Prevalence, virulence genes, and antimicrobial profiles of Escherichia coli O157:H7 isolated from healthy cattle in Tunisia

Introduction: Shiga toxin-producing Escherichia coli (STEC) O157:H7 is associated with intestinal infection in humans and is considered an important cause of food-borne diseases. The aim of the study was to assess the incidence of E. coli O157:H7 in fecal samples of healthy cattle collected in slaughterhouses (n = 160) and from five farms (n = 100). Methodology: E. coli isolates were detected on MacConkey agar. A total of 236 E. coli isolates were recovered from fecal samples of healthy cattle. We used sorbitol MacConkey medium to detect non-sorbitol fermenting colonies. These bacteria were examined for the presence of O157:H7 antigen by latex agglutination. The isolation of E. coli O157:H7 has been confirmed with PCR amplification of rfbEO157 and fliCH7 specific genes for serogroup O157 and with multiplex PCR of stx1, stx2 , eaeA, and ehxA . All isolates were examined for their susceptibility to 21 antibiotics by the disc diffusion method. Results: Of the 236 E. coli isolates, 4.2% (10/236) were positive for STEC O157:H7. Shiga toxin gene ( stx2 ) and ehxA were present in 70% of isolates, stx1 and eae were confirmed in 60% of the isolates. Other virulence factors screened ( fimH , sfa/focDE , cdt3 , traT , iutA, and hlyA ) were present among the 10 isolates. All E. coli O157:H7 isolates were sensitive to amoxicillin/clavulanic acid, cefotaxime, cefepime, aztreonam, colistin, and sulfamethoxazole/trimethoprim. All isolates belong to the phylo-group E. Conclusions: This is the first study of the incidence of E. coli O157:H7 in cattle in Tunisia. Our finding proves the existence of STEC O157:H7 in healthy animals producing food for human consumption which could be a source of food-borne disease.

Strains belonging to the subgroup of Shiga toxinproducing strains (STEC) are distinguished by certain EHEC serotypes, which are linked to outbreaks in humans and cause clinical sickness. STEC is a foodborne bacteria which have been associated with many epidemics across continents especially serotype O157:H7 [4]. Strains have been isolated from feces of healthy ruminants like cattle, goats, and sheep which can be natural reservoirs of these pathogens [5].
E. coli O157:H7 is the dominant serotype of the STEC group associated with human infections. The first identification of this serotype as a pathogen was in 1982 during an outbreak of hemorrhagic colitis in Oregon and Michigan, U.S.A. [6]. STEC O157:H7 can cause acute infections with a spectrum of human illnesses ranging from abdominal pain and bloody diarrhea to fatal diseases, like hemolytic-uremic syndrome (HUS) and hemorrhagic colitis (HC). The main STEC O157 infections are food-borne, particularly concerning cattle sources [7].
The STEC strains possess Shiga toxins (stx1 and stx2) genes that are considered the major virulence factors of these strains. Stx2 is associated more closely with the HUS sickness than stx1 [8]. Other important virulence determinants are: intimin protein, encoded by eae gene and important for attaching and effacing activity within the colonization of host intestinal mucosa and causing severe human infections, and enterohemolysin encoded by the plasmid-and phagecarried enterohemolysin (ehxA) gene [9]. STEC O157:H7 isolates have been detected in North Africa from humans, animals, and food products. An Algerian study identified a rate of 7% in the bovine carcasses [10]. In Morocco, the frequency of STEC O157:H7 was 9%, 9.1%, and 11.1% from raw meat products, dairy products, and marketed meat respectively [11,12]. A Tunisian study confirmed that 3.4% of E. coli isolates among human stool samples were STEC and the rate of E. coli O157:H7 was 0.3% [9]. In Egypt, a survey confirmed that the prevalence among beef samples, chicken samples, and lamb samples was 6%, 4%, and 4% respectively [13].
An increasing number of STEC O157 outbreaks are related to the human consumption of fruits and vegetables contaminated with domestic or wild animal feces. E. coli O157:H7 is transmitted to humans by the consumption of contaminated foods like raw meat, undercooked meat, and raw milk. Water and foods contaminated by fecal material and crosscontamination through food production and processing will lead to STEC infection [14]. Therefore, the objective of our study was to assess the incidence, virulence genes, and antimicrobial resistance profiles of E. coli O157:H7 in fecal samples of healthy cattle. To the best of our knowledge, this is the first report of E. coli O157:H7 in healthy cattle in Tunisia.

Samples Collection
The samples analysed in this study were collected as part of a research project dedicated to the study of antibiotic resistance of bacteria isolated from the main five slaughterhouses in the region of the greater Tunis and from cattle farms located in the governorate of Bizerte, which provides 11% of the national production of red meat.
All the feces samples were collected by rectal swabbing, by rolling-rubbing the rectal mucosa. There are two types of samples; firstly, fecal samples from 160 cattle intended for slaughter were collected between December 2016 and April 2017. These samples were collected from five slaughterhouses in the greater Tunis, designated as A, B, C, D, and E. In the second samples category, a total of 100 fecal samples were gathered from healthy cattle between March and November 2018 from cattle farms located in the governorate of Bizerte. Samples were transported appropriately to the laboratory in ice-cooled containers for bacterial isolation and further investigations.

Selective isolation of E. coli O157:H7
Fecal samples were enriched in buffered peptone water overnight at 37 °C, then 10 µL were cultured by the streak plate technique on MacConkey agar and incubated for 18 to 24 hours at 37 °C. One putative colony was subcultured from each plate onto brain heart infusion agar for confirmation as E. coli. The identification of E. coli colonies was performed by API 20E galleries (bioMérieux). The bacterial colonies were cultivated onto sorbitol MacConkey agar (Oxoid) supplemented with cefixime-tellurite (CT-SMAC) and incubated for 18-24 hours at 37 °C. On this medium, most STEC O157:H7 are distinct from other STEC by their inability to ferment sorbitol. On each plate with sorbitol non-fermenting (straw color or colorless) colonies, one colony was subcultured as probably E. coli O157.

Agglutination test of O157
Each non-sorbitol fermenting colony isolated on SMAC plates was examined for the existence of the O157 antigens by agglutination latex reagent (DrySpot™ E. coli O157 Latex Agglutination Test, Oxoid).

Affirmation of E. coli O157 by PCR
All non-sorbitol fermenting E. coli isolated and O157 agglutination-positive were examined for the existence of rfbEO157 gene and fliCH7 by simplex PCR [15]. The PCR condition was as follows: initial denaturation at 94 °C for 5 minutes; 35 cycles of denaturation at 94 °C for 45 seconds, annealing at a specific temperature for 45 seconds (Table 1), extension at 72 °C for 45 seconds; and a final extension (72 °C, 7 minutes).
A multiplex PCR for stx1, stx2, uidA, ehxA, and eae was achieved for the O157:H7 strains, and primers are listed in table 1 [16]. The thermal cycling program of multiplex PCR was as follows: the denaturation: 95 °C for 5 minutes, followed by 25 cycles of 95 °C for 1 minute, annealing at 56 °C for 1 minute, the extension at 72 °C for 1 minute, and the final extension at 72 °C for 5 minutes. The gel electrophoresis was used to separate PCR products by using a 2 % agarose gel in a TBE buffer containing ethidium bromide.
The stx1 and stx2 amplifications were sequenced in order to prove that the amplicon matched the stx1 and stx2 sequences. The gained sequences were aligned with the data sequences in NCBI (http://www.ncbi.nlm.nih.gov).

Detection of Phylogenetic groups
The phylogenetic groups (A, B1, B2, C, D, E, F) were detected among the isolates by the quadruplex PCR method developed by Clermont et al. [19]. The phylogroups were determined based on the presence of the chuA, yjaA genes, and TspE4-C2 fragment detected by quadruplex PCR (A, B1, B2, D), and C, E were further identified using specific primer sets (Table 1).

Results
In our study, 236 E. coli isolates were collected from the examination of 250 fecal samples of healthy cattle in Tunisia. Out of 236 E. coli isolates, 159 were from cattle in slaughterhouses and 77 from cattle farms. Of these E. coli strains, 100% were positive for methylred, lactose, and indol, and 100% were negative for urease, citrate, and H 2 S production. The results revealed that 10 E. coli were sorbitol nonfermenting on CT-SMAC and these 10 (4.2%) strains were E. coli O157:H7. Out the 10 strains; 6 isolates were isolated from healthy cattle in slaughterhouses and the other from healthy cattle on farms.
The confirmation of E. coli O157 by latex agglutination testing revealed that all isolates were O157 positive. All of these isolates were confirmed as E. coli O157:H7 via screening of rfbO157 and flicH7 genes by specific primers.
PCR analysis of the 10 E. coli O157 isolates reveals that uidA, flicH7, and O157 genes were present in all strains. Stx2 and ehxA genes were present in 7 isolates (70%), while stx1 and eae were confirmed in six isolates (60%).
The O157 isolates were further tested for 13 virulence factors. All isolates carried at least one virulence gene tested. Out of 10 isolates, 60% carried more than three virulence genes tested. FimH was the most frequent virulence gene and was detected in 90% (9/10) of the isolates, followed by sfa/focDE in 60%. The frequency of cdt3, traT, and iutA among the isolates was 50%, 50%, and 40% respectively, whereas, hly was found in one isolate ( Table 2). None of the isolates harbored cnf1, aer, papA, bfpA, papG allele III, ibeA, or fyuA.

Discussion
Human infections caused by STEC O157:H7 are associated with food of animal origin or plants contaminated with the feces of these animals. In particular, cattle, sheep, and goats have been demonstrated as the main natural reservoirs for STEC O157:H7 and play an important role in the public health concern [7]. This study was conducted to evaluate the incidence of E. coli O157:H7, antimicrobial profiles, and virulence genes in fecal samples of healthy cattle collected from slaughterhouses and cattle farms in  Tunisia. This is the first report concerning the presence of E. coli O157:H7 in cattle in Tunisia. Our findings demonstrated that among 236 E. coli isolates, ten E. coli O157:H7 were detected at a rate of 4.2%. These isolates were cultured on CT-SMAC agar as non-sorbitol fermenters and were confirmed as STEC O157 by using latex agglutination and PCR. This is consistent with previous research that found E. coli O157: H7 in cattle feces samples and carcass swabs in slaughterhouses, with rates of 4.7% and 2.7%, respectively, in Ethiopia [7]. In a study in United Arab Emirates, the frequency of E. coli O157:H7 among slaughtered cattle was 1.4% [20]. An Algerian study reported an incidence of E. coli O157 in more than 7% of bovine carcasses [10]. In Morocco, the incidence of E. coli O157:H7 in dairy products and marketed meat products was 9.1% and 11.1% respectively [12]. In Tunisia, 327 E. coli strains were isolated from diarrheic and non-diarrheic people. By using PCR techniques, it has been demonstrated that 11 isolates (3.4%) express the stx1 and stx2 genes encoding for STEC and only one (0.3%) was confirmed as E. coli O157:H7 [9].
In this study, the rate of these bacteria among healthy cattle in slaughterhouses was higher than that of healthy cattle from farms, maybe the stress of transporting cattle from the farm to the slaughterhouse results in increased bacterial excretion. McCluskey et al. [21] confirmed that there were significantly higher rates of E. coli O157 in lambs that were transported and held for ≥ 18 hours. Furthermore, animals were not perfectly fed during the transition and holding prior to slaughter; withholding of food has the potential effect on colonization with E. coli O157.
In the present study, one putative colony per sample was selected for confirmation as E. coli O157:H7. Examination of up to ten colonies per plate [22] may have led to more O157:H7 isolates and this criterion was confirmed by many studies. Furthermore, some characteristics of these bacteria can change during lab manipulations. For example, ehxA gene is located on a plasmid that could be lost either naturally in the animal host, or during lab manipulations resulting in increased sensitivity of the method. On the other hand, immunomagnetic separation would certainly have improved the rate of isolation of STECs.
In Africa, the highest incidence of cattle was 31.2% as represented in two studies [23,24]. In Asian countries, the highest rate was 12.22% in Jordanian cattle [25] and the lowest (0.13%) was evaluated in cattle from Taiwan [26]. According to a meta-analysis of 40 studies, in several states of the USA the estimated incidence was 7.60% [27] while in California it was highly variable, from 0 to 90% [28].
Healthy cattle can be the main reservoir for prospecting human infection and play an important role in the epidemiology of STEC infections. Moreover, most human diseases caused by STEC bacteria originate from cattle [29]. The existence of STEC O157:H7 in our study among animal feces in slaughterhouses highlighted the possible contamination of meat products prepared for human consumption. On the other hand, identifying the STEC O157:H7 in humans is very important for public health objectives, like identifying outbreaks. Underlining the great scarcity of studies in Tunisia, the identification of STEC O157:H7 and non-O157 by Al-Gallas et al. [9] among humans has important benefits for public health and proves the need for epidemiological surveys on STEC infection in this country. The detection of E. coli O157:H7 in cattle and humans in Tunisia calls for further epidemiological assessment to detect whether a case is part of an outbreak, the outbreak source, and the spread prevention of it.
Antimicrobial resistance is considered a global health threat. Animal products have been demonstrated as reservoirs of antimicrobial resistant bacteria because the same genes encoded for antimicrobial resistance were demonstrated in the bacteria of animal food and in humans [30].
Our results show that all E. coli O157:H7 isolates were susceptible to amoxicillin/clavulanic acid, cefotaxime, cefepime, aztreonam, colistin, and sulfamethoxazole/trimethoprim. Previous studies in animals reported different antibiotic resistance profiles of E. coli O157:H7 isolates. One study reported that in Hawassa (Ethiopia), all E. coli O157:H7 isolates were susceptible to cefotaxime, ceftriaxone, gentamycin, kanamycin, and nalidixic acid [7]. A further report from United Arab Emirates showed that all isolates were susceptible to cefotaxime, chloramphenicol, ciprofloxacin, norfloxacin, and polymyxin B [20]. However, a Saudia study reported that the isolates were resistant to all tested antibiotics [31]. One study from Iran revealed that the resistance rate to gentamycin, ampicillin, erythromycin, amoxicillin, and tetracycline was 56.0%, 48.0%, 40.0%, 16.0%, and 12.0% respectively [32]. A UK study in humans showed that the resistance profile among 327 STEC O157 to ampicillin, streptomycin, trimethoprim/sulphonamide, and tetracycline was 5.8% followed by the resistance rate to ciprofloxacin (2.6%) and chloramphenicol (2.1%) [33].
A study conducted in Latin American countries has documented 78.5% sensitivity to all the antimicrobial agents in 14 O157 STEC strains from cattle. Resistance to streptomycin, trimethoprim, and sulfonamide was found in three strains [34].
Antimicrobial resistance variation might be due to the expression of resistance genes among bacteria in animals, the environment, or humans and this variation in resistance rates may also be an indicator of animal husbandry and agricultural use of antibiotics and antimicrobials [35].
On the other hand, more than 40% of the isolates were resistant to cefuroxime and streptomycin, perhaps via co-resistance or cross-resistance or inappropriate or wide use of these drugs for prophylactic purposes and treating infections. In fact, no multi-drug resistance was observed in all the strains tested, in agreement with a Turkish study, where a low resistance rate to cephalothin, streptomycin, and nalidixic acid, was detected [36]. Vali et al. [37] showed no multi-drug resistance among E. coli O157 strains isolated from beef cattle farms and identified a low prevalence of resistance to cephalothin, sulphamethoxazole, streptomycin, sulphonamide compounds, and nalidixic acid. However, some studies have demonstrated that there has been an increase in the antimicrobial resistance of STEC O157:H7 [38,39]. Previous studies conducted around the world revealed that the majority of antibiotic-resistant isolates were discovered in animal farms, which more commonly used antibiotics for prophylactic and treating purposes, but our findings of less resistance and no MDR were unexpected.
It is debatable whether it is safe to use antimicrobial drugs in humans to prohibit HUS due to lysis of the bacteria and release of the Shiga toxins in the gut. However, reports have revealed that using some antimicrobials in the early phase of the disease may prevent HUS advancement [36].
In our study, most strains exhibited an intermediate resistance pattern, suggesting the possibility of future resistance. The intermediate susceptibility profiles should be elevated and taken into consideration with resistance results because it means the organism may be on the way to becoming resistant. In fact, antibiotics are not recommended for O157:H7 infections as they can induce the bacteria to express more Shiga toxin and make the disease worse with the risk of triggering hemolytic-uremic syndrome [40]. However, knowing the antibiotic resistance of O157:H7 strains can help track them in an outbreak and be a useful tool for selective isolation.
Shiga toxins (stx genotypes) are important clinical outcome factors that correlate with HC and HUS, as well as higher pathogenicity in strains carrying the stx2 genotype [41]. The eae gene encodes for an intimin protein, which is important for attaching and effacing activity in host intestinal cells and causes severe human illnesses, particularly HUS [42]. Furthermore, a hemolysin produced by STEC called enterohemolysin is encoded by the hlyA gene and causes erythrocyte lysis, which participates in iron intake in the intestine. This gene is commonly used as an epidemiological marker of STEC strains [43].
In the present study, the stx2 gene was present in most isolates (7/10), and eae and ehxA were found in more than half of the isolates. Many studies have found that the virulence factors stx2 and eaeA are clinically significant and are associated with the severity of human disease, particularly HUS [44,45]. In United Arab Emirates, Shiga toxin gene (stx2) was confirmed in all 24 E. coli O157 from camels, cattle, and goats. The eaeA and hlyA genes were present in 79.2% and 66.7%, respectively, whereas stx1 was absent in all isolates [20].
The presence of ehxA + eae and ehxA + eae + stx2 is significantly associated with HUS and O157:H7 isolates [46]. In our study, two isolates (T109 and T125) harbored stx2, eae, and ehxA, and one isolate (BS10) harbored stx1, eae, and ehxA suggesting that the existence of more than one virulence factor, particularly eae, and ehxA could be associated with more severe clinical outcomes in O157 infections. Hua et al. [46] mentioned that the presence of ehxA with stx and eae, can be used as a risk predictor for HUS in STEC infections.
Generally, all E coli O157:H7 possess intimin (eae). Some strains may lose the Shiga toxin phage and be negative for stx1 and sxt2, but the absence of eae in O157:H7 is very unusual. Furthermore, the lack of ehxA in O157:H7 is also unusual, although this gene is located on a plasmid that could be lost either naturally in the animal host or during lab manipulations. Negative results for the presence of eae and ehxA genes were verified by simplex PCRs with eae and ehxA primers instead of multiplex PCR.
The absence of eae and ehxA among our isolates recognizes these isolates as atypical O157:H7 which represents a less serious threat to public health. For typical STEC, the only reservoir is humans. However, atypical STEC can have both animal and human reservoirs and may also be associated with human diarrhoeal disease [47].
In the same way, the atypical EPEC (enteropathogenic E. coli) strains may be less virulent than the typical isolates. One reason may be the lack of the adherence factor (EAF) plasmid among the atypical strains. However, atypical strains have not been confirmed to be less pathogenic, and these bacteria have other virulence factors that may compensate for the absence of the EAF plasmid [48].
This study showed that 9 STEC strains harbored fimH and half the isolates harbored sfa/focDE, cdt3, traT, and iutA. These factors were identified in a previous study on E. coli from dairy farms in the USA [49]. In an Iranian study of STEC, they found papA, cnf1, traT, and cnf2 were the most common virulence genes [50]. The detected factors contribute to virulence, which affects host cell processes and contributes to bacterial pathogenesis. The findings of these virulence factors in our isolates in association with the high frequency of stx1, stx2, and ehxA suggest that STEC O157 in Tunisian calves may pose a serious public health concern.
The findings of our study revealed that all E. coli O157:H7 isolates belonged to phylogroup E which usually corresponds to commensal strains. This was identical to the report of Tenaillon et al. [51]. A study in Brazil demonstrated that E. coli belonging to phylogroups E and B1 were isolated from cattle, whereas phylogroups A and F were from poultry, and B2 and D were associated with isolates from water buffalo [52].

Conclusions
The frequency of E. coli O157:H7 in healthy cattle indicates a possible risk for a public health concern. The detection of STEC O157:H7 in this study among cattle and previously in humans in Tunisia have important benefits for public health and calls for the public health system in our country to track food-borne outbreaks. The existence of STEC O157:H7 in animals intended for slaughter highlighted the possible contamination of meat products prepared for human consumption. The high prevalence of stx1, stx2, and ehxA together with other virulence factors suggest that STEC O157 in Tunisian calves may pose a serious public health concern. Our study reveals the necessity of regular screening animals for E. coli O157:H7 in order to control this pathogen. Moreover, the frequency of O157:H7 in slaughterhouse animals indicates that the risk is significant for public health. Therefore, it is important to take the necessary precautions during the slaughter and skinning of animals to prevent cross contamination of meat by this pathogen. Clinical results must be obtained to evaluate the actual influence of food contamination on humans in Tunisia.

Funding
This work was supported by the research project PEER 7-349 funded by the USAID "Monitoring of antimicrobial resistance of bacteria for a better health of animals in Tunisia". Prof. Lilia Messadi is the recipient of the funding (number PEER 7-349).

Authors' Contributions
Ghassan Tayh designed the study, performed the experimental work (the microbiological and molecular tests), collected the data, analyzed and interpreted the data and drafted the manuscript. Salma Mariem Boubaker and Rym Ben Khedher collected samples and helped in performing the experimental part of the manuscript. Mounir Jbeli collected samples. Faten Ben Chehida, Aymen Mamlouk and Monia Dâaloul-Jedidi participated in the project design. Lilia Messadi designed and supervised the study, and contributed to final writing and editing the manuscript. All authors read and approved the final version of the manuscript.