Humanin Binds and Nullifies Bid Activity by Blocking Its Activation of Bax and Bak*

Recently, we discovered that Humanin (HN), a small endogenous peptide of 24 amino acids, binds to and inhibits the proapoptotic protein Bax. We show here that HN also interacts with the BH3-only Bcl-2/Bax family protein, Bid, as well as a truncated form of Bid (tBid) associated with protease-mediated activation of this proapoptotic protein. Synthetic HN peptide binds purified Bid and tBid in vitro and blocks tBid-induced release of cytochrome c and SMAC from isolated mitochondria, whereas mutant peptides that fail to bind Bid or tBid lack this activity. Moreover, HN peptide also retained protective activity on bax–/–mitochondria, indicating that HN can block tBid-induced release of apoptogenic proteins from these organelles in a Bax-independent manner. HN peptide inhibits tBid-induced oligomerization of Bax and Bak in mitochondrial membranes, as shown by experiments with chemical cross-linkers or gel filtration. Gene transfection experiments showed that HN (but not an inactive mutant of HN) also protects intact cells from apoptosis induced by overexpression of tBid. We conclude that Bid represents an additional cellular target of HN, and we propose that HN-mediated suppression of Bid contributes to the antiapoptotic activity of this endogenous peptide.

Mitochondria constitute one of the critical centers of life and death control within mammalian cells. Mitochondria undergo profound changes in membrane integrity during the apoptosis process, leading to the release of apoptogenic proteins, including cytochrome c, AIF, SMAC/Diablo, endonuclease G, Bit, and Omi/HtrA2 (1-7). Bcl-2 family proteins are central regulators of mitochondria-dependent cell death pathway, functioning as either anti-or proapoptotic factors that govern release of apoptosis-inducing proteins from these organelles (8 -11).
Sequence homology within Bcl-2 family proteins clusters into regions known as Bcl-2 homology (BH) 1 domains, with various family members containing as many as four conserved BH domains, BH1, BH2, BH3, and BH4 (8 -10). The six known human antiapoptotic members of the Bcl-2 family possess all four BH domains, and most also contain a C-terminal hydrophobic span of amino acids responsible for their insertion into intracellular membranes. In contrast, proapoptotic Bcl-2 family proteins are more diverse in their sequences, typically containing only one to three of the four conserved BH domains (2). In this regard, two principal classes of proapoptotic Bcl-2 family proteins have been delineated, known as "multidomain" proteins versus "BH3-only" proteins (BOPs). The multidomain proapoptotic proteins contain BH1, BH2, and BH3 domains and include Bax, Bak, and Bok, in mammals. Structural studies suggest that the multidomain proteins have essentially the same protein fold as the antiapoptotic Bcl-2 family proteins (12)(13)(14).
BOPs utilize their BH3 domains to interact with either antiapoptotic or proapoptotic Bcl-2 family proteins. Although the BH3 domains of all BOPs identified thus far are capable of binding and inhibiting antiapoptotic Bcl-2 family proteins, a few BOPs can also bind and activate proapoptotic multidomain proteins Bax and Bak. The activation of Bax and Bak is associated with their oligomerization in mitochondrial membranes, an event that results in release of apoptogenic proteins from mitochondria (15). Furthermore, gene ablation studies in mice indicate that either bax or bak is required for apoptosis induction by BOPs, such that double knock-out mice lacking both of these genes are insensitive to BOP-induced apoptosis (16 -18).
Bid is the first of the BOPs reported to bind and activate Bax and Bak directly, functioning as an agonist of these multidomain proteins. Under the normal physiological conditions, Bid is present in the cytosol of cells in an inactive state. However, upon proteolytic cleavage that removes the N-terminal ϳ50 amino acids, Bid becomes active, apparently exposing its BH3 domain, promoting insertion into mitochondrial membranes, and triggering interactions with other Bcl-2/Bax family proteins (15). The proteases known to cleave and activate Bid include caspase-8, which is activated following the engagement of death receptors Fas or tumor necrosis factor receptor 1 (15,19,20), as well as calpains and lysosomal proteases (21)(22)(23). Although sequence alignments reveal the presence only of the BH3 domain in Bid, the three-dimensional structure of Bid is highly similar to Bax, Bcl-2, and Bcl-X L (12)(13)(14).
Humanin (HN) was originally identified as an antiapoptotic peptide encoded in a cDNA which rescued neuronal cells from apoptosis induced by presenilin mutants associated with familial Alzheimer disease, during a functional screen of a cDNA library prepared from the brain of a patient with autopsy-confirmed Alzheimer disease (24,25). HN was found to protect neuronal cells from a variety of toxic insults (26 -28). Some reports suggest that HN is secreted from cells and binds receptors in the plasma membrane (29). However, our laboratory showed previously that HN is present within the cytosol of cells where it can bind Bax and prevent the translocation of Bax from cytosol to mitochondria, thus providing protection from death stimuli that depend on this multidomain member of the Bcl-2/Bax family (30). In this report, we further demonstrate that HN can bind Bid as well as N-terminally truncated Bid (tBid), nullifying the proapoptotic activity of tBid. The findings thus broaden the range of HN targets within the Bcl-2/Bax family and may provide the basis for eventually exploiting this knowledge toward the goal of devising novel cytoprotective strategies.
Fluorescence Polarization Assays (FPAs)-To determine the binding affinity of HN peptides to Bid and tBid proteins, FPAs were performed according to procedures published previously (35). Briefly, various concentrations of Bid, tBid (cleaved by caspase-8), or Bcl-X L ⌬TM (used as control) proteins were incubated with 20 nM FITC-conjugated synthetic purified HN peptide dissolved in water in the dark. Fluorescence polarization was measured using an Analyst TM AD Assay Detection System (LJL Biosystem, Sunnyvale, CA) in phosphate-buffered saline (PBS), pH 7.4. IC 50 determinations were performed using GraphPad Prism software (GraphPad, Inc., San Diego).
Cell Culture, Transfection, and Apoptosis Assays-HEK293T, HeLa, and MEF cells were maintained in Dulbecco's modified Eagle's medium (Irvine Scientific) supplemented with 10% fetal bovine serum, 1 mM L-glutamine, and antibiotics. For transient transfection apoptosis assays, cells (5 ϫ 10 5 ) in 6-well plates were cotransfected using Lipofectamine 2000 (Invitrogen) with 0.5 g each of pEGFP-C1 (Clontech) or pcDNA3-myc-tBid, together with various amounts of GFP-HN or pcDNA3-myc-Bcl-X L plasmids. For caspase assays, cell lysates were prepared after 20 h, normalized for protein content, and 10-g aliquots of cell lysates were incubated with 100 M DEVD-AFC, measuring enzyme activity by the release of AFC fluorescence. Data are reported as relative fluorescence units (RFU) of product produced/min/g of total protein.
Apoptosis was assessed by DAPI staining of cells. Briefly, at 20 h after transfection, both floating and adherent cells were collected, fixed with PBS containing 3.7% formaldehyde, and stained with 0.1 mg/ml DAPI in PBS. The percentages of apoptotic cells were determined by UV-microscopy, counting GFP-positive cells having nuclear fragmentation and/or chromatin condensation. All assays were performed in triplicate.
Protein Interaction Assays-Synthetic HN peptide (10 g) was incubated at 4°C for 4 h in 200 l of PBS together with recombinant His 6 -Bid, His 6 -tBid, or control His 6 -Bcl-X L ⌬TM (10 g each) and 10 l of Ni-NTA resin. Then the resins were washed three times with 1 ml of PBS, and bound proteins were analyzed by Tris-Tricine gel electrophoresis, following by Coomassie Blue staining.
Either recombinant tBid produced by cleavage with caspase-8 (100 ng) or His 6 -Bcl-X L ⌬TM was bound to Ni-NTA resin in PBS buffer containing 1% CHAPS and 2% ␥-globulins. HN peptide was preincubated with the resin for 15 min, then 100 ng of recombinant Bax was added. After an additional 4-h incubation, the resin was washed three times and then eluted with 0.5 M imidazole. The supernatants were analyzed by SDS-PAGE/immunoblotting using tBid or Bax antibody (Santa Cruz).
Coimmunoprecipitation and Immunoblotting Procedures-Transfected HEK293T cells (5 ϫ 10 5 ) were cultured with 50 M benzoyl-Val-Ala-Asp-fluoromethylketone (ZVAD-fmk; Bachem) to prevent cell death. At 24 h post-transfection, cells were collected and resuspended in lysis buffer (10 mM HEPES, pH 7.4, 142.4 mM KCl, 5 mM MgCl 2 , 0.5 mM EGTA, 0.2% Nonidet P-40, 1 mM phenylmethylsulfonyl fluoride, and a protease inhibitor mixture (Roche Applied Science)). After removing debris by centrifugation, lysates were normalized for protein content (50 g in 400 l) and incubated with 10 l of anti-myc (Santa Cruz) antibody-conjugated Sepharose beads overnight at 4°C. Beads were then washed four times in 1 ml of lysis buffer and boiled in an equal volume of Laemmli gel loading solution before analysis of proteins by SDS-PAGE/immunoblotting using polyclonal rabbit anti-GFP (Roche Applied Science) or monoclonal mouse anti-myc (Santa Cruz) antibodies. Alternatively, lysates were normalized for total protein content and analyzed directly by SDS-PAGE/immunoblotting.
Mitochondria Purification and Protein Release Assays-Cells were pelleted by centrifugation and then washed once in HM buffer (10 mM HEPES, pH 7.4, 250 mM mannitol, 10 mM KCl, 5 mM MgCl 2 , 1 mM EGTA), containing 1 mM phenylmethylsulfonyl fluoride and a mixture of protease inhibitors (Roche Applied Science). The cell pellet was then homogenized in HM buffer by 50 strokes of a Dounce homogenizer, using a B-type pestle. The homogenate was centrifuged twice at 600 ϫ g for 5 min to remove nuclei and debris. The resulting supernatant was centrifuged at 10,000 ϫ g for 10 min, and the resulting mitochondriacontaining pellet was washed twice with the HM buffer.
For mitochondrial protein release assays, 10 l of mitochondria (50 g) was added into a final volume of 50 l of HM buffer containing tBid or tBid preincubated with HN peptide at 30°C for 15 min. The reaction was incubated further at 30°C for 40 -60 min, then mitochondria were pelleted by centrifugation and resuspended in a volume of HM buffer equal to the volume of supernatant collected. The samples were then diluted 1/4 with 5ϫ Laemmli solution, boiled, and analyzed by SDS-PAGE/immunoblotting using cytochrome c (75981A Pharmingen) or SMAC antibodies (3).
Chemical Cross-linking-Oligomerization of Bak was assessed by chemical cross-linking. In brief, dimethyl sulfoxide or bismaleimidohexane cross-linker (Pierce) dissolved in dimethyl sulfoxide was added to mitochondria at a final concentration of 5 mM for 30 min at 30°C. The reactions were quenched with 50 mM cysteine for 15 min at room temperature. Mitochondria were then analyzed by SDS-PAGE/immunoblotting using anti-Bak antibody (Upstate).
Gel Filtration Chromatography-A Superdex 75 HR 10/30 column (Amersham Biosciences) was equilibrated with buffer containing 1% CHAPS, 20 mM HEPES, pH 7.4, and 150 mM NaCl. Then, 100 l of mitochondrial lysates dissolved in 2% CHAPS (200 g of total protein) was loaded onto the column, and 0.5-ml fractions were collected and analyzed by SDS-PAGE followed by immunoblotting using antibodies recognizing Bax (N20, Santa Cruz).
Coimmunoprecipitation Assays-HEK293T cells in 100-mm plates were cotransfected with 2 g each of plasmids encoding myc-tBid or myc-Bcl-X L with GFP-Bak in duplicate, and the cells were cultured with 50 M ZVAD-fmk to prevent cell death. Cell lysates were subjected to immunoprecipitation using anti-myc beads overnight at 4°C. Immune complexes were split into five equal aliquots, and various concentrations of HN peptide were added. After 6 h of incubation, the beads were washed three times with lysis buffer (10 mM HEPES, pH 7.4, 142.4 mM KCl, 5 mM MgCl 2 , 0.5 mM EGTA, 0.2% Nonidet P-40) containing 1 mM phenylmethylsulfonyl fluoride and a protease inhibitor mixture (Roche Applied Science)) (500 l each). Bound proteins were analyzed by SDS-PAGE/immunoblotting using tBid (36) or GFP antibodies (Roche Applied Science).
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, 1% triisopropylsilane for 2 h at room temperature; polyarginine peptides were treated for 6 h. The crude peptides were purified with a Gilson high performance liquid chromatography instrument and analyzed by matrix-assisted laser desorption ionization time-of-flight mass analysis with an Applied Biosystems Voyager System 6264.

RESULTS
HN Binds to Bid-We showed previously that HN can bind Bax but not Bak and not antiapoptotic Bcl-2 family proteins such as Bcl-2, Bcl-X L , Mcl-1, and Bcl-B (30). While testing other Bcl-2/Bax family proteins for interactions with HN, we observed that HN binds the proapoptotic BOP, Bid. Three different methods demonstrated binding of HN to Bid, including in vitro protein interaction experiments, immunoprecipitation from cell lysates, and FPAs ( Fig. 1). Interestingly, HN bound to both full-length and caspase-8-cleaved Bid. For example, we immobilized recombinant His 6 -Bid protein on nickel resin and then digested a portion of the immobilized protein with caspase-8. Synthetic HN was then incubated with Bid-coated beads, and the samples were analyzed by SDS-PAGE with Coomassie staining. By these pull-down assays, we observed that synthetic HN peptide bound equally well to either cleaved or uncleaved Bid (Fig. 1A). Note that after caspase cleavage, the N-and C-terminal portions of cleaved Bid remain physically associated, consistent with prior reports (13,37). In contrast, His 6 -Bcl-X L ⌬TM did not pull down HN peptide.
To explore further the binding of HN to Bid and cleaved Bid, we performed coimmunoprecipitation experiments where HN was expressed in HEK293T cells as a GFP fusion protein, together with myc-tagged Bid (Fig. 1B), myc-tagged truncated Bid, representing the last 60 -195 amino acids of the Bid protein corresponding to the C-terminal fragment produced in vivo by caspase cleavage (19,20), or myc-Bcl-X L . HN was expressed with GFP fused to either its N or C terminus and compared with GFP control protein. As shown in Fig. 1, GFP-HN and HN-GFP associated with both Bid and tBid (Fig. 1B), whereas GFP control protein did not. Immunoblot analysis of the transfected cell lysates demonstrated production of equivalent amounts of Bid or tBid in all samples and showed that GFP was produced at levels as high or higher than HN-GFP and GFP-HN, thus excluding differences in protein expression as a trivial explanation for the results. Using the same conditions, GFP-HN and HN-GFP did not coimmunoprecipitate with Bcl-X L , confirming the specificity of these results.
Finally, FPAs provided additional evidence that HN binds directly to Bid and tBid. For these experiments, FITC-conjugated HN peptide was incubated with various concentration of His 6 -Bid or caspase-8-cleaved Bid, and fluorescence polarization of the FITC-HN peptide was measured (Fig. 1D). Both Bid and tBid induced concentration-dependent, saturable fluorescence polarization of the FITC-HN peptide, whereas Bcl-X L did not. Binding of FITC-HN to tBid produced a sigmoidal curve, allowing estimation of the K d for the interaction of ϳ10 -20 nM (Fig. 1C).
HN Prevents tBid-induced Apoptosis-We tested whether HN can suppress caspase activation and apoptosis induced by overexpression of tBid in cells. For these experiments, HEK293T cells were transfected with plasmids encoding tBid in combination with plasmids encoding either GFP-HN or GFP control protein, then either caspase activity was measured in cell lysates or the cells were fixed and the percentage of apoptotic cells was determined by DAPI staining. Expression of tBid induced caspase activation and apoptosis (Fig. 2, A and B). Cotransfecting GFP-HN plasmid resulted in concentrationdependent reductions in caspase activity and apoptosis. Comparisons with a Bcl-X L -encoding plasmid revealed similar potency compared with GFP-HN with respect to suppression of tBid-induced caspase activation ( Fig. 2A). Because immunoblotting experiments confirmed that GFP-HN does not inter- fere with tBid expression (Fig. 1C), we conclude that HN is able to suppress tBid-induced apoptosis.
Synthetic HN Peptides Containing Membrane Penetration Sequences Suppress tBid-induced Apoptosis-Alternative theories have been proposed for explaining the cytoprotective mechanism of HN, involving either an extracellular role for HN via interactions with cell surface receptors (29) versus an intracellular role suppressing Bcl-2/Bax family proteins (30). To distinguish between these two mechanisms, we compared the activity of synthetic HN peptides containing versus lacking membrane-penetrating sequences with respect to their ability to inhibit caspase activation and apoptosis induced by overexpression of tBid. We exploited the ability of polyarginines to confer membrane-penetrating properties upon peptides, comparing four different designs, including polyarginine 8 at the N terminus, polyarginine 8 appended to the C terminus, polyarginine 6 added to the C terminus, and polyarginine 4 added to both the N-and C terminus (Fig. 3A). To reduce polarity of the peptides, the N termini were acetylated, and the C termini were amidated. Based on previous studies of similar peptides, we anticipated that all four of these arginine-modified versions of HN would be membrane-permeable (38 -40). Experiments with polyarginine peptides confirmed their ability to enter cells, as determined by confocal microscopy using indirect immunofluorescence microscopy with anti-HN antibody ( Fig. 3B and not shown). In contrast, the HN peptide lacking appended arginines did not penetrate cell membranes.
These synthetic peptides were added to cultures of HEK293 cells transfected with tBid-encoding plasmid, and then either caspase activity (Fig. 3C) or apoptosis (Fig. 3D) was measured. Although unmodified HN failed to suppress tBid-induced caspase activation and apoptosis at concentrations up to 20 M, all four of the arginine-modified HN peptides suppressed tBidinduced caspase activation and apoptosis in a concentrationdependent manner. We conclude therefore that addition of polyarginines improves the cytoprotective activity of HN peptide when applied to cell cultures.
HN Inhibits tBid-induced Release of Apoptogenic Proteins from Mitochondria-Expression of tBid in cells induces caspase activation by causing release of cytochrome c, which then binds cytosolic Apaf1 and triggers activation of caspase-9. Release of mitochondrial SMAC by tBid also facilitates caspase activation by negating the suppressive activity of IAP family proteins that bind and inhibit caspase-9 and selected downstream members of the caspase family (3,4). Because HN was found to suppress tBid-induced caspase activation and apoptosis, we explored whether HN inhibits release of apoptogenic proteins from isolated mitochondria.
For these experiments, we used recombinant His 6 -Bid cleaved by caspase-8. As expected, recombinant tBid protein (cleaved by caspase-8) induced cytochrome c and SMAC release from isolated mitochondria in a concentration-dependent manner (Fig. 4, A and C). Preincubation of tBid with synthetic HN peptide suppressed the mitochondrial release of cytochrome c and SMAC in a concentration-dependent manner, with a 100fold molar excess of HN completely suppressing tBid-induced release of these apoptogenic proteins (Fig. 4, B and D).
Note that we also used this assay to assess the activity of the HN peptides described above where polyarginines were added. All modified versions of HN peptides (HN1-HN4) remained active at suppression of tBid-induced release of SMAC from isolated mitochondria in vitro (not shown). Thus, addition of polyarginines does not interfere with tBid-suppressing activity of HN peptides.
Because HN is capable of binding both Bax and Bid, the observation that HN inhibits tBid-induced cytochrome c and SMAC release from mitochondria could be caused by interactions of this peptide with exogenous tBid or alternatively because of binding of HN to endogenous Bax protein resident within mitochondria preparations (30). To exclude an indirect effect caused by Bax, we repeated the above experiments using mitochondrial isolated from baxϪ/Ϫ cells (Fig. 5). Comparisons of mitochondria from baxϪ/Ϫ and baxϩ/ϩ MEFs revealed that tBid induced cytochrome c with equal efficiency, consistent with data indicating that either endogenous bax or bak is sufficient for tBid-induced apoptosis (15,18). Furthermore, HN peptide blocked tBid-induced cytochrome c release from both baxϪ/Ϫ and baxϩ/ϩ mitochondria (Fig. 5, B and D). We conclude therefore that HN is capable of directly suppressing the activity of tBid in vitro and does not require Bax for suppressing release of proteins from mitochondria.
HN Peptide Inhibits tBid Interaction with and Oligomerization of Bax and Bak-The BH3 domain of Bid is capable of binding either antiapoptotic (e.g. Bcl-2, Bcl-X L ) or proapoptotic (e.g. Bax; Bak) members of the Bcl-2/Bax family (10,41). We therefore tested the effects of HN peptide on association of Bid with other members of the Bcl-2/Bax family using either in vitro protein binding assays ("pull-downs") employing recombinant proteins or using coimmunoprecipitations employing ly- sates from transfected cells. For in vitro protein binding experiments, tBid was immobilized on nickel resin and incubated with recombinant Bax or Bcl-X L in the presence or absence of various concentrations of synthetic HN peptide in vitro. HN peptide caused a concentration-dependent reduction in the amounts of Bax associated with tBid (Fig. 6A). In contrast, HN did not block binding of Bcl-X L to Bax, confirming the specificity of these results (Fig. 6B).
Similarly, we tested the effects of addition of HN peptide to lysates of HEK293T cells that had been transfected with plasmids encoding myc-tBid and GFP-Bak (Fig. 6C). The myc-tBid protein was immunoprecipitated using myc-coated Sepharose beads, and associated GFP-Bak was identified by immunoblot analysis. These experiments revealed that HN peptide inhibited interaction of tBid with GFP-Bak (Fig. 6C). In contrast, the addition of HN peptide to cell lysates did not interfere with . At 12 h after transfection, either cell lysates were prepared for caspase activity measurements (C), or apoptosis was measured (D). For caspase assays, lysates were normalized for protein content (10 g) and incubated with 100 M Ac-DEVD-AFC. Enzyme activity was determined by the generation of fluorescent AFC product, and V max was calculated, expressing data as RFU/min/g (mean Ϯ S.D.; n ϭ 3). For apoptosis assays, both floating and adherent cells were collected, fixed, and stained with 0.1 g/ml DAPI. The percentages of apoptotic cells were determined by counting the GFP-positive cells having nuclear fragmentation and/or chromatin condensation (mean Ϯ S.D.; n ϭ 3).

FIG. 4. HN peptide blocks tBid-induced release of cytochrome c and SMAC from isolated mitochondria.
A and C, various concentrations of tBid (cleaved by caspase-8) were incubated with HeLa mitochondria (50 g in 50 l) for 1 h at 30°C, then samples were centrifuged to generate pellets and supernatants that were analyzed by SDS-PAGE/immunoblotting using anti-cytochrome c (Cyt c), anti-SMAC, or anti-Hsp60 antibodies. B and D, 100 ng of cleaved tBid was preincubated with various concentrations of HN peptide for 15 min in HM buffer before adding 50 g of HeLa mitochondria. Samples were processed as above. coimmunoprecipitation of Bcl-X L with Bak (Fig. 6D), consistent with the inability of HN to bind Bak (30).
The interaction of tBid with Bax and Bak normally induces oligomerization of these proteins in mitochondrial membranes, an event directly linked to release of apoptogenic proteins from these organelles (34). We evaluated the effects of HN peptide on oligomerization of Bax by gel sieve chromatography. Recombinant cleaved Bid was added to isolated mitochondria, then the membranes were solubilized in 2% CHAPS, and samples were analyzed by gel sieve chromatography (Fig. 7A). Because little endogenous Bax protein was associated with the mitochondria (isolated here from MEFs), we supplemented those preparations with recombinant full-length Bax protein (Fig.  7A). Prior to addition of tBid, Bax eluted from gel sieve columns predominantly as monomers. After the addition of cleaved Bid, however, higher molecular mass complexes of Bax were observed, consistent with oligomerization of these proteins. Preincubating cleaved Bid with 20 M HN peptide completely abrogated tBid-induced oligomerization Bax oligomerization.
The oligomerization of Bak can be detected using chemical cross-linkers (15). We therefore tested the effects of HN peptide on tBid-induced oligomerization of Bak, using isolated mitochondria in conjunction with the nonreversible cross-linker bismaleimidohexane. For Bak oligomerization, we used mitochondria isolated from baxϪ/Ϫ MEF cells, thus simplifying interpretation of the results. Addition of cleaved Bid to isolated mitochondria induced oligomerization of Bak, as evidenced by the presence of cross-linked protein complexes containing Bak (Fig. 7B). Preincubating cleaved Bid with HN peptide caused concentration-dependent suppression of tBid-induced oligomerization of Bak.
Taken together, these results indicate that HN interferes the binding of tBid to Bax and Bak, preventing the oligomerization of Bax and Bak, and thereby presumably preventing tBidinduced release of proteins from mitochondria.
Structure-Function Analysis of HN Peptide Interaction with Bid-Our previous data showed that the first 1-17 amino acids of HN are required for binding to Bax (30). To begin to delineate some of the structure-function relations of the HN peptide with respect to tBid binding and suppression, we compared the Then, mitochondria were solubilized in 2% CHAPS, and proteins were size fractionated by gel sieve chromatography, and the resulting fractions were analyzed by SDS-PAGE/immunoblotting using Bax antibody. Molecular mass standards are indicated. B, cleaved Bid was preincubated with various concentrations of HN peptide, then added to 50 g of mitochondria from baxϪ/Ϫ MEFs. After incubation at 30°C for 40 min, samples were treated with cross-linker bismaleimidohexane (5 mM), then detergent lysates were prepared and analyzed by SDS-PAGE/immunoblotting using anti-Bak antibody. Bands corresponding to monomeric and multimers of Bak are indicated by arrowheads, with asterisks denoting cross-linked multimers. The double asterisk corresponds to an intramolecular cross-linked Bak monomer that displays faster mobility in gels. 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 continue the structure-function analysis of the HN peptide, we performed comparisons with previously published HN mutants that have been reported either to retain or lose their cytoprotective activity. For example, Hashimoto and colleagues have reported that a Ser 14 3 Gly mutant of HN exhibits enhanced cytoprotective activity, whereas a Cys 8 3 Pro mutant is devoid of apoptosis-blocking activity (24,29). In addition, we compared the HN peptide with an alternative version that represents a translation of the open reading frame of HN cDNAs using the mitochondrial codon rules, because this mi- FIG. 9. Structure-function analysis of HN peptide reveals correlation between tBid binding and cytoprotection from tBid. A, HEK293T cells were transfected with 2 g of plasmids encoding GFP, GFP-HN, GFP-HN(S14G), GFP-HN(C8P), or GFP-HN(mito) together with 1 g of pcDNA3-myc-tBid. Cells were cultured for 24 h in 50 M ZVAD-fmk to prevent cell death, then lysates were prepared and used for immunoprecipitation (IP) experiments, employing anti-myc antibody. Immune complexes were analyzed by SDS-PAGE/immunoblotting (WB) using anti-GFP antibody (top panel). To verify expression of all proteins, equivalent volumes of lysates were also loaded directly in gels and analyzed by SDS-PAGE/immunoblotting using antibodies specific for GFP (middle panel) or myc (bottom panel). The asterisk denotes degradation product of GFP. B-D, for caspase assays, cells were transfected as above, then lysates were prepared 20 h after transfection, normalized for protein content (10 g), and incubated with 100 M Ac-DEVD-AFC. Enzyme activity was determined by release of fluorogenic AFC product, and V max was calculated, expressing data as RFU/min/mg (mean Ϯ S.D.; n ϭ 3). E and F, 100 ng of cleaved Bid protein was preincubated with various concentrations of HN(S14G) or HN(C8P) peptides for 15 min in HM buffer, then 50 g of isolated mitochondria was added. After incubation at 30°C for 1 h, samples were centrifuged to generate supernatant and pellet fractions, which were analyzed by SDS-PAGE/immunoblotting using anti-SMAC or anti-Hsp60 antibody, respectively. tochondrial translation of HN has been shown previously to retain its ability to bind and suppress Bax (30). First, we compared the ability of HN(S14G), HN(C8P), and HN(mito) to bind tBid by coimmunoprecipitation. As shown in Fig. 9A, tBid bound wild-type HN, HN(mito), and the gain-of-function HN(S14G) mutant, but not the loss-of-function HN(C8P) mutant. Second, we evaluated the effects of these HN mutants on caspase activation induced by overexpression of tBid. Although HN(S14G) and HN(mito) suppressed in a concentration-dependent manner the ability of tBid to induce increases in cellular caspase activity, the HN(C8P) mutant did not (Fig. 9, B,  C, and D). Third, we tested HN(S14G) and HN(C8P) mutants with respect to their ability to suppress tBid-induced release of SMAC from isolated mitochondria (Fig. 9, E and F). The HN(S14G) peptide suppressed SMAC release with potency comparable with that of wild-type HN. In contrast, the HN(C8P) peptide failed to suppress tBid-induced release of SMAC from isolated mitochondria.

DISCUSSION
In this report, we provide evidence that the cytoprotective peptide HN can bind and suppress the activity of proapoptotic Bcl-2/Bax family protein Bid. Bid is found in an inactive conformation in the cytosol of cells, undergoing activation in association with proteolytic removal of an N-terminal domain (for review, see Ref. 15). This cleavage of Bid can be mediated by caspases, calpains, and lysosomal proteases, thus linking Bid to both apoptotic and necrotic cell death pathways of relevance to stroke, neurodegeneration, hepatitis, and other ailments (19 -21, 23). We show here that HN binds both full-length and caspase-8-cleaved Bid, blocking tBid-induced release of cytochrome c and SMAC from isolated mitochondria and suppressing apoptosis and caspase activation induced in intact cells by overexpression of tBid. These observations thus extend our prior discovery that HN binds and suppresses Bax, demonstrating that Bid represents an additional likely intracellular target of the HN peptide.
Bid and Bax have in common that both of these proapoptotic proteins are initially present in the cytosol in an inactive conformation but translocate to mitochondrial membranes after activation (15). Although HN peptide inhibits translocation of Bax from cytosol to mitochondria in intact cells and interferes with Bax association with isolated mitochondria in vitro (30), we failed to detect an inhibitory effect of HN on association of tBid with mitochondria. Moreover, when expressed in cells, HN (with a GFP tag) and tBid colocalized on punctate intracellular structures (presumably mitochondria). 2 Of note, the mechanism of Bid activation differs from Bax. In contrast to Bax where it has been proposed that an intramolecular rearrangement exposes a C-terminal transmembrane domain for membrane insertion (12), the Bid protein becomes activated by proteolytic removal of its N-terminal domain and subsequent myristoylation (13,14,42). Thus, we infer that the cytoprotective mechanism of HN is not dependent on blocking proteolysis of Bid or translocation of cleaved Bid to mitochondria. Rather, HN peptide inhibits tBid association with Bax and Bak and suppresses tBid-induced oligomerization of Bax and Bak, presumably explaining its ability to suppress release of cytochrome c and SMAC from isolated mitochondria. Moreover, HN retains these functions even when applied to baxϪ/Ϫ mitochondria, indicating that its protective mechanism is not dependent on Bax. Because prior analysis of cells from baxϪ/Ϫ, bakϪ/Ϫ double knock-out mice have demonstrated that either the Bax or Bak protein is sufficient for tBid-induced apoptosis (18), our results imply that HN directly interferes with Bid, preventing Bid-induced oligomerization Bak, and thereby suppressing release of apoptogenic proteins from mitochondria. In addition, HN directly inhibits Bax, thus providing two mechanisms for suppressing mitochondria-dependent cell death.
The BH3 domain of Bid mediates its dimerization with other Bcl-2/Bax family proteins (43). This domain is comprised of a 20-amino acid amphipathic ␣-helix, the hydrophobic face of which is required for interactions with Bcl-2/Bax family members (for review, see Ref. 44). The three-dimensional structure of full-length Bid indicates that its BH3 domain is buried in the core of the protein, implying that a profound conformational change in the Bid protein is required to expose the hydrophobic surface of the BH3 domain for interaction with proteins such as Bax and Bak (13,14). Proteolytic removal of the N-terminal region of Bid exposes the BH3 domain, but rotation of the hydrophobic face of the BH3 ␣-helix would still presumably be required for effective dimerization with other Bcl-2/Bax family members. Consequently, we speculate that HN stabilizes the inactive conformation of tBid, preventing exposure of the BH3 domain, and thus accounting for our observation that HN inhibits binding of tBid to Bax and Bak. Structural studies are required to confirm this hypothesis.
The structure of the HN peptide is currently unknown. Prior mutagenesis studies indicated that substituting proline for the sole cysteine residue in the HN peptide abolishes its cytoprotective activity (24). This mutation also abolishes binding of HN to Bax (30) and, as shown here, to Bid. The HN(C8P) mutant peptide also failed to suppress tBid-induced release of apoptogenic proteins from isolated mitochondria and was defective at suppressing apoptosis induced by tBid overexpression. These data thus provide correlative evidence that binding of HN to proapoptotic Bcl-2/Bax family proteins is critical for the cytoprotective actions of the peptide. Cysteine 8 is conserved in all the four species including human, wheat, rat, and nematode. However, substitution of a proline at this position in the HN peptide could profoundly alter the structure of the peptide, thus explaining the loss of binding activity. Alternatively, the sulfhydryl side chain of the cysteine might participate directly in binding to target proteins such as Bid and Bax, but more extensive replacement mutagenesis using alternative amino acids (e.g. alanine, serine) is required before firm conclusions can be reached. Interestingly, we observed that the first 12 amino acids of the HN peptide are sufficient for binding and suppressing tBid activity in vitro and in intact cells. This finding raises hopes that the HN binding site on tBid might be sufficiently small to contemplate mimicking HN with membrane-permeable nonpeptidyl organic compounds. Such agents could prove useful for developing novel cytoprotective strategies for diseases where Bid has been implicated in animal models of disease (45). Of note, however, the S14G version of HN has been suggested to represent a gain of function mutation (29), suggesting that residues distal to the 1-12 segment may influence the conformation of the HN peptide.
Alternative theories have been advanced for explaining the cytoprotective mechanism of the HN peptide. One view holds that HN operates extracellularly after secretion from cells, binding to cell surface receptors, and triggering signaling events that promote cell survival (24,29). For instance, addition of 10 M HN peptide to the medium of cultured neurons was reported to rescue from cell death induced by cytotoxic amyloid-␤ peptide (24), and HN was also reported to bind insulin-like growth factor-binding protein 3 (46). Although we cannot exclude a role for HN as a ligand for an extracellular receptor, our experimental comparisons of using membranepermeable and -nonpermeable HN peptides suggest that cyto-protective activity is dependent on access to intracellular target, inasmuch as polyarginine-containing HN peptides provided protection from apoptosis induced by tBid overexpression, whereas unmodified HN peptide did not. In vitro experiments with isolated mitochondria showed that both unmodified HN and HN polyarginine synthetic peptides were active at suppressing tBid-induced release of SMAC, confirming that differences in the cytoprotective activity of unmodified and polyarginine-containing HN peptides do not provide a trivial explanation for the differential activity of these synthetic peptides when added to cell cultures. We cannot however exclude the possibility that addition of polyarginines to HN peptides improves their stability in serum-containing cultures or produces other unanticipated differences in their activity. Nevertheless, the hypothesis that HN targets intracellular rather than (or in addition to) extracellular proteins is independently bolstered by our observations that: (a) GFP-tagged HN expressed within cells from plasmids suppresses apoptosis and also inhibits SMAC release from mitochondria in intact cells that overexpress tBid; and (b) synthetic HN peptides suppress tBid-induced release of apoptogenic proteins from isolated mitochondria, in the absence of plasma membrane or extracellular proteins.
Our data do not address the issue of whether HN is a physiological inhibitor of Bid in vivo, and at present, little is known about the endogenous roles of the HN peptide. However, the discovery that HN is capable of binding and suppressing the activity of selected proapoptotic members of the Bcl-2/Bax family provides new insights into the mechanisms regulating these killer proteins and suggests that HN could provide a useful research tool for investigating the roles of Bax and Bid in models of cell death. Moreover, the observation that HN binds and suppresses the activity of Bax and Bid suggests a possible strategy for development of cytoprotective peptidomimetic or nonpeptidyl compounds that might prove useful for preserving cell survival ex vivo for bioproduction or in vivo for preventing tissue loss in settings of ischemia-reperfusion injury, inflammation, or neurodegenerative conditions where gene ablation studies in mice have suggested important roles of Bid and Bax (18,47).
In an accompanying paper, we present evidence that HN also binds to and inhibits Bim EL (48). Bim and Bid share in common the characteristics that they are (a) the only BOPs that function as agonists of Bax and Bak, and (b) translocation from cytosol to mitochondria during apoptosis induction. This observation, taken together with studies of HN peptide mutants and our previous investigations of Bax, suggest that Bax, Bid, and Bim EL possess shared structural features that render them targets of HN-mediated suppression. These conserved characteristics forecast possibilities for generating cytoprotective peptides or compounds that might simultaneously neutralize Bax, Bid, and Bim EL.