Abstract
A multiplex real-time polymerase chain reaction (PCR) was developed for the simultaneous detection of genes encoding intimin (eae) and all variants of Shiga toxins 1 and 2 (stx1 and stx2) in diagnostic samples. The uidA gene encoding a β-glucuronidase specific for Escherichia coli and Shigella spp. was included in the multiplex PCR assay as an internal amplification control. The multiplex PCR was tested on 30 E. coli reference strains and 174 diagnostic samples already characterized as harboring stx1, stx2, and eae genes. The multiplex PCR correctly detected the genes in all strains examined. No cross reaction was observed with 68 strains representing other gastrointestinal pathogens, normal gastrointestinal flora, or closely related bacteria, reflecting 100% specificity of the assay. The detection limits of the multiplex PCR were 5 genome equivalents for stx2 and 50 genome equivalents for eae and stx1.
Introduction
Attaching and effacing lesions are induced by both EHEC and EPEC strains. EHEC strains may cause hemorrhagic colitis, a bloody diarrhea that can lead to life-threatening hemolytic–uremic syndrome (Griffin and Tauxe, 1991; Paton and Paton, 1998). Food of animal origin such as meat or unpasteurized milk products, or contaminated vegetables are major vehicles for transmission of EHEC to humans (Karmali, 2004).
EPEC is a major cause of acute and persistent infantile diarrhea, particularly in developing countries (Kaper et al., 2004). EHEC strains possess a variety of different virulence factors. The presence of at least one of the two genes stx1 and stx2 encoding Shiga toxins is used as the common characteristic for molecular identification of EHEC. However, not all Shiga toxin-producing E. coli are indeed associated with human disease and therefore are not classified as EHEC. Among the additional virulence factors present in EHEC, the locus of enterocyte effacement (LEE) pathogenicity island is encountered most frequently (Kaper et al., 2004). LEE encodes a 94-kDa outer membrane protein called intimin (Eae), which mediates the intimate attachment of bacteria to epithelial cells (Jerse et al., 1990). The intimin gene eae is routinely used as marker for LEE-positive EHEC and all EPEC strains (Ogierman et al., 2000).
Various multiplex real-time polymerase chain reaction (PCR) methods for the detection of EHEC and EPEC strains have been reported (Sharma et al., 1999; Bélanger et al., 2002; Ibekwe et al., 2002; Reischl et al., 2002; Jinneman et al., 2003; López-Saucedo et al., 2003; Møller Nielsen and Andersen, 2003; Sharma and Dean-Nystrom, 2003; Perelle et al., 2004; Sekse et al., 2005; Yoshitomi et al., 2006; Brandal et al., 2007; Schuurman et al., 2007; Yang et al., 2007; Guion et al., 2008; Auvray et al., 2009; Chassagne et al., 2009). However, these assays lack an internal amplification control (IAC) and/or do not detect all the known variants of stx2. The Stx2 group is heterogeneous in its nucleic acid composition and consists of at least seven variants (Stx2, Stx2c, Stx2d, Stx2dact, Stx2e, Stx2f, and Stx2g) (Karch et al., 2005). The Stx2f variant, in particular, is not detected with probe-based multiplex real-time PCRs published up to now. Here we report the development of a 5′ nuclease-based multiplex real-time PCR assay targeting the stx1, stx2, and eae virulence genes and their variants. The E. coli and Shigella spp.-specific β-glucuronidase gene uidA is used as an IAC to detect any false-negative result that may occur because of the presence of PCR inhibitory components.
Materials and Methods
Bacterial strains
E. coli bacterial reference strains (n = 30; Table 1), other bacterial strains (n = 68; Table 2) including EIEC, ETEC, EAEC, and other species of the Enterobactericeae family as well as well-characterized diagnostic samples (n = 174, Table 3) were used in this study.
RKI: strain collection from Robert Koch Institut, Wernigerode.
LGL: Bavarian Health and Food Safety Authority.
EHEC, enterohemorrhagic Escherichia coli.
DSM-German Culture Collection.
LGL: Bavarian Health and Food Safety Authority.
RKI: Robert Koch Institut, Wernigerode.
EIEC, enteroinvasive E. coli; ETEC, enterotoxigenic E. coli.
PCR, polymerase chain reaction.
All strains belonging to the Enterobactericeae family were grown on endoagar at 37°C for 16 h. Strains from other taxonomic families were grown on Columbia sheep blood agar at 37°C overnight.
Clinical samples
For validation, 174 clinical stool samples were plated on Columbia sheep blood agar and grown at 37°C overnight. DNA of the mixed cultures on the agar plate was isolated by heat lysis. The presence of stx1, stx2, and eae was determined by the developed multiplex real-time PCR as well as our established laboratory reference assay (Reischl et al., 2002).
If stx1, stx2, or eae was detected in clinical samples, the LGL routine diagnostic bacteriology laboratory looked for the respective EHEC or EPEC strain for subsequent biochemical characterization.
Extraction of DNA
DNA was extracted from bacterial strains either using the High Pure Template Preparation Kit according to the manufacturer's instructions (Roche Applied Science, Mannheim, Germany) or by heat lysis. For this extraction, bacteria grown on appropriate agar media were resuspended in 1.5 mL physiological saline (0.9%). Twenty microliters of this solution was added to 400 μL sterile water and heated at 95°C for 15 min. After centrifugation (14,000 g, 5 min), the supernatant was used for amplification reactions.
Primer and probe design
For designing primers and TaqMan probes targeting the genes of interest (i.e., the variants of stx2 gene of EHEC and the eae gene of EHEC and EPEC), various virulence gene sequences from the GenBank including all hitherto known gene variants were aligned. Primers and probes for the genes were designed using FAST-PCR (
PCR, polymerase chain reaction.
Multiplex real-time PCR
Real-time PCR was performed in 20 μL reactions in a LightCycler® 480 Real-Time PCR System using a 96-well plate format (Roche Applied Science).
Primer–probe mixes were prepared for a particular primer–probe set (Table 4). Each primer–probe mix contained the respective primers and probes at a concentration of 4 μM.
The reaction mixtures contained 10 μL of 2 × QuantiTect Multiplex real time (RT)-PCR NoRox Mastermix (Qiagen, Hilden, Germany), 0.5 μL primer–probe mix specific for uidA, and 2.0 μL for each of the primer–probe mix for detection of stx1, stx2, and eae. Four microliters of template DNA and 1.5 μL sterile PCR-grade water were added to bring the final volume to 20 μL. The multiplex real-time PCR was performed in a LightCycler 480 instrument with detection channels for Cyan500, FAM, HEX/VIC, TexasRed, and Cy5. The PCR program consisted of an initial denaturation step for 15 min at 95°C, followed by 40 cycles of denaturation for 60 s at 95°C and annealing/extension for 60 s at the optimized temperature of 58°C. A color compensation experiment was performed according to the instructions of the LightCyler 480 instrument guide and used for analyses of each experiment.
Determination of the detection limit
To determine the detection limit of the multiplex real-time PCR and the corresponding singleplex PCR, template DNA was diluted from 50 ng/μL to 0.5 fg/μL and the amplifications were measured at each dilution three times. DNA of EHEC strain LGL 15314 for stx2 and DNA of EHEC strain LGL 5622/06 for stx1, eae, and uidA were used. The same dilution series were used for calculating PCR linearity and efficiencies from the following formula: E = 10−1/slope.
Results
Primer/probe design and optimization of the multiplex PCR
Initially, a primer–probe set for detection of stx1 and its variants was developed based on alignment of gene sequences. In singleplex PCRs, this primer–probe set showed good sensitivity, inclusivity, and exclusivity. However, after combining it with the primer–probe sets for eae, stx2, and uidA, cross reactions resulting in unspecific fluorescence signals in the channel for stx1 were detected. Therefore, this primer–probe set was replaced by the primer–probe set published by Sharma et al. (1999).
The primer–probe set for stx1 has already been shown to be specific for stx1 by Sharma et al. (1999). As variants of stx1 were not tested in their study, an alignment with several sequences from the NCBI database was performed, showing that the variants stx1c and stx1d could also be detected with the described primer–probe set.
For stx2, a new primer–probe set, including an additional probe and an additional reverse primer for stx2f, was designed for detection of stx2 and its hitherto known variants (stx2c, stx2d, stx2e, stx2f, and stx2g). Detection of those stx2 variants was verified with reference strains (see Selectivity section). A BLAST analysis of primers and probes for stx2 confirmed their exclusivity for detection of stx2 and its variants.
The available sequences of all known eae variants were aligned and primers as well as probes were selected within a highly conserved region. Test runs showed that the eae-α2 variant was not detected with good efficiencies by using eae_t_p1 as the only probe for eae detection. Development of a second probe for eae (eae_t_p2) solved this problem. The E. coli and Shigella spp.-specific gene uidA was used as IAC. Primer and probe sequences for uidA were adopted from Frahm and Obst (2003). For detection of the pathotype-specific genes, primer and probe concentrations were used as advised by the Multiplex RT-PCR NoRox Mastermix (Qiagen). To avoid competitive inhibitions, uidA primer and probe concentrations were decreased to 0.5 μM each. For stx1, stx2, and uidA, no differences in PCR efficiency and specificity were observed at 58°C or 60°C annealing temperatures. As PCR efficiency for some eae variants was improved at 58°C instead of 60°C, 58°C was chosen as annealing temperature for the multiplex PCR assay.
Verification of PCR products
To ensure that the primers described in Table 4 amplified the target genes, a conventional PCR with DNA from reference strains TU1 for detection of stx1, eae, and uidA, TU2 for detection of stx2, and EHEC T4/97 for detection of the stx2f variant was performed. Before sequencing, the appropriate size of each fragment was proven using agarose gel electrophoresis. Analysis of the obtained nucleotide sequences revealed that each amplicon matched the predicted sequence amplified by the respective primer pair. Cross reactions between the different primer pairs were not observed in conventional PCRs (data not shown).
Selectivity
Selectivity of the multiplex real-time PCR assay was examined using 88 EHEC, 22 EPEC, 95 non-EPEC/EHEC E. coli strains, and 58 non E. coli strains. All strains used were obtained directly from official culture collections (DSMZ), were well-characterized strains from the LGL strain collection, or were well-characterized E. coli pathotypes from Robert-Koch-Institut (Wernigerode), University of Hohenheim, or the Technische Universitaet München. E. coli pathotypes derived from clinical samples isolated in the LGL bacteriology laboratory were routinely characterized by the real-time PCR assay published by Reischl et al. (2002). In addition, genes encoding for the examined pathogenicity factors were sequenced to confirm the detected pathotype or gene variant, if required.
Inclusivity testing was divided into two parts. At first, ability of the assay to detect the different variants of stx1, stx2, and eae was examined using 12 EHEC and 8 EPEC reference strains (Table 1). All hitherto known gene variants of stx1 (stx1, stx1c, stx1d) and stx2 (stx2, stx2c, stx2d, stx2e, stx2f, stx2g) were detected. In case of the eae gene, the multiplex real-time PCR could detect all 11 probed variants.
Further, stx1, stx2, and eae of well-characterized diagnostic samples harboring EHEC (n = 76) and EPEC (n = 14) strains (Table 3) were tested. All stx1, stx2, and eae genes from all tested samples were detected using the quadruplex real-time PCR. In addition, quadruplex real-time PCR detected the stx1 gene of one sample harboring stx1/stx2, and eae genes of one sample harboring an EPEC strain and two stx2/eae samples harboring EHEC strains, which were not detected using the reference method by Reischl et al. (2002). Presence of the genes in the respective samples was confirmed by sequencing the PCR products generated by conventional PCR using primer pairs VS1/VS2 or eae_t_for1/eae_t_rev1.
As stx1, stx2, and eae of the tested strains were detected, the inclusivity of the quadruplex real-time PCR assay was 100%.
For exclusivity testing, 152 strains and samples were used. Ninety four of these were non-EHEC/EPEC E. coli pathotypes including 1 ETEC, 2 EIEC, 5 EAEC, 84 clinical E. coli-harboring samples, and 2 nonpathogenic reference strains. For all of these strains the E. coli-specific uidA gene was amplified, whereas no false-positive results for stx1, stx2, and eae were obtained.
Additionally, 58 non-E. coli strains including 45 strains of the Enterobacteriaceae family and 13 other bacterial strains occurring in the gastrointestinal flora were assayed (Table 2). As expected, none of the strains except Shigella spp. harbored the uidA gene. False-positive results for stx1, stx2, and eae were not obtained (Table 3). Therefore, exclusivity of the EHEC/EPEC multiplex PCR was 100% for all three genes.
Detection limit and PCR efficiency
Detection limits and PCR efficiencies of the quadruplex real-time PCR were determined by measuring DNA dilution series from EHEC strains LGL 15314 and LGL 5622/06 ranging from 50 ng/μL to 0.5 fg/μL. The detection limits of the multiplex PCR were compared with those of the respective singleplex real-time PCR assay (Table 5). For the stx1 and eae assays the detection limit was 500 fg/μL, equivalent to 50 genome copies per μL. The limits of detection for stx2 and uidA were lower and found to be 50 fg/μL, equivalent to five genome copies per μL. No differences in sensitivity were observed between multiplex PCR and the respective singleplex PCRs.
The multiplex real-time PCR also showed good linearity across the range of detection for all genes, with slopes between −3.33 and −3.35 and r 2 values of >0.99, resulting in PCR efficiencies between 1.97 and 1.99 (Table 5).
Discussion
In this study, a quadruplex real-time PCR assay was developed for the simultaneous detection of EHEC and EPEC pathotypes in diagnostic samples. The multiplex PCR was designed for detection of stx1, stx2, eae, and their variants. The β-glucuronidase gene uidA was used as IAC. Testing DNA from a broad range of EHEC and EPEC strains, other E. coli strains, and bacterial species verified the specific detection of these genes. In addition, there were no shifts in PCR sensitivity and efficiency when reactions were performed in multiplex instead of singleplex format. Excellent exclusivity as well as inclusivity was observed for all tested strains and gene variants.
Rapid and accurate methods are essential in identification of pathogenic E. coli. Many studies have employed PCR to target multiple genes associated with EHEC and EPEC virulence and infection, using the benefits of real-time PCR. The Shiga toxin genes stx1 and stx2 as well as the LEE pathogenicity island were primary targets for their detection.
These real-time PCR assays usually detect fewer variants of the stx1 and stx2 genes and do not include internal PCR inhibition controls. In particular, the stx2f variant does not appear to be detected by the probe-based assays published so far or its detection has been explicitly excluded in some of the mentioned studies (Sharma et al., 1999; Bélanger et al., 2002; Ibekwe et al., 2002; Reischl et al., 2002; Jinneman et al., 2003; López-Saucedo et al., 2003; Møller Nielsen and Andersen, 2003; Sharma and Dean-Nystrom, 2003; Perelle et al., 2004; Sekse et al., 2005; Yoshitomi et al., 2006; Brandal et al., 2007; Brendal et al., 2007; Schuurman et al., 2007; Stefan et al., 2007; Yang et al., 2007; Auvray et al., 2009; Guion et al., 2008; Chassagne et al., 2009). Until recently the stx2f variant was thought to amount to only a small fraction of all Shiga toxin-producing E. coli/EHEC strains found to be associated with human infections (Gannon et al., 1990; Friedrich et al., 2002; Jenkins et al., 2003; Isobe et al., 2004; Sonntag et al., 2005; Seto et al., 2007; Van Duynhoven et al., 2008). Thus, in most studies this variant has been considered to have no or rather very limited clinical relevance.
An alignment of the sequences of the primer–probe systems used in these studies with stx2f sequences from official databases clearly shows that detection of stx2f is not possible with the respective primer–probe sets or rather not possible assuming the need of a good PCR efficiency because of accumulating mismatches between reference sequences and primer–probe sets chosen. To our knowledge the only multiplex real-time PCR able to detect stx2f is the SYBR Green I method described by Bischoff et al. (2005), suggesting characteristic specificity problems of nonprobe-based real-time PCRs.
However, as common laboratory procedures used in testing for diarrheagenic E. coli fail to detect the stx2f variant, Prager et al. (2009) have examined 141 E. coli strains originally classified as atypical EPECs for the presence of the stx2f gene. Among those isolates, 32 E. coli strains were identified as stx2f positive, suggesting that stx2f-harboring E. coli are much more frequent among enteric infections in humans than earlier assumed. As a conclusion, stx2f-positive E. coli are considered as new emerging pathogens. Thus it has been strongly recommended to integrate detection of stx2f into routine diagnostics (Prager et al., 2009). However, our multiplex real-time PCR can detect the stx2f variant, pointing to a unique advantage inherent in our detection method.
Additionally, most of the published assays are not applicable as diagnostic tools as they lack an IAC to detect false-negative results due to PCR inhibitory compounds (Hoorfar et al., 2004).
We chose the E. coli and Shigella sp.-specific β-glucuronidase gene uidA as IAC. In fact, uidA was negative for one precharacterized EPEC and EHEC strain while the pathotype-specific genes were detected (Table 3). However, this is not problematic because efficient production of an IAC fluorescence signal is only required when stx1, stx2, and eae amplification has failed.
In comparison to the method currently used in our routine laboratory for detection of EHEC and EPEC (Reischl et al., 2002), our new TaqMan PCR could detect the presence of eae in two stx2/eae-positive diagnostic samples and the presence of stx1 in a stx1/stx2-positive diagnostic sample, which were not detected with the reference method (Table 3). By contrast the newly developed multiplex real-time PCR assay detected those genes as well.
Conclusion
In summary, the developed multiplex real-time PCR-assay is a fast, reliable, sensitive, and well-validated method for EHEC and EPEC detection in a routine laboratory using a single-tube reaction.
Footnotes
Acknowledgments
The authors thank Jasmin Fräßdorf, Katja Meindl, and Sabine Wolf for their technical assistance. This work was supported by a grant from the Bavarian State Ministry of the Environment and Public Health.
Disclosure Statement
No competing financial interests exist.