Skip to main content
. 2024 Apr 4;32(4):505–510.e4. doi: 10.1016/j.str.2024.01.012

Figure 2.

Figure 2

Comparison of Martini 3 and Martini 2 simulations

(A) Scrambling rate of MTCH2 in a POPC membrane shown as the number of “scrambled lipids” in time (running average over 200 ns). Each line corresponds to single 10 μs long Martini simulation. Lipid was considered to be scrambled when it was located in the opposite leaflet to its original one. Scrambling rate of lipids in a proteinless POPC membrane is zero in both models (on the simulated timescale).

(B) Density of phosphate beads around MTCH2 depicted as orange isosurface wireframe. Lipid phosphates readily insert into the cavity of MTCH2 in both models. In Martini 3, they are also able to easily enter the C-terminal pathway, while this is not the case for Martini 2. The position of the “hydrophobic gate” is indicated by the red arrow.

(C) Density of water beads around MTCH2 depicted as blue isosurface wireframe. In both models, a large number of molecules of water inserts into the cavity of MTCH2.

(D) Free energy profiles of lipid flip-flop in a proteinless POPC membrane (black) and along the MTCH2 scrambling pathway (red, blue, and purple). In both models, the presence of MTCH2 dramatically reduces the flip-flop free energy barrier. The red arrow in the Martini 2 chart indicates a peak in the free energy profile corresponding to the “hydrophobic gate.” The calculation error is <1 kJ mol−1. See Figure S3 for detailed view of free energy calculations convergence.