Prion Domain Initiation of Amyloid Formation in Vitro from Native Ure2p

See allHide authors and affiliations

Science  26 Feb 1999:
Vol. 283, Issue 5406, pp. 1339-1343
DOI: 10.1126/science.283.5406.1339


The [URE3] non-Mendelian genetic element of Saccharomyces cerevisiae is an infectious protein (prion) form of Ure2p, a regulator of nitrogen catabolism. Here, synthetic Ure2p1−65 were shown to polymerize to form filaments 40 to 45 angstroms in diameter with more than 60 percent β sheet. Ure2p1−65 specifically induced full-length native Ure2p to copolymerize under conditions where native Ure2p alone did not polymerize. Like Ure2p in extracts of [URE3] strains, these 180- to 220-angstrom-diameter filaments were protease resistant. The Ure2p1−65-Ure2p cofilaments could seed polymerization of native Ure2p to form thicker, less regular filaments. All filaments stained with Congo Red to produce the green birefringence typical of amyloid. This self-propagating amyloid formation can explain the properties of [URE3].

Genetic evidence identified [URE3] and [PSI], two nonchromosomal genes of Saccharomyces cerevisiae, as prions of Ure2p and Sup35p, respectively, which implies that proteins can be hereditary material (1). In response to a good nitrogen source (ammonia or glutamine), Ure2p blocks assimilation of poor nitrogen sources by blocking the action of the transcription regulator Gln3p (2). Sup35p is a subunit of the translation release factor (3). [URE3] (4) and [PSI] (5) are altered forms of Ure2p and Sup35p that have lost their normal functions but have acquired the ability to convert their normal forms into the altered (prion) form (1), a notion supported by genetic and biochemical data (6–10). The prion concept originates in studies of the spongiform encephalopathies (11), believed due to a self-propagating altered form of PrP that forms scrapie-associated filaments and amyloid deposits in brains of affected animals (12).

Amyloid is defined as a filamentous protein structure that stains with the dye Congo Red (CR) to produce green birefringence under polarized light and is characterized by protease resistance and an antiparallel β sheet structure (13). Amyloid deposits of the Aβ peptide accumulate in the brain of patients with Alzheimer's disease, and amyloid deposits of overproduced monoclonal immunoglobulins are found in patients with certain B cell neoplasms.

The 354-residue Ure2p is normally a dimer (14). Overexpression of the first 65 residues of Ure2p (the prion domain) induces de novo formation of the [URE3] prion at >1000-fold the spontaneous rate (6). Deletion of the first 65 residues leaves a COOH-terminal fragment (the nitrogen regulation domain) competent in nitrogen regulation but unaffected by the [URE3] prion (6). The prion domain can also propagate [URE3] in the absence of the nitrogen regulation domain (10). Thus, the NH2-terminal 65 residues are necessary for Ure2p to be altered to the prion form and sufficient to induce the change in normal Ure2p.

Synthetic Ure2p1−65 (15) in 6 M guanidine (16) was diluted 1:40 into buffer. A fine precipitate was first visible at 20 min and was complete after 1 hour, with >90% of the peptide in the precipitate. Negative staining electron microscopy of the precipitate showed thin straight filaments (Fig. 1A), which were variable in width because of lateral bundling of variable numbers of narrow protofilaments. Individual protofilaments are uniformly 40 to 45 Å in diameter.

Figure 1

Electron microscopy of negatively stained filaments formed by polymerization of Ure2p and its prion domain Ure2p1−65. (A) Filaments of Ure2p1−65, consisting of parallel bundles of 45 Å protofilaments. (B) Cofilaments of Ure2p and Ure2p1−65. (C) Cofilaments after digestion with proteinase K (20). Bar = 500 Å. (D) Filaments produced by seeding a solution of 66.4 μg of Ure2p with 1.7 μg of copolymer filaments. Drops of samples at 13 μM protein or peptide [6.5 μM for sample (C)] were applied to freshly glow-discharged carbon-collodion films mounted on electron microscope grids, negatively stained with 1% uranyl acetate, and observed in a Philips CM120 electron microscope. Although the axial regularity of filaments in (A) and (B) implies that they are ordered structures, no axial repeats were detected by Fourier analysis of the images (19). A report of filament formation by the NH2-terminal domain of Sup35p aged for 1 week at pH 2 (25) also alludes to filament formation by the Ure2p prion domain under these conditions, but these filaments were not described.

In an attempt to reproduce in vitro the in vivo induction of the [URE3] prion by Ure2p1−65 , we diluted denatured synthetic Ure2p1−65 into purified native Ure2p (17). Equimolar amounts of Ure2p1−65precipitated about 80% of Ure2p in <4 hours (Fig. 2). When 10- and 100-fold less Ure2p1−65 was added, precipitation of Ure2p decreased by 9.4- and 30-fold, respectively, indicating that the aggregate was roughly a 1:1 mixture of protein and peptide. When we added equimolar Ure2p1−65 to bovine serum albumin or a mixture of standard proteins (16), none of these proteins entered the peptide precipitate (Fig. 2), which suggests that Ure2p1−65 interacts specifically with Ure2p. When we added nonprion domain peptide residues 148 to 213 of Ure2p to Ure2p in equimolar amounts, no precipitate was formed. When we used Aβ peptide, the major component of amyloid in Alzheimer's disease, to generate amyloid filaments in vitro under the same conditions (16, 18), Ure2p was not incorporated into the precipitate (Fig. 2), again showing the specificity of the interaction of Ure2p1−65 with native Ure2p.

Figure 2

Filament formation by Ure2p1−65specifically precipitates intact native Ure2p. Native Ure2p alone, or with a mixture of other proteins, was incubated with Ure2p1−65 (16). Washed precipitates were analyzed by SDS-PAGE and stained with Coommassie blue. Lanes 12 to 14 show total amounts added to mixtures; other lanes show only the precipitate. BSA, bovine serum albumin.

Electron microscopy of Ure2p1−65-Ure2p copolymers (Fig. 1B) revealed filaments 180 to 220 Å wide, much wider than the prion domain filaments (Fig. 1A). The copolymer filaments were not hollow, as judged by the absence of a stain-penetrable lumen. No filaments were observed in soluble Ure2p without Ure2p1−65, and Ure2p formed only amorphous aggregates when heated at 100°C for 5 min (19).

Ure2p is protease-resistant in extracts of [URE3] strains compared to isogenic [ure-o] strains (6). We compared protease resistance of filaments of Ure2p1−65 and of Ure2p1−65-Ure2p with that of native Ure2p and heat-denatured Ure2p (20) (Fig. 3). Native Ure2p was digested in <2 min, whereas the filaments were digested more slowly. The 7-kD Ure2p1−65 persisted for >20 min. In the Ure2p1−65-Ure2p sample, fragments of 30 to 32 kD persisted for about 2 min and fragments of 7 to 10 kD—larger than the 7-kD Ure2p1−65—persisted for over 20 min. Heat-denatured Ure2p showed lower overall protease resistance and lacked the persistent 7- to 10-kD species. The pattern of protease resistance in the Ure2p1–65-Ure2p fiber is similar to that observed in [URE3] extracts (6), indicating that a change, similar to that found in [URE3] strains, occurred when Ure2p was precipitated by Ure2p1−65.

Figure 3

Proteinase K resistance of Ure2p incorporated into filaments with Ure2p1−65(20). Heat-denatured Ure2p, untreated native Ure2p, Ure2p1−65 filaments, cofilaments formed by an equimolar mixture of Ure2p1−65 and native Ure2p, and filaments of Ure2p seeded by cofilaments were treated with proteinase K for various times and the products were analyzed by immunoblotting with a Ure2p antibody specific for the NH2-terminus. The seeded filaments of Ure2p (last sample) consisted of 1.7 μg of seed and a further 38 μg of precipitated full-length Ure2p.

Electron microscopic examination of the protease-treated Ure2p1−65-Ure2p cofilaments showed filaments of variable width (Fig. 1C) but generally narrower than the starting material. Like the prion domain protofilaments (Fig. 1A), the narrowest segments were about 45 Å wide, straight, and lacked visible substructure. Thus a copolymer filament apparently had a backbone of a prion domain protofilament surrounded by the remainder of Ure2p. The latter portion was digested by proteinase K, with accumulation of the resistant 7- to 10-kD fragment.

Without the green birefringence, binding CR is not indicative of amyloid structure (21). When stained with CR (22), all Ure2p precipitates appeared red under bright field, indicating that they bound the dye, but under polarized light the heat-denatured Ure2p appeared dark (19), whereas Ure2p1−65-Ure2p and Ure2p1−65 filaments had an apple-green birefringence (Fig. 4). This showed that they were amyloid fibers, unlike the heat-denatured material, which, as observed by electron microscopy, was an amorphous precipitate.

Figure 4

CR birefringence of Ure2p1−65, Ure2p1−65- Ure2p, and Ure2p fibers seeded by mixed filaments. Filaments were stained with CR and observed at ×100 magnification under bright field (left) or by polarization microscopy (right) (22). Bar = 20 μm.

Ure2p1−65-Ure2p and Ure2p1−65 filaments bound 14.8 and 3.0 CR molecules per monomer, with dissociation constant (K d) values of 1.7 and 0.81 μM, respectively (23). Dye bound to both precipitates also exhibited a spectral shift with a maximum difference at 540 nm, which is typical of amyloid fibers.

Native Ure2p contains both α helix and β sheet (Table 1). The Ure2p1−65 filaments were predominantly β sheet with little or no helix structure. The Ure2p1−65-Ure2p copolymer was also high in β sheet. Assuming that Ure2p1−65 has the same structure in the copolymer as in the Ure2p1−65 filaments, we estimate that Ure2p in the cofilaments has significantly increased β sheet and decreased α helix compared with its native structure (Table 1).

Table 1

Secondary structure of Ure2p preparations. Raman spectroscopy, data collection, and analysis have been described (31). Protein solutions were 10 to 30 mg/ml in 50 mM phosphate buffer (pH 8.0) containing 0.2 M NaCl. The Raman amide I and amide III spectra were analyzed for secondary structure information by the nonnegative least-squares approach.

View this table:

Addition of 3 μl (1.7 μg) of Ure2p1−65-Ure2p cofilaments to 123 μl (66.4 μg) of native Ure2p resulted, after 2 weeks at 4°C, in precipitation of 62% of the native Ure2p. Filaments 290 to 400 Å in diameter (Fig. 1D) that showed green birefringence with CR were observed (Fig. 4). Proteinase K digestion of these filaments produced a pattern of resistant fragments similar to that observed with Ure2p1−65-Ure2p cofilaments (Fig. 3). The amounts of resistant fragments observed were far in excess of the small amount of seed filament used, showing that the protease-resistant material was not simply carried over from the seed filaments. Thus the Ure2p amyloid formation is a seeded process capable of continued propagation in vitro.

We suggest that the filament structure formed by Ure2p1−65also forms the core of the equimolar cofilament with intact Ure2p and of the propagated fiber composed mainly of intact Ure2p, with the prion domain stacking in β sheets and the COOH-terminal domain protruding to form the wavy, thicker structure.

Does the amyloid formation demonstrated here for Ure2p1−65and the Ure2p1−65-Ure2p mixture correspond to the [URE3] prion? The in vitro amyloid fibers show a pattern of protease resistance like that observed for Ure2p in extracts of [URE3] strains. Moreover, it is specifically the prion domain (residues 1 to 65) that induces the amyloid formation with intact native Ure2p. Neither Ure2p148–213 (part of the nitrogen regulation domain) nor the Alzheimer's Aβ peptide shows this activity, although the latter does form amyloid. The seeding by cofilaments of amyloid formation by native Ure2p suggests that, as in vivo, the in vitro reaction can propagate. Further, Ure2p–green fluorescent protein fusion proteins have been found to form intracellular aggregates in vivo specifically in [URE3] cells (24). We suggest that, in [URE3] cells, amyloid filaments recruit most of the Ure2p in the cytoplasm. Ure2p in filaments is inactive or unable to enter the nucleus. Mating of [URE3] and [ure-o] cells transmits filaments to the cytoplasm of all progeny cells, which seed further filament formation. However, final proof that this is [URE3] requires characterization of Ure2p from [URE3] strains and transmission of [URE3] to yeast cells by amyloid filaments formed in vitro.

At pH 2 the Sup35p prion domain forms filaments that show birefringence with CR and slightly increased protease resistance (25). Full-length Sup35p, made in Escherichia coli, also forms filaments, but these show no green birefringence with CR (26). Sup35p is aggregated in [PSI] strains (8, 27), but whether it is in an amyloid form has not been reported.

In Alzheimer's disease, a peptide fragment of a large precursor protein forms the amyloid. In vitro amyloid formation by the monoclonal light chains produced by some multiple myelomas occurs only after partial proteolysis (28). In parallel with in vivo results (6, 10), we found that native Ure2p did not form amyloid filaments (29) unless the prion domain peptide was provided. The full protein structure may prevent filament formation in these cases, and [URE3] induction by overproduction of the full-length protein may be initiated by proteolytic fragments. Our results imply that [URE3] is an infectious amyloidosis and suggest new approaches to the study of these diseases.

  • * To whom correspondence should be addressed. E-mail: wickner{at}


View Abstract

Navigate This Article