Finding needles in haystacks: Reranking DOT results by using shape complementarity, cluster analysis, and biological information

How to use not-always-reliable binding site information in protein-protein docking prediction

In many protein-protein docking algorithms, binding site information is used to help predicting the protein complex structures. Using correct and accurate binding site information can increase protein-protein docking success rate significantly. On the other hand, using wrong binding sites information should lead to a failed prediction, or, at least decrease the success rate. Recently, various successful theoretical methods have been proposed to predict the binding sites of proteins. However, the predicted binding site information is not always reliable, sometimes wrong binding site information could be given. Hence there is a high risk to use the predicted binding site information in current docking algorithms. In this paper, a softly restricting method (SRM) is developed to solve this problem. By utilizing predicted binding site information in a proper way, the SRM algorithm is sensitive to the correct binding site information but insensitive to wrong information, which decreases the risk of using predicted binding site information. This SRM is tested on benchmark 3.0 using purely predicted binding site information. The result shows that when the predicted information is correct, SRM increases the success rate significantly; however, even if the predicted information is completely wrong, SRM only decreases success rate slightly, which indicates that the SRM is suitable for utilizing predicted binding site information.

ASPDock: protein-protein docking algorithm using atomic solvation parameters model

Background

Atomic Solvation Parameters (ASP) model has been proven to be a very successful method of calculating the binding free energy of protein complexes. This suggests that incorporating it into docking algorithms should improve the accuracy of prediction. In this paper we propose an FFT-based algorithm to calculate ASP scores of protein complexes and develop an ASP-based protein-protein docking method (ASPDock).

Results

The ASPDock is first tested on the 21 complexes whose binding free energies have been determined experimentally. The results show that the calculated ASP scores have stronger correlation (r ≈ 0.69) with the binding free energies than the pure shape complementarity scores (r ≈ 0.48). The ASPDock is further tested on a large dataset, the benchmark 3.0, which contain 124 complexes and also shows better performance than pure shape complementarity method in docking prediction. Comparisons with other state-of-the-art docking algorithms showed that ASP score indeed gives higher success rate than the pure shape complementarity score of FTDock but lower success rate than Zdock3.0. We also developed a softly restricting method to add the information of predicted binding sites into our docking algorithm. The ASP-based docking method performed well in CAPRI rounds 18 and 19.

Conclusions

ASP may be more accurate and physical than the pure shape complementarity in describing the feature of protein docking.

How different from random are docking predictions when ranked by scoring functions?

Docking algorithms predict the structure of protein-protein interactions. They sample the orientation of two unbound proteins to produce various predictions about their interactions, followed by a scoring step to rank the predictions. We present a statistical assessment of scoring functions used to rank near-native orientations, applying our statistical analysis to a benchmark dataset of decoys of protein-protein complexes and assessing the statistical significance of the outcome in the Critical Assessment of PRedicted Interactions (CAPRI) scoring experiment. A P value was assigned that depended on the number of near-native structures in the sampling. We studied the effect of filtering out redundant structures and tested the use of pair-potentials derived using ZDock and ZRank. Our results show that for many targets, it is not possible to determine when a successful reranking performed by scoring functions results merely from random choice. This analysis reveals that changes should be made in the design of the CAPRI scoring experiment. We propose including the statistical assessment in this experiment either at the preprocessing or the evaluation step.

Coactosin-like protein functions as a stabilizing chaperone for 5-lipoxygenase: role of tryptophan 102

The activity of 5-LO (5-lipoxygenase), which catalyses two initial steps in the biosynthesis of pro-inflammatory LTs (leukotrienes), is strictly regulated. One recently discovered factor, CLP (coactosin-like protein), binds 5-LO and promotes LT formation. In the present paper we report that CLP also stabilizes 5-LO and prevents non-turnover inactivation of the enzyme in vitro. Mutagenesis of tryptophan residues in the 5-LO β-sandwich showed that 5-LO-Trp102 is essential for binding to CLP, and for CLP to support 5-LO activity. In addition, the stabilizing effect also depended on binding between CLP and 5-LO. After mutations which prevent interaction (5-LO-W102A or CLP-K131A), the protective effect of CLP was absent. A calculated 5-LO–CLP docking model indicates that CLP may bind to additional residues in both domains of 5-LO, thus possibly stabilizing the 5-LO structure. To obtain further support for binding between CLP and 5-LO in a living cell, subcellular localization of CLP and 5-LO in the monocytic cell line Mono Mac 6 was determined. In these cells, 5-LO associates with a nuclear fraction only when differentiated cells are primed with phorbol ester and stimulated with ionophore. The same pattern of redistribution was found for CLP, indicating that the two proteins associate with the nucleus in a co-ordinated fashion. The results of the present study support a role for CLP as a chaperoning scaffold factor, influencing both the stability and the activity of 5-LO.

Mechanism of alpha-synuclein oligomerization and membrane interaction: theoretical approach to unstructured proteins studies

Misfolding and oligomerization of unstructured proteins is involved in the pathogenesis of Parkinson's disease (PD), Alzheimer's disease, Huntington's disease, and other neurodegenerative disorders. Elucidation of possible conformations of these proteins and their interactions with the membrane is necessary to understand the molecular mechanisms of neurodegeneration. We developed a strategy that makes it possible to elucidate the molecular mechanisms of alpha-synuclein aggregation-a key molecular event in the pathogenesis of PD. This strategy can be also useful for the study of other unstructured proteins involved in neurodegeneration. The results of these theoretical studies have been confirmed with biochemical and electrophysiological studies. Our studies provide insights into the molecular mechanism for PD initiation and progression, and provide a useful paradigm for identifying possible therapeutic interventions through computational modeling.

Assessing the energy landscape of CAPRI targets by FunHunt

RosettaDock has repeatedly created high-resolution structures of protein complexes in the CAPRI experiment, thanks to the explicit modeling of conformational changes of the monomers at the side chain level. These models can be selected based on their energy. During the search for the lowest-energy model, RosettaDock samples a deep funnel around the native orientation, but additional funnels may appear in the energy landscape, especially in cases where backbone conformational changes occur upon binding. We have previously developed FunHunt, a Support Vector Machine-based classifier that distinguishes the energy funnels around the native orientation from other funnels in the energy landscape. Here we assess the ability of FunHunt to help in model selection in the CAPRI experiment. For all of 12 recent CAPRI targets, FunHunt clearly identifies a near-native funnel in comparison to the funnel around the lowest energy model identified by the RosettaDock global search protocol. FunHunt is also able to choose a near-native orientation among models submitted by predictor groups, demonstrating its general applicability for model selection. This suggests that FunHunt will be a valuable tool in coming CAPRI rounds for the selection of models, and for the definition of regions that need further refinement with restricted backbone flexibility.

Identification of near-native structures by clustering protein docking conformations

Most state-of-the-art protein-protein docking algorithms use the Fast Fourier Transform (FFT) technique to sample the six-dimensional translational and rotational space. Scoring functions including shape complementarity, electrostatics, and desolvation are usually exploited in ranking the docking conformations. While these rigid-body docking methods provide good performance in bound docking, using unbound structures as input frequently leads to a high number of false positive hits. For the purpose of better selecting correct docking conformations, we structurally cluster the docking decoys generated by four widely-used FFT-based protein-protein docking methods. In all cases, the selection based on cluster size outperforms the ranking based on the inherent scoring function. If we cluster decoys from different servers together, only marginal improvement is obtained in comparison with clustering decoys from the best individual server. A collection of multiple decoy sets of comparable quality will be the key to improve the clustering result from meta-docking servers.

Progress in computation and amide hydrogen exchange for prediction of protein-protein complexes

The macromolecular docking problem that must be solved for experimental biologists is prediction of the structures of complexes for which the components are known or reliably modeled in the unbound state, but the structure of the complex is unknown. The current state of the art in macromolecular docking is such that solving this problem usually requires supplementary experimental chemical and/or biological information to evaluate computational predictions. Amide (1)H/(2)H exchange measured by mass spectroscopy is a promising approach for obtaining such information, because it can reveal interfacial regions of each member of the complex and identify regions of conformational flexibility in the structure. In a previous article (Anand et al., Proc Natl Acad Sci USA 2003;100:13264-13269), we used (1)H/(2)H exchange data to predict the structure of a complex between regulatory and catalytic subunits of protein kinase A. Comparison of the prediction with a recent crystal structure determination (Kim et al., Science 2005;307:690-696) showed large conformational change in the regulatory subunit on formation of the complex. Analysis of the prediction, previous CAPRI results, novel data processing methods for the (1)H/(2)H exchange data, and new fragment docking computations give grounds for cautious optimism that this method can be useful even in cases of substantial conformational change.

Accounting for loop flexibility during protein-protein docking

Although reliable docking can now be achieved for systems that do not undergo important induced conformational change upon association, the presence of flexible surface loops, which must adapt to the steric and electrostatic properties of a partner, generally presents a major obstacle. We report here the first docking method that allows large loop movements during a systematic exploration of the possible arrangements of the two partners in terms of position and rotation. Our strategy consists in taking into account an ensemble of possible loop conformations by a multi-copy representation within a reduced protein model. The docking process starts from regularly distributed positions and orientations of the ligand around the whole receptor. Each starting configuration is submitted to energy minimization during which the best-fitting loop conformation is selected based on the mean-field theory. Trials were carried out on proteins with significant differences in the main-chain conformation of the binding loop between isolated form and complexed form, which were docked to their partner considered in their bound form. The method is able to predict complexes very close to the crystal complex both in terms of relative position of the two partners and of the geometry of the flexible loop. We also show that introducing loop flexibility on the isolated protein form during systematic docking largely improves the predictions of relative position of the partners in comparison with rigid-body docking.

Computational approaches to protein-protein interaction

The interactions between proteins allow the cell's life. A number of experimental, genome-wide, high-throughput studies have been devoted to the determination of protein-protein interactions and the consequent interaction networks. Here, the bioinformatics methods dealing with protein-protein interactions and interaction network are overviewed. 1. Interaction databases developed to collect and annotate this immense amount of data; 2. Automated data mining techniques developed to extract information about interactions from the published literature; 3. Computational methods to assess the experimental results developed as a consequence of the finding that the results of high-throughput methods are rather inaccurate; 4. Exploitation of the information provided by protein interaction networks in order to predict functional features of the proteins; and 5. Prediction of protein-protein interactions.

FastContact: rapid estimate of contact and binding free energies

Interaction free energies are crucial for analyzing binding propensities in proteins. Although the problem of computing binding free energies remains open, approximate estimates have become very useful for filtering potential binding complexes. We report on the implementation of a fast computational estimate of the binding free energy based on a statistically determined desolvation contact potential and Coulomb electrostatics with a distance-dependent dielectric constant, and validated in the Critical Assessment of PRotein Interactions experiment. The application also reports residue contact free energies that rapidly highlight the hotspots of the interaction.

AVAILABILITY

The program was written in Fortran. The executable and full documentation is freely available at http://structure.pitt.edu/software/FastContact

On proteins, grids, correlations, and docking

The activity of a living cell can be portrayed as a network of interactions involving proteins and nucleic acids that transfer biological information. Intervention in cellular processes requires thorough understanding of the interactions between the molecules, which can be provided by docking techniques. Docking methods attempt to predict the structures of complexes given the structures of the component molecules. We focus hereby on protein-protein docking procedures that employ grid representations of the molecules, and use correlation for searching the solution space and evaluating putative complexes. Geometric surface complementarity is the dominant descriptor in docking. Inclusion of electrostatics often improves the results of geometric docking for soluble proteins, whereas hydrophobic complementarity is more important in construction of oligomers. Using binding-site information in the scan or as a filter helps to identify and up-rank nearly correct solutions.

Identification of the protein kinase A regulatory RIalpha-catalytic subunit interface by amide H/2H exchange and protein docking

An important goal after structural genomics is to build up the structures of higher-order protein-protein complexes from structures of the individual subunits. Often structures of higher order complexes are difficult to obtain by crystallography. We have used an alternative approach in which the structures of the individual catalytic (C) subunit and RIalpha regulatory (R) subunit of PKA were first subjected to computational docking, and the top 100,000 solutions were subsequently filtered based on amide hydrogen/deuterium (H/2H) exchange interface protection data. The resulting set of filtered solutions forms an ensemble of structures in which, besides the inhibitor peptide binding site, a flat interface between the C-terminal lobe of the C-subunit and the A- and B-helices of RIalpha is uniquely identified. This holoenzyme structure satisfies all previous experimental data on the complex and allows prediction of new contacts between the two subunits.