A number of innovations are dramatically reducing the amount of time required to determine three-dimensional structures of biomolecules by X-ray crystallography. For example, on page 458 of this issue of Nature Structural Biology, Perrakis and colleagues present an automated method for model building into an electron density map. However, the first steps in any such enterprise are purification of the molecule of interest and crystallization, and despite the availability of crystallization 'kits', obtaining high-quality crystals can still be rate-limiting in many cases.

In his recently published book Crystallization of biological macromolecules (Cold Spring Harbor Laboratory Press; 1999), Alexander McPherson outlines the science of protein crystallization during its infancy. Perhaps not surprisingly, the first documented crystallization was of hemoglobin from the blood of earthworms, which was accomplished by F.L. Hünefeld in 1840. Other proteins, such as the reserve proteins of plant seeds and hen egg albumin, were also crystallized during the late 1800s, but some of the more notable achievements occurred during the early 1900s. One such accomplishment was that of J.B. Sumner, who was intent on purifying an enzyme to prove that enzymes were indeed proteins — a subject of debate at the time. Success came in 1926 when he crystallized, and thus purified, jack bean urease, and proved its protein nature. (Interestingly, in 1919 Sumner had already crystallized another protein, concanavalin B, which was not known to have enzymatic activity at that time but is now known to be a chitinase.) In 1927, insulin was crystallized by J.J. Abel and colleagues and in 1937, lysozyme was crystallized by E.P. Abraham and R. Robinson. Both proteins were destined to become, along with hemoglobin and myoglobin, among the first proteins with known three-dimensional structures. During the 1930s and 1940s, other enzymes such as trypsin and trypsinogen, chymotrysin and chymotrypsinogen, pepsin and pepsinogen, were crystallized by J.H. Northrup, R.M. Herriott, and M. Kunitz. For their work on crystallizing enzymes, Sumner and Northrup shared the 1946 Nobel prize along with W.M. Stanley, who crystallized tobacco mosaic virus in 1935.

Scientists in the field of X-ray diffraction took advantage of their colleagues' high quality crystals to study protein structures. For example, in 1934, J.D. Bernal and Dorothy Crowfoot (later Dorothy Hodgkin, see the book review on page 412 of this issue) presented X-ray photographs of pepsin, and a year later, Crowfoot published similar work with insulin. Today, although technological advances are making it possible to obtain useful diffraction data from previously unusable crystals, the science of crystallization will still play a major role in the success of future large scale efforts, such as structural genomics projects.