MEK Kinase 1 Induces Mitochondrial Permeability Transition Leading to Apoptosis Independent of Cytochrome cRelease*

Induction of apoptosis often converges on the mitochondria to induce permeability transition and release of apoptotic proteins into the cytoplasm resulting in the biochemical and morphological alteration of apoptosis. Activation of a serine threonine kinase MEK kinase 1 (MEKK1) is involved in the induction of apoptosis. Expression of a kinase-inactive MEKK1 blocks genotoxin-induced apoptosis. Upon apoptotic stimulation, MEKK1 is cleaved into a 91-kDa kinase fragment that further induces an apoptotic response. Mutation of a consensus caspase 3 site in MEKK1 prevents its induction of apoptosis. The mechanism of MEKK1-induced apoptosis downstream of its cleavage, however, is unknown. Herein we demonstrate that full-length and cleaved MEKK1 leads to permeability transition in the mitochondria. This permeability transition occurs through opening of the permeability transition (PT) pore. Inhibiting PT pore opening and reactive oxygen species production effectively reduced MEKK1-induced apoptosis. Overexpression of MEKK1, however, failed to release cytochrome c from the mitochondria or activate caspase 9. Since Bcl2 regulates changes in mitochondria and blocks MEKK1-induced apoptosis, we determined that Bcl2 blocks MEKK1-induced apoptosis when targeted to the mitochondria. This occurs downstream of MEKK1 cleavage, since Bcl2 fails to block cleavage of MEKK1. In mouse embryonic fibroblast cells lacking caspase 3, the cleaved but not full-length MEKK1 induces apoptosis and permeability transition in the mitochondria. Overall, this suggests that cleaved MEKK1 leads to permeability transition contributing to MEKK1-induced apoptosis independent of cytochrome c release from the mitochondria.

Apoptotic signals often lead to changes in the mitochondria. These changes consist of mitochondrial permeability transition, production of reactive oxygen species (ROS), 1 and release of proteins into the cytoplasm (1,2). These events could cause blockage of ATP production, damage to membranes, DNA condensation, and activation of caspases leading to apoptosis (1)(2)(3)(4). In response to apoptotic stimulation, the mitochondrial membrane opens to allow solutes and water to enter the mitochondria. This is controlled by a multiprotein complex found in the inner and outer membranes of the mitochondria known as the permeability transition (PT) pore (2,5,6). The PT pore consists of voltage-dependent anion channel/porin, adenine nucleotide translocator, cyclophilin D, creatine kinases, and other proteins (6,7). Upon PT pore opening, the mitochondria loses its membrane potential (⌬ m ) across the inner membrane. This often occurs following apoptotic stimuli. This is associated with increased production of ROS that further damages proteins and membranes (6,7). In addition, proteins are released from the mitochondria following apoptotic stimulus such as cytochrome c. When cytochrome c is released, it binds to an Apaf1 and caspase 9 complex (1,4,8). This leads to caspase 9 activation, causing cleavage of specific proteins and activation of other caspases. These mitochondrial events regulate the induction of apoptosis.
The Bcl2 family of proteins control mitochrondial apoptotic responses (9). Bcl2 family members consist of pro-and antiapoptotic proteins (10,11). Proapoptotic Bcl2 family members translocate to the outer membrane of the mitochondria from the cytosol following an apoptotic signal (11). This translocation results in release of cytochrome c into the cytosol and opening of the PT pore. The exact mechanism of how these proteins release mitochondrial cytochrome c or open the PT pore remains unclear. Anti-apoptotic Bcl2 family members such as Bcl2 itself can reverse the effects of the pro-apoptotic members (9). Bcl2 expression effectively blocks the PT pore opening and cytochrome c release. This is accomplished through Bcl2 binding to pro-apoptotic Bcl2 family members, forming heterodimers. This inhibits these pro-apoptotic proteins from opening the PT pore and releasing cytochrome c from the mitochondria (9,11). This interplay between pro-and anti-apoptotic Bcl2 family members regulates the role of the mitochondria in apoptosis.
MEK kinase 1 (MEKK1) is a serine threonine kinase that when overexpressed induces caspase activation and apoptosis (12,13). Many apoptotic stimuli including genotoxic agents activate MEKK1 kinase activity (12,14). Following genotoxin treatment, MEKK1 activation leads to increased expression of death receptors such as Fas and death receptors 4 and 5, presumably through activation of transcription factors such as NF-B (15). This up-regulation of death receptors contributes to genotoxin-induced apoptosis. Following treatment with apoptotic stimuli, MEKK1 is cleaved by caspases into a 91-kDa fragment containing the kinase domain (13,16). Overexpression of this cleaved product is more potent at activating caspases and inducing apoptosis than full-length MEKK1 (13). Cleavage of MEKK1 is dependent on caspase 3-like molecules, since MEKK1 contains a consensus site for caspase 3 cleavage, and caspase 3 inhibitors block MEKK1 cleavage. Mutation of this cleavage site prevents MEKK1-induced apoptosis following expression in cells (13). MEKK1-induced apoptosis is dependent on its kinase activity, since expression of a kinaseinactive form of MEKK1 fails to induce apoptosis (14,16). Indeed, kinase inactive MEKK1 acts as a dominant negative protein blocking apoptosis following genotoxin treatment (14,16). The downstream mechanism of MEKK1 induction of apoptosis following caspase cleavage, however, remains unknown.
Herein, we demonstrate that overexpression of MEKK1 suppresses mitochondrial ⌬ m mediated by opening of the PT pore. Inhibiting the PT pore and ROS formation effectively reduced MEKK1-induced apoptosis. Furthermore, full-length MEKK1 expression fails to suppress ⌬ m or induce apoptosis in mouse embryonic fibroblasts (MEFs) lacking caspase 3, whereas the 91-kDa kinase fragment still is capable of ⌬ m suppression and induction of apoptosis in these cells. However, MEKK1 fails to release cytochrome c from the mitochondria. This provides evidence that MEKK1 cleavage causes permeability transition, leading to apoptosis independent of cytochrome c release.

MATERIALS AND METHODS
Cell Culture-Human embryonic kidney (HEK) 293 cells, breast cancer cell line MCF-7, and human transformed cell line HeLa were maintained in a humidified 5.0% CO 2 , 37°C incubator in Dulbecco's modified medium supplemented with 100 units/ml penicillin, 100 g/ml streptomycin (Invitrogen). Medium for HEK 293 and MCF-7 cells was supplemented with 10% fetal bovine serum (Invitrogen). Primary MEF cells were maintained in minimum essential medium supplemented with 10% fetal bovine serum (kind gift from Dr. Tak Mak, Amigen Institute). HEK 293 cells expressing vector alone, MEKK1 KM and Bcl2 proteins were under selection with 1 mg/ml G418 (Invitrogen). MCF-7 cells expressing vector alone, Bcl2 wild type, and Bcl2 targeted to the mitochondria (Bcl2-mito) were under selection with 0.5 mg/ml G418 as previously described (17)  Transfections-HEK 293 cells were grown on glass coverslips. They were co-transfected with 2 g of pGFP containing cDNA for green fluorescent protein (GFP; CLONTECH), pcDNA containing the cDNA for MEKK1, kinase-inactive MEKK1, or the 91-kDa form of MEKK1 using the LipofectAMINE technique (Invitrogen) as per the manufacturer's instructions. The transfection efficiency was about 20%. MCF-7, HeLa, and MEF cells were also grown on coverslips and transfected with 8 g of plasmid DNA as described above using a Geneporter (GTS systems) as per the manufacturer's instructions. The transfection efficiency was 30% for MCF-7 cells, 20% for HeLa cells, and 10% for MEF cells. Cells were viable following transfection, and background apoptosis was 10%. ⌬ m Determination by Flow Cytometry and Fluorometry-HEK 293 and MEF cells were grown in 100-mm tissue culture plates (flow cytometry) or on large coverslips (fluorometry). HEK 293 cells grown on plates were transfected with cDNA for GFP alone, MEKK1, kinase-inactive MEKK1 (MEKK1 KM), or 91-kDa MEKK1 as described above. After 24 h of incubation, the cells were resuspended in 1ϫ Hepes-buffered saline solution (HBSS) and spun down at 1200 rpm for 2 min. One ml of 1ϫ HBSS was used to resuspend the cells, and the cells were counted. Cells (1 ϫ 10 6 ) were diluted in 1ϫ HBSS, and 500-l aliquots were placed into 1.5-ml centrifuge tubes. Tetramethylrhodamine (TMRM; Molecular Probes, Inc., Eugene, OR) was added to the appropriate tube at 150 M final concentration and incubated in the dark at room temperature for 15 min. The cells were then spun down and analyzed by a Becton Dickinson Facscaliber E2807 flow cytometer. FL-1 detects green fluorescence at 530 -585 nm, whereas FL-2 detects red fluorescence at 488 -635 nm. The compensation for FL1 was 18% of FL2, and compensation for FL2 was 58% of FL1. FL1 was gated for green fluorescent protein-positive events, and these positive events were further gated for FL2 (TMRM)-positive events. These gates remained unchanged for each condition tested. The number of FL1-positive compared with FL2-negative events was calculated as the percentage of ⌬ m suppression. Cells under each condition tested were analyzed for 20,000 events on the flow cytometer. HEK 293 cells stably expressing vector alone or MEKK1 KM were treated with 100 M etoposide and stained for TMRM as described above. For HEK 293 cells grown on coverslips, the cells were transfected with cDNA for GFP (2 g) or GFP fused to MEKK1 (2 g). For MEF cells, the cells were transfected with GFP alone (0.8 g) or in combination with MEKK1 (8 g) or 91-kDa MEKK1 (8 g) using Geneporter as per the manufacturer's instructions. After transfection, these cells were washed four times in 1ϫ HBSS, and vacuum grease was applied to the edges of the coverslip. The cells were then stained with 150 M TMRM in 1ϫ HBSS. The coverslip was washed four times with 1ϫ HBSS and analyzed on an Olympus IX70 inverted confocal laser microscope using Flouview 2.0 software.
PT Pore Opening-HEK 293 cells were grown on coverslips and transfected with cDNA for red fluorescent protein (RFP; 0.2 g) or in combination with MEKK1 (2 g) or 91-kDa MEKK1 (2 g) using Lipto-fectAMINE as per the manufacturer's instructions. The cells were washed in 1ϫ HBSS and stained with 5 mM CalceinAM (Molecular Probes) and 5 mM CoCl 2 (Sigma) for 15 min (18). The coverslips were then washed four times with 1ϫ HBSS and analyzed on an Olympus IX70 inverted confocal laser microscope using Flouview 2.0 software.
Immunohistochemistry and Apoptosis Measurement-Transfected HEK 293 cells were untreated or treated with 20 M cyclosporin A (CsA) or 5 mM 4,5-dihydroxy-1,3 benzene disulfonic acid (Tiron) for 24 h where indicated. Parental and superoxide dismutase-expressing HeLa cells along with MCF-7 cells expressing Bcl2 variants were also transfected as described above. The coverslips were collected and fixed in 3.7% formaldehyde in 1ϫ phosphate-buffered saline (PBS). They were then washed twice for 5 min with 500 l in 1ϫ PBS and 0.1% Nonidet P-40. In cells expressing HA-tagged-MEKK1, rabbit polyclonal anti-MEKK1 (Santa Cruz) at a dilution of 1:500 in 10% fetal bovine serum was incubated with the coverslips for 1.5 h with gentle shaking. Secondary antibody anti-rabbit fluorescein isothiocyanate (Chemicon) at a 1:5000 dilution in 10% fetal bovine serum, 1ϫ PBS, and 0.1% Nonidet P-40 was incubated with the coverslip for 1.5 h with gentle shaking in the dark. The exceptions were cells expressing GFP or GFP fused to MEKK1, which were untreated. The coverslips were washed twice in 500 l of 1ϫ PBS and 0.1% Nonidet P-40 and incubated for 6 min in 1:2000 dilution of 20 mM Hoechst stain (Sigma) in 1ϫ PBS and 0.1% Nonidet P-40 in the dark. Coverslips were mounted onto slides with 4 l of Fluoroguard antifade reagent (Bio-Rad). The condensed DNA was determined by intense local staining of the DNA in nuclei by Hoechst compared with the diffuse staining of the DNA in normal cells. The percentage of apoptotic cells was determined from cells containing normal DNA staining compared with cells with condensed DNA fragmentation and morphological changes consistent with apoptotic cells. No fewer than 200 cells were scored per sample. Fluorescence was visualized and captured using an Olympus fluorescent microscope equipped with a Spot camera.
Assessment of Cytochrome c Release-HEK 293 cells were grown on glass coverslips and transfected with GFP, MEKK1, truncated Bid (tBid), and 91-kDa MEKK1 cDNA as previously described. The coverslips were collected and fixed in 3.7% formaldehyde in 1ϫ PBS. They were then washed twice for 5 min with 500 l in 1ϫ PBS and 0.1% Nonidet P-40. Mouse anti-cytochrome c (Pharmingen) and rabbit anti-MEKK1 or anti-Bid (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) where indicated were diluted 1:500 in 10% fetal bovine serum, 1ϫ PBS, and 0.1% Nonidet P-40. The cells were then incubated for 1.5 h with gentle shaking. The coverslips were washed twice in 500 l of 1ϫ PBS and 0.1% Nonidet P-40. Secondary antibody anti-mouse cy3 or antirabbit fluorescein isothiocyanate (Chemicon) at a 1:5000 dilution in 10% fetal bovine serum, 1ϫ PBS, and 0.1% Nonidet P-40 was incubated with the coverslip for 1.5 h with gentle shaking in the dark. The secondary antibody was removed, and the coverslips were incubated for 6 min in a 1:2000 dilution of 20 mM Hoechst stain (Sigma) in 1ϫ PBS and 0.1% Nonidet P-40 in the dark. Coverslips were mounted onto slides with 4 l of Fluoroguard antifade reagent (Bio-Rad). No fewer than 200 cells were scored per sample. Fluorescence was visualized and captured using a Zeiss Axiphot microscope equipped with a cooled charged-coupled device camera. For biochemical analysis of cytochrome c release, transfected HEK 293 cells were resuspended in CFS buffer (10 mM Hepes, pH 7.4, 220 mM mannitol, 68 mM sucrose, 2 mM NaCl, 2. Immunoblots-Cells were lysed in Nonidet P-40 lysis buffer (50 mM HEPES, pH 7.25, 150 mM NaCl, 50 M ZnCl 2 , 50 M NaF, 2 mM EDTA, 1 mM sodium vanadate, 1.0% Nonidet P-40, 2 mM phenylmethylsulfonyl fluoride). Cell debris was removed by centrifugation at 8000 ϫ g for 5 min, and protein concentration was determined by a Bradford assay. 200 -400 g of cell lysate protein was subject to SDS-polyacrylamide electrophoresis and transferred to nitrocellulose membranes. The membranes were blocked in Tris-buffered saline and 5% milk (w/v). Blots were incubated with the appropriate antibody concentration overnight (anti-MEKK1, anti-cytochrome c, and anti-HA were purchased from Santa Cruz Biotechnology; anti-caspase 9 was purchased from New England Biolabs), washed three times with 1ϫ Tris-buffered saline, and incubated 1 h with the appropriate secondary antibody conjugated with alkaline phosphatase. Blots were visualized on x-ray film with enhanced chemiluminescence reagents (PerkinElmer Life Sciences). full-length MEKK1, and 91-kDa MEKK1. The cells were stained with the potentiometric dye TMRM, which detects ⌬ m levels in the mitochondria for 24 h as described under "Materials and Methods," and then analyzed by flow cytometry. This showed that cells expressing GFP have 20% ⌬ m suppression (defined as the reduction in mitochondrial membrane potential) while cells expressing MEKK1 or 91-kDa MEKK1 have 40 and 49% ⌬ m suppression, respectively (Fig. 1A). Expression of the kinase-inactive form of MEKK1 (MEKK1 KM) that prevents apoptosis failed to suppress ⌬ m (21%; Fig. 1A pressing only GFP showed mitochondrial membrane potential as indicated by red fluorescence of TMRM staining, whereas cells expressing GFP fused to full-length MEKK1 showed ⌬ m suppression as indicated by reduced TMRM staining (Fig. 1C). Some cells expressing the GFP-MEKK1 fusion protein showed TMRM staining. This correlates to the ability to MEKK1 to suppress ⌬ m in 40% of cells (Fig. 1A) and induce apoptosis in 30% of cells after 24 h (data not shown). Expression of kinaseinactive MEKK1 effectively blocked genotoxic agent etoposideinduced apoptosis in HEK 293 cells (14). Treatment of HEK 293 cells expressing MEKK1 KM with etoposide also blocked ⌬ m suppression compared with vector alone cells over a 48-h time course (Fig. 1D).

Overexpression of MEKK1 Suppresses
Loss of ⌬ m could be due to opening of the PT pore. To confirm if MEKK1 is opening the PT pore to suppress ⌬ m , HEK 293 cells were stained with the mitochondrial CalceinAM fluorescence dye. CalceinAM is freely permeable across cellular membranes but becomes fluorescent (green) and impermeable upon cleavage by intracellular esterases. This prevents its exit from the mitochondria until the PT pore opens. Once the PT pore is open, CalceinAM is released into the cytoplasm from the mitochondria, where it is quenched by CoCl 2 (added to the cells along with CalceinAM). Using confocal laser microscopy, HEK 293 cells expressing RFP alone or in combination with MEKK1 or 91-kDa MEKK1 were analyzed for calcein fluorescence. In cells expressing RFP alone, calcein fluorescence was punctate, indicating that CalceinAM is localized in the mitochondria. However, in cells expressing MEKK1 or 91-kDa MEKK1, cal- ceinAM fluorescence was eliminated, indicating a release of calceinAM into the cytoplasm, where its fluorescence is quenched ( Fig. 2A). Loss of ⌬ m could be due to membrane rupture or other open channels in the mitochondria. To determine if MEKK1-mediated loss of ⌬ m was due to PT pore opening, the PT pore opening inhibitor CsA was added to cells expressing MEKK1, and the amount of ⌬ m suppression was determined. GFP alone expression failed to increase ⌬ m suppression compared with cells expressing GFP fused to MEKK1 (12% versus 45%; Fig. 2B). In the presence of CsA, the amount of ⌬ m suppression was reduced to 12% (Fig. 2B). CsA treatment failed to change the amount of ⌬ m suppression in cells expressing GFP alone (data not shown). This indicates that the PT pore is open when MEKK1 is expressed, causing loss of ⌬ m .
MEKK1 Fails to Release Cytochrome c from the Mitochondria-In addition to changes in membrane potential, mitochondrial cytochrome c release occurs during apoptosis, leading to caspase activation (1). We investigated the ability of MEKK1 to release cytochrome c from the mitochondria. HEK 293 cells were transfected with vector alone, MEKK1, or 91-kDa MEKK1. The cells were then immunohistochemically stained for cytochrome c and MEKK1. In vector alone, cytochrome c staining was punctate, indicating that it is localized in the mitochondria. Using a mitochondria-specific stain (Mitotracker), a similar staining pattern was observed (data not shown). In cells expressing MEKK1 and 91-kDa MEKK1, the same punctate staining was present as in vector alone cells The cells were then fixed and stained with Hoechst. The amount of apoptosis was determined by DNA condensation on an Olympus fluorescent microscope. B, HEK 293 cells were also transiently transfected with GFP or MEKK1 fused with GFP and at the same time treated with the free radical scavenger Tiron (5 mM) for 24 h. After 24-h incubation, the cells were fixed and stained with Hoechst. The amount of apoptosis was determined as in A. C, HeLa cells parental or expressing superoxide dismutase (SOD) were transiently transfected with GFP or MEKK1 fused to GFP. The cells were incubated for 24 h. The cells were then fixed and stained with Hoechst. The amount of apoptosis was determined as described above. S.D. values were determined by three independent experiments. (Fig. 3A). As a positive control, truncated Bcl2 family member BID (tBid), which has previously been shown to release cytochrome c from the mitochondria, was expressed in HEK 293 cells (Fig. 3A). Upon expression, there was reduced staining for cytochrome c in cells, indicating release of cytochrome c into the cytosol. To further confirm these results, membrane and cytosolic fractions were isolated from these cells and Western blotted for the presence of cytochrome c. In cells expressing MEKK1, cytochrome c failed to be detected in the cytosolic fraction but was present in the membrane-bound fraction (Fig.  3B). Cytochrome c was, however, found in the cytosolic fraction following tBid expression (Fig. 3B). These same results were also observed over a 72-h time course, where the cells expressing MEKK1 were also observed undergoing apoptosis as determined by DNA condensation (data not shown). In addition, caspase 9 activation was determined by Western blotting for the inactive procaspase form of caspase 9. When vector alone or MEKK1 was transfected into HEK 293 cells, the level of caspase 9 remained unchanged, but when tBid was overexpressed, caspase 9 was cleaved into its active form (Fig. 3C). This indicates that MEKK1 fails to activate caspase 9.
Blockage of PT Pore Opening and of Reactive Oxygen Species Production Effectively Inhibits MEKK1-induced Apoptosis-Our data indicate that the overexpression of MEKK1 leads to ⌬ m suppression in the mitochondria mediated by PT pore opening. We further investigated if blockage of PT pore opening by treatment with CsA or blockage of reactive oxygen species production inhibits MEKK1-induced apoptosis. HEK 293 cells were transiently transfected with MEKK1 in the absence or presence of CsA. CsA effectively inhibited MEKK1-induced apoptosis (Fig. 4A). ⌬ m suppression also associated with increased production of ROS. Using a free radical scavenger agent, Tiron, HEK 293 cells were transiently transfected with MEKK1 in the presence and absence of Tiron. In the presence of Tiron, MEKK1-induced apoptosis was blocked (Fig. 4B). To further confirm the involvement of ROS in MEKK1 induction of apoptosis, HeLa cells overexpressing superoxide dismutase, which eliminates accumulation of ROS, were transiently transfected with MEKK1. As a control, parental HeLa cells were also transiently transfected with MEKK1. The cells were then stained for DNA and MEKK1, and the amount of apoptosis in MEKK1-expressing cells was determined by DNA condensation. When superoxide dismutase was overexpressed, MEKK1induced apoptosis was also blocked as compared with parental cells (Fig. 4C). These results indicate that PT pore opening and production of ROS is involved in MEKK1-induced apoptosis.

Bcl2 Targeted to the Mitochondria Prevents MEKK1-induced Apoptosis but Fails to Prevent Cleavage of MEKK1-
We have previously shown that Bcl2 prevents MEKK1-induced apoptosis. Bcl2 has also been shown to block permeability transition in the mitochondria. Using the breast cancer cell line MCF7 expressing vector alone, Bcl2 wild type (Bcl2-wt), and Bcl2 targeted to the mitochondria (Bcl2-mito) (17,19,20), the ability of GFP, MEKK1, or 91-kDa MEKK1 to induce apoptosis was determined. These cells were transiently transfected with GFP, MEKK1, or 91-kDa MEKK1, and the amount of apoptosis was determined by DNA condensation. Both MEKK1 and 91-kDa MEKK1 were able to induce apoptosis in MCF7 cells expressing vector alone (39 and 44%, respectively), whereas GFP failed to induce apoptosis (14%; Fig. 5A). MEKK1 and 91-kDa MEKK1 induction of apoptosis was inhibited in MCF7 cells expressing Bcl2-wt (21 and 25%, respectively) and Bcl2mito (19 and 28%) (Fig. 5A). MEKK1-induced ⌬ m suppression was also blocked in HEK 293 cells expressing Bcl2 (data not shown). This indicates that Bcl2 blocks MEKK1-induced apoptosis at the mitochondria.
MEKK1 is cleaved following treatment with genotoxins such as etoposide and following overexpression in HEK 293 cells (12)(13)(14). Etoposide treatment of HEK 293 cells expressing Bcl2 leads to the cleavage of MEKK1 similar to HEK 293 cells expressing vector alone (Fig. 5B). Increased Bcl2 expression, however, was effective at inhibiting etoposide-induced apoptosis (data not shown). Overexpression of MEKK1 in HEK 293 cells expressing Bcl2 also failed to block MEKK1 cleavage (Fig.  5C). This suggests that Bcl2 blocks MEKK1-induced apoptosis downstream of MEKK1 cleavage.
Cleavage of MEKK1 by Caspase 3 Is Required for ⌬ m Suppression-MEKK1 is cleaved during apoptosis mediated by caspase 3-like molecules (12)(13)(14). To determine whether caspase 3 plays a role in MEKK1-induced apoptosis and ⌬ m suppression, we investigated primary MEFs lacking caspase 3 expression. MEF cells lacking caspase 3 were transiently transfected with GFP, MEKK1, or 91-kDa MEKK1. The cells were stained for DNA and MEKK1 expression, and the amount of apoptotic cells was determined by DNA condensation. Expression of MEKK1 caused apoptosis in wild type MEF cells (33%) but was significantly reduced in MEF cells lacking caspase 3 (20%) as compared with GFP expression (9.5 and 15%, respec- tively; Fig. 6A). In contrast, 91-kDa MEKK1 expression was found to induce apoptosis in both wild type and caspase 3-lacking cells (49 and 43%, respectively; Fig. 6A). MEKK1 and 91-kDa MEKK1 induced ⌬ m suppression in wild type cells, but only 91-kDa MEKK1 induced ⌬ m suppression in MEF cells lacking caspase 3 (Fig. 6B). This correlates with the ability of 91-kDa MEKK1 to induce apoptosis in wild type and caspase 3-lacking MEF cells, whereas full-length MEKK1 only induces apoptosis in wild type MEF cells. Treatment of MEF cells lacking caspase 3 with etoposide revealed that MEKK1 failed to be cleaved, whereas MEF wild type cells treated with etoposide caused cleavage of MEKK1 (data not shown). This indicates that ⌬ m suppression in the mitochondria occurs downstream of MEKK1 cleavage. DISCUSSION Full-length MEKK1 is activated during apoptosis and has been implicated in the induction of apoptosis (13,16,21,22). This is partially mediated by MEKK1 up-regulating proapoptotic genes such as death receptors (15,21,23). This upregulation contributes to genotoxin-induced apoptosis. In contrast, activation of MEKK1 also occurs following treatment with nonapoptotic stimuli such as growth factors (24,25). Furthermore, MEKK1 is activated by and might protect cells from microtubule toxin-induced apoptosis (14,26). MEKK1, however, fails to be cleaved following treatment with growth factors or microtubule toxins. Thus, activation of full-length MEKK1 is not sufficient to induce an apoptotic response, suggesting that cleavage of MEKK1 is required for its pro-apoptotic function.
Regulation of MEKK1 induction of apoptosis involves the cleavage of MEKK1 into a 91-kDa kinase fragment (13). This fragment is a potent inducer of apoptosis, and blockage of the cleavage site in MEKK1 that produces the 91-kDa fragment fails to induce apoptosis (13). The mechanism of 91-kDa MEKK1 induction of apoptosis was, however, unknown. We have demonstrated that cleaved MEKK1 leads to permeability transition in the mitochondria. Blockage of PT pore opening inhibits MEKK1induced apoptosis. The importance of cleavage of MEKK1 was demonstrated in MEF cells lacking caspase 3. This showed that only 91-kDa MEKK1 induces apoptosis and permeability transition in these cells. In contrast, MCF-7 cells that lack caspase 3 expression undergo apoptosis following overexpression of MEKK1. MEKK1 is also cleaved following etoposide treatment in MCF-7 cells. 2 In addition, treatment of these cells with doxorubicin induces apoptosis where doxorubicin fails to induce apoptosis in MEF cells lacking caspase 3 (27,28). This suggests that MCF-7 cells have bypassed the requirement of caspase 3 to induce apoptosis and illustrates potential differences between primary cells (MEF cells) and transformed cell lines (MCF-7 cells) in inducing apoptosis. Overall, we have demonstrated that MEKK1 induction of apoptosis involves cleavage into its 91-kDa form that leads to ⌬ m suppression in the mitochondria. MEKK1, unlike other proapoptotic proteins, fails to release cytochrome c from the mitochondria. This suggests that MEKK1-mediated caspase activation is independent of mitochondrial cytochrome c release, and loss of ⌬ m is not sufficient to release cytochrome c. Full-length MEKK1 up-regulates death receptors that could activate caspases independent of the mitochondria (15,29). In addition, caspase 3 activation is required for full-length MEKK1-induced loss of ⌬ m and apoptosis, suggesting that caspase activation occurs prior to MEKK1-induced mitochondria permeability transition. Alternatively, cleavage of MEKK1 could also lead to the activation of signaling pathways that further activate caspases independent of mitochrondrial cytochrome c release. Nevertheless, MEKK1 induces ⌬ m suppression independent of cytochrome c release, indicating that cytochrome c release is not involved in MEKK1mediated caspase activation.
Bcl2 family members regulate mitochondrial permeability transition (11). We have demonstrated that Bcl2 negatively regulates MEKK1-induced apoptosis, and this is downstream of MEKK1 cleavage and occurs at the mitochondria. This suggests that pro-apoptotic Bcl2 family members are involved in MEKK1induced apoptosis. Indeed, these proteins are translocated from the cytoplasm to the mitochondria during the induction of apoptosis (11). It has been reported that expression of the kinase domain MEKK1 translocates pro-apoptotic Bcl2 family member Bak to the mitochondria. The kinase-inactive form of MEKK1 kinase domain blocks this translocation and prevents cisplatininduced apoptosis (30). Most of these pro-apoptotic proteins including Bak cause release of cytochrome c from the mitochondria (11), but MEKK1 overexpression fails to release cytochrome c, and the significance of pro-apoptotic Bcl2 family members such as Bak in regulating MEKK1-induced apoptosis is unclear. A new Bcl2 family member BNIP3 has been shown to localize to the mitochondria and induces permeability transition without the subsequent release of cytochrome c (31,32). BNIP3, however, fails to induce apoptosis but instead induces a necrotic-like cell death. Whatever the pro-apoptotic Bcl2 family member responsible for MEKK1-induced loss of ⌬ m and apoptosis, Bcl2 negatively regulates MEKK1-induced apoptosis at the mitochondria downstream of MEKK1 cleavage. The proapoptotic Bcl2 family member responsible for MEKK1-induced mitochondrial permeability transition will be further investigated.
The apoptotic signaling pathways leading to mitochondrial permeability transition are not well defined. MEKK1 is a serine threonine kinase that when cleaved following apoptotic stimulus leads to permeability transition, further committing cells to undergo apoptosis. We have demonstrated that kinase activity and cleavage of MEKK1 are required for mitochondrial permeability transition. MEKK1 activates many signaling pathways. In particular, Jun N-terminal kinase activation is regulated by MEKK1 (26,33). Jun N-terminal kinase has been shown to phosphorylate Bcl2 family members, inhibiting their anti-apoptotic functions (34,35). This could explain the changes in mitochondrial permeability transition. Overexpression of MEKK1, however, does not lead to the phosphorylation of Bcl2 or Bclx L proteins as determined by changes in gel mobility of these proteins. 3 Furthermore, this phosphorylation has also been demonstrated to induce the release of cytochrome c from the mitochondria but not ⌬ m suppression. Finally, expression of kinase inactive MEKK1 prevents genotoxin-induced apoptosis but fails to block Jun Nterminal kinase activation in various cell types (14,30). MEKK1 also activates other pathways leading to activation transcription factors such as NF-B (22,36). This could induce gene expression of a molecule that opens the PT pore such as pro-apoptotic Bcl2 family members without cytochrome c release. Alternatively, MEKK1 could directly phosphorylate proteins that control PT pore opening. The exact substrate(s) for MEKK1 activation specifically leading to changes in mitochondrial permeability transition remains unknown and will be the focus of further investigation.
Overall, these results indicate that MEKK1 apoptosis signaling involves mitochondrial permeability transition but is independent of cytochrome c release. This occurs downstream of MEKK1 cleavage, committing cells to undergo apoptosis. By pharmaceutically targeting the cleavage of MEKK1, MEKK1induced apoptosis could be effectively regulated without affecting the other functions of full-length MEKK1. This could be used in the treatment of various diseases such as cancer.