Multiscale modeling of ion exchange mechanism in the ClC antiporter /

Saved in:
Bibliographic Details
Author / Creator:Lee, Sang Yun, author.
Ann Arbor : ProQuest Dissertations & Theses, 2016
Description:1 electronic resource (135 pages)
Format: E-Resource Dissertations
Local Note:School code: 0330
URL for this record:
Hidden Bibliographic Details
Other authors / contributors:University of Chicago. degree granting institution.
Notes:Advisors: Gregory A. Voth Committee members: Aaron Dinner; Suri Vaikuntanathan.
Dissertation Abstracts International, Volume: 77-10(E), Section: B.
Summary:Ion channels or transporters are transmembrane proteins that conduct ions through a gated, water-filled pore between the intracellular and the extracellular environment. Such ion flux play a crucial role in controlling a wide range of biological functions, ranging from cell signaling, osmotic stress response to muscle contraction. Ion fluxes in some channel proteins are modulated by ligand binding, and the binding sites for the ligands are the attractive potential targets for diverse drugs, such as anesthetics, antiepileptics, and analgesics. Channels are distinguished from transporters depending on the need of energy input for transporting ions. In channels, ions diffuses down their electrochemical gradient through the membrane (passive transport), however in transporters, the uphill movement of ions against the gradient is driven by other energy sources from either the chemical energy, such as ATP hydrolysis (primary active transport), or the movement of other ions down the gradient (secondary active transport). ClC-ec1 is a bacterial transporter which mediates the uphill movement of H+ using the concentration gradient of Cl- through the membrane, and vice versa.
The physical structures of the proteins are the key source to understand their functions. Due to the recent progress in experimental techniques of recombinant expression, purification and crystallization, high-resolution crystal structures of ClC proteins have been revealed in several species. In addition to structural studies, the electrophysiological experiments along with site-directed mutation applied to some important residues provided some information for understanding how proteins work during the ion transport. However, despite the success accumulated over the past few decades, experimental approaches have some limitations, especially in elucidating all microscopic details. The system goes through multiple intermediate states which accompany conformational changes of the protein, for example, the channel pore controls the ion conduction by changing between open and close structures. However, in many cases, the crystal structures have been reported for conformations of only a part of all possible intermediate states. Another limitation in site-directed mutagenesis is the difficulty to identify the function of the target residues when they are involved in multiple intermediate steps and the modification in each step leads to a loss of the function.
All-atom molecular dynamics (MD) simulation is a useful tool to override such difficulties because one can generate trajectories with realistic models (force field) of all components in the system, such as protein, lipid membrane, ions and solvents, and directly monitor the ion permeation process and associated conformational changes of the protein. Although MD provides microscopic details of the motions of all atoms, it generally has some limitations in the classical force field and the sampling time of the simulation. Due to its simplicity, classical force field cannot describe neither any chemical reactions nor the effect of electronic polarization. The chemical bonds in classical force field are generally modeled by harmonic potentials as a function of bond distance, which do not allow any bond dissociations. All atoms are treated as point charged particles to evaluate the electrostatic interaction in classical force field, however the interactions between higher order multipoles are not considered. These two limitations should be treated more explicitly in H+ and Cl- transport in ClC-ec1. The movement of H+ through the hydrated cavity in the protein should be considered as a reactive process, where the excess proton migrates by Grotthuss shuttling mechanism between water molecules. The effect of induced polarization can be more significant for Cl- in the protein than in the bulk solution, where the protein interior is generally in low dielectric environment.
The hybrid QM/MM simulation, which is the combination of quantum mechanics (QM) and empirical force field (MM, molecular mechanics) is a well-known method to simulate chemical reactions in a large biological system. The formation and the cleavage of chemical bonds are described at the core region by QM, and the surrounding region, which is chemically inert, is treated at the classical force field level. However, the simulation time accessible by the current computing power is typically on a time scale of picoseconds, when a DFT-level QM/MM method is applied in such large system. The time scale of the permeation of one ion through the protein is approximately in the order of milliseconds. In order to access the dynamics of long time scale, enhanced sampling techniques, such as umbrella sampling or metadynamics, are employed, which apply the bias potential on the pre-determined reaction coordinate and explore the free energy landscape in more efficient way. However, the success of the enhanced sampling depends on the converged sampling of other degrees of freedom orthogonal to the reaction coordinate. Although the reaction coordinate is well chosen for the enhanced sampling, the QM/MM MD may not be feasible if other degrees of freedom are on a time scale slower than picoseconds.
The MS-RMD (multi-scale reactive molecular dynamics) model has been developed in the group for a long time. The model explicitly describes the deprotonation reaction of amino acids. It enables more efficient sampling without much loss of accuracy, compared to the QM/MM MD. The model parameters were obtained by the force fitting algorithm (FitRMD), which finds the best fit to the reference data, which is the forces calculated by a DFT-level QM/MM using configurations sampled around the transition state. The resulting MS-RMD model was shown to faithfully reproduce the free energy profiles (potential of mean forces, PMF) of the reference QM/MM Hamiltonian for H+ transport. The MS-RMD MD allowed us to calculate the two dimensional PMFs at an affordable computational cost. The two dimensional PMFs provided a more detailed information about other important degrees of freedom, strongly coupled with H+ transport, such as the increase of the hydration level in H + pathway and the changes of the orientation of the surrounding residues.