Exploring pH Dependent Landscape Shifts of Proteins

Abstract

Molecular dynamics(MD) simulations of proteins at physiological versus lower pH lead to separate stable conformations that do not interconvert. Although modulating pH requires adjusting the ionization states of many residues, we find this effect is achievable by introducing external force on a single one[1]. This finding has far reaching implications for controlling protein conformational dynamics.We thus perform all-atom MD simulations of 200ns under a series of pH and ionic strength conditions to determine the conformational distributions of calmodulin and ferric-binding protein. For example, in accord with FRET experiments, in calmodulin we find that pH of 5.0 encourages a more compact conformation where the two lobes directly interact, while at 7.4 the lobes are distal, communicating through the linker region[3]. Time scale of the conformational change between the two states is measured on milliseconds[3], way beyond what is observable through MD.At a coarse-grained level where each residue is treated as a single node, we scan the protein to determine forces that may be inserted on specific residues to cause the observed conformational change. This method, which operates in the linear response regime, points to a single charged residue with upshifted pKa in both proteins[1,4], although the latter information is not encoded in the model.We implement these findings into steered-MD, where the determined external force acts on the resolved residue, and the protein readily lends itself to the more compact form. Thus, the coarse-grained approach is not only an efficient method determining the main residue whose interactions lead to conformational arrest, but also suggests alternative scenarios to overcome factors hindering barrier crossing.[1] Atilgan et al. J Chem Phys, in press (2011).[2] Atilgan et al. Ann Rev Biophys, to appear (2012).[3] Slaughter et al. Biochemistry (2005).[4] Atilgan et al. PLoS Comput Biol (2009).