Rapid Quantification of Escherichia coli O157:H7 and Salmonella Typhimurium in Lettuce Using Immunomagnetic Separation and a Microfluidic System

Abstract

Escherichia coli O157:H7 and Salmonella spp. are common foodborne pathogens. Simple, rapid and accurate methods to detect and enumerate these pathogens are required to prevent outbreaks of foodborne illness. Here, a microfluidic system with on-chip staining and semi-automated counting functionality was combined with the use of immunomagnetic separation to collect E. coli O157:H7 and Salmonella Typhimurium without an enrichment step. The recovery of bacteria from lettuce at different spiking ratios of the two species was 61–83%. In the range 3.5 × 10² to 3.3 × 10⁵ cells/g of lettuce, there was a good linear relationship (R² = 0.9997–0.9998) between the average counts from the microfluidic quantification method and conventional immunofluorescence microscopy. Detection took <3 h. The limit of detection of the microfluidic device was 4 × 10¹ cells/g, comparable to that of plate culture and real-time PCR methods. Our microfluidic approach has potential for rapid on-site detection of multiple pathogenic bacteria in foods.

INTRODUCTION

Illnesses caused by foodborne pathogens have a significant impact on public health and the economy. According to estimates produced by the WHO, unsafe food causes approximately 600 million cases of illness and 420000 deaths worldwide each year.¹⁾ Escherichia coli O157:H7 and Salmonella enterica ssp. enterica serovar Typhimurium (referred to hereafter as Salmonella Typhimurium) are two of the most common foodborne pathogens, and they are studied as models to understand bacterial behavior.^2,3) E. coli and Salmonella have the highest frequency of occurrence with the highest number of outbreak cases in Europe and North America among foodborne pathogens.⁴⁾

Culture-based methods are used widely to detect foodborne pathogens because they are well established and very reliable.⁵⁾ However, culture-based methods can take 2–3 d for bacterial identification (and quantification) because it takes a long time for visible colonies to emerge on selective media and for the necessary biochemical or molecular tests. Thus, these methods require significant labor and trained personnel. Molecular diagnostic methods, including enzyme-linked immunosorbent assay (ELISA),^6,7) enzyme-linked fluorescence assay,⁸⁾ latex agglutination assay,⁹⁾ real-time PCR,^10,11) and loop-mediated isothermal amplification,^12,13) have high sensitivities and specificities, and are rapid, resulting in a significant decrease in the period required for analysis. However, there are challenges associated with these methods, including the need for skilled laboratory personnel and inhibitory effects of food components on detection.¹⁴⁾ New rapid, sensitive and reliable enumeration methods are desirable to overcome these challenges.

In the past few decades, microfluidic devices have been developed for use in microbiology.¹⁵⁾ The advantages of microfluidic devices include rapid analysis, low sample and reagent consumption, high versatility in design, portability, and disposability.¹⁶⁾ ELISA¹⁷⁾ and real-time PCR¹⁸⁾ combined with microfluidic devices are emerging technologies for rapid and sensitive detection of bacteria. In a previous study, we fabricated a microfluidic device that included on-chip staining, microbial counting, and portability for rapid quantification of bacteria.^19,20) Our developed microfluidic device combined with filtration through a 0.2-µm filter or two-step centrifugation quantified precisely the number of E. coli O157:H7 in lettuce and beef samples.²⁰⁾ However, the limit of detection (LOD) of current methods is not sufficient for the detection of low concentrations (such as 10–100 bacterial cells/g) in food samples without a sample enrichment step. In addition, to determine if food is contaminated, it is necessary to detect multiple types (species/strains) of bacteria rather than single pathogens.

In the present study, we investigated the microfluidic bacterial detection system coupled with an immunomagnetic separation (IMS) method for quantitative detection of two major foodborne pathogenic bacteria, E. coli O157:H7 and Salmonella Typhimurium, in one serial analysis. The IMS method was used to collect E. coli O157:H7 and Salmonella Typhimurium from food samples while removing the food matrix completely. E. coli O157:H7 and Salmonella Typhimurium captured by different magnetic beads were thus separated from food samples and microfluidic quantification was performed. The recovery at different spiking ratios was evaluated by comparing the numbers of E. coli O157:H7 and Salmonella Typhimurium measured using the microfluidic device and fluorescence microscope. The quantitative range was also evaluated using E. coli O157:H7 and Salmonella Typhimurium. Detection sensitivity tests were performed by comparing LODs from the microfluidic device, plate culture, and real-time PCR.

MATERIALS AND METHODS

Bacterial Strains

E. coli O157:H7 (ATCC43888, without the Shiga toxin-producing gene) and Salmonella Typhimurium (ATCC14028) were cultivated in Soybean-Casein Digest Broth (which has identical composition to trypticase soy broth; Daigo, Nihon Pharmaceutical Co., Ltd., Tokyo, Japan) and trypticase soy broth (Becton Dickinson, Franklin Lakes, NJ, U.S.A.) at 37 °C for at least 12 h, respectively. Cultured bacteria were enumerated by immunofluorescence microscopy as described below and diluted in phosphate-buffered saline (PBS; FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) to the appropriate concentration for each experiment.

Sample Preparation

Chopped iceberg lettuce was purchased from retail stores in Osaka City, Japan. Lettuce (25 g), 50 mL of buffered peptone water (Difco Laboratories, Detroit, MI, U.S.A.) containing 0.02% Tween 20 (SERVA, Heidelberg, Germany), and diluted E. coli O157:H7 and Salmonella Typhimurium cell suspensions were placed in a sterile stomacher bag with a coarse filter function (Stoma-Filter NEO, GSI Creos, Tokyo, Japan). Samples were homogenized (Pro-media SH-IIM, ELMEX, Tokyo, Japan) for at least 1 min before collection. The suspensions passed through the filter of the stomacher bag were collected in 50-mL tubes; 100 µL and 1 mL of these suspensions were collected for plate culture and real-time PCR analysis, respectively.

In some cases, the suspensions were centrifuged at 170 × g for 10 min. Subsequently, the supernatants were placed in fresh 50-mL tubes and filtered onto a white polycarbonate membrane filter (pore size 0.2 µm, diameter 47 mm; Toyo Roshi Kaisha, Tokyo, Japan). The bacterial cells collected on the membrane filters were resuspended in 3 mL of PBS and mixed well using a vortex mixer for IMS.

Immunomagnetic Separation

Two types of magnetic bead (Dynabeads M-280 Sheep anti-Rabbit immunoglobulin G (IgG) and Dynabeads M-280 Streptavidin; Thermo Fisher Scientific, Waltham, MA, U.S.A.), coated with anti-rabbit IgG and streptavidin respectively, were used for IMS. The protocol of IMS was described in previous report.²¹⁾ Streptavidin-coated beads were stained with 6 µg of biotin-conjugated anti-E. coli O157:H7 antibody (Goat-poly; SeraCare Life Science, Milford, MA, U.S.A.) and anti-rabbit IgG-coated beads were stained with 3 µg of AlexaFluor 488 (AF488)-conjugated anti-Salmonella Typhimurium polyclonal antibody (Rabbit IgG; Bioss Antibodies, Woburn, MA, U.S.A.). Before IMS, centrifugation (6600 × g, 3 min) was performed in some cases when the spiked bacterial count was low.

In some cases, resuspended E. coli O157:H7 and Salmonella Typhimurium that had not been inoculated into food samples were collected by IMS to evaluate the efficiency of recovery.

Cell Staining and Counting by Immunofluorescence Microscopy

Details of the bacterial cell staining and counting procedure were described in a previous report.²⁰⁾ The collected bacterial cells were stained with 2 µg of fluorescein isothiocyanate (FITC)-labeled anti-E. coli O157:H7 antibody (Goat-poly; SeraCare Life Science) and 3 µg of AF488-labeled anti-Salmonella Typhimurium antibody in the dark for 15 min at room temperature. According to the antibody manufacturers, cross-reactivity with other E. coli strains and Salmonella species has been minimized.

Cell Counting Using a Microfluidic System

The collected cells were stained and counted by using the on-chip staining²²⁾ and portable microfluidic bacterial quantification system (Supplementary Fig. S1). Details of the fabrication of the microfluidic device and the protocol for cell counting were described in previous reports.^20,21) FITC-labeled anti-E. coli O157:H7 antibody and 3 µg of AF488-labeled anti-Salmonella Typhimurium antibody were used as staining dyes. The flow-rate of the sample and staining fluid was 0.05 µL/min, and that of the sheath fluid was 0.01 µL/min. The amount of E. coli O157:H7 and Salmonella Typhimurium in each sample is presented in cells/mL, as determined from the cell count and flow volume.

Detection Sensitivity Test of E. coli O157:H7 in Lettuce

One hundred microliters of sample (after homogenization as described above) were spread over sorbitol MacConkey agar supplemented with cefixime (0.05 mg/mL) and potassium tellurite (2.5 mg/L; Nissui Pharmaceutical Co., Tokyo, Japan). The plates were incubated at 37 °C for 24 h, and the presence of transparent colonies of E. coli O157:H7 on the plates was confirmed visually.

One milliliter of sample (after homogenization) was subjected to genomic DNA extraction using a silica gel membrane spin column from the Qiagen DNeasy Blood & Tissue Kit (Qiagen, Valencia, CA, U.S.A.) according to the manufacturer’s instructions (final elution volume 400 µL). The concentration and purity of the extracted genomic DNA were estimated using a Multiskan SkyHigh spectrometer (Thermo Fisher Scientific). The primer and probe sequences used in real-time PCR assay were 5′-TTTCA CACTT ATTGG ATGGT CTCAA-3′ (forward), 5′-CGATG AGTTT ATCTG CAAGG TGAT-3′ (reverse), and 5′-6-carboxyfluorescein–AGGAC CGCAG AGGAA AGAGA GGAAT TAAGG–5-carboxytetramethylrhodamine-3′ (probe), which targets the rfbE gene encoding the lipopolysaccharide O antigen of E. coli O157:H7.^11,23) Real-time PCR was performed in a final volume of 10 µL, including 1 µL of extracted DNA, 0.12 µL of each primer (final concentration 60 nM), 0.12 µL of probe (final concentration 240 nM), 5 µL of FastStart Universal Probe Master (ROX, Roche Diagnostics Japan, Tokyo, Japan), and 3.64 µL of ribonuclease- and deoxyribonuclease-free water. The real-time PCR reactions were performed using a LightCycler^® 96 system (Roche Diagnostics Japan) with conditions 95 °C for 2 min, followed by 40 cycles of 95 °C for 15 s and 60 °C for 1 min.

Statistical Analysis

All data are expressed as the mean ± standard deviation of at least four independent determinations, except for the comparisons of detection sensitivity, for which n = 2. All statistical analyses were performed using Microsoft Excel. Real-time PCR data were analyzed using LightCycler^® 96 SW 1.1 software. The statistical significance of differences between experimental groups was analyzed using Student’s t-test, and a probability value of 0.05 was used as the criterion for significance.

RESULTS AND DISCUSSION

Collection of E. coli O157:H7 and Salmonella Typhimurium from Lettuce Samples Using the IMS Method

For successful quantification of foodborne pathogens using the microfluidic system, the effect of matrix contaminants from food flowing in the small chamber of the microfluidic device must be minimized to prevent clogging of the flow path in the microfluidic device and autofluorescence of the food matrix. IMS using magnetic nanoparticles enables one to specifically concentrate bacteria and is widely used in biomedicine²⁴⁾ and microbiology.¹⁵⁾ In this study, two types of magnetic beads were used, respectively coupled with specific antibodies for E. coli O157:H7 and Salmonella Typhimurium, to collect each type of bacterial cell. The recovery of the individual bacterial species without food inoculation was 70 ± 10% (E. coli O157:H7) and 69 ± 12% (Salmonella Typhimurium) (Table 1).

Table 1. Numbers and Recovery of Escherichia coli O157:H7 and Salmonella Typhimurium Cells Collected by Immunomagnetic Separation and Then Determined by Fluorescence Microscopy

	E. coli O157:H7	S. Typhimurium
Number of spiked cells	1.0 × 107	1.0 × 107
Number of recovered cells	7.0 (± 1.0) × 106	6.9 (± 1.2) × 106
Recovery (%)	70 (± 10)	69 (± 12)

Results shown are means ± standard deviations (n = 4).

Cross-contamination with multiple types of bacteria occurs in fresh produce, including vegetables such as lettuce, and is recognized as one of the main contributors to fresh-produce-associated disease outbreaks.²⁵⁾ Thus, we selected lettuce as a model contaminated with more than one foodborne pathogen. E. coli O157:H7 and Salmonella Typhimurium, whose cell numbers were counted by fluorescence microscopy, were inoculated onto lettuce individually or together and collected using the IMS method. The recoveries of E. coli O157:H7 and Salmonella Typhimurium when they were inoculated onto lettuce as the sole bacterial strain were 78 ± 6.0% (E. coli O157:H7) and 84 ± 5.6% (Salmonella Typhimurium), respectively (Table 2). The recoveries at different spiking ratios (E. coli O157:H7 to Salmonella Typhimurium) were 75 ± 10% (E. coli O157:H7) and 61 ± 11% (Salmonella Typhimurium) at spiking ratio 1 : 1 (cell : cell); 80 ± 7.4% (E. coli O157:H7) and 77 ± 2.7% (Salmonella Typhimurium) at spiking ratio 1 : 4; and 78 ± 3.6% (E. coli O157:H7) and 83 ± 6.7% (Salmonella Typhimurium) at spiking ratio 4 : 1 (Table 2). Comparisons of immunofluorescence microscopy images before and after IMS showed that the food matrix was effectively removed from the cell suspension collected using the IMS method (Fig. 1).

Table 2. Numbers and Recovery of Escherichia coli O157:H7 and Salmonella Typhimurium Cells Inoculated into Lettuce Samples at Different Ratios, Collected by Immunomagnetic Separation, and Then Determined by Fluorescence Microscopy

Spiking ratio*	Number of spiked cells		Number of collected cells		Recovery (%)
	E. coli O157:H7	S. Typhimurium	E. coli O157:H7	S. Typhimurium	E. coli O157:H7	S. Typhimurium
1 : 0	1.0 × 107	—	8.0 (± 0.6) ×106	—	78 (± 6.0)	—
1 : 1	1.0 × 107	1.0 × 107	7.5 (± 1.0) ×106	6.1 (± 1.1) ×106	75 (± 10)	61 (± 11)
1 : 4	1.0 × 107	4.0 × 107	8.1 (± 0.7) ×106	3.1 (± 0.1) ×107	80 (± 7.4)	77 (± 2.7)
4 : 1	4.0 × 107	1.0 × 107	3.1 (± 0.1) ×107	8.4 (± 0.1) ×106	78 (± 3.6)	83 (± 6.7)
0 : 1	—	1.0 × 107	—	8.4 (± 0.6) ×106	—	84 (± 5.6)

*Ratio E. coli O157:H7 to Salmonella Typhimurium. Results are means ± standard deviations (n = 4).

Fig. 1. Fluorescence Images of Escherichia coli O157:H7 and Salmonella Typhimurium Spiked into Lettuce Samples and Then Collected Using Immunomagnetic Separation

E. coli O157:H7 and Salmonella Typhimurium spiked into lettuce samples were collected using centrifugation and filtration (a). In addition, these cells were purified using the immunomagnetic separation method (b). Lettuce samples without bacterial spiking (c) were treated by centrifugation, filtration, and immunomagnetic separation. Collected cells were stained with E. coli O157:H7-specific fluorescein isothiocyanate-labeled antibodies and Salmonella Typhimurium-specific Alexa Fluor 488-labeled antibodies and observed by fluorescence microscopy.

IMS is used to concentrate bacteria from large-volume samples into a small volume that (1) shortens the time required for detection, and (2) decreases the effective LOD. In this study, we efficiently collected E. coli O157:H7 and Salmonella Typhimurium from lettuce within 1 h by using the two types of magnetic bead. Recently, automated IMS combined with a colorimetric assay²⁶⁾ and a microfluidic device included in the IMS system²⁷⁾ were developed. The use of automated IMS techniques will lead to simpler and more tractable detection methods.

Enumeration of E. coli O157:H7 and Salmonella Typhimurium in Lettuce Samples Using the Microfluidic Device with On-Chip Staining

E. coli O157:H7 and Salmonella Typhimurium collected using the IMS method were quantified using an on-chip type microfluidic system; the data were compared with enumeration by conventional immunofluorescence microscopy. As shown in Table 3, the counts of E. coli O157:H7 and Salmonella Typhimurium obtained by the microfluidic system and immunofluorescence microscopy were not significantly different at any tested spiking ratio of the two species (E. coli O157:H7 : Salmonella Typhimurium = 1 : 1, p = 0.15; 1 : 4, p = 0.98; 4 : 1, p = 0.089).

Table 3. Numbers of Escherichia coli O157:H7 and Salmonella Typhimurium Cells in Lettuce Samples Collected Using Immunomagnetic Separation and Determined by Fluorescence Microscopy and the Microfluidic System

Spiking ratio*	Microscopic count (cells)	Microfluidic count (cells)
1 : 1	1.4 (± 0.2) × 107	1.6 (± 0.2) × 107
1 : 4	3.9 (± 0.2) × 107	4.3 (± 0.2) × 107
4 : 1	3.9 (± 0.1) × 107	4.1 (± 0.2) × 107

*Ratio E. coli O157:H7 to Salmonella Typhimurium. Results are means ± standard deviations (n = 4).

The reliability of a measurement system is related to its quantification range. The counts of E. coli O157:H7 and Salmonella Typhimurium obtained by the microfluidic system and immunofluorescence microscopy were closely correlated in the range 10²–10⁵ bacterial cells/g of lettuce (Figs. 2, 3). For these low levels of bacteria (10²–10⁵ target bacterial cells/g in a lettuce sample), precentrifugation (4–67-fold concentration) was effective to enhance the number of bacteria flowing in the microchannel of the device. Although not performed in this study, high levels of bacteria (>10⁷ cells of target bacteria/g sample) require dilution because bacterial cells are not accurately counted by the portable system if too many bacterial cells flow rapidly in the microchannel of the device. Where the levels of bacteria in contaminated food were unknown, the presence of bacteria was identified by using the microfluidic device after a precentrifugation step (for the purpose of screening, i.e., preliminary detection of target bacteria), and adjustment of bacterial concentration can be performed if necessary to quantify the target bacteria.

Fig. 2. Comparison of Microfluidic and Conventional Fluorescence Microscopy Counts of E. coli O157:H7 Spiked into and Collected from Lettuce Samples

Serially diluted E. coli O157:H7 spiked into lettuce samples was collected using the immunomagnetic separation method. The cells were stained with E. coli O157:H7-specific fluorescein isothiocyanate-labeled antibody and enumerated by fluorescence microscopy or using the microfluidic system. The collected cells were concentrated 4–67-fold (depending on the spiking level) before enumeration.

Fig. 3. Comparison of Microfluidic and Conventional Fluorescence Microscopy Counts of Salmonella Typhimurium Spiked into and Collected from Lettuce Samples

Serially diluted Salmonella Typhimurium spiked into lettuce samples was collected using the immunomagnetic separation method. The cells were stained with Salmonella Typhimurium-specific Alexa Fluor 488-labeled antibody and enumerated by fluorescence microscopy or using the microfluidic system. The collected cells were concentrated 4–67-fold (depending on the spiking level) before enumeration.

Our microfluidic detection system captured the two stained bacterial species respectively bound to FITC- or AF488-conjugated antibodies. This method is thus effective when rapidly detecting multiple pathogens in food with the aim of screening for bacterial contamination. Using further fluorescent antibodies will increase the numbers of target bacterial species. However, our microfluidic detection system was constructed with a single-color laser and different target bacteria bound to antibodies with differently emitting fluorescent dyes cannot be detected by the system. Thus, our microfluidic method detects the total number of target bacteria using same fluorescent dye-conjugated antibody. Our method is effective as an initial screening approach for on-site assessment of contamination by foodborne pathogens. If foodborne pathogens are detected by our method, culture or other methods can be applied to determine the presence of particular species. Another limitation of our microfluidic method is that it cannot target bacteria for which antibodies have not been developed. Song et al. developed a single-stranded DNA aptamer that has broad reactivity with multiple bacteria by using a sequential toggle cell-SELEX method.²⁸⁾ Such aptamers conjugated with fluorescent dye overcome the drawback of needing to use multiple antibodies.

Detection Sensitivity Test of Microfluidic Device in Detection of E. coli O157:H7

The detection sensitivity of microfluidic detection from lettuce samples after IMS was compared with those of the bacterial plate counting method and real-time PCR without prior enrichment. Note, however, that the plate counting and real-time PCR methods used in this study were different from conventional plate counting and real-time PCR as described in various guidelines for detecting foodborne bacteria from food samples. Bacteria in lettuce samples inoculated with E. coli O157:H7 at 4 × 10⁰ to 4 × 10⁵ cells/g were enumerated using these methods (n = 2). The presence of 4 × 10¹ E. coli O157:H7 cells/g in lettuce samples was detectable using the IMS method and microfluidic system (Table 4). The real-time PCR method detected the presence of 4 × 10² E. coli O157:H7 cells/g in lettuce samples. The plate culture method detected the presence of 4 × 10² E. coli O157:H7 cells/g in lettuce samples (4 × 10¹ E. coli O157:H7 cells/g gave one positive and one negative result, n = 2). These results suggest that the LOD of the microfluidic detection system combined with IMS is comparable to those of real-time PCR and plate culture. However, the LODs would be reduced if pre-enrichment was performed before plating of bacteria (for the culture method) or before DNA extraction (for real-time PCR). Moreover, in these experiments, we evaluated only one bacterial species as a simple comparison of the microfluidic detection system with the real-time PCR and plate culture methods. Further comparisons using various bacterial species and environmental conditions are necessary to verify that our proposed method is more sensitive than other methods. A short-term pre-enrichment step in BPW-T medium will lead to the ability to detect low numbers of target bacteria using our microfluidic technique.

Table 4. Comparison of Conventional and Microfluidic Methods for the Detection of Escherichia coli O157:H7 Inoculated into Lettuce

Number of spiked cells (cells/g)	Real-time PCR	Plate culture	Microfluidic method
4 × 105	+	+	+
4 × 104	+	+	+
4 × 103	+	+	+
4 × 102	+	+	+
4 × 101	−	±	+
4 × 100	−	−	−

Detection results: +, E. coli O157:H7 detected; ±, different results for individual experiments; −, E. coli O157:H7 not detected (n = 2).

In this study, the LOD using our microfluidic technique was 3.6 × 10¹ cells/g and the required time was 3 h. This was better than the LOD of our previous method (1.1 × 10⁵ cells/g) without pre-enrichment and concentration steps.²⁰⁾ Our microfluidic method shows comparable performance with plate counting and current rapid and sensitive detection methods such as real-time PCR. However, the plate counting method evaluates numbers of colony-forming bacteria, which differs from our method (direct counting of flowing, stained bacteria). Foodborne pathogens such as Campylobacter jejuni and E. coli O157:H7 are rendered viable but nonculturable by long-term low-temperature storage.²⁹⁾ Direct counting of stained bacteria enables detection of such viable but nonculturable bacteria that cannot be detected by culture-dependent methods. Our method can be developed to evaluate the presence of viable but nonculturable bacteria by combination with fluorescent staining with 5-cyano-2,3-ditolyl-tetrazolium chloride, which stains respiring bacterial cells.

CONCLUSION

In this study, we propose a rapid and easy detection method for E. coli O157:H7 and Salmonella Typhimurium using IMS and a microfluidic system. The advantage of the IMS method and microfluidic system is the detection of target bacteria in a semi-automated system (for on-chip staining and counting), with portability (for on-site measurement). Our method can collect and purify foodborne pathogens from samples by removing food matrices and does not require sophisticated or long analysis procedures such as in traditional culture-based analysis. The target bacteria in food samples were effectively separated and concentrated from the food matrix by the IMS process using target-specific immunomagnetic beads. The presence of foodborne pathogens in food was confirmed within 3 h without a pre-enrichment process. There was a good linear relationship (R² = 0.9997–0.9998) between the average values determined using the microfluidic quantification method and conventional immunofluorescence microscopy at cell concentrations from 3.5 × 10² to 3.3 × 10⁵ cells/g of lettuce. We demonstrate that our technique can detect target bacteria at just 4 × 10¹ cells/g and quantify cells present at >4 × 10² cells/g in lettuce. As a result, we anticipate that our microfluidic method will prove valuable for on-site detection to identify and quantify pathogenic bacteria in foods.

Acknowledgments

This research was supported by Food Research Grants from the Toyo Institute of Food Technology. We thank Mamoru Saito, Ph.D., for helping to fabricate the microfluidic cell counting system.

Conflict of Interest

The authors declare no conflict of interest.

Supplementary Materials

This article contains supplementary materials.

REFERENCES

• 1) World Health Organization. “WHO estimates of the global burden of foodborne diseases: foodborne diseases burden epidemiology reference group 2007–2015.”: ‹https://www.who.int/publications/i/item/9789241565165›, accessed 6 January, 2024.
• 2) Xu M, Wang R, Li Y. Rapid detection of Escherichia coli O157:H7 and Salmonella Typhimurium in foods using an electrochemical immunosensor based on screen-printed interdigitated microelectrode and immunomagnetic separation. Talanta, 148, 200–208 (2016).
• 3) Yang L, Li Y. Simultaneous detection of Escherichia coli O157:H7 and Salmonella Typhimurium using quantum dots as fluorescence labels. Analyst, 131, 394–401 (2006).
• 4) Aiyedun SO, Onarinde BA, Swainson M, Dixon RA. Foodborne outbreaks of microbial infection from fresh produce in Europe and North America: a systematic review of data from this millennium. Int. J. Food Sci. Technol., 56, 2215–2223 (2021).
• 5) Wang Y, Salazar JK. Culture-independent rapid detection methods for bacterial pathogens and toxins in food matrices. Compr. Rev. Food Sci. Food Saf., 15, 183–205 (2016).
• 6) Cavaiuolo M, Paramithiotis S, Drosinos EH, Ferrante A. Development and optimization of an ELISA based method to detect Listeria monocytogenes and Escherichia coli O157 in fresh vegetables. Anal. Methods, 5, 4622–4627 (2013).
• 7) Zhu L, He J, Cao X, Huang K, Luo Y, Xu W. Development of a double-antibody sandwich ELISA for rapid detection of Bacillus Cereus in food. Sci. Rep., 6, 16092 (2016).
• 8) Rozand C, Feng PC. Specificity analysis of a novel phage-derived ligand in an enzyme-linked fluorescent assay for the detection of Escherichia coli O157:H7. J. Food Prot., 72, 1078–1081 (2009).
• 9) Medina MB, Shelver WL, Fratamico PM, Fortis L, Tillman G, Narang N, Cray WC Jr, Esteban E, Debroy A. Latex agglutination assays for detection of non-O157 Shiga toxin-producing Escherichia coli serogroups O26, O45, O103, O111, O121, and O145. J. Food Prot., 75, 819–826 (2012).
• 10) Kim JH, Oh SW. Optimization of bacterial concentration by filtration for rapid detection of foodborne Escherichia coli O157:H7 using real-time PCR without microbial culture enrichment. J. Food Sci., 84, 3241–3245 (2019).
• 11) Jenkins C, Lawson AJ, Cheasty T, Willshaw GA. Assessment of a real-time PCR for the detection and characterization of verocytotoxigenic Escherichia coli. J. Med. Microbiol., 61, 1082–1085 (2012).
• 12) Maruyama F, Kenzaka T, Yamaguchi N, Tani K, Nasu M. Detection of bacteria carrying the stx2 gene by in situ loop-mediated isothermal amplification. Appl. Environ. Microbiol., 69, 5023–5028 (2003).
• 13) Techathuvanan C, Draughon FA, D’Souza DH. Loop-mediated isothermal amplification (LAMP) for the rapid and sensitive detection of Salmonella Typhimurium from pork. J. Food Sci., 75, M165–M172 (2010).
• 14) Siala M, Barbana A, Smaoui S, Hachicha S, Marouane C, Kammoun S, Gdoura R, Messadi-Akrout F. Screening and detecting Salmonella in different food matrices in Southern Tunisia using a combined enrichment/real-time PCR method: correlation with conventional culture method. Front. Microbiol., 8, 2416 (2017).
• 15) Kant K, Shahbazi MA, Dave VP, Ngo TA, Chidambara VA, Than LQ, Bang DD, Wolff A. Microfluidic devices for sample preparation and rapid detection of foodborne pathogens. Biotechnol. Adv., 36, 1003–1024 (2018).
• 16) Zhang RQ, Liu SL, Zhao W, Zhang WP, Yu X, Li Y, Li AJ, Pang DW, Zhang ZL. A simple point-of-care microfluidic immunomagnetic fluorescence assay for pathogens. Anal. Chem., 85, 2645–2651 (2013).
• 17) Panraksa Y, Jang I, Carrell CS, Amin AG, Chailapakul O, Chatterjee D, Henry CS. Simple manipulation of enzyme-linked immunosorbent assay (ELISA) using an automated microfluidic interface. Anal. Methods, 14, 1774–1781 (2022).
• 18) Hua Z, Rouse JL, Eckhardt AE, Srinivasan V, Pamula VK, Schell WA, Benton JL, Mitchell TG, Pollack MG. Multiplexed real-time polymerase chain reaction on a digital microfluidic platform. Anal. Chem., 82, 2310–2316 (2010).
• 19) Yamaguchi N, Tokunaga Y, Goto S, Fujii Y, Banno F, Edagawa A. Rapid on-site monitoring of Legionella pneumophila in cooling tower water using a portable microfluidic system. Sci. Rep., 7, 3092 (2017).
• 20) Tokunaga Y, Yamaguchi N. Rapid quantification of Escherichia coli O157:H7 in lettuce and beef using an on-chip staining microfluidic device. J. Food Saf., 40, e12851 (2020).
• 21) Tokunaga Y, Wakabayashi Y, Yonogi S, Saito M, Yamaguchi N. Microfluidic rapid quantification of Salmonella enterica serovar Typhimurium collected from chicken meat using immunomagnetic separation after formaldehyde treatment. Int. J. Food Sci. Technol., 56, 5402–5408 (2021).
• 22) Yamaguchi N, Fujii Y. Rapid on-site monitoring of bacteria in freshwater environments using a portable microfluidic counting system. Biol. Pharm. Bull., 43, 87–92 (2020).
• 23) Savoye F, Feng P, Rozand C, Bouvier M, Gleizal A, Thevenot D. Comparative evaluation of a phage protein ligand assay with real-time PCR and a reference method for the detection of Escherichia coli O157:H7 in raw ground beef and trimmings. J. Food Prot., 74, 6–12 (2011).
• 24) Vidlarova M, Rehulkova A, Stejskal P, Prokopova A, Slavik H, Hajduch M, Srovnal J. Recent advances in methods for circulating tumor cell detection. Int. J. Mol. Sci., 24, 3902 (2023).
• 25) Possas A, Pérez-Rodríguez F. New insights into cross-contamination of fresh-produce. Curr. Opin. Food Sci., 49, 100954 (2023).
• 26) Park JY, Park K, Ok G, Chang HJ, Park TJ, Choi SW, Lim MC. Detection of Escherichia coli O157:H7 using automated immunomagnetic separation and enzyme-based colorimetric assay. Sensors (Basel), 20, 1395 (2020).
• 27) de Oliveira TR, Martucci DH, Faria RC. Simple disposable microfluidic device for Salmonella Typhimurium detection by magneto-immunoassay. Sens. Actuators B Chem., 255, 684–691 (2018).
• 28) Song MY, Nguyen D, Hong SW, Kim BC. Broadly reactive aptamers targeting bacteria belonging to different genera using a sequential toggle cell-SELEX. Sci. Rep., 7, 43641 (2017).
• 29) Dong K, Pan H, Yang D, Rao L, Zhao L, Wang Y, Liao X. Induction, detection, formation, and resuscitation of viable but non-culturable state microorganisms. Compr. Rev. Food Sci. Food Saf., 19, 149–183 (2020).