Survival of Escherichia coli O157:H7 in frozen foods: impact of the cold shock response

International Journal of Food Microbiology 64 (2001) 127–138 www.elsevier.nl / locate / ijfoodmicro

Survival of Escherichia coli O157:H7 in frozen foods: impact of the cold shock response Jill Bollman, Anne Ismond, Greg Blank* Department of Food Science, University of Manitoba, Winnipeg, Manitoba, Canada R3 T 2 NT Received 24 November 1998; received in revised form 25 August 2000; accepted 5 October 2000

Abstract The survival of Escherichia coli O157:H7 strains in both frozen foods and trypticase soy broth (TSB) was investigated following cold shocking at 108C for 1.5 h. Using both trypticase soy agar (TSA) and violet red bile agar (VRBA) as recovery media, it was demonstrated that survival levels between cold shocked (CS) and non-cold shocked (NS) E. coli in ground beef or pork were not significantly different (P # 0.05). In contrast, cold shocking E. coli in either milk, whole egg or sausage resulted in a significant(P # 0.05) enhancement in survival. For milk, survival levels of CS E. coli, by 28 days of frozen storage, were 1.89 and 1.66 log 10 cfu / ml higher on TSA and VRBA, respectively, when compared to NS cells. In egg these values were 0.64 and 1.31, while in sausage, values of 0.74 and 1.19 were obtained. In TSB (pH 7) survival of CS E. coli for some strains was about 3 log 10 cfu / ml higher when compared to NS cells. Acidification of TSB (pH 5), however, appeared to negate the protective effects of the cold shock treatment. In milk, increasing the differential between the growth and cold shock temperatures resulted in higher numbers of survivors. In this regard the growth-cold shock temperature protocol giving optimum protection was 37–108C. In contrast, growth of E. coli at 208C followed by cold shocking at 108C did not result in any significant freeze protection. In addition, increased protection due to cold shocking was correlated with the appearance of a novel protein appearing at pI 4.8 following isoelectric focusing analysis, thus demonstrating an alteration of protein synthesis.  2001 Elsevier Science B.V. All rights reserved. Keywords: Escherichia coli; Cold shock; Proteins; Foods; Frozen; Survival; Injury

1. Introduction Most microorganisms react to a wide range of environmental stresses by inducing the synthesis of proteins (Watson, 1990). Despite numerous studies which have examined stress responses (Graumann *Corresponding author. Tel.: 11-204-474-8742; fax: 11-204474-7630. E-mail address: [email protected] (G. Blank).

and Marahiel, 1994; Graumann et al., 1996), the function of these proteins is largely unknown. The cold shock response, initially described by Jones et al. (1987) for E. coli generally occurs when cells are exposed to an abrupt downshift in temperature. For example, temperature shifts of 138C or more normally result in complete induction of the cold shock response which culminates in the synthesis of at least 16 polypeptides or cold shock proteins (Csps) (Berry and Foegeding, 1997).

0168-1605 / 01 / $ – see front matter  2001 Elsevier Science B.V. All rights reserved. PII: S0168-1605( 00 )00463-3

128

J. Bollman et al. / International Journal of Food Microbiology 64 (2001) 127 – 138

The function(s) of the Csps have not been fully elucidated; however, a number of theories have been proposed including their possible role in protecting microorganisms during freezing. In this regard, antifreeze proteins could lower the freeze temperature and inhibit recrystallization and growth of ice crystals during frozen storage (Feeney and Yeh, 1998). In support of this proposal Goldstein et al. (1990) reported that when E. coli, grown at 378C, was frozen and thawed after pre-incubation at 108C (cold shock temperature) for 6 h, a 70-fold increase in survival resulted compared to a similar protocol without cold shocking. Additionally, Willimsky et al. (1992) constructed a Bacillus subtilis mutant in which the gene coding for a major Csp (CspB) was eliminated. Their research concluded that cells which were able to undergo the complete cold shock response had an increased survival rate over their mutant counterparts. Whyte and Inniss (1992) also reported a correlation between the increase in numbers and relative synthesis of Csps in Bacillus psychrophilus with lower cold shock temperatures. This suggested that the production of such proteins constituted an adaptive response by the bacterium to permit or aid in growth at low temperatures. In contrast, Cloutier et al. (1992) reported that arctic strains of rhizobia, following exposure to a variety of cold shock regimes, did not appear to exhibit any enhanced survival over their temperate counterparts during freezing. Rapid cooling and freezing of foods are important techniques used to maintain quality and safety particularly in the dairy and meat industries. Although cold shocking has not been previously thought to enhance the survival of pathogens such as E. coli in foods, cooling regimes do exist in the food industry which could potentially induce a cold shock phenomenon. Therefore, the purpose of the present study was to evaluate the impact of cold shocking on strains of E. coli O157:H7 in various foods and culture broths relative to their survival and injury during subsequent frozen storage.

2. Materials and methods

2.1. Bacterial strains and growth conditions E. coli O157:H7 strains 7283 and 7282 (beef isolates), 7110, 7128, 7236 (human isolates), and

7174 (bovine isolate) were kindly donated by the Laboratory Centre for Disease Control, Ottawa, Canada. A non-enterohemorrhagic E. coli strain (MY20) was obtained from the Food Development Center, Portage la Prairie, MB. All strains were maintained on trypticase soy agar plates (TSA, BBL, Cockeysville, MD; pH 7.0) at 48C. Cultures were activated by transferring loop inocula into 20 ml of trypticase soy broth (TSB, BBL; pH, 7.1) at 378C. Following two consecutive 24-h culture transfers, 10 ml were inoculated into TSB (20 ml) contained in 50-ml Erlenmeyer flasks. Incubation at 378C for 4 h using a rotary shaker (150 rpm) resulted in mid-log cultures.

2.2. Impact of cold shocking on survival and injury following freezing 2.2.1. Products Regular ground beef, uncooked breakfast sausages (casing removed after purchase) and medium ground pork (chub pack) were purchased at a local retail store; whole pasteurized egg was provided by Inovatech (Winnipeg, MB). The products (10 g) were placed into sterile, sampling bags (170 g capacity) and spread to a uniform thickness of about 2 mm. Following freezing (2208C) the samples were sterilized (25 kGy) using electron beam irradiation (Impela 1–10 / 1, AECL Pinawa, MB) and stored at 2 208C until used. Additional samples (5 ml) consisting of: milk (2%, UHT; retail), sterile TSB (BBL, pH 7.1) and TSB, pH adjusted to 5.0 with 1 M citric acid, were aseptically transferred to 15-ml sterile plastic tubes (Corning, Corning, NY) and held at 48C until use. 2.2.2. Product inoculation and cold shocking Mid-log cultures of the E. coli strains were separately inoculated into each product to achieve a final level of about 10 6 cfu per ml or per g with the exception of ground pork where a cocktail consisting of three (7282, 7110, 7283) strains was used as the inoculum (same total cell concentration). Following inoculation of the egg or meat products, the sample bags were throughly massaged by hand to distribute the organisms; liquid products were vigorously mixed for 30 s. Prior to cold shocking, the samples were tempered to 108C and inoculated. Cold shock treatment was carried out at 108C for 1.5 h. The samples were then placed in a freezer maintained at

J. Bollman et al. / International Journal of Food Microbiology 64 (2001) 127 – 138

2 20618C for up to 28 days. Controls (non-cold shock) consisted of samples which were tempered to room temperature ( | 228C); after inoculation they were placed in a freezer maintained at 2 20618C for up to 28 days.

2.2.3. Food product proximate analysis All food samples with the exception of UHT milk were analyzed for proximate composition. The moisture content was determined as the difference in weight before and after sample freeze-drying (Virtis, Gardiner, NY). The freeze-dried samples were used for all subsequent proximate analysis. The percent fat, total ash, protein and pH were determined according to AOAC (1990). A protein to nitrogen conversion factor of 6.25 was used. 2.2.4. Determination of survivors Survivors were assessed at 0, 1, 7 and 28 days of frozen storage. The food samples were thawed (liquid samples 30 min; solid samples 10 min) at room temperature (about 228C) before analysis. Sterile peptone water (0.1%, 90 ml) was added to the sample bags which were then massaged by hand. Samples were serially diluted using 0.1% sterile peptone water and surface plated in duplicate (0.1 ml) or quadruplicate (0.25 ml) on both TSA (BBL) and violet red bile agar (VRBA, BBL). Colonies were enumerated following incubation for 24 h at 378C. The square root of the number of colonies on one plate (highest dilution) per sample were streaked onto Sorbitol MacConkey agar (Difco, Detroit, MN) in order to assess culture purity. Survivors (log 10 colony forming units; cfu per g or per ml) and percent injury were determined. The latter was expressed as: cfu TSA 2 cfu VRBA ]]]]]] 3 100. cfu TSA

2.3. Temperature shifts inducing cold shock Mid-log cultures of strain 7282 which was arbitrarily chosen (10 7 cfu / ml), were grown in TSB as previously outlined. Following inoculation (10 ml) into a series of pre-warmed test-tubes containing 5 ml UHT milk maintained at 378C, replicate tubes were either transferred directly to a freezer maintained at 220618C (control) or rapidly cooled on ice to about 128C and then cold shocked at 10618C

129

for 1.5 h in a temperature-controlled incubator. The samples were subsequently placed in a freezer maintained at 2208C. At 0, 1 and 7 days, tubes were thawed (15 min) at room temperature. The milk samples were serially diluted and survivors were enumerated using both TSA and VRBA following incubation at 378C for 24 h. Similar studies were performed using the following growth-cold shock temperature protocols: 37–10, 30–10, 30–15, 20– 108C. In all cases the time for the cold shock treatment was maintained at 1.5 h.

2.4. Statistical analysis All experiments were performed in triplicate. To determine if differences between data sets existed, a Duncan’s multiple range test on all main effect means was performed using the ANOVA analysis procedure in the Statistical Analysis System (SAS Institute, Cary, NC).

2.5. Isoelectric focusing 2.5.1. Cell growth, cold shock treatment and cell extract preparation Mid-log cultures of E. coli strain 7283 grown in TSB (pH 7) as previously described were cold shocked and harvested by centrifugation (108C, 10 0003g, 20 min). Resultant pellets were washed twice in 10 mM Tris buffer (pH 6.8) and transferred to beakers containing 25 ml of 1% CHAPS in 10 mM Tris buffer (pH 6.8). The cell suspensions were chilled on ice, sonically disrupted (Braunsonic 1510, B. Braun, Melsungen, AG; 250 W; three 1-min treatments) and centrifuged (48C, 78003g, 15 min). The resultant supernatant was transferred to a sterile tube and kept at 2208C until required. Cold shocking consisted of placing flasks containing E. coli in an ice bath where the temperature was rapidly (about 5 min) dropped to 128C. The cultures were then transferred to a temperature-controlled incubator maintained at 10618C for 1.5 h. The time temperature was monitored using a digital thermometer (Omega Engineering, Stanford, CT) which was placed in a non-inoculated flask. Noncold shocked cultures (controls) were prepared in a similar manner, however, following mid-log growth they were not incubated at 108C for 1.5 h. Instead the cells were directly centrifuged (258C) and subsequently disrupted as previously described. Follow-

130

J. Bollman et al. / International Journal of Food Microbiology 64 (2001) 127 – 138

ing centrifugation, the supernatants were collected and maintained at 2208C.

2.5.2. Protein determination Sample protein concentrations were assessed prior to isoelectric focusing using the Pierce Protein Assay Reagent (Pierce Chemical Company, Rockford, IL) which is based on a reaction between protein and Coomassie Blue R-250. A series of protein standards (200–1500 mg / ml) were prepared using a stock solution (2 mg / ml) of bovine serum albumin (Pierce) diluted with 0.1 N NaOH. Standards and test protein samples (0.1 ml) were reacted with 5 ml of the protein assay reagent. Absorbance readings were made at 595 nm. 2.5.3. Isoelectric evaluation of sample Isoelectric focusing was carried out on an LKB 2117 Multiphor apparatus (Sweden) with an LKB 2197 Constant Power Supply. The temperature was controlled at 108C with a Haake circulating water bath (Germany). A Pharmacia Biotech Ampholine PAG (polyacrylamide gel) plate with a pH range of 4.0–6.5 was used. After the PAG plate was placed on the unit, electrode strips with 0.1 M b-alanine and

0.1 M glutamic acid in 0.5 M H 3 PO 4 were placed in the cathode and anode positions, respectively. Twenty ml of protein sample and Pharmacia Biotech standards (pH 2.5–6.5) were applied to applicator pieces and placed on the surface of the gel. The gel was focused for 2.5 h at a constant power of 25 W. The applicator pieces were removed after about 1.25 h. Upon completion, staining and preservation were carried out according to the method of Winter and Anderson (1997). After the gel was preserved, a plastic Mylar sheet soaked in preserving solution was placed over the gel and dried at room temperature. Isoelectric points of the proteins were determined in quadruplicate with reference to the standard calibration curve.

3. Results

3.1. Survival due to cold shock treatment The effect of cold shock treatment on E. coli in relation to survival and percent injury in ground beef and ground pork is presented in Figs. 1 and 2. In most cases, differences in the number of survivors

Fig. 1. Effect of cold shocking on the survival of E. coli (7110) in ground beef at 2208C. Results with the same superscript for each medium and storage time are not significantly different (P.0.05). Bars represent averages of triplicate results each performed in duplicate (n566S.D.). Non-cold shocked (NS); cold shocked (CS); tryptic soy agar (TSA); violet red bile agar (VRBA).

J. Bollman et al. / International Journal of Food Microbiology 64 (2001) 127 – 138

131

Fig. 2. Effect of cold shocking on the survival of E. coli (cocktail consisting of 7110, 7282, 7238) in ground pork at 2208C. Results with the same superscript for each medium and storage time are not significantly different (P.0.05). Bars represent averages of triplicate results each performed in duplicate (n566S.D.). Non-cold shocked (NS); cold shocked (CS); tryptic soy agar (TSA); violet red bile agar (VRBA).

Fig. 3. Effect of cold shocking on the survival of E. coli (7110) in milk at 2208C. Results with the same superscript for each medium and storage time are not significantly different (P.0.05). Bars represent averages of triplicate results each performed in duplicate (n566S.D.). Non-cold shocked (NS); cold shocked (CS); tryptic soy agar (TSA); violet red bile agar (VRBA).

J. Bollman et al. / International Journal of Food Microbiology 64 (2001) 127 – 138

132

Table 1 Typical injury levels in E. coli during frozen storage at 2208C Product (strain) Ground beef (7110)

Storage time (days)

1

28

7 28

Milk (7110)

7 28

7 28 Sausage (7110)

NS CS NS CS NS CS

NS CS NS CS NS CS

7 28

27.3612.8 68.364.1 58.969.5 63.6612.8 95.261.2 97.160.8 5.067.8

NS CS NS CS NS CS

0 1

15.3614.2 9.6610.6 54.661.4 47.0611.8 78.668.3 61.967.1 8.568.0

0 1

9.269.4 15.8615.6 14.7612.7 6.565.8 32.666.7 32.662.6 2.763.6

0 1

Egg (7110)

NS a CS b NS CS NS CS

0 1

% Injury

1.762.3 c

0

7

Ground pork (7282, 7110, 7283)

Treatment

23.665.3 13.966.5 72.465.4 4.563.5 82.166.4 20.762.1

period the number of survivors declined, but usually not more than 1 log 10 . Although the percent injury increased in both products with storage time no clear pattern was observed between cold shocked and control cells (Table 1). The extent of injury for both control and cold shocked cells at 7 days storage appeared higher in pork than in beef; however, in both cases the magnitude was less than 1 log 10 . In contrast to ground beef or pork, cold shocking E. coli in milk resulted in significant (P#0.05) increases in survival among strains during the entire storage period. A typical survival profile is presented in Fig. 3. The log 10 cfu / ml increase in survival levels as a result of the cold shock treatment among strains ranged from 0.51 to 1.89 and from 0.40 to 2.50 on TSA and VRBA, respectively (Table 2). Overall, the percent injury appeared slightly higher in milk and sausages than in ground beef or pork (Table 1). Although the level of injury increased with time, no clear pattern was observed between controls and cold shocked cells. In all cases injury levels were less than 90%. Table 2 Enhanced survival (log 10 cfu) of E. coli strains in food products at 28 days of storage (2208C) as a result of cold shock treatment Product

Strain

TSAa

VRBA

Milk

7128 7282 7110 7283 7174 7236 my20

0.51 1.58 1.89 1.52 1.68 0.98 0.89(1.29)b

0.40 1.13 1.66 0.89 2.50 1.80 1.44(1.40)

Egg

7128 7282 7110 7283 7174 7236 my20

0.97 0.41 0.64 0.51 0.51 0.29 0.68(0.57)

1.84 0.89 1.31 1.05 1.06 1.05 1.40(1.23)

Sausage

7128 7282 7110 7283 7174 7236 my20

1.34 0.66 0.74 0.71 1.53 0.85 1.22(1.00)

1.50 1.41 1.19 1.05 1.98 1.21 1.68(1.43)

52.168.8 NS CS NS CS NS CS

75.263.9 70.663.0 93.961.3 62.7629.1 94.462.1 81.2614.2

a

Non-cold shocked (Control). Cold shocked (108C, 1.5 h). c 6S.D. b

between cold shocked and control cells for these products, enumerated on either TSA or VRBA, were not significant (P#0.05). Similar results were obtained for the remaining strains in these products and therefore are not presented. Over the 28 days storage

a

Log 10 cfu CS2(Log 10 cfu NS). Values in parentheses are the means for each column for each food. b

J. Bollman et al. / International Journal of Food Microbiology 64 (2001) 127 – 138

133

Fig. 4. Effect of cold shocking on the survival of E. coli (7110) in whole egg at 2208C. Results with the same superscript for each medium and storage time are not significantly different (P.0.05). Bars represent averages of triplicate results each performed in duplicate (n566S.D.). Non-cold shocked (NS); cold shocked (CS); tryptic soy agar (TSA); violet red bile agar (VRBA).

Fig. 5. Effect of cold shocking on the survival of E. coli (7110) in sausage at 2208C. Results with the same superscript for each medium and storage time are not significantly different (P.0.05). Bars represent averages of triplicate results each performed in duplicate (n566S.D.). Non-cold shocked (NS); cold shocked (CS); tryptic soy agar (TSA); violet red bile agar (VRBA).

134

J. Bollman et al. / International Journal of Food Microbiology 64 (2001) 127 – 138

Significant protection resulting in survival of E. coli strains in whole egg was also observed throughout the storage period as a result of the cold shock treatment. A typical survival profile is illustrated in Fig. 4. By 28 days, the log 10 increase in survival as a result of cold shocking ranged from 0.29 to 0.97 and from 0.89 to 1.84 on TSA and VRBA, respectively (Table 2). As shown in Table 1, the percent injury in whole egg appeared higher for the controls at all sampling periods. At 28 days, the difference in injury level between cold and non-cold shocked cells was about 61%. Similar to previous products, the level of injury increased with storage time. Cold shocking E. coli in fresh raw sausage also resulted in protection. A typical survivor profile is presented in Fig. 5. By 28 days, the survivor levels for cold shocked E. coli strains ranged from 0.66 to 1.53 log 10 and from 1.05 to 1.98 log 10 higher on TSA and VRBA, respectively, compared to the controls (Table 2). In contrast to previous products, particularly ground beef and pork, the percent injury appeared relatively high, especially at day 1 of storage (Table 1). Although the percent injury increased during storage, a clear pattern between control and cold shocked cells was not observed, in part due to several observations with relatively high standard deviations. Overall, there was a significant (P#0.05) increase in the survival of all cold shocked strains when recovered on TSA (Table 3). By 28 days the survival of some cold shocked strains was approximately 2–3 log 10 higher than controls. Survival levels, particularly for some of the controls were less than 1 log 10 by 7 days. By 28 days the levels for all control strains recovered on VRBA was ,1 log 10 cfu / ml. In contrast, cold shocking E. coli in acidified TSB (pH 5) when compared to controls, resulted in a significant (P,0.05) decrease in survivors for several strains (Table 4). This effect was observed on both recovery media as early as day 1 (results not presented). By 7 days the survival of many strains approached 1 log 10 cfu / ml. Strains 7128 and 7282 appeared the most and least acid resistant, respectively.

3.2. Proximate analysis Proximate analysis was performed on all food products with the exception of UHT milk. Of the

Table 3 Survival of E. coli in TSB (pH 7.0) at 2208C Survivors (log 10 cfu / ml)a

Storage time (days)

E. coli (strain)

0

7128 7282 7110 7283 7174 7236 my20

5.8160.26 5.6360.03 5.4560.12 5.8360.27 5.6660.05 5.6860.04 5.5360.11

5.8560.23 5.6860.06 5.5260.08 5.8660.22 5.7260.03 5.6860.05 5.1660.07

7

7128 NS b 7128 CS c

3.3760.06 x 4.2260.07

,2.00 x 0.6360.59 y

7282 NS 7282 CS

0x 2.8360.07 y

0x 1.7760.07 y

7110 NS 7110 CS

,1.00 x 2.9060.07 y

0x 2.2360.59 y

7283 NS 7283 CS

2.7760.40 x 3.4860.16 y

7141 NS 7141 CS

,1.00 x 3.4860.12 y

0x 3.0360.01 y

7236 NS 7236 CS

1.1760.22 x 3.7160.03 y

0x 3.2860.12 y

my20 NS my20 CS

2.8660.03 y 4.1460.15 y

7128 NS 7128 CS

2.1660.12 x 3.2260.14 y

7282 NS 7282 CS

0x 2.2560.07 y

0 0

7110 NS 7110 CS

0.8760.29 x 2.3860.20 y

0 0

7283 NS 7283 CS

1.4660.06 x 3.0960.06 y

0x 0.7860.42 y

7174 NS 7174 CS

0x 2.9760.10 y

0x 1.2960.26 y

7236 NS 7236 CS

0x 3.0860.03 y

0x 1.7060.17 y

my20 NS my20 CS

1.3560.31 x 3.4760.21 y

28

a

TSA

VRBA

,2.00 ,2.00

,2.00 ,2.00 0x 2.0760.17 y

0 0

Results are means of triplicate determinations, each performed in duplicate (n566S.D.). Results in columns followed by the same superscript for each strain and storage time are not significantly different (P.0.05). b Non-cold shocked (Control). c Cold shocked. 0 indicates less than 10 cfu / ml; ,1.00 and ,2.00 indicate 10 21 and 10 22 were the lowest dilutions plated, respectively.

J. Bollman et al. / International Journal of Food Microbiology 64 (2001) 127 – 138 Table 4 Survival of E. coli in TSB (pH 5) at 2208C Survivors (log 10 cfu / ml)a

Storage time (day)

E. coli (strain)

TSA

VRBA

0

7128 7282 7110 7283 7174 7236 my20

5.7560.08 5.3560.08 5.5360.03 5.7460.08 5.5560.07 5.6760.03 5.4960.09

5.7360.07 5.3660.07 5.5760.05 5.7160.03 5.5960.01 5.7160.05 5.4960.13

7

7128 NS b 7128 CS c

3.8960.17 x 3.7660.19 x

2.6860.43 x 2.3360.06 x

7282 NS 7282 CS

0 0

0 0

7110 NS 7110 CS

1.3460.17 x 0y

0 0

7283 NS 7283 CS

2.2160.35 x 2.9160.14 y

0 0

7174 NS 7174 CS

2.0360.33 x 0y

0 0

7236 NS 7236 CS

3.2260.05 x 2.0360.33 y

1.0260.48 x 0y

my20 NS my20 CS

3.0460.16 x 3.5460.34 x

0 0

135

(30–15, 30–10 and 37–108C). Also, differences in survival appeared greater as the temperature differential increased (Table 5). These increases were observed on both recovery media and at each sampling period. In this regard maximal enhancement of survival resulting from the cold shock treatment was observed with the 37–108C protocol. Conversely, no significant enhancement of survivors due to cold shocking was observed with the 20–108C shift. Interestingly, the latter temperature combination did result in significantly higher numbers of survivors in the controls, at least for the first day of storage. Overall, no consistent pattern was observed between cold shocked and control cells regarding the percent injury. In part, the relatively high standard deviations of some of the data precluded a clear interpretation.

3.4. Isoelectric focusing The isoelectric focusing patterns of the cold and non-cold shocked E. coli in TSB are given in Fig. 6. A novel single band with a pI of 4.8 was produced consistently under cold shocking conditions.

4. Discussion

a

Results are means of triplicate determinations, each performed in duplicate (n566S.D.). Results in columns followed by the same superscript for each strain are not significantly different (P.0.05). b Non-cold shocked (Control). c Cold shocked. 0 indicates less than 10 cfu / ml.

samples analyzed, moisture level was highest and lowest in whole egg and sausage (54.4 and 73.8%), respectively. Ground beef and sausage each contained the highest level of fat (26%); the latter product also contained the highest ash concentration (2.4%). Ground beef and ground pork contained the highest levels of protein (16 and 17%, respectively). Both milk and whole liquid egg had pH values near neutral while ground beef and pork had pH values below 6. The pork sausages had a pH of 6.08.

3.3. Temperature shift to induce cold shock Survival was significantly greater for treated cells within each tested differential greater than 108C

Cold shocking significantly increased survival of E. coli O157:H7 in frozen milk, liquid eggs and raw sausage, but not in frozen ground beef or ground pork. Based on the results of this study, enhanced survival appeared to be linked, in part to substrate composition. For example, in both ground beef and ground pork the initial population of the controls for all strains decreased an average of only 1 and 0.77 log 10 , respectively, over 28 days of frozen storage (calculated using TSA data). In contrast the decrement in control survivors in both milk and sausage over the same time period was about double (2.07 and 2.16 log 10 , respectively) while in TSB (pH 7), on average, was over 4 log 10 . It would appear therefore that E. coli normally survives better in ground beef and pork than in milk, sausage and broth. Although survival of E. coli O157:H7 in frozen, ground muscle meats is relatively high (Doyle and Schoeni, 1984; Conner and Hall, 1994) factors contributing to its survival are not clear but may be related to the protein content of the food. Proteins exert a cryoprotective effect (Speck and

136

J. Bollman et al. / International Journal of Food Microbiology 64 (2001) 127 – 138

Table 5 The effect of growth-cold shock temperature combinations on survival of E. coli 7282 in frozen (2208C) milk a Growth cold shock temperature

Storage time / treatment (days)

37–108C

0 1 NS CS d

30–108C

30–158C

TSA

VRBA

5.1360.1 c

% Injury

5.0760.1 x

14.266.8 x

3.4760.1 4.9060.0 y (1.43)

2.9560.1 4.5060.1 y (1.55)

69.762.5 59. 667.8

7 NS CS

3.3560.2 x 4.7060.0 y (1.35)

2.5860.2 x 4.1660.1 y (1.58)

82.963.7 70.563.4

0

5.3560.0

5.2960.0 x

1.3610.2 x

1 NS CS

3.6160.2 4.8460.1 y (1.23)

3.6160.2 4.3960.1 y (1.22)

64.363.8 64.363.8

7 NS CS

3.5960.2 x 4.6860.0 y (1.09)

2.9760.1 x 4.3260.0 y (1.35)

70.862.1 56.462.4

0

5.2660.1

5.2360.1

1 NS CS 7 NS CS 20–108C

Survivors (log 10 cfu / ml)b

0

x

3.5760.1 4.6160.1 y (1.04) 3.1660.1 x 41.660.1 y (1.00) 5.4060.1 x

6.863.5 x

3.2660.2 4.1760.6 y (0.91)

49.3613.9 20.7613.9

2.8360.1 x 4.0260.0 y (1.19)

53.362.8 2.6367.0

5.3260.2

18.062.0 x

1 NS CS

4.5960.1 4.4760.1 y (0.12)

4.5560.1 4.2460.1 y (0.31)

14.267.3 32.265.6

7 NS CS

3.8560.0 x 3.9260.0 x

3.5860.2 x 3.4660.1 x

46.060.0 64.5610.2

a

Numbers in parentheses represent differences between cold and non-cold shocked values; given only for significant differences. Results are means of triplicate determinations, each performed in duplicate, n566S.D. Values in columns for each growth-cold shock temperature at each sampling time followed by the same superscript are not significantly different (P.0.05). c Non-cold shocked. c Cold shocked. b

Ray, 1977) ostensibly by forming hydrogen bonds with microbial proteins, thus stabilizing and protecting them against denaturation (Christophersen, 1968). In this study both ground beef and ground pork contained the highest levels of protein. Cold shocking effects on E. coli in these products may therefore have been overshadowed by the presence of a more favorable survival environment. In contrast, the effectiveness of cold shock treatment of E. coli in either milk, sausage or broth may be due to differences in cryoprotectant levels or the physical nature of the food matrix. It is possible that at least in the case of milk and broth, extracellular crystallization of water coupled with osmotic dehydration

may have been major stress factors which far outweighed the presence of cryoprotective agents including proteins. Alternatively, preservatives (salt and / or spices) in the breakfast sausage may have had a sensitizing effect on the cells towards freezing (Speck and Ray, 1977). This may also explain why this product, with a fat content similar to ground beef, did not provide E. coli equal protection. In recovering E. coli in frozen ground chicken meat, Conner and Hall (1994) also reported that the presence of salt severely reduced survivor numbers. The explanation that cold shocking appeared more effective in foods where E. coli had a poor survival rate can not be readily applied to either whole egg or

J. Bollman et al. / International Journal of Food Microbiology 64 (2001) 127 – 138

Fig. 6. Isoelectric Focusing of cold and non-cold shocked E. coli 0157:H7 (strain 7110) in TSB. Lanes: (A,D) standards; (B) non-cold shocked; (C) cold shocked.

TSB (pH 5). In the former product the decrease in the control population (0.79 log 10 ; calculated using TSA data) was comparable to that of ground pork (0.77 log 10 ; calculated using TSA data). Yet, despite the similarity in recovery levels, the protective effect of cold shock treatment in egg was significant for every strain. Interestingly, the level of protection due to cold shock treatment on TSA for this product was the lowest (0.57) compared to milk (1.29), sausage (1.00) and TSB, pH 7 (4.82). Also interesting is the observation that the average recovery levels of E. coli were higher for milk, sausage and whole egg on VRBA compared to TSA. For the latter product, the number of survivors more than doubled when enumerated on VRBA. Since this medium is selective in that it contains bile salts and crystal violet, it is generally assumed that many freeze-injured E. coli would fail to grow (Speck and Ray, 1977). Although the effect of substrate pH on cold shocking was not

137

specifically studied, it is interesting to note that all food products including TSB having a pH below 6 did not afford E. coli protection. In contrast, milk, egg, TSB and sausage having pH values above 6, all rendered E. coli protection. In TSB (pH 5) the results are somewhat unique in that the majority of the control strains exhibited significantly higher survival compared to strains which were cold shocked. In general, the survival of bacteria has been reported to depend on the pH of the substrate in which they are frozen; over a pH range of 2.6–5.0, the lethal effect increases as the pH decreases (Partman, 1975). Since the freezing protocol was identical for both control and treatment cells, it would appear that cold shocking per se contributed to the rapid reduction in survivors of some strains during freezing. Prompt freezing of cells as in the case of the controls, however, appeared to lessen the adverse effects in some strains caused by exposure to acid. E. coli O157:H7 strains are known to be aciduric and have been reported to exhibit enhanced survival in acidified TSB broth even at 108C (AbdulRaouf et al., 1993; Conner and Kotrola, 1995). Therefore, it is somewhat puzzling to explain why maintenance in an acid environment would sensitize E. coli to freezing. All downshifts in temperature with the exception of 20–108C provided some level of protection. Overall, the magnitude of protection appeared to decrease with a decrease in the range of the downshift temperature. Similar findings were reported by Jones and Inouye (1994). It is generally recognized that when cells are temperature shifted certain biochemical changes take place during their growth. These changes, of which many are known and primarily target the cell membrane (Rose, 1968) may also minimize or mask the cold shock response. With regards to E. coli, Marr and Ingraham (1962) reported that the proportion of unsaturated fatty acids in the cell membrane increased as the growth temperature decreased presumably as a means of maintaining membrane function. Paton et al. (1978) also reported that injury caused by cooling Bacillus amyloliquefaciens could be reduced if cells were grown at low temperatures. Recently, Jones and Inouye (1994) reported that when cells were grown at 308C and instantly chilled to 38C total viability was lost, however, cells grown at 208C could be chilled to 38C with almost no change in viability. It

138

J. Bollman et al. / International Journal of Food Microbiology 64 (2001) 127 – 138

should be pointed out that the temperatures chosen for cold shocking E. coli in this study were based on the fact that below 108C synthesis of cold shock proteins would be inhibited (Jones and Inouye, 1994). Alternatively, above 158C it is suspected that the population would have increased during cold shocking, thereby confounding the results. At 158C, the population of E. coli remained stable (results not included) during cold shocking (1.5 h). Statistical analysis to determine significant differences in percent injury between cold and non-cold shocked cells were not conducted because of the high variability between replicates. In general, cold shocking did not appear to reduce the incidence of injury, at least the type that could be discerned by selective plating on VRBA. In this study, isoelectric focusing was used to detect changes in protein synthesis in E. coli with cold shocking. A novel single band (pI 4.8) was detected that was produced consistently under cold shocking conditions. Although this protein did not correspond to CspA (pI 5.9; Goldstein et al., 1990) its presence confirmed that the protocol used to induce the cold shock response did cause an alteration in the pattern of protein synthesis. In conclusion, the results of this study indicated that cold shocking E. coli O157:H7 strains in specific foods resulted in freeze injury protection. Whether this phenomenon is of significant importance in regards to food safety will have to be further investigated. Specifically, cooling and freezing protocols, which are used in the food industry, will have to be examined and evaluated to determine whether the potential for cold shock-mediated survival of microorganisms exists. Overall there is a greater need to understand the impact of stress response on the physiology of pathogenic microorganisms in foods, particularly in regard to their survival.

References AOAC, 1990. In: Official Methods of Analysis, 15th Edition. Association of Official Analytical Chemists, Arlington, VA. Abdul-Raouf, U.M., Beuchat, L.R., Ammar, M.S., 1993. Survival and growth of Escherichia coli O157:H7 in ground, roasted beef as affected by pH, acidulants and temperature. Appl. Environ. Microbiol. 59, 2364–2368. Berry, E.D., Foegeding, P.M., 1997. Cold temperature adaptation and growth of microorganisms. J. Food Prot. 60, 1583–1594.

Christophersen, J., 1968. Effect of freezing and thawing on the microbial populations of foodstuffs. In: Hawthorn, J., Rolfe, E.J. (Eds.), Low Temperature Biology of Foodstuffs. Permagon Press, Oxford, pp. 251–269. Cloutier, J., Prevost, D., Nadeau, P., Antoun, H., 1992. Heat and cold shock protein synthesis in arctic and temperate strains of rhizobia. Appl. Environ. Microbiol. 58, 2846–2853. Conner, D.E., Hall, G.S., 1994. Efficacy of selected media for recovery of Escherichia coli O157:H7 from frozen chicken meat containing sodium chloride, sodium lactate or polyphosphate. Food Microbiol. 11, 337–344. Conner, D.E., Kotrola, J.S., 1995. Growth and survival of Escherichia coli O157:H7 under acid conditions. Appl. Environ. Microbiol. 61, 382–385. Doyle, M.P., Schoeni, J.L., 1984. Survival and growth characteristics of Escherichia coli associated with hemorrhagic colitis. Appl. Environ. Microbiol. 48, 855–856. Feeney, R.A., Yeh, Y., 1998. Antifreeze proteins: current status and possible food uses. Trends Food Sci. Technol. 9, 102–106. Goldstein, J., Pollitt, N.S., Inouye, M., 1990. Major cold shock protein of Escherichia coli. Proc. Natl. Acad. Sci. USA 87, 283–287. Grauman, P., Marahiel, M.A., 1994. The major cold shock protein of Bacillus subtilis CspB binds with high affinity to the ATTGG-CCAAT sequences in single stranded oligonucleotides. FEBS Lett. 338, 157–160. Graumann, P., Schroder, K., Schmid, R., Marahiel, M.A., 1996. Cold shock stress-induced proteins in Bacillus subtilis. J. Bacteriol. 178, 4611–4619. Jones, P.G., Inouye, M., 1994. The cold shock response — a hot topic. Mol. Microbiol. 11, 811–818. Jones, P.G., VanBogelen, R.A., Neidhardt, F.C., 1987. Induction of proteins in response to low temperature Escherichia coli. J. Bacteriol. 169, 2092–2095. Marr, A.G., Ingraham, J.L., 1962. Effect of temperature on the composition of fatty acids in Escherichia coli. J. Bacteriol. 84, 1260–1267. Partman, W., 1975. The effects of freezing and thawing on food quality. In: Duckworth, R.B. (Ed.), Water Relations of Foods. Academic Press, London, pp. 505–537. Paton, J.C., McMurchie, E.J., May, B.K., Elliot, W.H., 1978. Effect of growth temperature on membrane fatty acid composition and susceptibility to cold shock of Bacillus amyloliquefaciens. J. Bacteriol. 135, 754–759. Rose, A.H., 1968. Physiology of microorganisms at low temperatures. J. Appl. Bacteriol. 31, 1–11. Speck, M.L., Ray, B., 1977. Effects of freezing and storage on microorganisms in frozen foods: a review. J. Food Prot. 40, 333–346. Watson, K., 1990. Microbial stress proteins. In: Advances in Microbial Physiology, Vol. 31. Academic Press, New York, pp. 184–216. Winter, A., Anderson, V., 1997. Analytical electrofocusing in thin layers of polyacrylamide gels. LKB Application Note No. 250. Whyte, L.G., Inniss, W.E., 1992. Cold shock proteins and cold acclimation proteins in a psychrotrophic bacterium. Can. J. Microbiol. 38, 1281–1285. Willimsky, G., Bang, H., Fisher, G., Marahiel, M.A., 1992. Characterization of cspB, a Bacillus subtilis inducible cold shock gene affecting cell viability at low temperatures. J. Bacteriol. 170, 6326–6335.