Infectivity variation and genetic diversity among strains of Western equine encephalitis virus

Western equine encephalitis is a severe, mosquito-borne infection of the brain of humans and domestic animals in the Americas. Western equine encephalitis virus (WEEV) was first isolated in 1930 and has been responsible for large, periodic and extensive epizootics and epidemics of encephalitis in equines and humans in the USA and Canada (Reisen & Monath, 1988). It remains an endemic public-health concern in North and South America and is a potential biological-warfare agent, due to its potential aerosol transmissibility.

WEEV is a member of the genus Alphavirus, a group of enveloped viruses with a positive-sense, single-stranded RNA genome. All alphaviruses share a number of structural, sequence and functional similarities, including a genome with two polyprotein gene clusters. The non-structural proteins (nsPs) are translated directly from the 5' two-thirds of the genomic RNA. A subgenomic positive-strand RNA (26S RNA) is identical to the 3' one-third of the genome and serves as the translational template for the structural capsid (C), E3, E2, 6K and E1 proteins (reviewed by Schlesinger & Schlesinger, 1996; Strauss & Strauss, 1988, 1994). Sequence comparisons of short regions within the nsP4 gene and the E1 protein/3' non-translated region have been determined for many WEEV strains, allowing a preliminary assessment of the nucleic acid phylogenetic relationships within the WEEV antigenic complex (Weaver et al., 1997). These WEEV strains could be grouped into multiple lineages. Furthermore, genetic variation among numerous WEEV strains from California since 1938 has been studied by investigating the E2 protein, and four major lineages were identified (Kramer & Fallah, 1999).

To understand alphavirus pathogenesis in more detail, studies have been conducted on the molecular basis of virulence of the closely related Sindbis virus (SINV) (Davis et al., 1986; Lustig et al., 1988; Olmsted et al., 1984; Polo & Johnston, 1991; Polo et al., 1988). Certain essential amino acids for SINV virulence have been identified (Davis et al., 1986; Griffin et al., 1989; Tucker et al., 1997). However, comparable studies on the molecular determinants of virulence of WEEV have not been undertaken, other than a preliminary study of WEEV pathogenicity (Zlotnik et al., 1972) and a comparison of WEEV virulence among different strains (Bianchi et al., 1993). In this study, we compared eight strains of WEEV by using an intranasal route of infection in a mouse model and studied genetic diversity by comparison of the complete 26S subgenomic region. Relationships among WEEV infectivity, phenotypic changes and possible determinants of virulence at the molecular level are discussed.

Virus culture and purification.
Vero cells (CCL-81; ATCC) were grown in minimal essential medium containing 5 % fetal calf serum. A summary of the eight WEEV strains used in this study is shown in Table 1. A 10 % suckling mouse brain suspension of strain 71V-1658 was kindly provided by Dr Nick Karabatsos (Centers for Disease Control and Prevention, Fort Collins, CO, USA); the Fleming and California strains were purchased from ATCC; the B11 and CBA87 strains were kindly provided by Dr George Ludwig (United States Army Medical Research Institute of Infectious Disease, Frederick, MD, USA) and the McMillan, Mn520 and Mn548 strains were kindly provided by Drs Mike Drebot and Harvey Artsorb (National Microbiology Laboratory, Winnipeg, MN, USA). Seed stocks of WEEV strains were made by inoculation of Vero cells with virus suspensions at an m.o.i. of <0.1. The supernatants were clarified by centrifugation, aliquotted and stored at 70 °C. All experiments with live virus were carried out in the Defence Research and Development Canada Suffield (DRDC Suffield) biological level 3 containment facilities, in compliance with Health Canada and Canadian Food Inspection Agency guidelines. Plaque assays were performed as described previously in six-well plates and stained by using an agarose neutral red overlay (Greenway et al., 1995).

Table 1. Details of the eight WEEV virus strains used in this study SM, Suckling mouse; V, Vero cells; GP, guinea pig; M, mouse; TC, tissue culture; ?, unknown.

Mouse infectivity studies.
Female BALB/c mice (1725 g, 720 weeks old) were obtained from the pathogen-free mouse-breeding colony at DRDC Suffield, with the original breeding pairs purchased from Charles River Canada (St Constant, QC, Canada). The use of these animals was reviewed and approved by the Animal Care Committee at DRDC Suffield. Care and handling of the mice followed the guidelines set out by the Canadian Council on Animal Care. Virus was administered to the mice by an intranasal route. The volume of inoculum used was 50 µl, containing 1.5x10³ p.f.u. diluted in Hanks' balanced salts solution (HBSS). Briefly, mice were anaesthetized with sodium pentobarbital [50 mg (kg body weight)¹] given intraperitoneally. When the animals were unconscious, they were carefully supported by hand with their nose up and the virus suspension in HBSS was applied gently with a micropipette into the nostrils. The applied volume was inhaled naturally into the lungs. Infected animals were observed daily for up to 14 days post-infection (p.i.). The times to death among groups were compared by using a two-tailed Student's t-test and one-way ANOVA using GraphPad Prism version 4.0 (GraphPad Software). Differences were considered to be statistically significant at P<0.05.

Extraction of subgenomic 26S RNA and RT-PCR.
Subgenomic 26S RNA was prepared from cell lysates by using a Qiagen RNeasy Mini kit. RNA was precipitated with 2-propanol and stored at 70 °C. Prior to use, RNA was washed with 80 % (v/v) ethanol, dried and dissolved in nuclease-free water (Promega). RT-PCR was performed in an Eppendorf Mastercycler gradient by using a One-step RT-PCR kit (Qiagen) with 0.1 µg viral 26S RNA and a pair of primers flanking the open reading frame of the complete structural polyprotein gene encoding the C, E3, E2, 6K and E1 proteins. The primers used were 5'-AAGCTTCCGCCAAAATGTTTCCATACCCTCAG-3' (forward) and 5'-TCTAGAGTGTATATTAGAGACCCATAGTGAGTC-3' (reverse). The reverse-transcription reaction was performed in a total volume of 50 µl for 30 min at 45 °C, using Omniscript and Sensiscript reverse transcriptases (Qiagen). After the reverse-transcription step, HotStarTaq polymerase (Qiagen) was activated by an increase in temperature to 95 °C for 15 min, followed by 40 cycles of amplification (94 °C for 10 s, 68 °C for 30 s and 68 °C for 4 min) and final extension (72 °C for 10 min). PCR products (3.7 kb) were isolated from 1 % agarose gels and purified by using a QIAquick Gel Extraction kit (Qiagen).

Cloning and DNA sequencing.
The extracted PCR products were cloned into the pcDNA3.1 TA vector (Invitrogen) following the manufacturer's instructions and transformed into TOP10 chemically competent Escherichia coli (Invitrogen). Plasmid DNA was isolated with a QIAprep Miniprep kit (Qiagen). Selected clones were sequenced by using the plasmid-specific T7 promoter primer and the bovine growth hormone reverse primer combined with internal WEEV-specific primers (Table 2) (Netolitzky et al., 2000). Sequencing reactions were performed by using a CEQ DTCS Quick Start kit (Beckman). Reaction products were purified by using Centri-Sep columns (Princeton Separations) and run on a CEQ 8000 Genetic Analysis system (Beckman). Sequences were assembled and analysed by using LASERGENE DNA software (DNAStar).

Table 2. Primers used for DNA sequencing of the complete WEEV structural protein sequence

Sequence analysis.
Multiple sequence alignment of the complete structural polyprotein from different WEEV strains for creation of a cluster diagram was carried out by using the Jotun Hein algorithm in the MEGALIGN program of the LASERGENE software package (DNAStar) with the default of gap penalty set to 11, the gap length penalty set to 3 and using the PAM250 weight table. A genetic test of selection was conducted by estimating the synonymous and non-synonymous substitution rates for all eight strains (by pairwise analysis) by the NeiGojobori method using MEGA3.1 (Kumar et al., 2004). The sequence diversity levels among different WEEV strains were assigned values by scoring mismatched amino acids. The values were plotted in windows of 50 aa across all of the structural proteins. Lethal infection of mice by using the intranasal route of infectivity
When adult BALB/c mice were infected intraperitoneally with WEEV strains Fleming, CBA87 and 71V-1658 diluted in HBSS to 10⁴ p.f.u. in 100 µl, the virus did not induce encephalitis or show overt symptoms of disease. However, if adult BALB/c mice were inoculated by using an intranasal route of infection with eight different WEEV strains, 100 % of adult mice succumbed to a lethal infection from 1.5x10³ p.f.u. in 50 µl (Fig. 1). The different WEEV strains were shown to vary in their time to death of the infected mice and could be divided into two pathotypes (P<0.05). Pathotype A comprised strains California, Fleming and McMillan, which were more virulent (Fig. 1), with a time to death of 56 days p.i. Pathotype B consisted of strains CBA87, Mn548, B11, Mn520 and 71V-1658, which were less virulent in mice, with time to death ranging from 8 to 12 days p.i.

(18K): Fig. 1. WEEV infectivity in a mouse model. Groups of six mice were inoculated intranasally with 50 µl of different strains of WEEV. The mice were monitored for 14 days and the percentage survival was plotted.

Sequence analysis
Nucleotide sequence comparison of the WEEV structural protein genes showed a very high identity, ranging from 96.8 % between CBA87 and Mn548 to 99.4 % between 71V-1658 and Mn520 (Table 3). As nucleotide identity was slightly higher than amino acid identity among some pairs of WEEV strains, the synonymous and non-synonymous substitution rates were estimated for all eight strains (using pairwise analysis) by the NeiGojobori method using MEGA3.1. The number of non-synonymous substitutions per non-synonymous site (the non-synonymous distance) was greater than the number of synonymous substitutions per synonymous site (the synonymous distance) for most strains (P<0.05), indicating that the amino acids encoded by the WEEV gene for most strains have evolved by positive selection. A multiple alignment of deduced amino acid sequences of the structural proteins of the eight strains is shown in Fig. 2. A comparison of amino acid sequences revealed that <4 % of amino acids differed among these proteins (Table 3). Small regions of variability were found in the C, E1 and E2 proteins, whereas the sequences of the E3 and 6K proteins were relatively more conserved (Fig. 3). Specific sequences of E3 and 6K are required for directing E1 and E2 to the endoplasmic reticulum (Liljeström & Garoff, 1991; Schlesinger & Schlesinger, 1972). Relatively conserved regions were also noted in the C-terminal sequences of the E1 and E2 proteins, where there is a transmembrane segment responsible for insertion of E1 or E2 into the lipid envelope (Garoff & Simons, 1974; Garoff & Söderlund, 1978). Furthermore, some alphavirus-conserved sequences were highly conserved among the eight strains, including LAAQIEDLRRSIANLTFK in the C protein (aa 3754), a putative coiled-coil α-helix important for viral core assembly (Perera et al., 2001), KPGKRQRMCMKLESD in the C protein (aa 95109), which has been shown to bind to ribosomes and to lie within a region that binds genomic RNA (Strauss & Strauss, 1994; Wengler et al., 1992), and VFGGVYPFMWGGAQCFC in the E1 protein (aa 8096), which is thought to be involved in fusion of the viral envelope with cellular membranes to release the nucleocapsid into the cytoplasm of virus-infected cells (Strauss & Strauss, 1994; Takkinen, 1986). A cluster diagram for the structural protein sequences is shown in Fig. 4. Two major genotypes were identified. Genotype A comprised strains Fleming, California and McMillan, whilst genotype B was composed of strains CBA87, B11, 71V-1658, Mn520 and Mn548. This genotype grouping matched the pathotype grouping exactly. In genotype B, strains were clustered primarily by the year of isolation. The oldest strain CBA87 (1958) occupied the basal position, whilst the second oldest strain B11 (1961) was located next to CBA87 and the most recent strains, Mn548 (1984) and Mn520 (1981), occupied the terminal branches. Strain 71V-1658 (1971) occurred in the middle branch, indicating that this genotype has evolved overall as a single lineage since 1958. In genotype B, the WEEV strains were also distributed in both North and South America, suggesting that some WEEV strains are widespread. This is different from other New World alphaviruses, such as Eastern equine encephalitis virus and Venezuelan equine encephalitis virus (VEEV), in which the North and South American strains are genetically distinct (Weaver et al., 1992). There are 45 cysteines, potential disulphide-bond formation sites and seven AsnXSer/Thr sequences, potential N-glycosylation sites, in the complete WEEV structural protein sequence. All cysteines and AsnXSer/Thr sequences among the eight strains were conserved. Although we could not identify consistent amino acid substitutions between the two genotypes, there were potential amino acid differences in some positions, including aa 57, 89 and 250 in the C protein, aa 23 and 69 in the E2 protein, aa 30 in the 6K protein and aa 196, 349 and 374 in the E1 protein.

Table 3. Percentage nucleotide (lower left) and deduced amino acid (upper right) identities of the structural proteins among eight WEEV strains Comparison of 3711 aligned nucleotides and 1236 aligned amino acids.

(58K): Fig. 2. Multiple alignment of the deduced amino acid sequences of the eight WEEV strains.

(12K): Fig. 3. Sequence diversity levels among the different WEEV strains were assigned values and plotted in windows of 50 aa across the structural proteins.

(4K): Fig. 4. Cluster diagram of WEEV structural proteins determined by using the Jotun Hein algorithm of the MEGALIGN program in the DNASTAR package. Horizontal lines are proportional to the number of substitutions between branch points. The length of each pair of branches represents the distance between sequence pairs.

Much of the available information on the molecular details of the structure and the pathogenesis of the alphaviruses has come from extensive studies with two members, SINV and Semliki Forest virus (SFV) (Atkins et al., 1999; Caballero-Herrera & Nilsson, 2003; Davis et al., 1986; Gibbons et al., 2004; Lustig et al., 1988; Olmsted et al., 1984; Polo & Johnston, 1991; Polo et al., 1988; Tuittila & Hinkkanen, 2003). Relatively little is known about the other alphaviruses, particularly WEEV. Despite their similarities at the molecular level, alphaviruses are diverse in the severity of the diseases that they cause in humans and other vertebrates. WEEV can cause severe and sometimes fatal encephalitis in humans and is a significant pathogen in horses (Reisen & Monath, 1988). In contrast, SINV and SFV cause only mild fever and rash or an asymptomatic infection in humans (Shope, 1980). In order to gain insight into molecular determinants of WEEV pathogenesis, we studied viral infectivity with genetic variation among eight different strains.

An animal model is crucial to study the pathogenesis of viruses. So far, there has not been a reliable animal infection model for WEEV. When we first tried to inoculate WEEV by the intraperitoneal route, no symptoms of infection were seen in adult mice, which was different from the results of Bianchi et al. (1993). We chose BALB/c adult mice aged 720 weeks, whereas Bianchi et al. (1993) used outbred Swiss NIH adult mice aged 45 weeks. The difference in mouse strain and age could be the reason why we obtained different results. Research with C57 black mice has indicated that mice of less than 8 weeks old are susceptible to disease following intraperitoneal inoculation (unpublished results). Nevertheless, when adult mice were inoculated intranasally with WEEV at a similar or lower dose, all eight strains examined were 100 % lethal to the mice in this model. Based on the time to death, the eight strains could be divided into two pathotypes. Interestingly, the more virulent pathotype consisted of WEEV strains isolated during the major epidemic in the 1930s and 1940s. The less virulent pathotype was composed of the more recent isolates of WEEV. These results are consistent with the study by Bianchi et al. (1993), who examined the virulence of two CBA South American strains, one of which was strain CBA87. This strain, isolated 25 years earlier than CBA CIV180, developed higher brain titres and killed mice earlier when suckling mice were inoculated with CBA strain viruses by the intracranial or intraperitoneal routes of inoculation. Natural attenuation of WEEV strains over time is a trend that is observed when considering time to death as an indicator of virulence.

Virus virulence is reflected throughout the multiplication cycle and manifests itself in virus entry to cells, virus replication, viral interaction with host cells, interferon production, an immune response induced by viral infection and other factors. The structural proteins play an important role in the interaction between the virus and its surrounding environment and are potentially related to virus virulence. Thus, the complete structural protein sequence was chosen for analysis among the eight different strains to elucidate WEEV infectivity at the molecular level. We found relatively few differences in the 26S structural protein subgenomic region among the different strains. Nucleotide and amino acid identities were found to be >96 %, indicating the relative stability of WEEV within its natural environment. Small variable regions were concentrated in the C, E1 and E2 proteins, whilst the 6K and E3 proteins were relatively more conserved. Once the virions are exposed to the external environment, they are subject to selective pressure exerted by antibodies of infected or immunized animals. As a result, the E1 and E2 proteins have been shown to be relatively more diverse. The divergence was not distributed randomly over the C, E1 and E2 proteins. Segments in C, E1 or E2, including those important for WEEV basic biological functions such as E1/E2 insertion into a lipid envelope, C protein binding to ribosomes, viral core assembly and fusion of viral envelope with infected cells, were found to be conserved. Furthermore, the three-dimensional topography of the structural proteins of the eight strains is probably similar, as all cysteines and potential N-glycosylation sites in the structural proteins were conserved. The cluster diagram divided the eight strains into two closely related genotypes that matched the pathotype grouping exactly, suggesting that genetic variation of structural proteins contributes to infectivity changes. In addition to structural proteins, there are four non-structural proteins (nsP14) encoded by the 5' two-thirds of the WEEV genome; these are enzymes involved in virus replication. Thus, mutations in the non-structural proteins of viral genomes can play a significant role in virus attenuation by affecting virus replication. Attenuating mutations in non-structural proteins have been identified for SINV (Frolova et al., 2002), SFV (Tuittila et al., 2000) and VEEV (Kinney et al., 1989). We do not know whether differences in non-structural proteins contributed to the differences observed in virus infectivity. Further study is required to identify their contribution to virulence.

The eight strains of WEEV differed in their virulence for mice and this correlated with variations in the amino acid composition of their structural proteins. How does WEEV genetic variation influence its infectivity? A general feature of alphavirus pathogenesis is that remarkably small genetic changes, even a single nucleotide or amino acid substitution in structural proteins, can have dramatic effects on virus virulence. For example, the AR339 strain of SINV is not fatal in adult mice, but a neurovirulent strain that causes fatal encephalitis in adult mice has only four amino acid differences from AR339 in its structural proteins (Griffin et al., 1989). In the E2 protein of SINV, the residues at positions 55 and 172 determine the neurovirulence for mice of different ages and the efficiency of replication in tissue of the nervous system (Tucker et al., 1997). Further studies have shown that a single residue change from arginine to serine at aa 114 in the SINV E2 protein is sufficient to attenuate virulence in newborn mice and accelerate penetration of BHK cells (Davis et al., 1986). A similar result was obtained for VEEV, showing that a single residue substitution in E2 could change VEEV pathogenesis (Aronson et al., 2000). In our studies, we found that there was a potential amino acid substitution in some positions between the two groups, such as aa 57, 89 and 250 in the C protein, aa 23 and 69 in the E2 protein, aa 30 in the 6K protein and aa 196, 349 and 374 in the E1 protein, suggesting the potential relative importance of amino acid variation in virulence differences. Further studies need to be undertaken to identify the essential amino acids in WEEV infectivity and how these influence infectivity.