The effectiveness of domestic cook on inactivation of murine norovirus in experimentally infected Manila clams (Ruditapes philippinarum)




The aim of this work was to evaluate the efficacy of domestic cooking in inactivating Manila clams experimentally infected with murine norovirus (MNV).

Methods and Results

A cooking pan was modified to enable electronic temperature probes to be positioned to record both flesh and environment temperature. Manila clams were infected with 104 TCID 50% ml−1 of MNV. The infected whole-in-shell clams, divided into three replicates, were cooked on an electric stove, and groups of nine clams were removed from the pan at fixed intervals. Pools of three digestive glands were examined by virus isolation to ascertain residual viral load.


Results showed that 10 min of cooking by a traditional domestic method at a temperature close to 100°C, for at least 2 min, can completely devitalize the MNV in infected clams. This is generally the time needed for the majority of valves to open up.

Significance and Impact of the Study

At present, it is highly recommended to label all lagoon products as ‘requiring cooking before consumption’, but no specifications are given on how long and at what temperature they should be cooked. Our results can provide the consumer with useful indications on how to cook clams to prevent any risk of foodborne illness.


Noroviruses (NoVs) are members of the Caliciviridae family, including a group of nonenveloped icosahedral single-stranded positive-sense RNA viruses.

They are currently divided into five genogroups and 29 different genotypes. Three genogroups (GI, GII and GIV) infect humans, while genogroups III (GIII) and V (GV) infect cattle and mice, respectively (Zheng et al. 2006).

Human norovirus (HuNoV) is considered the major common cause of acute epidemic nonbacterial human gastroenteritis (Hutson et al. 2003; Lopman et al. 2003). They can easily contaminate food being ubiquitous, very small (27–30 nm), extremely stable, resistant to common disinfectants and highly infective. It only takes a few particles (<100) to cause infection and gastroenteritis (Koopmans et al. 2002). The disease caused by HuNoV is characterized by vomiting and diarrhoea, a short incubation period (15–48 h) and spontaneous resolution of symptoms within a few days.

Although fruits (i.e. strawberries, raspberries or blackberries) (Brassard et al. 2012) have recently been a major vehicle of infection for EU consumers, raw and/or undercooked shellfish continue to be responsible for several disease outbreaks (Prato et al. 2004). Bivalve seawater molluscs for instance can accumulate HuNoV by filtering water for food (Burkhardt and Calci 2000; Lees 2000; Hall et al. 2012). Oysters pose the highest risk for human health in Europe and in Italy, followed by mussels and clams, as oysters are routinely consumed raw (Le Guyader et al. 2011; EFSA 2011; Galmès Truyols et al. 2011; Wall et al. 2011). There are several potential transmission pathways: food and water contaminated with faecal matter (Gentry et al. 2009), contact with infected people (De Laval et al. 2011) and surfaces (Feliciano et al. 2012) and prolonged shedding of virus from both symptomatic and asymptomatic individuals (Rockx et al. 2002; Atmar et al. 2008). Infected as well as immunocompetent individuals can in fact shed large amounts of virus particles for two or more weeks after resolution of symptoms (Codex Alimentarius Commission, GL 79–2012). High rates of infection (>30%) are commonly observed during outbreaks of HuNoV in closed settings such as school canteens, healthcare facilities (e.g. hospitals or nursing homes) and cruise ships (Lopman et al. 2004; Verhoef et al. 2008) although airborne transmission cases have also been reported (Atmar and Estes 2006).

NoVs are present in several animal species such as swine, cattle and mice, and several of these have been used as surrogates for HuNoV (Duizer et al. 2004; Bae and Schwab 2008; Buckow et al. 2008). Some of these viruses have proved unsuitable as HuNoV surrogates (Slomka and Appleton 1998; Hewitt and Greening 2004), while others, that is MNV, are still in use due to failure to grow HuNoV in cell cultures.

A new cultivable norovirus affecting mice (murine norovirus, MNV; genogroup V) was reported and characterized in 2003 (Karst et al. 2003). This virus is readily cultured on the mouse macrophage cell line, RAW 264.7, and can be easily quantified using a plaque assay. Given its taxonomic relationship to HuNoV, MNV is considered a potential surrogate for increasing current knowledge of HuNoV properties and mode of replication (Wobus et al. 2004, 2006; Virgin et al. 2006; Ward et al. 2007). MNV has been extensively used in experimental studies: to test the efficacy of high-pressure processing (HPP) (Kingsley et al. 2007; Lou et al. 2011), inactivation by UV irradiation, pH, temperature and disinfection (Cannon et al. 2006; Belliot et al. 2008; Lee et al. 2008), and survival in water (Bae and Schwab 2008) and shellfish (Provost et al. 2011). MNV may not be as resistant to physical agents as other NoVs (Sattar et al. 2011): some authors have demonstrated that a pressure of 500 MPa or above completely inactivated MNV and feline calicivirus (FCV) (Kingsley et al. 2007; Leon et al. 2011; Lou et al. 2011; Arcangeli et al. 2012), while more recent studies have shown that HPP was able to destroy the HuNoV capsid only at 800 and 900 MPa for 15 and 2 min, respectively (Lou et al. 2012).

MNV was totally inactivated (5-log10) in PBS after 3 min at 72°C, but less thermal inactivation would be expected in viruses protected within shellfish tissues (Wolf et al. 2009). The digestive glands of mussels contaminated with FCV were still infected after cooking at 80°C for 15 min (Croci et al. 2012). However, further studies are warranted to better understand the heat resistance of MNV in different shellfish and in different tissues.

As mentioned above, shellfish can be easily contaminated by HuNoV. Between 2000 and 2010, Suffredini and colleagues conducted an investigation on 1793 Italian bivalve molluscs, of which 17·6% tested HuNoV positive. The lagoon environment has a little fresh water inflow and frequent injections of contaminated fresh water, so bivalves farmed in these areas are generally more susceptible to HuNoV infection (Suffredini et al. 2011, 2012). Recent studies conducted in Italy have confirmed the presence of HuNoV in different parts of the country (La Rosa et al. 2012; Terregino et al. 2012; Pavoni et al. 2013). In addition, the most common mollusc depuration processes are unable to rapidly inactivate HuNoV. Ueki et al. (2007), for example, found HuNoV in oysters up to 10 days after purification; Provost et al. (2011) showed that HuNoV is present in oyster haemocytes in acidic pH up to 12 days post infection; other authors have demonstrated that, to date, a routine period of depuration in commercial depuration systems does not completely free products from viral contamination (Savini et al. 2009; Terregino et al. 2012).

This study aims to give producers and consumers clear indications on how to cook clams to completely inactivate eventually present NoVs, as recommended by EFSA (EFSA 2011). The study was also designed to evaluate the effect of the time taken and the temperature traditionally used to cook Manila clams (Ruditapes philippinarum) on the survival rate of MNV.

Materials and methods


The MNV used to experimentally contaminate the samples (kindly provided by Prof. H. W. Virgin, Washington University, School of Medicine, MO, USA) was grown on a RAW 264.7 cell line, derived from murine macrophages (ATCC cat. numb TIB-71), cultured in Dulbecco's modified Eagle medium (Sigma-Aldrich cat. numb. D6429), supplemented with 10% foetal bovine serum (Euroclone, cat number ECS0180L) and 1% penicillin and streptomycin (Sigma P781, 10 000 IU penicillin and 10 mg streptomycin) and incubated at 37°C with 5% CO2. After cellular lysis, the lysate was frozen, thawed twice and then centrifuged at 1000 g for 10 min. The supernatant was filtered through a 0·45-μm filter, and the virus stock was stored at −80°C.

The infectious titre of the virus, expressed in TCID 50% ml−1 (50% tissue culture infectious dose) according to the Reed and Muench formula (Reed and Muench 1938), was calculated after preparing ten-fold serial dilutions of the virus and inoculating them into the RAW 264.7 cells seeded on 96-well plates immediately before infection. Each dilution was tested in quadruplicate using 0·05 ml to infect each well. The plates were incubated for 5 days at 37°C in a 5% CO2 atmosphere and the final cytopathic effect read.

Experimental design

A total of 204 Manila clams (R. philippinarum) were harvested from a lagoon located in the Po River Delta (north-east of Italy) in September, during a nonrainy period and with an E. coli load of <230 CFU per 100 g and analysed, in accordance with Reg. (EC) No. 854/2004, by the ISO/TS 16649-3:2005 microbiological method to calculate MPN. To avoid additional stress and preserve filtering ability, nondepurated clams were used for the purposes of this study.

Clams were acclimatized for 1 day prior to experimental infection in an aerated tank containing 100 l of artificial seawater, prepared using synthetic marine salt (Instant Ocean, Aquarius System, Sarrebourg, France). The seawater conditions were as follows: salinity 3·5%; water temperature 15°C; dissolved oxygen levels were higher than 90%. The seawater was contaminated with MNV at the final concentration of 10TCID 50% ml−1. The shellfish, weighing about 10 g per each (4–5 cm in length), were immersed in contaminated water for 24 h. The presence of excurrent siphons was observed, confirming that the clams had been in contact with the external water. Clams intended to be used as negative controls were kept in a tank with the same environmental conditions, but without virus contamination. After contamination, the clams were immediately collected and kept refrigerated until the heat treatment.

A total of 108 clams were divided into three groups (36 in each group) and used to replicate the same test. A commercial aluminium nonstick 28-cm-diameter pan was modified to accommodate three electronic temperature probes (Ellab Eval-Flex®) to record both flesh and environment temperature. One probe was set in an open clam, one in a closed clam and the third was left free in the pan. The probe sensor had an accuracy of <0·2°C and was calibrated to ±0·05°C, with a response time of 0·8 s. The data recording software was Ellab A/S Valsuite Pro® version 2.8 FDA for compliance with 21 CFR Part 11. Each group of 36 clams was put in the experimental pan covered with a glass lid and placed on an electric glass ceramic hob (A.D. mod 2F-6-500-450, power 4300W).

Groups of nine whole-in-shell clams were removed from the pan when the valves of 50% of the clams had opened up (T0). To maintain the right temperature inside the pan, the lid was opened only three times to collect the clams and at an interval of 2 min from each previous collection time (T1, T2 and T3, respectively). Immediately after collection, the clams were chilled with ice and water and then frozen and kept at −80°C until analysis. All clams were processed for virological analyses in pools of three clams. The entire experiment was repeated three times.

Two groups of thirty clams each, one infected but not heat-treated and one neither infected nor heat-treated, were used as positive and negative controls, respectively.

A preliminary test was conducted to check actual virus depletion at the time the valves opened up. Thirty-six whole-in-shell clams were treated as described above, but the first nine clams were collected individually at the time of valve opening (T0). T1, T2 and T3 indicate the cooking times after 50% of the valves had opened up (2, 4 and 6 min, respectively). Groups of nine clams were collected at each time point. All clams were processed for virological analyses in pools of three clams. This test was carried out once only to increase data on the heat resistance of MNV.

Virological and molecular analyses

The groups of clams (n = 9) collected at the different time points were processed in three pools of three clams each. Three digestive glands were removed by dissection, pooled and homogenized by mortar and pestle with sterile sand quartz and diluted 1 : 2 (w/v) with MEM supplemented with antibiotics (penicillin G 10 000 IU ml−1, streptomycin 10 mg ml−1, nystatin 5000 IU ml−1). The homogenized samples were then centrifuged at 2500/3000 g for 30 min, and the resulting supernatants were collected and kept at 4°C overnight to allow sufficient time for the antibiotics to have an effect. The supernatants were titred by inoculating, in quadruplicate, 0·05 ml of the ten-fold serial dilutions prepared on the RAW 264.7 cells seeded on 96-well plates immediately before infection. The infected monolayers were incubated for 5 days at 37°C in a 5% CO2 atmosphere and periodically observed under microscope to check for cytopathic effects (CPEs). At the end of the incubation period (5 days), where present, titres were expressed in TCID 50% ml−1 according to the Reed and Muench formula (Reed and Muench 1938). After 5 days, all negative plates were frozen at −80°C and then thawed. The lowest dilution wells were used for a second infection of new monolayers and observed for a further 5 days before considering the sample negative. Subcultivation was used to highlight the presence of a low viral load unable to produce a clear cytopathic effect in the first cell infection.

After the second passage on cell culture, two other subcultures were performed on the positive samples, as described above, for a total of four passages. The cell culture cryolysate of the final subculture was analysed by real-time PCR (Baert et al. 2008) and by negative staining electron microscopy (EM) observation (Doane and Anderson 1987) to confirm identification of MNV.


Cell culture cultivation performed on three pools of three digestive glands each, collected from infected but not heat-treated clams (positive control), showed CPE and an average viral titre of 104·07 TCDI 50% ml−1, confirming that experimentally infected clams filtered and concentrated MNV in their digestive glands. PCR and electron microscopy confirmed the identity of the virus.

Interestingly, virus growth was not observed in any of the three experimental replicates (Table 1 and Fig. 1), indicating that traditional domestic cooking methods can completely devitalize the virus after 7 min and 45 s of heating (±00:00:07 min), at an environment temperature of 97·44 (±3·138508)°C. Temperature data obtained in these experiments, during which the lid of the pan was opened only three times, confirmed the repeatability of the experiment as shown in Table 1.

Table 1. Summary of results of the three replicate cooking experiment. A total of 108 clams were used in this experiment. Nine clams were collected at each time point (T0, T1, T2, T3). T0 corresponds to the opening time of 50% of the clams. Virological analyses were performed on pools of three clams. The same experiment was repeated three times, and the merged results are shown in the table
Time pointDescriptionTime (hh:mm:ss)Probe 1 (°C) environment temperatureProbe 2 (C) open clam temperatureProbe 3 (°C) closed clam temperatureViral isolation (TCID 50% ml−1)
1st clam opening1st clam opening00:03:30 ± 00:00:2466·55 ± 23·1603635·28 ± 11·2945·95 ± 14·87753/
T 0 50% clams opening00:07:45 ± 00:00:0797·44 ± 3·13850894·47 ± 3·5997·99 ± 0·98845Absence of live virus
T 1 2 min after T000:10:00 ± 00:00:07100·59 ± 0·15874599·98 ± 0·29100·16 ± 0·123423Absence of live virus
T 2 2 min after T100:12:19 ± 00:00:07100·00 ± 0·880928100·15 ± 0·12100·23 ± 0·120554Absence of live virus
T 3 2 min after T200:14:38 ± 00:00:07100·79 ± 0·332916100·23 ± 0·07100·33 ± 0·095044Absence of live virus
Figure 1.

Observed time and temperature values plotted on a graph during one of the three replications of the experiment to cook infected Manila clams in a conventional pan. (a) Time the first clam opened up; (b) opening time of 50% of clams (T0); (c) end of the experiment. Environmental probe (continuous line); probe in the open clams (mixed line); probe in the closed clam (dotted line).

During the preliminary test, however, samples collected at T0 (corresponding to the exact valve-opening time) showed an average viral titre of 102·74 TCID 50% ml−1 (range 102·55–103·05 TCID 50% ml−1); viral identity was further confirmed by EM and real-time PCR. Nonetheless, samples collected at time points T1, T2 and T3 yielded negative results in all cases.

Detailed results are shown in Tables 1 and 2.

Table 2. Results from the preliminary cooking experiment. A total of 36 clams were used in this experiment. In T0, each clam was collected individually at the time of valve opening (for a total of nine clams); in T1, T2, T3 groups, nine clams were collected at the same time. This experiment was performed once. Virological analysis was performed on a pool of three clams
Time pointDescriptionTime (hh:mm:ss)Probe 1 (°C) environment temperatureProbe 2 (C) open clam temperatureProbe 3 (°C) closed clam temperatureViral isolation (TCID 50% ml−1)
T 0 1st clam opening00:03:1465·0531·4626·02103·05
T 0 2nd clam opening00:03:2367·9232·5228·52
T 0 3rd clam opening00:04:4876·2250·5762·13
T 0 4th clam opening00:04:4876·2250·5762·13102·75
T 0 5th clam opening00:06:0188·0475·1182·42
T 0 6th clam opening00:06:0488·1475·9683·12
T 0 7th clam opening00:06:4595·2385·3587·24102·55
T 0 8th clam opening00:07:2896·9092·9891·16
T 0 9th clam opening00:07:3197·5193·4791·44
T 1 50% clams opening00:09:31100·62100·1799·77Absence of live virus
T 2 2 min after T100:11:49100·71100·33100·23Absence of live virus
T 3 2 min after T200:14:06100·97100·36100·34Absence of live virus


Human gastroenteritis carries a high social and economic burden, and there has recently been an increase in the forms caused by HuNoV (Le Guyader et al. 2006; Fonager et al. 2013). Due to their biological behaviour, shellfish can frequently be contaminated by HuNoVs, particularly in environments as lagoons and coastal areas. Suffredini et al. (2011, 2012) and Pavoni et al. 2013 have reported the presence of HuNoV in clams and mussels in different parts of Italy, as Campania, Sardinia, Sicily and the Venetian lagoon. Given the high level of contamination of Italian shellfish, it is paramount to carefully assess the effects of time and temperature of traditional cooking methods on the survival rate of HuNoV. Unfortunately, HuNoV does not grow in cell culture and so MNV was employed as a surrogate to mimic natural conditions.

According to EU legislation (Regulations [EC] Nos. 854/2004, 853/2004 and 2073/2005) on heavy metals, biotoxic algae and bacterial load (E. coli and Salmonella sp), bivalve molluscs from class B areas must undergo purification. Conversely, the presence of enteric viruses is not regulated by any legislation, although it is common knowledge that the depurative period is not sufficient to completely eliminate all possible viruses (Schwab et al. 1998; Ueki et al. 2007; Savini et al. 2009; Terregino et al. 2012. Accordingly, the risk of contracting viral infection from consumption of raw or undercooked molluscs remains even after purification. To provide producers and consumers with clear cooking instructions, this study investigated the effects of the time and temperature, traditionally applied, on the survival rate of MNV when cooking Manila clams.

Results obtained in triplicate showed that cooking large clams in a pan for 7 min and 45 s (±00:00:07 min) to an environment temperature of 97·44°C (±3·138508) is an effective method to completely devitalize the virus. In our experiment, 50% of the clam valves opened up at these times. However, during the preliminary cooking test, in which each individual clam was collected while the valves were opening, a residual amount of virus was still present. This was probably due to the low internal temperature (average internal temperature <70°C) at which the first clam valves opened up. It should be stressed that the experimental cooking conditions were chosen to simulate the worst case scenario for temperature transmission: that is, the clams were cooked without adding water or oil (a normal practice in home cooking) to slow down heat transmission and avoid reaching overly high cooking temperatures. In any event, the virus titre detected when the clams first opened was lower than the one recorded in the untreated control group, suggesting temperature-dependent inactivation. Complete virus inactivation was achieved in all tests when both the internal and external temperature of the clam was close to 100°C.

These results are consistent with field data that indicate that heat inactivation processes (90°C for 90 s) applied to shellfish products in processing establishments in the UK are effective in inactivating HuNoV (as shown by a decrease in human illness resulting from consumption of these products) (Richards et al. 2010). Our results also confirm that while MNV was totally inactivated (5-log10) in PBS after 3 min at 72°C (Wolf et al. 2009), longer times and higher temperatures are required for the whole mollusc. This was further demonstrated in a recent work by Sow et al. 2011, in which MNV was completely inactivated in soft-shell clams subjected to an internal temperature of 90·8°C for 180 s (3 min).

As indicated in the ‘Codex Alimentarius Commission’ guidelines (Codex Alimentarius Commission, GL 79–2012), ‘Heat treatments of bivalve molluscs should be validated for their ability to inactivate viruses. An internal temperature of 85–90°C for at least 90 s is considered to be a virucidal treatment […]’. Even though cooking temperatures typically used by consumers may not achieve 90°C for at least 90 s and thus ensure inactivation of viruses, any cooking would reduce viral levels and depending on the initial level of contamination possibly would reduce the risk of causing foodborne infection in viruses protected within shellfish tissues, as corroborated by our own study.

At the present, it is highly recommended to label all lagoon products as ‘requiring cooking before consumption’, but without specifying for how long and at what temperature (EFSA 2011). The results of this investigation show that to completely inactivate a virus, thereby preventing the risk of foodborne diseases, it is sufficient to cook clams in a pan for 10 min, for at least two min at a temperature of c. 100°C (e.g. boiling the clam cooking liquid for 2 min), during which time the valves of the majority of clams will have opened up. This information should be reported on the label of mollusc products at the fish market to give consumers instructions on the correct clam cooking.


Authors acknowledge Mrs. Francesca Ellero for manuscript editing.

Conflict of interest

All authors have obtained permission from their employer or institution to publish, and there are no conflict of interests to report.