Structure comparison between Bax and Bcl-xL complexed to Bak BH3 peptide. Top left: ribbon representation of Bax, showing the tightly packed nine α-helices (α1–9). Each helix is designated with a different colour. Top right: ribbon representation of Bcl-xL complexed to the Bak BH3 peptide. A view similar to the Bax structure is chosen for comparison, and the same colour scheme as in Bax has been used to indicate the different helices (with the exception of the bound peptide, which is coloured yellow). Bottom panels: close-up views of the hydrophobic pocket occupied by the carboxy-terminal helix (α9) in Bax, and a similar view of the peptide-binding pocket in Bcl-xL. Note the reverse orientation of the bound peptide in the hydrophobic pocket relative to the carboxy-terminal helix of Bax. Image courtesy of Nico Tjandra. Credit: Nico Tjandra, NIH

There's a fine line between life and death — a tightrope walked by the Bcl-2 family of pro- and anti-apoptotic proteins. Tip the balance too far one way and, as discussed in two new papers, the cell slides helplessly to its death.

Members of the Bcl-2 family fall into three subfamilies. On one side of the death equation are the anti-apoptotic proteins Bcl-2 and Bcl-xL; on the other side are the pro-apoptotic members, including the Bax subfamily (Bax and Bak) and the 'BH3-only' proteins (such as Bid and Bad). But how do the interactions between these various proteins control apoptosis?

Reporting in Cell, Nico Tjandra and colleagues discuss the regulation of one pro-apoptotic Bcl-2 family member, Bax. They have used NMR to solve its solution structure — a structure, say the authors, that is strikingly similar to that of Bcl-xL. Both contain nine α-helices, with the first eight (α1–α8) occupying almost identical positions despite low sequence similarity. But whereas Bcl-xL (right-hand figure) contains a hydrophobic pocket that can accommodate another protein (here, the Bak BH3 peptide; yellow), in Bax this pocket is occupied by its own α9 helix (green). How, then, does Bax interact with other members of the Bcl-2 family?

Tjandra and colleagues believe that the answer lies in a conformational change. Early during apoptosis, Bax translocates from the cytosol to the mitochondria. Here it inserts into the outer mitochondrial membrane (OMM), where it is involved in the release of cytochrome c and, ultimately, apoptosis (see the Opinion articles on pages 63 and 67 of this issue for details). The authors propose that a conformational change, which allows Bax to insert into the OMM, also disengages the α9 helix from the hydrophobic pocket. This would expose the pocket, allowing it to bind other proteins.

Given its reported interaction with Bax, one candidate that might slip into this pocket is Bid — the subject of a report in Genes and Development by Craig Thompson, Stanley Korsmeyer and colleagues. They have studied tBid, a truncated, physiologically active form of Bid, which is involved in the release of cytochrome c from mitochondria. It could do this either by itself forming a pore through which cytochrome c can escape across the OMM, or by acting as a death ligand to activate another mitochrondrial protein with the same net effect.

Korsmeyer and co-workers favour the second possibility. They show that, for apoptosis to occur, tBid's BH3 domain (which is required for dimerization) must be present on the cytoplasmic face of the mitochondria. This indicates that tBid acts by binding other proteins, and the authors reveal at least one of its partners to be Bak. Not only do Bak and tBid interact physically, but this association is required for the release of cytochrome c. Finally, on binding tBid, Bak undergoes a conformational change and forms oligomers — indicating, speculate the authors, the possible formation of a cytochrome c-permeant pore.

The mitochondrial preparation used by Korsmeyer and colleagues did not contain abundant Bax, and the authors are now repeating these experiments in cells that contain both Bax and Bak. Indeed, the idea that Bid acts as a death ligand fits well with the proposed opening of Bax's hydrophobic pocket to accommodate other members of the Bcl-2 family. However, there is still much controversy over the functions of the various Bcl-2 family members in apoptosis, and it's likely to be some time before all of their molecular balancing acts are revealed to the watching scientific audience.