Contrasting Roles of Asparagine and Glutamine in the Aggregation of Prion-Like Proteins

Abstract

Sequences rich in glutamine (Q) and asparagine (N) are intrinsically disordered in monomeric form, but can aggregate into highly ordered amyloids, as seen in Q/N-rich prion domains (PrDs). Amyloids are fibrillar protein aggregates rich in β-sheet structures that can self-propagate through protein-conformational chain reactions. It has been shown that tuning the amount of Ns and Qs in yeast PrDs results in very different effects: N-rich mutants lead to non-pathological self-seeding amyloids while Q-rich mutants lead to toxic nonamyloid structures. These structural preferences have been explained in terms of an enhanced β- hairpin turn propensity of Ns over Qs. Here, we consider a variety of N/Q-rich peptides, including sequences found in the yeast Sup35 PrD, in parallel and antiparallel β-sheet aggregates, and probe all their possible steric-zipper interfaces to determine their relative stability. Our results show that polyglutamine aggregates are more stable than polyasparagine aggregates. The observation that Q-rich PrD mutants lack amyloid structure can be attributed to three facts. First, although once formed polyglutamine aggregates are more stable, their entropic contribution to the free energy is less favorable: Q-rich sequences have a larger phase space to sample. Second, N-rich sequences favor parallel β sheets, for which the formation of hairpin turns is irrelevant: indeed polyasparagine β-hairpins are more unstable than polyglutamine hairpins. Third, when other amino acids are present, such as in the Sup35 PrD, their shorter side chains favor steric-zipper formation for N but not Q, as they preclude the in-register association of the long Q side chains.