| Literature DB >> 26941048 |
Michael R Shortreed1, Brian L Frey1, Mark Scalf1, Rachel A Knoener1, Anthony J Cesnik1, Lloyd M Smith1,2.
Abstract
Proteomics is presently dominated by the "bottom-up" strategy, in which proteins are enzymatically digested into peptides for mass spectrometric identification. Although this approach is highly effective at identifying large numbers of proteins present in complex samples, the digestion into peptides renders it impossible to identify the proteoforms from which they were derived. We present here a powerful new strategy for the identification ofEntities:
Keywords: NeuCode; PTM; database search; proteoform; proteoform family; proteomics; top-down
Mesh:
Substances:
Year: 2016 PMID: 26941048 PMCID: PMC4917391 DOI: 10.1021/acs.jproteome.5b01090
Source DB: PubMed Journal: J Proteome Res ISSN: 1535-3893 Impact factor: 4.466
Figure 1Example intact protein chromatogram and spectrum. (A) An LC–MS chromatogram for one gel electrophoresis fraction of NeuCode SILAC yeast. (B) A full-scan mass spectrum obtained at a resolution of 100 000. (C) An expanded view of the mass spectrum showing one charge-state envelope containing multiple isotope peaks for each of the two isotopologues. (D) A further expanded view displaying the “Light” (left peak) and “Heavy” (right peak) isotopologues; the spacing between these two peaks is used to determine the number of lysines in this proteoform.
Figure 2Three-stage strategy for elucidating proteoform families. In the first stage, experimental intact masses, En, are compared to theoretical masses, Tn, (having the same lysine count) to create ET pairs for certain mass differences (e.g., 42 Da). In the second stage, EE pairs are similarly generated. In the third stage, the pairs are clustered together to produce proteoform families, two examples of which are shown here.
Figure 3Histograms of observed mass differences. (A) Mass differences between experimental masses and theoretical ones calculated from UniProt entries, which have the same number of lysines. (B) Mass differences between pairs of experimental observations, again stipulating the same lysine count. The most frequently observed mass differences correspond to common PTM or amino acid masses. A total of 31 of the 88 mass differences (highlighted in pink) were directly attributable to known modifications (e.g., oxidation, methylation, and acetylation) or amino acid losses at one of the proteoform termini. Another 34 peaks were adjacent to these 31 primary mass shifts, up to 2 Da away, and were attributed to misassignment of the proteoform monoisotopic mass. Several other mass shifts (e.g., 46 and 72 Da) were included in the construction of proteoform families because they exceeded the threshold but they remain unidentified. These two mass shifts in particular are absent in the compendium of modifications at unimod.org. They could arise from a combination of modifications. An FDR threshold is shown (green line). See also the Supplemental Tables S-9 and S-11 for the complete list of ET and EE mass differences to 500 Da.
Figure 4Proteoform families. (A) Display of 1178 proteoform families discovered in this work. (B–E) Expanded views of four example proteoform families. Theoretical unmodified proteins (red nodes) are labeled with their UniProt accession number. Theoretical modified proteins (green nodes) are labeled with their accession number and a PTM known to occur on that protein. Experimentally observed proteoforms (blue nodes) are labeled with their intact mass and the number of times it was detected. The area of each blue node is proportional in size to the number of times that proteoform was observed experimentally; however, to facilitate visualization, all nodes corresponding to 1–10 observations were given the same (minimum) size. Proteoforms are connected by select mass differences (edges) indicated by black lines with orange mass-difference values.
Figure 5Distribution of observed proteoforms in various types of proteoform families. Most of the observed proteoforms clustered with theoretical or other experimental proteoforms to make families, although some did not (i.e., “orphans”). The proteoform families are categorized as identified, ambiguous, or not yet identified based on containing one, two or more, or zero theoretical accession numbers, respectively. The term “Observation” here refers to each detection of a proteoform intact mass and lysine count in any of the 29 847 mass spectra collected in this study.