Predicting protein-protein interactions on a proteome scale by matching evolutionary and structural similarities at interfaces using PRISM

Abstract

Prediction of protein-protein interactions at the structural level on the proteome scale is important because it allows prediction of protein function, helps drug discovery and takes steps toward genome-wide structural systems biology. We provide a protocol (termed PRISM, protein interactions by structural matching) for large-scale prediction of protein-protein interactions and assembly of protein complex structures. The method consists of two components: rigid-body structural comparisons of target proteins to known template protein-protein interfaces and flexible refinement using a docking energy function. The PRISM rationale follows our observation that globally different protein structures can interact via similar architectural motifs. PRISM predicts binding residues by using structural similarity and evolutionary conservation of putative binding residue 'hot spots'. Ultimately, PRISM could help to construct cellular pathways and functional, proteome-scale annotation. PRISM is implemented in Python and runs in a UNIX environment. The program accepts Protein Data Bank-formatted protein structures and is available at http://prism.ccbb.ku.edu.tr/prism_protocol/.

Figure 1 |

Description of the prediction algorithm. (a) Schematic illustration of the concept of the prediction algorithm. If complementary partners (I_L and I_R) of a template interface are similar to surface regions of any two targets (T_L and T_R), these two targets can interact with each other via these regions. The red points are hot spots. These incorporate evolutionary information into the matching. (b) The flowchart of the algorithm. There are two data sets in the algorithm: the template data set and the target data set. First, the surface of the proteins in the target data set is extracted. Next, each partner of the template interface is aligned with the target surfaces. If the match passes the residue and hot spot matching thresholds, these targets are transformed on the template interface. If there are colliding residues (e.g., atoms of the residues penetrate into each other’s van der Waals radii after they are transformed onto the corresponding template interface) between the two partner targets, the putative complexes are eliminated. Otherwise, the predicted complexes are flexibly refined with their global energies computed. The best solution is chosen as the one with the lowest energy value.

Figure 2 |

The putative Falcipain-Cystatin complex predicted by PRISM using template 1stfEI.

Figure 3 |

The putative Chk1-p16^ink complex predicted by PRISM using templates 1blxAB and 1fmaDE. Template interfaces are represented by balls to show the matching parts of the target surface with the template interface partners.