Work Description

Date: 05 January 2024

Dataset Title: Giant viruses in early-diverging fungi: Mycodnaviridae

Dataset Creators: J. Myers and F. Schulz

Dataset Contact: Jillian Myers (jimyers@umich.edu)

Funding: DEB-1929738 (NSF), DE-AC02-05CH11231 (DOE).

Research Overview:
Giant viruses of the phylum Nucleocytoviricota have captured researchers' attention due to their increasingly recognized impacts on eukaryotic genome evolution. Their origins are hypothesized to predate or coincide with the diversification of eukaryotes and their natural hosts span the eukaryotic tree of life. But surprisingly, giant viruses have not been found in the fungal kingdom, though recent metagenomic work suggests a putative association. Here we show that early-diverging fungal lineages maintain both "viral fossils" and host active infections by giant viruses, which form a monophyletic clade that we name Mycodnaviridae. Viral genomes described here span up to 350kb and encode over 300 genes, including many genes characteristic of giant viruses. Interestingly, we observed variation in infection status among the isolates including apparent active infection and transcriptionally-suppressed states, suggestive that viral activation may be constrained to certain life stages of the host. Our experimental findings add to the scant few natural virus-host systems available in culture for the study of giant viruses. These viruses have likely shaped the early evolution of these fungal lineages and should prove a useful model for their study.

Methodology:
Fungal genomes prepared for a previous sequencing effort were mined for giant virus genomes. Mycodnaviridae dataset includes the five most complete viral genomes, as identified as follows:

We used the program ViralRecall (Aylward and Moniruzzaman 2021) to identify putative viral genes, viral contigs, and viral regions within contigs. We first ran ViralRecall 1.0 on all assemblies with both stringent (-s 15 -m 30 -v 10) and relaxed (-s 15 -m 2 -v 2) parameters–the latter for poorer quality genome assemblies–and identified isolates containing MCP homologs with e-value <1e-10. ViralRecall searches gene predictions generated using Prodigal (Hyatt et al. 2010); we also performed hmmsearches (HMMER3; S. Eddy 2009)with the ViralRecall 2.0 GVOG hmm on a subset of our Augustus-predicted genomes. We found no differences in the number of MCPs called, and so continued characterizations for the isolates that these initial ViralRecall results predicted at least one MCP homolog. For each MCP-containing contig, we calculated GC content, intron density per the Augustus annotation, and performed BLAST searches (Altschul et al. 1990) on both gene prediction datasets.

To aid in distinguishing viral genomes from within the fungal assemblies, we searched the prodigal annotated genomes the 16 genomes previously identified to contain MCPs with HMMER3 (Eddy 2009) as in Schulz et al. 2020: using hmms of a subset of the nucleocytoplasmic viral orthologous genes (NCVOGs)— the 20 NCVOGs most likely to have been vertically inherited (Yutin et al. 2009). We also ran the latest version of ViralRecall (2.0; Aylward and Moniruzzaman 2021) on these 16 assemblies and used these results in combination with genome statistics to discern putatively viral contigs, which we classified as belonging to a viral genome if they met the following criteria: ViralRecall score >2, viral region length (per ViralRecall) >70% of the total contig length, and intron density > 25% different from the host average. One isolate appeared to contain two distinct NCLDVs; we used MetaBAT2 (Kang et al. 2019) on the entire genome of Allomyces javanicus California12 under the default settings to separate the two NCLDV genomes.

To build the giant virus species tree the nsgtree pipeline was used (https://github.com/NeLLi-team/nsgtree) on genome representatives of the viral phylum Nucleocytoviricota (Aylward et al. 2021). In brief, 7 giant virus orthologous groups (GVOGs; Aylward et al. 2021) were identified using hmmsearch, extracted GVOGs were aligned with MAFFT (version 7.31; Katoh and Standley 2013), trimmed with trimal (-gt 0.1, v1.4; Capella-Gutiérrez, Silla-Martínez, and Gabaldón 2009) and concatenated. To limit the amount of missing data we removed genomes that had less than 4 out of 7 GVOGs. A species tree was built from the supermatrix alignment using IQ-Tree (version 2.03); (Minh et al. 2020; Kalyaanamoorthy et al. 2017) with LG+I+F+G4 and ultrafast bootstrap replications (Hoang et al. 2018) and visualized in iToL (Letunic and Bork 2021).:

Files contained here:
-Nucleotide fasta files for each of the 5 most complete giant virus genomes are given as "virus_name.fna". Viruses are named "Host_fungus_mycodnavirus_1"
-Files “latest_ncldv_Nov2023-frameworks-GVOG7-iqtree-perc5_red.ext” are the alignment (mafft ext) and tree (contree ext).

Related publication(s):
Myers, J., Schulz, F., Rahimlou, S., Amses, K., Simmons, D.R., Stajich,J.E., and James, T.Y. 2024. “Large DNA Viruses in Early Diverging Fungal Genomes Are Relics of Past and Present Infections.” bioRxiv, January, 2024.01.04.574182. https://doi.org/10.1101/2024.01.04.574182.

Use and Access:
This data set is made available under a CC-BY-NC 4.0 International license.

To Cite Data:
Myers & Schulz. 2024. Giant viruses in early-diverging fungi: Mycodnaviridae [Data set]. University of Michigan - Deep Blue. https://doi.org/10.7302/p5ay-ng61