Systematic characterization of site-specific proline hydroxylation using hydrophilic interaction chromatography and mass spectrometry

Abstract

AbstractTo improve understanding of the role of site-specific proline hydroxylation in controlling protein function, we have developed a robust workflow for the identification of proline hydroxylation sites in proteins using a combination of hydrophilic interaction chromatography (HILIC) enrichment and high-resolution nano-Liquid Chromatography-Mass Spectrometry (LC-MS). Using this approach, together with refining and filtering parameters during data analysis, by combining the results from cell lines being treated with either the prolyl hydroxylase inhibitor Roxadustat (FG-4592, FG) or the proteasome inhibitor MG-132 (MG), or DMSO, a total of 4,993 and 3,247 proline hydroxylation sites were identified in HEK293 and RCC4 samples, respectively. A subset of 1,954 and 1,253 high confident proline hydroxylation sites (non-collagen) from HEK293 and RCC4 samples were inhibited by FG-4592 treatment. A set of features characteristic of proline hydroxylated peptides were identified in both datasets, which differ from either unmodified peptides, or oxidised peptides. Peptides containing hydroxyproline were enriched in the more hydrophilic HILIC fractions and showed characteristic differences in charge and mass distribution, as compared with unmodified or oxidised peptides. Furthermore, we discovered that the intensity of the diagnostic hydroxyproline immonium ion was greatly influenced by parameters including the MS collision energy setting, parent peptide concentration and the sequence of adjacent amino acids neighbouring the modified proline. We show using synthetic peptides that a combination of retention time in LC and optimised MS parameter settings allows reliable identification of proline hydroxylation sites in peptides, even when multiple prolines residues are present. By matching all the proline hydroxylated, non-collagen proteins to the Pfam database, the most common protein family domains identified in both HEK293 and RCC4 datasets were RNA recognition motif (RRM_1), WD40 repeat (WD40), and protein kinase domain (Pkinase). Sequence analysis of the hydroxylated peptides showed enrichment for the motif GxPGxx, including the Gxx repeats found in collagen proteins, as well as the protein kinase domain GTP motif. Glycine (G), serine (S) and glutamic acid (E) residues were found frequently in the sequence window from the hydroxylated peptides. Reactome pathway analysis for the proteins of these newly identified proline hydroxylation sites (FG inhibited), showed enrichment for proteins involved in metabolism of RNA, mRNA splicing and cell cycle regulation, potentially mediated by prolyl hydroxylase enzymes (PHDs).