Báo cáo khoa học: Selecting cells with different Alzheimer’s disease c-secretase activity using FACS Differential effect of presenilin exon 9 deletion on c- and e-cleavage doc

Selecting cells with different Alzheimer’s disease c-secretase activity using FACS Differential effect of presenilin exon 9 deletion on c- and e-cleavage M.. Specific c-secretase inhibito

Selecting cells with different Alzheimer’s disease c-secretase activity using FACS

Differential effect of presenilin exon 9 deletion on c- and e-cleavage

M Fleur Sernee1, Genevie`ve Evin1, Janetta G Culvenor1, Jose´ A Villadangos2, Konrad Beyreuther3, Colin L Masters1and Roberto Cappai1

1 Department of Pathology, The University of Melbourne and The Mental Health Research Institute, Parkville, Victoria,

Australia; 2 The Walter and Eliza Hall Institute of Medical Research, Parkville, Victoria, Australia; 3 Center for Molecular Biology, ZMBH, University of Heidelberg, Heidelberg, Germany

The ultimate step in Alzheimer’s disease Ab generation

involves c-secretase, which releases Ab from its

membrane-bound precursor A similar presenilin-dependent proteolytic

activity is implicated in the release of the Notch intracellular

domain We have developed a novel assay for c-secretase

activity based on green fluorescent protein detection This

involves cotransfection of a substrate-activator based on the

amyloid precursor protein or the Notch sequence and a

fluorescent reporter gene Stable fluorescent cell populations

were selected by fluorescent activated cell sorting and

char-acterized This assay enabled the identification and sorting of

populations, which differ in their levels of c-secretase

acti-vity, with high fluorescent cells producing more Ab than low

fluorescent cells Specific c-secretase inhibitors, L-685,458 and MW167, reduced cell fluorescence in a dose-dependent manner that paralleled inhibition of Ab secretion Overex-pression of presenilin 1 increased the cell fluorescence Cells expressing presenilin with different aspartate mutations (D257A, D385A and D257A/D385A) or exon 9 deletion mutation showed reduced fluorescence The single aspartate mutations showed a concomitant reduction in Ab secretion, whereas the D257A/D385A and DE9 mutations had no effect on Ab secretion

Keywords: secretase; amyloid precursor protein; Notch; presenilin; fluorescence-assay

b-Amyloid (Ab) is the major constituent of Alzheimer’s

disease (AD) amyloid plaques and plays a key role in the

pathogenesis of AD The 4.5-kDa Ab peptide, is derived

from the type I integral membrane amyloid precursor

protein (APP) [1–3] Several groups have identified the

b-secretase activity that releases the N-terminus of Ab, as a

membrane-anchored aspartyl protease termed

b-site-APP-cleaving enzyme (BACE) [4] Cleavage of APP by BACE

generates a sAPPb ectodomain and a 99 amino acid

C-terminal fragment containing Ab (termed bCTF, C99 or

A4CT) that remains anchored in the membrane This bCTF

is cleaved by c-secretase, to produce Ab and the APP

intracellular domain (AICD or eCTF) that is released into

the cytosol [5–8] c-Secretase has a loose sequence specificity

as it can cleave its substrate at various sites to produce Ab

peptides of different lengths [9–11] c-Secretase activity is associated with a high-molecular weight complex that includes presenilin 1 (PS1) or presenilin 2 (PS2), nicastrin and PEN-2 [12–14] APH-1 is also required for c-secretase activity and may be part of the complex [15–17] Cells expressing PS1 with artificial mutations of the aspartate residues, amino acids 257 and 385, within the predicted transmembrane domains 6 and 7 show reduced c-secretase activity [18,19] This has led to the suggestion that PS’s are unusual aspartyl proteases The Notch family of type I membrane proteins are also processed within their trans-membrane domains, at site 3 (S3) and site 4 (S4) by a c-secretase-like activity that requires presenilin expression [20–24] Cleavage of Notch transmembrane domain releases the Notch intracellular domain (NICD), which traffics to the nucleus This event is critical for the function of Notch in the regulation of cellular proliferation and differentiation [25], therefore therapeutic approaches based on c-secretase inhibition will have to be selective for APP and should not alter Notch signaling

c-Secretase assays have generally been based on the detection of Ab secreted from cell culture media Recently, several groups have developed cell-free assays in which c-secretase activity was measured by detecting Ab with an enzyme-linked immunosorbent assay or by visualizing Ab and the corresponding 7-kDa CTF by immunoblotting In this report, we describe the development of a GFP-based cell fluorescence assay that is specific for c-secretase cleavages of either APP or Notch This assay involves cotransfection of eukaryotic cells with a substrate-activator

Correspondence to Dr Roberto Cappai, The University of Melbourne,

Parkville, Victoria 3010, Australia.

Fax: + 61 38344 4004, Tel.: + 61 38344 5882,

E-mail: r.cappai@unimelb.edu.au

Abbreviations: AD, Alzheimer’s disease; APP, amyloid precursor

protein; Ab, amyloid b protein; BACE, b-site-APP-cleaving enzyme;

CTF, C-terminal fragment; ECL, enhanced chemiluminescence;

FAD, familial Alzheimer’s disease; FACS, fluorescence activated cell

sorter; FBS, fetal bovine serum; GFP, green fluorescent protein; IP,

immunoprecipitation; NICD, Notch intracellular domain; NP-40,

nonidet P-40; PS1, presenilin 1; PS2, presenilin 2; WT, wild type.

(Received 15 October 2002, accepted 29 November 2002)

and a reporter gene The substrate-activator construct

mimics c-secretase substrates (based on Notch or APP)

fused to a transcription factor Upon proteolytic processing

by c-secretase the activator domain is released and promotes

expression of a green fluorescent protein (GFP) reporter

gene Thus c-secretase activity can be monitored by

measuring cell fluorescence We show the application of

this assay to the testing of c-secretase inhibitors and PS1

mutants This assay allows for the first time the selection of

cell populations and single cells by FACS based on their

differences in c-secretase activity that correlates with their

level of fluorescence Therefore this assay could be applied

to the screening of cDNA libraries to identify genes that

modulate c-secretase activity

Experimental procedures

DNA constructs

The APP-based substrate-activator plasmids: pcDNA3.1+/

SP-A4DCT-GV and pIRESpuro2/SP-A4DCT-GV Both

constructs consisted of the signal peptide (SP) of APP and

the APP695 (amino acids 597–653) sequence (A4DCT)

in frame with the GAL4 (G) and VP16 (V) sequences

SP-A4DCT (259 bp) was PCR amplified from pCEP4/

SP-A4CT (gift from S F Lichtenthaler, ZMBH, Germany),

using primer 1 (5¢-CCCAAGCTTGGGTGCCCCGCGC

AGGGTCGCG-3¢) and primer 2 (5¢-GTACTGTTTCTT

CTTCAGCATCACC-3¢) The GAL4-VP16 DNA

frag-ment (678 bp) was produced by PCR from pGAL4-VP16

[26] (a gift from G E O Muscat, University of Queensland,

St Lucia, Australia) with primer 3a (5¢-GGTGATGCTG

AAGAAGAAACAGTACATGAAGCTACTGTCTTC

TATCG-3¢) and primer 4 (5¢-GCTCTAGAGCTTCAC

fragments were mixed in equal molar concentrations

for splice-overlap PCR and cloned into pCDNA3.1+

(Invitrogen) or into pIRESpuro2 (Clontech) to enable the

control of gene transcription by adjusting the concentration

of puromycin

The Notch-based substrate-activator plasmid:

pCDNA3.1+/SP-NOTL-GV PCR was performed with

primer 3b (5¢-CCCAAGCTTATGAAGCTACTGTC

TTCTATCG-3¢) and primer 4 to amplify the GV-DNA

from the pCDNA3.1/SP-A4D CT-GV activator construct

The GV-DNA fragment was cloned into pBS-SP-NOTL, a

Notch construct in pBS(SK+) that consisted of the signal

peptide of APP and amino-acids 1648–1927 of human

Notch-1 (containing the S1 and S2 cleavage sites, but not the

entire N-terminal domain) This construct was kindly

provided by C Bergmann and T Hartmann (ZMBH,

Germany) The resulting SP-NOTL-GV was subsequently

cloned into pCDNA3.1+

The reporter plasmid: pSP72/5GAL-E1b-EGFP The

p5Gal-E1b-CAT plasmid ([27], kindly provided by G E

O Muscat) served as a PCR template to obtain the

5GAL-E1b-TATA promotor region (214 bp) using primer 5

(5¢-CCCAAGCTTGGGCATGCCTGCAGGTCGGAG-3¢)

and primer 6 (5¢-TTTAGCTTCCTTAGCTCCTGA-3¢)

The EGFP-DNA (750 bp) was amplified from

pSP64TK-EGFP (obtained from H Clarris, University of Melbourne, Parkville, Australia) with primer 7 (5¢-TCAGGAGCTAA GGAAGCTAAAATGGTGAGCAAGGGCGAG-3¢) and primer 8 (5¢-CCGCTCGAGTTACTTGTACAGCTCGT CCATGCC-3¢) The splice-overlap PCR-product was cloned into pUC18 (NEB) and subsequently cloned into pSP72 The hygromycin resistance gene was amplified from the pCEP4 plasmid (Invitrogen) with primer 9 (5¢-GGACCAGACCCCACGCAACG-3¢) and primer 10 (5¢-GCCCTGCTTCATCCCCGTGG-3¢) and cloned into the pSP72/5GAL-E1bEGFP construct at the NdeI site The presenilin constructs, pIRESpuro2/PS1 WT, PS1 D257A, PS1 D385A, PS1 D257A/385 A and PS1 DE9 All PS1 constructs were cloned into the pIRESpuro2 plasmid To obtain PS1 WT, RNA was extracted from SH-SY5Y cells with TRIzol (Life Technologies), cDNA was produced with the RNA PCR Core kit from Perkin Elmer (Roche) and primers 11 (5¢-CTAGCTAGCATGACAGA GTTACCTGCACC-3¢) and 12 (5¢-ATAGTTTAGCG GCCGCTAGATATAAAATTGATGGAATGC-3¢) were used to amplify presenilin DNA DNA sequencing revealed that some clones contained the sequence for the four amino acids (VRSQ) in exon 3 while some did not We used the PS1 WT containing the VRSQ sequence We used this construct to create the D385A mutation using the Quik-ChangeTMXL Site-Directed Mutagenesis Kit (Stratagene) PS1 D257A and PS1 D257A/D385A, were obtained from

A Weidemann and F Reinhard (ZMBH, Heidelberg, Germany) [8] and transferred from pCEP4 into pIRE-Spuro2 The PS1 DE9 DNA, was PCR amplified with primers 11 and 12 from pCDNA3.1/PS1 DE9 with an N-terminal flag sequence (kindly provided by F Reinhard) and was cloned into pIRESpuro2 We thereby removed the N-terminal flag sequence The last three mutations did not contain the VRSQ-sequence and therefore their mutations would be at position D253A and D381A, but we have kept the nomenclature similar to what is published in the literature to avoid confusion

Cell culture and transfection COS-7 cells were maintained in DMEM with high glucose (Life Technologies), and CHO cells were grown in

RPMI-1640 (ICN), supplemented with 10% (w/v) fetal bovine serum (FBS) (CSL, Parkville, Australia) and penicillin (50 UÆmL)1)/streptomycin (50 lgÆmL)1) (Life Technol-ogies) Substrate-activator and reporter plasmids were transfected in a 1 : 2 ratio, respectively, into COS-7 or CHO cells using Lipofectamine 2000 reagent (Life Technol-ogies) according to the manufacturer’s protocol Stable transfected cell lines were obtained after selection with Hygromycin B (300 lgÆmL)1), Geneticin (500 lgÆmL)1) (Life Technologies) or Puromycin (2.5–12.5 lgÆmL)1) (Sigma)

Antibodies The mouse monoclonal antibodies 1E8 [28], WO2, G2-10 (specific for Ab40) and G2-11 (Ab42) [29] were used for immunoprecipitation and Western blotting of Ab The rabbit polyclonals anti-Gal4 DNA binding region (Upstate

Biotechnology) and the anti-PS1 98/1 [30], were used for

immunoprecipitation of lysates Sheep

anti-mouse–horse-radish peroxidase conjugate (Amersham) was used as

secondary antibody in the blotting procedure Rabbit

anti-mouse Igs (Dako, CA, USA) were used to link the 1E8

and G210 mAbs to the protein A Sepharose CL-4B

(Pharmacia)

Radiolabeling, immunoprecipitation, gel electrophoresis

and Western blotting

To analyze protein expression and processing, cells were

starved for 45 min in methionine- and cysteine-free medium

(ICN), pulsed for 30 min in medium containing

1 mCiÆmL)1 35S translabel mix (ICN) and chased for

60 min Cell lysis and immunoprecipitation were performed

as described [31], with the modification that the samples

were equalized for their radioactive incorporation and

pre-cleared twice with 100 lL formalin-fixed, heat-inactivated

Staphylococcus aureus Cowan strain bacteria (Staph A,

10% v/v) before immunoprecipitation to reduce the

back-ground signal Proteins were separated on 12% Tris-Tricine

gels and transferred to polyvinylidene fluoride membranes

(Millipore) The membrane was either exposed to a

phosphorimaging screen and analyzed with theMACBAS V

2.0 imaging software (Fuji) or exposed to BioMax MR-1

film (Kodak) and the density of the bands quantified using

theNIH-IMAGE1.60 software For Ab–Western blotting, the

cell culture medium (1 mL) was harvested from 10 cm

dishes seeded with similar number of cells (approximately

90% confluent) and Ab was immunoprecipitated with mAb

WO2, 1E8, G210 and G211 The immunoprecipitates were

resolved on 10–20% Tris-Tricine gels (Novex, Invitrogen)

and transferred to nitrocellulose The membranes were

boiled for 5 min, blocked with 0.5% (w/v) casein, incubated

with primary antibody WO2, and developed by

chemi-luminescence reaction (ECL, Amersham) Ab release from

cells treated with inhibitors was determined from

radio-labeled cells Cells were grown in 24 well plates and

preincubated with the c-secretase inhibitors in starvation

medium After one hour incubation the medium was

replaced with labeling medium containing the inhibitors as

described above and incubated for 17 h

FACS analysis and sorting

Cells were trypsinized and resuspended in NaCl/Pi

contain-ing 10 mM EDTA and 1–2% FBS Propidium iodide

(50 lgÆmL)1) was added to stain dead cells Cells were kept

on ice until analysis with FACScan, or sorting using MoFlo,

Facs Star or FACS-II (Becton Dickinson) Analysis was

performed using the computer software programWAESEL

1.2.1 (F Battye, Walter and Eliza Hall Institute, Parkville,

Australia)

Analysis of presenilin 1 transfections

Cells stable transfected with PS1 were plated in triplicate (in

12-well plates) After 24 h medium was immunoprecipitated

with WO2 antibody to analyze Ab-secretion by Western

blotting as described above Cells were washed and

prepared for FACS analysis as described above The density

of the Ab-bands was quantitated and Ab-secretion was calculated relative to the protein concentration of the lysates prepared from the cells in each well, as determined by BCA protein assay (Pierce)

Protease inhibitor treatment Cells were plated into 12- or 24-well plates and incubated for 72 h with various concentrations of inhibitor in a final dimethylsulfoxide concentration of 0.5% After 24 h the medium was replaced with fresh medium containing inhibitor Inhibition of c-secretase activity was determined

by FACS analysis of the cells, using fluorescence as an indication of GFP expression, and by immunoprecipita-tion of radiolabeled-Ab from the culture media (as described above) L-685,458 [32,33] and compound 2 were obtained from M Shearman (Merck Sharp and Dohme, Terlings Park, UK) and MW167 [34] was purchased from Calbiochem E-64d (2S,3S)-trans-epoxysuccinyl-L -leucyl-amido-3-methyl-butane ethyl ester and lactacystin were from Sigma Calpain inhibitor I, N-acetyl-Leu-Leu-nor-leucinal (ALLN) and the caspase inhibitors Boc-D-FMK and z-DEVD-FMK were purchased from Calbiochem The signal peptide peptidase inhibitor (Z-LL)2ketone [35], was kindly provided by M Bogyo (UCSF, San Francisco,

CA, USA)

Results

Assay design This novel c-secretase assay involves cotransfection of a substrate-activator and a reporter gene into mammalian cells, as outlined in Fig 1A The substrate-activator construct mimics c-secretase substrates, either based on the APP or on the Notch sequence that are fused to the transcription activator factor Gal4-VP16 It is expressed with a signal peptide to ensure correct insertion and orientation into the membrane Proteolytic processing of the substrate-activator protein by c-secretase releases the activator domain into the cytosol, allowing it to promote expression of the enhanced green fluorescent protein (EGFP) reporter gene (Fig 1A) Therefore, cells cotrans-fected with the substrate-activator and reporter constructs can be distinguished for their c-secretase activity by their fluorescent appearance

The APP substrate consists of the APP signal peptide followed by two extra amino acids, leucine and glutamate (LE), and the b-secretase C-terminal fragment of APP (A4CT) minus the cytoplasmic domain It has been shown that the SP-LE-A4CT construct is a suitable c-secretase substrate [36,37] Our rationale behind the design of the substrate-activator construct was to develop an assay to screen specifically for the modulators of c-secretase activity The cytoplasmic domain of APP was shown to be sensitive

to caspase cleavage [38–40], thus we deleted most of the cytoplasmic domain from the SP-LE-A4CT construct but

we retained the triple lysine, glutamine and tyrosine motif (KKKQY) [41] The Gal4-VP16 (GV) transcription factor [42] binds to the reporter construct that contains five Gal4 binding sites The herpes simplex virus protein VP16 promotes the expression of the EGFP reporter gene through

the E1b viral transcription initiation codon [27] This

APP-based substrate-activator construct was named

SP-A4DCT-GV A similar Notch-construct, termed SP-NOTL-GV, was

prepared that includes the APP signal peptide followed by

the human Notch 1 sequence (residues 1648–1927; including

the S1, S2 and S3 cleavage sites for furin, TACE and

c-secretase-like activity, respectively [21,23,43,44], fused to

the GV domain

Co-transfection of both substrate-activator

(SP-A4DCT-GV) and reporter (5Gal-E1b-EGFP) DNA constructs into

COS-7 and CHO cells resulted in the expression of GFP

positive cells Figure 1B shows phase (panels 1 and 3) and

fluorescence microscopy (panels 2 and 4) of COS-7 cells

transfected with both constructs (panels 3 and 4) or with an

empty pCDNA3.1+ plasmid plus the reporter (panels 1

and 2) Control cells that were mock-transfected with the

empty plasmid and the reporter construct did not express GFP (Fig 1B, panel 2) whereas cells expressing SP-A4DCT-GV plus the reporter expressed GFP and were fluorescent (Fig 1B, panel 4) This demonstrates that expression of GFP is totally dependent upon the release

of the activator domain (GV) from the SP-A4DCT-GV substrate Similarly, cells transfected with both

SP-NOTL-GV and the reporter displayed green fluorescence, indica-ting release of the activator domain into the cytosol (data not shown)

c-Secretase activity correlates with GFP expression Correct membrane orientation and signal peptide cleavage

of the SP-A4DCT-GV construct was confirmed using

in vitro transcription/translation of the DNA constructs according to Bunnell et al [45] (data not shown) To characterize the expression of the SP-A4DCT-GV substrate-activator construct, metabolically labeled cells stably trans-fected with the SP-A4DCT-GV plus reporter plasmids were lysed and the proteins were immunoprecipitated with WO2 (anti-Ab) and anti-Gal Igs Both antibodies showed reac-tivity for a protein migrating at approximately 36 kDa, a molecular mass consistent with that expected for expression

of the full length protein with glycosylation of the Gal4 binding domain (Fig 2A; lanes 2–4) Immunoprecipitation with WO2 depleted the anti-Gal4 reactive protein species from the lysate, as shown by a marked reduction of the signal in subsequent anti-Gal4 immunoprecipitation This result confirms that both antibodies target the same protein (compare Fig 2A; lanes 2–4) The 36 kDa protein was not immunoprecipitated from control cells that do not express the A4DCT-GV protein (Fig 2A; lanes 1, 5 and 6) Immunoprecipitation with anti-Gal yielded an additional band of 31 kDa that was not detected by WO2 and is thus N-terminally truncated From its electrophoretic mobility, this would correspond to the C-terminal fragment produced

by c- or e-secretase cleavage (Fig 2A; lanes 3 and 4) The

31 kDa band was subjected to automated Edman degra-dation Counting of the fractions revealed a radioactive signal in fractions 1 and 2, suggesting the presence of Met at cycle 1 or/and 2 (data not shown) This data is consistent with recent reports showing that eCTF starts with Val50 and is sensitive to amino peptidase degradation [7,8,46–48] Immunoprecipitation of conditioned media from these cells with WO2 and 1E8 mAbs showed Ab secretion Immuno-precipitation with Ab C-terminal specific antibodies dem-onstrated that the predominant species secreted was Ab40 (immunoreactive to G2-10) whereas Ab42(immunoreactive

to G2-11) was undetectable (Fig 2B) Therefore correct metabolism of the A4DCT-GV construct into Ab was occurring Anti-Gal immunoprecipitations of lysates of cells stably transfected with Notch substrate-activator and reporter constructs detected full-length NOTL-GV as a 64-kDa species and a cleavage product migrating at 49-kDa,

as expected for a site-3/c-secretase cleavage product (Fig 2A; lane 7)

The doubly transfected cells that expressed GFP showed heterogeneity in their fluorescence intensity Therefore preparative FACS was used to sort stable low and high fluorescent populations for use in further experiments (Fig 2C) At least three rounds of FACS sorting were

Fig 1 The fluorescent reporter c-secretase assay (A) Principle of the

assay Cells are cotransfected with a substrate-activator and a reporter

cDNA constructs The substrates are based on APP and Notch1

sequences genetically fused to Gal4-VP16 transcription activators.

c-Secretase cleavage releases the activator domain in the cytosol, which

can then traffic to the nucleus to initiate the transcription of the green

fluorescence protein (GFP) gene by binding to the 5Gal-E1b domain

of the GFP reporter construct (B) Transfection of COS-7 cells with

both substrate-activator and reporter genes yielded fluorescent cells.

Double-transfected COS-7 cells were fixed and observed by phase

(panels 1 and 3) and fluorescence microscopy (panels 2 and 4) Phase

microscopy showed similar images of cells transfected with mock (1) or

SP-A4DCT (3) activator constructs Fluorescence was observed in cells

transfected with the c-secretase substrate (4) but not in cells containing

the empty control plasmid (2).

performed to obtain cell populations with a GFP expression level that remained stable over time, as determined by measurement of fluorescence intensity To determine if the fluorescence intensity paralleled Ab secretion we immuno-precipitated Ab from the medium of metabolically labeled cells expressing low and high levels of GFP Figure 2C shows that cells with a high level of fluorescence (as determined by FACS) secrete more Ab than cells with low fluorescence This clearly shows that fluorescence is dependent on c-secretase cleavage and correlates with Ab production

c-Secretase inhibitors decrease cell fluorescence

To confirm further the specificity of our assay, we tested the effect of specific c-secretase inhibitors on the doubly transfected cells After incubation for 72 h in the presence

of inhibitors, the cells were analyzed by FACS for GFP expression Propidium iodide staining was used to gate for live cells In each sample we gated for the same number of live cells Quantitation was performed using the mean values

of the fluorescence intensity of the gated cells and converted

to percentages, to allow comparison between individual experiments Mock-transfected cells were considered as nonfluorescent (0%) and the test cells treated with 0.5% dimethylsulfoxide were regarded as 100% fluorescent (see Fig 3A)

A dose-dependent reduction of GFP expression was observed with L-685,458, a potent inhibitor of c-secretase activity [32,33] Treatment of COS-7 cells expressing the APP substrate with 1 lML-685,458 resulted in a nearly total loss of fluorescence (Figs 3B–D) A concentration of 5 lM

of inhibitor was required to achieve similar results in the same cell line expressing the Notch substrate (Fig 3D) The control inactive compound 2 had no effect A marked reduction of fluorescence was also observed when COS-7 cells expressing either substrate were incubated with the difluoroketone inhibitor MW167 [34], at 50 lM concentra-tion (Fig 3D)

To confirm that c-secretase inhibition paralleled the loss

of fluorescence, the cell media were analyzed for Ab production Each inhibition experiment was initiated with the same number of cells, but we observed that after 72 h incubation the cells treated with MW167 were less confluent than the control cells or those incubated with L-685,458 Microscopic examination at 66 h confirmed that MW167 (‡ 50 lM) had a growth inhibiting/toxic effect on the cells as judged by their morphology and density (data not shown) Therefore, the effects of the inhibitors on Ab secretion and substrate cleavage were determined by immunoprecipitation after 17 h incubation in the presence of 35S label Ab secretion was decreased in a dose-dependent manner upon treatment with L-685,458 and was almost totally abolished

at 1 lM concentration (Fig 3B) MW167 also had a pronounced effect on Ab secretion at 50 lM, the same concentration that dramatically reduced the cell fluores-cence As expected, an accumulation of APP substrate, which corresponds to bCTF, was observed upon inhibitor treatment (data not shown)

The effect of the c-secretase inhibitors was also studied in CHO cells to confirm our findings in a different cell line The inhibitors were 2–10 times less potent in CHO cells

Fig 2 Characterization of the assay (A) The substrate-activator

constructs are correctly expressed and proteolytically processed in

mammalian cells COS-7 cells stably transfected with empty vector

(–, lanes 1, 5 and 6), SP-A4DCT-GV (A, lanes 2–4) or SP-NOTL-GV

(N, lane 7) constructs were pulsed for 30 min and chased for 1 h Cell

lysates were analyzed by immunoprecipitation (IP) with anti-Gal

(lanes 1–3, and 5 and 7) or anti-Ab (WO2, lanes 4 and 6) Igs Both

antibodies recognized the expected 36-kDa glycosylated protein

A4DCT-GV (lanes 2–4), while anti-GAL4 also recognized the 31-kDa

cleaved C-terminal fragment of A4DCT-GV (CTFc-GV) (lanes 2 and

3) Lanes 3 and 5 correspond to cell lysates immunoprecipitated

sequentially (seq-IP) with anti-Gal Ig after WO2; lanes 4 and 6

cor-respond to the WO2 IPs As expected, N OTL-GV is expressed as a

64-kDa protein and the anti-GAL4 reactive fragment of 49-kDa has

the correct size to represent the N-terminally truncated NICD-GV

product from site-3 cleavage (lane 7) (B) IP of the conditioned medium

from cells transfected with SP-A4DCT-GV with anti-Ab Igs followed

by Western blotting with WO2 detected the presence of Ab peptide

(lanes 5–7) G2-10 (Ab 40 ; lane 7) but not G2-11 (Ab 42 ; lane 8)

immu-noprecipitated the 4-kDa species, indicating that the cells secreted

mostly Ab40 Ab was undetectable in medium from mock-transfected

cells (lanes 1–4) (C) Fluorescence intensity varied between individual

cells, reflecting different levels of GFP expression Low and high

GFP-expressing populations were isolated by preparative FACS A typical

histogram is shown The dotted line are mock-transfected cells, the

grey solid line represents the population sorted for low expression of

GFP, while the black solid line are the cells expressing high levels of

GFP Levels of secreted Ab were analyzed from these low and high

fluorescent populations, which were grown overnight in six-well plates

in medium containing 1 mCiÆmL)1 35S using WO2.

transfected with the APP substrate than in COS-7 cells.

5 lM L-685,458 reduced the fluorescence levels nearly to zero, independent of the eukaryotic expression plasmid used (Fig 3C) MW167 also reduced the GFP expression, but was 5 times less potent than L-685,458

Effect of c-secretase inhibitors on the fluorescence

of cells expressing the Notch substrate Although it has been shown that a similar cleavage releases the intracellular domains of APP and Notch, it remains unclear whether the same proteolytic activity is involved Therefore we compared the effect of c-secretase inhibitors

on both substrates The effect of the L-685,458 and MW167 was less pronounced on cells expressing the Notch substrate than on the cells expressing the APP substrate but the relative order of potency was conserved, i.e L-685,458 was fivefold more potent than MW167 (Fig 3D) A fivefold higher concentration of L-685,458 was required to reduce the level of fluorescence in the Notch-substrate transfected cells to the same level as in APP substrate transfected cells At all concentrations tested (0.1 lM, 0.5 lM and

1 lM), L-685,458 caused a significant decrease in relative fluorescence, whereas the control inactive compound 2, had

a very marginal effect (Fig 3D)

Effect of cysteine protease, proteasome, caspase, and signal peptide peptidase inhibitors on GFP-expression The level of GFP expression could not be reduced to zero even after a 72-h incubation with the potent c-secretase inhibitor, L-685,458 This may reflect the stability of the GFP protein, which has a 24-h half-life Alternatively the activator domain could be released by more than one

Fig 3 Specific c-secretase inhibitors abolish cell fluorescence Stable populations of green-fluorescent cells expressing substrate-activator and reporter genes were incubated for 72 h in the presence of 0.5% dimethylsulfoxide containing various concentrations of c-secretase inhibitors then analyzed by FACS (A) Relative fluorescence was calculated as the ratio of mean linear fluorescence (MLF) of inhibitor-treated cells (grey solid line) to the MLF of DMSO inhibitor-treated cells (black solid line) after subtraction of the MLF of the mock-transfected cells (dotted line) (B) Comparison of the effect of the c-secretase inhibitors

on Ab secretion (open bars) and cell fluorescence (grey bars) Cells were metabolically labeled for 17 h in the presence of inhibitors Ab was immunoprecipitated from the media with WO2, resolved on 10–20% Tris-Tricine gels and analyzed by phosphoimaging using

MACBAS 2.0 software Relative c-secretase activity (Ab-secretion) was calculated for each experiment towards the signal obtained for cells treated with 0.5% dimethylsulfoxide only, after subtraction of the Ab-signal obtained for mock-transfected cells Relative fluorescence (relative c-secretase activity) was calculated as described in panel A (C) Effect of c-secretase inhibitors on fluorescence of CHO cells stably transfected with SP-A4DCT-GV cloned in pcDNA3.1 (open bars) or

in pIRESpuro2 plasmids (grey bars) (D) Comparison of the effect of c-secretase inhibitors on APP c-secretase (AICD-release) (open bars) and Notch S3 cleavage (NICD-release) (grey bars) in the fluorescent reporter assays *P < 0.05, **P < 0.005, ***P < 0.001; n represents the number of individual experiments analyzed.

proteolytic activity [31] To test the latter hypothesis other

inhibitors were also applied in the COS-7 cell fluorescence

assay, including in particular proteasome and caspase

inhibitors The inhibitors lactacystin (0.5 lM and 1 lM),

ALLN(10 lM) and MG132 (10 lM and 100 lM) had a

toxic effect on the cells as determined by microscopic

examination The dead cells were stained with propidium

iodide and were excluded during FACS analysis Table 1

summarizes the results of the inhibitor treatments Among

the inhibitors tested only the caspase inhibitor

DEVD-FMK showed a significant, but very slight (7%) decrease in

fluorescence in cells expressing the APP substrate

Lacta-cystin and the other inhibitors did not decrease the

fluorescence in cells transfected with the Notch substrate,

confirming that the fluorescence observed was mostly due to

c-secretase cleavage The cysteine protease inhibitor E-64d

significantly increased the fluorescence, particularly in the

cells expressing the Notch substrate As c-secretase cleavage

resembles the cleavage of signal peptides by signal peptide

peptidases (SPP), an inhibitor of SPP was also tested in our

assay Both c-secretase and SPP cleave their substrates

within the middle of the transmembrane region It was

recently shown that a gene identified for its homology to

presenilin and termed presenilin homologue 3 (PSH3) [49]

corresponds to SPP Thus the inhibitor of signal peptide

peptidase-activity (Z-LL)2 ketone, was tested in our assay

[35,50] This inhibitor showed no significant effect on the cell

fluorescence at 1, 2 and 10 l on both APP and Notch

substrates (Table 1) and did not affect Ab-secretion (data not shown)

Effect of presenilin 1 expression on fluorescence and Ab-secretion

Presenilin is required for 40–42 cleavage, or c cleavage (for the release of Ab40)42), and 49 cleavage, or e cleavage, of APP The precise cleavage mechanism is unknown and data using PS1 dominant-negative aspartate mutations have been controversial We determined the effect of transfecting wild type (WT) PS1 and PS1 mutants (PS1 D257, PS1 D385A, PS1 D257/385) on the cell fluorescence in our APP-based assay We also tested the effect of the PS1 exon-9 deletion mutation (PS1 DE9) This mutation prevents PS1 endoproteolysis and causes an aggressive form of early onset AD with abundance of amyloid positive cotton-wool plaques [51,52] We compared the effect of these mutations

on cell fluorescence and Ab-secretion For each PS1 mutant several transfections were performed and stable cell lines were obtained The cell lines with a high level of expression

of PS1 were selected for the study (Fig 4A) Results of nine individual experiments show that transfection with PS1 WT resulted in a 2.2-fold increase in fluorescence as compared to mock (Fig 4B), which was similar to the effect seen on Ab-secretion (a 2.8-fold increase in Ab-secretion compared

to mock) (Fig 4D) Transfection of cells with PS1 bearing the single (PS1 D257A or PS1 D385A) or the double aspartate mutations (PS1 D257A/D385A) caused a decrease

in cell fluorescence as compared to cells transfected with PS1

WT (n¼ 7) (Fig 4C) The level of fluorescence was below the level of mock-transfected cells, indicating displacement

of endogenous PS1 The PS1 DE9 mutation caused an increase in fluorescence, but not to the same level as PS1 WT (n¼ 7) (Fig 4C) We observed a significant reduction in Ab-secretion from the cells transfected with PS1 D257A and PS1 D385A as compared to those transfected with PS1 WT (Fig 4E) Expression of PS1 D257A/D385A and PS1 DE9 mutations did not change Ab-secretion significantly as compared to expression of PS1 WT (results from 4 individual experiments) (Fig 4E)

Discussion

We have developed a novel GFP-based assay to character-ize c-secretase The advantage of this assay over existing systems is that it allows the isolation of cells with stable differences in c-secretase activity We established that the differences in fluorescence correlated with Ab production with the high fluorescent cells expressing more Ab than the low fluorescent cells Therefore the release of the GAL4-VP16 domain provides a direct measure of c-secretase activity This assay has a definite advantage over traditional Ab-antibody based assays, such as ELISA and immuno-precipitation, for measuring c-secretase activity by allowing

a direct measure of c-secretase activity The Ab-antibody assays would be affected by factors which affect Ab turnover and clearance An alternative assay has recently been described that uses luciferase as the reporter molecule [53,54] The luciferase-based assay has the advantage over the GFP assay of being more quantitative, but it cannot compensate for dead cells that would clearly affect the

Table 1 Comparative effect of various protease inhibitors on the

fluor-escence of cells transfected with SPA4DCT-GAL-VP or

SP-NOTL-GVP plus the GFP reporter The number of independent experiments is

indicated by n.

Inhibitor

Relative fluorescence (%)

E-64d

10 l M 123 ± 10 (n ¼ 4)* 150 ± 16 (n ¼ 3)*

5 l M 116 ± 19 (n ¼ 5) 137 ± 34 (n ¼ 3)

N-Acetyl-Leu-Leu-norleucinal

10 l M 113 ± 7 (n ¼ 4) 149 ± 27 (n ¼ 3)

5 l M 106 ± 10 (n ¼ 4) 114 ± 12 (n ¼ 4)

Lactacystin

1 l M 62 ± 20 (n ¼ 4) 110 ± 39 (n ¼ 4)

0.5 l M 76 ± 24 (n ¼ 4) 97 ± 34 (n ¼ 4)

0.1 l M 115 ± 23 (n ¼ 2) 95 ± 20 (n ¼ 3)

Boc-D-FMK

10 l M 96 ± 3 (n ¼ 2) 97 ± 1 (n ¼ 2)

5 l M 97 ± 3 (n ¼ 2) 97 ± 2 (n ¼ 2)

DEVD-FMK

10 l M 93 ± 1 (n ¼ 2)* 97 ± 4 (n ¼ 2)

5 l M 93 ± 0 (n ¼ 2)*** 96 ± 3 (n ¼ 2)

(Z-LL) 2

10 l M 132 ± 13 (n ¼ 2) 98 ± 7 (n ¼ 2)

2 l M 105 ± 0 (n ¼ 2)*** 109 ± 6 (n ¼ 3)

1 l M 155 ± 26 (n ¼ 2) 110 ± 6 (n ¼ 2)

*P < 0.05, ***P < 0.001.

readout Furthermore, proteasome inhibitors can interfere

directly with luciferase reporter enzymes [55]

The GFP-assay was adapted for studying c-secretase

cleavage of Notch Both cleavages of APP- and Notch-based

substrates were modulated by known c-secretase inhibitors

We found that inhibition of Ab secretion correlated with the observed decrease in fluorescence determined by FACS analysis, but they were never identical (see Fig 3B) This discrepancy may reflect the different experimental condi-tions used (17 h vs 72 h incubation with inhibitors) and/or different methods of measurement (fluorescence assay, which measures expression of the GFP, as compared to immunoprecipitation of Ab secreted in the culture media) Our results suggest that the proteolytic activity required for cleavage at position Leu49 is not identical to that cleaving at position Val40 The fluorescent assay measures the release of the cytoplasmic domain into the cytosol, which we have shown is cleaved at Leu49 whereas the Ab assay measures a species that is cleaved at position 40 However, both methods clearly measured a dose-dependent decrease in c-secretase cleavage with specific c-secretase inhibitors and the relative potency of L-685,458 and MW167 (over 50-fold) was the same for both assays This is consistent with previous reports [32–34] and suggests that the same proteo-lytic machinery produces Ab and the C-terminal cytosolic fragment

Differences in potency of the inhibitors in the alternative cell types might reflect differences in processing between cell lines as observed by other groups [32,56] Our results show

an approximately 50-fold difference in potency between L-685,458 and MW167 which is consistent with a previous report using a luciferase reporter assay [53] These effective concentrations of both inhibitors on c-secretase activity also correspond to their potency as determined from Ab secretion [32,56,57] We observed a difference in potency

of inhibition between Notch and APP substrates The effective concentrations from our data did not correspond

to those obtained by Taniguchi and coworkers in HEK293 cells using a luciferase-based assay [54] This might reflect differences in cell lines and substrates, because they observed different effects of the L-685,458 inhibitor depending on the Notch substrate used (greater inhibition with Notch 3 than Notch 1) Furthermore, they show that there is a Notch receptor cleavage that depends on, but is not directly executed by presenilins, and cannot be inhibited by

Fig 4 Effects of presenilin 1 mutations on fluorescence and Ab-secre-tion Green fluorescent cells (transfected with the APP-based substrate-activator and reporter) were transfected with PS1 WT or mutants (A) Stable cell populations were produced and cell-lysates were analyzed for PS1 expression with the 98/1, anti-PS1-NTF, Igs (B) Analysis by FACS shows that transfection with PS1 WT increases the fluorescence

by an average of 2.2-fold as compared to mock-transfection (n ¼ 9) (C) Results of seven individual experiments show that each PS1 mutation tested decreases the fluorescence significantly as compared to PS1 WT (right panel) (D) PS1 WT transfection increased the Ab secretion to 2.8-fold as compared to mock (average of five experi-ments) (E) PS1 D257A and PS1 D385A caused a significant reduction

in Ab secretion as compared to PS1 WT PS1 D257A/D385A and PS1 DE9 had no significant effect on Ab secretion Conditioned medium was removed from the cells prior to FACS analysis and immunopre-cipitated with WO2 Densitometry values were determined for each

Ab band and standardized for amount of protein in each well The results of four individual experiments are shown **P < 0.005,

***P < 0.001.

c-secretase inhibitors and by immunoprecipitation with

anti-PS Igs

The signal peptide peptidase inhibitor (Z-LL)2-ketone,

was unable to inhibit c-secretase activity when tested at 1, 2

and 10 lM concentrations This could reflect the opposite

membrane orientation of the active-site motifs YD and

LGLGD in the predicted transmembrane 6 and 7 of this

protein compared to these motifs in PS [50] The slight

inhibitory effect of the caspase inhibitor DEVD-FMK

shows that its contribution to the cell fluorescence is minor

This result could explain why we never observed total loss of

fluorescence of the cells even when Ab secretion was nil The

increase in fluorescence observed with the E64-d inhibitor

suggests that a cysteine protease degrades the Notch

substrate, making less protein available for c-secretase

cleavage The Notch substrate contains the entire cytosolic

domain whereas the APP does not, thus it is likely that the

protease inhibited by E-64d processes the Notch construct

within the cytosolic domain Furthermore cysteine protease

activity has been reported to remove PS1 fragments that are

not incorporated into the complex as well as the holoprotein

itself [58] and therefore inhibition of this activity could

increase the levels of PS1 and therefore increase c-secretase

activity

To test whether the assay could identify differences in

c-secretase activity due to changes in components of the

c-secretase complex, we overexpressed WT PS1 and some

PS1 mutations The original data by Wolfe and coworkers

[18] that the aspartate residues were critical for PS-mediated

cleavage of APP were reproduced in our GFP-based assay

We were able to decrease the fluorescence levels below those

of the mock-transfected cells, indicating some displacement

of the endogenous PS However, we were unable to achieve

complete inhibition of cell fluorescence by these mutants, as

we were unable to replace all the endogenous PS with the

exogenously expressed PS1 In PS1 D257A/D385A

trans-fected cells we observed a decrease in fluorescence and

therefore a decrease in AICD release, while the Ab levels

remain unchanged Kim and coworkers [59] observed a

similar decrease in intracellular domain release and no effect

on Ab-secretion when they expressed PS1 D257A/D385A in

N2A cells Together with Yu and coworkers [60,61] they

also showed that aspartate mutations alter APP-trafficking

Overexpression of PS1 DE9 resulted in only a 40% increase

of fluorescence as compared to PS1 WT, but did not

significantly alter Ab-levels Chen and coworkers recently

reported that expression of PS1 DE9 increased Ab42levels,

but inhibited cleavage at the e-site and the release of AICD

[62] Therefore our data provide further evidence that c- and

e-cleavage can be differentially affected by PS1 mutations

This strengthens the hypothesis that the c-secretase complex

could have multiple active sites, multiple conformations or

one active site and at least two different substrate binding

sites for c- and e-cleavage [63,64] The PS1 DE9 deletion

mutation, like the single aspartate mutations, affects the

maturation of the high molecular weight complex

compo-nents that constitute the c-secretase activity [60] These

mutations could thus affect the components present in the

complex PS1 DE9 overexpression results in normal Ab

secretion, but reduced fluorescence, indicating reduced CTF

release from the membrane even in the presence of Ab

production This suggests that c-cleaved CTF remains

anchored in the membrane and can therefore not activate the reporter gene transcription We are currently investi-gating the presence of membrane-anchored c-cleaved CTF

in brain cortex of PS1 DE9 carriers, PS1 DE9 lympho-cytes and other cell-models Alternatively the intracellular domain could be released in a different compartment, because of altered trafficking of the substrate caused by the PS1 mutation, and is either rapidly degraded or unable to reach the reporter gene

In conclusion, these results show that our GFP reporter assays based on APP or Notch c-secretase substrates can be used to specifically study modulation of c-secretase activity

in parallel in various cell types Results of the inhibitor study suggest possible differences in the proteolytic activities or pathways that cleave the APP and Notch constructs The use of these parallel assays could facilitate the search for compounds that target APP processing and have a lesser effect on Notch The variations we observed between cell types might reflect physiological differences in protein processing and should be taken into account during the development of therapeutics Our GFP assay allows for direct readout of c-secretase activity without the use of antibodies and could be further developed into high throughput screens These can also be applied to the study

of other membrane protein substrates with cleavages regulated by presenilins, such as Erb-B4 [65,66], E-cadherin [67], LRP-receptor [68] and CD44 [69] An advantage of the GFP-based assay is that it facilitates the selection of different cell populations by FACS that vary in their fluorescence intensity This can be used to screen cDNA libraries for genes that modulate c-secretase activity Complete identification and further characterization of this activity is required for a better understanding of the development of Alzheimer’s disease in early and late onset cases

Acknowledgments

This study was supported by grants from the National Health and Medical Research Council of Australia, the Clive and Vera Ramaciotti Foundation and Merck Sharp and Dohme We thank Drs

S Lichtenthaler, G Muscat, H Clarris, C Bergmann, T Hartmann,

F Reinhard and A Weidemann for providing DNA-constructs and advice, and Ms F Katsis for protein sequencing We thank Dr

M Shearman and Dr M Bogyo for providing inhibitors We thank Drs

S Mok, A Hill and N Williamson for helpful discussions.

References

1 Nunan, J & Small, D.H (2000) Regulation of APP cleavage by a-, b- and c-secretases FEBS Lett 483, 6–10.

2 Selkoe, D.J (2001) Alzheimer’s disease: genes, proteins, and therapy Physiol Rev 81, 741–766.

3 Evin, G & Weidemann, A (2002) Biogenesis and metabolism of Alzheimer’s disease Ab amyloid peptides Peptides 23, 1285–1297.

4 De Strooper, B & Konig, G (1999) Alzheimer’s disease A firm base for drug development Nature 402, 471–472.

5 McLendon, C., Xin, T., Ziani-Cherif, C., Murphy, M.P., Findlay, K.A., Lewis, P.A., Pinnix, I., Sambamurti, K., Wang, R., Fauq,

A & Golde, T.E (2000) Cell-free assays for c-secretase activity FASEB J 14, 2383–2386.

6 Pinnix, I., Musunuru, U., Tun, H., Sridharan, A., Golde, T., Eckman, C., Ziani-Cherif, C., Onstead, L & Sambamurti, K.

(2001) A novel c-secretase assay based on detection of the putative

C-terminal fragment-c of amyloid b protein precursor J Biol.

Chem 276, 481–487.

7 Yu, C., Kim, S.H., Ikeuchi, T., Xu, H., Gasparini, L., Wang, R &

Sisodia, S.S (2001) Characterization of a presenilin-mediated

amyloid precursor protein carboxyl-terminal fragment c Evidence

for distinct mechanisms involved in c-secretase processing of the

APP and Notch1 transmembrane domains J Biol Chem 276,

43756–43760.

8 Weidemann, A., Eggert, S., Reinhard, F.B., Vogel, M., Paliga, K.,

Baier, G., Masters, C.L., Beyreuther, K & Evin, G (2002) A novel

epsilon-cleavage within the transmembrane domain of the

Alz-heimer amyloid precursor protein demonstrates homology with

Notch processing Biochemistry 41, 2825–2835.

9 Lichtenthaler, S., Wang, R., Grimm, H., Uljon, S., Masters, C.L.

& Beyreuther, K (1999) Mechanism of the cleavage specificty of

alzheimer’s disease c-secretase identified by

phenylalanine-scan-ning mutagenesis of the transmembrane domain of the amyloid

precursor protein Proc Natl Acad Sci USA 96, 3053–3058.

10 Golde, T.E., Eckman, C.B & Younkin, S.G (2000) Biochemical

detection of Ab isoforms: implications for pathogenesis, diagnosis,

and treatment of Alzheimer’s disease Biochim Biophys Acta

1502, 172–187.

11 Murphy, M.P., Hickman, L.J., Eckman, C.B., Uljon, S.N., Wang,

R & Golde, T.E (1999) c-Secretase, evidence for multiple

pro-teolytic activities and influence of membrane positioning of

sub-strate on generation of amyloid b peptides of varying length.

J Biol Chem 274, 11914–11923.

12 Fortini, M.E (2001) Notch and presenilin: a proteolytic

mech-anism emerges Curr Opin Cell Biol 13, 627–634.

13 Esler, W.P., Kimberly, W.T., Ostaszewski, B.L., Ye, W., Diehl,

T.S., Selkoe, D.J & Wolfe, M.S (2002) Activity-dependent

iso-lation of the presenilin-c-secretase complex reveals nicastrin and a

c substrate Proc Natl Acad Sci USA 99, 2720–2725.

14 Steiner, H., Winkler, E., Edbauer, D., Prokop, S., Basset, G.,

Yamasaki, A., Kostka, M & Haass, C (2002) PEN-2 is an

integral component of the c-secretase complex required for

coordinated expression of presenilin and nicastrin J Biol Chem.

277, 39062–39065.

15 Goutte, C., Tsunozaki, M., Hale, V.A & Priess, J.R (2002)

APH-1 is a multipass membrane protein essential for the Notch

sig-naling pathway in Caenorhabditis elegans embryos Proc Natl

Acad Sci USA 99, 775–779.

16 Francis, R., McGrath, G., Zhang, J., Ruddy, D.A., Sym, M.,

Apfeld, J., Nicoll, M., Maxwell, M., Hai, B., Ellis, M.C., Parks,

A.L., Xu, W., Li, J., Gurney, M., Myers, R.L., Himes, C.S.,

Hie-bsch, R., Ruble, C., Nye, J.S & Curtis, D (2002) aph-1 and pen-2

are required for Notch pathway signaling, c-secretase cleavage of

bAPP, and presenilin protein accumulation Dev Cell 3, 85–97.

17 Lee, S.F., Shah, S., Li, H., Yu, C., Han, W.G & Yu, G (2002)

Mammalian APH-1 interacts with presenilin and nicastrin, and

is required for intramembrane proteolysis of APP and Notch.

J Biol Chem 277, 45013–45019.

18 Wolfe, M.S., Xia, W., Ostaszewski, B.L., Diehl, T.S., Kimberly,

W.T & Selkoe, D.J (1999) Two transmembrane aspartates in

presenilin-1 required for presenilin endoproteolysis and c-secretase

activity Nature 398, 513–517.

19 Steiner, H., Duff, K., Capell, A., Romig, H., Grim, M.G., Lincoln,

S & Hardy, J., Yu, X., Picciano, M., Fechteler, K., Citron, M.,

Kopan, R., Pesold, B., Keck, S., Baader, M., Tomita, T.,

Iwat-subo, T., Baumeister, R & Haass, C (1999) A loss of function

mutation of presenilin-2 interferes with amyloid b-peptide

pro-duction and notch signalling J Biol Chem 274, 28669–28673.

20 Levitan, D & Greenwald, I (1995) Facilitation of lin-12-mediated

signalling by sel-12, a Caenorhabditis elegans S182 Alzheimer’s

disease gene Nature 377, 351–354.

21 De Strooper, B., Annaert, W., Cupers, P., Saftig, P., Craessaerts, K., Mumm, J.S., Schroeter, E.H., Schrijvers, V., Wolfe, M.S., Ray, W.J., Goate, A & Kopan, R (1999) A presenilin-1-depen-dent gamma-secretase-like protease mediates release of Notch intracellular domain Nature 398, 518–522.

22 Chan, Y.M & Jan, Y.N (1999) Presenilins, processing of beta-amyloid precursor protein, and notch signaling Neuron 23, 201–204.

23 Kopan, R & Goate, A (2000) A common enzyme connects notch signaling and Alzheimer’s disease Genes Dev 14, 2799– 2806.

24 Okochi, M., Steiner, H., Fukumori, A., Tanii, H., Tomita, T., Tanaka, T., Iwatsubo, T., Kudo, T., Takeda, M & Haass, C (2002) Presenilins mediate a dual intramembranous gamma-secretase cleavage of Notch-1 EMBO J 21, 5408–5416.

25 Mumm, J.S & Kopan, R (2000) Notch signaling: from the out-side in Dev Biol 228, 151–165.

26 Muscat, G.E., Downes, M & Dowhan, D.H (1995) Regulation of vertebrate muscle differentiation by thyroid hormone: the role of the myoD gene family Bioessays 17, 211–218.

27 Lillie, J.W & Green, M.R (1989) Transcription activation by the adenovirus E1a protein Nature 338, 39–44.

28 Christie, G., Markwell, R.E., Gray, C.W., Smith, L., Godfrey, F., Mansfield, F., Wadsworth, H., King, R., McLaughlin, M., Coo-per, D.G., Ward, R.V., Howlett, D.R., Hartmann, T., Lichtent-haler, S.F., Beyreuther, K., Underwood, J., Gribble, S.K., Cappai, R., Masters, C.L., Tamaoka, A., Gardner, R.L., Rivett, A.J., Karran, E.H & Allsop, D (1999) Alzheimer’s disease: correlation

of the suppression of beta-amyloid peptide secretion from cultured cells with inhibition of the chymotrypsin-like activity of the pro-teasome J Neurochem 73, 195–204.

29 Ida, N., Hartmann, T., Pantel, J., Schroder, J., Zerfass, R., Forstl, H., Sandbrink, R., Masters, C.L & Beyreuther, K (1996) Ana-lysis of heterogeneous A4 peptides in human cerebrospinal fluid and blood by a newly developed sensitive Western blot assay.

J Biol Chem 271, 22908–22914.

30 Culvenor, J.G., Evin, G., Cooney, M.A., Wardan, H., Sharples, R.A., Maher, F., Reed, G., Diehlmann, A., Weidemann, A., Beyreuther, K & Masters, C.L (2000) Presenilin 2 expression in neuronal cells: induction during differentiation of embryonic car-cinoma cells Exp Cell Res 255, 192–206.

31 Nunan, J., Shearman, M.S., Checler, F., Cappai, R., Evin, G., Beyreuther, K., Masters, C.L & Small, D.H (2001) The C-terminal fragment of the Alzheimer’s disease amyloid protein precursor is degraded by a proteasome-dependent mechanism distinct from c-secretase Eur J Biochem 268, 5329–5336.

32 Shearman, M.S., Beher, D., Clarke, E.E., Lewis, H.D., Harrison, T., Hunt, P., Nadin, A., Smith, A.L., Stevenson, G & Castro, J.L (2000) L-685,458, an aspartyl protease transition state mimic, is a potent inhibitor of amyloid b-protein precursor c-secretase acti-vity Biochemistry 39, 8698–8704.

33 Li, Y.M., Xu, M., Lai, M.T., Huang, Q., Castro, J.L., DiMuzio-Mower, J., Harrison, T., Lellis, C., Nadin, A., Neduvelil, J.G., Register, R.B., Sardana, M.K., Shearman, M.S., Smith, A.L., Shi, X.P., Yin, K.C., Shafer, J.A & Gardell, S.J (2000) Photoactivated c-secretase inhibitors directed to the active site covalently label presenilin 1 Nature 405, 689–694.

34 Wolfe, M.S., Citron, M., Diehl, T.S., Xia, W., Donkor, I.O.

& Selkoe, D.J (1998) A substrate-based difluoro ketone selectively inhibits Alzheimer’s c-secretase activity J Med Chem.

41, 6–9.

35 Weihofen, A., Lemberg, M.K., Ploegh, H.L., Bogyo, M & Martoglio, B (2000) Release of signal peptide fragments into the cytosol requires cleavage in the transmembrane region by a protease activity that is specifically blocked by a novel cysteine protease inhibitor J Biol Chem 275, 30951–30956.