Understanding Molecular Recognition: Thermodynamics and Binding Kinetics of Potent Thrombin Inhibitors
Khayat, Maan T.
MetadataShow full item record
Molecular recognition is the process of selective, reversible non-covalent binding between two molecules to form a complex. Numerous biological processes and cell pathways are precisely regulated and controlled by molecular recognition. Understanding molecular recognition at a fundamental level also translates into the challenge of being able to design ligands/drugs/small molecules that bind to specific sites on target biological macromolecules, most commonly proteins. Computer algorithms have been developed and utilized in attempt to predict ligand binding orientations and affinity with the aim of assisting rational drug design efforts. The accuracy of these algorithms has been generally limited because the predicted binding affinity typically only considers the individual non-covalent interactions, such as hydrogen bonding and van der Waals interactions, between the molecule and the active site in an additive fashion. Ligand functional group cooperativity, enthalpy-entropy compensation, and the detailed energetic participation of water molecules is often not included. Also the affect of the ligand conformation in solution before binding, and the kinetics of binding and dissociation, are typically not measured or considered. This thesis aims to address some of these deficiencies. The results presented in Chapter 1 demonstrate an apparent cooperativity among interacting groups with a series of thrombin ligands that provides an improvement in the binding affinity from a nanomolar level to a picomolar level. The demonstration of this level of apparent cooperativity can be instructive for the design of super potent ligands more generally. The results presented in Chapter 2 provide a kinetic analysis of the k on and k off for the picomolar thrombin sulfonamide inhibitor 9 and a series of close analogs. Higher affinity binding is usually achieved mainly by slower dissociation rates. Surprisingly the picomolar binding affinity of inhibitor 9relative to the nanomolar binding affinity of the close analogs, was mainly due to a three order of magnitude increase in k on (to a diffusion controlled binding rate) rather than a slower k off . Crystallographic studies showed that the thrombin bound conformations for these inhibitors are essentially the same. Isothermal calorimetry measurements showed that the three order of magnitude improvement in K dis largely due to an improvement in the entropy of binding. Detailed NMR studies of the ligands in aqueous buffer indicated that the accelerated kon is due to preformed conformations that are very close to the bioactive conformation within the protein cavity, along with an electrostatic directed binding into the S1 pocket. These structure-kinetic relationship results provide an important new insight into designing super potent ligands for target proteins. The results presented in Chapter 3 are focused on the contribution of rigidity in proline-containing thrombin inhibitors to binding affinity. Moreover, an investigation into how the proline moiety contributes to the picomolar activity of 9 is provided. We concluded that proline is an essential component for the potent activity of inhibitor 9 . It provides rigidity to the molecule that helps stabilize the H-bonds formed between the inhibitor and thrombin and reduces the conformational entropy loss upon binding. The results presented in Chapter 4 provide an extended analysis of the picomolar activity found in 9 by utilizing a variety of additional analogs. Chapter 5 presents an analysis of the effect of introducing a pyridyl cation into thrombin inhibitors for binding in the S1 subsite. The results indicated that a desolvation penalty reduces the binding affinity of these ligands relative to what they would likely be in the absence of this penalty. Overall, the results obtained from these projects provide a better understanding of the underlying interactions involved in biological molecular recognition. The data obtained may be valuable in rational drug design by providing a more accurate prediction of binding affinities.