December 2008 featured article
The structure of the pro-apoptotic protein BAX with a bound BIM BH3 'death helix' reveals that a previously unsuspected binding site on BAX triggers activation of the BAX cell-death pathway.
Surface diagram illustrating the BAX binding site. The side chains of hydrophobic, positively charged, negatively charged and hydrophilic residues are coloured yellow, blue, red and green, respectively. (PDB 2K7W)
Stressed cells mainly die by apoptosis, and a key player in this process is the protein BAX. Until now, little molecular information existed to explain how BAX is directly activated to induce cell death, but recent structural studies, supported by biochemical and cellular work, are helping to unravel BAX's activation site and trigger mechanism.
When cells are unable to withstand stress, the 'death proteins' BAX and BAK permeabilize the mitochondrial membrane to release a cascade of molecules that induce cell death by apoptotsis. BAX and BAK are pro-apoptotic members of the BCL-2 family, which contains pro- and anti-apoptotic proteins that maintain the balance between cellular life and death through competitive interactions with each other.
BID and BIM, which are members of the BCL-2 family, have been proposed to directly activate BAX, but exactly how they do so has been the subject of much debate. BID and BIM are so-called 'BH3-only' proteins, containing sequence homologous to only one of the conserved domains (BH domains) characteristic of the BCL-2 family, unlike BAX and BAK, which are 'multi-BCL-2 homology domain proteins'. Once activated, BAX self-oligomerizes, and these oligomers disrupt mitochondrial integrity by assembling within the outer mitochondrial membrane.
To establish a detailed mechanism of direct BAX activation, Gavathiotis et al. used a structurally stabilized form of the BIM BH3 domain known as BIM SAHB (stabilized α-helices of BCL-2 domains). They had previously shown that BID SAHB could bind to BAX and directly activate BAX-mediated mitochondrial apoptosis.
Gavathiotis et al. show that BIM SAHB binds to BAX at an interaction site that is on the opposite side from the groove on BAX where anti-apoptotic proteins bind BH3 domains. They provide the first structure of a BH3 α-helix bound to a full-length pro-apoptotic multi-BCL-2 homology domain protein.
NMR spectroscopy showed that when BIM SAHB binds, the NMR resonances of residues from helices α1 and α6 of BAX are shifted, as are the residues in the flexible loop between helices α1 and α2, indicating that these residues are likely to form the binding site. Although the location of the BH3-binding site on BAX is different from that of anti-apoptotic proteins, the site itself is a hydrophobic groove that shares some topological features.
The NMR spectroscopy also indicated that BAX undergoes a conformational change upon binding to BIM SAHB. Gavathiotis et al. followed up this observation by establishing that BIM SAHB triggers BAX oligomerization, as monitored by size exclusion chromatography, and that it induces the exposure of BAX's N-terminal residues, detected using a conformation-specific antibody.
They confirmed the functional activation of BAX by BIM SAHB by liposomal and mitochondrial assays, and concluded that BIM SAHB directly activates BAX in a dose-responsive manner.
The authors then carried out mutagenesis of BAX to link functional activity to binding specificity, and found that point mutations at the interface impaired BIM SAHB binding and BAX activation. For example, recombinant BAX with a point mutation at lysine 21, the residue that underwent the largest resonance shift upon binding of BIM SAHB, had significantly lower levels of oligomerization than the wild-type protein.
Finally, Gavathiotis et al. examined the significance of their findings in cells. They inserted either a wild-type or the mutant BAX(K21E) construct into mouse embryonic fibroblasts lacking both Bax and Bak function. On treatment with BIM SAHB or with a generic apoptotic stimulus, the wild-type BAX induced time-dependent apoptosis of BAX-reconstituted fibroblast cells, whereas BAX(K21E)-reconstituted cells had a reduced level of triggered apoptosis.
This work highlights the specificity of the interaction of BAX interaction with BIM BH3 and shows its functional relevance in vitro and in cells. The identification of a previously unknown binding site on BAX will have important implications for the development of apoptosis-based therapies for cancer and other diseases.