Splice isoform-specific analysis of endogenous NMD targets in human cells

Evangelos D. Karousis, from the University of Bern, spoke about using nanopore sequencing to study nonsense-mediated mRNA decay in human cells. He began by giving a brief overview of why RNA degradation is important in both prokaryotic and eukaryotic organisms. He said that it was important for disposing of defective RNA transcripts, protecting the host cells from viral infection, and can aid in regulating gene expression. He further stated that nonsense RNA degradation is important for all three of these processes.

Evangelos briefly touched on why nonsense RNAs can be potentially dangerous, where mutations, splicing or transcription errors cause premature termination codons to be incorporated into the newly-synthesised transcript, leading to malformed, potentially toxic proteins. To prevent a build-up of these potentially toxic proteins, eukaryotic cells have evolved ways of removing these transcripts before cell death can occur. This process, termed nonsense-mediated mRNA decay (NMD), involves a complex pathway. Evangelos focussed on the latter steps of this pathway, where targeted mRNAs are degraded by the activity of SMG6 and SMG7. Evangelos then stated that if SMG6 and SMG7 are depleted, this will result in an accumulation of mRNAs targeted by this pathway. Determining which transcripts are targeted for degradation is important, as many endogenous RNAs that encode full proteins are degraded in this fashion and changing these levels can have large effects on biological processes, ranging from neurodegenerative disorders to stress response.

Previously, Evangelos had used short-read sequencing to gain important information concerning NMD targeted RNAs at the gene level, however this approach fails to resolve things at the isoform level. He gave the analogy of considering the transcriptome to be the “news-stand of life” with RNA molecules being the newspapers. He said that nowadays, researchers can afford to buy the whole news stand, but the caveat is that it has all been put through a shredder first and needs to be pieced back together. Here, long read cDNA sequencing on the Oxford Nanopore platform provides a solution to do this.

Evangelos explained how NMD targeted transcripts could be detected once this activity had been essentially switched off. In essence, those targeted by NMD will not be in a standard annotation as they will have been degraded quickly and not previously sequenced. Therefore, once the NMD activity has been removed, novel isoform populations that differ from those previously seen, can be hypothesized to be targeted by NMD. Using both short read sequencing and Oxford Nanopore cDNA sequencing allows the gaps to be filled and a comprehensive curated NMD transcriptome.

SMG6 and SMG7 were knocked down in HeLa cells and validated by western blot. qPCR was used to validate that a known NMD targeted isoform was still produced under the knock-down conditions. RNA was extracted and polyA mRNA was selected prior to sequencing two replicates of each condition using Oxford Nanopore's Direct cDNA Sequencing Kit on the GridION device. Both short reads previously generated by the lab and these long reads were combined using a program called Squanti, which resulted in an annotation file of known and novel transcripts.

Using the above approach, Evangelos showed a novel transcript that contained an intronic region that is normally skipped in the standard protein coding isoform, making it a target for NMD. qPCR analysis targeting this isoform specifically showed significant upregulation under SMG6 and SMG7 knock-down conditions, providing orthogonal evidence that this transcript was indeed targeted by NMD.

Next, Evangelos spoke about finding isoforms that show differential expression levels under NMD conditions and furthermore differential exon-skipping patterns. Here, Evangelos found 2,247 differential exon-skipping events, 286 of which were novel. By far, the most common alternative splicing events were exon skipping, followed by intron retention; alternative 5’ and 3’ splice sites were rarest.

Using an MA plot, the next question was, to what extent does NMD target specific transcripts at the isoform level? Across all the data, approximately 30% of the genes that showed differential expression patterns also showed differential isoform usage patterns, with the trend being more prevalent in the most abundant isoforms.

Summarising, Evangelos said that: nanopore sequencing can facilitate the identification of so-far unknown splice isoforms that are targeted by NMD in human cells; many of the endogenous mRNAs that are targeted by NMD are specific isoforms of expressed genes; isoform expression levels change, even when gene levels stay the same; and the most common form of splicing that gives rise to NMD is exon skipping.