Evaluation of test-kits for the detection of Escherichia coli O 157 in raw meats and cattle faeces

Escherichia coli O157 detection limits in artificially contaminated beef and cattle faeces samples were determined using Dynabeads anti E. coli O157 immunomagnetic beads, VIDAS-UP, VIDAS-ICE, and real-time PCR (GeneDisc and LightCycler) systems. Dynabeads anti-E. coli O157 immunomagnetic separation (IMS) and the GeneDisc cycler were the most sensitive methods, and could detect an initial 1 CFU in 25g beef samples after 6h of incubation in modified tryptone soya broth with novobiocin (mTSB+n) or buffered peptone water (BPW). The VIDAS-UP method could detect an initial 10 CFU, while VIDAS-ICE and the LightCycler methods could only detect an initial 100 CFU. Higher detection rates were achieved with 18 hour incubations, where an initial 1 CFU in a 25g sample could be detected with all five methods. For cattle faeces enrichments, Dynabeads anti-E. coli O157 IMS could detect an initial 1 CFU after a 6 h incubation in mTSB+n, while the VIDAS-UP and VIDASICE methods could detect an initial 10 CFU and both PCR methods could only detect an initial 100 CFU. Detection rates were lower in BPW, compared to mTSB+n, with thresholds of 100 CFU for VIDAS-ICE, VIDAS-UP and GeneDisc methods, and >100 CFU for the LightCycler method.


Introduction
Escherichia coli was for many years considered to be a non-pathogenic inhabitant of human and animal intestinal tracts, and was only recognized as a cause of infantile gastroenteritis in the late 1940s (Smith & Fratamico, 2005;Fratamico, 2006).Pathogenic strains of Escherichia coli have since emerged as a major cause of documented outbreaks and sporadic cases of diarrhoea world-wide, and are reported by the Cen-ters for Disease Control and Prevention to be one of the three most common food-borne infections (CDC, 2005(CDC, , 2008)).Strains of E. coli inducing mild to serious gastrointestinal diseases in humans have been classified into six major categories; enterohaemorrhagic (EHEC), enterotoxigenic (ETEC), enteroinvasive (EIEC), enteroaggregative (EAEC), enteropathogenic (EPEC), and diffusely adherent (DAEC) E. coli.However, enterohaemorrhagic E. coli are the most significant group of diarrhoeagenic E. coli, based on frequency and severity of illness.The group includes O157:H7, O26, O111 and sorbitol-fermenting O157:NM serotypes associated with cases of hemorrhagic colitis (HC) and haemolytic uremic syndrome (HUS).All EHEC strains produce Shiga-like toxins (Stx ), also known as Verotoxins (VT), but require the presence of additional virulence markers to be fully pathogenic, such as the eae chromosomal gene associated with attachment and effacement, plasmid encoded enterohemolysin (ehx ), and an enterocyte effacement pathogenicity island (Buchanan & Doyle, 1997;Lindstedt, Heir, Gjernes, Vardund, & Kapperud, 2003).E. coli O157:H7 is the serotype accounting for the greatest proportion of EHEC disease cases, and is responsible for more than 73 000 cases of illness and 61 deaths each year in the U.S.A, in comparison to the 37 000 cases and 30 deaths each year attributed to non-O157 EHEC serotypes (CDC, 2008).While undercooked ground beef has been the food most often associated with E. coli O157 (Rangel, Sparling, Crowe, Griffin, & Swerdlow, 2005;Ateba & Bezuidenhout, 2008), outbreaks have also been linked to other food vehicles such as apple cider, mayonnaise, pea salad, cantaloupe, lettuce, hard salami, raw milk, goat cheese, and alfalfa and radish sprouts (Buchanan & Doyle, 1997;Meng, Doyle, Zhao, & Zhao, 2007;Stirling, Mc Cartney, Ahmed, & Cowden, 2007;Fonseca, Fallon, Sanchez, & Nolte, 2011).Fruit and vegetables implicated in outbreaks have in most cases been contaminated with cattle or other ruminant manure during harvesting and processing.Conventional methods for the detection of E. coli O157 commonly use enrichment, followed by isolation on Sorbitol MacConkey Agar with cefixime and potassium tellurite (CT-SMAC).Most E. coli O157 strains do not ferment sorbitol and colonies are colourless on Sorbitol Mac-Conkey Agar, distinguishing them from sorbitolfermenting non-O157 E. coli strains (Fratamico, 2006).However, traditional isolation methods are time-consuming, and competing flora may obscure O157 colonies thereby giving rise to false negative results.Selective capture with E. coli O157 specific magnetic beads significantly reduces background flora, and has been incorpo-rated in isolation methods for improved recovery rates and the reduction of false-negatives on selective agars (International Organization for Standardization, 2001;LeJeune, Hancock, & Besser, 2006).Isolation rates may be further improved by the use of chromogenic agars that can detect sorbitol-fermenting E. coli O157 strains.In recent years, a number of companies have developed methods for detection that are specific, faster and more sensitive than traditional culture methods.Enzyme-linked immunosorbent fluorescent assays (VIDAS-UP), immunoconcentration (VIDAS-ICE), and real-time PCR (GeneDisc, LightCycler) are examples of rapid methods that are commercially available for the detection of E. coli O157.Investigation of foods implicated in disease outbreaks indicate that the infectious dose for E. coli O157:H7 is lower than 50 cells (Meng et al., 2007).Since E. coli O157 may be present in food and environmental samples in small numbers (Reinders, Barna, Lipman, & Bijker, 2002) and the infectious dose is low, sensitive and rapid detection methods are required to ensure a safe supply of foods.The objectives of this study were therefore to determine the relative sensitivity and specificity of newly introduced E. coli O157 detection methods, and to assess the applicability of each detection method for routine laboratory use based on labour intensiveness, ease of use and interpretation, cost-effectiveness, and length of time required for results.1) were prepared, and cell counts determined on CT-SMAC agar plates prepared from Sorbitol MacConkey agar (Oxoid, Basingstoke, UK) supplemented with 0.05 mg L −1 cefixime and 2.5 mg L −1 tellurite (ITK Diagnostics BV, Uithoorn, Netherlands) .Immunomagnetic separation was carried out in duplicate on 0.1 mL of 10 −6 dilutions of the overnight cultures mixed with 0.9 mL of 6 h or 18 hour beef or cattle faeces enrichment.
For the IMS procedure (Invitrogen-Dynal, 2007), Dynabeads O157 suspensions (20 µL) were dispensed into 1.5 mL microcentrifuge tubes and 1 mL of the test sample added before incubation for 10 minutes in a Dynal MX4 sample mixer (Invitrogen-Dynal, Oslo, Norway).The beadbacteria complexes were separated in a magnetic separator rack (Dynal MPC-S) for three minutes, washed three times in 1 mL phosphate buffered saline with 0.05% Tween-20 (PBS-T) and resuspended in 100 µL PBS-T.Re-suspended samples were then spread on CT-SMAC agar plates, and typical colourless colonies after an 18-24 hour incubation at 37 O157 by a latex agglutination test for the O157 antigen (Oxoid, Basingstoke, UK).Recovery rates were calculated as a percentage of the initial inoculum.

Detection limits of E. coli O157 isolation kits
For each of the three ground beef and three cattle faeces samples, 25g were transferred to 225mL modified Tryptone Soya Broth (mTSB+n) or Buffered Peptone Water (BPW) and inoculated with 0.1 mL of either 10 −8 , 10 −7 or 10 −6 dilution of an overnight brain heart infusion broth culture of E. coli O157 (strain NCTC 12079) to give initial cellular counts of 1, 10 and 100 cells, respectively.Inoculated samples were incubated at 41.5 • C for 6h or the ISO recommended 18h, and the level and presence of E. coli O157 determined using immunomagnetic separation (Dynabeads anti-E.coli O157), VIDAS-ICE , VIDAS-UP , GeneDisc O157, and the LightCycler.Uninoculated samples served as negative controls for all detection methods.

Immunomagnetic separation
Dynabeads O157 suspensions (20 µL) were dispensed into 1.5 mL microcentrifuge tubes and 1 mL of 6 hour and 18 hour enrichment test samples added prior to immunomagentic separation as described in 2.3.Re-suspended samples (100µL) were divided into two equal portions and spread on CT-SMAC plates and on CT-CHROM agar O157 (Oxoid, Basingstoke, UK) supplemented with 0.025 mg L −1 cefixime and 1.25 mg L −1 tellurite.Typical colourless colonies on CT-SMAC, and pink colonies on 1 ⁄2 CT-CHROM after an 18-24 hour incubation at 37 • C were confirmed as E. coli O157 by a latex agglutination test for the O157 antigen (Oxoid, Basingstoke, U.K).

VIDAS -ICE
VIDAS immuno-concentration E. coli O157 (ICE) is a fully automated method for the selective concentration of E. coli O157 from enrichment broth, and consists of two components: IJFS October 2012 Volume 1 pages 126-134 a solid phase receptacle whose interior is coated with anti-E.coli O157 antibodies, and a reagent strip which contains all the wash and release solutions (BioMerieux, Inc. 2008).The enrichment broth (0.5 mL) was dispensed into the reagent strip and the sample cycled in and out of the solid phase receptacle in the VIDAS Immunoassay Analyzer (BioMerieux, Marcy-l'Etoile, France) in a 40 minute procedure.E. coli O157 cells present in the broth were bound to the anti-E.coli O157 antibodies coating the interior of the solid phase receptacle, and released in a final enzymatic step into one of the strip wells.The immunoconcentrated samples were then streaked on selective agar plates using sterile cotton swabs, before inspection for typical colourless colonies on CT-SMAC agar and pink colonies on 1 ⁄2 CT-CHROM agar after an 18-24 hour incubation period at 37 • C. Typical colonies were confirmed as E. coli O157 by a latex agglutination test for the O157 antigen.

VIDAS-UP O157
Enriched samples (1 mL) were transferred to 1.5 mL micro-centrifuge tubes and heated for 5 minutes at 95-96 • C. Cooled samples (0.5 mL) were then dispensed into VIDAS-UP reagent strips and the presence and level of E. coli O157 determined by automated assay in the VIDAS Immunoassay Analyzer (BioMerieux, Marcyl'Etoile, France).VIDAS-UP, unlike the VIDAS-ICE system, does not require plating on selective agar for the screening of E. coli O157, and is based on an enzyme-linked fluorescent assay.The interior of the solid phase receptacle is coated with recombinant phage tail fibre protein for the capture of E. coli O157, and the fluorescence of the final product measured at 450 nm at the end of the 50 minute assay (BioMerieux, Inc. 2009).

GeneDisc E. coli O157
A GeneSystems (Bruz, France) Food Extraction Pack was used for extracting DNA from beef and IJFS October 2012 Volume 1 pages 126-134 cattle faeces E. coli O157 enrichments.Enrichment broth samples (1 mL) were centrifuged at room temperature for 5 minutes at 500g to remove any cell fragments, before transferring 900 µL of supernatant to lysis tubes for a further 5 minute centrifugation at 10 000 g. Bacterial pellets were then re-suspended in 200 µL of dilution buffer supplied in the extraction pack and lysed by incubation for 10 minutes at 100 • C. Faecal sample extracts were further diluted ten-fold in sterile deionized water before analysis, to reduce PCR inhibition caused by bilirubin and bile salts in faeces (Wilson, 2000).For the PCR analysis, 37 µL of a Master Mix supplied with the GeneDisc pack was mixed with 37 µL of DNA sample in 200 µL micro-centrifuge tubes, and 72 µL transferred to each reservoir in the GeneDisc before analysis in the GeneDisc Cycler for presence of ehx, Stx 1, Stx 2, O157 and eae genes in a 55 minute assay (GeneSystems, 2009b(GeneSystems, , 2009a)).

LightCycler
The LightCycler protocol was designed for a 20 µL standard reaction (Roche Diagnostics, 2006).Volumes in the PCR mix for one reaction were: 10.5 µL PCR-grade water, 1 µL O157 forward primer, 1 µL O157 reverse primer, 1 µL O157 probe, 4 µL of Mastermix (Roche Diagnostics, Mannheim, Germany), and 2.5 µL of DNA sample extracted as outlined in Section 2.4.4.Each reaction mix was transferred to a capillary tube, loaded in a rotor, and the PCR analysis carried out in a Roche LightCycler 2.0 system in a 60 minute assay.

Specificity of VIDAS-UP O157 and GeneDisc O157
The specificity of VIDAS-UP O157 and the GeneDisc O157 was determined by analysis of identification results for 17 non-sorbitolfermenting E. coli O157 strains, 22 sorbitolfermenting E. coli O157 strains and 9 non-O157 strains.Overnight Brain Heart Infusion cultures (1 mL) were transferred to 1.5 mL microcentrifuge tubes and the presence of E. coli O157 determined by VIDAS-UP analysis (Section 2.4.3) and GeneDisc analysis (Section 2.4.4).
Strains were confirmed as E. coli O157 by latex agglutination tests for the O157 antigen.

IMS recovery rates in beef and cattle faeces enrichments
In six hour beef enrichments, mean recovery rates of E. coli O157 with Dyna beads anti-E.coli O157 were 79.6 ± 19.1% in mTSB+n and 54.6 ± 19.5% in BPW (Table 1) compared to mean recovery rates in 18h beef enrichments of 46.2 ± 16.0, and 17.8 ± 10.3%, in mTSB+n and BPW, respectively.Similar trends were noted in cattle faeces enrichments, where recovery rates were 59.4 ± 17.9, 39.1 ± 20.1, 46.5 ± 11.1, and 8.0 ± 4.5% in 6h mTSB+n, 6h BPW, 18h mTSB+n and 18 h BPW enrichments, respectively.Recovery rates were significantly lower in BPW than mTSB+n in 18h enriched beef and cattle faeces samples, and also lower for 18h enrichments compared to 6 h enrichments due to overgrowth by background flora on the CT-SMAC plates.It was also noted that recovery rates of inoculum in cattle faeces were significantly lower than in beef samples, presumably due to the higher level of background flora in faecal samples.

Relative sensitivity of E. coli O157 detection methods
Dynabeads and the GeneDisc were the most sensitive detection method for E. coli O157 and could detect an initial 1 CFU in 25 g beef samples after a 6 hour enrichment in both modified Tryptone Soya Broth (mTSB+n) and Buffered Peptone Water (Table 2).VIDAS-UP could detect an initial 10 CFU, while VIDAS-ICE and the LightCycler were less sensitive and could only detect an initial 100 CFU in 6 hour enrichments.In 6 hour cattle faeces enrichments, Dynabeads were the most sensitive isolation method and could detect an initial 1 CFU in two of the three mTSBn samples while VIDAS-ICE and VI-DAS UP could detect an initial 10 CFU, and the GeneDisc and LightCycler could only detect an initial 100 CFU (Table 3).Detection thresholds were higher in buffered peptone water, where an IJFS October 2012 Volume 1 pages 126-134    Heijnen and Medema (2006).

Specificity of VIDAS-UP E. coli O157 and GeneDisc O157
Both VIDAS-UP E. coli O157 and the Gene Disc O157 correctly identified 17 non-sorbitolfermenting and 22 sorbitol-fermenting E. coli O157 strains with a 0% false negative rate, and also correctly identified 9 non-E.coli strains.VIDAS-UP and the GeneDisc O157 are therefore highly specific for E. coli O157 and can be used as rapid and reliable pre-screening tests for the presence of E. coli O157.

Relative cost and applicability of E. coli O157 detection methods
Immunomagnetic separation was the least expensive of the five detection methods due to the lower cost of test equipment.However, IMS and the LightCycler methods were the most labour intensive, while the GeneDisc, VIDAS-UP, and VIDAS-ICE were automated and therefore less laborious and required less manipulation.In addition, the Gene Disc, VIDAS-UP and Light-Cycler were the most rapid, requiring only between 2-5 hours for preliminary test results after a 6 hour pre-enrichment, while IMS and VIDAS ICE required at least 26 hours before presumptive colony identification.It should however be noted that although VIDAS -UP and GeneDisc methods are rapid and convenient methods for E. coli O157 detection, the ISO procedure requires confirmation of presumptive positive samples by identification of typical colonies on selective agar.Thus, VIDAS -UP, GeneDisc, and LightCycler methods may serve only as rapid screening methods for the absence or presence of E. coli O157.

Conclusions
The GeneDisc and VIDAS-UP methods have a high sensitivity and may be used as convenient, reliable and rapid screening methods for E. coli O157 presence.VIDAS-ICE in addition to convenience, has the added advantage of colony isolation on selective agar for confirmatory tests, as required by ISO test protocols.Although CT-SMAC Agar is the ISO (International Organization for Standardization) recommended isolation medium, 1 ⁄2 CT-CHROM agar is more selective, resulting in a significant reduction of background flora, and may be used as a second isolation medium in combination with a mTSB+n enrichment to improve E. coli detection rates.

Table 1 :
Recovery rates for E. coli O157 with IMS in beef and cattle faeces enrichments.

Table 2 :
Relative sensitivity of E. coli O157 detection methods in spiked beef samples.

Table 3 :
Relative sensitivity of E. coli O157 detection methods in cattle faeces enrichments.