Detection and Circulation of a Novel Rabbit Hemorrhagic Disease Virus in Australia

The highly virulent rabbit hemorrhagic disease virus (RHDV) has been widely used in Australia and New Zealand since the mid-1990s to control wild rabbits, an invasive vertebrate pest in these countries. In January 2014, an exotic RHDV was detected in Australia, and 8 additional outbreaks were reported in both domestic and wild rabbits in the 15 months following its detection. Full-length genomic analysis revealed that this virus is a recombinant containing an RHDVa capsid gene and nonstructural genes most closely related to nonpathogenic rabbit caliciviruses. Nationwide monitoring efforts need to be expanded to assess if the increasing number of different RHDV variants circulating in the Australian environment will affect biological control of rabbits. At the same time, updated vaccines and vaccination protocols are urgently needed to protect pet and farmed rabbits from these novel rabbit caliciviruses.

descendants are genogroup G2 viruses. In the late 1990s, new genetically and antigenically distinct RHDV variants were reported and were designated as subtype RHDVa or genogroup G6 on the basis of differences in reactivity profiles with monoclonal antibodies and on sequence divergence (10,12,13). Despite this difference, vaccines developed for classic RHDV strains were still found to be protective against RHDVa when administered at the correct dose and within recommended intervals (10,12,13). RHD-Va strains were first detected in Italy and were subsequently detected in multiple countries, including Portugal, the United States, China, Cuba, and South Korea (12; reviewed in 1). In some geographic regions, it was suggested that RHDVa had replaced previous enzootic viruses (14)(15)(16)(17).
In addition to the virulent RHDV, RHDVa, and RHDV2, several nonpathogenic lagoviruses have been described that are genetically distinct and have historically been called rabbit caliciviruses (RCVs). RCVs have been reported from various geographic locations; in contrast to the pathogenic hepatotropic lagoviruses, most RCVs mainly cause a nonclinical infection of the small intestine (1). Until 2014, only the benign rabbit calicivirus Australia-1 (RCV-A1) and the RHDV biocontrol strain CAPM V-351 (and its descendants) were circulating in Australia (7,31). In May 2015, RHDV2 was detected in Australia for the first time (23). Before the detection of RHDV2, another exotic RHDV incursion was reported to the World Organisation for Animal Health (OIE) in January 2014 (32). We describe the detection, spread, and evolution of this first exotic virus, RHDVa, in Australia during December 2013-May 2015, before the arrival of RHDV2. For serologic analysis of healthy wild rabbits, we collected 20 wild rabbits from the ACT; these animals were shot from a vehicle using a 0.22-caliber rifle targeting the head or chest. We collected samples (blood, liver,  duodenum, and bile) postmortem; all sampling was conducted according to the Australian Code for the Care and Use of Animals for Scientific Purposes as approved by the CSIRO Ecosystem Sciences animal ethics committee (approval #12-15).

RHDV Antigen Capture ELISA
Liver samples submitted to EMAI were prepared by homogenizing ≈1 g of tissue in 5 mL of phosphate-buffered saline and centrifuging at 3,000 rpm for 10 minutes. Supernatant was tested in the RHDV antigen-capture ELISA as previously reported (33).

Real-time Reverse Transcription PCR Assay
We screened samples for RHDV and RHDVa using realtime reverse transcription PCR (rRT-PCR). We applied sterile cotton swabs to the freshly cut surface of liver samples and transferred them to 5-mL vials containing 3 mL of sterile phosphate-buffered gelatin saline, then extracted viral RNA from 50-µL samples from these vials using the MagMAX-96 Viral Isolation System (Ambion, Austin, TX, USA) on a Kingfisher 96 magnetic particle handling system (ThermoFisher Scientific, Waltham, MA, USA).
We tested purified viral RNA (5 µL) in an RHDVspecific Taqman rRT-PCR assay using the AgPath-ID One-Step RT-PCR kit (ThermoFisher Scientific), based on detection of a conserved sequence of the VP60 gene with modifications (Table 1) (34). We developed an in-house Taqman rRT-PCR to specifically detect the exotic RHDVa virus (Table 1). Cycling conditions are available from the authors upon request.

Sequencing
We extracted RNA from 20-30 mg of liver tissue or bile using the RNeasy Mini Kit (QIAGEN, Hilden, Germany); the Maxwell 16 LEV simplyRNA Tissue Kit (Promega, Madison, WI, USA); or the Invitrogen Purelink viral RNA/ DNA Mini Kit (ThermoFisher Scientific) following the manufacturers' instructions. We performed first-strand cDNA synthesis from 5 μL of RNA with 500 ng of Oli-godT (18mer) (Geneworks, Thebarton, South Australia, Australia) using Invitrogen Superscript III (Life Technologies, Carlsbad, CA, USA) following the manufacturer's directions. We conducted initial detection and typing of RHDV samples submitted to CSIRO (>1 isolate from each outbreak) using the universal lagovirus RT-PCR, as described previously (35), followed by Sanger sequencing of amplified products. RHDV G2 genomes were amplified as described previously (7). We amplified overlapping fragments of the RHDVa genome using Platinum Taq DNA polymerase high fidelity (Life Technologies) using primer sets detailed in Table 2 (cycling conditions available from the authors upon request). We purified PCR products using the QIAquick PCR purification kit (QIAGEN), quantified with the Qubit dsDNA BR assay (Life Technologies) and pooled in equimolar ratios.

Library Preparation and Illumina Sequencing
We prepared DNA libraries of PCR amplicons using the Nextera XT DNA Sample Prep Kit (Illumina, San Diego, CA, USA). Sequencing was performed on an Illumina MiSeq using the 300-cycle paired-end MiSeq Reagent Kit v2 (Illumina), as described previously (7).

NGS Genome Assembly and Data Analyses
We performed read quality assessment and trimming and merged overlapping paired-end reads, as described previously (7). We then mapped individual cleaned reads to the RHDV reference genome (Genbank accession no. M67473.1) using the Geneious in-built mapper tool as available in Geneious version 8.1.6 (39) and generated a majority consensus sequence. Primer sequences were trimmed from the 5′ and 3′ genomic termini. We named sequences using the syntax country/state/isolate identifier/ date, with date in the format yyyy/mm (e.g., AUS/NSW/ BER-1/2013/12). Consensus sequences were aligned with representative RHDV, RHDV2, and RCV-A1 sequences for recombination and phylogenetic analyses. We submitted all sequences generated in this study to GenBank (accession nos. KY628306-21. We inferred maximum likelihood phylogenies for both the nonstructural genes (44 sequences, 5,238 nt) and the VP60 gene (47 sequences, 1,743 nt) using PhyML (40). We estimated trees using the general time reversible model with rate heterogeneity among sites (4 discrete rate categories) and a proportion of invariant sites (determined as most appropriate using jModel test [41]), with a combination of nearest neighbor interchange and subtree pruning 24 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 24, No. 1, January 2018 and regrafting branch swapping methods employed to search for the optimal topology. We rooted trees using a European brown hare syndrome virus sequence (EBHSV; GenBank accession no. Z69620) and estimated branch support using 1,000 bootstrap replicates. We screened a full genome alignment containing the 15 newly acquired sequences and 12 reference genomes (27 sequences, 7,303 nt) for recombination using the RDP, GENECONV, MaxChi, and BootScan methods available in the RDP4 package (42), with p<0.05 representing significant evidence of recombination. We confirmed recombination events by phylogenetic analyses on either side of the proposed breakpoint.

Serologic Analysis
We tested serum specimens from the shot wild rabbits for the presence of RCV-A1 antibodies using a specific blocking ELISA (43). We also tested for antibodies against RHDV using a competition ELISA and used ELISAs for RHDV IgA and IgM subclass antibodies as described previously (44).

Results
In late December 2013 and January 2014, a property in northern Sydney, NSW, Australia, experienced the sudden deaths of 30 domestic rabbits from a population of 80. The treating veterinarian submitted liver samples to EMAI for routine RHDV testing in January 2014. The samples showed little or no reactivity in the RHDV antigen capture ELISA but tested strongly positive in the RHDV-specific rRT-PCR. This result was in contrast to the typical reactivity of field strains of RHDV from Australia, which react strongly in both the RHDV antigen capture ELISA and the RHDV-specific rRT-PCR (Table 3). To further characterize this calicivirus, a 326-bp fragment of the VP60 capsid gene was amplified using a universal lagovirus PCR (35). Sanger sequencing of this fragment identified the virus as a member of the RHDVa group (RHDV G6), a virus not previously reported in Australia.
Outbreak 11, from Mulligan's Flat Nature Reserve near Canberra, ACT, Australia, was suspected when rangers detected a sudden reduction in rabbit numbers during their rabbit eradication program, although no rabbit carcasses were recovered. Subsequent serologic analysis of 20 healthy shot rabbits revealed a high proportion of rabbits positive for RHDV-specific IgM (7/20) and IgA (15/20), indicating a recent virus outbreak (45). However, despite moderate to high IgA and IgM titers, the serum specimens tested low or negative in the RHDV competition ELISA and RCV-A1 blocking ELISA, which are known to be more strain-specific than the IgM and IgA ELISAs (data not shown). This finding indicated possible exposure of the Mulligan's Flat rabbit population to a lagovirus that was antigenically distinct from both RHDV and RCV-A1. Because RHDV RNA can be detected in the bile of recovering animals for as long as 15 weeks postinfection (34), we extracted RNA from the bile of IgM-positive individuals and  analyzed it using the universal lagovirus PCR (35). We amplified a lagovirus-specific fragment from 1 animal (AUS/ ACT/MF-109/2015/03; Table 4), and Sanger sequencing confirmed the presence of the exotic RHDVa in this rabbit. We also detected Australian field strains (RHDV G2) during the same sampling period on 3 occasions, 2 from wild rabbits (outbreaks 4 and 6; Table 4), and 1 from a domestic rabbit facility (outbreak 1; Table 4). The strain from the domestic rabbit facility was identified in samples from the first reported outbreak of the exotic RHDVa, indicating that both viruses were active simultaneously on this property.
We conducted full-genome sequencing for 15 RHDV isolates we sampled during December 2013-March 2015. We could not recover a full-length genome from AUS/ ACT/MF-109/2015/03. Recombination and phylogenetic analyses of the nonstructural genes indicated that all RHDVa viruses were recombinants between RHDVa and an RCV-A1-like virus, with strong bootstrap support (Figures 3, 4). Specifically, we found strong evidence for recombination with a putative breakpoint located between the RdRp and VP60 genes (genome position 5,304 in Genbank accession no. M67473.1 sequence numbering), detected by all recombination detection methods in RDP4 (Figure 3). Phylogenetic analysis of the complete VP60 capsid gene revealed that the RHDV G2 viruses clustered with other recent RHDV G2 samples from Australia and that the RHDVa samples clustered (99% nucleotide sequence identity) with exotic RHDVa variants, particularly a variant from China, XA/China/2010 (Genbank accession no. JN165234) (Figure 4, panel A) (14). In contrast, the nonstructural genes of the RHDVa viruses found in Australia clustered with the RCV-A1-like clade that contains several other lagoviruses, including RCV-A1/ RHDV2 recombinant viruses (46) and Michigan RCV (47), neither of which have been detected in Australia ( Figure 4, panel B). Although the nonstructural genes of these viruses cluster most closely with RCV-A1, they are quite distinct, sharing only 83.3%-84.4% nucleotide identity with RCV-A1 sequences in this region, and could therefore represent a novel lagovirus genotype.

Discussion
We describe the detection and characterization of RHD-Va in Australia, the first of 2 recent incursions of exotic lagoviruses (23). The route of entry of these viruses into Australia is unclear. The spread of RHDVa from Europe to distant locations such as Asia, the Americas, Reunion Island, and Australia (1) highlights the ease with which this virus can be disseminated, despite strict quarantine regulations. Caliciviruses are known to be highly environmentally stable, they replicate to very high titers in infected rabbits, and they are efficiently transmitted (1), suggesting that they would be relatively easy to disseminate inadvertently, analogous to the situation observed with the emergence of canine parvovirus (48). 26 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 24, No. 1, January 2018 *Samples are numbered according to the date on which they were collected. These numbers correspond to those used on the map in Figure 2. ACT, Australian Capital Territory; NSW, New South Wales; RHDV, rabbit hemorrhagic disease virus. †Samples for which the full genome was sequenced in this study. ‡Detection of RHDVa confirmed in January 2014. §Virus sequence obtained from a healthy shot rabbit.
Unexpectedly, full-genome sequencing revealed that all RHDVa viruses in Australia had recombinant genomes, with the capsid gene closely related to a 2010 virus from China and nonstructural genes related to RCV-A1-like viruses ( Figure 4). This finding highlights the importance of recombination in the generation of genetic diversity in this virus family. The emergence of this RCV-A1-like/RHDVa recombinant is of particular interest given the differences in tissue tropisms of RHDVa, which replicates in the liver, and nonpathogenic RCVs, which replicate in the small intestine (1,35). Recombinant viruses containing RCV-A1like nonstructural genes and RHDV2 structural genes have also been reported (46). Although RHDV is widely disseminated in an infected animal, there is no evidence to suggest it can actively replicate in the small intestine. Conversely, RCV-A1 has only been shown to infect the columnar epithelial cells of the small intestine (35). Therefore, we hypothesize that the recombination event likely occurred during early viral replication in macrophages, which has been suggested for both RHDV (1) and RCV-A1 (T. Strive, unpub. data). Although the exact parental virus and source of the RCV-A1-like nonstructural genes were not identified, the distance between the nonstructural genes of RHDVa in Australia and currently sampled RCV-A1 nonstructural genes (83.3%-84.4% nt identity, Figure 4B) suggests either that the recombination event occurred before the virus arrived in Australia with an as-yet uncharacterized calicivirus or that recombination did occur in Australia but the parental virus belongs to a divergent RCV-A1 lineage that has not yet been identified.
The exotic RHDVa variant was first detected in Australia in samples from an RHD outbreak in December 2013 in domestic rabbits in a suburb of northern Sydney (32). During January 2014-March 2015, this virus was detected in a further 8 outbreaks. Detections were localized to southeastern Australia, predominantly in the Sydney basin. In contrast, the initial spread of the CAPM V-351 virus in 1995 was estimated to be 50 km/wk (1). Seven of the 9 documented RHDVa outbreaks were in domestic rabbits, although this probably reflects sampling bias, because very limited surveillance occurred in NSW and ACT in wild rabbits during the sampling period.
The seroprevalence of RHDV and RCV-A1 antibodies was shown to be very high in wild rabbit populations in southeastern Australia (49), which may limit RHDVa infection and transmission rates in wild rabbits compared with unvaccinated domestic rabbits, due to partial or complete immunological cross-protection from previous infection with RHDV, RCV-A1, or both. In addition, RHDV outbreaks are much more readily detected in domestic rabbit facilities where animals are checked daily, whereas outbreaks in wild rabbits often go unnoticed, particularly as  Table 4), suggesting infection and transmission of RHDVa was occurring in wild populations. In Mulligan's Flat (outbreak 11), we recovered RHD-Va RNA from a healthy shot rabbit following an observed decline in rabbit numbers. Subsequent serologic analyses from a sample of healthy shot rabbits suggests that high levels of cross-reacting isotype antibodies (IgA, IgM, and IgG), in combination with low or negative test results for specific competition ELISAs for RHDV and RCV-A1, may be used as a tool to infer previous exposure of a rabbit population to RHDVa. This particular serologic profile may prove useful for large-scale field epidemiology studies attempting to track the spread and impact of RHDVa in the Australian landscape, even in the absence of a specific serologic test. Furthermore, our study demonstrates that PCR analysis of bile in recovering, recently RHDVinfected (IgM-positive) rabbits can be an additional tool to determine to which virus a wild rabbit population had been exposed.
Several of the RHDVa cases in domestic rabbits were reported to have occurred in previously vaccinated rabbits, although details about the vaccination regimens and intervals were difficult to retrieve. The RHDVa variant JN165234/XA/China/2010, which closely resembles the VP60 sequence of the RHDVa from Australia we describe, was reported to be able to partially overcome protection induced by vaccination against classic RHDV (14). However, the vaccine used in that study was not a commercial preparation, and dose rates were not provided (14). The same study demonstrated that vaccination against classic RHDV induced 100% protection against another RHDVa virus from China (14). RHDV2 is considered to be a second lagovirus serotype, but the OIE Terrestrial Manual does not discriminate between RHDV and RHDVa with respect to vaccination and conservatively recommends that breeder rabbits be vaccinated every 6 months to ensure protection in commercial rabbitries (11). Australian authorities have issued updated vaccination recommendations subsequent to the detection of RHDV2 to reflect those of the OIE (11). Ultimately, a polyvalent vaccine effective against all viruses circulating in Australia (RHDV and RHDV2) would be preferred for domestic rabbits.
Before the detection of RHDVa, the deliberately released CAPM V-351 and its descendants were the only RHDV variants circulating in Australia, despite both RHDVa and RHDV2 circulating overseas since 1997 and 2010, respectively (7,12,18). As such, the RHDV antigen capture ELISA was routinely used for diagnostic purposes. Given that the RHDVa virus we report here shows little to no reactivity in this ELISA, it is recommended that the monoclonal antibodies used in diagnostic ELISAs be updated, as described in the OIE Terrestrial Manual, or that alternative detection methods such as RT-PCR be used for RHD diagnosis in Australia (11). At this stage, it is unclear what effects the presence of this new virus will have on Australia's wild rabbit populations and on biocontrol efforts using RHDV. With the recent (March 2017) release of another RHDVa virus (an RHDVa from South Korea), it is critical that ongoing monitoring and surveillance systems are in place to explore the interactions 28 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 24, No. 1, January 2018  of these new viruses as they establish under Australian field conditions (50). In this context, it will be particularly interesting to determine if the recombinant RHDVa described here persists in the face of the exotic incursion of RHDV2, which has reportedly replaced RHDV and RHDVa viruses in many parts of Europe (24)(25)(26). Once again, the epidemiology of these viruses demonstrates the utility of Australia's rabbit populations serving as a model system for disease emergence, viral competition, and evolution.