A Rapid Method for Sequencing Double-Stranded RNAs Purified from Yeasts and the Identification of a Potent K1 Killer Toxin Isolated from Saccharomyces cerevisiae

Abstract

Mycoviruses infect a large number of diverse fungal species, but considering their prevalence, relatively few high-quality genome sequences have been determined. Many mycoviruses have linear double-stranded RNA genomes, which makes it technically challenging to ascertain their nucleotide sequence using conventional sequencing methods. Different specialist methodologies have been developed for the extraction of double-stranded RNAs from fungi and the subsequent synthesis of cDNAs for cloning and sequencing. However, these methods are often labor-intensive, time-consuming, and can require several days to produce cDNAs from double-stranded RNAs. Here, we describe a comprehensive method for the rapid extraction and sequencing of dsRNAs derived from yeasts, using short-read next generation sequencing. This method optimizes the extraction of high-quality double-stranded RNAs from yeasts and 3' polyadenylation for the initiation of cDNA synthesis for next-generation sequencing. We have used this method to determine the sequence of two mycoviruses and a double-stranded RNA satellite present within a single strain of the model yeast Saccharomyces cerevisiae. The quality and depth of coverage was sufficient to detect fixed and polymorphic mutations within viral populations extracted from a clonal yeast population. This method was also able to identify two fixed mutations within the alpha-domain of a variant K1 killer toxin encoded on a satellite double-stranded RNA. Relative to the canonical K1 toxin, these newly reported mutations increased the cytotoxicity of the K1 toxin against a specific species of yeast.

Keywords: dsRNA; killer toxin; mycovirus; sequencing; totivirus.

Figure 1

The extraction of dsRNAs from dried and actively growing cultures of killer and non-killer Saccharomyces yeasts. (A) Killer toxin production by different strains of Saccharomyces yeasts. The ability of yeasts to inhibit the growth of the lawn strain Saccharomyces bayanus CBS7001 indicate killer toxin production by strains Y8.5, BJH001, and EC-1118. (B) dsRNAs were separated by 0.8% agarose gel electrophoresis and stained with ethidium bromide. Molecular weight standards are DNA-specific and provide an approximate size of dsRNAs. Larger dsRNAs represent putative totiviruses, whereas the smaller heterogenous dsRNAs are strain-specific satellite dsRNAs.

Figure 2

Overview of cDNAs synthesis from dsRNAs using poly(A) polymerase for next-generation sequencing. (A) (1) Purified dsRNAs are denatured and rapidly cooled to separate RNA strands. (2) ssRNA is 3′ polyadenylated by poly(A) polymerase and anchored oligo(dT) primers are annealed (N = any nucleotide, V = A, G, C). (3) Reverse transcription is initiated from anchored oligo(dT) primers to create cDNAs that are complementary to both the positive and negative strand of the dsRNAs. (4) RNAs are removed by RNase H digestion followed by the annealing and repairing of cDNAs. (B) The increase in the molecular weight of dsRNAs by the 3′ addition of adenine nucleotides by poly(A) polymerase (PAP) as visualized by intensity trace. (C) High molecular weight cDNA synthesis from dsRNAs extracted from S. cerevisiae by anchored oligo(dT) priming and PCR amplification visualized by intensity trace.

Figure 3

Next-generation sequencing of totiviruses and dsRNA satellites from Saccharomyces cerevisiae strain BJH001. (A) Read quality (phred score) after Illumina QC shown along the length of the sequencing reads when using different concentrations of poly(A) polymerase and oligo(dT) or anchored oligo(dT) primers; 10%, 25%, 75%, and 90% quantile and median (50% quantile) read quality at each position along the reads are shown. (B) Sequence contigs after de novo assembled represented by contig coverage and contig length. BLAST analysis of the four contigs with the longest length and deepest coverage enabled their identification as totiviruses (ScV-L-A1 and ScV-L-BC) and a dsRNA satellite (ScV-M1), the latter was assembled as two separate contigs. Inset reverse transcriptase-PCR was used to confirm the presence of each type of dsRNA. Two primer pairs were used to amplify the ScV-L-BC. (C) Read depth coverage across the reference-assembled ScV-L-A, ScV-L-BC, and ScV-M1 contigs. Open reading frames present within each dsRNA are shown above the nucleotide position.

Figure 4

Natural variation in dsRNAs isolated from S. cerevisiae detected by NGS. A linear representation of the three dsRNAs isolated from the strain BJH001 showing the position of ORFs, relative to observed mutations (A) ScV-L-A1; 4575 bp, (B) ScV-L-BC; 4633 bp, and (C) ScV-M1; ~1700 bp. Bars above the indels represent the indel pairs that result in frameshifts of 18 and 4 amino acids in the Gag-Pol fusion protein. (D) Secondary structure prediction of the K1 killer toxin showing the position of the non-synonymous mutations in the strain BJH001.

Figure 5

Mutations within the K1 gene increases the ability of the killer toxin to inhibit the growth of the yeast Kazachstania africana in vitro. (A) The change in the area of growth inhibition around K1-expressing S. cerevisiae challenged with different strains of K1-sensitive yeasts measured in mm². Asterisks are indicative of a significant difference in the mean zone of the inhibition area (T-test, two-tailed, *** p < 0.01, ns indicates no significant difference). Error bars represent standard error of three independent repeats. (B) Representative images of the isogenic non-killer yeast strains expressing different K1 killer toxins (derived from the K1 reference sequence or K1 from S. cerevisiae BJH001), on agar seeded with yeasts known to be sensitive to K1 killer toxins.