November 2004 Contents Archive

Cell biology of protein misfolding: The examples of Alzheimer's and Parkinson's diseases

Figure 1. A model of key events underlying the pathogenesis of Alzheimer's disease, based on available evidence. APP molecules on the plasma membrane and in intracellular vesicles such as endosomes are cleaved by -secretase (BACE) and the presenilin−-secretase complex (PS−) to liberate the A region. A portion of A peptides can oligomerize, initially intravesicularly, and be released into the interstitial fluid of brain, where soluble oligomers may diffuse into synaptic clefts and interfere with synaptic function by unknown mechanisms. A oligomers can further polymerize into insoluble amyloid fibrils that aggregate into spherical plaques, resulting in tortuosity and dysfunction of adjacent axons and dendrites. A major accompaniment of such events is the activation of kinases in the neuronal cytoplasm, leading to hyperphosphorylation of the microtubule-associated protein, tau, and its polymerization into insoluble filaments that aggregate as neurofibrillary tangles. Activated microglia and reactive astrocytes surrounding plaques participate in a local inflammatory response that may contribute to neurotoxicity. Full Figure and legend (93K)

The other major lesion is the amyloid plaque. These are extracellular deposits of insoluble, 8−10-nm amyloid fibrils that are polymers of the amyloid- protein (A)^{8, 9}. One characteristic form of amyloid plaque is referred to as the neuritic plaque, namely, an extracellular core of amyloid fibrils intimately surrounded by dystrophic dendrites and axons — some of which contain bundles of PHF — as well as by activated microglia and reactive astrocytes¹⁰. Neuritic plaques occur in variable but generally high numbers in the molecular layer of the dentate gyrus of hippocampus, the amygdala, the association cortices of the frontal, temporal and parietal lobes, and within certain deep brain nuclei that project to these regions. The microscopic appearance of these 'mature' plaques suggests a focal deposit of amyloid protein that somehow incites a highly localized multicellular reaction.

In addition to neuritic plaques, an array of morphological variants of A deposits is found in the Alzheimer brain. The advent of A immunocytochemistry led to the recognition of innumerable 'diffuse plaques' in the limbic and association cortices and even in regions, such as striatum and cerebellum, that are not usually thought of as being targeted by the Alzheimer process. These diffuse deposits appear loose and granular by light microscopy and lack dystrophic neurites and altered astrocytes and microglia. The occurrence of sometimes abundant diffuse plaques in the brains of some neurologically normal late middle-aged and elderly humans — and also in children with Down's syndrome (trisomy 21) as early as the second decade of life — has led to the hypothesis that they represent precursor lesions to the neuritic plaques, which bear fibrillar amyloid. Neuropathological analyses in humans, aged monkeys and mice transgenic for human APP have all supported this interpretation. Alzheimer brains typically exhibit a complex mixture of A deposits that occur along a continuum from faint, wispy diffuse deposits to neurite- and glia-rich amyloid plaques, and even some dense amyloid cores devoid of surrounding cytopathology.

A is a 38−43-residue proteolytic fragment of APP¹¹ that is generated normally throughout life by virtually all mammalian cells^{12, 13, 14} (Fig. 2). APP is a ubiquitous Type 1 membrane glycoprotein that occurs as a heterogeneous group of polypeptides arising from alternative splicing, N- and O-linked glycosylation, phosphorylation, sulphation, glycosaminoglycan addition and complex proteolysis. A is released from APP through sequential cleavages by -secretase (also called BACE-1), a membrane-spanning aspartyl protease with its active site in the lumen¹⁵, and -secretase, an unusual intramembrane aspartyl protease containing presenilin at its catalytic site complexed with three other membrane proteins, nicastrin, Aph-1 and Pen-2 (refs 16, 17, 18). In addition, APP undergoes both constitutive and regulated secretory cleavages by certain metalloproteases referred to as -secretases, principally believed to be ADAM 10 and ADAM 17 (refs 19, 20). The -secretase cleavage occurs primarily between residues 16 and 17 of the A region of APP, and any precursors so cleaved cannot yield A peptides. Instead, processing by -secretase followed by -secretase generates a smaller hydrophobic fragment referred to as p3 (refs 12, 21), whose normal function and role in Alzheimer's disease (if any) are unclear

Figure 2. Schematic diagrams of the -amyloid precursor protein (APP) and its principal proteolytic derivatives. The first line depicts the largest of the known APP alternative splice forms, comprising 770 amino acids. Regions of interest are indicated at their correct linear positions. A 17-residue signal peptide occurs at the N-terminus. Two alternatively spliced exons of 56 and 19 amino acids are inserted at residue 289; the first contains a serine protease inhibitor domain of the Kunitz type (KPI). A single transmembrane domain (TM) at amino acids 700−723 is indicated. The amyloid -peptide (A) includes 28 residues just outside the membrane plus the first 12−14 residues of the TM. In the second line, the sequence within APP that contains the A and TM regions is expanded. The underlined residues represent the A1−42 peptide. The purple letters below the wild-type sequence indicate the known missense mutations identified in certain families with Alzheimer's disease and/or hereditary cerebral haemorrhage with amyloidosis. Three-digit numbers represent codon numbers (APP 770 isoform). In the third line, the first arrow indicates a site (after residue 687; same as white dot in A box in first line) of cleavage by -secretase that enables secretion of the large, soluble ectodomain APPs- into the medium and retention of the 83-residue C-terminal fragment (C83) in the membrane. C83 can undergo cleavage by -secretase principally at residue 711 or residue 713 to release the p3 peptides. The fourth line depicts the alternative proteolytic cleavage after residue 671 by -secretase that results in the secretion of the slightly truncated APP5- molecule and the retention of a 99-residue C-terminal fragment in the membrane. C99 can also undergo cleavage by -secretase to release the A peptides. Cleavage of both C83 and C99 by -secretase at the epsilon () site (line 2) releases the APP intracellular domain (AICD) into the cytoplasm. The order and interdependency of the - and cleavages have not been established Full Figure and legend (20K)

-secretase processing releases the large APP ectodomain (APPs-) from the cell (Fig. 2). Several biological activities have been ascribed to this secreted derivative in vitro, including as a trophic or neuroprotective factor, a serine protease inhibitor and a cell−substrate adhesion molecule. Sequential processing by -secretase and -secretase also releases the APP intracellular domain (AICD), a portion of which can bind to the cytoplasmic adaptor protein, Fe65, and apparently help mediate aspects of nuclear signalling^{22, 23, 24}. However, there has been no consensus so far as to which transcripts may be regulated by complexes shown to contain AICD, Fe65 and the histone acetylase Tip60 (ref. 22). Genetic deletion of APP in mice produces a mild, non-lethal phenotype that includes relatively minor alterations in the adult CNS as well as defects in neurite development observed in cultured primary neurons^{25, 26}. This lack of major consequence may be explained by an apparent functional redundancy with two APP homologues, APLP-1 (neuron-specific) and APLP-2 (ubiquitous). Whether transcriptional regulation is a bona fide function of APP and its two homologues remains to be seen.

A recurring controversy in the study of proteins implicated in human neurodegenerative disorders is whether the polypeptide in question undergoes a loss or gain of function in the disease. But as the actual biochemical effects of mutations in such proteins are deciphered, attempting to answer this question seems to miss the mark. A compelling example is provided by the presenilins, missense mutations in which constitute the most common cause of dominantly inherited AD^{27, 28}. The Caenorhabditis elegans homologue of presenilin, Sel-12, was identified as a critical mediator of Notch signalling, which is required for many cell-fate decisions in metazoans²⁹. It soon became apparent that loss of presenilin function prevents the proteolytic processing of Notch that releases its intracellular domain (NICD) to signal in the nucleus^{30, 31, 32} and likewise prevents the normal -secretase processing of APP to release A³³. The discovery that all presenilins contain two aspartate residues in transmembrane domains 6 and 7 of this putatively eight-spanning transmembrane polypeptide and that these represent the active site of -secretase revealed that presenilin functioned as an unprecedented intramembrane aspartyl protease^{34, 35, 36}. In this context, AD-causing missense mutations in human presenilin-1 or presenilin-2 alter the cleavage specificity of -secretase within the APP transmembrane domain, increasing the amount of a minor cleavage at the A42 residue while slightly reducing the major cleavage at A40. This subtle perturbation does not notably alter the additional 'epsilon' cleavage (Fig. 2) of the Notch and APP transmembrane domains near their cytoplasmic faces, and therefore does not seem to interfere with NICD and AICD release. In short, the AD-causing presenilin mutations produce a subtle biochemical alteration of presenilin proteolytic function that cannot easily be construed as either loss or gain of function.

Over time, it has become clear that presenilin has many substrates, all of which have so far been shown to be single-spanning transmembrane proteins that must first undergo -secretase-mediated shedding of their ectodomains before presenilin−-secretase can cleave within the membrane³⁷. It could be reasoned that presenilin missense mutations cause Alzheimer's disease by misprocessing one or more of these other substrates. However, when the effects of the other genes implicated so far in familial AD are considered — namely APP³⁸ and the 4 allele of apolipoprotein E³⁹ — it becomes clear that this is not the case. Most of the disease-causing missense mutations in APP are clustered at, or adjacent to, the - or -secretase cleavage sites (Fig. 2). These increase A42 production roughly twofold or so, as has been shown in transfected cells, transgenic mice and the affected individuals themselves^{40, 41, 42, 43, 44}. The remaining APP mutations occur near the middle of the A region and enhance the oligomerization and fibril formation of the resultant mutant A peptides^{45, 46, 47, 48}. Moreover, duplication of the APP gene in patients with Down's syndrome invariably leads to very early formation of diffuse A deposits, followed by the age-dependent development of amyloid-rich plaques containing altered neurites, microglia and astrocytes, as well as neurofibrillary tangles⁴⁹. In the case of apoE4, inheritance of one or two alleles sharply increases the likelihood of developing AD⁵⁰ and enhances amyloid plaque levels in the brain⁵¹. The mechanism seems not to involve increased A production but rather stabilization and decreased clearance of fibrillar A deposits, as revealed by expression of human ApoE4 compared with ApoE3 in ApoE-knock-out mice⁵².

Cell biological consequences of A accumulation

Taken together, the available information about genotype-to-phenotype conversions in familial AD supports the 'A hypothesis': a chronic imbalance between the production and clearance of a small hydrophobic peptide with a tendency to misfold and aggregate leads gradually to synaptic and neuritic compromise and glial activation⁵³. But even if this general model is accepted, and it is not by some, the details of how neuronal dysfunction ensues remain unclear. Looking at the Alzheimer brain, there are many possible consequences. The dilated and tortuous dendrites and axons in the immediate vicinity of amyloid plaques suggest that altered intracellular transport and slowed nerve conduction may well contribute to compromised information processing in the limbic and association cortices⁵⁴. APP travels by fast axonal transport⁵⁵, and there is recent evidence that intact APP molecules may interact with kinesins, either directly or indirectly, and thus help to mediate the axonal transport of various molecules, including presenilin and other secretases⁵⁶. An attack by A aggregates on local neuritic integrity and function could contribute to the development of cognitive failure.

But there is a potentially more subtle effect that could induce neuronal dysfunction early in the pre-clinical phase of the disease. The concept that subtle alterations of hippocampal synaptic efficacy may be caused by diffusible oligomeric assemblies of A has received support from the significant correlations between cortical levels of soluble A (which includes soluble oligomers) and the degree of cognitive impairment^{57, 58, 59}. Moreover, in certain APP transgenic mouse models, biochemical and electrophysiological evidence of synaptic alteration can be detected before visible plaque formation but after A levels start rising steadily^{60, 61, 62, 63}. Mechanistically, cerebral microinjection of naturally secreted A oligomers can both inhibit hippocampal long-term potentiation⁶⁴ and interrupt memory of a learnt behaviour in living rats⁶⁵. Moreover, behavioural deficits in aged APP transgenic mice were reversed overnight by injection of an anti-A antibody, without any decrease in amyloid deposits⁶⁶. These and other findings suggest that soluble oligomers of A can directly compromise synaptic function and that amyloid plaques may function as reservoirs of fibrous polymers that are in equilibrium with these diffusible species.

A major adverse consequence of the cerebral accumulation of various A assembly forms is the induction of neurofibrillary changes in myriad cell bodies and processes. Precisely how this comes about is not yet clear. In general, a variety of distinct neuronal insults (including, for example, brain trauma, rare viral infections of the brain and lipid storage disorders) can lead to secondary alterations of tau that result in NFT in diverse brain diseases. The genetics of AD suggest that A accumulation is the factor that ultimately triggers NFT formation in this disorder. Regardless of the specific precipitating factor, tau hyperphosporylation can lead to the detachment of tau from microtubules and the consequent destabilization of these vital organelles^{7, 67}. This tendency towards less stable microtubules represents a loss of tau function, but at the same time, a portion of the free tau molecules can now polymerize into insoluble cytoplasmic filaments (PHF). These can seemingly confer new, toxic effects that are likely to include disturbed axonal and dendritic transport, altered neuronal metabolism and structural changes in cell bodies and neurites. That this inexorable process of neurofibrillary degeneration is lethal to both the neuron and, ultimately, the patient, has been established by the discovery of missense and splicing mutations in tau that directly cause a group of less frequent dementing illnesses, which includes frontotemporal dementia with parkinsonism^{68, 69, 70}. In these disorders, the mutant tau molecules show decreased binding to microtubules, consequent loss of function, and florid NFT formation — all in the absence of A deposits. The lack of the latter lesions, coupled with the fact that inheritance of APP mutations invariably leads to NFT composed of wild-type tau, provides compelling genetic evidence that A build-up can induce tangles, but that the converse has not been established. This sequence is further supported by studies in transgenic mouse models^{71, 72, 73}. Precisely how A accumulation in AD leads to the hyperphosporylation of wild-type tau molecules is under intensive study and may include changes in calcium homeostasis and consequent activation of certain kinases⁷⁴.

A further important consequence of progressive A accumulation is the activation of local microglia and astrocytes in the vicinity of A deposits, with attendant expression and release of various pro- and anti-inflammatory cytokines and acute-phase proteins⁷⁵. This process seems to represent a response of the brain's innate immune system to the build-up of A. The epidemiological observation that prolonged usage of non-steroidal anti-inflammatory drugs such as ibuprofen offers partial protection against the development of AD could be explained by an anti-inflammatory effect, but such drugs have also been shown to function as direct modulators of -secretase that selectively lower A42 production⁷⁶.

Although progress in elucidating the biological mechanisms of AD has led to clinical trials of agents that might slow or even prevent the disease, numerous questions remain to be answered. Precisely where in the cell A is generated and how much of it is then secreted continue to elicit experimental interest. APP traffics from the endoplasmic reticulum (ER) to the cell surface, and there is evidence for the production of A peptides as early as in the ER but also in the Golgi and trans-Golgi network (TGN), at the plasma membrane and in recycling endosomes^{21, 77, 78, 79}. It has been postulated that a portion of any A peptides generated early in the secretory pathway (particularly A42) remains inside the cell, and these could contribute to intracellular A accumulation, which has been documented in neurons in culture and in the Alzheimer brain^{79, 80}. However, a substantial portion of A peptides is secreted and thus found in extracellular fluids (including CSF and plasma¹³), and A immunohistochemistry reveals extracellular deposits (that is, diffuse plaques of A42) very early in the development of AD-type neuropathology in Down's syndrome⁴⁹.

A closely related problem is determining where in the cell the presenilin−-secretase complex operates. This unique protease processes cell-surface Notch molecules after they make contact with ligands found on adjacent cells, but much of presenilin is found in the ER and Golgi, suggesting a 'spatial paradox'⁸¹. This issue has been clarified by detecting portions of presenilin⁸² and the other three -secretase components directly on the plasma membrane, where they complex with their cognate APP substrates (C83 and C99)⁸³. It is now important to understand where in the cell the -secretase complex first encounters its substrates, how APP substrates enter into the unusual active site of this channel-like protease, and how its many substrates manage to compete for processing in an orderly fashion. From a therapeutic perspective, it might be possible to inhibit - or -secretase on the plasma membrane without requiring the inhibitor to enter the cell.

Although the gene products implicated in causing AD are expressed ubiquitously, the disorder targets the brain and, in particular, the limbic and association cortices. If it turns out that A actually drives the disease, one explanation for this anatomical predilection could come from the realization that BACE-1 is expressed almost exclusively in brain and particularly in neurons. It is unclear to what extent the unusual asymmetry of neurons, with their very long axonal processes, is the key to understanding the development of the disease. The axonal transport of APP and its processing enzymes to synaptic terminals could be critical for the initiation of the disease, or else the generation of A in neuronal cell bodies and dendrites (as well as in glia and vascular cells) could function to elevate the extracellular pool of A42 (including secreted oligomers), thus allowing its subsequent precipitation into innumerable plaques.

Finally, it may be possible to further unravel the multifaceted responses of neurons and glia to the insidious accumulation of A during aging and in AD. Although transgenic mice may allow the dissection of elements of the reaction in a controlled temporal fashion⁷³, these models do not necessarily reproduce all aspects of the human disorder (for example, frank neuronal loss). Moreover, many cellular responses probably occur virtually simultaneously in humans and in transgenic mice. However, it may not be necessary to understand all the steps of pathogenesis in exquisite detail in order to intervene successfully with disease-modifying therapies.

Dopaminergic neurons in the substantia nigra and noradrenergic neurons in the locus coeruleus develop characteristic filamentous inclusions in PD called 'Lewy bodies'. The identification of the first familial PD gene as -synculein⁸⁵ led to the recognition that the principal subunit protein of the filaments comprising Lewy bodies in idiopathic PD was wild-type -synculein^{86, 87}. Although normally a 'natively unfolded' soluble protein, -synculein forms amyloid-like filaments under certain conditions that include mitochondrial complex I deficiency⁸⁸ and other forms of oxidative stress⁸⁹. Mice lacking -synculein are resistant to the neurotoxic effects of the complex I inhibitor, MPTP⁹⁰. -synculein can also form small protofibrillar intermediates that may contribute to its cytoxicity⁹¹. Very rare missense mutations in -synculein cause a dominant PD-like syndrome. Moreover, genomic triplication of the wild-type -synculein gene can cause PD⁹², a situation remarkably analogous to the duplication of APP in Down's syndrome that invariably induces the AD phenotype.

The second — and most common — mutant gene implicated in familial PD is parkin⁹³, and various loss-of-function mutations occurring in both alleles produce an aggressive, generally early form of PD in which dopaminergic midbrain neurons are lost, but the remaining cells usually lack Lewy bodies. Parkin is a cytoplasmic neuronal protein that contains two ring-finger domains, plus a ubiquitin-like domain at its amino terminus. This structure is consistent with parkin functioning as an E3 ubiquitin ligase, and such activity has been demonstrated^{94, 95}. Several substrates for the ubiquitin-ligase function of Parkin have been suggested, but there is no agreement as to which of these are bona fide substrates in vivo, as homozygous deletion of Parkin in mice does not seem to elevate brain levels of these proteins, although it also does not produce dopaminergic neuronal loss⁹⁶. In the brains of such mice, the down-regulation of several proteins implicated in the oxidative stress process suggests that Parkin could also function in this pathway⁹⁷ and fits well with considerable epidemiological and animal modelling evidence that idiopathic PD may be triggered in some cases by chronic exposure to environmental toxins (for example, rotenone, paraquat or MPTP) that can inhibit mitochondrial complex I, promote free radical stress and induce Lewy-body-like inclusions containing -synculein in dopaminergic neurons^{98, 99}.

At this early juncture in the study of genes that predispose to PD, it is unclear whether there will be a convergence of most or all of the implicated gene products in a common pathogenic pathway, analogous to the four confirmed AD genes that all lead to increased A accumulation in the brain. Much work is now needed to determine whether the misfolding and oligomerization of -synculein has a central role in the pathogenesis of dopaminergic neuronal dysfunction analogous to that of A in the Alzheimer cascade. There are several notable differences between the two syndromes: first, intraneuronal protein aggregates characterize PD, whereas both extra- and intra-neuronal aggregates accumulate abundantly in AD; second, the characteristic protein aggregates of PD are usually (but not always) absent in cases caused by the most commonly implicated gene (parkin); third, all genes implicated so far in familial AD are transmitted as autosomal-dominant traits, whereas familial PD can arise from both dominant and recessive inheritance; fourth, the Parkinson process targets primarily subcortical neurons, whereas the AD process principally compromises neurons in limbic and association cortices; and fifth, and perhaps most important, is the growing evidence that mitochondrial dysfunction, oxidative stress and the ubiquitin−proteasome pathway are centrally implicated in PD, whereas these may function in AD but seem to be farther downstream in the cascade. Despite these distinctions, it makes abundant sense to continue to look for opportunities to apply the lessons learnt from deciphering the role of misfolded proteins in AD to the effort to understand selective neuronal loss in the second most common neurodegenerative disease.

As already stated, a common feature of slowly progressive neurodegenerative disorders is the accumulation of protein aggregates rich in -pleated sheet conformation that can confer toxic properties on cells. Clearly, a full answer to the question of why this misfolding occurs is not yet possible, just as a detailed understanding of the normal mechanisms of protein folding is not at hand. But factors that predispose certain proteins to misfolding have been recognized, based in part on emerging knowledge of the biochemistry of systemic amyloid diseases, such as primary (immunoglobulin) and secondary (serum amyloid A) amyloidosis. One such factor is the occurrence of a missense mutation in a polypeptide that enhances the likelihood that it will self-aggregate into dimers and higher-order assemblies. The mutation may directly foster homotypic protein interactions, as is the case for the intra-peptide mutations within the central hydrophobic core of A. This process of folding into a -sheet-rich conformation and aggregating helps bury otherwise exposed hydrophobic residues. Alternatively, a mutation may alter the normal properties of a protein, leading secondarily to its accumulation to levels that exceed the critical concentration required for self-association, as in the case of tau mutations that decrease the normal sequestration of tau on microtubules.

Proteolysis is another factor that can contribute to amyloidogenic potential. In the case of serum amyloid A protein, limited proteolysis creates a 76-residue fragment that is more prone to self-association than the parent molecule. In a different way, altered proteolysis of APP caused by presenilin mutations increases the A42-to-A40 ratio, and the extra two hydrophobic residues (alanine and isoleucine) notably enhance the misfolding and aggregation of the peptide. In general, high local concentrations of a protein coupled with certain biochemical conditions (for example, low pH, exposure to hydrophobic surfaces on lipids and, in the example of -synculein, oxidative stress) may favour self-association. Deficiencies in protein chaperone systems may also contribute to a tendency for a polypeptide to lose its native structure and misfold. Of special importance is time, as achieving sufficiently high concentrations, overcoming energetic barriers to the loss of native structure and gain of -sheet-rich structure, and forming a 'seed' for aggregation (for example, a dimer) all seem to require great time. The neurodegenerative diseases — and amyloidoses in general — are strongly age-dependent, even though in some cases a mutant protein has been highly expressed since before birth. Of course, a considerable part of this lag time represents the gradual accrual of large protein deposits that function as reservoirs of cytotoxic species and also the accrual of sufficient cell dysfunction and loss to cause clinically detectable symptoms.

	Top

1. Selkoe, D. J., Ihara, Y. & Salazar, F. Alzheimer's disease: insolubility of partially purified paired helical filaments in sodium dodecyl sulfate and urea. Science 215, 1243−1245 (1982). | PubMed  | ISI | ChemPort |
2. Nukina, N. & Ihara, Y. Proteolytic fragments of Alzheimer's paired helical filaments. J Biochem (Tokyo) 98, 1715−1718 (1985). | PubMed  | ChemPort |
3. Grundke-Iqbal, I. et al. Abnormal phosphorylation of the microtubule-associated protein t (tau) in Alzheimer cytoskeletal pathology. Proc. Natl Acad. Sci. USA 83, 4913−4917 (1986). | PubMed  | ChemPort |
4. Kosik, K. S., Joachim, C. L. & Selkoe, D. J. Microtubule-associated protein, tau, is a major antigenic component of paired helical filaments in Alzheimer's disease. Proc. Natl Acad. Sci. USA 83, 4044−4048 (1986). | PubMed  | ChemPort |
5. Wischik, C. M. et al. Isolation of a fragment of tau derived from the core of the paired helical filament of Alzheimer's disease. Proc. Natl Acad. Sci. USA 85, 4506−4510 (1988). | PubMed  | ChemPort |
6. Lee, V. M.-Y., Balin, B. J., Otvos, L. & Trojanowski, J. Q. A68: A major subunit of paired helical filaments and derivatized forms of normal tau. Science 251, 675−678 (1991). | PubMed  | ISI | ChemPort |
7. Lee, V. M., Goedert, M. & Trojanowski, J. Q. Neurodegenerative tauopathies. Annu. Rev. Neurosci. 24, 1121−1159 (2001). | Article | PubMed  | ISI | ChemPort |
8. Glenner, G. G. & Wong, C. W. Alzheimer's disease: Initial report of the purification and characterization of a novel cerebrovascular amyloid protein. Biochem. Biophys. Res. Commun. 120, 885−890 (1984). | PubMed  | ISI | ChemPort |
9. Masters, C. L. et al. Amyloid plaque core protein in Alzheimer disease and Down syndrome. Proc. Natl Acad. Sci. USA 82, 4245−4249 (1985). | PubMed  | ChemPort |
10. Selkoe, D. J. Cell biology of the amyloid -protein precursor and the mechanism of Alzheimer's disease. Annu. Rev. Cell Biol. 10, 373−403 (1994). | Article | PubMed  | ISI | ChemPort |
11. Kang, J. et al. The precursor of Alzheimer's disease amyloid A4 protein resembles a cell-surface receptor. Nature 325, 733−736 (1987). | Article | PubMed  | ISI | ChemPort |
12. Haass, C. et al. Amyloid -peptide is produced by cultured cells during normal metabolism. Nature 359, 322−325 (1992). | Article | PubMed  | ISI | ChemPort |
13. Seubert, P. et al. Isolation and quantitation of soluble Alzheimer's -peptide from biological fluids. Nature 359, 325−327 (1992). | Article | PubMed  | ISI | ChemPort |
14. Shoji, M. et al. Production of the Alzheimer amyloid protein by normal proteolytic processing. Science 258, 126−129 (1992). | PubMed  | ISI | ChemPort |
15. Vassar, R. & Citron, M. A-generating enzymes: recent advances in - and -secretase research. Neuron 27, 419−422 (2000). | Article | PubMed  | ISI | ChemPort |
16. Kimberly, W. T. et al. -Secretase is a membrane protein complex comprised of presenilin, nicastrin, Aph-1 and Pen-2. Proc. Natl Acad. Sci. USA 100, 6382−6387 (2003). | Article | PubMed  | ChemPort |
17. Takasugi, N. et al. The role of presenilin cofactors in the -secretase complex. Nature 422, 438−441 (2003). | Article | PubMed  | ISI | ChemPort |
18. Edbauer, D. et al. Reconstitution of -secretase activity. Nature Cell Biol. 5, 486−488 (2003). | Article | PubMed  | ISI | ChemPort |
19. Buxbaum, J. D. et al. Evidence that tumor necrosis factor converting enzyme is involved in regulated -secretase cleavage of the Alzheimer amyloid protein precursor. J. Biol. Chem. 273, 27765−27767 (1998). | Article | PubMed  | ISI | ChemPort |
20. Parkin, E. T. et al. Structure−activity relationship of hydroxamate-based inhibitors on the secretases that cleave the amyloid precursor protein, angiotensin converting enzyme, CD23, and pro-tumor necrosis factor-. Biochemistry 41, 4972−4981 (2002). | Article | PubMed  | ISI | ChemPort |
21. Haass, C., Hung, A. Y., Schlossmacher, M. G., Teplow, D. B. & Selkoe, D. J. -amyloid peptide and a 3-kDa fragment are derived by distinct cellular mechanisms. J. Biol. Chem. 268, 3021−3024 (1993). | PubMed  | ISI | ChemPort |
22. Cao, X. & Sudhof, T. C. A transcriptionally active complex of APP with Fe65 and histone acetyltransferase Tip60. Science 293, 115−120 (2001). | Article | PubMed  | ISI | ChemPort |
23. Kimberly, W. T., Zheng, J. B., Guenette, S. Y. & Selkoe, D. J. The intracellular domain of the -amyloid precursor protein is stabilized by Fe65 and translocates to the nucleus in a Notch-like manner. J. Biol. Chem. 276, 40288−40292 (2001). | PubMed  | ISI | ChemPort |
24. Cao, X. & Sudhof, T. C. Dissection of amyloid- precursor protein-dependent transcriptional transactivation. J. Biol. Chem. 279, 24601−24611 (2004). | Article | PubMed  | ISI | ChemPort |
25. Zheng, H. et al. -amyloid precursor protein-deficient mice show reactive gliosis and decreased locomotor activity. Cell 81, 525−531 (1995). | Article | PubMed  | ISI | ChemPort |
26. Perez, R. G., Zheng, H., Van der Ploeg, L. H. & Koo, E. H. The -amyloid precursor protein of Alzheimer's disease enhances neuron viability and modulates neuronal polarity. J. Neurosci. 17, 9407−9414 (1997). | PubMed  | ISI | ChemPort |
27. Sherrington, R. et al. Cloning of a novel gene bearing missense mutations in early onset familial Alzheimer disease. Nature 375, 754−760 (1995). | Article | PubMed  | ISI | ChemPort |
28. Levy-Lahad, E. et al. Candidate gene for the chromosome 1 familial Alzheimer's disease locus. Science 269, 973−977 (1995). | PubMed  | ChemPort |
29. Levitan, D. & Greenwald, I. Facilitation of lin-12-mediated signalling by sel-12, a Caenorhabditis elegans S182 Alzheimer's disease gene. Nature 377, 351−354 (1995). | Article | PubMed  | ISI | ChemPort |
30. De Strooper, B. et al. A presenilin-1-dependent -secretase-like protease mediates release of Notch intracellular domain. Nature 398, 518−522 (1999). | Article | PubMed  | ISI | ChemPort |
31. Struhl, G. & Greenwald, I. Presenilin is required for activity and nuclear access of Notch in Drosophila. Nature 398, 522−525 (1999). | Article | PubMed  | ISI | ChemPort |
32. Ye, Y., Lukinova, N. & Fortini, M. E. Neurogenic phenotypes and altered Notch processing in Drosophila Presenilin mutants. Nature 398, 525−529 (1999). | Article | PubMed  | ISI | ChemPort |
33. De Strooper, B. et al. Deficiency of presenilin-1 inhibits the normal cleavage of amyloid precursor protein. Nature 391, 387−390 (1998). | Article | PubMed  | ISI | ChemPort |
34. Wolfe, M. S. et al. Two transmembrane aspartates in presenilin-1 required for presenilin endoproteolysis and -secretase activity. Nature 398, 513−517 (1999). | Article | PubMed  | ISI | ChemPort |
35. Li, Y. -M. et al. Photoactivated -secretase inhibitors directed to the active site covalently label presenilin 1. Nature 405, 689−694 (2000). | Article | PubMed  | ISI | ChemPort |
36. Esler, W. P. et al. Transition-state analogue inhibitors of -secretase bind directly to presenilin-1. Nature Cell Biol. 2, 428−434 (2000). | Article | PubMed  | ISI | ChemPort |
37. Struhl, G. & Adachi, A. Requirements for presenilin-dependent cleavage of notch and other transmembrane proteins. Mol. Cell 6, 625−636 (2000). | Article | PubMed  | ISI | ChemPort |
38. Goate, A. et al. Segregation of a missense mutation in the amyloid precursor protein gene with familial Alzheimer's disease. Nature 349, 704−706 (1991). | Article | PubMed  | ISI |&