Association of membrane-bound amyloid precursor protein APP with the apolipoprotein E receptor LRP

Molecular Brain Research 87 (2001) 238–245 www.elsevier.com / locate / bres

Research report

Association of membrane-bound amyloid precursor protein APP with the apolipoprotein E receptor LRP a,

b

a

c

G. William Rebeck *, Robert D. Moir , Stina Mui , Dudley K. Strickland , Rudolph E. Tanzi b , Bradley T. Hyman a a

Alzheimer Research Unit, 149 13 th Street, Massachusetts General Hospital, Charlestown, MA 02129, USA b Neurogenetics and Aging Unit, Massachusetts General Hospital, Charlestown, MA 02129, USA c American Red Cross, Rockville, MD 20855, USA Accepted 19 December 2000

Abstract In order to identify cell surface proteins that interact with the amyloid precursor protein (APP), we biotinylated H4 human neuroglioma cells in culture with a water soluble biotinylating agent, immunoprecipitated APP with an antibody specific to the intracellular domain, and probed the precipitated proteins with anti-biotin. In human neuroglioma cells overexpressing APP751, we found a high molecular weight protein that immunoprecipitated with APP. This band was identified as the low density lipoprotein receptor-related protein (LRP) by three criteria: first, the band immunolabeled with anti-LRP antibodies; second, the band bound the LRP receptor associated protein, RAP; and third, this band was present in LRP-expressing fibroblasts, but not LRP-deficient fibroblasts. In complementary experiments, we found that APP co-precipitated with LRP, with a preference for an isoform of APP containing the Kunitz protease inhibitor domain. Interaction of APP and LRP on the surface of living cells was demonstrated by crosslinking APP and LRP with the water-soluble cross-linking agent BS 3 . APP and LRP were shown by confocal microscopy to colocalize in perinuclear structures, but to primarily remain separate in vesicles and on the cell surface. We propose that full-length APP can transiently interact with the receptor LRP on the cell surface, affecting the processing and intracellular transport of APP.  2001 Elsevier Science B.V. All rights reserved. Theme: Disorders of the nervous system Topic: Degenerative disease: Alzheimer’s - beta amyloid Keywords: Alzheimer’s; Endocytosis; apoE; Lipoprotein; Kunitz protease inhibitor

1. Introduction The amyloid precursor protein (APP) is believed to be central to Alzheimer disease (AD) pathophysiology. APP is the precursor of Ab, a neurotoxic peptide that is the principal component of the amyloid plaques that characterize AD [24]. APP is a large single transmembrane spanning protein expressed as several isoforms by alternative splicing. The three major isoforms contain 695 (APP695), 751 (APP751), or 770 (APP770) amino acids [24]. APP751 and APP770 both contain a Kunitz proteinase inhibitor (KPI) domain. APP can be cleaved at the cell *Corresponding author. Tel.: 11-617-724-8329; fax: 11-617-7265677. E-mail address: [email protected] (G.W. Rebeck).

surface to release soluble extracellular domains (APPs). Mutations in APP lead to rare autosomal dominant forms of familial Alzheimer’s disease [31]. The normal functions of APP, its metabolic pathways, and its role in cell biology are a focus of intense research. KPI-containing isoforms of APP can act as protease inhibitors [27,34], and the soluble forms of KPI-APP can interact with the multifunctional receptor, the low density lipoprotein receptor-related protein (LRP) [14]. A number of other APP-binding proteins have been identified, including Fe65, X11, and Disabled [2,5,32], each of which interact with the C terminal intracellular domain. While LRP has been shown to mediate internalization and degradation of soluble KPI-APP, factors that affect the catabolism of full-length forms of APP are unclear. We [14] and others [10] have postulated that LRP may also

0169-328X / 01 / $ – see front matter  2001 Elsevier Science B.V. All rights reserved. PII: S0169-328X( 01 )00006-7

G.W. Rebeck et al. / Molecular Brain Research 87 (2001) 238 – 245

have an important role in the the catabolism of full length, potentially amyloidogenic forms of APP. In this study, we examined proteins that interact with full length APP on the cell surface. Using three different methods (cell surface labeling, cross-linking and co-immunoprecipitation), we found a physiological interaction between APP and LRP on the cell surface.

2. Materials and methods

239

2.4. Transient transfections The plasmids pHD-APP751 and pHD-APP695, containing the human APP cDNAs under the thymidine kinase promoter (the kind gifts of K.T. Beyreuther, University of Heidelberg) were used for expression of APP. Cells were plated at 6310 5 cells per 100-mm dish. Then 10–20 mg DNA was introduced into cells using lipofectamine (Gibco BRL, Rockville, MD), as per the company’s instructions. The cells were harvested 24–30 h after transfection, and solubilized in 20 mM Tris (pH 7.5), 2 mM CaCl 2 , 2 mM MgCl 2 , 150 mM NaCl, and 1% Triton X-100.

2.1. Cell culture 2.5. Immunoprecipitation Human neuroglioma (H4) cells were obtained from ATTC. H4 cells transfected with APP751 were the kind gift of K. Felsenstein [4]. CHO cells transfected with APP770 were the kind gift of E. Koo [11]. Mouse embryo fibroblasts, PEA13 (LRP deficient) and MEF (wild-type) were the kind gift of J. Herz [37]. Cells were grown at 378C in a humidified 5% CO 2 chamber in Dulbecco’s modified Eagle medium (DMEM) with 10% fetal bovine serum. Media for H4-APP751 and PEA13 was supplemented with G418 (200 mg / ml).

2.2. Antibodies, reagents The mouse monoclonal antibodies 7H5 (a-APP KPI) [8], 8E5 (a-APP extracellular domain, amino acids 444– 592), and 13G8 (a-APP intracellular domain) were the kind gifts of Dr Peter Seubert (Elan Pharmaceuticals). Two rabbit polyclonal antibodies to the C-terminus of APP were used: C7 [6] and CT15 [26] (the kind gift of Dr Eddie Koo, University of California, San Diego). The rabbit polyclonal antibody R777 [15] and the mouse monoclonal antibodies 5A6 and 8G1 [28] were raised against LRP isolated from human placenta. The mouse monoclonal antibody 7F1 recognizes the LRP receptor associated protein (RAP) [13]. Recombinant human RAP was isolated using a RAP-GST construct as previously described [36].

2.3. Biotinylation Subconfluent cells in culture were chilled on ice and labeled with 2 mg / ml of Sulfo-NHS-LC-Biotin (Pierce, Rockford, IL) in TEA buffer (10 mM triethandamine, 125 mM NaCl, 2 mM CaCl 2 , 0.5 mM MgCl 2 , pH 9.0) at 48C for 45 min [12]. Following washing (three times with DMEM), cells were incubated with lysis buffer (50 mM Tris, 130 mM NaCl, 0.02% NaN 3 , 1% NP40, 0.5% DOC, 0.1% SDS) and protease inhibitors (1.5 mM aprotinin; 200 mM PMSF; 2 mM pepstatin A; 4 mM leupeptin). DNA was sheared by repeated passage through a 22-gauge needle.

Immunoprecipitations were carried out with protein G (for LRP; Sigma, St. Louis, MO) or with BioMag Goat anti-Mouse IgG (for APP; PerSeptive Biosystems, Framingham, MA). Cell extracts were pre-incubated with Protein G-Sepharose or BioMag for 1 h at 228C to remove nonspecifically bound proteins. For APP immunoprecipitates, cell extracts were incubated with 2 ml of CT15 overnight at 48C, and precipitated with protein G after 1 h at 228C. Protein G-Sepharose was washed three times in lysis buffer, and once in 10 mM Tris and 0.1% NP40. For LRP, cell extracts were incubated with 2 ml 5A6 and BioMag beads overnight at 48C, and washed four times with phosphate buffered saline.

2.6. Immunoblotting Proteins were separated on 4–12% Tris–glycine polyacrylamide gels (Novex, San Diego, CA) under denaturing, reducing conditions and transfered to immobilon-P membrane (Millipore, Bedford, MA). Membranes were blocked in Superblock (Pierce, Rockford, IL) or 5% non-fat dried milk. For detection of biotinylated proteins, mouse antibiotin was used, and proteins visualized with 125 I-conjugated anti-mouse antibody (Amersham Pharmacia, Piscataway, NJ) using a Molecular Dynamics PhosphorImager. For detection of APP, both antibodies to the extracellular domain (8E5, 1:2500) and intracellular domain (13G8, 1:1000) were used. For detection of LRP, membranes were probed with R777 (1:1000). Secondary antibodies linked to horse radish peroxidase were visualized by chemiluminescence. Exposed films were scanned on a BioRad GS-700 densitometer.

2.7. Cross-linking H4-APP751 were pre-incubated in media with or without 2 mM RAP for 1 h at 378C. Following washing with cold PBS, cells were incubated at 48C in the presence or absence of 1 mg / ml of the membrane impermeable conjugative reagent bis (sulfosuccinimidyl) suberate (BS 3 , Pierce, Rockford, IL). The cross-linking reaction was

240

G.W. Rebeck et al. / Molecular Brain Research 87 (2001) 238 – 245

quenched with 25 mM Tris. Cells were lysed and nuclei and remaining whole cells removed by centrifugation (5003g for 2 min). Cell plasma membranes were pelleted by a second centrifugation (12 0003g for 30 min) and then extracted with SDS sample buffer (without b-mercaptoethanol). SDS membrane extracts, purified LRP (LRP) and partially purified membrane-associated APP (APP) were Western blotted and probed with 8G1 (anti-LRP) or C7 (anti-APP C-terminus) antibodies.

2.8. Immunostaining H4 cells on slides were fixed in 4% paraformaldehyde for 20 min; some slides were incubated in 1% Triton X-100 to permeablize cells. Cells were blocked in 1.5% normal goat serum, and probed with 7H5 (1:200) and R777 (1:200). Antibodies were detected with fluoresceinlabeled anti-rabbit and Cy3-labeled anti-mouse secondary antibodies and imaged on a BioRad 1024 confocal microscope.

3. Results

3.1. Interaction between full-length APP and cell surface proteins

biotinylating agent at 48C as described [12]. Cell extracts were then immunoprecipitated with CT15, an antibody to the C-terminal 15 amino acids of APP. Immunoprecipitated proteins were subjected to SDS–PAGE, transferred to membrane and immunoblotted or probed with an antibiotin antibody. In H4 human neuroglioma cells transfected with APP751, we found an anti-biotin immunoreactive band at 140 kDa (Fig. 1A, lane 2), indicating the form of APP on the cell surface. This band co-migrated with an anti-APP immunoreactive band consistent with the mature, fulllength form of APP (Fig. 1A, lane 1). We did not observe the lower molecular weight precursor (110 kDa) form of APP; thus the data are consistent with biotinylation of surface, but not intracellular forms of APP, with this protocol. In addition to the biotinylated band identified as APP, we observed anti-biotin immunoreactivity at apparent molecular weights of |250 and 50 kDa (Fig. 1A, lane 2). Experiments using CHO cells transfected with APP770 yielded similar results to those observed for APP751transfected H4 cells (Fig. 1B). However, co-precipitation of the high molecular weight biotinylated protein was not observed in the CHO cells. These data suggest that the interacting protein is present in H4 neuroglioma cells, but not CHO cells.

3.2. Identification of LRP in APP immunoprecipitates In order to identify cell surface APP and its cell surfaceassociated proteins, we treated cells with a water-soluble

We hypothesized that the high molecular weight species

Fig. 1. APP and cell surface proteins. Cells were treated with 2 mg / ml of Sulfo-NHS-LC-Biotin (a membrane-insoluble biotinylating agent) at 48C for 45 min. (A) Lane 1: extracts of H4-APP751 cells were probed with the anti-APP antibody 8E5, and peroxidase-linked secondary antibody. Lane 2: extracts of H4-APP751 cells were immunoprecipitated with the anti-C-terminal APP antibody CT15, and probed with mouse anti-biotin and 125 I-labeled anti-mouse antibodies. (B) Extracts of CHO-APP770 cells, immunoprecipitated with CT15, and probed with mouse anti-biotin and 125 I-labeled anti-mouse antibodies. Numbers indicate molecular weight markers, in kDa.

G.W. Rebeck et al. / Molecular Brain Research 87 (2001) 238 – 245

immunoprecipitated with APP in the H4 cells in Fig. 1 were forms of LRP. We had previously observed that on some polyacrylamide gradient gels, LRP (which is |500 kDa) runs anomalously, at slightly greater than 250 kDa. Moreover, we found that LRP is present at high levels in H4 cells, but at negligible levels in CHO cells. To test whether the high molecular weight protein was LRP, we immunoprecipitated H4-APP751 cell extracts with CT15 antibody, and probed with an anti-LRP antibody. A band of LRP immunoreactivity was detected (Fig. 2A), which co-migrated with the high molecular weight biotinylated protein detected in previous experiments (Fig. 1). Control precipitates from samples lacking cell extract or CT15 antibody did not contain this band. Furthermore, purified LRP was not immunoprecipitated by CT15. To confirm the identity of the high molecular weight band, two further analyses were performed. First, immunoprecipitated proteins on the membrane were probed with the 39-kDa RAP, which is a ligand for LRP [36]. The membrane was then probed with an anti-RAP antibody. Consistent with the identity of the band as LRP, anti-RAP immunoreactivity was detected at the same apparent molecular weight (Fig. 2B). In a second experiment, extracts from wild-type and LRP deficient mouse fibroblast cells were studied [37]. These fibroblast cells were first transiently transfected to overexpress high levels of APP751 (Fig. 2C, lower panel). A large LRP immunoreactive band was observed in the immunoprecipitate from the LRP-positive cells, but was lacking in extracts from the LRP-negative cells (Fig. 2C, upper panel). Thus, three lines of evidence support the

241

identification of LRP in APP immunoprecipitate: LRP immunoreactivity, RAP binding, and its presence in LRPpositive but not LRP-deficient cell extracts.

3.3. Identification of APP in LRP immunoprecipitates Previous data demonstrated that LRP is found after immunoprecipitation of APP. We reversed this experiment, and probed for the presence of APP after immunoprecipitation of LRP. H4 cell extracts showed prominent expression of both smaller, immature APP and larger, mature, glycosylated forms of APP (Fig. 3A, lane 2). After pre-clearing H4 cell extracts, we reproducibly observed low levels of both immature and mature forms of APP upon immunoprecipitation with anti-LRP antibody (Fig. 3A, lane 1). A 150-kDa band was also observed in the immunoprecipitates (Fig. 3A, lane 1), indicating low levels of non-reduced LRP antibody. H4 cells endogenously express primarily KPI-containing forms of APP, as evidenced by the larger bands of APP immunoreactivity in H4 cells compared to H4 cells transiently overexpressing APP695 (Fig. 3, lanes 6 and 7). In order to test whether APP interacted with LRP via its KPI domain, we immunoprecipitated LRP from extracts of H4 cells transiently transfected with APP695 or APP751. Cells transfected with APP695 demonstrated smaller APP immunoreactive bands than in cells transfected with APP751 or non-transfected cells, indicating that the predominant APP present was APP695 (Fig. 3C). When the LRP immunoprecipitates were probed for APP with an

Fig. 2. Co-immunoprecipitation of LRP with APP. Extracts from H4-APP751 cells were immunoprecipitated with anti-APP (C-terminus) antibody CT15, and probed with (A) anti-LRP antibody, R777 or (B) the LRP binding protein RAP, followed by an anti-RAP antibody. LRP is recognized as a .250-kDa immunoreactive band. (C) Extracts from MEF (lane 1, LRP1) and PEA13 (lane 2, LRP2) fibroblasts were transiently transfected with APP751. Lower panel: immunoblot of APP in cell extracts, using the APP antibody, 8E5. Upper panel: LRP immunoblot (R777) after APP immunoprecipitation of cell extracts.

242

G.W. Rebeck et al. / Molecular Brain Research 87 (2001) 238 – 245

Fig. 3. Co-immunoprecipitation of APP with LRP. Cell extracts were prepared from untransfected H4 cells (lanes 1,2,3, 6, and 9) or H4 cells transiently transfected with APP695 (lanes 4, 7, and 10) or APP751 (lanes 5, 8, and 11). Extracts were immunoprecipitated with anti-LRP antibody 5A6 and probed for APP with an antibody to the extracellular domain (8E5: (A) lane 1; (B) lanes 3–5) or to the cytoplasmic domain (13G8: (D) lanes 9–11). Cell extract proteins from H4 cells (lane 2) or proteins that were not immunoprecipitated after 5A6 treatment (C, lanes 6–8) were probed for APP with 8E5. Bands of immunoreactivity correspond to mature and immature forms of the APP splice variants.

N-terminal APP antibody, several prominent bands between 110 and 140 kDa were observed in cells overexpressing APP 695 and APP751 (Fig. 3B, lanes 4 and 5), with mature forms of APP predominating. Levels of APP immunoprecipitated from APP695- and APP751-transfected cells were significantly higher than from untransfected cells. The relative intensities of APP751 and APP695 bands were reversed in the extracts after immunoprecipitation, reflecting APP which did not immunoprecipitate with LRP (Fig. 3C). We used densitometry (within the linearity of the films) to compare the relative amounts of APP in the immunoprecipitates (Fig. 3B) with the APP in the post-IP fractions (Fig. 3C). APP751 was immunoprecipitated with LRP at levels nearly four-fold greater than APP695, consistent with our previous observations that LRP has a higher affinity for APP751 than APP695 [14]. Several APP immunoreactive bands precipitated with LRP, possibly representing soluble and full-length forms of APP. To test whether full-length APP was present, we probed the proteins immunoprecipitated with LRP with a C-terminal APP antibody (13G8). This antibody recognized a single band in the immunoprecipitate from APP751-expressing cells, but not APP-695 expressing cells (Fig. 3D, lanes 10 and 11). Thus, in confirmation of the data in Fig. 2, cellular LRP associates with full length APP, and this association is stronger for KPI-containing forms of APP.

3.4. Interaction of APP and LRP on the cell surface It is possible that the APP/ LRP complexes observed in immunoprecipitates formed after the proteins were extracted from plasma membranes. To test whether APP and LRP were associated in living cells, a membrane impermeable NHS-ester conjugative reagent (BS 3 ) was used. H4 cells were pre-incubated with or without RAP, which prevents binding of KPI-APP to LRP [14]. Cells were then incubated with or without BS 3 . Following repeated washing and treatment with quenching buffer, the cells were harvested, lysed, and soluble and membrane fractions were prepared and extracted with detergent. The membrane extracts were immunoblotted and probed with anti-LRP or anti-APP antibodies. Consistent with cell surface LRP binding to full-length integral membrane forms of APP, a high molecular weight band immunoreactive to both antibodies was observed in cells incubated with crosslinker (Fig. 4). In order to test whether this complex represented an association of APP with the ligand binding domain of LRP, we utilized RAP, which antagonizes the binding of all known ligands to LRP. This conjugated complex was substantially decreased in cells pre-incubated with RAP (Fig. 4), suggesting APP and LRP interact via their extracellular domains. These data are consistent with APP and LRP binding at the cell surface of neuroglioma cells.

G.W. Rebeck et al. / Molecular Brain Research 87 (2001) 238 – 245

243

Fig. 4. Cross-linking of LRP and full-length forms of APP. H4 cells stably transfected with APP751 were pre-incubated in media with (1) or without (2) 2 mM RAP for 1 h at 378C. Following washing with cold PBS, cells were incubated in the presence or absence of 1 mg / ml of the membrane impermeable conjugative reagent BS 3 at 48C. SDS membrane extracts, purified LRP (LRP) and partially purified membrane-associated APP (APP) were Western blotted and probed with 8G1 (anti-LRP) or C7 (anti-APP C-terminus) antibodies. Consistent with cell surface LRP binding to full-length integral membrane forms of APP, a high molecular weight band immunoreactive to both antibodies was observed in cells incubated with cross-linker.

To further extend our study of H4 cells, we used confocal microscopy to localize APP and LRP both in permeablized and non-permeablized cells. In permeablized cells, LRP and APP localized to cell surface and intracellular vesicles. APP and LRP co-localized to a great extent in the perinuclear structures, presumably endoplasmic reticulum and Golgi apparatus (Fig. 5, arrowheads with yellow indicating co-localization). APP and LRP remained primarily separate in vesicles (Fig. 5, arrows, APP in red, LRP in green), although some vesicles showed co-localization (yellow or orange). In non-permeablized cells (Fig. 5, upper right inserts), APP and LRP both appeared on the cell surface, but, again, they mostly remained separate. There were small regions in most cells of surface colocalization.

4. Discussion The findings described in this study support a model of direct binding between full-length plasma membrane APP and LRP in H4 neuroglioma cells. Full-length APP and LRP can be co-immunoprecipitated from cell extracts (Figs. 2 and 3), and can be cross-linked on the cell surface (Fig. 4). These observations, in combination with our previous data that LRP can bind and internalize soluble KPI-APP [14], and data from Knauer et al. [10] that LRP can internalize cell surface APP/ proteinase complexes, support the hypothesis that interactions between full length

APP and LRP are robust and functionally relevant. Since LRP is an important receptor for lipoprotein uptake, the binding of APP with LRP would provide a further link between APP processing and cholesterol metabolism [25]. LRP is a 600-kDa multi-functional receptor belonging to the family of low density lipoprotein receptors [29]. LRP has over 19 extracellular ligands, including lipoproteinassociated molecules and several proteinase inhibitors. Binding of ligands to LRP directs uptake and degradation via endosomal / lysosomal compartments. Interactions of APP and LRP could target APP to the endosomal / lysosomal pathway. This model is consistent with our observations by confocal microscopy (Fig. 5) that most of the APP and LRP do not co-localize in intracellular vesicles or at the cell surface. Presumably the APP/ LRP complex is short-lived, leading to endosomal and lysosomal degradation of APP. Data from several laboratories have shown that surface APP is recycled into endosomes and there acts as a precursor for Ab [6,11,21]. Recent support for a role of LRP in directing Ab generation comes from studies of cells containing LRP versus cells functionally deficient in LRP [33]. One series of experiments demonstrated that inhibition of LRP with RAP decreased the internalization of cell surface APP751 and the production of the Ab peptide. A second series of experiments demonstrated that introduction of LRP into LRP deficient cells expressing APP751 dramatically increased the production of Ab [33]. Thus, endosomal proteolysis of APP751 is promoted by interaction with LRP.

244

G.W. Rebeck et al. / Molecular Brain Research 87 (2001) 238 – 245

Fig. 5. APP and LRP subcellular localization. H4 cells were stained with 7H5 (a-APP, imaged in red), and R777 (a-LRP, imaged in green). Colocalization of APP and LRP is imaged in yellow. Inserts in upper right corner show stained H4 cells from non-permeablized cultures. Arrows indicate vesicles that contain APP (red), LRP (green) or both APP and LRP (yellow / orange). Arrow heads indicate areas of co-localization of APP and LRP in perinuclear structures.

with full-length APP was proposed by Trommsdorff et al., who showed that APP and LRP cytoplasmic tails can bind to separate domains of the cytoplasmic protein FE65 [32]. FE65 has two phosphotyrosine binding domains which bind NPXY motifs; the first of these binds to cytoplasmic LRP constructs, and the second binds to APP. The potential cytoplasmic bridge formed by FE65 between APP and LRP may underlie the weak association between APP695 and LRP shown in Fig. 3. We suggest that LRP activity in the brain may thus be important in Alzheimer pathophysiology. This hypothesis is supported by several lines of evidence from in vivo systems: (a) LRP and each of the known LRP ligands are present on senile plaques [23]; (b) soluble forms of KPIAPP [19] are also elevated in the AD brain; (c) other LRP ligands are known to be important in AD pathological mechanisms, including apoE, which is clear genetic risk factor [30] and which binds Ab [16,17], and a2-macroglobulin, which has been suggested as a genetic risk factor [1,18] and which also binds Ab and clears it via LRP [20,22]; (d) the LRP gene itself has a silent polymorphism which has been linked to risk for AD [7,9]. The interaction of LRP with apolipoprotein E, a2macroglobulin, and full-length APP places three molecules of interest in Alzheimer’s disease in a single metabolic pathway. This common LRP-mediated mechanism of uptake and degradation suggests that apoE and a2-macroglobulin could alter the metabolism of both APP and Ab. In principle, these types of interactions could underlie the pathophysiological mechanisms accounting for Ab deposition in Alzheimer’s disease.

Acknowledgements Our earlier work with the soluble forms of APP suggested that KPI-containing APP interacts with LRP [14]. Two findings in this study strongly support an interaction for full-length KPI-containing form of APP and LRP: the preferential co-immunoprecipitation of KPI-APP with LRP (Fig. 3); and the inhibition of APP-LRP crosslinking after addition of RAP. This interaction is likely to be similar to the binding demonstrated for another KPI-containing protein, tissue factor pathway inhibitor [35]. It is likely that the KPI domain of APP binds directly to LRP, but it is also possible that the presence of the KPI domain changes the structure of APP, exposing an otherwise cryptic binding site. Knauer et al. found that full-length KPI-APP would not be internalized via LRP until it complexed a serine protease, and then the protease / inhibitor complex would be cleared by LRP [10]. Thus, LRP may be a part of a protective system for inhibiting protease activity near the cell surface. This model suggests that KPI-APP would undergo differential processing compared to non-KPI-APP, which could explain the observation that KPI-APP produces more Ab than non-KPI-APP [3]. Another plausible mechanism for the association of LRP

We thank Dr Eddie Koo for his assistance with the cell surface labeling studies and for his gift of antibodies. We also thank Dr Peter Seubert for his gift of antibodies. This work was funded by a grant from the American Health Assistance Foundation, and NIH grants AG14473 (G.W.R.) and AG12406 (B.T.H.).

References [1] D. Blacker, M.A. Wilcox, N.M. Laird, L. Rodes, S.M. Horvath, R.C.P. Go, R. Perry, B. Watson, S.S. Bassett, M.G. McInnis, M.S. Albert, B.T. Hyman, R.E. Tanzi, Alpha-2 macroglobulin is genetically associated with Alzheimer disease, Nat. Genet. 19 (1998) 357–360. [2] J.-P. Borg, J. Ooi, E. Levy, B. Margolis, The phosphotyrosine interaction domains of X11 and FE65 bind to distinct sites on the YENPTY motif of amyloid precursor protein, Mol. Cell. Biol. 16 (1996) 6229–6241. [3] L. Do, K. Fukuchi, S.G. Younkin, The alternatively spliced Kunitz protease inhibitor domain alters amyloid b protein precursor processing and amyloid b protein production in cultured cells, J. Biol. Chem. 271 (1996) 30929–30934. [4] K.M. Felsenstein, L.W. Hunihan, S.B. Roberts, Altered cleavage and

G.W. Rebeck et al. / Molecular Brain Research 87 (2001) 238 – 245

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[20]

secretion of a recombinant beta-APP bearing Swedish familial Alzheimer’s disease mutation, Nat. Genet. 6 (1994) 251–255. S.Y. Guenette, J. Chen, P.D. Jondro, R.E. Tanzi, Association of a novel human FE65-like protein with the cytoplasmic domain of the b-amyloid precursor protein, Proc. Natl. Acad. Sci. USA 93 (1996) 10832–10837. C. Haass, E.H. Koo, A. Mellon, A.Y. Hung, D.J. Selkoe, Targeting of cell-surface b-amyloid precursor protein to lysosomes: alternative processing into amyloid-bearing fragments, Nature 357 (1992) 500– 503. E. Hollenbach, S. Ackermann, B.T. Hyman, G.W. Rebeck, Confirmation of an association between a polymorphism in exon 3 of the low density lipoprotein receptor-related protein gene and Alzheimer’s disease, Neurology 50 (1998) 1905–1907. B.T. Hyman, R.E. Tanzi, K.M. Marzloff, R. Barbour, D. Schenk, Kunitz protease inhibitor containing amyloid precursor protein immunoreactivity in Alzheimer’s disease: a quantitative study, J. Neuropathol. Exp. Neurol. 51 (1992) 76–83. D.E. Kang, T. Saitoh, X. Chen, Y. Xia, E. Masliah, L.A. Hansen, R.G. Thomas, L.J. Thal, R. Katzamn, Genetic association of the low-density lipoprotein receptor-related protein gene (LRP), an apolipoprotein E receptor, with late-onset Alzheimer’s disease, Neurology 49 (1997) 56–61. M.F. Knauer, R.A. Orlando, C.G. Glabe, Cell surface APP751 forms complexes with protease nexin 2 ligands and is internalized via the low density lipoprotein receptor-related protein (LRP), Brain Res. 740 (1996) 6–14. E.H. Koo, S.L. Squazzo, Evidence that production and release of amyloid b-protein involves the endocytic pathway, J. Biol. Chem. 269 (1994) 17386–17389. E.H. Koo, S.L. Squazzo, D.J. Selkoe, C.H. Koo, Trafficking of cell surface amyloid b-protein precursor. 1. Secretion, endocytosis and recycling as detected by labeled monoclonal antibody, J. Cell Sci. 109 (1996) 991–998. M.Z. Kounnas, W.S. Argraves, D.K. Strickland, The 39-kDa receptor-associated protein interacts with two members of the low density lipoprotein receptor family, a2-macroglobulin receptor and glucoprotein 330, J. Biol. Chem. 267 (1992) 21162–21166. M.Z. Kounnas, R.D. Moir, G.W. Rebeck, A.I. Bush, W.S. Argraves, R.E. Tanzi, B.T. Hyman, D.K. Strickland, LDL receptor-related protein, a multifunctional apoE receptor, binds secreted b-amyloid precursor protein and mediates its degradation, Cell 82 (1995) 331–340. M.Z. Kounnas, R.E. Morris, M.R. Thompson, D.J. FitzGerald, D.K. Strickland, C.B. Saelinger, The a2-macroglobulin receptor / low density lipoprotein receptor-related protein binds and internalizes Pseudomonas exotoxin A, J. Biol. Chem. 267 (1992) 12420–12423. M.J. LaDu, T.M. Pederson, D.E. Frail, C.A. Reardon, G. Getz, M.T. Falduto, Purification of apolipoprotein E attenuates isoform-specific binding to b-amyloid, J. Biol. Chem. 270 (1995) 9039–9042. A. Lalazar, K.H. Weisgraber, S.C. Rall Jr., H. Giladi, T.L. Innerarity, A.Z. Levanon, J.K. Boyles, B. Amit, M. Gorecki, R.W. Mahley, T. Vogel, Site-specific mutagenesis of human apolipoprotein E: receptor binding activity of variants with single amino acid substitutions, J. Biol. Chem. 263 (1988) 3542–3545. A. Liao, R.M. Nitsch, S.M. Greenberg, U. Finckh, D. Blacker, M. Albert, G.W. Rebeck, T. Gomez-Isla, A. Clatworthy, G. Binetti, C. Hock, T. Mueller-Thomsen, U. Mann, K. Zuchowski, U. Beisiegel, H. Staehelin, J.H. Growdon, R.E. Tanzi, B.T. Hyman, Genetic association of an alpha-2 macroglobulin (Val 1000 Ile) polymorphism and Alzheimer’s disease, Hum. Mol. Genet. 7 (1998) 1953– 1956. R.D. Moir, T. Lynch, A.I. Bush, S. Whyte, A. Henry, S. Portbury, G. Multhaup, D.H. Small, R.E. Tanzi, K. Beyreuther, C.L. Masters, Relative increase in Alzheimer’s disease of soluble forms of cerebral Ab amyloid protein precursor containing the Kunitz protease inhibitory domain, J. Biol. Chem. 273 (1998) 5013–5019. M. Narita, D.M. Holtzman, A.L. Schwartz, G. Bu, a2-Macroglobulin

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30] [31]

[32]

[33]

[34]

[35]

[36]

[37]

245

complexes with and mediates the endocytosis of b-amyloid peptide via cell surface low-density lipoprotein receptor-related protein, J. Neurochem. 69 (1997) 1904–1911. G.C. Peraus, C.L. Masters, K. Beyreuther, Late compartments of amyloid precursor protein transport in SY5Y cells are involved in b-amyloid secretion, J. Neurosci. 17 (1997) 7714–7724. Z. Qiu, D.K. Strickland, B.T. Hyman, G.W. Rebeck, a2-macroglobulin enhances the clearance of endogenous soluble b-amyloid peptide via low density lipoprotein receptor-related protein in cortical neurons, J. Neurochem. 73 (1999) 1393–1398. G.W. Rebeck, S.D. Harr, D.K. Strickland, B.T. Hyman, Multiple, diverse senile plaque-associated proteins are ligands of an apolipoprotein E receptor, the a 2 -macroglobulin receptor / low-densitylipoprotein receptor-related protein, Ann. Neurol. 37 (1995) 211– 217. D.J. Selkoe, Cell biology of the amyloid b-protein precursor and the mechanism of Alzheimer’s disease, Annu. Rev. Cell Biol. 10 (1994) 373–403. M. Simons, P. Keller, B. De Strooper, K. Betreuther, C.G. Dotti, K. Simons, Cholesterol depletion inhibits the generation of b-amyloid in hippocampal neurons, Proc. Natl. Acad. Sci. USA 95 (1998) 6460–6464. S.S. Sisodia, E.H. Koo, P.N. Hoffman, G. Perry, D.L. Price, Identification and transport of full-length amyloid precursor proteins in rat peripheral nervous system, J. Neurosci. 13 (1993) 3136–3142. R.P. Smith, D.A. Higuchi, G.J. Broze Jr., Platelet coagulation factor XIa-inhibitor, a form of Alzheimer amyloid precursor protein, Science 248 (1990) 1126–1128. D.K. Strickland, J.D. Ashcom, S. Williams, W.H. Burgess, M. Migliorini, W.S. Argraves, Sequence identity between the a2-macroglobulin receptor and low density lipoprotein receptor-related protein suggests that this molecule is a multifunctional receptor, J. Biol. Chem. 265 (1990) 17401–17404. D.K. Strickland, M.Z. Kounnas, W.S. Argraves, LDL receptorrelated protein: a multiligand receptor for lipoprotein and proteinase catabolism, FASEB J. 9 (1995) 890–898. W. Strittmatter, A. Roses, Apolipoprotein E and Alzheimer’s disease, Annu. Rev. Neurosci. 19 (1996) 53–77. R.E. Tanzi, G. Vaula, D.M. Romano, M. Mortilla, T.L. Huang, R.G. Tupler, W. Wasco, B.T. Hyman, J.L. Haines, B.J. Jenkins, M. Kalaitsidaki, A.C. Warren, M.G. McInnis, S.E. Antonarakis, H. Karlinsky, M.E. Percy, L. Connor, J. Growdon, D.R. CrapperMclachlan, F.J. Gusella, P.H. St. George-Hyslop, Assessment of amyloid b protein precursor gene mutations in a large set of familial and sporadic Alzheimer disease cases, Am. J. Hum. Genet. 51 (1992) 273–282. M. Trommsdorff, J.-P. Borg, B. Margolis, J. Herz, Interaction of cytosolic adaptor proteins with neuronal apolipoprotein E receptors and the amyloid precursor protein, J. Biol. Chem. 273 (1998) 33556–33560. P.G. Ulery, J. Beers, I. Mikhailenko, R.E. Tanzi, G.W. Rebeck, B.T. Hyman, D.K. Strickland, Modulation of b-amyloid precursor protein processing by the low density lipoprotein receptor-related protein (LRP), J. Biol. Chem. 275 (2000) 7410–7415. W.E. Van Nostrand, A.H. Schmaier, J.S. Farrow, D.D. Cunningham, Protease nexin-II (amyloid b protein precursor): a platelet a-granule protein, Science 248 (1990) 745–748. I. Warshawsky, G.J. Broze Jr., A.L. Schwartz, The low density lipoprotein receptor-related protein mediates the cellular degradation of tissue factor pathway inhibitor, Proc. Natl. Acad. Sci. USA 91 (1994) 6664–6668. S.E. Williams, J.D. Ashcom, W.S. Argraves, D.K. Strickland, A novel mechanism for controlling the activity of a2-macroglobulin receptor / low density lipoprotein receptor-related protein, J. Biol. Chem. 267 (1992) 9035–9040. T.E. Willnow, J. Herz, Genetic deficiency in low density lipoprotein receptor-related protein confers cellular resistance to Pseudomonas exotoxin A, J. Cell Sci. 107 (1994) 719–726.