Different virulence of porcine and porcine-like bovine rotavirus strains with genetically nearly identical genomes in piglets and calves

Direct interspecies transmissions of group A rotaviruses (RVA) have been reported under natural conditions. However, the pathogenicity of RVA has never been directly compared in homologous and heterologous hosts. The bovine RVA/Cow-tc/KOR/K5/2004/G5P[7] strain, which was shown to possess a typical porcine-like genotype constellation similar to that of the G5P[7] prototype RVA/Pig-tc/USA/OSU/1977/G5P9[7] strain, was examined for its pathogenicity and compared with the porcine G5P[7] RVA/Pig-tc/KOR/K71/2006/G5P[7] strain possessing the same genotype constellation. The bovine K5 strain induced diarrhea and histopathological changes in the small intestine of piglets and calves, whereas the porcine K71 strain caused diarrhea and histopathological changes in the small intestine of piglets, but not in calves. Furthermore, the bovine K5 strain showed extra-intestinal tropisms in both piglets and calves, whereas the porcine K71 strain had extra-intestinal tropisms in piglets, but not in calves. Therefore, we performed comparative genomic analysis of the K71 and K5 RVA strains to determine whether specific mutations could be associated with these distinct clinical and pathological phenotypes. Full-length sequencing analyses for the 11 genomic segments for K71 and K5 revealed that these strains were genetically nearly identical to each other. Two nucleotide mutations were found in the 5′ untranslated region (UTR) of NSP5 and the 3′ UTR of NSP3, and eight amino acid mutations in VP1-VP4 and NSP2. Some of these mutations may be critical molecular determinants for RVA virulence and/or pathogenicity.


Introduction
Group A rotavirus (RVA), a member of the Reoviridae family, is one of the major pathogens that causes severe and acute dehydrating diarrhea in young children and in a wide variety of domestic animals [1][2][3]. The rotavirus genome is enclosed in three concentric layers and is comprised of 11 segments of double-stranded (ds) RNA, encoding six structural proteins (VP1-4, VP6 and VP7) and five or six nonstructural proteins (NSP1-NSP5/6) [1,2,4]. This unique segmented nature of the RVA genomes allows for the occurrence of reassortment between different viruses during co-infection in a single cell and induction of progeny viruses with novel or atypical phenotypes [1,2,4].
A whole genome-based genotyping classification system for RVA has been proposed by the Rotavirus Classification Working Group (RCWG), which is based on nucleotide percentage identity cut-off values for each of the 11 RVA genomic segments [5][6][7]. Among these genotypes, the G (for glycoprotein) and P (for protease-sensitive) genotypes for the VP7 and VP4 outer capsid proteins, respectively, are the most important and frequently analyzed for rotavirus classifications, as they possess important immunogenic epitopes that are relevant for immune protection and vaccine development [1,8]. In extensive genomic studies, 27 G and 37 P genotypes have been identified globally, with various combinations of G and P genotypes distributed across humans and animals [7,9]. The most common G serotypes of RVA found in pigs are G3, G4, G5, G9 and G11 in combination with two dominant P genotypes, P [6] and P [7] [10][11][12]. In calves, the G6, G8, and G10 genotypes are the major types in combination with either the P [1], P [5] or P [11] genotype [13,14].
Porcine and bovine RVA are important pathogens because of their significant economic impact on the livestock industry as well as their crucial role as reservoirs for human infections [5,11,41]. Therefore, we previously investigated the causes of porcine and bovine diarrhea illnesses and characterized multiple RVA isolates from diarrheic fecal samples [11,19,20,32]. Among the Korean RVA isolated, the bovine RVA strains RVA/Cow-tc/KOR/ K5/2004/G5P [7] and RVA/Cow-tc/KOR/K8/2005/G5P [7] isolated from diarrheic calves have a remarkably similar genotype constellation as the porcine G5P [7] RVA strain OSU [32]. This prompted us to investigate whether this heterologous strain K5 could efficiently infect and induce diarrhea and pathology in the intestinal and extra-intestinal tracks of its homologous (piglets) and heterologous (calves) host. To assess the cross-species pathogenicity of the bovine K5 RVA strain in calves and piglets, the porcine G5P [7] OSU-like RVA strain RVA/Pig-tc/KOR/K71/2006/G5P [7] isolated from a diarrheic piglet was also used for comparison in this study. From these experimentally infected animals, various tissues and organs, fecal and nasal swabs and blood were sampled and analyzed for morphological changes, antigen distribution using an immunofluorescence assay (IFA) and viral RNA presence by RT-PCR and real-time RT-PCR. To better understand the potential consequences of viral genetic variations on infection characteristics, the genomic polymorphisms between the porcine K71 and bovine K5 strains were compared. The findings from our study provide insight into the possible functional contributions of viral genomic polymorphisms in the virulence and pathogenicity of heterologous RVA strains.

Origin of virus strains
The bovine K5 and porcine K71 RVA strains were isolated from diarrheic fecal samples of a piglet and a calf, respectively [11,32]. Both strains were passaged eight times in a monolayer of TF-104 cells (a cloned derivative of MA-104 monkey kidney cells) including the initial adaptation and triple plaque purification prior to characterization. Virus titers were assessed by cell cultured immunofluorescence (CCIF) assay using a monoclonal antibody against the VP6 protein of the OSU porcine strain, and were expressed as fluorescence focus units per milliliter (FFU/mL).
RT-PCR products amplified by primer pairs specific to each genomic segment were purified using a QIAEX II Gel Extraction kit (Qiagen, Valencia, CA, USA) according to the manufacturer's instructions. The extracted PCR products were ligated into the pGEM-T Easy Vector Systems (Promega), and were sub-cloned into home-made XL1-Blue competent cells. Individual colonies were grown and plasmid was purified using Hybrid-Q™ Plasmid (GeneAll, Seoul, South Korea). DNA sequencing was carried out using an ABI System 3700 automated DNA sequencer (Applied Biosystems, Foster City, CA, USA).
Determination of the 3′ and 5′ terminal sequences of K71 and K5 rotavirus strains Poly (A) tailing reaction and 3′ and 5′ cDNA synthesis Total RNA was extracted from a starting volume of 200 μL cell lysates by AccuPrep® Viral RNA extraction kit (Bioneer) according to the manufacturer's instructions. Poly (A) tailing reaction was done using Escherichia coli poly (A) polymerase (Ambion, Austin, TX, USA) according to the manufacturer's instructions. cDNA synthesis was performed using the SMARTer™ Rapid Amplification of cDNA Ends (RACE) cDNA amplification kit (Clontech, Mountain View, CA, USA). For the one reaction generation of 3′ RACE-ready cDNA, 3.75 μL of the poly A tailed RNA and 1.0 μL of 3′-cDNA Synthesis (CDS) Primer A were mixed and incubated at 72°C for 3 min, followed by cooling the tubes at 42°C for 2 min using a thermo cycler. The denatured RNA was mixed with a reaction mixture composed of 2.0 μL 5 X First-Strand Buffer, 1.0 μL dithiothreitol (DTT) (20 mM), 1.0 μL dNTP mix (10 mM), 0.25 μL RNase inhibitor (40 U/μL), and 1.0 μL SMARTScribe™ Reverse Transcriptase (100 U). Samples were incubated at 42°C for 90 min and were heated to 70°C for 10 min using a thermo cycler. The synthesized cDNA was diluted with 7 μL of Tricine-EDTA buffer and was used immediately or stored in −20°C prior to RACE PCR.
For a more efficient amplification of 5′ RACE cDNA for both strains, cDNA synthesis was carried out using random primers. In brief, 2.75 μL of total RNA was mixed with 1.0 μL of 10X random primer, incubated at 72°C for 3 min, and cooled at 22°C for 2 min. The preheated RNA was mixed with a reaction mixture consisting of 2 μL 5 X First-Strand Buffer, 1 μL DTT (20 mM), 1 μL dNTP mix (10 mM), 0.25 μL RNase inhibitor, and 1.0 μL SMARTer™ II A Oligonucleotide (12 uM). The reaction was incubated at room temperature (RT) for 10 min before adding 1.0 μL of SMARTer™ reverse transcriptase (100 U/μL). The reaction was incubated at 42°C for 90 min and was then heated to 70°C for 10 min. The cDNA was diluted with 7 μL of Tricine-EDTA buffer and was then used immediately or stored at −20°C prior to RACE PCR.
The RACE PCR products were separated on a 1.2% (m/v) agarose gel, and the bands were excised and purified using the Purelink™ Quick Gel Extraction kit (Invitrogen, Lohne, Germany). The RACE PCR products were ligated into a pGEM-T Easy Vector Systems (Promega) and were sub-cloned into home-made XL1-Blue competent cells. Individual colonies were grown and plasmid was purified using Hybrid-Q™ Plasmid (GeneAll). Sequencing was performed using an ABI System 3700 automated DNA sequencer (Applied Biosystems).

Molecular characterization
To generate complete nucleotide sequences for each genomic segment, both 5′ and 3′ end sequences of each genomic segment were assembled with the nucleotide sequence of the internal region sequenced with GSP. The full-length nucleotide sequences of each genomic segment of both porcine K71 and bovine K5 strains were compared with those of the other known RVA strains using the DNAsis MAX DNA Basic module (MiraiBio, Alameda, CA, USA). Phylogenetic and molecular evolutionary analyses based on the nucleotide alignments were constructed using the neighbor-joining method [47]. Genetic distances were calculated using the Kimura-2 correction parameter at the nucleotide level, and the phylogenetic trees were constructed using the neighborjoining method with 1000 bootstrap replicates. A sequence similarity search for the porcine K71 and bovine K5 strains with those of the other known RVA strains (see Additional file 2) was performed using the homology and distance matrices method of DNAMAN version 6.0 program (Lynnon, Vaudreuil, PQ, Canada).

Genotype assignment
The genotypes for the 11 genome segments of the porcine K71 and bovine K5 rotavirus strains were determined according to the genotyping recommendations of the Rotavirus Classification Working Group (RCWG) [6].

Genebank accession numbers
The Genbank accession numbers for the 11 genome segments of strains used in this study are shown in Additional file 2.
After inoculation, color and consistency of feces obtained from each calf and piglet were evaluated daily. The consistency of the feces was scored on a scale of 0-4, with 0 representing firm; 1, pasty; 2, semi-mucoid; 3, liquid and 4, profuse diarrhea [39,40,42,48]. Fecal and blood samples were collected daily from each calf and piglet before and after inoculation, as described previously [39,40,42,48]. The inoculated calves and piglets were euthanized at given times (Tables 1, 2, 3 and 4). Calves and piglets inoculated with the mock-or inactivated virus were euthanized at 2 or 3 day-post inoculation (dpi).
Necropsy was immediately performed after euthanasia. During necropsy, the intestinal tract, mesenteric lymph node (MLN), nasal turbinate, trachea, pancreas, lung, liver, spleen, brain, kidney, heart, urinary bladder, and choroid plexus were collected from experimental calves and piglets [39,40]. All organs were immediately placed in 10% buffered formalin for histological examination. Blood samples were collected from the jugular vein of the calves and via the intra-cardiac route for the piglets.

RT-PCR onset dpi (duration)
Nested-PCR onset dpi (duration) Feces Serum Feces Serum To evaluate the RVA-antigen distribution, intestinal and extra-intestinal organs from virus-inoculated and mock-inoculated experimental animals were sampled, embedded in Optimum Cutting Temperature compound, immediately snap-frozen in liquid nitrogen and stored at −80°C [39,40]. Cerebrospinal fluid (CSF) was collected after euthanasia from each experimental animal and stored at −80°C [39,40]. All samples collected for RT-PCR and real-time RT-PCR were immediately snapfrozen in liquid nitrogen, and kept at −80°C until use [39,40]. All procedures were approved by the Animal Care Committee of Chonnam National University (CNU IACUC-YB-2009-15).

Examination of rotavirus antigen by IFA
For assessment of the antigen distribution of rotavirus, IFA was performed in each organ sampled from virus-

RT-PCR onset dpi (duration)
Nested-PCR onset dpi (duration) Feces Serum Feces Serum Inoculation with a chloroform-inactivated bovine K5 strain. inoculated and mock-inoculated animals, as described elsewhere [35,39,40,42,48]. Briefly, cut frozen sections from experimental animals were fixed in 100% cold acetone for 10 min and were allowed to completely air-dry. Slides were washed twice with phosphate buffered saline (PBS, pH 7.2), and incubated for 2 h at RT with a 1:100 dilution of monoclonal anti-VP6 antibody diluted in PBS (pH 7.2). Slides were washed twice with PBS (pH 7.2), and incubated with goat anti-mouse Ig conjugated to fluorescein isothiocyanate (Jackson ImmunoResearch Labs, Baltimore, MD, USA) diluted 1:100 in PBS (pH 7.2) for 1 h at RT. Following incubation, the slides were washed twice with PBS (pH 7.2). Slides were incubated with propidium iodide diluted in 500 nM PBS (pH 7.2) for 10 min at RT as a nucleic acid stain. Slides were washed twice with PBS (pH 8.0), and covered with 60% glycerin in PBS (pH 8.0), and glass cover slips. Fluorescence of the samples was examined microscopically under ultraviolet illumination (Leica Microsystems, Wetzlar, Germany). To calculate the number of antigen-positive cells in the organs or tissues, 10 fields per section were analyzed, using a 40× objective and a 10× eyepiece, yielding a final magnification of 400×. Total counts of antigen-positive cells were calculated as mean values.

Examination of viral RNA in the fecal and serum samples by RT-PCR and nested-PCR
For the detection of viral RNA in the fecal and serum specimens from each experimental animal, RT-PCR and nested-PCR with primer pairs specific to the RVA VP6 gene were performed as previously described [39,40]. As negative controls, fecal and serum specimens sampled from mock-inoculated animals were used. The PCR products were visualized on 1.2% agarose gels stained with ethidium bromide.

Quantificaiton of viral RNA by real time RT-PCR using SYBR green chemistry
The one-step real-time RT-PCR assay was performed with a primer pair specific to the RVA VP6 gene, as described previously [39,40,49]. Briefly, all tissue and fluid samples from the experimental animals were individually weighed, homogenized or vortexed at a 1:10 dilution in 0.01 M PBS and were centrifuged (tissues 13 000 × g for 3 min; fecal samples 5000 × g for 10 min). The supernatants along with the remaining bulk samples were collected and stored at −80°C prior to analysis. After extracting total RNA from supernatants, each real time RT-PCR reaction was performed using a Rotor-Gene Real-Time Amplifica   and this threshold, a direct relation between the cycle number and the log concentration of RNA molecules initially present in the RT-PCR reaction was evident. By linear regression analysis of these data, Rotorgene 6000® software set up a standard curve, which enabled the determination of the concentration of RNA present in the samples.

Sequence and phylogenetic analyses
Three full-length (VP7, NSP2 and NSP5) and eight partial-length genomic sequences (VP1-VP4, VP6, NSP1, NSP3 and NSP4) from the bovine RVA strain K5 were characterized previously [32]. For the porcine K71 strain, the full-length VP7 and partial VP4 genomic sequences were also determined [11]. In this study, the full-length genomic sequences of both strains were sequenced by RT-PCR assays with GSP specific to both 5′ and 3′ end sequences. In addition, 5′ and 3′ RACE for the 11 genome segments of both strains was performed to obtain the full-length nucleotide sequences of both strains. The full-length nucleotide sequences of all 11 genomic segments of both porcine K71 and bovine K5 strains were analyzed and compared with those of the reference RVA strains. Generally, both strains shared 99.6-100% nucleotide identities of all 11 genomic segments ( Table 5). The gene segments for K71 and K5 had the highest nucleotide identities with porcine RVA strain OSU (VP1-VP4, VP7, NSP1-NSP3), porcine strain JL94 (VP6), porcine strain RMC321 (NSP4), and panda strain CH-1 (NSP5) ( Table 5). The genotype constellations of strains K71 and K5 were identical; G5-P[7]-I5-R1-C1-M1-A1-N1-T1-E1-H1, which is typical for porcine G5P [7] RVA strains, such as OSU (see Additional file 3). Phylogenetic analyses also demonstrated that all 11 genomic segments of both porcine K71 and bovine K5 strains clustered with those of porcine and porcine-like RVA strains (see Additional file 4).

Intestinal pathogenicity in piglets and calves
The above results indicate that the porcine K71 and bovine K5 strains possessed genomic constellations typical to porcine RVA strains such as OSU. Therefore, we determined whether these strains show different intestinal pathogenicity in piglets and calves. The porcine K71 strain induced continuous diarrhea in all inoculated piglets from 1 dpi to the termination of the experiment ( Table 1), but did not cause any diarrhea in the inoculated calves (Table 3). Nested PCR assay continuously detected fecal virus shedding from 1 dpi to the termination of the experiment in piglets (Table 1). Although the virus was also detected at 1 dpi in all calves, this viral shedding persisted for only 1 or 2 days (Table 3).
Small intestinal lesions typical of RVA infection included desquamation of villi epithelium resulting in villous atrophy and fusion, and crypt hyperplasia. Sequential histopathological lesion changes in the small intestines of piglets and calves infected with each strain are summarized in Additional files 5, 6, 7 and 8. The porcine K71 strain tended to cause a sequential increase of histopathological lesion changes in the small intestine of inoculated piglets, but induced only mild changes in the small intestine of calves (see Additional files 5, 6, 7 and 8 and Figure 1). RVA antigen-positive cell counts in the small intestinal villi were much higher in piglets than in calves (see Additional files 5, 6, 7 and 8 and Figure 2). Mock-inoculated or inactivated-virus-inoculated piglets and calves had no diarrhea, fecal virus shedding, or histopathological changes in the small intestine. From these results, it can be concluded that the porcine K71 strain efficiently infected the small intestines of piglets, but not calves. In contrast to the porcine K71 strain, the bovine K5 strain induced continuous diarrhea in both virus-inoculated piglets and calves from 1 dpi to the termination of the experiment (Tables 2 and 4). Fecal virus shedding was also continuously detected by nested PCR from K5 strain-inoculated piglets and calves (Tables 2 and 4). The bovine K5 strain tended to cause a sequential increase of small intestinal lesion changes in both virus-inoculated calves and piglets (see Additional files 7 and 8, and Figure 3). RVA antigen-positive cells in the small intestinal villi were detected at high levels at 1 dpi and gradually decreased in both virus-inoculated calves and piglets (see Additional files 7 and 8, and Figure 2). Mock-inoculated or inactivated-virus-inoculated piglets and calves had no diarrhea, fecal virus shedding, or histopathological changes in the small intestine. These results indicate that bovine K5 efficiently infected the small intestines of both calves and piglets.

Extra-intestinal histopathological lesion changes in piglets and calves
The above results indicate that the porcine K71 strain efficiently infected the small intestines of piglets, whereas the bovine K5 strain efficiently infected the small intestines of calves and piglets. Furthermore, histological changes in extra-intestinal organs and tissues were examined in piglets and calves inoculated with the porcine K71 or bovine K5 strains. Histopathological changes were absent in the MLN sampled from inactivated virusor mock-inoculated piglets or calves. The porcine K71 RVA strain induced lymphoid cell depletions with the infiltration of some macrophages and neutrophils in   the cortex of MLN obtained from piglets ( Figure 4) but not from calves (data not shown). However, the bovine K5 strains induced these histopathological changes in the MLN from both virus-inoculated piglets and calves (Figures 5 and 6). As was seen for the histopathological changes, porcine K71 RVA antigen-positive cells were detected in the MLN of piglets, but not from those of calves (see Additional files 9 and 10, and Figure 4). However, bovine K5 RVA antigen-positive cells were observed in the MLN of both piglets and calves (see Additional files 11 and 12, and Figures 5 and 6).
The porcine K71 strain induced hepatocyte necroses with some macrophages and lymphocyte infiltration in the liver sampled from the virus-infected piglets (Figure 4) but not from virus-infected calves (data not shown). Cells positive for RVA K71 antigen were only detected in the liver from the virus-infected piglets but not in that from the virus-infected calves (see Additional files 9 and 10, and Figure 4). In contrast, the bovine K5 strain induced hepatocyte necroses and antigen-positive cells in the livers from both virus-inoculated piglets and calves (see Additional files 11 and 12, and Figures 5 and 6).
The porcine K71 strain caused lung lesions characterized by mild to moderate thickening of alveolar walls by hyperplasia of type II pneumocytes, and infiltration of macrophages and lymphoid cells into the alveolar  Hematoxylin and eosin stain (A, B, D, E, G, H, J, and K). Indirect immunofluorescence assay with monoclonal against the VP6 protein of strain OSU (C, F, I and L). Bars denote 100 μm. interstitium in the virus-inoculated piglets (Figure 4) but not in the calves (data not shown). Cells positive for K71 RVA antigen were only detected in the lungs from the virus-inoculated piglets but not in those from the virus-inoculated calves. The bovine K5 strains elicited these lesions in the lungs of both virus-inoculated piglets and calves (see Additional files 11 and 12, and Figures 5  and 6).
In the choroid plexus, the porcine K71 strain induced limited epithelial degeneration and necroses with mild lymphoid cell infiltration in the tela choroidea of the piglets (Figure 4) but not in that of calves (data not shown). Cells positive for K71 porcine RVA antigen were only detected in the tela choroidea of piglets but not in that of calves (see Additional files 9 and 10, and Figure 4). The bovine K5 strain cited above induced histopathological changes and antigen-positive cells in the tela choroidea of both virus-inoculated piglets and calves (see Additional files 11 and 12, and Figures 5 and 6).

Quantification of viral RNA copy numbers
To quantify viral RNA copy numbers in the feces, sera, and extra-intestinal organs and tissues, real-time RT-PCR assays based on SYBR Green chemistry were performed with a primer pair specific to the RVA VP6 gene [39,40,49]. Real-time PCR results showed that viral RNA copy numbers of the porcine RVA strain K71 was observed in the feces of piglet at 1 dpi, reached a peak at 3 dpi, and then gradually declined after 5 dpi ( Figure 7A). As expected, viral RNA copy numbers of porcine K71 strain were much lower in the fecal samples of virusinoculated calves than those of virus-inoculated piglets throughout the experimental period ( Figure 7A). These results demonstrate that the porcine K71 strain efficiently replicated and induced pathology in the intestines of piglets, but not in those of calves. The bovine K5 strain replicated and induced pathology in the intestines of both virus-inoculated piglets and calves, but the viral RNA copy numbers were higher in the feces of piglets than in those of the calves, suggesting that strain K5 replicated more efficiently in piglets than in calves ( Figure 7B and C).
Viremia and extra-intestinal tropisms have been described in piglets and calves having typical gastrointestinal lesions [50][51][52][53]. As described above, the porcine K71 strain was pathogenic only to piglets but the bovine K5 strain was pathogenic to both piglets and calves. Therefore, viremia and extra-intestinal tropisms were evaluated in the sera and extra-intestinal specimens obtained from piglets inoculated with either porcine K71 or bovine K5 strains and calves inoculated only with the K5 strain. The viremia caused by the inoculation of the porcine K71 strain in piglets was detected by nested PCR at 2 dpi, and persisted for 2 or 3 days (Table 1), but no viremia was detected in the calves (Table 3). Real-time PCR results showed that viral RNA in the sera of virus-inoculated piglets were present at 1 dpi, reaching a peak at 3 dpi, and then gradually decreasing ( Figure 7A). The viral RNA of bovine K5 strain were detected by nested PCR at 1 or 2 dpi in both virus-inoculated piglets and calves, and lasted to 5 dpi in piglets and to 6 dpi in calves (Tables 2 and 4). Real-time PCR detected viral RNA of bovine K5 strain in the sera from both virus-inoculated piglets and calves at 1 dpi; the copy numbers increased at 3 dpi and then decreased at 5 dpi ( Figures 7B and C). Viral RNA copy numbers detected in the sera of piglets were, however, higher than that in calves throughout the whole experimental period (Figures 7B and C). These results indicate that both strains can cause viremia in piglets, but only K5 strain can induce viremia in calves.
Like the sequential changes of virus RNA copy numbers in the feces and sera, the viral copy numbers in the extraintestinal organs and tissues revealed similar changes in piglets infected with K71 strain and in calves infected with K71 or K5 strains (Figures 7A-C). The viral RNA were detected at 1 dpi, and its copy numbers reached a peak at 3 dpi and then decreased from 5 dpi. Among the extraintestinal organs and tissues, the MLN from the virusinfected piglet and calves show the highest RNA copy numbers by real-time PCR (Figures 7A-C).

Comparison of nucleotide and amino acid sequences
Although the bovine K5 and the porcine K71 strains shared the same genotype constellation, their virulence pattern differed between piglets and calves. Therefore, full-length nucleotide and deduced amino acid sequences of all 11 genomes of both strains were compared to each other. Genomic polymorphisms which were found in each of the two strains are listed in Table 6.
The comparison of the full-length nucleotide sequences of all 11 genomic segments of bovine K5 and porcine K71 strains identified 13 nucleotide changes. Among them, two nucleotide changes were evident; one in the 3′ untranslated region (UTR) of NSP3 and one in the 5′ UTR of NSP5 ( Table 6). The remaining 11 nucleotide changes were observed in the gene coding regions of VP1 (two nucleotides), VP2 (two nucleotides), VP3 (one nucleotide), VP4 (one nucleotide), and NSP2 (four nucleotides) segments. Among the 11 nucleotide changes in the gene coding regions, three were silent mutations, resulting in eight amino acid substitutions (Table 6). Next, we examined if these point mutations were present in other porcine or porcine-like bovine RVA strains, and for strain K5 we found three unique amino acid substitutions in the open reading frames of VP1, VP3 and VP4, along with one nucleotide substitution in the 3′ UTR of the NSP3 segment, which were not present in the other strains examined (Table 6).

Discussion
The high mutation rate in RVA, the ability of interspecies transmission and genome segments to reassort, and the large pool of RVA in mammals and birds ensure their continuously changing behavior [1]. Knowledge of the virulence and pathogenicity of RVA strains in different host species is key to understanding methods of prevention and control of RVA infections. Direct interspecies transmissions of RVA have been reported to occur in nature on a number of occasions [31,32,54]. However, the pathogenicity of such strains remains largely unknown in different host species. In our previous report [32], the bovine K5 strain has been isolated from a diarrheic calf, and characterized as a porcine-like RVA strain due to having the same genotype constellation as the porcine RVA reference strain OSU. In the present study, we further characterized its pathogenicity in its homologous (piglets) and heterologous (calves) hosts. In addition, the virulence of the bovine K5 strain in the homologous and heterologous hosts was compared with that of the porcine K71 strain. The porcine K71 strain was isolated from a diarrheic piglet and was identified as being an OSU-like strain with a genotype constellation identical to that of the OSU strain [11]. On the contrary to our expectations, the bovine K5 strain showed pronounced virulence and pathogenicity in the homologous (piglets) as well as heterologous (calves) hosts; piglets and calves infected with this strain had continuous diarrhea, high virus RNA copy numbers in the fecal samples, and severe and large histopathological lesions in the small intestinal mucosa. As expected, however, the porcine K71 strain had severe virulence and pathogenicity in piglets, but not in calves. When RVA cross the host species barrier, this is usually a dead-end infection since the virus is not able to properly spread in the new host species. However, if the interspecies transmitted virus is able to reassort with RVA inherent in that host species, the resultant reassortant strains may have a higher probability to efficiently infect and spread among the population of the new host [5,25,54,55]. On the contrary to this general belief, our results highlight the existence of different enteropathogenic virulence depending on the strains, even among those with the same genotype constellation in nature. To our knowledge, this is the first study on the virulence and host range restriction of a particular RVA strain in its homologous and heterologous hosts.
Although the virulence and pathogenicity of heterologous RVA infections have never been documented under natural conditions, they have been reported in experimental conditions. Simian RVA strain RRV and lapine strains share similar VP4, VP7 and NSP1, which have been implicated in host range restriction, and the simian RRV strain has been shown to replicate efficiently and to transmit horizontally in a rabbit model [56]. Therefore, it is possible that the possession of these three related genes may confer RRV the ability to replicate efficiently in rabbits [56]. Genetically, human DS-1-like strains are believed to have a common origin with bovine RVA strains because they share the same genotypes of six genomic segments (VP6, VP1, VP2, VP3, NSP2, and NSP4) [5]. Moreover, DS-1-like strains have been shown to successfully infect cattle [57,58], but not piglets [26]. On the contrary, the virulent human RVA strain Wa has a similar genomic constellation to those of porcine RVA, i.e., only the VP7, VP4, and VP6 genotypes are different with the porcine RVA strain OSU, VP7, VP4, and NSP1 with the porcine RVA strain Gottfried, and VP7, VP4, VP6, and VP2 genes with the porcine RVA strain A131 [5]. Moreover, the human RVA strain Wa induces diarrhea and pathology in the small intestine of piglets [38,59]. Based on these genomic and pathological properties, it has been speculated that human Wa-like RVA strains could have a common origin with the porcine Wa-like strains [5]. In the present study, both the bovine K5 and the porcine K71 RVA strains shared the genotype constellation with porcine RVA strain OSU, and induced diarrhea and intestinal pathology in piglets. In calves, however, the virulence of both strains was different; the bovine K5 strain induced diarrhea and intestinal pathology, but the porcine K71 did not. It is still unknown why the genetically nearly identical porcine K71 and bovine K5 strains had different virulence in calves. Interestingly, the human Wa-like strain D can also infect and induce mild disease in calves [5,57]. Like the human Wa-like strain D, the bovine K5 strain showed virulence not only in piglets, but also in calves.  Host specificity and pathogenicity of RVA may be the result of the interaction of numerous host factors and viral proteins. However, the molecular basis for virulence and/or host range restriction of the RVA strain is still unclear. To date, a number of gene segments have shown to be capable of affecting host range restriction and/or virulence (VP3, VP4, VP7, NSP1, NSP2, and NSP4) [22][23][24][25][26][27][28]. Most of these experiments were conducted using reassortant strains possessing quite different parental viruses. Our data show highly different virulence profiles in different hosts between two closely related RVA strains with identical genotype constellations. This indicates that differences in virulence and host range restriction cannot be fully explained by the genotype constellation.
The present sequence comparison between the bovine K5 and porcine K71 strains revealed that the bovine K5 strain harbored two nucleotide differences in the 5′ and 3′ UTR, and 11 nucleotide differences in the gene coding regions, resulting in eight amino acid substitutions. These unique mutations between the two strains may be possible candidates for virulence determinants in the heterologous host (Table 6). Since little information is available about the key virulence determinants (mutations) responsible for virulence and pathogenicity of RVA in a heterologous host, it is still uncertain whether the increased virulence and pathogenicity of bovine K5 strain in the heterologous host (calves) is induced by any single substitution in a certain gene or by a combination of these substitutions. Furthermore, three amino acid substitutions in VP1, VP3 and VP4, and one nucleotide substitution in the 3′ UTR of K5 were found to be unique in comparison with other porcine and bovine RVA reference strains. Since the virulence and pathogenicity in the heterologous and/or homologous hosts of these porcine and porcine-like bovine RVA reference strains has not been tested, it is still uncertain whether the increased virulence and pathogenicity of bovine K5 strain in the heterologous host (calves) are induced by the observed substitutions. Therefore, the molecular and biological properties of each genome segment, and its possible association with host range restriction and/or virulence will be needed to be clarified in further studies, using reverse genetics to swap individual genes of these viruses or introduce particular mutations. The exact mechanisms of how the bovine K5 strain acquired these specific mutations are unknown. The mutations of strain K5 could have been acquired by serial replication in vivo in calves, or alternatively, the mutations of strain K5 could have occurred during serial replication in vivo in piglets, and subsequently have been transmitted to calves.
In human RVA infections, RVA antigen and viral RNA have been sporadically detected in the serum and various extra-intestinal organs and tissues [34,60,61]. Moreover, viremia and extra-intestinal pathogenicity of RVA infection have been clearly demonstrated in experimental animal models [35][36][37]39,40,62]. However, little information is available as to whether heterologous RVA strains cause viremia and extra-intestinal pathogenicity in heterologous hosts. To address this point, we sequentially collected serum and extra-intestinal organs and tissues from piglets and calves inoculated with the bovine heterologous strain K5. Both piglets and calves inoculated with the K5 strain showed similar dynamics of viral RNA in the sera and extra-intestinal organs and tissues. Viral RNA was detected at 1 dpi and its concentration reached a peak at 3 or 5 dpi, and then gradually decreased. Viral antigen was detected in these extra-intestinal organs and tissues where degeneration and/or necrosis of the corresponding parenchymal cells were observed. Although this is the first report that the heterologous RVA strain causes viremia and extra-intestinal pathogenicity in its heterologous hosts, the patterns of viremia and extra-intestinal infection are consistent with those of other reports [35][36][37]39,40,62].
In conclusion, bovine heterologous K5 strain, despite having the same genomic constellation with the porcine G5P [7] prototype OSU and the Korean porcine K71 strains, displays pronounced virulence, and intestinal and extra-intestinal pathogenicities in a heterologous host. In addition, non-synonymous mutations of bovine K5 strain compared with those of porcine homologous K71 strain may contribute to virus replication ability, virulence, and pathogenicity in the heterologous host (calves). These preliminary observations contribute to our understanding of RVA infection in homologous and heterologous hosts. To further clarify the molecular and biological properties of each genomic segment associated with host range restriction and/or virulence, a combination of standard biochemical and molecular tools, together with the experimental inoculation studies with many heterologous strains in their natural homologous and heterologous hosts and new technologies such as reverse genetics system, will be required.

Additional files
Additional file 1: Oligonucleotide primers for sequencing or for 5′ and 3′ RACE PCR of all eleven genomic segments of the porcine K71 and bovine K5 rotavirus strains. The primer pairs used to generate the full-length sequence of porcine K71 and bovine K5 rotavirus strains are listed in the table. Also indicated are the gene-specific primers used for 5′ and 3′ RACE PCR.