Dimerization of the transmembrane domain of amyloid precursor protein is determined by residues around the γ-secretase cleavage sites

One of the hallmarks of Alzheimer's disease is the formation of extracellular amyloid plaques that consist mainly of abnormally aggregated forms of amyloid β (Aβ) peptides. These peptides are generated by γ-secretase-catalyzed cleavage of a dimeric membrane-bound C-terminal fragment (C99) of the amyloid precursor protein. Although C99 homodimerization has been linked to Aβ production and changes in the aggregation-determining Aβ42/Aβ40 ratio, the motif through which C99 dimerizes has remained controversial. Here, we have used two independent assays to gain insight into C99 homodimerization in the context of the membrane of live cells: bioluminescence resonance energy transfer and Tango membrane protein-protein interaction assays, which were further confirmed by traditional pull-down assays. Our results indicate a four-amino acid region within the C99 transmembrane helix (⁴³TVIV⁴⁶) as well as its local secondary structure as critical determinants for homodimerization. These four amino acids are also a hot spot of familial Alzheimer's disease-linked mutations that both decrease C99 homodimerization and γ-secretase cleavage and alter the initial cleavage site to increase the Aβ42/40 ratio.

Alzheimer's disease-associated mutations increase amyloid precursor protein resistance to γ-secretase cleavage and the Aβ42/Aβ40 ratio

Mutations in the amyloid precursor protein (APP) gene and the aberrant cleavage of APP by γ-secretase are associated with Alzheimer's disease (AD). Here we have developed a simple and sensitive cell-based assay to detect APP cleavage by γ-secretase. Unexpectedly, most familial AD (FAD)-linked APP mutations make APP partially resistant to γ-secretase. Mutations that alter residues N terminal to the γ-secretase cleavage site Aβ42 have subtle effects on cleavage efficiency and cleavage-site selectivity. In contrast, mutations that alter residues C terminal to the Aβ42 site reduce cleavage efficiency and dramatically shift cleavage-site specificity toward the aggregation-prone Aβ42. Moreover, mutations that remove positive charge at residue 53 greatly reduce the APP cleavage by γ-secretase. These results suggest a model of γ-secretase substrate recognition, in which the APP region C terminal to the Aβ42 site and the positively charged residue at position 53 are the primary determinants for substrate binding and cleavage-site selectivity. We further demonstrate that this model can be extended to γ-secretase processing of notch receptors, a family of highly conserved cell-surface signaling proteins.