Prion-Dependent Lethality of sup45 Mutants in Saccharomyces cerevisiae

Abstract

In yeast Saccharomyces cerevisiae, translation termination factors eRF1 (Sup45) and eRF3 (Sup35) are encoded by essential genes SUP45 and SUP35, respectively. Heritable aggregation of Sup35 results in formation of the yeast prion [PSI+]. It is known that combination of [PSI+] with some mutant alleles of the SUP35 or SUP45 genes in one and the same haploid yeast cell causes synthetic lethality. In this study, we perform detailed analysis of synthetic lethality between various sup45 nonsense and missense mutations on one hand, and different variants of [PSI+] on the other hand. Synthetic lethality with sup45 mutations was detected for [PSI+] variants of different stringencies. Moreover, we demonstrate for the first time that in some combinations, synthetic lethality is dominant and occurs at the postzygotic stage after only a few cell divisions. The tRNA suppressor SUQ5 counteracts the prion-dependent lethality of the nonsense alleles but not of the missense alleles of SUP45, indicating that the lethal effect is due to the depletion of Sup45. Synthetic lethality is also suppressed in the presence of the C-proximal fragment of Sup35 (Sup35C) that lacks the prion domain and can not be included into the prion aggregates. Remarkably, the production of Sup35C in a sup45 mutant strain is also accompanied by an increase in the Sup45 levels, suggesting that translationally active Sup35 up-regulates Sup45 or protects it from degradation.

Introduction

Two translation termination factors, eRF1 and eRF3, participate in termination of protein synthesis in eukaryotes (reviewed in ref. 1). In the yeast S. cerevisiae these termination factors are encoded by the SUP45 and SUP35 genes respectively (reviewed in ref. 2). Both proteins are essential, and they interact with each other.3,4 The major Sup45 binding site is located between amino acid (aa) positions 465 and 685 within the essential C-proximal region (Sup35C) of S. cerevisiae Sup35.5,6,7

In S. cerevisiae and also in some other budding yeast species, the N-proximal domain of Sup35 (Sup35N) is responsible for the generation and propagation of the [PSI⁺] prion, a self-perpetuating amyloid-like aggregate of Sup35 (reviewed in refs. 8 and 9). Aggregation of Sup35 in [PSI⁺] cells10,11 results in defective translation termination, that leads to read-through of nonsense-codons, or omnipotent nonsense suppression.12,13 [PSI⁺]-mediated read-through of terminator codons (nonsense-suppression) is dominant, while nonsense-suppression caused by sup35 or sup45 mutations is recessive (reviewed in ref. 14). [PSI⁺] variants of different suppressor efficiencies and mitotic stabilities, that are analogous to prion “strains” observed in mammals, can be induced by Sup35 overproduction and faithfully maintained in one and the same genotypic background.15

Excess Sup45 inhibits de novo induction of [PSI⁺] by overproduced Sup35 in yeast, but does not affect propagation of preexisting [PSI⁺].16 Possibly, interaction with Sup45 prevents Sup35 aggregation, needed for the formation of the initial prion “nuclei”. Once formed, the prion complex becomes resistant to disruptive influence of Sup45 and possibly other proteins interacting with eRF3. Consistent with the location of the major Sup45-binding site within Sup35C, excess Sup45 has little or no effect on [PSI⁺] induction by overproduced Sup35N (Chernoff Y and Newnam G, unpublished).

Previously, it was reported that a [PSI⁺] diploid heterozygous for the sup45 disruption grows slowly, compared to isogenic [psi^-] derivatives, and is completely unable to sporulate. Treatment with GuHCl that is known to convert Sup35 into a non-prion form ([psi^-]), restored growth and sporulation of such a diploid.17 It was suggested that either complete inactivation of one of the two SUP45 alleles in the diploid strain results in decreased levels of Sup45, or presence of the [PSI⁺] prion leads to inactivation of a significant proportion of the Sup45 protein. In accordance with the latter hypothesis, it was reported that in some [PSI⁺] cells, Sup45 is found mostly in the aggregated fraction, possibly due to its recruitment by Sup35^PSI+ aggregates.5,18 However, other experiments with different [PSI⁺] strains have not confirmed coaggregation of Sup35 and Sup45.10,19

Previously, it was shown that combinations of a specific variant of [PSI⁺] with some (but not all) mutant alleles of sup45 or sup35 are unviable in a haploid state, as a diploid [PSI⁺] strain heterozygous for such a mutant allele has not produced viable mutant ascospores after meiosis (reviewed in ref. 20). To our knowledge, none of the sup45 or sup35 mutant alleles exhibiting such a synthetic lethality with [PSI⁺] has ever been sequenced, so that the molecular basis of mutational alterations remained unknown. Synthetic lethality was also detected between some uncharacterized mutant alleles of SUP45 and non-mendelian factor [ETA⁺],21 which was later identified as a variant of [PSI⁺].22 Another variant of [PSI⁺] used in the same study did not show synthetic lethality, suggesting that such an effect could be variant-specific.21

In the present study, we demonstrate synthetic lethality of well-characterized missense and nonsense-alleles of SUP4523,24 with different variants of [PSI⁺]. Our data show that synthetic lethality is a common trend observed for various mutant/prion combinations. We have also observed that in some combinations, lethality is dominant and occurs at postzygotic stage, and that lethality coincides with a decreased level of Sup45 protein in the yeast cell.

Materials and Methods

Strains.

Escherichia coli strain used was XL1-Blue (recA1, endA1, gyrA96, thi-1, hsdR17, supE44, relA1 [F', proAB, lacIq, Δ(lacZ)M15, Tn10(tet)]).25

The Saccharomyces cerevisiae strains used in this study (Table 1) contain the UGA allele ade1-14 or UAA allele ade2-1 used as a reporter for [PSI⁺] detection. Strains OT55 [PSI⁺] and OT56 [PSI⁺] are derivatives of 74-D694.26 The strain BSC783/4c27 contains the ade2-1 (UAA) reporter allele and tRNA suppressor SUQ5 (SUP16), capable of suppressing ade2-1 only in the presence of [PSI⁺]. All [PSI⁺] and [psi^-] strains also contain a prion isoform of the Rnq1 protein, [PIN⁺],28,29 which facilitates de novo induction of [PSI⁺] but affects neither [PSI⁺] propagation nor translation termination per se. Previously characterized mutant alleles of sup4523,24 used in this study included nonsense alleles (sup45-n), such as sup45–101 (266E→TAA), sup45–102 (53Y→TAA), sup45–104 (283L→TAA), sup45–105 (385E→TAA) and sup45–107 (317L→TGA), as well as missense alleles (sup45-m), such as sup45–103 (L21S), sup45–111 (R65C), sup45–113 (M48I), sup45–115 (S70F) and sup45–116 (R62T).

Yeast plasmids.

Yeast plasmids used in this study are listed in Table 2. The centromeric (CEN) plasmids pRS315/SUP45, pRS315/sup45 and pRS316/SUP4523 expressing Sup45 or its mutant alleles from P_SUP45 promoter, pRSU1C expressing the C-terminal domain of Sup35 from the promoter of SUP35 gene,30 pYS-GAL10426 and pFL39GAL-SUP35N31 expressing Hsp104 or Sup35N, respectively, from the galactose-inducible (P_GAL) promoter were described earlier. The multicopy plasmid pYX242/SUP35MC was described previously as pYX242/SUP35-ΔN.32 Vectors pRS315, pRS316GAL, pYX242 were used as negative controls.

Genetic and microbiological procedures.

Standard rich YPD medium, synthetic complete SC medium and selective media lacking individual components of SC were used.33 Medium with galactose contained 2% galactose and 2% raffinose instead of glucose. Yeast strains were grown at 25°C. Yeast transformation was performed by the lithium acetate procedure.34 For E. coli transformation the high-efficiency procedure35 was used. For the plasmid shuffle36 selective medium containing 1 mg/ml 5-fluoroorotic acid (5-FOA, Sigma) was used. Quantitative mating assay was performed as described.37 Briefly, about 5 x 10⁶ cells from the exponentially growing cultures of each parent were mixed on the 45-µm nitrocellulose filter. These mixtures were then incubated for 2–3 h at 25°C. The cells were subsequently fixed in methanol: acetic acid (3:1) on ice for 1 h and washed two times with PBS (pH 7.4). 4′,6′-diamidino-2-phenylindole (DAPI) was added at 1 µg/ml for 5 min, and the cells were washed with PBS. Zygotes were then identified by differential interference contrast microscopy and fluorescence microscopy, using the ZEISS Axiolab system. For each mating, 200 cells identified as zygotes were analyzed. The standard error of proportion was calculated as √[p(1-p)/n], where p is a proportion of cells with the phenotype in question and n is a total number of the cells.38 Cell-to-cell mating test was performed by placing cells of opposite mating type into direct contact with each other on YPD agar medium using micromanipulator.

Assays for [PSI⁺] formation and curing.

The presence of [PSI⁺] was monitored by its ability to cause read-through of ade1-14 (UGA), which leads to growth on medium lacking adenine (-Ade) and lighter color on complete (YPD) medium due to nonsense read-through (see ref. 39). “Weak” ([PSI⁺]^w) and “strong” ([PSI⁺]^s) variants of [PSI⁺] were distinguished from each other by the efficiency of suppression of the ade1-14 reporter construct as described previously.15 De novo [PSI⁺] formation was induced by transient overexpression of Sup35N from the galactose-inducible P_GAL promoter as described previously.15,39 For this purpose, a [psi^- PIN ⁺] yeast strain bearing the plasmid was incubated on a galactose medium, selective for the corresponding plasmid, for 3–4 days to induce P_GAL, followed by transfer to the -Ade/glucose medium where [PSI⁺] is detected by growth. The inducing plasmid was lost from the resulting [PSI⁺] colonies via incubation in non-selective conditions. Dominance of the de novo obtained ]PSI⁺] derivatives of the strain 1A-D1628 was confirmed by mating to the [psi^-] strain 29V-P2156 (MATa ade1-14 his7-1 met13-A1). The non-mendelian nature of these [PSI⁺] derivatives was confirmed by guanidine-HCl (GuHCl) curability as described.39

To cure yeast cells of [PSI⁺] by the transient overproduction of the Hsp104 chaperone,26 a [PSI⁺] culture was transformed with the plasmid pYS-GAL104 bearing the HSP104 gene under the P^GAL promoter, and incubated on galactose medium selective for the plasmid for 3–5 days, followed by colony purification on YPD medium. As a result, the Ura- Ade- colonies that have lost both [PSI⁺] and plasmid were isolated.

DNA and protein analysis.

For DNA sequencing of both strands of the SUP45 ORF PCR-amplified from the strains 90-D201 and K62-90-D201, following previously described23 primers were used: 82 and 83 for amplification, and 82, 83, 87, 92, 93, 100 for sequencing. Protein isolation, SDS-PAGE electrophoresis, and Western blotting were performed as described previously.23 Antibodies specific to Sup45 and Sup35 were described previously.23,40 The monoclonal anti-α-tubulin antibodies were described before.41 The alkaline-phosphatase-coupled anti-rabbit immunoglobulin G (Jackson) (for Sup45 and Sup35) or alcaline-phosphatase-coupled anti-mouse immunoglobulin G secondary antibodies (Jackson) (for tubulin), used as secondary antibodies, were purchased by Amersham Biosciences ECF system. Signals were quantified with STORM 840 Phosphor-Imager (Molecular dynamics, USA) and ImageQuantNT 5.2 software.

Results

Nonsense and missense alleles of SUP45 exhibit synthetic lethality with both “strong” and “weak” variants of [PSI⁺] in a haploid strain.

Previously we have shown that some nonsense mutations (sup45-n) in the essential yeast gene SUP45, encoding translation termination factor Sup45 (eRF1), decrease but do not completely abolish accumulation of full-size Sup45.23 Strains with the nonsense mutations remain viable in various [psi^-] genotypic backgrounds. These mutants produce residual amounts of full size Sup45 due to occasional read-through of a nonsense-codon in the conditions when supply of this termination factor is reduced. Missense mutations in SUP45 (sup45-m) that do not affect viability in the [psi^-] background also do not decrease levels of Sup45, in contrast to nonsense mutations.24 To check whether sup45-n and sup45-m alleles are capable of maintaining viability of the [PSI⁺] strains, we employed the plasmid shuffle technique as described below. First, independent [PSI⁺] isolates were induced in the strain 1A-D1628 (Table 1) containing SUP45 deletion (sup45Δ) on the chromosome and wild-type SUP45 allele on the pRS316-based (URA3) plasmid. For this purpose, this strain was transformed with pFL39GAL-SUP35N (TRP1) plasmid that carries coding sequence for Sup35N-domain under the control of GAL1-promoter. The transformants were incubated on selective medium with galactose as a sole carbon source for four days in order to overproduce Sup35N, and shifted to -Ade/glucose medium. Ade⁺ clones were isolated, cured of the TRP1 plasmid and subdivided into two groups (”strong” and “weak”) based on their ability to grow on -Ade medium. “Strong” and “weak” [PSI⁺] variants grew on -Ade after 3–4 and 5–6 days of incubation, respectively. Dominance and GuHCl curability of [PSI⁺] in these variants were confirmed as described in Materials and Methods (data not shown). Representative isolates of the strong and weak groups, designated as [PSI⁺]^S and [PSI⁺]^W, respectively, were used in further experiments. At the next stage, both these isolates and the isogenic control [psi^-] strain were transformed with the pRS315 (LEU2) based plasmids carrying either wild-type or mutant alleles of SUP45. Resulting transformants were replica-plated onto 5-FOA medium that counter selects against the original URA3-based SUP45 plasmid (Fig. 1). All [psi^-] derivatives were able to lose the URA3 SUP45 plasmid in the presence of either LEU2 plasmid, confirming that both wild-type and all mutant alleles of SUP45 confer viability in the [psi^-] background. In contrast, the [PSI⁺]S derivatives were not able to lose the URA3 SUP45 plasmid in the presence of LEU2 plasmids carrying any mutant sup45 allele. This confirms that the [PSI⁺]^S prion variant is synthetically lethal with each mutant allele of SUP45 used in this work. By using the same approach, we have shown that the [PSI⁺]^W prion is also synthetically lethal with all missense alleles of SUP45, with the exception of two missense allele (sup45–113 and sup45–115) that cause only weak nonsense-suppressors. Combination of [PSI⁺]^W with the nonsense alleles of SUP45 resulted in heterogenous (papillated) growth, indicating that only a fraction of the cells are viable. These data show that both “weak” and “strong” variants of [PSI⁺] exhibit synthetic lethality with the majority of the missense and nonsense mutant alleles of SUP45.

Combination of two different Sup45 mutant alleles in a diploid [PSI⁺] strain leads to cell lethality.

Next, we checked the interaction between two different sup45 alleles in a [PSI⁺] background by using previously described strains 90-D201 and K62-90-D201. K62-90-D201 was selected as a derivative of the [PSI⁺] strain 90-D201, which bears the sup45 mutation increasing the suppression efficiency of [PSI⁺] (Tikhodeev ON, unpublished). The [psi^-] derivatives of K62-90-D201 and 90-D201 were obtained by “curing” of [PSI⁺] from the original [PSI⁺] strains. Sequencing confirmed that 90-D201 carries the wild-type SUP45 allele that is identical to one from the strain 1B-D1606,23 while K62-90-D201 carries a mutant allele designated as sup45-k62 that contains a single nucleotide substitution C112T that results in amino acid substitution P38S.

After mating the [PSI⁺] and [psi^-] variants of strains 90-D201 and K62-90-D201 individually to derivatives of the strain 1A-D1628 (Fig. 2), we observed that all [psi^-] diploids, as well as all [PSI⁺] diploids, bearing at least one wild-type SUP45 allele are viable. However, no viable diploids were detected in the [PSI⁺] background for any combination of the sup45-k62 allele with any nonsense or missense-allele of SUP45, including alleles that were compatible with [PSI⁺]^W in a haploid state. This confirms previous observations showing that some heteroallelic sup45 combinations are not viable in the [PSI⁺] background even when one of the alleles is compatible with the same variant of [PSI⁺] on its own in a haploid strain (Tikhodeev ON, unpublished) and demonstrates that such combinatorial lethality is observed for all combinations of the sup45-k62 allele with any other missense or nonsense allele of the SUP45 gene.

Strong [PSI⁺] variant is synthetically lethal with heterozygous mutant alleles of SUP45.

Next, we examined whether synthetic lethality between [PSI⁺] and mutant alleles of SUP45 is ameliorated in a diploid strain containing the wild-type allele of SUP45. For this purpose, we checked whether sup45 mutants are able to produce viable diploids after mating to a wild-type [PSI⁺] strain. Thus, derivatives of the strain 1A-D1628 were mated to the isogenic strains OT56 or OT55 carrying “strong” and “weak” variants of [PSI⁺] respectively (see Refs. 15,42). While all sup45 mutants tested produced viable diploids with [PSI⁺]^W, all nonsense-alleles and some missense-alleles (sup45–103, sup45–111, sup45–116) turned out to be synthetically lethal with [PSI⁺]^S even in the presence of wild-type copy of the SUP45 gene (Fig. 3A). Interestingly, all mutant alleles of SUP45 that produced synthetic lethality were previously characterized as strong suppressors.23,24 To determine whether the lethality was solely due to [PSI⁺], strains OT55 and OT56 were cured of [PSI⁺] rather than [PIN⁺] by Hsp104 overproduction,28 and matings were repeated. In all cases, viable diploids were formed (Fig. 3A).

In further experiments we ascertained that phenomenon of synthetic lethality is not strain-specific and does not depend on mating type. We used isogenic [PSI⁺]S strains GT81-1C and GT81-1D43 of opposite mating types that have genetic background different from the background of OT56. The haploid isogenic yeast strains 1A-D1628 and 1B-D1628, carrying sup45Δ on the chromosome and either the wild-type allele of SUP45 gene or its mutant alleles on centromeric LEU2 plasmid, were mated to these strains, as well as to their [psi^-] derivatives cured of [PSI⁺] by transient overproduction of Hsp104.26 The results (not shown) were the same as those observed in the 1A-D1628 and OT56 [PSI⁺]^S combination, confirming that all strong suppressor alleles of sup45 (but not some weak missense alleles) are synthetically lethal with strong [PSI⁺] variant even in a heterozygous state.

Synthetic lethality in diploids occurs at postzygotic stage.

There could be three alternative explanations for the inability of sup45 mutants to form viable diploids in crosses with the [PSI⁺] strains: (i) mating defect; (ii) death of zygotes; and (iii) lethality in the mitotic divisions following zygote formation.

To distinguish between these options, we performed quantitative mating experiments (see Materials and Methods). Mating mixtures were stained with DAPI, and the zygotes were visualized by microscopy. Although sup45 mutants produced zygotes less efficiently than wild-type strains, [PSI⁺] affected neither frequency of zygote formation (Table 3) nor karyogamy (data not shown). In all crosses, the majority of zygotes exhibited normal morphology. Thus, the lack of viable diploids in the crosses between the [PSI⁺] strain and sup45 mutants was not due to a mating defect.

In order to compare the viability of zygotes, the fate of the individual zygotes obtained in controlled cell-to-cell mating experiments was monitored under the microscope. For this purpose, the wild-type control strain and all sup45 mutants shown in Figure 3 were mated to the [PSI⁺]^S strain OT56. At least ten individual zygotes were analyzed for each combination. In all cases, diploid zygotes were formed and started budding after two hours of incubation, indicating that diploids generated by mating are initially alive and capable of mitotic proliferation. However, only zygotes generated by the wild-type strain or by the weak missense mutants sup45–113 and sup45–115 were able to produce microcolonies after two days of incubation (Fig. 3B). In all other cases, growth stopped after 2–6 cell divisions. These data show that sup45 mutants are able to form diploid cells by mating with [PSI⁺] strains, but mitotic progeny of these cells dies after only several division cycles.

Lethality depends on copy number of the wild-type SUP45 gene.

Next, we checked whether synthetic lethality could be modulated by copy number of the wild-type SUP45 gene. For this purpose, the same crosses as those in Figure 3A were performed by using derivative of the [PSI⁺]^S strain OT56 which carries an additional wild-type allele of the SUP45 gene on the centromeric URA3 plasmid under either its own (P_SUP45) or the galactose-inducible (P_GAL) promoter. Viability of diploids was restored in all crosses where SUP45 was expressed from P_SUP45 (data not shown). In the case of P_GAL constructs, sup45 mutants (with the exception of sup45–113 and sup45–115) produced viable diploids only on galactose where the P_GAL-SUP45 construct is induced (data not shown). Moreover, in all combinations that previously showed lethality in the absence of the extra-copy of SUP45, the URA3 SUP45 plasmid has become essential for the diploids and could not be lost on the 5-FOA medium counter selecting against URA3 (Fig. 3C and data not shown).

To measure the efficiencies of plasmid shuffle quantitatively, we have plated diploid cultures onto YPD medium at a density of 200 cells per plate, incubated them at 25°C for several days, and replica plated each plate onto two different media lacking uracil or leucine, respectively. In each case, a total of 400 to 1100 colonies from each of three independently obtained identical diploids were tested. While control dipoloid SUP45//sup45Δ [PSI⁺]^S was able to lose both [pRS316/SUP45] and [pRS315/SUP45] plasmids with the same efficiency (Fig. 3C), diploids bearing a mutant allele of SUP45 on the LEU2 plasmid produced very few (less than 1% in case sup45–105) or no (in case of sup45–103) colonies losing the URA3 plasmid with the wild-type SUP45 allele. Diploid Ura- colonies carrying sup45–105 allele on the LEU2 plasmid revealed suppressive phenotype before and after the GuHCl treatment probably due to the selection of the new suppressor or allosuppressor.

Taken together, these data show that synthetic lethality between the strong [PSI⁺] variant and mutant alleles of sup45 depends on of the number of wild-type SUP45 alleles per genome. [PSI⁺]^S diploids bearing one mutant allele and one wild-type allele are unviable, while isogenic and “isoprionic” diploids with one mutant allele and two wild-type alleles are viable. The simplest explanation of this phenomenon is that presence of [PSI⁺]^S leads to depletion of functional Sup45 that could be compensated by introduction of the extra copy of SUP45 (see Discussion for more detail).

Effects of suppressor tRNA on synthetic lethality.

Nonsense-suppressor tRNAs can rescue nonsense-mutations in sup45.23,44 To check whether mutant tRNA would also restore viability of diploids in our system, the 1A-D1628 derivatives bearing either wild-type SUP45 or various sup45 mutations were mated to the strain [PSI⁺] BSC783/4c27 bearing the dominant nonsense-suppressor allele of the gene SUQ5. SUQ512 which is also known as SUP1645 is an altered tRNA^Ser with a substitution in the anticodon (UCA → UAA),46,47 that is able to recognize the stop-codon UAA in mRNA.48 Suppressor efficiency of SUQ5 is known to be increased by [PSI⁺]12 at about ten fold.49 As expected, SUQ5 restored viability in all crosses that involved UAA sup45-n alleles but not in the crosses that involved UGA allele sup45–107 or missense alleles sup45–103, sup45–111 and sup45–116 (Fig. 4).

Synthetic lethality is compensated by Sup35C.

During termination of translation, Sup45 interacts with Sup35. Due to sequestration of Sup35 by prion aggregates, the [PSI⁺] cells contain decreased amount of soluble Sup35.10,11,22 To check whether levels of soluble Sup35 influence lethality, we have mated the derivatives of the strain 1A-D1628 [pRS316/SUP45], bearing either wild-type or mutant alleles of SUP45, to the strong [PSI⁺]^S strain OT56, containing a centromeric plasmid that produces Sup35C, a derivative of Sup35 lacking the prion (Sup35N) and middle (Sup35M) domains. Sup35C is fully functional in termination but cannot be sequestered into the prion aggregates due to lack of prion domain. Viable diploids were generated in all combinations, demonstrating that the presence of Sup35C in the zygote overcomes lethality. The same results were obtained with the [PSI⁺]^S strain that carried a multicopy plasmid producing Sup35MC, a derivative of Sup35 lacking the prion domain (Fig. 5). The difference in the incompatibility of the [PSI⁺]S and sup45 mutations in Figures 3 and 5 is plasmid-dependent. We suggest that the copy number of pRS315 and pRS316 plasmids in the cell could be different, that results in better viability of the diploid cells bearing sup45 alleles on the pRS316 plasmid compared to pRS315 plasmid (the most significant difference could be seen in the case of sup45–116 allele).

Earlier, it has been reported that overproduction of eRF3 (Sup35) results in stabilization of eRF1 (Sup45) in mammalian cells via rescuing eRF1 from degradation.50 To check whether the same mechanism could operate in the yeast cells, we measured amounts of Sup45 both in the absence and presence of overexpressed Sup35MC. For this purpose, the [PSI⁺] and [psi^-] variants of the strain OT56 were transformed with the plasmid pYX242 or its derivative encoding Sup35MC. In the presence of Sup35MC, the ratio of Sup45 to tubulin (used as a loading control) was increased 2.3- (in [psi^-] strain) to 3.5-fold (in [PSI⁺] strain) (Fig. 5B). These data demonstrate that an increased amount of Sup35MC is accompanied by over-accumulation of Sup45 in yeast cells.

Discussion

Previous reports indicated that some mutant alleles of SUP45 are synthetically lethal with some isolates of the [PSI⁺] prion,20,21 but it was not clear whether this phenomenon was of a general nature or restricted to only specific combinations. Our systematic analysis demonstrates that synthetic lethality is a typical characteristic of different variants of [PSI⁺] (both “strong” and “weak” ones) in combination with the majority of the sup45 mutant alleles ( essentially all nonsense-alleles and most missense-alleles).

Although SUP45 is an essential gene and even some small internal deletions in it are lethal,51,52 some nonsense-alleles of SUP45 are viable in the [psi^-] background.23,44,53 This is apparently due to the fact that lack of Sup45 causes read-through of its own mRNA, so that some residual amount of complete Sup45 protein could be generated. We observed that Sup45 production at 8% of the normal Sup45 level (for example, in case of nonsense allele sup45–102) is sufficient for viability.23 Also, it has been reported that reduction of the Sup45 or Sup35 amount to 10% of the wild-type level is still sufficient to maintain viability in 90% of the cells.54 However, combination of a nonsense allele of SUP45 with [PSI⁺] leads to lethality. One possible explanation could be that simultaneous shortage of both release factors, Sup45 (due to a nonsense-mutation) and Sup35 (due to its sequestration by prion aggregates) is harmful for the cell.

Decreased function of Sup45 in combination with decreased amount of functional Sup35 is also harmful, as the majority of the missense alleles of SUP45 also exhibited synthetic lethality with [PSI⁺] (see above), even though they do not affect Sup45 levels.24 Indeed, all the missense mutations studied in this work were located in the N-terminal part of Sup45 that is crucial for recognition of the stop codon.55

For the fist time, we report that some nonsense and missense alleles of SUP45 are synthetically lethal with the strong variant of [PSI⁺] even in a heterozygous state, that is, in the presence of the wild-type allele of SUP45. Remarkably, this lethality was ameliorated by increasing the copy number and/or expression level of the wild-type allele. This result resembles the previous observation that a diploid [PSI⁺] strain heterozygous for a disruption of SUP45 shows slow growth that is partially restored by reintroduction of the additional copy of SUP45.17 Taken together these results suggest that growth and viability of the [PSI⁺] strains depend on the amount of functional Sup45.

Indeed, UAA-suppressor tRNA^Ser, SUQ5, improves viability of the [PSI⁺] diploids heterozygous for UAA alleles (but not for UGA or missense alleles) of SUP45, apparently due to increased production of full-length Sup45. The suppressor effect of [PSI⁺] on the stop-codon read-through should have a self-regulatory mechanism. The stronger read-through is as a result of [PSI⁺] presence the more efficient translation termination is as a result of full-length Sup45 production.

Presence of tRNA suppressor shifts the balance toward the full-length Sup45 production that leads to improvement of the cells viability. Introduction of Sup35C, a derivative of Sup35 that lacks the prion domain and is not incorporated into prion aggregates, also ameliorated [PSI⁺]-dependent lethality of a heterozygous SUP45/sup45 diploid. Although it is possible that effect of Sup35C could be explained by the phenotypical restoration of the [psi^-] state due to presence of a non-prion derivative of the Sup35 protein, it is worth noting that production of Sup35C was also accompanied by an increase in the level of Sup45. It is likely that interaction with soluble Sup35C causes stabilization of Sup45 by protecting it from proteolytic degradation. This agrees with the previous observation that depletion of eRF3a destabilizes eRF1 in mammalian cells, while compensatory production of the complete or N-truncated eRF3 rescued the eRF1 levels.50 Increase in levels of Sup45 could provide an additional or even primary explanation for amelioration of the [PSI⁺] dependent synthetic lethality by Sup35C.

Taken together, our data confirm that there is a certain minimal threshold of the amount of functional release factors in the yeast cell, and that reduction of the release factor levels beyond this threshold results in drastic viability defect. Moreover, we demonstrate that simultaneous reduction in the amount of both release factors exhibits a synergistic defect in cell viability, compared to individual depletion of either release factor.

Abbreviations

sup45-n	=	nonsense sup45 mutant allele
sup45-m	=	missense sup45 mutant allele
[PSI+]S	=	“strong” [PSI+] variant
[PSI+]W	=	“weak” [PSI+] variant
PGAL	=	GAL1 promoter

Figures and Tables

Figure 1 Synthetic lethality between [PSI⁺] and mutant alleles of SUP45. [PSI⁺]^S, [PSI⁺]^W and [psi^-] derivatives of the strain 1A-D1628 [pRS316/SUP45] were transformed with the plasmids pRS315/SUP45 or pRS315/sup45. Resulting transformants were replica plated to 5-FOA medium to select against the URA3 plasmid pRS316/SUP45, and photographed after five days of incubation. Strong variant of [PSI⁺] ([PSI⁺]^S) reveals synthetic lethality with sup45 missense (m) and nonsense (n) alleles whereas weak variant of [PSI⁺] ([PSI⁺]^W) demonstrates lethality with some missense alleles and sublethality with all tested nonsense alleles of SUP45 gene.

Figure 2 Combination of two sup45 alleles in diploid [PSI⁺] strain is lethal. The [PSI⁺] and [psi^-] variants of haploid strains 90-D201 (wild-type SUP45) and K62-90-D201 (sup45-k62) (vertical lines) were mated to a haploid strain 1A-D1628 bearing either pRS315/SUP45 or pRS315/sup45 plasmid (horizontal lines). Diploids were selected by incubation on the medium lacking uracil and phenylalanine for five days. Combination of two different mutant alleles of SUP45 gene is lethal in diploid [PSI⁺] strain but not in [psi^-].

Figure 3 Heterozygous mutant alleles of SUP45 exhibit synthetic lethality with [PSI⁺]^S in SUP45 dose-dependent manner. (A) Isogenic strains OT55 [PSI⁺]^W (W) and OT56 [PSI⁺]^S (S) and their [psi^-] (-) derivatives (vertical lines) were mated to derivatives of the strain 1A-D1628 bearing the wild-type (WT), missense (m) or nonsense (n) allele of the SUP45 gene (horizontal lines). Plates were replica plated onto the medium selective for diploids, and incubated for four days. Synthetic lethality was observed in all [PSI⁺]^S SUP45/sup45 combinations, except two weak sup45 missense mutations (sup45–113 and sup45–115). (B) The cells of the [PSI⁺]^S strain OT56 were individually mated to the cells of the 1A-D1628 derivatives, bearing the wild-type (WT), missense (sup45–103) or nonsense (sup45–105) allele of the SUP45 gene on a plasmid. Photographs were made before the incubation (0) and after two hours (2h), four hours (4h) and two days (2d) of the incubation on YPD medium. At least ten individual zygotes were analyzed for each combination. Synthetic lethality in the individual zygotes was observed for [PSI⁺]^S SUP45/sup45 combinations after 2–6 cell divisions. (C) Plasmid loss assay was performed on diploids obtained in the cross of OT56 [PSI⁺]^S bearing pRS316/SUP45 plasmid with derivatives of 1A-D1628 (sup45Δ) transformed with pRS315/SUP45 (WT) or pRS315/sup45–103 (103) or pRS315/sup45–105 (105) plasmids. Diploids were tested by frequency of spontaneous plasmid loss at non-selective conditions. The mean values and standard deviations, calculated on the basis of three independent experiments, are presented in each case. [PSI⁺]^S diploids heterozygous for mutation in the SUP45 gene are viable only in the presence of additional copy of the SUP45 gene on the plasmid.

Figure 4 Mutant suppressor tRNA has a codon-specific effect on the synthetic lethality. [PSI⁺] and [psi^-] variants of BSC783/4c strain (vertical lines) carrying suppressor tRNA SUQ5 were mated with 1A-D1628 derivatives bearing missense or nonsense sup45 alleles (horizontal lines) on pRS315 plasmid and replica-plated on the medium selective for hybrids. Pictures were taken after four days of incubation.

Figure 5 Synthetic lethality of [PSI⁺]^S and mutant sup45 alleles is compensated by Sup35C. (A) OT56 [PSI⁺] strain and its [psi^-] derivative carrying centromeric or multicopy (*) plasmid that encodes C-terminal domain of Sup35 (vertical lines) were mated to derivatives of the strain 1A-D1628 (designed the same as in Fig. 2). Plates were replica plated onto the medium selective for diploids, and incubated for four days. Presence of Sup35C on centromeric or multicopy plasmid neutralize the synthetic lethality of [PSI⁺]^S and mutations in the SUP45 gene in diploid strain. (B) Levels of Sup45 are increased in the presence of C-terminal domain of Sup35. [PSI⁺]^S strain OT56 and its [psi^-] derivative were transformed with the plasmid pYX242/SUP35MC or control vector pYX242. Crude cell extracts was prepared, run on SDS-PAGE, transferred to nitrocellulose filter and reacted to antibodies against Sup45, Sup35 and tubulin. Blots were analyzed by densitometry, and Sup45 amounts were normalized by using tubulin as a loading control. Sup45/tubulin ratio in the control sample in the strain bearing an empty plasmid is taken as one in each case.

Table 1 Yeast strains

Strain	Genotype	Reference
1B-D1606	MATα ade1-14 his7-1 leu2-3, 112 ura3-52 trp1-289 lys9-A21	23
1A-D1628	MATα ade1-14 his3 leu2 ura3 trp1 lys2 SUP45::HIS3 [pRS316/SUP45]	23
1B-D1628	MATa ade1-14 his3 leu2 ura3 trp1 lys2 SUP45::HIS3 [pRS316/SUP45]	Moskalenko SE, unpublished
GT81-1C	MATa ade1-14 his3 leu2 ura3 trp1 lys2 *	43
GT81-1D	MATα ade1-14 his3 leu2 ura3 trp1 lys2 *	43
OT56	MATa ade1-14 his3Δ200 leu2-3,112 ura3-52 trp1-289 [PSI^+]^S*	15, 28, 42
OT55	MATa ade1-14 his3-Δ200 leu2-3,112 ura3-52 trp1-289 [PSI^+]W*	15, 28, 42
BSC783/4c	MATa ade2-1 his3-11,15 leu2-3,112 ura3-1 SUQ5 [PSI^+] *	27
90-D201	MATa ade2-144,717 his7-1 leu2-3,112 lys9-A21 pheA10	Tikhodeev ON, unpublished
k62-90D-201	MATa ade2-144,717 his7-1 leu2-3,112 lys9-A21 pheA10 sup45	Tikhodeev ON, unpublished

* Original strains were [PSI⁺]; [psi^-] derivatives were obtained as described in Materials and Methods.

Table 2 Yeast plasmids

Plasmid Name	Plasmid Type	Yeast Marker	Promoter/Expression Cassette
pRS316	CEN	URA3	None
pRS315	CEN	LEU2	None
pRS316/SUP45	CEN	URA3	P SUP45-SUP45
p RS315/SUP45	CEN	LEU2	P SUP45-SUP45
pRS315/sup45	CEN	LEU2	P SUP45-sup45
pYS-GAL104	CEN	URA3	P GAL-HSP104
pFL39GAL-SUP35N	CEN	TRP1	P GAL-SUP35N
pRSU1C	CEN	LEU2	P SUP45-SUP35C
pYX242	2 µ	LEU2	None
pYX242/SUP35MC	2 µ	LEU2	P TPI-SUP35MC

* pRS315/sup45 mean a series of plasmids bearing different mutant alleles of SUP45.

Table 3 Frequency of zygote formation in various strain combinations

Matinga	% of Zygotesb	% of Abnormal Zygotes Among all Zygotesc
[PSI^+] (wt) x [psi^-] (wt)	34.9 ± 1,5	0.8 ±0.5
[psi^-] (wt) x [psi^-] (wt)	34.1 ±1.5	0.3 ± 0.2
[PSI^+] (wt) x [psi^-] (sup45–103)	17.5 ± 1.2	4.7 ±1.6
[psi^-] (wt) x [psi^-] (sup45–103)	19.9 ±1,3	3.4 ±1.3
[PSI^+] (wt) x [psi^-] (sup45–105)	24.7 ± 1.4	6.4 ± 1.6
[psi^-] (wt) x [psi^-] (sup45–105)	22.0 ±1.3	4.5 ± 1.4

a Wild-type [PSI⁺] and [psi^-] cells (OT56) were mated on filters to either wild-type SUP45Δ [SUP45] cells (1A-D1628 [pRS315/SUP45]) or isogenic SUP45Δ [sup45-n] cells (1A-D1628 [pRS315/sup45])

b The mean percentages of the zygotes among the total cells in mating mixture are shown

c Zygotes were considered morphologically abnormal if they either showed greatly increased cell size or were composed of more than three cells.

Acknowledgements

We are grateful to S.E. Moskalenko for yeast strains and plasmids, G.P. Newnam for assistance, Y.O. Chernoff for very helpful discussion and assistance in the manuscript preparation, J P. Rousset and A.S. Borchsenius for the discussion of the experimental work. We also thank Michel Philippe for possibility to perform part of the work in his laboratory. This research was supported by grants from CRDF (RB1 2336 ST 02), NATO (CBP. NR.981898) and RFBR (07 04 00605).