MAP-1, a novel proapoptotic protein containing a BH3-like motif that associates with Bax through its Bcl-2 homology domains.

A novel Bax-associating protein, named MAP-1 (Modulator of Apoptosis), has been identified in a yeast two-hybrid screen. MAP-1 contains a BH3-like (BH: Bcl-2 homology) motif and mediates caspase-dependent apoptosis in mammalian cells when overexpressed. MAP-1 homodimerizes and associates with the proapoptotic Bax and the prosurvival Bcl-2 and Bcl-X(L) of the Bcl-2 family in vitro and in vivo in mammalian cells. Mutagenesis analyses revealed that the BH3-like domain in MAP-1 is not required for its association with Bcl-X(L) but is required for association with Bax and for mediating apoptosis. Interestingly, in contrast to other Bax-associating proteins such as Bcl-X(L) and Bid, which require the BH3 and BH1 domains of Bax, respectively, for binding, the binding of MAP-1 to Bax appears to require all three BH domains (BH1, BH2, and BH3) of Bax, because point mutation of the critical amino acid in any one of these domains is sufficient to abolish its binding to MAP-1. These data suggest that MAP-1 mediates apoptosis through a mechanism that involves binding to Bax.

Members of the Bcl-2 family of proteins are regulators of cell death that can be grouped into subfamilies of prosurvival and proapoptotic molecules (1,2). They are characterized by the presence of several conserved motifs, known as the Bcl-2 homology (BH) 1 domains, designated BH1, BH2, BH3, and BH4 (1). Although the N-terminal BH4 domain is restricted to some prosurvival members, BH1, BH2, and BH3 can generally be found among members of both the prosurvival and proapoptotic subfamilies (1,2). Mutagenesis and structural studies revealed that the BH1, BH2, and BH3 domains are important for dimerization function (1,2). Association of the prosurvival member Bcl-X L with the proapoptotic member Bax requires the BH1, BH2, and BH3 domains of the former but only the BH3 domain of the latter (3)(4)(5).
In addition to its role as a protein-protein interaction do-main, the BH3 domain of proapoptotic members appears to be important for mediating their proapoptotic function (3,4). This notion is further supported by the recent discovery of a new group of cell death agonists containing only the BH3 domain (6). Similar to other proapoptotic members of the Bcl-2 family, the BH3 domain in these molecules plays an important role in mediating proapoptotic function as well as association with the prosurvival members of the Bcl-2 family. The three-dimensional structure of Bcl-X L demonstrates that the BH1, BH2, and BH3 regions form an elongated hydrophobic cleft to which a BH3 amphipathic ␣ helix can bind (5,7). The other members of the Bcl-2 family that have similar domains are predicted to have a similar tertiary structure. Hence, proapoptotic members of the Bcl-2 family that contain multiple BH domains such as Bax are proposed to exist in two different conformations: one that is similar to Bcl-X L and the other with the BH3 domain rotated outside to allow its insertion into the hydrophobic cleft of a prosurvival protein (1). However, there is a lack of evidence to support the existence of a Bcl-X L -like conformation among the proapoptotic molecules that contain the BH1-3 domains.
In our effort to gain further understanding of the functions of the BH domains of Bax, we used the yeast two-hybrid screen to identify Bax-associating proteins. A novel protein, termed MAP-1, was identified. MAP-1 contains a putative BH3-like domain and induces caspase-dependent apoptosis in mammalian cells when overexpressed. MAP-1 forms homodimers and associates with Bax, Bcl-2, and Bcl-X L in vitro and in vivo in mammalian cells. The BH3-like domain in MAP-1 appeared to be required for mediating apoptosis and its binding to Bax, but not Bcl-X L . Interestingly, in contrast to other Bax-associating proteins, the binding of MAP-1 to Bax requires all three BH (BH1, BH2, and BH3) domains of Bax. MAP-1 thus represents a novel Bax-associating protein that may mediate apoptosis through binding to Bax.

EXPERIMENTAL PROCEDURES
Reagents and Cell Lines-Mono-and polyclonal antibodies against the Myc epitope (9E10, A14) and HA epitope (F7) were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Antibody against PARP (C2-10) was from Dr. G. Poirier, Laval University Medical Research Center, Québec Canada. MCF-7 (gift from Dr. V. Dixit), 293T, and SH-SY5Y cells were grown and maintained as described (8,9). HeLa cells were obtained from the American Type Culture Collection (ATCC), and the cells were grown and maintained according to the directions provided by the suppliers. The peptide protease inhibitors YVAD-fmk, ZVAD-fmk, and DEVD-fmk were from Enzyme System Products. -Zap human and mouse libraries were purchased from Stratagene (La Jolla, CA).
Plasmids Construction-Expression plasmids for hMAP-1, hMAP-1 deletion mutants, mBax, hBcl-X L , hBcl-2, mBimL, and mBid were generated by polymerase chain reaction (PCR) amplification of cDNA clones using the Expand High Fidelity PCR system (Roche Molecular Biochemicals, Germany), with primers incorporating appropriate re-striction sites, and subcloned into pXJ40HA or pXJ40Myc mammalian expression vectors (10) driven by the CMV promoter. Full-length MAP-1 cDNA was subsequently released from pXJHA-MAP-1 and cloned in-frame into GST fusion protein vector pGEX-TK4E (8). A coding region of each construct was sequenced to ensure that no error was introduced. All epitope tags were positioned at the N terminus. Substitution mutants of MAP-1 and Bax were created using the Transformer site-directed mutagenesis kit (CLONTECH). Deletion mutants of MAP-1 were generated by PCR with appropriate primers.
Yeast Two-hybrid Cloning-Yeast two-hybrid screening was done essentially as described previously (11). Briefly, mBax⌬C21 was cloned in-frame with Gal-4 DNA binding domain in pAS2(HA) vector and used to screen a yeast two-hybrid library derived from adult human brain (CLONTECH). Screening and subsequent protein-protein interaction studies were carried out in yeast Y190 reporter strain.
cDNA Library Screening-cDNA fragment of B26 was used to screen human brain (cerebellum) and mouse brain -ZAP cDNA libraries (Stratagene) using standard techniques as described (10).
Northern Blot Analysis-The human multiple tissue Northern blot (CLONTECH) was hybridized with a 32 P-labeled cDNA probe by using ExpressHyb hybridization solution (CLONTECH) according to the instructions of the manufacturer.
In Vitro Binding Assay-Bacterial GST-MAP-1 fusion protein was prepared as described previously (8,12), and the recombinant protein was immobilized onto glutathione-agarose beads. A GST binding assay was carried out according to a previously established procedure (8,12). Briefly, 35 S-labeled proteins were prepared by in vitro transcription/ translation of pXJ-Myc-cDNA constructs using a TNT T7-coupled reticulocyte lysate system (Promega). The integrity of the 35 S-labeled proteins was verified by SDS-polyacrylamide gel electrophoresis (PAGE). Equal amounts of total 35 S-labeled proteins (7 ϫ 10 5 cpm of trichloroacetic acid precipitable counts) were diluted into 0.2 ml of GST binding buffer (20 mM Tris (pH 7.5), 1 mM EDTA, 150 mM NaCl, 0.2% Nonidet P-40) and incubated for 1 h with GST-MAP-1 immobilized on the beads (ϳ2 g). Samples were subsequently washed six times with binding buffer and boiled for 3 min in loading buffer before fractionation on 12% SDS-PAGE. Bound proteins were visualized by autoradiography.
Coimmunoprecipitation-293T cells seeded on a 100-mm plate were transiently cotransfected with 10 g each of expression plasmids driven by the CMV promoter (pXJ40) (8,12) encoding the indicated N-terminal HA-and Myc-tagged proteins using the LipofectAMINE method (Life Technologies, Inc.). 16 h after transfection, cells were harvested for immunoprecipitation assay and analyzed as described (8,12). Briefly, the cells were harvested and lysed in 1 ml of lysis buffer (20 mM Tris, pH 7.5, 150 mM NaCl, 0.2% Nonidet P-40, 1 mM EDTA, 1 mM PMSF, 50 g/ml aprotinin, 10 g/ml leupeptin). An aliquot (1%) of the cell lysate was fractionated on SDS-PAGE for determining the expression of the proteins by Western blot analysis using monoclonal anti-Myc or anti-HA antibodies. The remaining lysates were subjected to immunoprecipitation using 1 g of polyclonal anti-Myc antibody for 1 h on ice and then mixed with 20 l of a 1:1 slurry of protein A-agarose and incubated for another 1 h at 4°C. The agarose beads were washed once in 1 ml of lysis buffer, twice in 1 ml of lysis buffer containing 500 mM NaCl, and once in 1 ml of lysis buffer before fractionation on SDS-PAGE followed by Western blotting analysis.
Apoptosis Assay-Apoptosis assays were carried out as described previously (8). Briefly, cells seeded on a 35-mm plate at 70% confluency were transiently cotransfected with 1.5 g each of the expression plasmids or vector and 0.3 g of pCMV-␤-galactosidase. Vector plasmid was supplemented to bring the total amount of plasmids for each transfection to 5 g. Transfections were carried out with LipofectAMINE for 6 h followed by change of media, and caspase inhibitors were added to the fresh media at this point where indicated. 18 h later, cells were fixed and incubated in X-gal buffer to mark the ␤-galactosidase-expressing cells. Data (mean Ϯ S.D.) are plotted as the percentage of apoptotic cells defined by the number of round blue cells over the total blue cells counted (500 -800 cells) from five randomly chosen fields.
Nuclear Staining of MAP-1-expressing Cells-MCF-7 cells were seeded onto glass coverslips at 70% confluency and cotransfected with 1.5 g of expression plasmids of MAP-1 or vector and 0.3 g of pEGFP (CLONTECH) using LipofectAMINE. 24 h after transfection, the cells were fixed, rinsed with phosphate-buffered saline and then incubated for 2 min with Hoescht 33342 dye (Molecular Probes Inc.) to enable nuclear staining. The cells were subsequently fixed and visualized using a Ziess Axioplan microscope as described previously (9).
Preparation of Nuclear Extracts-As described previously (13), cells were washed twice with phosphate-buffered saline and detached from plates by adding 1 ml of detachment buffer (150 mM NaCl, 1 mM EDTA, pH 8.0, 40 mM Tris, pH 7.6). The cells were centrifuged at 300 ϫ g, and the supernatant was discarded. The pellet was resuspended in 400 l of cold buffer A (10 mM Hepes, pH 7.9, 10 mM KCl, 0.1 mM EDTA, 1 mM dithiothreitol, 0.5 mM PMSF, 2 mM benzamidine, 10 g/ml aprotinin, and 10 g/ml leupeptin). Nuclei were pelleted by centrifuging at 2800 ϫ g and resuspended in 50 l of buffer B (20 mM Hepes, pH 7.9, 0.4 M NaCl, 1 mM EDTA, 1 mM dithiothreitol, 1 mM PMSF, 2 mM benzamidine, 10 g/ml aprotinin, and 10 g/ml leupeptin). The nuclei were lysed by vigorous vortexing, and the mixture was centrifuged at 15,000 ϫ g for 5 min and the supernatant was collected. Protein concentration was determined by Bradford assay, and equal amounts of proteins were used for SDS-PAGE followed by Western blotting analysis.
Reverse Transcriptase-linked PCR-Total RNA was prepared from various cell lines using TRIzol Reagent (Life Technologies, Inc.) according to the manufacturer's protocol. 10 g of total RNA from various cell lines was annealed with oligo(dT) [12][13][14][15][16][17][18] and reversed-transcribed using Moloney murine leukemia virus reverse transcriptase (Roche Molecular Biochemicals) for the synthesis of first strand cDNA. 2 l of the first strand cDNA was subjected to the following PCR conditions: Initial incubation at 94°C for 5 min, then 25 cycles of amplification (94°C for 1 min, 55°C for 1 min, and 72°C for 1 min), ending with a final extension step of 72°C for 10 min. For each experiment, PCR for glyceraldehyde-3-phosphate dehydrogenase was performed to ensure that equal quality and quantity of cDNA were used. Negative controls containing water instead of cDNA were included to ensure that the PCR product obtained was not due to cross-contamination of DNA. The sequence of the primers used are as follows: hMAP-1(ϩ) (5Ј-TAG AGG CTC TTC AGC CTG CCC TGC-3Ј), hMAP-1(Ϫ) (5Ј-ACT CGT TGC CAT ATC CCT TCG TGG-3Ј), glyceraldehyde-3-phosphate dehydrogenase(ϩ) (5Ј-TGA AGG TCG GTG TGA ACG GAT TTG-3Ј), and glyceraldehyde-3-phosphate dehydrogenase(Ϫ) (5Ј-GCC TAA ATG GCC TCC AAG GAG TAA-3Ј).

RESULTS AND DISCUSSION
A yeast two-hybrid screen was performed to identify Baxassociating proteins. Mouse cDNA encoding amino acids 1-171 of Bax (Bax⌬C21), with the hydrophobic C-terminal transmembrane domain deleted to reduce toxicity in yeast (14), was cloned in-frame with the GAL-4 DNA binding domain and used as "bait" to screen a yeast two-hybrid human brain library (CLONTECH). From approximately two million transformants, 12 clones from the primary screen were identified, 2 of which, B26 and B100, were found to interact specifically with Bax⌬C21 but not with other unrelated Gal-4 fusion proteins tested. Upon sequencing, these two clones were found to contain two independent partial complementary DNA (cDNAs) derived from the same gene. Using B26 cDNA fragment as probe, we screened a human brain (cerebellum) cDNA library and obtained several cross-hybridizing cDNA clones. The longest clone, B9, contained a 2.2-kb cDNA insert with a 1056nucleotide open reading frame beginning with a translational initiation consensus sequence and ending with an in-frame stop codon that encoded a protein of 351 amino acids with a predicted molecular mass of 39 kDa (Fig. 1A). We named this protein MAP-1 for modulator of apoptosis. A corresponding murine cDNA clone was subsequently obtained by screening a mouse brain cDNA library. The amino acid sequence deduced from the murine cDNA clone is 77% identical to the human clone (Fig. 1A). Data base searches revealed that the predicted human MAP-1 protein shares extensive amino acid sequence similarity with two other human proteins (Fig. 1A), Ma1 (58%) (15) and Ma2 (47%) (16), suggesting that MAP-1 is a member of a gene family. Ma1 and Ma2 were both initially identified as antigens recognized by autoantibodies present in the sera of some patients with paraneoplastic neurological syndromes with underlying malignancy (breast, colon, parotid for Ma1, and testicular cancer for Ma2) (15,16). At present, no function has been described for the Ma1 and Ma2 proteins.
Upon examination of the amino acid sequence of MAP-1, a region of eight amino acids (amino acids 120 -127) was found to be highly similar to the BH3 domains present among members of the Bcl-2 family (Fig. 1B). The BH3-like motif identified appears to be most similar to the recently identified BH3-B domain (12) present in the N-terminal region of Bid (Fig. 1B). The amino acid sequences in Ma1 and Ma2, corresponding to the region that contains the BH3-like domain, appear to be less similar to BH-3 motif (Fig. 1A).
Northern blot analysis revealed a single 2.8-kb transcript of MAP-1 to be ubiquitously expressed but at a higher level in heart and brain (Fig. 1C). The messages of MAP-1 in total RNA isolated from MCF-7 cells could not be detected in Northern blot analysis. However, MAP-1 messages were readily detectable in MCF-7 and other human cell lines, including HeLa and SH-SY5Y with reverse transcription-linked PCR. Interestingly, Ma1 and Ma2 have a highly restricted expression patterns. Ma1 protein is detected only in brain and testis (15), whereas Ma2 protein is found exclusively in brain (16).
The ability of MAP-1 to associate with Bax was further demonstrated in vivo in transient cotransfection experiments. HA-tagged MAP-1, Bax, Bid, Bcl-2, and Bcl-X L were transiently coexpressed with Myc-tagged MAP-1 into 293T cells, and lysates were prepared for immunoprecipitation. MAP-1 immunoprecipitated specifically with Bax, Bcl-2, and Bcl-X L but not Bid ( Fig. 2A). Other members of the Bcl-2 family, including BimL, Bak, Bad, and Bcl-w, also failed to interact with MAP-1 in the immunoprecipitation assay under the same experimental conditions (data not shown), suggesting that MAP-1 does not bind universally to all members of the Bcl-2 family. In addition, MAP-1 was able to homodimerize in vivo ( Fig. 2A).
It has been shown recently that the conformation of Bax can be affected by certain detergents such as Nonidet P-40 present in the lysis buffer, resulting in nonspecific interaction of it with some members of the Bcl-2 family (17). Because CHAPS (3-[(3cholamidopropyl)dimethylammonio]-1-propanesulfonic acid) was shown not to affect the N-terminal conformation of Bax (17), we performed the coimmunoprecipitation experiment as described in Fig. 2A in transient transfection analysis by lysing the 293T cells in buffer containing 0.2% CHAPS instead of Nonidet P-40. Similar to the results obtained with Nonidet P-40, MAP-1 self-associated and interacted with Bax, Bcl-2, and Bcl-X L , but not Bid under this condition (data not shown).
These interactions observed in the immunoprecipitation assays were further demonstrated in vitro by using the GSTpulldown assay. The in vitro translation of MAP-1 from the pXJ-Myc-MAP-1 plasmid in rabbit reticulocyte lysates generates MAP-1 products, which always appear as two bands when fractionated on SDS-PAGE. The upper band corresponds to the predicted molecular mass (39 kDa) of MAP-1 and appears to have the same mobility as the Myc-tagged MAP-1 expressed in 293T cells (data not shown). Furthermore, anti-Myc antibody was only able to immunoprecipitate efficiently the in vitro translation product of MAP-1 represented by the upper band (data not shown), suggesting that the protein represented by the lower band does not contain the Myc epitope, and it is likely to be generated from alternate usage of initiation codon. In agreement with the in vivo results, GST-MAP-1 was able to pulldown in vitro translated 35 S-labeled MAP-1, Bax, Bcl-2, and Bcl-X L but not the 35 S-labeled Bid (Fig. 2B).
We next examined the possible role of MAP-1 in modulating cell death by transient transfection experiments. MCF-7 cells were cotransfected with MAP-1 expression or control vector together with pCMV-␤-galactosidase or pEGFP as a marker for transfected cells. At 24 h post-transfection, a significant percentage of total ␤-galactosidase-positive (blue) cells cotransfected with MAP-1 displayed apoptotic morphology as evidenced by the appearance of shrunken and rounded cells (Fig.  3A). Nuclear condensation, another hallmark of apoptosis, was also observed in a significant percentage of total green fluorescence-positive (green) cells cotransfected with MAP-1 (Fig. 3B). Proteolytic cleavage of poly(ADP-ribose) polymerase (PARP), which serves as a marker for the activation of caspases in cells undergoing apoptosis (8,18), was also detected in cells transfected with MAP-1 (Fig. 3C). Furthermore, the cleavage of PARP to the signature 85-kDa apoptotic fragment in MCF-7 cells transfected with MAP-1 was completely blocked by treatment with ZVAD-fmk (10 M) (Fig. 3C).
A series of caspase inhibitors were tested for their ability to inhibit apoptosis triggered by MAP-1 overexpression. YVADfmk (50 M), a specific inhibitor of caspase-1-like caspases, and DEVD-fmk (50 M), an inhibitor of caspase-3-like caspases, conferred no and partial protection, respectively, against MAP-1-mediated killing (Fig. 3D). However, ZVAD-fmk (10 M), a wide-spectrum caspase inhibitor, was able to abrogate the apoptotic activity of MAP-1 (Fig. 3D), suggesting that MAP-1 mediates apoptosis through a caspase-dependent pathway. Similar to Bax-mediated apoptosis, overexpression of Bcl-X L effectively blocked the apoptotic effect of MAP-1 (Fig. 3E). The effect of MAP-1 in mediating apoptosis is not restricted to MCF-7 cells, because overexpression of MAP-1 was able to induce apoptosis (judged by round cell and nuclear condensation assays) in both HeLa and SH-SY5Y cells (data not shown). Similar to human MAP-1, mouse MAP-1 induced apoptosis in MCF-7 cells and associated with Bax and Bcl-X L in immunoprecipitation assay in transient cotransfection experiments (data not shown).
We next proceeded to evaluate the requirement of the BH3like domain in MAP-1 for mediating protein-protein interaction and proapoptotic functions. A series of Myc-tagged MAP-1 mutants was created (Fig. 4A). Deletion of the entire BH3-like domain (M1) or substitution mutations of several conserved amino acids (GHE/VLA) in the putative BH3-like domain (M2) of MAP-1 abrogated its interaction with Bax (Fig. 4B). The conserved leucine residue in the BH3 domains of Bax and Bad was reported to be critical for mediating interaction with Bcl-2 or Bcl-X L (19,20) as well as for the N-terminal BH3-B domain of Bid to interact with its C terminus (12). Substitution of this conserved leucine of the BH3 domain, but not the nonconserved histidine, to glutamic acid (M3, M4), in MAP-1 was found to substantially weaken its interaction with Bax (Fig. 4B). Interestingly, in contrast to most proapoptotic members of the Bcl-2 family, the BH3-like domain in MAP-1 did not appear to be required for interaction with Bcl-X L as deletion of the entire BH3-like domain (M1) of MAP-1 did not affect its binding to Bcl-X L (Fig. 4C). In addition, deletion mutant containing the BH3-like domain (M6) failed to interact with Bcl-X L , whereas the N-(M5) and C-(M7) terminal regions of MAP-1 were able to independently associate with Bcl-X L (Fig. 4C).
Similar to many BH3-domain containing molecules (1, 2), the BH3-like domain in MAP-1 appeared to be required for its cell death function as deletion of this domain (M1) or mutations of the highly conserved amino acids in this domain (M2, M3) resulted in nonapoptotic protein (Fig. 4D). The H126E (M4) mutant of MAP-1 that maintained binding to Bax also retained proapoptotic activity at a level similar to the wild-type MAP-1. Despite the presence of the BH3-like domain, the M6 mutant failed to induce apoptosis (Fig. 4D), suggesting that the BH3like domain alone is not sufficient for mediating apoptosis. Although it has been reported that peptides encompassing essentially the BH3 domain are capable of inducing apoptosis (21,22), the BH3 domain alone has also been shown to be insufficient for inducing apoptosis (23,24). On the other hand, because no functional activity such as protein-protein interaction or apoptosis induction has been demonstrated with the M6 mutant, we cannot rule out the possibility that M6 mutant may not fold properly at all, which could result in a totally inactive protein. Additional experiments involving other BH3 domaincontaining protein fragments or peptides of MAP-1 are clearly needed to resolve the issue whether the BH3-like region in MAP-1 is sufficient for mediating protein-protein interaction and/or inducing apoptosis. Taken together, these data suggest that the BH3-like domain in MAP-1 is required for interaction with Bax and mediating apoptosis. Although both the N-and C-terminal fragments of MAP-1 were able to associate with Bcl-X L , none of these mutants (M5 and M7) were found to be proapoptotic (Fig. 4D), suggesting that interaction with Bcl-X L is insufficient to render the molecule proapoptotic.
Several highly conserved amino acid residues of the BH1, BH2, and BH3 domains among Bcl-2 family members have been demonstrated to be critical for heterodimerization functions (20,25,26). To evaluate the requirement of the BH domains of Bax in mediating interaction with MAP-1, substi- 5. BH1, BH2, and BH3 domains of Bax are required for binding to MAP-1. A, Bcl-X L interacts with BH1 and BH2, but not BH3 mutant of Bax. 293T cells were transiently cotransfected with 10 g each of the indicated expression plasmids under the conditions as described in Fig. 2A. Myc-Bcl-X L was immunoprecipitated with the polyclonal anti-Myc antibody. Coprecipitating HA-tagged Bax and mutants were detected by Western blot. Aliquots (1%) of the total lysate (1 ml) were analyzed to estimate the levels of expressed HA-and Myctagged proteins (lower panels). B, BH1, BH2, BH3 domains in Bax are all required to interact with MAP-1. Experimental procedures for transient cotransfection and coimmunoprecipitation were performed as described in A. tution point mutations of these critical amino acid residues in the BH1, BH2, and BH3 domains were made. In agreement with previous data, mutation of the BH3 (L63E), but not the BH1 (G108A) or BH2 (W151A) domains of Bax, abolished its binding to Bcl-X L (Fig. 5A). However, none of the point mutants were able to bind MAP-1 (Fig. 5B), suggesting that the three BH domains (BH1, BH2, and BH3) of Bax are all required for mediating protein-protein interaction with MAP-1. Yeast twohybrid experiments, performed with these Bax mutations in the Bax⌬C21 context to examine their interaction with the B26 clone, yielded similar results (data not shown).
In this report, we describe the cloning and characterization of a Bax-associating protein, MAP-1. Several proapoptotic members of the Bcl-2 family have been identified by using prosurvival members of the Bcl-2 family as bait in the yeast two-hybrid system (27)(28)(29). However, there has been no report in the literature for the identification of protein partners of proapoptotic members of the Bcl-2 family such as Bax using this system. This in part may be due to the toxicity exhibited by some proapoptotic members of Bcl-2 family in yeast which limits its utility in this system (14,30). We took advantage of the observation that, by eliminating the hydrophobic C-terminal tail of Bax, the toxicity of Bax was substantially reduced in yeast, although it was still capable of mediating apoptosis in mammalian cells (14). Using Bax⌬C21 as bait, we managed to screen for protein partners of Bax. Interestingly, we failed to identify any prosurvival members of the Bcl-2 family, including Bcl-X L in our screen. Two possible reasons may account for this observation. First, cDNAs encoding prosurvival members of the Bcl-2 family may be in extremely low abundance in the library. Second, Bax may associate only weakly with the prosurvival members of the Bcl-2 family in this system. Indeed, we noted that interaction of Bax with Bcl-X L was relatively weak in the yeast two-hybrid system (data not shown).
MAP-1 contains a putative BH3 domain and associates with Bax, Bcl-X L , and Bcl-2 in vitro and in vivo in mammalian cells. It mediates caspase-dependent apoptosis when overexpressed. The BH3-like domain in MAP-1 is required for binding to Bax, but not Bcl-X L , and for mediating apoptosis. Interestingly, in contrast to other known Bax-associating proteins, the binding of Bax to MAP-1 requires all of the BH (BH1, BH2, and BH3) domains of Bax, because point mutations affecting any one of the BH domains abolished its binding to MAP-1. MAP-1 thus represents the first protein partner of Bax identified that requires all the BH domains of Bax for association.
Why do some proapoptotic members of the Bcl-2 family contain BH1, BH2, and BH3 domains resembling those of the prosurvival members of Bcl-2 family? The crystal structure of Bcl-X L suggested that it shares similarity to the pore-forming domains of bacterial toxins such as colicins A1 and E1 and diphtheria toxin (7). It has been reported that regions encompassing part of the BH1 and BH2 domains may have poreforming function (2). In addition, Bax was shown to form an ion channel in vitro (2). However, it still remains to be determined whether the Bcl-2 family of proteins actually forms channels in vivo and whether these proteins regulate apoptosis via the creation of ion channels (31).
Instead of the BH1, BH2, and BH3 domains forming a receptor structure, as in the case of Bcl-X L (6), the BH domains in Bax have been suggested to serve independent functions. The BH3 domain of Bax has been proposed to be involved in binding the permeability transition pore and inducing permeability transition and cytochrome c release from the mitochondria (31,32). In addition, it has been proposed that the BH3 domain of Bax mediates its proapoptotic effect by binding to the prosurvival molecule Bcl-X L and, hence, displacing the binding of Apaf-1 from Bcl-X L leading to caspase-9 activation (33,34). However, the validity of this model was recently challenged, because a physiologically relevant level of stable interaction between Apaf-1 and Bcl-X L cannot be demonstrated (35,36). Recently, the BH3-domain-only molecule Bid was found to associate with Bax through its BH1 domain (26). Thus, Bid is the only proapoptotic member of the Bcl-2 family known to associate with Bax prior to MAP-1. This association appears to alter the conformation of Bax and lead to the release of cytochrome c from the mitochondria (37). Hence, multiple mechanisms are likely to be responsible for the apoptotic effects mediated by Bax. Identification of MAP-1 as a novel Bax-associating protein may thus provide further opportunities for investigating the complex mechanisms operated by Bax.