Background

Japanese encephalitis virus (JEV) is a mosquito-borne member of the family Flaviviridae, genus Flavivirus, and a primary cause of viral encephalitis in humans within its range [1]. The positive-sense RNA viral genome is approximately 11 kb in length and is translated into three structural proteins [Capsid (C), Membrane (M), and Envelope (E)] and seven nonstructural (NS) proteins (NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5) with untranslated regions (UTR) at the 5' and 3' ends of the genome [2]. Historically, Culex tritaeniorhynchus was implicated as the primary vector of JEV in the Republic of Korea (ROK) and much of Asia [3, 4]. However, JEV has since been detected in additional culicine species throughout its range, including Cx. bitaeniorhynchus from the ROK [5]. JEV strains are generally classified into five genotypes (genotypes I, II, III, IV, and V) based on similarities in the E gene nucleotide sequence [6]. Previously, only genotype I was detected on the Korean peninsula [7]. Therefore, we characterized JEV-positive pools of Cx. tritaeniorhynchus and Cx. bitaeniorhynchus to determine whether the unexpected finding of JEV in Cx. bitaeniorhynchus in the ROK may have coincided with the appearance of an additional genotype.

Materials and Methods

Nineteen JEV-positive mosquito pools, from a total of 64 JEV-positive pools collected during 2008 and 2010 in the ROK (18 pools of Cx. tritaeniorhynchus and 1 pool of Cx. bitaeniorhynchus), and one JEV culture received from USAMRIID (United States Army Medical Research Institute of Infectious Diseases, USA) were genotypically characterized (Table 1, Figure 1). Total RNA was extracted from mosquito homogenate using Trizol reagent (Invitrogen, USA) in accordance with the manufacturer's instructions and was resuspended in 50 μl of RNase-free water containing 10 units of RNasin® Plus RNase Inhibitor (Promega, USA). RNA was used as the template for cDNA synthesis using the SuperScript III first strand synthesis system (Invitrogen, USA) with a random hexamer primer. The synthesized cDNA was then used for PCR amplification using iProof™ High-Fidelity DNA polymerase (Bio-Rad, USA). The NS5 gene/3' UTR and envelope (E) gene of 19 JEV-positive pools were amplified using EMF1/VD8 primers [8] and 940S/1720A primers [9], respectively. Products were purified using the QIAquick PCR purification kit (Qiagen, USA) and sequenced by AITBiotech Company (AITbiotech, Singapore).

Table 1 Locations and collection dates of JEV-positive mosquito pools analyzed in this study
Full size table
Figure 1
figure 1

Locations of JEV-positive mosquito pools collected during 2008 and 2010 in the Republic of Korea. Daeseongdong is a village near the military demarcation line (MDL) (center of the 4-Km wide demilitarized zone separating North and South Korea); Warrior Base training area is approximately 5 km north of Munsan; Camp Humphreys is in a rural area of Pyeongtaek; Gunsan Air Base is located near the small city of Gunsan; Gwangju Air Base is located near the metropolitan city of Gwangju; Haenam, Jinju, and Changnyeong sites are beef/swine farms near the small cities. Pool = number of sequenced samples/total JEV-positive samples.

Full size image

The sequences were edited and assembled using the Sequencer program v4.1.4 (Applied Biosystems, USA). Multiple sequence alignments and phylogenetic analysis were performed using ClustalX version 2.0.11 and MEGA version 5 programs [10, 11]. Percent sequence similarity/divergence was calculated using the MegAlign program found in the Lasergene v.8 software (DNASTAR, Inc., Madison, WI, USA). Phylogenetic analysis of the partial E gene (705 bp) was performed using the neighbor-joining method and Tamura-Nei model of nucleotide substitution. The maximum likelihood (ML) tree was constructed from the NS5/3'UTR nucleotide sequences (550 bp) by PhyML software v 3.0 [12] using the best fit model with aLRT branch support [13]. The ML tree for the complete E gene used the Tamura-Nei model with bootstrap analysis (2, 000 replicates) for testing the reliability of the tree using the MEGA5 (version 5) program (The Biodesign Institute, Tempe, Arizona) [11].

Results

The phylogenetic relationships among 19 JEV strains and JEV sequences retrieved from GenBank representing genotypes I-V were analyzed. The ML tree for the NS5/3'UTR (550 bp) and the neighbor-joining tree for the partial E gene (705 bp) showed similar branching patterns with high bootstrap support. Therefore, the ML tree is only presented in this report (Figure 2). Two genotypes were identified among the 19 JEV strains. JEV strains from 18 Cx. tritaeniorhynchus mosquitoes grouped into genotype I. These genotype I strains were closely-related to strains isolated from China, Korea, Japan, Vietnam, and Thailand from the early 1980s to the present (Figure 2). The remaining strain from Cx. bitaeniorhynchus (10-1827) grouped into genotype V together with the Muar strain which was isolated from an encephalitis patient in Malaysia in 1952.

Figure 2
figure 2

Maximum likelihood tree of JEV strains from the ROK using NS5/3'UTR base-sequence homologies. Phylogenetic analysis was performed using the GTR+G4+I model of nucleotide substitution (ln(L) = -3090.1 571) and aLRT branch support (indicated at major nodes). The phylogenetic tree has been rooted at its midpoint. Scale bar represents substitutions per site.

Full size image

The complete E gene was sequenced from a subset of strains in genotype I (A10.825, A10.881, A8.789) and genotype V (10-1827). The ML tree constructed from the complete E gene of these strains together with representative JEV genotype I-V sequences is shown in Figure 3. This ML tree supports the phylogenetic analysis results performed on the NS5/3'UTR (Figure 2) and the partial E gene previously mentioned. The ML tree in Figure 3 shows that the 10-1827 strain grouped with the Muar strain with 79% bootstrap support, while the remaining sequences clustered in genotype I together with K01-JN and K05-GS strains that were isolated from Cx. tritaeniorhynchus in the ROK in 2001 and 2005, respectively.

Figure 3
figure 3

Maximum likelihood tree of JEV strains from the ROK using complete E gene nucleotide sequence homologies. Phylogenetic analysis was performed using the MEGA5 program by the Tamura-Nei model of nucleotide substitution (ln(L) = -7791.2). Murray Valley encephalitis virus strain (MVE-1-51) was used as an outgroup. Bootstrap values at each major node were calculated using 2, 000 replicates. Scale bar represents the number of nucleotide substitutions per site.

Full size image

Sequence analysis of 18 strains shows minimal sequence variation among viruses in genotype I, with nucleotide sequence similarity of 97.5-100% for the NS5/3'UTR (Figure 2) and 99.6-100% for the E gene (Figure 3). In an earlier study, genetic stability was also observed among JEV strains isolated from mosquito vectors in the ROK between 1994 and 2005 [7]. Examination of the complete E sequence of 10-1827 strain (genotype V) showed less similarity to the other genotypes, with nucleotide similarity approximately 77.3% (91.3% for amino acids) to genotype I (K01-JN, K05-GS), 78.1% (91.0% for amino acids) to genotype II (FU strain), 77.7% (90.4% for amino acids) to genotype III (Nakayama), and 77.8% (91.0% for amino acids) to genotype IV (JKT6468) (Table 2). However, nucleotide and amino acid similarities to the Muar strain were 90.0% and 98.8%, respectively (Table 2). Likewise, the XZ0934 strain, a JEV genotype V recently isolated from China (2009), showed E gene nucleotide and amino acid sequence similarities to the Muar strain of 86.0% and 93.2%, respectively [14].

Table 2 Nucleotide sequence similarity and divergence of the complete E gene from ROK mosquito pools
Full size table

Figure 4 shows the amino acid sequence alignment of the complete E gene derived from strains A10.825, A10.881, A8.789, and 10-1827 and reference sequences (Muar, K01-JN, K05-GS). The E protein of the strains in genotype I is very conserved with few amino acid changes detected: A10.825 (from S = serine to N = asparagine at position 123) and A8.789 (from L = leucine to M = methionine at position 371). The alignment reveals differences in 6 amino acid residues between the Muar and 10-1827 strains (Figure 4). The eight Muar signature amino acid residues in domain III comprising a putative receptor binding region [15] were also identified in the 10-1827 strain along with the critical amino acid residue thought to be involved in receptor binding activity (Q = glutamine at position 327) [16]. Table 3 provides a complete listing of the strains that are referenced in this study.

Figure 4
figure 4

Amino acid sequence alignment of the full-length envelope gene from ROK JEV strains. ROK mosquito pools collected during 2008 and 2010 were aligned with the reference sequences GV (Muar) and GI (K01-JN, K05-GS). Dots indicate consensus. Differences in amino acids between Muar and 10-1827 strains are underlined. Residues enclosed by boxes represent 8 Muar signature amino acids in domain III.

Full size image
Table 3 Origin of 30 JEV strains referenced in this study
Full size table

Conclusion

This study is the first report of JEV genotype V in the ROK and represents the third report of genotype V in Asia, with the most recent findings from Cx. tritaeniorhynchus collected in Tibet, China (2009) [14]. The fact that JEV genotype V, first reported from an encephalitis patient in Malaysia in 1952 (Muar strain), came long before the discovery of its reemergence in China in 2009 and now its subsequent appearance in the ROK may mark the beginning of a genotypic shift in JEV within the region. Additionally, the emergence of this strain in Cx. bitaeniorhynchus, a mosquito species previously unknown to carry JEV in the ROK, underscores the need to step-up surveillance efforts within the ROK. The reemergence of this genotype after 57 years may have future implications with regard to JEV vaccination effectiveness and policy among civilian and military populations, as well as with preventive strategies designed to reduce the health impact and incidence of JEV among at risk Asian populations.