Portal de Periódicos da CAPES

Comparison of Methods for Profiling O-Glycosylation

2009; Elsevier BV; Volume: 9; Issue: 4 Linguagem: Inglês

10.1074/mcp.m900450-mcp200

1535-9484

Yoshinao Wada, Anne Dell, Stuart M. Haslam, Bérangère Tissot, Kévin Canis, Parastoo Azadi, Malin Bäckström, Catherine E. Costello, Gunnar C. Hansson, Yoshiyuki Hiki, Mayumi Ishihara, Hiromi Ito, Kazuaki Kakehi, Niclas G. Karlsson, Catherine Hayes, Koichi Kato, Nana Kawasaki, Kay‐Hooi Khoo, Kunihiko Kobayashi, Daniel Kolarich, Akihiro Kondo, Carlito B. Lebrilla, Miyako Nakano, Hisashi Narimatsu, Jan Novák, Miloš V. Novotný, Erina Ohno, Nicolle H. Packer, Elizabeth Palaima, Matthew B. Renfrow, Michiko Tajiri, Kristina A. Thomsson, Hirokazu Yagi, Shin‐Yi Yu, Naoyuki Taniguchi,

Advanced Proteomics Techniques and Applications

The Human Proteome Organisation Human Disease Glycomics/Proteome Initiative recently coordinated a multi-institutional study that evaluated methodologies that are widely used for defining the N-glycan content in glycoproteins. The study convincingly endorsed mass spectrometry as the technique of choice for glycomic profiling in the discovery phase of diagnostic research. The present study reports the extension of the Human Disease Glycomics/Proteome Initiative's activities to an assessment of the methodologies currently used for O-glycan analysis. Three samples of IgA1 isolated from the serum of patients with multiple myeloma were distributed to 15 laboratories worldwide for O-glycomics analysis. A variety of mass spectrometric and chromatographic procedures representative of current methodologies were used. Similar to the previous N-glycan study, the results convincingly confirmed the pre-eminent performance of MS for O-glycan profiling. Two general strategies were found to give the most reliable data, namely direct MS analysis of mixtures of permethylated reduced glycans in the positive ion mode and analysis of native reduced glycans in the negative ion mode using LC-MS approaches. In addition, mass spectrometric methodologies to analyze O-glycopeptides were also successful. The Human Proteome Organisation Human Disease Glycomics/Proteome Initiative recently coordinated a multi-institutional study that evaluated methodologies that are widely used for defining the N-glycan content in glycoproteins. The study convincingly endorsed mass spectrometry as the technique of choice for glycomic profiling in the discovery phase of diagnostic research. The present study reports the extension of the Human Disease Glycomics/Proteome Initiative's activities to an assessment of the methodologies currently used for O-glycan analysis. Three samples of IgA1 isolated from the serum of patients with multiple myeloma were distributed to 15 laboratories worldwide for O-glycomics analysis. A variety of mass spectrometric and chromatographic procedures representative of current methodologies were used. Similar to the previous N-glycan study, the results convincingly confirmed the pre-eminent performance of MS for O-glycan profiling. Two general strategies were found to give the most reliable data, namely direct MS analysis of mixtures of permethylated reduced glycans in the positive ion mode and analysis of native reduced glycans in the negative ion mode using LC-MS approaches. In addition, mass spectrometric methodologies to analyze O-glycopeptides were also successful. Recently, the Human Proteome Organisation Human Disease Glycomics/Proteome Initiative (HGPI) 1The abbreviations used are:HGPIHuman Disease Glycomics/Proteome InitiativeHexhexoseHexNAcN-acetylhexosamineNHexNAc (GalNAc)HHex (Gal)NANeuAc (N-acetylneuraminic acid)ABHE2-(2-aminobenzoylamino)-2-hydrazinocarbonylethanethiolLTQlinear trap quadrupole. coordinated an evaluation of methodologies that are widely used for defining the N-glycan content in glycoproteins (1.Wada Y. Azadi P. Costello C.E. Dell A. Dwek R.A. Geyer H. Geyer R. Kakehi K. Karlsson N.G. Kato K. Kawasaki N. Khoo K.H. Kim S. Kondo A. Lattova E. Mechref Y. Miyoshi E. Nakamura K. Narimatsu H. Novotny M.V. Packer N.H. Perreault H. Peter-Katalinic J. Pohlentz G. Reinhold V.N. Rudd P.M. Suzuki A. Taniguchi N. Comparison of the methods for profiling glycoprotein glycans—HUPO Human Disease Glycomics/Proteome Initiative multi-institutional study.Glycobiology. 2007; 17: 411-422Crossref PubMed Scopus (354) Google Scholar). Twenty laboratories around the world participated in the study in which the glycosylation of standard samples of transferrin and IgG was characterized by a variety of chromatographic and mass spectrometric techniques. Two clear messages emerged from this study. First, there was significant variance among the data sets from laboratories using chromatographic profiling that was likely due to incomplete derivatization with the fluorophores that had been used to "tag" the oligosaccharides to facilitate chromatography and provide a means of detection. Second, MS was shown to give consistent data in interlaboratory comparisons, and it was concluded that MS-based strategies provide the most effective means of both identification and quantitation of N-glycans in glycomics studies. Human Disease Glycomics/Proteome Initiative hexose N-acetylhexosamine HexNAc (GalNAc) Hex (Gal) NeuAc (N-acetylneuraminic acid) 2-(2-aminobenzoylamino)-2-hydrazinocarbonylethanethiol linear trap quadrupole. HGPI has now extended its comparison of analytical methodologies to encompass mucin-type O-glycosylation. In this type of glycosylation, O-glycans are attached by a GalNAc to the amino acids serine and threonine. They can occur as single O-glycans or clustered in mucin domains. Such domains are most abundant in the class of glycoproteins known as mucins, which typically have a great number of mucin domains arranged as tandem repeats, but are also found in many extracellular proteins like IgA, which is the subject of the current study. IgA1 represents one of two structurally and functionally distinct subclasses of IgA (2.Mestecky J. Immunobiology of IgA.Am. J. Kidney Dis. 1988; 12: 378-383Abstract Full Text PDF PubMed Scopus (45) Google Scholar). The heavy chains of IgA1 molecules contain a hinge region segment between the first and second constant region domains. This segment, which has a high content of Pro, Ser, and Thr, is the site of attachment of usually up to six O-linked glycan chains (3.Baenziger J. Kornfeld S. Structure of the carbohydrate units of IgA1 immunoglobulin. II. Structure of the O-glycosidically linked oligosaccharide units.J. Biol. Chem. 1974; 249: 7270-7281Abstract Full Text PDF PubMed Google Scholar, 4.Field M.C. Dwek R.A. Edge C.J. Rademacher T.W. O-linked oligosaccharides from human serum immunoglobulin A1.Biochem. Soc. Trans. 1989; 17: 1034-1035Crossref PubMed Scopus (82) Google Scholar, 5.Mattu T.S. Pleass R.J. Willis A.C. Kilian M. Wormald M.R. Lellouch A.C. Rudd P.M. Woof J.M. Dwek R.A. The glycosylation and structure of human serum IgA1, Fab, and Fc regions and the role of N-glycosylation on Fc alpha receptor interactions.J. Biol. Chem. 1998; 273: 2260-2272Abstract Full Text Full Text PDF PubMed Scopus (360) Google Scholar, 6.Renfrow M.B. Cooper H.J. Tomana M. Kulhavy R. Hiki Y. Toma K. Emmett M.R. Mestecky J. Marshall A.G. Novak J. Determination of aberrant O-glycosylation in the IgA1 hinge region by electron capture dissociation Fourier transform-ion cyclotron resonance mass spectrometry.J. Biol. Chem. 2005; 280: 19136-19145Abstract Full Text Full Text PDF PubMed Scopus (125) Google Scholar, 7.Renfrow M.B. Mackay C.L. Chalmers M.J. Julian B.A. Mestecky J. Kilian M. Poulsen K. Emmett M.R. Marshall A.G. Novak J. Analysis of O-glycan heterogeneity in IgA1 myeloma proteins by Fourier transform ion cyclotron resonance mass spectrometry: implications for IgA nephropathy.Anal. Bioanal. Chem. 2007; 389: 1397-1407Crossref PubMed Scopus (81) Google Scholar, 8.Tarelli E. Smith A.C. Hendry B.M. Challacombe S.J. Pouria S. Human serum IgA1 is substituted with up to six O-glycans as shown by matrix assisted laser desorption ionisation time-of-flight mass spectrometry.Carbohydr. Res. 2004; 339: 2329-2335Crossref PubMed Scopus (86) Google Scholar, 9.Tomana M. Kulhavy R. Mestecky J. Receptor-mediated binding and uptake of immunoglobulin A by human liver.Gastroenterology. 1988; 94: 762-770Abstract Full Text PDF PubMed Scopus (135) Google Scholar). In circulatory IgA1, these O-glycans consist of GalNAc with a β1,3-linked Gal; both saccharides may be sialylated (3.Baenziger J. Kornfeld S. Structure of the carbohydrate units of IgA1 immunoglobulin. II. Structure of the O-glycosidically linked oligosaccharide units.J. Biol. Chem. 1974; 249: 7270-7281Abstract Full Text PDF PubMed Google Scholar, 4.Field M.C. Dwek R.A. Edge C.J. Rademacher T.W. O-linked oligosaccharides from human serum immunoglobulin A1.Biochem. Soc. Trans. 1989; 17: 1034-1035Crossref PubMed Scopus (82) Google Scholar). The carbohydrate composition of the O-linked glycans in the hinge region of normal human serum IgA1 is variable. The prevailing forms include Gal-GalNAc disaccharide and its mono- and disialylated forms (4.Field M.C. Dwek R.A. Edge C.J. Rademacher T.W. O-linked oligosaccharides from human serum immunoglobulin A1.Biochem. Soc. Trans. 1989; 17: 1034-1035Crossref PubMed Scopus (82) Google Scholar, 5.Mattu T.S. Pleass R.J. Willis A.C. Kilian M. Wormald M.R. Lellouch A.C. Rudd P.M. Woof J.M. Dwek R.A. The glycosylation and structure of human serum IgA1, Fab, and Fc regions and the role of N-glycosylation on Fc alpha receptor interactions.J. Biol. Chem. 1998; 273: 2260-2272Abstract Full Text Full Text PDF PubMed Scopus (360) Google Scholar, 10.Novak J. Tomana M. Kilian M. Coward L. Kulhavy R. Barnes S. Mestecky J. Heterogeneity of O-glycosylation in the hinge region of human IgA1.Mol. Immunol. 2000; 37: 1047-1056Crossref PubMed Scopus (62) Google Scholar). Gal-deficient variants with terminal GalNAc or sialylated GalNAc are rarely found in the O-glycans of normal serum IgA1 (5.Mattu T.S. Pleass R.J. Willis A.C. Kilian M. Wormald M.R. Lellouch A.C. Rudd P.M. Woof J.M. Dwek R.A. The glycosylation and structure of human serum IgA1, Fab, and Fc regions and the role of N-glycosylation on Fc alpha receptor interactions.J. Biol. Chem. 1998; 273: 2260-2272Abstract Full Text Full Text PDF PubMed Scopus (360) Google Scholar) but are much more common in IgA nephritis patients (11.Moldoveanu Z. Wyatt R.J. Lee J.Y. Tomana M. Julian B.A. Mestecky J. Huang W.Q. Anreddy S.R. Hall S. Hastings M.C. Lau K.K. Cook W.J. Novak J. Patients with IgA nephropathy have increased serum galactose-deficient IgA1 levels.Kidney Int. 2007; 71: 1148-1154Abstract Full Text Full Text PDF PubMed Scopus (304) Google Scholar, 12.Suzuki H. Moldoveanu Z. Hall S. Brown R. Vu H.L. Novak L. Julian B.A. Tomana M. Wyatt R.J. Edberg J.C. Alarcón G.S. Kimberly R.P. Tomino Y. Mestecky J. Novak J. IgA1-secreting cell lines from patients with IgA nephropathy produce aberrantly glycosylated IgA1.J. Clin. Investig. 2008; 118: 629-639PubMed Google Scholar) in whom IgG autoantibodies reactive to the hypogalactosylated IgA1 form immune complexes and lead to mesangioproliferative glomerulonephritis (13.Suzuki H. Fan R. Zhang Z. Brown R. Hall S. Julian B.A. Chatham W.W. Suzuki Y. Wyatt R.J. Moldoveanu Z. Lee J.Y. Robinson J. Tomana M. Tomino Y. Mestecky J. Novak J. Aberrantly glycosylated IgA1 in IgA nephropathy patients is recognized by IgG antibodies with restricted heterogeneity.J. Clin. Investig. 2009; 119: 1668-1677PubMed Google Scholar). These pathological features clearly demonstrate the importance of determining the profile of the total glycan pool as an initial but essential step in tackling the complex O-glycan structures. In the present study, IgA1 preparations from three patients with multiple myeloma were delivered by HGPI to 15 experienced academic laboratories, and the results of O-glycomics analysis, especially the total glycoform profiles, obtained using different analytical methodologies were assessed. This study was designed to compare and evaluate various methods differing in sample preparation and analytical modes as well as to document levels of variance or consistency among the data. Three IgA1 myeloma proteins named "NUD," "VDS," and "SapII" were isolated from respective serum samples from three patients with multiple myeloma by precipitation with 50% saturated ammonium sulfate followed by gel filtration chromatography on Sepharose 6B and by ion exchange chromatography on DEAE-cellulose (14.Kobayashi K. Vaerman J.P. Heremans J.F. J-chain determinants in polymeric immunoglobulins.Eur. J. Immunol. 1973; 3: 185-191Crossref PubMed Scopus (34) Google Scholar, 15.Mestecky J. Hamilton R.G. Magnusson C.G. Jefferis R. Vaerman J.P. Goodall M. de Lange G.G. Moro I. Aucouturier P. Radl J. Cambiaso C. Silvain C. Preud'homme J.L. Kusama K. Carlone G.M. Biewenga J. Kobayashi K. Skvaril F. Reimer C.B. Evaluation of monoclonal antibodies with specificity for human IgA, IgA subclasses and allotypes and secretory component. Results of an IUIS/WHO collaborative study.J. Immunol. Methods. 1996; 193: 103-148Crossref PubMed Scopus (32) Google Scholar). Purity was assessed by immunoelectrophoresis with rabbit anti-human serum proteins. These samples were provided by one of the authors (K. K.). Each 250-μg sample was delivered with dry ice in a solution for NUD at 250 μg/ml in 6 m guanidine/HCl, 0.25 m Tris-HCl, 1% dithiothreitol (pH 8.6) and for Sap-II at 2 mg/ml in 20 mm Tris-HCl, 2% NaCl, 0.1% NaN3 (pH 8.0) or a lyophilized form for VDS. The analytical methods used by each participating laboratory are given in the supplemental material. Human IgA1 has both N- and O-glycosylation with the latter being the focus of this study. The O-glycans are located on the hinge region of the heavy chain, and five sites of glycosylation have been identified previously (Thr-225, Thr-228, Ser-230, Ser-232, and Thr-236), two of which (225 and 236) have been reported to be partially occupied (3.Baenziger J. Kornfeld S. Structure of the carbohydrate units of IgA1 immunoglobulin. II. Structure of the O-glycosidically linked oligosaccharide units.J. Biol. Chem. 1974; 249: 7270-7281Abstract Full Text PDF PubMed Google Scholar, 4.Field M.C. Dwek R.A. Edge C.J. Rademacher T.W. O-linked oligosaccharides from human serum immunoglobulin A1.Biochem. Soc. Trans. 1989; 17: 1034-1035Crossref PubMed Scopus (82) Google Scholar, 5.Mattu T.S. Pleass R.J. Willis A.C. Kilian M. Wormald M.R. Lellouch A.C. Rudd P.M. Woof J.M. Dwek R.A. The glycosylation and structure of human serum IgA1, Fab, and Fc regions and the role of N-glycosylation on Fc alpha receptor interactions.J. Biol. Chem. 1998; 273: 2260-2272Abstract Full Text Full Text PDF PubMed Scopus (360) Google Scholar, 6.Renfrow M.B. Cooper H.J. Tomana M. Kulhavy R. Hiki Y. Toma K. Emmett M.R. Mestecky J. Marshall A.G. Novak J. Determination of aberrant O-glycosylation in the IgA1 hinge region by electron capture dissociation Fourier transform-ion cyclotron resonance mass spectrometry.J. Biol. Chem. 2005; 280: 19136-19145Abstract Full Text Full Text PDF PubMed Scopus (125) Google Scholar). Tryptic cleavage yields the 38-amino acid hinge region O-glycopeptide: HYTNPSQDVTVPCPVPST225PPT228PS230PS232TPPT236PSPSCCHPR. The mass of core peptide containing carbamidomethylated cysteine residues is 4135.88 (monoisotopic) or 4138.56 (average). The 15 laboratories participating in the study analyzed tryptic glycopeptides, released oligosaccharides, or both. General work flows are shown in Schemes 1 and 2, and methodologies used by each laboratory are summarized in Table I. IgA1 derived from three patients with multiple myeloma (coded NUD, Sap-II, and VDS) (14.Kobayashi K. Vaerman J.P. Heremans J.F. J-chain determinants in polymeric immunoglobulins.Eur. J. Immunol. 1973; 3: 185-191Crossref PubMed Scopus (34) Google Scholar, 15.Mestecky J. Hamilton R.G. Magnusson C.G. Jefferis R. Vaerman J.P. Goodall M. de Lange G.G. Moro I. Aucouturier P. Radl J. Cambiaso C. Silvain C. Preud'homme J.L. Kusama K. Carlone G.M. Biewenga J. Kobayashi K. Skvaril F. Reimer C.B. Evaluation of monoclonal antibodies with specificity for human IgA, IgA subclasses and allotypes and secretory component. Results of an IUIS/WHO collaborative study.J. Immunol. Methods. 1996; 193: 103-148Crossref PubMed Scopus (32) Google Scholar) were prepared in Japan and dispatched to the participating laboratories in 250-μg aliquots. Several of the participating laboratories additionally analyzed a sample of IgA pooled from healthy individuals.Scheme 2Work flow of various methodologies used for O-glycan structural analysis. Numbers in parentheses are laboratory numbers. 2-AA, 2-aminobenzoic acid.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Table ISummary of the methodologies/instrumentation usedAnalyte/preparationAnalysis strategyMS instrumentLab 1Gp/RA, PrSep, MALDI-MS, ESI-MSABI Voyager MALDI-TOF DE ProOS/β-elimPe, MALDI-MSThermo LTQ with ETDLab 2Gp/RA, Pn, PrOn-line RP-LC ESI-MSThermo LTQ-FT-ICROS/HyOn-line Ca-LC ESI-MSThermo LTQ-FT-ICRLab 3Gp/RA, PrJac, HPLC, on-line RP-LC ESI-MSThermo LCQ Deca XPLab 4Gp/RA, Pr, PnSep, on-line RP-LC ESI-MSThermo OrbitrapLab 5Gp/RA, SDS-PAGE, PrOn-line AE-LC ESI-MSWaters Ultima Q-TOFOS/RA, SDS-PAGE, β-elimOn-line Ca-LC ESI-MSBruker HCT Ultra PTM Discovery SystemLab 6Gp/RA, SDS-PAGE, PrOn-line RP-LC ESI-MSThermo LTQ-FT-ICRLab 7OS/Hy, BlotglycoABHE, HPLCLab 8OS/AGC2-AA, HPLC, MALDI-MS, on-line LC ESI-MSABI Voyager DE Pro MALDI-TOF; Shimadzu LC MS-IT TOFLab 9OS/Pn, β-elimPe, MALDI-MSBruker Daltonics GmbH Reflex IV MALDI-TOF, Shimadzu AXIMA-QIT MALDI Quadrupole Ion Trap TOFLab 10OS/β-elimPe, MALDI-MSABI 4700 Proteomics analyzer; Waters MALDI Q-TOF UltimaLab 11OS/β-elimPe, MALDI-MS, ESI-MSABI 4700 Proteomics Analyzer; Thermo LCQ-MSLab 12OS/β-elimPe, MALDI-MSMALDILab 13OS/SDS-PAGE, β-elimOn-line Ca-LC ESI-MSThermo LTQLab 14OS/RA, SDS-PAGE, β-elimOn-line LC ESI-MSAgilent 3D Ion TrapLab 15OS/Pn, β-elimMALDI-MS, ESI-MSIonSpec FT-ICR Open table in a new tab MALDI and/or ESI mass fingerprinting of cysteine-alkylated tryptic hinge O-glycopeptides was carried out by six laboratories. Data of remarkable consistency were obtained from most of these laboratories despite the use of a variety of sample handling procedures and mass spectrometric instrumentation (see Table I and supplemental methods). Exemplar MALDI and ESI mass spectra from O-glycopeptides isolated from solution digests using hydrophilic affinity extraction (16.Wada Y. Tajiri M. Yoshida S. Hydrophilic affinity isolation and MALDI multiple-stage tandem mass spectrometry of glycopeptides for glycoproteomics.Anal. Chem. 2004; 76: 6560-6565Crossref PubMed Scopus (289) Google Scholar) by two laboratories (labs 1 and 4) are shown in Fig. 1 and supplemental Fig. S1, respectively. The enriched glycopeptide fraction was directly analyzed by MALDI-MS in linear TOF mode (lab 1) or subjected to on-line LC-ESI-MS (lab 4). Profiles obtained from on-line LC-ESI-MS of solution digests and by in-gel digestion followed by LC-ESI-MS are shown in Fig. 2 and supplemental Figs. S2 and S3 (labs 2, 5, and 6, respectively). Some laboratories additionally analyzed a pooled sample of IgA1 purified from healthy individuals. An exemplar mass fingerprint for the hinge glycopeptide of normal IgA1 is shown in supplemental Fig. S4. For simplicity, molecular ions of the glycopeptide glycoforms are annotated in all spectra with their predicted oligosaccharide compositions using N, H, and NA to represent HexNAc (GalNAc), Hex (Gal), and NeuAc (N-acetylneuraminic acid), respectively.Fig. 2ESI mass spectra of IgA O-glycopeptides of Sap-II, VDS, and NUD samples (lab 2). a, deconvoluted spectra of three samples. Monoisotopic masses of [M + H]+ ions are indicated in the mass spectrum of Sap-II. b, real ESI mass spectrum of Sap-II before deconvolution. The charge state of each peptide ion signal is given in parentheses. Samples were analyzed by LC-MS in positive ion mode using a Paradigm MS4 HPLC system coupled to a Thermo FT-ICR mass spectrometer. Mass accuracy was ±0.1 Da.View Large Image Figure ViewerDownload Hi-res image Download (PPT) Sialylation was prominent in Sap-II, and heterogeneity was small in NUD in all the data sets. However, some differences were also observed. Typically, sialylated glycans were less abundant in the MALDI mass spectra (lab 1) when looking at the ratio of 4N4H/4N4H1NA/4N4H2NA/4N4H3NA in Sap-II (Fig. 3) or that of 4N4H1/4N4H1NA in VDS in Fig. 1 and supplemental Figs. S1 and S2. This was likely due to low efficiency of the ionization of sialylated glycopeptides in the positive ion mode because the underestimation was less prominent in the MALDI mass spectra taken in the negative mode (see supplemental Fig. S5 from lab 1). Loss of sialyl residues is the major drawback of MALDI of oligosaccharides or glycopeptides, and this was observed, even when using the negative ionization mode, especially in the glycopeptide ions with multiple sialic acids: e.g. 5N4H3NA and 4N4H4NA (supplemental Fig. S5). However, the extent of sialic acid loss from O-glycans is less prominent than the cases of N-glycans. 2Y. Wada and M. Tajiri, unpublished observations. On the other hand, the results from ESI from three laboratories, lab 2 (Fig. 2), lab 4 (supplemental Fig. S1), and lab 5 (supplementary Fig. S2), also showed some discrepancies in the sialylated species. As indicated by the ratio of 4N4H to 4N4H1NA of NUD, sialylation levels were higher in the analysis of lab 4 than in those of lab 2 and lab 5. From the mass spectrometric point of view, one issue that should be clarified to address the disagreement is the charge state distribution of the original mass spectra from these laboratories or the charge state of the source mass spectra for deconvolution. Alternatively, the disagreement may be due to the usage of different types of mass analyzers and detectors. In addition, lab 5 reported predominance of small glycans probably due to different sample preparation (Fig. 3); lab 5 extracted glycopeptides from SDS-PAGE gels by in-gel digestion. One laboratory (lab 3) used jacalin lectin chromatography to isolate the glycopeptides, but the yields were found to be insufficient for a comprehensive MS study, and the limited data that were obtained differed substantially from the other laboratories. Considering that this laboratory has sufficient experience of the analysis of hinge O-glycopeptides obtained from a few mg of IgA1, their protocol was not suited for a smaller scale analysis; probably the volume (2 ml) of jacalin was too great to efficiently recover the glycopeptides from the lectin column. Additional analyses to decrease the complexity of the MS profiles were carried out by lab 1 (supplemental Figs. S6 and S7). Desialylation removed the negative effect of sialic acid on the ionization efficiency in the MALDI, and the small signals corresponding to highly glycosylated glycopeptides with six or seven HexNAc residues could be observed (supplemental Fig. S6). The number of O-glycans attached to the glycopeptides was clearly demonstrated by the mass spectra after eliminating peripheral glycans attached to the Ser/Thr-GalNAc units by incubation with trifluoromethanesulfonic acid (supplemental Fig. S7). However, the current protocol of reactions did not guarantee good reproducibility, and some removal of GalNAc residues was observed. Deglycosylation by glycosidase (galactosidase) would be an alternative to this chemical treatment, but the removal of galactose residues was incomplete in a preliminary experiment carried out by lab 1 (data not shown). The major advantages of glycopeptide analysis are clearly illustrated in this pilot study as follows. First, the average or maximum/minimum number of attached sites can be counted. Second, relative abundances of different glycoform compositions can be readily established at high sensitivity. Third, MS of glycopeptides does not miss the GalNAc-α-O-Ser/Thr (Tn epitope), which is characterized by components having a higher number of HexNAc than Hex such as 5N4H or 5N3H. Chromatography of reductively aminated oligosaccharides carrying fluorescent "tags" at their reducing ends is a well established procedure for quantitating mixtures of N-glycans. It remains a popular method despite issues of variance between laboratories (1.Wada Y. Azadi P. Costello C.E. Dell A. Dwek R.A. Geyer H. Geyer R. Kakehi K. Karlsson N.G. Kato K. Kawasaki N. Khoo K.H. Kim S. Kondo A. Lattova E. Mechref Y. Miyoshi E. Nakamura K. Narimatsu H. Novotny M.V. Packer N.H. Perreault H. Peter-Katalinic J. Pohlentz G. Reinhold V.N. Rudd P.M. Suzuki A. Taniguchi N. Comparison of the methods for profiling glycoprotein glycans—HUPO Human Disease Glycomics/Proteome Initiative multi-institutional study.Glycobiology. 2007; 17: 411-422Crossref PubMed Scopus (354) Google Scholar) because it does not require highly sophisticated and expensive instrumentation. Exploitation of this type of methodology in the O-glycan field is far less common because free reducing sugars are required for the tagging reaction, and no tools are available that are capable of liberating O-glycans efficiently and cleanly from glycopeptides or glycoproteins while retaining their reducing sugars. Unfortunately, no broad spectrum O-glycanase has yet been discovered. Therefore, the release of O-glycans requires base-catalyzed chemical elimination, and it has been known for over 40 years that the core 1 arm of O-glycans is readily "peeled" from the 3-position of the reducing GalNAc residue under basic conditions unless the latter is reduced. Hence, reductive elimination is the only way of obtaining an artifact-free O-glycan population. Nevertheless, chromatographic/tagging methodologies can play a useful role in O-glycan analysis, for example in the analysis of complex mucins (17.Xia B. Royall J.A. Damera G. Sachdev G.P. Cummings R.D. Altered O-Glycosylation and sulfation of airway mucins associated with cystic fibrosis.Glycobiology. 2005; 15: 747-775Crossref PubMed Scopus (135) Google Scholar), provided peeling products and other chemical artifacts are taken into account in data analysis. In the present study, three laboratories (labs 2, 7, and 8) analyzed base-eliminated fluorescently tagged glycans, and their data nicely encapsulate the issues that need to be considered when such methods are used. O-Glycans were released manually with anhydrous hydrazine (labs 2 and 7) or automatically with lithium hydroxide (lab 8) or were tagged with phenylhydrazine, 2-(2-aminobenzoylamino)-2-hydrazinocarbonyl-ethanethiol (ABHE), or 2-aminobenzoic acid, respectively. The ABHE tagging procedure involved a nanoparticle immobilization/derivatization step and subsequent release of the tagged O-glycans (see supplemental methods). Work flows and protocols are given in Scheme 2 and supplementary methods, and chromatographic profiles are given in supplemental Figs. S8–S10. Comparative quantitative data are shown in Table II. The stark discrepancies between the data sets highlight the artifact problems mentioned earlier. Nevertheless, where replicate experiments were carried out, reasonable consistency was observed, so it can be concluded that comparative chromatographic profiling is a valid technique under carefully controlled conditions.Table IIComparative data obtained by labs 2, 7 and 8 using chromatography analyses Nine labs analyzed reductively eliminated O-glycans using three general MS strategies: (i) positive ion mode MALDI-MS fingerprinting supplemented by MALDI and/or ESI-MS/MS sequencing of mixtures of permethylated glycans (labs 1 and 9–12), (ii) negative ion mode ESI-MS fingerprinting and ESI-MS/MS sequencing of native glycans that were purified and separated by graphitized carbon on-line LC (labs 5, 13, and 14), and (iii) negative and positive ion mode MALDI-FT-ICR-MS of mixtures of native glycans without on-line LC purification (lab 15). Work flows and protocols are given in Scheme 2 and supplemental methods, respectively, and typical data are shown in supplemental Figs. S11 and S12. Quantitative information was extracted from MALDI fingerprints and LC-MS profiles by measuring peak heights and peak areas, respectively. With one exception (lab 13; see supplemental methods), no corrections were made for differences in the response factors of the various glycans. Exemplar quantitative data are shown in Table III for permethylated and native samples, respectively, and data from all eight labs are collated in supplemental Figs. S8–S14. Collectively, these data are broadly in line with the conclusions of the glycopeptide profiling experiments (see above), namely that most of the glycans are core 1 type with monosialylated core 1 dominating in Sap-II and non-sialylated core 1 being the major glycan in VDS and NUD. For technical reasons, the neutral glycans were missed in some of the on-line experiments. On-line graphitized carbon LC will not detect the Tn antigen as single monosaccharides are not retained. Minor core 2 glycans were most reliably observed in the analyses of permethylated samples (supplemental Figs. S11 and S13 for confirmation via tandem MS). A weakness of both methodologies is that monosaccharides such as Tn are difficult to analyze reliably because of masking by matrix peaks in MALDI data and early elution together with impurities in on-line LC-MS.Table IIIO-glycans comparative results obtained by lab 9 (MALDI-TOF MS) and 13 (LC-ESI MS) In summary, the study has shown that both strategies i and ii are reliable strategies for semiquantitative O-glycan analysis except for the analysis of Tn. On the other hand, strategy iii in which native samples were analyzed as mixtures by MS without permethylation or on-line LC purification gave equivocal data. Notably, spectra were characterized by dominant matrix and artifact peaks, and only one molecular ion species was assignable to an authentic IgA O-glycan (see supplemental Fig. S14). It was also observed in some experiments that small quantities of glycopeptides (with two or more amino acids) are recovered from reductive elimination reactions, and their presence may need to be taken into account when data are analyzed. In this study, the O-glycan content of three IgA1 samples was analyzed by 15 participating laboratories. Six of these laboratories used either MALDI or ESI to obtain mass fingerprints of the hinge glycopeptide (which has up to five occupied O-glycosylation sites), three laboratories eliminated the O-glycans without reduction and then tagged the free reducing ends prior to chromatographic analysis, and nine laboratories used reductive elimination to release the glycans, which were then analyzed by MS after permethylation (five laboratories) or as native glycans by LC-MS (three laboratories) or MS alone (one laboratory). The main conclusions that can be drawn from the study are summarized below.