Prions of Yeast and Fungi

The genetic properties of [URE3] and [PSI], two non-chromosomal genetic elements of Saccharomyces cerevisiae, indicated that they were infectious proteins (prions) (1). Subsequent studies have supported this proposal, and the genetic criteria we proposed have been used in the discovery of another new prion, [Het-s], in the filamentous fungus Podospora anserina (2). The prion hypothesis has long been an intriguing explanation of the transmissible spongiform encephalopathies, such as scrapie, Creutzfeldt-Jakob disease, and “mad cow disease” (3–5) (reviewed in Refs. 6 and 7). Studies using Saccharomyces and Podospora have provided evidence of a type not available from studies of scrapie that there can be such a thing as an infectious protein. This work also revealed that prions can be the basis for inherited traits and initiated the use of the powerful yeast system to study this phenomenon. Here we review the basis for the proposal that [URE3], [PSI], and [Hets] are prions of the chromosomally encoded Ure2p, Sup35p, and Het-s protein, respectively. We also review the properties of [URE3] and [Het-s]. Further studies of [PSI] are reviewed by Liebman and Derkatch in the following minireview (8), and other reviews of these subjects have appeared (9–12).

The genetic properties of [URE3] and [PSI], two non-chromosomal genetic elements of Saccharomyces cerevisiae, indicated that they were infectious proteins (prions) (1). Subsequent studies have supported this proposal, and the genetic criteria we proposed have been used in the discovery of another new prion, [Het-s], in the filamentous fungus Podospora anserina (2). The prion hypothesis has long been an intriguing explanation of the transmissible spongiform encephalopathies, such as scrapie, Creutzfeldt-Jakob disease, and "mad cow disease" (3)(4)(5) (reviewed in Refs. 6 and 7). Studies using Saccharomyces and Podospora have provided evidence of a type not available from studies of scrapie that there can be such a thing as an infectious protein. This work also revealed that prions can be the basis for inherited traits and initiated the use of the powerful yeast system to study this phenomenon. Here we review the basis for the proposal that [URE3], [PSI], and [Hets] are prions of the chromosomally encoded Ure2p, Sup35p, and Het-s protein, respectively. We also review the properties of [URE3] and [Het-s]. Further studies of [PSI] are reviewed by Liebman and Derkatch in the following minireview (8), and other reviews of these subjects have appeared (9 -12).

[URE3], a Non-Mendelian Genetic Element Affecting Nitrogen Catabolism
Because ureidosuccinate (USA), 1 an intermediate in uracil biosynthesis, happens to resemble allantoate, a poor but usable nitrogen source, USA uptake is repressed by good nitrogen sources ( Fig. 1) (reviewed in Refs. 13 and 14).
Lacroute and co-workers (15, 16) discovered ure2 by selecting mutants able to take up USA on media with a good nitrogen source (ammonia). The USA uptake of one isolate was dominant, showed irregular segregation in meiosis (17), and was transmissible by cytoplasmic mixing (18), confirming its determination by a non-chromosomal gene that Lacroute named [URE3].

Three Genetic Criteria for a Prion Illustrated by [URE3]
Yeast viruses are found as nonchromosomal genetic elements (19) and do not venture out of the cell but are able to become widespread as a result of transmission through mating. We likewise expect infectious proteins of yeast to be found as non-chromosomal genetic elements. We proposed three genetic criteria that should distinguish a prion from a nucleic acid replicon (1) (Fig. 2). A prion is an altered form of a normal chromosomally encoded protein. This prion form has lost its normal function but has acquired the ability to convert the normal form of the protein into this same abnormal (prion) form.
Reversible Curability-If a prion can be cured, it should be possible for it to arise again in the cured strain at some low frequency, because the same event that initially gave rise to the prion could occur again.
[URE3] can be cured by the growth of cells on media containing 1-5 mM guanidine HCl, but from cured clones [URE3]-carrying colonies can again be isolated (1). These [URE3] clones arise at frequencies similar to the 10 Ϫ6 frequency with which [URE3] arises in most wild-type strains.
Overproduction of the Normal Protein Increases the Frequency of Prion Formation-Because the initial prion formation event is a spontaneous change of the protein, it is expected that overproducing the normal protein will increase the number of molecules that are candidates to undergo this change. Indeed, overproduction of Ure2p increases the frequency of de novo generation of [URE3] by 20 -200-fold (1).
Phenotype Relationship of Prion and Mutation of the Gene for the Protein-Because the prion is propagated by conversion of the normal form to the prion form, mutations of the chromosomal gene for the protein that prevent it being synthesized will prevent propagation of the prion. So the chromosomal gene for the protein will be seen (perhaps discovered) as a gene necessary for propagation of the non-Mendelian genetic element. But unlike such genes necessary for propagation of mitochondrial DNA (pet) or of the L-A dsRNA virus (mak), the phenotype of the mutant should be the same as the phenotype of the presence of the non-Mendelian genetic element. This is because either the presence of the prion or the mutation in the gene for the protein results in a deficiency of the normal protein.
The genetic criteria satisfied for yeast prions and, in two of three cases, by the Podospora prion are not yet satisfied by the transmissible spongiform encephalopathies. Thus, these new findings have provided a major foundation for the notion that there can be such a thing as an infectious protein.

[PSI] Satisfies the Genetic Criteria as a Prion Form of Sup35p
Sup35p is a subunit of the translation release factor that recognizes termination codons and releases the completed peptide from the last tRNA.
[PSI] is a non-Mendelian genetic element that, like sup35 mutations, increases the strength of weak suppressor tRNAs (20,21). We pointed out that [PSI], like [URE3], satisfies all three genetic criteria to be a prion form of Sup35p (1).
[PSI] may be cured by various agents (22,23) but from the cured strains may again be obtained clones in which [PSI] has arisen de novo (24). Overproduction of Sup35p increases the frequency with which [PSI] arises about 100-fold (25). Finally, the phenotype of sup35 mutants is similar to that of [PSI], and Sup35p is necessary for the propagation of [PSI] (26,27). Further discussion of [PSI] will be found in the minireview by Liebman and Derkatch (8).

Further Support for the Prion Model for [URE3]
is a prion form of Ure2p, it is expected that there will be a structural difference of some kind between Ure2p in wildtype and [URE3] strains. In fact Ure2p is more proteaseresistant in extracts of [URE3] strains than wild-type strains (28). The 40-kDa Ure2p is quickly and completely degraded, but Ure2p in extracts of [URE3] cells is first partially degraded to species of 30 and 32 kDa before eventually being completely degraded.
[URE3] Is Not a Stable Regulatory Circuit-Ure2p is a transcription regulator, and stable transcription circuits are known in some cases to produce inheritable ("epigenetic") traits (29). However, the frequency with which [URE3] arises is independent of the nitrogen-repressed or derepressed state of cells (30). Moreover, [URE3] can be propagated in cells that are either repressed or derepressed (30). Finally, as discussed below, the parts of the Ure2p molecule responsible for prion propagation and nitrogen regulation are completely distinct (28).
[URE3] Really Arises de Novo-[URE3] is isolated as an apparent mutant; it is not found in "wild-type" strains. Could [URE3] be a defective interfering derivative of a normal plasmid dependent on Ure2p for its propagation, analogous to a defective interfering virus or a "suppressive petite" deletion derivative of mitochondrial DNA? In that case, [URE3] would produce its phenotype by eliminating the normal version of the plasmid, thus explaining the dominance of [URE3]. The ure2 mutants would also produce their phenotype by resulting in loss of the normal plasmid, and so the identity of phenotypes of [URE3] and ure2 strains would be explained. However, replacing the URE2 gene in a ure2 mutant would then fail to correct its phenotypic defect, and this is not the case (31). This model would also predict that the [URE3] element could not be gen-erated in a ure2 strain because the parent "normal" plasmid would be already lost. In fact, introduction of a URE2 plasmid into a ure2 strain makes it able to become [URE3] at some low frequency (30).
This de novo induction of [URE3] is, in fact, because of overproduction of the Ure2 protein, not the URE2 mRNA or the gene itself (30). Thus, [URE3] arises de novo, not as a derivative of a pre-existing nucleic acid replicon. Moreover, it is the Ure2 protein that leads to [URE3] arising.

Prion Domain and Nitrogen Regulation Domains of Ure2p
Deletion analysis of URE2 showed that residues 66 -354 determine the nitrogen regulation function of Ure2p (28,31), whereas prion induction and propagation involve primarily residues 1-65 (Fig. 3) (28, 30). Although overexpression of intact Ure2p increases the frequency with which [URE3] arises by 20 -200-fold (1,28), the N-terminal 65 residues are sufficient to induce [URE3] at even higher efficiency (28). Thus, the two functions of Ure2p are attributed to separate parts of the molecule.
Deletions of parts of the C-terminal nitrogen regulation domain increase by 100-fold or more the frequency with which it converts to the prion form (28), suggesting that the C-terminal domain normally stabilizes the N-terminal domain and prevents it from converting to the prion form. Although they interact, the prion domain and nitrogen regulation domain can function completely separately in the cell when produced as separate molecules. The prion domain alone can propagate [URE3] and is not affected by the concurrent presence of separate molecules of the nitrogen regulation domain (30). Likewise, only when the nitrogen regulation domain has a covalently attached prion domain is the nitrogen regulation function affected by the presence of the [URE3] prion (30).

FIG. 1. Ure2p is a regulator of nitrogen catabolism that indirectly blocks uptake of ureidosuccinate, an intermediate in uracil biosynthesis.
Yeast growing on a rich nitrogen source, such as ammonia or glutamine, repress transcription of enzymes and transporters needed for utilization of poor nitrogen sources. Ure2p senses the availability of a rich nitrogen source and blocks the action of Gln3p, a positive transcription regulator of many genes whose products facilitate utilization of poor nitrogen sources (37,38). Among these genes is DAL5, encoding the allantoate importer (39). The incidental chemical similarity of ureidosuccinate to allantoate makes the former a substrate for the allantoate importer, Dal5p (40). USA (N-carbamyl aspartate) is an intermediate in uracil biosynthesis, the product of aspartate transcarbamylase. Cells blocked in this enzyme can take up ureidosuccinate to make uracil only if they are growing on a poor nitrogen source. Lacroute and co-workers (15,16) found that chromosomal ure2 mutants, selected for ability to take up USA on a rich nitrogen source (ammonia), were unable to repress enzymes for utilization of poor nitrogen sources.

[Het-s], a Prion of P. anserina, Carries Out a Normal Cellular Function
A fungal colony is a syncytium so that cells can exchange cytoplasm and even nuclei. When two colonies grow together, the cellular processes (hyphae) of one colony generally fuse (anastomose) with those of the other, so the colonies can share nutrients (32). This "hyphal anastomosis" has the disadvantage that viruses present in one colony are readily transmitted throughout the other. To limit this problem, the process is completed only by closely related colonies, likely to already carry the same viruses. Colonies of Podospora must be identical at 8 "het" loci to fuse. Colonies differing in alleles at one or more het locus will initially fuse hyphae, but the fused hyphae undergo a rapid degeneration process with formation of a barrier to further hyphal fusions. This is called "heterokaryon incompatibility" (Fig. 4).
Recently, Coustou et al. (2) showed that the product of one of these loci, het-s, only functions when it is in a prion form. As with [URE3] and [PSI], the Podospora prion was first detected as a non-Mendelian (non-chromosomal) genetic element. The het-s locus can have either of two alleles, called het-S and het-s. Whereas het-s cells fuse readily with other het-s cells and het-S cells fuse with het-S partners, a meeting of het-s and het-S cells results in the heterokaryon incompatibility reaction. Rizet (33) found that cells with genotype het-s could have either of two phenotypes. One, called [Het-s], shows this heterokaryon incompatibility when meeting a het-S strain. The other, called [Het-s*], is compatible with both het-s and het-S partners.
Coustou et al. (2) used two of the genetic criteria and the protease resistance of the Het-s protein in [Het-s] strains as evidence that [Het-s] is a prion form of the Het-s protein.
[Het-s] can be efficiently eliminated but will spontaneously return at low frequency. Moreover, overproduction of the Het-s protein results in an increase in the frequency with which [Het-s] arises de novo. The het-s gene is necessary for the propagation of [Het-s], as expected if [Het-s] is a prion form of its product.
These findings are important in showing that a prion need not be a cause of disease but rather may be the mediator of a normal cellular function. Heterokaryon incompatibility is a normal event in the life of a fungus, with a physiological purpose.

Cortical Inheritance in Ciliates and the Parallel with Prions
Sonneborn showed that the pattern of cilia on the surface (cortex) of Paramecium could be altered by microsurgery or accidents occurring during mating, events unlikely to alter the genome. These changes were transmitted to the offspring through mitosis and meiosis, in a kind of Lamarkian process (35). The implication of these studies is that the pattern of cilia on the cell surface acts as a template in the formation of the new cell. Similar phenomena have been demonstrated in other ciliates (36). This phenomenon has some formal similarity to the template function of prions believed to be the basis of prion propagation (36). Are other aspects of cell morphology heritable? Do other cellular structures act as templates in the generation of their offspring?

Comparison of Prion Systems
Whereas the prion domains of Ure2p and Sup35p are rich in asparagine and glutamine, there are no such regions in either PrP or the Het-s protein. This suggests that there will be differences in the detailed mechanisms of prion propagation, although there may yet be important similarities. Whereas [URE3] and [PSI] produce their phenotypes by inactivation of Ure2p and Sup35p, respectively, scrapie and [Het-s] are detectable because of positive activities of the altered forms of PrP and the Het-s protein, respectively.

Prospects for Future Work
The purification of Ure2p 2 will allow detailed structural studies of differences between the normal and prion forms of the molecule. Further dissection of prion domains and nitrogen regulation functions of Ure2p using molecular genetics are also under way. 3 How widespread is the prion phenomenon? The Saccharomyces and Podospora systems should be adaptable to general screens for prion domains of any organism, and we have begun such an approach. 4 Transmissible spongiform encephalopathies, [URE3] and [PSI], are diseases, but [Het-s] mediates a normal cellular function. We can anticipate that some other cellular functions (perhaps some aspects of differentiation and development) will be mediated by prions. The yeast prions determine inherited traits; could some inherited traits of mammals be determined by prions?