Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Filters applied. Clear all
. 2016 Apr;26(4):499-509.
doi: 10.1101/gr.199877.115. Epub 2016 Mar 2.

Genome-wide A-to-I RNA Editing in Fungi Independent of ADAR Enzymes

Affiliations
Free PMC article

Genome-wide A-to-I RNA Editing in Fungi Independent of ADAR Enzymes

Huiquan Liu et al. Genome Res. .
Free PMC article

Abstract

Yeasts and filamentous fungi do not have adenosine deaminase acting on RNA (ADAR) orthologs and are believed to lack A-to-I RNA editing, which is the most prevalent editing of mRNA in animals. However, during this study with the PUK1(FGRRES_01058) pseudokinase gene important for sexual reproduction in Fusarium graminearum, we found that two tandem stop codons, UA(1831)GUA(1834)G, in its kinase domain were changed to UG(1831)GUG(1834)G by RNA editing in perithecia. To confirm A-to-I editing of PUK1 transcripts, strand-specific RNA-seq data were generated with RNA isolated from conidia, hyphae, and perithecia. PUK1 was almost specifically expressed in perithecia, and 90% of transcripts were edited to UG(1831)GUG(1834)G. Genome-wide analysis identified 26,056 perithecium-specific A-to-I editing sites. Unlike those in animals, 70.5% of A-to-I editing sites inF. graminearum occur in coding regions, and more than two-thirds of them result in amino acid changes, including editing of 69PUK1-like pseudogenes with stop codons in ORFs.PUK1orthologs and other pseudogenes also displayed stage-specific expression and editing in Neurospora crassa and F. verticillioides Furthermore,F. graminearum differs from animals in the sequence preference and structure selectivity of A-to-I editing sites. Whereas A's embedded in RNA stems are targeted by ADARs, RNA editing inF. graminearum preferentially targets A's in hairpin loops, which is similar to the anticodon loop of tRNA targeted by adenosine deaminases acting on tRNA (ADATs). Overall, our results showed that A-to-I RNA editing occurs specifically during sexual reproduction and mainly in the coding regions in filamentous ascomycetes, involving adenosine deamination mechanisms distinct from metazoan ADARs.

Figures

Figure 1.
Figure 1.
Function, expression, and RNA editing of PUK1. (A) Mating cultures of the wild-type PH-1 (WT) and puk1 mutant were examined for cirrhus production (upper) and ascospore release (lower). Arrows point to cirrhi (ascospores oozing) and ascospore masses ejected from perithecia. Bar, 1 mm. (B) Asci and ascospores formed by PH-1 and the puk1 mutant. Deletion of PUK1 affected ascospore morphology. Bar, 20 μm. (C) The expression level of PUK1 in conidia (Coni), 24-h hyphae (Hyph), and perithecia collected at 8 dpf (Peri). The bar chart represents the absolute expression level (log2 FPKM) in RNA-seq data, and the line is the relative expression level (2−ΔΔCt) assayed by qRT-PCR (the expression level of PUK1 in conidia arbitrarily set to one). Error bars indicate standard deviation calculated from two biological replicates of RNA-seq data or three biological replicates for qRT-PCR. (D) The gene structure and editing sites of PUK1. The gene model and coding region of PUK1 is different between automated annotation (predicted) and actual cDNA sequence (observed). Rectangle boxes are coding regions, and the protein kinase domain region is in gray. The corrected gene model contains two tandem stop codons, UA1831G UA1834G (1830–1835, marked with a black vertical line) in its coding region that is part of an intron introduced erroneously by automated annotation. A1831and A1834 in the genomic DNA (gDNA) were changed to G's in cDNA sequences by RNA editing. WebLogo shows the frequency of A-to-G variants at each site in RNA-seq reads.
Figure 2.
Figure 2.
Properties of the A-to-I editing sites in F. graminearum. (A) The number of each type of RNA variant (gDNA → cDNA) sites per million mapped unduplicated reads identified in the two RNA-seq data of 8-dpf perithecia. (B) Histogram and box plot showing the frequency of RNA editing levels. The majority of editing sites have editing levels <30%. (C) The percentage of marked categories of genes that have editing sites of different editing levels. Genes that were specifically expressed (Perithecium-specific) or up-regulated (Perithecium-up) in perithecia were identified by comparative analysis of RNA-seq data of conidia, hyphae, and 8-dpf perithecia (see Supplemental Methods). (D) The distribution of 26,056 A-to-I editing sites. Because only a few genes have known UTRs in F. graminearum, we used the 500-bp region upstream of the start codon and the 500-bp region downstream from the stop codon to represent the 5′- and 3′-UTRs, respectively.
Figure 3.
Figure 3.
Functional consequences of the A-to-I editing sites in F. graminearum. (A) The percentage of editing events resulting in different types of changes in protein sequences or coding regions. The stop change category includes stop-loss and stop-retained editing events. (B) The number of genes with different numbers of recoding A-to-I editing events. (C) Box plots showing the editing levels of RNA editing sites with different types of functional consequences. The statistical significance (t-test) for each comparison is indicated: (****) P < 0.0001; (**) P < 0.01. (D) Percentage of missense A-to-I editing events resulting in different types of amino acid changes. (E) Numbers of genes with marked stop-loss or stop-retained RNA editing events. The nucleotides subjected to RNA editing are in bold.
Figure 4.
Figure 4.
Sequence and structure preferences of the A-to-I editing sites in F. graminearum. (A) Two Sample Logo showing the enriched (above the top line) and depleted (below the bottom line) nucleotides nearby the A's targeted for RNA editing (P < 0.01, t-test), with the level of preference or depletion proportional to the scale. A total of 30,000 adenosine sites randomly chosen from predicted cDNA sequences were used as the negative control. (B) The percentage of marked triplet sequences with A-to-I editing events. For each of the 16 possible triplets centered on the edited adenosine (NAN), the number of observed editing events was divided by its total occurrence in cDNA sequences. The horizontal dotted line marks the average percentage (0.56%) of editing events observed in these 16 NAN triplets. (C) Box plot comparing the editing levels of editing sites in different triplets: (****) P < 0.0001, t-test; (ns) not significant. (D) Stacked column showing the ratio of RNA editing events of marked editing levels in the five types of RNA secondary structure elements diagrammed on the right. The predicted RNA secondary structure is based on 30-nt upstream and 30-nt downstream sequences surrounding the edited A's. The statistical significance for hairpin loop ratio comparison is indicated: (****) P < 0.0001, χ2-test; (ns) not significant. (E) Box plot showing the minimum free energy (MFE) of predicted hairpin loops with A-to-I editing events of different editing levels. The statistical significance for each comparison is indicated: (****) P < 0.0001, t-test; (ns) not significant.
Figure 5.
Figure 5.
Evolution, expression, and function of ADATs in F. graminearum. (A) Phylogenetic tree of deaminase domain of fungal ADATs and ADARs constructed using PhyML3.1 (Guindon et al. 2010). The SH-like support of approximate likelihood ratios (aLRT-SH) is plotted as circles on the branches (only SH-like support >0.6 are shown). The prefixes for gene names or IDs are as follows: (ANID) Aspergillus nidulans; (FGRRES) Fusarium graminearum; (Hs) Homo sapiens; (NCU) Neurospora crassa; (Sc) Saccharomyces cerevisiae; (Sp) Schizosaccharomyces pombe. The domain structures are as follows: (A_deamin) adenosine-deaminase (PF02137); (dCMP_cyt_deam_1) cytidine and deoxycytidylate deaminase zinc-binding region (PF00383); (dsrm) double-stranded RNA binding motif (PF00035); (PRK) phosphoribulokinase (PF00485); (z-alpha) adenosine deaminase z-alpha domain (PF02295). (B) The expression level (fragments per kilobases of exons for per million mapped reads [FPKM]) of three ADAT genes of F. graminearum estimated with RNA-seq data of conidia, 24-h hyphae, and perithecia collected at 8 dpf. Error bars indicate standard deviations calculated from two biological replicates of RNA-seq data. (C) Mating cultures of the wild-type PH-1 (WT) and Fgtad1 deletion mutant were examined for ascospore release and cirrhus production. Arrows point to ascospore masses and cirrhi (ascospores oozing) ejected from perithecia. Bar, 1 mm. (D) Asci and ascospores formed by PH-1 and the Fgtad1 mutants were examined with 12-dpf perithecia. Bar, 20 μm. (E) Sequencing traces for the edited region of FgSSN3 (FGRRES_04484) amplified from RNA isolated from perithecia of PH-1 and Fgtad1 mutant. Black arrows mark the edited A's that have a mixed peak of A and G in sequencing traces.
Figure 6.
Figure 6.
A-to-I RNA editing in Neurospora crassa and Fusarium verticillioides. RNA editing in transcripts of the PUK1 orthologs in N. crassa (A) and F. verticillioides (B). The gene structure based on automated annotation (upper) differs from the actual cDNA sequence (lower) identified by RNA-seq analysis. Rectangle boxes are coding regions, and black arrowheads indicate the direction. In N. crassa, NCU03242 is specifically expressed in perithecia, and its UA1628G stop codon (marked with a black vertical line) corresponding to UA1831G of PUK1 was edited to UGG. In F. verticillioides, RNA editing of UA1836G in FVEG_01191 (black vertical line) corresponding to UA1834G of PUK1 was also only observed in perithecia. (C) Numbers of labeled RNA variant sites identified in RNA-seq data of perithecia in F. verticillioides.

Similar articles

See all similar articles

Cited by 27 articles

See all "Cited by" articles

Publication types

Feedback