• Development of Additive and Polarizable Force Field Parameters for Polypeptides

      Zhu, Xiao; MacKerell, Alexander D., Jr. (2012)
      Simulations of biological macromolecules are continually improving due to the extension and continued optimization of empirical force fields based on molecular mechanics. The quality of this "balls-on-springs" model is likely to increase, though limited by the nature of molecular mechanics, as more experimental data become available allowing for further optimization of the underlying parameters. Classically, force fields employed a fixed-charge, or additive, scheme where partial atomic charges on atoms cannot change during a simulation such that changes in molecular dipole (i.e. explicit polarization) as a function of environment are lost. Here, we describe the development, optimization, and application of a polarizable CHARMM force field that explicitly accounts for molecular polarizability via the inclusion of atomic polarizability based on the classical Drude oscillator. We begin by presenting an overview of the history of the CHARMM empirical force field, its various components, applications, as well as the formalisms for the inclusion of polarizability in force fields. This is followed by development efforts as part of this thesis towards the completion of the polarizable force field for simulations of protein. In Chapter 3, a systematic study of the side-chain conformations was performed using quantum mechanical methods. Through comparisons with a large-scale survey of the protein crystal structures, a relationship was drawn between the intrinsic energetics and occurrences in protein structures. The intrinsic energetics were then used as target data for the optimization of side-chain torsion parameters, as discussed in Chapter 4; the data and optimization approach is likely to be of value to other empirical force field development communities. Because one of the advantages of the polarizable force field is more accurate description of the electrostatics, the performance of the force field in the calculation of pKas was undertaken. In Chapter 5 we present data for the development of parameters for alternative protonation states of the Cys, Lys, and Tyr side-chains, opening doors to simulations of these species within protein systems. Finally, in Chapter 6 we applied the polarizable force field in the calculation of pKas for select residues in RNase A, demonstrating the advantages and potential challenges with respect to future optimization of the Drude polarizable force field.
    • The use of conformational sampling in CHARMM protein force field optimization and ligand-based drug design

      Shim, Jihyun; MacKerell, Alexander D., Jr. (2013)
      Sampling of the conformational space of biomolecules in computer simulations allows researchers to investigate atomistic details of biological phenomena such as protein folding and ligand binding. Conformational sampling based on empirical energy functions depends on the force field and is aided by enhanced simulation methods. This thesis discusses conformational sampling methods and force fields, along with application of conformational sampling to force-field optimization and ligand-based drug design. Extensive conformational sampling was performed for small peptides and drug-like molecules using temperature replica-exchange and Hamiltonian replica-exchange molecular dynamics. Obtained conformational ensembles were then used to improve peptide-backbone and side-chain parameters in the CHARMM protein force fields, thereby yielding more accurate conformational properties. Obtained ensembles were also applied to ligand-based drug design where a novel method based on the conformationally sampled pharmacophore approach was used to identify quantitative structure-activity relationships (SARs) of μ opioid receptor ligands. Based on the SARs, we proposed ligand-binding orientations related to receptor activation. The binding orientations were further investigated using simulations of selected ligands bound to the 3-dimensional -opioid receptor structures. Our studies validate ligand-based SARs and show atomistic details of ligand-receptor interactions and the mechanism of µ opioid receptor activation.