An efficient computational method for prediction of protein binding modes in multiprotein systems

Wednesday, September 24, 2014 — Poster Session IV
10:00 a.m. –12:00 p.m.	FAES Academic Center	NINDS	COMPBIO-3

* FARE Award Winner

Authors

• A. Cardone
• S.A. Hassan
• M. Brady
• R.D. Sriram
• H.C. Pant

Abstract

Understanding binding mechanisms in solution is essential to elucidate important physiological and pathological functions. Weak and ultra-weak binding modes play an important role in biological systems, especially in the crowded intracellular environment. These modes are difficult to detect with conventional NMR techniques and are a major challenge to sampling methods. In addition, accurate solvent effect computation is needed to characterize important physicochemical quantities such as binding affinity. We implemented a configurational-bias Monte Carlo (CBMC) technique to sample the spatial configuration of a multiprotein system efficiently, and to overcome important limitations of conventional sampling techniques. The method consists of an MC optimization of a conveniently-defined norm, which yields a statistical distribution of first-encounter binding modes. Adaptive bias MC simulations are then used based on this distribution, leading to the binding modes at equilibrium. The method accounts for conformational selection and induced fit. The method is used to predict the inhibitory mechanism of the pathological activator p25-induced cdk5 aberrant hyperactivity by a 33-residue peptide derived from p35, the endogenous cdk5 activator. We illustrate the entire computational protocol, starting from the sequence-to-structure prediction of the ligand conformations in solution to the final mode of binding of the complex at physiological conditions.

back to top