Molecular characterization of canine astrovirus, vesivirus and circovirus, isolated from diarrheic dogs in Turkey

Abstract

Background:

Canine astrovirus (CAstV) has been considered the primary cause of gastroenteritis in young animals worldwide, while canine vesivirus (CVeV) and canine circovirus (CCiV) are occasionally reported.

Aims:

This study aimed to investigate the existence and molecular characteristics for these three viruses in Turkey. Methods: Faecal samples from 150 shelter dogs with gastrointestinal problems (127 adults and 23 puppies) were collected and examined by reverse transcription-polymerase chain reaction (RT-PCR) analysis based on the partial sequence of RdRp gene (ORF1b) for CAstV, ORF2 gene of CVeV and capsid protein (Cap) and replication associated protein (Rep) gene of CCiV. Randomly selected positive samples were submitted to sequencing and molecular analyses were conducted based on partial sequences.

Results:

It was found that 66% (99/150) of diarrhoeic dogs were positive for CAstV, 3.33% (5/150) for CVeV, and 6% (9/150) for CCiV. Four sub-genotypes for CAstV and two sub-genotypes for CVeV were suggested according to molecular analyses. The phylogenetic relationship of CCiV with other strains obtained from various areas was further demonstrated.

Conclusion:

This study emphasizes the importance of emerging viruses for canids, classification of them and their proportional contribution in gastroenteritis cases. We concluded that astrovirus infection must be considered as the major cause of diarrhea in dogs; However, the prevalences of vesivirus and circovirus were relatively low in cases makes them less important in Turkey.

Introduction

The Astroviridae family consists of approximately 35 nm diameter icosahedral capsids, which comprise a positive, linear, single stranded RNA genome about 6.4-7.4 kb in length. Under the microscope, negative stained preparations show approximately 10% of the viral particles to be as astroviruses (Mendez and Arias, 2013 ▶; MacLachlan et al., 2017 ▶; Zhang et al., 2020 ▶). These viruses are classified under two genera as Mamastrovirus and Avastrovirus, whose genomes have three open reading frames (ORFs) (ORF1a, ORF1b, and ORF2) (Zhu et al., 2011 ▶; Caddy et al., 2015 ▶; Mihalov-Kovacs et al., 2017 ▶). Infectious viral RNA acts as both genomic and viral mRNA (Mendez and Arias, 2013 ▶; MacLachlan et al., 2017 ▶). Astroviruses were first identified in the faeces of children with diarrhea in 1975 (Appleton and Higgins, 1975 ▶; Madeley and Cosgrove, 1975 ▶) and are currently estimated to cause 10% of the gastroenteritis cases in children worldwide (Moser and Schultz-Cherry, 2005 ▶). Canine astroviruses (CAstV) have been identified in many countries of the world, and were first investigated in 1980 in faeces of diarrheic puppies in the USA (Williams, 1980 ▶; Martelle et al., 2011 ▶; Zhu et al., 2011 ▶; Grellet et al., 2012 ▶; Castro et al., 2013 ▶; Choi et al., 2014 ▶).

The host spectrum of caliciviruses comprise both wild and domestic animals, including dogs, cattle, pigs, felines, minks, monkeys, and humans. The family Caliciviridae consists of 11 genera, including Vesivirus, Lagovirus, Norovirus, Sapovirus, and Nebovirus; all of which share specifications such as having 27-40 nm diameter capsids and 7.4-8.3 kb single-stranded, positive-polarity RNA genome (Martella et al., 2015 ▶; Gutierrez-Escolano, 2017 ▶; Renshaw et al., 2018 ▶; Vinjé et al., 2019 ▶). The 5´ end of the viral genome is linked to a VPg protein, while the 3´ end has a poly (A) tail. The genome contains 2-4 ORFs depending on the genus. ORF1, located at the 5´ end, encodes a large polyprotein processed only by the viral protease to produce non-structural (NS) proteins. There are two additional ORFs (ORF2 and ORF3) in noroviruses and vesiviruses encoding major (VP1) and minor (VP2) capsid proteins, respectively (Martin-Alonso et al., 2005 ▶; Gutierrez-Escolano, 2017 ▶; Desselberger, 2019 ▶).

Vesiviruses were first detected in 1932, in domestic pigs with vesicular disease in the USA. Since then, they have been associated with abortion, hepatitis, respiratory disease, diarrhea, myocarditis, mucosal ulceration and vesicular lesions in human and various animal species (Smith et al., 1998 ▶; Smith et al., 2002 ▶; Radford et al., 2007 ▶). In the early 80s, canine vesivirus (CVeV) was isolated from a dog with glossitis for the first time (Evermann et al., 1981 ▶). In the following years gastroenteritis cases associated with this virus were reported worldwide (Evermann et al., 1985 ▶; Shaffer et al., 1985; Gabriel et al., 1996 ▶; Pratelli et al., 2000 ▶; Martella et al., 2002 ▶; Di Martino et al., 2009 ▶; Binn et al., 2018 ▶; Renshaw et al., 2018 ▶).

Circoviridae family contains 2 genera, Circovirus and Cyclovirus (Breitbart et al., 2017 ▶; Rosario et al., 2017 ▶). The virus has a small (15-25 nm diameter) capsid and circular ssDNA genomes. These genomes range from 1.7 to 2.1 kb and contain two main ORFs encoding replication (Rep) and capsid (Cap) associated proteins (Bexton et al., 2015 ▶; Breitbart et al., 2017 ▶). Circoviruses have been described as infectious agents of a group of animals such as birds, pigs, dogs, foxes, and wolves (Kapoor et al., 2012 ▶; Decaro et al., 2014 ▶; Hsu et al., 2016 ▶; Zaccaria et al., 2016 ▶; Dowgier et al., 2017 ▶; Rosario et al., 2017 ▶). Canine circovirus (CCiV) has been found in the blood serum of dogs with hemorrhagic enteritis in the USA (Kapoor et al., 2012 ▶). Since then, it has been reported in USA, Italy, Germany, Brazil, Taiwan, and Thailand (Li et al., 2013 ▶; Decaro et al., 2014 ▶; Hsu et al., 2016 ▶; Gentil et al., 2017 ▶; Piewbang et al., 2018 ▶; Cruz et al., 2020 ▶). A recent study demonstrated that CCiV was strongly related to the development of canine acute gastroenteritis, especially in case of co-infections with other etiological agents (Dowgier et al., 2017 ▶).

A large number of pathogens including CAstV, CVeV, and CCiV are responsible for gastroenteritis in dogs. However, none have been reported or investigated before in Turkey. This study, therefore, aimed to study the presence of these viruses in dogs with gastroenteritis and characterize viruses based on genomic data to understand the genetic relationship between strains isolated throughout the world.

Materials and Methods

Primer design

Oligonucleotides for the conventional polymerase chain reaction (PCR) assay were designed based on publicly available data from the database of The National Center for Biotechnology Information (NCBI). Sequences were aligned using MUSCLE alignment (Edgar et al., 2004 ▶), and the primers were designed using the Geneious Bioinformatics Software Platform (Kearse et al., 2012 ▶). The program was utilized to generate primer sets from well-conserved areas, which were thereby, capable of recognizing the most sequences in database.

Sampling

A total of 150 rectal swab specimens were collected from Sivas Municipality Animal Shelter, which accepts approximately a thousand dogs per year. All of the sampled dogs were manifesting gastro-intestinal problems, including 127 one-year-old adults and 23 two- to four-month-old puppies. Collected samples were transported to the laboratory immediately and stored at -80°C before being subjected to RNA and DNA isolation.

Virus investigation and phylogenetic analysis

Nucleic acid isolation

Faecal samples were diluted 1:10 with 1 M phosphate buffered saline (PBS) and centrifuged 10 min at 2876 g to remove large cellular debris. After centrifugation, supernatants were submitted to the total nucleic acid extraction procedure using a GF-1 Viral Nucleic Acid Extraction Kit (Vivantis Technologies, Malaysia) according to the manufacturer’s instructions. Eluted nucleic acids were stored at -80°C until use.

Reverse transcription (RT)

The cDNA synthesis was carried out in a 20 μL final volume containing 5 μL RNA extract, 10 mM deoxynucleoside triphosphate (dNTP), 2,5 μL 10x RT buffer (50 mM Tris-HCl (pH = 8.3 at 25°C), 75 mM KCl, 3 mM MgCl₂ and 10 mM DTT), 50 ng of the random hexamer, 40 U RNasin, and 200 U M-MuLV RT RNase H (Vivantis, Germany). The RT was performed at 37°C for 1 h.

PCR

The PCR was conducted in a 30 μL final volume using 3 μL of the RT reaction mixture or viral nucleic acid extract as template. The PCR mixture contained 3 μL 10x PCR buffer, 10 mM dNTP, 10 pmol/μL of each sense/antisense primer, and 5 U of Taq DNA polymerase (Vivantis, Germany).

Molecular detection of partial RdRp gene of CAstV, partial ORF2 gene of CVeV, partial ORF1 (Rep) gene and ORF2 (Cap) gene of CCiV were conducted by primer sets (Table 1). PCR conditions were adjusted as follows: 95°C for 2 min for pre-denaturaton, during 40 cycles, 94°C for 30 s denaturation, 55°C (CAstV) or 56°C (CVeV) or 51°C (both ORF1 and ORF2 of CCiV) for 30 s annealing, 72°C for 45 s for extension and lastly, 10 min for 72°C for the final extension.

Table 1.

Primers used in the study. Positions of the forward and reverse primer sets were indicated based on reference sequences (CAstV, NC_026814; CVeV, NC_004542; CCiV, NC_020904.1)

Primer name	Sequence (5´-3´)	Position	Target gene	Amplicon size
CAstV-3484F CAstV-3777R	GYACTATACCRTCTGATTTAATT AGACYAARGTGTCATAGTTCAG	3469-3491 3741-3762	RdRp gene (ORF1b) of canine astrovirus	294 bp
CVeV-6193F CVeV-6542R	ACCGMTGYCTTATGGCTGTGG CCAYCCWGTGTACATCTTSGC	6193-6213 6522-6542	ORF2 gene of canine vesivirus	359 bp
CCiV-1036F CCiV-1285R	CCCCCTTCGAGGCTGTWTATT AGGRGCTAACATGGTMTGGA	1035-1055 1265-1284	Capsid protein (Cap) gene of canine circovirus	250 bp
CCiV-241F CCiV-524R	GGTGGYCGCGGMCATTTTG ACBTBCACKTCCGTCTTCCA	241-259 505-524	Replication associated protein (Rep) gene of canine circovirus	284 bp

CAstV: Canine astrovirus, CVeV: Canine vesivirus, and CCiV: Canine circovirus

Sequencing and phylogenetic analysis

The PCR amplicons were purifed with Wizard SV Gel and a PCR Clean-Up System (Promega, Madison,WI) and were sequenced using the BigDye Terminator Cycle Sequencing Kit (Applied Biosystems, Foster City, CA) on an automated sequencer (ABI 3100; Applied Biosystems, Foster City, CA). All sequenced products were used to obtain phylogenetic data. Partial sequences were compared with other sequence data provided online by NCBI. Sequence alignment and phylogenetic analysis based on partial nucleotide sequences of 294 bp RdRp gene (canine astrovirus), 359 bp ORF2 gene (canine vesivirus), 250 bp ORF1 (Rep) and 284 bp ORF2 (Cap) genes (canine circovirus) were constructed using Geneious software (Kearse et al., 2012 ▶). The neighbour-joining (NJ) method and the Tamura-Nei genetic distance model (Tamura and Nei, 1993 ▶) were used to build the tree. The trees were drawn to scale, where branch lengths were measured based on the number of substitutions per site and bootstrapped with 1000 replicates.

Results

Molecular detections were conducted based on 294 bp partial RdRp gene (ORF1b) of the CAstV genome, 359 bp partial ORF2 gene of the CVeV genome, 284 bp partial Rep gene and 250 bp Cap gene of the CCiV. For CCiV, all positive samples gave exact bands for both Cap and Rep genes. Polymerase chain reaction assay showed that 82 samples for CAstV (64.57%), 4 samples for CVeV (3.15%) and 8 samples for CCiV based on both Rep and Cap genes (6.30%) were positive within the 127 adult dogs. For the puppies, 17 samples (73,91%) for CAstV, one in each sample (4.35%) for both CVeV and CCiV, were detected from faecal samples (Table 2).

Table 2.

Overall PCR results for detecting CAstV, CveV, and CCiV from adults and puppies

Age	Total Samples	CAstV	CVeV	CciV	Dual infection with CAstV and CVeV	Dual infection with CAstV and CCiV	Dual infection with CVeV and CCiV	Triple infection
Adults	127	64.57% (82/127)	3.15% (4/127)	6.30% (8/127)	1.58% (2/127)	5.51% (7/127)	0.79% (1/127)	0.79% (1/127)
Puppies	23	73.91% (17/23)	4.35% (1/23)	4.35% (1/23)	4.35% (1/23)	4.35% (1/23)	4.35% (1/23)	4.35% (1/23)
Overal	150	66.00% (99/150)	3.33% (5/150)	6.00% (9/150)	2% (3/150)	5.33% (8/150)	2% (3/150)	2% (3/150)

PCR: Polymrase chain reaction, CAstV: Canine astrovirus, CVeV: Canine vesivirus, and CCiV: Canine circovirus, x/y: Positive/total sample ratio

Ten amplicons of the 294 bp partial RdRp gene region of CAstV were randomly selected and sequenced from positive samples (GenBank Accession No.: MK507563.1-MK507572.1). Partial sequence data were compared with each other and the available nucleotide sequence data in GenBank. According to phylogenetic analyses based on nucleotide alignment, there were four distinct clades (Fig. 1). All Turkish isolates were part of clade 1, which also included the RefSeq strain (Gillingham/2012/UK). Clades (2, 3 and 4) were named according to the distance of the nucleotide identity from clade 1. The identity of clade 1 strains varies from 91.40 to 100.00%. Likewise, the nucleotide identity of Turkish CAstV strains varied from 94.22% to 100%. In comparison with clade 1 and 2, the identity level dramatically reduced between 73.21% to 80.08%, while clade 2 strains varied from 91.76 to 97.63%. Between clade 1 and 3, the identity level was varied from 68.20% to 73.53%, whereas clade 3 strains were 95.73% within the group. Lastly, clade 4 strains showed an identity between 91.80% to 100.00% within the group, and the nucleotide identity was between 58.04% to 63.16% when compared to clade 1.

Fig. 1.

Cladogram representing the consensus (1000 replicates) neighbour-joining phylogenic tree of canine astroviruses based on 294 bp partial RdRp gene sequences. Tree construction was built following the Tamura-Nei (1993) genetic distances model using Geneious Prime software version 2019.2.3 (available at https://www.geneious.com, Accessed Oct.11.2019). Novel strains are illustrated in bold text

From the positive samples, two 359 bp length sequence data were obtained from the partial ORF2 gene of CVeV and deposited to the GeneBank database (Acc. No.: MK783212.1 and MK783213.1). These data were compared with the other available CVeV sequences according to nucleotide identity level. The results showed that these two strains were segregated into two distinct clades (Fig. 2). The identity percentage between clade 1 and 2 varied between 67.04% to 71.75%. Clade 1 strains showed 84.59% to 100.00% identity within the group, whereas clade 2 strains were between 87.74% to 99.16%.

Fig. 2.

Cladogram representing the consensus (1000 replicates) neighbour-joining phylogenic tree of canine vesiviruses based on 359 bp partial ORF2 gene sequences. Tree construction was built following the Tamura-Nei (1993) genetic distances model using Geneious Prime software version 2019.2.3 (available at https://www.geneious.com, Accessed Oct.11.2019). Novel strains are illustrated in bold text

Phylogenetic analysis demonstrated that all Turkish CCiV strains (GenBank Acc. No.: MK783214.1-MK783223.1) were clustered with a Chinese CD17/2016 (MG266899.1), three USA 214 (JQ821392.1), UCD2-32162 (KC241984.1), and OH19098-1 (MF457592.1), two German Ha3 (KF887949.1) and FUBerlin-JRS (KT283604.1) strains which were isolated from dogs and 18 Italian strains, isolated from dogs, wolves and badgers, based on 250 bp partial cap gene of CCiV. Similar results were observed in the phylogenetic tree based on the replication associated protein (Rep) gene of CCiV with the contribution of five more Chinese strains including COS8/2016, COS2/2016, CDX2/2017, COM7/ 2016, and CDX8/2017 (MG266900-MG266904) in the same branch (Figs. 3A and B).

Fig. 3.

Trees representing the consensus (1000 replicates) neighbour-joining phylogenic tree of canine circoviruses based on (A) 250 bp partial capsid protein (Cap) gene and (B) 284 bp replication associated protein (Rep) gene of CCiV strains. Tree construction was built following the Tamura-Nei (1993) genetic distances model using Geneious Prime software version 2019.2.3 (available at https://www.geneious.com, Accessed Oct.11.2019). Novel strains are illustrated in red color

Moreover, the percentage of nucleotide identity of the novel strains ranged between 97.54% to 99.65% within the group, while the rest of the members in the clade were between 94.01% to 97.54%. On the other hand, the similarity between Turkish strains and the strains being drawn in the separate branches were 79.93% to 91.20%. While, similarity levels based on the Cap gene of CCiV seemed to be a little confusing, the tree substitution looked like a counterpart of the Rep gene. Nucleotide similarity of the novel strains varied from 97.20% to 100% based on the Cap gene comparison in the branches from 95.60% to 100%. On the other hand, similarity of Turkish strains were 83.60% to 92.40% compared to other branches. The most distant strains to our isolates were the Chinese group (MG279119 to MG279140) with similarity levels downs to 83.60%.

Discussion

The main goals of this study were to show the common circulation of CAstV, CVeV, and CCiV between dog populations in Turkey, and reveal molecular characterizations of the identified sequences. In our study, CAstV was found to be 66% (99/150). This ratio was 64.57% (82/127) in adults and 73.91% (17/23) in puppies. The presence of CAstV has been reported in countries such as China, Korea, England, Italy, the USA and Hungary. In these studies, the positivity rates of CAstV in diarrheic dogs ranged from 2.2% to 26.9% (Williams, 1980 ▶; Martella et al., 2011 ▶; Zhu et al., 2011 ▶; Choi et al., 2014 ▶; Caddy et al., 2015 ▶; Mihalov-Kovacs et al., 2017 ▶). Although there was no statistically significant difference in the positivity rates between adults and puppies (P>0.05), the positivity rates of both (73.91%) were slightly higher than those reported in previous studies. The main role of CAstV in viral enteric disorder still remains unknown (Mihalov-Kovacs et al., 2017 ▶), however, it is known to be a significant viral cause of gastroenteritis in children (Olortegui et al., 2018 ▶). The higher incidence of infection implies that CAstV might be considered as a factor causing diarrhoea in dogs. However, a comprehensive study of other geographic areas is needed to support this argument so that CAstV could be included as a component of future multivalent vaccines, similar to the work published by Xia et al. (2016) ▶ on the trivalent, recombinant subunit vaccine against human astrovirus infections, along with norovirus and hepatitis E (Xia et al., 2016 ▶). It must be noted that we designed a universal primer set for CAstV for this study. The higher incidence may also validate the accuracy of the designed primer sets, since cumulative data in the GenBank database had been utilized to create unique, degenerated oligonucleotides from the most conserved regions.

Sequence comparison in the ORF1b (RdRp) was calculated using a 294 nt (98 aa) long fragment located at 3484 to 3777 nt position (based on the RefSeq) from the 5´ UTR of CAstV genome which makes the last quarter of the ORF1b gene. Mihalov-Kovács et al. (2017) ▶ conducted a comprehensive study about CAstV and suggested that three distinct genetic sub-lineages existed based on the partial ORF1b gene sequences. We thus included all the sequence data deposited in the GenBank including novel strains (94 sequences in total) to evaluate the overall results precisely. The phylogenetic tree and similarity heat map showed that four distinct sub-lineages were available, and as shown in Supplementary Material 1 (SM1), the identity percentages gradually differed from >90% to <60%.

Canine calicivirus infections belonging to Norovirus (Ntafis et al., 2010 ▶; Martella et al., 2019 ▶) Sapovirus (Li et al., 2011 ▶) and Vesivirus (Roerink et al., 1999 ▶; Mochizuki et al., 2002 ▶; Castro et al., 2013 ▶; Martella et al., 2015 ▶; Renshaw et al., 2018 ▶) genera have been reported throughout the world. Canine vesiviruses were first included in the Vesivirus genus by Matsuura et al. (2002) ▶ and the incidence of CVeV literature reports variations between 1.1% and 64.8% (Mochizuki et al., 2002 ▶; Martella et al., 2015 ▶). In our study, this ratio was 3.33% (5/150). The positivity rates of adults (3.15%) and puppies (4.35%) were found not to be statistically significant (P>0.05). Nonetheless, this result is important, as it is the first investigation of CVeV in Turkey.

According to the phylogenetic analysis, two distinct sub-lineages of CVeV (Fig. 2) were demonstrated. As shown in Supplementary Material 2 (SM2), in comparison between clade 1 and 2, the nucleotide similarity level decreased from 67.04% to 71.75%, while clade 1 strains were from 84.59% to 100.00% within the group and clade 2 strains were from 87.74% to 99.16%.

Canine circovirus infections have been detected by various researchers in Europe (Matsuura et al., 2002 ▶; Kapoor et al., 2012 ▶; Decaro et al., 2014 ▶), the USA (Zaccaria et al., 2016 ▶; Anderson et al., 2017 ▶), Brazil (Weber et al., 2018 ▶; Kostias et al., 2019 ▶), Australia (Neef et al., unpublished; Bhatta et al., 2019 ▶) and the Far East (Hsu et al., 2016 ▶; Sun et al., 2019 ▶), and the incidence has been reported to range from 3.64% to 32.42%. Infection percentage was found to be 6% (9/150) in this study, and the positivity rates of both adults (6.30%) and puppies (4.35%) were found to be statistically insignificant (P>0.05). According to the phylogenetic analysis based on both Rep and Cap genes, our isolates substituted each other with other European and American strains, apart from fox circoviruses clustered together in a distinct branch. Despite the fact that Turkish strains included some of the far eastern strains, other branches did not include any European, American and Turkish strains.

In conclusion, we reported the presence of CCiV, CVeV, and CAstV from diarrheic dogs and provided initial genomic data for Turkey. We also suggested four sub-genotypes of CAstV and two sub-genotypes of CVeV. In addition, we also conducted the phylogenetic analysis of CCiV based on two gene regions according to molecular analyses of partial genomes. This is, to our knowledge, the first report showing proof of existence for these viruses in Turkey.

Conflict of interest

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Supporting Online Material

Refer to web version on PubMed Central® (PMC) for Supplementary Material.