Equilibrium folding properties of the yeast prion protein determinant Ure2¹

Abstract

The yeast non-Mendelian factor [URE3] propagates by a prion-like mechanism, involving aggregation of the chromosomally encoded protein Ure2. The [URE3] phenotype is equivalent to loss of function of Ure2, a protein involved in regulation of nitrogen metabolism. The prion-like behaviour of Ure2 in vivo is dependent on the first 65 amino acid residues of its N-terminal region which contains a highly repetitive sequence rich in asparagine. This region has been termed the prion-determining domain (PrD). Removal of as little as residues 2–20 of the protein is sufficient to prevent occurrence of the [URE3] phenotype. Removal of the PrD does not affect the regulatory activity of Ure2. The C-terminal portion of the protein has homology to glutathione S-transferases, which are dimeric proteins. We have produced the Ure2 protein to high yield in Escherichia coli from a synthetic gene. The recombinant purified protein is shown to be a dimer. The stability, folding and oligomeric state of Ure2 and a series of N-terminally truncated or deleted variants were studied and compared. The stability of Ure2, ΔG_{D-N, H2O}, determined by chemical denaturation and monitored by fluorescence, is 12.1(±0.4) kcal mol⁻¹ at 25 °C and pH 8.4. A range of structural probes show a single, coincident unfolding transition, which is invariant over a 550-fold change in protein concentration. The stability is the same within error for Ure2 variants lacking all or part of the prion-determining domain. The data indicate that in the folded protein the PrD is in an unstructured conformation and does not form specific intra- or intermolecular interactions at micromolar protein concentrations. This suggests that the C-terminal domain may stabilise the PrD against prion formation by steric means, and implies that the PrD does not induce prion formation by altering the thermodynamic stability of the folded protein.

Introduction

The concept of an infectious protein, i.e. a protein with an altered conformation that can coerce normal protein molecules to adopt a similar rogue conformation, has captured the imagination and directed the research efforts of scientists from a broad range of fields over a number of decades Pattison and Millson 1960, Alper et al 1966, Griffith 1967, Gajdusek 1977, Prusiner et al 1984, Chesebro et al 1985, Bueler et al 1993, Jarrett and Lansbury 1993, Kocisko et al 1994, Wickner 1994, Chernoff et al 1995, Riek et al 1996, Bruce et al 1997. Whether proteins truly have the ability to convey inheritable information remains a clouded issue Aguzzi and Weissman 1997, Chesebro 1998, Farquhar et al 1998. The mammalian diseases, which include kuru and Creutzfeld-Jacob disease (CJD) in humans, scrapie in sheep, and bovine spongiform encepholopathy (BSE), are all associated with the gene product, PrP. The recognition of an apparently similar phenomenon in fungi, represented by [URE3] and [PSI] in yeast (Wickner, 1994), and [HetS] in a filamentous fungus (Coustou et al., 1997), lends credence to the notion that such infectious proteins, or prions, might exist as a general phenomenon in nature.

The genetic properties of [URE3], a cytoplasmically inherited genetic element in bakers’ yeast, suggest that it is a prion form of the chromosomally encoded protein Ure2 (Wickner et al., 1999). In Saccharomyces cerevisiae, Ure2 has a role in negative regulation of nitrogen metabolism so that the yeast cannot utilise the compound ureidosuccinate in the presence of ammonia Lacroute 1971, Drillien and Lacroute 1972. Mutants in the URE2 gene cause this negative regulation to be lost. In genetic crosses [URE3] exhibits non-Mendelian segregation, and the determinant for the [URE3] phenotype has been shown to be carried in the cytoplasm Lacroute 1971, Aigle and Lacroute 1975, Wickner 1994. Cells can be cured of the [URE3] phenotype by a variety of treatments. Unlike non-Mendelian genetic elements which consist of nucleic acid, [URE3] can spontaneously recur after cure. Other aspects of [URE3] genetics at odds with a nucleic acid determinant are as follows: [URE3] requires the presence of a chromosomal gene URE2 for its propagation; the absence of the URE2 gene has the same phenotype as the presence of the [URE3] element (Aigle & Lacroute, 1975); and overproduction of Ure2 (or its N-terminal “prion-inducing” domain) increases the frequency with which [URE3] arises Wickner 1994, Masison and Wickner 1995.

Ure2 possesses an N-terminal region of unusual amino acid composition. Genetic studies have shown that the first 65 residues of the N-terminal region are required for expression of the prion phenotype (Masison & Wickner, 1995), and the prion phenotype has been shown to correspond to aggregation of Ure2 protein in the cell (Edskes et al., 1999). Overexpression of the region 1–65 induces the prion phenotype (Masison & Wickner, 1995), and so this region has been termed a prion-inducing or “prion-determining” domain (PrD). Overexpression of the slightly longer region 1–80 is 45 times more efficient at inducing the prion phenotype than just the PrD (Masison et al., 1997). Deletions of certain portions of this region (residues 2–20, 2–40 or 3–65), or a frameshift of one base to change the protein sequence between residues 44 and 80, are also sufficient to remove the prion-forming ability of the protein (Masison et al., 1997). The boundary at residue 65 was chosen at the DNA level on the basis of a convenient endonuclease restriction site (Masison & Wickner, 1995). The boundaries in the present study were chosen by examination of the amino acid sequence (see Figure 1). Sequence analysis suggests that the boundary between the N-terminal and C-terminal portions of the protein lies between residues 90 and 111 (Rossjohn et al., 1996). The SEG algorithm (Wootton, 1994) identifies the regions 1–14 and 43–73 as being “unusual” (i.e. highly repetitive) in sequence. These two regions are particularly rich in asparagine (approximately 50 %).

The C-terminal portion of Ure2 is homologous to the glutathione S-transferase (GST) family. GSTs were initially defined by their enzymatic activity, that of conjugating glutathione to electrophilic substances to reduce their toxicity. GST-like proteins with other functions have been identified on the basis of sequence analysis, and these include a variety of enzymes in bacteria (reviewed by Vuilleumier, 1997) and Ure2 in yeast (Coschigano & Magasanik, 1991). The PrD does not affect the activity of the GST region of Ure2 unless covalently attached, and the presence of the GST region appears to stabilise the PrD against prion formation. In addition, the presence of the PrD improves the nitrogen catabolite repression activity of the GST region Masison and Wickner 1995, Masison et al 1997. These observations suggest that there is a specific interaction between the PrD and GST regions of Ure2 which is facilitated by their covalent attachment.

The aim of this study was to investigate the properties of the Ure2 protein, by measuring its stability, folding and oligomeric state in vitro. In addition, we wished to establish whether the presence of the N-terminal prion-determining domain affects the properties of the protein in isolation from cellular co-factors, and hence whether prion-like behaviour can be attributed to the intrinsic properties of the protein. Ure2 protein was overproduced from a synthetic gene, allowing high-level expression in Escherichia coli. As there is ambiguity about where the PrD region ends and where the GST region begins, a series of N-terminally truncated variants were constructed beginning at residues 74, 90 and 111 (referred to as 74Ure2, 90Ure2 and 111Ure2, see Figure 1). Two further Ure2 variants were constructed in order to dissect out which regions of the N-terminal domain might affect aggregation or folding: 15Ure2, which lacks the first repetitive region of 14 amino acid residues, and Δ15-42Ure2 in which the island of “normal” amino acid sequence between residues 15 and 42 has been deleted (see Figure 1). The proteins were purified, and their properties studied by a variety of biophysical techniques.

Removal of the prion-determining region to residue 15, 74 or 90 did not affect the thermodynamic stability or oligomeric state of the protein. Deletion of residues 15–42, likewise, had no affect on the stability of the protein. This study demonstrates that any difference in behaviour conferred by the PrD is not due to a difference in thermodynamic stability of the dimeric proteins, as measured by guanidinium chloride (GdmCl) denaturation using intrinsic fluorescence as a structural probe. A variety of structural probes showed a single, coincident unfolding transition that was unaffected by a 550-fold change in protein concentration. This suggests that unfolding and dimer dissociation are not closely coupled. The data presented here provide the groundwork for a full elucidation of the folding pathway of Ure2, and to establish whether the prion behaviour is determined by the intrinsic properties of the protein.