Cholesterol-Dependent -Secretase Activity in Buoyant Cholesterol-Rich Membrane Microdomains Suzanne Wahrle,* ,1 Pritam Das,* ,1 Andrew C. Nyborg,* Chris McLendon,* Mikio Shoji, † Takeshi Kawarabayashi,* ,† Linda H. Younkin,* Steven G. Younkin,* and Todd E. Golde* ,2 *Mayo Clinic Jacksonville, Department of Neuroscience and Pharmacology, 4500 San Pablo Road, Jacksonville, Florida 32224; and † Department of Neurology, Gunma University, Gunma, Japan 371-8511 Received August 20, 2001; revised November 1, 2001; accepted for publication November 28, 2001 Buoyant membrane fractions containing presenilin 1 (PS1), an essential component of the -secretase complex, and APP CTF,a -secretase substrate, can be isolated from cultured cells and brain by several different fractionation procedures that are compatible with in vitro -secretase assays. Analysis of these gradients for amyloid  protein (A) and CTF production indicated that -secretase activity is predominantly localized in these buoyant membrane microdomains. Consistent with this localization, we find that -secretase activity is cholesterol dependent. Depletion of membrane cholesterol completely inhibits -secretase cleavage, which can be restored by cholesterol replace- ment. Thus, altering cholesterol levels may influence the development of Alzheimer’s disease (AD) by influencing production and deposition of A within cholesterol rich membrane microdomains. © 2002 Elsevier Science (USA) Key Words: A; Alzheimer’s disease; -secretase; presenilin; lipid raft; cholesterol. INTRODUCTION There is compelling evidence that abnormal accu- mulation of the amyloid  protein (A) plays a causal role in the development of Alzheimer’s disease (AD) (Hardy, 1997). Since inhibiting A production could be therapeutic, the proteases that produce A, as well as the factors that regulate their activity, are major targets for drug discovery. These proteases produce A from the amyloid  protein precursor (APP) through two sequential cleavages. APP is first cleaved by -secretase, a transmembrane aspartyl protease, at the amino terminus of A to generate a large secreted derivative (sAPP) and a membrane bound APP car- boxyl terminal fragment (CTF). Subsequent cleavage of CTF by -secretase results in production of the A and CTF. In a second pathway, APP is cleaved within the A sequence by -secretase, which generates an- other large secreted derivative (sAPP) and CTF (re- viewed in (Golde and Younkin, 2001)). A great deal of biochemical and genetic evidence now supports the concept that the polytopic mem- brane proteins presenilin 1 (PS1) and presenilin 2 (PS2) are -secretases, though definitive proof of this will require reconstitution of the proteolytic activity (reviewed in (Golde and Younkin, 2001)). In any case, it appears that PS1 serves as a biochemical marker for -secretase activity, as -secretase activity cofraction- ates with a high molecular weight presenilin-contain- ing complex and immunoprecipitation of PS1 results in immunoprecipitation of -secretase activity (Li et al., 2000). In this study, we have carried out a series of experiments to explore the possibility that -secretase activity localizes to buoyant membrane microdomains where PS1 and APP CTFs have been previously shown to reside (Lee et al., 1998). These studies dem- 1 These authors contributed equally to this work. 2 To whom correspondence and reprint requests should be ad- dressed. Fax: (904) 953-7370. E-mail: tgolde@mayo.edu. Neurobiology of Disease 9, 11–23 (2002) doi:10.1006/nbdi.2001.0470, available online at http://www.idealibrary.com on 0969-9961/02 $35.00 © 2002 Elsevier Science (USA) All rights reserved. 11