A novel BH3 ligand that selectively targets Mcl-1 reveals that apoptosis can proceed without Mcl-1 degradation

Abstract

Like Bcl-2, Mcl-1 is an important survival factor for many cancers, its expression contributing to chemoresistance and disease relapse. However, unlike other prosurvival Bcl-2-like proteins, Mcl-1 stability is acutely regulated. For example, the Bcl-2 homology 3 (BH3)-only protein Noxa, which preferentially binds to Mcl-1, also targets it for proteasomal degradation. In this paper, we describe the discovery and characterization of a novel BH3-like ligand derived from Bim, Bim(S)2A, which is highly selective for Mcl-1. Unlike Noxa, Bim(S)2A is unable to trigger Mcl-1 degradation, yet, like Noxa, Bim(S)2A promotes cell killing only when Bcl-x(L) is absent or neutralized. Furthermore, killing by endogenous Bim is not associated with Mcl-1 degradation. Thus, functional inactivation of Mcl-1 does not always require its elimination. Rather, it can be efficiently antagonized by a BH3-like ligand tightly engaging its binding groove, which is confirmed here with a structural study. Our data have important implications for the discovery of compounds that might kill cells whose survival depends on Mcl-1.

Figure 1.

Identifying residues in the BH3 region of Bim critical for binding prosurvival Bcl-2 proteins. (A–C) Individual residues on the phage-displayed BimBH3 (amino acids 51–76 of human Bim_S) were mutated to alanine except for native alanine or glycine, which were replaced instead with glutamic acid. Their affinities (in IC₅₀) for Bcl-x_L (A), Bcl-2 (B), and Mcl-1 (C) were determined by solution competition ELISAs (Kvansakul et al., 2007; Lee et al., 2007). The data represent the IC₅₀ (left) and fold reduction in binding compared with wild-type BimBH3 (right). The upper limits on the y axes were set to 1,000 nM (left) or at least a 40-fold reduction based on the IC₅₀ of a nonbinding mutant (G66E) independently confirmed in solution competition assays using an optical biosensor (Chen et al., 2005). (D–F) Validation of in vitro binding assays. Interactions between FLAG-tagged prosurvival Bcl-x_L (D), Bcl-2 (E), or Mcl-1 (F; black arrowheads) and HA-tagged Bim_S or selected BH3 mutants of it (white arrowheads) were assessed by coimmunoprecipitation. Equivalent ³⁵S-labeled lysates harvested from transiently transfected 293T cells were immunoprecipitated with antibodies to the FLAG (FL), HA, or control (C) tags. Wild-type Bim_S or mutants that bound with wild-type affinities (in A–C) to the prosurvival proteins are indicated by boxes. A nonbinding Bim_S mutant (Bim_S4E; Chen et al., 2005) served as a negative control. Data in A–C represent means ± SD (error bars) of three experiments; those in D–F are from representative experiments.

Figure 2.

Characterization of a novel Mcl-1–specific Bim variant. (A) A BimBH3 mutant that retains significant binding only to Mcl-1. The relative affinities of wild-type or the BimBH3 mutant 2A (L62A/F69A) peptides for prosurvival proteins were determined in solution competition assays using an optical biosensor. Compared with the wild-type sequence, the mutant 2A has reduced affinity for all prosurvival proteins except for Mcl-1. (B) Full-length Bim_S2A (top) interacts with Mcl-1 in cells (right), but, unlike wild-type Bim_S (bottom), it does not bind Bcl-x_L (left), which is consistent with the in vitro binding assays using purified components (A). Prosurvival proteins are indicated by black arrowheads, and Bim_S/Bim_S2A is indicated by white arrowheads. (C) Functional cooperation between the Mcl-1–specific Bim_S2A variant and Bad. Stable pools of wild-type MEFs were generated by hygromycin selection after retroviral infection with the parental pMIH vector or one expressing the full-length Mcl-1–specific Bim_S2A mutant. Long-term colony formation was assessed when these pools were reinfected with the control pMIG retrovirus or one expressing Bim_SBadBH3 (a Bim_S variant with its BH3 replaced with that of Bad; Chen et al., 2005). Bim_S2A was inert on its own, but there was a striking reduction in colonies formed if Bim_SBadBH3, which targets the prosurvival proteins (Bcl-x_L, Bcl-2, and Bcl-w) that Bim_S2A does not bind, is coexpressed. (D) The Mcl-1–specific BH3 ligands, full-length Bim_SNoxaBH3 or Bim_S2A, potently kill only when Bcl-x_L is absent, whereas Bim_SBadBH3 only kills the cells deficient in Mcl-1. Cell viability was assessed 24 h after infection with the indicated retroviruses. (E) HA-tagged wild-type Bim_S but neither the Bim_S2A nor the inert Bim_S mutant (Bim_S4E) coimmunoprecipitated with endogenous Bax. (F) FLAG-tagged Bak (middle) coimmunoprecipitates with HA-tagged Bcl-x_L and Mcl-1 but not with HA-tagged Bim_S4E, Bim_S, or Bim_S2A (top). Asterisks are the remnant signals from reprobing of the FLAG immunoprecipitation (top) with the FLAG antibody. Data in D represent means ± SD (error bars) of at least three experiments; those in A–C, E, and F) are from representative experiments. WCL, whole cell lysate.

Figure 3.

The Mcl-1–specific Bim variant does not trigger Mcl-1 degradation. (A) Immunoblots of equivalent lysates prepared from MEFs after infection with parental retroviral vector or ones expressing wild-type Bim_S, Bim_S4E, Bim_S2A, Bim_SNoxaBH3, or Noxa itself were probed with antibodies to Mcl-1 (top), Bcl-2 (middle), or HSP70 (loading control; bottom). Levels of Mcl-1 were significantly lower if Noxa or Bim_SNoxaBH3 (Bim_S with its BH3 domain replaced with that from Noxa) is expressed. In contrast, the Mcl-1–specific Bim_S2A variant stabilizes Mcl-1 levels. The images were assembled and cropped from a single gel. (B) Bim_S2A counters Mcl-1 degradation triggered by Noxa. Equivalent lysates were prepared from Noxa- or Bim_S4E-overexpressing MEFs after reinfection with the inert Bim_S4E, Noxa, or Bim_S2A retroviruses, and the blots were probed for Mcl-1 or HSP70 (loading control). Note that Mcl-1 levels are significantly higher in cells expressing the Mcl-1–specific Bim_S2A. (C) ITC analysis of BimBH3 2A and NoxaBH3 binding to Mcl-1. BimBH3 2A binds to Mcl-1 tighter than NoxaBH3.

Figure 4.

Bim_S2A efficiently inactivates Mcl-1 to induce cell killing. (A) Bim_S2A sensitizes cells to ABT-737, a BadBH3 mimetic. The viability of MEFs stably infected with the vector control or ones expressing Bim_S2A, Bim_SNoxaBH3, or Noxa was determined 24 h after treatment with 0–10 μM ABT-737. Although ABT-737, which targets Bcl-x_L, Bcl-2, and Bcl-w, was inert on its own, the three Mcl-1–selective BH3-only proteins were equipotent at cooperating with ABT-737 to efficiently kill cells. (B) The kinetics of killing with 1 μM ABT-737 were comparable regardless of whether Mcl-1 was degraded (with Noxa or Bim_SNoxaBH3) or neutralized (with Bim_S2A). (C) Bim_S2A promotes Bax/Bak-mediated cytochrome c release only when combined with ABT-737. Equivalent lysates prepared from wild-type (top) or bax^−/−bak^−/− (bottom) MEFs stably expressing the inert Bim_S4E or the Mcl-1–specific Bim_S2A were fractionated after incubation in vitro with 5 μM ABT-737 (+). Only the combination of Bim_S2A and ABT-737 caused cytochrome c release, which was abrogated in the absence of Bax and Bak (bottom). Blots for Bcl-2 (pellet fraction; p) and Apaf-1 (soluble fraction; s) served as markers for the subcellular fractionation. (D) Only when combined with ABT-737 does the Mcl-1–specific Bim_S2A kill cells, either through Bax or Bak without the need for the putative activator BH3-only proteins (Bim and Bid). Data in A, B, and D are means ± SD (error bars) of at least three experiments, whereas C is from a representative experiment.

Figure 5.

Long-term neutralization of Mcl-1 is not well tolerated. Wild-type MEFs (A) stably expressing Bim_S4E (black triangles), Bim_S2A (red circles), Noxa (blue diamonds), or Bad (green squares) were generated. Expression of the BH3-only proteins was linked to GFP expression through an IRES. Long-term expression of the inert Bim_S4E or Bad was maintained for nearly 40 d, as indicated by the proportion of cells that remained GFP^+ve. In contrast, targeting Mcl-1 (either with Bim_S2A or Noxa) was poorly tolerated unless the downstream mediators of apoptosis (Bax and Bak) were absent (B).

Figure 6.

The Mcl-1–specific Bim variant binds Mcl-1 like wild-type Bim. (A) Alignment of sequences of representative BH3 domains from BH3-only proteins or the Mcl-1–specific Bim_S variant Bim_S2A. The numbers in parentheses indicate the positions of residues in full-length proteins, and numbers in orange at the top indicate the positions in full-length Bim_S. The four conserved hydrophobic residues are boxed (blue) and labeled as h1–h4. Residues shaded in black are conserved in all proapoptotic BH3 domains. (B) Overlay of the crystal structures of human Mcl-1 (yellow) in complex with wild-type BimBH3 (orange) and that of human Mcl-1 (cyan) complexed with BimBH3 2A (blue). (C–E) Minor structural differences between the wild-type and mutant BimBH3 complexes were noted in the Mcl-1 regions proximal to L62 (C), F69 (D), and around residues Y72 and Y73 (E) in the BimBH3 peptide. (F) The effect on Mcl-1 binding of additional mutations (bold) in the context of the phage-displayed BimBH3 2A variant. The fold decrease in binding with respect to BimBH3 2A is shown. Residue I65 (position h3) but not I58 (position h1) appears critical for the binding of BimBH3 2A to Mcl-1.

Figure 7.

Physiological cell death initiated by Bim does not necessarily coincide with Mcl-1 degradation. (A) Down-regulating Bim attenuates thapsigargin-induced apoptosis. Viability of wild-type (−) or an MCF-7 subclone (+) stably expressing Bim shRNA treated with 1.5 μM thapsigargin for 0–48 h. (B) Treatment of wild-type MCF-7 cells with thapsigargin induced an increase in the levels of Bim (the isoforms include Bim_EL, Bim_L, and Bim_S; middle) that coincided with the onset of apoptosis seen in A. Mcl-1 levels increase upon treatment (top). Only low levels of Bim_EL were detectable in cells stably expressing Bim shRNA, whereas neither the Bim_L nor Bim_S isoforms were apparent. (C) Loss of Bim attenuates ionomycin-induced killing of thymocytes. The viability of wild-type (+/+) or Bim-deficient (−/−) thymocytes determined 0–24 h after treatment with 1 μg/ml ionomycin. Incubation with 50 μM of the pancaspase inhibitor qVD-OPH significantly reduced the killing induced by ionomycin. (D) Treatment of bim^+/+ thymocytes with ionomycin increases the levels of Bim_EL and Bim_L after 4 h (middle), which coincided with the onset of apoptosis seen in C. There was no significant concomitant loss of Mcl-1 (top). In B and D, immunoblots were reprobed with antibodies to HSP70 for loading control (bottom). In addition, the cells were treated with 50 μM qVD-OPH to avoid confounding effects caused by caspase cleavage of Mcl-1 (Herrant et al., 2004). Data in A and C represent means ± SD (error bars) of three experiments, and B and D are from representative experiments.