Prion properties and structural analysis of Candida albicans and Candida glabrata Ure2 proteins

Abstract

Prions are infectious proteins first implicated as causative agents of scrapie and kuru disease. These diseases are characterized by spongiform brain pathology caused by toxic amyloid filament deposition. Our goal is to understand how the conserved structural components of proteins contribute to prion amyloid formation. We used Saccharomyces cerevisiae as a model system because it harbors the prions [URE3] and [PSI+], which are amyloid prions of the proteins Ure2p and Sup35p, respectively. Two structural protein elements have been proposed to contribute to prion infectivity. One is a conserved amino acid domain in Ure2p found in many yeast species, argued to be important for the conservation of prion-forming ability. The other structural element consists of N/Q residue stretches, found in both Ure2p and Sup35p. We chose to examine Ure2 proteins from human pathogenic yeasts, Candida albicans and Candida glabrata, each containing one of the structural elements. Due to greater homology to Ure2p<italic>S. cerevisiae</italic></super>, presence of the conserved sequence, and the proposed idea that prions are evolutionarily conserved to benefit yeast, we hypothesized that the Ure2p<super><italic>C. glabrata</italic></super> will form a prion. To test this hypothesis, we compared the structures of Ure2p<super><italic>C. glabrata</italic></super> and Ure2p<super><italic>C.albicans</italic></super>, and their ability to form infectious amyloid.

Contrary to our hypothesis, our data showed that the more distantly related Ure2p<italic>C. albicans</italic></super> forms a prion, while Ure2p<super><italic>C.glabrata</italic></super> does not. Results showed that Ure2p<super><italic>C. albicans</italic></super> readily forms amyloid filaments <italic>in vitro</italic>, and that these filaments are infectious. In contrast, long-term incubation of Ure2p<super><italic>C. glabrata</italic></super> produced small amounts of amyloid, providing a sufficient explanation of its failure to form a prion <italic>in vivo</italic>. These results strongly suggest Ure2p is not evolutionarily conserved to become a prion and that homology and phylogenetic relationship cannot predict if a protein will generate a prion. We show that, like other yeast prion amyloids, Ure2p<super><italic>C. albicans</italic></super> amyloid has parallel, in-register beta-sheet architecture. Furthermore, data showed overexpression of the chaperone Hsp104 dramatically increased the frequency of Ure2p<super><italic>C. albicans</italic></super> prion formation, but did not allow Ure2p<super><italic>C. glabrata</italic></super> to form a prion. Thus, amino acid composition and not conserved sequence is an important determinant for prion generation.