The prevalence, distribution and characterization of Shiga toxin-producing Escherichia coli (STEC) serotypes and virulotypes from a cluster of bovine farms



Declan J. Bolton, Department of Food Safety, Teagasc Food Research Centre, Ashtown, Dublin 15, Ireland. E-mail:



To assess the prevalence of Shiga toxin-producing Escherichia coli (STEC) on a cluster of twelve beef farms in the north-east of Ireland.

Methods and Results

Samples were screened for stx1 and stx2 using PCR. Positive samples were enriched in mTSB and STEC O157 isolated using immunomagnetic separation. Enrichment cultures were plated onto TBX agar to isolate non-O157 STEC. All isolates were serotyped and examined for a range of virulence genes and their antibiotic resistance phenotype determined. Eighty-four isolates of 33 different serotypes were cultured from the 13·7% of samples that were stx positive. The most prevalent serotype was O157:H7, the most common Shiga toxin was stx2, and a variety of virulence factor combinations was observed. O-:H-, O26:H11, O76:H34, O157:H7, O157:H16 and OX18:H+ also carried eaeA and hlyA genes. Twenty-nine per cent of strains were resistant to at least one antibiotic, 48% of which had multiple drug resistance (MDR) with O2:H32 displaying resistance to five antibiotics.


The ubiquitous nature of STEC on beef farms, the detection of stx+ eaeA+ hlyA+ in the serotypes O-:H-, O157:H16 and OX18:H+ in addition to O157:H7 and O26:H11 and the widespread distribution of antibiotic resistance are of public health concern as new virulent STEC strains are emerging.

Significance and Impact of the Study

This study found no relationship between serotype and antibiotic resistance, therefore negating efforts to isolate serotypes using specific antibiotic supplemented media. The data presented provide further evidence of the emergence of new STEC virulotypes of potential public health significance.


Shiga toxin-producing Escherichia coli (STEC), also known as verocytotoxigenic Escherichia coli (VTEC), cause severe illnesses that are potentially fatal to humans. Clinical manifestations include diarrhoea, haemorrhagic colitis (HC), haemolytic uraemic syndrome (HUS) and thrombocytopenic purpura (TTP) (Griffin and Tauxe 1991). STEC was first documented as an outbreak pathogen in 1982 (Riley et al. 1983) and since then has become a ubiquitous foodborne pathogen worldwide. Between 2005 and 2009, there were 16 263 confirmed human STEC cases in EU member states, with 2–6 deaths reported annually (Anon 2011). While STEC O157 is the most commonly detected serogroup, non-O157 infections account for almost half of the currently confirmed STEC cases in Europe and outnumber O157 cases in the USA (Bosilevac and Koohmaraie 2011). Ireland has the highest per capita incidence (4·41 STEC cases per 100 000) in Europe (EFSA 2012).

Ruminants, especially cattle, are the primary source of both O157 and non-O157 STEC (Pennington 2010; Bolton 2011; Bosilevac and Koohmaraie 2011), and beef is considered to be an important source of STEC O157 and non-O157 human infection (Caprioli et al. 2005). Within the EU, 2·2–6·8% of bovine animals carry STEC, while the O157 prevalence ranges from 0·5 to 2·9% (Anon 2011). Infected cattle usually excrete between 102 and 105 CFU per gram of faeces although counts as high as 108 have been recorded in supershedders (Besser et al. 2001; Fukushima and Seki 2004). These STEC are transferred from faeces on the hide and the gastrointestinal tract to the carcass during slaughter. In Europe, the prevalence of non-O157 and O157 STEC on fresh meat ranges from 0·3–2·3% and 0·1–0·7%, respectively (Anon 2011). Contaminated meat, mainly beef, was identified as the source of human infection in 16 of the 40 STEC outbreaks in EU Member States where a source was identified, between 2007 and 2009.

Much ambiguity surrounds the dissemination of STEC on and between farms. Movement of cattle to/from farms, contaminated feed and water troughs (Sargeant et al. 2004; Zhao et al. 2006), environment (Avery et al. 2004), bedding, insects, birds, rodents (Nielsen et al. 2004) and watercourses (Cooley et al. 2007) have all been implicated as possible routes of contamination.

STEC infection is associated with the presence of one or both stx genes (stx1 and stx2) in conjunction with the putative virulence genes hlyA, eaeA and/or saa (Paton et al. 2001; Gyles 2007). Stx2 is more commonly associated with HUS (Kleanthous et al. 1990; Nataro and Kaper 1998; Boerlin et al. 1999). Intimin, encoded by the eaeA gene, is found on the locus of enterocyte effacement (LEE). It is responsible for the intimate attachment of STEC to the intestinal cells causing the formation of attaching and effacing lesions in the intestinal epithelium. Enterohaemolysin, encoded by the hlyA gene located on the pO157 plasmid, releases haemoglobin from red blood cells (Schmidt et al. 1994). STEC agglutinating adhesin (saa) located on the pO113 plasmid encodes an outer membrane protein, facilitating colonization of the intestinal gut wall by LEE-negative isolates (Paton et al. 2001).

The use of antibiotics in the treatment of STEC infections is controversial. Antibiotics are thought to induce the production and release of Shiga toxins by the bacterium causing the disease to progress to HUS (Dundas et al. 2001; Schroeder et al. 2002). However, it has also been reported that certain antimicrobial agents suppress toxin release, preventing HUS (Shiomi et al. 1999; Murakami et al. 2000). Regardless, increased resistance to therapeutically valuable antimicrobial agents will inevitably facilitate further spread and therapeutic failure in the future. Determining the antibiotic-resistant phenotype of STEC may indicate the distribution of antibiotic-resistant genes in the population and therefore allow the development of potential control strategies (Ateba and Bezuidenhout 2008).

In this study, we examined the incidence, prevalence, distribution and virulence factor profiles of STEC serotypes on twelve beef farms within the same geographical area. Resistance to fourteen clinically important antibiotics was also investigated.

Materials and methods

Sampling design and sample collection

Samples were collected from twelve beef farms in the north-east of Ireland in the same geographical area (Figure 1). The distance between farms ranged from 0·5 to ~6·5 km. Farms varied from beef suckler to dairy herds, with no known contact of personnel or animals between farms. Each farm was sampled four times, once per season with 10 samples taken per visit. A total of 650 samples (620 faecal and 30 slurry) were collected over a 16-month period between January 2008 and April 2009. Samples were obtained from freshly voided faeces and collected in 150-ml Sterilin jars (Dublin 15; Sterilin ltd., Medical Supply Company, Mulhuddart, Ireland) using Sterileware sampling scoops (Medical Supply Company). On three occasions, 10 ml of slurry was obtained directly from the slurry pit. All samples were stored at 4°C in a cool box and transported to the laboratory within 4 h of sampling. All samples were refrigerated at 4°C and processed within 24 h of collection.

Figure 1.

Map showing the location of the 12 farms in this study.

Sample processing

Faeces and slurry samples (10 g) were enriched in 90 ml tryptone soya broth (TSB; Merck, Whitehouse station, NJ, USA) supplemented with 4 μg ml−1 vancomycin (Sigma-Aldrich, St Louis, MO, USA), and each sample was homogenized and incubated overnight at 37°C. A 10 g sample was also inoculated with a positive control, 1 ml Jack in the Box (JIB) overnight culture (Bell et al. 1994). A 1 ml aliquot of each enrichment broth was centrifuged at 20 412 g for 2 min (Eppendorf, Davidson and Harvey, Belfast, UK), and the pellet was washed with 1 ml sterile distilled water (dH2O) and resuspended in 200 μl sterile dH2O. Genomic DNA was extracted by boiling the cells for 10 min and centrifuging to remove cell debris from the suspension. The remainder of the enrichment was stored at 4°C. Recovered DNA was screened for the presence of virulence factors, stx1, stx2, eaeA and hlyA, by multiplex PCR in a Peltier Thermal Cycler (PTC-200; MJ Research Inc., Watertown, MA, USA) using the components and conditions described by Paton and Paton (1998). The oligonucleotide primer sequences are shown in Table 1. The reaction mixture contained 2 μl of sample DNA, 250 nmol l−1 of each primer pair (Eurofins MWG GmbH, Ebersberg, Germany), 1× Green GoTaq® reaction buffer (Promega, Madison, WI, USA), 0·2 mmol l−1 deoxynucleotide triphosphate mix (dNTP) (Promega), 2 mmol l−1 MgCl2 (Promega) and PCR-grade water (Sigma-Aldrich), giving a final volume of 25 μl. All PCR cycles included a negative control replacing the DNA with 2 μl of PCR-grade water, and a positive control containing the JIB culture strain that is positive for stx1, stx2, eaeA and hlyA. Stx-positive samples were also screened for the presence of saa using the components and conditions described by Paton and Paton (2002). The positive control for saa was a sequenced strain from a previous study carried out in Teagasc Food Research Centre, Ashtown (Monaghan et al. 2011). PCR products (10 μl) were separated by electrophoresis on a 1·5% (wt/vol) agarose gel containing 4 μl of ethidium bromide and visualized under UV light (GelDoc 2000 system; Bio-Rad Laboratories, Hercules, CA). Product size was determined using a 100-bp DNA ladder (Qiagen, Hilden, Germany).

Table 1. Oligonucleotide primer sequences for virulence genes
PrimerTarget (gene)Sequence (5′-3′)ReferencesAmplification
Product size (bp)
stx 1F stx 1 ATAAATCGCCATTCGTTGACTACPaton and Paton (1998)180
stx 2 F stx 2 GGCACTGTCTGAAACTGCTCCPaton and Paton (1998)255
eaeA F eaeA GACCCGGCACAAGCATAAGCPaton and Paton (1998)384
hlyA F hlyA GCATCATCAAGCGTACGTTCCPaton and Paton (1998)534
TIR F tir CATTACCTTCACAAACCGACKobayashi et al. (2001)1550
EspAa espA CACGTCTTGAGGAAGTTTGGMcNally et al. (2001)299
EspBa espB CGATGGTTAATTCCGCTTCGMcNally et al. (2001)304
D1 etpD CGTCAGGAGGATGTTCAGSchmidt et al. (1997)1062
wkat-B katP CTTCCTGTTCTGATTCTTCTGGBrunder et al. (1996)2125
lpfO141-F lpf O157/OI-141 CTGCGCATTGCCGTAACSzalo et al. (2002)412
O154-FCT lpf O157/OI-154 GCAGGTCACCTACAGGCGGCToma et al. (2004)525
toxB.911F toxB ATACCTACCTGCTCTGGATTGATarr et al. (2002)602
iha-I iha CAGTTCAGTTTCGCATTCACCSchmidt et al. (2001)1305
saa F saa CGTGATGAACAGGCTATTGCPaton and Paton (2002)119
rfbO157 F rfbO157 CGGACATCCATGTGATATGGPaton and Paton (1998)259

Isolation of Shiga toxin-positive E. coli

A 1 ml aliquot of samples positive for stx1 and/or stx2 was serially diluted in 9 ml sterile maximum recovery diluent (MRD) (Oxoid, Hampshire, UK) and plated onto tryptone bile X-glucuronide agar (TBX) (Merck, Whitehouse station, NJ, USA). Following overnight incubation at 37°C, five colonies with differing colony morphology were taken from each sample and restreaked onto nutrient agar (NA) (Oxoid) and eosine methylene blue agar (EMB) (Oxoid) and incubated overnight at 37°C. Genomic DNA was extracted from presumptive E. coli (those with a metallic green sheen on EMB), and a confirmation multiplex PCR was carried out to determine the presence of stx1, stx2, eaeA and hlyA (Paton and Paton 1998). Stx-positive isolates were protected on cryoprotect beads (Technical Services Consultants, Lancashire, United Kingdom) and stored at −20°C.

Immunomagnetic separation

Immunomagnetic separation using immunomagnetic beads coated with anti-E. coli O157 antibody (Dynabeads® anti-E. coli O157, Dynal A.S., Norway) was used to isolate O157 from PCR-positive samples from the initial screening PCR. Briefly, 500 μl of phosphate-buffered saline (PBS) (Oxoid), 500 μl of enrichment broth and 10 μl of anti-E. coli O157 Dynabeads® were added into wells 1 and 2 of the immunomagnetic separation tray. Wells 3 and 4 contained 1 ml of PBS only, and well five contained 100 μl of PBS. The trays containing the samples were loaded into the Dynal Bead Retriever (Dynal, Oslo, Norway). The contents of well five containing the washed Dynabeads® were then plated in duplicate onto sorbitol MacConkey agar (SMAC, Oxoid) supplemented with cefixime–tellurite (CT) (Oxoid) and incubated at 37°C for 24 h. Five typical colonies were then selected and biochemically tested using latex auto agglutination kit (Oxoid) for the presence of O157. Isolates that had a positive reaction were confirmed as O157 using O157-specific primers (Table 1). O157-positive isolates were then protected on cryoprotect beads and stored at −20°C.


Serotyping of the O (lippopolysaccharide) and H (flagellar) antigens was performed by The E. coli Reference Center at Pennsylvania State University, University Park, PA, USA. The O antigen was determined using all available O antisera (O1–O181) with the exceptions of O31, O47, O72, O93, O94 and O121 as these are not designated. All antisera were obtained and absorbed with the corresponding cross-reacting antigens to remove the nonspecific agglutinins. The H antigen was established by polymerase chain reaction–restriction fragment length polymorphism (PCR-RFLP) analysis of fliC gene responsible for flagella (Machado et al. 2000).

Detection of putative virulence genes

One cryoprotect bead of each STEC isolate was individually cultured in 10 ml of TSB overnight at 37°C. Each STEC isolate was tested for the presence of putative virulence factors, including genes associated with STEC O157:H7 (katP, etpD and hlyA), genes thought to encode alternative attachment mechanisms to eaeA, including saa, iha, toxB, lpfAO157/OI-141 and lpfAO157/OI-154, and genes located on the LEE (tir, espA and espB). Primers for these PCRs are listed in Table 1 using the components and conditions set by each of the referenced authors, also listed in Table 1. PCR products (10 μl) were separated by electrophoresis on a 1·5% (w/v) agarose gel containing 4 μl of ethidium bromide and visualized under UV light (GelDoc 2000 system). Product size was determined using a 100-bp DNA ladder (Qiagen, Hilden, Germany).

Antimicrobial susceptibility

One cryoprotect bead of each serotyped STEC isolate was individually cultured in 10 ml of TSB overnight at 37°C. Eighty-four isolates were tested for antimicrobial susceptibility against 14 antibiotics using the agar disc diffusion method on Mueller-Hinton agar (Oxoid) following Clinical and Laboratory Standards Institute (CLSI) guidelines (CLSI 2008). Antimicrobial discs (Oxoid) used included the following: ampicillin (Amp 10 μg), cefachlor (Cec 30 μg), cefixime (Cfm 5 μg), chloramphenicol (C 30 μg), ciprofloxacin (Cip 5 μg), doxycycline (Do 30 μg), kanamycin (K 30 μg), minocycline (Mh 30 μg), nalidixic acid (Na 30 μg), norfloxacin (Nor 10 μg), streptomycin (S 10 μg), sulfonamides (Su 300 μg), tetracycline (T 30 μg) and trimethoprim (W 5 μg). These were applied to lawn cultures of each of the 84 isolates in triplicate and incubated at 37°C for 24 h. Zones of inhibition were measured, and the susceptibility (or resistance) of each isolate was determined according to the 2008 CLSI recommendations.


Of the 650 samples analysed, 13·7% (89/650) were stx1 and/or stx2 positive. Eight-four isolates were obtained, representing 33 different serotypes: O-:H-, O-:H10, O-:H11, O-:H12, O-:H14, O-:H16, O-:H18, O-:H21, O-:H46, O-:H48, O2:H+, O2:H25, O2:H27, O2:H32, O3:H12, O26:H11, O33:H11, O69:H-, O76:H34, O88:H8, O113:H4, O113:H36, O118:H16, O136:H12, O150:H8, O153:H+, O153:H40/44, O157:H7, O157:H16, O171:H2, OR:H18, OX18:H38 and OX18:H+, demonstrating a high overall serological diversity (Table 2). There was no seasonal peak in STEC isolation rates (data not shown). O157:H7 (26 isolates) was the most common serotype followed by O-:H18 (7), O-:H10 (6), O2:H27 and O171:H2 (4) with the remainder represented by 1–3 isolates. O157:H7 was also the most prevalent serotype, present on five of the 12 farms followed by O-:H10, O-:H18 and O2:H27, which were detected on four farms each. The STEC serotypes were randomly distributed over the 12 study farms with no discernable pattern.

Table 2. Shiga toxin-producing Escherichia coli source farm, serotypes, associated virulence genes and antibiotic resistance profiles
Farm(s)Serotype (n) stx1 stx2 eaeA hlyA tir saa espA espB iha lpfA O157/OI-141 lpfA O157/OI-154 toxB etpD katP rfbO157 AR1 profile
  1. AR1 = antibiotic resistance. chloramphenicol (C); kanamycin (K); minocycline (Mh); nalidixic acid (Na); norfloxacin (Nor); streptomycin (S); sulfonamides (Su); tetracycline (T) and trimethoprim (W).

3, 4, 8, 10O-:H10 (5)++ 
8O-:H11++++Nor, Su, T
6O-:H18++++Na, S, Su
8, 12O-:H18 (3)++++ 
8O-:H18+++++Mh, T
7O-:H21++K, Mh, T S
3O-:H48+++S, Su
7, 9O2:H27 (2)+ 
6O2:H32+C, Mh, S, T,W
12O88:H8++++K, Mh Su
8O113:H36++++K, Mh, S, W
7O150:H8++++K, Mh, T
12O153:H40/44++++K, Mh, Su
1O157:H7 (5)++++++++++++ 
1O157:H7 (3)++++++++++++C
7, 9, 10O157:H7 (3)++++++++++++ 
1O157:H7 (2)+++++++++++++ 
1O157:H7 (2)++++++++++ 
11O171:H2++S, Su, T
11OX18:H38+++T, W

O-:H12, O3:H12, O26:H11, O76:H34 and O136:H12 carried the stx1 gene only. Some of the O157:H7 and the O157:H16 strains carried both the stx1 and stx2 genes, while the remaining isolates were stx2 positive. O-H-, O26:H11, O76:H34, O157:H7, O157:H16 and OX18:H+ carried stx1 and/or stx2 and the eaeA gene. All of these isolates were also hlyA positive. Of the remaining virulence genes, tir, saa, espA, espB, iha, lpfAO157/OI-141, lpfAO157/OI-154, toxB, etpD and katP were present in 45, 17, 35, 40, 71, 40, 35, 27, 27 and 6% of isolates, respectively. When serotype and virulence profiles were combined, 76 different serotype–virulence profile combinations were obtained.

Twenty-five of the 84 isolates were resistant to at least one of the antibiotics tested (Table 2). Overall, the level of resistance to each of the 14 antibiotics assayed was as follows: ampicillin (0%), cefachlor (0%), cefixime (0%), chloramphenicol (7%; O2:H32 and O157:H7), ciprofloxacin (0%), doxycycline (0%), kanamycin (6%; O-:H21, O88:H8, O113:H36, O150:H8, O153:H40/44 and O157:H16), minocycline (8%; O-:H18, O-:H21, O2:H32, O88:H8, O113:H36, O150:H8, O153:H40/44 and O171:H2), nalidixic acid (2%; O-:H18 and OR:H18), norfloxacin (3%; O-:H11, O2:H25 and O157:H7), streptomycin (6%; O-:H18, O-:H21, O1:H48, O2:H32, O113:H36 and O171:H2), sulfonamides (7%; O-:H11, O-:H18, O-:H48, O88:H8, O153:H40/44 and O171:H2), tetracycline (9%; O-:H11, O-:H18, O-:H21, O2:H32, O113:H36, O150:H8, O171:H2 and OX18:H38) and trimethoprim (3%; O2:H32, O113:H36 and OX18:H38). Twelve isolates displayed multiple resistance with O-:H21 and O113:H36 being resistant to four antibiotics, while O2:H32 displayed resistance to five antibiotics (C, Mh, S, T and W).


Over 400 different STEC serotypes have been isolated from cattle (Blanco et al. 2004), a relatively small number of which are implicated in human disease (Nataro and Kaper 1998). Thirty-three different serotypes were obtained in this study of which 12 (O-:H46, O-:H48, O2:H+, O2:H32, O3:H12, O69:H-, O76:H34, O88:H8, O113:H36, O118:H16, O153:H40/44 and OX18: H+) have not been previously reported in cattle (Blanco et al. 2004; Hussein 2007). Furthermore, 22 of the strains were nontypable, highlighting the limitations of currently available antisera for typing emerging STEC.

There was no seasonal peak observed in this study, contrary to many similar studies where a late summer–early autumn surge is commonly reported for both O157 and non-O157 STEC (Pradel et al. 2000; Jenkins et al. 2002; Pearce et al. 2006).

The majority (94%) of our isolates were stx2 positive, with approximately half (48%) of this group carrying the eaeA gene. Of these, 32 isolates (86%) were also hlyA positive. Bovine STEC isolates, particularly those from older cattle, frequently carry the stx2 gene (Cobbold and Desmarchelier 2001; Beutin et al. 2007; Bosilevac and Koohmaraie 2011; Monaghan et al. 2011), while STEC strains from patients with more serious disease manifestations such as HC or HUS are frequently stx2 and eaeA positive and may also carry the hlyA gene (Friedrich et al. 2002). The high prevalence of the stx2+ eaeA+ hlyA+ combination is therefore of concern, and this study suggests in addition to O157:H7 and O26:H11, O-:H-, O157:H16 and OX18:H+ serotypes contain the necessary virulence genes to cause human disease and could be involved in future outbreaks. Suspect clinical cases, from which O157 is not detected, should be screened for these serotypes.

Most serotypes were eaeA negative, suggesting intimin has a minor if any role in the preferential colonization of the bovine gut (Ramachandran et al. 2003). From a human clinical perspective, these serotypes are also unlikely to be associated with HUS or large outbreaks (Cobbold and Desmarchelier 2001; Karmali et al. 2003). However, non-eaeA strains have been responsible for sporadic cases (Paton et al. 1999; Feng et al. 2001) as well as the O104:H4 German outbreak in 2011 (Anon 2011).

Alternative adherence factors, STEC autoagglutinating adhesion (Saa) and the long polar fimbriae (Lpf), were found in 17 and 49% of our isolates, respectively. Although saa has been previously reported in STEC O-:H18, O88:H8 and O113:H36 (Aidar-Ugrinovich et al. 2007), to the best of our knowledge, this is the first time saa has been reported in O-:H11, O153:H40/44, O171:H2 and OX18:H+. As previously reported (Paton et al. 2001; Jenkins et al. 2003; Kumar et al. 2004; Toma et al. 2004; Zweifel et al. 2005; Aidar-Ugrinovich et al. 2007; Monaghan et al. 2011), all of the saa-positive strains were eaeA negative; although given their locations on pO113 and the LEE region of the chromosome, respectively (Bolton 2011), there is no obvious explanation for this observation.

The hlyA, etpD and katP genes were present in 74, 27 and 6%, respectively, of our isolates despite all three genes being carried on pO157. Contrary to a previous report, which suggested that these three virulence factors were almost always present in O157:H7 strains (Fratamico et al. 2011), only 12% (3/26) of our strains had the hlyA+ etpD+ katP+ combination.

Different virulence gene profiles were detected in the same serotypes. For example, one O-:H- isolate carried the stx2, eaeA, hlyA, tir, espA, espB, iha, lpfAO157/OI-141, lpfAO157/OI-154, toxB and etpD genes, while another was only stx2 positive. This may be explained by the presence of the LEE and pO157 in the former but absent in the latter (Bolton 2011). LEE-positive and LEE-negative strains within the same serotypes may reflect different evolutionary lineages and provides supporting evidence that serotype is not always a good indicator of virulence. The presence of most of the virulence genes tested for in the O157:H7 strains also reflects the presence of the LEE and pO157.

Although the therapeutic intervention in the prevention of HUS is still unclear (von Baum and Marre 2005), the emergence and dissemination of antimicrobial resistance among STEC strains are still a cause for serious public health concern (Mora et al. 2005). Furthermore, antibiotic resistance data are essential for the further development of risk assessment and control strategies (Threlfall et al. 2000; Klein and Bülte 2003). In this study, 17% of O157 and 32% of non-O157 STEC were resistant to at least one antibiotic. These results are well within the range previously reported for STEC and are consistent with the higher prevalence commonly reported for non-O157 (Farina et al. 1996; Schroeder et al. 2002; Bettelheim et al. 2003). Furthermore, we are reporting resistance to chloramphenicol, kanamycin, minocycline, nalidixic acid, streptomycin, sulfonamides, tetracycline and trimethoprim. In a previous Irish study, Scott et al. (2009) also reported bovine STEC resistance to all of these antibiotics with the exception of minocycline, which was not tested for.

Nine strains displayed multiple drug resistance (MDR), supporting the hypothesis that MDR may be emerging across a range of STEC serotypes (Schroeder et al. 2002; Fukuyama et al. 2003, 2005; Walsh et al. 2006). Multiple antibiotic resistance may be acquired on mobile genetic elements such as plasmids, transposons and class I integrons (Mora et al. 2005; Singh et al. 2005) or because of overexpression of chromosomally encoded efflux pumps (Poole 2004). One strain, O2:H32, was resistant to five antibiotics, but only carried one virulence factor (stx2) and was therefore unlikely to be associated with serious human disease. Previous studies have reported multidrug resistance to up to seven antimicrobial agents in STEC O157 (Giammanco et al. 2002; Golding and Matthews 2004; Walsh et al. 2006).

This study found that there was no relationship between serotype and antibiotic resistance pattern, a finding that contradicts previous suggestions that antibiotics may aid the detection of specific serotypes in the future (Scott et al. 2009). Interestingly, our strains were resistant to the antibiotics most commonly used in veterinary medicine or as growth promoters prior to the European ban in 2006. This observation has been reported in other similar studies (Meng et al. 1998; Stephan and Schumacher 2001; Schroeder et al. 2002), suggesting a direct link between animal usage and the emergence of resistant phenotypes. Thus, antibiotic usage in animal production should be minimized.

It was concluded that STEC are ubiquitous on beef farms with a wide variety of serotype–virulence profile combinations and antibiotic resistance phenotypes. The detection of the stx2+ eaeA+ hlyA+ combination in rare serotypes, O-:H-, O157:H16 and OX18:H+, is a cause for concern as is the identification of a non-O157 STEC (O2:H32) with MDR. With such abundance of virulence genes in an environment that facilitates gene transfer within Enterobacteriaceae, it seems likely that new STEC serotypes with novel virulence gene combinations will emerge, some of which will inevitably pose a serious threat to humans.


The research was supported by the EU Framework VI project, ProSafeBeef (Food-CT-2006-36241).