Predicting permanent and transient protein–protein interfaces

Protein–protein interactions (PPIs) are involved in diverse functions in a cell. To optimize functional roles of interactions, proteins interact with a spectrum of binding affinities. Interactions are conventionally classified into permanent and transient, where the former denotes tight binding between proteins that result in strong complexes, whereas the latter compose of relatively weak interactions that can dissociate after binding to regulate functional activity at specific time point. Knowing the type of interactions has significant implications for understanding the nature and function of PPIs. In this study, we constructed amino acid substitution models that capture mutation patterns at permanent and transient type of protein interfaces, which were found to be different with statistical significance. Using the substitution models, we developed a novel computational method that predicts permanent and transient protein binding interfaces (PBIs) in protein surfaces. Without knowledge of the interacting partner, the method uses a single query protein structure and a multiple sequence alignment of the sequence family. Using a large dataset of permanent and transient proteins, we show that our method, BindML+, performs very well in protein interface classification. A very high area under the curve (AUC) value of 0.957 was observed when predicted protein binding sites were classified. Remarkably, near prefect accuracy was achieved with an AUC of 0.991 when actual binding sites were classified. The developed method will be also useful for protein design of permanent and transient PBIs. © Proteins 2013. © 2012 Wiley Periodicals, Inc.     

Using collections of structural models to predict changes of binding affinity caused by mutations in protein–protein interactions

Protein–protein interactions (PPIs) in all the molecular aspects that take place both inside and outside cells. However, determining experimentally the structure and affinity of PPIs is expensive and time consuming. Therefore, the development of computational tools, as a complement to experimental methods, is fundamental. Here, we present a computational suite: MODPIN, to model and predict the changes of binding affinity of PPIs. In this approach we use homology modeling to derive the structures of PPIs and score them using state‐of‐the‐art scoring functions. We explore the conformational space of PPIs by generating not a single structural model but a collection of structural models with different conformations based on several templates. We apply the approach to predict the changes in free energy upon mutations and splicing variants of large datasets of PPIs to statistically quantify the quality and accuracy of the predictions. As an example, we use MODPIN to study the effect of mutations in the interaction between colicin endonuclease 9 and colicin endonuclease 2 immune protein from Escherichia coli. Finally, we have compared our results with other state‐of‐art methods.     

Reassessing buried surface areas in protein–protein complexes

The buried surface area (BSA), which measures the size of the interface in a protein–protein complex may differ from the accessible surface area (ASA) lost upon association (which we call DSA), if conformation changes take place. To evaluate the DSA, we measure the ASA of the interface atoms in the bound and unbound states of the components of 144 protein–protein complexes taken from the Protein–Protein Interaction Affinity Database of Kastritis et al. (2011). We observe differences exceeding 20%, and a systematic bias in the distribution. On average, the ASA calculated in the bound state of the components is 3.3% greater than in their unbound state, and the BSA, 7% greater than the DSA. The bias is observed even in complexes where the conformation changes are small. An examination of the bound and unbound structures points to a possible origin: local movements optimize contacts with the other component at the cost of internal contacts, and presumably also the binding free energy.     

Interolog interfaces in protein–protein docking

Proteins are essential elements of biological systems, and their function typically relies on their ability to successfully bind to specific partners. Recently, an emphasis of study into protein interactions has been on hot spots, or residues in the binding interface that make a significant contribution to the binding energetics. In this study, we investigate how conservation of hot spots can be used to guide docking prediction. We show that the use of evolutionary data combined with hot spot prediction highlights near‐native structures across a range of benchmark examples. Our approach explores various strategies for using hot spots and evolutionary data to score protein complexes, using both absolute and chemical definitions of conservation along with refinements to these strategies that look at windowed conservation and filtering to ensure a minimum number of hot spots in each binding partner. Finally, structure‐based models of orthologs were generated for comparison with sequence‐based scoring. Using two data sets of 22 and 85 examples, a high rate of top 10 and top 1 predictions are observed, with up to 82% of examples returning a top 10 hit and 35% returning top 1 hit depending on the data set and strategy applied; upon inclusion of the native structure among the decoys, up to 55% of examples yielded a top 1 hit. The 20 common examples between data sets show that more carefully curated interolog data yields better predictions, particularly in achieving top 1 hits. Proteins 2015; 83:1940–1946. © 2015 The Authors. Proteins: Structure, Function, and Bioinformatics Published by Wiley Periodicals, Inc.     

Computational design of second‐site suppressor mutations at protein–protein interfaces

The importance of a protein–protein interaction to a signaling pathway can be established by showing that amino acid mutations that weaken the interaction disrupt signaling, and that additional mutations that rescue the interaction recover signaling. Identifying rescue mutations, often referred to as second‐site suppressor mutations, controls against scenarios in which the initial deleterious mutation inactivates the protein or disrupts alternative protein–protein interactions. Here, we test a structure‐based protocol for identifying second‐site suppressor mutations that is based on a strategy previously described by Kortemme and Baker. The molecular modeling software Rosetta is used to scan an interface for point mutations that are predicted to weaken binding but can be rescued by mutations on the partner protein. The protocol typically identifies three types of specificity switches: knob‐in‐to‐hole redesigns, switching hydrophobic interactions to hydrogen bond interactions, and replacing polar interactions with nonpolar interactions. Computational predictions were tested with two separate protein complexes; the G‐protein Gα_i1 bound to the RGS14 GoLoco motif, and UbcH7 bound to the ubiquitin ligase E6AP. Eight designs were experimentally tested. Swapping a buried hydrophobic residue with a polar residue dramatically weakened binding affinities. In none of these cases were we able to identify compensating mutations that returned binding to wild‐type affinity, highlighting the challenges inherent in designing buried hydrogen bond networks. The strongest specificity switches were a knob‐in‐to‐hole design (20‐fold) and the replacement of a charge–charge interaction with nonpolar interactions (55‐fold). In two cases, specificity was further tuned by including mutations distant from the initial design. Proteins 2010. © 2009 Wiley‐Liss, Inc.     

A minimal model of protein–protein binding affinities

A minimal model of protein–protein binding affinity that takes into account only two structural features of the complex, the size of its interface, and the amplitude of the conformation change between the free and bound subunits, is tested on the 144 complexes of a structure‐affinity benchmark. It yields K_d values that are within two orders of magnitude of the experiment for 67% of the complexes, within three orders for 88%, and fails on 12%, which display either large conformation changes, or a very high or a low affinity. The minimal model lacks the specificity and accuracy needed to make useful affinity predictions, but it should help in assessing the added value of parameters used by more elaborate models, and set a baseline for evaluating their performances.     

Discovery of binding proteins for a protein target using protein–protein docking‐based virtual screening

Target structure‐based virtual screening, which employs protein‐small molecule docking to identify potential ligands, has been widely used in small‐molecule drug discovery. In the present study, we used a protein–protein docking program to identify proteins that bind to a specific target protein. In the testing phase, an all‐to‐all protein–protein docking run on a large dataset was performed. The three‐dimensional rigid docking program SDOCK was used to examine protein–protein docking on all protein pairs in the dataset. Both the binding affinity and features of the binding energy landscape were considered in the scoring function in order to distinguish positive binding pairs from negative binding pairs. Thus, the lowest docking score, the average Z‐score, and convergency of the low‐score solutions were incorporated in the analysis. The hybrid scoring function was optimized in the all‐to‐all docking test. The docking method and the hybrid scoring function were then used to screen for proteins that bind to tumor necrosis factor‐α (TNFα), which is a well‐known therapeutic target for rheumatoid arthritis and other autoimmune diseases. A protein library containing 677 proteins was used for the screen. Proteins with scores among the top 20% were further examined. Sixteen proteins from the top‐ranking 67 proteins were selected for experimental study. Two of these proteins showed significant binding to TNFα in an in vitro binding study. The results of the present study demonstrate the power and potential application of protein–protein docking for the discovery of novel binding proteins for specific protein targets. Proteins 2014; 82:2472–2482. © 2014 Wiley Periodicals, Inc.     

Amino acid substitutions at protein–protein interfaces that modulate the oligomeric state

Despite similarities in their sequence and structure, there are a number of homologous proteins that adopt various oligomeric states. Comparisons of these homologous protein pairs, in terms of residue substitutions at the protein–protein interfaces, have provided fundamental characteristics that describe how proteins interact with each other. We have prepared a dataset composed of pairs of related proteins with different homo‐oligomeric states. Using the protein complexes, the interface residues were identified, and using structural alignments, the shadow‐interface residues have been defined as the surface residues that align with the interface residues. Subsequently, we investigated residue substitutions between the interfaces and the shadow interfaces. Based on the degree of the contributions to the interactions, the aligned sites of the interfaces and shadow interfaces were divided into primary and secondary sites; the primary sites are the focus of this work. The primary sites were further classified into two groups (i.e. exposed and buried) based on the degree to which the residue is buried within the shadow interfaces. Using these classifications, two simple mechanisms that mediate the oligomeric states were identified. In the primary‐exposed sites, the residues on the shadow interfaces are replaced by more hydrophobic or aromatic residues, which are physicochemically favored at protein–protein interfaces. In the primary‐buried sites, the residues on the shadow interfaces are replaced by larger residues that protrude into other proteins. These simple rules are satisfied in 23 out of 25 Structural Classification of Proteins (SCOP) families with a different‐oligomeric‐state pair, and thus represent a basic strategy for modulating protein associations and dissociations. Proteins 2010. © 2009 Wiley‐Liss, Inc.     

Linkers in the structural biology of protein–protein interactions

Linkers or spacers are short amino acid sequences created in nature to separate multiple domains in a single protein. Most of them are rigid and function to prohibit unwanted interactions between the discrete domains. However, Gly‐rich linkers are flexible, connecting various domains in a single protein without interfering with the function of each domain. The advent of recombinant DNA technology made it possible to fuse two interacting partners with the introduction of artificial linkers. Often, independent proteins may not exist as stable or structured proteins until they interact with their binding partner, following which they gain stability and the essential structural elements. Gly‐rich linkers have been proven useful for these types of unstable interactions, particularly where the interaction is weak and transient, by creating a covalent link between the proteins to form a stable protein–protein complex. Gly‐rich linkers are also employed to form stable covalently linked dimers, and to connect two independent domains that create a ligand‐binding site or recognition sequence. The lengths of linkers vary from 2 to 31 amino acids, optimized for each condition so that the linker does not impose any constraints on the conformation or interactions of the linked partners. Various structures of covalently linked protein complexes have been described using X‐ray crystallography, nuclear magnetic resonance and cryo‐electron microscopy techniques. In this review, we evaluate several structural studies where linkers have been used to improve protein quality, to produce stable protein–protein complexes, and to obtain protein dimers.     

Feature selection and classification of protein–protein complexes based on their binding affinities using machine learning approaches

Protein–protein interactions are intrinsic to virtually every cellular process. Predicting the binding affinity of protein–protein complexes is one of the challenging problems in computational and molecular biology. In this work, we related sequence features of protein–protein complexes with their binding affinities using machine learning approaches. We set up a database of 185 protein–protein complexes for which the interacting pairs are heterodimers and their experimental binding affinities are available. On the other hand, we have developed a set of 610 features from the sequences of protein complexes and utilized Ranker search method, which is the combination of Attribute evaluator and Ranker method for selecting specific features. We have analyzed several machine learning algorithms to discriminate protein‐protein complexes into high and low affinity groups based on their K_d values. Our results showed a 10‐fold cross‐validation accuracy of 76.1% with the combination of nine features using support vector machines. Further, we observed accuracy of 83.3% on an independent test set of 30 complexes. We suggest that our method would serve as an effective tool for identifying the interacting partners in protein–protein interaction networks and human–pathogen interactions based on the strength of interactions. Proteins 2014; 82:2088–2096. © 2014 Wiley Periodicals, Inc.     

Protein‐spanning water networks and implications for prediction of protein–protein interactions mediated through hydrophobic effects

Hydrophobic effects, often conflated with hydrophobic forces, are implicated as major determinants in biological association and self‐assembly processes. Protein–protein interactions involved in signaling pathways in living systems are a prime example where hydrophobic effects have profound implications. In the context of protein–protein interactions, a priori knowledge of relevant binding interfaces (i.e., clusters of residues involved directly with binding interactions) is difficult. In the case of hydrophobically mediated interactions, use of hydropathy‐based methods relying on single residue hydrophobicity properties are routinely and widely used to predict propensities for such residues to be present in hydrophobic interfaces. However, recent studies suggest that consideration of hydrophobicity for single residues on a protein surface require accounting of the local environment dictated by neighboring residues and local water. In this study, we use a method derived from percolation theory to evaluate spanning water networks in the first hydration shells of a series of small proteins. We use residue‐based water density and single‐linkage clustering methods to predict hydrophobic regions of proteins; these regions are putatively involved in binding interactions. We find that this simple method is able to predict with sufficient accuracy and coverage the binding interface residues of a series of proteins. The approach is competitive with automated servers. The results of this study highlight the importance of accounting of local environment in determining the hydrophobic nature of individual residues on protein surfaces. Proteins 2014; 82:3312–3326. © 2014 Wiley Periodicals, Inc.     

On the protein–protein diffusional encounter complex

When two proteins diffuse together to form a bound complex, an intermediate is formed at the end‐point of diffusional association which is called the encounter complex. Its characteristics are important in determining association rates, yet its structure cannot be directly observed experimentally. Here, we address the problem of how to construct the ensemble of three‐dimensional structures which constitute the protein–protein diffusional encounter complex using available experimental data describing the dependence of protein association rates on mutation and on solvent ionic strength and viscosity. The magnitude of the association rates is fitted well using a variety of definitions of encounter complexes in which the two proteins are located at up to about 17 Å root‐mean‐squared distance from their relative arrangement in the bound complex. Analysis of the ionic strength dependence of bimolecular association rates shows that this is determined to a greater extent by the (protein charge) – (salt ion) separation distance than by the protein–protein charge separation distance. Consequently, ionic strength dependence of association rates provides little information about the geometry of the encounter complex. On the other hand, experimental data on electrostatic rate enhancement, mutation and viscosity dependence suggest a model of the encounter complex in which the two proteins form a subset of the contacts present in the bound complex and are significantly desolvated. Copyright © 1999 John Wiley & Sons, Ltd.     

Organic dyes as small molecule protein–protein interaction inhibitors for the CD40–CD154 costimulatory interaction

It is becoming increasingly clear that small molecules can often act as effective protein–protein interaction (PPI) inhibitors, an area of increasing interest for its many possible therapeutic applications. We have identified several organic dyes and related small molecules that (i) concentration‐dependently inhibit the important CD40–CD154 costimulatory interaction with activities in the low micromolar (µM) range, (ii) show selectivity toward this particular PPI, (iii) seem to bind on the surface of CD154, and (iv) concentration‐dependently inhibit the CD154‐induced B cell proliferation. They were identified through an iterative activity screening/structural similarity search procedure starting with suramin as lead, and the best smaller compounds, the main focus of the present work, achieved an almost 3‐fold increase in ligand efficiency (ΔG⁰/nonhydrogen atom = 0.8 kJ/N_nHa) approaching the average of known promising small‐molecule PPI inhibitors (～1.0 kJ/N_nHa). Since CD154 is a member of the tumor necrosis factor (TNF) superfamily of cell surface interaction molecules, inhibitory activities on the TNF‐R1–TNF‐α interactions were also determined to test for specificity, and the compounds selected here all showed more than 30‐fold selectivity toward the CD40–CD154 interaction. Because of their easy availability in various structural scaffolds and because of their good protein‐binding ability, often explored for tissue‐specific staining and other purposes, such organic dyes can provide a valuable addition to the chemical space searched to identify small molecule PPI inhibitors in general. Copyright © 2009 John Wiley & Sons, Ltd.     

Interplay between binding affinity and kinetics in protein–protein interactions

To clarify the interplay between the binding affinity and kinetics of protein–protein interactions, and the possible role of intrinsically disordered proteins in such interactions, molecular simulations were carried out on 20 protein complexes. With bias potential and reweighting techniques, the free energy profiles were obtained under physiological affinities, which showed that the bound‐state valley is deep with a barrier height of 12 ? 33 RT. From the dependence of the affinity on interface interactions, the entropic contribution to the binding affinity is approximated to be proportional to the interface area. The extracted dissociation rates based on the Arrhenius law correlate reasonably well with the experimental values (Pearson correlation coefficient R = 0.79). For each protein complex, a linear free energy relationship between binding affinity and the dissociation rate was confirmed, but the distribution of the slopes for intrinsically disordered proteins showed no essential difference with that observed for ordered proteins. A comparison with protein folding was also performed. Proteins 2016; 84:920–933. © 2016 Wiley Periodicals, Inc.     

Weak conservation of structural features in the interfaces of homologous transient protein–protein complexes

Residue types at the interface of protein–protein complexes (PPCs) are known to be reasonably well conserved. However, we show, using a dataset of known 3‐D structures of homologous transient PPCs, that the 3‐D location of interfacial residues and their interaction patterns are only moderately and poorly conserved, respectively. Another surprising observation is that a residue at the interface that is conserved is not necessarily in the interface in the homolog. Such differences in homologous complexes are manifested by substitution of the residues that are spatially proximal to the conserved residue and structural differences at the interfaces as well as differences in spatial orientations of the interacting proteins. Conservation of interface location and the interaction pattern at the core of the interfaces is higher than at the periphery of the interface patch. Extents of variability of various structural features reported here for homologous transient PPCs are higher than the variation in homologous permanent homomers. Our findings suggest that straightforward extrapolation of interfacial nature and inter‐residue interaction patterns from template to target could lead to serious errors in the modeled complex structure. Understanding the evolution of interfaces provides insights to improve comparative modeling of PPC structures.     

Ligand loading at the surface of an optical biosensor and its effect upon the kinetics of protein–protein interactions

Optical biosensors are finding increasing use in the determination of kinetic and equilibrium constants for a variety of biomolecular interactions. Usually these biosensors require one biomolecule, the ligand, to be covalently attached to a hydrogel matrix which itself is bonded to the sensing surface. The ligands partner, the ligate, then binds from solution resulting in a measurable change in response which the instrument records as a function of time. Although in many cases, optical biosensors are used in order to obtain parameters that relate to interactions in solution, it is becoming clear that measurements involving the interaction of ligate with immobilized ligands on surfaces require careful experimental design. Here we report on how the density of ligand loading within the hydogel matrix affects the measured interaction kinetics. It is found that crowding of ligand within this matrix results in a significant reduction in the measured association rate constant, with a corresponding effect in the calculated overall affinity. However, measurements at low ligand loadings show association rate constants that are comparable to those measured in solution. Clearly, where this comparison is required, it is important to perform measurements under such conditions. © 1997 John Wiley & Sons, Ltd.     

Cold spots are universal in protein–protein interactions

Proteins interact with each other through binding interfaces that differ greatly in size and physico‐chemical properties. Within the binding interface, a few residues called hot spots contribute the majority of the binding free energy and are hence irreplaceable. In contrast, cold spots are occupied by suboptimal amino acids, providing possibility for affinity enhancement through mutations. In this study, we identify cold spots due to cavities and unfavorable charge interactions in multiple protein–protein interactions (PPIs). For our cold spot analysis, we first use a small affinity database of PPIs with known structures and affinities and then expand our search to nearly 4000 homo‐ and heterodimers in the Protein Data Bank (PDB). We observe that cold spots due to cavities are present in nearly all PPIs unrelated to their binding affinity, while unfavorable charge interactions are relatively rare. We also find that most cold spots are located in the periphery of the binding interface, with high‐affinity complexes showing fewer centrally located colds spots than low‐affinity complexes. A larger number of cold spots is also found in non‐cognate interactions compared to their cognate counterparts. Furthermore, our analysis reveals that cold spots are more frequent in homo‐dimeric complexes compared to hetero‐complexes, likely due to symmetry constraints imposed on sequences of homodimers. Finally, we find that glycines, glutamates, and arginines are the most frequent amino acids appearing at cold spot positions. Our analysis emphasizes the importance of cold spot positions to protein evolution and facilitates protein engineering studies directed at enhancing binding affinity and specificity in a wide range of applications.     

A non‐redundant protein–RNA docking benchmark version 2.0

We present an updated version of the protein–RNA docking benchmark, which we first published four years back. The non‐redundant protein–RNA docking benchmark version 2.0 consists of 126 test cases, a threefold increase in number compared to its previous version. The present version consists of 21 unbound–unbound cases, of which, in 12 cases, the unbound RNAs are taken from another complex. It also consists of 95 unbound–bound cases where only the protein is available in the unbound state. Besides, we introduce 10 new bound–unbound cases where only the RNA is found in the unbound state. Based on the degree of conformational change of the interface residues upon complex formation the benchmark is classified into 72 rigid‐body cases, 25 semiflexible cases and 19 full flexible cases. It also covers a wide range of conformational flexibility including small side chain movement to large domain swapping in protein structures as well as flipping and restacking in RNA bases. This benchmark should provide the docking community with more test cases for evaluating rigid‐body as well as flexible docking algorithms. Besides, it will also facilitate the development of new algorithms that require large number of training set. The protein–RNA docking benchmark version 2.0 can be freely downloaded from http://www.csb.iitkgp.ernet.in/applications/PRDBv2 . Proteins 2017; 85:256–267. © 2016 Wiley Periodicals, Inc.     

Solvated protein–protein docking using Kyte‐Doolittle‐based water preferences

HADDOCK is one of the few docking programs that can explicitly account for water molecules in the docking process. Its solvated docking protocol starts from hydrated molecules and a fraction of the resulting interfacial waters is subsequently removed in a biased Monte Carlo procedure based on water‐mediated contact probabilities. The latter were derived from an analysis of water contact frequencies from high‐resolution crystal structures. Here, we introduce a simple water‐mediated amino acid–amino acid contact probability scale derived from the Kyte‐Doolittle hydrophobicity scale and assess its performance on the largest high‐resolution dataset developed to date for solvated docking. Both scales yield high‐quality docking results. The novel and simple hydrophobicity scale, which should reflect better the physicochemical principles underlying contact propensities, leads to a performance improvement of around 10% in ranking, cluster quality and water recovery at the interface compared with the statistics‐based original solvated docking protocol. Proteins 2013. © 2012 Wiley Periodicals, Inc.