Fast rotation of EITC-labelled proteolipid after reconstitution of chloroplast ATP synthase in bilayer membranes

Abstract

The chloroplast ATP synthase (CF 0CF 1) has been isolated and labelled with the triplet probe eosin isothiocyanate (EITC) when dissolved in detergent solution. On SDS-PAGE gels of the labelled enzyme, eosin fluorescence was observed almost exclusively in a single band with a molecular mass of about 8000 Da, subunit III of CF 0. At saturation, about 10 mol EITC/mol CF 0CF 1 were bound, which may indicate, that the complex contains 10 copies of subunit III. The CF 0CF 1 complex labelled in subunit III was incorporated into liposomes using the dialysis technique and the enzyme rotational diffusion in the liposome membrane was measured using laser-flash-induced absorption anisotropy of the protein-bound EITC. At 30°C and a lipid/protein ratio of 8 · 10 4:1 mol/mol uniaxial rotation with a correlation time of 200 ns at 30°C was observed. Considering the model of Saffman and Delbrück (Saffman, P.G. and Delbrück, M. (1975) Proc. Natl. Acad. Sci. USA 72, 3111–3113) and the viscosity of the azolectin membrane (Wagner et al. (1989) Eur. J. Biochem. 182, 165–173), this rotational correlation time corresponds to uniaxial rotation of a protein cylinder in the membrane with the radius of 1.1 nm. The rotating unit with this dimensions can easily be formed by two transmembrane helices. We therefore conclude that the proteolipid oligomer in CF 0, probably due to binding of EITC at lysine-48, dissociated into monomers. The other subunits of CF 0CF 1 apparently remained stable parts of the ATP synthase complex and associated with the liposome membrane. In a second attempt, CF 0CF 1 was labelled with EITC when bound to the thylakoid membrane. In this case, on SDS-PAGE gels of the labelled enzyme, eosin fluorescence was observed in a single band with a molecular mass of about 57 kDa, corresponding to the β subunit of CF 1. After incorporation of the β-labelled CF 0CF 1 into liposomes the rotational correlation time of the enzyme was 1.3–1.5 μs at 30°C. This rotation time is compatible with the uniaxial rotation of a ‘protein cylinder’ in the membrane with the diameter of 6.1 nm. Obviously, this rotating unit represents the entire CF 0CF 1 complex. The rotation rate of the reconstituted enzyme complex appeared to be determined mainly by the friction experienced by CF 0 within the membrane and therefore the observed ‘rotating unit’ represents almost exclusively CF 0. Addition of 2 μM venturicidin or 2 μM tributyltin, which are known to block proton flow through CF 0, to proteoliposomes containing the β-subunit-labelled CF 0CF 1 increased the rotational correlation time of the enzyme to about 2 μs. This result shows that the both inhibitors block the H + flow through CF 0CF 1 by changing the overall structure of the F 0 part. In contrast to this result, with proteoliposomes containing the subunit-III-labelled CF 0CF 1 no changes in the rotational time were observed after addition of venturicidin and tributyltin. This indicates that binding of either of the two inhibitors does not affect the conformation of the monomeric subunit III to any remarkable extent.