Plasmid constructions.

Plasmid pJCB101 was constructed by cloning a 0.9-kb BssHII-SacII fragment containing SUI1 from p1200 into similarly digested pCFB03. The construction of plasmids pPMB01 through pPMB08 is described in the next section.

pPMB30 was created by inserting a 1.6-kb HindIII-SacI fragment containing sui1-L96P from pPMB03 into the corresponding sites of YCplac181.

To construct pPMB09, containing sui1-E48V,L51F, the QuikChange site-directed mutagenesis system (Stratagene) was used with the primers PM-74 (5′-CCCAGAGGTATATGATTTTAAGAGAATTCTTAAGGTC-3′) and PM-75 (5′-GACCTTAAGAATTCTCTTAAAATCATATACCTCTGGG-3′), using pPMB05 as a template.

To introduce the optimal sequence context A₋₃-A₋₂-A₋₁-AUG into the corresponding sc or hc plasmids containing either SUI1 or different sui1 mutant alleles, the QuikChange site-directed mutagenesis system (Stratagene) was used with the primers PM-181 (5′-TATAGCTGAAGCAAATAAAATGTCCATTGAGAATC-3′) and PM-182 (5′-GATTCTCAATGGACATTTTATTTGCTTCAGCTATA-3′).

To construct pPMB24, an 812-bp SUI1 fragment containing 577 bp upstream and 212 bp downstream of the ATG was amplified by PCR from pJCB101 using primers PM-169 (5′- GTGAGCTACGCGTCGACAGATCTGAATCTATTCTGGAC-3′) and PM-170 (5′- CGCGGATCCTTGACAATGTTACCATTACATGC-3′) introducing a novel SalI site in the 5′ end and a BamHI site in the 3′ end of the fragment, digested with SalI and BamHI, and inserted between the corresponding sites in the GCN4 plasmid p164, generating pPMB23. A BamHI fragment containing lacZ was excised from p180 and inserted into the BamHI site of pPMB23 to produce pPMB24, harboring a SUI1-lacZ fusion construct that includes the promoter region and first 215 bp of the SUI1 open reading frame (ORF), the lacZ ORF, and GCN4 sequences extending from the BamHI site through the 3′ untranslated region to the EcoRI site 3′ of GCN4.

pPMB25 and pPMB28 were constructed by cloning the same 812-bp SalI/BamHI fragment containing SUI1 just described into pCR2.1-TOPO (Invitrogen). Site-directed mutagenesis was conducted with the QuikChange system by using primer pairs (i) PM-181 (5′-TATAGCTGAAGCAAATAAAATGTCCATTGAGAATC-3′) and PM-182 (5′-GATTCTCAATGGACATTTTATTTGCTTCAGCTATA-3′) to introduce the A₋₃-A₋₂-A₋₁-AUGcontext and (ii) PM-206 (5′-TATAGCTGAAGCAAATTTTATGTCCATTGAGAATC-3′) and PM-207 (5′-GATTCTCAATGGACATAAAATTTGCTTCAGCTATA-3′) to introduce the U₋₃-U₋₂-U₋₁-AUG context. The resulting PCR products were digested with BamHI and SalI and cloned into p164, and the corresponding SUI1-lacZ fusions, pPMB25 and pPMB28, were generated as described above for pPMB24.

To construct pPMB26 and pPMB27, an ~4.1-kb SalI-KpnI fragment containing either SUI1-lacZ or SUI1-opt-lacZ from plasmids pPMB24 and pPMB25, respectively, was cloned between the corresponding sites of pHQ1303.

Selection of mutant SUI1 alleles.

Random mutagenesis of SUI1 was conducted by error-prone PCR using the SUI1 URA3 plasmid p1200 as a template with the primers JCO075 (5′-TGTACACTATGCATGCGCGT-3′) and JCO076 (5′-CCCCATGATAATGTACTCTCG-3′) and a GeneMorph II random mutagenesis kit (Stratagene). A 0.9-kb BssHII-SacII fragment from His-SUI1 plasmid pCFB03 was replaced with the resulting PCR products and ~67,000 bacterial transformants were pooled. Plasmid DNA was prepared from the pool for screening in the appropriate yeast strains. To identify Sui⁻ alleles, Leu⁺ yeast transformants of strain JCY03 were selected on SC-L and pooled, and dilutions were plated on SC-L supplemented with 5-FOA (to evict p1200) and 0.0015 mM histidine. Transformants with His⁺ phenotype were colony purified, and the resident plasmids were isolated.

Ssu⁻ mutations were identified as suppressors of the lethality of SUI5, as follows. Leu⁺ Trp⁺ transformants of strain PMY04 were selected on SGal-LW and pooled, and dilutions were plated on SC-LW supplemented with 5-FOA. Transformants that grew in the presence of glucose were colony purified, and the resident plasmids were isolated.

For each selection, the ability of the purified plasmids to confer the relevant phenotypes was verified after transformation of the same yeast strain used to screen the mutant library, and the complete DNA sequence of the 0.9-kb BssHII-SacII fragment was determined to identify the mutations present in SUI1. In most cases, more than one mutation was present either in the promoter or in the SUI1 coding region. In identify the mutations that confer the relevant phenotypes, site-directed mutagenesis was conducted to generate SUI1 alleles with single mutations using the QuikChange system and the appropriate primers to generate the following plasmids: (i) pPMB01, primers PM-43 (5′-AAGAGAATTCTTGAGGTCCTAAAGA-3′) and PM-44 (5′-TCTTTAGGACCTCAAGAATTCTCTT-3′); (ii) pPMB02, primers PM-47 (5′-CTTAAGGTCCTAAAGGAGGACTTTGCATG-3′) and PM-48 (5′-CATGCAAAGTCCTCCTTTAGGACCTTAAG-3′); (iii) pPMB03, primers PM-138 (5′-GAATTTATGATCTCCCAACCGGGATTGCAAAAGAAG-3′) and PM-139 (5′-CTTCTTTTGCAATCCCGGTTGGGAGATCATAAATTC-3′); (iv) pPMB04, primers PM-23 (5′-CCTTTCGCCGACGCAGGAGACGACG-3′) and PM-24 (5′-CGTCGTCTCCTGCGTCGGCGAAAGG-3′); (v) pPMB05, primers PM-25 (5′-GCAAGGTGTCCCAGAGGTATATGATTTAAAGAGAA-3′) and PM-26 (5′-TTCTCTTTAAATCATATACCTCTGGGACACCTTGC-3′); (vi) pPMB06, primers PM-33 (5′-CCCAGAGGAATATGATTTTAAGAGAATTCTTAAGGTC-3′ and PM-34 (5′-GACCTTAAGAATTCTCTTAAAATCATATTCCTCTGGG-3′); (vii) pPMB07, primers PM-132 (5′-GTCCTAAAGAAGGGCTTTGCATGTAATG-3′) and PM-133(5′-CATTACATGCAAAGCCCTTCTTTAGGAC-3′); and (viii) pPMB08, primers PM-27 (5′-CAGTTGCAGGGTGACCATAGAGCAAAGGTTTGC-3′) and PM-28 (5′-GCAAACCTTTGCTCTATGGTCACCCTGCAACTG-3′).

SUI1 Ssu⁻ mutations reduce SUI1 expression by exacerbating poor context.

We also screened our mutagenized SUI1 plasmids for eIF1/Sui1 substitutions that would suppress the Sui⁻ phenotypes of mutations in other eIFs, since this phenotype was not previously described for eIF1 mutants. Ssu⁻ substitutions were of interest because they should increase the accuracy of initiation and reinstate the requirement for an AUG start codon for efficient initiation in Sui⁻ mutants. It was of interest to determine whether such Ssu⁻ substitutions would also impose a stronger requirement for optimal AUG context for efficient initiation.

Ssu⁻ alleles of SUI1 were isolated by their ability to suppress the recessive lethality of the SUI5 mutation in eIF5 (Tif5 in yeast), the GTPase activating protein for eIF2. As diagrammed in Fig. 3A, we used a sui1Δ his4-301 strain containing plasmid-borne copies of SUI1⁺ and SUI5, plus a chromosomal TIF5 allele under the GAL1 promoter (P_GAL-TIF5) that expresses WT eIF5 only on galactose medium. This strain can grow on galactose medium lacking histidine owing to the dominant Sui⁻ phenotype of SUI5 and the complementation of its recessive lethality by P_GAL1-TIF5; however, the strain cannot grow on glucose medium because P_GAL1-TIF5 is repressed, and SUI5 is the only source of eIF5 under these conditions (Fig. 3A, left). Hence, we selected plasmids from the mutant library that, following eviction of the SUI1⁺ plasmid on 5-FOA medium, rescued the ability to grow on glucose medium (suppressing SUI5 lethality) and eliminated the His⁺ phenotype on galactose medium (suppressing the SUI5 Sui⁻ phenotype) (Fig. 3A, right). We focused on five such sui1 alleles that introduce single amino acid substitutions into eIF1—T15A, E48V, L51F, D61G, and Q84H—and by site-directed mutagenesis generated a sixth allele that combines two of the mutations, E48V and L51F (abbreviated as “48,51”). The results in Fig. 3B (right) document that these sui1 alleles suppress the His⁺/Sui⁻ phenotype of SUI5, while those in Fig. 3C (bottom) demonstrate cosuppression of SUI5 lethality. In the absence of SUI5, these sui1 alleles produce slight to moderate slow-growth phenotypes and, as expected, do not confer Sui⁻ phenotypes on their own (Fig. 3D).

Open in a separate window

Fig. 3.

Isolation of Ssu⁻ substitutions in eIF1. (A) Summary of genetic selection used to isolate Ssu⁻ alleles of SUI1 as suppressors of the recessive lethality of SUI5. The relevant genotype, the expression levels of his4-301 and P_GAL-TIF5 (ON or OFF), and the growth phenotypes on medium containing galactose or glucose as carbon source are indicated for the parental strain (two rectangles on the left) and for Ssu⁻ sui1 mutants (two rectangles on the right). See the text for further details. (B) Ssu⁻ phenotypes of derivatives of a sui1Δ his4-301 P_GAL-TIF5 strain with episomal SUI5 (PMY04) harboring the indicated SUI1 alleles. Tenfold serial dilutions of these strains plus strains PMY89 and PMY91 containing episomal TIF5 (p4119) or empty vector, versus SUI5 (last two rows) were spotted on SGal+His and SGal-His (containing 0.003 mM His). (C) In the top panel, the same strains as in panel B were streaked on SC containing 2% galactose, 1% raffinose as a carbon source, lacking Leu and tryptophan (Trp), and supplemented with 0.3 mM His (SGal+His). In the bottom panel, strains were streaked on SC lacking Leu and Trp and supplemented with 5-FOA (SC + 5-FOA). (D) Phenotypes of SUI1 Ssu⁻ mutations in the absence of SUI5. Derivatives of sui1Δ his4-301 strain JCY03 containing the indicated SUI1 alleles were analyzed as in Fig. 1A.

We verified the Ssu⁻ phenotypes of these mutants by Western analysis of the myc-tagged version of the his4-301 product expressed in these strains. Expression of WT His4-myc was measured in parallel in isogenic strains harboring HIS4-myc (with AUG start codon) versus his4-301-myc (Fig. 4A and B). Quantification of the Western data revealed that, compared to the SUI1⁺ strain, all of the mutants display significantly reduced levels of his4-301-myc versus His4-myc (Fig. 4C), indicating a decreased UUG to AUG initiation ratio for HIS4 mRNA in the manner expected for Ssu⁻ alleles. Note also that the mutants resemble the SUI1⁺ strain that overexpresses WT eIF1 from an hc SUI1⁺ plasmid (hc WT), which is known to confer an Ssu⁻ phenotype (Fig. 4C) (48).

Open in a separate window

Fig. 4.

SUI1 Ssu⁻ mutations reduce the HIS4 UUG/AUG initiation ratio in SUI5 and SUI3-2 cells. (A to C) Derivatives of sui1Δ his4-301-myc SUI5 (A) and sui1Δ HIS4-myc SUI5 (B) strains JCY04 and PMY16, respectively, harboring the indicated SUI1 alleles were cultured as in Fig. 1C, and WCEs were subjected to Western analysis with antibodies against myc epitope or Gcd6. Two different amounts of each extract differing by a factor of 2 were loaded in successive lanes. (C) Western signals from panels A and B were quantified, and the mean ratios of his4-301-myc to His4-myc (each normalized to Gcd6) are plotted with the SEM as error bars. (D) Derivatives of JCY03 containing episomal SUI3-2 (p4280/YCpSUI3-S264Y-W) and harboring the indicated SUI1 alleles were analyzed for Slg⁻ and His⁺/Sui⁻ phenotypes by spotting serial 10-fold dilutions on SC lacking Leu and Trp and supplemented with either 0.3 mM His (+His) or 0.003 mM His (−His). (E) Transformants of the SUI3-2 strains from panel D containing the AUG or UUG HIS4-lacZ reporters were analyzed as in Fig. 1B.

To confirm these last findings, we examined whether the sui1 Ssu⁻ mutations also suppress the Sui⁻ phenotype of the dominant eIF2β Sui⁻ allele SUI3-2 (21). Indeed, all of the sui1 alleles resemble hc SUI1⁺ in suppressing the His⁺ phenotype of SUI3-2 (Fig. 4D) and in lowering the UUG/AUG initiation ratio for the matched HIS4-lacZ fusions in the SUI3-2 strain (Fig. 4E). The fact that they mitigate the Sui⁻ phenotype of SUI3-2 confirms that all six sui1 alleles have Ssu⁻ phenotypes.

Remarkably, all of the Ssu⁻ sui1 alleles exhibit levels of eIF1 expression lower than that of the WT, with the T15A and 48,51 mutants decreasing expression the most, by a factor of ~10 (Fig. 5A). These results suggest that the Ssu⁻ substitutions exacerbate the effect of poor context at the SUI1 start codon. Supporting this interpretation, the reduction in eIF1 expression conferred by these sui1 mutations was substantially alleviated, or completely eliminated, by the presence of the optimal AUG context in the cognate SUI1-opt alleles (Fig. 5B). For example, whereas T15A and 48,51 reduced expression of WT SUI1 by ~10-fold (Fig. 5A), they reduced expression of SUI1-opt by only ~3-fold (Fig. 5B). Note also that whereas optimal context increases eIF1 expression from SUI1⁺ by a factor of 2.6, it provokes substantially larger increases, between 4.6- and 8.4-fold, in eIF1 expression from the Ssu⁻ alleles (Fig. 5C), as expected if the latter exacerbate the effect of poor context at the SUI1 AUG codon. The fact that T15A and 48,51 produce considerably lower than WT levels of eIF1 even in the presence of the optimal context (Fig. 5B) might indicate that these substitutions destabilize eIF1 protein in addition to decreasing AUG recognition. Alternatively, another unknown feature of SUI1 mRNA might be suboptimal, and these Ssu⁻ substitutions would discriminate against this hypothetical anti-determinant in addition to poor AUG context. For example, it was shown that coding sequences function coordinately with the AUG context to stimulate the efficiency of translation and stability of PGK1 mRNA in yeast (25).

Open in a separate window

Fig. 5.

SUI1 Ssu⁻ mutations reduce SUI1 expression by exacerbating poor context. (A to C) SUI1 Ssu⁻ mutations reduce eIF1 expression dependent on SUI1⁺ AUG context. Derivatives of JCY03 containing the indicated SUI1 alleles (A) or SUI1-opt alleles (B) were subjected to Western analysis as in Fig. 1C. Two different amounts of each extract differing by a factor of 2 were loaded in successive lanes. (C) Ratios of mean eIF1/Gcd6 values for SUI1 versus SUI1-opt alleles were calculated as in Fig. 1E. (D) Derivatives of JCY03 harboring the indicated SUI1 alleles and the SUI1-lacZ or SUI1-opt-lacZ reporter were analyzed for β-galactosidase activities as in Fig. 1G. (E) SUI1-lacZ and SUI1-opt-lacZ expression levels were determined in derivatives of JCY03 containing sc SUI1, hc SUI, or hc SUI1-opt, as indicated, as in Fig. 1G. (F) Western analysis of eIF1 expression in JCY03 derivatives containing the indicated sc or hc SUI1 alleles, conducted as in Fig. 1C.

To confirm the conclusions reached from Western analysis, we also examined the effects of the Ssu⁻ mutations in trans on SUI1-lacZ expression. All of the mutations decrease expression of the WT fusion by factors of 2 to 2.5 (Fig. 5D, Rel. SUI1-lacZ) but, importantly, have considerably less effect on expression of the SUI1-opt-lacZ fusion (Fig. 5D, Rel. SUI1-opt-lacZ). Furthermore, whereas introducing the optimal context increases fusion gene expression by a factor of 2.0 in WT cells, it produces larger increases, between 2.6- and 4.8-fold, in the Ssu⁻ mutants (Fig. 5D, SUI1-opt-lacZ/SUI1-lacZ). These findings support the conclusion that eIF1 Ssu⁻ substitutions reduce SUI1 expression by exacerbating its poor AUG context.

Consistent with their deleterious effects on eIF1 expression, Northern analysis revealed that the T15A and E48V mutations also reduced the level of SUI1 mRNA. However, since the magnitude of the decrease in mRNA is less than the reduction in eIF1 protein levels evoked by these mutations, they confer protein/mRNA ratios of less than unity (Fig. 2, rows 4 and 5, Protein/mRNA), supporting the conclusion that the mutations reduce SUI1 translational efficiency. Moreover, the eIF1 protein/SUI1 mRNA ratios are reduced by the T15A and E48V mutations to a greater extent for SUI1 versus SUI1-opt alleles (Fig. 2, rows 4 and 5 [Protein/mRNA] versus rows 15 and 16 [Norm. Protein/mRNA]), supporting the idea that they reduce SUI1 translational efficiency by intensifying the negative effect of its native, poor AUG context.

eIF1 moderately autoregulates translation by exacerbating the effects of poor AUG context at SUI1.

As noted above, overexpressing WT eIF1 confers an Ssu⁻ phenotype (41, 48). Consistent with this, hc SUI1⁺ phenocopies the sui1 Ssu⁻ mutations and reduces expression of the WT SUI1-lacZ fusion by a factor of 2 but diminishes expression of SUI1-opt-lacZ by only ~20% (Fig. 5E, SUI1 versus hc SUI1). The repression of SUI1-lacZ expression occurs in response to an ~7-fold increase in eIF1 expression from hc SUI1 versus sc SUI1 (Fig. 5F). We noted that ~2-fold greater eIF1 expression (16-fold above WT) was achieved with the hc version of SUI1-opt versus hc SUI1 (Fig. 5F), which is consistent with the 2-fold greater translational efficiency of SUI1-opt versus SUI1 determined above (Fig. 1C, D, and G and Fig. 2). Consistent with this, hc SUI1-opt conferred more extensive repression of SUI1-lacZ expression than did hc SUI1, plus an increased stimulatory effect of introducing the optimal context into the lacZ fusion (Fig. 5E, cf. SUI1-opt-lacZ/SUI1-lacZ ratios). These findings suggest that eIF1 negatively autoregulates its synthesis by exacerbating the deleterious effect of the poor AUG context in SUI1 mRNA when the eIF1 level increases in the cell. On the other hand, the autoregulation is not efficient enough to prevent considerable eIF1 overexpression and the attendant Ssu⁻ phenotype in response to elevated SUI1 dosage. As shown below, this might reflect the fact that the native SUI1 context is not fully suboptimal.

Sui⁻ mutations in eIF2β and eIF1A also suppress the poor AUG context at SUI1.

The results presented above indicate that Sui⁻ substitutions in eIF1 that reduce its discrimination against UUG codons in HIS4 mRNA likewise reduce the deleterious effect of poor AUG context in SUI1 mRNA. We hypothesized that other factors involved in rejecting near-cognate start codons, including eIF2β, eIF1A, and eIF5, also function in rejecting poor AUG context. Accordingly, we sought to determine whether Sui⁻ mutations in these factors provoke increased expression of SUI1 dependent on its poor context. To examine the effect of the Sui⁻ eIF2β mutation SUI3-2, we generated strains with chromosomal SUI3⁺ under the GAL1 promoter and harboring plasmid-borne SUI3-2 or SUI3, which express only the plasmid-encoded S264Y mutant, or WT, eIF2β on glucose medium.

Remarkably, the SUI3-2 transformant displayed a strong, 8.7-fold increase in eIF1 expression from WT SUI1 but only a 2.3-fold increase in eIF1 expression from SUI1-opt (Fig. 6A). In addition, optimizing the context produced no increase in eIF1 expression in the SUI3-2 cells (Fig. 6D). Analysis of mRNA levels revealed that SUI3-2 conferred a 2.6-fold increase in the eIF1 protein/SUI1 mRNA ratio for SUI1⁺ but had almost no effect on this ratio for SUI1-opt (Fig. 2, row 8, Protein/mRNA versus row 19, Norm. Protein/mRNA), confirming a significant increase in translational efficiency exclusively for the suboptimal, native AUG context. The same conclusion emerged from analysis of the SUI1-lacZ fusions, as SUI3-2 provoked 4.6-fold higher expression of SUI1-lacZ, but only 1.9-fold higher expression of SUI1-opt-lacZ, relative to the levels observed in SUI3⁺ transformants (Fig. 6B). Moreover, introducing the optimal context produced only a slight (~20%) increase in fusion gene expression in SUI3-2 cells, compared to the 3-fold higher expression seen in SUI3 cells (Fig. 6B). These findings indicate that SUI3-2 strongly suppresses the effect of poor context at SUI1. Thus, eIF2β (presumably in the context of the eIF2 holoprotein) discriminates against poor AUG context in addition to its known function in blocking non-AUG initiation.

Open in a separate window

Fig. 6.

Sui⁻ mutations in eIF2β and eIF1A suppress the poor AUG context at SUI1. (A) Derivatives of sui1Δ P_GAL-SUI3 strain PMY02 containing plasmid-borne SUI3 (p4450) or SUI3-2 (p4280) and either sc SUI1⁺ or sc SUI1-opt were cultured continuously in SD supplemented with His and Ura (with repression of chromosomal P_GAL-SUI3) and subjected to Western analysis of eIF1 expression, as in Fig. 1C. Two different amounts of each extract differing by a factor of 2 were loaded in successive lanes. (B) Analysis of SUI1-lacZ and SUI1-opt-lacZ expression in the SUI1⁺ strains from panel A conducted as in Fig. 1G. (C) Western analysis of eIF1 in derivatives of sui1Δ P_GAL-TIF11 strain PMY03 containing plasmid-borne TIF11⁺ (pDSO9), tif11-SE1*,SE2*+F131(pAS23), or tif11-17-21 (p4552), and either sc SUI1⁺ or sc SUI1-opt, conducted as in panel A. (D) Ratios of mean eIF1/Gcd6 values for SUI1 versus SUI1-opt alleles calculated for strains analyzed in panels A and C. (E) Analysis of SUI1-lacZ and SUI1-opt-lacZ expression in transformants of the strains from the SUI1⁺ strains from panel C, as in Fig. 1G.

To evaluate whether a Sui⁻ substitution affecting eIF1A also overcomes poor context, we took the approach described above and examined transformants harboring the plasmid-borne eIF1A Sui⁻ allele tif11-SE₁*,SE₂*+F131 (abbreviated SE below), or WT TIF11, in a strain where chromosomal TIF11⁺ is glucose repressible. The SE mutation, which impairs both scanning enhancer elements (SE₁ and SE₂) in the eIF1A C-terminal tail, greatly elevates the UUG/AUG ratio for the HIS4-lacZ fusion (41). We also examined the Ssu⁻ mutation tif11-17-21, which introduces substitutions into the scanning inhibitory (SI) element in the N-terminal tail of eIF1A and lowers the UUG/AUG ratio in cells harboring SUI3-2 (14, 41).

Consistent with our findings on other Sui⁻ mutations, the eIF1A SE mutation provoked an ~8-fold increase in eIF1 expression from WT SUI1 compared to an ~3-fold increase for the SUI1-opt allele (Fig. 6C), and introducing the optimal context increased eIF1 expression by only 1.3-fold in SE mutant cells compared to 2.9-fold in TIF11⁺ cells (Fig. 6D). We observed a more substantial increase in SUI1 mRNA in the SE mutant than in other Sui⁻ mutants, so that the eIF1 protein/SUI1 mRNA ratio was only 50% higher in SE versus TIF11⁺ cells (Fig. 2, row 11, Protein/mRNA). However, the SE mutation produced a larger increase in expression of SUI1-lacZ versus SUI1-opt-lacZ, of 3.1- versus 1.9-fold, respectively, and introducing the optimal context increased SUI1-lacZ expression only slightly (by 15%) in the SE mutant compared to the 2-fold increase observed in WT cells (Fig. 6E). Taken together, the results indicate that the eIF1A SE mutation mitigates the deleterious effect of poor AUG context, as described above for Sui⁻ mutations in eIF1 and eIF2β.

The eIF1A Ssu⁻ mutation tif11-17-21 reduced the level of eIF1 expressed from SUI1 but not from sui1-opt (Fig. 6C) and increased the stimulatory effect of introducing optimal AUG context on eIF1 expression from ~3-fold (in WT) to ~8-fold (Fig. 6D). The 17-21 mutation also produced an obvious reduction in the eIF1 protein/SUI1 mRNA ratio for SUI1, but not SUI1-opt (Fig. 2, row 10, Protein/mRNA, versus row 21, Norm. Protein/mRNA), and it decreased expression of the SUI1-lacZ fusion but not that of SUI1-opt-lacZ (Fig. 6E). These results indicate that, like the Ssu⁻ mutations in eIF1 discussed above, the eIF1A Ssu⁻ mutation 17-21 confers a significant decrease in SUI1 translation that depends on its poor AUG context.

Finally, we examined the effect of the eIF5 Sui⁻ substitution (G31R) encoded by the SUI5 allele of TIF5. As noted above, SUI5 is lethal as the only source of eIF5 but produces a dominant Sui⁻ phenotype and a large increase in the UUG/AUG ratio in cells coexpressing WT eIF5 (21). Hence, we first examined the effect of SUI5 on eIF1 and SUI1-lacZ expression in a strain containing chromosomal P_GAL-TIF5, and cultured cells on galactose to allow coexpression of WT eIF5 and attendant cell growth. Under these conditions, SUI5 produced little or no change in eIF1 expression from SUI1⁺ or SUI1-opt (Fig. 7A, Vec. versus SUI5 lanes), nor did it alter expression of SUI1-lacZ or SUI1-opt-lacZ (Fig. 7B). We repeated the analysis after shifting cells to glucose medium to repress eIF5 production from the chromosomal P_GAL-TIF5 allele. Western analysis confirmed the expected reduction in total eIF5 level in cells harboring empty vector versus episomal SUI5 or TIF5⁺ (Fig. 7C, SUI1 lanes, eIF5 blot). Under these conditions, we observed a reduction in eIF1 expression in the cells harboring episomal SUI5 versus TIF5⁺ or empty vector in the SUI1⁺, but not SUI1-opt, strains (Fig. 7C), which could indicate that SUI5 differs from other Sui⁻ mutants in exacerbating rather than suppressing the effect of poor context at SUI1. However, this effect was not observed when expression of the SUI1-lacZ and SUI1-opt-lacZ fusions were assayed under the same conditions (Fig. 7D). Similar results were obtained in a strain containing chromosomal TIF5⁺ instead of P_GAL-TIF5 (data not shown). Thus, despite the fact that SUI5 elevates initiation from the near-cognate UUG start codon at HIS4 to confer a Sui⁻ phenotype, it does not overcome the effect of poor context at SUI1 and thereby elevate eIF1 protein expression.

Open in a separate window

Fig. 7.

The eIF5 Sui⁻ mutation SUI5 does not suppress poor AUG context at SUI1. (A and C). Western analysis of eIF1 expression in derivatives of sui1Δ P_GAL-TIF5 strain PMY01 containing plasmid-borne TIF5-FL (p4119), empty vector (YCplac22) or SUI5 (p4281) and either sc SUI1⁺ or sc SUI1-opt, conducted as in Fig. 1C except that strains were cultured in synthetic minimal medium with 2% galactose as carbon source and histidine and uracil supplements (SGal+HU) (A) and then shifted to SD+HU for 15 h (C). Two different amounts of each extract differing by a factor of 2 were loaded in successive lanes. (B and D) Analysis of SUI1-lacZ and SUI1-opt-lacZ expression in transformants of SUI1⁺ strains from panels A and C, cultured as described there.

We thank Tom Dever and Jon Lorsch for many helpful suggestions and comments on the manuscript.

This study was supported in part by the Intramural Research Program of the National Institutes of Health and in part by International Human Frontiers of Science Program grant R6P0028.