Macromolecular Modeling Blog ™

This is the fourth post in the CAPRI series, summarizing the presentations of Paul Bates, Martin Zacharias, and Carlos Camacho, as provided by the speakers. More to appear in the continuation of the series.

Searching for Encounter Complexes: Taking Account of Macromolecular Crowding

Xiaofan Li, Iain Moal and Paul Bates

So far, docking has been a binary endeavour, with one receptor versus one ligand. However, in vivo, proteins interact with one another in a crowded macromolecular environment where specific and non-specific interactions occur simultaneously. We have developed BioSimz, a software package to simulate multiple interactions occurring in a crowded environment. Langevin dynamics is used in a box with periodic boundary conditions on a longer time-scale than traditional molecular dynamics allows. The behaviour of crowded molecular system can be studied in the context of lipids and ions (see figure to the right), and residues can be coloured to show how frequently they are involved in enduring (> 100ps) encounter complexes (see figure below). From these simulations, potential encounter complexes can be found and used to find starting positions for refinement. Our docking algorithm, SwarmDock, flexibly refines in from these starting positions using a linear combination of elastic network normal modes calculated at an atomistic level.

Residues Frequencies in Enduring Encounter Complexes

Flexible protein-protein docking with a coarse-grained protein model

Martin Zacharias

Focus of the talk was the efficient inclusion of protein flexibility during the protein-protein docking process. In a first part the coarse-grained model to describe proteins within the Attract docking program was presented [1,2]. The program is based on docking minimization with respect to orientational and translational degrees of freedom and additional degrees of freedom that represent protein flexibility. It included a description of an improved parameterization based on systematic optimization of the ranking of near-native docking minima with respect to decoy complexes. The new parameters showed improved ranking performance. The approximate inclusion of protein flexibility in terms of pre-calculated normal mode directions with small associated eigenvalue was presented and described for a few example applications [3,4]. A new combination of a stochastic search on docking minima including movements in normal mode directions and stochastic selection of side chain rotamers was also shown. This method could be useful to effectively sample docking minima near an approximately known binding site and to investigate the energy funnel structure around protein binding sites. The application of the docking approach to the recent CAPRI docking targets was presented and the performance was discussed in some detail for some of the target complexes. An effective refinement strategy employing a combination of molecular dynamics simulated annealing and energy minimization to translate the structures to atomic resolution was explained and discussed.

1. Zacharias M (2003). Protein-protein docking with a reduced protein model accounting for side-chain flexibility. Protein science : a publication of the Protein Society, 12 (6), 1271-82 PMID: 12761398
2. Saladin A, Fiorucci S, Poulain P, Prévost C, & Zacharias M (2009). PTools: an opensource molecular docking library. BMC structural biology, 9 PMID: 19409097
3. May A, & Zacharias M (2008). Protein-ligand docking accounting for receptor side chain and global flexibility in normal modes: evaluation on kinase inhibitor cross docking. Journal of medicinal chemistry, 51 (12), 3499-506 PMID: 18517186
4. May A, & Zacharias M (2008). Energy minimization in low-frequency normal modes to efficiently allow for global flexibility during systematic protein-protein docking. Proteins, 70 (3), 794-809 PMID: 17729269

Novel protein interactions decode a quasi exact code for assesing protein-DNA binding with no over-fitting.

Carlos Camacho

A major limitation in modeling protein interactions is the difficulty of assessing the over-fitting of the training set. Recently, an experimentally based approach that integrates crystallographic information of C2H2 zinc finger (ZF)-DNA complexes with binding data from 11 mutants, 7 from EGR finger I, was used to define an improved interaction code (no optimization). Here, we present a novel mixed integer programming (MIP) based method that transforms this type of data into an optimized code, demonstrating both the advantages of the mathematical formulation to minimize over-and-under fitting and the robustness of the underlying physical parameters mapped by the code. Based on structural models of feasible interaction networks for 35 mutants of EGR-DNA complexes, the MIP method minimizes the cumulative binding energy over all complexes for a general set of fundamental protein-DNA interactions. To guard against over-fitting, we use the scalability of the method to probe against the elimination of related interactions. From an initial set of 12 parameters (six hydrogen bonds, five desolvation penalties and a water factor), we proceed to eliminate five of them with only a marginal reduction of the correlation coefficient to 0.9983. Further reduction of parameters negatively impacts the performance of the code (under-fitting). Besides accurately predicting the change in binding affinity of validation sets, the code identifies possible context dependent effects in the definition of the interaction networks. Yet, the approach of constraining predictions to within a pre-selected set of interactions limits the impact of these potential errors to related low affinity complexes.