Identification and characterization of a spontaneously aggregating amyloid-forming variant of human PrP((90-231)) through phage-display screening of variants randomized between residues 101 and 112.

The N-terminal ‘unstructured’ region of the human prion protein [PrP(90–231)] is believed to play a role in its aggregation because mutations in this region are associated with seeding-independent deposition disorders like Gerstmann–Straussler–Scheinker disease (GSS). One way of examining the effects of such mutations is to search combinatorially derived libraries for sequence variants showing a propensity to aggregate and/or the ability to interact with prion molecules folded into a β-sheet-based conformation (i.e., β-PrP or PrPSc). We created a library of 1.8×107 variants randomized between positions 101 and 112, displayed it on filamentous bacteriophage, and ‘spiked’ it with a ∼25% population of phages-bearing wild-type prion (wt-PrP). Screening was performed through four rounds of biopanning and amplification against immobilized β-PrP, and yielded three β-PrP-binding populations: wt-PrP (26% representation) and two non-wt-PrP variants (∼10% and ∼64% representation, respectively). The remarkable enrichment of one non-wt-PrP variant (MutPrP) incorporating residues KPSKPKTNMKHM in place of KGVLTWFSPLWQ, despite its initial representation at a 5 million-fold lower level than wt-PrP, caused us to produce it and discover: (i) that it readily aggregates into thioflavin-T-binding amyloids between pH 6.0 and 9.0, (ii) that it adopts a soluble β-sheet based monomeric structure at pH 10.0, (iii) that it is less thermally stable and more compact than wt-PrP, and (iv) that it displays significantly greater resistance to proteolysis than wt-PrP. Our results suggest that sequence variations in the 101–112 region can indeed predispose the prion for aggregation.