Internal initiation in Saccharomyces cerevisiae mediated by an initiator tRNA/eIF2-independent internal ribosome entry site element

Abstract

Internal initiation of translation can be mediated by specific internal ribosome entry site (IRES) elements that are located in certain mammalian and viral mRNA molecules. Thus far, these mammalian cellular and viral IRES elements have not been shown to function in the yeast Saccharomyces cerevisiae. We report here that a recently discovered IRES located in the genome of cricket paralysis virus can direct the efficient translation of a second URA3 cistron in dicistronic mRNAs in S. cerevisiae, thereby conferring uracil-independent growth. Curiously, the IRES functions poorly in wild-type yeast but functions efficiently either in the presence of constitutive expression of the eIF2 kinase GCN2 or in cells that have two initiator tRNA(met) genes disrupted. Both of these conditions have been shown to lower the amounts of ternary eIF2-GTP/initiator tRNA(met) complexes. Furthermore, tRNA(met)-independent initiation was also observed in translation-competent extracts prepared from S. cerevisiae in the presence of edeine, a compound that has been shown to interfere with start codon recognition by ribosomal subunits carrying ternary complexes. Therefore, the cricket paralysis virus IRES is likely to recruit ribosomes by internal initiation in S. cerevisiae in the absence of eIF2 and initiator tRNA(met), by the same mechanism of factor-independent ribosome recruitment used in mammalian cells. These findings will allow the use of yeast genetics to determine the mechanism of internal ribosome entry.

Figure 1

A growth assay for the cricket IGR-IRES function in vivo. (A) A diagram of the dicistronic reporter construct that contains a CrPV IGR-IRES element preceding the second cistron. The expression of the reporter mRNA is directed by the copper promoter (Cup1). (B) Growth assay for yeast strains H1402 (isogenic wild type), H1692 (GCN2^c), and H1613 (GCN2^c) transformed with either pCup1 LEU2 IGR URA3 (IGR) or pCup1 LEU2 IGRmut14 URA3 (mut14) plasmids containing the dicistronic reporter. Transformants were streaked on minimal SD medium supplemented with 100 μM Cu²⁺, inositol, with or without uracil, as indicated.

Figure 2

Northern blot analysis of total RNA isolated from yeast strains transformed with various dicistronic reporter plasmids. (A) Yeast strains H1402, H1692, or H1613 were transformed with the plasmids described in Fig. 1 or with the pSal I (pSal) parental vector and grown in SD medium supplemented with (+) or without (−) 100 μM Cu²⁺, inositol, and uracil. Total RNA was isolated, separated on denaturing gels, transferred to nitrocellulose, and hybridized with a probe complementary to either the C-terminal FLAG epitope or the N-terminal coding region of URA3. The slight differences observed in mobility are also observed by ethidium bromide staining of the ribosomal RNAs and were not consistently observed, so it is unlikely that they reflect a true difference in mobility. (B) Yeast strains H1692 and H1613 transformed with the pCup1 LEU2 IGR URA3 plasmid were grown in SD medium supplemented with 100 μM Cu²⁺ and inositol. Northern analysis was performed as above by using the probe complementary to the N-terminal coding region of URA3.

Figure 3

Immunoblot analysis of CrPV IGR-dependent translation of tagged Ura3p and levels of eIF2α phosphorylation. (A) The transformed yeast strains are described in the legend to Fig. 1. An immunoblot was prepared and incubated with the anti-FLAG M2 monoclonal antibody to detect expression of Ura3p with a C-terminal FLAG tag. An image of the developed blot is shown. (B) Immunoblots from yeast strains H1402, H1692, and H1613 were prepared and incubated, as indicated, with anti-eIF2α polyclonal antibody, which detects only the phosphorylated form of eIF2α (eIF2α-P), with CM-217 polyclonal antibody, which detects both phosphorylated and nonphosphorylated forms of eIF2α (eIF2α) or without primary antibody (none). Images of the developed blots are shown. The 36-kDa eIF2α subunit migrates slightly above the 35-kDa marker protein.

Figure 4

Growth assays for strains H1402 (wild-type), H2545 (imt3 imt4), and H2546 (imt3 imt4 Δgcn2) transformed with the plasmids described in Fig.1B. The figure shows the growth of serially diluted yeast cells on plates containing minimal SD medium supplemented with 100 μM Cu²⁺, inositol, and with or without uracil, as indicated.

Figure 5

IGR-IRES activity in yeast extracts. (A) Diagram of dicistronic mRNAs containing the Renilla luciferase (R-Luc) as the first cistron and the firefly luciferase (F-Luc) as the second cistron. (B) Genotype of the wild-type (IGR) and various mutant IGR-IRES elements. Nucleotide sequences that differ from the wild-type IGR are shown in bold and underlined. (C) Translation-competent yeast extracts were programmed with various dicistronic RNAs, and the ratios of F-Luc to R-Luc production are indicated. (D) The primary data are shown. Δ represents the dicistronic mRNA without any IGR-IRES sequences inserted. Standard error of three experiments is indicated.

Figure 6

The in vitro translation activity of the wild-type CrPV IGR-IRES in the presence of edeine. Translation-competent yeast extract was preincubated for 5 min with the indicated concentrations of edeine and programmed with the IGR dicistronic mRNA diagramed in Fig. 5. Translation efficiencies of F-Luc (shaded) and R-Luc (unshaded) are shown. Error bars represent the standard error of three experiments.