Fast procedure for reconstruction of full-atom protein models from reduced representations

VarMod: modelling the functional effects of non-synonymous variants

Unravelling the genotype–phenotype relationship in humans remains a challenging task in genomics studies. Recent advances in sequencing technologies mean there are now thousands of sequenced human genomes, revealing millions of single nucleotide variants (SNVs). For non-synonymous SNVs present in proteins the difficulties of the problem lie in first identifying those nsSNVs that result in a functional change in the protein among the many non-functional variants and in turn linking this functional change to phenotype. Here we present VarMod (Variant Modeller) a method that utilises both protein sequence and structural features to predict nsSNVs that alter protein function. VarMod develops recent observations that functional nsSNVs are enriched at protein–protein interfaces and protein–ligand binding sites and uses these characteristics to make predictions. In benchmarking on a set of nearly 3000 nsSNVs VarMod performance is comparable to an existing state of the art method. The VarMod web server provides extensive resources to investigate the sequence and structural features associated with the predictions including visualisation of protein models and complexes via an interactive JSmol molecular viewer. VarMod is available for use at http://www.wasslab.org/varmod.

How good are simplified models for protein structure prediction?

Protein structure prediction (PSP) has been one of the most challenging problems in computational biology for several decades. The challenge is largely due to the complexity of the all-atomic details and the unknown nature of the energy function. Researchers have therefore used simplified energy models that consider interaction potentials only between the amino acid monomers in contact on discrete lattices. The restricted nature of the lattices and the energy models poses a twofold concern regarding the assessment of the models. Can a native or a very close structure be obtained when structures are mapped to lattices? Can the contact based energy models on discrete lattices guide the search towards the native structures? In this paper, we use the protein chain lattice fitting (PCLF) problem to address the first concern; we developed a constraint-based local search algorithm for the PCLF problem for cubic and face-centered cubic lattices and found very close lattice fits for the native structures. For the second concern, we use a number of techniques to sample the conformation space and find correlations between energy functions and root mean square deviation (RMSD) distance of the lattice-based structures with the native structures. Our analysis reveals weakness of several contact based energy models used that are popular in PSP.

Perspective: Coarse-grained models for biomolecular systems

By focusing on essential features, while averaging over less important details, coarse-grained (CG) models provide significant computational and conceptual advantages with respect to more detailed models. Consequently, despite dramatic advances in computational methodologies and resources, CG models enjoy surging popularity and are becoming increasingly equal partners to atomically detailed models. This perspective surveys the rapidly developing landscape of CG models for biomolecular systems. In particular, this review seeks to provide a balanced, coherent, and unified presentation of several distinct approaches for developing CG models, including top-down, network-based, native-centric, knowledge-based, and bottom-up modeling strategies. The review summarizes their basic philosophies, theoretical foundations, typical applications, and recent developments. Additionally, the review identifies fundamental inter-relationships among the diverse approaches and discusses outstanding challenges in the field. When carefully applied and assessed, current CG models provide highly efficient means for investigating the biological consequences of basic physicochemical principles. Moreover, rigorous bottom-up approaches hold great promise for further improving the accuracy and scope of CG models for biomolecular systems.

Interactions of the auxilin-1 PTEN-like domain with model membranes result in nanoclustering of phosphatidyl inositol phosphates

Auxilin-1 is a neuron-specific membrane-binding protein involved in a late stage of clathrin-mediated endocytosis. It recruits Hsc70, thus initiating uncoating of the clathrin-coated vesicles. Interactions of auxilin-1 with the vesicle membrane are crucial for this function and are mediated via an N-terminal PTEN-like domain. We have used multiscale molecular dynamics simulations to probe the interactions of the auxilin-1 PTEN-like domain with lipid bilayers containing differing phospholipid composition, including bilayers containing phosphatidyl inositol phosphates. Our results suggest a novel, to our knowledge, model for the auxilin/membrane encounter and subsequent interactions. Negatively charged lipids (especially PIP2) enhance binding of auxilin to lipid bilayers and facilitate its correct orientation relative to the membrane. Mutations in three basic residues (R301E/R307E/K311E) of the C2 subdomain of the PTEN-like domain perturbed its interaction with the bilayer, changing its orientation. The interaction of membrane-bound auxilin-1 PTEN-like domain with negatively charged lipid headgroups results in nanoclustering of PIP2 molecules in the adjacent bilayer leaflet.

Going Backward: A Flexible Geometric Approach to Reverse Transformation from Coarse Grained to Atomistic Models

The conversion of coarse-grained to atomistic models is an important step in obtaining insight about atomistic scale processes from coarse-grained simulations. For this process, called backmapping or reverse transformation, several tools are available, but these commonly require libraries of molecule fragments or they are linked to a specific software package. In addition, the methods are usually restricted to specific molecules and to a specific force field. Here, we present an alternative method, consisting of geometric projection and subsequent force-field based relaxation. This method is designed to be simple and flexible, and offers a generic solution for resolution transformation. For simple systems, the conversion only requires a list of particle correspondences on the two levels of resolution. For special cases, such as nondefault protonation states of amino acids and virtual sites, a target particle list can be specified. The mapping uses simple building blocks, which list the particles on the different levels of resolution. For conversion to higher resolution, the initial model is relaxed with several short cycles of energy minimization and position-restrained MD. The reconstruction of an atomistic backbone from a coarse-grained model is done using a new dedicated algorithm. The method is generic and can be used to map between any two particle based representations, provided that a mapping can be written. The focus of this work is on the coarse-grained MARTINI force field, for which mapping definitions are written to allow conversion to and from the higher-resolution force fields GROMOS, CHARMM, and AMBER, and to and from a simplified three-bead lipid model. Together, these offer the possibility to simulate mesoscopic membrane structures, to be transformed to MARTINI and subsequently to an atomistic model for investigation of detailed interactions. The method was tested on a set of systems ranging from a simple, single-component bilayer to a large protein-membrane-solvent complex. The results demonstrate the efficiency and the efficacy of the new approach.

Conformational changes of globular proteins upon adsorption on a hydrophobic surface

Coarse-grained Monte Carlo simulations are used to study thermal denaturation of small globular proteins adsorbed on a hydrophobic surface. Though helices are more stable than sheets, they are highly deformed in the adsorbed protein.

Aggregation of lipid-anchored full-length H-Ras in lipid bilayers: simulations with the MARTINI force field

Lipid-anchored Ras oncoproteins assemble into transient, nano-sized substructures on the plasma membrane. These substructures, called nanoclusters, were proposed to be crucial for high-fidelity signal transmission in cells. However, the molecular basis of Ras nanoclustering is poorly understood. In this work, we used coarse-grained (CG) molecular dynamics simulations to investigate the molecular mechanism by which full-length H-ras proteins form nanoclusters in a model membrane. We chose two different conformations of H-ras that were proposed to represent the active and inactive state of the protein, and a domain-forming model bilayer made up of di16:0-PC (DPPC), di18:2-PC (DLiPC) and cholesterol. We found that, irrespective of the initial conformation, Ras molecules assembled into a single large aggregate. However, the two binding modes, which are characterized by the different orientation of the G-domain with respect to the membrane, differ in dynamics and organization during and after aggregation. Some of these differences involve regions of Ras that are important for effector/modulator binding, which may partly explain observed differences in the ability of active and inactive H-ras nanoclusters to recruit effectors. The simulations also revealed some limitations in the CG force field to study protein assembly in solution, which we discuss in the context of proposed potential avenues of improvement.

Designing a polycationic probe for simultaneous enrichment and detection of microRNAs in a nanopore

The nanopore sensor can detect cancer-derived nucleic acid biomarkers such as microRNAs (miRNAs), providing a noninvasive tool potentially useful in medical diagnostics. However, the nanopore-based detection of these biomarkers remains confounded by the presence of numerous other nucleic acid species found in biofluid extracts. Their nonspecific interactions with the nanopore inevitably contaminate the target signals, reducing the detection accuracy. Here we report a novel method that utilizes a polycationic peptide-PNA probe as the carrier for selective miRNA detection in the nucleic acid mixture. The cationic probe hybridized with microRNA forms a dipole complex, which can be captured by the pore using a voltage polarity that is opposite the polarity used to capture negatively charged nucleic acids. As a result, nontarget species are driven away from the pore opening, and the target miRNA can be detected accurately without interference. In addition, we demonstrate that the PNA probe enables accurate discrimination of miRNAs with single-nucleotide difference. This highly sensitive and selective nanodielectrophoresis approach can be applied to the detection of clinically relevant nucleic acid fragments in complex samples.

High-quality protein backbone reconstruction from alpha carbons using Gaussian mixture models

Coarse-grained protein structure models offer increased efficiency in structural modeling, but these must be coupled with fast and accurate methods to revert to a full-atom structure. Here, we present a novel algorithm to reconstruct mainchain models from C traces. This has been parameterized by fitting Gaussian mixture models (GMMs) to short backbone fragments centered on idealized peptide bonds. The method we have developed is statistically significantly more accurate than several competing methods, both in terms of RMSD values and dihedral angle differences. The method produced Ramachandran dihedral angle distributions that are closer to that observed in real proteins and better Phaser molecular replacement log-likelihood gains. Amino acid residue sidechain reconstruction accuracy using SCWRL4 was found to be statistically significantly correlated to backbone reconstruction accuracy. Finally, the PD2 method was found to produce significantly lower energy full-atom models using Rosetta which has implications for multiscale protein modeling using coarse-grained models. A webserver and C++ source code is freely available for noncommercial use from: http://www.sbg.bio.ic.ac.uk/phyre2/PD2_ca2main/.

Interplay of physics and evolution in the likely origin of protein biochemical function

The intrinsic ability of protein structures to exhibit the geometric and sequence properties required for ligand binding without evolutionary selection is shown by the coincidence of the properties of pockets in native, single domain proteins with those in computationally generated, compact homopolypeptide, artificial (ART) structures. The library of native pockets is covered by a remarkably small number of representative pockets (∼400), with virtually every native pocket having a statistically significant match in the ART library, suggesting that the library is complete. When sequences are selected for ART structures based on fold stability, pocket sequence conservation is coincident to native. The fact that structurally and sequentially similar pockets occur across fold classes combined with the small number of representative pockets in native proteins implies that promiscuous interactions are inherent to proteins. Based on comparison of PDB (real, single domain protein structures found in the Protein Data Bank) and ART structures and pockets, the widespread assumption that the co-occurrence of global structure, pocket similarity, and amino acid conservation demands an evolutionary relationship between proteins is shown to significantly underestimate the random background probability. Indeed, many features of biochemical function arise from the physical properties of proteins that evolution likely fine-tunes to achieve specificity. Finally, our study suggests that a repertoire of thermodynamically (marginally) stable proteins could engage in many of the biochemical reactions needed for living systems without selection for function, a conclusion with significant implications for the origin of life.

Multiscale simulations reveal conserved patterns of lipid interactions with aquaporins

Interactions of membrane proteins with lipid molecules are central to their stability and function. We have used multiscale molecular dynamics simulations to determine the extent to which interactions with lipids are conserved across the aquaporin (Aqp) family of membrane proteins. Simulation-based assessment of the lipid interactions made by Aqps when embedded within a simple phospholipid bilayer agrees well with the protein-lipid contacts determined by electron diffraction from 2D crystals. Extending this simulation-based analysis to all Aqps of known structure reveals a degree of conservation of such interactions across the Aqp structural proteome. Despite similarities in the binding orientations and interactions of the lipids, there do not appear to be distinct, high-specificity lipid binding sites on the surface of Aqps. Rather Aqps exhibit a more broadly conserved protein/lipid interface, suggestive of interchange between annular and bulk lipids, instead of a fixed annular "shell" of lipids.

Three dimensional structure of the anthrax toxin translocon-lethal factor complex by cryo-electron microscopy

We have visualized by cryo-electron microscopy (cryo-EM) the complex of the anthrax protective antigen (PA) translocon and the N-terminal domain of anthrax lethal factor (LF(N) inserted into a nanodisc model lipid bilayer. We have determined the structure of this complex at a nominal resolution of 16 Å by single-particle analysis and three-dimensional reconstruction. Consistent with our previous analysis of negatively stained unliganded PA, the translocon comprises a globular structure (cap) separated from the nanodisc bilayer by a narrow stalk that terminates in a transmembrane channel (incompletely distinguished in this reconstruction). The globular cap is larger than the unliganded PA pore, probably due to distortions introduced in the previous negatively stained structures. The cap exhibits larger, more distinct radial protrusions, previously identified with PA domain three, fitted by elements of the NMFF PA prepore crystal structure. The presence of LF(N), though not distinguished due to the seven-fold averaging used in the reconstruction, contributes to the distinct protrusions on the cap rim volume distal to the membrane. Furthermore, the lumen of the cap region is less resolved than the unliganded negatively stained PA, due to the low contrast obtained in our images of this specimen. Presence of the LF(N) extended helix and N terminal unstructured regions may also contribute to this additional internal density within the interior of the cap. Initial NMFF fitting of the cryoEM-defined PA pore cap region positions the Phe clamp region of the PA pore translocon directly above an internal vestibule, consistent with its role in toxin translocation.

A dynamic model of the proteins that form the initial iron-sulfur cluster biogenesis machinery in yeast mitochondria

The assembly of iron-sulfur clusters (ISCs) in eukaryotes involves the protein Frataxin. Deficits in this protein have been associated with iron inside the mitochondria and impair ISC biogenesis as it is postulated to act as the iron donor for ISCs assembly in this organelle. A pronounced lack of Frataxin causes Friedreich's Ataxia, which is a human neurodegenerative and hereditary disease mainly affecting the equilibrium, coordination, muscles and heart. Moreover, it is the most common autosomal recessive ataxia. High similarities between the human and yeast molecular mechanisms that involve Frataxin have been suggested making yeast a good model to study that process. In yeast, the protein complex that forms the central assembly platform for the initial step of ISC biogenesis is composed by yeast frataxin homolog, Nfs1-Isd11 and Isu. In general, it is commonly accepted that protein function involves interaction with other protein partners, but in this case not enough is known about the structure of the protein complex and, therefore, how it exactly functions. The objective of this work is to model the protein complex in order to gain insight into structural details that end up with its biological function. To achieve this goal several bioinformatics tools, modeling techniques and protein docking programs have been used. As a result, the structure of the protein complex and the dynamic behavior of its components, along with that of the iron and sulfur atoms required for the ISC assembly, have been modeled. This hypothesis will help to better understand the function and molecular properties of Frataxin as well as those of its ISC assembly protein partners.

An atomistic model for assembly of transmembrane domain of T cell receptor complex

The T cell receptor (TCR) together with accessory cluster of differentiation 3 (CD3) molecules (TCR-CD3 complex) is a key component in the primary function of T cells. The nature of association of the transmembrane domains is of central importance to the assembly of the complex and is largely unknown. Using multiscale molecular modeling and simulations, we have investigated the structure and assembly of the TCRα-CD3ε-CD3δ transmembrane domains both in membrane and in micelle environments. We demonstrate that in a membrane environment the transmembrane basic residue of the TCR closely interacts with both of the transmembrane acidic residues of the CD3 dimer. In contrast, in a micelle the basic residue interacts with only one of the acidic residues. Simulations of a recent micellar nuclear magnetic resonance structure of the natural killer (NK) cell-activating NKG2C-DAP12-DAP12 trimer in a membrane further indicate that the environment significantly affects the way these trimers associate. Since the currently accepted model for transmembrane association is entirely based on a micellar structure, we propose a revised model for the association of transmembrane domains of the activating immune receptors in a membrane environment.

The road not taken: retreat and diverge in local search for simplified protein structure prediction

BACKGROUND

Given a protein's amino acid sequence, the protein structure prediction problem is to find a three dimensional structure that has the native energy level. For many decades, it has been one of the most challenging problems in computational biology. A simplified version of the problem is to find an on-lattice self-avoiding walk that minimizes the interaction energy among the amino acids. Local search methods have been preferably used in solving the protein structure prediction problem for their efficiency in finding very good solutions quickly. However, they suffer mainly from two problems: re-visitation and stagnancy.

RESULTS

In this paper, we present an efficient local search algorithm that deals with these two problems. During search, we select the best candidate at each iteration, but store the unexplored second best candidates in a set of elite conformations, and explore them whenever the search faces stagnation. Moreover, we propose a new non-isomorphic encoding for the protein conformations to store the conformations and to check similarity when applied with a memory based search. This new encoding helps eliminate conformations that are equivalent under rotation and translation, and thus results in better prevention of re-visitation.

CONCLUSION

On standard benchmark proteins, our algorithm significantly outperforms the state-of-the art approaches for Hydrophobic-Polar energy models and Face Centered Cubic Lattice.

Head-head interactions of resting myosin crossbridges in intact frog skeletal muscles, revealed by synchrotron x-ray fiber diffraction

The intensities of the myosin-based layer lines in the x-ray diffraction patterns from live resting frog skeletal muscles with full thick-thin filament overlap from which partial lattice sampling effects had been removed were analyzed to elucidate the configurations of myosin crossbridges around the thick filament backbone to nanometer resolution. The repeat of myosin binding protein C (C-protein) molecules on the thick filaments was determined to be 45.33 nm, slightly longer than that of myosin crossbridges. With the inclusion of structural information for C-proteins and a pre-powerstroke head shape, modeling in terms of a mixed population of regular and perturbed regions of myosin crown repeats along the filament revealed that the myosin filament had azimuthal perturbations of crossbridges in addition to axial perturbations in the perturbed region, producing pseudo-six-fold rotational symmetry in the structure projected down the filament axis. Myosin crossbridges had a different organization about the filament axis in each of the regular and perturbed regions. In the regular region that lacks C-proteins, there were inter-molecular interactions between the myosin heads in axially adjacent crown levels. In the perturbed region that contains C-proteins, in addition to inter-molecular interactions between the myosin heads in the closest adjacent crown levels, there were also intra-molecular interactions between the paired heads on the same crown level. Common features of the interactions in both regions were interactions between a portion of the 50-kDa-domain and part of the converter domain of the myosin heads, similar to those found in the phosphorylation-regulated invertebrate myosin. These interactions are primarily electrostatic and the converter domain is responsible for the head-head interactions. Thus multiple head-head interactions of myosin crossbridges also characterize the switched-off state and have an important role in the regulation or other functions of myosin in thin filament-regulated muscles as well as in the thick filament-regulated muscles.

eThread: a highly optimized machine learning-based approach to meta-threading and the modeling of protein tertiary structures

Template-based modeling that employs various meta-threading techniques is currently the most accurate, and consequently the most commonly used, approach for protein structure prediction. Despite the evident progress in this field, accurate structure models cannot be constructed for a significant fraction of gene products, thus the development of new algorithms is required. Here, we describe the development, optimization and large-scale benchmarking of eThread, a highly accurate meta-threading procedure for the identification of structural templates and the construction of corresponding target-to-template alignments. eThread integrates ten state-of-the-art threading/fold recognition algorithms in a local environment and extensively uses various machine learning techniques to carry out fully automated template-based protein structure modeling. Tertiary structure prediction employs two protocols based on widely used modeling algorithms: Modeller and TASSER-Lite. As a part of eThread, we also developed eContact, which is a Bayesian classifier for the prediction of inter-residue contacts and eRank, which effectively ranks generated multiple protein models and provides reliable confidence estimates as structure quality assessment. Excluding closely related templates from the modeling process, eThread generates models, which are correct at the fold level, for >80% of the targets; 40–50% of the constructed models are of a very high quality, which would be considered accurate at the family level. Furthermore, in large-scale benchmarking, we compare the performance of eThread to several alternative methods commonly used in protein structure prediction. Finally, we estimate the upper bound for this type of approach and discuss the directions towards further improvements.

Structure refinement of protein model decoys requires accurate side-chain placement

In this study, the application of temperature-based replica-exchange (T-ReX) simulations for structure refinement of decoys taken from the I-TASSER dataset was examined. A set of eight nonredundant proteins was investigated using self-guided Langevin dynamics (SGLD) with a generalized Born implicit solvent model to sample conformational space. For two of the protein test cases, a comparison of the SGLD/T-ReX method with that of a hybrid explicit/implicit solvent molecular dynamics T-ReX simulation model is provided. Additionally, the effect of side-chain placement among the starting decoy structures, using alternative rotamer conformations taken from the SCWRL4 modeling program, was investigated. The simulation results showed that, despite having near-native backbone conformations among the starting decoys, the determinant of their refinement is side-chain packing to a level that satisfies a minimum threshold of native contacts to allow efficient excursions toward the downhill refinement regime on the energy landscape. By repacking using SCWRL4 and by applying the RWplus statistical potential for structure identification, the SGLD/T-ReX simulations achieved refinement to an average of 38% increase in the number of native contacts relative to the original I-TASSER decoy sets and a 25% reduction in values of C(α) root-mean-square deviation. The hybrid model succeeded in obtaining a sharper funnel to low-energy states for a modeled target than the implicit solvent SGLD model; yet, structure identification remained roughly the same. Without meeting a threshold of near-native packing of side chains, the T-ReX simulations degrade the accuracy of the decoys, and subsequently, refinement becomes tantamount to the protein folding problem.

Assessing the relative stability of dimer interfaces in g protein-coupled receptors

Considerable evidence has accumulated in recent years suggesting that G protein-coupled receptors (GPCRs) associate in the plasma membrane to form homo- and/or heteromers. Nevertheless, the stoichiometry, fraction and lifetime of such receptor complexes in living cells remain topics of intense debate. Motivated by experimental data suggesting differing stabilities for homomers of the cognate human β1- and β2-adrenergic receptors, we have carried out approximately 160 microseconds of biased molecular dynamics simulations to calculate the dimerization free energy of crystal structure-based models of these receptors, interacting at two interfaces that have often been implicated in GPCR association under physiological conditions. Specifically, results are presented for simulations of coarse-grained (MARTINI-based) and atomistic representations of each receptor, in homodimeric configurations with either transmembrane helices TM1/H8 or TM4/3 at the interface, in an explicit lipid bilayer. Our results support a definite contribution to the relative stability of GPCR dimers from both interface sequence and configuration. We conclude that β1- and β2-adrenergic receptor homodimers with TM1/H8 at the interface are more stable than those involving TM4/3, and that this might be reconciled with experimental studies by considering a model of oligomerization in which more stable TM1 homodimers diffuse through the membrane, transiently interacting with other protomers at interfaces involving other TM helices.

Simple few-state models reveal hidden complexity in protein folding

Markov state models constructed from molecular dynamics simulations have recently shown success at modeling protein folding kinetics. Here we introduce two methods, flux PCCA+ (FPCCA+) and sliding constraint rate estimation (SCRE), that allow accurate rate models from protein folding simulations. We apply these techniques to fourteen massive simulation datasets generated by Anton and Folding@home. Our protocol quantitatively identifies the suitability of describing each system using two-state kinetics and predicts experimentally detectable deviations from two-state behavior. An analysis of the villin headpiece and FiP35 WW domain detects multiple native substates that are consistent with experimental data. Applying the same protocol to GTT, NTL9, and protein G suggests that some beta containing proteins can form long-lived native-like states with small register shifts. Even the simplest protein systems show folding and functional dynamics involving three or more states.

Genomics-aided structure prediction

We introduce a theoretical framework that exploits the ever-increasing genomic sequence information for protein structure prediction. Structure-based models are modified to incorporate constraints by a large number of non-local contacts estimated from direct coupling analysis (DCA) of co-evolving genomic sequences. A simple hybrid method, called DCA-fold, integrating DCA contacts with an accurate knowledge of local information (e.g., the local secondary structure) is sufficient to fold proteins in the range of 1-3 Å resolution.

Cavities determine the pressure unfolding of proteins

It has been known for nearly 100 years that pressure unfolds proteins, yet the physical basis of this effect is not understood. Unfolding by pressure implies that the molar volume of the unfolded state of a protein is smaller than that of the folded state. This decrease in volume has been proposed to arise from differences between the density of bulk water and water associated with the protein, from pressure-dependent changes in the structure of bulk water, from the loss of internal cavities in the folded states of proteins, or from some combination of these three factors. Here, using 10 cavity-containing variants of staphylococcal nuclease, we demonstrate that pressure unfolds proteins primarily as a result of cavities that are present in the folded state and absent in the unfolded one. High-pressure NMR spectroscopy and simulations constrained by the NMR data were used to describe structural and energetic details of the folding landscape of staphylococcal nuclease that are usually inaccessible with existing experimental approaches using harsher denaturants. Besides solving a 100-year-old conundrum concerning the detailed structural origins of pressure unfolding of proteins, these studies illustrate the promise of pressure perturbation as a unique tool for examining the roles of packing, conformational fluctuations, and water penetration as determinants of solution properties of proteins, and for detecting folding intermediates and other structural details of protein-folding landscapes that are invisible to standard experimental approaches.

Simultaneous single-structure and bundle representation of protein NMR structures in torsion angle space

A method is introduced to represent an ensemble of conformers of a protein by a single structure in torsion angle space that lies closest to the averaged Cartesian coordinates while maintaining perfect covalent geometry and on average equal steric quality and an equally good fit to the experimental (e.g. NMR) data as the individual conformers of the ensemble. The single representative 'regmean structure' is obtained by simulated annealing in torsion angle space with the program CYANA using as input data the experimental restraints, restraints for the atom positions relative to the average Cartesian coordinates, and restraints for the torsion angles relative to the corresponding principal cluster average values of the ensemble. The method was applied to 11 proteins for which NMR structure ensembles are available, and compared to alternative, commonly used simple approaches for selecting a single representative structure, e.g. the structure from the ensemble that best fulfills the experimental and steric restraints, or the structure from the ensemble that has the lowest RMSD value to the average Cartesian coordinates. In all cases our method found a structure in torsion angle space that is significantly closer to the mean coordinates than the alternatives while maintaining the same quality as individual conformers. The method is thus suitable to generate representative single structure representations of protein structure ensembles in torsion angle space. Since in the case of NMR structure calculations with CYANA the single structure is calculated in the same way as the individual conformers except that weak positional and torsion angle restraints are added, we propose to represent new NMR structures by a 'regmean bundle' consisting of the single representative structure as the first conformer and all but one original individual conformers (the original conformer with the highest target function value is discarded in order to keep the number of conformers in the bundle constant). In this way, analyses that require a single structure can be carried out in the most meaningful way using the first model, while at the same time the additional information contained in the ensemble remains available.

Further evidence for the likely completeness of the library of solved single domain protein structures

Recent studies questioned whether the Protein Data Bank (PDB) contains all compact, single domain protein structures. Here, we show that all quasi-spherical, QS, random protein structures devoid of secondary structure are in the PDB and are excellent templates for all native PDB proteins up to 250 residues. Because QS templates have a similar global contour as native, TASSER can refine 98% (90%) of those whose TM-score is 0.4 (0.35) to structures greater than or equal to the 0.5 TM-score threshold (0.74 (0.64) mean TM-score) for CATH/SCOP assignment. On the basis of this and the fact that, at a TM-score of 0.4, 83% (90%) of all (internal) core secondary structure elements are recovered, a 0.40 TM-score is an appropriate fold similarity assignment threshold. Despite the claims of Taylor, Trovato, and Zhou that many of their structures lack a PDB counterpart, using fr-TM-align, at a 0.45 (0.5) TM-score threshold, essentially all (most) are found in the PDB. Thus, the conclusion that the PDB is likely complete is further supported.

A scoring function based on solvation thermodynamics for protein structure prediction

We predict protein structure using our recently developed free energy function for describing protein stability, which is focused on solvation thermodynamics. The function is combined with the current most reliable sampling methods, i.e., fragment assembly (FA) and comparative modeling (CM). The prediction is tested using 11 small proteins for which high-resolution crystal structures are available. For 8 of these proteins, sequence similarities are found in the database, and the prediction is performed with CM. Fairly accurate models with average Cα root mean square deviation (RMSD) ∼ 2.0 Å are successfully obtained for all cases. For the rest of the target proteins, we perform the prediction following FA protocols. For 2 cases, we obtain predicted models with an RMSD ∼ 3.0 Å as the best-scored structures. For the other case, the RMSD remains larger than 7 Å. For all the 11 target proteins, our scoring function identifies the experimentally determined native structure as the best structure. Starting from the predicted structure, replica exchange molecular dynamics is performed to further refine the structures. However, we are unable to improve its RMSD toward the experimental structure. The exhaustive sampling by coarse-grained normal mode analysis around the native structures reveals that our function has a linear correlation with RMSDs < 3.0 Å. These results suggest that the function is quite reliable for the protein structure prediction while the sampling method remains one of the major limiting factors in it. The aspects through which the methodology could further be improved are discussed.

Template-based protein structure modeling using TASSER(VMT.)

Template-based protein structure modeling is commonly used for protein structure prediction. Based on the observation that multiple template-based methods often perform better than single template-based methods, we further explore the use of a variable number of multiple templates for a given target in the latest variant of TASSER, TASSER(VMT) . We first develop an algorithm that improves the target-template alignment for a given template. The improved alignment, called the SP(3) alternative alignment, is generated by a parametric alignment method coupled with short TASSER refinement on models selected using knowledge-based scores. The refined top model is then structurally aligned to the template to produce the SP(3) alternative alignment. Templates identified using SP(3) threading are combined with the SP(3) alternative and HHEARCH alignments to provide target alignments to each template. These template models are then grouped into sets containing a variable number of template/alignment combinations. For each set, we run short TASSER simulations to build full-length models. Then, the models from all sets of templates are pooled, and the top 20-50 models selected using FTCOM ranking method. These models are then subjected to a single longer TASSER refinement run for final prediction. We benchmarked our method by comparison with our previously developed approach, pro-sp(3) -TASSER, on a set with 874 easy and 318 hard targets. The average GDT-TS score improvements for the first model are 3.5 and 4.3% for easy and hard targets, respectively. When tested on the 112 CASP9 targets, our method improves the average GDT-TS scores as compared to pro-sp3-TASSER by 8.2 and 9.3% for the 80 easy and 32 hard targets, respectively. It also shows slightly better results than the top ranked CASP9 Zhang-Server, QUARK and HHpredA methods. The program is available for download at http://cssb.biology.gatech.edu/.

Improving the physical realism and structural accuracy of protein models by a two-step atomic-level energy minimization

Most protein structural prediction algorithms assemble structures as reduced models that represent amino acids by a reduced number of atoms to speed up the conformational search. Building accurate full-atom models from these reduced models is a necessary step toward a detailed function analysis. However, it is difficult to ensure that the atomic models retain the desired global topology while maintaining a sound local atomic geometry because the reduced models often have unphysical local distortions. To address this issue, we developed a new program, called ModRefiner, to construct and refine protein structures from Cα traces based on a two-step, atomic-level energy minimization. The main-chain structures are first constructed from initial Cα traces and the side-chain rotamers are then refined together with the backbone atoms with the use of a composite physics- and knowledge-based force field. We tested the method by performing an atomic structure refinement of 261 proteins with the initial models constructed from both ab initio and template-based structure assemblies. Compared with other state-of-art programs, ModRefiner shows improvements in both global and local structures, which have more accurate side-chain positions, better hydrogen-bonding networks, and fewer atomic overlaps. ModRefiner is freely available at http://zhanglab.ccmb.med.umich.edu/ModRefiner.

Molecular dynamics trajectory compression with a coarse-grained model

Molecular dynamics trajectories are very data-intensive thereby limiting sharing and archival of such data. One possible solution is compression of trajectory data. Here, trajectory compression based on conversion to the coarse-grained model PRIMO is proposed. The compressed data is about one third of the original data and fast decompression is possible with an analytical reconstruction procedure from PRIMO to all-atom representations. This protocol largely preserves structural features and to a more limited extent also energetic features of the original trajectory.

A multilayer evaluation approach for protein structure prediction and model quality assessment

Protein tertiary structures are essential for studying functions of proteins at molecular level. An indispensable approach for protein structure solution is computational prediction. Most protein structure prediction methods generate candidate models first and select the best candidates by model quality assessment (QA). In many cases, good models can be produced, but the QA tools fail to select the best ones from the candidate model pool. Because of incomplete understanding of protein folding, each QA method only reflects partial facets of a structure model and thus has limited discerning power with no one consistently outperforming others. In this article, we developed a set of new QA methods, including two QA methods for evaluating target/template alignments, a molecular dynamics (MD)-based QA method, and three consensus QA methods with selected references to reveal new facets of protein structures complementary to the existing methods. Moreover, the underlying relationship among different QA methods were analyzed and then integrated into a multilayer evaluation approach to guide the model generation and model selection in prediction. All methods are integrated and implemented into an innovative and improved prediction system hereafter referred to as MUFOLD. In CASP8 and CASP9, MUFOLD has demonstrated the proof of the principles in terms of both QA discerning power and structure prediction accuracy.

Automated protein structure modeling in CASP9 by I-TASSER pipeline combined with QUARK-based ab initio folding and FG-MD-based structure refinement

I-TASSER is an automated pipeline for protein tertiary structure prediction using multiple threading alignments and iterative structure assembly simulations. In CASP9 experiments, two new algorithms, QUARK and fragment-guided molecular dynamics (FG-MD), were added to the I-TASSER pipeline for improving the structural modeling accuracy. QUARK is a de novo structure prediction algorithm used for structure modeling of proteins that lack detectable template structures. For distantly homologous targets, QUARK models are found useful as a reference structure for selecting good threading alignments and guiding the I-TASSER structure assembly simulations. FG-MD is an atomic-level structural refinement program that uses structural fragments collected from the PDB structures to guide molecular dynamics simulation and improve the local structure of predicted model, including hydrogen-bonding networks, torsion angles, and steric clashes. Despite considerable progress in both the template-based and template-free structure modeling, significant improvements on protein target classification, domain parsing, model selection, and ab initio folding of β-proteins are still needed to further improve the I-TASSER pipeline.

Why not consider a spherical protein? Implications of backbone hydrogen bonding for protein structure and function

The intrinsic ability of protein structures to exhibit the geometric features required for molecular function in the absence of evolution is examined in the context of three systems: the reference set of real, single domain protein structures, a library of computationally generated, compact homopolypeptides, artificial structures with protein-like secondary structural elements, and quasi-spherical random proteins packed at the same density as proteins but lacking backbone secondary structure and hydrogen bonding. Without any evolutionary selection, the library of artificial structures has similar backbone hydrogen bonding, global shape, surface to volume ratio and statistically significant structural matches to real protein global structures. Moreover, these artificial structures have native like ligand binding cavities, and a tiny subset has interfacial geometries consistent with native-like protein-protein interactions and DNA binding. In contrast, the quasi-spherical random proteins, being devoid of secondary structure, have a lower surface to volume ratio and lack ligand binding pockets and intermolecular interaction interfaces. Surprisingly, these quasi-spherical random proteins exhibit protein like distributions of virtual bond angles and almost all have a statistically significant structural match to real protein structures. This implies that it is local chain stiffness, even without backbone hydrogen bonding, and compactness that give rise to the likely completeness of the library solved single domain protein structures. These studies also suggest that the packing of secondary structural elements generates the requisite geometry for intermolecular binding. Thus, backbone hydrogen bonding plays an important role not only in protein structure but also in protein function. Such ability to bind biological molecules is an inherent feature of protein structure; if combined with appropriate protein sequences, it could provide the non-zero background probability for low-level function that evolution requires for selection to occur.

Automated minimization of steric clashes in protein structures

Molecular modeling of proteins including homology modeling, structure determination, and knowledge-based protein design requires tools to evaluate and refine three-dimensional protein structures. Steric clash is one of the artifacts prevalent in low-resolution structures and homology models. Steric clashes arise due to the unnatural overlap of any two nonbonding atoms in a protein structure. Usually, removal of severe steric clashes in some structures is challenging since many existing refinement programs do not accept structures with severe steric clashes. Here, we present a quantitative approach of identifying steric clashes in proteins by defining clashes based on the Van der Waals repulsion energy of the clashing atoms. We also define a metric for quantitative estimation of the severity of clashes in proteins by performing statistical analysis of clashes in high-resolution protein structures. We describe a rapid, automated, and robust protocol, Chiron, which efficiently resolves severe clashes in low-resolution structures and homology models with minimal perturbation in the protein backbone. Benchmark studies highlight the efficiency and robustness of Chiron compared with other widely used methods. We provide Chiron as an automated web server to evaluate and resolve clashes in protein structures that can be further used for more accurate protein design.

From Coarse Grained to Atomistic: A Serial Multiscale Approach to Membrane Protein Simulations

Coarse-grained molecular dynamics provides a means for simulating the assembly and the interactions of membrane protein/lipid complexes at a reduced level of representation, allowing longer and larger simulations. We describe a fragment-based protocol for converting membrane simulation systems, comprising a membrane protein embedded in a phospholipid bilayer, from coarse-grained to atomistic resolution, for further refinement and analysis via atomistic simulations. Overall, this provides a method for generating an accurate and well equilibrated membrane protein/lipid complex. We exemplify the protocol using the acid-sensing/amiloride-sensitive ion channel protein (ASIC) channel protein, a trimeric integral membrane protein. The method is further evaluated using a test set of 10 different membrane proteins of differing size and complexity. Simulations are assessed in terms of protein conformational drift, lipid/protein interactions, and lipid dynamics.

TASSER_low-zsc: an approach to improve structure prediction using low z-score-ranked templates

In a variety of threading methods, often poorly ranked (low z-score) templates have good alignments. Here, a new method, TASSER_low-zsc that identifies these low z-score-ranked templates to improve protein structure prediction accuracy, is described. The approach consists of clustering of threading templates by affinity propagation on the basis of structural similarity (thread_cluster) followed by TASSER modeling, with final models selected by using a TASSER_QA variant. To establish the generality of the approach, templates provided by two threading methods, SP(3) and SPARKS(2), are examined. The SP(3) and SPARKS(2) benchmark datasets consist of 351 and 357 medium/hard proteins (those with moderate to poor quality templates and/or alignments) of length < or =250 residues, respectively. For SP(3) medium and hard targets, using thread_cluster, the TM-scores of the best template improve by approximately 4 and 9% over the original set (without low z-score templates) respectively; after TASSER modeling/refinement and ranking, the best model improves by approximately 7 and 9% over the best model generated with the original template set. Moreover, TASSER_low-zsc generates 22% (43%) more foldable medium (hard) targets. Similar improvements are observed with low-ranked templates from SPARKS(2). The template clustering approach could be applied to other modeling methods that utilize multiple templates to improve structure prediction.

Incorporation of local structural preference potential improves fold recognition

Fold recognition, or threading, is a popular protein structure modeling approach that uses known structure templates to build structures for those of unknown. The key to the success of fold recognition methods lies in the proper integration of sequence, physiochemical and structural information. Here we introduce another type of information, local structural preference potentials of 3-residue and 9-residue fragments, for fold recognition. By combining the two local structural preference potentials with the widely used sequence profile, secondary structure information and hydrophobic score, we have developed a new threading method called FR-t5 (fold recognition by use of 5 terms). In benchmark testings, we have found the consideration of local structural preference potentials in FR-t5 not only greatly enhances the alignment accuracy and recognition sensitivity, but also significantly improves the quality of prediction models.

Comprehensive structural and functional characterization of the human kinome by protein structure modeling and ligand virtual screening

The growing interest in the identification of kinase inhibitors, promising therapeutics in the treatment of many diseases, has created a demand for the structural characterization of the entire human kinome. At the outset of the drug development process, the lead-finding stage, approaches that enrich the screening library with bioactive compounds are needed. Here, protein structure based methods can play an important role, but despite structural genomics efforts, it is unlikely that the three-dimensional structures of the entire kinome will be available soon. Therefore, at the proteome level, structure-based approaches must rely on predicted models, with a key issue being their utility in virtual ligand screening. In this study, we employ the recently developed FINDSITE/Q-Dock ligand homology modeling approach, which is well-suited for proteome-scale applications using predicted structures, to provide extensive structural and functional characterization of the human kinome. Specifically, we construct structure models for the human kinome; these are subsequently subject to virtual screening against a library of more than 2 million compounds. To rank the compounds, we employ a hierarchical approach that combines ligand- and structure-based filters. Modeling accuracy is carefully validated using available experimental data with particularly encouraging results found for the ability to identify, without prior knowledge, specific kinase inhibitors. More generally, the modeling procedure results in a large number of predicted molecular interactions between kinases and small ligands that should be of practical use in the development of novel inhibitors. The data set is freely available to the academic community via a user-friendly Web interface at http://cssb.biology.gatech.edu/kinomelhm/ as well as at the ZINC Web site ( http://zinc.docking.org/applications/2010Apr/Brylinski-2010.tar.gz ).

FINDSITE-metal: integrating evolutionary information and machine learning for structure-based metal-binding site prediction at the proteome level

The rapid accumulation of gene sequences, many of which are hypothetical proteins with unknown function, has stimulated the development of accurate computational tools for protein function prediction with evolution/structure-based approaches showing considerable promise. In this article, we present FINDSITE-metal, a new threading-based method designed specifically to detect metal-binding sites in modeled protein structures. Comprehensive benchmarks using different quality protein structures show that weakly homologous protein models provide sufficient structural information for quite accurate annotation by FINDSITE-metal. Combining structure/evolutionary information with machine learning results in highly accurate metal-binding annotations; for protein models constructed by TASSER, whose average Cα RMSD from the native structure is 8.9 Å, 59.5% (71.9%) of the best of top five predicted metal locations are within 4 Å (8 Å) from a bound metal in the crystal structure. For most of the targets, multiple metal-binding sites are detected with the best predicted binding site at rank 1 and within the top two ranks in 65.6% and 83.1% of the cases, respectively. Furthermore, for iron, copper, zinc, calcium, and magnesium ions, the binding metal can be predicted with high, typically 70% to 90%, accuracy. FINDSITE-metal also provides a set of confidence indexes that help assess the reliability of predictions. Finally, we describe the proteome-wide application of FINDSITE-metal that quantifies the metal-binding complement of the human proteome. FINDSITE-metal is freely available to the academic community at http://cssb.biology.gatech.edu/findsite-metal/.

SwarmDock and the use of normal modes in protein-protein docking

Here is presented an investigation of the use of normal modes in protein-protein docking, both in theory and in practice. Upper limits of the ability of normal modes to capture the unbound to bound conformational change are calculated on a large test set, with particular focus on the binding interface, the subset of residues from which the binding energy is calculated. Further, the SwarmDock algorithm is presented, to demonstrate that the modelling of conformational change as a linear combination of normal modes is an effective method of modelling flexibility in protein-protein docking.

The utility of geometrical and chemical restraint information extracted from predicted ligand-binding sites in protein structure refinement

Exhaustive exploration of molecular interactions at the level of complete proteomes requires efficient and reliable computational approaches to protein function inference. Ligand docking and ranking techniques show considerable promise in their ability to quantify the interactions between proteins and small molecules. Despite the advances in the development of docking approaches and scoring functions, the genome-wide application of many ligand docking/screening algorithms is limited by the quality of the binding sites in theoretical receptor models constructed by protein structure prediction. In this study, we describe a new template-based method for the local refinement of ligand-binding regions in protein models using remotely related templates identified by threading. We designed a Support Vector Regression (SVR) model that selects correct binding site geometries in a large ensemble of multiple receptor conformations. The SVR model employs several scoring functions that impose geometrical restraints on the Cα positions, account for the specific chemical environment within a binding site and optimize the interactions with putative ligands. The SVR score is well correlated with the RMSD from the native structure; in 47% (70%) of the cases, the Pearson's correlation coefficient is >0.5 (>0.3). When applied to weakly homologous models, the average heavy atom, local RMSD from the native structure of the top-ranked (best of top five) binding site geometries is 3.1Å (2.9Å) for roughly half of the targets; this represents a 0.1 (0.3)Å average improvement over the original predicted structure. Focusing on the subset of strongly conserved residues, the average heavy atom RMSD is 2.6Å (2.3Å). Furthermore, we estimate the upper bound of template-based binding site refinement using only weakly related proteins to be ∼2.6Å RMSD. This value also corresponds to the plasticity of the ligand-binding regions in distant homologues. The Binding Site Refinement (BSR) approach is available to the scientific community as a web server that can be accessed at http://cssb.biology.gatech.edu/bsr/.

Multiscale investigation of chemical interference in proteins

We developed a multiscale approach (MultiSCAAL) that integrates the potential of mean force obtained from all-atomistic molecular dynamics simulations with a knowledge-based energy function for coarse-grained molecular simulations in better exploring the energy landscape of a small protein under chemical interference such as chemical denaturation. An excessive amount of water molecules in all-atomistic molecular dynamics simulations often negatively impacts the sampling efficiency of some advanced sampling techniques such as the replica exchange method and it makes the investigation of chemical interferences on protein dynamics difficult. Thus, there is a need to develop an effective strategy that focuses on sampling structural changes in protein conformations rather than solvent molecule fluctuations. In this work, we address this issue by devising a multiscale simulation scheme (MultiSCAAL) that bridges the gap between all-atomistic molecular dynamics simulation and coarse-grained molecular simulation. The two key features of this scheme are the Boltzmann inversion and a protein atomistic reconstruction method we previously developed (SCAAL). Using MultiSCAAL, we were able to enhance the sampling efficiency of proteins solvated by explicit water molecules. Our method has been tested on the folding energy landscape of a small protein Trp-cage with explicit solvent under 8M urea using both the all-atomistic replica exchange molecular dynamics and MultiSCAAL. We compared computational analyses on ensemble conformations of Trp-cage with its available experimental NOE distances. The analysis demonstrated that conformations explored by MultiSCAAL better agree with the ones probed in the experiments because it can effectively capture the changes in side-chain orientations that can flip out of the hydrophobic pocket in the presence of urea and water molecules. In this regard, MultiSCAAL is a promising and effective sampling scheme for investigating chemical interference which presents a great challenge when modeling protein interactions in vivo.

Interactions between a voltage sensor and a toxin via multiscale simulations

Gating-modifier toxins inhibit voltage-gated ion channels by binding the voltage sensors (VS) and altering the energetics of voltage-dependent gating. These toxins are thought to gain access to the VS via the membrane (i.e., by partitioning from water into the membrane before binding the VS). We used serial multiscale molecular-dynamics (MD) simulations, via a combination of coarse-grained (CG) and atomistic (AT) simulations, to study how the toxin VSTx1, which inhibits the archeabacterial voltage-gated potassium channel KvAP, interacts with an isolated membrane-embedded VS domain. In the CG simulations, VSTx1, which was initially located in water, partitioned into the headgroup/water interface of the lipid bilayer before binding the VS. The CG configurations were used to generate AT representations of the system, which were subjected to AT-MD to further evaluate the stability of the complex and refine the predicted VS/toxin interface. VSTx1 interacted with a binding site on the VS formed by the C-terminus of S1, the S1-S2 linker, and the N-terminus of S4. The predicted VS/toxin interactions are suggestive of toxin-mediated perturbations of the interaction between the VS and the pore domain of Kv channels, and of the membrane. Our simulations support a membrane-access mechanism of inhibition of Kv channels by VS toxins. Overall, the results show that serial multiscale MD simulations may be used to model a two-stage process of protein-bilayer and protein-protein interactions within a membrane.

Mapping the druggable allosteric space of G-protein coupled receptors: a fragment-based molecular dynamics approach

To address the problem of specificity in G-protein coupled receptor (GPCR) drug discovery, there has been tremendous recent interest in allosteric drugs that bind at sites topographically distinct from the orthosteric site. Unfortunately, structure-based drug design of allosteric GPCR ligands has been frustrated by the paucity of structural data for allosteric binding sites, making a strong case for predictive computational methods. In this work, we map the surfaces of the β₁ (β₁AR) and β₂ (β₂AR) adrenergic receptor structures to detect a series of five potentially druggable allosteric sites. We employ the FTMAP algorithm to identify ‘hot spots’ with affinity for a variety of organic probe molecules corresponding to drug fragments. Our work is distinguished by an ensemble-based approach, whereby we map diverse receptor conformations taken from molecular dynamics (MD) simulations totaling approximately 0.5 μs. Our results reveal distinct pockets formed at both solvent-exposed and lipid-exposed cavities, which we interpret in light of experimental data and which may constitute novel targets for GPCR drug discovery. This mapping data can now serve to drive a combination of fragment-based and virtual screening approaches for the discovery of small molecules that bind at these sites and which may offer highly selective therapies.

PRIMO/PRIMONA: a coarse-grained model for proteins and nucleic acids that preserves near-atomistic accuracy

The new coarse graining model PRIMO/PRIMONA for proteins and nucleic acids is proposed. This model combines one to several heavy atoms into coarse-grained sites that are chosen to allow an analytical, high-resolution reconstruction of all-atom models based on molecular bonding geometry constraints. The accuracy of proposed reconstruction method in terms of structure and energetics is tested and compared with other popular reconstruction methods for a variety of protein and nucleic acid test sets.