From conversion to aggregation: protofibril formation of the prion protein

Abstract

The ability to diagnose and treat prion diseases is limited by our current understanding of the conversion process of the protein from healthy to harmful isoform. Whereas the monomeric, benign species is well characterized, the misfolded conformations responsible for infectivity and neurodegeneration remain elusive. There is mounting evidence that fibrillization intermediates, or protofibrils, but not mature fibrils or plaques, are the pathogenic species in amyloid diseases. Here, we use molecular dynamics to simulate the conversion of the prion protein. Molecular dynamics simulation produces a scrapie prion protein-like conformation enriched in beta-structure that is in good agreement with available experimental data. The converted conformation was then used to model a protofibril by means of the docking of hydrophobic patches of the template structure to form hydrogen-bonded sheets spanning adjacent subunits. The resulting protofibril model provides a non-branching aggregate with a 3(1) axis of symmetry that is in good agreement with a wide variety of experimental data; importantly, it was derived from realistic simulation of the conversion process.

Fig. 1.

Simulated conversion of Syrian hamster D147N PrP^C to PrP^Sc at low pH levels. (a) The wild-type NMR structure is shown (Left) with the helices and strands labeled. A representative PrP^Sc-like structure (8-ns snapshot) is shown (Right). (b) Hydrogen-bonding network of the PrP^C β-strands S1 and S2 and the in silico PrP^Sc sheet E1–E3.

Fig. 4.

Dimensions of our PrP protofibril and higher-order oligomers. (a) A diglycosylated PrP^Sc-like trimer with circumferences (dashed circles) of the β-/extended core (magenta), all protein atoms (gray), and the diglycosylated protofibril (cyan). (b) Same view as in a of a 48-mer protofibril with the protein surface shown gray and the sugars shown in cyan. (c) Side view of a 48-mer protofibril. Bars at the top indicate diameters of the 35-Å extended β-core (magenta), 65-Å protein diameter (gray), and a 110-Å diglycosylated protofibril (cyan).

Fig. 2.

Modeling protofibril formation. (a) Building of a protofibril with 3₁ axis (viewed down the fiber axis). The oligomerization site occurs between E4 and E1 of the adjacent monomer. (b) Views of hexameric representation of protofibril showing maintenance of symmetry on oligomerization and propagation of the extended strands between monomers to form extended sheets. (c) The dimer with residues 113–141 is shown in green, corresponding to a peptide that inhibits scrapie formation in vitro. In the other dimer and corresponding space-filling model, a PrP^C epitope responsible for clearance of PrP^Sc is magenta (residues 132–156). (d) Main chain and space-filling model of a hexameric representation of the protofibril with the PrP^Sc selective epitope, residues 142–148 (orange), 162–170 (green), and 214–226 (magenta). The C terminus of one of the subunits (*) was extended from residue 219 to 231 to more accurately illustrate the epitope.

Fig. 3.

Comparison with EM images of two-dimensional PrP crystals. (a) Modeled protofibril (hexamer) with diglycosylated subunits. (b) Superimposition of two hexamers rotated 60° from one another around the fiber axis replicates the 6-fold symmetry of the crystals. (c–f) All images are of our protofibril except where noted. (c, e, and f) Residues 142–176 (magenta) represent the residues deleted from PrP 27–30 to form PrP^Sc106. helix B, helix C, and sugar groups are shown in white. The EM image is a difference map between PrP^Sc106 and PrP27–30 with statistically significant differences shown in magenta (24). (c) By superimposing two hexamers as in a, similar images are obtained. (d) Sugars are shown in cyan. The EM image is a difference map between PrP27–30 and PrP^sc106 with statistically significant differences in glycosylation shown in cyan (24). (e) EM image of the two-dimensional protofibril by Wille et al. (24). [Reproduced with permission from ref. (Copyright 2001, The National Academy of Sciences).] (f) Similar packing and dimensions of our modeled protofibril as compared with the EM image in e.