Several diverse projects are being pursued. These are the major ones pursued during the past year. pH dependance of a Na channel Na channels exhibit a decrease in conductance with lowering of pH. The Selectivity filter (SF) of bacterial Na channels consists of four glutamate residues. The protonation states in the SF has been shown to modulate the number of bound Na+ ions, and most likely the conductivity. We study the protonation states in the SF of a bacterial Na channel through molecular dynamics, free energy perturbation as well as constant pH simulations. The simulations show that the number of bound ions influences the protonation state of the filter. With 2 or 3 ions bound, at physiological pH, the SF is most likely in fully deprotonated state, and possibly also the singly protonated state. With 1 or 0 ions bound, the doubly protonated state can also get populated. The protonated states exhibit a level of structural asymmetry as opposed to fully deprotonated state of the SF. The crystal structures of the inactivated state are asymmetric, while those of the open channel are symmetric. We speculate that protonation of the SF, possibly triggered by the lowering of the Na concentration at the mouth of the channel could be involved in inactivation of the channel. We are currently investigating the pH dependence of a bacterial ion channel (NavMs) which is known to undergo structural changes between an open and closed conformation. To ensure we have sufficient structural sampling we are employing constant pH simulations combined with replica exchange MD on GPUs using the Amber MD software package. Protein kinases are dynamic and can adopt many conformational states, including active, inactive, and intermediate states which can represent an array of structural features that distinguish the ability of the protein to bind other molecules. Illuminating 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. Improper signaling of the nuclear factor-B (NF-B) pathway plays a critical role in many inflammatory disease states including cancer, stroke, and viral infections. Although the signaling pathways are known, how these molecular mechanisms respond to changes in the intracellular microenvironment such as pH, ionic strength, and temperature, remains elusive. Molecular dynamics simulations were employed to differentiate the structural dynamics of the NF-B Inducing Kinase (NIK), a protein kinase responsible for invoking the non-canonical NF-B pathway, in its native and mutant form, and in the absence and presence of salt concentration in efforts to probe whether changes in the ionic environment stabilize or destabilize the NIK dimer. Analyses of structureactivity and conformational-activity relationships indicate that the proteinprotein interactions are sensitive to changes in the ionic strength. Ligand binding pockets as well as regions between the oligomer interface either compress or expand, affecting both local and distal intermolecular interactions that result in stabilization or destabilization in the protein assembly. The use of different force field parameters for monovalent and divalent ions are being considered to further test the models. SAMPL6 prediction on CB8 binding free energies Binding free energy calculations for the cucurbit8uril (CB8) hostguest systems.The ability to accurately predict the binding free energies of drug-like small molecules to biological macromolecules would accelerate drug discovery and development, enabling in silico testing of binding affinity, selectivity and off-target interactions of leads. We participated in the Statistical Assessment of the Modeling of Proteins and Ligands 6 (SAMPL6) CB8 host-guest binding affinity prediction challenge which provides a unique platform for validating computational methods for predicting binding free energies of small molecules. For SAMPL6, eleven guests are included in the main challenge and three guests are offered as bonus cases. Learning from our experiences on previous SAMPL challenges, in the past year we focused on understanding the effects of two major sources of error 1) force-field parameters and 2) free energy simulation methods. Utilizing the classical molecular mechanics potential energy functions used in Chemistry at HARvard Molecular Mechanics (CHARMM), we generated parameters specific for each of host and guest molecules. Quantum mechanical implicit solvent calculations and quantum mechanical force matching were used to determine non-bonded (partial atomic charges) and bonded terms, respectively. Free energy calculations were carried out using the double-decoupling method (DDM) combined with Hamiltonian replica exchange method (HREM) and Bennett acceptance ratio (BAR) (method 1) and the umbrella sampling (US) method used with weighted histogram analysis method (WHAM) (method 2). In addition, a quantum mechanical/molecular mechanical free energy correction scheme was employed to achieve improvement over classical additive molecular mechanical force field parameters. The result obtained was satisfactory: We submitted two independent sets of predicted values. Our submissions for bonus cases were the top two in terms of error with respect to experiment and the overall correlations (including both main and bonus cases) between the predicted values and the experimental results were also top two of all submissions (36 in total). QM/MM binding free energy calculations of host-guest systems: We partook in the SAMPL6 blind challenge for the absolute binding free energy of several guest complexes to host Cucurbit8uril. Each of the 14 guest molecules (three of which was bonus cases, including the colorectal cancer drug oxaliplatin) was modeled by fitting charge distributions from QM implicit solvent models and parameterizing intramolecular degrees of freedom via QM force matching on classically generated ensembles. Following classically driven binding free energy calculations, corrections to ascertain QM/MM binding free energies were performed. The corrected results demonstrated some of the highest correlation to experimental values amongst both the submissions with only the standard guest set (11 guest molecules, 36 submissions) and the bonus guest set (14 guest molecules, 6 submissions). Of particular note is the result for oxaliplatin, a square planar Pt(II) complex presenting a significant challenge to describe classically, which was the most accurate prediction obtained (e.g., with a deviation from experiment of 1.4 kcal/mol). Structure and dynamics of human islet amyloid polypeptide. Islet amyloid polypeptide (IAPP), is a 37-long peptide that is the main constituent of amyloid aggregates of type-II diabetes. Certain species acquire the disease, whereas others dont. Human (hIAPP) and cat IAPP have been shown to aggregate, but rat and pig do not. Starting from the solid-state NMR structure for hIAPP, we perform single-point mutations towards the other species and employ molecular dynamics simulations on the resulting structures. By analyzing the dynamics of each structure, we infer the relative contributions of mutations on different structural elements, and establish which has a dominant modulating effect. Results from this study will hint towards possible novel therapeutics for the treatment of type-II diabetes.
Showing the most recent 10 out of 40 publications
