AN ALTERNATIVE CELLULAR MODEL FOR GENETIC VARIATIONS OF JAPANESE ENCEPHALITIS VIRUS IN THE TRANSMISSION PROCESS BETWEEN VERTEBRATES AND MOSQUITOES*

HOU Wen-jun JIANG Yu-ting ZHANG Heng-duan LI Chun-xiao XING Dan GUO Xiao-xia** ZHAO Tong-yan**

(1.Beijing Key Laboratory of Vector Borne and Natural Infectious Disease, State Key Laboratory of Pathogen and Biosecurity, Institute of Microbiology and Epidemiology, AMMS，Beijing 100071, China；2.Research Institute for Medical Prevention and Control of Public Health Emergency, Characteristic Medical Center of Armed Police，Beijing 102613, China)

Abstract To simulate alternative microenvirmonental habitats in vertebrates and mosquito hosts, an in vitro model was utilized to determine the gene variations′ pattern of the Japanese encephalitis virus (JEV). Different JEV-SA14 mutants were obtained and compared after sequenced, and the results showed that variable loci had been found in JEV during passaging in different microenviroments. All mutations were found single-based without any insertions or deletions, which suggests that the highly conserved evolution of JEV under the natural selections during viral transmission. The fitness of JEV to different hosts evaluated were shown comparable to previous studies along with corresponding genetic mutations. The in vitro alternative passaging protocol appeared to be reliable to simulate the variant infection microenviromental habitats between vertebrated hosts and vectorial mosquitoes.

Key words JEV；Alternative passage；Genetic variation；Fitness evaluation；Cell model

Japanese encephalitis (JE) is a mosquito-borne infectious disease with a disability rate of 20% ～ 40%, which was caused by the Japanese encephalitis virus (JEV) (Zhengetal., 2012; Chenetal., 2015; Connoretal., 2017). As the SA14-14-2 attenuated vaccine was incorporated into China′s national immunization program since 2007, it has been widely implemented in many Asian countries to decrease the incidence of JE (Zhengetal., 2012; Chenetal., 2015). However, JE is still a significant cause of viral encephalitis, severely threatening human health in most Asia and western Pacific parts (Zhengetal., 2012). There are about ten thousands JE cases and resulting in more than one thousand deaths annually worldwide (Chenetal., 2015).

Molecular epidemiological surveillances had provided timely detections of viral mutations and changes in virulence, antigenicity, and fitness (Panetal., 2011; Tengetal., 2013; Zhengetal., 2013; Suetal., 2014; Phametal., 2016). For example, Phametal.and Panetal.reported that the prevalence of JEV genotype Ⅰ(GⅠ) has increased significantly in recent years (Panetal., 2011; Phametal., 2016). Meanwhile, more than one attempts have been tried to characterize the genetic changes of JEV in the transmission cycles between vertebrate hosts and vector mosquitoes (McCurdyetal., 2011; Yangetal., 2014; Gromowskietal., 2015; Doetal., 2016; Huangetal., 2016). In 2014, Yangetal.found that genetic mutations were significantly increased when the SA14-14-2 strain was passaged successively in Vero cells compared to those in mouse brains (Yangetal., 2014). Similarly, McCurdyetal.discovered that genetic changes of the Nakayama strain occurred within just five passages in vivo culture (McCurdyetal., 2011). Most previous studies were imperfectual due to the limited one single host or one stable micorenviroment, fail to elucidate the JEV genetic variation patterns under selections by dual-hosts (McCurdyetal., 2011; Yangetal., 2014; Gromowskietal., 2015; Doetal., 2016; Huangetal., 2016; Oliveiraetal., 2018; Xiaoetal., 2018; Fanetal., 2019). As a matter of facts, JEV was tramsimitted within alternative microenviroments with vertebrate hosts and vector mosquitoes involved. Aninvitromodel comprised with cell lines from vertebrate hosts and mosquitoes and the repeated transferring processes of JEV was expected to produce variation patterns mostly match the natural cycles observed (Ciotaetal., 2013).

To simulate the natural cycle of JEV transmission between mosquito vectors and vertebrate hosts and thereby more closely agree with natural patterns of genetic variations and possible compensative fitness in JEV drived by natural selections from transmission cycles between verterbated host and vector mosquitoes, we set up an alternative passaging model between mosquitoes and mice for genetic studies of JEV in 2016 (Houetal., 2017). By more closely mimic the natural process of JEV transmission, the dual-host experimental model can provide an soundable theoretical basis for the molecular epidemiological surveillance of JEV (Ciotaetal., 2010). However, this model requires mosquito breeding and mice brain inoculation, which is labour-intensive and expensive (Houetal., 2017). The cell passaging model can effectively avoid these troubles, and reducing the unnecessary animal usage. Therefore, we alternately passaged the JEV-SA14 strain between mosquito C6/36 cells and mammalian BHK-21 cells to estible aninvitrodual-host model for laboratory studies of JEV.

1 Materials and Methods

1.1 Virus

The JEV-SA14 strain was provided by the viral laboratory of the Institute of Microbiology and Epidemiology of AMMS, which was initially isolated from mosquito specimens in Xi′an City, Shaanxi province, China and was subcultured 20 times in Kunming neonatal mice before stored in liquid nitrogen.

1.2 Cells cultures

TheAedesalbopictuscell line C6/36 was cultured in RPIM 1640 (Gibco, USA) culture medium supplemented with 10% heat-inactivated fetal bovine serum (FBS, Gibco, USA) and incubated at 28 ℃ without CO2. Baby hamster kidney (BHK-21) cells were maintained in Dulbecco′s minimal essential medium (DMEM, Gibco, USA) with 10% heat-inactivated fetal bovine serum, 1% penicillin and streptomycin (P/S, Gibco, USA), 0.5% L-Glutamine (Gibco, USA) and incubated at 37 ℃ with 5% CO2.

1.3 Alternative passage of the JEV-SA14 strain

The alternative passage was constructed between C6/36 (mosquito cell) and BHK-21 (mammalian cell) cells for 14 cycles, which is equivalent to the JEV transmission process in nature for more than two years, and the descendent strains at each cycle were named C1B1 for cycle 1, C2B2 for cycle 2, etc.

The cells were infected with JEV at an MOI of 0.01 and incubated at 37℃ for 1 hour. The cells were washed twice with phosphate-buffered saline (PBS) and then cultured at 37℃ after the virus was removed. When the cytopathic effect covered 75% of the monolayer, the cells were frozen at -80℃ and thawed at 37℃ three times. The supernatant was separated by centrifugation and frozen at -80℃ to serve as the source of subsequent virus inoculation, thereby completing a single alternating passage cycle.

1.4 Viral titrations

Following each passage, the JEV supernatant was quantified by a plaque assay on BHK-21 cells. 400 μL of serial diluted JEV supernatant (10-1～10-8) in DMEM containing 2% FBS was added to a monolayer of BHK-21 cells in 12-well plates. The cells were incubated at 37℃ with the viral supernatant for 1 hour, and the slides were rocked every 15 mins. After the virus was removed, each cell monolayer was coated with 2% soft agar (3 mL/well) and incubated for six days. Subsequently, cell monolayers were fixed with 4% formalin for 30 mins and stained with 1% crystal blue. The plaques were counted to calculate the viral titer.

1.5 Genomic sequencing

The SA14 parent and descendent strains were obtained and sequenced from every second alternative passage cycle as follows. Viral RNA was extracted from 200 μL of cell-free supernatant with 600 μL of TRIzol (Invitrogen, USA) according to the manufacturer′s protocol. RNA was precipitated with isopropanol, and centrifugated at 12 000 g for 15 min and then the pellets were washed once with precooled 75% ethanol. After air-drying, RNA was dissolved in 50 μL of RNase-free water and reverse transcripted with M-MLV Reverse Transcriptase from TaKaRa Ltd., Japan, followed by 35 rounds of amplification in a 50 μL reaction volume using kits from TaKaRa Ltd, Japan. The amplicons were purified from 1% agarose gels and submitted for sequencing by bidirections. The PCR primers were designed based on the SA14 sequence (GenBank accession: M5-5506) and are shown in Tab. 1.

Tab.1 Primers used to amplify JEV nucleotide sequences

1.6 Analysis of coding region sequence

Nucleotide and deduced amino acid sequences were analyzed and compared with the representative JEV strains (Tab. 2) deposited in GenBank using the MegAlign Pro software from the DNAStar package (Ver. 7.0). The phylogenetic tree was conducted using the neighbour-joining method of MEGA 5.0 software. The reliability of the clustering was assessed by the bootstrap test (1 000 replicates).

Tab.2 JEV sequences used for phylogenetic analysis in the present paper

表2续 Tab.2 Continued

2 Results

2.1 Cell passages and virus titer analysis

Cell pathological effects (CPE), such as shrinkage, were generally observed in C6/36 and BHK-21 cells and covered 75% of the monolayer about three days post-infection (Fig. 1). The mean value of the viral titer ranged from 7.3 to 8.4 log10PFU/mL, with an average value of 7.86 ± 0.09 log10PFU/mL. The titers of passage C2B2 and C6B6 were the highest, while the titer of passage C8B8 was the lowest. The viral titer varied among different passages but no significant trend founded (Fig. 2).

Fig. 1 Cytopathic changes in C6/36 and BHK-21 cell lines after inoculation with JEV

Fig. 2 The viral titers of the JEV-SA14 strains after alternative passages

2.2 Nucleotide mutations inE, prM and NS1 genes of JEV

Eight JEV-SA14 nucleotide sequences, including one parental strain and seven passage strains, were deposited in GenBank with the following accession numbers: KU871316, KU821122, KU871317, KU871318, KU871319, KU871320, KU871321 and KU871322. By comparing the seven viral genome sequences with the parental SA14 genome, we found 39 mutations at 36 sites, with no insertions or deletions. Among these mutations, 97.4% were transition mutations, and 2.6% were transversion mutations. The most common mutation was C→T (28.2%, 11/39) (Tab. 3). The mutations were scattered throughout the ORF, and a majority of the mutations were located in the E (41.7%) and NS1 (19.4%) regions (Tab. 3).

Tab.3 Statistics on nucleotide mutations found in 7 passaged SA-14 strains

2.3 Deduced amino acid mutation analysis

The protein deduced from the nucleic acid sequences were analyzed, and 11 amino acid mutations at 11 sites were found compared with the parental sequence (Tab. 4), which included 61.54% (24/39 nucleotide mutations) synonymous substitutions (dN/dS=0.625) (Tab. 4). The mutations′ highest rate (45.45%, 5/11 sites) was found in the E protein, significantly higher than that of other fragments. Notably, point mutations of I176 T and V340I in the E protein were located in regions related to antigenicity and virulence.

Tab.4 Deduced amino acid mutations and distribution of 7 passaged SA14 strains

2.4 Sequence homology alignment and evolution analysis

By comparing all the seven descendent sequences with the parental JEV-SA14 sequence, we found that the JEV viral genome showed a high homogeneity (98.2%-99.9%) among all passages, with a corresponding protein identity of 99.5%-99.9%. The passages C4B4 and C10B10 showed the nucleotides′ highest and lowest homogeneity to the parental sequence.

As expected, all the seven descendent sequences and the parental JEV-SA14 sequence were aligned with type III JEV genome sequences available in GenBank (Tab. 2) for phylogenetic analysis. Either at the nucleotide level or the amino acid sequence level, these sequences showed a high degree of homogeneity to type III JEV sequences, with homology rates ranged from 96.4%-99.9% for the nucleotide sequence and 98.1%-99.9% for the amino acid sequence (Fig.3) respectively. Remarkably, the parental SA14 and the passage C2B2 demonstrated a close evolutionary relationship to the strains CH13 (Sichuan, China), SW-GD-01-20 (Guangdong, China), YN and YUNNAN0901 (Yunnan, China) (Fig. 3).

Fig. 3 Phylogenetic analysis of the whole coding regions of the alternative passage JEV-SA14 strains and reference strains

3 Discussion

JEV is the leading cause of viral encephalitis and poses a severe health risk to the inhabitants of endemic areas in China. Of them, strain SA14 discovered from China mainland were adopted as the parent strain of attenuated vaccine (SA14-14-2) for its widely distributions in China and neighboring countries (Vratietal., 1999). As a representative strain, JEV SA14 strain were choosed in our studies, and the resultant squences from JEV mutants were shown with high similarities with wild strains, as CH13, SW/GD/01/2009 and YUNNAN0901 involved further indicated the representative of JEV SA14 in the endemic areas. From this opinion, our dual-host cell alternative passaging model well simulated the natural evolution and selections drived by veterbrated hosts and mosquito cycles.

Due to the error-prone nature of the RNA-dependent RNA polymerase, who lacks an exonuclease-proof reading activity in replications, the viral genome is prone to mutations naturally. In this study, nucleotides had a relatively high mutation rate over the entire coding genome of SA14 passage strains. Most of the mutations were single synonymous nucleotide substitutions without any amino acid changes. Neither insertions nor deletions was observed, agreed with the previous molecular epidemiological studies (Houetal., 2017).

It has been previously reported that 17 amino acid sites may be associated with viral virulence in JEV, such as E52, E76, E107, E135, E138, E176, E177, E232, E244, E264, E279, E315, E439, E447, NS2B-63, NS3-105 and NS4B-10 (Yangetal., 2014; Gromowskietal., 2015; Yangetal., 2017; Liuetal., 2018). We only observed one mutation (I176 T) of the above sites. However, the mutation presumably does not affect viral virulence (Gromowskietal., 2015), suggesting that mutations to decrease viral virulence are relatively infrequent under natural conditions. The observation supported this hypothesis that all viral strains produced typical pathological cell changes within about three days.

Among the gene fragments associated with antigenicity were reported as the key neutralization sites, with E337-345, E377-382, E397-403, E62, E327, E333, E373-395, E306, E331 and E387 were inolved (Lucaetal., 2012; Chenetal., 2013; Yangetal., 2014). In this study, only one V340I mutation was observed among these gene fragments, and it is not found in the neutralization site. These gene loci associated with antigenicity are not susceptible to mutate, that is why the live attenuated vaccine can provide broad-spectrum immune protection against JEV.

The survival fitness among the SA14 passage strains were observed with significant variants. It was speculated that the variations on some amino acid sites were related to the changes in survival fitness of JEV. For example, passages C2B2 and C6B6 had the highest virus titer, while C8B8 had the lowest virus titer. Sequencing results show that passages C2B2 and C6B6 are identical, and the amino acid sequence of the passage C8B8 has only one amino acid difference. The passage C8B8 acquired an E49K mutation in protein E, but it is not present in passages C2B2 and C6B6. This suggests that the mutation of the E protein (E49K) may be related to the decline of fitness for survival.

In summary, although the speed of genetic changes in JEV is rapid, the evolution of the virus is well conserved, especially in genetic loci associated with vital biological characteristics such as virulence and antigenicity. The genetic mutation and fitness compensations observed in this study agreed with previous studies (McCurdyetal., 2011; Yangetal., 2014; Houetal., 2017; Liuetal., 2018), which indicates that theinvitroalternative passaging model can be implemented to study the molecular biological characteristics of JEV, as well as other arboviruses with similar transmission cycles.