Scientists have long puzzled over the workings of membrane-spanning proteases, the molecular scalpels that cleave other membrane proteins into smaller segments. Their interest is due, in part, to the fact that this group of enzymes includes γ-secretase, the protease responsible for producing the amyloid β-peptide that forms damaging plaques in the brains of Alzheimer's sufferers. Understanding how γ-secretase works may lead to ways to stop its function and hinder the disease. On page 179 of this issue, Ya Ha of Yale University School of Medicine, New Haven, determines the crystal structure of another intramembrane protease from Escherichia coli. Although unrelated to γ-secretase in amino-acid sequence, the protease he describes is likely to act in the same way.

Proteases cleave proteins into smaller segments using water molecules to break amino-acid bonds. Thus, they typically operate in a watery environment. But the proteases embedded in the cell's membrane are an exception. They do the cutting inside the cell membrane.

Ha heard about this group of enzymes as a postdoc at Harvard University five years ago while searching for a project for his own lab. The intramembrane proteases seemed to fit the bill. “This area was ideal for a structural biologist like myself because there is no other way to figure out how these proteases work,” he explains. “You need direct visualization to do so.”

After setting up his lab at Yale in 2001, he set out, with the help of Yongcheng Wang and Yingjiu Zhang, to determine the crystal structures of several intramembrane proteases, including γ-secretase. He followed standard laboratory practice, engineering the genes encoding these proteases so that they were expressed in bacteria. He purified each protein to make crystals from it, and then bombarded the crystals with X-rays to reveal the relative positions of the protein's atoms.

“We were stuck at almost every step of the process,” laughs Ha. “It was a risky project.” Most of the proteins he had started with had to be discarded at various stages of the purification process because they did not yield either sufficient amounts or sufficiently pure or stable protein. But he persevered. “Structural biologists have a number of tricks to use,” he says. In the end, he was able to crystallize a bacterial rhomboid protein called GlpG.

With crystals in hand, he quickly obtained the protein's structure. Ha discovered that the amino acids involved in cleaving target proteins reside in a cavity smack in the middle of the protein, along with several water molecules. Although the membrane keeps water out, inside the enzyme there's a little watery pool where the proteins are cut. In the absence of a protein, the opening to this cavity is blocked by a string of amino acids shaped into a loop, which may act as a molecular gate.

Ha plans to use his experience to unravel the structure of other intramembrane proteases, including γ-secretase. “Now we know how difficult the project is, but that it can be done,” he says, while acknowledging that the process may not be any easier the second time around. “I have been in the field long enough to know that experience does not always translate into making things go faster.”