Empirical force field for cisplatin based on quantum dynamics data: case study of new parameterization scheme for coordination compounds.

Parameterization of molecular complexes containing a metallic compound, such as cisplatin, is challenging due to the unconventional coordination nature of the bonds which involve platinum atoms. In this work, we develop a new methodology of parameterization for such compounds based on quantum dynamics (QD) calculations. We show that the coordination bonds and angles are more flexible than in normal covalent compounds. The influence of explicit solvent is also shown to be crucial to determine the flexibility of cisplatin in quantum dynamics simulations. Two empirical topologies of cisplatin were produced by fitting its atomic fluctuations against QD in vacuum and QD with explicit first solvation shell of water molecules respectively. A third topology built in a standard way from the static optimized structure was used for comparison. The later one leads to an excessively rigid molecule and exhibits much smaller fluctuations of the bonds and angles than QD reveals. It is shown that accounting for the high flexibility of cisplatin molecule is needed for adequate description of its first hydration shell. MD simulations with flexible QD-based topology also reveal a significant decrease of the barrier of passive diffusion of cisplatin accross the model lipid bilayer. These results confirm that flexibility of organometallic compounds is an important feature to be considered in classical molecular dynamics topologies. Proposed methodology based on QD simulations provides a systematic way of building such topologies.

Experimental and simulation studies of unusual current blockade induced by translocation of small oxidized PEG through a single nanopore.

Detection of a single macromolecule based on the use of artificial nanopores is an attractive and promising field of research. In this work, we report a device based on a 5 nm single nanopore with a high length/diameter ratio, tailored by the track etching and atomic layer deposition techniques. The translocation of neutral polyethylene glycol (PEG) and charged polyethylene glycol-carboxylate (PEG-carboxylate) molecules of low molar masses (200 and 600 g mol(-1)) through this nanodevice was studied. It was shown that charged PEG-carboxylate molecules, which permeate through the pore, promote an unusual blockade of ionic current whereas the neutral PEG molecules do not show such behaviour. The molecular dynamics simulation shows that both neutral and charged PEGs permeate through the nanopore close to its inner surface. The main difference between the two macromolecules is the existence of a structured shell of cations around the charged PEG, which is likely to cause the observed unusual current blockade.

Determination of mean and Gaussian curvatures of highly curved asymmetric lipid bilayers: the case study of the influence of cholesterol on the membrane shape.

Although molecular dynamics simulations of highly curved lipid bilayers have become increasingly popular in recent years, there is no simple and general method of computing the shape and curvature of the bilayer, which is bent arbitrarily in three dimensions. In this work we propose a method, which allows computing local normal, mean and Gaussian curvatures at any point of an arbitrarily curved lipid membrane using molecular dynamics trajectories. The method is based on the analysis of local membrane patches and is applicable to the membranes of any shape and topology - bilayers, vesicles, micelles, bicelles, etc. The method is applied to a highly curved asymmetric DOPC/DOPS lipid bilayer simulated by means of extended coarse-grained molecular dynamics simulations. It is shown that addition of cholesterol makes the membrane more topologically heterogeneous by increasing the content of highly curved regions with either saddle-like or sphere-like topology. The topology of the DOPS lipid domains is more sensitive to the addition of cholesterol than DOPC domains.

Determination of the shape and curvature of nonplanar lipid bilayers that are bent in a single plane in molecular dynamics simulations.

The curvature of biological membranes is known to be an important influence on important phenomena such as membrane fusion, endocytosis, and the functioning of integral membrane proteins. There is a growing demand for analytical tools that are able accommodate molecular dynamics trajectories of significantly curved lipid bilayers. In this work, an algorithm for determining the shape and curvature of a nonplanar lipid bilayer in molecular dynamics simulations is proposed. The algorithm calculates the coordinates of the midline and the curvature of the bilayer as well as the local normal to it at any point on the membrane, which is bent arbitrarily in a single plane and is topologically equivalent to an infinite bilayer. The algorithm is implemented as a C++ program and tested by exploring the molecular dynamic trajectories of a highly curved meander-like asymmetric lipid bilayer. The algorithm is general enough to allow it to be easily applied to other geometries of nonplanar membrane systems.

Cholesterol induces uneven curvature of asymmetric lipid bilayers.

A remarkable flexibility is observed in biological membranes, which allows them to form the structures of different curvatures. We addressed the question of intrinsic ability of phospholipid membranes to form highly curved structures and the role of cholesterol in this process. The distribution of cholesterol in the highly curved asymmetric DOPC/DOPS lipid bilayer was investigated by the coarse-grained molecular dynamics simulations in the membrane patches with large aspect ratio. It is shown that cholesterol induces uneven membrane curvature promoting the formation of extended flattened regions of the membrane interleaved by sharp bends. It is shown that the affinity of cholesterol to anionic DOPS or neutral DOPC lipids is curvature dependent. The cholesterol prefers DOPS to DOPC in either planar or highly curved parts of the membrane. In contrast, in the narrow interval of moderate membrane curvatures this preference is inverted. Our data suggest that there is a complex self-consistent interplay between the membrane curvature and cholesterol distribution in the asymmetric lipid bilayers. The suggested new function of cholesterol may have a biological relevance.

Interdomain compactization in human tyrosyl-tRNA synthetase studied by the hierarchical rotations technique.

Aminoacyl-tRNA synthetases are key enzymes of protein biosynthesis which usually possess multidomain structures. Mammalian tyrosyl-tRNA synthetase is composed of two structural modules: N-terminal catalytic core and an EMAPII-like C-terminal domain separated by long flexible linker. The structure of full-length human cytoplasmic tyrosyl-tRNA synthetase is still unknown. The structures of isolated N-terminal and C-terminal domains of the protein are resolved, but their compact packing in a functional enzyme is a subject of debates. In this work we studied putative compactization of the N- and C-terminal modules of human tyrosyl-tRNA synthetase by the coarse-grained hierarchical rotations technique (HIEROT). The large number of distinct types of binding interfaces between N- and C-terminal modules is revealed in the absence of enzyme substrates. The binding propensities of different residues are computed and several binding "hot spots" are observed on the surfaces of N and C modules. These results could be used to govern atomistic molecular dynamics simulations, which will sample preferable binding interfaces effectively.

Hierarchical clustering of the correlation patterns: new method of domain identification in proteins.

New method of identification of dynamical domains in proteins - Hierarchical Clustering of the Correlation Patterns (HCCP) is proposed. HCCP allows to identify the domains using single three-dimensional structure of the studied proteins and does not require any adjustable parameters that can influence the results. The method is based on hierarchical clustering performed on the matrices of correlation patterns, which are obtained by the transformation of ordinary pairwise correlation matrices. This approach allows to extract additional information from the correlation matrices, which increases reliability of domain identification. It is shown that HCCP is insensitive to small variations of the pairwise correlation matrices. Particularly it produces identical results if the data obtained for the same protein crystallized with different spatial positions of domains are used for analysis. HCCP can utilize correlation matrices obtained by any method such as normal mode or essential dynamics analysis, Gaussian network or anisotropic network models, etc. These features make HCCP an attractive method for domain identification in proteins.

Towards realistic description of collective motions in the lattice protein folding models.

Collective motions and the formation of clusters of residues play an important role in the folding of real proteins. However, existing Monte Carlo (MC) techniques of the protein folding simulations based on highly popular lattice models provide only a schematic representation of collective motions, which is rather far from physical reality. The Clustering Monte Carlo (CMC) algorithm was developed with particular aim to provide a realistic description of collective motions on the lattice. CMC allows modeling the cluster dynamics and the effects of the solvent viscosity, which is impossible in conventional algorithms. In this study two 2D lattice peptides, with the ground states of hierarchical and non-hierarchical design, were investigated comparatively using three methods: Metropolis MC with the local move set, Metropolis MC with unspecific rigid rotations and the CMC algorithm. We present evidence that the folding pathways and kinetics of hierarchically folding clustered sequence are not adequately described in conventional MC simulations, and the account for cluster dynamics provided by CMC allows to capture essential features of the folding process. Our data suggest that the methods, which enable specific cluster motions, such as CMC, should be used for a more realistic description of protein folding.

Hierarchy of motions and quasi-particles in a simplified model of potassium channel selectivity filter.

We develop a simplified model of themultiply occupied Kcsa-like selectivityfilter based on the best availablestructural data. The existence of hierarchyof motions in the selectivity filter isshown. Fast fluctuations of the ion-iondistances may be considered adiabaticallydecoupled from the slow diffusive motion ofthe ions' center of masses. The latter canbe considered as a quasi-particle, called aquasi-ion, moving in an effectivepotential. In the Kcsa-like selectivityfilter occupied by three ions the effectivepotential allows free barrier-lessdiffusional motion of the quasi-ions. Theconcept of the quasi-ions performing iontranslocation through the channel may bevital in explaining barrier-less `knock-on' conduction postulated for real channels.