The effects of these group III Bcl-2 family members are inhibited by group I family members, including Bcl-2, Bcl-xL, and Mcl-1. Over the years, several different explanations for the antiapoptotic effects of these polypeptides have been proposed. First, it was suggested that group I polypeptides inhibit apoptosis by binding and neutralizing group II polypeptides, especially Bax (80,108,109) or Bak (110). The identification of Bcl-xL mutants that fail to bind Bax or Bak but still inhibit apoptosis (111) cast doubt on this model, at least as a universal mechanism of apoptosis inhibition. Second, it was proposed that Bcl-2 prevents apoptosis by increasing the antioxidant capacity of cells (112). Observations that place apoptosis-associated increases in reactive oxygen species downstream of caspase 3 activation (113,114) cast doubt on this model. Third, it was suggested that Bcl-2 prevents apoptosis by inhibiting the release of calcium from endoplasmic reticulum stores (115-119). Unfortunately, not all proapoptotic stimuli that are inhibited by Bcl-2 induce calcium release from intracellular stores. Finally, it has been hypothesized that group I polypeptides inhibit apoptosis by binding group III polypeptides, thereby preventing the release of cytochrome c (and other polypeptides) from mitochondria (120,121). This latter hypothesis has received the most attention.
Recent analysis of the binding between BH3 domains and group I polypeptides has demonstrated unanticipated specificity of these interactions (79). Whereas the group I polypeptides Bcl-xL and Bcl-w bind all tested BH3 peptides except Noxa with nanomolar affinity, other group I family members such as Mcl-1 and A1 bind Noxa, Puma, and Bim with much higher affinity than the remaining BH3 domain peptides. Coupled with the realization that different BH3-only polypeptides monitor different types of stress as described in the preceding section, these results suggest that different group I polypeptides should protect preferentially against certain types of proapoptotic stimuli. This has been documented to a certain extent (110,122) but requires further study.
As might be expected for polypeptides that are the final arbiters of cytochrome c release by various stresses, expression and activity of the group I Bcl-2 family members are highly regulated. Genes encoding Bcl-2, Bcl-xL, and Mcl-1 are activated by mitogen-activated protein (MAP) kinase signaling in some cell types (123,124), cytokine-mediated activation of STAT (signal transducer and activator of transcription) factors in others (125,126), and the phosphatidylinositol-3 kinase/Akt pathway, possibly acting through the transcription factor nuclear factor (NF)-kB, in others (127-130). In addition, the activity of group I polypeptides is regulated post-translationally by phosphorylation. Bcl-2 is activated by protein kinase Ca-mediated or extracellular signal-regulated kinase (ERK)-mediated phosphorylation on Ser70 (131,132) and other sites (133). Mcl-1 is likewise phosphorylated and stabilized in an ERK-dependent manner (134-136) as well as an ERK-independent manner (130,137). These observations provide at least a partial explanation for the antiapoptotic effects of MAP kinase activation (138,139). Further study will undoubtedly identify additional ways in which signal transduction pathways impinge on the antiapoptotic functions of Bcl-2 family members (e.g., 140,141), adding further complexity to the regulation of apoptosis in various cells. When the observed tissue-specific differences in expression of the antiapoptotic Bcl-2 family members (142,143) are also considered, it starts to become clear how the ability of various stimuli to induce cytochrome c release can vary widely from one cell type to another.
Although the model depicted in Fig. 3 provides an explanation for much of the available experimental data, it is important to realize that several issues require further investigation. For example, because extremely high concentrations of isolated BH3 domains rather than physiological concentrations of native polypeptides were utilized to detect the interactions in vitro, it remains to be demonstrated that all the proposed interactions occur in situ. Previous studies reporting inability to detect direct binding of Bim to Bax or Bak despite extensive efforts (102,144) highlight the importance of this issue. In addition, it is unclear what drives Bax or Bak into the mitochondrial membrane in the presence of BH3-only polypeptides such as Puma, Noxa, Bad, and Bmf, which are unable to directly activate Bax under cell-free conditions. One possibility is that there is another stimulus for group II polypeptide activation. It has been observed, for example, that activation of Jun N-terminal kinase, which commonly occurs during apoptosis, is accompanied by the release of Bax from cytoplasmic 14-3-3 proteins and subsequent localization to mitochondria (145). A second and related possibility is that group II polypeptides have an intrinsic propensity to insert in the outer mitochondrial membrane unless neutralized by group I polypeptides. Consistent with this possibility, it has been reported that Bak is sequestered by Mcl-1 and Bcl-xL but becomes active when released from both (110,146).
Was this article helpful?