Resumen de A theoretical and computational study of the mechanics of biomembranes at multiple scales
Alejandro Torres Sánchez
Lipid membranes are thin objects that form the main separation structure in cells. They have remarkable mechanical properties; while behaving as a solid shell against bending, they exhibit in-plane fluidity. These two aspects of their mechanics are not only interesting from a physical viewpoint, but also fundamental for their biological function. Indeed, the equilibrium shapes of different organelles in the cell rely on the bending elasticity of lipid membranes. On the other hand, the in-plane fluidity of the membrane is essential in functions such as cell motility, mechano-adaptation, or for the lateral diffusion of proteins and other membrane inclusions. The bending rigidity of membranes can be motivated from microscopic models that account for the stress distribution across the membrane thickness. In particular, the microscopic stress across the membrane is routinely computed from molecular dynamics simulations to investigate how different microscopic features, such as the addition of anesthetics or cholesterol, affect their effective mechanical response. The microscopic stress bridges the gap between the statistical mechanics of a set of point particles, the atoms in a molecular dynamics simulation, and continuum mechanics models. However, we lack an unambiguous definition of the microscopic stress, and different definitions of the microscopic stress suggest different connections between molecular and continuum models. In the first Part of this Thesis, we show that many of the existing definitions of the microscopic stress do not satisfy the most basic balance laws of continuum mechanics, and thus are not physically meaningful. This striking issue has motivated us to propose a new definition of the microscopic stress that complies with these fundamental balance laws. Furthermore, we provide a freely available implementation of our stress definition that can be computed from molecular dynamics simulations (mdstress.org). Our definition of the stress along with our implementation provides a foundation for a meaningful analysis of molecular dynamics simulations from a continuum viewpoint. In addition to lipid membranes, we show the application of our methodology to other important systems, such as defective crystals or fibrous proteins. In the second part of the Thesis, we focus on the continuum modeling of lipid membranes. Because these membranes are continuously brought out-of-equilibrium by biological activity, it is important to go beyond curvature elasticity and describe the internal mechanisms associated with bilayer fluidity. We develop a three-dimensional and non-linear theory and a simulation methodology for the mechanics of lipid membranes, which have been lacking in the field. We base our approach on a general framework for the mechanics of dissipative systems, Onsager's variational principle, and on a careful formulation of the kinematics and balance principles for fluid surfaces. For the simulation of our models, we follow a finite element approach that, however, requires of unconventional dicretization methods due to the non-linear coupling between shape changes and tangent flows on fluid surfaces. Our formulation provides the basis for further investigations of the out-of-equilibrium chemo-mechanics of lipid membranes and other fluid surfaces, such as the cell cortex.
