Breaking a Virus: Identifying Molecular Level Failure Modes of Viral Capsid Compression through Multi-Scale Simulation Techniques

Abstract

We use a systematically coarse-grained model for the protein capsid of Cowpea Chlorotic Mottle Virus (CCMV) to study its deformation under uniaxial compression, all the way from its initial elastic response to the capsid's ultimate structural failure. Our model amends the MARTINI force field with an iteratively refined elastic network, and we have previously shown that it reproduces the fluctuations of small fragments as well as the large-scale stress-strain response.We developed an automated identification method that classifies the contacting protein interfaces in the CCMV capsid into symmetry-classes and characterizes their structural changes upon deformation in residue-level detail. We observed that the symmetry-classes differ markedly in their stability, in a way that appears to backtrack the putative assembly pathway: interfaces that are believed to form last are most likely to break first. For instance, neither protein dimers (the first assembly step) nor pentamers of dimers (the second step) were ever seen to fail, while the hexameric association site (presumably the last to form) ruptures most readily. Interestingly, the wild type capsid fortifies this location with a cooperatively formed 6-stranded beta-barrel motif, which is missing in the mutant we employed in our studies. We hypothesize that interfacial binding strengths regulate the assembly order, but that later (and hence weaker) contacts may be reinforced by cooperative motifs that form post-assembly.