1. Introduction
Japanese encephalitis (JE) is the most significant vaccine-preventable neurological disease in Asia and the Western Pacific region [
1,
2]. The first confirmed cases of JE were detected in Japan in 1924, and it was subsequently reported in the rest of Asia, as well as in parts of Russia and Australia. At least 24 countries are currently considered to be at risk for JE transmission [
1,
3,
4]. Based on 2011 estimates, the global annual caseload of JE was approximately 68,000 cases, with at least 13,000 deaths [
5]. The global burden of JE is likely to increase as regions that were formerly less favorable to JE experience more hospitable conditions for its spread due to climate change [
6,
7,
8]. This is most evident in the recent JE outbreak in Australia, with 37 locally-acquired cases reported; the outbreak occurred 20 years after the last known human case [
9].
Japanese encephalitis virus (JEV), the causative agent of JE, is a positive-sense, single-stranded RNA virus of the Flaviviridae genus. JEV has five known genotypes (GI–GV) [
10,
11,
12,
13,
14], including two sub-genotypes of GI—GI-a and GI-b [
15]. The JEV genome is approximately 11 kilobases long and encodes ten proteins in an open reading frame flanked by two non-coding regions (NCRs) at the 5′ and 3′ ends. The ten proteins that are encoded are three structural proteins (capsid, membrane, and envelope) and seven non-structural (NS) proteins (NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5) [
16,
17].
JEV is transmitted by
Culex mosquitoes;
Culex tritaeniorhynchus is the primary vector [
18,
19], and other
Culex mosquitoes are secondary vectors. Other genera of mosquitoes, including
Aedes,
Anopheles, and
Mansonia, may act as regional or potential secondary vectors (reviewed by Pearce, et al. [
20]). The main reservoir hosts of JEV are birds, primarily of the family
Ardeidae, while swine act as amplifying hosts [
1,
21]. Humans are incidental, dead-end hosts [
2,
22]. Before 2000, GIII was the most prominent JEV genotype in all endemic areas, but it is currently being supplanted by GI in Asia [
23,
24,
25].
Prior to the abolishment of pig farming in 1992, Singapore had an average of 14 JE cases annually, but has had almost no cases since then [
26,
27]. However, the country has been identified as one of the 24 countries with a JE transmission risk [
3,
4]. A highly-urbanized island city-state, Singapore nevertheless retains much of its greenery and biodiversity through Nature Reserves (NRs), forest parks and urban parks, and mangroves [
28]. Its tropical climate, combined with its varied forest and urban habitats, provide a conducive environment for at least 180 species of mosquitoes, including
Cx. tritaeniorhynchus [
29]. The country is also home to wild boars [
30,
31,
32] and local wild birds. Additionally, migratory birds stopover in Singapore, which is located along the East-Asian Australasian Flyway (EAAF) [
33]. Ardeid birds are present in both the local (e.g., Grey Herons and Yellow Bitterns) and the migratory bird populations (e.g., Little Egrets, Chinese Pond Herons, and Black Bitterns) [
34].
A serological study that was carried out from 2010 to 2018, as well as an investigation of serum samples that were collected between 1996 and 1999, indicated JEV exposure in the wild boar and the wild bird populations in Singapore [
26,
35], suggesting that JEV may be enzootic in Singapore. Entomology-based surveillance that was initiated in 2011 detected five JEV-positive mosquito pools at a local Avian Sanctuary, further supporting this notion. The JEV sequences that were derived from those mosquito pools clustered with JEV GII isolates detected in Indonesia [
36]. The seroconversion of sentinel chickens during surveillance carried out from 2013 to 2014 further suggested active JEV transmission in the rural areas of Singapore [
35]. Although JEV had previously been detected in mosquito vectors in Singapore [
36,
37], virus isolation had not been successful [
36]. In the present study, we report the first isolation and detection of JEV GI from
Culex mosquitoes in Singapore. This study also describes the genetic characteristics of the newly detected JEV strains. The findings suggest that JEV has potentially established enzootic transmission locally, which has important implications on the risk assessment and vector control policies in Singapore.
2. Materials and Methods
2.1. Entomological and Virological Surveillance
Entomological surveillance and virological surveillance of JEV was carried out near-monthly; this occurred in two phases—a single-site phase from April 2014 to August 2016, and a multiple-site phase from September 2019 to May 2021. During the single-site phase, one surveillance site was set up at Park 1, which was in the north-eastern region of Singapore. During the multiple-site phase, four surveillance sites were set up at NR 1, NR 2-Park 2, Park 3, and Park 4 (
Figure 1). NR 2 and Park 2 were considered as one site due to their proximity to each other. These study sites were selected based on their proximity to slow-moving or still water bodies and the available information on the presence of JEV hosts, such as ardeid birds and wild boars, and the presence of JEV vectors. These selection criteria were adopted to maximize the possibility of capturing JEV-positive mosquitoes because of the lack of evidence on active transmission of JEV in Singapore.
Mosquitoes were trapped over two consecutive nights during the single-site phase, and for 18–20 h during the multiple-site phase. Each trapping session was conducted overnight, including crepuscular periods, using Centers for Disease Control (CDC) light traps baited with dry ice and incandescent light. Four to nine locations were selected at each surveillance site, and traps were positioned approximately 1.5–2.0 m above ground in shaded areas adjacent to water bodies where possible. Traps were set between 15:00 and 17:00 h and were collected the next day between 09:00 and 11:00 h.
Additionally,
Culex genus mosquitoes that were trapped from November 2020 to May 2021 as part of the malaria surveillance activities of the National Environment Agency (NEA) were also included in this study. The Malaria Surveillance Section (MSS) conducts on-going catches using the Night Catcher (a modified form of the CDC light trap [
38]) in 45-day cycles at 116 different locations in Singapore. Each MSS trapping session was carried out over a 12-h period overnight, including crepuscular periods. The two MSS trapping locations where JEV-positive pools were observed are indicated as MSS 1 and MSS 2 (
Figure 1).
2.2. Mosquito Identification
Mosquito specimens were transported on dry ice to the Environmental Health Institute (EHI), where they were sorted using mosquito identification keys [
39,
40,
41,
42,
43,
44,
45]. For the single-site phase, mosquito specimens were identified to species level where possible. For the multiple-site phase, mosquito specimens were initially identified to genus level, until the first detection of JEV, whereupon they were identified to species level where possible. In the case of morphologically similar species, the specimens were identified to the species group. Damaged specimens were identified to the genus level. For the start of the multiple-site phase, all female
Culex mosquitoes were pooled by trapping location and trapping month in pools of up to 100 specimens. Otherwise, females from six
Culex vector species—
Cx. bitaeniorhynchus,
Cx. gelidus,
Cx. quinquefasciatus,
Cx. sitiens,
Cx. tritaeniorhynchus, and
Culex spp.
vishnui subgroup (excluding
Cx. tritaeniorhynchus)—were pooled according to the species, trapping location, and trapping month, in pools of up to 100 specimens. All specimen pools were stored at −80 °C until further analysis.
2.3. RNA Extraction from Mosquito Specimens
Pooled mosquito samples were homogenized in 500 μL of universal transport media (Copan Diagnostics, Murrieta, CA, USA) using either the Mixer Mill MM 400 (Retsch Technology GmbH, Haan, Germany) or the 1600 MiniG homogeniser (Spex SamplePrep, Metuchen, NJ, USA), and centrifuged at 5000× g for 5 min. Viral RNA was extracted from the mosquitoes collected during the single-site phase using 140 μL of the homogenate and the QIAamp Viral RNA mini kit (Qiagen Group, Hilden, Germany), according to the manufacturer’s instructions. For mosquitoes collected during the multi-site phase, as well as the mosquitoes obtained from MSS, viral RNA was extracted from 200 µL of mosquito homogenate using the IndiMag Pathogen kit (Indical Bioscience GmbH, Leipzig, Germany) and the KingFisher Flex system (Thermo Fisher Scientific Inc., Waltham, MA, USA), as per manufacturer’s instructions. Eluted RNA and the remaining homogenates were stored at −80 °C until further use.
2.4. Detection of JEV by Reverse Transcriptase-PCR (RT-PCR)
Extracted RNA was screened for JEV, as described in Yap, Mailepessov, Lim, Chan, How, Humaidi, Yeo, Chong, Lam-Phua, Lee, Okumura, Vythilingam and Ng [
36]. Briefly, positive RNA controls of JEV (Nakayama strain) were extracted from Vero cell (ATCC CCL-81; ATCC, Manassas, VA, USA) supernatants using a QIAmp Viral RNA Mini kit (Qiagen), as described above. JEV was detected by using a modified RT-PCR assay adopted from Scherret, et al. [
46] and Santhosh, et al. [
47]. All reactions were performed using the QuantiTect
® SYBR
® Green RT-PCR kit (Qiagen) in a final volume of 20 µL, with 5 µL of extracted RNA and 10 µM of each primer. The cycling conditions were as follows: reverse transcription at 50 °C for 20 min and inactivation at 95 °C for 15 min; 40 cycles at 94 °C for 15 s, 58 °C for 30 s, and 72 °C for 30 s; melting curve analysis was conducted from 65 to 95 °C, with a slope of 0.1 °C/s, for validation of the amplified products; cycling conditions concluded with a cooling step at 37 °C for 20 s.
2.5. Sequencing and Analysis of the Envelope Gene of JEV
The envelope (
E) gene of JEV was amplified using two pairs of primers, as described in Schuh, et al. [
48]. Complementary DNA was first synthesized using the Maxima H Minus First Strand cDNA synthesis kit (Thermo Fisher Scientific Inc.), and subsequent amplifications were performed using Phusion Flash High-Fidelity PCR Master Mix (Thermo Fisher Scientific Inc.), as per the manufacturer’s recommendations. The amplified products (1700 and 800 bp) were purified using the FavorPrep GEL/PCR Purification Kit (Favorgen, Ping Tung, Taiwan) before sequencing at a commercial sequencing facility using the BigDye terminator Cycle Sequencing Kit protocol (Applied Biosystems, Thermo Fisher Scientific Inc., Waltham, MA, USA).
Raw sequence data were analyzed as described in Yap, Mailepessov, Lim, Chan, How, Humaidi, Yeo, Chong, Lam-Phua, Lee, Okumura, Vythilingam and Ng [
36]. A phylogenetic tree was constructed using 129 truncated (701 bp) JEV
E-gene sequences; this included the newly detected JEV sequences in the present study (n = 10) and 119 JEV sequences that were retrieved from the NCBI GenBank database (
Table S1). The sequences were aligned using MUSCLE in MEGA version X 10.2.5 [
49], and a phylogenetic tree was constructed using the maximum likelihood (ML) method. The best fitting model, which was the Tamura-Nei model [
50] with gamma-distributed substitution sites (TN93+G), was determined via the model selection tool in MEGA version X 10.2.5. The tree was constructed using TN93+G with 1000 bootstrap reiterations and rooted with the truncated (701 bp)
E-gene from Murray Valley Encephalitis virus (MVEV; NCBI Reference Sequence no. NC_000943.1). Amino acid sequences were also obtained via MEGA version X 10.2.5. Sequence identity matrices were generated with the MUSCLE tool in EMBL-EBI with default parameters [
51,
52]. The newly generated
E-gene sequences from this study were deposited in the GenBank nucleotide database (GenBank Accession Nos. KT346354, MW802631, MW802633 to MW802637, and ON804794 to ON804796).
2.6. JEV Isolation
Mosquito homogenates that were positive for JEV were filtered (Minisart 0.45 microns; Sartorius, Göttingen, Germany) before inoculation onto C6/36 cells (ATCC CRL-1660; ATCC) at 90% confluency. The C6/36 cells were incubated in Leibovitz L-15 media (Thermo Fisher Scientific Inc.) with 2% fetal bovine serum (FBS) (Thermo Fisher Scientific Inc.) at 33 °C and monitored for one week. Approximately 1/5 of cell culture suspension was then transferred to freshly prepared 90% confluent C6/36 cells, topped up with fresh 2% FBS Leibovitz L-15 media, and incubated again at 33 °C for one week. This step was repeated one more time for a maximum of three passages. If cytopathic effects or syncytia were observed, the cells were harvested and kept at −80 °C.
2.7. Sequencing and Phylogenetic Analysis of the Whole Genomes of JEV
Viral RNA was extracted from cell culture suspension using the QIAmp Viral RNA Mini kit (Qiagen), according to manufacturer’s instructions. Multiple amplicons covering the whole genome were then synthesized using the primers in
Table S3 and a protocol similar to that of
E-gene amplification, except with an annealing temperature of 60 °C. The amplification products were purified with the FavorPrep GEL/PCR purification kit (Favorgen) and sequenced at a commercial sequencing facility using the BigDye terminator Cycle Sequencing Kit protocol (Applied Biosystems, Thermo Fisher Scientific Inc.).
Raw sequence data were trimmed and assembled using SeqMan Pro 15.3.0 from the Lasergene software suite (DNAStar Inc., Madison, WI, USA) and BioEdit 7.2.5 to obtain consensus sequences. Where a consensus could not be reached for a sequence call in the contig, the position was represented by the appropriate IUPAC ambiguous base symbol.
A whole-genome phylogenetic tree was constructed using 50 JEV sequences; 3 sequences were generated during the present study and 47 sequences were obtained from the NCBI GenBank database (
Table S4). The sequences were aligned in MEGA version X 10.2.5; as above, the best fit model was determined to be the Tamura-Nei model, but with invariant and gamma-distributed substitution sites (TN93+G+I). The tree was constructed using TN93+G+I with 1000 bootstrap iterations and rooted with MVEV (NCBI Reference Sequence no. NC_000943.1). The amino acid translation was carried out via the Translate tools from Expasy [
53] and the Sequence Manipulation Suite [
54]. Sequence identity matrices of whole genome data sets were also generated as above. The whole genome sequences that were generated during the present study were deposited in the GenBank nucleotide database (GenBank Accession Nos. ON804797 to ON804799).
2.8. Mutation Analyses
Amino acid mutations that were unique to the study sequences were determined by comparing them with all of the JEV GI complete coding sequences available in the NCBI database (n = 158) as of 03 November 2022. Genotype I sequences were selected by constructing a phylogenetic tree consisting of all complete coding sequences of JEV (n = 404) available in the NCBI database. A consensus sequence of the GI complete coding sequences (n = 158) was generated using the PAM40 matrix with the
cons module from the EMBOSS suite [
55].
4. Discussion
All of the JEV sequences (n = 10) generated during the present study from the mosquito pools collected from 2014 to 2021 belonged to GI. This is in contrast to the detection of JEV belonging to GII locally in 2011 [
36]. This observation provides further empirical evidence for the establishment of genotype I as the predominant JEV genotype in Asia [
23,
24,
25,
59]. Nine of the ten new JEV
E-gene sequences belonged to sub-genotype GI-a, which is strongly associated with the tropics, and is often found in South-east Asia [
15]. The remaining sequence (SG/EHI-CT1372) belonged to sub-genotype GI-b, which is also known as a temperate sub-genotype, and is typically found in East Asia [
15].
The nine GI-a sequences formed two clusters (cluster A and cluster B,
Figure 2). Both of these clusters were closely related to JEV sequences from South-east Asian countries located to the north of Singapore. This suggested that local JEV strains were likely to be introduced by migrating ardeid birds via the EAAF, which stretches from Russia and Alaska to Australia and New Zealand through East Asia and South-east Asia.
Cluster A consisted of five sequences detected in
Culex mosquitoes within the same area during the same period (
Table 2). However, no subsequent instances of JEV belonging to cluster A were observed in the following years, despite continued surveillance in the same location. This suggested that cluster A could not sustain transmission in the local ecological setting and became extinct, similar to what was observed with JEV strains “B’/B/Ishikawa type” and “C’/C/Ishikawa + Kagawa type” in Toyama, Japan by Obara, et al. [
65]; this extinction could either have been due to stochastic reasons or non-favorable virus characteristics for its survival in a new environment. In contrast, the detection of cluster B four times from 2014 to 2021 may suggest its persistent presence through an enzootic transmission cycle involving the fauna in the northern and north-eastern parts of Singapore (
Figure 1 and
Figure 2,
Table 2). Although there have been no cases of JE in Singapore for almost two decades [
27], previous studies, including serological evidence of JEV infection in the local and migratory birds, and the wild boars, as well as seroconversion in sentinel chickens, have shown that JEV continues to be present in the animal populations in Singapore [
26,
35,
36] and suggest that it may currently be enzootic in Singapore. Alternatively, cluster B JEV strains could have been imported on repeated occasions by migratory birds. These assumptions appear to support the observations that were made by Kuwata, et al. [
66] on the importance of vector and/or host movement in transferring JEV strains between different regions and the maintenance of JEV strains in the local fauna within the same region over a period of time.
The detection of different variants of JEV suggested that multiple incursions of JEV have occurred in Singapore, not only over the course of this study (2014–2016 and 2019–2021), but also during the period that was studied previously [
36]. Besides the phylogenetic analyses, the mutation profiling also supported this notion. For example, uncommon mutation profiles suggested that SG/EHI-CT1372 was distinct and more likely to have originated from a different geographical area than SG/EHI-MS_CT261 and SG/EHI-CX135.
Although JEV has only been detected in the rural forested areas of Singapore thus far, increasing urbanization and the push for ecological conservation [
67,
68] result in close interactions between human dwellings and the habitats that are favored by culicine vectors and JEV hosts. A survey of bloodmeals of
Culex mosquitoes in Singapore showed that JEV vectors opportunistically feed on humans [
69], while a study of changing JEV epidemiological characteristics in metropolitan areas showed an apparent correlation between the presence of herons and the number of JE cases in the absence of pig farming [
70]. It is thus necessary to maintain entomological and virological surveillance to keep track of possible enzootic JEV variants, thereby allowing for more leeway in containing any vector–human transmission events.
Besides gaining insights into its molecular epidemiology, the genetic characterization of JEV is also important for understanding virus fitness and virulence. Both the coding and non-coding regions of JEV are important for its replication and virulence [
16,
17]. While the structural and non-structural proteins provide the backbone of virion structure and replication machinery, 5′- and 3′-NCRs are also involved in genome circularization and host immune response attenuation. Genome circularization, which involves the stem loop structure SLA, is crucial for flavivirus replication, and exoribonuclease-resistant RNA (xrRNA) structures are able to mitigate the degradation of flavivirus RNA [
56,
71,
72,
73].
Among the substitutions that were detected in the study sequences (
Table 6 and
Table 7, and
Table S5), two non-conservative substitutions E-T226A (T520A) and E-T231A (T525A) – detected in SG/EHI-CT1372 – reside within domain I, which is involved in E protein stability [
74,
75,
76]. An E-T226R substitution was also shown to increase JEV neurovirulence and neuroinvasiveness in mice [
77]; therefore, it may be useful to look into any effects of E-T226A on JEV virulence. Two E protein substitutions were also found at the neutralizing epitope position E-126 (residue 420 in the polyprotein) [
76] across the study sequences in cluster A (E-I126V;
Figure 2) and across the study sequences in cluster B (E-I126T;
Figure 2) – these substitutions may allow for immune escape. However, no substitutions were detected in the study sequences at the E protein positions commonly found to be important for JEV neurovirulence and/or neuroinvasion (
Table 6). Another important substitution was M-D40H (D167H), which was detected in SG/EHI-MS_CT261; a 40-residue region in flavivirus membrane (M) protein induces host cell apoptosis [
78], and an M-I36F substitution within that region impeded JEV virion assembly in mammalian cells – likely due to the impairment of M protein apoptosis activity [
79]. SG/EHI-MS_CT261 also carried an NS4b-A11T (A2283T) substitution that falls within the LIG-WD40 domain of the NS4b protein, and a mutant involving the LIG-WD40 domain has been shown to induce apoptosis and encephalitis in mouse brains [
80]. Positive selection at the serine residue 24 of NS4b has been reported in genotype III of JEV that was isolated from pigs [
81]. NS4b-P24S (polyprotein residue 2296) in SG/EHI-CX135 therefore warrants further investigations into its potential adaptation to swine—in particular, in the wild boar (
Sus scrofa) as is found in Singapore. The substitution NS5-K68R (K2595R) in SG/EHI-MS_CT261 resides in the N-terminal 83-residue region of NS5, which is required for blocking IFN-α signaling [
82]. However, NS5-K68R is a conservative change, and is less likely to affect the protein structure. Lastly, the most significant 3′-NCR mutations in the three whole genome sequences generated during this study were in pseudoknot 1 (PK1) of xrRNA1 (
Table 5). PKs have been shown to form a mechanical block to exoribonuclease 1 by wrapping around the 5′ end of the xrRNA [
83,
84,
85,
86], hence mitigating host immune response [
71,
72,
87]. Further research is needed to determine whether these mutations are important for adaptations to the local ecological niche.
The present study had several limitations. Firstly, the pooling of mosquitoes for screening might have confounded the detection of JEV in multiple mosquitoes of the same pool. Unfortunately, given the large number of specimens, it was not manpower- or cost-effective to screen individual mosquitoes. Secondly, the present study may not reflect the true prevalence of JE in Singapore due to the limited number of study sites, as the primary focus of this study was not to estimate the prevalence of JE in Singapore, but to determine whether there was any evidence of enzootic transmission of JEV in the country. Thirdly, although all sampling sites were selected based on the presence of ardeid birds, wild boars, water bodies, and vectors in their respective localities, vector control activities were strongly enforced in Park 2, 3, and 4 study sites (
Figure 1) due to the proximity to human activity. This could affect the trapping of the mosquitoes at those sites. Lastly, the use of CDC light traps may have biased specimen collection. Lord, et al. [
88] found that
Cx. tritaeniorhynchus was oversampled with light traps in contrast to
Cx. pseudovishnui and other potential JEV vectors and posited that this was due to the differences in the biting behaviors and host-seeking strategies rather than being indicative of the actual population density.