Improving experimentation and interpretation in solution nuclear magnetic resonance investigations.

Abstract

In the past three decades, roughly 100 fundamentally different Nuclear Magnetic Resonance (NMR) experiments have been developed for the determination of protein and nucleic acid structures, interactions, dynamics and kinetics. These experiments are now widely and successfully used in structural biology, proteomics, bioinformatics, metabolomics, and pharmaceutical sciences. However, in seeking to quantitatively understand biomolecules one must consider that the NMR experimentation may perturb the molecular biophysical quantity sought. Moreover, limitations on available instrument time unfavorably affect the validity of statistical analysis of the physical quantities obtained. Some of these limitations and concerns are considered and explored herein. In NMR relaxation experimentation, sample heating caused by the applied NMR pulses themselves is a major problem that can seriously bias the values of the dynamical parameters derived. A compensation and saturation scheme for constant sample heating is proposed and statistically evaluated. It is concluded that the scheme allows for reproducible, robust and rapid acquisition of NMR spin relaxation data sets. Molecular dynamics parameters derived from transverse Carr-Purcell-Meiboom-Gill relaxation experiments are seriously affected by the NMR frequency setting with respect to the protein's resonance frequencies. A novel phase cycle scheme is proposed and statistically evaluated; it is shown that it can greatly improve the relative reliability of these measured rates. Experimental Residual Dipolar Coupling (RDC) and structural data for large proteins have a high degree of uncertainty because of the intrinsic low sensitivity of the experiments. Herein it is systematically explored how these errors propagate through commonly used RDC analysis software and how they affect the derived structural and dynamical quantities. The RDC analysis is demonstrated to be remarkably robust even in unfavorable cases of much structural and experimental uncertainty. It is concluded that the RDC method is valid and powerful towards elucidating orientational information on large biomolecular systems and complexes. The issues described are just a few imperfections in NMR spectroscopy that are addressed in this thesis. These investigations help to address the confidence and reliability of conclusions drawn from these NMR techniques.