Mechanism and Regulation of Eukaryotic Protein Synthesis

A recent interest of the lab has been to study the regulation of translation start site selection. The small ribosomal subunit with bound Met-tRNAi associates with an mRNA at the 5′ cap and then scans down the mRNA in search of a start codon (typically AUG). In previous work with our collaborators at Stanford, we showed that scanning occurs rapidly, and that the ribosome can backtrack when encountering impediments. We also showed that precisely positioned secondary structures can enhance initiation at upstream weak start sites such as the near-cognate start codons CUG or UUG. Whereas translation typically initiates at an AUG codon, the context nucleotides flanking the AUG codon and the translation factors eIF1 and eIF5 influence the efficiency of initiation at a particular AUG codon. Ivaylo Ivanov, a member of our lab, previously showed that eIF1 and eIF5 auto- and trans-regulate the translation of their own mRNAs to control the levels of these factors in cells. We are exploiting this autoregulation to generate reporters to assess start-codon selection stringency, and we are also searching for mRNAs whose translation is sensitive to changes in stringency. In a previous search of mammalian genes, we identified five homeobox (Hox) gene paralogs initiated by AUG codons in conserved suboptimal context, as well as 13 Hox genes that contain evolutionarily conserved upstream open reading frames (uORFs) that initiate at AUG codons in poor sequence context [Reference 1]. We found that the conserved uORFs inhibit Hox reporter expression and that altering the stringency of start-codon selection by overexpressing translation initiation factors eIF1 or eIF5 modulates the expression of Hox reporters. We also showed that modifying ribosome homeostasis by depleting a large ribosomal subunit protein or treating cells with sublethal concentrations of puromycin lowers the fidelity of start-codon selection. As the Hox genes encode developmental regulators of animal body plans, our studies revealed that alterations in start-codon selection stringency could control body plan formation in animals.

In parallel with these studies, we are collaborating with Jody Puglisi and colleagues to molecularly characterize the key processes in translation start site selection, including ribosomal scanning, AUG selection, and ribosomal subunit joining. In recent work, we showed that overexpression of eIF1 or knockdown of eIF5 reduced initiation at near-cognate start codons, whereas knockdown of eIF1 and overexpression of eIF5 had the opposite effects. Consistent with these findings, we found that eIF1 and eIF5 rapidly and transiently sample initiation complexes, providing a rationale for how start-site selection is tuned to the levels of these two factors [Reference 2].

The translation factor eIF5B is a GTPase required for the last step of translation initiation: the joining of the large 60S ribosomal subunit to the small subunit poised on the start codon of an mRNA. The eIF5B binds to the 40S subunit and collaborates in the correct positioning of the initiator Met-tRNA_i^Met on the ribosome in the later stages of translation initiation, gating entrance into elongation. Our ongoing studies with the Puglisi lab reveal that, in addition to promoting 60S subunit joining, eIF5B controls a checkpoint that helps monitor the fidelity of translation start-site selection, a critical determinant in establishing the reading frame for translation on an mRNA, and we are currently characterizing mutations in eIF5B that cause ribosomes to scan past start codons without initiating.

The translation factor eEF2, like its bacterial ortholog EF-G, promotes translocation of tRNAs and mRNA from the A site to the P site on the ribosome following peptide bond formation. In most eukaryotes and archaea, a conserved histidine residue at the tip of eEF2 is post-translationally modified to diphthamide through the action of seven non-essential proteins. The function of diphthamide and the rationale for its evolutionary conservation are not well understood. The name diphthamide is derived from diphtheria, a disease of the nose and throat caused by the bacterium Corynebacterium diphtheriae. The bacterium expresses a toxin that ADP–ribosylates the diphthamide residue, leading to inactivation of eEF2.

Characterizing Saccharomyces cerevisiae mutants that lack diphthamide or that show synthetic growth defects in the absence of diphthamide, we found that diphthamide enhances translational fidelity [Reference 3]. Loss of diphthamide increases –1 ribosomal frameshifting at programmed frameshifting sites in the HIV and SARS-CoV-2 viruses. In addition, using reporter assays, we observed increased rates of frameshifting at non-programmed sites during normal translation elongation. Ribosome profiling of yeast and mammalian cells lacking diphthamide revealed increased ribosomal drop-off during elongation with fewer ribosomes translating to the end of mRNAs. As removal of out-of-frame stop codons restored ribosomal processivity on the ultra-long yeast MDN1 mRNA, we propose that increased rates of ribosomal frameshifting throughout elongation in cells lacking diphthamide leads to premature termination at out-of-frame stop codons. Thus, despite its non-essential nature in yeast or mammalian cells in culture, we propose that diphthamide has been conserved throughout evolution to maintain the fidelity of translation elongation and block spurious frameshifting events that would impair the production of native proteins and generate novel frameshifted proteins that might be deleterious to the cell. Moreover, we propose that the beneficial effects of diphthamide on translational fidelity have ensured its retention during evolution, despite its being a target for inactivation by bacterial toxins.

Our search of genes with poor start codons identified several mRNAs containing noncanonical uORFs initiated by near-cognate start codons that differ from AUG by a single nucleotide change or by AUG codons in poor context. In collaboration with Matthew Sachs, we are characterizing a uORF in the inl gene of the model organism Neurospora crassa, which encodes the first enzyme in inositol biosynthesis. Our preliminary data indicate that the uORF confers translational control in response to inositol. In parallel, we are characterizing a second candidate uORF identified in the plant mRNA encoding GDP-L-galactose phosphorylase (GGP), a control enzyme in the vitamin C biosynthetic pathway. Using reporter assays in mammalian cells and, in vitro, using rabbit reticulocyte lysates, we revealed that the uORF–like element in the GGP mRNA mediates translational control by vitamin C. We propose that interaction of vitamin C with the GGP uORF nascent peptide in the ribosome exit tunnel causes the ribosome to pause and that queuing of subsequent scanning ribosomes results in increased initiation on the uORF and prevents ribosome access to the GGP ORF. We hypothesize that a similar mechanism of nascent peptide recognition of a metabolite mediates translational control the inl mRNA in N. crassa, and that distinct uORFs in other mRNAs confer regulation in response to other metabolites.

The human disease MEHMO syndrome is caused by mutations in the translation initiation factor eIF2gamma. We are characterizing yeast, mammalian cell, and mouse models of the MEHMO syndrome to better understand how the mutations impair eIF2 function and cause disease. In previous studies, we showed that the MEHMO syndrome (named based on the patient phenotypes: mental [intellectual] disability, epilepsy, hypogonadism and hypogenitalism, microcephaly, and obesity) is caused by mutations in the EIF2S3 gene, which encodes the gamma subunit of eIF2. Using genetic and biochemical techniques in yeast, we showed that the mutations linked to the MEHMO syndrome impair eIF2 function, disrupt eIF2 complex integrity, and alter the stringency of translation start-codon selection. Over the past year, we have been characterizing additional novel EIF2S3 mutations identified in patients with the MEHMO syndrome, in part to confirm the pathogenicity of the mutations. Our studies aim to link genetic and biochemical impacts of the mutations to the broad clinical expressivity of MEHMO syndrome.

In previous studies, we characterized induced pluripotent stem (iPS) cells derived from a patient with the MEHMO syndrome. Our studies revealed defects in general protein synthesis, constitutive induction of the integrated stress response (ISR), a cellular stress-response pathway that alters protein synthesis to mount an adaptive response, and hyper-induction of the ISR under stress conditions. The EIF2S3 mutation also impaired neuronal differentiation by the iPS cells. We showed that the drug ISRIB, an activator of the eIF2 guanine nucleotide exchange factor, rescued the defects in cell growth, translation, and neuronal differentiation associated with the EIF2S3 mutation, offering the possibility of therapeutic intervention for the MEHMO syndrome [Reference 4]. Currently, we are generating and characterizing mouse models of the MEHMO syndrome. Whereas the EIF2S3 gene is exclusively on the X chromosome in humans, mice have eif2s3 genes on both the X and Y chromosomes. We introduced a patient mutation into the eif2s3x allele and are characterizing the phenotypes and pathologies of homozygous mutant females. We are also working to introduce the same mutation into the eif2s3y allele so that we can generate mutant males, likely a better model of the human disease.

Translation factor eIF5A is the sole cellular protein containing the unusual amino acid hypusine [N^e-(4-amino-2-hydroxybutyl)lysine]. We previously found that eIF5A promotes translation elongation and translation termination and that these activities are dependent on the hypusine modification. Moreover, using in vivo reporter assays and in vitro translation assays, we showed that eIF5A in yeast, like its bacterial homolog EF-P, is especially critical for the synthesis of proteins containing runs of consecutive proline residues. Given that we previously found that eIF5A binds in the ribosome E site with the hypusine residue projecting toward the acceptor stem of the P-site tRNA, we propose that eIF5A and its hypusine residue function to reposition the acceptor arm of the P-site tRNA to enhance reactivity towards either an aminoacyl-tRNA, for peptide bond formation, or a release factor, for translation termination.

In ongoing studies, we are focusing on three areas: (1) identifying the function of hypusine, (2) elucidating the role of eIF5A in controlling cellular polyamine levels, and (3) characterizing the fungal polyamine transporter Hol1.

To address the first aim, we are investigating the hypusine modification on eIF5A. The modification is formed in two steps: first, an n-butylamine moiety from spermidine is transferred to a specific Lys side chain on eIF5A, whereupon hydroxylation of the added moiety completes the formation of hypusine. In contrast to the essential deoxyhypusine synthase, which catalyzes the first step in hypusine formation, the LIA1 gene, encoding deoxyhypusine hydroxylase (DOHH), is non-essential in yeast. We identified and are now characterizing mutations in eIF5A that cause synthetic growth defects in cells lacking Lia1. The mutations map to the ribosome-binding face of eIF5A and near to magnesium ions that coordinate eIF5A binding to the ribosome. Our results are consistent with the notion that the hydroxyl modification helps bind and position eIF5A and its hypusine residue to effectively promote the reactivity of the peptidyl-tRNA on the ribosome. We are also exploiting these eIF5A mutants, which render hypusine hydroxylation critical for cell growth, so as to establish an assay to test the pathogenicity of mutations in DOHH. Patients with mutations in DOHH exhibit a neurodevelopmental disorder that includes global developmental delay, intellectual disability, facial dysmorphism, and microcephaly. Our yeast-based assays will provide a convenient tool to assess the pathogenicity of DOHH mutations identified in patients.

Towards the second aim, we have linked eIF5A to the regulation of polyamine metabolism in mammalian cells. The enzyme ornithine decarboxylase (ODC) catalyzes the first step in polyamine synthesis. ODC is regulated by a protein called antizyme, which, in turn, is regulated by another protein, called antizyme inhibitor (AZIN1). The synthesis of AZIN1 is inhibited by polyamines. We previously showed that polyamine inhibition of eIF5A triggers increased translation of an inhibitory uORF in the AZIN1 mRNA leader and thereby inhibits the synthesis of AZIN1. In ongoing studies, we found that polyamine control of eIF5A function underlies the translational control of the antizyme (OAZ1) and S-adenosylmethionine decarboxylase (AMD1) mRNAs, which encode other regulators and enzymes in the polyamine biosynthetic pathway. Thus, we propose that eIF5A is a sensor and effector for homeostatic regulation of cellular polyamines.

Regarding the third aim, we identified Hol1 as the high-affinity polyamine transporter in yeast [Reference 5]. The Hol1 protein is a member of the drug-proton antiporter (DHA1) family of transporters, and we showed that HOL1 was required for yeast growth under limiting polyamine conditions and for high-affinity polyamine uptake by yeast. Together with Anirban Banerjee’s lab, we showed that purified Hol1 transports polyamines. We also identified a conserved uORF in the leader of the HOL1 mRNA, and, like the uORF in the AZIN1 mRNA, we found that polyamine inhibition of eIF5A causes the uORF to repress Hol1 synthesis. Thus, polyamine transport, like polyamine biosynthesis, is under translational autoregulation by polyamines in yeast, highlighting the extensive control cells impose on polyamine levels. In ongoing studies, we are characterizing HOL1 homologs in the pathogenic yeast Candida albicans. Our preliminary data indicate that polyamines are critical for C. albicans pathogenesis, raising the possibility that combined inhibition of Hol1 and polyamine synthesis might be an effective means to block growth of this pathogenic yeast.