Discovery of an overarching design principle for icosahedral viral architecture solves open problems in structural virology
Article accepted for publication by Nature Communications
This paper places the seminal Caspar Klug theory into the wider context of Archimedean lattices, and thus explains virus architectures that have been structural puzzles for decades.
In Caspar Klug theory, the numbers of proteins in viral capsids are quantised. However, viruses with "forbidden" protein numbers have been observed, such as capsids formed from 120 proteins. The overaching theory explains their structures.
It has previously not been clear how larger and more complex viruses might have evolved from smaller viruses, given the large size gaps in Caspar Klug theory. Our theory reveals architectures of intermediate size that could serve as intermediate steps. It also suggests a mechanism: additional protein domains could have evolved occupying the triangular lattice positions, then have become independent protin units (minor capsid proteins) in larger viruses.
The new structural models provide a basis to study the biophysical properties of viruses following different lattice geometries. For example, percolation theory can be used to study the stability of viral capsids dependent on lattice type.
Discovery of RNA encoded virus assembly instructions
Mathematical concepts called Hamiltonian Paths have been instrumental in the discovery and characterisation of a viral assembly mechanism that has changed the existing paradigm.
Discovery of a molecular scaling principle
Affine extended symmetry groups encode constraints on the 3D architecture of viruses. These constraint sets have been used in the Hamiltonian Path Approach (see above)
Graph Theory in vaccination
An adaptation of Viral Tiling theory, see below, has enabled characterisation of self-assembling protein nanoparticles (SAPNs) with applications in the design of malaria vaccines.
Press release
Viral Tiling theory
The first mathematical theory correctly describing the surface architectures of the cancer-causing polyoma- and papillomaviridae, that fall out of the scope of the seminal Caspar Klug theory.
Packaging Signal Mediated Virus Assembly
The Hamiltonian path concept has been instrumental in analysing the outcomes of stochastic simulations of virus assembly, revealing the mechanism by which packaging signals solve a viral equivalent of Levinthal's Paradox, achieving efficent and selective packaging against a backdrop of cellular competitor RNAs.
Structural capsid transitions important for infection
Many viruses must undergo structural rearrangements of their capsids in order to become infectious. Such maturation events can be modelled as transitions between different geometric models. Recent work models the common cold virus.
Novel data analysis strategies
In collaboration with the Stockley lab at the University of Leeds, we have developed novel interdisciplinary strategies to characterise the packaging signal contacts with capsid protein to molecular detail (shown for Human Parechovirus on the left). These include analysis protocols for SELEX data, and most recently of X-ray footprints of viral RNAs in the coat protein free state and inside the virion (paper for Nature Methods in preparation).
I have also developed a dynamic network analysis method in collaboration with Polly Roy at the London School of Hygiene and Tropical Medecine in order to characterise RNA-RNA interactions in multi-segmented viruses.
Modelling of viral infection dynamics
Insights into virus assembly enable the modelling of viral infection dynamics inside an infected cell at unprecedented detail. Integration of such models into multi-scale models of a viral infection allows the impact of different anti-viral strtegies to be assessed.
Capsid disassembly via percolation theory
The classification of capsid geometries provides a basis for the analysis of capsid stabilty via percolation theory. Surface lattices formed from different capsid protein units (capsomers) differ in their stability, impacting on the mechanism of genome release.
RNA-Mediated Virus Assembly: Mechanisms and
Consequences for Viral Evolution and Therapy
Reidun Twarock and Peter Stockley
A modelling paradigm for RNA virus assembly
Reidun Twarock, Richard J Bingham, Eric C Dykeman and Peter G Stockley
Mathematical Modelling of Virus Architecture
Reidun Twarock
Geometry as a Weapon in the Fight Against Viruses
Reidun Twarock
1. New avenues in anti-viral drug design
The discovery of an RNA-encoded assembly instruction manual that is mediated by signals in the genomic viral RNA has opened up novel drug targets. We developed efficient tools for the identification of these signals and their binding sites on the capsid protein, and characterised the molecular details of these interactions. This enabled us to identify new drug targets, and we hold two patents (“Anti-viral Therapy” PCT/ GB2014/052696; “Viral Packaging Signals” GB1618094.5) for novel anti-viral strategies against these viruses. The spectrum of viruses covered by these patents includes Hepatitis B (HBV), a retro-transcribing DNA virus that packages its genome initially as a pre-genomic RNA. Using a bespoke high-throughput screen against >22,000 potential RNA binding ligands, in collaboration with the US NIH at Frederick, we identified >60 specific “hits” to the generic HBV genomic sites contacting the viral protein shell. These small molecular weight compounds were then screened for toxicity and assembly inhibition in collaboration with Professor Marcus Dorner, Imperial College, London. This identified 9 compounds with significant initial anti-viral activity, and we are now assaying their effects against virus in collaboration with colleagues at the NIH, who have also filed a patent, led by Dr Stuart LeGrice, NIH (“Methods of treating HBV”, U.S. Patent Application #62/685,145, 2018), that includes my experimental collaborator Peter G Stockley and myself as inventors. We have similar “hits” from the screen using the parechovirus genome targets. We are in discussions with industrial partners to exploit our invention commercially for HBV, and have plans to follow similar routes for other viruses.
2. New opportunities in synthetic vaccines and gene therapy
The mechanistic understanding of how packaging signals contribute to efficient formation of viral particles provides a basis to repurpose viral components in the design of virus-like particles (VLPs). We have demonstrated that the essential features of the RNA-encoded assembly manual can be used to enhance the efficiency of assembly around non-viral RNAs (see our patent “Virus-Like Particles” GB1708709.9), and we are in the process of extending this idea to major pathogens for which there are currently no vaccines, but for which novel forms of intervention are urgently required.
Virus-like particles are also of interest as transport vehicles in gene therapy. I am currently developing algorithms that support the development of technologies that render the assembly of VLPs vastly more efficient. Overcoming the assembly efficiency bottleneck is essential in order to make sufficient VLPs for successful treatment in patients.
Patents:
“Anti-viral Therapy” (4363P/US and 4363P/EP)
“Viral Packaging Signals” (GB1618094.5)
“Virus-Like Particles” (GB1708709)
“Methods of treating HBV” (U.S. Patent Application No. 62/685,145 filed June 14, 2018) – joint patent filing with the NIH in Washington