William of Occam contended that the simplest explanation is usually the best. Occam’s razor, as the maxim is now known, may not cut it every time, but as applied to γ-secretase—the proteolytic complex that slices and dices amyloid precursor protein (APP)—the fourteenth-century English friar and logician was presciently sharp. One of the last steps in the maturation of the secretase is the proteolytic cleavage of presenilin (PS) aspartyl protease. The simplest explanation for how this comes about is that PS does the job itself, and some researchers believe this is the case. However, there is also evidence that a separate protease is involved. The first, aka autoproteolysis, camp just got a boost. In the June 9 Journal of Neuroscience, researchers report that PS proteolysis proceeds in a stepwise fashion with blocks of three amino acids being nibbled away at a time. Interestingly, this is exactly how APP is processed, suggesting that PS and APP are cleaved by the same protease—presenilin.

Researchers led by Christian Haass at the University of Munich, Germany, discovered the stepwise cleavage with some twenty-first-century logic that Occam might have appreciated. Endoproteolysis of presenilin occurs in a cytoplasmic loop that stretches between transmembrane domains 6 and 7 of the protein. Processing splits the protease into two large, membrane-bound polypeptides. First author Akio Fukumori and colleagues realized that if they added a second protease site downstream of the usual break point, they could generate small polypeptides that could be easily analyzed by mass spectrometry (MS), uncovering the exact site, or sites, where normal presenilin cleavage takes place. Using amino acid substitution, they created a tobacco etch virus (TEV) protease site at the C-terminal end of the cytoplasmic loop, and after letting γ-secretase mature in HEK293 cells, they added TEV. The list of polypeptides they detected by MS was revealing.

The researchers found that endoproteolysis generates a major C-terminal protein fragment beginning with alanine 299, which is in agreement with previous findings. But they also found several longer, albeit less common, C-terminal fragments that start with amino acids as far upstream as methionine 292. Their presence suggests that presenilin endoproteolysis does not proceed by a simple exact cut, but results from successive cleavages, as in the case of APP processing. Several cleavage sites exist in APP (see Zhao et al., 2004) and Yasuo Ihara and colleagues at Doshisha University, Kizugawa, Japan, showed that γ-secretase cuts APP in a stepwise fashion, successively removing three-to-four amino acid peptides (see Takami et al., 2009).

If presenilin processing does proceed in an autocatalytic fashion, then one might expect that familial AD PS mutations, which alter the proteolytic processing of APP, would also alter PS endoproteolysis. In fact, this is exactly what the researchers found. Alanine 299 fragments still dominated the mix of peptides generated from FAD PS mutants, but the abundance of fragments starting with valine 296 and valine 293 grew when the most aggressive PS mutations were introduced. The analysis indicates that PS, like APP, is cleaved in successive steps, each removing three amino acids, strengthening the case that presenilin cleavage is autocatalytic. The researchers even found an explanation for one of the best arguments against autocatalysis. Mutations at glycine 384 in PS1 abolish γ-secretase activity but have no effect on PS1 cleavage. That alone would suggest a different protease is responsible for PS1 processing. But Fukumori and colleagues found that mutations at glycine 384 dramatically alter PS1 cleavage. Instead of the protein being fully processed to leave alanine 299 in the N-terminus of transmembrane loop 7, processing stops at either amino acid 292 or 296, rendering it a poor γ-secretase. The authors conclude that PS cleavage is autocatalytic, and furthermore, that stepwise endoproteolysis may be a common mechanism for processing intramembrane proteins.—Tom Fagan


  1. I recommend this paper. It is very nice work, clever use of molecular biology and mass spec, and very well-written paper. The work is for me the first that really convincingly demonstrates that presenilin gets cleaved by autocatalysis.

    View all comments by Bart De Strooper
  2. This is a solid study that provides important confirmation that presenilin is a zymogen that cuts itself (i.e., it is its own “presenilinase”) and that autoproteolysis occurs stepwise, every ~3 amino acids, similar to what is seen with γ-secretase substrates (albeit in the opposite direction, from N- to C-terminus, which is mechanistically puzzling). The finding that FAD mutations in PS1 can change the proportion of the various N-terminal variants of the PS1 C-terminal fragment subunit is a key piece of evidence, as is the observation that aspartate-scanning around the PS1 cleavage site can result in substantial increases in the putative proteolytic intermediates.

    The interpretation of the experiments resulted in a model for presenilin activation (Figure 7), in which the hydrophobic portion of the large loop (where the PS1 endoproteolytic cleavage site resides) is bound near the active site, with cleavage, deletion, or mutation of this region, resulting in active γ-secretase that allows substrate lateral entry. The idea of autoproteolysis had been originally put forward in Wolfe et al., 1999, in which it was demonstrated that the two conserved transmembrane aspartates in PS1 were absolutely required for PS1 endoproteolysis. In a follow-up Perspective article in Biochemistry several months later, we put forward the notion that cleavage (or deletion in the case of the ΔE9 PS1 FAD mutant) of the hydrophobic portion of the loop results in active γ-secretase that allows substrate lateral entry (see Wolfe et al., 1999; see especially Figure 2). It is gratifying to see that the new results by Fukumori and colleagues confirm these ideas.


    . Two transmembrane aspartates in presenilin-1 required for presenilin endoproteolysis and gamma-secretase activity. Nature. 1999 Apr 8;398(6727):513-7. PubMed.

    . Are presenilins intramembrane-cleaving proteases? Implications for the molecular mechanism of Alzheimer's disease. Biochemistry. 1999 Aug 31;38(35):11223-30. PubMed.

    View all comments by Michael Wolfe

Make a Comment

To make a comment you must login or register.


Paper Citations

  1. . Identification of a new presenilin-dependent zeta-cleavage site within the transmembrane domain of amyloid precursor protein. J Biol Chem. 2004 Dec 3;279(49):50647-50. PubMed.
  2. . gamma-Secretase: successive tripeptide and tetrapeptide release from the transmembrane domain of beta-carboxyl terminal fragment. J Neurosci. 2009 Oct 14;29(41):13042-52. PubMed.

Further Reading

No Available Further Reading

Primary Papers

  1. . Three-amino acid spacing of presenilin endoproteolysis suggests a general stepwise cleavage of gamma-secretase-mediated intramembrane proteolysis. J Neurosci. 2010 Jun 9;30(23):7853-62. PubMed.