| Glycobiology | Pages |
Structural characterization of the oligosaccharides of a human monoclonal anti-lipopolysaccharide immunoglobulin M
Introduction
Results
Discussion
Materials and methods
Acknowledgments
Abbreviations
Abbreviations
Structural characterization of the oligosaccharides of a human monoclonal anti-lipopolysaccharide immunoglobulin M
The oligosaccharide side chains of a human anti-lipopolysaccharide IgM produced by a human-human-mouse heterohybridoma were analyzed at each of its five conserved N-glycosylation sites. This antibody also has a potential sixth N-glycosylation site in the variable region of its heavy chain which is not glycosylated. The oligosaccharides were released by digestion with various endo- and exoglycosidases and analyzed by matrix-assisted laser desorption/ionization-time of flight mass spectrometry and fluorophore-assisted carbohydrate electrophoresis. The antibody has various complex- and hybrid-type oligosaccharide structures at Asn 171, various sialylated complex-type oligosaccharides at Asn 332 and 395, and high-mannose-type oligosaccharides at Asn 402 and 563. Of note is the presence in this human IgM of oligosaccharides containing N-glycolylneuraminic acid and N-acetylneuraminic acid in the ratio of 98:2 as determined using anion-exchange chromatography. Furthermore, we observed oligosaccharide structures containing Gal[alpha](1,3)Gal that have not been reported as components of human glycoproteins. Key words: site-specific N-glycosylation/human monoclonal IgM/FACE®/MALDI-TOF
Introduction
Cultured mammalian cells are used to produce proteins for therapeutic and diagnostic use because of the cells ability to perform complex posttranslational modifications that include glycosylation. Oligosaccharide moieties added posttranslationally can play an important role in defining the biological properties of glycoproteins, including clearance rate, immunogenicity, and specific biological activity (Wormald et al., 1991; Bazin et al., 1992; Kusakabe et al., 1994). It is now well-established that glycosylation can be species-specific (Kornfeld and Kornfeld, 1985; Rademacher et al., 1988; Kniskern et al., 1994; Matsumoto et al., 1995; Stimson et al., 1995) as well as tissue- and cell-specific within a given species (Parekh et al., 1987; Smith and Baenziger 1992; Skelton et al., 1992). This variability serves an important function within the cell, permitting fine regulation of clearance rate and biological activity for at least some proteins. Some carbohydrate structures have also been found to be antigenic, particularly terminal sialic acids and galactose [alpha](1,3)-linked to galactose [Gal[alpha](1,3)Gal]. Therefore, differences and/or modifications in oligosaccharide structures can dramatically affect the biological properties of glycoproteins (Hotchkiss et al., 1988).
Human immunoglobulin M is a glycoprotein containing 7-12% carbohydrate distributed at five conserved N-glycosylation sites located in the constant region of the heavy chain at positions Asn 171, Asn 332, Asn 395, Asn 402, and Asn 563. Some antibodies have also been reported with N-glycosylation sites in the variable region (Abel et al., 1968; Wallick et al., 1988; Wright et al., 1991; Leibiger et al., 1995).
All N-linked oligosaccharides share the same pentasaccharide core structure (GlcNAc[beta]1,4GlcNAc[beta]1,4Man[[alpha]1,6Man][alpha]1,3Man), but a wide variety and complexity of oligosaccharides can be attached to this core. N-linked oligosaccharides fall into one of three subtypes based on the carbohydrate residues attached to the core pentasaccharide (1) complex-type oligosaccharides, (2) high-mannose-type oligosaccharides, and (3) hybrid-type oligosaccharides that are a combination of the first two types.
Analysis of the oligosaccharide side chains of human IgM at individual glycosylation sites has been mostly conducted on pathological IgM derived from patients with Waldenström's macroglobulinemia (Hickmann et al., 1972). Pathological IgM was found to possess oligomannose structures at glycosylation sites Asn 402 and Asn 563 and complex-type structures at Asn 171, 332, and 395 (Wormald et al., 1991; Monica et al., 1993, 1995). The glycosylation of glycoproteins from an organism in a disease state may differ from that under normal conditions (Parekh et al., 1988, 1989). Cahour et al. (1983) reported differences for IgM from Waldenström's patients in terms of a predominance of multiantennary structures in comparison to that of serum IgM from normal individuals. For this reason the glycosylation of human monoclonal IgM is likely to be different from IgM derived from patients with Waldenström's macroglobulinemia or normal human serum IgM.
To date, monoclonal antibodies have been used mainly as diagnostic tools, but a significant number of monoclonal antibodies are being used as therapeutic agents in various stages of clinical trials. Given the expected influence of the oligosaccharide side chains on physicochemical properties, clearance rate from the bloodstream, antigenicity, and specific biological activity, it is important to expand the limited data currently available concerning human monoclonal IgM glycosylation. The primary object of the present study was to determine the oligosaccharide structures at each of the five N-glycosylation sites of a human monoclonal IgM. We report the complete structures of the major oligosaccharides of this human monoclonal IgM as determined by matrix-assisted laser desorption/ionization-time of flight (MALDI/TOF) mass spectrometry and fluorophore-assisted carbohydrate electrophoresis (FACE®).
Results
Sequence of the heavy and light chain variable regions
We deduced from cDNA sequence analysis that the gene of the heavy chain variable region (VH) belongs to the VHIII gene family and the gene of the light chain variable region (VL) belongs to the VL[lambda]1 gene family (Kabat et al., 1987). The variable region of the heavy chain has a potential N-glycosylation sequence at position Asn 56 (Asn-Lys-Ser). Asn 56 is in the complementary determining region 2 (CDR2). The light chain does not have any N-glycosylation sequence (Figure 1).
Figure 1. The amino acid sequences of the variable region genes of the heavy and light chain of the LPD5H4 antibody. The sequences are shown using one-letter codes. The complementary determining regions (CDR) are printed boldface type. The potential N-glycosylation site in the CDR2 region of the heavy chain is underlined. Monosaccharide composition of the LPD5H4 antibody
Monosaccharide composition of the oligosaccharides of the antibody LPD5H4 (Table I) was determined by gas chromatographic analysis of the TMS derivatives of the methyl glycosides and methyl esters. Glycosyl residues account for 11.14% of the mass of the LPD5H4 antibody. Mannose is 41% of the glycosyl residues, which may be present on hybrid- and/or high-mannose-type structures. No N-acetylgalactosamine (GalNAc) was obtained from this antibody suggesting that the LPD5H4 antibody does not contain O-linked oligosaccharides.
Table I
Monosaccharide
µg Monosaccharide/ 100 µg sample
Mol% total carbohydrates in LPD5H4
Fucose
0.7
7.2
Galactose
1.7
17.5
N-Acetylglucosamine
2.8
22.6
Mannose
4.1
41.1
N-Acetylneuraminic acid
0.04
0.4
N-Glycolylneuraminic acid
1.8
11.2
All oligosaccharide side chains appear to be located on the heavy chains of the LPD5H4 as sequence analysis established. There are no Asn-X-Ser/Thr N-glycosylation sequences in the light chain as demonstrated by sequence analysis, and there are no O-linked oligosaccharide side chains determined by the absence of GalNAc.
Lectin blots
The light and heavy chains of LPD5H4 were separated by electrophoresis, transferred onto nitrocellulose, and identified from their molecular weights. Lectin blots of the separated chains of the LPD5H4 antibody on nitrocellulose were performed to identify some of the glycosyl residue linkages in this human hybridoma IgM.
Sambucus nigra agglutinin (SNA) and Maakia amurensis agglutinin (MAA) blots of the heavy chain of the LPD5H4 antibody are positive indicating the presence of [alpha](2,6)-linked NeuAc/NeuGc and of [alpha](2,3)-linked NeuAc/NeuGc residues, respectively. The heavy chain also stained with Galanthus nivalis agglutinin (GNA) that binds strongly to non-reducing terminal Man[alpha](1,3)Man residues and binds weakly to Man[alpha](1,2)Man and Man[alpha](1,6)Man residues. In addition to a positive staining with Datura stramonium agglutinin (DSA) that specifically recognizes Gal[beta](1,4)GlcNAc residues on N-linked oligosaccharides, a positive staining was observed with Bandeiraea simplicifolia agglutinin-I B4 (BS-I B4) that specifically recognizes galactose [alpha](1,3)-linked to galactose residues. The positive staining for [alpha](2,3)- and [alpha](2,6)-linked NeuAc/NeuGc, [beta]-linked galactose, and mannosyl residues indicates the presence of sialylated complex oligosaccharide structures. These structures may terminate in neuraminic acids or be partially processed terminating in galactosyl or mannosyl residues on one or more branches. The presence of terminal mannosyl residues and the high amounts of mannose determined in the monosaccharide composition analysis also suggests the presence of high-mannose-and/or hybrid-type oligosaccharide structures in the LPD5H4 antibody. The positive staining with BS-I B4 may indicate the presence of complex oligosaccharides that carry the antigenic galactosyl epitope Gal[alpha](1,3)Gal.
Identification of glycopeptides
Glycopeptides were separated from nonglycosylated peptides in the tryptic digest of LPD5H4 on a Con A-Sepharose column. Unbound peptides were washed from the column in Tris-buffer. The Con A-bound glycopeptides were eluted in two fractions, the first was with buffer containing 200 mM [alpha]-methylglucopyranoside (fraction I), and the second was with buffer containing 500 mM [alpha]-methyl-mannopyranoside (fraction II). Fractions I and II were separately fractionated on a C18 reversed-phase column. The reversed-phase HPLC profiles of fractions I contained seven (T1-T7) and fraction II three (T7-T9) glycopeptides, respectively (Figure 2a,b). Glycopeptides were subjected to MALDI/TOF mass spectrometry and then treated with N-glycanase and reanalyzed by MALDI/TOF mass spectrometry to compare their molecular weights before and after N-glycanase treatment. Six peptides (T1, T2, and T4-T7) were confirmed as glycopeptides by their shift in molecular weight after N-glycanase treatment. An example of the effect of N-glycanase treatment is illustrated for glycopeptide T5, in Figure 3. Although glycopeptides T4 and T5 differed in their reversed-phase elution time by almost 5 min, they had the same molecular weight as determined by MALDI/TOF mass spectrometry. After treatment with N-glycanase the molecular weights of T4 and T5 shifted about 2083 Da. Amino acid sequence analysis confirmed that glycopeptides T4 and T5 have the same amino acid sequence. The reason for the difference in their reversed-phase HPLC elution times remains to be determined.
Figure 2. HPLC profile demonstrating the separation of glycopeptide containing Con A fractions I and II. Glycopeptides were separated by reversed-phase HPLC (C18 column). The column was eluted at 1 ml/min with a 0.1% TFA/0.085% TFA in 80% acetonitrile gradient. (A) Fraction I contained seven and (B) fraction II contained three glycopeptides.
Figure 3. Mass determination of glycopeptide T5 using MALDI/TOF mass spectrometry. The mass spectra shown represent glycopeptide T5 before treatment with N-glycanase (A) and after treatment with N-glycanase (B). Double peaks are due to Na- and K-adducts of the glycopeptide.
One glycopeptide was eluted at 42 min in the reversed-phase HPLC profiles of both Con A fractions. This glycopeptide (T6) was not completely eluted with buffer containing 200 mM [alpha]-methylglucopyranoside and, therefore, was also present in fraction II. The molecular weight of glycopeptide T6 lost more than 3600 Da after treatment with N-glycanase, which corresponds to the loss of a tetrasialylated tetra-antennary oligosaccharide. However, Con A binds only biantennary-type oligosaccharides (Cummings, 1994); thus, glycopeptide T6 seemed unlikely to contain a tetra-antennary oligosaccharides. Therefore, glycopeptide T6 is likely to contain two potential N-glycosylation sites carrying small complex- and/or hybrid-/high-mannose-type oligosaccharides. The only anticipated tryptic peptide of LPD5H4 that contains two potential N-glycosylation sites is a peptide from constant region 3 of the heavy chain, which possesses N-glycosylation sites Asn 395 and Asn 402. T6 was identified by amino acid sequence analysis. It contains the N-glycosylation sites Asn 395 and Asn 402. To identify the oligosaccharides of each of these sites, glycopeptide T6 was digested with endoproteinase Glu-C. The resulting peptides and glycopeptides were separated by reversed-phase HPLC (Figure 4). A small amount of T6 was not cleaved by endoproteinase Glu-C under the chosen conditions. The endoproteinase Glu-C digest contained four peptides, two of which (T6-1 and T6-4) were identified as glycopeptides by MALDI/TOF mass spectrometry analysis of each component before and after N-glycanase digestion (data not shown).
Figure 4. HPLC profile showing the separation of glycopeptide T6 after digestion with endoproteinase Glu C. Glycopeptides were separated by reversed-phase HPLC (C18 column). The column was eluted at 1 ml/min with a 0.1% TFA, 0.085% TFA in 80% acetonitrile gradient.
The points of attachment of N-linked carbohydrates in the heavy chain of the antibody were identified by amino acid sequence analysis of the glycopeptides. The glycopeptides and their corresponding amino acid sequences are shown in Table II.
Table II
Glycopeptide
Sequence
Glycosylation site
T1
NNSDISSTR
Asn 171
T2
YKNNSDISSTR
Asn 171
T4/T5
GLTFQQNASSMCVPDQDTAIR
Asn 332
T6
THTNISESHPNATFSAVGEASICEDDWNSGER
Asn 395 and 402
T6-1
THTNISE
Asn 395
T6-4
SHPNATFSAVGE
Asn 402
T7
STGKPTLYNVSLVMSDTAGTCY
Asn 563
Oligosaccharide profiles of glycopeptides
Profiles of the oligosaccharides of each glycopeptide were determined using fluorophore-assisted carbohydrate electrophoresis® (FACE®). Asparagine-linked oligosaccharides were released from glycopeptides using endoglycosidase H and N-glycanase. The released oligosaccharides were labeled with 8-aminonaphthalene-1,3,6,-trisulfonate (ANTS) and subjected to electrophoresis following the instructions of the manufacturer. Characteristic oligosaccharide band patterns were obtained for the carbohydrates released from each glycopeptide (Figure 5a).
Some oligosaccharides showed very similar electrophoretic mobilities and migrated close to each other in the gel (e.g., certain high-mannose-type glycans and certain complex-type sialylated glycans). Therefore, we initially released hybrid- and high-mannose-type oligosaccharides with endoglycosidase H and separated them from the glycopeptides. In a second step we released all remaining oligosaccharides from the glycopeptides with N-glycanase. The most heterogeneity in the glycosylation pattern was found on glycopeptide T2 which contains seven complex- and five hybrid- and/or high-mannose-type oligosaccharides (Figure 5a, lanes 6 and 7). Glycopeptide T5 contains four complex-type oligosaccharides (Figure 5a, lane 5) and glycopeptide T7 three hybrid- and/or high-mannose-type oligosaccharides (Figure 5a, lane 2). Glycopeptide T6-1 contains four complex-type oligosaccharides (Figure 5a, lane 4) and T6-4 five hybrid- and/or high-mannose-type oligosaccharides (Figure 5a, lane 3).a
| Figure 5. (a) Oligosaccharide profiles of the oligosaccharides derived from glycopeptides T2, T3, T5, T6-1, T6-4, and T7 of the LPD5H4 antibody using FACE®. Lane 1, Oligo Ladder Standard (glucose polymers); lane 2, T7 (Asn 563) oligosaccharides released by treatment with endoglycosidase H; lane 3, T6-4 (Asn 402) oligosaccharides released by treatment with endoglycosidase H; lane 4, T6-1 (Asn 395) oligosaccharides released by treatment with N-glycanase; lane 5, T5 (Asn 332) oligosaccharides released by treatment with N-glycanase; lane 6, T2 (Asn 171) oligosaccharides released by treatment with N-glycanase after removing hybrid-/high-mannose-type oligosaccharides; lane 7, T2 (Asn 171) oligosaccharides released by treatment with endoglycosidase H. (b) Sequencing profile of a purified monosialylated biantennary complex-type oligosaccharide derived from glycopeptide T6-1 (Asn 395) by sequential exoglycosidase treatment using FACE®. Lane 1, 4-[beta]-mannosyl-6-[alpha]-fucosyl chitobiose standard and 4-[beta]-mannosyl chitobiose standard; lane 2, purified oligosaccharide from T6-1 treated with sialidase, GALase III, HEXase III, and MANase II; lane 3, purified oligosaccharide from T6-1 treated with sialidase, GALase III, and HEXase III; lane 4, purified oligosaccharide from T6-1 treated with sialidase, GALase III; lane 5, purified oligosaccharide from T6-1 treated with sialidase; lane 6, purified oligosaccharide from T6-1; lane 7, oligosaccharide pool from T6-1; lane 8, oligo ladder standard (glucose polymers). (c) Sequencing profile of a purified high-mannose-type oligosaccharide containing six mannosyl residues derived from T7 (Asn 563) by sequential exoglycosidase treatment using FACE®. Lane 1, 4-[beta]-mannosyl-6-[alpha]-fucosyl chitobiose standard and 4-[beta]-mannosyl chitobiose standard; lane 2, purified oligosaccharide from T7 treated with sialidase, GALase III, HEXase III, and MANase II; lane 3, purified oligosaccharide from T7 treated with sialidase, GALase III, and HEXase III; lane 4, purified oligosaccharide from T7 treated with sialidase and GALase III; lane 5, purified oligosaccharide from T7 treated with sialidase; lane 6, purified oligosaccharide from T7; lane 7, oligosaccharide pool from T7; lane 8, oligo ladder standard (glucose polymers). |
|
Sequencing of oligosaccharides
ANTS labeled oligosaccharides were separated in a preparative electrophoresis gel, single oligosaccharide bands were excised, and the oligosaccharides extracted from the gel slices. The main oligosaccharides of each glycopeptide that were present in higher amounts (four complex- and two hybrid-/high-mannose-type oligosaccharides from T2; three complex-type oligosaccharides from both T5 and T6-1, respectively; four hybrid-/high-mannose-type oligosaccharides from T6-4 and two hybrid/high-mannose-type oligosaccharides from T7) were extracted for sequencing experiments. The remaining oligosaccharides (1-16% of the total oligosaccharide pool of each N-glycosylation site) were structures that were present in insufficient quantities for sequencing. Then, the isolated oligosaccharides were analyzed using a sequential, highly specific exoglycosidase treatment on the ANTS-labeled oligosaccharides which were subsequently separated by FACE®. The release of particular monosaccharide residues at the exposed nonreducing end can be deduced from the increased mobility of the remaining ANTS-labeled oligosaccharide in the gel relative to standards. (The migration values for oligosaccharides are defined by a degree of polymerization (DP) value.)
The four complex- and two main hybrid-/high-mannose-type oligosaccharides of glycopeptide T2 that were present in higher amounts were sequenced. The mobilities of the four complex-type oligosaccharides and one of the hybrid-/high-mannose-type oligosaccharides were increased by treatment with sialidase. Their mobilities increased by an amount equivalent the removal of one or two sialic acids, respectively. Therefore, glycopeptide T2 possesses three mono- and one disialylated complex-type oligosaccharides as well as mono- and nonsialylated hybrid-type oligosaccharides but no high-mannose-type structures. One of the monosialylated complex-type structures attached to T2 could not be enzymatically trimmed to the trimannosyl core. After treatment with a combination of sialidase and [beta]-galactosidase this oligosaccharide was additionally treated with [alpha]-galactosidase, resulting in an additional increase in its electrophoretic mobility. This result strongly suggests the presence of Gal[alpha](1,3)Gal in at least one branch of this oligosaccharide. However, with the set of glycosidases used, it was still not possible, to cleave this structure to the trimannosyl core structure. The exact structure remains to be determined. The complete structures of all sequenced oligosaccharides are given in Table III.
Table III
The most prevalent structures present on each of the five N-glycosylation sites of the LPD5H4 antibody were determined using sequential exoglycosidase treatment and FACE®. Values are calculated as the percentage of oligosaccharide present in the oligosaccharide pool obtained from that particular N-glycosylation site. R1, 6-[alpha]-fucosyl chitobiose; R2, chitobiose. #, Suggested structure.
A disialylated complex-type oligosaccharide is the most prevalent oligosaccharide attached to glycopeptide T5. As in glycopeptide T2, additional monosialylated complex-type oligosaccharides and the same complex-type oligosaccharide structure terminating in Gal[alpha](1,3)Gal are present. Glycopeptide T6-1 has a similar oligosaccharide pattern to that of glycopeptide T5. Over 80% of the oligosaccharides of glycopeptide T6-1 are disialylated, complex-type oligosaccharides. Furthermore, a monosialylated complex-type structure and a Gal[alpha](1,3)Gal-bearing structure were determined to be present in minor amounts (Table III). The sequence analysis of the monosialylated complex oligosaccharide derived from glycopeptide T6-1 is shown in figure 5b.
The oligosaccharides of glycopeptide T6-4 did not shift their electrophoretic mobilities in the gel after treatment with sialidase, galactosidase, and hexosaminidase (Figure 5c, lanes 3-6). A shift in the mobilities of these oligosaccharides equivalent to the release of three, four, five, or six mannose residues was observed after treatment with mannosidase (Figure 5c, lane 2). We concluded that glycopeptide T6-4 possesses almost equal amounts (22.5%:24.9%:25.5% of the total oligosaccharide pool at this glycosylation site) of high mannose-type oligosaccharides with four, five, and six mannosyl residues, respectively. A high-mannose-type oligosaccharide with seven mannosyl residues comprised 11.1% of the oligosaccharide pool of glycopeptide T6-4 (Table IV). Glycopeptide T7 possesses only high mannose-type oligosaccharides with five and six mannosyl residues in higher amounts. The sequence analysis of a six mannosyl residue containing high-mannose-type oligosaccharide derived from glycopeptide T7 is shown in Figure 5c.
Table IV
Discussion
We have described the structures of the most abundant oligosaccharide structures present at each of the five N-glycosylation sites of a human hybridoma IgM. Oligosaccharide sequencing analysis was done by sequential exoglycosidase treatment using FACE®. In addition, we used lectin blotting and MALDI/TOF mass spectrometry to identify single glycosylation sites and characterize their oligosaccharides.
Each of the five glycosylation sites of this antibody is occupied by a different set of oligosaccharides. The most heterogeneous mixture of oligosaccharides is attached to glycosylation site Asn 171 in the Fab fragment of the antibody. At this site 62% of the oligosaccharides are monosialylated structures. All complex-type oligosaccharides have a fucosylated trimannnosyl core structure. The N-glycosylation site at position Asn 171 is the only site that possesses monosialylated and nonsialylated, nonfucosylated hybrid-type oligosaccharides. They represent only a small portion of the complete oligosaccharide pool of this antibody. This is the first report for hybrid-type oligosaccharides attached to human monoclonal IgM. Hybrid-type oligosaccharides have not been reported on normal or pathological human serum IgM, but they have been reported on mouse monoclonal IgM (Brenckle and Kornfeld, 1980; Anderson et al., 1985).
N-Glycosylation sites at Asn 332 and Asn 395 of the human monoclonal antibody LPD5H4 carry mainly disialylated, biantennary, complex-type oligosaccharides, whereas the corresponding glycosylation sites in murine monoclonal antibodies contain triantennary sialylated oligosaccharides as described by Anderson et al. (1985). The presence of triantennary oligosaccharides was also reported for normal and pathological human serum IgM (Cahour et al., 1983). No triantennary oligosaccharides were found on the human monoclonal LPD5H4 antibody at any glycosylation site.
The presence of high-mannose-type oligosaccharides on N-glycosylation site Asn 402 of the LPD5H4 antibody is consistent with previous results for IgM obtained from patients with Waldenström's disease (Hickman et al., 1972). The N-glycosylation site Asn 563 of the LPD5H4 antibody was dominated by two high-mannose-type oligosaccharides with five and six mannosyl residues attached, respectively. High-mannose-type oligosaccharides were previously reported at this glycosylation site 563 for pathological IgM (Wormald et al., 1991). In contrast, Monica et al. (1995) reported complex sialylated oligosaccharides at Asn 563 on a human monoclonal IgM, a result which we did not obtain for the human monoclonal LPD5H4 antibody in the present study.
Normal human serum IgM, pathological serum IgM, and the monoclonal human LPD5H4 antibody also differ in sialylation and in linkage of galactose attached in these oligosaccharides. Lectin blotting with the lectin BS-I B4, which binds specifically to Gal[alpha](1,3)Gal, revealed the presence of Gal[alpha](1-3)Gal in the oligosaccharides of the LPD5H4 antibody. FACE® sequencing analysis also strongly suggested the presence of a complex-type oligosaccharide terminating in Gal[alpha](1,3)Gal since its electrophoretic mobility in the gel increased after treatment with [alpha]-galactosidase. The complete structure of this oligosaccharide could not be elucidated by sequential treatment with the set of exoglycosidases available to us. This particular oligosaccharide is present at N-glycosylation sites Asn 171, Asn 332, and Asn 395. Oligosaccharides containing Gal[alpha](1,3)Gal structures have not been reported as components of human monoclonal IgM or human serum IgM (Maiorella et al., 1993; Monica et al., 1995). The antigenic properties of Gal[alpha](1,3)Gal structures may stimulate an anti-Gal[alpha](1,3)Gal immune response in humans. Since 1% of the human serum IgG is already directed against the Gal[alpha](1,3)Gal epitope, changes in the pharmacokinetic properties of the Gal[alpha](1,3)Gal-expressing antibody are most likely to occur by interaction with these anti-Gal[alpha](1,3)Gal antibodies (Hamadeh et al., 1992).
About 11% of all monosaccharides on the LPD5H4 antibody are NeuGc. Glycoproteins in adult humans do not normally contain NeuGc, which is an oncofetal antigen (Muchmore et al., 1989). However, Monica et al. (1995) found NeuGc on one other human monoclonal hybridoma but in a much lesser amount. The antigenic properties of NeuGc in humans may stimulate an anti-NeuGc immune response dependent on the dosage (Noguchi et al., 1995). This immune response is likely to influence the in vivo effect of the antibody thereby limiting its therapeutic value. Furthermore, terminal NeuGc on recombinant proteins also seems to be correlated with a more rapid removal of the molecule from the circulatory system (Flesher et al., 1995). Besides the expected [alpha](2,6)-linked sialic acids we have found [alpha](2,3)-linked sialic acid as determined by binding of Maakia amurensis agglutinin. This result contrasts with reports from Anderson et al. (1985), Ohbayashi et al. (1989), and Monica et al. (1995) who did not detect the presence of [alpha](2,3)-linked sialic acids in human serum IgM, in IgM from Waldenström's patients, or in human monoclonal IgM. Table V summarizes the similarities and differences of normal and pathological human IgM, mouse monoclonal IgM and human monoclonal IgM (LPD5H4 antibody).
The LPD5H4 antibody has a potential N-glycosylation site in its CDR 2 at position Asn 56 of the heavy chain as determined by amino acid sequence analysis. This potential N-glycosylation site is in addition to the five conserved N-glycosylation sites in the constant domains of the heavy chain, all five of which were N-glycosylated. No oligosaccharides were found at Asn 56 during this analysis. The presence of the consensus sequence Asn-Lys-Ser does not guarantee glycosylation. The occupation of a potential glycosylation site with oligosaccharides may also depend on the site's position within the protein and its conformation, the host cell type used for the expression, and its physiological status (Pollack et al., 1983; Jenkins et al., 1996).
The analysis of the oligosaccharide side chains of the human monoclonal LPD5H4 antibody has shown its glycosylation pattern differs from those reported for normal and pathological human serum IgM as well as human monoclonal IgM (Cahour et al., 1983; Ohbayashi et al., 1989; Monica et al., 1993, 1995). The presence of NeuGc and the Gal[alpha](1,3)Gal epitope in the oligosaccharides of this antibody, which do not occur in human glycoproteins, suggests a dominance of the mouse glycosylation machinery in the human-human-mouse heterohybridoma LPD5H4. The presence of NeuGc and the Gal[alpha](1,3)Gal epitope in the oligosaccharides of this antibody may lead to an anti-NeuGc and anti-Gal[alpha](1,3)Gal immune response. Furthermore, the high proportion of NeuGc is likely to contribute to a significantly different clearance rate from the circulatory system as described for other recombinant glycoproteins (Flesher et al., 1995). Therefore, these two antigenic determinants for humans may lead to limitations in the therapeutic use of the LPD5H4 antibody.
Materials and methods
Purification of monoclonal antibody LPD5H4
A sample of the purified human monoclonal IgM LPD5H4 was obtained from the Department of Medical Immunology, Humboldt University, Berlin, Germany. The antibody was expressed in the human-human-mouse heterohybridoma cell line LPD5H4 (Schoenherr et al., 1996; Seifert et al., 1996). The antibody was purified by a two-step procedure using hydrophobic interaction chromatography on phenyl-Superose and size exclusion over Superose 12 (Pharmacia, Sweden) as described by Roggenbuck et al. (1994).
Cloning and sequencing of heavy and light chain variable region genes
Total RNA was prepared from hybridoma cells following standard protocols (Sambrook et al., 1989). For first-strand cDNA synthesis, total RNA (5 µg) was annealed to a random hexamer nucleotide mix (Boehringer GmbH, Germany) and extended using the AMV reverse transcriptase kit (Gibco BRL, Gaithersburg, MD). The gene coding region for the immunoglobulin variable region of the heavy chain (VH) was amplified using primer pairs with the 5[prime] framework 1 region (FR1) (5[prime]-GAGGTGCAGCTGCAGGAGTCTGG-3[prime]) and a JH-region stretch (5[prime]-CTTGGTGGA(AG)GAGACGGTGACC-3[prime]; Marks et al., 1991). The gene coding region for the immunoglobulin variable region of the light chain (VL) was amplified using primer pairs with 5` framework 1 region (5[prime]-GGTCCTGGGCCCAGTCTGTC-3[prime]; Küppers et al., 1995) and a J[lambda] stretch (5[prime]-GCCACTTACCTAGGACGGTGAC-3[prime]; Songsviali et al., 1990).
Table V
| Glycosylation | Human serum IgM | Pathological human serum IgM | Human monoclonal IgM (LPD5H4) | Mouse monoclonal IgM (MOPC 104E) |
| Site Asn 171 (mouse Asn 158) | Biantennary complex-type oligosaccharides | Biantennary complex-type oligosaccharides | Biantennary complex- and hybrid-type oligosaccharides | Biantennary complex-type oligosaccharides |
| Sites Asn 332 and Asn 395 (mouse Asn 333 and Asn 364) | Biantennary/triantennary complex-type oligosaccharides | Predominant multiantennary complex-type oligosaccharides | Biantennary complex-type oligosaccharides | Biantennary/triantennary complex- and high-mannose-type oligosaccharides |
| Site Asn 402 | High-mannose-type oligosaccharides | High-mannose-type oligosaccharides | High-mannose-type oligosaccharides | High-mannose-type oligosaccharides |
| Site Asn 563 | High-mannose-type oligosaccharides | High-mannose-type oligosaccharides | High-mannose-type oligosaccharides | Complex- and high-mannose-type oligosaccharides |
| Gal[alpha](1,3)Gal | No | No | Yes | Yes |
| NeuGc:NeuAc ratio | 0:100 | 0:100 | 98:2 | 35:65 |
| [alpha](2,3)-linked sialic acid | No | No | Yes | No |
The cDNA first-strand reaction sample (5 µl) was added to the polymerase chain reaction (PCR) mix containing 12.5 µM of each primer, 5 µl buffer (Promega, Madison, WI), and dNTP mix (200 µM each). One unit of Taq polymerase (Promega, Madison, WI) was added during the third step of the first PCR cycle. PCR was carried out in a thermal cycler running 35 cycles (60 s at 96°C, 60 s at 50°C, and 60 s at 72°C). Finally, after 35 cycles, reaction mixtures were incubated at 72°C for 10 min to ensure full extension of all PCR products. Fragments of the expected size were cloned into pUC 18 and sequenced from multiple clones in both directions following the method of Sanger et al. (1977).
SDS-PAGE and lectin blots
SDS-PAGE, protein transfer onto nitrocellulose, and lectin blots were conducted as described previously (Leibiger et al., 1995). In brief, nitrocellulose filters were blocked with PBS/0.1% Tween 20 and then incubated for 1 h with 1 µg/ml anti-human heavy chain (µ) HRP-labeled and anti-human light chain ([lambda]) HRP-labeled or 1 µg/ml biotin-labeled lectins in PBS/0.1% Tween 20 with additional 1 mM CaCl2, 1 mM MgCl2, and 1 mM MnCl2 (Sigma, St. Louis, MO). The filters were washed three times with 0.15 M PBS/0.1% Tween 20 and incubated with 0.2 µg/ml HRP-labeled streptavidin (Sigma, St. Louis, MO) for 30 min. The color reaction was developed as described by Kießig et al. (1992). Lectins used for lectin blotting and their corresponding specificities are: Sambucus nigra agglutinin (SNA) which binds terminal NeuAc/NeuGc [alpha](2,6)-linked to galactose (Shibuya et al., 1987); Datura stamonium agglutinin (DSA) which binds galactose [beta](1,4)-linked to N-acetylglucosamine (Crowley et al., 1984); Galanthus nivalis agglutinin (GNA); which binds terminal mannosyl residues present in high mannose-type or hybrid-type N-glycans (Shibuya et al., 1988); Grifonia simplicifolia lectin I (BS-I) which binds galactosyl residues [alpha](1,3)-linked to galactose (Hayes et al., 1974); Maakia amurensis agglutinin which binds terminal [alpha](2,3) NeuAc residues (Wang and Cummings, 1988). Standard glycoproteins (Boehringer GmbH, Germany) with known glycosylation were used as controls.
Monosaccharide composition analysis (GC)
Methylglycosides were prepared by methanolysis in anhydrous methanolic HCl (1 M, 80°C for 16 h), followed by N-acetylation with pyridine and acetic anhydride for 6 h at room temperature. The methylglycosides were then treated with trimethylsilane reagent (Tri-Sil, Pierce Chemical Co., Rockford, IL) to form TMS derivatives. These procedures were accomplished as described previously (Merkle and Poppe, 1994). GC analysis was performed on a 30-meter DB1 fused silica capillary column using a Hewlett-Packard 5890 GC. The GC temperature program was: initial temperature of 110°C, hold for 0.5 min then increase at 4°C/min to 260°C, then increase at 20°C/min to 300°C, hold for 5 min.
Desialylation and separation of sialic acids by HPAEC/PAD
Aliquots of IgM were diluted in water. Acetic acid was added to a final concentration of 2 M. The sample was incubated at 100°C for 30 min and then immediately frozen and lyophilized. Separation of the sialic acids was carried out at ambient temperature on a Dionex Bio-LC system (Dionex, Sunnyvale, CA) with a CarboPac PA-1 column (4.9 × 250 mm). The lyophilized samples were dissolved in water, and 0.2-1 nmol was injected. The column had been pre-equilibrated in 100 mM sodium hydroxide and 0.1 M sodium acetate. The flow rate was 1 ml/min. A standard mixture containing equimolar amounts of NeuAc and NeuGc was used to calibrate the HPAEC column. The samples were eluted with 100 mM sodium hydroxide and 0.15 M sodium acetate over 25 min.
Protease cleavage of the LPD5H4 antibody
Before protease treatment the antibody was reduced and S-carboxymethylated using published procedures (Fu and Van Halbeek, 1992). Trypsin (Boehringer GmbH, Germany) was added to aliquots of reduced and carboxymethylated IgM at an enzyme to substrate ratio of 1:50 at 0 and 2 h of incubation. Samples were incubated for 6 h at 37°C or for 24 h at ambient temperature. For digestion with endoproteinase Glu-C (Boehringer GmbH, Germany), the enzyme was added to aliquots of glycopeptides at an enzyme to substrate ratio of 1:20 at 0 and 24 h of incubation. Samples were incubated for 72 h at ambient temperature. The reactions were stopped by placing the samples in a boiling water bath for 5 min.
Separation of peptides by reversed-phase HPLC (RP-HPLC)
The RP-HPLC was run using a Bio-Rad Model 2800 HPLC system equipped with a Bio-Rad Model 1706 variable wave length detector and a C18 column (2.0 × 250 mm, Phenomenex, Torrance, CA). Absorbance was measured at 214 nm, and the flow rate was 0.2 ml/min. The RP-HPLC gradient was based on two solvents. Solvent A was 0.1% TFA, and solvent B was 0.085% TFA in 80% acetonitrile. The column was pre-equilibrated with 100% solvent A. The gradient consisted of 0-1.5 min 0% B; 1.5-3 min 0-15% B; 3-90 min 15-55% B; 90-95 min 55-90% B. Samples were mixed 1:1 with 0.1% TFA and centrifuged before injection.
Isolation of glycopeptides by Concanavalin A-Sepharose chromatography
Con A-Sepharose (1 ml) was used for 20 nmol of tryptic-digested antibody. The samples were dissolved in 0.5 ml of Tris-buffered saline (TBS/NaN3, 0.01 M Tris, pH 8.0, 0.15 M NaCl, 1 mM CaCl2, 1 mM MgCl2, and 0.02% NaN3) and applied to a Con A-Sepharose column pre-equilibrated in the same buffer. The column was washed with TBS/NaN3 and unbound material eluted. The bound material was eluted in two fractions. The first fraction was eluted by application of six times the column bed volume of TBS/NaN3 buffer containing 200 mM [alpha]-methylglucopyranoside (fraction I). Then about one-half the column bed volume of TBS/NaN3 buffer containing 500 mM [alpha]-methylmannopyranoside was drawn into the column at which time the bottom of the column was capped, and the column was immersed in a boiling water bath for 10 min. The column was eluted then with 5 bed volumes of TBS/NaN3 buffer containing 500 mM [alpha]-methylmannopyranoside that had been equilibrated to 60°C (fraction II). Fractions I and II were dialyzed against 0.05 M ammonium bicarbonate buffer (pH 7.8) and lyophilized.
Treatment of the antibody and the glycopeptides with endoglycosidases
The antibody or tryptic glycopeptide samples were dried in a centrifugal vacuum evaporator and resuspended in 0.05 M ammonium bicarbonate buffer, pH 7.8. Two mU of N-glycanase (Genzyme, Cambridge, MA) were added, and the samples were incubated at 37°C for 16 h. For treatment with endoglycosidase H (Boehringer Mannheim, Germany), the samples were diluted in 0.1 M sodium phosphate, pH 5.5, and incubated with 2.5 mU endoglycosidase H at 37°C for 16 h. Oligosaccharides were isolated from the protein or peptide by adding three volumes of cold 100% ethanol. The samples were incubated on ice for 10 min and centrifuged for 5 min to pellet the protein. The supernatants, which contained the oligosaccharides, were dried to a translucent pellet.
Fluorophore-assisted carbohydrate electrophoresis (FACE[)
Following the release of the oligosaccharides from the glycoprotein or the glycopeptides, the mixture of oligosaccharides was labeled with ANTS using the FACE® N-linked oligosaccharide profiling kit. The derivatized oligosaccharides were separated on an oligosaccharide profiling gel (Glyko, Novato, CA). Samples were electrophoresed in 0.5 mm thick gels with 15 mA at 4-8°C for 90 min. Imaging of the gel was performed using the FACE® imaging system.
For preparative electrophoresis 1.0 mm thick gels (20 mA) were used. The oligosaccharide bands were visualized using a long-wave UV light box transilluminator. The bands were excised carefully from the gel and transferred into separate microcentrifuge tubes. The gel slices were soaked in a minimum volume of 100% ethanol for 30 min. After removing the ethanol the gel slices were soaked in a minimum volume of water, vortexed gently, and held at 4°C for 12-16 h. Then supernatants were removed and dried in a centrifugal vacuum evaporator.
Sequencing of N-linked oligosaccharides by FACE[
Purified, labeled oligosaccharides (50-100 pmol) were diluted in 0.05 M sodium phosphate, pH 6.0, and treated with several exoglycosidase combinations. Exoglycosidases (2 µl each) were added to each sample. Exoglycosidases used were: sialidase from Arthrobacter ureafaciens (Oxford GlycoSystems, Abingdon, UK) that releases both NeuAc and NeuGc, GALase III that releases [beta](1,4)-linked galactose, HEXase III that releases [beta]-GalNAc and [beta]-GlcNAc, MANase II that releases [alpha]-linked Man, and [alpha]-GALase that releases [alpha](1,3)-linked galactose (Glyko, Novato, CA). The exoglycosidase combinations were: reaction 1, no enzyme; reaction 2, sialidase; reaction 3, sialidase, GALase III; reaction 4, sialidase, GALase III, and Hexase III; reaction 5, sialidase, GALase III, Hexase III, and MANase II. The samples were incubated at 37°C for 12-16 h, dried in a centrifugal vacuum evaporator, resuspended in loading buffer (Glyko, Novato, CA) and separated by FACE®.
The migration pattern and the known specificity of the exoglycosidases used to digest the oligosaccharide in combination with monosaccharide composition data allowed the oligosaccharide sequence to be determined. The migration values were defined as degree of polymerization value (DP).
Identification of glycopeptides by MALDI/TOF-mass spectrometry
Mass spectra were obtained using a Hewlett-Packard LDI 1700XP time-of-flight mass spectrometer operating at an accelerating voltage of 30 kV, an extractor voltage of 9 kV, and a pressure of approximately 2.0 × 10-6 Torr. The mass spectrometer was calibrated using an equimolar mixture of oxytocin, Arg-vasopressin, insulin [beta]-chain, and insulin. Peptides or glycopeptides obtained by chromatography with a Con A-Sepharose column and RP-HPLC were treated with endo- and exoglycosidases. Sinapinic acid or 2,5-dihydroxybenzoic acid solution were added alternately to these samples. One microliter of this solution was dried on the flat surface of a steel probe tip. Samples were ionized from the probe tip with a nitrogen laser ([lambda] = 337 nm) with a pulse width of 3 ns. Spectra were recorded over a mass-to-charge (m/z) range of 300-10,000. All mass spectra were averaged data of 50-70 laser shots.
Peptide identification by amino acid analysis
Peptides collected from reversed-phase HPLC were sequenced in the Macromolecular Structure Facility at the Department of Biochemistry, Michigan State University (East Lansing, MI) using a ABI 494 Procise protein/peptide sequencer.
Acknowledgments
We thank P. Siegel for assistance in sequencing of the antibody variable region genes, C. Bergmann for assistance in glycosylation site mapping, R. Merkle for assistance in oligosaccharide sequencing, E. Kannenberg for helpful comments and discussion, and R. Nuri and A. Dunn for their assistance in the preparation of the manuscript. This work was supported in part by Deutsche Forschungsgemeinschaft Grant LE 1000/1-1 and by National Institutes of Health NIH Grant 5P41RR05351.
Abbreviations
ANTS, 8-aminonaphthalene-1,3,6-trisulphonic acid; BS-I B4, Grifonia simplicifolia agglutinin I B4; CDR, complementary determining region; Con A, concanavalin A; DHB, 2,5-dihydroxybenzoic acid; DP, degree of polymerization; DSA, Datura stramonium agglutinin; FACE®, fluorophore-assisted carbohydrate electrophoresis; FR, framework region; GC, gas chromatography; HPAEC/PAD, high-pH anion-exchange chromatography with pulsed amperometric detection; HPLC, high-performance liquid chromatography; HRP, horseradish peroxidase; IgM, immunoglobulin M; MAA, Maakia amurensis agglutinin; MALDI/TOF, matrix-assisted laser desorption/ionization-time-of-flight; MS, mass spectrometry; PBS, phosphate-buffered saline; PCR, polymerase chain reaction; RNA, ribonucleic acid; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; SNA, Sambucus nigra agglutinin; TBS, Tris-buffered saline; TMS, trimethylsilane; VH, variable region of an immunoglobulin heavy chain; VL, variable region of an immunoglobulin light chain
References
2To whom correspondence should be addressed at: Göteborg University, Department of Medical Biochemistry, Medicinaregatan 9, S-41390 Gothenburg, Sweden
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 2 Apr 1998
Copyright© Oxford University Press, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. N. Arnold, M. R. Wormald, D. M. Suter, C. M. Radcliffe, D. J. Harvey, R. A. Dwek, P. M. Rudd, and R. B. Sim Human Serum IgM Glycosylation: IDENTIFICATION OF GLYCOFORMS THAT CAN BIND TO MANNAN-BINDING LECTIN J. Biol. Chem., August 12, 2005; 280(32): 29080 - 29087. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Fukuta, R. Abe, T. Yokomatsu, F. Omae, M. Asanagi, and T. Makino Control of Bisecting GlcNAc Addition to N-Linked Sugar Chains J. Biol. Chem., July 28, 2000; 275(31): 23456 - 23461. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||