Glycobiology vol. 16 no. 4 pp. 281293, 2006 doi:10.

1093/glycob/cwj067 Advance Access publication on December 8, 2005

Mass spectrometric approach for screening modifications of total serum N-glycome in human diseases: application to cirrhosis

Willy Morelle2, Christophe Flahaut3, Jean-Claude Michalski2, Alexandre Louvet4, Philippe Mathurin4, and Andr Klein1,2,5
2 Unit Mixte de Recherche CNRS/USTL 8576, Glycobiologie Structurale et Fonctionnelle, IFR 118, Universit des Sciences et Technologies de Lille 1, 59655 Villeneuve dAscq, France; 3Laboratoire de physiopathologie de la barrire hmato-encphalique, E.A. 2465, IMPRT-IFR 114 Universit dArtois, rue Souvraz, SP18, 62307 Lens Cedex, France; 4Services dHpato-Gastroentrologie, Hpital Huriez, CHRU Lille, France; and 5 Laboratoire de Biochimie et de Biologie Molculaire, UAM de glycopathologies, Hpital Calmette, CHRU Lille, Bld du Professeur Jules Leclerc, Lille 59037 Cedex, France

Introduction Changes in glycosylation associated with human diseases represent a rapidly growing field. Part of these diseases, directly related to a congenital defect, is termed congenital disorders of glycosylation (CDG). Since 1980, when Jaeken et al. (1980) described a new neurologic disorder, 17 different defective proteins have been shown to be involved in an N-glycan biosynthesis defect. These defects are classified into CDG Type I if there is a decrease in the biosynthesis or the transfer of the lipid-linked oligosaccharide precursor, dolichol-linked Glc3Man9GlcNAc2, or classified into CDG Type II when the defect occurs after the transfer to the glycoprotein (Grnewald et al., 2002). Other congenital defects can be included in this group of diseases, as defects of N-glycosylation such as I-cell disease, congenital dyserythropoietic anemia type II (hereditary erythroblastic multinuclearity with positive acidified-serum lysis test [HEMPAS]), congenital galactosemia, and congenital intolerance to fructose. Infants suffering from CDG usually have a severe phenotype with mental retardation, failure to thrive, gastrointestinal problems, and dysmorphia. Clinical presentations are extremely broad, with or without neurological symptoms. Therefore, CDG should be suspected in any multi-systemic disorders. Isoelectrofocusing of serum transferrin is the first step for the diagnostic of N-glycosylation defects, and this test is the most efficient diagnosis method available for the identification of these diseases. This assay is based on the number of sialic acid residues of the different transferrin glycoforms and reflects the absence of terminal sialic acids, variations in N-glycan branching or the complete absence of N-glycan on either one glycosylation transferrin site or both of these sites. The availability of this test in most clinical laboratories has made it possible to diagnose more than 500 patients suffering from CDG, even if the screening of this group of diseases is not generalized. It has also provided the opportunity of pointing out Nglycosylation variations in other congenital diseases such as galactosemia (Charlwood et al., 1998), hereditary fructose intolerance (Adamowicz and Pronicka, 1996), and cystic fibrosis (Larsson et al., 1998). This test is also used to demonstrate glycosylation alterations in serum glycoproteins in case of chronic alcohol intake. This is the only iatrogenic etiology resulting in acquired modification of glycosylation, and the alteration mechanism is poorly understood (Flahaut et al., 2003). Acquired glycosylation changes in human diseases also represent an extremely large field of study (Varki, 1999). N-glycosylation modifications in human serum glycoproteins have mainly been described in liver, inflammatory diseases, and cancers. In the group of liver diseases, many

Received on May 6, 2005; revised on December 5, 2005; accepted on December 6, 2005

Congenital and acquired modifications of glycosylation in diseases are a rapidly growing field that demonstrates the importance of glycosylation in human biology. Unfortunately, in clinical biochemistry, very few tests are available to explore oligosaccharide metabolism on a large scale. Such an assay needs to be of high throughput, rapid, and preferentially noninvasive. In the present study, we describe a method to analyze qualitative variations of N-glycosylation of human serum proteins. The method is based on direct release of N-linked oligosaccharides from patient serum samples, a single-step purification, and a matrix-assisted laser desorption ionization time of flight mass spectrometric analysis. A complementary structural study of the released oligosaccharides was achieved by enzymatic digestions, linkage analysis, and electrospray ionization ion trap mass spectrometry (ESI-IT-MS) of the permethylated N-glycome. A total of 26 oligosaccharide structures were individualized, their presence in human serum being the result of the combination of the biosynthesis and catabolic pathways. Application of the protocol to the serum of patients with cirrhosis demonstrates the ability of this assay to identify acquired modifications of glycosylation. Furthermore, we have analyzed the N-glycans and showed the increase in bisecting N-acetylglucosamine residue, core fucosylation, and the presence of an important population of neutral oligosaccharides. The study of total serum N-glycome modifications is a preliminary for the discovery of new noninvasive diagnostic or prognostic biomarkers resulting from the variations of the N-glycan metabolism during diseases. Key words: cirrhosis/glycosylation disorders/mass spectrometry/N-glycan/N-glycome

To whom correspondence should be addressed; e-mail: a-klein@chru-lille.fr

The Author 2005. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oxfordjournals.org 281

W. Morelle et al.

modifications in N-glycosylation have been demonstrated in hepatic cirrhosis (Anderson et al., 2002; Ryden et al., 2002), in hepatocellular carcinoma (Yamashita et al., 1989), and in case of chronic alcohol intake (Flahaut et al., 2003). Modification of branching, fucosylation, and sialylation of N-glycans on various serum glycoproteins, 1-acid glycoprotein (De Graaf et al., 1993), 1-antichymotrypsin (Hachulla et al., 1988), haptoglobin, and 1-antitrypsin (Goodarzi and Turner, 1998) have been characterized in chronic or acute inflammation. Modification of the immunoglobulin glycosylation has been demonstrated in rheumatoid arthritis (Parekh et al., 1985) and in other autoimmune diseases (Basset et al., 2000). However, none of these glycosylation modifications are used for diagnosis or prognosis, and the major reason is the non-availability of a simple, reproducible, sensitive, and high throughput assay which could be used in any clinical chemistry laboratory. Total serum N-glycome corresponds to the global qualitative study of all glycoprotein N-linked glycans present in the human serum and gives a general overview of the glycosylation of all serum glycoproteins that are essentially synthesized by the liver and by the immunoglobulin-secreting plasma cells. Any alterations in the liver and/or B-lymphocyte physiology might influence glycosylation mechanisms and show relative N-glycan differences, which would represent diagnosis or prognosis markers of diseases. Recently, Callewaert et al. have developed a protocol for profiling N-glycans on a DNA sequencer combined with a highthroughput multi-step method, capable of isolating N-linked oligosaccharides (Papac et al., 1998; Callewaert et al., 2001). They have validated their protocol with the diagnoses of CDG (Callewaert et al., 2003) and established the importance of total serum N-glycome in monitoring the progression of chronic liver diseases (Callewaert et al., 2004). The discovery of new biomarkers is crucial for the diagnosis or prognosis of diseases. Total serum N-glycome is a noninvasive test, in that blood sample can be obtained without inducing major trauma to the patient. However, such an assay needs to be of high throughput, sensitive, relatively simple, reproducible, and easy to interpret, so that it may be used in a clinical chemistry laboratory. In this study, we describe an extremely simple method for the acquisition of total serum N-glycomedesialylated or noton the basis of mass spectrometric approach of enzymatically released N-glycans purified by a single-step solidphase extraction. As a first application, we describe the use of this technique in cirrhosis. In the second part of this study, we elucidate the primary structure of N-glycans present in the total mix of serum glycoproteins in healthy donors and patients with cirrhosis, and the glycosylation changes observed are discussed.

(PGC) columns. Total serum N-glycome was analyzed by matrix-assisted laser desorption ionization time of flight (MALDI-TOF) mass spectrometry (MS) before and after chemical desialylation. The oligosaccharides were also characterized by MALDI-MS before and after exoglycosidase digestion, by linkage analysis and by electrospray ionization ion trap mass spectrometry (ESI-IT-MS). MALDI-TOF analysis of N-linked glycans present in total human serum The released N-glycans were either chemically desialylated or methylesterified and analyzed by MALDI-TOF MS. Methylesterification of the carboxyl group is used to avoid the loss of terminal sialic acids during MALDI-TOF analysis and therefore allows simultaneous analysis of neutral and sialylated oligosaccharides in the positive mode (Powell and Harvey, 1996). As shown in Figure 1A, the MALDITOF MS spectrum of the N-glycans from an healthy subject reveals the presence of major ions at m/z 2274 corresponding to an oligosaccharide with a chemical composition of NeuAcMe2Hex5HexNAc4. The monosialylated oligosaccharide is observed at m/z 1969, and the corresponding fucosylated oligosaccharides are observed at m/z 2420 and 2115. The ions at m/z 2944 correspond to a trisialylated oligosaccharide with a chemical composition of NeuAcMe3Hex6HexNAc5. The chemical compositions of each ion in the spectrum are summarized in Table 1. The reduction of sample heterogeneity by chemical desialylation of the N-glycans confirms this oligosaccharide distribution, and the major ions are observed at m/z 1664 (Hex5HexNAc4), 1810 (Hex5HexNAc4DeoxyHex1), and 2029 (Hex6HexNAc5) (Figure 1B). The MALDI-MS spectrum of N-glycans released from human serum of a patient with a decompensated cirrhosis is shown in Figure 1C. The sialic acids are stabilized by methylesterification of the carboxyl group. The spectrum presents three major differences with the MALDI-MS spectrum of the N-glycans of the healthy subject: (1) neutral oligosaccharide ions at m/z 1486, 1543, 1648, 1689, and 1851 represent 40% of the relative intensity of all the ions; (2) ions at m/z 1810, 2115, and 2420 corresponding to fucosylated oligosaccharides are more intense; and (3) sialylated oligosaccharides at m/z 2318 and 2623 that were present as trace amounts in the healthy subject N-glycome are major components. The MALDI-MS spectrum of the desialylated oligosaccharides (Figure 1D) confirms the previous observations, with an increase of the relative intensities of ions at m/z 1486, 1502, 1648, 1705, 1810, 1851, 1867, and 2013. Variations of total serum N-glycome of two other patients The comparison of the MALDI-TOF MS spectra of the methylesterified and desialylated total serum N-glycome of patient Y (Figure 1E and F) and of patient X (Figure 1C and D) both with a decompensated cirrhosis and with acute alcoholic hepatitis (Table 2) indicates differences in the relative intensities of two groups of ions: (1) ions at m/z 1486 and 1648 are less intense and (2) ions with an m/z superior to 2900 are more intense in patient Y. These ions correspond to three groups of three oligosaccharides: trisialotriantennary

Results After the denaturation of the proteins/glycoproteins present in human serum, the N-glycans were enzymatically released with peptide N-glycosidase F (PNGase F). The oligosaccharides were then purified in a single-step procedure by solid-phase extraction on porous graphitized carbon 282

Total serum N-glycome in human diseases

Fig. 1. MALDI-TOF MS analysis of total human serum N-glycome. Spectra obtained from (A) methylesterified N-linked oligosaccharides from normal subject; (B) desialylated N-linked oligosaccharides from normal subject; (C) methylesterified N-linked oligosaccharides from cirrhotic patient X; (D) desialylated N-linked oligosaccharides from cirrhotic patient X; (E) methylesterified N-linked oligosaccharides from cirrhotic patient Y; (F) desialylated N-linked oligosaccharides from cirrhotic patient Y; (G) methylesterified N-linked oligosaccharides from patient Z suffering from chronic hepatitis C; and (H) desialylated N-linked oligosaccharides from patient Z.

oligosaccharides at m/z 2944, 3090, and 3236; trisialotetraantennary oligosaccharides at m/z 3309, 3455, and 3601; and tetrasialotetraantennary oligosaccharides at m/z 3616, 3760, and 3906, the ions in each group corresponding to the non-, mono-, and difucosylated compounds respectively. MALDI-TOF MS spectra of the methylesterified total serum N-glycome of patient Z (Figure 1G and H) with chronic hepatitis C and with fibrosis differ from the control serum N-glycome by an increase of ions at m/z 1486, 1648, 2318, and 2623 (Figure 1A and G) confirmed by the increase of ions at m/z 2013 in the desialylated sample spectrum (Figure 1B and H). Methylation analysis and linkage analysis of total human serum N-glycans The N-glycans of human serum glycoproteins were permethylated after chemical desialylation or not and analyzed by MALDI-TOF MS. The distribution of the oligosaccharides was the same as the one observed previously and confirmed the important proportion of neutral oligosaccharides in the

N-glycome of patients with decompensated cirrhosis (ions at m/z 1836, 1866, 1907, 2040, 2081, 2111, and 2285) together with an increase of fucosylated oligosaccharides as attested by the ions at m/z 2244, 2605, and 2966 and the presence of oligosaccharide ions at m/z 2315, 2489, 2850, and 3212 almost absent in the normal N-glycome (Figure 2). The permethylated N-glycans and their desialylated counterparts were then hydrolyzed and analyzed using gas chromatography mass spectrometry (GC-MS) as their partially methylated alditol acetate derivatives. The data are presented in Table 3. Key features of the linkage analysis of the normal serum N-glycome are as follows. (1) Comparison of the relative abundance of the 2-linked mannose, representative of the biantennary complex glycans, with 2,4-linked mannose and 2,6-linked mannose suggests that minor tri- and/ or tetraantennary structures are present; (2) after desialylation, 3- and 6-linked galactose disappear and there is a concomitant increase of terminal galactose, indicating that sialic acid residues were mainly attached to the 6-position of galactose residues; (3) most of the N-acetylglucosamines are 4-linked, and a minor amount of terminal N-acetylglucosamine is 283

W. Morelle et al.

Table I. Structure of N-linked glycans from total human serum N-glycome m/z A 1486 1502 1543 1648 1664 B 1836 1866 1907 2040 2070 C 929 944 965 1032 1047 Chemical composition Hex3HexNAc4DeoxyHex Hex4HexNAc4 Hex3HexNAc5 Hex4HexNAc5DeoxyHex Hex5HexNAc4 Structure

1689

2081

1052

Hex3HexNAc5DeoxyHex Hex4HexNAc5 Hex5HexNAc4DeoxyHex Hex4HexNAc5DeoxyHex Hex5HexNAc5 NeuAcMeHex5HexNAc4

1705 1810 1851 1867 1969 2013 2029 2115 2175 2274 2318

2111 2244 2285 2315 2431 2489 2519 2605 2693 2792 2850

1067 1134 1154 1169

1256 1271

Hex5HexNAc5DeoxyHex Hex6HexNAc5 NeuAcMeHex5HexNAc4DeoxyHex

1358

Hex6HexNAc5DeoxyHex NeuAcMe2Hex5HexNAc4 NeuAcMeHex5HexNAc5DeoxyHex Hex7HexNAc6

2394 2420 2623 2966 3212

NeuAc2MeHex5HexNAc4DeoxyHex NeuAc2MeHex5HexNAc5DeoxyHex NeuAcMe2Hex6HexNAc5 NeuAc2MeHex6HexNAc5DeoxyHex NeuAc3MeHex6HexNAc5

2639

3241

2785 2944

3416 3603

284

Total serum N-glycome in human diseases

Table I. continued m/z A B C Chemical composition Structure

3090

3777

NeuAc3MeHex6HexNAc5DeoxyHex

3309 3614

4052 4413

NeuAc3MeHex7HexNAc6 NeuAc4MeHex7HexNAc6

Symbols for the structural formulae are defined as follows: , N-acetyl-D-glucosamine; , galactose; , mannose; , fucose; , NeuAc. A, m/z values refer to [M+Na]+ ions of methylesterified N-linked glycan; B, m/z values refer to [M+Na]+ ions of permethylated N-linked glycan; C, m/z values refer to [M+2Na]2+ ions of permethylated N-linked glycan. Table II. Routine chemistry and clinical data of the patients X, Y, and Z X Aspartate aminotransferase (U/L) (n =1040) Alanine aminotransferase (U/L) (n = 1040) Albumin (g/dL) (n = 3.85.1) Immunoglobulin G (g/dL) (n = 0.691.4) Total bilirubin (mg/dL) (n < 1) Fibrosis Inflammation 64 36 2.14 2.51 13.4 4 Severe Y 73 30 2.92 1.91 51.3 3 Moderate Z 22 16 3.94 1.56 0.4 3 Moderate

present; (4) 4,6-linked N-acetylglucosamine supports the presence of core 1,6-fucosylation, and traces of 3,4-linked N-acetylglucosamine correspond probably to low amounts of 3-fucosylation; and (5) 3,4,6-linked mannose is present as traces and indicates the rarity of bisected complex structures. Similar linkages are found in the serum N-glycome of patients with cirrhosis, but differences in the relative intensities are found: (1) terminal mannose, terminal galactose, and terminal N-acetylglucosamine have more elevated relative intensities, indicating the presence of neutral or incompletely sialylated complex-type oligosaccharides; (2) 3,4,6-linked mannose is present in considerable amount and suggests the existence of bisected complex-type oligosaccharides, and also contributes to the increase of terminal N-acetylglucosamine; (3) the intensity of terminal fucose is increased in the same proportion as 4,6-linked N-acetylglucosamine, indicating the presence of an increased relative amount of oligosaccharides with an 6-fucosylated core. Exoglycosidase digestions To define the anomeric configurations as well as to confirm tentative sequences, desialylated N-glycans released by PNGase F from normal human serum and from patients suffering from cirrhosis were subjected to digestion with -galactosidase alone or in combination with -N-acetylhexosaminidase and -fucosidase directly on the plate according to Mechref and Novotny (1998). These

exoglycosidase digestions have been performed during 6 or 8 h. Only the spectra obtained after 8-h digestion are shown in Figure 3. After -galactosidase treatment, the MALDI-MS spectra of the N-glycans from a normal human serum (Figure 3A) and from a patient suffering from cirrhosis (Figure 3B) were characterized by the presence of seven molecular ions at m/z 1340 (Hex3HexNAc4), 1486 (Fuc1Hex3HexNAc4), 1543 (Hex3HexNAc5), 1689 (Fuc1Hex3HexNAc5), 1745 (Hex3HexNAc6), 1851 (Fuc1Hex4HexNAc5), and 1892 (Fuc1Hex3HexNAc6). Therefore, the N-glycans were efficiently degalactosylated, indicating that the Gal residues were in normal -linkages. The MALDI-MS spectra (Figure 3A and B) present two major differences. The major molecular ion in the MALDI-MS spectrum of the Nglycans of the healthy subject and of the subject suffering from hepatic fibrosis was observed, respectively, at m/z 1340 (Hex3HexNAc4) and 1486 (Fuc1Hex3HexNAc4). These data confirm an increased relative amount of oligosaccharides with an 6-fucosylated core in the serum Nglycome of the patient suffering from cirrhosis. In addition, after 6-h digestions, ions at m/z 1689 (Fuc1Hex3HexNAc5) and 1851 (Fuc1Hex4HexNAc5) are more intense in the serum N-glycome of the patient suffering from cirrhosis (data not shown). Since linkage data indicated an increase of 3,4,6-linked Man in the serum N-glycome of the patient suffering from hepatic fibrosis, these exoglycosidase data confirm an increase of bisected complex-type N-glycans. After -galactosidase and -N-acetylhexosaminidase treatment (Figure 3C and D), two major molecular ions were observed at m/z 933 (Hex3HexNAc2) and 1079 (Fuc1Hex3HexNAc2), indicating that the GlcNAc residues were in normal -linkages. After -galactosidase, -N-acetylhexosaminidase, and -fucosidase treatment (Figure 3E and F), the molecular ion at m/z 1079 (Fuc1Hex3 HexNAc2) disappeared, and one molecular ion was observed at m/z 933 (Hex3HexNAc2) after 8-h digestion. However, after 6-h digestion, other molecular ions were still observed at m/z 1340 (Hex3HexNAc4), 1543 (Hex3Hex NAc5), and 1705 (Hex4HexNAc5) and were more intense in the serum N-glycome of the patient suffering from cirrhosis. Thus, possibly due to steric hindrance, the bisecting N-acetylglucosamine is partially resistant to the -N-acetylhexosaminidase (Yamashita et al., 1983). 285

W. Morelle et al.

Fig. 2. MALDI-TOF MS analysis of permethylated total serum N-glycome from a cirrhotic patient.

Table III. GC-MS analysis of partially methylated alditol acetates obtained from the PNGase F released N-glycans Relative abundance Retention time (min) 15.36 20.24 21.18a 24.32 25.02 25.24b 27.25b 28.25 29.59 30.27 31.42 32.42 35.09 35.50 37.14 38.24 Characteristic fragment ions 115, 118, 131, 162, 175 102, 118, 129, 145, 161, 162, 205 89, 102, 118, 162, 205, 278 129, 130, 161, 190 118, 233 118, 129, 161, 234 99, 102, 118, 129, 162, 189, 233 130, 190, 233 129, 130, 189, 190 118, 129, 189, 234 188, 333 117, 159, 203, 205 117, 159, 233 117, 159, 233 117, 159, 346 117, 159, 261 Assignment Terminal fucose Terminal mannose Terminal galactose 2-Linked mannose 4-Linked glucose 3-Linked galactose 6-Linked galactose 2,4-Linked mannose 2,6-Linked mannose 3,6-Linked mannose 3,4,6-Linked mannose Terminal GlcNAc 4-Linked GlcNAc 4-Linked HexNAc 3,4-Linked GlcNAc 4,6-Linked GlcNAc A 0.05 0.04 0.12 0.50 0.08 0.09 0.45 0.09 0.05 0.23 0.01 0.04 1.00 0.05 0.01 0.07 B 0.19 0.14 0.22 0.77 0.14 0.08 0.42 0.13 0.06 0.34 0.19 0.38 1.00 0.06 0.04 0.23

The 80% acetonitrile fractions from Sep-Pak purifications of permethylated glycans were hydrolyzed, reduced, acetylated, and analyzed by gas chromatography MS. A, normal total serum N-glycome; B, N-glycome of patient with decompensated cirrhosis. a Signals more intense after chemical desialylation of N-glycans. b Signals not observed after chemical desialylation of N-glycans.

ESI-IT-MS of human serum N-glycome After desialylation, the permethylated N-glycans were analyzed using ESI-IT-MS. The analytical capabilities of ESI-IT-MS allow to analyze oligosaccharides as complex mixture and also offer multiple stages of fragmentation (MSn) to elucidate their structural heterogeneity. When analyzed in this system, permethylated N-glycans form sodiated molecular adducts. All glycans were detected as their [M+2Na]2+ except oligosaccharide (Hex3HexNAc2 DeoxyHex1) which was detected by the ions at m/z 929 ([M+2Na]2+) and 1836 ([M+Na]+). The first stage (MS1) yields resolution of the methylated glycan mixture (Figure 4A) and gives the sugar chemical composition of each isobaric oligosaccharide (Table 1). The second and third stages (MS2 and MS3) give information on the nature of the antennae, the branching pattern, the core structure, and the 286

core fucosylation. The nomenclature describing the fragmentation of carbohydrates is that introduced by Domon and Costello (1988). Desialylated and methylated normal serum N-glycome was analyzed. The MS1 was characterized by the presence of abundant [M+2Na]2+ ions at m/z 1047 (data not shown). MS2 of ions m/z 1047 and MS3 of ions at m/z 1143.5 were characteristic of a biantennary complex Nglycan. Indeed, fragmentation was as described by Weiskopf et al. (1998), with the Y4 or Y4 ions resulting from the loss of disaccharide antenna at m/z 1607.5 for the singly sodiated, at m/z 815.5 for the doubly sodiated, and at m/z 1143.5 for the loss of both antennae. Other observed ions corresponded to [B5]2+ (B5/Y4,4). The MS3 spectrum of Y(4,4)2 ions gave all the characteristic fragment ions at m/z 939.2, 866.2, 662.2 and 639.3 corresponding

Total serum N-glycome in human diseases

Fig. 3. MALDI-TOF MS analysis of exoglycosidase digestions of total human serum desialylated N-glycome from normal subject (A, C, E) and from cirrhotic patient Y (B, D, F) (A, B: -galactosidase treatment; C, D: -galactosidase and -N-acetylhexosaminidase combined treatment; E, F: -galactosidase, -N-acetylhexosaminidase, and -fucosidase combined treatment).

respectively to Y3,3, B5, B5/Y3,3, and C4 ions (data not shown). The MS2 spectrum of [M+2Na]2+ ions at m/z 1134.5 corresponding to the chemical composition of Hex5HexNAc4 DeoxyHex1 was characterized by four major fragment ions at m/z 865.8, 902.1, 1317.7, and 1781.6 corresponding respectively to B5/Y(4,4)2, [Y4,4]2+, Y(4,4)2, and Y4,4 ions. The spectrum corresponds to a biantennary, core-fucosylated N-glycan. The MS2 spectrum of ions at m/z 1271.5, corresponding to the chemical composition of Hex6HexNAc5, was characterized by the fragment ions at m/z 1039.5, 1592.5, and 807.8 corresponding respectively to [Y4,4,4]2+, Y(4,4,4)2, and [Y(4,4,4)2]2+. Other characteristic ions B2,2, B5/Y4,4,4, and [B5]2+ were observed at m/z 486.2, 900.9, and 1132.4. MS3 of Y(4,4,4)3 ions at m/z 1129.0 corresponding to the core resulted in the presence of only two ions at m/z 939.6 and 852.1 corresponding to Y3 and B5 fragment ions. The spectrum corresponds to the fragmentation pattern of a triantennary complex N-glycan. The ESI-IT-MS analysis of the desialylated permethylated serum N-glycome of the patient with decompensated

cirrhosis is shown in Figure 4. In the MS spectrum, the more heterogeneous population of N-glycans observed in the MALDI-TOF MS analysis was confirmed (Figure 4A) and characterized by a relative increase of fucosylated oligosaccharides. MS2 was performed on all major ions. The MS2 spectrum of [M+Na]+ ions at m/z 1836.1 (Hex3HexNAc4DeoxyHex1) or of the doubly charged at m/z 929 is characterized by intense ions at m/z 1577.3 corresponding to the loss of an N-acetylhexosamine (Figure 4B). The ions at m/z 1318.9 correspond to the second Nacetylhexosamine removal. The presence of the B4 ions at m/z 1385.3 indicates the core fucosylation. The MS3 spectrum of ions at m/z 1318.3 is typical of a fucosylated pentasaccharide core (Weiskopf et al., 1998). Taken together, these results show that the N-glycan is an agalactobiantennary complex-type oligosaccharide (Figure 4B and C). The MS2 spectrum of the [M+2Na]2+ ions at m/z 1256.7 and the MS3 spectrum of the Y3/Y(4,4)2 ions at m/z 1303.3 are typical of a core-fucosylated biantennary complex N-glycan with a bisecting GlcNAc residue as described by Weiskopf et al. (1998) (Figure 4D and E). The MS2 spectrum (Figure 4F) of the [M+2Na]2+ ions at m/z 1154.7, corresponding to the chemical composition of 287

W. Morelle et al.

Fig. 4. ESI-IT-MS analysis of total human serum N-glycome of a cirrhotic patient. (A) MS1 of the permethylated oligosaccharides; (B) MS2 of singly charged parent ions at m/z 1836.1; (C) MS3 of the isolated pentasaccharide core of parent ions at m/z 1836.1; (D) MS2 of doubly charged parent ions at m/z 1256.7; (E) MS3 of isolated ions at m/z 1303.3 from the doubly charged parent ions at m/z 1257.1, representing the first three residues at the oligosaccharide reducing end; (F) MS2 of doubly charged parent ions at m/z 1154.7; (G) MS2 of doubly charged parent ions at m/z 1169.7; (H) MS2 of doubly charged parent ions at m/z 1359.0. The nomenclature describing the fragmentation of the carbohydrate is that introduced by Domon and Costello (1988), with a modification: instead of describing a fragmentation as Y4 or Y4, we condensed the abbreviation as Y4,.

288

Total serum N-glycome in human diseases

Hex5HexNAc4DeoxyHex1, was characterized by the loss of antennae observed with Y4 fragment ions at m/z 1822.2 and 923.1, Y4/Y3, Y4 fragment ions at m/z 1563.9 and 793.5, for singly and doubly charged ions, respectively, indicating either a biantennary structure with terminal Nacetylhexosamine residue, a N-acetyllactosamine, and a bisecting GlcNAc, or a triantennary complex oligosaccharide with two agalactosylated antennae. As previously reported, the B5 ions indicate the core fucosylation. Unfortunately, the MS3 spectrum of the ions at m/z 1304 was not enough informative to decide between these two isobaric structures. The MS2 spectrum of the [M+2Na]2+ ions at m/z 1169.7 (Figure 4G) indicates the presence of two Nacetyllactosamine units, but there is an uncertainty concerning the exact position of the N-acetylhexosamine residue. The same ambiguity was also found in the MS2 of 1154.7 ions, and any informative MS3 was obtained. Interpretation of the MS2 spectrum of [M+2Na]2+ ions at m/z 1359.0 (Figure 4H) indicates the presence of the three N-acetyllactosamine units and the core substitution by a fucose residue, suggesting a triantennary fucosylated structure. The structure of all N-glycans is summarized in Table 1. Characterization of N-linked glycans of normal human serum N-glycome Complex-type biantennary oligosaccharide represents the major component of healthy subject N-glycome; the ratio of biantennary to triantennary is about 6:1, as deduced from the relative intensities of ions at m/z 1664 and 2029 (Figure 1B), and is very close to the ratio observed in transferrin (Fu and van Halbeek, 1992). Among the fucosylated oligosaccharides, most of the fucose residues are core 1,6 linked. Traces of 1,3-linked fucose observed in the GCMS analysis (Table 3) were not detected by ESI-IT-MS. 2,6-Linked N-acetylneuraminic acid predominates, as deduced from the GC-MS analysis; the ratio of 2,6 to 2,3 is approximately 5 (Table 3). Trace amounts of neutral oligosaccharides are observed corresponding to agalactosylated biantennary complex glycans. Finally, glycans possessing a bisecting GlcNAc residue are present as very minor components. Qualitative and semi-quantitative aspects of the N-glycome observed by MALDI-TOF are in agreement with previously observed results on various glycoproteins isolated from serums (Wilson et al., 2002; Butler et al., 2003; Callewaert et al., 2003, 2004; Flahaut et al., 2003; Mills et al., 2003). Characterization of N-linked glycans of serum N-glycome isolated from patients with decompensated cirrhosis The distribution of the N-glycans is more heterogeneous and is characterized by the following: (1) the existence of an important subpopulation of neutral and truncated biantennary complex oligosaccharides (ions at m/z 1486, 1502, 1543, 1648, 1689, 1810, and 1851); and (2) the presence of a bisecting GlcNAc residue in the intense ions at m/z 2318 and 2630, as demonstrated by ESI-IT-MS. In ions at m/z 1543, 1689, 1705, 1851, and 1867, the same modification is probably present, based on the increase of 3,4,6-linked

mannose associated with the intensities of 2,4- and 2,6linked mannoses corresponding to triantennary oligosaccharides and on ESI-IT-MS results; and (3) the relative increased core fucosylation.

Discussion We have developed a rapid protocol that can provide total serum N-glycome in less than 24 h. The MALDI-TOF approach allows partial identification of the sialylated and neutral glycans through their sugar composition. The signal strength of each ionized glycan reflects the relative amount of material present in the sample (Harvey, 1999) and can be expressed as relative intensity. All three different phases of the protocolPNGase F digestion, purification of released oligosaccharides via solid-phase extraction on graphitized carbon column, and either methylesterification or chemical desialylationenabled the development of a high-throughput method with full automation. Also, we worked on less than 2% of the initial sample, and this analysis can be achieved on less than one microliter of serum. It is therefore fully suitable for the search of Nglycosylation modifications in congenital or acquired diseases and also for the comparison between the different states of a disease, and it is adapted to clinical laboratory. We are currently assessing this protocol to define prognosis and diagnosis markers in hepatic cirrhosis and hepatocarcinoma. Serum N-glycome was partly characterized by Callewaert et al. (2004). These authors have defined 11 desialylated oligosaccharides depending on their electrophoretic mobility and exoglycosidase digestions. In this study, the N-glycan mix constituting total serum N-glycome was elucidated by MS analyses of desialylated or methylesterified oligosaccharides and finally using permethylation; 26 Nglycans have been characterized. They partly reflect the extreme heterogeneity of total serum N-glycome. The number and position of the antennae, bisecting GlcNAc, the periphery defined by 1,3- and 1,6-linked fucose residues and by 2,3- and 2,6-linked N-acetylneuraminic acid, should result in hundreds of different N-glycan structures. In this study, we have defined the major ones; the MALDI-TOF approach only resolves this heterogeneity as a mixture of isobaric compounds; still it provides the opportunity of using these 26 groups of N-glycans as prognosis or diagnosis markers; a complementary structural analysis involving permethylation makes it possible to define the main changes. ESI-IT-MS allows the structural characterization of N-glycans present in the mixture, with the restriction of isobaric compounds and the sensitivity of the technique that did not enable us to identify 1,3-fucosylated compounds seen in the linkage analysis and present in smaller amounts. Total serum N-glycome in patients with cirrhosis is characterized by the presence of an important population of Nglycans with bisected GlcNAc (Callewaert et al., 2004). Modifications in the expression of N-acetylglucosaminyl transferases III and V (GnTIII and GnTV) are associated with many physiological and pathological processes in the liver (Taniguchi et al., 1999; Ito et al., 2001). It has been 289

W. Morelle et al.

shown that fatty liver is associated with an increased risk of progression to cirrhosis. In particular, GnTIII is involved in the dysfunction of the apolipoprotein metabolism, resulting in lipid accumulation that leads to fatty liver, one of the first steps of alcohol-induced liver injury (Ihara et al., 1998; Lee et al., 2004). The second characteristic of the Nglycome is the relative increase in 1,6-fucosylated oligosaccharides. 1,6-fucosyltransferase (FUT8) has also been involved in steatosis in the liver because of a downregulation of lysosomal acid lipase by the remodeling of its glycosylation (Wang et al., 2001). Alterations in the expression of GnTIII and FUT8 are probably one of the primary events leading to the lipid accumulation in hepatic steatosis. Another major glycosylation change appearing in the N-glycome of patients suffering from hepatic cirrhosis is the presence of neutral agalactosylated oligosaccharides; these N-glycans either result from altered biosynthesis or from the dysfunction of the asialoglycoprotein receptor (Burgess et al., 1992) resulting in the hyperasialoglycoproteinemia observed in cirrhosis and hepatocellular carcinomas (Sawamura et al., 1984). Total N-glycome is the sum of all serum glycoprotein glycosylations. To define what proteins are modified, complementary studies need to be done. Proteomic approaches have been used in liver cancer, where the increase in fucosylation has been associated with fetoprotein, 1-acid glycoprotein, 1-antitrypsin, and transferrin (Block et al., 2005), and only quantitative modifications were observed in cirrhotic patients by highresolution two-dimensional electrophoresis (Gravel et al., 1996). If the neutral agalactosylated oligosaccharides were present on liver glycoproteins such as transferrin, the isoelectrophoretic behavior should be modified since the loss of terminal sialic acids leads to a cathodic shift of the isoelectric point of glycoproteins. Therefore, the isoelectrophoretic pattern of transferrin should be close to the one observed in CDG Type II; curiously, these are normal for the three patients (data not shown). Finally, a positive correlation was observed between the relative intensity of ions at m/z 1486 and the immunoglobulin G (IgG) concentration (data not shown). The oligosaccharide structure corresponding to ions at m/z 1486 (Table 1) is present on immunoglobulin (Parekh et al., 1985; Fujii et al., 1990; Moore et al., 2005). One possible hypothesis is that the neutral oligosaccharides are present on pro-inflammatory galactose-deficient immunoglobulins, which are found in numerous diseases (Moore et al., 2005). Affinity chromatography purification of cirrhotic patient immunoglobulins will confirm the origin of the neutral oligosaccharides observed. The presence of a relative increase of multiantennary difucosylated oligosaccharides in patient Y as compared with patients X and Z might correspond to the modifications of the degree of branching of 1-acid glycoprotein oligosaccharides observed during acute and chronic inflammation (van Dijk et al., 1995; Higai et al., 2003). Further investigations, of additional patients with various inflammatory diseases treated or not, will precise the N-glycome modifications observed during inflammation. The point of developing a sensitive assay for screening modification of total serum N-glycome is to obtain diagnosis 290

and prognosis markers for chronic diseases such as the GlycoCirrhoTest (Callewaert et al., 2004). This test, obtained after desialylation of the N-glycome, relates to the logarithmic ratio of the relative intensity of two N-linked oligosaccharides, corresponding in our study to ions at m/z 2013 and 2029. From the sialylated or unsialylated N-glycome, we currently investigate into all the mathematical combinations offered by this MALDI-TOF MS approach, to define new noninvasive biomarkers.

Materials and methods Patient samples Serums were obtained from patient admitted in the Department of Gastroenterology and Hepatology of Lille University Hospital for decompensation of cirrhosis or hepatitis. Control serums were obtained from healthy volunteers. Patient X was suffering from decompensated cirrhosis with acute alcoholic hepatitis, Patient Y was suffering from liver fibrosis with acute alcoholic hepatitis, and Patient Z was suffering from chronic hepatitis C. Serum albumin was measured with the bromcresol green method; the total bilirubin, the aspartate aminotransferase (AST), and the alanine aminotransferase (ALT) were measured according to the International Federation of Clinical Chemistry (IFCC) recommendations; and the tests were performed on an Olympus AU600. IgG were measured by nephelemetry on a BN2 (dade Behring, Paris, France). Fibrosis and inflammation were evaluated after liver biopsy, and the grading and staging was done according to the METAVIR scoring system (The METAVIR cooperative group, 1994). Release and purification of the serum N-glycome Serums (10 L) were denatured at 100C in the presence of sodium dodecyl sulfate and -mercaptoethanol. After the addition of nonidet-P 40, the protein N-glycanase F (PNGase F, New England Biolabs, Herts, UK) was added (1 L) (500,000 U/mL) for 3 h. The oligosaccharides were purified by solid-phase extraction on a graphitized carbon adsorbent (Alltech Associates, Templemars, France), and the neutral and the sialylated oligosaccharides were eluted together with an acetonitrilewater solution (25/75; v/v) containing trifluoroacetic acid (0.1%) (Packer et al., 1998). Half of total N-glycan fraction was methylesterified, and the remaining part was chemically desialylated. Esterification of sialic acids To stabilize the sialic acid moiety under MALDI-MS conditions, the sialic acid residues of PNGase F-released oligosaccharides were stabilized by methylesterification of their carboxylic group (Powell and Harvey, 1996). Chemical desialylation Sialic acids were cleaved by treatment of N-glycan samples with acetic acid, 2 M at 80C, for 2 h (Varki and Diaz, 1984).

Total serum N-glycome in human diseases

Exoglycosidase digestions These were carried out on the released desialylated Nglycans using the following enzymes and conditions: Nacetyl--D-hexosaminidase (from jack bean, EC 3.2.1.52, Sigma, St-Quentin Fallavier, France), 0.2 U in 100 L of 50 mM ammonium formate buffer, pH 4.6; -galactosidase (from bovine testes, EC 3.2.1.23, Sigma), 10 mU in 100 L of 50 mM ammonium formate buffer, pH 4.6, -L-fucosidase (from bovine kidney, EC 3.2.1.51, Roche Molecular Biochemicals, Meylan, France), 0.2 U in 100 L of 100 mM ammonium acetate buffer, pH 4.5. Permethylation of N-glycans Permethylation using the sodium hydroxide procedure was performed according to Ciucanu and Kerek (1984). After derivatization, the reaction products were purified on C18Sep-Pak (Waters Ltd., St-Quentin en Yvelines, France) according to Dell et al. (1994). Linkage analysis of N-glycans Partially methylated alditol acetates were prepared from permethylated samples for GC-MS linkage analysis as described previously (Albersheim et al., 1967). GC-MS analysis were recorded using an Automass II 30 quadrupole mass spectrometer interfaced with a Carlo Erba 8000 Top gas chromatograph (Finnigan, Argenteuil, France). Electron ionization spectra were recorded using ionization energy of 70 eV. Gas chromatograph was equipped with a CP-Sil 5CB/MS capillary column (25 mm 0.32 mm, Chrompak, les Ullis, France). The gas vector was at a flow rate of 2 mL/min. The partially methylated alditol acetates were dissolved in methanol before on-column injection at 130C. The GC oven was held at 130C for 1 min before increasing to 180C at 2C/min and then to 240C at 4C/min. MALDI-TOF mass spectrometry All mass spectra were acquired on a Voyager Elite (DESTR) reflectron time-of-flight (TOF) mass spectrometer (Perseptive Biosystems, Framingham, MA) equipped with a pulsed nitrogen laser (337 nm) and a gridless delayed extraction ion source. Samples were analyzed in delayed extraction mode using an accelerating voltage of 20 kV, a pulse delay time of 200 ns, and a grid voltage of 66%. Detector bias gating was used to reduce the ion current below masses of 500 Da. After external calibration, between 100 and 200 scans were averaged for each spectrum shown. For all measurements, the dried droplet preparation technique was employed. Native and methylesterified Nglycans were co-crystallized with 2,5-dihydroxybenzoic acid (DHB) as matrix (10 mg/mL of DHB in an acetonitrile/ water solution [70 : 30] containing 0.1% of trifluoroacetic acid). For desialylated oligosaccharides, one part out of 100 was loaded on the target, whereas one part out of fifty was used for methylesterified N-glycans. Electrospray mass spectrometry Structural characterization of methylated N-glycan samples was carried out using electrospray ionization with an

LCQ Deca XP+ ion trap mass spectrometer equipped with a nanoelectrospray ionization source (Thermo-electron, San Jose, CA). Samples dissolved in a solution of 70% methanol and 1% formic acid were sprayed from a goldcoated medium-length borosilicate capillaries (Proxeon, Odense, Denmark). The spray voltage was 1 kV and the capillary temperature was 170C. No sheath gas was used. The total ionic current (TIC) signal was acquired for 1 min. For the generation of the MSn spectra, collision energies were set to 2040% of maximum. The isolation width was set to 2 U. All experiments were performed in the positive-ion mode.

Acknowledgments We are grateful to Prof. Philippe Delannoy for comments on the manuscript. We thank Adeline Page for assistance with the mass spectrometer. This research was supported by the Centre National de la Recherche (Unit Mixte de Recherche CNRS/USTL 8576; Director: Dr Jean-Claude Michalski) and the Ministre de la Recherche et de lEnseignement Suprieur. The Mass Spectrometry facility used in this study was funded by the European Community (FEDER), the Rgion Nord-Pas de Calais (France), the CNRS, and the Universit des Sciences et Technologies de Lille.

Abbreviations CDG, congenital disorders of glycosylation; ESI-IT-MS, electrospray ionization ion trap mass spectrometry; GC-MS, gas chromatography mass spectrometry; IgG, immunoglobulin G; MALDI-TOF MS, matrix-assisted laser desorption ionization time of flight mass spectrometry; PNGase F, protein N-glycosidase F.

References
Adamowicz, M. and Pronicka, E. (1996) Carbohydrate deficient glycoprotein syndrome-like transferrin isoelectric focusing pattern in untreated fructosaemia. Eur. J. Pediatr., 155, 347348. Albersheim, P., Nevins, D.J., English, P.D., and Karr, A. (1967) A method for the analysis of sugars in plant cell wall polysaccharides by gas liquid chromatography. Carbohydr. Res., 5, 340345. Anderson, N., Pollacchi, A., Hayes, P., Therapondos, G., Newsome, P., Boyter, A., and Smith, K. (2002) A preliminary evaluation of the differences in the glycosylation of alpha-1-acid glycoprotein between individual liver diseases. Biomed. Chromatogr., 16, 365372. Basset, C., Durand, V., Jamin, C., Clement, J., Pennec, Y., Youinou, P., Dueymes, M., and Roitt, I.M. (2000) Increased N-linked glycosylation leading to oversialylation of monomeric immunoglobulin A1 from patients with Sjogrens syndrome. Scand. J. Immunol., 51, 300306. Block, T.M., Comunale, M.A., Lowman, M., Steel, L.F., Romano, P.R., Fimmel, C., Tennant, B.C., London, W.T., Evans, A.A., Blumberg, B.S., and others. (2005) Use of targeted glycoproteomics to identify serum glycoproteins that correlate with liver cancer in woodchucks and humans. Proc. Natl. Acad. Sci. U. S. A., 102, 779784. Burgess, J.B., Baenziger, J.U., and Brown, W.R. (1992) Abnormal surface distribution of the human asialoglycoprotein receptor in cirrhosis. Hepatology, 15, 702706. Butler, M., Quelhas, D., Critchley, A.J., Carchon, H., Hebestreit, H.F., Hibbert, R.G., Vilarinho, L., Teles, E., Matthijs, G., Schollen, E., and

291

W. Morelle et al.

others. (2003) Detailed glycan analysis of serum glycoproteins of patients with congenital disorders of glycosylation indicates the specific defective glycan processing step and provides an insight into pathogenesis. Glycobiology, 13, 601622. Callewaert, N., Geysens, S., Molemans, F., and Contreras, R. (2001) Ultrasensitive profiling and sequencing of N-linked oligosaccharides using standard DNA-sequencing equipment. Glycobiology, 11, 275281. Callewaert, N., Schollen, E., Vanhecke, A., Jaeken, J., Matthijs, G., and Contreras, R. (2003) Increased fucosylation and reduced branching of serum glycoprotein N-glycans in all known subtypes of congenital disorder of glycosylation I. Glycobiology, 13, 367375. Callewaert, N., Van Vlierberghe, H., Van Hecke, A., Laroy, W., Delanghe, J., and Contreras, R. (2004) Noninvasive diagnosis of liver cirrhosis using DNA sequencer-based total serum protein glycomics. Nat. Med., 10, 429434. Charlwood, J., Clayton, P., Keir, G., Mian, N., and Winchester, B. (1998) Defective galactosylation of serum transferrin in galactosemia. Glycobiology, 8, 351357. Ciucanu, I. and Kerek, F. (1984) A simple and rapid method for the permethylation of carbohydrates. Carbohydr. Res., 131, 209217. De Graaf, T.W., Van der Stelt, M.E., Anbergen, M.G., and van Dijk, W. (1993) Inflammation-induced expression of sialyl Lewis X-containing glycan structures on alpha, 1-acid glycoprotein (orosomucoid) in human sera. J. Exp. Med. 177, 657666. Dell, A., Reason, A.J., Khoo, K.H., Panico, M., McDowell, R.A., and Morris, H.R. (1994) Mass spectrometry of carbohydrates-containing biopolymers. Methods Enzymol., 230, 108132. Domon, B. and Costello, C.E. (1988) Systematic nomenclature for carbohydrate fragmentations in FAB-MS/MS spectra of glycoconjuguates. Glycoconj. J., 5, 397409. Flahaut, C., Michalski, J.C., Danel, T., Humbert, M.H., and Klein, A. (2003) The effects of ethanol on the glycosylation of human transferrin. Glycobiology, 13, 191198. The French METAVIR, Cooperative Study Group (1994) Intraobserver and interobserver variations in liver biopsy interpretation in patients with chronic hepatitis C. Hepatology, 20, 1520. Fu, D., and van Halbeek, H. (1992) 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. 206, 5363. Fujii, S., Nishiura, T., Nishikawa, A., Miura, R., and Taniguchi, N. (1990) Structural heterogeneity of sugar chains in immunoglobulin G. Conformation of immunoglobulin G molecule and substrate specificities of glycosyltransferases. J. Biol. Chem., 265, 60096018. Goodarzi, M.T. and Turner, G.A. (1998) Reproducible and sensitive determination of charged oligosaccharides from haptoglobin by PNGase F digestion and HPAEC/PAD analysis: glycan composition varies with disease. Glycoconj. J., 15, 469475. Gravel, P., Walzer, C., Aubry, C., Balant, L.P., Yersin, B., Hochstrasser, D.F., and Guimon, J. (1996) New alterations of serum glycoproteins in alcoholic and cirrhotic patients revealed by high resolution twodimensional gel electrophoresis. Biochem. Biophys. Res. Commun., 220, 7885. Grnewald, S., Matthijs, G., and Jaeken, J. (2002) Congenital disorders of glycosylation: a review. Pediatr. Res., 52, 618624. Hachulla, E., Laine, A., and Hayem, A. (1988) Alpha, 1-antichymotrypsin microheterogeneity in crossed immunoaffinoelectrophoresis with free concanavalin A: a useful diagnostic tool in inflammatory syndrome. Clin. Chem. 34, 911915. Harvey, D.J. (1999) Matrix-assisted laser desorption/ionization mass spectrometry of carbohydrates. Mass Spectrom. Rev., 18, 349450. Higai, K., Azuma, Y., Aoki, Y., and Matsumoto, K. (2003) Altered glycosylation of alpha1-acid glycoprotein in patients with inflammation and diabetes mellitus. Clin. Chim. Acta, 329, 117125. Ihara, Y., Yoshimura, M., Miyoshi, E., Nishikawa, A., Sultan, A.S., Toyosawa, S., Ohnishi, A., Suzuki, M., Yamamura, K., Ijuhin, N., and Taniguchi, N. (1998) Ectopic expression of N-acetylglucosaminyltransferase III in transgenic hepatocytes disrupts apolipoprotein B secretion and induces aberrant cellular morphology with lipid storage. Proc. Natl. Acad. Sci. U. S. A., 95, 25262530.

Ito, Y., Miyoshi, E., Sakon, M., Takeda, T., Noda, K., Tsujimoto, M., Ito, S., Honda, H., Takemura, F., Wakasa, K., Monden, M., and others. (2001) Elevated expression of UDP-N-acetylglucosamine: alphamannoside beta1,6 N-acetylglucosaminyltransferase is an early event in hepatocarcinogenesis. Int. J. Cancer, 91, 631637. Jaeken, J., Vanderschueren-Lodeweyckx, M., Casaer, P., Snoek, L., Corbeel, L., Eggermont, E., and Eeckels, R. (1980) Familial psychomotor retardation with markedly fluctuating serum proteins, FSH and GH levels, partial TBG-deficiency, increased serum arylsulfatase A and increase CSF protein: a new syndrome? Pediatr. Res., 14, 179. Larsson, A., Flodin, M., and Kollberg, H. (1998) Increased serum concentrations of carbohydrate-deficient transferrin (CDT) in patients with cystic fibrosis. Ups. J. Med. Sci. 103, 231236. Lee, J., Song, E.Y., Chung, T.W., Kang, S.K., Kim, K.S., Chung, T.H., Yeom, Y.I., and Kim, C.H. (2004) Hyperexpression of N-acetylglucosaminyltransferase-III in liver tissues of transgenic mice causes fatty body and obesity through severe accumulation of Apo A-I and Apo B. Arch. Biochem. Biophys., 426, 1831. Mechref, Y. and Novotny, M.V. (1998) Mass spectrometric mapping and sequencing of N-linked oligosaccharides derived from submicrogram amounts of glycoproteins. Anal. Chem., 70, 455463. Mills, K., Mills, P.B., Clayton, P.T., Mian, N., Johnson, A.W., and Winchester, B.G. (2003) The underglycosylation of plasma alpha, 1-antitrypsin in congenital disorders of glycosylation type I is not random. Glycobiology, 13, 7385. Moore, J.S., Wu, X., Kulhavy, R., Tomana, M., Novak, J., Moldoveanu, Z., Brown, R., Goepfert, P.A., and Mestecky, J. (2005) Increased levels of galactose-deficient IgG in sera of HIV-1-infected individuals. AIDS, 19, 381389. Packer, N.H., Lawson, M.A., Jardine, D.R., and Redmond, J.W. (1998) A general approach to desalting oligosaccharides released from glycoproteins. Glycoconj. J., 15, 737747. Papac, D.I., Briggs, J.B., Chin, E.T., and Jones, A.J. (1998) A highthroughput microscale method to release N-linked oligosaccharides from glycoproteins for matrix-assisted laser desorption/ionization time-of-flight mass spectrometric analysis. Glycobiology, 8, 445454. Parekh, R.B., Dwek, R.A., Sutton, B.J., Fernandes, D.L., Leung, A., Stanworth, D., Rademacher, T.W., Mizuochi, T., Taniguchi, T., Matsuta, K., and others. (1985) Association of rheumatoid arthritis and primary osteoarthritis with changes in the glycosylation pattern of total serum IgG. Nature, 316, 452457. Powell, A.K. and Harvey, D.J. (1996) Stabilization of sialic acids in Nlinked oligosaccharides and gangliosides for analysis by positive ion matrix-assisted laser desorption/ionization mass spectrometry. Rapid Commun. Mass Spectrom., 10, 10271032. Ryden, I., Pahlsson, P., and Lindgren, S. (2002) Diagnostic accuracy of alpha(1)-acid glycoprotein fucosylation for liver cirrhosis in patients undergoing hepatic biopsy. Clin. Chem., 48, 21952201. Sawamura, T., Nakada, H., Hazama, H., Shiozaki, Y., Sameshima, Y., and Tashiro, Y. (1984) Hyperasialoglycoproteinemia in patients with chronic liver diseases and/or liver cell carcinoma. Asialoglycoprotein receptor in cirrhosis and liver cell carcinoma. Gastroenterology, 87, 12171221. Taniguchi, N., Miyoshi, E., Ko, J.H., Ikeda, Y., and Ihara, Y. (1999) Implication of N-acetylglucosaminyltransferases III and V in cancer: gene regulation and signaling mechanism. Biochim. Biophys. Acta., 1455, 287300. van Dijk, W., Havenaar, E.C., and Brinkman-van der Linden, E.C. (1995) Alpha1-acid glycoprotein (orosomucoid): pathophysiological changes in glycosylation in relation to its function. Glycoconj. J. 12, 227233. Varki, A. (1999) Acquired glycosylation changes in human disease. In Varki, A., Cummings, R., Esko, J., Freeze, H., Hart, J., and Marth, J. (eds), Essentials of Glycobiology. Cold Spring Harbor, New York, pp. 565580. Varki, A. and Diaz, S. (1984) The release and the purification of sialic acids from glycoconjuguates: methods to minimize the loss and migration of O-acetyl groups. Anal. Biochem. 137, 236247. Wang, W., Li, W., Ikeda, Y., Miyagawa, J.I., Taniguchi, M., Miyoshi, E., Sheng, Y., Ekuni, A., Ko, J.H., Yamamoto, Y., and others. (2001) Ectopic expression of alpha1,6 fucosyltransferase in mice causes

292

Total serum N-glycome in human diseases

steatosis in the liver and kidney accompanied by a modification of lysosomal acid lipase. Glycobiology, 11, 165174. Weiskopf, A.S., Vouros, P., and Harvey, D.J. (1998) Electrospray ionization-ion trap mass spectrometry for structural analysis of complex N-linked glycoprotein oligosaccharides. Anal. Chem., 70, 44414447 Wilson, N.L., Schulz, B.L., Karlsson, N.G., and Packer, N.H. (2002) Sequential analysis of N- and O-linked glycosylation of 2D-PAGE separated glycoproteins. J. Proteome Res. 1, 521529.

Yamashita, K., Hitoi, A., Matsuda, Y., Tsujii, A., Katunuma, N., and Kobata, A. (1983) Structural studies of the carbohydrate moieties of rat kidney gamma-glutamyltranspeptidase. An extremely heterogeneous pattern enriched with nonreducing terminal N-acetylglucosamine residues. J. Biol. Chem., 258, 10981107. Yamashita, K., Koide, N., Endo, T., Iwaki, Y., and Kobata, A. (1989) Altered glycosylation of serum transferrin of patients with hepatocellular carcinoma. J. Biol. Chem., 264, 24152423.

293