Unraveling the complexity of transcriptome in Ascosphaera apis spore: based on third-generation and next-generation sequencing

Abstract:

[Objective] Ascosphaera apis is a lethal fungal pathogen of honey bee larvae. This study aims to identify and analyze alternative splicing (AS) and alternative polyadenylation (APA) of genes and long non-coding RNAs (lncRNAs) in A. apis spore (AaS) by PacBio single molecular real-time (SMRT) sequencing, further unveiling the complexity of AaS's transcriptome. [Methods] AS events of genes in AaS were identified with Suppa, and AS events of various types were confirmed based on RT-PCR. TAPIS pipeline was utilized to explore APA sites of AaS genes. MEME was employed to investigate statistics of sequences at 50 bp upstream of poly(A) splicing sites and identify the motifs. The lncRNAs were predicted based on CPC, CNCI, and Swiss-prot database, and the intersection was regarded as lncRNA dataset. Then the transcript length, exon number and length, intron length, GC content, and AS event number were compared between lncRNAs and mRNAs. [Results] A total of 2 609 AS events were identified in AaS, including 1 227 (47.03%) retained introns (RI), 842 (32.27%) alternative 3ʹ splice sites (A3), 415(15.91%) alternative 5ʹ splice sites (A5), 85 (3.26%) alternative first exons (AF), 35 (1.34%) skipping exons (SE), 4 (0.15%) alternative last exons (AL), and 1 (0.04%) mutually exclusive exon (MX). The reliability of various AS events such as RI, A5, SE, and A3 was validated via RT-PCR. Additionally, 5 552 genes containing APA sites were identified, among which genes with more than 5 APA sites were the most abundant (2 197), followed by genes with 1 APA site (1 149). The numbers of genes with 2, 3, 4, and 5 APA sites were 782, 596, 477, and 351, respectively. Moreover, upstream and downstream sequences of full-length transcripts in AaS had apparent base bias, with U and A respectively enriching at upstream and downstream. In total, 953 lncRNAs were identified, including 247 bidirectional lncRNAs, 171 long intergenic RNAs, 154 anti-sense lncRNAs, 141 sense lncRNAs and 9 intronic lncRNAs. Compared with mRNAs, these lncRNAs had fewer and shorter exons, shorter introns, shorter transcripts, lower GC content and fewer AS events. [Conclusion] In this study, we analyzed the AS, APA, and lncRNA in AaS. The findings enrich the basic biological information of A. apis, unravel the complexity of transcriptome in AaS, and lay a foundation for further exploration of the molecular function of various splicing isoforms in pathogenic spores and during pathogen infection.