Pre- and post-docking sampling of conformational changes using ClustENM and HADDOCK for protein-protein and protein-DNA systems.

Incorporating the dynamic nature of biomolecules in the modeling of their complexes is a challenge, especially when the extent and direction of the conformational changes taking place upon binding is unknown. Estimating whether the binding of a biomolecule to its partner(s) occurs in a conformational state accessible to its unbound form ("conformational selection") and/or the binding process induces conformational changes ("induced-fit") is another challenge. We propose here a method combining conformational sampling using ClustENM-an elastic network-based modeling procedure-with docking using HADDOCK, in a framework that incorporates conformational selection and induced-fit effects upon binding. The extent of the applied deformation is estimated from its energetical costs, inspired from mechanical tensile testing on materials. We applied our pre- and post-docking sampling of conformational changes to the flexible multidomain protein-protein docking benchmark and a subset of the protein-DNA docking benchmark. Our ClustENM-HADDOCK approach produced acceptable to medium quality models in 7/11 and 5/6 cases for the protein-protein and protein-DNA complexes, respectively. The conformational selection (sampling prior to docking) has the highest impact on the quality of the docked models for the protein-protein complexes. The induced-fit stage of the pipeline (post-sampling), however, improved the quality of the final models for the protein-DNA complexes. Compared to previously described strategies to handle conformational changes, ClustENM-HADDOCK performs better than two-body docking in protein-protein cases but worse than a flexible multidomain docking approach. However, it does show a better or similar performance compared to previous protein-DNA docking approaches, which makes it a suitable alternative.

An overview of data-driven HADDOCK strategies in CAPRI rounds 38-45.

Our information-driven docking approach HADDOCK has demonstrated a sustained performance since the start of its participation to CAPRI. This is due, in part, to its ability to integrate data into the modeling process, and to the robustness of its scoring function. We participated in CAPRI both as server and manual predictors. In CAPRI rounds 38-45, we have used various strategies depending on the available information. These ranged from imposing restraints to a few residues identified from literature as being important for the interaction, to binding pockets identified from homologous complexes or template-based refinement/CA-CA restraint-guided docking from identified templates. When relevant, symmetry restraints were used to limit the conformational sampling. We also tested for a large decamer target a new implementation of the MARTINI coarse-grained force field in HADDOCK. Overall, we obtained acceptable or better predictions for 13 and 11 server and manual submissions, respectively, out of the 22 interfaces. Our server performance (acceptable or higher-quality models when considering the top 10) was better (59%) than the manual (50%) one, in which we typically experiment with various combinations of protocols and data sources. Again, our simple scoring function based on a linear combination of intermolecular van der Waals and electrostatic energies and an empirical desolvation term demonstrated a good performance in the scoring experiment with a 63% success rate across all 22 interfaces. An analysis of model quality indicates that, while we are consistently performing well in generating acceptable models, there is room for improvement for generating/identifying higher quality models.

Understanding the Binding Specificity of G-Protein Coupled Receptors toward G-Proteins and Arrestins: Application to the Dopamine Receptor Family.

G-Protein coupled receptors (GPCRs) are involved in a myriad of pathways key for human physiology through the formation of complexes with intracellular partners such as G-proteins and arrestins (Arrs). However, the structural and dynamical determinants of these complexes are still largely unknown. Herein, we developed a computational big-data pipeline that enables the structural characterization of GPCR complexes with no available structure. This pipeline was used to study a well-known group of catecholamine receptors, the human dopamine receptor (DXR) family and its complexes, producing novel insights into the physiological properties of these important drug targets. A detailed description of the protein interfaces of all members of the DXR family (D1R, D2R, D3R, D4R, and D5R) and the corresponding protein interfaces of their binding partners (Arrs: Arr2 and Arr3; G-proteins: Gi1, Gi2, Gi3, Go, Gob, Gq, Gslo, Gssh, Gt2, and Gz) was generated. To produce reliable structures of the DXR family in complex with either G-proteins or Arrs, we performed homology modeling using as templates the structures of the ß2-adrenergic receptor (ß2AR) bound to Gs, the rhodopsin bound to Gi, and the recently acquired neurotensin receptor-1 (NTSR1) and muscarinic 2 receptor (M2R) bound to arrestin (Arr). Among others, the work demonstrated that the three partner groups, Arrs and Gs- and Gi-proteins, are all structurally and dynamically distinct. Additionally, it was revealed the involvement of different structural motifs in G-protein selective coupling between D1- and D2-like receptors. Having constructed and analyzed 50 models involving DXR, this work represents an unprecedented large-scale analysis of GPCR-intracellular partner interface determinants. All data is available at www.moreiralab.com/resources/dxr.

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.

How an Inhibitor Bound to Subunit Interface Alters Triosephosphate Isomerase Dynamics.

The tunnel region at triosephosphate isomerase (TIM)'s dimer interface, distant from its catalytic site, is a target site for certain benzothiazole derivatives that inhibit TIM's catalytic activity in Trypanosoma cruzi, the parasite that causes Chagas disease. We performed multiple 100-ns molecular-dynamics (MD) simulations and elastic network modeling (ENM) on both apo and complex structures to shed light on the still unclear inhibitory mechanism of one such inhibitor, named bt10. Within the time frame of our MD simulations, we observed stabilization of aromatic clusters at the dimer interface and enhancement of intersubunit hydrogen bonds in the presence of bt10, which point to an allosteric effect rather than destabilization of the dimeric structure. The collective dynamics dictated by the topology of TIM is known to facilitate the closure of its catalytic loop over the active site that is critical for substrate entrance and product release. We incorporated the ligand's effect on vibrational dynamics by applying mixed coarse-grained ENM to each one of 54,000 MD snapshots. Using this computationally efficient technique, we observed altered collective modes and positive shifts in eigenvalues due to the constraining effect of bt10 binding. Accordingly, we observed allosteric changes in the catalytic loop's dynamics, flexibility, and correlations, as well as the solvent exposure of catalytic residues. A newly (to our knowledge) introduced technique that performs residue-based ENM scanning of TIM revealed the tunnel region as a key binding site that can alter global dynamics of the enzyme.

Coupling between catalytic loop motions and enzyme global dynamics.

Catalytic loop motions facilitate substrate recognition and binding in many enzymes. While these motions appear to be highly flexible, their functional significance suggests that structure-encoded preferences may play a role in selecting particular mechanisms of motions. We performed an extensive study on a set of enzymes to assess whether the collective/global dynamics, as predicted by elastic network models (ENMs), facilitates or even defines the local motions undergone by functional loops. Our dataset includes a total of 117 crystal structures for ten enzymes of different sizes and oligomerization states. Each enzyme contains a specific functional/catalytic loop (10-21 residues long) that closes over the active site during catalysis. Principal component analysis (PCA) of the available crystal structures (including apo and ligand-bound forms) for each enzyme revealed the dominant conformational changes taking place in these loops upon substrate binding. These experimentally observed loop reconfigurations are shown to be predominantly driven by energetically favored modes of motion intrinsically accessible to the enzyme in the absence of its substrate. The analysis suggests that robust global modes cooperatively defined by the overall enzyme architecture also entail local components that assist in suitable opening/closure of the catalytic loop over the active site.

Collective dynamics of the ribosomal tunnel revealed by elastic network modeling.

The collective dynamics of the nascent polypeptide exit tunnel are investigated with the computationally efficient elastic network model using normal mode analysis. The calculated normal modes are considered individually and in linear combinations with different coefficients mimicking the phase angles between modes, in order to follow the mechanistic motions of tunnel wall residues. The low frequency fluctuations indicate three distinct regions along the tunnel-the entrance, the neck, and the exit-each having distinctly different domain motions. Generally, the lining of the entrance region moves in the exit direction, with the exit region having significantly larger motions, but in a perpendicular direction, whereas the confined neck region has rotational motions. Especially the universally conserved extensions of ribosomal proteins L4 and L22 located at the narrowest and mechanistically strategic region of tunnel undergo generally anti- or non-correlated motions, which may have an important role in nascent polypeptide gating mechanism. These motions appear to be sufficiently robust so as to be unaffected by the presence of a peptide chain in the tunnel.

Conformational dynamics of bacterial trigger factor in apo and ribosome-bound states.

The chaperone trigger factor (TF) binds to the ribosome exit tunnel and helps cotranslational folding of nascent chains (NC) in bacterial cells and chloroplasts. In this study, we aim to investigate the functional dynamics of fully-atomistic apo TF and its complex with 50S. As TF accomodates a high percentage of charged residues on its surface, the effect of ionic strength on TF dynamics is assessed here by performing five independent molecular dynamics (MD) simulations (total of 1.3 micro-second duration) at 29 mM and 150 mM ionic strengths. At both concentrations, TF exhibits high inter- and intra-domain flexibility related to its binding (BD), core (CD), and head (HD) domains. Even though large oscillations in gyration radius exist during each run, we do not detect the so-called 'fully collapsed' state with both HD and BD collapsed upon the core. In fact, the extended conformers with relatively open HD and BD are highly populated at the physiological concentration of 150 mM. More importantly, extended TF snapshots stand out in terms of favorable docking onto the 50S subunit. Elastic network modeling (ENM) indicates significant changes in TF's functional dynamics and domain decomposition depending on its conformation and positioning on the 50S. The most dominant slow motions are the lateral sweeping and vertical opening/closing of HD relative to 50S. Finally, our ENM-based efficient technique -ClustENM- is used to sample atomistic conformers starting with an extended TF-50S complex. Specifically, BD and CD motions are restricted near the tunnel exit, while HD is highly mobile. The atomistic conformers generated without an NC are in agreement with the cryo-EM maps available for TF-ribosome-NC complex.

Correction: Ligand Docking to Intermediate and Close-To-Bound Conformers Generated by an Elastic Network Model Based Algorithm for Highly Flexible Proteins.

[This corrects the article DOI: 10.1371/journal.pone.0158063.].

Ligand Docking to Intermediate and Close-To-Bound Conformers Generated by an Elastic Network Model Based Algorithm for Highly Flexible Proteins.

Incorporating receptor flexibility in small ligand-protein docking still poses a challenge for proteins undergoing large conformational changes. In the absence of bound structures, sampling conformers that are accessible by apo state may facilitate docking and drug design studies. For this aim, we developed an unbiased conformational search algorithm, by integrating global modes from elastic network model, clustering and energy minimization with implicit solvation. Our dataset consists of five diverse proteins with apo to complex RMSDs 4.7-15 Å. Applying this iterative algorithm on apo structures, conformers close to the bound-state (RMSD 1.4-3.8 Å), as well as the intermediate states were generated. Dockings to a sequence of conformers consisting of a closed structure and its "parents" up to the apo were performed to compare binding poses on different states of the receptor. For two periplasmic binding proteins and biotin carboxylase that exhibit hinge-type closure of two dynamics domains, the best pose was obtained for the conformer closest to the bound structure (ligand RMSDs 1.5-2 Å). In contrast, the best pose for adenylate kinase corresponded to an intermediate state with partially closed LID domain and open NMP domain, in line with recent studies (ligand RMSD 2.9 Å). The docking of a helical peptide to calmodulin was the most challenging case due to the complexity of its 15 Å transition, for which a two-stage procedure was necessary. The technique was first applied on the extended calmodulin to generate intermediate conformers; then peptide docking and a second generation stage on the complex were performed, which in turn yielded a final peptide RMSD of 2.9 Å. Our algorithm is effective in producing conformational states based on the apo state. This study underlines the importance of such intermediate states for ligand docking to proteins undergoing large transitions.

ClustENM: ENM-Based Sampling of Essential Conformational Space at Full Atomic Resolution.

Accurate sampling of conformational space and, in particular, the transitions between functional substates has been a challenge in molecular dynamic (MD) simulations of large biomolecular systems. We developed an Elastic Network Model (ENM)-based computational method, ClustENM, for sampling large conformational changes of biomolecules with various sizes and oligomerization states. ClustENM is an iterative method that combines ENM with energy minimization and clustering steps. It is an unbiased technique, which requires only an initial structure as input, and no information about the target conformation. To test the performance of ClustENM, we applied it to six biomolecular systems: adenylate kinase (AK), calmodulin, p38 MAP kinase, HIV-1 reverse transcriptase (RT), triosephosphate isomerase (TIM), and the 70S ribosomal complex. The generated ensembles of conformers determined at atomic resolution show good agreement with experimental data (979 structures resolved by X-ray and/or NMR) and encompass the subspaces covered in independent MD simulations for TIM, p38, and RT. ClustENM emerges as a computationally efficient tool for characterizing the conformational space of large systems at atomic detail, in addition to generating a representative ensemble of conformers that can be advantageously used in simulating substrate/ligand-binding events.

Blind Dockings of Benzothiazoles to Multiple Receptor Conformations of Triosephosphate Isomerase from Trypanosoma cruzi and Human.

We aim to uncover the binding modes of benzothiazoles, which have been reported as specific inhibitors of triosephosphate isomerase from the parasite Trypanosoma cruzi (TcTIM), by performing blind dockings on both TcTIM and human TIM (hTIM). Detailed analysis of binding sites and specific interactions are carried out based on ensemble dockings to multiple receptor conformers obtained from molecular dynamics simulations. In TcTIM dimer dockings, the inhibitors preferentially bind to the tunnel-shaped cavity formed at the interface of the subunits, whereas non-inhibitors mostly choose other sites. In contrast, TcTIM monomer binding interface and hTIM dimer interface do not present a specific binding site for the inhibitors. These findings point to the importance of the tunnel and of the dimeric form for inhibition of TcTIM. Specific interactions of the inhibitors and their sulfonate-free derivatives with the receptor residues indicate the significance of sulfonate group for binding affinity and positioning on the TcTIM dimer interface. One of the inhibitors also binds to the active site, which may explain its relatively higher inhibition effect on hTIM.