RNA levers and switches controlling viral gene expression

Abstract

RNA viruses are diverse and abundant pathogens that are responsible for numerous human diseases. RNA viruses possess relatively compact genomes and have therefore evolved multiple mechanisms to maximize their coding capacities, often by encoding overlapping reading frames. These reading frames are then decoded by mechanisms such as alternative splicing and ribosomal frameshifting to produce multiple distinct proteins. These solutions are enabled by the ability of the RNA genome to fold into 3D structures that can mimic cellular RNAs, hijack host proteins, and expose or occlude regulatory protein-binding motifs to ultimately control key process in the viral life cycle. We highlight recent findings focusing on less conventional mechanisms of gene expression and new discoveries on the role of RNA structures.

Keywords: RNA structure; RNA viruses; ribosomal frameshifting.

Figure 1.. Common RNA secondary structures.

(A) A simple stem-loop with eight paired bases forming a stem and ten unpaired bases forming a loop. Also known as a hairpin. (B) A stem-loop containing a two-base bulge (C) A stem-loop containing a four-base interior loop. (D) A K-type pseudoknot (kissing loop) in which unpaired bases in one stem-loop pair with the unpaired bases in another. € An H-type pseudoknot in which the unpaired bases in a loop form base-pairing interactions with an adjacent single-stranded region. (F) A long-range interaction in which base-pairing interactions occur over large intragenomic distances.

Figure 2.. Programmed ribosomal frameshifting (PRF) and methods of quantification.

(A) A frameshift-stimulatory element (FSE) comprises a “slippery site” (N NNW WWH) and a downstream RNA structure that stimulates a fraction of ribosomes to frameshift backwards by 1 nt. For most FSEs, frameshifted ribosomes bypass a 0-frame stop codon and translate a longer polypeptide (top), while ribosomes that remain in frame stop earlier and translate a shorter polypeptide (bottom). (B) The fraction of ribosomes that frameshift (the “PRF efficiency”) can be measured using several methods. Western blotting can be applied to infected cells to measure the native PRF efficiency of live virus under physiological conditions but requires antibodies and lacks precision and sensitivity. Dual luciferase/fluorescence assays are precise and extremely sensitive and can be done in high throughput, but require artificial reporters such as firefly luciferase (FLuc) and either Renilla (RLuc) or Nano (NLuc) luciferase. Ribosome profiling can be applied to virus-infected cells, gives precise measurements, and provides additional information such as ribosome pause sites, but is cost- and labor-intensive compared to western blotting and luciferase assays.

Figure 3.. Roles of PRF in the life cycles of HIV-1 and SARS-CoV-2.

(A) (Top) The HIV-1 genome contains an FSE between the genes gag (structural proteins) and pol (enzymes) that causes 5 – 10% of ribosomes to frameshift. This ratio is optimal for assembling immature virions, which contain a roughly 20:1 ratio of Gag to Gag-Pol, and then mature via proteolysis using the Gag-Pol encoded protease. MA: matrix protein, CA: capsid protein, NC: nucleocapsid protein, PR: protease, RT: reverse transcriptase, IN: integrase. (Bottom) Mutants producing only Gag-Pol fail to generate virions; it is thought that overexpression of Gag-Pol activates the protease prematurely, which then cleaves the Gag-Pol proteins before virion assembly. Adapted from “HIV-1 Genome and Structure”, by BioRender.com (2022). Retrieved from https://app.biorender.com/biorender-templates. (B) The SARS-CoV-2 genome contains an FSE between the genes ORF1a and ORF1b. The PRF efficiency has been measured at 20 – 40% in artificial reporters and 50 – 70% in infected cells. Adapted from “Genomic Organization of SARS-CoV-2”, by BioRender.com (2022). Retrieved from https://app.biorender.com/biorender-templates.

Figure 4.. Diverse structural topologies of FSEs among different genera of viruses.

(A) Literature-based models of the presumed secondary structures of minimal FSEs from human coronavirus 229E (HCoV-229E) [56], SARS coronavirus 2 (SARS-CoV-2) [35], infectious bronchitis virus (IBV) [34], human immunodeficiency virus 1 (HIV-1) [60], Rous sarcoma virus (RSV) [32], and barley yellow dwarf virus (BYDV) [124]. Locations of the 5’ and 3’ ends and slippery site (SS) are indicated for each model. Colors indicate nucleotide identities. Created with VARNA [125]. (B) Models of the predominant secondary structures of the SARS-CoV-2 FSE in infected cells, based on DMS probing [36] and psoralen crosslinking [66]. Locations of alternative stem 1 (AS1), attenuator hairpin (AH), stem 2 (S2), stem 3 (S3), and the FSE Arch (named by Ziv et al. [66] ) are indicated. Proportion of each structure is based on data from Lan et al. [36]. Created with BioRender.com.

Figure 5.. Regulation of alternative splicing by RNA structures.

(A) When the splice sites and branch point are accessible (left), the small nuclear ribonucleoproteins (snRNPs) U1 and U2/U2AF bind to the 5’ and 3’ splice sites, respectively, and splice out the intron. When the splice sites are occluded by the RNA secondary structure (right), the snRNPs are unable to bind, and the intron remains in the mature mRNA. (B) The HIV-1 RNA surrounding the A3 splice acceptor site adopts two predominant conformations in infected T cells. Exposure of the splice site (left) permits splicing to generate mRNA for Tat protein, while occlusion (right) blocks splicing to maintain full-length RNAs.

Figure 6.. Discontinuous transcription of subgenomic RNAs in nidoviruses.

The plus-sense genomic RNA (top) contains one leader TRS (TRS-L) near the 5’ end and multiple body TRSs (TRS-Bs), each immediately upstream of the start codon of a structural protein, in the 3’ half of the genome. Discontinuous transcription begins with the viral replicase complex at the 3’ end of the genome and moves towards the 5’ end until reaching one of the TRS-Bs. The replicase and nascent minus-sense transcript detach, then re-anneal to the highly homologous TRS-L 20 – 30 kb upstream. The final 5’ portion (the “leader sequence”) is then appended to the growing end of the transcript. The finished minus-sense subgenomic RNA is used as a template for synthesizing plus-sense subgenomic RNAs, which are then translated into viral structural proteins.