Bax spontaneously kills cells in the presence of a BH3-only protein that occupies Bcl-2 (left and center), but not if the BH3-only protein cannot bind to Bcl-2 (right).


Our cells have a death wish, say Simon Willis, Jerry Adams, David Huang, and colleagues (Walter and Eliza Hall Institute, Melbourne, Australia). The apoptotic cascade, they show, is poised to destroy mammalian cells unless pro-survival proteins intervene.

The apoptotic executioners are Bax and Bak, which seem to bring about death by poking holes in mitochondria. But even thriving cells have plenty of Bax and Bak, suggesting these proteins are either inhibited in healthy cells or must be activated for apoptosis.

The balance between cellular life and death is shifted by proapoptotic BH3-only proteins and pro-survival Bcl-2 proteins. When certain BH3-only proteins were shown to bind to Bax, many researchers concluded that this interaction activated Bax. Liposome experiments supported this idea: the liposomes survived Bax but leaked when BH3 peptides were also added. But genetic evidence for the model was lacking, especially since BH3-only proteins were originally found to neutralize a Bcl-2 protein.

Having created the necessary knockout mice, Willis et al. now show that Bax and Bak do not require prior activation by the proposed BH3-only proteins. Instead, Bax and Bak seem to provoke apoptosis spontaneously when BH3-only proteins have soaked up protective Bcl-2 proteins.

A direct physiological interaction between Bcl-2 and Bax proteins has yet to be demonstrated convincingly, leaving Huang to wonder just how the executioners are regulated. “Bax and Bak,” he speculates, “might be intrinsically unstable proteins with a capacity to change [to an active, deadly] conformation. With lots of Bcl-2, the cell can block it.” But once BH3-only proteins occupy the Bcl-2, Bax and Bak might self-activate, resulting in suicide.

Death as a default pathway has precedents. “Developmental biologists would agree,” says Huang. They propose that “in developing tissues, cells die unless actively provided with a survival signal.” Huang thinks it might be “easier to communicate which cells should stay alive rather than which are to die.”


Willis, S.N., et al.