Peptide Identity by Mass Spectrometry Fragmentation: What the Published Literature Reports

Peptide-centric mass spectrometry workflows rest on determining the mass of biomolecules and of their fragments or related products with high accuracy, which a methodological review describes as conveying a highly specific assay for detection and identification, with the same ion currents also usable for quantitation [1]. In the conventional bottom-up configuration, protein identity is inferred from mass spectrometric analysis of peptides generated by proteolytic digestion, usually with trypsin, whereas so-called top-down approaches subject intact protein ions and large protein fragments directly to tandem mass spectrometry [2]. A review of top-down characterisation notes that a key enabling step was the ability to assign product ion identities, achieved either through high resolving power or through manipulation of product ion charge state, and that electrospray ionisation combined with high-field Fourier transform ion cyclotron resonance has been particularly powerful for detailed protein characterisation because of its mass resolution, mass accuracy, and access to electron capture-induced dissociation [2]. All of this work was carried out on purified analytes and biological extracts in vitro. The physics of how a selected precursor ion is broken apart constrains what identity information is recoverable. A review of collisional activation in FT-ICR instruments describes tandem mass spectrometry as mass selection of a primary ion, activation by collision or photon excitation, unimolecular decay into fragment ions characteristic of ion structure and internal excitation, and mass analysis of those fragments [3]. That review reports that for large ions two limits constrain fragmentation: the centre-of-mass collision energy, the upper limit of energy transfer in a single collision, falls as projectile ion mass increases at fixed kinetic energy, and the steep rise in density of states with internal degrees of freedom lowers dissociation rates by many orders of magnitude at a given internal energy, so practical experiments on complex ions use multiple-collision activation, multiphoton activation, or surface-induced dissociation [3]. Combining collision energy-resolved data with RRKM-based modelling in those in vitro experiments revealed an effect of peptide size and identity on energy transfer, and time-resolved surface-induced dissociation experiments identified a transition from smoothly varying dissociation kinetics to near-instantaneous "shattering" that generates many small fragment ions above a certain collision energy [3]. Fragmentation is also the route by which modifications are localised on a sequence. A review of glycopeptide analysis describes LC-ESI-MS2 and MS3 strategies used to pursue glycan structure, peptide identity, and attachment-site assignment for protease-digested glycoproteins, and presents CID and HCD MS2 spectra of a complex biantennary N-glycopeptide and a core 1 O-glycopeptide as worked in vitro examples [4]. For deamidation, capillary electrophoresis-tandem mass spectrometry was applied in vitro to the 4.5-kDa peptide TRI-1144, which carries five closely positioned potential deamidation sites; separation of deamidated, deacetylated, and positional isomer species allowed the number of modifications per degradation product to be determined, while collision-induced dissociation localised modification sites and comparison of theoretical with measured isotope distributions for specific y ions resolved the identity and relative abundance of comigrating isomers [5]. Several papers address how many fragments are actually needed to assign identity. A computational and in vitro study of selected reaction monitoring modelled interferences from other peptides in a proteome and defined deterministic "unique ion signatures"; the authors computed that signatures of only two transitions were diagnostic for more than 99% of Escherichia coli proteins and more than 96% of human proteins possessing a sequence-unique peptide, and demonstrated the approach experimentally on an E. coli cell lysate while proposing it as a functionally orthogonal check on peptide assignments from conventional MS/MS spectra [6]. A complementary instrument-control study implemented real-time recalibration of mass, retention time, and intensity on a quadrupole Orbitrap to predict precursor identity, reporting recognition and targeting of more than 25,000 peptides in single LC-MS runs and detection of precursors not visible in MS1 scans [7]. A de novo sequencing strategy instead used paired Lys-N and Lys-C digestions that yield peptides of identical mass and similar retention time but different tandem mass spectra; matched spectral pairs allowed fragment ion type, residues, and peptides to be assigned with residue-by-residue confidence, illustrated in vitro on an 18-mer using 14.5 T FT-ICR MS [8]. The literature also documents where assignment goes wrong. A peptide mapping study of therapeutic proteins and recombinant vaccine antigens reported incorrect peptide assignments produced by commercial software, including misidentifications arising from isobaric and near-isobaric dipeptides such as SA versus GT in antibody sequence analysis, and software-induced artefacts in SARS-CoV-2 spike protein variant data comprising artificial succinylation of aspartic acid residues to compensate for sequence mismatches and incorrect deamidation site assignment from misinterpreted isotopic peaks; the authors concluded that expert manual review of MS/MS data remains necessary even with validated commercial platforms [9]. Related methodological work notes that standard bottom-up workflows focus on internal tryptic peptides and often fail to capture N-terminal peptides, and describes a negative-selection enrichment protocol combined with data-independent acquisition for N-terminome profiling with annotation of cleavage sites and post-translational modifications [10]. A separate review emphasised sampling and dynamic range limitations and developments intended to improve translation of peptide fragmentation data into peptide and protein identities [1], and an organellar proteomics review described the stochastic nature of peptide selection for fragmentation as something addressed by biological replicates and prior gel-based protein separation [18]. Fragmentation-based readouts have been extended beyond sequence. Multidimensional mass spectrometry of alanine-rich peptides and their poly(ethylene glycol) conjugates used MS and MS2 to establish composition, sequence, and the site of PEG attachment at the peptide C-terminus, while ion mobility separation detected random coil and alpha-helical conformers in the gas phase whose identity was confirmed against simulated collision cross sections, with circular dichroism used as an independent in vitro corroboration [11]. Ultrahigh-resolution measurement of the fine structure of immonium ions produced by gas-phase fragmentation of polypeptides was validated for isotopic ratio analysis of aliphatic residues from proteins and cell lysates, with the same proteomics-type experiment providing sequence data used to verify sample purity and establish identity [12]. A review of FT-ICR developments reports ultra-high resolving power with isotopic resolution for peptides and small proteins up to about 20 kDa and a range of fragmentation strategies for sequence, structure, and post-translational modification characterisation, applied to human serum and saliva peptide profiles [13]. In biological studies, fragmentation-based identification has been used as the assignment step rather than the endpoint. A review of amyloid-beta characterisation describes the central role mass spectrometry has played in defining the molecular identity of these peptides and their higher-order assemblies in human brain, cerebrospinal fluid, blood, and plasma, while noting remaining knowledge gaps [14]. Direct MALDI peptide fingerprinting of single identified neurons in the mollusc Lymnaea stagnalis, an animal model, revealed trimmed variant peptides derived from precursor-encoded peptides, and the measured molecular masses were used as markers to guide purification before structural identity was confirmed by amino acid sequencing [15]. In a mouse retinal vascularisation model, mass spectrometry analysis showed that a recombinant C-terminal fibrillin-1 fragment altered the expression of various proteins including ADAMTS1 [16]. A commentary describes amyloid deposits dissected from affected human tissue, digested into fragments for proteomic analysis, and assessed by mass spectrometry to reveal both chaperone signatures and the identity of the amyloid-forming protein [17]. Proteomic identification was similarly used in vitro to identify 39 human proteins forming DNA-protein cross-links in HeLa nuclear extracts treated with 1,2,3,4-diepoxybutane, with cross-linking efficiency reported to depend on protein identity and HPLC-ESI-MS/MS of proteolytic digests revealing specific cysteine-guanine conjugates [19]. Outside peptides, a review compiling fragmentation schemes for around 40 small-molecule HIV antivirals and several phosphorylated anabolites tabulates the elemental composition and exact m/z of product ions used in selected reaction monitoring, illustrating the same principle that documented product ion identity underpins targeted assays [20].