The transmembrane domain of the amyloid precursor protein is required for antiamyloidogenic processing by α-secretase ADAM10

Abstract

Neurotoxic amyloid β-peptides are thought to be a causative agent of Alzheimer's disease in humans. The production of amyloid β-peptides from amyloid precursor protein (APP) could be diminished by enhancing α-processing; however, the physical interactions between APP and α-secretases are not well understood. In this study, we employed super-resolution light microscopy to examine in cell-free plasma membranes the abundance and association of APP and α-secretases ADAM10 (a disintegrin and metalloproteinase) and ADAM17. We found that both secretase molecules localize similarly closely to APP (within ≤50 nm). However, when cross-linking APP with antibodies directed against the GFP tag of APP, in confocal microscopy, we observed that only ADAM10 coaggregated with APP. Furthermore, we mapped the involved protein domain by using APP variants with an exchanged transmembrane segment or lacking cytoplasmic/extracellular domains. We identified that the transmembrane domain of APP is required for association with α-secretases and, as analyzed by Western blot, for α-processing. We propose that the transmembrane domain of APP interacts either directly or indirectly with ADAM10, but not with ADAM17, explaining the dominant role of ADAM10 in α-processing of APP. Further understanding of this interaction may facilitate the development of a therapeutic strategy based on promoting APP cleavage by α-secretases.

Keywords: ADAM17; Alzheimer's disease; amyloid beta; secretases; super-resolution microscopy.

Figure 1

Illustration of cell-freeplasmamembrane sheets (30).A, membrane sheets are generated by a brief 100 ms ultrasound pulse in ice-cold solution. Mechanical shearing forces applied to glass-adhered cells (left) remove the apical membrane and cytosolic structures; only the basal plasma membrane remains (right). B, about 1 day after transfection, HepG2 cells expressing APP-GFP are either directly fixed (cell; left) or membrane sheets are generated (right), followed by recording of APP-GFP fluorescence by confocal microscopy. APP, amyloid precursor protein.

Figure 2

APP processing in native plasma membrane sheets. Epifluorescence micrographs show images from membrane sheets generated from HepG2 cells that express mCherry-APP-GFP (see illustration to the right; arrows point towards the α- and β-cleavage site; yellow, Aβ region). Membrane sheets are either directly fixed (left images) or after incubation for 10 min at 37 °C in medium containing 10 μM DAPT (right images). Images from the same channels are shown at the same contrast adjustment (green, GFP channel; magenta, mCherry channel). Bar chart, quantification of GFP and mCherry on directly fixed (set to 100%) or incubated membranes. Values are given as means ± SD (n = 9 experiments, 10–20 membrane sheets per experiment and condition). Unpaired Student’s t tests compare incubation to direct fixation (ns, p >0.05; ∗∗∗∗p <0.0001). Aβ, amyloid β-peptide; APP, amyloid precursor protein; DAPT, N-[N-(3,5-difluorophenacetyl)-l-alanyl]-S-phenylglycine t-butyl ester; ns, not significant.

Figure 3

ADAM10 is the mainly involved α-secretase in APP processing.A, STED micrographs of membrane sheets from HepG2 (left) and SH-SY5Y cells (right) expressing APP-GFP, incubated without inhibitor (control; top), the broad inhibitor Batimastat (BATI; middle), or the ADAM10-specific inhibitor GI254023X (GI; bottom). B and C, Atto647 nanobody intensity quantified on (B) HepG2 and (C) SH-SY5Y cell membrane sheets. Values are given as means ± SD (n = 3 experiments per cell line, 13–40 membrane sheets per experiment and condition). Unpaired Student’s t tests compare the control to BATI or GI (∗∗∗∗p <0.0001; ∗∗∗p <0.001; ∗∗p <0.01; and ∗p <0.05). ADAM, a disintegrin and metalloproteinase; APP, amyloid precursor protein; STED, stimulated emission depletion.

Figure 4

Lateral organization of APP and secretases. STED micrographs of membrane sheets from (A) HepG2 cells expressing APP-GFP and (E) nonoverexpressing SH-SY5Y cells. Membrane sheets are double stained for secretases (magenta; left ADAM10; right ADAM17) and (A) overexpressed or (E) endogenous APP (green). Shown are channel overlays of overviews (top panels) and magnified views from the boxed regions (low panels). B and F, APP and secretase maxima density. C, D, G, and H, frequency distribution histograms of shortest intermaxima distances between (C and G) ADAM10 and APP or (D and H) ADAM17 and APP maxima in HepG2 (C and D) and SH-SY5Y (G and H) cells. For clarity, only distances ≤300 nm (C and D) or ≤550 nm (G and H) are included. Black and grey bars show frequencies of original and flipped images (the images of one channel were flipped vertically and horizontally as reference for randomized distribution). B and F, values are given as means ± SD ([B] n = 3 experiments; [F] n = 4 experiments, 10–40 membrane sheets per experiment and condition). C, D, G, and H, histograms include data from 43 to 61 membrane sheets collected from three to four experiments. ADAM, a disintegrin and metalloproteinase; APP, amyloid precursor protein; STED, stimulated emission depletion.

Figure 5

APP coaggregation with ADAM10 and α-cleavage depend on the transmembrane segment of APP.A, confocal micrographs of membrane sheets from HepG2 cells expressing APP-GFP. Membrane sheets were directly fixed (left), incubated without (control; middle) or with cross-linking antibodies (CoP; right), followed by immunostaining for ADAM10 (upper panels) or ADAM17 (lower panels). APP-GFP (green) and ADAM10/ADAM17 (magenta). B, overlap through cross-linking is quantified by the Pearson correlation coefficient (PCC) between ADAM10 or ADAM17 and the APP constructs as indicated (APP, APP-ΔN, APP-TMS, and APP-ΔC; for membrane sheets from cells expressing the variants, see Fig. S7A). C, analysis as in (B) of membrane sheets from APP-GFP–expressing HepG2 cells grown in the absence (−PMA) or the presence of 1 μM PMA (+PMA) (for images of the membrane sheets, see Fig. S4B). B and C, values are expressed as percentage of the condition 'fixed' (100% reference line). Values are given as the means ± SD ([B] n = 3 to 10 experiments for ADAM10 and APP/APP variants; n = 4 experiments for ADAM17 and APP; [C] n = 3 experiments; 10–20 membrane sheets per experiment and condition). D, Western blot quantification of sAPPα in lysate (Ly) and supernatant (Sn) of HepG2 cells grown in the presence of 10 μM DAPT, expressing either GFP-labeled APP or APP-TMS. The sum of sAPPα band intensities (Ly + Sn) is related to the sum of the band intensities of immature (APPi) and mature APP (APPm); APP-TMS is related to APP (set to 100%). Value is given as the mean ± SD (n = 3 experiments). Unpaired Student’s t tests compare (B and C) control and CoP to fixed or (D) APP-TMS to APP (∗∗∗∗p < 0.0001; ∗∗∗p < 0.001; ∗∗p < 0.01; and ∗p < 0.05). ADAM, a disintegrin and metalloproteinase; APP, amyloid precursor protein; DAPT, N-[N-(3,5-difluorophenacetyl)-l-alanyl]-S-phenylglycine t-butyl ester; PMA, phorbol 12-myristate 13-acetate; sAPPα, soluble APPα; TMS, transmembrane segment.