An Efficient and Rapid Method for Recovery of Norovirus from Food Associated with Outbreaks of Gastroenteritis

Transcription

1 504 Journal of Food Protection, Vol. 70, No. 2, 2007, Pages Copyright, International Association for Food Protection Research Note An Efficient and Rapid Method for Recovery of Norovirus from Food Associated with Outbreaks of Gastroenteritis INGEBORG L. A. BOXMAN, 1 * JEROEN J. H. C. TILBURG, 1 NATHALIE A. J. M. TE LOEKE, 1 HARRY VENNEMA, 2 ENNE DE BOER, 1 AND MARION KOOPMANS 2 1 Food and Consumer Product Safety Authority, Regionale dienst Oost, P.O. Box 202, 7200 AE, Zutphen, The Netherlands; and 2 Diagnostic Laboratory for Infectious Diseases and Perinatal Screening, National Institute for Public Health and the Environment, P.O. Box 1, 3720 BA, Bilthoven, The Netherlands MS : Received 30 March 2006/Accepted 10 July 2006 ABSTRACT Noroviruses have emerged as the most common cause of foodborne outbreaks of acute nonbacterial gastroenteritis. In this study, two methods for the extraction of viruses from deli ham were compared. Using both methods, as little as 1 to 10 reverse transcription (RT) PCR of inoculated norovirus and enterovirus could be detected by nested RT-PCR assays. The fastest and most efficient extraction method based on TRIzol LS Reagent was chosen to identify viruses in food items associated with three different outbreaks. Norovirus was detected using nested (real time) RT-PCR assays that target the genome region routinely used for diagnosis of human cases, thereby facilitating the comparison of sequences detected in food and clinical specimens. For one outbreak, a norovirus sequence (163/163 nucleotides) identical to those detected in clinical samples was found on salami sliced by a food handler with a recent history of gastroenteritis. For the other two outbreaks, norovirus was detected on leftovers of spareribs and ham, but fecal samples from affected persons were not available. The methods used in this study may be useful in future outbreak investigations because the extraction method is easy to perform and suitable for this particular type of food and the detection method facilitates direct comparison of patient and food data. Noroviruses, formerly known as Norwalk-like viruses, have emerged as the most common cause of outbreaks and sporadic cases of acute nonbacterial gastroenteritis in children and adults. These viruses are easily transmitted from person to person by the fecal-oral route, either directly or indirectly via contaminated surfaces, food, or water (11). Outbreaks of foodborne disease have been associated with the consumption of all kinds of foods (16, 19). Occasionally, cases have been linked to contaminated food items by identification of identical noroviral sequences (3 5, 14, 20). Most of the studies on viruses recovered from foods have focused on molluscan shellfish. Only a few methods, which include multiple steps and involve various reagents, have been tested on nonseafood (1, 6, 7, 10, 13, 17, 18). Some of these methods involve virus elution using a (glycine) buffer with high ph (9.0 to 9.5), whereas other methods involve rinsing with thiocyanate compounds, which directly releases viral RNA from the virus particles present. In this study, two methods were compared for the recovery of enteric viruses from inoculated deli meat, using ham as a representative food item. Subsequently, the method based on rinsing with thiocyanate compounds was applied to recover viruses from the food items associated with three outbreaks of gastroenteritis. The extraction method applied to deli meat samples was rapid, easy to perform, * Author for correspondence. Tel: ; Fax: ; ingeborg.boxman@vwa.nl. and very sensitive. This method was successfully applied previously for virus detection from shellfish (3). For the detection of norovirus, seminested reverse transcription (RT) PCR assays and a nested real-time PCR assay were used to facilitate direct linking of patient data and food data. These assays target the genome region (JV12Y/ JV13I) that is routinely used for diagnosis of norovirus infections in humans (21). MATERIALS AND METHODS Viral inoculation onto deli ham. A stool sample containing GGII.4 norovirus diluted to a 10% suspension in phosphate-buffered saline and a virus stock containing poliovirus Sabin type 1 (American Type Culture Collection, Manassas, Va.) were used for inoculation experiments. Titers of these virus stocks were determined by PCR unit end-point dilution using seminested RT-PCR assays for noroviruses (3) or enteroviruses (8). Ten-gram portions of deli ham were artificially contaminated with 10 l of serially diluted norovirus and 10 l of serially diluted poliovirus. After an incubation period of 20 min at room temperature, the food items were processed for virus extraction. Each virus extraction experiment (n 4) included six samples: five samples were inoculated with one of the serial dilutions and one sample was left untreated. Sample processing for the recovery of viruses from deli ham. Initially, two methods for recovery of viruses from deli ham were examined. Method 1 (C2 method) was developed from the method of Schwab et al. (18) with modifications (3). Treated and untreated ham (10-g samples) was rinsed with 8 ml of TRIzol LS

2 J. Food Prot., Vol. 70, No. 2 METHOD FOR VIRUS EXTRACTION FROM DELI MEAT 505 Reagent (Invitrogen, Life Technologies, Breda, The Netherlands) for 20 min by shaking at room temperature. After centrifugation of the rinsing solution at 8,000 g for 20 min at 4 C, the aqueous phase was collected and stored at 20 C until RNA extraction was performed. The remainder of the method 1 recovery procedure was identical to that of method 2 (E2 method) following virus recovery from the food. Method 2 was developed from the method of Lewis and Metcalf (15) with modifications. Treated and untreated ham (10-g samples) was rinsed with 50 ml of elution buffer (100 mm Tris and 50 mm glycine, ph 9.5) for 20 min. Thirty milliliters of the rinsing fluid was collected in a 50-ml tube, and the ph was set at 7.2 to 7.4. Virus was precipitated by adding polyethylene glycol 6000 and NaCl to final concentrations of 10% and 0.3 M, respectively, and incubating the mixture for 2 h at 4 C. The samples were centrifuged at 7,000 g for 30 min at 4 C, the supernatant was discarded, and the pellet was suspended in 15 ml of 0.15 M Na 2 HPO 4, ph 9.15 for 20 min by vortexing and shaking. After clarification by centrifugation at 8,000 g for 20 min at 4 C, the supernatant was collected into a new tube, and the ph was readjusted to 7.2 to 7.4. Virus particles were then concentrated by overnight ultracentrifugation at 3,000 g at 4 C using a Centriplusfilter (YM-10, Amicon, Beverly, Mass.). The concentrated virus was collected into a new tube by low-speed centrifugation at 2,000 g at 4 C for 4 min, and the virus extract was stored at 20 C until RNA extraction. RNA was extracted from 100 l of the aqueous phase obtained from method 1 or of the virus extract obtained from method 2 using a silica-based nucleic acid extraction method (2) as described previously (3) and stored at 80 C until further analysis. Outbreak descriptions: outbreak 1 (take-out food). A family of six persons consumed meals from an Egyptian take-out restaurant located in the northwestern region of The Netherlands. The father and daughter-in-law both ate kebabs in garlic sauce with vegetables. The father also consumed spareribs. The four sons (ages 1, 4, 7, and 18 years) consumed chicken kebabs in garlic sauce. All developed illness (3 January 2004) with frequent vomiting, diarrhea, abdominal pain, and nausea at 16 to 30 h after consumption. All patients had consumed different meals during the days before this outbreak. No other health complaints were reported by guests of the restaurant to the local Public Health Service. The local Public Health Service collected five fecal samples. For microbiological investigation of food samples, inspectors of the Food and Consumer Product Safety Authority collected two leftovers (spareribs and kebabs in garlic sauce) that had been stored in the refrigerator at the family s home. Outbreak descriptions: outbreak 2 (buffet at town hall). More than 40 people fell ill about 24 to 40 h after visiting an open house at a town hall located in the northern region of The Netherlands (16 February 2004). They suffered from diarrhea and vomiting and some had headache and moderate fever. During the event, vegetable soup and tomato soup with bread rolls were served. The food handler had been ill 6 days before the event. He recovered after 2 days, but his family members had been ill a few (1 to 4) days before the open-house event. The local Public Health Service collected four fecal samples from patients. The food handler was not willing to provide a fecal sample for norovirus testing. Inspectors of the Food and Consumer Product Safety Authority took eight food samples from the kitchen of the caterer for microbiological investigation, including roasted fillet of chicken, rolled meat, and blocks of ham, salami, and four different types of cheese that had been used for the preparation of the bread rolls. Outbreak descriptions: outbreak 3 (buffet at a private home). After a buffet that included bread, stuffed eggs and tomatoes, salads, and ham off the bone held at a private home in the southern part of The Netherlands (26 November 2004), 2 of 11 persons developed illness with frequent vomiting, diarrhea, abdominal pain, and nausea at about 48 h postconsumption (28 November). Meanwhile, the leftover ham off the bone was served for a second time without being heated to nine new guests (27 November 2004). All nine guests developed gastroenteritis within 26 h as did a neighbor who ate another part of the leftover ham at his own house. For microbiological investigation, inspectors of the Food and Consumer Product Safety Authority collected the leftovers of the ham off the bone, which had been stored in the refrigerator at the family s home. No stool samples were collected because patients did not agree to an etiological investigation. Bacteriological investigation. A total of nine stool specimens (outbreaks 1 and 2) were tested for bacterial enteropathogens, including Campylobacter, Salmonella, Shigella, and Escherichia coli O157 (9). Parts of all food samples were first examined in conformance with international standards for the presence of Clostridium perfringens (ISO 7937), Listeria monocytogenes (ISO 11290), Staphylococcus aureus (ISO 6888), Bacillus cereus (ISO 21871), and Salmonella (ISO 6579:1993 modified). The remaining parts were stored frozen and analyzed for the presence of norovirus. Virological investigation of outbreak samples: stool specimens. Stool specimens (outbreak 2 only) were examined for noroviruses at separate times and locations from food specimens as previously described (22) using modified primers JV12Y and JV13I (21). No fecal samples associated with the other outbreaks were available for viral analysis because they mistakenly had been thrown away after bacteriological analyses (outbreak 1) or no informed consent had been given (outbreak 3). Virological investigation of outbreak samples: food specimens. Virus particles and virus RNA were extracted from the collected food samples using method 1, and RNA was subjected to seminested RT-PCR assays on a MasterCycler (Eppendorf, Hamburg, Germany) with the NVI and NVII primer sets (3), which target the JV12Y/JV13I region of norovirus (21) in a genogroup-specific manner. Analysis of the amplification products was performed as previously described (3). RNA from food samples obtained during outbreaks 2 and 3 were also tested using a recently developed genogroup II specific nested real-time PCR with an ABI 7000 Sequence Detection System (Applied Biosystems, Nieuwerkerk aan de IJssel, The Netherlands); the amplification products of the JV12Y/JV13I RT-PCR were used as input. For this nested real-time PCR (N2pol), the following primers and probes were combined: two forward primers (NoroGII-FA: 5 -CYTGCACCTCMCAATGGA-3 and NoroGII-FB: 5 -CKTGCACCTCRCAATGGA-3 ), one reverse primer (NoroGII-Rb: 5 -TGTRACTTCAGAGAGYGCACAKA- 3 ), and a mix of three probes (NoroGIIA-P: FAM-CY- ATCGCCCACTGGCTYCTCA-TAMRA; NoroGIIB-P: FAM- CCATYRCCCACTGGCTCCTCA-TAMRA; and NoroGIIC-P: FAM-CTCCATTGCTCATTGGCTTCTCACG-TAMRA) (Applied Biosystems). After completion of the JV12Y/JV13I RT- PCR, 2.5 l of the amplification product was added to 22.5 l of reaction mixture consisting of 250 nm each primer, 125 nm probe, and 12.5 l of Taqman Universal PCR Master Mix (Applied Bio-

3 506 BOXMAN ET AL. J. Food Prot., Vol. 70, No. 2 TABLE 1. Detection of norovirus or poliovirus inoculated onto ham a Method 1 (C2 method) 2 (E2 method) No. of samples positive/no. of samples tested 1,000 RT-PCR 3/3 3/3 100 RT-PCR 1/4 10 RT-PCR 3/4 1 RT-PCR unit 0/4 1/4 a Ten grams of ham was seeded with decreasing numbers (seminested RT-PCR ) of both GGII.4 norovirus () and poliovirus (). After incubation for 20 min, samples were processed by methods 1 and 2, and extracted RNA was analyzed by seminested RT-PCR assays specific for norovirus GGII types or enterovirus. systems). After 2 min at 50 C and an initial 10 min of denaturation at 95 C, amplification was started with 40 cycles of denaturation at 95 C (15 s) and annealing and extension at 58 C (60 s). The NVII primer set (3) was used to sequence samples that tested positive with the N2pol set. Each analysis included negative controls (RNA without template), positive controls (GGI.7 Winchester and GGII.4 Lordsdale), and blanks. Each outbreak was investigated separately. Sequence analysis. Sequencing of positive samples was performed as previously described for clinical samples (22) and food samples (3). Sequences were compared with available sequences in GenBank using the BLAST program of the National Center for Biotechnology Information (Bethesda, Md.). Sequence comparisons were also performed using data of the database of the Foodborne Viruses in Europe network. Unweighted pair group method with arithmetic mean cluster analysis was followed by multiple sequence alignment with BioNumerics (Applied Maths, Sint-Martens-Latem, Belgium) for phylogenetic analysis (global clustering method). Detected sequences from food samples were submitted to GenBank (AY and DQ351906). RESULTS Recovery of norovirus and enterovirus inoculated onto deli ham. Viral titers of the virus stocks used for inoculation experiments were estimated to be 10 7 and 10 8 seminested RT-PCR per ml for norovirus and enterovirus, respectively (n 4). Ham portions, which were inoculated with both norovirus and poliovirus at numbers decreasing from 1,000 to 1 RT-PCR of virus, were processed by methods 1 and 2. Using method 1, norovirus and poliovirus could be detected on ham inoculated with virus counts as low as 1 RT-PCR unit (Table 1). Comparable levels of detection were obtained for norovirus (as few as 10 RT-PCR ) and poliovirus (as few as 1 RT-PCR unit) with the longer method 2. Method 1 was more rapid and more straightforward than method 2. For these reasons, food items associated with outbreaks were processed according method 1. Bacteriological investigation of outbreak samples. None of the fecal samples collected from outbreaks 1 and 2 were positive for Salmonella, Campylobacter, Shigella, or E. coli O157. None of the food samples contained bacterial pathogens that could explain the onset of the outbreaks. Virological investigation: outbreak 1. A GGIIb norovirus (AY604201) was detected only on the sparerib sample with the NVII seminested RT-PCR assay. No stool samples associated with this outbreak were collected for viral analyses. Virological investigation: outbreak 2. On the ham and the salami sample only, a GGII norovirus was detected using the N2pol nested real-time PCR assay and the NVII seminested RT-PCR assay. Threshold cycles of positive samples, representing the cycle at which the fluorescence signal generated during amplification passes above a threshold baseline with no fluorescence observed in the blanks, were Ct 32 and Ct 31, respectively. A GGII norovirus type also was detected in all four fecal samples tested. Sequence analyses revealed the presence of a GGIIb type in both clinical and food samples. Results obtained with a multiple sequence alignment program indicated that the sequence from the salami sample appeared identical (163/163 nucleotides) to sequences in four clinical samples as detected by the JV12Y/JV13I RT-PCR assay. The sequences detected on the ham and salami samples had a one-nucleotide difference. Although extraction and analyses were performed separately, an identical norovirus genotype was detected on the salami (outbreak 2) and on the leftover sparerib (outbreak 1). To determine the significance of finding identical sequences in seemingly unrelated outbreaks, the sequences of the nested amplification products (NVII) derived from the sparerib sample (outbreak 1) and the salami and ham samples (outbreak 2) were compared with GGIIb sequences detected in fecal samples from outbreak 2 (OB ), with GGIIb sequences detected in fecal samples of seven other outbreaks in The Netherlands in the same period of October 2003 to July 2004, and with a consensus sequence for GGIIb. In a 163-bp overlapping region (Fig. 1), the GGIIb sequences obtained from the spareribs and salami were similar to sequences detected in fecal samples obtained from four other outbreaks involving two institutions and two restaurants but diverged from sequences associated with three other outbreaks. Virological investigation: outbreak 3. A GGII norovirus type (DQ351906) was detected on the ham off the bone with the N2pol nested real-time PCR assay (Ct 25) and the NVII seminested RT-PCR assay. No stool samples associated with this outbreak were available for testing, but a multiple sequence alignment program (BioNumerics) revealed a strong similarity (161/163 nucleotides) between the detected sequence from the ham off the bone and the (Dutch) GGII outbreak variant (AY900231) (12). DISCUSSION In the present study, two methods for the recovery of virus from deli meats were compared, and the most straightforward method was applied to food items associated with three different outbreaks of gastroenteritis. The methods ap-

4 J. Food Prot., Vol. 70, No. 2 METHOD FOR VIRUS EXTRACTION FROM DELI MEAT 507 FIGURE 1. Phylogenetic tree based on a 163-bp region of the RNA-dependent RNA polymerase gene (ORF1). GGIIb sequences were derived from suspected food items associated with outbreak 1 (spareribs, AY604201) and outbreak 2 (salami, 6566SC2F; ham, 6515HC2-N2F) or from clinical specimens collected during outbreak 2 (OB ) and during seven other outbreaks in The Netherlands between October 2003 and July A consensus sequence (AIIb) based on multiple GGIIb entries was used as a reference. plied to the samples may be useful in future outbreak investigations. The extraction method is easy to perform, and the detection method facilitates direct comparison of RNA sequences obtained from clinical and food specimens. In this study, a norovirus sequence obtained from food samples was identical to that found in corresponding clinical samples. The three outbreaks described in the present study have at least two points in common. First, norovirus was detected on prepared deli meat in each outbreak, suggesting that the method works well with this particular type of food matrix, as also indicated using artificially contaminated deli meat samples. The TRIzol-based extraction method also was very efficient for recovering viruses from inoculated digestive glands of mussels and oysters (data not shown) and from naturally contaminated shellfish (3). In contrast, we would not recommend the TRIzol method for the recovery of viruses from soft fruit; preliminary experiments with inoculated strawberries indicated poor recovery (data not shown). Second, persons with infectious disease were likely involved in these outbreaks, because meat by itself is not considered a risky product compared with bivalve shellfish or soft fruit. Generally, workers with acute gastroenteritis should be excluded from food establishments until 48 to 72 h after symptoms have resolved. Even after this period, food handlers can be infectious. Infectious children of food handlers also may be a risk factor, as was determined in a comparable outbreak associated with deli meat in Texas in 1998 (5). In outbreak 2, the food handler and his family members had had a recent history of gastroenteritis. Because an identical norovirus sequence was recovered from the salami and the clinical samples, the bread rolls were very likely the source of outbreak 2. Although no stool samples were available for viral analyses for two of the three outbreaks investigated, data on norovirus sequences detected in food specimens should be documented in databases, e.g., the database of the European Food-borne Viruses Network ( uk) for retrospective comparisons. Finding identical norovirus RNA sequences in seemingly unrelated outbreaks could indicate that this virus variant had been circulating in the community at the time of the outbreaks or that outbreaks were related to a common source. These possibilities can be ruled out or definitively supported only if more systematic outbreak investigations include a food history. Investigations of foodborne outbreaks of gastroenteritis often are complicated by the absence of leftover food samples or the unwillingness of patients to cooperate. Successful linking of clinical cases to contaminated products may be enhanced by further education of persons involved in collection of food and clinical samples, possibly leading to the identification of new vehicles of infection and novel patterns of transmission. ACKNOWLEDGMENTS We thank the Public Health Service and inspectors of the Food and Consumer Product Safety Authority for their cooperation in collection of fecal samples and food items. REFERENCES 1. Bidawid, S., J. M. Farber, and S. A. Sattar Rapid concentration and detection of hepatitis A virus from lettuce and strawberries. J. Virol. Methods 88: Boom, R., C. J. A. Sol, M. M. M. Salimans, C. L. Jansen, P. M. E. Wertheim-Dillen, and J. van der Noordaa Rapid and simple method for purification of nucleic acids. J. Clin. Microbiol. 28: Boxman, I. L. A., J. J. H. C. Tilburg, N. A. J. M. te Loeke, H. Vennema, K. Jonker, E. de Boer, and M. Koopmans Detection of noroviruses in shellfish in The Netherlands. Int. J. Food Microbiol. 108: Brown, C. M., J. W. Cann, G. Simons, R. L. Fankhauser, W. Thomas, U. D. Parashar, and M. J. Lewis Outbreak of Norwalk virus in a Caribbean island resort: application of molecular diagnostics to ascertain the vehicle of infection. Epidemiol. Infect. 126: Daniels, N. A., D. A. Bergmire-Sweat, K. J. Schwab, K. A. Hendricks, S. Reddy, S. M. Rowe, R. L. Fankhauser, S. S. Monroe, R. L. Atmar, R. I. Glass, and P. Mead A food-borne outbreak of gastroenteritis associated with Norwalk-like viruses: first molecular trace-back to deli sandwiches contaminated during preparation. J. Infect. Dis. 181: Dubois, E., C. Agier, O. Traoré, C. Hennechart, G. Merle, C. Crucière, and H. Laveran Modified concentration method for the detection of enteric viruses on fruits and vegetables by reverse transcriptase polymerase chain reaction or cell culture. J. Food Prot. 65: Gouvea, V., N. Santos, M. d. C. Timenetsky, and M. K. Estes Identification of Norwalk virus in artificially seeded shellfish and selected foods. J. Virol. Methods 48: Häfliger, D., M. Gilgen, J. Luthy, and P. Hübner Seminested RT-PCR systems for small round structured viruses and detection of enteric viruses in seafood. Int. J. Food Microbiol. 37: Isenberg, H. D Processing and interpretation of bacterial fecal cultures, p In H. D. Isenberg (ed.), Clinical microbiology procedures handbook, vol. 1. American Society for Microbiology, Washington, D.C. 10. Kobayashi, S., K. Natori, N. Takeda, and K. Sakae Immunomagnetic capture RT-PCR for detection of norovirus from foods implicated in a food-borne outbreak. Microbiol. Immunol. 48:

5 508 BOXMAN ET AL. J. Food Prot., Vol. 70, No Koopmans, M., and E. Duizer Foodborne viruses: an emerging problem. Int. J. Food Microbiol. 90: Kroneman, A., H. Vennema, Y. van Duijnhoven, E. Duizer, and M. Koopmans High number of norovirus outbreaks associated with a GII.4 variant in The Netherlands and elsewhere: does this herald a worldwide increase? Eurosurveillance 8(52). Available at: Accessed 23 December Leggitt, P. R., and L. A. Jaykus Detection methods for human enteric viruses in representative foods. J. Food Prot. 63: Le Guyader, F. S., F. H. Neill, E. Dubois, F. Bon, F. Loisy, E. Kohli, M. Pommepuy, and R. L. Atmar A semiquantitative approach to estimate Norwalk-like virus contamination of oysters implicated in an outbreak. Int. J. Food Microbiol. 87: Lewis, G. D., and T. G. Metcalf Polyethylene glycol precipitation for recovery of pathogenic viruses, including hepatitis A virus and human rotavirus, from oyster, water and sediment samples. Appl. Environ. Microbiol. 54: Richards, G. P Enteric virus contamination of foods through industrial practices: a primer on intervention strategies. J. Ind. Microbiol. Biotechnol. 27: Sair, A. I., D. H. D Souza, C. L. Moe, and L. A. Jaykus Improved detection of human enteric viruses in foods by RT-PCR. J. Virol. Methods 100: Schwab, K. J., F. H. Neill, R. L. Fankhauser, N. A. Daniels, S. S. Monroe, D. A. Bergmire-Sweat, M. K. Estes, and R. L. Atmar Development of methods to detect Norwalk-like viruses (NLVs) and hepatitis A virus in delicatessen foods: application to a foodborne NLV outbreak. Appl. Environ. Microbiol. 66: Seymour, I. J., and H. Appleton Foodborne viruses and fresh produce. J. Appl. Microbiol. 91: Shieh, Y. S. C., S. S. Monroe, R. L. Fankhauser, G. W. Langlois, W. Burkhardt III, and R. S. Baric Detection of Norwalk-like virus in shellfish implicated in illness. J. Infect. Dis. 181:S360 S Vennema, H., E. de Bruin, and M. Koopmans Rational optimization of generic primers used for Norwalk-like virus detection by reverse transcriptase polymerase chain reaction. J. Clin. Virol. 25: Vinjé, J., H. Vennema, L. Maunula, C. H. von Bonsdorff, M. Hoehne, E. Schreier, A. Richards, J. Green, D. Brown, S. S. Beard, S. S. Monroe, E. de Bruin, L. Svensson, and M. P. G. Koopmans International collaborative study to compare reverse transcriptase PCR assays for detection and genotyping of noroviruses. J. Clin. Virol. 41: