Detection of 12 Common Food-Borne Bacterial Pathogens by TaqMan Real-Time PCR Using a Single Set of Reaction Conditions

Food safety has become an important public health issue worldwide. However, conventional methods for detection of food-borne pathogens are complicated, and labor-intensive. Moreover, the sensitivity is often low, and it is difficult to achieve high-throughput detection. This study developed a TaqMan real-time polymerase chain reaction (PCR) assay for the simultaneous detection and quantification of 12 common pathogens in a single reaction, including Escherichia coli O157:H7, Listeria monocytogenes/ivanovii, Salmonella enterica, Vibrio parahaemolyticus, β-streptococcus hemolyticus, Yersinia enterocolitica, Enterococcus faecalis, Shigella spp., Proteus mirabilis, Vibrio fluvialis, Staphylococcus aureus, and Campylobacter jejuni in food and drinking water. Based on published sequence data, specific primers, and fluorescently-labeled hybridization probes were designed targeting based on the virulence genes of the 12 pathogens, and these primers and probes were optimized to achieve consistent reaction conditions. The assay was evaluated using 106 pure bacterial culture strains. There was no cross-reaction among the different pathogens. The analytical sensitivity was 1 copy/μL for E. coli O157:H7, L. monocytogenes/ivanovii, β-streptococcus hemolyticus, Shigella spp., P. mirabilis, and V. fluvialis, 10 copies/μL for S. enterica, V. parahaemolyticus, Y. enterocolitica, E. faecalis, S. aureus, and C. jejuni, respectively. The limit of detection (LOD) was 296, 500, 177, 56, 960, 830, 625, 520, 573, 161, 875, and 495 CFU/mL for E. coli O157:H7, L. monocytogenes/ivanovii, S. enterica, V. parahaemolyticus, β-streptococcus hemolyticus, Y. enterocolitica, E. faecalis, Shigella spp., P. mirabilis, V. fluvialis, S. aureus, and C. jejuni, respectively. The limit of detection for the assay in meat samples was 103 CFU/g for V. parahaemolyticus and 104 CFU/g for other 11 strains. Together, these results indicate that the optimized TaqMan real-time PCR assay will be useful for routine detection of pathogenic bacteria due to its rapid analysis, low cost, high-throughput, high specificity, and sensitivity.


INTRODUCTION
Researchers have identified more than 250 known food-borne illnesses, most of which are infectious, and caused by a variety of bacteria, followed by viruses and parasites (Mangal et al., 2016). Bacterial food-borne diseases are becoming a growing public health concern for the whole world, especially for the developing countries (Fung et al., 2018). According to the World Health Organization (WHO) estimates, food-borne or water-based diarrhea are responsible for about 2.2 million deaths around the world each year (Johnson, 2011). In China, the situation is more optimistic, one in every 6.5 people is suffering from food-borne disease due to the intake of food contaminated with pathogens. E. coli O157:H7, L. monocytogenes, S. enterica, S. aureus, and C. jejuni contribute to most of the foodborne outbreaks (Mangal et al., 2016). Besides most outbreaks associated with contaminated foods, contaminated drinking water outbreaks have been reported (Park et al., 2011). Given the globally public health and economic burden due to food-borne illness, it is essential to develop reliable, and rapid methods for pathogen detection.
Conventional culture-based methods are still regarded as the "gold standard" for the identification of pathogenic bacteria, but the technique has several disadvantages. It is time-consuming, labor-intensive, and quantitative analysis is difficult to achieve (Wiemer et al., 2011). In addition, false-negative results may happen because of viable but non-culturable (VBNC) pathogens (Law et al., 2014). Furthermore, traditional cultural methods lack good sensitivity and are not capable of simultaneously detecting multiple food-borne pathogens. Thus a rapid, highly sensitive, and inexpensive detection techniques is required (Espy et al., 2006;Park et al., 2010;Ranjbar et al., 2014a;Valencia-Shelton and Loeffelholz, 2014).
Molecular-based methods included are polymerase chain reaction (PCR), real-time PCR, loop-mediated isothermal amplification (LAMP), and microarray which are used to detect food-borne pathogens widely (Zeng et al., 2016). In contrast to conventional single and multiplex PCR, real-time PCR technology that can monitor the products by measuring the fluorescent signal continuously, is most commonly used as a rapid and reliable tool because of its high sensitivity and specificity. Several fluorescent system such as TaqMan probes have been developed (Law et al., 2014). However, the design of primer/probes was a critical problem in TaqMan real-time PCR assays. Owing to less polymorphic loci in the gene, 16S rRNA sequences cannot be used to distinguish closely related bacteria at the species level such as Escherichia coli and Shigella spp. (Ohara-Nemoto et al., 1997), and cannot identify bacterial pathogenicity. The virulence genes have been used as target genes for nucleic acid-based assays (Finlay and Falkow, 1997;Wiemer et al., 2011;Van Lint et al., 2015). Namely, flic, hlyA, rfbE were selected from E. coli O157:H7, invA, staG from Salmonella spp., ail, nuc from Staphylococcus aureus and foxA genes from Yersinia enterocolitica to design primers and probes in real-time TaqMan PCR assay (Ranjbar et al., 2014b;Wang et al., 2014;Ding et al., 2017;Zhou et al., 2017). Additionally, a simultaneous quantifying detection of V. parahaemolyticus and L. monocytogenes by TaqMan-based real-time PCR using primers and probes that target tlh and hlyA genes, respectively, was developed (Zhang et al., 2015). Also, 23 individual TaqMan real-time PCR was developed to detect common food-borne pathogens by using TaqMan probes (Cremonesi et al., 2014).
In this study, the primers and probes were designed based on the specific virulence genes of E. coli O157:H7, L. monocytogenes/ivanovii, S. enterica, V. parahaemolyticus, β-streptococcus hemolyticus, Y. enterocolitica, E. faecalis, Shigella spp., P. mirabilis, V. fluvialis, S. aureus and C. jejuni, and a simple, fast, efficient, and economical real-time PCR method using the TaqMan probe for simultaneous detection and quantification of 12 common pathogenic bacteria was established. This assay was evaluated using genomic DNA from pure culture collection strains.

Bacterial Strains and Culturing Conditions
Escherichia coli spp., Listeria spp., Salmonella enterica, Yersinia enterocolitica, Shigella spp., Proteus spp., and Staphylococcus spp. were cultured in tryptic soy broth (TSB; BD, USA) at 37 • C. Vibrio parahaemolyticus were grown in tryptic soy broth supplemented with 3% NaCl at 37 • C. Streptococcus spp. were grown on tryptic soy agar (TSA; BD, USA) with 5% sheep blood at 37 • C to produce isolated colonies. Vibrio fluvialis were grown in 2216E liquid medium (Hope Bio-Technology, Qingdao, China) at 37 • C. Campylobacter jejuni were cultured on Columbia blood agar under microaerophilic conditions (5% O 2 , 10% CO 2 , and 85% N 2 ) at 42 • C for 24-48 h. Moreover, 12 bacterial strains were used to determine the specificity and sensitivity of the primer pairs (Table 1), and the strains were cultured on the appropriate growth media and under suitable culture conditions. Following overnight incubation, genomic DNA was extracted from supernatants of collected bacterial cells after centrifugation.

Genomic DNA Extraction
DNA used as template for PCR amplification was isolated from ∼0.5-1 mL (1.0-5.0E+09 cells) of a freshly grown (18-24 h) bacterial culture using MiniBEST Bacteria Genomic DNA Extraction Kit (TaKaRa, Japan) according to the manufacturer's instructions. The DNA quantity and purity were measured spectrophotometrically via absorbance measurements at A 260 and A 280 , as well as visualization on 1% agarose gel. A DNA Isolation Reagent for Meat and Meat Products (TaKaRa, Japan) was used to isolate DNA in meat as recommended by the manufacturer.

Primers and Probes
Virulence genes of the 12 bacterial pathogens were chosen as target genes whose sequences are available from NCBI and the pathogen-specific virulence factors (VFs) from our previous work were selected in priority (Niu et al., 2013). Furthermore, the pathogen-specific conservative sequence fragments of VFs were obtained by carrying out the designed similarity comparative algorithm approach (unpublished data, Yang Cao and Chao Niu). The 12 sets of different primers and TaqMan probes (FAM labeled) based on the pathogen-specific conservative sequence fragments were designed using the Primer Express 3.0 program (http://frodo.wi.mit.edu). Sequence features of each gene such as regions of high or low GC-content, and size were examined to ensure the same amplification conditions. The in silico specificity was also analyzed using nucleotide BLAST (http://blast.ncbi.nlm.nih.gov/) from the GenBank sequence

Assessment of Specificity of the Assay Using the Pure Cultures
Specificity of the assay was assessed with 106 bacterial strains that closely and distantly related to the 12 pathogens ( Table 2). All bacterial cultures were inoculated into the growth media and under the appropriate growth conditions. Bacterial DNA was extracted according to the procedure described above and was used as a template in the TaqMan real-time quantitative PCR.

Construction of Standard Plasmid and Standard Curve
Standard plasmids were constructed using clonal transformation. All purified fragment was cloned into the PMD19-T vector, and then transformed into E. coli strain DH5α. Plasmids constructed were isolated from cultures using the Plasmid Mini Kit (Omega, USA) according to the manufactures instruction. Purity and concentration of the DNA were checked using a Nanodrop-2000 Spectrophotometer (NanoDrop Technologies, USA). The copy number were calculated according to the formula: 6.02 * 10 23 * (ng/µL * 10 −9 ) bp * 660 Frontiers in Microbiology | www.frontiersin.org Frontiers in Microbiology | www.frontiersin.org  Plasmid DNA containing the target amplicon was diluted to contain 10 0 -10 6 copies/µL and used as the plasmid standard series. Linear relationship between Ct (threshold cycle) and log input DNA copies was established using the real-time PCR assay. The amplification efficiencies (E) was calculated by using the slope of standard curve and applying the following equation:

Staphylococcus aureus Stored in our laboratory
TaqMan Real-Time Quantitative PCR ABI 7500 FAST Cycler was used to carry out a real-time PCR. Amplification reactions were performed in a 25 µL reaction volume involving of 12.5 µL Premix Ex Taq TM (Probe qPCR, contains TaKaRa Ex Taq HS, dNTP Mixture, Mg 2+ , Tli RnaseH, Japan), 1 µL of each primer (10 µM), 1 µL of probe (5 µM), 1 µL of genomic DNA template, and 8.5 µL of H 2 O. Template DNA used in the designed assays is from target and non-target related strains. Each sample was tested in duplicate. The optimized PCR reaction conditions was 95 • C for 30 s, followed by 40 cycles of 95 • C for 5 s, 55 • C for 10 s, and 72 • C for 30 s. During the extension phase, fluorescence intensity was measured and the data were analyzed using 7500 Software, Version 2.0.6. For all PCR assays, positive controls were run along with a negative control, which consisted of sterile water. The test result was considered to be positive when Ct value was 36 or less.

Data Analysis
Data analysis was performed by Software of ABI 7500 FAST (7500 Software, Version 2.0.6). The Ct value is the cycle number at which the amplification curve crosses the fluorescence threshold.
Within the exponential phase of the run, the threshold is set manually. A signal for non-template control (NTC) was set at 15 %, thus only an increase in fluorescence intensity of at least 15 %, was considered a positive result.

Sensitivity Assays
The minimum amount of template DNA (copies/µL) was determined using a series of 10-fold dilutions of plasmid DNA containing the target amplicon by the TaqMan real-time PCR assay. Each reaction was amplified in duplicate. There is a linear relationship that the different template DNA concentrations (log10 copies/µL) were plotted against the corresponding Ct values.

Evaluation of the Limit of Detection (LOD)
Reference strains cultured overnight with gentle shaking (220 rpm/min) were serially 10-fold diluted in 0.9% NaCl during the logarithmic growth phase. Each dilution (1 mL) of bacterial culture was incubated in order to determine its concentration by plate counts in triplicate (CFU/mL). Then the dilutions were used for DNA extraction as described above followed by absolute quantitative real-time PCR.

Spiked Food Matrices
The bacterial cultures were diluted to the desired concentration with sterile phosphate buffered saline (PBS) (from 10 7 to 10 1 CFU/mL). The precise number of CFU in the dilutions was determined by plate counting method. The serial dilution concentrations of 12 available pure cultures were spiked onto 25 g of fresh minced meat purchased from a local super market. The samples were mechanically homogenized with a stomacher blender in 225 mL buffered peptone water (BPW; Merck, Germany). Non-inoculated food was subjected to the same procedure and used as a negative control. The food products with bacterial solution concentrations ranging from 10 7 to 10 1 CFU/g were concentrated, and the DNA from 1 mL homogenate was extracted using the method described above used as a template in real-time PCR experiments.

Probe Design
A BLAST search of GenBank in silico was used to test each TaqMan assay which revealed no identical sequences other than those targeted. The TaqMan real-time PCR assays developed in this study were performed under uniform conditions. Among the 12 bacteria cultures (106 strains), none of the nontargets were amplified in the detection assays. In addition, non-target bacterial species such as Klebsiella peneumoniae, Enterobacter Sakazakii, Bacillus cereus, Pseudomonas aeruginosa, and Clostridium perfringens did not produce an amplified signal.

Genomic DNA Extraction
The extracted DNA from 106 strains by TaKaRa Genomic DNA extraction kit yielded DNA concentrations ranging from 20 to 150 ng/µL with a mean λ 260/280 of 1.77 ± 0.09.

Analytical Specificity
Based on the analysis of 106 bacterial strains, the TaqMan realtime PCR was 100% specific ( Table 2). No false positive or negative results were found in the established TaqMan assays, confirming the exclusivity.

Standard Curve and Analytical Sensitivity
Correlation coefficients (R 2 ) of standard curves constructed were >0.99 and the reaction efficiencies ranged from 92 to 105% for the genes, indicating high linearity. As reported in Table 3, the analytical sensitivity was 1 copies/µL for E. coli O157:H7, L. monocytogenes/ivanovii, β-streptococcus hemolyticus, Shigella spp., P. mirabilis, and V. fluvialis, and 10 copies/µL for S. enterica, V. parahaemolyticus, Y. enterocolitica, E. faecalis, S. aureus, and C. jejuni. The full regression lines are reported in Table 3.

Limit of Detection and Sensitivity in Spiked Food Matrices
Based on the concentration of 12 bacteria by plate counting (CFU/mL) and absolute copy numbers (copies/µL) determined from the corresponding standard curves, according to the analytical sensitivity results (copies/µL), LODs (limit of detection) of the TaqMan real-time PCR assays were equivalent to 296,500,177,56,960,830,625,520,573,161,875, and 495 CFU/mL for E. coli O157:H7, L. monocytogenes/ivanovii, S.enterica, V. parahaemolyticus, β-streptococcus hemolyticus, Y. enterocolitica, E. faecalis, Shigella spp., P. mirabilis, V. fluvialis, S. aureus, and C. jejuni, respectively. The ability of all the TaqMan assays to detect 12 strains in spiked food samples was tested using artificial spiked serial dilutions of each pathogen into meat. The TaqMan assays revealed good amplification without inhibition of the amplification. V. parahaemolyticus was detected in spiked samples in the range of 10 3 -10 7 CFU/g, and the other 11 strains were detected in concentrations ranging from 10 4 to 10 7 CFU/g.

DISCUSSION
Culture-based methods are not sufficient to detect and prevent all outbreaks of food-borne illness. Methods for rapid, sensitive, and specific detection of food-borne pathogens are required. Nucleic acid-based methods such as PCR and loop-mediated isothermal amplification (LAMP) methods have advantages in sensitivity, specificity, and speed. Among these molecular methods, realtime PCR represents a powerful tool that is more sensitive and suitable for high-throughput analysis (Kralik and Ricchi, 2017). The major advantage of real-time PCR is that less time is required during the analysis. However, it could be affected by PCR inhibitors exist in DNA extracted from food matrix (Vital et al., 2017). In fact, in the analysis process DNA extraction is the first step and high-quality DNA is the most important element to assure subsequent real-time PCR (Cremonesi et al., 2014). For complex sample matrices, automated extraction methods do not reduce the yield of target DNA compared to well-established column-based extraction. The combination of the two methods may ensure reproducibility of the analysis and improve accuracy (Frickmann et al., 2015).
In this study, we established TaqMan real-time PCR assays to simultaneously detect 12 specific food-borne bacterial pathogens including E. coli O157:H7, L. monocytogenes/ivanovii, S. enterica, V. parahaemolyticus, β-streptococcus hemolyticus, Y. enterocolitica, E. faecalis, Shigella spp., P. mirabilis, V. fluvialis, S. aureus, and C. jejuni. The assays showed strong strain specificity and exclusivity ( Table 2). The results of cultured bacteria and spiked food samples exhibited highly efficient identification. Moreover, the manual protocol reduced the analysis time to ∼1 h, compared with the 7 days or more required for the conventional culture-based methods.
Primers and probes ensure sensitive, specific, and simultaneous detection through a single set of amplification conditions. Although 16S rRNA gene has been used for identification of bacteria at the species level, there still are some limitations. The 16S rRNA gene may not be sufficiently discriminative for species differentiation, and closely related bacteria cannot be discerned because of its low rate of base substitutions (Ruan et al., 2017). Although the homology of the genomes of various genus pathogens is very high and there is no difference in the biochemical reaction between the nonpathogenic types of the same genus, the specificity of virulence genes related to pathogenicity is high. In this investigation, specific primers and probes were designed targeting bacterial virulence genes to differentiate the closely related species. Then primers and probes were optimized based on the Tm and GC content to perform experiments under the same reaction conditions. Although TaqMan probes are more costly than SYBR green, the use of the sequence specific probes ensures high sensitivity, and specificity (Postollec et al., 2011;Law et al., 2014). Some target genes selected in the present study, as well as functional genes related to virulence or metabolism, have showed good specificity toward C. jejuni, Shigella spp., and Y. enterocolitica, when compared to standard methods (Wiemer et al., 2011;Wang et al., 2014;Van Lint et al., 2015). Three specific genes, namely, rfbE from E. coli O157:H7, hilA from S. enterica and prfA from L. monocytogenes were selected in our study to design primers and probes while fliC, invA, and hlyA were used in previous studies (Zhou et al., 2017).
The development of high-throughput qPCR platforms was a breakthrough which result in promising systems that are capable of processing a large number of samples simultaneously and also to perform a large number of assays per sample. The more target genes were amplified in a single set of reaction conditions in these high-throughput qPCR formats, the more pathogens were effectively detected or identified at the same time. A laboratory-developed TaqMan Array Card for simultaneous detection of 19 enteropathogens and high-throughput assay for rapid detection of nine pathogens directly from stools with diarrhea have been reported (Antikainen et al., 2013;Liu et al., 2013). In the work of Ishii et al. microfluidic quantitative PCR (qPCR) technology was applied for the simultaneous detection and quantification of multiple food-and waterborne pathogens as L. monocytogenes, Salmonella Typhimurium, V. parahaemolyticus, and Clostridium perfringens (Ishii et al., 2013). Rapid and high-throughput identification of more food-borne bacterial pathogens by multiple analysis platforms have been developed and combination of several detection methods is also possible (Cremonesi et al., 2014;Salihah et al., 2016;Nasrabadi et al., 2017;Thomas et al., 2017;Ahn et al., 2018;Carloni et al., 2018;Zhang et al., 2018).
In conclusion, real-time PCR assays offer the possibility of rapid detection of pathogenic bacteria with higher specificity, sensitivity, and reliability than traditional culture methods. We developed high-throughput qPCR assays that can be run in identical PCR conditions allow the inclusion of large number of assays and samples in one run. These assays enable processing of several samples simultaneously through high-throughput screening of multiple pathogens that brings great saving of time. In summary, the TaqMan real-time PCR assays performed well in contrast with conventional culture-based methods, indicating that it could be an rapid and effective alternative for the identification and diagnosis of food-borne outbreaks.

AUTHOR CONTRIBUTIONS
CN and JL conceived and designed the experiments. YL, QD, TW, and CN performed the experiments. YC and CN designed the probes and primers, and analyzed the data. YL, YC, CN, and JL wrote the manuscript. All authors read and approved the final manuscript.

FUNDING
This study is supported by the National Natural Science Foundation of China (Grant No. 81402681).