04/29/1996 07:00:00

Engineers Demonstrate Technique for Crystallizing Proteins

PASADENA—Chemical engineers have successfully demonstrated a new method of crystallizing proteins, an important and notoriously difficult problem in biochemistry. The new approach is reported by researchers from Caltech and the University of Washington in the April 30 issue of the Proceedings of the National Academy of Sciences.

Arranging protein molecules into a crystalline structure is an important scientific challenge; it is the crucial first step in studying a protein's molecular structure. Crystallization will also play an important role in building molecular devices such as biosensors or highly compact data storage mechanisms.

"We're excited about this method because it should provide a general mechanism for crystallizing a variety of both natural and engineered proteins," said Frances Arnold, associate professor of chemical engineering at Caltech and coauthor of the study. "Equally exciting, the new technique has possible applications in constructing molecular devices based on two-dimensional protein arrays." Arnold's Caltech colleagues in this work were graduate student Daniel Pack and research fellow Chao-Tsen Chen, both in chemical engineering.

The new approach uses specially designed lipids—fatty molecules— that act like molecular tugboats. Ordinarily, protein molecules in a liquid drift in disorder, but the lipids grab hold of the molecules and help assemble them into a two-dimensional crystalline layer.

Lipids are common biological molecules, but the special lipid used in this technique has been slightly altered by the addition of a metal-binding portion called iminodiacetate. The engineers then attached a metal ion, in this case copper (Cu2+), at one end. Almost all proteins contain the amino acid called histidine, to which metal ions such as Cu2+ bind. This binding between the metal ion in the lipid and the histidine in the protein is what enables the lipids to grab the protein molecules.

Researchers from the University of Washington collaborated for two years with the Caltech group to show that the new concept is valid. The Washington team—Viola Vogel, an assistant professor of bioengineering; Wolfgang Frey, a postdoctoral fellow; and William Schief, Jr., a graduate student in physics—used a special Brewster angle microscope to see how a naturally occurring protein, streptavidin, forms a two-dimensional crystal when bound to this lipid.

The presence of a histidine on the surface of streptavidin is essential for it to bind to the designer lipid. The histidine responsible for binding was identified with the help of Patrick Stayton, an assistant professor of bioengineering, and Ashutosh Chilkoti, a postdoctoral fellow in bioengineering, both at the University of Washington. They created and tested mutants of streptavidin that lacked histidine, and found that one of these variants did not bind to the lipid.

Lipids are "amphiphilic," which means that they simultaneously love and hate water. Part of the lipid (its long, oily tail) is repelled by water, while its metal-containing head is hydrophilic, or water-loving. Because of this amphiphilic property, the lipids line up in a monolayer—a film that is a single molecule thick—on the surface of water. The metal ions hang down into the water where they can grab the proteins, which are also hydrophilic, and pull them up to the lipid–water interface.

The lipids pack the molecules of protein—streptavidin in this case—into a dense, two-dimensional sheet. When the proteins are crowded together like this, they can spontaneously organize into large, two-dimensional crystals. Scientists can then use the two-dimensional crystals to seed three-dimensional ones.

"Scientists have used similar methods before to crystallize proteins in two dimensions, but in those cases they had to go to a lot of work custom-designing a lipid with just the right structure to bind to a particular protein," Arnold explained. "The beauty of the new technique is that the lipid-bound metal ions should bind to and crystallize many different proteins. It won't need to be redesigned for each new protein."

Written by John Avery