A novel and efficient tool for locating and characterizing protein cavities and binding sites

Systematic investigation of a protein and its binding site characteristics are crucial for designing small molecules that modulate protein functions. However, fundamental uncertainties in binding site interactions and insufficient knowledge of the properties of even well‐defined binding pockets can make it difficult to design optimal drugs. Herein, we report the development and implementation of a cavity detection algorithm built with HINT toolkit functions that we are naming Vectorial Identification of Cavity Extents (VICE). This very efficient algorithm is based on geometric criteria applied to simple integer grid maps. In testing, we carried out a systematic investigation on a very diverse data set of proteins and protein–protein/protein–polynucleotide complexes for locating and characterizing the indentations, cavities, pockets, grooves, channels, and surface regions. Additionally, we evaluated a curated data set of unbound proteins for which a ligand‐bound protein structures are also known; here the VICE algorithm located the actual ligand in the largest cavity in 83% of the cases and in one of the three largest in 90% of the cases. An interactive front‐end provides a quick and simple procedure for locating, displaying and manipulating cavities in these structures. Information describing the cavity, including its volume and surface area metrics, and lists of atoms, residues, and/or chains lining the binding pocket, can be easily obtained and analyzed. For example, the relative cross‐sectional surface area (to total surface area) of cavity openings in well‐enclosed cavities is 0.06 ± 0.04 and in surface clefts or crevices is 0.25 ± 0.09. Proteins 2010. © 2009 Wiley‐Liss, Inc.     

Transient pockets on XIAP-BIR2: toward the characterization of putative binding sites of small-molecule XIAP inhibitors

Protein-protein interactions are abundant in signal transduction pathways and thus of crucial importance in the regulation of apoptosis. However, designing small-molecule inhibitors for these potential drug targets is very challenging as such proteins often lack well-defined binding pockets. An example for such an interaction is the binding of the anti-apoptotic BIR2 domain of XIAP to the pro-apoptotic caspase-3 that results in the survival of damaged cells. Although small-molecule inhibitors of this interaction have been identified, their exact binding sites on XIAP are not known as its crystal structures reveal no suitable pockets. Here, we apply our previously developed protocol for identifying transient binding pockets to XIAP-BIR2. Transient pockets were identified in snapshots taken during four different molecular dynamics simulations that started from the caspase-3:BIR2 complex or from the unbound BIR2 structure and used water or methanol as solvent. Clustering of these pockets revealed that surprisingly many pockets opened in the flexible linker region that is involved in caspase-3 binding. We docked three known inhibitors into these transient pockets and so determined five putative binding sites. In addition, by docking two inactive compounds of the same series, we show that this protocol is also able to distinguish between binders and nonbinders which was not possible when docking to the crystal structures. These findings represent a first step toward the understanding of the binding of small-molecule XIAP-BIR2 inhibitors on a molecular level and further highlight the importance of considering protein flexibility when designing small-molecule protein-protein interaction inhibitors.     

Comparative surface geometry of the protein kinase family

Identifying conserved pockets on the surfaces of a family of proteins can provide insight into conserved geometric features and sites of protein–protein interaction. Here we describe mapping and comparison of the surfaces of aligned crystallographic structures, using the protein kinase family as a model. Pockets are rapidly computed using two computer programs, FADE and Crevasse. FADE uses gradients of atomic density to locate grooves and pockets on the molecular surface. Crevasse, a new piece of software, splits the FADE output into distinct pockets. The computation was run on 10 kinase catalytic cores aligned on the αF‐helix, and the resulting pockets spatially clustered. The active site cleft appears as a large, contiguous site that can be subdivided into nucleotide and substrate docking sites. Substrate specificity determinants in the active site cleft between serine/threonine and tyrosine kinases are visible and distinct. The active site clefts cluster tightly, showing a conserved spatial relationship between the active site and αF‐helix in the C‐lobe. When the αC‐helix is examined, there are multiple mechanisms for anchoring the helix using spatially conserved docking sites. A novel site at the top of the N‐lobe is present in all the kinases, and there is a large conserved pocket over the hinge and the αC‐β4 loop. Other pockets on the kinase core are strongly conserved but have not yet been mapped to a protein–protein interaction. Sites identified by this algorithm have revealed structural and spatially conserved features of the kinase family and potential conserved intermolecular and intramolecular binding sites.     

Optimal docking area: a new method for predicting protein-protein interaction sites

Understanding energetics and mechanism of protein-protein association remains one of the biggest theoretical problems in structural biology. It is assumed that desolvation must play an essential role during the association process, and indeed protein-protein interfaces in obligate complexes have been found to be highly hydrophobic. However, the identification of protein interaction sites from surface analysis of proteins involved in non-obligate protein-protein complexes is more challenging. Here we present Optimal Docking Area (ODA), a new fast and accurate method of analyzing a protein surface in search of areas with favorable energy change when buried upon protein-protein association. The method identifies continuous surface patches with optimal docking desolvation energy based on atomic solvation parameters adjusted for protein-protein docking. The procedure has been validated on the unbound structures of a total of 66 non-homologous proteins involved in non-obligate protein-protein hetero-complexes of known structure. Optimal docking areas with significant low-docking surface energy were found in around half of the proteins. The 'ODA hot spots' detected in X-ray unbound structures were correctly located in the known protein-protein binding sites in 80% of the cases. The role of these low-surface-energy areas during complex formation is discussed. Burial of these regions during protein-protein association may favor the complexed configurations with near-native interfaces but otherwise arbitrary orientations, thus driving the formation of an encounter complex. The patch prediction procedure is freely accessible at http://www.molsoft.com/oda and can be easily scaled up for predictions in structural proteomics.     

Anatomy of protein pockets and cavities: measurement of binding site geometry and implications for ligand design.

Identification and size characterization of surface pockets and occluded cavities are initial steps in protein structure-based ligand design. A new program, CAST, for automatically locating and measuring protein pockets and cavities, is based on precise computational geometry methods, including alpha shape and discrete flow theory. CAST identifies and measures pockets and pocket mouth openings, as well as cavities. The program specifies the atoms lining pockets, pocket openings, and buried cavities; the volume and area of pockets and cavities; and the area and circumference of mouth openings. CAST analysis of over 100 proteins has been carried out; proteins examined include a set of 51 monomeric enzyme-ligand structures, several elastase-inhibitor complexes, the FK506 binding protein, 30 HIV-1 protease-inhibitor complexes, and a number of small and large protein inhibitors. Medium-sized globular proteins typically have 10-20 pockets/cavities. Most often, binding sites are pockets with 1-2 mouth openings; much less frequently they are cavities. Ligand binding pockets vary widely in size, most within the range 10(2)-10(3)A3. Statistical analysis reveals that the number of pockets and cavities is correlated with protein size, but there is no correlation between the size of the protein and the size of binding sites. Most frequently, the largest pocket/cavity is the active site, but there are a number of instructive exceptions. Ligand volume and binding site volume are somewhat correlated when binding site volume is < or =700 A3, but the ligand seldom occupies the entire site. Auxiliary pockets near the active site have been suggested as additional binding surface for designed ligands (Mattos C et al., 1994, Nat Struct Biol 1:55-58). Analysis of elastase-inhibitor complexes suggests that CAST can identify ancillary pockets suitable for recruitment in ligand design strategies. Analysis of the FK506 binding protein, and of compounds developed in SAR by NMR (Shuker SB et al., 1996, Science 274:1531-1534), indicates that CAST pocket computation may provide a priori identification of target proteins for linked-fragment design. CAST analysis of 30 HIV-1 protease-inhibitor complexes shows that the flexible active site pocket can vary over a range of 853-1,566 A3, and that there are two pockets near or adjoining the active site that may be recruited for ligand design.     

Identification of protein-protein interaction sites from docking energy landscapes

Protein recognition is one of the most challenging and intriguing problems in structural biology. Despite all the available structural, sequence and biophysical information about protein-protein complexes, the physico-chemical patterns, if any, that make a protein surface likely to be involved in protein-protein interactions, remain elusive. Here, we apply protein docking simulations and analysis of the interaction energy landscapes to identify protein-protein interaction sites. The new protocol for global docking based on multi-start global energy optimization of an all-atom model of the ligand, with detailed receptor potentials and atomic solvation parameters optimized in a training set of 24 complexes, explores the conformational space around the whole receptor without restrictions. The ensembles of the rigid-body docking solutions generated by the simulations were subsequently used to project the docking energy landscapes onto the protein surfaces. We found that highly populated low-energy regions consistently corresponded to actual binding sites. The procedure was validated on a test set of 21 known protein-protein complexes not used in the training set. As much as 81% of the predicted high-propensity patch residues were located correctly in the native interfaces. This approach can guide the design of mutations on the surfaces of proteins, provide geometrical details of a possible interaction, and help to annotate protein surfaces in structural proteomics.     

Scoring optimisation of unbound protein-protein docking including protein binding site predictions

The prediction of the structure of the protein-protein complex is of great importance to better understand molecular recognition processes. During systematic protein-protein docking, the surface of a protein molecule is scanned for putative binding sites of a partner protein. The possibility to include external data based on either experiments or bioinformatic predictions on putative binding sites during docking has been systematically explored. The external data were included during docking with a coarse-grained protein model and on the basis of force field weights to bias the docking search towards a predicted or known binding region. The approach was tested on a large set of protein partners in unbound conformations. The significant improvement of the docking performance was found if reliable data on the native binding sites were available. This was possible even if data for single key amino acids at a binding interface are included. In case of binding site predictions with limited accuracy, only modest improvement compared with unbiased docking was found. The optimisation of the protocol to bias the search towards predicted binding sites was found to further improve the docking performance resulting in approximately 40% acceptable solutions within the top 10 docking predictions compared with 22% in case of unbiased docking of unbound protein structures.     

CO migration pathways in cytochrome P450cam studied by molecular dynamics simulations

Previous laser flash photolysis investigations between 100 and 300 K have shown that the kinetics of CO rebinding with cytochrome P450(cam)(camphor) consist of up to four different processes revealing a complex internal dynamics after ligand dissociation. In the present work, molecular dynamics simulations were undertaken on the ternary complex P450(cam)(cam)(CO) to explore the CO migration pathways, monitor the internal cavities of the protein, and localize the CO docking sites. One trajectory of 1 nsec with the protein in a water box and 36 trajectories of 1 nsec in the vacuum were calculated. In each trajectory, the protein contained only one CO ligand on which no constraints were applied. The simulations were performed at 200, 300, and 320 K. The results indicate the presence of seven CO docking sites, mainly hydrophobic, located in the same moiety of the protein. Two of them coincide with xenon binding sites identified by crystallography. The protein matrix exhibits eight persistent internal cavities, four of which corresponding to the ligand docking sites. In addition, it was observed that water molecules entering the protein were mainly attracted into the polar pockets, far away from the CO docking sites. Finally, the identified CO migration pathways provide a consistent interpretation of the experimental rebinding kinetics.     

Protein-protein interactions: hot spots and structurally conserved residues often locate in complemented pockets that pre-organized in the unbound states: implications for docking

Energetic hot spots account for a significant portion of the total binding free energy and correlate with structurally conserved interface residues. Here, we map experimentally determined hot spots and structurally conserved residues to investigate their geometrical organization. Unfilled pockets are pockets that remain unfilled after protein-protein complexation, while complemented pockets are pockets that disappear upon binding, representing tightly fit regions. We find that structurally conserved residues and energetic hot spots are strongly favored to be located in complemented pockets, and are disfavored in unfilled pockets. For the three available protein-protein complexes with complemented pockets where both members of the complex were alanine-scanned, 62% of all hot spots (DeltaDeltaG>2kcal/mol) are within these pockets, and 60% of the residues in the complemented pockets are hot spots. 93% of all red-hot residues (DeltaDeltaG>/=4kcal/mol) either protrude into or are located in complemented pockets. The occurrence of hot spots and conserved residues in complemented pockets highlights the role of local tight packing in protein associations, and rationalizes their energetic contribution and conservation. Complemented pockets and their corresponding protruding residues emerge among the most important geometric features in protein-protein interactions. By screening the solvent, this organization shields backbone hydrogen bonds and charge-charge interactions. Complemented pockets often pre-exist binding. For 18 protein-protein complexes with complemented pockets whose unbound structures are available, in 16 the pockets are identified to pre-exist in the unbound structures. The root-mean-squared deviations of the atoms lining the pockets between the bound and unbound states is as small as 0.9A, suggesting that such pockets constitute features of the populated native state that may be used in docking.     

Pocketome via comprehensive identification and classification of ligand binding envelopes

We developed a new computational algorithm for the accurate identification of ligand binding envelopes rather than surface binding sites. We performed a large scale classification of the identified envelopes according to their shape and physicochemical properties. The predicting algorithm, called PocketFinder, uses a transformation of the Lennard-Jones potential calculated from a three-dimensional protein structure and does not require any knowledge about a potential ligand molecule. We validated this algorithm using two systematically collected data sets of ligand binding pockets from complexed (bound) and uncomplexed (apo) structures from the Protein Data Bank, 5616 and 11,510, respectively. As many as 96.8% of experimental binding sites were predicted at better than 50% overlap level. Furthermore 95.0% of the asserted sites from the apo receptors were predicted at the same level. We demonstrate that conformational differences between the apo and bound pockets do not dramatically affect the prediction results. The algorithm can be used to predict ligand binding pockets of uncharacterized protein structures, suggest new allosteric pockets, evaluate feasibility of protein-protein interaction inhibition, and prioritize molecular targets. Finally the data base of the known and predicted binding pockets for the human proteome structures, the human pocketome, was collected and classified. The pocketome can be used for rapid evaluation of possible binding partners of a given chemical compound.     

Protein-protein docking by simulating the process of association subject to biochemical constraints

We present a computational procedure for modeling protein-protein association and predicting the structures of protein-protein complexes. The initial sampling stage is based on an efficient Brownian dynamics algorithm that mimics the physical process of diffusional association. Relevant biochemical data can be directly incorporated as distance constraints at this stage. The docked configurations are then grouped with a hierarchical clustering algorithm into ensembles that represent potential protein-protein encounter complexes. Flexible refinement of selected representative structures is done by molecular dynamics simulation. The protein-protein docking procedure was thoroughly tested on 10 structurally and functionally diverse protein-protein complexes. Starting from X-ray crystal structures of the unbound proteins, in 9 out of 10 cases it yields structures of protein-protein complexes close to those determined experimentally with the percentage of correct contacts >30% and interface backbone RMSD <4 A. Detailed examination of all the docking cases gives insights into important determinants of the performance of the computational approach in modeling protein-protein association and predicting of protein-protein complex structures.     

Role of flexibility and polarity as determinants of the hydration of internal cavities and pockets in proteins

Molecular dynamics simulations of Staphylococcal nuclease and of 10 variants with internal polar or ionizable groups were performed to investigate systematically the molecular determinants of hydration of internal cavities and pockets in proteins. In contrast to apolar cavities in rigid carbon structures, such as nanotubes or buckeyballs, internal cavities in proteins that are large enough to house a few water molecules will most likely be dehydrated unless they contain a source of polarity. The water content in the protein interior can be modulated by the flexibility of protein elements that interact with water, which can impart positional disorder to water molecules, or bias the pattern of internal hydration that is stabilized. This might explain differences in the patterns of hydration observed in crystal structures obtained at cryogenic and room temperature conditions. The ability of molecular dynamics simulations to determine the most likely sites of water binding in internal pockets and cavities depends on its efficiency in sampling the hydration of internal sites and alternative protein and water conformations. This can be enhanced significantly by performing multiple molecular dynamics simulations as well as simulations started from different initial hydration states.     

Identification of a small molecule that modulates platelet glycoprotein Ib-von Willebrand factor interaction

The von Willebrand factor (VWF) A1-glycoprotein (GP) Ibα interaction is of major importance during thrombosis mainly at sites of high shear stress. Inhibitors of this interaction prevent platelet-dependent thrombus formation in vivo, without major bleeding complications. However, the size and/or protein nature of the inhibitors currently in development limit oral bioavailability and clinical development. We therefore aimed to search for a small molecule protein-protein interaction inhibitor interfering with the VWF-GPIbα binding. After determination of putative small molecule binding pockets on the surface of VWF-A1 and GPIbα using site-finding algorithms and molecular dynamics, high throughput molecular docking was performed on both binding partners. A selection of compounds showing good in silico docking scores into the predicted pockets was retained for testing their in vitro effect on VWF-GPIbα complex formation, by which we identified a compound that surprisingly stimulated the VWF-GPIbα binding in a ristocetin cofactor ELISA and increased platelet adhesion in whole blood to collagen under arterial shear rate but in contrast inhibited ristocetin-induced platelet aggregation. The selected compound adhering to the predicted binding partner GPIbα could be confirmed by saturation transfer difference NMR spectroscopy. We thus clearly identified a small molecule that modulates VWF-GPIbα binding and that will now serve as a starting point for further studies and chemical modifications to fully characterize the interaction and to manipulate specific activity of the compound.     

Residue centrality, functionally important residues, and active site shape: analysis of enzyme and non-enzyme families

The representation of protein structures as small-world networks facilitates the search for topological determinants, which may relate to functionally important residues. Here, we aimed to investigate the performance of residue centrality, viewed as a family fold characteristic, in identifying functionally important residues in protein families. Our study is based on 46 families, including 29 enzyme and 17 non-enzyme families. A total of 80% of these central positions corresponded to active site residues or residues in direct contact with these sites. For enzyme families, this percentage increased to 91%, while for non-enzyme families the percentage decreased substantially to 48%. A total of 70% of these central positions are located in catalytic sites in the enzyme families, 64% are in hetero-atom binding sites in those families binding hetero-atoms, and only 16% belong to protein-protein interfaces in families with protein-protein interaction data. These differences reflect the active site shape: enzyme active sites locate in surface clefts, hetero-atom binding residues are in deep cavities, while protein-protein interactions involve a more planar configuration. On the other hand, not all surface cavities or clefts are comprised of central residues. Thus, closeness centrality identifies functionally important residues in enzymes. While here we focus on binding sites, we expect to identify key residues for the integration and transmission of the information to the rest of the protein, reflecting the relationship between fold and function. Residue centrality is more conserved than the protein sequence, emphasizing the robustness of protein structures.     

Travel depth, a new shape descriptor for macromolecules: application to ligand binding

Depth is a term frequently applied to the shape and surface of macromolecules, describing for example the grooves in DNA, the shape of an enzyme active site, or the binding site for a small molecule in a protein. Yet depth is a difficult property to define rigorously in a macromolecule, and few computational tools exist to quantify this notion, to visualize it, or analyze the results. We present our notion of travel depth, simply put the physical distance a solvent molecule would have to travel from a surface point to a suitably defined reference surface. To define the reference surface, we use the limiting form of the molecular surface with increasing probe size: the convex hull. We then present a fast, robust approximation algorithm to compute travel depth to every surface point. The travel depth is useful because it works for pockets of any size and complexity. It also works for two interesting special cases. First, it works on the grooves in DNA, which are unbounded in one direction. Second, it works on the case of tunnels, that is pockets that have no "bottom", but go through the entire macromolecule. Our algorithm makes it straightforward to quantify discussions of depth when analyzing structures. High-throughput analysis of macromolecule depth is also enabled by our algorithm. This is demonstrated by analyzing a database of protein-small molecule binding pockets, and the distribution of bound magnesium ions in RNA structures. These analyses show significant, but subtle effects of depth on ligand binding localization and strength.     

Examination of shape complementarity in docking of unbound proteins.

Here we carry out an examination of shape complementarity as a criterion in protein-protein docking and binding. Specifically, we examine the quality of shape complementarity as a critical determinant not only in the docking of 26 protein-protein "bound" complexed cases, but in particular, of 19 "unbound" protein-protein cases, where the structures have been determined separately. In all cases, entire molecular surfaces are utilized in the docking, with no consideration of the location of the active site, or of particular residues/atoms in either the receptor or the ligand that participate in the binding. To evaluate the goodness of the strictly geometry-based shape complementarity in the docking process as compared to the main favorable and unfavorable energy components, we study systematically a potential correlation between each of these components and the root mean square deviation (RMSD) of the "unbound" protein-protein cases. Specifically, we examine the non-polar buried surface area, polar buried surface area, buried surface area relating to groups bearing unsatisfied buried charges, and the number of hydrogen bonds in all docked protein-protein interfaces. For these cases, where the two proteins have been crystallized separately, and where entire molecular surfaces are considered without a predefinition of the binding site, no correlation is observed. None of these parameters appears to consistently improve on shape complementarity in the docking of unbound molecules. These findings argue that simplicity in the docking process, utilizing geometrical shape criteria may capture many of the essential features in protein-protein docking. In particular, they further reinforce the long held notion of the importance of molecular surface shape complementarity in the binding, and hence in docking. This is particularly interesting in light of the fact that the structures of the docked pairs have been determined separately, allowing side chains on the surface of the proteins to move relatively freely. This study has been enabled by our efficient, computer vision-based docking algorithms. The fast CPU matching times, on the order of minutes on a PC, allow such large-scale docking experiments of large molecules, which may not be feasible by other techniques. Proteins 1999;36:307-317.     

Automated docking of peptides and proteins by using a genetic algorithm combined with a tabu search.

A genetic algorithm (GA) combined with a tabu search (TA) has been applied as a minimization method to rake the appropriate associated sites for some biomolecular systems. In our docking procedure, surface complementarity and energetic complementarity of a ligand with its receptor have been considered separately in a two-stage docking method. The first stage was to find a set of potential associated sites mainly based on surface complementarity using a genetic algorithm combined with a tabu search. This step corresponds with the process of finding the potential binding sites where pharmacophores will bind. In the second stage, several hundreds of GA minimization steps were performed for each associated site derived from the first stage mainly based on the energetic complementarity. After calculations for both of the two stages, we can offer several solutions of associated sites for every complex. In this paper, seven biomolecular systems, including five bound complexes and two unbound complexes, were chosen from the Protein Data Bank (PDB) to test our method. The calculated results were very encouraging-the hybrid minimization algorithm successfully reaches the correct solutions near the best binded modes for these protein complexes. The docking results not only predict the bound complexes very well, but also get a relatively accurate complexed conformation for unbound systems. For the five bound complexes, the results show that surface complementarity is enough to find the precise binding modes, the top solution from the tabu list generally corresponds to the correct binding mode. For the two unbound complexes, due to the conformational changes upon binding, it seems more difficult to get their correct binding conformations. The predicted results show that the correct binding mode also corresponds to a relatively large surface complementarity score. In these two test cases, the correct solution can be found in the top several solutions from the tabu list. For unbound complexes, the interaction energy from energetic complementarity is very important, it can be used to filter these solutions from the surface complementarity. After the evaluation of the energetic complementarity, the conformations and orientations close to the crystallographically determined structures are resolved. In most cases, the smallest root mean square distance (r.m.s.d.) from the GA combined with TA solutions is in a relatively small region. Our program of automatic docking is really a universal one among the procedures used for the theoretical study of molecular recognition.