Novel Computational Techniques to Expand the Scope of Protein Nuclear Magnetic Resonance Spectroscopy

点击翻译按钮获取中文摘要

Project Summary: Understanding the structure and function of biological macromolecules is critical in countless biomedical disciplines, including cancer biology, drug design, and nanotechnology. It is often essential to understand molecular etiology to interpret a clinical presentation as well. Nuclear Magnetic Resonance Spectroscopy (NMR) is one of the principal techniques for investigation of protein structure and it is the primary technique for understanding the biology of proteins that lack fixed three-dimensional structures – termed intrinsically disordered proteins (IDPs) – a group that includes numerous proteins involved in biomineralization, cell signaling, and nucleic acid binding. However, NMR spectroscopy suffers from limitations that restrict the size and scope of proteins and IDPs that it can be used to investigate. The broad goal of this proposal is to develop and characterize improved techniques for analyzing NMR data to expand the set of feasible protein targets. One central limitation of NMR is the inherent resolution/sensitivity tradeoff in which resolution (the ability to discriminate signals with similar frequency) can be enhanced only by sacrificing sensitivity (the ability to distinguish signal from noise), or vice versa. Generally, an NMR spectroscopist may try to overcome these limitations through preparation of isotopically labeled samples or by using powerful spectrometers and sophisticated multidimensional experiments. Various mathematical manipulations can be applied to the raw data for further enhancement of sensitivity or resolution. Although useful, these techniques ultimately force a tradeoff between sensitivity and resolution in one way or another. Maximizing both resolution and sensitivity is critical in the biological applications of NMR, and therefore investigation of techniques with the potential to simultaneously enhance both is necessary. I have generated preliminary data, which strongly suggests that an innovative data processing technique called Maximum Entropy Reconstruction with linewidth deconvolution (deconvolution) may bypass the tradeoff by simultaneously enhancing resolution and sensitivity in multidimensional NMR spectra. Deconvolution functions by reducing signal overlap and scaling down spectral noise. This proposal details the first systematic comparison between conventional data processing techniques and deconvolution. I will conduct this comparison using a tripartite research strategy by first testing deconvolution in a precisely designed control scenario, in which the ideal outcome is known. Then I will quantify the abilities of deconvolution in unknown situations and finally I will use deconvolution to determine a protein structure and demonstrate its practical benefits. The quantitative results of these studies will definitively determine if deconvolution provides simultaneous enhancement of resolution and sensitivity. It would constitute a breakthrough for NMR spectroscopy and structural biology if deconvolution provides the benefits suggested by my data. Deconvolution is a cutting-edge technique that is inexpensive to implement and has the potential to provide the necessary spectral improvements for studying previously intractable proteins and IDPs.