Secondary-structure prediction revisited: Pβ and Pc represent structures of amyloids and aid elucidating phenomena in interspecies transmissions of prion.

Abstract

Prion is a unique infectious agent which consists solely of abnormally-folded prion protein (PrPSc) but possesses virus-like features, e.g. existence of strain diversity, adaptation to new hosts and evolutionary changes. These biological phenomena were attributed to the structural properties of PrPSc due to lack of genetic material of prion. Therefore, regardless of incompatibility with high-resolution structural analysis, many structural models of PrPSc have been hypothesized based on limited structural information and, recently, models consisting solely of {beta}-sheets and intervening loops/kinks have been suggested, i.e. parallel in-register {beta}-sheet models and {beta}-solenoid model. Given the relatively simple structural models of PrPSc, we utilized values of theoretical {beta}-sheet or random-coil propensity (P{beta} or Pc, respectively) calculated by secondary structure prediction with a neural network to analyze interspecies transmissions of prion, because numerical conversion of the primary structures would enable quantitative comparison between PrP with distinct primary structures. Reviewing experiments in the literature, we ascertained biological relevance of P{beta} and Pc and demonstrated how those parameters could aid interpretation and explain phenomena in interspecies transmissions. Our approach can lead to development of a versatile tool for investigation of not only prion but also other amyloids.