Spatial-temporal dissection of organelle-specific y-secretase interactomes using nanotechnology
From the moment Alzheimer’s disease (AD) causing mutations were discovered in the PSEN genes, its protein products, presenilins (PSEN), have been in the center of attention for AD research. While their function as the catalytic component of the tetrameric intramembrane protease, γ-secretase, was described, pieces of the puzzle started coming together. However, decades later, many questions remain about its biology, exact role in the development of AD, and potential as a drug target. γ-Secretase biology is a complex story that is influenced by many different factors. Further insight into these is necessary to discover new ways to modulate γ-secretase function and concomitantly the development of AD. Therefore, in this thesis, we set out to investigate both the formation of the tetrameric γ-secretase complex and its interactome.
The stoichiometric assembly of γ-secretase is a crucial step in its maturation. Using fractionation by complementary ultracentrifugation methods, we showed that while full complexes were present in intermediate compartments, in the endoplasmic reticulum (ER) only immature subcomplexes were observed. Through in vitro budding assays, we isolated coat protein complex II (COPII) vesicles, allowing us to study ER exit. We showed that γ-secretase does not exit the ER in full complex, but as monomeric subunits and dimeric subcomplexes (nicastrin (NCT) with anterior pharynx defective-1 (APH1) and PSEN1 with presenilin enhancer-2 (PEN-2)). Furthermore, NCT and PSEN1 have distinct preferences for different Sec24 isoforms, with NCT’s ER exit being mediated by Sec24C/D and PSEN1’s ER exit by Sec24A. The packaging of PSEN1 in COPII vesicles by Sec24A was achieved through a DPE motif that is lost in the AD-associated PSEN1∆E9 variant. Formation of subcomplexes increased the subunits stability and cell surface localization. Our data support a model wherein formation of dimeric subcomplexes and Sec24 paralog selectivity defines the stepwise assembly of γ-secretase and controls as such final levels of γ-secretase activity in the cell.
Then we turned to interactomics to identify localization specific interactors. Previous results from our lab described the different subcellular localizations of PSEN1/- and PSEN2/γ-secretase, with PSEN1/γ-secretase being broadly distributed throughout the cell and PSEN2/γ-secretase restricted to late endosomes and lysosomes. Mutating the PSEN2-specific trafficking motif responsible for its localization, led to the generation of two PSEN2 trafficking mutants with altered subcellular localizations. We first attempted to identify localization specific interactors by comparing the interactome of wild-type PSEN2 with both PSEN2 trafficking mutants. We found an interesting interaction with the lysosomal proton pump, but did not see any clear localization-dependent interactions. Therefore, we isolated late endosomes and lysosomes as a starting fraction for interactome isolation, to identify lysosomal specific interactions. This allowed us to identify ClC7, a lysosomal Cl-/H+ exchanger, as an interactor of PSEN2/γ-secretase and the lysosomal proton pump. γ-Secretase influences ClC7 localization, as PSEN double knockout (PSENdKO) resulted in decreased lysosomal ClC7, which could be rescued by either PSEN1 or PSEN2. We also found that ClC7 overexpression could rescue an endosomal pH defect in PSENdKO cells. With these results, we linked the PSEN/ClC7 interaction to intraluminal ion homeostasis.
To investigate the γ-secretase interactome more in depth we took advantage of two orthogonal interactomics methods: affinity purification with the GFP-tag and proximity-dependent biotinylation with the APEX2-tag. This approach allowed us to identify several proteins linking γ-secretase to lysosomal functions. We found a direct interaction between PSEN2/γ-secretase, Lamtor1 and RagC, both proteins involved in mTORC1 activation through nutrient sensing. We confirmed a role for PSEN2 in this process by measuring an increased mTORC1 activation in PSEN2KO cells. Interestingly, we found that lysosomal transport, which is linked to mTORC1 activation, was affected in PSEN2KO neurons. Finally, we also described colocalization of lysosomal PSEN2/γ-secretase and ER-localized reticulon-3 in ER-lysosome contact sites. This localization of PSEN2/γ-secretase at membrane contact sites can form an exciting link between ER-lysosome communication, nutrient sensing and ion homeostasis.