J Biol Chem, Vol. 274, Issue 34, 23687-23690, August 20, 1999

COMMUNICATION
A Novel BH3-like Domain in BID Is Required for Intramolecular Interaction and Autoinhibition of Pro-apoptotic Activity^*

Kuan Onn Tan, Karen Mei Ling Tan, and Victor C. Yu

From the Institute of Molecular and Cell Biology, National University of Singapore, 30 Medical Drive, Singapore 117609, Republic of Singapore

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

Upon activation of the Fas apoptotic signaling pathway, Bid, a "BH3 domain-only" pro-apoptotic molecule, is cleaved by caspase-8 into a 6.5-kDa N-terminal and a 15-kDa BH3 domain-containing C-terminal fragment, referred to as tⁿ-Bid and t^c-Bid, respectively. t^c-Bid is a more potent inducer of apoptosis than full-length Bid, suggesting that the N-terminal region of Bid has an inhibitory effect on its pro-apoptotic activity. Here, we report the identification of a novel BH3-like motif (amino acid residues 35-43) in tⁿ-Bid. Although Bid does not homodimerize, tⁿ-Bid is able to associate avidly with t^c-Bid. Site-directed mutagenesis revealed that both the novel BH3-like and BH3 domains are necessary for direct binding between tⁿ-Bid and t^c-Bid. While full-length Bid does not associate with tⁿ-Bid, substitution of Leu³⁵, a critical residue in mediating tⁿ-Bid/t^c-Bid interaction, with Ala in full-length Bid is sufficient to establish Bid/tⁿ-Bid interaction. Interestingly, the L35A Bid mutant is as effective as t^c-Bid in inducing apoptosis and binding Bcl-X_L. We propose that the intramolecular interaction involving the BH3-like and BH3 domains serves to regulate the pro-apoptotic potential of Bid.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

The Bcl-2 family of proteins are important regulators of cell death that can be grouped into subfamilies of pro-survival and pro-apoptotic molecules (1-3). Members of this family are characterized by the presence of several conserved motifs, known as the Bcl-2 homology (BH1-4) domains. Association of a pro-survival with a pro-apoptotic member, such as Bcl-X_L with Bak, requires the BH1, BH2, and BH3 domains of Bcl-X_L and the BH3 domain of Bak (4-5). In addition to being a protein-protein interaction domain, the BH3 domain in pro-apoptotic molecules is also required for their death-promoting function (6-8).

Recently, a new group of cell death promoters that possess only the BH3 domain has been identified. These proteins, termed collectively as the "BH3 domain-only" molecules, which include Bik, Bad, Bid, Hrk, Bim, Blk, and EGL-1 of Caenorhabditis elegans (9-15), interact with Bcl-2 or Bcl-X_L and induce cell death when overexpressed (9-14).

Other than the conserved amino acids of the BH3 domain, the BH3 domain-only molecules have no similarity in amino acid sequence among themselves, suggesting that the death-promoting functions of these molecules could be regulated by distinct upstream signaling mechanisms. Indeed, Bad was found to be a component of the interleukin 3 survival signaling pathway (16). Bad is phosphorylated on serine residues by activated Akt, a serine threonine protein kinase that can be activated by interleukin 3 (17, 18). Phosphorylated Bad is non-apoptotic and binds 14-3-3 instead of Bcl-X_L or Bcl-2 (16). Another BH3 domain-only protein, Bid, has recently been identified to be a proximal caspase-8 substrate of the tumor necrosis factor receptor 1 (TNFR1) and Fas (also known as CD95 or APO-1) signaling pathways (19-21). Upon receptor activation, Bid is cleaved by caspase-8 into two major fragments, a 6.5-kDa N-terminal and a 15-kDa BH3 domain-containing C-terminal fragment, referred to as tⁿ-Bid and t^c-Bid, respectively (19-21). t^c-Bid appears to bind Bcl-X_L more avidly than full-length Bid and is found to be rapidly translocated from the cytosol to the mitochondria (19-21). t^c-Bid is also more potent than full-length Bid in inducing cytochrome C release and apoptosis (19-21). It has been suggested that the cleavage of Bid unmasks the pro-apoptotic potential of Bid by removing the inhibitory N terminus and allowing the C-terminal BH3 domain to interact with receptors on the mitochondria (19-21). However, the molecular basis of this autoinhibitory function conferred by the N-terminal region is currently undefined. Here we report the identification of a novel BH3-like domain in the N-terminal region of Bid that is necessary for mediating interaction between the N and C termini of Bid. Furthermore, interaction through the novel BH3-like domain appears to be important for regulating the apoptotic activity of Bid. We propose that the pro-apoptotic potential of Bid is negatively regulated by an autoinhibitory mechanism involving intramolecular interaction.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

Yeast Two-hybrid Experiment-- The methodology and reagents employed for the yeast two-hybrid experiments were essentially the same as described (22). pAS2-1 and pGAD424 vectors were used for all expression constructs cloned in Gal4 DNA binding domain (GBD) and Gal4-transactivation domain (GAD), respectively.

Construction of Mutants-- Mutations in tⁿ-Bid, t^c-Bid, and Bid were introduced by site-directed mutagenesis (Transformer ^TM CLONTECH) method as described by the manufacturer. All deletion constructs were generated by polymerase chain reaction using appropriate primers and Pfu® polymerase enzyme (Stratagene). The DNA sequences of all constructs used were confirmed by sequencing.

In Vitro Binding Assay-- Bacterial GST¹ fusion proteins were prepared as described (23). The integrity of the in vitro translated ³⁵S-labeled protein was verified by SDS-PAGE. Aliquot (2-5 µl), equivalent to 700,000 cpm of trichloroacetic acid-precipitable counts of the ³⁵S-labeled protein, was diluted into 0.2 ml of binding buffer (20 mM Tris, pH 8, 1 mM EDTA, 150 mM NaCl, 0.2% Nonidet P-40, 1 mM PMSF, 50 µg/ml aprotinin, and 10 µg/ml leupeptin) and incubated 2-4 h at 4 °C with 2-4 µg of GST-fusion proteins immobilized on the beads. Samples were subsequently washed six times with binding buffer and boiled for 3 min in loading buffer before analysis on 12.5% SDS-PAGE. Gels were subsequently stained with Coomassie Blue, destained, and dried. Bound proteins were visualized following autoradiography.

Immunoprecipitation-- 293T cells seeded on 100-mm plate at 70% confluency were transiently co-transfected with 10 µg each of expression plasmids (pCMV) encoding the indicated N-terminal HA- and Myc-tagged proteins using the LipofectAMINE method (Life Technologies, Inc.). After 12 h of transfection, cells were incubated with fresh media for 6 h before harvesting, and the cell pellet was lysed in 1 ml of binding buffer. Cell lysates were incubated with 1 µg of polyclonal anti-HA antibody for 1 h on ice, mixed with 20 µl of a 1:1 slurry of protein A-agarose and incubated for another 1 h at 4 °C. The agarose beads were washed three times in 1 ml of lysis buffer containing 500 mM NaCl before the proteins were eluted and analyzed on SDS-PAGE. Western blotting analyses were performed using monoclonal anti-Myc antibody and enhanced chemiluminescent detection (24).

Apoptosis Assay-- Apoptosis assay was performed in MCF-7 cells. Cells seeded on 35-mm plate at 70% confluency were transiently co-transfected with 1.5 µg each of the expression plasmids pCMV-HA Bid, Bid mutants or vector control, and pCMV--galactosidase (0.25 µg). Transfections were carried out with LipofectAMINE for 6 h followed by change of media 18 h later, cells were fixed and incubated in 5-bromo-4-chloro-3-indolyl -D-galactopyranoside buffer to mark the -galactosidase-expressing cells. Percentage of apoptotic cells was defined as the round blue cells over the total blue cells counted (500-800 cells) from five randomly chosen fields.

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

To investigate the nature of the inhibitory activity conferred by the N-terminal region of Bid, we examined the amino acid sequence of Bid for clues. Interestingly, a stretch of nine amino acids (acids 35-43) highly similar to the core motif of the BH3 domain was identified within the tⁿ-Bid fragment (Fig. 1A). The putative BH3-like motif, referred to as BH3-B, from both mouse and human, was aligned against several established BH3 domains, including the BH3 domain present in t^c-Bid (Fig. 1B). With the exception of the conserved aspartic acid residue, the putative BH3-like motif aligned well with established BH3 domains.

View larger version (39K): [in this window] [in a new window]   Fig. 1.   Identification of a novel BH3-like motif in the N-terminal region of Bid. A, schematic representation of Bid and its cleavage products. Cleavage site for caspase-8 is at Asp59 (D59). The N- and the C-terminal cleavage products are referred to as tn-Bid and tc-Bid, respectively. Solid gray box denotes the established BH3 domain. Solid black box represents the N-terminal BH3-like (BH3-B) domain. B, amino acid sequence alignments of BH3 and BH3-B domains of Bid derived from both human and mouse with several BH3 domains from selected members of the Bcl-2 family. Amino acids that are identical or functionally conserved (I, V, L, and M) in at least nine of eleven sequences are shaded.

Because the BH3 domain is known to function as a protein-protein interaction domain, we tested the ability of the N-terminal region of Bid that contains the novel BH3-like motif to interact with the C-terminal region. GST-tⁿ-Bid was found to associate selectively with in vitro translated t^c-Bid, but not Bcl-X_L, or Bax (Fig. 2A). Interestingly, tⁿ-Bid failed to associate with in vitro translated full-length Bid (Fig. 2A), raising the possibility that the region of Bid required for interaction with tⁿ-Bid was masked. The association of tⁿ-Bid with t^c-Bid was also demonstrated in vivo in mammalian cells. In transient co-transfection analysis in 293T cells, tⁿ-Bid was readily co-immunoprecipitated with t^c-Bid, but not with Bid, Bax, or itself (Fig. 2B). Similar results were obtained with the yeast two-hybrid system, in which tⁿ-Bid was found to strongly interact with t^c-Bid, but not Bcl-X_L or Bax (Fig. 2C). The strength of tⁿ-Bid/t^c-Bid interaction detected in the yeast two-hybrid system was found to be similar to the interaction between Bcl-X_L and BimL (Fig. 2C).

View larger version (26K): [in this window] [in a new window]   Fig. 2.   tn-Bid interacts with tc-Bid in vitro and in vivo. A, in vitro binding. Equal amount of GST-tn-Bid fusion proteins were incubated with indicated 35S-labeled proteins as described under "Experimental Procedures." The same gel was stained with Coomassie Blue to demonstrate the equivalency of the amount of GST-fusion proteins used in each experiment (bottom panel). B, in vivo binding in mammalian cells. Cell lysates from 293T cells transiently expressing the indicated N-terminal HA- and Myc-tagged proteins were subjected to immunoprecipitation with the polyclonal HA-antibody (Y11, Santa Cruz Biotechnology). Co-immunoprecipitated Myc-tagged proteins were detected on Western blot with the monoclonal anti-Myc antibody (9E10, Santa Cruz Biotechnology). Aliquots (2% of total) of total lysates were analyzed by Western blots to document the levels of expression of the HA- and Myc-tagged proteins (bottom panels). C, interaction of tn-Bid and tc-Bid in the yeast two-hybrid assay. tn-Bid were co-transformed with tc-Bid or other fusion constructs into yeast strain Y190. Transformants were tested for -galactosidase activity in filter assay. Strength of interaction was graded as follows: +++, strong blue; ++, medium blue; and , no blue. GBD, Gal4 DNA binding domain; GAD, Gal4-transactivation domain.

Because the motifs found in the tⁿ-Bid and t^c-Bid are the respective BH3-like and BH3 domains and several conserved residues of BH3 domains have been shown to be necessary for interaction with Bcl-2 or Bcl-X_L in some pro-apoptotic molecules (5, 11, 25-28), we proceeded to evaluate the requirement of conserved residues of the BH3 domains in mediating interaction between the tⁿ-Bid and t^c-Bid by mutational analyses. Substitution of the Gly⁹⁴ with Glu (G94E) in Bid (11) or t^c-Bid was found to be sufficient to abolish interaction of these molecules with Bcl-X_L (data not shown). The G94E-t^c-Bid (t^c-Bid.m1) failed to bind tⁿ-Bid in the yeast two-hybrid and GST-pulldown assays (Fig. 3, B and C), suggesting Gly⁹⁴ in the BH3 domain of t^c-Bid is essential for interaction with both tⁿ-Bid and Bcl-X_L. Substitution or deletion of several conserved amino acids simultaneously at the core motif of the BH3 domain was found to be effective in abolishing its dimerization function (25, 28). Mutating the BH3 domain of t^c-Bid by substituting all three amino acid residues at the core region, Gly⁹⁴-Asp-Glu/Val-Leu-Ala (t^c-Bid.m3), diminished the binding of t^c-Bid to tⁿ-Bid (Fig. 3, B and C). These data demonstrate the importance of the BH3 domain of t^c-Bid in mediating interaction with tⁿ-Bid.

View larger version (34K): [in this window] [in a new window]   Fig. 3.   Mutational analyses of the BH3 and BH3-B domains of mBid. A, schematic representation of various mutations generated within the BH3 and BH3-like (BH3-B) domains of Bid. The indicated amino acids in the context of Bid, tn-Bid, or tc-Bid were substituted by site-directed mutagenesis. B, interaction of tn-Bid, tc-Bid, and mutants in the yeast two-hybrid assay. tn-Bid or tn-Bid mutants were co-transformed with tc-Bid or tc-Bid mutants into yeast strain Y190. Transformants were tested for -galactosidase activity in filter assay. Strength of interaction was graded as follows: +++, strong blue; +, weak blue; and , no blue. GBD, Gal4 DNA binding domain; GAD, Gal4-transactivation domain. C and D, in vitro binding assay. Equal amounts of 35S-labeled tc-Bid or tc-Bid mutants were incubated with GST-fusion proteins of tn-Bid or tn-Bid mutants as described under "Experimental Procedures." The same gels were stained with Coomassie Blue to verify that equivalent amount of the GST fusion proteins used in each sample (bottom panels).

The Asp⁹⁵ residue in the BH3 domain of Bid is invariant among essentially all BH3 domains (Fig. 1B). However, in the BH3-B domain of both human and mouse Bid, the corresponding position is occupied by positively charged amino acids, histidine and arginine, respectively (Fig. 1B). Mutations of the positively charged arginine to the negatively charged aspartic acid, R40D (tⁿ-Bid.m6) or glutamic acid, R40E (tⁿ-Bid.m7), in tⁿ-Bid had no effect on the strength of interaction between these mutants and t^c-Bid (Fig. 3, B and D). Similarly, a mutation of the negatively charged conserved aspartic acid in the BH3 domain of t^c-Bid to the positively charged histidine, the D95H (t^c-Bid.m2), also failed to weaken the binding of t^c-Bid to tⁿ-Bid (Fig. 3, B and C). It seems that the charged amino acid at the position of the conserved aspartic acid has no significant role in tⁿ-Bid/t^c-Bid interaction. The functional significance of the conserved aspartate residue of BH3 domains in pro-apoptotic molecules for mediating interaction with Bcl-X_L has also not been totally resolved. While NMR studies examining the BH3 domain-containing peptides of Bak documented the importance of the conserved aspartic acid for mediating interaction with Bcl-X_L (5), yeast two-hybrid and immunoprecipitation analyses examining the role of individual amino acid residues in the BH3 domain of Bax failed to demonstrate a requirement of the conserved aspartate residue for heterodimerization with Bcl-X_L (25, 26). It is possible that the conserved aspartate residue does not have the same role in all BH3 domain-containing pro-apoptotic molecules. In contrast to the G94E mutation of the BH3 domain in t^c-Bid, substitution of the conserved glycine to glutamic acid, G39E, in the BH3-B domain of tⁿ-Bid (tⁿ-Bid.m5) had no effect on the binding of tⁿ-Bid to t^c-Bid (Fig. 3, B and D), suggesting that the conserved glycine in the BH3-B domain of tⁿ-Bid is not required for interaction with t^c-Bid. Substitution of the conserved Leu¹⁵¹ of the BH3 domain to Ala in Bad is sufficient to abrogate association with Bcl-X_L (27). Substitution of the corresponding conserved Leu³⁵ in tⁿ-Bid to Ala, L35A (tⁿ-Bid.m4), dramatically reduced its ability to associate with t^c-Bid (Fig. 3, B and D). While the G39E and the R40D mutations of tⁿ-Bid did not affect the binding of tⁿ-Bid to t^c-Bid, substitution of all three amino acids residues constituting the core region of BH3 domain, G³⁹RE to VLA in tⁿ-Bid (tⁿ-Bid.m8), severely diminished its ability to bind t^c-Bid (Fig. 3, B and D). These results demonstrate that the various conserved amino acid residues of the novel BH3-like domain (BH3-B) are critical in mediating interaction between the tⁿ-Bid and t^c-Bid.

tⁿ-Bid associates with t^c-Bid, but not full-length Bid, suggesting that intramolecular interaction in Bid may be responsible for preventing it from interacting with tⁿ-Bid. If this is true, mutations such as L35A and G³⁹RE/VLA in tⁿ-Bid that disrupt interaction of tⁿ-Bid with t^c-Bid are likely to have a similar effect in Bid, resulting in disruption of the intramolecular interaction of Bid and allow the structurally exposed C terminus to interact with tⁿ-Bid. To test this hypothesis, we introduced L35A and G³⁹RE/VLA mutations into the BH3-B domain of full-length Bid and tested the ability of these mutants to interact with GST-tⁿ-Bid. As shown in Fig. 4A, the L35A and G³⁹RE/VLA Bid mutants, but not the G39E Bid, R40D Bid, or wild-type Bid, were able to associate with tⁿ-Bid effectively. These data, together with the observation that Bid does not form homodimers (11), suggest a model in which the N and C termini of Bid interact intramolecularly.

View larger version (29K): [in this window] [in a new window]   Fig. 4.   The novel BH3-B domain in Bid is required for intramolecular interaction and autoinhibition of pro-apoptotic activity of Bid. A, tn-Bid binds to L35A and G39RE/VLA Bid mutants but not Bid. Full-length Bid and mutants were in vitro translated, 35S-labeled, and incubated with GST-tn-Bid. Coomassie Blue-stained gel demonstrating the equivalency of GST-fusion protein used in each binding experiment is shown (bottom panel). B, mutations in the BH3-B domain of Bid enhance its apoptotic activity to that of tc-Bid. MCF-7 cells were transiently transfected with 1.5 µg of the indicated expression plasmids and 0.25 µg of pCMV--galactosidase. Apoptosis assay was performed as described under "Experimental Procedures." The data (mean ± S.D.) shown are percentage of round blue cells as a function of total number of blue cells counted (about 500-800 cells per sample from five randomly chosen fields). A minimum of three independent experiments were performed for each mutant, and similar results were observed. C, Bcl-XL antagonizes the apoptotic activity of tc-Bid more effectively than tn-Bid. tc-Bid expression plasmid (0.1 µg) was co-transfected with the indicated fold excess (µg) of plasmids. Total plasmids transfected were balanced with empty vector to 2 µg/plate. Apoptosis assay was performed as described in panel B. D, Bcl-XL binds to L35A and Gly39-Arg-Glu/Val-Leu-Ala Bid mutants more efficiently than to wild-type Bid. GST-Bcl-XL was incubated with in vitro translated, 35S-labeled Bid and Bid mutants and subjected to GST-pulldown assay as described in panel A. Bottom panel shows the amount of GST-Bcl-XL used in each binding experiment.

We then examined the possibility that the intramolecular interaction of Bid acts as an autoinhibitory mechanism to regulate the pro-apoptotic activity of Bid. In transient apoptosis assay, tⁿ-Bid is totally inactive, whereas t^c-Bid is much more potent than Bid in inducing cell death as previously reported (Refs. 19-21; Fig. 4B). Similar to the wild type Bid, the G39E Bid mutant that failed to bind tⁿ-Bid was also mildly apoptotic (Fig. 4B). However, the apoptotic activities of both the L35A and the G³⁹RE/VLA Bid mutants were substantially elevated to a level similar to the t^c-Bid (Fig. 4B). These observations suggest that the intramolecular interaction between the N and C termini of Bid regulates its apoptotic activity. However, in order for t^c-Bid to exert its potent apoptotic effect in vivo, mechanism(s) must exist to prevent the free tⁿ-Bid from interfering with t^c-Bid to perform its apoptotic function. Indeed, the apoptotic activity of t^c-Bid was not very efficiently blocked by tⁿ-Bid, though it was potently blocked by Bcl-X_L overexpression (Fig. 4C). This is not surprising as one would predict that, once cleaved, the N-terminal region of Bid should not be effective in its ability to interfere with the apoptotic function of t^c-Bid.

Because binding of t^c-Bid to Bcl-X_L is widely believed to be a key mechanism by which Bid exerts its apoptotic effect upon cleavage by caspase-8, it would be of value to examine the ability of N-terminal region of Bid in preventing Bid from binding to Bcl-X_L. We compared the binding of Bid, t^c-Bid, and Bid mutants to Bcl-X_L. Interestingly, wild-type Bid and the Bid mutants (Bid.mt5 and Bid.mt6) that were predicted to maintain intramolecular interaction associated very weakly with Bcl-X_L, whereas the Bid mutants (Bid.mt4 and Bid.mt8) that were predicted to lose the ability to interact intramolecularly were found to interact as strongly as t^c-Bid to Bcl-X_L (Fig. 4D). Similar results were obtained with co-immunoprecipitation experiment in transient co-transfection assay (data not shown). These data lend support to a model in which the N-terminal region serves to regulate the apoptotic activity of Bid by inhibiting the association of the C-terminal domain with Bcl-X_L.

While the manuscript was in preparation, the solution structure of Bid was reported (29, 30). Bid contains eight alpha helices arranged in such a way that two central hydrophobic helices are surrounded by six amphipathic helices. Two of the amphipathic helices (H1 and H2) are present in tⁿ-Bid. The BH3-B motif described here (Fig. 1A) forms part of the H2 amphipathic helix (29). Our study thus further defines the function of the H2 helix by demonstrating its role in mediating direct physical contact with the C terminus.

	ACKNOWLEDGEMENTS

We are grateful to Drs. Stanley J. Korsmeyer, Craig B. Thompson, Suzanne Cory, and David Huang for providing many of the cDNA clones employed in this study. We thank Drs. Shing-Ling Chan, Bor Luen Tang, and Karen Yee for critical reading of the manuscript.

	FOOTNOTES

^* This work was supported by grants from the National Science and Technology Board of Singapore.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence should be addressed. Tel.: 65-874-3740 and 874-6862; Fax: 65-779-1117; E-mail: mcbyuck@imcb.nus.edu.sg.

	ABBREVIATIONS

The abbreviations used are: GST, glutathione S-transferase; PMSF, phenylmethylsulfonyl fluoride; PAGE, polyacrylamide gel electrophoresis.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES