Abbreviations

aEPEC	Atypical Enteropathogenic Escherichia coli
CT	Cycle threshold
eae	intimin encoding gene
EHEC	Enterohaemorrhagic Escherichia coli
HUS	Hemolytic Uremic Syndrome
RT-PCR	Real Time-Polymerase Chain Reaction
STEC	Shiga toxin producing Escherichia coli
stx	shiga toxin
TSA	Tryptone Soya Agar

---

1. Introduction

In recent years there has been a progressive increase in the number of reported cases of non-O157 Shiga Toxin producing Escherichia coli (STEC) infections ^[1]. It is estimated that there are 168,000 cases of non-O157 STEC infections within the United States that compares with 96,000 for the more studied O157:H7 serotype ^[2]. The incidence of non-O157 STEC is thought to be significantly underestimated due to lack of reporting, misdiagnosis and lack of detection using current clinical diagnostics ^[2,3,4].

The majority of non-O157 STEC infections are associated with lower virulence than O157:H7 yet some can still lead to hemolytic uremic syndrome (HUS) if the strain harbors the full complement of virulence factors. Specifically, genes encoding for initimin attachment factor (eae), shiga toxin (stx) and occasionally, hemolysin (hylA), are frequently correlated to virulence ^[5]. The eae gene forms part of the chromosome-located LEE island whereas hylA is plasmid-encoded ^[6]. The stx gene can be subdivided into stx₁ and stx₂ that also exists in several variants ^[7]. It is commonly accepted that stx₂ is the more potent form of the toxin compared to stx₁ that is typically associated with lower virulent strains ^[7]. The stx gene is encoded on a prophage that integrates into the chromosome and when expressed produces new phages along with releasing the shiga toxin before cell lysis ^[8]. Consequently, the expression of shiga toxin in most cases ultimately leads to the death of the E. coli producing cell ^[8].

A major challenge in surveillance of non-O157 STEC is the wide distribution and diversity of the E. coli subgroup. The majority (> 70%) of non-O157 STEC encountered have low virulence due to restricted stx expression and lack of a full complement of virulence factors such as eae ^[9]. However, those STEC that express shiga toxin along with attachment factors can result in HUS development within a susceptible host and hence are classed as EHEC. Given that EHEC only makes up a small proportion of STEC encountered in animals, foods and the environment, there was a focus on those that represent the main risk to public health. To this end, it was noted that 70% of clinically confirmed non-O157 STEC could be attributed to 6 serotypes; O26, O45, O103, O111, O121 and O145 that became known as the Top or Big 6 ^[6].

It was assumed from the outset that the Top 6 STEC followed the same dissemination routes as O157:H7 in that cattle were the main source of the pathogens, along with beef being the main food vehicle for transmission. Consequently, legislation was passed in 2011 to designate the Top 6 non-O157 STEC as adulterants in beef trim then later in ground beef ^[10]. This in turn led to the development of diagnostics (primarily methods based on RT-PCR) to screen for the Top 6 non-O157 STEC in cattle, the environment and beef. The methods follow a similar format of enriching samples in TSB with novobiocin followed by screening with multiplex RT-PCR to detect stx and eae followed by a second panel that screens for the wzx gene that encodes LPS associated with the Top 6 serotypes. There have been several diagnostic platforms developed for screening for the Top 6 non-O157 STEC with the GeneDisk^® from the PALL Corporation and BAX from Dupont being two examples ^[5,11]. A common feature of surveillance studies, either performed on beef or cattle, is that a high prevalence (0.4-74%) and diversity of Top 6 non-O157 STEC are encountered ^{[12,13,14,15]}. However, attempting to recover isolates from presumptive positive samples remains a challenge. Specifically, the extent of recovery of culture positive can be < 1% of those samples testing positive by RT-PCR ^{[16,17,18,19,20,21]}. The underlying reasons for the low culture positive samples has been proposed to be due to the target being at low levels, virulence and LPS genes being present in different cells ^[22]. It has also been proposed that the lack of distinguishing phenotypes between non-O157 STEC and non-pathogenic E. coli can result in high background interference thereby making isolation of the former problematic during culturing ^[11,23].

The following reports on the incidence of the Top 6 non-O157 STEC associated with cattle and processing environment within Ontario beef processing operations. An RT-PCR platform was applied and problems with isolating culture positive samples addressed through extensive screening of colonies derived from presumptive positive samples. Specifically, whereas other studies select 2-5 colonies per plate the current study screened up to 20 colonies to increase the probability of isolating a Top 6 non-O157 STEC target.

---

2. Materials and Method

---

2.1. Experimental design

Samples were obtained from two different beef processing facilities in South-Western Ontario over 4 visits within a 5 month period. Hide, carcass, fecal and environmental samples were collected from individual animals for detection of E. coli non-O157:H7. Specifically, ten incoming cattle were selected at random and a saline pre-moistened sterile sponge (Nasco, Fort Atkinson, WI) was used to take a composite sample of the neck, flank and round. The same carcass was sampled following the de-hiding process. In addition, samples (10 cm² areas) were also collected using sponges from the holding pens and also the stun floor onto which the cattle rolled onto following stunning. All the samples were transferred to a cold box and processed within 24 h.

---

2.2. Top 6 STECscreening

The sponge samples were enriched in 100 mL TSB containing 20 mg/L novobiocin overnight at 42°C. The enriched samples were screened using RT-PCR GeneDisk system (GeneDisc^® Cycler, Pall Corporation, Port Washington, NY, USA) in accordance to the manufacturer's instructions. Here, 50μL of enriched sample was heated at 100°C for 10 min to extract the DNA. Aliquots (18μL) of the DNA template were added along with an equal volume of quantification master mix (QMM) containing primers targeting stx₁, stx₂, and eae genes. The PCR routine was performed over 45 cycles and reaction monitored through the increase in fluorescent signal. Samples testing positive for eae and stx were taken forward for a second round of screening for presence of genes for the Top 6 O-antigens, as per the manufacturer’s instructions.

The performance of the RT-PCR GeneDisc^® system was verified by inoculating homogenates of environmental, hide and carcass samples with a cocktail of Top 6 non-O157 STEC serotypes provided by the Public Health Agency of Canada (Guelph, ON, Canada). The individual isolates were cultivated 50 mL of TSB overnight at 37°C. The culture was used to inoculate (final cell density of ca. 3 log cfu/mL) TSB containing 20 mg/L novobiocin from a homogenate derived from hide, carcass or environment. The culture was enriched overnight at 42°C before removing an aliquot for screening using the GeneDisc® system.

For enriched samples testing positive for virulence factors and Top 6 O-antigens were put through a confirmatory second screening using the BAX RT-PCR STEC assay. Here, aliquots (20μL) of enriched samples were transferred to reaction tubes containing 200μL of lysis reagent that was then heated for 20 minutes at 37°C. The temperature was then increased to 95°C and held for 10 min and 30 µL of the DNA template mixed with reaction mix within PCR tubes containing primers targeted towards stx and eae genes. Samples testing positive were taken forward for a secondary screening for the O antigen gene (wzx) of the Top 6 STEC using two separate panels for O26, O111, and O121 (first panel) and O45, O103, O145 (second panel).

---

2.3. Culture confirmation testing

Enriched samples testing positive for eae, stx and Top 6 wzx were taken forward for culture confirmation testing. Here, the enriched samples were plated onto CHROMAgar^® (BioMed Diagnostics, Oregon, USA) and incubated for 24 h at 37°C. Ten presumptive STEC isolates (mauve colonies) were picked per plate and transferred to a new CHROMAgar^® plate (subsequently incubated at 37°C for 24 h) and the remainder of the inoculm to 500μL molecular grade water (Invitrogen Life Technologies, Mississauga, Ontario, Canada). All 10 selects colonies were pooled into the tube containing 500 µL water to make the composite sample and placed into a boiling water bath to extract the DNA. PCR protocol, primers, and thermal cycling conditions (MyCycler^® Thermal cycler, BIO-RAD, California, US) for stx, eae and hylA screening were followed as previously described by Paton & Paton ^[17]. DreamTaq Green PCR master mix (ThermoFisher, NY, USA) was used in the PCR reactions with nuclease-free water and 2 µM primer concentrations combined with 1 µL DNA template (Table 1). The PCR amplification products were separated on a 1.5% w/v agarose gel containing ethidium bromide. The gel bands were visualized under UV using a Bio-Rad Gel Doc EZ System (Bio-Rad, Hercules, CA).

In the event that composite samples tested positive for eae and stx the ten individual colonies that comprised the presumptive positive composite samples were individually screened by the same PCR technique to identify those that contributed to the positive result. Isolates harboring eae and stx were confirmed as E. coli by using the VITEK^® MS IND system (BioMerieux, Laval, QC, Canada).

---

2.4. Serotyping of culture positive Top 6 STEC isolates

Culture positive E. coli isolates harboring eae and stx were sub-cultured onto TSA and incubated for 24 h at 37°C. The Prolex^® E. coli non-O157 Identification latex agglutination test (Pro-Lab Diagnostics, Richmond Hill, Ontario, Canada) was used to screen for the Top 6 non-O157 STEC serotypes. Selected isolates were also submitted to the Public Health Agency of Canada (Guelph, Ontario) for confirmatory serotyping.

---

2.5. Enterobacterial Repetitive Intergenic Consensus (ERIC) sequence PCR

The isolated to be typed was cultivated on Luria-Bertani (LB) agar at 37°C for 24 h and a single colony suspended in 500 µL water. The sample was placed in a boiling water bath for 10 min followed by centrifugation at 13,000 g for 2 min to remove cell debris. Aliquots (1 µL) were mixed with 24 µL of DreamTaq Green PCR master mix (ThermoFisher, NY, USA) and 2 µM primers (Table 1). The PCR reaction was performed in a MyCycler^® Thermal cycler (Bio-Rad, California, US) with an initial cycle for 3 min at 94°C, then 35 cycles consisting of 30 s at 94°C, 1 min at 52°C, and 4 min at 65°C. The last cycle was for 8 min at 65°C ^[24]. The PCR amplification products were separated on a 1.5% w/v agarose gel and band patterns visualized using a Bio-Rad Gel Doc EZ System (Bio-Rad, Hercules, CA).

---

3. Results

---

3.1. Sampling and RT-PCR screening

Hide (n=186), de-hided carcasses (n=40) and environmental (n=62 holding pen n=40 stun floor) samples were collected from two participating beef processors over a 5 month period. The samples were enriched in TSB containing novobiocin then screened for the Top 6 STEC using RT-PCR GeneDisk^® system. It was found that positive samples (i.e. harboring stx, eae and wzx for Top 6 STEC serotypes) were returned for 92.5% (n=172) of the hide samples, 72.5% (n=29) de-hided samples and 84.3% (n=86) of the environmental samples screened. A secondary RT-PCR assay using BAX System confirmed the presence of stx and eae in enriched samples along with the occurrence of the LPS genes for the Top 6 STEC. The two RT-PCR platforms were in agreement with respect to stx and eae, along with the occurrence of the Top 6 STEC LPS encoding genes. With respect to the latter, a common feature in all the samples was the high diversity of the Top 6 serotypes encountered within the individual samples. The most commonly encountered serotypes were O103, O45 and O121 although all the Top 6 STEC serotypes were encountered to varying degrees (Figure 1). A further observation was the high diversity of Top 6 non-O157 STEC within the individual samples that ranged from three to all six serotypes.

Although the C_t values for the O antigens were low (< 20 cycles), those associated with the virulence factors had values ranging from 10-39 cycles (Figure 2). Control samples derived from enriched cultures derived from those inoculated with top 6 STEC strains returned positive results indicating no inhibitors were present in the different matrices (hide, carcass and manure).

The C_t values for eae and stx for each sample were within 5 cycles for 81% of samples, and within one cycle for 22% of cycles, suggesting that these targets were present in relatively equivalent amounts. Yet, the high C_t values would suggest the virulence genes were at relatively low concentration within the enriched samples (Figure 2).

---

3.2. Culture confirmation of presumptive positive samples

A sub-set of 120 presumptive positive enrichment samples were individually plated onto CHROMAgar^® that was subsequently incubated at 37°C for 24h. Mauve colonies (n=10) from individual plates were sub-cultured onto a new CHROMAgar^® plate and also combined to form a composite sample. The composite was taken forward to screen for the presence of stx₁, stx₂, eae and hlyA by conventional multiplex PCR. With composite samples testing positive for eae and stx the individual colonies making up the composite were screened to identify those contributing to the PCR positive result. After the first round of colony screening, 4 of 120 samples screened yielded composite sample pools containing virulence factors. When the individual colonies making up the sample composites were screened, at least one individual isolate harboring a full complement of virulent factors (eae & stx) was recovered from two of the samples, providing a rate of recovery of 1.7%. In the other two composite samples testing positive, the virulence factors were identified but in different isolates (Figure 3).

---

3.3. Effect of extended holding period at 4°C on apparent STEC prevalence

In an attempt to recover more isolates with a full complement of virulence factors, further colonies were recovered from the original plates that had been stored at 4°C for 14 days. Here, 10 further (different) colonies were taken to prepare the composite samples. Interestingly, on this occasion, the number of composite samples harboring the full compliment of virulence factors was 17 compared to only 2 in the first round. Isolates containing a full complement of virulence factors (i.e. eae and stx) were recovered from 16 of the composite samples (representing a recovery rate of 13%) with one sample having the virulence factors distributed between different cells.

In total, 39 individual isolates were recovered harboring full complement of virulence genes (Table 2). All the isolates were confirmed as E. coli with growth at 37°C but not 4°C. The most prominent strains harbored stx₁, stx₂, eae, hylA with stx₂ being encountered more so than the less potent stx₁ (Table 2). The majority of isolates were recovered from cattle hides but also the de-hided carcass but less from the environment of the slaughterhouses (Table 2).

As part of the culture maintenance the 39 isolates were sub-cultured onto TSA agar. It was found that the stx gene in all but 3 of the isolates was absent. Yet, when sub-cultured onto CHROMAgar^® the stx gene was retained. This was illustrated in isolate L1-17 that when cultured on CHROMAgar^® harbored eae, stx₂ and hylA but were devoid of the latter two virulence factors when grown on TSA agar (Figure 4). By using ERIC-PCR it was found that both isolates recovered from CHROMAgar^® and TSA had the same fingerprint indicating they belonged to the same strain (Figure 5).

The three isolates that exhibited stable virulence factor composition were identified as serotypes O172:H28, O76:H7 and O187:H52, all of which harbored stx₁, stx₂, eae and hylA. The other isolates screened did not belong to the Top 6 STEC as determined by latex agglutination assay.

---

4. Discussion

The original objective of the study was to determine the incidence of the Top 6 STEC within two Ontario beef processing facilities. In agreement with other studies, it was found that samples associated with cattle had a high prevalence and diversity of the Top 6 serotypes when initially screened using RT-PCR. Similar to other studies, serotype O103 was most frequently recovered, suggesting host related associations with cattle ^[14,19,25]. Yet, despite the apparent high prevalence of the Top 6 non-O157 STEC recorded using RT-PCR no culture positive isolates of the serotypes were recovered. The high prevalence of the Top 6 O serotypes could not be assigned to false positives given the results were confirmed using a different RT-PCR platform. Instead, it is more likely that genes encoding for the O antigens were present but within E. coli cells not harboring virulence factors. This represents a key limitation of screening for the Top 6 E. coli serotypes given that there is a poor correlation between O antigen and virulence factors unlike the more well characterized O157:H7 ^[26]. Yet, additional factors were also evident to explain the low recovery on plates from samples returning high prevalence according to RT-PCR. Specifically, the high C_T values for eae and stx would suggest that the target genes were in low abundance in the enriched cultures. Cooley et al. ^[22] suggested that only samples with C_T values > 27 cycles the probability of isolating culture positives would be < 1% ^[22]. Other researchers have encountered the same low recoveries of Top 6 STEC when C_T values were high ^[14,19]. In the current study, the C_T values were > 20 suggesting that the Top 6 non-O157 STEC could have been in the sample but the high background made isolation problematic.

An additional factor that contributed to the apparent high presumptive positive RT-PCR results was the presence of virulence factors in different cells. This was indeed found to be the case with virulence factors being present in composite samples taken from colonies from presumptive positive enrichment samples. The result is in agreement with other researchers that have also isolated virulence factors within different cells that constituted a presumptive PCR positive sample ^[27,28,29]. In the current study, the presence of virulence factors in different cells likely contributed to a minor extent given that the occurrence of stx and eae by RT-PCR was similar. If the virulence factors were in different cells then it would be expected that one of the factors would be disproportionate relative to the other. In addition, only two of the composites samples containing pooled isolates was demonstrated to have virulence factors in different cells.

A further factor that contributed to the apparent high prevalence of STEC by RT-PCR was the mobility of virulence factors (stx and to some extent hylA). Although E. coli did not grow at 4°C it was evident that stx genes (prophages) were exchanged between cells making up the colony. The acquisition of stx by aEPEC is well documented although the heterogeneous distribution of stx within the same colony has not been reported to any great extent ^[30]. The extent of stx exchange was found to be influenced by the growth medium with TSA being less likely to harbor stx and hylA compared to CHROMAgar^®. The heterogeneous distribution of stx within a population and influence of sub-culturing has been previously observed. Specifically, Bielaszewska et al. ^[31] demonstrated how Top 6 non-O157 STEC convert from EHEC to aEPEC by losing stx during cell division. The researchers observed loss of stx₂ in up to 14% of EHEC colonies tested following sub-culturing. Conversely, aEPEC were also found to acquire stx when incubated in the presence of shiga toxin phages ^[32]. The transient recovery of stx in E. coli was also observed from an isolate from a patient diagnosed with HUS ^[33]. The loss of stx has also been reported for STEC when sequentially sub-cultured ^[34]. However, there have been no reports on medium composition having an influence on the distribution of stx or hylA within E. coli populations. CHROMAgar^® selectivity is primarily based on cefixime and tellurite with differentiation by hydrolysis of chromogenic constituents. Therefore, it is possible that the mobility of stx prophage and/or hylA encoding plasmid was enhanced due to imposed stress by selective agents that were not present in TSA. In evolutionary terms, there is a strong incentive for E. coli to resist infection by stx prophage given expression ultimately means death of the producing cells.

Even though the study did not recover any Top 6 non-O157 samples there were three isolates that carried virulence factors that could potentially cause HUS. The serotypes recovered have not been previously implicated in non-O157 STEC infections, so their clinical significance is unclear. Yet, the research does underline the need for better diagnostics to identify STEC strains of importance to food safety above that of screening for serotypes.

---

5. Conclusion

The study has further highlighted the challenges in screening for the Top 6 non-O157 STEC in samples derived from cattle and beef processing environments. Although RT-PCR platforms provide a high throughput and convenient screening technique the results are only useful if an isolate is recovered in the end. The study illustrated that the disconnect between RT-PCR and culture based methods is due to a combination of targets being present at low levels, virulence factors being distributed between different cells and the transient acquisition of stx. In the broader sense, the study has further underlined the need for improved diagnostics to identify those STEC of significance to public health.

---

Acknowledgements

The authors wish to extend their appreciation to the Ontario Ministry of Agriculture, Food and Rural Affairs (Project # FS2011-1061) and the Beef Cattle Research Council of Canada (Project # FOS01.13) for funding the research along with the Saudi Arabian Ministry of Education. Thanks also goes to the participating slaughter houses for allowing access and PHAC for performing the serotyping of isolates.

---

Conflict of Interest

There is no conflict of interest relating to the research.

---