The R. rugulosa isolate, used in this study was isolated from ARD suffering, in vitro propagated M26 roots (Malus x domestica, Borkh.), which were grown for 8 weeks in soil from the ARD affected site Ellerhoop (Chamber of Agriculture Schleswig–Holstein, Germany, 53°42′51.7″N, 9°46′12.5″E) . Before isolation, fine root surfaces were disinfected (70% Ethanol 30 s, followed by 7.5 min 2% NaOCl). Root pieces of 1 cm length were plated on water agar (50 µg mL−1 penicillin, 10 µg mL−1 rifampicin, 25 µg mL−1 pimaricin). The outgrown endophytic fungi were separated and cultivated on 2% malt extract agar (MEA). For nucleic acid extraction, the fungi were propagated in 2% malt extract broth for 2 weeks and mycelium was ground in liquid nitrogen. R. rugulosa was identified by PCR and Sanger sequencing, as described by Crous et al. , using primers for the histone H3 gene (CYLH3F: 5ʹ AGGTCCACTGGTGGCAAG 3ʹ; CYLH3R: 5ʹ AGCTGGATGTCCTTGGACTG 3ʹ) . The sequencing was performed at Microsynth Seqlab (Göttingen, Germany). The resulting sequences were analyzed by NCBI BLASTn.
The R. rugulosa isolate investigated in this study was named No4. To test the infectivity of No4 after isolation, M26 plants were re-infected using a soil-free inoculation assay as described by Popp et al., 2019 . The plants showed reduced growth, as well as typical blackening in microscopic analyses of fine roots after 5 weeks.
Extraction of nucleic acids
DsRNA for Illumina sequencing was extracted from 20 g ground fungal material, stored at − 80 °C, based on a modified protocol of Morris and Dodds  as described by Lesker et al. , apart from using a different cellulose (acid-washed powder for column chromatography [Merck; Darmstadt, Germany; product nr. 22,184]). 20 mL eluate was digested first with 20 U Rnase T1 (Roche; Basel, Switzerland) and then with 40 U DNAse I (Roche; Basel, Switzerland) at 37 °C for 30 min each. DsRNA extracts were centrifuged and suspended in 25 µL Tris (5 mM). Subsequently, 20 µL extract was checked with 5 µL of GelRed® (Biotium; Fremont, CA, USA) dye in 1.5% agarose gel electrophoreses. For virus detection by RT-PCR and RNA end determination, a simpler protocol for whole nucleic acid extraction was used, following the protocol of Menzel et al. .
A Nextera XT Library Preparation Kit was used to prepare an Illumina library from double-stranded cDNA, obtained by cDNA synthesis of the dsRNA extract and second-strand synthesis with random octamer primers. The library was sequenced at the Leibniz-Institute DSMZ on a NextSeq instrument as paired-end reads (2 × 151 bp). The raw reads were trimmed and de novo assembled with Geneious v. R11.1 software (Biomatters; Auckland, New Zealand) using an in-house established workflow, followed by local BLASTn and BLASTp alignments of the assembled contigs against a custom database of NCBI nuclear-core reference sequences. The identified mycovirus contigs were ordered and trimmed according to reference sequences to determine the nearly complete genome sequences. The sequence information was used to design primers for virus detection by RT-PCR and determination of the extreme terminal ends of the genomes by RACE.
Virus detection with RT-PCR
RT-PCR protocols were adapted for the detection of each of the genomic viral segments. For cDNA synthesis, 4 µL dsRNA extract, 1 µL cDNA primer [10 µM (Table 1); salt-free; Eurofins Genomics; Ebersberg, Germany] and 5 µL A.bidest were mixed and heated up to 95 °C for three minutes to separate the dsRNA strands. 50 U Maxima H Minus Reverse Transcriptase (Thermo Fisher Scientific™; Waltham, MA, USA), 20 U RiboLock RNase Inhibitor (Thermo Fisher Scientific™; Waltham, MA, USA), 1 µL dNTPs (10 mM each; Thermo Fisher Scientific™; Waltham, MA, USA) and 4 µL 5X RT-buffer (250 mM Tris–HCl (pH 8.3), 250 mM KCl, 20 mM MgCl2, 50 mM DTT; Thermo Fisher Scientific™; Waltham, MA, USA) were added subsequently and adjusted to 20 µL with A.bidest. The cDNA synthesis started with 60 min at 50 °C, followed by 15 min at 55 °C, 15 min at 60 °C, and 5 min at 85 °C. PCR was performed with 5 µL 2 × Phusion Flash High-Fidelity PCR Master Mix (Thermo Fisher Scientific™; Waltham, MA, USA), 1 µL of each specific primer [forward and reverse, 10 µM each, salt-free; Eurofins Genomics; Ebersberg, Germany (Table 1)], 2 µL cDNA and 1 µL A.bidest. PCR was performed with an initial denaturation of 15 s at 98 °C, followed by 34 cycles (denaturation: 98 °C, 5 s; annealing: primer TA, 5 s; elongation: 72 °C, 15 s / 1000 bp amplicon), and a final elongation of 300 s at 72 °C. Amplicons were visualized by 1.0% agarose electrophoresis and sent to Microsynth Seqlab (Göttingen, Germany) for Sanger sequencing .
RNA end determination
The ends of dsRNAs were determined by RACE with an adapted protocol, based on the method described by Frohman et al. . 3ʹ-ends of both, dsRNA sense and antisense strands were analyzed. Reverse transcription followed the described protocol for virus detection, with different cDNA primers (10 µM each, salt-free; Eurofins Genomics; Ebersberg, Germany; Table 2). For tailing 3 µL cDNA was mixed with 20 U Terminal Deoxynucleotidyl Transferase (TdT; Thermo Fisher Scientific™; Waltham, MA, USA), 4 µL 5 × TdT Reaction Buffer (500 mM potassium cacodylate (pH 7.2), 10 mM CoCl2, 1 mM DTT; Thermo Fisher Scientific™; Waltham, MA, USA), 1 µL of either dATP, dCTP, dGTP, or dTTP (100 mM, Thermo Fisher Scientific™; Waltham, MA, USA) and 11 µL A.bidest. The mixture was incubated for 30 min at 37 °C followed by 10 min at 70 °C. For each RNA end, at least two different tails were used in different reactions. The subsequent PCR was performed as the one for virus detection with a different primer set (poly-n primer; nested primer; 10 µM each, salt-free; Eurofins Genomics; Ebersberg, Germany; Table 2).
Phylogenetic analyses were performed with several mycoviruses of the families Totiviridae, Quadriviridae, Chrysoviridae, Mitoviridae, and ten unclassified dsRNA Riboviria viruses. Before constructing a phylogenetic tree, amino acid sequences of the RNA-dependent RNA polymerase of all viruses were aligned, using the MUSCLE algorithm in MEGA X [27, 28]. Parameters were set to default (gap opening: − 2.9, gap extension: 0). After the initial alignment, highly conserved sequences were selected, referring to the segment A(679)–E(1066) of NC_016760 . The final alignment was performed with the set of chosen segments and default parameters, using the MUSCLE algorithm again. A maximum-likelihood tree was calculated, using the bootstrap method with 1000 replications and the Le_Gascuel_2008 substitution model with discrete Gamma distribution (LG + G) . The number of discrete gamma categories was set to 5 and for the data subset, all sites were used. Pairwise alignments of all segments of the families Quadriviridae, Mitoviridae, and the unclassified dsRNA viruses were done to calculate sequence identities by using the EMBOSS/Needle tool .
UTR alignment, secondary structures, and motifs
Conserved UTR-sequences of the proposed quadrivirus RrQV1 were analyzed by alignments of the four 5ʹ- and 3ʹ-ends and presented with the GeneDoc Software (National Resource for Biomedical Supercomputing, Pittsburgh, PA, USA). Secondary structure predictions of the proposed mitovirus RrMV1 were computed with the RNAfold WebServer (Institute for Theoretical Chemistry, University of Vienna) [32, 33]. Conserved motifs within the genomic RNAs were identified by using the NCBI Conserved Domain Search tool with default options (database: CDD v3.19–—58,235 PSSMs, expected value threshold: 0.01) .