Design of a self-regulating mRNA gene circuit.

Protein expression in vivo is predominately controlled via regulatory feedback mechanisms that adjust the level of mRNA transcription. However for positive sense single-stranded RNA viruses, protein expression is often controlled via secondary structural elements, such as internal ribosomal entry sites, that are encoded within the mRNA. The self-regulation of mRNA translation observed in this class of viruses suggests that it may be possible to design mRNAs that self-regulate their protein expression, enabling the creation of mRNAs for vaccines and other synthetic biology applications where protein levels in the cell can be tightly controlled without feedback to a transcriptional mechanism. As a proof of concept, I design a polycistronic mRNA based on bacteriophage MS2, where the upstream gene is capable of repressing synthesis of the downstream gene. Using a computational tool that simulates ribosome kinetics and the co-translational folding of the mRNA in response, I show that mutations to the mRNA can be identified which enhance the efficiency of the translation and the repression of the downstream gene. The results of this study open up the possibility of designing bespoke mRNA gene circuits in which the amount of protein synthesised in cells are self-regulated for therapeutic or antigenic purposes.

Modelling ribosome kinetics and translational control on dynamic mRNA.

The control of protein synthesis and the overall levels of various proteins in the cell is critical for achieving homoeostasis. Regulation of protein levels can occur at the transcriptional level, where the total number of messenger RNAs in the overall transcriptome are controlled, or at the translational level, where interactions of proteins and ribosomes with the messenger RNA determine protein translational efficiency. Although transcriptional control of mRNA levels is the most commonly used regulatory control mechanism in cells, positive-sense single-stranded RNA viruses often utilise translational control mechanisms to regulate their proteins in the host cell. Here I detail a computational method for stochastically simulating protein synthesis on a dynamic messenger RNA using the Gillespie algorithm, where the mRNA is allowed to co-translationally fold in response to ribosome movement. Applying the model to the test case of the bacteriophage MS2 virus, I show that the models ability to accurately reproduce experimental measurements of coat protein production and translational repression of the viral RNA dependant RNA polymerase at high coat protein concentrations. The computational techniques reported here open up the potential to examine the infection dynamics of a ssRNA virus in a host cell at the level of the genomic RNA, as well as examine general translation control mechanisms present in polycistronic mRNAs.

Genome-regulated Assembly of a ssRNA Virus May Also Prepare It for Infection.

Many single-stranded, positive-sense RNA viruses regulate assembly of their infectious virions by forming multiple, cognate coat protein (CP)-genome contacts at sites termed Packaging Signals (PSs). We have determined the secondary structures of the bacteriophage MS2 ssRNA genome (gRNA) frozen in defined states using constraints from X-ray synchrotron footprinting (XRF). Comparison of the footprints from phage and transcript confirms the presence of multiple PSs in contact with CP dimers in the former. This is also true for a virus-like particle (VLP) assembled around the gRNA in vitro in the absence of the single-copy Maturation Protein (MP) found in phage. Since PS folds are present at many sites across gRNA transcripts, it appears that this genome has evolved to facilitate this mechanism of assembly regulation. There are striking differences between the gRNA-CP contacts seen in phage and the VLP, suggesting that the latter are inappropriate surrogates for aspects of phage structure/function. Roughly 50% of potential PS sites in the gRNA are not in contact with the protein shell of phage. However, many of these sit adjacent to, albeit not in contact with, PS-binding sites on CP dimers. We hypothesize that these act as PSs transiently during assembly but subsequently dissociate. Combining the XRF data with PS locations from an asymmetric cryo-EM reconstruction suggests that the genome positions of such dissociations are non-random and may facilitate infection. The loss of many PS-CP interactions towards the 3' end of the gRNA would allow this part of the genome to transit more easily through the narrow basal body of the pilus extruding machinery. This is the known first step in phage infection. In addition, each PS-CP dissociation event leaves the protein partner trapped in a non-lowest free-energy conformation. This destabilizes the protein shell which must disassemble during infection, further facilitating this stage of the life-cycle.

The impact of local assembly rules on RNA packaging in a T = 1 satellite plant virus.

The vast majority of viruses consist of a nucleic acid surrounded by a protective icosahedral protein shell called the capsid. During viral infection of a host cell, the timing and efficiency of the assembly process is important for ensuring the production of infectious new progeny virus particles. In the class of single-stranded RNA (ssRNA) viruses, the assembly of the capsid takes place in tandem with packaging of the ssRNA genome in a highly cooperative co-assembly process. In simple ssRNA viruses such as the bacteriophage MS2 and small RNA plant viruses such as STNV, this cooperative process results from multiple interactions between the protein shell and sites in the RNA genome which have been termed packaging signals. Using a stochastic assembly algorithm which includes cooperative interactions between the protein shell and packaging signals in the RNA genome, we demonstrate that highly efficient assembly of STNV capsids arises from a set of simple local rules. Altering the local assembly rules results in different nucleation scenarios with varying assembly efficiencies, which in some cases depend strongly on interactions with RNA packaging signals. Our results provide a potential simple explanation based on local assembly rules for the ability of some ssRNA viruses to spontaneously assemble around charged polymers and other non-viral RNAs in vitro.

Assembly of infectious enteroviruses depends on multiple, conserved genomic RNA-coat protein contacts.

Picornaviruses are important viral pathogens, but despite extensive study, the assembly process of their infectious virions is still incompletely understood, preventing the development of anti-viral strategies targeting this essential part of the life cycle. We report the identification, via RNA SELEX and bioinformatics, of multiple RNA sites across the genome of a typical enterovirus, enterovirus-E (EV-E), that each have affinity for the cognate viral capsid protein (CP) capsomer. Many of these sites are evolutionarily conserved across known EV-E variants, suggesting they play essential functional roles. Cryo-electron microscopy was used to reconstruct the EV-E particle at ~2.2 Å resolution, revealing extensive density for the genomic RNA. Relaxing the imposed symmetry within the reconstructed particles reveals multiple RNA-CP contacts, a first for any picornavirus. Conservative mutagenesis of the individual RNA-contacting amino acid side chains in EV-E, many of which are conserved across the enterovirus family including poliovirus, is lethal but does not interfere with replication or translation. Anti-EV-E and anti-poliovirus aptamers share sequence similarities with sites distributed across the poliovirus genome. These data are consistent with the hypothesis that these RNA-CP contacts are RNA Packaging Signals (PSs) that play vital roles in assembly and suggest that the RNA PSs are evolutionarily conserved between pathogens within the family, augmenting the current protein-only assembly paradigm for this family of viruses.

An Intracellular Model of Hepatitis B Viral Infection: An In Silico Platform for Comparing Therapeutic Strategies.

Hepatitis B virus (HBV) is a major focus of antiviral research worldwide. The International Coalition to Eliminate HBV, together with the World Health Organisation (WHO), have prioritised the search for a cure, with the goal of eliminating deaths from viral hepatitis by 2030. We present here a comprehensive model of intracellular HBV infection dynamics that includes all molecular processes currently targeted by drugs and agrees well with the observed outcomes of several clinical studies. The model reveals previously unsuspected kinetic behaviour in the formation of sub-viral particles, which could lead to a better understanding of the immune responses to infection. It also enables rapid comparative assessment of the impact of different treatment options and their potential synergies as combination therapies. A comparison of available and currently developed treatment options reveals that combinations of multiple capsid assembly inhibitors perform best.

A stochastic model for simulating ribosome kinetics in vivo.

Computational modelling of in vivo protein synthesis is highly complicated, as it requires the simulation of ribosomal movement over the entire transcriptome, as well as consideration of the concentration effects from 40+ different types of tRNAs and numerous other protein factors. Here I report on the development of a stochastic model for protein translation that is capable of simulating the dynamical process of in vivo protein synthesis in a prokaryotic cell containing several thousand unique mRNA sequences, with explicit nucleotide information for each, and report on a number of biological predictions which are beyond the scope of existing models. In particular, I show that, when the complex network of concentration dependent interactions between elongation factors, tRNAs, ribosomes, and other factors required for protein synthesis are included in full detail, several biological phenomena, such as the increasing peptide elongation rate with bacterial growth rate, are predicted as emergent properties of the model. The stochastic model presented here demonstrates the importance of considering the translational process at this level of detail, and provides a platform to interrogate various aspects of translation that are difficult to study in more coarse-grained models.

Cryo-EM structure and in vitro DNA packaging of a thermophilic virus with supersized T=7 capsids.

Double-stranded DNA viruses, including bacteriophages and herpesviruses, package their genomes into preformed capsids, using ATP-driven motors. Seeking to advance structural and mechanistic understanding, we established in vitro packaging for a thermostable bacteriophage, P23-45 of Thermus thermophilus Both the unexpanded procapsid and the expanded mature capsid can package DNA in the presence of packaging ATPase over the 20 °C to 70 °C temperature range, with optimum activity at 50 °C to 65 °C. Cryo-EM reconstructions for the mature and immature capsids at 3.7-Å and 4.4-Å resolution, respectively, reveal conformational changes during capsid expansion. Capsomer interactions in the expanded capsid are reinforced by formation of intersubunit ß-sheets with N-terminal segments of auxiliary protein trimers. Unexpectedly, the capsid has T=7 quasi-symmetry, despite the P23-45 genome being twice as large as those of known T=7 phages, in which the DNA is compacted to near-crystalline density. Our data explain this anomaly, showing how the canonical HK97 fold has adapted to double the volume of the capsid, while maintaining its structural integrity. Reconstructions of the procapsid and the expanded capsid defined the structure of the single vertex containing the portal protein. Together with a 1.95-Å resolution crystal structure of the portal protein and DNA packaging assays, these reconstructions indicate that capsid expansion affects the conformation of the portal protein, while still allowing DNA to be packaged. These observations suggest a mechanism by which structural events inside the capsid can be communicated to the outside.

A modelling paradigm for RNA virus assembly.

Virus assembly, a key stage in any viral life cycle, had long been considered to be primarily driven by protein-protein interactions and nonspecific interactions between genomic RNA and capsid protein. We review here a modelling paradigm for RNA virus assembly that illustrates the crucial roles of multiple dispersed, specific interactions between viral genomes and coat proteins in capsid assembly. The model reveals how multiple sequence-structure motifs in the genomic RNA, termed packaging signals, with a shared coat protein recognition motif enable viruses to overcome a viral assembly-equivalent of Levinthal's Paradox in protein folding. The fitness advantages conferred by this mechanism suggest that it should be widespread in viruses, opening up new perspectives on viral evolution and anti-viral therapy.

RNA Virus Evolution via a Quasispecies-Based Model Reveals a Drug Target with a High Barrier to Resistance.

The rapid occurrence of therapy-resistant mutant strains provides a challenge for anti-viral therapy. An ideal drug target would be a highly conserved molecular feature in the viral life cycle, such as the packaging signals in the genomes of RNA viruses that encode an instruction manual for their efficient assembly. The ubiquity of this assembly code in RNA viruses, including major human pathogens, suggests that it confers selective advantages. However, their impact on viral evolution cannot be assessed in current models of viral infection that lack molecular details of virus assembly. We introduce here a quasispecies-based model of a viral infection that incorporates structural and mechanistic knowledge of packaging signal function in assembly to construct a phenotype-fitness map, capturing the impact of this RNA code on assembly yield and efficiency. Details of viral replication and assembly inside an infected host cell are coupled with a population model of a viral infection, allowing the occurrence of therapy resistance to be assessed in response to drugs inhibiting packaging signal recognition. Stochastic simulations of viral quasispecies evolution in chronic HCV infection under drug action and/or immune clearance reveal that drugs targeting all RNA signals in the assembly code collectively have a high barrier to drug resistance, even though each packaging signal in isolation has a lower barrier than conventional drugs. This suggests that drugs targeting the RNA signals in the assembly code could be promising routes for exploitation in anti-viral drug design.

Protein-mediated RNA folding governs sequence-specific interactions between rotavirus genome segments.

Segmented RNA viruses are ubiquitous pathogens, which include influenza viruses and rotaviruses. A major challenge in understanding their assembly is the combinatorial problem of a non-random selection of a full genomic set of distinct RNAs. This process involves complex RNA-RNA and protein-RNA interactions, which are often obscured by non-specific binding at concentrations approaching in vivo assembly conditions. Here, we present direct experimental evidence of sequence-specific inter-segment interactions between rotavirus RNAs, taking place in a complex RNA- and protein-rich milieu. We show that binding of the rotavirus-encoded non-structural protein NSP2 to viral ssRNAs results in the remodeling of RNA, which is conducive to formation of stable inter-segment contacts. To identify the sites of these interactions, we have developed an RNA-RNA SELEX approach for mapping the sequences involved in inter-segment base-pairing. Our findings elucidate the molecular basis underlying inter-segment interactions in rotaviruses, paving the way for delineating similar RNA-RNA interactions that govern assembly of other segmented RNA viruses.

A Model for Viral Assembly around an Explicit RNA Sequence Generates an Implicit Fitness Landscape.

Previously, a stochastic model of single-stranded RNA virus assembly was created to model the cooperative effects between capsid proteins and genomic RNA that would occur in a packaging signal-mediated assembly process. In such an assembly scenario, multiple secondary structural elements from within the RNA, termed "packaging signals" (PS), contact coat proteins and facilitate efficient capsid assembly. In this work, the assembly model is extended to incorporate explicit nucleotide sequence information as well as simple aspects of RNA folding that would be occurring during the RNA/capsid coassembly process. Applying this paradigm to a dodecahedral viral capsid, a computer-derived nucleotide sequence is evolved de novo that is optimal for packaging the RNA into capsids, while also containing capacity for coding for a viral protein. Analysis of the effects of mutations on the ability of the RNA sequence to successfully package into a viral capsid reveals a complex fitness landscape where the majority of mutations are neutral with respect to packaging efficiency with a small number of mutations resulting in a near-complete loss of RNA packaging. Moreover, the model shows how attempts to ablate PSs in the viral RNA sequence may result in redundant PSs already present in the genome fulfilling their packaging role. This explains why recent experiments that attempt to ablate putative PSs may not see an effect on packaging. This modeling framework presents an example of how an implicit mapping can be made from genotype to a fitness parameter important for viral biology, i.e., viral capsid yield, with potential applications to theoretical models of viral evolution.

Erratum: Genomic RNA folding mediates assembly of human parechovirus.

A correction has been published and is appended to both the HTML and PDF versions of this paper. The error has not been fixed in the paper.

HBV RNA pre-genome encodes specific motifs that mediate interactions with the viral core protein that promote nucleocapsid assembly.

Formation of the hepatitis B virus nucleocapsid is an essential step in the viral lifecycle, but its assembly is not fully understood. We report the discovery of sequence-specific interactions between the viral pre-genome and the hepatitis B core protein that play roles in defining the nucleocapsid assembly pathway. Using RNA SELEX and bioinformatics, we identified multiple regions in the pre-genomic RNA with high affinity for core protein dimers. These RNAs form stem-loops with a conserved loop motif that trigger sequence-specific assembly of virus-like particles (VLPs) at much higher fidelity and yield than in the absence of RNA. The RNA oligos do not interact with preformed RNA-free VLPs, so their effects must occur during particle assembly. Asymmetric cryo-electron microscopy reconstruction of the T = 4 VLPs assembled in the presence of one of the RNAs reveals a unique internal feature connected to the main core protein shell via lobes of density. Biophysical assays suggest that this is a complex involving several RNA oligos interacting with the C-terminal arginine-rich domains of core protein. These core protein-RNA contacts may play one or more roles in regulating the organization of the pre-genome during nucleocapsid assembly, facilitating subsequent reverse transcription and acting as a nucleation complex for nucleocapsid assembly.

Genomic RNA folding mediates assembly of human parechovirus.

Assembly of the major viral pathogens of the Picornaviridae family is poorly understood. Human parechovirus 1 is an example of such viruses that contains 60 short regions of ordered RNA density making identical contacts with the protein shell. We show here via a combination of RNA-based systematic evolution of ligands by exponential enrichment, bioinformatics analysis and reverse genetics that these RNA segments are bound to the coat proteins in a sequence-specific manner. Disruption of either the RNA coat protein recognition motif or its contact amino acid residues is deleterious for viral assembly. The data are consistent with RNA packaging signals playing essential roles in virion assembly. Their binding sites on the coat proteins are evolutionarily conserved across the Parechovirus genus, suggesting that they represent potential broad-spectrum anti-viral targets.The mechanism underlying packaging of genomic RNA into viral particles is not well understood for human parechoviruses. Here the authors identify short RNA motifs in the parechovirus genome that bind capsid proteins, providing approximately 60 specific interactions for virion assembly.

Bacteriophage MS2 genomic RNA encodes an assembly instruction manual for its capsid.

Using RNA-coat protein crosslinking we have shown that the principal RNA recognition surface on the interior of infectious MS2 virions overlaps with the known peptides that bind the high affinity translational operator, TR, within the phage genome. The data also reveal the sequences of genomic fragments in contact with the coat protein shell. These show remarkable overlap with previous predictions based on the hypothesis that virion assembly is mediated by multiple sequences-specific contacts at RNA sites termed Packaging Signals (PSs). These PSs are variations on the TR stem-loop sequence and secondary structure. They act co-operatively to regulate the dominant assembly pathway and ensure cognate RNA encapsidation. In MS2, they also trigger conformational change in the dimeric capsomere creating the A/B quasi-conformer, 60 of which are needed to complete the T=3 capsid. This is the most compelling demonstration to date that this ssRNA virus, and by implications potentially very many of them, assemble via a PS-mediated assembly mechanism.

Direct Evidence for Packaging Signal-Mediated Assembly of Bacteriophage MS2.

Using cross-linking coupled to matrix-assisted laser desorption/ionization mass spectrometry and CLIP-Seq sequencing, we determined the peptide and oligonucleotide sequences at the interfaces between the capsid proteins and the genomic RNA of bacteriophage MS2. The results suggest that the same coat protein (CP)-RNA and maturation protein (MP)-RNA interfaces are used in every viral particle. The portions of the viral RNA in contact with CP subunits span the genome, consistent with a large number of discrete and similar contacts within each particle. Many of these sites match previous predictions of the locations of multiple, dispersed and degenerate RNA sites with cognate CP affinity termed packaging signals (PSs). Chemical RNA footprinting was used to compare the secondary structures of protein-free genomic fragments and the RNA in the virion. Some PSs are partially present in protein-free RNA but others would need to refold from their dominant solution conformations to form the contacts identified in the virion. The RNA-binding peptides within the MP map to two sections of the N-terminal half of the protein. Comparison of MP sequences from related phages suggests a similar arrangement of RNA-binding sites, although these N-terminal regions have only limited sequence conservation. In contrast, the sequences of the C-termini are highly conserved, consistent with them encompassing pilin-binding domains required for initial contact with host cells. These results provide independent and unambiguous support for the assembly of MS2 virions via a PS-mediated mechanism involving a series of induced-fit viral protein interactions with RNA.

An implementation of the Gillespie algorithm for RNA kinetics with logarithmic time update.

In this paper I outline a fast method called KFOLD for implementing the Gillepie algorithm to stochastically sample the folding kinetics of an RNA molecule at single base-pair resolution. In the same fashion as the KINFOLD algorithm, which also uses the Gillespie algorithm to predict folding kinetics, KFOLD stochastically chooses a new RNA secondary structure state that is accessible from the current state by a single base-pair addition/deletion following the Gillespie procedure. However, unlike KINFOLD, the KFOLD algorithm utilizes the fact that many of the base-pair addition/deletion reactions and their corresponding rates do not change between each step in the algorithm. This allows KFOLD to achieve a substantial speed-up in the time required to compute a prediction of the folding pathway and, for a fixed number of base-pair moves, performs logarithmically with sequence size. This increase in speed opens up the possibility of studying the kinetics of much longer RNA sequences at single base-pair resolution while also allowing for the RNA folding statistics of smaller RNA sequences to be computed much more quickly.

Asymmetric genome organization in an RNA virus revealed via graph-theoretical analysis of tomographic data.

Cryo-electron microscopy permits 3-D structures of viral pathogens to be determined in remarkable detail. In particular, the protein containers encapsulating viral genomes have been determined to high resolution using symmetry averaging techniques that exploit the icosahedral architecture seen in many viruses. By contrast, structure determination of asymmetric components remains a challenge, and novel analysis methods are required to reveal such features and characterize their functional roles during infection. Motivated by the important, cooperative roles of viral genomes in the assembly of single-stranded RNA viruses, we have developed a new analysis method that reveals the asymmetric structural organization of viral genomes in proximity to the capsid in such viruses. The method uses geometric constraints on genome organization, formulated based on knowledge of icosahedrally-averaged reconstructions and the roles of the RNA-capsid protein contacts, to analyse cryo-electron tomographic data. We apply this method to the low-resolution tomographic data of a model virus and infer the unique asymmetric organization of its genome in contact with the protein shell of the capsid. This opens unprecedented opportunities to analyse viral genomes, revealing conserved structural features and mechanisms that can be targeted in antiviral drug design.

Revealing the density of encoded functions in a viral RNA.

We present direct experimental evidence that assembly of a single-stranded RNA virus occurs via a packaging signal-mediated mechanism. We show that the sequences of coat protein recognition motifs within multiple, dispersed, putative RNA packaging signals, as well as their relative spacing within a genomic fragment, act collectively to influence the fidelity and yield of capsid self-assembly in vitro. These experiments confirm that the selective advantages for viral yield and encapsidation specificity, predicted from previous modeling of packaging signal-mediated assembly, are found in Nature. Regions of the genome that act as packaging signals also function in translational and transcriptional enhancement, as well as directly coding for the coat protein, highlighting the density of encoded functions within the viral RNA. Assembly and gene expression are therefore direct molecular competitors for different functional folds of the same RNA sequence. The strongest packaging signal in the test fragment, encodes a region of the coat protein that undergoes a conformational change upon contact with packaging signals. A similar phenomenon occurs in other RNA viruses for which packaging signals are known. These contacts hint at an even deeper density of encoded functions in viral RNA, which if confirmed, would have profound consequences for the evolution of this class of pathogens.