Comprehensive and Rapid Real-Time PCR Analysis of 21 Foodborne Outbreaks

A set of four duplex SYBR Green I PCR (SG-PCR) assay combined with DNA extraction using QIAamp DNA Stool Mini kit was evaluated for the detection of foodborne bacteria from 21 foodborne outbreaks. The causative pathogens were detected in almost all cases in 2 hours or less. The first run was for the detection of 8 main foodborne pathogens in 5 stool specimens within 2 hours and the second run was for the detection of other unusual suspect pathogens within a further 45 minutes. After 2 to 4 days, the causative agents were isolated and identified. The results proved that for comprehensive and rapid molecular diagnosis in foodborne outbreaks, Duplex SG-PCR assay is not only very useful, but is also economically viable for one-step differentiation of causative pathogens in fecal specimens obtained from symptomatic patients. This then allows for effective diagnosis and management of foodborne outbreaks.


Introduction
The introduction of real-time PCR in foodborne outbreak investigations provides an opportunity for rapid detection of pathogens in food and clinical settings [1]. The benefits to public health administration from rapid real-time PCR assays are most notable after comprehensive and rapid detection of bacteria. The results can quickly inform a public health administrator about the causative pathogens of foodborne outbreak, allowing a more accurate, effective, and timely response. Abubakar et al. [2] implied in the Health Technology Assessment program (now part of the National Institute for Health Research, UK) that the feasibility of conversion to rapid methods such as multiplex PCR and DNA microarrays is dependent on localized considerations, including the community prevalence rates for specific pathogens, the skill base, and subsequent training costs for laboratory staff and spare capacity available to ensure adequate laboratory space for new equipment. Although these tests look promising, further studies are necessary to assess their usefulness [2]. Apart from saving time, real-time PCR is sensitive, highly specific and offers the potential for quantification [3]. The risk of cross-contamination is significantly reduced, and high-throughput performance and automation are possible since no post-PCR manipulations are required [4]. In principle, two different chemistries are available for realtime detection of PCR products: fluorescent probes that bind specifically to certain DNA sequences and fluorescent dyes that intercalate into any double-stranded DNA. Fluorescentprobe based real-time PCR (TaqMan PCR) studies to detect causative pathogens from foodborne outbreaks in feces using TaqMan probes have been carried out [3][4][5][6]. TaqMan PCR assays require the availability of primers and probes that must be selected according to very rigid criteria. Use of simple, cheaper double-stranded DNA-binding dye SYBR green I for detection of PCR amplicons (SG-PCR) overcomes this limitation. Therefore, real-time PCR could be applied without the need for fluorescent probes [7]. In the absence of probes, the specificity of the reaction is determined on the basis of the melting temperature (T m ). The advantages of SG-PCR over TaqMan PCR include the relative simplicity and 2 International Journal of Microbiology reduced cost of SYBR Green I compared to TaqMan probes [8]. Recently, the application of SG-PCR for the detection of foodborne bacteria in different samples has been increased [1,[9][10][11][12]. Duplex SG-PCR assays have been carried out to detect causative bacteria in feces from foodborne outbreaks [4,10,13].
We previously reported a set of four duplex SG-PCR assays for one-step differentiation of 8 genes of foodborne pathogens in DNA extracted from 5 feces using 32 capillary tubes of LightCycler (Roche). The first run was for the detection of 8 main foodborne pathogens and the second run was for the other pathogens. We reported here that improved diagnostic duplex SG-PCR assays were upgraded with new highly sensitive primer pairs for 11 foodborne pathogens. These assays successfully identified the causative pathogens of foodborne outbreaks caused by enteropathogenic  (Table 1). Bacterial cultures and viable-cell counting were described in a previous report [10]. For template DNA of each foodborne pathogen as a PCR control, 200 μL of each bacterial culture (10 8 CFU/mL) was treated with a QIAamp DNA Stool Mini kit (Qiagen) in the same procedure as the following stool treatments.  Table 2 were used for cases 19 to 21. In this study, 10 primer pairs (marked with * in Table 2) were newly designed or selected from earlier publications (see Table 2 references). The 4 primer pairs (ces, yadA-X, CCceuE, and aggR-Z) were newly designed. The ces primer was constructed from cereulide synthetase gene of emetic B. cereus [4], the yadA-X primer from yadA gene on the plasmid present in virulent Yersinia spp. [24], the CCceuE primer from ceuE gene encoding of a lipoprotein component of a binding-protein-dependent transport system for the siderophore enterochelin of C. coli [25], and the aggR-Z primer from aggR gene encoding of a transcriptional activator for EAEC aggregative adherence fimbria I expression [26]. To determine the specific primers ces, yadA-X, CCceuE, and aggR-Z, the genes of ces, yadA, ceuE, and aggR that were expected to be unique were selected with the Basic Local Alignment Search Tool (BLAST) program within GenBank and were designed by Biosearch Technologies Inc. (USA). Other primer pairs were those used in earlier publications (see Table 2 references). All oligonucleotide primers were synthesized by Invitrogen (Yokohama, Japan) or Biosearch Technologies Inc. (USA).

Duplex SG-PCR with Feces.
Feces (1 g) from 5 patients were weighed aseptically from the mass sample collected for virological inspection, placed into sterile tubes, and homogenized with 9 mL of distilled water. Then, 200 μL of stool suspension was treated with a QIAamp DNA Stool Mini kit. For real-time PCR, we used SYBR Premix EX Taq (Takara, Japan), 32 glass capillary tubes, and a LightCycler instrument (Roche Diagnostics, Mannheim, Germany) as described by the manufacturer. Duplex SG-PCR was performed using 32 glass capillary tubes with 4 groups of 2 primer sets on the LC instrument for each run. Analysis of each group of primer pairs was made in 8 glass capillary tubes; each of which included 1 negative DNA control consisting of PCRgrade water, 2 positive controls, and template DNA from 5 feces. The first run of duplex SG-PCR was analyzed using 4 primer sets selected from 11 primer sets described in our previous reports [10,13]. The newly first run primer set including eae plus FemB, AB plus EAST1, Tdh plus Ces-TM, and Styinva plus GAP (see Table 2) was used for analysis of cases 19 to 21. The second run was analyzed using 4 primer sets selected from the following primer sets: LT plus AHH1, STa plus PSG, aggR-Z plus virA, SG plus PAG and the third run using yadA-X plus CCceuE, and hlyA plus Trh. The eaeA-positive samples were analyzed by simple PCR using primers JMS1 and JMS2. Each reaction tube contained 10 μL of SYBR Premix EX Taq, 6.8 μL of PCR-grade H 2 O, 0.4 μL of both forward and reverse primers (10 μM) for the target gene of two foodborne pathogens, and 2 μL of template DNA in a 20 μL PCR mixture. The assay cycling profile was 95 • C for 10 minutes, followed by 30 cycles of denaturation at 95 • C for 5 seconds and then annealing at 60 • C for 20 seconds. Fluorescence signals were measured once per cycle at the end of the extension step. After PCR amplification, a melting temperature curve analysis was done. Next, the LightCycler PCR products were cooled to 65 • C and then heated to 95 • C at a rate of 0.1 • C per second. The fluorescence signals obtained were continuously monitored to confirm amplification specificity during 1 hour of analysis. The products' melting temperature peaks were calculated by performing 10 or more assays per sample and were based on the initial fluorescence curve found by plotting the negative derivative of fluorescence over temperature  Table 2  Bacterial strains Sources e PCR results with each primer set (see Table 2) eae JMS1 JMS2 LT STa EAST-1 aggR-Z virA Styinva yadA-X PAG PSG AB CC ceuE hlyA tdh trh AHH1 FemB ces-TM SG GAP

Duplex SG-PCR Procedures.
We previously reported duplex SG-PCR assays for detection of 19 species of foodborne pathogens using 22 primer pairs [10,13]. After that, more accurate duplex SG-PCR assays were designed by 10 more sensitive and specific primers including 6 primers (FemB, AB, ces-TM, Styinva, SG, and AHH1) selected from earlier publications (see references in  [27]. Using of 4 primer sets of 2 primer pairs, including newly selected or designed 6 primer pairs, for the detection of 7 main foodborne pathogens and astA-positive E. coli in the first run of duplex SG-PCR brought out the comprehensive, rapid, and sensitive detection of causative pathogens in foodborne pathogens to cases 19 to 21 (Table 2 and Figures 1 and 2). The second run of duplex SG-PCR used 4 primer sets and the final run utilized 2 primer sets selected from the remaining 4 primer pairs. The primers JMS1 and JMS2 were used for the single PCR detection of stx1 and/or stx2 genes from the eaeA gene-positive samples for the confirmation of EHEC. Figures 1 and 2 show the T m curves of the duplex SG-PCR products of the template DNA samples in each run. In duplex SG-PCR assay with two primer pairs, each PCR product was generated with a different T m curve. These could be resolved in a LightCycler by using T m curve analysis when a target bacterium was present in the reaction tube.  : 3, 12, and 14). Then the causative pathogens were later isolated in a routine laboratory. In cases 11 and 12, C. perfringens or C. jejuni was detected by duplex SG-qPCR with more than 10 5 CFU/g feces from only 1 sample and C. perfringens was then also isolated from only 1 of 46 samples and C. jejuni from only 1 of 16 samples by culture method. Therefore, the infections with both these pathogens were determined to be sporadic cases and they were immediately eliminated as causative pathogens in cases 11 and 12. It was confirmed that duplex SG-PCR analysis of 5 feces collected from symptomatic patients was ultimately the most effective screening method for foodborne pathogens in foodborne outbreaks [10,13].       Figure 1: Melting curve analysis of duplex SYBR Green I PCR products in the first run using four primer sets: FemB plus eaeA, AB plus EAST1, ces plus tdh, and GAP plus Styinva.

Using Duplex SG-PCR for Identification of the Causative Agent in 21 Foodborne Outbreaks.
Duplex SG-PCR rapidly and accurately demonstrated that 12 (57.1%) of 21 cases were caused with a single foodborne pathogen such as C. jejuni (6 cases), C. perfringens (3 cases), B. cereus (2 cases), and TDH-producing V. parahaemolyticus (one case). There were also 7 (33.3%) cases with plural foodborne bacterial pathogens (such as astApositive E. coli, EPEC, C. jejuni, C. perfringens, A. hydrophila, and P. shigelloides) and 2 (9.5%) cases with foodborne bacterial pathogens (astA-positive E. coli or EHEC O:26) and norovirus. In cases 2 and 10, although detection of norovirus is out of the scope of our work, norovirus and foodborne bacterial pathogens were concomitantly detected by conventional PCR analysis in our virological laboratory. In case 2 in which norovirus was detected in 6 of 7 feces, the astA gene of EAEC was detected from 7 of 10 feces and then astA-positive E. coli strains were isolated from 6 samples. In case 10 in which norovirus was detected from 20 of 22 feces, the eae gene of EPEC or EHEC was detected from 8 of 22 feces and EHEC O26 strains were isolated from 8 of 22 feces. In 7 cases (cases 1, 11, 12, 13, 16, 20, and 21), the pathogenic E. coli strains belonging to astA-positive E. coli and/or EPEC were concomitantly detected with other foodborne bacterial pathogens. In case 1, the eae gene of EPEC or EHEC was detected from 4 of 22 feces and the astA gene of EAEC was detected in 3 other feces. However, duplex SG-PCR could not detect other virulent genes, including the stx1 and stx2 genes of EHEC. Then EPEC strains were later isolated from 5 feces and astA-positive E. coli from 4 other feces. In case 12, the astA gene of EAEC was detected in all 5 feces and the eae gene of EPEC or EHEC in 2 feces, but duplex SG-PCR could not detect other E. coli virulent genes. The subsequent bacteriological examination could not isolate pathogenic E. coli among nonpathogenic E. coli flora. In case 16, the C. jejuni specific gene was detected in 6 of 9 feces and the astA gene of EAEC was detected in 5 feces (both genes from 3 feces). C. jejuni strains were then isolated from 9 of 14 feces, but we were not able to isolate the pathogenic E. coli strain among nonpathogenic E. coli flora. In cases 19 to 21 analyzed improved real-time PCR using 8 primers for the detection of 7 main foodborne bacteria and astA-positive E. coli, C. jejuni, EPEC, or astA-positive E. coliwere detected from 1 to 3 fecal samples on the first run and the absence of the other main foodborne bacteria in the analyzed samples was readily confirmed. In case 20, the eae gene of EPEC or EHEC was detected from 2 of 5 fecal samples on the first run and the gyrB gene of P. shigelloideswas detected separately from other 2 fecal samples on the second run. Then P. shigelloides strains were isolated from 2 feces, but isolation of the EPEC strain was very difficult due to the presence of large nonpathogenic E. coli flora in the feces.
In almost all cases, the duplex SG-PCR assay first run detected these causative agents from more than one of the five feces. Then, in almost all cases, the presence of a causative agent (presumed from duplex SG-PCR assay) was confirmed by the results of the final SG-PCR assay run and the bacteriological cultivation of additional feces. These findings confirmed that for foodborne outbreaks duplex SG-PCR is a useful tool for the rapid detection of both single and multiple pathogens.  Figure 2: Melting curve analysis of duplex SYBR Green I PCR products in the second run using four primer sets: ST plus PSG, aggR plus virA, LT plus AHH1, and PAG plus SG; the third run using two primer sets: CCcesE plus yadA and trh plus hlyA; simple PCR with primers JMS 1 and JMS2.

Quantification of the Causative Agent in 14 Foodborne
Outbreak Cases. Figure 2 shows the relationship between CFU and DNA copy of foodborne pathogens using SGquantitative PCR (qPCR) assay in 71 feces from 14 cases examined by viable cell counting. There was no correlation (r 2 = 0.1183) between CFU and DNA copy of foodborne pathogens in feces, although almost all pathogens were detected by SG-PCR from feces registering more than 10 3 CFU/g by viable cell counting. By using SG-qPCR assay combined with DNA extraction using the QIAamp DNA Stool Mini kit, Bibbal et al. [28] reported a significant correlation between CFU and DNA copy of ampicillinresistant Enterobacteriaceae in swine feces. Fu et al. [29] reported a significant correlation between CFU and DNA copy of Lactobacillus and total anaerobic bacteria in dog feces but found no correlation between CFU and DNA copy of  C. perfringens. Although accurate quantifications of foodborne pathogens, including C. jejuni and C. perfringens, in feces were not completely performed by SG-qPCR in this study, the presence of any foodborne pathogens at more than 10 3 CFU/g feces was certainly confirmed by melting curve analysis. There are two major problems for these differences.
One cause is different sample preparation that was used for CFU from the feces stored in the transport medium and for qPCR using the mass sample collected for virological inspection. Another cause is the approach used to construct the standard curves that were prepared from pure bacterial cultures. These curves do not relate with the "real" situation of a bacterial quantification in a faecal sample and can in part explain the absence of correlation between CFU and DNA copy of foodborne pathogens in faeces. In our routine bacteriological diagnostic laboratory, we used duplex SYBR Green I PCR assay combined with DNA extraction via QIAamp DNA Stool Mini kit for the detection of foodborne bacteria from 21 foodborne outbreak cases. The causative bacteria were detected in almost all cases in 2 hours or less. The first run was for the detection of 8 main foodborne bacteria and the second run was for the detection of other unusual suspect bacteria. The results proved that for comprehensive and rapid molecular diagnosis in foodborne outbreaks, duplex SG-PCR assay is not only very useful, but is also economically viable for one-step differentiation of causative bacteria in fecal specimens obtained from symptomatic patients. This then allows for effective diagnosis and management of foodborne outbreak.