the Vander Heiden study might be more significant for passage of uncharged organic
molecules.
Why do Bcl-2 and Bcl-xL form channels in synthetic lipid membranes with similar
electrical properties as Bax, but studies of outer mitochondrial membranes reveal only
Bax-associated channels? There are some existing examples of membrane proteins
with specific transport functions in situ that behave as nonspecific or unregulated
channels in artificial lipid membranes. The adenine nucleotide transporter (ANT)
has a strict requirement for cardiolipin for exchange activity. Calcium has been
proposed to convert ANT to a large nonspecific channel by binding to the phosphate
head groups of cardiolipin normally associated with ANT (Hoffmann et al., 1994;
Brustovetsky and Klingenberg, 1996). Uncoupling proteins 1, 2, and 3 exhibit Cl—
transport activity, but not H+ transport in the absence of the cofactor coenzyme Q
(Echtay et al., 2001). One may speculate that a specific transport function of Bcl-2/
Bcl-xL could require a cofactor missing in artificial lipid membranes, with purified
recombinant protein behaving as a nonspecific, and perhaps nonphysiologic, pore.
Similar questions can be raised for some variability in the reports concerning the
oligomerization state of mitochondria-localized Bax or Bid (Desagher et al., 1999;
Grinberg et al., 2002).
2.9
Antiapoptotic functions: what color is your parachute?If the antiapoptotic functions of Bcl-2/Bcl-xL are not attributable to Bcl-xL pores and,
at least under some circumstances, are independent of heterodimerizations with
proapoptotic BH members, what is left? An intriguing observation in several
laboratories is the ability of mitochondrial Bcl-2 to inhibit translocation of
cytoplasmic Bax to mitochondrial membranes. Nomura et al. (1999) reported that
cytosol from apoptotic cells facilitated Bax targeting to isolated mitochondria.
Mitochondria obtained from Bcl-2 transgenic mice were targeted as efficiently by
recombinant Bax in the presence of apoptotic cytosol as those from nontransgenic
littermates. Conversely, cytosol harvested from Bcl-2-transfected cells following an
apoptotic stimulus did not support Bax translocation to mitochondrial membranes.
As Bcl-2 protein was not detected in the cytosol samples, these results suggest that
mitochondrial Bcl-2 modifies one or more cytosolic factors (which may originate in
mitochondria) required for Bax translocation. The potential role of caspases was
discounted by the failure of caspase inhibitors to block Bax translocation in whole-
cell studies (Nomura et al., 1999), although others have obtained opposite results
(Goping et al., 1998). Murphy et al. (2000) further localized the Bcl-2-dependent
step as a conformational change in cytosolic Bax, detectable with antibodies to an N-
terminal epitope, that preceded Bax translocation. The numerous demonstrations of
a Bcl-2/Bcl-xL effect on cellular metabolism suggest some avenues for further
investigation of an intrinsic biochemical function (Hockenbery et al., 1993; Lam et
al., 1994; Shimizu et al., 1998).
MAKING SENSE OF THE BCL-2 FAMILY OF APOPTOSIS REGULATORS 59