Molecular dynamics simulations of O-linked Oligosaccharides from respiratory-mucus glycoproteins

Abstract

Mucins belong to the class of glycoproteins that contain oligosaccharide chains in which the N-acetylgalactosamine residue is covalently linked to the hydroxyl oxygen atom in the side chain of serine or threonine on the polypeptide chain. These O-linked oligosaccharides (O-glycans) constitute 85% to 90% of mucins and are characterized by molecular weights ranging from 1 x 10^6 to 15 x 10^6, depending on the source. The mucin oligosaccharides appear to be polydisperse, linear, and flexible threads and show extreme microheterogeneity. O-glycans in mucins are densely packed, as a result of which the mucin molecules assume an extended rod-like configuration and inhibit binding and helix formation of the polypeptide chain. The rod-like linear shape of the mucin molecule is believed to contribute to the major physicochemical property of mucins, namely, their viscoelasticity. The viscous and viscoelastic properties imparted to these O-glycans manifest in the protective and lubricative properties of mucins. Further, these molecules are also involved in a number of biological functions, such as virus infectivity, fertilization, and in the control of the immune system. They bind to pathogenic bacteria, parasites, and toxins and, at times, have been found to mimic the natural binding sites of pathogens, thus preventing the pathogens from binding to the mucosa.

The oligosaccharides in mucins may occur as neutral or acidic oligosaccharides. The sugar residues commonly found in O-glycans are N-acetyl-?-D-galactopyranose (GalNAc), ?-D-galactose (Gal), N-acetyl-?-D-glucosamine (GlcNAc), ?-L-fucose (Fuc), and sialic acid (Sia). Gal and GlcNAc residues may also be sulfated or phosphorylated, and thereby with Sia, contribute to the acidity of some mucin-type oligosaccharides. Generally, these sugar chains are terminated by the Fuc residue in a (1?2/3/4) and/or Sia in a (2?3/6) linkage. The O-glycan chains generally have Lewis blood group determinants at the nonreducing end. O-glycans are classified into six types based on the residues and the linkages in their core structure (designated as Core type 1, type 2, ..., type 6). Of these six Core types identified so far, Core type 4 (characterized by the trisaccharide sequence GlcNAc?(1?6)[GlcNAc?(1?3)]GalNAc) is mostly found in the human respiratory tracts.

A detailed knowledge of the conformational flexibility of oligosaccharides is an important step towards the understanding of the biological functions of these molecules. N-glycans (yet another class of oligosaccharides found in glycoproteins which are covalently linked to the nitrogen atom on the side chain of asparagine residue on the polypeptide chain) have been studied in some detail using computational and experimental methods. Till date, only the conformations of partial structures, such as the Lewis blood group determinants, that occur in O-glycans have been studied using both computational and experimental techniques.

In the present work, an attempt has been made to understand the conformational flexibility of O-glycans. All the O-glycans studied contain the trisaccharide sequence GlcNAc?(1?6)[GlcNAc?(1?3)]GalNAc (characteristic of the O-glycan Core type 4) and also have Lewis blood group determinants (Leb, Lex, and Ley) at the nonreducing end of the chain. The O-glycans extracted from the sputum of a patient (Blood group O) suffering from bronchiectasis due to Kartagener’s syndrome (Klein et al., 1991, Eur. J. Biochem. 198, 151) range from seven to nine residues in length. The glycosidic linkage and sequence information have been elucidated from chemical degradation, HPLC, fast-atom-bombardment mass spectrometry, and NMR methods. However, no 3-dimensional information is available for these large O-linked oligosaccharides.

In order to obtain insight into the conformation and dynamics of Core type 4 O-glycans, a number of these oligosaccharides have been subjected to Molecular Dynamics (MD) simulations. The list of oligosaccharide fragments investigated is given in the accompanying figure. The simulations, the analysis of the results, and discussions of the conformations of these oligosaccharides are the subject matter of this thesis.

The thesis is divided into 8 chapters. A general introduction to the types of oligosaccharides occurring in glycoproteins, their structural features, classification, importance of oligosaccharides in general, and O-glycans in particular, as well as different experimental and computational methods available for the study of these molecules, forms the content of Chapter 1.

The methodology of generating the initial structures, the minimization and molecular dynamics procedure details are discussed in Chapter 2.

The main aim of the present work is twofold: (i) Characterization of the structure and dynamics of O-glycans listed (Figure), and (ii) Analysis of the conformational behavior of smaller fragments in isolation and when they are part of bigger oligosaccharides. Hence, not only the simulations of the oligosaccharides but also those of the constituent disaccharides, Lewis blood group fragments, etc., have been performed. The MD simulation results of different disaccharide segments found in the selected O-glycan structures form the content of Chapter 3. Using the coordinate data obtained from the MD simulations, various conformational parameters (such as the glycosidic torsion angles, pyranose ring pucker parameters, conformations of the CH2OH group, conformations of the N-acetamido group in GlcNAc and GalNAc residues, hydrogen bond interactions) have been analyzed. These simulations show that the pyranose rings are flexible but, however, do not undergo significant transitions from one chair conformation to the other. The N-acetamido group in GlcNAc and GalNAc residues mostly assume conformations characterized by values for the torsion angle (C2 — C2 — N — H) around 180°. However, in certain simulations, a value for C(H2 — C2 — N — H) around 0° is also observed. The mean and root mean square (rmsd) values of the glycosidic torsion angles for the various disaccharide linkages have been computed. The respective disaccharide MD mean values for the glycosidic torsion angles ((?,?)) from these disaccharide simulations have been used to generate the initial structures of bigger fragments.

Chapter 4 presents the simulation results of Lewis blood group tetrasaccharide (Lea and Lex) and pentasaccharide (Leb and Ley) fragments. These simulations show that the Lewis blood group structures are fairly rigid and assume single conformations. While most of the interunit glycosidic torsion angles in these structures agree well with the values obtained from the MD simulations of respective disaccharide segments, a few exceptions are also found. For example, the segment Gal?(1?4)GlcNAc has values for the conformational angles (?,?) around (39°, -10°) in the isolated disaccharide simulation, while the values for this segment are around (-25°, -20°) in Lex simulation and around (63°, 13°) in Ley simulation. Similarly, the conformational angles of the segment Fuc?(1?3)GlcNAc are around (30°, -10°) in the isolated disaccharide simulation, while these are around (-24°, -22°) and (-32°, -26°) for this segment in Lex and Ley simulations respectively.

The behavior of the O-glycan Core type 4 fragment is studied using one trisaccharide sequence GlcNAc?(1?6)[GlcNAc?(1?3)]GalNAc, and two tetrasaccharide sequences: (i) GlcNAc?(1?6)[Gal?(1?4)GlcNAc?(1?3)]GalNAc and (ii) Gal?(1?4)GlcNAc?(1?6)[GlcNAc?(1?3)]GalNAc. The MD results of the Core type 4 tri- and tetrasaccharide sequences are discussed in Chapter 5. In these and other molecules discussed in Chapters 6 and 7, three different starting conformations of the (1?6) linkage corresponding to the torsion angle x equal to +60°, -60°, and 180° were considered for simulations. The simulation data are analyzed for the flexibility of the various glycosidic torsion angles in these fragments. The analysis shows that the Core type 4 fragments assume different conformations during the simulation period and their shapes range from an extended to a compact one. The compact shapes are often stabilized by hydrogen bond interactions.

The oligosaccharides listed in Figure I are divided into Set-I and Set-II depending on the absence or presence of Fuc?(1?2) linkage on the Gal3 residue. The results of Set-I oligosaccharides are presented in Chapter 6. The conformational behavior of smaller fragments in isolation and when they are present in the bigger oligosaccharides are extensively analyzed. For this purpose, a number of snapshots from the MD trajectories were selected and subjected to potential energy minimization. The various glycosidic torsion angles in the energy minimized structures were plotted in the corresponding disaccharide (?,?)-maps and the distributions of the conformations in the various (?,?)-maps were examined. This analysis showed minor shifts in the conformations accessed (which still lie well within the generally allowed region) in certain (?,?)-maps. The analysis of 3-dimensional shapes of Set-I oligosaccharides during the MD simulations reveals that these oligosaccharides assume compact conformations for most of the time during the simulation period. Also, the compact conformations are stabilized by stacking interactions between sugar residues and hydrogen bonding interactions. However, occasionally the Set-I oligosaccharides have also been found to assume extended shapes during the MD simulations. The behavior of the Lewis blood group fragment (i.e., Leb) when it is part of large structures is also examined.

Chapter 7 discusses the results of MD simulations of Set-II oligosaccharides. Analysis of the flexibility of the different glycosidic linkages in Set-II oligosaccharides is carried out by plotting the conformations corresponding to the various disaccharide segments in these oligosaccharides from the energy minimized snapshot structures. It is seen that the presence of Fuc?(1?2) linked to the Gal residue in the vicinity of the (1?6) linkage plays an important role in the conformational flexibility of Set-II oligosaccharides. This role could be: (i) to affect the flexibility of the (1?6) linkage; (ii) to shift the conformations accessed by certain disaccharide segments (in the allowed regions of the (?,?)-map) to regions that were found to be different in the Set-I oligosaccharides. In other words, the presence of Fuc?(1?2) linked to the Gal residue near the (1?6) linkage affects the overall conformation of the Set-II oligosaccharides. This is clearly seen from the 3-dimensional shapes assumed by the Set-II oligosaccharides during the MD simulations. Set-II oligosaccharides generally assume an extended conformation during most of the simulation period and are devoid of inter-residue hydrogen bonds. Compact conformations, however, are occasionally accessed during the simulations. The flexibility of Lewis blood group (Le6, Lex, and Ley) determinants has been looked into when they are part of large oligosaccharides, and observations are listed.

Chapter 8 summarizes the overall results obtained from the (vacuum) MD simulations of O-linked oligosaccharides from the respiratory mucus glycoproteins. This is followed by the list of references cited in the various chapters of the thesis.