Approaches to Macromolecular and Supramolecular Structure Determination

Citation

Agard, David Andrew (1981) Approaches to Macromolecular and Supramolecular Structure Determination. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/9psf-0a04. https://resolver.caltech.edu/CaltechTHESIS:03022018-140501638

Abstract

All of the common techniques for obtaining structural information from macromolecular and supramolecular objects suffer from one type of drawback or another. X-ray and neutron diffraction methods have the phase problem (only the amplitude and not the phase angle of the complex diffraction vector can be experimentally observed). Low-dose 3-dimensional electron microscopy suffers from the inability to collect a complete data set due to instrumental limitations. Although light microscopy allows the sample to be kept under biological conditions (compare with standard electron microscopic techniques), the low resolution has limited its appeal to more qualitative aspects of ultrastructure analysis.

All of these limitations can be considered as a lack of information in the Fourier space domain. A generalized mathematical approach is presented where global real-space constraints are utilized to ameliorate this loss of information. In simple cases (one-dimensional analysis) it is possible to solve the phase problem, and in more complex cases (two- and three-dimensional analyses) the added real-space information is used to augment the experimentally derived data. The iterative Fourier refinement scheme was chosen as it represents the most versatile means for incorporating global knowledge concerning the real-space behavior of the object.

This approach was used to solve the three-dimensional x-ray structure of the snake polypeptide neurotoxin α-Bungarotoxin, the one dimensional transmembrane electron density profile. for acetylcholine receptor containing membranes, and to solve the missing cone problem of electron microscopic structure analysis.

Real-space constraints were also incorporated into an iterative deconvolution scheme used both for image processing and for the quantitative analysis of overlapping gel electrophoretogram peaks. None of this work would have been possible if the information supplied by real-space constraints had not been utilized . The approaches presented to these problems are general ones and should be applicable to other systems.

Thesis Files