Two Novel Rhabdoviruses Related to Hypervirulence in a Phytopathogenic Fungus

ABSTRACT

Rhabdoviruses are ubiquitous and diverse viruses that propagate owing to bidirectional interactions with their vertebrate, arthropod, and plant hosts, and some of them could pose global health or agricultural threats. However, rhabdoviruses have rarely been reported in fungi. Here, two newly identified fungal rhabdoviruses, Rhizoctonia solani rhabdovirus 1 (RsRhV1) and RsRhV2, were discovered and molecularly characterized from the phytopathogenic fungus Rhizoctonia solani. The genomic organizations of RsRhV1 and RsRhV2 are 11,716 and 11,496 nucleotides (nt) in length, respectively, and consist of five open reading frames (ORFs) (ORFs I to V). ORF I, ORF IV, and ORF V encode the viral nucleocapsid (N), glycoprotein (G), and RNA polymerase (L), respectively. The putative protein encoded by ORF III has a lower level of identity with the matrix protein of rhabdoviruses. ORF II encodes a hypothetical protein with unknown function. Phylogenetic trees based on multiple alignments of N, L, and G proteins revealed that RsRhV1 and RsRhV2 are new members of the family Rhabdoviridae, but they form an independent evolutionary branch significantly distinct from other known nonfungal rhabdoviruses, suggesting that they represent a novel viral evolutionary lineage within Rhabdoviridae. Compared to strains lacking rhabdoviruses, strains harboring RsRhV2 and RsRhV1 showed hypervirulence, suggesting that RsRhV1 and RsRhV2 might be associated with the virulence of R. solani. Taken together, this study enriches our understanding of the diversity and host range of rhabdoviruses.

IMPORTANCE Mycoviruses have been attracting an increasing amount of attention due to their impact on important medical, agricultural, and industrial fungi. Rhabdoviruses are prevalent across a wide spectrum of hosts, from plants to invertebrates and vertebrates. This study molecularly characterized two novel rhabdoviruses from four Rhizoctonia solani strains, based on their genomic structures, transcription strategy, phylogenetic relationships, and biological impact on their host. Our study makes a significant contribution to the literature because it not only enriches the mycovirus database but also expands the known host range of rhabdoviruses. It also offers insight into the evolutionary linkage between animal viruses and mycoviruses and the transmission of viruses from one host to another. Our study will also help expand the contemporary knowledge of the classification of rhabdoviruses, as well as providing a new model to study rhabdovirus-host interactions, which will benefit the agriculture and medical areas of human welfare.

INTRODUCTION

Mycoviruses (or fungal viruses) are widespread in all major taxa of fungi, with wide-ranging impacts on fungi (1). Mycoviruses have increasingly been attracting attention, largely because some of them confer hypovirulence, thereby making virocontrol an effective strategy to combat fungal diseases. Cryphonectria parasitica hypovirus 1 is a classic example and a star in the mycovirus research arena due to its association with hypovirulence, and it has successfully controlled chestnut blight in some parts of Europe (2). Some mycoviruses can also confer hypervirulence, which is characterized by enhanced sporulation, aggressiveness, and growth (3–5). Although most mycoviruses have no apparently significant effects on their hosts, they could still potentially play essential, yet to be explored, roles in the natural ecological system (1).

Mycoviruses have diverse genomes with double-stranded RNA (dsRNA), positive-strand single-stranded RNA (+ssRNA), negative-strand ssRNA (−ssRNA), and single-stranded DNA. −ssRNA viruses in animals and plants have gained considerable attention because of their implications in dangerous medical and agricultural diseases (6–8). However, limited knowledge is available regarding mycoviruses with −ssRNA genomes. In 2013, the genomic information of the −ssRNA viruses was explored from public fungal transcriptome data, which suggests that −ssRNA mycoviruses potentially infect fungi (9). In 2014, the first −ssRNA mycovirus was characterized in a phytopathogenic fungus and confirmed to be associated with hypovirulence (10), which subsequently establishes a new family, Mymonaviridae (11). Since then, using high-throughput sequencing techniques, an increasing number of the novel −ssRNA mycoviruses are being discovered in phytopathogenic fungi and mushrooms (12, 13), implying that −ssRNA mycoviruses are much more abundant than previously thought. Although most of those −ssRNA mycoviruses are phylogenetically related to members of Mononegavirales and Bunyavirales or are undefined in their classification position, they usually represent viral evolutionary lineages that are significantly different from viruses known in animals and plants. Therefore, the discovery of novel −ssRNA mycoviruses will enrich existing knowledge and shed light on the diversity and evolution of the whole virosphere.

Rhabdoviridae is arguably the largest family within the order Mononegavirales and includes 20 genera and 144 species (14). Rhabdoviruses are distributed worldwide and have a wide array of hosts, including vertebrates, invertebrates, and plants, and some of them cause severe diseases in their hosts (15). To date, however, very few cases of rhabdoviruses have been recorded in the kingdom Fungi. The first fungal rhabdovirus was recently characterized in a phytopathogenic fungus (16). Classic rhabdovirus particles from vertebrates are characteristically bullet or cone shaped (17, 18), whereas rhabdovirus particles from plants are bacilliform-enveloped virions (19). Rhabdoviruses usually have an unsegmented genome that is 11 to 16 kb in length and encodes five typical proteins, including nucleoprotein (N), phosphoprotein (P), matrix protein (M), glycoprotein (G), and RNA-dependent RNA polymerase protein (L) (20, 21). These genes are linearly arranged in the genome in the order 3′-N-P-M-G-L-5′, and the transcription of their mRNA is also sequential and polar (22). However, genome organization in rhabdoviruses is highly diverse and complex because the five canonical genes may be overlapped or interspersed due to evolutionary gain and loss of accessory genes via continuous adaptation and purifying selection (8, 23).

Rhizoctonia is an important and complex genus of soilborne and semisaprophytic fungi and is responsible for various plant diseases. Rhizoctonia solani (teleomorph: Thanatephorus cucumeris) is regarded as the most important and complex population within the genus Rhizoctonia and is mainly divided into 14 anastomosis groups (AG-1 to AG-13 and AG-BI) according to the ability to anastomose with tester strains (24, 25). R. solani AG-1IA is one of the most devastating phytopathogenic fungi in rice, wheat, and maize production. Rice sheath blight caused by R. solani AG-1IA generally results in rice yield losses of 20% to 45% (26, 27). Like other phytopathogenic fungi, mycovirus infections are common phenomena in the R. solani population (28, 29). Coinfection of a single R. solani strain by multiple mycoviruses has been frequently observed (30). Increasing evidence suggests that the newly identified mycoviruses are extremely diverse in R. solani and some of them show unique genomic characteristics (28, 31), implying that our current understanding of the diversity of RNA viruses in R. solani remains far from complete. Most of the known mycoviruses in R. solani cause latent infections. However, some of them also confer hypovirulence (32, 33) or hypervirulence (34) to R. solani, indicating diverse interactions between mycoviruses and R. solani.

In the present study, we isolated strain XY175 of R. solani AG-1IA and used high-throughput sequencing to confirm its coinfection by six mycoviruses; two of them had unsegmented −ssRNA genomes, and their full-length genomes were determined. Phylogenetic analysis revealed that they were related to rhabdoviruses, providing solid evidence that rhabdoviruses can infect fungi besides animals and plants. The potential effects of the two rhabdoviruses on the biology of the fungal host were also investigated. Our data support the formation of a new genus within the family Rhabdoviridae to accommodate these phylogenetically related viruses.

RESULTS

Strain XY175 contains six mycoviruses with ssRNA genomes.

To effectively identify mycoviruses infecting strain XY175, metatranscriptome sequencing was conducted, which generated more than 12 GB of total sequence data. After removal of any low-quality reads, 3.6 × 10⁷ reads with lengths of >20 nucleotides (nt) (paired-end) were obtained. The reads that were well matched with the genome of R. solani were filtered and assembled de novo into larger contigs. Consequently, all contigs were subjected to BLASTx analysis against the NCBI nonredundant database. Sixty-four contigs that could represent partial or nearly complete genomes of the six potential mycoviruses were obtained and confirmed via reverse transcription (RT)-PCR with virus-specific primers (see Table S1 in the supplemental material). Based on the replication-related protein analysis, we deduced that the six mycoviruses contained ssRNA genomes, and we temporarily named them Rhizoctonia solani rhabdovirus 1 (RsRhV1), RsRhV2, Rhizoctonia solani endornavirus 9 (RsEV9), RsEV10, Rhizoctonia solani mitovirus 40 (RsMV40), and RsMV41. Thus, strain XY175 was coinfected by six ssRNA mycoviruses, including rhabdoviruses, endornaviruses, and mitoviruses, and their detailed information is listed in Table 1.

TABLE 1.

Detailed information on six mycoviruses coinfecting R. solani strain XY175

Name of putative virus	UniGene identification	Length (nt)	Best match via BLASTp	Identity (%)	Genome type	Family	E value
Rhizoctonia solani rhabdovirus 1	First_contig1571	3,471	Sclerotinia sclerotiorum rhabdovirus 1	57	−ssRNA	Rhabdoviridae	0
Rhizoctonia solani rhabdovirus 2	Contig107	11,415	Sclerotinia sclerotiorum rhabdovirus 1	54	−ssRNA	Rhabdoviridae	0
Rhizoctonia solani endornavirus 9	Contig4069	19,932	Rhizoctonia solani endornavirus 1	97	+ssRNA	Endornaviridae	6E−122
Rhizoctonia solani endornavirus 10	First_contig2675	16,137	Ceratobasidium endornavirus F	36	+ssRNA	Endornaviridae	5.00E−87
Rhizoctonia solani mitovirus 40	First_contig15	4,358	Macrophomina phaseolina mitovirus 3	41	+ssRNA	Mitoviridae	3.00E−77
Rhizoctonia solani mitovirus 41	First_contig6	4,208	Alternaria alternata mitovirus 1	53	+ssRNA	Mitoviridae	8.00E−125

Determination and analysis of the complete genomes of the two newly discovered rhabdoviruses.

Despite the increasing number of reports on rhabdoviruses in the kingdoms Plantae and Animalia, rhabdoviruses infecting fungi have rarely been reported. Therefore, we focused on the two rhabdovirus-related mycoviruses identified in the present study, RsRhV1 and RsRhV2, and determined their complete genomes by assembling sequences derived from the overlapping sequences of metatranscriptomic data, rapid amplification of cDNA ends (RACE) products, and a serial RT-PCR product (see Table S1).

(i) RsRhV1 genome.

The full-length genome of RsRhV1 comprised 11,716 nt and had five putative open reading frames (ORFs) (ORFs I to V) with a linear arrangement in the antigenome sense (Fig. 1A and Table 2). A shorter 3′ leader sequence and a longer 5′ trailer sequence, 84 and 244 nt in length, respectively, were determined (Table 2). ORF I (nucleotide positions 85 to 1558) was predicted to encode a 490-amino acid (aa)-long N protein. ORF IV (nucleotide positions 3463 to 4951) was predicted to encode a putative 495-aa-long G protein. A signal peptide (nucleotide positions 3463 to 3516) and a transmembrane domain (nucleotide positions 4825 to 4893) were predicted to be present in the ORF IV-encoded G protein. ORF V (nucleotide positions 5184 to 11469), the largest of the five ORFs, was predicted to encode a 2,094-aa-long replicase-related protein (L protein). Proteins encoded by ORF I, ORF IV, and ORF V have 37%, 45%, and 57% identity, respectively, with N, G, and L proteins of a recently reported fungal rhabdovirus (Sclerotinia sclerotiorum rhabdovirus 1 [SsRhV1]). ORF II (nucleotide positions 1610 to 2735) and ORF III (nucleotide positions 2764 to 3433) were predicted to encode two hypothetical proteins with unknown functions.

FIG 1.

Characterization of the genome structure of RsRhV1 and RsRhV2. (A and B) Genome size and organization of RsRhV1 (A) and RsRhV2. (B) ORF I putatively encodes N protein, ORFs IV and V putatively encode G and L proteins, and ORF II and III putatively encode two hypothetical proteins with unknown functions. (C) Alignment of the putative gene junction sequences shown in 3′-to-5′ orientation. TI sequences are represented by gray, and TPP sequences are represented by black. (D) Percent identity matrix, generated using Clustal Omega 2.1, of the typical species in each genus within Rhabdoviridae. The top half of the matrix, with the blue scale, is the percent identity of the RNA polymerase (L protein, LP) coding sequence, and the bottom half of the matrix, with the green scale, is the percent identity of the putative N protein (NP). For the sake of clarity, 100% identity along the diagonal was removed, and outlined red dots were used to note the junction point for the rows and columns of the identified viruses. Only LP values of ≥25% and NP values of ≥25% are indicated.

TABLE 2.

Features of ORFs and their encoded proteins in the RsRhV1 and RsRhV2 genomes

ORF	Putative gene function	Nucleotide position of ORF	Length (nt)	Molecular mass (kDa)	pI	No. of glycosylation sites		No. of phosphorylation sites
						N-linked	O-linked	Ser	Thr	Tyr
RsRhV1
1	N protein	85–1558	1,470	54.7	6.79	3	5	19	23	12
2	P2 (unknown)	1610–2735	1,122	41.7	5.04	1	36	36	14	4
3	P3 (unknown)	2764–3433	666	26.1	7.91	1	1	7	7	5
4	G protein	3463–4951	1,485	57.6	7.69	1	7	29	16	10
5	L protein	5184–11469	6,282	239.0	8.47	9	19	120	59	28
RsRhV2
1	N protein	111–1521	1,407	53.8	8.20	0	1	16	15	11
2	P2 (unknown)	1566–2508	939	35.4	4.73	1	14	19	16	0
3	P3 (unknown)	2545–3175	627	24.7	8.79	0	13	13	9	5
4	G protein	3203–4691	1,485	57.1	8.50	2	2	36	13	12
5	L protein	4923–11253	6,327	242.1	8.77	9	21	137	69	30

(ii) RsRhV2 genome.

RsRhV2 had a genomic organization similar to that of RsRhV1. The complete genome of RsRhV2 was 11,469 nt long and contained five large ORFs (Fig. 1B and Table 2). BLASTp analysis of the five ORF-encoded proteins (3′-N-PII-PIII-G-L-5′) was conducted on the NCBI website, and the results showed that the N (469 aa), G (495 aa), and L (2,190 aa) proteins displayed homology to known proteins with relatively low amino acid identity, except for SsRhV1. The N, G, and L proteins share 41%, 40%, and 56% identity, respectively, with the corresponding proteins of SsRhV1. No significant similarity of proteins encoded by ORF II and ORF III to known viral proteins was found in the BLASTp analysis.

(iii) Comparison of RsRhV1 and RsRhV2 genomes with a newly identified fungal rhabdovirus.

Proteins encoded by the five large ORFs of RsRhV1 and RsRhV2 were compared with corresponding sequences of plant and animal rhabdoviruses, and a percent identity matrix was constructed based on the L and N proteins (Fig. 1D). The sequence identities of N, G, L, and the ORF III-encoded protein between RsRhV1 and RsRhV2 were 36%, 45%, 55%, and 45%, respectively. The ORF II-encoded proteins did not have any identities between RsRhV1 and RsRhV2. Both RsRhV1 and RsRhV2 exhibited less than 35% identity to other classified nonfungal rhabdoviruses. These results suggest that RsRhV1 and RsRhV2 are more closely related to each other than to any other rhabdovirus. Additionally, proteins encoded by RsRhV1 and RsRhV1 contained multiple potential glycosylation and phosphorylation sites (Table 2).

(iv) Intergenic regions.

Similar to all reported rhabdoviruses, the five ORFs of both RsRhV1 and RsRhV2 were separated by intergenic regions consisting of a putative polyadenylation signal (for the former gene), an nontranscribed intergenic sequence, and a putative transcription start site (for the latter gene). Multiple alignment models revealed that highly conserved intact termination and polyadenylation (TTP) sequences (3′-AUACUUUUUUU-5′) and transcript initiation (TI) sequences (3′-UUGUC-5′) were observed in each intergenic region (Fig. 1C). This is consistent with the sequence of transcription termination regions in the typical rhabdovirus genome (35). However, the genomic structures of the gene junctions of RsRhV1 and RsRhV2 were different. Only the gene junctions between ORF I and ORF II of RsRhV1 and RsRhV2 conformed to the classic law. Other gene junctions of RsRhV1 and RsRhV2 included a TI sequence, nontranscribed nucleotides 14 to 190, and a TTP sequence. The orders of conservative TI and TTP motifs in RsRhV1 and RsRhV2 were different. We speculate that the transcriptional pattern of rhabdoviruses infecting fungi is quite different from that of animal rhabdoviruses.

(v) Characterization of accessory genes of the two fungal rhabdoviruses.

Apart from the five canonical larger ORFs in the two fungal rhabdoviruses, we also found nine smaller ORFs (accessory genes) >150 nt in length, and eight of them shared no detectable protein similarity with those in public databases (Fig. 2A). The RsRhV1 genome contains six potential accessory genes (Fig. 2B). Two accessory genes, Nx and Ny, were found in an alternative ORF located in the N gene, and another two, Gx and Gy, were detected in an alternative ORF in the G gene. The accessory gene U1 was an independent transcriptional unit located between the G and L genes. The accessory gene Lx was located within an alternative ORF in the region close to the end of the L gene and putatively encodes a U32 family peptidase (Photobacterium sp. strain CECT 9192) (Fig. 2B). Unlike RsRhV1, RsRhV2 contained only three potential accessory genes. Two accessory genes, Nx and Lx, were found in alternative ORFs located in the N and L genes, respectively. In addition, U1 of RsRhV2 was an independent transcriptional unit behind the L gene, encoding unknown hypothetical proteins (Fig. 2C).

FIG 2.

Comparative genomic organization of representative members within the family Rhabdoviridae. (A) The five canonical structural protein genes (N, P, M, G, and L) are shaded in different colors. The positions of the ORFs are indicated, and those encoding accessory proteins are shaded. The 3′-leader (le) and 5′-trailer (tr) regions are scaled. Virus names (abbreviation of the type species of the genera) and accession numbers are shown. (B) Comparison of RsRhV1 genomic organization with that of other rhabdoviruses. Genes encoding proteins with unknown functions are represented by small arrows in different colors. (C) Comparison of RsRhV2 genomic organization with that of other rhabdoviruses. Genes encoding proteins with unknown functions are represented by small arrows in different colors. Abbreviations of virus names are as listed in Table S1 in the supplemental material.

(vi) Transcriptional mode of RsRhV1 and RsRhV2.

The 5′ untranslated region (UTR) and the 3′ UTR of the five genes (ORFs I to V) from RsRhV1 and RsRhV2 were obtained and sequenced (Fig. 3A and C). The start and termination sites for each gene transcription were confirmed by resequencing. Based on sequences of the 5′ and 3′ UTRs, a transcript map of each gene of RsRhV1 and RsRhV2 is shown in Fig. 2. All five genes of RsRhV1 and RsRhV2 could be transcribed independently (Fig. 3B and D). In both RsRhV1 and RsRhV2, ORF III could be cotranscribed with ORF IV because two types of 3′-RACE products were amplified (Fig. 3A and C). To understand the expression profiles of all individual genes in RsRhV1 and RsRhV2, transcriptome analysis of R. solani strain XY175 was conducted. Similar to other reported rhabdoviruses (36), transcription of the five genes of RsRhV1 and RsRhV2 overall follows the gradient rule (Fig. 3E and F).

FIG 3.

Possible transcript maps of RsRhV1 and RsRhV2. (A and C) Agarose gel electrophoresis of the RACE products of ORFs I to VI (lanes 1 to 5) on 1.5% agarose gels for RsRhV1 (A) and RsRhV2 (C). Lane 6 is a negative control. (B and D) Possible transcript maps of RsRhV1 (B) and RsRhV2 (D). (E and F) Expression profiles of five major genes from RsRhV1 (E) and RsRhV2 (F), based on the transcriptome analysis of strain XY175.

RsRhV1 and RsRhV2 are related to but distant from the known rhabdoviruses, based on phylogenetic analysis.

To clearly define the relationship between the fungal rhabdoviruses (RsRhV1, RsRhV2, and SsRhV1) and nonfungal rhabdoviruses, multiple alignments of L proteins were constructed (Fig. 4A). MotifFinder revealed that the L proteins of RsRhV1 and RsRhV2 contained two conserved domains, i.e., Mononeg_RNA_pol (pfam00946), which is a typically conserved domain of members belonging to the order Mononegavirales (37, 38), and paramyx_RNAcap (TIGR04198), which is vital for the formation of the mRNA cap structures (39). The Mononeg_RNA_pol domains of the two fungal rhabdoviruses displayed six conserved motifs (Fig. 4A). Furthermore, an alignment of the G proteins from RsRhV1 and RsRhV2 with members of the Almendraviru genera indicated that RsRhV1 and RsRhV2 contained eight conserved cysteine residues in the ectodomain, with two additional cysteine residues that are absent in other rhabdoviruses (Fig. 4B).

FIG 4.

Multiple alignment analysis based on the L and G proteins of RsRhV1 and RsRhV2. (A) Amino acid sequence alignment of core L protein motifs of the selected RsRhV1 and RsRhV2 viruses belonging to Rhabdoviridae. The Roman letters (I to VI) represent the six conserved regions in the L protein sequences. (B) Amino acid sequence alignment of core G protein motifs of the selected RsRhV1 and RsRhV2 viruses belonging to Almendravirus. The signal domains and transmembrane domains are shaded gray. Cysteine residues in the ectodomain are shaded black. Fully conserved (*), strongly conserved (:), or weakly conserved (.) amino acids are indicated below the alignment. Abbreviations of virus names and their accession numbers in the GenBank database are as listed in Table S1 in the supplemental material.

Maximum likelihood (ML) phylogenetic analysis was conducted using the complete sequences of L, N, and G proteins from RsRhV1 and RsRhV2 with sequences from 85 other rhabdoviruses from 40 genera and other unclassified rhabdoviruses (Fig. 5A; also see Table S1 in the supplemental material). Although phylogenetic trees obtained from the L, G, and N proteins had different topologies, RsRhV1 and RsRhV2 always clustered in a distinct monophyletic group with SsRhV1 under solid bootstrap support. Based on L proteins, RsRhV1, RsRhV2, and SsRhV1 were clustered with members of the genus Sripuvirus and formed an independent clade with well-supported values. In contrast, the phylogenetic trees constructed using the G proteins and the N proteins revealed that RsRhV1, RsRhV2, and SsRhV1 were grouped in a clade related to members of the genus Almendravirus (Fig. 5B and C).

FIG 5.

Phylogenetic analysis of RsRhV1 and RsRhV2. The ML phylogeny of fungal rhabdoviruses and nonfungal rhabdoviruses belonging to 40 genera within Rhabdoviridae was determined. The ML tree was constructed based on the L protein (A), N protein (B), and G protein (C) alignment using PhyML v3.0 with the best-fit model LG with +I + G+F. Viruses assigned to the established genera are differentiated with different background colors. The host and transmission vectors of viruses are indicated. RsRhV1 and RsRhV2 are shown in red. Abbreviations of virus names and their accession numbers in the GenBank database are as listed in Table S1 in the supplemental material.

RsRhV1 and RsRhV2 are related to virulence in R. solani.

Three strains, 175A1VF13, 175A1VF15, and 175A1VF32, were single-protoplast regeneration derivatives of strain XY175, and their mycovirus contents were confirmed (Fig. 6A). Strain 175A1VF13 harbored four mycoviruses (RsEV9, RsEV10, RsMV39, and RsMV40) but lacked the rhabdoviruses RsRhV1 and RsRhV2. Besides the two mitoviruses and two endornaviruses coinfecting strains 175A1VF15 and 175A1VF32, strain 175A1VF15 was also infected by RsRhV1, while strain 175A1VF32 carried RsRhV2.

FIG 6.

Mycovirus content detection and biological features of strain XY175 and its protoplast-derived isolates. (A) Detection of virus contents in strain XY175 and its protoplast-derived isolates 175A1VF13, 175A1VF15, and 175A1VF32. Lanes 1 to 6 are specific PCR products with specific primers for RsRhV1, RsRhV2, RsEV9, RsEV10, RsMV40, and RsMV41, respectively. Lane M represents DNA molecular markers, lane 7 is a negative control, and lane 8 is positive control (glyceraldehyde-3-phosphate dehydrogenase [GAPDH] gene). (B and D) Colony morphology (B) and growth rate (D) of strain XY175 and three protoplast-derived isolates (28°C, 4 days postinoculation). (C and E) Virulence assays of strain 175A1 and three protoplast-derived isolates on rice leaves (28°C, 4 days). The lesion area on rice leaves was calculated using ImageJ software. (F) qRT-PCR analysis of the replication of four viruses, RsEV9, RsEV10, RsMV40, and RsMV41, in strain XY175 and its protoplast-derived isolates 175A1VF13, 175A1VF15, and 175A1VF32. The replication levels of the four viruses in strain XY175 were set as 1. *, significant difference, P = 0.05; **, significant difference, P = 0.01.

Four strains, XY175, 175A1VF13, 175A1VF15, and 175A1VF32, exhibited normal colony morphology and production of sclerotia. The growth rate of strain 175A1VF32 was significantly lower than that of strain XY175 but exhibited no significant difference from those of 175A1VF15 and 175A1VF32. Virulence assays of the four strains were conducted on the detached leaves of rice plants. Strains XY175 and 175A1VF32 displayed no significant difference in virulence on rice but showed greater virulence than strain 175A1VF15. However, the virulence of strain 175A1VF13, lacking the two rhabdoviruses, was significantly lower than that of the other strains (Fig. 6). Thus, we preliminarily inferred that RsRhV1 and RsRhV2 infections were involved in the virulence change of R. solani.

The abundance of four mycoviruses, including RsEV9, RsEV10, RsMV40, and RsMV41, in four strains were compared by quantitative RT-PCR (qRT-PCR) and transcriptome analysis. Compared with the original strain XY175, the abundance of two endornaviruses, RsEV9 and RsEV10, was significantly increased in strain 175A1VF13 (rhabdovirus-free strain) but was significantly inhibited in rhabdovirus-infected strains 175A1VF15 and 175A1VF32 (Fig. 6F). The change in abundance of two mitoviruses has a similar trend, compared with that of the two endornaviruses, although they are different in different strains (Fig. 6F). These results revealed that rhabdoviruses could inhibit the replication of endornaviruses and mitoviruses, suggesting that those mycoviruses likely interact with each other in R. solani.

To further define whether rhabdoviruses were related to hypervirulence on R. solani, experiments with dual cultures of strain XY175 (donor strain) and strain HG81R3 (recipient strain, belonging to AG-1IA) were conducted on potato dextrose agar (PDA) plates. Since strain HG81R3 harbors RsMV40 that shares 100% identity with that infecting strain XY175 (unpublished data), we determined whether the other five mycoviruses were able to transfer from strain XY175 to strain HG81R3. Unexpectedly, only RsRhV1 was horizontally transferred from XY175 to HG81R3, and two strains, HG81R3a and HG81R3c, that were picked up from the strain HG81R3 side were confirmed to be infected by RsRhV1. However, attempts with other viruses (RsRhV2, RsEV9, RsEV10, and RsMV41) were unsuccessful (Fig. 7A). In terms of virulence assays, the new RsRhV1-infected strains (HG81R3a and HG81R3c) showed stronger virulence than strain HG81R3 (Fig. 6C and D), suggesting that rhabdoviruses could potentially confer hypervirulence in R. solani.

FIG 7.

Mycovirus horizontal transmission. (A) Detection of virus contents in strains HG81R3, HG81R3a, and HG81R3c. Lanes 1 to 6 are specific PCR products with specific primers for RsRhV1, RsRhV2, RsEV9, RsEV10, RsMV40, and RsMV41, respectively. Lane M represents DNA molecular markers, lane 7 is a negative control (H₂O), and lane 8 is a positive control (GAPDH gene). (B) Colony morphology of strains HG81R3, HG81R3a, and HG81R3c (28°C, 4 days). (C) Virulence assay of strains HG81R3, HG81R3a, and HG81R3c on rice leaves of rapeseed (28°C, 4 days). (D) Lesion areas on rice leaves calculated using ImageJ software. **, significant difference, P = 0.01.

DISCUSSION

Rhabdoviruses constitute a diverse group of −ssRNA viruses and could successfully colonize a plethora of ecological niches, ranging from plants and insects to fish and mammals (22, 40). However, rhabdoviruses that infect fungi have not been well characterized. A contig sequence (GenBank accession number QHD64769.1) related to rhabdoviruses has recently been detected in the virome of Plasmopara viticola belonging to oomycetes (41); however, P. viticola does not belong to real fungi. The first complete genome of a fungal rhabdovirus, SsRhV1 (MT706019.1), was recently characterized (16). In the present research, we characterized two new rhabdoviruses, RsRhV1 and RsRhV2, based on genomic structures, transcription strategy, phylogenetic relationship, and biological impact on their host. The results of this study provide further evidence of rhabdoviruses infecting fungi and broaden the host range of rhabdoviruses.

Although genetic associations definitively placed RsRhV1 and RsRhV2 in the family Rhabdoviridae, within the order Mononegavirales, they were distinguished from known nonfungal rhabdoviruses and formed a distinct monophyletic group with another fungal rhabdovirus, SsRhV1, for multiple reasons. First, the extent of amino acid sequence identity between cognate proteins from the fungal rhabdoviruses and reported nonfungal rhabdoviruses was quite low, with less than 35% identity (Fig. 1D). Moreover, ORF II-encoded proteins in the fungal rhabdoviruses did not have any identities to known proteins in the NCBI database, whereas the second ORF usually encodes P proteins in nonfungal rhabdoviruses. Phylogenetic analysis also supported fungal rhabdoviruses forming an independent evolutionary lineage within Rhabdoviridae. Second, unlike the classic rhabdovirus, with each gene junction consisting of conserved sequence motifs (36), each gene junction of RsRhV1 and RsRhV2 was different. Only the gene junction between ORF I and ORF II of RsRhV1 and RsRhV2 conformed to the classic law. Other gene junctions of RsRhV1 and RsRhV2 were a TI sequence (3′-UUGUC-5′), nontranscribed bases 14 to 190, and a TTP sequence (3′-AUACUUUUUUU-5′). Interestingly, the order of conservative TI and TTP motifs differed between RsRhV1 and RsRhV2. Third, the host of RsRhV1 and RsRhV2 is a filamentous fungus, whereas other reported rhabdoviruses infect plants and animals. Therefore, it is quite significant that we report the isolation of two novel rhabdoviruses from a phytopathogenic fungus. Based on these genetic features, a new genus, Mycorhabdovirus, was proposed for the classification of fungal rhabdoviruses. It is notable, however, that the virions of SsRhV1 have bullet-shaped morphology (38), which is significantly different from that of RsRhV1 and RsRhV1.

Mycoviruses are very common in all major fungal groups, and some of them are related to plant viruses or animal viruses. Interestingly, a majority of mycoviruses with +ssRNA genomes are phylogenetically related to plant viruses (42), whereas −ssRNA mycoviruses are evolutionarily closer to animal viruses (43, 44). Since the communication between fungi and plants via complex mutualistic, symbiotic, or antagonistic relationships supplied opportunities for viruses to horizontally migrate in both directions, it is logical that a phylogenetically close relationship between plant viruses and mycoviruses with +ssRNA genomes exists. However, the evolutionary linkage between mycoviruses and animal viruses has not been reasonably explained. In the present study, we found that, based on their L proteins, the three fungal rhabdoviruses are phylogenetically positioned into animal virus groups and closely related sripuviruses, providing a novel incidence of −ssRNA mycoviruses being phylogenetically related to animal viruses. Previous reports suggested that the genus Sripuvirus comprises arthropod-transmitted rhabdoviruses (45, 46). The G proteins of vertebrate host rhabdoviruses contain 12 highly conserved cysteine residues (CI to CXII) (47), but G proteins in some insect viruses (such as sigmaviruses) lack two conserved cysteine residues (CVI and CVII) (20), which is similar to what is observed in RsRhV1 and RsRhV2. Therefore, we propose that fungal rhabdoviruses possibly play an intermediate role to undertake communication with viruses infecting invertebrates and vertebrates during the evolution of life history.

We found that RsRhV1 and RsRhV2 could be related to virulence changes in R. solani. Since a few rhabdoviruses are responsible for causing significant agricultural losses and pose a tremendous threat to public health (48–50), rhabdoviruses have garnered considerable scientific attention. They have become important model systems for studying molecular virology (51, 52). Rhabdoviruses were recently discovered in fungi, and fungus-rhabdovirus interactions could offer a novel platform to understand the biological features of rhabdoviruses, including pathogenic factors and immune escape mechanisms. Our group recently discovered a rhabdovirus, SsRhV1, in S. sclerotiorum but did not clearly define whether SsRhV1 was related to virulence (53). In the present study, we found that infection by RsRhV1 and/or RsRhV2 could enhance virulence in R. solani isolates, and we confirmed this conclusion via virus elimination and virus horizontal transmission experiments, which is the first instance linking fungal rhabdoviruses to virulence. However, the evidence that RsRhV1 and RsRhV2 enhance host pathogenicity is limited so far. To obtain direct evidence, we also attempted to introduce virions into the rhabdovirus-free strains, but we failed to induce infection due to lower virion concentrations or other unknown reasons. In addition, we attempted to transfer rhabdoviruses from strains XY175, 175A1VF15, and 175A1VF32 to strain 175A1VF13; however, we were confused that rhabdoviruses could not be successfully transferred even though those strains have the same genetic background. A possible explanation is that the other four mycoviruses (RsEV9, RsEV10, RsMV40, and RsMV41) could potentially interfere with rhabdovirus replication via unknown mechanisms, since the replication levels of all mycoviruses, especially two endornaviruses (RsEV9 and RsEV10), in individual isolates were the highest in the hypovirulent strain 175A1VF13. Previously reported MyRV1 or CHV1-Δp69 (lacking the RNA silencing suppressor) could highly activate RNA silencing to interfere with the replication of Rosellinia necatrix victorivirus 1 via transcriptional induction of Dicer expression (54). In the rhabdovirus-free strain XYA1VF13, the abundance of two endornaviruses significantly increased, which could be a potential reason to inhibit rhabdovirus infections; however, whether the antiviral RNA silencing mechanism involves the interaction among viruses infecting strain XY175 needs to be further explored. In addition, RsEV9 and RsEV10 might be potential hypovirulence factors due to their higher replication levels in the hypovirulent strain. However, we lack any direct evidence to support this assumption, because endornaviruses and/or mitoviruses failed to be transferred or transfected to other strains (e.g., HG81R3); in addition, they are stable in strain XY175 and cannot be eliminated by single-protoplast isolation or other strategies.

Although high-throughput sequencing has been used to confirm that −ssRNA mycoviruses widely exist in fungal populations (9, 55), only three fungal rhabdoviruses, RsRhV1, RsRhV2, and SsRhV1, were discovered and molecularly characterized with their complete genomes, suggesting that rhabdoviruses might have low abundance in the fungal population. Moreover, we found that it is difficult for RsRhV1, RsRhV2, and SsRhV1 to horizontally transfer between individual strains but it is easy for them to be eliminated via single-protoplast isolation. This feature could be a potential reason explaining rhabdoviruses not being widely distributed in fungi. In summary, we molecularly characterized two rhabdoviruses in R. solani and proposed to establish a new genus to accommodate fungus-infecting rhabdoviruses, including RsRhV1 and RsRhV2 in R. solani and SsRhV1 in S. sclerotiorum. The results of our study will help to expand the current knowledge regarding rhabdovirus classification and provide a model system to study virus-fungus and rhabdovirus-host interactions, which will profit the agricultural and medical areas of human welfare. Further work on the pathology and biology of fungal rhabdoviruses, including their potential transmission vectors, host ranges, and virion morphology, may be required to confirm their classification.

MATERIALS AND METHODS

Fungal strains and culture conditions.

R. solani strain XY175 was initially isolated from diseased (sheath blight) rice obtained from a field in Xiangyang County, Hubei Province, China, and was confirmed to belong to R. solani AG-1IA. Three isolates, 175A1VF13, 175A1VF15, and 175A1VF32, were derived from strain XY175 via single-protoplast isolation. In addition, strain HG81R3 was labeled with a hygromycin phosphotransferase gene. All strains were cultured on PDA plates at 25°C to 28°C and maintained on PDA at 4°C. Mature sclerotia of the individual strains were collected, dried, and stored at −20°C.

RNA extraction and cDNA synthesis.

To extract total RNA, each strain of R. solani was cultured at 28°C for 1 to 2 days on a cellophane membrane overlaying a PDA plate. Mycelium mass (about 1 g) was collected from each strain and ground to fine powder in liquid nitrogen with a mortar and pestle. Total RNA samples were prepared using the TRIzol RNA extraction kit (TaKaRa, Dalian, China) according to the manufacturer's instructions and treated with DNase I. Total RNA was then assessed using a Bioanalyzer 2100 (Agilent Technologies, USA). Total RNA (about 60 μg) of strain XY175 was sent to Shanghai Biotechnology Co. for metatranscriptomic sequencing on the HiSeq 2500 platform (Illumina, USA) based on the rRNA-depleted method. Virus-related sequence assembly and annotation were conducted following previously reported procedures (56).

To verify the presence of mycoviruses in each strain, cDNAs were synthesized using Moloney murine leukemia virus (M-MLV) reverse transcriptase (Promega, USA) and an oligo(dT) primer following the manufacturer’s instructions. Mycovirus-specific primers (see Table S1 in the supplemental material) were used to detect mycovirus content in individual strains of R. solani.

Complete genome cloning and Sanger sequencing.

RNA samples, either from virus particles or from mycelium mass, were used for cDNA cloning and terminal determination as described previously, with minor modifications (57). An anchor primer PC3-T7 loop (5′-p-GGATCCCGGGAATTCGGTAATACGACTCACTATATTTTTATAGTGAGTCGTATTA-OH-3′) was ligated to purified ssRNA using T4 RNA ligase and was used for the RT reaction. The primer PC2 (5′-CCGAATTCCCGGGATCC-3′) was designed based on the corresponding sequence of the PC3-T7 loop, and sequence-specific primers were designed based on the available sequence and proximal region sequences used for the amplification of terminal sequences. The expected PCR products were recovered and purified with a gel extraction kit (Omega Bio-Tek, USA) and cloned into the pMD18-T vector (TaKaRa, Dalian, China) for sequencing. To achieve high-quality consensus sequences and to avoid laboratory PCR artifacts, each nucleotide of full-length cDNA was sequenced at least three times. Primers used in this study are listed in Table S1 in the supplemental material.

To define transcription initiation and termination sites of the five genes in RsRhV1 and RsRhV2, the 5′- and 3′-RACE techniques were applied using previously described methods (58). The primers used for RT-PCR and RACE analyses are listed in Table S2 in the supplemental material.

To define the expression profiles of five genes of RsRhV1 and RsRhV2, the total RNA of strain XY175 was subjected to transcriptome sequencing. SAMtools v1.9 was used to count the number of reads for the mRNA sequences of genes from RsRhV1 and RsRhV1. The reads were mapped to RsRhV1 and RsRhV2 genomes using HISAT2 v2.1.0 (59). The aligned SAM files were sorted into BAM files, and then the idxstats command of SAMtools v1.9 (60) was used to count the number of reads for each mRNA sequence. The expression of these mRNAs in the sequencing data was calculated with the following formula: reads per kilobase of transcript per million mapped reads (RPKM) = total exon reads/(mapped reads × exon length). Three replications of transcriptome sequencing were conducted.

Bioinformatic analyses.

Full-length viral genome sequences were submitted to the ORFfinder on the NCBI website to determine their gene organization and the amino acid sequences of the gene products. The molecular weight and isoelectric point of each virus-encoded protein were calculated using the Compute pI/Mw tool available at ExPASy (https://web.expasy.org/compute_pi). Transmembrane α-helices and signal peptides were predicted using the TMHMM and SignalP programs, respectively (https://services.healthtech.dtu.dk). The potential glycosylation sites were predicted using the NetOGlyc v4.0 server (https://services.healthtech.dtu.dk/service.php?NetOGlyc-4.0) and NetNGlyc v1.0 server (https://services.healthtech.dtu.dk/service.php?NetNGlyc-1.0). The potential phosphorylation sites were determined using the NetPhos v3.1 server (https://services.healthtech.dtu.dk/service.php?NetPhos-3.1). The conserved domains were identified via MotifFinder (https://www.genome.jp/tools/motif). Multiple sequence alignment was computed using the ClustalW and BLAST software packages. Multiple alignments of the conserved domains in the proteins encoded by the viruses were projected using the MUSCLE program in MEGA v7. Phylogenetic analysis was carried out using the linked nucleotide sequences and translated amino acid sequences in the GenBank database, based on high levels of similarity with known viral nucleic acids and proteins. Phylogenetic trees based on the sequences of L, N, and G proteins of RsRhV1 and RsRhV2 were constructed using the ML method, with a bootstrap value of 1,000 replicates, through MEGA v7.0.18 (http://www.megasoftware.net/megamacBeta.php).

Phenotypic characteristics and virulence assay.

To assess the phenotypic characteristics of the different fungal isolates used in this study, agar plugs of fresh mycelia were transferred from colony margins of old cultures onto fresh PDA plates (9 cm in diameter) and incubated at 28°C. The colony diameter of each strain was measured 12 and 48 h postinoculation. Colony morphology was examined daily until mature sclerotia were produced. To evaluate the virulence of the fungal strains, fresh mycelial agar plugs were placed on detached leaves of rice and incubated at 28°C with 90% humidity. The development of diseased lesions on the rice leaves was examined and photographed every 12 h for 4 days. This experiment was conducted with more than three replicates for each treatment, and all biological characterization experiments were performed at least twice. Biological property data were analyzed by one-way analysis of variance using the SAS v8.0 program. Differences with P values of <0.01 were considered statistically significant.

Viral horizontal transmission and detection of mycoviruses by RT-PCR.

The experiment was carried out by using the pairing culture technique as described previously (54, 61). The strain XY175 (donor) containing RsRhV1 and RsRhV2 was cocultured with a vegetatively (somatic) compatible hygromycin-resistant R. solani strain (recipient) on a PDA plate (15 cm in diameter) for 10 days at 28°C. Mycelial agar plugs were picked up from the margin areas of recipient strains (at sites farthest from the mycovirus-infected strains) and then subcultured on fresh PDA (80 μg/mL hygromycin). All derivative isolates were tested for the presence of RsRhV1 and RsRhV2 by RT-PCR with specific primers.

Data availability.

The GenBank accession numbers for the two rhabdovirus genomes reported in this study are MW922511 and MW922512, and the raw sequence reads from the transcriptomic libraries of strain XY175 are available in the NCBI Sequence Read Archive (SRA) database under BioProject accession number PRJNA810759.