25 years of yeast prions.

In the early 1990s, Reed Wickner developed a novel hypothesis. For decades it was known that two phenotypes of the yeast Saccharomyces cerevisiae followed non-Mendelian patterns of inheritance [1,2]. These phenotypes were designated [PSI+] and [URE3]. When yeast strains were mated, if either parental strain had one of these phenotypes, all daughter spore clones would inherit the phenotype, although only half would be expected to if the phenotypes were governed by a nuclear gene. Instead, the genetic element resided in the cytoplasm as if it were one of the many yeast viruses. However, unlike yeast viruses, no nucleic acid could be identified. To explain these observations, Reed postulated that the genetic element was composed of protein, not nucleic acid. In 1994, Reed solo-authored an article in Science titled, ‘[URE3] as an altered URE2 protein: evidence for a prion analog in Saccharomyces cerevisiae’, where he described how the puzzling [URE3] and [PSI+] phenotypes could be explained simply as selfpropagating misshapen forms of the Ure2 and Sup35 proteins, respectively [3]. His experiments elegantly demonstrated that the Ure2 protein was itself the critical factor for the formation and propagation of the [URE3] prion, and he proposed that it was a yeast analog of mammalian prions. Noting the logical parallels with [PSI+] and the Sup35 protein, he extended his hypothesis to include [PSI+] as a prion analog of the Sup35 protein, opening the door for discovery of other prions in yeast. At that time, the prion concept – suggesting a form of the protein PrP was the infectious entity responsible for prion disease – was controversial and applied solely to the infectious species that caused transmissible spongiform encephalopathies of mammals (e.g. scrapie, Kuru and Mad Cow disease). Little else was known of PrP extracted from infectious brain aside from it being fibrous aggregates enriched in beta-sheet structure. Whether PrP was a prion component, the prion component, or merely a propagation factor for another pathological agent, was arguable. The question of whether prions existed in nature as defined (i.e. infectious proteins) remained unresolved. The enormous impact of Reed’s short paper is made obvious by the suddenly renewed and widespread interest in non-Mendelian genetic elements and the dramatic evolution of the scientific community’s view of prions in the 25 years since its publication. The broad acceptance of prion mechanisms is largely based on work and ideas pioneered by Reed and colleagues in the yeast model system. These studies provided the first confirmation of protein-only infectious elements and identified a common structural model that enabled a mechanism of protein infectivity: self-propagating amyloid with parallel in-register beta-sheet architecture [4]. This conceptual framework established how prion, or prion-like, mechanisms could be involved in human diseases, especially neurodegenerative disorders that commonly feature pathological protein aggregation in neuronal networks. This structural insight also explained how a variety of prion strains (or variants, with subtle differences in phenotype) could be ‘encoded’ and perpetuated by molecular variations in the underlying amyloid configuration. Reed thus provided a mechanism by which proteins could act as genes by templating their own conformation, just as DNA mediates inheritance by templating its sequence [5]. Recognizing that the misfolded proteins underlying prion phenotypes were a disease state [6], Reed also astutely identified that the domains facilitating prion formation were serving cellular functions that were independent of their amyloid-forming properties [7]. These intrinsically disordered domains with low-sequence complexity are found in dozens of different yeast proteins that appear to have at least some