Lipid-mediated Protein-protein Interactions Modulate Respiration-driven ATP Synthesis

Abstract

Energy conversion in biological systems is underpinned by membrane-bound proton transporters that generate and maintain a proton electrochemical gradient across the membrane which used, e.g. for generation of ATP by the ATP synthase. Here, we have co-reconstituted the proton pump cytochrome bo3 (ubiquinol oxidase) together with ATP synthase in liposomes and studied the effect of changing the lipid composition on the ATP synthesis activity driven by proton pumping. We found that for 100 nm liposomes, containing 5 of each proteins, the ATP synthesis rates decreased significantly with increasing fractions of DOPA, DOPE, DOPG or cardiolipin added to liposomes made of DOPC; with e.g. 5% DOPG, we observed an almost 50% decrease in the ATP synthesis rate. However, upon increasing the average distance between the proton pumps and ATP synthases, the ATP synthesis rate dropped and the lipid dependence of this activity vanished. The data indicate that protons are transferred along the membrane, between cytochrome bo3 and the ATP synthase, but only at sufficiently high protein densities. We also argue that the local protein density may be modulated by lipid-dependent changes in interactions between the two proteins complexes, which points to a mechanism by which the cell may regulate the overall activity of the respiratory chain.

Figure 1. Coupled bo₃-ATP synthase activity.

(A) Changes in luminescence of the luciferin-luciferase couple as a result of ATP synthesis driven by an electrochemical gradient generated by the bo₃ oxidase. The reaction was started at t = 0 upon addition of ubiquinol UQ₁ (20 μM after mixing) to a tube containing liposomes with co-reconstituted ATP synthase and bo₃ oxidase and DTT (2 mM). The reaction was stopped by addition of potassium cyanide (500 μM, at t ≅ 100 s), which inhibits the bo₃ oxidase. At t = −50 s, a small amount (5 pmol) of ATP was added to normalize the signals. The traces shown are for 100 nm liposomes with DOPC fractions of (the remaining part is DOPG): 100, 90, 80, 70, 60% from the top to the bottom. The traces are divided in 30-s segments for technical reasons. Measurements were performed at ~22 °C in 20 mM Tris-PO₄ at pH 7.5, 2.5 mM MgSO₄, 80 μM ADP. (B) The ATP synthesis rates (slopes of the traces in A) measured (n is the number of measurements) with increasing amounts of DOPG (black circles, n = 9–12), DOPA (black squares, n = 5–7), DOPE (white diamonds, n = 2 (points at 5% and 15%), 7 (remaining points)) or cardiolipin (white triangles, n = 4 (point at 5%), 12 (remaining points)). The relative ATP synthesis rates are shown as a fraction of the activity in 100 nm liposomes with 100% DOPC (~90 ATP × s⁻¹ × enzyme⁻¹ for dialysis reconstitution).

Figure 2. Individual activities of bo₃ oxidase and ATP synthase.

(A) Changes in the luminescence as a result of ATP synthesis. A transmembrane pH gradient was established as described in the Materials and Methods section (100 nm liposomes with ATP synthase, 60–100% DOPC (with DOPG). “a.u.” is “arbitrary units”. The inset shows the total amounts of ATP produced within the first 20 s (normalized to 100% DOPC, 145 pmol). Error bars: 4 measurements with two different samples. (B) Changes in oxygen concentration during turnover of the bo₃ oxidase in 100 nm liposomes. The reaction was initiated at t = 0 by addition of ubiquinol UQ₁ (20 μM) in the presence of 2 mM DTT (with valinomycin (10 μM) and FCCP (20 μM)) and stopped by addition of KCN (0.32 mM). The trace shown is for DOPC:DOPG = 80:20% (see panel D for data with other ratios). The rate was determined from the slope indicated as v. (C) Changes in fluorescence of (2.6 μM) ACMA (100% DOPC 100 nm liposomes with ATP synthase). Upon addition of 2.5 mM ATP (the rapid change is a mixing artifact) at t = 0 the fluorescence decreased with a slope, v₁, which reflects the rate of proton pumping into the liposomes. Upon addition of 130 nM valinomycin, the acidification rate increased (v₂). A “respiratory-control ratio” was calculated as v₂/v₁. (D) Summary of the data presented in panels (A–C). Within each data series, every point is normalized to that obtained with 100% DOPC (the remainder is DOPG). The (black squares) is the coupled bo₃-ATP synthase activity shown for reference (taken from Fig. 1B). (black triangles) Initial slope (~3 s) of the luminescence change for each of the traces in panel (A). This slope reflects the maximal ATP synthesis, i.e. the activity of the ATP synthase (2 measurements). (white circles) Oxygen consumption activities of the bo₃ oxidase (c.f. panel B) in the presence of uncouplers (3 measurements). (white diamonds) The ratio v₂/v₁ from panel C) (3 measurements).

Figure 3. ATP synthesis and O₂ consumption as a function of the fraction of DOPC for two different UQ₁ concentrations.

The remainder of the lipids is DOPG. The rates were measured with co-reconstituted bo₃-ATP synthase in 100 nm liposomes for 2 μM and 20 μM UQ₁, respectively (in the presence of 2 mM DTT). The activities were normalized to those obtained for 100% DOPC. Under these conditions, the ATP synthesis rates measured with 2 μM UQ₁ were a factor of 3 smaller than those measured with 20 μM UQ₁. The corresponding difference in the O₂-consumption rates was a factor of 10.

Figure 4. Coupled bo₃-ATP synthase activity as a function of the protein density and vesicle size.

(A) The coupled activity as a function of lipid composition for vesicles obtained with 100 nm (filled diamonds) and 200 nm (filled triangles) pore-size filters. The data is normalized to 100 nm vesicles containing 100% DOPC. The average number of proteins per vesicle was 5 cyt. bo₃ and 5 ATP synthase. For the 100 nm vesicles, DOPC was mixed only with DOPG, while for the 200-nm vesicles DOPC was mixed with both DOPG (filled triangles up) and cardiolipin (filled triangles down). (B) The coupled activity as a function of lipid composition for vesicles obtained with 200 nm pore-size filters and with different number of proteins (as indicated) keeping the 1:1 ratio of bo₃:ATP synthase. Data is normalized to 100% DOPC vesicles with 21 of each enzyme, which yields approximately the same lipid/enzyme ratio as for 5 of each enzyme in 100 nm vesicles (~2000 lipids/enzyme). The number of measurements was 2–4.

Figure 5. Schematic model illustrating the conclusions.

See text for explanation.