The evolution of novel fungal genes from non-retroviral RNA viruses

Abstract

Background: Endogenous derivatives of non-retroviral RNA viruses are thought to be absent or rare in eukaryotic genomes because integration of RNA viruses in host genomes is impossible without reverse transcription. However, such derivatives have been proposed for animals, plants and fungi, often based on surrogate bioinformatic evidence. At present, there is little known of the evolution and function of integrated non-retroviral RNA virus genes. Here, we provide direct evidence of integration by sequencing across host-virus gene boundaries and carry out phylogenetic analyses of fungal hosts and totivirids (dsRNA viruses of fungi and protozoans). Further, we examine functionality by tests of neutral evolution, comparison of residues that are necessary for viral capsid functioning and assays for transcripts, dsRNA and viral particles.

Results: Sequencing evidence from gene boundaries was consistent with integration. We detected previously unknown integrated Totivirus-like sequences in three fungi (Candida parapsilosis, Penicillium marneffei and Uromyces appendiculatus). The phylogenetic evidence strongly indicated that the direction of transfer was from Totivirus to fungus. However, there was evidence of transfer of Totivirus-like sequences among fungi. Tests of selection indicated that integrated genes are maintained by purifying selection. Transcripts were apparent for some gene copies, but, in most cases, the endogenous sequences lacked the residues necessary for normal viral functioning.

Conclusions: Our findings reveal that horizontal gene transfer can result in novel gene formation in eukaryotes despite miniaturized genomic targets and a need for co-option of reverse transcriptase.

Figure 1

A comparison of the simple genome of Totivirus, a dsRNA virus, with the fungal genome maps containing proposed endogenous Totivirus-like genes and open reading frames. Red arrows indicate capsid polypeptide-like genes and the blue arrows represent RNA-dependent RNA polymerase-like genes. These coloured arrows represent open reading frames). Gray lines represent an RNA based genome and black lines represent fungal chromosomal DNA. Gene copies that retain an exogenous viral genomic architecture are shown as solid coloured arrows with the coding strand indicated by the direction of the arrow. Sequences from the present study that reveal fungal-Totivirus-like genes boundaries are shown above the maps. Asterisks represent further gene boundary regions amplified successfully by polymerase chain reaction (PCR; see Figure 6A). Numbers below the gene annotations indicate the chromosomal positions of gene boundaries. Full sequences of the PCR products from these regions have the Genbank Accession numbers: GQ291318-GQ291321.

Figure 2

A comparison of the fungal genome maps of Penicillium and Candida proposed to contain endogenous Totivirus-like genes. Red arrows indicate capsid polypeptide-like genes and the blue arrows represent RNA-dependent RNA polymerase-like genes. These coloured arrows represent open reading frames. Black lines represent fungal chromosomal DNA. Gene copies that retain an exogenous viral genomic architecture are shown as solid coloured arrows with the coding strand indicated by the direction of the arrow. Numbers below the gene annotations indicate the contig positions of gene boundaries.

Figure 3

Maximum likelihood phylograms of totivirid-like RNA-dependent RNA polymerase genes based on an alignment of amino acid sequences. Bayesian posterior probability and nonparametric bootstrap values >85% are figured above branches. Fungal genomic sequences are highlighted by a blue box. The asterisk represents a putative fungal expressed sequence tag sequence.

Figure 4

Maximum likelihood phylogram of totivirid-like capsid polypeptide (Cp)-like genes based on an alignment of amino acid sequences. Bayesian posterior probability and nonparametric bootstrap values > 85% are figured above branches. Fungal genomic sequences are highlighted with a red box. Within-species copies are numbered as in Figure 1.

Figure 5

Binary rooted tree of capsid protein-like nucleotide sequences with Ka/Ks ratios plotted on branches. The tree is constrained with the assumption of tandem array monophyly. The relationships of gene family members are shown for Pichia stipitis (in red) and for Debaryomyces hansenii (in blue). Genes associated with the ancestral architecture are highlighted with a black oval.

Figure 6

A comparison of Totivirus and endogenous Totivirus-like fungal gene products. Pichia stipitis (lane 1), Saccharomyces cerevisiae with exogenous Totivirus infection (lane 2) and Debaryomyces hansenii (lane 3) (A) polymerase chain reaction (PCR) of total DNA (1.4% agarose gel). (B) Assay for dsRNA (1.4% agarose gel with EtBr staining). (C) Assay for viral particles (photograph of CsCl density gradients). (D) Reverse transcriptase (RT)-PCR of total RNA (1.4% agarose gel). (E) PCR of total RNA extractions treated with DNAse. Here, RT-PCR of Cp is attempted from primers that flank the Cp-RdRp boundary. Reverse transcriptase is marked as either present (+RT) or absent (-RT).

Figure 7

A matrix showing conservation patterns in capsid protein-like genes at residues in Saccharomyces cerevisiae virus L1 (L-A) capsid protein that are structurally important for the decapping function. Shading represents the seven amino acid categories assigned by size and charge in the GDE (Genetic Data Environment) format. Dashes are indels. Yellow-shaded names indicate fungal capsid-protein-like genes that lack conservation of functional residues for decapping (see Figure 1).