Crystal Reveals Unexpected Beginnings.

Birth is a tough process, particularly for crystals. Before a crystalline structure forms, atoms or molecules mingle in a solution, making and breaking attachments. After enough particles gather in a cluster, and surpass a critical size, crystal growth speeds along.

The initial clumping, called nucleation, helps determine a crystal's size, composition, and other properties. Scientists have developed theories about nucleation for decades, but they hadn't actually witnessed nuclei forming in solution.

At last, researchers report they've directly observed this early step in protein crystallization, and they were surprised at what they saw. The nucleus doesn't have the form that they expected, they report in the Aug. 3 NATURE.

Peter G. Vekilov and Siu-Tung Yau at the University of Alabama in Huntsville used an atomic force microscope to view the protein apoferritin as it crystallizes. Previous efforts at direct observation of nucleation in solution were thwarted because the crucial process can take place anywhere in the liquid and it happens very quickly.

To overcome these challenges, Vekilov and Yau waited for apoferritin clusters to approach the bottom of a glass cell where the researchers already had placed an apoferritin crystal as a docking point. When a cluster stuck, the team could study it at a resolution high enough for discerning individual molecules and for determining whether the cluster had the characteristics of a nucleus.

"The whole design of the experiment is very clever," comments David W. Oxtoby of the University of Chicago. "You're actually seeing the crucial step of something when it is first deciding, `Okay, yes, I will become a crystal.'"

At the critical moment, a nucleus can gain or lose a molecule with equal probability. Vekilov and Yau observed clusters immediately before and after this point. "We have some aggregates which are smaller than the nuclei and some which are larger," says Vekilov.

They found that all of these clusters had roughly the same shape--a flat layer of apoferritin molecules with a few more molecules on top. Each such cluster looked like a raft containing some 20 to 50 molecules.

"Since they all have this raftlike shape, we interpolate that the nucleus is also raftlike," says Vekilov.

Nucleation theory has assumed that atoms and round molecules form spherical nuclei. Since apoferritin is roughly spherical, it therefore ought to form spherical nuclei as well. "Obviously, this is not the case," says Vekilov.

"Everyone has always assumed that these would be little round balls, and therefore all of the theoretical predictions of when you would see crystallization were based on these models," says Oxtoby.

"Now, you have to doubt everything," agrees Vekilov. Researchers can't just assume they know the structure of the nucleus of any crystallizing material, he says. And, he adds, more experiments are needed to verify or modify models of how solids form.

"There are many, many technological and health-related issues in which you want to control the rate of nucleation," says Vekilov. Protein crystallization plays a key role in diseases such as sickle cell anemia and cataracts, and researchers use crystallization when developing drugs. Scientists also grow crystals to determine a protein's structure through X-ray crystallography.

What's more, understanding crystallization could help engineers improve the strength of materials such as metals. "If you can control nucleation, you can control the properties of the metal," says Vekilov. And that, he adds, could lead to many technological improvements, such as stronger turbine components for jet engines.