How the pro-apoptotic molecules Bak and Bax — which are potentially lethal — are maintained in an inactive, monomeric confirmation in viable cells is poorly understood. However, recent structural insights into the monomeric Bax molecule have provided a possible mechanism for its inactive status. And now, reporting in Science, Stanley Korsmeyer and colleagues have identified a protein — voltage-dependent anion channel 2 (Vdac2) — that keeps Bak in check.

Bak and Bax are required for mitochondrial apoptosis — 'BH3-only' members of the Bcl-2 family respond to death signals and subsequently trigger the activation of Bak and Bax, which leads to mitochondrial membrane permeabilization and the release of cytochrome c. This then initiates the caspase cascade.

To investigate whether Bak interacts with another mitochondrial protein that regulates its activity, Korsmeyer and colleagues used protein crosslinkers to identify a candidate protein (X) that complexes with Bak in purified mitochondria or whole cells. This Bak–X complex was lost when mitochondria were treated with the BH3-only protein tBid or when cells were treated with death stimuli. By testing various BH1- and BH3-domain mutants of both tBid and Bak, the authors concluded that X interacts with the Bak pocket that is formed by the BH1, BH2 and BH3 domains and can be displaced, directly or indirectly, by BH3-only molecules.

Protein X was identified as Vdac2, a low-abundance isoform of the Vdac outer-mitochondrial-membrane porin. Vdac2 was further implicated when the authors found that Vdac2-deficient embryonic stem cells lacked the Bak–X complex, which appeared when Vdac2 was re-expressed in these cells. Using haemagglutinin (HA)-tagged Vdac2, endogenous Bak — but not Bax — was coprecipitated. Similarly, HA-tagged Bak — but not Bax — coprecipitated endogenous Vdac2.

Bak in its active, oligomeric conformation is more susceptible to proteolysis than the inactive form, and the absence of Vdac2 increased its susceptibility. So, Vdac2 interacts with Bak, but not Bax, and regulates its conformation.

Next, Korsmeyer and co-workers set out to determine how Vdac2 modulates Bak-dependent apoptosis. Is Vdac2 an inhibitor of Bak-mediated apoptosis, or does it function as a proapoptotic factor itself when released from Bak? Vdac2 expression in Bax−/− cells inhibited apoptosis, but had no effect in Bak−/− cells, which argues against the latter possibility. Also, expression of Vdac2 inhibited tBid-induced apoptosis of Bax−/− cells, but not Bak−/− cells, which indicates that Vdac2 negatively regulates Bak-dependent apoptosis.

Cells deficient for Vdac2 were far more sensitive to death stimuli than Vdac1−/− and Vdac3−/− cells, both of which had similar sensitivities to wild-type cells. By re-expressing Vdac2, the susceptibility to apoptosis of Vdac2−/− cells reverted to normal. So, Vdac2 has a physiological role that is distinct from the other Vdac isoforms.

When analysing the apoptosis phenotype of Vdac2−/− cells, the authors noted an increased loss of mitochondrial transmembrane potential and the accelerated release of cytochrome c, compared with wild-type cells. After treatment with death stimuli, Vdac2−/− cells showed caspase activity and Bak oligomerization earlier than wild-type cells.

The authors concluded that Vdac2 is a specific inhibitor of Bak-dependent mitochondrial apoptosis, which, when absent, causes increased susceptibility to apoptotic death.