Low Density Lipoprotein Receptor-related Protein (LRP) Interacts with Presenilin 1 and Is a Competitive Substrate of the Amyloid Precursor Protein (APP) for γ-Secretase*

Presenilin 1 (PS1) is a critical component of the γ-secretase complex, which is involved in the cleavage of several substrates including the amyloid precursor protein (APP) and the Notch receptor. Recently, the low density receptor-related protein (LRP) has been shown to be cleaved by a γ-secretase-like activity. We postulated that LRP may interact with PS1 and tested its role as a competitive substrate for γ-secretase. In this report we show that LRP colocalizes and interacts with endogenous PS1 using coimmunoprecipitation and fluorescence lifetime imaging microscopy. In addition, we found that γ-secretase active site inhibitors do not disrupt the interaction between LRP and PS1, suggesting that the substrate associates with a γ-secretase docking site located in close proximity to PS1. This is analogous to APP-γ-secretase interactions. Finally, we show that LRP competes with APP for γ-secretase activity. Overexpression of a truncated LRP construct consisting of the C terminus, the transmembrane domain, and a short extracellular portion leads to a reduction in the levels of the Aβ40, Aβ42, and p3 peptides without changing the total level of APP expression. In addition, transfection with the β-chain of LRP causes an increase in uncleaved APP C-terminal fragments and a concomitant decrease in the signaling effects of the APP intracellular domain. In conclusion, LRP is a PS1 interactor and can compete with APP for γ-secretase enzymatic activity.

LRP is cleaved by furin in the trans-Golgi network, generating a 515-kDa ␣-subunit and a 85 kDa ␤-subunit that remain non-covalently associated as they are transported to the cell surface (14). LRP also undergoes proteolytic shedding of the extracellular domain by a metalloproteinase (15). It has recently been shown that the cytoplasmic tail of LRP can be processed intramembranously by a ␥-secretase activity that releases its intracellular domain (16). The ␥-secretase complex is a multiprotein complex that is composed of at least four members, namely presenilin 1 (PS1), which is believed to contain the catalytic site, nicastrin, Pen-2, and Aph1 (17). This complex is responsible for cleavage of at least 15 substrates, including APP and the Notch receptor. APP is a type-I transmembrane protein that is cleaved by ␣or ␤-secretase to remove its large extracellular domain. The remaining 83-residue (APPC83) or 99-residue (APPC99) fragments are cleaved intramembraneously by ␥-secretase to release the fragment p3 or the amyloid-␤ peptide (A␤), respectively (18). ␥-Secretase also cleaves near the inner leaflet to generate an APP intracellular domain (AICD), which may have signaling properties (19,20). LRP has a complex set of interactions with APP. LRP interacts both at the extracellular domain of the Kunitz protease inhibitor-containing isoform of APP (APP770 and APP751) and with the intracellular domain of all isoforms of APP via adaptor proteins such as Fe65 (9,10,21,22). LRP has also been shown to modulate A␤ clearance and APP processing (23)(24)(25)(26)(27).
In this study we explore whether LRP directly interacts with PS1 (28), and acts as a competitive substrate for ␥-secretase. We (29,30) and others (31,32) have shown previously that Notch and APP can compete for ␥-secretase activity. Here we extend the study to test whether LRP is also a competitive substrate of APP for ␥-secretase. We show that LRP interacts with PS1, binds to a docking site on PS1/␥-secretase, and can compete with APP for the enzymatic activity.

MATERIALS AND METHODS
Plasmids, Cell Lines, and Transfection Protocol-A summary of the LRP constructs used in this study is presented in Fig. 1. The LC construct contains the ␤-subunit of LRP with a Myc tag at the N terminus (33). The mutant LC consists of the ␤-subunit of LRP with a Myc tag at the N terminus in which asparagines and the tyrosine in both NPXY motifs have been substituted by alanines (34). The LRP-CT construct, consisting of the last 370 amino acids of the 601-amino acid LRP ␤-subunit, was subcloned into the pLHCX retroviral expression vector and transferred into the 293 GP packaging cell line (27).
Chinese hamster ovary (CHO) cells, human neuroblastoma (H4) cells, human embryonic kidney cells, and mouse embryonic fibroblasts lacking PS1 (MEF PS1Ϫ/Ϫ) or MEF PS1Ϫ/Ϫ stably expressing human wild-type PS1 (35) were used in this study. Cells were cultured in OPTI-MEM with 5% fetal bovine serum at 37°C with 5% CO 2 in a tissue culture incubator. For A␤ and APP internalization measurements, CHO cells stably expressing APP751 with the LRP C terminus (7WD10-CT) or without the LRP C terminus (7WD10) were used (27). Transient transfection was performed using Superfect reagents (Qiagen, Valencia, CA) according to the manufacturer's instructions. Cells were treated with 1 M DAPT or 1 M WPE-III-31C (a generous gift from M. Wolfe, Brigham and Women's Hospital, Boston, MA) or a vehicle control (0.1% dimethyl sulfoxide) for 24 h (36,37).
Western Blot and Coimmunoprecipitation-For Western blot analysis the cell lysates were adjusted to equal protein concentrations and electrophoresed on 10 -20% SDS-polyacrylamide Tris-glycine or 10% Bis-Tricine NuPAGE gels (Invitrogen). We used the following antibodies: the polyclonal anti-LRP antiserum 1704 (27); the monoclonal anti-LRP antibody 11H4; the monoclonal anti-PS1 antibody (PSN2) (38); the polyclonal PS1 antibodies X81 (a gift from Dr. D. Selkoe, Brigham and Women's Hospital, Boston, MA) and 4627 (39); the polyclonal anti-APP antiserum R3134 (25) or the monoclonal anti-APP antibody 26D6 (25), both of which recognize A␤; and the polyclonal anti-APP antibody CT15 (40) or the APP C8 antibody (41), both of which recognize the cytoplasmic domain of APP. For endogenous PS1-LRP coimmunoprecipitation from the normal rat brain (3-month-old Sprague-Dawley rats), we used anti-PS1 antibodies (X81 or 4627) for a pull-down assay and the 11H4 LRP antibody for blotting. The immunoblotting was followed by detection with a horseradish peroxidase-conjugated secondary antibody and enhanced chemiluminescence substrate (PerkinElmer Life Sciences). Bands on films were scanned and quantitated using Odyssey software (LI-COR, Lincoln, Nebraska).
A␤ Enzyme-linked Immunosorbent Assay Measurements and APP Internalization Assay-Media from 7WD10 and 7WD10-CT were collected after 48 h, debris was removed by centrifugation (13,000 rpm for 20 min), and the supernatants were subjected to A␤ 40 and A␤ 42 quan-tification using a standard sandwich enzyme-linked immunosorbent assay described previously (25). To measure internalization of cell surface APP, iodinated 1G7 antibody was added to confluent cultures of 7WD10 and 7WD10-CT cells at 37°C for 30 min. After incubation, the cells were rapidly chilled on ice, and the reaction was quenched by the addition of ice-cold binding medium. Cells were washed extensively, and the remaining surface antibody was detached by two acid washes. The cells were then lysed and collected for analysis. Acid-labile and acid-resistant radioactivity represents the surface and internalized pools of APP, respectively. The ratio of acid-resistant to acid-labile counts, therefore, provided a measure of the internalized versus cell surface pools of APP. The experiments were repeated three times.
Fe65-dependent APP Luciferase Transactivation Assay-Human embryonic kidney cells were cotransfected with the APP-Gal4 construct (a gift from Dr. Thomas Sü dhof, University of Texas Southwestern Medical Center, which was cloned from APP 695 with the Gal4 sequence spliced into the cytoplasmic tail), pG5E1B-Luc (a luciferase reporter plasmid with a Gal4 activation-dependent binding domain), Fe65, and a ␤-galactosidase plasmid (19). The generation of Fe65 with a C-terminal Myc tag has been reported elsewhere (42). Relative luciferase activity was measured and normalized to ␤-galactosidase activity.
Immunocytochemistry and Antibodies-Cells were fixed and immunostained 24 after transfection as described (43). Briefly, cells were fixed in 4% paraformaldehyde for 10 min, washed in Tris-buffered saline (pH 7.3), permeabilized by 0.5% Triton X-100 for 20 min, and blocked with 1.5% normal goat serum for 1 h. We used an antibody directed against the N terminus of PS1 (X81) and a mouse monoclonal antibody, 11H4, against the LRP C-terminal fragment (a hybridoma secreting 11H4 was obtained from the American Type Culture Collection, Manassas, VA). For the fluorescence lifetime imaging microscopy (FLIM) assays a FITC-labeled secondary antibody was used as a donor fluorophore, and a Cy3-labeled secondary antibody was used as an acceptor fluorophore.
Metabolic Labeling-Confluent cultures of CHO 7WD10 and CHO 7WD10 LRP-CT cells were incubated in methionine-free Dulbecco's modified Eagle's medium supplemented with 150 Ci/ml [ 35 S]methionine/cysteine for 1 h. Media were collected after 12 h, and total A␤ was immunoprecipitated with the polyclonal antibody R3134. The immunoprecipitates were fractionated by SDS-PAGE (6% Tris-glycine gels for full-length APP and 4 -12% Tris-Tricine gels for APP C-terminal fragments) and exposed to either film or phosphorimaging for quantitation.
Fluorescence Lifetime Imaging Microscopy Assay-FLIM has been recently described as a novel technique for the analysis of protein proximity (44 -46). The technique is based on the observation that fluorescence lifetimes of a donor fluorophore shorten in the presence of a FRET acceptor in close proximity (Ͻ10 nm). The decrease in lifetime is proportional to the distance between the fluorophores at R 6 . A modelocked titanium-sapphire laser (Spectra-Physics, Fremont, CA) emits a femtosecond pulse every 12 ns to excite the fluorophore. A high-speed Hamamatsu (Bridgewater, NJ) detector and hardware/software (SPC-830 Becker and Hickl, Berlin, Germany) were used to measure fluorescence lifetimes on a pixel-by-pixel basis. Donor fluorophore (FITC) lifetimes were fitted to two exponential decay curves as described (44,45). All samples were compared with a negative control in which the donor fluorophore (FITC) fluorescence lifetime was measured in the absence of the acceptor (no FRET) and was ϳ2300 ps. As a positive control, FITC lifetime was measured in the presence of a FRET acceptor (Cy3) in close proximity. In the positive control, FITC lifetime is shortened to Ͻ1000 ps.
Statistical Analysis-One-way analysis of variance (ANOVA) was performed to analyze differences in lifetime followed by least significant difference post hoc analysis. Levene's test was also performed to determine whether variances were equal.

Colocalization, FRET, and Coimmunoprecipitation Show
That LRP Interacts with Presenilin 1-Because LRP is cleaved by a ␥-secretase activity and PS1 contains the catalytic site of the multi-protein ␥-secretase complex (28), we first tested whether LRP interacts directly with PS1. We transfected CHO cells with wild-type PS1 and wild-type LRP/LC and double immunostained the cells with antibodies against PS1 and the C terminus of LRP. Using confocal microscopy, we observed that wild-type LRP/LC colocalized with PS1 mainly in the endoplasmic reticulum (Fig. 2). However, because colocalization has a spatial resolution of Ͼ250 nm, we used FLIM to test whether FIG. 1. Summary of the constructs used in this study. Because most ␥-secretase substrates require shedding of their extracellular domain, we used LC, a short form of LRP that contains the ␤-chain of LRP, and LRP-CT, the LRP C-terminal fragment that contains the last 370 amino acids of the ␤-chain of LRP. We used two antibodies to detect LRP, 11H4, and 1704, both directed against the C terminus of LRP. mutLC, mutant LC. PS1 and LRP interact more closely. FLIM is based on the observation that the lifetime of a donor fluorophore is shorter in the presence of a FRET acceptor in close proximity. We transfected CHO cells with wild-type PS1 and wild-type LC and immunostained the cells with a FITC-labeled antibody against the N terminus of PS1 and a Cy3-labeled antibody against the C terminus of LRP. In the absence of an acceptor, FITC lifetime was ϳ2300 ps. However, immunostaining of LRP with a Cy3-labeled antibody shortened the FITC lifetime markedly, indicating that the fluorophores are in close proximity (Table I). This shortening in FITC lifetime was specific, because immunostaining Bip, an endoplasmic reticulum-resident protein that colocalizes with PS1, with a Cy3-labeled antibody did not alter FITC lifetime. In cells transfected with PS1 and LRP/LC we observed that the strongest interactions occur close to the cell membrane (Fig. 3) in the same type of spatial patterns that we have observed previously for two other ␥-secretase substrates, APP and Notch (44,47). As expected, the lifetime pattern in cells stained for Bip did not differ from that of the negative control. Next, we examined whether the NPXY domains of LRP are necessary for the interaction of LRP with PS1, which might be the case if the interaction was mediated by an adaptor protein. We transfected cells with wildtype PS1 and a mutant LRP/LC construct in which both NPXY domains have been mutated and immunostained the cells for PS1 and LRP. Cells transfected with mutant LC showed similar patterns of expression and colocalization with PS1 as wildtype LC (Fig. 2), and FLIM analysis detected significant FRET between PS1 and mutant LC ( Table I). The main interacting molecules were localized close to the cell membrane, similar to what was observed in wild-type LC-expressing cells (Fig. 3). These data suggest that both wild-type LC and mutant LC interact with PS1 and that the NPXY domains of LRP are not essential for PS1-LRP interactions. The observed interaction between LRP and PS1 in transfected cells was also confirmed at the endogenous level in primary cultured neurons, as reflected by a significant shortening in donor lifetime when cells were double immunostained for LRP and PS1.
To further confirm this interaction, we performed coimmunoprecipitation experiments using an LRP-specific antibody (1704) for immunoprecipitation of an endogenous LRP and the N-terminal PS1 antibody (PSN2) for the subsequent Western blotting. These experiments were performed in mouse embryonic fibroblast PS1Ϫ/Ϫ cells expressing wild-type human PS1 and as control in PS1Ϫ/Ϫ cells. In cells expressing PS1, but not in PS1Ϫ/Ϫ cells, complexes of N-terminal fragments of PS1 can be recovered by immunoprecipitation with an LRP antibody (Fig. 4). Importantly, we did not detect any interaction of fulllength PS1 with LRP (Fig. 4), suggesting that LRP interacts primarily with the active heterodimeric form of PS1. We then performed a reverse coimmunoprecipitation experiment using PS1 N-terminal (X81) or PS1 C-terminal (4627) antibodies for immunoprecipitation of the endogenous PS1 from a normal rat brain and the LRP-CT antibody (11H4) for Western blotting. We found two LRP-positive bands representing LRP full-length ␤-chain and a C-terminal fragment (Fig. 4b) in a PS1-immunoprecipitated (right lane) sample, but not in a no-antibody immunoprecipitated (left lane) sample. We interpret this as an indication of a direct interaction between PS1 and LRP in mammalian cells.
␥-Secretase Inhibitors Do Not Disrupt the Interaction between LRP/LC and Presenilin-Next, we asked whether the relative proximity between LRP/LC and PS1 can be modified by conventional ␥-secretase inhibitors. H4 cells were transfected with LRP/LC and wild-type PS1 and treated with 1 M DAPT, 1 M WPE-III-31C, or a vehicle control. After 24 h, cells were immunostained with a FITC-labeled antibody against the C terminus of LRP and a Cy3-labeled antibody against the N terminus of PS1 for FLIM analysis. As expected, in cells treated with a vehicle FIG. 2. LRP/wild-type LC and LRP/mutant LC (mutLC) colocalize with presenilin 1. Both LRP/wild-type LC and LRP/mutant LC colocalize with presenilin 1, mainly in the endoplasmic reticulum, according to confocal microscopy imaging. CHO cells were transfected with wild-type LC or mutant LC and PS1 and immunostained for PS1 and the C terminus of LRP (11H4 antibody).

TABLE I LRP/wild-type LC and LRP/mutant L interact with presenilin1
We transfected CHO cells with wild-type LC or mutant LC (mutLC) and PS1. Twenty-four hours later the cells were fixed and immunostained for PS1 and LRP with a FITC-labeled antibody and a Cy3labeled antibody, respectively. We measured FITC lifetime in the absence or presence of a FRET acceptor. In the absence of an acceptor, FITC lifetime was ϳ2300 ps, but in the presence of LC or mutant LC the FITC lifetime was markedly shortened. Immunostaining of Bip, an endoplasmic reticulum-resident protein that colocalizes with PS1 to a great extent, did not change FITC lifetime, indicating that the shortening observed due to LRP was specific. Values presented are from a representative experiment (similar results were obtained in eight independent experiments, 6 -10 cells per condition).

FIG. 3. LRP/wild-type LC and LRP/mutant LC interact with presenilin 1 as assessed by FLIM.
CHO cells were transfected with wild-type LC or mutant LC (Mut-myc-LC) and PS1. Twenty-four hours later cells were immunostained for PS1 and LRP with a FITC-labeled antibody and a Cy3-labeled antibody, respectively. We measured FITC lifetime in the absence or presence of a FRET acceptor, and the results were fitted to two exponential curves. The values can also be represented in a pseudo-colored image, which ranges from blue (no FRET) to red (FRET). In the absence of an acceptor, FITC lifetime was ϳ2300 ps and appears pseudo-colored blue. In the presence of LC or mutant LC, FITC lifetime was markedly shortened, and pixels with short lifetimes appear pseudo-colored red. Immunostaining of Bip, an endoplasmic reticulum-resident protein, did not change FITC lifetime, indicating that the shortening was specific for LRP. control the FITC lifetime was shortened, indicating that both fluorophores are in close proximity (Table II). Treatment with either DAPT or WPE-III-31C did not significantly change FITC lifetime as compared with a vehicle treatment, suggesting that the availability of the active site is not critical for the LRP-PS1 interactions observed. This result is analogous to observations with APP-PS1 in which a stable docking site interaction can be detected when the active site is occupied (44).
Both Wild-type LRP/LC and Mutant LRP/LC Are Competitive Substrates for ␥-Secretase-Because both APP and LRP are substrates of ␥-secretase, we asked whether LRP overexpression could affect the ␥-secretase cleavage of APP. We metabolically labeled CHO cells stably transfected with APP751 and with LRP-CT (7WD10-CT) or without LRP-CT (7WD10) and measured the A␤ and p3 peptides, the two ␥-cleaved products of APP. We observed that the levels of A␤ and p3 were markedly reduced in cells overexpressing LRP-CT as measured by immunoprecipitation (Fig. 5a). Interestingly, when we analyzed different clones of 7WD10-CT cells we observed an inverse relation between the levels of expression of LRP-CT and the levels of A␤ secretion (Fig. 5, b and c). We also measured A␤ 40 and A␤ 42 using an enzyme-linked immunosorbent assay and observed that both A␤ species were decreased when LRP-CT was overexpressed (Fig. 5d). Next, we analyzed whether the difference in A␤ secretion could be due to different rates of APP internalization by using radiolabeled APP antibodies. We did not find differences in APP internalization rates between cells overexpressing APP751 and those overexpressing APP751 and LRP-CT (Fig. 5e). We also examined whether transfection with LRP could alter the levels of APP C-terminal fragments (CTFs), the direct substrates of ␥-secretase. We observed that, in CHO cells transfected with LRP/LC and APP770, a small increase in the levels of APP CTFs could be seen compared with cells transfected with an empty vector and APP770 (Fig. 5f). It has been shown that ␥-secretase also cleaves APP in the inner leaflet of the membrane to release the AICD (19). We asked whether transfection with LRP would have any impact on AICD generation. We postulated that, if LRP and APP competed for ␥-secretase, we should observe a reduction in AICD generation after transfection with LRP. To measure AICD generation, we measured the downstream signaling effects of APP-Gal4 (19). Human embryonic kidney cells were cotransfected with APP-Gal4, pG5E1B-Luc, Fe65, and either wild-type LC, mutant LC, or an empty vector. An expected, an increase in the ability of APP-Gal4 to stimulate transcription was observed when Fe65 was cotransfected (Fig. 6). In addition, we found that both wild-type LRP/LC and mutant LRP/LC both attenuated this phenomenon, decreasing the signaling effects of the Fe65-stabilized AICD generated from the Gal4-dependent luciferase reported plasmid. The attenuation was not due to differences in the levels of expression of APP or Fe65 (Fig. 6). Although LRP/LC might, in principle, have competed for Fe65, the mutant LRP/LC cannot interact with Fe65, suggesting that the inhibitory effect is at the ␥-secretase level. DISCUSSION Since the making of the initial discovery that APP is cleaved by PS1-dependent ␥-secretase (28, 48 -50), multiple other substrates that undergo ␥-secretase cleavage have been identified, including Notch, Erb-B4, E-and N-cadherins, CD44, nectin-1, the Notch ligands Delta and Jagged, and LRP (17,51). It is not clear whether these substrates are all cleaved by the same PS1dependent activity or by the same population of enzyme complexes. We have now examined this issue with regard to LRP; we find that LRP interacts with PS1 in distal cell compartments as assessed by FLIM and coimmunoprecipitation experiments and that presentation of a truncated form of LRP can inhibit A␤ production, increase APP CTFs, and decrease AICD generation. We interpret these data to be consistent with the idea that LRP is a competitive inhibitor of APP, implying that the two are cleaved by the same population of ␥-secretase complexes.
LRP is a transmembrane glycoprotein implicated in diverse biological processes both as an endocytic receptor and as a signaling molecule. Recent reports show that LRP undergoes cleavage by a ␥-secretase activity that results in the release of the LRP cytoplasmic domain (16). This LRP cytoplasmic domain also can interact with Fe65 and has been shown to translocate to the nucleus where it can interact with the transcription modulator Tip60 (52). Notch is also a single transmembrane domain protein whose ␥-secretase-dependent cleavage initiates a transcriptional cascade. We and others have demonstrated previously that Notch and APP compete for the same PS1-dependent ␥-secretase activity (29 -32). Our current data on LRP support the possibility that multiple substrates can compete for ␥-secretase activity.
Accumulating evidence suggests that ␥-secretase is an aspartyl protease and that two conserved aspartates in presenilins are catalytic residues (28,53). The direct binding of transition-state analogue ␥-secretase inhibitors to PS1 fragments strongly suggest that the active site is at the interface between the N-and C-terminal fragments and that each subunit contributes one of the two critical aspartates (37,53). In addition, FIG. 4. LRP coimmunoprecipitates with presenilin 1. a, fractions of mouse embryonic fibroblast PS1Ϫ/Ϫ cells expressing wild-type (wt) human PS1 and PS1Ϫ/Ϫ cells as a control were immunoprecipitated (IP) with a LRP-specific antibody (1704) and analyzed by immunoblotting (WB, Western blot) with the N-terminal PS1 antibody PSN2. In these cells, complexes of N-terminal fragments (NTF) but not of full-length (fl) PS1 can be recovered by immunoprecipitation. b, an endogenous LRP coimmunoprecipitated from a rat brain lysate using the PS1 antibody to pull down an endogenous PS1.  (36,37). After 24 h we immunostained for LRP and PS1 with a FITC-labeled antibody and a Cy3-labeled antibody, respectively. We measured FITC lifetime in the absence or presence of a FRET acceptor. As expected, in the absence of an acceptor FITC lifetime was ϳ2300 ps, but in the presence of LRP/LC the FITC lifetime was markedly shortened. Treatment with ␥-secretase inhibitors did not significantly change FITC lifetime. Values from a representative experiment are presented. Similar results were observed in four separate experiments (9 -10 cells per condition). recent biochemical evidence points to the existence of an initial substrate binding site in ␥-secretase distinct from the active site. Isolation of the ␥-secretase complex with an immobilized transition state analogue co-purifies with an endogenous APP substrate, suggesting that the substrate can bind at some other site (54). Moreover, APP and PS1 remain in close proximity in the presence of ␥-secretase transition state analogues as measured by FRET (44). In this study we found that LRP behaves much like APP in terms of its interactions with ␥-secretase; we can detect strong FRET between LRP and PS1, and this close proximity persists after treatment with ␥-secretase active site inhibitors. These data are consistent with the notion of an initial substrate binding site in or in close proximity to PS1. Because most ␥-secretase substrates require a prior shedding of the extracellular domain, in this study we used short forms of LRP, namely LC, which contains the entire ␤-chain of LRP, and the C-terminal fragment of LRP (LRP-CT), which contains the last 370 residues of the ␤-chain of LRP (27,34). First, we show that overexpressing LRP-CT reduces A␤ 40 , A␤ 42 , and the p3 peptide and that this effect was not due to a change in APP expression level or APP internalization. In these cell lines we observed an inverse relation between the levels of expression of FIG. 5. Overexpression of LRP-CT leads to a reduction in A␤ 40 , A␤ 42 , and p3 and an increase in APP C-terminal fragments. a, we measured the A␤ and p3 peptides, the two ␥-cleaved products of APP, in CHO cells stably transfected with APP751 and with LRP-CT (7.WD10-CT) or without LRP-CT (7.WD10) by immunoprecipitation. We observed that the levels of A␤ and p3 were markedly reduced in cells overexpressing LRP-CT. MW, molecular weight. b and c, we analyzed four different clones of 7WD10-CT cells and observed an inverse relation between the levels of expression of LRP-CT and the levels of A␤ secretion. d, we also measured A␤ 40 and A␤ 42 using an enzyme-linked immunosorbent assay and observed that both were decreased when overexpressing LRP-CT as compared with cells overexpressing only APP. e, the difference in A␤ secretion is not explained by different rates of APP internalization. We did not find differences in APP internalization between cells overexpressing APP751 and those overexpressing APP751 and LRP-CT using radiolabeled APP antibodies. f, CHO cells were transfected with APP770 and LC or an empty vector. In a Western blot, we observed a small increase in APP C-terminal fragments when we transfected with LC compared with an empty vector. No differences were seen in full-length (FL) APP. A representative blot is shown. Similar results were obtained in three separate experiments.
LRP-CT and the levels of A␤. Next, we found that transfection with LRP/LC, which contains the ␤-chain of LRP, or with a mutant form of LRP/LC impairs the transcriptional activity of the APP-Gal4 construct. Of note, the fact that the mutant form of LRP/LC also decreased the transcriptional activity of the APP-Gal4 construct suggests that APP and LRP do not compete for binding to the adaptor Fe65 or for binding to Tip60, which is Fe65-dependent. Finally, we also observed that cotransfection of CHO cells with LC was able to increase the levels of APP CTFs when compared with cells transfected with an empty vector. However, we cannot rule out the possibility that this increase is due to the effects of LRP on APP CTF stability rather than competition for ␥-secretase (27).
APP is also a transmembrane type I protein, which is processed by ␣-, ␤-, and ␥-secretase, and cleavage by the latter generates the A␤ and the p3 peptides. In addition to intramembraneous cleavage, ␥-secretase also cleaves APP at the inner leaflet of the membrane to generate AICD. The released AICD is thought to form a complex with the adaptor protein Fe65 and the histone acetyltransferase Tip60, which might have a role in transcriptional regulation (19,20,22). APP and LRP have been shown to display a complex set of interactions. In addition to the observations that LRP mediates the clearance of A␤ protein in vitro and possibly in vivo (23)(24)(25)55), LRP has been shown to interact with the Kunitz protease inhibitor-containing isoforms of APP (9). LRP can also modulate APP processing and A␤ generation (26). LRP also modulates the levels of APP CTFs and facilitates APP turnover and internalization (27). These effects were not limited to Kunitz protease inhibitor-containing isoforms of APP, and the critical region that modulates APP processing was mapped to a seven-peptide domain around the second NPXY domain of LRP (27). Our current observations show yet another level of interaction between APP and LRP as competitive substrates for ␥-secretase. Interestingly, the APP-LRP interactions noted above depend in great part on adaptor proteins like Fe65 forming a heterodimeric complex, and these interactions are blocked by NPXY mutants. By contrast, the interaction of LRP with PS1/␥-secretase is not dependent on the NPXY domains. It will be of interest to further examine the structural correlates of LRP-PS1/␥-secretase interactions.
In summary, we show that LRP colocalizes and directly interacts with PS1 by using confocal microscopy, FLIM, and coimmunoprecipitation assays. This result suggests that LRP ␥-secretase cleavage occurs via the same PS1-containing complex that is responsible for APP cleavage. We also found that the mutant LRP/LC, in which both NPXY domains have been substituted, colocalized and interacted with PS1 as assessed by confocal microscopy and FLIM. These data indicate that the NPXY domains are not required for the PS1-LRP interactions. In addition, we show that the ␥-secretase inhibitors WPE-III-31C and DAPT do not disrupt the observed interaction between LRP and PS1 in intact cells, consistent with the possibility that LRP/LC can occupy the docking site.
Finally, under some circumstances LRP is able to reduce the ␥-secretase-dependent cleavage of APP that generates the A␤ peptide, p3, and AICD. Taken together, these data support a model in which LRP and APP are competitive ␥-secretase substrates.