Publications

A signal property of connexin channels is the ability to mediate selective diffusive movement of molecules through plasma membrane(s), but the energetics and determinants of molecular movement through these channels have yet to be understood. Different connexin channels have distinct molecular selectivities that cannot be explained simply on the basis of size or charge of the permeants. To gain insight into the forces and interactions that underlie selective molecular permeation, we investigated the energetics of two uncharged derivatized sugars, one permeable and one impermeable, through a validated connexin26 (Cx26) channel structural model, using molecular dynamics and associated analytic tools. The system is a Cx26 channel equilibrated in explicit membrane/solvent, shown by Brownian dynamics to reproduce key conductance characteristics of the native channel. The results are consistent with the known difference in permeability to each molecule. The energetic barriers extend through most of the pore length, rather than being highly localized as in ion-specific channels. There is little evidence for binding within the pore. Force decomposition reveals how, for each tested molecule, interactions with water and the Cx26 protein vary over the length of the pore and reveals a significant contribution from hydrogen bonding and interaction with K(+). The flexibility of the pore width varies along its length, and the tested molecules have differential effects on pore width as they pass through. Potential sites of interaction within the pore are defined for each molecule. The results suggest that for the tested molecules, differences in hydrogen bonding and entropic factors arising from permeant flexibility substantially contribute to the energetics of permeation. This work highlights factors involved in selective molecular permeation that differ from those that define selectivity among atomic ions.

Aberrant opening of nonjunctional connexin hemichannels at the plasma membrane is associated with many diseases, including ischemia and muscular dystrophy. Proper control of hemichannel opening is essential to maintain cell viability and is achieved by physiological levels of extracellular Ca(2+), which drastically reduce hemichannel activity. Here we examined the role of conserved charged residues that form electrostatic networks near the extracellular entrance of the connexin pore, a region thought to be involved in gating rearrangements of hemichannels. Molecular dynamics simulations indicate discrete sites for Ca(2+) interaction and consequent disruption of salt bridges in the open hemichannels. Experimentally, we found that disruption of these salt bridges by mutations facilitates hemichannel closing. Two negatively charged residues in these networks are putative Ca(2+) binding sites, forming a Ca(2+)-gating ring near the extracellular entrance of the pore. Accessibility studies showed that this Ca(2+)-bound gating ring does not prevent access of ions or small molecules to positions deeper into the pore, indicating that the physical gate is below the Ca(2+)-gating ring. We conclude that intra- and intersubunit electrostatic networks at the extracellular entrance of the hemichannel pore play critical roles in hemichannel gating reactions and are tightly controlled by extracellular Ca(2+) Our findings provide a general mechanism for Ca(2+) gating among different connexin hemichannel isoforms.

Abnormal alteration of bone morphogenetic protein (BMP) signaling is implicated in many types of diseases including cancer and heterotopic ossifications. Hence, small molecules targeting BMP type I receptors (BMPRI) to interrupt BMP signaling are believed to be an effective approach to treat these diseases. However, lack of understanding of the molecular determinants responsible for the binding selectivity of current BMP inhibitors has been a big hindrance to the development of BMP inhibitors for clinical use. To address this issue, we carried out in silico experiments to test whether computational methods can reproduce and explain the high selectivity of a small molecule BMP inhibitor DMH1 on BMPRI kinase ALK2 vs. the closely related TGF-beta type I receptor kinase ALK5 and vascular endothelial growth factor receptor type 2 (VEGFR2) tyrosine kinase. We found that, while the rigid docking method used here gave nearly identical binding affinity scores among the three kinases; free energy perturbation coupled with Hamiltonian replica-exchange molecular dynamics (FEP/H-REMD) simulations reproduced the absolute binding free energies in excellent agreement with experimental data. Furthermore, the binding poses identified by FEP/H-REMD led to a quantitative analysis of physical/chemical determinants governing DMH1 selectivity. The current work illustrates that small changes in the binding site residue type (e.g. pre-hinge region in ALK2 vs. ALK5) or side chain orientation (e.g. Tyr219 in caALK2 vs. wtALK2), as well as a subtle structural modification on the ligand (e.g. DMH1 vs. LDN193189) will cause distinct binding profiles and selectivity among BMP inhibitors. Therefore, the current computational approach represents a new way of investigating BMP inhibitors. Our results provide critical information for designing exclusively selective BMP inhibitors for the development of effective pharmacotherapy for diseases caused by aberrant BMP signaling.

Gated solely by activity-induced changes in intracellular calcium, small-conductance potassium channels (SKs) are critical for a variety of functions in the CNS, from learning and memory to rhythmic activity and sleep. While there is a wealth of information on SK2 gating, kinetics, and Ca(2+) sensitivity, little is known regarding the regulation of SK2 subcellular localization. We report here that synaptic SK2 levels are regulated by the E3 ubiquitin ligase UBE3A, whose deficiency results in Angelman syndrome and overexpression in increased risk of autistic spectrum disorder. UBE3A directly ubiquitinates SK2 in the C-terminal domain, which facilitates endocytosis. In UBE3A-deficient mice, increased postsynaptic SK2 levels result in decreased NMDA receptor activation, thereby impairing hippocampal long-term synaptic plasticity. Impairments in both synaptic plasticity and fear conditioning memory in UBE3A-deficient mice are significantly ameliorated by blocking SK2. These results elucidate a mechanism by which UBE3A directly influences cognitive function.

Computational methodologies that couple the dynamical evolution of a set of replicated copies of a system of interest offer powerful and flexible approaches to characterize complex molecular processes. Such multiple copy algorithms (MCAs) can be used to enhance sampling, compute reversible work and free energies, as well as refine transition pathways. Widely used examples of MCAs include temperature and Hamiltonian-tempering replica-exchange molecular dynamics (T-REMD and H-REMD), alchemical free energy perturbation with lambda replica-exchange (FEP/lambda-REMD), umbrella sampling with Hamiltonian replica exchange (US/H-REMD), and string method with swarms-of-trajectories conformational transition pathways. Here, we report a robust and general implementation of MCAs for molecular dynamics (MD) simulations in the highly scalable program NAMD built upon the parallel programming system Charm++. Multiple concurrent NAMD instances are launched with internal partitions of Charm++ and located continuously within a single communication world. Messages between NAMD instances are passed by low-level point-to-point communication functions, which are accessible through NAMD's Tcl scripting interface. The communication-enabled Tcl scripting provides a sustainable application interface for end users to realize generalized MCAs without modifying the source code. Illustrative applications of MCAs with fine-grained inter-copy communication structure, including global lambda exchange in FEP/lambda-REMD, window swapping US/H-REMD in multidimensional order parameter space, and string method with swarms-of-trajectories were carried out on IBM Blue Gene/Q to demonstrate the versatility and massive scalability of the present implementation. Published by Elsevier B.V.

Presented is a polarizable force field based on a classical Drude oscillator framework, currently implemented in the programs CHARMM and NAMD, for modeling and molecular dynamics (MD) simulation studies of peptides and proteins. Building upon parameters for model compounds representative of the functional groups in proteins, the development of the force field focused on the optimization of the parameters for the polypeptide backbone and the connectivity between the backbone and side chains. Optimization of the backbone electrostatic parameters targeted quantum mechanical conformational energies, interactions with water, molecular dipole moments and polarizabilities, and experimental condensed phase data for short polypeptides such as (Ala)(5). Additional optimization of the backbone phi, psi conformational preferences included adjustments of the tabulated two-dimensional spline function through the CMAP term. Validation of the model included simulations of a collection of peptides and proteins. This first generation polarizable model is shown to maintain the folded state of the studied systems on the 100 ns time scale in explicit solvent MD simulations. The Drude model typically yields larger RMS differences as compared to the additive CHARMM36 force field (C36) and shows additional flexibility as compared to the additive model. Comparison with NMR chemical shift data shows a small degradation of the polarizable model with respect to the additive, though the level of agreement may be considered satisfactory. However, the polarizable model shows improvement for the residues with significantly underestimated S-2 order parameters in the additive model. Analysis of dipole moments associated with the peptide backbone and tryptophan side chains show the Drude model to have considerably larger values than those present in C36, with the dipole moments of the peptide backbone enhanced to a greater extent in sheets versus helices and the dipoles of individual moieties observed to undergo large variations during the MD simulations. Although there are still some limitations, the presented model, termed Drude-2013, is anticipated to yield a molecular picture of peptide and protein structure and function that will be of increased physical validity and internal consistency in a computationally accessible fashion.

The accuracy of empirical force fields is critical for meaningful molecular dynamics simulations of concentrated ionic solutions. Current models are typically developed on the basis of single ion properties such as the monohydrate energy in the gas phase, or the absolute hydration free energy at infinite dilution. However, the failure of these models to represent accurately the properties of concentrated solutions cannot be excluded. Here, these issues are illustrated for a polarizable potential based on classical Drude oscillators. To model accurately concentrated ionic solutions, the parameters of the potential functions are optimized to reproduce osmotic pressure data. The sodium-chloride potential of mean force in solution calculated from the empirically-adjusted model is consistent with the results from that calculated from ab initio CPMD simulations.

Overbinding of ions to lipid head groups is a potentially serious artifact in simulations of charged lipid bilayers. In this study, the Lennard-Jones radii in the CHARMM force field for interactions of Na+ and lipid oxygen atoms of carboxyl, phosphate, and ester groups were revised to match osmotic pressure data on sodium acetate and electrophoresis data on palmitoyloleoyl phosphatidylcholine (POPC) vesicles. The new parameters were then validated by successfully reproducing previously published experimental NMR deuterium order parameters for dimyristoyl phosphatidylglycerol (DMPG) and newly obtained values for palmitoyloleoyl phosphatidylserine (POPS). Although the increases in Lennard-Jones diameters are only 0.02-0.12 angstrom, they are sufficient of reduce Na+ binding, and thereby increase surface areas per lipid by 5-10% compared with the unmodified parameters.

An extremely scalable computational strategy is described for calculations of the potential of mean force (PMF) in multidimensions on massively distributed supercomputers. The approach involves coupling thousands of umbrella sampling (US) simulation windows distributed to cover the space of order parameters with a Hamiltonian molecular dynamics replica-exchange (H-REMD) algorithm to enhance the sampling of each simulation. In the present application, US/H-REMD is carried out in a two-dimensional (2D) space and exchanges are attempted alternatively along the two axes corresponding to the two order parameters. The US/H-REMD strategy is implemented on the basis of parallel/parallel multiple copy protocol at the MPI level, and therefore can fully exploit computing power of large-scale supercomputers. Here the novel technique is illustrated using the leadership supercomputer IBM Blue Gene/P with an application to a typical biomolecular calculation of general interest, namely the binding of calcium ions to the small protein Calbindin D9k. The free energy landscape associated with two order parameters, the distance between the ion and its binding pocket and the root-mean-square deviation (rmsd) of the binding pocket relative the crystal structure, was calculated using the US/H-REMD method. The results are then used to estimate the absolute binding free energy of calcium ion to Calbindin D9k. The tests demonstrate that the 2D US/H-REMD scheme greatly accelerates the configurational sampling of the binding pocket, thereby improving the convergence of the potential of mean force calculation.

The relative permeability of salt to water across an atomistic model of the FT-30 reverse osmosis (RO) membrane is studied using molecular dynamics (MD) simulations. The membrane model is built using a heuristic approach and gives a membrane density, water solubility and flux that are in good accord with the experimental values. The salt permeability is calculated from inhomogeneous solubility-diffusion theory using ion pathways from non-equilibrium targeted MD simulations, yielding an estimated salt rejection of 99.9% that is similar to the experimental value. The encouraging agreement with experimental data of FT-30 membrane suggests that MD simulations based on atomic models offer a useful way to support the experimental exploration of RD membrane development. (C) 2011 Elsevier B.V. All rights reserved.