Acute gastroenteritis (AGE) is the one of important causes of hospitalization in children in both developing as well as developed countries and Group A rotaviruses (RVA) are the most common cause of AGE among children. Recently, RV vaccines have been introduced in several countries including Korea to reduce the disease burden caused by RV infections. However, continuous surveillance of RV strains is necessary to detect the possible emergence of potential variants induced by vaccine-immune pressure because RV shows genetic diversity due to frequent point mutation and re-assortment. RVA belongs to the Reoviridae family, which is characterized by a segmented doublestranded RNA genome; RVA contains 11 gene segments encoding structural proteins (VP1-4, VP6, and VP7) and nonstructural proteins (NSP1-5/6)1). RVs are classified on the basis of their VP7 and VP4 gene sequences, and 27 G and 37 P genotypes have been reported2); the major genotypes reported worldwide are the G1P[8], G2P[4], G3P[8], G4P[8], G9P[8] combination genotypes, and G123). In Korea, the G1 genotype was predominant in the 2000's, followed by G9, G3 and G4 genotypes4). Recently, re-emergence of G9 and subsequent increase of G1 and G3 have been reported5). A classification system that incorporates all 11 gene segments has been proposed1), which would be useful for delineating the reassortment of genomes due to zoonotic transmission or immune selection pressure following the introduction of vaccination6). Although frequent detection of some genotypes such as G2 or G3 was has been reported after introduction of vaccines in some countries78), it is uncertain whether these findings were due to natural seasonal variation or immune pressure induced by mass campaign. RV vaccines have been introduced in Korea in June 2007 (RotaTeq® vaccine) and March 2008 (Rotarix® vaccine). The overall vaccination rate was 30% in 2009, and 34.1 % (Rotarix® vaccine) and 26.1% (RotaTeq® vaccine) in 2012910). However, the data for the change of RV genotypes after introduction of vaccines is still lacking in Korea. The purpose of this study was to investigate the epidemiology and genotypic change of RV following the introduction of vaccination in Korea during 2009–2013.

Stool samples were obtained from children with AGE who were treated at Sanggyepaik Hospital from August 2009 to June 2013. The initial identification of RV was carried out using an enzyme immunoassay (EIA) kit, Rotaclone® (Meridian Bioscience, Cincinnati, OH, USA) and RV-positive stool samples were included in this study. The study was approved by the Institutional Review Board of Sanggyepaik Hospital, Inje University, Seoul, Korea. Samples were obtained after receiving informed written consent from a patient's parents. The stool samples were diluted 1:10 (weight/volume) with phosphate-buffered saline, homogenized using a vortex mixer, and subjected to centrifugation at 5,000× g for 30 min at 4℃. Supernatants were recovered and stored at −70℃ until use.

Total RNA was extracted from stool samples using the QIAamp Viral RNA kit (Qiagen GmbH, Hilden, Germany), and RV sequences were then amplified via reverse transcription-polymerase chain reaction (RTPCR), using nested primers. RT-PCR was carried out on the 201 RV-positive samples in order to genotype the RV on the basis of VP7 and VP4 gene segment sequence. G genotyping was performed using a pool of different primers specific for G1-G4, G8, G9, G10, and G121112). The VP4 gene was characterized by using a pool of different P genotype-specific primers for P[4], P[6], and P[8]-P[10]13). An 881-bp fragment of the VP7 gene amplified in the first round of PCR, using VP7 consensus primers, was sequenced14). For PCR amplification of the VP4 gene, an 876-bp fragment was generated using the Con3 forward primer and Con2 reverse primer13). For G9-positive samples (G9P[4]: 10 cases, and G9P[8]: 3 cases), which were chosen by implying the status and amounts of remnant samples, characterization of other gene segments that encode VP1-3, VP6, and NSP1-5 was carried out via RT-PCR, using specific primers for each gene segment1). Amplified products of PCR reactions, which were randomly selected, were purified using the QIAquick® PCR Purification kit (Qiagen), and were then directly sequenced using the BigDye® Terminator Cycle Sequencing kit (version 3.1; Life Technologies, Carlsbad, CA, USA). Sequencing reaction products were resolved using an Applied Biosystems® ABI 3730 XL DNA analyzer (Life Technologies) . The nucleotide sequences of the VP7 and VP4 genes were aligned using the Clustal W, and phylogenetic trees were constructed using the MEGA software, version 4.014). We compiled the nucleotide sequences that were determined in this study with those available in the GenBank database by using the BLAST program (http://blast.ncbi.nlm.nih.gov/). The VP7 and VP4 sequences identified in this study have been submitted to the GenBank database under the accession numbers; VP7 from KF812568 to KF812615, and VP4 from KF812 616 to KF812663, respectively.

Categorical variables among study periods were compared using Chi-square test for trend. A P-value <0.05 was considered as statistically significant. All statistical analysis was performed by MedCalc version 15.6.1 (MedCalc software, Mariakerke, Belgium).

Between August 2009 and June 2013, the predominan t G genotype of human RV in Seoul, Korea successively changed. When compared to the results of our previous study6), it appears that the predominant G genotypes (G1, G3, and G9) naturally fluctuated, despite RV vaccination being introduced to Korea in 2007. Following the introduction of vaccination, a statistically significant decrease of the disease burden due to RV was reported in Korea16).

Interestingly, G1P[8] strains increased towards the end of this study. G1 lineage I was predominant throughout the study period and differences of amino acid residues at three positions (7-1a region) and one position (7-2 region) was revealed by comparison of VP7 antigenic regions of Korean G1 strains with both vaccine strains. Sublineage P[8]IIIA was most prevalent during the study period, consistent with previous findings in Brazil in which the definition of 1.5% cutoff value was proposed17). Continuous surveillance of the G1P[8] lineages are required to investigate any long term effect of vaccination on G1P[8] strain evolution.

combination genotype was observed and initially attributed to immune pressure following the introduction of the Rotarix® vaccine7), but it was later considered as a consequence of natural seasonal variation, in a study that showed frequent detection of the G2P[4] genotype in neighboring countries without a RV vaccination campaign1819). In Korea, the G2P[4] genotype was detected in 4.0–18.3% of cases between 2000 and 200920212223). In the present study, the G2P[4] genotype was detected in 4.9% of cases, which is similar to the incidence (4.0–4.3%) in other Korean studies that were carried out after the introduction of RV vaccination in 2007524). Frequent detection of the G3 genotype after the introduction of the RotaTeq® vaccine has been reported in Australia and the USA825). In the present study, the G3 genotype was mostly found in patients hospitalized in 2009/2010, but the detection rate of this genotype has fluctuated each year since the preintroduction of RV vaccine in Korea521). These results suggest that the prevalence of the G3 genotype might be unaffected by immune pressure following the introduction of RV vaccine.

In this study, the G4 genotype clustered in sublineage Ic, which was of interest because worldwide outbreaks due to this sublineage have been reported in several countries2627). The G12 genotype, a novel genotype recently detected in several countries, including Korea2528) was not found in the present study, however, another unusual genotype, G10, was found in three cases that occurred in 2011. This finding showed that re-assortment between human and bovine RV strains have occurred in Korea as in Africa29). The surveillance of animal RV strains is important because unusual strains could affect the diversity of human RV population. Collectively, our results for cases from the years 2009–2013 suggest that the emergence of unusual G genotypes due to vaccine pressure has not yet occurred in Korea, but it is necessary to remind that this study could not be the representatives of RV strain in the whole Korea during the same period.

In this study, circulating rotavirus isolates were highly diverse and mixed infections was most frequently found (26.7%) between August 2010 and July 2011. These results suggest the possible association between high rate of mixed infections of RV and the extent of strain diversity of circulating RV, which was consistent with those of a recent study30). In this study, 2.5% (5 cases) was non-typeable, which might be attributed to low concentration of virus particles or PCR inhibitors. The dominance of nontypeable isolates in adults with acute gastroenteritis was observed in a recent Indian study31).

Interestingly, the G9P[4] combination genotype was frequently detected in this study. Outbreaks of the unusual G9P[4] combination genotype have been reported in Mexico and Bangladesh3233), and were attributed to a possible reassortment in response to immune pressure. In Korea, the G9P[4] genotype was first identified in Gyeonggi province during 2003–2005 (detection rate: 1.7%)20), but this genotype was rarely found (range of detection rate: 0.04–0.7%) in later studies62124) prior to this study; these findings are similar to those of a recent study in Latin America18). In the present study, phylogenetic analysis of the G9 and P[4] genotypes showed that they clustered in lineages III and V, respectively. Further characterization of remnant gene segments in 10 G9P[4] isolates and three G9P[8] isolates revealed that all of them, except one isolate, belonged to the Wa-like genotype 1 constellation. These findings differ from those of a previous study in Mexico, which showed the predominance of the E6 genotype in the NSP4 gene33); the human-porcine reassortment of the NSP2 and NSP3 genes was not observed here, either. In the present study, we compared the amino acid residues in antigenic epitopes of VP7 and VP4 between the circulating isolates and the vaccine strains in Korea, which revealed less antigenic variation than in a comparable recent study in Bangladesh34). In a recent Finnish study35), there were no amino acid changes affecting the G1 genotype of VP7 and VP8^*, despite RotaTeq ® being exclusively used for vaccination after 2009, which suggests that there is no evidence of selection pressure. This study has some limitations: 1) lack of RV vaccination history and 2) using a primer amplifying about 40% of the VP4 segment. It had been suggested recently at least 50% of sequence were required from a segment for proper classification1).

In conclusion, periodic fluctuation of RV genotypes such as G1, G3, and G9 has been observed in Korea during the period 2009–2013, following the introduction of vaccination. Predominance of the unusual G9P[4] combination genotype in 2011 was initially considered to be evidence of immune pressure due to vaccination, but the subsequent re-emergence of G1P[8] in 2013 suggests the possibility of natural fluctuation of genotypes. In addition, we could not detect any significant differences in sublineages of either G9 or P[4], or in the entire constellation of RV gene segments. However , it is still necessary to continuously monitor and characterize RV strains, in order to detect novel strains that emerge due to the immune pressure associated with mass vaccination.