Image: Courtesy of Yue-Ming Li
Many neuroscientists think that the master criminal behind Alzheimer disease is AB-secretase-42, the 42-amino-acid peptide that forms amyloid plaques in the brain. Two accomplices, the enzymes B-secretase-secretase and g-secretase, consecutively cleave AB-secretase from the much larger B-secretase-amyloid precursor protein (APP). What baffles investigators about g-secretase is that its substrate is a stretch of amino acids within APP's seemingly inaccessible transmembrane domain.
Many genetic and cellular studies suggest that g-secretase's active site resides in the transmembrane protein presenilin-1 (PS1), which was first reported in 1995.1 Then in 2000, Merck & Co. labs in the United States and Great Britain provided compelling biochemical evidence in a Hot Paper2 favoring that hypothesis.
Nevertheless, even Yue-Ming Li, who worked on this project and now heads the biochemistry and molecular pharmacology laboratory at Memorial Sloan-Kettering Cancer Center, acknowledges the need for much more data. "We don't have the final proof yet that PS1 is really the g-secretase." Indeed, some scientists argue that PS1 cannot be g-secretase.
The Merck study relied on a potent g-secretase inhibitor that the company had identified in a high-throughput screen. Citing diverse biochemical considerations, Li claims that the inhibitor almost surely binds to the enzyme's active site. (Definitive crystallography results are lacking, however.) The hardest step was appropriately modifying that inhibitor, recalls Stephen J. Gardell, who directed the biological chemistry department study at Merck Research Laboratories in West Point, Pa. (He now heads obesity research at Bayer Pharmaceutical Division in West Haven, Conn.) The altered inhibitor had to retain potency as it gained the ability to bind covalently to an adjacent protein during photoactivation.
The solution was to place a benzophenone group at the inhibitor's N- or C- terminus. Upon exposure to light, this benzophenone group's double-bonded oxygen formed a triplet biradical that abstracted hydrogen from a nearby protein, ultimately causing the two molecules to fuse.
Using a method they had just developed, the Merck researchers solubilized HeLa cell membranes with CHAPSO detergent, which preserves g-secretase activity. When the solubilized fraction was incubated with inhibitor and exposed to light, PS1 was the protein that bound. A Harvard Medical School group reported the same finding after applying a lower-affinity inhibitor that bound covalently to its substrate without photoactivation.3 The Merck study, however, was not airtight. PS1, says Gardell, might have merely been contiguous to g-secretase's active site and been mistakenly snared by the chemical bonds cast by the photoactived inhibitor.
The study certainly does not convince Sangram S. Sisodia, a neurosciences professor at the University of Chicago. He doubts that PS1 is g-secretase and bases his opinion on a host of observations, ranging from the stoichiometry of the substrate-enzyme reaction to the different sites of PS1 (endoplasmic reticulum) and g-secretase (cell surface).4 Other researchers would take issue with many of Sisodia's contentions, but the fact remains that no one has mixed purified PS1 and APP in a test tube and shown g-secretase cleavage.
One reason this crucial experiment has not succeeded might be that b-secretase appears to operate as a large complex of proteins, including nicastrin, APH-1, and PEN-2. "If you simply have naked PS1, there's no guarantee that it, by itself, is capable of catalyzing the proteolytic activity," says Gardell. For Sisodia, "The gold standard is really to functionally reconstitute g-secretase activity in a membrane preparation" after all the components of the complex's have been identified.
One problem is that PS1 is normally sliced into two pieces and operates only if both pieces are bound together. Yet methods to separate components of the g-secretase complex from PS1 could also tear apart the PS1 pieces, rendering the enzyme inactive. Consequently, Li is contemplating another approach. He plans to study a PS1 mutant associated with familial Alzheimer disease that is active though not bisected. He wants to determine whether the purified mutant will cleave substrates such as APP.
Douglas Steinberg (firstname.lastname@example.org) is a freelance writer in New York City.
1. R. Sherrington et al., "Cloning of a gene bearing missense mutations in early-onset familial Alzheimer's disease," Nature, 375:754-60, 1995.
2. Y.-M. Li et al., "Photoactivated b-secretase inhibitors directed to the active site covalently label presenilin 1," Nature, 405:689-94, June 8, 2000. (Cited in 181 papers)
3. W.P. Esler et al., "Transition-state analogue inhibitors of g-secretase bind directly to presenilin-1," Nature Cell Biology, 2:428-34, 2000.
4. S.S. Sisodia, P.H. St. George-Hyslop, "g-secretase, Notch, AB-secretase and Alzheimer's disease: where do the presenilins fit in?" Nature Reviews Neuroscience, 3:281-90, April 2002.