by several amphipathic helices carrying the BH motifs. Of particular interest is a
structure of the antiapoptotic Bcl-xL protein in complex with the BH3 motif of the
proapoptotic Bax protein (Sattler et al., 1997). The interactions seen in this structure
can explain the propensity of Bcl-2-like proteins to heterodimerize.
The early observation that many proapoptotic Bcl-2 homologs bind to Bcl-2 and
other negative regulators leads to two contradicting models: do the antiapoptotic
Bcl-2-like members inhibit apoptosis by binding and sequestering the proapoptotic
members, or do they inhibit apoptosis by a different mechanism that can be prevented
by the proapoptotic Bcl-2 relatives? Initially, the latter idea seemed more likely, as
the nematode C.elegans contains only one Bcl-2 homolog, the antiapoptotic ced-9
gene product. Since there are no proapoptotic binding partners, a different inhibitory
pathway must be operative, at least in the nematode. However, there is accumulating
evidence that several proapoptic family members, such as Bax, are able to trigger the
mitochondrial permeability pore autonomously, although the mechanism is not
entirely understood (Martinou and Green, 2001). An interesting clue came from the
observation of a structural similarity between the Bcl-xL fold and that of bacterial
pore-forming toxins (Muchmore et al., 1996). A model wherein proapoptotic Bax-
like family members can form membrane pores by themselves is supported by the
observation of large structural changes in the Bcl-xL structure upon binding to
detergent micelles (Losonczi et al., 2000). Alternative explanations are also possible,
Table 8 Mammalian members of the Bcl-2 family
Domain abbreviations: BH1, BH2, BH3, and BH4, homology regions in the Bcl-2 family;
TM, transmembrane region.
92 GENETICS OF APOPTOSIS