Fusion Transcript Loci Share Many Genomic Features with Non-Fusion Loci

John Lai,Jiyuan An,Inge Seim,Carina Walpole,Andrea Hoffman,Leire Moya,Srilakshmi Srinivasan,Joanna L. Perry-Keene,Australian Prostate Cancer Bioresource,Chenwei Wang,Melanie L. Lehman,Colleen C. Nelson,Judith A. Clements,Jyotsna Batra
DOI: https://doi.org/10.1186/s12864-015-2235-4
IF: 4.547
2015-01-01
BMC Genomics
Abstract:Background Fusion transcripts are found in many tissues and have the potential to create novel functional products. Here, we investigate the genomic sequences around fusion junctions to better understand the transcriptional mechanisms mediating fusion transcription/splicing. We analyzed data from prostate (cancer) cells as previous studies have shown extensively that these cells readily undergo fusion transcription. Results We used the FusionMap program to identify high-confidence fusion transcripts from RNAseq data. The RNAseq datasets were from our (N = 8) and other (N = 14) clinical prostate tumors with adjacent non-cancer cells, and from the LNCaP prostate cancer cell line that were mock-, androgen- (DHT), and anti-androgen- (bicalutamide, enzalutamide) treated. In total, 185 fusion transcripts were identified from all RNAseq datasets. The majority (76 %) of these fusion transcripts were ‘read-through chimeras’ derived from adjacent genes in the genome. Characterization of sequences at fusion loci were carried out using a combination of the FusionMap program, custom Perl scripts, and the RNAfold program. Our computational analysis indicated that most fusion junctions (76 %) use the consensus GT-AG intron donor-acceptor splice site, and most fusion transcripts (85 %) maintained the open reading frame. We assessed whether parental genes of fusion transcripts have the potential to form complementary base pairing between parental genes which might bring them into physical proximity. Our computational analysis of sequences flanking fusion junctions at parental loci indicate that these loci have a similar propensity as non-fusion loci to hybridize. The abundance of repetitive sequences at fusion and non-fusion loci was also investigated given that SINE repeats are involved in aberrant gene transcription. We found few instances of repetitive sequences at both fusion and non-fusion junctions. Finally, RT-qPCR was performed on RNA from both clinical prostate tumors and adjacent non-cancer cells (N = 7), and LNCaP cells treated as above to validate the expression of seven fusion transcripts and their respective parental genes. We reveal that fusion transcript expression is similar to the expression of parental genes. Conclusions Fusion transcripts maintain the open reading frame, and likely use the same transcriptional machinery as non-fusion transcripts as they share many genomic features at splice/fusion junctions.
What problem does this paper attempt to address?