- Email: [email protected]
The behaviour of log phase Escherichia coli at temperatures below the minimum for sustained growth
Food Microbiology, 2002, 19, 83^90 Available online at http://www.idealibrary.com on
doi:10.1006/fmic.2001.0465
ORIGINAL ARTICLE
The behaviour of log phase Escherichia coli at temperatures below the minimum for sustained growth T. Jones1; *, C. O. Gill1 and L. McMullen2 The behaviour of cold-adapted, log phase Escherichia coli broth cultures during incubation at 21C or 61C for upto 8 days, and during subsequent incubation at 121C, was determined by measurement of absorbance values at 600 nm (A600), enumeration of colony forming units (cfu) on plate count agar (PCA) and violet red bile agar (VRBA), and measurement of the length of cells viewed under phase contrast illumination. The A600 values and the mean length of cells remained constant for cultures incubated at 21C; however, numbers of cfu recovered on PCA declined by about 1log cfu over 8 days, while the numbers of cfu recovered on VRBA declined by about 1log cfu during the ¢rst day, and by about a further log cfu by day 8. For cultures incubated at 61C, A600 values increased about 0?6 log A600 units during the ¢rst 4 days and declined by less than 0?1log A600 unit during the next 4 days. The numbers of cfu recovered on PCA increased by about 0?5 log cfu unit during the ¢rst day at 61C and declined by about 1log cfu during the subsequent 7 days. The numbers of cfu recovered on VRBA did not increase during the ¢rst day at 61C, and at that and subsequent times were between 0?3 and 0?8 log cfu less than the log numbers recovered on PCA. The mean lengths of cells declined from 5 to less than 4 mm during the ¢rst day at 61C, but increased to 8 mm between the fourth and eighth days, with the mean length of the longest 10% of cells increasing from 6 to 18 mm. For cultures incubated at 121C after incubation at 21C or 61C for 4 and 8 days, both A600 values and enumeration of colonies on PCA indicated the initiation of growth after about 15 h. However, cultures that had been incubated at 21C proceeded to sustained exponential growth, while cells in cultures that had been incubated at 61C elongated during incubation at 121C between 10 and 30 h. The division of elongated cells to cells of normal size resulted in numbers of cfu increasing at rates greater than the exponential growth rate at 121C. The observations may have implications for the control of mesophilic pathogen proliferation in raw meats and other chilled foods. # 2002 Elsevier Science Ltd. Received: 7 August 2001
Introduction Fabrication of raw meat at temperatures around 101C may result in Escherichia coli and related pathogens being in the logarithmic
*Corresponding author. Fax: 403 -782 - 6120. Email: [email protected] 0740 -0020/02/020083 +08 $35.00/0
phase of growth when the product is cooled to chiller cabinet temperatures for storage or display (Smith 1985). Temperatures of meat on retail display may £uctuate above and below the minimum temperatures for growth of E. coli and other mesophilic organisms (Olsson 1990). Although the proliferation of pathogens may be slow at temperatures near their minimum for growth, any increase in the numbers of r 2002 Elsevier Science Ltd.
1 Agriculture and Agri-Food Canada, Lacombe Research Centre, 6000 C&E Trail, Lacombe, AlbertaT4L1W1, Canada 2 Department of Agricultural, Food and Nutritional Sciences, University of Alberta, Edmonton, Alberta, T6G 2P5, Canada
84 T. Jones et al.
pathogens with very low infectious doses, such as E. coli O157:H7 (Willshaw et al. 1994), may increase risks to consumers’ health. Knowledge of the behaviour of cold-adapted, log phase cells that are exposed to temperatures near the minimum for growth is therefore required for realistic assessments of the risks that may arise from the growth of mesophilic pathogens during commercial processes for the storage and display of meat (Gill et al. 1998). The behaviour of cold-adapted, log phase E. coli at temperatures below 71C, the minimum temperature for growth of the organism, was recently examined (Gill et al. 2001). The ¢ndings of that study were that absorbance of broth cultures incubated at non-permissive temperatures increased at declining rates for several days, but no corresponding increases in numbers of colony forming units (cfu) as determined by plating were apparent. When E. coli cultures were returned from 21C to a growth-permitting temperature, the duration of the lag phase apparently increased, and initial growth rates apparently decreased with increasing times of incubation at 21C. The behaviour of log phase E. coli exposed to chiller temperatures appeared to be complex and variable with the chiller temperature, while data for numbers of cfu and absorbance values for the same broth cultures were contradictory. Therefore, a study was undertaken to resolve the discrepancies between counts of cfu and absorbance data for the same cultures, and to characterize better the lag phase that develops in log phase E. coli exposed to chiller temperatures.
tures were prepared by inoculating a £ask with 2 ml of a culture grown to the stationary phase at 251C, and incubating the £ask at 121C, with shaking at 100 rpm until an absorbance at 600 nm (A600 ) of 0?4^0?5 was reached. Absorbance values were determined using a spectrophotometer (UltraSpec III, Pharmacia LKB Biotechnology, Uppsala, Sweden). The cold-adapted, log phase cultures were used to inoculate £asks of BHI that were prechilled to the temperature at which each £ask would be incubated. Colony forming units were determined by preparing serial ten-fold dilutions in chilled 0?1% peptone water and spreading 0?1 ml of each appropriate dilution on to each of ¢ve plates of plate count agar (PCA; Difco) and on to ¢ve plates of violet red bile agar (VRBA; Difco). PCA and VRBA plates were incubated at 351C for 24 h.
Measurement of cell lengths For each determination of cell lengths, a drop of molten, 2% agar containing 4% formaldehyde (Fisher Scienti¢c, Edmonton, Canada) was added to 10 ml of culture on a microscope slide and the preparation was covered with a cover glass. Slides were stored at 21C for up to 7 days before they were examined. The slides were viewed by phase contrast microscopy using a microscope (Axioscope; Zeiss, Jena, Germany) connected to a colour video camera (DXC 930; Sony Corporation,Tokyo, Japan).The lengths of 100 randomly selected cells were measured on images from each slide, using Image Pro-Plus software Version 4.0 (Media Cybernetics, Silverspring, Maryland, USA).
Materials and Methods Culture preparation and enumeration of E. coli
Survival and growth of cold-adapted E. coli in BHI
A wild-type strain of E. coli isolated from a beef packing plant and maintained in cooked meat medium (Difco Laboratories, Detroit, Michigan, USA) was cultivated in 70 ml of half strength brain^heart infusion (BHI; Difco) in 125-ml £asks. A 0?1-ml portion of a stationary phase culture was inoculated into 70 ml BHI and was incubated overnight at 251C under static conditions. Cold-adapted, log phase cul-
Four £asks of BHI were each inoculated with a cold-adapted log phase culture to obtain an initial A600 of 0?2. Pairs of £asks were incubated at 61C or 21C in a shaking water bath for 8 days. Portions of each culture were removed immediately after inoculation and at daily intervals for determination of A600 values, measurement of cell lengths and the enumeration of colonies on PCA and VRBA.
Behaviour of chilled Escherichia coli 85
0.7 0.6 0.5
Change in log10 A600
To determine the behaviour of cultures at 121C after incubation at 61C or 21C, £asks of BHI were each inoculated to obtain an initial A600 of about 0?05. Five pairs of £asks were prepared and incubated, then a further ¢ve were prepared 12 h later to allow cultures to be sampled at convenient times. At each time, one pair of £asks was incubated at 21C for 4 days and three pairs of £asks were incubated at 61C, with one pair being incubated for 0?5, 4 or 8 days. After incubation at 21C or 61C, £asks were incubated at 121C for 48 h. A pair of £asks prepared at each time was incubated at 121C to serve as a control. Immediately after inoculation, at the time of transfer to 121C, and at 1-, 2- or 4-h intervals during incubation at 121C, portions of each culture were removed for determination of A600 values, measurement of cell lengths and enumeration of colonies on PCA and VRBA.
0.4 0.3 0.2 0.1 0 −0.1 1
0
2
3 4 5 Time (days)
6
7
8
Figure 1. Changes in absorbance values [log (A600 )] of cold-adapted, log phase Escherichia coli cultures incubated in 1/2 -strength brain^heart infusion broth at 21C (*) or 61C (*) for 8 days.
9.0
Analysis of data
Results Behaviour of cold-adapted E. coli at 61C and 21C The A600 values of cultures incubated at 21C for 8 days remained constant (Fig. 1). At 61C, the
Log cfu ml−1
For each culture, increases in log10 A600 values (log10 (A600 at tðxÞ /A600 at t0 )), where tðxÞ is the time of incubation in hours and t0 is 0 h, were plotted against the time of incubation. For cultures incubated at 121C, the duration of the lag was determined from the intercept with the x-axis of the backward extrapolation of the regression line to the steepest part of the log10 increase of A600 vs time curve. Colony counts were converted to log10 values and mean log values were calculated for each set of ¢ve plates for each culture. The duration of the lag was calculated from the intercept with the log cfu at 0 h of the backward extrapolation of the line to the steepest part of each log cfu vs time curve.The cell lengths were tested for normal distribution using the Wilk-Shapiro test in the PROC UNIVARIATE procedure of SAS (SAS Version 8, SAS Institute, Cary, North Carolina, USA). Mean lengths and the mean length of the longest 10% of cells were calculated using Microsoft Excel 97 (Microsoft Corp., Redmond,Washington, USA).
8.5 8.0 7.5 7.0 6.5 6.0 5.5
0
1
2
4 5 3 Time (days)
6
7
8
Figure 2. Numbers (log cfu ml1 ) of Escherichia coli recovered on plate count agar (circles) or violet red bile agar (squares) from log phase cultures incubated at 21C (open ¢gures) or 61C (closed ¢gures) for 8 days.
A600 values increased at a progressively diminishing rate for 5 days, then slowly declined. During the incubation of cultures at 21C, the numbers of colonies recovered on PCA declined progressively (Fig. 2). In contrast, when cultures were incubated at 61C, the numbers of colonies recovered on PCA increased during the ¢rst day, but subsequently decreased slowly. During incubation at 21C or 61C, the number of colonies recovered on VRBA was about 10% or between 60 and 80%, respectively, of the numbers of colonies recovered on PCA (Fig. 2).
86 T. Jones et al.
20
1.4
18 16
1.2 Change in log10 A600
Length (µm)
14 12 10 8 6 4 2
1.0 0.8 0.6 0.4 0.2
0 0
1
2
3
4
5
6
7
8
Time (days)
Figure 3. Mean lengths of all (open ¢gures) or the longest 10% (closed ¢gures) of cells in log phase cultures of Escherichia coli incubated at 21C (circles) or 61C (squares) for 8 days. The standard deviations within sets of log counts were o0?1 or o0?2 log unit for colonies recovered on PCA or VRBA, respectively. During incubation at 21C, the mean length of the cells remained nearly constant and the mean length of the longest 10% of cells remained below 10 mm (Fig. 3). During incubation at 61C, the mean length of cells decreased during the ¢rst day but gradually increased after the third day. The mean length of the longest 10% of cells increased about three-fold between 1 and 8 days at 61C. The mean lengths of the shortest 10% of cells were similar at all times at both temperatures.
Growth of E. coli at 121C after incubation at 61C or 21C When growth at 121C was determined from changes in A600 values, the exponential growth rate (EGR) of cultures incubated at 121C was 0?16 generations h1 (Fig. 4). For cultures that were incubated at 121C after being incubated at 21C for 4 days there was little increase in the A600 values during the ¢rst 13 h. Thereafter, the growth rate gradually increased until the maximum growth rate was attained after 34 h. After that time, the EGR was identical to that of the control culture. Back extrapolation of the log increase of A600 vs time curve for times after 34 h yielded an apparent lag time of 24 h. Cultures that were incubated at 61C for 0?5 or 4
0
10
20 30 Time (h)
40
50
Figure 4. Changes
in absorbance values [log10 (A600 )] of cold-adapted, log phase Escherichia coli cultures incubated at 121C (*) or at 121C after incubation at 21C for 4 days (*) or at 61C for 0?5 days (&), 4 days (&) or 8 days (~).
days before being incubated at 121C had short or long lag durations, respectively, but all attained an EGR of 0?16 generations h1 . The pairs of cultures incubated at 61C for 8 days before being incubated at 121C had di¡erent lag durations of 15 h or 20?5 h, and di¡erent EGRs of 0?12 or 0?09 generations h1 . When growth at 121C was determined from changes in the numbers of colonies recovered on PCA, the EGR of cultures incubated at 121C was 0?16 generations h1 (Fig. 5). For cultures that were incubated at 21C for 4 days before being incubated at 121C, an apparent lag of 10 h was followed by an EGR identical to that of the control culture. With cultures that were incubated at 61C for 0?5 day, the numbers of colonies recovered on PCA had not declined from the numbers recovered before incubation. When such cultures were incubated at 121C, growth occurred without lag. After that initial growth, growth ceased for 7 h.When growth resumed, the EGR was slightly lower than that of the control culture. The apparent lag was 4 h. Numbers of colonies recovered on PCA decreased in cultures incubated at 61C for 4 or 8 days. Cultures incubated at 61C for 4 days before being incubated at 121C had an apparent lag of 16 h. With such cultures, the initial rate of growth was 0?13 generations h1
Behaviour of chilled Escherichia coli 87
(a) 9.0 8.5 8.0 7.5 7.0 6.5 6.0 Log cfu ml−1
between 16 h and 24 h at 121C. However, between 32 and 36 h the growth rate accelerated to rates much higher than that of the control culture. A lag and EGR could not be calculated from cultures incubated at 121C after being incubated at 61C for 8 days because the culture did not approach a constant rate of growth. Inoculation of cultures into BHI at 21C or 61C resulted in di¡erences of 0?6^2?0 or 0^0?6 log cfu, respectively, between the numbers of colonies recovered on PCA or VRBA (data not shown). Numbers of colonies recovered on VRBA from cultures that were incubated at 21C for 4 days were 1 log cfu less than numbers recovered on PCA. When such cultures were incubated at 121C, the numbers of colonies recovered on VRBA decreased during the ¢rst 4 h, but the numbers of colonies recovered on PCA remained unchanged (Fig. 5). After 24 h at 121C, the numbers of colonies recovered from VRBA were similar to the numbers recovered on PCA, but at 34 h, 44 h and 48 h, the di¡erences between the numbers of colonies recovered on VRBA and PCA were about 1 log cfu. Numbers recovered on VRBA from cultures that were incubated at 61C for 0?5 or 4 days were less than the numbers recovered from PCA. After 18 h at 121C, the numbers recovered from VRBA were similar to the numbers recovered on PCA. Between 24 and 48 h at 121C, the numbers recovered from VRBA were less than the numbers recovered from PCA. Numbers of colonies recovered on VRBA from cultures that were incubated at 61C for 8 days were somewhat less than the numbers recovered on PCA (Fig. 6). Di¡erences between the numbers of colonies recovered on PCA and VRBA £uctuated, with maximum differences at 6 -h intervals, during incubation at 121C for times beyond 18 h.The standard deviations within sets of log counts were about 0?1 log unit or ranged up to 0?6 log unit for colonies recovered on PCA orVRBA, respectively, for cultures incubated at 121C after incubation at 21C for 4 days or at 61C for 0?5, 4 or 8 days. The mean lengths of cells for cultures incubated at 121C and for cultures incubated at 21C for 4 days before being incubated at 121C, ranged from 4 to 5 mm (Fig.7).The mean lengths of cells for cultures incubated for 0?5 days at 61C declined after 1 h at 121C and remained be-
5.5 0
4
8
12
24
28
32
36
16
20
24
9.0
(b) 8.5 8.0 7.5 7.0 6.5 6.0 5.5 40
44
48
Time (h)
Figure 5. Numbers (log cfu ml1 ) of Escherichia coli recovered on plate count agar from log phase cultures incubated at 121C (*) or at 121C after incubation at 21C for 4 days (*) or at 61C for 0?5 days (&), 4 days (&) or 8 days (~); growth rate of control culture (- - -). Incubation at 121C between 0 and 24 h (a), or between 24 and 48 h (b).
low 4 mm for the next 5 h. However, the mean lengths of cells for cultures incubated at 61C for 4 or 8 days before being incubated at 121C increased with increasing incubation time, to
88 T. Jones et al.
8.5
45 40
8.0
35 Length (µm)
Log cfu ml−1
7.5 7.0 6.5
30 25 20 15
6.0
10 5
5.5
0 0
10
5.0 0
10
20
30
40
40
50
50
Figure 8. Mean lengths of the longest 10% of
Time (h)
Figure 6. Numbers (log cfu ml1 ) of Escherichia coli recovered on plate count agar (*) or violet red bile agar (*) from log phase cultures incubated at 121C for 48 h after incubation at 61C for 8 days.
Length (µm)
30 20 Time (h)
cells in cold-adapted, log phase Escherichia coli cultures incubated at 121C for 48 h after incubation at 21C for 4 days (*) or after incubation at 61C for 0?5 days (*), 4 days (&) or 8 days (&).
cells incubated at 121C for 26 h after being incubated at 61C for 8 days were 410 mm long (Fig. 9). Branched and swollen cells were observed with increasing incubation time at 61C but not at 21C.
12 11 10 9 8 7 6 5 4 3 2
Discussion
0
10
20 30 Time (h)
40
50
Figure 7. Mean lengths of cells in cold-adapted, log phase Escherichia coli cultures incubated at 121C for 48 h after incubation at 21C for 4 days (*) or after incubation at 61C for 0?5 days (*), 4 days (&) or 8 days (&). maximum lengths after 26 h. After 44 h at 121C, the mean cell lengths had decreased to about 5 mm. The mean lengths of the longest 10% of cells in cultures incubated at 2, 6 or 121C were 11 mm (Fig. 8), but for cultures incubated at 121C after being incubated at 61C for 4 or 8 days, the mean lengths of the longest 10% of cell increased with time, to reach maximum lengths of 23 or 40 mm, respectively, after 26 h. The distribution of cell lengths at any sampling time was not normal. Most cells incubated at 121C were o10 mm long, but many
In a previous study, when log phase cultures of E. coli were incubated at 21C, the absorbance values of cultures remained unchanged at incubation times of up to 8 days, while the numbers of cfu recovered on PCA or VRBA were similar and did not change (Gill et al. 2001). When log phase cultures were incubated at 61C, the A600 values increased about fourfold at declining rates during the ¢rst 5 days but changed little thereafter, while the numbers of cfu recovered on PCA or VRBA were again similar and did not change. In this study, the A600 values of cultures were unchanged during incubation at 21C and increased during the ¢rst 5 days of incubation at 61C, as in the previous study. However, numbers of cfu recovered on PCA declined after the ¢rst day of incubation at 21C, or increased and then declined during incubation at 61C. Moreover, the numbers of cfu recovered on VRBA were less than those recovered on PCA for all times of incubation at either temperature. The discrepancy between the studies with respect to the numbers
Behaviour of chilled Escherichia coli 89
(a)
70 60 50 40 30 20 10
0 (b) 50 Percentage
40 30 20 10 0 (c) 70 60 50 40 30 20 10
0 0
10
20
30
40
50
60
70
Length (µm)
Figure 9. Distribution of the lengths of coldadapted, log phase Escherichia coli cells incubated at 121C (control, a), or after 26 h (b) or 44 h at 121C (c) after being incubated at 61C for 8 days.
recovered on PCA probably re£ects the better discrimination by ¢vefold sampling in the second study instead of duplicate sampling (Jarvis 1989).The discrepancy with respect to the numbers recovered on VRBA possibly re£ects differences of inhibitory e¡ects on sub-lethally injured cells between the batches of the commercial medium used for the two studies (Stephens et al. 1997). The ¢ndings of this study with regard to the numbers of cfu recovered from cultures and the lengths of cells permit a more extensive description of the behaviour of log phase E. coli exposed to temperatures of 21C or 61C than was previously possible.
When log phase E. coli were incubated at 21C growth apparently ceased with little or no delay, as A600 values, numbers of cfu recovered on PCA and cell lengths altered little during the ¢rst day of incubation at that temperature. However, about 90% of the cells were injured during that time, as indicated by the di¡erence in the numbers of cfu recovered on PCA or VRBA. Subsequently, the numbers of viable cells slowly declined while the fraction of injured cells remained approximately constant. Similar reduction in numbers of viable organisms during incubation at temperatures below the minimum for growth of various bacteria have been previously reported (Gill 2001). In contrast, at 61C, growth continued during the ¢rst day, with increases in both A600 values and the numbers of cfu recovered on PCA, and decreases of cell length, but with injury of about 50% of the cells. Subsequently, the numbers of cfu recovered on PCA declined while A600 values increased. These observations indicate that while a fraction of the cell population was growing and so increasing the total of viable and moribund cells that were detected by A600 measurements, a larger fraction of the cell population was losing viability. The heterogeneous condition of the cultures incubated at 61C was con¢rmed by the elongation of a fraction of the cells after 5 days of incubation. These ¢ndings disagree with reports of cultures losing viability more rapidly at higher than at lower temperatures below the minimum for growth (Gray and Cerf 1997, Muntada-Garriga et al. 1995). It has been shown previously that when E. coli is incubated at 21C for 4 h, a lag of about 2 h at 121C develops. After longer times of incubation at 21C, growth is still initiated after 2 h at 121C, but an accelerating phase of growth increases in duration with the time of incubation at 21C. The extended accelerating phases apparent from A600 measurements are likely to be the result of the fractions of injured and moribund cells increasing as incubation at 21C is extended. With cultures returned to 121C after incubation at 61C for 4 or 8 days, the A600 data indicate behaviour similar to that of cultures incubated at 21C. This is seemingly contradicted by the numbers of cfu recovered on PCA, and by the
90 T. Jones et al.
invariant mean cell length during that time. The cell length data give no indication that cells which elongated during incubation at 61C divided at early times after cultures were returned to 121C. Instead, those data show that cells tend to elongate when growth is resumed at 121C. Between 26 and 32 h the elongated cells then divided to give rates of increase of cfu recovered on PCA that exceeded the rate of exponential growth at 121C. The increases in cell length between 10 and 26 h and the apparently cyclic variations in the numbers of cfu recovered on VRBA as compared with the numbers recovered on PCA suggest that growth might have been synchronized when it resumed at 121C after incubation at 61C. Bacterial cells exposed to environmental stresses may be injured or killed, or they may grow into elongated or other unusual forms (Imlay and Linn 1987). Evidently, log phase E. coli that experience a temperature of 61C are diversely a¡ected in all three ways, and the behaviour of log phase E. coli at temperatures between 71C and 21C is probably similar (Gill et al. 2001). If such complex behaviour occurs in foods as well as broth cultures (Fedio 1986), then estimating the proliferation of E. coli and associated pathogens in chilled foods, which often experience marginal or £uctuating chiller temperatures (James 1996), may prove di⁄cult. The extent to which the ¢nding for broth cultures can be extended to describe the behaviour of E. coli in chilled foods will therefore require further investigation.
Acknowledgements We thank the Alberta Agricultural Research Institute for the funding of this study.
References Fedio, W. M. (1986) Fate of Salmonella typhimurium at suboptimal growth temperatures. MSc Thesis. University of Alberta, Edmonton, Alberta.
Gill, C. O. (2001) Microbial control with cold temperatures. In Control of foodborne microorganisms (eds V. K. Juneja and J. N. Sofos), pp. 55^73. New York. Marcel Dekker. Gill, C. O., Greer, G. G., and Dilts, B. D. (1998) Predicting the growth of Escherichia coli on displayed pork. Food Microbiol. 15, 235^242. Gill, C. O., Greer, G. G., Jones, T., Badoni, M. and Dilts, B. D. (2001) Induction of a lag phase by chiller temperatures in Escherichia coli growing in broth or on pork. Food Microbiol. 18, 141^149. Gray, A. and Cerf, O. (1997) Signi¢cance of temperature and pre-incubation temperature on survival of Listeria monocytogenes at pH 4.8. Lett. Appl. Microbiol. 25, 257^260. Imlay, J. A. and Linn, S. (1987) Mutagenesis and stress response induced in Escherichia coli by hydrogen peroxide. J. Bacteriol. 169, 2967^2976. James, S. (1996) The chill chain ‘from carcass to consumer’. Meat Sci. 43(Suppl), S203^216. Jarvis, B. (1989) Errors associated with colony count procedures. In Statistical Aspects of the Microbiological Analysis of Foods, pp. 95^116. Amsterdam, Elsevier Applied Science. Muntada-Garriga, J. M., Rodriguez-Jerez, J. J., Lopez-Sabater, E. I. and Mara-Ventara M. T. (1995) E¡ect of chill and freezing temperatures on survival of Vibrio parahemolyticus inoculated in homogenates of oyster meat. Lett. Appl. Microbiol. 20, 225^227. Olsson, P. (1990) Chill cabinet surveys. In Processing and Quality of Foods, Vol. 3. (eds P. Zeuthen, J. C. Chefter, C. Erikkson, T. R. Gromely, P. Linko and K. Paulus) pp. 3.279^3.288. London, Elsevier Applied Science. Smith, M. G. (1985) The generation time, lag time, and minimum temperature of growth of coliform organisms on meat, and implications for codes of practice in abbatoirs. J. Hyg. Camb. 94, 289^300. Stephens, P. J., Joynson, J. A., Davies, K. W., Holbrook, R., Lappin-Scott, H. M. and J. Humphrey, T. (1997) The use of an automated growth analyser to measure recovery times of single heat-injured Salmonella cells. J. Appl. Microbiol. 83, 445^ 455. Willshaw, G. A., Thirlwell, J., Jones, A. P., Parry, S., Salmon, R. L. and Hickey, M. (1994) Vero cytotoxin-producing Escherichia coli O157 in beef-burgers linked to an outbreak of diarrhoea, haemorrhagic colitis and haemolytic uraemic syndrome in Britain. Lett. Appl. Microbiol. 19, 304^307.
doi:10.1006/fmic.2001.0465
ORIGINAL ARTICLE
The behaviour of log phase Escherichia coli at temperatures below the minimum for sustained growth T. Jones1; *, C. O. Gill1 and L. McMullen2 The behaviour of cold-adapted, log phase Escherichia coli broth cultures during incubation at 21C or 61C for upto 8 days, and during subsequent incubation at 121C, was determined by measurement of absorbance values at 600 nm (A600), enumeration of colony forming units (cfu) on plate count agar (PCA) and violet red bile agar (VRBA), and measurement of the length of cells viewed under phase contrast illumination. The A600 values and the mean length of cells remained constant for cultures incubated at 21C; however, numbers of cfu recovered on PCA declined by about 1log cfu over 8 days, while the numbers of cfu recovered on VRBA declined by about 1log cfu during the ¢rst day, and by about a further log cfu by day 8. For cultures incubated at 61C, A600 values increased about 0?6 log A600 units during the ¢rst 4 days and declined by less than 0?1log A600 unit during the next 4 days. The numbers of cfu recovered on PCA increased by about 0?5 log cfu unit during the ¢rst day at 61C and declined by about 1log cfu during the subsequent 7 days. The numbers of cfu recovered on VRBA did not increase during the ¢rst day at 61C, and at that and subsequent times were between 0?3 and 0?8 log cfu less than the log numbers recovered on PCA. The mean lengths of cells declined from 5 to less than 4 mm during the ¢rst day at 61C, but increased to 8 mm between the fourth and eighth days, with the mean length of the longest 10% of cells increasing from 6 to 18 mm. For cultures incubated at 121C after incubation at 21C or 61C for 4 and 8 days, both A600 values and enumeration of colonies on PCA indicated the initiation of growth after about 15 h. However, cultures that had been incubated at 21C proceeded to sustained exponential growth, while cells in cultures that had been incubated at 61C elongated during incubation at 121C between 10 and 30 h. The division of elongated cells to cells of normal size resulted in numbers of cfu increasing at rates greater than the exponential growth rate at 121C. The observations may have implications for the control of mesophilic pathogen proliferation in raw meats and other chilled foods. # 2002 Elsevier Science Ltd. Received: 7 August 2001
Introduction Fabrication of raw meat at temperatures around 101C may result in Escherichia coli and related pathogens being in the logarithmic
*Corresponding author. Fax: 403 -782 - 6120. Email: [email protected] 0740 -0020/02/020083 +08 $35.00/0
phase of growth when the product is cooled to chiller cabinet temperatures for storage or display (Smith 1985). Temperatures of meat on retail display may £uctuate above and below the minimum temperatures for growth of E. coli and other mesophilic organisms (Olsson 1990). Although the proliferation of pathogens may be slow at temperatures near their minimum for growth, any increase in the numbers of r 2002 Elsevier Science Ltd.
1 Agriculture and Agri-Food Canada, Lacombe Research Centre, 6000 C&E Trail, Lacombe, AlbertaT4L1W1, Canada 2 Department of Agricultural, Food and Nutritional Sciences, University of Alberta, Edmonton, Alberta, T6G 2P5, Canada
84 T. Jones et al.
pathogens with very low infectious doses, such as E. coli O157:H7 (Willshaw et al. 1994), may increase risks to consumers’ health. Knowledge of the behaviour of cold-adapted, log phase cells that are exposed to temperatures near the minimum for growth is therefore required for realistic assessments of the risks that may arise from the growth of mesophilic pathogens during commercial processes for the storage and display of meat (Gill et al. 1998). The behaviour of cold-adapted, log phase E. coli at temperatures below 71C, the minimum temperature for growth of the organism, was recently examined (Gill et al. 2001). The ¢ndings of that study were that absorbance of broth cultures incubated at non-permissive temperatures increased at declining rates for several days, but no corresponding increases in numbers of colony forming units (cfu) as determined by plating were apparent. When E. coli cultures were returned from 21C to a growth-permitting temperature, the duration of the lag phase apparently increased, and initial growth rates apparently decreased with increasing times of incubation at 21C. The behaviour of log phase E. coli exposed to chiller temperatures appeared to be complex and variable with the chiller temperature, while data for numbers of cfu and absorbance values for the same broth cultures were contradictory. Therefore, a study was undertaken to resolve the discrepancies between counts of cfu and absorbance data for the same cultures, and to characterize better the lag phase that develops in log phase E. coli exposed to chiller temperatures.
tures were prepared by inoculating a £ask with 2 ml of a culture grown to the stationary phase at 251C, and incubating the £ask at 121C, with shaking at 100 rpm until an absorbance at 600 nm (A600 ) of 0?4^0?5 was reached. Absorbance values were determined using a spectrophotometer (UltraSpec III, Pharmacia LKB Biotechnology, Uppsala, Sweden). The cold-adapted, log phase cultures were used to inoculate £asks of BHI that were prechilled to the temperature at which each £ask would be incubated. Colony forming units were determined by preparing serial ten-fold dilutions in chilled 0?1% peptone water and spreading 0?1 ml of each appropriate dilution on to each of ¢ve plates of plate count agar (PCA; Difco) and on to ¢ve plates of violet red bile agar (VRBA; Difco). PCA and VRBA plates were incubated at 351C for 24 h.
Measurement of cell lengths For each determination of cell lengths, a drop of molten, 2% agar containing 4% formaldehyde (Fisher Scienti¢c, Edmonton, Canada) was added to 10 ml of culture on a microscope slide and the preparation was covered with a cover glass. Slides were stored at 21C for up to 7 days before they were examined. The slides were viewed by phase contrast microscopy using a microscope (Axioscope; Zeiss, Jena, Germany) connected to a colour video camera (DXC 930; Sony Corporation,Tokyo, Japan).The lengths of 100 randomly selected cells were measured on images from each slide, using Image Pro-Plus software Version 4.0 (Media Cybernetics, Silverspring, Maryland, USA).
Materials and Methods Culture preparation and enumeration of E. coli
Survival and growth of cold-adapted E. coli in BHI
A wild-type strain of E. coli isolated from a beef packing plant and maintained in cooked meat medium (Difco Laboratories, Detroit, Michigan, USA) was cultivated in 70 ml of half strength brain^heart infusion (BHI; Difco) in 125-ml £asks. A 0?1-ml portion of a stationary phase culture was inoculated into 70 ml BHI and was incubated overnight at 251C under static conditions. Cold-adapted, log phase cul-
Four £asks of BHI were each inoculated with a cold-adapted log phase culture to obtain an initial A600 of 0?2. Pairs of £asks were incubated at 61C or 21C in a shaking water bath for 8 days. Portions of each culture were removed immediately after inoculation and at daily intervals for determination of A600 values, measurement of cell lengths and the enumeration of colonies on PCA and VRBA.
Behaviour of chilled Escherichia coli 85
0.7 0.6 0.5
Change in log10 A600
To determine the behaviour of cultures at 121C after incubation at 61C or 21C, £asks of BHI were each inoculated to obtain an initial A600 of about 0?05. Five pairs of £asks were prepared and incubated, then a further ¢ve were prepared 12 h later to allow cultures to be sampled at convenient times. At each time, one pair of £asks was incubated at 21C for 4 days and three pairs of £asks were incubated at 61C, with one pair being incubated for 0?5, 4 or 8 days. After incubation at 21C or 61C, £asks were incubated at 121C for 48 h. A pair of £asks prepared at each time was incubated at 121C to serve as a control. Immediately after inoculation, at the time of transfer to 121C, and at 1-, 2- or 4-h intervals during incubation at 121C, portions of each culture were removed for determination of A600 values, measurement of cell lengths and enumeration of colonies on PCA and VRBA.
0.4 0.3 0.2 0.1 0 −0.1 1
0
2
3 4 5 Time (days)
6
7
8
Figure 1. Changes in absorbance values [log (A600 )] of cold-adapted, log phase Escherichia coli cultures incubated in 1/2 -strength brain^heart infusion broth at 21C (*) or 61C (*) for 8 days.
9.0
Analysis of data
Results Behaviour of cold-adapted E. coli at 61C and 21C The A600 values of cultures incubated at 21C for 8 days remained constant (Fig. 1). At 61C, the
Log cfu ml−1
For each culture, increases in log10 A600 values (log10 (A600 at tðxÞ /A600 at t0 )), where tðxÞ is the time of incubation in hours and t0 is 0 h, were plotted against the time of incubation. For cultures incubated at 121C, the duration of the lag was determined from the intercept with the x-axis of the backward extrapolation of the regression line to the steepest part of the log10 increase of A600 vs time curve. Colony counts were converted to log10 values and mean log values were calculated for each set of ¢ve plates for each culture. The duration of the lag was calculated from the intercept with the log cfu at 0 h of the backward extrapolation of the line to the steepest part of each log cfu vs time curve.The cell lengths were tested for normal distribution using the Wilk-Shapiro test in the PROC UNIVARIATE procedure of SAS (SAS Version 8, SAS Institute, Cary, North Carolina, USA). Mean lengths and the mean length of the longest 10% of cells were calculated using Microsoft Excel 97 (Microsoft Corp., Redmond,Washington, USA).
8.5 8.0 7.5 7.0 6.5 6.0 5.5
0
1
2
4 5 3 Time (days)
6
7
8
Figure 2. Numbers (log cfu ml1 ) of Escherichia coli recovered on plate count agar (circles) or violet red bile agar (squares) from log phase cultures incubated at 21C (open ¢gures) or 61C (closed ¢gures) for 8 days.
A600 values increased at a progressively diminishing rate for 5 days, then slowly declined. During the incubation of cultures at 21C, the numbers of colonies recovered on PCA declined progressively (Fig. 2). In contrast, when cultures were incubated at 61C, the numbers of colonies recovered on PCA increased during the ¢rst day, but subsequently decreased slowly. During incubation at 21C or 61C, the number of colonies recovered on VRBA was about 10% or between 60 and 80%, respectively, of the numbers of colonies recovered on PCA (Fig. 2).
86 T. Jones et al.
20
1.4
18 16
1.2 Change in log10 A600
Length (µm)
14 12 10 8 6 4 2
1.0 0.8 0.6 0.4 0.2
0 0
1
2
3
4
5
6
7
8
Time (days)
Figure 3. Mean lengths of all (open ¢gures) or the longest 10% (closed ¢gures) of cells in log phase cultures of Escherichia coli incubated at 21C (circles) or 61C (squares) for 8 days. The standard deviations within sets of log counts were o0?1 or o0?2 log unit for colonies recovered on PCA or VRBA, respectively. During incubation at 21C, the mean length of the cells remained nearly constant and the mean length of the longest 10% of cells remained below 10 mm (Fig. 3). During incubation at 61C, the mean length of cells decreased during the ¢rst day but gradually increased after the third day. The mean length of the longest 10% of cells increased about three-fold between 1 and 8 days at 61C. The mean lengths of the shortest 10% of cells were similar at all times at both temperatures.
Growth of E. coli at 121C after incubation at 61C or 21C When growth at 121C was determined from changes in A600 values, the exponential growth rate (EGR) of cultures incubated at 121C was 0?16 generations h1 (Fig. 4). For cultures that were incubated at 121C after being incubated at 21C for 4 days there was little increase in the A600 values during the ¢rst 13 h. Thereafter, the growth rate gradually increased until the maximum growth rate was attained after 34 h. After that time, the EGR was identical to that of the control culture. Back extrapolation of the log increase of A600 vs time curve for times after 34 h yielded an apparent lag time of 24 h. Cultures that were incubated at 61C for 0?5 or 4
0
10
20 30 Time (h)
40
50
Figure 4. Changes
in absorbance values [log10 (A600 )] of cold-adapted, log phase Escherichia coli cultures incubated at 121C (*) or at 121C after incubation at 21C for 4 days (*) or at 61C for 0?5 days (&), 4 days (&) or 8 days (~).
days before being incubated at 121C had short or long lag durations, respectively, but all attained an EGR of 0?16 generations h1 . The pairs of cultures incubated at 61C for 8 days before being incubated at 121C had di¡erent lag durations of 15 h or 20?5 h, and di¡erent EGRs of 0?12 or 0?09 generations h1 . When growth at 121C was determined from changes in the numbers of colonies recovered on PCA, the EGR of cultures incubated at 121C was 0?16 generations h1 (Fig. 5). For cultures that were incubated at 21C for 4 days before being incubated at 121C, an apparent lag of 10 h was followed by an EGR identical to that of the control culture. With cultures that were incubated at 61C for 0?5 day, the numbers of colonies recovered on PCA had not declined from the numbers recovered before incubation. When such cultures were incubated at 121C, growth occurred without lag. After that initial growth, growth ceased for 7 h.When growth resumed, the EGR was slightly lower than that of the control culture. The apparent lag was 4 h. Numbers of colonies recovered on PCA decreased in cultures incubated at 61C for 4 or 8 days. Cultures incubated at 61C for 4 days before being incubated at 121C had an apparent lag of 16 h. With such cultures, the initial rate of growth was 0?13 generations h1
Behaviour of chilled Escherichia coli 87
(a) 9.0 8.5 8.0 7.5 7.0 6.5 6.0 Log cfu ml−1
between 16 h and 24 h at 121C. However, between 32 and 36 h the growth rate accelerated to rates much higher than that of the control culture. A lag and EGR could not be calculated from cultures incubated at 121C after being incubated at 61C for 8 days because the culture did not approach a constant rate of growth. Inoculation of cultures into BHI at 21C or 61C resulted in di¡erences of 0?6^2?0 or 0^0?6 log cfu, respectively, between the numbers of colonies recovered on PCA or VRBA (data not shown). Numbers of colonies recovered on VRBA from cultures that were incubated at 21C for 4 days were 1 log cfu less than numbers recovered on PCA. When such cultures were incubated at 121C, the numbers of colonies recovered on VRBA decreased during the ¢rst 4 h, but the numbers of colonies recovered on PCA remained unchanged (Fig. 5). After 24 h at 121C, the numbers of colonies recovered from VRBA were similar to the numbers recovered on PCA, but at 34 h, 44 h and 48 h, the di¡erences between the numbers of colonies recovered on VRBA and PCA were about 1 log cfu. Numbers recovered on VRBA from cultures that were incubated at 61C for 0?5 or 4 days were less than the numbers recovered from PCA. After 18 h at 121C, the numbers recovered from VRBA were similar to the numbers recovered on PCA. Between 24 and 48 h at 121C, the numbers recovered from VRBA were less than the numbers recovered from PCA. Numbers of colonies recovered on VRBA from cultures that were incubated at 61C for 8 days were somewhat less than the numbers recovered on PCA (Fig. 6). Di¡erences between the numbers of colonies recovered on PCA and VRBA £uctuated, with maximum differences at 6 -h intervals, during incubation at 121C for times beyond 18 h.The standard deviations within sets of log counts were about 0?1 log unit or ranged up to 0?6 log unit for colonies recovered on PCA orVRBA, respectively, for cultures incubated at 121C after incubation at 21C for 4 days or at 61C for 0?5, 4 or 8 days. The mean lengths of cells for cultures incubated at 121C and for cultures incubated at 21C for 4 days before being incubated at 121C, ranged from 4 to 5 mm (Fig.7).The mean lengths of cells for cultures incubated for 0?5 days at 61C declined after 1 h at 121C and remained be-
5.5 0
4
8
12
24
28
32
36
16
20
24
9.0
(b) 8.5 8.0 7.5 7.0 6.5 6.0 5.5 40
44
48
Time (h)
Figure 5. Numbers (log cfu ml1 ) of Escherichia coli recovered on plate count agar from log phase cultures incubated at 121C (*) or at 121C after incubation at 21C for 4 days (*) or at 61C for 0?5 days (&), 4 days (&) or 8 days (~); growth rate of control culture (- - -). Incubation at 121C between 0 and 24 h (a), or between 24 and 48 h (b).
low 4 mm for the next 5 h. However, the mean lengths of cells for cultures incubated at 61C for 4 or 8 days before being incubated at 121C increased with increasing incubation time, to
88 T. Jones et al.
8.5
45 40
8.0
35 Length (µm)
Log cfu ml−1
7.5 7.0 6.5
30 25 20 15
6.0
10 5
5.5
0 0
10
5.0 0
10
20
30
40
40
50
50
Figure 8. Mean lengths of the longest 10% of
Time (h)
Figure 6. Numbers (log cfu ml1 ) of Escherichia coli recovered on plate count agar (*) or violet red bile agar (*) from log phase cultures incubated at 121C for 48 h after incubation at 61C for 8 days.
Length (µm)
30 20 Time (h)
cells in cold-adapted, log phase Escherichia coli cultures incubated at 121C for 48 h after incubation at 21C for 4 days (*) or after incubation at 61C for 0?5 days (*), 4 days (&) or 8 days (&).
cells incubated at 121C for 26 h after being incubated at 61C for 8 days were 410 mm long (Fig. 9). Branched and swollen cells were observed with increasing incubation time at 61C but not at 21C.
12 11 10 9 8 7 6 5 4 3 2
Discussion
0
10
20 30 Time (h)
40
50
Figure 7. Mean lengths of cells in cold-adapted, log phase Escherichia coli cultures incubated at 121C for 48 h after incubation at 21C for 4 days (*) or after incubation at 61C for 0?5 days (*), 4 days (&) or 8 days (&). maximum lengths after 26 h. After 44 h at 121C, the mean cell lengths had decreased to about 5 mm. The mean lengths of the longest 10% of cells in cultures incubated at 2, 6 or 121C were 11 mm (Fig. 8), but for cultures incubated at 121C after being incubated at 61C for 4 or 8 days, the mean lengths of the longest 10% of cell increased with time, to reach maximum lengths of 23 or 40 mm, respectively, after 26 h. The distribution of cell lengths at any sampling time was not normal. Most cells incubated at 121C were o10 mm long, but many
In a previous study, when log phase cultures of E. coli were incubated at 21C, the absorbance values of cultures remained unchanged at incubation times of up to 8 days, while the numbers of cfu recovered on PCA or VRBA were similar and did not change (Gill et al. 2001). When log phase cultures were incubated at 61C, the A600 values increased about fourfold at declining rates during the ¢rst 5 days but changed little thereafter, while the numbers of cfu recovered on PCA or VRBA were again similar and did not change. In this study, the A600 values of cultures were unchanged during incubation at 21C and increased during the ¢rst 5 days of incubation at 61C, as in the previous study. However, numbers of cfu recovered on PCA declined after the ¢rst day of incubation at 21C, or increased and then declined during incubation at 61C. Moreover, the numbers of cfu recovered on VRBA were less than those recovered on PCA for all times of incubation at either temperature. The discrepancy between the studies with respect to the numbers
Behaviour of chilled Escherichia coli 89
(a)
70 60 50 40 30 20 10
0 (b) 50 Percentage
40 30 20 10 0 (c) 70 60 50 40 30 20 10
0 0
10
20
30
40
50
60
70
Length (µm)
Figure 9. Distribution of the lengths of coldadapted, log phase Escherichia coli cells incubated at 121C (control, a), or after 26 h (b) or 44 h at 121C (c) after being incubated at 61C for 8 days.
recovered on PCA probably re£ects the better discrimination by ¢vefold sampling in the second study instead of duplicate sampling (Jarvis 1989).The discrepancy with respect to the numbers recovered on VRBA possibly re£ects differences of inhibitory e¡ects on sub-lethally injured cells between the batches of the commercial medium used for the two studies (Stephens et al. 1997). The ¢ndings of this study with regard to the numbers of cfu recovered from cultures and the lengths of cells permit a more extensive description of the behaviour of log phase E. coli exposed to temperatures of 21C or 61C than was previously possible.
When log phase E. coli were incubated at 21C growth apparently ceased with little or no delay, as A600 values, numbers of cfu recovered on PCA and cell lengths altered little during the ¢rst day of incubation at that temperature. However, about 90% of the cells were injured during that time, as indicated by the di¡erence in the numbers of cfu recovered on PCA or VRBA. Subsequently, the numbers of viable cells slowly declined while the fraction of injured cells remained approximately constant. Similar reduction in numbers of viable organisms during incubation at temperatures below the minimum for growth of various bacteria have been previously reported (Gill 2001). In contrast, at 61C, growth continued during the ¢rst day, with increases in both A600 values and the numbers of cfu recovered on PCA, and decreases of cell length, but with injury of about 50% of the cells. Subsequently, the numbers of cfu recovered on PCA declined while A600 values increased. These observations indicate that while a fraction of the cell population was growing and so increasing the total of viable and moribund cells that were detected by A600 measurements, a larger fraction of the cell population was losing viability. The heterogeneous condition of the cultures incubated at 61C was con¢rmed by the elongation of a fraction of the cells after 5 days of incubation. These ¢ndings disagree with reports of cultures losing viability more rapidly at higher than at lower temperatures below the minimum for growth (Gray and Cerf 1997, Muntada-Garriga et al. 1995). It has been shown previously that when E. coli is incubated at 21C for 4 h, a lag of about 2 h at 121C develops. After longer times of incubation at 21C, growth is still initiated after 2 h at 121C, but an accelerating phase of growth increases in duration with the time of incubation at 21C. The extended accelerating phases apparent from A600 measurements are likely to be the result of the fractions of injured and moribund cells increasing as incubation at 21C is extended. With cultures returned to 121C after incubation at 61C for 4 or 8 days, the A600 data indicate behaviour similar to that of cultures incubated at 21C. This is seemingly contradicted by the numbers of cfu recovered on PCA, and by the
90 T. Jones et al.
invariant mean cell length during that time. The cell length data give no indication that cells which elongated during incubation at 61C divided at early times after cultures were returned to 121C. Instead, those data show that cells tend to elongate when growth is resumed at 121C. Between 26 and 32 h the elongated cells then divided to give rates of increase of cfu recovered on PCA that exceeded the rate of exponential growth at 121C. The increases in cell length between 10 and 26 h and the apparently cyclic variations in the numbers of cfu recovered on VRBA as compared with the numbers recovered on PCA suggest that growth might have been synchronized when it resumed at 121C after incubation at 61C. Bacterial cells exposed to environmental stresses may be injured or killed, or they may grow into elongated or other unusual forms (Imlay and Linn 1987). Evidently, log phase E. coli that experience a temperature of 61C are diversely a¡ected in all three ways, and the behaviour of log phase E. coli at temperatures between 71C and 21C is probably similar (Gill et al. 2001). If such complex behaviour occurs in foods as well as broth cultures (Fedio 1986), then estimating the proliferation of E. coli and associated pathogens in chilled foods, which often experience marginal or £uctuating chiller temperatures (James 1996), may prove di⁄cult. The extent to which the ¢nding for broth cultures can be extended to describe the behaviour of E. coli in chilled foods will therefore require further investigation.
Acknowledgements We thank the Alberta Agricultural Research Institute for the funding of this study.
References Fedio, W. M. (1986) Fate of Salmonella typhimurium at suboptimal growth temperatures. MSc Thesis. University of Alberta, Edmonton, Alberta.
Gill, C. O. (2001) Microbial control with cold temperatures. In Control of foodborne microorganisms (eds V. K. Juneja and J. N. Sofos), pp. 55^73. New York. Marcel Dekker. Gill, C. O., Greer, G. G., and Dilts, B. D. (1998) Predicting the growth of Escherichia coli on displayed pork. Food Microbiol. 15, 235^242. Gill, C. O., Greer, G. G., Jones, T., Badoni, M. and Dilts, B. D. (2001) Induction of a lag phase by chiller temperatures in Escherichia coli growing in broth or on pork. Food Microbiol. 18, 141^149. Gray, A. and Cerf, O. (1997) Signi¢cance of temperature and pre-incubation temperature on survival of Listeria monocytogenes at pH 4.8. Lett. Appl. Microbiol. 25, 257^260. Imlay, J. A. and Linn, S. (1987) Mutagenesis and stress response induced in Escherichia coli by hydrogen peroxide. J. Bacteriol. 169, 2967^2976. James, S. (1996) The chill chain ‘from carcass to consumer’. Meat Sci. 43(Suppl), S203^216. Jarvis, B. (1989) Errors associated with colony count procedures. In Statistical Aspects of the Microbiological Analysis of Foods, pp. 95^116. Amsterdam, Elsevier Applied Science. Muntada-Garriga, J. M., Rodriguez-Jerez, J. J., Lopez-Sabater, E. I. and Mara-Ventara M. T. (1995) E¡ect of chill and freezing temperatures on survival of Vibrio parahemolyticus inoculated in homogenates of oyster meat. Lett. Appl. Microbiol. 20, 225^227. Olsson, P. (1990) Chill cabinet surveys. In Processing and Quality of Foods, Vol. 3. (eds P. Zeuthen, J. C. Chefter, C. Erikkson, T. R. Gromely, P. Linko and K. Paulus) pp. 3.279^3.288. London, Elsevier Applied Science. Smith, M. G. (1985) The generation time, lag time, and minimum temperature of growth of coliform organisms on meat, and implications for codes of practice in abbatoirs. J. Hyg. Camb. 94, 289^300. Stephens, P. J., Joynson, J. A., Davies, K. W., Holbrook, R., Lappin-Scott, H. M. and J. Humphrey, T. (1997) The use of an automated growth analyser to measure recovery times of single heat-injured Salmonella cells. J. Appl. Microbiol. 83, 445^ 455. Willshaw, G. A., Thirlwell, J., Jones, A. P., Parry, S., Salmon, R. L. and Hickey, M. (1994) Vero cytotoxin-producing Escherichia coli O157 in beef-burgers linked to an outbreak of diarrhoea, haemorrhagic colitis and haemolytic uraemic syndrome in Britain. Lett. Appl. Microbiol. 19, 304^307.