A tool for calculating binding-site residues on proteins from PDB structures

Background  

In the research on protein functional sites, researchers often need to identify binding-site residues on a protein. A commonly used strategy is to find a complex structure from the Protein Data Bank (PDB) that consists of the protein of interest and its interacting partner(s) and calculate binding-site residues based on the complex structure. However, since a protein may participate in multiple interactions, the binding-site residues calculated based on one complex structure usually do not reveal all binding sites on a protein. Thus, this requires researchers to find all PDB complexes that contain the protein of interest and combine the binding-site information gleaned from them. This process is very time-consuming. Especially, combing binding-site information obtained from different PDB structures requires tedious work to align protein sequences. The process becomes overwhelmingly difficult when researchers have a large set of proteins to analyze, which is usually the case in practice.     

Lactose-containing starburst dendrimers: influence of dendrimer generation and binding-site orientation of receptors (plant/animal lectins and immunoglobulins) on binding properties.

Starburst glycodendrimers offer the potential to serve as high-affinity ligands for clinically relevant sugar receptors. In order to define areas of application, their binding behavior towards sugar receptors with differential binding-site orientation but identical monosaccharide specificity must be evaluated. Using poly(amidoamine) starburst dendrimers of five generations, which contain the p-isothiocyanato derivative of p-aminophenyl-beta-D-lactoside as ligand group, four different types of galactoside-binding proteins were chosen for this purpose, i.e., the (AB)(2)-toxic agglutinin from mistletoe, a human immunoglobulin G fraction, the homodimeric galectin-1 with its two binding sites at opposite ends of the jelly-roll-motif-harboring protein and monomeric galectin-3. Direct solid-phase assays with surface-immobilized glycodendrimers resulted in obvious affinity enhancements by progressive core branching for the plant agglutinin and less pronounced for the antibody and galectin-1. High density of binding of galectin-3 with modest affinity increases only from the level of the 32-mer onwards points to favorable protein-protein interactions of the monomeric lectin and a spherical display of the end groups without a major share of backfolding. When the inhibitory potency of these probes was evaluated as competitor of receptor binding to an immobilized neoglycoprotein or to asialofetuin, a marked selectivity was detected. The 32- and 64-mers were second to none as inhibitors for the plant agglutinin against both ligand-exposing matrices and for galectin-1 on the matrix with a heterogeneous array of interglycoside distances even on the per-sugar basis. In contrast, a neoglycoprotein with the same end group was superior in the case of the antibody and, less pronounced, monomeric galectin-3. Intimate details of topological binding-site presentation and the ligand display on different generations of core assembly are major operative factors which determine the potential of dendrimers for applications as lectin-targeting device, as attested by these observations.     

Evolving specificity from variability for protein interaction domains

An important question in modular domain-peptide interactions, which play crucial roles in many biological processes, is how the diverse specificities exhibited by different members of a domain family are encoded in a common scaffold. Analysis of the Src homology (SH) 2 family has revealed that its specificity is determined, in large part, by the configuration of surface loops that regulate ligand access to binding pockets. In a distinct manner, SH3 domains employ loops for ligand recognition. The PDZ domain, in contrast, achieves specificity by co-evolution of binding-site residues. Thus, the conformational and sequence variability afforded by surface loops and binding sites provides a general mechanism by which to encode the wide spectrum of specificities observed for modular protein interaction domains.     

Exploring the Composition of Protein-Ligand Binding Sites on a Large Scale

The residue composition of a ligand binding site determines the interactions available for diffusion-mediated ligand binding, and understanding general composition of these sites is of great importance if we are to gain insight into the functional diversity of the proteome. Many structure-based drug design methods utilize such heuristic information for improving prediction or characterization of ligand-binding sites in proteins of unknown function. The Binding MOAD database if one of the largest curated sets of protein-ligand complexes, and provides a source of diverse, high-quality data for establishing general trends of residue composition from currently available protein structures. We present an analysis of 3,295 non-redundant proteins with 9,114 non-redundant binding sites to identify residues over-represented in binding regions versus the rest of the protein surface. The Binding MOAD database delineates biologically-relevant “valid” ligands from “invalid” small-molecule ligands bound to the protein. Invalids are present in the crystallization medium and serve no known biological function. Contacts are found to differ between these classes of ligands, indicating that residue composition of biologically relevant binding sites is distinct not only from the rest of the protein surface, but also from surface regions capable of opportunistic binding of non-functional small molecules. To confirm these trends, we perform a rigorous analysis of the variation of residue propensity with respect to the size of the dataset and the content bias inherent in structure sets obtained from a large protein structure database. The optimal size of the dataset for establishing general trends of residue propensities, as well as strategies for assessing the significance of such trends, are suggested for future studies of binding-site composition.     

14-3-3 protein C-terminal stretch occupies ligand binding groove and is displaced by phosphopeptide binding

14-3-3 proteins are important regulators of numerous cellular signaling circuits. They bind to phosphorylated protein ligands and regulate their functions by a number of different mechanisms. The C-terminal part of the 14-3-3 protein is known to be involved in the regulation of 14-3-3 binding properties. The structure of this region is unknown; however, a possible location of the C-terminal stretch within the ligand binding groove of the 14-3-3 protein has been suggested. To fully understand the role of the C-terminal stretch in the regulation of the 14-3-3 protein binding properties, we investigated the physical location of the C-terminal stretch and its changes upon the ligand binding. For this purpose, we have used Forster resonance energy transfer (FRET) measurements and molecular dynamics simulation. FRET measurements between Trp242 located at the end of the C-terminal stretch and a dansyl group attached at two different cysteine residues (Cys25 or Cys189) indicated that in the absence of the ligand, the C-terminal stretch occupies the ligand binding groove of 14-3-3 protein. Our data also showed that phosphopeptide binding displaces the C-terminal stretch from the ligand binding groove. Intramolecular distances calculated from FRET measurements fit well with distances obtained from molecular dynamics simulation of full-length 14-3-3zeta protein.     

A consensus-binding structure for adenine at the atomic level permits searching for the ligand site in a wide spectrum of adenine-containing complexes

Attempts to derive structural features of ligand-binding sites have traditionally involved seeking commonalities at the residue level. Recently, structural studies have turned to atomic interactions of small molecular fragments to extract common binding-site properties. Here, we explore the use of larger ligand elements to derive a consensus binding structure for the ligand as a whole. We superimposed multiple molecular structures from a nonredundant set of adenosine-5'-triphosphate (ATP) protein complexes, using the adenine moiety as template. Clustered binding-site atoms of compatible atomic classes forming attractive contacts with the adenine probe were extracted. A set of atomic clusters characterizing the adenine binding pocket was then derived. Among the clusters are three vertices representing the interactions of adenine atom N6 with its protein-binding niche. These vertices, together with atom C6 of the purine ring system, complete the set of four vertices for the pyramid-like structure of the N6 anchor atom. Also, the sequence relationship for the adenine-binding loop interacting with the C2-N6 end of the conjugated ring system is expanded to include a third hydrophilic cluster interacting with atom N1. A search procedure involving interatomic distances between cluster centers was formulated and applied to seek putative binding sites in test cases. The results show that a consensus network of clusters, based on an adenine probe and an ATP-complexed training set of proteins, is sufficient to recognize the experimental cavity for adenine in a wide spectrum of ligand-protein complexes.     

The relationship between metal-metal distance of two metal ions chelated complex and RNA cleavage activity.

We prepared a series of ligands possessing two binding sites for metal coordination: in each ligand molecule, two binding sites with the same functionality (2,2'-dipicolylamino group) were placed at the of various methyl arenes. Thus, the distances between the metal binding sites were different from ligand to ligand. We examined the rate of the hydrolysis of RNA dimer catalyzed by La3+ ion binuclear complexes of the ligands. The catalytic activity of the binuclear complexes increased as the distance between the metal binding sites was decreased.     

Multisite proteins and randomly coiled polymeric ligands. Chain length dependence of binding constants.

A model mechanism was developed for the binding of a rigid multisite protein with a randomly coiled multivalent ligand. Probabilities of the formation of chain loops between sites located at given distances at the protein were calculated by an extension of the concept of ring closure in coiled chain molecules. Expressions were derived for the dependence of overall equilibrium quantities, such as the binding constant between the protein and the ligand, on intrinsic parameters such as intrinsic binding constants, number of sites at the protein and their distances and on the chain length of the polymeric ligand. A pronounced chain length dependence of the overall binding constant was predicted even at chain lengths much longer than the size of the protein. Such a dependence was previously observed for the enzyme prolyl hydroxylase which acts on polymeric substrates like (ProProGly)n. This so far unexplained feature is quantitatively described by the model mechanism which is believed to be applicable to many other interactions of biological importance.     

Non-catalytic Binding Sites Induce Weaker Long-Range Evolutionary Rate Gradients than Catalytic Sites in Enzymes

Enzymes exhibit a strong long-range evolutionary constraint that extends from their catalytic site and affects even distant sites, where site-specific evolutionary rate increases monotonically with distance. While protein–protein sites in enzymes were previously shown to induce only a weak conservation gradient, a comprehensive relationship between different types of functional sites in proteins and the magnitude of evolutionary rate gradients they induce has yet to be established. Here, we systematically calculate the evolutionary rate (dN/dS) of sites as a function of distance from different types of binding sites in enzymes and other proteins: catalytic sites, non-catalytic ligand binding sites, allosteric binding sites, and protein–protein interaction sites. We show that catalytic sites indeed induce significantly stronger evolutionary rate gradient than all other types of non-catalytic binding sites. In addition, catalytic sites in enzymes with no known allosteric function still induce strong long-range conservation gradients. Notably, the weak long-range conservation gradients induced by non-catalytic binding sites in enzymes is nearly identical in magnitude to those induced by ligand binding sites in non-enzymes. Finally, we show that structural determinants such as local solvent exposure of sites cannot explain the observed difference between catalytic and non-catalytic functional sites. Our results suggest that enzymes and non-enzymes share similar evolutionary constraints only when examined from the perspective of non-catalytic functional sites. Hence, the unique evolutionary rate gradient from catalytic sites in enzymes is likely driven by the optimization of catalysis rather than ligand binding and allosteric functions.     

Similar binding sites and different partners: implications to shared proteins in cellular pathways

We studied a data set of structurally similar interfaces that bind to proteins with different binding-site structures and different functions. Our multipartner protein interface clusters enable us to address questions like: What makes a given site bind different proteins? How similar/different are the interactions? And, what drives the apparently less-specific association? We find that proteins with common binding-site motifs preferentially use conserved interactions at similar interface locations, despite the different partners. Helices are major vehicles for binding different partners, allowing alternate ways to achieve favorable association. The binding sites are characterized by imperfect packing, planar architectures, bridging water molecules, and, on average, smaller size. Interestingly, analysis of the connectivity of these proteins illustrates that they have more interactions with other proteins. These findings are important in predicting "date hubs," if we assume that "date hubs" are shared proteins with binding sites capable of transient binding to multipartners, linking higher-order networks.     

Multisite proteins and randomly coiled polymeric ligands. chain length dependence of binding constants

A model mechanism was developed for the binding of a rigid multisite protein with a randomly coiled multivalent ligand. Probabilities of the formation of chain loops between sites located at given distances at the protein were calculated by an extension of the concept of ring closure in coiled chain molecules. Expressions were derived for the dependence of overall equilibrium quantities, such as the binding constant between the protein and the ligand, on intrinsic parameters such as intrinsic binding constants, number of sites at the protein and their distances and on the chain length of the polymeric ligand. A pronounced chain length dependence of the overall binding constant was predicted even at chain lengths much longer than the size of the protein. Such a dependence was previously observed for the enzyme prolyl hydroxylase which acts on polymeric substrates like (ProProGly)_n. This so far unexplained feature is quantitatively described by the model mechanism which is believed to be applicable to many other interactions of biological importance.     

Analysis of sequence variation among legume lectins. A ring of hypervariable residues forms the perimeter of the carbohydrate-binding site.

Twelve plant lectins from the Papilionoideae subfamily were selected to represent a range of carbohydrate specificities, and their sequences were aligned. Two variability indices were applied to the aligned sequences and the results were analysed using the three-dimensional structures of concanavalin A and the pea lectin. The areas of greatest variability were located in the carbohydrate-binding site region, forming a perimeter around a well-conserved core. These residues are inferred to be specificity determining, in the manner of antibodies, and the most variable position corresponded to Tyr100 in concanavalin A, a known ligand contact residue. In addition to the five peptide loops known to form the binding site from crystallographic studies, a sixth segment with variable residues was located in the binding-site region, and this may contribute to oligosaccharide specificity. In their overall composition, the lectin sites resemble those of the sugar-transport proteins rather than antibodies. The prospects for modelling lectin binding sites by the methods used for antibodies were also assessed.     

Distance mapping of protein-binding sites using spin-labeled oligosaccharide ligands

The binding of a nitroxide spin-labeled analog of N-acetyllactosamine to galectin-3, a mammalian lectin of 26 kD size, is studied to map the binding sites of this small oligosaccharide on the protein surface. Perturbation of intensities of cross-peaks in the (15)N heteronuclear single quantum coherence (HSQC) spectrum of full-length galectin-3 owing to the bound spin label is used qualitatively to identify protein residues proximate to the binding site for N-acetyllactosamine. A protocol for converting intensity measurements to a more quantitative determination of distances between discrete protein amide protons and the bound spin label is then described. This protocol is discussed as part of a drug design strategy in which subsequent perturbation of chemical shifts of distance mapped amide cross-peaks can be used effectively to screen a library of compounds for other ligands that bind to the target protein at distances suitable for chemical linkage to the primary ligand. This approach is novel in that it bypasses the need for structure determination and resonance assignment of the target protein.     

Direct observation of protein-ligand interaction kinetics

Internal dynamics on the micro- to millisecond time scale have a strong influence on the affinity and specificity with which a protein binds ligands. This time scale is accessible through relaxation dispersion measurements using NMR. By studying the dynamics of a protein with different concentrations of a ligand, one can determine the dynamic effects induced by the ligand. Here we have studied slow internal dynamics of the N-terminal src homology 2 domain of phosphatidylinositide 3-kinase to probe the role of individual residues for the interaction with a tyrosine-phosphorylated binding sequence from polyoma middle T antigen. While slow dynamic motion was restricted to a few residues in the free SH2 and in the SH2 complex, motion was significantly enhanced by adding even small amounts of ligand. Kinetic rates induced by ligand binding varied between 300 and 2000 s(-1). High rates reflected direct interactions with the ligand or rearrangements caused by ligand binding. Large differences in rates were observed for residues adjacent in the primary sequence reflecting their individual roles in ligand interaction. However, rates were similar for residues involved in the same side chain interactions, reflecting concerted motions during ligand binding. For a subset of residues, exchange must involve structural intermediates which play a crucial role in high-affinity ligand binding. This analysis supports a new view of the dynamics of individual sites of a protein during ligand interaction.     

In silico modeling and experimental evidence of coagulant protein interaction with precursors for nanoparticle functionalization

The design of novel protein–nanoparticle hybrid systems has applications in many fields of science ranging from biomedicine, catalysis, water treatment, etc. The main barrier in devising such tool is lack of adequate information or poor understanding of protein–ligand chemistry. Here, we establish a new strategy based on computational modeling for protein and precursor linkers that can decorate the nanoparticles. Moringa oleifera (MO_2.1) seed protein that has coagulation and antimicrobial properties was used. Superparamagnetic nanoparticles (SPION) with precursor ligands were used for the protein–ligand interaction studies. The molecular docking studies reveal that there are two binding sites, one is located at the core binding site; tetraethoxysilane (TEOS) or 3-aminopropyl trimethoxysilane (APTES) binds to this site while the other one is located at the side chain residues where trisodium citrate (TSC) or Si₆₀ binds to this site. The protein–ligand distance profile analysis explains the differences in functional activity of the decorated SPION. Experimentally, TSC-coated nanoparticles showed higher coagulation activity as compared to TEOS- and APTES-coated SPION. To our knowledge, this is the first report on in vitro experimental data, which endorses the computational modeling studies as a powerful tool to design novel precursors for functionalization of nanomaterials; and develop interface hybrid systems for various applications.     

Homology-Modelling Protein-Ligand Interactions: Allowing for Ligand-Induced Conformational Change

Current homology-modelling methods do not consider small molecules in their automated processes. Therefore, the development of a reliable tool for protein-ligand homology modelling is an important next step in generating plausible models for molecular interactions. Two automated protein-ligand homology-modelling strategies, requiring no expert knowledge from the user, are investigated here. Both employ the “induced fit” concept with flexibility in side chains and ligand. The most successful strategy superimposes the new ligand over the original ligand before homology modelling, allowing the new ligand to be taken into consideration during protein modelling (rather than after), facilitating conformational change in the local backbone if necessary. We show that this approach results in successful modelling of the ligand and key binding-site residues of angiotensin-converting enzyme 2 (ACE2) from its homologue ACE, which is not possible via conventional homology modelling or by homology modelling followed by docking. Several other difficult target complexes are also successfully modelled, reproducing native protein-ligand contacts with significantly different biological substrates and different binding-site conformations. These include the modelling of Cdk5 (cyclin-dependent kinase 5) from Cdk2, thymidine phosphorylase from a bacterial homologue, and dihydrofolate reductase from a recombinant variant with a markedly different inhibitor. In terms of average modelling quality across 82 targets, the ligand RMSD with respect to the experimental structure is 1.4 Å (and 2.0 Å for the protein binding site) for “easy” cases and 2.9 Å for the ligand (and 2.7 Å for the protein binding site) in “hard” cases. This demonstrates the importance of selecting an optimal template. Ligand-modelling accuracy is strongly dependent on target-template ligand structural similarity, rather than target-template sequence identity. However, protein-modelling accuracy is dependent on both. Our automated protein-ligand homology-modelling strategy generates a higher degree of accuracy than homology modelling followed by docking, generating an average ligand RMSD that is 1-2 Å better than docking with homology models.     

Structural basis for the binding of compatible solutes by ProX from the hyperthermophilic archaeon Archaeoglobus fulgidus

Compatible solutes such as glycine betaine and proline betaine serve as protein stabilizers because of their preferential exclusion from protein surfaces. To use extracellular sources of this class of compounds as osmo-, cryo-, or thermoprotectants, Bacteria and Archaea have developed high affinity uptake systems of the ATP-binding cassette type. These transport systems require periplasmic- or extracellular-binding proteins that are able to bind the transported substance with high affinity. Therefore, binding proteins that bind compatible solutes have to avoid the exclusion of their ligands within the binding pocket. In the present study we addressed the question to how compatible solutes can be effectively bound by a protein at temperatures around 83 degrees C as this is done by the ligand-binding protein ProX from the hyperthermophilic archaeon Archaeoglobus fulgidus. We solved the structures of ProX without ligand and in complex with both of its natural ligands glycine betaine and proline betaine, as well as in complex with the artificial ligand trimethylammonium. Cation-pi interactions and non-classical hydrogen bonds between four tyrosine residues, a main chain carbonyl oxygen, and the ligand have been identified to be the key determinants in binding the quaternary amines of the three investigated ligands. The comparison of the ligand binding sites of ProX from A. fulgidus and the recently solved structure of ProX from Escherichia coli revealed a very similar solution for the problem of compatible solute binding, although both proteins share only a low degree of sequence identity. The residues involved in ligand binding are functionally equivalent but not conserved in the primary sequence.