Conformational analysis of oligosaccharides corresponding to the cell-wall polysaccharide of the Streptococcus group A by Metropolis Monte Carlo simulations.

Metropolis Monte Carlo simulations have been performed on four substructures from the cell-wall polysaccharide antigen of Streptococcus group A to explore the conformational behaviour of these compounds. The compounds examined are the trisaccharide, propyl 3-O-(2-acetamido-2-deoxy-beta-D-glucopyranosyl)-2-O-(alpha-L-rhamnopyranosyl)- alpha-L-rhamnopyranoside, 1, the tetrasaccharide, propyl 3-O-(3-O-(2-acetamido-2-deoxy-beta-D- glucopyranosyl)-2-O-(alpha-L-rhamnopyranosyl)-alpha-L-rhamnopyranosyl)-alpha-L- rhamnopyranoside, 2, the hexasaccharide, propyl 3-O-(2-O-(3-O-(3-O-(2-acetamido-2-deoxy-beta-D-glucopyranosyl)-alpha-L- rhamnopyranosyl)-alpha-L-rhamnopyranosyl)-3-O-(2-acetamido-2-deoxy-beta-D- glucopyranosyl)-alpha-L-rhamnopyranosyl)-alpha-L-rhamnopyranoside, 3, and the hexasaccharide, propyl 3-O-(2-acetamido-2-deoxy-beta-D-glucopyranosyl)-2-O-(3-O-(3-O-(2-acetamido-2- deoxy-beta-D-glucopyranosyl)-2-O-(alpha-L-rhamnopyranosyl)-alpha-L- rhamnopyranosyl)-alpha-L-rhamnopyranosyl)-alpha-L-rhamnopyranoside, 4. In general, the conformational flexibility of similar glycosidic linkages in different compounds is comparable. However, in a few cases, small differences in the conformations available to these linkages in different structural environments could be detected. Interestingly, a second conformation found for the beta-D-GlcNAc-(1-->3)-alpha-L-Rha linkage in three of the compounds was not populated in the hexasaccharide 4. Furthermore, a conformational locale of the alpha-L-Rha-(1-->3)-alpha-L-Rha linkage found to be populated in the trisaccharide 1, tetrasaccharide 2, and hexasaccharide 4 is negligibly populated in the hexasaccharide 3. Ensemble averaged proton-proton distances compare favourably with experimental average distances obtained from NMR spectroscopy. The trisaccharide branch point in the hexasaccharides is shown to be a highly defined conformational feature. The same unit has been found to be one of the crucial elements recognized by anti-Group A Streptococcus antibodies, a result that has implications for the design of improved immunodiagnostics and vaccines.

The structure of the carbohydrate backbone of the core-lipid-A region of the lipopolysaccharide from Vibrio cholerae strain H11 (non-O1).

Lipopolysaccharide from Vibrio cholerae strain H11 (non-O1) was de-O-acylated, dephosphorylated, reduced, de-N-acylated, N-acetylated, and the products were separated by high-performance anion-exchange chromatography (HPAE). A decasaccharide, 1, was isolated as the major product, representing the core oligosaccharide attached to the reduced GlcN-disaccharide lipid A backbone. Its structure was established by compositional and methylation analyses, and extensive NMR investigations including 1H,1H correlation spectroscopy (COSY), total correlation spectroscopy (TOCSY), and nuclear Overhauser enhancement spectroscopy (NOESY), as well as heteronuclear 13C,1H COSY. In another reaction sequence the lipopolysaccharide was hydrolysed with dilute acetic acid and reduced with NaBH4. The resulting core fractions were separated by HPAE giving seven individual octasaccharides differing at the reducing 3-deoxy-D-manno-octulosonic acid (Kdo) residue. A major product, 2, was isolated and investigated by the same methods as described for the decasaccharide 1. The following structures are proposed for compounds 1 and 2: alpha-D-GlcNAcp-(1-7)-[beta-D-Galp-(1-3)-]-alpha-Hepp-(1-2)- alpha-Hepp- (1-3)-[beta-D-Glcp-(1-4)-]- [alpha-D-Glcp-(1-6)-]-alpha-Hepp-(1-5)-R, where R is alpha-Kdop-(2-6)-beta-D-GlcNAcp-(1-6)-D-GlcNAcol in 1 and 4,8-anhydro-Kdool in 2, and Hep is L-glycero-D-manno-heptose. In lipopolysaccharide, the terminal residue of alpha-D-glucosamine possessed a free amino group, as proved by deamination with nitrous acid and the 1H-NMR spectrum of de-O-acylated lipopolysaccharide. The conformational preferences of the terminal core heptasaccharide was assessed by Monte Carlo simulations combined with restrained calculations of side chains based on experimentally determined proton-coupling constants. These calculations, confirmed by NOE data, displayed several long-range interactions, which resulted in a well-defined three-dimensional structure of the core oligosaccharide.

Preferred conformations and dynamics of five core structures of mucin type O-glycans determined by NMR spectroscopy and force field calculations.

Glycosyltransferases acting on O-glycans have been shown to exhibit distinct specificity for the carbohydrate and the peptide moiety of their substrates. As an approach to study the 3-dimensional interactions between enzymes and O-glycan substrates, we determined the preferred conformations of five oligosaccharide-core structures of mucin type glycoproteins by NMR spectroscopy and by static and dynamic force field calculations. Seven oligosaccharides, representing five basic core structures, were investigated: Gal beta (1-3)GalNAc alpha Bzl (1, core 1), GlcNAc beta (1-6)[Gal beta (1-3)]GalNAc alpha Bzl (2, core 2), GlcNAc beta (1-3)GalNAc alpha Bzl (3, core 3), GlcNAc beta (1-6)[GlcNAc beta (1-3)]GalNAc alpha Bzl (4, core 4), GlcNAc beta (1-6)GalNAc alpha Bzl (5, core 6), the elongated core 2, Gal beta (1-4)GlcNAc beta (1-6)[Gal beta (1-3)]GalNAc alpha pNp (6) and GalNAc alpha-Bzl (7). The dynamic behaviour of the molecules was studied by Metropolis Monte Carlo (MMC) simulations. Experimental coupling constants, chemical shift changes, and NOEs were compared with results from static energy minimizations and dynamic MMC simulations and show a good agreement. MMC simulations show that the (1-6) linkage is much more flexible than the (1-3) or the (1-4) linkages. The preferred conformations of the disaccharides (1) and (3) show only slight differences due to the additional N-acetyl group in (3). The conformational equilibrium of beta (1-3) glycosidic bonds of 1 and 3 was not affected by attaching a beta (1-6) linked GlcNAc unit to the GalNAc residue in 2 and 4. However, experimental and theoretical data show that the beta (1-6) linkages of the trisaccharides 2 and 4, which carry an additional beta (1-3) linked glycosyl residue, change their preferred conformations when compared with (5). The 6-branch also shows significant interactions with the benzyl aglycon altering the preferred conformation of the hydroxymethyl group of the GalNAc to a higher proportion of the gt conformer. The (1-6) linkage of 2, 4, and 6 can have two different families of conformations of which the lower energy state is populated only to about 20% of the time whereas the other state with a relative enthalpy of approximately 4 kcal mol-1 is populated to 80%. This fact demonstrates that the two conformational states have different entropy contents. Entropy is implicitly included in MMC simulations but cannot be derived from energy minimizations.

A Monte Carlo method for conformational analysis of saccharides.

A Metropolis Monte Carlo (MMC) algorithm was applied to explore conformational spaces spanned by the exocyclic dihedral angles of four disaccharides alpha-D-Man(1-->3)-alpha-D-Man(1-->O)Me (1), alpha-D-Man(1-->2)-alpha-D-Man(1-->O)Me (2), methyl beta-cellobioside (3), and methyl beta-maltoside (4). The simulation method uses the HSEA force field and randomly samples the conformational space with an automatic preference for low-energy states. In comparison to a systematic grid search, MMC offers a much more convenient and efficient protocol for the computation of ensemble average values of experimentally accessible NMR parameters such as NOE effects or 3J coupling constants. Energy barriers of a few kcal/mol were found to be surmounted easily when running the simulations with the temperature parameter set at room temperature, whereas passing significantly higher barriers required elevated temperature parameters. Ensemble average NOE values were calculated using the MMC technique and a conventional systematic grid search showing that the MMC method adequately samples the conformational spaces of 1-4. Theoretical NOEs derived for global or local minimum conformations are different from ensemble average values, and it is shown that averaged NOEs agree significantly better with experimental data. Ensemble average NOEs for 1 derived from MMC/HSEA, and previously reported MM2CARB and AMBER calculations all showed good agreement with experimental data, with MMC/HSEA giving the closest fit.

The solution conformation of sialyl-alpha (2----6)-lactose studied by modern NMR techniques and Monte Carlo simulations.

We present a comprehensive strategy for detailed characterization of the solution conformations of oligosaccharides by NMR spectroscopy and force-field calculations. Our experimental strategy generates a number of interglycosidic spatial constraints that is sufficiently large to allow us to determine glycosidic linkage conformations with a precision heretofore unachievable. In addition to the commonly used [1H,1H] NOE contacts between aliphatic protons, our constraints are: (a) homonuclear NOEs of hydroxyl protons in H2O to other protons in the oligosaccharide, (b) heteronuclear [1H,13C] NOEs, (c) isotope effects of O1H/O2H hydroxyl groups on 13C chemical shifts, and (d) long-range heteronuclear scalar couplings across glycosidic bonds. We have used this approach to study the trisaccharide sialyl-alpha (2----6)-lactose in aqueous solution. The experimentally determined geometrical constraints were compared to results obtained from force-field calculations based on Metropolis Monte Carlo simulations. The molecule was found to exist in 2 families of conformers. The preferred conformations of the alpha (2----6)-linkage of the trisaccharide are best described by an equilibrium of 2 conformers with phi angles at -60 degrees or 180 degrees and of the 3 staggered rotamers of the omega angle with a predominant gt conformer. Three intramolecular hydrogen bonds, involving the hydroxyl protons on C8 and C7 of the sialic acid residue and on C3 of the reducing-end glucose residue, contribute significantly to the conformational stability of the trisaccharide in aqueous solution.

Simulations of the static and dynamic molecular conformations of xyloglucan. The role of the fucosylated sidechain in surface-specific sidechain folding.

The hemicellulosic polysaccharide xyloglucan binds with a strong affinity to cellulosic cell wall microfibrils, the resulting heterogeneous network constituting up to 50% of the dry weight of the cell wall in dicotyledonous plants. To elucidate the molecular details of this interaction, we have performed theoretical potential energy calculations of the static and dynamic equilibrium conformations of xyloglucan using the GEGOP software. In particular, we have evaluated the preferred sidechain conformations of hexa-, octa-, deca- and heptadecasaccharide model fragments of xyloglucan for molecules with a cellulose-like, flat, glucan backbone, and a cellobiose-like, twisted, glucan backbone conformation. For the flat backbone conformation the determination of static equilibrium molecular conformations revealed a tendency for sidechains to fold onto one surface of the backbone, defined here as the H1S face, in the fucosylated region of the polymer. This folding produces a molecule that is sterically accessible on the opposite face of the backbone, the H4S face. Typically, this folding onto the H1S surface is significantly stabilized by favorable interactions between the fucosylated, trisaccharide sidechain and the backbone, with some stabilization from adjacent terminal xylosyl sidechains. In contrast, the trisaccharide sidechain folds onto the H4S face of xyloglucan fragments with a twisted backbone conformation. Preliminary NMR data on nonasaccharide fragments isolated from sycamore suspension-cultured cell walls are consistent with the hypothesis that the twisted conformation of xyloglucan represents the solution form of this molecule. Metropolis Monte Carlo (MMC) simulations were employed to assess sidechain flexibility of the heptadecasaccharide fragments. Simulations performed on the flat, rigid, backbone xyloglucan indicate that the trisaccharide sidechain is less mobile than the terminal xylosyl sidechains. MMC calculations on a fully relaxed molecule revealed a positive correlation between a specific trisaccharide sidechain orientation and the 'flatness' of the backbone glucosyl residues adjacent to this sidechain. These results suggest that the trisaccharide sidechain may play a role in the formation of nucleation sites that initiate the binding of these regions to cellulose. Based on these conformational preferences we suggest the following model for the binding of xyloglucan to cellulose. Nucleation of a binding site is initiated by the fucosylated, trisaccharide sidechain that flattens out an adjacent region of the xyloglucan backbone. Upon contacting a cellulose microfibril this region spreads by step-wise flattening of successive segments of the backbone. Self-association of xyloglucan molecules in solution may be prevented by the low frequency of formation of these nucleation sites and the geometry of the molecules in solution.

A new force-field program for the calculation of glycopeptides and its application to a heptacosapeptide-decasaccharide of immunoglobulin G1. Importance of 1-6-glycosidic linkages in carbohydrate.peptide interactions.

Energetically favored conformations of glycopeptide 1 were calculated using the newly developed force-field program, GEGOP (geometry of glycopeptides). The three-dimensional structure of glycopeptide 1, which is part of the Fc fragment of IgG1, has been calculated. 1 contains 27 amino acid residues from Pro291 to Lys317 and a biantennary decasaccharide N-linked to Asn297. The conformations of the peptide and the carbohydrate parts are shown to be mutually dependent. Single glycosyl residues of 1 exhibit interaction energies of up to -31.8 kJ/mol with the peptide portion. Generally, only a few of the glycosyl residues of the oligosaccharide moiety express significant interaction energies with the peptide part. No easy prediction is possible of glycosyl residues which exhibit favorable interaction energies. However, in all of the calculated structures, the glycosyl residues of the 1-6-linked branches show strong attractive forces for the peptide part. 1-6-glycosidically linked branches can adopt a larger number of conformations than other linkages due to their high flexibility which allows more favorable interactions with proteins. We developed the GEGOP program in order to be able to study the preferred conformations of large glycopeptides. The program is based on the GESA (geometry of saccharides) program and utilizes the HSEA (hard sphere exo anomeric) force field for the carbohydrate part and the ECEPP/2 (empirical conformation energy program for peptides) force field [Némethy, G., Pottle, M. S. & Scheraga, H. A. (1983) J. Phys. Chem. 87, 1883-1887] for the peptide part. The GEGOP program allows the simultaneous relaxation of all rotational degrees of freedom of these glycoconjugates during the energy optimization process. Thus, mutual interactions between glycosyl and amino acid residues can be studied in detail.