Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Mol Cell. Author manuscript; available in PMC 2010 July 6.
Published in final edited form as:
PMCID: PMC2897738

BAX and BAK Caught in the Act


In this issue of Molecular Cell, Kim et al. (2009) describe the steps involved in the direct activation of the proapoptotic proteins BAX and BAK by their BH3-only partners, resolving the controversy regarding direct versus indirect activation of these proteins.

Ever since apoptosis was defined as an active, inherently controlled homeostatic process of programmed cell death (Kerr et al., 1972), its molecular foundations have been the subject of intense study. The realization that apoptosis is not only a basic biological phenomenon but also plays key roles in cancer and many other major human diseases has propelled it to the forefront of biomedical research, with the hope that understanding its molecular mechanism will lead to new therapeutic approaches.

Description of the BCL-2 oncogene as a suppressor of cell death (Vaux et al., 1988), followed by the discoveries that its product localizes to mitochondrial membranes and belongs to a family of homologous life- and death-promoting proteins, provided the first insights to the molecular basis of apoptosis and established the first links between BCL-2 family proteins and the mitochondrial apoptotic pathway (Korsmeyer, 1995). These seminal findings paved the way for the subsequent studies that shape our current understanding of apoptosis (Youle and Strasser, 2008).

It is now well established that the BCL-2 family proteins are major apoptosis regulators whose activities are exerted through a network of intermolecular interactions that culminate in life or death decisions for the cell. These interactions are mediated by four conserved BCL-2 homology (BH1–BH4) domains: cytoprotective proteins (e.g., BCL-2 and BCL-XL) possess all four BH domains, while cytotoxic proteins can contain multiple BH domains (e.g., BAX and BAK) or just the BH3 domain (e.g., BID, BIM, and PUMA), which is highly conserved and essential for activating cell death.

Apoptosis can be inhibited when BCL-XL and BH3-only proteins associate, effectively neutralizing their conflicting activities. Sequestration of BCL-XL in this way could leave an opening for BAX or BAK to carry on their deadly activities unchecked (direct activation), but the data also indicate that apoptosis can be promoted when BH3 proteins directly bind and activate BAX/BAK, inducing mitochondrial outer membrane (MOM) permeabilization and apoptosis (Kuwana et al., 2005). Nevertheless, this has been the subject of controversy, because while the interaction of BH3s with BCL-XL is stable and well characterized (Sattler et al., 1997), their interaction with BAX/BAK is transient and has remained elusive until recently (Gavathiotis et al., 2008).

In this issue of Molecular Cell, Cheng and coworkers settle the question (Kim et al., 2009). They demonstrate that BAX and BAK directly interact with BH3 proteins (tBID, BIM, and PUMA) and dissect the mechanism by which this leads to BAX/BAK activation, MOM insertion, and oligomerization.

The NMR structures of BCL-XL (Sattler et al., 1997) and BAX (Suzuki et al., 2000) provided the initial framework for understanding the functions and interactions of these key apoptosis regulators. They exemplify the prototypical BCL-2 protein fold because they are remarkably similar to each other and to those of other multidomain BCL-2 proteins. A hydrophobic groove on the protein surface engages the helical BH3 domain of its proapoptotic partner to produce a stable complex that effectively sequesters the death signal. This BH3-binding groove is conserved across both anti- and proapoptotic BCL-2 family members and is crucial for mediating their interactions. However, in the cytosolic form of BAX, it is occupied by the hydrophobic C-terminal helix (α9) of the protein. This has important consequences for the interactions of BAX with BH3 proteins, which, instead, associate with BAX helices α1 and α6 at a site distinct from the canonical BH3-binding groove (Gavathiotis et al., 2008). Discovery of this new interaction site provided irrefutable evidence for the ephemeral BAX/BH3 interaction, but the question remained: how does BAX transition from a cytosolic, globular inactive state to a membrane-inserted active oligomer?

Kim et al. (2009) provide the answer. In this cell biology tour de force, the authors dissect the discrete steps involved in the BH3 activation of BAX from its cytosolic-inactive to membrane-active state, and in the activation of MOM-inserted BAK from monomeric-inactive to oligomeric-active state. The data reveal that activation is coupled with a conformational change that primes BAX for MOM insertion and oligomerization. Interaction of a BH3 activator with the newly discovered α1/α6 site leads to exposure of the BAX N terminus and release of the hydrophobic C terminus from the canonical BH3-binding groove at the opposite end of the molecule (Figure 1).

Figure 1
Stepwise Activation of BAX by BH3-Only Proteins

Notably, although the initial BH3/BAX interaction is transient and dynamic, the BH3 activator remains associated with BAX after its conformational change and translocation to the MOM, albeit at a different molecular site. Mutagenesis experiments reveal that the BAX/BH3 interaction rearranges to involve residues in the BAX BH1 and BH3 domains, and this helps drive BAX oligomerization. The apparent promiscuity of the BH3 activator protein is consistent with its inherent conformational flexibility (Yao et al., 2009).

In the cytosol, the BAX BH1 and BH3 domains participate in forming the canonical hydrophobic BH3-binding groove that is occupied by α9. Finding that they are actually involved with a BH3 protein in the membrane-inserted state suggests that the membrane conformation of BAX retains some vestiges of its cytosolic globular form, with potentially important implications for drug discovery.

In contrast to BAX, BAK is constitutively MOM associated. In this case, Kim et al. (2009) show that its N terminus is exposed prior to BH3 association, and activation involves only the late steps of BH3 interaction with the BAK BH1 and BH3 domains leading to oligomerization. Because BAK activation bypasses the initial steps of conformational change and MOM insertion, BAK expresses faster cell-killing kinetics than BAX.

By outlining the discrete steps in the BH3 activation of BAX/BAK, Cheng and coworkers not only resolve the controversy surrounding indirect versus direct activation and significantly advance our understanding of mitochondrion-dependent apoptosis but also set the stage for structural studies aimed at determining the molecular conformations and interactions of these key proteins in the membrane.


  • Gavathiotis E, Suzuki M, Davis ML, Pitter K, Bird GH, Katz SG, Tu HC, Kim H, Cheng EH, Tjandra N, Walensky LD. Nature. 2008;455:1076–1081. [PMC free article] [PubMed]
  • Kerr JF, Wyllie AH, Currie AR. Br. J. Cancer. 1972;26:239–257. [PMC free article] [PubMed]
  • Kim H, Tu H-C, Ren D, Takeuchi O, Jeffers JR, Zambetti GP, Hsieh JJ-D, Cheng E-YH. Mol. Cell. 2009;36(this issue):487–499. [PMC free article] [PubMed]
  • Korsmeyer SJ. Trends Genet. 1995;11:101–105. [PubMed]
  • Kuwana T, Bouchier-Hayes L, Chipuk JE, Bonzon C, Sullivan BA, Green DR, Newmeyer DD. Mol. Cell. 2005;17:525–535. [PubMed]
  • Sattler M, Liang H, Nettesheim D, Meadows RP, Harlan JE, Eberstadt M, Yoon HS, Shuker SB, Chang BS, Minn AJ, et al. Science. 1997;275:983–986. [PubMed]
  • Suzuki M, Youle RJ, Tjandra N. Cell. 2000;103:645–654. [PubMed]
  • Vaux DL, Cory S, Adams JM. Nature. 1988;335:440–442. [PubMed]
  • Yao Y, Bobkov AA, Plesniak LA, Marassi FM. Biochemistry. 2009;48:8704–8711. [PMC free article] [PubMed]
  • Youle RJ, Strasser A. Nat. Rev. Mol. Cell Biol. 2008;9:47–59. [PubMed]