at SciVerse ScienceDirect

Food Control 35 (2014) 274e283

Contents lists available

Food Control

journal homepage: www.elsevier .com/locate/ foodcont

Application of a novel pathogenicity marker in a multiplex real-timePCR method to assess total and pathogenic Vibrio vulnificus in food andenvironmental samples

Alejandro Garrido-Maestu, María-José Chapela, Belén Román, Juan M. Vieites,Ana G. Cabado*

Microbiology and Toxins Area, ANFACO-CECOPESCA, Campus Univ. 16, 36310 Vigo, PO, Spain

a r t i c l e i n f o

Article history:Received 13 February 2013Received in revised form25 June 2013Accepted 2 July 2013

Keywords:Vibrio vulnificusvcgCpilFPNCPNCCqPCR

* Corresponding author. Tel.: þ34 986 469 303; faxE-mail address: agcabado@anfaco.es (A.G. Cabado

0956-7135/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.foodcont.2013.07.007

a b s t r a c t

Pathogenic species of Vibrio genus, including Vibrio vulnificus, constitute a great challenge for foodcontrol agencies and a threat for consumers. V. vulnificus can appear in bivalve mollusks such as oysters,clams, and mussels. In addition, water, sediment, and plankton, have been described as reservoirs for thispathogen which constitutes the leading cause of death by consuming seafood in the United States.

The aim of this study was to develop and pre-validate a rapid and reliable multiplex real-time PCR(qPCR) method for total and pathogenic V. vulnificus detection. Peptone, Sodium Chloride, Cellobiose(PNC) broth, with and without Colistin (PNCC) was evaluated according to international methods (ISO).The capacity of these broths to recover low numbers of pathogenic V. vulnificus in the presence of highnumbers of interfering microorganisms was assessed. Finally, were used for food and environmentalsamples enrichment.

In addition three different DNA extraction protocols were compared, but one of them proved to bebetter than the others regarding DNA concentration and purity obtained, and also regarding Ct valuesand final fluorescence obtained by qPCR.

A qPCR efficiency above 90% was obtained, covering five orders of magnitude. The complete methodachieved low limit of detection (3 cfu/25 g). All quality parameters of the method (relative sensitivity,specificity, and accuracy) returned values over 90% after analyzing 45 spiked samples. These results wereobtained for all the targets analyzed (vvhA, vcgC and pilF).

In this study the complete qPCR method developed was applied to 28 natural samples including a widevariety of seafood types and environmental samples (water), but no positive samples were detected foreither target.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

The Gram-negative halophilic bacterium Vibrio vulnificus is anatural inhabitant of marine and estuarine environments world-wide (Baker-Austin, Stockley, Rangdale, & Martinez-Urtaza, 2010),and its presence is especially high in filtering feeding bivalvemollusks such as oysters (Han, Pu, Hou, & Ge, 2009; Panicker, Call,Krug, & Bej, 2004).

Historically V. vulnificus strains have been classified by bio-typing, a technique based on a combination of different phenotypic,serologic, and host range characteristics. Biotype 1 can be found in

: þ34 986 469 269.).

All rights reserved.

warm marine waters and was initially thought to be the onlybiotype associated with human infection. Biotype 2 was firstlythought to be pathogenic only to eels, but this was later disputedbased on human clinical evidence. In 1996, V. vulnificus biotype 3was first described in Israeli patients (Drake, DePaola, & Jaykus,2007; Harwood, Gandhi, & Wright, 2004).

V. vulnificus is an opportunistic human pathogen highly lethaland is responsible for the overwhelming majority of the reportedseafood-related deaths in the United States. In 2009 a total of 825vibriosis were reported in the US, from which 107 (13%) wereassociated with V. vulnificus. Out these 90% needed hospitalizationand 32% died (CDC, 2009, pp. 6). Outbreaks of V. vulnificus havebeen also reported in Europe and Asia (Canigral, Moreno, Alonso,Gonzalez, & Ferrus, 2010; DePaola et al., 2003; Gulig, Bourdage, &Starks, 2005; Harwood et al., 2004; Jones & Oliver, 2009). Fatality

A. Garrido-Maestu et al. / Food Control 35 (2014) 274e283 275

rate for primary septicemia has been reported at greater than 50%.Death can occur within a day or two of the onset of symptoms(Harwood et al., 2004). Additionally, it can produce serious woundinfection typically from exposure of open wounds to waterharboring the bacterium. Like systemic disease, wound infectionsprogress rapidly and they can lead to necrotizing fasciitis at the siteof infection. Mortality rate for wound infections is lower than thatfor systemic disease. Finally gastroenteritis caused by V. vulnificusmay go unreported since the disease is not usually life-threateningand symptoms are typically not severe enough to warrant medicalattention (Drake et al., 2007).

This organism possesses a wide array of virulence factors,including acid neutralization, capsular polysaccharide expression,iron acquisition, cytotoxicity, motility, and expression of proteinsinvolved in attachment and adhesion. Overall, V. vulnificus is acomplex microorganism with physiological characteristics thatcontribute to its survival in the marine environment and in thehuman host (Jones & Oliver, 2009).

Virulence factors are generally present in most strains and theydo not provide predictive value of pathogenicity. Thus, appropriatemarkers to screen the virulence potential of V. vulnificus in envi-ronmental reservoirs are needed (Chatzidaki-Livanis, Hubbard,Gordon, Harwood, & Wright, 2006). Recently, DNA sequence poly-morphisms at individual loci discriminated isolates with clinical oroysters origin in several independent studies. Polymorphic variantsgenerally included two genotypes, such as types A and B of the 16SrRNA gene, whose distribution significantly correlated with eitherenvironmental or clinical origin, respectively. Another importantvirulence factor is the capsular polysaccharide (CPS) which pre-sents 2 different alleles, 1 and 2, correlated with the clinical orenvironmental character (Chatzidaki-Livanis et al., 2006; Guliget al., 2005). Similar genetic distributions were reported for typesE (Environmental) and C (Clinical) of the vcg gene (virulencecorrelated gene, without known function) (Rosche, Yano, & Oliver,2005).

It has been shown that faecal indicator bacteria levels in wateror seafood do not correlate with the presence of pathogenic Vibriospp. Therefore, efficient and sensible techniques are necessary toincrease the rate of detection of this human pathogen (Canigralet al., 2010).

Since 1963, when Thiosulphate Citrate Bile Salt agar (TCBS) wasformulated for the isolation of pathogenic vibrios, several selectivemedia have been developed for the complementation and/orimprovement of TCBS in the isolation of vibrios, as SPS agar, VVagar, and CPC agar, among others; but little effort has been focusedon the development of enrichment broths for V. vulnificus. Hsu et al.optimized an alkaline peptone broth for rapid growth ofV. vulnificus and suppression of nontarget bacteria (Hsu, Wei, &Tamplin, 1998). The authors recommended PNCC broth, contain-ing peptone, sodium chloride, cellobiose, and colistin, as anenrichment broth for environmental V. vulnificus; however, the useof this medium has not been widely reported in the literature(Harwood et al., 2004).

The present study was focused on the development and pre-validation of a novel multiplex qPCR detection method for totaland pathogenic V. vulnificus. The method should be able to reliablyidentify not only biotype 1 strains, but also pathogenic strains ofbiotypes 2 and 3. Two individual objectives were assessed: first,evaluation of the suitability of the scarcely used PNC/PNCC broth forsample enrichment to achieve a low limit of detection in thepresence of interfering bacteria, and to compare it to the ISOreference broth; second, combination of a species specific gene forthe detection of total V. vulnificus (vvhA, (Campbell & Wright,2003)), a well known pathogenic related marker, the virulencecorrelated gene (vcgC, (Rosche et al., 2005)), and a novel

pathogenicity marker (pilF, (Roig, Sanjuan, Llorens, & Amaro,2010)). A previously applied Internal Amplification Control (IAC)(Calvo, Martinez-Planells, Pardos-Bosch, & Garcia-Gil, 2008) wasalso added to the qPCR method to ensure absence of false negativeresults due to PCR reaction inhibition.

2. Materials and methods

2.1. Bacterial strains and culture media

Spanish Type Culture Collection (CECT) strains used as referencestrains for evaluation of PNC and PNCC (PNC plus colistin, 1 U/mL)broth were: V. vulnificus CECT 529, V. vulnificus CECT 4608 andV. vulnificus CECT 4869. These reference strains were cultured inNutrient Broth (NB, Biokar diagnostics S.A., France) supplementedwith 1e2% of sodium chloride (SNB), and incubated at 37 �Covernight.

Other bacterial strains used as interfering microorganisms andfor the evaluation of the qPCR specificity, were cultured in TrypticSoy Broth (TSB, BioMérieux S.A., France) overnight at 37 �C, seeTable 1.

Productivity of PNC/PNCC was compared against Saline AlkalinePeptone Water (SAPW, Biolife Italiana S.r.l., Italy) the InternationalStandards Organization (ISO) recommended broth.

2.2. Enrichment broth comparison

Evaluation of different enrichment broths was performed basedon recommendations of ISO 11133-2 (ISO, 2003a) in two steps. First,bacterial growth was evaluated in pure culture. Second, mixedcultures of V. vulnificus along with interfering bacteria were done.For the evaluation of PNC, PNCC and SAPW, V. vulnificus CECT 4608,a pathogenic strain, was selected as the reference strain.

Inoculum level of target microorganism, was established at 10e100 cfu. To obtain the required inoculum concentration from theSNB overnight culture, ten-fold serial dilutions were done inAlkaline PeptoneWater (APW, Biolife Italiana S.r.l., Italy) and platedon Saline Nutrient Agar (SNA, Biokar diagnostics S.A., France) whichcontains 10 g/L of sodium chloride. Plateswere incubated overnightat 37 �C to calculate reference values. Productivity evaluation ofeach different step was done in triplicate.

2.2.1. Pure culturesIn the first step, tubes containing 10 mL of SAPW, PNC and PNCC

were inoculated with target bacterium. Tubes were incubated at35 �C for 22 � 2 h (optimal temperature) according to a previousstudy (Hsu et al., 1998). After enrichment ten-fold serial dilutionswere done in APW and plated in SNA. Plates were incubated at37 �C overnight.

2.2.2. Mixed culturesThe second step of the evaluation of the productivity consisted

on the inoculation of 100 mL of the different broths under studywith the same concentration of target microorganism as previouslymentioned and also non-target bacteria, see Table 1, in a higherconcentration. Bottles with mixed cultures were incubated underthe same conditions as in Materials and methods 2.2.1. Afterenrichment ten-fold serial dilutions were done in APW and platedon CHROMagar�Vibrio (CHROMagar Microbiology, Paris, France).Plates were incubated at 37 �C overnight.

2.3. Genes, primers and probes used for qPCR method

On the one hand, detection of total V. vulnificus was done tar-geting the species specific gene vvhA as previously described vvha-

Table 1Strains used to test: productivity and selectivity of media and specificity of primers and probes (qPCR result).

Bacteria Strain qPCR result vvhA, pilF, vcgCc Bacteria Strain qPCR result vvhA, pilF, vcgCc

V. vulnificusa,b CECT 4608 þ/þ/þ E. faecalisb CECT 481 �V. vulnificusb CECT 529 þ/�/� A. hydrophilab CECT 839 �V. vulnificusb CECT 4869 þ/�/� P. putidab CECT 324 �V. vulnificusb CAIM 611 þ/þ/� P. aeruginosab CECT 108 �V. parahaemolyticusb CECT 511 � P. fluorescensb CECT 378 �V. parahaemolyticusb CECT 5271 � E. colib CECT 516 �V. parahaemolyticusb CCUG 43362 � E. colib CECT 434 �V. parahaemolyticusb CCUG 43363 � C. freundiib CECT 401 �V. parahaemolyticusb CCUG 43364 � S. aureusb CECT 240 �V. parahaemolyticusb CCUG 43365 � S. aureusb CECT 435 �V. parahaemolyticusb CAIM 58 � S. entericab CECT 4594 �V. choleraeb CECT 514 (O1) � S. sonneib CECT 413 �V. choleraeb CCUG 47460 (O139) � S. flexnerib CECT 4804 �V. alginolyticusb CECT 586 � L. monocytogenesb CECT 935 �V. alginolyticusb CAIM 342 � L. innocuab CECT 910 �V. alginolyticusd,b (internal reference) � L. seeligerib CECT 917 �V. mimicusb CECT 4218 � L. ivanoviib CECT 913 �V. mimicusb BCCM/LMG 7896 �

CECT: Spanish Type Culture Collection, CCUG: Culture Collection University of Göteborg, CAIM: Collection of Aquatic Important Microorganisms, BCCM/LMG: Belgian Co-Orfinated Collections Of Micro-Organsims.

a Productivity.b Specificity of primers and probes.c Unless otherwise stated negative sign indicates no amplification of all three qPCR targets.d Strain identified in the laboratory of ANFACO-CECOPESCA.

A. Garrido-Maestu et al. / Food Control 35 (2014) 274e283276

F: TGTTTATGGTGAGAACGGTGACA, vvha-R: TTCTTTATCTAGG-CCCCAAACTTG, vvha-Probe: (5Cy5)-CCGTTAACCGAACCACCCG-CAA-(3BHQ_2) (Campbell & Wright, 2003). On the other hand, theassessment of the pathogenic character of V. vulnificus was donetargeting two genes, the newly described pilF polimorfism PilF-F:GATTGACTACGAYCCACACCG, PilF-R: GRCGCGCTTGGGTGTAG, PilFProbe: (56-FAM)-TGCTCAACCTCGCTAAGTTGGAAATCGATAC-(3BHQ_1)

as previously described (Baker-Austin et al., 2012).Regarding vcgC new primers and probe were designed for our

multiplex qPCR detection system within the original 505 bp frag-ment. For this purpose several sequences of the genewere obtainedfrom the GenBank (AY626575.1, AY626576.1, AY626577.1,AY626578.1, and EU851904.1), aligned with CLC Sequence Viewer 6(A/S). Primers and probe designwas carried out with the free onlineapplication Primer3 (Untergasser et al., 2007). Specificity of primersand probe designed was verified with BLAST (Basic LocalAlignment Search Tool, http://blast.ncbi.nlm.nih.gov/Blast.cgi), andagainst bacterial strains specified in Table 1. Primers and probedesigned were: PvcgC-F: AGTTCAAACATGGTCTCAAAAAGGAG,PvcgC-R: CGATAACTCATTGTTTTCGTTACTG, vcgC-Probe: (5HEx)-CACTAATGTGTCATCTGAACAGGCTATTG-(3IABkFQ).

Internal control based on chimerical DNA, published by Calvoet al. (2008) was selected to monitor qPCR reaction. Primersand probe were: IAC F: TCCAGGGCGAAAGTAAACGT, IAC R:GGCGAGCCGTACGAACAC, IAC Probe: (5TexRd-xN)-CCCAGTTGGCT-GATCACTTTCG-(3BHQ_2), IAC DNA: TCCAGGGCGAAAGTAAACGTNN-NCCCAGTTGGCTGATCACTTTCGNNGTGTTCGTACGGCTCGCC.

2.4. DNA extraction methods

DNA extraction from pure cultures was done as previouslydescribed by Blanco-Abad et al. (Blanco-Abad, Ansede-Bermejo,Rodriguez-Castro, & Martinez-Urtaza, 2009). Regarding extractionfrom food samples, threemethodswere compared: boiling (method1), a previously published method using Chelex 100 method wasalso tested (Malorny et al., 2004), (method 2) and a modification ofthe Chelex 100 method (method 3, modification consisted on anadditional washing step of bacterial pellet with PBS). An initialcentrifugation step of 2000 rpm for 2minwas done to eliminate bigfood particles, and supernatant was transferred to a new clean tube

in methods 1 and 3. DNA obtained from all 3 methods was storedat �20 �C until use. All the protocols are summarized in Fig. 2.

DNAwas quantified and purity assessed, using a NanoDrop 1000spectrophotometer (Thermo Fischer Scientific, Inc., USA) softwareND-1000 v3.7.1.Methodswere compared according to the amountofDNA obtained (ng/mL), 260/280, 260/230 ratios, Cycle Threshold (Ct)obtainedbyqPCRanddR last value (R is the rawfluorescence, and thedR value is dR baseline-corrected raw fluorescence which is calcu-lated as final fluorescence minus the initial fluorescence) for vvhA.

For the comparison of both methods, pure bacterial cultures ofV. vulnificus CECT 4608 were used and also one oyster sample wasinoculated with the same strain, and after enrichment in PNC allthree methods were applied. To compare pure cultures all param-eters previously mentioned were compared; for the inoculatedsample only the last two were taken into account. Three replicateswere done for each experiment.

2.5. Multiplex qPCR detection method for V. vulnificus

The qPCR reactionwas carried out in a final volume of 25 mL withthe following components: 12.5 mL of Brilliant III Ultra-Fast QPCRMaster Mix (Agilent Technologies, Inc., USA), 700 nM primers and250 nM probe were used for vvhA, 75 nM primers and 200 nMprobe were used for pilF; 300 nM primers and 200 nM probe wereused for vcgC and for Internal Amplification Control (IAC) 150 nMprimers, 45 nM probe and 8 � 102 copies of IAC DNA were addedper reaction. Two microliters of template DNA was added per re-action tube.

Stratagene Mx3005p thermocycler (Agilent Technologies, Inc.,USA) was used with the following thermal profile: 3 min at 95 �Cfor the activation of the polymerase (Hot Start), followed by 40cycles, each cycle consisted on a denaturation step of 15 s at 95 �C,and annealing-extension step at 60 �C for 60 s.

2.6. qPCR efficiency

Calculation of qPCR efficiency was done culturing overnightV. vulnificus CECT 4608 in 10 mL of SNB at 37 �C. This strain wasselected since it presents all three target genes. DNA extractionwas

Fig. 1. vcgC gene sequence alignment with primers and probe designed in the present study.

A. Garrido-Maestu et al. / Food Control 35 (2014) 274e283 277

accomplished as described in Materials and methods 2.3.2, andmeasured with a NanoDrop 1000 spectrophotometer.

Initial DNA templatewas ten-fold serially diluted in sterile milli-Q water, and two mL of each dilutionwas used as template for qPCR.Efficiency was determined in triplicate for total, and for eachpathogenicity gene of V. vulnificus.

The Mx3005pro software automatically calculates the stan-dard curve for each run based on the Cycle threshold (Ct)for each standard versus DNA concentration in ng. The formula

Fig. 2. Schematic representation of DN

from which the amplification efficiency was calculated ise ¼ 10�1/s�1, where “s” is the slope of the standard curve(Blackstone et al., 2007; Nordstrom, Vickery, Blackstone, Murray,& DePaola, 2007).

2.7. Sampling and sample preparation

All food samples were received at the laboratory from externalsuppliers, either frozen or refrigerated, and were kept in the same

A extraction methods compared.

A. Garrido-Maestu et al. / Food Control 35 (2014) 274e283278

conditions until analysis. Regarding water samples, were collectedin sterile 500 mL plastic bottles, and stored at 4 �C until analysis.

Food samples were prepared by weighting 25 g of sample plus225 mL of PNC. Complete matrix was homogenized for 30 s in alaboratory stomacher at normal speed. Water samples were pro-cessed as previously described (Garrido, Chapela, Ferreira, et al.,2012), briefly 500 mL or the whole sample volume, if less than500 mL were available, were filtered through a 0.45 mmmembrane(Millipore, Ma USA); this filter was placed in a stomacher bag with50 mL of PNC, homogenized for 30 s. All samples were incubated at35 �C for 22 � 2 h.

2.8. Limit of detection (LOD)

Raw oysters were inoculated for the evaluation of the LOD,which included a total of 22 samples. Two sets of 10 spiked sampleswith one black, all weighting 25 g, were used to determine the LOD.Raw oysters of each set of 11 samples, belonged to distinct batchesand were received different days.

V. vulnificus CECT 4608 was used for the evaluation of the LOD.Bacterium was grown in 10 mL of SNB at 37 �C overnight and ten-fold serially diluted in APW. All 20 samples were inoculated withone mL with anexpected concentration below 10 cfu/mL. Serialdilutions used to inoculate the samples, were also seeded on SNAplates in order to get a reference value of viable bacteria.

2.9. qPCR method evaluation

Evaluation of the method was done by calculating several pa-rameters: relative sensitivity (SE), relative specificity (SP), andrelative accuracy (AC), as previously described (Tomas, Rodrigo,Hernandez, & Ferrus, 2009).

Evaluation of these parameters was done by comparing the re-sults obtained by qPCR against the expected values of the spiked orblind, previously analyzed, samples.

Each sample with Positive (PA) and Negative (NA) Accordancewere defined as samples presenting the same result, positive ornegative, for the qPCR method and the expected results for spikedsamples. Negative Deviations (ND) include the number of samplesexpected positive with a negative result, and Positive Deviations(PD), represent the number of samples expected negative with apositive result.

2.10. Statistical analysis

One-way ANOVA, Tukey b (p< 0.05) was used to evaluate SAPW,PNC and PNCC broth productivity data, as well as DNA extractionmethods. These analyses were performed with SPSS 15.0 software(SPSS Inc., Chicago, IL, USA).

3. Results

3.1. Specificity of qPCR method

DNA obtained from bacteria in Table 1 was used as template forqPCR following specifications mentioned in Materials and Methods2.5.

Amplification for vvhA gene was obtained for all four strains ofV. vulnificus tested in the present study. Regarding vcgC gene pos-itive results were only observed for strain CECT 4608 (alignmentalong with primers and probe design can be seen in Fig. 1). FinallypilF was detected in strains CECT 4608 and CAIM 611.

According to BLAST tool, primers and probe designed for vcgCwere specific for V. vulnificus. When specificity was verified for thethree genes against the strains listed in Table 1, no amplification

was observed for any of the 30 non V. vulnificus strains tested,which covered 14 strains of the genus Vibrio, including the mainpathogenic species (Vibrio cholerae, Vibrio parahaemolyticus, Vibrioalginolyticus, and Vibrio mimicus) and 16 strains of other differentspecies which may be found in the same aquatic environment thanV. vulnificus. Amplification plots obtained with strain CECT 4608 areshown in Fig. 3.

3.2. Enrichment broth comparison

The evaluation of the different enrichment broths was per-formed based on the specifications of ISO 11133 (ISO, 2003b), in twoconsecutive steps. The first step consisted on the evaluation of eachbroth for the enrichment of pure cultures; and the second stepconsisted on the evaluation of each broth to recover V. vulnificus inthe presence of interfering bacteria.

3.2.1. Pure culturesTubes containing 10 mL of all three media under comparison

(SAPW, PNC and PNCC) were inoculatedwith a lowconcentration oftarget bacterium. Tubes were incubated, and after enrichment ten-fold serial dilutions were plated. Finally plates were incubated at37 �C overnight.

Viable bacterial counts, expressed as log cfu/mL, show signifi-cantly higher values for PNC and PNCC (8.72 � 0.09, and8.80 � 0.07 respectively) than for SAPW (8.05 � 0.42), howeverwithout differences among them, according to the statistical testapplied.

3.2.2. Mixed culturesEvaluation of the capacity of the media to recover V. vulnificus

in the presence of competing bacteria, was done inoculating100 mL of broth with target microorganism (10e100 cfu) andalso with non-target bacteria, see Table 1, in a higher concen-tration (more than 103 cfu). Bottles with mixed cultures wereincubated under the same conditions described in Materials andmethods 2.2.1. After enrichment ten-fold serial dilutions weredone, plated and incubated as previously described in Materialsand methods.

Viable counts, expressed as log cfu/mL, were statisticallycompared, and once more higher values were obtained for PNC andPNCC (7.11 � 0.32 and 7.40 � 0.04) than for SAPW (6.21 � 1.15).

3.3. DNA extraction methods

Three different DNA extraction methods were statisticallycompared, a simple boiling protocol (method 1) and two based onChelex 100 treatment (method 2) which differed, from one anotherbasically, in an additional PBS washing step (modified Chelex 100protocol, method 3).

Significantly higher DNA concentration was obtained formethod 1, followed bymethod 3, and the lowest concentrationwasobtained for method 2. Regarding purity, 260/280 ratio for methods1 and 3 obtained the highest values without differences amongthem, and again the lowest was obtained for method 2. Concerning260/230 ratio, the highest value was achieved with method 1, fol-lowed bymethods 2 and 3. No statistical differences were observedfor either Ct or dR last values obtained with all three methodsapplied to pure cultures.

qPCR analyses of spiked oyster showed that method 3 presentedstatistically lower Ct values than the other two protocols, which didnot show differences among them. Regarding dR last values,method 1 showed significantly lower data, and statistically highervalues were obtained for methods 2 and 3. All data are summarizedin Table 2.

Fig. 3. Amplification plots obtained by qPCR with strain CECT 4608 for genes vvhA, pilF, vcgC and the Internal Amplification Control (IAC).

A. Garrido-Maestu et al. / Food Control 35 (2014) 274e283 279

3.4. qPCR efficiency

The software automatically calculates the standard curve foreach run based on the Ct for each standard (ten-fold serial dilutionsfrom initial DNA extract). In order to assess the efficiency of themethod developed in the present study it was calculated in 3different experiments.

Average efficiency over 95% was obtained for every gene (vvhA97.8 � 1.2%, pilF 98.9 � 3.8%, vcgC 96.1 � 3.6%), covering 5 orders ofmagnitude. Amplification of all targets gave a correlation coeffi-cient higher than 0.99. Amplification efficiency plots obtained withstrains CECT 4608 are shown in Fig. 4.

Table 2Ct and dR values in pure cultures and inoculated oyster samples.

Pure culture Ct Pure culture dR last Pure culture DNA

Method 1 20.31 � 0.48a 7592.72 � 911.27a 44.11 � 1.28a

Method 2 22.86 � 2.40a 7339.06 � 3775.29a 14.74 � 3.31c

Method 3 19.42 � 0.31a 9580.98 � 769.58a 28.11 � 1.09b

Within same column, different letters indicate statistically significant differences amongMethod 1: boiling, Method 2: Chelex 100, Method 3: modified Chelex 100.DNA concentration expressed in ng/mL.dR Last indicates final fluorescence obtained.

3.5. Limit of detection (LOD)

Evaluation of the LOD was done twice by spiking raw oystersfrom two different batches and received at the laboratory differentdays.

Blank samples of each trial resulted negative for total andpathogenic V. vulnificus. Out of the 20 inoculated samples 18resulted positive (90%) for the three pathogens with an initialinoculation level of 3 cfu/mL; thus the LOD of the method wasestablished in 3 cfu in 25 g of sample.

A total of 28 natural, non inoculated samples were analyzedwith the method described. These samples covered oysters (main

260/280 260/230 Oyster Ct Oyster dR last

1.89 � 0.03a 1.26 � 0.03a 21.73 � 0.27b 2903.68 � 322b

1.43 � 0.14b 0.52 � 0.08c 23.31 � 1.26b 6050.89 � 1451.04a

1.93 � 0. 12a 0.93b 19.62 � 0.06a 7099.15 � 29.45a

data (p < 0.05).

Fig. 4. Efficiency plots generated for target genes vvhA, pilF, vcgC of strain CECT 4608 obtained by representing Ct values versus DNA quantity in nanograms.

A. Garrido-Maestu et al. / Food Control 35 (2014) 274e283280

source of infection with this bacterium), other bivalves, crusta-ceans, fish and water. None of the samples analyzed resulted pos-itive for the genes studied. Results obtained for both, inoculatedand natural samples, are listed in Table 3.

3.6. qPCR method evaluation

Evaluation of the method was done as previously described byTomas et al. (2009), thus relative sensitivity (SE), specificity (SP)and accuracy (AC) parameters were calculated. Samples inoculatedfor the evaluation of these parameters, were spiked either with one,two or three strains, as detailed in Table 3.

All three parameters showed values over 90%, even thoughslight differences were observed among the genes analyzed, due totwo ND for all three targets and three for pilF and vcgC. Valuesobtained for pilF and vcgCwere 93% for SE and AC, and 100% for SP;regarding vvhA values were higher achieving 95% for SE,100% for SPand 96% for AC. Only one ND was observed associated with inoc-ulation with mixed cultures as specified in Table 3. All 28 naturalsamples analyzed were negative for all three targets, see Table 4.

4. Discussion

V. vulnificus is an opportunistic human pathogen that may causegastroenteritis, severe necrotizing soft-tissue infections and pri-mary septicemia, with a high lethality rate. Illness is associated tothe ingestion of seafood or to the exposure of contaminated water

(Canigral et al., 2010). Currently shellfish industry faces manyconcerns and the most critical is the presence of V. vulnificus inoysters (Rosche et al., 2005).

The Centers for Diseases Control and Prevention (CDC) alongwith the FDA and the Gulf Coast states initiated a Cholera and OtherVibrio Illness Surveillance System (COVIS) in 1988, and by 1997nearly all states of USA were voluntarily reporting to this database.It was observed that in a five year period (2005e2009) the numberof vibriosis due to V. vulnificus was relatively stable, around 100cases per year with an average mortality rate of 32% (http://www.cdc.gov/nationalsurveillance/cholera_vibrio_surveillance.html).Even though the number of cases is relatively low, the mortalityrate is extremely high. So far, in Europe there has not been a similarinitiative. The European Union Regulation (EC) No. 2073/2005((EC), 2005) sets out the microbiological criteria for foodstuff. Thisregulation makes no provision for Vibrio controls in seafood tradedwithin the European Community. Thus currently, in the EU there isno legal basis for Vibrio testing. This is in part due to the recom-mendations of the EU expert scientific committee on veterinarymeasures related to public health (SCVMPH) who advised thatexisting internationally recognized methods were not sufficientlyfit for purpose and that available data did not support specificstandards for V. vulnificus in raw and undercooked seafood (Baker-Austin, Stockley, et al., 2010).

In the present study detection of V. vulnificuswas done targetingvvhA gene, known to encode for an extracellular cytolysin/hemo-lysin, which was originally thought to be involved in pathogenesis.

Table 3Spiked samples analyzed.

Number of samples Strain used for inoculation qPCR result Observations

CECT 529 CECT 4869 CECT 4608 vvha pilF vcgC

Oystersa 20 � � + + + + Two ND for all 3 genesOysters 3 � � � � � �

2 + � � + � �2 � � + + + +2 � + � + � �1 + + � + � �1 + � + � + � One ND for pilF and vcgC2 + + + + + +

Bivalves 1 � � + + + +1 � + � + � �

Water 2 � � � � � �2 + � � + � �3 � � + + + +2 � + � + � �

Crustaceans 1 � � + + + +

a Spiked samples used for evaluation of the LOD. ND: Negative Deviation.

A. Garrido-Maestu et al. / Food Control 35 (2014) 274e283 281

Later on this role was dismissed and proved to be species specificand was selected for total V. vulnificus detection (Campbell &Wright, 2003; Jones & Oliver, 2009). Even though previousstudies suggested a lack of specificity of primers and probedesigned by Campbell and Wright (2003) (Panicker & Bej, 2005),later studies by original authors confirmed originally describedspecificity by obtaining those strains from which the lack of spec-ificity was reported, and re-analyzing them (Wright et al., 2007).

Concerning pathogenicity, viuB was previously proposed as aclassical pathogenicity marker according to previous studies (El-Lathy, El-Taweel, El-Sonosy, Samhan, & Moussa, 2009; Han et al.,2009; Panicker et al., 2004), but later studies suggested that thepresence of this gene may not be related with pathogenic strains.Furthermore it was suggested that it was present in most strains ofV. vulnificus but showed allelic variation (Bogard & Oliver, 2007,2008; Swain et al., 2010), this was in agreement with the resultsobtained in our laboratory. Thus viuB was discarded and two otherdifferent targets were selected for the evaluation of the pathogeniccharacter of the bacteria. First vcgC gene, from which previousstudies have demonstrated that in V. vulnificus biotype 1 is stronglycorrelatedwith potential virulence in humans, andmayalso be usedfor typing biotypes 2 and 3 even though it has limited usefulnesswith these biotypes (Baker-Austin, Gore, et al., 2010; Baker-Austinet al., 2012; Rosche et al., 2005; Warner & Oliver, 2007, 2008). Thesecond pathogenicity target chosen was the variability in pilF gene,which codes for a protein required for pilus-type IV assembly,whosemutation in some bacterial pathogens implies attenuated virulence(Baker-Austin et al., 2012; Roig et al., 2010). The later target has beenscarcely applied for typing strains, and to our knowledge this is thefirst report to test it against food and/or water samples.

Recent studies have demonstrated the utility of combiningseveral molecular virulence testing approaches simultaneously

Table 4Natural samples analyzed by qPCR.

Type Number of samples qPCR result

vvhA pilF vcgC

Oysters 6 � � �Bivalves 10 � � �Fish 6 � � �Water 4 � � �Crustaceans 2 � � �

vvhA: extracellular hemolysin gene used for species specific detection. pilF: novelvirulence marker, PilF polymorphism. vcgC: conventional virulence marker, viru-lence correlated gene type C.

(Baker-Austin et al., 2012). In this regard the present study com-bined previously described primers and probes for specific detec-tion of a species specific target, vvhA, andwell knownpathogenicityindicator, vcgC (primers and probe designed in the present study),with a recently described one, pilF. Previous studies have appliedeither one or two of these targets (Baker-Austin et al., 2012;Campbell & Wright, 2003) but this is the first study to combinethese three genes with a previously developed and extensivelyapplied IAC (Calvo et al., 2008; Garrido, Chapela, Ferreira, et al.,2012; Garrido, Chapela, Roman, et al., 2012), being this last ahighly recommended criterion for diagnostic methods.

BLASTand qPCR verificationswere carried out against target andnon-target strains to evaluate the specificity of primers and probes,especially those designed in the present study (vcgC). Resultsconfirmed the correct designed and selection made in all casesstudied and the evaluation of multiplex qPCR efficiency reportedvalues between 96% and 99%, being these between previouslydescribed acceptable limits (90e110%) (Raymaekers, Smets, Maes,& Cartuyvels, 2009).

qPCR results obtained for the DNA extraction methodscompared are in agreement with previous studies, which reportedthat the addition of oyster tissue produced a significant loss ofsensitivity of qPCR methods (Wright et al., 2007). A modification ofa previous DNA extraction (Malorny et al., 2004) was evaluated. Itwas proved to be statistically better than the original one, regardingto DNA concentration and purity obtained in pure cultures; and alsoin Ct and dR values in spiked samples.

The complete method was able to specifically detect both, totaland pathogenic V. vulnificus, achieving a very low LOD (3 cfu/25 g),comparable to other studies performed with other pathogenicVibrio spp. by qPCR and traditional culture methods (Chapela et al.,2010; Garrido, Chapela, Ferreira, et al., 2012). Other authors havepreviously used three different concentrations for the evaluation ofthe LOD (high, medium and low). The present study was focusedonly in the detection of the lowest bacterial concentration. Otherspiked samples were inoculated above the LOD, achieving correctdetection. Also high SE, SP, and AC values were obtained (all over90%). Furthermore, this is the first report to completely evaluatePNC/PNCC broth according to ISO requirements (ISO, 2003b) (PNCCwas prepared with 1 U/mL of colistin as it has been reported thathigher concentrations may inhibit low concentrations ofV. vulnificus (Hsu et al., 1998)), and extensively apply it to naturaland spiked sample screening. Results obtained proved that thisbroth was suitable for the enrichment and detection of V. vulnificuseven in the presence of high numbers of interfering bacteria. A total

A. Garrido-Maestu et al. / Food Control 35 (2014) 274e283282

of 45 blind samples, including those of the LOD. Only three NDweredetected, but this results were explained due to the very lowinocula concentration used (lower than 5 cfu/mL, thus higher de-viations in bacterial counts may be expected, and probably thesesamples were below the LOD). No significant effect was observedassociated with the number of strains used to inoculate thesamples.

No natural food or water samples were found positive for, eithertotal or pathogenic V. vulnificus. This is in agreement with previousstudies, which reported that in contrast to Asian countries and theUSA, non-cholera Vibrio infections are less often reported in Europe(Baker-Austin, Stockley, et al., 2010; Kuhnt-Lenz, Krengel, Fetscher,Heer-Sonderhoff, & Solbach, 2004). V. vulnificus occurs naturally,and is not pollution associated, in temperate and coastal watersworldwide, but it is most frequently isolated when water temper-atures are above 20 �C and salinities are between 5 and 25%.Because of such environmental restrictions, there have been fewcases reported in the Mediterranean, presumably because the highsalinity (38%) of the water body (Oliver, 2005). Regarding Spain,according to data gathered up to 2010, the risk of vibriosis is verylow (AESAN, 2010), even though the presence of this bacterium hasbeen reported in different sources in Spain and other Europeancountries (Canigral et al., 2010; Oliver, 2005; Schaerer, Savioz,Cernela, Saegesser, & Stephan, 2011).

5. Conclusions

The present study demonstrated, after extensive evaluationaccording to internationally acceptance criteria, that PNC/PNCCmay be used for selective enrichment of food samples intended fordetection of V. vulnificus. Furthermore, it may be applied combinedwith a qPCR detection method for fast and reliable detection ofboth, total and pathogenic V. vulnificus in food and environmentalsamples.

Complete monitoring of total and pathogenic V. vulnificus can beachieved in less than two hours after enrichment, with a very lowLOD (including DNA extraction and qPCR) thus making this methodsuitable for sensitive analysis of food products with very shortshelf-life or any other products when fast results are needed. It maybe also used for environmental monitoring of this pathogenicbacterium in water samples, and for research purposes.

Acknowledgments

This work is financially supported by the Secretary General forthe Sea of the Spanish Ministry of Agricultural, Land and MarineResources (MARM), by order ARM/1193/2009.

Authors also thank Victoria Docampo and Angeles Marcote fortechnical assistance.

References

AESAN. (2010). Report from the Scientific Commitee of the Spanish Agency of FoodSafety and Nutrition about the microbiological criteria for the pathogenicspecies of Vibrio genus, as additional means of control in the border inspectionpoints, to imported fish products. In Scientific’s Commitee Magazine, 12.

Baker-Austin, C., Gore, A., Oliver, J. D., Rangdale, R., McArthur, J. V., & Lees, D. N.(2010). Rapid in situ detection of virulent Vibrio vulnificus strains in raw oystermatrices using real-time PCR. Environmental Microbiology Reports, 2(1), 76e80.

Baker-Austin, C., Lemm, E., Hartnell, R., Lowther, J., Onley, R., Amaro, C., et al. (2012).pilF polymorphism-based real-time PCR to distinguish Vibrio vulnificus strainsof human health relevance. Food Microbiology, 30(1), 17e23.

Baker-Austin, C., Stockley, L., Rangdale, R., & Martinez-Urtaza, J. (2010). Environ-mental occurrence and clinical impact of Vibrio vulnificus and Vibrio para-haemolyticus: a European perspective. Environmental Microbiology Reports, 2(1),7e18.

Blackstone, G. M., Nordstrom, J. L., Bowen, M. D., Meyer, R. F., Imbro, P., & DePaola, A.(2007). Use of a real time PCR assay for detection of the ctxA gene of Vibrio

cholerae in an environmental survey of mobile bay. Journal of MicrobiologicalMethods, 68(2), 254e259.

Blanco-Abad, V., Ansede-Bermejo, J., Rodriguez-Castro, A., & Martinez-Urtaza, J.(2009). Evaluation of different procedures for the optimized detection of Vibrioparahaemolyticus in mussels and environmental samples. International Journalof Food Microbiology, 129(3), 229e236.

Bogard, R. W., & Oliver, J. D. (2007). Role of iron in human serum resistance of theclinical and environmental Vibrio vulnificus genotypes. Applied and Environ-mental Microbiology, 73(23), 7501e7505.

Bogard, R. W., & Oliver, J. D. (2008). Role of iron in human serum resistance of theclinical and environmental Vibrio vulnificus genotypes (vol. 53, pg 671, 2007).Applied and Environmental Microbiology, 74(10), 3322e3322.

Calvo, L., Martinez-Planells, A., Pardos-Bosch, J., & Garcia-Gil, L. J. (2008). A newreal-time PCR assay for the specific detection of Salmonella spp. targeting thebipA gene. Food Analytical Methods, 1(4), 236e242.

Campbell, M. S., & Wright, A. C. (2003). Real-time PCR analysis of Vibrio vulnificusfrom oysters. Applied and Environmental Microbiology, 69(12), 7137e7144.

Canigral, I., Moreno, Y., Alonso, J. L., Gonzalez, A., & Ferrus, M. A. (2010). Detection ofVibrio vulnificus in seafood, seawater and wastewater samples from a Medi-terranean coastal area. Microbiological Research, 165(8), 657e664.

(CDC), C. f. D. C. a. P. (2009). COVIS annual summary. Atlanta, Georgia: US Depart-ment of Health and Human Services, CDC.

Chapela, M. J., Fajardo, P., Garrido, A., Cabado, A. G., Ferreira, M., Lago, J., et al. (2010).Comparison between a TaqMan polymerase chain reaction assay and a culturemethod for ctx-positive Vibrio cholerae detection. Journal of Agricultural andFood Chemistry, 58(7), 4051e4055.

Chatzidaki-Livanis, M., Hubbard, M. A., Gordon, K., Harwood, V. J., & Wright, A. C.(2006). Genetic distinctions among clinical and environmental strains of Vibriovulnificus. Applied and Environmental Microbiology, 72(9), 6136e6141.

CLC bio. A. S. CLC Sequence Viewer (6.6.2 ed.).DePaola, A., Nordstrom, J. L., Dalsgaard, A., Forslund, A., Oliver, J., Bates, T., et al.

(2003). Analysis of Vibrio vulnificus from market oysters and septicemiacases for virulence markers. Applied and Environmental Microbiology, 69(7),4006e4011.

Drake, S. L., DePaola, A., & Jaykus, L. A. (2007). An overview of Vibrio vulnificus andVibrio parahaemolyticus. Comprehensive Reviews in Food Science and Food Safety,6(4), 120e144.

(EC), C. R. (2005). Microbiological criteria for foodstuffs. Official Journal of the Eu-ropean Union, 2073/2005.

El-Lathy, M. A., El-Taweel, G. E., El-Sonosy, W. M., Samhan, F. A., & Moussa, T. A. A.(2009). Determination of pathogenic bacteria in wastewater using conventionaland PCR techniques. Environmental Biotechnology, 5(2), 73e80.

Garrido, A., Chapela, M. J., Ferreira, M., Atanassova, M., Fajardo, P., Lago, J., et al.(2012). Development of a multiplex real-time PCR method for pathogenicVibrio parahaemolyticus detection (tdhþ and trhþ). Food Control, 24(1e2),128e135.

Garrido, A., Chapela, M. J., Roman, B., Ferreira, M., Lago, J., Vieites, J. M., et al.(2012). Development of a multiplex real-time PCR method for simultaneousdetection of Salmonella enterica, Shigella flexneri and Listeria monocytogenesin processed food samples. European Food Research and Technology, 234(4),571e580.

Gulig, P. A., Bourdage, K. L., & Starks, A. M. (2005). Molecular pathogenesis of Vibriovulnificus. Journal of Microbiology, 43, 118e131.

Han, F. F., Pu, S. H., Hou, A. X., & Ge, B. L. (2009). Characterization of clinical andenvironmental types of Vibrio vulnificus isolates from Louisiana oysters. Food-borne Pathogens and Disease, 6(10), 1251e1258.

Harwood, V. J., Gandhi, J. P., & Wright, A. C. (2004). Methods for isolation andconfirmation of Vibrio vulnificus from oysters and environmental sources: areview. Journal of Microbiological Methods, 59(3), 301e316.

Hsu, W. Y., Wei, C. I., & Tamplin, M. L. (1998). Enhanced broth media for selectivegrowth of Vibrio vulnificus. Applied and Environmental Microbiology, 64(7),2701e2704.

ISO. (2003a). In Microbiology of food and animal feeding stuffs e Guidelines onpreparation and production of culture media e Part 2: Practical guidelines onperformance testing of culture media, 11133-2 (pp. 27).

ISO. (2003b). In (1st ed.).Microbiology of food and animal feeding stuffs e Guidelineson preparation and production of culture media e Part 2: Practical guidelines onperformance testing of culture media. Amendment 1, 11133-2 (pp. 27).

Jones, M. K., & Oliver, J. D. (2009). Vibrio vulnificus: disease and pathogenesis.Infection and Immunity, 77(5), 1723e1733.

Kuhnt-Lenz, K., Krengel, S., Fetscher, S., Heer-Sonderhoff, A., & Solbach, W. (2004).Sepsis with bullous necrotizing skin lesions due to Vibrio vulnificus acquiredthrough recreational activities in the Baltic Sea. European Journal of ClinicalMicrobiology & Infectious Diseases, 23(1), 49e52.

Malorny, B., Cook, N., D’Agostino, M., De Medici, D., Croci, L., Abdulmawjood, A.,et al. (2004). Multicenter validation of PCR-based method for detection ofSalmonella in chicken and pig samples. Journal of AOAC International, 87(4),861e866.

Nordstrom, J. L., Vickery, M. C. L., Blackstone, G. M., Murray, S. L., & DePaola, A.(2007). Development of a multiplex real-time PCR assay with an internalamplification control for the detection of total and pathogenic Vibrio para-haemolyticus bacteria in oysters. Applied and Environmental Microbiology, 73,5840e5847.

Oliver, J. D. (2005). Wound infections caused by Vibrio vulnificus and other marinebacteria. Epidemiology and Infection, 133(3), 383e391.

A. Garrido-Maestu et al. / Food Control 35 (2014) 274e283 283

Panicker, G., & Bej, A. K. (2005). Real-time PCR detection of Vibrio vulnificus inoysters: comparison of oligonucleotide primers and probes targeting vvhA.Applied and Environmental Microbiology, 71(10), 5702e5709.

Panicker, G., Call, D. R., Krug, M. J., & Bej, A. K. (2004). Detection of pathogenic Vibriospp. in shellfish by using multiplex PCR and DNA microarrays. Applied andEnvironmental Microbiology, 70(12), 7436e7444.

Raymaekers, M., Smets, R., Maes, B., & Cartuyvels, R. (2009). Checklist for optimi-zation and validation of real-time PCR assays. Journal of Clinical LaboratoryAnalysis, 23(3), 145e151.

Roig, F. J., Sanjuan, E., Llorens, A., & Amaro, C. (2010). pilF Polymorphism-based PCRto distinguish Vibrio vulnificus strains potentially dangerous to public health.Applied and Environmental Microbiology, 76(5), 1328e1333.

Rosche, T. M., Yano, Y., & Oliver, J. D. (2005). A rapid and simple PCR analysisindicates there are two subgroups of Vibrio vulnificus which correlatewith clinical or environmental isolation. Microbiology and Immunology, 49(4),381e389.

Schaerer, K., Savioz, S., Cernela, N., Saegesser, G., & Stephan, R. (2011). Occurrence ofVibrio spp. in fish and shellfish collected from the Swiss market. Journal of FoodProtection, 74(8), 1345e1347.

Swain, R. A., . J., M. K., Thiaville, P., Joseph, J., Gulig, P. A., & Wright, A. C. (2010). Theiron limitation response of Vibrio vulnificus: genetic variation, GacA regulation,and virulence. In Vibrios in the environment. Biloxi, Mississippi.

Tomas, D., Rodrigo, A., Hernandez, M., & Ferrus, M. A. (2009). Validation of real-timePCR and enzyme-linked fluorescent assay-based methods for detection of Sal-monella spp. in chicken feces samples. Food Analytical Methods, 2(3), 180e189.

Untergasser, A., Nijveen, H., Rao, X., Bisseling, T., Geurts, R., & Leunissen, J. A. M.(2007). Primer3Plus, an enhanced web interface to Primer3. Nucleic AcidsResearch, 35(Suppl. 2), W71eW74.

Warner, E., & Oliver, J. D. (2007). Refined medium for direct isolation of Vibriovulnificus from oyster tissue and seawater. Applied and Environmental Microbi-ology, 73(9), 3098e3100.

Warner, E., & Oliver, J. D. (2008). Population structures of two genotypes of Vibriovulnificus in oysters (Crassostrea virginica) and seawater. Applied and Environ-mental Microbiology, 74(1), 80e85.

Wright, A. C., Garrido, V., Debuex, G., Farrell-Evans, M., Mudbidri, A. A., &Otwell, W. S. (2007). Evaluation of postharvest-processed oysters by using PCR-based most-probable-number enumeration of Vibrio vulnificus bacteria. Appliedand Environmental Microbiology, 73(22), 7477e7481.