Novel Cadherin-related Membrane Proteins, Alcadeins, Enhance theX11-like Protein-mediated Stabilization of Amyloid �-ProteinPrecursor Metabolism*□S

Received for publication, June 9, 2003, and in revised form, September 10, 2003Published, JBC Papers in Press, September 12, 2003, DOI 10.1074/jbc.M306024200

Yoichi Araki‡§¶, Susumu Tomita‡¶�, Haruyasu Yamaguchi**, Naomi Miyagi‡, Akio Sumioka‡§,Yutaka Kirino§, and Toshiharu Suzuki‡ ‡‡

From the ‡Laboratory of Neuroscience, Graduate School of Pharmaceutical Sciences, Hokkaido University, Kita-ku Kita-12 Nishi-6, Sapporo 060-0812, the §Laboratory of Neurobiophysics, School of Pharmaceutical Sciences, The University ofTokyo, Bunkyo-ku Hongo 7-3-1, Tokyo 113-0033, and the **College of Medical Care and Technology, School of Medicine,Gunma University, Showa-machi 3-39-15, Maebashi 371-8514, Japan

Previously we found that X11-like protein (X11L) as-sociates with amyloid �-protein precursor (APP). X11Lstabilizes APP metabolism and suppresses the secretionof the amyloid �-protein (A�) that are the pathogenicagents of Alzheimer’s disease (AD). Here we found thatAlcadein (Alc), a novel membrane protein family thatcontains cadherin motifs and originally reported as cal-syntenins, also interacted with X11L. Alc was abundantin the brain and occurred in the same areas of the brainas X11L. X11L could simultaneously associate with APPand Alc, resulting in the formation of a tripartite com-plex in brain. The tripartite complex stabilized intracel-lular APP metabolism and enhanced the X11L-mediatedsuppression of A� secretion that is due to the retarda-tion of intracellular APP maturation. X11L and Alc alsoformed another complex with C99, a carboxyl-terminalfragment of APP cleaved at the �-site (CTF�). The for-mation of the Alc�X11L�C99 complex inhibited the inter-action of C99 with presenilin, which strongly sup-pressed the �-cleavage of C99. In AD patient brains, Alcand APP were particularly colocalized in dystrophicneurites in senile plaques. Deficiencies in the X11L-me-diated interaction between Alc and APP and/or CTF�enhanced the production of A�, which may be related tothe development or progression of AD.

The production, aggregation, and accumulation of amyloid�-protein (A�)1 in the brain are initial steps in the pathogene-

sis of Alzheimer’s disease (AD). A� is generated by the intra-cellular processing of amyloid �-protein precursor (APP). Majorproteolytic processing of APP generates a large extracellularamino-terminal domain (sAPP�) and a truncated carboxyl-ter-minal fragment (CTF�) by the digestion of �-secretase, whichcleaves APP at the �-site within the A� domain. Another formof proteolytic processing occurring at the �-site, by �-secretase(or BACE), gives rise to low levels of sAPP� and a carboxyl-terminal fragment (CTF�) including the entire A� domain.Both CTF� and CTF� are further cleaved by �-secretase (pre-senilin (PS) complex) and generate the p3 peptide and A�,respectively (1, 2).

A minority of AD cases fall into the familial category inwhich the overproduction of A� appears to be due to mutationsin genes encoding APP or PS (3). However, most AD cases areof the sporadic type (SAD) in which mutations in APP- orPS-encoding genes do not occur. The mechanisms that lead tothe overproduction of A� in SAD cases must therefore involvealternative mechanisms that cause the production, accumula-tion, and degradation of A� (4–6). These mechanisms are cur-rently being investigated intensively.

APP is a type I membrane protein (7). Immature APP (N-glycosylated form) is localized in the endoplasmic reticulumand cis-Golgi, whereas mature APP (N- and O-glycosylatedform) is localized to compartments following trans-Golgi and onthe plasma membrane. The cytoplasmic domain of APP (APP-cyt), through its interactions with cytoplasmic proteins and/orits conformational changes due to phosphorylation (8–13),plays an important role in the regulation of APP metabolism.X11-like protein (X11L) was originally isolated as a factor thatinteracts with APPcyt, and the interaction results in the sup-pression of A� production (14). X11L is a member of the X11protein family. The X11 was initially identified on the basis ofits possible link to Friedreich ataxia (15) and was originallydenoted as Mint (16, 17). X11 and X11L are thought to be

* This work was supported by Grants-in-aid for Scientific Researchon Priority Areas-Advanced Brain Science Project TS14017001 andTS15016002 and Grant-in-aid for Scientific Research TS 15659012 fromthe Ministry of Education, Science, Culture, Sports, and Science andTechnology, Japan, by The Mitsubishi Foundation, and by the TAKEDAScience Foundation. The costs of publication of this article were de-frayed in part by the payment of page charges. This article musttherefore be hereby marked “advertisement” in accordance with 18U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submittedto the GenBankTM/EBI Data Bank with accession number(s) AF438482.

□S The on-line version of this article (available at http://www.jbc.org)contains Supplemental Figs. 1–5.

¶ Both authors contributed equally to this work.� Recipient of a research fellowship from the Japan Society for the

Promotion of Science.‡‡ To whom correspondence should be addressed. Tel.: 81-11-706-

3250; Fax: 81-11-706-4991; E-mail: tsuzuki@pharm.hokudai.ac.jp.1 The abbreviations used are: A�, amyloid �-protein; Alc, Alcadein

(Alzheimer’s disease-related cadherin-like protein); APP, amyloid�-protein precursor; AD, Alzheimer’s disease; NTF, amino-terminalfragment; CTF, carboxyl-terminal fragment; CTF�, carboxyl-terminal

fragment of APP cleaved at �-site; C99, CTF� fragment expressed asCTF� construct; CHAPS, 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid; GM-130, 130-kDa Golgi matrix protein; GST,glutathione S-transferase; HA, hemagglutinin; KHC, kinesin heavychain; PDZ, repeated sequences in the brain-specific protein PSD-95,the Drosophila septate junction protein disks-large, and the epithelialtight junction protein ZO-1; PS, presenilin; sAPP, large extracellularamino-terminal domain truncated at the �- and/or the �-site; SYT,synaptotagmin; X11L, X11-like protein; h, human; XB31, X11L-bindingprotein clone number 31; SAD, sporadic AD; APPcyt, cytoplasmic do-main of APP; PI, phosphotyrosine interaction; HEK, human kidneyembryonic; PBS, phosphate-buffered saline; ELISA, enzyme-linked im-munosorbent assay.

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 278, No. 49, Issue of December 5, pp. 49448–49458, 2003© 2003 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

This paper is available on line at http://www.jbc.org49448

at Johns Hopkins U

niversity on May 22, 2007

ww

w.jbc.org

Dow

nloaded from

http://www.jbc.org/cgi/content/full/M306024200/DC1Supplemental Material can be found at:

homologues of LIN-10 in Caenorhabditis elegans (18, 19) anddX11L in Drosophila melanogaster (20). They are neuron-spe-cific adaptor proteins composed of a large amino-terminal re-gion, a central phosphotyrosine interaction (PI) domain, andtwo carboxyl-terminal PDZ domains. The 681GYENPTY687 mo-tif of APPcyt (numbering is based on the APP695 isoform)interacts with the PI domain of X11L (14).

To understand how APP metabolism is regulated, includingA� generation, it is of interest to isolate other proteins thatinteract with X11L. This will help elucidate the molecularmechanism by which X11L regulates APP metabolism.

In the present study, we isolated novel human cDNAs en-coding Alcadeins (Alcs), which are type I transmembrane pro-teins that interact with the PI domain of X11L. These genes areconserved in a wide variety of species, including D. melano-gaster and C. elegans. We found that the cytoplasmic domain ofAlc bound to the PI domain of X11L and initiated the formationof a tripartite complex comprised of Alc, X11L, and APP. Thiscomplex stabilized intracellular APP metabolism and signifi-cantly suppressed A� production by slowing APP maturation.We found that Alc and X11L also formed a tripartite complexwith C99, a CTF�. This complex inhibited the interaction ofC99 with PS1, a component of �-secretase that generates A�from C99. We could recover endogenous X11L�APP�Alc tripar-tite complexes from normal mouse brains, and normal mouseneurons displayed remarkable colocalization of these proteins.In AD brains, APP was found to colocalize with Alc in thedystrophic neurites of senile plaques. These observations indi-cate that APP exists in protein complexes composed of X11Land Alc that regulate APP metabolism, including A� produc-tion in neurons, and that this regulatory mechanism may beperturbed in AD.

MATERIALS AND METHODS

cDNA Cloning of Human Alc�1 and Plasmid Construction of Alc�Family cDNAs—The yeast two-hybrid system used in this study hasbeen described previously (14). Briefly, the human brain MATCH-MAKER cDNA library (Clontech) was screened with bait composed ofcDNA encoding part of the human X11L (hX11L) protein (amino acids129–555). This yielded a cDNA clone (X11L-binding protein clone num-ber 31 (XB31�)) that contained a partial open reading frame of Alc�1.Using this clone, a clone encoding the full-length Alc�1 protein wasisolated from a human brain �gt11 cDNA library (Clontech). This cDNAwas designated as human Alc�1 (GenBankTM accession numberAF438482, registered as XB31�). The full-length hAlc�1 open readingframe was subcloned into the HindIII and XbaI sites of pcDNA3 (In-vitrogen). We identified two similar cDNA clones in the HUGE (HumanUnidentified Gene-Encoded Large Proteins data base analyzed by theKAZUSA cDNA Project) Protein Database (21). The HUGE cDNAclones were denoted as human Alc�2 (accession number KIAA0911) andAlc� (accession number KIAA0726). We also found human Alc� (acces-sion number NM022131) in the GenBankTM/EBI Data Bank. Alc�cDNA, a generous gift from Dr. Nagase (KAZUSA DNA Research In-stitute, Chiba, Japan), was recloned into pcDNA3 at HindIII and XbaIsites to produce pcDNA3-hAlc�. The human Alc�1 and Alc� cDNAswere also recloned via an XbaI site into pcDNA3 with a FLAG tag on theamino-terminal end of the insert so that pcDNA3-FLAG-hAlc�1 andpcDNA3-FLAG-hAlc� could be produced.

Protein Interaction Assays in Yeast—The MATCHMAKER two-hy-brid system (Clontech) with pGBT9 and pGAD424 was used as de-scribed previously (22). �-Galactosidase activity in the yeast two-hybridsystem was measured in a liquid assay using o-nitrophenyl galactopy-ranoside and was expressed in Miller units.

Antibodies—The anti-Alc� polyclonal rabbit antibody UT83 wasraised against a peptide composed of Cys plus the sequence betweenpositions 954 and 971 of human Alc�1. The anti-Alc� polyclonal rabbitantibody BS7 was raised against a peptide composed of Cys plus thesequence between positions 954 and 968 of human Alc�. The anti-hX11L polyclonal antibodies, UT29 and UT30, were obtained as de-scribed previously (14), and the anti-X11L monoclonal antibody mint2was purchased from BD Biosciences. Anti-FLAG (M2, Sigma) andanti-HA (12CA5, Roche Diagnostics) monoclonal antibodies were pur-

chased. The anti-APP cytoplasmic domain polyclonal antiserum, G369,has been described previously (23). The anti-APP extracellular domainmonoclonal (22C11, Roche Diagnostics, and LN27, Zymed LaboratoriesInc.) and polyclonal (Sigma) antibodies were purchased. The 4G8 mono-clonal anti-A� antibody raised against the A�-(17–24) peptide waspurchased from Signet Laboratories (Dedham, MA). All polyclonal an-tibodies except G369 were affinity-purified before use. Monoclonal an-tibodies specific for protein-disulfide isomerase (1D3, Stressgen Bio-technologies, Victoria, British Columbia, Canada), 130-kDa Golgimatrix protein (GM-130) (clone no. 35, BD Biosciences), synaptotagmin(SYT) (clone no. 41, BD Biosciences), mouse kinesin heavy chain (KHC)(H2, Chemicon International; Temecula, CA), PS1 carboxyl-terminalfragment (PS1-CTF; Chemicon International), and PS1 amino-terminalfragment (PS1-NTF, Chemicon International) were purchased.

Coimmunoprecipitation and Western Blot Assays—COS7 and humanembryonic kidney 293 (HEK293) cells (�1 � 107 cells) were transfectedas described previously (14) with the indicated amounts of variouscombinations of the pcDNA3-hX11L, pcDNA3-hAlc�1, pcDNA3-hAlc�,pcDNA3APP695, and pcDNA3-PS1 plasmids. In another experiment,pcDNA3.1-HA-hX11, in which human X11 cDNA with an attachedHA tag at the 5�-end was inserted into the NheI and EcoRV sitesof pcDNA3.1, was transfected instead of pcDNA3-hX11L.pcDNA3-FLAG-hX11L and pcDNA3-FLAG-hX11, in which humanX11L and X11 cDNA with an attached FLAG tag at the 5�-end, werealso used in coimmunoprecipitation assay. Cells were harvested andlysed for 1 h on ice in CHAPS buffer, which consists of phosphate-buffered saline (PBS; 140 mM NaCl and 10 mM sodium phosphate (pH7.4)) containing 10 mM CHAPS, 5 �g/ml chymostatin, 5 �g/ml leupep-tin, 5 �g/ml pepstatin A, 1 mM Na3VO4, and 1 mM NaF. After centrif-ugation at 12,000 � g for 10 min at 4 °C, affinity-purified antibodies(indicated IgG amounts) or serum was added to the CHAPS lysatesupernatant. The immunoprecipitates were then subjected to Westernblot analysis using specific antibodies.

Subcellular Fractionation of Mouse Brain Proteins—The subcellularfractionation of mouse brains was performed as described previously(24, 25). Briefly, five brains of 8-week-old male C57BL/6 mice werehomogenized on ice in 5 volumes of buffer (10 mM HEPES (pH 7.4), 0.32M sucrose, 5 �g/ml chymostatin, 5 �g/ml leupeptin, and 5 �g/ml pep-statin) by 10 strokes of a loose-fitting (0.12-�m clearance) Teflon ho-mogenizer. The postnuclei supernatants were further centrifuged at100,000 � g for 1 h. The resulting precipitate (membrane fraction) wasresuspended and separated by centrifugation for 115 min at 41,000 rpmin a Beckman SW41 rotor at 4 °C on a 0–28% (w/v) iodixanol densitygradient. The supernatants were then fractionated into 13 tubes (0.8ml/tube), and the fractions (20 �l) were analyzed by Western blotanalysis with the indicated antibodies. Fraction 8 was solubilized byadding an equal amount of 2� CHAPS buffer consisting of 2� PBS (280mM NaCl and 20 mM sodium phosphate (pH 7.4)) plus 20 mM CHAPS.The mixture was allowed to rotate for 30 min at 4 °C and then centri-fuged at 100,000 � g for 60 min. The resulting supernatant was used inthe coimmunoprecipitation analysis of endogenous proteins.

Pulse-Chase Study—Pulse-chase labeling of cells was performedwith L-[35S] in vitro Cell Labeling Mix (0.4 mCi/ml; AGQ 0080, Amer-sham Biosciences). HEK293 cells were transfected with the indi-cated combinations of the pcDNA3APP695, pcDNA3-hX11L,pcDNA3-FLAG-hAlc�1, and pcDNA3 vector plasmids (transfection wasperformed with 9 �g of plasmids in total). After a 48-h transfection, thecells were metabolically labeled for 15 min. This was followed by a0–8-h chase period, which was initiated by replacing the labelingmedium containing excess amounts of unlabeled methionine. The cellswere then lysed and subjected to immunoprecipitation with G369 (anti-APP cytoplasmic domain antibody) and protein G-Sepharose. The re-covered APP was separated by SDS-PAGE (6% (w/v) polyacrylamide)and analyzed by Fuji BAS 2000 imaging analyzer.

Quantification of �-Amyloid—C99 protein was used as an intracel-lular �-secretase substrate (26). The construct containing Kozak se-quence, first methionine, signal peptide sequence of APP, and CTF�coding region was cloned into pcDNA3.1(�) at NheI and ApaI sites anddesignated as pcDNA3.1APP-C99. The C99 protein is expressed as anartificial CTF� once its signal sequence is cleaved off by signal pepti-dase (26).

Neuro-2a cells (�2 � 106 cells) were transiently transfected with theindicated combinations of pcDNA3APP695 or pcDNA3.1APP-C99,pcDNA3-hX11L, pcDNA3-FLAG-hAlc�1, and pcDNA3 vector plasmids(transfection was performed with 9 �g of plasmids in total). Cells weresupplied with fresh growth medium 5 h after the start of transfection,and conditioned medium was collected 48 h after the medium replace-ment. A�40 and A�42 were quantified with sandwich ELISA by using

Suppression of A� by X11L and Alcadein Association with APP 49449

at Johns Hopkins U

niversity on May 22, 2007

ww

w.jbc.org

Dow

nloaded from

three types of A�-specific monoclonal antibodies (27). IntracellularA�40 and A�42 were also extracted (28) and quantified by the sandwichELISA. Briefly, the cells were lysed by sonication in 40 �l of PBScontaining 6 M guanidine chloride and centrifuged at 20,000 � g for 15min at 4 °C. The resulting supernatant was diluted up to 12-fold byadding PBS and used in the ELISA. The single factor analysis ofvariance test followed by Tukey multiple comparisons was used toanalyze differences among groups of data. Data were presented asmeans � S.E.

Immunological Staining of Mouse Brain Sections—Experimentalprocedures were conducted in compliance with the guidelines of theAnimal Studies Committee of Hokkaido University. Adult C57BL/6mice (6 weeks, male) were perfused at 4 °C for 20 min with 4% (w/v)paraformaldehyde in 0.2 M sodium phosphate buffer, pH 7.5. The brainswere excised and postfixed with the same fixative at 4 °C overnightfollowed by treatment with 30% (w/v) sucrose in PBS for 2–3 days at4 °C. Brains were embedded into OCT compound (Miles Scientific), andfrozen sagittal sections (20 �m) were prepared. The sections wereincubated with 0.1% (v/v) Triton X-100 in PBS for 5 min at roomtemperature followed by treatment with 0.3% (v/v) H2O2 in PBS for 5min at room temperature to quench endogenous peroxidase activity.The sections were blocked with 3% (w/v) bovine serum albumin in PBSfor 10 min at room temperature and incubated for 3 h at 4 °C withantibodies in 1% (w/v) bovine serum albumin in PBS. After washing, thesections were further incubated with goat anti-mouse IgG coupled withAlexa Fluor 488 or goat anti-rabbit IgG coupled with Alexa Fluor 568 in1% (w/v) bovine serum albumin for 1 h at room temperature. Sectionswere viewed using the confocal laser scanning microscope LSM510(Carl Zeiss).

Immunological Staining of AD Brains—We used paraffin sections offrontal and temporal cortices from five AD subjects. Tissues were fixedwith Kryofix (a mixture of ethanol, polyethylene glycol, and water;Merck) for 1–7 days and embedded in paraffin. Dewaxed serial tissuesections (cut into 4-�m sections) were immunostained with the ABCelite kit (Vector Laboratories, Burlingame, CA). Some sections werepretreated with a microwave antigen retrieval method for 10 min in 10mM citrate buffer, pH 6.0 (29). Sections were incubated with antibodiesspecific for Alc� (UT83, 0.8 �g/ml), APP (22C11, 0.5 �g/ml), and A�(4G8, 1:1000 dilution). The peroxidase activity was visualized withdiaminobenzidine-H2O2 solution. For control analyses, tissue sectionswere incubated with anti-Alc� antibody UT83 in the presence of theantigen peptide (40 nM) or non-immune rabbit IgG (0.8 �g/ml).

For double immunofluorescence analyses, dewaxed sections wereincubated with a mixture of UT83 and 22C11 or of UT83 and 4G8 at thesame dilutions as above followed by incubation with a mixture offluorescein isothiocyanate-tagged goat anti-rabbit IgG (Jackson Immu-noResearch Laboratories, West Grove, PA; 1:30 dilution) and Cy3-tagged goat anti-mouse IgG (Jackson ImmunoResearch Laboratories,1:50 dilution). Prior to immunolabeling, the autofluorescence of thelipofuscin granules was blocked with Sudan black B staining (30).

RESULTS

Isolation, Identification, and Characterization of Novel Cad-herin-related Membrane Protein Genes—Yeast two-hybridscreening of a human brain cDNA library with cDNA encodinga part of the hX11L protein resulted in the isolation of a novelcDNA denoted XB31� (GenBankTM accession numberAF438482). The cDNA encoded a type I transmembrane pro-tein composed of 971 amino acids. In this paper, this proteinhas been entitled human Alcadein (Alzheimer-related cad-herin-like protein) �1 or Alc�1 (Fig. 1A). A thorough search ofthe human cDNA data base and genome data bases revealedthree similar genes. These encode 981-, 968-, and 955-aminoacid proteins, and we denoted these proteins as Alc�2, Alc�,and Alc�, respectively (Fig. 1A). Alc�2 is identical to humanAlc�1 except for 10 additional amino acids between the Alc�1residues at positions 71 and 72 (Fig. 1A, pink box). This extrasequence in Alc�2 is derived from one exon, and both proteinsare spliced variants of the XB31� (Alc�) gene. The Alc� andAlc� proteins are both �50% homologous to Alc�. They areencoded by different genes and belong to the same gene familyto which Alc� belongs. Further data base examinations re-vealed that there are Alc-like genes in D. melanogaster (Gen-

BankTM accession number AAF59384) and C. elegans (Gen-BankTM accession number NP495189) as well.

The Alc family members contain two cadherin motifs, a pu-tative Ca2�-binding sequence in their amino-terminal halves,and a single cytoplasmic domain composed of �110 amino acids(Fig. 1A). This cytoplasmic domain includes the X11L-bindingsite (see Fig. 2), which is highly conserved among differentspecies, and an acidic region at the carboxyl-terminal end of theX11L-binding site (Fig. 1, A and B).

Expression and Distribution of Alc�—X11L is specificallyexpressed in neural tissue (14). We examined which humantissues express Alc� and Alc� by Northern blot analysis, andwe found strong expression of the �5-kb Alc� transcript in thebrain and weaker expression in the heart, while a probe specificfor Alc� revealed high levels of a 4.4-kb transcript in the brain(data not shown, please see Supplemental Fig. 1).

Expression of Alc protein in mouse brain tissue was exam-ined by Western blot analysis. The anti-Alc� antibody UT83recognizes two endogenous proteins in mouse brain lysates(Fig. 1C) as well as recombinant human Alc�1 but not recom-binant Alc� (data not shown, please see Supplemental Fig. 2).The antigen peptide to which UT83 was raised competed withthe endogenous proteins for binding to UT83 (Fig. 1C, �). Thus,the UT83 antibody specifically recognizes two endogenous Alc�

proteins in the mouse brain. When UT83 was used to analyzeAlc� expression in a variety of murine tissues, the highestlevels were found in the brain, and small amounts were foundin the lung (Fig. 1D, upper panel). Assessment of X11L expres-sion with the UT30 X11L-specific antibody confirmed its brain-specific distribution (Fig. 1D, upper panel). These observationssuggest that, like X11L, Alc� is abundant in neurons. Thiscolocalization in the brain supports the notion that X11L andAlc� interact physiologically. This idea was further supportedwhen we examined Alc� and X11L expression in various partsof the brain by dissecting adult mouse brains into variouscomponent regions and blotting them with UT83 and UT30.Relatively high levels of Alc� were detected in the cerebralcortex, striatum, hippocampus, and thalamus; moderate levelswere found in the cerebellum; and low levels were observed inthe olfactory bulb, midbrain, and pons (Fig. 1D, lower panel).Expression in the sciatic nerve fiber was not observed. X11Lwas distributed in a very similar pattern (Fig. 1D, lower panel).Thus, X11L and Alc� are both largely expressed in neuraltissues apart from peripheral nervous system.

Alc and X11L Interaction—We investigated the interactionsbetween Alc and X11L by coimmunoprecipitation assays fromCOS7 cells expressing X11L together with Alc�1 or Alc� fol-lowed by Western blot analysis. The immunoprecipitates re-sulting from incubating the lysates with anti-Alc� (UT83) oranti-Alc� (BS7) antibodies included X11L as detected by theX11L antibody (UT30) (data not shown, please see Supplemen-tal Fig. 3). Conversely the immunoprecipitates generated bythe anti-X11L antibody contained Alc�1 and Alc�. In contrast,non-immune antibody (control IgG) immunoprecipitates didnot contain these proteins. Thus, both the Alc� and Alc� pro-teins bind to X11L within cells.

We identified the Alc�- and Alc�-binding domain of X11L byassessing the in vitro binding of various X11L domain con-structs to the putative cytoplasmic domains of Alc� and Alc�.Cell lysates containing X11L protein constructs were incubatedwith beads coupled to the cytoplasmic domains of human Alc�

or Alc� fused to GST. GST alone was used as a negative control.The pull-downs were subjected to Western blot analysis withthe anti-X11L amino-terminal antibody UT29. Full-lengthX11L and its construct encoding the X11L amino-terminal do-main attached to the PI domain (N � PI) bound the GST-Alc�

Suppression of A� by X11L and Alcadein Association with APP49450

at Johns Hopkins U

niversity on May 22, 2007

ww

w.jbc.org

Dow

nloaded from

and GST-Alc� fusion proteins (data not shown, please see Sup-plemental Fig. 4). However, the amino-terminal domain alonedid not bind either protein. GST alone also did not bind to anyof the X11L constructs. Thus, as with APP (14), the cytoplasmicdomains of Alc� and Alc� bind the PI domain of X11L.

We also identified the region in the cytoplasmic domain ofAlc� (Alc�CTF) that is required for the interaction with X11Lby yeast two-hybrid analysis. Truncated Alc�1CTF constructslacking the amino-terminal regions (CTF1, �871–902; CTF2,�871–908) or carboxyl-terminal regions (CTF3, �903–971;CTF4, �909–971) were examined for interaction with the N �PI domain (amino acids 129–555) of X11L. �-Galactosidaseactivity was measured in a liquid assay and calculated asMiller units. The CTF1 and CTF4 constructs, which have theconserved 904NPMETY909 (numbering is based on the hAlc�1isoform) sequence in common, bound X11L, but CTF2 andCTF3, which both lack the NPMETY sequence, did not interactwith X11L (Fig. 2). The NPXXXY sequence is a modified ver-sion of the NPXY motif of the cytoplasmic domain of APP thatbinds X11L (14). When we introduced mutations into the firsttwo amino acids of the NPMETY sequence generating

AAMETY, denoted as NP-AA, Alc�CTF did not interact withX11L. However, modification of NPMETY to NPMETA (Y-A)had no effect on the ability of the protein to bind to X11L (Fig.2). Thus, it appears that the first two amino acids of theNPMETY motif are essential for the association of Alc� withX11L. This is supported by the mutational analyses of the siteused by APP to bind X11 as it was shown that the end tyrosineresidue of the NPTY motif of APP is not important for bindingto X11 (31). Moreover, the first two Asp and Pro amino acidsare conserved in all the Alc proteins, whereas the end Tyrresidue is not conserved in Alc� or the D. melanogaster andC. elegans Alc proteins (Fig. 1B).

Alc, X11s, and APP Form a Tripartite Complex—The cyto-plasmic domains of Alc and APP both bind to the PI domain ofX11L. We thus investigated whether Alc and APP compete forbinding to X11L or act cooperatively. APP695 (the human APPisoform composed of 695 amino acids) and X11L were expressedtogether in HEK293 cells in the presence or absence of hAlc�1or FLAG-tagged hAlc�, and APP was immunoprecipitated fromcell lysates with the anti-APP cytoplasmic domain antibodyG369. The immunoprecipitates were analyzed by Western blot

FIG. 1. Comparison of structuresand cytoplasmic amino acid se-quences of the Alc family proteinsand protein expression of Alc� andX11L. A, schematic representation ofstructural features of the human Alc pro-teins. The pink box in the hAlc�2 se-quence indicates the amino acids derivedfrom an exon of the hAlc� gene. Yellowboxes include the predicted cadherin mo-tif, while the black box within the humanAlc� sequence represents the leucine zip-per motif. The green box includes a highlyconserved region containing the X11L-binding sequence. The red box indicatesan acidic region. The plasma membrane isindicated as a purple ladder-back struc-ture. The underlines indicate the regionsrecognized by the specific antibodiesUT83 and BS7. The numbers indicateamino acid (a.a.) residues. B, amino acidsequences of the cytoplasmic domains ofAlc family proteins. The arrow reveals theNP motif that is the X11L-binding site(blue). Gaps produced by the alignmentare indicated by a hyphen in the se-quence. The numbers indicate amino acidposition. C, detection of Alc proteins inthe mouse brain by the UT83 Alc� anti-body. Mouse brain lysates (50 �g of pro-tein) were subjected to SDS-PAGE (6%(w/v) polyacrylamide), transferred to amembrane, and probed with UT83 (0.16�g/ml IgG) in the presence (�) or absence(�) of the antigen peptide used to raise it(40 nM). UT83 specifically detected twoAlc� proteins. The numbers refer to themolecular masses (kDa) of the proteinstandards. D, analysis of the distributionof Alc� and X11L in various murine tis-sues and brain regions by Western blotanalysis. Mouse lysates (50 �g of protein)of tissues (upper panel) and brain regions(lower panel) were subjected to SDS-PAGE (6% (w/v) polyacrylamide), trans-ferred to the membrane, and probed withthe anti-Alc� (UT83) and anti-X11L(UT30) antibodies. Br, brain; Ht, heart;Lu, lung; Li, liver; Kid, kidney; Mus, mus-cle; OB, olfactory bulb, CC, cerebral cor-tex; ST, striatum; Hip, hippocampus; Ce,cerebellum; Mid, mid brain; Th, thala-mus; Sci, sciatic nerve.

Suppression of A� by X11L and Alcadein Association with APP 49451

at Johns Hopkins U

niversity on May 22, 2007

ww

w.jbc.org

Dow

nloaded from

analysis with anti-APP (G369), anti-X11L (UT29), anti-Alc(UT83), and anti-FLAG (M2) antibodies. As reported previ-ously (14), X11L was coimmunoprecipitated with the anti-APPantibody in the absence of Alc. Surprisingly, in the presence ofAlc�1 or FLAG-Alc�, higher levels of X11L were recovered inthe APP immunoprecipitates (Fig. 3A), and the two Alc pro-teins were found to be bound to X11L�APP complexes (Fig. 3A).However, when APP and FLAG-Alc�1 were coexpressed in theabsence of X11L, the G369 antibody did not coimmunoprecipi-tate Alc. The identical result was obtained when 22C11 anti-APP extracellular domain antibody was used (data not shown,please see Supplemental Fig. 5). Moreover M2 antibody did notcoimmunoprecipitate APP (Fig. 3B). Notably, in the cells thatcoexpress X11L and APP, the levels of immature APP (lowerband of APP, N-glycosylated form) were increased (Fig. 3, Aand B). This effect is related to the fact that X11L stabilizesAPP metabolism (see Fig. 5). These findings clearly indicatethat APP and Alc form a complex by interacting with X11L andthat the binding of APP and X11L is greatly stabilized by Alc.

We examined whether X11 is also able to mediate the for-mation of a tripartite APP�X11�Alc complex. Thus, Alc�1, HA-tagged X11, and APP695 were expressed in COS7 cells. WhenAlc�1 was immunoprecipitated with the anti-Alc� antibody inthe presence of X11, APP was coprecipitated (Fig. 3C). More-over, when APP was immunoprecipitated with the anti-APPantibody 22C11, Alc� was recovered along with X11. The bind-ing of APP and X11 is greatly stabilized by Alc as is X11L.Thus, both X11 and X11L can contribute to formation of atripartite complex comprised of APP�X11�Alc or APP�X11L�Alc.

It is usually documented that the PI domain does not inter-act at the same time with two proteins possessing a similarinteraction motif (32). Nevertheless the stoichiometry of X11L

bound to APP seems to be increased by addition of Alc (Fig. 3A),but is the APP�X11s�Alc complex stoichiometry 1:1:1 or 1:2 (ormore):1? Thus, we performed coimmunoprecipitation of APPand Alc with the M2 antibody from cells expressing APP,Alc�1, and FLAG-X11L or FLAG-X11 (Fig. 3D). If the stoichi-ometry of X11L bound to APP is increased by Alc, the amountof APP recovered by coimmunoprecipitation with anti-FLAGantibody is expected to decrease in the presence of Alc�1.However, FLAG-X11L and FLAG-X11 coprecipitated moreAPP in the presence of Alc�1 than in the absence of Alc�1. Thisresult suggests that the stoichiometry of X11s bound to APPdoes not increase by the addition of Alc. The most simpleexplanation is that the population of X11s bound to APP isincreased in cells by the expression of Alc. Alc may increase thechance that APP will interact with X11L or decrease the dis-sociation constant of the APP�X11L complex. Furthermore ourpreliminary result derived from studies to determine APP- andAlc-binding sites by introducing amino acid substitution in thePI domain of X11L supports the interaction of APP and Alcwith different sites of the single PI domain of X11L.2 However,we cannot rule out the possibility that more than one X11Lmolecule is bound to APP because both X11s could form homo-and/or heterodimers when the respective proteins were overex-pressed in cells.3

We examined whether APP, Alc, and X11L form similartripartite complexes in the brain. When we separated the post-nuclear fraction from the brains of five mice into cytoplasmicand membrane fractions and subjected both fractions to immu-noprecipitation assays, we found that APP and Alc� (which areboth membrane proteins) were not recovered in the cytoplasmicfraction (data not shown). However, the majority of X11L wasrecovered in the cytoplasmic fraction, although a moderateamount of X11L was also recovered in the membrane fraction(data not shown). To examine formation of these proteins intotripartite complexes in the brain, the membrane fraction wasfurther fractionated by iodixanol density gradient centrifuga-tion, and these fractions were subjected to Western blot anal-ysis with antibodies specific for APP, X11L, Alc�, the endo-plasmic reticulum protein protein-disulfide isomerase, theGolgi-resident protein GM-130, the synaptic vesicle proteinSYT, the vesicle transport motor protein KHC, and PS1. Thevesicles bearing APP, APPCTF, X11L, and Alc� all sedimentedin the same medium density membranous protein fractions7–9, which were a bit heavier than the SYT-containing vesiclesin fractions 9 and 10 (Fig. 4A). APP, KHC, and PS1 wereobserved in the same fractions, confirming earlier observations(33, 34). These data suggest that APP, X11L, and Alc� couldform tripartite complexes on organelles composed of mediumdensity membranes. To test this, the proteins in fraction 8 weresolubilized with CHAPS buffer, immunoprecipitated with anti-APP (G369) or non-immune control antibodies, and subjectedto Western blot analysis with antibodies specific for APP,X11L, Alc�, SYT, and rabbit IgG antibodies. Alc� and X11Lwere coimmunoprecipitated by anti-APP but not by non-im-mune antibody (Fig. 4B). Also SYT was not coimmunoprecipi-tated by anti-APP antibody. Thus, APP and/or APPCTF, Alc�,and X11L form a tripartite complex in vivo, probably on me-dium density membrane organelles that contain cargo proteinsincluding PS and that differ from the synaptic vesicles thatcontain SYT.

We next investigated whether APP, X11L, and Alc colocalizein neurons. Sagittal sections of the hippocampus from adultmouse brains were double stained with antibodies specific for

2 Y. Araki and T. Suzuki, unpublished observation.3 A. Sumioka and T. Suzuki, unpublished observation.

FIG. 2. Determination of the X11L-binding site of Alc�. Variousprotein constructs based on the cytoplasmic domain of Alc� (shownschematically in the upper panel) were expressed in yeast and exam-ined by the yeast two-hybrid system for their ability to bind X11L�1–125. The first 125 amino acids in the amino-terminal domain of X11Lwere deleted as they generate a nonspecific and constitutive positivesignal in this assay system (14). The following Alc� deletion and pointmutant constructs were tested: wild type (WT), the cytoplasmic domainof Alc� (amino acids 871–971 of Alc�1); CTF1, amino acids 903–971;CTF2, amino acids 909–971; CTF3, amino acids 871–902; CTF4, aminoacids 871–908; NP-AA, Ala-Ala substitution for Asn-Pro in NPMETYsequence; Y-A, Ala substitution for Tyr in NPMETY. The ability to growin selective medium was examined, and �-galactosidase activity wasquantified by a liquid assay and calculated in Miller units � S.D. (lowerpanel, n � 3). The plasmid alone was used as a control. GAL4ADrepresents the yeast Gal4 activator domain.

Suppression of A� by X11L and Alcadein Association with APP49452

at Johns Hopkins U

niversity on May 22, 2007

ww

w.jbc.org

Dow

nloaded from

APP and X11L, Alc� and X11L, or APP and Alc�. The threeproteins largely colocalized (Fig. 4C). Strong X11L immunore-activity was detected in the CA3 pyramidal cell somata andproximal dendrites where APP and Alc� were also localized(Fig. 4C). Colocalization was observed in the large pyramidalneurons of the cerebral cortex (data not shown). High powerviews of the cell bodies of these neurons confirmed the colocal-ization of APP, Alc�, and X11L (Fig. 4C). Thus, neurons ex-pressing X11L also express Alc� and APP, and the three pro-teins colocalize in these cells.

Alc Enhances the X11L-mediated Stabilization of APP Me-tabolism and Suppression of A� and sAPP Secretion—X11 sta-bilizes intracellular APP metabolism (35, 36), and we previ-ously reported that X11L suppresses the secretion of A�40 (14),indicating that X11L also stabilizes APP metabolism. Since Alcand APP form a complex with X11L through cytoplasmic inter-actions, and this complex formation strengthens the interac-tion between APP and X11L (Fig. 3), we speculated that Alcmay enhance the X11L-mediated stabilization of APP metabo-lism and thereby further suppress A� and sAPP production. Totest this, we assessed the intracellular APP metabolism inHEK293 cells that express APP with or without X11L in thepresence or absence of Alc�1 by pulse-chase assay (Fig. 5). Themetabolically radiolabeled APP was recovered by immunopre-cipitation at the indicated chase periods and separated bySDS-PAGE. The levels of immature APP were quantified byautoradiography and calculated with respect to the 0 h levels(1.0) (Fig. 5B). When APP was expressed alone, the immatureAPP levels decreased gradually with time due to the matura-tion of APP and the secretion of large extracellular domaintruncated at �- or �-site (sAPP) during the chase period. Asexpected, X11L coexpression slightly delayed this decrease inimmature APP levels, indicating that X11L stabilizes APPmetabolism. When Alc�1 was also coexpressed, the effect ofX11L was greatly enhanced. However, the increased stabiliza-tion of APP metabolism due to Alc�1 was not observed if X11Lwas not coexpressed.

We next investigated the production of A� and sAPP inNeuro-2a cells expressing APP with or without X11L in thepresence or absence of Alc�1. The amount of A� in mediumsecreted from the cells was quantified using sandwich ELISA.As we previously demonstrated (14), coexpression of X11L sup-presses the secretion of A�40 (Fig. 5C). X11L did tend to sup-press A�42 secretion, but it was not significant statistically.When X11L and Alc�1 were coexpressed, the effect of X11L wasremarkably enhanced. However, enhanced suppression ofA�40 secretion was not observed if only Alc�1 was expressed.Alc�1 and X11L coexpression did not significantly effect A�42secretion. Thus, Alc clearly enhances the X11L-mediated inhi-

with pcDNA3APP695 (2 �g), pcDNA3-HA-hX11 (0.25 �g), andpcDNA3-hAlc�1 (6.75 �g) in various combinations. To standardize theplasmid concentration, adequate amounts of pcDNA3 vector (�) wereadded (to yield 9 �g of plasmid in total). The cell lysates were immu-noprecipitated with anti-Alc� antibody (UT-83, upper panel) or anti-APP antibody (22C11, middle panel). The crude lysate (Lysate, lowerpanel) and immunoprecipitates (IP) were then analyzed by Westernblotting with anti-Alc� (UT83), anti-HA (12CA5), and anti-APP(G369) antibodies. D, Alc stabilizes APP binding to X11L and X11.COS7 cells (�1 � 107 cells) were cotransfected with pcDNA3APP695(2 �g), pcDNA3-Alc�1 (6.75 �g), and pcDNA3-FLAG-hX11L orpcDNA3-FLAG-hX11 (0.25 �g) in various combinations. To standardizethe plasmid concentration, adequate amounts of pcDNA3 vector (�)were added (to yield 9 �g of plasmid in total). The cell lysates wereimmunoprecipitated with anti-FLAG antibody (M2). The immunopre-cipitates were analyzed by Western blotting with anti-FLAG (M2, firstpanel), anti-APP (G369, second and third panels), and anti-Alc� (UT83,fourth panel) antibodies. The third panel shows a longer exposure (exp)of film than that of the second panel.

FIG. 3. Alc enhances APP-X11L binding, and the three proteinsform a tripartite complex. A, role of Alc in the binding of X11L toAPP. HEK293 cells (�1 � 107 cells) were transiently cotransfected withpcDNA3APP695 (2 �g) and pcDNA3-hX11L (0.25 �g) with or withoutpcDNA3-hAlc�1 (6.75 �g) or pcDNA3FLAG-hAlc� (6.75 �g) as indi-cated. To standardize the plasmid concentrations, the pcDNA3 vector(�) was added (to yield 9 �g of plasmid in total). The cells were lysed,and APP in the cell lysate was recovered by immunoprecipitation withthe anti-APP cytoplasmic domain antibody G369. The immunoprecipi-tates were then analyzed by Western blotting with anti-APP (G369),anti-X11L (UT29), anti-Alc� (UT83), and anti-FLAG (M2) antibodies.B, role of X11L in the formation of the APP�X11L�Alc tripartite complex.HEK293 cells (�1 � 107 cells) were transiently cotransfected withpcDNA3APP695 (3 �g) and pcDNA3-FLAG-hAlc�1 (3 �g) in the pres-ence or absence of pcDNA3-hX11L (3 �g). To standardize the plasmidamounts, 3 �g of pcDNA3 vector was added in the absence ofpcDNA3-hX11L (�). The cell lysates were immunoprecipitated with theanti-APP antibody G369 (upper three panels) or with the anti-FLAGantibody M2 (lower three panels). The immunoprecipitates were ana-lyzed by Western blot analysis with anti-APP (G369), anti-X11L(UT29), and anti-FLAG (M2) antibodies. C, X11 can complex withAPP and Alc�1. COS7 cells (�1 � 107 cells) were cotransfected

Suppression of A� by X11L and Alcadein Association with APP 49453

at Johns Hopkins U

niversity on May 22, 2007

ww

w.jbc.org

Dow

nloaded from

bition of A�, at least A�40, secretion from APP, which corre-lates with the finding that Alc stabilizes the interaction be-tween X11L and APP (Fig. 3A). The amount of sAPP in themedium secreted from the cells was detected by Western blotanalysis with anti-APP extracellular domain antibody 22C11(Fig. 5D). X11L suppressed sAPP secretion, and Alc�1 en-

hanced this effect. It is likely that the decreased A�40 andsAPP secretion by cells expressing APP, X11L, and Alc is due tothe slow-down of intracellular APP maturation caused by theirtripartite complex formation. Because we could not quantifythe A� levels in the cell lysates, we performed another studywith cells expressing C99 protein (see Fig. 6).

FIG. 4. APP, X11L, and Alc form a tripartite complex in the neurons of mouse brains. A, colocalization of APP, X11L, and Alc in the samebrain membrane fractions. The membrane fraction of five mouse brains was subjected to further fractionation by iodixanol density gradientcentrifugation. The density is indicated (upper panel). The fractions were analyzed by Western blotting with antibodies to APP and APPCTF (G369),X11L (mint2), Alc� (UT83), protein-disulfide isomerase (PDI)(ID3), GM-130 (clone no. 35), SYT (clone no. 41), KHC (H2), carboxyl-terminal halfof PS1 (PS1-CTF) (lower panel). B, presence of APP�X11L�Alc tripartite complexes in the brain. Fraction 8 in A was lysed, and APP was recoveredby immunoprecipitation with anti-APP (G369) antibody. As a control, non-immune antibody was used instead of G369. Crude lysate (Lysate) andthe immunoprecipitates (IP) were subjected to Western blot analysis with antibodies specific for APP and APPCTF (G369), X11L (mint2), Alc�(UT83), SYT (clone no. 41), and rabbit IgG antibodies. (H) indicates heavy chain of IgG. C, APP, Alc�, and X11L colocalize in mouse neurons.Sagittal sections of an adult mouse brain were double stained with antibodies against APP (22C11, red) and X11L (mint2, green) (panels 1–3), Alc�(UT83, red) and X11L (mint2, green) (panels 4–6), or APP (LN27, red) and Alc� (UT83, green) (panels 7–9). The merging of the signals (yellow)indicates the colocalization of the two proteins being assessed (panels 3, 6, and 9). CA3 region of the hippocampus is shown as are high power viewsof the cell bodies of the large pyramidal neurons of the cerebral cortex. The scale bar indicates 10 �m.

Suppression of A� by X11L and Alcadein Association with APP49454

at Johns Hopkins U

niversity on May 22, 2007

ww

w.jbc.org

Dow

nloaded from

X11L and Alc Form a Tripartite Complex with C99 andSuppress the �-Cleavage of C99—As we have shown in Fig. 4,APP not only colocalized with X11L and Alc� in the brain, the

tripartite complex also colocalized with PS1. This suggests thatX11L and Alc� could suppress the �-cleavage of the CTF ofAPP. We investigated this possibility by cotransfectingNeuro-2a cells with C99 instead of full-length APP with orwithout X11L and in the presence or absence of Alc�1. Thestability of C99 was examined by Western blot analysis, whichrevealed that X11L coexpression caused the intracellular C99to accumulate (Fig. 6A). This suggests that X11L protects theCTF from �-cleavage. This effect was enhanced when Alc�1was also coexpressed (Fig. 6A). However, the stabilization ofintracellular C99 was not observed if Alc� was coexpressed inthe absence of X11L.

When the A� in the medium from the transfected Neuro-2acells described above was quantified, it was found that X11Lcoexpression with C99 suppresses the secretion of both A�40and A�42 into medium (Fig. 6B). Moreover, when Alc�1 wasalso coexpressed, this effect was remarkably enhanced (Fig.6B). We could not quantify the level of intracellular A� derivedfrom full-length APP. Thus, we quantified the A� in cells ex-pressing C99. When the intracellular levels of A� in the trans-fected Neuro-2a cells were examined, the same effects of X11Land Alc�1 expression were observed for intracellular A�40generation (Fig. 6C). However, expression of X11L and Alc�1did not have a significant effect on the intracellular A�42levels, although the expression of both X11L and Alc�1 did tendto suppress the amount of intracellular A�42 (Fig. 6C). Thus,we concluded that X11L could suppress the �-cleavage of CTF�,and Alc enhances this effect.

PS is essential for �-secretase activity (37). We investigatedwhether X11L inhibits the interaction of CTF� with PS1 andwhether Alc enhances this inhibitory activity. HEK293 cellswere cotransfected with C99 and PS1 in the presence or ab-sence of X11L and with or without Alc�1. These cells weresubjected to immunoprecipitation assays with the anti-APPantibody G369, which recognizes C99, followed by Westernblotting to determine the interaction of C99 with PS1. A pre-vious report showed that the interaction between CTF� and PSis detectable in the presence of an aspartic protease inhibitorN-acetyl-leucyl-norleucinal (38), and thus we cultured thetransfected cells in the presence or absence of N-acetyl-leucyl-norleucinal. We observed that, in the presence of N-acetyl-leucyl-norleucinal, the anti-APP antibody coprecipitated PS1,confirming that C99 can interact with PS1 as expected (Fig.6D). In the presence of X11L, the anti-APP antibody recovereda smaller amount of PS1, indicating that X11L suppresses theinteraction between C99 and PS1. Surprisingly coexpression ofAlc�1 abolished the interaction between C99 and PS1. Thus, itappears that X11L blocks the association of CTF� with PS1and inhibits the �-cleavage of CTF�, thereby suppressing A�

generation. Furthermore Alc enhances this effect of X11L byforming a stable complex comprised of CTF�, X11L, and Alc.The mechanism by which CTF� processing is regulatedthrough its interactions with X11L and Alc is depicted sche-matically in Fig. 6E.

Localization of Alc in the AD Brain—To investigate thelocalization of Alc� in AD patient brains, we immunolabeledserial paraffin-embedded tissue sections with antibody spe-cific for Alc�. In ethanol (Kryofix)-fixed sections (Fig. 7, panel1) that had not been subjected to the antigen retrievalmethod, the anti-Alc� antibody consistently detected intenseAlc� immunolabeling of dystrophic neurites in the neuriticplaques together with faint neuronal labeling. The APP la-beling pattern was similar (Fig. 7, panel 2). Labeling was notobserved in sections stained with antigen-preabsorbed UT83(data not shown) or non-immune rabbit IgG used at the sameconcentration as UT83 (Fig. 7, panel 3). To confirm the colo-

FIG. 5. Alc� enhances the X11L-mediated stabilization of APPmetabolism and suppression of A� and sAPP secretion. A and B,effect of coexpressing Alc�1 and X11L on APP metabolism. HEK293cells (�1 � 107 cells) were transiently transfected with pcDNA3APP695(2 �g) in the presence or absence of pcDNA3-hX11L (0.25 �g) and withor without pcDNA3FLAG-hAlc�1 (6.75 �g). To standardize the plasmidamounts, pcDNA3 vector was added (to yield 9 �g of plasmid in total).After a 48-h transfection, the cells were metabolically pulse-labeledwith [35S]methionine for 15 min and chased for the indicated times.APP was immunoprecipitated from cell lysates with the anti-APP cy-toplasmic domain antibody G369, separated by SDS-PAGE (6% (w/v)polyacrylamide), detected by autoradiography (A), and quantified usinga Fuji BAS 2000 analyzer. The relative ratios of the levels of immatureAPP to the maximum level of immature APP at 0 h (the latter level wasassigned a reference value of 1.0) were calculated (B). mAPP, matureAPP695; imAPP, immature APP695. C, effect of coexpressing Alc�1and X11L on A� secretion. Neuro-2a cells (�1 � 107 cells) were tran-siently transfected with pcDNA3APP695 (3 �g) with or with-out pcDNA3-hX11L (0.3 �g) and in the presence or absence ofpcDNA3-FLAG-hAlc�1 (5.7 �g). To standardize the plasmid amounts,pcDNA3 vector was added (to yield 9 �g of plasmid in total). The culturemedium was collected and assessed for A�40 and A�42 levels using asandwich ELISA. The concentrations of A�40 and A�42 are presentedas means with S.E. (n � 6). The data were analyzed by one-way analysisof variance followed by the Tukey test (**, p 0.01; ***, p 0.001). D,effect of coexpressing Alc�1 and X11L on sAPP secretion. Neuro-2a cells(�1 � 107 cells) were transiently transfected with pcDNA3APP695 (3�g) with or without pcDNA3-hX11L (0.3 �g) and in the presence orabsence of pcDNA3-FLAG-hAlc�1 (5.7 �g). To standardize the plasmidamounts, pcDNA3 vector (�) was added (to yield 9 �g of plasmid intotal). The culture medium (2 ml) was collected, and sAPP was recov-ered by immunoprecipitation with anti-APP extracellular domain anti-body (22C11) as described previously (4). The immunoprecipitates (Me-dium) and cell lysates (Cell, �50 �g of protein) were analyzed byWestern blotting with 22C11. The levels of sAPP (a relative ratio) werenormalized to the amount of intracellular full-length mature and im-mature APP (APPFL) and indicated relative to the level of lane 1, whichwas assigned a reference value of 1.0. Results are the average ofduplicate assays, and error bars are indicated.

Suppression of A� by X11L and Alcadein Association with APP 49455

at Johns Hopkins U

niversity on May 22, 2007

ww

w.jbc.org

Dow

nloaded from

calization of Alc� and APP, paraffin-embedded tissue sec-tions from AD brains were double labeled for Alc� and APP(Fig. 7, panels 4–6). In a high power view of a neuritic plaque,most of the Alc� immunofluorescence (panel 4) colocalizedwith that of APP (panel 5). The merged view is shown inpanel 6. Furthermore double labeling of Alc� and A� revealedAlc�-positive neurites around the amyloid core of plaques(Fig. 7, panel 7). Unfortunately antibody specific for X11L(mint2) did not work on the paraffin-embedded tissue sec-tions (data not shown). These observations suggest that in

AD, Alc� and APP accumulate in dystrophic neurites aroundthe amyloid core of plaques.

DISCUSSION

The production, secretion, and aggregation of A� may causeneural cell death, resulting in the onset of AD. However, thecellular mechanisms involved in this neural cell death remainto be elucidated (2). The mechanisms involved in the develop-ment of familial AD involve mutations in APP and PS thatappear to increase A� production and cause the early onset of

FIG. 6. The Alc��X11L�C99 tripartite complex inhibits A� generation from C99 and blocks the association between C99 and PS1.A, effect of X11L and Alc on C99 cleavage. Neuro-2a cells (�1 � 107 cells) were transiently transfected with pcDNA3-APPC99 (3 �g) in the presenceor absence of pcDNA3-hX11L (0.3 �g) and with and without pcDNA3-hAlc�1 (5.7 �g). To standardize the plasmid concentrations, pcDNA3 vector(�) was added (to yield 9 �g of plasmid in total). The cells were lysed, and C99 was analyzed by Western blotting with anti-APP (G369) antibody.The lower panel indicates a shorter exposure (exp) of the film. B and C, effect of X11L and Alc on the generation of A� from C99. A�40 (left panel)and A�42 (right panel) in the medium (B) and the lysate (C) of the cells in A were quantified by a sandwich ELISA. The A�40 and A�42concentrations are presented as means with S.E. (n � 6). The data were analyzed by one-way analysis of variance followed by the Tukey test (*,p 0.05; **, p 0.01; ***, p 0.001). D, effect of X11L and Alc�1 on the interaction between C99 and PS1. HEK293 cells (�1 � 107 cells) weretransiently transfected with pcDNA3-APPC99 (2 �g) and pcDNA3-PS1 (1 �g) in the presence or absence of pcDNA3-hX11L (2 �g) and with orwithout pcDNA3-hAlc�1 (4 �g). To standardize the plasmid amounts, pcDNA3 vector (�) was added (to yield 9 �g of plasmid in total). The cellswere cultured for 24 h in the presence (�) or absence (�) of N-acetyl-leucyl-norleucinal (LLnL) (10 �M) and then lysed. C99 in the cell lysate wasrecovered by immunoprecipitation with the G369 anti-APP antibody. The immunoprecipitates were analyzed by Western blotting with antibodiesspecific for APP (G369), the full-length PS1 (PS1-NTF), X11L (mint2), and Alc� (UT83). E, schematic diagram showing how the tripartite complexcomposed of CTF�, X11s (X11L and X11), and Alc can block the �-cleavage of CTF� by PS. EC1 and EC2 in Alcadein indicate cadherin motifs 1and 2, respectively. Although the stoichiometry in the complex APP�X11L/X11�Alc is drawn as 1:1:1 for the sake of convenience, further analysisis needed to reveal the substance of the complex.

Suppression of A� by X11L and Alcadein Association with APP49456

at Johns Hopkins U

niversity on May 22, 2007

ww

w.jbc.org

Dow

nloaded from

dementia (1–3). However, the majority of AD cases are notassociated with these mutations, and these patients are there-fore classified as SAD patients. SAD also differs from familialAD by a late onset of dementia. The pathogenesis of SAD thusmust involve alternative mechanisms that increase A� produc-tion (39), prevent A� degradation (5), and/or accelerate A�aggregation (6, 40). To elucidate these pathogenic mechanisms,numerous investigators in this field, including our group, havefocused on understanding the role that APPcyt plays in regu-lating APP metabolism, including A� production (39). APPcytcontains at least three functional motifs, namely, 653YTSI656,667VTPEER672, and 681GYENPTY687 (human APP695 isoformnumbering). Mutations in these motifs alter the mechanismsthat regulate APP intracellular trafficking and/or metabolism(4, 8, 9). Several cytoplasmic proteins, through interacting withthe 681GYENPTY687 motif, are known to modify APP metabo-lism and consequently affect A� production (14, 35, 36, 41–46).One of these is the adaptor protein X11L. We found that whenX11L binds to the 681GYENPTY687 motif in APPcyt, the pro-duction of A� is suppressed (14). However, the molecular mech-anisms that regulate the effect of X11L on APP metabolism areas yet unclear, and thus we screened a human brain cDNAlibrary for proteins that bind to X11L. This led to the identifi-cation of the novel cadherin-related membrane protein family,Alc.

The first member of the Alc family we isolated was humanAlc�1. On the basis of its sequence, we also identified thecDNAs for human Alc�2, Alc�, and Alc� in the cDNA andgenome data bases. Homology searching also revealed Alc-likegenes in D. melanogaster and C. elegans. These proteins sharetwo cadherin motifs and a putative Ca2�-binding site in theamino-terminal ectodomain, a laminin G domain in the middleectodomain, and an X11L-binding site and an acidic region inthe cytoplasmic domain. It is likely that these proteins belongto the same family and play identical roles in neural function.However, the roles played by these putative Alc domains (apartfrom the X11L-binding site) in the metabolism and/or function

of APP require further investigation. Supporting the notionthat the Alc proteins participate in neural function(s) is that achicken protein that is homologous to Alc has been reportedrecently to be a postsynaptic membrane protein that may playa role in postsynaptic Ca2� signaling (47). It is possible that Alcmay transmit unidentified extracellular information throughan as yet unknown mechanism or that it may serve as areceptor, together with APP, of cargo proteins in membranetransport vesicles (39). Supporting the first possibility, we dem-onstrated in the present study that Alc couples with APPthrough cytoplasmic interactions bridged by X11L or X11. AP-Pcyt is thought to transmit some extracellular signals intonucleus by the mechanism of regulated intracellular proteoly-sis by coupling with adaptor proteins such as FE65 (48). Alcand X11s may moderate this signaling process by regulatingAPP processing by �-secretase complex including PS. Support-ing the possible cargo protein receptor function of Alc is ourdemonstration that APP, X11L, and Alc� were recovered to-gether with KHC in identical subcellular fractions. However,there is no direct evidence that APP and Alc operate as recep-tors of cargo proteins in membrane transport vesicles (39).

We found that the coupling of APP with Alc through X11Lsignificantly stabilized APP metabolism and enhanced the sup-pression of A� production. This effect was due to the suppres-sion of APP maturation, resulting in the suppression of the firstcleavage of APP at the �- and �-sites. Normally the majority ofAPP is subjected to non-amyloidogenic processing by �- and�-secretases that does not generate A� (2), although a smallproportion of APP is processed into A� by an intracellularamyloidogenic pathway in which the protein is cleaved by �-and �-secretases. We found that X11L could also associate withthe carboxyl-terminal fragments of APP that are metabolicproducts of APP cleavage at the �- or �-site. Thus, we examinedwhether X11L and Alc could also suppress the �-cleavage ofC99/CTF�. We found that the X11L-mediated association of Alcwith C99 enhanced the suppressive effects of X11L on A�

generation from C99. Moreover, since a recent report indicatesthat PS is an essential component of �-secretase (35), we ex-amined whether the interaction between C99/CTF� and PS1 isinhibited by formation of the C99�X11L�Alc complex. We foundthat X11L blocked PS from interacting with C99/CTF� andthat Alc enhanced this effect. The stable tripartite complexformed by C99/CTF�, X11L, and Alc may block the access of the�-secretase complex to CTF�. These observations suggest thatthe development of drugs that up-regulate X11L and Alc func-tion in AD patients and thereby down-regulate CTF� cleavagemay be useful in the treatment of AD.

Both APP and Alc recognized the PI domain of X11L, but thebinding of these two proteins to PI was cooperative, not com-petitive. The association of Alc with X11L may induce someconformational change of the PI domain of X11L and result inthe stable interaction of APP with X11L. The detailed analysisof the mechanism for interaction among three proteins, includ-ing the determination of the stoichiometry of proteins in thecomplex, is under consideration.

We found that the NPXXXY motif in the cytoplasmic domainof Alc was responsible for the interaction between Alc and thePI domain of X11L. The first and second residues (Asn and Pro)were essential for these interactions, unlike the end Tyr resi-due. Supporting this is that the Tyr residue is not conserved inall of the Alc family molecules. Moreover the NPXY motif inAPP, which is responsible for the interaction between APP andX11L, does not require conservation of the end Tyr residue (31).Thus, while the Tyr residue may be important, perhaps as apossible phosphorylation site, it is not required for interactionwith X11L.

FIG. 7. Localization of Alc�, APP, and A� in AD brain. Panels1–3, staining with single antibodies. Alc� (panel 1, UT83) localized inthe dystrophic neurites of disseminated senile plaques in a Kryofix-paraffin section of an AD brain. An adjacent section shows a similarpattern of APP labeling (panel 2, 22C11). This labeling was not ob-served in a control AD section stained with non-immune rabbit IgG(panel 3). Panels 4–6, double immunofluorescence staining for Alc�(panel 4, UT83, green) and APP (panel 5, 22C11, red) demonstrated thesimilar localization of the two proteins in a neuritic plaque of an ADbrain. A merged image is shown in panel 6 (yellow). Panel 7, doubleimmunofluorescence staining for Alc� (UT83, green) and A� (4G8, red)demonstrated accumulation of Alc�-positive dystrophic neurites aroundthe plaque core. Scale bar, 20 �m (white) and 100 �m (black).

Suppression of A� by X11L and Alcadein Association with APP 49457

at Johns Hopkins U

niversity on May 22, 2007

ww

w.jbc.org

Dow

nloaded from

APP immunolabeling is a sensitive method for detectingdisturbances in axonal transport (49, 50) and consistently iden-tifies dystrophic neurites in the senile plaques in both non-demented and AD brains (51, 52). In this study, we observedthat Alc� accumulated in the plaque neurites along with APP.However, X11L was reported to localize in the plaque cores(53), although the result requires confirmation. In contrast,X11L, Alc, and APP were found to colocalize largely in normalmurine neurons. It is likely that deficient interactions betweenAlc and APP, by implying axonal transport, generate greateramounts of A� in dystrophic neurons of AD patients.

Our studies and those of others reveal that various cytoplas-mic and membrane proteins such as X11L and Alc interact withAPPcyt to control A� production. This information may provehighly useful for the development of novel therapeutic drugsthat suppress A� production in SAD cases.

Acknowledgments—We thank Drs. T. Nakaya, H. Taru, H.Kawahara, Y. Saeki, and M. Watanabe (Hokkaido University) and T.Ozaki (Chiba Cancer Center) for helpful discussions; Professor S.Gandy (Thomas Jefferson University) for supplying the G369 anti-APPantibody; Dr. Nagase (KAZUSA DNA Research Institute) for gener-ously providing human Alc�cDNA (KIAA0726); and K. Arikawa fortechnical assistance. T. S. specifically thanks Dr. M. Fujino for encour-agement. We thank Dr. Olav R. Olsen (University of California, SanFrancisco) for critical reading of the manuscript.

Addendum—Alcadein is a molecule identical to calsyntenin, whichhas been published by Hintsch et al. (54).

REFERENCES

1. Price, D., Sissodia, S. S., and Borchelt, D. R. (1998) Science 282, 1079–10832. Selkoe, D. J. (2001) Physiol. Rev. 81, 741–7663. Hardy, J., and Gwinn-Hardy, K. (1998) Science 282, 1075–10794. Tomita, S., Kirino, Y., and Suzuki, T. (1998) J. Biol. Chem. 273, 19304–193105. Iwata, N., Tsubuki, S., Takaki, Y., Shirotani, K., Lu, B., Gerard, N. P., Gerard,

C., Hama, E., Lee, H. J., and Saido, T. C. (2001) Science 292, 1550–15526. Yanagisawa, K. (2002) J. Neurosci. Res. 70, 361–3667. Kang, J., Lemaire, H. G., Unterbeck, A. J., Salbaum, J. M., Masters, C. L.,

Grzeschik, K. H., Multhaup, G., Beyreuther, K., and Muller-Hill, B. (1987)Nature 325, 733–736

8. Koo, E. H., and Squazzo, S. L. (1994) J. Biol. Chem. 269, 17386–173899. Lai, A., Sisodia, S. S., and Trowbridge, I. S. (1995) J. Biol. Chem. 270,

3565–357310. Iijima, K., Ando, K., Takeda, S., Satoh, Y., Seki, T., Itohara, S., Greengard, P.,

Kirino, Y., Nairn, A. C., and Suzuki, T. (2000) J. Neurochem. 75, 1085–109111. Ando, K., Oishi, M., Takeda, S., Iijima, K., Isohara, T., Nairn, A. C., Kirino, Y.,

Greengard, P., and Suzuki, T. (1999) J. Neurosci. 19, 4421–442712. Ando, K., Iijima, K., Elliott, J. I., Kirino, Y., and Suzuki, T. (2001) J. Biol.

Chem. 276, 40353–4036113. Ramelot, T. A., and Nicholson, L. K. (2001) J. Mol. Biol. 307, 871–88414. Tomita, S., Ozaki, T., Taru, H., Oguchi, S., Takeda, S., Yagi, Y., Sakiyama, S.,

Kirino, Y., and Suzuki, T. (1999) J. Biol. Chem. 274, 2243–225415. Duclos, F., Boschert, U., Sirugo, G., Mandel, J. L., Hen, R., and Koenig, M.

(1993) Proc. Natl. Acad. Sci. U. S. A. 90, 109–11316. Butz, S., Okamoto, M., and Sudhof, T. C. (1998) Cell 94, 773–78217. Biederer, T., and Sudhof, T. C. (2000) J. Biol. Chem. 275, 39803–3980618. Rongo, C., Whitfield, C., Rodal, A., Kim, S. K., and Kaplan, J. M. (1998) Cell 94,

751–75919. Kaech, S. M., Whitfield, C. W., and Kim, S. K. (1998) Cell 94, 761–77120. Hase, M., Yagi, Y., Taru, H., Tomita, S., Sumioka, A., Hori, K., Miyamoto, K.,

Sasamura, T., Nakamura, M., Matsuno, K., and Suzuki, T. (2002) J. Neu-rochem. 81, 1223–1232

21. Kikuno, R., Nagase, T., Suyama, M., Waki, M., Hirosawa, M., and Ohara, O.(2000) Nucleic Acids Res. 28, 331–332

22. Tomita, S., Fujita, T., Kirino, Y., and Suzuki, T. (2000) J. Biol. Chem. 275,13056–13060

23. Oishi, M., Nairn, A. C., Czernik, A. J., Lim, G. S., Isohara, T., Gandy, S. E.,Greengard, P., and Suzuki, T. (1997) Mol. Med. 3, 111–123

24. Graham, J. M. (1997) in Subcellular Fractionation—A Practical Approach(Graham, J. M., and Rickwood, D., eds) pp. 205–242, Oxford UniversityPress, Oxford

25. Kohmura, N., Senzaki, K., Hamada, S., Kai, N., Yasuda, R., Watanabe, M.,Ishii, H., Yasuda, M., Mishina, M., and Yagi, T. (1998) Neuron 20,1137–1151

26. Lichtenthaler, S. F., Multhaup, G., Masters, C. L., and Beyreuther, K. (1999)FEBS Lett. 453, 288–292

27. Tomita, S., Kirino, Y., and Suzuki, T. (1998) J. Biol. Chem. 273, 6277–628428. Sawamura, N., Morishima-Kawashima, M., Waki, H., Kobayashi, K., Kuramo-

chi, T., Frosch, M. T., Ding, K., Ito, M., Kim, T.-W., Tanzi, R. E., Oyama, F.,Tabira, T., Ando, S., and Ihara, Y. (2000) J. Biol. Chem. 275, 27901–27908

29. Sherriff, F. E., Bridges, L. R., and Jackson, P. (1994) Neuroreport 5, 1085–108630. Schnell, S. A., Staines, W. A., and Wessendorf, M. W. (1999) Histochem.

Cytochem. 47, 719–73031. Russo, T., Faraonio, R., Minopoli, G., De Candia, P., De Renzis, S., and

Zambrano, N. (1998) FEBS Lett. 434, 1–732. Margolis, B., Borg, J.-P., Straight, S., and Meyer, D. (1999) Kidney Int. 56,

1230–123733. Kamal, A., Stokin, G. B., Yang, Z., Xia, C.-H., and Goldstein, L. S. (2000)

Neuron 28, 449–45934. Kamal, A., Almenar-Queralt, A., LeBlanc, J. F., Roberts, E. A., and Goldstein,

L. S. (2001) Nature 414, 643–64835. Borg, J.-P., Yang, Y., De Taddeo-Borg, M., Margolis, B., and Turner, R. S.

(1998) J. Biol. Chem. 273, 14761–1476636. Sastre, M., Turner, R. S., and Levy, E. (1998) J. Biol. Chem. 273, 22351–2235737. Takasugi, N., Tomita, T., Hayashi, I., Tsuruoka, M., Niimura, M., Takahashi,

Y., Thinakaran, G., and Iwatsubo, T. (2003) Nature 422, 438–44138. Verdile, G., Martins, R. N., Duthie, M., Holmes, E., St. George-Hyslop, P. H.,

and Fraser, P. E. (2000) J. Biol. Chem. 275, 20794–2079839. Sisodia, S. S. (2002) Science 295, 805–80740. Yanagisawa, K., Okada, A., Suzuki, N., and Ihara, Y. (1995) Nat. Med. 1,

1062–106641. Duilio, A., Faraonio, R., Minopoli, G., Zambrano, N., and Russo, T. (1998)

Biochem. J. 330, 513–51942. Fiore, F., Zambrano, N., Minopoli, G., Donini, V., Duilio, A., and Russo, T.

(1995) J. Biol. Chem. 270, 30853–3085643. Guenette, S. Y., Chen, J., Jondro, P. D., and Tanzi, R. E. (1996) Proc. Natl.

Acad. Sci. U. S. A. 93, 10832–1083744. Trommsdorff, M., Borg, J. P., Margolis, B., and Herz, J. (1998) J. Biol. Chem.

273, 33556–3356045. Taru, H., Iijima, K., Hase, M., Kirino, Y., Yagi, Y., and Suzuki, T. (2002)

J. Biol. Chem. 277, 20070–2007846. Taru, H., Kirino, Y., and Suzuki, T. (2002) J. Biol. Chem. 277, 27567–2757447. Vogt, L., Schrimpf, S. P., Meskenaite, V., Frischknecht, R., Kinter, J., Dino,

P. L., Ziegler, U., and Sonderegger, P. (2001) Mol. Cell. Neurosci. 17,151–166

48. Ebinu, J. O., and Yankner, B. A. (2002) Neuron 34, 499–50249. Kawarabayashi, T., Shoji, M., Harigaya, Y., Yamaguchi, H., and Hirai, S.

(1991) Brain Res. 563, 334–33850. Sherriff, F. E., Bridges, L. R., and Sivaloganathan, S. (1994) Acta Neuropathol.

(Berl.) 87, 55–6251. Shoji, M., Hirai, S., Yamaguchi, H., Harigaya, Y., and Kawarabayashi, T.

(1990) Brain Res. 512, 164–16852. Cras, P., Kawai, M., Lowery, D., Gonzalez-DeWhitt, P., Greenberg, B., and

Perry, G. (1991) Proc. Natl. Acad. Sci. U. S. A. 88, 7552–755653. McLoughlin, D. M., Irving, N. G., Brownlees, J., Brion, J. P., Leroy, K., and

Miller, C. C. (1999) Eur. J. Neurosci. 11, 1988–199454. Hintsch, G., Zurlinden, A., Meskenaite, V., Steuble, M., Fink-Widmer, K.,

Kinter, J., and Sonderegger, P. (2002) Mol. Cell. Neurosci. 3, 393–409

Suppression of A� by X11L and Alcadein Association with APP49458

at Johns Hopkins U

niversity on May 22, 2007

ww

w.jbc.org

Dow

nloaded from