Several diverse projects are being pursued. These are the major ones pursued during the past year. Using Enhanced Sampling Methods to Study Peptide Insertion in Membranes The insertion of peptides into lipid membranes is a crucial process in biology. The study of this process with molecular dynamics (MD) can provide atomic-level insight into this process. However typical MD simulations are unable to routinely reach the timescales needed to study insertion. We have employed a novel method for inserting peptides into membranes involving replica exchange molecular dynamics (REMD) coupled with high temperatures and strong electric fields. We are continuing to refine this method so it can be applied generally to different peptide/membrane combinations. Ligand modulation of ion channels. Hyperpolarization-activated cyclic nucleotide-gated (HCN) ion channels are expressed in the sinoatrial node, dorsal root ganglia and the basal ganglia. They play fundamental roles in electric signaling in nerve, muscle and synapse, but their function and gating mechanism are not completely understood. The overall goal of the project is to gain insight into the mechanism of the HCN channel activation upon binding of cyclic adenosine monophosphate (cAMP) and other small ligands to its intracellular C-terminal. Many mechanisms have been proposed for the opening motion propagation in the channel, but they do not completely explain the entire channel behavior. A novel theory states that upon cAMP binding, a part of the HCN C-terminal, called the C-helix, stabilizes its secondary structure and moves towards the binding pocket to make contacts with cAMP. Its movement is correlated with the opening conformational change of the channel pore. This theory is being tested using molecular dynamics simulations and self-guided Langevin dynamics (SGLD). The simulations enable sampling of conformations along this transition, giving insight into the occurring structural changes and ultimately into the HCN gating mechanism. Kinesin walking mechanism from SGLD simulations. Kinesin is a protein belonging to the class of Cytoskeletal motor proteins. Kinesin converts the energy of ATP hydrolysis into stepping movement along microtubules, which supports several vital cellular functions including mitosis, meiosis, and the transport of cellular cargo. Because kinesin is a fundamental protein, further research on the topic will provide important information as to how it functions. Combined with low resolution electron microscopic images, self-guided Langevin dynamics simulations are performed to study molecular motion and conformational change of kinesin motor domain in water and binding with microtubule. SGLD enable simulation to reach the time scale required for conformational change to understand the role of ATP binding and interaction with microtubules. Protein kinases are dynamic and can adopt many conformational states, including active, inactive, and intermediate states. These conformational states can represent an array of structural features that distinguish the ability of the protein to bind other molecules. Clarifying the transitions between the conformational states of protein complexes is critical for effective rational design, as it would allow deeper insights into the structure-function properties. In particular, the NF-B inducing kinase (NIK) and the inhibitor of B kinase-B (IKK), two protein kinases associated with inflammatory responses, were investigated. Molecular dynamics simulations were employed to consider how conformational changes and protein-protein interactions within multimeric assemblies are influenced by changes in the interacting subunits as well as the environment. Protein kinases complexed with small molecule activators or inhibitors were examined in the active, inactive, and mutant states to correlate structure-property and structure-function relationships as a function of intracellular ionic strength. Analyses indicate that the protein-protein interactions and the binding of small molecules are sensitive to changes in the ionic strength. The use of different force field parameters for monovalent and divalent ions were considered to further test the models. The results highlight that inconsistencies that require further investigation. Predicting partition coefficients Theoretical approaches are powerful tools to predict important physicochemical properties in biomolecular recognition processes as such predictive approaches can provide guidance in high-throughput screening for rational design. Although QSAR predictive models are well known for predicting various properties rapidly, the quality of these models are limited to the parametrization and poorly describe the properties coupled to changes in the electronic environment. Electronic structure methods are being used to predict the partition coefficients of a diverse set of several dozens of fragment- and drug-like molecules that resemble small molecule protein kinase inhibitors as a part of the SAMPL6 challenge. Building upon the lessons learned from the previous SAMPL challenges, additional efforts towards modeling electron correlation and solvation in diverse mediums are being investigated. Free energy calculations in a QM/MM model QM/MM model is also used besides QM/implicit solvent calculations in current SAMPL challenge to better account for the environmental effects from both aqueous and organic solvents. More advanced water models are used other than conventional TIP3P water model. Free energy calculations are expected to benefit from the new water models. For efficiency, QM/MM-NBB method is being used to connect the QM/MM surface to the MM surface at the endpoints to finally obtain the free energy difference in QM/MM level. Binding free energy calculations for the host-guest systems We are participating in the SAMPL6 blind challenge for the absolute binding free energy calculations of host-guest systems. This challenge involves two sets of host systems - Gibb Deep Cavity Cavitans (OA and TEMOA) and CB8 with 8 and 14 guest molecules respectively. For these calculations, the initial conformations are first obtained using docking. These candidate conformations are then refined by recalculating their energies with quantum chemical semi-empirical methods. Free energy calculations are performed using thermodynamic integration(TI) and Hamiltonian replica-exchange with Bennetts Acceptance Ratio(BAR) based on a serial insertion scheme. To avoid situations where derivative dU/dLambda would become discontinuous, potential for Lennard Jones transformations are treated via the soft-core potential. The binding free energies will be corroborated with experimental values. Protonation states in the selectivity filter of voltage activated Na channel The selectivity filter (SF) of bacterial voltage sensing sodium channels consists of four glutamate residues. Previous MD simulations of this channel were performed with fixed charges on the four Glu residues. We are using constant pH simulations to determine the most likely protonation state of the SF. The simulations show that the protonation state is coupled to the number of ions present in the pore of the channel. It is also suggestive that at physiological pH the channel could be in different protonation states. This result indicates a greater need for careful consideration of protonation states when looking at the conductivity of this channel.
Showing the most recent 10 out of 40 publications
