Leaky termination at premature stop codons antagonizes nonsense-mediated mRNA decay in S. cerevisiae

Abstract

The Nonsense-Mediated mRNA Decay (NMD) pathway mediates the rapid degradation of mRNAs that contain premature stop mutations in eukaryotic organisms. It was recently shown that mutations in three yeast genes that encode proteins involved in the NMD process, UPF1, UPF2, and UPF3, also reduce the efficiency of translation termination. In the current study, we compared the efficiency of translation termination in a upf1Delta strain and a [PSI(+)] strain using a collection of translation termination reporter constructs. The [PSI(+)] state is caused by a prion form of the polypeptide chain release factor eRF3 that limits its availability to participate in translation termination. In contrast, the mechanism by which Upf1p influences translation termination is poorly understood. The efficiency of translation termination is primarily determined by a tetranucleotide termination signal consisting of the stop codon and the first nucleotide immediately 3' of the stop codon. We found that the upf1Delta mutation, like the [PSI(+)] state, decreases the efficiency of translation termination over a broad range of tetranucleotide termination signals in a unique, context-dependent manner. These results suggest that Upf1p may associate with the termination complex prior to polypeptide chain release. We also found that the increase in readthrough observed in a [PSI(+)]/upf1Delta strain was larger than the readthrough observed in strains carrying either defect alone, indicating that the upf1Delta mutation and the [PSI(+)] state influence the termination process in distinct ways. Finally, our analysis revealed that the mRNA destabilization associated with NMD could be separated into two distinct forms that correlated with the extent the premature stop codon was suppressed. The minor component of NMD was a 25% decrease in mRNA levels observed when readthrough was >/=0.5%, while the major component was represented by a larger decrease in mRNA abundance that was observed only when readthrough was </=0.5%. This low threshold for the onset of the major component of NMD indicates that mRNA surveillance is an ongoing process that occurs throughout the lifetime of an mRNA.

FIGURE 1.

Readthrough reporter systems. (A) The β-galactosidase readthrough reporter constructs containing either the SXA or QXQ readthrough cassettes. (B) The dual luciferase readthrough reporter constructs.

FIGURE 2.

Northern blot analysis of mRNA from β-galactosidase SXA reporter constructs. Steady-state β-galactosidase mRNA levels were measured in the wild-type, [PSI⁺], upf1Δ, and [PSI⁺]/upf1Δ strains. The β-galactosidase SXA reporter constructs contained either a UAA, UAG, or UGA stop codon, or a UGG tryptophan codon in the readthrough position.

FIGURE 3.

Corrected readthrough levels determined using the β-galactosidase readthrough assay. The fold increase in readthrough in the [PSI⁺], upf1Δ, and [PSI⁺]/upf1Δ strains are represented relative to the basal level of readthrough measured in the wild-type strain at the corresponding stop codon. (A) SXA β-galactosidase reporter constructs. (B) QXQ β-galactosidase reporter constructs. WT, wild type.

FIGURE 4.

Context-dependent readthrough (QXQ/SXA) determined using the β-galactosidase readthrough assay. The data is expressed as the ratio of readthrough measured with the β-galactosidase QXQ reporter relative to the readthrough measured with the β-galactosidase SXA reporter. The QXQ/SXA ratios are shown for the wild-type, [PSI⁺], upf1Δ, and [PSI⁺]/upf1Δ yeast strains. WT, wild type.

FIGURE 5.

Corrected readthrough levels determined using the dual luciferase readthrough assay. The fold increase in readthrough measured in the [PSI⁺], upf1Δ, and [PSI⁺]/upf1Δ strains are represented relative to the basal level of readthrough measured in the wild-type strain at the corresponding stop codon.

FIGURE 6.

Effect of a premature stop codon on the translation efficiency of Renilla luciferase. Renilla specific activity (RLUs/sec/ng protein) per unit mRNA was determined for the indicated strains using dual luciferase reporter plasmids containing either the UGAC or CGAC tetranucleotides. The ratio of these values (UGAC/CGAC) serves as a measure of the relative translational efficiency. Assuming translation elongation rates are constant, this ratio should provide a measure of the relative levels of translation initiation in the UGAC and CGAC reporter plasmids.

FIGURE 7.

Correlation between mRNA levels and corrected readthrough of stop codons using the β-galactosidase readthrough assay. The corrected readthrough values obtained in both the β-galactosidase SXA and QXQ contexts were plotted against the relative level of LacZ mRNA for each construct as determined by northern blot analysis. (A) Plot of data obtained with the wild-type strain, SXA context (squares); wild-type strain, QXQ context (diamonds); [PSI⁺] strain, SXA context (circles); and [PSI⁺] strain, QXQ context (triangles). (B) Plot of data obtained with the upf1Δ strain, SXA context (squares); and the upf1Δ strain, QXQ context (triangles). WT, wild type.

FIGURE 8.

Comparison of relative mRNA levels (nonsense/sense). (A) The relative levels of LacZ mRNA in the wild-type, [PSI⁺], upf1Δ, and [PSI⁺]/upf1Δ strains expressed from the SXA reporters. (B) The relative levels of LacZ mRNA in the wild-type, [PSI⁺], upf1Δ, and [PSI⁺]/upf1Δ strains expressed from the QXQ reporters. (C) The relative levels of dual luciferase mRNA in the wild-type, [PSI⁺], upf1Δ, and [PSI⁺]/upf1Δ strains. (D) Representative Northern blot data from the wild-type and [PSI⁺]/upf1Δ strains expressing the β-galactosidase QXQ reporters. (E) Representative Northern blot data from the wild-type and [PSI⁺]/upf1Δ strains expressing the dual luciferase reporters. WT, wild type.