Normal mode analysis as a prerequisite for drug design: Application to matrix metalloproteinases inhibitors

Berberine ameliorates the progression of primary sclerosing cholangitis by activating farnesoid X receptor

Primary sclerosing cholangitis (PSC) is a rare cholestatic disease characterized by biliary infiltration, hepatic fibrosis and bile duct destruction. To date, treatment options for PSC are very limited. Therefore, the current study is aimed to investigate the therapeutic potential of berberine (BBR) against PSC. The disease was induced by feeding the mice with 3,5-diethoxycarbonyl-1,4-dihydro-2,4,6-collidine (DDC) for four weeks. The serum biochemistry and liver histology were analyzed. Furthermore, the expression of farnesoid X receptor (FXR) was also evaluated by real-time PCR. The results indicated that berberine prevents the progression of PSC by modulating the expression of FXR which ultimately regulates other genes (including Cyp7A1 and BSEP) thus maintaining bile acids homeostasis. Furthermore, the docking analysis showed that berberine interacts with the binding pocket of FXR to activate the protein thus acting as an FXR agonist. In conclusion, data indicate that berberine protects the liver from PSC-related injury. This effect might be due to the modulation of FXR activity.

Leveraging Fungal and Human Calcineurin-Inhibitor Structures, Biophysical Data, and Dynamics To Design Selective and Nonimmunosuppressive FK506 Analogs

Calcineurin is a critical enzyme in fungal pathogenesis and antifungal drug tolerance and, therefore, an attractive antifungal target. Current clinically accessible calcineurin inhibitors, such as FK506, are immunosuppressive to humans, so exploiting calcineurin inhibition as an antifungal strategy necessitates fungal specificity in order to avoid inhibiting the human pathway. Harnessing fungal calcineurin-inhibitor crystal structures, we recently developed a less immunosuppressive FK506 analog, APX879, with broad-spectrum antifungal activity and demonstrable efficacy in a murine model of invasive fungal infection. Our overarching goal is to better understand, at a molecular level, the interaction determinants of the human and fungal FK506-binding proteins (FKBP12) required for calcineurin inhibition in order to guide the design of fungus-selective, nonimmunosuppressive FK506 analogs. To this end, we characterized high-resolution structures of the Mucor circinelloides FKBP12 bound to FK506 and of the Aspergillus fumigatus, M. circinelloides, and human FKBP12 proteins bound to the FK506 analog APX879, which exhibits enhanced selectivity for fungal pathogens. Combining structural, genetic, and biophysical methodologies with molecular dynamics simulations, we identify critical variations in these structurally similar FKBP12-ligand complexes. The work presented here, aimed at the rational design of more effective calcineurin inhibitors, indeed suggests that modifications to the APX879 scaffold centered around the C₁₅, C₁₆, C₁₈, C₃₆, and C₃₇ positions provide the potential to significantly enhance fungal selectivity. IMPORTANCE Invasive fungal infections are a leading cause of death in the immunocompromised patient population. The rise in drug resistance to current antifungals highlights the urgent need to develop more efficacious and highly selective agents. Numerous investigations of major fungal pathogens have confirmed the critical role of the calcineurin pathway for fungal virulence, making it an attractive target for antifungal development. Although FK506 inhibits calcineurin, it is immunosuppressive in humans and cannot be used as an antifungal. By combining structural, genetic, biophysical, and in silico methodologies, we pinpoint regions of the FK506 scaffold and a less immunosuppressive analog, APX879, centered around the C₁₅ to C₁₈ and C₃₆ to C₃₇ positions that could be altered with selective extensions and/or deletions to enhance fungal selectivity. This work represents a significant advancement toward realizing calcineurin as a viable target for antifungal drug discovery.

Normal-mode driven exploration of protein domain motions

Domain motions involved in the function of proteins can often be well described as a combination of motions along a handfull of low-frequency modes, that is, with the values of a few normal coordinates. This means that, when the functional motion of a protein is unknown, it should prove possible to predict it, since it amounts to guess a few values. However, without the help of additional experimental data, using normal coordinates for generating accurate conformers far away from the initial one is not so straightforward. To do so, a new approach is proposed: instead of building conformers directly with the values of a subset of normal coordinates, they are built in two steps, the conformer built with normal coordinates being just used for defining a set of distance constraints, the final conformer being built so as to match them. Note that this approach amounts to transform the problem of generating accurate protein conformers using normal coordinates into a better known one: the distance-geometry problem, which is herein solved with the help of the ROSETTA software. In the present study, this approach allowed to rebuild accurately six large amplitude conformational changes, using at most six low-frequency normal coordinates. As a consequence of the low-dimensionality of the corresponding subspace, random exploration also proved enough for generating low-energy conformers close to the known end-point of the conformational change of the LAO binding protein, lysozyme T4 and adenylate kinase.

Protein vibrations and their localization behaviour. A numerical scaling analysis

Using a classical force field, we investigate the localization properties of protein normal modes. For a set of eighteen proteins that cover five classes of increasing size, we compute the participation ratio as a measure of the spatial extent of protein vibrations. In this scaling analysis, we find extended low-frequency far-infrared and Terahertz modes, in contrast to localized high-frequency near-infrared vibrations. These regimes are separated by a broad crossover around a wave number of 260 cm^-1. Biophysical and biochemical implications are discussed, and the vibrational localization properties are compared to those of amorphous solids.

Pharmmaker: Pharmacophore modeling and hit identification based on druggability simulations

Recent years have seen progress in druggability simulations, that is, molecular dynamics simulations of target proteins in solutions containing drug‐like probe molecules to characterize their drug‐binding abilities, if any. An important consecutive step is to analyze the trajectories to construct pharmacophore models (PMs) to use for virtual screening of libraries of small molecules. While considerable success has been observed in this type of computer‐aided drug discovery, a systematic tool encompassing multiple steps from druggability simulations to pharmacophore modeling, to identifying hits by virtual screening of libraries of compounds, has been lacking. We address this need here by developing a new tool, Pharmmaker, building on the DruGUI module of our ProDy application programming interface. Pharmmaker is composed of a suite of steps: (Step 1) identification of high affinity residues for each probe molecule type; (Step 2) selecting high affinity residues and hot spots in the vicinity of sites identified by DruGUI; (Step 3) ranking of the interactions between high affinity residues and specific probes; (Step 4) obtaining probe binding poses and corresponding protein conformations by collecting top‐ranked snapshots; and (Step 5) using those snapshots for constructing PMs. The PMs are then used as filters for identifying hits in structure‐based virtual screening. Pharmmaker, accessible online at http://prody.csb.pitt.edu/pharmmaker/, can be used in conjunction with other tools available in ProDy.

Druggability Simulations and X-Ray Crystallography Reveal a Ligand-Binding Site in the GluA3 AMPA Receptor N-Terminal Domain

Ionotropic glutamate receptors (iGluRs) mediate the majority of excitatory neurotransmission in the brain. Their dysfunction is implicated in many neurological disorders, rendering iGluRs potential drug targets. Here, we performed a systematic analysis of the druggability of two major iGluR subfamilies, using molecular dynamics simulations in the presence of drug-like molecules. We demonstrate the applicability of druggability simulations by faithfully identifying known agonist and modulator sites on AMPA receptors (AMPARs) and NMDA receptors. Simulations produced the expected allosteric changes of the AMPAR ligand-binding domain in response to agonist. We also identified a novel ligand-binding site specific to the GluA3 AMPAR N-terminal domain (NTD), resulting from its unique conformational flexibility that we explored further with crystal structures trapped in vastly different states. In addition to providing an in-depth analysis into iGluR NTD dynamics, our approach identifies druggable sites and permits the determination of pharmacophoric features toward novel iGluR modulators.

PLA2-like proteins myotoxic mechanism: a dynamic model description

Phospholipase A₂-like (PLA₂-like) proteins contribute to the development of muscle necrosis in Viperidae snake bites and are not efficiently neutralized by current antivenom treatments. The toxic mechanisms of PLA₂-like proteins are devoid of catalytic activity and not yet fully understood, although structural and functional experiments suggest a dimeric assembly and that the C-terminal residues are essential to myotoxicity. Herein, we characterized the functional mechanism of bothropic PLA₂-like structures related to global and local measurements using the available models in the Protein Data Bank and normal mode molecular dynamics (NM-MD). Those measurements include: (i) new geometric descriptions between their monomers, based on Euler angles; (ii) characterizations of canonical and non-canonical conformations of the C-terminal residues; (iii) accessibility of the hydrophobic channel; (iv) inspection of ligands; and (v) distance of clustered residues to toxin interface of interaction. Thus, we described the allosteric activation of PLA₂-like proteins and hypothesized that the natural movement between monomers, calculated from NM-MD, is related to their membrane disruption mechanism, which is important for future studies of the inhibition process. These methods and strategies can be applied to other proteins to help understand their mechanisms of action.

Jumping between protein conformers using normal modes

The relationship between the normal modes of a protein and its functional conformational change has been studied for decades. However, using this relationship in a predictive context remains a challenge. In this work, we demonstrate that, starting from a given protein conformer, it is possible to generate in a single step model conformers that are less than 1 Å (C_α -RMSD) from the conformer which is the known endpoint of the conformational change, particularly when the conformational change is collective in nature. Such accurate model conformers can be generated by following either the so-called robust or the 50 lowest-frequency modes obtained with various Elastic Network Models (ENMs). Interestingly, the quality of many of these models compares well with actual crystal structures, as assessed by the ROSETTA scoring function and PROCHECK. The most accurate and best quality conformers obtained in the present study were generated by using the 50 lowest-frequency modes of an all-atom ENM. However, with less than ten robust modes, which are identified without any prior knowledge of the nature of the conformational change, nearly 90% of the motion described by the 50 lowest-frequency modes of a protein can be captured. Such results strongly suggest that exploring the robust modes of ENMs may prove efficient for sampling the functionally relevant conformational repertoire of many proteins. © 2017 Wiley Periodicals, Inc.

Use of machine learning approaches for novel drug discovery

INTRODUCTION

The use of computational tools in the early stages of drug development has increased in recent decades. Machine learning (ML) approaches have been of special interest, since they can be applied in several steps of the drug discovery methodology, such as prediction of target structure, prediction of biological activity of new ligands through model construction, discovery or optimization of hits, and construction of models that predict the pharmacokinetic and toxicological (ADMET) profile of compounds.

AREAS COVERED

This article presents an overview on some applications of ML techniques in drug design. These techniques can be employed in ligand-based drug design (LBDD) and structure-based drug design (SBDD) studies, such as similarity searches, construction of classification and/or prediction models of biological activity, prediction of secondary structures and binding sites docking and virtual screening.

EXPERT OPINION

Successful cases have been reported in the literature, demonstrating the efficiency of ML techniques combined with traditional approaches to study medicinal chemistry problems. Some ML techniques used in drug design are: support vector machine, random forest, decision trees and artificial neural networks. Currently, an important application of ML techniques is related to the calculation of scoring functions used in docking and virtual screening assays from a consensus, combining traditional and ML techniques in order to improve the prediction of binding sites and docking solutions.

Exploring free energy landscapes of large conformational changes: molecular dynamics with excited normal modes

Proteins are found in solution as ensembles of conformations in dynamic equilibrium. Exploration of functional motions occurring on micro- to millisecond time scales by molecular dynamics (MD) simulations still remains computationally challenging. Alternatively, normal mode (NM) analysis is a well-suited method to characterize intrinsic slow collective motions, often associated with protein function, but the absence of anharmonic effects preclude a proper characterization of conformational distributions in a multidimensional NM space. Using both methods jointly appears to be an attractive approach that allows an extended sampling of the conformational space. In line with this view, the MDeNM (molecular dynamics with excited normal modes) method presented here consists of multiple-replica short MD simulations in which motions described by a given subset of low-frequency NMs are kinetically excited. This is achieved by adding additional atomic velocities along several randomly determined linear combinations of NM vectors, thus allowing an efficient coupling between slow and fast motions. The relatively high-energy conformations generated with MDeNM are further relaxed with standard MD simulations, enabling free energy landscapes to be determined. Two widely studied proteins were selected as examples: hen egg lysozyme and HIV-1 protease. In both cases, MDeNM provides a larger extent of sampling in a few nanoseconds, outperforming long standard MD simulations. A high degree of correlation with motions inferred from experimental sources (X-ray, EPR, and NMR) and with free energy estimations obtained by metadynamics was observed. Finally, the large sets of conformations obtained with MDeNM can be used to better characterize relevant dynamical populations, allowing for a better interpretation of experimental data such as SAXS curves and NMR spectra.

Binding of phenothiazines into allosteric hydrophobic pocket of human thioredoxin 1

Thioredoxins are multifunctional oxidoreductase proteins implicated in the antioxidant cellular apparatus and oxidative stress. They are involved in several pathologies and are promising anticancer targets. Identification of noncatalytic binding sites is of great interest for designing new allosteric inhibitors of thioredoxin. In a recent work, we predicted normal mode motions of human thioredoxin 1 and identified two major putative hydrophobic binding sites. In this work we investigated noncovalent interactions of human thioredoxin 1 with three phenotiazinic drugs acting as prooxidant compounds by using molecular docking and circular dichroism spectrometry to probe ligand binding into the previously predicted allosteric hydrophobic pockets. Our in silico and CD spectrometry experiments suggested one preferred allosteric binding site involving helix 3 and adopting the best druggable conformation identified by NMA. The CD spectra showed binding of thioridazine into thioredoxin 1 and suggested partial helix unfolding, which most probably concerns helix 3. Taken together, these data support the strategy to design thioredoxin inhibitors targeting a druggable allosteric binding site.

ENCoM server: exploring protein conformational space and the effect of mutations on protein function and stability

ENCoM is a coarse-grained normal mode analysis method recently introduced that unlike previous such methods is unique in that it accounts for the nature of amino acids. The inclusion of this layer of information was shown to improve conformational space sampling and apply for the first time a coarse-grained normal mode analysis method to predict the effect of single point mutations on protein dynamics and thermostability resulting from vibrational entropy changes. Here we present a web server that allows non-technical users to have access to ENCoM calculations to predict the effect of mutations on thermostability and dynamics as well as to generate geometrically realistic conformational ensembles. The server is accessible at: http://bcb.med.usherbrooke.ca/encom.

Recovery of the wild type atomic flexibility in the HIV-1 protease double mutants

The emergence of drug resistant mutations due to the selective pressure exerted by antiretrovirals, including protease inhibitors (PIs), remains a major problem in the treatment of AIDS. During PIs therapy, the occurrence of primary mutations in the wild type HIV-1 protease reduces both the affinity for the inhibitors and the viral replicative capacity compared to the wild type (WT) protein, but additional mutations compensate for this reduced viral fitness. To investigate this phenomenon from the structural point of view, we combined Molecular Dynamics and Normal Mode Analysis to analyze and compare the variations of the flexibility of C-alpha atoms and the differences in hydrogen bond (h-bond) network between the WT and double mutants. In most cases, the flexibility profile of the double mutants was more often similar to that of the WT than to that of the related single base mutants. All single mutants showed a significant alteration in h-bond formation compared to WT. Most of the significant changes occur in the border between the flap and cantilever regions. We found that all the considered double mutants have their h-bond pattern significantly altered in comparison to the respective single base mutants affecting their flexibility profile that becomes more similar to that of WT. This WT flexibility restoration in the double mutants appears as an important factor for the HIV-1 fitness recovery observed in patients.

On the relationship between low-frequency normal modes and the large-scale conformational changes of proteins

Normal mode analysis is a computational technique that allows to study the dynamics of biological macromolecules. It was first applied to small protein cases, more than thirty years ago. The interest in this technique then raised when it was realized that it can provide insights about the large-scale conformational changes a protein can experience, for instance upon ligand binding. As it was also realized that studying highly simplified protein models can provide similar insights, meaning that this kind of analysis can be both quick and simple to handle, several applications were proposed, in the context of various structural biology techniques. This review focuses on these applications, as well as on how the functional relevance of the lowest-frequency modes of proteins was established.

Importance of protein flexibility in ranking inhibitor affinities: modeling the binding mechanisms of piperidine carboxamides as Type I1/2 ALK inhibitors

Anaplastic lymphoma kinase (ALK) has gained increased attention as an attractive therapeutic target for the treatment of various cancers, especially non-small-cell lung cancer (NSCLC).

Conformational restriction of G-proteins Coupled Receptors (GPCRs) upon complexation to G-proteins: a putative activation mode of GPCRs?

GPCRs undergo large conformational changes during their activation. Starting from existing X-ray structures, we used Normal Modes Analyses to study the collective motions of the agonist-bound β2-adrenergic receptor both in its isolated "uncoupled" and G-protein "coupled" conformations. We interestingly observed that the receptor was able to adopt only one major motion in the protein:protein complex. This motion corresponded to an anti-symmetric rotation of both its extra- and intra-cellular parts, with a key role of previously identified highly conserved proline residues. Because this motion was also retrieved when performing NMA on 7 other GPCRs which structures were available, it is strongly suspected to possess a significant biological role, possibly being the "activation mode" of a GPCR when coupled to G-proteins.

Druggability Assessment of Allosteric Proteins by Dynamics Simulations in the Presence of Probe Molecules

Druggability assessment of a target protein has emerged in recent years as an important concept in hit-to-lead optimization. A reliable and physically relevant measure of druggability would allow informed decisions on the risk of investing in a particular target. Here, we define “druggability” as a quantitative estimate of binding sites and affinities for a potential drug acting on a specific protein target. In the present study, we describe a new methodology that successfully predicts the druggability and maximal binding affinity for a series of challenging targets, including those that function through allosteric mechanisms. Two distinguishing features of the methodology are (i) simulation of the binding dynamics of a diversity of probe molecules selected on the basis of an analysis of approved drugs and (ii) identification of druggable sites and estimation of corresponding binding affinities on the basis of an evaluation of the geometry and energetics of bound probe clusters. The use of the methodology for a variety of targets such as murine double mutant-2, protein tyrosine phosphatase 1B (PTP1B), lymphocyte function-associated antigen 1, vertebrate kinesin-5 (Eg5), and p38 mitogen-activated protein kinase provides examples for which the method correctly captures the location and binding affinities of known drugs. It also provides insights into novel druggable sites and the target’s structural changes that would accommodate, if not promote and stabilize, drug binding. Notably, the ability to identify high affinity spots even in challenging cases such as PTP1B or Eg5 shows promise as a rational tool for assessing the druggability of protein targets and identifying allosteric or novel sites for drug binding.

Hybrid approaches to molecular simulation

Molecular dynamics (MD) simulation is an established method for studying the conformational changes that are important for protein function. Recent advances in hardware and software have allowed MD simulations over the same timescales as experiment, improving the agreement between theory and experiment to a large extent. However, running such simulations are costly, in terms of resources, storage, and trajectory analysis. There is still a place for techniques that involve short MD simulations. In order to overcome the sampling paucity of short time-scales, hybrid methods that include some form of MD simulation can exploit certain features of the system of interest, often combining experimental information in surprising ways. Here, we review some recent hybrid approaches to the simulation of proteins.

Normal mode-based approaches in receptor ensemble docking

Explicitly accounting for target flexibility in docking still constitutes a difficult challenge due to the high dimensionality of the conformational space to be sampled. This especially applies to the high-throughput scenario, where the screening of hundreds of thousands compounds takes place. The use of multiple receptor conformations (MRCs) to perform ensemble docking in a sequential fashion is a simple but powerful approach that allows to incorporate binding site structural diversity in the docking process. Whenever enough experimental structures to build a diverse ensemble are not available, normal mode analysis provides an appealing and efficient approach to in silico generate MRCs by distortion along few low-frequency modes that represent collective mid- and large-scale displacements. In this way, the dimension of the conformational space to be sampled is heavily reduced. This methodology is especially suited to incorporate target flexibility at the backbone level. In this chapter, the main components of normal mode-based approaches in the context of ensemble docking are presented and explained, including the theoretical and practical considerations needed for the successful development and implementation of this methodology.

Pre-existing soft modes of motion uniquely defined by native contact topology facilitate ligand binding to proteins

Modeling protein flexibility constitutes a major challenge in accurate prediction of protein-ligand and protein-protein interactions in docking simulations. The lack of a reliable method for predicting the conformational changes relevant to substrate binding prevents the productive application of computational docking to proteins that undergo large structural rearrangements. Here, we examine how coarse-grained normal mode analysis has been advantageously applied to modeling protein flexibility associated with ligand binding. First, we highlight recent studies that have shown that there is a close agreement between the large-scale collective motions of proteins predicted by elastic network models and the structural changes experimentally observed upon ligand binding. Then, we discuss studies that have exploited the predicted soft modes in docking simulations. Two general strategies are noted: pregeneration of conformational ensembles that are then utilized as input for standard fixed-backbone docking and protein structure deformation along normal modes concurrent to docking. These studies show that the structural changes apparently "induced" upon ligand binding occur selectively along the soft modes accessible to the protein prior to ligand binding. They further suggest that proteins offer suitable means of accommodating/facilitating the recognition and binding of their ligand, presumably acquired by evolutionary selection of the suitable three-dimensional structure.

Free Energy Profiles along Consensus Normal Modes Provide Insight into HIV-1 Protease Flap Opening

Describing biological macromolecular energetics from computer simulations can pose major challenges, and often necessitates enhanced conformational sampling. We describe the calculation of conformational free-energy profiles along carefully chosen collective coordinates: "consensus" normal modes, developed recently as robust alternatives to conventional normal modes. In an application to the HIV-1 protease, we obtain efficient sampling of significant flap opening movements governing inhibitor binding from relatively short simulations, in close correspondence with experimental results.

Mutation D816V alters the internal structure and dynamics of c-KIT receptor cytoplasmic region: implications for dimerization and activation mechanisms

The type III receptor tyrosine kinase (RTK) KIT plays a crucial role in the transmission of cellular signals through phosphorylation events that are associated with a switching of the protein conformation between inactive and active states. D816V KIT mutation is associated with various pathologies including mastocytosis and cancers. D816V-mutated KIT is constitutively active, and resistant to treatment with the anti-cancer drug Imatinib. To elucidate the activating molecular mechanism of this mutation, we applied a multi-approach procedure combining molecular dynamics (MD) simulations, normal modes analysis (NMA) and binding site prediction. Multiple 50-ns MD simulations of wild-type KIT and its mutant D816V were recorded using the inactive auto-inhibited structure of the protein, characteristic of type III RTKs. Computed free energy differences enabled us to quantify the impact of D816V on protein stability in the inactive state. We evidenced a local structural alteration of the activation loop (A-loop) upon mutation, and a long-range structural re-organization of the juxta-membrane region (JMR) followed by a weakening of the interaction network with the kinase domain. A thorough normal mode analysis of several MD conformations led to a plausible molecular rationale to propose that JMR is able to depart its auto-inhibitory position more easily in the mutant than in wild-type KIT and is thus able to promote kinase mutant dimerization without the need for extra-cellular ligand binding. Pocket detection at the surface of NMA-displaced conformations finally revealed that detachment of JMR from the kinase domain in the mutant was sufficient to open an access to the catalytic and substrate binding sites.

Protein kinases are involved in a huge amount of cellular processes through phosphorylation, a crucial mechanism in cell signaling, and their misregulation often results in disease. The deactivation of protein tyrosine kinases (PTKs) or their oncogenic activation arises from mutations which affect the protein primary structure and the configuration of the enzymatic site apparently by stabilizing the activation loop (A-loop) extended conformation. Particularly, mutation D816V of receptor tyrosine kinase (RTK) KIT, found in patients with pediatric mastocytosis, acute leukemia or germ cell tumors, can be considered as the archetype of mutation inducing a displacement of the population equilibrium toward the active conformation. We present a comprehensive computational study of the activating mechanism(s) of this mutation. Our multi-approach in silico procedure evidenced a local alteration of the A-loop structure, and a long-range structural re-organization of the juxta-membrane region (JMR) followed by a weakening of the interaction network with the kinase domain. Our results provided a plausible conception of how the observed departure of JMR from kinase domain in the mutant promotes kinase mutant dimerization without requiring extra-cellular ligand binding. The pocket profiles we obtained suggested putative allosteric binding sites that could be targeted by ligands/modulators that trap the mutated enzyme.

A concerted mechanism for opening the GDP binding pocket and release of the nucleotide in hetero-trimeric G-proteins

G-protein hetero-trimers play a fundamental role in cell function. Their dynamic behavior at the atomic level remains to be understood. We have studied the Gi hetero-trimer through a combination of molecular dynamics simulations and normal mode analyses. We showed that these big proteins could undergo large-amplitude conformational changes, without any energy penalty and with an intrinsic dynamics centered on their GDP binding pocket. Among the computed collective motions, one of the modes (mode 17) was particularly able to significantly open both the base and the phosphate sides of the GDP binding pocket. This mode describing mainly a motion between the Ras-like and the helical domains of G(α) was in close agreement with some available X-ray data and with many other biochemical/biophysical observations including the kink of helix α5. The use of a new protocol, which allows extraction of the GDP ligand along the computed normal modes, supported that the exit of GDP was largely coupled to an opening motion along mode 17. We propose for the first time a "concerted mechanism" model in which the opening of the GDP pocket and the kink of the α5 helix occur concomitantly and favor GDP release from G(αβγ) complexes. This model is discussed in the context of the G-protein-coupled receptor/G-protein interaction close to the cell membrane.

L-arginine binding to human inducible nitric oxide synthase: an antisymmetric funnel route toward isoform-specific inhibitors?

Nitric oxide (NO) is an important signaling molecule produced by a family of enzymes called nitric oxide synthases (NOS). Because NO is involved in various pathological conditions, the development of potent and isoform-selective NOS inhibitors is an important challenge. In the present study, the dimer of oxygenase domain of human iNOS (iNOSoxy) complexed to its natural substrate L-arginine (L-Arg) and both heme and tetrahydro-L-biopterin (BH4) cofactors was studied through multiple molecular dynamics simulations. Starting from the X-ray structure available for that complex (PDB: 1NSI ), a 16 ns equilibration trajectory was first obtained. Twelve dynamics of slow extraction of L-Arg out from the iNOSoxy active site were then performed. The steered molecular dynamics (SMD) approach was used starting from three different points of the reference trajectory for a total simulation time of 35 ns. A probable unbinding/binding pathway of L-Arg was characterized. It was suggested that a driving force directed the substrate toward the heme pocket. Key intermediate steps/residues along the access route to the active site were identified along this "funnel shape" pathway and compared to existing data. A quasi-normal mode analysis performed on the SMD data suggested that large collective motions of the protein may be involved in L-Arg binding and that opening the route to the active site in one monomer promoted an inverse, closing motion in the second monomer. Finally, our findings might help to rationalize the design of human iNOS isoform competitive inhibitors.

Structural ensemble in computational drug screening

IMPORTANCE OF THE FIELD

Structure-based in silico drug screening is now widely used in drug development projects. Structure-based in silico drug screening is generally performed using a protein-compound docking program and docking scoring function. Many docking programs have been developed over the last 2 decades, but their prediction accuracy remains insufficient.

AREAS COVERED IN THIS REVIEW

This review highlights the recent progress of the post-processing of protein-compound complexes after docking.

WHAT THE READER WILL GAIN

These methods utilize ensembles of docking poses of compounds to improve the prediction accuracy for the ligand-docking pose and screening results. While the individual docking poses are not reliable, the free energy surface or the most probable docking pose can be estimated from the ensemble of docking poses.

TAKE HOME MESSAGE

The protein-compound docking program provides an arbitral rather than a canonical ensemble of docking poses. When the ensemble of docking poses satisfies the canonical ensemble, we can discuss how these post-docking analysis methods work and fail. Thus, improvements to the docking software will be needed in order to generate well-defined ensembles of docking poses.

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.

A flexible docking scheme to explore the binding selectivity of PDZ domains

Modeling of protein binding site flexibility in molecular docking is still a challenging problem due to the large conformational space that needs sampling. Here, we propose a flexible receptor docking scheme: A dihedral restrained replica exchange molecular dynamics (REMD), where we incorporate the normal modes obtained by the Elastic Network Model (ENM) as dihedral restraints to speed up the search towards correct binding site conformations. To our knowledge, this is the first approach that uses ENM modes to bias REMD simulations towards binding induced fluctuations in docking studies. In our docking scheme, we first obtain the deformed structures of the unbound protein as initial conformations by moving along the binding fluctuation mode, and perform REMD using the ENM modes as dihedral restraints. Then, we generate an ensemble of multiple receptor conformations (MRCs) by clustering the lowest replica trajectory. Using ROSETTALIGAND, we dock ligands to the clustered conformations to predict the binding pose and affinity. We apply this method to postsynaptic density-95/Dlg/ZO-1 (PDZ) domains; whose dynamics govern their binding specificity. Our approach produces the lowest energy bound complexes with an average ligand root mean square deviation of 0.36 A. We further test our method on (i) homologs and (ii) mutant structures of PDZ where mutations alter the binding selectivity. In both cases, our approach succeeds to predict the correct pose and the affinity of binding peptides. Overall, with this approach, we generate an ensemble of MRCs that leads to predict the binding poses and specificities of a protein complex accurately.

Activation of the ghrelin receptor is described by a privileged collective motion: a model for constitutive and agonist-induced activation of a sub-class A G-protein coupled receptor (GPCR)

Three homology models of the human ghrelin receptor (GHS-R1a) have been generated from the available X-ray structures of rhodopsin (RHO model), opsin (OPS model) and beta-2 adrenergic receptor (B2 model). The latter was used as a starting point for combined molecular dynamics simulation (MDS) and full atom normal modes analysis (NMA). A low-frequency normal mode (mode 16) perfectly reproduced the intracellular motions observed between B2 and RHO models; in the opposite direction along the same mode, the generated structures are closer to the OPS model, suggesting a direct link with GHS-R1a activation. This was in agreement with motions of the seven transmembranous segments, increase of the solvent accessibility of the 140-ERY-142 sequence, and flip of the Trp276 (C WLP) residue, some features related to GPCRs activation. According to our model, His280 was proposed to stabilize Trp276 in the active state; this was verified by site-directed mutagenesis and biochemical characterization of the resulting H280A and H280S mutants, which were fully functional but sharing an important decrease of their basal activities. Docking performed with short ghrelin derivatives Gly-Ser-Ser ([octa])-Phe-NH (2) and Gly-Ser-Ser ([octa])-Phe-Leu-NH (2) allowed the identification of a robust position of these peptides in the active site of the receptor. This model was refined by MDS and validated by docking experiments performed on a set of 55 ghrelin receptor ligands based on the 1,2,4- triazole scaffold. Finally, NMA performed on the obtained peptide-receptor complex suggested stabilization of the Trp276 residue and of the whole receptor in the active state, preventing the motion observed along mode 16 computed for the unbound receptor. Our results show that NMA offers a powerful approach to study the conformational diversity and the activation mechanism of GPCRs.

The intrinsic dynamics of enzymes plays a dominant role in determining the structural changes induced upon inhibitor binding

The conformational flexibility of target proteins continues to be a major challenge in accurate modeling of protein-inhibitor interactions. A fundamental issue, yet to be clarified, is whether the observed conformational changes are controlled by the protein or induced by the inhibitor. Although the concept of induced fit has been widely adopted for describing the structural changes that accompany ligand binding, there is growing evidence in support of the dominance of proteins' intrinsic dynamics which has been evolutionarily optimized to accommodate its functional interactions. The wealth of structural data for target proteins in the presence of different ligands now permits us to make a critical assessment of the balance between these two effects in selecting the bound forms. We focused on three widely studied drug targets, HIV-1 reverse transcriptase, p38 MAP kinase, and cyclin-dependent kinase 2. A total of 292 structures determined for these enzymes in the presence of different inhibitors and unbound form permitted us to perform an extensive comparative analysis of the conformational space accessed upon ligand binding, and its relation to the intrinsic dynamics before ligand binding as predicted by elastic network model analysis. Our results show that the ligand selects the conformer that best matches its structural and dynamic properties among the conformers intrinsically accessible to the protein in the unliganded form. The results suggest that simple but robust rules encoded in the protein structure play a dominant role in predefining the mechanisms of ligand binding, which may be advantageously exploited in designing inhibitors.

Consistent improvement of cross-docking results using binding site ensembles generated with elastic network normal modes

The representation of protein flexibility is still a challenge for the state-of-the-art flexible ligand docking protocols. In this article, we use a large and diverse benchmark to prove that is possible to improve consistently the cross-docking performance against a single receptor conformation, using an equilibrium ensemble of binding site conformers. The benchmark contained 28 proteins, and our method predicted the top-ranked near native ligand poses 20% more efficiently than using a single receptor. The multiple conformations were derived from the collective variable space defined by all heavy-atom elastic network normal modes, including backbone and side chains. We have found that the binding site displacements for best positioning of the ligand seem rather independent from the global collective motions of the protein. We also found that the number of binding site conformations needed to represent nonredundant flexibility was < 100. The ensemble of receptor conformations can be generated at our Web site at http://abagyan.scripps.edu/MRC.

Homology modeling and docking analysis of the interaction between polyphenols and mammalian 20S proteasomes

Molecular docking of small ligands to biologically active macromolecules has become a valuable strategy to predict the stability of complexes between potential partners and their binding modes. In this perspective, we applied this computational procedure to rationalize the reported role of polyphenols as inhibitors of the mammalian 20S proteasomes. In particular, polyphenols were shown to modulate each proteasomal activity at different extents both in the constitutive and the inducible enzyme. We performed a flexible molecular docking analysis between a set of polyphenols previously demonstrated to have the highest binding affinity and both the constitutive (from deposited PDB structures) and homology modeled active subunits of the IFN-gamma inducible proteasome, to provide insight into the possible mechanism of interaction. Among the tested polyphenols, (-)-epigallocatechin-3-gallate showed the highest affinity for the proteasome subunits, both in terms of intermolecular energy and predicted equilibrium constants, in particular for beta5 and beta5i subunits (E(total) = -66 kcal/mol, Ki = 81.3 microM and E(Total) = -83.9 kcal/mol, Ki = 0.29 microM, respectively), known to be related to the chymotrypsin-like and BrAAP activities. Collectively, polyphenols showed a higher affinity for the inducible subunits, in agreement with previous in vitro studies. Additionally, different contributions to the interaction energy (van der Waals, electrostatic, H-bond) of proteasome-polyphenols complexes were dissected.

Toward a molecular understanding of the interaction of dual specificity phosphatases with substrates: insights from structure-based modeling and high throughput screening

Dual-specificity phosphatases (DSPs) are important, but poorly understood, cell signaling enzymes that remove phosphate groups from tyrosine and serine/threonine residues on their substrate. Deregulation of DSPs has been implicated in cancer, obesity, diabetes, inflammation, and Alzheimer's disease. Due to their biological and biomedical significance, DSPs have increasingly become the subject of drug discovery high-throughput screening (HTS) and focused compound library development efforts. Progress in identifying selective and potent DSP inhibitors has, however, been restricted by the lack of sufficient structural data on inhibitor-bound DSPs. The shallow, almost flat, substrate binding sites in DSPs have been a major factor in hampering the rational design and the experimental development of active site inhibitors. Recent experimental and virtual HTS studies, as well as advances in molecular modeling, provide new insights into the potential mechanisms for substrate recognition and binding by this important class of enzymes. We present herein an overview of the progress, along with a brief description of applications to two types of DSPs: Cdc25 and MAP kinase phosphatase (MKP) family members. In particular, we focus on combined computational and experimental efforts for designing Cdc25B and MKP-1 inhibitors and understanding their mechanisms of interactions with their target proteins. These studies emphasize the utility of developing computational models and methods that meet the two major challenges currently faced in structure-based in silico design of lead compounds: the conformational flexibility of the target protein and the entropic contribution to the selection and stabilization of particular bound conformers.

Collective motions in glucosamine-6-phosphate synthase: influence of ligand binding and role in ammonia channelling and opening of the fructose-6-phosphate binding site

The large protein motions of the bacterial enzyme glucosamine-6-phosphate synthase have been addressed using full atom normal modes analysis for the empty, the glucose-6-phosphate and the glucose-6-phosphate+glutamate bound proteins. The approach that was used involving energy minimizations along the normal modes coordinates identified functional motions of the protein, some of which were characterized earlier by X-ray diffraction studies. This method made it possible for the first time to highlight significant energy differences according to whether none, only one or both of the active sites of the protein were occupied. Our data favoured a specific motion of the glutamine binding domain following the fixation of fructose-6-phosphate and suggested a rigidified structure with both sites occupied. Here, we show that most of the collective large amplitude motions of glucosamine-6-phosphate synthase that are modulated by ligand binding are crucial for the enzyme catalytic cycle, as they strongly modify the geometry of both the ammonia channel and the C-tail, demonstrating their role in ammonia transfer and ligand binding.

Docking to RNA via root-mean-square-deviation-driven energy minimization with flexible ligands and flexible targets

Structure-based drug design is now well-established for proteins as a key first step in the lengthy process of developing new drugs. In many ways, RNA may be a better target to treat disease than a protein because it is upstream in the translation pathway, so inhibiting a single mRNA molecule could prevent the production of thousands of protein gene products. Virtual screening is often the starting point for structure-based drug design. However, computational docking of a small molecule to RNA seems to be more challenging than that to protein due to the higher intrinsic flexibility and highly charged structure of RNA. Previous attempts at docking to RNA showed the need for a new approach. We present here a novel algorithm using molecular simulation techniques to account for both nucleic acid and ligand flexibility. In this approach, with both the ligand and the receptor permitted some flexibility, they can bind one another via an induced fit, as the flexible ligand probes the surface of the receptor. A possible ligand can explore a low-energy path at the surface of the receptor by carrying out energy minimization with root-mean-square-distance constraints. Our procedure was tested on 57 RNA complexes (33 crystal and 24 NMR structures); this is the largest data set to date to reproduce experimental RNA binding poses. With our procedure, the lowest-energy conformations reproduced the experimental binding poses within an atomic root-mean-square deviation of 2.5 A for 74% of tested complexes.

An NMA-guided path planning approach for computing large-amplitude conformational changes in proteins

This paper presents a new method for computing macromolecular motions based on the combination of path planning algorithms, originating from robotics research, and elastic network normal mode analysis. The low-frequency normal modes are regarded as the collective degrees of freedom of the molecule. Geometric path planning algorithms are used to explore these collective degrees of freedom in order to find possible large-amplitude conformational changes. To overcome the limits of the harmonic approximation, which is valid in the vicinity of the minimum energy structure, and to get larger conformational changes, normal mode calculations are iterated during the exploration. Initial results show the efficiency of our method, which requires a small number of normal mode calculations to compute large-amplitude conformational transitions in proteins. A detailed analysis is presented for the computed transition between the open and closed structures of adenylate kinase. This transition, important for its biological function, involves large-amplitude domain motions. The obtained motion correlates well with results presented in related works.