A Classifier Based on Accurate Mass Measurements to Aid Large Scale, Unbiased Glycoproteomics

Determining which glycan moieties occupy specific N-glycosylation sites is a highly challenging analytical task. Arguably, the most common approach involves LC-MS and LC-MS/MS analysis of glycopeptides generated by proteases with high cleavage site specificity; however, the depth achieved by this approach is modest. Nonglycosylated peptides are a major challenge to glycoproteomics, as they are preferentially selected for data-dependent MS/MS due to higher ionization efficiencies and higher stoichiometric levels in moderately complex samples. With the goal of improving glycopeptide coverage, a mass defect classifier was developed that discriminates between peptides and glycopeptides in complex mixtures based on accurate mass measurements of precursor peaks. By using the classifier, glycopeptides that were not fragmented in an initial data-dependent acquisition run may be targeted in a subsequent analysis without any prior knowledge of the glycan or protein species present in the mixture. Additionally, from probable glycopeptides that were poorly fragmented, tandem mass spectra may be reacquired using optimal glycopeptide settings. We demonstrate high sensitivity (0.892) and specificity (0.947) based on an in silico dataset spanning >100,000 tryptic entries. Comparable results were obtained using chymotryptic species. Further validation using published data and a fractionated tryptic digest of human urinary proteins was performed, yielding a sensitivity of 0.90 and a specificity of 0.93. Lists of glycopeptides may be generated from an initial proteomics experiment, and we show they may be efficiently targeted using the classifier. Considering the growing availability of high accuracy mass analyzers, this approach represents a simple and broadly applicable means of increasing the depth of MS/MS-based glycoproteomic analyses. Determining which glycan moieties occupy specific N-glycosylation sites is a highly challenging analytical task. Arguably, the most common approach involves LC-MS and LC-MS/MS analysis of glycopeptides generated by proteases with high cleavage site specificity; however, the depth achieved by this approach is modest. Nonglycosylated peptides are a major challenge to glycoproteomics, as they are preferentially selected for data-dependent MS/MS due to higher ionization efficiencies and higher stoichiometric levels in moderately complex samples. With the goal of improving glycopeptide coverage, a mass defect classifier was developed that discriminates between peptides and glycopeptides in complex mixtures based on accurate mass measurements of precursor peaks. By using the classifier, glycopeptides that were not fragmented in an initial data-dependent acquisition run may be targeted in a subsequent analysis without any prior knowledge of the glycan or protein species present in the mixture. Additionally, from probable glycopeptides that were poorly fragmented, tandem mass spectra may be reacquired using optimal glycopeptide settings. We demonstrate high sensitivity (0.892) and specificity (0.947) based on an in silico dataset spanning >100,000 tryptic entries. Comparable results were obtained using chymotryptic species. Further validation using published data and a fractionated tryptic digest of human urinary proteins was performed, yielding a sensitivity of 0.90 and a specificity of 0.93. Lists of glycopeptides may be generated from an initial proteomics experiment, and we show they may be efficiently targeted using the classifier. Considering the growing availability of high accuracy mass analyzers, this approach represents a simple and broadly applicable means of increasing the depth of MS/MS-based glycoproteomic analyses. N-Glycosylation is an important post-translational modification that affects cell-cell signaling and protein stability, and it has been implicated in various pathologies (1Varki A. Biological roles of oligosaccharides: all of the theories are correct.Glycobiology. 1993; 3: 97-130Crossref PubMed Scopus (4979) Google Scholar). Protein N-glycosylation is difficult to characterize, due to heterogeneity at the levels of glycosylation site occupancy, glycan composition, and glycan structure. A truly comprehensive analysis of protein glycosylation identifies glycans, maps occupied sites, and matches the glycans to specific sites on glycoproteins (2An H.J. Froehlich J.W. Lebrilla C.B. Determination of glycosylation sites and site-specific heterogeneity in glycoproteins.Curr. Opin. Chem. Biol. 2009; 13: 421-426Crossref PubMed Scopus (175) Google Scholar). This site-specific analysis can be performed via analysis of intact glycopeptides using mass spectrometry (MS). However, this technique is complicated by sensitivity, sample preparation, and fragmentation challenges (3Dodds E.D. Gas-phase dissociation of glycosylated peptide ions.Mass Spectrom. Rev. 2012; 31: 666-682Crossref PubMed Scopus (61) Google Scholar) that limit the throughput and depth of the results. The analysis of site-specific glycosylation is complicated in part because the ionization of glycopeptides is suppressed by any nonglycosylated peptides that are coproduced during protease digestion with specific proteases. Alternatively, digestion using nonspecific proteases has been implemented to eliminate competing peptide species (4Dalpathado D.S. Irungu J. Go E.P. Butnev V.Y. Norton K. Bousfield G.R. Desaire H. Comparative glycomics of the glycoprotein follicle-stimulating hormone: glycopeptide analysis of isolates from two mammalian species.Biochemistry. 2006; 45: 8665-8673Crossref PubMed Scopus (56) Google Scholar, 5Clowers B.H. Dodds E.D. Seipert R.R. Lebrilla C.B. Site determination of protein glycosylation based on digestion with immobilized nonspecific proteases and Fourier transform ion cyclotron resonance mass spectrometry.J. Proteome Res. 2007; 6: 4032-4040Crossref PubMed Scopus (62) Google Scholar). Notably, specific proteases yield highly predictable peptide footprints and have been utilized for analysis of complex mixtures. However, as noted in a recent publication, glycopeptides are often not selected for fragmentation in data-dependent analysis (DDA) 1The abbreviations used are:DDAdata-dependent analysisMDmass defectGRAEZglycopeptide-rich acquisition enhancement zoneTEABtriethyl ammonium bicarbonateACGautomatic gain control. 1The abbreviations used are:DDAdata-dependent analysisMDmass defectGRAEZglycopeptide-rich acquisition enhancement zoneTEABtriethyl ammonium bicarbonateACGautomatic gain control. (6Kolarich D. Jensen P.H. Altmann F. Packer N.H. Determination of site-specific glycan heterogeneity on glycoproteins.Nat. Protoc. 2012; 7: 1285-1298Crossref PubMed Scopus (157) Google Scholar), making glycopeptide identification impossible, as fragmentation is required for glycopeptide identification in nontrivial samples (7Desaire H. Hua D. When can glycopeptides be assigned based solely on high-resolution mass spectrometry data?.Int. J. Mass Spectrom. 2009; 287: 21-26Crossref Scopus (35) Google Scholar). To circumvent this issue, glycopeptide enrichment protocols using normal-phase hydrophilic interaction chromatography or lectin enrichment techniques have been established to enrich for glycopeptides (8Ito S. Hayama K. Hirabayashi J. Enrichment strategies for glycopeptides.Methods Mol. Biol. 2009; 534: 195-203PubMed Google Scholar). Although highly valuable, these purification approaches have varying specificities for glycopeptides, may preferentially isolate glycopeptides with certain types of glycans attached, and add additional sample handling steps. data-dependent analysis mass defect glycopeptide-rich acquisition enhancement zone triethyl ammonium bicarbonate automatic gain control. data-dependent analysis mass defect glycopeptide-rich acquisition enhancement zone triethyl ammonium bicarbonate automatic gain control. Because of these challenges, a classifier capable of quickly discriminating between peptide and glycopeptide signals in mass spectrometry would be valuable and may significantly complement existing purification techniques. Fragment ions have been found that are specific to glycopeptides (9Huddleston M.J. Bean M.F. Carr S.A. Collisional fragmentation of glycopeptides by electrospray ionization LC/MS and LC/MS/MS: methods for selective detection of glycopeptides in protein digests.Anal. Chem. 1993; 65: 877-884Crossref PubMed Scopus (370) Google Scholar, 10Jebanathirajah J. Steen H. Roepstorff P. Using optimized collision energies and high resolution, high accuracy fragment ion selection to improve glycopeptide detection by precursor ion scanning.J. Am. Soc. Mass Spectrom. 2003; 14: 777-784Crossref PubMed Scopus (73) Google Scholar); however, these are not useful if the glycopeptides were not selected for fragmentation or if they yielded low quality MS/MS spectra. As mass defect (MD) classifications have been applied to similar challenges in proteomics (11Bruce C. Shifman M.A. Miller P. Gulcicek E.E. Probabilistic enrichment of phosphopeptides by their mass defect.Anal. Chem. 2006; 78: 4374-4382Crossref PubMed Scopus (25) Google Scholar, 12Dodds E.D. An H.J. Hagerman P.J. Lebrilla C.B. Enhanced peptide mass fingerprinting through high mass accuracy: Exclusion of nonpeptide signals based on residual mass.J. Proteome Res. 2006; 5: 1195-1203Crossref PubMed Scopus (23) Google Scholar, 13Kirchner M. Timm W. Fong P. Wangemann P. Steen H. Nonlinear classification for on-the-fly fractional mass filtering and targeted precursor fragmentation in mass spectrometry experiments.Bioinformatics. 2010; 26: 791-797Crossref PubMed Scopus (13) Google Scholar), we investigated whether an MD classification would be useful for discriminating between peptides and glycopeptides. Notably, a lowering of the MD has been observed for glycopeptides, because of the relative increase of oxygen (and its negative MD value) in glycopeptides (14Lehmann W.D. Bohne A. von Der Lieth C.W. The information encrypted in accurate peptide masses-improved protein identification and assistance in glycopeptide identification and characterization.J. Mass Spectrom. 2000; 35: 1335-1341Crossref PubMed Scopus (41) Google Scholar). However, this original observation was made through comparison of tryptic peptides and small glycopeptides generated by nonspecific proteolysis. Since that original work, no systematic studies have been performed to maximize the analytical utility of this MD shift or to determine the true- and false-positive rates of such a classifier based on accurate mass measurements. Furthermore, no studies have determined whether the MD shift holds for peptides and glycopeptides generated by the same protease, a more pertinent comparison given typical sample preparation protocols. In this study, we have determined these glycopeptide-rich acquisition enhancement zones (GRAEZs) and validated their utility in identifying useful precursor m/z values for large scale glycopeptide assignment by tandem MS. This classification may be applied to identify likely N-glycopeptides without parallel proteomic or glycomic experiments and without any prior knowledge of the proteome or glycome present in a mixture. Targeted MS studies of species within the GRAEZ will increase selection of glycopeptides for fragmentation, leading to increased glycopeptide identification. This concept is presented schematically in Fig. 1. We further demonstrate the efficacy of GRAEZ classification by validating LC-MS/MS data from urinary proteomics analysis. The GRAEZ MD settings were determined using an in silico training dataset and evaluated using an in silico test dataset of peptides and glycopeptides. Training and test sets were generated from the HUPO plasma proteome database, which may be accessed on line. Entries were re-mapped to SwissProt Identifiers. A total of 1797 unique entries were generated. Six hundred random protein entries were selected and digested in silico with either trypsin or chymotrypsin using MS-Digest to form the training sets. The remaining 1197 proteins were used to form the test set. One missed cleavage was permitted; cysteine residues were considered as their carbamidomethyl derivatives, and peptide output was restricted to >3 amino acids and 400–5000 daltons. This range was chosen to select peptide sizes that were typically amenable to analysis on most MS instrumentation. MS-Digest reported singly protonated m/z values for all peptides. Peptide output was imported into Microsoft Excel for data analysis. Redundant peptide sequences were removed. Peptides containing potential N-glycosylation consensus site were identified by the presence of NXS or NXT sequences, where X is any amino acid except proline. Glycopeptides were then generated in silico by adding the monosaccharide masses of eight distinct N-glycan compositions to each consensus site peptide. The glycans utilized are shown in Table I and were chosen to represent common Homo sapiens N-glycans, without biasing the classifier for large N-linked glycans excessively. Because the MD shift is proportionally less for smaller N-glycans, a range of N-glycan masses was tested to challenge the classifier. Size for tryptic and chymotryptic peptides are shown in Fig. utilized to glycopeptides in in a Peptides and glycopeptides were on a mass defect to identify initial in and defect mass for each and were generated for each GRAEZ settings were between the for each and the accuracy of of the classifier was The initial and values were then optimized using an to maximize To the in silico a sample from a was obtained with an and using a published sample preparation for urinary proteomics Steen H. sample and for urinary Proteome Res. 2010; PubMed Scopus Google Scholar). was and on were and in the with and from the digestion with trypsin at a of at Peptides were with to the and with Peptides were into using an for and were in and by LC-MS/MS using a MS with an and A data-dependent LC-MS/MS was was to for and for A from to was collision was and the was for for and for In to the GRAEZ GRAEZ was peptides were generated as using a sample by a An initial run was performed on the sample all were using J. M. high peptide identification mass and protein 26: PubMed Scopus Google Scholar) and evaluated for GRAEZ A of unique was which were as glycopeptides by and targeted in two subsequent LC-MS were with similar except the collision was and the was for for and for spectra from the were for the presence of two the ion at or the ion at data were evaluated for the and identification of the precursor m/z and values was achieved by the of an that as an for the The all spectra for fragment ions and an and a as output that m/z selected for fragmentation, of the and the of the fragment has been into the D. M. D. P. for proteomics PubMed Scopus Google Scholar) for of and may any m/z this study, mass was and a of relative was required for the fragment The precursor m/z and values were used to values for GRAEZ and these GRAEZ classifications were the presence of the ions in spectra to the of as Because of the of N-linked glycans, N-glycopeptides are in as shown by the in Fig. 1. on the in silico all species may be from targeted analysis with in of tryptic peptides and of chymotryptic peptides were smaller daltons. However, the in silico specificity not the of these low mass species and are with to glycopeptide The GRAEZ settings are given in where is the mass of the of the singly protonated protonated and species and MD is the defect mass of the within the GRAEZ are more likely to be glycosylated as The GRAEZ determined by these are in Fig. and with the test The shown in are required the values for the of the GRAEZ or GRAEZ values that were their as MD by is between the values of and 1. A species that the is as a glycopeptide by a tryptic species with a of would between and and the fractional mass would be evaluated as shown in GRAEZ glycopeptide a large scale GRAEZ can be quickly and performed in any of of LC-MS The tryptic training a sensitivity of and a specificity of within the mass range of to daltons. m/z values the GRAEZ for the increased the tryptic test yielded an increase and the chymotryptic sets a The accuracy of GRAEZ classification of for tryptic sensitivity and specificity were achieved for the chymotryptic as in Table Furthermore, tryptic peptide and glycopeptide test sets were evaluated using the initial that an MD between these species (14Lehmann W.D. Bohne A. von Der Lieth C.W. The information encrypted in accurate peptide masses-improved protein identification and assistance in glycopeptide identification and characterization.J. Mass Spectrom. 2000; 35: 1335-1341Crossref PubMed Scopus (41) Google Scholar). Although the original achieved in identifying likely the rates of glycopeptide assignment to of tryptic peptides were as peptides in silico using the original MD classification GRAEZ classification is more for of the for the in silico data used in this in a The GRAEZ settings were further applied in silico to the remaining of 1197 proteins to their on The of peptides and glycopeptides utilized may be found in 1. the tryptic and chymotryptic test sets a in accuracy in the training that the GRAEZ classifier is In tryptic species were tested in silico and GRAEZ of these species. accuracy was achieved with the chymotryptic test species which The in silico training sets were evaluated as the and in to the species. The GRAEZ classification not with the of the in a for for which the or are the most The data shown this as the of glycopeptide in human at shift Notably, approaches to glycopeptides by observed m/z values to a of a glycan and a peptide We that of precursor m/z values will for large and from such approaches more often of the This be an of the GRAEZ approach relative to however, it additional of glycopeptide identified by fragment for glycan in a To further the in silico published proteomic and glycoproteomic data were GRAEZ of a published proteomic dataset of the proteome J. M. J. S. M. proteome and of a human Biol. 7: PubMed Scopus Google Scholar) of unique tryptic peptides between and as with a specificity of a GRAEZ classification of published site-specific glycoproteomic studies was performed to the sensitivity of the As studies have not the scale of proteomics distinct studies were to a of glycopeptides to test GRAEZ studies a of glycoprotein M.J. and of low glycopeptides via dissociation and mass Am. Mass Spectrom. 2012; PubMed Scopus Google Scholar), chromatography electrospray ionization Fourier transform ion cyclotron resonance mass of N-linked glycans and Chem. 2010; PubMed Scopus Google Scholar), human A. J. C. urinary site-specific analysis of and by and PubMed Google Scholar), glycoproteins D. M. S. M. M. A. identification using a of and Proteome Res. 2007; 6: PubMed Scopus Google Scholar), human D. M. of site-specific glycosylation of from plasma using ion dissociation mass Chem. Google Scholar), human acid glycoprotein Go E.P. Desaire H. of glycosylation heterogeneity in analysis of glycoproteins with to glycosylation Chem. PubMed Scopus Google Scholar), glycoprotein M. Mass of glycosylation of glycoprotein of Am. Soc. Mass Spectrom. PubMed Scopus Google Scholar), glycoprotein J. Go E.P. D.S. Desaire H. of MS MS for glycopeptide analysis of a highly glycosylated Am. Soc. Mass Spectrom. PubMed Scopus (61) Google Scholar), and human M. of from human 2007; 7: PubMed Scopus Google Scholar). In intact tryptic glycopeptides were identified in these studies within the mass range of daltons. GRAEZ was performed on m/z values given and on m/z values GRAEZ of these species as glycopeptides in for an sensitivity of This that the sensitivity of GRAEZ classification was these on samples. on these data from and we GRAEZ classification will be useful for a of studies to identify likely in The utility of GRAEZ was further evaluated using tryptic peptides from is a highly sample and it and To the of nonpeptide the peptides were with analysis. Using the spectra was for two fragment ions as at which was required for the and the which was required for without either of these ions were not considered in the GRAEZ was chosen to challenge the GRAEZ as it is a highly complex sample containing of We any glycopeptide to challenge the GRAEZ it was that the of peptides observed would glycopeptides, as peptides are to ionization of glycopeptides, and are present at higher stoichiometric Furthermore, a recent A. J. C. urinary site-specific analysis of and by and PubMed Google Scholar) are in which are to further challenge the accuracy of N-linked glycopeptide Although and glycopeptides can yield the same small ions higher dissociation fragmentation, they will not to GRAEZ classification because an glycan moieties are glycopeptides are to be missed by GRAEZ due to the small of most glycan the of precursor ions likely that are to be less for GRAEZ This the classification to challenge that can be to in sample these challenges, an analysis of spectra that of all species that yielded by were as N-glycopeptides by the GRAEZ of all peptide species were as The of any glycopeptide enrichment to a challenging analytical In unique peptide were selected by for glycopeptide To are studies that intact glycopeptides and peptides because peptides and glycopeptides have distinct optimal J. K. and for glycopeptide J. Mass Spectrom. 2009; PubMed Scopus Google Scholar, J.W. M. C. B.H. Lebrilla C.B. of glycopeptides by nonspecific and site-specific glycosylation Chem. PubMed Scopus Google Scholar). the presence of species in a glycopeptide enrichment sample is an common of enrichment protocols. In this study, we to demonstrate the utility of GRAEZ classification within a highly complex and challenging analytical In these samples were using MS and was a of low quality spectra However, high quality glycopeptide were and the glycan were assigned by the presence of the ion in A systematic for in spectra of J. 5: Scopus Google Scholar) and a of fragment of higher quality spectra are shown in Fig. In each a to the glycan moieties was by of monosaccharide In a ion was and in Fig. of the was identified the mass of the peptide in to the glycan to a total of glycopeptides were with quality to the glycan of the glycopeptides in the data-dependent and information is in Table species were glycopeptides with The peptide values were assignment of the MS/MS by and identifying the the remaining mass from the precursor was determined and N-glycan compositions to the glycan an N-glycan at less mass The peptide were not in this and are reported as their An sample of urinary peptides was by MS/MS and by targeted MS. A total of species from the initial analysis were as glycopeptides by the A total of spectra were and of these an ion at a of of the A less of increased the to or of all spectra fractionated sample a glycopeptide of by spectra in more a targeted based on GRAEZ classification significantly increased the glycopeptide MS/MS and The GRAEZ approach has specific it an capable of making precursor m/z measurements with high mass Furthermore, an of the in silico data that the GRAEZ approach will be useful the precursor mass was no shown in Fig. not enrichment however, the data in this that may be determined proteomics the of the of N-glycopeptides presented (61) is with recent using glycopeptide enrichment A. J. C. urinary site-specific analysis of and by and PubMed Google Scholar) of human a of the peptide was not performed in this GRAEZ was to be for N-glycopeptides with large glycan The glycopeptide was at for glycopeptides with small and in from the Desaire has optimized the MD of peptides and the of MD Desaire H. mass defect for human Proteome Res. 2010; PubMed Scopus (13) Google Scholar). found the of of the peptide at to be the GRAEZ settings for glycopeptides are for the same We this is due to two the of peptides with a consensus is a of the peptide the the of for glycopeptides is by the glycan if an of a mass of has an N-linked attached, the of MD from the peptide is to a peptide of daltons. The glycan not the MD as as are glycan Notably, the of mass is with small post-translational such as and as the MD not shift a small mass We will be useful glycopeptide signals are suppressed such that they are not or are fragmented and are in a common in data-dependent analysis of complex mixtures. no information the sample was required no glycomic or proteomic experiments to be knowledge of the protease is and it may be performed glycopeptide increasing the utility of by peptide and improving the of established glycopeptide enrichment strategies by increasing glycopeptide in MS/MS analysis. We demonstrate may be performed an initial proteomics in of glycopeptide The large scale analysis of glycopeptides a major analytical In this work, a and specific applied to proteomic data that discriminates N-glycopeptides from nonglycosylated peptides based on accurate mass measurements has been within the GRAEZ are in glycopeptides by an of By identifying an targeted from an initial data-dependent the may efficiently glycopeptides in a subsequent This the of glycopeptides in a in which the is to preferentially select likely glycopeptide masses for Because classification is based on the mass of the this can be applied to glycoproteomic without the for prior knowledge the proteome or glycome GRAEZ classification may be used to quickly the of glycopeptide sample where the of or of tandem MS spectra would be to GRAEZ classification of existing proteomic be used to quickly the of glycosylated peptides in existing with no additional This will to quickly the of a glycoproteomic on samples. this approach represents a simple means of significantly improving glycoproteomic as we that mass defect filtering will to the of site-specific