ABSTRACT

Shiga toxin–producing Escherichia coli (STEC) serogroups O26, O45, O103, O111, O121, and O145, called non-O157 STEC, are important foodborne pathogens. Cattle, a major reservoir, harbor the organisms in the hindgut and shed them in the feces. Although limited data exist on fecal shedding, concentrations of non-O157 STEC in feces have not been reported. The objectives of our study were (i) to develop and validate two multiplex quantitative PCR (mqPCR) assays, targeting O-antigen genes of O26, O103, and O111 (mqPCR-1) and O45, O121, and O145 (mqPCR-2); (ii) to utilize the two assays, together with a previously developed four-plex qPCR assay (mqPCR-3) targeting the O157 antigen and three virulence genes (stx1, stx2, and eae), to quantify seven serogroups and three virulence genes in cattle feces; and (iii) to compare the three mqPCR assays to a 10-plex conventional PCR (cPCR) targeting seven serogroups and three virulence genes and culture methods to detect seven E. coli serogroups in cattle feces. The two mqPCR assays (1 and 2) were shown to be specific to the target genes, and the detection limits were 4 and 2 log CFU/g of pure culture–spiked fecal samples, before and after enrichment, respectively. A total of 576 fecal samples collected from a feedlot were enriched in E. coli broth and were subjected to quantification (before enrichment) and detection (after enrichment). Of the 576 fecal samples subjected, before enrichment, to three mqPCR assays for quantification, 175 (30.4%) were quantifiable (≥4 log CFU/g) for at least one of the seven serogroups, with O157 being the most common serogroup. The three mqPCR assays detected higher proportions of postenriched fecal samples (P < 0.01) as positive for one or more serogroups compared with cPCR and culture methods. This is the first study to assess the applicability of qPCR assays to detect and quantify six non-O157 serogroups in cattle feces and to generate data on fecal concentration of the six serogroups.

Shiga toxin–producing Escherichia coli (STEC) is a major foodborne pathogen that causes human illnesses characterized by nonbloody and bloody diarrhea, with hemolytic uremic syndrome as a potential complication (17). Serotype O157:H7 is the most common STEC responsible for the majority of foodborne STEC illnesses; however, other STEC serogroups, particularly O26, O45, O103, O111, O121, and O145, have gained more recognition in recent years because they account for more than 70% of non-O157 STEC infections in the United States (33). In 2011, the U.S. Department of Agriculture, Food Safety and Inspection Service declared these six non-O157 STEC serogroups to be adulterants in ground beef and nonintact raw beef products (35). Cattle, a major reservoir of STEC, harbor the organisms in the hindgut and shed them in their feces; this serves as a major source of contamination of food and water (17).

Not much is known about fecal shedding of the six non-O157 E. coli pathogens in cattle because detection methods, PCR- and culture-based, have been developed and validated only recently (3, 5, 11, 21, 31). In addition to the presence of STEC pathogens in cattle feces, the concentration of these organisms plays a role in the spread between animals and subsequent hide and carcass contamination. A subset of cattle, known as “super shedders,” shed the STEC O157:H7 organism at high concentrations (≥104 CFU/g of feces) (7). Super-shedding cattle have been reported to be a major source of transmission of O157:H7 among cattle within the herd (28) and of subsequent contamination of hides and carcasses (2, 12, 19). Because there are no data on fecal concentration of non-O157 STEC in cattle, it is not known whether a subset of cattle that are super shedders of non-O157 STEC (as with O157) exists in a population. Fecal concentration data of the six non-O157 serogroups (O26, O45, O103, O111, O121, and O145) in cattle, when factored in microbial risk assessment models, allow estimation of the potential contamination burden that fecal shedding represents in upstream production stages. Although real-time PCR assays have been developed to detect non-O157 STEC in food matrices (14, 15, 24), the applicability of the real-time PCR assays to detect and quantify non-O157 E. coli in cattle feces has not been evaluated. Anklam et al. (1) have developed four separate multiplex qPCR assays to target the seven serogroups (O26, O45, O103, O111, O121, O145, and O157) and four virulence genes (stx1, stx2, eae, and ehxA) in cattle feces. The assays were validated using pure cultures and culture-spiked cattle feces, but applicability of the assays for the detection and quantification of E. coli serogroups and associated virulence genes in feces of naturally shedding cattle was not evaluated. Luedtke et al. (27) developed a multiplex qPCR to target enterohemorrhagic E. coli (EHEC)–specific attaching and effacing gene-positive conserved fragment 1, ecf1, and eae for enumeration of EHEC directly from cattle feces; however, the assay does not quantify individual serogroups. The objectives of our study were (i) to develop and validate two sets of multiplex quantitative PCR (mqPCR) assays to target O-antigen genes of O26, O103, and O111 (mqPCR-1) and O45, O121, and O145 (mqPCR-2) and (ii) to evaluate the applicability of the two assays, together with a previously developed four-plex qPCR assay (mqPCR-3) targeting the O157 antigen gene and three virulence genes (stx1, stx2, and eae (30)), to quantify six non-O157 and O157 serogroups and three virulence genes in cattle feces (n = 576) from a commercial feed yard. Additionally, the detection of the seven STEC serogroups and three virulence genes in cattle feces by the three mqPCR assays was compared with detection by a 10-plex conventional PCR (cPCR) that targets the same genes (seven serogroup-specific and three virulence genes) and by culture methods.

MATERIALS AND METHODS

Design of the assays.

Two sets of mqPCR assays (mqPCR-1 and mqPCR-2) were developed to detect and quantify O26, O45, O103, O111, O121, and O145 serogroups. The serogroups targeted by mqPCR-1 were O26, O103, and O111; those targeted by mqPCR-2 were O45, O121, and O145. The target gene for O26, O103, O111, O45, and O145 was wzx (3), which encodes for flippase involved in O-polysaccharide export (25). The target genes for O121 were wbqE and wbqF, which encode for putative glycosyl and acetyl transferases, respectively (3). The reporter dyes, FAM, VIC/MAX, and Texas Red, were conjugated at 5′ ends, and Black Hole Quencher dyes I and II were conjugated at 3′ ends of the probes to detect amplification products specific to each gene target. Primers and probes (Integrated DNA Technologies, Inc., Coralville, IA) designed to target O-antigen genes of the six non-O157 E. coli serogroups are shown in Table 1.

TABLE 1.

Primers and probes used in the multiplex quantitative PCR assaysa

Primers and probes used in the multiplex quantitative PCR assaysa
Primers and probes used in the multiplex quantitative PCR assaysa

Optimization of the assays.

The assays were optimized and validated with pure cultures of one strain each of six serogroups of non-O157 STEC (Table 2) individually and, subsequently, with pooled mixtures of two different combinations (O26, O103, and O111 for mqPCR-1 and O45, O121, and O145 for mqPCR-2). The strains of non-O157 STEC, stored at −80°C on cryobeads (CryoCare, Key Scientific Products, Round Rock, TX), were streaked onto blood agar plates (Remel, Lenexa, KS). Single colonies of each serogroup were inoculated and grown overnight in Luria-Bertani broth (Difco, BD, Sparks, MD) at 37°C, and then 100 μl of the culture was added to 10 ml of the broth and incubated at 37°C for 3 to 4 h until an absorbance of 0.4 at 600 nm (approximately 8 log CFU/ml) was reached. Equal volumes of the cultures of each serogroup were mixed into two combinations as described before. One milliliter of the culture suspension (individual serogroup and pooled mixtures) was boiled for 10 min and centrifuged at 9,300 × g for 5 min; the supernatant was used as DNA template for mqPCR assays. DNA was also subjected to 10-fold serial dilutions (10−1 to 10−7), and standard curves were generated with the mqPCR assays.

TABLE 2.

Sources and virulence gene profiles of Escherichia coli strains used for optimization of multiplex quantitative PCR assays

Sources and virulence gene profiles of Escherichia coli strains used for optimization of multiplex quantitative PCR assays
Sources and virulence gene profiles of Escherichia coli strains used for optimization of multiplex quantitative PCR assays

Running conditions of the assays.

The working concentrations of all primers in the primer mixture were 10 pmol/μl. The working concentrations of probes were 5 pmol/μl for O26 and O103, 3 pmol/μl for O121, 2.5 pmol/μl for O111 and O45, and 1.25 pmol/μl for O145. The PCR reaction mixture contained 10 μl of 2X Bio-Rad iQ Multiplex Powermix (Bio-Rad Laboratories, Life Science Group, Hercules, CA), 4 μl of nuclease-free water, 1 μl of primer mixture, 1 μl of each probe, and 2 μl of DNA template, making a total reaction volume of 20 μl. PCR was performed with the BioRad CFX96 mqPCR system, and data were analyzed using CFX Manager software version 3.1 (Bio-Rad Laboratories). The optimized PCR amplification protocol included a 10-min denaturation at 95°C, followed by 45 cycles of 95°C for 15 s, 56°C for 20 s, and 72°C for 40 s.

Specificity of the assays.

Specificity of the assays was evaluated with a number of E. coli and non–E. coli strains from our culture collection. A total of 170 strains (human clinical and bovine origin) belonging to six non-O157 E. coli serogroups (35 strains of O26, 40 strains of O103, 40 strains of O111, 25 strains of O45, 12 strains of O121, and 18 strains of O145) were used as a positive control, and another 100 strains representing 42 E. coli serogroups other than the six non-O157 serogroups and other bacterial species (Klebsiella pneumoniae, Proteus mirabilis, and Serratia marcescens) were used as a negative control.

In addition, pure cultures of target serogroups for each assay were pooled by mixing each serogroup at equal and different concentrations. Each serogroup of the assay was added at equal concentration and at decreasing concentrations (1:100, 1:10−1, 1:10−2, 1:10−3, 1:10−4, 1:10−5, 1:10−6, 1:10−7) to a mixture containing the other two targets at fixed concentration and was inoculated into a cattle fecal sample that tested negative for the six target non-O157 serogroups. One gram of spiked fecal samples was added to 9 ml of E. coli (EC) broth (Difco, BD), vortexed, and incubated at 40°C for 6 h. Extracted DNA from pre- and postenrichment fecal samples was subjected to mqPCR assays.

Sensitivity of the assays using pure cultures of non-O157 STEC.

Sensitivity of the assays was determined by using 10-fold serial dilutions of pure cultures of the six serogroups. Three different combinations of pooled pure cultures were prepared as described before: O26, O103, and O111 for mqPCR-1; O45, O121, and O145 for mqPCR-2; and O26, O45, O103, O111, O121, and O145 for both assays. Ten-fold serial dilutions (10−1 to 10−7) of each pooled pure culture grown in Luria-Bertani broth were performed, and 100-μl aliquots of the last three dilutions (10−5, 10−6, and 10−7) were spread plated onto blood agar plates (four plates per dilution) to determine initial cell concentrations (CFU per milliliter). One milliliter from each dilution (100 to 10−7) was removed, boiled for 10 min, and centrifuged at 9,300 ×g for 5 min. Supernatants were used as DNA templates for mqPCR, and standard curves were generated to determine correlation coefficients, amplification efficiencies, and detection limits for each assay. Both mqPCR assays (targeting three serogroups each) were compared to the corresponding single (targeting a single serogroup) and duplex (targeting two serogroups) assays. The experiment was repeated on a different day.

Sensitivity of the assays using feces spiked with pure cultures of non-O157 STEC.

Fifteen pen-floor fecal samples from cattle housed in the Kansas State University feedlot were collected and tested by both sets of mqPCR assays (1 and 2), and a sample that was negative for the six serogroups was selected to spike with pure cultures. Three different combinations of pooled pure cultures were prepared as described before: O26, O103, and O111 (mqPCR-1); O45, O121, and O145 (mqPCR-2); and O26, O45, O103, O111, O121, and O145 (both assays). Ten-fold serial dilutions (100 to 10−7) of each mixture were performed in Luria-Bertani broth, and initial concentrations of each were determined. Aliquots of 10 g of fecal sample were inoculated with 1 ml of different dilutions (10−1 to 10−7) of pooled pure cultures and were mixed as thoroughly as possible; 1 g of the spiked fecal sample was added to 9 ml of EC broth. The fecal suspension was vortexed and incubated at 40°C for 6 h. One-milliliter samples of pre- and postenrichment fecal suspensions were boiled for 10 min and centrifuged at 9,300 × g for 5 min. DNA cleanup of pre- and postenrichment fecal suspensions was performed using the GeneClean Turbo Kit (MP Biomedicals LLC, Solon, OH). Purified DNA from pre- and postenriched fecal suspensions and boiled DNA from pooled pure cultures used to spike the fecal sample were subjected to mqPCR, and standard curves were generated to determine the correlation coefficients, amplification efficiencies, and detection limits of each assay. The experiment was repeated with a different fecal sample.

Application of mqPCR assays to quantify E. coli serogroups and virulence genes in fecal samples from feedlot cattle.

A total of 576 pen-floor fecal samples from cattle housed in 24 pens (24 samples per pen) at a commercial feedlot in the central United States were collected the day before transportation of cattle to slaughter in the summer of 2013. Details regarding the design of the study and characteristics of the study population are available in Dewsbury et al. (11). Fecal samples were suspended in EC broth, and DNA was extracted as described above. The DNA was subjected to the two mqPCR assays (mqPCR-1 and mqPCR-2) and to a previously developed four-plex qPCR assay (mqPCR-3) targeting O157 serogroup (rfbEO157) and the three major virulence genes, eae, stx1, and stx2 (30). Cycle threshold (CT) values were determined to quantify seven major E. coli serogroups and three virulence genes. Concentration of each serogroup and virulence gene was determined based on standard curves obtained using cattle fecal samples spiked with known concentrations of pure cultures of E. coli O157 and non-O157 STEC serogroups.

Application of mqPCR assays and comparison with cPCR and culture methods for the detection of E. coli O157 and six non-O157 E. coli serogroups in fecal samples from feedlot cattle.

Fecal samples (n = 576) suspended in EC broth were enriched at 40°C for 6 h and then were subjected to three mqPCR assays (mqPCR-1, mqPCR-2, and mqPCR-3), 10-plex cPCR assay targeting the same genes (seven serogroups and three virulence genes) (3), and culture-based methods (11) for the detection of the seven serogroups. For the culture-based detection method, postenriched fecal samples were subjected to immunomagnetic separation with seven serogroup-specific beads (Abraxis, Warminster, PA), and 50 μl of post–immunomagnetic separation bead suspensions were spread plated onto chromogenic Possé medium (32) that was modified to include novobiocin at 5 mg/liter and potassium tellurite at 0.5 mg/liter for non-O157 serogroups, and sorbitol MacConkey with cefixime (0.05 mg/liter) and potassium tellurite (2.5 mg/liter) for E. coli O157. The plates were then incubated at 37°C for 18 to 24 h, and six chromogenic colonies (shades of blue, purple, mauve, or green) from modified Possé medium and sorbitol-negative colonies from sorbitol MacConkey with cefixime and potassium tellurite were picked and streaked onto blood agar and incubated for 18 to 24 h. For non-O157 serogroups, colonies from each of the six streaks on the blood agar plate were pooled in 50 μl of distilled water, boiled for 10 min, and used as a DNA template for seven-plex cPCR targeting the seven serogroups (31). For O157, colonies on blood agar were subjected to indole test, and indole-positive colonies were tested for O157 antigen using a latex agglutination assay (E. coli O157 latex test kit, Oxoid Ltd., Cheshire, England). Agglutination-positive colonies were then subjected to a six-plex cPCR assay (rfbEO157, fliCH7, stx1, stx2, eae, and ehxA) to confirm the O157 serogroup, H7 flagellar gene, and virulence genes (4).

Statistical analyses.

The proportion of positive samples for each serogroup and virulence gene based on the three detection methods (mqPCR, cPCR, and culture method) was calculated as the number of samples detected as positive for each gene by each detection method divided by the total number of samples tested by each detection method. The Cohen's kappa statistic was used to evaluate the agreement beyond that due to chance among mqPCR, cPCR, and culture methods for the detection of seven E. coli serogroups and three virulence genes. Interpretation of the kappa statistic was done based on the scale proposed by Landis and Koch (22). The McNemar's chi-square test was used to compare the proportion of positive samples obtained by the three detection methods (29). When the P value of McNemar's test is not significant (P > 0.05), there is little evidence to conclude that the tests are different; and, when the P value is significant (P < 0.05), there is a significant disagreement between tests, indicating that there is little value in further assessing agreement by Cohen's kappa statistic. In the latter case, Cohen's kappa statistics are provided for reference only. A receiver operating characteristic curve was generated by plotting the true positive against the false positive proportions across a range of reciprocal CT values (1/CT value) to determine the CT value that yields optimum sensitivity and specificity and the highest proportion of correctly classified samples by cPCR in relation to mqPCR. Statistical analyses were performed in Stata/MP version 12.0 (StataCorp LP, College Station, TX).

RESULTS

Specificity of the assays.

All 170 strains belonging to O26, O103, O45, O111, O121, and O145 serogroups were detected by the corresponding mqPCR-1 and mqPCR-2 assays (data not shown). No cross-amplification occurred with nontargeted serogroups. None of the E. coli strains belonging to serogroups other than the six non-O157 serogroups were detected by the mqPCR assays, nor were other bacterial species. Both assays (mqPCR-1 and mqPCR-2) correctly detected the target serogroups when performed with pooled pure cultures and spiked fecal samples of two different combinations containing equal or different concentrations of target serogroups (data not shown).

Sensitivity of the assays with pure cultures and pure culture–spiked fecal samples.

The detection limits of the two assays, mqPCR-1 and mqPCR-2, with pure cultures were 3 log CFU/ml, with mean CT values of 37.1 and 37.4 and amplification efficiencies from 99 to 104% and 99 to 102%, respectively. The correlation coefficient was >0.99 for both assays (Table 3). Detection limits, correlation coefficients, and amplification efficiencies of both assays were similar to the corresponding single (targeting a single serogroup) or duplex (targeting two serogroups) assays (data not shown). In fecal samples spiked with serially diluted, pooled pure cultures, detection limits of both assays (mqPCR-1 and mqPCR-2) were 4 log CFU/g of feces, with mean CT values of 37.2 and 37.4, respectively. After enrichment, detection limits of both assays improved to 2 log CFU/g of feces, with mean CT values of 37.6 and 37.9 for mqPCR-1 and mqPCR-2, respectively. The correlation coefficients and amplification efficiencies are shown in Table 3.

TABLE 3.

Detection limits, correlation coefficients, and amplification efficiencies of the multiplex quantitative PCR assays for pure cultures and cattle fecal samples spiked with pure cultures of non-O157 Shiga toxin–producing Escherichia coli strainsa

Detection limits, correlation coefficients, and amplification efficiencies of the multiplex quantitative PCR assays for pure cultures and cattle fecal samples spiked with pure cultures of non-O157 Shiga toxin–producing Escherichia coli strainsa
Detection limits, correlation coefficients, and amplification efficiencies of the multiplex quantitative PCR assays for pure cultures and cattle fecal samples spiked with pure cultures of non-O157 Shiga toxin–producing Escherichia coli strainsa

Application of mqPCR assays to quantify E. coli serogroups and virulence genes in cattle feces.

Preen-riched fecal samples that yielded CT values less than or equal to the mean cut-off CT value were considered positive (37.2 for mqPCR 1, 37.4 for mqPCR-2, and 38.3 for mqPCR-3) for the serogroups. Of the 576 fecal samples (before enrichment), 175 (30.4%) were quantifiable for at least one of the seven serogroups. Serogroup O157 (n = 66; 11.5%) was the predominant serogroup quantified by mqPCR, followed by O45 (n = 41; 7.1%), O103 (n = 41; 7.1%), O121 (n = 37; 6.4%), O26 (n = 29; 5%), O111 (n = 2; 0.3%), and O145 (n = 2; 0.3%) (Table 4). The concentrations of E. coli serogroups ranged from 4 to 7 log CFU/g of feces. A greater proportion of fecal samples tested positive for E. coli serogroups at concentrations of 4 (19.3%) than 5 log CFU/g (17.4%), and none of the non-O157 E. coli serogroups had concentrations >6 log CFU/g. Seven fecal samples (1.2%) were positive for O157 at concentrations of 6 to <7 log CFU/g. Of the 175 fecal samples that tested positive for at least one of the seven E. coli serogroups at quantifiable concentrations (≥4 log CFU/g), 141 (80.6%) were positive for one serogroup, 28 (16%) for two serogroups, and three (1.7%) each for three and four serogroups. The concentrations of stx1 and eae ranged from 4 to 8 log CFU/g, and that of stx2 ranged from 4 to 7 log CFU/g (Table 4).

TABLE 4.

Quantification of seven major Escherichia coli serogroups and three virulence genes in preenriched cattle feces using multiplex quantitative PCR assays

Quantification of seven major Escherichia coli serogroups and three virulence genes in preenriched cattle feces using multiplex quantitative PCR assays
Quantification of seven major Escherichia coli serogroups and three virulence genes in preenriched cattle feces using multiplex quantitative PCR assays

Application of mqPCR assays to detect E. coli serogroups and virulence genes in cattle feces.

Post-enriched fecal samples that yielded CT values less than or equal to the mean cut-off CT value were considered positive (37.6 for mqPCR-1, 37.9 for mqPCR-2, and 37.9 for mqPCR-3) for serogroups. Of 576 samples, 566 (98.3%) were positive for at least one of the seven serogroups, and all except one sample were positive for at least one of the three virulence genes. Serogroup O157 (89.8%) was the predominant E. coli detected, followed by O103 (84.7%), O26 (59.0%), O121 (57.8%), O45 (55.9%), and O145 (5.9%). Only four samples (0.7%) tested positive for serogroup O111. Among the virulence genes, eae (99.7%) was predominant, followed by stx1 (95.7%) and stx2 (94.6%) (Table 5).

TABLE 5.

Number of cattle fecal samples that tested positive for seven major Escherichia coli serogroups and three virulence genes by mqPCR, cPCR, and culture methods of detection, and agreement between testsa

Number of cattle fecal samples that tested positive for seven major Escherichia coli serogroups and three virulence genes by mqPCR, cPCR, and culture methods of detection, and agreement between testsa
Number of cattle fecal samples that tested positive for seven major Escherichia coli serogroups and three virulence genes by mqPCR, cPCR, and culture methods of detection, and agreement between testsa

Based on cPCR assay, 484 (84.0%) of the 576 fecal samples tested positive for at least one of the seven serogroups, and 571 (99.1%) tested positive for at least one of the three virulence genes. Of the 10 samples that tested negative for any of the seven serogroups by three mqPCR assays, one sample was positive for O157 by cPCR. E. coli O103 (56.6%) was the most commonly detected serogroup by cPCR, followed by O157 (54.7%), O26 (44.4%), O121 (22.9%), O45 (17.9%), O145 (1.9%), and O111 (0.7%). Among the virulence genes, eae (97.4%) was predominant, followed by stx2 (94.1%) and stx1 (64.4%) (Table 5).

Based on the culture method, 481 (83.5%) fecal samples tested positive for at least one of the seven serogroups. E. coli O103 (60.2%) was the most commonly detected serogroup, followed by O157 (43.1%), O26 (22.7%), O45 (16.7%), O145 (3.0%), O121 (2.3%), and O111 (0.2%) (Table 5). Because the pooled colonies were screened by a seven-plex PCR targeting only the seven major serogroups, virulence gene detection was not part of the culture detection method.

The McNemar's test indicated a significant (P < 0.01) difference between the proportions of positive fecal samples detected by mqPCR and cPCR for all the target genes except for wzxO111 and stx2. There was also a significant (P < 0.01) difference in the proportion of positive samples detected by mqPCR and the culture method for all target genes, except wzxO111. In both cases, the kappa statistics were provided for reference only (Table 5). The receiver operating characteristic curve analysis showed that a mean CT value of 32.3 yielded optimum sensitivity (83.3 to 100%), specificity (94.0 to 100%), and the highest number of correctly classified samples (93.0 to 100%) by cPCR in relation to mqPCR for all the seven serogroups.

Table 6 shows the number and proportion of fecal samples that tested positive or negative for each serogroup and virulence gene by three mqPCR assays, which tested positive or negative by cPCR or the culture method. Of the fecal samples that tested positive by three mqPCR assays for the seven serogroups, 0 to 68% tested negative by the cPCR, and 32.0 to 100% tested negative by the culture method, depending on the serogroup. Of the fecal samples that tested negative by the three mqPCR assays, the proportion of samples that were also negative by cPCR or culture method ranged from 83.0 to 100%. However, a few fecal samples that tested negative by mqPCR were positive by cPCR (4 for O157, 15 for stx2, 3 for stx1, and 1 for eae). Similarly, the culture method detected a few samples as positive that tested negative based on mqPCR (16 for O26, 15 for O103, 10 for O45, 8 for O145, 7 for O157, 1 for O111, and 1 for O121) (Table 6).

TABLE 6.

Comparison of mqPCR, cPCR, and culture method for the detection of seven Escherichia coli serogroups and three virulence genes in enriched fecal samplesa

Comparison of mqPCR, cPCR, and culture method for the detection of seven Escherichia coli serogroups and three virulence genes in enriched fecal samplesa
Comparison of mqPCR, cPCR, and culture method for the detection of seven Escherichia coli serogroups and three virulence genes in enriched fecal samplesa

DISCUSSION

We have developed two multiplex qPCR assays that target serogroup-specific O-antigen genes to detect and quantify E. coli O26, O103, O111, O45, O121, and O145 serogroups in cattle feces. Both assays were specific to their corresponding target genes, and the detection limits of both assays in pure cultures were 3 log CFU/ml. The detection limit increased to 4 log CFU/g when fecal samples were spiked with known concentrations of pure cultures. The enrichment of spiked fecal samples in EC broth for 6 h at 40°C improved the detection limit of both assays to 2 log CFU/g feces. These two assays, in combination with the four-plex assay targeting O157 serogroup and the three major virulence genes (30), can be used to quantify (before enrichment) and detect (after enrichment) seven E. coli serogroups and three major virulence genes in cattle feces.

Real-time PCR assays have been developed for the rapid detection of non-O157 E. coli in food matrices (14, 15, 24). Also, several commercially available real-time PCR-based detection systems have been evaluated for detection of non-O157 E. coli serogroups in beef and beef products (13, 15, 37). Conventional PCR assays for the detection of non-O157 E. coli serogroups in cattle feces have been reported (3, 6, 10, 31), but, to date, there has been no published study on the utility of mqPCR assays to detect or quantify the six non-O157 E. coli serogroups in feces of naturally shedding feedlot cattle. Quantitative PCR assays targeting serogroup-specific virulence genes have been developed to detect non-O157 E. coli serogroups in cattle feces. Sharma (34) developed two multiplex qPCR assays that target a region of eae gene specific to O26, O111, and O157 serogroups (assay 1) and stx1 and stx2 (assay 2). Luedtke et al. (27) developed a four-plex qPCR to detect EHEC in cattle feces by targeting ecf1, an EHEC-specific gene, and the three major virulence genes, eae, stx1, and stx2. That assay was designed to detect EHEC but not individual serogroups of EHEC. Anklam et al. (1) developed four separate mqPCR assays to target O-antigen genes of seven (O26, O45, O103, O111, O121, O145, and O157) serogroups and four virulence genes (stx1, stx2, eae, and ehxA) in cattle feces. One assay targeted O26 (wzyO26), O103 (wzxO103), and O145 (wzxO145); the second assay targeted O45 (wzyO45), O111 (manCO111), and O121 (wzxO121); the third assay targeted O157 (rfbEO157); and the fourth assay targeted stx1, stx2, eae, and ehxA (enterohemolysin). Our two mqPCR assays targeting the six non-O157 serogroups are, to some extent, similar to those of Anklam et al. (1): the same genes were targeted for two serogroups (O103 and O145), but for the other four serogroups (O26, O45, O111, and O121), the targeted genes were different (wzxO26 instead of wzyO26; wzxO45 instead of wzyO45; wzxO111 instead of manCO111; and wbqE and wbqF instead of wzxO121). We chose wzx genes for O26, O45, and O111 mainly because they worked better with the primers and probes of the other two O-antigen targets (O103 and O145) in the same reaction. As to O121, a Shigella dysenteriae strain (GenBank accession: AY380835.1) has an O-antigen gene cluster nearly identical to that of E. coli O121. We used minor differences in wbqE and wbqF genes to differentiate O121 from the Shigella strain. We used target regions of the wzx gene for O103 and O145 to make sure that there was no secondary structure or interference with each other; hence, the primers are different. Also, different combinations of the serogroups were included in the two assays, compared with Anklam et al. (1). We did not include ehxA in our assay because the gene is present in many of the non–Shiga toxigenic E. coli pathogens (9, 26). Anklam et al. (1) validated their assays in detecting E. coli serogroups and virulence genes by using pure cultures and cattle fecal samples spiked with pure cultures, but they did not evaluate the applicability of these assays to detect and quantify E. coli serogroups and virulence genes in feces of naturally shedding cattle. The detection limits of the assays reported by Anklam et al. (1) were 103 and 104 CFU/ml for pure cultures and spiked fecal samples, respectively, which improved to 100 after a 6-h enrichment of fecal samples. The detection limits of the pure cultures and spiked fecal samples are in agreement with the detection limits of our two assays. However, the detection limit for enriched samples was lower in their study than ours (100 versus 102), possibly because the fecal samples they used to inoculate with pure cultures were diluted 1:50 compared with the 1:10 dilution used in our assays.

To our knowledge, this is the first study to assess the applicability of mqPCR assays for detection and quantification of the six non-O157 E. coli serogroups in fecal samples from naturally shedding cattle. All three mqPCR assays detected more samples as positive for one or more of the serogroups compared with detection by cPCR and the culture method. McNemar's test indicated disagreement between the proportion of positive samples detected by mqPCR, cPCR, and the culture method. The disagreement between mqPCR and cPCR is explained by receiver operating characteristic curve analysis of CT values, which indicated that cPCR is less sensitive than mqPCR. It is known that real-time PCR assay is more sensitive than conventional PCR- or culture-based testing methods (23). As with any PCR assay, there is also a possibility of false positives because of amplification of DNA from nonviable cells in the feces. However, a few samples that were negative by mqPCR (16 for O26, 10 for O45, 15 for O103, 1 for O111, 1 for O121, 8 for O145, and 7 for O157) were positive by cPCR or the culture method. A likely reason for the difference could be the uneven distribution of target genes or serogroups in each aliquot that was subjected to the different methods of detection. Also, the misidentification of culture-positive samples by mqPCR is likely reflective of the difference in detection limits between the two methods. The mqPCR requires a concentration of 104 CFU/g for detection, whereas the culture method, which uses immunomagnetic separation beads, may randomly capture E. coli cells at lower concentrations. A similar disagreement between culture method and mqPCR assay has been reported for E. coli O157 in cattle feces (18, 30). The number of fecal samples testing positive for intimin and Shiga toxin genes was generally higher than the number positive for the seven E. coli serogroups, which is likely due to the presence of STEC serogroups other than the seven targeted in this study as well as free bacteriophages carrying Shiga toxin genes in fecal samples (8).

Apart from estimating the prevalence of E. coli serogroups and their virulence genes in cattle feces, determining their concentrations is essential for estimating the risk of foodborne illnesses associated with fecal shedding of E. coli serogroups and their associated virulence genes. Estimating the concentration of E. coli serogroups is beneficial for evaluating the efficacy of intervention strategies employed to reduce the pathogen load in feces so as to reduce hide and carcass contaminations (2). Data on the concentration of non-O157 E. coli serogroups in cattle feces are nonexistent. Quantitative PCR has been used to quantify E. coli O157:H7 and virulence genes (16, 18, 30, 36) in cattle feces. In our study, 175 (30.4%) of 576 fecal samples were positive for at least one of the seven E. coli serogroups, with a concentration of ≥4 log CFU/g of feces. A majority of the samples (80.6%) that were quantifiable were positive for one of the seven serogroups, with E. coli O157 being the most common serogroup. However, an inherent limitation of the mqPCR assay, similar to cPCR, is that the presence of virulence genes cannot be directly associated to any particular serogroup in the sample.

In conclusion, the two sets of mqPCR assays are rapid diagnostic tools for the detection and quantification of six major non-O157 E. coli serogroups in cattle feces. These two assays, together with the four-plex assay targeting E. coli O157 and three virulence genes (stx1, stx2, and eae), can be used to detect and quantify seven major E. coli serogroups and three virulence genes in cattle feces. This is the first study to provide data on concentrations of non-O157 E. coli serogroups in cattle feces and to identify the existence of a subset of cattle, similar to super shedders of O157, that shed non-O157 at high concentrations.

ACKNOWLEDGMENTS

This material is based upon work that is supported by the National Institute of Food and Agriculture, U.S. Department of Agriculture, under award no. 2012-68003-30155. The authors thank Neil Wallace, Melissa Juby, and Joy Miller for their assistance in this project.

REFERENCES

REFERENCES
1.
Anklam
,
K. S.
,
K. S. T.
Kanankege
,
T. K.
Gonzales
,
C. W.
Kaspar
, and
D.
Döpfer
.
2012
.
Rapid and reliable detection of Shiga toxin–producing Escherichia coli by real-time multiplex PCR
.
J. Food Prot
.
75
:
643
650
.
2.
Arthur
,
T. M.
,
J. E.
Keen
,
J. M.
Bosilevac
,
D. M.
Brichta-Harhay
,
N.
Kalchayanand
,
S. D.
Shackelford
,
T. L.
Wheeler
,
X.
Nou
, and
M.
Koohmaraie
.
2009
.
Longitudinal study of Escherichia coli O157:H7 in a beef cattle feedlot and role of high-level shedders in hide contamination
.
Appl. Environ. Microbiol
.
75
:
6515
6523
.
3.
Bai
,
J.
,
Z. D.
Paddock
,
X.
Shi
,
S.
Li
,
B.
An
, and
T. G.
Nagaraja
.
2012
.
Applicability of a multiplex PCR to detect the seven major Shiga toxin-producing Escherichia coli based on genes that code for serogroup-specific O-antigens and major virulence factors in cattle feces
.
Foodborne Pathog. Dis
.
9
:
541
548
.
4.
Bai
,
J.
,
X.
Shi
, and
T. G.
Nagaraja
.
2010
.
A multiplex PCR procedure for the detection of six major virulence genes in Escherichia coli O157:H7
.
J. Microbiol. Methods
82
:
85
89
.
5.
Baltasar
,
P.
,
S.
Milton
,
W.
Swecker
,
F.
Elvinger
, and
M.
Ponder
.
2014
.
Shiga toxin-producing Escherichia coli distribution and characterization in a pasture-based cow-calf production system
.
J. Food Prot
.
77
:
722
731
.
6.
Cernicchiaro
,
N.
,
C. A.
Cull
,
Z. D.
Paddock
,
X.
Shi
,
J.
Bai
,
T. G.
Nagaraja
, and
D. G.
Renter
.
2013
.
Prevalence of Shiga toxin-producing Escherichia coli and associated virulence genes in feces of commercial feedlot cattle
.
Foodborne Pathog. Dis
.
10
:
835
841
.
7.
Chase-Topping
,
M.
,
D.
Gally
,
C.
Low
,
L.
Matthews
, and
M.
Woolhouse
.
2008
.
Super-shedding and the link between human infection and livestock carriage of Escherichia coli O157
.
Nat. Rev. Microbiol
.
6
:
904
912
.
8.
Conrad
,
C. C.
,
K.
Stanford
,
T. A.
McAllister
,
J.
Thomas
, and
T.
Reuter
.
2014
.
Further development of sample preparation and detection methods for O157 and the top 6 non-O157 STEC serogroups in cattle feces
.
J. Microbiol. Methods
105
:
22
30
.
9.
Cookson
,
A. L.
,
J.
Bennett
,
F.
Thomson-Carter
, and
G. T.
Attwood
.
2007
.
Molecular subtyping and genetic analysis of the enter-ohemolysin gene (ehxA) from Shiga toxin-producing Escherichia coli and atypical enteropathogenic E. coli
.
Appl. Environ. Microbiol
.
73
:
6360
6369
.
10.
Dargatz
,
D. A.
,
J.
Bai
,
B. V.
Lubbers
,
C. A.
Kopral
,
B.
An
, and
G. A.
Anderson
.
2013
.
Prevalence of Escherichia coli O-types and Shiga toxin genes in fecal samples from feedlot cattle
.
Foodborne Pathog. Dis
.
10
:
392
396
.
11.
Dewsbury
,
D. M. A.
,
D. G.
Renter
,
P. B.
Shridhar
,
L. W.
Noll
,
X.
Shi
,
T. G.
Nagaraja
, and
N.
Cernicchiaro
.
2015
.
Summer and winter prevalence of Shiga toxin-producing Escherichia coli (STEC) O26, O45, O103, O111, O121, O145, and O157 in feces of feedlot cattle
.
Foodborne Pathog. Dis
.
12
:
726
732
.
12.
Fox
,
J. T.
,
D. G.
Renter
,
M. W.
Sanderson
,
A. L.
Nutsch
,
X.
Shi
, and
T. G.
Nagaraja
.
2008
.
Associations between the presence and magnitude of Escherichia coli O157 in feces at harvest and contamination of preintervention beef carcasses
.
J. Food Prot
.
71
:
1761
1767
.
13.
Fratamico
,
P. M.
, and
L. K.
Bagi
.
2012
.
Detection of Shiga toxin-producing Escherichia coli in ground beef using the GeneDisc real-time PCR system
.
Front. Cell Infect. Microbiol
.
2
:
152
.
14.
Fratamico
,
P. M.
,
L. K.
Bagi
,
W. C.
Cray
,
N.
Narang
,
X.
Yan
,
M.
Medina
, and
Y.
Liu
.
2011
.
Detection by multiplex real-time polymerase chain reaction assays and isolation of Shiga toxin-producing Escherichia coli serogroups O26, O45, O103, O111, O121, and O145 in ground beef
.
Foodborne Pathog. Dis
.
8
:
601
607
.
15.
Fratamico
,
P. M.
,
J. L.
Wasilenko
,
B.
Garman
,
D. R.
Demarco
,
S.
Varkey
,
M.
Jensen
,
K.
Rhoden
, and
G.
Tice
.
2014
.
Evaluation of a multiplex real-time PCR method for detecting Shiga toxin-producing Escherichia coli in beef and comparison to the U.S. Department of Agriculture Food Safety and Inspection Service Microbiology laboratory guidebook method
.
J. Food Prot
.
77
:
180
188
.
16.
Guy
,
R. A.
,
D.
Tremblay
,
L.
Beausoleil
,
J.
Harel
, and
M.-J.
Champagne
.
2014
.
Quantification of E. coli O157 and STEC in feces of farm animals using direct multiplex real time PCR (qPCR) and a modified most probable number assay comprised of immunomagnetic bead separation and qPCR detection
.
J. Microbiol. Methods
99
:
44
53
.
17.
Gyles
,
C. L.
2007
.
Shiga toxin-producing Escherichia coli: an overview
.
J. Anim. Sci
.
85
:
E45
62
.
18.
Jacob
,
M. E.
,
J.
Bai
,
D. G.
Renter
,
A. T.
Rogers
,
X.
Shi
, and
T. G.
Nagaraja
.
2014
.
Comparing real-time and conventional PCR to culture-based methods for detecting and quantifying Escherichia coli O157 in cattle feces
.
J. Food Prot
.
77
:
314
319
.
19.
Jacob
,
M. E.
,
D. G.
Renter
, and
T. G.
Nagaraja
.
2010
.
Animal- and truckload-level associations between Escherichia coli O157:H7 in feces and on hides at harvest and contamination of preevisceration beef carcasses
.
J. Food Prot
.
73
:
1030
1037
.
20.
Jacob
,
M. E.
,
X.
Shi
,
B.
An
,
T. G.
Nagaraja
, and
J.
Bai
.
2012
.
Evaluation of a multiplex real-time polymerase chain reaction for the quantification of Escherichia coli O157 in cattle feces
.
Foodborne Pathog. Dis
.
9
:
79
85
.
21.
Kalchayanand
,
N.
,
T. M.
Arthur
,
J. M.
Bosilevac
,
J. E.
Wells
, and
T. L.
Wheeler
.
2013
.
Chromogenic agar medium for detection and isolation of Escherichia coli serogroups O26, O45, O103, O111, O121, and O145 from fresh beef and cattle feces
.
J. Food Prot
.
76
:
192
199
.
22.
Landis
,
J. R.
, and
G. G.
Koch
.
1977
.
The measurement of observer agreement for categorical data
.
Biometrics
33
:
159
174
.
23.
Lemmon
,
G. H.
, and
S. N.
Gardner
.
2008
.
Predicting the sensitivity and specificity of published real-time PCR assays
.
Ann. Clin. Microbiol. Antimicrob
.
7
:
18
.
24.
Lin
,
A.
,
O.
Sultan
,
H. K.
Lau
,
E.
Wong
,
G.
Hartman
, and
C. R.
Lauzon
.
2011
.
O serogroup specific real time PCR assays for the detection and identification of nine clinically relevant non-O157 STECs
.
Food Microbiol
.
28
:
478
483
.
25.
Liu
,
D.
,
R. A.
Cole
, and
P. R.
Reeves
.
1996
.
An O-antigen processing function for Wzx (RfbX): a promising candidate for O-unit flippase
.
J. Bacteriol
.
178
:
2102
2107
.
26.
Lorenz
,
S. C.
,
I.
Son
,
A.
Maounounen-Laasri
,
A.
Lin
,
M.
Fischer
, and
J. A.
Kase
.
2013
.
Prevalence of hemolysin genes and comparison of ehxA subtype patterns in Shiga toxin-producing Escherichia coli (STEC) and non-STEC strains from clinical, food, and animal sources
.
Appl. Environ. Microbiol
.
79
:
6301
6311
.
27.
Luedtke
,
B. E.
,
J. L.
Bono
, and
J. M.
Bosilevac
.
2014
.
Evaluation of real time PCR assays for the detection and enumeration of enterohemorrhagic Escherichia coli directly from cattle feces
.
J. Microbiol. Methods
105
:
72
79
.
28.
Matthews
,
L.
,
I.
McKendrick
,
H.
Ternent
,
G.
Gunn
,
B.
Synge
, and
M.
Woolhouse
.
2006
.
Super-shedding cattle and the transmission dynamics of Escherichia coli O157
.
Epidemiol. Infect
.
134
:
131
142
.
29.
McNemar
,
Q.
1947
.
Note on the sampling error of the difference between correlated proportions or percentages
.
Psychometrika
12
:
153
157
.
30.
Noll
,
L. W.
,
P. B.
Shridhar
,
X.
Shi
,
B.
An
,
N.
Cernicchiaro
,
D. G.
Renter
,
T. G.
Nagaraja
, and
J.
Bai
.
2015
.
A four-plex real-time PCR assay, based on rfbE, stx1, stx2, and eae genes, for the detection and quantification of Shiga toxin-producing Escherichia coli O157 in cattle feces
.
Foodborne Pathog. Dis
.
12
:
787
794
.
31.
Paddock
,
Z.
,
X.
Shi
,
J.
Bai
, and
T. G.
Nagaraja
.
2012
.
Applicability of a multiplex PCR to detect O26, O45, O103, O111, O121, O145, and O157 serogroups of Escherichia coli in cattle feces
.
Vet. Microbiol
.
156
:
381
388
.
32.
Possé
,
B.
,
L.
De Zutter
,
M.
Heyndrickx
, and
L.
Herman
.
2008
.
Novel differential and confirmation plating media for Shiga toxin-producing Escherichia coli serotypes O26, O103, O111, O145 and sorbitol-positive and -negative O157
.
FEMS Microbiol. Lett
.
282
:
124
131
.
33.
Scallan
,
E.
,
R. M.
Hoekstra
,
F. J.
Angulo
,
R. V.
Tauxe
,
M.-A.
Widdowson
,
S. L.
Roy
,
J. L.
Jones
, and
P. M.
Griffin
.
2011
.
Foodborne illness acquired in the United States—major pathogens
.
Emerg. Infect. Dis
.
17
:
7
15
.
34.
Sharma
,
V. K.
2002
.
Detection and quantitation of enterohemorrhagic Escherichia coli O157, O111, and O26 in beef and bovine feces by real-time polymerase chain reaction
.
J. Food Prot
.
65
:
1371
1380
.
35.
U.S. Department of Agriculture, Food Safety and Inspection Service
.
2011
.
Shiga toxin-producing Escherichia coli in certain raw beef products
.
Fed. Regist
.
76
:
58157
58165
.
36.
Verstraete
,
K.
,
E.
Van Coillie
,
H.
Werbrouck
,
S.
Van Weyenberg
,
L.
Herman
,
J.
Del-Favero
,
P.
De Rijk
,
L.
De Zutter
,
M.-A.
Joris
,
M.
Heyndrickx
, and
K.
De Reu
.
2014
.
A qPCR assay to detect and quantify Shiga toxin-producing E. coli (STEC) in cattle and on farms: a potential predictive tool for STEC culture-positive farms
.
Toxins (Basel)
6
:
1201
1221
.
37.
Wasilenko
,
J. L.
,
P. M.
Fratamico
,
C.
Sommers
,
D. R.
DeMarco
,
S.
Varkey
,
K.
Rhoden
, and
G.
Tice
.
2014
.
Detection of Shiga toxin-producing Escherichia coli (STEC) O157:H7, O26, O45, O103, O111, O121, and O145, and Salmonella in retail raw ground beef using the DuPont™ BAX® system
.
Front. Cell. Infect. Microbiol
.
4
:
81
.

Author notes

† This publication is contribution no. 15-378-J Kansas Agricultural Experiment Station.