LAUSR

GAGs exhibit a high level of conformational and configurational diversity, which remains untapped in terms of the recognition and modulation of proteins. Although GAGs are suggested to bind to more than 800 biologically important proteins, very few therapeutics have been designed or discovered so far. A key challenge is the inability to identify, understand and predict distinct topologies accessed by GAGs, which may help design novel protein-binding GAG sequences. Recent studies on chondroitin sulfate (CS), a key member of the GAG family, pinpointing its role in multiple biological functions led us to study the conformational dynamism of CS building blocks using molecular dynamics (MD). In the present study, we used the all-atom GLYCAM06 force field for the first time to explore the conformational space of all possible CS building blocks. Each of the 16 disaccharides was solvated in a TIP3P water box with an appropriate number of counter ions followed by equilibration and a production run. We analyzed the MD trajectories for torsional space, inter- and intra-molecular H-bonding, bridging water, conformational spread and energy landscapes. An in-house phi and psi probability density analysis showed that 1→3-linked sequences were more flexible than 1→4-linked sequences. More specifically, phi and psi regions for 1→4-linked sequences were held within a narrower range because of intra-molecular H-bonding between the GalNAc O5 atom and GlcA O3 atom, irrespective of sulfation pattern. In contrast, no such intra-molecular interaction arose for 1→3-linked sequences. Further, the stability of 1→4-linked sequences also arose from inter-molecular interactions involving bridged water molecules. The energy landscape for both classes of CS disaccharides demonstrated increased ruggedness as the level of sulfation increased. The results show that CS building blocks present distinct conformational dynamism that offers the high possibility of unique electrostatic surfaces for protein recognition. The fundamental results presented here will support the development of algorithms that help to design longer CS chains for protein recognition.