Molecular Characterization of Shiga Toxin-Producing Escherichia coli Isolated from Ruminant and Donkey Raw Milk Samples and Traditional Dairy Products in Iran

The aims of the current study were to detect the virulence factors and antibiotic resistance of Shiga toxin-producing E. coli, in animal milk and dairy products in Iran. After E. coli dentification with culture method, PCR assay were developed for detection of pathogenic genes, serotypes and antibiotic resistance genes of E. coli. Results showed that out of 719 samples, 102 (14.18%) were confirmed to be positive for E. coli and out of 102 positive samples, 17.64% were O26 and 13.72% were O157 and 1.96% were O91 and 1.96% were O145 serotypes. Totally, the prevalence of stx1 and papA genes were the highest while the prevalence of sfaS and fyuA were the lowest in the positive samples. PCR results showed that tetA, tetB were the highest (64.70%) and aac(3)-IV were the lowest (27.45%) antibiotic resistant genes in E. coli positive samples. Our study indicated that the isolated E. coli trains in these regions had a highest antibiotic resistance to tetracycline (58.82%) and the lowest to nitrofurantoin (3.92%). tetA gene and E. coli O157 serotype had highest and aac(3)-IV gene, and E. coli O145 serotype had a lowest frequency rates of antibiotics resistance genes, in the region.

The Shiga toxin-producing E. coli (STEC) are the causative agents of severe clinical syndromes in humans like HUS. Four of the most important virulence factors of STEC are the two phage-encoded cytotoxins, called Shiga toxin 1 (stx1) and Shiga toxin 2 (stx2), the protein intimin (eae), and the plasmid-encoded enterohaemolysin or enterohaemorrhagic E. coli haemolysin (ehly) [7]. Gastrointestinal diseases, food poisoning, and even death in some cases make this microorganism an insidious threat to human health and food safety [4,8]. Trough outbreaks in the European Union [9] and several reports of septicemia in human and animals [10,11] up to recent studies in England [12], all of them have doubled the importance of addressing the issue.

Sampling and Detecting E. coli .
Overall 424 milk samples were collected: bovine (n = 110), caprine (n = 85), ovine (n = 98), buffalo (n = 52), camel (n = 44), and donkey (n = 35) raw milk samples were collected from farm bulk tanks and milk collection centers from several geographic regions of Iran, from March 2010 to March 2011. Bovine and buffalo milk samples were collected throughout this time period. Because the lactating periods of ewes, donkeys, and goats in Iran is seasonal (from March through May and September to November of the subsequent year), caprine, donkey, and ovine milk samples were only available through these months within the forementioned time frame. At each site, sampling of milk was performed according to the International Dairy Federation Guidelines [34]. Samples (100 mL, in sterile glass containers) were transported to the laboratory at ca. 4 • C within a maximum of 6-12 h after sampling.
For dairy products, 106 samples of soft cheese, 99 soft butter, and 90 soft ice-cream made up of raw milk were purchased from Shiraz, Isfahan, and Shahrekord-three big cities of Iran supermarkets. All of these dairy products were made traditionally and after collection were kept under refrigeration in plastic bags; information about dates of production and of assigned shelf-lives was not presented. Dairy product samples were collected over a period of six months between May and November 2010 and were analyzed on the day of acquisition. Samples were transported under refrigeration (4-6 • C) in thermal boxes containing ice packs and were tested immediately after collection. A 25 g portion of each sample was blended with 225 mL of nutrient broth (Merck, Germany) for 2 min at normal speed, using a Stomacher lab blender and incubated at 37 • C for 24 h. A 1 mL sample of the nutrient broth culture was mixed with 9 mL of MacConkey broth (Merck, Germany) and further incubated at 37 • C for 24 h. One loop of each tube was streaked on MacConkey agar (Merck, Germany). Four colonies from each plate with typical E. coli morphology were selected and examined by biochemical tests, including hydrogen sulphide, citrate, urease, and indol.

DNA Isolation.
Bacterial strains were overnight grown in trypticase soy agar (TSA-Merck, German) at 37 • C. One colony was suspended in 100 μL of sterile distilled water. After boiling the suspension for 13 min, this was followed by freezing and subsequently centrifuged at 14,000 rpm for 15 min to pellet the cell debris [35]. The supernatant was used as a template for amplification reaction.

Polymerase Chain
Reaction. PCR assays for detection of STEC serotypes, virulence factors, and antibiotic resistance genes in E. coli isolated from animal milk samples and traditional dairy products were performed in 3 states as follows.
Another PCR was performed for detection of E. coli serotypes by application of specific primers shown in Table 1  The presence of genes associated with resistance to streptomycin, tetracycline, trimethoprim, fluoroquinolone, gentamicin, sulfonamide, cephalothin, ampicillin, and chloramphenicol was determined by PCR, and a set of primers was used for each gene showed in Table 2. PCR reactions were performed in a total volume of 25 μL, including 2 mM MgCl 2 , 50 mM KCl, 10 mM Tris-HCl (pH 9.0), 0.1% Triton X-100, 150 μM dNTPs each, 0.4 μM primers, 1 U of Taq DNA polymerase, and 4 μL (40-260 ng/μL) of DNA. Amplification reactions were carried out using a DNA thermocycler as follows: three min at 95 • C, 35 cycles each consisting of 1 min at 94 • C, 90 s at ∼55 • C (show in Table 2), 1 min at 72 • C, and a final extension step of 8 min at 72 • C.
Finally, a DNA thermocycler (Eppendorf Mastercycler, Eppendorf-Nethel-Hinz GmbH, Hamburg, Germany) was used. The amplified products were visualized by ethidium bromide staining after gel electrophoresis of 10 μL of the final reaction mixture in 1.5% agarose.

Antimicrobial Susceptibility
Testing. Antimicrobial susceptibility tests were performed by the Kirby-Bauer disc diffusion method using Mueller-Hinton agar (HiMedia Laboratories, Mumbai, India, MV1084), according to the Clinical and Laboratory Standards Institute guidelines [41]. After incubating the inoculated plate aerobically at 37 • C for 18-24 h in an aerobic atmosphere, the susceptibility of the E. coli isolates to each antimicrobial agent was measured, and the results were interpreted in accordance with interpretive criteria provided by CLSI (2006). E. coli ATCC 25922 was used as quality control organisms in antimicrobial susceptibility determination.

Statistical
Analysis. Statistical analysis were performed using SPSS/16.0 software for significant relationship between incidences of virulence factors and antibiotics resistance genes of E. coli isolated from various animal milk samples and dairy products. Statistical significance was regarded at a P value < 0.05.

Results
In the current study, all E. coli colonies were tested by applying PCR method in order to detect 16S rRNA gene of bacterium. According to data, out of 424 animal milk samples, 57 (13.44%) and out of 295 traditional dairy products 45 (15.25%) were positive for presence of E. coli (Table 3). Therefore, it was shown that incidences of E. coli in different traditional dairy products were higher than animal milk species. Our results showed that bovine (20.9%) and soft cheese (23.58%) had a highest and camel (6.81%) and soft ice cream (10%) had a lowest rate of E. coli in raw milk and traditional dairy products, respectively (Table 3). Statistical analysis of data showed significant differences (P < 0.05) between bovine, ovine, caprine, and camel samples as well as donkey and camel milk samples for presence of E. coli. On the other hand, significant differences (P < 0.05) were observed in soft cheese, soft butter, and ice cream in incidence of this bacterium. It was concluded that the total prevalence of E. coli in all samples in Iran was 14.18% (102 out of 719).
On the other hand, our results revealed that out of 102 E. coli-positive samples, 34 (33.33%) were non-STEC and 68 (66.66%) were diagnosed as STEC serotypes (Table 5). Our results showed that raw bovine milk samples and traditional soft cheese had several virulence genes especially stx1 and traT, so were the main sources of pathogenic E. coli for human in Iran.
The last part of our study was allotted to antibiotic resistance patterns in E. coli that isolated from raw milk samples and traditional dairy products which evaluated by application of PCR for presences of genes associated with resistance to various antibiotics and disc diffusion method. PCR assays were successfully developed for detection of twelve different resistance genes of genomic DNA of E. coli isolates from milk and dairy products.
Therefore totally, the prevalence of antimicrobial resistance especially for tetracycline, streptomycin, sulfonamide, chloramphenicol, and finally cephalothin was so high in this survey.
Besides, statistical analysis revealed a statistical significant association (P < 0.05) between tetA and tetB genes in dairy production, tetB with aac(3)-IV, qnr, cat1, and cmlA in bovine, ovine, caprine, and buffalo milk samples, aadA1 with qnr, dfrA1, tetB, blaSHV, CITM,cat1, and cmlA in soft cheese, butter, and ice cream. Moreover, our results showed that all STEC serotypes which were isolated from raw animal milk samples and traditional dairy products had resistance to more than one antibiotic. Also, based on our data, out of 102 E. coli-positive isolates, all of them were resistant to one or more antimicrobial agent.
Finally, antibiogram tests, beside PCR assays, revealed the highest resistance to tetracycline, penicillin, and enrofloxacin with the frequency rate of 58.82%, 46.07%, and 45.09%, Table 3: Prevalence of STEC serotypes isolated from raw animal milk samples and traditional dairy products in Iran. The Scientific World Journal Table 4: Distribution of virulence genes in E. coli isolated from raw animal milk samples and traditional dairy products in Iran.

Samples
Virulence genes    respectively, as well as the lowest resistance to nitrofurantoin and ciprofloxacin with the frequency rate of 3.92% and 7.84% (Table 7). On the other hand, in the current situations in Iran, the use of nitrofurantoin and ciprofloxacin due to low antibiotic resistance, can be more effective for treatment of diseases caused by E. coli. This survey indicated the highest antimicrobial resistance in O157, O26 and O111 serotypes. Unfortunately, our results showed that nondetected serotypes had a highest incidence of antibiotics resistance. Besides, bovine had highest and camel had lowest rates of resistance to antibiotics. Statistical analysis showed significant differences (P < 0.05) between O157, O26, and nondetected serotypes with O103, 145, 91, 113, 121 and 128 serotypes and between resistance to trimethoprim, streptomycin, lincomycin, penicillin, and tetracycline with nitrofurantoin and ciprofloxacin antibiotics. Totally E. coli antibiotic resistances against common antibiotics which use in veterinary in Iran were so high. Based on Table 6, between these ten serotypes, E. coli O157 and E. coli O145 had a highest and lowest prevalence of antibiotic resistance genes. Results showed that in nondetected serotypes, prevalence of antibiotic resistance genes were so high too.

Discussion
The present study like the previous studies which introduced ruminants as reservoirs of E. coli O157 [42], for the first time, has been introduced donkeys, camels, and buffaloes as the other sources of E. coli serotypes especially O157. Despite the highest prevalence of E. coli O157 in recent studies [2,43], our results showed that E. coli O26 had a highest frequency rate in traditional dairy products. On the other hand, our study indicated that in spite of raw animal milk samples, pathogenic E. coli had a highest prevalence in traditional dairy products. Probably using contaminated water in processing traditional dairy products, infected people who produce dairy products, using contaminated equipments, basic infections of dairy sources which are necessary for producing dairy products, and finally lack of public and individual hygiene, are the factors that can be responsible for the higher prevalence of E. coli in traditional dairy products than raw animal milk samples. Only one recent study showed that presence of E. coli O157 in foods may be derived from polluted water [44]. There are several studies which have detected E. coli species in yogurt [45], soft cheeses [46], cheddar cheese [47], 8 The Scientific World Journal Table 7: Antibiotic resistance properties in STEC serotypes isolated from raw animal milk and traditional dairy products in Iran (disk diffusion method). and Feta cheese [48]; however in this survey it is for the first time that we observed that this bacterium can be detected in soft cheese, soft butter, and soft ice cream which were produced in traditional conditions too.

STES serotypes
According to Meshref in 2010 from Egypt on cooking butter, 31.7% of the samples were contaminated by E. coli [49], in which this rate was higher than ours (11.11%). Another study, performed on dairy products, showed that when the assay was applied with 60 market dairy food samples, one sample each of raw milk, paneer, and ice cream was found to be positive for E. coli O157:H7 [50] that was lower than our study.
Another study in Iran showed that only E. coli O157 subtype was isolated from traditional cheese (4.2%), ice cream (3.33%), and yoghurt (0%) [51] that was more limited than our study. The other study in Pantnagar revealed that out of 135 dairy product samples, 11 samples (8.14%) were positive for E. coli [52] that was lower than our study (14.18%). Recent study from Iran indicated that E. coli O157:H7 can be detected in high rates from white cheese and probiotic cheese [53]. Besides, a 3-year study of E. coli O157:H7 in cattle, camel, sheep, goat, chicken, and beefminced meat in Iran revealed that 1.7% and 2.1% of samples were positive for E. coli O157:H7 and E. coli O157:NM, respectively [54]. According to a study in Kerman on raw milk cheese producing, stx genes detected in 6.4% of the samples, but STEC strains could be isolated in only five of them (4%) [55]. Another study from Iran showed that about 21.8% of E. coli isolates from cattle were positive for ehxA, stx1, and/or stx2 genes [56], but our study indicated that stx1, cnf1, and eae with incidences of 24.5%, 21.56%, and 17.64% had a highest frequency rate of virulence genes of E. coli isolates from raw bovine milk samples.
Therefore, according to the results, the food contamination with animal fecal origin is so obvious in Iran. Based on the other surveys, the presence of E. coli O157 in raw milk and dairy products is proven, however in this investigation we found O26, O111, O103, and even O45 in addition to O157 as a predominant serotypes of E. coli in animal's raw milk and dairy products with the frequency of 17.64%, 9.8%, 5.88%, 5.88%, and 13.72%, respectively. So in our study E. coli O26 was the predominant serotype, AEEC was the predominant subtype, and papA and stx1 were the predominate genes.
Another study in Swiss in 2006 on 432 cheese samples and in 2007 on 364 cheese samples showed 16 (3.7%) and 23 (6.3%) E. coli-positive samples, respectively, so their results were lower than ours in cheese (23.58%) specimens. Further more in this study the investigators found non-O157 STEC strains in 16 samples in which 11 samples based on their O groups typed into 7 categories (O2, O15, O22, O91, O109, O113, O174), whereas 5 strains were nontypeable. They detect stx1 in only 1 sample and stx2 in 15 samples of 16 non-O157 STEC specimens. Furthermore this study showed that 3 STEC strains harbored E-hlyA as a further putative virulence factor and none of the strains tested positive for eae gene [14]. In another study in Irish bovine dairy herds, 85 out of 117 milk filters and 43 out of 120 bulk-tank milk samples were positive for vt genes of E. coli. Out of vt-positive samples, 58, 162, 1, 324, and 26 samples were positive for O26, O103, O111, O145, and O157 serotypes, and 10 of these isolates contained at least one of the four virulence genes of vt1, vt2, eaeA, and hlyA [57].
Recent study in United State reported that EHEC subtype of E. coli was detected in 23% of strains and eaeA gene was found in 4.2%, and one or both Shiga-like toxin genes (stx1 and stx2) were positive in bulk-tank milk too. Based on realtime PCR, they proved that 5 samples were contaminated with O157:H7 [58] while our study showed that out of 102 positive samples for E. coli, 13.72% were EHEC, 17.64% had eaeA gene, and 35.29% had both Shiga-like toxin genes.
The prevalence of Shiga toxin in our study (66.66%) was obviously higher than Lowa (4.08%) [59]. Another study on lamb food chain showed that all three virulence genes, eae, stx1, and stx2, were the most prevalent genes in slaughterhouses (69%), processing plant (32%), and butcheries (9-10%) [60], but our results indicated that in raw milk and dairy products, stx1 and papA with frequency rate of 24.5% had a highest prevalence of E. coli's virulence genes.
Detection of bacteria in milk samples has some merits like being available and economical. In some geographic places of Iran because of bad climate conditions, bovines, ovine, caprine, and buffaloes are not able to survive, so camels are the main milk supplier for people. So camel's milk hygiene can be essential for human health especially in those especially in that area where the climate is arid and desert. Donkeys' milk samples due to high similarity with human milk and high-sweet components due to high presence of lactose sugar sometimes can be used instead of human milk. Buffaloes' milk hygiene due to high fat and protein contents is very important especially for producing some dairy products.
Studies showed that the antibiotic resistance bacteria which exist in the milk of infected animals can be transmitted to human by the ingestion of unpasteurized milk or dairy products such as cheese, butter, yoghurt, and ice cream. However, the mechanism of spreading antibiotic resistance from food animals to humans remains controversial [61,62].
Our study showed that E. coli O157 was the most prevalent serotype in donkey's milk. Furthermore, this serotype had the highest antibiotic resistance to penicillin and enrofloxacin. To our knowledge, veterinary service-like treatments are very low for donkeys in Iran and unfortunately if a veterinarian wants to apply antibiotical treatments, in most of cases uses penicillin and enrofloxacin for donkeys. Therefore, the high antibiotic resistance in donkeys to these two antibiotics (in our study) may be occurred due to these successive and ectopic treatments, but in bovines the best choices for antibiotic treatment were enrofloxacin, trimethoprim, and tetracycline. So due to the high administration of antibiotics, the prevalence of antibiotic resistance against these antibiotics was so high in bovine's milk samples in the current study. Veterinary interventions and antibiotic administrations are not common in camels in Iran. In most regions of Iran, camel's distribution is so low and reports about the occurrences of diseases in camels usually were low in Iran. Therefore based on our results, it seems logical that 10 The Scientific World Journal the prevalence of antibiotics resistance in camel samples was low.
Our results showed that 58.82% of isolated E. coli from raw milk and traditional dairy products were tetracycline resistant. As well as an other study, substantial numbers of human E. coli isolates were tetracycline resistant [63]. Recent study indicated that 61% of E. coli which isolated from humans, cattle, swine, and food during 1985 to 2000 was susceptible to all 13 antimicrobials tested, while 27% of them were resistant to tetracycline, 26% to sulfamethoxazole, 17% to cephalothin, and 13% to ampicillin [63]. As the present study indicates, there is a high antibiotic resistance in E. coli which isolated from raw milk and traditional dairy products.
A report from Scotland revealed that the frequency of antibiotic-resistant E. coli was 10.6% in milk samples and suggested that the prevalence of antibiotic-resistant strains in milk was higher in housed bovines than in outdoor bovines. This study showed that out of 1125 E. coli isolates tested, 22.2% were antibiotic resistant, and out of these resistant strains, 42.4% were resistant to more than one antibiotic [64]. Another study in Iran indicated that the resistance of the E. coli strains isolated from traditional cheese, ice cream, and yoghurt to ampicillin and gentamycin was the most common findings (44.4%), followed by resistance to erythromycin (33.3%), amoxicillin (11.1%), tetracycline (11.1%), and nalidixic acid (11.1%) [51]. Our study showed that two of the most prevalent resistant genes of E. coli were tetA + tetB (64.70%), and this is similar to previous study that was performed on cow's milk [65].
Unfortunately, indiscriminate prescribing of antibiotics in veterinary is so high and in a report from Netherland 300 000 kg of antibiotics are used annually on veterinary prescription in animals [66]. These high amounts of antibiotics can be kept in animals and lead to antibiotics resistance in bacteria. Then resistance genes of bacteria shedding from milk and meat can be used in humans. So these resistance genes will be causing drug resistance in human too. Our results showed that simultaneous application of culture and PCR can easily be effective in diagnosis of pathogens and their resistance factors especially in food for human consumption.
Our results revealed that (i) using raw milk without pasteurization and milking with unsanitary methods and using traditional dairy products that are produced in unsanitary conditions and probably from unpasteurized milk are the main resources for growth, proliferation, and survival of E. coli and cause several disorders for human. Therefore, vaccinating dairy animals (if necessary), improving methods of milking, checking milking halls in order to detect E. coli especially in the animal feces monthly, fumigating milking halls frequently, observing hygiene during milking, boiling milk, using pasteurized milk for dairy products, keeping dairy products in cool and dry places, away from sunlight, and finally preventing contamination of dairy products with extrinsic factors like insects and dust are the best ways to prevent E. coli infections. (ii) Chloramphenicol is a forbidden antibiotic, and the high antibiotic resistance to drug in our study indicated the irregular and unauthorized use of this antibiotic, in veterinary treatment in Iran. Unfortunately, veterinarians in many fields of veterinary such as large animal internal medicine, poultry, and even aquaculture use this antibiotic as a basic one. Therefore, in a very short period of time, antibiotic resistance will appear. (iii) We recommend PCR method as an accurate, safe, and fast diagnostic one for detection of pathogens in milk and dairy products (iv) due to antibiotic resistance especially in E. coli, the veterinarians should pay more attention to prescribing the antibiotics.
(v) E. coli virulence genes especially stx1, stx2, eae, cnf1, cnf2, iutA, papA, and finally traT are well distributed in milk and dairy products in Iran. (vi) STEC serotypes and especially AEEC are the predominant strains and O157, and O26 serotypes are the predominant serotypes of bacterium in animal raw milk and traditional dairy products in Iran. (vii) Bovine and camel raw milk and soft cheese and ice cream had a highest and lowest frequency rate of pathogenic E. coli in raw milk and dairy product samples in Iran, respectively. (viii) In order to prevent antibiotic resistance in bacteria, we should apply antibiotics more cautiously in animals, detect resistance genes, and finally use different antibiotics periodically. Our results recommended the use of PCR for detection of antibiotic resistance genes of bacteria as a safe, rapid, and accurate method in laboratories.