Dynamics of metastable ß-hairpin structures in the folding nucleus of amyloid ß-protein.

The amyloid ß-protein (Aß), which is present predominately as a 40- or 42-residue peptide, is postulated to play a seminal role in the pathogenesis of Alzheimer's disease (AD). Folding of the Aß(21-30) decapeptide region is a critical step in the aggregation of Aß. We report results of constant temperature all-atom molecular dynamics simulations in explicit water of the dynamics of monomeric Aß(21-30) and its Dutch [Glu22Gln], Arctic [Glu22Gly], and Iowa [Asp23Asn] isoforms that are associated with familial forms of cerebral amyloid angiopathy and AD. The simulations revealed a variety of loop conformers that exhibited a hydrogen bond network involving the Asp23 and Ser26 amino acids. A population of conformers, not part of the loop population, was found to form metastable ß-hairpin structures with the highest probability in the Iowa mutant. At least three ß-hairpin structures were found that differed in their hydrogen bonding register, average number of backbone hydrogen bonds, and lifetimes. Analysis revealed that the Dutch mutant had the longest ß-hairpin lifetime (≥500 ns), closely followed by the Iowa mutant (≈500 ns). Aß(21-30) and the Arctic mutant had significantly lower lifetimes (≈200 ns). Hydrophobic packing of side chains was responsible for enhanced ß-hairpin lifetimes in the Dutch and Iowa mutants, whereas lifetimes in Aß(21-30) and its Arctic mutant were influenced by the backbone hydrogen bonding. The data suggest that prolonged ß-hairpin lifetimes may impact peptide pathogenicity in vivo.

Elucidation of amyloid beta-protein oligomerization mechanisms: discrete molecular dynamics study.

Oligomers of amyloid beta-protein (Abeta) play a central role in the pathology of Alzheimer's disease. Of the two predominant Abeta alloforms, Abeta(1-40) and Abeta(1-42), Abeta(1-42) is more strongly implicated in the disease. We elucidated the structural characteristics of oligomers of Abeta(1-40) and Abeta(1-42) and their Arctic mutants, [E22G]Abeta(1-40) and [E22G]Abeta(1-42). We simulated oligomer formation using discrete molecular dynamics (DMD) with a four-bead protein model, backbone hydrogen bonding, and residue-specific interactions due to effective hydropathy and charge. For all four peptides under study, we derived the characteristic oligomer size distributions that were in agreement with prior experimental findings. Unlike Abeta(1-40), Abeta(1-42) had a high propensity to form paranuclei (pentameric or hexameric) structures that could self-associate into higher-order oligomers. Neither of the Arctic mutants formed higher-order oligomers, but [E22G]Abeta(1-40) formed paranuclei with a similar propensity to that of Abeta(1-42). Whereas the best agreement with the experimental data was obtained when the charged residues were modeled as solely hydrophilic, further assembly from spherical oligomers into elongated protofibrils was induced by nonzero electrostatic interactions among the charged residues. Structural analysis revealed that the C-terminal region played a dominant role in Abeta(1-42) oligomer formation whereas Abeta(1-40) oligomerization was primarily driven by intermolecular interactions among the central hydrophobic regions. The N-terminal region A2-F4 played a prominent role in Abeta(1-40) oligomerization but did not contribute to the oligomerization of Abeta(1-42) or the Arctic mutants. The oligomer structure of both Arctic peptides resembled Abeta(1-42) more than Abeta(1-40), consistent with their potentially more toxic nature.

Effects of the Arctic (E22-->G) mutation on amyloid beta-protein folding: discrete molecular dynamics study.

The 40-42 residue amyloid beta-protein (Abeta) plays a central role in the pathogenesis of Alzheimer's disease (AD). Of the two main alloforms, Abeta40 and Abeta42, the longer Abeta42 is linked particularly strongly to AD. Despite the relatively small two amino acid length difference in primary structure, in vitro studies demonstrate that Abeta40 and Abeta42 oligomerize through distinct pathways. Recently, a discrete molecular dynamics (DMD) approach combined with a four-bead protein model recapitulated the differences in Abeta40 and Abeta42 oligomerization and led to structural predictions amenable to in vitro testing. Here, the same DMD approach is applied to elucidate folding of Abeta40, Abeta42, and two mutants, [G22]Abeta40 and [G22]Abeta42, which cause a familial ("Arctic") form of AD. The implicit solvent in the DMD approach is modeled by amino acid-specific hydropathic and electrostatic interactions. The strengths of these effective interactions are chosen to best fit the temperature dependence of the average beta-strand content in Abeta42 monomer, as determined using circular dichroism (CD) spectroscopy. In agreement with these CD data, we show that at physiological temperatures, the average beta-strand content in both alloforms increases with temperature. Our results predict that the average beta-strand propensity should decrease in both alloforms at temperatures higher than approximately 370 K. At physiological temperatures, both Abeta40 and Abeta42 adopt a collapsed-coil conformation with several short beta-strands and a small (<1%) amount of alpha-helical structure. At slightly above physiological temperature, folded Abeta42 monomers display larger amounts of beta-strand than do Abeta40 monomers. At increased temperatures, more extended conformations with a higher amount of beta-strand (approximately < 30%) structure are observed. In both alloforms, a beta-hairpin at A21-A30 is a central folding region. We observe three additional folded regions: structure 1, a beta-hairpin at V36-A42 that exists in Abeta42 but not in Abeta40; structure 2, a beta-hairpin at R5-H13 in Abeta42 but not in Abeta40; and structure 3, a beta-strand A2-F4 in Abeta40 but not Abeta42. At physiological temperatures, the Arctic mutation, E22G, disrupts contacts in the A21-A30 region of both [G22]Abeta peptides, resulting in a less stable main folding region relative to the wild type peptides. The Arctic mutation induces a significant structural change at the N-terminus of [G22]Abeta40 by preventing the formation of structure 3 observed in Abeta40 but not Abeta42, thereby reducing the structural differences between [G22]Abeta40 and [G22]Abeta42 at the N-terminus. [G22]Abeta40 is characterized by a significantly increased amount of average beta-strand relative to the other three peptides due to an induced beta-hairpin structure at R5-H13, similar to structure 2. Consequently, the N-terminal folded structure of the Arctic mutants closely resembles the N-terminal structure of Abeta42, suggesting that both Arctic Abeta peptides might assemble into structures similar to toxic Abeta42 oligomers.

Role of electrostatic interactions in amyloid beta-protein (A beta) oligomer formation: a discrete molecular dynamics study.

Pathological folding and oligomer formation of the amyloid beta-protein (A beta) are widely perceived as central to Alzheimer's disease. Experimental approaches to study A beta self-assembly provide limited information because most relevant aggregates are quasi-stable and inhomogeneous. We apply a discrete molecular dynamics approach combined with a four-bead protein model to study oligomer formation of A beta. We address the differences between the two most common A beta alloforms, A beta 40 and A beta 42, which oligomerize differently in vitro. Our previous study showed that, despite simplifications, our discrete molecular dynamics approach accounts for the experimentally observed differences between A beta 40 and A beta 42 and yields structural predictions amenable to in vitro testing. Here we study how the presence of electrostatic interactions (EIs) between pairs of charged amino acids affects A beta 40 and A beta 42 oligomer formation. Our results indicate that EIs promote formation of larger oligomers in both A beta 40 and A beta 42. Both A beta 40 and A beta 42 display a peak at trimers/tetramers, but A beta 42 displays additional peaks at nonamers and tetradecamers. EIs thus shift the oligomer size distributions to larger oligomers. Nonetheless, the A beta 40 size distribution remains unimodal, whereas the A beta 42 distribution is trimodal, as observed experimentally. We show that structural differences between A beta 40 and A beta 42 that already appear in the monomer folding, are not affected by EIs. A beta 42 folded structure is characterized by a turn in the C-terminus that is not present in A beta 40. We show that the same C-terminal region is also responsible for the strongest intermolecular contacts in A beta 42 pentamers and larger oligomers. Our results suggest that this C-terminal region plays a key role in the formation of A beta 42 oligomers and the relative importance of this region increases in the presence of EIs. These results suggest that inhibitors targeting the C-terminal region of A beta 42 oligomers may be able to prevent oligomer formation or structurally modify the assemblies to reduce their toxicity.

In silico study of amyloid beta-protein folding and oligomerization.

Experimental findings suggest that oligomeric forms of the amyloid beta protein (Abeta) play a critical role in Alzheimer's disease. Thus, elucidating their structure and the mechanisms of their formation is critical for developing therapeutic agents. We use discrete molecular dynamics simulations and a four-bead protein model to study oligomerization of two predominant alloforms, Abeta40 and Abeta42, at the atomic level. The four-bead model incorporates backbone hydrogen-bond interactions and amino acid-specific interactions mediated through hydrophobic and hydrophilic elements of the side chains. During the simulations we observe monomer folding and aggregation of monomers into oligomers of variable sizes. Abeta40 forms significantly more dimers than Abeta42, whereas pentamers are significantly more abundant in Abeta42 relative to Abeta40. Structure analysis reveals a turn centered at Gly-37-Gly-38 that is present in a folded Abeta42 monomer but not in a folded Abeta40 monomer and is associated with the first contacts that form during monomer folding. Our results suggest that this turn plays an important role in Abeta42 pentamer formation. Abeta pentamers have a globular structure comprising hydrophobic residues within the pentamer's core and hydrophilic N-terminal residues at the surface of the pentamer. The N termini of Abeta40 pentamers are more spatially restricted than Abeta42 pentamers. Abeta40 pentamers form a beta-strand structure involving Ala-2-Phe-4, which is absent in Abeta42 pentamers. These structural differences imply a different degree of hydrophobic core exposure between pentamers of the two alloforms, with the hydrophobic core of the Abeta42 pentamer being more exposed and thus more prone to form larger oligomers.

Small assemblies of unmodified amyloid beta-protein are the proximate neurotoxin in Alzheimer's disease.

Pioneering work in the 1950s by Christian Anfinsen on the folding of ribonuclease has shown that the primary structure of a protein "encodes" all of the information necessary for a nascent polypeptide to fold into its native, physiologically active, three-dimensional conformation (for his classic review, see [Science 181 (1973) 223]). In Alzheimer's disease (AD), the amyloid beta-protein (Abeta) appears to play a seminal role in neuronal injury and death. Recent data have suggested that the proximate effectors of neurotoxicity are oligomeric Abeta assemblies. A fundamental question, of relevance both to the development of therapeutic strategies for AD and to understanding basic laws of protein folding, is how Abeta assembly state correlates with biological activity. Evidence suggests, as argued by Anfinsen, that the formation of toxic Abeta structures is an intrinsic feature of the peptide's amino acid sequence-one requiring no post-translational modification or invocation of peptide-associated enzymatic activity.

Identification and characterization of key kinetic intermediates in amyloid beta-protein fibrillogenesis.

Amyloid beta-protein (Abeta) assembly into toxic oligomeric and fibrillar structures is a seminal event in Alzheimer's disease, therefore blocking this process could have significant therapeutic benefit. A rigorous mechanistic understanding of Abeta assembly would facilitate the targeting and design of fibrillogenesis inhibitors. Prior studies have shown that Abeta fibrillogenesis involves conformational changes leading to the formation of extended beta-sheets and that an alpha-helix-containing intermediate may be involved. However, the significance of this intermediate has been a matter of debate. We report here that the formation of an oligomeric, alpha-helix-containing assembly is a key step in Abeta fibrillogenesis. The generality of this phenomenon was supported by conformational studies of 18 different Abeta peptides, including wild-type Abeta(1-40) and Abeta(1-42), biologically relevant truncated and chemically modified Abeta peptides, and Abeta peptides causing familial forms of cerebral amyloid angiopathy. Without exception, fibrillogenesis of these peptides involved an oligomeric alpha-helix-containing intermediate and the kinetics of formation of the intermediate and of fibrils was temporally correlated. The kinetics varied depending on amino acid sequence and the extent of peptide N- and C-terminal truncation. The pH dependence of helix formation suggested that Asp and His exerted significant control over this process and over fibrillogenesis in general. Consistent with this idea, Abeta peptides containing Asp-->Asn or His-->Gln substitutions showed altered fibrillogenesis kinetics. These data emphasize the importance of the dynamic interplay between Abeta monomer conformation and oligomerization state in controlling fibrillogenesis kinetics.

The 'Arctic' APP mutation (E693G) causes Alzheimer's disease by enhanced Abeta protofibril formation.

Several pathogenic Alzheimer's disease (AD) mutations have been described, all of which cause increased amyloid beta-protein (Abeta) levels. Here we present studies of a pathogenic amyloid precursor protein (APP) mutation, located within the Abeta sequence at codon 693 (E693G), that causes AD in a Swedish family. Carriers of this 'Arctic' mutation showed decreased Abeta42 and Abeta40 levels in plasma. Additionally, low levels of Abeta42 were detected in conditioned media from cells transfected with APPE693G. Fibrillization studies demonstrated no difference in fibrillization rate, but Abeta with the Arctic mutation formed protofibrils at a much higher rate and in larger quantities than wild-type (wt) Abeta. The finding of increased protofibril formation and decreased Abeta plasma levels in the Arctic AD may reflect an alternative pathogenic mechanism for AD involving rapid Abeta protofibril formation leading to accelerated buildup of insoluble Abeta intra- and/or extracellularly.

Presenilin-dependent gamma-secretase processing of beta-amyloid precursor protein at a site corresponding to the S3 cleavage of Notch.

The presenilin (PS)-dependent site 3 (S3) cleavage of Notch liberates its intracellular domain (NICD), which is required for Notch signaling. The similar gamma-secretase cleavage of the beta-amyloid precursor protein (betaAPP) results in the secretion of amyloid beta-peptide (Abeta). However, little is known about the corresponding C-terminal cleavage product (CTFgamma). We have now identified CTFgamma in brain tissue, in living cells, as well as in an in vitro system. Generation of CTFgamma is facilitated by PSs, since a dominant-negative mutation of PS as well as a PS gene knock out prevents its production. Moreover, gamma-secretase inhibitors, including one that is known to bind to PS, also block CTFgamma generation. Sequence analysis revealed that CTFgamma is produced by a novel gamma-secretase cut, which occurs at a site corresponding to the S3 cleavage of Notch.

Amyloid beta-protein oligomerization: prenucleation interactions revealed by photo-induced cross-linking of unmodified proteins.

Assembly of the amyloid beta-protein (Abeta) into neurotoxic oligomers and fibrils is a seminal event in Alzheimer's disease. Understanding the earliest phases of Abeta assembly, including prenucleation and nucleation, is essential for the development of rational therapeutic strategies. We have applied a powerful new method, photoinduced cross-linking of unmodified proteins (PICUP), to the study of Abeta oligomerization. Significant advantages of this method include an extremely short reaction time, enabling the identification and quantification of short lived metastable assemblies, and the fact that no pre facto structural modification of the native peptide is required. Using PICUP, the distribution of Abeta oligomers existing prior to assembly was defined. A rapid equilibrium was observed involving monomer, dimer, trimer, and tetramer. A similar distribution was seen in studies of an unrelated amyloidogenic peptide, whereas nonamyloidogenic peptides yielded distributions indicative of a lack of monomer preassociation. These results suggest that simple nucleation-dependent polymerization models are insufficient to describe the dynamic equilibria associated with prenucleation phases of Abeta assembly.

In vitro studies of amyloid beta-protein fibril assembly and toxicity provide clues to the aetiology of Flemish variant (Ala692-->Gly) Alzheimer's disease.

In a Flemish kindred, an Ala(692)-->Gly amino acid substitution in the amyloid beta-protein precursor (AbetaPP) causes a form of early-onset Alzheimer's disease (AD) which displays prominent amyloid angiopathy and unusually large senile plaque cores. The mechanistic basis of this Flemish form of AD is unknown. Previous in vitro studies of amyloid beta-protein (Abeta) production in HEK-293 cells transfected with cDNA encoding Flemish AbetaPP have shown that full-length [Abeta(1-40)] and truncated [Abeta(5-40) and Abeta(11-40)] forms of Abeta are produced. In an effort to determine how these peptides might contribute to the pathogenesis of the Flemish disease, comparative biophysical and neurotoxicity studies were performed on wild-type and Flemish Abeta(1-40), Abeta(5-40) and Abeta(11-40). The results revealed that the Flemish amino acid substitution increased the solubility of each form of peptide, decreased the rate of formation of thioflavin-T-positive assemblies, and increased the SDS-stability of peptide oligomers. Although the kinetics of peptide assembly were altered by the Ala(21)-->Gly substitution, all three Flemish variants formed fibrils, as did the wild-type peptides. Importantly, toxicity studies using cultured primary rat cortical cells showed that the Flemish assemblies were as potent a neurotoxin as were the wild-type assemblies. Our results are consistent with a pathogenetic process in which conformational changes in Abeta induced by the Ala(21)-->Gly substitution would facilitate peptide adherence to the vascular endothelium, creating nidi for amyloid growth. Increased peptide solubility and assembly stability would favour formation of larger deposits and inhibit their elimination. In addition, increased concentrations of neurotoxic assemblies would accelerate neuronal injury and death.

A de novo designed helix-turn-helix peptide forms nontoxic amyloid fibrils.

We report here that a monomeric de novo designed alpha-helix-turn-alpha-helix peptide, alpha t alpha, when incubated at 37 degrees C in an aqueous buffer at neutral pH, forms nonbranching, protease resistant fibrils that are 6-10 nm in diameter. These fibrils are rich in beta-sheet and bind the amyloidophilic dye Congo red. alpha t alpha fibrils thus display the morphologic, structural, and tinctorial properties of authentic amyloid fibrils. Surprisingly, unlike fibrils formed by peptides such as the amyloid beta-protein or the islet amyloid polypeptide, alpha t alpha fibrils were not toxic to cultured rat primary cortical neurons or PC12 cells. These results suggest that the potential to form fibrils under physiologic conditions is not limited to those proteins associated with amyloidoses and that fibril formation alone is not predictive of cytotoxic activity.

An improved method of preparing the amyloid beta-protein for fibrillogenesis and neurotoxicity experiments.

Synthetic amyloid beta-protein (A beta) is used widely to study fibril formation and the physiologic effects of low molecular weight and fibrillar forms of the peptide on cells in culture or in experimental animals. Not infrequently, conflicting results have arisen in these studies, in part due to variation in the starting conformation and assembly state of A beta. To avoid these problems, we sought a simple, reliable means of preparing A beta for experimental use. We found that solvation of synthetic peptide with sodium hydroxide (A beta x NaOH), followed by lyophilization, produced stocks with superior solubility and fibrillogenesis characteristics. Solubilization of the pretreated material with neutral buffers resulted in a pH transition from approximately 10.5 to neutral, avoiding the isoelectric point of A beta (pI approximately 5.5), at which A beta precipitation and aggregation propensity are maximal. Relative to trifluoroacetate (A beta x TFA) or hydrochloric acid (A beta x HCl) salts of A beta, yields of "low molecular weight A beta" (monomers and/or dimers) were improved significantly by NaOH pretreatment. Time-dependent changes in circular dichroism spectra and Congo red dye-binding showed that A beta x NaOH formed fibrils more readily than did the other A beta preparations and that these fibrils were equally neurotoxic. NaOH pretreatment thus offers advantages for the preparation of A beta for biophysical and physiologic studies.

A furin-like convertase mediates propeptide cleavage of BACE, the Alzheimer's beta -secretase.

The novel transmembrane aspartic protease BACE (for Beta-site APP Cleaving Enzyme) is the beta-secretase that cleaves amyloid precursor protein to initiate beta-amyloid formation. As such, BACE is a prime therapeutic target for the treatment of Alzheimer's disease. BACE, like other aspartic proteases, has a propeptide domain that is removed to form the mature enzyme. BACE propeptide cleavage occurs at the sequence RLPR downward arrowE, a potential furin recognition motif. Here, we explore the role of furin in BACE propeptide domain processing. BACE propeptide cleavage in cells does not appear to be autocatalytic, since an inactive D93A mutant of BACE is still cleaved appropriately. BACE and furin co-localize within the Golgi apparatus, and propeptide cleavage is inhibited by brefeldin A and monensin, drugs that disrupt trafficking through the Golgi. Treatment of cells with the calcium ionophore, leading to inhibition of calcium-dependent proteases including furin, or transfection with the alpha(1)-antitrypsin variant alpha(1)-PDX, a potent furin inhibitor, dramatically reduces cleavage of the BACE propeptide. Moreover, the BACE propeptide is not processed in the furin-deficient LoVo cell line; however, processing is restored upon furin transfection. Finally, in vitro digestion of recombinant soluble BACE with recombinant furin results in complete cleavage only at the established E46 site. Taken together, our results strongly suggest that furin, or a furin-like proprotein convertase, is responsible for cleaving the BACE propeptide domain to form the mature enzyme.

Separation of presenilin function in amyloid beta-peptide generation and endoproteolysis of Notch.

Most of the genetically inherited Alzheimer's disease cases are caused by mutations in the presenilin genes, PS1 and PS2. PS mutations result in the enhanced production of the highly amyloidogenic 42/43 amino acid variant of amyloid beta-peptide (Abeta). We have introduced arbitrary mutations at position 286 of PS1, where a naturally occurring PS1 mutation has been described (L286V). Introduction of charged amino acids (L286E or L286R) resulted in an increase of Abeta42/43 production, which reached almost twice the level of the naturally occurring PS1 mutation. Although pathological Abeta production was increased, endoproteolysis of Notch and nuclear transport of its cytoplasmic domain was significantly inhibited. These results demonstrate that the biological function of PS proteins in the endoproteolysis of beta-amyloid precursor protein and Notch can be separated.

Monitoring protein assembly using quasielastic light scattering spectroscopy.

This article discussed the principles and practice of QLS with respect to protein assembly reactions. Particles undergoing Brownian motion in solution produce fluctuations in scattered light intensity. We have described how the temporal correlation function of these fluctuations can be measured and how mathematical analysis of the correlation function provides information about the distribution of diffusion coefficients of the particles. We have explained that deconvolution of the correlation function is an "ill-posed" problem and therefore that careful attention must be paid to the assumptions incorporated into data analysis procedures. We have shown how the Stokes-Einstein relationship can be used to convert distributions of diffusion coefficients into distributions of particle size. In the case of fibrillar polymers, this process allows direct determination of fibril length, enabling nucleation and elongation rates to be calculated. Finally, we have used examples from studies of A beta fibrillogenesis to illustrate the power these quantitative capabilities provide for understanding the molecular mechanisms of the fibrillogenesis reaction and for guiding the development of fibrillogenesis inhibitors.

Protofibrillar intermediates of amyloid beta-protein induce acute electrophysiological changes and progressive neurotoxicity in cortical neurons.

Alzheimer's disease (AD) is a progressive neurodegenerative disorder that is thought to be caused in part by the age-related accumulation of amyloid beta-protein (Abeta). The presence of neuritic plaques containing abundant Abeta-derived amyloid fibrils in AD brain tissue supports the concept that fibril accumulation per se underlies neuronal dysfunction in AD. Recent observations have begun to challenge this assumption by suggesting that earlier Abeta assemblies formed during the process of fibrillogenesis may also play a role in AD pathogenesis. Here, we present the novel finding that protofibrils (PF), metastable intermediates in amyloid fibril formation, can alter the electrical activity of neurons and cause neuronal loss. Both low molecular weight Abeta (LMW Abeta) and PF reproducibly induced toxicity in mixed brain cultures in a time- and concentration-dependent manner. No increase in fibril formation during the course of the experiments was observed by either Congo red binding or electron microscopy, suggesting that the neurotoxicity of LMW Abeta and PF cannot be explained by conversion to fibrils. Importantly, protofibrils, but not LMW Abeta, produced a rapid increase in EPSPs, action potentials, and membrane depolarizations. These data suggest that PF have inherent biological activity similar to that of mature fibrils. Our results raise the possibility that the preclinical and early clinical progression of AD is driven in part by the accumulation of specific Abeta assembly intermediates formed during the process of fibrillogenesis.

Beta-secretase cleavage of Alzheimer's amyloid precursor protein by the transmembrane aspartic protease BACE.

Cerebral deposition of amyloid beta peptide (Abeta) is an early and critical feature of Alzheimer's disease. Abeta generation depends on proteolytic cleavage of the amyloid precursor protein (APP) by two unknown proteases: beta-secretase and gamma-secretase. These proteases are prime therapeutic targets. A transmembrane aspartic protease with all the known characteristics of beta-secretase was cloned and characterized. Overexpression of this protease, termed BACE (for beta-site APP-cleaving enzyme) increased the amount of beta-secretase cleavage products, and these were cleaved exactly and only at known beta-secretase positions. Antisense inhibition of endogenous BACE messenger RNA decreased the amount of beta-secretase cleavage products, and purified BACE protein cleaved APP-derived substrates with the same sequence specificity as beta-secretase. Finally, the expression pattern and subcellular localization of BACE were consistent with that expected for beta-secretase. Future development of BACE inhibitors may prove beneficial for the treatment of Alzheimer's disease.