|Home | About | Journals | Submit | Contact Us | Français|
White root rot, caused by the ascomycete Rosellinia necatrix, is a devastating disease worldwide, particularly in fruit trees in Japan. Here we report on the biological and molecular properties of a novel bipartite double-stranded RNA (dsRNA) virus encompassing dsRNA-1 (8,931 bp) and dsRNA-2 (7,180 bp), which was isolated from a field strain of R. necatrix, W779. Besides the strictly conserved 5′ (24 nt) and 3′ (8 nt) terminal sequences, both segments show high levels of sequence similarity in the long 5′ untranslated region of approximately 1.6 kbp. dsRNA-1 and -2 each possess two open reading frames (ORFs) named ORF1 to -4. Although the protein encoded by 3′-proximal ORF2 on dsRNA-1 shows sequence identities of 22 to 32% with RNA-dependent RNA polymerases from members of the families Totiviridae and Chrysoviridae, the remaining three virus-encoded proteins lack sequence similarities with any reported mycovirus proteins. Phylogenetic analysis showed that the W779 virus belongs to a separate clade distinct from those of other known mycoviruses. Purified virions ~50 nm in diameter consisted of dsRNA-1 and -2 and a single major capsid protein of 135 kDa, which was shown by peptide mass fingerprinting to be encoded by dsRNA-1 ORF1. We developed a transfection protocol using purified virions to show that the virus was responsible for reduction of virulence and mycelial growth in several host strains. These combined results indicate that the W779 virus is a novel bipartite dsRNA virus with potential for biological control (virocontrol), named Rosellinia necatrix megabirnavirus 1 (RnMBV1), that possibly belongs to a new virus family.
Viruses are found ubiquitously in major groups of filamentous fungi (40), and an increasing number of novel mycoviruses are being reported (3, 36). Mycoviruses with RNA genomes are now classified into 10 families, of which four accommodate double-stranded RNA (dsRNA) viruses and the remaining six comprise single-stranded RNA (ssRNA) viruses (23). While many ssRNA mycoviruses, like hypoviruses and endornaviruses, do not produce particles, dsRNA virus genomes, whether undivided (the family Totiviridae) or divided (11 or 12 segments for the family Reoviridae, 4 segments for the family Chrysoviridae, and 2 segments for the family Partitiviridae), are encapsidated in rigid particles. Most mycoviruses are considered to cause cryptic infections, while some cause phenotypic alterations that include hypovirulence and debilitation. However, the lack of artificial introduction methods for most mycoviruses has greatly hampered progress in exploring mycovirus-host interactions (23, 40). Thus, a virus etiology of altered fungal phenotypes was established only for a limited number of examples, including hypovirus-C. parasitica and mycoreovirus-C. parasitica.
White root rot is one of the most devastating diseases of perennial crops worldwide, particularly highly valued fruits in Japan like apple, Japanese pear, and grapevine. The causal fungus, Rosellinia necatrix, is an ascomycete with a wide range of host plants of >197 species spanning 50 families (31) and is difficult to control by conventional methods, as is often the case for soilborne pathogens. Fungicide application, though it may be effective, is labor-intensive and raises environmental concerns, while cultural practices may not be effective. Successful biocontrol of chestnut blight disease in Europe with hypovirulent strains (25, 38) inspired a group of Japanese researchers to conduct an extensive search of a large collection of >1,000 field fungal isolates for mycoviruses that might serve as virocontrol agents. Virocontrol or virological control refers to one form of biological control utilizing viruses that infect organisms pathogenic to useful organisms (23). Approximately 20% of the collected isolates of R. necatrix were found to be dsRNA positive and presumed to be infected by mycoviruses (4, 29). Agarose gel profiles of dsRNAs suggested infections by members in the families Totiviridae, Partitiviridae, Reoviridae, and Chrysoviridae, as well as unassigned viruses (S. Kanematsu and A. Sasaki, unpublished results). Among those dsRNAs, the genomic segments of Mycoreovirus 3 (MyRV3) (55) and Rosellinia necatrix partitivirus 1 (RnPV1) (44) were well characterized. However, many other dsRNAs remain uncharacterized.
Artificial virion introduction protocols, which are often unavailable for mycoviruses, have been developed for specific viruses infecting the white root rot fungus. Using a polyethylene glycol (PEG)-mediated method, as established for MyRV1 and MyRV2 infecting C. parasitaca (27, 28), RnPV1 and MyRV3 were shown to be infectious as particles (45, 46). Subsequently, the cause-effect relationship was established: MyRV3 was demonstrated to confer hypovirulence (attenuated virulence) on an isogenic strain and a few vegetatively incompatible virulent strains of R. necatrix (33, 45), while RnPV1 was shown to be associated with symptomless infection. Protoplast fusion is also available for introduction of partitiviruses and uncharacterized viruses into recipient fungal strains that are vegetatively incompatible with virus-containing ones (A. Sasaki, unpublished results). Furthermore, DNA transformation systems are available for foreign gene expression in R. necatrix (33, 42). These technical advances have made the R. necatrix-mycovirus systems attractive for studies of virus-host interactions and virocontrol (23, 37).
R. necatrix strain W779 was isolated by Ikeda et al. (29, 30) from soil in Ibaraki Prefecture as a dsRNA-positive strain that had yet to be characterized. Here we describe the purification and biological and molecular properties of a novel virus isolated from W779. Particles ~50 nm in diameter isolated from strain W779 consist of two dsRNA elements termed dsRNA-1 and -2 of approximately 9 and 7 kbp and a major protein of 135 kDa encoded by one of two open reading frames (ORFs) on dsRNA-1. Importantly, purified virus particles were shown to be infectious and confer hypovirulence on vegetatively incompatible fungal strains. The two dsRNA segments share the conserved terminal sequences at both ends, and both possess extremely long (>1.6 kb) 5′ untranslated regions (UTRs) similar to each other, two ORFs, and relatively short 3′ UTRs. The 3′-proximal ORF of dsRNA-1 encodes an RNA-dependent RNA polymerase (RdRp) showing low levels (22 to 32%) of sequence identity to those of members of the families Totiviridae and Chrysoviridae. A phylogenetic analysis with RdRp sequences revealed that the W779 virus is placed into a separate clade from the recognized virus families. These attributes indicate that dsRNA-1 and -2 represent the genome segments of a novel bipartite virus, designated Rosellinia necatrix megabirnavirus 1 (RnMBV1), with virolocontrol agent potential. We propose the establishment of a new family, Megabirnaviridae, to accommodate RnMBV1 as the type species.
R. necatrix strain W779 was isolated by the baiting method with mulberry twigs from the soil of a Japanese pear orchard infested with white root rot in Ibaraki prefecture (29). Although asexual sporulation is often used to obtain virus-free isolates of filamentous fungi, asexual spores of R. necatrix are known to germinate but never develop into mycelia under laboratory conditions (39). Thus, an isogenic virus (dsRNA)-free strain, W1015, was prepared from spheroplasts of W779 and used as a reference strain. Two virus-free fungal strains, W97 and W370T1, were previously described (33, 45) and used in transfection with W779 virus particles. Strains W779, W97, and W370T1 are vegetatively incompatible with each other and belong to different mycelial compatibility groups (MCGs), MCG351, MCG80, and MCG139, respectively (4, 30). All strains were cultured at 25°C on Difco potato dextrose agar (PDA; Becton Dickinson, Sparks, MD) or in Difco potato dextrose broth (PDB) in the dark and kept at 5°C until used.
dsRNA was extracted as described by Sun and Suzuki (50). W779 was cultured for 10 days in PDB at room temperature. Mycelia were homogenized in liquid nitrogen. Nucleic acid fractions were obtained by treatments with phenol, phenol-chloroform, and chloroform. This was followed by digestion with S1 nuclease and subsequently with DNase I as described by Suzuki et al. (52). dsRNA was further purified by CC41 cellulose column chromatography. A cDNA library of total dsRNA was constructed by using a classic non-PCR-based method to minimize misincorporations during cDNA synthesis (28). After denaturation at 65°C in 90% dimethyl sulfoxide (DMSO) (5), dsRNA was used as a template for cDNA synthesis with random hexamers and a TimeSaver cDNA synthesis kit (Amersham). The resulting cDNA was cloned into the vector pGEM-T Easy (Promega) following addition of an A by Taq polymerase and used for transformation of Escherichia coli strain DH5α.
Two methods were used to determine the terminal sequences of the genome segments. A classic 5′-RACE (rapid amplification of cDNA ends) protocol using terminal deoxynucleotidyl transferase, was performed as described by Suzuki et al. (53). Approximately 20 ng of purified dsRNA, along with specific primers denatured by the DMSO method, was reverse transcribed in a reaction mixture containing 50 mM Tris-HCl (pH 8.3), 50 mM KCl, 4 mM MgCl2, 10 mM dithiothreitol, 40 U of Moloney murine leukemia virus reverse transcriptase (Fermentas), and 20 U of RNase inhibitor (Toyobo, Osaka, Japan). The primers used for the respective dsRNA segments were SC1 and SC2 for dsRNA-1 and SC3 and SC4 for dsRNA-2 (for the primer sequences, see Table S1 at http://www.rib.okayama-u.ac.jp/pmi/2003/html/JV09suppl.html). After d(C) tailing, cDNA products were used as the template with the 5′ abridged anchor primer and a nested specific primer.
For RNA ligase-mediated RACE (RLM-RACE) (53), a 5′-phosphorylated oligodeoxynucleotide, 3RACE-adaptor (5′-PO4-CAATACCTTCTGACCATGCAGTGACAGTCAGCATG-3′), was ligated to each of the 3′ termini of the dsRNA-1 and -2 segments (single-stranded form) with T4 RNA ligase at 16°C for 16 h. Ligated DNA-RNA strands were denatured in 90% DMSO together with the oligonucleotide 3RACE-1st (see Table S1 at the URL mentioned above) as described above and used as templates for cDNA synthesis. 3RACE-1st is complementary to the 3′ half of 3RACE-adaptor. The resulting cDNA was amplified by PCR with the primer set 3RACE-2nd (see Table S1 at the URL mentioned above), which is complementary to the 5′ half of 3RACE-adaptor, and gene-specific primers. Four PCRs were set for the plus- and minus-sense strands of the two segments. Amplified DNA fragments were cloned into pGEM-T Easy for sequence analysis.
Plasmid DNA was prepared with spin columns (LaboPass Mini; Hokkaido System Science, Sapporo, Japan) and sent to Macrogen Japan, Inc. (Tokyo), for sequencing. Sequences were analyzed with the Genetyx DNA-processing software (SDC, Tokyo, Japan). Database searches were performed with the BLAST program (1) at http://www.ncbi.nlm.nih.gov/ provided by the National Center for Biotechnology Information and with the FASTA 3 program (41) at http://www.ddbj.nig.ac.jp/Welcome-j.html from the DNA Data Bank of Japan (DDBJ). Motif searches were carried out against the Pfam database at http://pfam.sanger.ac.uk/ (20) and the PROSITE database at http://www.expasy.ch/. Multiple sequence alignments and phylogenetic trees were constructed with the CLUSTAL_X program (54). RNA structures and motifs were analyzed at http://www.bioinfo.rpi.edu/applications/mfold/ and http://wilab.inha.ac.kr/fsfinder2/ with the Mfold (version 3.2) (56) and FSFinder (8) programs, respectively.
Total nucleic acid preparations were obtained by the method of Suzuki and Nuss (51). Northern blot analysis of total RNA isolated as described above was done as described by Suzuki et al. (52). RNA was separated by electrophoresis in 1.0% agarose gels under denaturing conditions and capillary transferred onto Hybond-N+ nylon membrane (Amersham Biosciences, Buckingham, England). The probes used were digoxigenin (DIG)-11-dUTP-labeled DNA fragments amplified by PCR as recommended by the manufacturer (Roche Diagnostics, Mannheim, Germany). Chemiluminescent signals of probe-RNA hybrids were detected with a DIG detection kit and a CDP star kit (Roche).
Fungal strain W779 was grown in PDB or cellophane-PDA for 10 days. Mycelia (approximately 30 g, wet weight, for PDB cultures and 10 g for cellophane-PDA cultures) were harvested and ground to powder in the presence of liquid nitrogen. The homogenates were mixed with 150 ml of 0.1 M sodium phosphate, pH 7.0, with a Waring blender and clarified twice with CCl4 for PDB culture or 13% Vertrel XF (Du Pont-Mitsui Fluorochemicals Co., Tokyo, Japan) for PDA-cellophane cultures. NaCl and PEG 6000 were added to final concentrations of 1 and 8%, respectively, for PDB culture. After being stirred for 3 h at 4°C, the suspension was centrifuged at 16,000 × g for 20 min. Resultant pellets, suspended in 10 ml of 0.05 M sodium phosphate buffer, pH 7.0, were centrifuged at 7,000 × g for 20 min. Centrifugation at 119,000 × g for 2 h was substituted for precipitation by PEG-NaCl when cellophane-PDA cultures were used. The supernatant was recentrifuged through a 20% sucrose cushion (3 ml) in a Beckman SW41 Ti rotor at 80,000 × g for 2 h. The pellet was suspended in 1 ml 0.05 M Na phosphate. After centrifugation at 7,000 rpm for 5 min, the supernatant was fractionated through a 10 to 40% sucrose gradient by centrifugation at 70,000 × g for 2 h. Fractions in the middle portion were subjected to measurement of UV absorbance and ultracentrifugation. Recovered virus particles were resuspended in 100 μl of 0.05 M sodium phosphate buffer, pH 7.0.
Protein preparation was done according to the protocol recommended by the manufacturer (Bruker Daltonics) based on the method of Charoenpanich et al. (10). Purified virus preparations were electrophoresed in a preparative sodium dodecyl sulfate (SDS)-polyacrylamide (10%) gel. The major coat protein band of 135 kDa, stained by Coomassie brilliant blue, was excised, decolorized, and treated with dithiothreitol to break disulfide bonds, followed by carboxymethylation of cysteine residues with iodoacetamide. Modified protein was subjected to in-gel digestion with trypsin (sequencing grade modified trypsin; Promega), elution in acetonitrile-trifluoroacetic acid solution, and desalting by Zip-Tips (Cleanup C18 pipette tips; Millipore). After being redissolved in acetonitrile-trifluoroacetic acid solution, cleaved peptides were analyzed by matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometry (MS) and MALDI-TOF tandem MS (MS/MS) on an UltraFLEX mass spectrometer (BRUKER Daltonics) in which α-cyano-4-hydroxycinnamic acid was spotted as a matrix on an α-cyano-4-hydroxycinnamic acid AnchorChip (Bruker Daltonics) target plate along with protein preparations. MS/MS spectra were obtained from some selected tryptic peptides in a reflector mode. Peptide masses were compared to those deduced from sequences of ORFs on dsRNA-1 and -2 with the MASCOT program (Matrix Science).
Transfection is defined as inoculation of protoplasts with virus virions (28). Protoplasts derived from two virus-free strains of R. necatrix, W97 and W370T1, were prepared according to the method of Kanematsu et al. (33). These two trains are vegetatively incompatible with strain W779. Purified virus particles (from PDA-cellophane cultures) were passaged through Ultrafree-MC sterile centrifugal filter units (Millipore, Tokyo, Japan) and introduced into protoplasts by using PEG as described by Sasaki et al. (45, 46).
Virulence was assessed by the method of Kanematsu et al. (33), with minor modifications. R. necatrix strains were cultured on PDA at 25°C in the dark for 2 weeks. Autoclaved apple twig segments were added to each plate, and the plates were incubated for an additional 2 weeks. Five apple nursery plants (Malus prunifolia var. ringo), frequently used as rootstocks in apple cultivation, were grown in a single plastic container (20 by 65 by 18 cm) containing autoclaved orchard soil and pearlite for 43 days. Two segments of the incubated twigs covered with freshly grown mycelia were placed as inocula in the soil, approximately 7 cm below the soil surface, in contact with the taproots of apple plants. Inoculated containers were kept in a greenhouse at approximately 25°C.
The complete nucleotide sequences of dsRNA-1 and -2 reported here have been deposited with the EMBL/GenBank/DDBJ databases under accession no. AB512282 and AB512283.
A previously conducted screen of field isolates revealed the presence in strain W779 of two differently sized dsRNAs, termed dsRNA-1 and dsRNA-2, of approximately 9 and 7 kbp, respectively (see Fig. S1A at the URL mentioned above). The present study showed that virus particles (see below) and the two dsRNAs occurred concomitantly. However, no such particles or dsRNAs were found in isogenic strain W1015, which was derived from spheroplasts of W779 and which was cured of the virus (dsRNA) (see Fig. S1A at the URL mentioned above). The two segments are thought to make up the bipartite genome of a novel virus (see below).
The two strains, W779 and W1015, are distinguishable in colony morphology and virulence. W779 showed a much smaller colony size on PDA than did W1015 (Fig. (Fig.1A).1A). Comparison of vegetative mycelial growth on the surface (Fig. (Fig.1B)1B) and inside of the bark of apple rootstocks (data not shown) clearly indicated a higher level of mycelial growth in virus-free W1015 than in W779. That is, the whitish mycelia (shown by arrows) of strain W1015 grew and expanded from the site of inoculation much faster than virus-containing W779 on apple rootstocks buried in the soil. No apparent mycelial growth of strain W779, however, was observed in the 2 weeks postinoculation (Fig. (Fig.1B1B).
To further confirm the effects of the virus in the W779 host background, W1015 hyphae were fused as a recipient with W779 to transfer dsRNA-1 and -2 back to W1015. Interestingly, the phenotype of W1015, upon receipt of the virus, was converted to that of W779 (data not shown). This suggests that the virus comprising dsRNA-1 and -2 is likely responsible for the phenotypic differences between the two strains.
Particles could be purified by a conventional method entailing differential centrifugation and subsequent sucrose density gradient centrifugation. Twenty fractions (fractions 1 to 20, from top to bottom) were obtained from a sucrose gradient. A profile of absorbance at 260 nm showed a single peak at fraction 14. Electron micrographs of negatively stained virus particles purified from the peak fraction showed spherical particles ~50 nm in diameter with a rough surface (Fig. (Fig.2A).2A). The estimated size of the particles was slightly greater than that of those reported for members of the families Partitiviridae (30 to 40 nm), Chrysoviridae (30 to 40 nm), and Totiviridae (30 to 40 nm) and smaller than that of those of members of the genus Mycoreovirus (approximately 80 nm) that are known to infect fungi. When compared, on the same grid, to particles (approximately 40 nm) of the victorivirus HvV190S (provided by S. A. Ghabrial), the W779 particles were slightly larger (data not shown).
Extraction of nucleic acids by the SDS-phenol method provided an RNA profile indistinguishable from that of nucleic acids prepared from mycelia of the same W779 culture as was used for particle preparations (Fig. (Fig.2B).2B). dsRNA-1 and -2 were resistant to DNase I and S1 nuclease (see Fig. S1A at the URL mentioned above). The larger dsRNA-1 segment migrated slightly faster than the dsRNA replicative forms of CHV3-GH2 (9.8 kbp; reference 48) or CHV1-EP713 mutant Δp69b, which lacks 96.1% of the p69 coding domain (10.9 kbp; reference 51) (see Fig. S1B at the URL mentioned above). Together with the resistance of the segments to S1 nuclease and DNase I (see Fig. S1 at the URL mentioned above), their ability to bind CC41 cellulose strongly suggested their dsRNA nature. This was further confirmed by their resistance to RNase A at a high salt concentration (0.3 M NaCl) and susceptibility in water (data not shown). When examined by band intensity on ethidium bromide-stained agarose gels, interestingly, the molar ratio of dsRNA-1 to dsRNA-2 varied, depending on the cultures (compare Fig. S1 at the URL mentioned above with Fig. Fig.2B).2B). Although dsRNA-1 is generally less abundant than dsRNA-2, the dsRNA profiles from purified particles and mycelia from single cultures are consistently similar (Fig. (Fig.2B,2B, compare lanes mycelia and VP).
SDS-polyacrylamide gel electrophoresis analysis of purified virions showed a single major band corresponding to 135 kDa and a minor protein migrating slightly slower than the size standard of 250 kDa (shown by black and gray arrowheads, Fig. Fig.2C,2C, lane VP). No major protein was detected in preparations obtained by the same procedure as for particle purification from virus-free mycelia of W1015 (Fig. (Fig.2C,2C, lane VF). Only a few minor protein bands were occasionally detected. The 135-kDa protein possibly represents the major capsid protein. We performed PMF analysis to determine the sequence of peptides derived by digestion of the 135-kDa protein with trypsin. The analysis provided peptide sequences of a total of 24 peptide fragments, which perfectly matched those from the deduced amino acid sequence coded for by one of the dsRNA-1 ORFs, as discussed in detail below (see Table S2 at the URL mentioned above).
The complete sequence of each segment was obtained by sequencing recombinant plasmid clones derived from the cDNA library, reverse transcription (RT)-PCR, 5′-RACE, and RLM-RACE. For the sequencing strategy used, see Fig. S2 at the URL mentioned above. From the cDNA library prepared from a mixture of dsRNA-1 and -2, cDNA plasmid clones with relatively large inserts of >700 bp were randomly chosen for sequence assembly. Assembly of sequences of the cDNA clones and gap-filling RT-PCR clones resulted in the generation of two contigs, 1 and 2. More than 75 cDNA clones were analyzed, and yet no clones were found to constitute an additional contig. This suggested that no dsRNAs other than dsRNA-1 and -2 were present in strain W779 that might comigrate with them in gel. At least two cDNA clones were sequenced in both directions for a single site (see Fig. S2 at the URL mentioned above). Contigs 1 and 2 were 8,894 and 7,135 nucleotides (nt) long, respectively, and were considered to be from dsRNA-1 and -2 and cover more than 95% of the dsRNA-1 and dsRNA-2 sequences based on the agarose gel electrophoresis profile of the dsRNAs of MyRV1 (53), CHV3-GH2 (48), and the Δp69b mutant of CHV1 (51) as size markers (see Fig. S1B at the URL mentioned above). Interestingly, the 5′-terminal regions of the coding strand from both contigs showed high levels of sequence similarity (approximately 70% identity), with gaps of 9 to 27 nt (see below). Many identical sequence stretches of >10 nt were found between the 5′ UTRs of the two segments (see Fig. S3 at the URL mentioned above). Sequence similarities could be attributed to template switches during cDNA synthesis and/or errors during sequence assembly. To eliminate this possibility and assign the contigs experimentally, we conducted Northern blotting in which two cDNA clones representing contigs 1 (probe 1) and 2 (probe 3) and one cDNA clone of dsRNA-1 spanning a region similar in sequence to the corresponding portion of dsRNA-2 (probe 2) were used as probes. As shown in Fig. Fig.3A,3A, probes 1 and 3 detected single bands of dsRNA-1 and -2 on a Northern blot, respectively. However, probe 2 hybridized to both segments (Fig. (Fig.3A),3A), confirming the sequence similarities between them. The relative positions of these probes are shown in Fig. Fig.3B.3B. When cDNA fragments corresponding to regions of unique sequences of the 5′ UTRs of dsRNA-1 (map positions 1402 to 1639) and dsRNA-2 (map positions 60 to 253) were used as probes, the respective segments were detected specifically (data not shown), providing further evidence of the correctness of the sequences.
Terminal sequences of the two segments were determined by a combination of classic 5′-RACE and 3′-RLM-RACE. Six clones gained by 5′-RACE were sequenced for each 5′ end of both strands of dsRNA-1 and -2. Most of the clones (four out of six), obtained by d(C) tailing of cDNA to the 5′ terminus of the plus-sense strand, had the 5′-terminal sequence 5′-(G)nGCATAAAAA, which was shared by dsRNA-1 and -2. The 5′ sequence of the minus-sense strands was 5′-(G)nGCGAAAAA, which is common to the two segments. In order to determine the extreme terminal nucleotide sequence, a total of four 3′-RLM-RACE reactions were designed for the two strands each of dsRNA-1 and -2. Fragments were amplified whose sizes agreed with the ones presumed from the positions of the primers and the expected lengths of regions uncovered by the contigs (see Fig. S4 at the URL mentioned above). The resultant six 3′-RLM-RACE clones derived from a single RLM-RACE reaction were identical in sequence. The terminal sequences of dsRNA-1 and -2 were determined by RLM-RACE to be 5′-GCA and CGC-3′ for the plus strand. As shown in Fig. Fig.4A,4A, the 5′ 24-mer and 3′ octamer are strictly conserved among RACE clones of dsRNA-1 and -2. Except for the conserved terminal heptamer, no conserved sequence stretches were detected in the 3′-terminal regions (Fig. (Fig.4A).4A). In addition to three potential tandem stem-loop structures commonly found within the 3′-terminal 90-nt sequences of dsRNA-1 and -2 by Mfold at default settings (Fig. (Fig.4B),4B), a 9-bp inverted repeat could be formed by the 5′- and 3′-terminal sequences of either strand of each segment (Fig. (Fig.4C4C).
The complete nucleotide sequences of dsRNA-1 and -2 were deposited in the GenBank/EMBL/DDBJ databases. The schematic genetic organization of the coding strands of dsRNA-1 and -2 is shown in Fig. Fig.3B.3B. dsRNA-1 and -2 are 8,931 and 7,180 nt in length, each possessing 5′ UTRs of >1.6 kb, two tandem nonoverlapping ORFs, and relatively short 3′ UTRs. dsRNA-1 ORF1 and -2 could encode polypeptides of 1,240 aa (P1) and 1,111 aa (P2), respectively, while dsRNA-2 ORF3 and -4 encode polypeptides of 1,426 aa (P3) and 227 aa (P4) (Fig. (Fig.3B).3B). P1 was unequivocally shown by PMF analysis to be the major capsid protein of 135 kDa (see Table S2 at the URL mentioned above). That is, the MALDI-TOF MS analysis allowed the identification of 24 specific tryptic peptides with 38.8% coverage of the 135 kDa and a highly significant MOWSE score of 262 in the Mascot program. Examples included FHNQLSGVYER (amino acid positions 130 to 140), QPLLDEAFGAGVVQPGNMDLVGAGIDFTR (amino acid positions 505 to 533), DLGDFTAGTVDAAASGYEWDNYVYR (amino acid positions 557 to 581), MLSLTFGLAGQLR (amino acid positions 711 to 723), and AAHTWDTFVQNPR (amino acid positions 1078 to 1090). No peptide sequences matched those coded for by other ORFs of dsRNA-1 and -2.
Sequence heterogeneities among cDNA clones were noted. Two transitions were detected at position 8358 of dsRNA-1 (T to C) and position 3702 of dsRNA-2 (C to T) that were silent, while two transversions were found at dsRNA-2 positions 2183 (C to A) and 5275 (A to T) that would cause amino acid changes on the ORF 3-encoded protein (P3) from Pro176 to His and Ser1207 to Cys, respectively. Moreover, an interesting 40-nt deletion was detected at the 5′ UTR of dsRNA-2 that extended from position 172 to position 211. The deletion was observed in RT-PCR clones, as well as RLM-RACE clones (see the slightly broader band of RLM-RACE products in Fig. S4 at the URL mentioned above, lane dsRNA-2 minus sense). We analyzed a total of 21 independent cDNA clones, including RACE clones, and 8 of them carried the same deletion. This suggested that two dsRNA-2 segments that differ from each other in size exist in W779, as recently found for a novel RNA mycovirus in Phytophthora infestans (9).
As shown in Fig. Fig.3B,3B, dsRNA-1 and -2 each possess two ORFs coding for a total of four proteins. The deduced amino acid sequence of P1, P3, or P4 did not yield any significant hits, with E values below 0.01, in BLAST or FASTA3 searches. A BLAST search detected low levels of sequence similarities only between the dsRNA-1 ORF2-encoded P2 and RdRps of members of the families Totiviridae and Chrysoviridae. The highest sequence identity scores were found by a BLAST search between P2 and RdRps encoded by virus-like dsRNAs isolated from two different basidiomycetes, Phlebiopsis gigantea (PgV1) (30%) (35) and Lentinula edodes (LeV-HKB) (25%). The partial sequences of those dsRNAs were recently published or are publicly available only in the database. Table Table11 summarizes the results of a BLAST search with dsRNA-1 P2 (RdRp), indicating sequence identities found between aligned regions, bit scores, and E values among 17 mycoviruses. FASTA3 searches with P2 yielded hits similar to those listed in Table Table11.
A search of the PROSITE motif database with dsRNA-1 ORF2-encoded protein P2 detected functional sequence motifs typical of RdRps that included the SG—-TS/T and GDD motifs. This domain was readily identified by a Pfam search as the RdRp_4 motif. In addition, a search of the PROSITE database revealed that ORF2-encoded P2 contains a carbamoyl phosphate synthase subdomain signature 2 motif (INECNLKV) in the terminal region (amino acid positions 927 to 934) when default settings were used to exclude common motifs. No other sequence motif was observed in the polypeptides (P1, P3, and P4) coded for by ORF1, -3, and -4, except for frequently found motifs like phosphorylation sites and glycosylation sites.
A multiple amino acid sequence alignment of RdRp motifs I to VIII and their flanking regions of P2 and those of related viruses was constructed with CLUSTAL_X (Fig. (Fig.5),5), which was modified manually based upon the alignments reported by Bruenn (7), Jiang and Ghabrial (32), and Hillman et al. (26). All of the mycoviruses listed in Table Table11 were subjected to phylogenetic analysis. As reported earlier (7, 32), a few amino acid residues were strictly conserved in most of the motifs, e.g., K-E-G—R in motif III and D-DFN-H in motif IV. Motifs I and II were not found in members of the families Partitiviridae and Hypoviridae. Therefore, alignments of sequences from motifs III to VIII were employed for the construction of a phylogenetic tree in which the RdRp sequences of members of the family Hypoviridae were used as the outgroup. As expected from the results of a BLAST search (Table (Table1),1), the dendrogram showed that W799 RdRp clusters with dsRNA elements from P. gigantea and L. edodes. Importantly, their clade is separate from those of the other known dsRNA mycovirus families (Fig. (Fig.6).6). W799 RdRp is evolutionarily more closely related to members of the family Chrysoviridae than to members of the families Totiviridae and Partitiviridae. Grouping of chrysoviruses and members in the W779 RdRp clade is supported by a relatively strong bootstrap value, 839 in a total of 1,000 trials.
R. necatrix strain W1015, which was cured of the W779 virus, showed enhanced mycelial growth and virulence (Fig. (Fig.1),1), while introduction of the virus back into W1015 by anastomosis resulted in reduced virulence (data not shown). These results strongly suggest that the phenotypic changes were induced by virus infection. To further examine the possibility that the W779 virus can cause phenotypic alterations in host fungal strains other than W779, we attempted to transfect purified particles into virus-free spheroplasts derived from virus-free isolates W97 and W370T1 of R. necatrix. W370T1 was previously shown to be susceptible to a mycoreovirus (MyRV3) and a partitivirus (RnPV1) and to be stable phenotypically when transfected with virus particles (33, 44, 45; S. Kanematsu, unpublished data). Mycelial incompatibility between W779 (MCG351) and W370T1 (MCG139) or W97 (MCG80) impeded lateral transfer of the W779 virus when the two strains were cocultured (data not shown). However, transfected colonies of W97 and W370T1 became infected with the virus and showed mycelial growth rates decreased by 55 to 65%, as well as reduced growth of aerial hyphae relative to that of the virus-free strain (Fig. (Fig.7A).7A). Virus infection was stably maintained in transfectants during repeated subculturings. Subcultured colonies showed a phenotype identical to that shown in Fig. Fig.7A7A and were found by dsRNA extraction and gel electrophoretic analysis to carry dsRNA-1 and -2. Moreover, virus particles shown in Fig. Fig.2A2A were purified from these fungal colonies. Therefore, the morphological consequence of W779 virus infection is similar in three host backgrounds, W97, W370T1, and W1015 (compare Fig. Fig.1A1A and and7A7A).
The absence of residual mycelia in the virion fractions used for transfection was confirmed by culturing them on PDA. Furthermore, transfected strains were able to transmit the virus to their virus-free counterparts W97 and W370T1 through anastomosis but not to virus-free W1015, which is isogenic to W779, which is vegetatively incompatible with strains W97 and W370T1 (data not shown). These results eliminated the possibility that fungal colonies transfected with W779 particles originated from presumable contaminants.
Figure Figure7B7B shows the levels of virulence exhibited by untransfected [virus (−)] and transfected [virus (+)] strains W97 and W370T1. Representative apple plants inoculated with virus-free and virus-containing strain W370T1 are shown in Fig. Fig.7B7B (top). Virus-free fungal strains induced a lethal phenotype much more frequently in aerial parts of plants than did virus-containing fungal strains. More-severe symptoms were also evident in roots inoculated with virus-free fungal strains. In apple rootstocks inoculated with virus-containing strains, lesions on the taproot were small and superficial, allowing vigorous growth of healthy root hairs. By contrast, lesions induced by virus-free counterparts girdled the entire taproot, destroying most lateral roots. The mortality of apple rootstocks was much greater when they were inoculated with virus-free strains (7/10 for W97 and 10/10 for W370T1) than when they were inoculated with transfected strains (0/10 for W97 and 1/10 for W370T1) (Fig. (Fig.7B,7B, bottom). Taken together, these results clearly show that the W779 virus is responsible for hypovirulence and repressed mycelial growth in R. necatrix, regardless of the fungal host strain.
Fruit crops-R. necatrix-mycoviruses is an interesting pathosystem that attracts scientific and public attention in Japan. It is similar to the chestnut-C. parasitica-mycoviruses system that has contributed to a better understanding of fungal and viral pathogenesis (13, 16, 18, 19, 49, 50), including insights into RNA silencing in fungi as an antiviral response (47), viral replication (15, 27, 24), and biocontrol with mycoviruses (2, 25, 38). R. necatrix is a destructive soilborne fungal phytopathogen of economically important fruit crops. Although most fungal viruses cause symptomless infections (23), several viruses are known to induce phenotypic alterations in this fungus (33, 45). Thus, R. necatrix and the associated mycoviruses provide a good system with which to explore virus-host and virus-virus interactions based on established methods of host fungus transformation and artificial virus inoculation, the latter being still limited to a few viruses. Furthermore, the potential of mycoviruses infecting R. necatrix as biocontrol (virocontrol) agents has been suggested (37). Biological and molecular characterization of the novel bipartite dsRNA virus infecting a field strain, W779, of R. necatrix is important from the perspectives of fundamental fungal virology and practical virocontrol with mycoviruses.
Several lines of evidence have been presented in support of the conclusion that dsRNA-1 and -2 represent the genome of a novel virus species. Firstly, only dsRNA-1, not dsRNA-2, encodes an RdRp, suggesting that dsRNA-1 and -2 are not derived from two independent viruses. Both segments show similar sequences at the 5′ and 3′ ends (Fig. (Fig.4A;4A; see Fig. S3 at the URL mentioned above), a fact considered to be important for virus replication. The conserved terminal property is frequently observed in the genome segments of multipartite viruses. The MALDI-TOF MS analysis assigned peptide sequences in the 135-kDa capsid protein to those of dsRNA-1 ORF1-encoded P1 (see Table S2 at the URL mentioned above) but not to those of P2, P3, or P4, indicative of the absence of capsid proteins other than the 135-kDa protein that might show a similar migration position. Thus, the two segments are encapsidated into particles ~50 nm in diameter (Fig. (Fig.2A)2A) consisting of a single major structural protein (P1) of 135 kDa. A clue to addressing whether the two segments are contained in single particles comes from inspection of Fig. Fig.2B2B and of Fig. S1 at the URL mentioned above. Variable and unequal molar ratios of dsRNA-1 and dsRNA-2 in purified virus particles favor the idea that the two segments are separately packaged into particles composed of a single major capsid protein. These findings led us to conclude that the two elements dsRNA-1 and -2 represent the genome of a novel virus for which we propose the name Rosellinia necatrix megabirnavirus 1 strain W779 (RnMBV1/W779). Megabirna is from the much greater (mega) size (approximately 16 kbp) of its bisegmented dsRNA genome (birna for bipartite dsRNA genome) than those of members of the family Birnaviridae (approximately 6 kbp) (14) or Picobirnaviridae (approximately 4 kbp) (http://www.expasy.ch/viralzone/all_by_protein/603.html).
dsRNA-1 and -2 are similar to each other in genetic organization; that is, each has an extremely long 5′ UTR, two ORFs, and a relatively short 3′ UTR. Their genetic organization raises an interesting question as to how the ORFs are expressed. The extremely long (>1.6 kb) 5′ UTRs of dsRNA-1 and -2 are similar in sequence (see Fig. S3 at the URL mentioned above) and contain a number of minicistrons (see Fig. S5 at the URL mentioned above). Those minicistrons would hamper canonical translation of 5′-proximal large ORF1 and -3, according to the scanning model (34). The 5′-proximal ORFs are likely to be expressed by a noncanonical translational mechanism. A parallelism in size and the presence of mini-ORFs are found between the 5′ UTRs of RnMBV1 and members of the genus Aphthovirus of the family Picornaviridae. Picornaviruses utilize internal ribosome entry site (IRES)-mediated translation for expression of their polyproteins, requiring different sets of initiation factors, depending on the IRES type (6). Picornavirus IRES sequences have well-defined structural and sequence motifs. Although no IRES is identified in fungal or mycoviral mRNAs, IRES activities are implicated in the 5′ UTRs of genome size mRNAs of hypoviruses (27) and totiviruses (22). It will be interesting to examine whether the 5′ UTRs of RnMBV1 dsRNA-1 and dsRNA-2 serve as IRES sites. Guo et al. (24) developed a reporter assay system to assess low expression levels of a downstream gene in a dicistronic configuration in the filamentous fungus C. parasitica. A similar approach could be utilized to delineate the possible expression strategy for ORF1 and -3.
How 3′-proximal ORF2 (P2) and ORF4 (P4) are expressed is also largely unknown. Preliminary experiments failed to identify subgenomic RNAs corresponding to the downstream ORFs, while full-length mRNAs of dsRNA-1 and -2 were detected (data not shown), suggesting a polycistronic nature of those mRNAs. It should be noted that dsRNA-1 ORF2 is in −1 frame relative to ORF1 and an in-frame UGA triplet is located at map positions 5332 to 5334, which is 69 codons upstream of the start codon of ORF2 at positions 5539 to 5541 and 25 nt upstream of the stop codon of ORF1 at 5357 to 5359 (Fig. (Fig.3B).3B). Moreover, the FSFinder program readily identified a possible slippery sequence, 5′-A-AAA-AAC-3′ (consensus sequence: X-XXN-NNZ; X, any; N, A or U; Z, any but G) and a stem-loop structure at bases 5350 to 5356 and 5368 to 5425. These two elements are situated immediately upstream and downstream of the termination codon of ORF1, respectively. These findings may suggest that ORF2 is expressed by a −1 frameshift mechanism. This strategy is widely used by many animal (e.g., members of the family Retroviridae and the order Nidovirales), plant (e.g., members of the family Luteoviridae and the genus Dianthovirus), and fungal (e.g., members of the genus Totivirus) viruses. In almost all of these examples, the catalytic domains (pol or RdRp), which are required for their genome replication, are expressed by −1 frameshifting. In support of this notion, a minor protein band of 270 kDa was detected (Fig. (Fig.2C,2C, gray arrowhead), although it can be a dimer form of the 135-kDa capsid protein. An immunological and/or biochemical study is needed to test this possibility.
dsRNA-2 ORF4 is situated in the same frame as ORF3 without termination codons between the ORF3 stop and ORF4 start codons. A readthrough mechanism may allow expression of ORF4 as a fusion protein with P3 from the dsRNA-2 full-length mRNA if ORF4 (P4) is translated. As an alternative possibility, IRES-mediated translation of ORF4 is not ruled out.
While proteins P1, P3, and P4 do not show any significant sequence similarity to known protein sequences, dsRNA-1 ORF2-encoded P2 shows low levels of sequence identity to RdRps encoded by virus-like dsRNAs and dsRNA mycoviruses belonging to the families Totiviridae and Chrysoviridae (Table (Table1).1). However, RnMBV1 is readily differentiated from these related mycoviruses in genome size, genome segment number, and/or genome organization. Moreover, RnMBV1 particles are larger than those of any other International Committee for the Taxonomy of Viruses-approved mycoviruses, except mycoreoviruses, which are known to form rigid particles (30 to 40 nm in diameter). The phylogenetic tree generated based on RdRp sequence alignment (Fig. (Fig.5)5) reveals that RnMBV1 RdRp clusters with the partially characterized dsRNA elements PgV1 and LeV-HKB from basidiomycetes (Fig. (Fig.6),6), with a high bootstrap value of 951, and is placed separately from other known dsRNA mycoviruses. The members of this cluster and members of the family Chrysoviridae might have diverged from an immediate ancestral virus. The grouping of these members is supported by a relatively high bootstrap value of 839. Based on these data, we propose a new virus family, Megabirnaviridae, containing RnMBV1/W779 as the type species. At least one tentative member of the proposed family occurs in strain W8 of R. necatrix that is coinfected with a partitivirus, RnPV1 (44; Sasaki, unpublished). Corresponding proteins between RnMBV1 and the tentative member exhibit moderate levels of sequence identity (Sasaki, unpublished). Furthermore, another strain of RnMBV1 was detected at a geographically distant location from Ibaraki Prefecture where W779 was collected (Kanematsu, unpublished). These results show that members of the proposed Megabirnaviridae family occur widely in R. necatrix.
It is often difficult to establish the etiology of mycovirus because of the lack of an introduction method (22). One of the solutions to this problem is the construction of infectious viral cDNA clones or preparation of in vitro-synthesized transcripts derived from virus cDNA clones. Examples include plus-sense ssRNA fungal viruses like hypoviruses and narnaviruses for which virus infection can be initiated by transcripts from chromosomally integrated full-length viral cDNA or by synthetic transcripts directly introduced into the cytoplasm of host cells (11, 12, 17). Another approach is virion transfection using purified virus particles. However, this approach has rarely been successful. Thus, mycoviruses were once regarded as “noninfectious, endogenous, or heritable” (21). Recently, reproducible PEG-mediated transfection methods with virus particles were developed for mycoreoviruses (27, 28). Since then, the approach was applied to not only members of the genus Mycoreovirus (45) but also a member of the family Partitiviridae (46). This study provided the third example of such viruses that are infectious to R. necatrix as particles. Particles of dsRNA viruses are believed to contain all of the enzymatic activities necessary for transcription and the synthesis of viral mRNA, leading to the onset of infection. The virion transfection technique may allow examination of the experimental host range of not only RnMBV1 but also other dsRNA mycoviruses producing rigid particles like members of the families Chrysoviridae and Totiviridae.
This study clearly established a virus etiology for the reduced virulence and vegetative mycelial growth of R. necatrix. Curing of RnMBV1 in W779 results in the generation of an isogenic fungal strain, W1015, that shows an enhanced growth rate and virulence level (Fig. (Fig.1).1). The virus is able to infect strains W370T1 and W97 via virion transfection and induce a similar set of phenotypic alterations, i.e., reduced mycelial growth and attenuated virulence (hypovirulence) (Fig. (Fig.7).7). From these RnMBV1-infected strains, virus particles encapsidating dsRNA-1 and -2 were isolated. The transfected fungal strains, W97 and 370T1, belonging to MCG80 and MCG139 (4), are vegetatively incompatible with W779 (MCG351) (30) and unable to receive RnMBV1 when cocultured with W779. It was confirmed that the transfected strains shown in Fig. Fig.77 carried the phenotypic markers of their parents. Thus, a possibility of contamination with the W779 fungal cells during transfection is ruled out. These combined results represent the completion of Koch's postulates for this novel virus and provide a basis for future identification of viral and host factors associated with symptom expression.
This study was conducted as part of a project entitled Development of Introduction Protocols of Virocontrol Agents into Phytopathogenic Fungi toward a feasible virocontrol of primarily root rot of fruit trees in Japan (23). To achieve virocontrol (one form of biological control) of the disease white root rot, we need to breach three barriers: (i) isolation of mycoviruses with virocontrol potential (virocontrol agents), (ii) transfection of fungal strains of interest with virocontrol agents, and (iii) efficient spread of virocontrol agents ideally at individual and population levels, as well as the conversion of virulent fungal strains to hypovirulent ones in the soil once they are introduced. As discussed above, the novel virus RnMBV1 is infectious as particles and able to confer hypovirulence, under laboratory conditions, on virus-free fungal strains that otherwise are virulent. Therefore, we have overcome the first two constraints by developing an artificial introduction method for RnMBV1 with the ability to attenuate virulence. From the practical point of view, issue iii is critical and may be influenced by many factors, like the ecological fitness of virocontrol agents and the genetic structures of causal fungal pathogens. One crucial factor is the diversity of MCGs in a given population, since lateral transmission of mycoviruses is expected to be inversely correlated with MCG diversity (38). This notion is exemplified by the successful biocontrol of chestnut blight disease in Europe. An opposing example is the failure of attempts at biocontrol of the same disease in eastern North America, where complex populations of the causal pathogen are presumed to impede the spread of virulence-attenuating viruses via anastomosis in a plot. Interesting ecological studies of the white root rot fungus were carried out previously that showed simple population structures of the fungus in single infested orchards of Japanese pears (H. Nakamura, personal communication). That is, only a few strains belonging to different MCGs dominate single infested fields. The less diverse structure of fungal populations will undoubtedly stimulate future research on the virocontrol of white root rot disease.
We are grateful to Yomogi Inc. (N.S.) and the Program for Promotion of Basic and Applied Researches for Innovations in Bio-Oriented Industries (S.K., A.S., and N.S.) for financial support during this study.
We thank Said A. Ghabrial for fruitful discussion and critical reading of the manuscript. We acknowledge generous gifts of fungal strains and plasmid clones from Said A. Ghabrial, Bradley I. Hillman, and Donald L. Nuss. Technical assistance in PMF by Akio Tani is greatly appreciated.
Published ahead of print on 14 October 2009.