Simultaneous Screening of 24 Target Genes of Foodborne Pathogens in 35 Foodborne Outbreaks Using Multiplex Real-Time SYBR Green PCR Analysis

A set of 8 multiplex real-time SYBR Green PCR (SG-PCR) assays including 3 target primers and an internal amplification control (IAC) primer was simultaneously evaluated in 3 h or less with regard to detection of 24 target genes of 23 foodborne pathogens in 7 stool specimens of foodborne outbreak using a 96-well reaction plate. This assay, combined with DNA extraction (QIAamp DNA Stool Mini kit), offered detection of greater than 103-104 foodborne pathogens per g in stool specimens. The products formed were identified using melting point temperature (Tm) curve analysis. This assay was evaluated for the detection of foodborne pathogens in 33 out of 35 cases of foodborne outbreak, using 4 different PCR instruments in 5 different laboratories. No interference from the multiplex real-time SG-PCR assay, including IAC, was observed in stool specimens in any analysis. We found multiplex real-time SG-PCR assay for simultaneous detection of 24 target genes of foodborne pathogens to be comprehensive, rapid, inexpensive, accurate, of high selectivity, and good for screening probability.


Introduction
Technological advances in the past 2 decades have substantially increased the possibility of rapid diagnostic testing for many diseases. However, for bacterial pathogens which cause foodborne infections or foodborne outbreaks, traditional culture methods, which can take up to 1 week, are still the only method many microbiology laboratories routinely use for diagnosis [1]. Real-time PCR is one of the principle methodologies emerging for rapid diagnosis of foodborne outbreak. We previously reported a duplex real-time SYBR Green PCR (SG-PCR) screening system of 8 specific genes of foodborne pathogens in 5 fecal samples [2][3][4]. The realtime SG-PCR assay combined with DNA extraction using a QIAamp DNA Stool Mini kit offered detection of greater than 10 3 -10 4 foodborne pathogens per g in fecal samples. For diagnosis and management of foodborne outbreaks, this could distinguish patients infected with foodborne pathogens from healthy carriers. The introduction of this screening system in foodborne outbreak investigations provides an opportunity for comprehensive and rapid detection of pathogens in fecal samples. The results can quickly inform a public health administrator about the causative pathogens of foodborne outbreak, allowing a more accurate, effective and timely response. If it is possible to test for almost all foodborne pathogens including enteric and toxin-producing bacteria at a time, real-time PCR tests will certainly be useful for multiplex screening of foodborne pathogens.       With multiplex PCR tests, if multiple bacteria could be simultaneously detected in the same reaction tube or during the same run, molecular diagnosis may prove to be very costeffective. However, at present, published evaluations of these assays are insufficient.
One of the risks associated with testing samples by PCR is the occurrence of a false negative resulting from PCR inhibition [5,6]. While positive and negative controls are normally run with every PCR master mix to ensure integrity of the reagents, PCR inhibition by the sample matrix can prevent amplification of the target template, resulting in false-negative reporting [5,6]. Therefore, it is necessary to include an internal amplification control (IAC) in each individual reaction mixture to prevent reporting of false negatives [5]. Previous works have utilized various methods for developing and using an IAC for detection of a single target gene, except in the case of 4target TaqMan multiplex PCR to detect V. parahaemolyticus [7].
The objective of the present study was to establish simple and specific methods to simultaneously detect 24 specific genes of foodborne pathogens in 7 stool specimens in a realtime SG-PCR assay using a 96-well reaction plate containing a universal, noncompetitive IAC.  Table 2 are used as control for PCR analysis. DNA was isolated from cultured bacteria to test the specificity of the primers used in this study. Viable counts were obtained by culturing each dilution (10 µL) overnight at 37 • C on tryptic soy agar (TSA) plates for aerobic bacteria and TSA plates containing 3% NaCl for Vibrio spp. Yersinia spp. strains were cultured at 28 • C for 48 h. The Clostridium perfringens strains were cultured on TSA overnight at 37 • C using anaerobic conditions. The Campylobacter jejuni strains were cultured at 37 • C for 48 h on Skirrow agar plates under microaerobic conditions.

Internal Amplification Control (IAC) and IAC Primers for PCR.
An IAC was included in the assay by adding a small amount of PCR products using IAC primers from the bacterium Yersinia ruckeri (JCM15110), which is the causative agent of enteric red-mouth disease in salmonid fish species [8] and the presence of this bacterium in human fecal samples and food samples is never reported. Bacterium used for DNA extraction was grown on brain heart infusion broth (Difco) at 30 • C for 2 days. Two IAC primer pairs with different Tm of PCR products were used for amplifying 16S rRNA gene (GenBank accession no. X75275) of Y. ruckeri. One IAC primer was yers described by Lund et al. [9] and the Tm value of PCR product used for this primer was 77.3 ± 0.15 • C. Another IAC primer sequence of yersH2-F and yersH2-R were chosen by alignment of 16S rRNA gene sequence from foodborne pathogens shown in Table 1 using the BLAST program within GenBank and was designed by Biosearch Technologies Inc. The Tm value of PCR product used for this primer was 86.0 ± 1.5 • C.

DNA Extraction.
For the DNA isolation from bacterial cultures, one milliliter of broth culture was centrifuged at 12,000 × g for 3 minutes. The pellet was then washed in 1 mL of distilled water, centrifuged, and suspended into 1 mL of distilled water. Each 200 microliters of suspension, containing 10 8 foodborne bacterial cells, was treated with the QIAamp DNA Stool Mini kit (Qiagen) according to manufacturer instructions. DNA preparations were used immediately for PCR amplification and stored at −20 • C. Four µL of DNA sample were used for PCR assay. For the DNA isolation from stool samples, stool samples (1 g) were weighed aseptically, placed into sterile tubes, and homogenized with 9 mL of distilled water. Two-hundred µL of this stool suspension was treated with the QIAamp DNA Stool Mini kit according to manufacturer instructions in 1 h or less.  Table 1. The specificity and sensitivity of PCR assay using each primer were confirmed in each referred report. The primer pairs of tdh-F176 and tdh-R422 for the detection of tdh-positive V. parahaemolyticus, yadA-F1757 and yadA-R1885 for the detection of Y. enterocolitica and Y. pseudotuberculosis, PSG-F64 and PSG-R313 for the detection of P. shigelloides, ipaH1672-F and ipaH1761-R for the detection of Shigella spp., and EIEC, daaD-F31 and daaD-R263 for the detection of DAEC were chosen by alignment of virulent or specific gene sequences from foodborne pathogens shown in Table 2 using the BLAST (Basic Local Alignment Search Tool) program within GenBank and was designed by Biosearch Technologies Inc. (Tokyo). The Tm values of these primers varied from 74.5 to 88.7.

Primers
International Journal of Microbiology Presence of PCR target gene was determined by another conventional PCR primer sets before this test.  Table 2 for primer sets.

Real-Time Multiplex SG-PCR.
Real-time multiplex SG-PCR and data analysis were performed for a total volume of 20 µL using 96-well reaction plates and an ABI7500 or ABI7500 Fast Real-Time PCR system (Applied Biosystems), LightCycler 480 (Roche) or Thermal Cycler Dice Real-Time System (Takara, Japan). Each reaction tube contained 10 µL of SYBR Premix DimerEraser (Takara, Japan), 0.4 µL of ROX Reference Dye II (50×) (for ABI 7500 and ABI7500 Fast), 0.8 µL (for ABI 7500 and ABI 7500 Fast), or 1.2 µL (for LightCycler 480 and Thermal Cycler Dice) of PCR-grade H 2 O, each 1.2 µL of a 10 µM primer set for 3 target genes, 1.2 µL of a 10 µM IAC primer set, 2 µL of IAC DNA, and 2 µL of sample DNA in a 20 µL PCR mixture. In each of 8 lines (12 wells per line) on a 96-well reaction plate, the samples were set as negative control (4 µL of dH 2 O) in the 1st well, each 2 µL of IAC, and dH 2 O in the 2nd well, each 2 µL of IAC, and 1 out of 3 target positive controls in the 3rd to 5th wells, and 2 µL of IAC and each 7 stool DNA samples in the 6th to 12th wells. The PCR amplicons resulting from foodborne pathogens and Y. ruckeri were used for the positive controls and IAC, respectively. The concentrations of positive control (equal 10 5 to 10 6 cfu/g) were adjusted to become the C t values to 17 to 21 by dilution of 10 3 -to 10 4 -fold with Easy Dilution (Takara, Japan) and two IACs (equal 10 1 to 10 2 cfu/g) were adjusted to become the C t values to 27 to 29 by dilution of 10 6 -to 10 7 -fold with Easy Dilution. The assay cycling profile was one cycle of 95 • C for 30 s followed by 30 cycles of denaturation at 95 • C for 5 s (3 s for ABI 7500 Fast), annealing at 55 • C for 30 s (34 s for ABI 7500) and then 72 • C for 30 s (34 s for ABI 7500), and a dissociation stage of 1 cycle at 95 • C for 15 s, 60 • C for 60 s, and 95 • C for 15 s. The specificity of the reaction was found by the detection of the Tms of the amplification products immediately after the last reaction cycle. These reactions were finished in 2 hours or less. Results were analyzed with SDS software provided with each real-time PCR system.

Multiplex Real-Time SG-PCR Analysis in 35 Foodborne
Outbreaks. Multiplex real-time SG-PCR analysis of foodborne outbreak was experimentally tested using the ABI 7500 in Shimane (22 cases Table 3). The DNA samples were extracted with the QIAamp DNA Stool Mini kit from patient fecal samples (within 1 hour) and were set on a 96-well reaction plate as described above. The samples before 2008 were used after 1 to 3 years store at −20 • C. The multiplex PCR assay was evaluated with regard to detection (in 2 hours or less) of 24 specific genes of foodborne pathogens in 7 stool specimens. Each PCR product was generated with a different Tm curve among 4 Tm curves of PCR target gene products. These could all be resolved using each software and Tm curve analysis whenever target bacteria were present in the reaction well.

Noncompetitive IAC and Two IAC Primers.
In this study, the Y. ruckeri bacterium was successfully used as a noncompetitive IAC and for 2 pairs of IAC primer for routine detection of 24 target genes of foodborne pathogens. Using an IAC with real-time PCR detection is important to identify false negative results and to control for the presence of amplification inhibitors [26]. It is important to take into account that components of the sample or the competing microflora may influence the effectiveness of the PCR, especially by reducing the detection limit and producing false negative results. The consequences of false negative results in the detection of a pathogenic microorganism may potentially be life threatening [27]. The European Standardization Committee, in collaboration with International Standard Organization (ISO), has proposed a general guideline for PCR testing of foodborne pathogens that requires presence of IAC in the reaction mixture [28].
While some design approaches such as cloning require substantial technical skills, others can be done using basic PCR methodology. There are two main strategies for use of an IAC in a diagnostic real-time PCR assay. Their difference lies in whether the IAC is to be used competitively or noncompetitively [5]. By using the composite primer technique, the target and the competitive IAC are amplified with one common set of primers and under the same conditions and in the same real-time PCR tube. The competitive IAC method was used for TaqMan PCR to detect S. enterica [26,29,30], E. coli O157 [31], and C. botulinum [32] and real-time SG-PCR to detect C. botulinum [33]. However, these competitive IAC methods can lower the amplification efficiency, which results in a lower detection limit [5]. In the noncompetitive IAC method, the target and IAC are amplified using a different primer set for each. The disadvantage is that amplification of the IAC may not accurately reflect amplification of the target. This method was used for TaqMan PCR to detect Campylobacter spp. [9,34], B. cereus [35], C. botulinum [36], and V. parahaemolyticus [7]. These assays were used for the detection of single target gene except for the four-target TaqMan multiplex PCR to detect V. parahaemolyticus [7]. Although the main advantage of the noncompetitive IAC method is that it can be used for many different assays in the same laboratory [5], we do not have a unique real-time PCR assay for the detection of almost all foodborne pathogens using universal IAC.
Each previously described method for introduction of an IAC is limited due to primer competition or because it requires the presence of a specific substrate or organism. The new approach presented in this paper comprises a separate amplification of target DNA and noncompetitive IAC-DNA using each specific target primer set and two different IAC primer sets on the detection of each foodborne pathogens. The latter is based on 16S rRNA of Y. ruckeri, which is not found naturally in human stool and food samples.
The IAC primer yersH2 was used for detection of 15 target genes of foodborne pathogens which Tm values of amplicons were lower than 83 • C and shown as primer sets A to E (described in the next section), and the IAC primer     yers was used for detection of 9 target genes of foodborne pathogens which Tm values of amplicons were over than 80 • C and shown as primer sets F to H (described in the next section). The IAC-specific low peak on a Tm curve analysis was present in all reaction tubes added with IAC and IAC primers and in all the negative results of target PCR in reaction tubes added with IAC primers (Figures 1 and 2).

Development of PCR Procedures for a Set of 8 Multiplex
Assays. We developed the ultimately new PCR screening system for foodborne pathogens in stool specimens. One can simultaneously analyze 24 pathogenic or specific genes of foodborne pathogens in 7 stool specimens by using multiplex real-time SG-PCR containing IAC and 96-well reaction plate. Single or multiple real-time PCR assays were reported for detection of one species among foodborne pathogens, such as E. coli [11,17,18,22], Salmonella [26,28,29], C. jejuni [13,37], V. cholerae [19], V. parahaemolyticus [38], and S. aureus [21]. Comprehensive, rapid real-time SG-PCR procedures, which used 24 primer pairs for detection of 15 bacterial species including: 6 groups of E. coli, 2 subgroups each of B. cereus and V. parahaemolyticus, and 2 primer pairs for an IAC, were developed using a set of 8 multiplex PCR assays with 3 primer pairs for foodborne pathogens and an IAC primer pair. Nineteen pairs of primers for foodborne pathogens were selected from earlier publications (Table 1), and 5 pairs of primers for tdh gene of TDHpositive V. parahaemolyticus, yadA gene of Y. enterocolitica and Y. pseudotuberculosis, gyrB gene of P. shigelloides, ipaH gene of EIEC and Shigella spp., and daaD gene of DAEC were constructed. This was done to make all 24 SG-PCR methods suitable for the same PCR conditions (an annealing temperature of 60 • C). The sequence, target, PCR product size, threshold cycle (C t ) values, and Tm values of 24 primer pairs for target genes and 2 primer pairs for IAC are listed in Table 1. The specificity of the PCR assay was confirmed on 659 strains listed in Table 2. The STa-F and STa-R primer pair could not detect st gene from 5 of 18 st-positive ETEC strains. The ipaH1672-F and ipaH1761-R primer pair crossreacts with Shigella spp. and EIEC. The SG-F and SG-R primer pair cross-reacts with enterotoxigenic and emetic B. cereus. As same as previous studies [4], the eae-F2 and eae-R primer pair cross-reacts with EPEC and EHEC, and the EAST-1S and EAST-1AS primer pair cross-reacts with EAEC and some strains of EPEC, ETEC, and DAEC. The yadA667-F and yadA851-R2 for Yersinia adhesion reacts with virulent Y. enterocolitica and Y. pseudotuberculosis, but not with nonpathogenic strains of Yersinia spp. A Foodborne Outbreak Investigation Report (http:// www.mhlw.go.jp/topics/syokuchu/), by the Ministry of Health, Labor and Welfare, Japan, during 2005 to 2008, shows that 97% of foodborne outbreaks were caused by the following 7 species of foodborne pathogens: C. jejuni (56.5%), S. enterica (16.0%), TDH-positive V. parahaemolyticus (10.0%), S. aureus (6.8%), C. perfringens (3.4%), emetic B. cereus (2.0%), and EHEC (2.4%), and other virulent E. coli (2.1%) which include astA-positive E. coli which is a strain of E. coli that does not possess any diarrheagenic characteristics except the EAEC heatstable toxin 1 (EAST1) gene and is frequently isolated in diarrhea outbreaks [39]. Each primer set was combined with 4 primer pairs designed for 1 of 8 main foodborne pathogens and were also designed for IAC and 2 of 16 target genes of other foodborne pathogens (  primer for 8 main foodborne pathogens was carefully set in 8 different primer sets for keeping away from the presence of multiple primers for main foodborne pathogens in the same reaction well. Really the plural target genes were detected from 26 stool samples in 15 cases of foodborne outbreaks but these target genes were, respectively, detected from different reaction wells (Table 3). The fluorescent amplification curves and Tm curves of the multiplex SG-PCR products of the DNA of foodborne pathogens and IAC were shown in Figure 1. The C t values of the amplicons resulting from foodborne pathogens were 17 to 21 and those of IAC were 27 to 29. In each analysis, the T m distance was from 0.8 to 6.2 • C among the target gene's products. Looking at the short variations among some T m distance in set C and set E, the presence of ces gene in set C could be decided altogether with the presence of nheB gene of B. cereus in set H and the presence of stx1 gene in set E could be decided altogether with the presence of eaeA gene of EHEC in set D. The IAC-specific low peak was present in all samples with added IAC (C t : 27 to 29) using real-time SG-PCR analysis of the 24 target genes of primer sets A to E including primer yersH2 and of primer sets F, G, and H including primer yers. IAC was certainly amplified in the IAC-only samples. These could be resolved in the ABI 7500 by using Tm curve analysis when a target bacterium was present in the reaction tube. The Tm values of PCR products of stool samples, including each foodborne pathogens, could be identified with that of control bacteria in the same run based on a Tm curve analysis.

Multiplex SG-PCR for Identification of the Causative
Pathogens in Foodborne Outbreaks. In foodborne outbreaks, stool specimens from patients infected with enteric bacteria with acute severe disease may contain large numbers of causative bacterial species [2,11]. In most cases of foodborne outbreak, we found that causative bacteria can be rapidly detected and that a presumptive diagnosis of the causative agent of foodborne outbreak could be made within 3 hours. We used a combination of the multiplex real-time SG-PCR assay with DNA extraction with the QIAamp DNA Stool International Journal of Microbiology Mini kit used for detection. Almost all bacterial pathogens are detectable in stool specimens at a concentration of 10 3 to 10 4 bacteria per g. This is because the concentration of DNA extracted from stool specimens using this DNA extraction kit was finally diluted to 6 × 10 4 -fold in the reaction mixture. The PCR sensitivity for bacteria inoculated in stool samples may be as low as the presence of 10 cells in the reaction well, as described in our previous report [2]. The real-time SG-PCR assay is a rapid, specific, and sensitive detection technique. The DNA extraction of 7 stool specimens with this DNA extraction kit was carried out within 1 hour. Then, the multiplex real-time SG-PCR assay was also carried out within 2 hours, and we could then specifically identify the products based on a Tm curve analysis. For example, Figure 2 shows those of case 21, in which C. jejuni and astApositive E. coli strains were isolated from 4 and one of 7 stool specimens of symptomatic patients, respectively. Two panels show detection of target genes of foodborne pathogens by primer sets B and G, but it was not detected by the other 6 primer sets. In multiplex PCR analysis, the C. jejuni-specific gene and the astA gene were simultaneously detected by primer sets B and G from the same culture-positive stool specimens.  Enteritidis (1 case). There were also 19 (54.2%) cases with plural foodborne bacterial pathogens and 3 (2.9%) cases with foodborne bacterial pathogens (astA-positive E. coli, EHEC O:26, or C. perfringens) and norovirus. The causative pathogens had been isolated from 125 (62.8%) of 199 PCR examined samples and from 216 (56.7%) of 381 total samples in all 35 cases. Although the target genes of EPEC, EAEC, ETEC, astA-positive E. coli, P. alcalifaciens, and A. hydrophila were detected by SG-PCR, the isolation of these pathogens from the stool samples containing much normal E. coli flora was difficult. This analysis may be a very useful tool for the detection of these unusual pathogens which are generally difficult to isolate. We previously reported that the presence of any foodborne pathogens at more than 10 3 CFU/g feces was certainly confirmed by melting curve analysis in duplex SG-PCR [2,4]. In this multiplex PCR analysis including IAC, the presence of any foodborne pathogens at more than 10 5 CFU/g feces was certainly confirmed in 40 (97.6%) of 41 samples by melting curve analysis, 10 4 CFU/g feces was confirmed in 7 (63.6%) of 11 samples and 10 3 CFU/g feces in 3 (50%) of 6 samples ( Table 4). The sensitivity of this multiplex SG-PCR including IAC might became slightly lower than that of duplex SG-PCR (absent IAC), caused by the interference among 4 primer pairs including IAC primer in the same reaction well. In 2 cases (5.7%), in which S. enteric serovar Enteritidis was isolated by direct culture (unknown cfu) from one patient in case 23 and 10 4 cfu/g of feces from 2 patients in case 26, the target gene of Salmonella was not detected by multiplex SG-PCR, because the sensitivity of invA2 primer may be slightly lower than those of other primers. The choice or design of more sensitive primer for the detection of Salmonella spp. is indispensable in future studies.

Usefulness of Multiplex SG-PCR for the Rapid Diagnostic
Test in Foodborne Outbreaks. Systematically reviewing clinical implications, public health considerations, and costeffectiveness of rapid diagnostic tests for detection and identification of bacterial intestinal pathogens in feces and food [1], economic modeling suggests that adoption of rapid test methods, especially for PCR, in combination with a routine culture is unlikely to be cost-effective, however, as the cost of rapid technologies decreases, total replacement with rapid technologies may be feasible. Despite the relatively poor quality of reporting of studies evaluating rapid detection methods, the reviewed evidence shows that PCR for Campylobacter, Salmonella, and E. coli O157 is potentially very successful in identifying pathogens. It is possibly detecting more than the numbers currently being reported using cultures. Less is known about the benefits of testing for B. cereus, C. perfringens, and S. aureus. This review pointed out that further investigation is needed on how clinical outcomes may be altered if test results are available more quickly and at greater precision than the current practice of using bacterial culture [1]. In the present study, simple and specific methods were established to detect comprehensive and simultaneously 24 specific genes of foodborne pathogens including main bacterial pathogens such as Campylobacter, Salmonella, E. coli O157, B. cereus, C. perfringens, and S. aureus in 7 stool specimens in a real-time SG-PCR assay using a 96-well reaction plate containing a universal noncompetitive IAC. The usefulness of this method for the rapid diagnostic tests was confirmed by the successful detection of causative bacteria in 33 foodborne outbreak cases.
In conclusion, the multiplex real-time method described here for simultaneous screening of 24 target genes of foodborne pathogens were comprehensive, rapid, inexpensive, highly selective, accurate, and demonstrated detection probability. Due to the use of IAC and 2 IAC primers, the assay is suitable for accurate and rapid diagnosis of almost all foodborne pathogens in stool specimens of foodborne outbreak outbreaks. In future studies, workers should improve the kit of multiplex real-time PCR and select more suitable primers for foodborne pathogens.