Fate of Escherichia coli O157:H7 during the processing and storage of Ergo and Ayib, traditional Ethiopian dairy products

International Journal of Food Microbiology 103 (2005) 11 – 21 www.elsevier.com/locate/ijfoodmicro

Fate of Escherichia coli O157:H7 during the processing and storage of Ergo and Ayib, traditional Ethiopian dairy products Mekonnen Tsegayea, Mogessie Ashenafib,T b

a Department of Biology, Addis Ababa University, P.O. Box 1176, Addis Ababa, Ethiopia Institute of Pathobiology, Addis Ababa University, P.O. Box 1176, Addis Ababa, Ethiopia

Received 30 September 2003; received in revised form 18 May 2004; accepted 30 December 2004

Abstract The long-term survival of E. coli O157:H7 in acid foods is well documented. This prompted us to evaluate the organism’s survival during the making of Ergo, a traditional Ethiopian sour milk and Ayib, a traditional Ethiopian cottage cheese and during storage of these products at both ambient and refrigeration temperatures. E. coli O157:H7 test strains were separately inoculated into milk at initial levels of log10 3 cfu/ml. Levels of E. coli O157:H7 in the absence of lactic acid bacteria (LAB) reached log10 8.4 cfu/ml at 24 h as the pH dropped to 5.6F1.0. In milk inoculated with LAB (log10 6.4 cfu/ml) and E. coli O157:H7 (log10 3 cfu/ml), levels of LAB were log10 9.6 cfu/ml and 9.4 cfu/ml at 24 h and 72 h, respectively, and the pH values were 3.5 and 3.9, respectively. In the presence of LAB, E. coli O157:H7 grew to log10 6.5 cfu/ml at 24 h, with the levels decreasing to log10 3.2 cfu/ml at 72 h. Post-fermentation inoculation of E. coli O157:H7 in Ergo at an initial level of log10 3 cfu/ ml, resulted in complete elimination of test organisms at 6 h at ambient temperature storage, but they were recovered until 72 h at refrigerated storage. At a higher initial inoculum level (log10 6 cfu/ml) of the E. coli O157:H7, the counts decreased by 4 logs within 12 h (pH 4.2) at ambient temperature storage, and complete elimination was observed at 36 h (pH 4.0). At refrigeration temperature, counts at 72 h were between 2.2 and 3.5 log cfu/ml for the different strains. During Ayib processing, E. coli O157:H7 in souring milk increased by 3 logs in 24 h, with a slight reduction being observed at 36 h. The pH dropped to 4.3 during this time. The numbers of E. coli O157:H7, immediately after curd cooking, were below detectable levels, but could be recovered by enrichment. Complete inactivation was observed on 24 h after curd cooking. When E. coli O157:H7 was inoculated into steam-treated Ayib (pH 4.2–4.3) at levels of log10 6.0–6.7 cfu/g and maintained at ambient temperatures, there was a gradual decrease in numbers to 3.7–4.3 logs by day 3. After day 7, pH values were 3.8–3.9 and E. coli O157:H7 was only detectable after enrichment. E. coli O157:H7 strain MF-1847 was completely inactivated by day 9, but the other two strains could still be recovered by enrichment at this time. At refrigeration storage, decrease in count of the test strains was gradual and counts at day 8 and 9 were N4 log cfu/g. As Ergo is preferably consumed soon after (24 h) fermentation, traditional fermentation

T Corresponding author. E-mail address: [email protected] (M. Ashenafi). 0168-1605/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.ijfoodmicro.2004.12.006

12

M. Tsegaye, M. Ashenafi / International Journal of Food Microbiology 103 (2005) 11–21

of Ergo would not guarantee that E. coli O157:H7 can be controlled and, therefore, Ergo can be a potential health hazard if prepared from milk contaminated with E. coli O157:H7. D 2005 Elsevier B.V. All rights reserved. Keywords: E. coli O157:H7; Ergo; Traditional sour milk; Ayib; Cottage cheese; Fermentation; Storage temperatures

1. Introduction Since its identification as a human pathogen in 1982 (Riley et al., 1983), Escherichia coli O157:H7 has become a pathogen of major concern for the food and dairy industries because of its ability to cause severe illness such as haemorrhagic colitis (HC), haemolytic uremic syndrome (HUS), and thrombotic thrombocytopenic purpura (TTP). The diseases affect all age groups and the pathogen is exceptional in its severe consequences of infection, its low infectious dose and unusual acid resistance (Buchanan and Doyle, 1997). Survival of E. coli O157:H7 in acidic foods such as yoghurt (Morgan et al., 1993) and salami (Tilden et al., 1996), and its growth in unpasteurized apple cider (Zhao et al., 1993) prompted food microbiologists to investigate its behavior in different foods. The growth of E. coli O157:H7 cells in apple wounds (Janisiewice et al., 1999) its isolation from frozen strawberry after 30 days of storage (Knudsen et al., 2001) and its survival in fermented as well as fermenting dairy foods for long periods of time (Dineen et al., 1998; McIngvale et al., 2000) makes its hardy nature apparent. Many animals, including cattle, sheep, and goats are known to harbor E. coli O157, however, cattle are most often implicated as the zoonotic source of human infection (Trevena et al., 1996). Many outbreaks of E. coli O157:H7 are usually associated with foods from these animals or fecal contamination of water and vegetables by these animals. Raw milk (Mcdonough et al., 1991) and yoghurt (Morgan et al., 1993) have been implicated in food-borne illnesses caused by E. coli O157:H7. The known fermented animal products in Ethiopia are milk products. The major ones produced by smallholder farmers using traditional methods are Ergo or Ititu (fermented milk), Qibe (traditional butter), Neter Qibe (traditional ghee), Ayib (traditional cottage cheese), Arera (sour defatted milk) and Agwat (whey) (Gonfa et al., 2001). Ergo is a popular

fermented milk product consumed in all parts of the country. In rural areas Ergo can be stored for 15–20 days depending on the temperature while in the urban areas the product may be stored for relatively short periods unless refrigerated (Gonfa et al., 2001). The consistency and the taste of Ergo vary from place to place and from time to time. This is due to the large variation in the initial count and the composition and rate of development of the microflora during souring (Ashenafi, 1995). Traditionally, Ergo is produced from fermented raw milk and, during its fermentation, different types of lactic acid bacteria (LAB) and other microorganisms are involved in the souring process. In one study, coccobacillus shaped lactobacilli were the dominant microflora in most of the samples whereas in some samples streptococci species were dominant (Ashenafi, 1995). Gonfa et al. (1999) found that cocci shaped lactic acid bacteria were dominant and played a major role in the fermentation of Ergo. Ayib is a traditional Ethiopian cottage cheese made from sour milk after the fat is removed by churning. Raw milk is collected in a clay pot and kept in a warm place (about 30 8C) for 24–48 h to sour spontaneously. The pH of sour milk is usually about 4.0. Churning of sour milk is carried out by slowly shaking the contents of the pot until the fat is separated. The fat is then removed and the defatted milk is heated to about 50 8C until a distinct curd mass forms and floats over the whey. Temperature, however, can vary between 40 8C and 70 8C without markedly affecting product composition and yield (O’Mahoney, 1988). Since temperatures higher than 80 8C are reported to give the product a cooked flavor (O’Mahoney, 1988), heat precipitation of curd at 70 8C (pH 4.0) was recommended as it resulted in a wholesome product with low microbial load (Ashenafi, 1990a,b). After gradual cooling, the curd is recovered from the whey and is ready for consumption without any further processing. The pH of Ayib collected from market ranged from 3.3 to 4.6 and contained high counts of aerobic mesophilic bacteria

M. Tsegaye, M. Ashenafi / International Journal of Food Microbiology 103 (2005) 11–21

(log 8 cfu/ml) (Ashenafi, 1990a,b, 1994a,b). However, since cooking of the curd is expected to decrease the count of microorganisms, fresh Ayib is supposed to have a lower microbial load. Its low pH value should also assist in maintaining the low count for a certain period of time. The high microbial load in market Ayib would, thus, come from handlers and plant parts used for packaging and for imparting flavor. Previous studies showed that Salmonella spp. and Listeria monocytogenes could survive during the souring of Ergo for 48 to 60 h (Ashenafi, 1992, 1994a,b) and Staphylococcus aureus and Bacillus cereus for 24 to 48 h (Ashenafi, 1993). As Ergo is preferably consumed soon after 24 h of fermentation, survival of pathogens for more than 24 h would be undesirable. Traditional Ayib processing requires a heat treatment of the curd, and it would be useful to evaluate the safety of the product when prepared from milk contaminated with E. coli O157:H7. The purpose of this study was, therefore, to evaluate the fate of E. coli O157:H7 during traditional souring of Ergo and processing of Ayib and to assess its survival during the storage of the products at both ambient and refrigeration temperatures.

2. Materials and methods 2.1. Bacterial strains Three E. coli O157:H7 strains, 4595-52B, 93-30 and MF-1847, isolated from hamburger meat, were obtained from the Food Microbiology Laboratory, Howard University, USA. These were maintained as stock cultures on Nutrient Agar (Oxoid) slants at refrigeration temperature. The inoculum for each experiment was prepared by transferring the cultures from slants to nutrient broth and allowing them to grow overnight at 37 8C. 2.2. Isolation of lactic acid bacteria For the experimental fermentation of Ergo, lactic acid bacteria were isolated from naturally souring milk. The raw milk was allowed to ferment at ambient temperature (18–23 8C) and appropriate dilutions of the souring milk were surface plated on MRS (Oxoid)

13

medium at regular intervals. MRS plates were incubated anaerobically using anaerobic jar (Oxoid, HP 11, Basingstoke, England) and incubated at 32 8C for 3 days. The colonies on the media were counted and typical LAB colonies were transferred into MRS broth for further purification. Pure colonies were examined for catalase production and microscopic appearance as in Collins et al. (1991). LAB strains consisting of cocci and rods, isolated from naturally fermenting Ergo, were used to prepare experimental souring milk. The strain mixtures were inoculated into sterilized milk (121 8C for 15 min) and the milk was allowed to ferment at ambient temperature overnight. Laboratory personnel tasted the product and confirmed that the sour milk had an acceptable natural flavor and aroma of Ergo. 2.3. Enumeration of LAB starter cultures and E. coli O157:H7 strains in souring milk Raw milk purchased from small dairy farms was dispensed in 200 ml amounts into 250 ml screw capped bottles. The milk was sterilized at 121 8C for 15 min. Appropriate dilutions of overnight cultures of each E. coli O157:H7 strain was separately inoculated into the milk in duplicates to give an initial inoculum level of log10 3 cfu/ml. Milk containing the test strains was also inoculated with LAB starter cultures to give an initial inoculum level of log10 6 cfu/ml. Similarly, control bottles were prepared by separately inoculating duplicate sterilized milk samples only with E. coli O157:H7. Enumeration of E. coli O157:H7 and LAB was done at 0, 12, 24, 36,48, 60 and 72 h. Samples were serially diluted in 0.1% peptone water and appropriate dilutions were surface plated onto Sorbitol MacConkey agar (SMAC) (Oxoid) for counting E. coli O157:H7 and MRS agar for LAB count. SMAC Plates were incubated at 37 8C for 24 h and MRS plates were incubated anaerobically at 32 8C using Oxoid anaerobic jar for 3 days. The pH of fermenting milk was also measured during the sampling times using a digital pH meter. 2.4. Enumeration of the test strains in Ergo Ergo was prepared by inoculating sterilized milk with LAB starter cultures as described above. Fresh

14

M. Tsegaye, M. Ashenafi / International Journal of Food Microbiology 103 (2005) 11–21

Ergo samples were inoculated separately with the three E. coli O157:H7 test strains at an initial inoculum level of 106 cfu/ml. An equal amount of each strain was also separately inoculated in Phosphate Buffer Saline and served as a control. Samples were prepared in duplicates for incubation at refrigerator and ambient temperatures. Samples were analyzed for E. coli O157:H7 at 0, 6, 12, 24, 36 and 48 h for samples stored at ambient temperatures and at 0,24,48, and 72 h for samples stored at refrigeration temperature. Enrichment in Nutrient broth was done when counts were b10 cfu/ml. 2.5. Enumeration of E. coli O157:H7 during Ayib production Appropriate dilutions of LAB starter cultures and separate overnight cultures of E. coli O157:H7 test strains were added to sterilized milk to give an initial inoculum level of log10 6 cfu/ml and log10 3 cfu/ml, respectively. This was allowed to sour at ambient temperatures. At souring, the milk coagulated and the curd was defatted by churning using a shaker at 102-rotation cycles per minute for 1 h. The floating fat was removed aseptically. Bottles containing defatted sour milk were kept in a water bath at a temperature of 70 8C. They were maintained at this condition for 55 min after the internal temperature of defatted milk reached 70 8C. During heating, the curd separated from the liquid and floated over the whey. At completion of curd cooking, the bottles were removed from the water bath and allowed to cool. The curd was separated from the whey aseptically using sterile cheesecloth. The product, Ayib was transferred into a sterile stomacher bag aseptically for convenient mixing during microbiological analysis. Enumeration of E. coli O157:H7 was done at 0, 24, and 36 h for souring milk. Samples were serially diluted in 0.1% peptone water and appropriate dilutions were surface plated on Sorbitol MacConkey agar (SMAC) (Oxoid) for counting E. coli O157:H7. SMAC Plates were incubated at 37 8C for 24 h. Counting of E. coli O157:H7 was also done immediately after the casein was separated from the whey and at 24-h intervals thereafter. Ayib sample (10 g) was homogenized with 90 ml of

0.1% peptone water using a stomacher lab blender for 2 min. Samples were serially diluted and appropriate dilutions were spread-plated on SMAC. Plates were incubated at 37 8C for 24 h. When counts were blog 1 cfu/g, Ayib (1 g) was enriched in Nutrient broth and incubated at 37 8C overnight. Any growth was checked by streak plating on SMAC. 2.6. Determination of survival of E. coli O157:H7 in processed Ayib Samples of Ayib were purchased from the market, dispensed in 200 g amounts into 250 ml screw capped bottles and steam treated for 5–7 min to eliminate the vegetative forms of microbes that contaminate the product. The steam-treated samples were transferred into a sterile stomacher bag aseptically for ease of sample analysis. 2.7. Inoculation of Ayib samples and enumeration of E. coli O157:H7 To study the rate of inhibition of our test strains in the product, E. coli O157:H7 strains were separately inoculated into Ayib at a level of log10 6 cfu/g and blended using a stomacher lab blender for 5 min. Samples were prepared in replicates for incubation at ambient and refrigerated temperatures. E. coli O157:H7 test strains were enumerated at 24-h intervals for 9 days. When counts were blog 1 cfu/ g, Ayib (1 g) was enriched in Nutrient broth and incubated at 37 8C overnight. Any growth was checked by streak plating on SMAC. 2.8. Determination of pH The pH of fermenting and fermented products was measured at every sampling time during the fermentation by dipping the electrode of the pH meter into portions of samples in a test tube. 2.9. Statistical analysis Experiments were conducted four times. Counts are, thus, average values. Coefficient of variation (C.V) was calculated to see significant differences between individual values at any sampling time.

M. Tsegaye, M. Ashenafi / International Journal of Food Microbiology 103 (2005) 11–21

15

3. Results and discussion 3.1. Survival of E. coli O157:H7 in fermenting Ergo In the control milk, where no LAB was added, the three E. coli O157:H7 test strains grew well. They reached counts of log 8.4 cfu/ml at 24 h, while the counts at 72 h were between log 7.7 and 8.1 cfu/ml. There was no significant variation in the counts of each strain at any particular sampling time in the four experiments (CVb10%). The growth of the test strains was accompanied by a gradual decrease in pH from an initial mean value of 6.45 to mean values between 5.6F1.0 at 24 h and to around 4.8 at 72 h. Variation in pH values during sampling times was also not significant (b10%) (Fig. 1). In milk co-inoculated with LAB and the test strains, LAB increased from initial mean counts of log 6.41 to log 9.64 cfu/ml at 24 h and log 9.29 cfu/ml at 72 h. This resulted in a faster drop in pH to values of 4.5 at 24 h and 3.8–3.9 at 72 h. The three E. coli O157:H7 test strains also showed growth in LABfermenting Ergo, reaching mean counts of log 6.47– 6.58 cfu/ml within 24 h and gradually decreasing thereafter to log 3.06–3.43 cfu/ml at 72 h (Fig. 1). Counts of LAB and all E. coli O157:H7 strains did not have significant variation at any sampling time (CVb10%), except at 72 h, where counts of strains 4595-52B and 93-30 showed variations as high as 21% and 25%, respectively (data not shown). All E. coli O157:H7 strains showed similar growth and inactivation pattern in fermenting and control milk. It is worth noting, however, that complete inactivation of the test strains was not achieved and that levels of E. coli O157:H7 in fermenting dErgoT were never less than log 3 cfu/ml, even at 72 h. Rate of growth of E. coli O157:H7 in our study was lower than that observed in the fermentation of traditional African yoghurt at 25 8C (Ogwaro et al., 2002). At 24 h, the pH dropped from 6.5 to 5.7 during growth of E. coli O157:H7 strains alone and to 4.5 during growth of our test strains in the presence of LAB. At 48 h, these values were 5.7 and 4.5, respectively. The observed values were much lower than those noted by Ogwaro et al. (2002) in the fermentation of traditional African yoghurt at comparable temperatures. Similar to our observation, Chang et al. (2000) reported that growth

Fig. 1. Changes in pH (A) and counts (B) of E. coli O157:H7 strains 93-30 (squares), 4595-52B (triangles) and NF-1847 (circles) during Ergo souring in the absence of LAB (filled symbols) and in association with LAB (open symbols). Diamond symbols depict LAB counts.

16

M. Tsegaye, M. Ashenafi / International Journal of Food Microbiology 103 (2005) 11–21

of E. coli O157:H7 alone in skim milk resulted in a slight change in pH during the 48-h fermentation period (6.3 to 5.6). Trends in reduction of pH in milk inoculated with E. coli O157:H7 test strains and LAB starter cultures were also similar to the observations of Chang et al. (2000). Although in the absence of LAB, our E. coli O157:H7 reached levels of log10 8.4 cfu/ml at 24–36 h, the highest population they reached in LAB-fermenting Ergo was log 6.5–6.6 cfu/ml at 24 h. This is different from the observations of other workers where E. coli O157:H7 strains coinoculated with LAB reached counts of around 8 logs at 36 h in fermenting skim milk (Chang et al., 2000) and at 18 h in fermenting traditional African yoghurt (Ogwaro et al., 2002). In contrast to our observation that E. coli O157:H7 survived Ergo fermentation and was present at levels of log10 3.2 cfu/ml at a pH value of 3.9 at 72 h, Dineen et al. (1998) reported that E. coli O157:H7 was not recovered after the curd formation step in yoghurt fermentation. McIngvale et al. (2000), however, recovered E. coli O157:H7 after the fermentation of buttermilk inoculated prior to fermentation. The decrease in the population of the E. coli O157:H7 test strains in fermenting Ergo could have been due to the antagonistic activities of the LAB through their metabolites and resulting low pH. The growth of E. coli O157:H7 in LAB-fermenting Ergo until 24 h at pH values of 5.5–4.5 could have contributed to the development of acid tolerance, which eventually resulted in survival at pH values ofb4.0 for more than 24 h. According to Buchanan and Doyle (1997), E. coli cells in the stationary phase of growth are substantially more acid tolerant than cells in the exponential phase. The souring of milk is generally considered as a preferable process for improving the keeping quality of milk and for ensuring its safety. Ergo fermentation is a natural process and traditionally proceeds at ambient temperatures (20–25 8C). Previous studies have shown that S. aureus and B. cereus were eliminated within 24–38 h during Ergo fermentation (Ashenafi, 1992), while it took 48–60 h to inactivate Salmonella typhimurium, Salmonella enteritidis and L. monocytogenes in fermenting Ergo (Ashenafi, 1993, 1994a,b). However, as Ergo develops the optimum flavor and odor for consumption at 24 h of fermentation, survival of pathogens at 24 h would

make Ergo consumption potentially hazardous to health. Unlike other food-borne pathogens studied previously, which could not survive in fermenting Ergo for more than 60 h (Ashenafi, 1992, 1993, 1994a,b), our E. coli O157:H7 test strains could survive at 72 h with counts much higher than the infective dose. Fermentation of Ergo does not control E. coli O157:H7, and, therefore, Ergo can be a potential health hazard if it is prepared from milk contaminated with E. coli O157:H7. Fermented milk contaminated with this pathogen did not show any off odor. Thus, it is difficult to depend on the sensory evaluation of the product as a sign of safety of the product. This was also observed during the fermentation of buttermilk and buffalo raw milk inoculated with E. coli O157:H7 (Dineen et al., 1998; McIngvale et al., 2000, El-Kosi et al., 2000). 3.2. Survival of E. coli O157:H7 in Ergo product In a preliminary study, Ergo samples were inoculated with E. coli O157:H7 at levels of ca. log10 3 cfu/ ml. The E. coli O157:H7 test strains were undetectable within 6 h of storage at ambient temperature while they were recovered at 72 h in samples maintained at refrigerator temperature (data not shown). To evaluate the rate of inactivation of our test strains in Ergo, a study was conducted with a higher initial inoculum level of E. coli O157:H7 (log 6 cfu/ ml). At ambient temperature, E. coli O157:H7 decreased by 4 logs within 12 h and the decrease in pH during this time period was b0.1 units. Complete inactivation of the test strains was observed at 36 h at pH values of around 4.0 (Fig. 2). This pattern was similar for the three E. coli O157:H7 test strains considered in this study and variations in count at the different sampling times were not significant for all the strains (CVb10%). Counts of the test strains in the control phosphate buffer saline did not show significant variation throughout the incubation period (CVb10%). Survival of E. coli O157:H7 at pHb4.0 between 48 and 72 h in fermenting Ergo, but its elimination at a similar pH at 36 h in Ergo product stored at ambient temperature is worth noting. This may be due to the possible development of acid tolerance where E. coli

M. Tsegaye, M. Ashenafi / International Journal of Food Microbiology 103 (2005) 11–21

Fig. 2. Changes in pH (A) and counts (B) of E. coli O157:H7 strains 93-30 (squares), 4595-52B (triangles) and NF-1847 (circles) in Ergo maintained at ambient (open symbols) and refrigeration (closed symbols) temperatures.

17

O157:H7 had the chance to grow at pH values of 5.0 in fermenting Ergo. Development of acid tolerance would have been less likely in the fermented product because they were inoculated in a product with a pH value of 4.3. In addition, the concentration of antimicrobial substances produced by LAB would be considerably higher in the Ergo product than in the fermenting milk. The faster rate of inactivation observed in our study was different from the rate of inactivation of E. coli O157:H7 inoculated in a traditional African yoghurt, where it required over 96 h for complete elimination of the pathogen (Ogwaro et al., 2002). Inhibition of E. coli O157:H7 was also observed at refrigeration temperature but at a much lower rate (Fig. 2). Reduction in mean counts of 0.7 to 1.3 logs was observed for the different strains at 24 h, and mean counts at 72 h were lower by 4.3 and 3.1 logs for strains MF-1847 and 93-30, respectively, than the initial inoculum level. Although the general trend of reduction in numbers was evident for all strains, variations in counts were significant during the sampling times (CVN10%). Noticeable changes in the pH of Ergo were not observed during storage at refrigeration temperature, mainly due to the absence of activity from LAB or E. coli O157:H7 strains at this temperature. During a traditional yoghurt fermentation, Massa et al. (1997) reported a slight decrease in number of E. coli O157:H7 after 7 days of storage of yoghurt at 4 8C. Their strains had still higher counts at day 7 probably because they were survivors of the fermentation process and had developed acid tolerance which became more pronounced at cold temperatures. In Yakult, a diluted fermented milk drink (pH 3.6) and yoghurt (pH 3.9) stored at 7 8C, the levels of E. coli O157:H7 were not reduced by more than 1 log at 72 h, whereas the counts of another strain were reduced by about 4 logs at 72 h (Hsin-Yi and Chou, 2001). Ogwaro et al. (2002) observed no marked change in numbers of E. coli O157:H7 in fermented milk kept at 4 8C for 144 h. The different results obtained by various workers might be due to differences in strain sensitivity to pH and temperature as observed by Hsin-Yi and Chou (2001) and Ogwaro et al. (2002), or to the type of starter cultures used which, according to Dineen et al. (1998), could differ in their ability to reduce E. coli O157:H7 in fermentation systems.

18

M. Tsegaye, M. Ashenafi / International Journal of Food Microbiology 103 (2005) 11–21

Under normal circumstances, Ergo would likely not be contaminated at the levels used in the present study. As low initial inoculum sizes were followed by complete inhibition with 6 h during storage, postfermentation contamination of Ergo may not be considered as a potential health hazard. 3.3. Survival and inactivation of E. coli O157:H7 during Ayib processing During the fermentation of milk with LAB to produce Ayib, E. coli O157:H7 counts increased by about 3 logs in 24 h with a slight reduction at 36 h (Table 1). The pH dropped to 4.3 during this time. Variation in counts of the test strains or pH values in fermenting milk was not significant (CVb10%). Removal of milk fat reduced the count of the test strains by 0.4 logs. This would mean that E. coli O157:H7 would be present in the recovered fat. However, as this traditional butter would be further heat-processed before being used for consumption, the presence of E. coli O157:H7 in the recovered fat may not be of public health significance. The levels of E. coli O157:H7 immediately after curd cooking were below the detectable limit, but they could be recovered after enrichment. Complete elimination was, however, observed 24 h after curd cooking (Table 1). Changes in pH after curd cooking were not noticeable, and this was basically due to the absence of microbial activity, as the cooking temperature would have inactivated both LAB and E. coli

O157:H7, which were active in fermenting milk. Salmonella spp. has been shown not to survive the curd-cooking temperatures used during Ayib making (Abdella et al., 1996). The survival of some injured E. coli O157:H7 cells after curd cooking is of some significance. Stringer et al. (2000) reviewed the published data on heat inactivation of E. coli O157:H7 and concluded that there was no evidence that a heat treatment of 70 8C for 2 min failed to deliver a 6 log reduction in cell numbers. According to Rhee et al. (2003), when ground beef was cooked to an internal temperature of 71.1 8C in a one turn-over single-sided grill, the time required to reach 71.1 8C was 10.9 min, and some survivors were recovered, by enrichment, after cooking. Turner (2002) also reported that the temperature required to inactivate E. coli depended on, among other things, substrate composition and moisture content. There could have been some sublethal heating during the long time needed to bring the internal temperature of the casein to 70 8C in our study, which could have resulted in heat shocking. Heat shocking allowed E. coli O157:H7 to survive longer at 60 8C in beef gravy and ground beef (Juneja et al., 1998). In addition, its survival for over 36 h when present at high levels in fermenting milk for Ayib making could have resulted in the development of acid tolerance. This phenomenon could confer some cross-protection effect against heat stress to E. coli O157:H7 (Cheville et al., 1996). Similarly, Duffy et al. (2000) studied the effect of

Table 1 Counts (log cfu/g) of Escherichia coli O157 test strains during milk souring and Ayib processing Strain

Milk souring 0h

9303 1847 4595 Mean S.D. %CV

24 h

Defatted milk

After curd cooking

36 h

0h

24 h

pH

Count

pH

Count

pH

Count

pH

Count

pH

Count

pH

6.43 6.46 6.46 6.45 0.02 0.3

3.33 3.65 3.82 3.60 0.25 7.0

4.34 4.43 4.41 4.39 0.05 1.1

6.21 6.51 6.16 6.29 0.19 3.0

4.32 4.35 4.30 4.32 0.03 0.6

6.12 6.00 6.17 6.09 0.09 1.4

4.30 4.31 4.30 4.30 0.01 0.2

5.77 5.57 5.82 5.72 0.13 2.3

4.36 4.31 4.29 4.32 0.04 0.8

+ + +

4.30 4.30 4.30 4.30 0 0

+, detectable only after enrichment. , not detectable after enrichment. S.D., standard deviation. CV, coefficient of variation.

Count

M. Tsegaye, M. Ashenafi / International Journal of Food Microbiology 103 (2005) 11–21

19

growth pH and fermentation on the thermotolerance of E. coli O157:H7 at 55 8C, and concluded that growth at low pH could confer cross-protection against heat. 3.4. Survival of E. coli O157:H7 in Ayib product Preliminary studies showed that at low initial inoculum levels (log10 3 cfu/g), E. coli O157:H7 could not survive for a long period in the Ayib product stored at ambient temperature (data not shown). The inoculum levels used (ca 103 cfu/ml) are generally considered to be similar to contamination via unclean utensils or hands under normal food handling conditions in the household (Svanberg et al., 1992). Therefore, at these levels, Ayib product might be considered safe when maintained at ambient temperatures. When our E. coli O157:H7 test strains were inoculated into steam-treated Ayib (pH 4.2–4.3) at levels of log10 6.0–6.7 cfu/ml and maintained at ambient temperature, there was a gradual decrease in counts to 3.7–4.3 log cfu/g at day 3 for the different strains (Fig. 3). The pH dropped by 0.1 to 0.2 units during this time. On day 4, the count of all strains showed a slight increase. This growth could be due to development of acid adaptation during survival for 3 days at pH values which did not decrease markedly. Increase in metabolic activity of the test strains due to growth resulted in a further decrease in pH to values V4.0, followed by subsequent reduction in count of the test strains. After day 7, pH values were 3.8–3.9 and the test strains were detectable only after enrichment. E. coli O157:H7 strain MF-1847 was completely eliminated by day 9, but the other two strains could be recovered by enrichment at this time. Variations in counts and pH values during any sampling time were not significant (CVb10%). At room temperature storage, E. coli O157:H7 survived (counts Nlog 4 cfu/ml) for over 20 days in tomato juice (pH 4.4), but was completely eliminated by day 6 in tomato ketchup (pH 4.2) (Eribo and Ashenafi, 2003). Similarly, long survival of E. coli O157:H7 at room temperature storage was observed in asparagus juice (pH 3.6) where counts of the test strains were Nlog10 3 cfu/ml on day 13, but complete elimination was observed in mango juice (pH 3.2) within 5–6 days at room temperature (Hsin-Yi and Chou, 2001). Therefore, elimination during storage at 20–25 8C

Fig. 3. Changes in pH (A) and counts (B) of E. coli O157:H7 strains 93-30 (squares), 4595-52B (triangles) and NF-1847 (circles) in Ayib maintained at ambient (open symbols) and refrigeration (closed symbols) temperatures.

20

M. Tsegaye, M. Ashenafi / International Journal of Food Microbiology 103 (2005) 11–21

would depend on the pH and composition of the medium. At refrigeration temperatures, the highest reduction in number of E. coli O157:H7 (N1 log) was observed on day 1 of storage. A decrease in counts was very gradual thereafter, and reduction in pH by day 7 was not marked. Counts at days 8 and 9 were still N4 log cfu/g. Variations in counts and pH values during any sampling time were not significant (CVb10%). Similar high numbers of E. coli O157:H7 were also observed in refrigerated tomato ketchup (pH 4.2) (Eribo and Ashenafi, 2003). According to Buchanan and Doyle (1997), acid tolerance in E. coli O157:H7 is growth phase dependent and inducible. As our test strains were inoculated in Ayib samples while they were in the stationary phase of growth, they would likely be more acid tolerant. An acid tolerant state can persist for extended periods (z28 days) if cells are stored at refrigeration temperatures (Cheville et al., 1996). Ashenafi (1990a,b) analyzed the microflora of market Ayib and found out that the product had high mesophilic bacterial counts (log10 8 cfu/g) and pH values that varied from 3.3 to 4.6. Thus, contamination of Ayib with high levels of E. coli O157:H7, particularly for products with high pH values, would make it potential hazardous. It would be worthwhile to study the survival of E. coli O157:H7 in Ayib made with different curd-cooking temperatures, in order to recommend a safe processing procedure.

Acknowledgements MT acknowledges the financial support of Sida/ SAREC obtained through the School of Graduate Studies, Addis Ababa University. MA thanks the DAAD for financial support and the Institute of Immunology and Molecular Biology, Faculty of Veterinary Medicine, FU Berlin for making the required facilities and support available during the preparation of the manuscript.

References Abdella, M., Becker, H., Terplan, G., 1996. Comparative studies on the detection of salmonellae in Ethiopian cottage cheese (ayib) using different culture methods. Arch. Lebensm.hyg. 47, 83 – 90.

Ashenafi, M., 1990a. Microbiological quality of dAyibT, a traditional Ethiopian cottage cheese. Int. J. Food Microbiol. 10, 263 – 268. Ashenafi, M., 1990b. Effect of curd cooking temperature on the microbiological quality of dAyibT, a traditional Ethiopian cottage cheese. World J. Microbiol. Biotechnol. 6, 159 – 162. Ashenafi, M., 1992. Growth potential and inhibition of Bacillus cereus and Staphylococcus aureus during the souring of dErgoT a traditional Ethiopian fermented milk. Ethiop. J. Health Dev. 6, 23 – 29. Ashenafi, M., 1993. Fate of Salmonella enteritidis and Salmonella typhimurium during the fermentation of dErgoT, a traditional Ethiopian sour milk. Ethiop. Med. J. 31, 91 – 98. Ashenafi, M., 1994. The aerobic microflora and lactic acid bacteria of market dAyibT. Ethiop. J. Agric. Sci. 14, 104 – 111. Ashenafi, M., 1994. Fate of Listeria monocytogenes during the souring of dErgoT, a traditional Ethiopian fermented milk. J. Dairy Sci. 77, 696 – 702. Ashenafi, M., 1995. Microbial development and some chemical changes during the making of dErgoT, a traditional Ethiopian fermented milk. Bull. Anim. Health Prod. Afr. 43, 171 – 176. Buchanan, R.L., Doyle, M.P., 1997. Food borne disease significance of Escherichia coli O157:H7 and other enterohemorrhagic Escherichia coli. Food Technol. 51, 61 – 76. Chang, J.-H., Chou, C.-C., Li, C.-F., 2000. Growth and survival of Escherichia coli O157:H7 during the fermentation and storage of diluted cultured milk drink. Food Microbiol. 17, 579 – 587. Cheville, A.M., Arnold, K.W., Buchreiser, C., Cheng, C.-M., Kaspar, C.W., 1996. rpoS regulation of acid, heat and salt tolerance in Escherichia coli O157:H7. Appl. Environ. Microbiol. 62, 1822 – 1824. Collins, C.H., Lyne, P.M., Grange, J.M., 1991. Microbiological methods, 6th ed. Butterworth-Heinemann Halley Court, Jordan Hill, Oxford. Dineen, S.S., Takeuchi, K., Soudah, J.E., Boor, K.J., 1998. Persistence of Escherichia coli O157:H7 in dairy fermentation systems. J. Food Prot. 61, 1602 – 1608. Duffy, F., Riordan, D.C.R., Sheridan, J.J., Call, J.E., Whiting, R.C., Blair, I.S., McDowell, D.A., 2000. Effect of pH on survival, thermotolerance, and verotoxin production of Escherichia coli O157:H7 during simulated fermentation and storage. J. Food Prot. 63, 12 – 18. El-Kosi, O.H.R., Abdel-Hakiem, E.H., Saad, H.H., 2000. Fate of enterohemorrhagic Escherichia coli O157:H7 in buffalo’s milk and some of its manufacturing products. http:// www.wiz.uni-kassel.de/gear/symp2000/proceed99/vm.2.pdf. Eribo, B., Ashenafi, M., 2003. Behaviour of Escherichia coli O157:H7 in tomato and processed tomato products. Food Res. Int. 36, 823 – 830. Gonfa, A., Fite, A., Urga, K., Gashe, B.A., 1999. The microbiological aspects of dErgoT (Ititu) fermentation. SINET: Ethiop. J. Sci. 22, 283 – 289. Gonfa, A., Foster, H.A., Holzapfel, W.H., 2001. Field survey and literature review on traditional fermented milk products of Ethiopia. Int. J. Food Microbiol. 68, 173 – 186. Hsin-Yi, C., Chou, C.-C., 2001. Acid adaptation and temperature effect on the survival of E. coli O157:H7 in acidic fruit juice

M. Tsegaye, M. Ashenafi / International Journal of Food Microbiology 103 (2005) 11–21 and lactic fermented milk product. Int. J. Food Microbiol. 70, 1.5 – 112. Janisiewice, W.J., Conway, W.S., Brown, M.W., Sapers, G.M., Fratamico, P., Buchanan, R.L., 1999. Fate of Esherichia coli O157:H7 on fresh-cut apple tissue and its potential for transmission by fruit flies. Appl. Environ. Microbiol. 65, 1 – 5. Juneja, V.K., Klein, P.G., Marmer, B.S., 1998. Heat shock and thermotolerance of Escherichia coli O157:H7 in a model beef gravy system and ground beef. J. Appl. Microbiol. 84, 677 – 684. Knudsen, D.M., Yamamato, S.A., Linda, J.H., 2001. Survival of Salmonella spp. and Escherichia coli O157:H7 on fresh and frozen strawberries. J. Food Prot. 64, 1483 – 1488. Massa, A., Altieri, C., Quaranta, V., De Pace, R., 1997. Survival of Escherichia coli O157:H7 in yoghurt during preparation and storage at 4 degrees C. Lett. Appl. Microbiol. 24, 347 – 350. Mcdonough, S., Heer, F., Shirley, L., 1991. Foodborne outbreak of gastroenteritis caused by Escherichia coli O157:H7 North Dakota, 1990. Morb. Mortal. Wkly. Rep., CDC Surveill. Summ. 40, 265 – 267. McIngvale, S.C., Chen, X.Q., McKillip, J.L., Drake, M.A., 2000. Survival of Escherichia coli O157:H7 in buttermilk as affected by contamination point and storage temperature. J. Food Prot. 63, 441 – 444. Morgan, D., Newman, C.P., Hutchinson, D.M., Walker, A.M., Rowe, B., Majid, F., 1993. Verotoxin producing Escherichia coli O157:H7 infections associated with the consumption of yoghurt. Epidemiol. Infect. 111, 181 – 187. Ogwaro, B.A., Gibson, H., Whitehead, M., Hill, J.D., 2002. Survival of Escherichia coli O157:H7 in traditional African yoghurt fermentation. Int. J. Food Microbiol. 79, 105 – 112. O’Mahoney, F., 1988. Rural Dairy Technology—Experience in Ethiopia. ILCA Manual No.4. International Livestock Centre for Africa, Addis Ababa.

21

Rhee, M.S., Lee, S.Y., Hillers, V.N., McCurdy, S.M., Kang, D.H., 2003. Evaluation of consumer style cooking methods for reduction of Escherichia coli O157:H7 in ground beef. J. Food Prot. 66, 1030 – 1034. Riley, L.W., Remis, R.S., Helgerson, S.D., McGee, H.B., Wells, J.G., Davis, B.R., Hebert, R.J., Olcott, E.S., Johnson, L.M., Hargrett, N.T., Blake, P.A., Cohen, M.L., 1983. Hemorrhagic colitis associated with a rare Escherichia coli serotype. N. Engl. J. Med. 308, 681 – 685. Stringer, S.C., George, S.M., Peck, M.W., 2000. Thermal inactivation of Escherichia coli O157:H7. Soc. Appl. Microbiol. Symp. Ser. 29, 79S – 89S. Svanberg, U., Sjogren, E., Lorri, W., Svennenholm, A.-M., Kaifser, B., 1992. Inhibited growth of common enteropathogenic bacteria in lactic fermented cereal gruels. World J. Microbiol. Biotechnol. 8, 601 – 606. Tilden, J.J., Young, W., McNamara, A.M., Custer, C., Boesel, B., LamberFlair, M., Majikowski, J., Vugia, D., Werner, S.B., Hollingsworth, J., Morris, J.G., 1996. A new route of transmission for Escherichia coli: infection from dry fermented salami. Am. J. Public Health 86, 1142 – 1144. Trevena, W., Hooper, R., Wray, C., Willshaw, G., Cheasty, T., Domingue, G., 1996. Verotoxin producing Escherichia coli O157 associated with animals. Vet. Rec. 136, 400. Turner, C., 2002. The thermal inactivation of E. coli in straw and pig manure. Bioresour. Technol. 84, 57 – 61. Zhao, T., Doyle, M.P., Bessar, R.E., 1993. Fate of Enterohemorrhagic Escherichia coli O157:H7 in apple cider with and without preservatives. Appl. Environ. Microbiol. 59, 2526 – 2530.