Prion-dependent proteome remodeling in response to environmental stress is modulated by prion variant and genetic background

ABSTRACT A number of fungal proteins are capable of adopting multiple alternative, self-perpetuating prion conformations. These prion variants are associated with functional alterations of the prion-forming protein and thus the generation of new, heritable traits that can be detrimental or beneficial. Here we sought to determine the extent to which the previously-reported ZnCl2-sensitivity trait of yeast harboring the [PSI+] prion is modulated by genetic background and prion variant, and whether this trait is accompanied by prion-dependent proteomic changes that could illuminate its physiological basis. We also examined the degree to which prion variant and genetic background influence other prion-dependent phenotypes. We found that ZnCl2 exposure not only reduces colony growth but also limits chronological lifespan of [PSI+] relative to [psi−] cells. This reduction in viability was observed for multiple prion variants in both the S288C and W303 genetic backgrounds. Quantitative proteomic analysis revealed that under exposure to ZnCl2 the expression of stress response proteins was elevated and the expression of proteins involved in energy metabolism was reduced in [PSI+] relative to [psi−] cells. These results suggest that cellular stress and slowed growth underlie the phenotypes we observed. More broadly, we found that prion variant and genetic background modulate prion-dependent changes in protein abundance and can profoundly impact viability in diverse environments. Thus, access to a constellation of prion variants combined with the accumulation of genetic variation together have the potential to substantially increase phenotypic diversity within a yeast population, and therefore to enhance its adaptation potential in changing environmental conditions.


Introduction
Fungal prions are protein-based epigenetic elements capable of adopting a spectrum of self-propagating conformations that can lead to a variety of heritable phenotypes [1,2]. Nearly one dozen fungal prions have been identified to date [2]. In their native states, many of the fungal prion-forming proteins have roles in gene expression and signaling [2,3], and thus the changes in their functional states that occur upon switching to prion conformations lead to the acquisition of new heritable traits [4,5]. In contrast to the prion conformations of the mammalian prion protein, which are invariably associated with disease [6], these traits can be detrimental or beneficial; indeed, all known fungal prions have the potential to produce beneficial phenotypes in some environments [2]. In many instances, these beneficial traits relate to environmental stress resistance [7][8][9], and prion formation rates increase significantly during exposure to environmental stress conditions [10][11][12][13].
These observations, along with the conservation of prion-forming domains throughout evolution [14] and the widespread existence of prions in wild yeast populations [4,15] have led some researchers to propose that fungal prions serve as a mechanism for propagating altered cellular states that may facilitate adaptation by enhancing survival in rapidly fluctuating environments [4,16,17]. An alternative hypothesis, which suggests that fungal prions are diseases, is based on several observations. For example, one of the earliest studies examining the prevalence of prions in wild yeast strains failed to find either the [PSI + ] or [URE3] prions among 70 wild strains examined [18], suggesting these prions have a net deleterious effect on the host cells (notably, the [RNQ + ] prion was found in 11 of the 70 strains, and a later study found [PSI + ] in nearly 1.5% of strains examined and [RNQ + ] in more than 6% of wild strains [4] represent disease states [19], as does the existence of barriers to prion transmission both within and between yeast species [20]. A number of genes have been identified that, when introduced to cells in which they had been deleted, can cure prion variants that arose in their absence, leading to the proposal that yeast possess an array of anti-prion systems [21][22][23]. Additionally, nonprion roles for some prion-forming domains could explain their evolutionary conservation, and may suggest that prion formation could be an artifact of the nonprion functions of these domains [24][25][26]. One of the most intensively studied fungal prions is the [PSI + ] prion, which results from alternative conformations of the translation termination factor Sup35 [27][28][29]. Sequestration of Sup35 into amyloid aggregates compromises its function in translation termination, leading to an increased frequency of nonsense suppression. Read through of stop codons allows cells to sample genetic variation within the 3ʹ untranslated regions (UTRs) of genes to produce proteins with novel C-terminal extensions, which could potentially alter protein function, stability and abundance [16,17]. Since Sup35 has the potential to adopt a spectrum of distinct amyloid structures (prion variants) that impact the strength of the nonsense suppression phenotype [30][31][32], the phenotypic outcomes of the [PSI + ] prion depend on the unique prion variant present in each strain. Although the overall decrease in translational fidelity and many of the traits it bestows are detrimental [10,17], occasional new beneficial phenotypes conferred by the [PSI + ] prion can ultimately become genetically fixed and thus prion-independent, facilitating rapid adaptation to new environments [12,17].
In an attempt to assess the impact of the [PSI + ] prion on the yeast proteome, we previously employed Stable Isotope Labeling by Amino acids in Cell culture (SILAC)based quantitative mass spectrometry to compare the proteomes of isogenic [PSI + ] and [psi − ] strains [33]. Only a relatively small number of proteins (~3% of proteins quantified) exhibited significant changes in abundance between the isogenic [PSI + ] and [psi − ] strains in the standard laboratory growth conditions we used in that study. However, since many prion-dependent phenotypes manifest not in standard laboratory growth conditions but rather under conditions of environmental stress [16], more widespread prion-dependent remodeling of the proteome might only become evident in cells exposed to stress. Moreover, since prion-dependent phenotypes are heavily influenced by the genetic background of the strain [16], any prion-dependent changes in protein expression that occur during environmental stress are likely to be idiosyncratic to each strain.
In their seminal study examining [PSI + ]-dependent phenotypes in multiple genetic backgrounds across more than 150 growth conditions, True and Lindquist [16] observed constellations of prion-dependent phenotypes unique to each of the seven genetic backgrounds they studied. Only one environmental conditiongrowth on agar medium supplemented with 5 mM ZnCl 2consistently conferred a fitness advantage to [psi − ] cells among all genetic backgrounds tested, suggesting that altered zinc metabolism may be a universal feature of [PSI + ] cells. We therefore sought to determine the extent to which the previously-reported ZnCl 2 -sensitivity trait of strains harboring the [PSI + ] prion is modulated by genetic background and prion variant. We found [PSI + ] strains to be sensitive to ZnCl 2 relative to isogenic prion-free [psi − ] strains, with reduced colony growth and chronological lifespan. The degree of sensitivity was influenced by prion variant and genetic background, and quantitative proteomic analysis of cells exposed to ZnCl 2 during growth indicates that [PSI + ] cells exhibit increased cell stress and reduced energy metabolism compared to [psi − ] cells. Finally, we show that prion-dependent phenotypes and changes in protein abundance are profoundly influenced by prion variant and genetic background, thus enhancing phenotypic diversity within a population and potentially providing a mechanism for enhancing survival in fluctuating environments.

ZnCl 2 significantly reduces colony size of [PSI + ] relative to isogenic [psi − ] yeast
To confirm that [PSI + ] cells are more sensitive to ZnCl 2 exposure compared to isogenic [psi − ] cells, we used an S288C-based strain we previously constructed [33] with an induced [PSI + ] variant that produces a strong nonsense-suppression phenotype (determined by the ade1-14-based colony color assay), hereafter referred to as [PSI + ] SI-str (for S288C Induced strong variant). First, we cured the strain of the [PSI + ] prion by passage on medium containing 5 mM GuHCl to generate an isogenic [psi − ] strain. We then confirmed that the [psi − ] strain has a fitness advantage compared to the isogenic [PSI + ] SI-str strain by measuring colony size on synthetic defined (SD) agar supplemented with 5 mM ZnCl 2 . While the [PSI + ] SI-str strain produced significantly larger colonies than the isogenic [psi − ] strain on SD agar in the absence of ZnCl 2 (Figure 1(a)), the opposite was observed in the presence of ZnCl 2 (Figure 1(b)). Notably, we did not detect any fitness advantage for [psi − ] cells when grown in liquid SD medium supplemented with 5 mM ZnCl 2 (data not shown), and thus the ZnCl 2 -sensitivity trait for [PSI + ] SI-str may be confined to growth on solid media. The reduction in fitness as measured by colony size may be indicative of a reduction in lifespan and viability for the S288C [PSI + ] SI-str cells when exposed to ZnCl 2 . We therefore asked whether exposure to ZnCl 2 differentially affects chronological lifespan in S288C [PSI + ] SI-str and isogenic [psi − ] cells. We assessed viability of cells in colonies grown on SD agar with and without 5 mM ZnCl 2 over a 20-day time-course by staining cells with propidium iodide (PI) and measuring the fraction of inviable cells by flow cytometry. We found that ZnCl 2 reduces viability in both [psi − ] and [PSI + ] SI-str cells; however, by the end of the timecourse we observed a small but significant exacerbation of the ZnCl 2 -sensitivity phenotype in [PSI + ] SI-str cells (Figure 2(a)). By day 20 there was no significant difference in the relative viability of [psi − ] and [PSI + ] cells in the absence of ZnCl 2 , yet a nearly 50% reduction in viability of [PSI + ] relative to [psi − ] cells in the presence of ZnCl 2 . Taken together, our data indicate that the reduction in colony growth rate and viability induced by exposure to ZnCl 2 on solid media is modulated by the prion status of the yeast cells.
Is the [PSI + ] SI-str strain simply less fit than the isogenic [psi − ] strain under all stress conditions? To test if this is the case, we examined whether the [PSI + ] SI-str prion confers reduced viability in two other stress conditions: mild heat stress, and mild oxidative stress. Since the largest differences in viability between the isogenic [psi − ] and [PSI + ] strains (both with and without ZnCl 2 ) were observed at 15 days (Figure 2(a)), we examined 15-day viability in SD supplemented with 100 µM H 2 O 2 , and in SD at 37°C (Figure 2(b)). In contrast to growth at 30°C in SD and SD + ZnCl 2 , no significant differences in 15-day survival were observed between isogenic S288C [psi − ] and S288C [PSI + ] SI-str strains in the presence of H 2 O 2 or at 37°C (Figure 2(b)). Thus, the sensitivity phenotype of the [PSI + ] SI-str strain when exposed to ZnCl 2 is not a general stress sensitivity phenotype, but rather it is specific for this environmental condition.

ZnCl 2 enhances expression of stress response proteins and reduces expression of proteins involved in energy metabolism in [PSI + ] cells relative to isogenic [psi − ] cells
The reduction in fitness and viability for the [PSI + ] SI-str strain relative to the isogenic [psi − ] strain when exposed Colonies were incubated for~72 hours on SD agar or 96 hours on SD agar + ZnCl 2 . Center lines show the medians; box limits indicate the 25th and 75th percentiles as determined by R software; whiskers extend to 5th and 95th percentiles, outliers are represented by dots. Notches represent the 95% confidence interval for each median. Non-overlapping notches give~95% confidence that two medians differ. **p < 0.001 (Z-test). n = 236, 326, 166, 161 sample points.
to ZnCl 2 suggests that [PSI + ] SI-str cells experience more stress than [psi − ] cells in that environment. To determine whether the proteome of [PSI + ] SI-str cells is remodeled differently than that of isogenic [psi − ] cells following growth in the presence of ZnCl 2 , we used SILAC (Stable Isotope Labeling by Amino acids in Cell culture) followed by quantitative mass spectrometry to detect prion-dependent changes in protein abundance in colonies growing on SD agar supplemented with 5 mM ZnCl 2 . We collected data for 1267 proteins (~one quarter of the proteome) that were quantified in at least two of the three experiments using the Perseus platform integrated to MaxQuant (Table S1). Of those 1267 proteins, the levels of 37 and 87 were significantly increased or decreased in the [PSI + ] SI-str strain in comparison to the isogenic [psi − ] strain, respectively ( Figure 3), representing changes in abundance for~10% of the proteins quantified. Notably, we observed more than two-fold enrichment of some proteins in [PSI + ] cells, including the heat shock-induced Hsp30 and the acid stress response Yro2, whereas the Nop1 histone glutamine methyltransferase, and the Ina17 F1F0 ATPase synthase peripheral stalk assembly factor, were both reduced by more than 50% ( Figure 3).
Consistent with the slower growth and reduced viability in [PSI + ] cells compared to [psi − ] cells exposed to ZnCl 2 , proteins involved in stress responses (protein folding) were significantly more abundant (with >25% change in abundance) in [PSI + ] SI-str cells relative to isogenic [psi − ] cells (Table 1), while proteins involved in energy metabolism were significantly less abundant ( Table 2). Components of the translation machinery were also less abundant in [PSI + ] SI-str cells, though the false discovery rate (FDR) for this category was >0.05.

Prion variant and genetic background profoundly influence viability in diverse environments
Is the prion-dependent reduction in fitness and lifespan on SD agar supplemented with ZnCl 2 modulated by genetic background and prion variant? Previous studies [16] found ZnCl 2 sensitivity to be a universal trait among seven genetic backgrounds tested. We used an established protein transformation protocol [34] (Figure 4(b)). Indeed, the W303 strains all exhibited greater survival on SD supplemented with ZnCl 2 compared to the S288C strains. We also observed survival differences between strains in other growth conditions. For example, the S288C strains were only marginally viable after 15 days at 37°C, whereas approximately one quarter of W303 cells Error bars indicate SEM, n = 3 independent biological replicates per strain. * p < 0.05, ** p < 0.001 (student's t-test). (b). There is no significant difference in 15-day viability between [PSI + ] SI-str and isogenic [psi − ] exposed to H 2 O 2 or at 37°C. The [PSI + ] SI-str strain exhibits reduced 15-day viability compared to the isogenic [psi − ] strain on SD and SD + ZnCl 2 at 30°C, but no significant change in viability on SD agar supplemented with H 2 O 2 or on SD at 37°C. Error bars indicate SEM, n = 3-4 independent biological replicates per strain. * p < 0.05, ** p < 0.001 (student's t-test).
survived. Furthermore, we observed significant differences in viability based on prion status in both the presence and absence of ZnCl 2 . For example, at 30°C in the absence of ZnCl 2 , in both genetic backgrounds we observed a consistent relationship between survival and prion strength, with [ Prion-dependent changes in protein abundance are modulated by prion variant and by genetic background The substantial impact of genetic background and prion variant on survival phenotypes (Figure 4)   . Proteins with significantly increased or decreased abundance in the [PSI + ] SI-str strain relative to the isogenic [psi − ] strain when grown on SD agar supplemented with 5 mM ZnCl 2 . The SILAC approach followed by quantitative mass spectrometry was utilized to examine the relative abundance of proteins in the [PSI + ] SI-str strain relative to an isogenic [psi − ] strain. Proteins with <25% change in abundance (vertical shading) or for which changes in abundance were not significant (P > 0.05; horizontal shading) are indicated. Examples of proteins exhibiting significant changes in abundance of >50% (increase or decrease in [PSI + ] SI-str ) are identified by name (see Table S1 for a complete list).

Discussion
Yeast prions can produce a range of phenotypes that vary depending on environment and on the genotype of the strain. The [PSI + ] prion results in elevated levels of nonsense suppression, and thus has the potential to remodel the proteome to produce new traits. Our previous proteomic analysis, however, identified only relatively small changes between the proteomes of isogenic [PSI + ] and [psi − ] strains in standard laboratory growth conditions [33]. One [PSI + ]-dependent trait, sensitivity to ZnCl 2 , was previously found to be exhibited in all genetic backgrounds examined [16], suggesting that altered zinc metabolism may be a universal feature of [PSI + ] cells. We therefore examined the extent to which this phenotype, and other prion-dependent phenotypes, are influenced by genetic background and prion variant. We found that the ZnCl 2 -sensitivity trait of [PSI + ] cells is indeed modulated by genetic background and prion variant. Quantitative proteomic analysis of cells grown in the presence of ZnCl 2 identified significant priondependent changes in abundance for~10% of proteins quantified, with [PSI + ] cells exhibiting significant enrichment of stress response proteins and significant reduction of proteins involved in energy metabolism. More generally, we found that prion variant and genetic background profoundly influenced prion-dependent phenotypes and changes in protein abundance.
Our quantitative proteomic analysis revealed that relative to isogenic [psi − ] cells, [PSI + ] SI-str cells exposed to ZnCl 2 are enriched for heat shock proteins and exhibit reduced abundance of proteins involved in energy metabolism. These findings are consistent with the activation of stress responses and reduced growth in [PSI + ] cells compared to [psi − ] cells exposed to ZnCl 2 , and with previous studies suggesting that the [PSI + ] prion can lead to cell stress and enhance expression of heat shock proteins [38]. Whether the same subsets of proteins would exhibit altered abundance in other genetic backgrounds, or for different [PSI + ] variants, is unclear; however, our flow cytometry analysis of Ssa4-GFP, Hsp26-GFP, and Yro2-GFP indicate that priondependent changes in protein abundance that occur when cells are grown in the presence of ZnCl 2 are modulated by these factors (Figure 5). Since the phenotypic strength of the nonsense suppression phenotype of [PSI + ] is influenced by prion variant [30][31][32], the variant-dependent changes in protein abundance we observed are likely due to differences in stop codon read-through frequencies. Moreover, since the identity of the stop codon and the sequence context surrounding it impacts the frequency of nonsense suppression [39], the impact of the [PSI + ] prion on gene expression will vary from gene to gene. This natural variation in [PSI + ]dependent effects on gene expression was not captured by our flow cytometry-based protein abundance assay, as the natural stop codons and 3ʹUTR sequences of the Ssa4, Yro2, and Hsp26 genes were replaced by identical sequences due to the C-terminal GFP tag. It is also unclear if the [PSI + ]-dependent changes in the proteome contribute to the ZnCl 2 -sensitivity phenotype, or are instead a response to it. Previous studies have found that short-term exposure of yeast to ZnCl 2 leads to oxidative stress, through consumption of low molecular mass thiols and increased production of reactive oxygen species (ROS), and also leads to the induction of genes encoding chaperones (among others) [40]. Thus, additive effects on protein abundance due to induction of stress response genes by [PSI + ] [38] and by ZnCl 2 [40] could account for some of our observations.
We assessed three environmental conditions (30°C, ZnCl 2 , and 37°C) in two genetic backgrounds (S288C and W303) with three different prion states ([psi − ], [PSI + ] Sc37 , and [PSI + ] SI-str ). Although neither of the [PSI + ] variants in either genetic background conferred a significant survival advantage in these environments, at least one of them (the S288C [PSI + ] SI-str strain) did have an apparent growth advantage over the isogenic [psi − ] strain in the absence of ZnCl 2 in that it produced larger colonies (Figure 1(a)). Intriguingly, the cells in these colonies exhibited significantly reduced survival compared to the [psi − ] strain following extended incubation (Figure 2(b)). The larger colony size could potentially be due to the ade1-14 allele present in these strains, which enables [PSI + ] cells (but not [psi − ] cells) to synthesize adenine; however, since we used synthetic defined medium (and not YPD) for the colony size assay, which contained an abundant supply of adenine, this is unlikely to be the cause. Additionally, the extent to which [PSI + ] cells exhibited reduced survival varied considerably depending on environment, prion variant, and genetic background. Indeed, previous work has found that other [PSI + ] prion variants can prolong chronological lifespan in the 74-D694 and 5V-H19 genetic backgrounds [41]. Taken together, our data demonstrate that environment, genetic background and prion variant all interact to modulate prion-dependent proteome remodeling and phenotypic outcomes. Since many environmental stress conditions have been shown to enhance rates of prion formation [10][11][12][13], and different prion variants can spontaneously arise in the population [32,[42][43][44][45][46], the resulting phenotypic diversity could provide a mechanism for enhancing survival in fluctuating environments.

Yeast viability assays
Viability assays were performed as previously described [50,51]. Briefly, colonies on SD agar medium (with or without ZnCl 2 ) were harvested in phosphate-buffered saline and incubated at 30°C for 30 minutes with 2 µM propidium iodide (Molecular Probes). Viability was assessed by flow cytometry with a BD FACS Canto II flow cytometer using the PE 'area' parameter. For each yeast strain, three or four independent replicates with 50,000 cells per replicate were analyzed.
Proteomic profiling of [PSI + ] and [psi − ] strains exposed to ZnCl 2 Proteomic differences between [PSI + ] and [psi − ] strains exposed to 5 mM ZnCl 2 were assessed by SILAC (Stable Isotope Labeling by Amino acids in Cell culture) followed by quantitative mass spectrometry. Cells were diluted for single colonies on plates with SD-Lys-Arg medium supplemented with 5 mM ZnCl 2 and either heavy-(Lys 4 , Arg 6 ) or light-labeled arginine (20 mg/L) and lysine (30 mg/L), and then incubated for 4 days to allow for colony formation. Approximately 100 colonies were resuspended in sterile H 2 O, cell density in each suspension was determined by counting using a hemocytometer, and then suspensions of heavy and light-labeled cells were diluted to identical densities. Cells were lysed in SDS sample buffer in a Precellys bead beater and the lysates were cleared by centrifugation (16,200g for 20min at 4°C). About 60 µg of protein was run on a SDS-PAGE for in-gel trypsin digestion [52]. Quantitative mass spectrometry were then performed as previously described on an Impact II (Bruker Daltonics) on-line coupled to an EASY Nano-LC 1000 nanoflow HPLC (Thermo Scientific) [33]. Three independent replicates were performed and proteins were quantified using the Perseus platform (1.6.2.1) integrated to MaxQuant (1.6.2.1). The search was performed against the Saccharomyces Genome Database (Date stamp: 20110203). The search was configured with the following MaxQuant parameters: peptide mass accuracy 10 ppm with trypsin as the protease (K/R cleavage specificity), allowing a maximum of two missed cleavages, carbamidomethyl as fixed modification, and methionine oxidation, N-terminal acetylation, and asparagine and glutamine deamination as variable modifications. The false discovery rate was set below 1% at both the peptide and protein level.

Protein transformation with [PSI + ] SI-str and [PSI + ] Sc37 prion variants
Protein transformation to introduce the [PSI + ] SI-str and [PSI + ] Sc37 prion variants into the S288C and W303 genetic backgrounds was performed as described previously [34], with some modifications. Partially purified prion particles were prepared by harvesting mid-log phase cultures, washing cells in sterile water, then resuspending in lysis buffer (40mM Tris-HCl pH 7.4, 150mM KCl, 15mM MgCl 2 , protease inhibitor cocktail mini tablet (Pierce)). Cells were lysed by vortexing with glass beads, lysates were centrifuged at 10,000g for 5 min at 4°C and the supernatant subjected to ultra-centrifugation in a Beckman Coulter Airfuge at 30 psi for 25 min. Pellets were resuspended in 1 M lithium acetate, incubated on ice for 30 min with gentle agitation, and then again subjected to ultra-centrifugation in a Beckman Coulter Airfuge at 30 psi for 25 min. Pellets were resuspended in 5 mM potassium phosphate buffer (pH 7.4) containing 150 mM NaCl and sonicated on ice for 20 seconds (20% amplitude; 10 pulses of 1 second on, 1 second off).
Cells to be transformed were first grown sequentially (three times) on agar medium supplemented with 5 mM guanidine hydrochloride (ACROS Organics) to eliminate [PSI + ] and [RNQ + ] prions (elimination of [RNQ + ] by this treatment in these strains has been previously confirmed by lack of visible GFP foci in cells over-expressing Rnq1-GFP [48]; and our unpublished data). Cells were then treated with 100U of lyticase (Sigma) in 1 M sorbitol + 10 mM Tris pH 7.5 at 30°C for 1 hour to generate spheroplasts. Spheroplasts were collected by gentle centrifugation (400g, 4 min) and washed with 10 ml of 1 M sorbitol, harvested again and washed with 10 ml of STCbuffer (1 M sorbitol, 10 mM CaCl 2 , 10 mM Tris, pH 7.5), then collected once more and resuspended in 1 ml of STC-buffer. Spheroplasts (100 µl) were mixed with partially purified prions (final concentration~20-40 µg), URA3 marked plasmid (~3 µg) and salmon sperm DNA (15 µg), and incubated for 30 min at room temperature. Following addition of 9 volumes of PEG-buffer (20% [w/ v] PEG 3350, 10 mM CaCl 2 , 10 mM Tris, pH 7.5) and incubation at room temperature for 30 min, cells were collected by gentle centrifugation (400g, 4 min), resuspended with 150 µl of SOS-buffer (1 M sorbitol, 7 mM CaCl 2 , 0.25% yeast extract, 0.5% bacto peptone), and incubated at 30°C for 30 min. Cells were added tõ 7.5 ml molten SD-URA + 2.5% agar + 1 M sorbitol (held at~46°C), immediately mixed and plated over SD-URA agar. Colonies arising after several days were screened for prion status by color phenotype on ¼YEPD agar medium and confirmed by their ability to be cured to [psi − ] following growth on medium supplemented with 5 mM guanidine hydrochloride (ACROS Organics). Subsequent passage of cells on 5-Fluoroorotic acid (5-FOA; USBiologicals) agar medium selected for cells that had lost the URA-marked plasmid used during the protein transformation protocol.

Determination of relative protein abundance by flow cytometry
Relative protein abundance was determined by flow cytometry of strains expressing proteins with C-terminal GFP fusions as described previously [48], with the exception that cells were harvested from SD agar plates (with or without 5 mM ZnCl 2 ) grown for 3 days at 30°C. For each strain, GFP fluorescence intensity was measured for 50,000 cells using a BD FACS Canto II flow cytometer and analyzed using the FITC 'area' parameter. For each genetic background (S288C or W303), mean GFP fluorescence was normalized to the isogenic [psi − ] strain grown in the same condition.