Assembly of polypeptide and protein backbone conformations from low energy ensembles of short fragments: development of strategies and construction of models for myoglobin, lysozyme, and thymosin beta 4.

Recently we developed methods for the construction of knowledge-based mean fields from a data base of known protein structures. As shown previously, this approach can be used to calculate ensembles of probable conformations for short fragments of polypeptide chains. Here we develop procedures for the assembly of short fragments to complete three-dimensional models of polypeptide chains. The amino acid sequence of a given protein is decomposed into all possible overlapping fragments of a given length, and an ensemble of probable conformations is calculated for each fragment. The fragments are assembled to a complete model by choosing appropriate conformations from the individual ensembles and by averaging over equivalent angles. Finally a consistent model is obtained by rebuilding the conformation from the average angles. From the average angles the local variability of the structure can be calculated, which is a useful criterion for the reliability of the model. The procedure is applied to the calculation of the local backbone conformations of myoglobin and lysozyme whose structures have been solved by X-ray analysis and thymosin beta 4, a polypeptide of 43 amino acid residues whose structure was recently investigated by NMR spectroscopy. We demonstrate that substantial fractions of the calculated local backbone conformations are similar to the experimentally determined structures.     

PPM: a side-chain and backbone chemical shift predictor for the assessment of protein conformational ensembles

The combination of the wide availability of protein backbone and side-chain NMR chemical shifts with advances in understanding of their relationship to protein structure makes these parameters useful for the assessment of structural-dynamic protein models. A new chemical shift predictor (PPM) is introduced, which is solely based on physical?Cchemical contributions to the chemical shifts for both the protein backbone and methyl-bearing amino-acid side chains. To explicitly account for the effects of protein dynamics on chemical shifts, PPM was directly refined against 100?ns long molecular dynamics (MD) simulations of 35 proteins with known experimental NMR chemical shifts. It is found that the prediction of methyl-proton chemical shifts by PPM from MD ensembles is improved over other methods, while backbone C??, C??, C??, N, and H^N chemical shifts are predicted at an accuracy comparable to the latest generation of chemical shift prediction programs. PPM is particularly suitable for the rapid evaluation of large protein conformational ensembles on their consistency with experimental NMR data and the possible improvement of protein force fields from chemical shifts.     

Correlation between local structural dynamics of proteins inferred from NMR ensembles and evolutionary dynamics of homologues of known structure

Conformational changes in proteins are extremely important for their biochemical functions. Correlation between inherent conformational variations in a protein and conformational differences in its homologues of known structure is still unclear. In this study, we have used a structural alphabet called Protein Blocks (PBs). PBs are used to perform abstraction of protein 3-D structures into a 1-D strings of 16 alphabets (a–p) based on dihedral angles of overlapping pentapeptides. We have analyzed the variations in local conformations in terms of PBs represented in the ensembles of 801 protein structures determined using NMR spectroscopy. In the analysis of concatenated data over all the residues in all the NMR ensembles, we observe that the overall nature of inherent local structural variations in NMR ensembles is similar to the nature of local structural differences in homologous proteins with a high correlation coefficient of .94. High correlation at the alignment positions corresponding to helical and β-sheet regions is only expected. However, the correlation coefficient by considering only the loop regions is also quite high (.91). Surprisingly, segregated position-wise analysis shows that this high correlation does not hold true to loop regions at the structurally equivalent positions in NMR ensembles and their homologues of known structure. This suggests that the general nature of local structural changes is unique; however most of the local structural variations in loop regions of NMR ensembles do not correlate to their local structural differences at structurally equivalent positions in homologues.     

Solution structure and proposed domain domain recognition interface of an acyl carrier protein domain from a modular polyketide synthase

Polyketides are a medicinally important class of natural products. The architecture of modular polyketide synthases (PKSs), composed of multiple covalently linked domains grouped into modules, provides an attractive framework for engineering novel polyketide-producing assemblies. However, impaired domain-domain interactions can compromise the efficiency of engineered polyketide biosynthesis. To facilitate the study of these domain-domain interactions, we have used nuclear magnetic resonance (NMR) spectroscopy to determine the first solution structure of an acyl carrier protein (ACP) domain from a modular PKS, 6-deoxyerythronolide B synthase (DEBS). The tertiary fold of this 10-kD domain is a three-helical bundle; an additional short helix in the second loop also contributes to the core helical packing. Superposition of residues 14-94 of the ensemble on the mean structure yields an average atomic RMSD of 0.64 +/- 0.09 Angstrom for the backbone atoms (1.21 +/- 0.13 Angstrom for all non-hydrogen atoms). The three major helices superimpose with a backbone RMSD of 0.48 +/- 0.10 Angstrom (0.99 +/- 0.11 Angstrom for non-hydrogen atoms). Based on this solution structure, homology models were constructed for five other DEBS ACP domains. Comparison of their steric and electrostatic surfaces at the putative interaction interface (centered on helix II) suggests a model for protein-protein recognition of ACP domains, consistent with the previously observed specificity. Site-directed mutagenesis experiments indicate that two of the identified residues influence the specificity of ACP recognition.     

Clustering algorithms for identifying core atom sets and for assessing the precision of protein structure ensembles

An important open question in the field of NMR-based biomolecular structure determination is how best to characterize the precision of the resulting ensemble of structures. Typically, the RMSD, as minimized in superimposing the ensemble of structures, is the preferred measure of precision. However, the presence of poorly determined atomic coordinates and multiple "RMSD-stable domains"--locally well-defined regions that are not aligned in global superimpositions--complicate RMSD calculations. In this paper, we present a method, based on a novel, structurally defined order parameter, for identifying a set of core atoms to use in determining superimpositions for RMSD calculations. In addition we present a method for deciding whether to partition that core atom set into "RMSD-stable domains" and, if so, how to determine partitioning of the core atom set. We demonstrate our algorithm and its application in calculating statistically sound RMSD values by applying it to a set of NMR-derived structural ensembles, superimposing each RMSD-stable domain (or the entire core atom set, where appropriate) found in each protein structure under consideration. A parameter calculated by our algorithm using a novel, kurtosis-based criterion, the epsilon-value, is a measure of precision of the superimposition that complements the RMSD. In addition, we compare our algorithm with previously described algorithms for determining core atom sets. The methods presented in this paper for biomolecular structure superimposition are quite general, and have application in many areas of structural bioinformatics and structural biology.     

The Three-Dimensional Solution Structure of the Src Homology Domain-2 of the Growth Factor Receptor-Bound Protein-2

A set of high-resolution three-dimensional solution structures of the Src homology region-2 (SH2) domain of the growth factor receptor-bound protein-2 was determined using heteronuclear NMR spectroscopy. The NMR data used in this study were collected on a stable monomeric protein solution that was free of protein aggregates and proteolysis. The solution structure was determined based upon a total of 1439 constraints, which included 1326 nuclear Overhauser effect distance constraints, 70 hydrogen bond constraints, and 43 dihedral angle constraints. Distance geometry-simulated annealing calculations followed by energy minimization yielded a family of 18 structures that converged to a root-mean-square deviation of 1.09 Å for all backbone atoms and 0.40 Å for the backbone atoms of the central -sheet. The core structure of the SH2 domain contains an antiparallel -sheet flanked by two parallel -helices displaying an overall architecture that is similar to other known SH2 domain structures. This family of NMR structures is compared to the X-ray structure and to another family of NMR solution structures determined under different solution conditions.     

Ab initio construction of polypeptide fragments: efficient generation of accurate,representative ensembles

We describe a novel method to generate ensembles of conformations of the main-chain atoms [N, C(alpha), C, O, Cbeta] for a sequence of amino acids within the context of a fixed protein framework. Each conformation satisfies fundamental stereo-chemical restraints such as idealized geometry, favorable phi/psi angles, and excluded volume. The ensembles include conformations both near and far from the native structure. Algorithms for effective conformational sampling and constant time overlap detection permit the generation of thousands of distinct conformations in minutes. Unlike previous approaches, our method samples dihedral angles from fine-grained phi/psi state sets, which we demonstrate is superior to exhaustive enumeration from coarse phi/psi sets. Applied to a large set of loop structures, our method samples consistently near-native conformations, averaging 0.4, 1.1, and 2.2 A main-chain root-mean-square deviations for four, eight, and twelve residue long loops, respectively. The ensembles make ideal decoy sets to assess the discriminatory power of a selection method. Using these decoy sets, we conclude that quality of anchor geometry cannot reliably identify near-native conformations, though the selection results are comparable to previous loop prediction methods. In a subsequent study (de Bakker et al.: Proteins 2003;51:21-40), we demonstrate that the AMBER forcefield with the Generalized Born solvation model identifies near-native conformations significantly better than previous methods.     

Improving the accuracy of protein stability predictions with multistate design using a variety of backbone ensembles

Multistate computational protein design (MSD) with backbone ensembles approximating conformational flexibility can predict higher quality sequences than single‐state design with a single fixed backbone. However, it is currently unclear what characteristics of backbone ensembles are required for the accurate prediction of protein sequence stability. In this study, we aimed to improve the accuracy of protein stability predictions made with MSD by using a variety of backbone ensembles to recapitulate the experimentally measured stability of 85 Streptococcal protein G domain β1 sequences. Ensembles tested here include an NMR ensemble as well as those generated by molecular dynamics (MD) simulations, by Backrub motions, and by PertMin, a new method that we developed involving the perturbation of atomic coordinates followed by energy minimization. MSD with the PertMin ensembles resulted in the most accurate predictions by providing the highest number of stable sequences in the top 25, and by correctly binning sequences as stable or unstable with the highest success rate (≈90%) and the lowest number of false positives. The performance of PertMin ensembles is due to the fact that their members closely resemble the input crystal structure and have low potential energy. Conversely, the NMR ensemble as well as those generated by MD simulations at 500 or 1000 K reduced prediction accuracy due to their low structural similarity to the crystal structure. The ensembles tested herein thus represent on‐ or off‐target models of the native protein fold and could be used in future studies to design for desired properties other than stability. Proteins 2014; 82:771–784. © 2013 Wiley Periodicals, Inc.     

Hash: a program to accurately predict protein Hα shifts from neighboring backbone shifts

Chemical shifts provide not only peak identities for analyzing nuclear magnetic resonance (NMR) data, but also an important source of conformational information for studying protein structures. Current structural studies requiring H^α chemical shifts suffer from the following limitations. (1) For large proteins, the H^α chemical shifts can be difficult to assign using conventional NMR triple-resonance experiments, mainly due to the fast transverse relaxation rate of C^α that restricts the signal sensitivity. (2) Previous chemical shift prediction approaches either require homologous models with high sequence similarity or rely heavily on accurate backbone and side-chain structural coordinates. When neither sequence homologues nor structural coordinates are available, we must resort to other information to predict H^α chemical shifts. Predicting accurate H^α chemical shifts using other obtainable information, such as the chemical shifts of nearby backbone atoms (i.e., adjacent atoms in the sequence), can remedy the above dilemmas, and hence advance NMR-based structural studies of proteins. By specifically exploiting the dependencies on chemical shifts of nearby backbone atoms, we propose a novel machine learning algorithm, called Hash, to predict H^α chemical shifts. Hash combines a new fragment-based chemical shift search approach with a non-parametric regression model, called the generalized additive model, to effectively solve the prediction problem. We demonstrate that the chemical shifts of nearby backbone atoms provide a reliable source of information for predicting accurate H^α chemical shifts. Our testing results on different possible combinations of input data indicate that Hash has a wide rage of potential NMR applications in structural and biological studies of proteins.     

Protein φ and ψ dihedral restraints determined from multidimensional hypersurface correlations of backbone chemical shifts and their use in the determination of protein tertiary structures

The chemical shifts of the backbone atoms of proteins can be used to obtainrestraints that can be incorporated into structure determination methods. Eachchemical shift can be used to define a restraint and these restraints can besimultaneously used to define the local, secondary structure features. Theglobal fold can be determined by a combined use of the chemical shift basedrestraints along with the long-range information present in the NOEs ofpartially deuterated proteins or the amide–amide NOEs but not from suchlimited NOE data sets alone. This approach has been demonstrated to be capableof determining the overall folding pattern of four proteins. This suggeststhat solution-state NMR methods can be extended to the structure determinationof larger proteins by using the information present in the chemical shifts ofthe backbone atoms along with the data that can be obtained on a small numberof labeled forms.     

Protein structure prediction assisted with sparse NMR data in CASP13

CASP13 has investigated the impact of sparse NMR data on the accuracy of protein structure prediction. NOESY and ¹⁵N-¹H residual dipolar coupling data, typical of that obtained for ¹⁵N,¹³C-enriched, perdeuterated proteins up to about 40 kDa, were simulated for 11 CASP13 targets ranging in size from 80 to 326 residues. For several targets, two prediction groups generated models that are more accurate than those produced using baseline methods. Real NMR data collected for a de novo designed protein were also provided to predictors, including one data set in which only backbone resonance assignments were available. Some NMR-assisted prediction groups also did very well with these data. CASP13 also assessed whether incorporation of sparse NMR data improves the accuracy of protein structure prediction relative to nonassisted regular methods. In most cases, incorporation of sparse, noisy NMR data results in models with higher accuracy. The best NMR-assisted models were also compared with the best regular predictions of any CASP13 group for the same target. For six of 13 targets, the most accurate model provided by any NMR-assisted prediction group was more accurate than the most accurate model provided by any regular prediction group; however, for the remaining seven targets, one or more regular prediction method provided a more accurate model than even the best NMR-assisted model. These results suggest a novel approach for protein structure determination, in which advanced prediction methods are first used to generate structural models, and sparse NMR data is then used to validate and/or refine these models.     

Influence of non-bonded parameters on the quality of NMR structures: A new force field for NMR structure calculation

The effects of different non-bonded parameters of force fields for NMR structure calculation on the quality of the resulting NMR solution structures were investigated using Interleukin 4 as a model system. NMR structure ensembles were calculated with an ab initio protocol using torsion angle dynamics. The calculations were repeated with five different non-bonded energy functions and parameters. The resulting ensembles were compared with the available X-ray structures, and their quality was assessed with common structure validation programs. In addition, the impact of torsion angle restraints and dihedral energy terms for the sidechains and the backbone was studied. The further improvement of the quality by refinement in explicit solvent was demonstrated. The optimal parameters, including those necessary for water refinement, are available in the new version of the PARALLHDG force field.     

Solution NMR and molecular dynamics reveal a persistent alpha helix within the dynamic region of PsbQ from photosystem II of higher plants

The extrinsic proteins of photosystem II of higher plants and green algae PsbO, PsbP, PsbQ, and PsbR are essential for stable oxygen production in the oxygen evolving center. In the available X‐ray crystallographic structure of higher plant PsbQ residues S14‐Y33 are missing. Building on the backbone NMR assignment of PsbQ, which includes this “missing link”, we report the extended resonance assignment including side chain atoms. Based on nuclear Overhauser effect spectra a high resolution solution structure of PsbQ with a backbone RMSD of 0.81 Å was obtained from torsion angle dynamics. Within the N‐terminal residues 1–45 the solution structure deviates significantly from the X‐ray crystallographic one, while the four‐helix bundle core found previously is confirmed. A short α‐helix is observed in the solution structure at the location where a β‐strand had been proposed in the earlier crystallographic study. NMR relaxation data and unrestrained molecular dynamics simulations corroborate that the N‐terminal region behaves as a flexible tail with a persistent short local helical secondary structure, while no indications of forming a β‐strand are found. Proteins 2015; 83:1677–1686. © 2015 The Authors. Proteins: Structure, Function, and Bioinformatics Published by Wiley Periodicals, Inc.     

Simple and versatile restraints for the accurate modeling of alpha-helical coiled-coil structures of multiple strandedness, orientation and composition

We present a minimalist approach for the modeling of the three-dimensional structure of multistranded alpha-helical coiled coils. The approach is based on empirical principles introduced by F. H. C. Crick (F. H. C. Crick, Acta Crystallogr, 1953, Vol. 6, pp. 689-697). Crick hypothesized that keeping the distance between the residues at the interacting interface of alpha-helices constant would lead to supercoiling or the formation of a coiled coil through the knobs-into-holes mode of packing. We have implemented the latter hypothesis in a simulating annealing protocol in the simple form of interhelical distance restraints (two per heptad) between Calpha at the interfacial positions and. To demonstrate the authenticity of Crick's hypothesis and the precision and accuracy of our approach, we have modeled the crystal structures of six synthetic coiled coils in dimeric, trimeric, and tetrameric states. The mean root mean square deviations (RMSDs) between the backbone atoms of the ensemble of structures calculated and those of the corresponding geometric averages is always below 0.76 A, indicating that our protocol has an excellent degree of convergence and precision. The RMSDs between the backbone atoms of each of the six geometric average structures and the backbone of the corresponding crystal structures all range between 0.43 and 0.95 A, indicative of excellent accuracy and proving the authenticity of Crick's hypothesis. Moreover, without specifying any dihedral angles, we found that in 81% of the occurrences, the most populated conformer of the side chains at positions and in the ensembles calculated were identical to those observed in the crystal structures. This shows that our simple approach, which is the simplest reported so far, can generate accurate results for the backbone and side chains. Finally, as a test case for a wider application of our approach in the field of structural proteomics, we describe the successful modeling of the overall structure of SNARE and the organization of its interfacial ionic layer known to play an important functional role.     

REMO: A new protocol to refine full atomic protein models from C‐alpha traces by optimizing hydrogen‐bonding networks

Protein structure prediction approaches usually perform modeling simulations based on reduced representation of protein structures. For biological utilizations, it is an important step to construct full atomic models from the reduced structure decoys. Most of the current full atomic model reconstruction procedures have defects which either could not completely remove the steric clashes among backbone atoms or generate final atomic models with worse topology similarity relative to the native structures than the reduced models. In this work, we develop a new protocol, called REMO, to generate full atomic protein models by optimizing the hydrogen‐bonding network with basic fragments matched from a newly constructed backbone isomer library of solved protein structures. The algorithm is benchmarked on 230 nonhomologous proteins with reduced structure decoys generated by I‐TASSER simulations. The results show that REMO has a significant ability to remove steric clashes, and meanwhile retains good topology of the reduced model. The hydrogen‐bonding network of the final models is dramatically improved during the procedure. The REMO algorithm has been exploited in the recent CASP8 experiment which demonstrated significant improvements of the I‐TASSER models in both atomic‐level structural refinement and hydrogen‐bonding network construction. Proteins 2009. © 2009 Wiley‐Liss, Inc.     

Three-dimensional structure of the neurotoxin ATX Ia from Anemonia sulcata in aqueous solution determined by nuclear magnetic resonance spectroscopy

With the aid of 1H nuclear magnetic resonance (NMR) spectroscopy, the three-dimensional structure in aqueous solution was determined for ATX Ia, which is a 46 residue polypeptide neurotoxin of the sea anemone Anemonia sulcata. The input for the structure calculations consisted of 263 distance constraints from nuclear Overhauser effects (NOE) and 76 vicinal coupling constants. For the structure calculation several new or ammended programs were used in a revised strategy consisting of five successive computational steps. First, the program HABAS was used for a complete search of all backbone and chi 1 conformations that are compatible with the intraresidual and sequential NMR constraints. Second, using the program DISMAN, we extended this approach to pentapeptides by extensive sampling of all conformations that are consistent with the local and medium-range NMR constraints. Both steps resulted in the definition of additional dihedral angle constraints and in stereospecific assignments for a number of beta-methylene groups. In the next two steps DISMAN was used to obtain a group of eight conformers that contain no significant residual violations of the NMR constraints or van der Waals contacts. Finally, these structures were subjected to restrained energy refinement with a modified version of the molecular mechanics module of AMBER, which in addition to the energy force field includes potentials for the NOE distance constraints and the dihedral angle constraints. The average of the pairwise minimal RMS distances between the resulting refined conformers calculated for the well defined molecular core, which contains the backbone atoms of 35 residues and 20 interior side chains, is 1.5 +/- 0.3 A. This core is formed by a four-stranded beta-sheet connected by two well-defined loops, and there is an additional flexible loop consisting of the eleven residues 8-18. The core of the protein is stabilized by three disulfide bridges, which are surrounded by hydrophobic residues and shielded on one side by hydrophilic residues.     

In silico predictions of 3D structures of linear and cyclic peptides with natural and non‐proteinogenic residues

We extended the use of Peplook, an in silico procedure for the prediction of three‐dimensional (3D) models of linear peptides to the prediction of 3D models of cyclic peptides and thanks to the ab initio calculation procedure, to the calculation of peptides with non‐proteinogenic amino acids. Indeed, such peptides cannot be predicted by homology or threading. We compare the calculated models with NMR and X‐ray models and for the cyclic peptides, with models predicted by other in silico procedures (Pep‐Fold and I‐Tasser). For cyclic peptides, on a set of 38 peptides, average root mean square deviation of backbone atoms (BB‐RMSD) was 3.8 and 4.1 Å for Peplook and Pep‐Fold, respectively. The best results are obtained with I‐Tasser (2.5 Å) although evaluations were biased by the fact that the resolved Protein Data Bank models could be used as template by the server. Peplook and Pep‐Fold give similar results, better for short (up to 20 residues) than for longer peptides. For peptides with non‐proteinogenic residues, performances of Peplook are sound with an average BB‐RMSD of 3.6 Å for ‘non‐natural peptides’ and 3.4 Å for peptides combining non‐proteinogenic residues and cyclic structure. These results open interesting possibilities for the design of peptidic drugs. Copyright © 2011 European Peptide Society and John Wiley & Sons, Ltd.     

Antibody modeling assessment

A blinded study to assess the state of the art in three‐dimensional structure modeling of the variable region (Fv) of antibodies was conducted. Nine unpublished high‐resolution x‐ray Fab crystal structures covering a wide range of antigen‐binding site conformations were used as benchmark to compare Fv models generated by four structure prediction methodologies. The methodologies included two homology modeling strategies independently developed by CCG (Chemical Computer Group) and Accerlys Inc, and two fully automated antibody modeling servers: PIGS (Prediction of ImmunoGlobulin Structure), based on the canonical structure model, and Rosetta Antibody Modeling, based on homology modeling and Rosetta structure prediction methodology. The benchmark structure sequences were submitted to Accelrys and CCG and a set of models for each of the nine antibody structures were generated. PIGS and Rosetta models were obtained using the default parameters of the servers. In most cases, we found good agreement between the models and x‐ray structures. The average rmsd (root mean square deviation) values calculated over the backbone atoms between the models and structures were fairly consistent, around 1.2 Å. Average rmsd values of the framework and hypervariable loops with canonical structures (L1, L2, L3, H1, and H2) were close to 1.0 Å. H3 prediction yielded rmsd values around 3.0 Å for most of the models. Quality assessment of the models and the relative strengths and weaknesses of the methods are discussed. We hope this initiative will serve as a model of scientific partnership and look forward to future antibody modeling assessments. Proteins 2011; © 2011 Wiley‐Liss, Inc.     

Evidence of Conformational Selection Driving the Formation of Ligand Binding Sites in Protein-Protein Interfaces

Many protein-protein interactions (PPIs) are compelling targets for drug discovery, and in a number of cases can be disrupted by small molecules. The main goal of this study is to examine the mechanism of binding site formation in the interface region of proteins that are PPI targets by comparing ligand-free and ligand-bound structures. To avoid any potential bias, we focus on ensembles of ligand-free protein conformations obtained by nuclear magnetic resonance (NMR) techniques and deposited in the Protein Data Bank, rather than on ensembles specifically generated for this study. The measures used for structure comparison are based on detecting binding hot spots, i.e., protein regions that are major contributors to the binding free energy. The main tool of the analysis is computational solvent mapping, which explores the surface of proteins by docking a large number of small “probe” molecules. Although we consider conformational ensembles obtained by NMR techniques, the analysis is independent of the method used for generating the structures. Finding the energetically most important regions, mapping can identify binding site residues using ligand-free models based on NMR data. In addition, the method selects conformations that are similar to some peptide-bound or ligand-bound structure in terms of the properties of the binding site. This agrees with the conformational selection model of molecular recognition, which assumes such pre-existing conformations. The analysis also shows the maximum level of similarity between unbound and bound states that is achieved without any influence from a ligand. Further shift toward the bound structure assumes protein-peptide or protein-ligand interactions, either selecting higher energy conformations that are not part of the NMR ensemble, or leading to induced fit. Thus, forming the sites in protein-protein interfaces that bind peptides and can be targeted by small ligands always includes conformational selection, although other recognition mechanisms may also be involved.