The use of a neutral peptide aptamer scaffold to anchor BH3 peptides constitutes a viable approach to studying their function

Abstract

The B-cell CLL/lymphoma-2 (Bcl-2) family of proteins are important regulators of the intrinsic pathway of apoptosis, and their interactions, driven by Bcl-2 homology (BH) domains, are of great interest in cancer research. Particularly, the BH3 domain is of clinical relevance, as it promotes apoptosis through activation of Bcl-2-associated x protein (Bax) and Bcl-2 antagonist killer (Bak), as well as by antagonising the anti-apoptotic Bcl-2 family members. Although investigated extensively in vitro, the study of the BH3 domain alone inside cells is more problematic because of diminished secondary structure of the unconstrained peptide and a lack of stability. In this study, we report the successful use of a novel peptide aptamer scaffold - Stefin A quadruple mutant - to anchor and present the BH3 domains from Bcl-2-interacting mediator of cell death (Bim), p53 upregulated modulator of apoptosis (Puma), Bcl-2-associated death promoter (Bad) and Noxa, and demonstrate its usefulness in the study of the BH3 domains in vivo. When expressed intracellularly, anchored BH3 peptides exhibit much the same binding specificities previously established in vitro, however, we find that, at endogenous expression levels, Bcl-2 does not bind to any of the anchored BH3 domains tested. Nonetheless, when expressed inside cells the anchored PUMA and Bim BH3 α-helices powerfully induce cell death in the absence of efficient targeting to the mitochondrial membrane, whereas the Noxa helix requires a membrane insertion domain in order to kill Mcl-1-dependent myeloma cells. Finally, the binding of the Bim BH3 peptide to Bax was the only interaction with a pro-apoptotic effector protein observed in this study.

Figure 1

Structural analysis of the SQT scaffold presenting four different pro-apoptotic BH3 domains. (a) Amino-acid sequence of the four BH3 peptides used in this study. In the Puma sequence, the underlined amino acids are deleted in pepPumaDel. (b) Nuclear magnetic resonance (NMR) structure of Stefin A (adapted from pdb ref: 1DVC). The three sites for peptide presentation in the SQT peptide aptamer scaffold are highlighted in red. (c) The CD spectrum of SQT alone and with N-terminal insertions of the HA tag (a negative control not expected to adopt an alpha helical configuration) or with each BH3 peptide. The single inflection point around 218 nm indicates predominantly β-strands in the SQT and SQT-HA spectra. A shift in the inflection point to 220–222 nm, the deeper inflection of the spectrum and an additional shoulder signal the alpha helix formed by the BH3 peptides. SQT with Noxa BH3 and the HA peptide have been reported previously

Figure 2

PepBH3 expression and identification of their binding partners. (a) Following induction, pepBH3 expression is detectable after 30 min. (b) Co-IP of pro- and anti-apoptotic Bcl-2 proteins with pepBH3 following their expression in A375 (top panel) and Hela cells (middle panel). There is no interaction between Bcl-2 and pepBim in A375 or Hela cells, however, when Bcl-2 is overexpressed (in A375) binding to pepBim is significant (bottom panel, underneath dashed line). (c) IP of endogenous full-length Bad from A375 cell lysate reveals binding of Bcl-xl, but not Bcl-2 (top panel). Equally in 293T cells, the Bad:Bcl-2 interaction does not occur (bottom panel). PARP cleavage indicates that apoptosis was induced successfully in these cells

Figure 3

PARP cleavage in A375 cells expressing pepBH3 proteins. (a) PepBim/Puma expression results in cleavage of PARP. (b) Cleavage of PARP occurs as early as 30 min after PepBim/Puma expression is initiated and downregulation of anti-apoptotic proteins starts after 2–4 h

Figure 4

Cell viability and localisation of pepPuma/Bim in cells. (a) PepPuma/Bim expression in A375 cells reduces cell viability by approximately 80%. The western blot shows levels of SQT with increasing amounts of Shield (error bars are ±S.D.; n=3). (b) PepPumaDel does not cause cell death (error bars are ±S.D.; n=3). (c) Immunofluorescence of Hela cells expressing SQT reveals no specific localisation pattern. PepBim or pepPuma (green) colocalise with mitotracker (red) in some but not all cells. DAPI-stained nuclei are shown in blue

Figure 5

PepPuma expression results in cytochrome c release from the mitochondria. (a) Shown are three time points in the process of apoptosis. In the left-hand panel, both cells in the main field of view have their mitochondria intact. In the central panel, the right-hand cell is starting to undergo apoptosis with mitochondrial membrane permeabilisation leading to the GFP signal becoming more diffuse. In the right-hand panel, the right-hand cell has completely lost mitochondrial integrity and GFP–cytochrome c has translocated to the nucleus. (b) White arrows indicate cells before (left panel) and after (right panel) mitochondrial permeabilisation

Figure 6

Apoptotic potential and cellular localisation of pepNoxa and pepBad. (a) Expression of pepNoxa in U266 cells has no effect on cell viability and does not cause PARP cleavage. PepNoxa with the BHR domain efficiently induces cell death and leads to PARP cleavage (error bars are ±S.D.; n=3). The inducible expression system is leaky in this cell type and thus pepNoxa-BHR expression occurs even in the absence of shield. The set of columns labelled ‘U266' denotes viability measurements of untransfected cells (which are not 100% viable). PepBH3 degradation products are indicated by an asterisk (*). (b) Immunofluorescence analysis of Hela cells expressing pepNoxa or pepBad (green) reveals no mitochondrial (red) localisation. PepNoxa-BHR colocalises very specifically with mitotracker (red). (c) Co-IP experiment using an anti-Stefin A antibody and lysate from cells expressing pepNoxa or pepNoxa-BHR. PepNoxa-BHR is expressed at lower levels, but purifies with a larger amount of Mcl-1 than pepNoxa lacking a BHR. (d) PepBad and MG-132 synergise to reduce viability in A375 cells (most prominently at 100 nM MG-132; error bars are ±S.D.; n=3)