Norwalk-like calicivirus genes in farm animals.

Viruses closely related to Norwalk-like viruses (NLVs) were recently found in stored stool samples from two calves (United Kingdom and Germany) and four pigs (Japan), sparking discussions about the potential for zoonotic transmission. To investigate if NLVs are commonly present in farm animals, pooled stool samples from 100 pig farms, 48 chicken farms, 43 dairy cow herds, and 75 veal calf farms from the Netherlands were assayed by reverse transcription- polymerase chain reaction amplification, using primers specific for the detection of NLVs from humans. NLV RNA was detected in 33 (44%) of the specimens from veal calf farms and two (2%) specimens from pig farms. Our data show that NLV infections until recently thought to be restricted to humans occur often in calves and sometimes in pigs. While zoonotic transmission has not been proven, these findings suggest that calves and pigs may be reservoir hosts of NLVs.

general population, causing sporadic cases and outbreaks. However, occasionally epidemics occur in which most outbreaks are caused by a single genotype (e.g., Lordsdale-like virus in the Netherlands in 1996) (5,11). Hypotheses for the mechanisms behind the emergence of epidemic types range from large-scale foodborne transmission of a single strain to introduction from a nonhuman reservoir. An indication for the latter was a recent report from Japan of the finding of NLV-like sequences in stool specimens from pigs (12). Enteric caliciviruses have also been detected in other animal species (calves, dogs, and chickens) (13)(14)(15)(16). However, except for canine calicivirus (17), these viruses lack definitive sequence evidence linking them to the Caliciviridae (18). Molecular characterization of three calicivirus strains from cattle has recently been reported. The first virus (Tillamook virus; BCV-Bos1), described by Neill et al. (19), caused respiratory symptoms and was phylogenetically related to San Miguel sea lion virus and vesicular exanthema of swine virus, both within the genus vesivirus. Viruses in this genus have a broader host range (20). Two other viruses closely related to GGI NLVs (Jena virus 117/80 and Newbury agent type 2) were detected in feces from newborn calves with diarrhea (21,22). The

Norwalk-Like Calicivirus Genes in Farm Animals
Viruses closely related to Norwalk-like viruses (NLVs) were recently found in stored stool samples from two calves (United Kingdom and Germany) and four pigs (Japan), sparking discussions about the potential for zoonotic transmission. To investigate if NLVs are commonly present in farm animals, pooled stool samples from 100 pig farms, 48 chicken farms, 43 dairy cow herds, and 75 veal calf farms from the Netherlands were assayed by reverse transcription-polymerase chain reaction amplification, using primers specific for the detection of NLVs from humans. NLV RNA was detected in 33 (44%) of the specimens from veal calf farms and two (2%) specimens from pig farms. Our data show that NLV infections-until recently thought to be restricted to humansoccur often in calves and sometimes in pigs. While zoonotic transmission has not been proven, these findings suggest that calves and pigs may be reservoir hosts of NLVs.

Resear Resear Resear
Resear Research ch ch ch ch similarity of the porcine and bovine sequences with those of NLVs found in humans suggests that a reservoir for these human pathogens may exist in farm animals. We studied enteric caliciviruses in recently collected fecal samples from cattle, chicken, and swine in the Netherlands by using polymerase chain reaction (PCR) assays specific for NLVs found in humans. In addition, we analyzed (by sequence analysis) the genetic variation of detected calicivirus strains and compared it with that of prototype GGI and GGII NLV strains.

Fecal Specimens
Stool specimens from animals were collected at farms in the Netherlands as part of ongoing surveillance for potential zoonotic microorganisms associated with gastroenteritis in humans. From October 10, 1998, to April 21, 1999, samples were collected from 3-to 9-month-old fattening pigs at 100 pig farms of 22 to 1600 pigs. From January 1 to April 1, 1998, stool samples were collected from 75 veal calf farms of 38 to 930 calves and 43 dairy cattle herds of 25 to 180 cows. The age of the calves was 1 to 52 weeks (average 12 weeks), and of dairy cattle herds 4 to 6 years. From January 6 to April 21, 1998, samples were collected from < 6-week-old chickens at 48 broiler farms, with 5,000 to 70,000 per farm.

Sampling
The sampling strategy allowed monitoring for pathogens in a large number of animals and detection of microorganisms at farm level with a prevalence of 5% and 95% confidence (23). Fecal samples from calves, pigs, and chickens (20 to 60 per farm) were collected from animals housed in one randomly chosen farm building. Dairy cattle of one farm were regarded as a single herd, and stool samples were collected from randomly selected cows. Until tested, fecal samples were stored at -70°C in 15 g/L of Trypton Soya broth (TSB) (Oxoid CM 129) and 10% glycerol.

Molecular Detection of NLVs by RT-PCR
For extraction of viral RNA, stool samples were resuspended in HBBS Hanks (Gibco BRL, Breda, Netherlands) to a final concentration of approximately 10%. These suspensions were centrifuged at 3,000 x g for 20 minutes, and 100 µl was used for RNA extraction by addition of a high-molarity solution of guanidinium isothiocyanate (GuSCN). Bound RNA was washed and eluted as previously described (24). To reduce the risk for contamination, specimens from different species were analyzed separately, and one negative control sample was included for every two stool specimens. A human stool sample positive for NLV by electron microscopy was included as positive control.
We used a single-round reverse transcription (RT)-PCR assay with a broadly reactive primer pair, which had been developed for the detection of NLVs in stool specimens from humans (5). For RT, 5 µl RNA was mixed with 4 µl 50 pmol JV13 primer. The solution was heated to 94°C for 2 minutes, cooled, and 6 µl RT buffer was added. The RT reaction was performed in a final volume of 15 µl, consisting of 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 3 mM MgCl 2 , 1 mM each of deoxynucleoside triphosphates (dNTPs), and 5 units of avian myeloblastoma virus RT (Boehringer Mannheim, Almere, Netherlands). The mixture was incubated for 1 hour at 42°C, heated for 5 minutes at 94°C to denature the enzyme, and then cooled. Five µl of the RT mixture was added to the PCR mix, containing 10 mM Tris-HCl (pH 9.2), 75 mM KCl, 1.5 mM MgCl 2 , 0.2 mM dNTPs, 2.5 units AmpliTaq (Perkin Elmer, Nieuwerkerk a/d IJssel, Netherlands), and 15 pmol primer JV12. Mineral oil was added, and 40 amplification cycles, 1 minute at 94°C, 1.5 minute at 37°C, and 1 minute at 74°C each, were performed. The amplification products were analyzed by 2% agarose gel electrophoresis and visualized with UV after ethidium bromide staining.
For Southern blots, the RT-PCR products in the agarose gel were denatured by incubating in 0.5 M NaOH for 30 minutes and transferred to a positively charged nylon membrane (Boehringer, Almere, Netherlands) by vacuum blotting (Millipore, Etten-Leur, Netherlands). Hybridization of NLV RT-PCR products was performed as described previously (5).

Sequence and Phylogenetic Analysis
The NLV RT-PCR products of expected size (326 bp) and 21 products of smaller size of the calf herd samples were excised from a 2% agarose gel and purified with a Qiaquick gel extraction kit (Qiagen, Hilden, Germany). Purified RT-PCR products were sequenced with the BigDye Terminator Cycle Sequencing Ready Reaction  (25).
The confidence values of the internal nodes were calculated by performing 100 bootstrap analyses. Evolutionary trees for nucleotide sequences were drawn by using the Neighbor-joining method.

Electron Microscopy
Electron microscopy was performed as recommended by Flewett (26) and Doane and Anderson (27), with model Philips 400T (Philips, Eindhoven, Netherlands) at 80 kV. Identification of virus particles was based on morphologic criteria, i.e., the size and characteristic surface morphology (28).

Results
Twenty-five (33%) of 75 veal calf farm samples and 2 (2.0%) of 100 pig farm samples showed visible products of the expected size (326 bp). The gel electrophoresis and hybridization results are shown for eight strains (Figure 1). Hybridization with a set of probes used for detection of NLVs in humans yielded no positive reactions ( Figure 1B); therefore, 46 products (25 of the expected size and 21 of a smaller size [ Figure 1A, lane 7]) were sequenced. All 25 RT-PCR products of the expected size contained the GLPSG amino acid motif characteristic of viral RNA polymerases (29). The 21 products of smaller size revealed no viral sequences. The NLV-RT-PCR was negative for all pooled specimens from 43 dairy herds and 48 chicken farms.
By phylogenetic analysis, all calf calicivirus sequences formed a tight cluster closely related to NLVs from GGI with the highest similarity with the Newbury calf calicivirus (Figure 2) (22). Nucleotide sequence identities between the calf NLVs and GGI NLVs were 63% to 70% (75% to 77% amino acids). The swine virus sequences clustered as a separate lineage within GGII NLVs, with 69% to 71% nucleotide and 79% to 83% amino acid identity. In addition, the swine sequences strongly resemble (94% nucleotides, 100% amino acids) the swine calicivirus sequences from Japan. Partial polymerase sequences of Bo/NLV/176/1998/NET and Sw/ NLV/34/1998/NET have been assigned GenBank accession numbers AF194183 and AF194184, respectively.
Since the calf and pig strains did not hybridize to the probe mixture for detection of NLVs from humans, specific probes were developed based on the consensus sequence of RT-PCR products of 10 calf herd samples and on the pig farm samples, respectively. The probe sequences were: RH1(calf) (5'-GGATGTGGTGCAGGCAA AC-3') and RH2(pig) (5'-TCCGCATCTCTATCGT GG-3'). Hybridizations with the calf probe (RH1) confirmed all RT-PCR-positive calf samples after 15 minutes exposure to an ECL hyperfilm ( Figure 1C). Overnight exposure to ECL hyperfilm revealed eight more positives, for 33 (44%) positive calf farm samples.
All veal calf farm samples were screened by electron microscopy for viruses (Table). Particles with NLV morphologic features were found in only one specimen (Figure 3), in which the presence of calicivirus sequence was confirmed by RT-PCR and hybridization (Table).

Conclusions
Until recently, humans were considered the sole host of NLVs. Recent studies in Japan and the United Kingdom, however, demonstrated calicivirus sequences in the caecum of pigs and in stored calf stool samples containing caliciviruslike particles by electron microscopy (12,21,22). Molecular characterization of calf enteric caliciviruses, named Newbury agent (22) and Jena virus (21), revealed that they were genetically related and more closely associated with GGI NLVs than with any other known calicivirus. In this study, we detected NLV nucleotide sequences in 33 (44%) of 75 of the pooled samples from veal calf farms.
Phylogenetic analysis of the detected sequences showed that they formed a tight cluster with the Newbury sequence (22) and were closely related to NLVs from GGI. Bridger et al. (30) demonstrated that calves can be infected with Newbury calicivirus. The detection of Newbury-like NLVs in pooled veal calf samples of 33 farms from different regions suggests that the species is a normal NLV host. Similarly, the close genetic relationship between calicivirus sequences detected in the pig samples in our study and those from Japan suggests that these viruses commonly infect pigs. The close similarity of porcine and bovine sequences to the NLVs infecting humans indicates the possibility of an animal reservoir for human infection.   However, these findings by no means prove zoonotic transmission. Genetically related viruses are commonly found in different species without resulting in apparent widespread interspecies transmission (e.g., rotavirus). On the other hand, zoonotic transmission cannot be excluded. The genetic distances between the animal and human NLVs are similar to the distances between the GGI and GGII strains, and epidemic spread of some NLV strains (5) led to the widespread-possibly global-emergence of a single predominant strain in the human population in 1995-96 (11). This epidemiologic observation resembles early descriptions of the vesicular exanthema of swine virus (VESV) epidemics (31). These viruses, belonging to another genus within the Caliciviridae, the vesiviruses, are known for their broad host range (20). The occasional epidemic spread of VESV was later linked to introduction of caliciviruses from an ocean reservoir (20). Therefore, interspecies transmission of NLVs is possible; the occasional widespread NLV epidemics caused by a single strain may result from introductions of new NLV strains from an animal reservoir. Reynolds et al. reported electron microscopy detection of caliciviruses in calf diarrhea outbreaks (32), and the bovine caliciviruses Newbury agent (22) and Jena virus (21) were pathogenic for calves under experimental conditions and in field studies (33). In our study, we did not record disease or death on the surveyed farms and therefore cannot determine whether these caliciviruses were associated with disease. Despite repeated attempts, experimental infection of different animal species with the prototype Norwalk virus has not been successful, with the exception of infection in chimpanzees (3). Our findings may lead to the development of an animal model for the NLVs of humans; for example, a pig model would enable studies of (mucosal) immunity following NLV infection (1).
The low prevalence of Norwalk-like calicivirus in swine farms (2 per 100) is in agreement with the findings reported from Japan (12). Electron microscopy confirmed the calicivirus origin of the detected NLV sequences in only one calf sample. This low number of positives by electron microscopy was not surprising, given the greater sensitivity of RT-PCR and the use of pooled specimens for screening. NLVs in humans typically are shed at relatively low levels.
The absence of NLV sequences in all 43 dairy herd specimens may be explained by use of an inappropriate primer pair, which was optimized for detection of NLVs of humans (4) or by an acquired strain-specific immunity in older animals that prevents repeated infections. A third possibility is that levels of shedding in adult animals are very low, precluding virus detection in pooled specimens. The swine farm samples were all from fattening pigs at least 3 months old. We may have found a higher prevalence in recently weaned pigs, as the prevalence of other enteric pathogens is highest in young piglets (34).
Our findings raise important questions about the host range of NLVs. It is unclear if animal NLVs form genetically distinct stable lineages or are part of a common pool of viruses circulating between animals and humans. If calves or swine indeed are reservoirs, the prevalence of calicivirus should be determined in both healthy and diseased cattle and swine and methods to detect such interspecies transmissions at an early stage should be developed, in collaborative research by public health and veterinary sciences.