Localization of PostTranslational Modifications i

Authors: Matthew A Baird Alexandre A Shvartsburg

Publish Date: 2016/09/19

Volume: 27, Issue: 12, Pages: 2064-2070

PDF Link

Abstract

Precise localization of posttranslational modifications PTMs on proteins and peptides is an outstanding challenge in proteomics While electron transfer dissociation ETD has dramatically advanced PTM analyses mixtures of localization variants that commonly coexist in cells often require prior separation Although differential or field asymmetric waveform ion mobility spectrometry FAIMS achieves broad variant resolution the need for standards to identify the features has limited the utility of approach Here we demonstrate full a priori characterization of variant mixtures by highresolution FAIMS coupled to ETD and the procedures to systematically extract the FAIMS spectra for all variants from such dataMany proteins incorporate posttranslational modifications PTMs that govern critical biological functions 1 2 For example the phosphorylation pattern of human τ protein is thought to control the aggregation of paired helical filaments in Alzheimer’s disease by influencing the cis/trans ratio of prolyl bonds adjacent to T231 and S235 and thereby the protein conformation 3 Phosphorylated sites also have distinct regulatory functions for nonamyloid proteins such as ERK and CDK family members and adapter proteins like SRC and Fak 4 The same PTMs often attach in different locations on the backbone creating “localization variants” that coexist in vivo and have different and even opposite biological activities 5 6 7 8 For instance demethylation of H3 histone is associated with transcription activation when on the K4 residue but with heterochromatin and repression when on the closest lysine K9 6As the technologies for identifying and quantifying primary protein sequences mature the proteomics frontier is moving to the characterization of PTMs and their roles in disease states 1 2 9 That advance is hindered by the openended diversity of PTMs with hundreds discovered to date and substoichiometric modification 9 Many PTMs are labile including perhaps the most biologically consequential—the phosphorylation considered here 2 5 7 8 10 Ergodic methods such as collisioninduced dissociation CID usually eject them from the peptide and/or shift them to another residue before severing the backbone obliterating or falsifying the attachment site information 2 10 Thus comprehensive tandem mass spectrometry MS/MS analysis of PTMs requires a direct fragmentation mechanism such as the electron capture dissociation ECD or electron transfer dissociation ETD that cut the peptide backbone at every residue without abstracting PTMs 2 5 11 12 However MS/MS methods are fundamentally unable to distinguish some variants in a mixture of three or more due to the absence of unique product masses 12 This can in theory be overcome by subsequent fragmentations but unlike with CID in MSn protocols successive ETD application is impractical in view of 1 charge reduction in the first step which produces few or no multiply charged ions that could be precursors for the second step and 2 low ETD yield into each fragment normally ~01–1 because of limited ETD efficiency and nonspecific dissociation along ~10–100 channelsHence peptide variants generally need to be separated prior to the MS/MS step While proteolytic digests are routinely fractionated by liquid chromatography LC prior to MS/MS it often fails to resolve the localization variants especially those with nearby alternatively occupied sites that also produce the most similar fragmentation patterns 7 13A growing substitute to LC is ion mobility spectrometry IMS where ionized compounds are driven through gas by electric field and fractionated depending on the transport properties 14 The original approach of linear IMS is based on the absolute ion mobility K at moderate fields and achieves resolving powers R up to ~200 in the drift tube DT 15 16 and ~50 in the travelingwave TW 17 implementations These techniques can partly separate the phosphopeptide and unmodified peptide domains and resolve some variants 18 19 20 Unfortunately those could not be assigned using ETD the extended times needed for a suitable yield of reaction between the ions and ETD reagent commonly ~100–300 ms greatly exceed the temporal peak width of ion packets upon IMS separation typically ~01–1 ms and thus implementing ETD on the fly after dispersive DT or TW IMS separations would obliterate the established resolution Whereas the IMS resolving power can be augmented somewhat using the cyclotron paradigm 21 to pull apart the variants better that would counteract matching the IMS peak width and the ETD timescaleA newer approach of differential or field asymmetric waveform IMS FAIMS leverages the nonlinear ion motion in fields of high intensity E to sort ions by the mean derivative of KE function over a certain range 22 23 That quantity is elicited employing an asymmetric waveform of some amplitude dispersion voltage DV set up across a gap between two electrodes through which ions are carried by the gas flow Ions injected into the gap are pushed to either electrode but a given species can be balanced and thus pass and be detected by superposing a particular fixed compensation voltage CV on the waveform Scanning CV commonly expressed as the compensation field E C to normalize for the gap width reveals the spectrum of species present As the KE derivative is correlated to ion mass substantially weaker than the mobility itself FAIMS is much more orthogonal to MS than linear IMS 24 25 26 The gain depends on the analyte nature and is about 4fold for peptides 27 Hence FAIMS tends to resolve isomers finer than linear IMS with equal R metric

Keywords: