The role of Bcl-2 and Bax in regulating apoptosis has been controversial since the discovery of the two molecules. Expression of Bcl-2 is sufficient to prevent programmed cell death (PCD) induced by any one of an assortment of stimuli in a wide variety of cell types. Bax was initially detected as a polypeptide that co-precipitates with Bcl-2 in cells expressing high levels of Bcl-2. Unlike Bcl-2, increased expression of Bax makes cells more susceptible to chemotherapy agents and if expression is high enough is sufficient to initiate PCD. Furthermore, Bax and Bcl-2 antagonize each other when expressed in the same cell. Thus the most popular model for regulation of PCD by Bax/Bcl-2 envisions Bax homodimers as toxic while the Bax molecules in Bcl-2-Bax heterodimers are inactive (no one has ever shown dimers so hetero-oligomers would be a more appropriate term). Insults that cause cell death are believed to do so via an accumulation of Bax monomers due to increased expression, or from release of pre-exisiting Bax molecules from an endogenous binding partner. It is the formation of Bax homodimers from these monomers that in turn initiates PCD. The mechanism by which Bax homodimers initiate PCD is the subject of this paper.
Recent data suggest that under some circumstances Bcl-2, its anti-apoptotic homologue Bcl-XL and Bax can all form pores in lipid bilayers. The electrophysiologic properties of these channels differs from group to group. However, in most published and unpublished reports pore formation is measured using truncated versions of the molecules and exposure to low pH (generally pH 4 or less) is required to get insertion into the lipid bilayer. This data together with the observed structural similarity between Bcl-XL and the pore-forming bacterial toxins, colicins and diptheria toxin have lead to the current models in which pore formation by Bcl-2 family members accounts for their pro- or anti- apoptotic activity. One model suggests that pore formation in the outer- mitochondrial membrane regulates the release of cytochrome C from mitochondria that in turn directly and indirectly activates the Caspases and nucleases that execute the cell.
In the present paper the authors present data that suggests Bax can form pores in membranes at physiological pH and that Bax pore formation is antagonized by Bcl-2. They suggest that pore forming activity is intrinsic to Bax and may account for the pro-apoptotic effects of Bax.
This paper is important not so much because of the model proposed but because of the questions raised by the data and the interpretation of the data presented. Pore formation by Bcl-2 family members is the latest 'fad' in PCD. Similar to other published and unpublished reports the authors of this paper demonstrate that under some circumstances fragments of Bcl-2 and Bax can form pores in membranes. The novelty in this particular paper is that pores are formed in 'real biological' membranes. Such a result should advance considerably biological relevance of pore formation. However, as mentioned above it was the questions raised by the data that intrigued me about this paper.
First the authors contend that Bcl-2 and Bax molecules lacking the carboxyl terminus are biologically active. Indeed the caption for Fig. 1 begins with the sentence "Characterization and purification of biologically active Bax and Bcl-2 lacking the hydrophobic COOH- terminal domain. " The data supporting this contention is that the fragments of Bax and Bcl-2 have effects similar to the full length proteins when injected into neurons. They also site a reference in which a truncated form of Bcl-2 was shown to be more functional than wild type (data that cannot be replicated in my own and other laboratories). Furthermore the micro-injection data provided in the paper contradicts that of virtually every other published analysis of the proteins. It also has not been possible to demonstrate anti-apoptotic activity for truncated Bcl-2 molecules in vitro. Even localization to different sub-cellular membranes has been shown to greatly effect the efficacy of Bcl-2 in a cell type dependent manner. Therefore, assuming that a micro-injection based assay provides a biologically relevant assay for Bcl-2 function it is very unlikely that truncated Bcl-2 or Bax molecules function in other than very limited circumstances (e.g. in neurons deprived of NGF). This is important because many of the other assays in the paper do not use neurons deprived of NGF.
Most intriguing is the demonstration that Bax exhibits lytic effects on neurons and red blood cells. In these experiments when Bax was added to the outside of cells hemoglobin release was clearly evident within 30-60 min. Moreover, addition of truncated Bcl-2 together with the truncated Bax prevented the lytic activity of Bax. However, since Bcl-2 and Bax form hetero-oligomers this result is difficult to interpret. This result becomes even more ambiguous when one considers that the Youle group has demonstrated that inside cells, Bcl-2 and Bax do not form hetero-oligomers unless the 'normal' cellular conformations are perturbed. I wonder if the truncated forms of Bcl-2 and Bax purified from E. coli form hetero-oligomers?
The truncated form of Bax is also shown to release lipid-encapsulated carboxyfluorescein. Furthermore, "the release of carboxyfluorescein induced by 2.5 nM Bax increased in a pH- dependent manner and was eight times greater at pH 4.0 than at pH 7.5." Similarly Bcl-2 "was as efficient as Bax at pH 4.0...". Since Bcl-2 does not release lipid-encasulated carboxyfluorescein at neutral pH and during 'normal' cellular physiology neither Bax nor Bcl-2 would ever be exposed to pH 4.0 (unlike the toxins they are compared to) my interpretation of this data is that denatured Bax and Bcl-2 disrupt lipid bilayers. If the truncated Bax purified from E. coli contains a population of molecules which are already suitably 'denatured' while Bcl-2 survives the purification (or refolds) then this interpretation would fit all of the data presented.
The final set of experiments examine pore formation by the truncated Bax protein electrophysiologically with planar lipid bilayers. I am sure many of the readers of this journal club are more familiar with these types of experiments than I. Therefore, I hope that some of you will look at Fig. 4 in the paper and report on your own interpretation of the data presented. There are a lot of different conductances present in the Bax containing liposomes and the conductances vary over a pretty wide range. Furthermore, the pores exhibit ion selectivity, albeit only a ratio of 2.1 for Na+ to Cl-. This seems to me a very surprising result. How can a pore exhibit ion selectivity yet be large enough to release lipid-encasulated carboxyfluorescein? Or for that matter hemoglobin?
Thus we are left with some peptide domains with pretty remarkable properties and a variety of questions about how one can examine membrane proteins using biophysical techniques in vitro and relate the data to a complex biological process in cells. It is of course, an extremely difficult problem encountered with most membrane proteins. Membrane proteins are difficult to study in the complex environment of the endogenous membrand but more often than not they are insoluble when removed from membranes. The approach used in this paper was to remove the one domain of the protein that has been shown to be a bonafide transmembrane segement. But how can one reasonably study a membrane protein if the domain responsible for membrane assembly and subcellular targeting is removed? In nature if these molecules do form pores then it is likely that there will be some contribution due to the hydrophobic carboxyl-terminus of the molecule on pore formation and or function.
Proteins of the Bcl-2 family are intracellular membrane-associated proteins that regulate programmed cell death (apoptosis either positively or negatively by as yet unknown mechanisms. Bax a pro-apoptotic member of the Bcl-2 family, was shown to form channels in lipid membranes. Bax triggered the release of liposome-encapsulatied carboxylfluorescein at both neutral and acidic pH. At physiological pH, release could be blocked by Bcl-2. Bcl-2, in contrast, triggered carboxylfluorescein release at acidic pH only. In planar lipid bilayers, Bax formed pH- and voltage-dependent ion-conducting channels. Thus, the pro-apoptotic effects of Bax may be elicited through an intrinsic pore-forming activity that can be antagonized by Bcl-2.