Heterologous cross-seeding mimics cross-species prion conversion in a yeast model

Background Prions are self-perpetuating, infectious, aggregated proteins that are associated with several neurodegenerative diseases in mammals and heritable traits in yeast. Sup35p, the protein determinant of the yeast prion [PSI+], has a conserved C terminal domain that performs the Sup35p function and a prion domain that is highly divergent. Prions formed by chimeras of the prion domain of various species fused to the C domain of Saccharomyces cerevisiae exhibit a 'species barrier', a phenomenon first observed in mammals, and often fail to transmit the prion state to chimeras with prion domains of other species. Results We focus on the chimera containing the prion domain of Pichia methanolica and examine how tight the 'species barrier' is between the chimera and S. cerevisiae. Although either of two Q/N-rich prions, [PSI+] or [PIN+], enhances the formation of the chimeric prion, [CHI+PM], neither a non-Q/N-rich prion nor a non-prion Q-rich aggregate promotes the formation of [CHI+PM]. [CHI+PM] has many features characteristic of yeast prions: aggregation, cytoplasmic transmission and a two-level protein structure. [CHI+PM] formed in the presence of [PSI+] can propagate independently of [PSI+] and forms at least two different variants of the prion, suggesting the generation and not transmission of new prion seeds. Conclusion Although the sequence similarity between the S. cerevisiae Q/N-rich prion determinants and the P. methanolica prion domain is low, we find that the chimera containing the prion domain of P. methanolica can occasionally be cross-seeded by [PSI+] to mimic crossing the species barrier, to form the [CHI+PM] prion. Our data suggests that crossing the barrier occurs by a de novo formation of the foreign chimeric prion. Thus, the species barrier appears to be crossed by a heterologous seeding mechanism, wherein the infected prion protein uses the pre-existing seed as an inefficient template.


Background
The idea that only nucleic acid elements could transfer genetic information was challenged by the discovery of prions [1]. Prions are the causative agents of several neurodegenerative diseases such as Creutzfeldt-Jakob disease (CJD) in humans, scrapie in sheep and bovine spongi-form encephalopathy in cattle [2]. Once the prion protein (PrP) converts to its prion PrP Sc form, which is in a largely β-sheet-rich, aggregated, amyloid state, it induces the conversion of normally folded, soluble, mainly α-helical PrP C into the PrP Sc form [3]. Thus, the protein structure is passed at the protein, and not at the nucleic acid level.
Although PrP Sc converts PrP C into the prion form, the propagation of the PrP Sc amyloid is highly specific. PrP Sc from one species can rarely convert PrP C from another species into the prion form. This phenomenon, known as the species barrier, has been used to explain the lack of transfer of the disease from scrapie-infected sheep to humans [4]. However, a novel form of CJD is believed to have emerged from the conversion of human PrP to its prion form by ingested bovine prions [5]. Studies of laboratory PrP Sc have shown that the some species barriers can be crossed whereas others are more rigid [4]. The barrier has been attributed not just to the difference in the primary sequence of PrP between species, but more importantly to the protein conformations that primary sequence is capable of adopting [6,7]. This difference in conformation of PrP Sc is believed to give rise to different phenotypes of prion diseases, known as strains, that vary in characteristics such as time of incubation and patterns of neuropathology [8,9]. An in vitro study of mammalian proteins suggests that the barrier is crossed when the PrP primary sequence of a certain species is capable of adopting the conformation of a PrP Sc strain of another species [7].
The phenomenon of information transfer through a protein-only process is not limited to mammals. Indeed, several prions have been studied in yeast [10]. Unlike mammals, where only one prion protein of unknown function has been described, yeast contain several prion protein determinants [11,12]. They share no homology with the primary sequence of PrP, but share common features such as high β-sheet content, infectivity and amyloid characteristics [13]. In addition, yeast prions also exist as different strains, called variants, that appear to differ in their amyloid conformations, which leads to phenotypic and biochemical variations [14][15][16][17][18][19][20][21][22].
Two extensively studied yeast prions [PSI + ] and [PIN + ] are altered, aggregated forms of Sup35p and Rnq1p, respectively [11,23,24]. Sup35p is a translation termination factor and the function of Rnq1p is unknown [23,25]. [PSI + ] cells have reduced translation termination efficiency, as Sup35p is mainly in the aggregated form, whereas [psi -] cells that have non-prion Sup35p terminate efficiently [26]. Mendelian mutants of SUP35 also have reduced efficiency of translation termination, not because of aggregated Sup35p, but due to the mutation [27]. However, [PSI + ] can be distinguished from sup35 mutants, as Sup35p is aggregated only in [PSI + ] strains [28,29]. Sup35p has three distinct domains: a C-terminal domain (C) that performs the function of translation termination, and a N-terminal (N) and middle (M) domain that are required for the induction and faithful propagation of [PSI + ] [14,[29][30][31][32].
The [PIN + ] prion is not a loss-of-function prion, but rather [PIN + ] cells have the ability to induce [PSI + ] more efficiently than [pin -] cells [33]. A cross-seeding model has been proposed to explain this phenomenon where the [PIN + ] prion acts as an inefficient seed for the de novo formation of the [PSI + ] prion [24,34]. Furthermore, different variants of [PSI + ], analogous to mammalian PrP Sc strains, are induced in [PIN + ] cells by the overexpression of either the full length Sup35p or more efficiently by the Sup35NMp [14,29,35].
Like mammalian prions, yeast prions exhibit barriers across species [36][37][38][39][40][41]. [PSI + ] formed by Sup35p from one species rarely passes the prion conformation to Sup35p from other species. Sup35p itself has retained its modular architecture in many species. Various species of yeast have the N, M and C domains where the sequence of the C domain is highly conserved between species, in contrast to the NM domains, which are often highly divergent [38]. However, the N domains of various species share the common features of having a high Q/N-rich content and oligopeptide repeats [38].
Chimeras of the NM from various species fused to the C domain of Saccharomyces cerevisiae, although able to form and propagate as prions in S. cerevisiae, fail to transmit the prion state to S. cerevisiae Sup35p or other chimeras [36,38,39]. This failure to pass the prion state is attributed to the highly divergent sequence of the NM domains. However, a certain variant of S. cerevisiae [PSI + ] is able to transmit the prion state to the chimera of the NM of Candida albicans and C domain of S. cerevisiae (NM CA -C SC ) [6] even though S. cerevisiae and C. albicans share only around 40% similarity in their NM domains. Thus, the variant of the prion is important for determining transmission across a species barrier. Other studies have shown that prion domains of other species, such as S. bayanus and S. paradoxus, which are much more similar to NM SC , and that co-aggregate with S. cerevisiae [PSI + ] ([PSI + ] SC ), still exhibit low transmission of the prion state to these foreign species [39].
In this study we focus on the chimera of the NM of Pichia methanolica and C domain of S. cerevisiae (NM PM -C SC ). NM PM -C SC as the sole copy of Sup35p can be induced and propagated in its prion state, [CHI + PM ] (for chimeric [PSI + ], PM for P. methanolica), in S. cerevisiae [36,37]. Like other chimeras it does not transmit its prion state en masse to S. cerevisiae Sup35p, although overexpression of NM PM can induce the formation of S. cerevisiae [PSI + ], albeit at a lower frequency than by the overexpression of NM SC . However, overexpression of NM SC fails to induce NM PM -C SC into a prion. Prion conversion by overexpressed NM PM is specific to S. cerevisiae Sup35p as it fails to induce the prion form of chimeras from other species [38].
When NM PM -C SC or NM CA -C SC are expressed at the same level as S. cerevisiae Sup35p, neither the NM PM -C SC nor the NM CA -C SC chimeras are frequently infected by [PSI + ] SC to become [CHI + PM ] [38]. We examine how tight the barrier is and find that NM PM -C SC but not NM CA -C SC occasionally converts into [CHI + PM ] in the presence of either of at least two Q/N-rich prions, [PSI + ] or [PIN + ]. We propose that the species barrier can be crossed by a heterologous seeding mechanism similar to that of the cross-seeding between the [PIN + ] prion and Sup35p.

NM PM -C SC but not NM CA -C SC is inactivated in the presence of [PSI + ] at a frequency of 10 -4 to 10 -3
As previously observed, chimeras of the prion domains of either P. methanolica (NM PM ) or C. albicans (NM CA ) fused to the C domain of S. cerevisiae (C SC ) were functional in the presence of [PSI + ] aggregates [38] ( Figure 1A). This is phenotypically monitored using a yeast strain that has the ade1-14 allele with a suppressible nonsense mutation [26,42]. When the foreign fusions (containing an HA tag between NM and C) were ectopically expressed from a plasmid in [PSI + ][pin -] ade1-14 cells at a moderate, constitutive level, the fusions remained functional and these cells were red on low adenine media (see Methods) and

A B
could not grow on -Ade ( Figure 1A). To determine whether the fusions get inactivated occasionally, [PSI + ] cells containing the fusions were plated onto plasmid selective -Ade media. The formation of Ade + colonies showed that the NM PM -C SC fusion was inactivated in [PSI + ] cells at a frequency of about 10 -4 to 10 -3 ( Figure 1B, Table 1). In contrast, the NM CA -C SC fusion was not inactivated in [PSI + ] cells ( Figure 1B). As a control these fusions were expressed in the presence of a sup35 mutant strain that, like [PSI + ], can grow on -Ade, but unlike [PSI + ] does not cause Sup35p to aggregate. The NM PM -C SC fusion was not inactivated in the sup35 mutant yeast ( Figure 1A and 1B).

NM PM -C SC Ade + colonies have prion properties
To determine if the inactivation of NM PM -C SC was due to its prionization, we tested for features characteristic of prions: aggregation and cytoplasmic inheritance. We denote NM PM -C SC in its active form as [chi -PM ] and in its inactive form as [CHI + PM ].
If NM PM -C SC were in its prion form, most of the protein would be expected to be aggregated. The aggregation state was tested both biochemically and visually. When [PSI + ] or [PIN + ] cells are subjected to high-speed centrifugation most of the protein is in the pellet fraction, whereas in cells lacking the prion most of the protein is in the supernatant fraction [17,28,29]. Similarly, NM PM -C SC in [CHI + PM ] lysates subjected to high-speed centrifugation was present in the pellet fraction and absent from the supernatant fraction ( Figure 2A). In contrast, lysates from [chi -PM ] cells had most of the NM PM -C SC in the supernatant fraction ( Figure 2A).
The [PSI + ] and [PIN + ] prions are not dissolved into monomers when treated with sodium dodecyl sulfate (SDS) in the absence of boiling, but break into SDS-stable sub-particles that can be resolved on agarose gels [43,44]. Like-wise, NM PM -C SC in [CHI + PM ] lysates was not dissolved into monomers when treated with unheated 2% SDS, whereas NM PM -C SC from [chi -PM ] lysates remain as monomers (Figure 2B).
In vivo, transiently overexpressed green fluorescent protein (GFP)-tagged prion domains are observed as distinct puncta in prion-containing cells, whereas cells lacking the prion show diffuse fluorescence [23,29,45]. Similarly, when NM PM -GFP was overexpressed transiently [CHI + PM ] cells showed punctuate dots, whereas the fluorescence in [chi -PM ] cells was diffuse ( Figure 2C). Thus the chimeric protein (NM PM -C SC ) is in an aggregated state in the Ade + colonies.
One of the characteristic features of prions is that they are passed from cell to cell through cytoplasmic mixing without a nuclear contribution. This is achieved by cytoduction, which involves mating donor and recipient yeast in the presence of a kar1 mutation that inhibits efficient nuclear fusion. Daughter cells with the recipient haploid nucleus and a mixture of the parental cytoplasms can be selected (cytoductants) [46].   Figure 4). These variants, as in the case of [PSI + ], could sometimes be differentiated biochemically by the size of their sub-particles [43] ( Figure 4B). Unlike [PSI + ] variants but like at least one strong hybrid [CHI+ PM ] variant (where NM PM -C SC is the sole copy of Sup35p in the cell), the strong chimeric variant was associated with the larger sub-particles whereas the small sub-particles were associated with the weaker variant ( Figure 4B) [43]. Some strong and weak [CHI + PM ] variants distinguished on the basis of color could not be differentiated by a change in the size of subparticles, but did differ in the amount of sub-particles present ( Figure 4A).  : HtQ72, HtQ103), which is associated with Huntington's disease, fused to GFP was expressed in a sup35 mutant strain along with NM PM -C SC . HtQ72-GFP and HtQ103-GFP, shown to aggregate in wild type strains [47], also aggregated in the sup35 mutant, but failed to enhance the conversion of NM PM -C SC from a [chi -PM ] to a [CHI + PM ] state (Table 1). We also tested whether a non-Q/ N-rich prion, [Het-s] y , a prion from Podospora anserine that can propagate as a prion in yeast [48], could enhance the formation of [CHI + PM ]. The prion domain of HET-s fused to GFP (the protein determinant of [Het-s] y ), was induced into the prion form ( [48] and see Methods) in a sup35 mutant expressing NM PM -C SC , but failed to enhance the formation of [CHI + PM ] (Table 1).

Discussion
Although the degree of similarity is not high between the prion domains of S. cerevisiae and either P.   [43,44]: NM PM -C SC appears to be assembled into large aggregates (resolved by high-speed centrifugation) that can be broken into SDS-stable sub-particles. Additionally, we show the [CHI + PM ] phenotype can be transferred via cytoplasm, which is common to all known yeast prions [11,23,24,42,49]. Thus, we show that although the transmission of the prion state across species barriers is not very efficient, the prion state can be transferred occasionally to give rise to a foreign prion.  [39], suggesting that interaction between endogenous Sup35p and the chimeric protein is limited. Species that have NM domains that are much more similar to S. cerevisiae, such as S. bayanus and S. paradoxus, have been shown to co-aggregate but there is no transmission of the prion state [39]. This suggests that tight interactions might actually hamper the formation of the heterologous prion and that heterologous prion formation might be mediated by transient interactions between the seed and the prionizing protein.
Indeed, in vitro PrP in the non-fiber form from one species is capable of binding PrP in the fiber form of another species, but there is no conversion of the non-fiber form to the fiber form [51]. Thus, stable interactions might actually be inhibitory to the process of heterologous prion conversion.  PM ] formation seems limited to the two Q/Nrich prions that we tested. Aggregates of GFP fused to HtQ103p, the mutated first exon of the Huntingtin protein (HtQ103), whose aggregation is associated with Huntington's disease [52] and enhances de novo formation of [PSI + ] [34], fail to enhance the formation of [CHI + PM ]. Furthermore, a non-Q/N-rich prion [Het-s] y , a prion from P. anserine that can propagate in yeast, that is induced twofold higher in [PIN + ] cells [48], also failed to enhance [CHI + PM ] formation, suggesting that although Q/ N-rich prions can enhance the formation of non-Q/N-rich prions, non-Q/N-rich prions do not always enhance the formation of Q/N-rich prions in vivo. In fact, in vivo non-Q/N-rich amyloids failed to enhance the de novo formation of [PSI + ] [34], suggesting that cross talk between Q/ N-rich and non-Q/N-rich prions might occur only in one direction.

NM PM -C SC is not incorporated into [PSI + ] sub-particles
Two models have been proposed to explain the ability of heterologous prions to enhance the de novo formation of other prions: titration of inhibitory factors by the heterologous prion, and direct cross-seeding by the heterologous prion [24,34,53]. One or both of these mechanisms could be playing a role in the cross-seeding activity. In vitro evidence supports the direct cross-seeding model, as many Q/N-rich and non-Q/N-rich amyloids have been shown to stimulate the aggregation of Sup35p [34]. Since [CHI + PM ] formation is specific to Q/N-rich prions, we suggest that de novo formation of this prion requires interactions between Q/N-rich domains. Several studies have shown that Q and N residues play an important role in the initial step of amyloid formation [54], which might be essential for the formation of [CHI + PM ]. Since we see no increase in [CHI + PM ] formation in the presence of the nonprion Q/N-rich HtQ103 amyloid, we propose that interaction of resident prion propagating/enhancing factors that associate with the prion cross-seed might help stabilize the newly forming [CHI + PM ] seed. For example Hsp104, a chaperone required for the propagation of all known yeast prions [12,26,33,55], binds preferentially to [PSI + ] aggregates versus non-prion Sup35p [56] Although prion variants play an important role in the transmission of the prion state across a species barrier, two studies show that variants cause slightly differing results. In the case of mammalian in vitro-made fibers, Syrian hamster PrP (23-144) fibers were able to cross-seed mouse PrP (23-144) protein but not vice versa. However, Syrian hamster-seeded mouse fibers had properties of the Syrian hamster seed and not that of spontaneously formed mouse fiber [7]. In the case of yeast, one specific variant of [PSI + ] SC was able to cross the barrier and infect the NM domain of C. albicans (NM CA ) to give rise to a novel variant of the prion form of NM CA -C SC [6]. We see that although NM PM and NM SC have very little similarity, NM PM -C SC is infected by [PSI + ] SC to give rise to not one but at least two variants of [CHI + PM ]. We suggest that the chimeric foreign protein NM PM -C SC is heterologously crossseeded to form [CHI + PM ] de novo, giving rise to different variants. In the case of homologous seeding of Sup35p to form [PSI + ], studies have shown that short peptide sequences mediate initial nucleation to give rise to amyloid fibers [57]. We propose that the variant of the infecting prion is important as different peptide sequences may be exposed that allow different foreign protein sequences to interact to lead to prionization. Thus the [PIN + ] prion, that has such low similarity with NM PM , might have short stretches of peptides that can interact with NM PM giving rise to [CHI + PM ]. Our data suggest that heterologous seeding events between proteins from different species might mimic a crossing of the species barrier.

Conclusion
We show here that in spite of low sequence similarity between the P. methanolica prion domain and the S. cerevisiae Q/N-rich prion determinants, the chimera can convert to its prion form, [

Plasmids
Centromeric plasmids pNM PM -C SC (p1180) and pNM CA -C SC (p1072) (kindly provided by Jonathan Weissman) have NM domains of the following species -PM: P. methanolica, CA: C. albicans, fused to the C domain of S. cerevisiae (C SC ), with an HA tag between the NM and C domains, under the SUP35 promoter, in pRS316 (URA3) [58]. The control vector used was pRS316 [58]. pNM PM -GFP (p1680) is a centromeric (URA3) plasmid in which the copper-inducible promoter controls the expression of NM PM -GFP (kindly provided by Yury Chernoff). NM PM -GFP was induced using 50 μM copper for 4 hours and observed using a Zeiss AxioScope2.

Yeast strains and media
The following yeast strains are derivatives of 74-D694 (MATa ade1-14 leu2-3,112 his3-Δ200 trp1-289  Standard yeast media and cultivation procedures were used [60]. Transformants were grown on synthetic dextrose (SD) lacking the appropriate amino acid. To monitor the efficiency of translational read through of ade1-14 transformants containing SUP35 fusion proteins, the color of cells grown on plasmid-selective, low adenine media with 0.13% yeast nitrogen base, 0.5% ammonium sulfate, 1% casamino acids, and 2% glucose, tryptophan, one quarter the required amount of adenine and no uracil was determined. This same media but with additional adenine and 5 mM guanidine hydrochloride (GuHCl) was used to cure the [CHI + PM ] prion. Synthetic glycerol (SG) -Ura containing 3 mg/liter cyclohexamide was used to select for cytoductants.
To isolate can1 R mutants, cells were plated on SD-Arg containing 60 mg/liter canavanine and resistant colonies were picked. To make the strain [rho -], cells were grown on complex glucose media (YPD) containing 0.05 mg/ml ethidium bromide [61]. To select for sup35 mutant cyto-ductants with different [PIN + ] variants, SG-Arg + canavanine was used.

Scoring for the formation of [CHI + PM ]
Previously described suppression assays were used to score for the formation of [CHI + PM ] [26]. Briefly, in [PSI + ] strains the premature stop codon in the ade1-14 allele is read through, allowing ade1-14 cells to grow on -Ade and causing them to be white on YPD. This is because most of the Sup35p is inactivated in [PSI + ] cells because it is sequestered into the prion aggregate. In [psi -] cells, Sup35p is available for efficient translation termination and thus [psi -] ade1-14 cells do not grow on -Ade and are red on YPD. A sup35 mutant (L2333) is also able to read through the ade1-14 premature stop codon, due to a Mendelian mutation in the SUP35 gene that impairs the activity of the Sup35p protein, and thus can grow on -Ade. L1763 (strong [PSI + ][pin -]), L2333 (sup35 mutant) or L2333 with [PIN + ] variants were transformed with plasmids (pNM PM -C SC , pNM CA -C SC , and control vector) and transformants were dissolved in water and spotted onto SD-Ura (to maintain the plasmid), low adenine (to monitor read through via color) or SD-Ura-Ade media (to monitor read through with growth on -Ade). This determined the functionality of NM PM -C SC in the [PSI + ] yeast. However, if a few NM PM -C SC molecules were inactivated, it would be difficult to monitor this in the above spot test. To test if NM PM -C SC was occasionally inactivated, cells taken from SD-Ura were dissolved in water and serially diluted. Larger numbers of cells (10 4 to 10 7 ) were plated onto SD-Ade-Ura and lower dilutions (10 1 to 10 3 ) were plated onto SD-Ura. Viability was determined by the number of cells on SD-Ura and the frequency of formation of [CHI + PM ] was determined by comparing the number of colonies on SD-Ura-Ade to the number on SD-Ura.

Effect of [Het-s] y on [CHI + PM ] formation
[Het-s] y was essentially induced and maintained as previously reported [48]. Briefly, pHET-s(PrD)-GFP was transformed into a sup35 mutant and transformants were grown on synthetic raffinose (SR)-Trp + 2% galactose to maintain the plasmid and induce HET-s(PrD)-GFP. This was then transferred to SR-Ura-Trp + 0.05% galactose to maintain [Het-s] y as dots. Either dot or diffuse HETs(PrD)-GFP cells were micromanipulated and propagated on SR-Ura-Trp + 0.05% galactose. NM PM -C SC or a control vector were transformed into [Het-s] y -containing strains and plated on SR-Ura-Trp + 0.05% galactose with or without adenine to monitor the formation of [CHI + PM ]. Cells with Het-s(PrD)-GFP dots retained dots in 70 to 90% of the cells whereas diffuse cells remained diffuse.

Donor [PSI + ][CHI + PM ] or [PSI + ][chi -
PM ] strains were mated to L2598, a cyh R strain defective in nuclear fusion (kar1). Both the donor and recipient strains contained pNM PM -C SC , expressing NM PM -C SC . Mating was done on SD-Ura to maintain pNM PM -C SC . Cytoductants were selected on SG-Ura + 3 mg/liter cyclohexamide. This media selects against the donor and diploids, as they cannot grow in the presence of cyclohexamide. Recipient cells cannot grow on SG as they lack mitochondria ([rho -]). Only cytoductants, recipient cells that have acquired mitochondria (and therefore cytoplasm) from the donor, can grow.

Protein analysis
To prepare cell lysates, strains were grown in appropriate media and harvested at an optical density of 1.5 to 2.0 (A 600 ). Crude cell lysates were prepared by physical disruption using glass beads (0.5 mm, Biospec, Bartlesville, OK, USA) in 750 μl of lysis buffer containing 50 mMTris/ HCl, pH 7.5, 50 mM KCl, 10 mM MgCl 2 and 5% (w/v) glycerol with protease inhibitor cocktail (P8215, 1:50, Sigma, St. Louis, MO, USA) and 5 mM PMSF. Cells were lysed by vortexing (Vortex-Genie 2) at high speed three times for 2 min each with 1 min in between in ice at 4°C. Crude lysates were pre-cleared by centrifuging at 3000 g for 5 min at 4°C to remove unlysed cells. The aqueous layer was used for further analysis.
To perform high-speed centrifugation analysis, approximately 600 to 800 μg of crude cell lysate in 300 μl was spun at 100,000 g for 30 min at 4°C. The supernatant was separated from the pellet fraction and the pellet fraction was dissolved in 300 μl of lysis buffer with protease inhibitor cocktail (Sigma, St. Louis, MO, USA) supplemented with 5 mM PMSF. Approximately 30 μl of the supernatant, the pellet and total protein each were mixed with 4× sample buffer (final concentration 62.5 mM Tris pH 6.8, 5% glycerol, 2% SDS and 0.2% bromophenol blue) and 2% β-mercaptoethanol and boiled for 5 min. This was then subjected to polyacrylamide gel electrophoresis using BioRad 10% Tris-HCl ready gels and transferred to a PVDF membrane. The NM PM -C SC was detected using a monoclonal mouse anti-HA tag antibody (1:10,000, Sigma Aldrich, St. Louis, MO, USA).
To perform semi-denaturing detergent agarose gel electrophoresis (SDD-Age) analysis, crude lysate (40 to 80 μg of total protein) was treated with 2% SDS in sample buffer for 7 min at room temperature. The lysates were subjected to agarose electrophoresis on a 1.5% agarose gel in running buffer to resolve the [CHI + PM ] sub-particles, and were transferred to a PVDF membrane using a wider mini-gel cassette or a semi-dry blot. Native Sup35p was detected by rabbit anti-NM SC antibody (kindly provided by S. Lindquist). The PVDF membrane was stripped prior to probing for the native Sup35p using the Applied Biosystems protocol. A preparation of chicken pectoralis extract (a kind gift from T. Keller) was used to estimate molecular weight [62]. When stained with Coomassie, chicken pectoralis extract reveals several abundant muscular proteins: titin (3,000 kDa), nebulin (750 kDa), and myosin heavy chain (200 kDa). Although this ladder cannot be used for precise determination of molecular mass, it does provide an estimate.