“Shiga toxin-producing Escherichia coli (STEC),” “Enterohemorrhagic E. coli (EHEC),” and “E. coli O157” are often used interchangeably when discussed as threats to food establishments. Are these three targets really the same? The answer is complicated because it can be both yes and no.

STEC strains are categorized by two characteristics: their ability to produce verotoxins or Shiga toxins (stx) and their ability to adhere to cell membranes. While there are several stx variants, two have been associated as the causative agents for human illness: stx1 and stx2. Likewise, STEC has multiple ways to adhere to cell membranes. The most common pathway is associated with a set of genes, eae, that produce intimin, an outer-membrane adhesin essential for attachment. Not all STEC strains contain eae genes, as evidenced by the 2011 E. coli O104 STEC sprouts outbreak in Europe that resulted in 31 deaths. This strain was eae-negative and used alternative pathways for membrane adherence.

EHEC is a subset of STEC strains that can cause hemorrhagic colitis (HC). HC and the strains that cause it are of primary concern to food establishments because the disease can progress to hemolytic uremic syndrome (HUS), a potentially fatal illness. When discussing EHEC and STEC, it is important to remember that all EHEC strains can be considered STEC strains, but not all STEC strains are EHEC strains—only those that cause HC.

E. coli O157:H7 is the leading serogroup associated with HUS. An E. coli O157 strain that produces stx, adheres to cell membranes, and causes HUS would be classified as both an STEC and an EHEC. A nontoxin-producing E. coli O157 would be considered neither an STEC nor an EHEC. Thus, the classification of these three terms remains open-ended and warrants that food establishments have a strong understanding of each.

Challenges for STEC Detection
E. coli O157 has long been considered an adulterant in foods, and detection methods have been used in the food industry for nearly 20 years. STEC strains are estimated to cause over 250,000 illnesses annually, with around one-third resulting from O157. But there are many more STEC strains than just O157. Six additional serogroups—O26, O45, O103, O111, O121, and O145—have been identified as the source of 70 percent of the non-O157-related illnesses and have been identified as adulterants by the U.S. Department of Agriculture Food Safety and Inspection Service (USDA-FSIS). In June 2012, USDA-FSIS implemented requirements for the expansion of STEC testing in raw beef trimmings, targeting these six serogroups, and made public a new reference standard for analysis.

The U.S. was not the only country to increase testing requirements. As a result of the deadly sprout outbreak, the European Union (EU), in 2013, issued new microbiological testing criteria for sprouts. Those recommendations included additional testing for STEC. In the years since, several EU countries have also increased the testing for STEC strains in beef and dairy products. All testing in the EU would be conducted using the ISO TS 13136 STEC testing standard. While both USDA-FSIS and the EU made methods available for STEC analysis, these methods were limited, with few rapid technologies available commercially.

Variants of STEC strains are closely related and can often prove difficult to isolate for detection. Common techniques employed by labs to solve these problems are often impractical. Supplementation of media with antibiotics can be used to select for O157, but the recommended antibiotic for aiding in the selection of O157, novobiocin, can completely or severely inhibit the growth of non-O157 STEC strains. Molecular-based detection methods, traditionally the methods of choice for O157, would need to be designed to detect genetic markers for multiple targets, yet still be user-friendly. The creation of a complete polymerase chain reaction (PCR) solution requires the skill of highly trained analysts. In addition to detecting the correct serogroup, for the strain to be considered an adulterant, it would need to be both virulent and able to adhere to cell membranes. Additional testing is required to confirm the presence of these virulence factors (stx) and adhesion genes (eae). After presumptive identification, laboratories must then be able to isolate the target strain. Isolation of the serogroups using antibody-coated beads is technique-sensitive, and cross-reactivity between closely related species can lead to inconsistent results. Media used to isolate the strains often lack effective selectivity or differentiation ability, making it hard to discern one serogroup from the next. These methods require skilled analysts to complete complex work flows and can often be labor-intensive and time-consuming. A strong need existed for improving scientific approaches and expanding into new methods of detection and differentiation.

Improvements in Detection
Since the implementation of regulations by USDA-FSIS and the EU, advances have occurred in all phases of STEC analysis from enrichment to identification. These changes have simplified work flows and increased the accuracy of the data being generated. The first opportunity for improvements occurred with the enrichment of the target analytes. Modifications to enrichment media formulations along with focusing on the use of elevated temperatures and shorter incubation times to selectively emphasize the growth of the target organisms while reducing the growth of background microflora resulted in better performance of rapid detection methods.

In addition to more effective enrichment media, improvements were made to the most common form of detection, molecular PCR kits. As the understanding of the genomic profiles of these organisms increased, molecular methods were refined with new and more specific targets, increasing the specificity and selectivity of the assays. Ready-made test kits, with premixed amplification reagents (in both liquid and pellet form), simplified a complex molecular process. The assays were designed to detect multiple target gene sequences, looking for both virulence and adhesion factors, and/or specific serotypes. However, not all the problems with the original assays were solved. Due to the number of targets, a single PCR assay to test for the virulence factors as well as each serotype has not been developed. Instead, system suites were created, with one assay screening for the target virulence and adhesion factors followed by detection of specific serogroups using a second and/or third assay as needed. Multiplex kits were developed and validated from multiple method developers, allowing laboratories to seamlessly integrate analysis into their current systems.

In parallel to the advances in the detection of these organisms, improvements were made in the isolation process. Antibodies used for the isolation of organisms continue to be modified to reduce cross-reactivity and reduce background microflora on agar plates. Combinations of chromogens, antibiotics, and other growth factors have been included in new agar formulations, improving their ability to not only select for the target strains but to also allow for easier discrimination between the top seven serogroups. Wide ranges of colony color morphologies make it easier to identify the serogroups compared with discernment of shades of a single color (purple, purple-blue, purple-red, etc.). Multiple approaches are now available for isolation, ranging from the use of a single agar that allows for differentiation between all seven serogroups to the use of multiple agars that target single serogroups. Both approaches provide advantages and drawbacks to the end-user. Cost savings for a single agar may be offset by the skill required to discern between colony morphologies. Reduced labor costs in identifying strains can be obtained with the multiple agar use, but material costs may increase. Ultimately, these improvements in isolation allow end-users more flexibility in the choice that best fits their needs.

Since the mandate for additional testing, improvements have been made. Ready-to-use assays are being employed in labs globally, and isolation of presumptive positive isolates is easier than ever. Even with these improvements, however, problems still exist. The use of multiple test kits does not guarantee that a presumptive positive sample contains an STEC strain. For example, an enriched sample may contain one of the top seven serogroups that is nonvirulent. That sample may also contain E. coli strains that have stx or the eae genes. Test kits are not able to discern between multiple isolates and require additional cultural work to confirm the presence of a true STEC. This detection and isolation process can be lengthy and expensive and has pushed food establishments to look at alternative technologies.

Next Steps in Detection
The end goal for many in the food industry is a method designed to detect the top seven serogroups that contain both virulent and adhesion factors in a single, easy-to-use, and affordable test kit. The last half-decade has brought this closer to a reality with the implementation of genetic sequencing technologies such as next-generation sequencing (NGS) and whole-genome sequencing (WGS) in more and more laboratories. Reduced footprints, simplified work flows, and improved databases have increased the adoption of NGS/WGS. These technologies offer the ability to detect serotype, virulence, and other genetic factors based on the profiles produced. Advances in WGS have resulted in its being implemented for the use of outbreak detection and investigation due to its ability to provide an advanced level of strain discrimination in days versus weeks.

Method developers continue to explore and expand the possibilities of sequencing by combining it with other types of technology. The use of PCR-based detection, followed by WGS, combines the advantages of both technologies, allowing for rapid detection and then further discrimination of isolates. However, NGS and WGS are not without their own challenges. The food industry has been slow to adopt WGS. Cost per sample analysis can make it prohibitive for adoption, especially when screening results in a lack of detection. Preparation and analysis of samples often requires more skill than the ready-made molecular assays do. Issues can also exist in the interpretation of the bioinformatics data produced during analysis, requiring skilled technicians, not microbiologists, to interpret the data.

Where to Now?
With new mandates to conduct testing by USDA-FSIS and the EU, technology to detect and differentiate STEC has progressed rapidly. Advantages and drawbacks exist both with commonly used methodologies and with those gaining traction. While the continued development of new and novel technologies may make the technologies of today seem archaic, they bring their own drawbacks. Laboratories are not one size fits all, and a need for multiple options is vital. A broad approach to improving our current methods as well as investigating new, novel techniques is important to the food industry.  

Patrick Bird, M.Sc., is the owner of PMB BioTek Consulting and a technical consultant with AOAC.