RT-PCR Assays for Seven Serotypes of Epizootic Haemorrhagic Disease Virus & Their Use to Type Strains from the Mediterranean Region and North America

Epizootic haemorrhagic disease virus (EHDV) infects wild ruminants, causing a frequently fatal haemorrhagic disease. However, it can also cause bluetongue-like disease in cattle, involving significant levels of morbidity and mortality, highlighting a need for more rapid and reliable diagnostic assays. EHDV outer-capsid protein VP2 (encoded by genome-segment 2 [Seg-2]) is highly variable and represents the primary target for neutralising antibodies generated by the mammalian host. Consequently VP2 is also the primary determinant of virus “serotype”, as identified in virus neutralisation tests (VNT). Although previous reports have indicated eight to ten EHDV serotypes, recent serological comparisons and molecular analyses of Seg-2 indicate only seven EHDV “types”. Oligonucleotide primers were developed targeting Seg-2, for use in conventional RT-PCR assays to detect and identify these seven types. These assays, which are more rapid and sensitive, still show complete agreement with VNT and were used to identify recent EHDV isolates from the Mediterranean region and North America.

The current report by the International Committee for the Taxonomy of Viruses (ICTV) [2] indicates that there are eight serotypes of EHDV [2], which can be identified by the specificity of neutralising antibody interactions with the outer-surface of the virus (particularly with VP2) in 'virus neutralisation tests' (VNT) [14][15][16]. VP2 is the most variable of the EHDV proteins, showing sequence differences that correlate with both the serotype and geographic origin (topotype) of the virus lineage [2,15,17]. However, recent molecular and serological data presented by Anthony et al [15] showed that the EHDV isolate from Nigeria (NIG1967/01) that is used at Institute for Animal Health (IAH) as a reference strain for EHDV-3, cross-neutralises in VNTs with the reference isolate of EHDV-1 (New Jersey -USA1955/01), indicating that they in fact belong to the same 'serotype'. EHDV strain '318', which was first isolated in Bahrain in 1983 (BAR1983/01) and subsequently recovered from sentinel calves in the Sudan [15,18,19] was (for both isolates) identified as a western strain of EHDV-6 (EHDV-6[W]). Consequently only seven distinct types of EHDV were identified [15].
Since 1998, multiple different BTV types have caused widespread disease outbreaks in Europe [36][37][38][39][40]. These events, which demonstrate the presence of vector-competent Culicoides species across the entire region, have been linked to climate change [41]. During the same period, multiple, previously exotic BTV types, were also isolated in the southern states of the USA [42]. The primary involvement of EHDV with wild rather than domesticated species, and a lack of rapid/reliable molecular assays, has lead to lack of information concerning the prevalence and distribution of different EHDV types, particularly in the Middle East and the Indian sub-continent. However, it is clear that EHDV has recently also expanded its geographic distribution and economic importance, with outbreaks in livestock in Israel, Morocco, Algeria, Turkey, the USA and Reunion Island during 2004-2009 [34,35,43-54]. These events suggest that EHD represents an increasing threat to southern Europe and the USA, and are reflected by its declaration as a notifiable disease by OIE [55].
Conventional procedures for the detection and typing of EHDV involve virus isolation in embryonated chicken eggs (ECE) or sheep, and/or adaptation to cell culture, followed by 'typing' using reference antiserum against the established serotypes in 'virus neutralisation tests' (VNT) [56]. Alternatively antiserum from infected animals can also be tested for its ability to neutralise reference strains of each virus serotype in 'serum neutralisation tests' (SNT) [57]. However, these procedures are costly, time consuming and require access to reference viruses and/or antisera. Experience with BTV suggests that they are also prone to crossreactions between types, particularly in areas where multiple virus serotypes are co-circulating [58,59]. RT-PCR based molecular assays have recently been developed for individual BTV types [60][61][62][63] that are more rapid and more reliable than these conventional serological methods and have become standard diagnostic tests in many BTV reference laboratories. However, to improve EHDV surveillance, there is a need for similar molecular assays to identify and 'type' EHDV infections.
Detection of EHDV has previously been reported using either real time or conventional RT-PCR assays, containing virusspecies/serogroup specific primers and probes targeting conserved genome segments 5, 7 and/or 10 (NS1, VP7 and/or NS3 genes respectively) [14,[64][65][66][67][68]. This report describes the development of a new panel of EHDV 'type-specific' primers targeting genome segment 2 (Seg-2 -the VP2 gene), and their use to type and characterise EHDV strains involved in recent disease outbreaks in the Mediterranean region and North America.

Field isolates
The different EHDV and other orbivirus isolates (listed in Table 1 and 2) were grown in BHK-21 cell monolayers, until 100% cytopathic effect (CPE) was observed. Each isolate is stored in the IAH 'dsRNA virus reference collection' (dsRNA-VRC) with a collection number composed of 'country code, year, and the number of the isolate in that year from that country [48,49]. The samples used in these studies were taken from naturally infected animals in the field, by qualified veterinarians, as part of normal veterinary care and diagnostic testing procedures in the respective countries.
Moroccan Full details of these isolates are presented in [15], and are also available on the dsRNA virus reference collection website [48]. *Institute for Animal Health, Pirbright. **Serotypes listed are as reclassified by Anthony et al [15], with NIG1967/01 as serotype 1, not serotype 3. They have also characterised the unclassified strain '318' as a western topotype of serotype 6 (EHDV-6[W]). ***The only available sequence for a western strain of EHDV-7 was derived from a mixed virus isolate. doi:10.1371/journal.pone.0012782.t001 Seg-9, as described by the manufacturer (Laboratoire Service International -LSI).

Extraction of Viral dsRNA
The dsRNA extracted either from EHDV infected cell pellets using the TrizolH (Invitrogen) [69], or from cell-free supernatant fluid, using the QIAamp Viral RNA Mini Kit from Qiagen according to manufacturer's instructions. Each isolate was processed separately to minimize the risk of sample-to-sample contamination.

Selection of serotype-specific oligonucleotide primers
Comparisons of full-length VP2 sequences already obtained for strains of the different EHDV serotypes [14,15]  intra-typic conservation and hetero-typic variation) in the Seg-2, for the design of serotype-specific RT-PCR primers. Wherever possible, these analyses included data for multiple isolates of each type, from different geographical origins, to help ensure that the resulting primers would amplify Seg-2 from different topotypes within the same serotype.
At least two primer-pairs were designed for each serotype (Table S1 -provided as online supporting information file). Primer-pairs (identified by the letter 'A', indicating 'all') were designed to detect any isolate of each EHDV serotype. Other primers were designed targeting Seg-2 of either eastern or western topotypes of each type (identified by the letters 'E' and 'W', respectively; Table S1 -supporting information). Individual primers were named as follows: 'EHDV' representing Epizootic haemorrhagic disease virus, followed by a number indicating the serotype; 'S2' to indicate they target Seg-2; 'F' representing forward, or 'R' representing reverse primers; and a number corresponding to the nucleotide binding position of the primer in Seg-2.

Reverse transcription and amplification of Seg-2
Complementary DNA (cDNA) copies were synthesised (reverse transcribed) from denatured dsRNA templates, then amplified by a 'single-tube' method, using the One-Step RT-PCR kit (Qiagen, Courtaboeuf, France), using conditions described for amplification of cDNAs from Seg-2 of bluetongue virus (BTV) by Mertens et al [63]. Alternatively a single-tube reaction contained the Super-Script TM III one-step RT-PCR system (Invitrogen) and high fidelity platinumH Taq was used with RNA templates extracted from blood or cell culture supernatant, as described by Maan et al [37]. In some cases, full length cDNA copies of EHDV Seg-2 were also synthesised and sequenced, as described by Maan et al [70].

Sequencing of Seg-2 amplicons
For confirmation of RT-PCR typing results, amplified cDNA products from Seg-2 were gel-purified and sequenced directly on an Applied Biosystems BigDye ddNTP 3730 capillary sequencer, using the PCR primers, as sequencing primers. The resultant sequences were analysed and aligned using the Clustal W program [71] and MAFFT (www.ebi.ac.uk/Tools/mafft/) and then subjected to either BLAST analysis and/or in silico comparisons to Seg-2 sequence data available for the 7 EHDV serotypes (including both western and eastern strains). Phylogenetic comparisons were made using Neighbour-Joining (NJ) algorithms and the trees were constructed in MEGA ver. 4.1 software [72], using nucleotide sequences in a pair-wise deletion, p-distance algorithm, and bootstrapped using 500 replicates.

Development of RT-PCR assays and their evaluation for type specificity
Multiple primer-pairs were designed based on sequence data for Seg-2, targeting each of the seven EHDV serotypes (Table 1 and  Table S1 -supporting information). These primers were tested in RT-PCR reactions containing RNA extracted from BHK-21 cells infected with the different EHDV serotypes (including both eastern and western topotypes, where available - Table 1). The resulting cDNAs were analysed by 1% agarose gel electrophoresis (AGE), to check both their sizes and relative level of synthesis. In each case, cDNAs of the expected size were generated from RNA belonging to the homologous EHDV serotype ( Fig. 1 and 2). However, in most cases, significant amounts of cDNA of the expected size were only amplified in reactions containing RNA of the same topotype. None of the primer-pairs tested amplified cDNAs from RNA of even the most closely related of the heterologous EHDV serotypes, or from the other related orbiviruses tested (Table 1 and 2) (data not shown). In some serotypes it was possible to design primers using the combined sequences from both eastern and western strains within an individual serotype, which successfully amplified Seg-2 sequences from both major topotypes. These initial evaluations confirm in silico analyses, indicating that within the range of EHDV strains tested (Table 1 and 2) the primers listed in Table S1 (supporting information) are 'type-specific'. EHDV-1 (and 3). Three primer-pairs ('1W1', '1W2' and '1W3') were designed, based on sequence data for Seg-2 of two western strains of EHDV-1 (USA1955/01 [Ac. No. AM744978] and NIG1967/01 [Ac. No. AM745008]). NIG1967/01 was previously classified as EHDV-3 but is now included in EHDV-1 [15]. Eastern topotype specific primer-pairs ('1E1' and '1E2') were also designed and evaluated, using Seg-2 sequence data generated for EHDV-1/AUS1995/02 (Ac. No. HM156728). By combining Seg-2 sequence data from both eastern and western strains, two primer-pairs ('1A1' and '1A2') were designed for amplification of cDNAs from Seg-2 of any of the three EHDV-1 isolates.

Typing of North African field isolates of EHDV from 2006
RNA was extracted from BHK-21 cells infected with Algerian or Moroccan orbivirus isolates from 2006 (ALG2006/02 and MOR2006/05, respectively). These viruses were distinct from those used in the initial design of the EHDV 'type-specific' primers described above. The RNAs were initially tested using an EHDV 'group-specific' RT-PCR assay, containing primers targeting Seg-7 ( [14], Table S1 -supporting information). In each case cDNA bands of the expected size were generated (results not shown) providing the initial identification of both isolates as EHDV. These RNA preparations were also tested using a commercially available EHDV 'group-specific' real-time RT-PCR (from LSI), giving cycle threshold (CT) values of 12 and 14 respectively, confirming their identity as EHDV.
ALG2006/02 and MOR2006/05 were subsequently tested by RT-PCR, using 'type-specific' primers listed in Table S1 (supporting information). cDNAs of the predicted sizes (573 and 468 base pairs [bps] respectively) were generated from either RNA preparation with primer-pairs '6W1' and '6W2'. No cDNAs were generated with any of the other primer-pairs, including those directed against eastern strains of EHDV-6 ('6E1'), providing the first indication that both isolates belong to the western topotype of EHDV-6 (EHDV-6[W]).
Full-length cDNA were also synthesised from Seg-2 of ALG2006/02 and MOR2006/05, using an 'anchor-primer' ligated to the RNA 39 termini, then sequenced, as described by Maan et al [70]. The Seg-2 sequences generated were identical, reflecting the origin of both isolates from the same North African outbreak in 2006. These full length Seg-2 sequences were compared to EHDV strains belonging to each of the different Figure 3. Neighbour-joining tree, showing relationships in Seg-2 between reference and field strains of EHDV. The tree was constructed using distance matrices, generated using the p-distance determination algorithm in MEGA 4.1 (500 bootstrap replicates) [72]. EHD viruses split into four distinct groups (A-D) based on Seg-2 sequences, reflecting serological relationships between virus strains [15]. Reference strains of EHDV are shown in black and field isolates are shown in blue font. Isolate designations: IAH 'dsRNA virus reference collection number composed of 'country code, year, and the number of the isolate in that year from that country [48,49] Fig. 3). This not only confirms that both virus isolates belong to EHDV-6[W], but also that primer-pairs '6W1' and '6W2' can be used to identify field isolates of this serotype and topotype. Subsequent comparisons to the reference strains of EHDV serotypes 1-8 [15] showed highest levels of identity (72.6% nt and 77% aa) to Seg-2 of EHDV-6[E] (AUS1981/07) and lowest levels (46% nt and 31% aa) to Seg-2 of EHDV-2/CAN1962/01.

EHDV from Israel 2006
Ten orbivirus isolates were made in embryonated chicken eggs (ECE) from the blood of cows showing clinical signs of disease in Israel during 2006 (ISR2006/01 to ISR2006/09 and ISR2006/ 13), then grown in BHK-21 cells. RNA extracted from the infected cell pellet was tested using an EHDV group-specific RT-PCR assay (targeting genome Seg-7 - [14], Table S1 -supporting information), generating a cDNA band of the expected size (1162 bp), and providing the first identification of these viruses as EHDV (result not shown). ISR2006/13 was tested by VNT, showing a reduction in titre of 2.5 Log 10 TCID 50 with reference antisera to EHDV-7[E] (AUS1981/06), but no significant reduction in titre with reference sera to the other EHDV serotypes. A reduction of .2.0 Log 10 is usually considered to be indicative of the homologous serotype in this type of assay, confirming the identity of ISR2006/13 as EHDV-7.
cDNA copies were subsequently synthesised from Seg-2 of ISR2006/13, then sequenced using an 'anchor spacer' ligated to 39 termini of the RNA, as described by Maan et al [70]. The upstream terminal 2867 nt/880 aa of Seg-2/VP2 were compared to strains of different EHDV serotypes (Fig. 3), showing highest levels of identity (75% nt/79% aa respectively) with the Australian reference strain of EHDV-7 (AUS1981/06). The detection of EHDV-7 sequences using anchor ligation indicates that EHDV-7 represented at least the dominant component of the virus population in the sample tested.
Recent studies of BTV Seg-2 [73] have detected up to 31% nt variation within a single serotype. Although the levels of nucleotide sequence variation between and within individual EHDV serotypes have not yet been fully explored, initial studies suggest that the same may be true for EHDV [15]. The relationship between Seg-2 of ISR2006/13 and that of other EHDV strains (,30.5% nt identity), is therefore consistent with its serological identification as EHDV-7 (with which it shares 75% nt identity). However, the relationship is quite distant, suggesting that the Australian and Israeli viruses (AUS1981/06 and ISR2006/13) belong to different topotypes (E and W respectively). Seg-2 of ISR2006/13 also showed slightly lower levels of identity with three strains of EHDV-2 (69.3% nt/68.2% aa identity to AUS1979/05; 69.5% nt and 68% aa identity to JAP1959/01; and 67.8% nt/68% aa identity to CAN1962/01). EHDV-2 is the most closely related heterologous serotype to EHDV-7 (Group A -as described by Anthony et al [15] , Fig. 3).
An antiserum sample from Israeli cattle (from 2006), which neutralised 100 TCID 50 of the reference strain of EHDV-7[E] (AUS1981/06) at a dilution of 1/320, gave a similar result in SNT with EHDV-2 (the most closely related serotype to EHDV-7). These serological typing results were therefore regarded as inconclusive, but suggest that the animal could previously have been infected with both EHDV-2 and EHDV-7. No results are available for the reactivity of this serum sample with EHDV-6.

EHDV isolates from the USA 2006
The USA 2006 isolates were initially tested at the Southeastern Cooperative Wildlife Disease Study (SCWDS), USA, for virus neutralisation (VNT) using antisera to EHDV-1 and 2 [17], with negative results. The virus isolates were then tested at USDA National Veterinary Services Laboratories (NVSL) for virus neutralisation using antisera to different EHDV serotypes. A significant reduction in cytopathology was observed in wells inoculated with these viruses and antiserum to EHDV-6 (CSIRO 753). No neutralisation was observed with antisera to EHDV serotypes 1 (previously regarded as type 3), 4, 5, 7 and 8.

EHDV isolates from Turkey 2007
RNAs extracted from six virus isolates derived from diagnostic blood samples taken from cattle in Turkey during 2007 (TUR2007/01 to TUR2007/06, listed in Table 1) and a bovine lung and spleen tissue homogenate sample (TUR2007/07 - Table 1), were tested by conventional group-specific RT-PCR assays (targeting Seg-7, [14], Table S1 -supporting information). Although these samples were contaminated with bacteria and appeared to be somewhat degraded, a positive result (with a cDNA band of the expected size) was obtained. Real time diagnostic RT-PCR assays for BTV [74], were negative for all of the samples, although group-specific real time RT-PCR assay for EHDV (from LSI) also gave a positive result, with CT values of 20.32-27.5. One serum sample (number A162/07 13) from a different animal also gave a weak +ve result for EHDV specific antibodies by ELISA [75][76][77]. Subsequent RT-PCR assay, using primers '6W1' and '6W2' (Table S1 -supporting information), amplified cDNAs of the appropriate sizes from isolates TUR2007/01, TUR2007/02, TUR2007/03, TUR2007/04, and TUR2007/07, providing the first identification of both the virus type and topotype (Fig. 6 -Panels A-D). No amplification was detected with the other serotype specific primers including '6E1'.
Sequencing and phylogenetic analyses of the cDNAs generated (nt 120-684) from TUR2007/02 and TUR2007/07, showed highest levels of nt sequence identity with BAR1983/01 (96.  (Fig. 3). The different isolates of EHDV-6 from Turkey from 2007 showed a minimum level 98.9% nt sequence identity among themselves and 97.3% identity to Seg-2 (120-684 nt) of EHDV-6 ISR2006/13, indicating that they all share a recent common ancestry and suggest they represent continuation of a single outbreak.

Discussion
Serological 'typing' methods (VNT and SNT), based on the specificity of interactions between neutralising antibodies and the outer virion surface (particularly VP2), have been used for many years and with considerable success, as primary diagnostic assays to monitor outbreaks of orbiviral diseases and identify the 'serotype' of the virus involved. However, these assays are slow and time consuming (taking weeks), and require access to stocks of well characterised reference antisera and reference virus strains for each serotype [78]. Serological assays are also unable to distinguish individual strains within a single serotype, and are therefore unsuitable for high precision epidemiological studies that need to differentiate closely related virus lineages, or topotypes.
Until recently the limited availability of reference 'antisera and isolates' of individual EHDV serotypes and topotypes, have made it difficult to determine the precise serological relationships that exist between different orbivirus strains, by conventional serological methodologies (VNT and SNT assays). Consequently there is some uncertainty in the literature concerning the total number of EHDV serotypes [2,64,[85][86][87]. Between 8-10 serotypes/strains of EHDV (serotypes 1-8, EHDV-318 and Ibaraki virus (IBAV) have been suggested in the literature [2], although the most recent study by Anthony et al [15] has grouped type 1 and 3 together (as EHDV-1) and strain 318 as a western strain of type 6, making a total of only seven distinct EHDV serotypes.
In contrast 25 serotypes of BTV have been identified by serological assays as well as phylogenetic comparisons of Seg-2/ VP2 [63,73,[88][89][90]. Recently a sequence database for Seg-2 of BTV [73], was used to design oligonucleotide primers and probes for conventional and real-time RT-PCR assays to identify individual BTV types (conventional assays -Mertens et al [63]; in preparation; real-time assays available from LSI).
Anthony et al [15,91,92] compared full-genome sequence data for 'reference' strains of EHDV, confirming that (like BTV) variations in Seg-2/VP2 correlate with virus serotype. These data have supported the design of EHDV 'serotype-specific' primers for use in the RT-PCR assays described here. Multiple primer sets targeting Seg-2 of each EHDV serotype were evaluated, in particular for their ability to identify and discriminate between reference strains or available field isolates of different EHDV serotypes and topotypes, as well as other related orbiviruses, confirming their specificity.
These Seg-2 specific RT-PCR assays and primers, together with associated sequencing studies, allow individual EHDV serotypes (and topotype) to be identified with greater accuracy, sensitivity and more rapidly than by conventional serological typing methods. It is now also possible to identify individual strains and lineages within a single EHDV serotype in a manner that is impossible by conventional serological typing methods, facilitating intense surveillance. These assays can also provide Seg-2 amplicons for sequencing -allowing comparisons to identify serotype, topotype and strain with great precision. Although primer sets were identified that were specific for Seg-2 of each serotype, these vary in their ability to prime cDNA synthesis from different isolates. This reflects high levels of sequence variability in Seg-2 of BTV and related orbiviruses, even within a single serotype, depending largely on the geographical origin (topotype) of the isolate [15,36].
In recent years (2006)(2007), several EHD outbreaks caused by different serotypes that have occurred in different locations around the world, have been identified using the type-specific assays described here. These include EHDV-6[W] recently recognized as an emerging pathogen of cattle in Algeria, Morocco (in 2006) and Turkey (2007) [34]. EHDV-7[W] was also identified in Israel in 2006 [35]. EHDV isolates from white-tailed deer in Indiana and Illinois [17], were confirmed as type 6 and for the first time identified as an eastern topotype of type 6 (EHDV-6[E]), a serotype and topotype originally isolated in Australia. Genetic characterization of EHDV-6[E] Indiana (304-06) [17] showed that it is a reassortant virus, with outer capsid proteins that determine serotype (VP2 and VP5) derived from an exotic strain of EHDV-6[E], although how this strain arrived in the USA remains uncertain. The remaining structural and non-structural proteins are derived from an indigenous strain of EHDV-2 (Alberta), making it difficult to distinguish this strain from other American strains of EHDV using group specific serological or RNA sequence based assays (RT-PCR) that target these more conserved genome segments and proteins.
These studies highlight a need to establish reference strains for both eastern and western topotypes of each EHDV serotype, with associated sequence data for Seg-2 (as listed in Table 1; [48]). The new EHDV 'type-specific' assay systems described here provide a valuable tool for future studies, to provide more information concerning the persistence and movements of individual virus strains in the field. The assays were designed based on sequence data from a limited pool of available EHDV strains, primarily to distinguish between isolates of the different EHDV serotypes, and the resulting assays are successful in this respect. However, it is possible that EHDV strains will be identified in the future that cross-react by RT-PCR assay with heterologous serotypes, or fail to amplify using these type-specific primers. This will require further sequence analyses/comparison and primer redesign to maintain the efficacy and specificity of these assays. The fine-scale differentiation between individual EHDV strains within each serotype and topotype depends on the availability and characterisation of well documented isolates, such as those stored and maintained in the reference collection at IAH Pirbright [48,49].
Conventional and real-time RT-PCR assays will inevitably increase the speed and reliability of laboratory-based diagnostic testing for EHDV. It is not easy to compare the sensitivity and specificity of type-specific and group-specific, conventional and real-time RT-PCR assays, targeting different orbivirus genome segments, in a meaningful way. However, in our experience of routine detection of BTV and EHDV RNA in diagnostic tissue samples, or cell-culture materials [14,[64][65][66][67][68]74], those samples that give CT values of ,30 by real-time PCR (equivalent to ,100 copies -data supplied by LSI for their Seg-9 based EHDV groupspecific assay), will also reliably generate cDNA amplicons by conventional RT-PCR assay targeting the appropriate Seg-2 [63]. The serotype specific assays described here can be used to detect and 'type' individual strains of EHDV in the absence of reference antisera for each serotype. Unlike real-time RT-PCR assays, the conventional type-specific primers can also be used to sequence any of the cDNA amplicons that are generated, providing further and definitive proof of the relationships between individual virus strains in Seg-2. Phylogenetic comparisons of the resulting EHDV Seg-2 sequences will help determine the worldwide prevalence and distribution of different EHDV serotypes and topotypes, improving our understanding of the epidemiology and transmission of the virus. The data generated will inform the design and implementation of future control strategies, including vaccination.