Voltage-dependent Anion Channel 1-based Peptides Interact with Hexokinase to Prevent Its Anti-apoptotic Activity*

In brain and tumor cells, the hexokinase isoforms, HK-I and HK-II, bind to the voltage-dependent anion channel (VDAC) in the outer mitochondrial membrane. The VDAC domains interacting with these anti-apoptotic proteins were recently defined using site-directed mutagenesis. Now, we demonstrate that synthetic peptides corresponding to the VDAC1 N-terminal region and selected sequences bound specifically, in a concentration- and time-dependent manner, to immobilized HK-I, as revealed by real time surface plasmon resonance technology. The same VDAC1-based peptides also detached HK bound to brain or tumor-derived mitochondria. Moreover, expression of the VDAC1-based peptides in cells overexpressing HK-I or HK-II prevented HK-mediated protection against staurosporine-induced release of cytochrome c and subsequent cell death. One loop-shaped VDAC1-based peptide corresponding to a selected sequence and fused to a cell-penetrating peptide entered the cell and prevented the anti-apoptotic effects of HK-I and HK-II. This peptide detached mitochondrial-bound HK better than did the same peptide in its linear form. Both cell-expressed and exogenously added cell-penetrating peptide detached mitochondrial-bound HK-I-GFP. These results point to HK-I and HK-II as promoting tumor cell survival through binding to VDAC1, thereby inhibiting cytochrome c release and apoptotic cell death. Moreover, VDAC1-based peptides interfering with HK-mediated anti-apoptotic activity may potentiate the efficacy of conventional chemotherapeutic agents.

located at the outer membrane of mitochondria, where they gain preferential access to mitochondrially generated ATP (6). In this manner, HK-I and HK-II drive the high glycolytic rates typical of tumor cells, even under aerobic conditions (1,3). Such elevated levels of mitochondria-bound HK is suggested to play a pivotal role in promoting cell growth and survival in rapidly growing, highly glycolytic tumors and in protecting against mitochondria-mediated cell death (1,4,7). The elevated levels of HK-I and HK-II allow tumor cells to evade apoptosis, thereby allowing proliferation to continue (4). Indeed, overexpression of mitochondria-bound HK in the tumor-derived cell line U-937, in T-REx-293, or in vascular smooth muscle cells suppressed staurosporine-induced apoptotic cell death (8,9). A decrease in apoptotic cell death and concomitant increase in cell proliferation have also been reported to be induced upon HK-II expression in NIH-3T3 (7), rat 1a fibroblasts (10), and WEHI 7.1 cells (11). In addition, binding of HK-II to mitochondria inhibited Bax-induced cytochrome c release and apoptosis (12).
Mitochondria-mediated apoptosis results in the efflux of a number of potential apoptotic regulators, such as cytochrome c, to the cytosol, triggering caspases activation and cell destruction. The molecular mechanism of HK-mediated protection against cell death involves HK-I and HK-II binding to the outer mitochondrial membrane protein, the voltage-dependent anion channel (VDAC) (8,9,(13)(14)(15)(16). Indeed, VDAC was shown to be up-regulated in cancer cells exhibiting elevated binding to HK-I and/or HK-II (17). Also, it has been recently shown that acute viruses, such as HEV, can protect infected cells from apoptotic death via enhanced VDAC and HK expression (18).
VDAC is a key channel protein that integrates cellular energy metabolism. VDAC has also been recognized as a key protein in mitochondria-mediated apoptosis, as the proposed target for the pro-and anti-apoptotic Bcl2 family of proteins and due to its function in the release of apoptotic proteins (19 -21). Purified HK-I has been shown to directly interact with purified VDAC reconstituted into a planar lipid bilayer, leading to channel closure (8). HK-I also prevented opening of the permeability transition pore and suppresses the release of cytochrome c, thus inhibiting the mitochondrial phase of apoptosis (8), and thereby contributing to cell survival (12,22).
Recent studies have shown that the interaction of HK with VDAC can be regulated via VDAC phosphorylation (23). It has also been shown that Akt activation partly inhibits apoptosis by promoting the binding of HK-II to mitochondria (23). This effect of Akt was shown to be mediated by negative regulation of glycogen synthase kinase 3␤ activity. Glycogen synthase kinase 3␤-phosphorylated VDAC is unable to bind HK-II, shifting the equilibrium toward cytosolic HK and thus allowing apoptosis induction (23). These results suggest that HK-I or HK-II, via their interaction with VDAC1, prevent key events in mitochondria-mediated apoptosis.
Despite these advances, the precise VDAC1 sequences interacting with HK-I or HK-II remain unknown. Recently, using site-directed mutagenesis, we identified VDAC1 domains that interact with HK-I (24), located in two cytoplasmic domains, according to previously proposed VDAC1 topology models (25,26). However, according to recently published NMR-based VDAC1 structure (27,28) only part of these domains are exposed to the cytosol. In addition, the amphipathic N-terminal domain of VDAC1 (24,27,28) was identified as the HK-I interaction site. 3 These VDAC1 domains are involved in HK-Imediated protection against cell death via the inhibition of cytochrome c release. In this study, we have used VDAC1-based synthetic and recombinant peptides to define the boundaries of VDAC1 sequences essential for HK-I and HK-II binding, and show their abilities to disturb HK mediating protection against cell death.

EXPERIMENTAL PROCEDURES
Materials-Glucose 6-phosphate, Hepes, lactate dehydrogenase, leupeptin, mannitol, phenylmethylsulfonyl fluoride, propidium iodide, pyruvate kinase, staurosporine, sucrose, and Tris were purchased from Sigma. A Cibacron blue HiTrap TM Blue HP column was purchased from Amersham Biosciences. Monoclonal anti-VDAC antibodies came from Calbiochem-Novobiochem (Nottingham, UK). Monoclonal antibodies against actin and polyclonal goat antibodies against HK-I and HK-II were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Monoclonal anti-cytochrome c antibodies were obtained from BD Biosciences. Horseradish peroxidase-conjugated anti-mouse and anti-goat antibodies were obtained from Promega (Madison, WI).
Plasmids and Site-directed Mutagenesis-Five mVDAC1 peptide-encoding inserts were generated by standard PCR and cloned into the BamHI/EcoRI sites of the tetracycline-inducible pcDNA4/TO vector (Invitrogen), using mVDAC1 cDNA as a template and the primers indicated in Table 1. The N-terminal-, LP1-, LP2-, LP3-, and LP4-coding VDAC1 sequences ( Fig. 1) were generated using primers F-N and R-N and F-1 and R-1 to F-4 and R-4, respectively ( Table 1).
The four VDAC1 loop-based peptides, LP1 to LP4, were flanked by a tryptophan zipper motif, namely, the SWTWE amino acid sequence at the N terminus of the peptide and KWTWK sequence at the C terminus. Correct constructs were confirmed by sequencing. The pcDNA3-HK-II and pcDNA3-HK-I plasmids were kindly provided by J. E. Wilson (Michigan State University), and the HK-I-GFP fusion protein, in which GFP was connected to the HK-I C-terminal, was generated as described previously (24). Construction of the native and mutated mVDAC1 vectors, mVDAC1-, mVDAC1/E72Q-, mVDAC1/E202Q-, and mVDAC1/K109L-pcDNA4/TO, was described previously (24).
Peptides Synthesis-The synthetic N-terminal, LP1, LP2, LP3, and LP4 VDAC1-based peptides were synthesized by the Oligonucleotides and Peptide Synthesis Unit, Weizmann Institute, Rehovot, Israel. The LP4-Antp looped-shaped peptide, including the cell penetrating peptide, Antp, the peptide Antp, and other VDAC1-based peptides, were synthesized by GL Biochem (Shanghai, China). The N-terminal and LP4 peptides were water-soluble, whereas the other peptides were dissolved in Me 2 SO.
Cancer Cell Lines and Tumors-CT26 is a murine colon carcinoma cell line, B16 is a murine melanoma cell line, BCL1 is a murine B cell leukemia cell line, MCF7 represents a human breast carcinoma cell line, and Molt4 is a human T lymphoblastic leukemia cell line. All were purchased from ATCC (Rockville, MD) or provided by Dr. Eliezer Flescher (Tel Aviv University). T-REx-293 cells correspond to a transformed primary human embryonal kidney cell line expressing the tetracycline repressor (Invitrogen). Cell growth was maintained in a humidified atmosphere, at 37°C with 5% CO 2 . CT26, B16, MCF-7, and T-REx-293 cells were maintained in Dulbecco's modified Eagle's medium (Biological Industries), supplemented with 10% fetal calf serum, 2 mM L-glutamine, 100 units/ml penicillin, 100 g/ml streptomycin, 1 mM sodium pyruvate and nonessential amino acids (all from Biological Industries). Molt4 and BCL1 cells were maintained in RPMI 1640 medium (Biological Industries), supplemented with 10% fetal calf serum, 2 mM L-glutamine, 100 units/ml penicillin, and 100 g/ml streptomycin. Peripheral blood lymphocytes from venous blood of healthy donors were prepared by Ficoll-Hypaque (GE Healthcare) density gradient centrifugation, as previously described (30).
To generate BCL1 leukemia-bearing C57BL BALB/c mice (Tel-Aviv University), animals were inoculated intraperitoneally with BCL1 cells freshly extracted from BCL1-bearing mice. Four weeks later, the mice were killed under anesthesia, their  FEBRUARY 6, 2009 • VOLUME 284 • NUMBER 6 spleens were removed, and the cells were extracted by spleen dispersion. Red blood cells were disrupted using red blood cell lysis buffer (Sigma). Cell Transfection-Logarithmically growing T-REx-293 cells were transiently co-transfected with plasmids pcDNA3-HK-II or pcDNA3-HK-I alone or with one of the following vectors: each of the mVDAC1-peptide-encoding pcDNA4/TO plasmids or the native or mutated mVDAC1-encoding pcDNA4/TO plasmid (or the corresponding control plasmid). Transfections were performed using FuGENE HD transfection reagent (Roche), according to the manufacturer's instructions. After 48 h growth in the presence of tetracycline (1 g/ml), cells were subjected to the desired treatment.

VDAC1-based Peptides Interact with HK
Acridine Orange/Ethidium Bromide Staining of Cells-To determine the extent of apoptosis, cells were subjected to staining with 100 g/ml acridine orange and 100 g/ml ethidium bromide (EtBr) in PBS (see figure legends). Isolation of Rat Brain, Mammalian Cells, and Tumor-derived Mitochondria-Rat brain, mammalian cells, and tumor-derived mitochondria were prepared by mechanical cell homogenization and differential centrifugation, as previously described (31).
Detachment of Mitochondriabound HK and HK Activity Assay-To induce detachment of mitochondria-bound HK, increasing amounts of synthetic peptides (0.0025-0.2 mM) were added to isolated mitochondria (1.5-2 mg/ml) and released HK was assayed following its activity (32) or by immunoblotting using anti-HK-I antibodies.
Rat Brain HK-I Purification-HK-I was purified from rat brain mitochondria (8), with the modification of using a Cibacron Blue HiTrap TM Blue HP column and a AKTAbasic purifier chromatography system (Amersham Biosciences).
Real-time Surface Plasmon Resonance (SPR)-SPR, using the ProteOn-XPR36 system (Bio-Rad) was employed to study the interaction of VDAC1-based synthetic peptides with purified HK-I. Purified HK-I and rabbit IgG were immobilized onto a GLM sensor surface, according to the manufacturer's instructions. The peptides were diluted in running buffer (150 mM NaCl, 0.005% Tween 20, 4% (v/v) Me 2 SO, 10 mM PBS, pH, 7.4) and injected onto the sensor chip at varying concentrations (flow rate, 40 l/min). Response units were monitored using the ProteOn imaging system and related software tools. Signals were normalized using appropriate controls.

VDAC1-based Peptides Interact with HK
with a cut off of 1000 Da (Cellu. Sep H1, Membrane Filtration Products, Seguin, TX). T-REx-293 cells were incubated for 90 min with 20 M FITC-labeled-LP4-Antp or Antp peptide in serum-free Dulbecco's modified Eagle's medium, washed with PBS, fixed with 4% formaldehyde, and viewed under a Zeiss LSM 510 meta-confocal microscope.

LP4-Antp and Antp Peptide Treatment of HK-I Overexpressing Cells
-T-REx-293 cells were transiently transfected with plasmid pcDNA3-HK-I or control plasmid pcDNA3. 48 h after transfection, cells were incubated for 2 h with 20 M of the LP4-Antp or the Antp peptide in serum-free Dulbecco's modified Eagle's medium and then exposed to STS (1.25 M) for 2 h. Apoptosis was analyzed quantitatively by acridine orange/EtBr staining, as described above.
Confocal Microscopy-T-REx-293 cells (5 ϫ 10 4 ) were grown on poly-D-lysine-coated coverslips, and transfected with plasmid pEGFP-HK-I alone or with plasmid pEGFP-HK and pcDNA4/TO-LP4, stabilized with the tryptophan zipper motif. After 48 h, control cells were incubated for 90 min with a solution containing 0.4% Me 2 SO or LP4-Antp (15 M). Cells were fixed for 15 min using 4% paraformaldehyde prepared in PBS, rinsed for 30 min in PBS, and cell imaging was carried out by confocal microscopy (Olympus 1X81).

RESULTS
We have demonstrated that the pro-survival protein HK-I, expressed at high levels in many types of cancer (1, 2, 5), interacted with VDAC1 and protected against cell death as induced by STS or VDAC1 overexpression (8,9). We also demonstrated that E72Q-mVDAC1 no longer retains its capacity to bind HK-I (9, 24). The domains and amino acid residues involved in the interaction of VDAC1 with HK-I were also defined (9, 24). These domains, according to previously proposed VDAC1 membrane topology models (25,26), are exposed to the cytosol. However, according to recently published NMR-based recombinant human VDAC1 structure (27,28), these sequences are only partially exposed to the cytosol (Fig. 1). Based on certain mutations in VDAC1 found to diminish HK-I protection against apoptosis (24), VDAC1-based peptides were synthesized and their interactions with HK-I and HK-II, as well as the effects of their expression on HK-mediated anti-apoptotic activity, were studied.
HK-II Overexpression Protects against Cell Death-The antiapoptotic effect of HK-II was tested by its overexpression in T-REx-293 cells expressing native or mutated mVDAC1 and induction of apoptosis by STS. STS-induced apoptotic cell death was reduced by 64% in cells overexpressing HK-II and native mVDAC1, as compared with the values obtained in control cells. However, the protective effect of HK-II was abolished in cells overexpressing E72Q-, E202Q-, or K109L-mVDAC1 (Fig. 2, A and B). The level of HK-II in cells transfected to express the protein was about 6-fold higher than in control cells (Fig. 2C). These results indicate that like HK-I (9, 24), the interaction of HK-II with VDAC1 involves glutamate and lysine residues located in 2 different regions of VDAC1. These results also show that although the cells express endogenous hVDAC1, the presence of mVDAC1 mutants, nonetheless, completely prevented the anti-apoptotic effect of HK-II. This is in agree-ment with previous studies (9) showing that no protection mediated by HK-I or ruthenium red against STS-induced apoptosis was observed in cells expressing E72Q-VDAC1, although endogenous VDAC1 was present in these cells. This indicates that E72Q-VDAC1 possesses a dominant negative effect and implies that VDAC1 molecules are capable of intermolecular interactions to form a homo-oligomer, involved in apoptosis induction (9).
VDAC1-based Synthetic Peptides Interact with HK-I-Five VDAC1-based peptides (labeled in gray in Fig. 1) (LP1, LP2, LP3, LP4, and the N terminus) were synthesized and their interactions with purified rat brain HK-I was demonstrated using SPR technology (Fig. 3). Purified HK-I or rabbit IgG were coupled to a SPR biosensor chip. Increasing concentrations (20 to 300 M) of the VDAC1-based peptides were injected onto the T-REx-293 cells were transiently co-transfected with plasmid pcDNA3-HK-II and pcDNA4/TO encoding native, E72Q-, E202Q-, or K109L-mVDAC1 (or the corresponding control plasmids) and were grown for 48 h in the presence of tetracycline (1 g/ml, to induce their expression) prior to exposure to STS (1.25 M) for 4 h. A, apoptotic cell death was visualized by acridine orange/ethidium bromide staining ( Fig. 2A). Fluorescent images were viewed under an Olympus IX51 microscope and recorded with an Olympus DP70 camera, using a SWB filter (scale bars ϭ 50 m). In each independent experiment, an early apoptotic state, represented as degraded nuclei (stained by acridine orange) and a late apoptotic state, as revealed by the presence of degraded nuclei (stained with both acridine orange/EtBr), as well as nonapoptotic cells, were counted (ϳ250 cells were counted for each treatment). B, quantitative analysis of apoptosis was performed by analysis of variance and t test with p Ͻ 0.001 (***) being considered statistically significant. Data shown in B are the mean Ϯ S.E. of four independent experiments. C, Western blot analysis of HK-II and mVDAC1 expression levels in the different transfected cells using polyclonal anti-HK-II and monoclonal anti-VDAC antibodies, respectively. Actin level assessment confirmed that equal amounts of cells were used. FEBRUARY 6, 2009 • VOLUME 284 • NUMBER 6 sensor chips and binding to HK-I was monitored. The N-terminal, LP1, and LP4 peptides strongly bound to immobilized HK-I in a concentration-and time-dependent manner (Fig. 3). By contrast, two other VDAC1-based peptides, i.e. LP2 and LP3, did not interact with HK-I (Fig. 3). Binding of the LP1, LP4, and N-terminal peptides to HK-I was specific, because no signal was obtained with the IgG-immobilized chip (Fig. 3A, data shown only for LP1).

VDAC1-based Peptides Interact with HK
All HK-I-interacting VDAC1-based peptides associated relatively rapidly with the immobilized HK-I. However, whereas the LP1 and the N-terminal peptides showed rapid dissociation, the LP4 peptide showed slow dissociation from the immobilized HK-I. The apparent binding affinities of the LP1, LP4, and N-terminal peptides to HK-I were derived from dose-response curves obtained using steady-state values and calculated to be 65, 70, and 140 M for LP1, LP4, and N-terminal peptides, respectively (Fig. 3D, data shown only for the N-terminal peptide). The results thus demonstrate direct and specific interaction of selected VDAC1-based peptides with HK-I.
VDAC1-based Synthetic Peptides Detach Bound HK from Mitochondria-HK-I and HK-II are highly expressed in cancer cells, where over 70% of the protein is bound to mitochondria (3,4). Brain mitochondria also express very high levels of mitochondria-bound HK-I (34). The effects of the synthetic VDAC1-based peptides on HK-I bound to rat brain mitochondria and on HK-I/HK-II bound to mitochondria iso-lated from BCL1 leukemia cells were tested. The LP1 peptide, containing the Glu 72 residue, was able to release mitochondria-bound HK from both rat brain and BCL1 mitochondria, as reflected in the appearance of HK activity in the supernatant (Fig. 4, A-C). The N-terminal peptide was also able to release HK from mitochondria from both sources, but to a lesser extent than LP1 (Fig. 4, A, B, and D). HK was released from rat brain mitochondria by both the LP1 and N-terminal peptides in a dose-response manner. Apparent affinities (K 0.5 ) of 2 and 20 M were derived for the LP1 and N-terminal peptides from the saturation curve fits (Fig. 4, C and  D). The LP2 and LP4 peptides also caused release of a small fraction of mitochondria-bound HK-I (Fig.  4A). By contrast, LP3 peptide did not detach HK from either mitochondrial preparation. The observations that the peptides induced detachment of mitochondriabound HK suggest that a dynamic equilibrium exists between free and mitochondria-bound HK.
Expression of VDAC1-based Peptides in Cells Overexpressing HK-I or HK-II Prevents Their Anti-apoptotic Effects-The synthetic VDAC1-based peptides used in the SPR and HK detachment experiments are linear, flexible in solution, and most likely do not adopt the same loop-like structure as those predicted in the membrane topology model (Fig. 1). Thus, to better mimic the native loop-like structure, we designed VDAC1-based peptides flanked by the tryptophan zipper motif consisting of the amino acids sequences SWTWE at the N terminus and KWTWK at the C terminus of the peptide, inducing the formation of stable ␤-hairpins by tryptophan-tryptophan cross-strand pairing (35).
HK-I and HK-II overexpression was shown to prevent apoptotic cell death induced by STS or VDAC1 overexpression ( Fig. 2 and Refs. 9 and 24). The results presented in Fig. 5 show that HK-I or HK-II overexpression protected against STS-induced apoptosis in T-REx-293 cells. When cells expressing HK-I or HK-II were also transfected with mammalian expression vectors encoding the VDAC1-based peptides, the expression of the N-terminal, LP1, LP2, or LP4, but not the LP3, peptides stabilized in a loop-form by an induced tryptophan zipper motif, no protection by HK-I or HK-II against STS-induced cell death was obtained (Fig. 5). These results suggest that the VDAC1 sequences represented by the recombinant loop-peptides, LP1, LP2, and LP4, or part thereof, as well as the N terminus recombinant peptide, interact with HK-I and HK-II and thus prevent interaction of these proteins with VDAC1, in turn preventing their protec-

VDAC1-based Peptides Interact with HK
tion against cell death, as induced by STS. Because an inducible plasmid was used for the expression of the VDAC1based peptides, inhibition of the anti-apoptotic effects of HK-I and HK-II was observed only when peptide expression was induced by tetracycline (Fig. 5B).

VDAC1-based Peptides Prevent HK-I-and HK-II-mediated Inhibition of STS-induced Cytochrome c
Release-Next, we tested whether the interaction of HK-I and HK-II with the VDAC1-based peptides would prevent HK-I-and HK-II-mediated inhibition of cytochrome c release, as activated by STS in T-REx-293 cells. Cytochrome c released from mitochondria to the cytosol upon STS-activated apoptosis in these cells was analyzed in the cytosolic fraction by immunoblotting, using anticytochrome c antibodies (Fig. 5C). HK-I or HK-II overexpression (Fig. 5D) inhibited STS-induced cytochrome c release in control cells but not in cells expressing the LP1 loop peptide (Fig. 5C). These findings are in correlation with the inability of HK-I or HK-II to protect against cell death induced by STS in cells expressing the LP1 peptide (Fig. 5, A and B). To rule out mitochondrial contamination in the cytosolic fraction as the source of cytochrome c, the absence of VDAC was confirmed by immunoblotting (data not shown).
The Synthetic Loop-shaped LP4-Antp Peptide Penetrates the Cell, Interferes with HK-I-mediated Anti-apoptotic Activity, and Detaches Bound HK-I from Rat Brain Mitochondria-To allow passive penetration of the peptide into the cell, we designed a LP4 peptide flanked by the tryptophan zipper motif (see above) and attached to the Drosophila homeobox protein Antennapedia homeodomain 16-amino acid sequence (Antp). Antp is a well known, non-toxic cell-penetrating peptide that is able to facilitate the translocation of fused peptides across the cell membrane (36 -39). A schematic representation of the loop-shaped LP4-Antp peptide is shown in Fig. 6A. FITC-labeled LP4-Antp and Antp peptides were used to assess their cell penetration (Fig. 6B). Fluorescence microscope images of cells incubated with either peptide show that both penetrate the cell (Fig. 6B).
Next, we tested the effects of the LP4-Antp peptide on apoptosis induced by STS in T-REx-293 cells overexpressing HK-I (Fig. 6C). The LP4-Antp peptide prevented HK-Imediated protection against cell death induced by STS. Although in cells exposed to the Antp peptide HK-I protected against STS-induced apoptosis by about 75%, in the presence of the LP4-Antp peptide the protection was only 13%. Moreover, the LP4-Antp peptide alone (i.e. in the absence of STS) induced apoptotic cell death.
Like the LP4 peptide (Fig. 4), the LP4-Antp peptide also induced the displacement of HK-I from rat brain mitochondria. As demonstrated by Western blot analysis, incubation of isolated brain mitochondria with the loop-shaped LP4-Antp peptide but not the control Antp peptide resulted in the appearance of HK in the supernatant (Fig. 6D).
Finally, to demonstrate that the peptides can detach mitochondria-bound HK-I at the cellular level, we expressed the HK-I-GFP fusion protein alone or together with the LP1, LP2, LP3 or LP4 peptides or when incubated with a synthetic peptide LP4-Antp, and visualized HK-I-GFP cellular distribution (Fig.  7). Confocal fluorescence microscopy showed that in control cells, HK-I-GFP fluorescence is punctuated, as expected for a mitochondrial distribution (see also Ref. 24). However, when cells expressing HK-I-GFP were exposed to the synthetic peptide LP4-Antp, HK-I-GFP was detached from the mitochondria, as reflected in the diffuse fluorescence (Fig. 7A). Similarly, HK-I-GFP fluorescence in cells expressing the LP1 or LP2 but not the LP3 peptide was diffused throughout the cytosol (Fig.  7B). These results indicate that both expressed LP1, LP2, and LP4 but not the LP3 peptide, as well as the membrane-penetrating LP4-Antp peptide, detach or prevent HK-I binding to the mitochondria, as also demonstrated with isolated mitochondria (Fig. 4) and in the prevention of HK-I-or HK-II-mediated protection against STS-induced cell death (Fig. 5). Mitochondria isolated from rat brain (A) or BCL1 leukemia tumor cells (B) were incubated without or with VDAC1-based peptides at the indicated concentrations. After a 1-h incubation at 25°C, the samples were centrifuged (15,000 ϫ g, 10 min) and soluble and mitochondria-bound HK fractions were separated. HK activity in the supernatant was measured spectrophotometrically at room temperature at 340 nm by coupling NADH oxidation by lactate dehydrogenase to the production of ADP by HK, and its subsequent phosphorylation by pyruvate kinase with phosphoenolpyruvate (1 mM) as the substrate. Lactate dehydrogenase and pyruvate kinase were added in excess, i.e. about 50 milliunits of each enzyme being added for 5 milliunits of HK. Glu-6-P (G6P) served as a positive control in this assay (8). In A, C, and D, detachment of HK-I from rat liver mitochondria was assayed as a function of LP1 and N peptide concentrations. E, Western blot analysis of HK-I released from rat brain mitochondria following the indicated treatments (0.5 mM Glu-6-P, 0.05 mM LP1, and 0.1 mM LP3 or N peptide). FEBRUARY 6, 2009 • VOLUME 284 • NUMBER 6

JOURNAL OF BIOLOGICAL CHEMISTRY 3951
The Loop-shaped LP4-Antp Peptide Triggers Human and Murine Cancer Cell Death-To further assess the effect of the LP4-Antp peptide on cell viability, different human and murine cancer and transformed cell lines (T-REx-293, MCF7, Molt4, CT26, and B16 cell lines) were exposed to the loop-shaped LP4-Antp peptide and cell death was analyzed by flow cytometry of PI-positive cells (Fig. 8 and Table 2). As shown, the peptideinduced cell death in all these cell lines, each originating from a different type of cancer (or transformed cells, such as T-REx-293 cells) but not in human peripheral blood lymphocytes.
Interestingly, quantitative analysis of HK-I and HK-II levels in these cells showed that the higher the level of bound HK, the lesser the degree of cell death induced by the LP4-Antp peptide ( Table 2). For example, the MCF7 cell line, which contains the highest level of bound HK, showed 24% cell death and only after 3 h of incubation, whereas Molt4 cells with 40% the HK level of MCF7 cells showed 65% cell death following a 1.5-h incubation with the same concentration of LP4-Antp peptide.

DISCUSSION
Recently (9, 24), we have identified those domains in VDAC1 involved in the interaction of the VDAC1 with HK-I. Relying on these identified VDAC1 sequences, we have now used synthetic peptides corresponding to these and other VDAC1 sequences and assessed their success in binding to and detaching of mitochondrial HK-I and HK-II. We also demonstrated their success in diminishing the protective activities of HK-I and HK-II against cell death, with the aim of introducing these VDAC1based peptides into cancer therapy.
HK-I and HK-II interaction with the VDAC1-based peptides was analyzed using 4 different approaches, namely direct interaction of the synthetic peptides with purified HK-I as monitored using SPR (Fig. 3), detachment of HK-I bound to mitochondria isolated from brain or tumor cells (Fig. 4), both possessing high levels of mitochondria-bound HK, testing the effects of the expression of the VDAC1-based peptides on HK-I-and HK-II-mediated protection against cell death (Fig. 5), and detachment of HK-I-GFP in cells expressing VDAC1-based peptides (Fig. 7). The results obtained using the various approaches indicate that some (i.e. LP1, LP2, and LP4) but not all (LP3) of the VDAC1-based peptides interact, to differing extents, with HK-I and HK-II (Figs. 3-5 and  7). The finding that LP2 detached mitochondrial-bound HK-I (Figs. 4 and 7) and interfered with HK-I-and HK-II-mediated protection against cell death (Fig. 5), yet attained no significant binding to immobilized HK-I (Fig. 3) could be due to differences in the Homogenates were centrifuged (1,000 ϫ g) at 4°C for 5 min and the supernatants were re-centrifuged at 15,000 ϫ g for 15 min at 4°C. Aliquots (10 l) of the resultant supernatants, designated as cytosolic fractions, were immediately boiled in SDS-PAGE sample buffer and resolved by SDS-PAGE on Tris-Tricine gels (13% polyacrylamide) and immunoblotted using monoclonal anti-cytochrome c antibodies (1:1000), followed by incubation with horseradish peroxidase-conjugated anti-mouse IgG. Secondary antibody binding was detected by chemiluminescence using a kit obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Although supernatants from the different treated cells were derived from an equal number of cells, protein concentrations in the total cell lysates were estimated and actin levels were analyzed by immunoblotting, using an anti-actin antibody. D, HK-I and HK-II levels in control cells and cells transfected to overexpress HK-I or HK-II and/or VDAC1based peptide. Cell lysates (50 g) were analyzed using polyclonal anti-HK-I and anti-HK-II antibodies, respectively. As a loading control, actin levels in the cell lysates were compared using an anti-actin antibody.

VDAC1-based Peptides Interact with HK
assay conditions. In SPR, HK-I was immobilized so the binding site may not be available, and the duration of the experiment is relatively short, on the minute scale. In the cell-based HK-I and peptide expression experiments, the assay was performed 48 h following expression of both the LP2 peptide and HK-I, thus allowing HK-I/peptide steady-state interactions.
Some of the VDAC1 amino acids, such as Glu 72 , mutation of which diminished HK-mediated protections against apoptosis, are located in the ␤-strands. Based on NMR study, it was proposed that Glu 73 (counting from the first methionine) influenced the structural instability of the N-terminal segment (27) because Glu 73 , as a charged residue, was found to break the alternating pattern of hydrophobic and hydrophilic residues in strand ␤-4 and introduce structural instability of the N-terminal segment. Replacing Glu 73 with glutamine or valine resulted in a stable conformation of the N-terminal part of the protein (27), and this may affect its accessibility to HK. As shown here and in previous studies (9,24), substitution of Glu 73 by glutamine has important consequences for the function of VDAC1, including abolishing ruthenium red-and HK-I-mediated inhibition of VDAC channel activity and ruthenium red-and HK-Imediated protection against apoptosis.
In addition to the VDAC1 loops facing the cytoplasm, the N-terminal region of VDAC1 also interacts with HK. These finding thus suggest that the interaction of HK with VDAC1 is mediated by multiple interaction sites.
The interaction of the N terminus of VDAC1 with HK, as demonstrated by the loss of HK binding upon its removal 3 as well as the results of this study, suggests that this region is accessible to HK and hence, exposed to the cytoplasm. This is in accord with the suggestion that the VDAC1 N-terminal region is a mobile component involved in the interaction with antiapoptotic proteins. 3 Various proposed VDAC1 membrane topology models predict differences in the location of the N terminus (25,40,41). Recent NMR-based studies of recombinant VDAC1 structure suggest that the N-terminal tail of VDAC1 is not part of the barrel wall but rather is located inside the pore, with only part of it in the form of ␣-helix (27,28). However, because this segment is involved in voltage gating (42), 3 it might adopt different conformations depending on cell conditions. Able to interact with HK, it is implied that this VDAC1 region is not permanently embedded in the pore.
Although the N-terminal region of HK has been defined as being essential for its anchoring to mitochondria (3, 8, 12, 43,

LP4-Antp but not Antp peptide induced cell death of human and murine cancer cells as measured by PI staining using flow cytometry
T-REx-293, Molt4, B16, and CT26 cells were incubated for 90 min with 20 M LP4-Antp or Antp peptides in serum-free DMEM. Peripheral blood leukocytes (PBL) and MCF7 cells were incubated with the peptides for 2 and 3 h, respectively. The cells were then stained with PI. The level of cell death, reflected by the percentage of PI-positive cells, was determined by flow cytometry. Data are mean Ϯ S.E. of three to four independent experiments. The LP4-Antp-induced cell death was obtained by subtracting cell death in the control. Control indicates an average of the percentage of PI-positive cells incubated without (0.4% dimethyl sulfoxide) and with Antp peptide. The level of mitochondria-bound HK was analyzed by Western blot of isolated mitochondria using polyclonal anti-HK antibodies recognizing both HK-I and HK-II. Immunoblotting using anti-cytochrome c and anti-VDAC antibodies were used as loading controls.  (1,3,4,7). Various studies have demonstrated the anti-apoptotic activity of these proteins and that the HK-VDAC interaction is critical for preventing the induction of apoptosis in tumor-derived cells (8,9,45,46). Thus, the greatly increased expression of HK-II in aggressive tumors, such as gliomas (45), or of HK-II and HK-I in hepatomas (3), in comparison to their limited expression in normal tissues, makes HK-I and HK-II attractive targets for cancer therapy. Apoptogenic compounds acting directly on the HK-mitochondria interaction in cancerous cells have been extensively studied (47). Among these are HK inhibitors, such as glucose 6-phosphate and its analogs (48), enzymes cleaving the HK N terminus (34), or those that act by unknown mechanisms of action, such as hypericin or TH-070 (lonidamine) (12,49,50). In addition, intra-arterial injection of 3-bromopyruvate, an inhibitor of mitochondria-bound HK, into tumors implanted in rabbit liver resulted in death of up to 90% of the tumor cells (14,33). In addition, detachment of HK-II from the mitochondria by activated glycogen synthase kinase 3␤ and the subsequent phosphorylation of VDAC promoted a synergistic increase in cell death brought about by suboptimal doses of the chemotherapeutic drugs, doxorubicin and paclitaxel (23). Recently (29), we have shown that a plant stress hormone of the jasmonate family, which induces cancer cell death, methyl jasmonate binds in a specific manner to HK and detached both HK-I and HK-II from mitochondria isolated from several cancer cell types.

Cell type
These agents, however, suffer from various disadvantages, such as low specificity or poor cell permeability. In the present study, the interaction of the VDAC1-based LP1, LP2, LP4, and N-terminal peptides with HK-I and HK-II prevented their inhibition of cytochrome c release and protection against apoptotic cell death (Fig. 5). The ability of the LP1, LP2, and LP4 peptides to detach mitochondria-bound HK is clearly demonstrated in this study using HK-I-GFP (Fig. 7). Therefore, in cells overexpressing HK-I or HK-II, such as cancer cells, inhibition of the interaction of HK-I or HK-II with VDAC1 by VDAC1-based peptides would initiate mitochondrial apoptotic signaling cascades. In addition, our proposed approach to inhibiting the anti-apoptotic activity of HK-I and HK-II using the cell-penetrating VDAC1-based LP4 peptide fused to the Antennapedia internalization sequence, previously fused to peptides modulating apoptosis, such as BH3 domains or p53 peptides (36,37,39), overcomes the problem of cell permeabilization. We demonstrate that the LP4-Antp peptide alone, in the form of a synthetic loop-shaped peptide, easily penetrates into cells, is able to detach HK from mitochondria, and induce cell death in several cancer cell lines but not in normal cells, such as peripheral blood lymphocytes (Fig. 8 and Table 2). It is evident that the LP4-Antp peptide acts to detach HK (Fig. 6D) or HK-I-GFP (Fig. 7) bound to the mitochondria, with a capacity to induce cell death dependent on the level of HK in the cell (Table 2). Thus, we suggest that promoting the detachment of HK from its mitochondrial binding site by VDAC1-based peptides represents a promising cancer strategy. VDAC1-based peptides would have strong therapeutic potential because they affect mitochondria-bound HK, a characteristic of cancer cells, yet have no inhibitory effect on HK enzymatic activity (data not shown). In addition, because the VDAC1 sequence is highly conserved and found in all human cells, designed VDAC1based peptides would be expected to be non-immunogenic and have no toxic effects. Moreover, VDAC1-based peptides, as pro-apoptotic compounds, can be used to potentiate the efficacy of conventional chemotherapeutic agents. Targeting VDAC1-based peptides to tumor cells overexpressing such anti-apoptotic proteins may prove an effective VDAC1-based cancer therapy, acting to minimize the self-defense mechanisms of cancer cells, thereby promoting apoptosis.