Current Alzheimer Research (www.bentham.org/car), 2004, 1, 1-10
Bentham Science Publishers Ltd.(www.bentham.org)

---

Protein Trafficking and Alzheimer’s Disease

Kengo Uemura, Akira Kuzuya and Shun Shimohama*

^*Department of Neurology, Graduate School of Medicine, Kyoto University, Kyoto 606-8507, Japan

*Address correspondence to this author at the Department of Neurology, Graduate School of Medicine, Kyoto University, 54 Shogoin-Kawaharacho, Sakyo-ku, Kyoto 606-8507, Japan; Tel: 81-75-751-3771; Fax: 81-75-751-3265; E-mail: i53367@sakura.kudpc.kyoto-u.ac.jp

Abstract: Mutations in presenilin 1 (PS1) cause early-onset familial Alzheimer`s disease (FAD). Although FAD accounts for less than 5% of all cases of Alzheimer`s disease (AD), extensive analyses of PS1 function have elucidated an important neuronal mechanism underling AD pathogenesis. PS1 is considered to be an essential component of g-secretase, which cleaves amyloid precursor protein (APP) at the transmembrane region and releases amyloid b (Ab) peptide. In addition to this well-documented function, a growing amount of evidence suggests that PS1 is involved in  the intracellular trafficking of selected membrane proteins (i.e. APP, nica-

strin, trkB, telencephalin). Recently, we have also shown that PS1 is involved in the trafficking of N-cadherin from the endoplasmic reticulum to the plasma membrane via the microtubule network. N-cadherin is localized at the synaptic junctional complex, providing an adhesive force across the synaptic cleft, and the its regulation is crucial for the neuron to exert its specific function, i.e. synaptic activity. In a mature neuron, polarized targeting of proteins from the cell body to the axonal and dendritic processes is essential for its proper function, especially, for the maintenance of synaptic function. Alterations in protein transport caused by a dysfunction in PS1 could lead to a disturbance in synaptic transmission and finally to neurodegeneration. This article will review the current knowledge of PS1 function in protein trafficking and discuss its potential role in AD pathogenesis.

Keywords: Alzheimer’s disease, Presenilin 1, Amyloid Precursor Protein, Amyloid b peptide, Protein Trafficking.

Introduction

Pathological features of Alzheimer’s disease (AD) are characterized by neurofibrillary tangles (NFT) and amyloid plaques. Amyloid plaques are composed of amyloid b peptide (Ab), which is derived from the sequential cleavage of amyloid precursor protein (APP), whereas NFTs are characterized by the accumulation of hyperphosphorylated tau. These deposits are thought to disrupt normal function in the specific regions of the CNS which are essential for learning and memory [1-3]. However, the mechanism involved in the generation and development of these deposits has remained unclear until three genes responsible for the inherited form of AD were discovered. Although the early onset familial form of Alzheimer’s disease (FAD) accounts for less than 5% of all AD cases, it has similar pathological features to those observed in sporadic cases. Intensive research focusing on the function of the mutant genes in FAD has helped to elucidate the pathological mechanism of AD. The first gene identified as a causative gene of FAD is APP [4, 5]. APP is normally cleaved by proteases called the a-, b-, and g- secretases (Fig. 1).

The activity of a-secretase cleaves APP in the middle of the amyloid region, thereby preventing the formation of amyloid. The a-secretase-mediated processing of APP consists of a constitutive and a regulated component that can be activated by protein kinase C [6]. Possible contributions of prohormone convertase PC7 to the constitutive a-secretase pathway have been demonstrated [7]. Two zinc-dependent metalloproteinases are shown to be involved in this type of cleavage - ADAM 10 [8] and ADAM17/TACE [9]. The first proteolytic step in the amyloidogenic processing of APP is catalysed by an aspartyl protease, b-secretase or BACE 1 [10]. This is followed by intramembranous cleavage of the APP C-terminal fragment by elusive g-secretase to yield either the Ab₄₀ or Ab₄₂ peptide. Mutations in the APP cluster near these processing sites promote the release of Ab through the acceleration of the proteolytic processing by b- and g- secretases. However, there are several reports on AD-linked APP mutations, which cause decrease rather than increase of Ab formation [11, 12]. These mutations could produce more pathogenic species of Ab, rather than increasing the overall amount of Ab production. Subsequently, other FAD-linked genes were cloned - presenilin 1 (PS1) [13], and its homologue, presenilin 2 (PS2) [14]. These are involved in the processing of APP to produce Ab. FAD-linked mutations in PS1 and PS2 yield more amyloidgenic Ab₄₂ compared to the wild-type [15, 16]. Accumulating experimental data, showing that mutations in APP and PS genes alter the APP processing in such a way as to make Ab deposition is more likely, support the idea that an increase in Ab deposition is the primary event in AD pathogenesis. This is called the amyloid hypothesis (Fig. 2), which is becoming the focus of AD research [1].

PRESENILIN 1 AS A COMPONENT OF g-SECRETASE

PS1 was cloned in 1995 by genetic linkage study and was mapped to chromosome 14q24.3. [13]. It was found to be expressed predominantly in the neuronal cells of the central nervous system [17]. PS1 is highly conserved in evolution and present in Caenorhadditis elegans (as its ortholog Sel-12), facilitating signaling mediated by the Notch/lin-12 family [18]. PS1 is a membrane protein with six to eight predicted transmembrane (TM) domains (Fig. 3) and a hydrophobic loop [19, 20]. It is localized mainly in the endoplasmic reticulum (ER) and Golgi membranes [21].

PS1 is also located at the plasma membrane [22], especially at the synaptic and epithelial cell-cell contact sites, where it associates with the cadherin/catenin adhesion complex [23, 24]. After translation, PS1 is proteolytically cleaved into 27-28 kDa N-terminal and approximately 16-17 kDa C-terminal derivatives by an unknown protease [25, 26]. These two fragments remain associated with each other, forming a functional complex [27].

FAD-linked mutations of PS1 are widely scattered along the entire coding region and the majority of them are mis-sense mutations [28]. These mutations in PS1 consistently induce an increase in the formation of Ab₄₂ [15, 16], suggesting that PS1 is involved in APP metabolism. There is an important exceptional mutation (insR352) of PS1, which is associated with fronto temporal dementia and acts as dominant negative PS1, decreasing the Ab production [29]. The essential role of PS1 in Ab formation and g-secretase activity is further supported by studies using PS1 knockout mice. PS1 knockout mice die shortly after natural birth because of defects in Notch signaling, which determines the cell fate during development [30]. Neural cells derived from PS1 knockout mouse embryos show impaired Ab production and g-secretase activity [31]. The hypothesis that PS1 itself might be the elusive g-secretase was proposed by Wolfe et al. [32], and is supported by a growing amount of evidence. Mutations in two aspartyl residues in TM6 and/or TM7 (D257A/D385A) have a dominant-negative effect on both PS1 endoproteolysis and g-secretase activity, suggesting that PS1 might be a unique aspartyl proteinase [32, 33]. Moreover, photoaffinity labeling and cross-linking study showed that PS1 fragments were directly labeled by potent g-secretase inhibitors that were designed to function as transition state analogue inhibitors directed to the active site of an aspartyl protease. These results strongly support the idea that PS1 is an aspartyl protease zymogen [34]. Although all this evidence suggests that PS1 carries the active site of g-secretase, PS1/g-secretase hypothesis is still controversial, since Ab production is detected in cells deficient in PSs [35, 36].

OTHER MEMBERS AND SUBSTRATES OF g-SECRETASE

Although plenty of evidence suggests that PS1 is an essential component of g-secretase, it may not be the only one. After proteolytic cleavage, PS1 fragments form a high-molecular-weight complex with additional molecules to exert g-secretase activity [37, 38]. Efforts to isolate the other constituents of g-secretase led to the identification of Nicastrin, which is a type1 transmembrane glycoprotein [39]. Artificial mis-sense mutations in a conserved hydrophilic domain of nicastrin increase Ab₄₂ and Ab₄₀ peptide secretion, although these mutations have not been identified in FAD patients. Moreover, suppression of nicastrin expression in C. elegans induces a Notch phenotype similar to those induced by null mutations in presenilin orthologs (Sel-12) of C.elegans, strongly suggesting that it is a functional component of g-secretase [39]. Most recently, a genetic screen of C. elegans revealed two genes, aph-1/APH-1 and pen-2/PEN-2, encoding multipass transmembrane proteins that interact strongly with sel-12/presenilin and aph-2/nicastrin. RNAi-mediated inactivation of aph-1, pen-2, or nicastrin in cultured Drosophila cells reduces g-secretase cleavage of bAPP and Notch substrates and also reduces the level of processed presenilin [40]. Another study revealed that Drosophila APH-1 (Aph-1) increases the stability of Drosophila presenilin (Psn) holoprotein in the complex, whereas depletion of PEN-2 (Pen-2) prevents endoproteolysis of presenilin (Psn) and promotes stabilization of the presenilin holoprotein. Thus, both APH-1 and PEN-2 are required for g-secretase activity. Therefore, recent studies suggest the idea that g-secretase activity resides in a complex of at least four proteins, namely PS1 (processed into N- and C- terminal fragments), nicastrin, APH-1 and PEN-2.

Recent advances in the research revealed that the g-secretase-mediated cleavage occurs in a growing number of membrane proteins. These proteins include membrane receptor protein (ErbB-4 [41], low-density lipoprotein receptor-related protein (LRP) [42]), adhesion molecules (CD44 [43], E- and N-cadherin [44], Nectin-1a [45]), ER resident proteins (Ire1a and b [46]), Notch ligands (Delta and Jagged [47, 48]), heparin sulfate glycoprotein (syndecan 3 [49]) and members of the APP family (amyloid b precursor like protein (APLP) -1 and 2 [50]). Most of these cleavages occur intramembranously or at the membrane-cytosol interface and it seems that PS1 is involved in a variety of cellular signalling mechanisms through release of the intracellular domains of these proteins, which presumably translocate to the nucleus and regulate transcription [51]. However, the further details of this will not be discussed in this manuscript.

PS1 in Membrane Protein Trafficking

Although efforts to reveal the biological functions of PS1 have focused on its role in APP metabolism (which is of most significance with respect to AD pathogenesis), surprisingly little is known about the physiological functions of PS1, APP and especially Ab. With respect to PS1’s biological functions, analogy has been made to another C.elegans ortholog, spe-4, which is an integral membrane protein residing in Golgi/ER derived organelles and required for the proper localization of macromolecules that are subject to asymmetric partitioning during spermatogenesis [52, 53]. This led to the speculation that PS1 may be involved in intracellular protein trafficking. In accordance with this finding, studies using PS1-deficient neurons demonstrated that maturation of the BDNF receptor-TrkB was severely compromised in the absence of PS1 [54]. Also, PS1 interacts with the cell adhesion molecule ICAM-5/telencephalin [55]. Deletions in PS1 caused striking alterations in the subcellular distribution of endogenous telencephalin, which accumulated in intracellular structures [55].

More recently, we demonstrated that PS1 is involved in the maturation and trafficking of N-cadherin to the plasma membrane, thus regulating cell-cell contact (Fig. 4) [56, 57]. In the human neuroblastoma cell line, SH-SY5Y, the absence of functional PS1 led to impaired maturation and aberrant subcellular localization of N-cadherin predominantly located in the ER). In neurons, N-cadherin is normally localized at the synaptic junctional complex, providing an adhesive force across the synaptic cleft [58-60]. It is involved in the control of synaptic contact strength, modulating synaptic function [61, 62]. Cadherin also regulates dendritic spine morphogenesis and related synaptic functions [63]. Therefore, regulation of N-cadherin is crucial for the neuron to exert its specific function, i.e. synaptic activity. AD is pathologically characterized by extensive synapse loss, which is strongly correlated to a decline in cognitive function [64, 65]. Therefore, the ability of PS1 to regulate the membrane trafficking of these adhesion molecules may underlie the pathogenic mechanism of synapse loss observed in AD. Conversely, trafficking of another adhesion molecule N-CAM, is not altered in PS1 FAD variants [66], supporting the idea that PS1 mediates the trafficking of selected membrane proteins. This selectivity is particularly important in mature neurons, in which polarized targeting of proteins from the cell body to the axonal and dendritic processes is essential for its proper function [67], especially the maintenance of synaptic function [68, 69]. Perturbation of this system could lead to synapse dysfunction and neurodegeneration.

PROTEIN TRAFFICKING AND Ab SYNTHESIS

It has recently been demonstrated that maturation and trafficking of another component of g-secretase, nicastrin is dependent upon PS1 [70, 71]. In the absence of PS1, nicastrin is largely retained in the ER. Once PS1 and nicastrin have become associated, they travel together towards the Golgi and perhaps to the plasma membrane [71, 72]. Conversely, PS1 deficiency enhances the maturation and trafficking of the g-secretase substrate, APP, to the plasma membrane [66, 70]. Moreover, our preliminary data suggests that a deficiency of functional PS1 causes a decrease in mature b-secretase, beta-site APP cleaving-enzyme 1 (BACE1) [10]. In addition, our preliminary data using the neuroblastoma cell line, SH-SY5Y, shows that in the absence of functional PS1, BACE1 is located predominantly in the ER, suggesting that it is unable to enter the Golgi apparatus and late compartments out of the ER. Conversely, in cells overexpressing wild-type PS1, the amount of mature BACE1 is increased compared to that in native cells. Moreover, endogenous PS1 and BACE1 co-localize in the extended neuritic processes and distal compartments in the retinoic acid-induced differentiated SH-SY5Y cells. Another group have already shown that full-length PS1 interacts directly with immature BACE1 [73]. These findings suggest that PS1 regulates BACE1 activity via direct interaction and facilitated trafficking of BACE1 to the late compartments.

These new data that PS1 is involved in the transport/trafficking of nicastrin [70, 71], APP [66, 70] and, possibly, BACE1 to ensure their maturation and precise cell compartment localization has implications concerning the mechanism of Ab formation. Investigation of APP sorting in Madin-Darby canine kidney (MDCK) cells, epithelial cells known to possess physiologically distinct apical and basolateral plasma membranes, revealed a preferential localization of APP on the basolateral cell surface [74]. Ab was also secreted basolaterally [74]. In contrast, BACE1 is mainly located apically or in neurons, targeted to the axon [75]. Neurons are highly polarized cells with axonal and dendritic compartments and elaborate targeting of both APP and b-secretase occurs in a neuron, which might regulate the amount of Ab production. Indeed, a recent report demonstrated that targeted expression of BACE in a specific cellular compartment called ‘lipid raft’ extensively up-regulated APP processing to yield Ab [76], supporting the idea that intracellular trafficking of BACE1 is the key to the regulation of Ab generation.

It has been pointed out that endogenous PS1 is most abundantly expressed in the ER, the intermediate compartment, the cis-Golgi network, and in some transport vesicles [21], whereas most of the PS-mediated g-secretase cleavage of APP occurs in the late compartments [77]. Given the data that the major pool of PS1 in the ER is not active as a g-secretase [78], transport of g-secretase and its substrate to precise subcellular compartment may be indispensable for the formation of an active g-secretase complex. Several studies suggest that there is preferential Ab₄₂ production in the early compartment of the secretory pathway [15, 16]. More recently, it was shown that the generation of Ab₄₂ may take place in the absence of presenilins in the early secretory pathway [35]. Taken together, derangement in the transport/trafficking of APP, b- and g-secretases could make neurons prone to produce more amyloidgenic Ab₄₂, which could be the molecular basis of AD pathology, resulting in the deposition of senile plaques.

APP in Axonal Transport

Despite plenty of studies regarding the cell-biological roles of APP, the real functions of APP have remained unclear. Recent research provided evidence of a novel function of APP as a membrane cargo receptor for kinesin-I, a microtubule motor protein. It was demonstrated by co-immunoprecipitation, sucrose gradient, and direct in vitro binding that the axonal transport of APP in neurons is mediated by the direct binding of the carboxy-terminus of APP to the kinesin light chain (KLC) subunit of kinesin-I [79]. This finding was subsequently extended to an in vivo study. Inhibition of Drosophila APP-like protein (APPL) expression caused axonal transport phenotypes of vesicle accumulation, ‘organelle jams’, similar to kinesin mutants. Genetic reduction of kinesin-I expression exacerbated these phenotypes [80]. Conversely, overexpression of the presumed cargo receptor APP (and APPL) also led to vesicle accumulation by depleting the cellular kinesin- I pool, which, in turn, diminished the transport of non-APP-dependent vesicles [80]. These data indicated a role for APP as a kinesin- I cargo receptors in vivo. In a further study, an axonal membrane compartment containing APP, b-secretase, PS1, the neurotrophin receptor TrkA, GAP43, and synapsin I, but not synaptophysin or synaptotagmin was identified in a mouse sciatic nerve and cortex [81]. The fast anterograde axonal transport of this compartment was shown to be mediated by APP and kinesin-I. Moreover, proteolytic processing of APP was found to occur in this axonal compartment both in vitro and in vivo, leading to the production of Ab, and liberating kinesin-I from the membrane [81]. These findings led authors to speculate that APP processing, g-secretase-mediated cleavage in particular, might serve as a cellular mechanism to detach the vesicle from the motor protein once it has arrived at the nerve terminal [82]. It was demonstrated, and generally accepted that Ab is a normal product of cellular metabolism [83]. Whether this physiological generation of Ab is directly linked to axonal transport or not, and whether APP-dependent axonal transport of vesicles is actually disturbed in AD or not, should be elucidated in the future.

PS1 in Axonal Transport

In an approach to understanding the physiological role of PS1 in normal nerve cells, there has been extensive effort to identify the molecules that interacts with PS1. Glycogen synthase kinase 3b (GSK3b) is a component of the WNT signaling pathway, which is involved in cell fate determination and has been shown to interact with PS1 [84]. In vivo substrates for GSK3b include PS1, b-catenin, tau and kinesin-I light chains (KLCs) [85, 86]. Through the phosphorylation of KLCs, active GSK3b inhibits anterograde transport in axons and reduces the amount of kinesin bound to membrane-bound organelles (MBOs) [86]. A recent study concerning the role of PS1 in axonal transport showed that GSK3b activity was increased in cells both in the presence of FAD-linked mutant PS1 and in the absence of PS1 [87]. It was also demonstrated that relative levels of KLC phosphorylation were increased, concomitant with increased GSK3b activity, whereas the amount of kinesin-I bound to MBOs was reduced. In addition, the density of synaptophysin- and syntaxin-I-containing vesicles and mitochondria was reduced in the hippocampal neurons of FAD-linked mutant PS1 knock-in mice, which is consistent with a deficit in kinesin-I-mediated fast axonal transport [87]. These findings suggest that PS1 modulates GSK3b activity and regulates kinesin-I-mediated intracellular transport. They also advocate the idea that FAD-linked mutations in PS1 may compromise neuronal function by affecting GSK-3 activity and kinesin-I-mediated transport.

It should be pointed out that GSK3b can also phosphorylate the microtubule associate protein tau to generate a precursor of NFTs [88, 89]. Phosphorylated tau has a reduced affinity for microtubules and a reduced ability to promote microtubule assembly [90, 91]. However, it has been reported that tau itself inhibits kinesin-dependent transport of peroxisomes, neurofilaments, Golgi-derived vesicules and, in particular, transport of APP into axons and dendrites, causing its accumulation in the cell body [92]. Therefore, given the recent finding that FAD-linked PS1 mutation enhances GSK3b activity [87], kinesin-I v.s. tau phosphorylation could have opposing effects, from the viewpoint of kinesin-mediated transport, because tau phosphorylation is supposed to liberate the ‘traffic jam’ caused by the inhibition of kinesin-driven transport [92].

The findings above suggest that axonal transport may be a crossroad between APP and tau pathology. By elucidating the complex regulatory mechanism of axonal transport, especially with regard to the roles of APP and tau, we might be able to understand further the molecular background of AD pathogenesis.

PS1-interacting Proteins and Intra-cellular Trafficking

It has also been demonstrated that PS1 interacts with several other proteins, which play essential roles in intracellular trafficking. For example, it has been reported that PS1 and PS2 associate with Rab 11, a small GTPase and a membrane trafficking regulators implicated in protein transport of the biosynthetic and endocytic pathway [93]. The function of Rab11 is not yet fully understood, but several studies indicate its role in trafficking and recycling of internalized proteins [94], or in protein transport from the trans-Golgi network through post-Golgi vesicles to the plasma membrane [95]. Although it was demonstrated that overexpression of Rab11 or its constitutively active mutant did not have functional influence on APP physiopathological maturation, these findings do not preclude the possibility for Rab11 to modulate other presenilin-mediated functions in human cells [96]. In addition, the N-terminus of PS1 binds to the rab GDP dissociation inhibitor (rabGDI), a regulatory factor in vesicle transport [97]. RabGDI regulates precise vesicular membrane transport by coordinating the events leading to membrane fusion [98], and PS1 deficiency leads to impaired rabGDI-membrane association [97].

Moreover, both full-length PS1 and PS2 interact with the cytoplasmic linker protein-170/Restin (CLIP170/Restin) [99, 100], which is believed to conjoin membrane organelles to microtubules [101]. Indeed, PS1 co-localizes with microtubules in the human neuroblastoma-derived cell line, SH-SY5Y (Fig. 5). The cytoskeletal association of PS1 is suggested by several other studies [102]. Disruption of the PS1/CLIP-170 complex is associated with both decreased secretion of endogenous Ab and decreased uptake of exogenous Ab from the medium. It was also suggested that PS1 may serve as an anchor for specific subcellular fractions containing amyloidogenic fragments to the microtubules via CLIP-170 [99].

Thus, PS1 seems to regulate intracellular protein trafficking by interacting with various trafficking regulators, thereby contributing to the maintenance of normal cellular function.

Concluding Remarks

In this review, the focus has been mainly on the functions of PS1, especially with regard to intracellular trafficking. PS1 selectively regulates intracellular protein transport in multiple steps in the trafficking pathway, including protein budding, sorting and maturation in the ER/Golgi compartment, axonal transport of vesicles to the distal compartment and, possibly, endocytosis and recycling of membrane proteins (Fig. 6).

These functions of PS1 may contribute to the formation and maintenance of cellular polarity in a neuron. Although experimental evidence is accumulating concerning the roles of PS1 and APP in protein trafficking, many questions remain to be answered. Among them, the protein trafficking mechanism specifically affected in AD should be elucidated. Certain brain areas involved in learning and memory are more vulnerable to neurodegeneration in AD. Do neurons in these areas have a higher transport burden? Do they have specific traffic mechanisms prone to be affected by AD pathology? Where does the `protein trafficking hypothesis` meet the `amyloid hypothesis`?

Hippocampal neurons, vulnerable to AD pathology, are highly polarized cells, in which elaborate protein trafficking in certain areas of the cells is necessary to maintain normal cellular function, whereas less polarized cerebellar cells are resistant to AD pathology. Future research in this field should reveal a new strategy for the treatment of AD.

References

[1]             Hardy J and Selkoe DJ. The amyloid hypothesis of Alzheimer's disease: progress and problems on the road to therapeutics. Science 297: 353-356 (2002).

[2]             Selkoe DJ. Translating cell biology into therapeutic advances in Alzheimer's disease. Nature 399: A23-31 (1999).

[3]             Hardy J. Amyloid, the presenilins and Alzheimer's disease. Trends Neurosci 20: 154-159 (1997).

[4]             Levy E, Carman MD, Fernandez-Madrid IJ, Power MD, Lieberburg I, van Duinen SG. et al. Mutation of the Alzheimer's disease amyloid gene in hereditary cerebral hemorrhage, Dutch type. Science 248: 1124-1126 (1990).

[5]             Van Broeckhoven C, Haan J, Bakker E, Hardy JA, Van Hul W, Wehnert A et al. Amyloid beta protein precursor gene and hereditary cerebral hemorrhage with amyloidosis (Dutch). Science 248: 1120-1122 (1990).

[6]             Buxbaum JD, Gandy SE, Cicchetti P, Ehrlich ME, Czernik AJ, Fracasso RP et al. Processing of Alzheimer beta/A4 amyloid precursor protein: modulation by agents that regulate protein phosphorylation. Proc Natl Acad Sci USA 87: 6003-6006 (1990).

[7]             Lopez-Perez E, Seidah NG and Checler F. Proprotein convertase activity contributes to the processing of the Alzheimer's beta-amyloid precursor protein in human cells: evidence for a role of the prohormone convertase PC7 in the constitutive alpha-secretase pathway. J Neurochem 73: 2056-2062 (1999).

[8]             Lammich S, Kojro E, Postina R, Gilbert S, Pfeiffer R, Jasionowski M et al. Constitutive and regulated alpha-secretase cleavage of Alzheimer's amyloid precursor protein by a disintegrin metalloprotease. Proc Natl Acad Sci USA 96: 3922-3927 (1999).

[9]             Buxbaum JD, Liu KN, Luo Y, Slack JL, Stocking KL, Peschon JJ et al. Evidence that tumor necrosis factor alpha converting enzyme is involved in regulated alpha-secretase cleavage of the Alzheimer amyloid protein precursor. J Biol Chem 273: 27765-27767 (1998).

[10]          Vassar R, Bennett BD, Babu-Khan S, Kahn S, Mendiaz EA, Denis P et al. Beta-secretase cleavage of Alzheimer's amyloid precursor protein by the transmembrane aspartic protease BACE. Science 286: 735-741 (1999).

[11]          Nilsberth C, Westlind-Danielsson A, Eckman CB, Condron MM, Axelman K, Forsell C et al. The 'Arctic' APP mutation (E693G) causes Alzheimer's disease by enhanced Abeta protofibril formation. Nat Neurosci 4: 887-893 (2001).

[12]          Ancolio K, Dumanchin C, Barelli H, Warter JM, Brice A, Campion D et al. Unusual phenotypic alteration of beta amyloid precursor protein (betaAPP) maturation by a new Val-715 --> Met betaAPP-770 mutation responsible for probable early-onset Alzheimer's disease. Proc Natl Acad Sci U S A 96: 4119-4124 (1999).

[13]          Sherrington R, Rogaev EI, Liang Y, Rogaeva EA, Levesque G, Ikeda M et al. Cloning of a gene bearing missense mutations in early-onset familial Alzheimer's disease. Nature 375: 754-760 (1995).

[14]          Rogaev EI, Sherrington R, Rogaeva EA, Levesque G, Ikeda M, Liang Y et al. Familial Alzheimer's disease in kindreds with missense mutations in a gene on chromosome 1 related to the Alzheimer's disease type 3 gene. Nature 376: 775-778 (1995).

[15]          Scheuner D, Eckman C, Jensen M, Song X, Citron M, Suzuki N et al. Secreted amyloid beta-protein similar to that in the senile plaques of Alzheimer's disease is increased in vivo by the presenilin 1 and 2 and APP mutations linked to familial Alzheimer's disease. Nat Med 2: 864-870 (1996).

[16]          Borchelt DR, Thinakaran G, Eckman CB, Lee MK, Davenport F, Ratovitsky T et al. Familial Alzheimer's disease-linked presenilin 1 variants elevate Abeta1-42/1-40 ratio in vitro and in vivo. Neuron 17: 1005-1013 (1996).

[17]          Suzuki T, Nishiyama K, Murayama S, Yamamoto A, Sato S, Kanazawa I et al. Regional and cellular presenilin 1 gene expression in human and rat tissues. Biochem Biophys Res Commun 219: 708-713 (1996).

[18]          Levitan D and Greenwald I. Facilitation of lin-12-mediated signalling by sel-12, a Caenorhabditis elegans S182 Alzheimer's disease gene. Nature 377: 351-354 (1995).

[19]          Li X and Greenwald I. Additional evidence for an eight-transmembrane-domain topology for Caenorhabditis elegans and human presenilins. Proc Natl Acad Sci USA 95: 7109-7014 (1998).

[20]          Lehmann S, Chiesa R and Harris DA. Evidence for a six-transmembrane domain structure of presenilin 1. J Biol Chem 272: 12047-12051 (1997).

[21]          Annaert WG, Levesque L, Craessaerts K, Dierinck I, Snellings G, Westaway D et al. Presenilin 1 controls gamma-secretase processing of amyloid precursor protein in pre-golgi compartments of hippocampal neurons. J Cell Biol 147: 277-294 (1999).

[22]          Ray WJ, Yao M, Mumm J, Schroeter EH, Saftig P, Wolfe M et al. Cell surface presenilin-1 participates in the gamma-secretase-like proteolysis of Notch. J Biol Chem 274: 36801-36807 (1999).

[23]          Georgakopoulos A, Marambaud P, Efthimiopoulos S, Shioi J, Cui W, Li HC et al. Presenilin-1 forms complexes with the cadherin/catenin cell-cell adhesion system and is recruited to intercellular and synaptic contacts. Mol Cell 4: 893-902 (1999).

[24]          Georgakopoulos A, Marambaud P, Friedrich Jr VL, Shioi J, Efthimiopoulos S and Robakis NK. Presenilin-1: a component of synaptic and endothelial adherens junctions. Ann N Y Acad Sci 920: 209-214 (2000).

[25]          Thinakaran G, Borchelt DR, Lee MK, Slunt HH, Spitzer L, Kim G et al. Endoproteolysis of presenilin 1 and accumulation of processed derivatives in vivo. Neuron 17: 181-190 (1996).

[26]          Podlisny MB, Citron M, Amarante P, Sherrington R, Xia W, Zhang J et al. Presenilin proteins undergo heterogeneous endoproteolysis between Thr291 and Ala299 and occur as stable N- and C-terminal fragments in normal and Alzheimer brain tissue. Neurobiol Dis 3: 325-337 (1997).

[27]          Thinakaran G, Regard JB, Bouton CM, Harris CL, Price DL, Borchelt DR et al. Stable association of presenilin derivatives and absence of presenilin interactions with APP. Neurobiol Dis 4: 438-453 (1998).

[28]          Fraser PE, Yang DS, Yu G, Levesque L, Nishimura M, Arawaka S et al. Presenilin structure, function and role in Alzheimer disease. Biochim Biophys Acta 1502: 1-15 (2000).

[29]          Amtul Z, Lewis PA, Piper S, Crook R, Baker M, Findlay K et al. A presenilin 1 mutation associated with familial frontotemporal dementia inhibits gamma-secretase cleavage of APP and notch. Neurobiol Dis 9: 269-273 (2002).

[30]          Shen J, Bronson RT, Chen DF, Xia W, Selkoe DJ and Tonegawa S. Skeletal and CNS defects in Presenilin-1-deficient mice. Cell 89: 629-639 (1997).

[31] &nb