OUP user menu

N-terminal extension of Saccharomyces cerevisiae translation termination factor eRF3 influences the suppression efficiency of sup35 mutations

Kirill Volkov, Kirill Osipov, Igor Valouev, Sergey Inge-Vechtomov, Ludmila Mironova
DOI: http://dx.doi.org/10.1111/j.1567-1364.2006.00176.x 357-365 First published online: 1 May 2007


The eukaryotic translation termination factor eRF3 stimulates release of nascent polypeptides from the ribosome in a GTP-dependent manner. In most eukaryotes studied, eRF3 consists of an essential, conserved C-terminal domain and a nonessential, nonconserved N-terminal extension. However, in some species, this extension is required for efficient termination. Our data show that the N-terminal extension of Saccharomyces cerevisiae eRF3 also participates in regulation of termination efficiency, but acts as a negative factor, increasing nonsense suppression efficiency in sup35 mutants containing amino acid substitutions in the C-terminal domain of the protein.

  • nonsense suppression
  • translation termination
  • eRF3
  • yeast


Termination of translation in Saccharomyces cerevisiae is governed by two interacting factors, eRF1 and eRF3, encoded respectively by the SUP45 and SUP35 genes. eRF1 is responsible for recognizing stop codons and activating the peptidyl-tRNA hydrolysis catalyzed by the ribosome; eRF3 is a GTPase that stimulates eRF1 (Frolova, 1994; Stansfield, 1995b; Zhouravleva, 1995) (reviewed in Bertram, 2001; Inge-Vechtomov, 2003; Kisselev, 2003). However, many aspects of translation termination and release factor functioning remain to be established. In particular, there are several intriguing problems related to eRF3 structure and function. In most eukaryotes studied, with the exception of some protists (Inagaki & Doolittle, 2000), this protein comprises two modules: an evolutionarily conserved eEF-1A-like C-terminal part, and a nonconserved N-terminal extension. In some species, including Saccharomyces cerevisiae, this extension may be subdivided into the N and M domains, which differ in amino acid composition. In contrast to the C-terminal domain, which is essential for both termination and viability, the N-terminal extension is not essential for viability (Ter-Avanesyan, 1993). However, the fact that the eRF3 modular structure is conserved may indicate that it is functionally important.

In some species, e.g. Podospora anserina, the N-terminal extension contributes positively to termination efficiency, as its deletion causes nonsense suppression (Gagny & Silar, 1998). The N-terminus of mouse GSPT1, one of two mammalian eRF3 homologs, impairs its function, at least when this protein is used to complement mutant yeast eRF3. Deletion of the GSPT1 N-terminal extension makes such complementation possible (Le Goff, 2002). In other species, including Saccharomyces cerevisiae, termination efficiency is not changed in cells in which eRF3 lacks its N-terminal extension (Ter-Avanesyan, 1993). Nevertheless, some data indicate that the N-terminal extension participates in determining the efficiency of termination in Saccharomyces cerevisiae. In particular, termination efficiency is impaired in strains containing the cytoplasmically inherited determinant [PSI+], the aggregated form of eRF3 representing a specific class of protein-based genetic determinants, the so-called yeast prions (for review, see Liebman & Derkatch, 1999; Wickner, 1999; Chien, 2004). The prion properties of eRF3 are determined by its N-terminal extension, first of all by the N domain (Doel, 1994; Ter-Avanesyan, 1994; Derkatch, 1996; DePace, 1998; Liu & Lindquist, 1999), although the M domain also plays a role in [PSI+] propagation (Liu, 2002; Bradley & Liebman 2004). As aggregation partially inactivates eRF3, [PSI+] cells manifest a nonsense suppressor phenotype.

Other recently-obtained data show that regulation of termination efficiency by the N-terminal extension may be related to reasons other than the prionization of eRF3; for example, it may occur in strains with a decreased level of eRF1. The termination efficiency in such strains may be significantly improved when the full-length eRF3 is replaced with eRF3C (Valouev, 2002). Indirectly, this observation supports data from crystal structure analysis of Schizosaccharomyces pombe eRF3 (Kong, 2004). This work showed that the N-terminus of eRF3 may bind to the region at its C-terminal domain, where the eRF1 binding site is located. This intramolecular interaction may block eRF1 binding under certain physiologic conditions, and if similar competition takes place in Saccharomyces cerevisiae, elimination of the N-terminal extension may compensate for eRF1 depletion.

We have also observed that the eRF3 N-terminus may inhibit translation termination under special conditions (Volkov, 2002). This effect was revealed in two sup35 mutants, sup35-10 and sup35-25, containing amino acid substitutions in the C-terminal domain of eRF3. Deletion of the N-terminal extension decreased the suppressor effect of these mutations. This was the first indication that the N-terminal extension of eRF3 contributes to the suppression manifested by sup35 mutations. In this article, we describe a further examination of this effect.

Materials and methods


The Saccharomyces cerevisiae strains are listed in Table 1. The full name of 2V-P3982 is du8-132-L28-2V-P3982. ade1-14 (UGA), his7-1 (UAA) and lys2-87 (UGA) are nonsense mutations used as indicators of suppression efficiency. The sup35 mutants 10-2V-P3982, 25-2V-P3982, 110-2V-P3982 and 112-2V-P3982 have been obtained and characterized previously (Volkov, 2002). The SUP35 form encoding eRF3 lacking its N-terminal extension is designated SUP35C. The construction of the SUP35C allele and the substitution of chromosomal SUP35 with SUP35C or its mutant derivatives have been described previously (Volkov, 2002). The chromosomal SUP35 in strains pRSU1-16A-D1608, pRSU1C-16A-D1608 and pRSU2-16A-D1608 is disrupted by a TRP1 insertion; the corresponding plasmids are pRSU1, pRSU1C and pRSU2 (see below). The strains initially used for isolating the sup35 mutants were the [psi] strains 10-2V-P3982 and 25-2V-P3982, lacking the prion-like antisuppressor determinant [ISP+] (Volkov, 2002).

View this table:
Table 1

Genotypes of yeast strains

2V-P3982MATαade1-14 his7-1 lys2-87 thr4-B15 ura3Δleu2-1
10-2V-P3982MATαade1-14 his7-1 lys2-87 thr4-B15 ura3Δleu2-1sup35-10
25-2V-P3982MATαade1-14 his7-1 lys2-87 thr4-B15 ura3Δleu2-1sup35-25
110-2V-P3982MATαade1-14 his7-1 lys2-87 thr4-B15 ura3Δleu2-1sup35-110
112-2V-P3982MATαade1-14 his7-1 lys2-87 thr4-B15 ura3Δleu2-1sup35-112
SUP35C-2V-P3982MATα ade1-14 his7-1 lys2-87 thr4-B15 ura3Δleu2-1 SUP35C
10C-2V-P3982MATαade1-14 his7-1 lys2-87 thr4-B15 ura3Δleu2-1sup35-10C
25C-2V-P3982MATαade1-14 his7-1 lys2-87 thr4-B15 ura3Δleu2-1sup35-25C
pRSU1-16A-D1608MATαade1-14 his7-1 lys2-87 met13-A1 ura3Δthr4-B15 leu2-3,112 trp1-289 cyh2-1 SUP35::TRP1 (pRSU1)
pRSU1C-16A-D1608MATαade1-14 his7-1 lys2-87 met13-A1 ura3Δthr4-B15 leu2-3,112 trp1-289 cyh2-1 SUP35::TRP1 (pRSU1C)
pRSU2-16A-D1608MATαade1-14 his7-1 lys2-87 met13-A1 ura3Δthr4-B15 leu2-3,112 trp1-289 cyh2-1 SUP35::TRP1 (pRSU2C)
2V-P2156MATa ade1-14 his7-1met13-A1
14V-D1622MATa ade1-14 his7-1 lys2-87 ura3Δthr4-B15 leu2-1 trp5


The centromeric plasmids pRSU1 and pRSU2 carry the SUP35 gene and contain LEU2 and URA3 as selective markers, respectively; pRSU1C is the derivative of pRSU1 containing SUP35C. The construction of these plasmids has been described previously (Volkov, 2002). To convert the mutant sup35C alleles obtained in the pRSU1C plasmid into full-length alleles, the StuI–XbaI fragments isolated from the mutant pRSU1C were used for substitution of the same region in pRSU1. The substitution was confirmed by sequencing the corresponding region. pYex4 is a multicopy URA3-based plasmid; its derivative pYex4-SUP45 contains the SUP45 gene. Both plasmids were kindly gifted by V. Kushnirov.

Cultivation procedures and genetic manipulations

Standard yeast media and techniques were used (Sherman, 1986). Strains were grown at 25°C. sup35 mutations were obtained as spontaneous suppressors of the ade1-14 and his7-1 mutations on synthetic medium lacking adenine and histidine. Allelism of suppressor mutations was established by crossing the mutants with the sup35 and sup45 derivatives of strains 2V-P2156 (for mutants obtained from 2V-P3982) or 14V-D1622 (for mutants obtained from 16A-D1608). YPD containing 5 mM guanidine hydrochloride (Sigma) was used to cure the yeast strains of the prion-like determinant [ISP+] (GuHCl test) (Tuite, 1981). Treatment and subsequent examination of the clones obtained were performed as described previously (Derkatch, 1997). SC medium containing 1 mg mL−1 5-fluoro-orotic acid (Boeke, 1984), purchased from Angus, was used to eliminate the URA3-carrying plasmids.

LacZ read-through assay

To quantify suppression efficiency, the sup35 and sup35C mutants were transformed with the centromeric nonsense codon read-through assay plasmids pUKC815, pUKC817 and pUKC818/819 (Stansfield, 1995a). Six independent transformants of every mutant with each of the four plasmids were used for evaluation of β-galactosidase activity (Miller, 1972). The read-through level for each of the three stop codons was calculated as the percentage of the β-galactosidase activity in cells transformed with pUKC817 (UAA), pUKC818 (UAG) and pUKC819 (UGA) in comparison with that in cells transformed with the control plasmid pUKC815. The differences in read-through between pairs of sup35 and sup35C mutants were evaluated by the Wilcoxon rank sum test.

SUP35 sequencing


Western blot analysis

Yeast lysates were prepared as follows. Aliquots of overnight yeast cultures were washed with buffer (25 mM Tris-HCl, pH 7.4, 100 mM NaCl, 5 mM MgCl2, 1 mM dithiothreitol) containing 1 mM phenylmethylsulfonyl fluoride to limit proteolytic degradation, and lysed in the same buffer by vortexing with glass beads. Cell debris was removed by centrifugation at 15 000 g for 5 min. Lysate samples were separated on a 10% polyacrylamide gel containing 0.1% sodium dodecyl sulfate (Laemmli, 1970) and transferred to nitrocellulose sheets (Towbin, 1979). The protein concentration was determined by the microbiuret method (Herbert, 1971). The Western blots were probed at room temperature for 1–4 h, depending on antibody titer, with polyclonal affinity-purified rabbit antibodies against eRF3, eRF3C and eRF1 in TBS buffer (30 mM Tris-HCl, pH 7.5, 150 mM NaCl) containing 0.4% casein (Sigma) and 0.02% Tween-20. For visualization, species-specific horseradish peroxidase-linked antibodies (Promega) were used. Bound antibodies were detected using the Amersham ECL system. The protein expression level was evaluated by probing serially diluted lysates with the corresponding antibody.


sup35 and sup35C mutants differ phenotypically

The work began with the isolation of omnipotent suppressor mutations in two strains, 2V-P3982 and SUP35C-2V-P3982, as described in ‘Materials and methods’. These strains differ only in the SUP35 allele (Table 1). 2V-P3982 contains a full-length SUP35, whereas SUP35C-2V-P3982 contains a truncated version of this gene, encoding an eRF3 form that lacks its N-terminal extension, designated hereafter as eRF3C. After the sup35 mutants had been identified by crossing with tester strains (see ‘Materials and methods’), the phenotypes of the mutants obtained in both strains were compared. Although the selection method was the same, the two groups of mutants differed with respect to nonsense suppression. The sup35 mutants obtained in 2V-P3982 efficiently suppressed not only ade1-14 and his7-1, used to select the suppressor mutations, but also lys2-87. However, only ade1-14 was suppressed efficiently in any of the sup35C mutants obtained in SUP35C-2V-P3982; his7-1 was suppressed weakly, and lys2-87 not at all (Table 2). These results are in agreement with our previous findings regarding the sup35-10 and sup35-25 mutations (Volkov, 2002). Mutants obtained in SUP35C-2V-P3982 displayed the same suppression pattern as strains carrying a 5′-terminally truncated version of the sup35-10 and sup35-25 alleles.

View this table:
Table 2

Nonsense suppression in strains containing sup35 and sup35C mutations

sup35 alleleNonsense mutations
ade1-14 (UGA)his7-1 (UAA)lys2-87 (UGA)
  • sup35-501–sup35-518, mutants obtained in 2V-P3982; sup35C-301-sup35C-378, mutants obtained in SUP35C-2V-P3982;+, rapid growth on medium lacking adenine, histidine or lysine;+/−, weak growth; −/+, impaired growth;−, no growth.

Additional sets of sup35 and sup35C mutants were obtained in strains pRSU1-16A-D1608 and pRSU1C-16A-D1608. As the chromosomal SUP35 in these strains is disrupted by a TRP1 insertion, suppressor mutations affected the plasmid copy of the gene. All eight mutants containing N-terminally truncated SUP35 showed the same suppression pattern as the mutants containing chromosomal sup35C mutations – efficient suppression of ade1-14, weak suppression of his7-1 and no lys2-87 suppression – whereas 20 mutants obtained in pRSU1-16A-D1608 demonstrated the same phenotype as mutants obtained in 2V-P3982 (not shown).

As we showed previously, some sup35 mutants may contain a prion-like determinant that decreases the efficiency of suppression (Volkov, 2002). Treating this determinant with the universal antiprion agent 5 mM GuHCl restores suppression efficiency. Thirteen chromosomal sup35C mutants and eight plasmid sup35C mutants were examined for this determinant by the GuHCl test (see ‘Materials and methods’). None of them showed a phenotypic change (not shown).

Plasmid sup35C mutants were also examined for secondary chromosomal mutations that might decrease suppression efficiency. To this end, plasmids bearing mutant sup35C alleles were isolated and used to transform strain pRSU2-16A-D1608 containing wild-type SUP35 in a URA3-based vector. After the pRSU2 plasmid was eliminated on 5-fluoro-orotic acid-containing medium, the transformants showed no greater efficiency of suppression than was observed initially for every mutant (not shown). Thus, weak suppression is an intrinsic property of sup35C mutations, unrelated to secondary chromosomal mutations or nonchromosomal determinants that decrease the efficiency of suppression.

Sequencing of mutant sup35C alleles

Four mutant sup35 alleles obtained in strain 2V-P3982 have been sequenced previously (Volkov, 2002). They carry D363N (sup35-10), T378I (sup35-25), V413L (sup35-112) and P575H (sup35-110) amino acid replacements in eRF3. Here, four chromosomal (301–378) and three plasmid (826–845) sup35C mutants were sequenced (Table 3).

View this table:
Table 3

Sequencing of sup35C alleles

Mutant alleleNucleotide substitutionAmino acid substitution
sup35C-845C1235T; C1236TThr412Ile

Two of the chromosomal mutations (307 and 316) were shown to carry identical nucleotide substitutions. Of the six mutated sites revealed in these mutants, five (340, 367, 386, 412 and 423) are close to the positions previously identified in the sup35-10, sup35-25 and sup35-112 mutants (see above). The sixth substitution, Pro575His, is identical to that revealed in sup35-110; nevertheless, strains expressing sup35-110 and sup35C-301 alleles differ drastically in phenotype (Fig. 1).

Figure 1

Growth of strains expressing SUP35 and SUP35C alleles and their mutant derivatives sup35-110 and sup35C-301 on SC-Lys medium and SC as a control.

Adding the N-terminal extension to the mutant eRF3C increases the suppression efficiency of sup35 mutations

Further evidence for the contribution of the N-terminal extension to nonsense suppression was obtained by converting sup35C mutant alleles into sup35 alleles. To this end, the sequenced sup35C alleles, sup35C-826, sup35C-827 and sup35C-845, located in the pRSU1C plasmid, were used. As all three mutant sites were situated within the StuI–XbaI fragments of SUP35, these fragments were isolated from the mutant sup35C alleles and used to substitute the same region of the full-length SUP35 cloned in pRSU1 (see ‘Materials and methods’). As a result, the ‘extended’ mutant alleles designated sup35C-826-N, sup35C-827-N and sup35C-845-N were obtained. Plasmids carrying ‘extended’ alleles and the original sup35C mutant alleles were used to transform strain pRSU2-16A-D1608. After elimination of pRSU2 containing the wild-type SUP35, phenotypes of transformants were examined. Four transformants were selected for phenotypic analysis in each case. The transformants containing ‘extended’ alleles displayed an efficient suppression pattern, similar to that observed for sup35 mutants. The most obvious difference between strains expressing sup35C-N alleles and those expressing sup35C alleles was observed again for lys2-87 suppression (Fig. 2).

Figure 2

Growth of strains expressing the sup35C alleles (top line) and the same alleles in the ‘extended’ form (bottom line) on SC-Lys medium (left) and on SC as a control (right).

Quantification of read-through efficiency of stop codons in sup35C, sup35 and sup35C-N mutants

To verify that the phenotypic differences between the sup35 and sup35C mutants are related to different levels of stop codon read-through, the nonsense codon read-through assay was applied (Stansfield, 1995a). We used three pairs of sup35 and sup35C alleles (sup35-10 and sup35-10C, sup35-25 and sup35-25C, and sup35-110 and sup35C-301), and two pairs of sup35C and sup35C-N alleles (sup35C-826 and sup35C-826-N, and sup35C-827 and sup35C-827-N), containing identical mutations.

Mutants containing these alleles were transformed by a set of pUKC read-through assay plasmids (see ‘Materials and methods’). The set included the control vector pUKC815 and three read-through vectors, pUKC817–819, which carried stop codons TAA, TAG and TGA fused in-frame with the LacZ coding sequence. The read-through efficiency manifested by transformants was calculated by evaluating β-galactosidase activity as described in ‘Materials and methods’, and expressed as a percentage of control β-galactosidase level. In each pair of mutants, the read-through efficiency of UAA, UAG and UGA was significantly higher in strains containing the full-length sup35 allele than in strains containing the sup35C alleles (Table 4).

View this table:
Table 4

Efficiency of read-through of stop codons in the full-length and sup35C mutants

SUP35 allelePlasmid
pUKC817 (UAA)pUKC818 (UAG)pUKC819 (UGA)
sup35-11010.24 ± 2.1210.95 ± 1.3512.63 ± 2.26
sup35C-3016.35 ± 1.377.31 ± 1.874.11 ± 1.21
sup35-1010.76 ± 2.8112.23 ± 3.5214.09 ± 3.36
sup35-10C5.43 ± 1.135.88 ± 0.726.9 ± 1.46
sup35-2512.62 ± 3.8517.19 ± 4.0215.29 ± 3.57
sup35-25C2.97 ± 0.763.64 ± 0.615.34 ± 1.17
sup35C-826-N10.71 ± 3.0617.29 ± 4.0312.74 ± 2.97
sup35C-8264.69 ± 0.873.25 ± 0.607.21 ± 1.54
sup35C-827-N16.65 ± 4.0911.91 ± 3.3713.81 ± 3.89
sup35C-8275.00 ± 1.016.10 ± 0.856.67 ± 1.54

Differences in suppression efficiency between sup35 and sup35C mutations do not correlate with differences in the amount of eRF3

The differences in suppression efficiency between mutants containing the full-length sup35 alleles and sup35C alleles may be related to different amounts of eRF3. The eRF3 level could be changed, for example, if the mutant proteins containing the N-terminal extension were less stable, or if an unidentified regulatory effect increased the amount of eRF3C and, as a result, decreased the efficiency of suppression in sup35C mutations.

The amount of eRF3 was evaluated by Western blotting (see ‘Materials and methods’). In mutants expressing the sup35-10 and sup35-25 alleles, the eRF3 level was approximately the same as in the corresponding wild-type strain (Fig. 3a, top). Similarly, strains carrying sup35C-N alleles did not differ significantly by eRF3 level from wild-type strains (Fig. 3b, top). Thus, the high suppression efficiency of these mutations is not related to a decreased amount of eRF3.

Figure 3

Amount of eRF3 in strains expressing the full-length and the 5′-terminally truncated SUP35 alleles. (a) Mutants obtained in 2V-P3982. (b) Mutants obtained in pRSU1C-16A-D1608. The ratio of the amount of eRF3 in mutant strains to that in the corresponding wild-type strains (taken as 1.0) is indicated.

The eRF3 level in strains expressing sup35-10C and, especially, sup35-25C was higher than in the initial strain SUP35C-2V-P3982 expressing nonmutant SUP35C (Fig. 3a, bottom). However, among plasmid sup35C mutants obtained in pRSU1C-16A-D1608, only the sup35C-845 mutant contained more eRF3 than the nonmutant strain; in two other sup35C mutants, the eRF3 level was decreased (Fig. 3b, bottom). Nevertheless, as we have shown, these two mutants displayed a low level of read-through efficiency, comparable with that displayed by sup35-10C and sup35-25C mutants (Table 3).

Overexpression of SUP45 does not influence the suppression efficiency of sup35 and sup35C mutants

In Schizosaccharomyces pombe, the negative regulatory effect of the eRF3 N-terminal extension on translation termination is probably mediated by competition with eRF1 for the same binding site in the eRF3 C-terminal domain (Kong, 2004). If similar competition takes place in Saccharomyces cerevisiae, overproduction of eRF1 should cause a decrease of suppressor efficiency in sup35 mutants. To examine this possibility, three strains expressing the full-length alleles, sup35-25, sup35C-826-N and sup35C-827-N, were transformed with pYex4-SUP45 or with pYex4 as a control. The amount of eRF1 in the transformants was checked by Western blotting. In the pYex4-SUP45 transformants, the amount of eRF1 was approximately four-fold higher than in the pYex4 transformants (Fig. 4).

Figure 4

Increase in eRF1 level following SUP45 expression from the multicopy plasmid pYex4-SUP45 as determined by Western blot analysis. The control strain transformed with pYex4 expresses chromosomal SUP45 only. Serially diluted lysates with eRF1 antibodies are presented.

Six transformants of every mutant were examined for their ability to grow on C-Ade, C-His and C-Lys (no medium contained uracil, to prevent plasmid loss). No phenotypic differences were observed between the transformants obtained using pYex4-SUP45 and pYex4 (not shown). Thus, excess eRF1 did not counteract the negative effect of the eRF3 N-terminal extension on termination efficiency observed in sup35 mutants.


We have studied the dependence of suppression efficiency of sup35 mutations, which cause amino acid substitutions in the C-terminal domain of eRF3, on the N-terminal extension of the protein. The possibility of such dependence has been ascertained previously (Volkov, 2002), using 5′-deleted forms of two mutant sup35 alleles that produce eRF3 lacking its N-terminal extension. Strains expressing the truncated alleles demonstrated a decreased level of suppression compared with the original strains, which express the full-length mutant alleles. The further studies described in the present article showed that this effect was consistent and that sup35C mutants show decreased suppression levels in comparison with sup35 mutants, even if they carry identical amino acid substitutions in the eRF3 C-terminal domain. We also showed that modification of mutant sup35C alleles to make full-length alleles increases the suppression efficiency. In total, dependence on the eRF3 N-terminus was demonstrated for six pairs of sup35 and sup35C mutations, obtained either by truncation of sup35 alleles or by extension of sup35C alleles.

It is important that neither the decreased nor the increased levels of suppression manifested by different forms of mutant sup35 alleles depend on the prion-like suppressor or antisuppressor determinants, as GuHCl treatment did not change the phenotypes of the mutants studied. Moreover, the effect observed is not related to differences in the amount of mutant protein between cells expressing different alleles of SUP35. The amount of the full-length mutant protein is not markedly different from that in the wild type; the amount of mutant eRF3C may be either higher or lower than that of the nonmutant eRF3C (Fig. 3). Thus, differences in suppression efficiency between the mutations studied depend on the structure of eRF3 but not on its abundance. Quantification of this effect showed that the contribution of the N-terminal extension to read-through efficiency is comparable with the efficiency manifested by the mutations per se, or even exceeds it several-fold (Table 3). Interestingly, it has recently been reported that the frequency of sup35 mutations in the strain containing the SUP35C allele is lower than in the isogenic strain containing the normal SUP35 gene (Borchsenius, 2005). This can be explained by the low suppression efficiency of sup35C mutations, which makes them difficult to detect.

Our data indicate that the contribution of the N-terminal extension to suppression efficiency is probably not mediated by competition with eRF1 for the binding site in the C-terminal domain, as described for Schizosaccharomyces pombe (Kong, 2004). This conclusion is based on the finding that four-fold overproduction of eRF1 did not decrease the suppression efficiency of the sup35 mutations (Fig. 4). Therefore, we propose that either there is no intramolecular interaction preventing eRF1 binding to eRF3 in Saccharomyces cerevisiae, or its contribution to the efficiency of suppression is not significant.

If this is so, then how can the observed effect be explained? It is known that in the wild-type protein the N-terminal extension has neutral function; therefore, its deletion does not affect termination efficiency in Saccharomyces cerevisiae (Ter-Avanesyan, 1993). Why does its role change when amino acid substitutions are present in the eRF3 C-terminal domain? Unfortunately, the exact role of eRF3 in translation termination is still elusive, so it is difficult to answer this question at present. Of the three hypotheses proposed for the function of eRF3, i.e. eRF1 delivery to the ribosomal A site (Stansfield, 1995b), stimulation of eRF1 recycling (Zavialov, 2001), or direct involvement in termination through enhancement of the eRF1-decoding properties (Salas-Marco & Bedwell, 2004), only the latter two have received experimental support. Strikingly, eight of the nine mutant positions identified in this study are located in the GTPase fold of eRF3 (Fig. 5a). Although the ninth position (Pro575His) is located outside this region, it is close to the GTPase fold in the tertiary structure model of eRF3C (Fig. 5b), and may influence its function through intramolecular interactions. Therefore, we propose that the sup35 mutations selected as suppressors of the ade1-14 and his7-1 nonsense mutations most likely affect the GTPase function of eRF3.

Figure 5

Location of amino acid substitutions caused by the mutations studied. (a) Schematic map of eRF3 showing domain boundaries and positions of the G1–G5 GTP-binding sites; the sup35 and sup35C mutations whose effects were shown to depend on the N-terminal extension are indicated by bold type. (b) eRF3C tertiary structure generated on the basis of the tertiary structure of the Schizosaccharomyces pombe eRF3 C-terminal domain with the swiss-prot program (http://tw.expasy.org/sprot/) and visualized with deepview/swiss-pdbviewer software (http://www.expasy.org/spdbv/).

Our results show that the N-terminal extension of eRF3 acts as an additional factor increasing the nonsense suppression efficiency of sup35 mutations. The N-terminal extension of eRF3 might further decrease the GTPase activity of this protein in strains that contain amino acid substitutions in its C-terminal domain. As a result, the termination efficiency in sup35 mutants is lower than in sup35C mutants; consequently, the read-through efficiency of stop codons is increased. Although the mechanisms underlying the intramolecular interaction revealed in this study remain to be established, the data obtained indicate that the N-terminal extension of the eRF3 translation termination factor in Saccharomyces cerevisiae is specifically involved in the regulation of translation fidelity.


We thank M.D. Ter-Avanesyan for helpful discussions. This work was supported by grants from the Russian Foundation for Basic Research and by the BRHE program executed jointly by the Civilian Research and Development Foundation (USA) and the Russian Federation Ministry of Education and Science.


  • Editor: Monique Bolotin-Fukuhara


View Abstract