Rotation-Direction-Dependent Mechanism of the Inhibitor Protein IF1 for Mitochondrial ATP Synthase from Atomistic Simulations

Abstract

ATPase inhibitory factor 1 (IF₁) is an endogenous regulatory protein for mitochondrial F_oF₁-ATP synthase. It blocks the catalysis and rotation of the F₁ part by deeply inserting itself into the rotor-stator interface. Recent single-molecule manipulation experiments have elucidated that forcible rotations only in the ATP-synthesis direction eject IF₁, rescuing F₁ from the IF₁-inhibited state. However, the molecular mechanism of the rotation-direction-dependent process at an atomic resolution is still elusive. Here, we have performed all-atom molecular dynamics (MD) simulations of the IF₁-bound F₁ structure with a torque applied to the rotor γ subunit. In the torque-applying simulations, we first found that the core part of the γ subunit rotated more in response to an external torque in the synthesis direction than in the hydrolysis direction. Further rotations of the γ subunit up to 120° revealed that the conformational change of the IF₁-bound αβ was only allowed in the synthesis direction. Also, the 120° rotation in the synthesis direction disrupted its contacts with IF₁, destabilizing the short helix of IF₁. After additional rotation up to the synthetic 240° state, the closed-to-open conformational change of the IF₁-bound β subunit pulled IF₁ outwardly, deforming the long helix of IF₁. These stepwise destabilizations of the IF₁ helices should be crucial for IF₁ ejection. Our simulations also provided insight into the nullification mechanism of the hydrolytic rotation, highlighting the steric clash between F22 of IF₁ and the β_TP subunit. Finally, we discuss a sufficient proton motive force to rescue F_oF₁-ATP synthase from the IF₁-inhibited state.

Keywords: ATP synthase; ATPase inhibitory factor 1; F1-ATPase; molecular dynamics simulation.

1

Characterization of IF₁-inhibited structures. (A) The top view of the F₁–IF₁ structure (PDB: 2v7q). The α, β, γ, and IF₁ are shown in orange, blue, red, and green, respectively. The δ and ε subunits are omitted in this figure. (B) Principal Component Analysis (PCA) of αβ pairs. The 78 αβ pairs from 26 bMF₁ structures were projected on PC1 and PC2. The nucleotide state of each αβ pair, Empty, TP, and DP, are shown in red, blue, and green, respectively. The values in the parentheses of the axis labels refer to the 1st and 2nd eigenvalue contributions. The units of each axis are in nanometers (nm). (C) Structural comparison of the αβ_DP in the IF₁-inhibited structure (PDB: 2v7q, orange, blue, and green) with the ground-state structure (PDB: 2jdi, gray/transparent). Arrows represent the interface motion of IF₁-bound αβ_DP compared to the catalysis-waiting state. (D) The rotary angle of the γ subunit in 26 bMF₁ structures, where that of the 1bmf structure was defined as 0 degree. The orange points represent the IF₁-bound structures.

2

40° torque simulation with various force constants. (A) Overview of the torque-applying simulation. The γ subunit was rotated in hydrolysis (counterclockwise; CCW) or synthesis (clockwise; CW) direction at ω = 1°/ns. (B) The structure of the γ subunit. The core part (residues 1–26 and 228–272) is colored purple, and the protruded part (residues 27–227) is colored red. (C) The γ subunit rotation of the core part and the protruded part. The protruded and core angles show linear and jiggling behaviors, respectively. The core angles are shown in gradation concerning different force constants. (D) The final core angles from the 40 ns simulations. The mean values and SD error bars are calculated from three independent simulations. (E) Work profiles over the simulation time. Blue and orange lines represent the simulation with IF₁, and cyan and pink lines represent IF₁-free simulations. (F) The final work from the simulations. The mean values and SD error bars are calculated from three independent simulations.

3

The 120° torque simulation. (A) Conformational change of the IF₁-bound αβ pair upon the CW (left) and CCW (right) rotation, projected onto the PC1-PC2 plane. The gray dots correspond to the X-ray crystal structures (Figure B). (B) The relative amount of accomplished conformational change of the IF₁-bound αβ₁ pair upon the γ rotation quantified from the PC2 value. Different shades of color represent five independent trajectories. (C) A typical example of the short helix (residues 14–18) and the long helix (residues 21–45) formation during the CW 120° simulation. Helix formation at time 0 was calculated from the simulation result before the torque-applying simulation. Inset is the initial structure of IF₁. (D) Average helix formation in the final 10 ns (110–120 ns) of the CW simulation. The dotted lines represent the value calculated from the simulation result before the torque-applying simulation. (E) Time-dependent changes in contacts between the short helix and the γ subunit during the CW simulation. The relative contact values were calculated by setting the contact number from the equilibrium simulation as 100%. The data points at 30, 60, 90, and 120 ns were calculated from the number of contacts in the preceding 30 ns interval. (F) The snapshots at 0 ns (left) and 120 ns (right) during the CW rotation. The bright green region represents the short helix of IF₁, while the rest is shown as transparent. Representative residues are shown in both figures, γR8, γS12, and γN15. γM242 from the C-terminus is shown in the right figure, which approaches the short helix at the end of some simulations. (G) The time-dependent contact changes of the specific residue pairs between the short helix and the N-terminus of γ subunit; IF₁(A14)-γ­(S12), IF₁(V15)-γ­(S12), IF₁(V15)-γ­(I16). Dark blue represents a high contact ratio, while light blue represents a low one. The same simulation trajectory was used to illustrate (C,F,G).

4

Deformation of the long helix at CW 240°. (A) Overview of the CW 240° state, where the bound nucleotides are also shown. (B) Snapshots from the targeted MD at 0 ns (Left) and 49 ns (Right). The inset in the left lower corner in each panel is IF₁, where residues 21–50 are shown. The inset in the right lower corner in each panel is the enlarged view of the R408 of IF₁-bound β (β₁) and E30 of IF₁. (C) A typical example of the long helix (residues 21–45) deformation. The long helix was divided into three parts: residues 21–30 (pink), 31–37 (blue), and 38–45 (green). Helix formation at time 0 was estimated from the simulation result before the targeted MD. (D) The bending angle of the long helix during the targeted MD, which was defined by the angle formed by the Cα atoms of residues 21, 38, and 45. (E) The minimum distance between E30 of IF₁ and R408 of IF₁-bound β (β₁). Four independent trajectories are shown in different colors, while the gray line represents the result of the equilibrium run before the targeted MD in (D,E).

5

The 120° torque simulation in CCW direction. (A) Average helix formation in the final 10 ns (110–120 ns) of the CCW simulation. The dotted lines represent the equilibrium value calculated from the simulation before applying torque. (B) Time-dependent contact changes between the short helix and the γ subunit during the CCW simulation. See also Figure D,E for reference of the CW rotation. (C) Snapshots focusing on the short helix of IF₁ and the γ subunit. (D) The side-chain dihedral angle χ₁ of F22 of IF₁. The blue and orange lines represent the 120° forcible simulation in CCW and CW, respectively. (E) The ratio of χ₁ = 200° state and the χ₁ = 300° state in CCW and CW are shown in green and purple, respectively. The dihedral angle that was not assigned to either χ₁ = 200° or χ₁ = 300° is shown in gray. (F) Snapshots from the CCW simulation. The left snapshot represents the χ₁ = 200° state, where the phenyl group of F22 in IF₁ was oriented toward the I16 of the γ subunit. The right snapshot represents the χ₁ = 300° state, where the phenyl group of F22 in IF₁ was oriented toward the S383 of the β_TP subunit. Arrows represent the side-chain directions of F22.