Cytoprotective Peptide Humanin Binds and Inhibits Proapoptotic Bcl-2/Bax Family Protein BimEL*

Humanin (HN) is a recently identified endogenous peptide that protects cells against cytotoxicity induced by various stimuli. Recently, we showed that HN binds to and inhibits Bax, a proapoptotic Bcl-2 family protein, suggesting a mechanism for HN action. In this study, we identified Bim, a Bcl-2 homology 3-only member of the Bcl-2/Bax family, as an additional HN target protein. Using in vitro protein binding, immunoprecipitation, and coimmunolocalization assays, we demonstrated that HN binds directly to the extra long isoform of Bim (BimEL) but not the long (BimL) or short (BimS) isoforms. HN also protects cells against apoptosis induced by BimEL but not BimL and BimS in gene transfection studies. In contrast, mutants of HN which failed to bind BimEL failed to protect from BimEL-induced cell death. Moreover, HN inhibited BimEL-induced release of SMAC and cytochrome c from mitochondria isolated from bax–/–cells, indicating that HN can suppress BimEL independently of its effect on Bax. Finally, we demonstrate that HN prevents BimEL-induced oligomerization of Bak using isolated mitochondria. Taken together, our results indicate that the inhibition of BimEL may contribute to the antiapoptotic properties of the HN peptide.

Apoptosis or programmed cell death is a genetically regulated cellular suicide mechanism that plays critical roles in normal development, tissue homeostasis, and elimination of infected or damaged cells (1,2). Mitochondria represent crucial organelles for the integration of various cell death stimuli and execution of the cell death program (3,4). Mitochondria are capable of releasing several apoptogenic proteins into the cytosol, including cytochrome c, AIF, SMAC/Diablo, endonuclease G, and Omi/HtrA2 (5).
The integrity of mitochondrial membranes is controlled primarily by a balance between the antagonistic actions of the proapoptotic and antiapoptotic members of the Bcl-2 family. Bcl-2 family proteins comprise three principal subfamilies: (a) antiapoptotic members, such as Bcl-2/Bcl-X L , which possess the Bcl-2 homology (BH) 1 domains, BH1, BH2, BH3, and BH4; (b) proapoptotic members, including Bax, Bak, and Bok, which have the BH1, BH2, and BH3 domains; and (c) BH3-only proteins (BOPs), such as Bid, Bim, Bad, Bik, Puma, and Bmf, which generally possess only the BH3 domain (6). The BH3 domain mediates interactions among Bcl-2 family proteins, allowing for networks of protein interactions (7,8). Most BOPs function as antagonists of antiapoptotic Bcl-2 family proteins. However, some BOPs operate as both antagonists of the antiapoptotic proteins and as agonists of proapoptotic family members Bax and Bak. These agonists bind via their BH3 domains, inducing oligomerization of Bax and Bak. The pore-forming capability of oligomerized Bax and Bak results in the destabilization of the mitochondrial outer membrane and the subsequent release of the death molecules from the confines of these organelles. Cells derived from bax/bak double knock-out mice are resistant to a wide spectrum of apoptotic stimuli, including BOPs (9,10), formally demonstrating that Bax and Bak mediate death signals from various BOPs.
The BOP Bim was first identified as a Bcl-2-interacting protein (11) but was subsequently shown to bind both antiapoptotic and proapoptotic members of the Bcl-2 family. Bim is expressed in hematopoietic, epithelial, neuronal, and germ cells (12), and alternative mRNA splicing generates three major isoforms: short (BimS), long (BimL), and extra long (BimEL), with BimEL representing the predominant isoform in most tissues. Recent studies using gene knock-out mice have identified the bim gene as a major regulator of apoptosis in the lymphoid system. bim plays a critical role in thymic education and peripheral deletion of T lymphocytes, as well as myeloid cell homeostasis (13)(14)(15). The bim gene is also required for death of B lymphocytes in response to cross-linking of surface immunoglobulin (16). Bim proteins also trigger apoptosis in neurons, hematopoietic progenitors, and other types of cells (13,(17)(18)(19), where their activity is regulated both at the level of protein expression and by post-translational modifications such as phosphorylation (20 -22). In healthy cells, BimEL and BimL are sequestrated on the microtubular dynein motor complex through direct interaction with the dynein light chain LC8 (23). In contrast, BimS does not bind to LC8, a situation that may explain the superior potency of BimS compared with BimEL and BimL in gene transfection-based apoptosis assays.
Certain stress conditions, such as cytokine deprivation, ␥-irradiation, and exposure to microtubule-targeting drugs, cause release of BimEL and BimL from the dynein motor complex, resulting in their translocation to mitochondria. Bim is among the few BOPs that can function as agonists of Bax and Bak, inducing oligomerization of these proteins in a BH3-dependent manner and triggering apoptogenic changes in mitochondrial membrane permeability (10).
By functional screening of a brain cDNA library derived from an Alzheimer disease patient, Hashimoto et al. (24,25) identified an endogenous peptide, termed Humanin (HN), which suppresses neuronal cell death initiated by Alzheimer diseaserelated insults, including amyloid ␤-peptide and mutant presenilins in primary neurons and neuronal cell lines. HN is a short 24-residue polypeptide, MAPRGFSCLLLLSEIDLPVK-RRA, which is apparently produced at the highest levels in testis, colon, Alzheimer disease-affected brain tissue, and some tumor cell lines (27,28). The mechanisms responsible for regulating levels of HN peptide in vivo are unknown, although post-translational control of HN protein stability by interactions with other proteins is among the reported possibilities (23). Interestingly, an open reading frame corresponding to HN is found embedded in a ribosome gene within the mitochondrial genome, in addition to HN-encoding genes within the nuclear genomes of several animal species (11), raising the possibility of gene transfer from mitochondrial to nuclear genome.
The cytoprotective mechanism of HN has been controversial. Early data suggested that HN is secreted from cells (24,29) and implied that a cell surface receptor is targeted by this peptide. More recently, however, our laboratory discovered an alternative mechanism of action for HN, showing that this peptide binds to Bax and prevents the translocation of this proapoptotic protein from cytosol to mitochondria, thereby protecting cells from apoptosis-inducing insults that trigger the mitochondrial (but not nonmitochondrial) pathway for cell death (28). In contrast, HN does not bind multiple other Bcl-2 family members, including Bak, Bok, Bcl-2, Bcl-X L , Mcl-1, and Bcl-B. Here, we extended the protective action of HN by identifying BimEL as a new target of HN. We demonstrate that the specific binding of HN to BimEL abolishes its proapoptotic activity by preventing BimEL-induced activation of Bax and Bak and protecting cells from BimEL-induced cell death. Although the normal in vivo targets of HN remain to be clarified, these results broaden the scope of proapoptotic Bcl-2/Bax family proteins that the HN peptide is capable of antagonizing, suggesting expanded opportunities for exploiting this antiapoptotic peptide directly or indirectly for possible development of cytoprotective therapies.
Plasmid Constructions and Directed Mutagenesis-pEF-PGKhygro expression plasmids incorporating the N-terminal EE (EYMPME) epitope tag and encoding BimEL, BimL, and BimS have been described previously (22). The plasmid pcDNA3-myc-Bcl-X L , incorporating a myc epitope tag, has also been described previously (28). A cDNA encoding human Bak was subcloned into the BamHI and EcoRI sites of pEGFP-C1 plasmid (Clontech). A cDNA containing the open reading frame of HN without additional flanking sequences was generated by PCR using an expressed sequence tag clone encoding full-length HN as a template (BE899497). The resulting PCR products were digested with restriction endonucleases and subcloned into the XhoI and HindIII sites of pEGFP-C1 (Clontech). Site-specific mutants of HN were created by PCR using QuikChange site-directed mutagenesis (Stratagene). For replacing cysteine 8 with proline (HN(C8P)), the following primers (XX IDT) were used: sense, 5Ј-CCACGAGGGTTCAGCCCTCTCTTACTTT-TAACC-3Ј; antisense, 5Ј-GGTTAAAAGTAAGAGAGGGCTGAACCC-TCGTGG-3Ј. Mutation was verified by DNA sequencing. The pRSET-BimEL plasmid encoding the recombinant His 6 -tagged BimEL protein was a generous gift of Yoshide Tsujimoto (30).
Fluorescence Polarization Assays (FPAs)-For FPAs, various concentrations of His 6 -BimEL or His 6 -SMAC fusion proteins were incubated with 40 nM FITC-conjugated synthetic purified HN peptide for 10 min in the dark in phosphate-buffered saline (PBS), pH 7.4, as described previously (28). Fluorescence polarization was measured in PBS buffer, using an Analyst AD Assay Detection System (LJL Biosystem).
Caspase Activity Measurements-The caspase-3-like protease activity in cell lysates was measured using a fluorometric substrate Ac-DEVD-AFC (Alexis Corporation, San Diego), which was dissolved in dimethyl sulfoxide and stored as a 10 mM solution at Ϫ20°C. HEK293T cells were collected, washed in cold PBS, and lysed at 4°C in buffer A (50 mM HEPES, pH 7.5, 150 mM NaCl, 20 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 10 g/ml aprotinin, and 0.3% Nonidet P-40). Homogenates were centrifuged (13,000 rpm, 10 min), and the protein concentrations of the supernatants were determined using a Bradford protein assay kit (Bio-Rad). Equal amounts of supernatants (normalized for protein content) were mixed in wells of 96-well microtiter plates in a 100-l final volume, in buffer B (50 mM HEPES, pH 7.4, 100 mM NaCl, 1 mM EDTA, 10% sucrose, 0.5% CHAPS, 5 mM dithiothreitol, and 100 M Ac-DEVD-AFC substrate (caspase-3/ 7)). Substrate cleavage was monitored continuously by spectrofluorometry, in kinetic mode, using Fluorolite 1000 (Dynatech Laboratories). Data are reported as relative fluorescence units of product produced/ min/10 g of total protein. Data were analyzed using the PRISM Statistics software package employing an unpaired t test method.
Apoptosis Assay-HEK293T cells were transfected as described above. After 18 h, both floating and adherent cells (after trypsinization) were collected, washed with PBS, fixed with PBS containing 3.7% formaldehyde, and stained with 2 g/ml DAPI in PBS to visualize the nuclei by ultraviolet microscopy. The percentages of apoptotic cells were determined by counting 200 GFP-positive cells, scoring cells having nuclear fragmentation and/or chromatin condensation as apoptotic (mean Ϯ S.D.; n ϭ 3).
Protein Purification-The recombinant His 6 -tagged fusion protein BimEL was expressed from pRSET-BimEL plasmid transformed in Escherichia coli BL21 DE3. Briefly, cells were grown in 2 liters of Luria-Bertani broth with 50 g/ml ampicillin at 37°C to a A 600 of 1.0, then 1 mM isopropyl ␤-D-thiogalactopyranoside was added, and the cultures were incubated at 25°C for 4 h. Cells were then recovered and incubated with 1 mg/ml lysozyme in 20 mM HEPES buffer, pH 7.4, 1 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride, followed by sonication. Cellular debris were pelleted by centrifugation at 16,000 ϫ g for 15 min, and the resulting supernatants were applied to nickelnitrilotriacetic acid (Ni-NTA)-agarose resin (Qiagen, Inc., Valencia, CA).
The His 6 -BimEL⅐Ni-NTA complex was washed five times with 20 column volumes of 20 mM HEPES buffer, pH 7.4, and eluted with a linear gradient of 0 -0.25 M imidazole in 20 mM HEPES, pH 7.4, and 1 mM dithiothreitol.
Peptide Synthesis-Peptides were synthesized using Fmoc synthesis and DIC/HOBt coupling with an Advanced Chem Tech 350 multiple peptide synthesizer. HN and its derivatives were synthesized on Fmoc-Alanine Wang resin to give peptides with N-terminal amino groups and C-terminal carboxyl groups. HN-Ahx-r8, its derivatives, and other peptides were synthesized on Rink amide resin. The HN-Ahx-r8 peptides were amidated on the C termini and had N-terminal amino ends. The Bak BH3 peptide was acetylated and amidated on the ends. FITC-HN is MAPRGFSCLLLLSEIDLPVK(FITC)RRA with FITC linked to the Lys side chain. Fmoc amino acids and resins were purchased from Nova Biochem.
In the synthesis of HN and its derivatives, Fmoc-(FmocHmb)Leu-OH was used for assembling Leu 11 , and the symmetric anhydride of Fmocleucine was used for adding Leu 10 to the sterically hindered (Hmb)Leu. Fmoc-Lys(Mtt) was used in addition to the above for the synthesis of FITC-HN.
The peptides were deprotected and cleaved from the resin by treatment with 94% trifluoroacetic acid, 2.5% H 2 O, 2.5% 1,2-dithioethane, and 1% triisopropysilane for 2 h at room temperature; r8 peptides were treated for 6 h. The crude peptides were obtained by the addition of cold diethyl ether and purified with a Gilson high performance liquid chromatography instrument. The purified peptides was analyzed and confirmed by matrix-assisted laser desorption ionization time-of-flight mass analysis with an Applied Biosystems Voyager System 6264.
Mitochondria Purification-Mitochondria were isolated from HeLa, MEF baxϩ/ϩ, and baxϪ/Ϫ cells. Mitochondria were purified using a differential centrifugation method. Briefly, cells were centrifuged at 1,000 ϫ g for 5 min at 4°C, washed, and resuspended in fractionation buffer (10 mM HEPES-KOH, pH 7.4, 250 mM D-mannitol, 0.5 mM EGTA) supplemented with protease inhibitors (20 g ml Ϫ1 leupeptin, 20 g ml Ϫ1 aprotinin, and 1 mM phenylmethylsulfonyl fluoride). Cells were homogenized by 70 strokes in a Dounce homogenizer. Homogenates were centrifuged twice at 1,000 ϫ g for 5 min at 4°C to pellet nuclei and unbroken cells, which were discarded. The resulting supernatants were centrifuged at 10,000 ϫ g for 10 min at 4°C, and the mitochondria pellet was washed twice with fractionation buffer. Mitochondria were finally resuspended in fractionation buffer at 5 g/l and used within 2 h.
Mitochondria Protein Release Assays-Purified recombinant human His 6 -BimEL (10 -600 ng) was preincubated with or without various concentrations of synthetic HN peptide or mutant HN(C8P) peptide for 10 min at room temperature. The untreated or peptide pretreated BimEL protein samples were then mixed with 50 g of mitochondria in a total volume of 50 l of fractionation buffer at 30°C for 1 h. Samples were then centrifuged at 10,000 ϫ g for 10 min to obtain pellet and supernatant fractions, measuring cytochrome c, SMAC, and Hsp60 by immunoblotting using monoclonal cytochrome c antibody (BD Biosciences), rabbit polyclonal SMAC antibody, and monoclonal Hsp60 antibody (Stressgen).
Bak Oligomerization Assays-Bak oligomers were detected using the noncleavable protein cross-linker, bismaleimidoxane (BMH). BMH stock solution was prepared at 100 mM dissolved in dimethyl sulfoxide. Various amounts of recombinant His 6 -BimEL alone or together with various concentrations of HN peptide were preincubated for 10 min at room temperature in fractionation buffer, then 50 g of isolated mitochondria from baxϪ/Ϫ MEF cells was added to a final volume of 50 l. After incubation at 30°C for 1 h, 10 mM BMH cross-linker was added for 30 min at room temperature. Reactions were stopped by the addition of Laemmli buffer. Samples were analyzed by SDS-PAGE/immunoblotting (12% gels) using polyclonal anti-Bak antibody (Upstate Biotechnology).
Bax Activation Assays-To examine conformational changes in Bax associated with activation of this protein, 100 g of mitochondria was lysed in fractionation buffer supplemented with 1% CHAPS and subjected to immunoprecipitation with anti-Bax 6A7 monoclonal antibody (Sigma). Then, the resulting immune complexes were analyzed by immunoblotting using polyclonal Bax (N-20) antibody (Santa Cruz Biotechnology).
SMAC Antibody Production-Polyclonal antisera recognizing SMAC were generated in New Zealand White rabbits using peptide (AR-49) or recombinant protein (AR-50A; B) immunogens. SMAC recombinant protein was produced as a GST fusion protein and affinity-purified as described previously (31). The SMAC anti-peptide antibody (AR-49) was generated using a synthetic peptide (NH 2 -AVPLAQKSEPHSLS-SEALC-amide) corresponding to residues 55-73 of human (hu) SMAC, which was synthesized with an N-terminal cysteine appended to permit conjugation to maleimide-activated carrier proteins keyhole limpet hemocyanin and ovalbumin (Pierce, Inc.) (32). The monospecificity of all antibodies for their intended protein targets was confirmed by SDS-PAGE/immunoblot analysis.

RESULTS
HN Binds to BimEL-We showed previously that HN can bind the proapoptotic protein Bax but not several other antiapoptotic Bcl-2 family members (28). During attempts to extend studies of HN to other Bcl-2 family proteins, we noticed binding to BimEL by both coimmunoprecipitation from cell lysates containing the expressed proteins and by in vitro protein binding assays. For example, using lysates from HEK293T cells cotransfected with plasmids encoding BimEL and HN (expressed with an N-terminal GFP), we performed immunoprecipitation experiments using anti-Bim polyclonal antibody and analyzed the resulting immune complexes for associated GFP-HN by SDS-PAGE/immunoblotting (Fig. 1A). These experiments demonstrated that GFP-HN can interact with both endogenous (lane 3) or transfected (lane 4) BimEL (expressed here with an N-terminal EE epitope tag). In contrast, GFP control protein did not associate with BimEL. Similar results were obtained using HN expressed with a C-terminal GFP (data not shown).
Direct interaction of HN with BimEL was demonstrated by FPAs using recombinant purified BimEL protein and a FITCconjugated synthetic HN peptide. Mixing various concentrations of His 6 -BimEL with a fixed amount of FITC-HN peptide resulted in concentration-dependent increases in fluorescence polarization, allowing an estimate of affinity of the interaction of K d ϭ 50 nM. In contrast, His 6 -SMAC protein did not bind the FITC-HN peptide, serving as a specificity control (Fig. 1B).
We then compared BimEL with the two other Bim isoforms, BimL and BimS, generated by alternative splicing (Fig. 1C). For these experiments, we coexpressed the three isoforms of Bim, BimEL, BimL, and BimS, as EE-tagged proteins along with GFP-HN in HEK293T cells by cotransfection. With these cell lysates, we preformed immunoprecipitations using monoclonal anti-EE antibody and analyzed the resulting immune complexes using anti-GFP antibody to detect associated GFP-HN (Fig. 1D). Although BimEL, BimL, and BimS were all recovered from lysates efficiently by immunoprecipitation, only BimEL immune complexes contained GFP-HN (Fig. 1D).
To support further the hypothesis that HN can associate with BimEL in cells, we performed immunofluorescence microscopy experiments to localize the various EE-tagged isoforms of Bim in COS-7 cells cotransfected with plasmids encoding GFP-HN or GFP control protein. Unlike GFP, which was found diffusely throughout cells and especially prominent in the nucleus, GFP-HN was localized to punctate structures in the cytosol of transfected COS-7 cells (Fig. 2). Similarly, EEtagged BimEL was localized to punctate cytosolic structures, colocalizing with GFP-HN, as determined by merging images. In contrast, EE-BimL and EE-BimS were distributed diffusely throughout the cytosol of transfected COS-7 cells and did not colocalize with GFP-HN. Unlike GFP-HN, the GFP control protein did not colocalize with EE-BimEL in cotransfected cells (not shown). Thus, among the various isoforms of Bim, the BimEL protein selectively colocalizes with HN in cells.
HN Selectively Prevents BimEL-induced Apoptosis-BimEL, BimL, and BimS can each induce apoptosis when overexpressed in cells (11,13). We therefore investigated whether the direct interaction of HN with BimEL can interfere with its ability to trigger apoptosis. For these experiments, HEK293T cells were transiently transfected with plasmids encoding BimEL in combination with GFP or GFP-HN. After 20 -24 h, caspase activity was measured in cell lysates as a surrogate indicator of apoptosis. BimEL induced strong caspase activation (ϳ5-fold above background), whereas neither GFP nor GFP-HN alone triggered caspase activation (Fig. 3A). When cotransfected with BimEL, GFP-HN but not GFP reduced BimEL-induced caspase activation significantly.
We also compared the effects of GFP-HN on caspase activation induced by BimEL, BimL, and BimS to determine whether binding to GFP-HN correlates with suppression of apoptosis. As shown in Fig. 3B, all three Bim isoforms induced similar -fold increases in caspase activity in transfected HEK293T cells. However, GFP-HN only suppressed caspase activation induced by BimEL but not by BimL or BimS (Fig. 3B). We conclude therefore that the ability of HN to suppress Bim-induced caspase activation correlates with binding to Bim protein, such that only the isoform of Bim which binds HN is suppressed.
To confirm the specificity of these results, we performed complementary experiments in which a mutant version of HN lacking cytoprotective activity was tested for binding to BimEL. In this regard, substitution of proline for cysteine at position 8 of the HN peptide has been shown to abrogate its ability to rescue neuronal cells from apoptosis (24,25). Moreover, we showed previously that the C8P mutant HN fails to bind Bax and is unable to protect cells against Bax-induced cell death (28). We therefore compared the ability of wild-type HN and the C8P mutant of HN to protect cells against apoptosis induced by BimEL overexpression, correlating the results with protein binding data. Overexpression of BimEL induced caspase activation (Fig. 4A) and apoptosis (Fig. 4B), which were suppressed in a concentration-dependent manner by cotransfection of plasmid encoding GFP-HN but not GFP-HN(C8P) mutant. As a positive control, cotransfections were also performed with Bcl-X L -encoding plasmid, demonstrating suppression of BimEL-induced caspase activation and apoptosis. Immunoblot analysis confirmed production of the GFP-HN and GFP-HN(C8P) proteins at comparable levels, excluding trivial explanations for the differential cytoprotective activity of these proteins (Fig. 4A, bottom panel). The protein blot analysis also showed that GFP-HN does not prevent production of BimEL, despite suppressing BimEL-induced caspase activation and apoptosis. Finally, in coimmunoprecipitation assays using anti-Bim antibodies, wild-type GFP-HN was recovered in association with BimEL-containing immune complexes much more efficiently than GFP-HN(C8P) (Fig. 4C), despite equivalent production of these proteins. Thus, binding of HN to BimEL correlates with its ability to suppress apoptosis induced by BimEL overexpression, similar to previous observations concerning HN interactions with Bax (28).
HN Peptide Blocks BimEL-induced Apoptosis from within Cells-Two alternative mechanisms have been proposed for explaining the cytoprotective action of BimEL. One mechanism envisions HN acting as a secreted peptide that protects by binding a cell surface receptor and delivering survival signals (24,25,33). The alternative mechanism envisions an intracel- lular role of HN as an antagonist of Bax/Bcl-2 family proteins (28). To distinguish between these two mechanisms, we compared the effects of HN peptides synthesized with or without a membrane-penetrating polyarginine sequence (Fig. 5A). Immunolocalization studies confirmed differential update of these peptides by cells (34). These peptides were added in vast molar excess relative to Bcl-2/Bax family target proteins to cultures of HEK293T cells that had been transfected with apoptosis-inducing plasmids encoding either BimEL or Bak. Bak was employed here as a control, based on previous observations that HN neither binds Bak nor suppresses Bak-induced apoptosis (28). Addition to cultures of wild-type HN peptide lacking a membrane-penetrating polyarginine sequence failed to suppress caspase activation induced by either BimEL or Bak (Fig. 5B). In contrast, HN polyarginine at 20 M suppressed caspase activation induced by overexpression of BimEL but not Bak. Addition of these peptides to cultures of control-transfected HEK293T cells had no effect on basal levels of caspase activity. These results support the hypothesis that HN peptide must penetrate cells for its cytoprotective activity, consistent with a mechanism involving antagonism of Bax/Bcl-2 family protein such as BimEL.
HN Peptide Prevents BimEL-induced SMAC and Cytochrome c Release from Isolated Mitochondria-We showed previously that HN peptide can block the release of apoptogenic proteins from isolated mitochondria induced by recombinant Bax protein (28). BimEL has also been reported to induce release of apoptogenic proteins when added directly to isolated mitochondria (30,35). We therefore investigated the effects of HN peptide on BimEL-induced release of SMAC and cytochrome c using isolated mitochondria. First, we tested various concentrations of recombinant His 6 -BimEL to evaluate its ability to induce SMAC release from isolated mitochondria purified from HeLa cells (Fig. 6A). At 1 h after BimEL release, mitochondria were recovered by centrifugation, and the relative levels of SMAC remaining in the mitochondria were compared with SMAC released into the supernatant by immunoblotting, using Hsp60 antibody as control to ensure that each sampled contained the same amount of mitochondria. These experiments showed that His 6 -BimEL provoked SMAC release at 300 -600 ng of recombinant protein/50 g of mitochondria (Fig. 6A). In contrast to BimEL, control protein His 6 -SMAC did not induce endogenous SMAC release from isolated mitochondria (not shown), thus confirming the specificity of these results.
Using 300 ng of His 6 -BimEL, we performed similar assays where His 6 -BimEL was preincubated for 15 min with or without various concentrations of synthetic HN peptide (1-50 M) before addition to mitochondria (Fig. 6B). HN peptide at Ն25 M abolished SMAC and cytochrome c release from isolated mitochondria.
To confirm the specificity of these results obtained with wild-type HN peptide, we compared the effects of mutant HN(C8P) peptide, using isolated mitochondria. In contrast to wild-type HN peptide, mutant HN(C8P) peptide did not block BimEL-induced SMAC release even at 50 M (Fig. 7A). In addition, we tested the ability of HN peptide to block mitochondrial release of SMAC induced by an alternative stimulus predicted to be resistant to HN. Namely, we used a BH3 peptide of Bak, which is known to induce cytochrome c release from isolated mitochondria (35), presumably by displacing antiapoptotic Bcl-2 family proteins from mitochondria-resident proapoptotic proteins. In contrast to BimEL, SMAC release induced by Bak BH3 peptide was not blocked significantly by HN peptide (Fig. 7B). Taken together, these results confirm the specificity of HN-based suppression of BimEL-induced release of apoptogenic proteins from isolated mitochondria. Note that the synthetic HN peptide used here is the same as the peptide used for cell culture experiments above, thus confirming that this non-Arg 8 -tagged peptide is an active inhibitor of BimEL in vitro even though it cannot penetrate cell membranes.
HN Peptide Prevents BimEL-induced Bax Activation-We performed additional experiments to explore the mechanism by which HN interferes with the proapoptotic action of BimEL. BimEL is one of the few BOPs that can operate as either an antagonist of antiapoptotic or as an agonist of proapoptotic Bcl-2/Bax family proteins (36). With respect to antagonism of antiapoptotic Bcl-2/Bax family proteins, we examined the effects of HN on the association of BimEL with antiapoptotic proteins Bcl-2 and Mcl-1 by coimmunoprecipitation experiments, but we observed no effect (data not shown). Thus, HN is unlikely to interfere with the ability of BimEL to antagonize antiapoptotic members of the Bcl-2 family.
We therefore turned our attention to an analysis of the effects of HN on proapoptotic protein Bax, asking whether HN interferes with the ability of BimEL to induce activation of Bax, as defined by conversion of this protein from the inactive to active conformation. Using a monoclonal antibody specific for the active conformation of Bax, we performed immunoprecipitation experiments where HeLa cell-derived mitochondria were treated with BimEL in the presence or absence of HN peptide. (Note that it has been shown previously that Bax is preassociated with mitochondria in HeLa cells, unlike some other types of cells where Bax resides in the cytosol prior to translocation to mitochondria (37).) As expected, BimEL induced conversion of Bax to the active conformation, as shown by increased immunoprecipitation with the epitope-specific antibody (Fig. 8). In contrast, preincubating BimEL with HN peptide blocked conversion of Bax to the active conformation, although also preventing BimEL-induced release of SMAC from mitochondria. Immunoblot analysis using a confirmation-independent anti-Bax antibody confirmed the presence of equivalent total amounts of Bax in all samples, thus confirming the validity of these results (Fig. 8). Thus, HN interferes with the ability of BimEL to activate proapoptotic protein Bax.

HN Peptide Can Prevent BimEL-induced SMAC and Cytochrome c Release
Independently of Bax-Because HN peptide binds both BimEL (this report) and Bax (28), the interference of BimEL-induced activation of Bax could be the result of a direct effect on either BimEL or Bax. To distinguish between these two possibilities, we performed experiments using mitochondria isolated from baxϪ/Ϫ cells (9,38,39). BimEL induced release of SMAC and cytochrome c from both baxϩ/ϩ (left panel) and baxϪ/Ϫ (right panel) mitochondria (Fig. 9), consistent with the known ability of BimEL to bind and activate the mitochondria-resident protein Bak independently of Bax (39). The addition of HN peptide inhibited BimEL-induced release of SMAC and cytochrome c from both baxϩ/ϩ and baxϪ/Ϫ mitochondria. We conclude therefore that HN can inhibit BimEL independently of Bax, consistent with a direct effect of HN on the BimEL protein.
HN Peptide Inhibits BimEL-induced Oligomerization of Bak-Release of apoptogenic proteins from mitochondria is associated with oligomerization of Bax and Bak, which can be detected by use of chemical cross-linkers (40). Bim proteins are among the BOPs capable of inducing Bax and Bak oligomerization in mitochondrial membranes (35,39). We therefore assessed the effects of HN peptide on oligomerization of Bak, using baxϪ/Ϫ mitochondria to simplify interpretation of the data. For these experiments, BimEL was incubated with or without HN peptide, then applied to mitochondria, followed by the addition of chemical cross-linker BMH, and analysis of the resulting samples by SDS-PAGE/immunoblotting using anti-Bak antibody. In untreated mitochondria, Bak was present mostly as a monomer of ϳ22 kDa (Fig. 10A). Treatment of mitochondria with BimEL induced multimerization of Bak, as revealed by the presence of slower migrating forms of Bak in cross-linked samples analyzed by gel electrophoresis. Pretreatment of BimEL with HN peptide resulted in concentrationdependent inhibition of Bak oligomerization (Fig. 10A). To control for specificity, we repeated these Bak oligomerization assays using HN(C8P) control peptide. Unlike wild-type HN, which inhibited Bax oligomerization at concentrations as low as 3-10 M, the HN(C8P) peptide was inactive at concentrations as high as 50 M (Fig. 10B).
Structure-Function Analysis of HN Peptide Interaction with BimEL-Our previous data showed that the first 1-17 amino acids of HN are required for binding to Bax (28). To begin to delineate some of the structure-function relations of the HN peptide with respect to BimEL binding and suppression, we compared the effects of fragments of the HN peptide representing either the first or the last 12 amino acids of this peptide (e.g. 1-12 and 13-24).
To determine whether either the N-or C-terminal half of HN can bind BimEL, we expressed these segments of HN as GFP fusion proteins in HEK293T cells along with EE-tagged BimEL and performed coimmunoprecipitation experiments, making comparisons with full-length HN protein. Immunoprecipitation of BimEL using anti-Bim antibody revealed the presence of associated GFP-HN and GFP-HN(1-12) but not GFP-HN (13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24). Immunoblot analysis of the lysates confirmed production of the GFP -HN(1-12) and GFP-HN(13-24) proteins at comparable levels, thus excluding differences in the levels of these FIG. 4. HN C8P mutant fails to bind BimEL or protect against BimEL-induced cell death. A, HEK293T cells were transfected with 1-3 g of pEGFP-C1, pEGFP-HN, pEGFP-HN(C8P), or myc-Bcl-X L together with 0.3 g of plasmid DNA encoding BimEL. After 18 h, cells were lysed, and caspase activity was assessed using DEVD-AFC substrate. For assessing protein expression (lower panels), lysates were prepared from transfected cells, normalized for total protein content, and analyzed by SDS-PAGE/immunoblotting using Bim, Hsp60, and GFP antibodies. Hsp60 antibody served as a loading control. B, HEK293T cells were transfected as described above. After 18 h, cells were trypsinized, washed with PBS, fixed with 3.7% formaldehyde, and stained with DAPI to visualize nuclei by ultraviolet microscopy. The percentages of apoptotic cells were determined by counting 200 GFP-positive cells, scoring cells having nuclear fragmentation and/or chromatin condensation. Data are reported as the mean Ϯ S.E. (n ϭ 3). C, HEK293T cells were transfected with pEGFP-C1, pEGFP-HN, or pEGFP-HN(C8P) in combination with pEF-PGK-BimEL. Cells lysates were prepared, and immunoprecipitations were performed with monoclonal anti-EE mAb. The immunoprecipitates (IP) or the lysates were blotted with anti-GFP or Bim-antibodies. fusion proteins as a trivial explanation for the differential binding to BimEL (Fig. 11A).
Next, we tested the effects of synthetic HN(1-12) and HN(13-24) peptides on BimEL-induced release of SMAC from isolated mitochondria (Fig. 11B). Although HN(1-12) peptide suppressed in a concentration-dependent manner the amount of SMAC released from mitochondria after treatment with recombinant BimEL protein, HN(13-24) did not.
Finally, the ability of GFP-HN(1-12) to suppress apoptosis induced by expression of BimEL was tested, using caspase activity as a surrogate for apoptosis. In transfected HEK293T cells, BimEL induced an increase in caspase activity, which  6. HN blocks BimEL-induced cytochrome c and SMAC release from isolated mitochondria. A, various amounts of recombinant His 6 -BimEL protein (10 -600 ng) were preincubated with 50 g of isolated mitochondria purified from HeLa cells. B, 300 ng of His 6 -BimEL protein were preincubated with or without various concentrations of HN peptide for 10 min, as indicated. Then, 50 g of isolated mitochondria purified from HeLa cells was added. A and B, after incubation at 30°C for 1 h, mitochondria were pelleted by centrifugation. The mitochondria-containing pellets and the supernatants were analyzed by SDS-PAGE/ immunoblotting using Hsp60, cytochrome c, or SMAC antibodies.

FIG. 7. Comparison of effects of wild-type and C8P mutant HN on mitochondrial SMAC release induced by BimEL or Bak BH3 peptide.
A, 300 ng of His 6 -BimEL protein was preincubated with various concentrations of HN or HN(C8P) peptide for 10 min, then 50 g of isolated mitochondria was added. After incubation at 30°C for 1 h, the pellets containing mitochondria and the supernatants were analyzed by SDS-PAGE/immunoblotting using anti-SMAC and anti-Hsp60 antibodies. B, 300 ng of His 6 -BimEL or various concentrations of Bak BH3 peptide were preincubated with 30 M HN peptide for 10 min, then 50 g of isolated mitochondria was added. Mitochondria were processes and analyzed as above. was suppressed significantly by coexpression of GFP-HN or GFP-HN(1-12) but not GFP-HN(13-24) (Fig. 11C).
We conclude therefore that the first 12 amino acids of the HN peptide are necessary and sufficient to bind BimEL and to suppress BimEL activity in vitro and in cells. DISCUSSION Previously, we reported a new mechanism of action for the endogenous cytoprotective peptide HN, showing that it binds and inhibits proapoptotic protein Bax (28). In this study, we extended the protective action of HN by identifying BimEL as a new target of HN regulation. HN binds BimEL and inhibits its proapoptotic activity, as demonstrated in intact cells and in experiments using isolated mitochondria where HN blocks BimEL-induced oligomerization of Bak and release of apoptogenic proteins cytochrome c and SMAC. Thus, BimEL can be inhibited directly by HN, preventing it from activating Bax and Bak and thereby negating the agonist activity of BimEL on these proapoptotic members of the Bcl-2 family.
Previously we reported that HN binds and suppresses Bax, and in an accompanying paper, we also show that HN binds to and inhibits Bid (34). The Bax, Bid, and BimEL proteins have in common that they are normally present in the cytosol in an inactivate state, subsequently translocating to mitochondria where they interact with other Bcl-2/Bax family members and induce release of apoptogenic proteins. Thus, it is intriguing that HN peptide can target all three of these proapoptotic proteins, perhaps stabilizing their inactive conformations. In addition, Bid and BimEL are the only BOPs (among the Ͼ15 identified in mammals) that function as agonists of Bax and Bak. Presumably Bax, Bid, and BimEL share structural char-acteristics that allow these proteins to bind HN. However, because the three-dimensional structure of BimEL is not yet known, it is not possible to speculate where HN might bind. This notion of conserved structural features is also supported by (a) the observation that the first 12 amino acids of the HN peptide are sufficient to bind and suppress both BimEL (this report) and Bid (accompanying report (34)) and by (b) experimental findings using FPAs that show a similar affinity of full-length HN for Bax, Bid, and BimEL. Downstream residues in the HN peptide however probably also either directly or indirectly make contributions to binding, perhaps by influencing peptide conformation, because our previous data suggested that amino acids 1-17 in HN are required for binding Bax (28).
Expression studies suggest that BimEL is the predominant isoform of Bim produced in most tissues, with far less of the BimL and BimS proteins present (12). However, to the extent that cells produce abundant amounts of the shorter BimL and BimS proteins through alternative mRNA splicing, then HN would be unable to rescue from cell death induced under circumstances where BimEL no longer remains the dominant isoform. Interestingly, both BimEL and BimL are reportedly sequestered on microtubules in a complex with the dynein motor complex (23). It will be interesting therefore to determine the effects of HN on the stability of the BimEL-dynein complex and whether HN acts at the level of these protein complexes or post-release of BimEL from the complex. Our in vitro experiments demonstrate clearly that HN is capable of suppressing the activity of isolated BimEL, and thus we speculate a role for this cytoprotective peptide downstream of the FIG. 8. HN blocks BimEL-induced activation of Bax. 300 ng of His 6 -BimEL was preincubated with or without 50 M HN peptide for 10 min, then 50 g of isolated mitochondria was added. After incubation at 30°C for 1 h, samples were centrifuged, and the resulting mitochondria-containing pellets and supernatants were analyzed by SDS-PAGE/ immunoblotting using SMAC, Bax, and Hsp60 polyclonal antibodies. To examine Bax conformational change, mitochondrial pellets were lysed in 1% CHAPS and subjected to immunoprecipitation with anti-Bax 6A7 monoclonal antibody. Immunoprecipitates (IP) were analyzed by SDS-PAGE/immunoblotting using polyclonal anti-Bax antibody (bottom panel). dynein motor complex, after release of BimEL.
It has been suggested that the HN peptide can operate extracellularly via cell surface receptors to convey cytoprotective signals into cells. For example, the Hashimoto group (24,29) reported that to exert neuroprotective activity against Alzheimer disease-relevant insults, HN must be secreted from cells, suggesting that a cell surface receptor is the target of HN action. In particular, it was shown that adding 10 M HN peptide into medium can rescue neurons from cell death induced by amyloid ␤-peptide (33). Consistent with this notion, it was reported recently that HN can interact with insulin-like growth factor-binding protein 3 (41), providing further evidence that HN may act outside cells. Although we cannot exclude the possibility that HN has a dual function as both an intracellular and extracellular regulator of cell survival, our findings thus far provide support for an intracellular role of HN peptide as an antagonist of selective Bcl-2/ Bax family proapoptotic proteins. The evidence supporting this intracellular role for HN comes from this study, the accompanying report on Bid (34), and our prior work on Bax (28), and includes (a) perfect correlation of binding with cytoprotective function, where HN suppresses apoptosis induced only by members and isoforms of the Bcl-2/Bax family that it binds and where inactive mutants of HN fail to bind Bcl-2/Bax family proteins; (b) failure of synthetic HN peptide to suppress apoptosis unless equipped with a membrane-penetrating sequence that allows access to intracellular compartments; (c) retained activity of HN when expressed intracellularly as a GFP fusion protein; and (d) ability of HN peptide to act on isolated mitochondria, preventing release of apoptogenic proteins in the absence of cell membrane receptors.
The structure of the HN peptide has not yet been determined. The cysteine residue at position 8 of this 24-amino acid peptide appears to be important for its cytoprotective activity, inasmuch as substitution of a proline at this site abolishes antiapoptotic activity (25) and negates binding to Bax and BimEL. Proline substitution conceivably could distort the overall structure of the HN peptide. Alternatively, the unique cysteine residue in HN could play a specific role in HN binding to cellular targets, but a survey of other types of amino acid substitution (e.g. alanine, serine, glycine) must be performed before conclusions can be drawn. In this regard, it has been proposed that the cysteine of HN mediates dimerization of this peptide (29). However, the reducing environment of the cytosol is not conducive to disulfide bond formation, and thus this notion is unlike to apply to the role of HN as an intracellular antagonist of BimEL, Bid, and Bax. If HN is secreted from cells and also possesses extracellular functions, then disulfide bond formation mediated by the unique cysteine in HN could be envisioned. Determination of the three-dimensional structure of the HN peptide and extensive structure-based mutagenesis FIG. 11. HN(1-12) peptide is sufficient to bind and block the proapoptotic function of BimEL. A, HEK293T cells were transfected with pEGFP-C1, pEGFP-HN(1-24), pEGFP-HN(1-12), or pEGFP-HN (13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24) in combination with pEF-PGK-BimEL. Cells lysates were prepared, and immunoprecipitations were performed with polyclonal anti-Bim mAb. The immunoprecipitates (IP) or the lysates were blotted with anti-GFP or Bimantibodies. B, 300 ng of His 6 -BimEL protein was preincubated with various concentrations of HN(1-24), HN(1-12), or HN(13-24) peptide for 10 min, then 50 g of isolated mitochondria was added. After incubation at 30°C for 1 h, the pellets containing mitochondria and the supernatants were analyzed by SDS-PAGE/immunoblotting using anti-SMAC and anti-Hsp60 antibodies. C, HEK293T cells were transfected with 1 g of pEGFP-C1, pEGFP-HN(1-24), pEGFP-HN(1-12), pEGFP-HN (13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24), or myc-Bcl-X L together with 0.03-0.3 g of plasmid DNA encoding BimEL. After 18 h, cells were lysed, and caspase activity was assessed using DEVD-AFC substrate (mean Ϯ S.D.; n ϭ 3).
analysis are needed to provide additional insights into the structure-function activities of this interesting peptide.
Although the in vivo physiological functions of the endogenous HN peptide remain to be defined, the discovery that HN can bind and inhibit Bax, Bid, and BimEL raises the possibility of exploiting this peptide or chemical compounds that mimic it for the purpose of cell preservation during disease. Thus, to the extent that BimEL, Bid, and Bax possess conserved structural characteristics explaining their ability to bind the N-terminal segment of the HN peptide, possibilities may exist for generating cytoprotective peptides or compounds that simultaneously neutralize all three of these proapoptotic proteins. In this regard, several studies have implicated Bax, Bid, and Bim in neuronal cell death (17,26,(42)(43)(44)(45)(46)(47)(48)(49). Thus, it may be particularly attractive to consider HN-based therapies for prevention of neuronal cell death during conditions such as stroke, axotomy, and neurodegeneration. Further preclinical studies of HN-mediated cytoprotection thus are warranted to ascertain the spectrum of cytoprotective activities and the detailed mechanism of this naturally occurring peptide.