Suppression of ribosomal reinitiation at upstream open reading frames in amino acid-starved cells forms the basis for GCN4 translational control

Regulation of translation initiation and modulation of cellular physiology

Translational control of gene expression is an important regulatory mechanism in cellular physiology. In eukaryotes, ribosomes can initiate translation by two different mechanisms: a majority of mRNAs undergo cap-dependent initiation at their extreme 5'-ends, but initiation can occur internally in some mRNAs. A number of important cellular responses, such as entry into a proliferative state and adaptation to changing nutrient levels, are mediated by changes in the mechanism of translation initiation of specific mRNAs. This article discusses new insights into control of gene expression gained through studies of regulation of eukaryotic translational initiation.

Functional expression of the transcriptional activator Opaque-2 of Zea mays in transformed yeast

The aim of this research was to determine whether the structural homology between the O2 gene, a maize transcriptional activator, and the GCN4 gene, a yeast transcriptional factor, is reflected at the level of function. The O2 cDNA was cloned in the yeast expression vector pEMBLyex4 under the control of a hybrid inducible promoter, and used to transform the yeast Saccharomyces cerevisiae. Transformed yeast cells produced O2 mRNA and a polypeptide immunoreactive with anti-O2 antibodies during growth in galactose. The heterologous protein was correctly translocated into the yeast nuclei, as demonstrated by immunofluorescence, indicating that the nuclear targeting sequences of maize are recognized by yeast cells. Further experiments demonstrated the ability of O2 to rescue a gcn4 mutant grown in the presence of aminotriazole, an inhibitor of the HIS3 gene product, suggesting that O2 activates the HIS3 gene, gene normally under control of GCN4. It was shown that the O2 protein is able to trans-activate the HIS4 promoter in yeast cells and binds to it in vitro. The sequence protected by O2, TGACTC, is also the binding site for GCN4. Finally, the expression of O2 protein in yeast did not produce alterations during batch growth at 30 degrees C, while transformants expressing O2 protein showed a conditionally lethal phenotype when grown in galactose at 36 degrees C; this phenotype mimics the behaviour of gcd mutants. The results support the idea that basic mechanisms of transcription control have been highly conserved in eukaryotes.

Gene-specific translational control of the yeast GCN4 gene by phosphorylation of eukaryotic initiation factor 2

Phosphorylation of the alpha subunit of eukaryotic initiation factor 2 (eIF-2 alpha) is one of the best-characterized mechanisms for down-regulating total protein synthesis in mammalian cells in response to various stress conditions. Recent work indicates that regulation of the GCN4 gene of Saccharomyces cerevisiae by amino acid availability represents a gene-specific case of translational control by phosphorylation of eIF-2 alpha. Four short open reading frames in the leader of GCN4 mRNA (uORFs) restrict the flow of scanning ribosomes from the cap site to the GCN4 initiation codon. When amino acids are abundant, ribosomes translate the first uORF and reinitiate at one of the remaining uORFs in the leader, after which they dissociate from the mRNA. Under conditions of amino acid starvation, many ribosomes which have translated uORF1 fail to reinitiate at uORFs 2-4 and utilize the GCN4 start codon instead. Failure to reinitiate at uORFs 2-4 in starved cells results from a reduction in the GTP-bound form of eIF-2 that delivers charged initiator tRNA(iMet) to the ribosome. When the levels of eIF-2.GTP.Met-tRNA(iMet) ternary complexes are low, many ribosomes will not rebind this critical initiation factor following translation of uORF1 until after scanning past uORF4, but before reaching GCN4. Phosphorylation of eIF-2 by the protein kinase GCN2 decreases the concentration of eIF-2.GTP.Met-tRNA(iMet) complexes by inhibiting the guanine nucleotide exchange factor for eIF-2, which is the same mechanism utilized in mammalian cells to inhibit total protein synthesis by phosphorylation of eIF-2.

Nonlinear ribosome migration on cauliflower mosaic virus 35S RNA

Cauliflower mosaic virus 35S RNA contains a 600 nt leader with several small open reading frames that by themselves inhibit translation of downstream coding regions. In the context of the whole leader and in certain plant cells, however, translation of downstream coding regions is allowed. This translation is dependent on the RNA 5' terminus and other elements of the leader. However, its central portion is dispensable or can be modified by insertion of an energy-rich stem-loop structure or long coding region with many internal AUG codons. We conclude that this region can be by-passed (shunted) by the scanning complex. Shunting was also observed in trans between two separate RNA molecules.

Mutations in POL1 increase the mitotic instability of tandem inverted repeats in Saccharomyces cerevisiae

Tandem inverted repeats (TIRs or hairpins) of 30 and 80 base-pair unit lengths are unstable mitotically in yeast (Saccharomyces cerevisiae). TIR instability results from deletions that remove part or all of the presumed hairpin structure from the chromosome. At least one deletion endpoint is always at or near the base of the hairpin, and almost all of the repaired junctions occur within short direct sequence repeats of 4 to 9 base pairs. The frequency of this event, which we call "hairpin excision," is influenced by chromosomal position, length of the inverted repeats, and the distance separating the repeat units; increasing the distance between the inverted repeats as little as 25 base pairs increases their chromosomal stability. The frequency of excision is not affected by representative rad mutations, but is influenced by mutations in certain genes affecting DNA synthesis. In particular, mutations in POL1/CDC17, the gene that encodes the large subunit of DNA polymerase I, increase the frequency of hairpin deletions significantly, implicating this protein in the normal maintainance of genomic TIRs.

Molecular homology among members of the R gene family in maize

The R gene family determines the timing, distribution and amount of anthocyanin pigmentation in maize. This family comprises a set of regulatory genes, consisting of a cluster of several elements at the R locus, on chromosome 10, the Lc and Sn gene lying about two units R distal and B on chromosome 2. Each gene determines a tissue-specific pigmentation of different parts of the seed and plant. The proposed duplicated function of R, Sn, Lc and B loci is reflected in cDNA sequence similarity. In this paper an extensive analysis of the predicted proteins of the R, Sn, Lc and B genes together with a search for putative sites of post-translational modification is reported. A comparison with the prosite database discloses several N-glycosylation and phosphorylation sites, as well as the basic Helix-Loop-Helix (HLH) domain of transcriptional activators. Sn, Lc, and R-S show a high conservation of these sites, while B is more divergent. Analysis of the 5' leader of mRNA sequences discloses the presence of five ATG triplets with two upstream open reading frames (uORFs) of 38 and 15 amino acids and a loop structure indicating a possible mechanism of control at the translational level. It is conceivable that possible mechanisms acting at the translational and post-translational level could modulate the expression and the activation of these transcription factors. Northern analysis of various tissues of different R alleles highlights a strict correlation between pigment accumulation in different tissues and the expression of the regulatory and structural genes suggesting that the pattern of pigmentation relies on a mechanism of differential expression of the members of the R family. Analysis of the Sn promoter discloses the presence of several sequences resembling binding sites of known transcription factors (as GAGA and GT) that might be responsible for the spatial and light-induced expression of this gene. Two regions include a short sequence homologous to the consensus binding site of the B-HLH domain suggesting a self-regulatory control of the Sn gene.

Diversity of mechanisms in the regulation of translation in prokaryotes and lower eukaryotes

Regulation of translation is used to control the expression of many essential and highly expressed genes. The known repertoire of molecular mechanisms for translational regulation is expanding. Recently elucidated mechanisms involve alterations in mRNA structure and modulation of the activity of translation initiation factors.

Molecular biology of translation in yeast

The combination of genetic, molecular and biochemical approaches have made the yeast Saccharomyces cerevisiae a convenient organism to study translation. The sequence similarity of translation factors from yeast and other organisms suggests a high degree of conservation in the translational machineries. This view is also strengthened by a functional analogy of some proteins implicated in translation. Beautiful genetic experiments have confirmed existing models and added new insights in the mechanism of translation. This review summarizes recent experiments using yeast as a model system for the analysis of this complex process.

Control of translation initiation in Saccharomyces cerevisiae

The first observations regarding the control of translation initiation in the yeast Saccharomyces cerevisiae were made by Fred Sherman and his colleagues in 1971. Elegant genetic studies of the CYC1 gene resulted in the formulation of 'Sherman's Rules' for translation initiation as follows: (i) AUG is the only initiator codon. (ii) the most proximal AUG from the 5' end of a message will serve as the start site of translation; and (iii) if the upstream AUG codon is mutated then initiation begins at the next available AUG in the message. Hidden within these rules is the mechanism of eukaryotic translation initiation, as these very same rules were later shown to apply to higher eukaryotic organisms and were formulated into the scanning model. However, only in the past five years has yeast been taken seriously as an organism for studying the mechanism of eukaryotic translation initiation. The basis for this is that the yeast genes for at least four mammalian translation initiation factor homologues have been identified and the number is growing. Similar factors suggest similar mechanisms for translation initiation between yeast and mammals. For some translation initiation factors, the genetics of yeast has provided new insights into their function. A mechanism for regulating translation initiation in mammalian cells is now evident in yeast. It seems clear that the molecular genetics of yeast coupled with the available in vitro translation system will provide a wealth of information in the future regarding translational control and regulatory mechanisms. The purpose of this review is to summarize what is known about translational control in S. cerevisiae.

Phosphorylation of initiation factor 2 alpha by protein kinase GCN2 mediates gene-specific translational control of GCN4 in yeast

We show that phosphorylation of the alpha subunit of eukaryotic translation initiation factor 2 (eIF-2) by the protein kinase GCN2 mediates translational control of the yeast transcriptional activator GCN4. In vitro, GCN2 specifically phosphorylates the alpha subunit of rabbit or yeast eIF-2. In vivo, phosphorylation of eIF-2 alpha increases in response to amino acid starvation, which is dependent on GCN2. Substitution of Ser-51 with alanine eliminates phosphorylation of eIF-2 alpha by GCN2 in vivo and in vitro and abolishes increased expression of GCN4 and amino acid biosynthetic genes under its control in amino acid-starved cells. The Asp-51 substitution mimics the phosphorylated state and derepresses GCN4 in the absence of GCN2. Thus, an established mechanism for regulating total protein synthesis in mammalian cells mediates gene-specific translational control in yeast.

Minigene rescues acetylcholinesterase lethal mutations in Drosophila melanogaster

The gene encoding acetylcholinesterase in Drosophila melanogaster is over 34,000 base-pairs long. We have constructed a 5800 base-pair minigene containing 1500 base-pairs of genomic sequence upstream from the transcription start spliced to the coding sequence, but lacking the nine introns. After germline genetic transformation, this minigene rescues acetylcholinesterase lethal mutants. Tissue-specific distribution appears normal. This allows us to test site-directed mutations of acetylcholinesterase. In a first effort, deletion of most of the unusual 1000 bases leader and its intriguing short open reading frames showed no effect on gene expression. The way is open to study in vivo the structure-function relationships of acetylcholinesterase and insecticide resistance.

A consideration of alternative models for the initiation of translation in eukaryotes

Although recent biochemical and genetic investigations have produced some insights into the mechanism of initiation of translation in eukaryotic cells, two aspects of the initiation process remain controversial. One unsettled issue concerns a variety of functions that have been proposed for mRNA binding proteins, including some initiation factors. The need to distinguish between specific and nonspecific binding of proteins to mRNA is discussed herein. The possibility that certain initiation factors might act as RNA helicases is evaluated along with other ideas about the functions of mRNA- and ATP-binding factors. A second controversial issue concerns the universality of the scanning mechanism for initiation of translation. According to the conventional scanning model, the initial contact between eukaryotic ribosomes and mRNA occurs exclusively at the 5' terminus of the message, which is usually capped. The existence of uncapped mRNAs among a few plant and animal viruses has prompted a vigorous search for other modes of initiation. An "internal initiation" mechanism, first proposed for picornaviruses, has received considerable attention. Although a large body of evidence has been adduced in support of such a mechanism, many of the experiments appear flawed or inconclusive. Some suggestions are given for improving experiments designed to test the internal initiation hypothesis.

AUG codons in the RNA leader sequences of the yeast PET genes CBS1 and SCO1 have no influence on translation efficiency

We report that the major transcription start sites of the yeast PET gene SCO1 are located at positions -149 and -125 relative to the AUG initiation codon of the SCO1 reading frame. The leader sequences of the resulting mRNAs possess a single AUG codon at position -49, which initiates a short open reading frame of three amino acids. The recent finding of a similar situation in the case of the PET gene CBS1 prompted us to address the question as to whether these AUG codons might play some role in the expression of these PET genes. After removal of the upstream AUG codons by site-directed mutagenesis, expression was monitored by use of lacZ fusions and compared to the respective wild-type constructs. Our data show that under all growth conditions tested the leader-contained AUG initiation codons have no significant influence on the expression of both PET genes.

Isolation and nucleotide sequence of the ribosomal protein S16-encoding gene from Aspergillus nidulans

A genomic clone has been isolated from Aspergillus nidulans which is homologous to the ribosomal (r) protein S16-encoding gene of Saccharomyces cerevisiae (S16A) and the r-protein S19-encoding gene of rat (S19). The amino acid (aa) sequences, deduced from nucleotide (nt) sequence analysis, show that in both cases more than 63% of the aa are conserved. The proposed A. nidulans r-protein S16 gene (rps16) differs from that of S. cerevisiae in that it occurs as a single copy in the haploid genome (rather than two copies as in yeast) and contains two putative introns (rather than one). The mRNA leader is long compared to many Aspergillus genes, commencing 293 nt upstream from the coding region, and contains an open reading frame of 13 codons.