Scoring docked conformations generated by rigid-body protein-protein docking

Rigid-body methods, particularly Fourier correlation techniques, are very efficient for docking bound (co-crystallized) protein conformations using measures of surface complementarity as the target function. However, when docking unbound (separately crystallized) conformations, the method generally yields hundreds of false positive structures with good scores but high root mean square deviations (RMSDs). This paper describes a two-step scoring algorithm that can discriminate near-native conformations (with less than 5 A RMSD) from other structures. The first step includes two rigid-body filters that use the desolvation free energy and the electrostatic energy to select a manageable number of conformations for further processing, but are unable to eliminate all false positives. Complete discrimination is achieved in the second step that minimizes the molecular mechanics energy of the retained structures, and re-ranks them with a combined free-energy function which includes electrostatic, solvation, and van der Waals energy terms. After minimization, the improved fit in near-native complex conformations provides the free-energy gap required for discrimination. The algorithm has been developed and tested using docking decoys, i.e., docked conformations generated by Fourier correlation techniques. The decoy sets are available on the web for testing other discrimination procedures. Proteins 2000;40:525-537.     

Protein-protein docking with simultaneous optimization of rigid-body displacement and side-chain conformations

Protein-protein docking algorithms provide a means to elucidate structural details for presently unknown complexes. Here, we present and evaluate a new method to predict protein-protein complexes from the coordinates of the unbound monomer components. The method employs a low-resolution, rigid-body, Monte Carlo search followed by simultaneous optimization of backbone displacement and side-chain conformations using Monte Carlo minimization. Up to 10(5) independent simulations are carried out, and the resulting "decoys" are ranked using an energy function dominated by van der Waals interactions, an implicit solvation model, and an orientation-dependent hydrogen bonding potential. Top-ranking decoys are clustered to select the final predictions. Small-perturbation studies reveal the formation of binding funnels in 42 of 54 cases using coordinates derived from the bound complexes and in 32 of 54 cases using independently determined coordinates of one or both monomers. Experimental binding affinities correlate with the calculated score function and explain the predictive success or failure of many targets. Global searches using one or both unbound components predict at least 25% of the native residue-residue contacts in 28 of the 32 cases where binding funnels exist. The results suggest that the method may soon be useful for generating models of biologically important complexes from the structures of the isolated components, but they also highlight the challenges that must be met to achieve consistent and accurate prediction of protein-protein interactions.     

Divergence of interdomain geometry in two-domain proteins

For homologous protein chains composed of two domains, we have determined the extent to which they conserve (1) their interdomain geometry and (2) the molecular structure of the domain interface. This work was carried out on 128 unique two-domain architectures. Of the 128, we find 75 conserve their interdomain geometry and the structure of their domain interface; 5 conserve their interdomain geometry but not the structure of their interface; and 48 have variable geometries and divergent interface structure. We describe how different types of interface changes or the absence of an interface is responsible for these differences in geometry. Variable interdomain geometries can be found in homologous structures with high sequence identities (70%).     

Effect of local shape modifications of molecular surfaces on rigid-body protein-protein docking

Geometric complementarity is the most dominant term in protein-protein docking and therefore, a good geometric representation of the molecules, which takes into account the flexibility of surface residues, is desirable. We present a modified geometric representation of the molecular surface that down-weighs the contribution of specified parts of the surface to the complementarity score. We apply it to the mobile ends of the most flexible side chains: lysines, glutamines and arginines (trimming). The new representation systematically reduces the complementarity scores of the false-positive solutions, often more than the scores of the correct solutions, thereby improving significantly our ability to identify nearly correct solutions in rigid-body docking of unbound structures. The effect of trimming lysine residues is larger than trimming of glutamine or arginine residues. It appears to be independent of the conformations of the trimmed residues but depends on the relative abundance of such residues at the interface and on the non-interacting surface. Combining the modified geometric representation with electrostatic complementarity further improves the docking results.     

On docking, scoring and assessing protein-DNA complexes in a rigid-body framework

We consider the identification of interacting protein-nucleic acid partners using the rigid body docking method FTdock, which is systematic and exhaustive in the exploration of docking conformations. The accuracy of rigid body docking methods is tested using known protein-DNA complexes for which the docked and undocked structures are both available. Additional tests with large decoy sets probe the efficacy of two published statistically derived scoring functions that contain a huge number of parameters. In contrast, we demonstrate that state-of-the-art machine learning techniques can enormously reduce the number of parameters required, thereby identifying the relevant docking features using a miniscule fraction of the number of parameters in the prior works. The present machine learning study considers a 300 dimensional vector (dependent on only 15 parameters), termed the Chemical Context Profile (CCP), where each dimension reflects a specific type of protein amino acid-nucleic acid base interaction. The CCP is designed to capture the chemical complementarities of the interface and is well suited for machine learning techniques. Our objective function is the Chemical Context Discrepancy (CCD), which is defined as the angle between the native system's CCP vector and the decoy's vector and which serves as a substitute for the more commonly used root mean squared deviation (RMSD). We demonstrate that the CCP provides a useful scoring function when certain dimensions are properly weighted. Finally, we explore how the amino acids on a protein's surface can help guide DNA binding, first through long-range interactions, followed by direct contacts, according to specific preferences for either the major or minor grooves of the DNA.     

Automated docking of substrates to proteins by simulated annealing

The Metropolis technique of conformation searching is combined with rapid energy evaluation using molecular affinity potentials to give an efficient procedure for docking substrates to macromolecules of known structure. The procedure works well on a number of crystallographic test systems, functionally reproducing the observed binding modes of several substrates.     

Incorporation of flexibility into rigid-body docking: applications in rounds 3-5 of CAPRI

We have submitted models for all 9 targets in Rounds 3-5 of CAPRI and have predicted at least 30% of the correct contacts for 4 of the targets and at least 10% of the correct contacts for another 4 targets. We have employed a variety of techniques but have had the greatest success by combining established rigid-body docking with a variety of initial conformations generated by molecular dynamics.     

Structural requirements for membrane assembly of proteins spanning the membrane several times

We have investigated the structural requirements for the biogenesis of proteins spanning the membrane several times. Proteins containing various combinations of topological signals (signal anchor and stop transfer sequences) were synthesized in a cell-free translation system and their membrane topology was determined. Proteins spanning the membrane twice were obtained when a signal anchor sequence was followed by either a stop transfer sequence or a second signal anchor sequence. Thus, a signal anchor sequence in the second position can function as a stop transfer sequence, spanning the membrane in the opposite orientation to that of the first signal anchor sequence. A signal anchor sequence in the third position was able to insert amino acid sequences located COOH terminal to it. We conclude that proteins spanning the membrane several times can be generated by stringing together signal anchor and stop transfer sequences. However, not all proteins with three topological signals were found to span the membrane three times. A certain segment located between the first and second topological signal could prevent stable membrane integration of a third signal anchor segment.     

Structural studies of a two-domain chitinase from Streptomyces griseus HUT6037

Chitinase C (ChiC) from Streptomyces griseus HUT6037 was the first glycoside hydrolase family 19 chitinase that was found in an organism other than higher plants. An N-terminal chitin-binding domain and a C-terminal catalytic domain connected by a linker peptide constitute ChiC. We determined the crystal structure of full-length ChiC, which is the only representative of the two-domain chitinases in the family. The catalytic domain has an alpha-helix-rich fold with a deep cleft containing a catalytic site, and lacks three loops on the domain surface compared with the catalytic domain of plant chitinases. The chitin-binding domain is an all-beta protein with two tryptophan residues (Trp59 and Trp60) aligned on the surface. We suggest the binding mechanism of tri-N-acetylchitotriose onto the chitin-binding domain on the basis of molecular dynamics (MD) simulations. In this mechanism, the ligand molecule binds well on the surface-exposed binding site through two stacking interactions and two hydrogen bonds and only Trp59 and Trp60 are involved in the binding. Furthermore, the flexibility of the Trp60 side-chain, which may be involved in adjusting the binding surface to fit the surface of crystalline chitin by the rotation of chi2 angle, is shown.     

Narrowing the conformational space sampled by two-domain proteins with paramagnetic probes in both domains

Calmodulin is a two-domain protein which in solution can adopt a variety of conformations upon reorientation of its domains. The maximum occurrence (MO) of a set of calmodulin conformations that are representative of the overall conformational space possibly sampled by the protein, has been calculated from the paramagnetism-based restraints. These restraints were measured after inclusion of a lanthanide binding tag in the C-terminal domain to supplement the data obtained by substitution of three paramagnetic lanthanide ions to the calcium ion in the second calcium binding loop of the N-terminal domain. The analysis shows that the availability of paramagnetic restraints arising from metal ions placed on both domains, reduces the MO of the conformations to different extents, thereby helping to identify those conformations that can be mostly sampled by the protein.     

The affinity-enhancing roles of flexible linkers in two-domain DNA-binding proteins.

Recently many attempts have been made to design high-affinity DNA-binding proteins by linking two domains. Here a theory for guiding these designs is presented. Flexible linkers may play three types of roles: (a) linking domains which by themselves are unfolded and bind to DNA only as a folded dimer (as in a designed single-chain Arc repressor), (b) connecting domains which can separately bind to DNA (as in the Oct-1 POU domain), and (c) linking a DNA-binding domain with a dimerization domain (as in the lambda repressor). In (a), the linker keeps the protein as a folded dimer so that it is always DNA-binding-competent. In (b), the linker is predicted to enhance DNA-binding affinity over those of the individual domains (with dissociation constants K(A) and K(B)) by p(d(0))/K(B) or p(d(0))/K(A), where p(d(0)) = (3/4pil(p)bL)(3/2) exp(-3d(0)(2)/4l(p)bL)(1 - 5l(p)/4bL +...) is the probability density for the end-to-end vector of the linker with L residues to have a distance d(0). In (c), the linker is predicted to enhance the binding affinity by K(d)(C)/p(d(0)), where K(d)(C) is the dimer dissociation constant for the dimerization domain. The predicted affinity enhancements are found to be actually reached by the Oct-1 POU domain and lambda repressor. However, there is room for improvement in many of the recently designed proteins. The theoretical limits presented should provide a useful guide for current efforts of designing DNA-binding proteins.     

Comparison of Segger and other methods for segmentation and rigid-body docking of molecular components in cryo-EM density maps

Segmentation and docking are useful methods for the discovery of molecular components in electron cryo-microscopy (cryo-EM) density maps of macromolecular complexes. In this article, we describe the segmentation and docking methods implemented in Segger. For 11 targets posted in the 2010 cryo-EM challenge, we segmented the regions corresponding to individual molecular components using Segger. We then used the segmented regions to guide rigid-body docking of individual components. Docking results were evaluated by comparing the docked components with published structures, and by calculation of several scores, such as atom inclusion, density occupancy, and geometry clash. The accuracy of the component segmentation using Segger and other methods was assessed by comparing segmented regions with docked components.     

Structural relationships between clathrin assembly proteins from the Golgi and the plasma membrane

We have established by peptide mapping and immunochemical analysis of purified clathrin assembly protein preparations from bovine brain, that the cluster of components of mol. wt 100-120 kd fall into four classes, which we term alpha, beta, beta' and gamma. The beta and beta' proteins are immunologically related and generate a series of common tryptic peptides. The same criteria reveal no such homologies between the alpha, beta(beta') and gamma polypeptides. The so-called HA-II assembly protein group contains equimolar amounts of alpha and beta class polypeptides, which are shown to interact with each other. In the HA-I group assembly protein complex gamma and beta' class polypeptides form a stoichiometric complex. Immunofluorescence microscopy reveals that the HA-I complex is specifically associated with clathrin-coated membranes in the Golgi region of cultured cells, whereas the HA-II complex appears to be restricted to coated pits on the plasma membrane. The data lead to the tentative conclusion that the clathrin assembly proteins are involved in the recognition of the intracellular targets by uncoated vesicles.     

Iterative assembly of helical proteins by optimal hydrophobic packing

We present a method for the computer-based iterative assembly of native-like tertiary structures of helical proteins from alpha-helical fragments. For any pair of helices, our method, called MATCHSTIX, first generates an ensemble of possible relative orientations of the helices with various ways to form hydrophobic contacts between them. Those conformations having steric clashes, or a large radius of gyration of hydrophobic residues, or with helices too far separated to be connected by the intervening linking region, are discarded. Then, we attempt to connect the two helical fragments by using a robotics-based loop-closure algorithm. When loop closure is feasible, the algorithm generates an ensemble of viable interconnecting loops. After energy minimization and clustering, we use a representative set of conformations for further assembly with the remaining helices, adding one helix at a time. To efficiently sample the conformational space, the order of assembly generally proceeds from the pair of helices connected by the shortest loop, followed by joining one of its adjacent helices, always proceeding with the shorter connecting loop. We tested MATCHSTIX on 28 helical proteins each containing up to 5 helices and found it to heavily sample native-like conformations. The average rmsd of the best conformations for the 17 helix-bundle proteins that have 2 or 3 helices is less than 2 A; errors increase somewhat for proteins containing more helices. Native-like states are even more densely sampled when disulfide bonds are known and imposed as restraints. We conclude that, at least for helical proteins, if the secondary structures are known, this rapid rigid-body maximization of hydrophobic interactions can lead to small ensembles of highly native-like structures. It may be useful for protein structure prediction.     

Lipid interaction networks of peripheral membrane proteins revealed by data-driven micelle docking

Many signaling and trafficking proteins contain modular domains that bind reversibly to cellular membranes. The structural basis of the intermolecular interactions which mediate these membrane-targeting events remains elusive since protein-membrane complexes are not directly accessible to standard structural biology techniques. Here we report a fast protein-micelle docking methodology that yields three-dimensional model structures of proteins inserted into micelles, revealing energetically favorable orientations, convergent insertion angles, and an array of protein-lipid interactions at atomic resolution. The method is applied to two peripheral membrane proteins, the early endosome antigen 1 (EEA1) FYVE (a zinc finger domain found in the proteins Fab1, YOTB/ZK632.12, Vac1, and EEA1) and Vam7p phagocyte oxidase homology domains, which are revealed to form extensive networks of interactions with multiple phospholipid headgroups and acyl chains. The resulting structural models explain extensive published mutagenesis data and reveal novel binding determinants. The docking restraints used here were based on NMR data, but can be derived from any technique that detects insertion of protein residues into a membrane, and can be applied to virtually any peripheral membrane protein or membrane-like structure.     

Polyglutamylation of nucleosome assembly proteins

Polyglutamylation is an original posttranslational modification, discovered on tubulin, consisting in side chains composed of several glutamyl units and leading to a very unusual protein structure. A monoclonal antibody directed against glutamylated tubulin (GT335) was found to react with other proteins present in HeLa cells. After immunopurification on a GT335 affinity column, two prominent proteins of approximately 50 kDa were observed. They were identified by microsequencing and mass spectrometry as NAP-1 and NAP-2, two members of the nucleosome assembly protein family that are implicated in the deposition of core histone complexes onto chromatin. Strikingly, NAP-1 and NAP-2 were found to be substrates of an ATP-dependent glutamylation enzyme co-purifying on the same column. We took advantage of this property to specifically label and purify the polyglutamylated peptides. NAP-1 and NAP-2 are modified in their C-terminal domain by the addition of up to 9 and 10 glutamyl units, respectively. Two putative glutamylation sites were localized for NAP-1 at Glu-356 and Glu-357 and, for NAP-2, at Glu-347 and Glu-348. These results demonstrate for the first time that proteins other than tubulin are polyglutamylated and open new perspectives for studying NAP function.     

Monte Carlo docking of oligopeptides to proteins.

A new two-step procedure has been developed for the docking of flexible oligopeptide chains of unknown conformation to static proteins of known structure. In the first step positions and conformations are sampled and the association energy minimized starting from an approximate preselected docking position. The resulting conformations are further optimized in the second step by a Metropolis Monte Carlo minimization, which optimizes each of these structures. The method has been tested on the HIV-1 aspartic proteinase complex with an inhibitor, whose crystallographic structure is known at 2.3 A resolution. Furthermore, the application of this method to the docking of the hendecapeptide 58-68 of the influenza A virus matrix protein to the HLA-A2 molecule produced results which are in agreement with experimental observations in identifying side chains critical for T cell recognition and residues responsible of MHC protein binding.     

Structural determinants of purple membrane assembly

The purple membrane is a two-dimensional crystalline lattice formed by bacteriorhodopsin and lipid molecules in the cytoplasmic membrane of Halobacterium salinarum. High-resolution structural studies, in conjunction with detailed knowledge of the lipid composition, make the purple membrane one of the best models for elucidating the forces that are responsible for the assembly and stability of integral membrane protein complexes. In this review, recent mutational efforts to identify the structural features of bacteriorhodopsin that determine its assembly in the purple membrane are discussed in the context of structural, calorimetric and reconstitution studies. Quantitative evidence is presented that interactions between transmembrane helices of neighboring bacteriorhodopsin molecules contribute to purple membrane assembly. However, other specific interactions, particularly between bacteriorhodopsin and lipid molecules, may provide the major driving force for assembly. Elucidating the molecular basis of protein-protein and protein-lipid interactions in the purple membrane may provide insights into the formation of integral membrane protein complexes in other systems.     

Synthesis and assembly of connecting-piece proteins as revealed by radioautography

The synthesis and assembly of connecting-piece proteins have been studied during spermiogenesis in the rat by electron microscopy and radioautography following intratesticular injection of radiolabeled amino acids [3H]proline and [3H]cystine. Early in spermiogenesis (steps 1-7) the two centrioles that give rise to the connecting piece are essentially unmodified. During the 6.5-day period between steps 8 and 15, the major elements of the connecting piece (striated columns and capitulum) gradually become assembled from an electron-dense material that is deposited around the walls of the centrioles; throughout this period, protein molecules containing proline and cystine are synthesized by the step 8-15 spermatids and incorporated into the developing neck region. These proteins subsequently become permanent structural components of the connecting piece. Following completion of the major elements in step 15, few additional proteins are added to the connecting piece during the final steps 16-19 of spermiogenesis.     

Structural characterisation of the native fetuin-binding protein Scilla campanulata agglutinin: a novel two-domain lectin

The three-dimensional structure of a 244-residue, multivalent, fetuin-binding lectin, SCAfet, isolated from bluebell (Scilla campanulata) bulbs, has been solved at 3.3 A resolution by molecular replacement using the coordinates of the 119-residue, mannose-binding lectin, SCAman, also from bluebell bulbs. Unlike most monocot mannose-binding lectins, such as Galanthus nivalis agglutinin from snowdrop bulbs, which fold into a single domain, SCAfet contains two domains with approximately 55% sequence identity, joined by a linker peptide. Both domains are made up of a 12-stranded beta-prism II fold, with three putative carbohydrate-binding sites, one on each subdomain. SCAfet binds to the complex saccharides of various animal glycoproteins but not to simple sugars.