Genome-wide and subcellular nonsense mediated decay
Nonsense mediated decay (NMD) is a translation-dependent RNA quality control pathway that selectively degrades aberrant transcripts before they lead to truncated proteins. NMD causes rapid degradation of target RNAs by assembling multiple factors to drive poly-deadenylation and decapping of the mRNA. NBAS (neuroblastoma amplied sequence) protein was identied to be required for NMD in an RNAi screen in Caenorhabditis elegans. Subsequent evidence has suggested it is a conserved NMD factor in zebrash and human cells. NBAS is localised to the endoplasmic reticulum (ER) suggesting it could be involved in a novel branch of the NMD pathway specic to ER. The key NMD factor UPF1 has been shown to partially localise to the ER and interact with NBAS. NMD targets harbour premature termination codons (PTC)s, upstream open reading frames or long 3' UTRs. However not all transcripts with these features activate NMD, and many NMD targets appear not to contain any of these features. Therefore, the previously dened rules regarding initiation of the NMD pathway do not account for all variation seen in NMD eciency. This thesis presents the creation and investigation of a CRISPR cell line containing an endogenously tagged NBAS protein. Immunoprecipitation mass spectrometry on these cells revealed that NBAS interacts with the ER translocon, where the ribosome docks for translation at the ER membrane, suggesting the presence of an NBAS-UPF1-translocon complex. This interaction was conrmed by proximity ligation assay, a microscopy-based approach. In order to identify the eect of NBAS on transcript expression and stability, RNA-sequencing was performed on total, nascent, cytoplasmic, and membraneassociated fractions, upon siRNA depletion of either NBAS or UPF1 mRNA. Differential expression analysis revealed a correlation in dierentially expressed transcripts when NBAS or UPF1 were depleted with an enrichment in ER-associated RNAs specically in NBAS targets. UPF1 was found to aect expression of a similar number of transcripts in the cytoplasmic and membrane-associated fractions conrming its role in NMD at the ER. Whereas depletion of NBAS led to a 5-fold higher proportion of membrane-associated genes being increased in expression than of non-membrane genes, showing its subcellular specicity. Using both nascent and total RNA-sequencing datasets, estimates of dierential transcript stability upon depletion of either factor were made. Approximately 20% of transcripts increased in expression upon depletion of UPF1 were also found to increase in stability, whereas the remaining 80% of expression changes could be explained by changes in transcription, suggesting these were due to secondary eects. Of those transcripts that changed in expression when NBAS was depleted, none were identied as changed in stability in this study, suggesting NBAS has subtle or negligible eect on transcript stability and is therefore likely to be a peripheral factor in the NMD pathway. Finally, whole genome sequencing was used in conjunction with total and nascent expression datasets to identify correlation between genome variants and NMD eciency. I show that it is possible to identify heterozygous variants with allelespeci c processing and measure the eect of NMD to identify direct NMD activity. This approach can be used in the future to systematically explore the rules of NMD targeting in human cells. In sum, this thesis reveals that NBAS interacts with the ER translocon and aects the expression of hundreds of ER-associated transcripts. However, NBAS does not appear to regulate transcript stability suggesting its role at the translocon is distinct from NMD. My ndings add to the understanding of NMD quality control at the ER and the genome-wide rules governing NMD eciency in human cells.