Nicastrin is heavily glycosylated contributing KDa to the mo

Nicastrin is heavily glycosylated, contributing 40–70KDa to the molecular weight of human γ-secretase. Among the 16 predicted N-linked glycosylation sites in the ECD, 11 have been unambiguously identified. The glycans appear to mostly stabilize the local structure, thus contributing to the folding and stability of nicastrin. Selective elimination of glycosylation through mutation of specific Asn residues mostly gave rise to insoluble or aggregated protein. Nicastrin was reported to play an important role in substrate recruitment, with Glu333 and Tyr337 directly involved in substrate binding [17, 18]. Notably, both residues are buried in a hydrophilic pocket that is formed by polar and charged residues []. Access to Glu333 and Tyr337 is blocked by a surface loop called lid (Figure 3i). In this scenario, the lid must be dislocated in order for the substrate to be recognized by residues in the pocket. However, the lid was recently shown to be dispensable for γ-secretase activity because mutations in the lid or vitamin d3 sources of the lid had little impact on substrate cleavage [54]. In addition, the cleavage of an elongated Notch substrate by γ-secretase was found to be reduced, and the reduced cleavage was restored by the presence of a reducing agent [55]. These experiments were interpreted to suggest nicastrin as a regulator of substrate entry, although this hypothesis must be scrutinized by additional studies.
Perspective The structures at near-atomic resolutions differ markedly from the other reported EM structures of γ-secretase that were determined at lower resolutions. At resolutions lower than 10Å, separation of the detergent micelles away from the transmembrane domain of γ-secretase becomes quite challenging. Consequently, the structures determined in these resolution ranges usually contain information of the detergent micelles. In our case, 3D classifications at resolutions lower than 10Å always yield a duck-shaped structure (Figure 1e), with the duck body representing the transmembrane domain surrounded by a shell of detergents and the duck head the ECD of nicastrin. These features are exactly captured in one study [23] and similarly observed in others [21]. As the resolution improves, the noise from the detergents is gradually peeled off and the signal from the TMs begins to dominate. Comparison of two structures determined at vastly different resolutions is difficult and should be restricted to features that are apparent at both resolutions. For the inhibitor-bound γ-secretase structures, the conformational change induced by DAPT mainly occurs at the active site of PS1 [], whereas compound E appears to induce a marked positional shift of the ECD relative to the transmembrane domain [23]. Such differences should be ideally examined by high-resolution structures of various inhibitor-bound γ-secretase. Using γ-secretase expressed from insect cells, different antibodies and inhibitors were applied to study the conformational changes [56, 57]. Limited by the resolutions coupled with the presence of detergent micelles, the observed structural changes should be interpreted with caution. Obviously, capturing distinct conformations of γ-secretase is of great importance; but such information must be derived from structures that allow faithful interpretation. Structural information on γ-secretase helps guide functional investigation. In a recent study [], 138 AD-associated mutations in PS1 were evaluated for their consequences on the in vitro production of Aβ40 and Aβ42 in the context of assembled γ-secretase. This represents the most comprehensive effort to date on potential relationship between AD mutations and production of Aβ by γ-secretase. Quite unexpectedly, no statistically significant correlation was found between the Aβ42/Aβ40 ratio and the age at onset (AAO) for familial AD, or between the combined Aβ42 and Aβ40 production and the AAO []. With the caveats that these experiments were performed on recombinant γ-secretase variants, this result appears to contradict the causal relationship between the Aβ42/Aβ40 ratio and AD development. As AD is a complex disease caused by multiple factors, these preliminary findings should be scrutinized by rigorous follow-up studies and in vivo investigations.