Virus-Contaminated Oysters: a Three-Month Monitoring of Oysters Imported to Switzerland

Abstract

Molluscan shellfish are known to be carriers of viral and bacterial pathogens. The consumption of raw oysters has been repeatedly linked to outbreaks of viral gastroenteritis and hepatitis A. Switzerland imports over 300 tons of oysters per year, 95% of which originate in France. To assess the level of viral contamination, a 3-month monitoring study was conducted. Therefore, the sensitivities of several previously described methods for virus concentration were compared, and one protocol was finally chosen by using dissected digestive tissues. Eighty-seven samples consisting of five oysters each were analyzed for Norwalk-like viruses (NLVs), enteroviruses, and hepatitis A viruses from November 2001 to February 2002. The oysters were exported by 31 French, three Dutch, and two Irish suppliers. Eight oyster samples from six French suppliers were positive for NLVs, and four samples from four French suppliers were positive for enteroviruses; two of the latter samples were positive for both viral agents. No hepatitis A viruses were detected. The sequences of NLV and enterovirus amplicons showed a great variety of strains, especially for the NLVs (strains similar to Bristol, Hawaii, Mexico, and Melksham agent). The data obtained indicated that imported oysters might be a source of NLV infection in Switzerland. However, further studies are needed to determine the quantitative significance of the risk factor within the overall epidemiology of NLVs.

An association of infectious diseases with shellfish consumption was first documented with cases of typhoid fever in the United States in 1894 (15). Since then, the etiologic agents could not be identified in about 50% of over 15,000 documented cases of infectious diseases caused by shellfish because of a lack of appropriate detection methods. Before the PCR was developed in the 1980s (32), the most frequently identified etiologic agents for shellfish-borne diseases were Salmonella enterica serovar Typhi and hepatitis A virus (HAV). The first cloning and sequencing of the Norwalk virus in 1990 (20) and the subsequent development of diagnostic tools (reverse transcription [RT]-PCR) revealed the importance of the noncultivable Norwalk-like viruses (NLVs) (28) and enteroviruses (EVs) (28) as viral contaminants in shellfish, including oysters, and thus considerably reduced the number of idiopathic cases whose etiology was unknown. Accumulation of different pathogens in shellfish occurs because each of these filter-feeding animals filters several liters of seawater daily (1). Thus, the microbial contamination levels in oyster tissues are considerably greater than those in the overlying water (29, 36). This bioaccumulation of viruses during feeding is probably assisted by the ionic bonding of viral particles to the mucopolysaccharide moiety of shellfish mucus (13). As the level of mucus production generally corresponds to the glycogen content of the connective tissue (16), which is highest in oysters from November through March (7), there could be a connection to the fact that viral contamination is mainly observed during the winter months (28).

Concentration of viruses from oysters and removal of inhibitors by purifying and isolating the viral nucleic acid are essential for a sensitive and reproducible PCR (7). We compared different proposed methods for isolation of viruses from oysters based on whole shellfish or dissected tissue (3, 12, 18, 19, 21, 25, 31), and we adapted a method that was originally used for whole shellfish (31) to a dissected tissue using digestive tissues.

In the present study, the adapted method was used to assess the viral contamination of oysters exported from France, The Netherlands, and Ireland to Switzerland. The screening program took the main pathogenic enteric viruses (namely, NLVs, EVs, and HAV) into consideration (24).

MATERIALS AND METHODS

Oyster sampling.

Oysters (Crassostrea gigas and Ostrea edulis, also known as the European flat oyster) were collected weekly from November 2001 to February 2002 by the Swiss Federal Veterinary Office at three border inspection posts (Muttenz, Bardonnex, and Zurich Airport) and sent on ice to the Cantonal Food Laboratory of Solothurn. A pool consisting of five oysters from a producer was defined as a sample. Altogether, samples from 31 French, three Dutch, and two Irish suppliers were collected, stored at 4°C, and analyzed within 2 days.

Positive control samples.

The following virus strains were used for experimental inoculation of oyster tissues. Poliovirus strains at a known concentration were used to estimate the efficiency of the detection method (RT-PCR, as described by Häfliger et al. [19]). The poliomyelitis vaccine Poloral Berna (Serum- und Impfinstitut, Berne, Switzerland) used consisted of three poliovirus strains, Sabin I (2 × 10⁶ 50% tissue culture infective doses [TCID₅₀]/ml; 1 TCID₅₀ corresponded to 10 to 100 viral particles), Sabin II (2 × 10⁵ TCID₅₀/ml), and Sabin III (6 × 10⁶ TCID₅₀/ml). We established that the detection limit of our RT-PCR was a 10⁻⁶-fold dilution, which corresponded to 8.2 TCID₅₀ (if there were 10 to 100 viral particles per TCID₅₀). Detection of approximately 100 copies was based on results obtained by Mullendore et al. (31).

NLV strains belonging to genogroups I and II were isolated from different stool specimens. The NLV genogroup I positive control sample contained Southampton-related viruses (accession number L07418), and the NLV genogroup II positive control sample contained Camberwell-related viruses (accession number AF145896). The detection limits for seeded NLVs in 1.5 g of oyster tissue were established, and dilutions of 10⁻³ for NLV genogroup I and 10⁻² for NLV genogroup II remained positive (results not shown).

For HAV vaccine (7 × 10⁶ TCID₅₀/ml; University of Berne, Berne, Switzerland), we established a detection limit of a 10⁻⁶-fold dilution, which corresponded to 7 TCID₅₀ (if there were 10 to 100 viral particles per TCID₅₀) for our RT-PCR.

Negative control samples.

To exclude possible cross-contamination or inhibitory effects, we included two negative control samples per run; these samples consisted of 1.5-g tissue samples from previously negative oysters.

Oyster processing and virus isolation.

Oysters and all positive and negative control samples were analyzed by using a modified protocol of the virus isolation method described by Mullendore et al. (31). As external positive controls, four oyster tissue samples (1.5 g each) from previously negative samples were each experimentally contaminated by direct inoculation of 10 μl of NLV genogroup I (10⁻³-fold dilution), 10 μl of NLV genogroup II (10⁻²-fold dilution), 10 μl of HAV (10⁻⁶-fold dilution), and 10 μl of poliomyelitis vaccine (10⁻⁶-fold dilution). After vortexing for 60 s and incubation for 10 min at room temperature, the samples were analyzed by using the same protocol that was used for test samples.

Oysters (five oysters per sample) were shucked, the stomach and the digestive diverticula of each oyster were removed by dissection to obtain 1.5 g, and the tissues were transferred to 50-ml Falcon tubes. Fifteen milliliters of chilled sterile 0.05 M glycine-0.14 M NaCl buffer (pH 7.5) was added to the tissue in each tube, and the tissue was then homogenized twice on ice with a Waring blender at high speed. In order to prevent cross-contamination, we sterilized the homogenizing part of the Waring blender with ethanol (70%) and heating for 60 s with a Bunsen burner after each sample was processed. After centrifugation at 5,000 × g for 20 min at 4°C, the supernatant was collected in a second Falcon tube and stored at 4°C. The pellet was resuspended in 15 ml of 0.5 M threonine-0.14 M NaCl (pH 7.5) by vortexing for 60 s. After centrifugation at 5,000 × g for 20 min at 4°C, the supernatant was combined with the first supernatant in a third Falcon tube, and the pellet was discharged. Subsequently, 15 ml of a polyethylene glycol 6000 solution (12% [wt/vol] polyethylene glycol 6000, 0.3 M NaCl; 4°C) was added, and the suspension was precipitated at 4°C for 2 h. The resulting floc was sedimented by centrifugation at 6,700 × g for 30 min at 4°C and resuspended in 15 ml of phosphate-buffered saline (pH 7.5). After addition of 15 ml of chloroform to the pellet and vortexing for 60 s, the virus suspension was centrifuged at 1,900 × g for 30 min at 4°C. The supernatant was transferred and precipitated a second time by adding 7.5 ml of the 12% [wt/vol] polyethylene glycol 6000-0.3 M NaCl solution at 4°C for 2 h. After centrifugation at 14,000 × g for 15 min at 4°C, each of the polyethylene glycol pellets was resuspended in 1 ml of a 6 M guanidinium isothiocyanate solution. The suspension was incubated for 10 min at room temperature and centrifuged at 12,000 × g for 10 min. After this, 560 μl of the supernatant was used for RNA extraction.

RNA extraction.

RNA extraction was performed by using reagents from a QIAamp viral RNA mini kit (QIAGEN GmbH, Hilden, Germany) according to the manufacturer's protocol for a sample volume of 560 μl. Ten microliters of the 60-μl RNA solution obtained was used for each RT experiment.

Oligonucleotides.

Three NLV RT-PCR systems, one HAV RT-PCR system, and one EV RT-PCR system were used to perform all analyses. Table 1 shows the sequences and localizations of all of the oligonucleotides used. All primers were synthesized by Microsynth (Balgach, Switzerland) and stored freeze-dried at −40°C until they were used for the first time.

TABLE 1.

NLV, HAV, and EV primers used in this study

Oligonucleotide	Region	Sequence (5′-3′)a	Polarity	Localizationb
Genogroup-specific NLV, HAV, and EV oligonucleotidesc
NLV genogroup I
SRI-2 (PCR)	Capsid	AAA TGA TGA TGG CGT CTA AG	Sense	5344-5361
SRI-3 (RT-PCR)		AAA AYR TCA CCG GGK GTA T	Antisense	5584-5566
NLV genogroup II
SRII-2 (PCR)	RNA pol	TWC TCY TTY TAT GGT GAT GAT GA	Sense	4844-4866
SRII-3 (RT-PCR)		TTW CCA AAC CAA CCW GCT G	Antisense	5046-5028
HAV
HAV1 (PCR)	3D	TTT GGT TGG ATG AAA ATG GTT	Sense	6305-6325
HAV4 (RT-PCR)		ATT CTA CCT GCT TCT CTA ATC	Antisense	6716-6696
EV
EV05 (RT-PCR)	5′ UTR	CAC GGA CAC CCA AAG TAG T	Antisense	563-184
EV06 (PCR)		CAA GCA CTT CTG TTT CCC	Sense	448-467
Generic NLV oligonucleotidesd
Mon431 (PCR)	RNA pol	TGG ACI AGR GGI CCY AAY CA	Sense	5093-5112
Mon432 (PCR)	RNA pol	TGG ACI CGY GGI CCY AAY CA	Sense	5093-5112
Mon433 (RT-PCR)	RNA pol	GAA YCT CAT CCA YCT GAA CAT	Antisense	5285-5305
Mon434 (RT-PCR)	RNA pol	GAA SCG CAT CCA RCG GAA CAT	Antisense	5285-5305

^aMixed bases in degenerate primers: W = A or T; Y = C or T; K = G or T; R = A or G; and S = C or G.

^bLocalizations are localizations in reference to the positions in Norwalk virus (accession number M87661), HAV strain 18F (M59808), and poliovirus type Sabin 3 (X00596).

^cSee reference 19.

^dMonroe, personal communication.

The genogroup-specific NLV RT-PCR system used for separate detection of NLV genogroups I and II was based on degenerate primers described by Häfliger et al. (19) and located in highly conserved regions of the capsid gene for NLV genogroup I and of the RNA polymerase for NLV genogroup II. The predicted product sizes were 241 bp for NLV genogroup I and 203 bp for NLV genogroup II.

The generic NLV RT-PCR system used was based on primers for region B in ORF1 (3′ end) (S. S. Monroe, Centers for Disease Control and Prevention, Atlanta, Ga., personal communication). The predicted product size was 213 bp.

The HAV RT-PCR system was based on primers described by Häfliger et al. (19). The predicted product size was 412 bp.

The EV RT-PCR system was based on primers in the 5′ nontranslated region described by Häfliger et al. (19). The predicted product size was 400 bp.

RT.

Ten microliters of RNA was reverse transcribed for all four RT-PCR systems by using a Sensiscript RT kit (QIAGEN GmbH) according to the manufacturer's protocol. The RNAs were reverse transcribed by incubation for 60 min at 37°C, followed by incubation for 5 min at 95°C to inactivate the reverse transcriptase. Each reaction mixture (final volume, 20 μl) contained 1× RT Sensiscript buffer, each deoxynucleoside triphosphate at a concentration of 0.5 mM, 1 μM primer (SRI-2, SRII-2, Mon431 or Mon32, or EV06), 10 U of RNasin RNase inhibitor (N211A; Promega, Madison, Wis.), 1 μl of RT-Sensiscript, and 10 μl of template RNA.

PCR.

All 20 μl of a completed RT reaction mixture was mixed with 60 μl of a PCR mixture containing (final concentrations) 1× PCR buffer for a Taq DNA polymerase recombinant (Invitrogen, Basel, Switzerland), 2 μg of bovine serum albumin (Fluka, Buchs, Switzerland) per ml, each deoxynucleoside triphosphate at a concentration of 0.2 mM, 0.25 μM forward primer, MgCl₂ at a concentration of 1.5 mM (for all three NLV RT-PCRs) or 3 mM (for HAV and EV RT-PCRs), and 1 U of recombinant Taq polymerase (Invitrogen). Cycling was performed with a Biometra UNO II thermocycler (Biometra, Göttingen, Germany). The PCR program consisted of denaturation for 180 s at 95°C and 40 cycles of 40 s at 95°C, 90 s at 50°C, and 60 s at 72°C. Finally, an extension step consisting of 420 s at 72°C was performed. All three NLV RT-PCRs, as well as the HAV and EV RT-PCRs, were performed by using a one-tube RT-PCR in two steps; the same protocol (except the MgCl₂ concentration) was used in all cases.

Analyses of PCR products.

Ten microliters of each amplicon was mixed with 10 μl of loading buffer and analyzed on 2.6% agarose (agarose MP; Roche, Reinach, Switzerland) gels. Products were visualized by ethidium bromide staining and UV transillumination. Fragment sizes were compared with a commercially available size standard (100-bp DNA ladder; Promega Corp.). To determine the viral origins of amplicons, both strands of each positive RT-PCR product, including all positive controls, were sequenced.

DNA sequencing and sequence analysis.

Both strands of RT-PCR products were directly analyzed by cycle sequencing (ABI PRISM 377 DNA sequencer; Perkin-Elmer) carried out by Microsynth GmbH. A 132-bp amplicon of both NLV primer systems (specific and generic) and of the EV primer system for the viral RNA polymerase region were compared with GenEMBL data bank entries by using the Fasta program of the GCG software (Wisconsin Package, version 9.1; Genetics Computer Group, Madison, Wis.). Phylogenetic relationships were calculated by using Clustal W (39), and dendrograms were constructed by using the TreeView drawing software (34).

RESULTS

Weekly analyses of oyster samples.

To assess the percentage of virus-contaminated oysters imported into Switzerland, 87 samples consisting of five oysters each, resulting in 435 oysters, were analyzed for the presence of NLVs, HAV, and EVs. Sixty-one of the samples were exported by 31 different French suppliers, 12 of the samples were exported by three Dutch suppliers, and 14 of the samples were exported by two Irish suppliers. Eight (9.4%) of the 87 samples, from 6 (19.4%) of the 31 French suppliers, were positive for NLVs, 4 (4.9%) of the 87 samples, from 4 (12.9%) of the 31 French suppliers, were positive for EVs, and 2 (2.3%) of the 87 samples were positive for both viruses. All samples were negative for HAV as determined by this analysis. Positive results were obtained mainly at the end of January 2002. Figure 1 shows ethidium bromide-stained NLV genogroup II sequences (obtained with genogroup-specific primers) from positive oyster samples at week 8 on a 2.6% agarose gel. Three of 10 samples tested on this agarose gel contained 203-bp amplicons of NLV genogroup II genomes.

FIG. 1.

RT-PCR analysis of NLVs in 10 oyster samples in week 8 by using genogroup-specific primers and agarose gel electrophoresis. The arrow indicates the position of the 203-bp amplicon. Lanes 3, 7, and 8, NLV genogroup II sequence positive; lane P, positive control; lanes N, negative controls; lane M, marker.

Sequence analyses.

The sequences of NLVs belonged exclusively to genogroup II viruses, and EVs were represented by coxsackie- and echoviruses. Individual regions of France with higher contamination risks could be identified by comparing the viral sequences amplified and the locations of the suppliers. Three regions (regions 1, 2, and 3) of the eight French regions considered were positive for NLVs, and two of them (regions 1 and 2) were also positive for EV. The results of phylogenetic analyses of NLV sequences related to sequences of reference strains with GenBank accession numbers are shown in Fig. 2 for genogroup-specific NLV primers and in Fig. 3 for generic NLV primers. Hawaii virus (84 to 96% identity) and Toronto-Mexico virus (90% identity) sequences related to NLV sequences were detected in oyster samples from region 1. Bristol virus (94% identity), Camberwell virus (93 to 94% identity), and Hawaii virus (92% identity) sequences related to NLV sequences were found in samples from region 2. One sequence related to Snow Mountain virus (97% identity) was detected in samples from region 3. EV sequences related to echovirus 30 strain Bastianni were found in samples from region 1, and sequences related to human coxsackieviruses A were detected in samples from region 2 (results not shown).

FIG. 2.

Phylogenetic analysis performed with a 132-bp region of the RNA polymerase of NLV genogroup II detected with genogroup-specific primers (19), in which seven oyster samples were used. Two of the eight French exporting regions considered were positive for NLV genogroup II, and 6 of 31 French suppliers provided positive samples. Samples obtained from regions 1 and 2 are indicated by rounded rectangular and rectangular boxes, respectively. The previously published sequences of Lordsdale virus (GenBank accession number X86557), Camberwell virus (AF145896), Bristol virus (X76716), Mexico virus (U22498), Melksham virus (X81879), and Hawaii virus (U07611) were also included.

FIG. 3.

Phylogenetic analysis performed with a 132-bp region of the RNA polymerase of NLV genogroup II detected with generic primers. Weekly analyses of oyster samples were performed, and each sample consisted of five oysters from a supplier. One of the eight French exporting regions (samples from 6 of the 31 French suppliers investigated), in addition to the two regions detected by using genogroup-specific primers (Fig. 2), was positive for NLV genogroup II. Samples obtained from regions 2 and 3 are indicated by rectangular and hexagonal boxes, respectively. The previously published sequences of Lordsdale virus (GenBank accession number X86557), Camberwell virus (AF145896), Bristol virus (X76716), Melksham virus (X81879), and Hawaii virus (U07611) were also included.

DISCUSSION

Isolation of viruses from oysters by different techniques has been described several times (3, 12, 18, 19, 21, 25, 31). The choice of the virus extraction method which we used was supported by an experimental comparison of some of the available methods. The highest sensitivity was obtained with a protocol adapted from that of Mullendore et al. (31). The only modification consisted of using dissected digestive tissues instead of whole organisms. As the amount of tissue per organism was considerably reduced, this technique also enabled us to pool five oysters to obtain one sample in order to increase the sample volume and to approach the quantity usually consumed. The reproducibility and effectiveness of virus recovery when digestive tissues were used were demonstrated previously (37, 38) and were supported by a greater accumulation of viruses and lower concentrations of RNA extraction and RT-PCR inhibitors (28) in these tissues.

Cross-contamination was prevented by working in separate rooms and areas with sterilized reagents and equipment, by limiting the number of samples to 11 at one time and weekly, and by introducing two negative process controls. Positive process controls (digestive oyster tissues inoculated with viruses) allowed us to control the efficiency of virus recovery and inhibitor removal. External sequencing of both strands of each positive amplicon from a sample or control eliminated the risk of false-positive results and allowed a useful phylogenetic analysis to be performed.

The decision to screen oyster samples for NLVs, HAV, and EVs was based on different criteria. NLVs are the predominant gastrointestinal viruses worldwide (18, 24) and have been found to be the primary etiologic agents (52%) in reported cases of infectious disease associated with shellfish consumption (8-11, 14, 22). Furthermore, contamination of shellfish by HAV is a controversial issue, and opinions differ on whether contamination has been found (12, 17, 23, 28, 35). Also, little is known about contamination of shellfish by EVs. In a few publications, the authors (27, 28) have suggested that these agents also play an important role among viral contaminants in shellfish, which was a good reason to include them in these analyses.

The main time of oyster import into Switzerland determined the period of analysis, which was during 3 winter months. Furthermore, studies performed by Le Guyader et al. (28) with French oysters confirmed that these months are the months in which the rate of contamination of shellfish by viruses is highest.

The fact that no viral contamination was found in the oysters obtained from the Dutch and Irish suppliers, compared to the 19.4% of the French suppliers that provided oysters which were positive for NLVs and the 12.9% of the French suppliers that provided oysters which were positive for EVs, could have been a consequence of the low sample numbers due to the low import rates from these countries. Unfortunately, we could not find any study focusing on these regions with which to compare our results. However, the percentage of French suppliers who provided positive oysters is in accordance with results published in 2000 by Le Guyader et al. (28), who performed a 3-year assessment which revealed that 23% of French oysters were contaminated by NLVs and 19% were contaminated by EVs. We also confirmed the great diversity of NLV sequences observed by these authors (28) with the NLV strains related to different prototype strains (Bristol, Hawaii, Mexico, and Melksham). Moreover, strains related to Bristol virus, predominant in French region 2, were closely related to distinct strains of NLV identified by Noel et al. (33) and described as a 95/96-US subset having a global distribution. NLV sequences detected in previous studies (4, 5) in 2000 and 2002 focusing on European mineral waters and strains detected by Bon et al. (6) in French children and by Le Guyader et al. (28) in oysters suggest that NLV strains related to the 95/96-US subset are predominant in France and adjacent countries. Phylogenetic analysis of NLV strains in oysters from region 1 also showed a level of similarity of 95% to the American NLV strain NLV/Richmond/283/1994/US, and strains from region 2 exhibited a level of similarity of 94% with the US NLV strain NLV/Westover/302/1994/US; both of these strains were described by Ando et al. (2). Oyster samples exported from region 2 and surrounding French regions are known to be occasionally contaminated by NLVs. Miossec et al. (30) demonstrated that there was sequence similarity between RT-PCR NLV amplicons from stool samples collected from persons infected by consuming oysters from a supplier in region 2 and RT-PCR NLV amplicons from oysters from the same lot harvested 1 week later. They also linked contamination of the estuarine water concerned during this season with exceptional contamination of an inflowing river by wastewater as a result of exceptional precipitation in region 2 which resulted in inundation and overflows. As we obtained most of our NLV- and EV-positive results at the end of January, with oysters harvested at the end of December 2002, there could also have been a correlation between these results and the high levels of precipitation that occurred during this season in all three described regions (http://www.meteo.fr). The fact that the greatest accumulation of viruses in shellfish was in winter could also have been a consequence of the best virus survival occurring in estuarine water at the lowest water temperatures (7), supported by the highest shedding rates of viruses due to the winter seasonality of viral gastrointestinal illnesses (28). An additional reason for the high bioaccumulation rates in the winter could also be the fact that the highest concentration of glycogen, which seems to be responsible for the ionic binding of viral particles (13), in the connective tissue of shellfish occurs from November to March (7).

Estuarine waters of all three regions which we found to be occasionally contaminated with viruses are normally classified as waters of good bacteriological quality, not requiring any depuration step for oysters before export and consumption (30). This classification is made by using thermostable coliforms as indicators. Viral parameters are not taken into consideration, although no solid correlation could be demonstrated between bacterial contamination and viral contamination in oyster cultures (28, 35). The only viral parameter included in European regulations for testing the quality of food is the presence of EVs in water (35). Our study and a study of Le Guyader et al. (28) showed that there is no correlation between the occurrence of EVs and NLVs, thus eliminating EVs as indicators for pathogenic viruses. Coliphages were also proposed as indicators of fecal contamination, but no correlation was found by Legnani et al. (26) between coliphage concentrations and bacterial fecal indicators in shellfish. Furthermore, the highly variable bioaccumulation of coliphages (7) eliminates these agents as appropriate indicators in shellfish.

In summary, our results showed that a significant percentage of imported oysters was contaminated with NLVs. The direct impact on food safety might not be too important, since the overall consumption of oysters in Switzerland is rather low. However, it is well known that secondary cases easily occur after transmission from individuals infected with NLVs. In this respect, the epidemiological role of contaminated oysters could be more important than it seems to be. Ongoing investigations of nonbacterial food-borne outbreaks by case control study should support the assessment of risk due to consumption of raw oysters in Switzerland and help define preventive measures.