An Outstandingly Rare Occurrence of Mycoviruses in Soil Strains of the Plant-Beneficial Fungi from the Genus Trichoderma and a Novel Polymycoviridae Isolate

ABSTRACT In fungi, viral infections frequently remain cryptic causing little or no phenotypic changes. It can indicate either a long history of coevolution or a strong immune system of the host. Some fungi are outstandingly ubiquitous and can be recovered from a great diversity of habitats. However, the role of viral infection in the emergence of environmental opportunistic species is not known. The genus of filamentous and mycoparasitic fungi Trichoderma (Hypocreales, Ascomycota) consists of more than 400 species, which mainly occur on dead wood, other fungi, or as endo- and epiphytes. However, some species are environmental opportunists because they are cosmopolitan, can establish in a diversity of habitats, and can also become pests on mushroom farms and infect immunocompromised humans. In this study, we investigated the library of 163 Trichoderma strains isolated from grassland soils in Inner Mongolia, China, and found only four strains with signs of the mycoviral nucleic acids, including a strain of T. barbatum infected with a novel strain of the Polymycoviridae and named and characterized here as Trichoderma barbatum polymycovirus 1 (TbPMV1). Phylogenetic analysis suggested that TbPMV1 was evolutionarily distinct from the Polymycoviridae isolated either from Eurotialean fungi or from the order Magnaportales. Although the Polymycoviridae viruses were also known from Hypocrealean Beauveria bassiana, the phylogeny of TbPMV1 did not reflect the phylogeny of the host. Our analysis lays the groundwork for further in-depth characterization of TbPMV1 and the role of mycoviruses in the emergence of environmental opportunism in Trichoderma. IMPORTANCE Although viruses infect all organisms, our knowledge of some groups of eukaryotes remains limited. For instance, the diversity of viruses infecting fungi—mycoviruses—is largely unknown. However, the knowledge of viruses associated with industrially relevant and plant-beneficial fungi, such as Trichoderma spp. (Hypocreales, Ascomycota), may shed light on the stability of their phenotypes and the expression of beneficial traits. In this study, we screened the library of soilborne Trichoderma strains because these isolates may be developed into bioeffectors for plant protection and sustainable agriculture. Notably, the diversity of endophytic viruses in soil Trichoderma was outstandingly low. Only 2% of 163 strains contained traces of dsRNA viruses, including the new Trichoderma barbatum polymycovirus 1 (TbPMV1) characterized in this study. TbPMV1 is the first mycovirus found in Trichoderma. Our results indicate that the limited data prevent the in-depth study of the evolutionary relationship between soilborne fungi and is worth further investigation.

T richoderma spp. (Hypocreales, Ascomycota) are mycoparasitic filamentous fungi that occur on other fungi, dead wood, and in soils. Some of them inhabit a diverse range of ecological niches in plants as endo-and epiphytes and thus are considered environmental opportunists (1). Trichoderma spp. are commonly used as agents of biological control of fungal pests (biocontrol) (2)(3)(4)(5) in sustainable agriculture because of their ability to parasitize plant-pathogenic fungi. They are especially effective in controlling soilborne diseases, such as root rot caused by Rhizoctonia spp., Verticillium spp., and Fusarium spp. (6)(7)(8)(9). There is also evidence to suggest that Trichoderma species can promote plant growth and induce systemic resistance (10)(11)(12)(13)(14). The comparative genomic studies of several opportunistic strains of Trichoderma revealed numerous properties, which can be associated with their ecological versatility but did not explain them (1,15,16). In this study, we tested whether the soil strains of Trichoderma contain viruses as our previous studies showed that this may influence their phenotype (12,14).
Most mycoviruses have no significant influence on the colony morphology and mycelial growth rate of Trichoderma spp. (20,21,25,26), except for the research of Wang et al. on Trichoderma harzianum partitivirus 2, or ThPV2 (14), which caused faster growth of the host strain and triggered abundant aerial hyphae and dark green pigmentation of the conidia. However, the infection of T. harzianum with ThPV2 did not change the efficiency of the biocontrol product based on this fungus. In other reports, viruses affected the secondary metabolism of Trichoderma (26) or influenced the activity of lignocellulolytic enzymes (20). Some mycoviruses, such as ThPV2 (14), altered the biomass and spore production by Trichoderma spp. (14,25,28,29); similar to the effects of Fusarium graminearum virus-china 9 (FgV-ch9) (28) and Rhizoctonia solani dsRNA virus 5 (RsRV5) (29) on their hosts.
Soil-borne Trichoderma spp. may be putatively suitable bioeffectors for agriculture (36). To study the diversity of mycoviruses, we tested 163 soil isolates. Notably, only four strains contained mycoviruses, including the new virus from T. barbatum. The genome and predicted protein products of this mycovirus were characterized and the relationships with other mycoviruses were analyzed.

RESULTS
One strain of Trichoderma out of 160 was infected by a virus. One hundred sixty Trichoderma species isolates were tested for the presence of RNA viruses, and only a single strain-HB40111-shows a positive result ( Fig. 1A and B). The three-loci DNA barcoding analysis as specified by Cai and Druzhinina (37) identified the strain HB40111 of T. barbatum as tef1 (NCBI GenBank accession number OP978144) and rpb2 (NCBI GenBank accession number OP978145) with sequences that were . 97% and 99% similar to the sequences of the ex-type strain of this species (CBS 125733), respectively; ITS (NCBI GenBank accession number OP978249) was also identical to the sequence of CBS 125733. Diversity analysis using ITS rRNA sequences revealed that there are no other strains with identical or highly similar phylotypes, meaning that only one isolate of T. barbatum was sampled (Table S1 and Fig. S1). The detailed molecular identification of the 159 virus-free strains is presented elsewhere. In our previous study, we detected mycoviruses in the three other strains from the same data set, including two distinct strains of T. harzianum infected with Trichoderma harzianum bipartite mycovirus 1 (25) and Trichoderma harzianum mycovirus 1 (12). Furthermore, Trichoderma sp. HB40444 contained a putative dsRNA mycovirus, which could not be identified (data not shown). Thus, the 163 Trichoderma strains isolated from grassland soils in Inner Mongolia contained a maximum of four isolates with putative viral infections.  T. barbatum HB40111 was infected with the dsRNA virus. The metagenomic sequencing and the electrophoretic analysis revealed that the sizes of the mycovirus fragments are approximately 2.4 kb, 2.1 kb, 1.9 kb, and 1 kb (Fig. 1). We screened dsRNA contigs with a similar band size and homology to fungal mycoviruses. Based on the sizes of the fragments (2.4 kb, 2.1 kb, 1.9 kb, and 1 kb), four virus contigs with similar sizes were screened and searched for similar isolates from the BLASTN of NCBI (https://blast.ncbi.nlm.nih.gov/Blast.cgi?PROGRAM=tblastn&PAGE_TYPE=BlastSearch& LINK_LOC=blasthome), and three fungal mycoviruses had homology with it: Aspergillus fumigatus tetramycovirus-1, Cladosporium cladosporioides virus 1, and Botryosphaeria dothidea virus 1. RT-PCR analysis verified that the extracted dsRNA contains the screened contigs (Fig. 1). The sequencing results were consistent with the metagenomics sequencing results. The 59 end sequence and 39 end sequence of the cDNA were obtained using the classical 59 RACE and 39 RACE cloning methods, and the 39 end of the virus contained a poly(A) structure for every fragment. The complete genome sequencing of the virus was assembled: dsRNA1 was 2420 bp, dsRNA2 was 2127 bp, dsRNA3 was 1909 bp, and dsRNA4 was 1041 bp. The GC content of each segment is 56.7%, 58.3%, 55.7%, and 61.5%. The NCBI ORF Finder predicted that the dsRNA1 and dsRNA2 coding strands contain one ORF each on the positive strand, whereas dsRNA3 and dsRNA4 have one ORF each on the negative strand. ORF1 (nt positions 33 to 2333) was predicted to encode a protein of 766 amino acids (aa) in length and ;84.7 kDa with RNA-dependent RNA polymerase (RdRP) activity. ORF2 (nt positions 75 to 2051) was predicted to encode a hypothetical protein of 678 aa in length and ;71.1 kDa, and ORF3 (nt 1857 to 16) was predicted to encode a protein of 613 aa in length and ;65.5 kDa with methyl transferase activity. Finally, ORF4 (nt positions 932 to 150) was predicted to encode a hypothetical protein of 260 aa in length and ;27.7 kDa. These nucleotide sequences and amino acid sequences have been submitted to GenBank under the accession numbers OM307406, OM307407, OM307408, and OM307409. The 59 UTRs of dsRNA1, dsRNA2, dsRNA3, and dsRNA4 are 32, 74, 52, and 109 bp long, respectively (Fig. S2, S4, S6, and S8); the 39UTRs are 87, 76, 15, and 149 bp ( Fig. 1; Fig. S3, S5, S7, S9, and S8).
Finally, we constructed a phylogenetic tree based on the concatenated ORF1, ORF2, ORF3, and ORF4 alignment (Fig. 2). The resulting tree topology was consistent with those of the single ORF-based trees. Based on these results, we conclude that the novel mycovirus from T. barbatum BH40111 is a polymycovirus, which is the first mycovirus of T. barbatum; therefore, we propose the name Trichoderma barbatum polymycovirus 1 (TbPMV1). However, TbPMV1 has a distinct phylogenetic position surrounded by viruses isolated from fungi from another class (Eurotiomycetes) or Magnaportales order of Sordariomycetes. The viruses from the phylogenetically close Trichoderma hosts belonged to the same subclade of the Polymycoviridae but are genetically distinct (Fig. 2).

DISCUSSION
Members of the Polymycoviridae family are unique and possess multisegmented dsRNA genomes (from 3 to 11 segments) (34,(41)(42)(43) and noncapsid proteins (44). In this study, we discovered a novel mycovirus species, TbPMV1, a Polymycoviridae member present in the T. barbatum strain isolated from the soil. This is the first polymycovirus identified in the genus Trichoderma, enriching the known diversity of mycoviruses associated with these fungi (Table 1). Before our study, 11 Trichoderma mycoviruses have been reported, which have shown diversity in population, comprising one Fusagraviridae (TaMV1-NFCF377), two Hypoviridae (ThHV1 and ThHV2), three Partitiviridae (ThPV1, TaPV1, and TaPV2), one totivirus (TkTV1/Mg10), and four unclassified (TaMV1, TaRV1, ThBMV1, and ThMV1). Though phylogenetic analysis was carried out for the 12 mycoviruses, the best alignment of each possible protein could not be obtained, which means it is difficult to demonstrate the evolutionary relationship using the limited isolates from Trichoderma spp.; we need to explore more mycovirus isolates from this soilborne fungus. Moreover, the discovery of TbPMV1 also provides evidence for the classification and diversity of mycoviruses associated with Trichoderma spp.
Because the previous characterizations of Trichoderma mycoviruses are limited, there is a dearth of experimentally validated functional data from which the functions of proteins In addition to the characterization of the novel Polymycoviridae virus and the first detection of these viruses in Trichoderma spp., the most interesting result of this study is the outstandingly rare occurrence of viruses in these fungi. The phylogenetic analysis suggests that these viruses are unlikely to be specific to T. barbatum or that the data are not enough to speculate on the possible coevolution between the virus and the host. This is the fourth mycovirus to be isolated from a collection of 163 Trichoderma strains in our laboratory, making the positive rate 2.45%. This demonstrates that associations between mycoviruses and Trichoderma strains are outstandingly rare. Further screening should reveal whether strictly mycoparasitic strains of Trichoderma are also largely virus-free and whether the same pattern can be reproduced for soil Trichoderma from another geographic region.

MATERIALS AND METHODS
Isolation and molecular identification of Trichoderma spp. One hundred and sixty-three Trichoderma species isolates were obtained from grassland soils in Xinjiang and Inner Mongolia provinces from 2014 to 2016 (Table S1). Pure single-spored cultures were stored on potato dextrose agar at 4°C. The diversity of the strains was assessed using the primary fungal DNA barcoding locus (ITS of the rRNA gene cluster) (37,45), while the precise and accurate molecular identification of these strains is presented elsewhere. The strain containing the new mycovirus was identified using the combination of the ITS and the two secondary DNA barcoding markers (partial fragments encoding the translation elongation factor 1-alpha tef1 and RNA-polymerase subunit B2 rpb2) using the protocols of Cai and Druzhinina (46).
Extraction and purification of dsRNA. The diversity of mycoviruses with RNA genomes was tested in the strains that were cultivated in the liquid potato glucose broth at 28°C with 200 rpm shaking for 2 days. Mycelia were collected from the liquid cultures via vacuum filtration (using a 0.22-mm pore size). A standard phenol-chloroform extraction method (47) was used to isolate the RNA, which was then purified using CF11 cellulose column chromatography as previously described (48). The isolated RNA was treated with RNase-free DNase I (TaKaRa, Dalian, China) and S1 Nuclease (TaKaRa) according to the manufacturer's instructions to remove DNA and single-stranded nucleic acids, respectively. Finally, the dsRNA was visualized on a 1% agarose gel. Moreover, the dsRNA was also degraded by DNase and RNase III; DNase can degrade dsDNA and ssDNA while RNase III can only degrade ssRNA. Therefore, after the degradation, the nucleotide type should be checked out. After the application of DNase, if the fragments disappear the mycovirus should be the ssDNA or dsDNA nucleotide style; if the fragments are retained, the mycovirus should be the dsRNA nucleotide style. After the degradation by RNase III, if the fragments disappear, the mycovirus should be the ssRNA nucleotide style; if the fragments are retained, the mycovirus should be the dsRNA, ssDNA, or dsDNA nucleotide style. Therefore, the nucleotide style of the mycovirus needs to be screened after the degradation by DNase and RNase III.
Characterization of the mycovirus genome. Unigenes were annotated using RefSeq, and proteins were annotated using the NAR virus database (https://digitalworldbiology.com/blog/bio-databases-2020-viruses -and-covid-19). Summed contig reads per kilobase per million mapped reads values were calculated for each species based on the annotations. Oligonucleotide primers were then designed based on the mycovirus contigs. Reverse transcription (RT)-PCR was performed to determine whether the target regions could be amplified. Amplicons were purified using a gel extraction kit (TaKaRa, Dalian, China), ligated to the PMD18-T vector, and sent to the Shanghai Biotechnology Corporation for sequencing. Most of the sequences of the four dsRNA fragments of the mycovirus isolated from Trichoderma strain HB40111 were obtained. The missing 59and 39-terminal end sequences were resolved using 59-and 39-rapid amplification of cDNA ends (RACE) cloning (52,53). Using Vector NTI Advance 11.5.4, contiguous sequences for the single RNAs were assembled into four complete sequences (dsRNA1, dsRNA2, dsRNA3, and dsRNA4), yielding the complete viral genome. The PCR products were sent to TsingKe Biological Technology (Beijing, China) for sequencing.
Phylogenetic analysis. Putative open reading frames (ORFs) were predicted with the ORF Finder tool (https://www.bioinformatics.org/sms2/orf_Find) (54). The nucleotide mycovirus sequences were used as queries in BLASTN analysis (https://blast.ncbi.nlm.nih.gov/ BLAST.cgi) on the United States National Center for Biotechnology Information (NCBI) website (https://blast.ncbi.nlm.nih.gov/Blast.cgi) to identify similar mycoviruses for inclusion in the phylogenetic analysis. All sequences that were derived from fungal dsRNA viruses, had similar sizes to the genomic fragments observed on the agarose gel, and had a similarity higher than 40% were selected for further processing. Phylogenetic trees were constructed with the best models based on the amino acid on the single ORFs (ORF1, ORF2, ORF3, and ORF4) and a concatenated ORF11ORF21ORF31ORF4 of the 12 mycovirus isolates from Trichoderma spp. alignment using the maximum likelihood (ML) method implemented in MEGA v10.0 (https://megasoftware.net/) (55) with 1000 bootstrap replicates. The best amino acid substitution models were LG1G1I1F (ORF1), LG1G1F (ORF2, ORF3, and a concatenated sequence), and WAG1G (ORF4) (Table S2).
Viral nomenclature. The name of the novel mycovirus was assigned according to the International Code of Virus Classification and Nomenclature (ICVCN) assessed through the website of the International Committee on Taxonomy of Viruses (https://ictv.global/about/code) and their guidelines (https://ictv.global/ faq/names).
Data availability. The nucleotide and amino acid sequences of ORF1, ORF2, ORF3, and ORF4 of the novel mycovirus isolated in this study have been submitted to GenBank under the accession numbers OM307406, OM307407, OM307408, and OM307409, respectively. The proteins encoded by ORF1, ORF2, ORF3, and ORF4 were also assigned the accession numbers WAA18813, WAA18814, WAA18815, and WAA18816, respectively.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. SUPPLEMENTAL FILE 1, PDF file, 0.5 MB. The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.