Homepage bandeau_genéral

  Director : KEAN, Katherine M. (kathiemb@pasteur.fr)



This newly created research unit is interested in the molecular mechanisms of eukayotic protein synthesis and the translational control of gene expression that can be brought to bear, for example upon infection by pathogens such as RNA viruses. Current research is focused towards translation initiation, with a particular emphasis on recruitment of the 40S ribosomal subunit onto the mRNA. The roles of both mRNA 5'-noncoding and 3'-end sequences in this process are studied, along with cellular factor requirements.



Communication between the cap and the poly(A) tail of cellular mRNAs

We previously showed that efficient translation initiation on cellular mRNAs requires a functional interaction between the 5' cap and the 3' poly(A) tail, mediated by interactions between the proteins that bind the ends of the mRNA (eIF4E and PABP respectively, both bound to the scaffold protein eIF4G). These interactions allow the physical circularisation of cellular mRNAs, and it has been proposed that such circularisation of an mRNA would facilitate ribosome recycling upon translation termination. Our results this year refute this hypothesis, and even question the role of mRNA circularisation per se in efficient translation initiation. Indeed, we have found that the translation efficiency of a capped but non-polyadenylated mRNA is stimulated by the presence of exogenous poly(A) chains. This trans stimulation is as efficient as when the mRNA carries its own poly(A) tail, provided that the exogenous poly(A) chain length is physiological, and by all biochemical criteria examined to date behaves mechanistically in the same way as 3' poly(A) tail co-operation with the 5' cap (Borman et al., submitted to EMBO J).

Mechanism of translation initiation on uncapped mRNA

We have also studied the mechanism of translation initiation on uncapped or non-polyadenylated mRNAs. Recent results show that the translation efficiency on an uncapped, polyadenylated mRNA increases as the length of the 5'-noncoding region is increased. This contrasts with the case of capped, polyadenylated mRNA, where translation efficiency decreases as the length of the 5'-noncoding region is increased. However, translation initiation on uncapped, polyadenylated mRNA remains 5'-end dependent. We have been able to show clearly that it is the 3' poly(A) tail that is responsible for this property of translation initiation (Paulous et al., in preparation for Mol.Cell.Biol.).

Cleavage of the eIF4F complex by picornaviral proteases

It has long been known that one of the components of the eukaryotic translation initiation complex 4F, eIF4G, is cleaved by different picornaviral proteases. However, the physiological consequence of this is still disputed, and several authors claim that eIF4G cleavage cannot completely explain the selective inhibition of host cell translation observed upon infection by most picornaviruses. It was recently shown that PABP, which associates with the EIF4F complex, is a target for the enteroviral 2A and 3C proteases, and it was proposed that its cleavage could contribute to the inhibition of cellular translation. We have shown this year that PABP can also be cleaved by the human rhinovirus 2A protease, but not by another protease that recognises eIF4G, the foot-and-mouth disease virus Lb protease (Borman et al., 2001, J.Virol. 75, 7864-7871).

Translation initiation on an atypical picornavirus RNA, that of hepatitis A virus (HAV)

Picornavirus mRNAs are amongst those translated by an alternative mechanism of initiation, from an IRES (for Internal Ribosome Entry Segment). We have previously shown that the HAV IRES is inactive upon cleavage of eIF4G, whereas all other picornaviral IRESes can use the C-terminal cleavage product of this protein for 40S ribosomal subunit recruitment. We have now shown that the HAV IRES can bind eIF4G, and that the binding affinity is comparable whether eIF4G is cleaved or intact. This suggests that we were not witnessing the results of a problem of protein recognition depending on its form. However, eIF4G is a scaffold protein, which in the cell binds several other factors. We have now found that the displacement of eIF4E or of PABP from eIF4G specifically inhibits HAV IRES activity. Given that the interaction of eIF4E and PABP with eIF4G induces major conformational changes in this protein, we suggest that, unlike other picornaviral IRESes, the HAV IRES is extremely sensitive to the conformation of eIF4G, and is active only with exactly the same conformation as cellular mRNAs. This, coupled to the fact that HAV does not provoke a shut-off of host cell translation during infection, could contribute an explanation of the slow, inefficient infectious cycle of this virus (Borman et al., 2001, J.Virol. 75, 7864-7871).

Structure-function analysis of the poliovirus (PV) IRES

Picornavirus IRESes are approximately 450 nt long, and structured into several domains. Domain E is accepted to be important in PV neurovirulence, as it contains residues that are mutated in the three vaccine strains developed by Albert Sabin and which have been implicated as strong determinants of attenuation. This domain also comprises a large lateral bulge-loop of ill-defined secondary structure geographically juxtaposed relative to these attenuation mutations. The sequence conservation pattern observed within entero- and rhinoviruses would be compatible with conserved intra-bulge base pairing and the possible presentation of a GNRA tetraloop. To gain insight into the RNA structure of domain E of the PV IRES and its role in IRES function, we have undertaken a site-directed mutagenesis programme, coupled with biochemical and virological analyses. The study of mutants within the large lateral bulge-loop provided clear evidence of intra-domain tertiary interactions, and underlined the participation of this entire region of the RNA in PV neurovirulence (Malnou et al., in preparation for J.Virol.). When we placed the potential attenuation determinants of the different vaccine strains in the same genomic context, that of wild-type PV type 1, we found that only one of them resulted in a significantly defective IRES. Recent results concerning this mutant force a considerable revision concerning the mechanism of eIF4G cleavage during viral infection (Malnou et al., in preparation for J.Biol.Chem.).

Regulation of the translational capacity of hepatitis C virus (HCV)

HCV translation is also IRES-dependent, but this IRES is significantly different from those of picornaviruses. Furthermore, the 3' end of HCV mRNA consists of about 100 highly structured nt (the X region) rather than a poly(A) tail. Last year we showed that translation from the HCV IRES is stimulated by the presence of the X region at the 3' end of the RNA, and proposed that HCV RNA would be circularised via a different protein complex from classical cellular mRNAs (Michel et al., 2001, Mol.Cell.Biol. 21, 4097-4109). We have now begun to characterise the quasi-species distribution of the HCV IRES in infected individuals, to address the question of the sequence variability tolerated within the HCV IRES, and the functional consequences of its evolution during infection in man (in collaboration with JM Pawlotsky and M Soler, hôpital Henri Mondor, Créteil). The phylogenetic analysis of sequences from 6 patients before and after interferon alpha treatment, and the determination of RNA secondary structures, showed that the genetic complexity of the 5'-noncoding region is high, similar to that seen for other regions of the genome. In contrast, the genetic diversity is low, indicating that this region is subject to strong conservatory pressures, and most sequence changes do not affect the IRES predicted secondary structure (Soler et al., submitted to J.Virol.). Recent results derive from a functional study concerning patients infected with HCV genotype 1b, a majority of which harbour variants in which the JIIIabc cloverleaf at the heart of the IRES is predicted to be transformed into a simple stem-loop. These variant IRESes have negligible activity. Furthermore, a specific base substitution in the eIF3 binding site results in a hyper-stimulation of IRES activity by the X region, whereas other sequence changes that correlate with destabilisation of the IRES secondary structure are accompanied by an insensitivty to X.


puce Publications of the unit on Pasteur's references database


  Office staff Researchers Scientific trainees Other personnel

KEAN, Katherine M., CNRS, (research director, head of unit, kathiemb@pasteur.fr)

BORMAN, Andrew M., Pasteur Institute, (researcher)

MICHEL, Yanne M., PhD student

MALNOU, Cécile E., PhD student

QUESNOIT, Mélanie, pre-doctoral student

BEAULIEUX, Fréderik, visiting PhD student

NORA, Tamara, undergraduate rotation student

LHUILLIER, Pierre, undergraduate rotation student

PAULOUS, Sylvie, (technician, Pasteur Institute)

FANZONE, Nathalie, (technician, Pasteur Institute fixed term contract)


Page Top research Institut Pasteur homepage

If you have problems with this Web page, please write to rescom@pasteur.fr.