Selection of Collision Energies in Proteomics Mass Spectrometry Experiments for Best Peptide Identification: Study of Mascot Score Energy Dependence Reveals Double Optimum

Abstract

Collision energy is a key parameter determining the information content of beam-type collision induced dissociation tandem mass spectrometry (MS/MS) spectra, and its optimal choice largely affects successful peptide and protein identification in MS-based proteomics. For an MS/MS spectrum, quality of peptide match based on sequence database search, often characterized in terms of a single score, is a complex function of spectrum characteristics, and its collision energy dependence has remained largely unexplored. We carried out electrospray ionization-quadrupole-time of flight (ESI-Q-TOF)-MS/MS measurements on 2807 peptides from tryptic digests of HeLa and E. coli at 21 different collision energies. Agglomerative clustering of the resulting Mascot score versus energy curves revealed that only few of them display a single, well-defined maximum; rather, they feature either a broad plateau or two clear peaks. Nonlinear least-squares fitting of one or two Gaussian functions allowed the characteristic energies to be determined. We found that the double peaks and the plateaus in Mascot score can be associated with the different energy dependence of b- and y-type fragment ion intensities. We determined that the energies for optimum Mascot scores follow separate linear trends for the unimodal and bimodal cases with rather large residual variance even after differences in proton mobility are taken into account. This leaves room for experiment optimization and points to the possible influence of further factors beyond m/ z.

Site-specific N-glycosylation of HeLa cell glycoproteins

“…Peptide identification was performed under conventional conditions (see the Experimental part for details) and evaluated against Swissprot human and bovine databases. A known limitation of data-dependent analysis is that some compounds are not sampled in the LC-MS/MS run 5,32 . Furthermore, glycopeptides are among the low abundance peaks, so this problem is even more pronounced in the case of glycopeptide analysis.…”

“…It is a human cervical adenocarcinoma cell line expressing over 10 000 proteins 2 . The tryptic digest of HeLa S3 cells is commercially available and is the most widely characterized mass spectrometry standard 2–5 in proteomics. The mixture contains various post-translational modifications that have been studied as well 6–8 .…”

Collision energies on QTof and Orbitrap instruments: How to make proteomics measurements comparable?

“…Collision energy dependent nano‐LC‐MS/MS measurements of HeLa tryptic digest on the Bruker Maxis II ETD QTof instrument were recorded earlier as part of a previous work on the Mascot score‐collision energy relationship 8 . The detailed description of the experiments can be found in the previous publication; therefore, only a brief summary is presented here.…”

“…Recently, we have studied collision energy effects in a large scale proteomics study using a QTof mass spectrometer, which is one of the most frequently used instrument in this field 8 . We have characterized the quality of a CID spectrum by the identification score of the database search using the widely accepted Mascot search engine.…”

“…The b ions' intensity shows maximum at the lower energy optimum, whereas y ions are most abundant at the higher energy maximum, although the latter are dominant in the whole energy range. These findings led us to develop novel proteomics workflows via energy optimization, resulting in significantly better protein coverage, higher sequence coverage, 8,9 and more efficient sequence validation 10 than typical protocols using generic collision energy setup.…”