INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Examination of frozen foods for the presence of pathogenic bacteria has been increasing in recent years because food service operations and consumers use frozen foods and food ingredients frequently. Furthermore, food samples are often frozen as test samples for investigations of food poisoning. Selective reagents are often used for enrichment culturing of food samples, including frozen food samples, because these reagents are required for preserving small numbers of the target bacteria by suppressing the growth of other contaminating bacteria. However, it has been observed that these reagents can inhibit the growth of injured pathogens (7). Thus, a method that both resuscitates injured target bacteria and suppresses the growth of other contaminating bacteria is required to isolate pathogens from food samples that may be contaminated with injured target bacteria. Since Escherichia coli O157:H7 was recognized as a food-poisoning agent in 1982, there have been many outbreaks linked to ingestion of not only beef but also vegetables and fruits, including lettuce, cantaloupe, cabbage, alfalfa sprouts, radish sprouts, and apple juice (2, 4, 14, 15, 23, 25; M. Ackers, B. Mahon, E. Leahy, T. Damrow, L. Hutwagner, T. Barrett, W. Bibb, P. Hayes, P. Griffin, and L. Slutsker, Abstr. 36th Intersci. Conf. Antimicrobial Agents Chemother., abstr. K43, 1996). Many selective enrichment broth media have been used for isolation of E. coli O157:H7 from foods (5, 6, 8, 17). We have shown previously that an enrichment method in which modified E. coli broth supplemented with bile salts and novobiocin (mEC+n) (16) is used is better than other methods for isolating E. coli O157:H7 from beef and radish sprouts artificially contaminated with the organism (10). However, we later found that resuscitation performed with nonselective broth media prior to selective enrichment is effective for isolating E. coli O157:H7 from foods that are artificially contaminated with freeze-injured E. coli O157:H7 cells. In order to develop an effective enrichment method for frozen foods that may be contaminated with injured cells, we first examined whether E. coli O157:H7 cells in foods are injured by freezing of the foods and then tried to isolate E. coli O157:H7 from foods that were contaminated with freeze-injured cells.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Comparison of freeze injuries of five E. coli O157:H7 strains. Five E. coli O157:H7 strains (strains 212, 970056, ATCC 43889, ATCC 43890, and ATCC 43894) were used to compare freeze injuries in different strains (Fig. 1). Strains 970056 and 212 were isolates obtained from beef and a patient in Japan, respectively. The five E. coli O157:H7 strains were grown overnight at 35°C on tryptic soy agar (TSA) (Difco, Detroit, Mich.). Colonies were suspended to a turbidity equivalent to a no. 4 McFarland standard in 5 ml of chilled sterilized reagent grade water obtained with a Milli-Q Plus filter (Nihon Millipore Ltd., Tokyo, Japan) and were sedimented by centrifugation at 2,500 × g for 20 min. The cells were washed three times with reagent grade water and finally were resuspended in reagent grade water at a density of 10⁴ or 10⁶ CFU/ml. After the cells were kept in a freezer at 20°C for 24 h and then thawed, a cell suspension or a dilution of a cell suspension was spread onto TSA and sorbitol MacConkey agar (Oxoid, Unipath Ltd., Hampshire, United Kingdom) supplemented with cefixime (0.05 mg/liter) and potassium tellurite (2.5 mg/liter) (CT-SMAC). The number of freeze-injured E. coli O157:H7 cells was estimated by subtracting the number of CFU on CT-SMAC (a selective medium) from the number of CFU on TSA (a nonselective medium). After 18 h of incubation at 37°C, the numbers of colonies on the media were counted. The percentage of injured cells was calculated by dividing of the number of injured cells by the number of noninjured cells plus the number of injured cells.

View larger version (20K): [in this window] [in a new window]   FIG. 1.   Method used to compare freeze injuries in five E. coli O157:H7 strains.

Detection of freeze-injured or noninjured E. coli O157:H7 cells in various frozen foods inoculated with E. coli O157:H7. Strains 970056 and 212 were used to enumerate freeze-injured or noninjured E. coli O157:H7 cells in frozen foods (Fig. 2). Each strain of E. coli O157:H7 was cultured in tryptic soy broth (TSB) (Difco) at 37°C for 18 h. Each culture was diluted with phosphate-buffered saline (PBS) so that the concentration was 10⁸ CFU/ml. Portions (0.1 ml) of the 10⁸-CFU/ml suspension were inoculated into 25-g portions of various food samples, including samples of sliced cabbage, salted (3%) sliced cabbage, sliced cucumbers, raw ground beef, milk, boiled potatoes, grated radishes, fresh tomatoes, fresh seaweed, fresh strawberries, and vegetable juice. The inoculated food samples were then stored in a freezer at 20°C for 2 to 16 weeks. These samples were used to enumerate freeze-injured and noninjured E. coli O157:H7 cells in frozen foods.

View larger version (25K): [in this window] [in a new window]   FIG. 2.   Method used to detect freeze-injured or noninjured E. coli O157:H7 cells in various frozen foods inoculated with E. coli O157:H7.

Growth of artificially freeze-injured cells. Strain 970056 was used to prepare artificially freeze-injured E. coli O157:H7 (Fig. 3). After a cell suspension (10⁶ CFU/ml) prepared by the method that was used for comparisons of freeze injuries in five E. coli O157:H7 strains was kept in a freezer at 20°C for more than 24 h, there were more than 10³ CFU per ml on TSA but no cells on CT-SMAC.

View larger version (17K): [in this window] [in a new window]   FIG. 3.   Method used to grow artificially freeze-injured cells.

Detection of E. coli O157:H7 in food extracts or food samples inoculated with artificially freeze-injured E. coli O157:H7 cells. In order to assess the efficiency of a resuscitation step in which we used nonselective broth to recover freeze-injured cells in food samples, we attempted to isolate E. coli O157:H7 from food extracts and foods inoculated with freeze-injured E. coli O157:H7 by a procedure that included incubation in nonselective broth (Fig. 4). Radish sprouts and strawberries purchased from retail shops and 3% salted cabbage prepared by adding NaCl to sliced cabbage from retail shops were used as the food samples. The extracts of these foods, which were prepared by homogenizing the foods in equal amounts (1:1, wt/vol) of sterilized distilled water with a stomacher, were used as food extracts. TSB and modified E. coli broth without bile salts (b-mEC) (20 g of peptone, 5 g of lactose, 4 g of K₂HPO₄, 1.5 g of KH₂PO₄, 5 g of NaCl, 1 liter of distilled water; Eiken Co. Ltd., Tokyo, Japan) were used as resuscitating broth media. mEC+n, which was prepared by adding sodium novobiocin (20 mg/liter) (Sigma Chemical Co. Ltd., St. Louis, Mo.) and bile salts no. 3 (1.12 g/liter) (Difco) to b-mEC, was used as a selective enrichment broth. These broth media were added to the food and food extract samples. Freeze-injured E. coli O157:H7 cells (strain 970056) prepared by the method that was used to grow artificially freeze-injured cells were inoculated into 1-ml portions of food extracts and 25-g portions of food to obtain 20 and 12 to 50 CFU/25 g, respectively, and then 9- and 225-ml portions of each broth were added to the inoculated food extract and food samples, respectively. The foods in the broth media were homogenized with a stomacher prior to incubation. The samples in mEC+n were then incubated statically at 37 or 42°C for 18 h. The samples in b-mEC and TSB were incubated statically at 25°C for 2 h in order to resuscitate injured cells and then at 37 or 42°C for 18 h in the presence of sodium novobiocin and bile salts no. 3, which were added to the b-mEC and TSB cultures in order to obtain concentrations of 20 mg/liter and 1.12 g/liter, respectively. These three enrichment methods were designated the mEC+n, b-mEC (b,n), and TSB (b,n) methods, respectively. The food extract cultures were plated directly onto CT-SMAC and CHROMagar O157, and the food cultures were plated onto the same agar media directly or after immunomagnetic separation (IMS) with Dynabeads anti-E. coli O157 (Dynal, Oslo, Norway) performed according to the manufacturer's instructions. At least three suspected colonies on each plate that formed after 24 h of incubation at 37°C were confirmed to be E. coli O157:H7 colonies with a UNI kit and CLIG agar.

View larger version (19K): [in this window] [in a new window]   FIG. 4.   Method used to detect E. coli O157:H7 in food extracts or food samples inoculated with artificially freeze-injured E. coli O157:H7 cells.

Detection of E. coli O157:H7 in frozen foods samples inoculated with E. coli O157:H7. Strain 970056 was cultured in TSB (Difco) at 37°C for 18 h. The culture was diluted with PBS to obtain a concentration of 10² CFU/ml. Portions (0.1 ml) of the 10²-CFU/ml suspension were inoculated into 25-g portions of grated radishes, strawberries, salted cabbage, and ground beef. The inoculated food samples were then stored in a freezer at 20°C for 2 to 4 weeks or 10 months. These samples were used for detection of E. coli O157:H7 in food samples that were frozen after inoculation with the organism (Fig. 5). The frozen food samples (25 g) were thawed in 225 ml of mEC+n or b-mEC and then homogenized with a stomacher. The samples in mEC+n were incubated at 42°C for 18 h. The samples in b-mEC were incubated statically at 25°C for 2 to 3 h and then incubated at 42°C for 18 h in mEC+n containing sodium novobiocin and bile salts no. 3. The cultures were plated onto CT-SMAC and CHROMagar O157 directly or after IMS. After 24 h of incubation at 37°C, at least three suspected colonies that formed on each plate were confirmed to be E. coli O157:H7 colonies with a UNI kit and CLIG agar.

View larger version (24K): [in this window] [in a new window]   FIG. 5.   Method used to detect E. coli O157:H7 in frozen food samples inoculated with E. coli O157:H7.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Comparison of freeze injuries in five E. coli O157:H7 strains. Strains 970056 and ATCC 43890 but not strain 212, ATCC 43889, or ATCC4 3894 was sensitive to freezing (Table 1).

Detection of freeze-injured or noninjured E. coli O157:H7 cells in various frozen foods inoculated with E. coli O157:H7. Figure 6 shows the occurrence of freeze-injured cells in 11 types of food. The colonies of E. coli O157:H7 on TSA that were visualized on CHROMagar O157 were considered colonies that contained both freeze-injured and noninjured cells, and the colonies of E. coli O157:H7 on CT-SMAC were considered colonies of noninjured cells. The proportion of freeze-injured cells was estimated by subtracting the number of CFU on CT-SMAC from the number of CFU on TSA. The proportion of injured strain 970056 cells increased significantly in salted cabbage, grated radish, and tomato samples within 2 weeks (Fig. 6A). In grated radishes, the size of the population decreased significantly, and a large proportion of the cells was injured within 2 weeks. Although most of the cells in strawberry samples were not injured by week 2, the proportion of injured cells increased significantly by week 16. In vegetable juice samples, the size of the population slowly decreased during storage. The number of injured strain 212 cells did not increase in any sample during the storage period (Fig. 6B).

View larger version (39K): [in this window] [in a new window]   FIG. 6.   Detection of E. coli O157:H7 cells in various frozen foods inoculated with E. coli O157:H7. Symbols: , TSA; , CT-SMAC. An asterisk indicates that the proportion of injured E. coli O157:H7 cells was significant at a level of 95% when the value was compared with the zero-time value by using ANOVA. A dagger indicates that the number of E. coli O157:H7 cells on TSA was significant at a level of 95% when the value was compared with the zero-time value by using ANOVA.

Growth of artificially freeze-injured cells. The number of freeze-treated cells on CT-SMAC increased during 1 h of incubation to a level close to the level on TSA, indicating that resuscitation of freeze-injured cells occurred in TSB within 1 h at 25°C (Fig. 7). The number increased only slightly from 1 to 3 h under the same conditions (Fig. 7). Identical results were obtained when we used b-mEC instead of TSB (data not shown).

View larger version (16K): [in this window] [in a new window]   FIG. 7.   Colony counts on TSA and CT-SMAC for untreated and freeze-treated E. coli O157:H7 grown in TSB at 25°C. Symbols: , freeze-treated cells on TSA; , nontreated cells on TSA; , freeze-treated cells on CT-SMAC; , nontreated cells, on CT-SMAC.

Detection of E. coli O157:H7 in food extracts or food samples inoculated with artificially freeze-injured E. coli O157:H7 cells. E. coli O157:H7 was isolated from 2 to 8 out of 10 samples of each food extract after 18 h of incubation at 42°C in selective broth media after 2 h of incubation at 25°C in nonselective broth media, whereas the organism was not isolated from any food extract sample after 18 h of incubation at 37 and 42°C in mEC+n without 2 h of incubation in nonselective broth media (Table 2). Enrichment culturing at 37°C after 2 h of incubation in nonselective broth media was less efficient for recovering E. coli O157:H7 than enrichment culturing at 42°C was (Table 2).

Detection of E. coli O157:H7 in frozen food samples inoculated with E. coli O157:H7. Enrichment with the b-mEC (b,n) method was more effective than enrichment with the mEC+n method for recovering E. coli O157:H7 in grated radishes and strawberries inoculated with ca. 8 CFU of E. coli O157:H7 (range, 5 to 13 CFU). The organism was isolated from 27.8% of the grated radish samples by enrichment with the b-mEC (b,n) method but not by enrichment with the mEC+n method (Table 4). However, the organism was isolated from all samples by enrichment with the mEC+n method when the amount of E. coli O157:H7 inoculated was approximately 163 CFU, (range, 100 to 210 CFU). When strawberry samples were inoculated with ca. 8 CFU of E. coli O157:H7, the organism was isolated from 14.3% of the samples by enrichment with the b-mEC (b,n) method and from 5.6% of the samples by enrichment with the mEC+n method (Table 4). With salted cabbage samples, enrichment in mEC+n resulted in recovery of more E. coli O157:H7 than enrichment with the b-mEC (b,n) method resulted in E. coli O157:H7 was recovered from all ground beef samples inoculated with ca. 4 CFU of the organism (range, 1 to 7 CFU) after enrichment in mEC+n (Table 4).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial cells are injured by being heated or frozen or by other stresses (13, 15, 18). Such injury creates an important problem for detection of pathogens. The injured bacteria may not grow in the presence of selective reagents in enrichment broth that support selective growth of a small number of target bacteria in the presence of other competing bacteria. For E. coli O157:H7, Stephens and Johnson (22) reported that bile salts and antibiotics inhibited the growth of acid- or salt-stressed cells. Therefore, a resuscitation procedure may be required before enrichment culturing in order to successfully isolate the target bacteria in food samples when it is thought that the target bacteria in the samples may have been injured. For resuscitation in enrichment broth, it has been reported previously that freeze-injured E. coli cells are recovered by incubation in TSB at 25°C for 1 to 6 h (11, 14, 19, 24). Freeze-injured E. coli O157:H7 has been recovered on plating media (7, 8, 20), but recovery in liquid medium has not been reported yet. In this study we tried to detect freeze-injured E. coli O157:H7 in foods by resuscitating freeze-injured cells with liquid media.

We demonstrated the difference in the freeze injuries of five E. coli O157:H7 strains in our first experiment. Based on the results, a strain that was sensitive to freezing and a strain that was not sensitive to freezing were used to detect freeze-injured or noninjured E. coli O157:H7 cells in various frozen foods inoculated with E. coli O157:H7.

The fate of E. coli O157:H7 in foods during frozen storage has not been determined except for ground beef or meat products (1, 9, 21). In the present study, we examined E. coli O157:H7 injury caused by freezing in various foods, such as vegetables and fruits. Our results showed that E. coli O157:H7 cells were injured in some foods, although not in all foods. Thus, we investigated suitable methods of recovering freeze-injured E. coli O157:H7 cells from foods by using a strain that was sensitive to freezing.

Artificially freeze-injured cells were used in this study in order to develop an effective enrichment procedure. The presence of even a few noninjured cells along with the freeze-injured cells might have interfered with the effectiveness of the resuscitation step. Therefore, it was necessary to prepare food samples that contained only freeze-injured cells. Because it is difficult to prepare food samples that contain only freeze-injured cells by freezing foods, we added artificially prepared freeze-injured cells to the foods. First, the length of the resuscitation period for artificially injured cells in broth at 25°C was determined to be 1 to 3 h by performing a growth experiment with artificially freeze-injured cells. Next, it was demonstrated that a resuscitation step consisting of 2 h of incubation at 25°C in nonselective broth prior to selective enrichment was an efficient method for isolating freeze-injured E. coli O157:H7 cells from foods naturally contaminated with numerous competitive bacteria. Selective enrichment at 42°C for 18 h with mEC+n after 2 h of nonselective enrichment with b-mEC at 25°C resulted in good recovery of E. coli O157:H7 from food extract and food samples.

Moreover, we tried to isolate E. coli O157:H7 by this method from foods that were inoculated with the organism and then kept in a freezer at 20°C. E. coli O157:H7 was recovered from grated radishes by enrichment culturing in selective medium after resuscitation in nonselective medium. Enrichment with resuscitation improved recovery of E. coli O157:H7 from strawberries and grated radishes, but a similar improvement was not observed with salted cabbage. The low detection rates for these samples (14.3 to 27.8%) may be attributed to death of the organism during frozen storage (Fig. 6A). E. coli O157:H7 was detected in most ground beef samples regardless of whether the resuscitation step was used or not, which presumably showed that in ground beef the organism is not injured by freezing (Fig. 6A).

Thus, our results demonstrated that E. coli O157:H7 is killed or injured in some types of food during frozen storage and that injured E. coli O157:H7 cells in foods can be detected more successfully by selective enrichment culturing after 2 h of incubation of in nonselective broth at 25°C.