Figure 1.
Sortilin and APP are allied in vitro and in vivo.
(A) Colocalization of sortilin and APP. Mouse cortical neuron (upper) and brain cortex (lower) were immunostained for APP (red) and sortilin (green). Colocalization of sortilin and APP was indicated in merged panels (yellow). DAPI stained cell nuclei (blue). Plotted colocalization: 92% mean± SEM, n = 20, in cortical neurons and 95% mean± SEM, n = 3, in brain cortexes. Scale bar 25 µm for cortical neurons, 5 µm for enlarged image and 100 µm for brain cortex. (B) Co-IP of sortilin with APP in co-transfected HEK293 cells. HEK293 cells growing in 10 cm culture dishes were co-transfected with APP770-YFP/Sort-FL-myc/His (lane 1, 4) or Sort-T-myc/His (lane 2, 5). Cell lysates were immunoprecipitated with rabbit anti-GFP (α-GFP) for APP and blotted with mouse anti-Myc (α-Myc) for sortilin. Mixed lysates were used for IgG (lane 3). Sort-FL-myc.His (Sort-myc), Sort-T-myc.His (sort-T-myc) and APP770-YFP (APP-YFP) are indicated by arrows. (C) Co-IP of sortilin with APP in APPSwe/PS1dE9 transgenic mouse brain lysate. Mouse brain lysates were subjected to immunoprecipitation with rabbit anti-APP C’ (α-APP C’) and blotted with rabbit anti-sortilin (α-Sort) (left panel) or immunoprecipitation with α-Sort and blotted with α-APP C’ (right panel). Sortilin and APP are indicated by arrows. (D) Control for Co-IP using pEYFP. HEK293 cells were co-transfected with pEYFP/Sort-myc. Co-IP was performed using α-GFP and blotted with α-sort and α-GFP. Rabbit IgG (IgG) was used as a control for non-specific binding.
Figure 2.
A schematic diagram of APP-YFP and sortilin-CFP constructs.
The region of 78–831 aa is chosen for cloning based on that mature sortilin receptor, a bioactive form is released by furin cleavage from its proform at amino acid (aa) residue 77 in neurons (see results). The region cloned into pEYFP-N1 or pECFP-N1 is indicated by open and solid boxes. The line indicates the deleted region. Abbreviation: signal peptides: SP; extracellular domain: ECD; transmembrane domain: TMD; intracellular domain: ICD; amino acid: aa; gamma secretase: γ-Sec; furin cleavage site: FS; Motif 1: MS1 (787FLVHRY792); Motif 2: MS2 (823HDDSDEDLL831); deletion: del.; Arrowhead: α or β secretase cleavage site.
Figure 3.
FRET analysis of the interaction between sortilin and APP.
HEK293 cells co-transfected with APP770-YFP (APP-YFP) and different sortilin-CFP constructs were used for FRET. (A) FRET efficiency representing the protein-protein interaction was determined from a photobleached region of interest (ROI). (B) Expression levels for constructs used in FRET were determined by Western blot with gt-α-GFP. β-Actin was blotted with mouse-α-actin for sample equal loading. NC: negative control. PC: positive control. Three independent experiments were performed. Bars represent mean±SEM (n = 6 ROI×3). The star (*) indicates p<0.01.
Figure 4.
Mapping sortilin binding to APP.
(A) Identifying sortilin binding sites to APP. HEK293 cells were co-transfected with APP695-pcDNA3.1 (APP695) and Sort 78–385-CFP (lane 1) or Sort 385–786-CFP (lane 2), or Sort 756–831 (lane 3), or Sort 779–831-CFP (lane 4). Cell lysates were immunoprecipitated with mouse anti-APP, 22c11, (α-APP, lanes 1–4) and mouse IgG (mixed lysates, lane 5), and blotted with goat-anti-GFP (gt-α-GFP). Lysates were used as input (lanes 6–9). Sort 78–385CFP (61 kDa), Sort 385–786CFP (71 kDa), Sort 756–831CFP (35 kDa) Sort 779–831-CFP (32 kDa), APP695 (100 kDa) and IgG heavy chain (HC, 50 kDa) are indicated by arrows. (B) Determining sortilin C-terminal FLVHR motif (MS1) binding to APP. APP695 was used with Sort del.MS2-CFP (lanes 1, 6) or Sort del.MS1-CFP (lanes 2, 7) or Sort del.MS1/MS2-CFP (lanes 3, 8) or Sort MS1/MS2 (lanes 4, 9) for co-transfection. Co-IP was performed as (A). Sort del.MS2-CFP (34 kDa), Sort del.MS1-CFP (34 kDa), Sort-del.MS1/MS2-CFP (32 kDa), Sort MS1/MS2-CFP (33 kDa) and APP695 in precipitated samples and lysate are indicated by arrows. (C) Control for Co-IP using pECFP. HEK293 cells were co-transfected with pECFP/APP695. Co-IP was performed using α-APP and mouse IgG (IgG) as indicated. The blot was probed with α-GFP (upper) and re-probed with α-APP (lower). Lysate was used as input. APP695 and CFP (27 kDa) are indicated by arrows. Mouse IgG was used as a control for non-specific binding in all co-IP. (D and E) Sort-T-myc/His and MS1 peptide (CGGRFLVHRYSVLQQ; Peptide-2.0, Chantilly, VA) corresponding to the amino acids 783–797 of sortilin (GenBank Accession No. NM_002959) prevent sortilin N or C terminal constructs binding to APP in the competition binding assay, respectively. Sort-78–385CFP and Sort-del.MS2CFP (D) and two new sortilin constructs (termed Sort-sp-78–385-CFP and Sort-sp-del.MS2CFP) created by overlapping PCR and cloned in pECFP (E) were used in the competition binding assays. The integrity of the constructs was confirmed by DNA sequencing. To purify Sort-T-myc/His, HEK293 cells growing in 3×10 cm culture dishes were transfected with Sort-T-myc/His for 48 h. Cells were harvested in cold PBS containing 1 mM PMSF and protease inhibitor (Roche), sonicated on ice and centrifuged at 15000 rpm at 4°C for 10 minutes. The supernatant was loaded into Ni-NTA column (Qiagen) and eluted in elution buffer after washing. The elute protein was concentrated by Amicon Ultra- Centrifugal Filters (Millipore) and the protein was recovered in 50 µl of PBS pH 8, containing 1 mM PMSF and protease inhibitor. For competition binding assays, HEK293 cells (1×10 cm dish) were transfected with APP695 or each sortilin construct for 48 hours and harvested for preparing cell lysates. The lysate containing APP695 was incubated with 50 µl (+) of the purified Sort-T-myc/His protein or with 10 µM (+) of MS1 peptide at 4°C for 1 hour and then combined with the lysate containing Sort-78-385CFP or Sort-del.MS2CFP or Sort-sp-78-385-CFP or Sort-sp-del.MS2-CFP, and the incubation was continued for 2 hours at 4°C with rotation. The lysates were then immunoprecipitated with α-APP and blotted for Sort-78-385 and Sort-del.MS2 by gt-α-GFP, respectively. Cell lysate was used as input. Scramble MS1 peptide (Scramble, 10 µM) and BSA (10 µM) were used as control for non-specific competition. Sortilin constructs and APP695 from co-IP samples and inputs are indicated by arrows.
Figure 5.
Mapping APP binding to sortilin and the binding sequences.
(A) Determining APP 1-287 and APP 713-770 binding to sortilin. HEK293 cells were co-transfected with sort-FL-myc and APP 1-287YFP or APP 1-542YFP or APP 541-671YFP or APP713-770YFP (lanes 1–4). Sort-T-myc was used for the co-transfection with the same APP-YFP constructs (lanes 5–8). Lysates from each co-transfection were used as input (lane 9–12). Cell lysates immunoprecipitated with α-Myc (lanes 1–8) and inputs were blotted with goat (gt)-α-GFP. The co-IP blot was re-probed with α-Myc for sortilin constructs (lower). APP 1-542YFP (87 kDa), APP 1-287YFP (58 kDa), APP 541-671YFP (42 kDa), APP 713-770YFP (33 kDa), Sort-myc (115 kDa) and Sort-T-myc (110 kDa) are indicated by arrows. (B) Determining N terminal binding sites between APP and sortilin. HEK293 cells were co-transfected with APP 1-287YFP and Sort 78-385CFP. Cell lysates were immunoprecipitated with α-APP-N’ for APP1-287YFP (lane 1) and mouse IgG (mIgG) as a control for non-specific binding (lane 2), and blotted with gt-α-GFP. Lysate was used as input (lane 3). APP1-287YFP (58 kDa) and Sort78-385CFP (61 kda) are indicated by arrows. (C) Determining APP 1-141 binding to sortilin. HEK293 cells were co-transfected with APP 1-141YFP and Sort-FL-myc. Cell lysates were immunoprecipitated with α-Myc for Sort-FL-myc (lane 1) and mIgG (lane 2), and blotted with gt-α-GFP for APP1-141YFP. The direct Immunoprecipitated APP1-141YFP by α-GFP was used as input (lane 3). Also, cell lysates were immunoprecipitated with rabbit-α-GFP for APP1-141YFP (lane 4) and rabbite IgG (rIgG) as control (lane 5), and blotted with α-Myc for sortilin. The direct immunoprecipitated Sort-FL-myc by α-Myc was used as input (lane 6). APP1-141YFP (41 kDa) and Sort-FL-myc (115 kDa) are indicated by arrows. (D) Determining C terminal binding sites between APP and sortilin. HEK293 cells were co-transfected with APP 713-770YFP and Sort del.MS2-CFP. Cell lysates were immunoprecipitated with α-APP-C’ for APP 713-770YFP (lane 1) and rIgG (lane 3), and blotted with α-Sort C’. Lysate was used as input (lane 2). Sort del.MS2-CFP (34 kDa) is indicated by arrow. (E) Determining Sort-MS1 (del.MS2) interaction with APP NPTYKFFE motif. Sort del.MS2-CFP was used with APP 713-770-YFP or APP 713-770 mut-YFP for the co-transfection of HEK293 and then subjected to FRET. APP 713-770 mut-YFP construct contains a mutated NPTYKFFE motif where Y and F (underlined) are substituted with A. FRET efficiency representing the protein-protein interaction was determined from a photobleached region of interest (ROI). Three independent experiments were performed. Bars represent mean± SEM (n = 6 ROI×3). The star (*) indicates p<0.01. Abbreviation: negative control: NC; positive control: PC.
Figure 6.
Lack of sortilin reduces APP distribution in lysosome and increases APP distribution in lipid rafts.
Colocalization of APP with cell organelles in cortical neurons. Wild type (WT) and sortilin knockout (KO) mouse cortical neurons were immunostained for APP with mouse anti-APP-N’ (22c11) and followed by staining with Cy3 conjugated secondary antibodies (red) and cell organelles: Golgi, Early endosome, late endosome and lysosome immunostained with Giantin, EEA1, anti-mannose 6 phosphate receptor (for late endosome) and Lamp1, followed by staining with Alexa 488 conjugated secondary antibodies (green). Cell nuclei are stained by DAPI (blue). The colocalization is indicated in merged panels (yellow). Snapshots of colocalization are shown on the right of each panel. The percentage of colocalization (Col.) is plotted and is represented as mean± SEM (n = 20). The colocalization of APP with cell organelles is compared between WT and KO neurons. The star (*) indicates p<0.01. Scale bar 10 µm.
Figure 7.
Lack of sortilin increases APP distribution in lipid rafts in cortical neurons.
Wild type (WT) and sortilin knockout (KO) mouse cortical neurons were immunostained for APP with mouse anti-APP-N’ (22c11) and followed by staining with Cy3 conjugated secondary antibodies (red), and lipid rafts were immunostained with anti-flotillin, followed by staining with Alexa 488 conjugated secondary antibodies (green). Cell nuclei are stained by DAPI (blue). Colocalization is analyzed by counting the number of merged APP/raft lipids (yellow), and is plotted as mean of fold increase ± SEM (n = 20 neurons). The star (*) indicates p<0.01. Scale bar 7.5 µm.
Figure 8.
Effect of sortilin binding domain on APP lysosomal targeting.
A: APP colocalization with lysosome. HEK293 cells were co-transfected with APP770-YFP and different sortilin constructs. The expressed APP and sortilin constructs excluding Sort-FL were visualized by either YFP or CFP fluorescence. Sort-FL-myc/His was immunostained with anti-sortilin, followed by staining with Cy3 conjugated secondary antibodies. Lysosomes were immunostained with lysosomal antibody (Lamp1), followed by staining with Cy5 conjugated secondary antibodies. B: Sortilin constructs sorting ASM to lysosomes. HEK293 cells were transfected with different sortilin constructs. The endogenous ASM was immunostained with mouse anti-ASM (Abcam), followed by staining with Cy3 conjugated secondary antibodies. Lysosome were immunostained with lysosomal antibody (Lamp1), followed by staining with Alexa 488 conjugated secondary antibodies. Sortilin constructs excluding Sort-FL were visualized by CFP fluorescence. Sort-FL-myc/His was immunostained with anti-sortilin, followed by staining with Cy5 conjugated secondary antibodies. Plotted colocalization is indicated at bottom. The percentage of colocalization is represented as mean± SEM (n = 20). The colocalization is compared with Sort-FL. The star (*) indicates p<0.01. Scale bar 7.5 µm.
Figure 9.
Effect of sortilin on APP distribution in lipid rafts.
A: HEK293 cells co-transfected with APP695 and sortilin constructs were lysed in Triton-X-100 buffer and subjected to discontinuous sucrose density gradient ultracentrifugation fractionation. Equal volumes from each fraction were examined by WB for APP and flotillin-1. B: Lack of sortilin increases APP distribution within lipid rafts in sortilin KO mice. Mouse brains were homogenized in Triton-X-100 buffer and subjected to discontinuous sucrose density gradient ultracentrifugation fractionation. Equal volumes from each fraction were examined by WB for APP and flotillin-1. Bars represent mean± SEM (n = 3) from three independent experiments. The star (*) indicates p<0.01.
Figure 10.
Effect of sortilin on APP lysosomal degradation.
APP lysosomal targeting. HEK293 cells were co-transfected with APP-YFP and sortilin-pcDNA3.1 constructs or pcDNA3.1 (mock DNA) plasmid in 1∶1 molar ratio for 24 hours and then treated with DMSO or Bafilomycin AI, a lysosomal inhibitor, (BafA1, 4 µM, Sigma) for 6 h [30]. Cell lysates were harvested and APP level was examined by WB with mouse anti-APP-N’ (22c11). Transfected APP-YFP served to monitor transfection efficiency. The APP level was plotted after correction as to the corresponding β-actin and transfection efficiency.