Available online at www.sciencedirect.com

biology 120 (2007) 179–185www.elsevier.com/locate/ijfoodmicro

International Journal of Food Micro

Evaluation of an extracting method for the detection of Hepatitis A virus inshellfish by SYBR-Green real-time RT-PCR

Nerea Casas ⁎, Félix Amarita, Iñigo Martínez de Marañón

AZTI-Tecnalia, Food Research Division, Txatxarramendi Ugartea z/g, 48395 Sukarrieta (Bizkaia), Spain

Abstract

Consumption of virus-contaminated shellfish has caused numerous outbreaks of gastroenteritis and hepatitis worldwide. In the present study,we evaluated a rapid and simple extraction method to concentrate and purify enteric viruses from shellfish tissues for their detection by real-timeRT-PCR. This procedure consists of an alkaline elution with a glycine buffer, solids removal by slow speed centrifugation, purification bychloroform extraction and virus concentration by ultracentrifugation. The efficiency of this method to recover Hepatitis A virus (HAV) fromoysters seeded with this virus, was assessed by real-time RT-PCR and conventional RT-nested PCR after extracting viral RNA by a commercialisolation kit. Real-time RT-PCR yielded higher detection sensitivity than the obtained by conventional RT-nested PCR. Besides the improvementsin detection sensitivity, the real-time RT-PCR, by quantifying HAV RNA, allowed to check the overall extraction procedure and the recoveryefficiency after each processing step. After the last phase, i.e. virus concentration by ultracentrifugation, the RNA purity was high but theestimated HAV recovery efficiency was however low, probably due to virus losses and the presence of RT-PCR inhibitors in sample concentrates.In contrast, the HAV recovery percentage was higher after the virus elution step while the RNA purity was lower. Real-time RT-PCR detectioncould allow to eliminate some purification and concentration steps that are required for conventional RT-nested PCR detection. The overallprocedure for detecting HAV could be then simplify avoiding virus losses during manipulation.© 2007 Elsevier B.V. All rights reserved.

Keywords: HAV; Oyster; Extracting method; Detection sensitivity; Quantification; RT-PCR

1. Introduction

Bivalve molluscan shellfish are filter feeders that readilybioconcentrate human pathogens, such as enteric viruses, fromfaecally contaminated growing waters. Therefore, shellfishexposed to faecal contamination may cause outbreaks of hepatitisor gastroenteritis if they are eaten raw or insufficiently cooked(Richards, 1987; Leoni et al., 1998; Koopmans et al., 2002).

Current EU standards for the evaluation of the sanitary qualityof shellfish rely entirely on bacterial indicators of faecalcontamination (Anonymous, 1991). It has been well documentedthat such indicators are not correlated with the presence of viralpathogens (Doré et al., 1998; Pina et al., 1998; Le Guyader et al.,2000; Croci et al., 2000; Muniain-Mujika et al., 2003). Therefore,Council Regulation (EC, No. 2073/2005) recently proposedthat sanitary controls of shellfish should include viral parametersto guarantee safety for human consumption. This document

⁎ Corresponding author. Tel.: +34 94 6029400; fax: +34 94 6870006.E-mail address: ncasas@suk.azti.es (N. Casas).

0168-1605/$ - see front matter © 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.ijfoodmicro.2007.01.017

specifies that standardised methods should be developed beforethe establishment of virological criteria. However, the methodscurrently available are too expensive and too time-consuming forroutine screening of foods (Loopman et al., 2002; Koopmans andDuizer, 2004). Therefore, there is a clear need for a sensitive,reliable and rapid method for the detection and quantification ofenteric viruses in shellfish.

Molecular techniques such as reverse transcription-polymer-ase chain reaction (RT-PCR) have been successfully utilized forthe detection of Hepatitis A virus (HAV) in shellfish (Pina et al.,1998; LeGuyader et al., 2000; Casas and Suñen, 2001). Real-timeRT-PCR has been recently adopted for the quantification of HAV,Noroviruses (NoV) and other enteric viruses in foods (Shan et al.,2005; Jothikumar et al., 2005; Costafreda et al., 2006; Duboiset al., 2006). The application of both techniques for virusdetection in shellfish is hindered by two main disadvantages: thepresence of PCR inhibitory substances in shellfish and the lowconcentration of virus recovered. For this reason, viral extractionfrom shellfish tissues is a critical step for the detection of entericviruses by molecular methods.

180 N. Casas et al. / International Journal of Food Microbiology 120 (2007) 179–185

The aim of this work was therefore to evaluate the efficiencyof an extracting method for the concentration and purification ofHAV from shellfish and for viral detection through real-timeRT-PCR. We also compared detection sensitivities betweenconventional RT-nested PCR and real-time RT-PCR for theoverall procedure. Moreover, HAV was quantified after eachprocessing step to point out the minimal steps required to detectHAV RNA by real-time RT-PCR.

2. Materials and methods

2.1. Viruses

The HM-175 strain of HAV was used throughout this studyand it was supplied by Dr. Peter Müller from the Canadian FoodInspection Agency.

2.2. Seeding assays in shellfish samples

Serial ten-fold dilutions of HAV, ranging in concentration from2.5×106 to 25 Tissue Culture Infectious Dose50 (TCID50) per mlwere prepared and aliquots of 100 μL were inoculated onto theshellfish homogenates in the F1 processing step. The finalconcentration of HAV ranged from 8.3×104 to 0.85 TCID50 inoyster concentrates (equivalent to 8 g of oyster samples).

2.3. Shellfish processing for recovering of viruses

Oyster samples were collected from different sites along theBasque Country coast between October 2004 and June 2005.Oyster samples negative for HAVand NoV detection were usedfor seeding experiments. Un-inoculated oysters were utilized asnegative controls.

The extracting method was based on the one described byPina et al. (1998) with some modifications. Twenty-five gramsof the whole shellfish tissues was homogenized in glycinebuffer (0.25 N, pH 10) (F1) and virus were eluted by incubationfor 15 min at room temperature (F2). Solids were pelleted bycentrifugation at 2500 ×g for 20 min at 4 °C (F3) and samplewas clarified at 40,000 ×g for 45 min at 4 °C (F4), and thesupernatant was adjusted to pH 7. Viruses were further purifiedby extraction with chloroform (vol/vol) and centrifugation at

Table 1Primers used in the conventional RT-nested PCR and SYBR-Green real-time RT-PC

Name Orientation Location a

Conventional RT-PCRHAV-1 + 332–352HAV-2 – 680–700

Inner primersHAV-3 + 371–391HAV-4 – 641–661

Real-time RT-PCRHAV-1Q + 396–419HAV-2Q – 463–483a Nucleotide positions based on the HAV Genotype IB HM-175 strain (accession

1800 ×g for 30 min at 4 °C (F5). The chloroform and interfacefractions were re-extracted with a glycine solution (0.05 N–0.15 N NaCl, pH 7.5) (v/v) (F6). After centrifugation (at 1800 ×gfor 30 min at 4 °C), the supernatant was combined with thesupernatant of the previous step and concentrated by ultracentri-fugation at 100,000 ×g for 1 h at 4 °C (F7). The final pellet wasresuspended with 200 μL of lysis buffer supplied in the TotalQuick RNA isolation kit (Talent™, Italy) by vigorous vortexmixing and pipeting.

2.4. Viral RNA extraction from shellfish concentrates

Total RNA was extracted using Total Quick RNA isolationkit (Talent™, Italy) following the manufacturer's instructions.Extracted viral RNA was suspended in 25 μL of RNase-freeH2O.

2.5. Determination of RNA purity

A volume of 7 μL of the extracted viral RNA was analysedby the spectrophotometer Gene Quant Pro (AmershamBiosciences, USA). RNA purity was determined by calculatingthe ratio between the absorbance measured at 260 nm and themeasured at 280 nm.

2.6. Primers

The highly conserved 5′-end untranslated (5′NTR) region ofHAV was the target for the 368-pb cDNA and 290-pb for thenested PCR (Pina et al., 1998). For real-time RT-PCR detection,the selected region was also in the 5′NTR region and theprimers were designed using the Primer Express v. 2.0. software(Applied Biosystems, USA) following the information provid-ed. The sequences of the oligonucleotide primers used in theconventional and in the real-time RT-PCR are listed in Table 1.

All primers have been tested for primer–dimer formationusing the Pubmed NCBI Blast software.

2.7. Reverse transcription (RT)-PCR and nested PCR

A final volume of 20 μL of RT reaction consisted of 2 mMMgCl2, 0.2 mM dNTP mix, 10× PCR buffer (500 mM KCl,

R

Sequence (5′–3′) Ref.

TTGGAACGTCACCTTGCAGTG Pina et al. (1998)CTGAGTACCTCAGAGGCAAAC Pina et al. (1998)

ATCTCTTTGATCTTCCACAAG Pina et al. (1998)GAACAGTCCAGCTGTCAATGG Pina et al. (1998)

AGGCTACGGGTGAAACCTCTTAG This workGCCGCTGTTACCCTATCCAA This work

no. M14707) sequence.

Table 2Detection of HAV RNA in oyster samples, initially seeded at different inputs, byconventional and real-time RT-PCR

Input a

(TCID50/g)No. of positives/No. of samplesanalysed

Ctvalues b

Quantified RNAc

(TCID50/PCR reaction)

RT-nestedPCR

Real-timeRT-PCR

104 3/3 3/3 32.15 4.2×10103 3/3 3/3 33.83 1.4×10102 3/3 3/3 34.54 9.3101 3/3 3/3 35.07 6.65 2/3 3/3 35.64 4.61 0/3 1/3 36.12 3.40.1 ND ND ND –

ND: Not detected.a Twenty-five grams of oyster tissueswas seeded and processed by the described

method.Virus input is based on the titre of HAV stock seeded in F1 processing step.b Ct values correspond to HAV RNA detected.c HAV RNA of initially seeded sample was extracted from F7 processing step

and quantified by real-time RT-PCR by means of a standard curve (see Fig. 2).

181N. Casas et al. / International Journal of Food Microbiology 120 (2007) 179–185

100 mM Tris–HCl, pH 8.3), 2 μM Random Hexamer, 1 μLRNase inhibitor (20 U/μL) and 1.25 μL MuLV Reversetranscriptase (50 U/μL) (Applied Biosystems, USA) and 5 μLof RNA (equivalent to 0.4 μg) was incubated at 42 °C for45 min and then the reverse transcriptase was inactivated at95 °C for 5 min.

PCR amplification was performed by combining 2 mMMgCl2, 1 μM for each primer, 0.5 μL Amplitaq DNApolymerase (5 U/μL) (Applied Biosystems, USA) and H2O toa final volume of 100 μL. The PCR reaction was carried out byforty thermal cycles of annealing at 95 °C for 1 min, extensionfor 1.5 min at 55 °C, denaturation 1 min at 72 °C and a finalextension for 7 min at 72 °C.

For the second amplification (nested PCR), 2 μL of theamplified DNA from the first PCR reaction was added to anew batch of 50μLof PCR reactionmix containing 2mMMgCl2,1× PCR buffer (500 mMKCl, 100 mM Tris–HCl, pH 8.3), 1 μMfor each inner primers and 0.25 μL Amplitaq DNA polymerase(5 U/μL) (Applied Biosystems, USA). The amplification cycleswere as described previously. The amplified PCR products wereseparated by electrophoresis in a 2%agarose gel and visualized byethidium bromide staining.

2.8. SYBR-Green real-time RT-PCR

HAV RT reaction was performed at 48 °C for 30 min on theABI Prism 7000 sequence detection system (Applied Biosys-tems, USA) consisting of 2 mM MgCl2, 0.2 mM dNTP mix,10× PCR buffer (500 mM KCl, 100 mM Tris–HCl, pH 8.3),2 μM Random Hexamer, 1 μL RNase inhibitor (20 U/μL) and1.25 μL MuLV Reverse transcriptase (50 U/μL) (AppliedBiosystems, USA) and 2 μL of RNA (equivalent to 0.2 μg).

Real-time amplification was performed using a total reactionvolume of 25 μL. Real-time reactions were carried out with12.5 μL SYBR-Green PCR Master Mix (Applied Biosystems,USA), 300 nM of each specific primers described in Table 1 and2 μL of the cDNA. Reactions were run on the ABI Prism 7000sequence detection system (Applied Biosystems, USA) with thefollowing thermal conditions: 50 °C for 2 min, 95 °C for 10 minfollowed by 40 cycles of 95 °C for 15 s and 60 °C for 1 min.Readings were taken every cycle, and the logarithm of theincrement in fluorescence was plotted versus the cycle number.The threshold level was fixed at the same midexponentialposition for all runs.

For data analysis, the melting temperature, fluorescence(−dF / dT) under the melting curves window, and Ct, which isdefined as the fractional cycle number where the fluorescencepassed the fixed threshold, were selected as the evaluatingparameters in the quantification window. Each Ct value wasobtained by the means of three replicates with a standarddeviation of b0.2.

For the evaluation of the specificity of the SYBR-Green real-time RT-PCR, a melting temperature analysis was carried out bymeans of the study of the dissociation curves obtained from thePCR products which were monitored in the same instrument,and the derivatives (−dF / dT) of fluorescence values wereplotted at 0.3 °C intervals from 60 °C to 95 °C.

2.9. Quantification with the real-time RT-PCR assay

To evaluate the quantitative SYBR-Green real-time PCRassay for HAV, serial 10-fold dilutions of the HM-175 HAVstrain were amplified to generate a standard curve.

2.10. PCR efficiency of the SYBR-Green real-time RT-PCR

The PCR efficiency (E) was calculated from the slope of thestandard curve described above and using the followingequation: E=101/− s where E = PCR efficiency and S = slope.

3. Results

3.1. Sensitivity of detection of HAV RNA by conventionalRT-nested PCR and real-time RT-PCR for seeded oyster samples

Known concentrations of HAV stock were experimentallyspiked into oyster homogenates, processed and analysed byconventional RT-nested PCR or real-time RT-PCR for the deter-mination of the overall detection sensitivity.

Table 2 shows that real-time RT-PCR assay, could detectHAV RNA up to 5 TCID50/g in all seeded oyster samples.However, HAV RNA was detected in two-third of the samplesby RT-nested PCR. This difference in sensitivity was alsopointed out for samples inoculated with 1 TCID50/g: whenusing real-time RT-PCR, HAV RNA was detected in one-thirdof the samples, whereas no HAV RNA was detected byconventional RT-nested PCR. The difference in virus detectionsensitivity between conventional RT-nested PCR and real-timeRT-PCR was about 1 Log in this complex matrix (oyster homo-genates) and was therefore similar to the one pointed out fordistilled water (see above).

Concerning the evaluation of the specificity of the real-timeRT-PCR for the detection of HAVRNA in seeded oyster samples,an analysis of themelting temperaturewas performed bymeans of

Fig. 1. Melting temperature analysis of standard HAV cDNA fragment amplifiedby SYBR-Green real-time PCR detection, in serial dilutions ranging from3.2×102 to 3.2 TCID50. The average Tm value (°C) is indicated at the peakposition.

Fig. 3. Standard curve for quantification of HAV using real-time RT-PCR:standard DNA fragment amplified from HM-175 strain, in serial dilutionsranging from 3.2×105 to 3.2 TCID50.

182 N. Casas et al. / International Journal of Food Microbiology 120 (2007) 179–185

the dissociation curve study obtained from the real-time PCRproduct. The melting curve profiles obtained from the real-timePCR amplification of HAV cDNAare shown in Figs. 1 and 2. Themajor peak for the HAV amplicon was observed at 75.1 °C±0.2 °C. As it is shown in Fig. 1, the major peak was constant atdifferent concentrations of HAV cDNA and the size of this majorpeak correlated well with the different amounts of HAV cDNAfragment amplified by SYBR-Green real-time PCR detection.Moreover, the melting curve analysis revealed that the positionand specificity of the major peak was not affected by the presenceof non-target DNAwhen un-inoculated oysters were analysed and103 PCR units (PCRU) per gram of Norovirus Genogroup I (NoVGI) were seeded in HAV-inoculated oyster samples as shown inFig. 3.

Fig. 2. Melting temperature analysis of the SYBR-Green real-time RT-PCRproduct in oyster samples seeded with 2.5×104 TCID50 of HAV. The wider lineshows the melting profile of HAV DNA in oysters seeded with HAV(2.5×104 TCID50) and 103 PCRU per gram of NoV GI. The dashed lineshows the melting profile of the un-inoculated oyster samples used as negativecontrols. The average Tm value (°C) is indicated at the peak position.

Afterwards, we evaluated the real-time PCR HAV quantifi-cation assay by means of the generation of a standard curve withserial 10-fold dilutions of the HM-175 HAV strain. A linearregression from 3.2×105 to 3.2 TCID50 was fitted (R2 =0.996)(Fig. 3). The slope (−3.63) of this standard curve was similar tothe one (−3.41) estimated by Shan et al. (2005). The detectionlimit of the assay was 10 copies of HAV genome per PCRreaction although sometimes b10 copies were detected aspreviously reported (Shan et al., 2005). A PCR efficiency of100% is ideally achieved when the slopes are close to thetheoretical value of −3.32. According to the slope obtainedfrom the standard curve shown in Fig. 1, the calculated PCRefficiency was 88.5%.

Table 2 shows that recovered HAV RNA, quantified by real-time RT-PCR by means of the standard curve (Fig. 3), was notwell related to the initial inocula: the slopeLog quantified/Log input

was 0.199 (R2 =0.978). HAV inputs extended from 104 to5 TCID50/g, i.e. 4 Log, whereas the range for recovered HAVRNA was 1 Log (from 42 to 4.6 TCID50/PCR reaction).

This lack of relation in quantification could be due to theinterference of remaining inhibitors of RT-PCR reaction fromsample concentrates. To verify that these inhibitors limited theHAVdetection, oyster samples initially seeded with 104 TCID50/g of HAV were processed and serially 10-fold diluted after RNAextraction. Serial dilutions were then analysed by real-time RT-PCR.When RNA extracted from sample concentrate was 10-folddiluted (from 104 to 103 TCID50/PCR reaction), the Ct valueshould be higher than the one estimated for the undiluted sample,i.e. the recoveredRNAcontent should be lower. However, Table 3shows just the opposite, i.e. the Ct value is lower for 10-folddiluted sample than for the undiluted one. Those Ct values wouldindicate the HAV RNA content of the undiluted sample(22 TCID50/PCR reaction) would be lower than the 10-folddiluted sample one (54 TCID50/PCR reaction). This discordancewould confirm the presence of inhibitors of the RT-PCR reaction.For additional serial dilutions, a good enough relationship(slopeLog quantified/Log input=0.709; R

2=0.992) was pointed outbetween RNA inputs (from 103 to 1 TCID50/PCR reaction) andRNA quantified. For a 3 Log increase in RNA inputs (from thelower RNA input), a 2 Log raise is detected after the PCR reaction(Table 3) while a 0.62 Log raise was previously detected (Table 2)for identical viral inputs. These results also confirm the presenceof RT-PCR inhibitors in oyster extracts.

Table 5Effect of inhibition in HAV RNA analysis by real-time RT-PCR

Processingsteps

RNA purity(A260/A280) a

Ct values b HAV (TCID50/PCR reaction c)

Mean SD Mean SD Seeded Quantified d %Recovery

F1 e ND – 30.59 0.16 1.5×103 7×10 4.6F2 1.04 0.06 34.97 0.11 1.5×103 1.1×103 73.3F3 1.15 0.08 33.30 0.09 1.5×103 2.0×102 13.3F4 ND – 35.22 0.08 1.5×103 6.0×10 4.0F5 1.29 0.11 33.19 0.07 1.5×103 2.2×102 14.6F6 ND – ND – – ND NDF7 f 1.54 0.07 31.26 0.12 7.5×102 7.4×10 9.9

ND: Not detected.a Average values for RNA purity of two replicate samples.b Ct values were obtained per triplicate.c 0.5 μL-volume sample was analysed per PCR reaction.d HAV RNA was extracted from 100 μL of initially seeded sample and

quantified by real-time RT-PCR by means of a standard curve (see Fig. 2).e 7.5×104 TCID50 of HAV stock initially seeded in 10 g of oyster sample. An

aliquot was extracted corresponding to an inoculum of 1.5×103 TCID50 perPCR reaction.f 7.5×104 TCID50 of HAV stock initially seeded in 10 g of oyster sample. An

aliquot was extracted corresponding to an inoculum of 7.5×102 TCID50 perPCR reaction.

Table 3Quantification ofHAVRNAafter serial dilution of initial seeded oysters by real-timeRT-PCR

Input (TCID50/PCR reaction) a

Ct values b Quantified RNAc

(TCID50/PCR reaction)

104 33.11 2.2×10103 31.76 5.4×10102 34.87 7.5101 37.45 1.51 39.51 0.4a Twenty-five grams of oyster tissues was seeded with 104 TCID50/PCR

reaction, processed, 10-fold diluted and analysed by real-time RT-PCR.b Ct values correspond to HAV RNA detected.c HAV RNA of initially seeded sample and 10-fold diluted was quantified by

real-time RT-PCR by means of a standard curve (see Fig. 2).

183N. Casas et al. / International Journal of Food Microbiology 120 (2007) 179–185

3.2. Recovery of HAV seeded onto shellfish

Firstly, the RNA purity was determined in PBS after the viralRNA extraction, obtaining a reference value of 1.82 (data notshown). Afterwards, the RNA purity of the seeded sample wasestimated by extracting viral RNA after each processing step(Table 4). RNA extracted from the elution (F2) and the con-centration (F7) steps showed the lowest purity. An increase ofthe RNA purity value was achieved after the centrifugation (F3)and chloroform extraction (F5) steps. However, since somemolecules could hinder the RNA purity value determination,samples obtained after each processing step were 10-fold diluted(Table 5). The determination of RNA purity showed a noticeableincrease for elution (F2) and concentration (F7) steps and, inminor degree, for solid separation (F3). For the chloroformextraction step (F5) the RNA purity showed similar value thanobtained from undiluted seeded sample.

Table 4Quantification of HAV RNA in oyster samples from each processing step

Processingsteps

RNA purity(A260/A280) a

Ct values b HAV (TCID50/PCR reaction c)

Mean SD Mean SD Seeded Quantified d Recovery %e

F1 f ND – 29.48 0.13 1.5×103 3.2×101 2.2F2 0.65 0.05 33.04 0.09 1.5×103 1.9×102 13.3F3 1.03 0.08 31.06 0.14 1.5×103 7.9×101 5.5F4 ND – 33.03 0.16 1.5×103 2.5×101 1.6F5 1.25 0.11 31.61 0.07 1.5×103 6.3×101 3.9F6 ND – ND – – ND NDF7 g 0.76 0.08 33.49 0.18 7.5×102 1.8×101 2.4

ND: Not detected.a Average values for RNA purity of two replicate samples.b Ct values were obtained per triplicate.c 5 μL-volume sample was analysed per PCR reaction.d HAV RNA was extracted from 100 μL of initially seeded sample and

quantified by real-time RT-PCR by means of a standard curve (see Fig. 2).e Recovery % was determined by dividing the seeded HAV by the recovered

HAV (TCID50/PCR reaction).f 7.5×104 TCID50 of HAV stock initially seeded in 10 g of oyster sample. An

aliquot was extracted corresponding to an inoculum of 1.5×103 TCID50 perPCR reaction (see Materials and methods).g 7.5×104 TCID50 of HAV stock initially seeded in 10 g of oyster sample. An

aliquot was extracted corresponding to an inoculum of 7.5×102 TCID50 perPCR reaction (see Materials and methods).

Concerning HAV recovery efficiency (Table 4), calculatedfrom HAV RNA quantified by real-time RT-PCR, elution step(F2) achieved the highest recovery yield (13.3%), followed bycentrifugation (F3), chloroform extraction (F5), homogeniza-tion (F1) and clarification (F4) processing steps. The last phaseof the overall procedure, i.e. the concentration step (F7),reported low recovery efficiency (2.4%) probably due to thepresence of RT-PCR inhibitors in the sample concentrates.Samples obtained after each processing step were 10-folddiluted (Table 5) to point out the inhibition effect on RT-PCR.Recovery yields showed a minimum of 2-fold increase in allprocessing steps when samples were 10-fold diluted. Thehighest efficiency in HAV recovery was observed in the elutionstep (F2: 73.3%) where a ∼5-fold increase was reported afterdilution of seeded sample. The concentration step (F7) achieveda 4-fold increase with respect to undiluted sample concentrates,reaching a recovery yield of ∼10%. The low recovery reportedat this phase (i.e. after the overall extraction procedure) wouldindicate that viruses were lost during the previous processingsteps. The re-extraction step (Tables 4 and 5: F6) did not yieldany virus detectable by real-time RT-PCR. Since this stepincreases the final volume of shellfish extract, this phase wasfinally excluded from the overall procedure.

4. Discussion

To date, there is an absence of a standardised methodologyfor extracting enteric viruses from shellfish tissues due toseveral factors that hamper their detection by molecularmethods. The elimination of inhibitors which hinder the reversetranscription and amplification reaction of PCR and the lowconcentration levels of viruses in food are the main problems tobe solved (Le Guyader et al., 1996; Traore et al., 1998; Kingsley

184 N. Casas et al. / International Journal of Food Microbiology 120 (2007) 179–185

and Richards, 2001; Casas and Suñen, 2001; Shan et al., 2005;Rzezutka et al., 2006).

This study was firstly focused on the evaluation of theefficiency of an extracting method to purify and concentrateHAV from shellfish tissues for detection by real-time RT-PCR.The overall procedure involves simplicity of manipulations,absence of risk linked with cross-contamination and the use ofcommercially available agents and common laboratory chemi-cals. Detection limits obtained after seeding experimentsshowed that real-time RT-PCR yielded higher detection sen-sitivity than the obtained by conventional RT-nested PCR. Inaddition, real-time RT-PCR is more rapid, simple and reducesrisks for carry-over contamination from the second amplifica-tion required in RT-nested PCR. The total time required toperform the viral extraction and the real-time RT-PCR detectionwas approximately 6–7 h.

We compared the sensitivity of detection of HAV RNAwhenoyster samples processed by the extracting method wereanalysed by conventional RT-nested PCR and real-time RT-PCR. The HAV detection limit was initially determined usingreal-time RT-PCR or RT-nested PCR in distilled water byperforming assays with 10-fold serial dilutions of HAV stock.The detection limit for real-time RT-PCR assay was almost 10-fold lower than obtained by RT-nested PCR (data not shown).

Since the real-time RT-PCR allowed to quantify HAV RNA,some discordances between initial inoculum and quantified HAVwere reported by seeding oyster samples at different levels ofHAV. This lack of relationship could be due to the fact that ourstudy included seeding experiments at low levels of HAVand thematrix analysed (shellfish) was complex and presented inhibitorsof RT-PCR reaction. By diluting the HAV RNA obtained fromsample concentrate, the relationship was improved and was closeto the one found by Shan et al. (2005) who reported proportionalvalues when determining HAV in green onion and strawberryrinses. Nevertheless, the differences between the HAV RNAinputs and the detected ones (for diluted HAV RNA samples)were weaker (0.4 to 1.3 Log) than the ones (1.4 to 1.7 Log)reported by Shan et al. (2005) in PBS and produce rinses withoutimmunomagnetic capture.

Regarding recovery efficiency of the extracting method(overall procedure), HAV recovery yields achieved in this study(recovery percentage: 10%) differed from those reported inpreviously published methods. Casas and Suñen (2001) reporteda recovery percentage b2% measured by using a cell cultureinfectivity assay. This low recovery is possibly due to the fact thatthe procedure includes numerous steps and consequently, highlosses of viruses could occur during manipulation. In that study,HAV was extracted from mussel tissues by alkaline elution,concentrated by polyethylene glycol (PEG), purified by a chloro-form extraction and re-concentrated by a second PEG precipita-tion. Otherwise, Rutjes et al. (2006) used an ultracentrifugationmethod to concentrate Canine calicivirus (CaCV) from lettuce,reporting high recovery percentages (90%). Even though anultracentrifugation step was used in our work, the low recoveryefficiency could be explained by the complexity of the shellfishmatrix with regard to the vegetable one. In fact, the presenceof remaining inhibitors in shellfish concentrates affected the

recovery efficiency value and a 10-fold sample dilution allowedan increase of quantified HAV for the overall procedure.

So far, viral extraction methodologies involved the inclusionof purification and concentration steps to avoid false negativeresults on detection of enteric viruses by conventional RT-PCRin shellfish (Pina et al., 1998; Mullendore et al., 2001; Suñenet al., 2004). However, most developed procedures are time-consuming and involve multiple steps, resulting in virus lossesduring manipulation and a reduction of the detection sensitivity.

In this study, each processing step was analysed by real-timeRT-PCR to point out the minimal steps required to efficientlydetectHAVRNA. It was shown that recovery efficiency after viralelution was very high (73%) and was similar to the one reportedby Shieh et al. (1999), by performing infectivity assays, to recoverpoliovirus from elution step in oysters. Moreover, Dubois et al.(2006) achieved different values of HAV recovery yields after theelution step in butter lettuce: recoveries ranged from12% to 182%depending on the buffer solutions used.

The determination of RNA purity after each processing steprevealed that inclusion of purification phases, as chloroformextraction, significantly increased the quality of the final sampleconcentrate. However, RNA extracts from the elution step alsopresented enough purity to be analysed by real-time RT-PCR,which allowed us to report that the highest efficiency in HAVrecoverywas achieved after viral elution.After this elution step, inparticular during the phase separating solid particles, a lot of virusparticles were lost. Furthermore, the application of real-time RT-PCR allowed the HAV detection and quantification in allundiluted seeded shellfish concentrates in spite of the presenceof RT-PCR inhibitors. Those advantages then make possible theapproach of a more simple viral extraction method, mainly basedon a unique alkaline elution step, for HAV detection by real-timeRT-PCR. However, if RNA extracts are utilized for sequencing orassays that require more purified nucleic acids, the overallextraction procedure is needed.

Future studies will focus on the application of the describedprocedure for the detection and quantification of low levels ofHAVRNA seeded in shellfish and the evaluation of the efficiencyof this procedure for the analysis of other enteric viruses. Furtherapproaches should be also developed to evaluate the minimalprocessing steps required to detect enteric viruses in other foodmatrices, such as vegetables.

Acknowledgements

This work was supported by funds from the Department ofAgriculture, Fisheries and Food from the Basque Government.We sincerely thank Peter Müller from the Canadian FoodInspection Agency for providing the HAV strain HM-175.

References

Anonymous, 1991. Council Directive 91/492/EEC. Laying down the healthconditions for the production and placing on the market of live bivalvemolluscs (91/492/EEC). Official Journal of EuropeanCommunities 268, 1–14.

Casas, N., Suñen, E., 2001. Detection of enterovirus and hepatitis A virus RNAin mussels (Mytilus spp.) by reverse transcriptase-polymerase chain reaction.Journal of Applied Microbiology 90, 89–95.

185N. Casas et al. / International Journal of Food Microbiology 120 (2007) 179–185

Costafreda, M.I., Bosch, A., Pinto, R.M., 2006. Development, evaluation, andstandardization of a real-time TaqMan reverse transcription-PCR assay forquantification of hepatitis A virus in clinical and shellfish samples. Appliedand Environmental Microbiology 72, 3846–3855.

Croci, L., De Medici, D., Scalfaro, C., Fiore, A., Divizia, M., Donia, D.,Cosentino, A.M., Moretti, P., Costantini, G., 2000. Determination ofenteroviruses, hepatitis A virus, bacteriophages and Escherichia coli inAdriatic Sea mussels. Journal of Applied Microbiology 88, 293–298.

Doré, W.J., Henshilwood, K., Lees, D.N., 1998. The development of manage-ment strategies for control of virological quality in oysters. Water Scienceand Technology 38, 29–35.

Dubois, E., Hennechart, C., Deboose're, N., Merle, G., Legeay, O., Burger, C.,Le Calve, M., Lombard, B., Ferre, V., Traore, O., 2006. Intra-laboratoryvalidation of a concentration method adapted for the enumeration ofinfectious F-specific RNA coliphage, enterovirus, and hepatitis Avirus frominoculated leaves of salad vegetables. International Journal of FoodMicrobiology 108, 164–171.

Jothikumar, N., Lowther, J.A., Henshilwood, W., Lees, D.N., Hill, V.R., Vinje,J., 2005. Rapid and sensitive detection of noroviruses by using TaqMan-based one-step reverse transcription-PCR assays and application to naturallycontaminated shellfish samples. Applied and Environmental Microbiology71, 1870–1875.

Kingsley, D.H., Richards, G.P., 2001. Rapid and efficient extraction method forreverse transcription-PCR detection of hepatitis A and Norwalk-like virusesin shellfish. Applied and Environmental Microbiology 67, 4152–4157.

Koopmans, M., Duizer, E., 2004. Foodborne viruses: an emerging problem.International Journal of Food Microbiology 90, 23–41.

Koopmans, M., Von Bondsdorff, C.H., Vinjé, J., De Medici, D., Monroe, S.,2002. Foodborne viruses. FEMS Microbiology Reviews 26, 187–205.

Le Guyader, F., Neill, F.H., Estes, M.K., Monroe, S.S., Ando, T., Atmar, R.L.,1996. Detection and analysis of a small round-structured virus strain inoysters implicated in an outbreak of acute gastroenteritis. Applied andEnvironmental Microbiology 62, 4268–4272.

Le Guyader, F., Haugarreau, L., Miossec, L., Dubois, E., Pommepuy, M., 2000.Three year study to assess human enteric viruses in shellfish. Applied andEnvironmental Microbiology 66, 4152–4157.

Leoni, E., Bevini, C., Degli Esposti, S., Graziano, A., 1998. An outbreak of intra-familiar hepatitis A associated with clam consumption: epidemic transmissionto a school community. European Journal of Epidemiology 14, 187–192.

Loopman, B., Van Duynhoven, Y., Hanon, F.X., Reacher, M., Koopmans, M.,Brown, D., 2002. Laboratory capability in Europe in foodborne viruses.Eurosurveillance 7, 61–65.

Mullendore, J.L., Sobsey, M.D., Shieh, Y.C., 2001. Improved method for therecovery of hepatitis A virus from oysters. Journal of Virological Methods94, 25–35.

Muniain-Mujika, I., Calvo, M., Lucena, F., Girones, R., 2003. Comparativeanalysis of viral pathogens and potential indicators in shellfish. InternationalJournal of Food Microbiology 83, 75–85.

Pina, S., Puig, M., Lucena, F., Jofre, J., Girones, R., 1998. Viral pollution in theenvironment and in shellfish: human adenovirus detection by PCR as an indexof human viruses. Applied and Environmental Microbiology 64, 3376–3382.

Richards, G.P., 1987. Shellfish associated enteric virus illness in the UnitedStates, 1934–1984. Estuaries 10, 84–85.

Rutjes, S.A., Lodder-Verschoor, F.M., Van Der Poel, W.H.M., Van Duijnhoven,Y.T.H.P., De Roda Husman, A.M., 2006. Detection of Noroviruses in foods:a study on virus extraction procedures in foods implicated in outbreaks ofhuman gastroenteritis. Journal of Food Protection 69, 1949–1956.

Rzezutka, A., D'Agostino, M., Cook, N., 2006. An ultracentrifugation-basedapproach to the detection of hepatitis A virus in soft fruits. InternationalJournal of Food Microbiology 108, 315–320.

Shan, X.C., Wolffs, P., Griffiths, M.W., 2005. Rapid and quantitative detectionof hepatitis A virus from green onion and strawberry rinses by use of real-time reverse-transcription-PCR. Applied and Environmental Microbiology71, 5624–5626.

Shieh, Y.S.C., Calci, K.R., Baric, R., 1999. A method to detect low levels ofenteric viruses in contaminated oysters. Applied and EnvironmentalMicrobiology 65, 4709–4714.

Suñen, E., Casas,N.,Moreno, B., Zigorraga, C., 2004. Comparison of twomethodsfor the detection of hepatitis A virus in clams samples (Tapes spp.) by reversetranscription-nested PCR. International Journal of Food Microbiology 91,147–154.

Traore, O., Arnal, C., Mignotte, B., Maul, A., Laveran, H., Billaudel, S.,Schwartzbrod, L., 1998. Reverse transcriptase PCR detection of astrovirus,hepatitis A virus, and poliovirus in experimentally contaminated mussels:comparison of several extraction and concentration methods. Applied andEnvironmental Microbiology 64, 3118–3122.