TMEM16F is a Ca 2+ -activated ion channel and lipid scramblase. Lipid scrambling dissipates the plasma membrane asymmetry and activates multiple signaling pathways related to processes such as blood coagulation and membrane fusion. Previously published structures of the fungal and ER scramblases revealed that Ca 2+ binding favors a conformational transition from a closed to an open membrane-exposed state. In contrast, structural studies of TMEM16F in detergent or nanodiscs did not capture a similar transition. Recently, we used cryoEM to image TMEM16F in liposomes and found it adopts a novel conformation where the groove forms a protein-delimited ion permeation pore and that thins the membrane on its exterior to enable lipid scrambling. Here, we sought to elucidate the energetic landscape underlying TMEM16F gating transitions using a combination of unbiased and biased MD simulations. First, we carried out ensembles of unbiased MD simulations starting from previously identified states of TMEM16F (the experimentally determined Ca 2+ -bound closed and active as well as the computationally predicted Ca 2+ bound open). We used these trajectories as training data for a machine-learning (ML) pipeline to identify collective variables (CV) that describe the observed gating transitions. We used these CVs in metadynamics simulations with multiple walkers to efficiently sample TMEM16F’s conformational space. We found that as expected the Ca 2+ -bound closed state is the most stable state in landscape. Then, to investigate how mutations perturb the free-energy gating landscape of TMEM16F, we performed comparative metadynamics simulations using established mutations identified from the literature and novel mutants based on exploratory unbiased MD simulations. From these we found several mutations that modify the gating energetics of activation.
Alvarenga-Cruz et al. (Sun,) studied this question.