Insertases Scramble Lipids: Molecular Simulations of MTCH2

Scramblases play a pivotal role in facilitating bidirectional lipid transport across cell membranes, thereby influencing lipid metabolism, membrane homeostasis, and cellular signaling. MTCH2, a mitochondrial outer membrane protein insertase, has a membrane-spanning hydrophilic groove resembling those that form the lipid transit pathway in known scramblases. Employing both coarse-grained and atomistic molecular dynamics simulations, we show that MTCH2 significantly reduces the free energy barrier for lipid movement along the groove and therefore can indeed function as a scramblase. Notably, the scrambling rate of MTCH2 in silico is similar to that of VDAC, a recently discovered scramblase of the outer mitochondrial membrane, suggesting a potential complementary physiological role for these mitochondrial proteins. Finally, our findings suggest that other insertases which possess a hydrophilic path across the membrane like MTCH2, can also function as scramblases.

Table S1: Distribution of umbrella sampling windows along the collective variable with biasing force constants used for Martini 3 simulations of lipid flip-flop through membrane with MTCH2.Reference distances are in nm, force constants in kJ mol  S6: Distribution of umbrella sampling windows along the collective variable used for atomistic simulations of lipid flip-flop through membrane with MTCH2."Origin" corresponds to the pulling simulation from which the initial configuration for the given umbrella sampling window was taken.R marks windows where Hamiltonian replica exchange was applied.
Reference distances are in nm, force constants in kJ mol −1 nm −2 .9 Figure S3: Convergence of free energy calculations captured by calculating additional free energy profiles from the first (azure) and the second (lime) halves of the production phase of umbrella sampling."MTCH2 down" and "MTCH2 up" correspond to umbrella sampling simulations in which MTCH2 was present, and the initial configurations for the umbrella sampling windows were obtained from steered molecular dynamics, during which the selected lipid phosphate was pulled from the upper to the lower leaflet and from the lower to the upper leaflet, respectively."MTCH2 HREX" corresponds to simulation in which umbrella sampling was enhanced by Hamiltonian replica exchange.For more details, refer to the Methods section.Note that an additional estimate of the error can be obtained from the asymmetry of the free energy minima located on each side of the profile.
10 12 Figure S6: A) Comparison of free energy profiles calculated using umbrella sampling windows with initial configurations obtained from steered molecular dynamics simulations, where the selected lipid phosphate was either pulled from the upper to the lower leaflet (down pull, blue profile) or from the lower to the upper leaflet (up pull, red profile).The HREX, orange profile was calculated using umbrella sampling windows from both pulling directions (by alternating the origin of initial configurations) while employing Hamiltonian replica exchange.Note the hysteresis between the red and blue profiles, which prompted the use of Hamiltonian replica exchange.B-C) Representative simulation snapshots captured from Martini 2 umbrella sampling windows.B corresponds to the set of windows originating from the "down pull", while C corresponds to windows originating from the "up pull"."HREX" windows are not shown.The color scheme is the same as in Figure S5.The difference between the "down pull" and "up pull" profiles likely originates from the lipid using a different scrambling pathway in the C-terminal part of the protein when pulled from the lower to the upper leaflet (compare snapshots B5 and C5).Note that this alternative pathway is likely an artifact of pulling, as it was not observed in the windows with applied Hamiltonian replica exchange after initial equilibration.13 Figure S7: Representative simulation snapshots of a selected lipid translocating through a membrane containing MTCH2 captured from atomistic umbrella sampling windows.MTCH2 is depicted as a molecular surface colored according to the character of its residues (hydrophilic = green, hydrophobic = white, positively charged = blue, negatively charged = red).The pulled lipid is shown in orange, while other lipids are only partially represented with their phosphate groups illustrated as gray beads.Water and ions are not shown. 14

Figure S1 :Figure S2 :
Figure S1: Simulation snapshots obtained from unbiased Martini 3 (A), Martini 2 (B), and atomistic (C) simulations showing insertion of lipid headgroups into the hydrophilic cavity of MTCH2 and the disorder in the lipid tails.Only lipids in proximity to the cavity are shown.Lipid headgroups are depicted as orange beads while lipid tails as orange tubes.MTCH2 is represented as a molecular surface colored according to the character of its residues (hydrophilic = green, hydrophobic = white, positively charged = blue, negatively charged = red).

Figure S4 : 11 Figure S5 :
Figure S4: Representative simulation snapshots of a selected lipid translocating through a proteinless membrane captured from Martini 3 umbrella sampling windows.The snapshots are centered on the phosphate of the translocating lipid.The translocating lipid is shown in orange, while only phosphate groups are shown for the other lipids, represented by gray beads.Only water beads close to the membrane are shown, and these are depicted as blue beads.

Table S2 :
−1 nm −2 .Distribution of umbrella sampling windows along the collective variable with biasing force constants used for Martini 2 simulations of lipid flip-flip through membrane with MTCH2.Reference distances are in nm, force constants in kJ mol −1 nm −2 .

Table S3 :
Distribution of umbrella sampling windows along the collective variable with biasing force constants used for Martini 2 and 3 simulations of lipid flip-flop through proteinless membrane.Reference distances are in nm, force constants in kJ mol −1 nm −2 .

Table S4 :
Position and dihedral restraints applied during the individual stages of atomistic equilibration.All values are in kJ mol −1 nm −2 or kJ mol −1 rad −2 (for lipid dihedrals).

Table S5 :
Distribution of umbrella sampling windows along the collective variable with biasing force constants used for atomistic simulations of lipid flip-flop through proteinless membrane.R marks windows where Hamiltonian replica exchange was applied.Reference distances are in nm, force constants in kJ mol −1 nm −2 .