Automated Prediction of Protein-Protein Association Rate Constants

Abstract

The association of two proteins generally occurs in two steps [1]. The first is the formation of a transient complex, in which the two proteins have near-native separation and relative orientation. In the subsequent step, the proteins undergo conformational rearrangement to form the native short-range interactions. Orientational restraints of the transient complex limit its formation to a “basal” rate constant ∼105 M−1s−1, but long-range electrostatic attraction can enhance the rate by three to four orders of magnitude [1]. In our recently developed transient-complex theory [1], the electrostatically enhanced association rate is predicted as ka = ka0exp(-<Uel>∗/kBT), where ka0 is the basal rate constant and <Uel>∗ is the average electrostatic interaction energy between the proteins in the transient complex. The predictive power of the transient-complex theory has been demonstrated over a large number of protein-protein and protein-RNA complexes [1-3]. To make our method widely accessible for predicting protein association rate constants, here we developed it into a web server. The “TransComp” server is available at http://pipe.sc.fsu.edu/transcomp/, with the structure of the native complex formed by two proteins as input. The generation of the transient complex and the calculation of ka0 and <Uel>∗ are done automatically. Test run of the server on 14 protein-protein complexes was carried out to demonstrate its capability. Association rate constants calculated by the server for these complexes agree well with experimental data, which range from 5.5 × 104 to 2.4 × 108 M−1s−1. We expect that the TransComp server will become a valuable tool for kinetic characterization of protein-protein and protein-nucleic acid association. [1] R. Alsallaq and H.-X. Zhou (2008) Proteins 71, 320-335. [2] S.B. Qin and H.-X. Zhou (2008) J. Phys. Chem. B 112, 5955-5960. [3] S.B. Qin and H.-X. Zhou (2009) Proc. Natl. Acad. Sci. USA 106, 6974-7979.