We perform a replica exchange molecular dynamics simulation corresponding to a 2.8 mu s total time for the extensive enhanced sampling of the conformational space of the C-terminal part (residues 124-226) of the mouse prion protein (PrP); 1.3% of the conformations sampled display a high level of beta-structure (>= 19 residues), allowing the assessment of beta-propensities along the sequence and high-lighting the roost structurally labile hot spots. A clustering algorithm is applied to sort the structures of this pool according to their fold. Ten beta-rich folds are thus defined and analyzed with regard to their topology, accumulation temperatures, and structural characteristics. In contrast to the so-called spiral and beta-helix models suggesting that the beta-rich core of the scrapie isoform (PrPSc) comprises the N-terminal tail and part of the C-terminal domain up to helix 1 (H1), we present putative structural models for monomeric precursors of PrPSc and PrP beta-oligomers that are characterized by a C-terminal beta-rich core, in agreement with the suggestions of a series of recent experiments.