An algorithm for predicting protein-protein interaction sites: Abnormally exposed amino acid residues and secondary structure elements

Multiprotein systems mediate most regulatory processes in living organisms. Although the structures of the individual proteins are often defined, less is known of the structures of multiprotein systems. Computational methods for predicting interfaces, using evolutionary conservation and/or physicochemical data, have been developed. Here we consider the use of solvent accessibility, residue propensity, and hydrophobicity, in conjunction with secondary structure data, as prediction parameters. We analyze the influence of residue type and secondary structure on solvent accessibility and define a measure of "relative exposedness." Clustering abnormally high scoring residues provides a basis for predicting interaction sites. The analysis is extended to investigate abnormally exposed secondary structure elements, particularly beta-sheet strands. We show that surface-exposed beta-strands lacking protective features are more likely to be found at protein-protein interfaces, allowing us to create an algorithm with approximately 68% and approximately 75% accuracy in differentiating between interacting and edge strands in isolated beta-strands and beta-sheet strands, respectively. These methods of identifying abnormally exposed surface regions are combined in an algorithm, which, on a data set of 77 unbound and disjoint (single chain extracted from complex) structures, predicts 79% of the protein-protein interfaces correctly. If enzyme-inhibitor complexes, where the inhibitor mimics a nonprotein substrate, are excluded, the accuracy increases to 85%.     

Assessing methods for identifying pair-wise atomic contacts across binding interfaces

An essential step in understanding the molecular basis of protein-protein interactions is the accurate identification of inter-protein contacts. We evaluate a number of common methods used in analyzing protein-protein interfaces: a Voronoi polyhedra-based approach, changes in solvent accessible surface area (DeltaSASA) and various radial cutoffs (closest atom, Cbeta, and centroid). First, we compared the Voronoi polyhedra-based analysis to the DeltaSASA and show that using Voronoi polyhedra finds knob-in-hole contacts. To assess the accuracy between the Voronoi polyhedra-based approach and the various radial cutoff methods, two sets of data were used: a small set of 75 experimental mutants and a larger one of 592 structures of protein-protein interfaces. In an assessment using the small set, the Voronoi polyhedra-based methods, a solvent accessible surface area method, and the closest atom radial method identified 100% of the direct contacts defined by mutagenesis data, but only the Voronoi polyhedra-based method found no false positives. The other radial methods were not able to find all of the direct contacts even using a cutoff of 9A. With the larger set of structures, we compared the overall number contacts using the Voronoi polyhedra-based method as a standard. All the radial methods using a 6-A cutoff identified more interactions, but these putative contacts included many false positives as well as missed many false negatives. While radial cutoffs are quicker to calculate as well as to implement, this result highlights why radial cutoff methods do not have the proper resolution to detail the non-homogeneous packing within protein interfaces, and suggests an inappropriate bias in pair-wise contact potentials. Of the radial cutoff methods, using the closest atom approach exhibits the best approximation to the more intensive Voronoi calculation. Our version of the Voronoi polyhedra-based method QContacts is available at .     

Dissecting protein-protein recognition sites

The recognition sites in 70 pairwise protein-protein complexes of known three-dimensional structure are dissected in a set of surface patches by clustering atoms at the interface. When the interface buries <2000 A2 of protein surface, the recognition sites usually form a single patch on the surface of each component protein. In contrast, larger interfaces are generally multipatch, with at least one pair of patches that are equivalent in size to a single-patch interface. Each recognition site, or patch within a site, contains a core made of buried interface atoms, surrounded by a rim of atoms that remain accessible to solvent in the complex. A simple geometric model reproduces the number and distribution of atoms within a patch. The rim is similar in composition to the rest of the protein surface, but the core has a distinctive amino acid composition, which may help in identifying potential protein recognition sites on single proteins of known structures.     

Dissecting subunit interfaces in homodimeric proteins

The subunit interfaces of 122 homodimers of known three-dimensional structure are analyzed and dissected into sets of surface patches by clustering atoms at the interface; 70 interfaces are single-patch, the others have up to six patches, often contributed by different structural domains. The average interface buries 1,940 A2 of the surface of each monomer, contains one or two patches burying 600-1,600 A2, is 65% nonpolar and includes 18 hydrogen bonds. However, the range of size and of hydrophobicity is wide among the 122 interfaces. Each interface has a core made of residues with atoms buried in the dimer, surrounded by a rim of residues with atoms that remain accessible to solvent. The core, which constitutes 77% of the interface on average, has an amino acid composition that resembles the protein interior except for the presence of arginine residues, whereas the rim is more like the protein surface. These properties of the interfaces in homodimers, which are permanent assemblies, are compared to those of protein-protein complexes where the components associate after they have independently folded. On average, subunit interfaces in homodimers are twice larger than in complexes, and much less polar due to the large fraction belonging to the core, although the amino acid compositions of the cores are similar in the two types of interfaces.     

Prediction of interface residues in protein-protein complexes by a consensus neural network method: test against NMR data

The number of structures of protein-protein complexes deposited to the Protein Data Bank is growing rapidly. These structures embed important information for predicting structures of new protein complexes. This motivated us to develop the PPISP method for predicting interface residues in protein-protein complexes. In PPISP, sequence profiles and solvent accessibility of spatially neighboring surface residues were used as input to a neural network. The network was trained on native interface residues collected from the Protein Data Bank. The prediction accuracy at the time was 70% with 47% coverage of native interface residues. Now we have extensively improved PPISP. The training set now consisted of 1156 nonhomologous protein chains. Test on a set of 100 nonhomologous protein chains showed that the prediction accuracy is now increased to 80% with 51% coverage. To solve the problem of over-prediction and under-prediction associated with individual neural network models, we developed a consensus method that combines predictions from multiple models with different levels of accuracy and coverage. Applied on a benchmark set of 68 proteins for protein-protein docking, the consensus approach outperformed the best individual models by 3-8 percentage points in accuracy. To demonstrate the predictive power of cons-PPISP, eight complex-forming proteins with interfaces characterized by NMR were tested. These proteins are nonhomologous to the training set and have a total of 144 interface residues identified by chemical shift perturbation. cons-PPISP predicted 174 interface residues with 69% accuracy and 47% coverage and promises to complement experimental techniques in characterizing protein-protein interfaces. .     

Quantification of protein surfaces,volumes and atom-atom contacts using a constrained Voronoi procedure

MOTIVATION: Geometric representations of proteins and ligands, including atom volumes, atom-atom contacts and solvent accessible surfaces, can be used to characterize interactions between and within proteins, ligands and solvent. Voronoi algorithms permit quantification of these properties by dividing structures into cells with a one-to-one correspondence with constituent atoms. As there is no generally accepted measure of atom-atom contacts, a continuous analytical representation of inter-atomic contacts will be useful. Improved geometric algorithms will also be helpful in increasing the speed and accuracy of iterative modeling algorithms. RESULTS: We present computational methods based on the Voronoi procedure that provide rapid and exact solutions to solvent accessible surfaces, volumes, and atom contacts within macromolecules. Furthermore, we define a measure of atom-atom contact that is consistent with the calculation of solvent accessible surfaces, allowing the integration of solvent accessibility and inter-atomic contacts into a continuous measure. The speed and accuracy of the algorithm is compared to existing methods for calculating solvent accessible surfaces and volumes. The presented algorithm has a reduced execution time and greater accuracy compared to numerical and approximate analytical surface calculation algorithms, and a reduced execution time and similar accuracy to existing Voronoi procedures for calculating atomic surfaces and volumes.     

A "solvated rotamer" approach to modeling water-mediated hydrogen bonds at protein-protein interfaces

Water-mediated hydrogen bonds play critical roles at protein-protein and protein-nucleic acid interfaces, and the interactions formed by discrete water molecules cannot be captured using continuum solvent models. We describe a simple model for the energetics of water-mediated hydrogen bonds, and show that, together with knowledge of the positions of buried water molecules observed in X-ray crystal structures, the model improves the prediction of free-energy changes upon mutation at protein-protein interfaces, and the recovery of native amino acid sequences in protein interface design calculations. We then describe a "solvated rotamer" approach to efficiently predict the positions of water molecules, at protein-protein interfaces and in monomeric proteins, that is compatible with widely used rotamer-based side-chain packing and protein design algorithms. Finally, we examine the extent to which the predicted water molecules can be used to improve prediction of amino acid identities and protein-protein interface stability, and discuss avenues for overcoming current limitations of the approach.     

Protein surface mapping by chemical oxidation: structural analysis by mass spectrometry

The solvent-accessible surface area of proteins is important in biological function for many reasons, including protein-protein interactions, protein folding, and catalytic sites. Here we present a chemical technique to oxidize amino acid side chains in a model protein, apomyoglobin, and subsequent elucidation of the effect of solvent accessibility on the sites of oxidation. Under conditions of low protein oxidation (zero to three oxygen atoms added per apomyoglobin molecule), we have positively identified five oxidation sites by liquid chromatography-tandem mass spectrometry and high-resolution Fourier transform mass spectrometry. Our results indicate that all oxidized amino acids, with the exception of methionine, have highly solvent-accessible side chains, but the rate of oxidation may not be dictated solely by solvent accessibility and amino acid identity.     

Two different proteins that compete for binding to thrombin have opposite kinetic and thermodynamic profiles

Thrombin binds thrombomodulin (TM) at anion binding exosite 1, an allosteric site far from the thrombin active site. A monoclonal antibody (mAb) has been isolated that competes with TM for binding to thrombin. Complete binding kinetic and thermodynamic profiles for these two protein-protein interactions have been generated. Binding kinetics were measured by Biacore. Although both interactions have similar K(D)s, TM binding is rapid and reversible while binding of the mAb is slow and nearly irreversible. The enthalpic contribution to the DeltaG(bind) was measured by isothermal titration calorimetry and van't Hoff analysis. The contribution to the DeltaG(bind) from electrostatic steering was assessed from the dependence of the k(a) on ionic strength. Release of solvent H(2)O molecules from the interface was assessed by monitoring the decrease in amide solvent accessibility at the interface upon protein-protein binding. The mAb binding is enthalpy driven and has a slow k(d). TM binding appears to be entropy driven and has a fast k(a). The favorable entropy of the thrombin-TM interaction seems to be derived from electrostatic steering and a contribution from solvent release. The two interactions have remarkably different thermodynamic driving forces for competing reactions. The possibility that optimization of binding kinetics for a particular function may be reflected in different thermodynamic driving forces is discussed.     

Predicting where Small Molecules Bind at Protein-Protein Interfaces

Small molecules that bind at protein-protein interfaces may either block or stabilize protein-protein interactions in cells. Thus, some of these binding interfaces may turn into prospective targets for drug design. Here, we collected 175 pairs of protein-protein (PP) complexes and protein-ligand (PL) complexes with known three-dimensional structures for which (1) one protein from the PP complex shares at least 40% sequence identity with the protein from the PL complex, and (2) the interface regions of these proteins overlap at least partially with each other. We found that those residues of the interfaces that may bind the other protein as well as the small molecule are evolutionary more conserved on average, have a higher tendency of being located in pockets and expose a smaller fraction of their surface area to the solvent than the remaining protein-protein interface region. Based on these findings we derived a statistical classifier that predicts patches at binding interfaces that have a higher tendency to bind small molecules. We applied this new prediction method to more than 10 000 interfaces from the protein data bank. For several complexes related to apoptosis the predicted binding patches were in direct contact to co-crystallized small molecules.     

Cavities and Atomic Packing in Protein Structures and Interfaces

A comparative analysis of cavities enclosed in a tertiary structure of proteins and interfaces formed by the interaction of two protein subunits in obligate and non-obligate categories (represented by homodimeric molecules and heterocomplexes, respectively) is presented. The total volume of cavities increases with the size of the protein (or the interface), though the exact relationship may vary in different cases. Likewise, for individual cavities also there is quantitative dependence of the volume on the number of atoms (or residues) lining the cavity. The larger cavities tend to be less spherical, solvated, and the interfaces are enriched in these. On average 15 ų of cavity volume is found to accommodate single water, with another 40–45 ų needed for each additional solvent molecule. Polar atoms/residues have a higher propensity to line solvated cavities. Relative to the frequency of occurrence in the whole structure (or interface), residues in β-strands are found more often lining the cavities, and those in turn and loop the least. Any depression in one chain not complemented by a protrusion in the other results in a cavity in the protein–protein interface. Through the use of the Voronoi volume, the packing of residues involved in protein–protein interaction has been compared to that in the protein interior. For a comparable number of atoms the interface has about twice the number of cavities relative to the tertiary structure.     

Wet and dry interfaces: the role of solvent in protein-protein and protein-DNA recognition

Water molecules are found in abundance in protein-protein and protein-DNA interfaces. Although interface solvent molecules exchange quickly with the bulk solvent, structural and biochemical data suggest that water-mediated interactions are as important as direct hydrogen bonds in the stability and specificity of recognition.     

Statistical analysis and prediction of protein-protein interfaces

Predicting protein-protein interfaces from a three-dimensional structure is a key task of computational structural proteomics. In contrast to geometrically distinct small molecule binding sites, protein-protein interface are notoriously difficult to predict. We generated a large nonredundant data set of 1494 true protein-protein interfaces using biological symmetry annotation where necessary. The data set was carefully analyzed and a Support Vector Machine was trained on a combination of a new robust evolutionary conservation signal with the local surface properties to predict protein-protein interfaces. Fivefold cross validation verifies the high sensitivity and selectivity of the model. As much as 97% of the predicted patches had an overlap with the true interface patch while only 22% of the surface residues were included in an average predicted patch. The model allowed the identification of potential new interfaces and the correction of mislabeled oligomeric states.     

Rapid identification of protein-protein interfaces for the construction of a complex model based on multiple unassigned signals by using time-sharing NMR measurements

Protein-protein interactions are necessary for various cellular processes, and therefore, information related to protein-protein interactions and structural information of complexes is invaluable. To identify protein-protein interfaces using NMR, resonance assignments are generally necessary to analyze the data; however, they are time consuming to collect, especially for large proteins. In this paper, we present a rapid, effective, and unbiased approach for the identification of a protein-protein interface without resonance assignments. This approach requires only a single set of 2D titration experiments of a single protein sample, labeled with a unique combination of an (15)N-labeled amino acid and several amino acids (13)C-labeled on specific atoms. To rapidly obtain high resolution data, we applied a new pulse sequence for time-shared NMR measurements that allowed simultaneous detection of a ω(1)-TROSY-type backbone (1)H-(15)N and aromatic (1)H-(13)C shift correlations together with single quantum methyl (1)H-(13)C shift correlations. We developed a structure-based computational approach, that uses our experimental data to search the protein surfaces in an unbiased manner to identify the residues involved in the protein-protein interface. Finally, we demonstrated that the obtained information of the molecular interface could be directly leveraged to support protein-protein docking studies. Such rapid construction of a complex model provides valuable information and enables more efficient biochemical characterization of a protein-protein complex, for instance, as the first step in structure-guided drug development.     

Standard atomic volumes in double-stranded DNA and packing in protein--DNA interfaces

Standard volumes for atoms in double-stranded B-DNA are derived using high resolution crystal structures from the Nucleic Acid Database (NDB) and compared with corresponding values derived from crystal structures of small organic compounds in the Cambridge Structural Database (CSD). Two different methods are used to compute these volumes: the classical Voronoi method, which does not depend on the size of atoms, and the related Radical Planes method which does. Results show that atomic groups buried in the interior of double-stranded DNA are, on average, more tightly packed than in related small molecules in the CSD. The packing efficiency of DNA atoms at the interfaces of 25 high resolution protein-DNA complexes is determined by computing the ratios between the volumes of interfacial DNA atoms and the corresponding standard volumes. These ratios are found to be close to unity, indicating that the DNA atoms at protein-DNA interfaces are as closely packed as in crystals of B-DNA. Analogous volume ratios, computed for buried protein atoms, are also near unity, confirming our earlier conclusions that the packing efficiency of these atoms is similar to that in the protein interior. In addition, we examine the number, volume and solvent occupation of cavities located at the protein-DNA interfaces and compared them with those in the protein interior. Cavities are found to be ubiquitous in the interfaces as well as inside the protein moieties. The frequency of solvent occupation of cavities is however higher in the interfaces, indicating that those are more hydrated than protein interiors. Lastly, we compare our results with those obtained using two different measures of shape complementarity of the analysed interfaces, and find that the correlation between our volume ratios and these measures, as well as between the measures themselves, is weak. Our results indicate that a tightly packed environment made up of DNA, protein and solvent atoms plays a significant role in protein-DNA recognition.     

A dissection of the protein-protein interfaces in icosahedral virus capsids

We selected 49 icosahedral virus capsids whose crystal structures are reported in the Protein Data Bank. They belong to the T=1, T=3, pseudo T=3 and other lattice types. We identified in them 779 unique interfaces between pairs of subunits, all repeated by icosahedral symmetry. We analyzed the geometric and physical chemical properties of these interfaces and compared with interfaces in protein-protein complexes and homodimeric proteins, and with crystal packing contacts. The capsids contain one to 16 subunits implicated in three to 66 unique interfaces. Each subunit loses 40-60% of its accessible surface in contacts with an average of 8.5 neighbors. Many of the interfaces are very large with a buried surface area (BSA) that can exceed 10,000 A(2), yet 39% are small with a BSA<800 A(2) comparable to crystal packing contacts. Pairwise capsid interfaces overlap, so that one-third of the residues are part of more than one interface. Those with a BSA>800 A(2) resemble homodimer interfaces in their chemical composition. Relative to the protein surface, they are non-polar, enriched in aliphatic residues and depleted of charged residues, but not of neutral polar residues. They contain one H-bond per about 200 A(2) BSA. Small capsid interfaces (BSA<800 A(2)) are only slightly more polar. They have a similar amino acid composition, but they bury fewer atoms and contain fewer H-bonds for their size. Geometric parameters that estimate the quality of the atomic packing suggest that the small capsid interfaces are loosely packed like crystal packing contacts, whereas the larger interfaces are close-packed as in protein-protein complexes and homodimers. We discuss implications of these findings on the mechanism of capsid assembly, assuming that the larger interfaces form first to yield stable oligomeric species (capsomeres), and that medium-size interfaces allow the stepwise addition of capsomeres to build larger intermediates.     

Membrane model for the G-protein-coupled receptor rhodopsin: hydrophobic interface and dynamical structure

Rhodopsin is the only member of the pharmacologically important superfamily of G-protein-coupled receptors with a known structure at atomic resolution. A molecular dynamics model of rhodopsin in a POPC phospholipid bilayer was simulated for 15 ns, revealing a conformation significantly different from the recent crystal structures. The structure of the bilayer compared with a protein-free POPC control indicated hydrophobic matching with the nonpolar interface of the receptor, in agreement with deuterium NMR experiments. A new generalized molecular surface method, based on a three-dimensional Voronoi cell construction for atoms with different radii, was developed to quantify cross-sectional area profiles for the protein, lipid acyl chains and headgroups, and water. Thus, it was possible to investigate the bilayer deformation due to curvature of the individual lipid monolayers. Moreover, the generalized molecular surface derived hydrophobic interface allowed benchmarking of the hydropathy sequence analysis, an important structural genomics tool. Five water molecules diffused into internal hydration sites during the simulation, yielding a total of 12 internal waters. The cytoplasmic loops and the C-terminal tail, containing the G-protein recognition and protein sorting sequences, exhibited a high mobility, in marked contrast to the extracellular and transmembrane domains. The proposed functional coupling of the highly conserved ERY motif to the lipid-water interface via the cytoplasmic loops provides insight into lipid effects on G-protein-coupled receptor activation in terms of a flexible surface model, involving the spontaneous monolayer curvature.     

A new, structurally nonredundant, diverse data set of protein-protein interfaces and its implications

Here, we present a diverse, structurally nonredundant data set of two-chain protein-protein interfaces derived from the PDB. Using a sequence order-independent structural comparison algorithm and hierarchical clustering, 3799 interface clusters are obtained. These yield 103 clusters with at least five nonhomologous members. We divide the clusters into three types. In Type I clusters, the global structures of the chains from which the interfaces are derived are also similar. This cluster type is expected because, in general, related proteins associate in similar ways. In Type II, the interfaces are similar; however, remarkably, the overall structures and functions of the chains are different. The functional spectrum is broad, from enzymes/inhibitors to immunoglobulins and toxins. The fact that structurally different monomers associate in similar ways, suggests "good" binding architectures. This observation extends a paradigm in protein science: It has been well known that proteins with similar structures may have different functions. Here, we show that it extends to interfaces. In Type III clusters, only one side of the interface is similar across the cluster. This structurally nonredundant data set provides rich data for studies of protein-protein interactions and recognition, cellular networks and drug design. In particular, it may be useful in addressing the difficult question of what are the favorable ways for proteins to interact. (The data set is available at http://protein3d.ncifcrf.gov/~keskino/ and http://home.ku.edu.tr/~okeskin/INTERFACE/INTERFACES.html.)