On the use of constraints in molecular dynamics simulations of shocked membrane systems

The plasma membrane of a cell exerts tight control over the intracellular milieu through an extensive network of transport pathways. Although these mechanisms nominally protect the cell, they also present a formidable barrier to intracellular drug accessibility. Transient perforation induced by compression waves has shown promise as a low-toxicity alternative to current-best chemical methods for enhancing drug transport. Differences in membrane composition and mechanical properties between healthy and cancerous cells result in a diverse response to compression waves that provides an opportunity for physics-based engineering: Can we exploit these differences to deliver a shock with tailored properties to form pores of a specific size?

Molecular dynamics simulations of shock-induced poration have helped to probe the (nm and ps) spatiotemporal domain associated with the underlying physics. To improve computational efficiency, constraints that fix bond lengths or angles to a nominal value are often imposed. Although algorithmic removal of these degrees of freedom proxies the quantum “freezing out” behavior, the validity of this approximation has not been assessed in the context of the localized pressure and temperature spikes in compression waves. First, we investigate the influence of constraints on the MSST-derived Hugoniot as well as equilibrium properties of TIP3P water. Comparison with data and an experimentally derived EOS highlight deficiencies in the water model and classical MD descriptions. We then comment on this in the context of solvated membrane systems.