Spatial segregation of gamma-secretase and substrates in distinct membrane domains

Abstract

Gamma-secretase facilitates the regulated intramembrane proteolysis of select type I membrane proteins that play diverse physiological roles in multiple cell types and tissue. In this study, we used biochemical approaches to examine the distribution of amyloid precursor protein (APP) and several additional gamma-secretase substrates in membrane microdomains. We report that APP C-terminal fragments (CTFs) and gamma-secretase reside in Lubrol WX detergent-insoluble membranes (DIM) of cultured cells and adult mouse brain. APP CTFs that accumulate in cells lacking gamma-secretase activity preferentially associate with DIM. Cholesterol depletion and magnetic immunoisolation studies indicate recruitment of APP CTFs into cholesterol- and sphingolipid-rich lipid rafts, and co-residence of APP CTFs, PS1, and syntaxin 6 in DIM patches derived from the trans-Golgi network. Photoaffinity cross-linking studies provided evidence for the preponderance of active gamma-secretase in lipid rafts of cultured cells and adult brain. Remarkably, unlike the case of APP, CTFs derived from Notch1, Jagged2, deleted in colorectal cancer (DCC), and N-cadherin remain largely detergent-soluble, indicative of their spatial segregation in non-raft domains. In embryonic brain, the majority of PS1 and nicastrin is present in Lubrol WX-soluble membranes, wherein the CTFs derived from APP, Notch1, DCC, and N-cadherin also reside. We suggest that gamma-secretase residence in non-raft membranes facilitates proteolysis of diverse substrates during embryonic development but that the translocation of gamma-secretase to lipid rafts in adults ensures processing of certain substrates, including APP CTFs, while limiting processing of other potential substrates.

Fig. 1. DIM association of BACE-1, γ-secretase, and APP CTFs in mouse brain

12-Month-old adult mouse brain was homogenized and solubilized in a buffer containing 0.5% Lubrol WX at 4 °C for 30 min. The lysates were then subject to flotation sucrose density gradient centrifugation as described under “Experimental Procedures.” The gradients were harvested from the top, and equal volume of each fraction was analyzed by Western blotting with antibodies against APP C terminus, BACE-1, PS1, nicastrin, and PEN-2. Fractions 4 and 5 represent the interface between 5 and 35% sucrose in the gradient and are enriched in lipid raft marker flotillin-2. Fractions 8–12 contain detergent-soluble proteins, marked by the presence of non-raft proteins γ-adaptin and calnexin.

Fig. 2. DIM accumulation of APP CTFs in cells lacking γ-secretase activity

wt and PS1^−/−/PS2^−/− fibroblasts were solubilized in 0.5% Lubrol WX and analyzed by sucrose gradient fractionation and Western blotting with antibodies against APP, flotillin-2, and cavelolin-1. In the absence of PS1 and PS2 expression, endogenous wt APP CTFs become Lubrol WX-resistant and were recovered mainly in raft fractions 4 and 5, identified by the presence of raft markers.

Fig. 3. DIM localization of γ-secretase components and APP CTFs in stable PS1 D385A cells

Stably transfected N2a APPSwe cells co-expressing wt PS1 or D385A PS1 were subject to flotation sucrose density gradient centrifugation and analyzed by Western blotting with antibodies against PS1, nicastrin, APH-1, PEN-2, and APP. Note that APP CTF in PS1 wt cells is detectable in non-raft fractions upon longer exposure of the immunoblots to the film. mat, mature; imm, immature.

Fig. 4. APP CTFs associate with DIM in cholesterol-sensitive manner and co-reside with syntaxin 6 and PS1

A, N2a Swe.10 cells were treated with Me₂SO or 1 μM L685,458 for 16 h. Lipid raft association of full-length APP and APP CTFs was assessed by flotation sucrose density gradient centrifugation. Note the marked accumulation of APP CTFs in raft fractions in cells treated with the inhibitor. B, Swe.10 cells were exposed to 10 nM CompE for 16 h and then treated with Me₂SO (vehicle control) or 5 mM MβCD for 2 h at 37 °C prior to fractionation. Note that DIM accumulation of APP CTFs is sensitive to cholesterol depletion. C, pooled fractions 4 and 5 from the indicated cell lines were fractionated on 16.5% Tris-Tricine gels and sequentially probed with APP C-terminal polyclonal antibody 369 and mAb 26D6 (raised against epitopes 1–12). Antibody 369 reacts with β-, β′ (+11)-, and α-CTFs, whereas mAb 26D6 only reacts with 3-CTF. D, DIM fractions from Swe.10 cells treated with CompE were incubated with magnetic beads coated with syntaxin 6 or OKT8 antibody. Bound DIMs were analyzed by Western blotting using 369, mAb 26D6, and PS1_NT. An aliquot of the input (1/30th volume) was also fractionated in the same gel for comparison. Note that APP CTFs, but not full-length APP, co-reside with syntaxin 6 in DIMs.

Fig. 5. Covalent labeling of PS1 by active site directed 3-secretase inhibitor L-852,505

Pooled raft and non-raft fractions from wt mouse embryonic fibroblasts (A), N2a wt.11 cells (B), or adult mouse brain (C) were incubated with Me₂SO (− inhibitor) or L-852,505 (+ inhibitor), followed by irradiation with 365 nm UV light on ice for 90 min to allow photoaffinity cross-linking. Biotinylated proteins were captured with streptavidin-agarose beads and eluted by incubation at 95 °C for 5 min in Laemmli/SDS sample buffer. The eluates (lanes 5–9) as well as 1/40th of the solubilized membranes used for the streptavidin (Strep.) precipitation (input) were analyzed by immunoblotting using αPS1Loop antibody. The left panels show the distribution of PS1 and flotillin-2 in the flotation density gradient fractions used for the in vitro cross-linking experiment. IP, immunoprecipitation.

Fig. 6. Flotation density gradient analysis of Notch1 and Jagged2 derivatives

A, mouse embryonic fibroblasts stably expressing Notch-6mycGFP were treated or not with 10 nM CompE for 18 h and incubated with 10 mM EDTA for 30 min to induce ligand-independent activation of Notch processing (51) prior to lysis in 0.5% Lubrol WX and fractionation. Full-length Notch- and C-terminal derivatives were visualized by immunoblotting with mAb 9E10. Note that the Notch cleavage products S1/TMIC (indicated by filled arrowhead) and S2/NEXT (open arrowhead) migrate as a close doublet, and S3/NICD (open circle) is absent in cells treated with CompE. Endogenous APP CTF was analyzed by combined immunoprecipitation and immunoblotting with APP antibody 369. Lower panels represent immunoblots probed with PS1 and nicastrin antibodies. mat, mature; imm, immature. B, 3T3 fibro-blasts stably transfected with Myc-tagged Jagged2 were incubated with CompE, and distribution of Jagged2 and derivatives was analyzed by fractionation and immunoblotting with mAb 9E10. CompE-treated cells lack JICD but have increased levels of Jagged2 CTF in non-raft fractions.

Fig. 7. Non-raft association of Notch1, DCC, and N-cadherin CTFs

Brain tissue from 12-month-old adult mouse or E15.5 embryos were homogenized in 0.5% Lubrol WX lysis buffer and subject to flotation sucrose density gradient centrifugation. Aliquots of the gradients were analyzed by Western blotting with antibodies against C terminus of Notch1, DCC, N-cadherin, and syntaxin 6. Notch cleavage products S1/TMIC (indicated by filled arrowhead), S2/NEXT (open arrowhead), and S3/NICD (open circle) are indicated. Note that in PS1^−/− embryos S3/NICD is absent, and the levels of S2/NEXT as well as DCC and N-cadherin CTFs are higher relative to wt embryonic brain.

Fig. 8. Non-raft association of PS1 and APP CTFs in embryonic mouse brain

Raft and non-raft distribution of PS1, nicastrin, APP, flotillin-2, calnexin, and γ-adaptin in flotation density gradient fractions from E15.5 embryonic brains were analyzed by Western blotting. Note the predominant non-raft localization of PS1, nicastrin, and APP CTFs. mat, mature; imm, immature.