Orchid fleck virus is a rhabdovirus with an unusual bipartite genome

Transcription

1 Journal of General Virology (2006), 87, DOI /vir Orchid fleck virus is a rhabdovirus with an unusual bipartite genome Hideki Kondo, 1 Takanori Maeda, 1 3 Yukio Shirako 2 and Tetsuo Tamada 1 Correspondence Hideki Kondo hkondo@rib.okayama-u.ac.jp 1 Research Institute for Bioresources, Okayama University, Kurashiki , Japan 2 Asian Center for Bioresources and Environmental Sciences, University of Tokyo, Tokyo Japan Received 5 January 2006 Accepted 21 March 2006 Orchid fleck virus (OFV) has an unusual bipartite negative-sense RNA genome with clear sequence similarities to those of nucleorhabdoviruses. The OFV genome consists of two single-stranded RNA molecules, RNA1 and RNA2 that are 6413 and 6001 nt long, respectively, with open reading frame (ORF) information in the complementary sense. RNA1 encodes 49 (ORF1), 26 (ORF2), 38 (ORF3), 20 (ORF4) and 61 kda (ORF5) proteins, and RNA2 encodes a single protein of 212 kda (ORF6). ORF1, ORF5 and ORF6 proteins had significant similarities (21 38 % identity) to the nucleocapsid protein (N), glycoprotein (G) and polymerase (L) gene products, respectively, of other rhabdoviruses, especially nucleorhabdoviruses, whereas ORF2, ORF3 and ORF4 proteins had no significant similarities to other proteins in the international databases. Similarities between OFV and rhabdoviruses were also found in the sequence complementarity at both termini of each RNA segment (the common terminal sequences are 39-UGUGUC---GACACA-59), the conserved intergenic sequences and in being negative sense. It was proposed that a new genus Dichorhabdovirus in the family Rhabdoviridae of the order Mononegavirales should be established with OFV as its prototype member and type species. INTRODUCTION Orchid fleck virus (OFV) was first found in Japan in Cymbidium plants with chlorotic or necrotic fleck symptoms (Doi et al., 1977). It has subsequently been reported in Australia, Brazil, Denmark, Germany, Korea and USA, causing chlorotic or necrotic spots and rings in many genera of the Orchidaceae (Blanchfield et al., 2001; Gibbs et al., 2000; Kitajima et al., 2001). OFV is thus widespread in orchids and is recognized as an important viral pathogen of orchids. OFV is sap-transmissible to Dendrobium and a few species in the families Chenopodiaceae, Solanaceae, Leguminosae and Aizoaceae (Chang et al., 1976; Doi et al., 1977). Its virion has been observed by electron microscopy in OFV-infected leaf tissue (Chang et al., 1976). They are bacilliform and helically constructed measuring nm in diameter, nm in length and with a pitch of 4?5 nm. Partially purified virions from OFV-infected leaves of Cymbidium and Odontoglossum are infectious by sap inoculation to Dendrobium (Chang et al., 1976). 3Present address: College of Bioresource Sciences, Nihon University, Fujisawa , Japan. The GenBank/EMBL/DDBJ accession numbers of the sequences reported in this paper are AB and AB Hydrophobicity plots of the G proteins of OFV and RYSV are available as supplementary material in JGV Online. In ultrathin sections of infected orchid leaf tissue, OFV or OFV-like agents induce a characteristic, intranuclear, electron-lucent viroplasms and a large number of virions have been found in nuclei (Chang et al., 1976; Kitajima et al., 1974; Lesemann & Doraiswamy, 1975). Virions could be seen scattered throughout the viroplasm, and they are often found associated perpendicularly with the inner membrane of the nuclear envelope (Kitajima et al., 2001). In situ immunolabelling assays revealed that the intranuclear viroplasm contains viral structural antigens and may be the site of virion assembly (Kitajima et al., 2001). A number of virions have also been found in spokewheel -like structures inside of extrusion from the nuclear membrane toward the cytoplasm (Chang et al., 1976; Kitajima et al., 1974; Lesemann & Doraiswamy, 1975). Rarely, enveloped virions measuring nm have been found in the endoplasmic reticulum (Chang et al., 1976; Lesemann & Doraiswamy, 1975), but the relationship of these and virions is uncertain. OFV is transmitted by the false-spider mite Brevipalpus californicus Banks in a persistent manner (Kondo et al., 2003). Citrus leprosis virus (CiLV), Coffee ringspot virus (CoRSV) and some other viruses have also been reported to be associated with the Brevipalpus mite (Brevipalpus phoenicis Geijskes), and to have virions that resemble those of OFV in morphology, and to induce similar cytopathic effects (Chagas et al., 2003; Kitajima et al., 2003; Rodrigues G 2006 SGM Printed in Great Britain 2413

2 H. Kondo and others et al., 2003). Based on the shape and size of the virions, OFV and other mite-borne viruses (CiLV and CoRSV) were tentatively placed as unassigned plant rhabdoviruses in the sixth ICTV Report (Wunner et al., 1995), but OFV has subsequently been classified as a totally unassigned virus in the seventh ICTV Report because of the fact that its genome is bipartite (Calisher et al., 2000). In a previous paper, we described briefly that the OFV genome consisted of two molecules: RNA1 and RNA2 (Kondo et al., 2003). In this study, we report detailed data of the complete nucleotide sequence of the OFV genome and show that it resembles that of rhabdoviruses except that the genome is divided. We propose a new genus, Dichorhabdovirus, with OFV as its type species. METHODS Virion purification and RNA extraction. The original OFV isolate (So) from Cymbidium species collected in Okayama Prefecture, Japan was used in this study (Kondo et al., 1995). This virus was propagated in Tetragonia expansa in glasshouses. Locally infected leaves of T. expansa were harvested approximately 324 weeks after inoculation and stored at 280 uc until used. OFV virions were partially purified from infected leaves using the method described by Miranda et al. (2000) for Rice grassy stunt virus. Nucleic acids were extracted from purified virions using SDS followed by phenol extraction and ethanol precipitation. Electrophoretic analysis of viral nucleic acids. Electrophoresis of viral nucleic acids extracted from purified virions was done under non-denaturing and formaldehyde denaturing conditions. Approximately 0?120?5 mg viral nucleic acids and marker DNA (lambda DNA/HindIII digests) or RNA (0?2826?58 kb RNA markers; Promega) were applied to each lane of the gel. Viral RNAs released directly from the virions by heating in 1 % SDS were also analysed by electrophoresis under non-denaturing conditions. Construction of random cdna library of the OFV genome. The heat-denatured RNA was fractionated by electrophoresis in agarose gel and purified from an excised gel slice using RNaid reagents (Bio 101) as instructed by the manufacturer. It was then used as template. Genomic RNA (approx. 2 mg) was transcribed into cdna by AMV reverse transcriptase (Life Sciences) at 42 uc for 1 h using random hexamers as primer. cdna was transcribed into double-stranded DNA using Escherichia coli DNA polymerase I in the presence of RNase H at 12 uc for 2 h and at 21 uc for 1 h. The synthesized double-stranded cdna was made blunt-ended using T4 DNA polymerase. After EcoRI methylase reaction, EcoRI linkers were added to the ends of the cdna and it was then hydrolysed by EcoRI, and cdna molecules larger than 0?5 kb were selected by electrophoresis in agarose gel, as described above. The double-stranded cdna with linkers was ligated into pgem3z vector (Promega) that had been hydrolysed by EcoRI and dephosphorylated. Ninety-seven sequence fragments were obtained from 40 ampicillin-resistant clones using vector primers of both orientations and/or nestedspecific primers. The sequences of the primers used in this study are available upon request. Determination of the sequences of the 3 and 5 termini of OFV genome RNA. The terminal sequences of OFV were identified by rapid amplification of cdna ends (RACE). A ligation-anchored PCR method (Wetzel et al., 1994) was used to determine the 39- terminal sequences of both RNA1 and RNA2. Purified viral genomic RNA was ligated to the 59-phosphorylated primer LF (59-GAGCT- CTGCAGCGGCCGCGAATTCT-39) using T4 RNA ligase (Takara). The reaction mixture was incubated at 16 uc overnight. First-strand synthesis was primed with the primer LR (59-GAATTCGCGGCC- GCTGCAGAGCTC-39; antisense primer of LF). The ligated region of 39-ends were amplified by PCR using the primer LR together with primers specific for OFV RNA1 or RNA2, respectively. The sequences of the 59-terminal regions of both RNA1 and RNA2 were determined by a circular RACE method using 59 Full RACE Core Set (Takara). The cdna was synthesized using the 59-phosphorylatedspecific primer complementary to the region near the 59-end of RNA1 or RNA2, and circularized as single-stranded cdna using T4 RNA ligase. The ligation products were amplified from the junction of 59-terminal cdna by PCR using specific primer pairs for RNA1 or RNA2. The PCR was done using Taq DNA polymerase in a Gene Amp 2400 Thermal Cycler (PE Applide Biosystems). PCR products of the 59- and 39-terminal regions of both RNA1 and RNA2 were cloned into a pgem-t vector (Promega) and the nucleotide sequences were determined from five independent isolates. DNA sequencing and sequence analysis. Plasmid DNA was sequenced by the dideoxynucleotide chain-termination method using the DNA sequencer 377 models (PE Applied Biosystems). The nucleotide sequences were assembled with the AutoAssembler program (PE Applied Biosystems). The nucleotide and amino acid sequences were analysed with GENETYX-MAC software package (Software Development). Databases were searched using the BLAST suite of programs from National Center for Biotechnology Information. Multiple sequence alignments and phylogenetic analyses were done using CLUSTAL W (Thompson et al., 1994). Phylograms were produced using the TREEVIEW program (Page, 1996). The additional sequences used for comparative studies were collected from GenBank as follows. Rhabdoviridae, Nucleorhabdovirus: Maize mosaic virus (MMV; AY618418), Rice yellow stunt virus (RYSV; AB011257), Sonchus yellow net virus (SYNV; L32603), Maize fine streak virus (MFSV; AY618417) and Taro vein chlorosis virus (TaVCV; AY674964); Cytorhabdovirus: Lettuce necrotic yellows virus (LNYV; AJ867584), Northern cereal mosaic virus (NCMV; AB030277) and Strawberry crinkle virus (SCV; AY005146); Lyssavirus: Rabies virus (RABV; M13215); Vesiculovirus: Vesicular stomatitis Indiana virus (VSIV; M20166); Ephemerovirus: Bovine ephemeral fever virus (BEFV; AF234533); Novirhabdovirus: Infectious hematopoietic necrosis virus (IHNV; X89213) and Viral hemorrhagic septicemia virus (VHSV; AF143863); Bornaviridae: Borna disease virus (BDV; U04608); Filoviridae: Marburg virus (MBGV; M92834) and Ebola virus (EBOV; U23458); Paramyxoviridae: Human respiratory syncytial virus (HRSV; M75730), Newcastle disease virus (NDV; X05399), Human parainfluenza virus 2 (HPIV2; X57559) and Sendai virus (SeV; D00053); Morbillivirus: Measles virus (MeV; D10575); Unclassified rhabdovirusrelated virus: Taastrup virus (TV; AY423355); Varicosavirus: Lettuce big-vein associated virus (LBVaV; AB075039). Northern blot analyses. Partially purified OFV genomic RNA was fractionated in a 1?4 % denaturing formaldehyde agarose gel by electrophoresis and transferred onto a nylon membrane (Hybond-N + ; Amersham). The RNA gel blots were hybridized with the strand-specific digoxigenin (DIG)-labelled riboprobes as described below. RNA1-specific riboprobes were synthesized by in vitro run-off transcription with T7 polymerase on pou42-1 [(+) probe] or pou42-2 (in reverse order of pou42h1 insert) [(2) probe]. Similarly RNA2- specific riboprobes were synthesized run-off transcription with T7 polymerase on OF25 [(+) probe] or pof25 (in reverse order of OF25 insert) [(2) probe]. These plasmids were derived from pgem3z. Hybridization and washing were done as advised in the protocols provided by Roche. Unlabelled strand-specific transcripts from the T7 RNA polymerase were used as standards to measure the amounts of genomic and antigenomic RNAs Journal of General Virology 87

3 Genome structure of orchid fleck virus Electron microscopy. Purified virions from sucrose density-gradient fractions were negatively stained with 2 % uranyl acetate, examined in a Hitachi model H-7100 transmission electron microscope and photographed. RESULTS Nucleic acid components of OFV virions OFV virions were purified by differential centrifugation, followed by sucrose density-gradient centrifugation, and the resulting virion preparation gave a single major lightscattering band in the gradient. Electron microscopy of purified preparations showed non-enveloped bacilliform or bullet-shaped virions of approximately nm (Fig. 1a). The purified virions of OFV were infectious when sap inoculated to T. expansa. The morphology and structure of these virions were closely similar in detail to that of OFV and OFV-like virions found in several orchid genera (Begtrup, 1972; Chang et al., 1976; Lesemann & Doraiswamy, 1975). The virions also showed morphological features resembling those found in subviral structures (nucleocapsids) of classical rhabdovirus virions from which the outer membrane had been removed by detergent treatment, even though the OFV virions were smaller than the cores of the rhabdovirus virions (Ziemiecki & Peters, 1976). The nucleic acid preparations extracted from OFV virions using SDS and phenol (Fig. 1b, lane 2) gave one major and two minor bands when they were fractionated by electrophoresis in a non-denaturing gel. The nucleic acid of the major band (named ss) was identified as single-stranded RNA (ssrna) and the minor bands (named ds1 and ds2) were identified as double-stranded RNA (dsrna) based on their susceptibility to digestion with RNase A and resistance to digestion with DNase I (data not shown). When denatured phenol-extracted viral RNAs were fractionated by electrophoresis, two species of ssrna, designated ss1 (6?4 kb) and ss2 (6?0 kb), were detected (Fig. 1c, lanes 2 and 3). ss2 stained less intensely than ss1. Interestingly, the dsrnas could not be detected in the RNA preparations prepared by disrupting the OFV virions in SDS immediately before non-denaturing electrophoresis (Fig. 1d, lane 3). These observations indicate that the detected dsrnas (Fig. 1b) were formed by annealing of complementary viral ssrnas during RNA extraction. Nucleotide sequence analysis of the OFV genome From a random cdna library of OFV RNAs, we obtained two independent continuous sequences (RNA1 and RNA2), and their termini were determined by 39 and 59 RACE. The total number of nucleotides of OFV RNA1 and RNA2 are 6413 and 6001 nt, respectively. Computer-assisted sequence analyses identified that there were significant ORFs in only one orientation of the RNA1 and RNA2 cdnas (Fig. 2 and see below for details). Fig. 1. (a) Transmission electron micrograph of purified preparations of OFV virions. An enlargement of the area outlined by a rectangle. (b and c) Agarose gel electrophoretic patterns of phenol-extracted viral nucleic acids in non-denaturing (b, lane 2) or denaturing conditions (c, lanes 2 and 3). (d) Agarose gel electrophoretic patterns of viral nucleic acids in non-denaturing conditions. Lane 2, phenol-extracted viral RNAs containing 1 % SDS; lane 3, SDS (1 %)-disrupted virion. Lane 1 in (b) and (d), lambda DNA/HindIII digest. Lane 1 in (c), single-stranded RNA markers. In order to determine the polarities of the OFV RNA1 and RNA2, Northern hybridization was done with strand-specific riboprobes. Both coding (positive) and negative-sense riboprobes were hybridized with each RNA species (Fig. 3). To estimate the amounts of negative and positive RNAs in virions, hybridization signals were compared with in vitrotranscribed RNA standards. More negative-sense RNAs were present in virions than positive-sense RNAs. The ratios of negative to positive RNAs were nearly identical for each RNA segment (80 90 % minus and % plus), and the amount of RNA1 was nearly three- to fourfold greater than

4 H. Kondo and others Fig. 2. Schematic representation of the OFV genome. Boxes represent ORFs on the virus-complementary RNA (VC). Position of the translational start and stop of each ORF indicated by numbers above left and right sides of the boxes, respectively. Arrows indicate the putative gene-junction regions on the viral RNA (V; negative) sense. The terminal sequences of the viral RNA are shown. The conserved region of the L protein (RNA-dependent RNA polymerase) of mononegaviruses is indicated by a grey box. pou42-1, pou42-2, OF25 and pof25 (grey bars) indicate locations of probes used for Northern blot hybridization (see the legend to Fig. 3). RNA2 (data not shown). These results show that although both positive and negative RNAs are contained in virions, the negative-sense molecules of both RNA1 and RNA2 are present in excess, indicating that the OFV genome, namely the viral RNAs in the virions, is negative sense. OFV RNA1 contains antisense information for five major non-overlaping ORFs (Fig. 2), which were designated ORF1, Fig. 3. Northern blot analysis of OFV RNAs contained in virions. RNA extracted from purified virions was hybridized to DIG-labelled riboprobes that were transcribed in vitro from inserts of subclones for RNA1 (pou42-1 and pou42-2) and RNA2 (OF25 and pof25), in either the coding (+) or the antisense strand ( ).V, Viral RNA; VC, virus-complementary RNA. ORF2, ORF3, ORF4 and ORF5. These are flanked and separated by untranslated sequences (Fig. 2). RNA2 encodes a single-large ORF (designated ORF6) (Fig. 2). The first AUG codons of ORF3 and ORF5 are located at positions 2577 and 4550, respectively, which are upstream of, or overlap, the putative gene-junction sequences (arrows in Fig. 2) of each ORF. Primer-extension analysis indicated that the transcription initiation sites of the ORF3 and ORF5 genes are located at the putative gene-junction sequences (data not shown). Thus, the first AUG codons of ORF3 and ORF5 are probably non-functional, and the second AUG codons at positions 2682 and 4580 may be the actual translational initiators. Analysis of the encoded proteins of OFV RNA1 A BLAST database search with the amino acid sequence of the ORF1 (49 kda) protein showed similarity to the nucleocapsid (N) proteins of plant nucleorhabdoviruses [TaVCV (27 % identity), RYSV (26 %), MMV (25 %), MFSV (23 %) and SYNV (23 %)] and a cytorhabdovirus LNYV (21 %). Thus, ORF1 probably encodes the rhabdovirus N protein. The ORF5 (61 kda) protein also showed similarity to the glycoproteins (G) of RYSV (24 %). Moreover, the hydropathy profile of the ORF5 protein suggests an overall structure similar to those of plant rhabdoviruses (see Supplementary Fig. S1 available in JGV Online). This protein has two major hydrophobic regions at both its N and C termini, which probably constitute the signal peptide and transmembrane domains of rhabdovirus G proteins. The protein also contains two potential glycosylation sites (Asn X Ser/Thr) located at positions 353 and 391. Multiple alignment of the ORF5 protein and G proteins of seven plant rhabdoviruses indicated that 11 cysteine residues are similarly positioned in the aligned sequences (data not shown), and these could potentially form structurally 2416 Journal of General Virology 87

5 Genome structure of orchid fleck virus important disulfide bridges (Walker & Kongsuwan, 1999). These results suggest that OFV ORF5 corresponds to the rhabdovirus G protein. ORF2 (26 kda), ORF3 (38 kda) and ORF4 (20 kda) proteins showed no apparent similarities to other viral sequences in the international gene sequence databases. Analysis of the protein encoded by the OFV RNA2 The deduced amino acid sequence of the ORF6 (212 kda) showed significant similarity to the polymerase (L) protein (RNA-dependent RNA polymerase) of nucleorhabdoviruses [MFSV (38 % identity), MMV (37 %), RYSV (36 %), TaVCV (36 %) and SYNV (35 %)] and cytorhabdoviruses [SCV (32 %), NCMV (31 %) and LNYV (30 %)]. Moreover, the ORF6 protein showed 28 % identity to the L protein of LBVaV, which belongs to the genus Varicosavirus (Sasaya et al., 2002). Lower similarity matches were also obtained with L proteins of the animal and fish rhabdoviruses (22 26 % identity) and other viruses within the families Bornaviridae, Filoviridae and Paramyxoviridae (less than 21 % identity), all of which have non-segmented, negativestranded RNA genomes. However, the OFV ORF6 protein showed no detectable similarity with the L proteins of viruses within the families Arenaviridae, Bunyaviridae and Orthomyxoviridae, all of which have segmented, negativestranded RNA genomes. The L proteins of viruses belonging to the order Mononegavirales have been found to share six conserved domains (I VI) (Poch et al., 1989, 1990). The OFV ORF6 L protein also has these domains, except domain VI (data not shown). The similarity of the L proteins of OFV and rhabdoviruses is particularly great in domain III (Fig. 4a). This domain contains four very distinct conserved motifs (motifs A, B, C and D) (Poch et al., 1990) that are thought to play an essential role for the RNA polymerase activity (Schnell & Conzelmann, 1995). Motif C contains the conserved tetrapeptide GDNQ (Gly Asp Asn Gln), which may have a similar function to the GDD motif in the polymerases of positive-strand RNA viruses (Poch et al., 1989, 1990). A phylogenetic tree calculated from the deduced amino acid sequences of the four motifs located within domain III (Fig. 4b) suggests that OFV is most closely related to nucleorhabdoviruses, especially SYNV and MFSV, and that LBVaV is more closely related to cytorhabdoviruses than to nucleorhabdoviruses (Fig. 4b) Fig. 4. (a) Alignments of the proposed catalytic domains of the L proteins of OFV and selected rhabdoviruses. The four motifs (A, B, C and D) within the catalytic domain (III) of the L proteins are indicated. Strictly invariant residues are shown in grey boxes. Numbers at the beginning of the lines indicate the position of the first displayed amino acid. Numbers within the angle brackets indicate the numbers of amino acids not represented in Fig. (b) Phylogenetic relationship of OFV and non-segmented, negativestranded RNA viruses based on amino acid sequences in the conserved region (block III) of the L protein. The tree was constructed using the neighbour-joining method (CLUSTAL W) and based on its alignment obtained using default parameters (data not shown). Values on the branches represent the percentage of trees containing each cluster of 1000 bootstrap replicates. Scale bar represents the number of amino acid replacements per site. The abbreviated names of virus species are described in Methods

6 H. Kondo and others Analysis of 3 - and 5 -untranslated region (UTR) sequences of the OFV genome The 39-UTR sequences of RNA1 and RNA2 were of 227 (58 % A+U content) and 182 nt (60 % A+U content), respectively. The 59-UTR sequences of RNA1 and RNA2 were 205 (61 %) and 185 nt (56 %), respectively. The 39-terminal sequences of OFV RNA1 and RNA2 (negative sense) shared the first 13 nt with a single mismatch, whereas the 59- terminal sequences shared the first 6 nt (Fig. 2). The 39- and 59-ends of both RNA1 and RNA2 had perfect complementarity for the first six and 10 residues, respectively (Fig. 5). The two RNAs also had the same sequences for the terminal six residues (39-UGUGUC----GACACA-59). Such complementarities are a common feature among plant rhabdoviruses (Choi et al., 1994; Revill et al., 2005; Tanno et al., 2000; Tsai et al., 2005; Wetzel et al., 1994) (Fig. 5). The 39- and 59-terminal sequences of OFV were identical to that of MFSV for the first 5 nt and to RYSV for the first 4 nt. The 39- and 59-terminal complementary sequences of cytorhabdoviruses LNYV and NCMV have overhanging nucleotides in the 39-terminal end; however, similar nucleotides were not found in the OFV genome and they are not found in nucleorhabdovirus genomes (Fig. 5). Analysis of UTRs of the OFV genome The untranslated sequences between the ORFs of OFV RNA1 were similar to one another both in length ( nt) and A+U content (61 69 %). A consensus sequence (39-UAA- AUUUAUUUUGUU/AG/AUU-59) was found in all these untranslated sequences (Fig. 6a). This sequence indicates that these are probably gene-junction sequences, which are a feature of other plant rhabdovirus genomes (Heaton et al., 1989; Huang et al., 2003; Luo & Fang, 1998; Reed et al., 2005; Revill et al., 2005; Tanno et al., 2000; Tsai et al., 2005; Wetzel et al., 1994). The consensus sequence of the OFV RNA1 was mostly identical to the MFSV genejunction sequences and was also similar to the gene-junction sequences of the other plant rhabdoviruses (Fig. 6b). Similar consensus sequences were also present at the 39- and 59-UTRs of the OFV genome (Fig. 6a). Both OFV RNA1 and RNA2 possess short UTRs at the 39- and 59-ends of the Fig. 5. Complementarity between the 39- and 59-ends of OFV and plant rhabdoviruses. The 39- and 59-end sequences of the genome are shown in viral RNA (39 59, negative) sense. Vertical lines between the sequences indicate complementary nucleotides. Sequences conserved between the complementary areas of the genome ends are in grey boxes. The underlined nucleotides correspond to overhang in the genome sequence. Fig. 6. (a) Alignment of the putative gene-junction sequences within the OFV RNA1 and RNA2. The regions between each ORF and non-coding regions at 39 and 59 termini of OFV RNAs are shown in the viral RNA sense (39 59, negative sense). Sequences conserved between the intergenic regions are within the grey boxes. First AUG codons of ORF3 and ORF5 (italic and bold letters) are upstream of, or overlapping, their putative gene-junction sequences. (b) Comparison of the gene-junction sequences of OFV and plant rhabdoviruses. Short non-transcribed sequences that separate each gene in intergenic gene-junction regions are indicated in italics and underlined. (n), Number of variable nucleotides Journal of General Virology 87

7 Genome structure of orchid fleck virus genome. In plant rhabdoviruses, the 39 and 59 extremities of the genome have similar UTRs, termed the 39-leader and 59-trailer regions. The putative 39-leader regions of OFV RNA1 and RNA2 are composed of 158 and 133 nt, respectively, while those of plant rhabdoviruses range from 84 (LNYV) to 203 nt (RYSV). The putative 59-trailer regions of OFV RNA1 and RNA2 are of 148 and 150 nt, respectively, while those of plant rhabdoviruses range from 64 (TaVCV) to 273 nt (NCMV). DISCUSSION We have sequenced the complete genome of OFV and showed that it is bipartite, negative sense. Comparative sequence analyses showed that the OFV N (ORF1), G (ORF5) and L (ORF6) proteins have significant sequence similarities to those of viruses belonging to the genus Nucleorhabdovirus in the family Rhabdoviridae. However, the deduced amino acid sequences of ORF2, ORF3 and ORF4 showed no apparent similarities to other known viral genes or sequences. The genomes of monopartite plant rhabdoviruses, such as SYNV and LNYV, consist of six genes in the negative sense in the order of N, P, sc4 or 4b (depending on the virus), M, G and L (Dietzgen et al., 2006; Jackson et al., 2005; Redinbaugh & Hogenhout, 2005) (Fig. 7). The genome of OFV has genes in the same order, but is split between the G protein and L polymerase genes, with the L gene present on RNA2 and the other five genes present on RNA1 (Fig. 2). In Northern hybridization tests, no larger size of RNA than RNA1 or RNA2 was detected in the total RNA from OFV-infected plants (data not shown). The bacilliform or bullet-shaped virions of OFV resemble those of plant rhabdoviruses, but they are smaller and lack an envelope (Fig. 1a). Their fine structure is similar to that of the internal component of plant rhabdoviruses. The fact that OFV virions are nm in length, namely about half the size of typical plant rhabdoviral virions, suggests that RNA1 and RNA2 are packaged in separate virions, because the length of the bullet-shaped rhabdovirions is known to be determined by the length of the RNA they contain (Jayakar et al., 2004; Schnell et al., 1996). The N protein (ORF1) of OFV may be the major protein forming the rhabdoviral nucleocapsid (Jackson et al., 2005). The ORF2 protein (gene 2) contains five candidate consensus sites for phosphorylation by casein kinase II (data not shown), which are known to be important for primary phosphorylation of the phosphoprotein (P) of Vesicular stomatitis virus (Barik & Banerjee, 1992), suggesting that the OFV ORF2 corresponds to the P protein. The ORF4 protein (gene 4) may be a counterpart of the rhabdovirus matrix protein (M), although the ORF4 protein had neither the consensus phosphorylation sites of casein kinase II nor the characteristic sequence motifs of rhabdovirus M proteins. The ORF3 protein (gene 3) may correspond to the third non-structural protein gene of plant rhabdoviruses, which has been proved to be involved in viral cell-to-cell movement in plants (Huang et al., 2005). The ORF5 protein shows similarities to G proteins of nucleorhabdoviruses, which forms the surface spikes of the mature membrane-bound rhabdovirions (Jackson et al., 2005). However, the enveloped virions were rarely seen in Fig. 7. Comparison of the genome organizations of OFV, SYNV, LNYV and LBVaV. The genomic locations in the negativesense arrangement are shown, also the relative sizes of the viral genes. Arrowheads indicate the gene-junction regions in the viral genome.

8 H. Kondo and others the infected plant cells (Chang et al., 1976), suggesting that the budding process of these virions may be incomplete. One explanation is that the OFV G protein probably has only two potential glycosylation sites, whereas nucleorhabdovirus G proteins contain five to 10 glycosylation sites, suggesting that fewer glycosylation sites or less glycosylation of this protein may attenuate the membrane-budding function of OFV. Another possible explanation is that the putative M protein (ORF4) of OFV may have lost an important role(s) in virus budding as it lacks a hydrophilic region of about aa near the C terminus of the M proteins of nucleorhabdoviruses (data not shown). Similarities between OFV and rhabdoviruses were also found in the conserved intergenic region sequences of RNA1. The sequences of gene junctions, which are broadly conserved among rhabdoviruses, contain the regulatory signals for termination/polyadenylation of upstream mrna transcription and initiation of transcription of the adjacent downstream mrna (Jackson et al., 2005; Redinbaugh & Hogenhout, 2005). The termination/polyadenylation sequences found in plant rhabdoviruses comprise an AUrich region and a poly(u) tract (Heaton et al., 1989; Huang et al., 2003; Luo & Fang, 1998; Tsai et al., 2005; Wetzel et al., 1994). The putative gene-junction sequences of OFV RNAs also contain a possible termination/polyadenylation sequence (39-UAAAUUUA/CUUUU-59) (Fig. 6b). This sequence was always followed by the semi-consensus sequence GUUG(G)UU, which may be involved in the initiation of transcription of neighbouring genes of OFV genome. The similarity at the putative gene-junction regions indicates that the regulatory functions in the genome of OFV and rhabdoviruses may be similar. Based on the similarities between OFV and nucleorhabdovirus in their genome structure, virion morphology, subcellular distribution patterns and formation of nuclear viroplasms, OFV can be regarded as a nucleorhabdovirus with a divided genome. Therefore, we propose that OFV be designated the type species of a new genus Dichorhabdovirus in the family Rhabdoviridae, from the Greek root dikho- means apart or asunder, or split into two, and here refers to the nucleorhabdovirus that is split in two. It is also noteworthy that LBVaV, the type member of the genus Varicosavirus, has a two-segmented, negativestranded RNA genome and its virions are not enveloped (Sasaya et al., 2001, 2002, 2004) (Fig. 7). In addition, phylogenetic analyses of the L polymerase genes show that LBVaV is most closely related to the plant rhabdoviruses (Sasaya et al., 2002), but sequence relationships between OFV and LBVaV are more distant than those between OFV and plant rhabdoviruses (Fig. 4b). Furthermore LBVaV is transmitted in soil by the zoospores of a fungus vector (Sasaya et al., 2001), whereas most plant rhabdoviruses are transmitted by aphids, leafhoppers or planthoppers, and several of them have been shown to replicate in those arthropod vectors (Jackson et al., 2005). OFV is transmitted by the false-spider mite B. californicus in a persistent manner (Kondo et al., 2003). CiLV and CoRSV, which resemble OFV in virion morphology and cytopathic effects, are also transmitted by Brevipalpus mites (Chagas et al., 2003; Kitajima et al., 2003; Rodrigues et al., 2003), but it is not known whether these mitetransmitted viruses have a bipartite genome. Several related plant viruses have positive-stranded RNA genomes that are monopartite or bipartite, but nonetheless have similar vectors. It seems that, among plant viruses, vector associations may survive evolutionary change better than genome structure. ACKNOWLEDGEMENTS We thank Koji Mitsuhata for technical assistance; Drs Narinobu Inouye and Salah Bouzoubaa for helpful comments; Dr Nobihiro Suzuki for critically reading the manuscript; Dr Adrian Gibbs for critically reading the manuscript and for his suggestion of the name Dichorhabdovirus. This work was supported by Grant-in-Aid for Encouragement of Young Scientists from the Ministry of Education, Culture, Sports, Science and Technology of Japan. REFERENCES Barik, S. & Banerjee, A. K. (1992). Phosphorylation by cellular casein kinase II is essential for transcriptional activity of vesicular stomatitis virus phosphoprotein P. Proc Natl Acad Sci U S A 89, Begtrup, J. (1972). Structure of a bacilliform virus in Dendrobium as revealed by negative staining. Phytopathol Z 75, Blanchfield, A. L., Mackenzie, A. M., Gibbs, A., Kondo, H., Tamada, T. & Wilson, C. R. (2001). Identification of Orchid fleck virus by reverse transcriptase-polymerase chain reaction and analysis of isolate relationships. J Phytopathol 149, Calisher, C. H., Carstens, E. B., Christian, P., Mahy, B. W. J., Mayo, M. A. & Shope, R. E. (2000). Unassigned viruses. In Virus Taxonomy, Seventh Report of the International Committee on Taxonomy of Viruses, pp Edited by M. H. V. van Regenmortel, C. M. Fauquet, D. H. L. Bishop, E. B. Carstens, M. K. Estes, S. M. Lemon, J. Maniloff, M. A. Mayo, D. J. McGeoch, C. R. Pringle & R. B. Wickner. San Diego: Academic Press. Chagas, C. M., Kitajima, E. W. & Rodrigues, J. C. V. (2003). Coffee ringspot virus vectored by Brevipalpus phoenicis (Acari: Tenuipalpidae) in coffee. Exp Appl Acarol 30, Chang, M. U., Arai, K., Doi, Y. & Yora, K. (1976). Morphology and intracellular appearance of orchid fleck virus. Ann Phytopathol Soc Jpn 42, Choi, T. J., Wagner, J. D. & Jackson, A. O. (1994). Sequence analysis of the trailer region of sonchus yellow net virus genomic RNA. Virology 202, Dietzgen, R. G., Callaghan, B., Wetzel, T. & Dale, J. L. (2006). Completion of the genome sequence of Lettuce necrotic yellows virus, type species of the genus Cytorhabdovirus. Virus Res 118, Doi, Y., Chang, M. U. & Yora, K. (1977). Orchid fleck virus. CMI/AAB Descriptions of Plant Viruses, no Gibbs, A., Mackenzie, A., Blanchfield, A., Cross, P., Wilson, C., Kitajima, E. W., Nightingale, M. & Clements, M. (2000). Viruses of orchids in Australia; their identification, biology and control. Aust Orchid Rev 65, Journal of General Virology 87

9 Genome structure of orchid fleck virus Heaton, L. A., Hillman, B. I., Hunter, B. G., Zuidema, D. & Jackson, A. O. (1989). Physical map of the genome of sonchus yellow net virus, a plant rhabdovirus with six genes and conserved gene junction sequences. Proc Natl Acad Sci U S A 86, Huang, Y., Zhao, H., Luo, Z., Chen, X. & Fang, R. X. (2003). Novel structure of the genome of Rice yellow stunt virus: identification of the gene 6-encoded virion protein. J Gen Virol 84, Huang, Y. W., Geng, Y. F., Ying, X. B., Chen, X. Y. & Fang, R. X. (2005). Identification of a movement protein of rice yellow stunt rhabdovirus. J Virol 79, Jackson, A. O., Dietzgen, R. G., Goodin, M. M., Bragg, J. N. & Deng, M. (2005). Biology of plant rhabdoviruses. Annu Rev Phytopathol 43, Jayakar, H. R., Jeetendra, E. & Whitt, M. A. (2004). Rhabdovirus assembly and budding. Virus Res 106, Kitajima, E. W., Blumenschein, A. & Costa, A. S. (1974). Rodlike particles associated with ringspot symptoms in several orchid species in Brazil. Phytopathol Z 81, Kitajima, E. W., Kondo, H., Mackenzie, A., Rezende, J. A. M., Gioria, R., Gibbs, A. & Tamada, T. (2001). Comparative cytopathology and immunocytochemistry of Japanese, Australian and Brazilian isolates of orchid fleck virus. J Gen Plant Pathol 67, Kitajima, E. W., Chagas, C. M. & Rodrigues, J. C. V. (2003). Brevipalpus-transmitted plant virus and virus-like diseases: cytopathology and some recent cases. Exp Appl Acarol 30, Kondo, H., Matsumoto, J., Maeda, T. & Inouye, N. (1995). Host range and some properties of orchid fleck virus isolated from oriental Cymbidium in Japan. Bull Res Inst Bioresour Okayama Univ 3, (in Japanese with English abstract). Kondo, H., Maeda, T. & Tamada, T. (2003). Orchid fleck virus: Brevipalpus californicus mite transmission, biological properties and genome structure. Exp Appl Acarol 30, Lesemann, D. & Doraiswamy, S. (1975). Bullet-shaped virus-like particles in chlorotic and necrotic leaf lesions of orchids. Phytopathol Z 83, Luo, Z. L. & Fang, R. X. (1998). Structure analysis of the rice yellow stunt rhabdovirus glycoprotein gene and its mrna. Arch Virol 143, Miranda, G. J., Azzam, O. & Shirako, Y. (2000). Comparison of nucleotide sequences between northern and southern Philippine isolates of rice grassy stunt virus indicates occurrence of natural genetic reassortment. Virology 266, Page, R. D. M. (1996). TREEVIEW: an application to display phylogenetic trees on personal computers. Comput Appl Biosci 12, Poch, O., Sauvaget, L., Delarue, M. & Tordo, N. (1989). Identification of four conserved motifs among the RNA-dependent polymerase encoding elements. EMBO J 8, Poch, O., Blumberg, B. M., Bougueleret, L. & Tordo, N. (1990). Sequence comparison of five polymerases (L proteins) of unsegmented negative-stranded RNA viruses: theoretical assignment of functional domain. J Gen Virol 71, Redinbaugh, M. G. & Hogenhout, S. A. (2005). Plant rhabdoviruses. Curr Top Microbiol Immunol 292, Reed, S. E., Tsai, C. W., Willie, K. J., Redinbaugh, M. G. & Hogenhout, S. A. (2005). Shotgun sequencing of the negative-sense RNA genome of the rhabdovirus Maize mosaic virus. J Virol Methods 129, Revill, P., Trinh, X., Dale, J. & Harding, R. (2005). Taro vein chlorosis virus: characterization and variability of a new nucleorhabdovirus. J Gen Virol 86, Rodrigues, J. C. V., Kitajima, E. W., Childers, C. C. & Chagas, C. M. (2003). Citrus leprosis virus vectored by Brevipalpus phoenicis (Acari: Tenuipalpidae) on citrus in Brazil. Exp Appl Acarol 30, Sasaya, T., Ishikawa, K. & Koganezawa, H. (2001). Nucleotide sequence of the coat protein gene of Lettuce big-vein virus. J Gen Virol 82, Sasaya, T., Ishikawa, K. & Koganezawa, H. (2002). The nucleotide sequence of RNA1 of Lettuce big-vein virus, genus Varicosavirus, reveals its relation to nonsegmented negative-strand RNA viruses. Virology 297, Sasaya, T., Kusaba, S., Ishikawa, K. & Koganezawa, H. (2004). Nucleotide sequence of RNA2 of Lettuce big-vein virus and evidence for a possible transcription termination/initiation strategy similar to that of rhabdoviruses. J Gen Virol 85, Schnell, M. J. & Conzelmann, K. K. (1995). Polymerase activity of in vitro mutated rabies virus L protein. Virology 214, Schnell, M. J., Buonocore, L., Kretzschmar, E., Johnson, E. & Rose, J. K. (1996). Foreign glycoproteins expressed from recombinant vesicular stomatitis viruses are incorporated efficiently into virus particles. Proc Natl Acad Sci U S A 93, Tanno, F., Nakatsu, A., Toriyama, S. & Kojima, M. (2000). Complete nucleotide sequence of Northern cereal mosaic virus and its genome organization. Arch Virol 145, Thompson, J. D., Higgins, D. G. & Gibson, T. J. (1994). CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res 22, Tsai, C. W., Redinbaugh, M. G., Willie, K. J., Reed, S., Goodin, M. & Hogenhout, S. A. (2005). Complete genome sequence and in planta subcellular localization of maize fine streak virus proteins. J Virol 79, Walker, P. J. & Kongsuwan, K. (1999). Deduced structural model for animal rhabdovirus glycoproteins. J Gen Virol 80, Wetzel, T., Dietzgen, R. G. & Dale, J. L. (1994). Genomic organization of lettuce necrotic yellows rhabdovirus. Virology 200, Wunner, W. H., Calisher, C. H., Dietzgen, R. G. & 8 other authors (1995). Family Rhabdoviridae. InVirus Taxonomy, Sixth Report of the International Committee on Taxonomy of Viruses, pp Edited by F. A. Murphy, C. M. Fauquet, D. H. L. Bishop, S. A. Ghabrial, A. W. Jarvis, G. P. Martelli, M. A. Mayo & M. D. Summers. Vienna: Springer. Ziemiecki, A. & Peters, D. (1976). The proteins of sowthistle yellow vein virus: characterization and location. J Gen Virol 32,