Optimal identification of semi-rigid domains in macromolecules from molecular dynamics simulation

Fast, Accurate, and System-Specific Variable-Resolution Modeling of Proteins

In recent years, a few multiple-resolution modeling strategies have been proposed, in which functionally relevant parts of a biomolecule are described with atomistic resolution, with the remainder of the system being concurrently treated using a coarse-grained model. In most cases, the parametrization of the latter requires lengthy reference all-atom simulations and/or the usage of off-shelf coarse-grained force fields, whose interactions have to be refined to fit the specific system under examination. Here, we overcome these limitations through a novel multiresolution modeling scheme for proteins, dubbed coarse-grained anisotropic network model for variable resolution simulations, or CANVAS. This scheme enables a user-defined modulation of the resolution level throughout the system structure; a fast parametrization of the potential without the necessity of reference simulations; and the straightforward usage of the model on the most commonly used molecular dynamics platforms. The method is presented and validated with two case studies, the enzyme adenylate kinase and the therapeutic antibody pembrolizumab, by comparing the results obtained with the CANVAS model against fully atomistic simulations. The modeling software, implemented in Python, is made freely available for the community on a collaborative github repository.

Molecular dynamics identifies semi-rigid domains in the PD-1 checkpoint receptor bound to its natural ligand PD-L1

Cells in danger of being erroneously attacked by leucocytes express PD-L1 on their surface. These cells activate PD-1 on attacking leucocytes and send them to death, thus curbing erroneous, autoimmune attack. Unfortunately, cancer cells exploit this mechanism: By expressing PD-L1, they guard themselves against leucocyte attack and thereby evade immune clearance. Checkpoint inhibitors are drugs which re-enable immune clearance of cancer cells by blocking the binding of PD-L1 to PD-1 receptors. It is therefore of utmost interest to investigate these binding mechanisms. We use three 600 ns all-atom molecular dynamics simulations to scrutinize molecular motions of PD-1 with its binding partner, the natural ligand PD-L1. Usually, atomic motion patterns are evaluated against whole molecules as a reference, disregarding that such a reference is a dynamic entity by itself, thus degrading stability of the reference. As a remedy, we identify semi-rigid domains, lending themselves as more stable and reliable reference frames against which even minute differences in molecular motion can be quantified precisely. We propose an unsupervised three-step procedure. In previous work of our group and others, minute differences in motion patterns proved decisive for differences in function. Here, several highly reliable frames of reference are established for future investigations based on molecular motion.

A study on allosteric communication in U1A-snRNA binding interactions: network analysis combined with molecular dynamics data

The allosteric regulation during the binding interactions between small nuclear RNAs (snRNAs) and the associated protein factors is critical to the function of spliceosomes in alternative RNA splicing. Although network models combined with molecular dynamics simulations have shown to be powerful tools for the analysis of protein allostery, the atomic-level simulations are, however, too expensive and with limited accuracy for the large-size systems. In this work, we use a residual network model combined with a coarse-grained Gaussian network model (GNM) to investigate the binding interactions between the snRNA and the human U1A protein which is a major component of the spliceosomal U1 small nuclear ribonucleoprotein particle, and to identify the residues that play an important role in the allosteric communication in U1A during this process. We also utilize the Girvan-Newman method to detect the structural organization in U1A-snRNA recognition and interactions. Our results reveal that: (Ι) not only the residues at the binding sites that are traditionally considered to play a major role in U1A-snRNA association, but those residues that are far away from the RNA binding interface participate in the U1A's allosteric signal transmission induced by the RNA binding; (Π) the structure of U1A protein is well organized with different communities acting different roles for its RNA binding and allosteric regulation. The study demonstrates that the combination of the residual network and elastic network models is an effective and efficient method which can be readily extended to the investigation of the allosteric communication for other macromolecular interaction systems.

Coupling of Conformational Switches in Calcium Sensor Unraveled with Local Markov Models and Transfer Entropy

Proteins often have multiple switching domains that are coupled to each other and to the binding of ligands in order to realize signaling functions. Here we investigate the C2A domain of Synaptotagmin-1 (Syt-1), a calcium sensor in the neurotransmitter release machinery and a model system for the large family of C2 membrane binding domains. We combine extensive molecular dynamics (MD) simulations with Markov modeling in order to model conformational switching domains, their states, and their dependence on bound calcium ions. Then, we use transfer entropy to characterize how the switching domains are coupled via directed or allosteric mechanisms and give rise to the calcium sensing function of the protein. Our proposed switching mechanism contributes to the understanding of the neurotransmitter release machinery. Furthermore, the methodological approach we develop serves as a template to analyze conformational switching domains and the broad study of their coupling in macromolecular machines.

Discovery of 6-Arylurea-2-arylbenzoxazole and 6-Arylurea-2-arylbenzimidazole Derivatives as Angiogenesis Inhibitors: Design, Synthesis and in vitro Biological Evaluation

We embarked on a structural optimization campaign aimed at the discovery of novel anti-angiogenesis agents with previously reported imidazole kinase inhibitors as a lead compound. A library of 29 compounds was synthesized. Several title compounds exhibited selective inhibitory activities against vascular endothelial growth factor receptor 2 (VEGFR-2) over epidermal growth factor receptor (EGFR) kinase; these compounds also displayed selective and potent antiproliferative activity against three cancer cell lines. The newly synthesized compounds were evaluated for anti-angiogenesis activity by chick chorioallantoic membrane (CAM) assay. Among them, 1-(2-(2-chlorophenyl)benzo[d]oxazol-5-yl)-3-(4-(trifluoromethoxy)phenyl)urea (compound 5 n) showed the most potent anti-angiogenesis capacity, efficient cytotoxic activities (in vitro against human umbilical vein endothelial cells (HUVEC), H1975, A549, and HeLa cell lines, with respective IC₅₀ values of 8.46, 1.40, 7.61, and 0.28 μm), and an acceptable level of VEGFR-2 kinase inhibition (IC₅₀ =0.25 μm). Molecular docking analysis revealed 5 n to be a type II inhibitor of VEGFR-2 kinase. In general, these results indicate that these 6-arylurea-2-arylbenzoxazole/benzimidazole derivatives are promising inhibitors of VEGFR-2 kinase for potential development into anti-angiogenesis drugs.

Spatiotemporal multistage consensus clustering in molecular dynamics studies of large proteins

The aim of this work is to find semi-rigid domains within large proteins as reference structures for fitting molecular dynamics trajectories. We propose an algorithm, multistage consensus clustering, MCC, based on minimum variation of distances between pairs of Cα-atoms as target function. The whole dataset (trajectory) is split into sub-segments. For a given sub-segment, spatial clustering is repeatedly started from different random seeds, and we adopt the specific spatial clustering with minimum target function: the process described so far is stage 1 of MCC. Then, in stage 2, the results of spatial clustering are consolidated, to arrive at domains stable over the whole dataset. We found that MCC is robust regarding the choice of parameters and yields relevant information on functional domains of the major histocompatibility complex (MHC) studied in this paper: the α-helices and β-floor of the protein (MHC) proved to be most flexible and did not contribute to clusters of significant size. Three alleles of the MHC, each in complex with ABCD3 peptide and LC13 T-cell receptor (TCR), yielded different patterns of motion. Those alleles causing immunological allo-reactions showed distinct correlations of motion between parts of the peptide, the binding cleft and the complementary determining regions (CDR)-loops of the TCR. Multistage consensus clustering reflected functional differences between MHC alleles and yields a methodological basis to increase sensitivity of functional analyses of bio-molecules. Due to the generality of approach, MCC is prone to lend itself as a potent tool also for the analysis of other kinds of big data.

Quantitative Characterization of Domain Motions in Molecular Machines

The work of molecular machines such as the ribosome is accompanied by conformational changes, often characterized by relative motions of their domains. The method we have developed seeks to quantify these motions in a general way, facilitating comparisons of results obtained by different researchers. Typically there are multiple snapshots of a structure in the form of pdb coordinates resulting from flexible fitting of low-resolution density maps, from X-ray crystallography, or from molecular dynamics simulation trajectories. Our objective is to characterize the motion of each domain as a coordinate transformation using moments of inertia tensor, a method we developed earlier. What has been missing until now are ancillary tools that make this task practical, general, and biologically informative. We have provided a comprehensive solution to this task with a set of tools implemented on the VMD platform. These tools address the need for reproducible segmentation of domains, and provide a generalized description of their motions using principal axes of inertia. Although this methodology has been specifically developed for studying ribosome motion, it is applicable to any molecular machine.

Unifying view of mechanical and functional hotspots across class A GPCRs

G protein-coupled receptors (GPCRs) are the largest superfamily of signaling proteins. Their activation process is accompanied by conformational changes that have not yet been fully uncovered. Here, we carry out a novel comparative analysis of internal structural fluctuations across a variety of receptors from class A GPCRs, which currently has the richest structural coverage. We infer the local mechanical couplings underpinning the receptors’ functional dynamics and finally identify those amino acids whose virtual deletion causes a significant softening of the mechanical network. The relevance of these amino acids is demonstrated by their overlap with those known to be crucial for GPCR function, based on static structural criteria. The differences with the latter set allow us to identify those sites whose functional role is more clearly detected by considering dynamical and mechanical properties. Of these sites with a genuine mechanical/dynamical character, the top ranking is amino acid 7x52, a previously unexplored, and experimentally verifiable key site for GPCR conformational response to ligand binding.

The biological functionality of several receptors and enzymes depends on their capability to sustain large-scale structural fluctuations and adopt different conformational states in response to ligand binding. This is the case for G protein-coupled receptors (GPCRs), the largest superfamily of signaling proteins in mammals and a primary pharmaceutical target. To better understand the functional dynamics of GPCRs, we have analysed the inter-residue distance variations across the available structures for several receptors of the rhodopsin-like family (class A). We first reconstructed the network of mechanical, rigid-like couplings between nearby amino acids and then identified those acting as dynamical/mechanical hubs. These were the sites whose virtual removal led to a significant softening of the overall mechanical network. After validating the biological relevance of these sites by comparison against known key functional sites, we singled out those regions which emerge as prominent mechanical hubs and yet have an otherwise still unknown functional role. The most relevant of such novel putative functional sites, which could be probed by mutagenesis experiments, is at interface of two transmembrane helices and we expect it to be crucial for assisting GPCRs conformational response to agonist binding.

Effect of Inactivating Mutations on Peptide Conformational Ensembles: The Plant Polypeptide Hormone Systemin

As part of their basal immune mechanism against insect/herbivore attacks, plants have evolved systemic response mechanisms. Such a systemic wound response in tomato was found to involve an 18 amino acid polypeptide called systemin, the first polypeptide hormone to be discovered in plants. Systematic alanine scanning and deletion studies showed differential modulation in its activity, particularly a major loss of function due to alanine substitution at positions 13 and 17 and less extentive loss of function due to substitution at position 12. We have studied the conformational ensembles of wild-type systemin along with its 17 variants by carrying out a total of 5.76 μs of replica-exchange molecular dynamics simulation in an implicit solvent environment. In our simulations, wild-type systemin showed a lack of α-helical and β-sheet structures, in conformity with earlier circular dichroism and NMR data. On the other hand, two regions containing diproline segments showed a tendency to adopt polyproline II structures. Examination of conformational ensembles of the 17 variants revealed a change in the population distributions, suggesting a less flexible structure for alanine substitutions at positions 12 and 13 but not for position 17. Combined with the experimental observations that positions 1-14 of systemin are important for the formation of the peptide-receptor complex, this leads to the hypothesis that loss of conformational flexibility may play a role in the loss of activity of systemin due to the P12A and P13A substitutions, while T17A deactivation probably occurs for a different reason, most likely the loss of the threonine phosphorylation site. We also indicate possible structural reasons why the substitution of the prolines at positions 12 and 13 leads to a loss of conformational freedom in the peptide.

ResiCon: a method for the identification of dynamic domains, hinges and interfacial regions in proteins

MOTIVATION

Structure of most proteins is flexible. Identification and analysis of intramolecular motions is a complex problem. Breaking a structure into relatively rigid parts, the so-called dynamic domains, may help comprehend the complexity of protein's mobility. We propose a new approach called ResiCon (Residue Contacts analysis), which performs this task by applying a data-mining analysis of an ensemble of protein configurations and recognizes dynamic domains, hinges and interfacial regions, by considering contacts between residues.

RESULTS

Dynamic domains found by ResiCon are more compact than those identified by two other popular methods: PiSQRD and GeoStaS. The current analysis was carried out using a known reference set of 30 NMR protein structures, as well as molecular dynamics simulation data of flap opening events in HIV-1 protease. The more detailed analysis of HIV-1 protease dataset shows that ResiCon identified dynamic domains involved in structural changes of functional importance.

AVAILABILITY AND IMPLEMENTATION

The ResiCon server is available at URL: http://dworkowa.imdik.pan.pl/EP/ResiCon.

CONTACT

pawel@bioexploratorium.pl

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Relative Movements of Domains in Large Molecules of the Immune System

Molecular dynamics was used to simulate large molecules of the immune system (major histocompatibility complex class I, presented epitope, T-cell receptor, and a CD8 coreceptor.) To characterize the relative orientation and movements of domains local coordinate systems (based on principal component analysis) were generated and directional cosines and Euler angles were computed. As a most interesting result, we found that the presence of the coreceptor seems to influence the dynamics within the protein complex, in particular the relative movements of the two α-helices, Gα₁ and Gα₂.

SPECTRUS: A Dimensionality Reduction Approach for Identifying Dynamical Domains in Protein Complexes from Limited Structural Datasets

Identifying dynamical, quasi-rigid domains in proteins provides a powerful means for characterizing functionally oriented structural changes via a parsimonious set of degrees of freedom. In fact, the relative displacements of few dynamical domains usually suffice to rationalize the mechanics underpinning biological functionality in proteins and can even be exploited for structure determination or refinement purposes. Here we present SPECTRUS, a general scheme that, by solely using amino acid distance fluctuations, can pinpoint the innate quasi-rigid domains of single proteins or large complexes in a robust way. Consistent domains are usually obtained by using either a pair of representative structures or thousands of conformers. The functional insights offered by the approach are illustrated for biomolecular systems of very different size and complexity such as kinases, ion channels, and viral capsids. The decomposition tool is available as a software package and web server at spectrus.sissa.it.

Motion Tree Delineates Hierarchical Structure of Protein Dynamics Observed in Molecular Dynamics Simulation

Molecular dynamics (MD) simulations of proteins provide important information to understand their functional mechanisms, which are, however, likely to be hidden behind their complicated motions with a wide range of spatial and temporal scales. A straightforward and intuitive analysis of protein dynamics observed in MD simulation trajectories is therefore of growing significance with the large increase in both the simulation time and system size. In this study, we propose a novel description of protein motions based on the hierarchical clustering of fluctuations in the inter-atomic distances calculated from an MD trajectory, which constructs a single tree diagram, named a “Motion Tree”, to determine a set of rigid-domain pairs hierarchically along with associated inter-domain fluctuations. The method was first applied to the MD trajectory of substrate-free adenylate kinase to clarify the usefulness of the Motion Tree, which illustrated a clear-cut dynamics picture of the inter-domain motions involving the ATP/AMP lid and the core domain together with the associated amplitudes and correlations. The comparison of two Motion Trees calculated from MD simulations of ligand-free and -bound glutamine binding proteins clarified changes in inherent dynamics upon ligand binding appeared in both large domains and a small loop that stabilized ligand molecule. Another application to a huge protein, a multidrug ATP binding cassette (ABC) transporter, captured significant increases of fluctuations upon binding a drug molecule observed in both large scale inter-subunit motions and a motion localized at a transmembrane helix, which may be a trigger to the subsequent structural change from inward-open to outward-open states to transport the drug molecule. These applications demonstrated the capabilities of Motion Trees to provide an at-a-glance view of various sizes of functional motions inherent in the complicated MD trajectory.

Dynamic architecture of a protein kinase

Protein kinases are dynamically regulated signaling proteins that act as switches in the cell by phosphorylating target proteins. To establish a framework for analyzing linkages between structure, function, dynamics, and allostery in protein kinases, we carried out multiple microsecond-scale molecular-dynamics simulations of protein kinase A (PKA), an exemplar active kinase. We identified residue-residue correlated motions based on the concept of mutual information and used the Girvan-Newman method to partition PKA into structurally contiguous "communities." Most of these communities included 40-60 residues and were associated with a particular protein kinase function or a regulatory mechanism, and well-known motifs based on sequence and secondary structure were often split into different communities. The observed community maps were sensitive to the presence of different ligands and provide a new framework for interpreting long-distance allosteric coupling. Communication between different communities was also in agreement with the previously defined architecture of the protein kinase core based on the "hydrophobic spine" network. This finding gives us confidence in suggesting that community analyses can be used for other protein kinases and will provide an efficient tool for structural biologists. The communities also allow us to think about allosteric consequences of mutations that are linked to disease.

Finding semirigid domains in biomolecules by clustering pair-distance variations

Dynamic variations in the distances between pairs of atoms are used for clustering subdomains of biomolecules. We draw on a well-known target function for clustering and first show mathematically that the assignment of atoms to clusters has to be crisp, not fuzzy, as hitherto assumed. This reduces the computational load of clustering drastically, and we demonstrate results for several biomolecules relevant in immunoinformatics. Results are evaluated regarding the number of clusters, cluster size, cluster stability, and the evolution of clusters over time. Crisp clustering lends itself as an efficient tool to locate semirigid domains in the simulation of biomolecules. Such domains seem crucial for an optimum performance of subsequent statistical analyses, aiming at detecting minute motional patterns related to antigen recognition and signal transduction.

Network-based models as tools hinting at nonevident protein functionality

Network-based models of proteins are popular tools employed to determine dynamic features related to the folded structure. They encompass all topological and geometric computational approaches idealizing proteins as directly interacting nodes. Topology makes use of neighborhood information of residues, and geometry includes relative placement of neighbors. Coarse-grained approaches efficiently predict alternative conformations because of inherent collectivity in the protein structure. Such collectivity is moderated by topological characteristics that also tune neighborhood structure: That rich residues have richer neighbors secures robustness toward random loss of interactions/nodes due to environmental fluctuations/mutations. Geometry conveys the additional information of force balance to network models, establishing the local shape of the energy landscape. Here, residue and/or bond perturbations are critically evaluated to suggest new experiments, as network-based computational techniques prove useful in capturing domain movements and conformational shifts resulting from environmental alterations. Evolutionarily conserved residues are optimally connected, defining a subnetwork that may be utilized for further coarsening.

Rigid-CLL: avoiding constant-distance computations in cell linked-lists algorithms

Many of the existing molecular simulation tools require the efficient identification of the set of nonbonded interacting atoms. This is necessary, for instance, to compute the energy values or the steric contacts between atoms. Cell linked-lists can be used to determine the pairs of atoms closer than a given cutoff distance in asymptotically optimal time. Despite this long-term optimality, many spurious distances are anyway computed with this method. Therefore, several improvements have been proposed, most of them aiming to refine the volume of influence for each atom. Here, we suggest a different improvement strategy based on avoiding to fill cells with those atoms that are always at a constant distance of a given atom. This technique is particularly effective when large groups of the particles in the simulation behave as rigid bodies as it is the case in simplified models considering only few of the degrees of freedom of the molecule. In these cases, the proposed technique can reduce the number of distance computations by more than one order of magnitude, as compared with the standard cell linked-list technique. The benefits of this technique are obtained without incurring in additional computation costs, because it carries out the same operations as the standard cell linked-list algorithm, although in a different order. Since the focus of the technique is the order of the operations, it might be combined with existing improvements based on bounding the volume of influence for each atom.