Novel bioinformatics method for identification of genome-wide non-canonical spliced regions using RNA-Seq data.

Yongsheng Bai,Philip Li,Randal J Kaufman,Rajasree Menon,James D Cavalcoli,Zachary Wright,Maureen A Sartor,Justin Hassler,Gilbert S Omenn,Ahdad Ziyar,Arnab Pain

doi:10.1371/journal.pone.0100864

Abstract

SettingDuring endoplasmic reticulum (ER) stress, the endoribonuclease (RNase) Ire1α initiates removal of a 26 nt region from the mRNA encoding the transcription factor Xbp1 via an unconventional mechanism (atypically within the cytosol). This causes an open reading frame-shift that leads to altered transcriptional regulation of numerous downstream genes in response to ER stress as part of the unfolded protein response (UPR). Strikingly, other examples of targeted, unconventional splicing of short mRNA regions have yet to be reported.ObjectiveOur goal was to develop an approach to identify non-canonical, possibly very short, splicing regions using RNA-Seq data and apply it to ER stress-induced Ire1α heterozygous and knockout mouse embryonic fibroblast (MEF) cell lines to identify additional Ire1α targets.ResultsWe developed a bioinformatics approach called the Read-Split-Walk (RSW) pipeline, and evaluated it using two Ire1α heterozygous and two Ire1α-null samples. The 26 nt non-canonical splice site in Xbp1 was detected as the top hit by our RSW pipeline in heterozygous samples but not in the negative control Ire1α knockout samples. We compared the Xbp1 results from our approach with results using the alignment program BWA, Bowtie2, STAR, Exonerate and the Unix “grep” command. We then applied our RSW pipeline to RNA-Seq data from the SKBR3 human breast cancer cell line. RSW reported a large number of non-canonical spliced regions for 108 genes in chromosome 17, which were identified by an independent study.ConclusionsWe conclude that our RSW pipeline is a practical approach for identifying non-canonical splice junction sites on a genome-wide level. We demonstrate that our pipeline can detect novel splice sites in RNA-Seq data generated under similar conditions for multiple species, in our case mouse and human.

Highlights

The endoplasmic reticulum (ER) is the cellular organelle responsible for protein folding and assembly of approximately one-third of all eukaryotic cellular proteins [1]
We only examined one time point, it was during the peak of IRE1a-dependent Xbp1 mRNA splicing; the lack of additional non-Xbp1 mRNA splice targets is consistent with what was reported in yeast, where HAC1 was found to be the sole splicing substrate of IRE1a [16]
Either the number of reads supporting these splices was much lower than observed for Xbp1, the splice length was likely too small or large to be a true positive, the results were only found in one treatment or the gap was genomic in origin

Summary

Introduction

The endoplasmic reticulum (ER) is the cellular organelle responsible for protein folding and assembly of approximately one-third of all eukaryotic cellular proteins [1]. Upon activation of the UPR, the protein Ire1a (Inositol Requiring Enzyme 1 a); known as ERN1 (endoplasmic reticulum to nucleus signaling 1), initiates the cytosolic (noncanonical) splicing of Xbp (X-box binding protein 1) mRNA to remove a 26 nucleotide (nt) sequence that results in a shift of the translational-open reading frame [7]. Spliced Xbp mRNA encodes a potent transcriptional activator of hundreds of UPR genes encoding functions that facilitate protein folding, secretion, and degradation in response to ER stress. U2and U12- dependent splicing events occur within nuclear premRNAs whereas IRE1a-dependent Xbp mRNA splicing occurs within the cytosol and has its own sequence recognition requirements [9,10], it is non-canonical in the sense of not being U2-dependent; it is uncharacteristic of any classification

Methods

Results

Discussion

Conclusion