The paradox of Bcl-2: How does paclitaxel convert Bcl-2 function from antiapoptotic to proapoptotic?

Bcl-2 is an antiapoptotic protein often overexpressed in human cancer and exerts an antiapoptotic function at mitochondrial level, through the opening suppression of the permeability transition pore complex (PTPC). This process prevents the consequent leakage of mitochondrial transmembrane potential and effluxing from mitochondria of proapoptotic factors such as Cytochrome C, AIF and APAF-1. Recently, in ovarian cancer cells paclitaxel (PTX)-resistance has been associated to downregulation of Bcl-2. In order to investigate this paradox, human Bcl-2 was stably transfected in PTX-resistant ovarian cancer cells and mitochondria were isolated from these cells. Mitochondria prepared from untransfected control cells showed undetectable levels of Bcl-2 and PTX was unable to modulate the opening of PTPC, as demonstrated by the uptake of the cationic fluorochrome Rhodamine 123. In the same cells transfected with Bcl-2, sensitivity to paclitaxel was restored and the drug was able to induce the opening of PTPC and the leakage of mitochondrial transmembrane potential. Among Bcl-2 family members, the peculiar difference of Bcl-2 consists in the disordered loop domain. Therefore, we prepared stable Bcl-2 transformed cells with a construct devoid of the loop domain (Bcl-2-Δ). In mitochondria prepared from these cells, PTX was unable to modulate the opening of PTPC. These findings suggest that the disordered loop of Bcl-2 is involved in the PTX binding, but did not clarify if this occurs in a direct or indirect way, with the involvement of other proteins. In order to address this issue, we used plasmon resonance-based optical biosensor technology. In vitro Bcl-2 and Bcl-2-Δ translated proteins were immobilised on a optical biochip to permit a real time monitoring of the potential binding of paclitaxel to Bcl-2. The experimental output pointed out that PTX directly binds to Bcl-2, and not to Bcl-2-Δ, this signaling that the PTX binding site is located within the disordered loop domain of the protein. With the aim of better understanding the binding mechanism, we performed molecular modelling investigations, taking as template the experimentally determined structure of class I beta-tubulin in complex with PTX. Results indicated that the putative binding site of PTX in the disordered loop domain is indeed very similar to that present in beta tubulin. In beta-tubulin PTX binds to a pocket with strong interactions with His229, Pro360-Pro359 and the sequence Arg278-Ser277-Thr276-Leu275-Pro274. Similar interactions were found in our paclitaxel/Bcl-2 docked complex: His58, Pro39-Pro40 and the sequence Thr69-Arg68-Ala-67-Val66-Pro65. Due to the similarity in the binding, it is likely that PTX is a peptide-mimicking structure with a structure similar to endogenous agonist(s) able to bind to both beta-tubulin and Bcl-2, and consequently transferring a death signal from microtubules to mitochondria.