In-Source Collision-Induced Dissociation (CID) Improves Higher-Energy Collisional Dissociation (HCD)-Dependent Fragmentation of ADP-Ribosyl Peptides

Abstract

Rationale: ADP-ribosylation is a posttranslational modification whose higher-energy collisional dissociation (HCD) products are dominated by complete or partial modification losses, complicating peptide sequencing and acceptor site localization efforts. We tested whether in-source collision-induced dissociation (CID) performed on a quadrupole-Orbitrap could convert ADPr to the smaller phosphoribose-H₂O derivative to facilitate HCD-dependent peptide sequencing.

Methods: ADP-ribosyl (ADPr) peptides derived from the human macrophage-like cell line THP-1 were analyzed on a quadrupole-Orbitrap. We monitored the dissociation of ADPr (+ 541.061 Da) to phosphoribosyl-H₂O (+ 193.997 Da) peptides while varying the source and high-field asymmetric waveform ion mobility mass spectrometry (FAIMS) compensation voltages. Xcorr and ptmRS were used to evaluate peptide sequencing and acceptor site confidence, respectively. Phosphoribosyl-H₂O acceptor sites were compared with those determined by electron-transfer higher-energy collision dissociation (EThcD), performed on a quadrupole-ion trap-Orbitrap.

Results: In-source CID of ADPr peptides to their phosphoribosyl-H₂O derivatives increased with increasing source voltage (up to 50 V), as judged by monitoring the corresponding modification loss ([adenosine monophosphate/AMP]⁺) and the number of identified phosphoribosyl-H₂O peptide identifications. The average Xcorr increased from 1.36 (ADPr) to 2.26 (phosphoribosyl-H₂O), similar to that achieved with EThcD for ADPr peptides (2.29). The number of high-confidence acceptor sites (> 95%) also increased, from 31% (ADPr) to 70% (phosphoribosyl-H₂O), which was comparable to EThcD (70%).

Conclusions: In-source CID converts ADP-ribosyl to phosphoribosyl-H₂O peptides that are more amenable to HCD-dependent peptide sequencing, providing an alternative method for acceptor site determination when ETD-based methods are not available.

Keywords: ADP‐ribosylation; electron transfer dissociation; in‐source collision‐induced dissociation; ribosylome.

FIGURE 1

The dissociation properties of adenosine diphosphate ribose (ADPr). (A) HCD‐dependent fragment ions of ADPr peptides. “P” indicates intact peptide after the modification (m‐ion) loss; “p” indicates sequential dissociation of the peptide backbone. The P5‐m6 fragment ions are the targets of this study. (B) A schematic of summarizing that ADPr m‐ions result from the sequential dissociation of ADPr or AMP (Kuraoka et al., 2021).

FIGURE 2

Rationale for in‐source CID of ADPr peptides. (A) Dissociation curves for ADPr, ADP, and AMP small molecules (source: mzCloud), demonstrating their precursor to product relationship in HCD. (B) Strategy and hypothesized fragmentation results for in‐source CID‐HCD of ADPr peptides on the benchtop quadrupole‐Orbitrap (Exploris 480) mass spectrometer. (C) MS1 scans (m/z 130–550) of ADPr peptides enriched from THP‐1 cells acquired with increasing source voltage. Peak intensity was calculated using the monoisotopic peak extracted over a 70‐min gradient. (D) The average MS1 intensity (across source voltages) for each ADPr ion.

FIGURE 3

An evaluation of source voltage and FAIMS on ADPr and phosphoribosyl* peptide yields. (A,C) The number of ADPr peptides identified at three test source voltages with or without FAIMS. (B, D) The number of phosphoribosyl* peptides identified at three test source voltages with or without FAIMS. *, —H₂O.

FIGURE 4

Further in‐source CID optimization with FAIMS. (A) The number of unique ADPr vs. phosphoribosyl* peptide (P5) sequences at increasing source voltage. Overlap, common peptides sequences (inclusive to the ADPr and P5 plots). (B) Retention time (RT) distributions for common peptides. Source‐induced P5 dissociation products will have identical RTs as their ADPr peptide precursors. (C) Peptide scores (Xcorr values from SEQUEST) are higher for the P5 precursors demonstrating improved b/y dissociation with the smaller modification (phosphoribosyl* vs. ADPr). *, —H₂O.

FIGURE 5

In‐source CID‐HCD‐generated phosphoribosyl* precursors improve ADPr peptide sequence quality. (A) Venn diagram displaying extent of overlapping sequence identifications across the acquisition strategies. (B) A summary of Xcorr and acceptor site probabilities across the acquisition strategies. (C) Distribution of amino acid acceptor sites with > 95% localization confidence (ptmRS in Proteome Discoverer). *, —H₂O.

FIGURE 6

An example ADPr peptide identified as its ADPr and phosphoribosyl* forms using either EThcD performed on a Fusion Lumos or in‐source CID‐HCD performed on an Exploris 480. Both EThcD of the ADPr form, and in‐source CID‐HCD of the phosphoribosyl* form confirmed the acceptor site to be the lysine (y5); however, when analyzed by HCD, upstream Asp and Glu are also candidate acceptor sites. *, —H₂O. Highlighted peaks are ADPr peptide fragment ions annotated using RiboMaP.

FIGURE 7

In‐source CID‐HCD with GPS + FAIMS provides the maximum number of P5/phosphoribosyl* peptide. (A) Venn diagram displaying extent of overlapping ADPr and P5/phosphoribosyl* peptide and (B) proteins. *, —H₂O. Inset (A) the number of unique acceptor sites identified using the phosphoribosyl* precursor. (C) The number and distribution of acceptor sites identified in the THP‐1 ADP‐ribosylome when capitalizing on in‐source CID‐HCD‐derived phosphoribosyl* precursors.