Structure Determination
Determining the structure of a molecule, especially a protein, is an impor­tant step in determining its function. In many diseases, changes in molecu­lar structure are involved in the pathologic process, and understanding these changes can help in the design of therapy. The three-dimensional arrange­ment of atoms in a molecule can be determined using a variety of physical techniques. For large biological macromolecules the most common ex­perimental techniques for structure determination include X-ray crystallog­raphy, electron microscopy, and nuclear magnetic resonance (NMR). Finally, approximate models depicting the three-dimensional arrangement of atoms can be built using computer modeling.
The principles behind the use of X rays for the determination of the structures of biological macromolecules are quite different from the use of “X rays” in the practice of medicine. A medical X-ray film shows a shadow revealing internal body parts depending on how easily the X rays penetrated them. In contrast, X-ray crystallography looks at how X rays are diffracted, or scattered, by the atoms in a sample and determines what the three­dimensional arrangement of the atoms must be to give rise to the observed pattern of scattering. (This is somewhat akin to determining the structure of a jungle gym by bouncing a tennis ball off it and recording the pattern of bounces.)
The distances between atoms in a molecule are very small, on the or­der of 10^-10 meters, and the wavelength of the radiation used to determine their relative positions must be correspondingly small. X rays have the nec­essary small wavelength. The amount of radiation scattered by one mole­cule is too small to measure; therefore, it is necessary to combine the diffraction from a large number of molecules. Crystals are used because they contain an ordered arrangement of many molecules. Computers are used to reconstruct an image of the molecules in the crystal. The technique of X- ray crystallography provides the most detailed and accurate information on the structure of biological macromolecules.
Electron microscopy uses an electron beam to study the structure of bi­ological materials. By using a magnet to focus electrons scattered from a sample, an electron microscope can form an image in a manner similar to a conventional microscope. One problem is that the electron beam used in such a microscope has a very high energy and can destroy sensitive biolog­ical samples. To aid in its preservation, the sample is often maintained at a very low temperature (this is called cryo-electron microscopy). As in the case of X-ray diffraction, it is an advantage to combine the electrons scattered from many molecules to get an average image. This can be done using or­dered samples such as two-dimensional crystals or by orienting and averag­ing many images. Electron microscopy is especially useful for large complexes of macromolecules.
Nuclear magnetic resonance is not a scattering technique, but a spec­troscopic technique that depends on the interaction of atomic nuclei with radio-frequency radiation and a magnetic field. This interaction is very sen­sitive to the environment surrounding an atom, and therefore can be used to determine what other atoms are nearby a given atom. Once the features in an NMR spectrum have been associated with specific atoms it is possi­ble to combine this experimental information of the local arrangement of atoms with knowledge of the chemical structure of a molecule to derive a three-dimensional structure. An advantage of NMR is that it examines mol­ecules in solution and can also provide information about their dynamic properties, or motion.
In addition to the experimental techniques for determining the three-dimensional structures of molecules, it is possible to use computational techniques  to  predict  structures.  The  most  successful  approach  to  structure prediction utilizes the observation that proteins with similar amino acid sequences have similar three-dimensional structures. This allows one to predict an approximate structure if a structure of a related protein is already known.  This  starting  point  can  then  be  combined  with  knowledge  of  the chemical structure and physical principles to improve the model.
References
Alberts, Bruce, et al. Molecular Biology of the Cell, 4th ed. New York: Garland Pub­lishing, 2000.
Branden, Carl, and John Tooze. Introduction to Protein Structure, 2nd ed. New York: Garland Publishing, 1999.
Stryer, Lubert. Biochemistry, 4th ed. New York: W. H. Freeman and Company, 1995.