Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Colloquium Interactions among prions and prion "strains" in yeast Michael E. Bradley*, Herman K. Edskest, Joo Y. Hong*, Reed B. Wicknert, and Susan W. Liebman*: *Laboratory for Molecular Biology, Department of Biological Sciences, University of Illinois, 900 South Ashland Avenue, Chicago, IL 60607; and Laboratory of Biochemistry and Genetics, National Institute of Diabetes, Digestive and Kidney Diseases, Building 8, Room 225, National Institutes of Health, 8 Center Drive MSC0830, Bethesda, MD 20892-0830 Prions are "infectious" proteins. When Sup35, a yeast translation termination factor, is aggregated in its lPSI+] prion form its function is compromised. When Rnq1 is aggregated in its [PIN+l prion form, it promotes the de novo appearance of [PSI+]. Heritable variants (strains) of [PSI+] with distinct phenotypes have been isolated and are analogous to mammalian prion strains with different pathologies. Here, we describe heritable variants of the [PIN+] prion that are distinguished by the efficiency with which they enhance the de novo appearance of [PSI+]. Unlike [PSI+] variants, where the strength of translation termination corresponds to the level of soluble Sup35, the phenotypes of these [PIN+] variants do not correspond to levels of soluble Rnq1. However, diploids and meiotic progeny from crosses between either different [PSI+], or different [PIN+] variants, always have the phenotype of the parental variant with the least soluble Sup35 or Rnq1, respectively. Apparently faster growing prion variants cure cells of slower growing or less stable variants of the same prion. We also find that YDJ1 overexpression eliminates some but not other [PIN+] variants and that prions are destabilized by meiosis. Finally, we show that, like its affect on [PSI+] appearance, [PIN+] enhances the de novo appearance of [URE3]. Surprisingly, [PSI+] inhibited [URE3] appearance. These results reinforce earlier reports that heterologous prions interact, but suggest that such interactions can not only positively, but also negatively, influence the de novo generation of priors. Prions are best known as the infectious agents proposed to be responsible for the mammalian transmissible spongiform en- cephalopathies including scrapie in sheep, mad cow disease in cattle, and Creutzfeldt-Jakob disease in humans (1~. The prion form of the PrP protein is proposed to propagate its abnormal form to other "normal" PrP protein molecules with the same primary sequence (2, 3~. Evidence suggesting that self-propagating prion proteins are not limited to PrP was presented in 1994 by Wickner (4~. Three yeast proteins with self-perpetuating, alternate confor- mations have now been well described: tPSI+] (5), the prion form of the translational termination factor Sup35 (6-8~; fURE3] (9), the prion form of the nitrogen catabolite repressor Ure2 (4, 8~; and tPIN+] (10, 11), the prion form of Rnql (12, 13~. In tPSI+] and fURE3] cells, Sup35 and Ure2 are respectively inactivated by aggregation, causing the same phenotypes as mutations in the SUP35 and URE2 genes. No phenotype has yet been associated with inactivation of RNQ1. The de novo appearance of each of these yeast priors, tPSI+], fU^3i, and tPIN+], is enhanced by overproducing the correspond- ing prion domains (12, 14, 1S). The increased number of protein molecules presumably enhances the chance that a prion seed will form de novo. However, de novo appearance of some prions depends on the presence of other prions or prior-like aggregates (13, 16~. We first described tPIN+] as a prior-like element having the phenotype of allowing overproduction of Sup35 to convert [psi-] cells to tPSI+] (10, 11), and later showed that tPIN+] is equivalent to the prion form of Rnql (13~. The presence of [URE3] 16392-16399 1 PNAS 1 December 10, 2002 1 vol. 99 1 suppl. 4 (13) or the artificial fusion prion tNU+] (16) also permitted overexpression of Sup35 to induce the appearance of tPSI+J. The existence of different heritable forms or strains of prions is a fascinating chapter in the biology of priors. Prion diseases exhibit variable incubation times, neurodegenerative patterns, and PrP prion deposits, all of which remain distinct on transmission in inbred mammals (17~. Recent evidence supports the idea that prion strain variation is a result of the PrP protein's ability to propagate in different heritable prion forms (18-20~. Others, however, view the existence of prion strains as more compatible with a viral model for prion disease (21~. The finding of tPSI+] strains (14), and more recently [URE3] strains (22), in yeast, where the viral hypothesis is unreasonable, supports the idea that prion strains result from multiple prion protein forms. Distinct strains of tPSI+] have different mitotic stabilities (fre- quencies of tPSI+] loss), translational termination activities as measured by suppression of nonsense codons (14), and levels of nonaggregated Sup35 (234. Weak tPSI+] are less stable than strong tPSI+] in mitotic division (14), and the levels of nonaggregated Sup35 and accompanying translational termination are higher in weak tPSI+ J cells than in strong tPSI+] cells (23~. Strains of tPSI+] have also been distinguished by their differential responses to mutations in the SUP35 gene (24, 25) and by their responses to chaperones (26~. Several tPSI+] strains have been shown to be dominant, non-Mendelian traits when crossed with [psi-] (5, 27~. In addition, the strong tPSI+] phenotype appears in diploids made from mating strong and weak tPSI+] cells (23, 24~. Several in vitro studies support the hypothesis that strains of [PSI+] result from distinct protein conformations of Sup35. Purified Sup35 prion domain (Sup35NM) forms fibers in vitro with either wavy or straight structures (28~. Also, a purified chimeric Sup35NM was shown to form aggregates with distinct conformations and distinct in vitro seeding activities (29~. Most convincingly, protein extracts from strong [PSI+] cells converted purified Sup35NM into fibers more efficiently than did protein extracts from weak tPSI+] cells (30, 314. The continued propagation of prions depends on normal chap- erone protein levels. The finding that either deleting or overex- pressine the HSP104 chaperone gene causes the elimination of [PSI+] (32) supported the prion model for tPSI+] because it suggested that tPSI+] formation involved a conformational change. Deleting, but not overexpressing, HSP104 eliminates tURE3] (33) and tPIN+] (10, 34), and overexpressing ~J1, which encodes an Hsp40 family chaperone, promotes the loss of tURE3] (33) and a weak strain of Pichia methanolica [PSI+] in Saccharomyes cerevisiae This paper results from the Arthur M. Sackler Colloquium of the National Academy of Sciences, "Self-Perpetuating Structural States in Biology, Disease, and Genetics," helcl March 22-24, 2002, at the National Acaclemy of Sciences in Washington, DC. Abbreviations: GuHCI, guanicJine~hydrochloride; Cyh, cycloheximide; GFP, green fluores- cent protein; USA, ureidosuccinate. tTo whom reprint requests should be acidressed. E-mail: SUEL@uic.eclu. www.pnas.org/cgi/doi/10. 1 073/pnas. 152330699
(264. HsplO4 functions together with Hsp40 and Hsp70 (35) to promote the renaturation of denatured or aggregated proteins (36~. The effect of HsplO4 on tPSI+] is modified by levels of, and alterations in, the Hsp70 family members Ssa and Ssb (26, 37-39~. Maintaining the prion form of Rnql requires specific domains of the Hsp40 family member Sisl (34~. In addition, a specific mutation of SIS1 caused an altered aggregation pattern of the Rnql prion that appeared to be heritable even in the absence of the SIS1 mutation (34~. Here, we demonstrate the existence of different strains of tPIN+] with distinct phenotypes. We determine the relative competitive- ness of these tPIN+] prion strains and of [PSI+] prion strains and find that one factor foretells the outcomes of competitions between two variants of the same prion. We also find that, whereas the tPIN+] prion enhances the de novo appearance of LURED, the presence of the tPSI+] prion inhibits tURE3] appearance. The non-Mendelian segregation of prions has been reported to deviate from the classical 4:0 ratio (4, 5, 9, 27, 40~. Here, we carefully document this effect for tPSI+] and tURE3i, and show that it is due to meiosis and not the conditions used to stimulate sporulation. Finally, we show that overproducing the chaperone Ydjl promotes the elimination of some tPIN+] strains. Materials and Methods Cultivation Procedures. Standard yeast media and cultivation pro- cedures were used (41~. Yeast extract/pep/one/dextrose (YPD) with 5 mM guanidine hydrochloride (YPD + GuHCl) was used to eliminate prions (42~. YPD with 10 mg/liter cycloheximde (YPD + Cyh) was used to select for cycloheximide-resistant mutations (cyhR) and for random spores in tPSI+] experiments. Synthetic medium with 3 mg/liter cycloheximide was used to select for cytoductants. Growth on YPD plates containing 40 ,ug/ml of ethidium bromide converted strains to frho-] (43~. Casamino acid (CA) medium contained 0.13% yeast nitrogen base, 0.5% ammo- nium sulfate, 1% casamino acids, and 2% glucose or 2% glycerol. Uracil, adenine, and tryptophan were added to CA when necessary. CA with 5 mg/liter cycloheximide (CA + Cyh) was used to select for random spores in tPIN+] experiments. Synthetic medium con- taining galactose and raffinose (S Gal + Raf) was used to overex- press ~J1 or URE2 from the galactose-inducible promoter GAL1. Copper sulfate (Cu; 50 ,uM) was added to medium to induce expression of RNQ1 or SUP35 under the control of the inducible CUP1 promoter. Strains. Opposite mating type yeast strains that are isogenic, sporu- late efficiently when intercrossed, and contain the adel-14 allele, which permits weak and strong [PSI+] to be distinguished, were constructed by mating 74-D694 (MA Ta adel-14 trpl-289his3-~200 leu2-3,112 ura3-52) (44) with an efficiently sporulating strain, NKY292 (MA Ta Iys2 ura3 leu2::hisG ho::LYS2) (45) (kindly sup- plied by D. Bishop, University of Chicago) and backcrossingMATc~ meiotic progeny to 74-D694 four times. Progeny from the final backcross were diploidized by transforming them with pGAL-HO (46) (kindly provided by R. Esposito, University of Chicago). The diploids were sporulated and dissected to obtain MATa and MATor segregants that otherwise had the same genotype as 74-D694. L1842 (MA Ta) and L1843 (MA To`) are one pair of opposite mating type progeny from a single tetrad, as are L1844 (MA Ta) and L1845 (MA Tor). L2176 is a spontaneous cyhR derivative of L1845 in which the cyhR mutation was shown to be recessive. To induce weak and strong tPSI+], L1842, L1843, L1844, and L1845 were transformed with pEMBL-SUP35. Transformants were grown on plasmid selective medium for ~14 generations and subsequently on YPD for ~ 14 generations to promote plasmid loss before plating the cultures for individual colonies on YPD. L2010 and L2012 are, respectively, weak and strong [PSI+] derivatives of L1844. YHE711 (MA Tor ura2 leu2 [psi-] [ure-o] [PIN+]) (47) was scored Bradley et a/. as [PIN+] because expression of an RNQ1-GFP fusion in the strain gave rise to discrete aggregates characteristic of tPIN+] (12, 13). In addition, as expected in a tPIN+] background, overexpression of SUP35NM-GFP in YHE711 led to the appearance of ribbon and curve aggregates characteristic of newly induced tPSI+] (48). De- rivatives of YHE711 grown in YPD + GuHCl failed to give rise to either of these types of aggregates and are therefore Lpin-~. An RNQ1 deletion derivative of the tPIN+] version of YHE711 was constructed by transformation with a PCR product of the RNQl::kanMX4 insertion from yeast strain "American Type Cul- ture Collection (ATCC) no. 4003435; ref. 49] amplified with DNA primers HE230 (RNQ1 5' UTR,5'-CACGTATTTCAGTTGTCC- 3') and HE231 (RNQ1 3' UTR, 5'-CCACTCTTACATTGT- CATT-3'). Transformants were selected on YPD containing 300 ,ug/ml G418, after a recovery period in YPD. To confirm the disruption of RNQ1, candidate mutants were analyzed by PCR using primers HE265 (RNQ1 5' UTR, 5'-GAATGATCCATCGT- TCTTAC-3'), HE266 (RNQ1 3' UTR, 5'-GATGGCTTATATC- CTGCTC-3'), HE267 (kanMX4 pointing to 5' end,5'-CTGCAGC- GAGGAGCCGTAAT-3'), and HE268 (kanMX4 pointing to 3' end, 5'-TGATTTTGATGACGAGCGTAAT-3'). GuHCl-treated versions of yeast strains A3099 (MATor ade2-1 SUQS Iysl-1 his3-11,15 leaf karl-1 ura3::KanMX4 [psi-][rho-]) (12), clOB-H49 (MATcx ade2-1 SUQ5 Iysl-1 his3-11,15 leul karl-1 cyhR [psi-][rho-]) (50), BY4741 (MATa his3-l\1 leu2-~\ metlS-l\ ura3-~\ tpsi-] [PIN+]) (from Research Genetics), and 3385 (MATa ura2 leu2 karl his- [psi-] [ure-o][PIN+]) (4) were used in cytoduc- tion experiments. Plasmids. A 2,u plasmid (pEMBL-SUP35) with URA3 leu2-d mark- ers and SUP35 under its native promoter (51) was used to over- produce Sup35 at moderate levels on synthetic medium lacking uracil (-Ura) or high levels (on -Leu). The defective LEU2 promoter present in the leu2-d allele on this plasmid selects for overamplification of the plasmid on -Leu. Moderate overexpres- sion of SUP35 was used to induce tPSI+~; high-level overexpression was used to distinguish different variants of ~PIN+] on the basis of growth inhibition. YDJ1 under the control of the GAL1 promoter is present on a CEN LEU2 plasmid (p901); the parent plasmid without YDJ1 is pH316 (33). Plasmids pRNQ1-GFP and pSUP35NM-GFP, which respectively contain the fusions of either RNQ1 or the NM domains of SUP35 to green fluorescent protein (GFP) under the control of the CUP1 promoter, were used to score for ~PIN+] as described previously (13). Plasmids used for the [URE3] induction experiments were 2,u-based, with a LEU2 marker and the GAL1 promoter to express URE2 (pH376), URE2 (1-65) (pH382), URE27\~s~_~5s (pH377), or no expression of URE2 as a negative control (pH317) (52). Cytoduction. Cytoductions were performed by crossing tRHO+] donors to frho-] recipients. Either the donor or the recipient carried the karl-1 allele, which inhibits nuclear fusion (53). When the recipient was cycloheximide-resistant (cyhR), cytoductants were selected on synthetic glycerol medium containing cycloheximide. Otherwise, diploids and cytoductants were selected on synthetic glycerol medium deficient in a nutrient required by the donor strain for growth. Cytoductants were then identified by subcloning the population and screening colonies for the recipient mating type and auxotrophic markers. Analyses of [PSI+] Variants. After inducing weak and strong [PSI+] in L1842, L1843, L1844, and L1845, the L1842 and L1843 deriva- tives (shown in Fig. 1) and the L1844 and L1845 derivatives were crossed in all possible combinations. We assayed diploids for tPSI+] strength by color on YPD and level of growth on synthetic medium lacking adenine (-Ade). The diploids were sporulated after prop- agating for approximately 42 generations. Meiotic progeny from PNAS | December 10, 2002 | vol. 99 | suppl. 4 | 16393
P 1 [psi-] ~psi-] [psil strong weak x x x x x x P 2 [psi-] weak strong strong strong G. diploid. · ~ melO IC progeny Fig. 1. Meiotic inheritance of [PSI+] variants. Isogenic haploid yeast, parent 1 (P1; L1842 derivatives) and parent 2 (P2; L1843 derivatives) carrying the indicated [PSI+] variantswere mated. One representative diploid and twotetradsfrom each cross are shown. each diploid were assayed for tPSI+] strength, mating type, and curability of tPSI+] by growth on YPD + GuHCl. To perform the random spore analyses of tPSI+] loss frequencies, weak and strong tPSI+] derivatives of L1844 (L2010 and L2012, respectively) were mated to a GuHCl-treated c~hR derivative of L1845 (L2274) to produce diploids SL-1142 and SL-1143, respec- tively. The resulting diploids were sporulated, and frequencies of tPSI+] loss were determined in at least three independent trials by counting red vs. total number of colonies after suspending cells in 100 ,ul of 10% gluculase, vortexing for 20 s, diluting 1 x 10-4 in water, and plating to YPD + Cyh. Analyses of [PIN+] Variants. Phenotypes of tPIN+] variants in 74- D694, L1844, and L1845 were determined by overexpressing SUP35 from the pEMBL-SUP35 plasmid. Transformants were patched to -Ura, where the plasmid is present in moderate copy number, and then spotted to -Ura, -Leu, and -Ace. The induction of tPSI+] was identified as GuHCl-curable nonsense suppression of adel-14 that was independent of the plasmid used to induce the appearance of tPSI+~. Phenotypes of tPIN+] variants in clOB-H49 were deter- mined by overexpressing SUP35 and spreading transformants to -Ace. To determine the relative competitiveness of the tPIN+] variants, we first generated isogenic opposite mating type yeast carrying each of the tPIN+] variants. Independent tPIN+] derivatives of 74-D694 were cytoduced into a GuHCl-treated derivative of clOB-H49 and from there into a GuHCl-treated spin-] derivative of L1844. Diploids made from crosses of the final cytoductants to a GuHCl- treated spin-] cyhR derivative of L1845 were transformed with pEMBL-SUP35 and sporulated. The tPIN+] phenotypes of cyhR progeny that maintained pEMBL-SUP35 (obtained by plating sporulated cultures to CA-Ura + Cyh) were determined by growth on -Leu and -Ace. MATa progeny carrying the different tPIN+] variants were backcrossed to the original [PIN+] derivatives of 74-D694. Diploids were tested for the tPIN+] phenotypes and sporulated after propagating for approximately 28 generations. Random spores were selected on CA-Ura + Cyh, and their tPIN+] phenotypes were determined. Comparison of Rnq1 Among lPIN+] Variants. Derivatives of 74-D694 were grown in liquid YPD to mid-log (OD600 ~ 1.0~. Harvested cells were resuspended in lysis buffer [50 mM Tris, pH 8.0/150 mM NaCl/0.2% Triton X-100/1.9 ,ug/ml aprotinin/3.5 ,ug/ml E-64/5 ,ug/ml leupeptin/5 ,ug/ml pepstatin/400 ,ug/ml 1,10 phenanthro- line/500 ,ug/ml PMSF/50 ,ug/ml N-(p-tosyl~lysine chloromethyl 16394 1 www.pnas.org/cgi/doi/10.1073/pnas.152330699 ketone (TLCK)], and mixed with 750 ,ul of glass beads/lysis buffer slurries. Total protein lysates were obtained by vortexing each tube eight times for 10 s, with intermittent incubations on ice, and removing cell debris at 10,000 x g for 10 min. The protein concentrations of the cleared lysates were measured (Bio-Rad Protein Assay), and lysis buffer was added to normalize the samples. Total protein lysate (1-2 ma) was fractionated at 280,000 x g for 30 min in a Sorvall TLA100.1 rotor. Pelleted proteins were resus- pended in 200,ul of lysis buffer. Rnql was detected with a polyclonal antibody (kind gift from S. Lindquist, University of Chicago). Influence of [PIN+] and [PSI+] on [URE3] Appearance. tPIN+] and tPSI+] derivatives from 74-D694 were cytoduced into a GuHCl- treated version of clOB-H49, and from there into a GuHCl-treated version of 3385, and finally from 3385 into a GuHCl-treated version of YHE711. The tURE3] prion induction assay was performed as described (52~. Briefly, fure-o] strains were transformed with pH317, pH376, or pH382 and transformant colonies were individ- ually grown to saturation in SGal + Raf-Leu. Starting from 107 cells per plate, serial dilutions were plated on synthetic, dextrose-based medium containing 100 ,ug/ml ureidosuccinate (USA). Colonies appearing after 5 days at 30°C were recorded. Elimination of [PIN+] by YDJ1 Overexpression. Various [psi-] [PIN+] derivatives of 74-D694, as well as one [psi-] [pin-] control, were transformed with p901 (33) or the control plasmid pH316 (lacking YDJ1 but containing the GAL1 promoter). Transformants were grown on media containing galactose to induce YDJ1 expression. Two transformants for each strain and plasmid combination were subcloned on SGal + Raf-Leu 2-3 times consecutively by picking 3-4 same-sized colonies for each successive colony purification step. Colonies underwent an average of 20 cell doublings before being purified again or tested for fLPIN+~. The elimination of tPIN+] was scored by checking for the loss of aggregated Rnql. Colonies were patched to YPD and crossed to tester strains: GuHCl-treated versions of 64-D697 (`MATa adel-14 trpl-289 Iys9-A21 leu2-3, 112 ura3-52) or SL1010-lA (<MA T(x adel-14 met8-1 trpl-1 his5-2 leu2-1 ura3-52) transformed with pRNQ1-GFP (13~. Diploids from these crosses were selected by complementation on medium (SC-His,Lys,Ura + Cu) that selects for maintenance of pRNQ1-GFP and contains 50 ~M Cu to induce expression of the fusion protein. In this assay, [PIN+] colonies have bright green aggregates in the majority of cells, and [pin-] colonies have evenly distributed green fluorescence. Results Genetic Analysis of Strains of [PSI+l. To avoid confusion of yeast strains with prion strains we sometimes refer to the latter as variants. Yeast bearing strong or weak [PSI+] prion variants were previously distinguished (14) by the efficiency with which they could suppress the adel-14 nonsense allele, which contains a premature stop codon that prevents the protein from being completely trans- lated. Yeast that are [psi-] do not grow on -Ace because trans- lation termination at the premature adel-14 stop codon is efficient, and they are red on YPD due to the accumulation of a metabolic intermediate of the adenine biosynthesis pathway. Yeast strains with weak tPSI+] grow poorly on -Ace and are pink on YPD because they maintain lower levels of functional Sup35 compared with [psi-~; consequently, they partially suppress termination at the premature adel-14 stop codon and produce some functional Adel. Yeast strains with strong tPSI+] grow well on -Ace and are white on YPD because they maintain even less functional Sup35 than weak fLPSI+], and they efficiently suppress termination at the premature adel-14 stop codon (23, 30~. Diploids resulting from pairwise matings of isogenic yeast that were either [psi-], weak tPSI+], or strong tPSI+] displayed nonsense suppression levels equal to that of the tPSI+] parent with the "strongest" phenotype (Fig.1~. Tetrad analyses revealed that strong Bracl~ey et a/.
Table 1. Tetrad analysis of [PSI+] variants Parent 1 [psi-] Weak Weak Strong Strong Weak Strong Tetrads Viable Parent 2 dissected progeny [psi-] [psi-] Weak [psi-] Strong Strong Weak 14 59 33 33 16 18 21 56 200 118 121 60 67 77 [PSI+] phenotypes of meiotic progeny [psi-] Weak Strong 56 10 2 o o o o o 190 116 o o o o Each row represents the sum of progeny obtained from two to four independent diploids. At least one diploid was from a cross between deriva- tives of L1842 and L1843, and one was from a cross between derivatives of L1844 and L1845. Weak and strong refer to the [PSI+] variants. [PSI+] diploids always segregated strong [PSI+] in a 4:0 ratio and weak [PSI+] diploids usually segregated weak [PSI+] in a 4:0 ratio (Fig. 1 and Table 1~. Meiosis Eliminates [PSI+] and [URE3] Prions. Approximately 2-5% of the spores from weak [PSI+] diploids were [psi-] (Table 1, rows 2 and 3), which is significantly more than the 0.7% loss of weak [PSI+] among the mitotic progeny from one of the parents of these diploids (average of ~2,800 colonies from three independent, equally represented trials). This, together with an earlier finding that a weak [PSI+] (then called [ETA+ was very unstable in meiosis (27), led us to investigate this phenomenon further. Haploid cells did not exhibit an enhanced loss of weak [PSI+] when exposed to the same conditions that induce meiosis in an isogenic weak [PSI+] diploid. We incubated three yeast strains on sporulation medium: a weak [PSI+] haploid, L2010; a weak [PSI+] diploid isogenic to L2010 but heterozygous for cyhR, SL-1142; and a week tPSI+] cyhR meiotic segregant from this diploid, SL1142-lA. Random spores were selected from the diploid culture on YPD + Cyh. SL1142-lA was also plated on this medium, and L2010 was plated on YPD lacking cycloheximide. Ihe frequency of D?si-] among the random spore colonies was 5.7% (average of ~2,400 colonies from three independent, equally represented trials). The frequency of Lpsi-] among mitotic colonies from either of the haploid controls was only 0.4% (7 Epsi-] of 1,670 colonies from L2010, and 15 Lpsi-] of 3,530 colonies from SL1142-lA). There- fore, the observed effect was not a result of the conditions used to induce sporulation. Ihe effect was also not due to heightened instability in the diploid phase because the frequency of weak tPSI+] loss from mitotic diploid progeny was only 0.07% (average of ~3,450 colonies from three independent, equally represented trials). Ibus, some aspect of meiosis interferes with the inheritance of tPSI+~. Likewise, although tURE3] is highly stable during mitotic growth, it is frequently lost in meiotic segregants (4, 9,40~. We examined the effect of meiosis-inducing conditions on the stability of tURE33. tURE3] was stable in mitotic growth: it was efficiently cytoduced from strain 1735 to 1019 (16 of 16 cytoductants examined were USA+~. Growing 1735 on sporulation medium did not decrease the stability of tURE33. Furthermore, a diploid made by crossing 1735 with 1019 stably maintained tURE3] (100 of 100 colonies examined after growth on YPD were USA+~. However, sporulation of this diploid produced mostly USA- spores (39 of 48 USA- spores). Thus, the process of meiosis causes loss of tURE3~. [PIN+] Variants. We previously described the isolation of spontane- ously appearing tPIN+] colonies after prolonged incubation of a [pin-] [psi-] derivative of 74-D694 (11~. Rare tPIN+] cells were detected by selecting for the appearance of tPSI+] after overex- Bradley et a/. Originals ~ _ _..~... _ _ _ _ O .~E 121 ~ ·- ~ 67 _ ._ '_ _ _ __ 77 _ ~ _ ~ _ _ __ ~ ~ ~ _ ~ _ _ -Ura -Ad e -Leu ~in -] low med. high v. h. Cytocluctants - - _ _ . -Ura -Ad e -Leu Fig. 2. Characterization of [PIN+] variants and their inheritance through cyto- duction. Independent derivatives of the original 74-D694 are shown (Originals). These were each cytoduced into a [pin-] version of c10B-H49, a karl-1 yeast strain, and from there back into a [pin-] version of 74-D694 (Cytoductants). Both the originals and cytoductants carry the pEMBL-SUP35 plasmid. When the plas- mid is maintained at moderate level on - Ura distinct levels of [PSI+] induction are observed on transferto -Ace. When the plasmid is amplified on -Leu, different levels of growth inhibition are observed. The different [PIN+] variants are cyto- plasmically inherited because cytoductants and donors exhibit the same [PIN+] phenotypes. Row 1 is the [pin-] control; rows 2, 3, and 5 are the low, medium (med.), and very high (v.h.) spontaneous [PIN+] variants obtained in [pin-] 74- D694; row four is high [PIN+] from the original 74-D694. pression of SUP35. We eliminated tPSI+], but not tPIN+], by overexpressing HSP104. Later, the tPIN+] status of these isolates was shown to be a consequence of the prion form of Rnql because the loss of tPIN+] correlated with the loss of Rnql aggregates (13~. The phenotypes of tPIN+] were originally described as allowing moderate overproduction of Sup35 to convert [psi-] cells to tPSI+] and as inhibiting growth in the presence of extreme overproduction of Sup35 (10~. We now distinguish tPIN+] variants with different levels (low, medium, high, and very high) of tPSI+] induction and growth inhibition (Fig. 24. Because tPIN+] is cytoducible (12, 13), cytoduction should transfer the distinct phenotypes if they result from heritable tPIN+] variants, but not if they are the result of Mendelian modifier mutations. Derivatives of 74-D694 carrying the different [PIN+] isolates and a [pin-] control were cytoduced into a [pin-] derivative of clOB-H49, and from there back into a [pin-] derivative of 74-D694. The cytoductants displayed the donor's tPIN+] pheno- types (Fig. 2~. This result proves that Mendelian mutations do not cause the variable phenotypes. Genetic Analysis of Variants of [PIN+]. To determine the relative competitiveness of the different [PIN+] variants, we crossed pairs of opposite mating type yeast harboring each of the different tPIN+] variants in all possible combinations and determined the tPIN+] phenotypes of the resulting diploids by measuring tPSI+] induction and growth inhibition levels. High tPIN+] outcompetes low and medium tPIN+] (data not shown), and medium [PIN+] outcom- petes low (Fig. 3~; however, very high tPIN+] was outcompeted by high, medium, and low ~PIN+] (Fig. 3 and data not shown). Therefore, the winner in these competitions is not always the one with the "strongest" tPIN+] phenotype. Meiotic progeny always exhibited the same phenotype as the diploid parent (Table 2~. This result was true even when two different tPIN+] variants were crossed. For example, crosses between low tPIN+] and very high [PIN+] produced low tPIN+] diploids whose meiotic progeny always inherited low tPIN+~. Because very high ~PIN+] was outcompeted by low, medium, and PNAS I December 10, 2002 1 vol. 99 I suppl. 4 1 16395
P2 P1\ [Ping low med. v.h. .= ~ ~ s .= O ~ ~ -Ura -Ace -Leu Fig. 3. Genetic analysis of variants of [PIN+]. Independent diploids carrying the pEMBL-SUP35 plasmid reveal the outcome of crosses between haploid parents (P1 and P2) carrying the [pin-], low, medium (med.), or very high (v.h.) [PIN+] variants. The [PIN+] phenotypes are scored as in Fig. 2. high tPIN+], it was possible that the latter phenotypes were caused by a combination of [PIN+] and an additional "modifier" prion distinct from tPIN+~. In this case, diploids formed from crosses of very high [PIN+] to, e.g., low tPIN+] would contain both the ~PIN+] prion and the modifier prion, resulting in the low tPIN+] phenotype. To test this possibility, we cytoduced low and medium tPIN+] derivatives of clOB-H49 into a llrnql derivative of BY4741 that, while unable to maintain [PIN+], should be able to maintain other cytoduced modifiers. The Arnql recipient was then cytoduced into a very high tPIN+] derivative of clOB-H49 to test whether the hypothesized prion modifier would convert very high into low or medium ~PIN+~. Because most cytoductants remained very high tPIN+] regardless of whether [pin-] control (11 of 12 independent cytoductants) or presumptive modifier-containing donor cytoplasm (11 of 12 independent cytoductants) was used, the data do not support the prion modifier hypothesis. In control experiments, we cytoduced weak or medium tPIN+] derivatives of BY4741 into a very high tPIN+] derivative of clOB-H49 and found that the majority of the recipients (10 of 15) were indeed converted into the phenotype of the donor (low or medium). Thus, it appears that the distinct phenotypes result from heritable differences in the tPIN+] aggregates themselves. Comparison of Rnq1 Aggregation Among [PIN+] Variants. We com- pared the amounts of soluble and aggregated Rnql in the high, medium, low and very high tPIN+] variants. Each maintained indistinguishable amounts of aggregated Rnql, but the levels of soluble Rnql showed strain-specific differences (Fig. 4~. Strikingly, the hierarchy of ~PIN+] variants determined by the competition Table 2. Meiotic inheritance of [PIN+] variants [PIN+] phenotypes of melotic progeny Parent 1 Parent 2 [pin-] [pin-] [pin-] [pin-] Low Med. V.H. Low Med. Med. [pin-] Low Med V.H. Low Med V.H. V.H. V.H. Low Low Med. 26 30 29 30 30 30 30 62 60 60 o 29 o o 29 O o 60 o o o o 29 o o 30 o o 60 58 V.H o o o 30 o o 30 o o o high med. low v.h. [pirl~] S p ~ Rnq1 p ~:~: Rnq1 p Fig. 4. Comparison of levels of soluble and aggregated Rnq1 among [PIN+] variants. Lysates were fractionated into soluble (S) and pellet (P) fractions by ultra-centrifugation. Rnq1 was detected by Western blotting with polyclonal Rnq1 antibody (kind gi ft from S. Lindquist). Stripped blots were then re-probed with monoclonal Sup35 antibody (Control) as a loading control. Note, the soluble Rnq1 in this figure were exposed twice as long as the Rnq1 in the pellet. The gradient depicted was generally reproducible; however, in three of eight inde- pendent protein isolations, the level of soluble Rnq1 in the v.h. and low variants appeared similar. experiments described above exactly corresponded to the gradient, from least to most, of soluble Rnql exhibited by these strains. For example, the variant with the most soluble Rnql, very high, was lost when crossed with each of the other tPIN+] variants, whereas the variant with the least soluble Rnql, high, outcompeted all of the other [PIN+] variants. 26 1 o o Effect of [PIN+] and [PSI+] on lURE3] Appearance. Because [PIN+] facilitates the appearance of tPSI+] (10, 11), we asked whether it has a similar effect on the appearance of another prion, [URE34. We compared the frequency with which overexpression of URE2 can induce the appearance of tURE3] in the [PIN+] yeast strain YHE711 and a [pin-] derivative of this strain obtained by growth on GuHCl. The original YHE711 strain had enhanced tURE3] induction relative to the [pin-] derivative, which had a drastically reduced frequency of tURE3] appearance (Table 3~. Because deleting RNQ1 from the original tPIN+] strain also abolished the ability to induce tURE3], YHE711 does not harbor other elements in addition to [PIN+] that independently allow for tURE3] induction (Table 3~. Clearly, tPIN+] facilitates ~URE3] appearance. We further showed that [PIN+] did not cause greater overproduction of Ure2, nor did it stabilize newly appearing tURE3] (data not shown). To determine whether the different variants of tPIN+] that were characterized above in terms of their distinct [PSI+] induction frequencies could also be distinguished on the basis of their effect on tURE3] induction, we cytoduced several different variants of tPIN+] into a GuHCl-treated derivative of YHE711. Surprisingly, the original YHE711 [PIN+] facilitated tURE3] appearance more efficiently than even the very high tPIN+] (Table 3), yet the original [p/n ] YHE711 tPIN+] was less efficient than very high tPIN+] in the Sup35-based tPSI+] induction phenotypes (data not shown). Whereas the presence of the low, medium, or very high fLPIN+] elements enhanced the frequency of tURE3] induction relative to the [pin-] control,we could not consistently observe any differences in tURE3] induction levels among strains with the low, medium and very high tPIN+] variants (Table 3~. Although it generally appears that prions enhance the appear- ance of other prions (13), we find here that [PSI+] does not facilitate tURE3] induction, but rather inhibits its appearance. By overex- pressing the URE26~s~_~5s allele, which efficiently induces tURE3] (15, 54), we found that a tPSI+] spin-] derivative of YHE711 is less inducible to tURE3] compared with a [psi-] spin-] derivative (averages of 346 + 107 vs. 989 + 394 tURE3~/106 cells from six independent experiments). o 2 2 [PIN+] phenotypes were scored using the SUP35 overexpression assays shown in Fig. 2. For each row, ~10 progeny were obtained from 3 or 6 independent diploids. Low, medium (Med.), and very high (V.H.) variants of [PIN+] were used. 16396 1 www.pnas.org/cgi/doi/ 10.1 073/pnas. 1 52330699 Bradley et a/.
Table 3. Influence of [PIN+] on [URE3] appearance Frequency of [URE3] as % of [pin-] control [PIN+] Variants Experiment [pin-] /\rnql Original Veryhigh Medium Low 1 100 + 59 4,600 + 1,100 2 100 + 49 4,300 + 2,400 3 100 + 50 37,000 + 12,500 3,500 + 1,200 4 100 + 39 9,200 + 4,600 550 + 80 5* 100 19 23,000 3,500 6* 100 23 1,100 1,100 7* 100 830 650 720 8 100 + 50 800 + 360 2,100 + 1,100 1,200 + 650 9 100 + 142 15,00 + 6,400 2,800 + 1,300 1,000 + 1,100 10 100 + 143 39,000 + 19,000 510 + 230 5,700 + 3,800 Average 100 21 16,700 1,700 2,400 960 Full-length Ure2 was overproduced in derivatives of YHE711. The frequency of [URE3] colonies appearing in the [pin-] strain per 106 cells plated is normalized to 100, and other values were normalized accordingly. The YHE711 derivatives were: GuHCI-treated ([pint]), RNQ1 deletion (Arnq1), not GuHCI-treated (Original), GuHCI-treated and cytoduced with very high, medium, or low [PIN+]. Averages and standard deviations are shown when three or more transformants were assayed. Induction of [URE3] by overproducing just the U re2 prion domain (1 -65 aa) was also facilitated by the presence of [PIN+] (data not shown). Each experiment also included controls in which URE2 was not overexpressed, where the numbers of [URE3] colonies was very low (not shown). *Indicates that only two measurements were performed, the average is shown. Effect of YDJ1 Overexpression on SPIN+]. Various [psi-] tPIN+] derivatives of 74-D694, as well as a p?si-] [pin-] control, were transformed with p901, carrying YDJ1 under the control of the GAL1 promoter. Transformants were grown on media containing galactose to induce YDJ1 expression, after which Rnql aggregation was used to detect [PIN+~. Some tPIN+] variants were readily cured by overproducing Ydjl, whereas other [PIN+] variants were not cured (Table 4~. tPIN+] was never lost in control experiments by using the pH316 empty vector (Table 4~. The experiment gave similar results when repeated with two of the curable and three of the incurable tPIN+] variants (Table 44. Thus, we find that over- expressing YDJ1 promotes the loss of some tPIN+] variants. Discussion According to the protein-only prion model, two explanations of the prion strain phenomena are possible. Prion variants may result from inherent flexibility of the tertiary structure allowing one chain of amino acids to have two or more self-perpetuating conformations that are stably inherited. Alternatively, variants might result from a single tertiary conformation arranged into Table 4. YDJ1 cures some lPlN+] variants Expt. 1 Expt. 2 Control YDJ1 Control YDJ 1 Strain [pin-] [PIN+] [pin-] [PIN+] [pin-] [PIN+] [pin-] [PIN+] [pin-] 24 V.H. 0 Low 0 Med. 0 High 0 L1941 0 L1952 0 L1947 0 L1949 0 L1956 0 0 24 24 0 24 0 24 8 24 2 24 3 24 14 24 22 24 0 24 0 o o o o o 0 35 0 48 0 24 24 16 22 21 10 2 24 24 52 61 66 49 46 o 29 11 4 o 64 32 56 62 66 Strains are derivatives of 74-D694. The very high (V.H.), low, medium (Med.), and high [PIN+] variants are indicated as such. Other independent [PIN+] isolates are also indicated (L1941-1956). Colonies checked from the control (pH316) and YDJ1 bearing (p901) transformants are shown. Bradley et al. two or more different quaternary arrangements that are stable and self-perpetuating. Here, the existence of distinct heritable variants of the .[PIN+] prion is described, and the phenotypic differences between them are shown not to be due to either nuclear or cytoplasmic modifiers. It is now clear that each of the well characterized yeast prions tPSI+], [URE3i, and [PIN+] can exist as different distinct heritable variants. Although different priors, e.g., tPSI+] and [PIN+], or [PSI+] and f URE3], can be maintained together in a single cell (11, 13, 16), such coexistence may not be possible for two variants of the same prion. Indeed, diploids formed by mating cells with different tPSI+] variants that could be distinguished from each other by Sup35-GFP staining retained only one of the tPSI+] variants (25~. Furthermore, if prion variants could coexist, one might expect diploids formed from crosses between isogenic cells bearing weak and strong tPSI+] to exhibit a phenotype more extreme than strong tPSI+], and to occasionally segregate out both strong and weak tPSI+] in mitotic and meiotic growth. We show here that this is not the case. Rather, except for rare cases of loss of [PSI+], the diploids and all mitotic and meiotic progeny were indistinguishable from the strong tPSI+] parent. Furthermore, weak tPSI+] did not emerge from the diploids created by crossing weak and strong tPSI+] even after stimulation of tPSI+] loss by short-term growth in medium containing GuHCl (M.E.B. and S.W.L., unpublished work). Although one could argue that strong and weak tPSI+] coexist, but that the phenotype of strong tPSI+] cannot be made any stronger, similar results obtained for crosses between cells bearing the different tPIN+] variants cannot easily be explained in this way. This result is because, unlike crosses between weak and strong fLPSI+], where the variant with the strongest phenotype prevails, tPIN+] variants with less dramatic phenotypes (low, medium, and high [PIN+~) prevailed over very high tPIN+], which has the most dramatic phenotype. Thus, if the different tPIN+] variants coex- isted, the very high [PIN+] phenotype could not reasonably be expected to be hidden by the tPIN+] variants with milder phenotypes. Because two variants of the same prion compete for the same pool of newly synthesized protein to reproduce and be heritable, faster replicating variants should eventually outcompete slower or less stable variants by starving them for convertible protein (Fig. 5~. Indeed we found that [PIN+] variants, which maintain little soluble Rnql but abundant aggregated Rnql, indicating fast reproduction, PNAS | December ~o, 2002 | vo~. g9 | supp~.4 | 16397
~3 \ ~ ~~ l MANY GENERATIONS _1\`LOSS O F 'WEAK'' ^ PRION STRAIN DIPLOID\~ ~ ~ 4:0 'STRONG" Fig. 5. Cartoon depicting a competition for soluble protein between variants of the same prion. Haploids carrying soluble protein (free circles) and aggregat- ed-prion protein as either "weak" (rhomboids) or "strong" (squares) variants are mated. Mitotic growth of the diploid results in loss of the weak variant, presum- ably because it is starved for convertible soluble protein. When the diploid sporulates, each spore inherits the strong prion variant. did take over when combined with tPIN+] variants that maintain more soluble Rnql and less abundant aggregated Rnql. Previous observations indicate that tPSI+] variants show a similar pattern (23, 24~. Recent work has demonstrated that certain prions facilitate the appearance of other priors: tPIN+] and [URE3] (13), or the artificial fusion protein prion tNU+] (16), permit overexpression of SUP35 to induce the appearance of tPSI+~; and tPSI+] and fURE3] facilitate the appearance of tPIN+] (13~. Here, we show that tPIN+] facilitates, but tPSI+] inhibits, the de novo appearance of [URE3~. Two mechanisms of prior-facilitated prion appearance have been proposed (13, 164. According to the seeding model, heterol- ogous prions provide a template for initial cross-seeding of a de novo prion aggregate. The titration model hypothesizes that pre- existing heterologous prion aggregates sequester a protein that normally inhibits prion appearance, thereby allowing other prions to appear more easily. Our current finding of no correlation between the efficiencies with which the different tPIN+] variants promote the appearance of tPSI+] and the efficiencies with which they promote the appearance of tURE3] can most easily be ex- plained by the cross-seeding model. We propose that some tPIN+] variants cross-seed Sup35 better than Ure2, whereas others exhibit the opposite preference (Fig. 6~. More complicated scenarios involve combinations of the seeding and titration models, or multiple inhibitors with distinct binding properties. The tPIN+] variants described here cannot easily be distinguished by the amounts of aggregated Rnql (Fig. 44. There is also no correlation between the levels of soluble Rnql and the phenotypes of the tPIN+] variants: the order of increasing soluble Rnql levels 16398 1 www.pnas.org/cgi/cloi/10.1073/pnas.152330699 -' ae~bi,~ [URE~ ~ , Ve *| - 1PS~ ~[PI~ e~5~ 1PS~ 4,,5~9 .~1URE~ re Original.| `_ - 1PINI1 Fig. 6. Models illustrating the different seeding preferences proposed for two [PIN+] variants. (A) The very high [PIN+] variant inefficiently seeds Ure2 (red circles), but efficiently seeds Sup35 (green triangles) converting them into the [PSI+] shape. (B) The original YHE711 [PIN+] variant inefficiently seeds Sup35, but efficiently seeds Ure2, converting them into the [URE3] shape. Both [PIN+] vari- ants propagate theirforms by converting Rnq1 (blue rectangles) with the highest efficiency. is high, medium, low, then very high. One possibility to explain this conundrum is that different prion conformations of Rnql are better at influencing tPSI+] appearance, and these conformations are only coincidentally distinguishable by soluble Rnql levels. Another possibility is that accessory proteins, such as Sisl (34), are associated in different amounts with each of the tPIN+] variants. The presence of such proteins may hinder the action of tPIN+] or, if these are chaperone proteins, they might be essential for creating the action of tPIN+~. It is also possible that other variants of tPIN+] that do not facilitate the induction of tPSI+] or tURE3] may exist. We have unexpectedly found that the presence of one prion can inhibit the de novo appearance of another, because tPSI+] inhibited the appearance of tURE3~. Whereas the effect of tPSI+] on the induction of tURE3] is inhibitory, it still suggests that heterologous prions interact. [PSI+] may inhibit de novo tURE3] appearance by occasionally joining and "poisoning" fURE3] seeds thereby inhib- iting fURE3] propagation as previously proposed to explain the [PSI+] curing effect of certain SUP35 mutants and the fURE3] curing effect of URE2-GFP (~55-57~. Alternatively, ~PSI+] may sequester Ure2l\~5~_~5s thereby reducing the amount of protein available to form fURE3] seeds. That tPSI+] stimulates tPIN+] appearance but inhibits f URE3] appearance is inconsistent with the inhibitor model and more compatible with the idea that heterolo- gous prions directly interact sometimes causing cross-seeding and sometimes causing inhibition. Prions are stable, heritable elements. Yeast cells should therefore acquire multiple different prions by mating with other cells even though the prions may be disadvantageous under some circum- stances. Our finding that prions are occasionally eliminated by meiosis may be an indication that yeast have evolved mechanisms of ridding themselves of priors. Several possibilities might explain how prions are eliminated by meiosis. Because overexpressing HSP104 is known to eliminate ~PSI+] (32), the elevated HSP104 levels associated with sporulating cultures (58) might disrupt the inheritance of tPSI+~. Alternatively, because meiotic progeny re- BradIey et a/.
ceive less cytoplasm than mitotic daughter cells (59), they might lose tPSI+] more frequently as a result of inheriting fewer tPSI+] seeds. Because most Ure2 amyloid filaments are found in a single cyto- plasmic network in tURE3] cells (60), it is not surprising that tURE3] is often lost in meiosis. Overexpression of YDJ1, which interacts with HsplO4 and Hsp70 to rescue denatured proteins (35), was previously shown to cause the loss of a tURE3] variant (33~. Here, we show that overexpressing DJ1 eliminates some, but not other, tPIN+] variants. tU03] and tPIN+] are both eliminated by deletion, but not overexpression, of HSP104 (10, 33, 34~. The spontaneous tPIN+] variants described in this paper, including those that were not cured by Ydjl overex- pression, were cured by deleting HSP104 (M.E.B. and S.W.L., unpublished work). Possibly, YDJ1 may cure some tPIN+] variants by sequestering HsplO4 (33), but other tPIN+] variants may be less sensitive to the reduction of HsplO4. Alternatively, Ydjl may stimulate protein refolding and therefore cure tPIN+] by disaggre- gating it. If this result were true, it could mean that some tPIN+] variants maintain too many seeds, or replicate too quickly to be 9. 10. 1. Prusiner, S. B., Scott, M. R., DeArmond, S. J. & Cohen, F. E. (1998) Cell 93, 337-348. 2. Griffith, J. S. (1967) Nature (London) 215, 1043-1044. 3. Prusiner, S. B. (1982) Science 216,136-144. 4. Wickner, R. B. (1994) Science 264, 566-569. 5. Cox, B. S. (1965) Heredity 20, 505-521. 6. Liebman, S. W. & Derkatch, I. L. (1999) J. Biol. Chem. 274, 1181-1184. 7. Serio, T. R. & L~ndquist, S. L. (1999) Annul Rev. Cell. Dev. Biol. 15, 661-703. 8. Wickner, R. B., Taylor, K. L., Edskes, H. K., Maddelein, M. L., Moriyama, H. & Roberts, B. T. (2001) Adv. Protein Chem. 57, 313-334. . Lacroute, F. (1971) J. Bacter~ol. 106, 519-522. Derkatch, I. L., Bradley, M. E., Zhou, P., Chernoff, Y. O. & Liebman, S. W. (1997) Genetics 147, 507-519. 11. Derkatch, I. L., Bradley, M. E., Masse, S. V., Zadorsl~, S. P., Polozkov, G. V., Inge-Vechtomov, S. G. & Liebman, S. W. (2000) EMBO J. 19, 1942-1952. 12. Sondheimer, N. & L~ndquist, S. (2000) Mol. Cell 5, 163-172. 13. Derkatch, I. L., Bradley, M. E., Hong, J. Y. & Liebman, S. W. (2001) Cell 106, 171-182. 14. Derkatch, I. L., Che~noff, Y. O., Kushnirov, V. V., Inge-Vechtomov, S. G. & Liebman, S. W. (1996) Genetics 144,1375-1386. 15. Masison, D. C. & Wickner, R. B. (1995) Science 270, 93-95. 16. Osherovich, L. Z. & Weissman, J. S. (2001) Cell 106, 183-194. 17. Mabbott, N. A. & Bruce, M. E. (2001) J. Gen. Virol. 82, 2307-2318. 18. Bessen, R. A., Kocisko, D. A., Raymond, G. J., Nandan, S., Lansbury, P. T. & Caughey, B. (1995) Nature (London) 375, 698-700. 19. Caughey, B., Kocisko, D. A., Raymond, G. J. & Lansbury, P. T. (1995) Chem. Biol. 2, 807-817. 20. Safar, J., Wille, H., Itri, V., Groth, D., Serban, H., Torchia, M., Cohen, F. E. & Prusiner, S. B. (1998) Nat. Med. 4, 1157-1165. 21. Mestel, R. (1996) Science 273, 184-189. 22. Schlumpberger, M., Prusiner, S. B. & Herskowitz, I. (2001) Mol. Cell. Biol. 21, 7035-7046. 23. Zhou, P., Derkatch, I. L., Uptain, S. M., Patino, M. M., Lindquist, S. & Liebman, S. W. (1999) EMBO ]. 18, 1182-1191. 24. Derkatch, I. L., Bradley, M. E., Zhou, P. & Liebman, S. W. (1999) Cu~r. Genet. 35, 59-67. 25. King, C. Y. (2001) ]. Mol. Biol. 307, 1247-1260. 26. Kushnirov, V. V., Kryndushkin, D. S., Boguta, M., Smirnov, V. N. & Ter- Avanesyan, M. D. (2000) Curr. Biol. 10,1443-1446. 27. Liebman, S. W. & All-Robyn, J. A. (1984) Curr. Genet. 8, 567-573. 28. Glover, J. R., Kowal, A. S., Schirmer, E. C., Patino, M. M., Liu, J. J. & Lindquist, S. (1997) Cell 89, 811-819. 29. Chien, P. & Weissman, J. S. (2001) Nature (London) 410, 223-227. 30. Uptain, S. M., Sawicki, G. J., Caughey, B. & L~ndquist, S. (2001) EMBO J. 20, 6236-6245. 31. Kochneva-Pervukhova, N. V., Chechenova, M. B., Valouev, I. A., Kushnirov, V. V., Smirnov, V. N. & Ter-Avanesyan, M. D. (2001) Yeast 18, 489-497. 32. Chernoff, Y. O., Lindquist, S. L., Ono, B., Inge-Vechtomov, S. G. & Liebman, S. W. (1995) Sc~ence 268, 880-884. Bradley et al. eliminated by overexpressing just Ydjl. Possibly overexpressing HsplO4 or Hsp70 along with Ydjl would eliminate some of these incurable tPIN+] variants. Indeed, overexpression of Ydjl together with Ssal or Ssbl did cause loss of a weak tPSI+] (264. Our results show that tPIN+] prions exhibit strain variation, as do tPSI+] and tU^3] (14, 22), further supporting the hypothesis that strain variation is compatible with the protein-only model for priors. The recent suggestions that prions affect the appearance of other prions (13, 16) and that prion strain variation and species barrier are related phenomena (29) indicate that elucidating the molecular basis of tPIN+] strain vanation may yield impodant clues about the transmissibility of prion diseases between different species. We thank D. Bishop, R. Esposito, and S. Lindquist for strains, plasmids, and antibodies; I. Derkatch, J. Gavin-Smyth, and Y. Vitrenko for helpful comments about the manuscript; and J. Gavin-Smyth for help with the figures. This work was partially supported by a grant from the National Institutes of Health (GM56350) to S.W.L. 33. Moriyama, H., Edskes, H. K. & Wickner, R. B. (2000) Mol. Cell. Biol. 20, 8916-8922. 34. Sondheimer, N., Lopez, N., Craig, E. A. & Lindquist, S. (2001) EMBO J. 20, 2435-2442. 35. Glover, J. R. & Lindquist, S. (1998) Cell 94, 73-82. 36. Parsell, D. A., Kowal, A. S., Singer, M. A. & Lindquist, S. (1994) Nature (London) 372, 475-478. 37. Chernoff, Y. O., Newnam, G. P., Kumar, J., Allen, K. & Zink, A. D. (1999) Mol. Cell. Biol. 19, 8103-8112. 38. Newnam, G. P., Wegrzyn, R. D., Lindquist, S. L. & Chernoff, Y. O. (1999) Mol. Cell. Biol. 19,1325-1333. 39. Jung, G., Jones, G., Wegrzyn, R. D. & Masison, D. C. (2000) Genetics 156, 559-570. 40. Aigle, M. & Lacroute, F. (1975) Mol. Gen. Genet. 136, 327-335. 41. Sherman, F., Fink, G. R. & Hicks, J. B. (1986) Methods in Yeast Genetics (Cold Spring Harbor Lab. Press, Plainview, NY). 42. Tuite, M. F., Mundy, C. R. & Cox, B. S. (1981) Genetics 98, 691-711. 43. Goldring, E. S., Grossman, L. I., Krupnick, D., Cryer, D. R. & Marmur, J. (1970) J. Mol. Biol. 52, 323-335. 44. Chernoff, Y. O., Derkach, I. L. & Inge-Vechtomov, S. G. (1993) Curr. Genet. 24, 268-270. 45. Kane, S. M. & Roth, R. (1974) J. Bacteriol. 118, 8-14. 46. Herskowitz, I. & Jensen, R. E. (1991) Methods Enzymol. 194, 132-146. 47. Edskes, H. K., Hanover, J. A. & Wickner, R. B. (1999) Genetics 153, 585-594. 48. Zhou, P., Derkatch, I. L. & Liebman, S. W. (2001) Mol. Microbiol. 39, 37-46. 49. Winzeler, E. A., Shoemaker, D. D., Astromoff, A., Liang, H., Anderson, K., Andre, B., Bangham, R., Benito, R., Boeke, J. D., Bussey, H., et al. (1999) Science 285, 901-906. 50. Kochneva-Pervukhova, N. V., Poznyakovski, A. I., Smirnov, V. N. & Ter- Avanesyan, M. D. (1998) Cu~. Genet. 34,146-151. 51. Ter-Avanesyan, M. D., Kushnirov, V. V., Dagkesamanskaya, A. R., Didichenko, S. A., Chernoff, Y. O., Inge-Vechtomov, S. G. & Smirnov, V. N. (1993) Mol. Microbiol. 7, 683-692. 52. Edskes, H. K. & Wickner, R. B. (2000) Proc. Natl. Acad. Sci. USA 97, 6625-6629. 53. Conde, J. & Fink, G. R. (1976) Proc. Natl. Acad. Sci. USA 73, 3651-3655. 54. Maddelein, M. L. & Wickner, R. B. (1999) Mol. Cell. Biol. 19, 4516-4524. 55. Kochneva-Pervukhova, N. V., Paushkin, S. V., Kushnirov, V. V., Cox, B. S., Tuite, M. F. & Ter-Avanesyan, M. D. (1998) EMBO J. 17, 5805-5810. 56. DePace, A. H., Santoso, A., Hillner, P. & Weissman, J. S. (1998) Cell 93, 1241-1252. 57. Edskes, H. K., Gray, V. T. & Wickner, R. B. (1999) Proc. Natl. Acad. Sci. USA 96, 1498-1503. 58. Sanchez, Y., Taulien, J., Borkovich, K. A. & Lindquist, S. (1992) EMBO J. 11, 2357-2364. 59. Brewer, B. J. & Fangman, W. L. (1980) Proc. Natl. Acad. Sci. USA 77, 5380-5384. 60. Speransky, V. V., Taylor, K. L., Edskes, H. K, Wickner, R. B. & Steven, A. C. (2001) J. Cell. Biol. 153, 1327-1336. PNAS | December 10, 2002 1 vol. 99 | suppl. 4 | 16399