THE PRESENCE OF VEROTOXINOGENIC E. COLI IN SOME FOODS*

In this study 30 samples each of ready-to-cook meatballs and white cheese as well as 96 samples of various ready-to-eat foods, obtained from different sales outlets in Istanbul, were analysed for the presence of verotoxins (consequently verotoxigenic E. coli) E. coli with the aid of the enzyme immunoassay technique. Additionally, total coliform and chromogenic E. coli count were determined by cultural methods for all food samples. E. coli growth was detected in all ready-tocook meatballs (100%), in 27 of the white cheese samples (90%) and in 69 of the other ready-to-eat food samples (71.9%). Verotoxins, however, could not be detected in any of the samples examined with the aid of the ELISA technique. The findings of this study indicate a low microbiological quality of the analysed meatball, white cheese and ready-toeat food samples; a considerable part of them did not conform to legal standards. However, within the sensitivity limits of the method applied no verotoxinogenic E. coli could be detected.

ble for some cases of very severe poisoning that ended in death Š6, 23, 9, 20, 3¹.EHEC includes E. coli strains (E. coli O157:H7) that produce verotoxins (VT1 and VT2).Verotoxins have a cytotoxic effect and also produced by some other E. coli serotypes Š15¹.Verotoxin-producing E. coli (VTEC) are considered a major cause of gastrointestinal disease in developed countries.The nature of illness can range from a mild form of diarrhea to more severe forms known as hemorrhagic colitis and hemolytic uremic syndrome Š30, 28¹.Although the number of organisms required to cause disease is not known, it is suspected to be very small.
The main sources of infection are meat and meat products from cattle, sheep and goats.Among other known sources of infection are consumption of sprouts, vegetable salad, unpasteurized milk and juice.Insufficient cooking or raw consumption of contaminated products increases the risk of poisoning Š36, 28¹.The presence of verotoxigenic E. coli in various foods has been investigated by several researchers.The general finding was a low rate of incidence in animal based food Š5, 26, 17, 13, 87¹.There are various methods for isolation of verotoxigenic E. coli from food and other material.However, these methods are rather onerous and time consuming.In contrast, detection of toxins (verotoxin) from an enrichment culture using immunoenzymatic assay as a way of identifying the VTEC in foods is more easy, fast and sensitive Š32, 14, 27, 25¹.This study has been realised with the aim of determining with the aid of the ELISA technique the presence of verotoxinogenic E. coli in ready-to-cook meatballs, cheese and ready-to-eat foods marketed in Istanbul.

Materials / Materijal
In this study a total of 156 food samples, which are composed of 30 ready-to-cook meatball, 30 white pickled cheese and 96 ready-to-eat food (traditional Turkish cold dishes) and marketed without packaging, have been analysed.The samples were obtained from various markets and other sales outlets in Istanbul.The samples were placed in sterile plastic bags and were transported immediately to the laboratory under hygienic conditions.

Microbiological analyses / Mikrobiolo{ke analize
A 10 g sample of each food sample was homogenised 1:10 (w/v) with sterile 0.1 % peptone water for 2 min in a Stomacher Lab-Blender (Seward).Serial 10-fold dilutions were prepared from the same solution and inoculated in appropriate growth media Š12¹.
Total coliform and E. coli count were determined on Chromogenic E. coli / Coliform Medium Agar (Oxoid, CM 0956) by using pour plate technique.The plates were incubated for 24 hours at 37°C and evaluated.Red to pink colonies were considered total coliform and dark-blue to violet colonies presumptive E. coli Š10, 19¹.

Materials and methods / Materijal i metode rada
Detection of verotoxin / Ustanovljavanje verotoksina The presence of verotoxins was determined by ELISA-based Ridascreen Verotoxin Test Kit (Biopharm, R5701).Toxin detection was carried out after cultural enrichment.For this purpose, food samples (25 g) were added into 225 ml modified Tryptic Soy broth (mTSB, Oxoid CM129) supplemented with 1.0 ml of a 0.45% novobiocin (Biochemika 71882) and incubated overnight at 37°C.This was folloÓed by short centrifugation and the supernatant was directly applied into the assay.The assay was performed following the manufacturer's instructions to detect VT1 and VT2.The readings were taken at 450 nm wavelength using an automated micro plate reader (ELx-800, Universal Microplate Reader, BIO-TEK).
In this study, coliform and E. coli growth were detected in all ready-tocook meatball samples (Tables 1 and 3).The average coliform and E. coli counts were high and determined as 1.2x10 5 cfu/g and 3.4x10 4 cfu/g, respectively.The high contamination rate and average counts may be explained as a sign of the low hygienic quality of the meatballs.Another reason for the high count of microorganisms may be due to the fact that they were collected during a relatively hot season (from May to August) of the year.Chapman et al.Š7¹ who investigated the presence of E. coli in meat and raw meat products found microorganism counts in lamb and beef collected during July and August exceeding 10 4 cfu/g.The Turkish Food Codex Directive on prepared meat mixtures Š31¹ sets a limit for E. coli of 5.0x10 3 cfu/g.In 24 of the samples (80%), the E. coli count was above the legal limit.Of the cheese samples, only 3 (10%) were negative for coliform and E. coli growth (Tables 1 and 3).In twenty five samples the coliform and E. coli count exceeded 100 cfu/g.According to the cheese standard issued by the Institute for Turkish Standards Š34¹, the coliform count in 1g of white cheese must not exceed 100, while E. coli must not be present at all.This means that 80% of the examined white cheese samples did not conform to that standard with respect to their coliform count while 90% failed to meet the E. coli condition.The average coliform count was found to be 7.9x10 4 cfu/g while the average E. coli count was Results and discussion / Rezultati i diskusija 5.3x10 4 cfu/g.The ratio of E. coli among the total of coliform bacteria is rather high (67.1%).For meatballs and the ready-to-eat foods investigated, the ratios were considerably lower, 28.3% and 23.1% respectively.Another study carried out in Turkey reported high coliform and E. coli counts in white cheese Š35¹, which failed to meet the E. coli condition.Tawfek et al Š33¹ found coliform bacteria in all of the 100 Kariesh cheese samples with average coliform counts of 18.9x10 3 /g.Reibnitz et al.Š29¹ found, in a study carried out in Brasil that 79% of the examined cheese samples did not conform to legal requirements while 20% exceeded the limits for E. coli.Aleksieva Š2¹ reported that 3.1% of the analysed white pickled cheese samples contained between 10-100 cfu/g, a finding which is far below the results obtained in this study.The high contamination with coliform and E. coli of the examined cheese samples may be an indicator that most products have been produced from raw milk and had not matured sufficiently.
In the 96 samples comprising thirteen different food varieties of readyto-eat products, the average coliform count was 2.2x10 4 cfu/g and the average E. coli count 5.1x10 3 cfu/g (Tables 2 and 4).Coliform and E. coli growth were not determined at a countable level in 22.9% and 28.1% of the samples, respectively.The E. coli count in 12 samples (12.5%) was in the range of 10-100 cfu/g, in 41 of the samples (42.7%) between 100 and 10.000 cfu/g, and in 16 samples (16.7%) it exceeded 10.000 cfu/g.The highest average coliform count, 8.8x10 4 cfu/g, was detected in kýsýr samples -a cold dish made with cracked wheat, salad vegetables, and olive oil -while the lowest count, 3.5x10 2 cfu/g, was found in tarator samples -an appetizer made with yoghurt, garlic, mashed corn, chopped carrot, sesame oil, mashed chickpea and mashed eggplant.With respect to E. coli, the highest count with 1.4x10 4 cfu/g was detected in paste with hot pepper samplesand the lowest count with 2.9x10 2 cfu/g in tarator samples.
Phls Š24¹ classified ready-to-eat foods with respect to their E. coli count in four categories: <10 cfu/g -suitable; <100 cfu/g -acceptable; 100-10 000 cfu/g -insufficient; (>10 000 cfu/g) -not acceptable.On the basis of that classification 28.1% of the ready-to-eat foods investigated in this study conform to the quality of "suitable", 12.5% are "acceptable", 42.7% "insufficient" and 16.7% "not acceptable".The values and ratios reported in other studies on ready-to-eat foods are lower than in this work.Gillespie et al Š11¹ examined the microbiological quality of 2354 samples of ready-to-eat foods.They reported that 157 samples (6%) were of insufficient quality and only 2 samples (<1%) with E. coli counts exceeding 10 4 cfu/g were not acceptable.Mosupye and Holy Š21¹, who analysed 51 ready-to-eat food samples in South Africa, reported an average Enterobacteriaceae count of 2.0 log cfu/g.Only in 2% of the samples inspected could E. coli be isolated.Kaneko et al.Š16¹ found in their study of ready-to-eat foods in Tokyo a coliform count range of 0.1-2.3log cfu/g.Despite the fact that high counts of coliform and E. coli have been detected in the meatball, white cheese and ready-to-eat food samples investigated in this study, no verotoxins and consequently no verotoxinogenic E. coli could be found with the ELISA technique.Meat and meat products, which may be easily contaminated with verotoxin producing E. coli strains during processing, are ideal medium supporting the bacterial growth.Abdul-Raouf et al Š1¹ investigated the presence of verotoxin 1 produced by E. coli O157:H7 in roasted beef.They incu-bated beef (pH 5.9) at different temperatures and found that samples incubated at 37°C had a 15-24% higher verotoxin content than samples incubated at 21°C.A decrease of the pH to 5.4 (with the addition of acetic acid and citric acid) reduced the amount of toxins considerably.Weeratna and Doyle Š37¹ reported the highest concentrations of verotoxin 1 in milk and ground meat samples incubated at 37°C for 48 hrs.Samples incubated at 25°C and 30°C were found to produce much less toxin.In this study no toxin was detected despite the fact that all samples were incubated in an enriched medium at 37°C.It can, therefore, be concluded that non of the examined samples contained verotoxinogenic E. coli.All studies on the incidence rate of E. coli O157:H7 in risk foods reported only very few cases.Coia et al.Š8¹ who analysed 1190 raw meat products in Scotland, found VTEC O157 in only 2 beef burgers.Chapman et al.Š7¹, found E. coli O157:H7 in 21 out of 1500 cattle carcasses (1.4%), in 10 out of 1500 lamb carcasses (0.7%) an in 22 out of 4983 raw meat samples (0.44%).Heuveling et al.Š13¹ tested, in 1996 and 1997 in The Netherlands, 2,941 meat products for the presence of verotoxin producing E. coli.They isolated O157 VTEC in 6 out of 571 samples of ground beef (1.3%), in 2 out of 402 samples of raw beef and ground pork (0.5%), in 1 out of 76 samples of raw ground pork (1.3%) and in 1 out of 393 other raw pork products (0.3%).The results for all other raw beef products as well as all poultry, lamb and game products were negative.Little and Louvois (1998) who analysed 2330 samples of raw meat in Great Britain reported O157 VTEC in 3 samples while in 2192 cooked samples no traces of the organism could be detected.However low the incidence rate of verotoxinogenic E. coli in meat products may be, the agent survives in meat and meat products for extended periods of time.Ansay et al.Š5¹ reported that E. coli O157:H7 remained active for a long time in contaminated ground beef kept at different storage temperatures.
Due to their low pH, cheeses are not an ideal medium for the development and support of E. coli.However, E. coli remains present in cheese produced from contaminated raw milk which is not sufficiently matured.Maher et al.Š18¹ investigated the behaviour of E. coli O157:H7 in cheese prepared from contaminated raw milk.They found that a count of 1.52 log cfu/g in milk increased in one day old cheese to 3.4 log cfu/g while rapidly decreasing during maturing; after 21 days the count fell below 10 cfu/g.However, even on the 90 th day of maturing the bacteria could still be detected in an enriched medium.Quinto and Capeda Š26¹ investigated the presence of E. coli in 221 samples of soft cheese made from raw milk and in 75 samples of soft cheese made from pasteurised milk.They reported that 3 samples of the raw milk cheese (1.4%) were contaminated with toxigenic E. coli, one of which contained verotoxin.Coia et al.Š8¹ examined 500 samples of raw milk and 739 samples of cheese made from raw milk.They reported no presence of the bacteria in any of the samples.
In the present study E. coli O157:H7 was not directly determined in the foods with high E. coli count.And even after enrichment according to procedures no verotoxin could be detected.This means that the samples did not contain vero-toxinogenic E. coli.Ansay and Kaspar Š5a¹ also found contamination with E. coli in 58% of the cheese samples tested, however, they did not find E. coli O157:H7 in any of the samples.The findings of this study indicate that the analysed meatball, cheese and ready-to-eat samples have a low microbiological quality, but that samples, however, were reliable from the point of view of verotoxinogenic E. coli.