Autosomal recessive hypercholesterolemia protein interacts with and regulates the cell surface level of Alzheimer's amyloid beta precursor protein.

The familial Alzheimer's disease gene product amyloid beta protein precursor (A beta PP) is sequentially processed by beta- and gamma-secretases to generate the A beta peptide. Although much is known about the biochemical pathway leading to A beta formation, because extracellular aggregates of A beta peptides are considered the cause of Alzheimer's disease, the biological role of A beta PP processing is only recently being investigated. Cleavage of A beta PP by gamma-secretase releases, together with A beta, a COOH-terminal A beta PP intracellular domain, termed AID. Hoping to gain clues about proteins that regulates A beta PP processing and function, we used the yeast two-hybrid system to identify proteins that interact with the AID region of A beta PP. One of the interactors isolated is the autosomal recessive hypercholesterolemia (ARH) adapter protein. This molecular interaction is confirmed in vitro and in vivo by fluorescence resonance energy transfer and in cell lysates. Moreover, we show that reduction of ARH expression by RNA interference results in increased levels of cell membrane A beta PP. These data assert a physiological role for ARH in A beta PP internalization, transport, and/or processing.

familial forms of AD are found in A␤PP itself and in the highly homologous genes PS1 and PS2, which are key components of a multimolecular complex with ␥-secretase activity (6 -15).
Although the role of A␤ peptides in the pathogenesis of AD has been extensively studied, reports as to the role of AID are very recent. AID-like peptides have been identified in human brains from cases of sporadic AD (16) and play a role in apoptosis (16), transcription (17)(18)(19)(20), and Ca 2ϩ release from the endoplasmic reticulum (21).
A␤PP processing and AID signaling can be regulated by A␤PP-interacting proteins (17)(18)(19)(20)(21)(22). Thus, to find A␤PP-binding proteins we employed the yeast two-hybrid selection system. This screening resulted in the identification of several proteins that bind the intracellular domain of A␤PP. Here we report the novel A␤PP interactor ARH, an adapter protein that has been shown to regulate cholesterol uptake by genetic studies (22). These data suggest that ARH may be a mediator of the well described effect of cholesterol metabolism on A␤PP processing.

MATERIALS AND METHODS
Yeast Two-hybrid System-The two-hybrid screening was conducted using the Matchmaker system from Clontech according to the manufacturer's instruction. For library screening, Yeast190 expressing GAL4BD-AID fusion proteins were transformed with a human fetal brain cDNA library cloned in the pACT2 vector (Clontech). 2 ϫ 10 6 clones were analyzed. Transformed yeast were selected in synthetic drop-out plates lacking tryptophan, leucine, and histidine in the presence of 50 mM 3-aminotriazole (Sigma) and grown for 5 days at 30°C. Colonies growing on selective media were scored as positive. Assays were done for eight independent transformants.
cDNA Cloning and Constructs-The GAL4BD-A␤PP bait was constructed using the pAS2 vector (Clontech) and consisted of the COOHterminal 58 amino acids of A␤PP fused to the DNA binding domain of GAL4, respectively. A␤PP, A␤PPNcas, and AID were made as described previously (16). GST fusion proteins were made in pGEX vectors (Amersham Biosciences). Mutations were introduced by using the transformer site-directed mutagenesis kit (Clontech).
Cell Lines and Transfections-Human embryonic kidney (HEK) 293T cells were grown in RPMI 1640 media (Invitrogen) supplemented with glutamine and with 10% heat-inactivated fetal calf serum (Biofluids; Rockville, MD). Transfections were performed in 6-well plates either using Metafectene (Biontex Laboratories Gmbh) with 3 l per 1 g of DNA.
Northern Blot Analysis-A multitissue blot was purchased from Clontech, and it was hybridized with a 32 P-labeled ARHPTB probe following the manufacturer's instruction. After washing, the blot was developed by autoradiography.
GST Pull-down, Immunoprecipitation, and Immunoblot Analysis-* This work was supported by the Irene Diamond Foundation grant (to L. D'A.) and the Telethon Italia grant (to C. N.).The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Recombinant GST fusion proteins were expressed in Escherichia coli strain BL21 (Invitrogen) to make non-phosphorylated proteins and strain TKB1 (Stratagene) to make tyrosine-phosphorylated proteins using the pGEX system (Amersham Biosciences) (23). Proteins were purified using glutathione-Sepharose beads.
For GST pull down of proteins produced in vitro (Fig. 2a), [ 3 H]leucine-labeled proteins were made using the TNT-coupled in vitro transcription/translation system (Promega). After synthesis of the radiolabeled protein for 1.5 h, aliquots of the protein were incubated with GST fusion proteins bound to glutathione-Sepharose beads for 2 h at room temperature. The beads were then washed three times with lysis buffer T (1% Triton X-100, 50 mM Tris-HCl, pH 7.4, 45 mM NaCl) and boiled with SDS loading buffer with DTT. The proteins were separated by SDS-PAGE, and the gel were fixed with 50% methanol, 40% H 2 O, 10% acetic acid. The gel was incubated in Amplify (Amersham Biosciences) for 20 min and dried, and signals were detected using autoradiography.
For GST-pull down of proteins produced in vivo (Fig. 2b), HEK293 cells were lysed in lysis buffer T containing a protease inhibitor tablet (Roche Applied Science) 24 -48 h following transfection. Lysis was allowed to continue for 10 min on ice and was then spun down at full speed at 4°C for 10 min. Some lysate representing the total lysate was removed and boiled with SDS loading buffer with DTT, whereas the rest was pulled-down with GST fusion proteins. The beads were washed five times with lysis buffer and boiled with SDS loading buffer with DTT. The proteins were separated by SDS-PAGE and blotted onto nitrocellulose (Bio-Rad). Membranes were probed with either ␣FLAG or rabbit polyclonal ␣ARH (Zymed Laboratories Inc., new product catalog number 36-0400) followed by horseradish peroxidase-conjugated secondary antibodies (Southern Biotech). Proteins were detected using the Supersignal West Pico chemiluminescent system (Pierce).
For immunoprecipitation from transfected cells, lysates were immunoprecipitated for 2 h at room temperature with the ␣FLAG monoclonal antibody bound to agarose beads (Sigma). After washing and SDS-PAGE, membranes were probed with either the 22C11 monoclonal antibody directed against the ectodomain of A␤PP (Calbiochem) or rabbit polyclonal ␣ARH.
For co-precipitation of endogenous proteins, untransfected HEK293 cells were lysed as above. One mg of protein was used for immunoprecipitation with rabbit polyclonal ␣APP COOH-terminal (Zymed Laboratories Inc., new product), rabbit polyclonal ␣ARH, or rabbit antimouse IgG antibodies (ICN, Aurora, OH). Immunoprecipitations were performed overnight at 4°C, followed by incubation of immunoprecipitates with protein A-agarose beads, washing, and immunoblot as described above.
FRET Analysis-HEK293T cells were plated in 24-well plates and co-transfected with the CFP and YFP fusion proteins using Metafectene. All transfections contained ratios between the two cDNAs determined empirically to yield the best co-expression as follows. Cells were harvested between 18 and 24 h after transfection in their conditioned media. The analysis was performed as described previously (24).

RESULTS AND DISCUSSION
To identify proteins interacting with the cytoplasmic domain of A␤PP, we have used the yeast two-hybrid selection system. This screening resulted in the identification of two independent clones coding for a novel protein (the interaction in yeast between AID and the representative AT60 clone is shown in Fig.  1a). Northern blot analysis showed that AT60 is expressed, albeit at different levels, in all human tissues analyzed (Fig.  1b). A recent blast search showed that AT60 is identical to the recently identified ARH adapter protein (22) (Fig. 1c). Homozygous ARH mutations cause familial forms of autosomal recessive hypercholesterolemia (ARH). ARH contains a phosphotyrosine-binding domain (PTB), which is known to interact with the PTB-binding motif present in the cytoplasmic tail of A␤PP ( 682 YENPTY 687 ) (19, 23, 24, 28 -32). Tyrosine phosphorylation of Tyr 682 of A␤PP is important for its interaction with ShcA and ShcC (23). On the contrary, other binding partners interact with A␤PP regardless of Tyr 682 phosphorylation. Therefore, we sought to determined whether the PTB domain of ARH mediated binding to the 682 YENPTY 687 of A␤PP and whether phosphorylation of A␤PP was necessary for binding of ARHs. First, 3 H-labeled ARH (ARHPTB), Fe65 COOH-terminal (Fe65C-PTB, which binds A␤PP), and NH 2 -terminal (Fe65N-PTB, which does not bind A␤PP, Refs. 28 and 30) PTB domains were produced and incubated with GST-AID or GST alone (all bound to glutathione-Sepharose beads). The protein-protein com-FIG. 1. ARH interacts with the A␤PP cytoplasmic region in yeast. a, the interaction between AT60 and AID in yeast is specific because AT60 does not interact with SV40 large T and AID does not bind p53. The interaction between p53 and large T is shown as a positive control. b, Northern blot analysis showing that AT60/ARH is ubiquitously expressed in adult human tissues. c, sequence alignment showing that AT60 is identical to ARH. The AT60 clone is missing the NH 2terminal 32 amino acids of ARH. The PTB domain is indicated by the lighter shading. The Pro to Ser mutation may represent a polymorphism of ARH. plexes were purified, resolved by gel electrophoresis, and visualized by autoradiography. Fig. 2a shows that GST-AID specifically and directly binds both ARHPTB and Fe65C-PTB but not Fe65N-PTB. Next, HEK293 cells were transfected with constructs expressing either ARH or ARH⌬PTB (a mutant lacking the PTB domain) of FLAG-tagged Fe65. Cell lysates were incubated with GST-AID, GST-AID phosphorylated on Tyr 682 (GST-AID P ), GST-AID Y682G (in which tyrosine 682 was mutated to glycine), or GST, and pull downs were resolved by gel electrophoresis. Western blot analysis (Fig. 2b) using either ␣ARH or ␣FLAG antibodies revealed that ARH, like Fe65, interacted with AID, whereas ARH⌬PTB did not. We also found that, similarly to Fe65, A␤PP binds ARH independent of Tyr 682 phosphorylation, but this binding is abolished by the Y682G mutation.
In further experiments, HEK293 cells were co-transfected with FLAG-ARH and the following A␤PP constructs: wild type A␤PP, A␤PPNcas (an A␤PP mutant lacking the COOH-terminal 31 amino acids that include the YENPTY motif), A␤PP Y682F (in which Tyr 682 was mutated to phenylalanine), A␤PP Y687A (in which Tyr 687 was mutated to alanine), or A␤PP T653F (in which Thr 653 was mutated to phenylalanine). Cell lysates were immunoprecipitated with ␣FLAG monoclonal antibody and analyzed by Western blot. Fig. 3a shows that although FLAG-ARH immunoprecipitates A␤PP, it does not immunoprecipitate A␤PPNcas which lacks the YENPTY motif. Additionally, A␤PP Y682F mutation significantly affects the interaction with ARH.
To detect interaction between ARH and A␤PP in living cells, we used fluorescence resonance energy transfer (FRET) (25)(26)(27). HEK293T cells were co-transfected with yellow fluorescent protein-AID (Y-AID) and either cyan fluorescent protein-ARH (C-ARH) or C-ARH⌬PTB fusion proteins. In FRET on living cells, if the proteins are in close proximity, on the order of 10 nm or less, the energy from the excitation of CFP will be transferred to YFP, and emission at the wavelength of YFP will be detected. If the proteins are not within this proximity, excitation of CFP is not transferred, and only emission at the wavelength of CFP is detected. Importantly, FRET was detected when Y-AID was co-transfected with C-ARH but not with C-ARH⌬PTB. Altogether, these data indicate that ARH interacts with the YENPTY motif of A␤PP through its PTB domain. Furthermore, this interaction is independent of Tyr 682 phosphorylation but requires the presence of Tyr 682 .
To determine whether endogenous A␤PP and ARH interact, we immunoprecipitated HEK293 lysates either with an ␣A␤PP or an ␣ARH polyclonal antibody. As shown in Fig. 4, which is representative of data from two independent experiments, A␤PP was immunoprecipitated with both ␣A␤PP and ␣ARH antibodies, whereas A␤PP was not immunoprecipitated with a control rabbit anti-mouse IgG antibody (R␣M). Similar findings were obtained with the reverse experiment, that is ARH was  immunoprecipitated by the ␣A␤PP and the ␣ARH antibodies but not by the R␣M IgG control. Altogether, these experiments indicate that endogenous A␤PP and ARH associate.
To determine whether ARH physiologically regulates A␤PP biology, we have repressed ARH protein expression in HEK293 cell using RNA interference (RNAi), which involves the transfection of small interfering RNA (siRNA) duplexes in cells. We have designed siRNAs duplex specific for human ARH, A␤PP, and a duplex of unrelated scrambled sequence. As shown in Fig. 5a, ARH and A␤PP siRNAs specifically reduce expression of ARH and A␤PP proteins, respectively. Because ARH is involved in of low density lipoprotein receptors endocytosis, we next assessed whether ARH may regulate the cell surface levels of A␤PP. To this end, we have performed FACS analysis experiments. The monoclonal antibody P2-1, which is specific for the ectodomain of human A␤PP, was used in this experiment. The isotype-matched and unrelated monoclonal antibody P3 was used as a negative control. Viable cells were initially stained with either P2-1 or P3. After washing out the unbound antibody, cells were incubated with a phycoerythrin-labeled anti-mouse IgG secondary antibody and analyzed by FACS. P2-1 specifically binds to the cell surface A␤PP of HEK293 cells as determined by three complementary facts: (i) the isotypematched P3 antibody does not bind HEK293 cells (Fig. 5b); (ii) overexpression of A␤PP in HEK293 cells increases P2-1 binding (Fig 5b); (iii) conversely, reduction of A␤PP protein levels by RNAi results in decreased P2-1 binding (Fig. 5c). Of interest, reduction of ARH protein levels by RNAi results in increased amounts of A␤PP on the cell membrane of HEK293 cells (Fig.  5c). These data indicate that ARH physiologically regulates the cell surface level of A␤PP.
In this study we have demonstrated in vitro, in vivo, in living cells, and for endogenous proteins the interaction between A␤PP and ARH. Moreover, we have shown that ARH regulates A␤PP cell membrane levels. These changes in A␤PP cell surface levels in ARH-low cells reflect a physiological role for ARH in either A␤PP internalization, transport to the cell membrane, and/or shedding of the ectodomain by secretases (␣ and/or ␤). Although further work will be required to discriminate among these possibilities, our data nevertheless prove a function for ARH in A␤PP biology.
Genetic defects in ARH impair endocytosis of low density lipoprotein receptors family members (22,33,34), thereby reducing cholesterol uptake and increasing plasma concentration of cholesterol. Of interest, biochemical, epidemiological, and genetic evidence have involved cholesterol metabolism in the regulation of A␤PP processing and the pathogenesis of AD (35)(36)(37)(38). The finding that ARH interacts with A␤PP and regulates the cell membrane levels of A␤PP is provocative and entices us to speculate that ARH may be a molecular link between A␤PP processing and cholesterol.