PCR-Based DNA Amplification and Presumptive Detection of Escherichia coli O157:H7 with an Internal Fluorogenic Probe and the 5′ Nuclease (TaqMan) Assay

Abstract

Presumptive identification of Escherichia coli O157:H7 is possible in an individual, nonmultiplexed PCR if the reaction targets the enterohemorrhagic E. coli (EHEC) eaeA gene. In this report, we describe the development and evaluation of the sensitivity and specificity of a PCR-based 5′ nuclease assay for presumptively detecting E. coli O157:H7 DNA. The specificity of the eaeA-based 5′ nuclease assay system was sufficient to correctly identify all E. coli O157:H7 strains evaluated, mirroring the previously described specificity of the PCR primers. The SZ-primed, eaeA-targeted 5′ nuclease detection assay was capable of rapid, semiautomated, presumptive detection of E. coli O157:H7 when ≥10³ CFU/ml was present in modified tryptic soy broth (mTSB) or modified E. coli broth and when ≥10⁴ CFU/ml was present in ground beef-mTSB mixtures. Incorporating an immunomagnetic separation (IMS) step, followed by a secondary enrichment culturing step and DNA recovery with a QIAamp tissue kit (Qiagen), improved the detection threshold to ≥10² CFU/ml. Surprisingly, immediately after IMS, the sensitivity of culturing on sorbitol MacConkey agar containing cefeximine and tellurite (CT-SMAC) was such that identifiable colonies were demonstrated only when ≥10⁴ CFU/ml was present in the sample. Several factors that might be involved in creating these false-negative CT-SMAC culture results are discussed. The SZ-primed, eaeA-targeted 5′ nuclease detection system demonstrated that it can be integrated readily into standard culturing procedures and that the assay can be useful as a rapid, automatable process for the presumptive identification of E. coli O157:H7 in ground beef and potentially in other food and environmental samples.

Enterohemorrhagic Escherichia coli O157:H7 is an important human pathogen that is predominantly associated with hemorrhagic colitis and the more severe complications of hemolytic uremic syndrome. Although human-to-human transmission of E. coli O157:H7 has been demonstrated (25), most infections have been associated with the consumption of contaminated ground beef, milk, water, produce, and apple juice products that have been improperly handled, stored, or cooked (1, 5, 7, 15, 18, 19, 25, 43). The primary reservoir is believed to be cattle (24, 27). However, a clear understanding of the farm ecology of E. coli O157:H7 is lacking, partly because of the low detected prevalence in individual cattle and herds (20, 40, 55) and the low infectious dosage required for human infections (25, 53).

Established methods for recovering and identifying E. coli O157:H7 from foods and feces have been hindered by the inability to specifically and rapidly detect small numbers of organisms from complex matrices and background microflora. Although the inclusion of preenrichment incubations and immunomagnetic separation (IMS) (6, 9, 11, 16, 22, 31, 44, 48) and additional selective subculturing or secondary enrichment incubations (12, 14, 30) have been reported to increase the detection rate of E. coli O157:H7 from foods and fecal specimens, these methods are dependent on isolating individual colonies from selective and/or indicator media and then characterizing them in immunological and biochemical/fermentation reactions. Immunological assays are used to determine if the O157 somatic and H7 flagellar antigens are present, while the biochemical/fermentation reactions determine in classical taxonomic fashion the genus and species of the isolate. Combined with the initial replication steps in the isolation process, the current E. coli O157:H7 identification process takes 5 or more days to complete. This adds considerably to the costs required to determine whether a sample contains E. coli O157:H7 and is a limiting factor in doing more E. coli O157:H7 tests.

Rapid methods for identifying E. coli O157:H7 in foods or fecal specimens have been directed at immunological or genetic targets. Antigenic targets have included the E. coli somatic (O157) or flagellar (H7) antigens (21, 50), two low-molecular-weight antigens (30, 45), and the virulence-associated Shiga-like toxin (SLT) types I and II (3, 17, 33). However, these assays are occasionally unable to distinguish certain other E. coli strains from E. coli O157:H7 strains (30, 49) and/or toxigenic from nontoxigenic E. coli O157 strains (46).

PCR-based detection procedures have been used to identify E. coli O157:H7 and have targeted the sltI and sltII genes (32, 47, 54), the enterohemorrhagic E. coli (EHEC) uidA gene (10), and a portion of a 60-MDa plasmid (23). Because similar genes are present in some nonpathogenic E. coli and in other bacteria, individual PCRs that target these genes are unable to confirm the identity of an isolate as E. coli O157:H7. Identification by PCR requires that multiple genes be targeted in separate PCRs on the DNA from a suspect organism or that the DNA from that organism be subjected to a multiplex PCR that targets the multiple genes simultaneously (10, 23, 42, 51).

Presumptive identification of E. coli O157:H7 is possible in an individual, nonmultiplexed PCR if the reaction targets the EHEC eaeA gene. In separate studies targeting two different regions of the eaeA gene, every E. coli O157:H7 reference strain evaluated demonstrated the predicted PCR product. Louie et al. (36) targeted the 3′ end of the EHEC eaeA gene, whereas Meng et al. (41) amplified a 633-bp product upstream of the 5′ end of the EHEC eaeA gene by using PCR primers SZ-I and SZ-II. Both reactions were limited as confirmatory PCRs for identifying E. coli O157:H7, because similar PCR products were evident with some E. coli O157:NM strains and some enteropathogenic E. coli O55:H7 and O55:NM strains (36, 41).

Although PCR can amplify DNA molecules thousands-fold, the specifically amplified product must be detected in order to prove its presence; a variety of methods have been developed for this purpose. The most commonly used research technique, gel electrophoresis, does not show the specificity of the PCR and lacks sensitivity. Southern blots or dot blot hybridizations with probes will demonstrate the specificity of the PCR, but they require multistep processing and add considerable time and expense to the detection process. Neither of these PCR detection processes is conducive to rapid, high-throughput, automated PCR detection schemes.

Recently, 5′ nuclease assays (TaqMan; PE Applied Biosystems, Foster City, Calif.) that allow the automated PCR amplification, detection, and analysis of Salmonella spp. (13, 39), Listeria monocytogenes (2, 4), and SLT genes (28, 54) in various foods have been described. The 5′ nuclease assay exploits the 5′→3′ exonuclease activity of Thermus aquaticus (Taq) DNA polymerase (29, 37) to hydrolyze an internal TaqMan probe labeled with a fluorescent reporter dye and a quencher dye (34). For the intact probe, the quencher dye suppresses the fluorescent emission of the reporter dye because of its spatial proximity on the probe. During PCR, the probe anneals to the target amplicon and is hydrolyzed during extension by the Taq DNA polymerase. The hydrolysis reduces the quenching effect and allows for an increase in emission of the reporter fluorescence. This increase is a direct consequence of a successful PCR, whereas the emission of the quencher dye remains constant irrespective of amplification.

Because development of fluorogenic reporter signals occurs only with a successful PCR, detection of specific DNA sequences can be based on monitoring for an increase in reporter fluorescence following PCR with a fluorometer (ABI Prism sequence detection system; PE Applied Biosystems). Interpretation of the fluorometric data can be automatically read and interpreted by using a 96-sample format and presented as a “yes” or “no” conclusion as to the presence or absence of the DNA within 15 min of the completion of the PCR.

In this report, we describe the development and evaluation of the sensitivity and specificity of a 5′ nuclease assay for amplifying and presumptively detecting E. coli O157:H7 DNA. Evaluation of the 5′ nuclease assay included the comparison of two DNA extraction procedures for recovering E. coli O157:H7 DNA from pure broth cultures; from pure broth cultures immediately before IMS, where the DNA was then retrieved on sorbitol MacConkey agar supplemented with cefeximine and tellurite (CT-SMAC); from pure broth cultures following IMS plus 18-h secondary enrichment, where the DNA was then retrieved on CT-SMAC; and from spiked broth cultures containing ground beef.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

The strains of bacteria that were evaluated are listed in Table 1. The E. coli O157:H7 strain MD 380-94, originally recovered from salami by the U.S. Department of Agriculture (USDA) Food Safety Inspection Service (FSIS) after an associated disease outbreak, was used as the reference strain in all optimization and sensitivity experiments. The E. coli strains were cultured in modified E. coli broth (mEC; BBL, Cockeysville, Md.) containing 20 mg of novobiocin per liter or in tryptic soy broth (TSB; Difco, Detroit, Mich.) modified (mTSB) to contain 100 mg of novobiocin, 10 mg of cefsulodin, 8 mg of vancomycin, and 0.05 mg of cefeximine per liter. All antibiotics were obtained from Sigma Chemical Co., St. Louis, Mo. Cultures were transferred to CT-SMAC plates containing sorbitol MacConkey agar (Difco) supplemented with cefeximine and tellurite (Dynal, Lake Success, N.Y.) and then incubated (18 to 24 h, 37°C).

TABLE 1.

Bacterial strains evaluated and ΔRQ values generated in PCR with the SZ-I and SZ-II primers and the SZI-97 fluorogenic probe

Bacterial strain evaluated	ΔRQa	Interpretationb
E. coli O157:H7
43888c	3.84	Yes
43889c	3.74	Yes
43890c	4.51	Yes
43894c	4.41	Yes
43895c	4.50	Yes
88.1558d	3.98	Yes
MD 380-94e	2.77	Yes
Meat isolatee (KSU label, Eh 7-1)	4.19	Yes
15753 (meat isolate)e	4.16	Yes
45756 (meat isolate)e	3.48	Yes
4731A (meat isolate)e	4.09	Yes
45953 (clinical isolate)f	4.29	Yes
45959 (clinical isolate)f	3.72	Yes
EDL 933 (clinical isolate)f	4.47	Yes
Other E. coli strains
E. coli O157:NM, MF7123Ae	−0.02	No
E. coli O157:NM, MF13180-25e	4.01	Yes
E. coli O111:NM, 91.1030d	0.06	No
E. coli O11:NM, 90-1772d	0.13	No
E. coli O55:H7, Dec 5Ag	4.07	Yes
E. coli O55:H7, Dec 5Bg	4.68	Yes
E. coli 11775c	0.01	No
E. coli RCR93h	0.07	No
E. coli 4221 (canine isolate)i	0.08	No
E. coli 4225 (canine isolate)i	0.35	No
E. coli 17/2j	0.10	No
E. coli 2748/69j	0.23	No
E. coli R555-1j	0.17	No
E. coli 274-1j	0.39	No
E. coli H10407j	−0.02	No
E. coli O26:H1j	0.08	No
E. coli O5:NMj	−0.02	No
E. coli O6:H31j	−0.03	No
E. coli O26:H11j	−0.06	No
E. coli O103:H2j	−0.05	No
E. coli O111:NMj	−0.02	No
E. coli 862 (porcine isolate)i	0.03	No
E. coli (O157 +; H7 −; sorbitol +; MUG +) 35Ae	0.24	No
E. coli (O157 −; H7 +) WS-41f	0.07	No
Other bacteria
H. alvei 13337c	−0.08	No
C. freundii CL 787B-75k	0.09	No
C. freundii CL 350B-77k	0.01	No

^aDetermined with ∼150 ng of template DNA from each strain in PCR conditions with 40 nM SZI-97 probe and 2 mM MgCl₂ in two-step PCR in MOT with optical caps. 

^bA “yes” score is assigned when the ΔRQ is greater than the ΔRQ threshold value (1.04) calculated at the 99% confidence level. 

^cAmerican Type Culture Collection. 

^dE. coli Reference Center, Pennsylvania State University. 

^eFSIS, USDA, Beltsville, Md. 

^fCenters for Disease Control and Prevention, Atlanta, Ga. 

^gMichael Doyle, University of Georgia, Athens. 

^hStan Bailey, Russell Research Center, Athens, Ga. 

ⁱBradley Fenwick, Kansas State University. 

^jDavid Acheson, Tufts University School of Medicine, Boston, Mass. 

^kM. M. Chengappa, Veterinary Diagnostic Laboratory, Kansas State University. 

Developing fluorogenic probes for E. coli O157:H7.

Fluorogenic probes to the eaeA gene contained within the SZ-primed amplicon (41, 42) (GenBank accession no. U32312) were synthesized as previously described (2). The efficiency of individual probes was determined by PCR using purified DNA templates (QIAamp tissue kit; Qiagen, Inc., Chatsworth, Calif.) from reference strains of E. coli and other bacteria. Approximately 150 ng of DNA (DNA DipStick kit; Invitrogen, Carlsbad, Calif.) from each reference strain of bacteria was PCR amplified by using the E. coli O157:H7-specific primers SZ-I and SZ-II (41, 42). Following the two-step PCR, the fluorescence intensities of the fluorescent reporter dye (6-carboxy-fluorescein; λ_em = 518) and the fluorescent quencher dye (6-carboxytetram-ethylrhodamine; λ_em = 582) were determined for each tube by using a luminescence spectrometer with a 96-tube reader accessory (TaqMan LS-50B PCR detection system; Perkin-Elmer). The degree of hydrolysis was calculated by using the equation ΔRQ = RQ⁺ − RQ⁻ (2), where RQ⁺ = emission of reporter dye/emission of quencher dye and RQ⁻ = emission of reporter dye (no DNA template)/emission of quencher dye (no DNA template).

The ΔRQ threshold for determining the presence (“yes”) of E. coli O157:H7 DNA in individual MicroAmp optical tubes and caps (MOT; Perkin-Elmer) was based on a 99% confidence interval based on the standard deviation of RQ⁻ values from no-template controls from >20 plates (three no-template controls per plate). A “yes” interpretation of a ΔRQ value obtained by using MicroAmp optical 96-well reaction plates with MicroAmp caps (MORP) (Perkin-Elmer) was determined at 99% confidence intervals, by using the standard deviation of RQ⁻ values of three no-template controls per plate. Data were collected and analyzed with the Fluorescence Data Manager (Perkin-Elmer) and Excel spreadsheet (Microsoft Corporation, Redmond, Wash.) programs on a personal computer.

PCR conditions.

The PCR amplification conditions were modified from those described by Meng et al. (41) and included fluorogenic probes. Briefly, 5 μl of sample containing the DNA template to be evaluated was added to 45 μl of PCR master mix (5 μl of 1× PCR buffer II [Perkin-Elmer], 1.5 to 4.0 mM MgCl₂, 200 nM each primer [SZ-I and SZ-II], 200 μM deoxynucleoside triphosphate, 0.025 U of AmpliTaq DNA polymerase [Perkin-Elmer], 25 to 50 nM fluorogenic probe, 26 μl of water) in 200-μl capacity MOT or in individual wells of a MORP. Each set of reactions included a single tube (well) of TE buffer (10 mM Tris-HCl [pH 8.0], 1 mM EDTA) for the autozero control and triplicate tubes (wells) that were no-template controls (containing no E. coli O157:H7 DNA templates). Each assay also included DNA collected from other bacterial species: Salmonella choleraesuis, Hafnia alvei, and Citrobacter freundii, and NovaBlue, a general-purpose cloning host with the genotype endA1 hsdR17(r_K⁻ m_K⁺) supE44 thi-1 gyrA96 relA1 lac [F′ proAB lacI^qZΔM15 Tn10(Tet^r)] recA1 (Novagen, Madison, Wis.). All non-E. coli organisms were cultivated in TSB (37°C, overnight). The PCR had an initial denaturization step (95°C, 5 min) followed by 35 amplification cycles of a two-step PCR (95°C, 20 s; 60°C, 60 s), with a final extension (72°C, 10 min or longer) on a thermocycler (GeneAmp PCR System 9600; Perkin-Elmer).

DNA recovery procedures.

To evaluate the sensitivity of an E. coli O157:H7 5′ nuclease assay system, different DNA recovery procedures were evaluated with different types of samples.

(i) QIAamp tissue kit.

The ability of the QIAamp tissue kit to recover E. coli O157:H7 DNA from pure broth cultures was evaluated according to the manufacturer’s instructions. Tenfold dilutions were made of E. coli O157:H7 in mTSB. A 1.0-ml aliquot from each dilution was removed and pelleted by centrifugation (16,000 × g, 10 min), and the supernatant was carefully removed and discarded. The pellets were centrifuged again (16,000 × g, 1 min), and each pellet was resuspended in 180 μl of Buffer ATL (Qiagen) and 20 μl of proteinase K stock solution (Qiagen). Tubes were vortexed and incubated in a 55°C water bath for 30 min or until the sample was lysed. Twenty microliters of RNase A (20 mg/ml) was added to each tube, and the tubes were vortexed (5 s) and incubated (room temperature, 2 min). Then 200 μl of Buffer AL (Qiagen) was added to each tube, and tubes were vortexed (5 s) and incubated (70°C in a dry heating block, 10 min). Next, 210 μl of 100% ethanol was added to each tube, and the contents were mixed by vortexing and transferred to a QIAamp spin column that was placed in a 2-ml collection tube.

The spin column and collection tube were centrifuged (5,200 × g, 1 min). The spin column then was placed on a new collection tube, and 500 μl of Buffer AW (Qiagen) was added to the column. The spin column and a new collection tube then were centrifuged (5,200 × g, 1 min), the spin column was removed and placed into a new 2-ml collection tube, and 500 μl of Buffer AW (Qiagen) was added. The spin column and a new collection tube were centrifuged (5,200 × g for 1 min, followed by 16,000 × g for 2 min), the spin columns then were placed onto marked 1.5-ml microcentrifuge tubes, and 200 μl of 70°C buffer AE (Qiagen) was added. The liquid in the collection tubes was precipitated with ethanol, placed in QIAamp spin columns, and processed according to the manufacturer’s instructions. The DNA was eluted from the QIAamp spin columns by adding 200 μl of Buffer AE preheated to 70°C, incubating the columns at room temperature for 1 min, and then centrifuging the columns (6,000 × g for 1 min). The resulting elution samples were stored at −20°C or used in 5′ nuclease assays.

(ii) DNA-ER procedure.

One of the DNA recovery procedures evaluated utilized the chelating properties of chelex resin as previously described for recovering Salmonella DNA from foods (DNA extraction reagent [DNA-ER] method; TaqMan Salmonella PCR amplification detection kit; Perkin-Elmer). Briefly, 10-fold serial dilutions of mTSB and mEC broth cultures were made. Aliquots of 500 and 1,000 μl from each dilution were centrifuged (16,000 × g, 3 min), the supernatant was decanted carefully, and the pellet was resuspended in 200 μl of thoroughly mixed DNA-ER solution (DNA Extraction Reagent, product no. N808-0087; Perkin-Elmer). The tubes were vortexed for 5 to 10 s or as long as required to resuspend the pellet. The tubes then were incubated in a water bath (56°C, 30 min), floated in boiling water (7 min), and chilled on ice (5 min). The tubes then were centrifuged (16,000 × g, 3 min), and the supernatants were carefully transferred to new microcentrifuge tubes. A 5-μl aliquot of the supernatant served as the template for each PCR amplification in the 5′ nuclease assay.

Sample types.

To evaluate the sensitivity of an E. coli O157:H7 5′ nuclease assay, different sample types were evaluated.

(i) Pure cultures.

Pure cultures of E. coli O157:H7 were grown in mTSB or mEC (35°C, 12 to 15 h). Tenfold serial dilutions of the cultures were made with broth as the diluent, and aliquots were taken for E. coli O157:H7 enumeration (CFU/milliliter) on CT-SMAC and for immediate DNA recovery.

(ii) IMS from pure cultures.

E. coli O157:H7 was detected and enumerated following recovery by IMS from mTSB, by using the beads as inoculum for direct plating onto CT-SMAC, by using the beads as inoculum for secondary enrichment and subsequent plating onto CT-SMAC, or by recovering DNA from each of the previous steps and subjecting it to E. coli O157:H7 5′ nuclease reactions.

Specifically, 10-fold dilutions were made of pure cultures of E. coli O157:H7 while they were in logarithmic growth (between 4 and 8 h of culture). Aliquots of broth (100 μl) were plated on CT-SMAC plates in triplicate (incubated at 37°C for 18 h). Similarly, 0.5- and 1.0-ml aliquots were collected and subjected to the DNA-ER and QIAamp tissue kit DNA extraction methods. Simultaneously, 1.0-ml aliquots were collected from each dilution and transferred to microcentrifuge tubes containing 20 μl of anti-E. coli O157 immunomagnetic beads (Dynabeads Anti-E. coli O157; Dynal). After being separated by using a stationary magnet and washed according to the manufacturer’s instructions, the beads were placed in 200 μl of phosphate-buffered saline. These bead solutions then were equally divided: 100 μl was plated on CT-SMAC (incubated at 37°C for 18 h), and 100 μl was added to a tube containing 9.0 ml of mTSB for secondary enrichment culturing (37°C, 18 h). Following secondary enrichment incubation, 100-μl aliquots of each dilution were plated on CT-SMAC (37°C, 18 h), and 0.5- and 1.0-ml aliquots were collected and subjected to both the DNA-ER and QIAamp tissue kit DNA recovery procedures. DNA also was recovered from 0.5- and 1.0-ml aliquots immediately prior to IMS and following 18-h secondary enrichment of IMS beads and were then subjected to 5′ nuclease assay. The resulting ΔRQ values obtained from each dilution were compared with the CT-SMAC culture results. All 5′ nuclease assay reactions in the IMS study were conducted in 96-well MORP.

(iii) Ground beef samples spiked with E. coli O157:H7.

The ability to detect E. coli O157:H7 in ground beef was evaluated by spiking ground beef samples with different numbers of bacteria and then recovering DNA from each spiked sample and subjecting it to the 5′ nuclease detection assay. Ground beef (containing 20% fat) was obtained from the Kansas State University Meat Processing Laboratory on three different occasions and confirmed to be culture negative for E. coli O157:H7 before inclusion in the study. E. coli O157:H7 strain MD 380-94 was collected from mTSB cultures while it was in logarithmic growth (4 to 8 h), and 10-fold dilutions were made in mTSB. One milliliter of each dilution was added to tubes containing 9.0 ml of ground beef-mTSB mixture (10 g of ground beef, 90 ml of mTSB) that had been previously incubated (6 h, 37°C). Aliquots (0.5 ml) were immediately removed from each dilution for DNA recovery by the DNA-ER and QIAamp tissue kit procedures and then subjected to the 5′ nuclease assay. All 5′ nuclease assay reactions containing ground beef were conducted in 96-well MORP.

RESULTS

E. coli O157:H7 fluorogenic TaqMan probe design, 5′ nuclease assay performance, and specificity.

The adaptation of a 5′ nuclease fluorogenic detection process to a PCR specific for E. coli O157:H7 required the evaluation of DNA sequences upstream of the eaeA gene. Recommended guidelines for designing fluorogenic probes for 5′ nuclease assays were followed (35), and four fluorogenic probes (SZI-97, SZI-107, SZII-194, SZII-200) targeting both complementary DNA strands of the SZ-primed amplicon were constructed by using standard procedures (Table 2). Unless otherwise noted, all specificity evaluations were conducted in MOT.

TABLE 2.

Primers and fluorogenic probes

Probe or primera	Sequence (5′→3′)	Denaturation temp (°C)b	Location within eaeA genec
Primers
SZ-I	CCATAATCATTTTATTTAGAGGGA	61.7	28–51
SZ-II	GAGAAATAAATTATATTAATAGATCGGA	61.6	632–659
Probesd
SZI-97	TTGCTGCAGGATGGGCAACTCTTGAp	78	97–121
SZI-107	ATGGGCAACTCTTGAGCTTCTGTAAp	70.3	107–131
SZII-194	ATTGTCGCTTGAACTGATTTCCTCp	74.3	582–605
SZII-200	TAATGTTTATTGTCGCTTGAACTGATp	66.4	588–613

^aPrimer sequences as previously described by Meng et al. (41). 

^bCalculated by nearest-neighbor algorithm. 

^cE. coli O157:H7 intimin (eaeA) gene as described in GenBank accession no. U32312. 

^dPhosphoramidites added to the 5′ end (fluorescent reporter dye [6-carboxy-fluorescein; λ_em = 518] and fluorescent quencher dye [6-carboxytetram-ethylrhodamine; λ_em = 582]). The 3′ end also contained a phosphate cap (p). 

The efficiency of the probes was evaluated in PCRs with DNA recovered from reference strains of E. coli O157:H7, other E. coli strains, and other bacteria. In preliminary evaluations, the SZI-97 and SZI-107 fluorogenic probes generated higher ΔRQ values (Table 3). Based on these findings, the SZI-97 probe was selected as the probe for optimization and further evaluation of the 5′ nuclease assay. Monitoring ΔRQ values in PCRs with various concentrations of the SZI-97 probe (25 to 50 nM) identified 35 nM as the optimal probe concentration for all subsequent PCRs (data not shown). Similarly, the MgCl₂ concentration was shown to influence the ΔRQ value, and subsequent PCRs used a final concentration of 4 mM MgCl₂ (Table 4).

TABLE 3.

Evaluation of probes for the E. coli O157:H7 PCR-based fluorogenic 5′ nuclease assay

Bacterial DNA template	ΔRQ value for probea/interpretationb
	SZI-97	SZI-107	SZII-94	SZII-200
NovaBlue competent cells	−0.02/no	0.03/no	0.03/no	−0.12/no
E. coli O157:H7, MD380-94c	1.28/yes	1.21/yes	1.02/yes	1.19/yes
E. coli O157:NM, MF7123Sc	0.08/no	−0.05/no	−0.05/no	−0.08/no
E. coli O157:H7, 88.1588d	2.03/yes	2.24/yes	1.53/yes	1.74/yes
E. coli O111:NM, 91.1030d	0.07/no	−0.03/no	0.03/no	0.00/no
E. coli O157:NM, MF13180-25c	2.51/yes	2.00/yes	1.56/yes	1.75/yes
E. coli O111:NM, 90-1772d	0.04/no	−0.02/no	−0.01/no	−0.11/no
S. choleraesuis 94-00041e	0.04/no	0.03/no	0.12/no	−0.05/no

^aProbes at 40 nM (final concentration) in MOT for two-step PCR using SZ primers and ca. 150 ng of purified template DNA (QIAamp tissue kit) in each reaction. 

^bA “yes” score is assigned when the ΔRQ is greater than the ΔRQ threshold value (1.04) calculated at the 99% confidence level. 

^cFSIS, USDA, Beltsville, Md. 

^dE. coli Reference Center, Pennsylvania State University. 

^eVeterinary Diagnostic Laboratory, Kansas State University. 

TABLE 4.

Effect of MgCl₂ concentration on ΔRQ values with the SZ primers and SZI-97 fluorogenic probe

Bacterial strain	ΔRQ values with [MgCl2] ofa:
	1.5 mM	2.5 mM	4.0 mM
E. coli O157:H7, MD380-94b	1.42	1.64	4.82
E. coli O157:H7, 88,1558c	0.65	1.32	5.58
E. coli O157:NM, MF7123Ad	0.15	0.09	0.6
E. coli O111:NM, 91.1030c	−0.18	−0.02	0.47
E. coli O157:NM, MF13180-25b	1.22	1.46	4.74
E. coli O11:NM, 90-1772d	−0.35	0.17	0.06
NovaBlue competent cells	−0.17	0.0	−0.03
S. choleraesuis 94-00041e	−0.18	0.34	0.0

^aDetermined by using ca. 150 ng of DNA extracted with QIAamp tissue kit in two-step, SZ-primed PCR with a 40 nM concentration of SZI-97 probe in MOT. 

^bFSIS, USDA, Beltsville, Md. 

^cAmerican Type Culture Collection. 

^dE. coli Reference Center, Pennsylvania State University. 

^eVeterinary Diagnostic Laboratory, Kansas State University. 

Similar to the PCR results described by Meng et al. (41), elevated ΔRQ values were demonstrated in all PCRs that used the SZ primers, 35 nM SZI-97 probe, and ∼150 ng of E. coli O157:H7 DNA (Table 1). With a ΔRQ detection threshold level determined at 99% confidence limits, all E. coli O157:H7 strains evaluated in MOT demonstrated ΔRQ values that were greater than the ΔRQ threshold level of 1.04 and therefore resulted in a “yes” conclusion for the presence of E. coli O157:H7 DNA in the sample. The ΔRQ detection threshold level indicative of a “yes” conclusion with MORP at 99% confidence limits was ≥0.34.

In evaluations of the specificity of the SZ primers and the fluorogenic SZI-97 probe for organisms other than E. coli O157:H7, ΔRQ values above the detection threshold were observed with some E. coli O157:NM strains and two enteropathogenic E. coli O55:H7 strains (Table 1). This cross-reactivity was similar to that described by Meng et al. (41, 42) using the SZ primers in single and multiplex PCRs. No elevated ΔRQ values above threshold detection limits were detected for the other reference bacteria, including H. alvei and C. freundii strains.

Sensitivity of the E. coli O157:H7 5′ nuclease detection system.

Tests of the sensitivity of the E. coli O157:H7 fluorogenic 5′ nuclease detection system using DNA recovered with the QIAamp tissue kit from overnight cultures of E. coli O157:H7 grown in either mTSB or mEC and with PCRs conducted and analyzed in MOT resulted in ΔRQ values greater than the detection threshold level when ≥10⁴ CFU/ml was present (Fig. 1). The sensitivity of the 5′ nuclease assay was increased when the PCRs, detection, and analysis were conducted in MORP (Table 5). The minimum ΔRQ levels indicative of a “yes” conclusion with the MORP were recorded when E. coli O157:H7 in logarithmic growth, present at ≥10³ CFU/ml, was evaluated. These detection levels were demonstrated with both the QIAamp tissue extraction kit and the DNA-ER procedure (Table 5).

FIG. 1.

Sensitivity of the fluorogenic 5′ nuclease assay for detecting E. coli O157:H7 in individual MOT (Perkin-Elmer). Tenfold dilutions of E. coli O157:H7 were made in mTSB and mEC in triplicate. One-milliliter aliquots from each dilution were subjected to QIAamp tissue kit DNA recovery, and 5 μl of the recovered DNA solution was PCR amplified with the SZ-I and SZ-II primers in the presence of the SZI-97 fluorogenic probe. Detection and analysis were completed with the ABI Prism sequence detection system. All amplification and detection reactions were completed in MOT. The average ΔRQ values for each dilution were plotted against the average CFU/milliliter as determined by plating each dilution on CT-SMAC. The adjusted ΔRQ threshold value was calculated to be 1.04 (dashed line). Error bars indicate the standard deviation of the mean (n = 6).

TABLE 5.

Comparison of standard culture and the E. coli O157:H7 5′ nuclease assay ΔRQ values for detecting E. coli O157:H7, using two different DNA extraction procedures and different assay conditions

CFU/ml (avg of 3 plates)			DNA-ERd ΔRQe values				QIAampf ΔRQ value
Immediately before IMSa	After IMSb	IMS + 18 hc	Before IMS		IMS + 18 h		Before IMS		IMS + 18 h
			1.0 mlg	0.5 ml	1.0 ml	0.5 ml	1.0 ml	0.5 ml	1.0 ml	0.5 ml
3.3 × 107	1.7 × 106	2.8 × 106	7.79	7.79	4.72	6.07	6.97	7.16	8.07	7.54
1.7 × 106	7.2 × 105	1.8 × 106	7.31	6.90	0.41	1.18	5.98	6.14	7.61	7.48
2.1 × 105	4.8 × 104	9.4 × 103	5.66	5.34	0.32	0.27	4.21	4.86	7.42	6.79
1.6 × 104	2.0 × 103	2.7 × 105	2.99	3.04	6.27	6.23	2.53	2.01	8.30	7.96
1.8 × 103	NDh	1.5 × 103	0.71	1.05	0.04	0.03	0.93	0.64	6.76	6.31
9.8 × 101	ND	2.2 × 101	0.06	0.19	0.00	0.05	0.15	0.15	6.91	6.25
2.0 × 101	ND	ND	−0.01	0.02	−0.02	0.02	0.02	0.09	0.03	−0.03
1.0 × 101	ND	ND	−0.03	0.07	−0.01	0.00	0.03	−0.01	0.05	0.02

^aDetermined by plating 0.1 ml of mTSB with E. coli O157:H7 strain MD 380-94 on CT-SMAC agar plates (incubated at 37°C for 18 h). 

^bDynabead Anti-O157 (Dynal) (20 μl) washed beads resuspended to 100 μl with TE buffer. 

^cFollowing IMS, beads were incubated in 9.0 ml of mTSB (incubated at 37°C for 18 h). 

^dProcedure as described in Salmonella TaqMan kit (Perkin-Elmer). 

^eAll PCRs and analyses were conducted in MORP (Perkin-Elmer). ΔRQ detection threshold for “yes” respondents had ΔRQ of ≥0.34. 

^fQIAamp tissue extraction kit (Qiagen). 

^gVolume of mTSB extracted. 

^hND, not detected. 

Sensitivity of the 5′ nuclease assay to identify E. coli O157:H7 was found to be approximately 10⁴ CFU/ml in ground beef-mTSB mixtures (Fig. 2). The ΔRQ values obtained from various dilutions of ground beef-mTSB mixtures were indicative of a “yes” conclusion (ΔRQ values of ≥0.34) when ≥10⁴ CFU/ml was processed in 0.5-ml aliquots with the QIAamp tissue extraction kit. This sensitivity was obtained without including an IMS step. With the same dilutions, the DNA-ER procedure demonstrated similar “yes” and “no” conclusions but gave consistently lower average ΔRQ values than the QIAamp tissue extraction kit when ≥10⁴ CFU/ml was present (Fig. 2).

FIG. 2.

Sensitivity of the fluorogenic 5′ nuclease assay for detecting E. coli O157:H7 in mTSB containing ground beef. Tenfold dilutions of E. coli O157:H7 were made in mTSB in triplicate, and 1.0 ml was added to a tube containing 9.0 ml of ground beef-mTSB mixture (10 g of ground beef, 90 ml of mTSB; incubated for 6 h at 37°C). Aliquots (0.5 ml) were collected for DNA recovery by using the DNA-ER (□) and QIAamp tissue kit (▿) DNA extraction methods, and 5 μl of the recovered DNA solution was amplified with the SZ-I and SZ-II primers in the presence of the SZI-97 fluorogenic probe. Detection and analysis were completed with the ABI Prism sequence detection system. All amplification and detection reactions were completed in MORP. The average ΔRQ values determined from DNA recovered from both DNA extraction methods were plotted against the average CFU/milliliter determined by plating each ground beef dilution on CT-SMAC. The ΔRQ threshold value at 99% confidence limits was calculated to be 0.34. Error bars indicate the standard deviation from the mean (n = 3).

A difference was noted in the ΔRQ values when samples containing older cultures of E. coli O157:H7 were processed by the two DNA recovery methods. The DNA recovery with the QIAamp tissue extraction kit after IMS–18-h secondary enrichment demonstrated ΔRQ values interpreted as “yes” when the starting inoculum was ≥10² (9.8 × 10¹) CFU/ml (Table 5). This was also the limiting dilution at which E. coli O157:H7 was recovered on CT-SMAC following IMS–18-h secondary enrichment. The ΔRQ values obtained from the same dilutions but with the DNA-ER recovery process gave “yes” conclusions when the inoculum prior to IMS was ≥10⁴ CFU/ml. However, the ΔRQ values obtained by using the DNA-ER procedure were inconsistent and resulted in numerous false “no” interpretations, as determined by plating of IMS–18-h secondarily enriched dilutions on CT-SMAC and recovering organisms from the dilutions when the starting inoculum contained ≥10² CFU/ml. Decreasing the volume of the secondary enrichment medium processed for DNA recovery did not appear to improve the efficiency of the DNA-ER procedure, because interpretations of ΔRQ values were similar whether 1.0 or 0.5 ml was processed (Table 5).

DISCUSSION

The ability to presumptively detect E. coli O157:H7 DNA from diverse kinds of samples by using an automated amplification/detection procedure would be beneficial to food producers, food processors, food safety regulatory agencies, and clinical microbiologists. In this report, we describe the development and evaluation of a PCR-based 5′ nuclease (TaqMan) assay for automatically amplifying and then detecting E. coli O157:H7 DNA. The assay was successful in detecting E. coli O157:H7 DNA, with “yes” interpretations for all reference strains of E. coli O157:H7 evaluated, for pure cultures of E. coli O157:H7 grown in mTSB and mEC, and for ground beef-mTSB mixtures spiked with E. coli O157:H7. Detection sensitivities were improved if IMS and 18-h secondary enrichments were incorporated into the process.

The data indicate that the automated eaeA-based E. coli O157:H7 5′ nuclease detection assay, when integrated with an effective DNA recovery process, is capable of rapid, semiautomated, presumptive detection of E. coli O157:H7 if ≥10³ CFU/ml is present in mTSB or mEC or if ≥10⁴ CFU/ml is present in a ground beef-mTSB mixture. These data also indicate that incorporating preenrichment and secondary enrichment steps and an IMS recovery process could further improve the sensitivity of the detection procedure. Because only 1/40 of a 1.0-ml sample (concentrated to 200 μl following DNA recovery) was actually evaluated in any PCR (5-μl sample/PCR), evaluating a larger fraction of the sample per PCR or further concentrating the DNA in the sample might increase the sensitivity of the procedure. However, both steps would increase costs or add additional manipulations to the procedure.

The strategy for developing an E. coli O157:H7-specific fluorogenic probe was to target an internal region of the SZ-primed amplicon that could detect subtle sequence differences between the eaeA gene of E. coli O157:H7 and similar genes in enteropathogenic E. coli, C. freundii, and H. alvei. Specificity of the fluorogenic probe in this PCR was based on the presumptions (i) that an increase in reporter fluorescence emission from the probe would occur only if the PCR primers annealed specifically to E. coli O157:H7 DNA templates as previously described (41, 42) and, simultaneously, the fluorogenic probe annealed specifically to the same target that contained the eaeA gene of E. coli O157:H7 and (ii) that digestion of the probe would occur only if extension of the new complementary strand of DNA proceeded in a 5′→3′ direction as predicted in a successful PCR. Based on our observations and the conclusions from previous studies using the SZ primers (41, 42), the SZ primers and SZI-97 fluorogenic probe appear to be highly specific for presumptively identifying E. coli O157:H7 DNA. Additional evaluations on a larger reference collection of E. coli O157:H7 will be required to confirm this hypothesis.

By utilizing the SZ-I and SZ-II primers in PCR in the presence of the fluorogenic SZI-97 probe (in MOT), it was possible to presumptively determine the presence of E. coli O157:H7 DNA in broth cultures of mTSB and mEC with similar sensitivities (≥10⁴ CFU/ml). The sensitivity was increased to ≥10³ CFU/ml by using MORP and was comparable to the previously reported sensitivity of the SZ-primed eaeA-based PCR as determined by visually interpreting ethidium bromide-stained agarose gels (41). We attribute the difference in sensitivities between MOT and MORP to improved reproducibility of the optical characteristics of MORP, resulting in more consistent ΔRQ⁻ values, which, in turn, reduced the ΔRQ detection threshold level. Regardless, all E. coli O157:H7 strains evaluated demonstrated ΔRQ values above the detection threshold levels and resulted in “yes” interpretations at 99% confidence levels.

Elevated ΔRQ values above the detection threshold were also observed with some E. coli O157:NM strains and two enteropathogenic E. coli O55:H7 strains. These findings were not unexpected, because similar cross-reactivity has been demonstrated in assays using the SZ primers in single and multiplex PCRs (41, 42). In most circumstances, cross-reactivity in a PCR would be sufficient to limit the usefulness of those primers for detection purposes; however, several points suggest that the SZ-primed, eaeA-based 5′ nuclease detection system would be useful as a rapid, sensitive, semiautomated, presumptive detection process for E. coli O157:H7 in food and environmental samples. First, the eaeA-based 5′ nuclease assay system was able to correctly identify all E. coli O157:H7 strains evaluated and mirrored the previously described specificity of the primers (41, 42). Second, all cross-reactivity was limited to organisms that would be considered human pathogens and undesirable in foods. Specifically, E. coli O157:NM has been increasingly isolated from hemolytic uremic syndrome patients in Europe (8, 26), and the enteropathogenic E. coli O55:H7 is associated with worldwide outbreaks of infantile diarrhea (52). Phylogenetic analyses have also suggested that E. coli O157:H7 may have originated from an E. coli O55:H7 clone (52).

Regardless, definitive confirmation could be completed by standard culture techniques on all presumptive “yes” samples identified by the SZ-primed 5′ nuclease assay. Similarly, all presumptive “yes” samples could be confirmed by PCR amplification/detection approaches targeting the eaeA and the sltI and sltII genes as previously described by Meng et al. (41, 42). The latter approach would eliminate the need to complete biochemical or immunological analyses to correctly confirm the identity of E. coli O157:H7 isolates. Once organisms were identified by the presumptive eaeA-based 5′ nuclease assay, efforts to recover viable organisms from the broth culture could be initiated by using standard techniques. For example, once sorbitol-non-fermenting colonies were identified on CT-SMAC, confirmatory E. coli O157:H7 genetic testing could be initiated on colonies by using fluorogenic 5′ nuclease detection systems that individually targeted the eaeA and the sltI and sltII genes with specific fluorogenic probes (28, 41, 42).

Perhaps the most important attribute of the eaeA-based 5′ nuclease assay for presumptively detecting E. coli O157:H7 is that the entire identification process (disregarding culture incubation times) takes approximately 2.5 h (about 20 min for DNA recovery, <2 h for PCR preparation and thermal cycling, and <15 min for PCR product detection and analysis).

Sample processing is an important component of any DNA-based detection system. To optimize sample preparation, we compared the abilities of two extraction procedures to recover E. coli O157:H7 DNA for the 5′ nuclease detection system. When E. coli O157:H7 cultures in logarithmic growth were evaluated, the DNA-ER and QIAamp tissue kit DNA extraction methods were equally effective in demonstrating “yes” ΔRQ values (read in MORP) if ≥10³ CFU/ml was present. However, when older cultures (IMS plus 18-h secondary enrichment) were evaluated, the efficiency of the DNA-ER procedure, as determined by ΔRQ values, demonstrated differences that could be associated only with the DNA recovery processes. When DNA was recovered with the DNA-ER procedure, ΔRQ values of samples containing older cultures were highly variable. This variability hindered interpretation of the endpoint sensitivity of the assay and, more importantly, resulted in false-negative interpretations when DNA-ER results were compared to CT-SMAC culture results. When the QIAamp tissue kit was evaluated in assays using the same dilutions, the ΔRQ values and “yes” interpretations were identical to the IMS–18-h secondary enrichment culture results on CT-SMAC, indicating a detection capability when ≥10² CFU/ml was present in the original dilution.

The 5′ nuclease detection system was able to detect E. coli O157:H7 when ≥10³ CFU/ml was present in pure culture in mTSB or mEC and yielded results comparable to those that Meng et al. (41) obtained by visually interpreting agarose gels. However, incorporating an IMS step, followed by a secondary enrichment culturing step and QIAamp tissue kit DNA recovery, improved the detection threshold to ≥10² CFU/ml in the original sample. This detection level is closer to the suggested minimum infectious dosage for humans (25, 53).

Surprisingly, the sensitivity of CT-SMAC culturing immediately after IMS was such that identifiable colonies resulted only if ≥10⁴ CFU/ml was present in the original sample. Several factors might be involved in creating these false-negative culture results. First, the detection limit of the IMS procedure using Dynabeads Anti-E. coli O157 beads (Dynal) is approximately 10² organisms/ml of preenriched sample, indicating that E. coli O157:H7 concentrations of <10² CFU/ml could go undetected. Also, the successful isolation of some strains of E. coli O157:H7 could be affected adversely by plating the strains onto CT-SMAC. A recent study indicated that plating unstressed, laboratory-reared E. coli O157 isolates on CT-SMAC adversely affected their isolation by delaying their growth, resulting in false-negative conclusions (38). Surprisingly, we were able to detect organisms in dilutions that originally contained >10² and <10⁴ CFU/ml only when they were subjected to IMS and an additional 18-h secondary enrichment prior to being plated on CT-SMAC. To overcome this delay in growth and improve the recovery for some E. coli O157 strains, others have suggested reducing the levels of antibiotics in preenrichment broths prior to IMS and plating on CT-SMAC (9, 38). Regardless, additional research will be required to address the inhibitory effect of CT-SMAC and the effects that such inhibition would have on accurately detecting E. coli O157:H7 in food and environmental samples.

To demonstrate the applicability of a presumptive E. coli O157:H7 5′ nuclease detection system in food production and processing facilities, throughput capability of the system must be enhanced by integrating automated liquid handling capability into the process, automating the DNA recovery process, and linking all phases of the integrated detection process with a computer-based system that would allow increased retrieval and storage of data. As demonstrated in this study, presumptive pathogen detection systems using 5′ nuclease assay components can be integrated readily into standard culture/detection procedures to reduce the time required to presumptively detect E. coli O157:H7 or other microbial contaminants. Cost and time savings could also be realized by reducing the need for complete bacterial isolation procedures and biochemical/immunological characterizations on every sample, unless suggested by an eaeA-based 5′ nuclease assay presumptive “yes” result.