HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 279, Issue 23, 24343-24354, June 4, 2004

This Article Abstract Full Text (PDF) Supplemental Data All Versions of this Article: 279/23/24343    most recent M401925200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Araki, Y. Articles by Suzuki, T. Search for Related Content PubMed PubMed Citation Articles by Araki, Y. Articles by Suzuki, T. Social Bookmarking                      What's this?

Coordinated Metabolism of Alcadein and Amyloid -Protein Precursor Regulates FE65-dependent Gene Transactivation^*

Yoichi Araki, Naomi Miyagi, Naoko Kato, Tomohiro Yoshida, Sachiyo Wada, Masaki Nishimura¶, Hiroto Komano||, Tohru Yamamoto**, Bart De Strooper, Kazuo Yamamoto, and Toshiharu Suzuki¶¶

From the Laboratory of Neuroscience, Graduate School of Pharmaceutical Sciences, Hokkaido University, Kita-ku Kita-12 Nishi-6, Sapporo 060-0812, Japan, the Graduate School of Pharmaceutical Sciences, University of Tokyo, Hongo 7-3-1, Bunkyo-ku, Tokyo 113-0033, Japan, the ^¶Neurology Unit, Molecular Neuroscience Research Center, Shiga University of Medical Science, Seta, Otsu, Shiga 520-2192, Japan, the ^||Department of Dementia Research, National Institute for Longevity Science, 36-3, Gengo, Morioka, Obu, Aichi 474-8522, Japan, the ^**Laboratory for Cell Asymmetry, Center for Developmental Biology, RIKEN, 2-2-3 Minatojima Minatomachi, Chuo-ku, Kobe 650-0047, Japan, the Neuronal Cell Biology Laboratory, K. U. Leuven and Flanders Interuniversitary Institute for Biotechnology, Herestraat 49, B-3000 Leuven, Belgium, and the Department of Integrated Biosciences, Graduate School of Frontier Sciences, University of Tokyo, Kashiwanoha 5-1-5, Kashiwa, Chiba 277-8562, Japan

Received for publication, February 22, 2004

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
The Alcadeins (Alcs)/calsyntenins and the amyloid -protein precursor (APP) associate with each other in the brain by binding via their cytoplasmic domains to X11L (the X11-like protein). We previously reported that the formation of this APP-X11L-Alc tripartite complex suppresses the metabolic cleavages of APP. We show here that the metabolism of the Alcs markedly resembles that of APP. The Alcs are subjected to a primary cleavage event that releases their extracellular domain. Alcs then undergo a secondary presenilin-dependent -cleavage that leads to the secretion of the amyloid -protein-like peptide and the liberation of an intracellular domain fragment (AlcICD). However, when Alc is in the tripartite complex, it escapes from these cleavages, as does APP. We also found that AlcICD suppressed the FE65-dependent gene transactivation activity of the APP intracellular domain fragment, probably because AlcICD competes with the APP intracellular domain fragment for binding to FE65. We propose that the Alcs and APP are coordinately metabolized in neurons and that their cleaved cytoplasmic fragments are reciprocally involved in the regulation of FE65-dependent gene transactivation. Any imbalance in the metabolism of Alcs and APP may influence the FE65-dependent gene transactivation, which together with increased secretion of amyloid -protein may contribute to neural disorders.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
The deposition and accumulation of amyloid -protein (A)¹ in the human brain are hallmarks of Alzheimer's disease (AD) (1). Amyloid -protein precursor (APP) is the precursor of A. It has a receptor-like transmembrane protein structure that consists of an extracellular domain, a transmembrane domain, and a short carboxyl-terminal cytoplasmic domain (2). The cytoplasmic domain of APP controls its metabolism and various physiological functions by interacting with cytoplasmic adaptor proteins (3–8). One of these adaptor proteins is X11L (the X11-like protein), which associates with the cytoplasmic domain of APP and stabilizes APP metabolism (5, 9). During our previous research that aimed to reveal the molecular mechanism by which X11L regulates APP metabolism, we found that the Alcadeins, which form cadherin-related membrane protein family, are X11- and X11L-binding proteins (9). These proteins are also known as calsyntenins, which were originally isolated as postsynaptic Ca²⁺-binding membrane proteins, but whose functions were not identified (10, 11). The Alcadeins (Alcs) consist of two Alc isoforms (Alc1 and Alc2) and Alc and Alc, all of which are type I transmembrane proteins and contain a conserved X11L-binding motif in their single cytoplasmic domains, similar to APP (9).

Alc does not directly interact with the cytoplasmic domain of APP. Rather, the association between the two molecules is bridged by the phosphotyrosine interaction domain of X11L. This results in the formation of a tripartite complex in the brain (9). The formation of this complex enhances the X11L-mediated stabilization of APP metabolism and suppresses the generation of A. Through its interaction with X11L, Alc also forms another tripartite complex with a carboxyl-terminal fragment of APP (CTF) that resulted from its cleavage at the - and/or -sites. The Alc-X11L-CTF complex promotes the X11L-dependent inhibition of the -secretase cleavage of the CTF because the complexing of CTF interferes with its interaction with presenilin (PS) (9). Moreover, it has been shown that APP and Alc co-localize in neurons (9) and that calsyntenin-1 is extracellularly cleaved (10). This suggests that APP and Alc may be metabolized in a coordinated fashion by a secretase(s).

In the present study, we found that, like APP, the Alcs are cleaved by the PS-dependent -secretase after a primary cleavage in their extracellular juxtamembrane region. These cleavages lead to the secretion of an A-like short peptide (-Alc) and liberate the intracellular domain of Alc (AlcICD). AlcICD down-regulates the gene transactivation ability of the APP intracellular domain (AICD). These observations suggest that APP and Alc are coordinately metabolized to regulate the gene transactivation ability of APP. This indicates that these two proteins play an important role in neural functions. Impairment of the coordinated metabolic regulation of APP and Alc and the consequent loss of the reciprocal regulation by APP and Alc in the gene transactivation may lead to neural disorders and disease progression in AD.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Plasmid Construction—Human Alcadein cDNAs were inserted into pcDNA3 vector to construct pcDNA3-hAlc1, pcDNA3-hAlc, and pcDNA3-hAlc as previously described (9). The FLAG sequence was inserted between Leu³⁸ and Glu³⁹ of hAlc1cDNA (pcDNA3-FLAG-hAlc1) and between Ile⁴¹ and Glu⁴² of hAlccDNA (pcDNA3-FLAG-hAlc), whereas the HA sequence was inserted between Ile⁴⁴ and Glu⁴⁵ of hAlccDNA (pcDNA3-HA-hAlc). The amino acids between Glu³⁹ and Ala⁸¹⁵ of pcDNA3-FLAG-hAlc1 was deleted to produce pcDNA3-AlcE. The pcDNA3-PS1 and pcDNA3-PS1D385A (12), pcDNA3APP695 (13), and pcDNA3-hX11L (5) plasmids have been described previously. cDNAs encoding human AICD (residues 652–695 of human APP695) and human AlcICD (residues 871–971 of human Alc1) were produced by PCR, and the products bearing the linker sequence encoding EFPGIPPG at their 5'-end were cloned into the pBIND vector (Promega) at the KpnI/BamHI sites. This generated pBIND-Gal4BD-AICD and pBIND-Gal4BD-AlcICD, which encode fusion protein composed of Gal4BD linked to AICD or AlcICD. The same procedure was used to generate the pBIND-Gal4BD-AlcICD/NPAA plasmid, which expresses a mutant hAlcICD molecule in which Asn⁹⁰³ and Pro⁹⁰⁴ were substituted by Ala. The AlcICD and AlcICD/NPAA cDNAs followed by the signal peptide sequence composed of 28 amino acid residues were also inserted into pcDNA3 to produce pcDNA3-hAlcICD and pcDNA3-hAlcICD/NPAA. FE65 cDNA (8) was inserted into pcDNA3.1 at the NheI/NotI sites to generate pcDNA3.1NFLAG-FE65. This construct produces FE65 bearing a FLAG tag at its amino terminus.

Antibodies—The anti-Alc polyclonal rabbit antibody UT83 has been described (9). The anti-Alc polyclonal rabbit antibody UT99 was raised against a peptide composed of Cys plus the sequence between positions 951 and 968 of human Alc (C+DSPSSDERRIIETPPHRY). The anti-Alc polyclonal rabbit antibody UT105 was raised against a peptide composed of Cys plus the sequence between positions 938 and 955 of human Alc (C+QNGARQAQLEWDDSTLPY). These Alc-specific antibodies all recognize the cytoplasmic domain of the Alc proteins to which they had been raised. As a result, they recognized the full-length Alc protein, the CTF1 derived from Alc, and AlcICD. These antibodies did not react with other Alc proteins apart from the one to which they had been raised. The anti-APP cytoplasmic G369 antibody has been described (14), and another anti-APP antibody was purchased from Sigma (catalog number A8717). The anti-X11L (mint2; BD Biosciences), anti-FLAG (M2; Sigma), anti-HA (12CA5; Roche Applied Science), anti-PS1 amino-terminal fragment (PS-NTF; Chemicon International), and anti-PS1 carboxyl-terminal fragment (PS-CTF; Chemicon International) antibodies were purchased.

In Vitro Generation of AlcICD—The cells were homogenized in 8 volumes of buffer A (10 mM triethanolamine-acetic acid buffer, pH 7.8, 0.25 M sucrose) containing 5 µg/ml chymostatin, 5 µg/ml leupeptin, and 5 µg/ml pepstatin and centrifuged at 1,000 x g for 10 min at 4 °C. The supernatant was then centrifuged at 100,000 x g for 15 min at 4 °C to obtain the membrane fraction (15). The membrane fraction was resuspended in buffer A containing 5 µg/ml chymostatin and 5 µg/ml leupeptin and incubated at 37 °C for the indicated periods in the presence or absence of 1 µM L-685,458, an aspartic protease inhibitor (Calbiochem). These samples were separated by electrophoresis through a Tris-Tricine gel (15% (w/v) polyacrylamide) and subjected to Western blotting.

Western Blot Assays—Human embryonic kidney (HEK) 293 cells (1 x 10⁷ cells) were transfected as described previously (9) with the indicated amounts of various combination of plasmids. The cells were harvested and lysed in Hepes-buffered saline with Triton X-100 (HBST; 10 mM Hepes, pH 7.4, 0.5% (v/v) Triton X-100, 150 mM NaCl, 5 µg/ml chymostatin, 5 µg/ml leupeptin, and 5 µg/ml pepstatin). The cell lysate or immunoprecipitates were separated on Tris-glycine or Tris-Tricine gels, transferred to nitrocellulose membrane, and probed with specific antibodies. PS1-deficient, PS1- and PS2-deficient, and wild type fibroblast cells (16) were also transfected by the same procedure.

Protein Sequence Analysis—Alc1 and its CTF1 fragment were separated on SDS-PAGE and blotted on polyvinylidene difluoride membrane. The proteins corresponding Alc1 and CTF1 on the membrane were cut off and then subjected to the gas phase protein sequencer Procise 492HT (Applied Biosystems).

Immunostaining of Cells—HEK293 cells expressing FLAG-FE65 and AlcICD were fixed in phosphate-buffered saline containing 4% (w/v) paraformaldehyde and 4% (w/v) sucrose, permeabilized in phosphate-buffered saline containing 0.2% (v/v) Triton X-100, blocked in phosphate-buffered saline containing 3% (w/v) bovine serum albumin, and incubated with the anti-FLAG monoclonal M2 antibody (1:500 dilution) and the anti-Alc polyclonal antibody UT83 (0.8 µg/ml IgG). After washing, the cells were further incubated with goat anti-rabbit IgG coupled with Alexa Fluor 488 or goat anti-mouse IgG coupled with Alexa Fluor 546 antibody. The cells were viewed using the confocal laser scanning microscope LSM510 (Carl Zeiss).

Gene Transactivation Assay—The indicated cells (3 x 10⁴ cells) plated in 96-multiwell plate were transiently transfected with various combinations of the following plasmids in LipofectAMINE 2000 (Invitrogen): pBIND-Gal4BD-AICD or pBIND-Gal4BD-AlcICD (20 ng), pG5luc (600 ng) (Promega), and pcDNA3.1NFLAG-FE65 or pcDNA3.1NFLAG vector (40 ng) in the presence or absence of pcDNA3-hAlcICD or pcDNA3-hAlcICD/NPAA. To standardize the plasmid amount, pcDNA3 or pBIND-Gal4 vector (-) was added (to yield 0.7 µg of plasmid in total). In a separate study, pBIND-APP695-Gal4BD was used instead of pBIND-Gal4BD-AICD. The cells were harvested 36 h after transfection, and the transcriptional activity of the reporter gene was analyzed by using the dual luciferase assay system (Promega). All of the combinations were tested in triplicate, and the luciferase activity was normalized according to the manufacturer's protocol to eliminate the effect of transfection efficiency differences.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Alcs Are Metabolized in a Similar Manner to APP and Are Novel Substrates for PS-dependent -Secretase—We previously reported that Alc and APP form a tripartite complex by cytoplasmic interactions with X11L, that the three proteins colocalize in neurons, and that tripartite complex formation stabilizes APP metabolism (9). Alcs are type I membrane proteins, as is APP, and a previous report has suggested that the Alc/calsyntenin proteins are cleaved extracellularly (10). These observations led us to speculate that the Alcs are metabolized in a similar manner to APP.

It is now well understood that APP is first cleaved at major - and minor -sites in its extracellular juxtamembrane region and that this releases the large extracellular amino-terminal domain (sAPP/) and generates the intracellular carboxyl-terminal fragments (CTF/) (17). We first examined whether the Alcs also undergo cleavage at a primary site that generates a large extracellular amino-terminal domain (sAlc) and a short cytoplasmic carboxyl-terminal fragment (CTF1). Thus, FLAG-Alc1, FLAG-Alc, and HA-Alc were overexpressed in HEK293 cells (Fig. 1, B–D, middle lanes), as was APP (Fig. 1A, middle lane). The cell lysates were analyzed by Western blotting with the indicated antibodies (Cell). The sAlcs and sAPP were also recovered from the medium by immunoprecipitation with tag-specific antibodies and analyzed by Western blotting with the same antibody (Medium). The 30-kDa CTF1 for Alc and Alc and the 21-kDa CTF1 for Alc were detected in the cell lysates along with the intact Alc proteins. At the same time, the sAlc fragments were detected in the medium. Thus, the Alcs are cleaved by a primary cleavage event, similar to APP.

View larger version (29K): [in this window] [in a new window]   FIG. 1. Cleavage of Alcs and APP by PS-dependent -secretase. A–D, generation of carboxyl-terminal fragments and secretion of amino-terminal domain from the Alcs and APP. HEK293 cells (1 x 106 cells) were transiently transfected with 1.5 µg of the plasmids pcDNA3-FLAG-APP695 (A), pcDNA3-FLAG-hAlc1 (B), pcDNA3-FLAG-hAlc (C), pcDNA3-HA-hAlc (D), or pcDNA3 vector (-). The cells were cultured for 24 h in the presence (+) or absence (-) of L-685,458 (1 µM) and then lysed. The cell lysates (20 µg of protein) were analyzed by Western blotting with the G369 (A), UT83 (B), UT99 (C), or UT105 (D) polyclonal antibody. The culture media (1 ml) were also collected, and sAPP or the sAlcs were recovered by immunoprecipitation with the anti-FLAG antibody M2 (A–C) or the anti-HA antibody 12CA5 (D). The immunoprecipitates were then analyzed by Western blotting with same antibody (A–D, bottom panels). E, dependence of the metabolism of CTF1 on PS. The cell lysates (20 µg of protein) of wild type (PS1 +/+, PS2 +/+), PS1-deficient (PS1 -/-, PS2 +/+) and PS1- and PS2-deficient (PS1 -/-, PS2 -/-) fibroblasts (1 x 106 cells) were analyzed by Western blotting with the UT83, PS1-CTF, or PS2-CTF antibody. Long exp indicates a longer exposure of the 30-kDa area of the film. F, effect of blocking endogenous PS activity on the metabolism of CTF1. HEK293 cells (1 x 106 cells) that stably express wild type PS1 (WT) (S. Oguchi and T. Suzuki, unpublished observations) or the dominant negative PS1D385A mutant (D385A) (12) were transfected with or without pcDNA3-FLAG-hAlc1 (1.5 µg). The cell lysates (20 µg of protein) were analyzed by Western blotting with UT83. The numbers refer to the molecular masses (kDa) of the protein standards.

View larger version (25K): [in this window] [in a new window]   FIG. 3. Generation of AlcICD by -secretase in vitro and characterization of Alc metabolites in the brain. A, generation of AlcICD. HEK293 cells (3 x 106 cells) were transfected with pcDNA3-hAlc1 (top panel, +), pcDNA3-hAlc (middle panel, +), pcDNA3-hAlc (bottom panel, +) or pcDNA3 (-) (3 µg), and their membrane fractions were incubated at 37 °C for 1 or 3 h in the presence (+) or absence (-) of L-685,458 (1 µM). The cell lysates (50 µg of protein) were analyzed by Western blotting with the anti-Alc UT83 (top panel), the anti-Alc UT99 (middle panel), or the anti-Alc UT105 (bottom panel) antibodies. B, presence of endogenous AlcCTF1 and ICD fragments in the mouse brain. Mouse brain tissue was homogenized in HBST, and the lysate (25 µg of protein) was analyzed by Western blotting with the anti-Alc antibody UT83. The lower panel indicates a longer exposure (Long exp) of the film. The asterisks may indicate a glycosylated form of CTF1 (see text). The numbers refer to the molecular masses (kDa) of the protein standard.

Secretion of A-like Peptide and Release of the Intracellular Domain Fragment by the -Cleavage of Alc CTF1—The above observations indicate that Alc is subjected to a primary cleavage event that generates CTF1 and that CTF1 is then cut by a secondary PS-dependent -cleavage. It is likely that the later event results in the secretion of a short peptide (-Alc) and the release of AlcICD into the cytoplasm. We first determined the primary cleavage site of Alc by subjecting the CTF1 of Alc1 to automated sequence analysis. The amino-terminal amino acid sequence was XXXPXFVX. Based on the protein sequence that was deduced from the Alc1 cDNA sequence, it appears that the amino-terminal sequence of the Alc1 CTF1 is ⁸¹⁶AAQPQFVH⁸²³. Thus, the most likely primary cleavage site of Alc1 is the peptide bond between Met⁸¹⁵ and Ala⁸¹⁶ (for the Alc2 isoform, which is identical to Alc1 apart from an additional 10 amino acids between the Alc1 residues at position 71 and 72 (9), the cleavage site that would generate the Alc2 CTF1 would be between Met⁸²⁵ and Ala⁸²⁶). We designated this cleavage site, which is in the extracellular domain, (Fig. 2A). We also determined the amino-terminal amino acid sequence of holo Alc1. The determined sequence was ARVNK, which is identical to the Ala²⁹-Arg-Val-Asn-Lys³³ sequence. This indicates that the signal peptidase cleaves the peptide bond between Ala²⁸ and Ala²⁹ of Alc, which means that the amino-terminal 28 residues constitute the signal peptide.

View larger version (21K): [in this window] [in a new window]   FIG. 2. Secretion of an A-like peptide by the -secretase cleavage of Alc CTF1. A, schematic structure of the Alc1 constructs used in this study. Alc1 is cleaved between Ala28 and Ala29 by the signal peptidase, between Met815 and Ala816 by the primary cleaving enzyme (indicated as ), and within the membrane by the PS/-secretase complex (indicated as ). A FLAG sequence was inserted between Leu38 and Glu39 to generate FLAG-Alc1. To mimic the primary cleaved product of Alc1, the sequence between Gln39 and Met815 was deleted. This generated AlcE. The overlines indicate the region recognized by the UT83 antibody (9). TM, transmembrane domain. B, expression of FLAG-tagged Alc1 and AlcE proteins in HEK293 cells. HEK293 cells (1 x 106 cells) were transiently transfected with pcDNA3-FLAG-hAlc1 (FLAG-Alc1), pcDNA3-FLAG-hAlcE (AlcE), or pcDNA3 (Vector) (1.5 µg plasmid). The cell lysates (20 µg of protein) were analyzed by Western blotting with the anti-FLAG antibody M2 (left panel) or the anti-Alc antibody UT83 (right panel). C, secretion of -Alc from AlcE. HEK293 cells (1 x 107 cells) were transfected with (+) or without (-) pcDNA3-FLAG-hAlcE (6 µg) and cultured for 24 h in the presence or absence (-) of the -secretase inhibitors, compound E (Cpd.E; 50 nM), DAPT (1 µM), or L-685,458 (1 µM). The medium (6 ml) and cell lysate (Cell, 200 µg of protein) were subjected to immunoprecipitation and then analyzed by Western blotting with the M2 antibody. The arrows labeled -Alc indicate the peptides that were generated by the -secretase cleavage of AlcE. D, -Alc is not generated in PS-deficient cells. Wild type (PS1 +/+, PS2 +/+), PS1-deficient (PS1 -/-, PS2 +/+), and PS1- and PS2-deficient (PS1 -/-, PS2 -/-) fibroblasts (2 x 107 cells) were transfected with pcDNA3-FLAG-AlcE(+,6 µg). The medium (6 ml) and cell lysates (Cell, 200 µg of protein) were subjected to immunoprecipitation and analyzed by Western blotting with the M2 antibody (left panel). The levels of -Alc (a relative ratio) were normalized by the amount of intracellular AlcE, and these values were then converted so that they were relative to the level in wild type fibroblasts, which was assigned a reference value of 1.0 (right panel). The results shown are the average of duplicate assays, and the error bars are indicated.

We then examined whether the A-like peptide (we termed it -Alc) is generated from AlcE by -cleavage. HEK293 cells expressing AlcE were cultured with or without various -secretase inhibitors, namely compound E (23), N-[N-(3,5-difluorophenacetyl-L-alanyl)]-S-phenylglycine t-butyl ester (DAPT) (24), or L-685,458. The medium and cell lysates were subjected to immunoprecipitation with the anti-FLAG antibody, after which the immunoprecipitates were separated on a Tris-Tricine gel (17.5% (w/v) polyacrylamide) and detected by Western blotting with the same antibody (Fig. 2C). In the absence of inhibitors, -Alc was detected in the medium as a pair of 5-kDa fragments (upper panel), but it was not recovered from the cell lysate (data not shown). However, in the presence of the inhibitors, the -Alc was not detected, which indicates that -Alc is generated from AlcE by -cleavage. Because two bands were observed, Alc may be cleaved at two different sites by -secretase. That the production of -Alc is dependent on -secretase was confirmed by using PS-deficient cells (Fig. 2D). AlcE was expressed in PS1-deficient, PS1- and PS2-deficient, or wild type fibroblasts and the secretion of -Alc into the medium was examined. The secretion was decreased in PS1-deficient cells because the ratio of the secreted -Alc levels to the intracellular AlcE levels was 30% of the ratio in wild type fibroblasts. Moreover, -Alc secretion was almost completely suppressed in the PS1- and PS2-deficient cells. These observations demonstrate that -Alc is generated by PS-dependent -secretase cleavage and then secreted into the extracellular milieu.

When the membrane fraction of cells expressing APP is incubated in vitro, the intracellular domain fragment of APP (AICD) is generated by /-cleavage. AICD can be visualized in this case because it is not subjected to rapid intracellular degradation (25, 26). We performed the same procedure to detect AlcICD. Thus, the membrane fractions of HEK293 cells expressing Alc1, Alc, or Alc were incubated for 1 or 3 h at 37 °C in the presence or absence of L-685,458. The Alc-, Alc-, and Alc-ICD molecules were then detected by Western blot analysis with the Alc-specific cytoplasmic domain antibodies (Fig. 3A). The 25-kDa AlcICD, 18-kDa AlcICD, and 25-kDa AlcICD fragments were generated from the membranes. The molecular masses on SDS-PAGE did not agree with the AlcICD sizes that were deduced from their amino acid sequences (please see supplementary Fig. 2), probably because these molecules all contained a large acidic region (9). L-685,458 inhibited the generation of these AlcICD molecule, which suggests that they were generated by the -secretase cleavage of the Alcs. Notably, Fig. 1C contradicts this result because it indicates that the CTF1 of Alc does not accumulate in Alc-expressing cells when they are treated with L-685,458. One possibility is that there may be two types of -secretase enzymes in the cell, only one of which is sensitive to L-685,458. Alc may be the substrate for the L-685,458-insensitive enzyme in vivo, but the membrane isolation and/or incubation may disturb the localization of the two enzymes. As a result, Alc may be preferentially cleaved by the L-685,458-sensitive enzyme in vitro.

It is important to demonstrate that the generation of CTF1 and AlcICD is physiological metabolism of Alc in the brain. Thus, mouse brain lysate was homogenized, and the endogenous Alc protein and its metabolic products were detected by Western blotting with the anti-Alc cytoplasmic antibody UT83 (Fig. 3B). We detected the endogenous CTF1 and AlcICD fragments along with the holo Alc protein. The asterisks in Fig. 3B may be an N-glycosylated form of CTF1 because these bands disappeared after treatment with N-glycosidase F (data not shown). These data demonstrate that the primary and secondary cleavages of Alc that generate CTF1 and AlcICD are physiological events that occur in vivo.

Synergic Stabilization of Alc and APP Metabolism Because of the Formation of a Tripartite Complex with X11L—We found previously that Alc enhances the X11L-dependent stabilization of APP metabolism (9). Because the above experiments demonstrated that the Alc proteins are metabolized in a similar manner as APP, we examined whether APP can enhance the X11L-dependent stabilization of Alc metabolism by forming a tripartite complex comprised of Alc, X11L, and APP. If it can, we would suggest that the mechanisms of Alc and APP are regulated coordinately by the formation and dissociation of the tripartite complex.

To examine this possibility, FLAG-Alc1 was expressed with or without X11L in the presence or absence of APP, and the secretion of sAlc1 was quantified along with the holo Alc1 levels in the cell lysates (Fig. 4). The expression of X11L without APP suppressed sAlc1 secretion to 60% of the levels secreted by cells that were only transfected with FLAG-Alc1 (compare lanes 2 and 3). However, co-expression of X11L and APP strongly suppressed sAlc1 secretion to 15% (compare lanes 2 and 5). Expression of APP on its own did not affect sAlc1 secretion (compare lanes 2 and 4). In cells expressing X11L, the levels of intracellular Alc1 increased (lanes 3 and 5), which suggests that X11L stabilizes the intracellular metabolism of Alc1. That X11L and APP were indeed expressed was confirmed (Fig. 4, left panels, lower two panels). As described previously (9), the intracellular metabolism of APP is stabilized by the co-expression of X11L and Alc1 (compare lanes 4 and 5). The endogenous X11L and APP695 levels in the cells were below the detection level (lane 1). These observations together show that the primary event that metabolizes Alc1 and produces sAlc1 is regulated by X11L in an identical manner as sAPP secretion (9). Thus, the primary metabolic events of APP and Alc may be coordinated.

View larger version (17K): [in this window] [in a new window]   FIG. 4. Effect of co-expressing X11L and APP on sAlc1 secretion. HEK293 cells (1 x 106 cells) were transiently transfected with pcDNA3-FLAG-hAlc1 (0.5 µg) (see Fig. 2A for the construct) with or without pcDNA3-hX11L (0.5 µg) and in the presence or absence of pcDNA3APP695 (0.5 µg). To standardize the plasmid amounts, pcDNA3 vector (-) was added. The culture medium (1 ml) was collected, and sAlc1 was recovered by immunoprecipitation with the anti-FLAG antibody M2. The immunoprecipitates (Medium) and cell lysates (Cell, 20 µg of protein) were analyzed by Western blotting with the M2 (sAlc1 and Alc1), mint2 (X11L), and G369 (APP) antibodies (left panel). The levels of sAlc1 (a relative ratio) were normalized by the amounts of intracellular holo Alc1, and these values were then converted so that they were relative to the level in lane 2, which was assigned a reference value of 1.0 (right panel). The results shown are the averages of duplicate assays, and the error bars are indicated.

We also previously demonstrated that Alc enhances the X11L-dependent inhibition of the -cleavage of APP CTF (9). Thus, we next examined whether APP can enhance the X11L-dependent inhibition of the -cleavage of Alc CTF1. AlcE and PS1 were expressed in HEK293 cells with or without X11L in the presence or absence of APP (Fig. 5A). The production of -Alc was decreased by X11L co-expression, and APP enhanced this effect (compare lanes 3 and 5 with lane 2). The secreted -Alc levels were quantified and indicated as a relative ratio. Expression of APP alone does not have any effect on -Alc secretion (lane 4).

View larger version (17K): [in this window] [in a new window]   FIG. 5. X11L suppresses -Alc generation and inhibits the interaction between PS1 and Alc CTF1, and this is enhanced by the formation of a tripartite complex between X11L, Alc CTF1, and APP. A, effect of co-expressing X11L and APP on the generation of -Alc. HEK293 cells (1 x 107 cells) were transiently transfected with pcDNA3-FLAG-hAlcE (1.5 µg) and pcDNA3-PS1 (1.5 µg) in the presence or absence of pcDNA3-hX11L (1.5 µg) with or without pcDNA3APP695 (1.5 µg). To standardize the plasmid amounts, pcDNA3 vector (-) was added to yield 6.0 µg of plasmid in total. The cells were cultured for 24 h, and the culture medium (6 ml) was immunoprecipitated with the anti-FLAG antibody. The immunoprecipitates were then analyzed by Western blotting with M2. The levels of -Alc (a relative ratio) were normalized by the amount of intracellular AlcE (data not shown), and these values were then converted so that they were relative to the levels shown in lane 2, which was assigned a reference value of 1.0. The results are the averages of duplicate assays, and the error bars are indicated. B, interaction of AlcE with PS1. HEK293 cells (1 x 106 cells) were transiently transfected with pcDNA3-FLAG-hAlcE (0.75 µg) and pcDNA3-PS1 (0.75 µg). To standardize the plasmid amounts, pcDNA3 vector (-) was added to yield 1.5 µg of plasmid in total. The cells were cultured for 24 h in the presence (+) or absence (-) of LLnL (10 µM) and then lysed. The cell lysates were immunoprecipitated with the anti-FLAG antibody M2, and the IP and cell lysates (20 µg of protein) were analyzed by Western blotting with the anti-FLAG M2 (AlcE), the anti-PS1 amino-terminal fragment (FL and NTF), and the anti-PS1CTF antibodies. FL, full length of PS1; NTF, amino-terminal fragment of PS1; CTF, carboxyl-terminal fragment of PSI. C, effect of X11L and APP on the interaction of AlcE with PS1. HEK293 cells (1 x 106 cells) were transiently transfected with pcDNA3-FLAG-hAlcE (0.375 µg) and pcDNA3-PS1(0.375 µg) in the presence or absence of pcDNA3-hX11L (0.375 µg) with or without pcDNA3APP695 (0.375 µg). To standardize the plasmid amounts, pcDNA3 vector (-) was added to yield 1.5 µg of plasmid in total. The cells were cultured for 24 h in the absence of LLnL and then lysed. The cell lysates were immunoprecipitated with the anti-FLAG antibody M2, and the immunoprecipitates were analyzed by Western blotting with antibodies specific for FLAG (AlcE), the amino-terminal half of PS1 (PS1), X11L (mint2), and APP (G369).

Stabilization and Localization of AlcICD by Interacting with FE65—It is well documented that AICD associates with FE65 and that this interaction stabilizes intracellular AICD metabolism (28). Like AICD, AlcICD possess a X11L-binding motif to which FE65 is expected to bind. To assess this, HEK293 cells were transfected with pcDNA3.1NFLAG-FE65 and pcDNA3-AlcICD or pcDNA3-AlcICD/NPAA. In the latter construct, the NPMETY sequence of AlcICD has been altered to AAMETY, which should prevent AlcICD from interacting with X11L (9). The overexpressed AlcICD protein was identical in size to the AlcICD that was generated endogenously from Alc (please see supplementary Fig. 2). A co-immunoprecipitation assay using the anti-FLAG antibody recovered AlcICD, but not AlcICD/NPAA, along with FLAG-FE65 (Fig. 6A, left panel). This result demonstrates that like AICD, AlcICD can associate with FE65. We also confirmed that FE65 associates with endogenously generated AlcICD by immunoprecipitating lysates of HEK293 cells expressing FLAG-FE65 and Alc1 or APP with the anti-FLAG antibody. The IP were analyzed by Western blotting with the anti-Alc1 cytoplasmic UT83 or the anti-APP cytoplasmic G369 antibody (Fig. 6A, right panel). As expected, FE65 associated with APP (middle lane). The anti-FLAG antibody also co-precipitated the full-length Alc1, Alc1 CTF1, and AlcICD along with FLAG-FE65 (right lane). This indicates that FE65 can interact with endogenously generated AlcICD within cells.

View larger version (34K): [in this window] [in a new window]   FIG. 6. FE65-dependent stabilization of AlcICD metabolism and alteration of AlcICD localization. A, interaction of Alc1, Alc CTF1, and AlcICD with FE65. Left panel, HEK293 cells (1 x 106 cells) were transfected with pcDNA3.1NFLAG-FE65 (+, 0.75 µg) in the presence of pcDNA3-hAlcICD (+, 0.75 µg) or pcDNA3-hAlcICD/NPAA (+, 0.75 µg). To standardize the plasmid amounts, pcDNA3 vector (-) was added to yield 1.5 µg of plasmid in total. The cell lysates were immunoprecipitated with the anti-FLAG antibody M2. The IP and cell lysates (Lysate, 20 µg of protein) were analyzed by Western blotting with the M2 and anti-Alc UT83 antibodies. Right panel, HEK293 cells (1 x 106 cells) were transfected with pcDNA3.1NFLAG-FE65 (+, 0.5 µg) in the presence of pcDNA3APP695 (+, 0.5 µg) or pcDNA3-hAlc1 (+, 0.5 µg). To standardize the plasmid amounts, pcDNA3 or pcDNA3.1 vector (-) was added to yield 1.5 µg of plasmid in total. The cell lysate were immunoprecipitated with the anti-FLAG antibody M2. The IP and cell lysates (Lysate, 20 µg of protein) were analyzed by Western blotting with the M2 (FE65), G369 (APP), or UT83 (Alc1, CTF1, and AlcICD) antibodies. Alc1, holo Alc1; CTF1, endogenously generated Alc CTF1 (see Fig. 2A); ICD, endogenously generated AlcICD. B, stabilization of AlcICD by FE65. Left panel, HEK293 cells (1 x 106 cells) were transfected with pcDNA3-hAlc1 (0.75 µg) with (+) or without (-) pcDNA3.1NFLAG-FE65 (0.75 µg) in the presence (+) or absence (-) of L-685,458 (1 µM). To standardize the plasmid amounts, pcDNA3 or pcDNA3.1 vector (-) was added to yield 1.5 µg of plasmid in total. Left panel, the cell lysates (20 µg of protein) were analyzed by Western blotting with UT83. The lower panel indicates a longer exposure (Long exp) of the film. Right panel, in another study, the cells were subjected to subfractionation. Membrane (P100, 10 µg of protein) and cytoplasmic (S100, 30 µg of protein) fractions were analyzed by Western blotting with UT83. C, effect of FE65 on the localization of AlcICD. HEK293 cells (7 x 104 cells) were transfected with pcDNA3-AlcICD (0.18 µg) with (AlcICD/FE65) or without (AlcICD) pcDNA3.1NFLAG-FE65 (0.18 µg). The cells were stained with the UT83 (AlcICD, green) and M2 (FLAG-FE65, red) antibodies along with nuclear staining with 4',6'-diamidino-2-phenylindole dihydrochloride (DAPI, blue). The signals have also been merged. Cell1 and Cell2 indicate different fields of microscopy. The scale bar indicates 10 µm.

It has been reported that AICD translocates into the nucleus followed by its association with FE65 (29). Thus, we explored the intracellular localization of AlcICD in the presence or absence of FE65 (Fig. 6C). AlcICD overexpressed on its own showed a predominantly cytoplasmic localization (upper panels). However, when AlcICD was co-expressed with FE65, Alc-ICD localized in the nucleus (lower panels), which suggests that AlcICD is translocated into the nucleus because of its association with FE65. Thus, AlcICD is very similar to AICD with regard to its metabolism and localization.

AICD- and FE65-dependent Gene Transactivation Is Regulated by AlcICD—AICD is able to transactivate gene(s) after it interacts with FE65 (30). Because AlcICD associates with FE65 and translocates into the nucleus, as does AICD, we assessed whether AlcICD can also transactivate genes. When the Gal4-AICD fusion protein is expressed in SHSY5Y cells with or without FLAG-FE65 and the reporter gene activity is quantified, as reported elsewhere (30), reporter gene transactivation activity is observed in the presence of FE65 (Fig. 7A, compare columns 3 and 4). When we substituted Gal4-AICD in this experiment with Gal4-AlcICD, however, FE65-dependent gene transactivation activity was not observed (compare columns 7 and 8). This indicates that AlcICD itself does not possess gene transactivation activity. Surprisingly, however, when AlcICD was co-expressed with Gal4-AICD and FE65, the gene transactivation mediated by AICD was almost completely suppressed (compare columns 4 and 5). When the same experiment was performed with AlcICD/NPAA (which lacks FE65 binding ability) instead of AlcICD, the AICD-mediated gene transactivation was not inhibited (compare columns 5 and 6).

View larger version (29K): [in this window] [in a new window]   FIG. 7. Down-regulation of AICD- and FE65-dependent gene transactivation by AlcICD. A, effect of AlcICD on the reporter gene transactivation mediated by AICD and FE65. SHSY5Y cells (1.5 x 106 cells) were transfected with pG5luc (Promega) and pBIND-Gal4-AICD (AICD, +) or pBIND-Gal4 (AICD, -) with or without pcDNA3.1NFLAG-FE65 in the presence or absence of pBIND-Gal4-hAlcICD or pBIND-Gal4-hAlcICD/NPAA. The transactivation values indicated are the ratios of the transactivation of the indicated plasmid combination to that of the pBIND vector alone (column 1). The error bars indicate the standard deviation. The data were analyzed by one-way analysis of variance followed by Tukey's test. ***, p < 0.001. B, suppression by AlcICD of the interaction between AICD and FE65. HEK293 cells (1 x 106 cells) were transfected with pcDNA3.1NFLAG-FE65 (0.5 µg) and pcDNA3-GFP-AICD (0.5 µg) in the presence or absence of pcDNA3-AlcICD or pcDNA3-AlcICD/NPAA (0.5 µg). To standardize the plasmid amounts, pcDNA3 or pcDNA3.1 vector (-) was added to yield 1.5 µg of plasmid in total. The cell lysates were immunoprecipitated with the anti-FLAG M2 (top set of panels) or the anti-APP G369 (middle set of panels) antibody. The IP and cell lysates (Lysate, bottom set of panels, 20 µg of protein) were then analyzed by Western blotting with the M2 (FE65), G369 (AICD), or UT83 (AlcICD) antibodies.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
We previously demonstrated that Alc forms a complex with APP in the brain through the cytoplasmic interaction with X11L, which acts to bridge Alc and APP (9). Once this tripartite complex is formed, APP becomes very stable with regard to its metabolism, whereas in contrast the naked APP molecule is readily cleaved by the - and/or -secretases in the cell. Moreover, the CTF also forms a tripartite complex with X11L and Alc. This blocks the access to CTF by the -secretase component PS1 and consequently decreases the generation of A and AICD from CTF by -secretase. Without this suppressive regulation, the naked CTF molecule generates large amount of A because of cleavage by -secretase (9). Alc is a type I transmembrane protein that bears the same receptor-like structure as APP. We and another group have found that the Alcs/calsyntenins are, like APP, cleaved extracellularly (this paper and Ref. 10). In the study reported here, we determined the primary cleavage site of Alc1 and located it in the extracellular juxtamembrane region. This suggests that CTF1, the cytoplasmic domain fragment of Alc1 generated by this primary cleavage, can subsequently be cleaved by -secretase, as is APP. Supporting this, we demonstrated here that Alc is a novel substrate of -secretase. Recent publications have reported that a number of membrane proteins can act as -secretase substrates (31–41). Thus, it is now clear that there are a variety of -secretase substrates and that -secretase is not specific for APP and Notch (42).

A novel indication of our observation that Alc is cleaved by -secretase is that the -secretase-mediated cleavages of APP and Alc should be coordinated. We found that when Alc forms tripartite complex with X11L and APP, its primary cleavage is suppressed, and the generation of sAlc and CTF1 is strongly inhibited. In contrast, the naked Alc molecule is quickly cleaved at the primary -site. This metabolic pattern is identical to the pattern of APP metabolism. Furthermore, when AlcE, which mimics the CTF1 product of the primary cleavage, was expressed in cells, we could detect A-like short peptides (-Alc) in the medium. These peptides are likely to be the amino-terminal products of the -cleavage of AlcE. However, when AlcE formed a tripartite complex with X11L and APP, the cleavage of AlcE by -secretase was suppressed because the interaction between AlcE and PS1 was suppressed. CTF cleavage by PS1 is regulated by an identical mechanism (9). These observations together support the hypothesis that APP and Alc are coordinately metabolized because when both proteins form a complex through their intracellular interaction with X11L, their primary and secondary cleavages are remarkably suppressed. However, when the tripartite complex is dissociated into free components, APP and Alc are readily subjected to be cleaved (Fig. 8). This coordinated regulation suggests that if the complex formation in the neuron is disturbed in some way, as it may occur in the AD brain, both proteins will be subjected to unregulated cleavage that leads the increased production of A and -Alc. It will be of interest to assess whether there is a difference in the formation of complexes composed of APP or APP CTFs, X11L or X11, and Alcs or Alc CTF1 in AD and healthy brains.

View larger version (18K): [in this window] [in a new window]   FIG. 8. Schematic diagram showing the coordinated metabolism of APP and Alc and the reciprocal regulation of gene transactivation by AICD and AlcICD. APP metabolism and Alc metabolism are both suppressed by associating with each other through their interactions in the cytoplasm with X11L. Thus, the tripartite complex synergically stabilizes APP and Alc metabolism. When X11L dissociates from the complex, APP and Alc are simultaneously cleaved in a coordinated fashion at first their primary cleavage sites and then at their secondary sites. The naked APP protein thus generates A (or p3) and AICD followed by the production of sAPP and CTF (or CTF), whereas the naked Alc protein generates -Alc and AlcICD followed by the production of sAlc and CTF1. AICD associates with FE65 and shows gene transactivation activity. AlcICD also associates with FE65 and can inhibit the FE65-dependent gene transactivation activity of AICD by blocking the association between AICD and FE65 in the cytoplasm and/or the nucleus. Thus, FE65-dependent gene transactivation was reciprocally regulated by association of AICD or AlcICD with FE65.