Molecular modeling and simulation of pH effects in lipid bilayers

Tese de doutoramento, Bioquímica (Bioquímica Teórica), Universidade de Lisboa, Faculdade de Ciências, 2015Biological membranes are the first barrier between cell and the extracellular environment, and also their first via of interaction. These biological structures are complex and diverse but share a common feature: they are all supported by a lipid bilayer. This bilayer is often negative and sensitive to pH and ionic strength. Computational / theoretical methods have been used to understand how these two properties inuence the bilayer structure. However, the methods available at the beginning of our work had two significant bottlenecks that we attempted to get rid of. First, none of the available Poisson{Boltzmann (PB) solvers were able to deal with both periodicity and pKa calculations. The first project was developed to overcome this technical limitation: we used an available PB solver in a new approach to perform pKa calculations taking into account the system periodicity. Secondly, to simulate highly charged membranes, a proper treatment of the ionic strength is crucial and a full neutralization of the system is probably too rough an approximation. Hence, we developed a PB-based method to determine the number of ions that should be added to the simulations. With these two problems solved, we developed a new constant-pH molecular dynamics method to deal with charged lipid bilayers (CpHMD-L). This method allows a proper treatment of the periodicity and ionic strength in model membranes. Finally, we applied our methods to three model systems to illustrate the importance of taking into account protonation / conformation coupling in molecular dynamics simulations, in particular when looking at pH dependent phenomena. To the best of our knowledge, this is the only available method that can deal simultaneously with pH considering the protonation / conformation coupling, periodic boundary conditions in protonation free energy calculations, and a careful treatment of ionic strength. This represents a significant improvement in simulations of model biological membranes. It is now possible to move a step forward in the direction of biological membranes since we are now able to simulate a lipid mixture with several different lipids