Fig 1.
Classes of backbone cleavage on peptides.
An example peptide, SALFA, demonstrating the locations of the three types of backbone cleavage. Type A cleavage (red dashed lines) occurs between the α-carbon and the carbonyl carbon of the same amino acid. Type B cleavage (blue narrow lines) occurs at the peptide bond between two amino acids. Type C cleavage (green thick lines) occurs between the α-carbon and the nitrogen of the same amino acid and it results in the release of the aromatic side chain. Type C has only been observed in the aromatic residues of Phe, Trp, and Tyr [3].
Table 1.
Relative mass fragmentation eventsa.
Table 2.
Non-relative mass fragmentation events (charged side chain losses)a.
Fig 2.
Comparison of reproducible Gln and Asn fragmentation behavior during DEP-EI-MS.
Three dominant fragmentation mechanisms were observed in the Gln (blue trace) and Asn (red trace) mass spectra. A) Peaks associated with an “Ammonia loss” fragmentation are specific to the amide-containing amino acids. Based on the relative intensity, having this be the sole fragmentation event is relatively rare. B) Peaks associated with loss of the C-terminal carboxy group is a common fragmentation event seen in most amino acids and peptides tested. C) Peaks associated with the combination of both a loss of the C-terminal carboxy group and the “Ammonium loss” fragmentation event. This is particularly favored in Gln. D) Peaks at around 74 m/z which corresponds with the loss of side chain and is held in common with any amino acid side chain loss. Once again, this fragmentation peak is more pronounced in Gln. The maximum peak intensity is 2.86 X 108 counts for Gln, and 1.38 X 108 counts for Asn. Proposed structures are shown along with the resulting fragment formula and monoisotopic m/z. adegradation type also observed in [11], bdegradation type also observed in [3].
Fig 3.
Observed cationic alkene side chain fragments from aromatic residues.
For three of the aromatic amino acids, a peak was observed which corresponds to the loss of the side chain with a gain of charge along with the formation of a double bond between the α-carbon and the first carbon of the side chain. The above structures, each displaying extended conjugation from the aromatic side chain to the α-carbon, represent the likely fragment produced.
Fig 4.
QFA fragment peak generation flow chart.
The current model produces 55 fragments including the original peptide with unique mechanism flow paths after duplicates are removed. Note that the presence of non-relative fragmentation events does result in a few duplicated m/z predictions because those events are not combinatorial with the fragmentations that might have preceded them in a given degradation pathway since they will result in the same end fragment.
Table 3.
Crude synthetic tri/tetra peptides used for prediction comparison.
Table 4.
Major peaks arising from protecting groups used in peptide synthesis.
Fig 5.
QFA DEP-EI-mass spectrum compared with predicted peaks.
In blue are all plotted peaks observed in the spectrum for the QFA peptide. In green are peaks that match with predictions generated using our model. Red squares mark peaks associated with protecting groups used in the synthesis of the peptides. A peak was considered a match if it was within the max instrumental error (+/- 0.25 m/z) of the mass spectrometer. The maximum peak intensity is 9.12 X 107 counts.
Fig 6.
QFA DEP-EI-mass spectrum compared with predicted peaks for the sequence CIA.
As a counter example for evaluating the program’s predictions, peak matching was done between an incorrect sequence, CIA, and the QFA mass spectrum. In blue are all plotted peaks observed in the spectrum for the QFA peptide. In green are peaks that match with CIA fragmentation predictions generated using our model. Red squares mark peaks associated with protecting groups used in the synthesis of the peptides. A peak was considered a match if it was within the max instrumental error (+/- 0.25 m/z) of the mass spectrometer. The maximum peak intensity is 9.12 X 107 counts.
Fig 7.
Percent matches of proteinogenic amino acids.
Percentages of prediction-matching peaks relative to all peaks were calculated using peaks whose intensity met or exceeded the threshold intensity relative to the most intense peak in the spectrum.
Fig 8.
Percent matches of crude synthetic peptides.
Percentages of prediction-matching peaks relative to all peaks were calculated using peaks whose intensity met or exceeded the threshold intensity relative to the most intense peak in the spectrum.
Fig 9.
The average run times for the 20 proteinogenic amino acids as well as the di-, tri-, and tetrapeptides (SEQA) investigated in this work were evaluated by timing how long it took the algorithm to generate and filter its predictions for each sequence 10 times.