Variable region heavy chain glycosylation determines the anticoagulant activity of a factor VIII antibody

BACKGROUND

N-glycosylation occurs in the variable region of about 10% of antibodies but the role of carbohydrate at this location is still poorly understood.

OBJECTIVES

We investigated the function of N-glycosylation in the variable region of the heavy chain of a human monoclonal antibody, mAb-LE2E9, that partially inhibits factor VIII (FVIII) activity during coagulation.

METHODS AND RESULTS

Enzymatic deglycosylation indicated that the oligosaccharides do not determine the affinity of the antibody but enhance its FVIII neutralizing activity. A mutant antibody lacking the N-glycosylation site in the variable region of the heavy chain inhibited FVIII activity by up to 40%, while inhibition by the native antibody was 80%. To evaluate the physiological effect of such a FVIII inhibition, we investigated the ability of the mutant antibody devoid of N-glycosylation in the variable region to prevent thrombosis in mice with a strong prothombotic phenotype resulting from a type II deficiency mutation in the heparin binding site of antithrombin. Despite its moderate inhibition of FVIII activity, the mutant antibody significantly prevented thrombosis in treated animals. We also carried out glycan analysis of native and mutant antibodies.

CONCLUSIONS

Modification of glycosylation in the variable region of antibodies contributes to the diversity of FVIII type II inhibition possibly by steric hindrance of the active site of FVIII by glycans, and may provide a novel strategy to modulate the functional activity of therapeutic antibodies.

Complement regulation at the molecular level: the structure of decay-accelerating factor

The human complement regulator CD55 is a key molecule protecting self-cells from complement-mediated lysis. X-ray diffraction and analytical ultracentrifugation data reveal a rod-like arrangement of four short consensus repeat (SCR) domains in both the crystal and solution. The stalk linking the four SCR domains to the glycosylphosphatidylinositol anchor is extended by the addition of 11 highly charged O-glycans and positions the domains an estimated 177 A above the membrane. Mutation mapping and hydrophobic potential analysis suggest that the interaction with the convertase, and thus complement regulation, depends on the burial of a hydrophobic patch centered on the linker between SCR domains 2 and 3.

Matrix remodelling enzymes, the protease cascade and glycosylation

Glycosylation influences the specific activities of serine proteases including tissue-type plasminogen activator and plasmin which act together in a ternary complex with fibrin. Serine proteases and matrix metalloproteinases (MMPs), including gelatinase B, participate in a protease cascade to remodel the extracellular matrix. In addition to the recognition and targeting functions of carbohydrates and the fact that they confer protease resistance on glycoproteins, oligosaccharides may extend particular protein domains of matrix remodelling enzymes and fine-control their activities within the context of the extracellular matrix. For example, the sialic acids of gelatinase B influence the catalytic activity of this enzyme in a complex with the tissue inhibitor of metalloproteinases-1 (TIMP-1).

Antibodies inhibit prion propagation and clear cell cultures of prion infectivity

Prions are the transmissible pathogenic agents responsible for diseases such as scrapie and bovine spongiform encephalopathy. In the favoured model of prion replication, direct interaction between the pathogenic prion protein (PrPSc) template and endogenous cellular prion protein (PrPC) is proposed to drive the formation of nascent infectious prions. Reagents specifically binding either prion-protein conformer may interrupt prion production by inhibiting this interaction. We examined the ability of several recombinant antibody antigen-binding fragments (Fabs) to inhibit prion propagation in cultured mouse neuroblastoma cells (ScN2a) infected with PrPSc. Here we show that antibodies binding cell-surface PrPC inhibit PrPSc formation in a dose-dependent manner. In cells treated with the most potent antibody, Fab D18, prion replication is abolished and pre-existing PrPSc is rapidly cleared, suggesting that this antibody may cure established infection. The potent activity of Fab D18 is associated with its ability to better recognize the total population of PrPC molecules on the cell surface, and with the location of its epitope on PrPC. Our observations support the use of antibodies in the prevention and treatment of prion diseases and identify a region of PrPC for drug targeting.

A family of novel, acidic N-glycans in Bowes melanoma tissue plasminogen activator have L2/HNK-1-bearing antennae, many with sulfation of the fucosylated chitobiose core

A family of about 20 novel acidic bi- and tri-antennary N-glycans, amounting to almost half those expressed on Bowes melanoma tissue-plasminogen activator (t-PA) were found to possess Galbeta1-->4GlcNAcbeta1-->, sulfated and sialylated GalNAcbeta1-->4GlcNAcbeta1--> or sulfated GlcAbeta1--> 3Galbeta1-->4GlcNAcbeta1--> antennae, of which those containing sulfated GlcA, depicting the L2/HNK-1 carbohydrate epitope, were preferentially located on the 6 arm. A proportion of the glycans were highly charged, because of multiple and variously distributed sulfation, some of which was located on the fucosylated chitobiose core. Multiple expression of the L2/HNK-1 epitope on a single glycan was observed. The most abundant compound was a biantennary glycan carrying sulfated GlcA on the 6-branched antenna and an alpha2-->6 sialylated GalNAc on the other. The N-glycosylation sequon containing Asn448, which is known to express all of the sulfate-carrying N-glycans contains, unusually, an arginine residue. An electrostatic interaction between this cationic amino acid and the core-sulfate group of the N-glycan is proposed to reduce mobility of the carbohydrate in the region of the t-PA active site. Because of the 'brain-type' nature of the N-glycans described in this neuro-ectodermal cell line, the possibility of neural t-PA interacting with the L2/HNK-1-recognizing molecule, laminin, of the central nervous system extracellular matrix is discussed.

Prion glycoprotein: structure, dynamics, and roles for the sugars

The prion protein contains two N-linked glycosylation sites and a glycosylphosphatidylinositol (GPI) anchor. The large size of the N-linked sugars, together with their dynamic properties, enables them to shield two orthogonal faces of the protein almost completely. Thus, the sugars can protect large regions of the protein surface from proteases and from nonspecific protein-protein interactions. Immunoprecipitation of prion protein with calnexin suggests that in the ER the oligosaccharides may provide a route for protein folding via the calnexin pathway. Major questions relate to the relevance of the glycoform distribution (as defined by glycan site occupancy) to strain type and disease transmission. Glycan analysis has shown that prion protein contains at least 52 different sugars, that these consist of a subset of brain sugars, and that there is site specific glycan processing. PrP(Sc) from the brains of Syrian hamsters contains the same set of glycans as PrP(C), but a higher proportion of tri- and tetra-antennary sugars. This may be attributed to a decrease in the activity of GnTIII. The GPI anchor, which is modified with sialic acid, may allow the prion protein to be mobile in the lipid bilayer. Potentially, this provides a possible means for translocating the prions from one cell to another.

O-glycan analysis of natural human neutrophil gelatinase B using a combination of normal phase-HPLC and online tandem mass spectrometry: implications for the domain organization of the enzyme

Gelatinase B is a matrix metalloproteinase (MMP-9) expressed under strict control by many cell types including neutrophils, monocytes, macrophages, and tumor cells. MMP-9 is a key mediator in the physiological maintenance of the extracellular matrix both in tissue remodeling and development, while uncontrolled enzyme activity contributes to pathologies such as cancer and inflammation. Neutrophils release MMP-9 from granules in response to IL-8 stimulation. Human MMP-9 has three potential N-linked glycosylation sites and contains a Ser/Pro/Thr rich domain, known as the type V collagen-like domain, which is expected to be heavily O-glycosylated. Indeed, approximately 85% of the total sugars on human neutrophil MMP-9 are O-linked. This paper presents the detailed analysis of picomole amounts of these O-glycans using a novel HPLC-based strategy for O-glycan analysis that provides linkage and arm specific information in addition to monosaccharide sequence. The initial structural assignments were confirmed using HPLC with online MS/MS fragmentation analysis. Twelve sugars were identified that contained from two to nine monosaccharide residues. Most of these contained type 2 core structures with Galbeta1-4GlcNAc (N-acetyl lactosamine) extensions, with or without sialic acid or fucose. The O-glycans were modeled using the oligosaccharide structural database. On the basis of the structure of gelatinase A (MMP-2), a model of MMP-9 suggests that the type V collagen-like domain in gelatinase B is located on a loop remote from the active site. Fourteen potential O-glycosylation sites are multiply presented on this loop of 52 amino acids. Many of the O-glycans identified contain terminal galactose residues that may provide recognition epitopes. Importantly, heavy glycosylation of this loop region, absent in gelatinase A, has considerable implications for the domain organization of MMP-9.

Mutations at critical N-glycosylation sites reduce tyrosinase activity by altering folding and quality control

Tyrosinase is a copper-containing enzyme that regulates melanin biosynthesis in mammals. Mutations at a single N-glycosylation sequon of tyrosinase have been reported to be responsible for oculocutaneous albinism type IA in humans, characterized by inactive tyrosinase and the total absence of pigmentation. To probe the role that each N-glycosylation site plays in the synthesis of biologically active tyrosinase, we analyzed the calnexin mediated folding of tyrosinase N-glycosylation mutants. We have determined that four of the six potential N-glycosylation sites, including that associated with albinism, are occupied. Analysis of the folding pathway and activity of 15 tyrosinase mutants lacking one or more of the occupied N-glycosylation sites shows that glycans at any two N-glycosylation sites are sufficient to interact with calnexin and give partial activity, but a specific pair of sites (Asn(86) and Asn(371)) is required for full activity. The mutants with less than two N-glycosylation sites do not interact with calnexin and show a complete absence of enzyme activity. Copper analysis of selected mutants suggests that the observed partial activity is due to two populations with differential copper content. By correlating the degree of folding with the activity of tyrosinase, we propose a local folding mechanism for tyrosinase that can explain the mechanism of inactivation of tyrosinase N-glycosylation mutants found in certain pigmentation disorders.

Hybrid and complex glycans are linked to the conserved N-glycosylation site of the third eight-cysteine domain of LTBP-1 in insect cells

Covalent association of LTBP-1 (latent TGF-beta binding protein-1) to latent TGF-beta is mediated by the third eight-cysteine (also referred to as TB) module of LTBP-1, a domain designated as CR3. Spodoptera frugiperda (Sf9) cells have proved a suitable cell system in which to study this association and to produce recombinant CR3, and we show here that another lepidopteran cell line, Trichoplusia niTN-5B1-4 (High-Five) cells, allows the recovery of large amounts of functional recombinant CR3. CR3 contains an N-glycosylation site, which is conserved in all forms of LTBP known to date. When we examined the status of this N-glycosylation using MALDI-TOF mass spectrometry and enzymatic analysis, we found that CR3 is one of the rare recombinant peptides modified with complex glycans in insect cells. Sf9 cells mainly processed the fucosylated paucomannosidic structure (GlcNAc)(2)(Mannose)(3)Fucose, although hybrid and complex N-glycosylations were also detected. In High-Five cells, the peptide was found to be modified with a wide variety of hybrid and complex sugars in addition to paucomanosidic oligosaccharides. Most glycans had one or two fucose residues bound through alpha1,3 and alpha1,6 linkages to the innermost GlcNAc. On the basis of these results and on the structure of an eight-cysteine domain from fibrillin-1, we present a model of glycosylated CR3 and discuss the role of glycosylation in eight-cysteine domain protein-protein interactions.

Glycosylation differences between the normal and pathogenic prion protein isoforms

Prion protein consists of an ensemble of glycosylated variants or glycoforms. The enzymes that direct oligosaccharide processing, and hence control the glycan profile for any given glycoprotein, are often exquisitely sensitive to other events taking place within the cell in which the glycoprotein is expressed. Alterations in the populations of sugars attached to proteins can reflect changes caused, for example, by developmental processes or by disease. Here we report that normal (PrP(C)) and pathogenic (PrP(Sc)) prion proteins (PrP) from Syrian hamsters contain the same set of at least 52 bi-, tri-, and tetraantennary N-linked oligosaccharides, although the relative proportions of individual glycans differ. This conservation of structure suggests that the conversion of PrP(C) into PrP(Sc) is not confined to a subset of PrPs that contain specific sugars. Compared with PrP(C), PrP(Sc) contains decreased levels of glycans with bisecting GlcNAc residues and increased levels of tri- and tetraantennary sugars. This change is consistent with a decrease in the activity of N-acetylglucosaminyltransferase III (GnTIII) toward PrP(C) in cells where PrP(Sc) is formed and argues that, in at least some cells forming PrP(Sc), the glycosylation machinery has been perturbed. The reduction in GnTIII activity is intriguing both with respect to the pathogenesis of the prion disease and the replication pathway for prions.

Roles for glycosylation of cell surface receptors involved in cellular immune recognition

The majority of cell surface receptors involved in antigen recognition by T cells and in the orchestration of the subsequent cell signalling events are glycoproteins. The length of a typical N-linked sugar is comparable with that of an immunoglobulin domain (30 A). Thus, by virtue of their size alone, oligosaccharides may be expected to play a significant role in the functions and properties of the cell surface proteins to which they are attached. A databank of oligosaccharide structures has been constructed from NMR and crystallographic data to aid in the interpretation of crystal structures of glycoproteins. As unambiguous electron density can usually only be assigned to the glycan cores, the remainder of the sugar is then modelled into the crystal lattice by superimposing the appropriate oligosaccharide from the database. This approach provides insights into the roles that glycosylation might play in cell surface receptors, by providing models that delineate potential close packing interactions on the cell surface. It has been proposed that the specific recognition of antigen by T cells results in the formation of an immunological synapse between the T cell and the antigen-presenting cell. The cell adhesion glycoproteins, such as CD2 and CD48, help to form a cell junction, providing a molecular spacer between opposing cells. The oligosaccharides located on the membrane proximal domains of CD2 and CD48 provide a scaffold to orient the binding faces, which leads to increased affinity. In the next step, recruitment of the peptide major histocompatibility complex (pMHC) by the T-cell receptors (TCRs) requires mobility on the membrane surface. The TCR sugars are located such that they could prevent non-specific aggregation. Importantly, the sugars limit the possible geometry and spacing of TCR/MHC clusters which precede cell signalling. We postulate that, in the final stage, the sugars could play a general role in controlling the assembly and stabilisation of the complexes in the synapse and in protecting them from proteolysis during prolonged T-cell engagement.

Glycosylation of natural human neutrophil gelatinase B and neutrophil gelatinase B-associated lipocalin

Gelatinase B is a matrix metalloproteinase (MMP-9) involved in tissue remodeling, development, cancer, and inflammation. Neutrophils produce three major forms of (pro)gelatinase B: 92 kDa monomers, homodimers, and complexes of gelatinase B covalently bound to neutrophil gelatinase B-associated lipocalin (NGAL). In contrast to the case for other proteinases, little information about the glycosylation of any natural human MMP is available. Here, both gelatinase B and NGAL were purified from human peripheral blood neutrophils, and the entire contents of the released N- and O-glycan pools were analyzed simultaneously using recently developed high-performance liquid chromatography-based technology. The results are discussed within the context of the domain structure of gelatinase B and a molecular model of NGAL based on data from this study and the three-dimensional nuclear magnetic resonance (NMR) structure of the protein. More than 95% of the N-linked glycans attached to both gelatinase B and NGAL were partially sialylated, core-fucosylated biantennary structures with and without outer arm fucose. The O-linked glycans, which were estimated to comprise approximately 85% of the total sugars on gelatinase B, mainly consisted of type 2 cores with Galbeta1,4GlcNAc (lactosamine) extensions, with or without sialic acid or outer arm fucose. This paper also contains the first report of O-linked glycans attached to NGAL. Although both proteins were isolated from neutrophils and contained O-linked glycans mainly with type 2 cores, the glycans attached to individual serine/threonine residue(s) in NGAL were significantly smaller than those on gelatinase B. In contrast to NGAL, gelatinase B contains a region rich in Ser, Thr, and Pro typical of O-glycosylated mucin-like domains.

Tyrosinase folding and copper loading in vivo: a crucial role for calnexin and alpha-glucosidase II

Tyrosinase is the key enzyme of melanin biosynthesis. It is a multiply glycosylated metalloenzyme, which has a long maturation time making it an ideal in vivo model system to probe protein folding and metal loading events. The use of NB-DNJ, an alpha-glucosidase I and II inhibitor has allowed us to dissect these processes. Here we show that tyrosinase folds through several inactive intermediates, at least two of which are recognised by the ER chaperone, calnexin. If the association with calnexin is prevented, more rapid folding occurs, the resulting protein fails to bind copper and is inactive. If dissociation from calnexin is inhibited, folding is prevented; the protein does not go through the normal secretory pathway and is targeted for degradation. Thus, tyrosinase folds off calnexin, giving alpha-glucosidase II a critical role, but the association with calnexin is essential to promote the correct folding which enables it to acquire copper.

Glycoproteins: glycan presentation and protein-fold stability

Glycosylation of proteins has been shown to play a role in a variety of cellular events. Thanks to recent advances in obtaining conformational constraints across glycosidic linkages, structural characterisation of glycoproteins has improved considerably. It is now becoming apparent that N-glycosylation of a folded protein can have a significant stabilising effect on large regions of the backbone structure.

Oligosaccharide analysis and molecular modeling of soluble forms of glycoproteins belonging to the Ly-6, scavenger receptor, and immunoglobulin superfamilies expressed in Chinese hamster ovary cells

Most cell surface molecules are glycoproteins consisting of linear arrays of globular domains containing stretches of amino acid sequence with similarities to regions in other proteins. These conserved regions form the basis for the classification of proteins into superfamilies. Recombinant soluble forms of six leukocyte antigens belonging to the Ly-6 (CD59), scavenger receptor (CD5), and immunoglobulin (CD2, CD48, CD4, and Thy-1) superfamilies were expressed in the same Chinese hamster ovary cell line, thus providing an opportunity to examine the extent to which N-linked oligosaccharide processing might vary in a superfamily-, domain-, or protein-dependent manner in a given cell. While we found no evidence for superfamily-specific modifications of the glycans, marked differences were seen in the types of oligosaccharides attached to individual proteins within a given superfamily. The relative importance of local protein surface properties versus the overall tertiary structure of the molecules in directing this protein-specific variation was examined in the context of molecular models. These were constructed using the 3D structures of the proteins, glycan data from this study, and an oligosaccharide structural database. The results indicated that both the overall organization of the domains and the local protein structure can have a large bearing on site-specific glycan modification of cells in stasis. This level of control ensures that the surface of a single cell will display a diverse repertoire of glycans and precludes the presentation of multiple copies of a single oligosaccharide on the cell surface. The glycans invariably shield large regions of the protein surfaces although, for the glycoproteins examined here, these did not hinder the known active sites of the molecules. The models also indicated that sugars are likely to play a role in the packing of the native cell surface glycoproteins and to limit nonspecific protein-protein interactions. In addition, glycans located close to the cell membrane are likely to affect crucially the orientation of the glycoproteins to which they are attached.

A statistical analysis of N- and O-glycan linkage conformations from crystallographic data

We have generated a database of 639 glycosidic linkage structures by an exhaustive survey of the available crystallographic data for isolated oligosaccharides, glycoproteins, and glycan-binding proteins. For isolated oligosaccharides there is relatively little crystallographic data available. A much larger number of glycoprotein and glycan-binding protein structures have now been solved in which two or more linked monosaccharides can be resolved. In the majority of these cases, only a few residues can be seen. Using the 639 glycosidic linkage structures, we have identified one or more distinct conformers for all the linkages. The O5-C1-O-C(x)' torsion angles for all these distinct conformers appear to be determined chiefly by the exo-anomeric effect. The Manalpha1-6Man linkage appears to be less restrained than the others, showing a wide degree of dispersion outside the ranges of the defined conformers. The identification of distinct conformers for glyco-sidic linkages allows "average" glycan structures to be modeled and also allows the easy identification of distorted glycosidic linkages. Such an analysis shows that the interactions between IgG Fc and its own N-linked glycan result in severe distortion of the terminal Galbeta1-4GlcNAc linkage only, indicating the strong interactions that must be present between the Gal residue and the protein surface. The applicability of this crystallographic based analysis to glycan structures in solution is discussed. This database of linkagestructures should be a very useful reference tool in three-dimensional structure determinations.

Polyhydroxylated pyrrolidine and pyrrolizidine alkaloids from Hyancinthoides non-scripta and Scilla campanulata

Aqueous ethanol extracts from the immature fruits and stalks of bluebell (Hyacinthoides non-scripta) were subjected to various ion-exchange column chromatographic steps to give 1,4-dideoxy-1,4-imino-D-arabinitol (1),2(R),5(R)-bis(hydroxymethyl)-3(R),4(R)-dihydroxypyrrolidine (DMDP) (2), 6-deoxy-6-C-(2,5-dihydroxyhexyl)-DMDP (3),2,5-dideoxy-2,5-imino-DL-glycero-D-manno-heptitol (homoDMDP)(4),homoDMDP-7-O-apioside (5), homoDMDP-7-O-beta-D-xylopyranoside (6), (1S*,2R*,3R*,5R*,7aR*)-1,2-dihydroxy-3,5- dihydroxymethylpyrrolizidine (7), and (1S*,2R*,3R*,5R*,6R*,7R*,7aR*)-3-hydroxymethyl-5-methyl-1,2,6,7 tetrahydroxypyrrolizidine (8). Bulbs of Scilla campanulata (Hyacinthaceae) yielded (1S*,2R*,3R*,5S*,7aR*)-1,2-dihydroxy-3,5-dihydroxy-methylpyrrol izidine (9) in addition to compounds 1-7. Compounds 3,6,7,8, and 9 are new natural products. Compound 4 is a potent competitive inhibitor with K(i) values of 1.5 microM for Caldocellum saccharolyticum beta-glucosidase and 2.2 microM for bovine liver beta-galactosidase. The 7-O-beta-D xyloside 6 was a stronger competitive inhibitor than 4 of C saccharolyticum beta-glucosidase and rat intestinal lactase, with K(i) values of 0.06 and 0.07 microM, respectively, but a weaker inhibitor of bovine liver beta-galactosidase. Furthermore, compound 4 is also a competitive inhibitor (K(i) = 1.8 microM) of porcine kidney trehalase, but 6 was inactive against this enzyme.

The high degree of internal flexibility observed for an oligomannose oligosaccharide does not alter the overall topology of the molecule

The conformational properties of oligosaccharides are important in determining their biological properties, such as recognition by proteins. The structural and dynamic properties of many oligosaccharides are poorly understood both because of a lack of experimental data (usually obtained from solution NMR parameters) and because of gross approximations frequently invoked in theoretical models. To characterise the oligomannose oligosaccharide Man,GlcNAc2 we have acquired a more extensive NMR data set and performed the first unrestrained molecular dynamics (MD) simulation in water of this large oligosaccharide (employing the GLYCAM_93 parameter set with the AMBER force field). Good agreement is seen between the computed dynamics data and the results of both an isolated spin pair (ISPA) analysis of short mixing time NOE data and NOE build-up curves for mixing times from 100 to 2000 ms. The number of experimental conformational constraints obtained in this study are in principle sufficient to fully define a rigid structure. The fact that this could not be done indicates a high degree of internal flexibility and/or the presence of multiple conformations about the glycosidic linkages. Independently, the same conclusions are reached from an analysis of the MD results. In addition, the theoretical results allow the overall topology of the molecule and its intra-molecular and solvent-mediated hydrogen bonding pattern to be defined. Extensive re-organisation of solvent and inter-residue hydrogen bonds is shown to be required for significant conformational changes to occur, resulting in relatively long life-times for distinct glycosidic linkage conformations, despite the high local flexibility of the glycosidic linkages. This factor is also seen in the overall topology of the molecule, where the considerable internal flexibility is not translated into gross changes in structure. The control exerted by the solvent over both the flexibility and overall topology of an oligosaccharide has important implications for recognition processes and for the conformational properties of glycans attached to glycoproteins.

The glycosylation and structure of human serum IgA1, Fab, and Fc regions and the role of N-glycosylation on Fcα receptor interactions

The human serum immunoglobulins IgG and IgA1 are produced in bone marrow and both interact with specific cellular receptors that mediate biological events. In contrast to IgA1, the glycosylation of IgG has been well characterized, and its interaction with various Fc receptors (Fc Rs) has been well studied. In this paper, we have analyzed the glycosylation of IgA1 and IgA1 Fab and Fc as well as three recombinant IgA1 molecules, including two N-glycosylation mutants. Amino acid sequencing data of the IgA1 Fc O-glycosylated hinge region indicated that O-glycans are located at Thr228, Ser230, and Ser232, while O-glycan sites at Thr225 and Thr236 are partially occupied. Over 90% of the N-glycans in IgA1 were sialylated, in contrast to IgG, where < 10% contain sialic acid. This paper contains the first report of Fab glycosylation in IgA1, and (in contrast to IgG Fab, which contains only N-linked glycans) both N- and O-linked oligosaccharides were identified. Analysis of the N-glycans attached to recombinant IgA1 indicated that the Cα 2 N-glycosylation site contained mostly biantennary glycans, while the tailpiece site, absent in IgG, contained mostly triantennary structures. Further analysis of these data suggested that processing at one Fc N-glycosylation site affects the other. Neutrophil Fcα R binding studies, using recombinant IgA1, indicated that neither the tailpiece region nor the N-glycans in the C alpha 2 domain contribute to IgA1-neutrophil Fcα R binding. This contrasts with IgG, where removal of the Fc N-glycans reduces binding to the Fcγ R. The primary sequence and disulfide bond pattern of IgA1, together with the crystal structures of IgG1 Fc and mouse IgA Fab and the glycan sequencing data, were used to generate a molecular model of IgA1. As a consequence of both the primary sequence and S-S bond pattern, the N-glycans in IgA1 Fc are not confined within the inter-α-chain space. The accessibility of the Cα 2 N-glycans provides an explanation for the increased sialylation and galactosylation of IgA1 Fc over that of IgG Fc N-glycans, which are confined in the space between the two Cγ 2 domains. This also suggests why in contrast to IgG Fc, the IgA1 N-glycans are not undergalactosylated in rheumatoid arthritis.

The solution NMR structure of glucosylated N-glycans involved in the early stages of glycoprotein biosynthesis and folding

Glucosylated oligomannose N-linked oligosaccharides (Glc(x)Man9GlcNAc2 where x = 1-3) are not normally found on mature glycoproteins but are involved in the early stages of glycoprotein biosynthesis and folding as (i) recognition elements during protein N-glycosylation and chaperone recognition and (ii) substrates in the initial steps of N-glycan processing. By inhibiting the first steps of glycan processing in CHO cells using the alpha-glucosidase inhibitor N-butyl-deoxynojirimycin, we have produced sufficient Glc3Man7GlcNAc2 for structural analysis by nuclear magnetic resonance (NMR) spectroscopy. Our results show the glucosyl cap to have a single, well-defined conformation independent of the rest of the saccharide. Comparison with the conformation of Man9GlcNAc2, previously determined by NMR and molecular dynamics, shows the mannose residues to be largely unaffected by the presence of the glucosyl cap. Sequential enzymatic cleavage of the glucose residues does not affect the conformation of the remaining saccharide. Modelling of the Glc3Man9GlcNAc2, Glc2Man9GlcNAc2 and Glc1Man9GlcNAc2 conformations shows the glucose residues to be fully accessible for recognition. A more detailed analysis of the conformations allows potential recognition epitopes on the glycans to be identified and can form the basis for understanding the specificity of the glucosidases and chaperones (such as calnexin) that recognize these glycans, with implications for their mechanisms of action.

The glycosylation of the complement regulatory protein, human erythrocyte CD59

Human erythrocyte CD59 contains N- and O-glycans and a glycosylphosphatidylinositol (GPI) anchor, all of which have been analyzed in this study. The anchor consists principally of the minimum core glycan sequence Manalpha1-2Manalpha1-6Manalpha1-4GlcN-linked to a phosphatidylinositol moiety with the structure sn-1-O-alkyl(C18:0 and C18:1)-2-O-acyl(C20:4)glycerol-3-phospho-1-(2-O-palmitoyl(C16:0))myo- inositol. This structure is essentially identical to that of human erythrocyte cholinesterase (Deeg, M. A., Humphrey, D. R., Yang, S. H. , Ferguson, T. R., Reinhold, V. N., and Rosenberry, T. L. (1992) J. Biol. Chem. 267, 18573-18580). This first comparison of GPI anchors from different proteins expressed in the same tissue suggests that human reticulocytes produce only one type of anchor structure. The N- and O-glycans were sequenced using a novel approach involving digestion of the total glycan pool with multiple enzyme arrays. The N-glycan pool contained families of bi-antennary complex-type structures with and without lactosamine extensions and outer arm fucose residues. The predominant O-glycans were NeuNAcalpha2-3Galbeta1-3GalNAc and Galbeta1-3[NeuNAcalpha2-3]GalNAc. Inspection of a molecular model of CD59, based on the NMR solution structure of the extracellular domain and the structural data from this study, suggested several roles for the glycans, including spacing and orienting CD59 on the cell surface and protecting the molecule from proteases. This work completes the initial structural analysis of CD59, providing the most complete view of any cell surface glycoprotein studied to date.

Variations in oligosaccharide-protein interactions in immunoglobulin G determine the site-specific glycosylation profiles and modulate the dynamic motion of the Fc oligosaccharides

Glycoproteins, such as immunoglobulin G (IgG), consist of an ensemble of glycosylated variants, or glycoforms, which have different oligosaccharides attached to a common peptide. Alterations in the normal glycoform populations of IgG are associated with certain disease states, notably rheumatoid arthritis and its remission during pregnancy. In this paper, we show that two sets of IgG Fc glycoforms have quite different physical properties. The first set has 1,6 arm terminal galactose residues which interact with the protein, resulting in glycan binding to the protein surface, in agreement with the crystal structure. In contrast, the second set of glycoforms which lack galactose does not bind to the protein surface. Recently developed HPLC techniques combined with enzymatic digestion and mass spectrometry have been used to assign the glycan structures on IgG, Fab, and Fc. Comparison of Fab with Fc shows that glycosylation is site-specific. Two major glycan structures are present on Fab (fucosylated digalacto-bianntenary with and without bisect) and three on Fc (fucosylated agalacto-, 1,6 arm monogalacto-, and digalacto-bianntenary). In comparison to Fab, Fc glycans contain (i) lower levels of bisecting GlcNAc, (ii) lower levels of galactose, (iii) higher than expected levels of 1,6 arm galactose relative to 1,3 arm, and (iv) no 1,6 arm sialylation. We interpret these differences to indicate a role for both the protein quaternary structure and specific protein-glycan interactions in determining the glycoform populations. NMR relaxation measurements have been used to probe the mobility of the glycans in the Fc. By comparing two samples with different glycoform populations, we conclude that this mobility is dependent on the primary sequence of the glycan. Glycans carrying a galactose residue on the 1,6 arm have relaxation properties very similar to those of the peptide backbone and thus do not have independent motion. Glycans lacking galactose have relaxation rates 30 times slower than that of the peptide and thus a higher degree of mobility. These agalactosyl glycans do not interact with the protein, resulting in exposure of previously covered regions of the peptide surface and making the glycan more accessible. This implies that at the early stages of glycan processing the Fc glycans are mobile and only partially protected by the protein quaternary structure. Immobilization of the glycans occurs as a consequence of addition of galactose to the 1,6 arm and results in increased protection.

Short communication: selective placental transport of maternal IgG to the fetus

During pregnancy there is a dramatic reduction in the serum levels of agalactosyl IgG (G0IgG) in both normal women and those with rheumatoid arthritis. In order to determine if a similar reduction in G0IgG were apparent in fetal serum, a comparison of the galactose content of IgG from nine paired samples of umbilical vein or fetal blood and peripheral maternal serum, at gestational ages ranging from 16-41 weeks was performed. The full-term maternal IgG samples were highly galactosylated, so confirming previous observations of reduced G0IgG levels during pregnancy. In addition every paired sample of fetal IgG had a higher level of galactosylation than the corresponding maternal IgG. Therefore, during pregnancy there is both a reduced biosynthesis of the G0IgG glycoform by the mother, and a restriction of its transport across the placenta. The ratio of estimated G0IgG in fetal and maternal serum was found to be related to changes in IgG transport, and in particular the active transport of IgG1 across the placenta during gestation. Our data suggest that the placental IgG transport mechanism is either carbohydrate independent by discriminating for IgG1, or is carbohydrate dependent selecting for highly galactosylated IgG glycoforms. This study emphasizes the need for further investigations on the biological function of G0IgG in normal physiological states, in addition to disease states, such as juvenile and adult rheumatoid arthritis, where elevated G0IgG levels correlate with disease activity.

Glycobiology: 'the function of sugar in the IgG molecule'

Immunoglobulin G (IgG) is glycosylated in both the Fc and the Fab regions of the protein with a heterogeneous ensemble of structures (glycoforms) that is both highly reproducible (i.e. nonrandom) and site specific. In normal IgG, the 2 highly conserved oligosaccharides of the Fc region are found buried between the CH2 domains, forming specific protein-saccharide interactions with the Fc protein surface. One of the functions attributed to the Fc oligosaccharides of normal IgG is to maintain the conformational arrangements of the Fc domains as well as the hinge regions. These structural features are necessary for Fc effector functions such as Clq and monocyte binding. A hallmark of rheumatoid arthritis (RA) patients is a dramatic increase in the presence of serum IgG containing Fc oligosaccharides lacking an outer arm galactose residue (termed 'G0' glycoforms). The increased level of G0 has been shown to be directly related to the pathogenesis of RA. Nuclear magnetic resonance relaxation studies of the Fc region from normal and RA IgG, as well as examination of x-ray structures, show that the G0 oligosaccharides have an increased mobility resulting from the loss of binding between the G0 oligosaccharide and the Fc protein surface. From these observations it follows that regions of the protein surface that are normally covered by the oligosaccharide are revealed. The newly accessible protein surface could have lectin-like activity and also be inherently antigenic. In addition, the more mobile G0 oligosaccharide can be recognised by mannose binding protein. As the mannose binding protein can activate complement, and the Fc oligosaccharide would not normally be accessible to protein recognition, this finding might suggest a specific role for the G0 glycoform in inflammation when the appropriate IgG glycoforms are clustered.

The effects of variable glycosylation on the functional activities of ribonuclease, plasminogen and tissue plasminogen activator

The relatively large size and dynamics of oligosaccharides can result in substantial shielding of functionally important areas of proteins to which they are attached, modulate the interactions of glycoconjugates with other molecules and affect the rate of processes which involve conformational changes. This review focuses on the occupancy of N-linked glycosylation sites on three enzymes, ribonuclease, plasminogen and tissue plasminogen activator. Each of these proteins occurs naturally as two populations of molecules, distinguished from each other only by the presence or absence of an oligosaccharide at one glycosylation site. The presence of an oligomannose sugar on ribonuclease (at Asn-34) alters its overall dynamics, increases its stability towards proteinases and decreases its functional activity towards double-stranded RNA. The N-linked sugar on plasminogen (at Asn-288) within kringle 3 reduces the rate of the beta- to alpha-conformational change, modulates the transport of plasminogen into the extravascular compartment, decreases plasminogen binding to U937 cells and downregulates the activation of plasminogen by both urokinase and tissue plasminogen activator. Additionally, in fibrinolysis, within a ternary complex of fibrin, plasminogen and tissue plasminogen activator, the N-linked sugar of plasminogen hinders the initial interaction with tissue plasminogen activator (i.e., it alters Km). The presence of an N-linked glycan (at Asn-184) in the kringle 2 domain of tissue plasminogen activator hinders the rearrangement of this ternary complex, decreasing the turnover rate (Kcat).

Glycosylation changes of IgG associated with rheumatoid arthritis can activate complement via the mannose-binding protein

The glycosylation of the circulating immunoglobulin-gamma (IgG) antibody molecules changes in rheumatoid arthritis. The extent of the changes correlates with the disease severity and reverses in remission. We demonstrate here that the alteration in glycosylation associated with rheumatoid arthritis can create a new mode for the interaction of IgG with complement through binding to the collagenous lectin mannose-binding protein (MBP). Rheumatoid arthritis is associated with a marked increases in IgG glycoforms that lack galactose (referred to as G0 glycoforms) in the Fc region of the molecule and that terminate in N-acetyl glucosamine (GlcNAc). We show, using nuclear magnetic resonance (NMR) and X-ray data, that these terminal GlcNAc residues become accessible for MBP binding. We further demonstrate that multiple presentation of IgG-G0 glycoforms to MBP results in activation of the complement. This suggests that a contribution to the chronic inflammation of the synovial membrane could arise from the localization of the IgG-G0 glycoforms in the affected joint and from resulting activation of complement.

Binding of sugar ligands to Ca(2+)-dependent animal lectins. I. Analysis of mannose binding by site-directed mutagenesis and NMR

The Ca(2+)-dependent carbohydrate-recognition domain (CRD) of rat serum mannose-binding protein has been subjected to site-directed mutagenesis to determine the importance of individual residues in ligation of mannose and related sugars. The effects of the mutations were assessed by direct binding assays, competition binding studies, partial proteolysis, and NMR analysis of sugar-CRD titrations. As suggested by the crystal structure of the mannose-binding CRD complexed with oligosaccharide ligand, asparagine and glutamic acid residues that interact with hydroxyl groups 3 and 4 of the sugar, as well as with one of the two bound Ca2+, are critical for ligand binding. In addition, the beta-carbon of His189 contributes substantially to the binding affinity, apparently through a van der Waals contact with C-4 of the sugar ligand. van der Waals contacts between the imidazole ring of His189 and the 2 hydroxyl group of mannose, and between Ile207 and C-6 of mannose, observed in the crystal structure, contribute less to stability of the ligand complex. The effects of changes at positions 189 and 207 on the ability of the CRD to distinguish between alpha-and beta-methyl L-fucosides suggest that fucose may bind in an alternative orientation compared to the arrangement originally proposed based on the mannose-CRD complex.

Fast sequencing of oligosaccharides: the reagent-array analysis method

A method of oligosaccharide analysis involving controlled fragmentation resulting from enzymatic digestion is presented. The principle involves generating a set of fragments from the original oligosaccharides, characterizing them in terms of their hydrodynamic volumes, determining their molar proportions, and identifying the oligosaccharides by comparison with a computer-generated data base. Experimentally, this technique involves incubation of aliquots of a sample with a set of defined mixtures of exoglycosidases followed by pooling of the products and a single analysis on the product pool. This method has several practical advantages over current techniques, including speed and the ability to use smaller amounts of starting material. The detection of the intensity-versus-hydrodynamic volume profile is limited only by the specific activity of the labeling method. The ability to perform the enzyme digestions is limited by the individual Km values of the enzymes.

The solution conformation of the Le chi group

The solution conformation of the non-reducing terminal Gal beta 1----4 (Fuc alpha 1----3)GlcNAc (Lewis X or Le chi) group in the oligosaccharide Lacto-N-fucopentaose (LNFP) III has been determined by high resolution 1H NMR spectroscopy and semi-empirical quantum mechanical calculations. The two methods give the same single conformer for the Le chi group showing close packing of the Gal and Fuc rings. The metal binding properties and homotypic oligomer formation of LNFP III have also been investigated by NMR spectroscopy. No evidence for metal binding or high-affinity homotypic oligomer formation has been found.

Characterisation of oligosaccharides from Drosophila melanogaster glycoproteins

An analysis of the released oligosaccharides from a membrane glycoprotein preparation of third instar larvae (3rdIL), and purified larval serum protein 2 (LSP2) from Drosophila melanogaster was performed. Sequential exoglycosidase digestion in combination with high-resolution gel permeation chromatography and partial acetolysis indicated the presence of two series of oligomannosides; one of these series was unusual and characterized by the presence of a core alpha 1-6 linked fucose, the other was a typical mammalian oligomannose series containing the following isomers -D1, -D2, -D12, -D123 and -CD123 as well as the unprocessed Man9GlcNAc2 structure. Conventional oligomannose could only be detected in the LSP2 sample. This study opens the way to use powerful molecular and classical genetic techniques to analyse the control and functional significance of glycosylation in higher organisms.

The conformational effects of N-glycosylation on the tailpiece from serum IgM

1H-NMR spectroscopy has been used to study the conformation and dynamics of the isolated tailpiece from human serum immunoglobulin M, a 22-residue peptide containing a single asparagine glycosylation site. The peptide is isolated as a set of glycoforms, varying only in the sequence of the oligosaccharide attached at the glycosylation site. The oligosaccharides present have the general formula (Man)n(GlcNAc)2, with 45% having n = 6, 45% having n = 8 and 10% having n = 7 and/or 9. They have been identified and their NMR parameters compared to those found for the isolated oligosaccharides in free solution. The conformation and dynamics of the peptide component have also been studied, using NOE data and hydrogen-exchange experiments, and the results compared to those obtained from the aglycosyl peptide of the same sequence. The presence of the peptide is found to have no measurable effect on the conformation of the oligosaccharides. However, the presence of oligosaccharide causes a decrease in the conformational mobility of the backbone and sidechains of the peptide in the region of the glycosylation site. This is proposed to result from interactions between the oligosaccharide core and the amino acid side chains. Further, the conformation of the N-glycosidic linkage has been shown to be both rigid and planar. Thus, the conformational space available to an N-linked oligosaccharide in a glycoprotein relative to the protein may depend to a large extent upon the flexibility of the asparagine side chain. Various roles for the different glycoforms of the tail peptide are discussed.

Some structural features of the iron-uptake regulation protein

An extensive proton nuclear magnetic resonance study of the iron-uptake regulation protein (Fur) from Escherichia coli has been made. Considerable difficulties were experienced in the NMR experiments in 1H2O which may be due unfavourable proton exchange rates in the pH range greater than 6.2, where the protein is soluble. Even in 2H2O, the two-dimensional NMR spectra were not easily interpreted due to widely differing line widths, as a result of the protein side-chains having very differing mobilities. Despite these problems, virtually all the 20 aromatic amino acids have been assigned. Small regions of the protein core were assigned by taking advantage of the approximately 20 non-exchanging peptide-NH resonances in 2H2O. Using two-dimensional J-correlated, homonuclear Hartmann-Hahn and NOE spectroscopies, we have been able to give some assignments in which there is considerable confidence for about one third of the amino acids. Taking advantages of two series of probe experiments, using Mn(II) and a spin label, together with longer range NOE data and result from structure predictions and CD data, we have put forward a tentative fold for the protein which is seen to have a relatively rigid series of interior strands and more flexible exterior strands, many of which are likely to be helical. The Mn(II) probe experiments have also allowed us to define the Fe(II) binding site.

Structure and function of X-Pro dipeptide repeats in the TonB proteins of Salmonella typhimurium and Escherichia coli

The TonB protein is required for several outer membrane transport processes in bacteria. A short 33-residue peptide segment of TonB has been studied by 1H and 13C nuclear magnetic resonance spectroscopy. The sequence of this peptide segment contains multiple Glu-Pro and Lys-Pro dipeptide repeats that maintain rigid, elongated structures and flank a short connecting segment that adopts a beta-strand configuration. This TonB peptide is shown to interact specifically with the FhuA protein, the outer membrane receptor for ferrichrome-iron, providing the first direct evidence that the TonB protein interacts with outer membrane receptors. Interaction with the FhuA protein involves the extended structural element containing positively charged Lys-Pro repeats, and suggests a functional role for this segment of the TonB protein. As TonB is anchored in the cytoplasmic membrane the protein must, uniquely, span the periplasm. These data, together with studies described in the accompanying paper, suggest a model by which TonB serves to transduce conformational information over extended distances, from the cytoplasmic membrane to the outer membrane.

Solution NMR studies of colicin E1 C-terminal thermolytic peptide. Structural comparison with colicin A and the effects of pH changes

The aqueous solution structure of the C-terminal thermolytic peptide of colicin E1 has been investigated using both one- and two-dimensional NMR techniques. The NMR data are consistent with a fold for the peptide very similar to that reported for the colicin A C-terminal peptide in the crystalline state, although some differences have been noted. The one-dimensional NMR spectrum of the peptide has been used to follow changes in both the structure and dynamics of the peptide on changing pH. The in vitro functionally competent form of the peptide (present in solution only below pH 6) does not differ in structure significantly from the higher pH form. However, small local conformational changes are observed together with an increase in mobility in some of the more hydrophilic regions. This suggests that the effect of lower pH is to change the ease with which the major conformational changes during insertion into a membrane can occur.

Identification and description of beta-structure in horse muscle acylphosphatase by nuclear magnetic resonance spectroscopy

Nuclear magnetic resonance spectra of acylphosphatase were searched for signs of beta-structure, i.e. characteristic nuclear Overhauser enhancement patterns displayed in the two-dimensional spectra, typical chemical shifts, coupling constants and slow 2H-H exchange. The results provided identification of the main-chain resonances of amino acid residues involved in the beta-structure. The full sequential assignment of this region was gained by identification of some amino acid spin systems and their alignment with the primary sequence. The assignment of the side-chains was virtually completed subsequently and a list produced of nuclear magnetic resonance (n.m.r.) constraints derived from the spectra. The beta-structure consists of a beta-sheet with four antiparallel chains, one attached parallel chain, three tight turns and a beta-bulge. The conformation of the beta-sheet was determined by distance geometry calculation using the n.m.r. constraints (174 intraresidual, 107 sequential and 226 long-range distances, 32 torsion angles, phi, and 28 hydrogen bonds) as input. Observation of some interactions between the sheet and previously identified alpha-helical regions made it possible to give an outline of the three-dimensional structure of the enzyme.

Mapping space by NMR using susceptibility changes at phase boundaries

A technique for differentiating high-resolution NMR signals from different regions of small objects is outlined and some initial results on model systems are given. This method uses inorganic paramagnetic or diamagnetic ions to create magnetic field gradients at phase boundaries.