(493f) The Relationship of Glycoside Hydrolase Ligand Binding Free Energy to Processive Ability

In nature, organisms secrete synergistic glycoside hydrolase (GH) enzyme cocktails to deconstruct polysaccharides to sugars. The cocktails consist of multiple classes of processive, non-processive, and accessory enzymes that aid in substrate accessibility and reduce product inhibition. Non-processive enzymes attack the amorphous regions of polymer crystals and cleave glycosidic linkage once or twice to create accessible free chain ends; whereas, processive enzymes usually attach to chain ends and hydrolyze many glycosidic linkages in sequence to produce disaccharide units prior to dissociation. Processive enzymes are responsible for the majority of hydrolytic bond cleavages, and hence, are the most logical targets for activity improvements towards efficient and economical biomass conversion. However, the molecular-level mechanism and factors responsible for GH processivity still elude description. From structural data and analysis of reaction products it has been suggested that the presence of a long tunnel or deep enzyme active site cleft in GH indicates processivity and lack thereof indicates non-processivity. However, this strict delineation is not always accurate. We hypothesize that the processive GHs active site architecture contributes to relatively high ligand binding affinities and low substrate dissociation rates, while non-processive enzymes will demonstrate comparatively weak binding allowing the enzyme to more freely dissociate. To investigate our hypothesis, we have chosen the model GH system, Serratia marcescens Family 18 chitinases, which includes two processive chitinases, ChiA and ChiB, and a non-processive endo-chitinase, ChiC. In line with our hypothesis, we expect that ChiA and ChiB will exhibit stronger ligand binding than non-processive ChiC.  Free energies of binding a chito-oligosaccharide to each chitinase calculated using free energy perturbation with Hamiltonian replica exchange molecular dynamics support our hypothesis. Further, we find that interaction of the protein with pyranose rings in the substrate side of the binding cleft contribute the most to affinity. Comparison to experimental values obtained with Isothermal Titration Calorimetry reveals the formation of a catalytically active conformation significantly affects heat evolution, and thus affinity. Molecular dynamics simulations were used to characterize protein dynamics contributing to ligand binding affinity. Overall, our characterization of ligand binding affinity in S. marcescens Family 18 chitinases suggests that binding in the pre-hydrolytic substrate binding sites significantly contributes to affinity and likely processivity.