Detection of Vibrio parahaemolyticus in Shellfish by Use of Multiplexed Real-Time PCR with TaqMan Fluorescent Probes

Abstract

We developed a multiplexed real-time PCR assay using four sets of gene-specific oligonucleotide primers and four TaqMan probes labeled with four different fluorophores in a single reaction for detection of total and pathogenic Vibrio parahaemolyticus, including the pandemic O3:K6 serotype in oysters. V. parahaemolyticus has been associated with outbreaks of food-borne gastroenteritis caused by the consumption of raw or undercooked seafood and therefore is a concern to the seafood industry and consumers. We selected specific primers and probes targeting the thermostable direct hemolysin gene (tdh) and tdh-related hemolysin gene (trh) that have been reported to be associated with pathogenesis in this organism. In addition, we targeted open reading frame 8 of phage f237 (ORF8), which is associated with a newly emerged virulent pandemic serotype of V. parahameolyticus O3:K6. Total V. parahaemolyticus was targeted using the thermolabile hemolysin gene (tlh). The sensitivity of the combined four-locus multiplexed TaqMan PCR was found to be 200 pg of purified genomic DNA and 10⁴ CFU per ml for pure cultures. Detection of an initial inoculum of 1 CFU V. parahaemolyticus per g of oyster tissue homogenate was possible after overnight enrichment, which resulted in a concentration of 3.3 × 10⁹ CFU per ml. Use of this method with natural oysters resulted in 17/33 samples that were positive for tlh and 4/33 samples that were positive for tdh. This assay specifically and sensitively detected total and pathogenic V. parahaemolyticus and is expected to provide a rapid and reliable alternative to conventional detection methods by reducing the analysis time and obviating the need for multiple assays.

Vibrio parahaemolyticus is a gram-negative halophilic bacterium that is indigenous to coastal marine waters throughout the world (29, 53). As a common cause of acute gastroenteritis and the source of some cases of septicemia, it is also an organism of public health concern. Typically, human infections with this organism result from the consumption of raw or undercooked seafood (9, 46). This is of particular concern with molluscan shellfish, such as oysters, which, in addition to commonly being consumed raw, are also known to concentrate microorganisms from the environment during the filter-feeding process (45). In recent years, pathogenic strains of V. parahaemolyticus have been associated with gastroenteritis outbreaks around the world at locations that include Spain, Taiwan, Japan, Russia, India, North America, and Southeast Asia (14, 17, 33, 39, 49, 59, 61). In the United States, a 1997 outbreak in the Pacific Northwest resulted in 209 illnesses and one death (11). Outbreaks also occurred in 1998 in New York and Texas. The Galveston Bay, Texas, outbreak involved 416 illnesses and was the largest outbreak that has been reported so far in the United States (12, 15). These facts have prompted interest in the development of fast, sensitive methods for detecting this organism in seafood, particularly oysters. In particular, there is a desire to improve the turnaround time and decrease the amount of labor associated with the more traditional culture-based or biochemical assays. Such assays can take several days to provide positive results, and often the number of samples that can be detected simultaneously is limited (32).

In previous studies, the ability of a strain of V. parahaemolyticus to cause disease was associated with β-type hemolysis on Wagatsuma blood agar, which is known as the Kanagawa phenomenon (36, 56) and is caused by the tdh-encoded thermostable direct hemolysin (23, 36, 48). However, there have been cases of illness that were associated with strains that did not produce thermostable direct hemolysin and were therefore Kanagawa negative. Many of these strains were found to produce a different hemolysin, the trh-encoded thermostable direct-related hemolysin (22, 24, 25, 40). Studies have shown that there is a strong association between pathogenicity and the presence of either the tdh gene or the trh gene or both, indicating that in V. parahaemolyticus both of these genes are virulence factors (36, 48, 50).

Beginning in 1996, an increase in the incidence of V. parahaemolyticus around the world was observed. A large number of the clinical cases seen in the years after 1996 were associated with a unique clone of V. parahaemolyticus serotype O3:K6 (35, 39, 59). It was determined that many of the O3:K6 strains isolated since 1996 contain a filamentous phage, f237, and that this phage contains a unique open reading frame, ORF8. Two other V. parahaemolyticus serotypes that were subsequently isolated, O4:K68 and O1:KUT, have also been shown to contain the f237 phage and therefore the ORF8 gene (26). It has been demonstrated that serovars O3:K6, O4:K68, and O1:KUT are closely related, and molecular biology studies have suggested that O4:K68 and O1:KUT diverged from O3:K6 by genetic alteration of the O and K antigens (7, 35). Together, these serotypes have been designated the pandemic group (41). It has been suggested that ORF8 could play a role in the increased virulence of the recently described pandemic group strains and could be a useful genetic marker for identification of these strains (37, 38).

An additional genetic target, the thermolabile hemolysin gene (tlh), has been proposed and used in some V. parahaemolyticus studies (6). The product of this gene has not been associated with pathogenicity. However, it has been observed in all of the V. parahaemolyticus strains identified so far (6, 51, 52). This gene is therefore a useful target for detection of total V. parahaemolyticus.

In recent years, workers have developed PCR-based assays that target one or more of the genes described above for identification of V. parahaemolyticus in various types of samples. One conventional PCR assay developed in our lab targeted the tlh, tdh, and trh genes simultaneously in a multiplexed format (6, 31), while another assay was designed to detect ORF8 (37). Yet another assay included the tlh target in a multiplex designed to detect multiple microorganisms in shellfish samples (10). These types of tests, while successfully detecting total and/or pathogenic strains, require postamplification analysis in the form of agarose gel electrophoresis. Real-time PCR platforms, in contrast, provide the ability to detect products as they accumulate and obviate the need for further analysis. In addition, real-time PCR assays require less time and labor to accomplish and have been shown to be accurate and reliable (58). Recently, using real-time PCR with SYBR green I dye, tdh and trh nucleotide sequences were included in a group of primers designed to be used in sets of two to detect multiple microorganisms in fecal samples (19). However, real-time PCR assays in which fluorescently labeled TaqMan probes are used provide greater specificity than conventional or SYBR green assays. Successful real-time PCR for V. parahaemolyticus using a TaqMan probe have targeted a single locus (tdh) (8) or dual loci in a multiplexed format (tdh and toxR) (27). For a recent food-borne outbreak that originated from a Chinese buffet in Polk County, FL, V. parahaemolyticus strains containing tlh and tdh were identified in mussels using real-time PCR with TaqMan fluorescent probes (16). However, in this study the workers tested each targeted gene in separate reactions and did not test for the pandemic strain of the V. parahaemolyticus O3:K6 serotype. In order to establish a comprehensive method for detection of total and pathogenic V. parahemolyticus, however, it is necessary to include more than two loci in a single reaction in a multiplexed format. While all pathogenic strains studied so far have at least one of the virulence factors mentioned above, no pathogenic strain with all of the virulence factors has been isolated. A single assay that targets multiple virulence factors to detect any pathogenic strain would therefore be useful, less time-consuming, and economical. In this paper, we describe a real-time PCR assay that uses four sets of oligonucleotide primers and four TaqMan probes labeled with four different fluorophores, which target all four loci (tlh, tdh, trh, and ORF8) in a single multiplexed reaction. An assay like this assay provides the ability to more rapidly detect total and pathogenic V. parahaemolyticus in seafood, particularly oysters, which should benefit not only the seafood industry but also the consumer by ensuring the safety of the food provided for consumption.

MATERIALS AND METHODS

Bacterial strains and growth media.

The following strains of V. parahaemolyticus were used in this study: V. parahaemolyticus F113-A (oyster, Washington), V. parahaemolyticus serotype O3:K6 strain BAC 98-3064 (clinical isolate, FDA), and V. parahaemolyticus serotype O3:K6 strain TX2116 (clinical isolate, Texas). These strains were used for optimization of the real-time multiplexed PCR, for sensitivity studies using purified DNA and pure bacterial cultures, for studies involving detection of V. parahaemolyticus in seeded oysters, and for experiments involving detection of V. parahaemolyticus in cold-stressed conditions. V. parahaemolyticus strains were cultured in T₁N₁ broth (1% [wt/vol] Bacto tryptone [Becton-Dickinson, Franklin Lakes, NJ], 1% [wt/vol] NaCl) at 35°C in a rotary incubator shaker or on T₁N₃ agar plates (1% [wt/vol] Bacto tryptone [Becton-Dickinson], 3% [wt/vol] NaCl) at 35°C overnight (2). The following other strains of bacteria were used in this study: Vibrio furnissii CDC 1958-83, Vibrio fluvialis CDC 1954-82, Vibrio campbellii ATCC 25920, Vibrio alginolyticus ATCC 17749, Vibrio vulnificus MO6-024, and Plesiomonas shigelloides ATCC 14029. These strains were cultured on half-strength marine agar (18.7 g [wt/vol] per liter) (Becton-Dickinson) at 35°C overnight or in full-strength marine broth at 35°C with shaking.

DNA purification.

DNA was extracted from pure cultures of bacterial strains using alkaline lysis and treatment with cetyltrimethylammonium bromide-NaCl and phenol-chloroform, followed by ethanol precipitation (4). Each DNA pellet was washed with cold 70% ethanol, dried, and resuspended in 50 μl Tris-EDTA (pH 8.0) buffer. The DNA concentration and purity were determined using a Lambda II spectrophotometer (Perkin-Elmer, Shelton, CT) at wavelengths of 260 and 280 nm.

Selection of oligonucleotide primers and probes.

The primers and probes used in this study are shown in Table 1. A 450-bp segment of the tlh gene located between nucleotides 781 and 1230 was amplified for detection of total V. parahaemolyticus (6, 10). A 369-bp region of ORF8 located between nucleotides 824 and 1192 was amplified using primers F-03MM823 and R-03MM1192 (37). The primers for tdh (F-tdh170DG and R-tdh403) were designed to amplify the 229-bp region between nucleotides 170 and 438 of this gene, and the primers for the trh region were designed to amplify the 207-bp region between nucleotides 82 and 287. All information for the oligonucleotide probes used for the targeted genes labeled with various fluorescent dyes is shown in Table 1. All primers were analyzed using the IDT Oligoanalyzer software (Integrated DNA Technology, Inc., Coralville, IA) to determine the G+C content and self-dimer and hairpin structures. The oligonucleotide probes used for the TaqMan real-time PCR were analyzed by using the BioTools Primer Quest ScoTools Primer Quest software (IDT), and they had a G+C content of <50%, lacked strong hairpin or self-dimer structures, and did not have 2′-deoxyribosylguanine residues at the 5′ end. The melting temperatures (T_m) for the primers were determined by using the following formula: T_m (˚C) = 2(A+T) + 4(G+C) (47). The specificity of sequences was evaluated using the National Center for Biotechnology Information GenBank database (http://www.ncbi.nlm.nih.gov) and the BLAST search program. All oligonucleotide primers and probes were custom synthesized by Integrated DNA Technology, Inc., Coralville, IA.

TABLE 1.

Target genes and oligonucleotide primers and probes used for multiplexed PCR detection of total and pathogenic V. parahaemolyticus

Target gene	Primer or probe	Sequencea	Length (bp)	G+C content (%)	Melting temp (°C)	Amplicon size (bp)	Reference(s)
tlh	F-tlb	5′-AAA GCG GAT TAT GCA GAA GCA CTG-3′	24	45.8	70	450	6, 10
	R-tlc	5′-GCT ACT TTC TAG CAT TTT CTC TGC-3′	24	41.7	68
	P-tl952d	5′-TexR-AAG AAC TTC ATG TTG ATG ACA CT-BHQ2-3′e	23	34.8	62		This study
ORF8	F-O3MM824b	5′-AGG ACG CAG TTA CGC TTG ATG-3′	21	52.4	64	369	37
	R-O3MM1192c	5′-CTA ACG CAT TGT CCC TTT GTA G-3′	22	45.5	64
	P-ORF8-853d	5′-6-FAM-AAG CCA TTA ACA GTT GAA GGC GTT GAC T-BHQ1-3′f	28	42.9	80		A. Rizvi and A. Bej, unpublished
tdh	F-tdh170DGb	5′-GTA RAG GTC TCT GAC TTT TGG AC-3′	23	43.5	66	229	6; this study
	R-tdh403c	5′-CTA CAG AAT YAT AGG AAT GTT GAA G-3′	25	32.0	66
	P-tdh-341Rd	5′-Cy5-ATT TTA CGA ACA CAG CAG AAT GA-Iowa Black-RQ-3′g	23	34.8	62		This study
trh	F-trh82b	5′-CCA TCM ATA CCT TTT CCT TCT CC-3′	23	43.5	66	207	This study
	R-trh287c	5′-ACY GTC ATA TAG GCG CTT AAC-3′	21	42.9	60
	P-trh275d	5′-TET-TAT TTG TYG TTA GAA ATA CAA CAA T-BHQ1-3′h	25	20.0	60		This study

^aY = C or T; M = A or C; R = G or A.

^bForward primer.

^cReverse primer.

^dOligonucleotide probe.

^eTexR, sulforhodamine 101 (Texas Red) fluorescent dye; BHQ2, Black Hole-2 quencher dye.

^f6-FAM, 6-fluorescein fluorescent dye; BHQ1, Black Hole-1 quencher dye.

^gCy5, carbocyanine fluorescent dye; Iowa Black-RQ, Iowa Black quencher dye.

^hTET, tetrachloro-6-carboxyfluorescein fluorescent dye.

Optimization of PCR.

For real-time PCR experiments, a Cepheid SmartCycler II (Cepheid, Sunnyvale, CA) was employed. For optimization of the real-time PCR system, purified DNA from V. parahaemolyticus strains F113-A and BAC 98-3064 (serotype O3:K6) were used. The optimal concentrations of reagents were determined using the following conditions: 1× PCR buffer (20 mM Tris-Cl [pH 8.4], 50 mM KCl), 3 to 5 mM MgCl₂, each deoxynucleoside triphosphate at concentrations of 200 to 400 μM, each primer at concentrations of 0.4 to 0.8 μM, each fluorophore-labeled probe at concentrations of 0.2 to 0.4 μM, 20 to 400 ng bovine serum albumin, and 1.5 to 2.5 U of Platinum Taq DNA polymerase (Invitrogen, Carlsbad, CA) per reaction mixture. Sterile MilliQ water (Millipore, Bedford, MA) was used to adjust the volume of each reaction mixture to 25 μl. The PCR cycling conditions included an initial denaturation step of 95°C for 120 s, followed by 45 cycles of amplification. Each cycle consisted of template DNA denaturation at 95°C for 20 s, primer annealing at 56°C for 20 s, and primer extension at 72°C for 30 s. The increase in fluorescence for each channel was measured and recorded during the annealing step of each cycle. The cycle threshold (C_t) value used was 15 fluorescence units.

Sensitivity of detection.

The sensitivity of the assay was determined using purified genomic DNA (4) and pure cultures of V. parahaemolyticus. Equal concentrations of purified DNA from V. parahaemolyticus strains F113-A and BAC 98-3064 (serotype O3:K6) were combined and 10-fold serially diluted in Tris-EDTA buffer (pH 8.9) to obtain amounts ranging from 1 μg to 1 pg, and PCR amplification was performed using the optimum reaction conditions. A PCR mixture containing no DNA was used as a negative control.

To determine the sensitivity of detection for pure cultures, V. parahaemolyticus F113-A and TX2116 (serotype O3:K6) were combined and 10-fold serially diluted to obtain concentrations ranging from 10⁷ CFU/ml (optical density at 450 nm [OD₄₅₀], 0.5) to extinction in T₁N₁ broth. The dilutions were then centrifuged at 12,000 × g for 10 min with a microcentrifuge to pellet the bacterial cells. The supernatants were discarded, and the cell pellets were resuspended in 100 μl Instagene matrix (Bio-Rad, Hercules, CA). The resuspended samples were incubated for 20 min at 56°C and boiled at 100°C for 10 min. Samples were then briefly centrifuged to pellet the Instagene matrix and cellular debris. For analysis, 3 μl of each treated sample was included in a real-time PCR amplification mixture.

Detection of V. parahaemolyticus DNA in a background of non-V. parahaemolyticus DNA.

The ability of the system to detect a known detectable amount of V. parahaemolyticus DNA (2 ng) in the presence of external nonspecific DNA was tested by adding different amounts of an equimolar mixture of DNA from five bacterial strains to the system. These strains were V. furnissii, V. fluvialis, V. campbellii, V. alginolyticus, and P. shigelloides strains. The total amounts of external bacterial DNA added to the system ranged from 100 ng to 7 μg. Experiments in which 5 μg to 100 μg herring sperm DNA (Gibco BRL, Carlsbad, CA) was added to the system were also done, and the ability to detect 2 ng V. parahaemolyticus DNA was tested.

Detection of V. parahaemolyticus in a background of nonspecific bacteria.

Pure cultures of V. parahaemolyticus F113-A and TX2116, V. vulnificus, V. campbellii, and V. furnissii were grown until the OD₄₅₀ was 0.2 (10⁶ CFU/ml). For each experiment, a mixture of F113-A and TX2116 was 10-fold serially diluted to extinction as described above. To each dilution, approximately 1 × 10⁶ to 1 × 10⁹ CFU of an equal mixture of V. vulnificus, V. campbellii, and V. furnissii was added as background. The cell mixtures were centrifuged for 10 min at 12,000 × g, and the cell pellets were treated with 100 μl Instagene matrix (Bio-Rad) to release the DNA. PCR amplification was performed with 3 μl of the released DNA using primers and probes specific for V. parahaemolyticus.

Detection of V. parahaemolyticus in seeded oysters.

Oyster samples were purchased from a local seafood store and were processed by using the standard methods described by the American Public Health Association (1). Briefly, oysters were shucked and homogenized using a ConAir Waring stainless steel blender (Fisher Scientific, Pittsburgh, PA). To test for the presence of background Vibrio species, aliquots of oyster homogenate were plated onto CHROMagar Vibrio (DRG International, Mountainside, NJ) plates. The plates were monitored for mauve colonies, which indicated the presence of V. parahaemolyticus.

For seeded oyster experiments, V. parahaemolyticus F113-A and TX2116 (serotype O3:K6) cultures were grown in T₁N₁ broth at 35°C until the OD₄₅₀ was 0.3 (10⁷ CFU/ml). A mixture of equal amounts of the two strains containing 10⁵ CFU/ml was 10-fold serially diluted to extinction and inoculated into 250-ml culture flasks containing 50 ml T₁N₁ and 1 g oyster tissue homogenate. Unseeded oyster homogenate in T₁N₁ broth was used as a negative control. The flasks were enriched overnight (approximately 16 h) at 35°C in a rotary shaker at 170 rpm. Following enrichment, a 100-μl aliquot of each sample was treated with Instagene matrix as described above. For real-time PCR amplification, 3 μl of the sample extract was used in a 25-μl reaction mixture. Also, viable plate counts of V. parahaemolyticus were determined by plating serial dilutions of overnight enriched samples onto CHROMagar Vibrio plates.

Detection of V. parahaemolyticus in natural oysters.

Oysters were collected from the Gulf of Mexico near Bayou LaBatre, Alabama, at 2- to 4-week intervals from November 1997 to January 1999. Collected oysters were immediately chilled, cleaned, and processed by using the standard methods (1). Samples of oyster homogenates were enriched overnight as described above. Enriched samples were aliquoted and stored at −80°C until they were used in experiments. For the experiments described here, a 100-μl aliquot of each of 33 separate samples was extracted using Instagene matrix as described above, and 3 μl of each sample was used in a real-time PCR amplification mixture. In order to determine whether negative TaqMan PCR results for the oyster homogenates were due to reaction inhibition, an aliquot (3 μl) of the samples spiked with purified genomic DNA targeting all four genes was amplified.

Detection of V. parahaemolyticus after extended incubation under low-temperature conditions.

V. parahaemolyticus strains F113-A and TX2116 (serotype O3:K6) were grown in autoclaved (121°C for 20 min at 15-lb/in² pressure) Gulf of Mexico water (salinity, 20 ppt) supplemented with 0.2% (wt/vol) Bacto peptone (Becton Dickinson) at 35°C until the OD₄₅₀ was approximately 0.3, and viable plate counts were determined on T₁N₃ agar. Equal volumes of the cultures were distributed into two sterile 250-ml culture flasks; one of these flasks was stored at 4 ± 1°C, and the other was stored at room temperature (23 ± 1°C). An aliquot (1 ml) was removed from each culture after 0, 1, 7, 14, 21, 28, 35, 42, and 48 days, centrifuged at 12,000 × g for 30 min at 4°C, and treated with 100 μl Instagene matrix to release the DNA. For PCR amplification, 3 μl of an extract was used in a 25-μl reaction mixture. Viable plate counts at the times examined were also determined on T₁N₃ agar plates. For cultures whose initial plate counts were below detectable levels (approximately 100 cells/ml), a 10-ml aliquot was filtered through a sterile Millipore type HA filter (Millipore, Bedford, MA), which was then incubated overnight at 30°C on a T₁N₃ agar plate with the cell side up. Samples from cold-stressed and newly grown cultures were subjected to staining with a LIVE/DEAD BacLight bacterial viability kit and observed using a Leitz Diaplan epifluorescence microscope. When this system was used, cells that stained green were considered to be viable, while cells that stained red were considered to be nonviable. Direct counting of bacterial cells in samples exposed to low temperatures was performed as described by Hobbie et al. (21).

RESULTS

Optimization and specificity of PCR.

After optimization of the reactions with various concentrations of the reagents described in Materials and Methods, the following conditions were found to be the best conditions for consistent amplification of all four targeted genes in a multiplexed format: 1× PCR buffer (20 mM Tris-Cl [pH 8.4], 50 mM KCl), 5 mM MgCl₂, each deoxynucleoside triphosphate at a concentration of 400 μM, each of the eight gene-specific oligonucleotide primers at a concentration of 0.4 μM, each of the four fluorescence-labeled probes at a concentration of 0.2 μM, 100 ng bovine serum albumin, and 2 U of Platinum Taq DNA polymerase (Invitrogen, Carlsbad, CA) per 25-μl reaction mixture. Also, the optimal primer annealing temperature was found to be 56°C, and the best results were obtained with the following thermal cycling times: 20 s for denaturation, 20 s for primer annealing, and 30 s for primer extension. Agarose gel electrophoresis results confirmed the presence of amplicons at the expected molecular weights (450 bp for tlh, 369 bp for ORF8, 229 bp for tdh, and 207 bp for trh) (Table 1). Primers and probes were shown to be specific using the BLAST search program, the NCBI GenBank genome database, and the results of previously reported conventional PCR tests conducted in our laboratory with numerous V. parahaemolyticus strains, other Vibrio spp., and non-Vibrio spp. (6, 37).

Sensitivity of detection.

The minimum level of detection of purified V. parahaemolyticus genomic DNA at all targeted gene loci was 200 pg. This level of detection of genomic DNA is equivalent to approximately 10⁴ CFU V. parahaemolyticus, assuming that single copies of target loci are present in the genome (3). The average C_t values for each locus at this level, as well as at other concentrations of DNA, are shown in Table 2. As the concentration of DNA in a reaction mixture decreased, the C_t value increased (Table 2). The levels of amplification product determined by agarose gel electrophoresis also decreased with decreasing amounts of initially added template DNA (data not shown). The sensitivity of detection in pure cultures of V. parahaemolyticus was 10⁴ CFU/ml, which is comparable to the level at which purified DNA is detected. The C_t value for each of the targeted genes increased as the concentration of viable bacteria increased (Table 3), which was shown by the optics graphs and was confirmed by agarose gel electrophoresis (Fig. 1). Standard curves showed that there was a good linear correlation between the C_t values and the concentrations of input DNA or bacteria; the r² values for the different loci ranged from 0.988 to 0.993 for purified DNA and from 0.984 to 0.999 for bacterial cultures (Fig. 2).

TABLE 2.

Sensitivity of detection of V. parahamolyticus purified DNA using multiplexed real-time PCR with TaqMan probes^a

DNA concn	Ct value
	tlh	ORF8	tdh	trh
2 μg	15.30 ± 0.18	15.10 ± 0.40	14.05 ± 0.34	15.54 ± 0.55
200 ng	17.46 ± 0.30	17.01 ± 0.34	15.96 ± 0.32	17.69 ± 0.38
20 ng	21.21 ± 0.38	21.13 ± 1.14	19.93 ± 0.87	21.45 ± 0.69
2 ng	24.70 ± 0.12	24.30 ± 0.23	23.29 ± 0.31	24.93 ± 0.57
200 pg	28.15 ± 0.65	28.41 ± 1.15	26.98 ± 0.99	28.53 ± 1.10
20 pg	NDb	ND	ND	31.74c
2 pg	ND	ND	ND	ND
Negative control	ND	ND	ND	ND

^aThe data are means ± standard deviations for four independent experiments.

^bND, not detected.

^cDetected in one reaction.

TABLE 3.

Sensitivity of detection for V. parahamolyticus pure cultures using multiplexed real-time PCR with TaqMan probes^a

Sampleb	V. parahaemolyticus concn (CFU/ml)	Ct value
		tlh	ORF8	tdh	trh
1	107	17.25 ± 1.45	16.24 ± 3.70	14.73 ± 0.44	16.12 ± 0.24
2	106	19.96 ± 0.40	20.14 ± 0.36	18.61 ± 0.10	20.14 ± 0.08
3	105	23.53 ± 0.26	24.14 ± 0.85	22.21 ± 0.20	23.30 ± 0.06
4	104	28.29 ± 0.29	28.92 ± 1.09	25.95 ± 0.48	27.18 ± 0.26
5	103	NDc	ND	30.16 ± 0.41	30.47 ± 0.13
6	102	ND	ND	ND	ND
7	101	ND	ND	ND	ND
8	100	ND	ND	ND	ND
9	Positive controld	18.44 ± 0.58	18.60 ± 0.49	17.48 ± 0.25	19.11 ± 0.41
10	Negative controle	ND	ND	ND	ND

^aThe data are means±standard deviations for three independent experiments.

^bThe sample numbers correspond to the lane numbers in Fig. 1E.

^cND, not detected.

^dThe positive control included 100 ng purified V. parahaemolyticus DNA.

^eThe negative control included 3 μl extract of T₁N₁ growth medium with no added cells.

FIG. 1.

Sensitivity of detection of V. parahaemolyticus in pure culture using real-time multiplexed PCR with TaqMan probes. The numbers in the graphs correspond to the sample numbers in Table 3, as follows: 1, 10⁷ CFU/ml; 2, 10⁶ CFU/ml; 3, 10⁵ CFU/ml; 4, 10⁴ CFU/ml; 5, 10³ CFU/ml; 6, 10² CFU/ml; 7, 10¹ CFU/ml; 8, 10⁰ CFU/ml; 9, positive control (100 ng V. parahaemolyticus DNA); 10, negative control. (A) Optics graph for the tlh locus with the Texas Red (Tx Red) fluorophore-labeled probe. The C_t values are 16.62 (sample 1), 18.33 (sample 9), 20.17 (sample 2), 23.75 (sample 3), and 28.40 (sample 4). (B) Optics graph for the ORF8 locus with the 6-fluoresein (FAM) fluorophore-labeled probe. The C_t values are 16.82 (sample 1), 18.12 (sample 9), 20.24 (sample 2), 24.87 (sample 3), and 29.27 (sample 4). (C) Optics graph for the tdh locus with the Cy5 fluorophore-labeled probe. The C_t values are 15.08 (sample 1), 17.19 (sample 9), 18.72 (sample 2), 22.39 (sample 3), 26.32 (sample 4), and 30.45 (sample 5). (D) Optics graph for the trh locus with thetetrachloro-6-carboxyfluorescein (TET) fluorophore-labeled probe. The C_t values are 15.87 (sample 1), 18.65 (sample 9), 20.05 (sample 2), 23.27 (sample 3), 26.90 (sample 4), and 30.37 (sample 5). (E) Agarose gel (1.8%, wt/vol) showing the electrophoresis results for real-time PCRs. Lane 1, 10⁷ CFU/ml; lane 2, 10⁶ CFU/ml; lane 3, 10⁵ CFU/ml; lane 4, 10⁴ CFU/ml; lane 5, 10³ CFU/ml; lane 6, 10² CFU/ml; lane 7, 10¹ CFU/ml; lane 8, 10⁰ CFU/ml; lane 9, positive control (100 ng V. parahaemolyticus DNA); lane 10, negative control; lane SS, “Clone-Sizer” 100-bp DNA ladder (Norgen Biotek Corporation, Ontario, Canada). The sizes of amplified DNA bands are indicated on the right.

FIG. 2.

Standard curves for the numbers of V. parahaemolyticus cells or amounts of purified DNA versus C_t values of the fluorescent signals for the tlh locus (A), the ORF8 locus (B), the tdh locus (C), and the trh locus (D). The error bars indicate standard deviations for four independent experiments.

Effect of background nonspecific DNA or bacteria.

When 2 ng purified V. parahaemolyticus was tested with a background of 5 μg nonspecific bacterial DNA, the C_t values for each locus increased by approximately one cycle or less (Table 4). The increases in the values in the presence of 5 μg were within 1 standard deviation of the values obtained when no background DNA was present, and therefore no significant inhibition occurred in the presence of 5 μg background DNA. When background DNA was included at levels below this level (5 μg), no significant change in the C_t value was observed (data not shown). For the tlh and ORF8 loci, the C_t values again increased by approximately one cycle when the amount of background DNA was increased from 5 μg to 6 μg. The difference in C_t values between samples with 6 μg background DNA and samples with no background DNA was significant for these loci, and this finding suggests that there was some inhibition of the reaction; however, at this level, all loci were still consistently detected in 100% of the samples tested. When the level of background DNA was increased to 7 μg, inhibition, as shown by increased reaction inconsistency, began to occur. This effect was observed to occur first and most prominently for tlh (450 bp), followed by ORF8 (369 bp). Similar inhibition began to be seen for tdh and trh with 8 μg background DNA. As the level of background DNA increased above this level, the inhibitory effect also increased. With more than 10 μg background DNA, inhibition was complete for all loci (Table 4). The observed effects were the same whether the background DNA was bacterial or nonbacterial (herring sperm DNA).

TABLE 4.

Effect of background nonspecific DNA on detection of V. parahaemolyticus DNA^a

Amt of nonspecific DNA (μg)b	tlh		ORF8		tdh		trh
	Ct value	% Positivec	Ct value	% Positive	Ct value	% Positive	Ct value	% Positive
0	20.51 ± 0.74	100	20.31 ± 0.27	100	19.97 ± 1.39	100	21.33 ± 0.60	100
5	21.06 ± 0.89	100	20.74 ± 1.96	100	21.12 ± 1.00	100	21.57 ± 0.67	100
6	22.08 ± 1.02	100	21.71 ± 1.12	100	20.59 ± 0.92	100	21.47 ± 0.56	100
7	21.16	33	24.30 ± 5.69	100	21.21 ± 0.58	100	21.52 ± 0.26	100
8	21.08	33	20.26 ± 0.57	67	23.79 ± 3.71	100	22.84 ± 1.28	100
9	21.76	33	22.29	33	23.17 ± 0.18	67	23.73 ± 0.96	100
10	NDd	0	22.99	33	25.03 ± 2.42	67	24.63 ± 0.74	100
25	ND	0	ND	0	ND	0	ND	0
50	ND	0	ND	0	ND	0	ND	0
100	ND	0	ND	0	ND	0	ND	0

^aThe data are means ± standard deviations for three independent experiments. The amount of V. parahaemolyticus DNA included in each reaction mixture was 2 ng.

^bThe nonspecific DNA included herring sperm DNA and DNA purified from V. furnissii, V. fluvialis, V. campbellii, V. alginolyticus, and P. shigelloides.

^cPercentage of reactions that yielded positive results with the TaqMan real-time PCR system.

^dND, not detected.

When the ability to detect V. parahaemolyticus bacterial cells in the presence of nonspecific bacteria (V. vulnificus, V. campbellii, and V. furnissii) was tested, the presence of background bacteria at a level of 10⁶ CFU did not decrease the sensitivity of detection (Table 5). The C_t values in the presence of this level of background bacteria increased slightly compared to the values when no external bacteria were present. This effect was apparently not due to cross-reactivity, since no positive results were obtained for negative controls or for samples containing less than 10⁴ V. parahaemolyticus CFU/ml. The higher C_t values could have been due to a coprecipitation effect, in which the presence of other cells improved retention and collection of relatively small amounts of the targeted cells (37, 49). When the level of nonspecific background bacteria was increased to 10⁹ CFU/ml, inhibition similar to that seen with increasing amounts of background DNA was observed. Increased incidence of reaction inconsistency was seen under these conditions, and this was again most prominent for the loci with higher molecular weights, tlh (450 bp) and ORF8 (369 bp). For these two loci, detection of any amount of V. parahaemolyticus DNA was affected, and the C_t values increased by approximately 3 cycles compared to samples with no background bacteria. Inhibition was not observed at the tdh and trh loci in the presence of this amount of background bacteria unless the levels of V. parahaemolyticus were below the sensitivity level (10⁴ CFU/ml) (Table 5).

TABLE 5.

Effect of nonspecific bacteria at different concentrations on the sensitivity of detection of V. parahaemolyticus in pure culture^a

Amt of V. parahaemolyticus (CFU/ml)	Ct values with different amt of nonspecific background bacteriab
	tlh			ORF8			tdh			trh
	No bacteria	106 CFU/ml	109 CFU/ml	No bacteria	106 CFU/ml	109 CFU/ml	No bacteria	106 CFU/ml	109 CFU/ml	No bacteria	106 CFU/ml	109 CFU/ml
107	17.25 ± 1.45	15.72 ± 0.35	18.81d	16.24 ± 3.70	15.38 ± 0.44	17.52d	14.73 ± 0.44	14.22 ± 0.68	15.22 ± 0.43	16.12 ± 0.24	16.55 ± 0.13	16.56 ± 0.15
106	19.96 ± 0.40	19.40 ± 0.26	22.22d	20.14 ± 0.36	19.91 ± 0.68	23.10 ± 2.10	18.61 ± 0.10	18.00 ± 0.23	18.88 ± 0.47	20.14 ± 0.08	20.20 ± 0.49	20.72 ± 0.61
105	23.53 ± 0.26	22.75 ± 0.16	26.49d	24.14 ± 0.85	22.92 ± 0.38	28.55d	22.21 ± 0.20	21.67 ± 0.40	22.54 ± 0.04	23.30 ± 0.06	23.62 ± 0.25	23.88 ± 0.14
104	28.29 ± 0.26	25.56 ± 0.19	NDc	28.92 ± 1.09	25.97 ± 0.39	ND	25.95 ± 0.48	24.70 ± 0.20	25.91 ± 0.35	27.18 ± 0.26	26.66 ± 0.16	27.59 ± 0.11
103	ND	ND	ND	ND	ND	ND	30.16 ± 0.41	28.43d	ND	30.47 ± 0.13	ND	ND
Positive controle	18.44 ± 0.58	18.42 ± 0.59	19.06 ± 0.27	18.60 ± 0.49	18.53 ± 0.52	19.01 ± 0.76	17.48 ± 0.25	17.37 ± 0.14	17.12 ± 0.31	19.18 ± 0.64	19.76 ± 0.25	19.16 ± 0.37
Negative control	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND

^aThe data are means ± standard deviations for two independent experiments.

^bThe background was an equal mixture of three organisms, V. vulnificus, V. campbellii, and V. furnissii.

^cND, not detected.

^dOnly one positive result was obtained with two samples.

^eThe positive control was 100 ng purified V. parahaemolyticus genomic DNA.

Detection of total and pathogenic V. parahaemolyticus in seeded oyster tissue homogenate following enrichment.

After overnight enrichment, V. parahaemolyticus in samples in which the initial inoculum was 1 CFU V. parahaemolyticus and the final amount was approximately 3.3 × 10⁹ CFU/ml could be detected in 1 g seeded oyster tissue homogenate (Table 6). This level of initial cell detection meets the detection limit requirements outlined by the Interstate Shellfish Sanitation Conference (28), and this was achieved only after enrichment. There was not a linear correlation for the C_t values observed with different concentrations of seeded bacteria, and the C_t values were not significantly different from each other. This was likely due to the fact that enrichment occurred overnight and the seeded cultures were allowed to grow to saturation. No V. parahaemolyticus was detected in negative controls or in oyster samples seeded with T₁N₁ broth containing no V. parahaemolyticus. No V. parahaemolyticus colonies were detected when unseeded oyster homogenates were cultured on CHROMagar plates. The viable plate counts for V. parahaemolyticus on CHROMagar following overnight enrichment were approximately 10⁹ CFU/ml (Table 6).

TABLE 6.

Detection of V. parahaemolyticus in seeded oysters with overnight enrichment^a

Amt of V. parahaemolyticus seeded (CFU)	Final V. parahamolyticus concn (109 CFU/ml)b	Ct value
		tlh	ORF8	tdh	trh
105	3.88	18.94 ± 1.38	17.87 ± 2.03	16.75 ± 1.41	18.18 ± 1.57
104	4.20	19.33 ± 0.85	19.55 ± 0.59	17.24 ± 0.84	18.57 ± 1.22
103	2.80	19.36 ± 0.75	24.37 ± 2.26	17.55 ± 0.92	18.16 ± 1.03
102	4.65	20.08 ± 0.49	19.02 ± 0.97	17.40 ± 0.70	19.54 ± 0.93
101	4.00	19.55 ± 0.67	18.38 ± 0.86	17.18 ± 0.88	19.56 ± 1.31
100	3.30	20.84 ± 0.93	19.71 ± 0.50	18.61 ± 0.95	19.21 ± 1.44
0	NDc	NDc	ND	ND	ND
Positive controld		18.98 ± 0.42	18.16 ± 0.40	17.23 ± 0.40	19.29 ± 0.41
Negative controle		ND	ND	ND	ND

^aThe data are means ± standard deviations for four independent experiments.

^bThe viable plate count is an average for three dilutions.

^cND, not detected.

^dThe positive amplification control included 100 ng purified V. parahaemolyticus genomic DNA.

^eThe negative amplification control included 3 μl sterile water.

Detection of V. parahaemolyticus in natural oysters.

Of 33 collected oysters tested in this study, 17 were positive for the tlh gene, indicating that V. parahaemolyticus was present in the samples (Table 7). Of the 17 tlh-positive samples, 4 were also positive for the tdh gene, indicating that these samples contained a tdh-bearing pathogenic strain of V. parahaemolyticus. None of the samples tested were positive for trh or ORF8. Agarose gel electrophoresis of the real-time PCR-amplified products revealed that 17 samples were positive for tlh (450 bp), and 4 of these 17 samples were positive for tdh (235 bp). The gel electrophoresis results for the 16 TaqMan PCR-negative oyster samples spiked with purified genomic DNA showed that there was PCR amplification of all four targeted genes, suggesting that the lack of detection was not due to inhibition of the reaction (data not shown).

TABLE 7.

Summary of real-time PCR detection of V. parahaemolyticus in natural oyster samples

Locus	No. positive/ no. tested (%)
tlh	17/33 (51.5)
ORF8	0/33 (0)
tdh	4/33 (12.1)a
trh	0/33 (0)

^aThe samples positive for tdh were also positive for the tlh locus.

Detection of cold-stressed V. parahaemolyticus.

The initial viable plate count for the mixed-strain culture was 1.3 × 10⁸ CFU/ml. The viable plate count for the cultures maintained at 4°C was reduced to approximately 4 CFU/ml by day 48, while the viable plate count for the sample maintained at room temperature remained high and was only reduced to 1.7 × 10⁷ CFU/ml by day 48. The C_t values for samples from the mixed culture maintained at room temperature (not shown) did not vary significantly over time. The initial C_t values at time zero were 21.82 (tlh), 21.76 (ORF8), 20.38 (tdh), and 22.37 (trh), while the values on day 48 were 22.67 (tlh), 22.81 (ORF8), 20.69 (tdh), and 21.51 (trh). The C_t values obtained from real-time PCR experiments performed with cold-stressed cultures maintained at 4°C at various times are summarized in Table 8. The C_t values for the cold-stressed samples tended to increase over time, indicating that the concentrations of viable cells were decreasing or that cellular debris from dead cells had inhibitory effects on reactions; however, all loci could still be detected in this culture, even after the viable plate counts of the cultures dropped below the determined sensitivity level (10⁴ CFU/ml). Positive real-time PCR results were obtained up to day 48, when the plate counts were <5 CFU/ml. When observed using BacLight staining, ∼97% of the cells in the 48-day, cold-stressed cultures appeared to be Syto9 Green-stained viable cells. These viable cells had the typical coccoid morphology seen in cultures maintained at low temperatures in growth medium with minimal nutrients, which suggested that the culture was in a viable but physiologically compromised state of growth (30). Enumeration of the viable coccoid cells in these cold-stressed cultures yielded a value of approximately 1.2 × 10⁸ CFU/ml.

TABLE 8.

Detection of V. parahaemolyticus at various times after exposure to cold temperatures^a

Time (days)	Ct value				Concn of culturable cells (CFU/ml)b
	tlh	ORF8	tdh	trh
0	20.92 ± 0.39	19.95 ± 0.32	19.13 ± 0.56	21.79 ± 1.43	1.30 × 108
1	20.01 ± 0.43	19.62 ± 0.90	18.03 ± 0.53	20.40 ± 1.02	3.13 × 107
14	21.70 ± 0.12	25.82 ± 0.56	20.52 ± 0.25	21.26 ± 0.81	3.94 × 106
21	23.02 ± 0.63	28.30 ± 0.00e	21.71 ± 0.53	22.87 ± 0.64	5.06 × 104
28	23.77 ± 0.19	28.91 ± 0.94f	22.77 ± 0.54	24.04 ± 0.61	4.74 × 103
35	25.40 ± 0.22	30.49 ± 3.13f	24.25 ± 0.67	25.40 ± 0.56	<10
42	25.42 ± 3.48	20.75f	23.19 ± 4.29	24.30 ± 3.84	<10
48	27.67 ± 0.43	32.08 ± 2.73	26.35 ± 0.59	27.32 ± 1.11	<5
Positivec	23.95 ± 0.31	22.95 ± 0.46	22.35 ± 0.34	25.52 ± 0.03
Negatived	NDg	ND	ND	ND

^aThe data are means ± standard deviations for three experiments in which samples were incubated at 4°C.

^bThe viable plate count is an average for three dilutions.

^cThe positive amplification control included 10 ng purified V. parahaemolyticus genomic DNA.

^dThe negative amplification control included 3 μl sterile water.

^eOne of three samples yielded positive results.

^fTwo of three samples yielded positive results.

^gND, not detected.

After day 21, detection of ORF8 in the cold-stressed mixture was inconsistent, and some experiments yielded negative results; however, this was consistent with the behavior of the ORF8-positive TX2116 strain, which was found to be more susceptible to the cold-stressed conditions than the F113-A strain when the strains were tested individually. The viable plate count value for the TX2116 culture maintained at 4°C was <10 CFU/ml by day 24, while that for the F113-A culture did not reach this level until after day 35. This indicates that the viable plate counts for cultures of the TX2116 strain reached threshold detection levels or below more rapidly than the viable plate counts for cultures of the F113-A strain, which was still detected consistently with real-time PCR at day 48.

DISCUSSION

In this study we developed and optimized a “four-color” multiplexed TaqMan probe-based real-time PCR assay for detection of total and pathogenic V. parahaemolyticus in oysters. A species-specific gene locus (tlh) (6, 51, 52) and two established virulence gene loci (tdh and trh) (36, 48, 50), as well as a DNA sequence (ORF8), to detect strains of the recently emerged, increasingly virulent pandemic strain of the V. parahaemolyticus O3:K6 serotype were selected for this study (37, 38). We targeted these genes in order to obtain comprehensive detection of total and pathogenic V. parahaemolyticus in oysters, in which not all virulent strains of this microorganism contain all three gene segments (tdh, trh, and ORF8) (22, 24, 25). Also, while the ORF8 gene has been used as a marker to detect recent serotype O3:K6 and other V. parahaemolyticus pandemic group strains (37, 38), one study identified eight O3:K6 isolates obtained between 1998 and 2000 that were not positive for the ORF8 sequence (7). The primers used for the ORF8 DNA were different from primers previously reported by our laboratory (37), and none of the strains were available to us. The previously reported study in which the toxRS/new or toxRS/old primers were used did not distinguish the newly emerged pathogenic V. parahaemolyticus O3:K6 serotype from the pre-1996 nonpathogenic O3:K6 serotype strains (7, 41). Although the pandemic strain has not been isolated since the outbreak in the United States (11, 12, 17), this pathogen remains prevalent in Asian countries (13, 35, 39-41, 61). Therefore, for detection of pathogenic pandemic V. parahaemolyticus O3:K6 serotype strains, it seems that ORF8 is the best target. However, these isolates were positive for tdh. So far, the evidence suggests that any pathogenic strain, including an O3:K6 serotype strain, contains one of the gene targets (tdh, trh, or ORF8). Therefore, identification of any of the three loci individually or in combination would be necessary to establish the presence of a pathogenic strain of V. parahaemolyticus in oysters. Also, our assay allows results to be obtained within approximately 2.5 to 3 h after enrichment, since the cycle times are short and there is no need for postamplification analysis, such as agarose gel electrophoresis or DNA-DNA hybridization, for confirmation of real-time detection.

While a four-color assay designed to detect the factor V Leiden and prothrombin G20210A mutations using two sets of primers has been described (54), to our knowledge only one other assay that uses four separate primer-probe sets in a real-time PCR TaqMan multiplex for bacterial detection has been reported. Vickery et al. (55) described a four-color multiplexed real-time PCR assay that targeted the tlh, tdh, and trh genes of V. parahaemolyticus and included a PCR internal control. The internal control was reported to prevent false-negative detection, to allow determination of the level of PCR inhibition, and to potentially allow quantitation (55). However, in our study our intent was to use the four-channel capability of the Cepehid Smart Cycler II instrument to target all four known genetic loci that are important for the detection of V. parahaemolyticus in oysters. Since it is expected that in most cases enrichment is a necessary step for detection of microbial pathogens in seafood and since enrichment potentially leads to growth levels that could confound the attempt to accurately determine the initial amounts of bacteria, it was not our goal to provide an internal control for quantitative purposes. Also, we found that postenriched cultures do not significantly inhibit the PCR for nonquantitative detection of this pathogen. Therefore, an internal PCR control does not seem to make a significant contribution to our assay. Quantitative enumeration of total and pathogenic V. parahaemolyticus in unenriched postharvest-treated or natural oysters using an internal control described by Vickery et al. (55) with a minimum level of detection of <30 CFU/g is yet to be achieved. Moreover, descriptions of the internal standard and the primers for the other targeted genes currently are not available from these authors for further evaluation and comparison.

Our real-time PCR assay was capable of detecting a minimum of 200 pg of purified genomic DNA or an equivalent level (10⁴ CFU) of V. parahaemolyticus in pure cultures. However, the agarose gel electrophoresis analysis resulted in faint DNA bands at expected molecular weights, indicating that the presence of amplification products at a level of 20 pg or 10³ CFU could be detected in the gel (Fig. 1E) when negative C_t values were obtained with the real-time PCR with TaqMan probes. Although occasionally positive results were obtained at this level with real-time PCR for tdh and/or trh (data not shown), not all of the products were detected by agarose gel electrophoresis. A consistent level of detection for all four loci with multiplexed real-time TaqMan PCR and corresponding gel electrophoresis methods was achieved for 200 pg of purified DNA or 10⁴ CFU V. parahaemolyticus in pure cultures. Therefore, these levels of detection were established as the sensitivity of detection in this study. However, individual amplification of each of the loci resulted in 10-fold-higher sensitivity (20 pg of purified DNA or 10³ CFU) for detection, and consistency was noticed for both the TaqMan PCR and gel electrophoresis methods (data not shown). This level of sensitivity is approximately 10-fold lower than the level of sensitivity for a TaqMan probe-based real-time PCR duplex assay targeting tlh and ORF8 developed in our lab (unpublished data). The reduced sensitivity of the TaqMan PCR with addition of more targets may be related to the increased background fluorescence that is an inevitable by-product of the presence of multiple TaqMan fluorescent probes. In each TaqMan PCR experiment, the C_t values remained within the range described in this paper. Studies addressing the factors that affect the performance of 5′ nuclease assays using TaqMan fluorescent probes have shown that increasing the fluorescent probe concentration above a certain critical value leads to a decrease in the ratio of reporter fluorescence to quencher fluorescence (34). In a multiplexed assay, it is necessary to include a sufficient concentration of each probe so that the concentration does not become a limiting factor for the reaction; however, with an increased number of heterologous probes, the background fluorescence level with four-color multiplexed assay remains a challenge. Background fluorescence that is too great can lead to difficulty in detecting the smaller fluorescence increases observed with low concentrations of the template DNA in the sample. This phenomenon may also explain the occasional presence of faint bands in agarose gels at a level of sensitivity higher than the levels obtained from real-time PCR. We also found during our experiments that the position of the probe relative to the forward or reverse primer could increase the efficiency of detection. This effect has also been observed by other investigators (34). An additional consideration that was important in optimizing this real-time PCR assay was the use of the hot-start Platinum Taq DNA polymerase. Thermostable DNA polymerases designed to become active only after incubation at a high temperature (95°C) have been recommended in order to improve specificity and sensitivity (20). In some of our previous studies we found that use of a regular thermostable DNA polymerase is sufficient to achieve the required specificity and sensitivity of detection of a targeted pathogen (43, 44). When a larger number of targets and multiple fluorophore-labeled probes were included, however, as in this assay, the presence of a hot-start enzyme was found to significantly improve the sensitivity and consistency of the results.

In order to obtain the minimum level of detection recommended by the Interstate Shellfish Sanitation Conference (<30 CFU per g of oyster tissue) (28), we used the recommended standard overnight enrichment method (32) for seeded and natural oyster homogenates. In some of our previous studies, we reported that a 5- to 8-h enrichment period is sufficient for conventional and real-time PCR assays to detect V. parahaemolyticus and V. vulnificus in oyster tissue at the recommended level of sensitivity (6, 43, 44); however, V. parahaemolyticus has been shown to be able to exist in a physiologically compromised state in the environment, particularly when it is subjected to stress conditions, such as low temperatures (17, 18). In these conditions, the compromised bacteria exhibit low growth rates. Studies have shown that slowly growing or compromised V. parahaemolyticus cells can be restored by overnight incubation at temperatures and with nutrient conditions that promote growth (5). Overnight enrichment, therefore, ensures that V. parahaemolyticus can be detected with the proper sensitivity and reliability.

Our experiments also showed that a multiplexed real-time PCR assay could be used to detect V. parahaemolyticus cultures following exposure to a low temperature (4°C) and minimal nutrient conditions for 48 days at a level of <5 CFU/ml, as determined by the viable plate count method. However, total microscopic counts for these cultures determined using BacLight LIVE/DEAD fluorescence staining showed that they contained 1.2 × 10⁸ CFU/ml, a value which is much higher than the detection limit for a pure culture (10⁴ CFU/ml). This can be explained by the fact that a number of viable but physiologically compromised cells in cold-stressed cultures fail to form active colonies on agar plates. Detection of cold-stressed, physiologically compromised cultures is apparently not as robust as detection of exponentially grown cultures. While the number of cells in cold-stressed cultures at day 48 was determined to be on the order of 10⁸ CFU/ml, the C_t values of samples, which were between 26.35 and 32.08, more closely resembled the C_t values of fresh cultures containing 10⁴ CFU/ml than the C_t values of cultures containing 10⁸ CFU/ml (Table 3 and Table 8). Although the reasons for this are unclear, it can be postulated that most of the coccoid nonculturable cells in these cold-stressed cultures were not affected by the Instagene-boil-lysis method and thus insufficient template DNA was available for the real-time PCR. This hypothesis is supported by previous reports (13, 42, 57, 60) in which the rigidity of the membrane was described. The difference in the responses to cold stress conditions of different strains of V. parahaemolyticus is not understood and requires further investigation.

The inhibitory effect of oyster tissue homogenate has always been an important issue for TaqMan-based real-time PCR-based detection of microbial pathogens. However, the inhibition control experiments in a previous study in our laboratory performed with oyster tissue homogenate spiked with V. vulnificus along with treatment with Instagene matrix showed that the sensitivity of detection by the TaqMan PCR was not significantly affected (43, 44). Moreover, even if there is some inhibition of the reaction, enrichment provides enough templates to overcome this effect, which leads to nonquantitative detection of this pathogen. We also explored the potential inhibitory effect of background nonspecific DNA or bacteria on the TaqMan PCR assay. The results indicated that the presence of background DNA or bacteria did not prevent detection of V. parahaemolyticus as long as the background DNA or bacteria were present at concentrations above the established sensitivity thresholds, 200 pg DNA and 10⁴ CFU. Relatively large amounts of nonspecific bacteria did not change the sensitivity of the reaction. However, once the concentration of background DNA or bacteria reached a certain level (approximately 7 to 8 μg DNA or 10⁹ bacteria), inhibition of the reaction was observed. Results from this study indicate that this effect is caused by overloading of the system, since similar inhibition was observed when only V. parahaemolyticus was present in bacterial samples at saturation levels (10⁹ CFU/ml) (data not shown). It has been established previously that adding too much template can lead to mispriming or loss of amplification in a PCR (62). In our experiments, dilution (1:1, vol/vol) or the use of a smaller sample (3 μl from a 100-μl sample) eliminated the inhibition seen with overconcentrated samples. Our results indicate that V. parahaemolyticus can be detected in a background of other bacteria as long as it is present at levels above the sensitivity limit, as determined in this study, and as long as the total amount of DNA or bacteria does not exceed the inhibitory amounts determined in this study.

Results of our experiments indicate that our real-time TaqMan probe-based PCR system can successfully detect four targets simultaneously in natural oyster samples. This assay can provide information about the presence of V. parahaemolyticus, whether the strain detected is likely to be pathogenic, and whether the strain belongs to the recently emerged pandemic group without the necessity of performing multiple assays or using postamplification analysis methods. Rapid detection like this should provide the seafood industry with early warning of potential health risks associated with potentially contaminated seafood and allow appropriate measures to prevent disease outbreaks to be swiftly undertaken. The use of an assay such as this assay should be beneficial to both industry and consumer health.