Characterization of Protein-Protein Interfaces through a Protein Contact Network Approach

Head or tail? A molecular dynamics approach to the complex structure of TNF-associated factor TRAF2

Tumor necrosis factor receptor-associated factor proteins (TRAFs) are trimeric proteins that play a fundamental role in signaling, acting as intermediaries between the tumor necrosis factor (TNF) receptors and the proteins that transmit the downstream signal. The monomeric subunits of all the TRAF family members share a common tridimensional structure: a C-terminal globular domain and a long coiled-coil tail characterizing the N-terminal section. In this study, the dependence of the TRAF2 dynamics on the length of its tail was analyzed in silico. In particular, we used the available crystallographic structure of a C-terminal fragment of TRAF2 (168 out of 501 a.a.), TRAF2-C, and that of a longer construct, addressed as TRAF2-plus, that we have re-constructed using the AlphaFold2 code. The results indicate that the longer N-terminal tail of TRAF2-plus has a strong influence on the dynamics of the globular regions in the protein C-terminal head. In fact, the quaternary interactions among the TRAF2-C subunits change asymmetrically in time, while the movements of TRAF2-plus monomers are rather limited and more ordered than those of the shorter construct. Such findings shed a new light on the dynamics of TRAF subunits and on the protein mechanism in vivo, since TRAF monomer-trimer equilibrium is crucial for several reasons (receptor recognition, membrane binding, hetero-oligomerization).

Assessing Combinatorial Diversity of Aureochrome Basic Leucine Zippers through Genome-Wide Screening

Aureochromes are unique blue light-responsive light-oxygen-voltage (LOV) photoreceptors cum basic leucine zipper (bZIP) transcription factors (TFs), present exclusively in photosynthetic marine stramenopiles. Considering the availability of the complete genome sequence, this study focuses on aureochromes from Ectocarpus siliculosus. Aureochromes mediate light-regulated developmental responses in these brown photosynthetic algae. Both the LOV sensor and the bZIP effector show overall sequence-structure conservation. The structurally similar LOV + bZIP modules of aureochrome homologs/paralogs prefer a dimeric state. Besides a heterogeneous linker connecting the sensor-effector and a flexible N-terminal region, the sequence composition of both domains is vital. Aureochromes execute diverse cellular responses in different photosynthetic stramenopiles - though their activities can vary even within a given algal species. Therefore, it is important to understand whether aureochromes select dimerization partners from the same family or interact with other bZIPs as well. To regulate multifarious biological activities, it is possible that aureochromes activate the global TF interaction network. Following homo/heterodimer modeling, we address the compatibility of dimerization partners by screening through heptad repeats. We evaluate the dimer interface area in terms of gain in solvation energy and the number of hydrogen bonds/salt bridge interactions. We further explore the relative stability of these structures from a graph-theoretic perspective through well-studied measures such as the energy of the graph, average participation coefficient, and betweenness centrality. Furthermore, we also conduct an information-theoretic analysis using hitherto understudied measures such as network information centrality and Kullback-Leibler divergence. We find that all our investigations into the relative stability of the dimers using diverse methods from bioinformatics, network science, and information theory are in harmonious agreement. Coupling preferences of monomers in aureochromes can be further translated to design novel optogenetic tools useful for understanding human development and disease.

In-Silico Characterization of von Willebrand Factor Bound to FVIII

Factor VIII belongs to the coagulation cascade and is expressed as a long pre-protein (mature form, 2351 amino acids long). FVIII is deficient or defective in hemophilic A patients, who need to be treated with hemoderivatives or recombinant FVIII substitutes, i.e., biologic drugs. The interaction between FVIII and von Willebrand factor (VWF) influences the pharmacokinetics of FVIII medications. In vivo, full-length FVIII (FL-FVIII) is secreted in a plasma-inactive form, which includes the B domain, which is then proteolyzed by thrombin protease activity, leading to an inactive plasma intermediate. In this work, we analyzed through a computational approach the binding of VWF with two structure models of FVIII (secreted full-length with B domain, and B domain-deleted FVIII). We included in our analysis the atomic model of efanesoctocog alfa, a novel and investigational recombinant FVIII medication, in which the VWF is covalently linked to FVIII. We carried out a structural analysis of VWF/FVIII interfaces by means of protein–protein docking, PISA (Proteins, Interfaces, Structures and Assemblies), and protein contact networks (PCN) analyses. Accordingly, our computational approaches to previously published experimental data demonstrated that the domains A3-C1 of B domain-deleted FVIII (BDD-FVIII) is the preferential binding site for VWF. Overall, our computational approach applied to topological analysis of protein–protein interface can be aimed at the rational design of biologic drugs other than FVIII medications.

Biophysical Insight into the SARS-CoV2 Spike-ACE2 Interaction and Its Modulation by Hepcidin through a Multifaceted Computational Approach

At the center of the SARS-CoV2 infection, the spike protein and its interaction with the human receptor ACE2 play a central role in the molecular machinery of SARS-CoV2 infection of human cells. Vaccine therapies are a valuable barrier to the worst effects of the virus and to its diffusion, but the need of purposed drugs is emerging as a core target of the fight against COVID19. In this respect, the repurposing of drugs has already led to discovery of drugs thought to reduce the effects of the cytokine storm, but still a drug targeting the spike protein, in the infection stage, is missing. In this work, we present a multifaceted computational approach strongly grounded on a biophysical modeling of biological systems, so to disclose the interaction of the SARS-CoV2 spike protein with ACE2 with a special focus to an allosteric regulation of the spike-ACE2 interaction. Our approach includes the following methodologies: Protein Contact Networks and Network Clustering, Targeted Molecular Dynamics, Elastic Network Modeling, Perturbation Response Scanning, and a computational analysis of energy flow and SEPAS as a protein-softness and monomer-based affinity predictor. We applied this approach to free (closed and open) states of spike protein and spike-ACE2 complexes. Eventually, we analyzed the interactions of free and bound forms of spike with hepcidin (HPC), the major hormone in iron regulation, recently addressed as a central player in the COVID19 pathogenesis, with a special emphasis to the most severe outcomes. Our results demonstrate that, compared with closed and open states, the spike protein in the ACE2-bound state shows higher allosteric potential. The correspondence between hinge sites and the Allosteric Modulation Region (AMR) in the S-ACE complex suggests a molecular basis for hepcidin involvement in COVID19 pathogenesis. We verify the importance of AMR in different states of spike and then study its interactions with HPC and the consequence of the HPC-AMR interaction on spike dynamics and its affinity for ACE2. We propose two complementary mechanisms for HPC effects on spike of SARS-CoV-2; (a) HPC acts as a competitive inhibitor when spike is in a preinfection state (open and with no ACE2), (b) the HPC-AMR interaction pushes the spike structure into the safer closed state. These findings need clear molecular in vivo verification beside clinical observations.

Emerging landscape of molecular interaction networks: Opportunities, challenges and prospects

Network biology finds application in interpreting molecular interaction networks and providing insightful inferences using graph theoretical analysis of biological systems. The integration of computational bio-modelling approaches with different hybrid network-based techniques provides additional information about the behaviour of complex systems. With increasing advances in high-throughput technologies in biological research, attempts have been made to incorporate this information into network structures, which has led to a continuous update of network biology approaches over time. The newly minted centrality measures accommodate the details of omics data and regulatory network structure information. The unification of graph network properties with classical mathematical and computational modelling approaches and technologically advanced approaches like machine-learning- and artificial intelligence-based algorithms leverages the potential application of these techniques. These computational advances prove beneficial and serve various applications such as essential gene prediction, identification of drug–disease interaction and gene prioritization. Hence, in this review, we have provided a comprehensive overview of the emerging landscape of molecular interaction networks using graph theoretical approaches. With the aim to provide information on the wide range of applications of network biology approaches in understanding the interaction and regulation of genes, proteins, enzymes and metabolites at different molecular levels, we have reviewed the methods that utilize network topological properties, emerging hybrid network-based approaches and applications that integrate machine learning techniques to analyse molecular interaction networks. Further, we have discussed the applications of these approaches in biomedical research with a note on future prospects.

Effect of Lithium Drug on Binding Affinities of Glycogen Synthase Kinase-3 β to Its Network Partners: A New Computational Approach

Finding new methods to study the effect of small molecules on protein interaction networks provides us with invaluable tools in the fields of pharmacodynamics and drug design. Lithium is an antimanic drug that has been used for the treatment of bipolar disorder for more than 60 years. Here, we utilized a new approach to study the effect of lithium as a drug on the protein interaction network of GSK-3β as a hub protein and computed the affinities of GSK-3β to its partners in the presence of lithium or sodium ions. For this purpose, ensembles of GSK-3β protein structures were created in the presence of either lithium or sodium ions using adaptive tempering molecular dynamics simulations. The protein binding patches of GSK-3β for its partners were determined, and finally, the affinity of each binding patch to the related partner was computed for structures of ensembles using a monomer-based approach. Besides, by comparing structural dynamics of GSK-3β during MD simulations in the presence of LiCl and NaCl, we suggested a new mechanism for the inhibitory effect of lithium on GSK-3β.

Centrality Measures in Residue Interaction Networks to Highlight Amino Acids in Protein–Protein Binding

Residue interaction networks (RINs) describe a protein structure as a network of interacting residues. Central nodes in these networks, identified by centrality analyses, highlight those residues that play a role in the structure and function of the protein. However, little is known about the capability of such analyses to identify residues involved in the formation of macromolecular complexes. Here, we performed six different centrality measures on the RINs generated from the complexes of the SKEMPI 2 database of changes in protein–protein binding upon mutation in order to evaluate the capability of each of these measures to identify major binding residues. The analyses were performed with and without the crystallographic water molecules, in addition to the protein residues. We also investigated the use of a weight factor based on the inter-residue distances to improve the detection of these residues. We show that for the identification of major binding residues, closeness, degree, and PageRank result in good precision, whereas betweenness, eigenvector, and residue centrality analyses give a higher sensitivity. Including water in the analysis improves the sensitivity of all measures without losing precision. Applying weights only slightly raises the sensitivity of eigenvector centrality analysis. We finally show that a combination of multiple centrality analyses is the optimal approach to identify residues that play a role in protein–protein interaction.

Disclosing Allostery Through Protein Contact Networks

Proteins are located in the twilight zone between chemistry and biology, where a peculiar kind of complexity starts. Proteins are the smallest 'devices' showing a sensible adaptation to their environment by the production of appropriate behavior when facing a specific stimulus. This fact qualifies (from the 'effector' side) proteins as nanomachines working as catalysts, motors, or switches. However (from the sensor side), the need to single out the 'specific stimulus' out of thermal noise qualifies proteins as information processing devices. Allostery corresponds to the modification of the configuration (in a broad sense) of the protein molecule in response to a specific stimulus in a non-strictly local way, thereby connecting the sensor and effector sides of the nanomachine. This is why the 'disclosing' of allostery phenomenon is at the very heart of protein function; in this chapter, we will demonstrate how a network-based representation of protein structure in terms of nodes (aminoacid residues) and edges (effective contacts between residues) is the natural language for getting rid of allosteric phenomena and, more in general, of protein structure/function relationships.

Molecular Dynamics Simulation Techniques as Tools in Drug Discovery and Pharmacology: A Focus on Allosteric Drugs

Allosteric drugs are ligands that when bound to an allosteric site modify the conformational state of the pharmacological target, leading then to a modification of functional response upon binding of the endogenous ligand. Pharmacological targets are defined as biological entities, to which a ligand/drug binds and leads to a functional effect. Pharmacological targets can be proteins or nucleic acids. Computational approaches such as molecular dynamics (MD) sped up discovery and identification of allosteric binding sites and allosteric ligands. Classical all-atom and hybrid classical/quantum MD simulations can be generalized as simulation techniques aimed at analysis of atoms and molecular motion. Main limitations of MD simulations are related to high computational costs, that in turn limit the conformational sampling of biological systems. Indeed, other techniques have been developed to overcome limitations of MD, such as enhanced sampling MD simulations. In this chapter, classical MD and enhanced sampling MD simulations will be described, along with their application to drug discovery, with a focus on allosteric drugs.

ANCA: A Web Server for Amino Acid Networks Construction and Analysis

Amino acid network (AAN) models empower us to gain insights into protein structures and functions by describing a protein 3D structure as a graph, where nodes represent residues and edges as amino acid interactions. Here, we present the ANCA, an interactive Web server for Amino Acids Network Construction and Analysis based on a single structure or a set of structures from the Protein Data Bank. The main purpose of ANCA is to provide a portal for three types of an environment-dependent residue contact energy (ERCE)-based network model, including amino acid contact energy network (AACEN), node-weighted amino acid contact energy network (NACEN), and edge-weighted amino acid contact energy network (EACEN). For comparison, the C-alpha distance-based network model is also included, which can be extended to protein–DNA/RNA complexes. Then, the analyses of different types of AANs were performed and compared from node, edge, and network levels. The network and corresponding structure can be visualized directly in the browser. The ANCA enables researchers to investigate diverse concerns in the framework of AAN, such as the interpretation of allosteric regulation and functional residues. The ANCA portal, together with an extensive help, is available at http://sysbio.suda.edu.cn/anca/.

Surveying the Side-Chain Network Approach to Protein Structure and Dynamics: The SARS-CoV-2 Spike Protein as an Illustrative Case

Network theory-based approaches provide valuable insights into the variations in global structural connectivity between different dynamical states of proteins. Our objective is to review network-based analyses to elucidate such variations, especially in the context of subtle conformational changes. We present technical details of the construction and analyses of protein structure networks, encompassing both the non-covalent connectivity and dynamics. We examine the selection of optimal criteria for connectivity based on the physical concept of percolation. We highlight the advantages of using side-chain-based network metrics in contrast to backbone measurements. As an illustrative example, we apply the described network approach to investigate the global conformational changes between the closed and partially open states of the SARS-CoV-2 spike protein. These conformational changes in the spike protein is crucial for coronavirus entry and fusion into human cells. Our analysis reveals global structural reorientations between the two states of the spike protein despite small changes between the two states at the backbone level. We also observe some differences at strategic locations in the structures, correlating with their functions, asserting the advantages of the side-chain network analysis. Finally, we present a view of allostery as a subtle synergistic-global change between the ligand and the receptor, the incorporation of which would enhance drug design strategies.

The Discovery of a Putative Allosteric Site in the SARS-CoV-2 Spike Protein Using an Integrated Structural/Dynamic Approach

SARS-CoV-2 has caused the largest pandemic of the twenty-first century (COVID-19), threatening the life and economy of all countries in the world. The identification of novel therapies and vaccines that can mitigate or control this global health threat is among the most important challenges facing biomedical sciences. To construct a long-term strategy to fight both SARS-CoV-2 and other possible future threats from coronaviruses, it is critical to understand the molecular mechanisms underlying the virus action. The viral entry and associated infectivity stems from the formation of the SARS-CoV-2 spike protein complex with angiotensin-converting enzyme 2 (ACE2). The detection of putative allosteric sites on the viral spike protein molecule can be used to elucidate the molecular pathways that can be targeted with allosteric drugs to weaken the spike-ACE2 interaction and, thus, reduce viral infectivity. In this study, we present the results of the application of different computational methods aimed at detecting allosteric sites on the SARS-CoV-2 spike protein. The adopted tools consisted of the protein contact networks (PCNs), SEPAS (Affinity by Flexibility), and perturbation response scanning (PRS) based on elastic network modes. All of these methods were applied to the ACE2 complex with both the SARS-CoV2 and SARS-CoV spike proteins. All of the adopted analyses converged toward a specific region (allosteric modulation region [AMR]), present in both complexes and predicted to act as an allosteric site modulating the binding of the spike protein with ACE2. Preliminary results on hepcidin (a molecule with strong structural and sequence with AMR) indicated an inhibitory effect on the binding affinity of the spike protein toward the ACE2 protein.

Novel indole derivatives targeting HuR-mRNA complex to counteract high glucose damage in retinal endothelial cells

The ELAVL1 (or human antigen R - HuR) RNA binding protein stabilizes the mRNA, with an AU-rich element, of several genes such as growth factors (i.e. VEGF) and inflammatory cytokines (i.e. TNFα). We hereby carried out a virtual screening campaign in order to identify and test novel HuR-mRNA disruptors. Best-scored compounds were tested in an in-vitro model of diabetic retinopathy, namely human retinal endothelial cells (HRECs) challenged with high-glucose levels (25 mM). HuR, VEGF and TNFα protein contents were evaluated by western-blot analysis in total cell lysates. VEGF and TNFα released from HRECs were measured in cell medium by ELISA. We found that two derivatives bearing indole moiety, VP12/14 and VP12/110, modulated HuR expression and decreased VEGF and TNF-α release by HREC exposed to high glucose (HG) levels. VP12/14 and VP12/110 inhibited VEGF and TNF-α release in HRECs challenged with high glucose levels, similarly to dihydrotanshinone (DHTS), a small molecule known to interfere with HuR- TNFα mRNA binding. The present findings demonstrated that VP12/14 and VP12/110 are innovative molecules with anti-inflammatory and anti-angiogenic properties, suggesting their potential use as novel candidates for treatment of diabetic retinopathy.

Soft regions of protein surface are potent for stable dimer formation

By having knowledge about the characteristics of protein interaction interfaces, we will be able to manipulate protein complexes for therapies. Dimer state is considered as the primary alphabet of the most proteins' quaternary structure. The properties of binding interface between subunits and of noninterface region define the specificity and stability of the intended protein complex. Considering some topological properties and amino acids' affinity for binding in interfaces of protein dimers, we construct the interface-specific recurrence plots. The data obtained from recurrence quantitative analysis, and accessibility-related metrics help us to classify the protein dimers into four distinct classes. Some mechanical properties of binding interfaces are computed for each predefined class of the dimers. The computed mechanical characteristics of binding patch region are compared with those of nonbinding region of proteins. Our observations indicate that the mechanical properties of protein binding sites have a decisive impact on determining the dimer stability. We introduce a new concept in analyzing protein structure by considering mechanical properties of protein structure. We conclude that the interface region between subunits of stable dimers is usually mechanically softer than the interface of unstable protein dimers. AbbreviationsAABaverage affinity for bindingANManisotropic network modelAPCaffinity propagation clusteringASAaccessible surface areaCCDinter residues distanceCSCcomplex stability codeDMdistance matrixΔG_dissPISA-computed dissociation free energyGNMGaussian normal mode analysisNMAnormal mode analysisPBPprotein binding patchPISAproteins, interfaces, structures and assembliesrASArelative accessible area in respect to unfolded state of residuesRMrecurrence matrixrPrelative protrusionRPrecurrence plotRQArecurrence quantitative analysisSEMstandard error of meanCommunicated by Ramaswamy H. Sarma.

Aflibercept regulates retinal inflammation elicited by high glucose via the PlGF/ERK pathway

Diabetic retinopathy (DR) is a secondary complication of diabetes. DR can cause irreversible blindness, and its pathogenesis is considered multifactorial. DR can progress from non-proliferative DR to proliferative DR, characterized by retinal neovascularization. The main cause of vision loss in diabetic patients is diabetic macular edema, caused by vessel leakage and blood retinal barrier breakdown. Currently, aflibercept is an anti-VEGF approved for diabetic macular edema. Aflibercept can bind several members of vascular permeability factors, namely VEGF-A, B, and PlGF. We analyzed the aflibercept-PlGF complex at molecular level, through an in silico approach. In order to explore the role of PlGF in DR, we treated primary human retinal endothelial cells (HRECs) and mouse retinal epithelial cells (RPEs) with aflibercept and an anti-PlGF antibody. We explored the hypothesis that aflibercept has anti-inflammatory action through blocking of PlGF signaling and the ERK axis in an in vitro and in vivo model of DR. Both aflibercept and the anti-PlGF antibody exerted protective effects on retinal cells, by inhibition of the ERK pathway. Moreover, aflibercept significantly decreased (p < 0.05) the expression of TNF-α in an in vitro and in vivo model of DR. Therefore, our data suggest that inhibition of PlGF signaling, or a selective blocking, may be useful in the management of early phases of DR when the inflammatory process is largely involved.

Comparative Study of Elastic Network Model and Protein Contact Network for Protein Complexes: The Hemoglobin Case

The overall topology and interfacial interactions play key roles in understanding structural and functional principles of protein complexes. Elastic Network Model (ENM) and Protein Contact Network (PCN) are two widely used methods for high throughput investigation of structures and interactions within protein complexes. In this work, the comparative analysis of ENM and PCN relative to hemoglobin (Hb) was taken as case study. We examine four types of structural and dynamical paradigms, namely, conformational change between different states of Hbs, modular analysis, allosteric mechanisms studies, and interface characterization of an Hb. The comparative study shows that ENM has an advantage in studying dynamical properties and protein-protein interfaces, while PCN is better for describing protein structures quantitatively both from local and from global levels. We suggest that the integration of ENM and PCN would give a potential but powerful tool in structural systems biology.

GH32 family activity: a topological approach through protein contact networks

The application of Protein Contact Networks methodology allowed to highlight a novel response of border region between the two domains to substrate binding. Glycoside hydrolases (GH) are enzymes that mainly hydrolyze the glycosidic bond between two carbohydrates or a carbohydrate and a non-carbohydrate moiety. These enzymes are involved in many fundamental and diverse biological processes in plants. We have focused on the GH32 family, including enzymes very similar in both sequence and structure, each having however clear specificities of substrate preferences and kinetic properties. Structural and topological differences among proteins of the GH32 family have been here identified by means of an emerging approach (Protein Contact network, PCN) based on the formalization of 3D structures as contact networks among amino-acid residues. The PCN approach proved successful in both reconstructing the already known functional domains and in identifying the structural counterpart of the properties of GH32 enzymes, which remain uncertain, like their allosteric character. The main outcome of the study was the discovery of the activation upon binding of the border (cleft) region between the two domains. This reveals the allosteric nature of the enzymatic activity for all the analyzed forms in the GH32 family, a character yet to be highlighted in biochemical studies. Furthermore, we have been able to recognize a topological signature (graph energy) of the different affinity of the enzymes towards small and large substrates.