Pro-domain removal in ASP-2 and the cleavage of the amyloid precursor are influenced by pH
© Sidera et al. 2002
Received: 16 May 2002
Accepted: 31 August 2002
Published: 31 August 2002
Skip to main content
© Sidera et al. 2002
Received: 16 May 2002
Accepted: 31 August 2002
Published: 31 August 2002
One of the signatures of Alzheimer's disease is the accumulation of aggregated amyloid protein, Aβ, in the brain. Aβ arises from cleavage of the Amyloid Precursor protein by β and γ secretases, which present attractive candidates for therapeutic targeting. Two β-secretase candidates, ASP-1 and ASP-2, were identified as aspartic proteases, both of which cleave the amyloid precursor at the β-site. These are produced as immature transmembrane proteins containing a pro-segment.
ASP-2 expressed in HEK293-cells cleaved the Swedish mutant amyloid precursor at different β-sites at different pHs in vitro. Recent reports show that furin cleaves the pro-peptide of ASP-2, whereas ASP-1 undergoes auto-catalysis. We show that purified recombinant ASP-2 cleaves its own pro-peptide at ph 5 but not pH 8.5 as seen by mass spectrometry, electrophoresis and N-terminal sequencing.
We suggest that ASP-2 processing as well as activity are influenced by pH, and hence the cellular localisation of the protein may have profound effects on the production of Aβ. These factors should be taken into consideration in the design of potential inhibitors for these enzymes.
Alzheimer's disease is a common age-related dementia which is characterised pathologically by the appearance of brain senile plaques [1, 2] composed primarily of aggregated forms of Aβ. These are 39–43 residue peptides released following proteolytic processing of the transmembrane precursor glycoprotein, APP. The amyloidogenic pathway requires the APP to be sequentially cleaved by β and γ secretases [3, 4]. β-Secretase cleaves APP close to the membrane to produce βAPPs (secreted), and the 12-kDa, C100 transmembrane stub, subsequently cleaved by γ-secretase to produce the Aβ peptide and a cytoplasmic fragment with very short half life. α-Secretase cleaves APP within the Aβ sequence thus preventing its formation producing the N-terminal αAPPs domain and the 10-kDa membrane-localised C-terminal stub, C83. As aggregated Aβ is thought to promote neuronal death [5–7], the secretases represent potential drug targets for the treatment and/or prevention of AD. Presenilin-1 was suggested to be the ideal candidate for γ-secretase  whereas α-secretase has been characterised as ADAM10 disintegrin and metalloprotease . Recently, several groups used expression cloning, genomic search, or purification and proteomic analysis [10–13] to clone and identify β-secretase as an aspartic endopeptidase (EC 3.4.23) named BACE (β-site APP clevage enzyme), ASP-2 (aspartic protease 2), or memapsin 1. An additional candidate, ASP-1, BACE-2 or memapsin-2, has also been cloned [14–16]. The ASP-2 gene codes for a signal peptide, a pro-peptide (22-TQHGIRLPLRSGLGGAPLGLRLPR-46), followed by the catalytic domain, a transmembrane segment and a cytoplasmic C-terminal tail. Several cysteine residues are present, six of which are in the lumenal domain which may form intramolecular disulphide bridges contributing to the folding of the active site . Both ASP-1 and -2 are extensively glycosylated  and phosphorylated , and contain S-palmitoyl groups which may aid membrane anchorage.
Nearly all aspartic proteinases (EC 3.4.23.X) are synthesized as zymogens which are converted to active enzymes at acidic pH by proteolytic cleavage of the pro-segment [21, 22]. This process is autocatalytic for some pro-enzymes such as pepsinogen [23–25] and cathepsins D  and E . Furin is thought to cleave the ASP-2 pro-domain [20, 28, 29], though other pro-protein convertases were effective as well [20, 30]. In contrast, ASP-1 has been demonstrated to cleave its own pro-peptide . Although the pro-sequences of ASP-1 and ASP-2 are dissimilar we investigated whether ASP-2 also cleaves its pro-peptide. The activity of ASP-2 was assessed using the Swedish mutant form of APP  which is more readily cleaved by β-secretase than the wild type . We investigated the effects of pH on ASP-2 processing and activity. We report that pH affects the pro-domain removal of ASP-2 in vitro as well as its site of β-secretase cleavage in APP to produce Aβ.
Understanding what governs the activity of ASP-2 toward the APP substrate at a cellular and molecular level could facilitate the discovery of compounds that could inhibit the development of AD. It was therefore important to study the maturation, processing and activity of ASP-2 and relate that to the proteolytic events that lead to Aβ production. In agreement with several other reports [15, 17, 18, 37] glycosylated proteins were successfully produced in HEK293 cells displaying mobility of ~55–70 and 65–80-k, for ASP-1 and ASP-2, respectively. The broadness of these bands suggested a heterogeneous protein population, as a result of variable translational modification which was mainly due to glycosylation and possibly pro-processing and as also detected by others in the fully glycosylated, endoglycosidase H-resistant ASP-2 forms [17, 18, 31, 34]. Using a new anti-ASP-2 IgG, we detected the presence of higher molecular weight proteins in the purified preparations, by analysis under denaturing SDS-PAGE, which corresponded in size to ASP-2 homo-multimers. Oligomerisation has also been identified  for pepsin and cathepsin E [27, 40]. A putative dimer was clearly noted at ~150-k, corresponding to (ASP-2)2, which was also sensitive to deglycosylation giving rise to a change in mobility which corresponded to two monomers. The appearance of the largest ASP-2 forms was partly dependent on concentration and length of storage suggesting this is partly due to non-specific aggregation. In contrast, the putative dimers (~140 kDa) were evident even in fresh samples and were mainly insensitive to reducing agent providing evidence for the existence of strong non-disulphide protein-protein interactions not broken by high salt and non-ionic detergents used during purification or the strong denaturing conditions of SDS-boiling used during SDS-PAGE. Recently a laboratory in Germany reported the isolation of native active ASP-2 dimers from human brain homogenates, (Multhaup, G. and colleagues from Germany, unpublished data). These native dimers were similar to those observed in our preparations as they were SDS-resistant.
Activation of acidic proteases by removal of a pro-peptide occurs either by autocatalysis as in the case of pepsinogen [23, 25] and other aspartic proteases, or the action of other proteases, like Cathepsins S and L which activate Cathepsin C . Several reports by other laboratories, showed furin to successfully cleave the ASP-2 pro-segment [20, 29] also demonstrating that recombinant ASP-2 does not cleave its own synthetic or recombinant fusion pro-segments. However, Hussain et al. (2001), have recently shown that minimal cleavage of ASP-1 pro-domain is achieved by ASP-2 whereas ASP-1 displayed a unimolecular pro-peptide auto-removal. Here we demonstrated in vitro cleavage of the ASP-2 pro-peptide in full-length, pure, recombinant proteins instigated by incubation of ASP-2 at acidic pH, implying autocatalysis. We propose that cleavage occurs at two sites RLPLR↓S32 and RLPR↓E46 within the pro-domain, suggesting a two-step cleavage resembling the mechanism of maturation of other aspartic proteinases by autocatalysis (for a review see Richter et al., 1998). We have not investigated to see if the pro-sequence removal from ASP-2 is a unimolecular reaction, or due to the attack of a second ASP-2 molecule. APP cleavage at both the Asp1 (β1-site) as well as the Glu11 (β2-site) positions shows that ASP-2 has a preference to cleave N-terminally to an acidic residue as is the case for the pro-peptide cleavage. However, although the RXXR↓ found in the ASP-2 pro-sequence is thought to be a minimal sequence required for a pro-peptide convertase  furin prefers to cleave C-terminal to two adjacent basic residues. Recent results  showed that ASP-2 processing takes place in furin-deficient cell lines and suggested that other pro-peptide convertases are involved whereas Bennett et al., 2000 observed that a proportion of recombinant ASP-2 is processed in cells untransfected with furin. Further to those findings we suggest that ASP-2 can remove its pro-domain by autocatalysis in vitro, as the BACE inhibitor inhibits this process whereas EDTA does not, so that an additional pro-peptide convertase may not be required. Hence, it is possible that autocatalysis could be preferred under conditions of acidic pH, or in cells or cellular subcompartments where pro-peptide convertases are not expressed. Although other reports have shown that ASP-2 does not cleave its pro-domain in vitro[20, 31] the difference between those and our observations could be due to differences in the protein preparations. We used full-length pure proteins for enzyme and substrate whereas, synthetic peptides as well as truncated fusion proteins were employed in those studies. These factors in addition to differences in expression systems (and modifications) could influence the results observed herein as encountered for pepsinogen, which displays different activation kinetics as a fusion than as a whole protein . To conclude we suggest that ASP-2 may cleave its pro-peptide in a two-step mechanism in the absence of any other protease. Multiple-step removal of pro-sequences has been seen with other aspartic proteinases, where complete activation sometimes requires cleavage at more than one site .
We showed in vitro β-secretase activity of ASP-2, producing a C-terminal APP fragment, Mw 12-kDa, consistent with APPswe cleavage at the β1-site. This cleavage occurs efficiently at ph 8.5 although the optimum pH of ASP-2 activity is acidic. We must mention though that a proportion of ASP-2 in the ph 8.5 preparation is processed (as purified from the cells) and could contribute to the β-cleavage of APP we observe at ph 8.5. Cleavage of APPswe at ph 5, however was different to pH 8.5 as it resulted in two C-terminal APP fragments each corresponding to cleavage at the Aβ-producing site (β1), and the β2-site described by Creemers et al. (2000), to be the most favourable producing N-terminally truncated, highly cytotoxic , Aβ peptides (11–42) found in patients with the Flemish mutation. Creemers et al. (2000) suggested that cleavage at these sites is governed by the expression levels of ASP-2, in vivo, rather than its maturation. In addition to that we report here that our results show that pH can also determine the site of APPswe β-cleavage by ASP-2. Results from our laboratory  have localised β-cleavage of APP in the TGN and early endosomes, and recently Huse et al. (2000) showed TGN localisation for ASP-2 indicating that these compartments are loci for β2-site cleavage of APP as their pH is acidic. On the other hand, localisation in the endoplasmic reticulum and early Golgi imply processing at neutral to alkaline pH that according to our results gives cleavage at the β1-site and production of Aβ (1–40/42).
Taken together, our results of ASP-2 pro-peptide processing and diverse APP proteolysis, both happening at ph 5, imply an additional mode of cellular control of APP β-cleavage whereby subcellular localisation of APP and ASP-2 in different compartments where the pH is either acidic or alkaline influences the proteolytic processing of both ASP-2 as well as of APP by ASP-2 itself and consequently the production of Aβ. Thus APP cleavage at ph 5 is relevant to Alzheimer's in patients with the Flemish and Dutch mutations who demonstrate increased deposition of N-terminally truncated Aβ, resulting from APP processing at the β2-site.
Culture media and antibiotics were from Life Technologies, Paisley, UK. All culture plasticware, were from Nalge Nunc International, Loughborough, UK. Amino Acids for peptide synthesis were from Nova Biochem, Nottingham, UK. Freunds Adjuvant, standard chemicals, 9E10 clone ascites fluid (anti-myc), Triton X-100, Nonidet P40, and Kodak X-Omat film were purchased from Sigma-Aldrich Company Ltd, Dorset, UK, 6E10 (anti-Aβ) from ID labs, P.O. Box 3556, Glasgow, Scotland, United Kingdom. C-18 ZIP tips, Centricon-30 concentrators, PVDF membrane were purchased from Millipore. Automated N-terminal peptide sequencing was carried out by AltaBiocscience, University of Birmingham, Edgbaston, UK. All secondary horseradish peroxidase-linked anti-species IgG, Enhanced chemiluminescence reagents, ECL Hyper film, Superdex 200HR, Sephadex G-10, were from Amersham Pharmacia Biotech, UK Ltd, Buckinghamshire, UK, whereas the Ni2+-affinity resin NiNTA was from Novagen, Nottingham, UK. Pfx2 Lipids were initially purchased from Invitrogen Life Technologies, Paisley, UK. Ready-made Tris-Tricine gels and low molecular weight Kaleidoscope Polypeptide Standards were from BIORAD, Hemel Hempstead, UK. Prestained protein molecular weight markers and Acrylamide mixture (30% acrylamide: 0.8% (wt/vol) bisacrylamide) from National Diagnostics, East Riding of Yorkshire, UK. Protease Inhibitors and N-glycosidase-F were purchased from Roche Molecular Biochemicals, East Sussex, UK.
pCEP4APP695 Hygromycin resistant, APPswe Neomycin resistant. pcDNA3.1MycHis (A) ASP-1 and 2 were kindly donated by the GlaxoSmithKline laboratories, Harlow, UK. The clones were engineered in frame with the Histidine tag and the Myc epitope, which in that order are downstream of the coding sequence of the Aspartic proteases (ASP).
For the production of stable colonies HEK293 cells were transfected using pFX2 lipids as described by the manufacturers for 2.5 × 104 cells per transfection reaction. Positive transfectants were identified and selected using Geneticin. Colonies were screened by SDS-PAGE and Western Blotting analysis of cell lysates prepared by direct suspension of the cells in SDS, 0.1% (wt/vol) followed by sonication and suspension in gel loading buffer. HEK293 cells were cultured and propagated as described in Frears et al., 1999. Cell lysates for affinity- or immuno-purification were prepared by lysis of the cells with frequent agitation for 2 hrs at 4°C in phosphate buffered saline (PBS: 150 mM NaCl 2.7 mM KCl, 10 mM Na2HPO4 and 1.75 mM KH2PO4, at pH7) supplemented with protease inhibitors, 0.5 mM EDTA, 1 μM Leupeptin, 1 μM Pepstatin and 1 mM PMSF, and Triton X-100, 1% (wt/vol) and NP40, 1% (wt/vol). For affinity purification of ASP-2 EDTA and pepstatin were omitted. Prior to purification the lysate was diluted to 0.5% (wt/vol) detergent with the above buffer/inhibitor solution. Ni2+ – affinity purification was carried out as recommended by the manufacturers using a pre-charged Ni2+ column, 1 ml of settled bed resin for 50 × 106 cells with minor alterations; all binding and washes (with increasing imidazole concentrations, pH 7.4) were carried out in PBS buffer supplemented with 0.2% (wt/vol) NP40 to reduce non-specific binding. Elution was achieved with 300 mM imidazole/0.2% (wt/vol/) detergent.
Freshly purified ASP-2 was dialysed immediately into 50 mM CH3COONa, 20 mM NaCl, 0.2% (wt/vol) NP-40, at either pH 8.5 or 5, at 4°C overnight, in the presence of 1 μM Leupeptin and 1 mM PMSF. The protein solution was then incubated at room temperature for two hours before SDS-PAGE, in vitro cleavage reactions, or N-terminal sequencing. For Mass Spectrometry the proteins were first dialysed in the above buffer at pH 8.5, and then the pH was adjusted with dilute acetic acid to ph 5. For N-terminal peptide sequencing the samples were immobilised on a PVDF membrane.
Anti-CT15 (2 μg/ml) was used to immunoprecipitate (Stephens and Austen 1996) APPswe from HEK293 cell lysates/extracts prepared as above. The pellet was resuspended in reaction buffer (20 mM CH3COONa, ph 5, 50 mM NaCl, 0.2% NP40, 1 mM PMSF and 1 μM Leupeptin) prior to the addition of ASP-2 (purified and pre-incubated in reaction buffer at designated pH). Reactions were carried out for 2 hours at 37°C. Samples were analysed by SDS-PAGE on a 16.5% Tris-Tricine gel as described by the manufacturers followed by western blotting and detection with anti-CT15 IgG (0.5 μg/ml) or 6E10 (4 μg/ml).
Purified ASP-2 which had been dialysed at pH 8.5 as described above was pre-incubated in the reaction buffer titrated at the designated pH without detergent prior to absorption to a C-18 ZIP tip (Millipore) equilibrated with 0.1% (vol/vol) trifluorocetic acid, (TFA). Peptides were eluted with 50% (vol/vol) acetonitrile and analysed on an Axima-CFR KRATOS Mass Spectrometer after addition of an equal volume of 10 mg/ml α-cyano-4-hydroxycinnamic acid (Sigma) in 50% (vol/vol) acetonitrile. Angiotensin 1, the dimeric form of α-cyano-4-hydroxycinnamic acid, and neuropeptide Y were used as external calibrants spotted on an adjacent spot on the chip.
The C-terminal peptide of ASP-2 (CLRQQHDDFADDISLLK residues 482–501) was synthesised on a Milligen 9050 synthesizer using Fmoc N-terminal protection and, after deprotection and release from resin was purified by HPLC on a column of Vydac C4 with gradients of acetonitrile in 0.1% TFA. In brief, bovine thyroglobulin was activated with succinimidyl 4-(N-maleimido-methyl) cyclohexane-1-carboxylate, desalted on a Sephadex G-10 column and coupled to the HPLC-purified synthetic peptide utilising its free-cysteine as described in . The coupled peptide was used to immunise rabbits and its immunoreactivity was screened by ELISA using immobilised antigenic peptide. Reactive serum was purified on a peptide-conjugated Sepharose column, aliquoted and stored at -20°C. Pure anti-ASP-2 IgG was tested for its specificity towards ASP-2 by western blotting of extracts of recombinant clones as well as mock-transfected cells using purified IgG pre-incubated or not with excess antigenic peptide.
80% Confluent flasks of HEK293 APPswe-trasfected cells (~5 × 106 cells) were depleted of Methionine by incubation in Met-free media supplemented with FCS and Glutamax and pyruvate for 1 hr. The same media was then supplemented with [35S]-Met at 50 μCi/ml in 5 mls, and incubated for 3 hrs. Cells were first rinsed with PBS and then lysed and extracted as described earlier. The [35S]-labelled protein was immunoprecipitated as described earlier and reactions with ASP-2 were carried out. These were analysed by 7% SDS-PAGE, soaked in AmplifyTR for 15 minutes and dried before exposure to Kodak X-Omat film.
Proteins were first denatured by boiling in SDS sample buffer (0.25 M Tris-HCl, pH 8.8, 2.2% (wt/vol) SDS, 10% (v /v) Glycerol, 0.05% (wt/vol) bromophenol blue) with or without reducing agent (1% β-mercaptoethanol and/or 10 mM DTT) and electrophoresed on SDS-PAGE gels prepared as described by the manufacturers of the acrylamide solution (National Diagnostics) and run using a BioRad Mini-Gel system at 20 mA per gel. The gels were subsequently transferred onto PVDF membrane using a Biorad semi-dry blotter in 25 mM Tris, 192 mM Glycine, and 20% (vol/vol) methanol. Antibody detection was carried out by immunoblotting as described in Stephens and Austen (1996). For silver staining, gels were fixed with methanol, 40% (vol/vol) and acetic acid 10% (vol/vol), for one hour, before they were soaked for 30 min in DTT (0.5 μg/ml). The gels were rinsed twice with water and soaked for one hour in AgNO3, 0.1% (wt/vol) prior to development with a solution of Na2CO3, 3% (wt/vol) and formaldehyde, 0.0185% (vol/vol).
Cellular extracts or purified ASP-2 were first heat-denatured in the presence of 0.5% (wt/vol) SDS in PBS, pH8 and the rest of the procedure was carried out as described by the manufacturers.
β amyloid peptide
Mutations in APP in which KM is replaced by NL in Swedish family with familial Alzheimer's dementia
Dulbecco's modified eagle medium
Nonident P 40
Sodium dodecyl sulphate polyacrylamide gel electrophoresis
We thank the Medical Research Council, UK, for funding this research and the GlaxoSmithKline laboratories, Harlow, UK for their donation of the ASP-1 and ASP-2 clones.