In late February 2011, neurologic disease was reported in several horses in northwestern and southwestern NSW. The number of cases and geographic distribution gradually increased. By mid-June 2011, specimens from ≈300 horses were submitted to the virology laboratory at Elizabeth Macarthur Agriculture Institute (Menangle, NSW, Australia). Many more horses probably were affected. Diseased horses were located throughout most of NSW, west of the Great Dividing Range, but also extending through the Hunter River Valley region, Sydney Basin, and Illawarra coastal region immediately south of Sydney (Figure 1, panel B). Cases also occurred in other southern Australia states. Clinical signs were generally consistent with those described in horses infected with WNV in the United States (5). A detailed report of the clinical signs, virology, and pathology of equine cases will be published elsewhere.

Whole brains were removed from 12 horses at postmortem examination. Half of each brain was fixed in 10% neutral buffered formalin; the other half was held fresh at 4°C. Upon receipt, we collected small pieces of fresh and formalin-fixed tissue from several locations in the cerebrum and cerebellum and along the brain stem and cervical spinal cord. If virus isolation could not be performed on fresh samples within 24 h after receipt, we held the samples at −80°C until tested. Before testing, we prepared 10% tissue homogenates in RPMI medium (Life Technologies, Carlsbad, CA, USA) containing antimicrobial drugs.

Mosquitoes were collected throughout NSW, as part of the NSW Arbovirus Surveillance and Mosquito Monitoring Program, by using dry ice–baited light traps. The mosquitoes were submitted live to the Medical Entomology Laboratory at Westmead Hospital (Westmead, NSW, Australia) for species identification, arbovirus isolation, and virus identification (6).

We propagated Vero 76 cells in Dulbecco modified minimum essential medium (DMEM; Life Technologies) supplemented with 10% fetal bovine serum (FBS). C6/36 Aedes albopictus mosquito cells were maintained in RPMI medium supplemented with 10% FBS, and BHK21 cells were maintained in DMEM containing 5% FBS. All cell lines were obtained from the American Type Culture Collection (Manassas, VA, USA). We used prototype WNV_KUN MRM61C (7) and WNV New York 99 (WNV_NY99) 4132 strains (8) for comparison with WNV_NSW2011. Stocks of WNV_KUN MRM61C and WNV_NY99 4132 that most closely resembled the low-passage level WNV_NSW2011 were prepared by electroporation of BHK21 cells with WNV_KUN or WNV_NY99 RNA (prepared from corresponding infectious cDNA clones) and passaging them 1× in C6/36 cells. Viral supernatants were harvested 5 days later. Viral titers for each viral stock were determined by plaque assay on Vero 76 cells.

We conducted microneutralization tests (9) in Vero cells by using 25–100 infectious units (measured as 50% tissue culture infective doses) of WNV_NSW2011, WNV_KUN, and WNV_NY99; we used 2-fold dilutions of serum from an initial dilution of 1:20. Results were scored as 80% reduction in virus growth or 100% inhibition of virus growth. Reduction in virus growth was determined by assessing the extent of cytopathic effect in each well. Inhibition of virus growth was determined by the absence of viral antigen in the cells of each well when tested with a WNV-reactive monoclonal antibody (mAb) in ELISA.

We used the MagMax-96 Viral RNA Isolation Kit (Ambion, Austin, TX, USA) on a magnetic particle handling system (Kingfisher 96; Thermo Electron Corporation, Vantaa, Finland) to extract total nucleic acid from clarified 10% brain homogenate (50 μL) or tissue culture fluid. Purified nucleic acids were eluted in 50 μL of kit elution buffer and used immediately for PCR amplification or stored frozen at ≈−20°C

We used a published WNV real-time reverse transcription PCR (rRT-PCR) (10 [assay 2]) with the following variations: Black Hole Quencher (Biosearch Technologies, Novato, CA, USA) was used instead of TAMRA on the probe, the internal control system was not used, and 5 μL of RNA was used as template. The AgPath-ID One-Step RT-PCR Kit (Applied Biosystems, Foster City, CA, USA) was used for the rRT-PCR on a 7500HT Fast Real-Time PCR System (Applied Biosystems). We ran rRT-PCR reactions in standard mode, according to conditions recommended by the mastermix manufacturer.

Supernatant from the 10% brain homogenate was placed on monolayers of A. albopictus C6/36 mosquito cells in cell culture tubes. The cultures, which were maintained in RPMI medium containing antimicrobial drugs at standard concentrations and supplemented with 2% FBS, were incubated at 28°C for 5–7 days. Culture supernatants were then passaged up to 3× on BHK21 cells maintained in Basal Medium Eagle (MP Biomedicals, Sydney, Australia) containing antimicrobial drugs and 2% FBS. We regularly examined cultures by light microscopy for cytopathic effects. We used rRT-PCR to confirm the identity of virus isolates in culture supernatants or to confirm that there was no virus replication in the absence of cytopathology. Viral RNA recovered from culture fluid at the first or second passage in BHK21 cells was sequenced as described below. A virus isolate obtained from the first brain examined was designated WNV_NSW2011. Virus isolation was conducted on homogenates of mosquitoes by similar methods but with passage onto BHK and PSEK cells after initially being placed on C6/36 cells. We identified virus isolates by using immunoassays with generic and specific mAbs.

WNV_NSW2011 virus harvested from the first passage in C6/36 cells was used to examine plaque morphology and virulence in mice. The virus was passaged 1× in Vero76 cells for 4 d and 1× in C6/36 cells for 5 d. Virus supernatant was centrifuged at 500 × g at 4°C for 5 min before being stored at −80°C

Reactivity of the new isolate with a panel of mAbs was compared with that of WNV_KUN and WNV_NY99 by using a fixed-cell ELISA (11). The mAbs and their characteristics follow: mAb 10C6, anti–nonstructural protein (NS) 1 (reactive with MVEV); mAb 10A1, anti-envelope glycoprotein (anti-E; specific for WNV_KUN); mAb 2B2, anti-E (reactive with WNV); mAb 3.1112G, anti-NS1 (reactive with WNV); mAb 5H1, anti-NS5 (strong reaction with WNV_KUN strains, weak for WNV_NY99, nonreactive with WNV strains from other lineages); mAb 17D7, anti-E (specific for strains of WNV with glycosylated E); mAb 3.101C, anti-E (specific for strains of WNV with unglycosylated E); and the pan–flavivirus-reactive mAbs (i.e., 4G4, anti-NS1; and 4G2, anti-E) (12–15; J. Hobson-Peters et al., unpub. data).

For sequencing of the whole genome, we used total nucleic acid purified from virus-infected cell culture supernatant as template in 5 RT-PCRs with primers designed to cover the coding regions of any WNV genome (Table 1). RT and amplification were performed by using the SuperScript III One-Step RT-PCR System (Invitrogen, Carlsbad, CA, USA) with primers at 20 μM. RT was performed at 50°C for 30 min, followed by denaturation at 94°C for 2 min. PCR amplification involved 40 cycles (95°C for 30 s, 55°C for 1 min, 68°C for 4.5 min) followed by a final extension at 68°C for 10 min. Reaction products were visualized after electrophoresis on a ethidium bromide–stained 1% agarose gel. Reaction products were purified directly (MinElute PCR Purification Kit; QIAGEN, Valencia, CA, USA) or excised from the gel and cleaned (Gel MinElute PCR Purification Kit; QIAGEN). Purified nucleic acid was sequenced at the Australian Genome Research Facility (Sydney) by using the primers used to generate the PCR product. Each RT-PCR was run 3× and sequenced in both forward and reverse orientation. Sequence data were assembled by using Sequencher software (Gene Codes Corp., Ann Arbor, MI, USA). For subsequent isolates from horses and mosquitos, we used primers designed from the sequence of the WNV_NSW2011 genome (Table 1) to amplify and sequence the NS3 and NS5 regions, in which changes had been identified. The same RT-PCR and sequencing methods were used, except that the annealing temperature was 50°C and extension time was 1 min. The nucleic acid sequences were translated and then aligned with WNV_NSW2011 and WNV_KUN (GenBank accession nos. JN887352 and D00246.1, respectively) by using ClustalW (www.clustal.org).

Complete coding regions of selected WNV isolates, representing all lineages and clades and including all complete KUN sequences, were aligned with the WNV_NSW2011 sequence as described (16). This alignment was transferred to BioEdit (www.mbio.ncsu.edu/BioEdit/bioedit.html) for manual editing before construction of phylogenetic trees. Maximum-likelihood trees were constructed by using PhyML (17). Trees were rooted by using the Japanese encephalitis virus Nakayama sequence (GenBank accession no. EF571853), which was removed from the final tree for clarity.

To examine glycosylation of the E protein, viral proteins from cultures of infected C6/36 cells were digested as described (18). Proteins were separated and analyzed by Western blot. Samples were loaded with reducing sodium dodecyl sulfate–polyacrylamide gel electrophoresis buffer (NuPAGE LDS Sample Buffer; Invitrogen) on a 4%–12% NuPAGE Gel (Invitrogen). Electrophoresed proteins were electroblotted onto nitrocellulose paper (Hybond C; GE Healthcare, Little Chalfont, UK) and immunostained with anti-E mAb (11).

We allowed the virus to adsorb to monolayers of Vero 76 cells in 6-well plates for 2 h at 37°C. The cells were overlaid with DMEM containing 0.75% low melting point agarose and 2% FBS. Four days after infection, the cells were fixed with 4% formaldehyde solution and stained with 0.2% crystal violet.

Groups of 10 weanling (18–19 days old) or young adult (4 weeks old) Swiss outbred CD1 mice were injected intraperitoneally with 10-fold dilutions of virus. The mice were monitored for 21 days after injection and euthanized when signs of encephalitis were evident. All animal procedures had received prior approval from The University of Queensland Animal Ethics Committee.

Viral RNA was detected by WNV-specific rRT-PCR in fresh brain tissue from 6 of 12 horses showing signs of encephalitis. Viruses were isolated from 4 of these samples; each showed distinct cytopathology in BHK21 and Vero cells. rRT-PCR of the culture fluids and immunoperoxidase staining of the cells with pan–flavivirus-reactive and WNV-specific mAbs confirmed the isolation of a West Nile–like virus. The first isolate was named NSW2011 and designated WNV_NSW2011. Eight isolates of WNV_NSW2011 were isolated from Culex annulirostris Skuse mosquitoes during the 2011 vector season. Of the 8 isolates, 5 were from mosquitoes collected in the Riverina region of southwestern NSW (Hanwood, 4 isolates; Barren Box, 1 isolate); 2 were from the Murray region in southern NSW (Mathoura, 1 isolate; Moama, 1 isolate); and 1 was collected at Lower Portland in the outer western Sydney region of NSW. No other isolates of WNV were obtained.

To antigenically type WNV_NSW2011 in a fixed-cell ELISA (11), we used a panel of mAbs previously shown to differentiate between strains of WNV_KUN and other WNVs (11,13–15,19,20). The recognition patterns showed that the WNV_NSW2011 isolate most closely resembled Australian WNV_KUN strains; the WNV_KUN-specific mAb 10A1 reacted strongly with prototype WNV_KUN and WNV_NSW2011 but not with WNV_NY99 (Table 2). However, the anti-NS5 mAb 5H1, which is also specific for WNV_KUN isolates from Australia (15), failed to react with WNV_NSW2011 and WNV_NY99, but it bound strongly to WNV_KUN. The reaction patterns of mAbs 17D7 and 3.101C, which react specifically with glycosylated and unglycosylated WNV E antigens, respectively (11,14; J. Hobson-Peters et al., unpub. data), indicated that, unlike the E protein of WNV_KUN, the E protein of WNV_NSW2011 is glycosylated.

To assess the level of antigenic crossreactivity between WNV_NSW2011, WNV_KUN, and WNV_NY99, we assessed neutralization titers for homologous and heterologous viruses in immune serum samples from the following sources: horses infected during the 2011 outbreak, horses infected with WNV_KUN in the Northern Territory of Australia several years earlier, and horses infected with WNV in the United States. Convalescent-phase serum samples from WNV_NSW2011-immune horses had neutralizing titers similar to those of the homologous virus (WNV_NSW2011) and of WNV_KUN and WNV_NY99 (Table 3). Serum samples from WNV_KUN-immune horses from Northern Territory and from WNV-immune animals from the United States showed a similar pattern of cross-neutralization. However, serum samples from horses from Northern Territory and the United States showed slightly less neutralizing efficiency of WNV_NSW2011; this was likely due to a higher dose (≈4-fold) of virus in the assay (Table 3). Overall, these results are consistent with those in our previous reports showing a high level of cross-neutralization between WNV_KUN and WNV_NY99 strains (21,22).

A comparison of the nucleotide sequence of the complete coding region of the first isolate of WNV_NSW2011 with sequences available in GenBank confirmed that WNV_NSW2011 was genetically most closely related to Australian WNV_KUN isolates (Figure 2). A detailed comparison of deduced amino acid sequences for the entire coding region of WNV_NSW2011, WNV_KUN, and WNV_NY99 further confirmed a closer relationship between WNV_NSW2011 and WNV_KUN than between WNV_NSW2011 and WNV_NY99. There was a 42-aa difference (18 nonconserved changes) between WNV_NSW2011 and WNV_KUN and an 89-aa difference (38 nonconserved changes) between WNV_NSW2011 and WNV_NY99 (Table A1). At least 2 of the known WNV virulence markers present in WNV_NY99 but not in WNV_KUN were found in WNV_NSW2011 (Table A1). The glycosylation tripeptide (N-Y-S) at residues E_154–156, which allows N-linked glycosylation at a conserved site on the E protein, domain I, and is associated with virulence in most WNV strains (23), was present in WNV_NSW2011; its presence is consistent with the mAb recognition profile. A phenylalanine residue at aa 653 in NS5, which also is associated with enhanced virulence of WNV strains (24), was present in WNV_NSW2011. Because the WNV_KUN strain–specific mAb 5H1 did not react with WNV_NSW2011, we also examined the predicted amino acid sequence of the NS5 protein that corresponded to the linear epitope previously mapped for 5H1 to residues 41–53 in the methyltransferase domain (15). WNV_NSW2011 and WNV_NY99 contain a substitution (I→V) at residue 49 that is not contained in WNV_KUN, confirming the critical role of this residue for 5H1 binding. Together these data suggest that WNV_NSW2011 represents a virulent WNV strain that has emerged in Australia. However, the amino acid substitution in NS3, which is believed to be associated with increased virulence of WNV_NY99 in birds in North America (25), was not present in isolate WNV_NSW2011. Sequencing of the key regions of the other WNV isolates obtained during this outbreak (3 from horses and 8 from mosquitoes) showed that each was indistinguishable from WNV_NSW2011.

The average plaque size of WNV_NSW2011 (Figure 3, panel A) was 4.2 mm ± 0.5 mm, much closer to WNV_NY99 (4.7 mm ± 0.8 mm) than to the prototype WNV_KUN (2.8 mm ± 0.4 mm). To confirm the presence of an N-linked glycan on the E protein of WNV_NSW2011, we used Western blot to analyze endoglycosidase-digested viral protein. Analysis showed that the E protein of WNV_NSW2011 and WNV_NY99 migrated slightly faster than the undigested control protein. This result is consistent with N-linked glycosylation (Figure 3, panel B). However, consistent with the lack of a potential glycosylation site on the E protein of most WNV_KUN isolates, we found no evidence of N-linked glycosylation for WNV_KUN (11,20,26).

Injection of 18- to 19-day-old (weanling) mice with 10-fold dilutions of virus showed that substantially lower doses of WNV_NSW2011 (50% lethal dose [LD₅₀] 0.5 PFU), compared with WNV_KUN (LD₅₀ 13.4 PFU), induced neurologic signs (Table 4); and the time to disease onset was substantially shorter (Table 4; Figure 3, panel C). In contrast, the LD₅₀ for WNV_NY99 (0.1 PFU) was lower than that for WNV_NSW2011, and neurologic signs developed more rapidly (Table 4; Figure 3, panel C). Only WNV_NY99 and WNV_NSW2011 caused a substantial number of deaths among 4-week-old (young adult) mice. Compared with WNV_NSW2011, WNV_NY99 exhibited a lower LD₅₀ (240 PFU vs. 0.7 PFU, respectively) and a shorter time to death (10.7 days vs. 6.6 days, respectively, at 1,000 PFU) (Table 5; Figure 3, panel C).