The VDAC1 N-terminus is essential both for apoptosis and the protective effect of anti-apoptotic proteins

Programmed cell death – apoptosis – is a genetically predetermined mechanism elicited by several molecular pathways (Kroemer et al., 2007). Defects in the regulation of apoptosis are often associated with disease, such as neuronal degenerative diseases, tumorigenesis, autoimmune disorders, and viral infections (Reed, 2002; Thompson, 1995). Mitochondria are potent integrators and coordinators of apoptosis by releasing pro-apoptotic agents and/or disrupting cellular energy metabolism. Following an apoptotic stimulus, various proteins that normally reside in the intermembrane space of the mitochondria, including cytochrome c and apoptosis-inducing factor (AIF), are released to initiate the activation of pro-caspases, the protease mediators of cell death (Green and Reed, 1998; Halestrap et al., 2000). It remains unclear, however, how these apoptotic initiators cross the outer mitochondrial membrane (OMM) and are released into the cytosol. Some models predict that release is facilitated by swelling of the mitochondrial matrix and subsequent rupture of the OMM, whereas other models predict the formation of protein-conducting channels that are large enough to allow the passage of cytochrome c and other proteins into the cytosol, without compromising the integrity of the OMM (Halestrap et al., 2000; Shoshan-Barmatz et al., 2008). The finding that cytochrome c can leak from intact mitochondria (Doran and Halestrap, 2000; Martinou et al., 2000; Shoshan-Barmatz et al., 2006) supports the latter models. Release of cytochrome c is regulated by different proteins, some of which interact with the OMM protein, VDAC (voltage-dependent anion channel), also known as mitochondrial porin. VDAC is thus believed to participate in the release of cytochrome c and to interact with anti- and pro-apoptotic proteins, and is considered a probable candidate for such an OMM pore-forming protein (Crompton, 1999; Halestrap et al., 2000; Kroemer et al., 2007; Lemasters et al., 1999; Shoshan-Barmatz et al., 2006; Tsujimoto and Shimizu, 2002; Zamzami and Kroemer, 2003).

Indeed, a variety of apoptotic stimuli were shown to trigger apoptosis by modulation of VDAC and implicate VDAC1 as a component of the apoptosis machinery (Shoshan-Barmatz et al., 2008; Shoshan-Barmatz et al., 2006; Tsujimoto and Shimizu, 2002; Zheng et al., 2004). By contrast, VDAC proteins have been reported to be dispensable for Ca²⁺- and oxidative stress-induced permeability transition pore (PTP) opening (Baines et al., 2007). Nonetheless, recent studies have elucidated that VDAC1 is an indispensable protein for PTP opening and induction of apoptosis (Tajeddine et al., 2008; Yuan et al., 2008). It was found that reducing VDAC1 levels by siRNA attenuates endostatin-induced apoptosis in endothelial cells, and that endostatin-induced PTP opening is accompanied by upregulation of VDAC1 expression (Yuan et al., 2008). Similarly, siRNA-mediated depletion of VDAC1 strongly reduced cisplatin-induced cytochrome c release and apoptotic cell death (Tajeddine et al., 2008). These recent reports and previous accumulated evidence strongly suggest that VDAC serves a key function in apoptosis (Crompton, 1999; Halestrap et al., 2000; Kroemer et al., 2007; Lemasters et al., 1999; Shoshan-Barmatz et al., 2008; Shoshan-Barmatz et al., 2006; Tsujimoto and Shimizu, 2002; Zamzami and Kroemer, 2003). VDAC is the major OMM transporter and plays an important role in the controlled passage of adenine nucleotides, Ca²⁺ and other metabolites into and out of mitochondria, thereby assuming an important role in energy production through the control of metabolite traffic (Colombini, 2004; Lemasters and Holmuhamedov, 2006). Recently, the three-dimensional (3D) structures of recombinant human and murine VDAC1 were determined by NMR spectroscopy and X-ray crystallography (Bayrhuber et al., 2008; Hiller et al., 2008; Ujwal et al., 2008). VDAC1 was shown to adopt a β-barrel architecture comprising 19 β-strands and an α-helix located within the pore. The α-helix of the N-terminal segment is oriented against the interior wall, causing a partial narrowing at the center of the pore. As such, this helix is ideally positioned to regulate the conductance of ions and metabolites passing through the VDAC pore. These structural and other studies (De Pinto et al., 2007) further indicate that the helical region of the VDAC N-terminal domain forms a distorted α-helix that does not span the whole sequence but only part of it. Although these recently determined VDAC1 structures suggest that the hydrophilic N-terminus of the protein is nestled within the pore (Bayrhuber et al., 2008; Hiller et al., 2008; Ujwal et al., 2008), other approaches point to the N-terminal α-helix as being exposed to the cytoplasm (De Pinto et al., 2003), to cross the membrane (Colombini, 2004) or to lie on the membrane surface (Reymann et al., 1995). Further evidence suggesting that the N-terminal region of VDAC1 in fact constitutes a mobile component of the protein comes from the observations that the N-terminal α-helix exhibits motion during voltage gating (Mannella, 1997), that anti-VDAC antibodies raised against the N-terminal region of the protein interact with membranal VDAC (Abu-Hamad et al., 2006; Junankar et al., 1995; Shoshan-Barmatz et al., 2004) and that mobility of the N-terminal α-helix may modulate the accessibility of apoptosis-regulating proteins of the Bcl2 family (i.e. Bax and Bcl-xL) to their binding sites on VDAC1 (Shi et al., 2003).

In the present study, the role of the VDAC1 N-terminal domain in regulating VDAC1 activity in cell life and apoptotic cell death was investigated. Using Δ(26)mVDAC1, an N-terminal truncated form of murine VDAC1 expressed in human (h)VDAC1-shRNA cells expressing low levels of endogenous hVDAC1, the N-terminal region of VDAC1 was shown to be required for the release of cytochrome c and apoptotic cell death. Furthermore, the VDAC1 N-terminal α-helix was shown to mediate the interaction of VDAC1 with the anti-apoptotic, pro-survival factors, hexokinase (HK)-I, HK-II and Bcl2. The findings of this study thus indicate that VDAC1 is a critical component in mitochondria-mediated apoptosis and its regulation and point to the N-terminal domain of VDAC1 as being a critical component in the regulation of cytochrome c release and apoptosis.

The N-terminal domain of VDAC1 has been the object of considerable interest and was proposed to possess functionally relevant properties (Colombini et al., 1987; De Pinto et al., 2007; Peng et al., 1992; Stanley et al., 1995; Ujwal et al., 2008; Yehezkel et al., 2007). VDAC provides the major pathway for nucleotides and other metabolites moving across the OMM. To verify that the N-terminal region of VDAC1 is necessary for such transport activity, N-terminally truncated murine VDAC1 [Δ(26)mVDAC1] was expressed in human T-REx293 cells in which endogenous VDAC1 expression had been suppressed by ∼85% using a single hVDAC1-shRNA (Abu-Hamad et al., 2006). Such cells proliferated slowly, compared with normal cells (Fig. 1A), owing to low levels of ATP and reduced ATP synthesis capacity, both of which could be restored upon expression of mVDAC1 under the control of tetracycline (Abu-Hamad et al., 2006). Δ(26)mVDAC1 expression in hVDAC1-silenced cells (shRNA-hVDAC1) also restored cell growth (Fig. 1A), suggesting that the 26 N-terminal residues are not essential for VDAC1 function in cell growth and that Δ(26)mVDAC1 can substitute for native VDAC1 in supporting the metabolite exchange between the mitochondria and the cytosol in cells expressing low levels of VDAC1 (Abu-Hamad et al., 2006). Expressing N-truncated VDAC1 in porin-less yeast restored their growth to levels of porin-less yeast expressing native VDAC (data not shown).

The oxygen consumption rates of control T-REx cells, shRNA-hVDAC1 cells with a low VDAC1 expression level, and cells stably expressing native or N-terminal-truncated mVDAC1 were measured as an index of in situ mitochondrial metabolism (Fig. 1B). The oxygen consumption rate of shRNA-VDAC1 cells was about 67% that of control cells, suggesting that VDAC1 is required for normal mitochondrial respiration. When such cells were transfected to express either native or Δ(26)mVDAC1, the respiration rate was increased to almost the level obtained by control cells. Rotenone inhibited oxygen consumption in all cell types (Fig. 1B). The expression levels of VDAC, as analyzed by immunoblotting using polyclonal anti-VDAC antibodies, clearly show that the level of VDAC expression in VDAC1-shRNA cells was dramatically reduced by over 90%. That value could be restored to control levels by transfecting the cells with plasmids encoding mVDAC1 or Δ(26)mVDAC1 (Fig. 1C). These results suggest that Δ(26)mVDAC1 is as active as native mVDAC1.

We then demonstrated, by confocal microscopy, that mVDAC1-GFP and Δ(26)mVDAC1-GFP expressed in T-REx293 cells showed similarly distributed punctuated fluorescence, which was confined to the mitochondria (Fig. 1D). The expression of Δ(26)mVDAC1-GFP in these cells was also shown by immunoblot analysis (Fig. 1E). The results indicate that both native and truncated mVDAC1 were localized to the mitochondria, suggesting that the 26 N-terminal residues do not act as a targeting sequence.

We further demonstrated that Δ(26)mVDAC1 functions as a channel-forming protein. Δ(26)mVDAC1 or mVDAC1 were expressed in the mitochondria of porin-less yeast, purified and reconstituted into a planar lipid bilayer (PLB), where VDAC channel activity was analyzed. Single-channel recording gave a maximal conductance of 4 nS (at 1 M NaCl) for both the native and the N-terminally truncated proteins. However, whereas bilayer reconstituted-mVDAC1 showed a typical voltage-dependence conductance, with the highest values being obtained at transmembrane potentials between –20 and +20 mV and decreasing at both high negative and positive potentials (Fig. 2A), Δ(26)mVDAC1 showed no voltage dependence and exhibited high conductance at all the tested voltages (Fig. 2A) in agreement with previous results (De Pinto et al., 2008; Koppel et al., 1998; Popp et al., 1996). Since there is no significant membrane potential across the OMM, the loss of voltage sensitivity of Δ(26)mVDAC1 may not have physiological significance, as reflected in the finding that its ability to support cell growth was similar to that of the intact protein (Fig. 1A).

In our previous studies (Abu-Hamad et al., 2008; Azoulay-Zohar et al., 2004), we demonstrated that hexokinase (HK) interacts with VDAC to decrease it channel activity. The interaction of the VDAC1 N-terminal region with purified rat brain HK-I was studied using bilayer-reconstituted mVDAC1 or Δ(26)mVDAC1. Addition of HK-I to bilayer-reconstituted native mVDAC1 induced the channel, fluctuating between the fully open and a sub-conducting state, to adopt a stable, long-lived low-conducting state (Fig. 2B). Although Δ(26)mVDAC1 assumed a stable highly conductive state, as would be expected for a voltage-insensitive channel (Fig. 2A), HK-I had no effect on the channel activity of Δ(26)VDAC1. These results indicate that the VDAC1 N-terminal region is required for HK-I modification of channel conductance.

It has been previously demonstrated that overexpression of human, murine, yeast or rice VDAC1, VDAC1-GFP or mutated VDAC1 induces apoptotic cell death (Abu-Hamad et al., 2006; Bae et al., 2003; Zaid et al., 2005). Accordingly, upon overexpression of mVDAC1 in T-REx293 cells silenced for hVDAC expression by hVDAC1-shRNA (using 2.5 μg/ml tetracycline), growth continued for 3 days before cells starting to die, with all cells having died by the fifth day (Fig. 3A). By contrast, cells expressing Δ(26)mVDAC1 continued to grow for >8 days, with only 5-8% of the cells succumbing to apoptosis, as reflected by enhanced nuclear fragmentation, visualized by Acridine Orange and ethidium bromide staining (Fig. 3B). The overexpression of native and N-terminal truncated mVDAC1 in cells stably expressing these proteins is shown in Fig. 3C.

In the light of the findings that cells overexpressing N-terminal-truncated VDAC1 had lost their ability to undergo mitochondria-mediated apoptosis (Fig. 3), we tested for the involvement of the N-terminal region of VDAC1 in apoptosis, as induced by various stimuli. hVDAC1-shRNA-T-REx293 cells expressing mVDAC1 or Δ(26)mVDAC1 were challenged with various apoptosis-inducing agents, all of which affect the mitochondria but produce their effect by a different mechanisms (Fig. 4). Staurosporine (STS), curcumin, As₂O₃ and cisplatin, all induced death in cells expressing native mVDAC1 but not in those expressing Δ(26)mVDAC1. By contrast, tumor necrosis factor alpha (TNFα), an agent that activates apoptosis via a mitochondria-independent pathway (Schmitz et al., 2000), induced a similar extent of cell death in native mVDAC1- and Δ(26)mVDAC1-expressing cells (Fig. 4A).

Similar results were obtained by flow cytometric analysis, using annexin V-FITC and propidium iodide (PI; Fig. 4B,C). The results showed that induction of mVDAC1 overexpression by a high concentration of tetracycline (2.5 μg/ml) resulted in an increase in cell death of approximately three- to fourfold, but reduced to 1.4-fold or less (Fig. 4B,C) when Δ(26)mVDAC1 was overexpressed. Exposure of cells expressing mVDAC to As₂O₃ resulted in 47.5% annexin V-FITC-positive and PI-positive cells (i.e. late apoptotic cells), as compared with the 22.5% value obtained with cells expressing Δ(26)mVDAC1 (Fig. 4C). These results indicate that in cells expressing Δ(26)mVDAC1, apoptosis is inhibited.

To test whether the apoptotic resistance of Δ(26)mVDAC1-expressing cells is due to an inability of the cells to release cytochrome c, cells expressing native mVDAC1 or Δ(26)mVDAC1 were exposed to As₂O₃, and the localization of cytochrome c, before and after exposure to the apoptosis inducers, was analyzed by immunocytochemistry (Fig. 5A). Although in cells expressing native mVDAC1 the punctuate mitochondrial localization of cytochrome c (i.e. colocalized with MitoTracker, a dye that specifically labels mitochondria in living cells) changed upon exposure to apoptosis stimuli to a diffused cytoplasmic location (Fig. 5B), no such change in cytochrome c distribution was observed in cells expressing Δ(26)mVDAC1 (Fig. 5C). Cytochrome c release into the cytosol was also analyzed by western blotting using anti-cytochrome c monoclonal antibodies (Fig. 5D). STS or cisplatin induced cytochrome c release in cells expressing native mVDAC1 but not those expressing Δ(26)mVDAC1. These findings clearly show that the N-terminal region of VDAC1 is essential for cytochrome c release.

To test whether the resistance to apoptosis conferred by N-terminal-truncated mVDAC1 has a dominant negative effect, Δ(26)mVDAC1 was expressed in three different cell lines expressing endogenous VDAC, and apoptosis was induced by various stimuli. As with cells possessing low level of VDAC1 (Fig. 4), all cell types showed insensitivity to apoptosis induction by reagents that act through activation of cytochrome c release, but not by TNFα, when expressing Δ(26)mVDAC1 (Fig. 6). These results show that although the cells express endogenous hVDAC1, the presence of the N-terminal-truncated mVDAC1 mutant, completely prevented the induction of apoptosis.

HK-I has been shown to interact with VDAC and protect against apoptotic cell death as induced by STS or VDAC1 overexpression (Abu-Hamad et al., 2008; Arzoine et al., 2009; Azoulay-Zohar et al., 2004; Zaid et al., 2005). As shown in Fig. 2B, purified HK-I did not modify the channel conductance of bilayer-reconstituted Δ(26)mVDAC1. Thus, to further assess the interaction of HK-I and Bcl2 with the VDAC1 N-terminal region, HK-I (Fig. 7A) and recombinant Bcl2(Δ23) (Fig. 7B) or, as a control, rabbit IgG, were coupled to surface plasmon resonance (SPR) biosensor chips. Increasing concentrations of a synthetic peptide corresponding to the N-terminal region of VDAC1 were injected onto the sensor chips, and binding to purified immobilized HK-I (Fig. 7C), Bcl2(Δ23) (Fig. 7D) or IgG was monitored. The N-terminal peptide bound to immobilized HK-I or Bcl2(Δ23) in a concentration- and time-dependent manner, rapidly associating with and dissociating from the immobilized proteins. The binding of the VDAC1 N-terminal peptide to HK-I and Bcl2(Δ23) was specific, since no signal was obtained with the IgG-immobilized chip. Moreover, a different VDAC1-based peptide (designated BP), in a loop configuration (18 amino acids, E157 to T174), did not interact with either HK-I (Fig. 7A) or Bcl2(Δ23) (Fig. 7B). The results thus demonstrate the direct and specific interaction of VDAC1-N-terminal peptide with HK-I and Bcl2.

The interaction of the VDAC1-N-terminal peptide with HK-I was further demonstrated at the cellular level by expressing the HK-I–GFP fusion protein alone or together with the N-terminal peptide (1-26 amino acids) or other VDAC1-based peptides (i.e. peptides A, 63W-N76, and B, 157E-T174) and then visualizing HK-I–GFP cellular distribution (Fig. 8A). Confocal fluorescence microscopy showed that in transformed control cells, HK-I–GFP fluorescence was punctuated, as expected for a mitochondria distribution. However, HK-I–GFP fluorescence in cells expressing the N-terminal peptide was diffuse throughout the cytosol (Fig. 8A). Similar results were obtained when peptide A (AP), but not peptide B (BP) was expressed (Fig. 8A). These results indicate that the N-terminal peptide can detach or prevent HK-I binding to the mitochondria.

Next, we tested whether expression of the VDAC1 N-terminal 26-amino-acid peptide in cells overexpressing Bcl2-GFP would interfere with Bcl2- or HK-mediated protection against apoptosis, as induced by STS. Upon exposure of control cells to STS for 5 hours, about 50% of the cells underwent apoptosis (Fig. 8B). By contrast, cells transformed to overexpress Bcl2-GFP (Fig. 8D) had complete protection against STS-induced apoptosis (Fig. 8B). However, in cells overexpressing both Bcl2-GFP and the VDAC1 N-terminal peptide, Bcl2-GFP did not confer protection against STS-induced cell death (Fig. 8B). Since an inducible plasmid for VDAC1 N-terminal peptide expression was used, inhibition of the anti-apoptotic effect of Bcl2-GFP by the peptide was observed only when its expression was induced by tetracycline (Fig. 8B). As with Bcl2, overexpression of HK-I or HK-II (Fig. 8C,E) also conferred protection against cell death induced by STS. This protection was prevented upon expression of the VDAC1 N-terminal peptide (Fig. 8C).

In a previous study, we employed chemical cross-linking with the membrane-permeable cross-linker, EGS, to show that VDAC assumes an oligomeric form in isolated mitochondria (Zalk et al., 2005). Fig. 9 shows that upon apoptosis induction by STS, the dynamic equilibrium between monomeric and oligomeric forms of VDAC was shifted towards the formation of dimers, trimers, tetramers and multimers. VDAC oligomerization was, moreover, dramatically increased (up to sixfold) in both HEK293 and B-16 cells upon exposure to STS (Fig. 9A,B).

VDAC1 overexpression not only encouraged apoptotic cell death (Figs 3 and 4), but also led to enhanced VDAC oligomerization. Upon mVDAC1 overexpression, VDAC oligomers (dimers to hexamers and multimers) became apparent even in the absence of chemical cross-linking (Fig. 9D). The appearance of trimers, tetramers (and possibly hexamers and multimers) upon overexpression of mVDAC1 was, nonetheless, apparent upon chemical cross-linking with EGS (Fig. 9C). Similar results were obtained upon expression of Δ(26)mVDAC1, as visualized using polyclonal anti-VDAC antibodies (Fig. 9C). The Δ(26)mVDAC1 dimers (as expected, with a lower molecular mass) and higher oligomeric states were obtained upon cross-linking with EGS. Using monoclonal anti-VDAC1 antibodies that specifically interact with the N-terminal region of VDAC1, no VDAC cross-linked products were detected in cells expressing Δ(26)mVDAC1 (Fig. 9D). These results suggest that the N-terminal region of VDAC1 is not required for VDAC1 oligomerization but is essential for apoptosis induction involving a mechanism proposed below.

We have shown that cells expressing a N-terminal deletion mutant of VDAC1 are resistant to mitochondria-mediated apoptosis. While the channel still localizes to mitochondria and functions normally, releases of cytochrome c and apoptotic cell death were inhibited. Moreover, the anti-apoptotic proteins HK and Bcl2 were shown to confer protection against apoptosis through interaction with the VDAC1 N-terminal region. Finally, we demonstrated that apoptosis induction is associated with VDAC oligomerizaion. In this Discussion section, we consider these findings in light of a model we propose for VDAC1 and its N-terminal function in apoptosis induction and regulation.

Lying in the OMM, VDAC appears to be a convergence point for a variety of cellular survival and death signals. Here, we have shown that Δ(26)mVDAC1 functions as a channel and, like native mVDAC1, could restore the growth and oxygen consumption of cells silenced for hVDAC1 expression by hVDAC1-shRNA, suggesting that the N-terminal region of VDAC1 is not essential for VDAC1 transport activity and, hence, for mitochondrial energy production.

However, the N-terminal α-helix of VDAC1 is required for cytochrome c release and apoptosis when induced by various stimuli (Figs 3, 4, 5, 6). Overexpression of human, murine, yeast or rice VDAC1, VDAC1-GFP and mutated VDAC1 was shown to induce apoptotic cell death, regardless of the cellular host (Abu-Hamad et al., 2006; Bae et al., 2003; Godbole et al., 2003; Muller et al., 2000; Zaid et al., 2005), suggesting that VDAC1 is a conserved mitochondrial element of the death machinery. In accordance with previous results, overexpression of mVDAC dramatically triggers cell death. By contrast, overexpression of Δ(26)mVDAC1, although distributed to the mitochondria (Fig. 1B), triggered no such apoptotic cell death (Figs 3 and 4). Moreover, induction of apoptosis by STS, curumin, cisplatin and As₂O₃, each relying on different pathways (Duvoix et al., 2005; Schmitz et al., 2000; Zaid et al., 2005; Zhang et al., 2005), all ultimately led to cytochrome c release (Fig. 5) and cell death (Figs 3, 4 and 6) in cells expressing native but not Δ(26)mVDAC1. Thus, by expressing N-terminal-truncated VDAC1, we have constructed an apoptosis-resistant cell line, namely, a line for which a highly active apoptosis stimulus did not induce cytochrome c release or cell death.

Recently, it has been shown that overexpression of hVDAC1 in COS cells increased the number of cells with depolarized mitochondria, a phenomenon associated with apoptosis, and that this effect could not be obtained upon overexpression of (Δ19)hVDAC1 (De Pinto et al., 2007). These findings are consistent with our results showing that the VDAC1 N-terminal region is necessary for mitochondria-mediated apoptosis.

HK-I and HK-II are highly expressed in cancer cells, with over 70% of the protein being bound to mitochondria (Arora et al., 1990; Pedersen et al., 2002), or more precisely, to VDAC (Abu-Hamad et al., 2008; Arzoine et al., 2009; Azoulay-Zohar et al., 2004; Pastorino et al., 2002). Similarly, Bcl2 is overexpressed in more than half of all human cancers, with its broad expression in tumors being linked to a conferring of resistance to chemotherapy-induced apoptosis (Ding et al., 2000). Both HK (Abu-Hamad et al., 2008; Arzoine et al., 2009; Azoulay-Zohar et al., 2004; Pastorino et al., 2002) and members of the Bcl2 family (Pastorino et al., 2002; Shi et al., 2003; Shimizu et al., 1999; Tsujimoto and Shimizu, 2002) are though to act via interaction with VDAC.

Recently, those VDAC1 amino acid residues important for HK-I binding and conferring the anti-apoptotic effects of the protein were identified (Abu-Hamad et al., 2008). Clearly, the results presented here indicate that the N-terminal region of VDAC1 is essential for the HK-mediated reduction in channel conductance (Fig. 2B) and for HK-I and HK-II protection against apoptosis (Fig. 8C). Both HK-I and Bcl2 interact with a synthetic peptide corresponding to the VDAC1 N-terminus, as revealed using SPR technology (Fig. 7A,B). Moreover, expression of this N-terminal peptide in cells overexpressing HK-I, HK-II or Bcl2 prevented their protective effects against apoptosis, as induced by STS (Fig. 8). Our findings suggest that the expressed peptide competes with VDAC1 for binding to Bcl2, HK-I and HK-II and thus interferes with their binding to VDAC1, an interaction that prevents apoptosis. It has been shown that the N-terminal region of VDAC1 modulates the accessibility of VDAC1 to Bax and Bcl-xL (Shi et al., 2003). Our observations, therefore, suggest that the anti-apoptotic activity of HK and Bcl2 is mediated through interactions with the VDAC1 N-terminal region, pointing to the N-terminal region of VDAC1 as participating in the regulation of apoptosis by anti-apoptotic proteins. This novel mechanism is presented in Fig. 10.

Numerous studies have focused on the importance of the N-terminal α-helical segment in channel function. These studies have yielded a wide range of predictions as to the functional disposition of this domain, ranging from it forming a segment of the channel wall to acting as the voltage sensor (Colombini et al., 1996; Koppel et al., 1998), to regulating the conductance of ions and metabolites passing through the VDAC1 pore, based on its position in the pore (Bayrhuber et al., 2008; Ujwal et al., 2008). In addition, conformational instability of the N-terminal part of hVDAC1 was suggested, with the domain being predicted to switch between different conformations, during which hydrogen bonds within the N-terminus and the strands of the barrel are transiently formed and broken (Bayrhuber et al., 2008; Ujwal et al., 2008). Therefore, the N-terminal domain may adopt different conformations, depending on various factors, such as protein-protein interactions, ion interactions and lipid environment. Re-arrangement and coordination of the N-terminal helix is suggested here to be involved in VDAC1 function in apoptosis.

The precise mechanisms regulating cytochrome c release, a key initial step in the apoptotic process, remain unknown. Similarly, the molecular architecture of the cytochrome-c-conducting channel has yet to be determined. We hypothesize that a protein-conducting channel is formed within a VDAC1 homo-oligomer or a hetero-oligomer containing VDAC1 and pro-apoptotic proteins (Shoshan-Barmatz et al., 2008; Shoshan-Barmatz et al., 2006; Zalk et al., 2005; Zheng et al., 2004). Indeed, further support for the supramolecular organization of VDAC, comprising monomers to tetramers, hexamers and higher oligomers, has come from recent studies applying atomic force microscopy (Goncalves et al., 2007; Hoogenboom et al., 2007) and NMR (Bayrhuber et al., 2008; Malia and Wagner, 2007). It was also demonstrated that VDAC oligomerization is encouraged in the presence of cytochrome c (Zalk et al., 2005). The function of this VDAC oligomerization in apoptosis is demonstrated here, with the demonstration that the supramolecular assembly of VDAC in cultured cells is highly enhanced upon apoptosis induction (Fig. 9). Apoptosis induction by STS or by overexpression of mVDAC1 resulted in several-fold increase in VDAC oligomerization. Similarly, the apoptosis-inducing effect of As₂O₃ was attributed to an induction of VDAC homodimerization that was prevented by overexpression of the anti-apoptotic protein, Bcl-XL (Yu et al., 2007; Zheng et al., 2004). However, since (Δ26)mVDAC1 is able to undergo oligomerization (Fig. 9) but not to induce apoptosis (Figs 3, 4 and 6), it appears that the N-terminal region is not required for VDAC oligomerization but is important for cytochrome c release and subsequent apoptosis.

Together, these observations can be explained by the following model (Fig. 10). Once formed, an oligomeric VDAC1 intermolecular pore can serve as a conduit for a protein. Oligomers of VDAC1 β-barrels would, however, form a hydrophobic channel that would not allow charged proteins, such as cytochrome c to pass. However, if the amphipathic N-terminal region (containing three positive and two negative residues, but with an overall charge of zero) of each VDAC1 molecule is moved to line the external surface of the β-barrel, a hydrophilic pore would form (Fig. 10). In such a scenario, the now hydrophilic pore, lined by several α-helix N-terminal regions, would permit protein transport across the OMM (Fig. 10B). In this regard, it should be noted that a recent study suggested that the N-terminus of VDAC in aqueous medium is uncoiled, requiring a hydrophobic and negatively charged environment to gain its α-helical structure (De Pinto et al., 2007). With N-terminally truncated VDAC1, such a hydrophilic protein-conducting pore could not be formed, and hence, cytochrome c would not be released (Fig. 5). The finding that the VDAC1 N-terminal α-helix specifically interacts with cytochrome c (Mannella et al., 1987) supports the proposed model.

The ability of a synthetic peptide, corresponding to the N-terminal region of VDAC1, to bind both HK-I and Bcl2 (Fig. 7), to detach mitochondrial-bound HK-I–GFP (Fig. 8A) and to prevent Bcl2-, HK-I- and HK-II-mediated protection against apoptosis (Fig. 8B,C), suggests that the function of the α-helix N-terminal portion of VDAC1 in apoptosis can be modulated by anti-apoptotic proteins. According to our proposal, the anti-apoptotic activities of HK-I, HK-II and Bcl2 result from their interaction with the mobile N-terminal region of VDAC1 and preventing its reorientation into the cytochrome c-releasing pore (Fig. 10C).

Taken together, the requirement for the VDAC1 N-terminal region for cytochrome c release and apoptosis induction, as well as its interaction with anti-apoptotic proteins, point to the VDAC1 N-terminal region as being a key element in regulating apoptosis and the target for anti-apoptotic proteins (Fig. 10).

Finally, mitochondria-mediated apoptosis plays a crucial role in the pathophysiology of many diseases, including heart attack and stroke, neurodegenerative disorders, such as Parkinson's disease and Alzheimer's disease, as well as mitochondrial encephalomyopathies and cancer (Mattson, 2000; Thompson, 1995). Although some diseases, such as cancer, are caused by suppression of apoptosis, other diseases, such as Alzheimer's disease, are characterized by increased apoptosis. The VDAC1 N-terminal region offers a tempting target for the development of therapeutic agents designed to inhibit apoptosis. In cancer therapeutics, targeting VDAC1 N-terminal peptide to tumor cells overexpressing anti-apoptotic proteins, such as Bcl2 and HK-I, would minimize the self-defense mechanisms of the cancer cells, thereby promoting apoptosis and increasing sensitivity to chemotherapy.

Staurosporine (STS), poly-D-lysine (PDL) and propidium iodide were purchased from Sigma (St Louis, MO). Mito-Tracker red dye CMXPos was from Molecular Probes. Annexin V-FITC kit was purchased from Beender MedSystem. Monoclonal anti-VDAC antibodies raised to the N-terminal region of VDAC1 came from Calbiochem-Novobiochem (Nottingham, UK), and rabbit polyclonal anti-VDAC antibodies, prepared against residues 150-250 of human (h)VDAC1, were from Abcam. Monoclonal anti-cytochrome c antibodies for immunoblots (1:2000; cat. no. 556433) and for immunostaining (1:800; cat. no. 556432) were obtained from BD Biosciences Pharmingen. Monoclonal anti-GFP antibodies were obtained from Santa Cruz Biotechnology. Metefectene was purchased from Biotex (Munich, Germany). RPMI 1640 and DMEM growth media and the supplements, fetal calf serum, L-glutamine and penicillin-streptomycin were purchased from Biological Industries (Beit Haemek, Israel). Blasticidin and zeocin were purchased from InvivoGen (San Diego, CA). Puromycin was purchased from ICN Biomedicals (Eschwege, Germany).

Vectors expressing N-terminally truncated murine (m)VDAC1 or the N-terminal 26-amino-acid peptide were generated by PCR and cloned into the BamHI-NotI or BamHI-EcoRI sites of the tetracycline-inducible pcDNA4/TO vector (Invitrogen), using mVDAC1 cDNA as template and the following primers: 5′-CGGGATCCATGATAAAACTTGATTTGAAAACG-3′ (F) and 5′-GCGGCCGCTTATGCTTGAAATTCCAGTCC-3′ (R) for (Δ26)mVDAC1 and 5′-GCCAGATCTATGGCTGTGCCACCCACGTA-3′ (F) and 5′-CGAATTCTCATAAGCCAAATCCATAGCCCTT-3′ (R) for the 26 amino acid peptide.

The pcDNA3–HK-II and pcDNA3–HK-I plasmids were kindly provided by J. E. Wilson (Michigan State University). HK-I–GFP was generated as described previously(Abu-Hamad et al., 2008).

For construction of C-terminally truncated Bcl2, Bcl2-GFP cDNA was sub-cloned from plasmid pcDNA3 using the primers: 5′-CTCCATGGCGATGGCGCACGCTGGGAGAACG-3′ (F) containing a NcoI restriction site and 5′-CGGAATTCTTACAGAGACAGCCAGGAGAAATC-3′ (R), containing a stop codon and an EcoRI restriction site. The generated product, encoding Bcl2(Δ23), was then cloned into a pHISparallel vector for expression in E. coli BL21 cells.

MCF7 cells are a human breast carcinoma cell line, HEK293 cells are a transformed primary human embryonal kidney cell line, T-REx293 cells are tetracycline repressor-expressing HEK293 cells and HeLa cells are a cervical cancer-derived cell line. These and hVDAC1-shRNA T-REx293 cells were grown in Dulbecco's modified Eagle's medium (Biological Industries), supplemented with 10% fetal calf serum, 2 mM L-glutamine, 100 IU/ml penicillin, 100 μg/ml streptomycin and nonessential amino acids (all from Biological Industries) and maintained in a humidified atmosphere, at 37°C with 5% CO₂. hVDAC1-shRNA T-REx293 cells stably expressing the pSUPERretro vector encoding shRNA-targeting hVDAC1 were transfected with plasmids pcDNA4/TO-mVDAC1 or pcDNA4/TO-Δ(26)mVDAC1 and grown as described previously (Abu-Hamad et al., 2006). T-REx293 cells were transiently co-transfected with plasmids pEGFP-Bcl2, pcDNA3.1–HK-I or pcDNA3.1–HK-II and pcDNA/4TO encoding the N-terminal 26 amino acids and grown for 48 hours with tetracycline (1 μg/ml), prior to a 5 hour exposure to STS (1.25 μM). Overexpression of mVDAC1- or Δ(26)mVDAC1-pcDNA4/TO was induced by tetracycline (2.5 μg/ml) for 96-100 hours. Apoptosis was analyzed by Acridine Orange and ethidium bromide staining (Zaid et al., 2005) or by FACS.

Oxygen utilization was measured polarographically with a thermostatically controlled (37°C) Clark oxygen electrode (Strathkelvin 782 Oxygen System). T-REx293, hVDAC1-shRNA T-REx293 cells expressing stable native or Δ(26)mVDAC1 (2×10⁶) were washed, suspended in a serum-free DMEM medium [1000 mg/l glucose pyridoxine, HCl and NaHCO₃, without phenol red (Sigma D-5921)], supplemented with L-glutamine (4 mM) and L-pyruvate (1 mM), and then added to a 1 ml water-jacketed chamber. Cells respiration was determined in absence and presence of rotenone (5 μM, added directly into the respiration chamber) for up to 10 minutes.

Cells (5×10⁴) were grown on PDL-coated coverslips in a 24-well dish and transfected with plasmid pEGFP–HK-I alone or with DNA encoding the N-terminal peptide (NP), peptide A (amino acids 63W-78N, AP, previously named LP1) and peptide B (157E-174T, BP, previously named LP3) that were generated and cloned into the tetracycline-inducible pcDNA4/TO vector (Invitrogen), as described previously (Arzione et al., 2009). Peptides A and B were flanked by a tryptophan zipper motif, i.e. the SWTWE amino acid sequence, at the N-terminus of the peptide, and the KWTWK sequence at the C-terminus, to form a loop (Arzione et al., 2009).

Cells were fixed for 15 minutes with 4% paraformaldehyde and rinsed with PBS for 30 minutes prior to imaging by confocal microscopy (using an Olympus 1X81 microscope).

For immunoblot analysis, cells expressing native or Δ(26)mVDAC1 were grown for 72 hours, exposed to STS, cisplatine or As₂O₃ and analyzed for cytochrome c release (Abu-Hamad et al., 2008). Aliquots of the cytosolic fraction were separated on Tris-Tricine gels (13% polyacrylamide) and immunoprobed using anti-cytochrome c antibodies. To rule out mitochondrial contamination, the cytosolic fraction was immunoprobed for VDAC using anti-VDAC antibodies, and then with HRP-conjugated anti-mouse IgG as a secondary antibody (1:10,000). An enhanced chemiluminiscent substrate (Pierece Chemical, Rockford, IL) was used for detection of HRP.

For immunostaining, cells (5×10⁴) were grown on poly-D-lysine (PDL)-coated coverslip in a six-well plate. After 48 hours, cells were treated with As₂O₃ (60 μM, 24 hours). Then, the cells were treated with MitoTracker red dye (25 nM) for 15 minutes in an incubator (37°C, 5% CO₂), washed three times with PBS, fixed for 15 minutes with 4.0% paraformaldehyde and rinsed three times with PBS. Cells were treated with methanol for 15 minutes (–20°C), then incubated for 1 hour with blocking solution (5% normal goat serum (NGS) and 0.2% Triton X-100), followed by incubation with anti-cytochrome c antibodies (cat. no. 556432, 1:800, diluted in blocking solution) for 1 hour. After washing with PBS, cells were incubated for 1 hour with Alexa-Fluor-488-conjugated antibodies (1:500, diluted in blocking solution) and washed with PBS. Coverslips were mounted on glass slides with mounting medium and sealed with nail polish. Cells were visualized by confocal microscopy (Olympus 1X81).

Apoptotic cells were estimated using a commercial annexin V-FITC and PI detection kit or following PI uptake and FACS analysis (FACScan, Beckton-Dickinson, San Jose, CA) and ModFIT-lt2.0 software.

mVDAC1 and Δ(26)mVDAC1 were expressed in the Saccharomyces cerevisiae por1-mutant strain M22-2 (Zaid et al., 2005) and purified from mitochondria (Shoshan-Barmatz and Gincel, 2003). HK-I was purified from rat brain mitochondria (Azoulay-Zohar et al., 2004). Bcl2(Δ23) was expressed in E. coli BL21, induced with isopropyl β-D-1-thiogalactopyranoside. Cells were grown for 30-60 minutes at 20°C, sonicated, and the soluble fraction (5-10% of the total) was purified by Ni-NTA chromatography.

For cross-linking, T-REx293, HEK293 or B-16 cells (2-3 mg/ml) were incubated with different concentrations of EGS [ethylene glycol bis(succinimidylsuccinate)] in PBS, pH 8.3 (15 minutes, 30°C). Samples (50 μg) were subjected to SDS-PAGE and immunoblotting using anti-VDAC antibodies.

Reconstitution of purified native mVDAC1 and Δ(26)mVDAC1 into a planar lipid bilayer (PLB), channel recording, and analysis were carried out as described previously (Gincel et al., 2001).

Purified rat brain HK-I, recombinant Bcl2 and rabbit IgG were immobilized on a CM5 sensor surface according to the manufacturer's instructions. N-terminal VDAC1 peptide was diluted in running buffer [150 mM NaCl, 0.005% Tween 20, 4% (v/v) DMSO, PBS, pH 7.4] and injected onto the sensor chip at varying concentrations (flow rate of 40 μl/minute, 25°C). Experiments were carried out using the SPR ProteOn-XPR36 system (Bio-Rad). Peptides were synthesized by the oligonucleotide and peptide synthesis unit, Weizmann Institute, Israel.