(BACE2) Cleaves the Amyloid Precursor Protein at the - Science Direct

Molecular and Cellular Neuroscience 16, 609 – 619 (2000) doi:10.1006/mcne.2000.0884, available online at http://www.idealibrary.com on

MCN

ASP1 (BACE2) Cleaves the Amyloid Precursor Protein at the ␤-Secretase Site I. Hussain,* D. J. Powell, † D. R. Howlett,* G. A. Chapman,* L. Gilmour, † P. R. Murdock, ‡ D. G. Tew, § T. D. Meek, § C. Chapman, ‡ K. Schneider, ¶ S. J. Ratcliffe, ¶ D. Tattersall,* T. T. Testa, ‡ C. Southan, 㛳 D. M. Ryan,** D. L. Simmons,* F. S. Walsh,* C. Dingwall,* and G. Christie* ,1 *Department of Neuroscience Research, †Department of Molecular Screening Technologies, ‡ Department of Biotechnology and Genetics, ¶Department of Discovery Chemistry, and Department of Bioinformatics, SmithKline Beecham Pharmaceuticals, New Frontiers Science Park, Harlow, Essex CM19 5AW, United Kingdom; and **Department of Medicinal Chemistry and §Department of Molecular Screening Technologies, SmithKline Beecham Pharmaceuticals, 709 Swedeland Road, King of Prussia, Pennsylvania 19406

Sequential proteolytic processing of the Amyloid Precursor Protein (APP) by ␤- and ␥-secretases generates the 4-kDa amyloid (A␤) peptide, a key component of the amyloid plaques seen in Alzheimer’s disease (AD). We and others have recently reported the identification and characterisation of an aspartic proteinase, Asp2 (BACE), as ␤-secretase. Here we describe the characterization of a second highly related aspartic proteinase, Asp1 as a second ␤-secretase candidate. Asp1 is expressed in brain as detected at the mRNA level and at the protein level. Transient expression of Asp1 in APP-expressing cells results in an increase in the level of ␤-secretase-derived soluble APP and the corresponding carboxy-terminal fragment. Paradoxically there is a decrease in the level of soluble A␤ secreted from the cells. Asp1 colocalizes with APP in the Golgi/endoplasmic reticulum compartments of cultured cells. Asp1, when expressed as an Fc fusion protein (Asp1-Fc), has the N-terminal sequence ALEP. . . , indicating that it has lost the prodomain. Asp1-Fc exhibits ␤-secretase activity by cleaving both wild-type and Swedish variant (KM/NL) APP peptides at the ␤-secretase site.

INTRODUCTION Alzheimer’s disease (AD) is a progressive neurodegenerative disorder characterized by the deposition of

1 To whom correspondence and reprint requests should be addressed. Fax: 01279-622555. E-mail: [email protected].

1044-7431/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.

the 4-kDa amyloid protein (A␤) in the brain as amyloid plaques (Glenner and Wong, 1984; Masters et al., 1985). A␤ is generated by the proteolytic processing of the type I transmembrane glycoprotein, the Amyloid Precursor Protein (APP; Kang et al., 1987; Selkoe, 1994). In the amyloidogenic pathway, cleavage of APP by ␤-secretase at the amino terminus of the amyloid domain, results in the release of a large soluble ⬃110-kDa N-terminal fragment, sAPP␤, and generation of a 12kDa membrane-anchored C-terminal fragment, CTF␤ (Haass et al., 1992). Subsequent cleavage of CTF␤ by ␥-secretase at the C-terminus of the amyloid domain gives rise to the 4-kDa A␤ peptide (Citron et al., 1992; Cai et al., 1993). In contrast, cleavage of APP in the nonamyloidogenic pathway prevents the release of intact A␤ (Esch et al., 1990; Sisodia et al., 1990). In this pathway, cleavage of APP by ␣-secretase within the amyloid domain results in the release of the large soluble N-terminal fragment, sAPP␣, and generation of a 10-kDa membrane-anchored C-terminal fragment, CTF␣ (Selkoe et al., 1988; Weidemann et al., 1989). Subsequent cleavage of CTF␣ by ␥-secretase gives rise to the 3-kDa p3 peptide. The identity of the APP secretases has eluded scientists for over a decade. However, recently we and others have reported that ␤-secretase is a novel membranebound aspartic proteinase (Asp2/Beta APP-cleaving enzyme, BACE/Memapsin 2; Hussain et al., 1999; Vasser et al., 1999; Yan et al., 1999; Sinha et al., 1999; Lin

609

610 et al., 2000). This proteinase, which we refer to as Asp2, has many of the characteristics expected of ␤-secretase. Asp2 is present in brain and in cultured cells colocalizes with APP in the Golgi compartments. Transient expression of Asp2 in APP-expressing cells results in cleavage of APP at the ␤-secretase site and, hence, an increase in the level of sAPP␤ secreted into the medium and a concomitant accumulation of CTF␤ in the cells. We report here that the Asp2 homologue, Asp1 (BACE2, DRAP; Yan et al., 1999; Acquati et al., 2000; Saunders et al., 1999; Bennett et al., 2000), also exhibits ␤-secretaselike activity. Transient expression of Asp1 in APP-expressing cells results in an increase in the levels of ␤-secretase derived cleavage products, sAPP␤ and CTF␤. We also show that Asp1 expressed as a recombinant Fc chimera cleaves both wild-type and Swedish variant APP-derived peptides at the ␤-secretase site.

RESULTS We have identified Asp1 using a proprietory EST database and subsequently cloned the full-length cDNA from a melanoma Marathon-Ready cDNA preparation (Clontech Laboratories, Inc.). Asp1 is highly homologous to Asp2 at the amino acid level as shown in Fig. 1. The amino acid sequence of Asp1 has two conserved active site motifs characteristic of aspartic proteinases at amino acid positions 110 to 113 (DTGS) and 303 to 306 (DSGT). In common with Asp2, but unlike all other mammalian aspartic proteinases identified to date, Asp1 has a C-terminal extension, which includes a transmembrane spanning domain. Asp1 is a type I transmembrane glycoprotein with the extracellular domain of the protein residing in the lumen of vesicles and the short C-terminal tail being cytoplasmic (Acquati et al., 2000). In order to evaluate the relative levels of Asp1 and Asp2 mRNA in brain we have used TaqMan analysis on brain regions from four individuals (Fig. 2). The mRNA’s for Asp1 and Asp2 are detectable in all the brain regions examined and have similar distribution profiles, with clear differences between brain regions. Asp1 mRNA is highest in the spinal cord, medulla oblongata, substantia nigra, and locus coruleus. The mRNA for Asp2 is highest in the substantia nigra, locus coruleus, and medulla oblongata. To demonstrate the presence of Asp1 protein in brain we have raised a rabbit polyclonal antibody to Asp1. The affinity purified rabbit polyclonal anti-Asp1 antiserum detects Asp1 as a single band of ⬃62-kDa (Fig. 3a, lane 1) in Asp1-transfected cells and does not cross-

Hussain et al.

react with overexpressed Asp2 (Fig. 3a, lanes 3 and 4). Incubation of the Asp1-transfected cell lysate with Nglycosidase F results in a decrease in the molecular weight of Asp1 to ⬃55-kDa (Fig. 3a, lane 2), indicating that Asp1 is glycosylated in cells. The specificity of the Asp1 bands is confirmed by preabsorption of the antiserum with the immunising peptide (Fig. 3a, lanes 5 and 6). Expression of Asp2 is confirmed by Western blot analysis of the same samples used in lanes 3 and 4 with the anti-His antibody (Fig. 3a, lanes 7 and 8). Treatment of the Asp2-transfected cell lysate with N-glycosidase F resulted in a reduction in the molecular weight of the Asp2, thus confirming earlier reports (Bennett et al., 2000) that Asp2 is glycosylated in cells. The affinity purified anti-Asp1 and anti-Asp2 antisera (Hussain et al., 1999) have been used to determine protein expression in human brain by Western blot analysis (Fig. 3b). Immunoreactive bands at the correct molecular weight for Asp1 and Asp2 are present in homogenates of frontal cortex from AD patients. Asp1 immunoreactivity is observed in lanes 2 and 3 and in lane 1 upon prolonged exposure of the blot. An Asp2 immunoreactive band is detected in lane 1 and in lanes 2 and 3 upon prolonged exposure of the blot. The Asp1 and Asp2 immunoreactive bands are abolished by preincubation with the immunising peptides. The Asp1 antiserum has also been used to determine Asp1 protein expression in human brain sections from 12 AD patients and aged controls by immunohistochemistry. All samples show labeling of pyramidal neurons throughout the hippocampus and temporal cortex, representative data from one AD patient is shown (Figs. 4a, 4b, and 4d). Only background staining is evident with preimmune serum (Fig. 4c) and preincubation of the Asp1 antibody with its immunogenic peptide results in a loss of specific cell labeling (Figs. 4e and 4f ). Immunoreactivity is largely neuronal with little evidence of labeling of other cell types, such as microglia or oligodendrocytes. Distribution of Asp1 staining in aged control brain shows a very similar pattern to that observed in the AD tissue (data not shown). Asp1 protein is also detected in cortical neurons in close proximity to A␤ plaques in sections from Trisomy 21 patients (data not shown). The Role of Asp1 in the Processing of APP Given the expression of Asp1 in brain and sequence homology to Asp2, we investigated the potential role of Asp1 in the processing of APP. Transient expression of Asp1mycHis or Asp2mycHis in SK-N-SH cells enables detection of the protein by Western blot analysis with

Processing of APP by Asp1

611

FIG. 1. Alignment of the amino acid sequences of Asp1 (Genbank Accession No. AF204944) with Asp2 (Genbank Accession No. AF204943). Regions of identity are shown in black. The predicted signal sequence at the N-terminus is shown green. The prodomain of Asp2 is shown in red. The N-terminal sequence of the Asp1-Fc is shown by the red arrow. The active site motifs DTGS and DSGT are shown in yellow. The membrane spanning domain of Asp1 and Asp2 is shown on dark blue. Potential N-linked glycosylation sites are also shown in pink.

an anti-His antibody (Fig. 5a). Expression of Asp1 is approximately twofold lower than Asp2. Asp1 migrates as a glycoprotein band of ⬃62-kDa (Fig. 5a), whereas Asp2 migrates as a glycoprotein of ⬃70-kDa as reported previously (Hussain et al., 1999). Overexpression of Asp1 results in an increase in the secretion of sAPP␤ into the medium, as detected using a specific antibody (Hussain et al., 1999), which recognizes the neo-epitope revealed after ␤-secretase cleavage of APP (Fig. 5b, lanes 4 – 6) when compared to the vector transfected cells (Fig. 5b, lanes 1–3). The increase in the secretion of

sAPP␤ caused by overexpression of Asp1 is similar to that observed with Asp2 (Fig. 5b, lanes 7–9) if the level of expression of the two proteinases are taken into account (band density for sAPP␤/band density for His tagged protein; 0.97 ⫾ 0.19 for Asp1, 0.77 ⫾ 0.11 for Asp2). In order to show that the effect of overexpressing Asp1 on the level of sAPP␤ is not restricted to SK-N-SH cells we have overexpressed Asp1 in SH-SY5Y cells stably expressing APP695 (Fig. 5c). Similar to the results obtained in SK-N-SH cells, overexpression of Asp1 in the SH-SY5Y APP695 cells causes a 2.6-fold increase in

612

Hussain et al.

FIG. 2. TaqMan distribution of Asp1 (top panel), Asp2 (middle panel), and ␤-actin (bottom panel) over a range of tissues from human brain isolated from the tissues of 4 individuals (2 males, 2 females). Expression is presented as mRNA detected per ng mRNA pool.

the secretion of sAPP␤ (Fig. 5d; lanes 1–3 band density 1.09 ⫾ 0.41; lanes 4 – 6 band density 2.8 ⫾ 0.1). In agreement with the report by Bennett et al. (2000), overexpression of Asp1 does not increase A␤ secretion, there is a sevenfold reduction in the concentration of A␤x-40 in the media from SK-N-SH cells overexpressing Asp1 (Fig. 5e), this is confirmed by Western blot analysis of the media for A␤x-40 (data not shown). Similar results were obtained in other cell systems (data not shown). It has been shown with Asp2 that the increase in secretion of sAPP␤ is reflected in an accumulation of the corresponding intracellular C-terminal fragment (CTF␤) in cells. Thus, we transiently expressed Asp1 in COS-7 cells overexpressing APP751 with the Swedish variant (KM 651,652/NL, APP751 numbering; Mullan et al., 1992), which has been shown to be a better substrate for Asp2 compared to the wild-type sequence (Vasser et al., 1999). There is a slight decrease in full length APP upon overexpression of both Asp1 and Asp2 compared to vector transfected cells (Fig. 5f ). Overexpression of Asp1 results in the accumulation of the 12-kDa C-terminal fragment (CTF␤), as previously reported with Asp2 (Hussain et al., 1999). Overexpression of Asp1 also results in an increase in the level of a second CTF of ⬃10-kDa. This CTF is most likely gen-

FIG. 3. Asp1 immunoreactivity in cells and in human brain. (a) Asp1 immunoreactivity in cell lysates transfected with Asp1mycHis (lanes 1 and 2) or Asp2mycHis (lanes 3 and 4), untreated (⫺) or treated (⫹) with N-glycosidase F. Asp1 immunoreactivity is abolished by preincubation of the antiserum with the immunizing peptide (lanes 5 and 6). Expression of the Asp2 protein, as detected using the anti-His antibody, is shown in lanes 7 and 8. (b) Asp1 (upper panel) and Asp2 (lower panel) immunoreactivity in AD brain (lanes 1–3), and after preincubation with the immunizing peptides (lanes 4 – 6).

613


FIG. 4. Asp1 immunoreactivity in the hippocampus and temporal cortex of a 62-year-old female AD patient: (a and b) hippocampal pyramidal cells; (c) similar region to (a) incubated with preimmune serum; (d) temporal cortex, outer pyramidal layer; (e) as (d) but Asp1 antiserum preabsorbed with peptide immunogen prior to incubation with section; (f ) phase contrast of image (e) showing pyramidal cell layer. Scale bars represent 100 ␮m (a), 25 ␮m (b), or 50 ␮m (c–f ).

erated following Asp1 cleavage of APP at a second site downstream from the major ␤-secretase site, for instance, following cleavage between residues Tyr10Glu11 of the A␤ domain as has been reported for Asp2 (Vasser et al., 1999). Asp1 Colocalizes with APP If APP is a physiological substrate for Asp1, both proteins must colocalize in cells. To determine the intracellular localisation of Asp1, we transiently transfected COS-7 cells expressing Swedish variant APP751 with mycHis tagged Asp1 allowing detection of Asp1 with an anti-myc antibody. Figure 6a shows that Asp1 localizes to the Golgi as revealed by the strong juxtanuclear staining pattern throughout the cell. A similar staining pattern is seen for APP in these cells (Fig. 6b). Merging of the confocal images for Asp1 and APP (Fig. 6c) indicates that the two protein colocalize in the Golgi. Confirmation of this localization is obtained using markers for the endoplasmic reticulum and Golgi, the KDEL epitope (Fig. 6d) and the Golgi 58K protein (Fig. 6e), respectively, which show staining very similar to that seen for Asp1 and APP. However, the distribution of Asp1 and APP is quite distinct from that seen for the

early endosome marker EEA1, which shows peripheral staining and no staining in the juxtanuclear region (Fig. 6f ). Asp1 Cleavage of APP ␤-Secretase Cleavage Site Peptides In order to obtain direct evidence that Asp1 exhibits ␤-secretase-like enzymatic activity, we determined if recombinant Asp1 could cleave synthetic APP peptides at the ␤-secretase site. Asp1 lacking the native signal sequence was expressed as a soluble Fc fusion protein in COS-7 cells and the protein purified from the medium using protein A–sepharose. The Asp1-Fc migrates as a band at ⬃100-kDa (Fig. 7a), a second band at ⬃30kDa is also present. Western blot analysis of this material with the anti-Asp1 antibody shows that the major Coomassie stained band (⬃100-kDa) is Asp1, whereas both Coomassie stained bands were immunoreactive with an anti-Fc antibody (data not shown), suggesting that the latter represent degraded forms of the Asp1-Fc chimera. N-terminal sequencing of the 100-kDa band shows ⬎90% of the protein commences at amino acid 63 (Ala) of Asp1. Incubation of Asp1-Fc with the ␤-secretase cleavage site peptides results in cleavage at a single

614

Hussain et al.

FIG. 5. Effect of Asp1 and Asp2 overexpression on APP processing. (a) Expression of the transfected proteins, Asp1mycHis and Asp2mycHis in SK-N-SH cells. (b) Increase in sAPP␤ secreted into the medium from the Asp1- and Asp2-transfected SK-N-SH cells compared to vector-transfected cells. (c) Expression of Asp1mycHis in SH-SY5Y APP695 cells compared to vector transfected cells. (d) Increase in sAPP␤ secreted into the medium from the Asp1-transfected SH-SY5Y APP695 cells compared to vector-transfected cells. (e) Decrease in the level of A␤x-40 secreted into the medium from the Asp1 transfected SK-N-SH cells compared to vector and Asp2 transfected cells. Data are expressed as mean ⫾ SD ng A␤x-40 per mg of cell protein. (f ) Full-length APP (APP FL) and CTFs in Asp1-, Asp2-, or vector-transfected COS-7 APP751swe cells.

site in both the wild-type (ISEVKMDAEFRHDKdnpG) and the Swedish variant (ISEVNLDAEFRHDKdnpG) APP peptides (Fig. 7b). Under the incubation conditions used the Swedish variant is cleaved to a greater extent than the wild-type peptide. Mass spectrometric analysis shows that this cleavage occurs at the expected ␤-secretase site, between the Met-Asp residues in wild-type and Leu-Asp residues in Swedish variant APP peptides.

DISCUSSION The production and deposition of A␤ in the brain is widely accepted as a key pathological event in the aetiology of AD. Until recently, the enzymes that are involved in the production of A␤ from APP have remained elusive. Intensive research has now shown that presenilin-1 may be an essential diaspartyl cofactor for ␥-secretase or maybe ␥-secretase itself (Wolfe et al., 1999a,b). We (Hussain et al., 1999) and others (Vasser et al., 1999; Yan et al., 1999; Sinha et al., 1999; Lin et al.,

2000) have recently reported that ␤-secretase is a novel aspartic proteinase. This proteinase differs from all other human aspartic proteinases identified to date in that it has the shortest prodomain and a C-terminal extension which includes a transmembrane domain. A second aspartic proteinase, which shows high homology to Asp2, has also been identified (Yan et al., 1999; Acquati et al., 2000). This proteinase, which we refer to as Asp1 (BACE2), has been localized to the Down syndrome obligate region of chromosome 21 at 21q22.2– 21q22.3 (Saunders et al., 1999). Down syndrome (Trisomy 21) sufferers have been shown to deposit A␤ in the brain at a relatively early age; this has been attributed to overexpression of APP, which also resides in the Down syndrome critical region of chromosome 21 (Murphy et al., 1990; Mann et al., 1990). Here we show that Asp1 exhibits many of the characteristics expected of ␤-secretase. Northern blotting has shown that the messenger RNA for Asp1 is present in a range of tissues and at a low level in brain when compared to Asp2 (Yan et al., 1999; Bennett et al., 2000).


615

FIG. 6. Subcellular localization of Asp1 and APP in COS-7 APP751swe cells. (a) Staining for Asp1. (b) Staining for APP. (c) Merged image of a and b showing significant colocalisation (yellow) of Asp1 and APP. (d) Staining for endoplasmic reticulum marker, KDEL. (e) Staining for Golgi 58K protein. (f ) Staining for Early Endosome Antigen 1 (EEA1).

Using sensitive TaqMan analysis we show that the mRNA for Asp1 and Asp2 are highest in the same brain regions, for example in the substantia nigra. In addition, those brain areas which are low in Asp2 mRNA also tend to be low for Asp1 mRNA, for example, in the striatum (Fig. 3). Despite the mRNA level detected for Asp1 being lower than that for Asp2 we show by Western blot analysis that Asp1 protein is present in brain. In addition, immunohistochemical analysis demonstrates that Asp1 protein is present in neurons within the hippocampus, frontal cortex, and temporal cortex in brain from both AD and aged control subjects (data not shown). Transient expression of Asp1 in cells results in cleavage of APP at the expected ␤-secretase site leading to an increase in the level of sAPP␤ secreted into the medium and an accumulation of CTF␤ in the cells. Overexpression of ␤-secretase would be expected to result in an increase in the production of A␤. However, we report that overexpression of Asp1 results in a paradoxical decrease in the level of A␤ in the cell conditioned medium. A reduction in the level of A␤ in the medium has previously been reported following overexpression of Asp2 (Yan et al., 1999). The formation of A␤ is not a direct measure of ␤-secretase activity as its generation relies on both ␤- and ␥-secretase cleavage activities,

whereas the accumulation of CTF␤ upon overexpression of Asp1 and Asp2 is diagnostic for cleavage of APP at the ␤-secretase site. The decrease in A␤ from cells overexpressing Asp1 may be a consequence of inefficient processing of the accumulated CTF␤ by ␥-secretase. Alternatively the A␤ may not be secreted from the cells. It has been reported that overexpression of Asp2 results in the generation of CTF’s which start at residues glutamic acid 11 and aspartic acid 1 of A␤ (Vasser et al., 1999). However, A␤ species which start at glutamic acid 11 are not detected in the medium (Vasser et al., 1999), suggesting that any N-terminally truncated A␤ species do not leave the cell. Despite this drop in extracellular A␤ the accumulation of intracellular CTF␤ may be important in the aetiology of AD as it has been shown that CTF␤ (otherwise called C100) is directly toxic to neuronal cells (Yankner et al., 1989; Fukuchi et al., 1993; Sopher et al., 1994). Furthermore CTF␤ transgenic mice show AD-like pathology (Neve et al., 1996). In confirmation of the capacity of Asp1 to cleave APP we have shown that Asp1 colocalizes with APP within the intracellular Golgi compartments, where ␤-secretase cleavage is known to occur (Haass et al., 1995; Martin et al., 1995; Thinkaran et al., 1996). We also show, by expressing Asp1 as a Fc chimera, that Asp1 lacking the prodomain can cleave peptide substrates that span

616

Hussain et al.

EXPERIMENTAL METHODS Cloning of Asp1 An EST encoding the partial gene sequence of the aspartic proteinase, Asp1, which contained the characteristic active site motif of an aspartic proteinase (hydrophobic-hydrophobic-Asp-Ser/Thr-Gly) was identified by searching databases with the sequence of mammalian aspartic proteinases. The missing 5⬘ end of the gene was generated by 5⬘ nested RACE PCR from a melanoma Marathon-Ready cDNA template (Clontech). The complete Asp1 coding region was assembled in pBluescript (Stratagene), using a unique NcoI site in the area of overlap between the EST and the novel 5⬘ sequence, and fully sequenced. FIG. 7. Asp1-Fc cleaves APP peptides at the ␤-secretase site. (a) Coomassie-stained gel of Asp1-Fc. (b) Hydrolysis of wild-type (WT) and Swedish variant (SWE) APP peptides by Asp1. Reverse phaseHPLC profile showing the dnp labeled substrate and product peaks at 360 nm. The amino acid sequence of the peptides as determined by mass spectrometry are shown.

the ␤-secretase cleavage site. Asp1-Fc cleaves peptides that have both the wild-type and the Swedish variant sequence and this cleavage occurs at the known ␤-secretase site (Met/Leu-Asp). Asp1 cleaves the Swedish variant APP peptide with greater efficiency than the wildtype APP peptide in agreement with that reported for Asp2 (Vasser et al., 1999). In conclusion, we report that Asp1 is able to cleave APP at the ␤-secretase site and thus may function as a second ␤-secretase. Despite low levels of the mRNA for Asp1 in brain we report that Asp1, like Asp2 has many of the characteristics expected of a candidate ␤-secretase. Although the physiological relevance of Asp1 activity in the brain remains to be confirmed, the presence of two closely related proteinases which exhibit ␤-secretase like activity raises many questions. For example, it may indicate the presence of a new family of aspartic proteinases, which may be involved in the processing of membrane bound substrates, such as APP. Given the chromosomal localization of Asp1 in the Down syndrome obligate region of chromosome 21, cleavage of APP by Asp1 may be more relevant in this syndrome where there may be 1.5⫻ the gene dosage of Asp1. In order to elucidate the potential redundancy in the APP processing pathway and the physiological relevance of APP cleavage by Asp1 it will be necessary to develop transgenic overexpressing and knockout mice and specific inhibitors of these two highly related proteinases.

TaqMan Analysis Poly(A) ⫹ RNA was prepared from brain subregion tissue from four different individuals (two males and two females, obtained from the Netherlands Brain bank), 1 ␮g was reverse transcribed and TaqMan mRNA analysis performed as described previously (Sarau et al., 1999), using gene-specific PCR reagents. Quantitation of mRNA-derived TaqMan signal was obtained using known plasmid/genomic DNA standards. AspI gene-specific reagents: forward primer 5⬘GCAACCATGAACTCAGCTATTAAGAA-3⬘, reverse primer 5⬘-AGAAAGCGCCACCATCGA-3⬘, TaqMan probe 5⬘-CCCGGCTGCTGCCCTGGAA-3⬘. Asp2 gene specific reagents: forward primer 5⬘-TATCATGGAGGGCTTCTACGTTG-3⬘, reverse primer 5⬘-GTCCTGAACTCATCGTGCACAT-3⬘, TaqMan probe 5⬘CCCGAAAACGAATTGGCTTTGCTGTC-3⬘. ␤-actin gene specific reagents as described previously (Sarau et al., 1999). Plasmids and Transfections AspI was cloned into pcDNA3.1mycHis (Invitrogen) for transient expression in mammalian cells. Transfections were performed as described previously (Hussain et al., 1999). Briefly, cells were transfected using LipofectAMINE PLUS Reagent (Life Technologies) and 24 h posttransfection the medium was changed to OptiMEM-1 (10 ml). The medium was collected 24 h later for the immunoassay and for concentration using Centriprep 10 concentrators (Amicon). Cells were harvested and lysed by incubation for 30 min at 4°C in 50 mM Tris/HCl pH 7.4, 1% Triton X-100 containing a cocktail of “complete” protease inhibitors (Boehringer-Mann-

617


heim). Cell lysates were deglycosylated using N-glycosidase F (Boehringer-Mannheim) as described by the manufacturer. To generate recombinant Asp1 base pairs 72–1398 were subcloned into the HindIII/XhoI sites in Signal pIgPlus vector (R&D Systems) to generate a C-terminal Fc chimera with a CD33 signal sequence. This was subsequently used to transfect COS-7 cells using DEAE dextran. Six days posttransfection, the Asp1 Fc fusion protein was purified from the medium of these cells using protein A–Sepharose beads and eluted with 0.1 M Glycine, pH 2.8, neutralized with 1 M Tris, pH 8.0, and dialyzed overnight versus PBS. Antiserum and Immunoblotting Rabbits were immunized with a KLH peptide conjugate of Asp1 amino acid sequence 47–59 (PGPGTPAERHADG). The antiserum was affinity purified using the peptide conjugated to Sulfolink (Pierce and Warriner Ltd) according to the manufacturer’s instructions. Proteins in the cell lysate (20 ␮g) or medium (15 ␮l of 20-fold concentrated media) were resolved on 10% Trisglycine or 10 –20% Tris-tricine SDS–polyacrylamide gels (Novex) for Western blot analysis as described previously (Hussain et al., 1999). Peptide competition was performed by incubating a 10-fold molar excess of the immunising peptide with the affinity purified antibody for 2 h at room temperature, prior to immunodetection. Asp1 and Asp2 proteins were detected by Western blotting using the affinity purified anti-Asp1 and Asp2 antisera (Hussain et al., 1999) in frontal cortex from AD patients. The frontal cortex was homogenized under liquid nitrogen and 200 ␮g of this powder was lysed at 4°C for 20 min in 500 ␮l of lysing buffer (0.1% Triton X-100, 20 mM Hepes, 25 mM NaCl, 2 mM EDTA, 0.5 mM DTT, 0.1 mM PMSF, and EDTA-free protease inhibitor cocktail (Boehringer-Mannheim). After centrifugation (20,800g, 4°C, 20 min) the supernatants were stored at ⫺80°C until used. Western blotting was performed as above with 12.5 ␮l of the lysate. A␤ Immunoassay The level of A␤ in the medium was measured by an immunoassay employing a specific A␤ C-terminal antibody configured in a sandwich format. For the assessment of A␤40, the monoclonal antibody 2F12 (to residues 4 – 8 of A␤) served as the capture antibody for A␤ in the medium and biotinylated monoclonal antibody G210 (to residues 35– 40 of A␤; Ida et al., 1996) provided the detection antibody. A standard curve was constructed using A␤1– 40 (California Peptide Research)

dissolved in dimethyl sulphoxide at 1 mg/ml and subsequently diluted in assay buffer (50 mM Tris, 150 mM sodium chloride, and 0.05% Tween 20). In brief the method used was as follows. After coating 96-well plates overnight at 4°C with 200 ␮l of 2F12 (5.64 ␮g/ ml), the plates were blocked for 1 h at 37°C in assay buffer containing gamma globulin (0.5%) and gelatin (1%). Plates were then washed and unconcentrated media or standard peptide was added together with the secondary biotinylated detection antibody G210 (4.76 ␮g/ml) in assay buffer and plates were incubated for a minimum of 24 h at 4°C. The amount of A␤ bound to the plates was subsequently determined by successive incubations with streptavidin-europium and enhancer solution (both Wallac) before measurement by timeresolved fluorescence using a DELFIA plate reader. Data is expressed as mean ⫾ SD ng of A␤x-40 normalized for protein in cells. Subcellular Localization cDNA encoding Asp1 with a mycHis epitope tag was transfected into COS-7 APP751 cells. Cells were fixed and processed for indirect immunofluorescence as described (Spector et al., 1998). APP was visualized using the anti C-terminal antibody Ab54. Control antibodies were, monoclonal Early Endosome Antigen 1 (EEA1; Transduction Labs), monoclonal anti-KDEL (Stressgen), monoclonal anti-myc (SantaCruz Biotechnology) and monoclonal anti-Golgi 58K protein (Sigma). Detection was with Alexa 488 labeled anti mouse or Alexa 568 labeled anti-rabbit antibodies (Molecular Probes). Microscopy was carried out using a Leica confocal microscope. Immunohistochemistry Ten-micrometer sections of paraformaldehyde-fixed hippocampus and temporal cortex from twelve AD patients and aged controls were rehydrated and labeled with the affinity purified antiserum to Asp1 by incubation overnight at 4°C. Subsequent processing used the biotin-avidin system (with biotinylated goat anti-rabbit antibody) and the chromogen diaminobenzidine (Vector Laboratories). Peptide Cleavage Studies Peptide (50 ␮M), WT ⫽ ISEVKMDAEFRHDK(dnp)G or SWE ⫽ ISEVNLDAEFRHDK(dnp)G was incubated with Asp1 (20 nM) in buffer containing 50 mM sodium acetate, 20 mM NaCl, pH 4.5, for 2 h (SWE) or 16 h (WT)

618 in a final assay volume of 20 ␮l. The reaction was stopped by addition of 4 volumes of 5% trifluoroacetic acid (TFA). The assay components were loaded onto a POROS R1 column (Applied Biosystems) in 0.08% TFA and eluted with a linear gradient of 0.08% TFA in acetonitrile (linear gradient from 20 –35% acetonitrile over 8 min). The chromogenic dnp group on the peptide was followed by monitoring at 360 nm. The second nonlabeled peak is not detectable at 360 nm due to loss of the dnp group.

ACKNOWLEDGMENTS We thank D. Dewar (Wellcome Surgical Institute, Glasgow, Scotland) for the AD brain sections and D. Mann (University of Manchester) for the Down syndrome sections used in the immunohistochemistry studies, N. Cairns (Institute of Psychiatry, London) for the brain samples used in the Western blotting, T. Hartmann and K. Beyreuther (University of Heidelberg, Germany) for the monoclonal antibody WO2, Dr. R. Ravid (Netherlands Brain Bank, The Netherlands), for arrangement/donation of brain tissue, G. Moore for provision of the TaqMan probes and W. Neville for N-terminal sequencing. We also thank C. W. Gray for critical reading of the manuscript.

REFERENCES Acquati, F., Accarino, M., Nucci, C., Fumagalli, P., Jovine, L., Ottolenghi, S., and Taramelli, R. (2000). The gene encoding DRAP (BACE2), a glycosylated transmembrane protein of the aspartic protease family, maps to the Down critical region. FEBS Lett. 468: 59 – 64. Bennett, B. D., Babu-Kahn, S., Loeloff, R., Louis, J-C., Curran, E., Citron, M., and Vasser, R. (2000). Expression analysis of BACE 2 in brain and peripheral tissues. J. Biol. Chem. 275: 20647–20651. Cai, X-D., Golde, T. E., and Younkin, S. G. (1993). Release of excess amyloid ␤ protein from a mutant amyloid ␤ protein precursor. Science 259: 514 –516. Citron, M., Oltersdorf, T., Haass, C., McConlogue, L., Hung, A. Y., Seubert, P., Vigo-Pelfrey, C., Lieberburg, I., and Selkoe, D. J. (1992). Mutation of the ␤-amyloid precursor protein in familial Alzheimer’s disease increases ␤-protein production. Nature 360: 672– 674. Esch, F. S., Keim, P. S., Beattie, E. C., Blacher, R. W., Culwell, A. R., Oltersdorf, T., McClure, D., and Ward, P. J. (1990). Cleavage of amyloid ␤ peptide during constitutive processing of its precursor. Science 248: 1122–1124. Fukuchi, K., Sopher, B., Furlong, C. E., Smith, A. C., Dang, N. T., and Martin, G. M. (1993). Selective neurotoxicity of COOH-terminal fragments of the ␤-amyloid precursor protein. Neurosci. Lett. 154: 145–148. Glenner, G. G., and Wong, C. W. (1984). Alzheimer’s disease: Initial report of the purification and characterisation of a novel cerebrovascular amyloid protein. Biochem. Biophys. Res. Commun. 120: 885– 890. Haass, C., Koo, E. H., Mellon, A., Hung, A. Y., and Selkoe, D. J. (1992). Targeting of cell-surface ␤-amyloid precursor protein to lysosomes: Alternative processing into amyloid-bearing fragments. Nature 357: 500 –503.

Hussain et al.

Haass, C., Lemere, C. A., Capell, A., Citron, M., Seubert, P., Schenk, D., Lannfelt, L., and Selkoe, D. J. (1995). The Swedish mutation causes early-onset Alzheimer’s disease by beta-secretase cleavage within the secretory pathway. Nature Med. 1: 1291–1296. Hussain, I., Powell, D., Howlett, D. R., Tew, D. G., Meek, T. D., Chapman, C., Gloger, I. S., Murphy, K. E., Southan, C. D., Ryan, D. M., Smith, T. S., Simmons, D. L., Walsh, F. S., Dingwall, C., and Christie, G. (1999). Identification of a novel aspartic protease (Asp2) as ␤-secretase. Mol. Cell. Neurosci. 14: 419 – 427. Ida, N., Hartmann, T., Pantel, J., Schroder, J., Zerfass, R., Forstl, H., Sandbrink, R., Masters, C. L., and Beyreuther, K. (1996). Analysis of heterogeneous ␤A4 peptides in human cerebrospinal fluid by a newly developed sensitive Western blot assay. J. Biol. Chem. 271: 22908 –22914. Kang, J., Lemaire, H-G., Unterbeck, A., Salbaum, J. M., Masters, C. L., Grzeschik, K-H., Multhaup, G., Beyreuther, K., and Mu¨ller-Hill, B. (1987). The precursor of Alzheimer’s disease amyloid A4 protein resembles a cell-surface receptor. Nature 325: 733–736. Lin, X., Koelsch, G., Wu, S., Downs, D., Dashti, A., and Tang, J. (2000). Human aspartic protease memapsin 2 cleaves the ␤-secretase site of ␤-amyloid precursor protein. Proc. Natl. Acad. Sci. USA 97: 1456 –1460. Mann, D. M., Jones, D., Prinja, D., and Purkiss, M. S. (1990). The prevalence of amyloid (A4) protein deposits within the cerebral and cerebellar cortex in Down’s syndrome and Alzheimer’s disease. Acta Neuropathologica 80: 318 –327. Martin, B. L., Schrader-Fischer, G., Busciglio, J., Duke, M., Paganetti, P., and Yankner, B. A. (1995). Intracellular accumulation of ␤-amyloid in cells expressing the Swedish mutant amyloid precursor protein. J. Biol. Chem. 270: 26727–26730. Masters, C. L., Simms, G., Weidemann, N. A., Multhaup, G., McDonald, B. L., and Beyreuther, K. (1985). Amyloid plaque core protein in Alzheimer disease and Downs syndrome. Proc. Natl. Acad. Sci. USA 82: 4245– 4249. Mullan, M., Crawford, F., Houlden, H., Axelman, K., Lulius, L., Winblad, B., and Lannfelt, L. (1992). A pathogenic mutation for probable Alzheimer’s disease in the APP gene at the N-terminus of beta-amyloid. Nature Genet. 1(5): 345–347. Murphy, G. M. Jr., Eng, L. F., Ellis, W. G., Perry, G., Meissner, L. C., and Tinklenberg, J. R. (1990). Antigenic profile of plaques and neurofibrillary tangles in the amygdala in Down’s syndrome: A comparison with Alzheimer’s disease. Brain Res. 537: 102–108. Neve, R. L., Boyce, F. M., McPhie, D. L., Greenan, J., and OsterGranite, M. L. (1996). Transgenic mice expressing APP-C100 in the brain. Neurobiol. Aging 17: 191–203. Sarau, H. M., Ames, R. S., Chambers, J., Ellis, C., Elshourbagy, N., Foley, J. J., Schmidt, D. B., Muccitelli, R. M., Jenkins, O., Murdock, P. R., Herrity, N. C., Halsey, W., Sathe, G., Muir, A. I., Nuthulaganti, P., Dytko, G. M., Buckley, P. T., Wilson, S., Bergsma, D. J., and Hay, D. W. P. (1999). Identification, molecular cloning, expression, and characterization of a cysteinyl leukotriene receptor. Mol. Pharmacol. 56: 657– 663. Saunders, A. J., Kim, T-W., and Tanzi, R. E. (1999). BACE maps to chromosome 11 and BACE homolog, BACE 2, reside in the obligate Down Syndrome region of chromosome 21. Science 286: 125–1255a. Selkoe, D. J., Podlisny, M. B., Joachim, C. L., Vickers, E. A., Lee, G., Fritz, L. C., and Oltersdorf, T. (1988). Beta-amyloid precursor protein of Alzheimer disease occurs as 110- to 135-kilodalton membrane-associated proteins in neural and nonneural tissues. Proc. Natl. Acad. Sci. USA 85: 7341–7345. Selkoe, D. J. (1994). Cell biology of the amyloid beta-protein precursor and the mechanism of Alzheimer’s disease. Annu. Rev. Cell Biol. 10: 373– 403.


Sinha, S., Anderson, J. P., Barbour, R., Basi, G. S., Caccavello, R., Davis, D., Doan, M., Dovey, H. F., Frigon, N., Hong, J., JacobsonCroak, K., Jewett, N., Keim, P., Knops, J., Lieberburg, I., Power, M., Tan, H., Tatsuno, G., Tung, J., Schenk, D., Seubert, P., Suomensaari, S. M., Wang, S., Walker, D., Zhao, J., McConlogue, L., and John, V. (1999). Purification and cloning of amyloid precursor protein ␤-secretase from human brain. Nature 402: 537–540. Sisodia, S. S., Koo, E. H., Beyreuther, K., Unterbeck, A., and Price, D. L. (1990). Evidence that beta-amyloid protein in Alzheimer’s disease is not derived by normal processing. Science 248: 492– 495. Sopher, B. L., Fukuchi, K., Smith, A. C., Leppig, K. A., Furlong, C. E., and Martin, G. M. (1994). Cytotoxicity mediated by conditional expression of a carboxyl-terminal derivative of the ␤-amyloid precursor protein. Mol. Brain Res. 26: 207–217. Thinakaran, G., Teplow, D. B., Siman, R., Greenberg, B., and Sisodia, S. S. (1996). Metabolism of the “Swedish” amyloid precursor protein variant in Neuro2a (N2a) cells. J. Biol. Chem. 271: 9390 –9397. Spector, D. L., Goldman, R. D., and Reinweld, L. A. (1998). In Cells, a Laboratory Manual, Vol. 3, p. 105.3. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Vasser, R., Bennett, B. D., Babu-Khan, S., Kahn, S., Mendiaz, E. A., Denis, P., Teplow, D. B., Ross, S., Amarante, P., Loeloff, R., Luo, Y., Fisher, S., Fuller, J., Edenson, S., Lile, J., Jarosinski, M. A., Biere, A. L., Curran, E., Burgess, T., Louis, J-C., Collins, F., Treanor, J.,

619 Rogers, G., and Citron, M. (1999). ␤-secretase cleavage of Alzheimer’s amyloid precursor protein by the transmembrane aspartic protease BACE. Science 286: 735–741. Weidemann, A., Konig, G., Bunke, D., Fischer, P., Salbaum, J. M., Masters, C. L., and Beyreuther, K. (1989). Identification, biogenesis, and localization of precursors of Alzheimer’s disease A4 amyloid protein. Cell 57: 115–126. Wolfe, M. S., Xia, W., Ostaszewski, B. L., Diehl, T. S., Kimberly, W. T., and Selkoe, D. J. (1999a). Two transmembrane aspartates in presenilin-1 required for presenilin endoproteolysis and ␥-secretase activity. Nature 398: 513–517. Wolfe, M. S., De Los Angelos, J., Miller, D. D., Xia, W., and Selkoe, D. J. (1999b). Are presenilins intramembrane-cleaving proteases? Implications for the molecular mechanism of Alzheimer’s disease. Biochemistry 38: 11223–11230. Yan, R., Bienkowski, M. J., Shuck, M. E., Miao, H., Tory, M. C., Pauley, A. M., Brashler, J. R., Stratman, N. C., Mathews, W. R., Buhl, A. E., Carter, D. B., Tomasselli, A. G., Parodi, L. A., Heinrikson, R. L., and Gurney, M. E. (1999). Membrane-anchored aspartic protease with Alzheimer’s disease ␤-secretase activity. Nature 402: 533–537. Yanker, B. A., Dawes, L. R., Fisher, S., Villa-Komaroff, L., OsterGranite, M. L., and Neve, R. L. (1989). Neurotoxicity of a fragment of the amyloid precursor associated with Alzheimer’s disease. Science 245: 417– 420. Received May 17, 2000 Revised July 7, 2000 Accepted July 11, 2000