Although nobody doubts that the release of cytochrome c from mitochondria is a crucial step in cell death, quite how this happens is a source of controversy. Some say that the pro-apoptotic molecule Bax punches holes directly in the outer mitochondrial membrane (OMM) through which cytochrome c pours out. But others believe that Bax acts more indirectly, cooperating with components of the permeability transition pore (PTP). This could cause the swelling and rupture of mitochondria (and the subsequent release of cytochrome c), or it might allow the pore to open widely enough to allow cytochrome c to pass through.

A report by Donald Newmeyer and co-workers in Cell now provides evidence for the first theory, showing that activated Bax can mediate the formation of supramolecular openings in the OMM. The authors did this by developing cell-free systems that allowed them to reproduce the behaviour of whole mitochondria.

Newmeyer and co-workers first isolated Xenopus OMMs and resealed them to form vesicles. Although these vesicles probably contained most of the proteins found in the OMM (including components of the PTP), they were devoid of inner-mitochondrial-membrane (IMM) and mitochondrial-matrix proteins. The authors then showed that fluoroscein dextrans — ranging in size from 10–2,000 kDa — could be released from these vesicles on treatment with either oligomerized (and therefore active) Bax (OG-Bax), or a cleaved, active form of another pro-apoptotic molecule Bid (N/C-Bid). Conversely, dextran release could be blocked by treatment with the anti-apoptotic protein Bcl-XL.

There was no gross disruption of the lipid bilayer, which indicates that supramolecular openings might somehow have been created in the membrane. But could N/C-Bid and OG-Bax do this in the absence of other proteins? To test this the authors generated two types of protein-free liposome — one from extracted mitochondrial lipids ('M liposome') and one from microsomal lipids ('ER liposome'). Dextran release triggered by OG-Bax was much more efficient from M liposomes than from ER liposomes. N/C-Bid alone did not trigger dextran release from either type of liposome; however, when N/C-Bid and monomeric Bax were added together there was efficient release from the M (but not the ER) liposomes. N/C-Bid can therefore activate monomeric Bax to produce the membrane openings.

The main difference between the two types of liposome is the additional presence of cardiolipin in the M liposomes. So the authors suggest that cardiolipin is important for both the membrane-targeting ability of Bid and the permeabilization function of Bax. Interestingly, however, the IMM — which is not permeabilized by Bax during apoptosis — is also rich in cardiolipin. One explanation, proposed by Newmeyer and co-workers, is that because the cardiolipin might be in the inner leaflet of the IMM, it would not be accessible to Bax. Further studies — using the new cell-free systems — should be able to address this question.