N-Glycosylation Site Occupancy in Serum Glycoproteins Using Multiple Reaction Monitoring Liquid Chromatography-Mass Spectrometry
2007; Elsevier BV; Volume: 6; Issue: 12 Linguagem: Inglês
10.1074/mcp.m700361-mcp200
ISSN1535-9484
AutoresAndreas J. Hülsmeier, Patricie Paesold-Burda, Thierry Hennet,
Tópico(s)RNA and protein synthesis mechanisms
ResumoCongenital disorders of glycosylation (CDGs) are a family of N-linked glycosylation defects associated with severe clinical manifestations. In CDG type-I, deficiency of lipid-linked oligosaccharide assembly leads to the underoccupancy of N-glycosylation sites on glycoproteins. Although the level of residual glycosylation activity is known to correlate with the clinical phenotype linked to individual CDG mutations, it is not known whether the degree of N-glycosylation site occupancy by itself correlates with the severity of the disease. To quantify the extent of underglycosylation in healthy control and in CDG samples, we developed a quantitative method of N-glycosylation site occupancy based on multiple reaction monitoring LC-MS/MS. Using isotopically labeled standard peptides, we directly quantified the level of N-glycosylation site occupancy on selected serum proteins. In healthy control samples, we determined 98–100% occupancy for all N-glycosylation sites of transferrin and α1-antitrypsin. In CDG type-I samples, we observed a reduction in N-glycosylation site occupancy that correlated with the severity of the disease. In addition, we noticed a selective underglycosylation of N-glycosylation sites, indicating preferential glycosylation of acceptor sequons of a given glycoprotein. In transferrin, a preferred occupancy for the first N-glycosylation site was observed, and a decreasing preference for the first, third, and second N-glycosylation sites was observed in α1-antitrypsin. This multiple reaction monitoring LC-MS/MS method can be extended to multiple glycoproteins, thereby enabling a glycoproteomics survey of N-glycosylation site occupancies in biological samples. Congenital disorders of glycosylation (CDGs) are a family of N-linked glycosylation defects associated with severe clinical manifestations. In CDG type-I, deficiency of lipid-linked oligosaccharide assembly leads to the underoccupancy of N-glycosylation sites on glycoproteins. Although the level of residual glycosylation activity is known to correlate with the clinical phenotype linked to individual CDG mutations, it is not known whether the degree of N-glycosylation site occupancy by itself correlates with the severity of the disease. To quantify the extent of underglycosylation in healthy control and in CDG samples, we developed a quantitative method of N-glycosylation site occupancy based on multiple reaction monitoring LC-MS/MS. Using isotopically labeled standard peptides, we directly quantified the level of N-glycosylation site occupancy on selected serum proteins. In healthy control samples, we determined 98–100% occupancy for all N-glycosylation sites of transferrin and α1-antitrypsin. In CDG type-I samples, we observed a reduction in N-glycosylation site occupancy that correlated with the severity of the disease. In addition, we noticed a selective underglycosylation of N-glycosylation sites, indicating preferential glycosylation of acceptor sequons of a given glycoprotein. In transferrin, a preferred occupancy for the first N-glycosylation site was observed, and a decreasing preference for the first, third, and second N-glycosylation sites was observed in α1-antitrypsin. This multiple reaction monitoring LC-MS/MS method can be extended to multiple glycoproteins, thereby enabling a glycoproteomics survey of N-glycosylation site occupancies in biological samples. In humans, disturbances in the biosynthesis of glycoconjugates lead to diseases with heterogeneous biochemical and clinical characteristics (1Ohtsubo K. Marth J.D. Glycosylation in cellular mechanisms of health and disease.Cell. 2006; 126: 855-867Abstract Full Text Full Text PDF PubMed Scopus (2049) Google Scholar). Deficiencies in the biosynthesis of serine- or threonine-linked (O-linked), asparagine-linked (N-linked), glycosylphosphatidylinositol, and glycosphingolipid glycans have been described (2Topaz O. Shurman D.L. Bergman R. Indelman M. Ratajczak P. Mizrachi M. Khamaysi Z. Behar D. Petronius D. Friedman V. Zelikovic I. Raimer S. Metzker A. Richard G. Sprecher E. Mutations in GALNT3, encoding a protein involved in O-linked glycosylation, cause familial tumoral calcinosis.Nat. Genet. 2004; 36: 579-581Crossref PubMed Scopus (459) Google Scholar, 3Freeze H.H. Aebi M. Altered glycan structures: the molecular basis of congenital disorders of glycosylation.Curr. Opin. Struct. Biol. 2005; 15: 490-498Crossref PubMed Scopus (203) Google Scholar, 4Almeida A.M. Murakami Y. Layton D.M. Hillmen P. Sellick G.S. Maeda Y. Richards S. Patterson S. Kotsianidis I. Mollica L. Crawford D.H. Baker A. Ferguson M. Roberts I. Houlston R. Kinoshita T. Karadimitris A. Hypomorphic promoter mutation in PIGM causes inherited glycosylphosphatidylinositol deficiency.Nat. Med. 2006; 12: 846-851Crossref PubMed Scopus (183) Google Scholar, 5Simpson M.A. Cross H. Proukakis C. Priestman D.A. Neville D.C. Reinkensmeier G. Wang H. Wiznitzer M. Gurtz K. Verganelaki A. Pryde A. Patton M.A. Dwek R.A. Butters T.D. Platt F.M. Crosby A.H. Infantile-onset symptomatic epilepsy syndrome caused by a homozygous loss-of-function mutation of GM3 synthase.Nat. Genet. 2004; 36: 1225-1229Crossref PubMed Scopus (320) Google Scholar). Inherited deficiencies in the N-glycosylation biosynthetic pathway are referred to as congenital disorders of glycosylation (CDGs). 1The abbreviations used are: CDG, congenital disorder of glycosylation; HFI, hereditary fructose intolerance; MRM, multiple reaction monitoring; PNGaseF, peptide N-glycanase F; EPI, enhanced product ion. 1The abbreviations used are: CDG, congenital disorder of glycosylation; HFI, hereditary fructose intolerance; MRM, multiple reaction monitoring; PNGaseF, peptide N-glycanase F; EPI, enhanced product ion. CDGs represent multiorgan diseases involving central and peripheral nervous defects and are often associated with endocrine and coagulation disorders (6Jaeken J. Matthijs G. Congenital disorders of glycosylation: a rapidly expanding disease family.Annu. Rev. Genomics Hum. Genet. 2007; 8: 261-278Crossref PubMed Scopus (232) Google Scholar). The heterogeneous symptoms associated to CDG and the structural diversity of glycoconjugate render the identification of glycosylation defects a difficult task. After initial clinical suspicion of CDG, the serum glycoprotein transferrin serves as a biomarker for possible defects of N-glycosylation. The integrity of transferrin N-glycosylation is tested by isoelectric focusing electrophoresis, Western blot analysis, and MS measurements of the purified glycoprotein (7Lacey J.M. Bergen H.R. Magera M.J. Naylor S. O'Brien J.F. Rapid determination of transferrin isoforms by immunoaffinity liquid chromatography and electrospray mass spectrometry.Clin. Chem. 2001; 47: 513-518Crossref PubMed Scopus (169) Google Scholar, 8Mills K. Mills P. Jackson M. Worthington V. Beesley C. Mann A. Clayton P. Grünewald S. Keir G. Young L. Langridge J. Mian N. Winchester B. Diagnosis of congenital disorders of glycosylation type-I using protein chip technology.Proteomics. 2006; 6: 2295-2304Crossref PubMed Scopus (30) Google Scholar, 9Wada Y. Mass spectrometry for congenital disorders of glycosylation, CDG.J. Chromatogr. B Anal. Technol. Biomed. Life Sci. 2006; 838: 3-8Crossref PubMed Scopus (46) Google Scholar). Deficiencies of N-glycosylation site occupancy are classified as CDG type-I (CDG-I), whereas defects of N-glycan trimming and elongation are classified as CDG type-II (CDG-II) (10Aebi M. Helenius A. Schenk B. Barone R. Fiumara A. Berger E.G. Hennet T. Imbach T. Stutz A. Bjursell C. Uller A. Wahlstrom J.G. Briones P. Cardo E. Clayton P. Winchester B. Cormier-Dalre V. de Lonlay P. Cuer M. Dupre T. Seta N. de Koning T. Dorland L. de Loos F. Kupers L. Fabritz L. Hasilik M. Marquardt T. Niehues R. Freeze H. Grünewald S. Heykants L. Jaeken J. Matthijs G. Schollen E. Keir xKeir G. Kjaergaard S. Schwartz M. Skovby F. Klein A. Roussel P. Körner C. Lübke T. Thiel C. von Figura K. Koscielak J. Krasnewich D. Lehle L. Peters V. Raab M. Saether O. Schachter H. Van Schaftingen E. Verbert A. Vilaseca A. Wevers R. Yamashita K. Carbohydrate-deficient glycoprotein syndromes become congenital disorders of glycosylation: an updated nomenclature for CDG. First International Workshop on CDGS.Glycoconj. J. 1999; 16: 669-671Crossref PubMed Google Scholar). To date, 12 distinct genetic defects, which affect the N-glycosylation output to different levels, have been identified as the cause of CDG-I. Mutations in the phosphomannomutase-2 gene represent the most frequent form of CDG-I. The level of residual enzymatic activity associated to individual mutations in the phosphomannomutase-2 gene has been correlated with the severity of the clinical presentations (11Grünewald S. Schollen E. Van Schaftingen E. Jaeken J. Matthijs G. High residual activity of PMM2 in patients' fibroblasts: possible pitfall in the diagnosis of CDG-Ia (phosphomannomutase deficiency).Am. J. Hum. Genet. 2001; 68: 347-354Abstract Full Text Full Text PDF PubMed Scopus (92) Google Scholar). However, the exact degree of protein underglycosylation in CDG-I has not been determined yet, and it is unclear whether the degree of underglycosylation directly relates to the severity of the disease. Also the degree of N-glycosylation site occupancy found in healthy individuals has not been defined yet. To better understand the functional significance of N-glycosylation in humans, a precise measure for the glycosylation site occupancy in healthy and in CDG-related conditions is necessary. In proteomics, the most widely applied experiment for the determination of peptide sequences relies on product ion scanning mass spectrometry (MS/MS). Precursor ions of selected peptides are selected in the first mass analyzer and fragmented in a collision cell, and the generated product ion fragments are scanned in a second mass analyzer. Triple quadrupole mass spectrometers are particularly suitable for quantitative determinations of changes in serum protein levels. These instruments allow multiple reaction monitoring (MRM) experiments, a sequential selection of peptide specific precursor ions in the first mass analyzer together with a characteristic fragment ion in the second mass analyzer upon fragmentation in the collision cell. Transitions of precursor to fragment ions lead to signals in the detector and can be recorded as a function of time during chromatographic elution. Additionally the sequence of eluting peptides can be confirmed by the acquisition of corresponding product ion spectra. The selectivity of the MRM transitions together with short scan times for each transition allow for sensitive, quantitative analyses of complex samples (12Domon B. Aebersold R. Mass spectrometry and protein analysis.Science. 2006; 312: 212-217Crossref PubMed Scopus (1591) Google Scholar, 13Stahl-Zeng J. Lange V. Ossola R. Aebersold R. Domon B. High sensitivity detection of plasma proteins by multiple reaction monitoring of N-glycosites.Mol. Cell. Proteomics. 2007; 6: 1809-1817Abstract Full Text Full Text PDF PubMed Scopus (312) Google Scholar). In the present study, we describe a MRM LC-MS/MS approach to quantify the N-glycosylation site occupancy of the serum glycoproteins transferrin and α1-antitrypsin. We analyzed 28 serum samples and quantified glycosylation site occupancies in healthy and in disease-related samples. For the first time, we found a clear correlation in CDG-I between the degree of N-glycosylation site occupancy and the severity of the disease. Serum transferrin was saturated for 30 min with ferric citrate (0.4 mm) in the presence of sodium hydrogen carbonate (20 mm) and separated on a rehydrated Immobiline DryPlate gel (GE Healthcare) with a pH range of 4.0–7.0 using the PhastSystem (Amersham Biosciences). Transferrin was visualized by gel immunoprecipitation using polyclonal rabbit anti-human transferrin antibodies (DakoCytomation). The antibody solution was directly applied onto the gel and incubated for 1 h. To remove excess antibodies and non-bound proteins, the gel was washed overnight in 0.9% NaCl followed by fixation and staining with PhastGel Blue R (GE Healthcare) according to the manufacturer's protocol. Serum glycoproteins were purified with a custom-made HU-6 multiple affinity removal system column (4.6 × 50 mm, Agilent Technologies) targeting transferrin, IgG, and α1-antitrypsin. The purification was performed from 5 μl of human serum using the buffers and protocol recommended by the manufacturer. The glycoproteins were eluted into 1-ml fractions, and the elution was monitored by UV absorption at 280 nm. The proteins in the collected fractions were precipitated with TCA and pooled with reduction buffer (0.57 m Tris-HCl, pH 8.5, 50 mm DTT) to a final volume of 250 μl. The proteins were reduced at 80 °C for 5 min and cooled to room temperature, and 63 μl of 1 m iodoacetamide in water were added. The samples were incubated for 40 min at room temperature in the dark, and the alkylated proteins were desalted by TCA precipitation. The reduced and alkylated samples were redissolved in 50 μl of 20 mm ammonium bicarbonate, 10% ACN containing Asp-N and trypsin (0.5 μg each, Roche Applied Science) and digested for 16 h at 37 °C. The digests were heat-inactivated at 80 °C for 5 min. Then 0.5 μmol of EDTA, 100 μg of aprotinin, 8 pmol of the standard peptides corresponding to the transferrin N-glycosylation sites (TFP1NK, TFP1NKS, TFP1DK, TFP1DKS, TFP2NVT, and TFP2DVT; see Table I), and 2 pmol each of the standard peptides corresponding to the α1-antitrypsin N-glycosylation sites (ATP1N, ATP1D, ATP2N, ATP2D, ATP3N, and ATP3D; see Table I) were added. The samples were lyophilized for at least 16 h. The samples were redissolved in 20 mm ammonium bicarbonate in 95 atom % H218O (Sigma), and 1.5 units of peptide N-glycanase F (PNGaseF; Roche Applied Science) reconstituted with 95 atom % H218O were added. The samples were digested for 8 h at 37 °C, heat-inactivated at 80 °C for 5 min, and stored at −20 °C until further use.Table ICustom-synthesized standard peptidesDesignationSequenceQ3 ionATP1NQLAHQSNSTNIFFSPVSIATAFAMLSLGTKy9ATP1DQLAHQSDSTNIFFSPVSIATAFAMLSLGTKy12ATP2NDEILEGLNFNLTEIPEAQIHEGFQELLRy2+14ATP2DDEILEGLNFDLTEIPEAQIHEGFQELLRy2+14ATP3NYLGNATAIFFLPy1ATP3DYLGDATAIFFLPb11TFP1NKCGLVPVLAENYNKy9TFP1NKSCGLVPVLAENYNKSy10TFP1DKCGLVPVLAENYDKy9TFP1DKSCGLVPVLAENYDKSy10TFP2NVTQQQHLFGSNVTb5-NH3TFP2DVTQQQHLFGSDVTb9 Open table in a new tab An Applied Biosystems QTRAP 3200 LC-MS/MS system equipped with a NanoSpray II ion source and an Eksigent nanoLC-2D system was used. The sample reaction mixture (2–10 μl) was filtered with 0.22-μm centrifugal filter devices, adjusted to 20 μl with 0.1% formic acid in 2% ACN, and loaded on an LC Packings C18 PepMap 100 trap column (300-μm inner diameter × 5 mm, Dionex). The trap column was washed with 60 μl of loading buffer and switched in line to a ProteoPep2 column (75-μm inner diameter × 15 cm, C18, IntegraFrit, New Objective). A binary gradient at 250 nl/min was applied. Buffer A was 0.1% formic acid, and buffer B was 80% ACN containing 0.1% formic acid. Buffer B was held at 14% for 3 min, increased to 54% over 102 min and to 100% over 8 min, held at 100% for 12 min, and decreased to 14% over 12 min, and the column was re-equilibrated at 14% buffer B for 16 min. MRM transitions and product ion scans were acquired sequentially, and the LC gradient was divided into three acquisition periods. The first 23 min, MRM transitions and product ion scans corresponding to the second, C-terminal transferrin N-glycosylation site were recorded (TFP2NVT, TFP2DVT, and corresponding sample peptides). During the second period, data corresponding to the first N-terminal transferrin N-glycosylation site were acquired for 25 min (TFP1NK, TFP1NKS, TFP1DK, TFP1DKS, and corresponding sample peptides). During the third period, data corresponding to the α1-antitrypsin N-glycosylation sites were acquired until completion of the LC run (ATP1N, ATP1D, ATP2N, ATP2D, ATP3N, ATP3D, and corresponding sample peptides). Fragment ions and fragmentation settings for the sample-derived peptides were set according to the MRM transitions established for the standard peptides (Table I). The average counts per second of the sample peptide peaks were divided by the average counts per second of the standard peptide and multiplied with the amount of standard peptides added to the sample preparation. The sum of the amounts of previously glycosylated and unglycosylated peptides was set to 100%. The percentage of site occupancy was calculated as the part of previously glycosylated, i.e. PNGaseF-sensitive, peptides. The serum glycoprotein transferrin is the most widely used biomarker for the clinical diagnosis of CDG-I and -II. Normal serum transferrin typically contains two sialylated, biantennary complex type N-glycans attached to the glycosylation sites at Asn-413 and Asn-611 (14Fu D. van Halbeek H. N-Glycosylation site mapping of human serotransferrin by serial lectin affinity chromatography, fast atom bombardment-mass spectrometry, and 1H nuclear magnetic resonance spectroscopy.Anal. Biochem. 1992; 206: 53-63Crossref PubMed Scopus (68) Google Scholar). Serum samples from CDG-I, CDG-II, hereditary fructose intolerance (HFI), and healthy individuals were analyzed by transferrin isoelectric focusing gel electrophoresis, and the typical deviations of the band pattering of CDG sera from the healthy controls could be observed (Fig. 1). A major band corresponding to tetrasialylated, fully glycosylated transferrin was observed in control samples. In CDG sera, undersialylation was detected with more intense bands corresponding to tri-, di-, mono-, and asialylated transferrin. In CDG-II, this pattern can be explained by altered trimming and elongation of transferrin N-glycan structures, which leads to decreased transfer of terminal sialic acid. In CDG-I, decreased N-glycosylation site occupancy results in the undersialylation as observed in the transferrin isoelectric focusing gel (Fig. 1). In HFI the metabolite fructose 1-phosphate accumulates in the liver, which leads to inhibitory effects in early steps of N-glycosylation biosynthesis. Therefore, HFI can be considered as a secondary CDG syndrome, and CDG-I typical patterns are observed in transferrin isoelectric focusing gel electrophoresis (15Jaeken J. Pirard M. Adamowicz M. Pronicka E. van Schaftingen E. Inhibition of phosphomannose isomerase by fructose 1-phosphate: an explanation for defective N-glycosylation in hereditary fructose intolerance.Pediatr. Res. 1996; 40: 764-766Crossref PubMed Scopus (70) Google Scholar). Polypeptide polymorphisms of transferrin also lead to altered gel patterning, which can be revealed by sialidase treatment prior to isoelectric focusing electrophoresis. To precisely measure the N-glycosylation site occupancy in healthy as well as in pathological conditions, sera from 17 CDG patients, one alcohol abuse patient, and one HFI patient and nine sera from healthy individuals were selected for analysis. The serum of an alcohol abuse patient was included as a positive control. Alcohol abuse leads to reduced N-glycosylation site occupancy of transferrin, thus mimicking CDG-I (16Peter J. Unverzagt C. Engel W.D. Renauer D. Seidel C. Hösel W. Identification of carbohydrate deficient transferrin forms by MALDI-TOF mass spectrometry and lectin ELISA.Biochim. Biophys. Acta. 1998; 1380: 93-101Crossref PubMed Scopus (52) Google Scholar). The samples were subjected to glycoprotein purification using a custom-made multiple affinity column directed against human transferrin, α1-antitrypsin, and IgG. The reduced and alkylated proteins were digested with a combination of trypsin and Asp-N, and isotopically labeled standard peptides were added (17Gerber S.A. Rush J. Stemman O. Kirschner M.W. Gygi S.P. Absolute quantification of proteins and phosphoproteins from cell lysates by tandem MS.Proc. Natl. Acad. Sci. U. S. A. 2003; 100: 6940-6945Crossref PubMed Scopus (1536) Google Scholar) (Table I). An isotope label was introduced into N-glycosylated sample peptides with PNGaseF in H218O buffer. PNGaseF deglycosylates Asn-linked glycopeptides and converts the corresponding Asn to Asp. In the presence of H218O buffer 18O is incorporated into the carboxyl group of the newly formed Asp (18Küster B. Mann M. 18O-Labeling of N-glycosylation sites to improve the identification of gel-separated glycoproteins using peptide mass mapping and database searching.Anal. Chem. 1999; 71: 1431-1440Crossref PubMed Scopus (139) Google Scholar, 19Kaji H. Saito H. Yamauchi Y. Shinkawa T. Taoka M. Hirabayashi J. Kasai K. Takahashi N. Isobe T. Lectin affinity capture, isotope-coded tagging and mass spectrometry to identify N-linked glycoproteins.Nat. Biotechnol. 2003; 21: 667-672Crossref PubMed Scopus (564) Google Scholar). The added standard peptides were synthesized according to the transferrin and α1-antitrypsin N-glycosylation site sequences with either Asn or Asp introduced into the N-glycosylation sequon. The precursor and fragment ion pairs used for each MRM transition were selected empirically based on maximal signal intensities and signal specificities as judged from the chromatography of extracted ion pairs (Table II). For the sample peptides, the corresponding precursor ion and fragment ion types as selected for the standard peptides were used. The inclusion of the standard peptides facilitated the identification and quantitation of N-glycosylation site occupancy with LC-MS/MS. The peptide mixture was separated on a reverse-phase nano-LC column and introduced to the mass spectrometer via a nanoelectrospray ion source. The chromatography was monitored from the total ion chromatogram (Fig. 2). MRM transitions for each target peptide and the fragment ion spectra of the corresponding sample peptides were recorded sequentially. The acquisition was grouped into three time periods according to the elution positions of the second and the first transferrin target peptides and the elution positions of the α1-antitrypsin target peptides, respectively (Fig. 2).Table IITriple quadrupole settings for the target peptidesDesignationQ1Q3DwellDPEPCEPCECXPmsVVVVVATP1N1064.56977.56200359.555447ATP1D1066.891220.68200359.5605049ATP2N1091.89836.94200467.555456ATP2D1093.56833.93200439.060456ATP3N667.37116.0750217.0372750ATP3D1337.721222.6620859.5404550TFP1NK742.391054.56100409.0404145TFP1NKS785.901141.60100419.5434448TFP1DK744.881059.56100419.0404145TFP1DKS788.401146.59100419.5434448TFP2NVT633.32625.29150409.0353743TFP2DVT635.321051.50150409.0353744 Open table in a new tab The MRM transitions together with the co-elutions of the sample peptides with the labeled standard peptides served as primary criteria for peptide identification. In addition, the enhanced product ion spectra of the sample peptides were acquired, and the peptide sequences were deduced from the fragment spectra by comparison with previously recorded fragment ion spectra of the corresponding standard peptides (Fig. 3). The average signal intensities of the sample and standard peptides were used to calculate the site occupancies of each glycosylation site (Fig. 4). Overall the sensitivity for detection of each peptide was dependent on their ionization characteristics and was partially accounted for in adjusted dwell times of the respective MRM transition (between 20 and 200 ms; Table II). The results are summarized in Table III. The site occupancy for all detected N-glycosylation sites in the healthy control samples ranged between 98 and 100%. The transferrin glycosylation sites are equally occupied under normal physiological conditions. In α1-antitrypsin, the first and the third N-glycosylation sites were always occupied to 100%, whereas the second N-glycosylation site appeared slightly less stringently occupied with an average of 99 ± 1.4% under normal physiological conditions.Table IIIN-Glycosylation site occupancy of transferrin and α1-antitrypsinGroupGene defectOMIM entrynTFP1DTFP2DATP1DATP2DATP3DAverageS.D.AverageS.D.AverageS.D.AverageS.D.AverageS.D.Healthy controls3–14991.1990.71000.0991.41000.2CDG-IaPMM26017853963.394100971000.0CDG-IaPMM26017853–6893.3750.91000.01000.0981.2CDG-IaPMM26017853–6704.2593.710094962.0CDG-IaPMM26017853–57210.8416.29410.8655.4944.0CDG-IaPMM26017853–4885.2564.51000.0922.3982.2CDG-IaPMM26017853–4724.2412.19882903.8CDG-IaPMM26017853953.38910099990.6CDG-IaPMM26017853–4792.0672.4100100952.3CDG-IaPMM26017854912.9762.394980.5CDG-IbMPI1545503–4932.891100970.4990.3CDG-IcALG66045664901.8622.5100100932.2CDG-IcALG66045663–4834.4577.89979892.3CDG-IeDPM16035033912.88510099990.2CDG-IfMPDU16040414–5921.1851.792980.7CDG-IxUnknown3869.3799895982.0CDG-IIaMGAT26026164–5990.2971.71000.2990.21000.3CDG-IIxUnknown3–41000.0980.3990.11000.1HFIALDOB2296003–4960.9912.51000.1981.8990.2Alcohol abuseAcquired4–5961.3911.41000.0970.9990.2 Open table in a new tab The level of site occupancy in CDG-I samples was decreased to variable extents depending on the of N-glycosylation sites analyzed. In transferrin, the first N-glycosylation site was occupied from 70 to 96%, whereas the occupancy of the second N-glycosylation site was reduced down to 41% in CDG-I samples (Table III). Generally the second transferrin glycosylation site was less occupied than the first site, indicating a preferred glycosylation of the first site in this glycoprotein. The underglycosylation level of α1-antitrypsin was less pronounced than that of transferrin in CDG-I. In α1-antitrypsin, the second N-glycosylation site showed the strongest reduction down to 65% occupancy, whereas the first and third sites were hardly altered, showing occupancies between 94–100% and 89–100%, respectively (Table III). The site occupancy of the first α1-antitrypsin N-glycosylation site could not be determined in every sample possibly due to low ionization efficiencies of the triply charged peptide in the ion source. Additionally this peptide contains N-terminal glutamine, which can circularize to pyroglutamic acid, and a methionine residue, which can oxidize to methionine sulfoxide, thus evading detection via the MRM transitions applied. The occupancy of transferrin N-glycosylation sites in HFI and in alcohol abuse samples were not as pronounced as in CDG-I samples, yet like in CDG-I, the second site was less glycosylated than the first site. The N-glycosylation of α1-antitrypsin was hardly reduced in the HFI and alcohol abuse samples investigated (Table III). As expected, N-glycosylation site occupancy was unchanged in CDG-II samples, which are characterized by defects of glycan trimming and elongation. The N-glycosylation site occupancies of transferrin and α1-antitrypsin were determined with high precision as illustrated by the standard deviations obtained. The maximum deviation reached 10.8% in one of the 28 serum samples analyzed. The standard deviations calculated for the other samples ranged from 0.0 to 7.8% per glycosylation site. The highest accuracy was obtained in the CDG-II, HFI, alcohol abuse, and healthy control samples. In these samples the maximum standard deviation was calculated to be 2.5% (Table III). The reasons for the higher deviations in the measurements of the N-glycosylation site occupancies of the CDG-I sera are not clear. Possibly lower serum protein levels as observed in CDG-I patients contribute to this observation. From all N-glycosylation sites investigated the second transferrin site appeared to respond most sensitively to defects of glycosylation. Clinical information was available for 11 CDG-I cases investigated here, and we grouped these cases into three groups of patients displaying severe, moderate, and mild clinical manifestations (11Grünewald S. Schollen E. Van Schaftingen E. Jaeken J. Matthijs G. High residual activity of PMM2 in patients' fibroblasts: possible pitfall in the diagnosis of CDG-Ia (phosphomannomutase deficiency).Am. J. Hum. Genet. 2001; 68: 347-354Abstract Full Text Full Text PDF PubMed Scopus (92) Google Scholar). A clear correlation between the occupancy of the second transferrin N-glycosylation site and the severity of CDG-I condition could be observed (Fig. 5). The lowest level of site occupancy detected in CDG-I samples never reached less than 40%. A minor reduction of N-glycosylation site occupancy to about 85% of normal levels as seen in transferrin was already compatible with a disease state, indicating that N-glycosylation site occupancy is stringently required for proper physiological function of glycoproteins. Routinely isoelectric focusing gel electrophoresis of transferrin is applied for the diagnosis of CDG. Individual transferrin glycoforms are separated according to charge as mediated by terminal sialic acid residues attached to the complex type N-glycans. In combination with Western blotting analysis of transferrin, CDG-I and -II can be easily discriminated. Potential transferrin peptide polymorphisms can be revealed by sialidase treatment prior to IEF. Additional molecular weight determinations of transferrin by ESI-MS or SELDI-MS can further corroborate the diagnosis (7Lacey J.M. Bergen H.R. Magera M.J. Naylor S. O'Brien J.F. Rapid determination of transferrin isoforms by immunoaffinity liquid chromatography and electrospray mass spectrometry.Clin. Chem. 2001; 47: 513-518Crossref PubMed Scopus (169) Google Scholar, 8Mills K. Mills P. Jackson M. Worthington V. Beesley C. Mann A. Clayton P. Grünewald S. Keir G. Young L. Langridge J. Mian N. Winchester B. Diagnosis of congenital disorders of glycosylation type-I using protein chip technology.Proteomics. 2006; 6: 2295-2304Crossref PubMed Scopus (30) Google Scholar). However, these indirect approaches do not discriminate between individual N-glycosylation sites, and quantitative information relies on densitometric measurements of glycoprotein bands in IEF or SDS-PAGE. In this study we developed a method for directly determining the presence and absence of N-glycosylation site occupancy, and fine deviations from normal N-glycosylation site occupancy could be determined. In a small survey we analyzed sera from healthy individuals and authentic serum samples from patients suffering from disease conditions associated with reduced N-glycosylation site occupancy. We used MRM LC-MS/MS, which allows for sensitive and specific detection of target peptides by taking advantage of the dynamic quantitative range capabilities inherent of quadrupole mass analyzers (13Stahl-Zeng J. Lange V. Ossola R. Aebersold R. Domon B. High sensitivity detection of plasma proteins by multiple reaction monitoring of N-glycosites.Mol. Cell. Proteomics. 2007; 6: 1809-1817Abstract Full Text Full Text PDF PubMed Scopus (312) Google Scholar, 20Hopfgartner G. Varesio E. Tschappat V. Grivet C. Bourgogne E. Leuthold L.A. Triple quadrupole linear ion trap mass spectrometer for the analysis of small molecules and macromolecules.J. Mass Spectrom. 2004; 39: 845-855Crossref PubMed Scopus (224) Google Scholar). Target peptides can be detected and quantified reliably even in complex mixtures. Additional confidence in peptide sequence assignment is provided by acquisition of enhanced product ion spectra in the ion trapping mode of the quadrupole ion trap mass spectrometry (QTRAP) system used. For the first time using this approach, we quantitatively mapped N-linked glycosylation sites of selected serum glycoproteins. A striking feature of the glycoproteins analyzed, transferrin and α1-antitrypsin, is the narrow range of N-glycosylation site occupancy under physiological conditions. To warrant normal physiological function N-glycosylation site occupancy of 99 and 100% is preserved. Symptomatically mild CDG cases show only a slight reduction in N-glycosylation site occupancy, indicating a small threshold from healthy to pathological conditions. Our data show a tendency of preferential glycosylation of the first N-linked glycosylation sites in transferrin and α1-antitrypsin under conditions of limited dolichol-linked oligosaccharide substrate availability. In addition, transferrin was more prone to underglycosylation than was α1-antitrypsin. The reasons for this observation are not clear. Transferrin is about 4 times more abundant in human serum and almost 2-fold larger than α1-antitrypsin and therefore might require different subsets of the translation and glycosylation machinery in the cell (21Chavan M. Lennarz W. The molecular basis of coupling of translocation and N-glycosylation.Trends Biochem. Sci. 2006; 31: 17-20Abstract Full Text Full Text PDF PubMed Scopus (58) Google Scholar). Of all N-glycosylation sites analyzed the second transferrin site showed the greatest dynamics in site occupancy in pathological conditions. This feature could be exploited for the development of an assay targeting the second transferrin site and might be applied in a postnatal screen upon suspicion of CDG-I. Information about the glycosylation site occupancy at this stage would be beneficial in providing estimates about the developmental prospects of the affected individuals. With the clinical information available, we classified CDG-I samples according to previously suggested criteria for mild clinical expressions of this disease (11Grünewald S. Schollen E. Van Schaftingen E. Jaeken J. Matthijs G. High residual activity of PMM2 in patients' fibroblasts: possible pitfall in the diagnosis of CDG-Ia (phosphomannomutase deficiency).Am. J. Hum. Genet. 2001; 68: 347-354Abstract Full Text Full Text PDF PubMed Scopus (92) Google Scholar). The remaining samples were grouped into moderate and severe clinical expressions. Most samples analyzed were derived from patients with a moderate pathology, comprising the classical CDG-I symptoms like failure to thrive, seizures, developmental delay, hypotonia, ataxia, or mental regression. The severe cases comprised conditions such as death in early childhood in addition to the moderate CDG-I pathology. When plotted against the N-glycosylation site occupancy, we found a clear correlation between the presence of glycosylation and clinical expression of the disease, confirming the causative coherency of the absence of N-glycosylation and clinical expression in CDG-I. These findings underline the significance of N-glycosylation in health and development of the human organism. Overall our data validate transferrin as a sensitive marker protein for CDG-I and alcohol abuse conditions. The method developed in this study will be useful for unambiguously mapping the sensitivity of proteins to defects of glycosylation. Furthermore this method can be extended by including additional serum glycoproteins, which will contribute to a better understanding of the significance of N-glycosylation in human physiology. We thank Dr. Dirk Lefeber, University of Nijmegen, for providing serum samples. We gratefully acknowledge the support from the Functional Genomics Center Zurich (FGCZ) and the Zurich Glycomics Initiative (GlycoInit).
Referência(s)