Bookmark and Share

From: Food Quality & Safety magazine, April/May 2006

Elusive E. Coli

Ferreting Out the Deadly O157:H7 Strain Means having the Tools to Find and Expose What Makes it Tick

by Frank Burns, PhD

Enterohemorrhagic E. coli O157:H7 gained great attention as a foodborne pathogen in 1982 after an outbreak from contaminated hamburgers in a fast-food chain. Although anyone can get ill from ingesting even low levels of this organism, the very young and the elderly are at greater risk for developing hemolytic uremic syndrome (HUS) as a result of an E. coli O157:H7 infection. HUS is a severe illness that can lead to permanent loss of kidney function and even death. According to estimates by the Centers for Disease Control (CDC), E. coli O157:H7 is the cause of 73,000 cases of shigatoxigenic E. coli (STEC) infections and 61 deaths each year in the United States.

In 2002, the USDA’s Food Safety and Inspection Service (FSIS) mandated that beef plants reexamine HACCP plans and implement necessary measures to eliminate or significantly reduce the risk of E. coli O157:H7 in their products. FSIS also enhanced training programs for its inspectors and compliance officers, toughened its enforcement policies and held a series of workshops for small and very small plant operators.

It appears that these interventions have been effective. Data from the FSIS shows that the E. coli O157:H7 positive samples collected between 2000 and 2004 declined by more than 80 percent.1 Recalls have also shown a steady decline from 21 in 2002 to only six in 2004.1 And, most importantly, the CDC reported a 40 percent reduction in human illnesses associated with E. coli O157:H7 in 2005.2

Despite these encouraging results, we should not be lulled into a false sense of security. Improved prevention methods are a crucial part of the food safety solution, but they go hand-in-hand with reliable testing for the potentially deadly pathogen. Recent research indicates that E. coli O157:H7 is more complex and variable than previously understood, and its elusive nature challenges detection by the testing approaches in routine use today.

What Detection Methods Are Available?

Differences between E. coli O157:H7 and other types of E. coli can be detected directly by looking at the DNA or indirectly by looking at the products of the genes.

Polymerase chain reaction (PCR) is an example of a direct genetics-based approach that tests for actual fragments of target DNA. Lateral flow devices and other enzyme-linked methods rely on antigen-antibody responses that produce a chromogenic or fluorogenic signal. Culture methods look for color distinctions on plating media arising from E. coli O157:H7’s inactive b-glucaronidase, its inability to ferment sorbitol and/or its resistance to telluride.

The difficulties encountered in testing for E. coli O157:H7 are due to its highly dynamic nature, where the DNA readily undergoes change. This variability affects both the genetic structure and the genetic products.

Antigen-Antibody Reaction Problems

Sometimes this variability can affect the antigens that were used to name the pathogen: the O157 antigen on the cell surface and the H7 antigen on the flagella. Certain DNA changes within the organism can “turn off” expression of the H7 antigen and the production of flagella. The resulting variants are called O157:HNM (H-Non-Motile), and they have been explicitly included in the USDA food testing regulations.

A more recent discovery is that the same thing can happen with the O157 antigen on the cell surface. As first reported in 1998,3 the organism can undergo a genetic change that turns off expression of the O157 antigen. The resulting variants in this case are called “rough.” Rough isolates can be invisible not only to antigen-antibody tests, but they may also be missed by enrichment or confirmation methods that rely on antigen-mediated schemes such as antibody-coated bead capture for cell concentration prior to analysis.

These rough variants are not explicitly regulated by the USDA, but there is no reason to presume that they differ significantly in pathogenic potential from those that express O157. In fact, rough isolates have been isolated from clinical cases.4 In 2003, the USDA animal research center examined 1,697 E. coli O157:H7 isolates from cattle.5 They determined that approximately 10 percent were non-motile (did not express the H7 antigen) and approximately 1 percent were rough (did not express the O157 antigen).

Biochemical Reaction Problems

Biochemical or culture-based methods are also stymied by this organism. Because E. coli O157:H7 is believed to be telluride-resistant, many plating methods include telluride as a selective agent. (The T in CT-SMAC stands for telluride.) Although field-derived data is lacking, perhaps because telluride sensitivity precludes isolation with current plating procedures, data from laboratory strains indicate that telluride-sensitive isolates do exist.6

Some chromogenic agars used to isolate E. coli O157:H7 rely on the inability of the organism to ferment sorbitol (Sorbitol-MacConkey or SMAC agar). Others depend on the organism’s inability to metabolize 4-methylumbelliferryl-b-D- glucuronide (MUG) due to inactive b-glucaronidase. However, sorbitol-fermenting O157:H7 and O157:HNM isolates have been isolated from patients with severe disease7,8 and have emerged as a major health problem in Europe.9 Further complicating the situation is the fact that most of the sorbitol-fermenting O157:HNM isolates do produce a functional b-glucaronidase and are also sensitive to telluride.

Thus, the three common culture methods for identifying E. coli 0157:H7 that rely on biochemical reactions are at risk to miss variants.

Genetic Detection Problems

The phenotypic variations that elude indirect techniques result from the E. coli O157:H7 genome being in a state of flux. In order to avoid similar misses, direct techniques that examine DNA structure must identify consistent markers that will not be lost as the genome shifts.

Several test developers have targeted virulence factors, following the logic that if the virulence factor isn’t present, the sample would not be pathogenic. Because the virulence factors occur not only in E. coli O157:H7, but also in other bacteria commonly associated with cattle, these methods must use an antigen-antibody screening prior to PCR to avoid widespread cross-reaction.

Two common markers for virulence are the shiga-like toxin genes (stx1 and stx2). E. coli O157:H7 may have one, both or neither of these genes. Furthermore, the stx genes are on prophages that pop in and out of the genome with such frequency that the stx genes become unreliable as markers.10 This explains why isolates from some patients with the most severe illness do not have either stx gene11 while others contain both stx-negative and stx-positive genes.

Another commonly targeted virulence marker is the attachment and effacement gene (eae). This gene encodes intimin, a protein used by the bacteria to attach to the intestinal wall. However, at least one documented outbreak of HUS has been caused by an eae-negative strain of E. coli O157:H7,12 indicating that lack of the eae strain does not necessarily mean lack of virulence.

Genetic Detection – An Improved Approach

Since the virulence factors may not be sufficiently consistent on the E. coli O157:H7 genome to give reliable results, a different approach is needed.

With new understanding of the variants described above, researchers at DuPont developed a multiplex PCR assay for detecting E. coli O157:H7 that provides more genetic information for additional cross-checks and greater reliability.

First, the developers looked for some rare genetic sequences in the more stable regions of the genome, rather than virulence factors that are carried on mobile genetic elements. Next, they examined those areas against a large panel of E. coli O157:H7, including rough and HNM varieties, along with closely related organisms. Finally, they added several sequences that were common to all the E. coli O157:H7 in the panel to the revised assay. This approach aims at providing a very high confidence level that variant strains will not be missed, and it is cross-reactive with only one non-O157:H7 serotype (O55:H7, the nearest genetic neighbor). This confidence has resulted in the USDA FSIS adopting the DuPont developed E. coli O157:H7 MP assay as its regulatory screening test for raw ground beef and beef trim.

Will a new variant of E. coli O157:H7 that defeats even this multiplex assay be identified? Although no one can predict the future, past experience has shown that the dynamic nature of E. coli O157:H7 tends to find an elusive spot in every detection method used for it so far. The challenge for scientists is to stay one step ahead of this potentially deadly pathogen. The challenge for food companies is to stay informed, focus on prevention and rely on proven, science-based detection methods to verify their food safety systems. n


  1. USDA FSIS press release Feb. 28, 2005.
  2. USDA FSIS press release Feb. 23, 2006.
  3. Feng et al. (1998). J. Clin. Micro. (36): 2339-2341i.
  4. Vila et al. (1997). J. Clin. Micro. (35): 2279-2282.
  5. Barcoky-Gallagher et al. (2003). J. Food Protection (67):993-998.
  6. Taylor et al. (2002). J. Bacteriology (184): 4690-4698.
  7. Gunzer, F., Bohm, H., Russman, H., Bitzan, M., Aleksic, S., Karch, H. (1992). J. Clin. Micro. (30):1807-10.
  8. Fratamico, P.M., Buchanan, R.L., Cooke, P.H. (1993). Appl. Environ. Microbiol. (59): 4245-52.
  9. Karch and Bielaszewska (2001). J. Clin. Micro. (39):2043-2049.
  10. Shiakh and Tarr (2003). J. Bacteriology (185):3596-3605.
  11. Schmidt et al. (1999). J. Clin. Micro. (37):3491-3496.
  12. Paton et al. (1999). J. Clin. Micro. (37):3357-3361.

Frank Burns, Ph.D., is a senior scientist for DuPont Qualicon (Wilmington, Del.). Reach him at 302-695-5300.



Current Issue

Current Issue

February/March 2015

Site Search

Site Navigation