Files
Abstract
This dissertation focuses on the characterization of interactions of glycoproteins with glycosaminoglycans (GAGs) using primarily nuclear magnetic resonance (NMR) methodology. Glycoproteins are proteins carrying covalently linked glycans; many glycoproteins play crucial roles in human physiology and disease. Many function by interacting with other glycans, including the highly sulfated and structurally diverse glycans found in the extracellular matrix (GAGs). The characterization of these systems is best performed on properly glycosylated forms produced by the expression of the proteins in mammalian cell culture. However, mammalian protein characterization by traditional NMR methodology is challenging since the uniform isotopic labeling with isotopes needed for NMR observation (13C, 15N and 2H) becomes extraordinarily expensive and deuteration is very detrimental to cell growth. I describe an alternative methodology based on sparse labeling with single isotopically enriched amino acids. The primary limitation of sparse labeling is that the connectivities between isotopically labeled residues are lost. As a result, traditional triple resonance assignment approaches are no longer applicable. To overcome this obstacle, a new strategy is developed to assign the crosspeaks in a heteronuclear single quantum coherence (HSQC) spectrum of a sparsely labeled protein sample. This strategy uses a genetic algorithm to search for an optimal pairing of HSQC crosspeaks with labeled sites based on the experimental and predicted values of chemical shifts, nuclear Overhauser effects and residual dipolar couplings. This methodology has been validated on a set of previously assigned proteins and a sparsely labeled two-domain construct from a glycosylated signaling protein, Robo1-Ig1-2. Using available NMR assignments, I have characterized Robo1-Ig1-2 interacting with two heparan sulfate tetramers and an octamer using a series of high structural content NMR experiments. A model of this complex has been generated and used to rationalize how heparan sulfate may modulate interaction with Robols signaling partner, Slit2. This methodology was then applied to study the interaction between glycoprotein LAR and heparan sulfate, another glycoprotein-GAG interaction that is important for signal transduction. I present a model for LAR-Ig1-2 interacting with a particular heparan sulfate pentasaccharide, and use this to assess heparan sulfate modifications that may lead to enhanced binding and induced LAR dimerization.