Degradation of Mcl-1 by Granzyme B

Recent studies have suggested that in the absence of Bid, granzyme B (GrB) can utilize an unknown alternative pathway to mediate mitochondrial apoptotic events. The current study has elucidated just such a pathway for GrB-mediated mitochondrial apoptotic alterations. Two Bcl-2 family members have been identified as interactive players in this newly discovered mitochondrial response to GrB: the pro-survival protein Mcl-1L and the pro-apoptotic protein, Bim. Expression of Mcl-1L, which localizes mainly to the outer mitochondrial membrane, decreases significantly in cells subjected to CTL-free cytotoxicity mediated by a combination of GrB and replication-deficient adenovirus. The data suggest that Mcl-1L is a substrate for GrB and for caspase-3, but the two enzymes appear to target different cleavage sites. The cleavage pattern of endogenous Mcl-1L resembles that of in vitro translated Mcl-1L subjected to similar proteolytic activity. Co-immunoprecipitation experiments performed with endogenous as well as with in vitro translated proteins suggest that Mcl-1L is a high affinity binding partner of the three isoforms of Bim (extra-long, long, and short). Bim, a BH3-only protein, is capable of mediating the release of mitochondrial cytochrome c, and this activity is inhibited by the presence of exogenous Mcl-1L. The findings presented herein imply that Mcl-1L degradation by either GrB or caspase-3 interferes with Bim sequestration by Mcl-1L.

induction of a mitochondrial apoptotic cascade is controversial, it has been documented that a deficiency in both Bax and Bak attenuates the mitochondrial response to GrB (13,22). Furthermore, embryonic fibroblasts (MEFs) from Bid Ϫ/Ϫ mice or Bax/Bak double knockout mice still have disrupted mitochondrial transmembrane potential in response to GrB (23). These findings imply that one or more cytosolic mediators other than Bid may act as a link between GrB and the mitochondria. Alternatively, GrB may act directly on mitochondrial outer membrane proteins providing an explanation for the ability of GrB to disrupt the mitochondrial transmembrane potential in a caspase-and Bid-independent manner (10,13).
Mitochondrial response to apoptotic stimuli is determined by the balance between pro-and anti-apoptotic Bcl-2 family members. Anti-apoptotic Bcl-2 family members, including Bcl-2, Bcl-XL, and Mcl-1L (myeloid cell leukemia-1) protect against mitochondrial apoptotic events, whereas pro-apoptotic Bcl-2 family members, including Bax, Bak, and BH3-only proteins, promote the release of apoptogenic proteins from the mitochondria. Mcl-1 is an anti-apoptotic Bcl-2 family protein that was discovered as an early induction gene during myeloblastic leukemic cell differentiation (24). The biological significance of Mcl-1 has been elucidated by recent studies demonstrating that a Mcl-1 deficiency results in peri-implantation embryonic lethality (25). Increased expression of endogenous full-length Mcl-1 is associated with the maintenance of cell viability and decreased expression with cell death (24,26,27). A short splice variant of Mcl-1, Mcl-1S, has recently been identified as a BH3 domain only pro-apoptotic protein (28,29). Despite numerous studies on the induction of Mcl-1, scarce information is available regarding the mechanisms involved in its down-regulation. Based on the effects of proteasome inhibitors, two recent studies proposed that Mcl-1L is degraded by proteasomes in response to UV or actinomycin D, in HeLa or multiple myeloma cell lines, respectively (30,31).
Our studies, described below, are the first to report that Mcl-1L is a direct substrate for caspase-3 and GrB. We have also identified Bim, a BH3-only protein, as a high affinity binding partner for Mcl-1. The Mcl-1L/Bim cooperation may constitute an alternative mitochondrial apoptotic pathway that could be activated by the direct effect of GrB on the mitochondria, independent of Bid.
Cell Lines, Cell Lysates, and Cell Extracts-Jurkat T leukemic cell line and the HeLa tumor cell line were obtained from American Type Culture Collection (ATCC, Rockville, MD). Jurkat cells were grown in RPMI 1640 medium containing 10% fetal calf serum, 20 mM HEPES, 2 mM L-glutamine, and 100 units/ml each of penicillin and streptomycin. The Bak-deficient Jurkat clonal cell line was obtained from wild-type Jurkat cells (22). HeLa cells were grown in Dulbecco's modified Eagle's medium containing 15% fetal calf serum, 20 mM L-glutamine, and 100 units/ml each of penicillin and streptomycin. Cell lysates were prepared with 1% Nonidet P-40, 20 mM Tris base, pH 7.4, 137 mM NaCl, 10% glycerol, 1 mM PMSF, 10 g/ml leupeptin and 10 g/ml aprotinin. Cellular lysates were obtained as the supernatants resulting from centrifugation at 14,000 ϫ g at 4°C for 30 min. To prepare cell extracts for GrB or caspase-3 reactions, cultured Jurkat cells were washed twice with phosphate-buffered saline and then resuspended in ice-cold buffer (20 mM HEPES, pH 7.0, 10 mM KCl, 1.5 mM MgCl 2 , 1 mM sodium EDTA, 1 mM sodium EGTA, 1 mM dithiothreitol, 250 mM sucrose, and protease inhibitors). After incubation on ice for 20 min, cells (2.5 ϫ 10 6 /0.5 ml) were disrupted by Dounce homogenization. Nuclei were removed by centrifugation at 650 ϫ g for 10 min at 4°C.
Molecular Cloning of Mcl-1L-Total RNA was isolated from HeLa cells using RNA STAT-60 Reagent (Tel-Test "B", Inc.). Reverse transcription was carried out with 5 g of total RNA using the oligonucleotide primer Mcl-R, 5Ј-TACAGCTTGGAGTCCAACTGC-3Ј, which is complementary to nucleotides (nt) 1779 -1799 in the 3Ј-untranslated region (UTR) of the Mcl-1L mRNA sequence and SuperScript III RNase H Ϫ reverse transcriptase (Invitrogen). PCR was performed using the GC-Rich PCR System kit (Roche Applied Science). Amplicons containing the entire open reading frame (ORF) of Mcl-1L were generated using the forward primer Mcl-F, 5Ј-CTGGCAATGTTTGGCCTCAAA-3Ј, which corresponds to nt 658 -678 and thus extends by 6 nt into the 5Ј-UTR and the reverse primer Mcl-R (see above). The Mcl-1L amplicon was size-selected and purified using a 1% agarose gel and the QIAquick gel extraction kit (Qiagen  [12][13][14][15][16][17][18]. PCR was performed with the Expand Long Template PCR system (Roche Applied Science) and a forward and reverse primer pair specific for the amino and carboxylterminal ends, respectively, of BimEL , BimL, and BimS ORFs. The sequence of the forward Bim primer is 5Ј-GCCACCATGGCAAAG-CAACCTTCTGAT-3Ј, whereas the reverse primer is 5Ј-TCAATGCAT-TCTCCACACCAG-3Ј. Reaction products were separated on a 1% agarose gel, and a band corresponding in size (ϳ600 bp) to the BimEL ORF was excised and DNA extracted as described for Mcl-1L. The putative BimEL amplicon was subcloned into the vector pCR3.1 as above. Random clones were sequenced also as above to confirm the presence of the BimEL ORF.
Preparation of His-tagged BimL-Mouse amino-terminal histidinetagged BimL was expressed from a recombinant plasmid produced by ligating an NdeI-XhoI-digested BimL amplicon produced with the Expand Long Template PCR system kit (Roche Applied Science) utilizing mouse BimL cDNA and the primers BimL-F, 5Ј-GGAATTCCATATG-GCCAAGCAACCTTCT-3Ј, and BimL-R, 5Ј-CCGCTCGAGTCAATGC-CTTCTCCATACCAG-3Ј, into the NdeI-XhoI-digested bacterial expression vector pET-14b (Novagen). Escherichia coli strain BL21(DE3) cells were transformed and cultured at 37°C in Terrific Broth. The induction of expression was started at an A 600 of 0.8 -1.0 by the addition of 0.4 mM, isopropyl ␤-D-thiogalactoside with continued incubation of the culture at 37°C for 2-3 h. The bacterial pellets were resuspended and sonicated in a buffer containing 5 mM imidazole, 500 mM NaCl, and 20 mM Tris-HCl, pH 7.9. After centrifugation, the cleared supernatants were passed through a His-Bind nickel agarose affinity chromatographic column pre-charged with 50 mM NiSO 4 (Novagen). The columns were washed with wash buffer containing 60 mM imidazole, 500 mM NaCl, and 20 mM Tris-HCl, pH 7.9. His-tagged Bim was eluted with elution buffer containing 400 mM imidazole, 500 mM NaCl, and 20 mM Tris-Cl, pH 7.9, and were further purified using a Sephadex G-50 column equilibrated with phosphate-buffered saline. A single major band was detected by SDS-PAGE stained with Coomassie Blue.
GrB-induced Apoptosis-CTL-free apoptosis was induced by incubation of target cells with GrB (33-330 nM) and replication-deficient adenovirus type V (Ad; 10 -100 pfu/ml) for 6 h. The cells were then washed to remove excess exogenous GrB. To avoid the enzymatic activity of GrB during the lysis procedure, the GrB inhibitor, Ac-IETD-CHO (500 M) was added to the lysis buffer.
Cellular Fractionation and Mitochondria Purification-To obtain an enriched mitochondrial fraction, Jurkat or HeLa cells were suspended in mitochondrial buffer (MIB) composed of 0.3 M sucrose, 10 mM MOPS, 1 mM EDTA, and 4 mM KH 2 PO 4 , pH 7.4, and lysed by Dounce homogenization as previously described (32,33). Briefly, nuclei and debris were removed by a 10-min centrifugation at 650 ϫ g, and a pellet containing mitochondria was obtained by two successive spins at 10,000 ϫ g for 12 min. To obtain the S-100 fraction, the postnuclear supernatant was further centrifuged at 100,000 ϫ g for 1 h at 4°C. To obtain the enriched mitochondrial fraction, the mitochondria containing pellet was resuspended in MIB and layered on a Percoll gradient consisting of four layers of 10, 18, 30, and 70% Percoll in MIB. After centrifugation for 30 min at 15,000 ϫ g, the mitochondrial fraction was collected at the 30/70 interface. Mitochondria were diluted in MIB containing 1 mg/ml bovine serum albumin (at least a 10-fold dilution required to remove Percoll). The mitochondrial pellet was obtained by a 40-min spin at 20,000 ϫ g and used immediately. Purity was assessed by electron microscopy and by enzyme marker analysis (33). For enzyme analysis, the following enzymes were assayed: aryl sulfatase (lysosomes/granules); N-acetyl-␤-D-glucosaminidase, ␣-L-fucosidase, and ␤-glucuronidase (lysosome); lactate dehydrogenase (cytosol); cytochrome oxidase or monoamine oxidase (mitochondria); thiamine pyrophosphatase (Golgi); NADH oxidase (endoplasmic reticulum); and dipeptidyl peptidase IV (plasma membrane). The purity was assessed at 95%, with ϳ5% or less contamination from the microsomal fraction. When indicated, mitochondria were pelleted then incubated in 0.1 M Na 2 CO 3 , pH 11.5, for 20 min on ice to remove loosely attached proteins (34). The alkali treatment was not associated with the release of mitochondrial intermembrane proteins, such as SMAC or AIF. Following alkali treatment, supernatants and mitoplasts were separated by centrifugation and boiled in SDS sample buffer. The fractions were analyzed by immunoblotting.
Release of Mitochondrial Apoptogenic Proteins-Purified mitochondria (50 g of protein) were incubated with His-BimL at various doses as indicated, in 25 l of MIB at 37°C for 30 min. Mitochondria were pelleted by centrifugation at 10,000 ϫ g for 10 min. The resulting supernatants or mitochondria were mixed with SDS sample buffer and analyzed by SDS-PAGE and immunoblotting for the presence of mitochondrial apoptogenic proteins.
In Vitro Transcription-Translation-Mcl-1L, Mcl-1S, and BimEL cDNAs were expressed in the TNT T7 transcription-translation reticulocyte lysate system (Promega). Each coupled transcription-translation reaction contained 1 g of plasmid DNA in a final volume of 50 l in a methionine-free reticulocyte lysate reaction mixture supplemented with 35 S-labeled methionine according to the manufacturer's instructions. After incubation at 30°C for 90 min, the reaction products were immediately used or stored at Ϫ70°C.
In Vitro Cleavage Reaction with Caspase-3 or GrB-In vitro cleavage reactions were performed in total volume of 20 l. The reaction buffer consisted of 20 mM HEPES (pH 7.4), 10 mM KCl, 1.5 mM MgCl 2 , 1 mM EDTA, 1 mM EGTA, 20% glycerol, 1 mM PMSF, 10 g/ml leupeptin, and 10 g/ml aprotinin. Each reaction also contained 3 l of reticulocyte lysate containing 35  Immunoprecipitation-For Mcl-1 and Bim immunoprecipitation experiments cells (5-10 ϫ 10 6 ) and mitochondria (200 g of protein) were lysed in 1% CHAPS buffer (20 mM HEPES, 10 mM KCl, 1.5 mM MgCl 2 , 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 0.1 mM PMSF, and 1% CHAPS). The lysates were precleared with Protein A-or G-Sepharose beads and incubated with anti-Mcl-1 or anti-Bim Abs at 4°C for 4 h. The immune complexes were then precipitated with Protein A-or G-Sepharose beads at 4°C overnight. The pellets were washed four times with the appropriate lysis buffer and boiled for 5 min in SDS sample buffer.
Western Blot Analysis-Proteins in cell lysates, cell extracts, mito-chondria, or S-100 were resolved by SDS-PAGE and transferred to polyvinylidene difluoride membranes, as previously described (35). Following probing with a specific primary Ab and horseradish peroxidaseconjugated secondary Ab, the protein bands were detected by enhanced chemiluminescence (Pierce, Rockford, IL).

Bid-dependent and Bid-independent Mechanisms of Mitochondrial Release of Apoptogenic Proteins in Response to Direct
Application of GrB-GrB has been reported to have a direct effect on mitochondria that results in the release of cytochrome c (10). Such GrB dose-dependent release of apoptogenic proteins, including, cytochrome c, SMAC, and AIF ( Fig. 1A) may be mediated by full-length Bid that in addition to its cytoplasmic localization, is also associated with purified mitochondria. However, full-length Bid is loosely attached to purified mitochondria, because it is entirely removed by treatment with alkali, and therefore it is probably not anchored to the mitochondrial outer membrane (Fig. 1B). Exposure of purified mitochondria to GrB results in the processing of mitochondria-associated Bid as indicated by its reduced level of detection (Fig. 1C). To investigate the significance of mitochondria-attached full-length Bid in the response to direct application of GrB, we utilized liver mitochondria purified from Bid Ϫ/Ϫ mice. Direct application of various doses of GrB (33-330 nM) to murine wild-type liver mitochondria resulted in a significant release of cytochrome c (Fig. 1D). We confirmed that some full-length Bid was associated with mito- chondria purified from the wild-type liver (data not shown). Unexpectedly, Bid-deficient mitochondria also released cytochrome c, albeit at a significantly reduced level. These results suggest that, when GrB is applied directly to the mitochondria, it probably acts on the outer membrane and induces cytochrome c release by two mechanisms that are Biddependent and Bid-independent.
Down-regulation of Mcl-1 in GrB/Ad-treated Target Cells-The involvement of one or more intracellular factors different from Bid in the induction of mitochondrial apoptogenic events has recently been indicated in a study demonstrating the ability of GrB to induce mitochondrial depolarization in both Bid Ϫ/Ϫ MEF and Bax/Bak double knockout MEF (23). In a search for mitochondrial substrates for GrB, we examined the expression of anti-apoptotic Bcl-2 family members in tumor cells exposed to GrB. We observed a significant down-regulation in the expression of Mcl-1 in cells treated with a combination of GrB and replication-deficient adenovirus (Ad). The expression of Mcl-1 was significantly reduced in wild-type Jurkat cells treated with various doses of GrB ( Fig. 2A). Wild-type Jurkat cells express Bak but are deficient in all Bax isoforms that contain the BH3 domain (36 -38). 2 The GrB effect on the level of Mcl-1 expression was also examined in a clonal Jurkat cell line that demonstrates a significant resistance to GrB, as indicated by delayed apoptosis and reduced release of mitochondrial apoptogenic proteins in response to GrB/Ad (22). This clonal cell line, derived originally from the wild-type Jurkat cells, expresses a significantly reduced level of Bak in addition to its inherent deficiency in Bax␣ (36 -38). The expression of Mcl-1 in this Bax/Bak-deficient cell line was also down-regulated in response to GrB/Ad, although an increased dose of GrB was required to achieve this effect ( Fig. 2A). GrB-mediated reduction in expression of Mcl-1 was also observed in Jurkat cells stably transfected with Bcl-2 or Bcl-XL suggesting that Mcl-1 loss is not inhibited by the mechanisms known to attenuate the mitochondrial release of apoptogenic proteins (data not shown). To ensure that the loss in Mcl-1 was mediated by the function of GrB in situ, rather than in the cell lysate, the GrB/Ad-treated cells were lysed in a buffer containing Ac-IETD-CHO (500 M) to inhibit potential activity of excess GrB during the lysis procedure (13). To differentiate between direct GrB activity in situ and activity of caspases stimulated by GrB, wild-type Jurkat cells were treated with GrB/Ad in the presence of a potent caspase inhibitor, Z-VAD-Fmk (100 M), which does not inhibit GrB. In the presence of this inhibitor, the processing of caspase-3 p20 to the p19 and p17 subunits as well as its enzymatic activity were blocked (Fig. 2B).  (Fig. 3A, lanes 4 -6). In Jurkat cellular extract, Ab-1 detected only Mcl-1L, whereas both Ab-2 and Ab-3 detected also a very faint protein band that corresponds in its SDS-PAGE migration to Mcl-1S (Fig. 3A,  lane 1). Additional cross-reactive protein bands with a slightly faster migration than that of Mcl-1S were detected by Ab-2 and Ab-3. To determine the sub-cellular localization of endogenous Mcl-1 proteins, lysates of mitochondria or S-100 fractions were included in the assessment for co-migration with the in vitro translated Mcl-1L and Mcl-1S (Fig. 3A, lanes 2 and 3). The results obtained with the three anti-Mcl-1 Abs suggest that Mcl-1L is associated mainly with the mitochondria (Fig. 3A). In the experiment shown in Fig. 3A, the mitochondria and S-100 fractions were loaded on the SDS-PAGE in cell equivalents, i.e. each lane corresponds to the same number of cells. Thus, the detection level of ␤-actin in the S-100 fraction was similar to that of the cellular extract (lanes 1 and 3), and the Cox IV detection level in the mitochondrial fraction was comparable to the one detected in the cellular extract (lanes 1 and 2). Therefore, the low detection of Mcl-1L in the S-100 fraction is a true representation of the in situ situation. However, Mcl-1L is also detected in the cytosol (mainly by the polyclonal Ab-3) probably due to its partial membrane attachment (Fig. 3B). In contrast to Cox IV, a mitochondrial matrix protein, or VDAC, an integral outer membrane protein that resides on the contact sites between outer and internal membranes, which were not de-tected in the alkali supernatant, Mcl-1 was only partly alkaliresistant. Nonetheless, relative to the mitochondria, there is only a low level of Mcl-1L in the cytosol. Through reverse transcription-PCR and immunoblot analyses performed with HeLa or Jurkat cells, we confirmed that the expression of Mcl-1S is significantly reduced relative to that of Mcl-1L, and therefore it was not consistently detected in cell lysates or S-100 by Abs capable of detecting in vitro translated Mcl-1S. The cross-reactive protein bands detected by Ab-2 in the S-100 fraction and by Ab-3 in the cell extract migrate faster than Mcl-1S and currently remain unidentified.
Direct Degradation of Mitochondrial Mcl-1L by GrB-To investigate whether GrB has a direct effect on Mcl-1L, purified mitochondria from either Jurkat or HeLa tumor cell lines were treated with various doses of purified GrB (Fig. 4, A-C). A dose-dependent loss in the level of Mcl-1L was detected by three Mcl-1-specific Abs. No change was detected in the expression levels of Bcl-XL, Bcl-2 or the three isoforms of Bim. Similar GrB-mediated Mcl-1L loss was detected in HeLa cells (Fig. 4B) and in breast carcinoma cell lines (data not shown). Utilizing a polyclonal anti-Mcl-1 Ab (Ab-3), a p26 Mcl-1 cleavage product was detected in GrB-treated mitochondria (Fig. 4A). To ascertain that the observed loss in Mcl-1L expression was mediated by GrB rather than by GrB-stimulated caspases, the experiment was repeated in the presence of Z-VAD-Fmk (Fig. 4C). This inhibitor did not have any effect suggesting that the reduction in the detection level of mitochondrial Mcl-1L was mediated by GrB directly or by non-caspase enzymatic activity associated with the mitochondria and stimulated by GrB. GrB also reduced the detection level of Mcl-1L in the S-100 fraction (Fig. 4D). These results suggest that endogenous Mcl-1L present in the mitochondrial or cytosolic fractions is a substrate for GrB activity. (Fig. 2) indicated that the loss in Mcl-1 expression was mediated by both GrB and caspase activity. To confirm this observation, Jurkat cell extracts were treated with recombinant caspase-3 (Fig. 5). Endogenous Mcl-1L and Mcl-1S were sensitive to caspase-3 activity as their expression was reduced at a dose of caspase-3 as low as 10 nM, whereas at a dose range of 5-100 nM neither a loss in the level of pro-caspase-8 was detected (not shown), nor was there a significant level of processing of XIAP (Fig. 5A). However, XIAP and pro-caspase-8 are susceptible to a higher dose range (100 -1000 nM) of recombinant caspase-3 (Fig. 5B).  (Fig. 7). Immunoblotting with Mcl-1-specific Ab (Ab-3) revealed a similarity between cleavage products obtained from in vitro trans-  4, 10, and 13). Likewise, treatment with GrB also generated the same size products from in vitro translated Mcl-1L, cell extract, and S-100 fractions (lanes 5, 11, and 14). Endoge-nous Mcl-1S cleavage products similar to those produced by protease digestion of in vitro translated Mcl-1S were not detected in the cellular fractions, potentially due to its low level of expression in situ. Although the Mcl-1 cleavage patterns mediated by GrB and caspase-3 appear to be different, the exact aspartic acid residues targeted by these enzymes remain to be identified.

Co-immunoprecipitation of Mcl-1 with Bim-Because
Mcl-1L is an anti-apoptotic Bcl-2 family member, we reasoned that the biological significance of its degradation might be manifested by the release of a pro-apoptotic Bcl-2 family member from a physical interaction with Mcl-1L. To this end, we screened the immunoprecipitated pellet of endogenous Mcl-1 for the presence of BH3-only proteins. Immunoblot analysis of immunoprecipitated Mcl-1 detected the co-presence of three Bim isoforms (Extra-Long (EL), Long (L), and Short (S)) (Fig. 8, A and B) but no Bid or Bad (data not shown). The binding of Bim to Mcl-1L was specific as the immunoprecipitation of Mcl-1L co-immunoprecipitated Bim (Fig. 8A), and vice versa, the immunoprecipitation of Bim co-immunoprecipitated Mcl-1L (Fig. 8B). Immunoprecipitation studies were performed with cell extract, lysate of purified mitochondria, or S-100 fractions. Because the anti-Mcl-1 Ab utilized for the immunoprecipitation and immunoblotting (Ab-2) can detect in the cell extract both a crossreactive protein and Mcl-1S, the identity of the protein that migrates faster than Mcl-1L, remains unclear and designated as cross-reactive. Mcl-1L was efficiently detected by Ab-2 in the cell lysate and in mitochondria, and was also effectively immunoprecipitated from these cellular fractions (note depletion of Mcl-1L in the remaining supernatant). Because Ab-2 does not detect a significant level of Mcl-1L in the S-100 fraction, no immunoprecipitated Mcl-1L was detected (Fig.  8A, left panel). BimEL, -L, and -S were effectively coimmunoprecipitated with Mcl-1L (Fig. 8A, right panel). Interestingly, a reduced level of Bim was also co-immunoprecipitated with the Mcl-1 cross-reactive protein present in the S-100 fraction (Fig. 8A, right panel, bottom). Immunoprecipitation of Bim from the cell extract or purified mitochondria, effectively co-immunoprecipitated Mcl-1L (Fig. 8B, right  panel). However, immunoprecipitation of Bim from the S-100 fraction was not accompanied by precipitation of Mcl-1L, rather, the cross-reactive protein was co-immunoprecipitated. Because the cross-reactive protein is detected by Mcl-1-specific mAb (Ab-2) and is also capable of binding Bim, it may represent a Mcl-1 isoform not yet identified. Bim has been reported to bind to anti-apoptotic Bcl-2 family members, Bcl-2 and Bcl-XL, in cells transfected with these proteins or in cells stimulated to undergo apoptosis (21,40). However, immunoprecipitation of naturally expressed, endogenous Bim from non-apoptotic cells, was associated with Mcl-1L, but not with Bcl-2, Bcl-XL, or Bcl-w (data not shown), further indicating the selectivity and high affinity of Bim for Mcl-1L.
To further investigate the specificity and affinity of binding between Mcl-1 and Bim, we examined whether the corresponding in vitro translated products will be co-immunoprecipitated with each other. To this end, 35  Bim-mediated Release of Mitochondrial Cytochrome c Is Inhibited by Mcl-1L-To investigate the ability of Mcl-1L to repress the apoptotic activity of Bim, we employed a cell-free cytochrome c releasing assay. Initially, the ability of Bim to mediate the release of mitochondrial apoptogenic proteins, was assessed by applying various doses of recombinant Histagged BimL to mitochondria purified from wild-type Jurkat cells. Recombinant BimL mediated the release of cytochrome c in a dose-dependent manner (Fig. 10A). The detection of cytochrome c in the mitochondrial supernatant coincided with its reduced presence in the mitochondrial pellet. However, the presence of exogenous McL-1L resulted in a dosedependent inhibition of BimL-mediated cytochrome c release (Fig. 10B). These results suggest that the interaction of BimL with Mcl-1L interferes with the cytochrome c releasing capability of BimL.

DISCUSSION
Current understanding of GrB apoptotic activity does not fully explain the direct effect of GrB on mitochondrial permeability transition or the ability of GrB to cause mitochondrial depolarization in cells deficient in Bid, or Bax and Bak. The inability to satisfactorily elucidate the mechanisms underlying the function of GrB led to the assumption that one or more unknown intracellular factors are involved in the mitochondrial response to GrB. In the current study we identified two interactive Bcl-2 family members, Mcl-1L and Bim, as novel mediators of the mitochondrial response to GrB.
The interaction of Mcl-1L and Bim at the outer mitochondrial membrane fits into the current dogma for regulation of mitochondrial apoptosis by heterodimerization between prosurvival and pro-death Bcl-2 family members. The Bcl-2 gene family includes three groups of proteins that regulate stressinduced apoptosis. The pro-survival members include Bcl-2, Bcl-XL, Bcl-w, A1, and Mcl-1 that share three or four regions of homology (BH1-4) (41). The two pro-death sub-groups of the Bcl-2 family include Bax/Bak-like proteins, which are structurally similar to Bcl-2, as well as the more distantly related BH3-only proteins (41,42). The BH3 domain is essential for the binding of the BH3-only proteins to the anti-apoptotic members of the family and for their ability to kill cells. Heterodimerization is mediated by the insertion of the BH3 domain of the pro-apoptotic molecules into a hydrophobic cleft formed by the BH1, BH2, and BH3 domains on the surface of the anti-apoptotic proteins. Such a mechanism of heterodimerization and potential neutralization of Bim apoptotic activity may underlie the binding observed between endogenous as well as in vitro translated products for Mcl-1L and Bim. In vitro translated BimEL has the ability to also interact with Mcl-1S, but it is not clear whether such binding occurs in situ.
Like many pro-and anti-apoptotic members of the Bcl-2 family, Mcl-1L and the three Bim isoforms have a C-terminal transmembrane domain, which can target them to the cytoplasmic side of intracellular membranes. In contrast to Mcl-1L, Mcl-1S does not possess a C-terminal transmembrane domain and therefore is expected to localize mainly to the cytoplasm. Due to low expression of endogenous Mcl-1S (at the RNA and protein levels) in both Jurkat and HeLa cells it was not consistently detected by immunoblotting performed with various sub-cellular fractions.
Current literature suggests that down-regulation in expression of Mcl-1L would facilitate the apoptotic process. In human neutrophils (43), differentiating human myeloblastic leukemia (44), and multiple myeloma cell lines (30) down-regulation of Mcl-1 by antisense oligonucleotides causes a rapid entry into apoptosis. Although Mcl-1 small interference RNA did not induce apoptosis in HeLa cells (31), Mcl-1 antisense therapy chemosensitized human melanoma to subsequent treatment with decarbazine (45). The outcome of Mcl-1 elimination in different cells may vary because of the co-expression of redundant anti-apoptotic Bcl-2 family members, including Bcl-2, Bcl-XL, and Bcl-w. However, its significance for cell survival is suggested by the apoptotic sensitization mediated through its down-regulation (45). The specific apoptotic mechanisms induced by Mcl-1 down-regulation are not clearly understood. Release of Bim from Mcl-1 sequestration may lead to Bim activation. Bim-Mcl-1 interaction appears to be both specific and of a high affinity nature. The specificity in Mcl-1-Bim binding was indicated by the co-immunoprecipitation of the three Bim isoforms, but not Bid or Bad, with endogenous Mcl-1L from either Jurkat cell extracts or purified mitochondria. Although under apoptotic conditions or when overexpressed, Bim is capable of interaction with various anti-apoptotic Bcl-2 family members, upon immunoprecipitation from extracts of non-apoptotic cells, it preferentially co-precipitated Mcl-1L, but not Bcl-2, Bcl-XL, or Bcl-w. The binding between in vitro translated BimEL and Mcl-1 that occurs at 4°C with no need for prior co-incubation suggests the high affinity nature of this interaction.
Alternative splicing generates at least three isoforms of Bim: BimS, BimL, and BimEL (46). All three isoforms are potent inducers of apoptosis but may be subjected to differential regulation. The additional regions present in the longer isoforms attenuate their activity, by allowing Bim sequestration to the microtubule-associated dynein motor complex through binding to dynein light chain LC8/DLC1 (46,47). Certain apoptotic stimuli, such as taxol, induce the release of a BimL-DCL1 complex from the dynein motor complex. The BimL-DCL1 complex then relocalizes to the membranes of intracellular organelles, such as mitochondria, where it binds to and inhibits the function of Bcl-2. According to our sub-cellular fractionation experiments, the three Bim isoforms are present in both S-100 and purified mitochondria from non-apoptotic Jurkat and HeLa cells. Other studies also reported the constitutive presence of BimEL in the mitochondria (48). In support of these observations several studies have demonstrated the co-localization of Bim and LC8 to heavy membrane fractions that are enriched in mitochondria (47,48). Currently, the role of LC8/ DCL1 in regulating mitochondrial attached Bim is unknown (48). Also, several studies suggested that the apoptotic activity of Bim is regulated by its phosphorylation status (48,49). The role of Bim phosphorylation in its interaction with Mcl-1 has not yet been investigated. Noteworthy, we did not observe changes in the SDS-PAGE migration pattern of Bim following exposure of cells to GrB/Ad or treatment of mitochondria with GrB (data not shown).
The constitutive, rather than apoptosis-induced binding of Bim to a pro-survival protein suggests that Bim proteins are the target for Mcl-1L inhibition. The ability of exogenous Mcl-1 to inhibit recombinant Bim from releasing mitochondrial cytochrome c, provides further support to a potential regulatory role that Mcl-1L plays in Bim function. Thus, the cleavage of Mcl-1L by either GrB or GrB-stimulated caspases is a mitochondrial apoptotic event that may alter the activation status of Bim. However, this hypothesis requires further investigation.
The mitochondrial mechanisms underlying the response to Bim are not yet resolved. Studies conducted with Bax Ϫ/Ϫ Bak Ϫ/Ϫ MEF suggested that MEF cell death induced by a retrovirus expressing BimS or tBid requires the function of either Bax or Bak (20,50). Another study conducted with purified liver mitochondria from wild-type or Bak Ϫ/Ϫ mice suggested that Bim, but not tBid, induces cytochrome c release in both Bak-dependent and Bakindependent manners (51). Moreover, Bim, but not tBid, was able to induce cytochrome c release from yeast mitochondria that do not possess Bcl-2 family members. Bim-mediated mitochondrial permeability transition and cytochrome c release are VDAC-dependent, because they are both inhibited by a VDAC-specific Ab (51). Regardless of the underlying functional mechanisms, the ability of Bim that is free from Mcl-1 sequestration to mediate mitochondrial permeability transition and release of apoptogenic proteins is predicted to be the basis of the novel pathway we identified for GrBmediated mitochondrial apoptosis.
The cleavage of Mcl-1L by either GrB or caspase-3 eliminates the presence of full-length Mcl-1L while producing several cleavage products. One of the cleavage products generated by GrB and one of those generated by caspase-3 appear to be sustained in the presence of increased doses of either enzyme (Fig. 6A) or a prolonged reaction period (data not shown). Thus, GrB or caspase-3 proteolytic activities may abrogate the protective effect of Mcl-1L and at the same time produce proapoptotic cleavage products. In an analogous scenario, Bcl-2 and Bcl-XL that are cleaved by caspase-3 are known to possess potent pro-apoptotic activity (52,53). However, the Mcl-1 cleavage sites targeted by GrB and caspase-3 need to be identified to further investigate the nature and the biologic significance of its cleaved products.