Poliovirus translation: a paradigm for a novel initiation mechanism

All eukaryotic cellular mRNAs, and most viral mRNAs, are blocked at their 5' ends with a cap structure (m7GpppX, where X is any nucleotide). Poliovirus, along with a small number of other animal and plant viral mRNAs, does not contain a 5' cap structure. Since the cap structure functions to facilitate ribosome binding to mRNA, translation of polio-virus must proceed by a cap-independent mechanism. Consistent with this, recent studies have shown that ribosomes can bind to an internal region within the long 5' noncoding sequence of poliovirus RNA. Possible mechanisms for cap-independent translation are discussed. Cap-independent translation of poliovirus RNA is of major importance to the mechanism of shut-off of host protein synthesis after infection. Moreover, it is likely to play a role in determining poliovirus neurovirulence and attenuation.     

Minimum internal ribosome entry site required for poliovirus infectivity.

Translation initiation by internal ribosome binding is a recently discovered mechanism of eukaryotic viral and cellular protein synthesis in which ribosome subunits interact with the mRNAs at internal sites in the 5' untranslated RNA sequences and not with the 5' methylguanosine cap structure present at the extreme 5' ends of mRNA molecules. Uncapped poliovirus mRNAs harbor internal ribosome entry sites (IRES) in their long and highly structured 5' noncoding regions. Such IRES sequences are required for viral protein synthesis. In this study, a novel poliovirus was isolated whose genomic RNA contains two gross deletions removing approximately 100 nucleotides from the predicted IRES sequences within the 5' noncoding region. The deletions originated from previously in vivo-selected viral revertants displaying non-temperature-sensitive phenotypes. Each revertant had a different predicted stem-loop structure within the 5' noncoding region of their genomic RNAs deleted. The mutant poliovirus (Se1-5NC-delta DG) described in this study contains both stem-loop deletions in a single RNA genome, thereby creating a minimum IRES. Se1-5NC-delta DG exhibited slow growth and a pinpoint plaque phenotype following infection of HeLa cells, delayed onset of protein synthesis in vivo, and defective initiation during in vitro translation of the mutated poliovirus mRNAs. Interestingly, the peak levels of viral RNA synthesis in cells infected with Se1-5NC-delta DG occurred at slightly later times in infection than those achieved by wild-type poliovirus, but these mutant virus RNAs accumulated in the host cells during the late phases of virus infection. UV cross-linking assays with the 5' noncoding regions of wild-type and mutated RNAs were carried out in cytoplasmic extracts from HeLa cells and neuronal cells and in reticulocyte lysates to identify the cellular factors that interact with the putative IRES elements. The cellular proteins that were cross-linked to the minimum IRES may represent factors playing an essential role in internal translation initiation of poliovirus mRNAs.     

A viral sequence in the 3''-untranslated region mimics a 5'' cap in facilitating translation of uncapped mRNA.

For recognition by the translational machinery, most eukaryotic cellular mRNAs have a 5' cap structure [e.g. m7G(5')ppp(5')N]. We describe a translation enhancer sequence (3'TE) located in the 3'-untranslated region (UTR) of the genome of the PAV barley yellow dwarf virus (BYDV-PAV) which stimulates translation from uncapped mRNA by 30- to 100-fold in vitro and in vivo to a level equal to that of efficient capped mRNAs. A four base duplication within the 3'TE destroyed the stimulatory activity. Efficient translation was recovered by addition of a 5' cap to this mRNA. Translation of both uncapped mRNA containing the 3'TE in cis and capped mRNA lacking any BYDV-PAV sequence was inhibited specifically by added 3'TE RNA in trans. This inhibition was reversed by adding initiation factor 4F (eIF4F), suggesting that the 3'TE, like the 5' cap, mediates eIF4F-dependent translation initiation. The BYDV-PAV 5'UTR was necessary for the 3'TE to function, except when the 3'TE itself was moved to the 5'UTR. Thus, the 3'TE is sufficient for recruiting the translation factors and ribosomes, while the viral 5'UTR may serve only for the long distance 3'-5' communication. Models are proposed to explain this novel mechanism of cap-independent translation initiation facilitated by the 3'UTR.     

Translational efficiency of poliovirus mRNA: mapping inhibitory cis-acting elements within the 5'' noncoding region.

Poliovirus mRNA contains a long 5' noncoding region of about 750 nucleotides (the exact number varies among the three virus serotypes), which contains several AUG codons upstream of the major initiator AUG. Unlike most eucaryotic mRNAs, poliovirus does not contain a m7GpppX (where X is any nucleotide) cap structure at its 5' end and is translated by a cap-independent mechanism. To study the manner by which poliovirus mRNA is expressed, we examined the translational efficiencies of a series of deletion mutants within the 5' noncoding region of the mRNA. In this paper we report striking translation system-specific differences in the ability of the altered mRNAs to be translated. The results suggest the existence of an inhibitory cis-acting element(s) within the 5' noncoding region of poliovirus (between nucleotides 70 and 381) which restricts mRNA translation in reticulocyte lysate, wheat germ extract, and Xenopus oocytes, but not in HeLa cell extracts. In addition, we show that HeLa cell extracts contain a trans-acting factor(s) that overcomes this restriction.     

Selective destabilization of short-lived mRNAs with the granulocyte-macrophage colony-stimulating factor AU-rich 3'' noncoding region is mediated by a cotranslational mechanism.

The 3' noncoding region element (AUUUA)n specifically targets many short-lived mRNAs for degradation. Although the mechanism by which this sequence functions is not yet understood, a potential link between facilitated mRNA turnover and translation has been implied by the stabilization of cellular mRNAs in the presence of protein synthesis inhibitors. We therefore directly investigated the role of translation on mRNA stability. We demonstrate that mRNAs which are poorly translated through the introduction of stable secondary structure in the 5' noncoding region are not efficiently targeted for selective destabilization by the (AUUUA)n element. These results suggest that AUUUA-mediated degradation involves either a 5'-->3' exonuclease or is coupled to ongoing translation of the mRNA. To distinguish between these two possibilities, we inserted the poliovirus internal ribosome entry site, which promotes internal ribosome initiation, downstream of the 5' secondary structure. Translation directed by internal ribosome binding was found to fully restore targeted destabilization of AUUUA-containing mRNAs despite the presence of 5' secondary structure. This study therefore demonstrates that selective degradation mediated by the (AUUUA)n element is coupled to ribosome binding or ongoing translation of the mRNA and does not involve 5'-to-3' exonuclease activity.     

Effect of viral infection on host protein synthesis and mRNA association with the cytoplasmic cytoskeletal structure

We studied the association of several eucaryotic viral and cellular mRNAs with cytoskeletal fractions derived from normal and virus-infected cells. We found that all mRNAs appear to associate with the cytoskeletal structure during protein synthesis, irrespective of their 5' and 3' terminal structures: e.g., poliovirus that lacks a 5' cap structure or reovirus and histone mRNAs that lack a 3' poly A tail associated with the cytoskeletal framework to the same extent as capped, polyadenylated actin mRNA. Cellular (actin) and viral (vesicular stomatitis virus and reovirus) mRNAs were released from the cytoskeletal framework and their translation was inhibited when cells were infected with poliovirus. In contrast, actin mRNA was not released from the cytoskeleton during vesicular stomatitis virus infection although actin synthesis was inhibited. In addition, several other conditions under which protein synthesis is inhibited did not result in the release of mRNAs from the cytoskeletal framework. We conclude that the association of mRNA with the cytoskeletal framework is required but is not sufficient for protein synthesis in eucaryotes. Furthermore, the shut-off of host protein synthesis during poliovirus infection and not vesicular stomatitis virus infection occurs by a unique mechanism that leads to the release of host mRNAs from the cytoskeleton.     

Cap-independent translation of poliovirus mRNA is conferred by sequence elements within the 5'' noncoding region.

Poliovirus polysomal RNA is naturally uncapped, and as such, its translation must bypass any 5' cap-dependent ribosome recognition event. To elucidate the manner by which poliovirus mRNA is translated, we have determined the translational efficiencies of a series of deletion mutants within the 5' noncoding region of the mRNA. We found striking differences in translatability among the altered mRNAs when assayed in mock-infected and poliovirus-infected HeLa cell extracts. The results identify a functional cis-acting element within the 5' noncoding region of the poliovirus mRNA which enables it to translate in a cap-independent fashion. The major determinant of this element maps between nucleotides 320 and 631 of the 5' end of the poliovirus mRNA. We also show that this region (320 to 631), when fused to a heterologous mRNA, can function in cis to render the mRNA cap independent in translation.     

Internal binding of eucaryotic ribosomes on poliovirus RNA: translation in HeLa cell extracts.

Translation initiation on poliovirus mRNA in poliovirus-infected cells has been shown to occur by internal binding of ribosomes to the 5' noncoding region (J. Pelletier and N. Sonenberg, Nature [London] 334:320-325, 1988). Here we show that internal ribosome binding can occur in HeLa cell extracts in vitro. Internal binding to the 5' noncoding region of poliovirus mRNA in a bicistronic context was independent of the upstream open reading frame and did not require poliovirus proteins.     

Protein synthesis in eukaryotes: the growing biological relevance of cap-independent translation initiation

Ribosome recruitment to eukaryotic mRNAs is generally thought to occur by a scanning mechanism, whereby the 40S ribosomal subunit binds in the vicinity of the 5'cap structure of the mRNA and scans until an AUG codon is encountered in an appropriate sequence context. Study of the picornaviruses allowed the characterization of an alternative mechanism of translation initiation. Picornaviruses can initiate translation via an internal ribosome entry segment (IRES), an RNA structure that directly recruits the 40S ribosomal subunits in a cap and 5' end independent fashion. Since its discovery, the notion of IRESs has extended to a number of different virus families and cellular RNAs. This review summarizes features of both cap-dependent and IRES-dependent mechanisms of translation initiation and discusses the role of cis-acting elements, which include the 5' cap, the 5'-untranslated region (UTR) and the poly(A) tail as well as the possible roles of IRESs as part of a cellular stress response mechanism and in the virus replication cycle.     

In vitro mutational analysis of cis-acting RNA translational elements within the poliovirus type 2 5'' untranslated region.

Initiation of translation on poliovirus RNA occurs by internal binding of ribosomes to a region within the 5' untranslated region (UTR) of the mRNA. This region has been previously roughly mapped between nucleotides 140 and 631 of the 5' UTR and termed the ribosome landing pad. To identify cis-acting elements in the 5' UTR of poliovirus type 2 (Lansing strain) RNA that confer cap-independent internal initiation, we determined the in vitro translational efficiencies of a series of deletion and point mutations within the 5' UTR of the mRNA. The results demonstrate that the 3' border of the core poliovirus ribosome landing pad is located between nucleotides 556 and 585, whereas a region extending between nucleotides 585 and 612 confers enhanced translation. We studied two cis-acting elements within this region of the 5' UTR: a pyrimidine stretch which is critical for translation and an AUG (number 7 from the 5' end) that is located approximately 20 nucleotides downstream from the pyrimidine stretch and augments translation. We also show that the stem-loop structure which contains this AUG is not required for translation.     

The 3'-untranslated region of apolipoprotein II mRNA contains two independent domains that bind distinct cytosolic factors

The 3'-untranslated region of apolipoprotein II (apoII) mRNA contains target sites for mRNA breakdown (Binder, R., Hwang, S.-P. L., Ratnasabapathy, R., and Williams, D. L. (1989) J. Biol. Chem. 264, 16910-16918). Degradation occurs via endonucleolytic cleavage at 5'-AAU-3'/5'-UAA-3' elements in single-stranded loop domains of the 3'-untranslated region. Degradation target sites occur in two clusters that are localized within two larger domains of secondary structure. In this study, gel shift and label transfer assays were used to identify liver cytosolic factors that recognize the 3'-untranslated region of apoII mRNA. The results show preferential binding of cytosolic factors to the 3'-untranslated region as compared to the coding region. UV cross-linking experiments confirmed that cytosolic factors labeled by the 3'-untranslated region are a subset of proteins labeled by the entire mRNA. Two distinct binding domains were identified within the 3'-untranslated region. The upstream domain encompassing nucleotides 400-547 extends from the translation stop codon through the complex stem-loop D structure described previously. This domain labeled primarily a 34-kDa protein in UV cross-linking experiments. The downstream binding domain encompassing nucleotides 568-643 includes another region of secondary structure and terminates within the universal polyadenylation signal. The downstream domain labeled primarily a 60-kDa protein in UV cross-linking experiments. The upstream and downstream binding domains did not compete with each other in gel shift or cross-linking experiments. These results indicate that the 3'-untranslated region can form two independent messenger ribonucleoprotein complexes localized to domains that include target sites for apoII mRNA degradation. We speculate that these messenger ribonucleoprotein complexes may play a role in the degradation of apoII mRNA or in the regulation of this process.     

Specific protein binding to a conserved region of the ornithine decarboxylase mRNA 5'-untranslated region.

An RNA gel retardation assay was used to identify one or more cellular protein(s) (ornithine decarboxylase mRNA 5'-UTR binding protein (ODCBP)) that bind specifically to a conserved region of the 5'-untranslated region (5'-UTR) of rat ornithine decarboxylase (ODC) mRNA. Ultraviolet light cross-linking demonstrated that this protein has an apparent Mr = 58,000 in mammalian cells. Treatment with the oxidizing agent diamide prevented binding of the ODCBP to ODC mRNA; addition of beta-mercaptoethanol reversed this inhibition and permitted mRNA.ODCBP complex formation. Cytoplasmic extracts from a variety of animal cells and tissues demonstrated similar binding activities; however, there was marked tissue-specific expression of the protein in the rat, with brain, heart, lung, and testis containing large amounts, and kidney, spleen, and skeletal muscle expressing negligible amounts. Binding was completely prevented by several mutations within a highly conserved heptanucleotide region (CCAU/ACUC) that was within 61 bases of the initiation codon in ODC mRNAs from mammals, Xenopus, and Caenorhabditis elegans; mutations 5' and 3' of the conserved heptanucleotide domain had no effect on binding activity. Binding was not affected by manipulation of cellular polyamine levels or by treatment of cells with agents that stimulate ODC biosynthesis. Thus, we have identified a widely distributed cellular protein that binds to a conserved domain within the 5'-UTR of ODC mRNA from many animal species; functional consequences of this binding remain to be determined.     

Translation of poliovirus RNA: role of an essential cis-acting oligopyrimidine element within the 5'' nontranslated region and involvement of a cellular 57-kilodalton protein.

Translation of poliovirus RNA is initiated by cap-independent internal entry of ribosomes into the 5' nontranslated region. This process is dependent on elements within the 5' nontranslated region (the internal ribosomal entry site) and may involve novel translation factors. Systematic mutation of a conserved oligopyrimidine tract has revealed a cis-acting element that is essential for translation in vitro. The function of this element is related to its position relative to other cis-acting domains. This element is part of a more complex structure that interacts with several cellular factors, but changes in protein binding after mutation of this element were not detected in a UV cross-linking assay. A 57-kDa protein from the ribosomal salt wash fraction of HeLa cells was identified that binds upstream of the oligopyrimidine tract. Translation of poliovirus mRNA in vitro was strongly and specifically inhibited by competition with the p57-binding domain (nucleotides 260 to 488) of the 5' nontranslated region of encephalomyocarditis virus, indicating a probable role for p57 in poliovirus translation. p57 is likely to be identical to the ribosome-associated factor that binds to and is necessary for the function of the internal ribosomal entry site of encephalomyocarditis virus RNA.     

Poly(A)-binding protein facilitates translation of an uncapped/nonpolyadenylated viral RNA by binding to the 3' untranslated region

Viruses employ an alternative translation mechanism to exploit cellular resources at the expense of host mRNAs and to allow preferential translation. Plant RNA viruses often lack both a 5' cap and a 3' poly(A) tail in their genomic RNAs. Instead, cap-independent translation enhancer elements (CITEs) located in the 3' untranslated region (UTR) mediate their translation. Although eukaryotic translation initiation factors (eIFs) or ribosomes have been shown to bind to the 3'CITEs, our knowledge is still limited for the mechanism, especially for cellular factors. Here, we searched for cellular factors that stimulate the 3'CITE-mediated translation of Red clover necrotic mosaic virus (RCNMV) RNA1 using RNA aptamer-based one-step affinity chromatography, followed by mass spectrometry analysis. We identified the poly(A)-binding protein (PABP) as one of the key players in the 3'CITE-mediated translation of RCNMV RNA1. We found that PABP binds to an A-rich sequence (ARS) in the viral 3' UTR. The ARS is conserved among dianthoviruses. Mutagenesis and a tethering assay revealed that the PABP-ARS interaction stimulates 3'CITE-mediated translation of RCNMV RNA1. We also found that both the ARS and 3'CITE are important for the recruitment of the plant eIF4F and eIFiso4F factors to the 3' UTR and of the 40S ribosomal subunit to the viral mRNA. Our results suggest that dianthoviruses have evolved the ARS and 3'CITE as substitutes for the 3' poly(A) tail and the 5' cap of eukaryotic mRNAs for the efficient recruitment of eIFs, PABP, and ribosomes to the uncapped/nonpolyadenylated viral mRNA.     

Mechanism of mRNA binding to bovine mitochondrial ribosomes

The binding of mRNA to bovine mitochondrial ribosomes was investigated using triplet codons, homopolymers and heteropolymers of various lengths, and human mitochondrial mRNAs. In the absence of initiation factors and initiator tRNA, mitochondrial ribosomes do not bind triplet codons (AUG and UUU) or homopolymers (oligo(U] shorter than about 10 nucleotides. The RNA binding domain on the 28 S mitoribosomal subunit spans approximately 80 nucleotides of the mRNA, judging from the size of the fragments of poly(U,G) and natural mRNAs protected from RNase T1 digestion by this subunit, but the major binding interaction with the ribosome appears to occur over a 30-nucleotide stretch. Human mitochondrial mRNAs coding for subunits II and III of cytochrome c oxidase and subunit 1 of the NADH-ubiquinone oxidoreductase (complex I) were used in studying in detail the binding of mRNA to the small subunit of bovine mitochondrial ribosomes. We have determined that these mRNAs have considerable secondary structure in their 5'-terminal regions and that the initiation codon of each mRNA is sequestered in a stem structure. Little mRNA was bound to ribosomes in a manner conferring protection of the 5' termini from RNase T1 digestion, under standard conditions supporting the binding of artificial templates, but such binding was greatly stimulated by the addition of a mitochondrial extract. Initiation factors and tRNAs from Escherichia coli were unable to stimulate the 5' terminus protected binding of these mRNA molecules, demonstrating a requirement for homologous factors. Our results strongly suggest that mitochondrial initiation factors are required for the proper recognition and melting of the secondary structure in the 5'-terminal region of mitochondrial mRNAs, as a prerequisite for initiation of protein synthesis in mammalian mitochondria.     

SARS coronavirus nsp1 protein induces template-dependent endonucleolytic cleavage of mRNAs: viral mRNAs are resistant to nsp1-induced RNA cleavage

SARS coronavirus (SCoV) nonstructural protein (nsp) 1, a potent inhibitor of host gene expression, possesses a unique mode of action: it binds to 40S ribosomes to inactivate their translation functions and induces host mRNA degradation. Our previous study demonstrated that nsp1 induces RNA modification near the 5'-end of a reporter mRNA having a short 5' untranslated region and RNA cleavage in the encephalomyocarditis virus internal ribosome entry site (IRES) region of a dicistronic RNA template, but not in those IRES elements from hepatitis C or cricket paralysis viruses. By using primarily cell-free, in vitro translation systems, the present study revealed that the nsp1 induced endonucleolytic RNA cleavage mainly near the 5' untranslated region of capped mRNA templates. Experiments using dicistronic mRNAs carrying different IRESes showed that nsp1 induced endonucleolytic RNA cleavage within the ribosome loading region of type I and type II picornavirus IRES elements, but not that of classical swine fever virus IRES, which is characterized as a hepatitis C virus-like IRES. The nsp1-induced RNA cleavage of template mRNAs exhibited no apparent preference for a specific nucleotide sequence at the RNA cleavage sites. Remarkably, SCoV mRNAs, which have a 5' cap structure and 3' poly A tail like those of typical host mRNAs, were not susceptible to nsp1-mediated RNA cleavage and importantly, the presence of the 5'-end leader sequence protected the SCoV mRNAs from nsp1-induced endonucleolytic RNA cleavage. The escape of viral mRNAs from nsp1-induced RNA cleavage may be an important strategy by which the virus circumvents the action of nsp1 leading to the efficient accumulation of viral mRNAs and viral proteins during infection.