LEAP: Highly accurate prediction of protein loop conformations by integrating coarse‐grained sampling and optimized energy scores with all‐atom refinement of backbone and side chains

Prediction of protein loop conformations without any prior knowledge (ab initio prediction) is an unsolved problem. Its solution will significantly impact protein homology and template‐based modeling as well as ab initio protein‐structure prediction. Here, we developed a coarse‐grained, optimized scoring function for initial sampling and ranking of loop decoys. The resulting decoys are then further optimized in backbone and side‐chain conformations and ranked by all‐atom energy scoring functions. The final integrated technique called loop prediction by energy‐assisted protocol achieved a median value of 2.1 Å root mean square deviation (RMSD) for 325 12‐residue test loops and 2.0 Å RMSD for 45 12‐residue loops from critical assessment of structure‐prediction techniques (CASP) 10 target proteins with native core structures (backbone and side chains). If all side‐chain conformations in protein cores were predicted in the absence of the target loop, loop‐prediction accuracy only reduces slightly (0.2 Å difference in RMSD for 12‐residue loops in the CASP target proteins). The accuracy obtained is about 1 Å RMSD or more improvement over other methods we tested. The executable file for a Linux system is freely available for academic users at http://sparks‐lab.org . © 2013 Wiley Periodicals, Inc.     

Prediction of Protein Loop Conformations using the AGBNP Implicit Solvent Model and Torsion Angle Sampling

The OPLS-AA all-atom force field and the Analytical Generalized Born plus Non-Polar (AGBNP) implicit solvent model, in conjunction with torsion angle conformational search protocols based on the Protein Local Optimization Program (PLOP), are shown to be effective in predicting the native conformations of 57 9-residue and 35 13-residue loops of a diverse series of proteins with low sequence identity. The novel nonpolar solvation free energy estimator implemented in AGBNP augmented by correction terms aimed at reducing the occurrence of ion pairing are important to achieve the best prediction accuracy. Extended versions of the previously developed PLOP-based conformational search schemes based on calculations in the crystal environment are reported that are suitable for application to loop homology modeling without the crystal environment. Our results suggest that in general the loop backbone conformation is not strongly influenced by crystal packing. The application of the temperature Replica Exchange Molecular Dynamics (T-REMD) sampling method for a few examples where PLOP sampling is insufficient are also reported. The results reported indicate that the OPLS-AA/AGBNP effective potential is suitable for high-resolution modeling of proteins in the final stages of homology modeling and/or protein crystallographic refinement.     

Comparison between self-guided Langevin dynamics and molecular dynamics simulations for structure refinement of protein loop conformations

This article presents a comparative analysis of two replica‐exchange simulation methods for the structure refinement of protein loop conformations, starting from low‐resolution predictions. The methods are self‐guided Langevin dynamics (SGLD) and molecular dynamics (MD) with a Nosé–Hoover thermostat. We investigated a small dataset of 8‐ and 12‐residue loops, with the shorter loops placed initially from a coarse‐grained lattice model and the longer loops from an enumeration assembly method (the Loopy program). The CHARMM22 + CMAP force field with a generalized Born implicit solvent model (molecular‐surface parameterized GBSW2) was used to explore conformational space. We also assessed two empirical scoring methods to detect nativelike conformations from decoys: the all‐atom distance‐scaled ideal‐gas reference state (DFIRE‐AA) statistical potential and the Rosetta energy function. Among the eight‐residue loop targets, SGLD out performed MD in all cases, with a median of 0.48 Å reduction in global root‐mean‐square deviation (RMSD) of the loop backbone coordinates from the native structure. Among the more challenging 12‐residue loop targets, SGLD improved the prediction accuracy over MD by a median of 1.31 Å, representing a substantial improvement. The overall median RMSD for SGLD simulations of 12‐residue loops was 0.91 Å, yielding refinement of a median 2.70 Å from initial loop placement. Results from DFIRE‐AA and the Rosetta model applied to rescoring conformations failed to improve the overall detection calculated from the CHARMM force field. We illustrate the advantage of SGLD over the MD simulation model by presenting potential‐energy landscapes for several loop predictions. Our results demonstrate that SGLD significantly outperforms traditional MD in the generation and populating of nativelike loop conformations and that the CHARMM force field performs comparably to other empirical force fields in identifying these conformations from the resulting ensembles. Published 2011 Wiley Periodicals, Inc. J Comput Chem, 2011     

Protein structure prediction aided by geometrical and probabilistic constraints

Database-assisted ab initio protein structure prediction methods have exhibited considerable promise in the recent past, with several implementations being successful in community-wide experiments (CASP). We have employed combinatorial optimization techniques toward solving the protein structure prediction problem. A Monte Carlo minimization algorithm has been employed on a constrained search space to identify minimum energy configurations. The search space is constrained by using radius of gyration cutoffs, the loop backbone dihedral probability distributions, and various secondary structure packing conformations. Simulations have been carried out on several sequences and 1000 conformations have been initially generated. Of these, 50 best candidates have then been selected as probable conformations. The search for the optimum has been simplified by incorporating various geometrical constraints on secondary structural elements using distance restraint potential functions. The advantages of the reported methodology are its simplicity, and modifiability to include other geometric and probabilistic restraints.     

Determination of the major solution conformation of tyrocidine A,using molecular mechanics energy minimization and NMR-derived distance and torsion angle constraints

Molecular mechanics energy calculations coupled with nuclear magnetic resonance-determined distance and torsion angle constraints have been used to determine the three-dimensional structure of tyrocidine A, a cyclic decapeptide which exists largely as a single conformation in solution. Two open-chain polyalanine models were used to represent separate halves of the peptide backbone and a combinatorial method of searching conformation space used to generate candidate structures consistent with experimental distance constraints. These structures were energy-minimized using the AMBER molecular mechanics forcefield and the resulting conformations classified by factor analysis of their Cartesian coordinates. Representative low-energy conformers of the two halves of the backbone were fused together and two candidate conformations of the completed backbone refined by further minimization using both distance and torsional constraints. Side chains were then added as their experimentally preferred rotamers and the whole molecule minimized without constraints to give the final model structure. This shows type II' and III ß turns at residues 4–5 and 9–10, respectively, coupled by twisted antiparallel strands which show hydrogen bonds between all four pairs of opposing peptide groups. The backbone conformation of residues 2–6 closely resembles that found in the crystal structure of gramicidin S.     

An evaluation of molecular models of the cytochrome P450 Streptomyces griseolus enzymes P450SU1 and P450SU2

Summary P450SU1 and P450SU2 are herbicide-inducible bacterial cytochrome P450 enzymes from Streptomyces griseolus. They have two of the highest sequence identities to camphor hydroxylase (P450cam from Pseudomonas putida), the cytochrome P450 with the first known crystal structure. We have built several models of these two proteins to investigate the variability in the structures that can occur from using different modeling protocols. We looked at variability due to alignment methods, backbone loop conformations and refinement methods. We have constructed two models for each protein using two alignment algorithms, and then an additional model using an identical alignment but different loop conformations for both buried and surface loops. The alignments used to build the models were created using the Needleman-Wunsch method, adapted for multiple sequences, and a manual method that utilized both a dotmatrix search matrix and the Needleman-Wunsch method. After constructing the initial models, several energy minimization methods were used to explore the variability in the final models caused by the choice of minimization techniques. Features of cytochrome P450cam and the cytochrome P450 superfamily, such as the ferredoxin binding site, the heme binding site and the substrate binding site were used to evaluate the validity of the models. Although the final structures were very similar between the models with different alignments, active-site residues were found to be dependent on the conformations of buried loops and early stages of energy minimization. We show which regions of the active site are the most dependent on the particular methods used, and which parts of the structures seem to be independent of the methods.     

Modeling loop backbone flexibility in receptor‐ligand docking simulations

The relevance of receptor conformational change during ligand binding is well documented for many pharmaceutically relevant receptors, but is still not fully accounted for in in silico docking methods. While there has been significant progress in treatment of receptor side chain flexibility sampling of backbone flexibility remains challenging because the conformational space expands dramatically and the scoring function must balance protein–protein and protein–ligand contributions. Here, we investigate an efficient multistage backbone reconstruction algorithm for large loop regions in the receptor and demonstrate that treatment of backbone receptor flexibility significantly improves binding mode prediction starting from apo structures and in cross docking simulations. For three different kinase receptors in which large flexible loops reconstruct upon ligand binding, we demonstrate that treatment of backbone flexibility results in accurate models of the complexes in simulations starting from the apo structure. At the example of the DFG‐motif in the p38 kinase, we also show how loop reconstruction can be used to model allosteric binding. Our approach thus paves the way to treat the complex process of receptor reconstruction upon ligand binding in docking simulations and may help to design new ligands with high specificity by exploitation of allosteric mechanisms. © 2012 Wiley Periodicals, Inc.     

A method for conformational sampling of loops in proteins based on adiabatic decoupling and temperature or force scaling

A method for conformational Boltzmann sampling of loops in proteins in aqueous solution is presented that is based on adiabatic decoupling molecular dynamics (MD) simulation with temperature or force scaling. To illustrate the enhanced sampling, the loop from residues 33 to 43 in the bovine protein ribonuclease A is adiabatically decoupled from the rest of the protein and the solvent with a mass scaling factor s(m) =1000 and the sampling is enhanced with a scaling of the temperature using s(T) =2 or of the force using s(V) =0.667. Over 5 ns of simulation the secondary structure of the protein remains unaltered while a combined dihedral-angle conformational cluster analysis shows an increase of conformations outside the first most populated cluster of loop conformations for adiabatic decoupling MD with temperature scaling using s(T) =2 or force scaling using s(V) =0.667 compared to the standard MD simulation. The atom-positional root-mean-square fluctuations of the C(α) atoms of the loop show an increase in the movement of the loop as well, indicating that adiabatic decoupling MD with upscaling of the temperature or downscaling of the force is a promising method for conformational Boltzmann sampling.     

Geometric algorithms for the conformational analysis of long protein loops

The efficient filtering of unfeasible conformations would considerably benefit the exploration of the conformational space when searching for minimum energy structures or during molecular simulation. The most important conditions for filtering are the maintenance of molecular chain integrity and the avoidance of steric clashes. These conditions can be seen as geometric constraints on a molecular model. In this article, we discuss how techniques issued from recent research in robotics can be applied to this filtering. Two complementary techniques are presented: one for conformational sampling and another for computing conformational changes satisfying such geometric constraints. The main interest of the proposed techniques is their application to the structural analysis of long protein loops. First experimental results demonstrate the efficacy of the approach for studying the mobility of loop 7 in amylosucrase from Neisseria polysaccharea. The supposed motions of this 17-residue loop would play an important role in the activity of this enzyme.     

Supervised consensus scoring for docking and virtual screening

Docking programs are widely used to discover novel ligands efficiently and can predict protein-ligand complex structures with reasonable accuracy and speed. However, there is an emerging demand for better performance from the scoring methods. Consensus scoring (CS) methods improve the performance by compensating for the deficiencies of each scoring function. However, conventional CS and existing scoring functions have the same problems, such as a lack of protein flexibility, inadequate treatment of salvation, and the simplistic nature of the energy function used. Although there are many problems in current scoring functions, we focus our attention on the incorporation of unbound ligand conformations. To address this problem, we propose supervised consensus scoring (SCS), which takes into account protein-ligand binding process using unbound ligand conformations with supervised learning. An evaluation of docking accuracy for 100 diverse protein-ligand complexes shows that SCS outperforms both CS and 11 scoring functions (PLP, F-Score, LigScore, DrugScore, LUDI, X-Score, AutoDock, PMF, G-Score, ChemScore, and D-score). The success rates of SCS range from 89% to 91% in the range of rmsd < 2 A, while those of CS range from 80% to 85%, and those of the scoring functions range from 26% to 76%. Moreover, we also introduce a method for judging whether a compound is active or inactive with the appropriate criterion for virtual screening. SCS performs quite well in docking accuracy and is presumably useful for screening large-scale compound databases before predicting binding affinity.     

Finite-state and reduced-parameter representations of protein backbone conformations

This article studies the representation of protein backbone conformations using a finite number of values for the backbone dihedral angles. We develop a combinatorial search algorithm that guarantees finding the global minima of functions over the configuration space of discrete protein conformations, and use this procedure to fit finite-state models to the backbones of globular proteins. It is demonstrated that a finite-state representation with a reasonably small number of states yields either a small root-mean-square error or a small dihedral angle deviation from the native structure, but not both at the same time. The problem can be resolved by introducing limited local optimization in each step of the combinatorial search. In addition, it is shown that acceptable approximation is achieved using a single dihedral angle as an independent variable in local optimization. Results for 11 proteins demonstrate the advantages and shortcomings of both the finite-state and reduced-parameter approximations of protein backbone conformations. © 1994 by John Wiley & Sons, Inc.     

A sampling-based method for ranking protein structural models by integrating multiple scores and features

One of the major challenges in protein tertiary structure prediction is structure quality assessment. In many cases, protein structure prediction tools generate good structural models, but fail to select the best models from a huge number of candidates as the final output. In this study, we developed a sampling-based machine-learning method to rank protein structural models by integrating multiple scores and features. First, features such as predicted secondary structure, solvent accessibility and residue-residue contact information are integrated by two Radial Basis Function (RBF) models trained from different datasets. Then, the two RBF scores and five selected scoring functions developed by others, i.e., Opus-CA, Opus-PSP, DFIRE, RAPDF, and Cheng Score are synthesized by a sampling method. At last, another integrated RBF model ranks the structural models according to the features of sampling distribution. We tested the proposed method by using two different datasets, including the CASP server prediction models of all CASP8 targets and a set of models generated by our in-house software MUFOLD. The test result shows that our method outperforms any individual scoring function on both best model selection, and overall correlation between the predicted ranking and the actual ranking of structural quality.     

Closing the side-chain gap in protein loop modeling

The success of structure-based drug design relies on accurate protein modeling where one of the key issues is the modeling and refinement of loops. This study takes a critical look at modeled loops, determining the effect of re-sampling side-chains after the loop conformation has been generated. The results are evaluated in terms of backbone and side-chain conformations with respect to the native loop. While models can contain loops with high quality backbone conformations, the side-chain orientations could be poor, and therefore unsuitable for ligand docking and structure-based design. In this study, we report on the ability to model loop side-chains accurately using a variety of commercially available algorithms that include rotamer libraries, systematic torsion scans and knowledge-based methods. Electronic supplementary material  The online version of this article (doi:) contains supplementary material, which is available to authorized users.     

Conformational memories with variable bond angles

Conformational Memories (CM) is a simulated annealing/Monte Carlo method that explores peptide and protein dihedral conformational space completely and efficiently, independent of the original conformation. Here we extend the CM method to include the variation of a randomly chosen bond angle, in addition to the standard variation of two or three randomly chosen dihedral angles, in each Monte Carlo trial of the CM exploratory and biased phases. We test the hypothesis that the inclusion of variable bond angles in CM leads to an improved sampling of conformational space. We compare the results with variable bond angles to CM with no bond angle variation for the following systems: (1) the pentapeptide Met-enkephalin, which is a standard test case for conformational search methods; (2) the proline ring pucker in a 17mer model peptide, (Ala)(8)Pro(Ala)(8); and (3) the conformations of the Ser 7.39 chi(1) in transmembrane helix 7 (TMH7) of the cannabinoid CB1 receptor, a 25-residue system. In each case, analysis of the CM results shows that the inclusion of variable bond angles results in sampling of regions of conformational space that are inaccessible to CM calculations with only variable dihedral angles, and/or a shift in conformational populations from those calculated when variable bond angles are not included. The incorporation of variable bond angles leads to an improved sampling of conformational space without loss of efficiency. Our examples show that this improved sampling leads to better exploration of biologically relevant conformations that have been experimentally validated.     

An off-lattice frustrated model protein with a six-stranded β-barrel structure

We apply a global optimization method, which is conformational space annealing (CSA) to a challenging problem of the 69-residue protein with the sequence B(9)N(3)(LB)(4)N(3)B(9)N(3)(LB)(4)N(3)B(9)N(3)(LB)(5)L, where B, L, and N designate hydrophobic, hydrophilic, and neutral residues, respectively. The 69-residue BLN protein folds into a six-stranded β-barrel structure. The CSA method always maintains the diversity of sampling and is able to cross the high energy barriers between local minima. The CSA successfully located the global minimum of the 69-residue BLN protein for all 100 independent runs. For a single run, it takes about 3 h and 30 min on average to obtain the global minimum on a Linux PC. Also, we investigate the properties of the 69-residue BLN protein, and the general behavior of the M-residue BLN protein for CSA runs.     

Statistical Analysis of Ion Mobility Spectrometry. II. Adaptively Biased Methods and Shape Correlations

Following a recent effort [J. Am. Soc. Mass Spectrom. 23, 386-396 (2012)], we continue to explore computational methodologies for generating molecular conformations to support collisional cross sections suggested by ion mobility measurements. Here, adaptively biased molecular dynamics (ABMD) simulations are used to sample the configuration space and to achieve flat-histogram sampling along the reaction coordinates of the first two moments of the gyration tensor. The method is tested and compared with replica-exchange simulations on triply-protonated bradykinin and on a larger 25-residue peptide. It is found to have a much higher efficiency for producing large sets of conformations in a broad range of diffusion cross-sections, whereas it does not compete with conventional replica-exchange molecular dynamics in locating the lowest-energy structure. Nevertheless, the broad sampling obtained from the ABMD method allows to quantitatively correlate the diffusion cross-section Ω with other geometric order parameters that have simpler interpretation. The strong correlations found between the diffusion cross-section and the radius of gyration, the surface area and the volume of the convex hull suggest an optimal template for accurately mimicking the variations of Ω in a broad range of conformations, using only geometrical information and doing so at a very moderate computational cost. The existence of such a correlation is confirmed on the much larger protein α-lactalbumin.     

A kinematic view of loop closure

We consider the problem of loop closure, i.e., of finding the ensemble of possible backbone structures of a chain segment of a protein molecule that is geometrically consistent with preceding and following parts of the chain whose structures are given. We reduce this problem of determining the loop conformations of six torsions to finding the real roots of a 16th degree polynomial in one variable, based on the robotics literature on the kinematics of the equivalent rotator linkage in the most general case of oblique rotators. We provide a simple intuitive view and derivation of the polynomial for the case in which each of the three pair of torsional axes has a common point. Our method generalizes previous work on analytical loop closure in that the torsion angles need not be consecutive, and any rigid intervening segments are allowed between the free torsions. Our approach also allows for a small degree of flexibility in the bond angles and the peptide torsion angles; this substantially enlarges the space of solvable configurations as is demonstrated by an application of the method to the modeling of cyclic pentapeptides. We give further applications to two important problems. First, we show that this analytical loop closure algorithm can be efficiently combined with an existing loop-construction algorithm to sample loops longer than three residues. Second, we show that Monte Carlo minimization is made severalfold more efficient by employing the local moves generated by the loop closure algorithm, when applied to the global minimization of an eight-residue loop. Our loop closure algorithm is freely available at http://dillgroup. ucsf.edu/loop_closure/.     

A comparison of different initialization protocols to obtain statistically independent molecular dynamics simulations

We study how the results of molecular dynamics (MD) simulations are affected by various choices during the setup, e.g., the starting velocities, the solvation, the location of protons, the conformation of His, Asn, and Gln residues, the protonation and titration of His residues, and the treatment of alternative conformations. We estimate the binding affinity of ligands to four proteins calculated with the MM/GBSA method (molecular mechanics combined with a generalized Born and surface area solvation energy). For avidin and T4 lysozyme, all variations gave similar results within 2 kJ/mol. For factor Xa, differences in the solvation or in the selection of alternative conformations gave results that are significantly different from those of the other approaches by 4-6 kJ/mol, whereas for galectin-3, changes in the conformations, rotations, and protonation gave results that differed by 10 kJ/mol, but only if residues close to the binding site were modified. This shows that the results of MM/GBSA calculations are reasonably reproducible even if the MD simulations are set up with different software. Moreover, we show that the sampling of phase space can be enhanced by solvating the systems with different equilibrated water boxes, in addition to the common use of different starting velocities. If different conformations are available in the crystal structure, they should also be employed to enhance the sampling. Protonation, ionization, and conformations of Asn, Gln, and His may also be used to enhance sampling, but great effort should be spent to obtain as reliable predictions as possible close to the active site.     

Statistical potential for modeling and ranking of protein-ligand interactions

Applications in structural biology and medicinal chemistry require protein-ligand scoring functions for two distinct tasks: (i) ranking different poses of a small molecule in a protein binding site and (ii) ranking different small molecules by their complementarity to a protein site. Using probability theory, we developed two atomic distance-dependent statistical scoring functions: PoseScore was optimized for recognizing native binding geometries of ligands from other poses and RankScore was optimized for distinguishing ligands from nonbinding molecules. Both scores are based on a set of 8,885 crystallographic structures of protein-ligand complexes but differ in the values of three key parameters. Factors influencing the accuracy of scoring were investigated, including the maximal atomic distance and non-native ligand geometries used for scoring, as well as the use of protein models instead of crystallographic structures for training and testing the scoring function. For the test set of 19 targets, RankScore improved the ligand enrichment (logAUC) and early enrichment (EF(1)) scores computed by DOCK 3.6 for 13 and 14 targets, respectively. In addition, RankScore performed better at rescoring than each of seven other scoring functions tested. Accepting both the crystal structure and decoy geometries with all-atom root-mean-square errors of up to 2 ? from the crystal structure as correct binding poses, PoseScore gave the best score to a correct binding pose among 100 decoys for 88% of all cases in a benchmark set containing 100 protein-ligand complexes. PoseScore accuracy is comparable to that of DrugScore(CSD) and ITScore/SE and superior to 12 other tested scoring functions. Therefore, RankScore can facilitate ligand discovery, by ranking complexes of the target with different small molecules; PoseScore can be used for protein-ligand complex structure prediction, by ranking different conformations of a given protein-ligand pair. The statistical potentials are available through the Integrative Modeling Platform (IMP) software package (http://salilab.org/imp) and the LigScore Web server (http://salilab.org/ligscore/).