Protein-protein docking

From Freepedia

Protein-protein docking is a field of theoretical biochemistry aimed at predicting properties of the complexes formed by two or more proteins. Specifically, for any given set of proteins, it aims to answer the following questions:

• Do the proteins bind in vivo?

If they bind,

• What is the spatial configuration which they adopt in their bound state?
• How strong or weak is their interaction?

The field of protein-protein docking is highly computationally oriented, and it shares approaches with molecular docking. Molecular docking is sometimes referred to as small-molecule docking, to draw a distinction from protein-protein docking. Proteins complexed with polynucleotide molecules are widely studied using similar or identical approaches to protein-protein docking, although if the nucleotide molecule is small enough, the case may be framed as a molecular docking problem.

Contents 1 Generating putative complex structures 2 Rigid-Body docking vs. Flexible docking 3 Ranking a generated set of structures 4 Deciding whether or not a complex actually occurs in nature and measuring its affinity

Generating putative complex structures

The structures for the components of the complex must be available individually; if not, and their sequence is known, they may be homology modelled. Then one of various geometrical techniques is used to generate possible structures for the complex itself. These include:

A simple series of discrete translations and rotations of the components with respect to each other and to one fixed component (usually the largest, to save time).

A series of molecular mechanics runs.

Where the complex is a homomultimer, each generated structure may be required to possess a symmetry axis, since the vast majority of real homomultimeric protein complexes possess a symmetry axis.

Rigid-Body docking vs. Flexible docking

If the bond angles, bond lengths and torsion angles of the components are not modified at any stage of complex generation, it is known as rigid body docking. A subject of debate is whether or not rigid-body docking is sufficiently good to find most complexes. When substantial conformational change occurs within the components at the time of complexation, rigid-body docking is clearly seen to be inadequate. However, exahaustively accounting for all possible conformational change is prohibitively expensive in computer time; it can be argued such cases are fundamentally beyond calculation anyway. Docking procedures which permit conformational change, or flexible docking procedures necessarily explore only a small subset of conformational change.

Flexible docking methods fall into two main classes:

Methods which use libraries of possible conformations of sidechains (rotamer libraries) to generate new configurations, and

Methods which use physical laws to generate new configurations.

An example of the former method is the SCWRL algorithm, which prunes a graph of interacting rotamers. The latter family includes molecular dynamics-based and normal mode-based methods.

Ranking a generated set of structures

Heuristic scores are used in protein-protein docking to compare the suitability of a set of putative complexes. Examples are based on residue contacts, shape complementarity of molecular surfaces, and free energies estimated using parameters from molecular mechanics force fields developed by theoretical chemists, such as CHARMM or AMBER. Evolutionary history of amino-acid sequences of the associating proteins is also examined for clues about the functional site.

Deciding whether or not a complex actually occurs in nature and measuring its affinity

A reliable method for affinity prediction has the potential to transform biochemistry and cell biology. This may be considered the as the ultimate aim of the protein-protein docking community. With the scope current methods, however, affinity predictions remain a distant prospect.