Genetics of Apoptosis

that disrupt heterodimeric interactions without disabling their respective apoptotic
functions (Cheng et al., 1996; Simonian et al., 1996; Kelekar et al., 1997).

2.2

BH3 domains

One approach to distinguish death-promoting functions of the Bax side of the BH
family from antagonism of Bcl-2-related survival functions is to isolate these activities
as independent protein domains. The BH3 homology region was first identified in
Bak as a domain required for both Bak cytotoxicity and formation of Bak: Bcl-xL
heterodimers (Chittenden et al., 1995). Synthetic Bak BH3 peptides induced
apoptosis and mitochondrial cytochrome c release in cell-free systems and when added
to cells as a fusion protein with the Antennapedia homeoprotein internalization
domain (Cosulich et al., 1997; Holinger et al., 1999). A mutated Bak BH3 peptide
(Leu^78 Ala) with reduced Bcl-xL affinity is nontoxic, suggesting that these isolated
peptides act by inhibiting Bcl-xL function (Narita et al., 1998). This interpretation
is complicated, however, by the observation that BH3 peptides interfere with the
binding of Bcl-xL to full-length partners (Diaz et al., 1997). Thus, BH3 domains
removed from the context of a proapoptotic BH protein may act by either directly
inhibiting an intrinsic Bcl-xL function or indirectly stimulating the activity of Bcl-xL
proapoptotic binding partners by promoting their dissociation (Ottilie et al., 1997;
Kelekar and Thompson, 1998). Since the structure of dimeric BH proteins is not yet
available, the possibility remains that dimerization-dependent and -independent
functions can be approached in a cellular system by domain analysis. Short of this,
further progress in this area must await a new generation of hypotheses for intrinsic
functions of the BH proteins independent of other BH-binding partners. In this vein,
it is noteworthy that the intrinsic pore-forming activities of Bcl-xL and Bcl-2 are
suppressed by Bak BH3 peptides (Tzung et al., 2001).

2.3

Mouse genetic models

The study of gene function in genetic backgrounds lacking known binding partners
is a classic approach to determine epigenetic relationships. Thus, in the case of BH
proteins, an agonist-antagonist relationship would predict loss of Bax-dependent
phenotypes in the absence of Bcl-2 or Bcl-xL, or vice versa. Studies of murine knockout
strains for Bcl-2 and Bax do not support this model, but rather suggest that each
protein can function independently of the other (Knudson and Korsmeyer, 1997).
Bcl-2-deficient mice developed accelerated thymic involution within the first month
after birth (Veis et al., 1993). Loss of Bax function in double-knockout mice leads to
a nearly complete suppression of thymic hypoplasia and excess thymocyte apoptosis
(Knudson and Korsmeyer, 1997). Conversely, Bcl-2 function is not redundant in the
absence of Bax, as lck promoter-driven Bcl-2 expression enhances survival of
thymocytes and mature T cells in Bax+/+ or Bax-/-backgrounds. These conclusions are

52 GENETICS OF APOPTOSIS