Atomic-Resolution Experimental Structural Biology and Molecular Dynamics Simulations of Hyaluronan and Its Complexes

Abstract

This review summarizes the atomic-resolution structural biology of hyaluronan and its complexes available in the Protein Data Bank, as well as published studies of atomic-resolution explicit-solvent molecular dynamics simulations on these and other hyaluronan and hyaluronan-containing systems. Advances in accurate molecular mechanics force fields, simulation methods and software, and computer hardware have supported a recent flourish in such simulations, such that the simulation publications now outnumber the structural biology publications by an order of magnitude. In addition to supplementing the experimental structural biology with computed dynamic and thermodynamic information, the molecular dynamics studies provide a wealth of atomic-resolution information on hyaluronan-containing systems for which there is no atomic-resolution structural biology either available or possible. Examples of these summarized in this review include hyaluronan pairing with other hyaluronan molecules and glycosaminoglycans, with ions, with proteins and peptides, with lipids, and with drugs and drug-like molecules. Despite limitations imposed by present-day computing resources on system size and simulation timescale, atomic-resolution explicit-solvent molecular dynamics simulations have been able to contribute significant insight into hyaluronan's flexibility and capacity for intra- and intermolecular non-covalent interactions.

Keywords: NMR; conformation; crystallography; flexibility; hyaluronan; hyaluronate; hyaluronic acid; interaction; molecular dynamics.

Figure 1

Chemical structure of the glycosaminoglycan biopolymer hyaluronan. GlcA = β-D-glucuronate; GlcNAc = N-acetyl-β-D-glucosamine. The GlcA carboxylic acid moiety is expected to be deprotonated at physiological pH and is represented as such. Rotatable dihedrals in β1-4 and β1-3 glycosidic linkages are in red. The index n indicates the overall length of the polymer. The index i uniquely identifies each monosaccharide in the polymer. IUPAC conventions are used to define (ϕ₁₄, ψ₁₄) and (ϕ₁₃, ψ₁₃) and assign to each an index i [41]. For example, (ϕ₁₄, ψ₁₄)_i is defined by (O5_i-C1_i-O4_i−1-C4_i−1, C1_i-O4_i−1-C4_i−1-C3_i−1) and (ϕ₁₃, ψ₁₃)_i+1 by (O5_i+1-C1_i+1-O3_i-C3_i, C1_i+1-O3_i-C3_i-C2_i). Numbering of carbon atoms follows the convention for GlcNAc in the figure, with oxygen atoms assuming the number of the carbon to which they are attached. The ring oxygen is considered O5, as it is attached solely to C5 in the linear aldose form of the monosaccharide. Glycosidic linkage oxygen atoms assume the number of the attached carbon atom of the monosaccharide having the lower index i in the relevant disaccharide unit.

Figure 2

Conformational ensemble of hyaluronan 50-mers created by extrapolating from glycosidic linkage conformations sampled in molecular dynamics simulations of hyaluronan oligosaccharides. The star indicates a continuous linear stretch in one of the 50-mers. Reprinted with permission from Ref. [55]. Copyright 2006 Elsevier.

Figure 3

Radii of gyration (R_g) of modeled hyaluronan random coils of varying molecular mass (M_m) constructed using data from molecular dynamics simulations with either 0.2 M NaCl (white circles) or neutralizing sodium (blue circles) as compared to various experimental data (other symbols; see [118] for references to experimental data). Reprinted with permission from Ref. [118]. Copyright 2017 Elsevier.

Figure 4

Computed hyaluronan disaccharide conformational free energies compared with all available PDB hyaluronan structural data for (a) (ϕ₁₄, ψ₁₄) and (b) (ϕ₁₃, ψ₁₃) glycosidic linkage dihedrals. Free energy data, in kJ/mol and shown as contours, are adapted with permission from Figure S1 from reference [123], copyright 2021 American Chemical Society, and PDB data, shown as +’s, are from Table 2 in the present work. Free energies were computed from all-atom explicit-solvent molecular dynamics simulations of (a) GlcNAcβ1-4GlcA and (b) GlcAβ1-3GlcNAc disaccharides using the CHARMM force field.

Figure 5

Hairpin-like turns induced by Na⁺ binding to hyaluronan in atomic-resolution explicit-solvent molecular dynamics simulation. Panels (A–E) are successive frames from a single molecular dynamics trajectory demonstrating that Na⁺ binding either immediately precedes or coincides with turn formation. Na⁺ ions are drawn as orange spheres, and arrows point to the glucuronate residues where the events occur (“GCU37”, “GCU23”, “GCU29”). The elapsed simulation time, in nanoseconds, is in red font. Reprinted with permission from Ref. [126]. Copyright 2020 Elsevier.

Figure 6

Ensemble of computationally predicted binding poses for hyaluronan 4-mers (“HA dp4”) and 6-mers (“HA dp6”) with IL-10. Dynamic Molecular Docking (DMD) [165] was used to generate the ensemble. DMD is an atomic-resolution explicit-solvent molecular dynamics method where full flexibility of both the protein and ligand are allowed throughout a molecular dynamics trajectory during which the ligand is gradually pulled toward the protein. Reprinted with permission from Ref. [154]. Copyright 2015 Elsevier.

Figure 7

Molecular dynamics snapshot of a hyaluronan oligosaccharide interacting with the C-terminal region of Streptococcus equisimilis hyaluronan synthase. Probability densities computed from molecular dynamics snapshots show that R406 and R413 sidechains interact strongly with hyaluronan while the K414 sidechain interaction is weak. Reprinted with permission from reference [175], copyright 2017 American Chemical Society.

Figure 8

”Picket fence” on the cell membrane created by a single molecule of extracellular hyaluronan participating in polyvalent binding with multiple transmembrane CD44 molecules. Reprinted with permission from Ref. [43]. Copyright 2018 Elsevier.