We are in an exciting era when new structural information for membrane proteins is being collected at an increasingly rapid pace. A major challenge to the computational biophysics community is to develop novel techniques that can utilize these structural information to provide new insights about the function of membrane proteins. In additional to fundamental impacts, meeting this challenge has great biomedical relevance considering that 30% of all genes encode membrane proteins and 60% of approved drug targets are membrane proteins.
Many membrane proteins either function as a sensor to changes in the environment or their normal function is gated by the environmental condition. The environmental stimulation can be either physical (e.g., membrane tension or temperature) or chemical (e.g., ligand binding or pH variation) in nature. Therefore, to achieve a deeper understanding in the working mechanism of membrane proteins, it is important to characterize their structural response to external stimulation, which is often difficult to do at sufficient spatial and temporal resolution using experiments alone. Despite their increasing importance, atomistic studies are often limited in this context by the short time-scale accessible in typical simulations. Therefore, this is an important area where the development of simplified but effective computational models is sorely needed. In the immediate future, the external stimulation we will focus on is mechanical perturbation in the membrane. However, the computational framework will be developed with our long-term goal borne in mind, which is to apply such models to investigate the structural response of membrane proteins to diverse environmental stimulations such as variations in pH and temperature. Since both membrane and protein are treated explicitly, the computational framework is also applicable to membrane remodeling (e.g., bending, fusion and tubulation), another important class of coupled membrane/protein process that implicates multiple scales and therefore will benefit tremendously from novel computational approaches.