2 Introduction


2.1 Cell death

2.1.1 Necrosis

Cell death occurring passively, in an unregulated and uncontrolled fashion is called necrosis (figure 1). This process is initiated when cell damage is too severe to overcome and when the cell is not able to activate the energy-dependent apoptotic pathway. Necrosis is marked by cell and organelle swelling due to augmentation of the membrane permeability. This leads to the release of lysosomes into the extra cellular matrix  (Kerr, et al., 1994 ; (Wyllie, et al., 1980). Intracellular contents are thrown into the extra cellular matrix often causing inflammation in the neighbouring cells.

Figure 1: morphological changes during apoptosis and necrosis

1) Normal cell
2) Shrinking of the cell, condensation of the chromatin and membrane blebbing
3) Nuclear fragmentation and formation of apoptotic bodies
4) Phagocytosis of the apoptotic bodies
A) Swelling of the organelles
B Disruption of the cell membrane
  (Kerr, et al., 1994 )

2.1.2 Apoptosis


The term apoptosis was established by Kerr (Kerr, et al., 1972 ). Apoptosis is a selective form of cell death that consists of morphological and biochemical changes. It is a highly conserved process, from C. elegans to man. It describes a controlled process consisting of shrinkage of the cell, blebbing of the plasma membrane, condensation of the chromatin leading to DNA fragmentation and eventually formation of so-called apoptotic bodies (Wyllie, et al., 1984 ) (figure 1). These are eliminated by phagocytosis to avoid inflammation (Fadok, et al., 2001). Apoptosis allows elimination of unwanted, aged, misplaced or damaged cells. Therefore, apoptosis is a crucial process in embryonic development, tissue homeostasis, differentiation, proliferation and immune response (Vaux and Korsmeyer, 1999 ). Developmental cell death occurs in a programmed fashion and is referred to as “programmed cell death” (type I cell death). However, this process can also be induced by exogenic factors like DNA-damaging agents, UV-irradiation and nutrition depletion. Inhibition or abnormal regulation of apoptosis underlies many disorders such as cancer and autoimmunity (Bakhshi, et al., 1985 ).

2.1.3 Autophagy

In cell death type II or autophagy, cell death occurs by degradation, where the cytoplasm or damaged organelles are recycled (Shintani and Klionsky, 2004 ). Autophagy is defined by the formation of the autophagosome, a double- or multi-membrane bound vacuole (Dunn, 1990 ). Fusion with lysosomes results in autophagolysosomes degrading their contents. Apart from this function, autophagy is engaged in cellular remodelling during differentiation and metamorphosis as well as aging, muscular disorder and neurodegeneration (Shintani and Klionsky, 2004 ). Autophagy might be an attempt to protect the cells against mitochondrial permeability transition during oxidative stress or mitochondrial calcium overkill (Rodriguez-Enriquez, et al., 2004 ).

2.2 Apoptotic pathways

Three main starting points of apoptosis are known, death receptors, mitochondria and the endoplasmic reticulum (Nakagawa and Yuan, 2000 ). These pathways are largely independent but can establish crosstalk since there are some junctions along their pathways.

2.2.1 The death receptor pathways


Extra cellular receptors transmit external signals into the cell. Besides signals that lead to proliferation and differentiation, cytotoxic signals are taken up by these receptors e.g. to eliminate target cells or excess immune cells after an immune response (Ashkenazi and Dixit, 1998, Schulze-Osthoff, et al., 1994 ). Initiation of the extrinsic or death receptor pathway (figure 2) occurs upon binding of extra cellular death ligands to the tumour necrosis factor (TNF) super-family of plasma membrane death receptors. Members of this receptor family are Fas/CD95, TNFR1, TRAIL receptors DR3, DR4 and DR5. They possess a cysteine rich extra cellular domain (Smith, et al., 1994 ). Upon binding of the death ligand, the death receptors form trimers. The intracellular death domain (DD) of the receptor recruits adaptor proteins like FADD (Fas Associated Death Domain) or TRADD (TNF-receptor associated protein with death domain) (Chinnaiyan, et al., 1995 ). In turn, these adapter proteins bind, with their N-terminal death effector domain (DED), to procaspase-8 leading to formation of a complex called death inducing signalling complex (DISC) (Muzio, et al., 1996 ). In this complex, monomeric caspase-8 becomes at least a dimer leading to autoprocessing and activation. Activated caspase-8 can proteolytically process and activate the effector caspase-3, -6, and -7 to amplify the death signal. Furthermore, caspase-8 can cleave Bid, a pro-apoptotic protein, which can initiate the mitochondrial apoptotic pathway (Li, et al., 1998 ; Luo, et al., 1998 ). Regulation is mediated by the caspase-8 inhibitory proteins FLIP (flice inhibitory protein) that exists in two splicing variants, the long form FLIPL and the short one FLIPS. They contain two DEDs, but they do not have proteolytic activity (Thome, et al., 1997 ; Thome and Tschopp, 2001 ). Both cellular isoforms possess significant anti-apoptotic activity once they are part of the DISC. Overexpressed FLIP can block autoproteolytic activation of procaspase-8 by binding competitively to FADD (Thome and Tschopp, 2001 ).

2.2.2 The mitochondrial death pathway

The intrinsic pathway is regulated by proteins of the Bcl-2 family and is therefore referred to as the mitochondrial pathway, because the mitochondria play a central role (figure 2). Several stimuli can lead to activation of the mitochondrial cell death pathway such as cytotoxic drugs, heat shock, ionising, DNA damage and growth factor withdrawal. These stimuli trigger Bax and Bak activation, which subsequently mediate the permeabilization of the outer mitochondrial membrane and release of distinct proteins. Among them are cytochrome c, Smac/Diablo, apoptosis inducing factor (AIF), endonuclease G and Omi/HtrA2 (van Loo, et al., 2002 ). Released cytochrome c mediates the formation of a protein complex called apoptosome, which consists of Apaf-1 (apoptosis protease activating factor 1), initiator caspase-9 and (d)ATP (Li, et al., 1997 ; Liu, et al., 1996 ; Saleh, et al., 1999 ). Binding of Apaf-1 to cytochrome c increases the affinity of Apaf-1 to ATP, which is providing the energy for apoptosome formation (Hu, et al., 1999 ). During apoptosome formation procaspase-9 is activated and, in turn, causes the activation of the effector caspase-3 and caspase-7 (Rodriguez and Lazebnik, 1999 ). Further, cytochrome c release is accompanied by a mitochondrial permeability shift, acidification of the cytosol and ion fluxes (Daniel, et al., 2003 ). Released endonuclease G translocates to the nucleus, where it digests DNA caspase-independently (van Gurp, et al., 2003 ). Moreover, activation of released Smac/Diablo results in dimerization. Smac/Diablo contributes to caspase activation by restraining inhibitor of apoptosis proteins (IAPs). As implied by the name, these proteins inhibit activation of pro-caspases. In his task Smac/Diablo is supported by Omi/HtrA2 (van Gurp, et al., 2003 ). The death receptor signalling pathway is connected to the mitochondrial cell death pathway by the BH3-only protein Bid when cleaved by caspase-8 (Gross, et al., 1999 ). This truncated Bid may bind to Bax or Bak thereby inducing a conformational change in these proteins and the insertion of them into the outer mitochondrial membrane. It also sequesters anti-apoptotic Bcl-2 family proteins. This is followed by the permeabilization of the mitochondrial membrane and consequently the amplification of the mitochondrial pathway. Despite of this connection the pathways function mostly independently form each other (Bouchon, et al., 2000 ). The anti-apoptotic proteins of the Bcl-2 family can block the intrinsic pathway. These proteins can prevent cytochrome c release and caspase activation (Kluck, et al., 1997 ).

Figure 2: The mitochondrial and death receptor apoptotic pathway

The mitochondrial (intrinsic) apoptotic signalling pathway is initiated by cell damaging events, upon which pro-apoptotic members of the Bcl-2 family are activated and translocate to the mitochondria to neutralize anti-apoptotic proteins. Permeabilization of the mitochondria causes the release of cytochrome c. Released cytochrome c associates with Apaf-1 and procaspase-9 in the presence of dATP to form the apoptosome. Activated caspase-9 triggers a caspase cascade leading to apoptosis.
The death receptor (extrinsic) pathway is activated when ligands of the TNF family bind to their receptors on the cell surface. Binding of the ligand induces trimerization of the receptor and recruitment of the adaptor protein FADD and caspase-8. Within this complex, caspase-8 is activated and in turn cleaves and activates casapse-3. The two pathways are mostly independent, but in type II cells the two pathways can be linked via cleavage of Bid by caspase-8, caspases-3 or -10. Truncated Bid activates the mitochondrial apoptotic pathway.

2.2.3 The endoplasmic reticulum pathway


The endoplasmic reticulum is responsible for the maintenance of the calcium homeostasis and is also the major intracellular calcium storage. The uptake of Ca2+ into the lumen of the ER is managed by energy-dependent SERCA (sarcoplasmic/endoplasmic reticulum Ca2+-ATPase). The release is handled by IP3 (Inositol 1,4,5-Triphosphat (IP3)-regulated receptors or ryanodin (RyR) Ca2+ receptors (Berridge, et al., 2000 ). Furthermore, the ER is the main compartment for protein synthesis, folding, targeting and trafficking. The ER contains numerous chaperone proteins, a high level of calcium and an oxidative environment to carry out these functions efficiently (Rao, et al., 2004 ). Changes in Ca2+ levels or accumulation and aggregation of un- or misfolded proteins lead to ER stress, which is gauged by the ER stress sensors IRE1, PERK and ATF6. To restore normal ER function, the unfolded protein response (UPR) is initiated (Szegezdi, et al., 2006 ). Excessive ER stress forces the unfolded protein response to activate diverse pathways that eventually lead to apoptosis (Ferri and Kroemer, 2001 ). Also, prolonged ER stress is involved in the pathogenesis of some neurodegenerative disorders that feature misfolded proteins (Rao, et al., 2002 ). ER stress can also be elicited by several agents including tunicamycin a specific N-glycosylation inhibitor, Brefeldin A, an inhibitor of the protein transport from ER to Golgi and thapsigargin, which blocks Ca2+ uptake by inhibiting the SERCA (Lee and East, 2001 ). The answer to these stresses is the upregulation of ER chaperons, including the glucose regulated protein GRP78 also referred to as BiP (immunoglobulin heavy chain-binding protein) and the transcription factor CHOP (C/EBP homologous protein). They have both anti-apoptotic features and regulate the ER stress sensors (Lee, 2005 ). To relief the ER stress, GRP78/BiP accelerates protein folding in the ER lumen (Momoi, 2004 ). CHOP sensitizes cells to ER stress by downregulation of Bcl-2 and activation of GADD34 (protein phosphatase 1 (PP1)-interacting protein) and ERO1alpha, an ER oxidase (Li, et al., 2006 ). In summary, these proteins facilitate protein folding and prevent aggregation.

In mice, chaperones present the signal to capase-12, which is on the cytoplasmic site of the ER membrane (Szegezdi, et al., 2003 ). The role of caspase-12 in human is not known since it is expressed in a truncated form and is not functional (Nakagawa, et al., 2000 ). Caspase-4 was proposed to fulfil this function in human, but it is still under debate. It was shown to be localized at the ER and be activated by ER stress. Its mechanism though, is still not fully understood (Hitomi, et al., 2004 ). Capase-12 is activated upon ER stress, but neither upon death receptor nor mitochondrial apoptotic signals (Nakagawa, et al., 2000 ). Apart from the discussion about activation of murine caspases-12 upon ER stress, there is also the possible involvement of caspases-3 being involved in Ca2+ homoeostasis. Caspase-3 cleaves IP3 receptors in Jurkat cells and thus decreases the activity of the IP3 receptor under apoptotic circumstances (Hirota, et al., 1999 ).

2.3 Calcium mobilization

Calcium from the ER is not just an ER stress signal it could also be involved in the regulation of Ca2+ influx, ER protein folding and chaperone interaction, gene expression and regulation of nuclear pore opening. Calcium elevations may be mediated by ER proteins, BAP31 (Breckenridge, et al., 2003 ), or tBid, which enhance transmission of IP3 mediated Ca2+ signals to mitochondria (Csordas, et al., 2002 ). Ca2+ released from the ER is taken up by the mitochondria and accumulates in the matrix and this may reset in cytochrome c release. Ca2+ traverses the outer mitochondrial membrane primarily through the voltage dependent anion channel (VDAC). Ca2+ interacts with cyclophilin D, a component of the permeability transition pores, to induce their opening of the permeability transion pore (Basso, et al., 2005 ). Calcium overload of the mitochondria has been linked to the recruitment of Drp1, that has been implicated in mitochondrial fragmentation (Breckenridge, et al., 2003 ; Rudner, et al., 2002 ). Consequences of the pore opening are the loss of membrane potential and re-release of Ca2+. When the elevated cytoplasmic Ca2+ level persists, the permeability transion pore stays open and allows accumulation of solutes in the mitochondrial matrix. The entry of solutes leads to the extension of the mitochondrial matrix and to rupture of the outer mitochondrial membrane, thereby to releasing the intermembrane space content (Green and Kroemer, 2004 ). Changes in Ca2+ levels appear to be regulated by both pro-and anti-apoptotic members of the Bcl-2 family. Bax, Bak and Bcl-2 are localized at the outer mitochondrial as well as at the ER membrane. At both membranes they regulate fluxes of calcium, meaning the release from the ER and the subsequent uptake by the mitochondria, while Bax and Bak support calcium release and Bcl-2 and Bcl-xL antagonize it (Garrido, et al., 2006 ). Bcl-2 was shown, however, to diminish calcium levels in the ER and to stimulate accelerated re-uptake into the ER or into the mitochondria. In this way, the Ca2+ concentration does not cross an intracellular threshold after an apoptotic insult. It might be that blockade of Bcl-2 and Bcl-xL interaction with IP3 (inositol 1,4,5-trisphosphate) receptors helps the effect of Bax and Bak (Chen, et al., 2004 ; White, et al., 2005 ). IP3 receptors are found on the ER membrane, where they regulate the mobilization of Ca2+ stores (Berridge, 2005 ). Recent studies point to a link between the function of cytochrome c and IP3 triggered Ca2+ mobilization (Boehning, et al., 2005 ). Overexpression of Bcl-2 partially prevents rise of cytoplasmic Ca2+ levels upon exposure to exogenous ceramide, staurosporin, thapsigargin or growth factor depletion (Distelhorst and McCormick, 1996 ). Moreover, Bcl-2 targeted to the ER (Zhu, et al., 1996 ) averted apoptosis stimulated by ceramide, irradiation, thapsigargin and the upregulation of Bax and Bad (Annis, et al., 2001 ; Rudner, et al., 2001 ; Thomenius, et al., 2003 ; Wang and Spector, 2001 ). A model for calcium release from the ER is given in figure 3.


Figure 3: ER stress and Calcium release

Apoptotic signals targeting the ER may induce Ca2+ release to regulate mitochondrial activation by opening its permeability transition pore (PTP).

2.4 Activation of the mitochondria

Mitochondria produce ATP, maintain the equilibrium in-between ions and regulate apoptosis. As already mentioned (see 1.2.2), upon apoptotic signals mitochondrial proteins are release form the intermembrane space into the cytosol. Among these proteins are cytochrome c  (Liu, et al., 1996 ), Smac/Diablo (Adrain, et al., 2001 ), Omi/HtrA2 (Suzuki, et al., 2001 ), AIF (Cande, et al., 2002 ) and endonuclease G (Li, et al., 2001 ). In order to release these proteins the outer mitochondrial membrane needs to be permeabilized.

2.4.1 Mitochondrial outer membrane permeabilization

The mechanisms of mitochondrial outer membrane permeabilization have not yet been illuminated in all details. This leaves room for several models. The two most popular ones are the permeabilization of the inner membrane by so far enigmatic mechanisms, and the insertion of the multi-domain Bcl-2 proteins Bax and Bak into the outer mitochondrial membrane.


Permeabilization of the inner mitochondrial membrane is achieved by formation of a channel spanning through both membranes of the mitochondria (Crompton, 2000 ). Each of these channels, so-called permeability transition pores, has been suggested to consists of the voltage-dependent anion channel (VDAC), an outer membrane protein, the adenine nucleotide translocator (ANT), in the inner membrane, and cyclophilin D in the matrix (Green and Kroemer, 2004 ). Increased cytosolic Ca2+ or reactive oxygen species promote the opening of these more or less functionally defined permeability transition pores. Once the pore is open it allows the osmotic driven influx of water, ions and other small molecules into the matrix. These activities leading to swelling of the mitochondria, and are termed mitochondrial permeability transition (Zamzami and Kroemer, 2001 ). Now, rupture of the outer mitochondrial membrane is induced and eventually the outer membrane is permeabilized, spilling the mitochondrial content into the cytosol (Armstrong, 2006 ). This goes along with the loss of the mitochondrial membrane potential. VDAC, the core of the permeability transition pore, is regulated by Bax, Bak and Bcl-2 (Chandra, et al., 2005 ). In another model, pro-apoptotic Bax and Bak form pores in the outer mitochondrial membrane by oligomerization. But first, BH3-only proteins induce an allosteric conformational change in Bax and Bak. Conformational change exposes the N-terminal region of Bax and is needed for integration into the membrane (Korsmeyer, et al., 2000 ). Now, Bax is able to translocate to the mitochondria and stably insert into the membrane. They form large multimeric pores in the outer mitochondrial membrane. Intermembrane space proteins can escape through these pores into the cytosol. In this model, the inner membrane and the matrix would not be affected and remain intact. The crucial role of Bax and Bak in mitochondrial membrane permeabilization becomes evident by the fact that double knockout cells do not undergo outer membrane permeabilization in response to apoptotic stimuli (Armstrong, 2006 ). The pore forming model embarked, when structural similarities of pore forming bacterial toxins and multi-domain proteins were found (Muchmore, et al., 1996 ). But not only Bax and Bak are included in this model. Truncated Bid is also believed to form a pore into the outer mitochondrial membrane by oligomerization. Therefore it might function like Bax and Bak regarding permeabilization of the mitochondria (Henry-Mowatt, et al., 2004 ). While Bax and Bak induce the release of apoptotic factors from the mitochondria, Bcl-2 and Bcl-xL block outer membrane permeabilization and thereby avoid the release of these factors (Daniel, et al., 2003 ). Due to interactions between pro- and anti-apoptotic proteins, the binding affinity and the stochiometry in-between them is a crucial factor in induction of apoptosis. Release of cytochrome c

It is not exactly known how cytochrome c is translocated through the mitochondrial outer membrane, but there is evidence that the Bcl-2 family members are the main regulators of cytochrome c release. Bcl-xL, Bax and Bid share similarities in their transmembrane domains, which they have in common with bacterial colicin toxins and diphtheria toxins. This led to the assumption that Bcl-2 proteins are also able to form pores. On one hand, there is the model already described above, implying that Bax forms pores upon oligomerization and integration into the outer mitochondrial membrane. On the other hand it has been suggesting that Bax and Bak interact with VDAC and thereby stimulates the opening of this channel, although the mechanism is not known (Marzo, et al., 1998 ; Saito, et al., 2000 ). This model is supported by the finding that inactive Bak is associated to VDAC, but can be freed by anti-apoptotic proteins that bind to VDAC2 (Chandra, et al., 2005 ). Other reports doubt the pore formation of Bax and Bak, although confirming their ability to oligomerize. Additionally to these two models, it was shown that the fragmentation of the mitochondrial network during apoptosis is an essential event for mitochondrial breakdown. Moreover, it was proposed that Bax plays a critical role in mitochondrial fragmentation (Karbowski, et al., 2002 ). Also, it was shown that the truncated form of Bid can release apoptotic factors from the mitochondria by forming pores at the outer membrane, when the extrinsic apoptotic pathway is activated (Zamzami, et al., 2000 ). Further, it was reported that the pro-apoptotic activity of BH3-only proteins induce cytochrome c release without permeability transition (Shimizu and Tsujimoto, 2000 ).

There are arguments for and against the models mentioned. Most likely, different apoptotic stimuli induce distinct mechanisms for permeabilization of the outer mitochondrial membrane. Release of pro-apoptotic factors


Not only cytochrome c is set free upon disruption of the outer mitochondrial membrane. Also other apoptosis-inducing factors are liberated from the mitochondria.

AIF (apoptosis inducing factor) is a flavoprotein, which translocates to the nucleus and is participating in chromatin condensation and DNA-fragmentation in a caspase-dependent manner (Daugas, et al., 2000 ; Lorenzo, et al., 1999 ; Susin, et al., 1999 ). Its redox activity is not important for its apoptotic effect (Miramar, et al., 2001 ). These data are, however, discussed in a highly controversial fashion.

Endonuclease G is a DNA degrading enzyme important for DNA-repair in mitochondria. It is released from the matrix and migrates to the nucleus and degrades DNA in a caspase independent way (van Loo, et al., 2002 ).


Omi/HtrA2 (high temperature requirement A2) is a serine protease, which acts as chaperone for correction of misfolded proteins or their degradation, but also inhibits IAPs (inhibitor of apoptosis proteins). Its mitochondrial signal sequence targets it to the intermembrane space, where this targeting-signal is being cleaved off and, the remaining protein is transformed into a mature protease. In the cytosol, Omi/HtrA2 facilitates caspase-dependent apoptosis, catalytic cleavage of IAPs, and permeabilization of the outer mitochondrial membrane (Faccio, et al., 2000 ; Hegde, et al., 2002 ; Suzuki, et al., 2004 ; Verhagen, et al., 2002 ; Yang, et al., 2003 ). Generally speaking, this protein appears to exert similar functions as Smac/Diablo (second mitochondria derived activator of caspases / Direct IAP binding protein with low pI). In a caspase-independent and Apaf1-independent fashion, Omi/HtrA2 is able to induce cell death using its serine protease activity (Hegde, et al., 2002 ; Suzuki, et al., 2004 ). Structural analyses showed that cytosolic Smac/Diablo is a symmetric homo-dimer (Chai, et al., 2000 ). After the release from the mitochondria Smac/Diablo binds IAPs in the cytosol. In this way, Smac/Diablo regulates IAPs by inhibiting their activity (Vaux and Silke, 2003 ). Therefore, it amplifies caspase activity, since IAPs inhibit the activity of processed caspases (Du, et al., 2000 ).

2.5 Crosstalk between ER and mitochondria

The mechanisms and the part of crosstalk between the mitochondria and the endoplasmic reticulum are not entirely illuminated. But it seems that cytochrome c induced apoptosis activated by ER - mitochondria crosstalk is important for ER stress mediated cell death (Momoi, 2004 ). Moreover, the main signal in ER-mitochondria crosstalk is thought to be calcium. After activation of the death receptor pathway and consequent activation of caspase-8, BAP31 (Bcl-2-associated protein 31) is cleaved to a p20 fragment. BAP31 is an integral ER membrane protein that seems to be a mediator of crosstalk between the two organelles and it has pro-apoptotic capacities. It is cleaved and activated by a unique isoform of caspase-8 (Breckenridge, et al., 2002 ). The p20 cleavage product of BAP31 causes the release of Ca2+ from the ER. Liberated Ca2+ is taken up by the mitochondria inducing the recruitment of Drp1. Drp1 mediates the scission of the outer mitochondrial membrane, resulting in dramatic fragmentation and fission of the mitochondrial network and cytochrome c release (Breckenridge, et al., 2003 ). Calcium signals from the ER regulate the opening of the permeability transition pore. Absorbed Ca2+ in the matrix causes at a certain level the opening of the mitochondrial transion pore, which leads to loss of the mitochondrial membrane potential and hence to the release of cytochrome c and apoptotic factors. This process is followed by translocation of cytochrome c to the ER, where it interacts with IP3 receptors to induce a positive feedback loop (Boehning, et al., 2003 ). Recent studies of the ER-mitochondrial communication presented some evidence for regulation of the IP3 receptor-mediated Ca2+ release by both pro- and anti-apoptotic proteins. Pro-apoptotic factors have been described to facilitate the mobilization of ER Ca2+ thereby strengthening the calcium signal propagation to the mitochondria. This may lead to Ca2+-dependent mitochondrial membrane permeabilization (Hajnoczky, et al., 2006 ). ER membrane proteins also interact with Bcl-2 proteins and thereby influence the apoptotic events. Further, it was reported that Bax and Bak induced cytochrome c release by interacting with VDAC. With its cytosolic domain BAP31 can interact with procaspase-8, Bcl-2 and Bcl-xL, but the BH3-only protein Spike can block the formation of the BAP31 – Bcl-xL complex (Mund, et al., 2003 ). Another BH3-only protein, Bad, might also be involved in ER activation. It is dephosphorylated by calcineurin, a Ca2+/calmodulin dependent protein phosphatase and induces the release of pro-apoptotic factors from the mitochondria (Wang, et al., 1999 ).

2.6 Caspases

Caspases (cysteine-aspartate proteases) are key proteins in the apoptotic process. They specifically cleave their substrates following an aspartate residue (Earnshaw, et al., 1999 ). Caspases are synthesized as catalytically inactive pro-caspases i.e. zymogens. They need an apoptotic stimulus to be activated by either proteolytic cleavage by other caspases or by autocatalysis to become an active enzyme (Cohen, 1997 ; Shi, 2002 ). The caspase family comprises in human 11 members (13 in mammalian). Based on observations made in C. elegans, the first member of this family was described in 1992 as the CED-3 homologue interleukin-1β-converting enzyme (ICE, later caspase-1) (Cerretti, et al., 1992 ; Thornberry, et al., 1992 ). According to there structure, several other caspases were identified. The amino acid sequence QAC(R/Q/G)G, found in the active site of caspases is highly conserved (Alnemri, et al., 1996 ). The substrate recognition motif varies among caspases, making them target-specific substrates (Thornberry, 1997 ). Nevertheless, all caspases cleave at the peptide bond C-terminal of aspartate residues. Zymogens consist of 3 domains, one N-terminal pro-domain, followed by one large (~20kDa) and one small subunit (~10kDa), which are separated by linkers. After proteolytic cleavage both subunits form heterodimers, made of 2 subunits of each proform, with a catalytic site. Caspases can undergo autoactivation or be activated by other caspases.

2.6.1 Initiator and executioner caspases


Caspases can be divided into two groups (figure 4), the initiator caspases (caspase-2, -8, -9, -10) and the executioner caspases (caspase-3, -6, -7) (Thornberry and Lazebnik, 1998 ). These two groups differ in the length and structure of their prodomain. Initiator caspases carry a long prodomain with a DED and a CARD (caspase recruitment domain) domain and are the link between cell signalling and apoptosis. The CARD domain is binding the adapter molecules, the DED cares for the hydrophobic binding of the adapter protein (Earnshaw, et al., 1999 ). Caspases-8 is the main mediator in the extrinsic cell death pathway induced by the TNF family members. Binding of death ligands (CD95, Fas) leads to trimerization of the death receptors and aggregation of death domains (DD). Together with FADD (Fas associated death domain) which recruits procaspase-8 the death-inducing signalling complex (DISC) is formed, where caspase-8 is activated  (Muzio, et al., 1996 ). Also in the intrinsic cell death pathway the formation of a complex, the apoptosome, consisting of procaspase-9 and its adapter protein Apaf-1 is required. The pro-domain of the initiator caspase interacts with the adaptor protein leading to dimerization and finally to autocatalytic activation (Earnshaw, et al., 1999 ). Autoproteolytic cleavage yields in two small and two large subunits forming an active caspase as a homodimer of two heterodimers (Nicholson, 1999 ).

The effector caspases share a short pro-domain and act downstream of the apoptotic pathway. Initially, they are cleaved by the initiator caspases at a specific asparagin residue, followed by autoproteolytic cleavage. Active effector caspases cleave substrates escalating the death signal and executing apoptosis (Savill and Fadok, 2000 ). Caspase-3 is the main effector caspase that cleaves the majority of the cellular substrates in apoptotic cells. Caspase-7 is highly similar to caspase-3 and has comparable substrate specificity (Degterev, et al., 2003 ). They both amplify mitochondrial caspase activation signalling (Lakhani, et al., 2006 ).

The position of caspases in the pathway is not just defined by their structure and length, but also by their subcellular localization (Zhivotovsky, et al., 1999 ).


Figure 4: The caspase family

There are three major groups of caspases, group 1: initiator caspases, group 2: effector caspases and group 3: inflammatory caspases. Caspases are synthesized as pro-caspases with an N-terminal pro-domain. The active caspase is a heterotetramer of two large and two small subunits.

2.6.2 Substrates of the caspases

Caspases target substrates, which are either maintaining the metabolism of the cells or proteins involved in the apoptotic machinery (Fischer, et al., 2003 ). Caspases cleave different substrates in a different manner. However, processing of the substrate can be directly connected to the morphological changes during apoptosis. In the final steps of apoptosis the cell loses its shape. This is indicated by caspases that degrade structural proteins, which maintain e.g. the cytoskeleton (Brockstedt, et al., 1998 ). Among them are actin filaments and nuclear lamins, which are cleaved by caspase-6 (Orth, et al., 1996 ). Caspases cleave proteins involved in cell cycle regulation and DNA repair (Cohen, 1997 ). In healthy cells, ICAD (inhibitor of caspase-activated DNase) is in a complex with the inhibitor DFF45. Activation requires cleavage of the inhibitor protein by caspase-3 resulting in active CAD-nuclease which then translocates to the nucleus (Enari, et al., 1998 ; Sakahira, et al., 1998 ). PARP is a nuclear enzyme and is activated upon cleavage by caspase-3 at an early stage of apoptosis. It is cleaved into a small, N-terminal fragment and a large, C-terminal (catalytical) subunit abrogating its DNA repair and ATPase activity. PARP catalyzes the transfer of ADP-ribose polymers to other nuclear proteins, which are in involved in DNA repair and stabilization. If caspases are not active, PARP is excessively consuming energy during the apoptotic process and cells end up dying by necrosis (Kaufmann, et al., 1993 ). Thus, it seems that PARP may act as a molecular switch between apoptosis and necrosis (Fischer, et al., 2003 ).

2.7 The Bcl-2 family

The members of the Bcl-2 (B-cell lymphoma (gene) 2) family are important regulators of apoptosis. They transmit external survival or death signals inside the cell. The Bcl-2 family members are grouped into two subfamilies of anti-apoptotic and pro-apoptotic proteins (figure 5). They all share at least one of the four conserved α-helical so-called Bcl-2 homology (BH) domains (BH1, BH2, BH3, BH4). These domains are conserved motifs (Cory and Adams, 2002 ). The pro-apoptotic proteins involve a subgroup, which comprises the BH3-only proteins (Huang, et al., 2002 ). The anti-apoptotic members share all four BH domains, whereas the pro-apoptotic members do not a possess BH4. Instead, they are subdivided into the multi-domain group carrying BH1-3 and the BH3-only group. BH1 to BH3 are responsible for binding by forming a hydrophobic pocket. They give the proteins the ability to form homo- and heterodimers (Borner, 2003 ). Anti-apoptotic proteins prevent permeabilization of the mitochondria, which is on the other hand promoted by the multi-domain group of the pro-apoptotic proteins. Some of the anti-apoptotic proteins interact directly with members of the pro-apoptotic group (Gross, et al., 1999 ; Rosse, et al., 1998 ). The balance in-between anti-apoptotic and pro-apoptotic proteins decide about life or death of the cell.


Members of the Bcl-2 family can regulate apoptosis in different manner: by forming protein channels in mitochondrial membranes or by changing the membrane structure by interactions with lipids. They regulated ion fluxes and release of other proteins by the permeabilization of the mitochondrial membrane and the endoplasmic reticulum (Sharpe, et al., 2004 ).

Figure 5: The Bcl-2 family

Bcl-2 family members possess at least one of four BH (Bcl-2 homology) domains and are grouped according to their ability to inhibit or activate cell death. The pro-apoptotic proteins bind with their BH3 domain to the hydrophobic pocket formed by the BH1-BH3 domains of the anti-apoptotic proteins. The C-terminal transmembrane (TM) region is a hydrophobic, single membrane spanning alpha helix that mediates localization to intracellular membranes.
Anti-apoptotic members of the Bcl-2 family
Pro-apoptotic members of the Bcl-2 family, subgrouped into multi-domain proteins and BH3-only proteins.

2.7.1 The anti-apoptotic proteins

The anti-apoptotic members of the Bcl-2 family share three or four Bcl-2 homology (BH) domains, which are essential for apoptosis (figure 5). Proteins such as Bcl-2, Bcl-xL  (Boise, et al., 1993 ), Bcl-w (Gibson, et al., 1996 ), A1/Bfl-1 (Choi, et al., 1995 ; Lin, et al., 1996 ) and Mcl-1 (Kozopas, et al., 1993 ) belong to this group. Bcl-2 and Bcl-xL possess all four BH domains, where as other anti-apoptotic family members such as Mcl-1 carry only BH1 and BH2. The BH-domain serves as binding site for interaction with other proteins. BH1, BH2 and BH3 form a hydrophobic groove, which is stabilized by the N-terminal BH4-domain (Huang, et al., 2002 ) . This groove represents the binding site for the BH3 α-helix of the BH3-only proteins (Sattler, et al., 1997 ). This interaction neutralizes the anti-apoptotic family members. In healthy cells, the anti-apoptotic proteins primarily might hold back Bax and Bak from disturbing intracellular membranes, especially the outer mitochondrial membrane (Liu, et al., 2003 ). Anti-apoptotic proteins can become pro-apoptotic after N-terminal proteolytic cleavage of the BH4 domain (Cheng, et al., 1997 ). Furthermore, their anti-apoptotic effect seems to depend solely on their BH4 domain, since its deletion led to loss of their activity (Huang, et al., 1998 ). With their C-terminal tail the anti-apoptotic proteins such as Bcl-2, Bcl-w, Bcl-xL and Mcl-1 are able to insert themselves into sub cellular membranes, including the outer mitochondrial membrane, the endoplasmic reticulum or the nuclear envelope (Cory and Adams, 2002 ; Duriez, et al., 2000 ). High expression of Bcl-2 and Bcl-xL is found in several types of cancer. Bcl-2 is an integral membrane protein, even in healthy cells, whereas Bcl-w and Bcl-xL only incorporate into membranes after death signals. Bcl-2 and Bcl-xL inhibit apoptosis by insertion into the outer mitochondrial membrane, they may maintain the membrane integrity (Sharpe, et al., 2004 ). Bcl-xL was reported to inhibit cytochrome c release (Daniel, et al., 2003 ). There is some discussion about how and if the anti-apoptotic members are involved in mitochondrial membrane permeabilization. Structural features seem to make it likely that they have pore forming abilities, since they show similarities with several bacterial toxins, colchicines A, E1 and diphtheria toxin (Antonsson, et al., 1997 ; Minn, et al., 1997 ; Schendel, et al., 1997 ; Schlesinger, et al., 1997 ). Others say that VDAC is involved in the regulation of the anti-apoptotic group members. VDAC is located at the outer mitochondrial membrane. Interaction of VDAC and the anti-apoptotic proteins causes the exchange of anions between cytosol and the intermembrane space of the mitochondria (Heiden, et al., 2000 ). Moreover, Bcl-2 and Bcl-xL keep pro-apoptotic BH3-only proteins inactive by binding to them in a complex.


Bcl-2, Bcl-xL and Bcl-w strongly inhibit apoptosis in response to many, but not all, cytotoxic insults. Every cell requires protection by at least one anti-apoptotic family member, as they are chief guardians of tissue homeostasis (Cory and Adams, 2002 ). Efficient apoptosis calls for neutralization of several anti-apoptotic proteins and imply that not all the anti-apoptotic proteins may have the same function (Adams and Cory, 2007 ).

2.7.2 The pro-apoptotic proteins

The pro-apoptotic family members are subdivided in multi-domain proteins and BH3-only proteins. Pro-apoptotic proteins Bax (bcl-2 associated protein X) (Oltvai, et al., 1993 ), Bak (Chittenden, et al., 1995 ) and Bok (Hsu, et al., 1997 ) contain BH1-BH3 but lack BH4 and are therefore also called multi-domain proteins. In healthy cells Bak is held in an inactive monomeric state in the outer mitochondrial membrane through its association with VDAC2 (Cheng, et al., 2003 ). Bax is a monomeric cytosolic protein, inactive through interactions with several proteins (Breckenridge and Xue, 2004 ). Upon activation they both undergo a conformational change. First Bax has to translocate to the mitochondria (Hsu, et al., 1997 ; Wolter, et al., 1997 ) before inserting into the outer membrane (Goping, et al., 1998 ), whereas Bak stays in the mitochondrial membrane, where they both oligomerize (Antonsson, et al., 1997 ). Their carboxyl-terminus is essential for targeting to mitochondria. One model is that Bax and Bak form pores to release pro-apoptotic factors from the mitochondria. The other hypothesis says that they associate with parts of the permeability transition pore (Cory and Adams, 2002 ). But both models agree on permeabilization of the outer mitochondrial membrane by Bax and Bak. And both multi-domain proteins stimulate the release of cytochrome c form the mitochondria using the voltage-dependent anion channel (VDAC) (Daniel, et al., 2003 ). Bax and Bak are indispensable for apoptotic signalling, since double knockout experiments showed that development and proliferation is not possible without these proteins, either Bax or Bak has to be present in order to fulfil the assignment properly (Lindsten, et al., 2000 ). Cells without Bax and Bak are resistant to most apoptotic stimuli (Cheng, et al., 2001 ). Bax and Bak mainly regulate the intrinsic pathway, being localized at the mitochondria, but they also operate at the ER (Scorrano, et al., 2003 ; Wei, et al., 2001 ). Currently two models are discussed: 1. Direct activation, holds that certain BH3-only proteins, termed activators, Bim and tBid, can bind to Bax and Bak directly and promote their activation, in this model, the remaining BH-only proteins, termed sensitizers, bind only to the pro-survival proteins and purportedly act by displacing any bound Bim or tBid, allowing them to directly activate Bax and Bak. 2. Indirect activation, on the other hand, suggests that all the BH3-only proteins engage only their pro-survival relatives and act by preventing them from countering Bax or Bak activation, on this model, Bim and tBid are potent inducers of apoptosis simply because they can engage all the pro-survival proteins (Willis and Adams, 2005 ; Willis, et al., 2007 ). In the extrinsic pathway, truncated Bid initiates oligomerization of Bax or Bak leading to cytochrome c release from mitochondria (Wei, et al., 2000 ), since cells lacking both of these proteins did not undergo apoptosis after Bid activation. But if at least one of these two multi-domain proteins is present, cell death will be initiated (Wei, et al., 2001 ). Interestingly, high levels of anti-apoptotic proteins block Bax oligomerization and pore formation, but no Bcl-2 - Bax complex could be detected (Mikhailov, et al., 2001 ). The BH3-only group comprise quite a lot of members, such as Bid (Wang, et al., 1996 ), Noxa (Oda, et al., 2000 ), Puma (Nakano and Vousden, 2001 ; Yu, et al., 2001 ), Bim (O'Connor, et al., 1998 ), Nbk (Boyd, et al., 1995 ; Han, et al., 1996 ), Bad (Yang, et al., 1995 ) and some more (figure 5). The BH3 motif is a short sequence of nine amino acids and the only commonality of these proteins. The BH3-only proteins function as sensors of apoptotic stimuli and are charged to trigger apoptosis (Huang and Strasser, 2000 ). The crucial decision on life or death seems to be fought on these membranes, although most family members are recruited to these sites upon apoptotic signal (Adams and Cory, 2007 ). In healthy cells the BH3-only proteins are in an inactive state (Huang and Strasser, 2000 ). Only upon apoptotic signal they are activated and perform their duty. Distinct apoptotic stimuli activate different proteins of this group, which then deliver the death signal to the mitochondria by engaging Bax/Bak or Bcl-2/Bcl-xL  (Puthalakath and Strasser, 2002 ). For example, Noxa and Puma are induced by transcription with p53 as their transcription factor. P53 is activated upon DNA-damage, irradiation and cytotoxic drugs (Han, et al., 2001 ; Lakin and Jackson, 1999 ). Other proteins of the BH3-only group are activated by posttranslational modifications, e.g. Bad, which is dephosphorylated (Harada, et al., 1999 ). Bid is activated by proteolytic cleavage. In its inactive form Bid is a cytosolic protein, but once it is cleaved, truncated Bid (tBid), it translocates to the mitochondria (Li, et al., 1998 ). There, it stimulates the permeabilization of the mitochondrial membrane and therefore the release of apoptotic factors. Expression of several BH3-only proteins, such as Bim, Bad and Bid in Bax/Bak double knockout cells could not induce apoptosis, suggesting that BH3-only proteins require Bax or Bak to mediate apoptotic signals (Zong, et al., 2001 ). Most members have additionally a hydrophobic sequence at the C-terminus, which helps them to integrate into organelle membranes.

Bid is activated upon cleavage by caspase-8 after initiation of the extrinsic apoptotic pathway (Li, et al., 1998 ). Normally, Bid is localized as an inactive form in the cytosol. After activation, truncated Bid (tBid) translocates to the mitochondria to enhance the apoptotic signal (Gross, et al., 1999 ). Additionally, Bid activates Bax by inducing a conformational change, followed by the translocation of Bax to the mitochondria (Eskes, et al., 1998 ). Truncated Bid seems to amplify the perturbation of the mitochondria by forming homotrimers in the membrane (Cory and Adams, 2002 ). Apart from amplification of the mitochondrial pathway, tBid can also convey death signals form other organelles. Bid has an unique place in apoptosis, since it connects the extrinsic and intrinsic pathway (Daniel, 2000 ). The BH3-only protein Nbk/Bik localizes at the ER and not at the mitochondria, it mediates the activation of the mitochondria in a Bax-dependent manner (Gillissen, et al., 2003 ). Puma and Noxa and Hrk/DP5, are controlled primarily at the transcriptional level. Puma and Noxa are expressed upon p53 activation and they were shown to cause outer membrane permeabilization. Co-immunoprecipitation studies showed that Noxa binds to Bcl-2 and Bcl-xL, but not to Bax (Oda, et al., 2000 ).


If the balance between anti-apoptotic and pro-apoptotic Bcl-2 family members is shifted in the favour of the pro-apoptotic proteins, they bind to and occupy the anti-apoptotic proteins, thereby liberating Bax and Bak (Gogvadze and Orrenius, 2006 ). Bim

This BH3-only protein is subject to this thesis and will therefore be introduced in more details. While screening for proteins that bind to Bcl-2, a novel protein, named Bim (Bcl-2 interacting mediator), was discovered (O'Connor, et al., 1998 ). The only common feature with other known proteins was its BH3 domain, placing it into the BH3-only group of the Bcl-2 family. Bim plays a major role in embryogenesis, in the control of haematopoietic cell death and as a barrier against autoimmunity (Bouillet, et al., 1999 ). Originally three different Bim splicing variants were found (O'Connor, et al., 1998 ), but there are also additional splicing variants, all encoded by the same gene (figure 6). These various isoforms differ in size and apoptotic strength. The best characterized are the three main isoforms BimEL, BimL and BimS is constitutively pro-apoptotic and appears to be the most toxic, whereas BimEL and BimL are held in an inactive form in healthy cells. Binding to the dynein motor complex light chain LC8 (DLC1) of the microtubule sequesters BimEL and BimL. In response to cytokine removal, calcium flux, microtubule perturbation (by taxol a microtubule polymerizing drug) or cellular damage by UV-irradiation, BimEL and BimL are released from the microtubule in a complex with LC8 (Puthalakath, et al., 1999 ). The LC8-Bim complex can bind to and antagonize the anti-apoptotic proteins (Borner, 2003 ) and is recruited to the mitochondria. There are controversial reports on BimS about its function and expression. Tissue distribution studies for Bim demonstrated expression of BimEL and BimL in haematopoietic, epithelial, neuronal and germ cells, while BimS was not detected by Immunoprecipitation (IP) or Western blot analysis. It was speculated that BimS is only expressed in specific cells that need to be extinguished quickly or that BimS is not expressed under physiological conditions (Bouillet, et al., 1999 ; Puthalakath, et al., 1999 ). On the other hand, there are data showing expression of all three isoforms at similar levels in DLD-1 and HSC-2 cells (Adachi, et al., 2005 ). Besides, colony-forming assays and cell death studies illustrated the differences of all three isoforms in their cytotoxicity. BimS was found to block colony formation and to stimulate apoptosis in the most effective way (O'Connor, et al., 1998 ).

Figure 6: The three main isoforms of Bim

Schematic structures of the three main isoforms are shown. BimEL and BimL contain the dynein binding site by which they are attached to the microtubule, BimS lacks this region.


Bim is regulated by transcriptional induction since upregulation results in an elevated amount of Bim-Bcl-xL complexes (Bouillet, et al., 1999 ) as well as by phosphorylation by c-Jun N-terminal kinase (JNK) (Lei and Davis, 2003 ) and is subject to several types of post-translational modifications (Puthalakath, et al., 1999 ). The BH3-region is essential for the interaction of Bim with anti-apoptotic Bcl-2 family members and for most of its deadly effect. Its hydrophobic C-terminus enables Bim to localize to cytoplasmic membranes (O'Connor, et al., 1998 ). Furthermore, it cannot trigger cell death in bax/bak double knockout cells. Hence, it might be that multi-domain pro-apoptotic proteins are activated by their release from anti-apoptotic factors upon BH3-only protein interference (Cheng, et al., 2001 ). Bim seems to act upstream of Bax and Bak, since this BH3-only protein was not able to induce neither cytochrome c release nor apoptosis in cells deficient for Bax and Bak (Wei, et al., 2001 ). Being at the microtubule, Bim has the perfect place to serve as stress sensor and communicator of stress signal to Bax and Bak. Overexpression of Bim highlighted its cytotoxity in several cell types, although it was shown that a few molecules of Bim are enough to induce apoptosis even in the presence of increased levels of anti-apoptotic proteins (Strasser, et al., 2000 ). Indeed, Bim possibly is the most potent killer of any other Bcl-2 family protein. Also, to date, Bim is not only the sole BH3-only protein expressing three different splicing products, but all of these main isoforms interact with Bcl-2 and Bcl-xL.

© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.
DiML DTD Version 4.0Zertifizierter Dokumentenserver
der Humboldt-Universität zu Berlin
HTML-Version erstellt am: