Why do regulators of apoptosis look like bacterial toxins?

Correspondence Why do regulators of apoptosis look like bacterial toxins? Yuri Lazebnik The Bcl-2 family of proteins includes key regulators of apoptosis (reviewed in [1,2]), some of which promote cell death while others prevent it. The observation that the Bcl-2 family regulates the release of pro-apoptotic proteins from mitochondria led to the concept that members of the family control cell survival by regulating the permeability of these organelles [3,4]. Several models explain how this regulation might be achieved (reviewed in [5]). One of the models comes from the intriguing observation that the tertiary structure of Bcl-2 proteins resembles that of colicins and diphtheria toxin (DT) [6]. Colicins are toxins made by bacteria to kill other bacterial strains (reviewed in [7]). Colicins contain three domains: the receptor binding, the translocation and the toxic domain. Bcl-2 proteins are similar in structure to the toxic domains, that kill by making pores in the bacterial membrane. Perforation dissipates ion gradients, which leads to the release of essential metabolites, the influx of water and cell rupture. This structural similarity suggested that Bcl-2 proteins might function by making pores in mitochondrial membranes (reviewed in [5,8]). Although some Bcl-2 proteins can indeed form pores in synthetic lipid membranes, the model faces substantial challenges (reviewed in [2]). For example, the structures of Bax and Bcl-x are remarkably similar [9], suggesting that if one of the proteins makes the pores, the other would likely make them as well. Yet Bax promotes the release of pro-apoptotic molecules from mitochondria, whereas Bcl-x prevents it. This puzzle has not been made easier by the difficulty to explain how Bcl-x or Bcl-2 could prevent release of proteins from mitochondria by making pores. These and other unexplained observations argue that either the similarity between the Bcl-2 family and the toxins is a coincidence, or they share a common function unrelated to the formation of pores. I would like to argue in favor of the second possibility by exploring the homology of the Bcl-2-like proteins to diphtheria toxin. Diphtheria toxin kills human cells by a mechanism that is different from that employed by colicins to kill bacteria. DT belongs to the AB family of toxins, which includes bacterial and plant toxins that are made of two moieties: A and B (reviewed in [10]). The A moieties are usually enzymes that are toxic to the cell. For example, the A moiety of DT is an ADP-ribosyl …