PCR Screening for Shigatoxin- and Intimin-Encoding Genes of Escherichia coli and for Salmonella in Beef Products in Processing Establishments


Samples from 971,389 lots (trim, ground beef and variety meats) were analyzed during 2005 to 2008 by using an 8h enrichment followed by PCR detection of E. coli virulence-related genes. Of the lots sampled, 15% were positive for the stx1 and/or stx2, 9.1% for the eae, 3.0% for rfb, and 1.7% for Salmonella by the PCR assay. Generally, lots of ground beef showed the lowest frequency of contamination and variety meats the highest. Overall, 15%, 4.6%, 4.6% and 0.81% samples were screen-positive for Shigatoxin-producing E. coli, enteropathogenic E. coli, enterohemorrhagic E. coli and E. coli O157, receptively. Of the E. coli O157 positive samples, 14% were also Salmonella positive. The frequency of screen-positive samples increases during the summer months. The higher incidence of non-O157 pathogenic E. coli and Salmonella in these beef products suggest that they could sere as indicator organisms signaling an increased risk for the presence E. coli O157.


Pathogenic Escherichia coli strains on raw or insufficiently cooked foods are of public health concern as serious disease may result from their ingestion. Therefore, most commercial producers of beef products screen for E. coli O157:H7 prior to shipment. While Salmonella is not considered an adulterant on raw beef products, it is used as an indication of process control. To detect these microorganisms, rapid screening methods are often used to provide results within 8-24 hours after sampling. Samples from several commercial beef production plants were tested using a rapid screening method based on the polymerase chain reaction (PCR) to determine if they were prescriptively positive for bacterial cells carrying E. coli genes related to virulence, if the frequency of positive samples varied over time and if the contamination rates differed between trim, variety meat and ground beef samples.


Samples were analyzed during 2005-2008 by using an 8h enrichment followed by PCR detection of the stx1, stx2 (shiga toxins), eae (attaching and effacing encoding intimin), and rfb (encoding the O157 specific O side chain polysccharide) genes plus a lateral flow antigen detection kit for E. coli O157, plus two Salmonella-specific genes. The significance of differences between frequencies of positive results was determined using Fisher’s Exact Test. These are screening results so signals may arise from different strains in a sample.


Table 1. Number and distribution of samples analyzed




Table 2: Frequency of samples yielding positive PCR



derived results




Table 3: Association of stx+ and sal signals





Table 4. Comparison of the rfb and LF screening tests




Table 5. Percent STEC+ trim samples by percent



Figure 3. Distribution of STEC+ samples during the
calendar year in product classes



Figure 4. Frequency of virulence genes during the
calendar year


Figure 5. Frequency of positive samples for pathogen
groups in product classes by year.





Samples (971,389) from ground beef, trim and variety meats were collected and screened by PCR over a four-year period for genes associated with E. coli O157 virulence and for Salmonella. Of the samples analyzed, 44.4% were trim from 008. Overall, trim accounted for 92% ground beef 1.7% and variety meats 6.5% of the samples. Forty seven percent of the samples were collected during 2008.

The most frequent PCR signal detected was for the stx2 gene (12%) followed by eae (9.1%) and stx1 (5.7%). PCR signals were combined to generate derived classes (Table 2 footnote). Fifteen percent of the samples were screen-positive for STEC strains and 4.6% of the samples could harbor EHEC and up to 0.81% E. coli O157.

The frequency of STEC samples in ground beef was higher than that of trim and variety meats during most of  2005 but then was similar for the duration of the study (Figure 1). The spike in positive frequency was higher for all product classes in the summer of 2008.

As the presence of Salmonella and E. coli are associated with fecal contamination, it is not surprising that stx, eae and sal signals would often be found together at a higher frequency that if they occurred independently (Figure 2). In fact, stx and eae appear together in trim 3.3 times more that expected if random and 3.0 times in variety meats (Table 3). For ground beef, this ratio increases to 8.4.

While the rfb and LF tests are both used to detect Ecoli O157, rfb appears more sensitive (Table 4). As all O157 cells are presumed to contain rfb, samples testing rfb- LF+ could be false positives resulting from a cross reaction of LF with non-O157 strains. The rfb+ LF- could result from PCR-based tests being more sensitive than the immunological one. About 25% of all rfb+ samples preemptively harbor EHECs regardless of the LF result. Samples with a higher fat content tend to have a higher percentage of STEC-positive samples (Table 5). As fat is often on the outside of meat, it is more likely to become contaminated with fecal matter during processing.

All product classes have a higher contamination rate during the summer months than other seasons (Figure 3). As bacterial growth is temperature dependent, this is not surprising. Ground beef and variety meats are more likely to be contaminated than is trim. Positive tests for stx, eae, and rfb and/or LF reflect this seasonal increase (Figure 4) and can be considered indicators of pathogens.

When the frequency of positive samples for pathogen groups are examined by year, the results vary (Figure 5). Generally, samples from 2005 had a higher frequency of contamination than those in later years. In 2005, variety meats were most often contaminated while in 2008, positive samples of ground beef were most prevalent, except when considering products harboring Salmonella.