RNA-Seq, and ye shall find: Sexual-stage-specific A-to-I RNA editing in fungi.

ABSTRACT

Fusarium graminearum, a pathogen of wheat and barley, is a haploid homothallic ascomycete filamentous fungus (Goswami and Kistler 2004). It overwinters as saprophytic hyphae in plant debris and undergoes the sexual cycle in spring to produce fruiting bodies (perithecia) bearing the progeny ascospores. The genome sequence of the F. graminearum PH-1 strain was reported last year (King et al. 2015). This year, Jin-Rong Xu and colleagues in Northwest A&F University, China, and Purdue University, USA, re-sequenced the PH-1 genome and also performed RNA-Seq analysis on two independent biological replicates each of RNA from conidia, hyphae, and 8-day post-fertilization perithecia (Liu et al. 2016). Alignment of the two replicate perithecial RNA-Seq reads with the reference genome sequence revealed 23,041 and 19,764 single-nucleotide variants (SNVs), of which, respectively, 22,578 and 19,261 corresponded to A (adenosine)- to-G (guanosine) transitions, and 17,613 A-to-G transitions were common to both the replicates. Non- A-to-G variants were far fewer (463 and 503) and only 35.9% were common between the two perithecial replicates, suggesting that the non-A-to-G variants were false-positives. In sum, 26,056 A-to-G variants were identified as putative A-to-I RNA editing sites at which hydrolytic deamination at the C6 position of the purine ring of A produces I (inosine). Since I preferentially base-pairs with C (cytidine), an I within a transcript is read as G by the translation machinery. Also, during reverse transcription, I directs the incorporation of C; thus, it appears as a G in double-stranded cDNA. The conidial and hyphal RNA-Seq data showed only 68 and 112 A-to-G transitions and 335 and 452 non-A-to-G conversions, indicating that the A-to-I RNA editing is specific to the sexual stage. Xu and colleagues had initially set out to do a yeast two-hybrid experiment to identify proteins that interact with a protein kinase named Puk1 (perithecium unique kinase 1). Ascospores from the puk1 mutant have an abnormal morphology. Additionally, qRT-PCR showed that PUK1 transcription is markedly upregulated in perithecia, suggesting that PUK1 expression and function might be restricted to the sexual stage. Therefore, to generate the PUK1 ORF bait, they synthesized cDNA using RNA isolated from perithecial cultures of the PH-1 strain. Sequencing of the construct revealed that two tandem stop codons – UAG UAG – in the PUK1 ORF were changed to UGG UGG in the cDNA, presumably via A-to-I RNA editing. More than 90% of PUK1 reads in perithecial RNA-Seq showed the A-to-I editing, and experiments with site-specific mutant alleles showed that the editing was essential for PUK1 function. This was an unexpected discovery because fungi lack orthologs of the Adenosine Deaminase Acting on RNA (ADAR) family of enzymes that in metazoans converts A to I in double-stranded RNA. Presumably, A-to-I RNA editing in fungi uses different enzymes than animals. This discovery motivated Xu and colleagues to expand the search for RNA editing genome-wide in transcriptomes from vegetative and sexual-stage tissues (conidia, hyphae, and perithecia). The percentage of reads with the A-to-G variant was taken as the RNA editing level at the site. Editing levels varied from 3% to 100%. Strikingly, 47% of genes bearing sites with editing levels >60% tended to be up-regulated or specifically expressed in perithecia compared to conidia and hyphae. A majority of the editing events resulted in amino acid substitutions, which suggested that Ato- I editing might be important for adaptation of protein functions during sexual reproduction. Editing events similar to those in PUK1 were found 69 other genes, including the rid (RIP defective) ortholog and genes important for meiosis (see below). All these genes displayed UAG-to-UGG change in exons that automated annotation had incorrectly predicted as introns.