Prion-Inducing Domain of Yeast Ure2p and Protease Resistance of Ure2p in Prion-Containing Cells

The genetic properties of the [URE3] non-Mendelian element of Saccharomyces cerevisiae suggest that it is a prion (infectious protein) form of Ure2p, a regulator of nitrogen catabolism. In extracts from [URE3] strains, Ure2p was partially resistant to proteinase K compared with Ure2p from wild-type extracts. Overexpression of Ure2p in wild-type strains induced a 20- to 200-fold increase in the frequency with which [URE3] arose. Overexpression of just the amino-terminal 65 residues of Ure2p increased the frequency of [URE3] induction 6000-fold. Without this “prion-inducing domain” the carboxyl-terminal domain performed the nitrogen regulation function of Ure2p, but could not be changed to the [URE3] prion state. Thus, this domain induced the prion state in trans, whereas in cis it conferred susceptibility of the adjoining nitrogen regulatory domain to prion infections.

27. Cyclin A-H6 is a histidine-tagged truncated bovine cyclin A that can bind Cdk2 (R. Y. C. Poon and T. Hunt, unpublished data) and was expressed in bacteria and purified as described (21). GST-Cdk2 was cleaved with thrombin and treated with hirudin as above. The cleaved Cdk2 (100 ng) was incubated with cyclin A-H6 (100 ng) and GST-KAP (1 pLg) with bovine serum albumin (5 ,ug) as a carrier at 230C for 30 min; the GST-KAP was then recovered with GSHagarose.
28. GST-Cdk2 was phosphorylated with CAK immunoprecipitates in the presence of 1 mM ATP and 15 mM Mg2+; the phosphorylated GST-Cdk2 was immobilized on GSH-agarose, cleaved with thrombin, and treated with hirudin as above. The cleaved Cdk2 (100 ng) was incubated with cyclin A-H6 (1 ,ug) and then with GST-KAP or C140S mutant (1 ,ug), followed by cyclin A-H6 (1 ,ug) as indicated; each incubation was at 230C for 15 min. The histone Hi kinase assay was as described (4) A prion is an infectious protein, an altered form of a normal protein that may have lost its normal function but has acquired the ability to convert the normal form into the altered (prion) form. This concept originated in studies of scrapie of sheep, kuru and Creutzfeldt-Jakob diseases of humans, and bovine spongiform encephalopathy (mad cow disease) (1, 2). These diseases are believed to be caused by a self-propagating conformational change of a highly conserved protein denoted PrP. We previously suggested that [URE3] and [PSI], two non-Mendelian elements of yeast, are prion forms of the chromosomally encoded Ure2p and Sup35p, respectively (3). [URE3] (4,5) or mutations in the chromosomal URE2 gene (6) each produce derepression of nitrogen catabolic enzymes that would normally be repressed by a good nitrogen source (7).
[PSI] (8,9) or mutations in SUP35 (10) (13). The presence of more of the normal form increases the likelihood that the spontaneous prion change will occur. The relative protease resistance of PrP from diseased animals compared with that from normal animals was an early indication that the mammalian scrapie agent was an altered form of this protein (17). It has also been critical in the recent demonstration of in vitro conversion of PrPC to PrPSc (2). By immunoblotting with a polyclonal antibody to Ure2p (3), we detected equal amounts of similarly migrating Ure2p in extracts of strains with and without [URE3] (Fig. LA, lanes marked X). Ure2p in extracts of normal strains was digested by proteinase K in less than 1 min to products that run off the gel, whereas Ure2p from isogenic [URE3] (prion-con- by growth on rich medium containing 5 mM guanidine HCl (3), showed the same proteinase K sensitivity of Ure2p as the parental wild-type strain (Fig. 1, A and B). All three [URE3] isolates tested showed increased protease resistance of Ure2p (Fig. 1, A and B). Ure2p was stable in both wild-type and [URE3] extracts unless protease was added (Fig. IA), and treatment of an equal mixture of extracts from wild-type and [URE3] strains showed an essentially additive result (Fig. IA), indicating that the difference in degradation was not due to proteases in the wild-type extract or inhibitors in the [URE3] extract, but to a difference in the structure or associations of Ure2p.
[URE3] arises at a frequency of about 10-5 (3,5). Overexpression of Ure2p increased this frequency by 20to 200-fold (3) (Fig. 2). Deletions from the COOHterminus of Ure2p eliminated complementation of a chromosomal ure2 deletion, but the truncated protein could induce [URE3] in a strain with a normal chromosomal URE2 with an efficiency 100-fold greater than that of the intact protein, that is, 3000-fold above the spontaneous rate (Fig. 2). The NH2-terminal 65 amino acids sufficed for this increased [URE3]inducing activity, so we call this the "prion-inducing" domain. In-frame deletion of eight residues within the nitrogen regulatory domain around the Apa I site (Fig. 2) mimicked the COOH-terminal deletions. The prioninducing domain is 40% asparagine and 20% serine 94 and threonine (Fig. 2) (18).
Deletion of part or all of the prioninducing domain eliminated the ability of the overexpressed truncated Ure2p to induce [URE3] (Fig. 2), but the remaining COOH-terminal part could complement a chromosomal ure2 deletion (Fig. 2) (18). We therefore refer to it as the nitrogen regulatory domain (Fig. 2); it contains the region related to glutathione-S-transferases (18).
Overexpression of the nitrogen regulatory domain alone failed to induce the were expressed from the GAL1 promoter and tested for complementation of ure2A (24) (Fig. IC), although the difference was less marked than when the entire Ure2 protein was examined. Although only fragments of Ure2p were resistant to proteinase K, the apparent size of the prion-inducing domain did not change on digestion, suggesting that it is within the proteaseresistant domain of the prion form of Ure2p.  (16). Whether the COOH-terminal region essential for cell growth is also needed for propagation of [PSI] could not be tested. The [PSI] phenotype of increased nonsense suppression was only expressed when the NH2-terminal domain was expressed in cis with the essential COOHterminal domain, that is, as one molecule (16). Although the authors propose that "the Sup35p serves as a trans acting factor required for the maintenance of [psi+]" (16, p. 675), our prion model for [PSI] (3) is more consistent with the results. Our results likewise indicate that the COOH-terminal portion of Ure2p is not inactivated by the NH2-terminal prion-inducing domain unless it is covalently attached. Whether overproduction of part of PrP could induce the prion change is not yet known.
The Ure2p prion domain resembles neither PrP nor the region of Sup35p needed to propagate [PSI]. Although both Sup35p and PrP have similar octapeptide repeats (19), these repeats seem to be dispensable for propagation of scrapie (1). Mutations resulting in increased prion formation (familial Creutzfeldt-Jakob disease) are distributed through most of the PrP gene (1). Similarly, we find that deletions in the large COOH-terminal domain result in a markedly increased frequency of [URE3] generation. The absence of apparent structural similarity among these prions, PrP, Sup35p, and Ure2p, suggests that prions can arise in various ways, producing analogous phenomena by substantially different detailed mechanisms.