In Alzheimer's disease, the processing of the amyloidβ precursor protein (AβPP) to yield the amyloidogenic Aβ42 peptide is mediated by the intramembrane proteolytic activity of γ-secretase, making this protease a potential target for drug development. One impediment to such a strategy, however, is that the protease is also involved in the processing of notch, a membrane receptor and signal transduction molecule, which, when activated, regulates many intracellular pathways and has a profound impact on cell differentiation.

In yesterday's PNAS online, researchers from the University of Toronto and New York University School of Medicine, report that mutations in the γ-secretase-associated protein, nicastrin, can lead to a selective enhancement of notch processing. This finding suggests that the proteolytic cleavage of AβPP and notch may be independently regulated, and offers hope that specific inhibitors of the AβPP pathway can be developed.

First author Richard Rozmahel, working with Peter St. George-Hyslop and colleagues, discovered the selective activation of notch in presenilin 1 (PS1) null mice that had been crossed with wild-type animals. Presenilin 1 and γ-secretase are considered by many to be one and the same (see related news items 1, 2, and 3). PS1 null mice typically have severe deformities that correlate with a lack of notch processing. However, some of the offspring of the cross had a much milder phenotype and notch processing was increased to about 43 percent that of wild-type levels-notch processing in PS1 nulls is about 34 percent that of wild type. AβPP processing, on the other hand, didn't seem to be affected, as determined by western blotting and ELISA measurements of brain Aβ40 and Aβ42.

The researchers mapped the locus responsible for the rescue to a narrow region of chromosome 1 on some of the wild-type mice. This region contains two strong candidates that could be responsible for the rescue, presenilin 2 and nicastrin. DNA sequence analysis showed that the partially rescued mice hade no changes in the PS2 coding region, nor in PS2 protein levels. The authors did find two missense mutations in the nicastrin gene. One of these (phenylalanine 21 to serine) occurs in the N-terminal hydrophobic region, and could be part of a signal peptide or a membrane-associated domain, the other (isoleucine 678 to threonine) occurs within the transmembrane domain. The authors predict that either of these mutations could subtly alter interactions between nicastrin and other partners in the γ-secretase complex.—Tom Fagan


  1. Using genetic cross-breeding, Rozmahel et al. presented a novel finding that the presenilin 1 (PS1) developmental activity could be modified by genes located on the distal arm of chromosome 1. They further showed that one of the modifiers could be nicastrin, as the polymorphisms between 129 and C57BL/6J strains result in two missense substitutions in the nicastrin protein.

    It would be very interesting to see how the two missense substitutions modify the functions of nicastrin and presenilins. Since the PS1 null phenotype does not seem to be affected by genetic background (our unpublished observations), it is likely that the phenotypic differences are caused by differential modification of the PS1 hypomorphic allele used in this study. While the authors clearly show that both the Notch-severe and Notch-mild animals exhibit the same defect in AβPP g-secretase cleavage, the Notch assay and the AβPP assay were done using different tissues (fibroblasts vs. brains) and at different ages (post-natal day 6 versus adult). As embryonic expression of PS1 is a critical factor for its notch and developmental activity (Qian et al., 1998), it can be argued that the differential effect of the modifier on notch and AβPP could be due to the temporal-restricted modification of the hypomorphic PS1 allele. Therefore, absent a mechanism for the modifier, the author’s statement that “these results indicate that Notch S3-site cleavage and APP g-site cleavage are distinct presenilin-dependent processes” warrants further investigation.


    . Mutant human presenilin 1 protects presenilin 1 null mouse against embryonic lethality and elevates Abeta1-42/43 expression. Neuron. 1998 Mar;20(3):611-7. PubMed.

  2. The results of this paper are very interesting and emphasize the critical role of nicastrin in PS-dependent γ-secretase activity. The problem is that the genetic findings are interpreted without adequate consideration of the biochemistry of γ-secretase. The principal conclusion is that the γ-secretase cleavage of AβPP and the S3 cleavage of Notch are distinct PS-dependent processes. In the discussion, emphasis is placed on the different sites of cleavage: in the middle of the transmembrane domain for AβPP and near the cytosolic edge for Notch. This apparent difference, however, has been resolved by the recent findings that AβPP is also cleaved near the cytosolic edge and Notch is also cleaved in the middle of their respective transmembrane domains. In other words, the transmembrane domains of these proteins are cleaved at least twice: a γ-secretase cleavage in the middle and an S3 (also called ε) cleavage near the cytosolic edge. These cuts are all PS-dependent. Moreover, in our hands all these cuts occur even after co-immunoprecipitation using PS-specific antibodies from solublized membrane preparations. Unfortunately, the analysis of the AβPP and Notch proteolytic products in the present study is incomplete: Aβ is examined, but not the AICD; NICD is examined but not the Notch counterpart of Aβ. To conclude that the nicastrin mutations affect Notch and AβPP differentially, one must analyze both cleavage products generated from both Notch and AβPP.

    But even assuming that there is indeed differential processing, this still does not mean that the γ-secretase cleavage of AβPP and the S3 cleavage of Notch are distinct PS-dependent processes. There is a perfectly reasonable biochemical explanation that is also consistent with other biochemical observation pointed out above: These nicastrin mutations (and also some reported PS mutations) lead to alterations in the substrate binding site(s) that have differential effects on the ability of the two substrates to interact with the protease.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Presenilin—Guilty of Proteolysis by Association?
  2. "Cat"egorical Evidence for Presenilin's Proteolytic Role?
  3. Stockholm: There Is No Spatial Paradox!

Further Reading

No Available Further Reading

Primary Papers

  1. . Alleles at the Nicastrin locus modify presenilin 1- deficiency phenotype. Proc Natl Acad Sci U S A. 2002 Oct 29;99(22):14452-7. PubMed.