Presenilin Roles in APP Processing

By Michele Arduengo
Promega Corporation

One of the hallmarks of AD pathology is the appearance of amyloid plaques in the brains of affected individuals. These plaques are extracellular deposits of a variety of proteins, in particular the Ab_1-42(43) fragment. This secreted polypeptide is produced by cleavage of the Amyloid Precursor Protein (APP; for review of this process see 25). Two pathways exist in cells through which APP can be processed, the alpha pathway, which does not produce the Ab fragment, and the beta pathway, which does. Two enzyme activities are involved in the beta pathway processing of APP. An activity called b-secretase cleaves APP outside of the membrane region between residues 671 and 672 producing the C99 fragment of APP; the enzyme responsible for this activity has recently been isolated (40-42). A second activity, called g-secretase is responsible for cleaving APP within the membrane between residues 712-713, producing the secreted Ab_1-40 peptide fragment. This fragment does not seem to self-aggregate, and is found in the tissues and plasma of unaffected individuals. However, when the g-secretase activity cleaves after residue 714, a longer Ab peptide fragment results (Ab_1-42(43)). These long peptide fragments are found in the amyloid plaques and tend to self aggregate (are amyloidogenic). Thus the placement of the g-secretase activity is critical for the production of the amyloidogenic Ab_1-42(43) fragment. The enzyme responsible for g-secretase activity has not been definitively described.

In neurons, the presenilins appear to form part of the macromolecular complex that is associated with the g-secretase activity. Several studies have implicated the presenilins in the g-secretase processing of APP. First, plasma from individuals with PS mutations contains more of the amyloidgenic Ab fragment than plasma from unaffected individuals (26). Additionally brain tissues from AD patients who have PS mutations contain increased Ab_1-42(43) fragments in the amyloid deposits (27). Transgenic animals that over express dominant negative mutations in PS make increased amounts of Ab_1-42(43) (28); however transgenic mice that lack PS1 activity (null mutants) show reduced g-secretase processing of APP (29,30). More detailed study of PS1 in mice has shown that two aspartate residues in PS1 are critical for g-secretase activity, and that CHO cells transfected with PS1 containing alanine substitutions at these positions produce increased amounts of the C83 (alpha APP pathway intermediate product) and C99 (beta APP pathway intermediate product), indicating that the a- and b-secretase activities are intact, but that the g-secretase activity is disrupted (31).

Additional support for the role of the PS1 in g-secretase processing comes from studies of Notch protein processing. Like APP, Notch is a Type I integral membrane protein. It is cleaved in the Golgi into two fragments that associate to form the functional plasma membrane receptor. Notch is involved in critical cell fate decisions during development, and Notch signaling relies upon ligand-induced cleavage of Notch within the transmembrane domain to produce the Notch intracellular domain (NCID). The NCID translocates to the nucleus and activates transcription of target genes. Researchers have used g-secretase inhibitors to show that the processing of the intracellular domain of Notch is disrupted by the same protease inhibitors that block g-secretase processing of APP (30). The influence of presenilin mutations on Notch signalling  and processing in flies and worms coupled with the sensitivity of Notch processing to g-secretase inhibitors has led to the the conclusion that presenilins are involved in a Notch processing event that is similar to APP processing. Another line of support for the involvement of presenilins in g-secretase activity is that photo-activated g-secretase inhibitors specifically label human PS1 and PS2 in vitro (32).

In mice, Drosophila, and C. elegans, PS1^-/- embryos show defects in developmental events that are mediated by Notch (21,22, 24, 33). Studies of PS1^-/- mice show that Notch protein is translated, cleaved to its mature form in the Golgi (30), so PS mutations do not appear to affect the synthesis and localization of Notch to the plasma membrane. To investigate whether the ligand-induced processing of Notch to release the intracellular domain was dependent on PS1, researchers studied the localization of two different Notch constructs. One construct, mNotchDE, codes only for the transmembrane and intracellular domain of the protein (the ligand binding domain is absent) and is cleaved to release the intracellular domain. The second construct, NICV, encodes only the intracellular domain of Notch. With the NICV, the signaling domain of Notch, NCID was translocated to the nucleus with equal efficiency both PS1^-/- cells and in wildtype cells (30). These results are consistent with the observation that sel-12 mutations do not interfere with LIN-12 signaling if only the LIN-12 intracellular domain is expressed (23).

Pulse-chase experiments to analyze the production of NICD from mNotchDE, showed that NICD was greatly reduced in PS1^-/- neurons (30). A similar set of experiments in Drosophila embryos also showed that Notch constructs that encoded either wildtype Notch or constitutively activated membrane bound Notch, did not produce a fragment that was localized to the nucleus in PS null embryos. The nuclear localization of the protein produced from construct that contained only the intracellular domain was unaffected by null PS mutations (24). In contrast however, another group reports that a Notch-derived construct similar to the mNotchDE construct can signal in PS null Drosophila (33).

These studies indicate that presenilins are involved in the processing of APP and Notch. Additional studies have shown that PS1 modulates the processing of another protein, Ire1 (34,35). However, the precise mechanism by which the presenilins regulate those processing events remains to be delineated. One major caveat of any theory suggesting direct presenilin involvement in the g-secretase processing events of APP or Notch is that these events are proposed to occur at the plasma membrane. Most studies localize presenilins to the Golgi/ER or other internal membranes in cells and not the plasma membrane. However a recent study reporting that presenilins localize to endocytic compartments suggests that processing events of APP and Notch could take place in these compartments (14).

Site-directed mutagenesis of transmembrane aspartate residues in the presenilins reduced g-secretase activity, but these changes also affected the interaction of other proteins with PS1 (36). This indicates that the presenilins may be necessary but not sufficient for the g-secretase activity. A second line of evidence for this suggestion comes from the analysis of the effect of the APOE e4 allele on the development of AD. The presence of this allele influences the age of onset of AD caused by a particular APP mutation that is thought to directly affect g-secretase processing of APP. However, the e4 allele does not influence age of onset in AD caused by presenilin mutations (37,38). These observations indicate that the actual processing of APP and the role of the presenilins in that processing are genetically separable events.

A recent paper describes a gene encoding a novel Type I transmembrane protein, Nicastrin, that could be a g-secretase partner for the presenilins (4). Nicastrin interacts with both PS1 and PS2 in vitro and is highly conserved in multicellular organisms ranging from C. elegans to Arabidopsis. Nicastrin in worms was shown to be identical to the C. elegans aph-2 gene (39), a gene that influences signaling in notch/glp-1 pathways. The authors propose that Nicastrin is an integral part of a presenilin “secretosome” and either enhances binding of substrates such as C99-bAPP to the presenilins or that it regulates presenilin catalytic activity in some way (4).

This review has summarized information on the role of the presenilins in APP and Notch processing and intracellular trafficking. Studies on the presenilins and their mutant phenotypes have also suggested potential roles in apoptosis, cytoskeleton defects, and other cellular processes. A precise understanding of the function of the presenilins will depend on a more thorough understanding of all the interacting proteins as well as more sensitive means and technologies for investigating fundamental biological questions of development.