Solid-state nuclear magnetic resonance (SSNMR) is an important tool for the structure, function and dynamics study of many chemical and biological systems, especially powerful in studying membrane proteins, whose structures have been difficult to analyze by traditional x-ray crystallography or solution NMR techniques. In this thesis, various NMR techniques are used to study the structure and dynamics of membrane proteins within lipid bilayers.

The main technique applied in this thesis is spin diffusion experiments. We study the structural rearrangement upon membrane binding of colicin Ia by the proton-driven 13C spin diffusion (PDSD) 13C-13C 2D correlation experiment. Membrane bound colicin Ia turns out to have a more extended structure compared to the soluble state. Then a 1D 1H detected 1H spin diffusion experiment is developed to provide the same membrane protein topology information as the 2D 13C detected version, but with significant sensitivity enhancement. We demonstrated this new technique on the colicin Ia channel-forming domain and achieved about 200 fold time saving. Further, the data analysis method is developed to extract the intermolecular distance as long as 12 y from 19F spin diffusion experiment CODEX, where the oligomeric state is obtained at the same time. Demonstrated on the M2 proton channel system, this method is applied to extract the intermolecular distances between a key residue Trp41 in different states of the M2 proton channel. Finally, the water accessibility of the M2 proton channel in different states is studied by the 1H spin diffusion experiment and 3D low resolution models are proposed for this proton channel system by simulating the 1H spin diffusion process between the water and protein.

The second focus of this thesis is the dynamics of the M2 peptide in a complex membrane system. Compared to the single component model lipid bilayers, this composite membrane is shown to reduce the rotational rate of the membrane protein by 2 orders of magnitude, which is explained by a rotational diffusion model. The advantage of this immobilization is the ability to acquire high resolution SSNMR spectra at physiological temperatures.