CORE Metadata, citation and similar papers at core.ac.uk

Provided by Elsevier - Publisher Connector

Virology 433 (2012) 236–244

Contents lists available at SciVerse ScienceDirect

Virology

journal homepage: www.elsevier.com/locate/yviro

Malakal virus from Africa and Kimberley virus from are geographic variants of a widely distributed ephemerovirus

Kim R. Blasdell a, Rhonda Voysey a, Dieter M. Bulach a, Lee Trinidad a, Robert B. Tesh b, David B. Boyle a, Peter J. Walker a,n

a CSIRO Livestock Industries, Australian Health Laboratory, 5 Portarlington Road, Geelong VIC 3220, Australia b Center for Biodefense and Emerging Infectious Diseases, Department of Pathology, University of Texas Medical Branch, Galveston, TX, USA

article info abstract

Article history: Kimberley virus (KIMV) is an -borne rhabdovirus that was isolated in 1973 and on several Received 7 June 2012 subsequent occasions from healthy cattle, mosquitoes (Culex annulirostris) and biting midges (Culicoides Returned to author for revisions brevitarsis) in Australia. Malakal virus (MALV) is an antigenically related rhabdovirus isolated in 1963 1 August 2012 from mosquitoes ( uniformis) in . We report here the complete genome sequences of Accepted 3 August 2012 KIMV (15442 nt) and MALV (15444 nt). The genomes have a similar organisation (30-l-N-P-M-G-G - Available online 25 August 2012 NS a1-a2-b-g-L-t-50) to that of bovine ephemeral fever virus (BEFV). High levels of amino acid identity in Keywords: each gene, similar gene expression profiles, clustering in phylogenetic analyses of the N, P, G and L Kimberley virus proteins, and strong cross-neutralisation indicate that KIMV and MALV are geographic variants of the Malakal virus same ephemerovirus that, like BEFV, occurs in Africa, Asia and Australia. Rhabdovirus Crown Copyright & 2012 Published by Elsevier Inc. All rights reserved. Ephemerovirus Genome sequence Gene expression

Introduction from Australia (Blasdell et al., 2012; Bourhy et al., 2005; Calisher et al., 1989; Kuzmin et al., 2006; Schmidt et al., 1965). Ephemeroviruses are arthropod-borne rhabdoviruses of ungu- Assignment of these viruses as species in the genus Ephemerovirus lates, primarily infecting cattle and other ruminants. Bovine is based primarily on serological cross-reactions and genome ephemeral fever virus (BEFV) causes a debilitating febrile illness sequence data. Ephemeroviruses have characteristically large and in cattle and water buffaloes (St. George and Standfast, 1988; complex genomes comprising the five structural protein genes that Walker, Blasdell, and Joubert, 2012) and has been isolated on are common to all rhabdoviruses (N, P, M, G and L), and up to six many occasions from biting midges (Culicoides spp.) and mosqui- additional ORFs located between the G and L genes (Walker et al., toes (Anopheles bancrofti), and from cattle in Africa, Asia and 2012). In all ephemeroviruses, the Ggeneisfollowedbyasecond

Australia (van der Westhuizen, 1967; Walker et al., 2012). Other non-structural glycoprotein gene (GNS) which encodes a protein with known ephemeroviruses include Berrimah virus (BRMV) and significant homology to rhabdovirus G proteins and appears to have Adelaide River virus (ARV), which were each isolated in 1981 arisen by gene duplication or recombination (Walker et al., 1992,

from healthy sentinel cattle in Australia but have not been 2011; Wang and Walker, 1993). The GNS gene is followed by a gene associated with disease (Gard et al., 1983; Gard et al., 1984). encoding a putative viroporin (a1) and two genes of unknown Kotonkan virus (KOTV) and Obodhiang virus (OBOV) have also function (a2andb)(Wang et al., 1994). In BEFV and KOTV, but recently been characterised as ephemeroviruses (Blasdell et al., not other ephemeroviruses sequenced to date, the b gene is followed 2012). KOTV was isolated from biting midges (Culicoides spp.) in by a third gene (g) of unknown function (McWilliam et al., 1997). and shown to cause an ephemeral fever-like illness in KOTV is unique in that the g gene is followed by the d gene which cattle (Kemp et al., 1973a; Tomori et al., 1974). OBOV was isolated encodes a protein with significant amino acid sequence similarity to from mosquitoes (Mansonia uniformis) in Sudan and has been the pleckstrin homology domain of coactivator-associated arginine shown to be closely related antigenically and genetically to ARV methyltransferase 1 (CARM1) (Blasdell et al., 2012). Serological and phylogenetic data suggest at least two other viruses may be members of the genus Ephemerovirus (Bourhy et al., 2005; Calisher et al., 1989). Kimberley virus (KIMV) was n Corresponding author. Fax: þ61 3 5227 5000. first isolated from a pool of mosquitoes (Culex annulirostris) E-mail address: [email protected] (P.J. Walker). collected in the remote Ord River Valley of Western Australia in

0042-6822/$ - see front matter Crown Copyright & 2012 Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.virol.2012.08.008 K.R. Blasdell et al. / Virology 433 (2012) 236–244 237

1973 (Liehne et al., 1981) and has subsequently been isolated on exception of a 20–21 nt overlap at g–L gene junction, as also several occasions in Australia from biting midges (Culicoides occurs in BEFV (Table S1). brevitarsis) and healthy sentinel cattle (Cybinski and Muller, A pairwise alignment of MALV and KIMV genomes indicated 1990; Cybinski and Zakrzewski, 1983; Zakrzewski and Cybinski, overall 90.6% nucleotide sequence identity. The 2 nt difference in 1984). Although it has not been associated with disease out- the genome length occurred as a consequence of various inser- breaks, KIMV neutralising antibodies have been detected in cattle, tions and deletions which were located primarily in the IGRs and water buffaloes, goats and horses in Australia, , Papua the sequences corresponding to 30 untranslated regions of the and (Cybinski and Zakrzewski, 1983; Jiang and mRNAs. In most cases, the corresponding ORFs were identical in Yan, 1989; Soleha et al., 1993). In Australia, its distribution is length but a single nucleotide substitution (A to G) in the a2 gene similar to that of BEFV and there is evidence that KIMV infection of KIMV (strain CS368) at the position corresponding to the MALV may offer some short-term protection against ephemeral fever a2 initiation codon has shifted the start of the ORF 84 nt down- during BEFV outbreaks (Cybinski, 1987). Malakal virus (MALV) stream to the next available methoinine codon (Fig. 2A). Similarly, was isolated from mosquitoes (M. uniformis) collected along the a single nucleotide substitution in the KIMV (strain CS368) b ORF banks of the White Nile River near Malakal in Sudan in 1963 has introduced a termination codon such that it is truncated by (Schmidt et al., 1965). MALV cross-reacts strongly in indirect 225 nt relative to that of the b ORF of MALV (Fig. 2B). In order to fluorescent antibody tests with BEFV, KIMV and several other determine if these variations are distinguishing features of KIMV, ephemeroviruses, including OBOV which was isolated from the accessory gene regions were analysed in four other KIMV M. uniformis collected one month after MALV at the same site in isolates with more limited passage histories. In three of these Sudan, and Puchong virus (PUCV) which was isolated from isolates (CS756, CS1186 and CS1501), the a2 and b ORFs were M. uniformis in Selangor, in 1965 (Calisher et al., 1989). identical in length to those of MALV. In the fourth isolate No other information is currently available on the geographic (CS1413), the b ORF was intact but the a2 ORF was truncated distribution, vertebrate host range or pathogenicity of MALV. by a single nucleotide deletion close to the N-terminus. In this paper, we report the complete genome sequences, gene The analysis suggests that the a2 and b ORFs KIMV CS368 and expression profiles and serological and phylogenetic relationships the a2 ORF of KIMV CS1413 have been corrupted by mutation of MALV and KIMV. The data indicate that the viruses should be during adaptation to cell culture. considered to be geographic variants of the same ephemerovirus species which appears to have a widespread distribution in Africa, Asia and Australia. Deduced amino acid sequences of MALV and KIMV proteins

Analysis of the deduced amino acid sequences indicated that Results (other than the a2 and b proteins of KIMV CS368, see above), the predicted molecular weights of each of the MALV and KIMV Nucleotide sequences of MALV and KIMV genomes proteins are very similar (Table S1). Furthermore, a global comparison of all available ephemerovirus amino acid sequences The complete genome sequences of MALV strain SudAr 1149-64 (Table S2) indicated that MALV and KIMV are the most closely (15444 nt) and KIMV strain CS368 (15442 nt) were determined by related viruses, sharing 485% identity in all proteins and 497% de novo assembly using Illumina sequencing (101 bp reads, paired identity in the N and L proteins. Lowest sequence identity was in ends). The average read depth coverage was 2283 for MALV and the P (86.5%) proteins and a1 proteins (88.9%) which generally 3889 for KIMV, with coverage exceeding 1000 reads for most of are less well conserved amongst the ephemeroviruses. As is each genome. Sequences at the extreme termini were confirmed by reflected in the similar genome organisations, MALV and RACE and Sanger sequencing. The genome organisations of the KIMV are most closely related to BEFV in all encoded proteins viruses are identical with 10 open reading frames (ORFs) arranged (Table S2). In pairwise comparisons, OBOV and ARV are the next 0 0 in the order 3 -l-N-P-M-G-GNS-a1-a2-b-g-L-t-5 in negative polar- most closely related ephemeroviruses but they are far more ity (Fig. 1). This genome organisation is similar to that of BEFV divergent than MALV and KIMV, sharing highest sequence iden- except for the absence of alternative ORFs in the a2(a3) and tity in the N proteins (87.3%) and lowest identity in the a2 P(P0) genes (McWilliam et al., 1997). In each virus, the ORFs are proteins (30.1%). flanked by highly conserved transcription initiation (UUGUCC) and In BEFV, the G protein is the target of neutralising antibodies, transcription termination/polyadenylation (DUAC[U]7) sequences, containing multiple independent neutralisation sites that are both except for a1anda2 which occur as consecutive ORFs within linear (G1) and non-linear (G2 and G3) (Kongsuwan et al., 1998). the a gene (Table S1). The MALV and KIMV leader and trailer Alignment of the MALV and KIMV G protein amino acid sequences sequences are identical (51 nt and 57 nt, respectively). The inter- (94.9 % identity) indicated that all 18 cysteine residues and all five genic regions (IGRs) are relatively long (22–65 nt), with the potential N-glycosylation sites are conserved and that most

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15kb

G α1 α2 βγ BEFV 3’ N P M G L 5’ P’ α3 G α1 α2 βγ MALV 3’ N P M G L 5’

G α1 α2 βγ KIMV 3’ N P M G L 5’

Fig. 1. Schematic representation of the genome organisations of BEFV, MALV and KIMV. Block arrows indicate the location of ORFs, including overlapping ORFs (light shading) and the characteristic ephemerovirus accessory genes (darker shading). 238 K.R. Blasdell et al. / Virology 433 (2012) 236–244

Fig. 2. ClustalX alignment of the nucleotide and deduced amino acid sequences of (A) the a gene region and (B) the b gene region of the genomes of MALV strain SudAr 1149-64, KIMV strain CS368 and KIMV strain CS1501. Initiation codons and termination codons are indicated as corrupted sequences in KIMV strain CS368 that correspond to the locations of the a2 ORF initiation codon and the b ORF termination codon of the other strains.

variations occurred in the N-terminal signal peptide and the neutralising BEFV MAb found to cross-react with KIMV (Cybinski acidic C-terminal endodomain (Fig. 3). Furthermore, alignment et al., 1990). with the BEFV G protein sequence indicated that, in the regions As is characteristic of ephemeroviruses, the MALV and KIMV G corresponding to the major neutralisation sites, the MALV and proteins also share significant amino acid homology with the

KIMV sequences share almost complete identity, with only single corresponding GNS proteins (34.6%). Alignment of the G and GNS amino acids variations in each of the three sites. In contrast, proteins (92.6 % identity) indicated that, of the 18 predicted MALV and KIMV shared poor amino acid sequence identity with cysteine residues found in the G protein ectodomain, 14 are

BEFV in neutralisation sites G1 and G2, and in two of the three conserved in each of the GNS proteins, suggesting that they adopt elements of the major conformational site, G3 (Fig. 3). However, a similar folded structure (data not shown). Although KIMV 217 in one element of site G3, corresponding to amino acids F to GNS lacks one of the 10 potential N-glycosylation sites found in 231 P of the BEFV G protein (Kongsuwan et al., 1998), the level of MALV GNS, the variation occurs at a locus at which the two sites in amino acid sequence identity was observed to be relatively high MALV overlap, suggesting they are not both utilised. Most (73.3%). In this region, BEFV, MALV and KIMV share a stretch N-glycosylation sites are located in the C-terminal region of the 229 of identical amino acids including residue E which occurs at GNS protein ectodomains which is also the site of most amino acid the binding site of monoclonal antibody (MAb) 8D3, the only sequence variations between MALV and KIMV. K.R. Blasdell et al. / Virology 433 (2012) 236–244 239

Fig. 3. Alignment of BEFV, MALV and KIMV G proteins illustrating the conservation of cysteine residues and the location of regions corresponding to neutralising antigenic sites G1, G2 and G3. The alignment was generated in Clustal X using default parameters. Cysteine residues are shaded and numbered according to the system described in Walker and Kongsuwan (1999). Antigenic sites are boxed. Predicted N-glycosylation sites are underlined. Predicted signal peptides and transmembrane domains are also indicated. Other fully conserved (n) amino acids are indicated between the relevant sequences. The location of BEFV amino acid E229 that was selected using antigenic site G3b MAb 8D3 is shaded in black.

The MALV and KIMV small accessory proteins (a1, a2, b and g)    share generally low but identifiable amino acid sequence identity with the cognate proteins of other ephemeroviruses (Table S2). Each a1 protein, like those of other ephemeroviruses, is predicted to contain a central hydrophobic transmembrane domain and a highly basic C-terminal domain that is characteristic of viropor- ins. The sequences of the other putative accessory proteins were not indicative of their likely functions.

MALV transcription profile

An anchor PCR utilising oligo[dT] and gene-specific primers was used to identify polyadenylated viral transcripts expressed in MALV-infected Vero cells (Fig. 4). The sizes of the amplified products were consistent with transcription initiation and tran- Fig. 4. Detection of MALV transcripts in infected Vero cells by anchor PCR using an scription termination/polyadenylation at each of the identified oligo(dT) primer and a sequence-specific primer in each ORF. The dots mark the TI and TTP sequences, generating nine monocistronic mRNAs and positions of amplified products of the size anticipated for mRNAs termination at one bicistronic (a1–a2) transcript. Although products of smaller the TTP site associated with each gene. 240 K.R. Blasdell et al. / Virology 433 (2012) 236–244 sizes were also amplified, sequence analysis demonstrated that Phylogenetic relationships they were produced by mis-priming of oligo[dT] on A-rich regions within the respective ORFs. As identical TI and TTP sequences are Neighbour-joining trees were generated from ClustalW align- present in the KIMV genome, its transcription profile was not ments of the complete G protein sequences of MALV, KIMV and a analysed. large set of available G protein sequences of other animal rhabdoviruses. The analysis placed MALV and KIMV within the ephemerovirus clade, supported by high bootstrap values Expression of MALV and KIMV proteins in infected cells (Fig. 6A). Phylogenetic analyses were also conducted using ClustalW alignments of partial nucleotide and deduced amino Proteins expressed in MALV- and KIMV-infected Vero cells at acid sequences of the G gene (1305 nt; 435 aa) and P gene 1, 18, 24 and 48 hpi were analysed by SDS-PAGE and immuno- (432 nt; 144 aa) of MALV, six Australian KIMV isolates, six blotting using MALV mouse ascitic fluid (MAF) and a BEFV M Australian BEFV isolates, and the ephemeroviruses BRMV, ARV, protein monoclonal antibody (MAb 20A6) that was shown pre- OBOV and KOTV. Single BEFV isolates from Israel, , viously to cross-react with KIMV (Cybinski et al., 1990). Across all and Turkey were also included in the partial G protein analysis. time points for each virus, specific reactions were detected with Wongabel virus (WONV) was used as the outgroup. In each proteins migrating at sizes corresponding approximately to those analysis, MALV consistently clustered in a clade with KIMV, with predicted for the major structural proteins (N, P, M and G) (Fig. 5). the sequence diversity between MALV and KIMV isolates appar- As expected from deduced amino acid sequences, the cognate ently similar to that of BEFV isolates from different geographic proteins of each virus were similar in size. BEFV MAb 20A6 cross- origins and far less than that observed between the ephemer- reacted with KIMV M protein (Fig. 5) but failed to cross-react with oviruses BEFV and BRMV, or between OBOV and ARV (Fig. 6B, C). MALV M protein (not shown). For each virus, strong reactions Estimates of sequence diversity based on analysis of the partial G were also detected with 25 kDa proteins. In MALV-infected genes indicated that the MALV and KIMV isolates displayed less cells, a 17 kDa protein was also detected and there was a very divergence at the nucleotide level than the selected BEFV isolates weak reaction with a 13 kDa protein at 18 hpi (not evident in (7.6% and 9.6%, respectively), but slightly more divergence at the the illustration). The sizes of the 17 kDa and 13 kDa proteins amino acid level (5.1% and 4.7%, respectively) (Table 2). In the P correspond approximately to the predicted molecular weight of gene, sequence divergence was higher but the variation at both the b and g proteins. In KIMV-infected cells, the upper of these nucleotide and amino acid levels between KIMV isolates and bands (designated as ‘y’ in Fig. 5) appeared to be smaller than the MALV (14.3% and 19.9%, respectively) was much lower than 17 kDa KIMV protein and the 13 kDa band was not detected. between BEFV and BRMV (43.0% and 52.7%- respectively), the

Table 1 Serological relationships Virus neutralisation tests of MALV, KIMV and other BEF serogroup viruses.

Virus Titre of antibody Virus neutralisation tests were conducted to determine the serological relationships between KIMV, MALV and other ephe- BRMV BEFV MALV KIMV ARV OBOV KOTV meroviruses (Table 1). Solid two-way cross-neutralisation was detected between MALV and KIMV, which were indistinguishable BRMV 440 480 – – – – – BEFV 20 1280 20 – – – – by this method. As observed by Cybinski and Zakrzewski (1983), MALV –n – 1280 80 – – – there was no cross-neutralisation between KIMV and BEFV. KIMV – – 1280 80 – – – However, a very weak one-way cross-reaction (1/20) was ARV – – – – 640 80 – detected against BEFV using MALV MAF. No cross-neutralisation OBOV ––––4041280 – was detected between MALV and KIMV with either BRMV, ARV, KOTV –––––– 4640 OBOV or KOTV. As reported previously, there was partial two-way BRMV, Berrimah virus; BEFV, bovine ephemeral fever virus; MALV, Malakal virus; cross-neutralisation between BEFV and BRMV and between ARV KIMV. Kimberley virus; ARV, Adelaide River virus; OBOV, Obodhiang virus; KOTV, n and OBOV (Blasdell et al., 2012; Gard et al., 1983). kotonkan virus. Dash signifies titre o 20.

MALV - MALV MAF KIMV - MALV MAF KIMV - BEFV M MAb hpi hpi hpi kDa kDa BEFV 1182448C 1182448C 1182448C M MAb 188 188 98 98 G/ 62 GNS 62 49 N 49

38 P 38 28 M 28 M x 17 y 17 14 14

Fig. 5. SDS-PAGE and immunoblotting of proteins expressed in MALV- and KIMV-infected Vero cells or mock-infected cells (C). Proteins were detected at 1, 18, 24 and 44 hpi by immunoblotting using MALV-specific specific mouse ascetic fluid (MAF) or BEFV M protein monoclonal antibody (MAb) 20A6. The reaction of MAb 20A6 with semi-purified BEFV is also shown. Bands corresponding approximately in size to the estimated molecular weights of the major viral structural proteins (G, N, P and M) and the non-structural glycoprotein (GNS) are indicated. Smaller bands that appear to be induced or were detected by immunoblotting in infected cells but not mock-infected cells are also indicated (x, y). K.R. Blasdell et al. / Virology 433 (2012) 236–244 241

STRV Vesiculovirus SCRV

Novirhabdovirus Sigmavirus

0.1

TUPV

Lyssavirus

DURV 100 100 100 97 95 97 94 MOUV OVRV 100 99 41 69

100 79 100 BEFV Hart Park group 100

100 BRMV

Ephemerovirus 100 100 KIMV Tibrovirus MALV

OBOV ARV KOTV

66 BEFV CS967 KIMV CS368 48 BEFV CS968 KIMV CS756 56 BEFV CS660 43 83 0.05 BEFV BB7721 0.1 KIMV CS1413 69 65 BEFV V8363 47 KIMV CS1501 BEFV CS1865 99 100 83 KIMV CS1186 BEFV Japan Hirado-9 100 KIMV CS1542 BEFV Taiwan TN/2004/124 100 98 BEFV Israel ISR04 61 MALV 96 BEFV Turkey CP62 BRMV 100 90 BRMV BEFV CS1865 MALV 100 100 BEFV BB7721 KIMV CS1186 97 BEFV CS660 KIMV CS368 48 BEFV CS968 67 KIMV CS756 17 31 KIMV CS1501 32 BEFV V8363 26 KIMV CS1413 81 BEFV CS967 23 KIMV CS1546 KOTV 100 ARV 60 100 ARV OBOV KOTV OBOV WONV WONV

Fig. 6. Phylogenetic analysis of animal rhabdovirus proteins. Panel A: Full-length G protein amino acid sequences rhabdoviruses representing the genera Vesiculovirus, Novirhabdovirus, Lyssavirus,proposedgeneraTibrovirus and Sigmavirus, Wongabel virus, Ngaingan virus and Flanders virus (Hart Park group), as well as unclassified viruses Oak Vale virus (OVRV), Moussa virus (MOUV), Durham virus (DURV), tupaia rhabdovirus (TUPV), Simiperca chuatsi rhabdovirus (SCRV) and sea trout rhabdovirus (STRV), and the ephemeroviruses bovine ephemeral fever virus (BEFV), Berrimah virus (BRMV), Malakal virus (MALV), Kimberley virus (KIMV), kotonkan virus (KOTV), Adelaide River virus (ARV) and Obodhiang virus (OBOV). Clusters of viruses representing the assigned or recently proposed genera and the Hart Park group have been collapsed and shaded to simplify the illustration. Panel B: Phylogenetic analysis of a 435 aa region of the G protein ectodomain sequences of selected isolates of MALV, KIMV, BEFV, BRMV, ARV, OBOV, KOTV and outgroup WONV. Panel C: Phylogenetic analysis of a 144 aa region of the P protein sequences of selected isolates of MALV, KIMV, BEFV, BRMV, ARV, OBOV, KOTV and outgroup WONV. The neighbour-joining consensus bootstrap trees (1000 repetitions) were constructed in MEGA5 from a Clustal W sequence alignments using default parameters. most closely related ephemeroviruses that are currently classified set of viruses (data not shown) confirmed that the nucleotide and as different species (Table 2). Analysis of partial sequences of the amino acid sequence divergence amongst MALV and KIMV N gene (699 nt; 233 aa) and L gene (402 nt; 134 aa) of the same isolates is similar to that detected between BEFV isolates. 242 K.R. Blasdell et al. / Virology 433 (2012) 236–244

Table 2 Nucleotide and deduced amino acid sequence divergence of partial G and P genes of BEFV, BRMV, MALV and KIMV.

Virus comparison Sequence divergence (%)

Partial G Partial P

Nucleotide Amino Acid Nucleotide Amino Acid

KIMV vs MALV 7.6 5.1 14.6 20.1 Australian KIMV isolates only 1.5 0.9 4.6 8.3 All BEFV isolates 10.3 4.9 Not available Not available Australian BEFV isolates only 5.1 3.0 6.7 10.7 BEFV vs BRMV 24.3–25.9 17.1–19.2 43.0 52.9–55.1

Discussion the continent (Cybinski and Zakrzewski, 1983). KIMV has never been shown to be associated with disease outbreaks. Indeed, KIMV has The genus Ephemerovirus currently comprises the species been isolated from healthy cattle during an outbreak of ephemeral Bovine ephemeral fever virus (type species), Berrimah virus and fever in an Australian dairy herd and it has been suggested that KIMV Adelaide River virus (Dietzgen et al., 2011). Kotonkan virus and and other ephemeroviruses may limit BEF epizootics by providing Obodhiang virus have also recently been proposed to be members temporary protection through viral interference (St. George, 1985). of the genus (Blasdell et al., 2012). Data presented here indicate Challenge experiments may resolve this question. However, clinical that the MALV and KIMV genomes are typical of ephemeroviruses, ephemeral fever is also known to occur in Australia in the absence of encoding, in addition to the five canonical rhabdovirus structural seroconversion to BEFV and it is possible that some strains of KIMV proteins, a non-structural class I transmembrane glycoprotein contribute to the epizootiology of the disease. The pathogenicity of

(GNS), a putative viroporin (a1) and three other putative accessory MALV is unknown and, although BEFV has been reported to occur in proteins (a2, b and g) of unknown function. The genomes share Egypt, and Nigeria (Davies et al., 1990; Kemp et al., 1973b; 490% overall nucleotide sequence identity and are almost iden- Madbouly et al., 2006), little is currently known of the epizootiology tical in length (2 nt variation between the sequenced isolates), of ephemeral fever in Sudan or other parts of Central or East Africa. employ the same genome organisation and transcription strategy, Experimental infections may assist in establishing the pathogenic and encode proteins that are almost identical in size. The sequence potential of MALV and KIMV for cattle and other ruminants. divergence between cognate genes and proteins of MALV and A previous experimental challenge of cattle with KIMV resulted in KIMV isolates from Africa and Australia is similar to that observed seroconversion but no disease (M.F. Uren and T.D. St. George, between BEFV isolates from Australia, East Asia and the Middle unpublished data). However, the experiment was conducted using East, and is far less than occurs between BEFV and BRMV which the cell culture-adapted CS368 strain and it is known that adaptation represent the most closely related of the currently designated of BEFV to cell culture rapidly attenuates its pathogenicity in cattle. ephemerovirus species. Furthermore, MALV and KIMV are indis- Comparison of the accessory gene regions of MALV and the five tinguishable from each other in virus neutralisation tests, demon- available KIMV isolates indicated that the sequence of strain CS368 strating solid two-way cross-neutralisation, but they are clearly has been corrupted by point mutations that eliminate the a2 distinguishable serologically from other ephemeroviruses. Current initiation codon and introduce a truncating termination codon in species demarcation criteria for ephemeroviruses are based on low the b ORF. Similar modifications to the accessory genes have been or no cross-neutralisation, up to 91% sequence identity in the N reported for the BEFV strain BB7721, in which the b ORF is also protein, and variations in genome organisation and transcription truncated and expression of the g gene is attenuated by corruption of control sequences (Dietzgen et al., 2011). There is, therefore, the upstream b gene TTP sequence (McWilliam et al., 1997). In ARV sufficient evidence here to establish that MALV and KIMV repre- strain DPP61, expression of the GNS, a1anda2 genes appears to be sent geographic variants of the same virus and should be classified attenuated by corruption of upstream TTP sequences resulting in as a single ephemerovirus species. Although MALV was the first to expression of a G–GNS–a polycistronic transcript (Wang and Walker, be isolated, we propose the name Kimberley virus for the new 1993). Like KIMV strain CS368, these BEFV and ARV isolates have species on the basis that it is better characterised biologically. been passaged extensively and plaque cloned in BHK and/or Vero Based on genome organisation and nucleotide and amino acid cells, indicating that the accessory proteins are redundant and sequence identity data, KIMV and MALV are most closely related to potentially detrimental to growth in cell culture. Conversely, the BEFV. There was also evidence of low-level cross-neutralisation of accessory proteins may well have important functions in vivo and BEFV by MALV polyclonal mouse ascites fluid. This is consistent attenuation of their expression during cell culture adaptation may with a previous observation that natural BEFV infection of a cow explain the accompanying loss in pathogenicity. with pre-existing KIMV antibody induced a secondary KIMV neu- Similar expressed protein profiles were detected in cells tralising antibody response (Cybinski, 1987). A single neutralising infected with MALV and KIMV using MALV MAF. In addition to BEFV MAb (8D3) directed at antigenic site G3b has also been shown proteins of similar size to the major structural proteins (G, N, to cross-react with KIMV (Cybinski et al., 1990). Indeed, the specific P and M), several bands corresponding to smaller proteins were amino acid that was shown to vary in BEFV escape mutants selected detected. However, these could not be assigned unambiguously as using MAb 8D3 (E229) is conserved in both KIMV and MALV, as are accessory proteins. An intensely staining 25 kDa band detected elements of the surrounding sequence of the G3 antigenic site in MALV-infected cells (and less prominently in KIMV-infected (Kongsuwan et al., 1998). Amongst Australian BEFV isolates, we cells) was significantly larger that the predicted sizes of the a1, have also observed that amino acid K203 (which also forms part of a2, b or g proteins. A 17 kDa band detected in MALV-infected the G3 site) is under strong positive selection and this may be driven cells was similar in size to that predicted for the b protein but it by pre-existing KIMV antibody (P.J. Walker, unpublished data). was also detected in KIMV-infected cells (strain CS368) in which In Australia, BEFV and KIMV both infect cattle and are known to the b ORF was shown to be severely truncated. A 13 kDa band have a similar distribution throughout most of the north and east of detected in MALV-infected cells was similar in size to the a1 and K.R. Blasdell et al. / Virology 433 (2012) 236–244 243 g proteins but it reacted very weakly and could not be detected at Virus neutralisation tests all in KIMV-infected cells. Thus, although the transcription pro- files indicated that the small accessory genes are transcribed in Virus neutralisation tests were performed as described infected cells, further studies using mono-specific antisera will be previously (Blasdell et al., 2012; Tian et al., 1987). The method required to confirm expression at the protein level. of Reed and Muench (1938) was used to calculate infectivity titres Interestingly, the KIMV M protein was detected using BEFV which were determined as the lowest dilution of antibody at MAb 20A6 which had previously been shown to cross-react with which neutralisation was observed. KIMV, BRMV and ARV (Cybinski et al., 1990) but the MAb failed to react with MALV M protein. There are only seven amino acid RNA extraction, PCR-select cDNA subtraction and RACE differences between the KIMV and MALV M protein sequences. Two of these are in a short sequence of six residues in the Total RNA extraction from infected cells and the modified C-terminal domain (P193PDLDK) that are fully conserved in BEFV, PCR-select cDNA subtraction for enrichment for viral sequences BRMV and KIMV, suggesting this may be the site of the cross- were conducted as described by Gubala et al. (2010; 2008).ThePCR- reactive epitope. However, this site is not conserved in ARV and select method used WONV as the driver and three different scanning of the sequence failed to identify any other linear sites restriction enzymes (Alu I, Hae III and Rsa I) were used to construct that could account for the reported pattern of cross-reactivity. the libraries. The protocol for the rapid amplification of cDNA ends The known geographic distribution of KIMV includes Australia, has been described previously (Blasdell et al., 2012; Li et al., 2005). , Indonesia and at least two provinces of China Primers used in this protocol are shown in Table S4. (Cybinski and Zakrzewski, 1983; Daniels et al., 1995; Jiang and Yan, 1989; Soleha et al., 1993). Evidence presented here indicates that RT-PCR of partial N, P, G, L and accessory protein region sequences geographic distribution of this virus also extends to Central Africa and that the range of potential vectors includes biting midges cDNA was synthesised from total RNA using the Superscript III (C. brevitarsis) and at least two species of (C. annulirostris first-strand synthesis system (Invitrogen) and 50 mM random and M. uniformis). BEFV has also been isolated from biting midges hexamers at 50 1C for 60 min followed by 70 1C for 15 min. (including C. brevitarsis) and several species of mosquito, and has a Primers used in this protocol are shown in Table S5. PCR geographic distribution that includes Africa, Asia and Australia amplification of partial N and P genes was performed using the (St. George and Standfast, 1988; Walker et al., 2012). Although the GoTaq PCR system (Promega) using 2.5 ml cDNA, 5 ml 5X GoTaq role of biting midges in ephemerovirus transmission is questionable Buffer, 3 ml MgCl2, 0.5 ml 10 mM dNTPs, 1 ml each primer (10 mM), (Kirkland, 1993; St George, 1993; Walker et al., 2012), BEFV and KIMV 0.125 ml GoTaq polymerase and 11.5 ml nuclease-free H2O. The do appear to share a similar ecology, perhaps originally evolving in cycling conditions were as follows: 1 cycle 95 1C for 2 min; 35 Africa as separate lineages in different wild ruminant species and cycles 95 1C30s,551C30s,721C 1 min; 1 cycle 72 1C 7 min; 4 1C then adapting to cattle and spreading to Asia and Australia through hold. PCR amplifications for the partial G and L genes and for the the eastward movement of livestock. Further studies are required to accessory protein region were performed using the LongAmp Taq better define the vertebrate host range and extent of the geographic PCR system (New England BioLabs) using 3 ml cDNA, 5 ml5X distribution of KIMV, and to determine if other African ephemer- LongAmp Buffer, 0.75 ml 10mM dNTPs, 1 ml each primer (10 mM), oviruses, such as KOTV and OBOV, also occur throughout the same 1 ml LongAmp Taq and 13.25 ml nuclease-free H2O. Cycling con- geographic range and contribute to the epizootiology of ephemeral ditions for these PCRs were as follows: 1 cycle 95 1C for 2 min; fever in Asia and Australia. 35 cycles 95 1C30s,X1C30s,651C Y s; 1 cycle 65 1C 10 min; 4 1C hold. Temperatures and durations represented by X and Y are shown in Table S5. Materials and methods DNA sequencing and sequence analysis Viruses and cells Sanger sequencing was performed using the ABI PRISM 3130xl The origins of virus isolates used in his study are shown in Genetic Analyzer in conjunction with the BigDye Terminator v. 3.1 kit Table S3. Passage histories of ARV, BEFV, BRMV, KIMV, KOTV and (Applied Biosystems, USA). High-throughput DNA sequencing was OBOV isolates used for the virus neutralisation tests (VNTs) have performed using the Illumina GAIIx platform at Micromon (Monash been described previously (Blasdell et al., 2012). Prior to sequen- University, Clayton, Australia) according to the manufacturer’s proto- cing KIMV (strain CS368) was passaged 13 times in BHK-21cells col and analyses were performed as described previously (Blasdell and twice in Vero cells, plaque-cloned three times in Vero cells et al., 2012). The sequences described in this paper have been and passaged a further 14 times in BHK-BSR cells. MALV (strain deposited in GenBank under accession numbers JQ941664- SudAr 1149-64) was passaged once in suckling mice and twice in JQ941707. Alignments and phylogenetic analyses were conducted Vero cells. Viruses were grown in Vero or BHK-BSR cells at 37 1C. using MEGA version 5.0 (Tamura et al., 2011). BHK-BSR cells were cultured in Basal Medium Eagle (BME) and Vero cells grown in Eagle’s Minimum Essential Medium (EMEM), Transcription profiling each supplemented with 10 mM HEPES, 2 mM L-glutamine, 137 mM streptomycin, 80 U/ml penicillin, and either 5% (growth Transcription profiling was conducted by PCR as described medium) or 2.5% (maintenance medium) foetal calf serum. previously (Blasdell et al., 2012). Primers and annealing tempera- tures are shown in Table S6. Antisera SDS-PAGE and immunoblotting of viral proteins Bovine BEFV immune serum, ARV rabbit antiserum, KOTV, OBOV, BRMV and KIMV MAFs, and negative control bovine serum have been SDS-PAGE and immunoblotting were performed as described described (Blasdell et al., 2012). MALV MAF was produced as previously (Blasdell et al., 2012; Gubala et al., 2008) except that described by Palacios et al. (2011). BEFV M protein mouse MAb primary antibodies were used at dilutions of 1:1000 (MALV MAF), 20A6 has been described previously (Cybinski et al., 1990). 1:500 (KIMV MAF) and 1:10000 (MAb 20A6), and membranes 244 K.R. Blasdell et al. / Virology 433 (2012) 236–244 were developed using the NOVEX Chemiluminescent Substrate Young, D., Hoffmann, D. (Eds.), Bovine Ephemeral Fever and Related Rhabdo- Reagent Kit (Invitrogen) at room temperature for 5 min. viruses. ACIAR, Canberra, pp. 33–37. Kongsuwan, K., Cybinski, D.H., Cooper, J., Walker, P.J., 1998. Location of neutralizing epitopes on the G protein of bovine ephemeral fever rhabdovirus. J. Gen. Virol. 79, 2573–2581. Acknowledgments Kuzmin, I.V., Hughes, G.J., Rupprecht, C.E., 2006. Phylogenetic relationships of seven previously unclassified viruses within the family Rhabdoviridae using The support of the Australian Biosecurity CRC for Emerging partial nucleoprotein gene sequences. J. Gen. Virol. 87, 2323–2331. Li, Z., Yu, M., Zhang, H., Wang, H.Y., Wang, L.F., 2005. Improved rapid amplification Infectious Diseases is gratefully acknowledged. We thank Steven of cDNA ends (RACE) for mapping both the 5’ and 3’ terminal sequences of Davis and Lorna Melville for the transfer of reagents from paramyxovirus genomes. J. Virol. Methods 130, 154–156. Berrimah Veterinary Laboratory, Darwin. RBT was supported by Liehne, P.F.S., Anderson, S., Stanley, N.F., Liehne, C.G., Wright, A.E., Chan, K.H., the U.S. National Institutes of Health contract HHSN27220 Leivers, S., Britten, D.K., Hamilton, N.P., 1981. Isolation of Murray Valley encephalitis virus and other arboviruses in the Ord River Valley 1972–1976. 1000040I/HHSN27200004/DO4. Aust. J. Exp. Biol. Med. Sci. 59, 347–356. Madbouly, H.M., Hegazi, A.G., El Shabrawy, M., Tamam, S.M., Gamil, G.S., 2006. Isolation and identification of bovine ephemeral fever virus from cattle in Appendix A. Supporting information middle Egypt. Vet. Med. J. Giza 54, 877–886. McWilliam, S.M., Kongsuwan, K., Cowley, J.A., Byrne, K.A., Walker, P.J., 1997. Genome organization and transcription strategy in the complex GNS-L inter- Supplementary data associated with this article can be found in genic region of bovine ephemeral fever rhabdovirus. J. Gen. Virol. 78, the online version at http://dx.doi.org/10.1016/j.virol.2012.08.008. 1309–1317. Palacios, G., Tesh, R., Travassos da Rosa, A., Savji, N., Sze, W., Jain, K., Serge, R., Guzman, H., Guevara, C., Nunes, M.R., Nunes-Neto, J.P., Kochel, T., Hutchison, S., Vasconcelos, P.F., Lipkin, W.I., 2011. Characterization of the Candiru antigenic References complex (Bunyaviridae: Phlebovirus), a highly diverse and reassorting group of viruses affecting humans in tropical America. J. Virol. 85, 3811–3820. Reed, L.J., Muench, H., 1938. A simple method of estimating fifty percent end- Blasdell, K.R., Voysey, R., Bulach, D., Joubert, D.A., Tesh, R.B., Boyle, D.B., Walker, P.J., points. Am. J. Hyg. 27, 493–497. 2012. Kotonkan and Obodhiang viruses: African ephemeroviruses with large Schmidt, J.R., Williams, M.C., Lulu, M., Mivule, A., Mujombe, E., 1965. Viruses and complex genomes. Virology 425, 143–153. Bourhy, H., Cowley, J.A., Larrous, F., Holmes, E.C., Walker, P.J., 2005. Phylogenetic isolated from mosquitoes collected in the Southern Sudan and Western relationships among rhabdoviruses inferred using the L polymerase gene. . East African Virus Research Institute Report vol. 15, 24. J. Gen. Virol. 86, 2849–2858. Soleha, E., Daniels, P.W., Sukarsih, Sendow, I., 1993. A study of bovine ephemeral Calisher, C.H., Karabatsos, N., Zeller, H., Digoutte, J.-P., Tesh, R.B., Shope, R.E., fever group rhabdoviral infections in West Java, Indonesia. In: St. George, T.D., Travassos da Rosa, A.P.A., St. George, T.D., 1989. Antigenic relationships among Uren, M.F., Young, P.L., Hoffmann, D. (Eds.), Bovine Ephemeral Fever and rhabdoviruses from vertebrates and hematophagous . Intervirology Related Rhabdoviruses,ACIAR Proceedings, vol. 44. ACIAR, Canberra, pp. 45–50. 30, 241–257. St. George, T.D., 1993. The natural history of ephemeral fever of cattle. In: Cybinski, D.H., 1987. Homologous and heterologous antibody reactions in sera St. George, T.D., Uren, M.F., Young, P.L., Hoffmann, D. (Eds.), Bovine Ephemeral from cattle naturally infected with bovine ephemeral fever group viruses. Vet. Fever and Related Rhabdoviruses, ACIAR Proceedings, vol. 44. ACIAR, Canberra, Microbiol. 13, 1–9. pp. 13–19. Cybinski, D.H., Muller, M.J., 1990. Isolation of arboviruses from cattle and at two St. George, T.D., 1985. Studies on the pathogenesis of bovine ephemeral fever in sentinel sites in Queensland, Australia, 1979–1985. Aust. J. Zool. 38, 25–32. sentinel cattle. I. Virology and serology. Vet. Microbiol. 10, 493–504. Cybinski, D.H., Walker, P.J., Byrne, K.A., Zakrzewski, H., 1990. Mapping of antigenic St. George, T.D., Standfast, H.A., 1988. Bovine ephemeral fever. In: Monath, T.P. sites on the bovine ephemeral fever virus glycoprotein using monoclonal (Ed.), The Arboviruses: Epidemiology and Ecology, vol. 2. CRC Press, Boca antibodies. J. Gen. Virol. 71, 2065–2072. raton, pp. 71–86. Cybinski, D.H., Zakrzewski, H., 1983. The isolation and preliminary characteriza- Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M., Kumar, S., 2011. MEGA5: tion of a rhabdovirus in Australia related to bovine ephemeral fever virus. Vet. molecular evolutionary genetics analysis using maximum likelihood, evolu- Microbiol. 8, 221–235. tionary distance, and maximum parsimony methods. Mol. Biol. Evol. 28, Daniels, P.W., Sendow, I., Soleha, E., Hunt, S.N.T., Bahri, S., 1995. Australian– 2731–2739. Indonesian collaboration in veterinary arbovirology – a review. Vet. Microbiol. Tian, F.G., Jiang, C.L., Zakrzewski, H., Davis, S.S., 1987. A comparison of a Chinese 46, 151–174. and an Australian strain of bovine ephemeral fever virus. Aust. Vet. J. 64, 159. Davies, F.G., Ochieng, P., Walker, A.R., 1990. The occurrence of ephemeral fever in Tomori, O., Fagbami, A., Kemp, G., 1974. Kotonkan virus: experimental infection of Kenya, 1968–1988. Vet. Microbiol. 22, 129–136. white Fulani calves. Bull. Epizoot. Dis. Afr. 22, 195–200. Dietzgen, R.G., Calisher, C.H., Kurath, G., Kuzman, I.V., Rodriguez, L.L., Stone, D.M., van der Westhuizen, B., 1967. Studies on bovine ephemeral fever. I. Isolation and Tesh, R.B., Tordo, N., Walker, P.J., Wetzel, T., Whitfield, A.E., 2011. Rhabdoviridae. preliminary characterization of a virus from natural and experimentally produced In: King, A.M.Q., Adams, M.J., Carstens, E.B., Lefkowitz, E.J. (Eds.), Virus casesofbovineephemeralfever.OnderstepoortJ.Vet.Res.34,29–40. Taxonomy, Ninth Report of the International Committee on Taxonomy of Walker, P.J., Blasdell, K.R., Joubert, D.A., 2012. Ephemeroviruses: arthropod-borne Viruses. Elsevier, San Diego, pp. 654–681. rhabdoviruses of ruminants, with large and complex genomes. In: Dietzgen, Gard, G.P., Cybinski, D.H., St. George, T.D., 1983. The isolation in Australia of a new R.G., Kuzman, I.V. (Eds.), Rhabdoviruses: Molecular Taxonomy, Evolution, virus related to bovine ephemeral fever virus. Aust. Vet. J. 60, 89–90. Genomics, Ecology, Cytopathology and Control. Horizon Scientific Press, Gard, G.P., Cybinski, D.H., Zakrzewski, H., 1984. The isolation of a fourth bovine Norwich. ephemeral fever group virus. Aust. Vet. J. 61, 332. Walker, P.J., Byrne, K.A., Riding, G.A., Cowley, J.A., Wang, Y., McWilliam, S., 1992. Gubala, A., Davis, S., Weir, R., Melville, L., Cowled, C., Walker, P., Boyle, D., 2010. Ngaingan virus, a macropod-associated rhabdovirus, contains a second glyco- The genome of bovine ephemeral fever rhabdovirus contains two related protein gene and seven novel open reading frames. Virology 399, 98–108. glycoprotein genes. Virology 191, 49–61. Gubala, A.J., Proll, D.F., Barnard, R.T., Cowled, C.J., Crameri, S.G., Hyatt, A.D., Boyle, Walker, P.J., Dietzgen, R.G., Joubert, D.A., Blasdell, K.R., 2011. Rhabdovirus D.B., 2008. Genomic characterisation of Wongabel virus reveals novel genes accessory genes. Virus Res. 162, 110–125. within the Rhabdoviridae. Virology 376, 13–23. Walker, P.J., Kongsuwan, K., 1999. Deduced structural model for animal rhabdovirus Jiang, C.L., Yan, J.D., 1989. Evidence of Kimberley virus infection of cattle in China. glycoproteins. J. Gen. Virol. 80, 1211–1220. Trop. Anim. Health Prod. 21, 85–86. Wang, Y., McWilliam, S.M., Cowley, J.A., Walker, P.J., 1994. Complex genome Kemp, G.E., Lee, V.H., Moore, D.L., Shope, R.E., Causey, O.R., Murphy, F.A., 1973a. organization in the GNS-L intergenic region of Adelaide River rhabdovirus. Kotonkan, a new rhabdovirus related to Mokola virus of the rabies serogroup. Virology 203, 63–72. Am. J. Epidemiol. 98, 43–49. Wang, Y., Walker, P.J., 1993. Adelaide River rhabdovirus expresses consecutive Kemp, G.E., Mann, E.D., Tomori, O., Fabiyi, A., O’Connor, E., 1973b. Isolation of glycoprotein genes as polycistronic mRNAs: new evidence of gene duplication bovine ephemeral fever virus in Nigeria. Vet. Record 93, 107–108. as an evolutionary process. Virology 195, 719–731. Kirkland, P.D., 1993. The epidemiology of bovine ephemeral fever in south-eastern Zakrzewski, H., Cybinski, D.H., 1984. Isolation of Kimberley virus, a rhabdovirus, Australia: evidence for a mosquito vector. In: St. George, T.D., Uren, M.F., from Culicoides brevitarsis. Aust. J. Exp. Biol. Med. Sci. 62, 779–780.