Introduction

The ade6 gene of fission yeast encodes a phosphoribosylaminoimidazole carboxylase required for the de novo synthesis of purines (Szankasi et al. 1988). Cells with mutations in ade6 are auxotrophic for adenine, but grow as efficiently as wild type when adenine is added to the media (Gutz et al. 1974). On media with limiting amounts of adenine, wild-type cells form white colonies, whereas ade6 mutants accumulate a red pigment (Gutz et al. 1974). Early work based on these properties identified nearly 400 different alleles of ade6 and about 46,000 additional alleles were created in the following five decades [e.g. (Gutz 1971; Grimm et al. 1994; Steiner et al. 2009)]. These have provided powerful tools for analyses of many broadly conserved biological processes such as transcription; RNA processing and stability; chromatin remodeling; genetic recombination; DNA damage repair; and genome stability [e.g, (Hottinger and Leupold 1981; Bernardi et al. 1991; Allshire et al. 1994; Nimmo et al. 1994; Szankasi and Smith 1995; Kon et al. 1997; Mizuno et al. 1997; Kon et al. 1998; Mansour et al. 2001; Huang et al. 2005; Niwa et al. 2006; Hirota et al. 2008; Leem et al. 2008; Osman and Whitby 2009; Zhao et al. 2009; Chino et al. 2010; Gao et al. 2013)].

The ade6-M375 and ade6-M26 alleles are of particular interest. They each have a single G-to-T substitution that converts a glycine codon (GGA) to an opal stop codon (UGA), with M375 and M26 affecting adjacent codons (Fig. 1b) (Szankasi et al. 1988). Serendipitously, the M26 mutation also creates a cyclic AMP responsive element (CRE)-like DNA site (5′-ATGACGT-3′) (Schuchert et al. 1991) that is bound avidly by Atf1-Pcr1 (Mts1-Mts2) (Wahls and Smith 1994), which is a heterodimeric basic leucine zipper (bZIP) transcription factor of the ATF/CREB family (Wahls and Smith 1994; Kanoh et al. 1996; Kon et al. 1997). The M375 allele does not have a binding site for Atf1-Pcr1 heterodimer and thus serves as a nucleotide substitution type-matched, codon type-matched negative control. Such controls supported discoveries that the Atf1-Pcr1-M26 protein-DNA complex directly regulates multiple biological processes including chromatin remodeling (Davidson et al. 2004; Jia et al. 2004; Kim et al. 2004; Yamada et al. 2004), transcription of stress-responsive genes (Hirota et al. 2003; Davidson et al. 2004; Eshaghi et al. 2010), and the positioning of meiotic recombination at hotspots (Kon et al. 1997; Gao et al. 2008; Wahls and Davidson 2010).

Fig. 1
figure 1

Graphical summary and model. a Diagram of translation release factor Sup35 (eRF3) and its GTP elongation factor Tu domains (D1–D3). We identified an allele of sup35 that has a single base pair substitution (numbered from start codon in cDNA) and encodes a protein with a single amino acid substitution. b Reporter alleles of ade6 (M375 and M26) each have a stop codon (*) in place of a glycine codon. c Efficient termination of translation by Sup35 at the internal stop codons abolishes production of Ade6 protein. Inefficient termination by Sup35-F592S causes read-through of each stop codon (insertion of a random amino acid, X). The phenotypes of identical stop codons at adjacent positions are differentially suppressed, revealing that the local context of the stop codon affects the efficiency of suppression

Full size image

Phenotypes caused by stop codons can be suppressed by mutations in tRNA genes that change the sequence of the anticodon loop (Rafalski et al. 1979; Willis et al. 1984, 1986). For example, a single nucleotide substitution in the anticodon of a serine tRNA (5′-UGA-3′ to 5′-UCA-3′) allows it to bind to a UGA (opal) stop codon and insert a serine into the nascent polypeptide, thus allowing the ribosome to read through the stop codon (Willis et al. 1984). Such suppressors have provided useful tools to analyze the biological functions of specific nonsense alleles, such as ade6-M26 (Ponticelli et al. 1988).

In this study we used the opal stop codons of ade6-M26 and its control allele, ade6-M375, as tools to elucidate the molecular basis of two nonsense codon suppressors. There are two key findings. First, we localized the sup9 suppressor gene physically to the spctrnaser.11 locus and confirmed that the mechanism of suppression involves mutation of the anticodon of a serine tRNA (Willis et al. 1984). Second, we localized another suppressor gene to the sup35 locus and we determined that the mechanism of suppression involves an amino acid substitution in a GTP elongation factor Tu domain of the ubiquitous eukaryotic translation release factor 3 (eRF3). And as has been reported for translational read-through suppression in other organisms (Dalphin et al. 1997; Poole et al. 1998; Namy et al. 2001), we show that the local context of the stop codon also affects the efficiency of suppression in fission yeast.

Materials and methods

Yeast strains and culture

The genotypes of strains used in this study are provided in Table 1. Relevant sequences of the ade6 (Szankasi et al. 1988) and sup35 alleles are provided in the main text. The ade6-D1 allele lacks the ade6 ORF, allowing us to use ade6 as a selectable marker in plasmids and eliminating potential effects of chromosomal alleles on selection. Culture media, culture conditions, genetic crosses, and scoring of genetic markers were as described (Gutz et al. 1974; Forsburg and Rhind 2006; Gao et al. 2008; Kan et al. 2011). For rich media we used yeast extract liquid (YEL) or agar (YEA); for minimal media we used nitrogen base liquid (NBL) or agar (NBA) supplemented as necessary with amino acids and nucleobases at 100 µg/ml; and for mating we used sporulation agar (SPA). For initial strain constructions and experiments prior to identifying the gene responsible for the sup† phenotype, we used test crosses and tetrad spore colony phenotyping (ability to suppress adenine auxotrophy caused by nonsense codons in ade6) to follow the markers. Once we identified the sup† mutation (sup35-T1775C of cDNA sequence from start codon), we used PCR and DNA sequencing to genotype alleles at the sup35 locus. Alleles of spctrnaser.11 were also analyzed by PCR and DNA sequencing.

Table 1 Genotypes of S. pombe strains used in this study
Full size table

Tetrad dissection

Procedures for mating, isolation of conjugants and tetrad dissection using a Singer MSM300 microdissection apparatus were according to instructions of the manufacturer (Singer Instrument Co. Ltd., Somerset, UK). Spores from each tetrad were plated in grids on YEA and incubated for four to five days at 32 °C.

Molecular biology

Standard methods were used for PCR, for constructing plasmids, and for DNA sequencing. Oligonucleotide primers were designed using tools of, and were synthesized by, Integrated DNA Technologies (Coralville, Iowa). Primer sequences are available upon request. Relevant features of all plasmids are a fission yeast origin of replication and an allele of ade6 (wild-type pade6, pM375 and pM26). A subset of the plasmids contain in addition the ura4 gene (pade6-ura4, pM375-ura4 and pM26-ura4). Cells were transformed using the LiOAc procedure (Ito et al. 1983), transformants were plated on NBA lacking adenine or uracil, and plates were incubated at 32 °C for four to six days.

Results

The sup9-e (sup9-UGA) phenotype is due to a mutation of spctrnaser.11 that alters the anticodon of tRNASer.11

We obtained, from different sources, strains that harbored alleles of sup9, a well-characterized opal stop codon suppressor (Willis et al. 1984, 1986). Genetic and haploidization mapping experiments placed sup9 close to (5.8 cM away from) arg1 on Chromosome III (Kohli et al. 1977).

The first allele, designated sup9-UGA or sup9-e, was cloned by plasmid transformation and found to harbor a mutation in a gene encoding a serine tRNA located adjacent to a gene encoding a methionine tRNA (Willis et al. 1984). Our inspection of the genome sequence (Wood et al. 2002) revealed that, of the seven genes encoding a tRNASer that are located on Chromosome III, only spctrnaser.11 is located next to a tRNAMet gene (spctrnamet.07). This physical location is consistent with the reported genetic map location (Kohli et al. 1977). We therefore amplified DNA from the spctrnaser.11 locus and identified a single base pair substitution that alters the anticodon of tRNASer.11 from 5′-UGA-3′ to 5′-UCA-3′. Our sequence of the mutated tRNASer.11 gene matches that reported for the serine tRNA encoded by the plasmid clone (Willis et al. 1984), thus localizing the original mutation to the spctrnaser.11 locus. This mutation, which provides a molecular basis for recognition of UGA (opal) stop codons and for suppression, was not observed in wild-type strains.

The second strain, originally designated as sup9 (Ponticelli et al. 1988), was reported to harbor an opal nonsense suppressor. However, when we sequenced the spctrnaser.11 (sup9) locus of this strain we did not detect any mutations within the spctrnaser.11 gene. Thus, nonsense codon suppression in this second strain is not due to changes in the tRNASer.11. We therefore designated this strain temporarily as genotype sup† and sought to determine the molecular basis for its sup† phenotype.

Strains with the sup† phenotype express a mutated translation release factor, Sup35 (eRF3)

We found that sup† strains do not harbor any mutations in the spctrnaser.11 locus. However, our genetic mapping data (some of which are presented below) indicated that the sup† marker also resides in that region of chromosome III. Inspection of the genome sequence (Wood et al. 2002) revealed six candidate genes in that region, including five encoding tRNAs (tRNALys.12, tRNASer.11, tRNAMet.07, tRNAAsn.06 and tRNASer.12). Whole-genome sequencing of the sup† strain that we conducted for another project (Gao et al. 2013) revealed no mutations in any of the tRNA candidate genes.

However, whole-genome sequencing detected a mutation in the sup35 gene of strains with the sup† phenotype. The encoded protein, Sup35, is also known as eRF3 (eukaryotic translation release factor 3). PCR amplification and conventional sequencing of the locus from strains with the sup† phenotype confirmed the presence of the mutation first identified by whole genome sequencing. The mutation is a single base pair substitution (T1775C, numbered in the cDNA from the start codon) that results in a phenylalanine to serine substitution at position 592 of the protein (Fig. 1a). We therefore refer to this allele as sup35-F592S.

To determine whether suppression is due to this mutation, we set up heterozygous sup35-F592S crosses, prepared genomic DNA from 16 individual spore colonies, used PCR to amplify the sup35 locus, and sequenced the PCR products. In each case, sequences from spore colonies without suppression were wild-type at sup35 and sequences from spore colonies exhibiting suppression contained the T1775C substitution within the sup35 gene.

Together, these findings provided strong evidence that the Sup35-F592S protein is responsible for nonsense codon suppression of sup† strains. The Sup35 (eRF3) protein is the universal eukaryotic translation release factor that ensures the fidelity of termination at stop codons recognized by eRF1 (Kisselev et al. 2003; Salas-Marco and Bedwell 2004; Alkalaeva et al. 2006). The suppression phenotype exhibited by the mutant strain is therefore a logical consequence of the amino acid substitution, which affects a functional domain involved in GTP-binding and translation termination (Fig. 1, for additional details see “Discussion”).

The degree of phenotype suppression by Sup35-F592S varies for opal codons at different locations in the same gene

The mutation type-matched, codon type-matched M375 and M26 alleles of ade6 harbor UGA stop codons at amino acid positions 45 and 46, respectively (Fig. 1b). To determine the suppressibility of these alleles, we constructed strains that harbored sup35-F592S together with ade6-M375 or ade6-M26. We then plated serial dilutions of strains on rich media with limiting amounts of adenine, on minimal media supplemented with adenine, and on minimal media lacking adenine (Fig. 2). As long as adenine was present, all strains plated efficiently. As expected, strains of genotype ade6-M375 and ade6-M26 failed to plate in the absence of adenine. Interestingly, while sup35-F592S ade6-M26 strains plated efficiently in the absence of adenine (efficient suppression), the sup35-F592S ade6-M375 strains did not.

Fig. 2
figure 2

Adenine auxotrophy caused by stop codons of ade6-M375 and ade6-M26 is differentially suppressed by both sup35-F592S and sup9-UGA. Serial dilutions of cells taken from rich liquid medium were plated on the indicated solid media. The ade6 mutants produce a red pigment when adenine is limiting (YEA) and fail to grow without adenine (NBA). Efficient suppression confers both adenine prototrophy (NBA) and white colony color (YEA); whereas weak suppression confers weak prototrophy and nearly white colony color

Full size image

The differential suppression of phenotypes for ade6 alleles was also apparent using a colorimetric readout. On media with limiting adenine, the ade6-M375 and ade6-M26 mutants each accumulated a red pigment (Fig. 2) due to defects in the adenine biosynthesis pathway (Gutz et al. 1974). Cells of genotype sup35-F592S ade6-M26 formed white colonies (as do wild-type cells), indicating suppression of the biosynthetic defect caused by the ade6-M26 mutation. However, cells of genotype sup35-F592S ade6-M375 remained red (Fig. 2). We conclude that while adenine auxotrophy caused by the opal stop codon of ade6-M26 is efficiently suppressed by sup35-F592S, that of ade6-M375 is at best weakly suppressed.

Differential suppression also occurs for the sup9-UGA (tRNASer.11) suppressor

To determine whether the differential suppression is specific to Sup35-F592, we also constructed strains combining sup9-UGA with the reporter alleles ade6-M26 and ade6-M375. Similar results were obtained with this tRNA suppressor as with the Sup35 suppressor (Fig. 2). Cells of genotype sup9-UGA ade6-M26 were able to grow efficiently on media without adenine, whereas sup9-UGA ade6-M375 cells were not. However, the sup9-UGA ade6-M375 cells plated slightly more efficiently than the sup35-F592S ade6-M375 cells, indicating partial (albeit very weak) suppression of ade6-M375 by sup9-UGA. Plating of strains on media with limiting adenine, which provides a colorimetric readout of suppression, provided further support of this conclusion. While strains of genotype sup9-UGA ade6-M375 were severely hypomorphic for the biosynthesis of adenine (inefficient plating without adenine), the low level of suppression was sufficient to render an almost white colony color when plated on media with limiting adenine (Fig. 2).

Differential suppression is not due to synthetic toxicity

To further characterize the differential suppression detailed above, a different assay was used. A strain harboring a deletion of the ade6 gene and the sup35-F592S mutation was transformed with plasmids bearing wild-type ade6 (pade6), ade6-M26 (pM26) or ade6-M375 (pM375). The resulting transformants were then plated on media lacking adenine (Fig. 3a). Transformations with pade6 provided a positive control and, as expected, yielded Ade+ colonies. Transformations with plasmid-born ade6-M26 always produced Ade+ colonies, whereas transformations with plasmid-born ade6-M375 never produced Ade+ colonies. We resequenced the plasmids and confirmed that their ade6 DNA sequences contain only the M26 or M375 mutations, excluding the possibility that the differential plating efficiencies were due to other mutations in or flanking ade6. Again, sup35-F592S could efficiently suppress the phenotype of ade6-M26 but not that of ade6-M375.

Fig. 3
figure 3

Test for synthetic toxicity between sup35-F592S and ade6-M375. a Strains of genotype sup35-F592S ade6-D1 ura4-D18 were transformed with plasmids pade6 (wild-type), pM375, pM26, pade6-ura4, pM375-ura4 and pM26-ura4. Data are transformation efficiencies when plated on media lacking adenine (Ade+ colonies) or on media lacking uracil (Ura+ colonies). N.a. not applicable. b Serial dilutions of Ura+ transformants harboring pM375-ura4 and pM26-ura4 were plated on media without or with adenine. Note that while cells harboring both sup35-F592S and pM375-ura4 are viable, they plate inefficiently in the absence of adenine

Full size image

For the plating experiments with endogenous alleles (previous section) and for the plasmid transformation experiments, it remained possible that some feature of the genetic background in our sup35-F592S strains was synthetically toxic when coupled with ade6-M375. To test this possibility, we transformed sup35-F592S ade6-D1 ura4-D18 cells with plasmids bearing the different ade6 alleles together with a ura4 marker (pade6-ura4, pM26-ura4 and pM375-ura4). We then selected for transformants using media lacking uracil but containing adenine (Fig. 3a). All three plasmids were capable of producing Ura+ transformants, eliminating the possibilities that plasmids bearing ade6-M375 fail to enter cells or that the ade6-M375 allele is toxic in the presence of sup35-F592S. When the Ura+ transformants were subsequently transferred to media lacking adenine, the cells harboring plasmids with ade6-M26 plated efficiently while cells harboring plasmids with ade6-M375 plated inefficiently (Fig. 3b). This recapitulated the results obtained with ade6 alleles at the endogenous chromosomal locus (Fig. 2).

Our finding that phenotypes caused by identical stop codons at different (adjacent) positions in the same reporter gene are differentially suppressed by both sup9-UGA and sup35-F592S was intriguing, but not unprecedented (see “Discussion”). More importantly, understanding the nature of the reporter allele-specific differences allowed us to make sense of suppression phenotypes that seemed, initially, to be complex.

Nonsense codon suppression is due exclusively to the single amino acid substitution in the Sup35 protein (Sup35-F592S)

Sequence analyses of strains with and without the sup† phenotype indicated, but did not demonstrate at high resolution, that sup35-F592S is responsible for suppression (above). In addition, the differential suppressibility of reporter alleles suggested that additional loci might be involved. For example, extragenic modifiers in addition to sup35-F592S might affect the efficiencies of suppression for ade6-M26 and ade6-M375. To test this possibility and to strengthen our linkage assignments, we crossed an ade6-M26 sup35-F592S strain (efficient suppression) to an ade6-M375 strain (no suppressor). We then determined the segregation patterns of phenotypes and of genotypes at the ade6 and sup35 loci (Fig. 4).

Fig. 4
figure 4

The Sup35-F592S protein is the suppressor. a Diagram of genetic cross and summary of linkage mapping. b Tetrads were dissected and spores were plated on media with limiting adenine to provide a colorimetric readout (white, efficient suppression; red, inefficient or no suppression). Phenotypes of spores from representative parental ditype (PD), tetratype (TT) and non-parental ditype (NPD) tetrads are shown; totals of each class are below images. Genotypes inferred phenotypically were confirmed by additional phenotyping (e.g, Fig 2) and by direct analyses of alleles at the ade6 and sup35 loci. The tetrad data demonstrate inheritance of traits controlled by alleles of only two, weakly linked loci

Full size image

We dissected tetrads and plated their spores on rich media with limiting amounts of adenine (YEA), which provides a colorimetric readout for phenotypes of ade6 alleles (Gutz et al. 1974). All spores must inherit an ade6 allele (ade6-M375 or ade6-M26). Spores without suppression produce red colonies and those with suppression produce white colonies (Fig. 2). The phenotypes of 48 tetrads with four viable spores fell into three classes (Fig. 4) whose genotypes were determined by analyses of alleles at ade6 and sup35:

  • Parental ditype (PD, n = 19): 2 red (ade6-M375) and 2 white (sup35-F592S ade6-M26) colonies.

  • Tetratype (TT, n = 26): 2 red (ade6-M26 or ade6-M375), 1 small red (sup35-F592S ade6-M375), and 1 white (sup35-F592S ade6-M26) colonies.

  • Non-parental ditype (NPD, n = 3): 2 red (ade6-M26) and 2 small red (sup35-F592S ade6-M375) colonies.

These data support four important conclusions. First, the suppression and adenine phenotypes (and corresponding genotypes) are each inherited in a Mendelian fashion (2:2). Second, application of genetic mapping functions to the segregation data revealed that the sup35 locus is weakly linked (46 cM) to the ade6 locus. This is in good agreement with existing genetic and physical maps (Wood et al. 2012). Third, at the resolution of our tetrad (and all other) analyses the suppressor phenotype is inseparable from the sup35-F592S mutation, indicating that the Sup35-F592S protein is responsible for suppression. Fourth, all phenotypes were attributable unambiguously to the physical segregation of alleles at only two loci (ade6 and sup35), demonstrating that there are no additional, unlinked modifiers.

Target allele-specific effects of on adenine auxotrophy/prototrophy map exclusively to ade6

The tetrad dissection data also revealed that the differential suppressibility of adenine auxotrophy for the ade6-M375 and ade6-M26 alleles (Figs. 2, 3) is intrinsic to (inseparable genetically from) the alleles themselves (Fig. 4). While spores of genotypes sup35-F592S ade6-M375 and sup35-F592S ade6-M26 each formed colonies when plated on media containing adenine, only the sup35-F592S ade6-M26 cells continued to grow when subsequently patched to media lacking adenine. Similarly, when we plated random spore populations directly on dropout media, all 50 of the Ade+ spore colonies tested were of genotype sup35-F592S ade6-M26. This occurred even though spores of genotype sup35-F592S ade6-M375 were generated and were proficient for germination and colony formation when plated on nonselective media.

Discussion

We report that the sup9-e/sup9-UGA nonsense codon suppressor, known to involve mutation of the anticodon of a serine tRNA (Willis et al. 1984), is due to mutation of the spctrnaser.11 locus. Our findings on the nature and mechanisms of this suppressor are consistent with those described previously (Willis et al. 1984). Mutation of the tRNASer.11 anticodon from 5′-UGA-3′ to 5′-UCA-3′ allows the tRNA to recognize a UGA (opal) stop codon and add a serine to the growing polypeptide chain.

We also report that a suppression phenotype attributed previously to an allele of sup9 (Ponticelli et al. 1988) is actually due to a single amino acid substitution in the Sup35 protein, resulting from a single base pair substitution in the coding region of the sup35 gene (Fig. 1). Sup35 (eRF3) is a broadly conserved subunit of the eukaryotic translation termination complex that catalyzes the release of the nascent polypeptide chain when the ribosome encounters a stop codon (Kisselev et al. 2003; Salas-Marco and Bedwell 2004; Alkalaeva et al. 2006). Our finding that Sup35-F592S is a nonsense codon suppressor is consistent with evidence that mutations in, or changes in the dosage of, Sup35 in other species can confer suppression [e.g, (Kushnirov et al. 1990; Wakem and Sherman 1990; Gagny and Silar 1998; Chao et al. 2003; Janzen and Geballe 2004; Liu et al. 2014)]. When the functions of Sup35 are compromised, the stop codons are no longer utilized efficiently and the ribosome can insert an amino acid into the nascent polypeptide. In the case of nonsense alleles within coding regions (e.g, ade6-M375), the “wrong” amino acid is almost always inserted (Fig. 1c) and then translation continues on to the normal stop codon, at which point it can either terminate or read through. The net effect is the production of some full-length protein with amino acid substitutions corresponding to the position of the nonsense allele.

The Sup35 protein of some organisms is capable of a conformational change and self-aggregation to form prions, an infectious amyloid state [extensive body of work reviewed by (Serio and Lindquist 1999; Uptain and Lindquist 2002; Tuite and Cox 2007; Prusiner 2013)]. The Sup35 aggregates are less soluble and less functional, and thereby cause an increased rate of read-through at stop codons (increased nonsense suppression). Notably, Sup35 proteins with the prion conformation are capable of triggering other, regularly folded Sup35 molecules to adopt the amyloid state, so the prions are “infectious proteins” that exhibit the remarkable property of non-Mendelian inheritance. This raises the intriguing possibility that the fission yeast Sup35-F592S protein adopts a prion-like state that compromises its normal functions in translation termination and thereby contributes to elevated suppression of nonsense codons.

However, we think that Sup35-F592S is unlikely to form prions for three reasons. First, the amino acid substitution (Fig. 1a) is outside of the domains necessary and sufficient for prion formation by Sup35 proteins (Liu et al. 2002; Bradley and Liebman 2004; Chang et al. 2008; Bateman and Wickner 2012). Second, we did not detect any evidence for an infectious state: crossing of sup35 and sup35-F592S strains never produced any progeny of genotype sup35 that displayed elevated suppression (Figs. 2, 4). Third, it was reported recently that the ability of Sup35 proteins to form prions varies widely among species and does not occur for the fission yeast protein (Edskes et al. 2014). Given that the sup35 gene is essential (Kim et al. 2010), we infer that the Sup35-F592S protein is expressed and is hypomorphic for its functions in translation termination. In support of this idea, the amino acid substitution F592S affects a conservative residue within the GTP elongation factor Tu domain 3 (Fig. 1a).

In the course of our studies, and complicating our efforts to identify the gene responsible for the sup† phenotype, we discovered reporter allele-specific effects. Phenotypically, the opal nonsense codon of ade6-M26 is efficiently suppressed by sup35-F592S and forms a basis for positive selection (prototrophy for adenine), but that of ade6-M375 is at best weakly suppressed (Figs. 2, 3, 4). The sup35-F592S ade6-M375 genotype is so severely hypomorphic that it cannot be used for the selection of spore colonies from genetic crosses or for the maintenance of plasmids bearing ade6-M375. Genetic mapping revealed that the differential suppressibility of adenine auxotrophy is intrinsic to the ade6 alleles themselves (Fig. 4). Correspondingly, the differential suppression is neither attributable to nor unique to the Sup35 suppressor because it also occurs for a tRNA suppressor (Fig. 2).

The fact that identical stop codons at adjacent positions are differentially suppressed, and that this effect is intrinsic to the alleles themselves, revealed that the position or context in which the stop codon resides affects the efficiency of suppression. This property is not unique to fission yeast and it is known that the efficiency of translation termination is influenced by sequences surrounding stop codons [(Dalphin et al. 1997; Jacobs et al. 2009) and refs. therein]. The type and position of amino acid substitution can also affect the phenotype. The best-studied model systems, such as E. coli and S. cerevisiae, have revealed complex determinants extending at least six base pairs downstream of the stop codon itself [e.g, (Poole et al. 1998; Namy et al. 2001; Poole et al. 2003; Hatin et al. 2009)]. Our findings are consistent with such mechanisms operating in S. pombe, too.

In conclusion, we report that two different nonsense codon suppressors of fission yeast are due to mutations in the spctrnaser.11 and sup35 genes. The mechanisms of suppression involve, respectively, alteration of a serine tRNA anticodon such that it recognizes opal stop codons, and increased read-through of stop codons by attenuating the activity of a ubiquitous eukaryotic translation release factor. Secondarily, we report that local context-dependent suppression discovered in other organisms also applies to fission yeast and we provide additional evidence that fission yeast Sup35 does not form prions.