* Neuronal Cell Biology Laboratory, CME, KU Leuven and Flanders Interuniversitary Institute for Biotechnology (VIB4), Herestraat 49, 3000 Leuven, Belgium† Thromb–X, NV, Herestraat 49, 3000 Leuven, Belgium

‡ e–mail: Bart. Destrooper@ med. kuleuven. ac. be he question of how proteins such as the β-amyloid precursor protein (β-APP), Notch-1 and others can be cleaved in the plane of the membrane has challenged cell biologists and researchers into Alzheimer’s disease for some time. Processing of β-APP by γ-secretase is considered an important therapeutic target, as it constitutes the final step in the release of the amyloid βpeptide (Aβ), the principal constituent of the amyloid plaques in the brains of Alhzeimer’s patients. Recently, γ-secretase activity was linked to two proteins, presenilin-1 (PS1) and presenilin-2 (PS2; refs 1, 2). Presenilins are strongly hydrophobic proteins embedded in intracellular membranes, and missense mutations in their genes cause a rare, aggressive form of Alzheimer’s disease3. It has been proposed that two aspartate residues in the transmembrane domains of presenilins constitute the catalytic site of γsecretase2. One argument against this hypothesis is the important residual γ-secretase activity observed in cells derived from PS1-or PS2-deficient mice1, 4. The question is therefore whether cells that are completely devoid of PS1 and PS2 maintain this γ-secretase activity or not.

Unfortunately, PS-null mice die early in embryogenesis (after roughly 9.5 days of development; E9. 5), making it impossible to generate sufficient cells to carry out the required biochemical experiments4. We therefore generated pluripotent lines of embryonic stem cells from PS-null blastocysts obtained by mating of PS1+/–PS2–/–mice4. We genotyped the cell lines using the polymerase chain reaction (PCR) 4 and Southern blotting (Fig. 1a). The 32% yield of double-deficient (PS-null) cell lines was slightly higher than the expected 25%; we confirmed the absence of PS1 (Fig. 1b) and PS2 (data not shown) by western blotting. We observed the accumulation of endogenous β-and α-secretase-cleaved β-APP carboxy-terminal stubs, the direct substrates for γ-secretase (Fig. 1b and ref. 1). We then expressed β-APP695 harbouring the K595N/M596L (Swedish) mutation in the embryonic stem cells, using recombinant Semliki Forest virus (SFV). This mutation markedly increases production of Aβ5, 6, facilitating its detection in our assays. In contrast to results obtained with single-knockout cells, no Aβ was detected in the conditioned media of PS1–/–PS2–/–cells (Fig. 1c). Using serial dilutions, we established that we would have detected Aβ even if its production was< 0.5% of that in wildtype cells (Fig. 1c). We further confirmed the absence of γ-secretase activity using an enzyme-linked immunosorbent assay (ELISA) specific for the 42-amino-acid form of Aβ1. mNotch∆ E is another substrate for PS1-dependent γsecretase7, 8. Processing of this construct mimics that of full-length Notch-1, and results in release of the Notch intracellular domain (NICD). NICD translocates to the nucleus and activates, together with transcription factors of the CBF-1 family, several genes, including HES-1 (ref. 9). We therefore co-transfected the embryonic stem cells with plasmids encoding mNotch∆ E and luciferase under the control of the HES-1 promoter (kindly provided by R. Kopan and A. Israel), and measured induction of luciferase activity10. As a control, we also transfected embryonic stem cells with plasmids encoding the NICD. Induction of HES-1 activity with mNotch∆ E was observed in PS1+/+ PS2–/–cells, but not in PS1–/–