|Home | About | Journals | Submit | Contact Us | Français|
Alphatronlike (genogroup IV [GIV]) noroviruses (NoVs) have been recently identified in carnivores. By screening a collection of 183 fecal samples collected during 2007 from dogs with enteric signs, the overall NoV prevalence was found to be 2.2% (4/183). A unique strain, Bari/91/07/ITA, resembled GIV.2 NoVs in its ORF1 (polymerase complex), while it was genetically unrelated in its full-length ORF2 (capsid gene) to GIV animal and human NoVs (54.0 to 54.4% amino acid identity) and to any other NoV genogroup (<54.7% amino acid identity). It displayed the highest identity (58.1% amino acid identity) to unclassified human strain Chiba/040502/04/Jp. Interestingly, the very 5′ end of ORF2 of the canine virus matched short noroviral sequences (88.9% nucleotide identity and 98.9% amino acid identity) identified from oysters in Japan, indicating that similar viruses may be common environmental contaminants.
The Caliciviridae are small nonenveloped viruses approximately 35 nm in diameter with single-stranded, positive-polarity RNA genomes of 7.4 to 8.3 kb (16). The family includes the genera Vesivirus, Lagovirus, Norovirus, and Sapovirus and unassigned recently identified bovine and simian caliciviruses (14, 17, 43). Noroviruses (NoVs) are considered important gastroenteric pathogens in humans of all age groups (16). The viruses are highly contagious and are transmitted by direct contact or by contaminated water and food, such as raw shellfish (4, 22, 24, 26). Based on full-length sequence analysis of the VP2-coding gene (ORF2), NoVs are classified into genogroup I (GI) to GV. Strains within the same genotype (or cluster) share >85% amino acid identity, while strains of different genotypes within the same genogroup share 55 to 85% amino acid identity (47). Human NoVs are GI, GII, and GIV (45). In addition, NoVs classified in GII and GIII have been detected in pigs and cows (36, 44, 46), and a NoV proposed as GV has been detected in mice (23).
Caliciviruses have been occasionally identified in specimens from dogs. Most isolates were found to be feline caliciviruses (6, 12, 13, 30, 40) or remained uncharacterized (41). Another calicivirus, found to be antigenically and genetically unrelated to feline calicivirus, has been tentatively proposed as a “true” canine calicivirus and was included in the Vesivirus genus (31, 32, 39). More recently, we identified a mixed infection by calicivirus and type 2a parvovirus in a dog with gastroenteritis (29). The calicivirus strain GIV.2/Bari/170/07-4/ITA was found to resemble a lion NoV strain, GIV.2/Pistoia/387/06/ITA, previously detected in a captive lion cub with enteritis (28), and to a lesser extent Alphatronlike human NoVs (GIV). Amino acid identity between the canine and lion NoVs was 90.1%, while amino acid identity to GIV human NoVs was 68.2% to 69.4% and to non-GIV NoVs was <53.3%. Accordingly, the animal GIV strains were classified as a genotype (GIV.2) distinct from the human GIV NoVs (GIV.1). The detection of NoVs in carnivores has raised interesting questions regarding the diffusion, pathogenic potential, and, given the interactions between pets and humans, the zoonotic potential and possible association with the evolution of human NoVs.
In order to assess the epidemiology of NoVs in dogs, a surveillance study was initiated by implementing, with NoV-specific assays, the diagnostic algorithms of all the cases of gastroenteric disease reported to our laboratories during the year 2007. A total of 183 stool samples were collected between January and December 2007 from young dogs (ages of 1 to 6 months) with signs of mild to severe gastroenteritis. All fecal samples were screened for the presence of common canine viral pathogens by either gel-based PCR or quantitative PCR and reverse transcription-PCR (7, 8, 9, 10, 11, 15, 19). To assess the presence of NoV RNA, the samples were screened using a broadly reactive primer pair, p289-p290, targeted to highly conserved motifs DYSKWDST and YGDD of the RNA-dependent RNA polymerase (RdRp) region of the polymerase complex (21). This primer pair amplifies a band of 315 bp for NoV and of 330 bp for Vesivirus and Sapovirus. In the samples yielding amplicons of the expected sizes, the presence of NoV was confirmed using norovirus-specific primer pair JV12Y-JV13I (45), which targets the same RdRp region as primer pair p289-p290. Out of 183 samples, 72 (39.3%) contained canine parvovirus type 2 (CPV-2), either alone (51 [28%]) or in a mixed infection (21 [11.5%]). Thirty-five samples (19.1%) contained canine coronavirus RNA, alone (17 [9.2%]) or in conjunction with other viruses (18 [9.9%]). Canine adenovirus type 2, canine respiratory coronavirus, and canine distemper virus were detected sporadically in 2/183 (1%), 1/183 (0.5%), and 1/183 (0.5%) samples, respectively, while rotaviruses were not detected. Along with strain GIV.2/Bari/170/07-4/ITA, an additional three NoV strains were detected (strains Lecce/445/07-A, Lecce/445-07-B, and Bari/91/07/ITA). Three of those four samples also contained CPV-2 but did not contain other viral pathogens (Table (Table1).1). By sequence comparison, in the short RdRp fragment, the viruses Lecce/445/07-A, Lecce/445-07-B, and Bari/91/07/ITA were closely related to the prototype canine NoV strain, GIV.2/Bari/170/07-4/ITA (91.7 to 94.0% nucleotide identity). Overall, the prevalence of NoV in the 2007 sample collection was rather low (2.1%, 4/183). Whether this value reflects the real prevalence of NoV infection remains to be determined, since we did not use specific primers for animal GIV.2 NoVs.
The sequence of a 3.5-kb fragment at the 3′ end of the genome of strain Bari/91/07/ITA (the 3′ end of ORF1, the full-length ORF2, ORF3, and the noncoding region through the poly[A] tail) was determined as described by Wang et al. (46) (Fig. (Fig.1).1). Sequence editing and multiple alignments were performed with the BioEdit software package, version 2.1 (18). Phylogenetic analysis (neighbor joining and the unweighted pair group method using average linkages) with bootstrap analysis (1,000 replicates) and the Kimura two-parameter correction were conducted by using the MEGA software package, version 3.0 (25). In spite of several attempts, we were unable to amplify the 3′ ends of the genomes of strains Lecce/445/07-A and Lecce/445-07-B. A 14-nucleotide (nt) overlap was present in the ORF1-ORF2 junction region of strain Bari/91/07/ITA, as with most described human and animal NoVs. The 3′ partial sequence of ORF1 spanned 808 nt, encoding 268 amino acids at the COOH terminus of the polymerase complex. By BLAST and FASTA analysis, the highest identity was found to be to the canine strain GIV.2/Bari/170/07-4/ITA (91.8% nucleotide identity and 99.3% amino acid identity) and to the lion NoV strain GIV.2/Pistoia/387/06/ITA (85.7% nucleotide identity and 97.0% amino acid identity), and this genetic relatedness appeared clear in the polymerase-based tree (Fig. (Fig.2).2). There was a single-nucleotide overlap between ORF2 and ORF3; also, there was a 131-nt nontranslating region between ORF3 and the poly(A) tail. ORF3 was 837 nt and encoded a 279-amino-acid (aa) polypeptide.
ORF2 was 1,749 nt and contained an open reading frame encoding a capsid protein with a predicted size of 582 aa. For ORF2, by preliminary analysis with BLAST and FASTA, the highest sequence matches (88.9% nucleotide and 98.9% amino acid identities) were found to be three short sequences (281 nt in length) of NoVs detected from oysters in Japan from 2003 to 2005, Yamaguchi/C34/03/JP, Yamaguchi/24B/02/JP, and Yamaguchi/24C/02/JP (GenBank accession numbers AY353927, AB262094, and AB262095, respectively) (34) (Fig. (Fig.3).3). These short sequences overlapped the 5′ end of ORF2. Nucleotide identity to any other NoV sequence in this short fragment was <78.0%, and the corresponding amino acid identity was <85%, with the best match being to strain Chiba/040502/04/Jp (35). This intriguing finding poses some questions. It is tempting to hypothesize that NoVs similar to the canine strain Bari/91/07/ITA are common in other geographical settings and that they can contaminate the coastal environmental and accumulate at detectable levels in bivalve mollusks destined for raw consumption. Contamination of shellfish by animal (porcine and bovine) enteric caliciviruses, alone or in conjunction with human viruses, has been demonstrated in 22% of oysters in United States (5). However, while the impact of sewage pollution on the water environment by livestock may be relevant, especially in the areas of high livestock production, it is difficult to explain the presence of caninelike NoVs in oysters, since sewage pollution, or illegal discharge of canine sewage into harvest areas, is not a feasible hypothesis. A possible explanation for this is that similar viruses are harbored in other animal species or in settled human populations.
In order to follow strictly the outlines of Zheng's classification (47), a pairwise distance matrix was generated by the uncorrected distance method using a 172-sequence alignment without removing the gaps, including also the sequences of GII.18 and GII.19 porcine NoVs and the sequences of GIV.2 canine and lion NoVs. The complete VP1 sequence of the Japanese strain Chiba/040502/04/Jp was available in GenBank and was included. A detailed list of all the NoV strains used in the analysis is provided in Table S1 in the supplemental material. In this analysis, the VP1 protein of canine NoV Bari/91/07/ITA was more related (58.1% amino acid identity) to the human strain Chiba/040502/04/Jp than to GIV human and animal NoVs (54.0 to 54.4% amino acid identity). The viruses Bari/91/07/ITA and Chiba/040502/04/Jp displayed <54.7% and <53.9% amino acid identities, respectively, to the other NoV strains within GI, GII, GIII, GIV, and GV (see Table S2 in the supplemental material), below the proposed intergenogroup cutoff value (55% amino acid identity) (47). Accordingly, the canine virus Bari/91/07/ITA and the human virus Chiba/040502/04/Jp cannot be classified within any existing NoV genogroups. A phylogenetic tree was constructed using the capsid proteins of the 172 selected human and animal NoVs of the various NoV genogroups (GI to GV). In the tree (Fig. (Fig.3),3), the canine calicivirus Bari/91/07/ITA was grouped with the human virus Chiba/040502/04/Jp, intermingled with but distinct from GIV and GII NoVs.
Based on the polymerase- and VP1-based analysis (Fig. (Fig.22 and and3),3), strain Bari/91/07/ITA displayed inconsistencies suggestive of a recombination event. A nucleotide identity plot of the genome of strain Bari/91/07/ITA (from the 3′ end of ORF1 to the poly[A] tail) was compared with those for the canine strain GIV.2/Bari/170/07-4/ITA, the lion strain GIV.2/Pistoia/387/06/ITA, and the human GIV.1 NoVs Fort Lauderdale/560/98/US and Saint Cloud/624/US (Fig. (Fig.1)1) using Simplot (27), with a window size of 200 and a step size of 20 and with gap strip off and Hamming correction on. The canine NoV Bari/91/07/ITA displayed high nucleotide conservation with the strain GIV.2/Bari/170/07-4/ITA throughout ORF1, while after the ORF1-ORF2 junction region, the viruses markedly differed. Accordingly, strain Bari/91/07/ITA displayed the same polymerase type as the GIV.2 animal NoV but a completely different capsid type, and the recombination event appeared to have occurred in the junction region. Along with accumulation of point mutations, recombination is a powerful mechanism driving the evolution of NoVs (38, 42). NoV recombination, notably between genetically related strains, has been described frequently, and the crossover points are usually located at highly conserved sequence stretches, such as the ORF1-ORF2 junction region (1, 2, 20, 38). Recombination between genetically unrelated NoV strains has also been observed, but rarely (33).
In conclusion, the findings of this study demonstrate that (i) NoVs circulate in dogs, (ii) that canine NoVs are genetically heterogeneous, (iii) that humans may be exposed to these newly recognized strains via certain transmission routes, e.g., by consumption of raw bivalve mollusks, since viruses similar to the canine strain Bari/91/07/ITA have been identified in oysters destined for raw consumption in Japan (34), and (iv) that exploring the genetic diversity of animal NoVs is paramount to understand the evolution of human NoVs.
The GenBank accession number of the sequence of the 3.5-kb fragment at the 3′ end of the genome of strain Bari/91/07/ITA is FJ875027.
The work was funded with grants Progetto di Ateneo 2007, Università degli studi di Bari, Studio dei calicivirus animali ed implicazioni zoonosiche; Progetto di Ateneo 2008, Università degli studi di Bari, Infezioni virali del cane a carattere zoonosico; and MIUR/Ateneo 2007, Norovirus nei carnivori: caratterizzazione molecolare, epidemiologia, implicazioni zoonosiche.
The work was founded by the grant “Progetto di Ateneo 2007, Università degli studi di Bari, Studio dei calicivirus animali ed implicazioni zoonosiche.
Published ahead of print on 26 August 2009.
†Supplemental material for this article may be found at http://jvi/asm.org/.