The B-cell lymphoma 2 (Bcl-2) family of proteins are key regulators of cell death pathways. The first identified Bcl-2 family member, the Bcl-2 proto-oncogene, was originally identified as a gene linked to an immunoglobulin locus at the chromosomal breakpoint of t(14:18) in follicular lymphoma where its transcription is upregulated by the immunoglobulin heavy chain promoter [Bakhshi et al., 1985; Cleary and Sklar, 1985; Tsujimoto et al., 1985]. Contrary to the zeitgeist of a proto-oncogene at the time, Bcl-2 was found to inhibit apoptosis following pathological and physiological stimuli and not to promote cell proliferation [Vaux et al., 1988]. The observation that Bcl-2 localised to the mitochondrion was unexpected and nominated this organelle as an apoptotic regulator [Hockenbery et al., 1990].
Bcl-2 protein regions
Bcl-2 proteins contain between one to four regions of homology, known as the Bcl-2 homology (BH) domains. The BH domains roughly correspond to -helices which determine the structure and function of Bcl-2 family members [Chittenden et al., 1995; Danial and Korsmeyer, 2004; Yin et al., 1994]. Members of the Bcl-2 family can be characterised functionally into pro-apoptotic and anti-apoptotic members. The pro-apoptotic family members are further divided into two functional subgroups based on structure and function [Strasser, 2005].
Anti-apoptotic Bcl-2 family members
The anti-apoptotic Bcl-2 family members include, Bcl-2, Bcl-XL (extra long), Bcl- W, Mcl-1 (myeloid cell leukaemia-1), A1 and BOO (Bcl-2 homologue of ovary) [Youle and Strasser, 2008]. Anti-apoptotic Bcl-2 family members contain four BH domains (BH1-4) with the exception of Mcl-1 which only contains BH1 and BH3.
Overexpression of either family member can protect cultured cells from a range of cell death stimuli including UV irradiation, growth factor withdrawal and DNA damage [Strasser, 2005]. Bcl-2 is essential for the embryonic development of mice, and bcl2-/- mice show growth retardation and early postnatal mortality [Veis et al., 1993]. Mcl-1 and Bcl-2 are also required for the long term lifespan of mature B-cells and T-cells [Nakayama et al., 1993; Opferman et al., 2003]. Bcl-XL is required for the development of mice, and bcl-XL -/- mice die around embryonic day 13 due to extensive cell death of the spinal cord, hepatocytes and the developing brain [Motoyama et al., 1995]. The anti-apoptotic Bcl-2 family have been found to predominately reside at the mitochondria where they prevent MOMP by sequestering pro-apoptotic Bcl-2 family members.
Pro-apoptotic Bcl-2 family members
The pro-apoptotic Bax/Bak-like proteins which include Bax (Bcl-2 associated X protein), Bak (Bcl-2 agonist of killer), Bok (Bcl-2 ovarian killer) and Bcl-Xs (extra short) contain three BH domains (BH1-3) with the exception of Bcl-XS which contains BH3 and BH4 [Strasser, 2005]. Bax and Bak are essential for cell death induced by developmental cues, cytotoxic drugs or cytokine withdrawal [Ren et al., 2010]. In healthy cells anti-apoptotic Bcl-2 family members restrain Bak and Bax pro-apoptotic activity through heterodimerisation [Youle and Strasser, 2008].
The tertiary structure of Bax consists of 9 -helices, where -helices 1 through 8 show structural similarity to Bcl-XL, however, why Bax is a pro-apoptotic protein and Bcl-2 is a anti-apoptotic protein has been a matter of intrigue. Recently, replacement of -helix 5 of Bcl-XL with the corresponding helix from Bax has been shown to alter Bcl-XL from an anti-apoptotic to a pro-apoptotic protein [George et al., 2007].
Following apoptotic stimuli Bak and Bax control the commitment of cells to undergo apoptosis by either homo or heterodimerisation at the outer mitochondrial membrane (OMM) which regulates MOMP [Wei et al., 2001]. MOMP formation is proposed to occur in an ordered series of events, whereby the cleavage product of the pro-apoptotic protein Bid, tBid, binds to the mitochondrial membrane where it interacts with Bax and causes the insertion of Bax into the mitochondria culminating in MOMP [Lovell et al., 2008]. Recently a report has proposed a model whereby Bcl-XL maintains Bax in a cytosolic state through retro-translocation from the mitochondria, thus preventing Bax accumulation at the mitochondria and MOMP formation [Edlich et al., 2011].
Puma & Noxa
Puma and Noxa were identified through gene-expression profiling for targets of the tumour suppressor, p53, and from yeast two-hybrid screens using Bcl-2 as bait [Han et al., 2001; Nakano and Vousden, 2001; Oda et al., 2000; Yu et al., 2001]. The transcription factor, p53, transcriptionally upregulates Puma and Noxa following DNA damage, thus triggering apoptosis [Villunger et al., 2003a]. Puma and Noxa can bind and inhibit the anti-apoptotic Bcl-2 family members, Mcl-1, Bcl-2 and Bcl-XL, resulting in apoptosis [Nakano and Vousden, 2001; Oda et al., 2000; Yu et al., 2001]. Recently, Noxa has been implicated in H-Ras mediated autophagic cell death through increased expression and the displacement of Mcl- 1 from the autophagy regulator, Beclin-1 [Elgendy et al., 2010].
The BH3-only protein, Bad, was originally identified as Bcl-XL interactor through yeast two-hybrid and -phage library screens, and was found to selectively dimerise with Bcl-2 and Bcl-XL, but not Mcl-1 [Yang et al., 1995]. In the presence of the survival factor, IL-3, Bad becomes phosphorylated by the Ser/Thr kinase, Akt, resulting in its binding to the 14-3-3 proteins, thus inhibiting its apoptotic function [Datta et al., 1997; Zha et al., 1996]. Phosphorylation of Bad by Cdk1 in post-mitotic neurons results in the dissociation of Bad from 14-3-3 proteins and subsequent cell death [Konishi et al., 2002].
Bid was discovered as an interactor with both Bcl-2 and Bax [Wang et al., 1996]. Even though Bid contains a BH3 domain, it does not show sequence homology to either anti-apoptotic Bcl-2 proteins or multi-domain pro-apoptotic Bcl-2 proteins; however, it is structurally similar to Bax and Bcl-XL [McDonnell et al., 1999]. Full length Bid shows little apoptotic activity; however, it can induce cell death following glutamate receptor activation in rat hippocampal neuronal cultures [Konig et al., 2007]. Full activation of Bid activity requires proteolysis by either caspase-8 or the cytotoxic granule protease, granzyme B, forming truncated Bid (tBid) which subsequently undergoes N-myristoylation and translocation to the mitochondria where it can initiate apoptosis by activating Bax [Zha et al., 2000].
Bik is the founding BH3-only family member [Boyd et al., 1995] and is found to interact with Bcl-XL and the virally encoded Bcl-2 like family members, E1B-19K and EBV-BHRF1 [Chinnadurai et al., 2008]. Bik predominately localises to the endoplasmic reticulum (ER) where it mediates apoptotic induction through the release of Ca2+ from the ER and the recruitment of the mitochondrial fission protein, dynamin-related protein 1 (DRP1), from the cytosol, resulting in remodelling of the mitochondrial cristae and the release of cytochrome C [Germain et al., 2005; Germain et al., 2002].
Bmf binds to the myosin V motor complex through association with dyenin-light chain 2 in healthy cells [Puthalakath et al., 2001]. Bmf interacts with the antiapoptotic Bcl-2 family members, Bcl-2, Mcl-1, Bcl-XL and Bcl-W, however, it is not found to co-immunoprecipitate with Bax [Puthalakath et al., 2001]. During anoikis (the absence of cell attachment and intergrin signalling) Bmf is found to be released from the actin cytoskeleton and interact with Bcl-2 [Puthalakath et al., 2001]. UV irradiation also results in release of Bmf from the myosin V motor complex following JNK1/2-dependent phosphorylation resulting in its translocation to the mitochondria and subsequent cell death activation [Lei and Davis, 2003].
The BH3-only protein Bim has been extensively characterised as a mediator of intrinsic cell death pathways. Following cellular stresses such as UV irradiation, Taxol treatment and serum growth factor withdrawal, Bim localises to the mitochondria and inhibits the anti-apoptotic members of the Bcl-2 family, Bcl-2, Bcl-XL and Mcl-1 and activates Bak and Bax resulting in MOMP and subsequent cell death [Kutuk and Letai, 2010].
In vivo Bim is critical for cell homeostasis during mouse development. Nearly half of all Bim-/- mice exhibit developmental defects such as splenomegaly, persistence of interdigital webs, and an increase in the numbers of monocytes, granulocytes, B cells and T cells [Ren et al., 2010]. Furthermore, Bim-/- B and T cells show resistance to cell death induced by growth factor withdrawal [Bouillet et al., 1999; Ren et al., 2010]. In 1999, Bouillet et al. discovered that Bim -/- pre-T cells were 10 to 30 times more resistant to Taxol-induced cell death than wildtype cells [Bouillet et al., 1999]. Transient siRNA knockdown of Bim has also been shown to decrease Taxol-mediated cell death in the breast cancer cell lines, CRI-702, MDA-MB-231, SK-BR-3, and MCF-7, thus identifying Bim as aÿkey regulator in Taxol-mediated cell death [Li et al., 2005; Sunters et al., 2003]. The Bim gene, BCL2L11, encodes three major isoforms, BimEL, BimL and BimS [Ley et al., 2005a]. The transcriptional regulation of Bim is important for cell fate in a range of cellular processes from cell death to cell differentiation and cell survival [Ley et al., 2005a]. The Leukaemia/Lymphoma-related transcription factor (LRF) downregulates Bim and leads to the suppression of apoptosis and the continued production of red blood cells [Maeda et al., 2009]. Insulin-like growth factor 1 (IGF1) regulates Bim in multiple myeloma through a diverse range of mechanisms such as activation of the Akt pathway, inactivation of Forkhead box O3A (FOXO3A), and epigenetic regulation of the Bim and FOXO3A promoters[ De Bruyne et al., 2010].
ERK1/2has been shown to phosphorylate BimEL at multiple sites [Weston et al., 2003]. Phosphorylation of BimEL by ERK1/2 following serum withdrawal results in its degradation by 20S proteasomes [Wiggins et al., 2011]. BimEL is phosphorylated by ERK1/2 at up to six different residues, three of which are within exon 3, an exon only expressed in BimEL. The phosphorylation of BimEL by ERK1/2 requires the DEF-domain (docking site for ERK, FXFP) within exon 3.
Deletion of the DEF-domain within BimEL inhibits the phosphorylation of BimEL by ERK1/2 following serum withdrawal [Ley et al., 2005b]. Furthermore, deletion of the DEF-domain or mutation of Ser65 to alanine within BimEL results in its stabilisation following serum withdrawal and potentiates cell death [Ley et al., 2005b]. In addition, it was found that the ERK1/2-dependent phosphorylation of Bim at Ser65 was an essential priming step for the phosphorylation of Bim on other residues following serum withdrawal [Ley et al., 2004].
While the phosphorylation of BimEL by ERK1/2 is well characterised the role of ERK1/2 in the phosphorylation BimL remains controversial [Ley et al., 2005a]. BimL becomes phosphorylated in a MEK/MAPK dependent manner following treatment of PC12 cells with nerve growth factor [Biswas and Greene, 2002]. Moreover, BimL has been shown to become phosphorylated by ERK1/2 followingÿtreatment of the murine polylymphocytic cell line, FL5.12, with the haematopoietic survival factor, IL-3 [Shinjyo et al., 2001]. However, BimL does not contain a DEF-docking domain required for ERK1/2 docking. Moreover, BimL is not phosphorylated by ERK1/2 in an in vitro kinase assay [Ley et al., 2004]. Thus it is proposed that ERK1/2 may phosphorylate BimL while ERK1/2 is bound to a BimL partner protein such as dyenin light chain 1 (DLC1) or Bcl-2 Ley et al., 2005a].
Proteasomal degradation of Bim
The phosphorylation of BimEL by ERK1/2 leads to the dissociation of BimEL from the BimEL/MCL-1 and BimEL/Bcl-XL complexes, resulting in its polyubiquitination and subsequent proteasomal degradation [Ewings et al., 2007; Wiggins et al., 2010]. The identity of the E3 ubiquitin ligase responsible for the degradation of BimEL has been a matter of intense debate. Casillas B-lineage lymphoma (CBL) oncogene was initially proposed as the E3 ligase responsible for BimEL ubiquitination [Akiyama et al., 2003]. However, substrates of CBL classically contain phosphorylated tyrosine residues, whereas ERK1/2- dependent phosphorylation and degradation of BimEL has been shown to require serine phosphorylation. In addition, ERK1/2-dependent BimEL turnover proceeds normally in CBL-deficient cells, indicating the CBL is not required for BimEL destruction [Wiggins et al., 2007].
Intrinsic cell death & Bcl-2 family proteins
The sensitiser model of MOMP regulation
It was originally thought that BH3-only proteins indiscriminately bind and neutralise all Bcl-2 family members [Danial and Korsmeyer, 2004; Huang and Strasser, 2000]. However, Chen and colleagues demonstrated that BH3-only proteins show varying specificity and affinity towards their anti-apoptotic Bcl-2 counterparts. Bim, truncated Bid (tBid) and Puma can engage with all anti-apoptotic members and with the pro-apoptotic proteins Bax and Bak. Bad can only bind Bcl-2, Bcl-XL and Bcl-W, whereas Noxa is only capable of binding Mcl-1 and A1 and resulting in mitochondrial membrane permeabilisation (MOMP) [Chen et al., 2005].
Bax/Bak mediation of MOMP
The molecular mechanism by which BH3-only proteins regulate both anti-apoptotic Bcl-2 family members and Bax/Bak to initiate cell death has been extensively debated and two main models have been proposed. In the sensitiser model (also known as the indirect model) it is proposed that BH3-only proteins indirectly activate cell death through the inhibition of anti-apoptotic Bcl-2 family members which liberates Bak and Bax to homo-oligomerise resulting in MOMP formation and cell death initiation [Willis et al., 2007].
Bax/Bak mediation of MOMP
Mcl-1 and Bcl-XL constrain Bak until the suppression is lifted through the binding of Noxa to Mcl-1 and Bad to Bcl-XL, thereby eliciting Bak dependent apoptosis. Furthermore, Bax is restrained by binding to either Mcl-1, Bcl-2 and either Bcl-XL or Bcl-W, or collectively by all four anti-apoptotic Bcl-2 members, and activation of apoptosis requires neutralisation of anti-apoptotic Bcl-2 members [Chen et al., 2005; Willis et al., 2007]. Furthermore Bim and Bid are unable to interact directly with Bax and Bak. In addition, Bim -/- and Bid -/- double knockout mice undergo apoptosis following UV irradiation and etoposide treatment, suggesting that Bim and Bid are not necessary to activate Bak and Bax mediated apoptosis [Willis et al., 2007].
Binding affinities of BH3-only proteins towards their pro-survival counterparts
The hierarchical model of MOMP regulation
In the hierarchical activation model (also known as the direct model) it is proposed that BH3-only proteins can be divided into two subgroups, sensitisers and activators, which act at different stages in the hierarchical activation of Bcl-2 family members.
In a manner similar to the activation of the sensitiser model, the sensitiser subset of BH3-only proteins (Bad, Noxa, Bmf) engage anti-apoptotic Bcl-2 family members; however, in contrast to the sensitiser model, the hierarchical activation model suggests that the activator BH3-only subset (Bim, Puma and tBid) become released from the anti-apoptotic Bcl-2 family members to directly activate Bax and Bak resulting in MOMP [Galonek and Hardwick, 2006].
Recently it has been reported that triple null Bim-/-, Bid-/- and Puma-/- deficient mice display similar developmental defects to Bax-/- and Bak-/- mice and are refractory to apoptotic stimuli, thus supporting the hierarchical activation model [Ren et al., 2010]. The work of Ren et al. provides strong support for the proposal that Bim, Bid and Puma are required for Bax/Bak activation, while the remaining BH3-only subsets act as sensitisers.
MOMP Sensitiser Hierarchy
In the sensitiser model it is proposed that BH3-only family members neutralise the anti-apoptotic effect of the pro-survival Bcl-2 family members resulting in Bak/Bax-dependent mitochondrial outer membrane permeabilisation (MOMP). In the hierarchical model it is proposed that sensitiser BH3-only proteins (Bad, Bmf and Noxa) allow for the release of activator BH3-only proteins (Bim, Puma and Bid) to directly activate Bax and Bak and induce MOMP [Kim et al., 2006; Willis et al., 2007].