Multiplexed Bead-Based Mesofluidic System for Detection of Food-Borne Pathogenic Bacteria

Abstract

In the present study, a simple and rapid multiplexed bead-based mesofluidic system (BMS) was developed for simultaneous detection of food-borne pathogenic bacteria, including Staphylococcus aureus, Vibrio parahaemolyticus, Listeria monocytogenes, Salmonella, Enterobacter sakazakii, Shigella, Escherichia coli O157:H7, and Campylobacter jejuni. This system is based on utilization of isothiocyanate-modified microbeads that are 250 μm in diameter, which were immobilized with specific amino-modified oligonucleotide probes and placed in polydimethylsiloxane microchannels. PCR products from the pathogens studied were pumped into microchannels to hybridize with the oligonucleotide-modified beads, and hybridization signals were detected using a conventional microarray scanner. The short sequences of nucleic acids (21 bases) and PCR products characteristic of bacterial pathogens could be detected at concentrations of 1 pM and 10 nM, respectively. The detection procedure could be performed in less than 30 min with high sensitivity and specificity. The assay was simple and fast, and the limits of quantification were in the range from 500 to 6,000 CFU/ml for the bacterial species studied. The feasibility of identification of food-borne bacteria was investigated with samples contaminated with bacteria, including milk, egg, and meat samples. The results demonstrated that the BMS method can be used for effective detection of multiple pathogens in different foodstuffs.

Bacterial food-borne pathogens pose a significant threat to human and animal heath. These organisms are the leading causes of illness and death in less developed countries, killing approximately 1.8 million people annually, and they are the third leading cause of death. In developed countries, food-borne pathogens are responsible for millions of cases of infectious gastrointestinal diseases each year, costing billions of dollars in medical care and lost productivity (42). The Centers for Disease Control and Prevention estimates that food-borne diseases cause approximately 76 million illnesses, 325,000 hospitalizations, and 5,000 deaths in the United States each year (25). The Food and Drug Administration's 2005 Food Code states that the estimated cost of food-borne illness is $10 billion to $83 billion annually (29). The major food-borne pathogens include Salmonella, Listeria monocytogenes, Staphylococcus aureus, Vibrio parahaemolyticus, Campylobacter jejuni, Escherichia coli O157:H7, Enterobacter sakazakii, and Shigella (14, 32, 41).

Rapid detection and identification of pathogens and other microbial contaminants in food are needed by the food industry, food safety agencies, and public health bureaus. Traditional methods to detect food-borne bacteria often rely on time-consuming growth in culture media, isolation of bacteria, biochemical identification, and sometimes serology (5, 8). There have been many attempts to develop faster, more convenient, more sensitive, and more specific techniques for detection and diagnosis of pathogenic bacteria, including immunological methods, biosensors, and microarray technologies. Although immunological methods, such as enzyme immunoassays (33), immunofluorescence techniques (13), and enzyme-linked immunosorbent assays (2, 21), are specific and sensitive, they often result in many false positives due to washing and cross-reactions. Several nucleic acid-based methods have been developed for rapid and simultaneous detection of multiple pathogenic bacteria (7, 30). PCR is the method used most commonly for specific detection of food-borne pathogens (12, 16, 31, 40). The biosensor method (24, 28) and oligonucleotide microarray technology have been used for analysis of microbial pathogens (4, 6, 18, 26, 36). Despite the recent advances in detection of food pathogens, there are still many challenges and opportunities for improving the current technologies, including improving the integration and automation of operations, increasing the throughput and robustness, and decreasing the cost. Therefore, it is highly desirable to develop a method that can provide simple, practical, and high-throughput routine detection of pathogens in food samples.

Over the past decade, in order to improve the microanalysis method, much effort has been devoted to continued development of the micro total analysis system concept, such as a laboratory on a chip (39). So far, microfluidic and mesofluidic chips have been developed and used for various biological and chemical processes by taking advantage of automation, large surface-to-volume ratios, low solvent consumption, sensitivity, short separation time, miniaturization, and portability, and these chips can be controlled by the fluid velocity (3, 9, 11, 17). Mesofluidic chip (channel diameter, >100 μm) analysis provides better fluidics control and maneuverability than the microfluidic system (channel diameter, <1 μm). It allows accurate control of required conditions on beads and flow of reagents to facilitate the hybridization reactions using a peristaltic pump and an injection pump. Here, we report development of a simple, multiplexed, bead-based mesofluidic system (BMS) for simultaneous detection of eight of the major food-borne pathogens. The BMS method is based on nucleic acid hybridization on the surface of different beads in polydimethylsiloxane (PDMS) microchannels, which makes it suitable for high throughput and parallel analysis of food samples (Fig. 1). First, eight specific oligonucleotide probes corresponding to eight different bacterial pathogens were designed. Glass microbeads (diameter, 250 μm) precoated with the different probes were arranged in the PDMS microchannels (diameter, 300 μm) using a predetermined order (NUC, VP, PRS, C20, SG, IAL, HLY, and HIP beads). Then the fluorescence-labeled PCR products of target genes amplified from the pathogens studied were infused into the microchannels, hybridized, and captured by the corresponding probes on the beads. The beads were scanned, and the fluorescence intensities were employed to identify the pathogens. The complete procedure can be completed within 30 min. The analytical performance of the method in terms of specificity, sensitivity, and preliminary validation is discussed below. Moreover, 184 contaminated food samples were analyzed to illustrate use of this method.

FIG. 1.

Operation of the BMS. (A) Profile of microchannels filled with glass beads; (B) side view of microchannels formed by PDMS and slides; (C) schematic diagram of the flow in the BMS. The arrows indicate the direction of fluid flow.

MATERIALS AND METHODS

Reagents and materials.

Glass microbeads (average diameter, 250 μm), 3-aminopropyltrimethoxysilane, a 2,4,6-trinitrobenzene sulfonic acid solution, and 1,4-phenylene diisothiocyanate were purchased from Sigma-Aldrich (St. Louis, MO). PDMS was purchased from Dow Corning (Midland, MI). Other chemicals and solvents were analytical grade and were purchased from Sinopharm Chemcial Reagent Co. Ltd. (Shanghai, China). All reagents were used directly without additional purification. The water was double distilled, processed to 18.2 MΩ with a Milli-Q water purification system (Millipore, Bedford, MA), and sterile. Oligonucleotide probes and primers (Table 1) were purchased from Songon Inc. (Shanghai, China), synthesized by standard phosphoramidite chemistry, and purified by reverse-phase high-performance liquid chromatography. Some strains, purified genomic DNA, and the food samples were kindly provided by Shanghai Entry-Exit Inspection & Quarantine Bureau and Shanghai Jiao Tong University. Purified genomic DNA from 27 reference strains, including Salmonella sp. strains CMCC50098, CMCC50001, CMCC50004, CMCC50017, CMCC50335, and CMCC50770, S. aureus ATCC 8095, ATCC 6538, ATCC 13565, ATCC 27664, ATCC 12600, ATCC 25923, ATCC 27661, and ATCC 29213, L. monocytogenes ATCC 7644, ATCC 27708, ATCC 13313, and ATCC 13932, V. parahaemolyticus ATCC 17802 and ATCC 33846, Shigella sp. strains CMCC51334, AS1.1869, and AS1.1868, E. sakazakii ATCC 29544 and ATCC 50205, E. coli O157:H7 strain ATCC 43889, and C. jejuni ATCC 29428, were used in this study.

TABLE 1.

Sequences of primers and probes for eight pathogens

Pathogen	Primer			Probe
	Primer	Sequence (5′-3′)a	GenBank accession no.	Probe	Sequence (5′-3′)b
S. aureus	NUC-F	TTGGTAGCAGCGGTAGAGTTGATCATTATTGTAGGTGTATTAGC	BA000017	NUC	GTCCCCTTTTTGAAAGGACC
	NUC-R	ATGGTGTAAACTTGTACCAGCAGGCGTATTCGGTTTC
V. parahaemolyticus	VP-F	ATGGTGTAAACTTGTACCAGAACTGACGCTTGTTGAGG	BA000031	VP	CAGTGAAGAGCACGGTTTCA
	VP-R	TTGGTAGCAGCGGTAGAGTTGAAACTCCTGCCTGACGAT
L. monocytogenes	PRS-F	ATGGTGTAAACTTGTACCAGGCTGAAGAGATTGCGAAAGAAG	EF110565	PRS	CGAAACTGCTGGTGCAACTA
	PRS-R	TTGGTAGCAGCGGTAGAGTTGCAAAGAAACCTTGGATTTGCGG
Salmonella	C20-F	ATGGTGTAAACTTGTACCAGACCGCTGGTGAAACGACA	AE008786	C20	TAACCTCCTCTTTCCAGCGA
	C20-R	TTGGTAGCAGCGGTAGAGTTGGCGACGGCAGTGCTTATT
E. sakazakii	SG-F	ATGGTGTAAACTTGTACCAGGGGTTGTCTGCGAAAGCGAA	AY702093	SG	CGTAATAAGAAATGCGCGGT
	SG-R	TTGGTAGCAGCGGTAGAGTTGGTCTTCGTGCTGCGAGTTTG
Shigella	IAL-F	TTGGTAGCAGCGGTAGAGTTGCTGGATGGTATGGTGAGG	AY206439	IAL	TGTCCATCAAACCCCACTCT
	IAL-R	ATGGTGTAAACTTGTACCAGGGAGGCCAACAATTATTTCC
E. coli O157:H7	HLY-F	ATGGTGTAAACTTGTACCAGCAGTAGGGAAGCGAACAGAG	AY495950	HLY	ACAGGAGGAAGCGGTAATGA
	HLY-R	TTGGTAGCAGCGGTAGAGTTGAAGCTCCGTGTGCCTGAAGC
C. jejuni	HIP-F	TTGGTAGCAGCGGTAGAGTTGGGCAATGATAGAAGATGG	Z36940	HIP	TTTGCCTTTTCTGGAGCACT
	HIP-R	ATGGTGTAAACTTGTACCAGATTAGCCTGTGCAAGACC

^aUnderlining indicates sequences that are universal primers for the second round of PCR amplification.

^bEach oligonucleotide probe is modified with 16-mer poly(T) and an amino group at 5′ end.

Bead modification and oligonucleotide probe immobilization.

The isothiocyanate-modified beads and probe-immobilized beads were prepared using a previously described method (34). Prepared probe-modified beads were stored in a vacuum at room temperature.

BMS.

The operating principle of the BMS is shown schematically in Fig. 1. The core component of the BMS was the PDMS mesofluidic reaction chamber, which was filled with glass beads in PDMS microchannels. A silicon wafer with a pattern made of SU-8 by photolithography was used to cast the PDMS mesofluidic mold. PDMS was mixed well with a curing agent at a ratio of 10:1 (wt/wt), and then the mixture was poured onto the silicon wafer, which had been fumigated by fluoroalkyl silanes. Subsequently, the master mold with the PDMS was then placed in a vacuum desiccator for approximately 15 min to help remove air bubbles from the PDMS that were introduced when the curing agent was stirred in. Then the master mold with PDMS was removed from the desiccator and placed on an 80°C hot plate for 1 h to cure. After the PDMS on the master mold cured, it was lifted off. Then the PDMS device was bonded to a glass slide using an oxygen plasma bonder (PDC-002; Harrick Scientific Corp., United States). Inside the plasma bonder, the bonding surfaces of the slide (25 mm by 75 mm by 1 mm) and the PDMS chip were exposed to high-energy plasma, which stripped away electrons on the surface. This caused the surfaces to become hydrophilic. When the two hydrophilic surfaces came into contact, they formed a strong bond. The integrated mesofluidic chamber device was fabricated by sealing with a glass slide, which contained access holes for connection of mesofluidic fittings (Fig. 1A and B). The entire process was performed using the methods described previously (43).

The BMS was assembled with a programmable multichannel peristaltic pump (KH-07550; Cole-Parmer, United States), a valve, the mesofluidic chamber, and some reagent vessels using silicone tubing (Fig. 1C).

The entire analysis procedure in the mesofluidic system, including bead loading, washing, hybridization, and reagent injection, was automatically carried out by using a multichannel peristaltic pump. First, the glass beads attached with oligonucleotide probes were pumped in triplicate into the PDMS microchannels in a predetermined order. The inlet capillary was connected to the sample reservoir, and the outlet capillary was connected to the waste reservoir through the peristaltic pump. The inner wall of the microchannels and the glass beads were washed with 0.2% sodium dodecyl sulfate (SDS) and double-distilled H₂O for 10 min, respectively. The short sequences of nucleic acids (21 bases) or PCR products in hybridization solution (4× SSC-0.1% SDS [1× SSC is 0.15 M NaCl plus 0.0.15 M sodium citrate]) were injected into the microchannels, and the hybridization fluid was allowed to flow back and forth in the microchannels at 42°C for several minutes. After the reaction, 1× SSC-0.03% SDS, 0.2× SSC, and 0.05× SSC were sequentially infused again into the microchannels to wash the beads (5 min each). The flow rate of the solution was 20 μl/min.

Detection of complementary oligonucleotides.

One 21-mer synthetic target oligonucleotide, T_ial (Cy5-AGAGTGGGGTTTGATGGACAA), which was specific for the IAL probe, was selected to investigate the performance of the system. To characterize the response of the system to the synthetic target oligonucleotide T_ial, solutions with various concentrations of T_ial were flowed through the modified beads in microchannels. The hybridization time (T_ial retention time in microchannels) could be controlled using the flow rate.

Detection of PCR products of pathogens.

Eight target genes were employed to specifically identify eight pathogens as described previously (23). The prs gene encoding phosphoribosyl pyrophosphate synthetase is specific for L. monocytogenes (10), the hyd gene encoding a putative cell wall-associated hydrolase is specific for Salmonella spp. (23), the nuc gene encoding a micrococcal nuclease is specific for S. aureus (35), the VP1316 gene encoding a transcriptional regulator of the LysR family is specific for V. parahaemolyticus (44), the hipO gene encoding a hippuricase is specific for C. jejuni (19), the hlyA gene encoding hemolysin is specific for E. coli O157:H7 (38), the sg fragment of the 16S-23S rRNA intergenic spacer is specific for E. sakazakii (22), and the ial fragment encoding a component of the Mxi-Spa secretion machinery is specific for Shigella spp. (15).

Portions (25 g or 25 ml) of food samples were homogenized with 225 ml 2YT liquid medium (16.0 g/liter tryptone, 10.0 g/liter yeast extract, 5.0 g/liter NaCl; pH 7.0), and the mixtures were incubated at 37°C for 8 h. Genomic DNAs of pathogens were extracted by the boiling method (1). The cultured bacteria in 1 ml were put into a 1.5-ml tube and centrifuged for 5 min at 12,000 rpm, and the supernatant was discarded. Then the bacterial precipitate was washed in Tris-HCl-EDTA buffer twice and centrifuged at 12,000 rpm for 3 min, and the supernatant was discarded. Then the bacterial precipitate was added to 1 ml diethyl pyrocarbonate-treated H₂O to redisperse the cells and boiled for 10 min. The tube was centrifuged for 5 min at 12,000 rpm to obtain a supernatant for PCR. Amplification of the target genes was carried out using a 50-μl reaction mixture containing 1× PCR buffer (Mg²⁺ Plus), 0.2 μM of each primer (Table 1), 0.2 mM of each deoxynucleoside triphosphate, 2.5 U of polymerase (rTaq; TaKaRa, Japan), and 50 to ∼100 ng of genomic DNA. PCR amplification was performed with a thermal cycler (PTC-225; MJ Research), using the following program: initial denaturation at 96°C for 5 min, followed by 30 cycles of 96°C for 45 s, 56°C for 45 s, and 72°C for 45 s and then a final extension at 72°C for 10 min. PCR products were labeled by second-round PCR amplification with universal Cy5-modified primers FL1-1 (5′-ATGGTGTAAACTTGTACCAG-3′) and FL1-2 (5′-Cy5-TTGGTAGCAGCGGTAGAGTTG-3′). The PCR was performed as follows. A 5-μl PCR mixture was used, and sequences were amplified using 12 to 15 cycles of 96°C for 30 s, 55°C for 1 min, and 72°C for 30 s. All PCR products were analyzed by 1% agarose gel electrophoresis. The fluorescent PCR products were used directly without further purification. PCR products were purified by using a QIAquick PCR purification kit (Qiagen, Germany) and then quantified with a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, DE) to determine the sensitivity of the BMS.

Beads precoated with different probes were serially pumped into a single microchannel in the following order: NUC, VP, PRS, C20, SG, IAL, HLY, HIP. Each bead set was processed in triplicate. Multiple prepared microchannels were operated in parallel simultaneously. Hybridization solutions (50 μl; 4× SSC-0.1% SDS) containing different pathogen PCR products were injected into prepared microchannels at a flow rate of 20 μl/min. The hybridization reaction mixture was kept at 42°C for 30 min. After hybridization, the beads in microchannels were washed with 1× SSC-0.03% SDS, 0.2× SSC, and then 0.05× SSC (5 min each).

Scanning and data analysis.

The mesofluidic chip was scanned with a Genepix 4000B scanner (Axon Instruments, Foster City, CA) at a resolution of 5 μm for a Cy5 optical filter. The laser power and photomultiplier tube voltage were set to obtain optimum signal intensities. Spot analysis and quantification of the original 16-bit TIF images were performed with the Genepix software (v5.0). In addition, the mesofluidic chip could also be scanned in line by integrating the fluorescence detection device into the system.

RESULTS AND DISCUSSION

To demonstrate the potential and performance of the BMS method, an initial set of experiments was carried out to detect the targets of oligonucleotides and PCR products complementary to different probes on the beads. Synthetic target oligonucleotide T_ial and the PCR product of ial specific for the IAL probe were used to illustrate the specificity and sensitivity of the hybridization on beads.

PCR amplification and design of probes.

Primers with the universal sequences were selected to amplify eight target genes, and the products (264 to 465 bp) were specific for the pathogens tested (23). The results demonstrated that these primers can specifically amplify the target genes (Fig. 2). Amplification of target genes can also be carried out using multiplex PCR (data not shown). The probes were also carefully selected to achieve unambiguous identification of each pathogen. Eight 20-bp oligonucleotide probes, NUC, VP, PRS, C20, SG, IAL, HLY, and HIP, were designed with the Primer3 program to hybridize specifically with the target genes representing eight pathogens (Table 1). The specificity of probes was also investigated. As shown in Fig. 3, when oligonucleotide T_ial specific for the IAL probe was used for hybridization, only the beads carrying the IAL probe produced a strong signal, whereas the other beads showed very low signals close to the background signal. The other probes exhibited similar signal patterns. The specificities of all of the probes were also examined and confirmed with a microarray using PCR products of 27 reference strains (data not shown). The results also strongly indicated that probes NUC, VP, PRS, C20, SG, IAL, HLY, and HIP were specific for S. aureus, V. parahaemolyticus, L. monocytogenes, Salmonella spp., E. sakazakii, Shigella spp., E. coli O157:H7, and C. jejuni, respectively.

FIG. 2.

PCR products of eight pathogens visualized by agarose gel electrophoresis. Lane M, DL2000 marker; lane 1, S. aureus; lane 2, V. parahaemolyticus; lane 3, L. monocytogenes; lane 4, Salmonella; lane 5, E. sakazakii; lane 6, Shigella; lane 7, E. coli O157:H7; lane 8, C. jejuni.

FIG. 3.

Fluorescence values for the eight types of beads hybridized with synthetic target oligonucleotide T_ial.

Surface density of probes on beads.

The effect of the probe concentration on probe immobilization was also studied. If different probe concentrations generate different densities of immobilized probes on the bead surface, the hybridization efficiency is directly affected. Synthetic target oligonucleotide T_ial and the PCR product of ial were hybridized with the beads precoated with five different concentrations of probe IAL. As shown in Fig. 4, the hybridization signal increased with the probe concentration from 0.1 μM to 3.0 μM for targets of both the oligonucleotide and PCR products. However, the hybridization signal decreased when the probe concentration was 5.0 μM. The data clearly demonstrated that the highest value for hybridization signal was obtained with beads prepared with 3.0 μM probe, while the amounts of hybridized targets decreased and the signals were relatively low when beads with higher probe concentrations (5.0 μM) were used. These observations were consistent with the finding of Le Berre et al. (20) and Peterson et al. (27) that a higher probe density could reduce the efficiency of duplex formation and kinetics of the target capture procedure. The reason for the decreased hybridization intensity may be steric and electrostatic interference that prevents target access for hybridization with probes. Therefore, a probe concentration of 3.0 μM was chosen as the working concentration in the following experiments. Based on our previous work (34), the concentration of the DNA probes on the surface was estimated to be 5 × 10¹³ to 7 × 10¹³ probes/cm² when 3.0 μM probe was used for immobilization.

FIG. 4.

Influence of probe concentration. Beads that were precoated with five different concentrations of probe IAL were hybridized with synthetic target oligonucleotide T_ial (▪) at room temperature for 20 min and with the PCR product of ial (•) at 42°C for 30 min.

Fast hybridization of oligonucleotide and PCR products.

The influence of hybridization time was also investigated. The synthetic target oligonucleotide T_ial (50 μl of a 10⁻⁸ M preparation) was hybridized with beads precoated with oligonucleotide probe IAL at room temperature for different times (range, 1 min to 20 min). As shown in Fig. 5, the hybridization was faster in microchannels than in bulk solutions. The signal obtained at 5 min was about 95% of the saturated value. Besides oligonucleotide samples, the PCR product of the ial gene (361 bp), amplified from pathogen cultures, was also tested. Compared with oligonucleotides, PCR products usually require more hybridization time (such as overnight incubation). Accordingly, PCR products (50 μl) were pumped at a flow rate of 20 μl/min and hybridized to beads in microchannels at 42°C for 5 min to 45 min by using stop-flow incubation. It was found that longer hybridization times resulted in stronger signals, and the maximum hybridization signal was observed at 30 min. The results obtained demonstrated that the hybridization efficiency could be enhanced greatly by the mesofluidic flow assay. Remarkably, the hybridization time required in the BMS assay was reduced to 30 min, whereas for a traditional slide microarray the hybridization time is up to 4 h or overnight. The fast hybridization of BMS was attributed mainly to a higher mass transfer velocity in the flow reaction.

FIG. 5.

Effect of the hybridization time for synthetic target oligonucleotide T_ial (▪) at room temperature and for the PCR product of ial (•) at 42°C.

Reproducibility and number of beads.

An attractive feature of the BMS method with a series of beads is the possibility of parallel detection of several microorganisms in a single microchannel. The number of beads used in the BMS method represents the throughput for analysis (like spots of a microarray). The effect of the number of beads was also investigated. PCR products (50 μl) of the ial gene in 4× SSC-0.1% SDS were injected into microchannels with different numbers of beads precoated with the IAL probe (10, 20, 50, 60, and 100 beads). The results are shown in Fig. 6. When the number of beads was in the range from 10 to 100, no significant change of the fluorescence intensity was observed. The method revealed that there was good reproducibility with a variation coefficient of 11.0%.

FIG. 6.

Average fluorescence intensity of all beads in a channel, showing the influence of the number of beads. Each channel contained a different number of beads. PCR products were hybridized with beads at 42°C for 30 min.

Sensitivity of the BMS method for DNA targets.

Analytic sensitivity is one of the most critical factors determining the practicability of a detection method. In our work, the sensitivity of the BMS assay for detection of a synthetic target oligonucleotide and a PCR product was tested. Dilutions of the synthetic target oligonucleotide and PCR product were used to determine the limits of detection. Figure 7 shows the correlation of the signal intensity with the concentration of DNA targets. The fluorescence intensity increased linearly with the concentration from 10⁻¹² to 10⁻⁶ M and then reached equilibrium for synthetic target oligonucleotide. Clearly, the BMS method showed a strong response even for target oligonucleotides at subnanomolar concentrations. The results demonstrated that the target oligonucleotides can be detected at a concentration of 10⁻¹² M. This is a remarkable improvement compared to the biosensor method (28) and the microarray method (37), which detected 200 pM and 1.0 nM. The PCR products were diluted to obtain a series of concentrations from 10⁻¹² to 10⁻⁶ M. The minimum concentration measured in this work was 10⁻⁸ M with a hybridization time of 30 min.

FIG. 7.

Effects of target concentrations of the synthetic target oligonucleotide (▪) from 10⁻¹⁴ to 10⁻⁵ M, of the PCR product (▴) from 10⁻¹² to 6.45 × 10⁻⁶ M, and of control DNA (•).

Detection of pathogenic bacteria.

To evaluate the analytical performance and to assess the applicability of the method, PCR products of pathogenic bacteria were employed with the BMS method. As shown in Fig. 8, only the corresponding beads exhibited high fluorescence intensities. Two or more pathogens can also be detected simultaneously with the BMS method. In most cases, the ratio of the intensities of the positive signals to the intensities of the negative signals (beads with the 20-bp random oligonucleotide) in all experiments exceeded 5:1. Here we used an intensity ratio of 5:1 as the threshold to categorize the positive and negative signals. The results demonstrated that the BMS method has excellent specificity for eight pathogenic bacteria. The detection limits of the BMS were determined with the following eight strains: Salmonella enterica serovar Typhi CMCC50098, S. aureus ATCC 8095, L. monocytogenes ATCC 7644, V. parahaemolyticus ATCC 17802, Shigella sonnei CMCC51334, E. sakazakii ATCC 29544, E. coli O157:H7 strain ATCC 43889, and C. jejuni ATCC 29428. Serially diluted bacterial cultures (1 ml) were used to extract genomic DNAs, which were used as templates to identify the bacteria by the BMS method. Meanwhile, 200-μl serially diluted bacterial cultures were employed for cell counting on plates. It was found that the detection limits of the BMS method were about 500 to 6,000 CFU/ml for these pathogens (Table 2).

FIG. 8.

Detection of PCR products of one, two, three, or four pathogens by the BMS method.

TABLE 2.

Detection results for 184 naturally contaminated food samples

Pathogen	Detection limit of BMS (CFU/ml)a	No. of samples	Results of:
			BMSb	Traditional culture and biochemical identificationc
Salmonella	5 × 102	73	+	+
S. aureus	1 × 103	46	+	+
L. monocytogenes	2 × 103	15	+	+
V. parahaemolyticus	6 × 103	40	+	+
Shigella	1.5 × 103	3	+	+
E. sakazakii	1.3 × 103	5	+	+
Enterohemorrhagic E. coli O157:H7	1 × 103	1	+	+
C. jejuni	1 × 103	1	+	+

^aThe detection limits of the BMS were determined using the following eight strains: S. enterica serovar Typhi CMCC50098, S. aureus ATCC 8095, L. monocytogenes ATCC 7644, V. parahaemolyticus ATCC 17802, S. sonnei CMCC51334, E. sakazakii ATCC 29544, E. coli O157:H7 strain ATCC 43889, and C. jejuni ATCC 29428. After bacterial DNA was extracted from serially diluted bacterial cultures, it was subjected to PCR amplification for BMS detection of pathogens.

^b+, pathogen detected.

^cDetermined using China National Standard GB/T 4789-2003.

As a proof of principle for detection of microorganisms in contaminated food, we examined the detection of pathogenic bacteria in endogenously infected samples (>500 to 6,000 CFU/ml). A total of 184 contaminated food samples from different matrices were tested by the BMS method; these samples included 43 egg samples, 46 pork samples, 42 chicken samples, 20 shellfish samples, 23 fish samples, 5 ice cream samples, and 5 milk power samples. The results obtained with the BMS method were completely consistent with the results obtained with traditional culture and biochemical identification methods (Table 2). The pathogens in all food samples were accurately determined and identified by our method, whereas traditional bacteriology was considered the “gold standard.” This indicated that the BMS method can successfully detect pathogens in contaminated food samples.

In this study, an integrated mesofluidic system comprising a bead-based mesofluidic device, a valve, and a peristaltic pump was developed for pathogen detection. This system provides a platform for injection, transport, and manipulation of probe-modified beads in PDMS channels to facilitate hybridization. All operations, including bead-infusing, hybridization, and washing steps, can be carried out simply by controlling the peristaltic pump. The hybridization reaction could be completed in 30 min. This process, conducted in microchannels, could reduce the sample volume required and protect the liquids from evaporation and cross contamination. A practical application for the mesofluidic system was confirmed based on its ability to determine microorganisms in food samples. This method is fast, has high sensitivity, and can be automated to carry out parallel and high-throughput detection. The microbeads improved the assay sensitivity by increasing the active capture area and also decreased the variability and hybridization times. The flexible nature of the BMS method allows beads precoated with new probes to be added easily. Indeed, the use of different multiplex probe-modified beads in mesofluidic devices should open up new flexible, high-throughput approaches for bioanalysis.