Christopher U. T. Hellen

Viruses employ diverse mechanisms to gain access to the cellular translation apparatus and to restrict translation of cellular mRNAs to suppress innate immune responses to infection, whereas cells have evolved various countermeasures to prevent viral translation. This Special Issue consists of a series of research papers and reviews from leading experts that advance understanding of the dynamic interplay between viruses and their hosts that account for the structure of viral mRNAs and their capacity to support non-canonical translation mechanisms, and that underlie aspects of infection, including viral replication, pathogenesis and host range.

Viruses depend on the cellular translation apparatus, and the translation process is consequently a focus for competing mechanisms to promote and to restrict viral translation

Members of the Picornaviridae family of viruses have single-stranded positive-sense RNA genomes, and are conventionally considered to contain a single open reading frame (ORF) preceded by a long 5′ untranslated region (UTR). The 5′UTR contains an internal ribosomal entry site (IRES) that mediates end-independent initiation of translation of ORF1, resulting in synthesis of a polyprotein that is proteolytically processed to yield the individual structural and nonstructural proteins. However, recent studies indicate that the dogma that picornavirus genomes contain a single ORF (ORF1) is an oversimplification, and that members of many picornavirus genera contain an additional ORF that is translated to yield an accessory protein that may influence pathogenicity and tissue specificity. This review summarizes the structural organization of picornavirus genomes that contain a second ORF. Most commonly, these ORFs overlap the 5′-end of ORF1, but they may also be wholly distinct from ORF1 or located entirely within it in a different reading frame. Ribosomal access to these alternate ORFs has been shown or is hypothesized to depend on a remarkable variety of non-canonical translation mechanisms, including ribosomal recruitment by a second IRES, leaky scanning after IRES-mediated internal ribosomal entry, cap-independent initiation, frame-shifting and termination–reinitiation.

SARS-CoV-2 nonstructural protein 1 (Nsp1) binds to 40S ribosomal subunits and induces host protein synthesis shut off by inhibiting translation initiation and triggering endonucleolytic cleavage of cellular mRNAs. Irrespective of the mode of initiation, Nsp1-mediated cleavage is induced by the cooperative action of the N-terminal domain of Nsp1, the RRM domain of eIF3g and 40S subunits. Using in vitro reconstitution, we determined that cleavage occurs by transesterification following intramolecular nucleophilic attack of the 2'OH of the ribose on the adjacent phosphodiester bond yielding 5'OH and 2',3'-cyclic phosphate termini. Cleavage requires a guanosine ~10-22 nucleotides from the 5' end of mRNA, occurs within a narrow window upstream of this G, is most efficient between nucleotides at positions -6/-7 and -7/-8 relative to G, and shows a preference for Pu at positions -7 or -8 which provides the 2'OH for the nucleophilic attack. Zero-length UV cross-linking of Nsp1 to nucleotides at positions -1 and -2 suggests that the critical guanosine may be recognized by Nsp1. Resistance to Nsp1-mediated cleavage of SARS-CoV-2 mRNA was ensured both by the relatively long distance between its G23G24 and the 5'end and by the preceding oligoPy stretch lacking purines at positions -7 or -8 upstream of G23G24.

Viral internal ribosomal entry sites (IRESs) are highly structured cis-acting RNAs that mediate end-independent initiation of translation. Their origin remains obscure. The simplest IRESs (type 6) occur in the intergenic region of Dicistroviridae genomes (order Picornavirales), consist of two pseudoknots, and initiate translation by factor-independent binding to ribosomes. Larger variants contain a third pseudoknot that modifies the mechanism of IRES function by engaging with the ribosomal head and promoting binding to the ribosomal aminoacyl site. Metagenomic analyses undertaken to identify structurally distinct type 6 IRESs identified subsets ranging from ∼120–260 nt in length. They differ by the cumulative addition of structural elements, suggesting an accretion mechanism for the structural elaboration of IRESs. Insertions occurred at specific loci, possibly reflecting non-templated nucleotide insertion during replication, and form additional subdomains. Biochemical analysis showed that these novel classes of type 6 IRES all bound directly to the ribosomal peptidyl site. Identification of chimeric IRESs implicates recombinational exchange of domains as a second mechanism for the diversification of IRES structure. Recombination likely also accounts for the presence of type 6 IRESs at the 5′-end of dicistrovirus-like genomes and in families other than Dicistroviridae, including Marnaviridae (order Picornavirales) and Tombusviridae (order Tolivirales).

We utilized the nedicistrovirus (NediV) intergenic region (IGR) internal ribosomal entry site (IRES)-mediated, initiation factor-independent translation initiation system and determined high-resolution structures of 80S ribosome complexes with the NediV IRES in various functional states, including binary complexes, aminoacyl–transfer RNA (tRNA)-bound complexes, and complexes with elongation factor eEF2. In binary complexes, the NediV IRES primarily occupies the ribosomal P site, exhibiting conformational flexibility and engaging the ribosome at multiple interaction sites. Upon translocation, the IRES undergoes structural rearrangements, including destabilization of its PKI domain, facilitating the transition to canonical elongation. Crucially, we captured an eEF2-bound complex, along with an eEF1A-bound failed decoding complex featuring a mismatched tRNA, the latter representing the first instance of a canonical elongation complex visualized in the presence of a natural, hydrolysable nucleotide and without the addition of any trapping agents. These findings provide a comprehensive structural overview of IGR IRES-mediated translation initiation and its transition to elongation, revealing key mechanistic details of viral translation and proofreading.

The encephalomyocarditis virus (EMCV) internal ribosomal entry side (IRES) and other Type 2 IRESs favor translation of the viral genome during infection. The domains H-L of these IRESs specifically interact with the cellular translation initiation factors eIF4G/eIF4A through their essential JK domain. However, the JK domain is not sufficient for IRES activity, which also strictly requires the preceding domain I of unknown function. To identify interactions that drive ribosomal attachment to eIF4G/eIF4A-bound Type 2 IRESs, we determined the cryo-EM structure of 48S initiation complexes formed on the EMCV IRES. The apical cloverleaf of domain I contacts ribosomal proteins uS13 and uS19 via its subdomain Id, whereas the essential GNRA tetraloop in subdomain Ic interacts with the TψC domain of initiator tRNA. The IRES-tRNA interaction also provides a mechanism for release of the IRES after eIF2 is replaced by eIF5B during subunit joining to allow attachment of 60S subunits. Functional assays supported the exceptional role of these interactions for initiation on this IRES. The strong conservation of the apex of domain I amongst Type 2 IRESs suggests that the reported interactions provide a common general mechanism of ribosomal attachment on them all.

We utilized the Nedicistrovirus (NediV) intergenic region (IGR) IRES-mediated, initiation factor-independent translation initiation system and determined high-resolution structures of 80S ribosome complexes with the NediV IRES in various functional states, including binary complexes, aminoacyl-tRNA-bound complexes, and complexes with elongation factor eEF2. In binary complexes, the NediV IRES primarily occupies the ribosomal P site, exhibiting conformational flexibility and engaging the ribosome at multiple interaction sites. Upon translocation, the IRES undergoes structural rearrangements, including destabilization of its PKI domain, facilitating the transition to canonical elongation. Crucially, we captured an eEF2-bound complex, along with an eEF1A-bound post-proofreading complex featuring a mismatched tRNA, the latter representing the first instance of a canonical elongation complex visualized in the presence of a natural, hydrolysable nucleotide and without the addition of any trapping agents. These findings provide a comprehensive structural overview of IGR IRES-mediated translation initiation and its transition to elongation, revealing key mechanistic details of viral translation and proofreading.

The DExH-box helicase DHX29 plays a critical role in mammalian translation initiation. It is required for the scanning of mRNAs with complex 5’UTRs. Despite its importance, the detailed mechanism of DHX29’s action has remained debated. Using structural models derived from AlphaFold and cryo-EM structure of DHX29 bound to the ribosomal 43S pre-initiation complex, we provide a revised structural framework that clarifies the interplay between DHX29, the 40S ribosomal subunit, and eIF3. Our findings suggest that the 40S subunit regulates DHX29’s NTPase activity through an activation mechanism resembling the G-patch protein regulation of DEAH helicases. Moreover, our model supports a 3’ to 5’ translocase mechanism, in which DHX29 transiently pulls the mRNA opposite to the scanning direction, destabilizing stable stem-loops trapped in the mRNA channel and allowing scanning to resume. This structural analysis refines our understanding of DHX29’s function and provides new hypotheses regarding its role in mRNA unwinding during scanning and start codon selection.

The encephalomyocarditis virus (EMCV) IRES and other Type 2 IRESs comprise domains H-L and specifically interact with eIF4G/eIF4A through their essential JK domain. However, the JK domain is not sufficient for IRES function, which also requires the preceding domain I of unknown function. To identify interactions that drive ribosomal recruitment of eIF4G/eIF4A-bound Type 2 IRESs, we determined the cryo-EM structure of 48S initiation complexes formed on the EMCV IRES. It revealed that the apical domain I cloverleaf contacts ribosomal proteins uS13 and uS19 via its Id subdomain and that the essential GNRA tetraloop in subdomain Ic interacts directly with the TψC domain of initiator tRNA. Functional assays supported the exceptional role of these interactions for initiation on this IRES. The strong conservation of primary and secondary structures of the apex of domain I among Type 2 IRESs suggests that the reported interactions are a common essential feature of them all.