Structural aspects of the plasmalemmal Amyloid Precursor Protein cluster

Abstract

Alzheimer’s Disease (AD) is a major socio-economic threat to health care systems and society in general. Involvement of the amyloid-beta peptide (Aβ) and its parent, the amyloid precursor protein (APP), in AD seemed obvious since the first time it was described by Alois Alzheimer in 1906 and is corroborated by ample genetic, anatomical and pathophysiological evidence. Unfortunately, to date all efforts to cure, slow or even effectively treat AD have failed, indicating that there is still plenty to be learned about its initiation and progression. Two of the most obvious unknowns that are imperative to resolve are the biological function of APP and its cleavage products as well as the underlying molecular mechanisms that coordinate whether APP undergoes β-secretase-mediated amyloidogenic or α-secretase-mediated non-amyloidogenic processing. Previous publications show that at least part of the APP is present in the plasma membranes of many cell types as higher-order oligomers and that plasmalemmal clustering of APP is a primary cause for amyloidogenic processing. Whether APP oligomerization itself can prevent non-amyloidogenic cleavage simply by sterically prohibiting α-secretases access to the molecules in the center of the cluster or APP is able to evade α-processing by other means and oligomerization and subsequent amyloidogenic processing than merely take place as a secondary effect remains to be uncovered. To answer this question, we set out to elucidate the molecular density of the plasmalemmal APP cluster in a relevant human neuronal cell model, the SH-SY5Y cell line. By employing an elaborate combination of various biochemical techniques and state-of-the-art microscopy, it was found that the cellular copy number of APP is about 63,900 molecules, of which 17.5% or about 11,200 reside in the plasma membrane. The surface area of SH-SY5Y cells was determined to be 1,244 μm², giving a plasmalemmal molecular density of 9 molecules per μm². Around 34% of the detected APP signals constituted monomeric entities, revealing that over 98% of APP molecules had to be oligomerized. Finally, it could be demonstrated that most clusters have a diameter of between 65 and 85 nm and typically contain between 21 and 31 molecules, with considerably bigger entities of well over 100 nm diameter and more than 50 molecules being detected as well. This means that each molecule occupies an area with a radius of just over 7 nm within the cluster, assuming they are distributed evenly over the cluster area. However, it is plausible that the molecules in the center of the complex are in closer proximity to each other than the ones on the periphery and it has been implied that parts of APP curl back onto the plasma membrane. Besides, it is likely that the found molecular density is an underestimate and that other proteins are part of the cluster complex as well. Taken together, molecular crowding within an APP cluster is presumably substantially higher than our results alone suggest, making it increasingly reasonable that α-secretases are indeed unable to approach APP molecules residing within such a cluster. <br /> Furthermore, preceding research has shown that the tiny region consisting of the first 5 amino acids of the Aβ domain is essential for efficient cluster formation of full-length APP. This is extremely interesting, as the only mutation of APP that has been found to be protective against AD, the Icelandic or A673T variant, occurs on the second amino acid of this region. To probe whether its safeguarding effect is linked to the plasmalemmal oligomerization dynamics of APP, the plasma membrane clusters of APP harboring this mutation were characterized and compared to wild-type clusters. We could show that the A673T mutation results in less compact clusters and a more homogenous distribution in the plasma membrane, as confirmed by both cluster size and relative standard deviation measurements, respectively. Moreover, the plasmalemmal lateral mobility of the full-length proteins was significantly increased for the A673T mutant. A mild but nonsignificant trend towards a decreased colocalization with clathrin-related structures was observed, indicating that the protective variant might also be less prone to be reinternalized into the endocytic sorting machinery. Finally, it could be demonstrated that the A673T mutation is indeed more efficiently non-amyloidogenically processed than the wild-type protein, exemplifying that a decreased tendency to oligomerize may be a mechanism by which it could exert its protective influence. By extent, these results also reinforce the hypothesis that APP’s plasmalemmal clustering regulates more than just clathrin-mediated endocytosis. Our work revealed a novel potential mechanism by which protection against amyloidogenic processing can be granted should it be feasible to address pharmacologically in a specific manner, which could possibly have far reaching consequences in the battle against AD.