< Terug naar vorige pagina

Publicatie

Visualizing γ-secretase at single-molecule resolution

Boek - Dissertatie

In the Alzheimer's disease (AD) field, the amyloid hypothesis remains the dominant hypothesis wherein an overproduction and/or clearance defect of toxic, aggregation-prone amyloid peptides is considered a major perpetrator of neurodegeneration and cognitive decline. Amyloid-β peptides are the result of a sequential cleavage of the amyloid precursor protein (APP).This amyloidogenic process starts when APP encounters Bace1, which sheds the extracellular domain of APP giving rise to membrane-bound APP C-terminal fragment (CTF), a direct substrate for γ-secretase proteolysis. Alternatively, APP can also undergo a non-amyloidogenic processing by ADAM10 and γ-secretase, releasing shorter non-amyloidogenic peptides. Recently, it was suggested that γ-secretase forms pre-existing complexes with the respective sheddases. However, these were biochemical studies, so whether they actually exist in cells is unclear.γ-secretase consists of four transmembrane proteins including nicastrin (NCT), PEN2, APH1 and presenilin (PSEN), the last being the catalytic subunit. Active γ-secretase complexes reside in post-Golgi compartments with PSEN1-complexes being distributed at the cell surface and endosomal compartments while PSEN2-complexes are restricted to late endosomes/lysosomes.Interestingly, mutations in PSENs resulting in familial early onset AD (FAD), appear to affect this processivity increasing the production of relatively longer/more toxic Aβ peptides.In the past years, a 3.5Å atomic resolution of γ-secretase structure was resolved revealing three conformational states and tight association of phosphatidylcholine molecules. More recently, the 3D structure of the complex including its substrates APP and Notch provided insight in how and where γ-secretase processing occurs. Based on this, computational modelling studies predicted conformational changes dependent on the lipid microenvironment, including cholesterol.Moreover, lipidomic studies on γ-secretase showed high sensitivity of the complex to lipid changes. Overall, γ-secretase structure is very dynamic and association to cholesterol or other lipid molecules determine changes in structure. Therefore, alterations to lipid environment or mutations that change affinity to certain lipids could therefore contribute as well to amyloidogenesis.Amyloidogenic processing has been proposed to occur preferentially in cholesterol and sphingomyelin rich nanodomains or lipid rafts. These are considered signaling platforms that include or exclude proteins based on their biophysical characteristics. Historically, lipid rafts were defined by biochemical fractionation with detergents, however, this approach is prone to induce association artifacts. As lipid rafts are also too small to be resolved by confocal microscopy, advanced microscopy techniques including single molecule microscopy have madepossible to analyze in situ lipid rafts.In parallel, microscopy studies on actin gave rise to an alternative mechanism to explain the heterogeneous distribution of proteins and lipids, the 'picket-fence' hypothesis. Here, actin forms an intracellular mesh that can hinder the movement of proteins and direct their movement to specific locations. To be effective to all proteins and lipids across the bilayer, this actin 'fence' would require 'pickets' or transmembrane proteins associated to actin that help in the movement of proteins and lipids bound only to the exoplasmic leaflet of the plasma membrane (PM). For long time no such pickets have being found, until recently thanks to single particle tracking (SPT) on living cells CD44 was shown to act as a picket, helping in organizing proteins and lipids at the PM. Therefore, evidence suggests that if existing, lipid rafts would be controlled by the actin cytoskeleton. To our knowledge, the use of super-resolution microscopy has not been exploited to study γ-secretase at the PM, including its dynamics and relationships to the lipid microenvironment. So far, most of our knowledge on γ-secretase is gathered essentially using biochemical approaches.Thus, a major objective of my research has been to setup advanced microscopy techniques, including structured illumination microscopy (SIM), photoactivated localization microscopy (PALM)/stochastic reconstruction microscopy (STORM) to study the biology of γ-secretase incellular lipid bilayers. First we addressed the distribution of PSEN1/γ-secretase complexes with respect to cluster formation, size and density. We found that strikingly instead of finding clusterization claimed by lipid rafts, we mostly found monomeric or dimeric associations of the complex. Moreover, by tagging two subunits of the same complex, we could visualize and confirm the stochiometry of the complex subunits. Next we addressed its association with a substrate or sheddase, were we found no significant associations for γ-secretase and its sheddases as recent biochemical studies have proposed. However, we did found association of γ-secretase with its substrate APP, which was expected. Finally, we went to living cells to address PSEN1/γ-secretase lateral diffusion (single particle tracking or SPT-PALM) and how this is influenced by internal (e.g. FAD mutations) and external factors such as cholesterol and actin. Although we found, as expected, an increase in mobility after actin depletion, cholesterol depletion turned out to immobilize, even at short time-points, the complex. This result indicate that cholesterol is indeed associated to γ-secretase and plays a critical role in its mobility, most likely affecting its conformation and thus, its conformation. But the precise mechanism is still under study.
Toegankelijkheid:Closed