Novel Mycoviruses Discovered from a Metatranscriptomics Survey of the Phytopathogenic Alternaria Fungus

Alternaria fungus can cause notable diseases in cereals, ornamental plants, vegetables, and fruits around the world. To date, an increasing number of mycoviruses have been accurately and successfully identified in this fungus. In this study, we discovered mycoviruses from 78 strains in 6 species of the genus Alternaria, which were collected from 10 pear production areas using high-throughput sequencing technology. Using the total RNA-seq, we detected the RNA-dependent RNA polymerase of 19 potential viruses and the coat protein of two potential viruses. We successfully confirmed these viruses using reverse transcription polymerase chain reaction with RNA as the template. We identified 12 mycoviruses that were positive-sense single-stranded RNA (+ssRNA) viruses, 5 double-strand RNA (dsRNA) viruses, and 4 negative single-stranded RNA (−ssRNA) viruses. In these viruses, five +ssRNA and four −ssRNA viruses were novel mycoviruses classified into diverse the families Botourmiaviridae, Deltaflexivirus, Mymonaviridea, and Discoviridae. We identified a novel −ssRNA mycovirus isolated from an A. tenuissima strain HB-15 as Alternaria tenuissima negative-stranded RNA virus 2 (AtNSRV2). Additionally, we characterized a novel +ssRNA mycovirus isolated from an A. tenuissima strain SC-8 as Alternaria tenuissima deltaflexivirus 1 (AtDFV1). According to phylogenetic and sequence analyses, we determined that AtNSRV2 was related to the viruses of the genus Sclerotimonavirus in the family Mymonaviridae. We also found that AtDFV1 was related to the virus family Deltaflexivirus. This study is the first to use total RNA sequencing to characterize viruses in Alternaria spp. These results expand the number of Alternaria viruses and demonstrate the diversity of these mycoviruses.

In recent years, metatranscriptomics has been widely used in virus discovery. Many novel viruses have been discovered in fungi, which has significantly promoted the progress of viromics research and has sped up the discovery and understanding of unknown viruses [13,44,[50][51][52][53][54][55][56][57][58][59]. In addition, transcriptomes data of fungi provided evidence of the existence of negative-sense RNA viruses in fungi before the first negative-sense RNA mycovirus was isolated [59].
In this study, we followed a metatranscriptomics approach to identify the mycovirus communities of six species of the genus Alternaria, which are associated with pear spot disease in China. We also identified near-full-length sequences of putative mycoviruses. We isolated a novel −ssRNA mycovirus from an A. tenuissima strain HB-15, which we designated as Alternaria tenuissima negative-stranded RNA virus 2 (AtNSRV2). We also characterized a novel +ssRNA mycovirus from an A. tenuissima strain SC-8, which we designated as Alternaria tenuissima deltaflexivirus 1 (AtDFV1). Through phylogenetic and sequence analyses, we found AtNSRV2 to be related to the viruses of the genus Sclerotimonavirus in the family Mymonaviridae. We also found that AtDFV1 is related to the family Deltaflexivirus. In this study, we identified viruses in Alternaria spp., for the first time, using total RNA sequencing. These results significantly expanded the number of Alternaria viruses and revealed the abundant diversity of the mycoviruses.

RNA Extraction and Sample Preparation for High-Throughput Sequencing
We prepared total RNA for high-throughput sRNA sequencing. We cultured 78 strains on cellophane membranes that overloaded PDA plates at 28 • C for 7 days. By mixing the mycelia in equal proportion, we ground the samples using liquid nitrogen. We used a TRIzol RNA extraction kit (Thermo Fisher, Waltham, MA, USA) to extract the total RNA, which we then clarified with chloroform in 2 mL tubes. We used ethanol precipitation to obtain the total nucleic acid fractions. We used 75% alcohol to wash the sample two times, and then dissolved them in water treated with diethylpyrocarbonate (DEPC). The total RNA was detected by agarose gel electrophoresis with 1.2% (w/v). Then, we determined the concentration and selected qualified samples for high-throughput sequencing.

High-Throughput Sequencing and Data Analysis
We sent the total RNA of the mixed strains to the Beijing Biomarker Technologies Company (Beijing, China), where it was constructed and sequenced for cDNA library. First, rRNA was removed, and then double-stranded cDNAs were synthesized using random hexamers (N6). The cDNA ends were repaired, and the A-tail was added and sequenced. The final library was obtained using the polymerase chain reaction (PCR). Then, the quality of the library was tested and sequenced using an Illumina HiSeq XTen platform. A certain proportion of low-quality data is inevitable when obtaining raw data through sequencing. To ensure the reliability and accuracy of the analysis results, we had to preprocess the raw data, which were strictly controlled, and filtered as follows: we first removed the reads using adaptors, and then removed the low-quality reads for clean data. We analyzed the potential sequences of mycoviruses according to the splicing and assembly of clean data.

Validation of Virus-like Contigs by RT-PCR and Viral Sequencing Amplification
We isolated the total RNA of 78 strains and used reverse transcription polymerase chain reaction (RT-PCR) to investigate viruses in the fungal strains using specific primers, which were designed based on the assembled contigs (Supplementary Table S2). For cDNA synthesis, we added 1 µL random hexamers (N6), 2 µL DEPC-treated water, and 7 µL total RNA sample extracted from each tested strain to a 500 µL tube (RNase-free). After 8 min in boiling water bath, the sample was placed on ice for 3-5 min, the following system was added: 4 µL of DEPC-treated water, 4 µL of 5×M-MLV reverse transcription buffer, 1 µL of 2.5 mM dNTP, 0.5 µL of RNase inhibitor (TaKaRa, Dalian, China), 0.5 µL of M-MLV reverse transcriptase (PROMEGA, Madison, WI, USA). We mixed and centrifuged the sample, and then reverse transcribed the mixture for 1-2 h at 37 • C in a 20 µL reaction mixture. After reaction, obtained products were used for PCR or stored at −20 • C. The PCR products were electrophoresed in a 1.2% agarose gel, and stained with ethidium bromide (EB, 0.1 µg/Ml) for visualization on a UV trans-illuminator. We used a Pmd18-T vector (TaKaRa, Dalian, China) to ligate the purified PCR product, which was transformed into competent cells of Escherichia coli DH5α. We conduct the sequencing at Sangon Biotech Co. (Shanghai, China).
We used RT-PCR to obtain full-length sequences of the viruses. We determined virus sequences using primers designed from the data obtained from the assembled contigs (Supplementary Table S3). We also amplificated the cDNA ends sequence of virus using a SMARTer RACE 5 /3 kit (TaKaRa, Dalian, China) according to the manufacturer's instructions. We ligated the purified PCR product into a pMD18-T vector, which was transformed into E. coli DH5α. The sequencing was completed at Sangon Biotech Co. We determined three or more independent clones for each product in both orientations.

Sequences Alignment and Phylogenetic Analysis
We calculated sequence similarities using BLAST program on the NCBI database (https://www.ncbi.nlm.nih.gov (accessed on 22 October 2022)). According to the alignment information, the viruses related information (e.g., nucleic acid type, virus species, and the number of open reading frames [ORFs]) was preliminarily confirmed. We used DNAMAN (Lynnon Corporation, Pointe-Claire, Quebec, Canada) to conduct the sequence assembly. We used the ORF Finder website (http://www.ncbi.nlm.nih.gov/gorf/gorf.html (accessed on 22 October 2022)) to predict ORFs. We searched the CDD database (http://www.ncbi. nlm.nih.gov/Structure/cdd/wrpsb. cgi (accessed on 22 October 2022)) and the Pfam database (http://pfam.xfam.org/ (accessed on 22 October 2022)) to predict the conserved domains in sequences. We used MEGA 10 and applied the maximum likelihood method to construct the phylogenetic trees, which we tested with 1000 bootstrap replicates. We used MAFFT (http://www.ebi.ac.uk/Tools/msa/maft/ (accessed on 22 October 2022)) to conduct multiple sequence alignments of proteins encoded by the contigs and the reference mycovirus.

Diversity of Alternaria Viruses
Reportedly, several phytopathogenic fungi have fungal viruses. We identified the presence of mycoviruses in the fungi of Alternaria and generated an RNA sequencing library with 78 strains ( Table 1). The library was sequenced and assembled at a considerable depth. After removing rRNA from the assembled data, we obtained 67,011,714 reads by fragment reverse transcription, and screened out 66,709,434 clean reads by quality control data. Overall, a total of 131,203 contigs with a total length of 82,578,727 bp were yielded by transcriptional splicing, database alignment, and coding region prediction. The N50 and N90 values were 812 and 291, respectively. The results showed that the quality of the spliced sequences was good. We obtained 26 contigs that were derived from mycoviruses from this sequence ( Table 2).  We subjected the contigs to BLAST analysis and assigned the contigs to 21 putative novel mycoviruses. These mycoviruses were characterized by eight different viral families, including Botourmiaviridae, Botybirnaviridae, Chrysoviridae, Deltaflexiviridae, Hypoviridae, Partitiviridae, Mybuviridea, Mymonaviridae, and Narnaviridae. Among them, 12 viruses belonged to +ssRNA viruses, 5 viruses belonged to dsRNA viruses, and 4 viruses belonged to −ssRNA viruses (Table 2).

Detection and Validation of Alternaria Viruses by RT-PCR
According to the sequence of contigs obtained by high-throughput sequencing, specific primers were designed for RT-PCR detection. The results showed that 21 viruses could be detected in 22 strains of Alternaria spp. (Figure 1), including 16 strains of A. tenuissima, 3 strains of A. alternata, 1 strain each of A. arborescens, A. gossypina, and A. gaisen. These results indicated that the putative viral sequences were reliable. Strain G-21-2 of A. tenuissima and KEL-4-4 of A. arborescens were infected by four viruses (Figure 1). Strain HB-15 was infected by three viruses. Strains G-9 and AH-25 of A. tenuissima were infected by two viruses (Figure 1). Others harbored only one virus.

Positive-Sense Single-Stranded RNA Virus
We used the obtained contigs and identified 12 positive-sense single-stranded RNA viruses. These sequences could be classified into four families, including Hypoviridae, Narnaviridae, Deltaflexiviridae, and Botourmiaviridae ( Table 2).
The family Mitoviridae had only one genome and encoded one ORF, including genera Duamitovirus, Kvaramitovirus, Triamitovirus, and Unuamitovirus. The contig 5012 (2170 nt) had a large ORF (72-2048 nt) that encoded a 658 aa protein. We found that a predicted amino acid sequence of the protein best resembled the polyprotein of Alternaria arborescens mitovirus 1 (AaMV1, GenBank: YP_009270635), which had a 91.04% homology. Therefore, this virus was a strain of AaMV1 that belonged to the genus Duamitovirus in the family Mitoviridae (Table 2). We detected this virus in A. tenuissima strains G-9 and G-21-2 ( Figure 1) and called it Alternaria tenuissima mitovirus 1 (AtMV1).
The family Narnaviridae had only one genome and encoded one ORF, including one genus Narnavirus. The contig 5919 (2004 nt) had a 98.04% homology with the RdRp of Neofusicoccum parvum narnavirus 2 (NpNV2, GenBank: QDB74995), which meant this virus was a strain of NpNV2 and belonged to the genus Narnavirus. The results of RT-PCR revealed that A. tenuissima strain AH-29 harbored this virus and was called Alternaria tenuissima narnavirus 1 (AtNV1) (Figure 1).
We detected this virus in A. tenuissima strain AH-25 ( Figure 1) and called it Alternaria 283 tenuissima ourmia-like virus 5 (AtOLV5). AtOLV5 was also detected in A. gaisen strain 284 AH-20 ( Figure 1). According to the BLASTx results, contig 19628 (946 nt) and contig 49207 285 (482 nt) best resembled the RdRp of Colletotrichum fructicola ourmia-like virus 2 286 (CfOLV2, GenBank: UOV22974) at 89.04% and 86.44% similarity, respectively. We de-287 tected this virus in A. tenuissima strains SC-12 and HB-15 ( Figure 1). Therefore, we called 288 it Alternaria tenuissima ourmia-like virus 6 (AtOLV6) and AtOLV6 might be a novel vi-289 rus. Unfortunately, not constructing phylogenetic tree for lacking the GDD motif in the 290 amino acid sequence of RdRp of AtOLV6 encoded by contig 19628. 291 We analyzed the relationships among three novel viruses and other mycoviruses in 292 the family Botourmiaviridae. Then, we constructed a phylogenetic tree based on the RdRp 293 amino acid sequence of AarOLV1, AalOLV1, AtOLV4, and other related viral sequences, 294 such as ourmiaviruses, sclerouliviruses, magouliviruses, botouliviruses, and mitoviruses. 295 However, this was not used in phylogenetic analysis due to the lack of GDD motif in the 296 amino acid sequence of RdRp of AtOLV6 encoded by contig 19628. In the obtained phy-297 logenetic tree, AalOLV1 was grouped with some botoliviruses (Figure 2). AtOLV4 and 298 some betabotoliviruses were grouped in a branch (Figure 2). AarOLV1 was clustered with 299 some deltasclerouliviruses in a group (Figure 2). As a result, we identified AalOLV1 and 300 AtOLV4 as the new members of the recent genus Botolivirus and Betabotolivirus, and Aa-301 rOLV1 as a new member of the genus Deltasclerouliviruses within the family Botourmiaviri-302 dae. 303 304 Figure 2. Phylogenetic analysis of AarOLV1, AalOLV1, AtOLV4. The novel viruses obtained by high 305 throughput sequencing were indicated by a red circle (•). The data coverage percentages were 306 shown by the numbers on the left of branches. We constructed a phylogenetic tree using the 307 Figure 2. Phylogenetic analysis of AarOLV1, AalOLV1, AtOLV4. The novel viruses obtained by high throughput sequencing were indicated by a red circle (•). The data coverage percentages were shown by the numbers on the left of branches. We constructed a phylogenetic tree using the maximum likelihood method. We based the 1000 bootstrap replications on the best-fit protein evolution (LG+G+I+F) model. We set the gamma value at 2.

Phylogenetic Analysis of AtDFV1
We constructed a maximum-likelihood phylogenetic tree using the entire replicase of AtDFV1 and members of Alphaflexiviridae, Deltaflexiviridae, and Gammaflexiviridae. We verified the existence of motifs I-VI in AtDFV1 and members of the deltaflexivirus family due to the amino acid alignment with the predicted RdRp ( Figure 3D). According to this phylogenetic analysis, AtDFV1, AaDFV1, EnDFV1, and TpDFV1 were related to viruses in the family Deltaflexiviridae (Figure 4).

Double Strand RNA Virus
In this study, five mycoviruses belonged to dsRNA viruses, of which four viruses could be classified into the families Chrysoviridae, Partitiviridae, and the genus Botybirnavirus, and one had no classification status.
According to the BLASTx results, four contigs were homologous at 96.90-100.00% with the corresponding proteins encoded by Alternaria alternata chrysovirus 1 (AaCV1), which was a betachrysovirus in the family Chrysoviridae ( Table 2). The BLASTx search also revealed that contig 1410 (3617 bp) had a 99.37% identity with the RdRp of AaCV1 (GenBank: QJW39304). Contig 2756 (2814 bp) shared a 100.00% similarity with the CP of AaCV1. Therefore, these contigs might belong to a new strain of the AaCV1. As shown in Figure 1, we detected the virus using the special primers based on contig 2756 in A. tenuissima strain SC-8 and called it as Alternaria tenuissima chrysovirus 1 (AtCV1).
The BLASTn results revealed that the contig 10,828 (1417 bp) and contig 13,903 (1210 bp) were notable for regions that had a very strong similarity to the sequences of Alternaria longipes dsRNA virus 1 (AlRV1; GenBank: KJ817371, 91; 88% and 90.58% identity). According to BLASTx, contig 10,828 best resembled the hypothetical protein of AlRV1 with 97.96% homology and contig 13,903 best resembled the RdRp of AlRV1 with 97.84% homology (Table 2). Therefore, the virus was a strain of AlRV1. This virus was detected in A. arborescens strain KEL-4-4 and called it Alternaria arborescens dsRNA virus 1 (AaRV1).
Therefore, we determined that dsRNA viruses in the pooled RNA-Seq sample were not new mycoviruses.

Negative-Sense Single-Stranded RNA Viruses
Based on their RdRp amino acid sequences, we identified four negative-stranded RNA viral sequences in our samples. We assigned these viruses to two taxonomical groups of negative-strand viruses: two were assigned to the family Mymonaviridea, and two were assigned to the family Discoviridae ( Table 2).
We found that contig 35 ( According to BLASTp, this protein had low homology with the RdRp of Coniothyrium diplodiella negative-stranded RNA virus 1 (CdNSRV1) with 68.79% similarity. As shown in Figure 1, we detected this virus in A. arborescens strain KEL-4-4 and called the novel virus Alternaria arborescens negative-stranded RNA virus 1 (AaNSRV1).

Characterization of the Virus AtNSRV2 Genome
We determined the complete cDNA sequence of AtNSRV2 using sequence verification and terminal cloning. The full-length sequence of AtNSRV2 was 9067 nt. The GC content of the whole genome was 53.3%, the 5 and 3 -untranslated regions (UTRs) were 300 nt and 85 nt long, respectively. We predicted that the full length of the AtNSRV2 had five major ORFs (ORFs I-V). These non-overlapping ORFs were arranged in a line along the viral genome ( Figure 5A). We also identified the conserved noncoding sequences (3 -AUUU/AAAUAAAACUUAGGA-5 ), which were downstream the ORFs ( Figure 5B). Because the gene-junction sequences were ubiquitous in the viral genome, we determined that they were a characteristic feature of the mononegaviruses. We deposited the nucleotide sequence of AtNSRV2 in GenBank (accession number: OP566533).
We found that ORF I encoded a protein with 253 amino acid residues and had a mass of 29 kDa. According to BLASTp, the protein had 33.62% homology with a hypothetical protein of Botrytis cinerea negative-stranded RNA virus 4 (BcNSRV4) (GenBank: QJW39407). We also obtained three motifs from the hypothetical protein encoded ORF I, including Family of unknown function (DUF5798; position 118 to 179 aa; pfam19111, E-value = 0.0032), LRRC37A/B like protein 1 C-terminal domain (LRRC37AB_C; position 198 to 251 aa; pfam14914, E-value = 0.073), and LMBR1-like membrane protein (LMBR1; position 100 to 230 aa; pfam04791, E-value = 0.098). We found that ORF II encoded a protein of 402 amino acid residues and had a mass of 44 kDa. According to BLASTp, it was 47.46% similar to the N protein (nucleoprotein) of the SsNSRV-1 (GenBank: YP_009094314). We discovered that ORF III encoded a protein of 52 amino acid residues and had a mass of 6 kDa. We did not find any significant similarity protein sequences in the BLASTp search. We found that ORF IV encoded the largest protein of 1931 amino acids in length and had a molecular mass of 220 kDa. According to BLASTp, the L protein was 56.48% similar to the RdRp of CpSMV1 (GenBank: QMP84020). The P4 protein of AtNSRV2 also had high homology with the RdRp of other −ssRNA mycoviruses (see Supplementary Table S6). The conserved domain predicted that the protein contained a mononegavirales mRNA-capping region V domain (Mononeg_mRNAcap; position 1127 to 1272 aa; pfam14318, E-value = 1 × 10 −6 ) and a mononegavirales RdRp domain (Mononeg_RNA_pol; position 21 to 1041 aa; pfam00946, E-value = 3.3 × 10 −110 ). Based on the multiple alignment of the sequences of the RdRp amino acid of AtNSRV2 and other related viruses from Sclerotimonavirus, we identified four conserved motifs (I-IV) ( Figure 5C). We found that ORF V encoded a protein of 186 amino acid residues and had a mass of 20 kDa. According to BLASTp, the protein best resembled ORF4 of Plasmopara viticola lesion associated mononega virus 2 (PvLAMV2), with a 41.67% similarity (GenBank: QHD64788). The conserved domain predicted that the protein contained a Mannosidase Ig/CBM-like domain (Mannosidase_ig; position 61 to 102 aa; pfam17786, E-value = 0.16).
According to the motif scan results, this protein contained a conservative Bunyavirus 453 RNA-dependent RNA polymerase domain from 545 aa to 1,224 aa. In addition, we de-454 tected a Saccharopine dehydrogenase C-terminal domain from amino acids 1,938 to 2,012 455 (E-value = 0.85) in the hypothetical protein. According to BLASTp, this protein had low 456 homology with the RdRp of Coniothyrium diplodiella negative-stranded RNA virus 1 457 (CdNSRV1) with 68.79% similarity. As shown in Figure 1, we detected this virus in A. 458 arborescens strain KEL-4-4 and called the novel virus Alternaria arborescens negative-459 stranded RNA virus 1 (AaNSRV1). 460 3.5.1. Characterization of the Virus AtNSRV2 Genome 461 We determined the complete cDNA sequence of AtNSRV2 using sequence verifica-462 tion and terminal cloning. The full-length sequence of AtNSRV2 was 9,067 nt. The GC 463 content of the whole genome was 53.3%, the 5′ and 3′-untranslated regions (UTRs) were 464 300 nt and 85 nt long, respectively. We predicted that the full length of the AtNSRV2 had 465 five major ORFs (ORFs I -V). These non-overlapping ORFs were arranged in a line along 466 the viral genome ( Figure 5A). We also identified the conserved noncoding sequences (3′-467 AUUU/AAAUAAAACUUAGGA-5′), which were downstream the ORFs ( Figure 5B). Be-468 cause the gene-junction sequences were ubiquitous in the viral genome, we determined 469 that they were a characteristic feature of the mononegaviruses. We deposited the nucleo-470 tide sequence of AtNSRV2 in GenBank (accession number: OP566533).

Phylogenetic Analysis of the Novel −ssRNA Viruses
We constructed a phylogenetic tree based on the RdRp aa sequences of AtNSRV2, AaNSRV1, and AtNSRV4 as well as members of the families Mymonaviridae and Discoviridae. We found that AtNSRV2 formed a supported clade with the following viruses: Soybean leafassociated negative-stranded RNA virus 1 (SlaNSRV-1), Fusarium gramineartm negativestranded RNA virus 1 (FgNSRV1), AtNRV1, BcNSRV3, SlaNSRV2, and CpSMV1 ( Figure 6). They were close to the clade and included members of the genus Sclerotimonavirus but were separate from other genera in the family Mymonaviridae ( Figure 6). Therefore, according to genomic characteristics and phylogenetic analysis, we proposed that AtNSRV2 should be a new member of the genus Sclerotimonavirus in the family Mymonaviridae. We also identified novel −ssRNA mycoviruses that should be grouped with members in the family Discoviridae. We included AaNSRV1 in a group with CdNSRV1, BcNSRV2, and Fusarium poae negative-stranded virus 2 (FpNSRV2), which were classed a genus Orthodiscovirus in the family Discoviridae. We also included AtNSRV4 in a clade with the members in the family Discoviridae, but we considered it to be part of different groups. Therefore, AaNSRV1 and AtNSRV4 are in two highly supported groups inside the family Discoviridae ( Figure 6). The genomic −ssRNA of a novel mycovirus called AtNSRV2 was from strain HB-15. It was completely sequenced, and we found the virus to have a similar genome structure and high sequence similarity with members of the genus Sclerotimonavirus in the family Mymonaviridae.

Discussion
To search for novel mycoviruses, we used fungal strains from different regions to apply viral metagenomics to a mixed pool. This method allowed us to identify a great variety of new mycoviruses with different classes of genomes. In this study, we collected a total of 78 strains, including 58 A. tenuissima, 13 A. alternata, 3 A. arborescens, 2 A. gaisen, 1 A. gossypina, and 1 A. longipes, from pear spot disease from different regions of China, and mixed a fungal pool used to search for novel mycoviruses. As a result, we identified 21 putative viral sequences, most of which were nearly the full-length genome. We found 12 +ssRNA viruses, 5 dsRNA, and 4 −ssRNA viruses. We also identified sequences that were near full-length sequences in putative mycoviral genomes. Notably, the following eight distinct lineages were similar to these viral families: Botourmiaviridae, Chrysoviridae, Deltaflexiviridae, Discoviridae, Hypoviridae, Partitiviridae, Mitoviridae, Mymonaviridae, and Narnaviridae. In these putative viruses, nine were new mycoviruses. Through our analysis, we found that most of these viruses were dsRNA and ssRNA, and we did not find any DNA virus in the 78 strains. We used RT-PCR with specific primers, which had been designed based on the assembled contigs, and confirmed this variety of viruses. Two viruses from strains of A. tenuissima were provided a complete genome sequence using RT-PCR and RACE in our study.
Viruses are commonly present in Alternaria spp. Recently, 14 mycoviruses in 7 families have been associated with fungi from Alternaria. Of these fungi, we also identified AaMV1, AaHV1, AaCV1, and AlV1 in our study. These four viruses were founded from the different isolates of A. arborescens, A. alternata, and A. longipes [24][25][26]34,39,40]. In our study, 10 viruses were discerned initially from other hosts with more than 90% similarity were detected in Alternaria isolates. A larger body of research has found that mycovirus occurs in two fungi hosts that are taxonomically. Some dsRNA and ssRNA mycoviruses were previously identified as infecting other fungal species or fungal genera. For example, Bipolaris maydis botybirnavirus 1 (BmBBV1) was found in Bipolaris maydis and Botryosphaeria dothidea [60]. Ophiostoma novo-ulmi mitovirus 3a-Ld (OnuMV3a-Ld) was found in Sclerotinia homoeocarpa and Ophiostoma novo-ulmi [61], Previously, Helminthosporium victoriae virus 190S was identified in Helminthosporium victoriae and Bipolaris maydis [62,63], and the mitovirus Hymenoscyphus fraxineus mitovirus 1 (HfMV1) was found in the Hymenoscyphus fraxineus and H. albidus [64,65]. Another study idengified a mycovirus that infected different fungal strains through metatranscriptomics. Two viruses, Macrophomina phaseolina mitovirus 4 and Rhizoctonia solani mitovirus 10, were different strains of the same virus infected different hosts [59] and Botrytis fuckeliana totivirus 1 was found in Botrytis cinerea and B. fuckeliana [14]. Additionally, Sclerotinia sclerotiorum hypovirus 1-A, Sclerotinia sclerotiorum hypovirus 2, Sclerotinia sclerotiorum negativestranded RNA virus 5, Sclerotinia sclerotiorum partitivirus 2 were found in in B. cinerea and S. sclerotiorum [14]. Both S. sclerotiorum and B. cinerea are necrotrophic fungi. They have wide hosts ranges, and their genomes show high sequence identity as well as a similar arrangement of genes [66]. Given that these mycoviruses can be found in coinfections in common plant hosts, these mycoviruses may transfer horizontally between coinfecting fungi. An essential resource may be transmitted between different fungi hosts. For example, Sclerotinia sclerotiorum hypovirulence associated DNA virus 1 (SsHADV-1) could infect a mycophagous insect, Lycoriella ingenua, which could act as a transmission vector [67]. Cryphonectria hypovirus 1 (CHV1) could replicate and spread in Nicotiana tabacum, which is a model plant [68]. Plant-fungal-mediated routes may disseminate the same viruses between different fungi in nature.
Recently, deep sequencing has been used to identify the diversity of fungal viruses within fungal species from diverse geographic regions [52,55,70,71,82,[84][85][86][87][88][89][90]. Unlike other dsRNA extraction and cloning methods, this approach could obtain viral information for different classified viruses regardless of their genome types [13,14,56,86]. Deep sequencing also has been used to identify mycoviruses from diverse fungal strains in an experimental project. For instance, using a high-throughput sequencing-based metatranscriptomic approach, 66 previously undescribed mycoviruses were obtained from five fungal species, including Colletotrichum truncatum, Diaporthe longicolla, Macrophomina phaseolina, R. solani, and S. sclerotiorum [13]. Multiple mycoviruses, which were coinfected in a fungal strain, were often efficiently identified by deep sequencing. For example, using deep sequencing, 17 different mycovirus that were assigned to different families were found in an R. solani isolate DC-17 [70]. Eleven mycoviruses were identified as being part of 3 families in a single F. mangiferae strain SP1 [71]. Eight new fungal viruses that co-infected a single isolate of P. juniperi were found using high-throughput sequencing [72]. In addition, high-throughput sequencing has been used to detect mycoviruses on the phyllosphere and arbuscular mycorrhizal fungi in the roots [54,85]. For example, 22 putative mycovirus genomes have been organized into 10 taxonomic groups and assembled from soybean leaf metatranscriptomes [54]. The diversity, evolution, and annual variation of mycovirus in S. sclerotiorum within a single field for three years were investigated using the metatranscriptomic approach [86].
Alternaria spp. have been reported to have several RNA viruses, including dsRNA viruses, positive-sense ssRNA viruses, and other unidentified viruses . The virus SsNSRV1 was the first member of the family Mymonaviridae [12]. Recently, two sclerotimonaviruses in this family have been found in A. tenuissima and A. dianthicola, respectively [44,45]. In this study, four negative-stranded RNA viruses have been identified. AtNSRV2 and AtNSRV3 were members of the genus Sclerotimonavirus in the family Mymonaviridae. The genome of the virus AtNSRV2 also was obtained. According to BLASTp, the L protein of AtNSRV2 was similar to the RdRp of SlaNRV2 with 55.85% identity. The virus SlaNRV2 was assembled from soybean leaf metatranscriptomes [54]. AaNSRV1 and AtNSRV4 were new members of the family Discoviridae. More research is needed to address these differences and to confirm their molecular and biological characterization. A lot of −ssRNA mycoviruses were classed into the order Discoviridae [3,4]. A recently reported BcNSRV1, RsNSRV4, and Macrophomina phaseolina negative-stranded RNA virus 1 (Mp-NSRV1) are new members of the order [8,13]. This is the first time that a mycovirus with a negative-stranded ssRNA genome has been reported to have infected an Alternaria strain.

Conclusions
In conclusion, diversity analysis of mycoviruses from 78 strains was executed by using high-throughput sequencing technology. The used stains were collected from 10 pear production areas and belonged to six species of the genus Alternaria. We excavated at least 21 different mycoviruses. The +ssRNA viruses belonged to the families Mitoviridae, Narnaviridae, Deltaflexiviridae, Hypoviridae, and Botourmiaviridae. The dsRNA viruses belonged to the families Chrysoviridae, Partitiviridae, and a genus Botybirnavirus, and one unclassified virus. The −ssRNA viruses belonged to the families Mymonaviridae of order Mononegavirales, and Discoviridae of order Bunyavirales. We also identified near-full-length sequences of mycoviral genomes that are putative. We isolated a novel −ssRNA mycovirus from an A. tenuissima strain HB-15, which we designated as AtNSRV2. We also characterized a novel +ssRNA mycovirus from an A. tenuissima strain SC-8, which we designated as AtDFV1. Through phylogenetic and sequence analyses, we found AtNSRV2 to be related to the viruses of the genus Sclerotimonavirus in the family Mymonaviridae. We also found that AtDFV1 is related to the family Deltaflexivirus. The results of this study significantly enhanced the number of Alternaria viruses and identified their abundant diversity.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/v14112552/s1, Table S1: Origin of the strains of Alternaria species used in this study; Table S2: RT-PCR primers used for detection viruses in this study; Table S3: RT-PCR primers used for obtaining full-length sequences of the viruses in this study; Table S4: Abbreviations of virus names and viral protein accession numbers used in alignment analysis in this study; Table S5: Best BLASTp matches of P1 of Alternaria tentissima deltaflexivirts 1; Table S6: Best BLASTp matches of P4 of Alternaria tentissima negative-stranded RNA virus 1.