Unit: Eukaryotic and Viral Translational Control - CNRS URA 1966
Director: KEAN, Katherine M.
Our research concerns the study of molecular mechanisms of protein synthesis by the eukaryotic ribosome, and translational regulation that can occur according to mRNA structure and in response to infection with pathogenic RNA viruses. Current work concentrates on the initiation step of translation, particularly the role of mRNA 5'-noncoding and 3'-end sequences in 40S ribosomal subunit recruitment.
1) A revision of the poliovirus Internal Ribosome Entry Segment (IRES) structure-function relationship
Picornavirus mRNAs are amongst those translated by an alternative mechanism of initiation, from an IRES (for Internal Ribosome Entry Segment). Picornavirus IRESes are approximately 450 nucleotides long, and are structured into several domains, including domain E that is accepted to be important in poliovirus (PV) neurovirulence. This domain contains residues that are mutated in all three Sabin vaccine strains which have played a crucial role in the control of poliomyelitis since the 1950s and which have been a keystone of the imminent world-wide poliovirus eradication goal. Domain E also encompasses a large lateral bulge of ill-defined structure, juxtaposed to the Sabin mutations. To better understand the structure of the RNA of domain E of the PV IRES and the role of this domain in IRES function, we have carried out a programme of targeted mutagenesis, coupled to biochemical and virological analyses. Our studies have previously shown clear evidence for intra-domain tertiary RNA interactions and highlighted that the whole region plays a role in poliovirus neurovirulence. This year we have accumulated enough data to propose the existence of a critical C motif RNA tertiary structure element, and an equilibrium model for the RNA structure. Furthermore, we can propose more attractive vaccine candidate mutations than the originals, in terms of phenotype and stability, supported by a structural rationale. Finally, the analysis of phenotypic revertants has shown that a defective IRES can be functionally compensated by gainer function mutations in the 2A viral protease.
2) Analysis of the effects of oxygen starvation on morphology of liver cells
This year, we have begun to analyse the effects of oxygen deprivation on gene expression in hepatocytes. Effectively, the mammalian liver, the target of several hepatotrophic viruses including hepatitis C virus (HCV) has an extremely low oxygen pressure compared to other organs. Such deprivation of oxygen supply to cells should result in the onset of numerous metabolic responses destined to conserve energy and enhance survival, including major changes in translation (general suppression; selective maintenance of translation of certain specific mRNAs). Our first results show that the rate of cell growth is not affected by incubation at low oxygen concentration, at least over several days. However, dramatic differences in cell morphology are observed. The habitual confluent cell sheet is replaced by a web-like filamentous network, in which cell-to-cell contact is maintained by the development of prominent cytoplasmic extrusions. The cytoplasmic lipid granules that characterise hepatocytes in culture essentially re-localise to the tips of these extrusions, while the normally large and condensed nuceoli are fragmented into much smaller structures. This correlates with a reduction of approximately 50% in the ribosomal RNA content of the cells. These accumulated changes do not seem to reflect serious cell suffering, since cells begin to regain their usual morphology within 2hrs of return to standard cultivation conditions, "recovery" being complete in under 6hrs.
3) Genetic analysis of HCV translational control
We previously developed in vitro translation systems that recapitulate physiological translation properties such as the 5' cap-3' poly(A) synergy observed in vivo that is the basis of the closed-loop RNA model of eukaryotic translation initiation. These translation systems have allowed us to show that efficient translation of HCV RNA also relies on a closed-loop model, even though the structure of this mRNA is atypical. Effectively, HCV translation is IRES-dependent, and the 3'-end of the RNA also carries a highly structured sequence (the X region, a 100-nt cloverleaf structure) rather than a poly(A) tail. We have found the interplay between the X region and the IRES to be incredibly subtle, and indeed the X region is a remarkable modulator of translation, depending on the exact sequence of the HCV IRES. This year we have followed our policy of advancing collaborative projects by hosting foreign researchers to determine whether domain II of the HCV IRES is involved in the 5'-3' RNA cross-talk. To date, the work carried out jointly with P. Mavromara's laboratory (Institut Pasteur Hellénique) does not indicate a key role for the apical part of domain II.
4) Regulation of murine FGF6 translation
We chose to examine translational control of a member of the fibroblast growth factor family - mitogenic peptides implicated in physiological processes including oncogenesis, angiogenesis, morphogenesis, tissue regeneration and survival. The FGF6 gene was selected because of the nature of the mRNA, which encompasses three potential translation initiation sites. We have found evidence indicating that the initiation codon usage profile varies according to the cell type and differentiation state. In addition, we showed that two FGF6 mRNAs exist, which differ in the length of their 5' noncoding regions. Now we have determined the conditions required for translation of these mRNAs. Unlike other cellular mRNAs, those of murine FGF6 are not translatable in vitro in classical acellular systems, even upon supplementation with limiting cytoplasmic factors. Furthermore, the translation of FGF6 mRNAs in vivo requires their nuclear synthesis, again in contrast to many other mRNAs where an appropriate experimental cytoplasmic transcription can suffice. The results acquired suggest that the mRNAs encoding FGF6 must be loaded with nuclear factors for them to be competent translation templates. Finally, using bicistronic mRNAs, we have been able to show that their 5' noncoding regions contain an IRES.
Keywords: RNA virus, translation initiation, IRES, RNA structure-function, mRNA 5’-3’ cross-talk