Whole-Genome Analysis of a Rare Human Korean G3P Rotavirus Strain Suggests a Complex Evolutionary Origin Potentially Involving Reassortment Events between Feline and Bovine Rotaviruses

A rare human rotavirus, G3P[9] strain RVA/Human-tc/KOR/CAU12-2-51/2013/G3P[9], was isolated from the stool of a 9-year-old female hospitalized with acute watery diarrhea in August 2012 in South Korea using a cell culture system, and its genome was analyzed. The complete genomic constellation of the CAU12-2-51 strain revealed a novel genotype constellation for human rotavirus, G3-P[9]-I2-R2-C2-M2-A3-N2-T3-E3-H3. Phylogenetic analysis revealed that the CAU12-2-51 strain originated from feline- and bovine-like reassortment strains. The genes encoding VP4, VP7, NSP1, NSP3, NSP4, and NSP5 were related to human/feline-like and feline rotavirus strains, whereas the remaining five genes encoding VP1, VP2, VP3, VP6, and NSP2 were related to the human/bovine-like and bovine rotavirus strains. This novel strain was identified for the first time, providing evidence of feline/bovine-to-human transmission of rotavirus. The data presented herein provide information regarding rotavirus diversity and evolution.


Introduction
Rotaviruses are the major cause of acute gastroenteritis in young children worldwide. Each year, these viruses are associated with deaths of approximately 453,000 children younger than 5 years of age, most of whom reside in low-income countries in Africa and Asia, where access to potable water, sanitation, and medical care is often limited [1]. Rotavirus, a member of the family Reoviridae, is classified into eight species/groups (A-H) [2]. Its genome is composed of 11 gene segments of double-stranded RNA that encode six structural viral proteins (VP1-VP4, VP6, and VP7) and six nonstructural proteins (NSP1-NSP6) [3].
In South Korea, rotavirus is the most common viral agent causing acute diarrhea and gastroenteritis in young children, especially during the winter season. The distributions of the G and P genotypes show temporal and geographical fluctuations [22]. The major human G types are G1-G4 and G9, combined with the P types P [4], P [6], and P [8] [23,24]. Furthermore, unusual G and P combination genotypes, including G3P [3], G3P [9], G3P [10], G8P [8], G11P [4], G11P [25], and G12P [6], have been isolated from South Korea [25]. From 2003 to 2005, the G3 genotypes were predominant [26]. Although the P [9] genotype has been detected in some rotavirus surveillance studies in South Korea, no significant emergence of P [9] genotypes has been observed in the Korean population [22]; however, no studies have specifically focused on this genotype. Therefore, further analysis of the circulating Korean genotype P [9] is needed to understand its origin, genetic variation, and potential impact on the human population.

Ethics statement
The stool samples used in this study were collected and analyzed according to methods detailed in our study protocol (number #2011-10-06), which was approved by the Human Subjects Institutional Review Board (IRB) of Chung-Ang University College of Medicine, Seoul, Korea. Written informed consent was obtained from all participants in this study. For the children enrolled in this study, written informed consent was obtained from their parents. The provision of informed consent also included permission to use the data for future research purposes under conditions of anonymity.

Patient history and rotavirus identification
The stool specimen was collected from a 9-year-old girl who was hospitalized for severe gastroenteritis in August 2012 at Chung-Ang University Hospital, Seoul, South Korea. The patient was hospitalized for three days with episodes of watery diarrhea (three times/day), pain in the right upper quadrant of the abdomen, and vomiting (nonprojectile and nonbilious, containing consumed food). At that time, the patient had a hemoglobin level of 13.8 g/ mL, a white blood cell (WBC) count of 8,058 WBCs/mL, and a platelet count of 236,000 cells/mL. The stool WBC count was 0-1/high-power field. The patient had not received any prior rotavirus vaccines and had not had any contact with animals (e.g., cats, dogs, or cows) at least 30 days before hospitalization. Human group A rotavirus strain was then detected from the stool sample by both enzyme-linked immunosorbent assay (ELISA) and reverse transcription-polymerase chain reaction (RT-PCR). This strain was named CAU12-2-51. The stool culture produced no growth for pathogenic bacteria, such as Salmonella and Shigella, which commonly cause diarrhea.

Rotavirus isolation
MA104 cells were acquired from the Korean Cell Line Bank (Seoul, South Korea) and grown in minimum essential mediumalpha (MEM-a; Gibco BRL, Grand Island, NY, USA) containing 5% fetal bovine serum (FBS; Gibco BRL) and 0.1% gentamicin (Gentamicin Reagent Solution; Gibco BRL) at 37uC in the presence of 5% CO 2 . A G3P[9]-positive stool sample was diluted 10-fold in phosphate-buffered saline (PBS; pH 7.4) and clarified by centrifugation at 10,0006g for 10 min. The supernatant was filtered using a 0.45-mm sterile syringe filter (Corning Costar, Corning, NY, USA), treated with 10 mg/mL Trypsin 250 (Becton Dickinson, Sparks, MD, USA) for 30 min at 37uC. The supernatant was then inoculated onto MA104 cells in glass tubes and incubated with MEM-a in the presence of trypsin (5 mg/mL), with constant rotation during incubation. Cells were harvested 5-7 days after infection and subsequently passaged with MA104 cells until a cytopathic effect (CPE) was achieved. The rotavirus remaining in the culture fluid after the final passage was examined by immunofluorescence and RT-PCR.

RNA extraction and RT-PCR
Viral double-stranded RNA was extracted using TRIzol reagent (Gibco BRL Life Technologies) according to the manufacturer's instructions. Extracted RNA was resuspended in RNase-free water and stored at 280uC until use. The extracted RNA was then subjected to one-step RT-PCR using primer sets for the full amplification of all 11 gene segments (Table S1) [12,28,29,30,31].

Nucleotide sequencing and sequence analysis
All amplified PCR products were purified using a QIAquick PCR Purification Kit (Qiagen, Westburg, Germany) and then sequenced using a BigDye Terminator Cycle Sequencing Kit and an automated DNA sequencer (Model 3730; Applied Biosystems, Foster City, CA, USA). The resulting open reading frame (ORF) of the genes and deduced amino acid (aa) sequences were aligned using the CLUSTAL_X 1.81 program [28] and Lasergene software (DNASTAR, Madison, WI, USA). Data were compared with the corresponding rotavirus sequences from the National Center for Biotechnology Information GenBank database. The nucleotide (nt) sequences obtained in this study were deposited in GenBank under the accession numbers KJ187594-KJ187604.

Phylogenetic analysis
The nt sequences of the 11 segments of the CAU12-2-51 strain were compared with representative rotavirus sequences available from the GenBank database. Phylogenetic trees were constructed using neighbor-joining algorithms [25] from the PHYLIP suite [29] and the Kimura two-parameter model using MEGA5.03 software [30,31]. Evolutionary distances for the neighbor-joining analysis were based on the model described by Jukes et al. [32]. Tree topology was evaluated using the bootstrap resampling method with 1000 replicates of the neighbor-joining dataset with the SEQBOOT and CONSENSE programs from the PHYLIP suite.

Rotavirus genotyping and isolation
The VP7 and VP4 nt sequences of the CAU12-2-51 strain obtained from the first-round PCR products were identical to the sequences of G3 and P [9] genotypes available in the GenBank database. CPEs were observed after two culture passages when MA104 cells exhibited obscure borders, fusion, rounding, and detachment from the walls of the tubes (data not shown).

Gene sequence comparisons and phylogenetic analysis
The complete full-length nt and deduced aa sequences of the CAU12-2-51 strain were obtained. For reference, only the strains that contained complete or at least 50% sequence availability of ORFs, as suggested by Matthijnssens et al. [27], were used.
VP4 gene. Phylogenetic analysis of the P[9] genotypes revealed five distinct clusters, in which the VP4/P [9] genotype of the prototype strain AU-1 was assigned to cluster a. Phylogenetic analysis showed that the VP4/P [9] genotype of the CAU12-2-51 strain was classified into cluster c, together with the recently identified human and feline P [9] genotypes, including the feline P [9] strain isolated in Italy (BA222) and human/feline-like strains isolated in France (R1320), Japan (KF17), and Russia (O211) (Figure 1a). Interestingly, VP4/P [9] of the CAU12-2-51 strain showed the highest nt and aa sequence identities to the G3P [9] feline/human reassortant BA222 strain (97.9% and 99.2%, respectively). On the other hand, VP4/P [9] of the CAU12-2-51 strain showed lower sequence identity to VP4/P [9] of the prototype strain AU-1 (95.4% at the nt level and 97.3% at the aa level).

Rotavirus infection causes acute diarrhea and is the most common cause of gastroenteritis affecting infants and young children in South
The P [9] genotype is frequently detected in cats [47], and the first P [9] human rotavirus, the K8 strain with G1 specificity, was isolated from a 14-year-old boy in Hokkaido, Japan in 1977 [48]. Although P [9] genotypes have been detected sporadically in most of the rotavirus surveillance studies conducted in South Korea, there has been no significant emergence of P [9] genotypes in the Korean population. The relative global frequency of the P [9] genotypes represents less than 2.5% of total rotavirus infections, and the majority of these genotypes carry G1, G2, and G3 specificity, except for a few isolates that have been found to have G4, G6, G9, and G12 specificity [20]. The present analysis of VP4/P [9] genes revealed a genetic relationship between the CAU12-2-51 strain and the recently identified P [9] strains, demonstrating that this strain had the highest identity to the feline rotavirus strain BA222.
There are two approved vaccines for rotavirus, which were developed based on different strategies. Rotarix (GlaxoSmithKline Biologicals, Rixensart, Belgium) is a single, live-attenuated human G1P [8] rotavirus strain [50], while RotaTeq (Merck and Co. Inc., West Point, PA) is a pentavalent, live bovine-human reassortant vaccine containing the G1, G2, G3, G4, and P [8] genotypes [51]. RotaTeq and Rotarix were introduced in South Korea in September 2007 and July 2008, respectively. The patient in this study had not received any rotavirus vaccines and showed severe symptom of diarrhea. Comparison of the aa sequence of the VP4 gene of vaccines and CAU12-2-51 strains revealed low sequence identity (34.2%-34.9%). Novel rotaviruses may contain aa changes in regions that have known biological functions, and therefore, we expect that the effectiveness of vaccines raised against these regions may be poor when attempting to prevent infection of such reassortment rotavirus strains [52].
In conclusion, the repeated detection of G3P [9] strains in many surveillance studies suggests that these strains may represent important strains currently circulating in South Korea. Whole genome sequencing of all 11 segments of the CAU12-2-51 strain revealed a unique pattern of genetic diversity. Phylogenetic analysis of the CAU12-2-51 strain indicated that this strain had a complex evolutionary origin, potentially involving reassortment events between feline and bovine rotaviruses. The VP7, VP4, NSP1, NSP3, NSP4, and NSP5 genes were related to human/felinelike and feline rotavirus strains, and the remaining five genes, i.e., VP6, VP1, VP2, VP3, and NSP2, were the most related to human/ bovine-like and bovine rotavirus strains. Taken together, our current data enhance our understanding of the diversity of rotavirus through molecular, genetic, evolutionary, and epidemiological analyses.

Author Contributions
Conceived and designed the experiments: WK. Performed the experiments: SJ. Analyzed the data: VTT WK. Contributed reagents/materials/ analysis tools: IL WK. Wrote the paper: SJ WK.