Amyloids can be formed from many different proteins. In Alzheimer disease (AD), amyloid-β (Aβ) is the culprit, but in other diseases it can be transthyretin, serum amyloid A, prions, or any of up to about 15 different proteins that form amyloid (see the comprehensive list prepared for our recent live discussion). Familial British (FBD) and Danish (FDD) dementias, for example, are caused by mutations in the gene for the type II membrane protein BRI2. In these diseases, furin-catalyzed proteolytic cleavage of BRI2, which normally generates an 23-amino-acid peptide, instead releases a peptide that is 34 residues long. It is these longer peptides that form the ABri and ADan amyloids. The resulting plaques, neurofibrillary tangles, and rampant neurodegeneration that are characteristic of FBD/FDD bear an uncanny resemblance to the pathology seen in AD. There, the similarity between these different diseases might end. But this week a paper in press in the Journal of Biological Chemistry reveals that wild-type BRI2 not only interacts with AβPP, but that it attenuates production of Aβ.
Luciano D’Adamio and colleagues at Albert Einstein College of Medicine, New York, the New York University School of Medicine, and the Mayo Clinic in Jacksonville, Florida, made the connection when they set up a screen to find proteins that might regulate cleavage of AβPP by secretases. Notch processing, which is very similar to that of AβPP, is known to be regulated by membrane-bound proteins, prompting D’Adamio and colleagues to look for similar regulators of AβPP proteolysis.
First author Shuji Matsuda and colleagues utilized the split ubiquitin system to detect membrane-bound binding partners for human AβPP. This system brings together N- and C-terminal ends of ubiquitin when bait and prey hybrid proteins bind. The subsequent cleavage of the bait, in this case an AβPP hybrid, by a ubiquitin-specific protease releases a soluble transcription factor that then activates a reporter gene, signaling that protein-protein interaction has taken place at the membrane.
Matsuda and colleagues tested their bait in HeLa cells co-transfected with vectors harboring human brain library DNA. The two DNA clones that tested positive in the screen were found to express BRI2 and BRI3. Because of the obvious link between BRI2 and AD, the authors characterized the APP/BRI2 interaction.
Using immunoprecipitation experiments, Matsuda found that in HeLa cells, BRI2 interacts with full-length AβPP and with the β-secretase (BACE)-generated C-terminal 99 amino acids (C99) of AβPP, but not with the α-secretase-generated C83. The finding suggests that the amino acids corresponding to the first 17 residues of Aβ are essential for the interaction between the two proteins. When the authors performed similar experiments to test tissue isolated from human brain, they achieved almost identical results, suggesting that the interaction may also be important in vivo.
So what might be the consequences of these interactions? One thing Matusda and colleagues noticed in their experiments was that the addition of BRI2 always resulted in a dramatic increase in the levels of C99. This could mean that BRI2 increases BACE activity, or that BRI2 prevents further processing of C99 by γ-secretase. Another alternative would be that by attenuating α-secretase cleavage, BRI2 shifts the balance toward BACE processing. In fact, Matsuda and colleagues found that BRI2 seems to do all three. When they expressed BRI2 in cells harboring an AβPP-Gal4 reporter system that depends on γ-secretase cleavage, they found reporter activity was attenuated. They also found that BRI2 reduced the amount of Aβ40-42 released into the medium by HEK293 cells. Both these observations suggest a reduction in γ-secretase activity. Also, using pulse chase experiments, the authors found that BRI2 increases the release of sAPPβ and decreases the release of sAPPα, the soluble products of BACE and α-secretase, respectively, suggesting that BACE activity is augmented while α-secretase is attenuated.
“…However, the finding that BRI2 regulates APP processing is intriguing and prompt[s us] to speculate that altered APP processing is also a pathogenic factor in FBD and FDD,” write the authors. The speculation is backed up by recent findings from the same group that Aβ coaggregates with ADan in vascular deposits (see Holton et al., 2002). What now needs to be addressed is what impact FBD/FDD BRI2 mutations have on AβPP processing.—Tom Fagan