Computationally Projecting the Influence of Nucleic Acid on Pathways of Nucleation-Limited Virus Capsid Assembly

Elucidating assembly pathways of complex macromolecular structures, such as virus capsids, is an important problem for understanding the many cellular processes dependent on self-assembly but also challenging given limited experimental technologies for observing such systems. We have previously addressed this problem through simulation-based data fitting, learning rate parameters of coarse-grained stochastic simulation models to match light scattering data from bulk assembly of purified coat protein in vitro providing an unprecedented view of the fine-scale reaction pathways that might have produced those data. A key question raised by such models, though, is how well they might reflect assembly under more natural cellular conditions where factors such as local concentration changes, non-specific crowding, and often the influence of nucleic acid during assembly become relevant. In the present study, we examine the latter issue, how using analytical models of various contributions of RNA folding to assembly would influence overall pathways and kinetics, primarily with reference to cowpea chlorotic mottle virus (CCMV). We find a surprising complexity and synergy of interaction effects. Energetic effects that gain or lower free energy tend to disrupt successful assembly relative to the in vitro model individually, while the full combination of positive and negative effects collectively promotes greatly accelerated assembly without loss of yield. Furthermore, it accomplishes this change in kinetics while substantially altering the ensemble of assembly pathways open to the system. These simulation results help us understand how RNA viral coat and genome may interact in assembly to promote rapid growth while avoiding kinetic traps expected from much prior theory, bringing us a step closer to the goal of understanding how viral assembly in the cell may differ from our current conception based largely on in vitro models.