Norovirus contamination in different shellfish species harvested in the same production areas

Authors


Correspondence

Luciana Croci, Dipartimento di Sanità Pubblica Veterinaria e Sicurezza Alimentare, Istituto Superiore di Sanità, Viale Regina Elena, 299, 00161 Rome, Italy. E-mail: luciana.croci@iss.it

Abstract

Aims

To investigate Norovirus (NoV) contamination of mussels, clams and oysters harvested in two class B harvesting areas of the delta of the Po river, to choose a species as an indicator.

Methods and Results

Environmental parameters (temperature and salinity) and hydrometric levels of the tributary river were measured. Seventy shellfish samples (35 samples per area) were examined for Escherichia coli and NoV (GI and GII). NoV contamination was found in 51·4% of samples, of which, 2·9% contained only NoV GI, 14·3% only NoV GII, while the majority of the samples (34·3%) contained both genogroups. Most of the positive results (90·0%) were obtained in the period between November 2008 and April 2009.

Conclusions

No significant differences were found between the results from the two harvesting areas and the three shellfish species. However, on the basis of the average Ct values, the recovery rate (from 0·46 to 1·15%) and the distribution of positive results in the samplings, mussels seem to be a suitable indicator species to monitor viral contamination in these areas.

Significance and Impact of the Study

The data allow the optimization of monitoring plans to improve the prevention strategies in terms of money and time, by the intensification of controls in the cold season and the use of one species as indicator.

Introduction

Noroviruses (NoV) belong to the family Caliciviridae and are considered the leading cause of nonbacterial human gastroenteritis worldwide (Koopmans and Duizer 2004; Atmar and Estes 2006; Svraka et al. 2007). NoV display a broad genomic diversity, with about 40 genotypes clustering in five genogroups (GI to GV), of which GI and GII are most commonly associated with human infections. NoV GI and GII are further subdivided into different clusters or genotypes (eight for GI and 17 for GII) based on capsid protein sequences (Zheng et al. 2006). Although many genotypes may co-circulate in a specific area, the GII genotypes seem to have a prominent epidemiological role in human gastroenteritis and, in particular, the GII.4 variants (Kroneman et al. 2006; Phan et al. 2006; Mesquita and Nascimento 2009). Shellfish play an important role in the transmission of NoV, as filtering of large volumes of seawater contaminated by faecal waste may result in the accumulation of these pathogens to considerable levels in shellfish tissue (Metcalf et al. 1979; Rippey 1994; Burkhardt and Calci 2000). Thus, consumption of raw or improperly cooked shellfish is a major risk factor for food-borne outbreaks (Koopmans and Duizer 2004; Prato et al. 2004; Huppatz et al. 2008; Le Guyader et al. 2008; Guillois-Becel et al. 2009). NoV illnesses because of shellfish consumption present a seasonal pattern, generally showing a peak incidence during the wintertime (Rippey 1994; Rohayem 2009). This seasonality could be attributed to several factors, including increased stability of viruses at low water temperature, reduced solar inactivation and selective bioaccumulation of these pathogens by the shellfish. Moreover, it has been reported that the bioaccumulation and elimination kinetics of bacteria or viruses by bivalve molluscs vary with the shellfish species, type of micro-organism and environmental conditions (Burkhardt and Calci 2000). In particular, the bioaccumulation of viruses by shellfish during feeding is influenced by many factors as mucus production, seasonal physiological changes, etc. (Di Girolamo et al. 1977; Burkhardt and Calci 2000), and for NoV, strain-specific binding to shellfish tissue glycans (Maalouf et al. 2011) and seasonal variation of ligands expression has been demonstrated (Maalouf et al. 2010; Le Guyader et al. 2012).

Previous monitoring studies report that the prevalence of NoV in shellfish from the Italian harvesting areas, classified as B, is about 31·1%, and the different species are not equally involved in viral contamination (Croci et al. 2010). The objective of the present study was to evaluate the contamination of NoV in different seasons of the year in shellfish, and in two class B harvesting areas influenced by the Po River at the northern Adriatic Sea. The presence of NoV in three shellfish species (mussels, clams and oysters) harvested in the same area and the circulation of NoV GI and GII were examined to choose a species as an indicator of contamination and to optimize monitoring plans controlling the products' safety.

Materials and methods

Harvesting areas

Two harvesting areas (Area 1 and Area 2) classified as B according to the European Regulation 854/2004 were selected in the delta area of Po River (Fig. 1). Environmental parameters (temperature and salinity) in the two areas were continuously monitored using multiparameter probes (IDROLAB H 20) and continuous data transmission through GSM communication. Hydrometric levels of the tributary river were measured by the AIPO (Agenzia Interregionale per il Fiume Po; Interregional Agency for the Po river).

Figure 1.

Selected harvesting areas (Area 1 and Area 2) in the delta of Po River.

Samples

Samples of mussels (Mytilus galloprovincialis), clams (Tapes philippinarum) and oysters (Crassostrea gigas) were collected from the two harvesting areas each month for 1 year and were shipped to the laboratory in insulated boxes. A total of 70 shellfish samples (35 for each area; 23 mussels, 24 clams and 23 oysters) were examined for the determination of Escherichia coli and the presence of NoV GI and GII.

Bacteriological analysis

Fifty grams of shellfish (meat and liquor) was homogenized in a blender for 30 s at maximum speed (Osterizer Pulse Magic 16, Milwaukee, WI, USA). Escherichia coli enumeration was performed in accordance with the ISO/TS 16649-3:2005 method.

Viral analysis

Viral analysis was performed using a standardized protocol (Lees and CEN WG6 TAG4 2010). Ten individuals (30 for clams) were selected from each sample, and the digestive gland was aseptically dissected and finely chopped with a sterile blade and used for the nucleic acids extraction.

Nucleic acid extraction

A 2-g aliquot of digestive tissue and 10 μl of process control were mixed, and lysis of tissue was performed with 2 ml of proteinase K solution (0·1 mg ml−1). Suspensions were vigorously agitated using a vortex for about 1 min, incubated at 37°C in a shaking incubator for 60 min to allow digestion and then placed in a water bath at 60°C for 15 min to inactivate the enzyme. The samples were then centrifuged at 3000 g for 5 min, and the supernatant was collected. The volume of supernatant was recorded for each sample (range from 2·3 to 3·0 ml; median value, 2·7 ml) and was normalized to 3·0 ml by the addition of sterile phosphate-buffered saline (pH 7·3).

Viral RNA extraction and purification was performed using the MiniMag NucliSens Magnetic Extraction kit (bioMerieux, Marcy l'Etoile, France) according to the manufacturer's instructions. The method is based on guanidine thiocyanate lyses coupled with the nucleic acid–binding properties of silica particles (Boom et al. 1990). For each sample, 500 μl of supernatant was subjected to lysis with the appropriate buffer; binding to magnetic silica and washing steps with the specific solutions of the kit to remove amplification reaction inhibitors were successively performed. Nucleic acids were recovered in 100 μl of elution buffer and were stored at −80°C until they were subjected to real time RT-PCR analysis.

Process control

The efficiency of the extraction procedure was determined by the addition to each sample of a titrated suspension (TCID50 3·2 × 105 ml−1) of Mengovirus (avirulent clone mutant vMC0, kindly provided by Albert Bosch, Department of Microbiology, University of Barcelona, Spain) as a process control (Costafreda et al. 2006; Le Guyader et al. 2008). Recovery of Mengovirus was determined by real time RT-PCR, comparing Ct values obtained on shellfish samples extracts and on viral stock, taking into account the dilution factor due to the extraction procedure and the aliquot of sample subjected to the analysis. Real time RT-PCR for Mengovirus and calculations were performed according to Costafreda et al. (2006).

Real time RT-PCR

The analyses were performed by a one-step real time RT-PCR method based on two separate reactions for the qualitative detection of GI and GII strains respectively (da Silva et al. 2007). The primers and probes used to detect GI were forward primer QNIF4 (5′-CGC TGG ATG CGN TTC CAT-3′), reverse primer NVILCR (5′-CCT TAG ACG CCA TCA TCA TTT AC-3′) and probe NVILCRpr (5′-TGG ACA GGA GAY CGC RAT CT-3′). Primers and probe targeting GII were forward primer QNIF2 (5′-ATG TTC AGR TGG ATG AGR TTC TCW GA-3′), reverse primer COG2R (5′-TCG ACG CCA TCT TCA TTC ACA-3′) and probe QNIFS (5′-AGC ACG TGG GAG GGC GAT CG-3′) (Y = C or T; R = A or G; N = A, C, G, or T; W = A or T). The two probes were labelled with 6-carboxyfluorescein (FAM) at the 5′ end and with 6-carboxytetramethylrhodamine (TAMRA) at 3′ end. Primers and probe targeting the Mengovirus were forward primer Mengo110 (5′-GCG GGT CCT GCC GAA AGT-3′), reverse primer Mengo209 (5′-GAA GTA ACA TAT AGA CAG ACG CAC AC-3′) and probe Mengo147 (5′-ATC ACA TTA CTG GCC GAA GC-3′) labelled with 5′ FAM and 3′ minor groove binder (MGB).

Reverse transcription and PCR were performed on an ABI Prism 7700 Sequence Detector System (Applied Biosystems, Monza, Italy) using the Ultrasense One-step qRT-PCR System (Life Technologies, Monza, Italy) and the following reaction mix: 1× Ultrasense reaction mix, 900 nmol l−1 reverse primer, 500 nmol l−1 forward primer, 250 nmol l−1 TaqMan probe, 1× ROX reference dye, 1·25 μl of Ultrasense enzyme mix and 5 μl of sample nucleic acid (final reaction volume 25 μl). The amplification conditions were reverse transcription for 60 min at 55°C followed by 5 min at 95°C and 45 cycles of 15 s at 95°C, 1 min at 60°C and 1 min at 65°C. Amplifications were performed in triplicate using the described protocols.

A control for RT-PCR inhibition in individual samples was carried out by co-amplification of an external control (EC) RNA (approximately 103 copies of target sequence), and amplification efficiency was calculated by comparison with the Ct value of EC-RNA alone (math formula). Results from samples were considered acceptable if amplification efficiency was ≥90%. Samples failing the criterion (three mussel, four clam and eight oyster samples) were retested and tested also in their dilutions (1 : 2 and 1 : 10) to confirm either absence of amplification or detection of target sequences. In the case of positive samples (one mussel, one clam and three oyster samples), theoretical Ct value for the undiluted sample was calculated by subtraction of 1·0 or 3·33 to the Ct obtained on 1 : 2 or 1 : 10 dilution. In each run, two negative controls (molecular grade water) and a positive control (EC RNA, expected Ct value 25·1 ± 0·7 for NoV GI reaction and 26·1 ± 0·8 for NoV GII) were added. Runs were considered acceptable when Ct values for positive controls fell within the defined range and no amplification was detected in the negative control samples.

Statistical analysis

The results obtained from the two areas and from the different mussel species were analysed by chi-square (χ2) test with Yates' correction for continuity.

Results

The results of the environmental monitoring showed that the Po River had less impact on the salinity of Area 1 than on Area 2. Salinity in Area 1 ranged between 21 and 36‰ (average 26·7‰, variability K = 15), while in Area 2, which was directly connected to the river by inland channels, salinity ranged between 13 and 31‰, with an average of 22·2‰ and higher variability (K = 18). The temperature during monitoring varied between 4°C in January 2009 and 26°C in July 2009 but was substantially equivalent in the two areas under observation during the same monitoring time. The hydrometric levels of the Po River were below zero during the whole period and ranged between a maximum level in February 2009 (−1·0 m) and two minimum levels registered in October 2008 and August 2009 (−5·1 m).

The bacteriological analysis of the shellfish showed that, in all species, the number of E. coli was generally below the European legislation limit (4600 MPN per 100 g in samples from category B areas), with fluctuations between values <20 and 1300 MPN per 100 g. A number of E. coli above the limit was observed in two samplings, performed in November 2008 and February 2009 in Area 2, in which the samples with the highest contamination reached 9200 MPN and 5400 MPN per 100 g respectively.

The viral analysis demonstrated that among a total 70 shellfish samples analysed, 36 (51·4%) resulted positive for the presence of NoV (Table 1); NoV GII was present in 34 samples (48·6%), whereas NoV GI was present in 26 samples (37·1%). About one-third of the samples (34·3%) contained both genogroups (GI and GII), 10 samples (14·3%) were positive only for NoV GII, and only two samples (2·9%) from Area 2 were positive for NoV GI only. There was no statistically significant difference (P = 0·811) between the frequency of positive samples from the two harvesting areas.

Table 1. Ct values obtained in real time RT-PCR detection of Norovirus
 NoV GINoV GII
Area 1Area 2Area 1Area 2
MusselClamOysterMusselClamOysterMusselClamOysterMusselClamOyster
  1. −, negative sample; NT, not tested; Avg, average; NoV, Norovirus.

October 200839·4738·93
November 200836·2936·5038·2436·3235·3536·0236·83
December 200832·8634·6635·3933·6235·1537·4430·4230·8332·8330·7231·1433·04
January 200935·4037·8333·9434·6931·0533·5434·7329·0933·91
February 200934·8232·4134·9235·2234·2434·3631·9627·6430·3331·4229·3930·77
March 200937·2436·8833·3140·6934·1034·7834·23
April 200938·0336·7638·2438·2139·78-34·0531·5834·4133·17
May 2009NTNT
June 2009
July 2009
August 2009NTNT
September 200936·8638·93
Avg35·85 ± 5·3334·81 ± 1·8236·60 ± 1·6836·09 ± 1·7436·27 ± 2·6037·09 ± 2·5733·46 ± 2·0833·11 ± 4·7633·04 ± 1·6334·73 ± 2·9131·76 ± 2·6633·76 ± 2·19
Avg mussel35·97 ± 1·8034·14 ± 2·54
Avg clam35·62 ± 2·2832·50 ± 3·89
Avg oyster36·81 ± 1·9233·40 ± 1·86

Avg all

species

36·07 ± 1·9933·39 ± 2·89

NoV presence was detected in samples covering the whole range of E. coli contamination levels (from 20 to 9200 MPN per 100 g), although frequency of NoV-positive samples increased in relation to E. coli concentrations: 35% in shellfish below 230 E. coli MPN per 100 g, 87% in samples ranging from 230 to 4600 and 100% for the samples above 4600 MPN per 100 g (data not shown).

Mussels provided a higher number of positive samples (60·9%), whereas 45·8% of clams and 47·8% of oysters were found positive (Table 1); differences were not statistically significant (P = 0·459 and P = 0·554, respectively). In all three species, the majority of positive samples (9/14 for mussels, 9/11 for clams and 6/11 for oysters) contained both NoV GI and GII.

Virus recovery varied in the three species from an average of 0·23 ± 0·07% in clams (range 0·14–0·36%) to 0·58 ± 0·36% in oysters (range 0·19–1·27%) and 0·80 ± 0·23% in mussels (range 0·46–1·15%), and no significant difference was detected between positive and negative samples in the recovery rate or in the inhibition of real time RT-PCR.

Most of the positive results to either GI or GII (90·0%) were obtained in the period between November 2008 and April 2009 (Table 1). Positive samples were detected in 15 samplings. In nine of them, all shellfish species (mussel, clams and oysters) were contaminated, while in six samplings, only one (either mussels or clams) or two species (mussels and oyster or mussels and clams) were involved.

The average Ct values obtained for the three shellfish species were similar, varying from 35·62 ± 2·28 of clams to 36·81 ± 1·92 of oysters in NoV GI detection and from 32·50 ± 3·89 (clams) to 34·14 ± 2·54 (mussels) in NoV GII analysis.

Discussion

The measures of the environmental parameters in the two harvesting areas showed predictable fluctuations related to seasonal variations (temperature) and to the different impact of the Po River (salinity). Most of the contaminations (90·0%) from both areas were detected during the period from November 2008 and April 2009 and no sample resulted positive in the summer, showing a clear seasonal pattern.

The high percentage of NoV-contaminated shellfish (48·6% in Area 1 and 54·3% in Area 2) condensed in 6 months, and the coexistence of two genogroups in the majority of positive samples (37·1% from Area 1 and 31·4% from Area 2) indicates a widespread diffusion of NoV, without significant differences between the two harvesting areas.

With regard to the circulation of the two NoV genogroups, GII was detected more frequently than GI (48·6% vs 37·1%). Meanwhile, a high prevalence of NoV GI was noticed, confirming the data of other authors on marine environment (Gentry et al. 2009) and shellfish (Le Guyader et al. 2008), where multiple contamination is frequently detected also in outbreak implicated samples (Le Guyader et al. 2012).

In the majority of cases, NoV were found in all three species of shellfish collected during the same sampling. Based on these data, the three shellfish species seem to be similarly involved in the viral contamination in the two areas under investigation, despite the different cultivation procedures (Tapes on the sand of the lagoon, Mytilus on ropes attached to stocks and Crassostrea in floating cages) and the variations on the bioaccumulation.

Therefore, to minimize the costs of a monitoring plan towards viral contamination, it could be possible to use a single species as indicator also for the other shellfish species growing in the same area. In our study, mussels showed the potential to serve as an indicator species. In fact, even though RT-PCR of clams provided average Ct values lower than the other two species (presumably indicating a higher contamination level), mussels showed a higher average recovery compared to the other species (0·80% vs 0·23% of clams and 0·58% of oysters), a characteristic that could aid reducing the chance of false negative results because of low recovery efficiency. Besides, mussels showed also the highest frequency of positive results (60·9%), combined with a distribution of such results covering almost the totality of samplings (mussels were positive in 14 of the 15 samplings where NoV were detected), therefore showing representatively of contamination in the areas during the observed period. Finally, mussels seem a good choice as an indicator species both from a laboratory approach (relative easiness of sample preparation for viral analysis, especially when compared to clams) and from a production management point of view (commercial value of the species), combining the aim to obtain reliable results with the needs for a practical approach to monitoring. The data gathered in this study could contribute to the setting-up of specific monitoring programs, which could be further improved through the acquisition of quantitative data on sample contamination (virus genome copies per gram) once reliable standards for real-time calibration will be available. Taking into account the seasonal trend of contamination, and the intensification of samplings during the cold season by using mussels as indicator species for NoV contamination, could have a great impact on money and timewise. Optimization of monitoring plans could lead to better management of the shellfish harvest.

Acknowledgements

This work has been partially supported by 7th European Framework projects, Grant agreement 222738; ‘Selection and improving of fit-for-purpose sampling procedures for specific foods and risks’ (http://www.baselineeurope.eu/). We acknowledge the Agenzia Interregionale del Fiume Po – Section Porto Tolle (RO) for the data on the hydrometric levels of the Po river.

Ancillary