| Summary: | Ubiquitin-like proteins (UBLs) have broad activities including modification of proteins, lipids, and tRNA. Analysis of both UBL conjugation sites and characterisation of system wide proteome responses is essential to understanding the impact UBLs have on regulating cellular systems. While mass spectrometry is a powerful analytical technique, identifying UBL isopeptides can be particularly challenging due their low abundance and the spectral complexity of UBL isopeptides with very long tryptic remnants. Furthermore, the complexity of whole proteomes delivers limitations to the throughput and depth of quantitative discovery proteomics. This thesis addresses some of the major technical challenges in UBL research. Biochemical and bioinformatic methodologies have been developed for the selective enrichment proteomics datasets, with a common theme of overcoming proteome complexity. In the first instance, enrichment of diglycine isopeptides has been achieved through a novel method exploiting the polyglycine specificity of the bacterial protease- transpeptidase Sortase A (SrtA). Using a mutant with increased catalytic activity, SrtA mediates biotinylation of diglycine tryptic remnants, and also acts as a specific protease for release of isopeptides from streptavidin for analysis by mass spectrometry. This cost-effective approach to isopeptide enrichment is also applicable to linear N-terminal ubiquitylation. The method is demonstrated to offer greater than 100x enrichment and is exemplified on an in vitro ubiquitylation of MIRO1 by PINK1-activated PARKIN. In contrast to a physical enrichment, a dataset enrichment of isopeptide identifications has been achieved through UBL isotope labelling. In vitro substrate modification with isotopically light and heavy UBL generates a characteristic isotopic doublet enabling isopeptides to be distinguished at the MS1 level. Candidate peptide identity can be assigned using high-resolution precursor mass and complementary MS2- level spectral interpretation with SUMmOn adds further confidence to isopeptide identities. Application to SUMO2 modification of putative substrate RNA guanine-7 methyltransferase (RNMT) revealed widespread SUMOylation at 16 different lysines by UBC9 despite lacking a consensus motif. Finally, a quantitative proteomics workflow is presented that enriches whole proteome datasets by combining peptide identifications from unlabelled and SILAC proteomes. A software implementation, with additional tools for improved data quality management, is demonstrated to significantly improve proteome coverage and quantitative precision in unfractionated proteomes. An exemplification on elp3Δ/wt unfractionated yeast proteome reveals a subset of ELP-dependent uridine-34 tRNA modifications to be particularly important for efficient translation. Interestingly, the three mcm5s2U tRNAs which are co-modified by the UBL URM1, have a much greater impact on protein translation efficiency than the ELP-only modified mcm5U tRNAs.
|
|---|
