2D IR spectroscopy and computational modeling : application to protein folding and binding

Abstract

In this thesis, dynamics experiments are developed that can be used to study protein conformational changes such as folding and binding. Every functional motion of a protein is inextricably linked to conformational dynamics. However, most of our insight into protein folding and binding is indirectly obtained through kinetics experiments that measure reaction rates and reveal how fast populations of stable states interconvert. Two-dimensional infrared spectroscopy (2D IR) is the central tool developed in this thesis for protein dynamics experiments due to its combination of time and structural resolution. As a vibrational spectroscopy, 2D IR potentially offers femtosecond time resolution. Its advantages over linear, absorption spectroscopy come through correlating excitation and emission frequencies to allow for a separation of homogenous and inhomogeneous line shape components, and to give rise to structurally sensitive cross-peaks. One general problem was repeatedly addressed in this thesis: how can 2D IR spectra best be modeled to reveal atomistic structural information? The key feature that now sets 2D IR apart from other fast protein probes is that the data can readily be calculated from an atomistic structure or molecular dynamics simulation using the methods developed in this thesis work. Demonstrative applications are presented for the amide 1-11 spectroscopy of NMA, the amide 1'-II' spectroscopy of poly-L-lysine, isotope-edited 2D IR spectroscopy of trpzip2, and transient 2D JR spectroscopy of ubiquitin unfolding after a temperature jump. The emerging paradigm is to interpret 2D IR spectra with the aid of an atomistic, molecular dynamics simulation. The applications to protein binding use the monomer-dimer transition of insulin as a model system. Using a combination of experiments and simulations, this equilibrium was characterized as a function of protein concentration, temperature, and solvent. Finally, as a complement to the structural information provided by 2D IR, dye-labeling and intrinsic tyrosine fluorescence experiments on insulin are described.