By David N. Leff

Every crime buff knows by heart how a well-documented murder story unfolds: First, the body, cordoned off by police. Then the pathologist, scoping the corpse for cause of death, contents of stomach, signs of violence, DNA clues. Next, the detective, beating the bushes for witnesses, motive, suspects, material evidence. Finally, the trial - arraignment, indictment, jury selection, defendant, defense, prosecution, verdict, sentence.

Many of these elements fit the death of neurons implicated in Alzheimer's disease (AD). Wrapped around these brain cells are the smoking-gun bullets - the amyloid-beta plaques - that are the crucial evidence of AD. The molecular weapon that fires these fatal missiles is a lengthy chain of 700 amino acids called bAPP - beta-amyloid precursor protein - which lurks holed up on human chromosome 21.

Pulling the triggers that cleave off the stubby, sinister, amyloid-beta fragment from bAPP are a pair of co-conspiratorial proteolytic enzymes called beta-secretase and gamma secretase. Gamma had been known for some time, but beta-secretase was identified only a year ago by biochemist Jordan Tang, at the Oklahoma Medical Research Foundation in Oklahoma City, where he heads the Protein Studies Program.

Tang is senior author of a paper in the current issue of Science, dated Oct. 6, 2000, titled: "Structure of the protease domain of memapsin 2 (b-secretase) complexed with inhibitor."

"In this Science article," Tang told BioWorld Today, "we reported how we determined the crystal structure of beta-secretase bound to a potent inhibitor. So that structure is available now for the first time to drug designers. What we can learn from it is that there is interaction of inhibitor with the active site of beta-secretase - which we called memapsin 2." That funky word, he explained, "is named after membrane-anchored aspartic protease - formerly known as pepsin.

"Memapsin-2's activity," he said, "is cutting bAPP. We reported cloning and identifying it in February 2000 in PNAS, where we demonstrated that it is actually the beta-secretase that people have been looking for for some time."

Tang continued: "When memapsin 2 cleaves bAPP at the beta-secretase site, it calls on that second enzyme, gamma-secretase, to cut at a different site. And these two cuts together make the spin-off amyloid-beta peptide. This accumulates in some brains in the form of senile plaques, which then cause dementia and the onset of AD. If you do not have this cleavage of the beta-secretase," he pointed out, "you shouldn't have amyloid-beta produced. And then presumably one can stop the AD progression - inhibit it." (See BioWorld Today, June 12, 2000, p. 1.)

"Almost every major pharmaceutical company has a very active program on this project," Tang observed. "They have been searching and screening their compound chemical libraries for this protease inhibitor. I understand," he added, "that many of them have found some lead inhibitor compounds, and are in the process of trying to improve their potency and other pharmacokinetic properties."

Atomic Pattern Offers Industry Leg Up

"This is where our crystal structure comes in," Tang said. "In order to design their lead compound into a better inhibitor, suitable for a drug against AD, they need to know the 3-dimensional structure of the enzyme, as well as how their own inhibitor will bind to its active site. So we feel this structure in Science will be able to help everybody move forward toward the goal of making drugs."

Tang described the molecular mechanism underlying potential inhibitors of plaque-forming amyloid-beta. "Most proteases," he pointed out, "such as serine protease, thioprotease and metalloproteases, are working in either neutral or slightly alkaline pH. Aspartic proteases specialize in more acidic media, in the lysomes, endosomes and the stomach - that kind of environment.

"They also utilize a very different route to do the cutting of the protein. In this case, there are two aspartic acids in the active site. So their mechanism of hydrolysis - target of inhibition - is somewhat different from other groups of proteases, such as the antiretroviral proteases against the AIDS virus.

"The inhibitor we created in April of this year, and the one that binds the active site," Tang observed, "are very potent, but they are not yet designed to go to the brain, or have other pharmacological properties."

As the first step toward their structural analysis, Tang and his co-authors cloned the protease domain of the human memapsin 2 gene and - using recombinant DNA technology - put it into the genome of Escherichia coli bacteria. "Then," he recounted, "we asked E. coli to make recombinant memapsin 2 enzyme as a protein, which we isolated from the bacterial culture. We put the inhibitor together with the protein, to bind very tightly, and crystallized this protein complex. Then we used X-ray crystallography to solve that structure at 1.9 Angstroms, which is very high resolution."

To connect the dots between this atomic structural evidence and real-life clinical pathology of Alzheimer's disease, Tang described the situation of an extended family in Sweden, afflicted with inherited early-onset AD.

Two Mutations Speed Up AD Symptoms

"The Swedish family members have early onset of AD," he noted. "It usually starts at 60 years of age, or older. These familial types may start at 50 or 55. One of the pedigrees in Sweden has two mutations - amino acid replacements - in their bAPP next to the site where memapsin 2 cuts. That change causes the enzyme to cleave the precursor protein more rapidly. So when it cuts faster, more amyloid-beta is produced, which leads to earlier onset of AD.

"This is so," he explained, "because the chain of two amino acids makes the substrate bind much more favorably in the active site. We were able to demonstrate the interaction, and give the reason why the change of the structure caused different enzymic efficiency."

Connecting the final dot, Tang observed, "We've been trying to use this information to improve our inhibitor. That's what this structure is for. It gives you a road map of what the active site looks like. Now we can design better inhibitors. By publishing that, we give this structure to the scientific world, so they can do the same." n