al. 1992; Kennedy et al. 2005).

This study establishes the thermal resistance characteristics of Staph. aureus at a constant pH (pH 7·3) and examines the effect of simulated domestic refrigeration storage in stress hardening Staph. aureus to more effectively survive the thermal challenges associated with subsequent mild reheating/cooking processes.

Methods and materials

Chilled storage temperature profile

The chilled storage temperature profile used to treat the Staph. aureus cocktail was obtained by recording the air temperature in 50 domestic refrigerators at different locations throughout Ireland over a 72 h period using Testo 175TM temperature data loggers (Testo Ltd, Alton, Hampshire, UK). The average temperature profile at 6 h intervals is shown in Fig. 1.

Figure 1.

The average temperature of 50 refrigerators taken at 6 h intervals

Inoculum preparation

The Staph. aureus cocktail suspension used in these experiments included five strains isolated from the internal base and sides of in-use domestic refrigerators in the Republic of Ireland. Suspect Staph. aureus were isolated by plating onto Baird Parker medium (Baird Parker Agar Base with Egg Yolk tellurite emulsion, Oxoid). The plates were incubated at 37°C for 48 h. Five typical colonies of Staph. aureus were removed from each plate and subject to a series of confirmatory tests including the Gram stain procedure, testing for the production of coagulase, catalase, DNAse (DNase agar, Oxoid), testing for the fermentation of mannitol (Mannitol Salt agar, Oxoid), heamolysis (Columbia Base Agar and 5% Lysed Horse Blood, Oxoid) and oxidation. Colonies that displayed typical characteristics of Staph. aureus (i.e. positive for all the aforementioned tests) were maintained on TSA and confirmed by testing for the clumping factor (Staphylase Test Kit, Oxoid). The strains exhibiting the clumping factor were maintained on ProtectTM Stock Culture Beads (Protect, Technical Consultants Ltd, UK) at −18°C. To prepare the inoculum for the thermal inactivation experiments, a single bead of each strain was resuscitated in 30 ml Tryptone Soya Broth (TSB, CM0876, Oxoid, Unipath Ltd, Basingstoke, UK) at 37°C for 24 h. Following incubation, a 1 ml aliquot from the cultures was transferred into 99 ml TSB in a sterile bottle (Duran Schott, Mainz, Germany) and was incubated for a further 8 h in an oscillating incubator at 37°C and 200 rev min−1. The cultures were then combined and centrifuged (Eppendorf AG, Hamburg) at 4800 g for 10 min at 4°C. The recovered pellet was washed three times with, and re-suspended in, maximum recovery diluent (MRD, CM 733, Oxoid, Unipath Ltd, Basingstoke, UK). The numbers of Staph. aureus cells per millilitre of the cocktail suspension was estimated using the Acridine Orange method of Walls and Sheridan (1989) and diluted as required to a final concentration of about 8 log10 CFU ml−1.

Within 5 min, duplicate 1 ml samples of the adjusted cocktail suspension were; (a) placed under simulated domestic chilled storage for 72 h using a temperature controlled polyethylene glycol LoudaTM bath (Louda DR.R. Wobser, GMBH & Co. KG) programmed as per the temperature profile in Fig. 1 and then subjected to thermal challenge as described below or (b) subjected to thermal challenge (i.e. without a chill holding period).

Thermal challenge

Data on resistance to thermal challenge were obtained by inoculating 1 ml volumes of the bacterial suspension into 99 ml volumes of sterilized TSB, as carried out by Ugborogho and Ingham (1994), equilibrated to 50, 55 or 60°C by immersion in water baths (Grant Instruments, Cambridge, UK and Techne, Tempette Junior TE-8J) adjusted to these temperatures. Water bath temperatures were monitored using thermocouples inserted into ‘blank’ samples, attached to a temperature microprocessor (Ellab A/S, Oslo, Norway). Samples of inoculated TSB (2 ml) were withdrawn immediately after inoculation, and at 60, 5 or 2 min intervals, chilled to 4°C immediately in an ice-bath and examined as described below. The pH of the broth remained constant at pH 7·3 ± 0·2.

Each cooled sample was serially diluted in MRD, and plated onto BPA (Baird Parker Agar Base CM0275 with Egg Yolk Tellurite emulsion SR0054, Oxoid, Unipath Ltd, Basingstoke, UK) and TSA (Tryptone Soya Agar, CM0876, Oxoid). BPA plates were incubated at 37°C for 48 h to obtain a direct selective plate count. TSA plates were incubated at 25°C for 2 h to allow injured cells to recover, over-poured with BPA and incubated for a further 48 h at 37°C to obtain an overlay recovery plate count.

Calculation of D- and z-values and statistical analysis

Data on thermal death rates at 50, 55 and 60°C were obtained in duplicate (i.e. two inoculated samples of TSB at each temperature). The experiment was repeated three times.

D-values were obtained using two different methods. (i) Data on surviving cell numbers (log10) were plotted against treatment times at each temperature. The average slope and standard error (SE) of the resultant curves were obtained using linear regression analysis (Genstat 5, Statistics Department, Rothamsted Experimental Station, Hertfordshire, UK), and used to calculate D-values for each temperature treatment using eqn (1): D = −1/slope. This method is widely used in the scientific literature and its’ application in this study facilitated the comparison of the results with other relevant studies. (ii) The rate coefficient for inactivation (k) was calculated from a plot of ln N/N0vs t (from eqn (2): ln N/N0 = −kt), which gave a straight line through the origin with a slope of value (k) and the D-values subsequently calculated using eqn (3): D = 2·303 k−1. This method provides more information about the thermal inactivation kinetics and is more useful for modeling purposes. All D-values are expressed in °C.

Differences between D-values were determined by analysing differences between the average slopes using the t-test (b1 − b2/v(SE1)2 + (SE2)2 with degrees of freedom (n1 + n2) − 4). Each z-value was calculated from the slope of an individual curve of a plot of log10D-values against temperature by linear regression of the slopes of the plots. Thermal destruction times were calculated using the formula: Dx = log−1 (logD60 − ((t2 − t1)/z)) (Bolton et al. 2003).


The thermal inactivation data for the Staph. aureus cocktails are presented in Table 1. At 50°C the thermal inactivation reaction followed first order kinetics and r2 values of 0·97–0·99 were obtained. The rate coefficients for inactivation ranged from 0·018 to 0·023 min−1 and the D-values calculated using eqn (1) (ranging from 94·3 to 112·4 min) were similar to those obtained using eqn (3) (ranging from 100·1 to 127·9 min). Neither prechilling nor culture medium significantly (P > 0·05) affected the D-values obtained.

Table 1.  The r2 values for the plots of surviving cell numbers (log10) vs treatment times, the corresponding D-values calculated using the equation D = −1/slope, coefficients of inactivation (k) obtained from the plots of ln N/N0vs t and the corresponding D-values calculated using the equation D = 2·303 k−1 for Staph. aureus isolates with and without chilling at 50, 55 and 60°C
Temperature (°C)Storage conditionsCulture mediumr2D-value (min) (D = −1/slope)k (min−1)D-value (min) (D = 2·303 k−1)
50No chillingTSA/BPA0·99104·20·021109·7
50No chillingBPA0·9794·30·022104·7
55No chillingTSA/BPA0·9420·40·12818·0
55No chillingBPA0·9113·00·1713·6
60No chillingTSA/BPA0·985·90·3566·5
60No chillingBPA0·974·80·455·1

At 55°C the thermal inactivation data were also log-liner (r2 values ranged from 0·91 to 0·97). The rate coefficients of inactivation ranged from 0·109 to 0·17 min−1 and D-values ranged from 13 to 21·7 min, regardless of calculation method. Prechilling did not toughen the Staph. aureus against subsequent thermal inactivation. However, bacterial counts on the TSA/BPA plates were significantly (P < 0·05) higher than those on the BPA media for the untreated (no chilling) cells.

Thermal inactivation at 60°C also followed first order kinetics. D-values ranged from 4·8 to 6·6 min regardless of calculation method and k from 0·35 to 0·45 min−1. The chilling treatment had no affect on thermal resistance. The TSA/BPA plates recovered significantly (P < 0·05) more bacteria than the BPA selective media for both sets of treatments (no chilling and prechilling).

Z-values were calculated from the D-values obtained using eqn (1) and ranged from 7·7 to 8·0 (Table 2). Equivalence calculations for D-values using the formula Dx = log−1(logD60 − ((t2 − t1)/z)) of 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71 and 72°C were calculated (Fig. 2). These suggested, for example, that approx. 24 s would be required to achieve a 1 log reduction at 70°C.

Table 2.  The z-values for Staphylococcus aureus isolates with and without chilling and resuscitation of injured cells at 50, 55 and 60°C
Storage conditionsCulture mediumz-value
UnchilledTSA and BPA7·7
ChillingTSA and BPA8·0
Figure 2.

Time/temperature treatments (derived from equivalence calculations) required to achieve a 1 log10 reduction of unchilled (bsl00000) and prechilled (bsl00001) Staph. aureus using the formula Dx = log−1 (logD60 − ((t2 − t1)/z))


The D-values found in this study are in general agreement with the limited data available in the literature. For example, Ugborogho and Ingham (1994) reported D56 values of 10·2 min for Staph. aureus ATCC 13565 and 13·2 min for Staph. aureus ATCC 14458, in casein peptone soymeal peptone plus yeast extract. These authors also reported D56 values of 9·2 min (Staph. aureus ATCC 13565) and 9·8 min (Staph. aureus ATCC 14458) in skim milk.

There was no difference in the D-values calculated using eqn (1) (D = −1/slope from a plot of log10 survivors vs time) and those determined using eqn (2) (ln N/N0 = −kt (a plot of N/N0vs t gave a straight line through the origin with a slope of value k)) in combination with eqn (3) (D = 2·303 k−1). This was attributed to the fact that the thermal inactivation of Staph. aureus observed in this series of experiments were first order chemical reactions and the number of cells inactivated with exposure time was directly proportional to the number of cells at time = t. Thus the criticisms of the first model (eqn (1)) as an accurate description of the inactivation of microorganisms in liquids by Davey (1982) did not apply.

The first order chemical kinetics and the lack of shouldering and tailing suggests that neither heat shock protein synthesis (Lindquist 1986; Hutchison et al. 2005) nor the presence of a resistant sub-population (Peleg and Cole 1998; Murphy et al. 1999) played a role in the thermal inactivation of the Staph. aureus.

In this study, chilling did not enhance the resistance of Staph. aureus to subsequent thermal inactivation. Pathogenic bacteria can develop systems that assist them to survive and adapt to environmental stresses such as heat and cold in a variety of ways. These responses entail the production of protective proteins, some of which offer protection to more than one type of stress (Sheridan and McDowell 1998). Nearly all cells respond to a decrease in temperature by inducing a set of cold shock proteins (CSPs). These CSPs are small monomeric proteins (Perl et al. 1998) and these proteins are thought to play a role in the protection of cells against damage caused by temperature reductions (Lee 2003). Staphylococcus aureus are known to produce these cold shock proteins, although the ability to produce and intensity of production may vary from strain to strain (Cordwell et al. 2002). It is therefore possible that the cocktail of strains used in these experiments did not have a cold shock protein capacity.

The results obtained in this study suggest that Staph. aureus has a greater D-value than L. monocytogenes (Toora et al. 1992; Embarek and Huss 1993; Bolton et al. 2000; Lihono et al. 2001), the target microorganism on which many cooking regulatory guidelines for the thermal destruction of vegetative cells are based (Anonymous 2001). While, the currently recommended mild cooking conditions of 70°C for 2 min would still achieve a 5 log reduction in Staph. aureus, (which is generally accepted to be the target reduction for food-borne pathogenic bacteria (Anonymous 2001)) it would not allow for the error which is normally incorporated into such recommendations. Thus pasteurization at a time-temperatures combination of anything less than 70°C for 2·0 min might result in the survival of Staph. aureus allowing for subsequent growth and toxin production if the product was not chilled after cooking. Reheating would have little effect on the heat stable SE's (Fung et al. 1973). If Staph. aureus were to be the target microorganism, higher time-temperature recommendations would be required to include a degree of over-treatment thus allowing a margin for error and ensuring the protection of public health. A combination of at least 75°C for at least 1 min is recommended.

In a few cases, this study observed significant differences between D-values calculated from estimates of surviving bacterial numbers from direct plating on selective medium (BPA) and D-values similarly calculated from estimates from recovery plating on TSA with subsequent BPA overlay. This observation is in agreement with the report of Stranzt and Zottola (1989), who obtained significantly higher estimates, and therefore significantly higher D-values, from recovery plating than direct plating in the detection/enumeration of Salmonella enterica Typhimurium. Conversely, a pattern of ‘death rather than damage’ was found in studies by Bolton et al. (2000,2003) who noted the absence of significant differences between direct and overlay techniques in heat treated samples of L. monocytogenes, Yersinia enterocolitica and Salmonella spp. The results of this study suggest that Staph. aureus cells in the inoculation cocktail might have been injured by some of the temperature time combinations applied to them in the thermal treatments.

The temperature equivalence calculations (Dx = log−1(logD60 − ((t2 − t1)/z)) predict that increasing the temperature of the heat treatment by 1 or 2° would greatly influence the survival of Staph. aureus. For example, the equivalence calculations suggest that increasing the thermal treatment from 60 to 62°C would decrease the D-value by 1·8-fold for prechilled or unchilled Staph. aureus. Such calculations are in agreement with Sorquist and Danielsson-Tham (1990) who estimated that similar minor changes in the thermal treatments induced significant changes in D-values for Salmonella enterica Montevideo (two-fold).

In summary, this study provides thermal inactivation data for Staph. aureus that strongly suggests that this bacterium and not L. monocytogenes should be the target microorganism for mild cooking or pasteurization. The current temperature-time recommendation for such cooks should be increased from 70°C for 2 min minimum core temperature to 75°C for 1 min minimum, which would include a margin for error. This study demonstrates that D-values can be obtained using either; D = −1/slope from a plot of log10 survivors vs time or a plot of ln N/N0vs t to obtain k in combination with the equation D = 2·303 k−1, to describe thermal inactivation of bacteria when this process is a first order reaction and other parameters, such as pH, are constant. Finally, this study found no evidence that chilled storage toughened Staph. aureus to subsequent cooking.


The authors thank safefood, The Food Safety Promotion Board (Ireland) for funding this research.