Integrative Modelling of Biomolecular Complexes.

In recent years, the use of integrative, information-driven computational approaches for modeling the structure of biomolecules has been increasing in popularity. These are now recognized as a crucial complement to experimental structural biology techniques such as X-ray crystallography, nuclear magnetic resonance (NMR) spectroscopy and cryo-electron microscopy (cryo-EM). This trend can be credited to a few reasons such as the increased prominence of structures solved by cryo-EM, the improvements in proteomics approaches such as cross-linking mass spectrometry (XL-MS), the drive to study systems of higher complexity in their native state, and the maturation of many computational techniques combined with the widespread availability of information-driven integrative modeling platforms. In this review, we highlight recent works that exemplify how the use of integrative and/or information-driven approaches and platforms can produce highly accurate structural models. These examples include systems which present many challenges when studied with traditional structural biology techniques such as flexible and dynamic macromolecular assemblies and membrane-associated complexes. We also identify some key areas of interest for information-driven, integrative modeling and discuss how they relate to ongoing challenges in the fields of computational structural biology. These include the use of coarse-grained force fields for biomolecular simulations-allowing for simulations across longer (time-) and bigger (size-dimension) scales-the use of bioinformatics predictions to drive sampling and/or scoring in docking such as those derived from coevolution analysis and finally the study of membrane and membrane-associated protein complexes.

Performance of HADDOCK and a simple contact-based protein-ligand binding affinity predictor in the D3R Grand Challenge 2.

We present the performance of HADDOCK, our information-driven docking software, in the second edition of the D3R Grand Challenge. In this blind experiment, participants were requested to predict the structures and binding affinities of complexes between the Farnesoid X nuclear receptor and 102 different ligands. The models obtained in Stage1 with HADDOCK and ligand-specific protocol show an average ligand RMSD of 5.1 Å from the crystal structure. Only 6/35 targets were within 2.5 Å RMSD from the reference, which prompted us to investigate the limiting factors and revise our protocol for Stage2. The choice of the receptor conformation appeared to have the strongest influence on the results. Our Stage2 models were of higher quality (13 out of 35 were within 2.5 Å), with an average RMSD of 4.1 Å. The docking protocol was applied to all 102 ligands to generate poses for binding affinity prediction. We developed a modified version of our contact-based binding affinity predictor PRODIGY, using the number of interatomic contacts classified by their type and the intermolecular electrostatic energy. This simple structure-based binding affinity predictor shows a Kendall's Tau correlation of 0.37 in ranking the ligands (7th best out of 77 methods, 5th/25 groups). Those results were obtained from the average prediction over the top10 poses, irrespective of their similarity/correctness, underscoring the robustness of our simple predictor. This results in an enrichment factor of 2.5 compared to a random predictor for ranking ligands within the top 25%, making it a promising approach to identify lead compounds in virtual screening.

Sense and simplicity in HADDOCK scoring: Lessons from CASP-CAPRI round 1.

Our information-driven docking approach HADDOCK is a consistent top predictor and scorer since the start of its participation in the CAPRI community-wide experiment. This sustained performance is due, in part, to its ability to integrate experimental data and/or bioinformatics information into the modelling process, and also to the overall robustness of the scoring function used to assess and rank the predictions. In the CASP-CAPRI Round 1 scoring experiment we successfully selected acceptable/medium quality models for 18/14 of the 25 targets - a top-ranking performance among all scorers. Considering that for only 20 targets acceptable models were generated by the community, our effective success rate reaches as high as 90% (18/20). This was achieved using the standard HADDOCK scoring function, which, thirteen years after its original publication, still consists of a simple linear combination of intermolecular van der Waals and Coulomb electrostatics energies and an empirically derived desolvation energy term. Despite its simplicity, this scoring function makes sense from a physico-chemical perspective, encoding key aspects of biomolecular recognition. In addition to its success in the scoring experiment, the HADDOCK server takes the first place in the server prediction category, with 16 successful predictions. Much like our scoring protocol, because of the limited time per target, the predictions relied mainly on either an ab initio center-of-mass and symmetry restrained protocol, or on a template-based approach whenever applicable. These results underline the success of our simple but sensible prediction and scoring scheme. Proteins 2017; 85:417-423. © 2016 Wiley Periodicals, Inc.

The DisVis and PowerFit Web Servers: Explorative and Integrative Modeling of Biomolecular Complexes.

Structure determination of complex molecular machines requires a combination of an increasing number of experimental methods with highly specialized software geared toward each data source to properly handle the gathered data. Recently, we introduced the two software packages PowerFit and DisVis. These combine high-resolution structures of atomic subunits with density maps from cryo-electron microscopy or distance restraints, typically acquired by chemical cross-linking coupled with mass spectrometry, respectively. Here, we report on recent advances in both GPGPU-accelerated software packages: PowerFit is a tool for rigid body fitting of atomic structures in cryo-electron density maps and has been updated to also output reliability indicators for the success of fitting, through the use of the Fisher z-transformation and associated confidence intervals; DisVis aims at quantifying the information content of distance restraints and identifying false-positive restraints. We extended its analysis capabilities to include an analysis of putative interface residues and to output an average shape representing the putative location of the ligand. To facilitate their use by a broad community, they have been implemented as web portals harvesting both local CPU resources and GPGPU-accelerated EGI grid resources. They offer user-friendly interfaces, while minimizing computational requirements, and provide a first interactive view of the results. The portals can be accessed freely after registration via http://milou.science.uu.nl/services/DISVIS and http://milou.science.uu.nl/services/POWERFIT.

Defining the limits and reliability of rigid-body fitting in cryo-EM maps using multi-scale image pyramids.

Cryo-electron microscopy provides fascinating structural insight into large macromolecular machines at increasing detail. Despite significant advances in the field, the resolution of the resulting three-dimensional images is still typically insufficient for de novo model building. To bridge the resolution gap and give an atomic interpretation to the data, high-resolution models are typically placed into the density as rigid bodies. Unfortunately, this is often done manually using graphics software, a subjective method that can lead to over-interpretation of the data. A more objective approach is to perform an exhaustive cross-correlation-based search to fit subunits into the density. Here we show, using five experimental ribosome maps ranging in resolution from 5.5 to 6.9Å, that cross-correlation-based fitting is capable of successfully fitting subunits correctly in the density for over 90% of the cases. Importantly, we provide indicators for the reliability and ambiguity of a fit, using the Fisher z-transformation and its associated confidence intervals, giving a formal approach to identify over-interpreted regions in the density. In addition, we quantify the resolution requirement for a successful fit as a function of the subunit size. For larger subunits the resolution of the data can be down-filtered to 20Å while still retaining an unambiguous fit. We leverage this information through the use of multi-scale image pyramids to accelerate the search up to 30-fold on CPUs and 40-fold on GPUs at a negligible loss in success rate. We implemented this approach in our rigid-body fitting software PowerFit, which can be freely downloaded from https://github.com/haddocking/powerfit.

The HADDOCK2.2 Web Server: User-Friendly Integrative Modeling of Biomolecular Complexes.

The prediction of the quaternary structure of biomolecular macromolecules is of paramount importance for fundamental understanding of cellular processes and drug design. In the era of integrative structural biology, one way of increasing the accuracy of modeling methods used to predict the structure of biomolecular complexes is to include as much experimental or predictive information as possible in the process. This has been at the core of our information-driven docking approach HADDOCK. We present here the updated version 2.2 of the HADDOCK portal, which offers new features such as support for mixed molecule types, additional experimental restraints and improved protocols, all of this in a user-friendly interface. With well over 6000 registered users and 108,000 jobs served, an increasing fraction of which on grid resources, we hope that this timely upgrade will help the community to solve important biological questions and further advance the field. The HADDOCK2.2 Web server is freely accessible to non-profit users at http://haddock.science.uu.nl/services/HADDOCK2.2.

DisVis: quantifying and visualizing accessible interaction space of distance-restrained biomolecular complexes.

UNLABELLED: We present DisVis, a Python package and command line tool to calculate the reduced accessible interaction space of distance-restrained binary protein complexes, allowing for direct visualization and quantification of the information content of the distance restraints. The approach is general and can also be used as a knowledge-based distance energy term in FFT-based docking directly during the sampling stage. AVAILABILITY AND IMPLEMENTATION: The source code with documentation is freely available from https://github.com/haddocking/disvis. CONTACT: a.m.j.j.bonvin@uu.nl SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

1H, 13C and 15N assignment of the GNA1946 outer membrane lipoprotein from Neisseria meningitidis.

GNA1946 (Genome-derived Neisseria Antigen 1946) is a highly conserved exposed outer membrane lipoprotein from Neisseria meningitidis bacteria of 287 amino acid length (31 kDa). Although the structure of NMB1946 has been solved recently by X-Ray crystallography, understanding the behaviour of GNA1946 in aqueuos solution is highly relevant for the discovery of the antigenic determinants of the protein that will possibly lead to a more efficient vaccine development against virulent serogroup B strain of N. meningitidis. Here we report almost complete (1)H, (13)C and (15)N resonance assignments of GNA1946 (residues 10-287) in aqueous buffer solution.

Two-rung model of a left-handed beta-helix for prions explains species barrier and strain variation in transmissible spongiform encephalopathies.

In this study, a new beta-helical model is proposed that explains the species barrier and strain variation in transmissible spongiform encephalopathies. The left-handed beta-helix serves as a structural model that can explain the seeded growth characteristics of beta-sheet structure in PrP(Sc) fibrils. Molecular dynamics simulations demonstrate that the left-handed beta-helix is structurally more stable than the right-handed beta-helix, with a higher beta-sheet content during the simulation and a better distributed network of inter-strand backbone-backbone hydrogen bonds between parallel beta-strands of different rungs. Multiple sequence alignments and homology modelling of prion sequences with different rungs of left-handed beta-helices illustrate that the PrP region with the highest beta-helical propensity (residues 105-143) can fold in just two rungs of a left-handed beta-helix. Even if no other flanking sequence participates in the beta-helix, the two rungs of a beta-helix can give the growing fibril enough elevation to accommodate the rest of the PrP protein in a tight packing at the periphery of a trimeric beta-helix. The folding of beta-helices is driven by backbone-backbone hydrogen bonding and stacking of side-chains in adjacent rungs. The sequence and structure of the last rung at the fibril end with unprotected beta-sheet edges selects the sequence of a complementary rung and dictates the folding of the new rung with optimal backbone hydrogen bonding and side-chain stacking. An important side-chain stack that facilitates the beta-helical folding is between methionine residues 109 and 129, which explains their importance in the species barrier of prions. Because the PrP sequence is not evolutionarily optimised to fold in a beta-helix, and because the beta-helical fold shows very little sequence preference, alternative alignments are possible that result in a different rung able to select for an alternative complementary rung. A different top rung results in a new strain with different growth characteristics. Hence, in the present model, sequence variation and alternative alignments clarify the basis of the species barrier and strain specificity in PrP-based diseases.

Data-driven docking: HADDOCK's adventures in CAPRI.

We have shown previously that given high-resolution structures of the unbound molecules, structure determination of protein complexes is possible by including biochemical and/or biophysical data as highly ambiguous distance restraints in a docking approach. We applied this method, implemented in the HADDOCK (High Ambiguity Driven DOCKing) package (Dominguez et al., J Am Chem Soc 2003;125:1731-1737), to the targets in the fourth and fifth rounds of CAPRI. Here we describe our results and analyze them in detail. Special attention is given to the role of flexibility in our docking method and the way in which this improves the docking results. We describe extensions to our approach that were developed as a direct result of our participation in CAPRI. In addition to experimental information, we also included interface residue predictions from PPISP (Protein-Protein Interaction Site Predictor; Zhou and Shan, Proteins 2001;44:336-343), a neural network method. Using HADDOCK we were able to generate acceptable structures for 6 of the 8 targets, and to submit at least 1 acceptable structure for 5 of them. Of these 5 submissions, 3 were of medium quality (Targets 10, 11, and 15) and 2 of high quality (Targets 13 and 14). In all cases, predictions were obtained containing at least 40% of the correct epitope at the interface for both ligand and receptor simultaneously.

Solution structure and DNA-binding properties of the C-terminal domain of UvrC from E.coli.

The C-terminal domain of the UvrC protein (UvrC CTD) is essential for 5' incision in the prokaryotic nucleotide excision repair process. We have determined the three-dimensional structure of the UvrC CTD using heteronuclear NMR techniques. The structure shows two helix-hairpin-helix (HhH) motifs connected by a small connector helix. The UvrC CTD is shown to mediate structure-specific DNA binding. The domain binds to a single-stranded-double-stranded junction DNA, with a strong specificity towards looped duplex DNA that contains at least six unpaired bases per loop ("bubble DNA"). Using chemical shift perturbation experiments, the DNA-binding surface is mapped to the first hairpin region encompassing the conserved glycine-valine-glycine residues followed by lysine-arginine-arginine, a positively charged surface patch and the second hairpin region consisting of glycine-isoleucine-serine. A model for the protein-DNA complex is proposed that accounts for this specificity.

Rapid protein fold determination using secondary chemical shifts and cross-hydrogen bond 15N-13C' scalar couplings (3hbJNC').

The possibility of generating protein folds at the stage of backbone assignment using structural restraints derived from experimentally measured cross-hydrogen bond scalar couplings and secondary chemical shift information is investigated using as a test case the small alpha/beta protein chymotrypsin inhibitor 2. Dihedral angle restraints for the phi and psi angles of 32 out of 64 residues could be obtained from secondary chemical shift analysis with the TALOS program (Corneliscu et al., 1999a). This information was supplemented by 18 hydrogen-bond restraints derived from experimentally measured cross-hydrogen bond 3hbJNC' coupling constants. These experimental data were sufficient to generate structures that are as close as 1.0 A backbone rmsd from the crystal structure. The fold is, however, not uniquely defined and several solutions are generated that cannot be distinguished on the basis of violations or energetic considerations. Correct folds could be identified by combining clustering methods with knowledge-based potentials derived from structural databases.

Hydration dynamics of the collagen triple helix by NMR.

The hydration of the collagen-like Ac-(Gly-Pro-Hyp)(6)-NH(2) triple-helical peptide in solution was investigated using an integrated set of high-resolution NMR hydration experiments, including different recently developed exchange-network editing methods. This approach was designed to explore the hydration dynamics in the proximity of labile groups, such as the hydroxyproline hydroxyl group, and revealed that the first shell of hydration in collagen-like triple helices is kinetically labile with upper limits for water molecule residence times in the nanosecond to sub-nanosecond range. This result is consistent with a "hopping" hydration model in which solvent molecules are exchanged in and out of solvation sites at a rate that is not directly correlated to the degree of site localization. The hopping model thus reconciles the dynamic view of hydration revealed by NMR with the previously suggested partially ordered semi-clathrate-like cylinder of hydration. In addition, the nanosecond to sub-nanosecond upper limits for water molecule residence times imply that hydration-dehydration events are not likely to be the rate-limiting step for triple helix self-recognition, complementing previous investigations on water dynamics in collagen fibers. This study has also revealed labile proton features expected to facilitate the characterization of the structure and folding of triple helices in collagen peptides.

Changes in dynamical behavior of the retinoid X receptor DNA-binding domain upon binding to a 14 base-pair DNA half site.

The retinoid X receptor (RXR) is a prominent member of the nuclear receptor family of ligand-inducible transcription factors. Many proteins of this family exert their function as heterodimers with RXR as a common upstream partner. Studies of the DNA-binding domains of several nuclear receptors reveal differences in structure and dynamics, both between the different proteins and between the free- and DNA-bound receptor DBDs. We investigated the differences in dynamics between RXR free in solution and in complex with a 14 base-pair oligonucleotide, using (1)H and (15)N relaxation studies. Nano- to picosecond dynamics were probed on (15)N, employing Lipari-Szabo analysis with an axially symmetric tumbling model to estimate the exchange contributions to the transverse relaxation rates. Furthermore, milli- to microsecond dynamics were estimated qualitatively for (1)H and (15)N, using CPMG-HSQC and CPMG-T(2) measurements with differential pulse spacing. RXR shows hardly any nano- to picosecond time-scale internal motion. Upon DNA binding, the order parameters show a tiny increase. Dynamics in the milli- to microsecond time scale is more prevalent. It is localized in the first and second zinc fingers of the free RXR. Upon DNA-binding, exchange associated with specific/aspecific DNA-binding of RXR is observed throughout the sequence, whereas conformational flexibility of the D-box and the second zinc finger of RXR is greatly reduced. Since this DNA-binding induced folding transition occurs remote from the DNA in a region which is involved in protein-protein interactions, it may very well be related to the cooperativity of dimeric DNA binding.

Localisation and dynamics of sodium counterions around DNA in solution from molecular dynamics simulation.

The localisation and dynamics of sodium counterions around the DNA duplex d(AGCGTACTAGTACGCT)2 corresponding to the trp operator fragment used in the crystal structure of the half site complex has been studied by a 1.4 ns molecular dynamics simulation in explicit solvent. A continuous and well-defined counterion density is shown to be present around the minor groove, while density patches are found in the major groove in regions where DNA bending is observed. A residence time analysis reveals the dynamic nature of these distributions. The resulting picture agrees with previous theoretical and experimental studies of A-tract DNA sequences, and is consistent with the polyelectrolyte condensation model.

beta-hairpin stability and folding: molecular dynamics studies of the first beta-hairpin of tendamistat.

The stability and (un)folding of the 19-residue peptide, SCVTLYQSWRYSQADNGCA, corresponding to the first beta-hairpin (residues 10 to 28) of the alpha-amylase inhibitor tendamistat (PDB entry 3AIT) has been studied by molecular dynamics simulations in explicit water under periodic boundary conditions at several temperatures (300 K, 360 K and 400 K), starting from various conformations for simulation lengths, ranging from 10 to 30 ns. Comparison of trajectories of the reduced and oxidized native peptides reveals the importance of the disulphide bridge closing the beta-hairpin in maintaining a proper turn conformation, thereby insuring a proper side-chain arrangement of the conserved turn residues. This allows rationalization of the conservation of those cysteine residues among the family of alpha-amylase inhibitors. High temperature simulations starting from widely different initial configurations (native beta-hairpin, alpha and left-handed helical and extended conformations) begin sampling similar regions of the conformational space within tens of nanoseconds, and both native and non-native beta-hairpin conformations are recovered. Transitions between conformational clusters are accompanied by an increase in energy fluctuations, which is consistent with the increase in heat capacity measured experimentally upon protein folding. The folding events observed in the various simulations support a model for beta-hairpin formation in which the turn is formed first, followed by hydrogen bond formation closing the hairpin, and subsequent stabilization by side-chain hydrophobic interactions.

The solution structure of Lac repressor headpiece 62 complexed to a symmetrical lac operator.

BACKGROUND: Lactose repressor protein (Lac) controls the expression of the lactose metabolic genes in Escherichia coli by binding to an operator sequence in the promoter of the lac operon. Binding of inducer molecules to the Lac core domain induces changes in tertiary structure that are propagated to the DNA-binding domain through the connecting hinge region, thereby reducing the affinity for the operator. Protein-protein and protein-DNA interactions involving the hinge region play a crucial role in the allosteric changes occurring upon induction, but have not, as yet, been analyzed in atomic detail. RESULTS: We have used nuclear magnetic resonance (NMR) spectroscopy and restrained molecular dynamics (rMD) to determine the structure of the Lac repressor DNA-binding domain (headpeice 62; HP62) in complex with a symmetrized lac operator. Analysis of the structures reveals specific interactions between Lac repressor and DNA that were not found in previously investigated Lac repressor-DNA complexes. Important differences with the previously reported structures of the HP56-DNA complex were found in the loop following the helix-turn-helix (HTH) motif. The protein-protein and protein-DNA interactions involving the hinge region and the deformations in the DNA structure could be delineated in atomic detail. The structures were also used for comparison with the available crystallographic data on the Lac and Pur repressor-DNA complexes. CONCLUSIONS: The structures of the HP62-DNA complex provide the basis for a better understanding of the specific recognition in the Lac repressor-operator complex. In addition, the structural features of the hinge region provide detailed insight into the protein-protein and protein-DNA interactions responsible for the high affinity of the repressor for operator DNA.

Water molecules in DNA recognition I: hydration lifetimes of trp operator DNA in solution measured by NMR spectroscopy.

The present NMR study investigates the residence times of the hydration water molecules associated with uncomplexed trp operator DNA in solution by measuring intermolecular nuclear Overhauser effects (NOE) between water and DNA protons, and the nuclear magnetic relaxation dispersion (NMRD) of the water 2H and 17O resonances. Both methods indicate that the hydration water molecules exchange with bulk water on the sub-nanosecond time scale at 4 degreesC. No evidence was obtained for water molecules bound with longer residence times. In particular, the water molecules at the sites of interfacial hydration in the trp repressor/operator complex do not seem kinetically stabilized in the uncomplexed DNA. Analysis of the crystal structures of two different trp repressor/operator complexes shows very similar structural environments for the water molecules mediating specific contacts between the protein and the DNA, whereas much larger variations are observed for the location of corresponding water molecules detected in the crystal structure of an uncomplexed trp operator DNA duplex. Therefore, it appears unlikely that the hydration characteristics of the uncomplexed DNA target would be a major determinant of trp repressor/operator recognition.

Water molecules in DNA recognition II: a molecular dynamics view of the structure and hydration of the trp operator.

The structure and hydration of the DNA duplex d-(AGCGTACTAGTACGCT)2 corresponding to the trp operator fragment used in the crystal structure of the half site complex (PDB entry 1TRR) was studied by a 1.4 ns molecular dynamics simulation in water. The simulation, starting from a B-DNA conformation, used a non-bonded cutoff of 1.4 nm with a reaction field correction and resulted in a stable trajectory. The average DNA conformation obtained was closer to the ones found in the crystal structures of the complexes (PDB entries 1TRO and 1TRR) than to the crystal structure of unbound trp operator (Nucleic Acid Database entry BDJ061). The DNA hydration was characterized in terms of hydrogen bond percentages and corresponding residence times. The residence times of water molecules within 0.35 nm of the DNA non-exchangeable protons were calculated for comparison with NMR measurements of intermolecular water-DNA NOEs and nuclear magnetic relaxation dispersion measurements. No significant difference was found between major and minor groove hydration. The DNA donors and acceptors were hydrogen bonded to water molecules for 77(+/-19)% of the time on average. The average residence time of the hydrogen bonded water molecules was 11(+/-11) ps with a maximum of 223 ps. When all water molecules within NOE distance (0.35 nm) of non-exchangeable protons were considered, the average residence times increased to an average of 100(+/-4) ps and a maximum of 608 ps. These results agree with the experimental NMR results of Sunnerhagen et al. which did not show any evidence for water molecules bound with more than 1 ns residence time on the DNA surface. The exchange of hydration water from the DNA occurred in the major groove primarily through direct exchange with the bulk solvent, while access to and from the minor groove frequently proceeded via pathways involving ribose O3' and O4' and phosphate O2P oxygen atoms. The most common water diffusion pathways in the minor groove were perpendicular to the groove direction. In general, water molecules visited only a limited number of sites in the DNA grooves before exiting. The hydrogen bonding sites, where hydrogen bonds could be formed with donor and acceptor groups of the DNA, were filled with water molecules with an average B-factor value of 0.58 mn2. No special values were observed at any of the sites, where water molecules were observed both in the trp repressor/operator co-crystals and in the crystal structure of unbound DNA.