The Fusarium oxysporum strain LJ3-3 used in this study was recovered in 2020 from a capsicum root rot sample (Luohe, Henan Province, China). The strain AJ3-8 of F. oxysporum was used as a control. The diseased root was cut into 0.5-cm² samples and soaked for 30 s in 75% ethanol. Then, the samples were rinsed with sterilized water three times and dried on sterilized blotting paper. Finally, the samples were cultured on potato dextrose agar (PDA) medium at 25°C in the dark for 2 days. A small amount of mycelium was scraped off the culture and washed with 2 ml sterile water. Then, the mixture was pooled and filtered through three-layer lens wiping paper. The spore liquid was diluted to 10³ conidia ml^–1, and 100 μl was smeared on a PDA plate and incubated overnight at 25°C. The next day, a single colony was selected and considered a purified strain. Five mycelial agar plugs were inoculated on fresh PDA medium covered with cellophane membranes and cultured at 25°C for 4–5 days. Mycelium in each dish were harvested and stored at –70°C until use. Genomic DNA was extracted from the fungi using the CTAB method. The primers for translation elongation factor 1-alpha (EF-1α), RNA polymerase II subunit I gene (RPB1), and RNA polymerase II subunit II gene (RPB2) were used to confirm the Fusarium species identification (Mishra et al., 2003; O’Donnell et al., 2010). The EF-1α, RPB1, and RPB2 sequences were analyzed by Blast search against data in the Fusarium ID: Cyber-infrastructure for Fusarium database (fusariumdb.org). Mycelial growth and conidial production were evaluated according to the procedures described by Wu et al. (2007). A transformant of F. oxysporum strain B9 was used as a recipient strain in a horizontal transmission test. The B9 strain was isolated from a fusarium root rot sample of tobacco (Xuchang, Henan Province, China, 2020). This B9 strain has a hygromycin-resistance gene (Hygromycin B phosphor-transferase), a normal colony morphology, and high virulence in its hosts. The strain AJ3-8 and strain B9 were confirmed as virus-free strains by high-throughput sequencing and RT-PCR detection (data not shown). All strains were cultured on PDA medium at 25°C and then stored at –70°C in 25% glycerol.

Total RNA of twenty-two F. oxysporum strains were extracted from 1.0 g of mycelium using an RNAiso Plus Kit (TaKaRa, Dalian, China) following the manufacturer’s instructions. Next, total RNA was purified using an RNAClean XP Kit (Cat A63987, Bechman Coulter, Inc., Brea, CA, United States) and RNase-Free DNase set (Cat79254, QIAGEN, GmBH, Hilden, Germany), and rRNA was depleted by a Ribo-ZeroTm rRNA Removal Kit (Illumina, San Diego, CA, United States). Finally, the qualified samples were mixed into one sample and used for pair-end sequencing on an Illumina HiSeq 2500 platform at Shanghai Bohao Biotechnology Co., Ltd. One sequencing library was constructed by the qualified total RNA of F. oxysporum strains. The raw reads were filtered base on default parameters, 1 × 10⁸ bp clean reads were obtained and mapped against genome sequence of F. oxysporum using Bowtie (1.0) software. Then, unmapped reads were assembled de novo using CLC Genomics Workbench (version: 6.0.4) with scaffolding contig algorithm, word-size = 45, and minimum contig length ≥ 200. Consequently, 44,679 contigs were achieved. After Blasted by the non-redundant protein sequences (nr) database in NCBI¹, 12 contigs which represented partial genome segments of “virus” or “viral” were retrieved. Finally, the contig 1028 that was identical or complementary to mymonavirus genomic sequences were extracted and subjected to further analysis.

The cDNA of each F. oxysporum strains were synthesized using a PrimerScript™ 1st Strand cDNA synthesis Kit (TaKaRa, Dalian, China) following the manufacturer’s instructions. The occurrence of putative virus sequence in the F. oxysporum strains included in the RNA-Seq sample was investigated using RT-PCR with specific primers designed based on the assembled contig 1028. The RT-PCR results suggested that contig 1028 is the sole viral agent of the LJ3-3 strain (data not shown). The 5′- and 3′-terminal sequences of the FoMyV1 genome were completed using a SMARTer RACE Amplification Kit (Clontech, Mountain View, CA, United States) following the manufacturer’s instructions using gene-specific primers (GSPs). GSP-1028F1 (GAGCAAGAACATAGATTCACCT) and GSP-1028F2 (TGGTTGTGGAGAAATGGGGCTGGTA) were used as the inner and outer primers, respectively, for 3′-RACE. GSP-1028R1 (5′-CTGGCTGGTTTGGTAGGG-3′) and GSP-1028R2 (5′-CTTCGTCGTCTGCCCAAT-3′) were used as the inner and outer primers, respectively, for 5′-RACE. Meanwhile, seven pairs of specific primers were designed to verify the almost full length of FoMyV1 by RT-PCR (Supplementary Table 1 and Supplementary Figure 1). All RT-PCR products of expected size were purified and cloned into the pMD19-T vector (TaKaRa, Dalian, China) and introduced into Escherichia coli Trelief 5α (TSINGKE Biotech, Zhengzhou, China) by transformation. At least three recombinant clones were sent to TSINGKE Biotech for sequencing. In addition, the DNA of strain LJ3-3 was used as a template to examine whether the viral sequence was integrated into the host. Moreover, 143 F. oxysporum strains from five counties or cities in Henan province of China were used to test for the presence of FoMyV1.

The putative open reading frames (ORFs) of FoMyV1 were deduced using the ORF Finder program on the website of the National Center for Biotechnology Information (NCBI).² Homologous sequences were searched for full-length cDNA sequences and deduced polypeptides of FoMyV1 in the NCBI database using BlastN and BlastP, respectively. A search for the predicted domains present in the polypeptide sequence was conducted using the Conserved Domain Database (CDD).³ Multiple sequence alignments of the RdRp sequences were performed using DNAMAN (Version 9) and ClustalX (Version 2.0) (Thompson et al., 1997). A phylogenetic tree was constructed using the maximum-likelihood (ML) method in MEGA-X (Version 10.1.8) with 1,000 bootstrap replicates (Kumar et al., 2018).

To investigate the vertical transmissibility of FoMyV1, 40 single-conidium isolates were obtained from parental strain LJ3-3. Then, the presence of FoMyV1 was determined using RT-PCR with the primer pair ct1028RT-F2/R2 (Supplementary Table 1), which was designed to amplify a 746-bp product.

The pairing-culture technique (Wu et al., 2007; Zhang and Nuss, 2008) was used to investigate the horizontal transmissibility of FoMyV1 between F. oxysporum strains. In the contact culture in each plate (9 cm in diameter), the strain LJ3-3 served as the donor, whereas the strain B9 (a hygromycin-resistance-gene transformant of F. oxysporum) served as the recipient. The mycelial agar plugs of two strains were cultured at a distance of 2 cm in PDA medium. After incubation of the contact cultures at 25°C for 8 days, three mycelial derivative isolates were obtained from three colonies of the recipient strain in the contact cultures. Then, derivative isolates were cultured in hygromycin-resistance (50 mg/ml) PDA medium three times. Finally, the primer pair ct1028RTF7/F7 with a 1,024-bp amplicon was used to verify the presence of FoMyV1 in the derivative isolates. Two derivative isolates of B9 (B9-VI) infected by FoMyV1 both contained the mycovirus FoMyV1 (Supplementary Figure 2). Compared with strain B9, the colony of B9-VI was irregular and the aerial hyphae were rare.

To assess the effects of FoMyV1 on its host biological characteristics and plant pathogenicity, two isogenic strains B9-VI (virus-infected) and B9 (virus-free) were used. Each strain was individually tested for mycelial growth rate (PDA, 25°C) and conidium production. Five mycelial plugs were inoculated into 100 ml carboxymethylcellulose sodium (CMC) fluid medium and cultured for 4 days (28°C, 180 rpm). Then, the mycelium solution was filtered through two layers of sterile gauze and the precipitate was resuspended with 50 ml sterile water. Finally, the concentration of conidium was counted using a blood counting plate and the conidium production was calculated. Tobacco cultivar ‘Zhongyan 100’ and Nicotiana benthamiana were grown to the third or fourth leaf stage and then individuals with the same growth status were selected. The roots were injured and then tobacco seedlings were transplanted into new pots (9 cm × 7 cm × 6 cm, top width × bottom width × height), inoculated with 30 ml spore suspension (1 × 10⁷ml^–1), and cultured at 25°C under fluorescent light (16 h light/8 h dark). One month later, investigate the disease incidence, severity, index, and phenotypic values of plants. Disease incidence was defined as the percentage of infected plants, and disease severity was rated on a scale of 0–9 as follows: level 0, no symptoms; level 1, the plant growth is basically normal or slightly dwarfing, a few roots are necrotic and dark brown, middle and lower leaves are chlorosis or discoloration; level 3, the disease plants are 1/4–1/3 lower than the healthy ones, half of the roots are necrotic and black, 1/2–2/3 of the leaves are wilting; level 5, the disease plants are 1/3–1/2 lower than the healthy ones, most of the roots are necrotic and black, more than 2/3 of the leaves are wilting, tip and margin of the middle and lower leaf are slightly withered; level 7, the disease plants are more than 1/2 lower than the healthy ones, all of the leaves are wilting, all of the roots are necrotic and black, and the secondary roots near the surface were obviously damaged; level 9, plant is dead. Pathogen was re-isolated from seedlings inoculated with strain B9-VI and detected to carry fungal virus FoMyV1. The assay treatments were repeated three times with three seedlings each. A one-way factorial analysis of variance (ANOVA) (SAS Institute, Cary, NC, United States, Version 8.0, 1999) was used to determine the differences in growth rate, conidial production, phenotypic values, and disease index of each strain.

Based on RT-PCR and EF-1α, RPB1, and RPB2 sequencing, we identified strain LJ3-3 as F. oxysporum (Supplementary Table 3). The sample also contained the (–)ssRNA mycovirus FoMyV1 (Supplementary Figure 2). We cultured F. oxysporum strain LJ3-3 at 25°C on PDA for 10 days to observe its morphology (Figure 1A). The average radial mycelial growth of LJ3-3 was 8.75 mm/day, which was significantly (p < 0.05) slower than that of the virus-free strain AJ3-8 (12.00 mm/day) (Figure 1B). The average conidial production of LJ3-3 was 6.45 × 10⁷ml^–1, which was also significantly (p < 0.05) lower than that of strain AJ3-8 (21.50 × 10⁷ml^–1; Figure 1C). In the virulence assay using Zhongyan 100 leaves, the average lesion diameter (7 mm) caused by strain LJ3-3 was significantly (p < 0.05) smaller than that caused by strain AJ3-8 (14 mm; Figures 1D,E). For comparation, strain LJ3-3 infected with FoMyV1 showed slow growth rate, low conidial production, and weak virulent.

The complete genome sequence of FoMyV1 (GenBank accession no. OM049502) was 10,114 nt, with a GC% content of 47%, possessing five non-overlapping ORFs (ORF1–5) and two untranslated regions (UTR) of 129 and 291 nt at the 5′- and 3′-termini, respectively (Figure 2A). ORF1–ORF5 encode proteins 263, 412, 190, 193, and 1,952 amino acids (aa) in length, located in the reading frames +1, +1, +1, +2, and + 2, respectively. The conserved motif search showed that ORF5 contained a Mononegavirales RNA-dependent RNA polymerase domain (pfam00946; aa location 142–980; e-value 1.61e-112), Mononegavirales mRNA-capping region V (pfam14318; aa location 1,077–1,231; e-value 3.43e-12), and paramyxovirus_RNAcap (TIGR04198; aa location 1,114–1,462; e-value 3.82e-06; Figure 2A). We did not find any conserved domains in the other four ORFs. BlastP analysis showed that the putative protein L encoded by ORF5 of FoMyV1 was similar to the RdRp of Hubei rhabdo-like virus 4 (HbRLV4) with 65% identity (Shi et al., 2016). In addition, the putative protein L also showed 26–36% identity with the RdRp encoded by other mymonaviruses in the family Mymonaviridae (Supplementary Table 3). The proteins encoded by ORF1 and ORF2 were similar to the hypothetical proteins 1 and 2 of HbRLV4, with 43 and 67% identity, respectively (Table 1). However, the putative protein encoded by ORF3 and ORF4 of FoMyV1 was not significantly similar to any other protein in the search to characterize. In addition, the semi-conserved AU-rich sequences are finds in the putative untranslated sequences between ORFs in the FoMyV1 genome (Figure 2B). The putative gene-junction sequence of ORF1/2 and ORF2/3 (viral RNA strand, 3′-UAAAUUGUUUUG-5′) was identical to those of HbRLV4. We also found several complementary nucleotides near the end of the FoMyV1 genome sequence (Figure 2C). We identified four conserved motifs (I–IV) from Mononegavirales in protein L encoded by FoMyV1 ORF5 (Figure 2D). Meanwhile, FoMyV1 was not detected in the DNA template of LJ3-3 strain (Supplementary Figure 3). In summary, FoMyV1 genome structural characteristics were consistent with the typical characteristics of members in order Mononegavirales, and belong to the family Mymonaviridae.

To examine the relationship between FoMyV1 and other mymonaviruses (Supplementary Table 4), we performed a maximum-likelihood phylogenetic analysis based on the amino acid sequences of the L protein of FoMyV1 and 38 other (–)ssRNA viruses, including representative members of six families in order Mononegavirales (Nyamiviridae, Bornaviridae, Rhabdoriridae, Paramyxoviridae, Filoviridae, and Pneumoviridae), and representative strains of nine genus in family Mymonaviridae. FoMyV1 clustered with HbRLV4 and H2BulkLitter 1223 virus (Starr et al., 2019) to form a distinct clade with a bootstrap support value of 100%, indicating a close evolutionary relationship. These three viruses belong to the genus Hubramonavirus and clustered with 22 other mymonaviruses, forming a large independent clade of family Mymonaviridae (Figure 3). The other 17 (–)ssRNA viruses also formed corresponding viral family clades. These results confirm that FoMyV1 is a novel member of the genus Hubramonavirus, family Mymonaviridae.

We used F. oxysporum strain B9 as a recipient for horizontal transmission of FoMyV1. We obtained one mycelial derivative isolate—B9-VI—from one B9 recipient colony in the two contact cultures of LJ3-3/B9 (Figure 4A). The average growth rate of B9-VI was 8.29 mm/day, which was significantly slower than that of B9 (10.00 mm/day; Figure 4B). Similarly, the conidium production of B9-VI was 3.35 × 10⁷ ml^–1, significantly lower than that of B9 (6.76 × 10⁷ ml^–1; Figure 4C). Furthermore, RT-PCR indicated that FoMyV1 was successfully transmitted from LJ3-3 to the virus-free strain B9Hyg^R (Supplementary Figure 2). In summary, colony morphology, growth rate, and conidium production were significantly affected by the FoMyV1 infection (Figure 4).

To determine the frequency of FoMyV1 transmission by conidium progeny in the laboratory, we obtained 40 single-conidium isolates from parent strain LJ3-3 and tested for FoMyV1 using RT-PCR. All the single-conidium isolates were FoMyV1 positive (Supplementary Figure 4). The average growth rate of the isolates was 6.6–9.6 mm/day in PDA plate. There were no significant differences in growth rate among the 40 conidium progeny (p < 0.05).

To study the effect of FoMyV1 on the virulence of its fungal host, we evaluated the pathogenicity of two isogenic strains, B9-VI (virus-infected) and B9 (virus-free), in two different tobacco cultivars. The disease index of B9-VI and B9 in tobacco cultivar ‘Zhongyan 100’ was 35.80 and 38.27, respectively, and those in N. benthamiana were 8.64 and 9.87, respectively (Supplementary Table 5). There were no significant differences in disease index between the two strains on either cultivar (p < 0.05). Furthermore, there were no significant differences (p < 0.05) in plant height, fresh weight, or root length between plants inoculated with the two strains in either plant cultivar (Figure 5). At the same time, we re-isolated the pathogen from the root of B9-VI infected plants for FoMyV1 detection. The two re-isolated strains were detected as FoMyV1 positive (Supplementary Figure 5). These results suggest that FoMyV1 does not alter the pathogenicity of its host.

To investigate the incidence of FoMyV1 in Henan Province, 143 F. oxysporum strains (Supplementary Table 6) were tested for the presence of FoMyV1 using RT-PCR with primer pair ct1028RT-F7/R7 (Supplementary Table 1). The result showed that only two strains harbored FoMyV1 in the tested 143 F. oxysporum strains (Supplementary Figure 2), and the incidence of FoMyV1 was only 1%. The FoMyV1 infected strain LJ3-3 and LJ4-1 were both collected from same location (Xuchang city, Henan Province, China).