Rapid protein-ligand docking using soft modes from molecular dynamics simulations to account for protein deformability: binding of FK506 to FKBP

Most current docking methods to identify possible ligands and putative binding sites on a receptor molecule assume a rigid receptor structure to allow virtual screening of large ligand databases. However, binding of a ligand can lead to changes in the receptor protein conformation that are sterically necessary to accommodate a bound ligand. An approach is presented that allows relaxation of the protein conformation in precalculated soft flexible degrees of freedom during ligand-receptor docking. For the immunosuppressant FK506-binding protein FKBP, the soft flexible modes are extracted as principal components of motion from a molecular dynamics simulation. A simple penalty function for deformations in the soft flexible mode is used to limit receptor protein deformations during docking that avoids a costly recalculation of the receptor energy by summing over all receptor atom pairs at each step. Rigid docking of the FK506 ligand binding to an unbound FKBP conformation failed to identify a geometry close to experiment as favorable binding site. In contrast, inclusion of the flexible soft modes during systematic docking runs selected a binding geometry close to experiment as lowest energy conformation. This has been achieved at a modest increase of computational cost compared to rigid docking. The approach could provide a computationally efficient way to approximately account for receptor flexibility during docking of large numbers of putative ligands and putative docking geometries.     

New insights into the mechanism of protein-protein association

Association of a protein complex follows a two-step mechanism, with the first step being the formation of an encounter complex that evolves into the final complex. Here, we analyze recent experimental data of the association of TEM1-beta-lactamase with BLIP using theoretical calculations and simulation. We show that the calculated Debye-Hückel energy of interaction for a pair of proteins during association resembles an energy funnel, with the final complex at the minima. All attraction is lost at inter-protein distances of 20 A, or rotation angles of >60 degrees from the orientation of the final complex. For faster-associating protein complexes, the energy funnel deepens and its volume increases. Mutations with the largest impact on association (hotspots for association) have the largest effect on the size and depth of the energy funnel. Analyzing existing evidence, we suggest that the transition state along the association pathway is the formation of the final complex from the encounter complex. Consequently, pairs of proteins forming an encounter complex will tend to dissociate more readily than to evolve into the final complex. Increasing directional diffusion by increasing favorable electrostatic attraction results in a faster forming and slower dissociating encounter complex. The possible applicability of electrostatic calculations for protein-protein docking is discussed.     

An algorithm to filter out packing arrangements based on steric clashes

This document outlines the use of an algorithm to filter out impossible crystal-packing arrangements based on steric considerations. Within an exhaustive grid search frame, the space sample is reduced by analysis of spherical areas where atom pairs from different rigid units might clash.This technique finds areas in the state space where the global energy minimum might lie. The minimum can then be found by the usual methods of molecular modeling restricted to these particular areas.Only a tiny fraction of atom pair distances need to be tested; usually a single quantity on average per one state of model space! For example, a crystal of three rigid molecules, each containing 12 atoms, has 3×12×12=432 atom pairs just in one unit cell but our method needs to test on average 1 to 4 atom pairs per state.Using modern computers, about 10^12–15 models can be tested within several hours or days. For example, a crystal model with six rotational degrees of freedom (two rigid molecules in the unit cell) each with step 3° can be tested in a few hours on a 1-GHz x86 processor-based machine.The method presented here has been implemented in the SUPRAMOL program.     

SIDEpro: a novel machine learning approach for the fast and accurate prediction of side-chain conformations

Accurate protein side-chain conformation prediction is crucial for protein modeling and existing methods for the task are widely used; however, faster and more accurate methods are still required. Here we present a new machine learning approach to the problem where an energy function for each rotamer in a structure is computed additively over pairs of contacting atoms. A family of 156 neural networks indexed by amino acid and contacting atom types is used to compute these rotamer energies as a function of atomic contact distances. Although direct energy targets are not available for training, the neural networks can still be optimized by converting the energies to probabilities and optimizing these probabilities using Markov Chain Monte Carlo methods. The resulting predictor SIDEpro makes predictions by initially setting the rotamer probabilities for each residue from a backbone-dependent rotamer library, then iteratively updating these probabilities using the trained neural networks. After convergences of the probabilities, the side-chains are set to the highest probability rotamer. Finally, a post processing clash reduction step is applied to the models. SIDEpro represents a significant improvement in speed and a modest, but statistically significant, improvement in accuracy when compared with the state-of-the-art for rapid side-chain prediction method SCWRL4 on the following datasets: (1) 379 protein test set of SCWRL4; (2) 94 proteins from CASP9; (3) a set of seven large protein-only complexes; and (4) a ribosome with and without the RNA. Using the SCWRL4 test set, SIDEpro's accuracy (χ(1) 86.14%, χ(1+2) 74.15%) is slightly better than SCWRL4-FRM (χ(1) 85.43%, χ(1+2) 73.47%) and it is 7.0 times faster. On the same test set SIDEpro is clearly more accurate than SCWRL4-rigid rotamer model (RRM) (χ(1) 84.15%, χ(1+2) 71.24%) and 2.4 times faster. Evaluation on the additional test sets yield similar accuracy results with SIDEpro being slightly more accurate than SCWRL4-flexible rotamer model (FRM) and clearly more accurate than SCWRL4-RRM; however, the gap in CPU time is much more significant when the methods are applied to large protein complexes. SIDEpro is part of the SCRATCH suite of predictors and available from: http://scratch.proteomics.ics.uci.edu/.     

Hybrid simulation approach incorporating microscopic interaction along with rigid body degrees of freedom for stacking between base pairs

Stacking interaction between the aromatic heterocyclic bases plays an important role in the double helical structures of nucleic acids. Considering the base as rigid body, there are total of 18 degrees of freedom of a dinucleotide step. Some of these parameters show sequence preferences, indicating that the detailed atomic interactions are important in the stacking. Large variants of non‐canonical base pairs have been seen in the crystallographic structures of RNA. However, their stacking preferences are not thoroughly deciphered yet from experimental results. The current theoretical approaches use either the rigid body degrees of freedom where the atomic information are lost or computationally expensive all atom simulations. We have used a hybrid simulation approach incorporating Monte‐Carlo Metropolis sampling in the hyperspace of 18 stacking parameters where the interaction energies using AMBER‐parm99bsc0 and CHARMM‐36 force‐fields were calculated from atomic positions. We have also performed stacking energy calculations for structures from Monte‐Carlo ensemble by Dispersion corrected density functional theory. The available experimental data with Watson–Crick base pairs are compared to establish the validity of the method. Stacking interaction involving A:U and G:C base pairs with non‐canonical G:U base pairs also were calculated and showed that these structures were also sequence dependent. This approach could be useful to generate multiscale modeling of nucleic acids in terms of coarse‐grained parameters where the atomic interactions are preserved. This method would also be useful to predict structure and dynamics of different base pair steps containing non Watson–Crick base pairs, as found often in the non‐coding RNA structures. © 2015 Wiley Periodicals, Inc. Biopolymers 105: 212–226, 2016.     

Distance-constrained molecular docking by simulated annealing

An optimized method based on the principle of simulated annealing is presented for determining the relative position and orientation of interacting molecules. The spatial relationships of these molecules are described by intermolecular distance constraints between specific pairs of atoms, such as found in hydrogen bonds or from experimentally determined data. The method makes use of a random walk through six rotational and translational degrees of freedom where the constituent molecules are treated as rigid bodies. Van der Waals repulsions are used only to define a lower bound on distances between constrained atom pairs within the docking procedure. A cost function comprised of purely geometric constraints is optimized via simulated annealing, in order to search for the best orientation and position of the two molecules. Our docking procedure is applied to eight serine proteinase complexes from the Brookhaven Protein Data Bank. For each simulation 100 computations were performed. A typical docking computation requires only a few seconds of CPU time on a VAXserver 3500. The influence of the number of constraints on the final docked positions was studied. The sensitivity of the docking procedure to a ligand structure which is not well defined is also addressed. Possible applications of this method include using approximate distances incorporating complete energy functions.     

Configurational temperature density of states simulations of proteins

A novel method has been implemented to compute the density of states of proteins. A united atom representation and the CHARMM (Brooks et al., 1983) force-field parameters have been adopted for all the simulations. In this approach, an intrinsic temperature is computed based on configurational information about the protein. A random walk is performed in potential energy space and the configurational temperature is collected as a function of potential energy of the system. The density of states is then calculated by integrating the reciprocal of temperature. Unlike previously available methods, this approach does not involve calculations based on histograms of stochastic visits to distinct energy states. It is found that the proposed method is more efficient than earlier, related schemes for simulation of protein folding. Furthermore, it directly provides thermodynamic information, including free energies. The usefulness of the method is discussed by presenting results of simulations of the 16-residue beta-hairpin taken from the C-terminal fragment (41-56) of protein G.     

距离约束的HMC采样算法在蛋白质结构预测中的运用

蛋白质结构预测中,采样是指在构象空间中生成具有最小自由能的状态。传统的采样方法是对自由度直接赋值。这种方法在处理较少的残基时能取得好的效果。但是对于包含100个残基以上的蛋白质结构,由于构象空间的急剧增长,难以得到理想的结构。本文引入深度学习中的HMC(Hybrid Monte Carlo)采样方法,以概率分布为依据对蛋白质的自由度进行采样,能够对包含100、200甚至更多个残基的蛋白质结构进行采样。并且,在采样的过程中加入残基间的距离约束,使得一个结构中,相对于Rosetta的ab initio最多有75%(平均40%)的残基对得到优化,满足距离约束。     

Protein flexibility predictions using graph theory

Techniques from graph theory are applied to analyze the bond networks in proteins and identify the flexible and rigid regions. The bond network consists of distance constraints defined by the covalent and hydrogen bonds and salt bridges in the protein, identified by geometric and energetic criteria. We use an algorithm that counts the degrees of freedom within this constraint network and that identifies all the rigid and flexible substructures in the protein, including overconstrained regions (with more crosslinking bonds than are needed to rigidify the region) and underconstrained or flexible regions, in which dihedral bond rotations can occur. The number of extra constraints or remaining degrees of bond-rotational freedom within a substructure quantifies its relative rigidity/flexibility and provides a flexibility index for each bond in the structure. This novel computational procedure, first used in the analysis of glassy materials, is approximately a million times faster than molecular dynamics simulations and captures the essential conformational flexibility of the protein main and side-chains from analysis of a single, static three-dimensional structure. This approach is demonstrated by comparison with experimental measures of flexibility for three proteins in which hinge and loop motion are essential for biological function: HIV protease, adenylate kinase, and dihydrofolate reductase.     

Modeling protein-induced configurational changes in DNA minicircles

A method is offered for obtaining minimum energy configurations of DNA minicircles constrained by one or more DNA-binding proteins. The minicircles are modeled as elastic rods, while the presence of bound protein is implied by rigidly fixing portions of these chains. The configurations of the geometrically constrained circular rods are sampled stochastically and optimized according to a simple elastic energy model of nicked DNA. The shapes of the minimum energy structures identified after a simulated annealing process are analyzed in terms of relative protein orientation and writhing number. The procedure is applied to minicircles 500 base pairs in length, bound to two evenly spaced DNA-wrapping proteins. The presence of histone octamers is suggested by rigidly fixing the two protein-bound portions of each minicircle as small superhelices similar in dimension to nucleosomal DNA. The folded minimum energy forms of sample chains with different degrees of protein wrapping are noteworthy in themselves in that they offer a new resolution to the well-known minichromosome linking number paradox and point to future minicircle simulations of possible import. © 1997 John Wiley & Sons, Inc.     

Prediction of protein-protein interactions based on surface patch comparison

A method to predict if two proteins interact, based on their three-dimensional structures, is presented. It consists of five steps: (i) the surface of each protein, represented by the solvent accessible atoms, is divided into small patches; (ii) the shape of each patch is described by the atom distributions along its principal axes; (iii) the shape complementarity between two patches is estimated by comparing, through contingency table analysis, their atom distributions along their principal axes; (iv) given protein A, with nA surface patches, and protein B, with nB surface patches, nA x nB shape complementarity values are obtained; and (v) the distribution of the latter allows one to discriminate pairs of interacting and of noninteracting proteins. Only a few seconds are necessary to predict if two proteins interact, with accuracy close to 80%, sensitivity over 70% and specificity close to 50%.     

A method to predict muscle control in the kinematically and mechanically indeterminate human masticatory system

A method is proposed to generate muscle activation patterns for goal-directed movements of the human masticatory system. This system is special because apart from a larger amount of muscles than degrees of freedom its joints do not restrict its movements a priori. Therefore, each muscle is able to influence all six degrees of freedom which makes the system kinematically and mechanically indeterminate. Furthermore, its working space is principally determined by the dynamical properties of its muscles and not by passive constraints.The presented method determines the contribution of each degree of freedom to a movement of a reference point on the mandible. It avails of straightforward mathematical techniques like Linear Programming. It does not require a separate trajectory planning step. It was applied in a six degrees of freedom dynamical mathematical model of the human masticatory system. This model which was based upon rigid-body dynamics incorporating skull morphology and muscle architecture including dynamical properties. Movements were exclusively defined by a goal position of the mandibular reference point.The method proved to be robust in generating muscle activation patterns for both feasible and infeasible movement tasks. Generally, they were accomplished faster than habitually observed. If the task was infeasible the movement stopped at the outer boundary of the working space at the side of the unreachable goal. The method, therefore, enables to explore the working space of the mandible and the factors that are relevant for its boundaries.     

Linking 3D and 2D binding kinetics of membrane proteins by multiscale simulations

Membrane proteins are among the most functionally important proteins in cells. Unlike soluble proteins, they only possess two translational degrees of freedom on cell surfaces, and experience significant constraints on their rotations. As a result, it is currently challenging to characterize the in situ binding of membrane proteins. Using the membrane receptors CD2 and CD58 as a testing system, we developed a multiscale simulation framework to study the differences of protein binding kinetics between 3D and 2D environments. The association and dissociation processes were implemented by a coarse‐grained Monte‐Carlo algorithm, while the dynamic properties of proteins diffusing on lipid bilayer were captured from all‐atom molecular dynamic simulations. Our simulations show that molecular diffusion, linker flexibility and membrane fluctuations are important factors in adjusting binding kinetics. Moreover, by calibrating simulation parameters to the measurements of 3D binding, we derived the 2D binding constant which is quantitatively consistent with the experimental data, indicating that the method is able to capture the difference between 3D and 2D binding environments. Finally, we found that the 2D dissociation between CD2 and CD58 is about 100‐fold slower than the 3D dissociation. In summary, our simulation framework offered a generic approach to study binding mechanisms of membrane proteins.     

Sequence-dependent DNA structure: a database of octamer structural parameters

We have constructed the potential energy surfaces for all unique tetramers, hexamers and octamers in double helical DNA, as a function of the two principal degrees of freedom, slide and shift at the central step. From these potential energy maps, we have calculated a database of structural and flexibility properties for each of these sequences. These properties include: the values of each of the six step parameters (twist roll, tilt, rise, slide and shift), for each step of the sequence; flexibility measures for both decrease and increase in each property value from the minimum energy conformation for the central step; and the deviation from the path of a hypothetical straight octamer. In an analysis of structural change as a function of sequence length, we observe that almost all DNA tends to B-DNA and becomes less flexible. A more detailed analysis of octamer properties has allowed us to determine the structural preferences of particular sequence elements. GGC and GCC sequences tend to confer bistability, low stability and a predisposition to A-form DNA, whereas AA steps strongly prefer B-DNA and inhibit A-structures. There is no correlation between flexibility and intrinsic curvature, but bent DNA is less stable than straight. The most difficult deformation is undertwisting. The TA step stands out as the most flexible sequence element with respect to decreasing twist and increasing roll. However, as with the structural properties, this behavior is highly context-dependent and some TA steps are very straight.     

A mezoscopic model of nucleic acids. Part 2. An effective potential energy function for DNA

In this study we present an effective Potential of Mean Force (PMF) designed for Lagrangian and Quaternion Molecular Dynamics (LQMD) of DNA. The DNA model is built from pseudoatoms as well as rigid and pseudo-elastic bodies described by a limited number of selected Cartesian and internal degrees of freedom. Phosphate groups, deoxyribose rings and nucleic acid bases are represented by pseudoparticles, some of them with internal degrees of freedom. PMF is defined as the sum of effective bonded and long-range potentials. The potentials were fitted to numerical free energy surfaces. Over 50 free energy surfaces, each depending on a conformational variable (pseudobond length, angle or dihedral angle) and the pseudorotation phase of the nearest neighbour deoxribose ring, were computed. The numerical free energy surfaces were obtained from probability distributions derived from a 1.5 ns conventional, microscopic MD simulation of the d(GpC)9 double helical DNA molecule. An umbrella sampling method was used to simulate transitions between the A and B DNA forms, and PMF reproduces these transitions.     

Prediction of atomic structure from sequence for double helical DNA oligomers

DNA can adopt different conformations depending on the base sequence, solvent, electrolyte composition and concentration, pH, temperature, and interaction with proteins. Here we present a model for calculating the three-dimensional atomic structure of double-stranded DNA oligomers. A theoretical energy function is used for calculating the interactions within the base steps and an empirical backbone function is used to restrict the conformational space accessible to the bases and to account for the conformational coupling of neighboring steps in a sequence. Conformational searching on large structures or a large number of structures is possible, because each base step can be described by just two primary degrees of freedom (slide and shift). A genetic algorithm is used to search for low-energy structures in slide-shift space, and this allows very rapid optimization of DNA oligomers. The other base step parameters have been previously optimized for all possible slide-shift sequence combinations, and a heuristic algorithm is used to add the atomic details of the backbone conformation in the final step of the calculation. The structures obtained by this method are very similar to the corresponding X-ray crystal structures observed experimentally. The average RMSD is 2.24 Angstroms for a set of 20 oligomer structures. For 15 of these sequences, the X-ray crystal structure is the global energy minimum. The other 5 are bistable sequences that have B-form global energy minima but crystallize as A-DNA.     

Identifying 3D maximal common substructures using transputer networks

This paper describes the use of a multiprocessor system for identifying the maximal common substructure (MCS) between pairs of three-dimensional (3D) chemical structures. The system is constructed from Transputers, 32-bit RISC microprocessors produced by Inmos Ltd., linked together in a tree network. The MCS algorithm used, developed by Crandell and Smith, identifies the MCS by a breadth-first search in which individual atoms common to the structures are extended one atom at a time until no further extension of the common substructure can be obtained. Experiments using a Pascal-based simulation package demonstrate the feasibility of using a multiprocessor system to increase the speed of MCS identification. Experiments with networks of Transputers demonstrate that substantial increases in speed can be achieved in practice if, and only if, the MCS is large.     

Generalized-ensemble algorithms for molecular simulations of biopolymers.

In complex systems with many degrees of freedom such as peptides and proteins, there exists a huge number of local-minimum-energy states. Conventional simulations in the canonical ensemble are of little use, because they tend to get trapped in states of these energy local minima. A simulation in generalized ensemble performs a random walk in potential energy space and can overcome this difficulty. From only one simulation run, one can obtain canonical-ensemble averages of physical quantities as functions of temperature by the single-histogram and/or multiple-histogram reweighting techniques. In this article we review uses of the generalized-ensemble algorithms in biomolecular systems. Three well-known methods, namely, multicanonical algorithm, simulated tempering, and replica-exchange method, are described first. Both Monte Carlo and molecular dynamics versions of the algorithms are given. We then present three new generalized-ensemble algorithms that combine the merits of the above methods. The effectiveness of the methods for molecular simulations in the protein folding problem is tested with short peptide systems.