The C99 domain of the amyloid precursor protein is a disordered membrane phase-preferring protein

Processing of the amyloid precursor protein (APP) via the amyloidogenic pathway is associated with the etiology of Alzheimer’s disease. The cleavage of APP by β-secretase to generate the transmembrane 99-residue C-terminal fragment (C99) and subsequent processing of C99 by γ-secretase to yield amyloid-β (Aβ) peptides are essential steps in this pathway. Biochemical evidence suggests amyloidogenic processing of C99 occurs in cholesterol- and sphingolipid-enriched liquid ordered phase membrane raft domains. However, direct evidence that C99 preferentially associates with rafts has remained elusive. Here, we test this idea by quantifying the affinity of C99-GFP for raft domains in cell-derived giant plasma membrane vesicles. We find that C99 is essentially excluded from ordered domains in HeLa cells, SH-SY5Y cells and neurons, instead exhibiting a strong (roughly 90%) affinity for disordered domains. The strong association of C99 with disordered domains occurs independently of its cholesterol binding activity, homodimerization, or the familial Alzheimer disease Arctic mutation. Finally, we confirm previous studies suggesting that C99 is processed in the plasma membrane by α-secretase, in addition to the well-known γ-secretase. These findings suggest that C99 itself lacks an intrinsic affinity for raft domains, implying either that amyloidogenic processing of the protein occurs in disordered regions of the membrane, that processing involves a marginal sub-population of C99 found in rafts, or that as-yet-unidentified protein-protein interactions involving C99 in living cells drive it into rafts to promote its cleavage therein.