Use of Multiple Ion Fragmentation Methods to Identify Protein Cross-Links and Facilitate Comparison of Data Interpretation Algorithms

Multiple ion fragmentation methods
involving collision-induced
dissociation (CID), higher-energy collisional dissociation (HCD) with
regular and very high energy settings, and electron-transfer dissociation
with supplementary HCD (EThcD) are implemented to improve the confidence
of cross-link identifications. Three different S. cerevisiae proteasome samples cross-linked by diethyl suberthioimidate (DEST)
or bis­(sulfosuccinimidyl)­suberate (BS3) are analyzed. Two
approaches are introduced to combine interpretations from the above
four methods. Working with cleavable cross-linkers such as DEST, the
first approach searches for cross-link diagnostic ions and consistency
among the best interpretations derived from all four MS2 spectra associated with each precursor ion. Better agreement leads
to a more definitive identification. Compatible with both cleavable
and noncleavable cross-linkers such as BS3, the second
approach multiplies scoring metrics from a number of fragmentation
experiments to derive an overall best match. This significantly increases
the scoring gap between the target and decoy matches. The validity
of cross-links fragmented by HCD alone and identified by Kojak, MeroX, pLink, and Xi was evaluated using multiple fragmentation data. Possible ways to
improve the identification credibility are discussed. Data are available
via ProteomeXchange with identifier PXD018310.