Protonated peptides are dissociated in the gas phase into fragments as shown above. Here, a single molecular ion of the peptide MASSSPECR is isolated (left) and then ‘heated’ until dissociation (middle). Cleavage at the amide bond is common (see mobile proton model), and produces sequence-informative fragments. In most cases, only one amide bond per peptide ion is cleaved. However, each copy of the same peptide sequence will cleave at a different amide bond (right), leading to fragmentation that is statistically (randomly) distributed in sequence position.
Peptide fragments are named according to whether the fragment contains the N- or C- terminus, the backbone bond that is cleaved, and the position along the peptide chain.1,2 For example, the b- and y- designations correspond to cleavage at the amide bond. The b- ion type contains the N-terminus, whereas the y- ion type contains the C-terminus. The subscript number is the residue number relative to the terminus that the fragment contains (N-terminus for b-, C-terminus for y-). The remaining letters correspond to cleavage at other bonds.
Shown below is the CID spectrum of doubly protonated IGGIGTVpVGr, where p and r are “heavy” proline and arginine, respectively. The fragmentation spectrum is dominated by the b- and y- ion series, providing good coverage of the peptide sequence. If the identity of this peptide were unknown, then this fragmentation spectrum would be searched against a database of predicted fragmentation spectra. A unique peptide “hit” is found when the similarities between the predicted and observed fragmentation match to a significantly greater degree than any other peptide candidates.
Alternatively, peptide sequences can be determined “de novo” by analyzing the peak differences in the fragmentation spectrum. These approaches, which do not require database searching, utilize the fact that the mass difference between neighboring peaks of the same ion type can be used to identify the bracketed amino acid residue. For example, analysis of the mass differences among the b2 and b5 ion series results in a partial sequence of G(I/L)G. In principle, “de novo” sequencing is capable of identifying sequences that are not predicted from the genome (e.g. protein SNP isoforms), which is a unique advantage over database searching.
1 Roepstorff, P.; Fohlman, J. Biol. Mass Spectrom. 1984, 11, 601.
2 See http://www.matrixscience.com/help/fragmentation_help.html for more information on peptide fragmentation nomenclature.