Modelling the Self-Assembly of Virus Capsids

We use computer simulations to study a model, first proposed by Wales [1], for the reversible and monodisperse self-assembly of simple icosahedral virus capsid structures. The success and efficiency of assembly as a function of thermodynamic and geometric factors can be qualitatively related to the potential energy landscape structure of the assembling system. Even though the model is strongly coarse-grained, it exhibits a number of features also observed in experiments, such as sigmoidal assembly dynamics, hysteresis in capsid formation and numerous kinetic traps. We also investigate the effect of macromolecular crowding on the assembly dynamics. Crowding agents generally reduce capsid yields at optimal conditions for non-crowded assembly, but may increase yields for parameter regimes away from the optimum. Finally, we generalize the model to a larger triangulation number T = 3, and observe more complex assembly dynamics than that seen for the original T = 1 model.

💡 Research Summary

This paper presents a computational study of a highly simplified, coarse‑grained model for the reversible, monodisperse self‑assembly of icosahedral virus capsids, originally proposed by Wales. The authors employ a combination of Monte Carlo and molecular dynamics simulations to explore how thermodynamic parameters (temperature) and geometric factors (bond strength) influence the efficiency and pathway of capsid formation. By scanning a broad range of temperatures and interaction strengths, they identify an optimal “assembly window” where the process follows a sigmoidal time course: an initial nucleation phase, a rapid growth phase, and a final saturation phase. Within this window the system reliably reaches the global energy minimum corresponding to a complete capsid.

A key insight is the relationship between the observed assembly behavior and the underlying potential energy landscape (PEL). The authors use basin‑hopping searches to map the landscape, revealing a funnel‑like topology that guides the system toward the capsid minimum when the parameters are well‑chosen. Outside the optimal window, the landscape is riddled with local minima that act as kinetic traps: at low temperatures the system becomes stuck in partially assembled intermediates, while at high temperatures frequent bond breakage prevents stable growth. The authors also demonstrate hysteresis: once a capsid has formed, raising the temperature does not immediately dissolve it, indicating that the assembled structure resides in a metastable basin separated by a sizable energy barrier.

The study further investigates macromolecular crowding by adding inert spherical “crowders” at varying volume fractions. Crowding generally reduces capsid yields under optimal assembly conditions because it limits the free volume available for subunit diffusion and collision. However, in parameter regimes far from the optimum (e.g., low bond strength or sub‑optimal temperature), crowding can enhance yields by increasing effective collision frequencies, thereby compensating for the weaker driving forces. This dual effect mirrors experimental observations in cellular environments, where crowding can both hinder and facilitate viral assembly depending on the specific biochemical context.

Beyond the original T = 1 model (60 subunits forming a single icosahedron), the authors extend the framework to a larger T = 3 capsid (180 subunits). The T = 3 system exhibits markedly more complex dynamics: multiple intermediate oligomers (e.g., 30‑mers, 45‑mers) appear, the assembly pathway branches, and the overall time to completion is significantly longer. The PEL for T = 3 contains many more local minima and higher energy barriers, reflecting the increased combinatorial possibilities of arranging a larger number of subunits. Consequently, the optimal temperature‑bond strength window narrows, and the system is more susceptible to kinetic traps.

Overall, the paper demonstrates that even a strongly coarse‑grained model can capture essential features of viral capsid assembly observed experimentally: sigmoidal kinetics, hysteresis, kinetic traps, and the nuanced role of crowding. By linking these phenomena to the shape of the underlying energy landscape, the authors provide a conceptual framework that can guide the design of antiviral strategies (e.g., small molecules that raise energy barriers) and the engineering of synthetic nanocontainers. The work also highlights the importance of systematic parameter scanning and landscape analysis for predicting assembly outcomes in more realistic, heterogeneous environments.