Bcl-2 family protein including anti-apoptotic (Bcl-2) or pro-apoptotic (Bax) members can form ion channels when incorporated into synthetic lipid bilayers. This contrasts with the observation that Bcl-2 stabilizes the mitochondrial membrane barrier function and inhibits the permeability transition pore complex (PTPC). Here we provide experimental data which may explain this apparent paradox. Bax and adenine nucleotide translocator (ANT), the most abundant inner mitochondrial membrane protein, can interact in arti®cial lipid bilayers to yield an e cient composite channel whose electrophysiological properties di er quantitatively and qualitatively from the channels formed by Bax or ANT alone. The formation of this composite channel can be observed in conditions in which Bax protein alone has no detectable channel activity. Cooperative channel formation by Bax and ANT is stimulated by the ANT ligand atractyloside (Atr) but inhibited by ATP, indicating that it depends on the conformation of ANT. In contrast to the combination of Bax and ANT, ANT does not form active channels when incorporated into membranes with Bcl-2. Rather, ANT and Bcl-2 exhibit mutual inhibition of channel formation. Bcl-2 prevents channel formation by Atr-treated ANT and neutralizes the cooperation between Bax and ANT. Our data are compatible with a meÂnage aÁ trois model of mitochondrial apoptosis regulation in which ANT, the likely pore forming protein within the PTPC, interacts with Bax or Bcl-2 which in¯uence its pore forming potential in opposing manners.

The ionophore properties of magainin I, an antimicrobial and amphipathic peptide from the skin of Xenopus, were investigated in planar lipid bilayers. Circular dichroism studies, performed comparatively with alamethicin, in small or large unilamellar phospholipidic vesicles, point to a smaller proportion of alpha-helical conformation in membranes. A weakly voltage-dependent macroscopic conductance which is anion-selective is developed when using large aqueous peptide concentration with lipid bilayer under high voltages. Single-channel experiments revealed two main conductance levels occurring independently in separate trials. Pre-aggregates lying on the membrane surface at rest and drawn into the bilayer upon voltage application are assumed to account for this behaviour contrasting with the classical multistates displayed by alamethicin.

In spite of the fact that planar lipid bilayers are still the best-suited artificial membrane system for the study of reconstituted ion channels and receptors, data dealing with their physical characterization, especially as regards dynamics, are scanty. A combined electrical and optical chamber was designed and allowed fluorescence recovery after photobleaching recovery curves to be recorded from stable virtually solvent-free bilayers. D, the lateral diffusion coefficient of N-(7-nitrobenzoyl-2-oxa-1,3-diazol-4-yl)-1,2-dihexadecanoyl-sn- glycero-3-phosphoethanolamine, was found to be relatively insensitive to the phospholipid composition (headgroup, chain unsaturation, etc.), whereas inclusion of 33-50% cholesterol in the membrane reduced D by a factor of 2. Divalent cations significantly reduced D of negatively charged bilayers. These results compare well with data gathered on other model and natural systems. In addition, the incorporation of the voltage-dependent pore-former alamethicin did slightly reduce lipid lateral mobility. This study demonstrates the feasibility of such experiments with planar bilayers, which are amenable to physical constraints, and thus offers new opportunities for systematic studies of structure-function relationships in membrane-associating molecules.

No abstract