Double-stranded RNA partitiviruses identified in S. cerevisiae

To take an unbiased approach to understanding virus diversity in S. cerevisiae, dsRNAs were purified from a random selection of 520 strains from the 1,002 Yeast Genomes Project collection (S1 Table) [44]. These dsRNAs were analyzed by a short-read sequencing pipeline that identified 110 contigs with significant sequence similarity to known virus families (Fig 1A and S2 Table) [45,46]. Contigs with sequences similar to the Totiviridae were most frequently identified. Contigs with similarity to satellite dsRNAs were also identified with ORFs corresponding to the S. cerevisiae killer toxins K1, K2, Klus, and S. paradoxus killer toxins K66 and K21. After totiviruses and satellites, the next highest coverage score was for viruses of the Partitiviridae [19]. Other virus families identified by these analyses were represented by three or fewer contigs and were not investigated further (S2 Table). Assembled contigs identified sequences similar to previously identified virus proteins in Saccharomyces yeasts that were not fully characterized [47,48].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. The identification of PVs in S. cerevisiae.

(A) Virus contigs were assembled from the sequencing of dsRNAs extracted from 520 strains of S. cerevisiae. (B) Agarose gel electrophoresis analysis of dsRNAs extracted from two PV-infected strains of S. cerevisiae. (C) DsRNAs extracted in panel B were subjected to incubation at 37°C with and without RNase III, and the resulting products were analyzed by agarose gel electrophoresis. (D) Two-step RT-PCR of total nucleic acids extracted from S. cerevisiae targeting the CP or RdRP genes with (+RT) or without (-RT) reverse transcriptase. Strain BJH001 was used as a negative control as it does not harbor PVs. Primers targeting the yeast gene UBC6 were used as a PCR-positive control for genomic DNA.

https://doi.org/10.1371/journal.ppat.1011418.g001

To confirm the presence of PVs in S. cerevisiae, 161 strains from the 1,002 Yeast Genomes Project collection, 146 strains from the collection of Ludlow et al., and three additional strains from other culture collections were screened using cellulose chromatography to extract dsRNAs (S3 Table) [44,49]. This approach identified PVs by the characteristic electrophoretic mobility of their dsRNAs, with 14% of strains containing PVs, mainly from the collection of Ludlow et al (Table 1). For example, purifying dsRNAs from S. cerevisiae strains YO858 and Y-5509 identified three distinct nucleic acid molecules with similar electrophoretic mobilities in both strains (Fig 1B). The RNAs were confirmed to be dsRNAs by RNase III digestion (Fig 1C). These dsRNAs likely represent a monopartite totivirus genome (≈4.6 kbp) and the bipartite genome of a PV (≈1.5 kbp) encoding an RdRP (dsRNA1) and CP (dsRNA2). Finally, reverse transcriptase-PCR (RT-PCR) confirmed PVs in total nucleic acid extracts of S. cerevisiae strains YO858 and Y-5509 (Fig 1D). PCR alone did not result in the amplification of PV sequences, and all samples were positive for the yeast chromosomal gene UBC6 (Fig 1D). S. cerevisiae BJH001 was included as a negative control as it is not infected with PVs [45].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. The prevalence of viruses in 310 strains of S. cerevisiae.

https://doi.org/10.1371/journal.ppat.1011418.t001

In Saccharomyces yeasts, dsRNAs smaller than 4 kbp have been associated with non-autonomously replicating satellites [1,2]. These satellites are often responsible for encoding antifungal killer toxins maintained by the replicative machinery of totiviruses. One strain of killer yeast (S. cerevisiae CYC1172 (Fig 2A)) was discovered to harbor dsRNAs corresponding to a totivirus and three species of low molecular weight dsRNAs (Fig 2B). This suggested the coexistence of a totivirus, PV, and a satellite dsRNA. To confirm that one of the three low molecular weight dsRNAs was a satellite, strain CYC1172 was grown on agar containing cycloheximide. This resulted in the curing of the satellite dsRNA and the loss of the killer phenotype, but not the curing of the totivirus or PV (Fig 2A and 2B). Moreover, screens of S. cerevisiae strains for dsRNAs identified 15 strains with PVs that lacked totiviruses (4.8% of all yeasts screened; Table 1), supporting the autonomy of PVs (Fig 2C).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. The presence of ScPVs is independent of satellite dsRNAs and totiviruses.

(A) Top Schematic of a killer toxin production by a killer yeast strain with a zone of growth inhibition. Bottom Killer activity of strain CYC1172 before and after treatment with cycloheximide (B) Electrophoretic mobilities of dsRNAs extracted from CYC1172 before and after treatment with cycloheximide. (C) Agarose gel electrophoresis of dsRNAs extracted from selected strains of S. cerevisiae that contained either a totivirus and PV (lane 1) or only a PV (lanes 2–5).

https://doi.org/10.1371/journal.ppat.1011418.g002

PVs from S. cerevisiae are most closely related to Cryptosporidium parvum virus 1

ScPVs were significantly enriched in strains isolated from coffee (25/44, Fisher’s exact test p<0.01) and cacao (16/44, Fisher’s exact test p<0.01) (Fig 3A) (S3 Table). Only three strains of S. cerevisiae with PVs (Y-5509, CYC1172, and ICV D254 Lalvin) were isolated from other environments (coconut pods and grape must). To characterize the genome sequence and organization of PVs from S. cerevisiae (ScPVs), complete ORFs were assembled for 14 ScPVs that corresponded to both an RdRP (486–491 amino acids) and a CP (280–310 amino acids) using short read sequencing (S1 Fig and S4 Table). ScPV proteins were smaller than PVs from plants, fungi, and protozoa, except for the RdRPs from the Deltapartitivirus genus (Fig 3B). An alignment of the RdRP from 14 ScPVs and 45 representative sequences from five official and two unofficial Partitiviridae genera was constructed for phylogenetic analysis using maximum likelihood. All ScPVs clustered as a single monophyletic group, with CSpV1 from the protozoan C. parvum as the most closely related outgroup (Fig 3C). The percent amino acid identity between ScPV and CSpV1 RdRPs was between 33–38%, and the CPs between 17–23% (S5 Table). The clustering of ScPVs suggests that there are three distinct species, which have been named Saccharomyces cerevisiae partitivirus-1, -2, and -3, represented by ScPVs isolated from strains Y-5509 (ScPV1-5509), YO858 (ScPV2-858), and CYC1172 (ScPV3-1172), respectively. The amino acid identities of the RdRP or CP from these three species are less than 90% identical, supporting their categorization as distinct species (S5 Table). The partial sequence of an ScPV (named ScCV1) was previously identified in a European oenological yeast strain [48], which has 98.78% amino acid sequence identity to ScPV3-1172, isolated from a European wine strain (La Rioja, Spain) in grape must.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3.

ScPVs are closely related to C. parvum PVs (A) Strains of S. cerevisiae isolated from different sources and the proportion infected with ScPVs (dotted regions). (B) Comparing the protein lengths of RdRPs and CPs of diverse PVs to ScPV1-5509, SCPV2-858, and ScPV3-1172. (C) A PhyML maximum likelihood phylogenetic model of the relatedness of PVs based on the amino acid sequence of the RdRP [50]. The RdRP from the picobirnavirus OaPV was used as an outgroup. The VT amino-acid exchange rate matrices were selected with a gamma distribution (G), a proportion of invariable sites (I), and empirical amino acid frequencies in the alignment fit these data best, as judged by PROTTEST [51]. The numbers at each node are the bootstrap values from 1000 iterations. The scale bar represents the distance of one amino acid substitution per site. The amino acid sequences in the phylogeny are from the viruses listed in S7 Table.

https://doi.org/10.1371/journal.ppat.1011418.g003

Having identified many strains of S. cerevisiae with ScPVs, RT-PCR, and short-read sequencing were used to determine which species of ScPV were present in each yeast strain (S6 Table and Table 1). Eight strains harbored ScPV1, 16 strains ScPV2, and 16 strains were co-infected with ScPV1 and ScPV2. ScPV3 was only found in two European wine strains. Co-infection of strains with ScPV1 and ScPV2 was confirmed using gel electrophoresis by the resolution of doublets of dsRNA1 (S2 Fig).

The complete genome sequences of ScPV1-5509, ScPV2-858, and ScPV3-1172 were determined by 5’ rapid amplification of cDNA ends (RACE) (Fig 4A). Overall, the bipartite organization of ScPVs was the same as CSpV1, with similar length genome segments and untranslated regions (UTRs). Analysis of the 5’ UTRs of these viruses by mFold revealed commonalities in RNA secondary structure between genome segments of each ScPV and extensive structure in dsRNA2 (Fig 4A and S3 Fig). The sequence analysis of the 3’ UTRs of ScPVs identified conserved sequence elements (CSEs) in dsRNA1 and dsRNA2 that were unique to each virus species, including CSpV1 (Fig 4B) [52]. The CSEs of ScPV1-5509 and ScPV2-858 were 32 bp and 40 bp, and 100% identical when comparing the UTRs of dsRNA1 and dsRNA2. ScPV3-1172 encoded a CSE of 36 bp but with 84% identity. There was no homology between CSEs from different ScPV species, and all were predicted to form small stem-loop structures that could be involved in viral RNA replication or packaging.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Genome organization of PVs from S. cerevisiae.

(A) Schematic representation of dsRNA1 and dsRNA2 of three species of PV from S. cerevisiae as well as CSpV1. The ORFs are represented as rectangles that encode the RdRP and CP. Stem-loop structures are annotated to represent similar structures in the 5’ UTRs of each species (S3 Fig). ^#/*/**/*** represent the pairs of CSE sequences in the terminal 3’ UTRs (B) Secondary structure of RNA sequence present in the 3’ CSE of dsRNA1 and dsRNA2 of ScPV1-5509, ScPV2-858, and ScPV3-1172.

https://doi.org/10.1371/journal.ppat.1011418.g004

ScPVs assemble spherical particles with structural similarity to the PV shell domain

Previously described PVs formed spherical particles that package viral RNAs for replication and transcription. To determine whether ScPVs assemble similar particles, a strain of S. cerevisiae only infected with ScPVs was used for viral particle purification. Spherical virus-like particles with a diameter of ≈30 nm that lacked obvious P-domain surface protrusions were observed by TEM after sucrose gradient fractionation (Fig 5A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. The ScPV CP and RdRP share structural similarities with PVs and poliovirus.

(A) Visualization of ScPV1 particles by TEM as indicated by arrowheads. Scale bars are 500 nm (main image) and 25 nm (inset). (B) Molecular modeling using AlphaFold2 and energy minimization of the S-domain of three species of ScPV and CSpV1 compared to the solved structure of PCV1 (PDB:7ncr). The α-helices (blue rectangles) and β-sheets (magenta arrows) are labeled according to their common positioning within the S-domains. (C) An overlay of the five structures represented in panel B. (D) Molecular models of CSpV1, ScPV1, ScPV2, and ScPV3 RdRPs generated by comparison to poliovirus 1 RdRP (PDB:1ra6). Red, fingers domain; Blue, thumb domain; White, palm domain. (E) A superimposition of the RdRP conserved catalytic motifs A-F of poliovirus 1, CSpV1, ScPV1, ScPV2, and ScPV3. Red, motif A; Blue, motif B; Green, motif C; Purple, motif D; Cyan, motif E; Yellow, motif F.

https://doi.org/10.1371/journal.ppat.1011418.g005

To determine the tertiary structure of the CP monomer of ScPVs, ab initio models were generated using AlphaFold2. The N-terminus of all ScPV CPs and the Cryspovirus CSpV1 were modeled with confidence of up to 80% that sharply declined after ≈150 amino acids (S4 Fig). Energy minimization improved the models resulting in an average molprobity score of 2.2 Å and an average of 81.7% Ramachandran favored residues (S5 Fig and S8 Table). The N-terminal domain of the ScPV CP was mostly α-helical and with a long α3-helix cradled by four shorter α-helices, (α2, α4, α5, and α6) forming a rhomboid shell domain (S-domain) prototypical of PVs (Fig 5B). DALI analysis of the ScPV CP revealed structural similarity to the CP of pepper cryptic virus 1 (PCV1; PDB:7ncr) with an average root mean squared deviation (RMSD) of 4.34 Å (Fig 5C). Superimposition of the ScPV CP in the context of a virus particle from PCV1 showed that the ScPV S-domains likely form the majority of the particle structure (S6 Fig). Due to the lack of a confident modeling prediction for the C-terminal domain, it was unclear whether ScPVs encode P-domains.

The RdRPs of ScPVs share structural homology with the polymerases of the Picornaviridae

The same molecular modeling approach as performed with the ScPV CPs was also applied to predict the tertiary structure of the RdRPs of ScPVs and CSpV1, with all models having 95% confidence across the entire sequence (S4 Fig). Energy minimization resulted in an average 93.5% Ramachandran favored residues and an average molprobity of 1.20 Å (S5 Fig and S8 Table). The tertiary structure organization of the modeled structures included the characteristic fingers, palm, and thumb domains, with a tertiary structure similar to human poliovirus RdRP (PDB:2ijd-1 and PDB:1ra6, average RMSD <3.1 Å) (Fig 5D). As expected, motifs A-F aligned well between ScPVs, CSpV1, and poliovirus (Fig 5E and S7 Fig). The close similarity of these RdRPs supports the common ancestry of the Partitiviridae and the Picornaviridae and their membership within the phylum Pisuviricota.

ScPVs are maintained in S. cerevisiae under laboratory conditions and inherited in a non-mendelian manner after sporulation

To leverage the power of S. cerevisiae as a model system to study PVs, it was important to determine whether these viruses are stably maintained under laboratory conditions. Four strains of S. cerevisiae containing ScPVs were maintained on agar plates incubated at either room temperature or 4°C for 6 weeks. The presence of ScPVs was monitored by the extraction of dsRNAs and indicated that ScPVs were stably maintained (S8 Fig). PVs have been shown to transmit inefficiently during spore formation in certain fungal species [31], which would limit genetic approaches to studying virus-host interactions in S. cerevisiae. To determine whether ScPVs can segregate efficiently to daughter cells during meiosis, sporulation was induced in two diploid strains infected with ScPVs. Total nucleic acids were extracted from each haploid progeny that were assayed for the presence of ScPVs by RT-PCR and by extracting dsRNAs (S8 Fig). ScPVs were identified in all four meiotic progenies. As ScPVs segregated in a non-Mendelian fashion, virus-containing strains were mated with the laboratory strain S. cerevisiae BY4741 to create ScPV-infected hybrid strains with the expected genotype (S9 Fig).

ScPVs are resistant to the antiviral effects of Xrn1p

Prior studies have shown that pathways of RNA metabolism in yeasts are important for controlling the replication of totiviruses [9,53–58]. Specifically, the 5’ exoribonuclease Xrn1p cures dsRNA totiviruses when overexpressed, likely because of the degradation of uncapped viral RNAs. To determine if ScPVs are susceptible to attack by Xrn1p, a high copy plasmid encoding XRN1 was used to transform a strain of S. cerevisiae harboring totiviruses (L-A-lus and L-BC), an M2 satellite dsRNA, and ScPV3. After transformation, colonies were assayed for the loss of killer toxin production as an indication of Xrn1p restriction, as described previously [56]. The expression of XRN1 led to 50% of clones losing K2 killer toxin expression (S10 Fig), which was comparable to the loss of K1 expression (57%) [56]. Extraction of dsRNAs from non-killer yeasts revealed the expected loss of M2 in all strains, whereas killer yeast always maintained M2. Conversely, dsRNAs from ScPV3 and totiviruses were still observable, even after the loss of M2 (S10 Fig). This demonstrated that ScPV3 appears resistant to the overexpression of XRN1, which is effective at curing the L-A totivirus [56,57].

Double-stranded RNA extraction

Double-stranded RNA was purified according to the rapid method of Crabtree et al. [45]. The yield was increased by recovering the supernatant from multiple phenol-chloroform extractions and loading it onto a single cellulose column. All strain information and their dsRNA content are compiled in S3 Table.

Sequencing of dsRNAs

Samples were prepared for sequencing by adding 3’ poly(A) tails to purified dsRNAs and used as priming sites for cDNA synthesis, as previously described [45,46]. DNAs were prepared for Illumina sequencing using tagmentation [84], and sequencing of cDNAs was carried out according to Crabtree et al. [45]. For determining the sequence of the ends of PV genomes, the 5’ RACE kit was used from Invitrogen as per the manufacturer’s instructions. Completing the ScPV1 RdRP ORF required the following modifications to the manufacturer’s protocol: Superscript IV reverse transcriptase (Invitrogen) was used instead of Superscript II. A longer GSP1 was designed to suit the higher temperatures of cDNA synthesis. And the 37°C TdT incubation time was decreased to 2.5 min. Primers used to amplify the 5’ ends of the segments of the various PVs are listed in S9 Table. 5’ RACE products were cloned into the pCR8 vector using the TOPO-TA cloning kit (Thermo Fisher Scientific). All complete genome sequences of ScPVs have been deposited to NCBI Genbank under the accession numbers OP555746 (ScPV2-858 dsRNA2; CP), OP555747 (ScPV2-858 dsRNA1; RdRp), OP555748 (ScPV1-5509 dsRNA1; RdRP), OP555749 (ScPV1-5509 dsRNA2; CP), OP555750 (ScPV3-1172 dsRNA1; RdRP), and OP555751 (ScPV3-1172 dsRNA2; CP). All ScPV sequence accession numbers are listed in S10 Table.

Sequencing analysis of dsRNA metagenomic data from S. cerevisiae

109 contigs (>300 nt in length) were identified as matching "Viruses (taxid: 10239)" using BLASTx (Max targets 10, Threshold 1 x 10⁻⁶). Each hit was classified by virus family, and the coverage scores were totaled.

RNAse III digestions

Double-stranded RNAs were digested with ShortCut RNase III (New England Biolabs, Ipswich, MA, USA) according to the manufacturer’s protocol.

Total nucleic acid extraction from S. cerevisiae

Total nucleic acids were extracted from saturated liquid cultures of yeasts for RT-PCR as described previously [85].

Reverse transcription-PCR

Superscript IV (Invitrogen) was used for all RT-PCR reactions as directed by the manufacturer. Primers used to detect PVs are listed in S8 Table.

Killer phenotype assays

Killer toxin production by yeasts was measured using killer yeast agar plates (yeast extract, peptone, and dextrose (YPD) agar plates with 0.003% w/v methylene blue buffered at pH 4.6), as described previously [46].

Cycloheximide curing of satellite dsRNAs

10 μL of an overnight YPD culture of S. cerevisiae strain CYC1172 was transferred to YPD plates containing up to 1.5 mg/L of cycloheximide and incubated at ambient temperature for 3–5 days. Colonies growing on the highest concentration of cycloheximide were streaked to single colonies on standard YPD media and incubated at ambient temperature for 3–5 days. Ten colonies were assayed for the loss of the killer phenotype as described under “killer phenotype assays.”

Phylogenetic analysis

RdRP protein sequences from ScPVs and other PVs from previously described genera (S7 Table) were aligned using MUSCLE and were inspected manually for accuracy. PROTTEST3 v3.4.2 was used to determine each dataset’s appropriate amino acid substitution matrix [51]. The search space used consisted of all possible matrices and decorations with eight categories for the +I and +I+G models. The best-fit model was determined to be VT+I+G+F, with the next best model having a delta AIC of 28.52. PhyML v3.3.20220408 was used to create phylogenetic models [50]. PhyML parameters were chosen based on the results of PROTTEST3. The analysis consisted of 1000 bootstrap replicates, the VT amino acid substitution model was used, the proportion of invariable sites was determined by maximum likelihood, amino acid frequencies were estimated empirically by the frequency of occurrence in the dataset, the BEST tree search method was used for tree estimation, and the analysis was started with five random trees.

Virus particle purification

This protocol was adapted from that of Naitow et al. [86]. Cells were inoculated in 1000 mL yeast peptone dextrose (YPD) liquid media and incubated at 30°C for 24 hours. Cells were harvested by centrifugation for 5 min at 3800 × g and washed with double-deionized water. Washed cells were suspended in spheroplasting buffer (100 mM Tris-HCl, 1M D-sorbitol, 20 mM β-mercaptoethanol, 0.5 mg/mL Zymolyase 20T, pH 7.6) and incubated for 90 min at ambient temperature, stirring gently. Digested cells were harvested by centrifugation for 5 min at 3,800 × g and suspended in buffer A (50 mM Tris-HCl, 150 mM NaCl, 5 mM EDTA, 1 mM dithiothreitol, pH 7.6), after which they were lysed by passing the cell suspension twice through a French pressure cell press (Thermo Scientific) at 20,000 psi. Cell debris was removed by centrifugation for 20 min at 9,600 × g, and the supernatant was adjusted to 3% PEG-8000 and 0.5 M NaCl before rocking gently for 60 min at 4°C. The precipitate was harvested by centrifugation for 20 min at 9,600 × g, gently suspended in 15 mL of 50 mM sodium phosphate buffer (pH 7.0), and incubated on ice overnight with rocking. Insoluble precipitates were removed by centrifugation at 7,000 × g for 10 min. The supernatant was layered onto 5 mL 20% sucrose cushion in an ultracentrifuge tube (Thermo Scientific, Nalgene High-Speed Round-Bottom PPCO Centrifuge Tubes) and centrifuged at 80,000 × g for 2 hours at 4°C. The pellet was recovered and suspended in 500 μL sodium phosphate buffer and layered onto a discontinuous sucrose gradient (10–50% in 5% increments of 1 mL each). This was centrifuged at 70,000 × g for 2 hours at 4°C.

Transmission electron microscopy

For the negative staining of VLPs, 5 μl of purified particles from strain YO1126 were applied to a 400-mesh copper grid coated with carbon-Formvar (Ted Pella) for 20 seconds, wicking with a filter paper, and washed three times on 50 μl water droplets. A staining solution of 2% uranyl acetate (pH ≈ 3) was added to the sample-side of the grid, wicking, and reapplied for 20 seconds before removal by wicking and lateral aspiration. Images were obtained on an FEI Tecnai F20 transmission electron microscope, captured with a BM UltraScan camera, and stored in digital micrograph 3 format.

Molecular modeling

Collabfold version 1.4 was used to generate models using the open reading frame of the CP and RdRP of viruses CSpV1, ScPV1-5509, ScPV2-858, ScPV3-1172 with the following parameters (msa_mode: MMseqs2 (uniref+environment); model_type: auto; pair_mode: unpaired+paired; num_recyles: 3). Five unrelaxed structures of each CP and RdRP were generated. RdRP models were generated with similar local distance difference test (pLDDT) values per residue (S4 Fig). The highest average pLDDT score for each model was chosen for energy minimization and analysis. Energy minimization was carried out for each model using standard energy minimization protocol described in our previous study [17]. A dodecahedron box was generated around each model and solvated using a 10 Å layer of TIP3P. Ions were added to maintain charge neutrality at a concentration of 0.15 mol/L using Na⁺ and Cl^- ions. The force field parameter used for protein and ions was AMBER99SB*-ILDNP. Each system underwent energy minimization using the steepest descent algorithm for 10,000 steps via the GROMACS package. Stereochemical analysis was carried out with the SWISS-MODEL structure assessment tool for each model after the energy minimization (https://swissmodel.expasy.org/) to verify that model sterics were physically reasonable after energy minimization. All PDB formatted files for each relaxed model can be found in S1 File. The final energy minimized RdRP and CP model structures were submitted to the DALI server to identify structural homologs (DALI data presented in S2 File). The cealign command in the PyMOL visualization software package was used to identify regions of homology between CPs and homologous CP structures identified by DALI [87].

Yeast tetrad dissection

S. cerevisiae cells were grown overnight at 30°C in 2 mL YPD broth. 250 μL were added to 2 mL spo media (0.25% yeast extract, 0.25% dextrose, 1.5% potassium acetate, 0.002% (histidine, leucine, lysine, tryptophan, methionine, arginine), 0.004% (adenine, uracil, tyrosine), 0.01% phenylalanine, 0.035% threonine, filter-sterilized). The sporulation rate was monitored by microscopy. When the sporulation rate was at least 25%, 50 μL of cell suspension was centrifuged, and the supernatant was removed. Cells were suspended in 50 μL 1 mg/mL yeast lytic enzyme (A. G. Scientific) and digested for precisely 12 min, after which the reaction was quenched by adding 0.5–1 mL sterile water. Asci were dissected with a 25 μm fiber optic cable glued to a micromanipulator mounted onto a Nikon Eclipse 50i microscope. Ascospores were germinated at 30°C, grown to single colonies, and used for additional analysis.

Transmission of ScPVs

Stable haploids of strains YO1126 and CYC1172 were obtained by HO disruption and sporulation to create the diploid strains YTag006 and Ytag015, respectively. Primers NTT033 and NTT034 were used to amplify the ho::HygMX cassette from strain YJM975 MAT⍺ ho::HygMX, using purified chromosomal DNA as a template. 1 μg of the amplicon was used to transform YO1126 and CYC1172. After selection on hygromycin (0.3 mg/mL), PCR verification was performed by probing HO with the same primers. HO/ho::HygMX diploid cells were sporulated and dissected. The dissection plate was replica-plated onto YPD containing hygromycin. Mating type was determined by PCR, using primers NTT041, NTT042, and NTT043. This procedure created the stable haploid strains YTag009 and YTag058 from YTag006 and Ytag015, respectively. YTag009 and YTag058 were each crossed with BY4741 using a method from the Meneghini laboratory (personal communication). Briefly, YTag009 was mixed with BY4741 in a ratio of ≈ 1:100 and spotted onto YPD agar and incubated at 30°C overnight. Cells were streaked to single colonies on YPD plates on which had been spread 125 μL 1 mg/mL alpha factor pheromone and incubated at 30°C for 1–2 days; this process was repeated once more. Diploids were verified by PCR, using primers NTT041, NTT042, and NTT043. This method created YTag121 and YTag085. ScPV infection was verified by RT-PCR and dsRNA extraction followed by agarose gel electrophoresis.

ScPV stability

Yeast strains were streaked onto two YPD agar plates each and incubated at either 25°C or 4°C for 6 weeks. Cells were cultured in 12 mL YPD broth and incubated at 25°C for 2 days, and their dsRNAs were extracted and analyzed by agarose gel electrophoresis.

Halo mating type assays

Strains DBY7730 and DBY7442 were streaked onto YPD agar and incubated at 30°C for 1–2 days. One colony from each was used to inoculate overnight cultures in YPD broth, shaken at 30°C. The cell suspensions were diluted 1:20, and 400 μL of each was spread onto a 100 mm YPD agar plate and incubated at 30°C for 30 min. Strains of interest were pinned onto the lawn using sterile toothpicks, and the plates were incubated overnight at 30°C. A zone of growth inhibition surrounding a colony indicated that it is the opposite mating type of the lawn strain.

Curing of viruses and dsRNA satellites from S. cerevisiae by XRN1 expression

The strain S. cerevisiae YTag085 (an M2 killer yeast) was transformed with the high copy plasmid pPAR219 to express XRN1 cloned from S. cerevisiae [56]. The loss of the killer phenotype in 183 clones was confirmed by a killer assay using two different K2-susceptible strains of yeast (BY4741 and DBY7730). Cured clones were then assayed for dsRNAs by cellulose chromatography.

Strains of yeast

All strains of yeast used in this manuscript are listed in S11 Table.