Oncogene (2006) 25, 4717–4724. doi:10.1038/sj.onc.1209605

The many shapes of mitochondrial death

G M Cereghetti1 and L Scorrano1

1Dulbecco-Telethon Institute, Venetian Institute of Molecular Medicine, Padova, Italy

Correspondence: Dr L Scorrano, Dulbecco-Telethon Institute, Venetian Institute of Molecular Medicine, Via Orus 2, I-35129 Padova, Italy. E-mail:



Mitochondria integrate apoptotic signalling by releasing cytochrome c and other proapoptotic cofactors needed for activation of effector caspases. Previously overlooked morphological changes, mitochondrial fragmentation and cristae remodelling, emerged as subroutines of the mitochondrial programme of apoptosis in mammalian cells, as well as in developmental cell death of Caenorhabditis elegans. Mitochondrial morphology results from fusion and fission processes, controlled by a growing set of 'mitochondria-shaping' proteins. Their levels and function appear to influence mitochondrial pathways of cell death, but mechanisms are largely unknown. An emerging model implicates different signals converging on mitochondria-shaping proteins to activate or deactivate them during apoptosis. In turn, these proteins can orchestrate changes in mitochondrial shape to insure cytochrome c release and progression of the apoptotic cascade. These therefore appear an appealing novel therapeutic target to modulate cell death in cancer.


mitochondria, cytochrome c release, fusion, fission, cristae, apoptosis



Mitochondria are the site of the tricarboxylic acid cycle and oxidative phosphorylation, the last steps of cellular respiration producing most energy required by the cell. The number of mitochondrial units in a particular cell type is determined by its energetic demand, so for example, hepatocytes possess thousands of mitochondria, whereas erythrocytes have lost all of them as their energy production relies exclusively on glycolysis. These 'mitochondrial units' have a major axis of 2–5 mum, as observed in classical electron micrographs of isolated mitochondria. In the cytosol of a living cell, these display an elongated, tubular morphology. Their shape is dynamic, continuously remodelled by cycles of fusion and fission events. These do not occur randomly, but are coordinated to respond to specific cellular needs, such as 'on-demand' transport of mitochondria at specific subcellular sites and equal division of the mitochondrial progeny between the two daughter cells during mitosis (Shaw and Nunnari, 2002). At the same time, during the early steps of cell death, mitochondria undergo dramatic structural changes (fragmentation of the reticulum and remodelling of the cristae) that are required to insure progression of the apoptotic cascade (Scorrano, 2005).

The search for the molecular mechanisms involved in the structural modification of mitochondria during apoptosis has raised interest on the molecular mechanisms working behind the scenes of mitochondrial shape changes, eventually leading to increased knowledge on the proteins controlling the morphology of this organelle. Several crucial questions remain open: what is the relative role of mitochondrial shape changes in the apoptotic pathways? Are these epiphenomena of the many functional changes occurring at the mitochondrial level during cell death or are these really required for apoptosis to proceed? What is their relationship with known players of the mitochondrial pathways of apoptosis, such as Bcl-2 family members? Are all these changes somehow coordinated and if yes, how? Can we identify signals that recruit these mitochondria-shaping proteins in the apoptotic cascade? Recent evidences have started to clarifying some of these questions.


The many shapes of mitochondria

In the last years, some of the principal regulators of mitochondrial shape have been identified and characterized. Pioneer work in Saccharomyces cerevisiae has been followed by the identification of orthologues in mammalian systems (Yaffe, 1999). Both fusion and fission processes are controlled by evolutionarily conserved large GTPases belonging to the dynamin family. Dynamins are large, ubiquitous mechanoenzymes that use the free energy generated by guanine 5'-triphosphate (GTP) hydrolysis to exert mechanical forces on biological membranes. The prototype dynamin I is essential for synaptic vesicle endocytosis in nerve terminals, by mediating the fission of the nascent vesicle from the plasma membrane. These have evolved to also regulate the dynamic changes of intracellular membranes, including mitochondria (Praefcke and McMahon, 2004).

In mammals, Opa1, mitofusin 1 (Mfn1) and mitofusin 2 (Mfn2) regulate fusion, whereas dynamin-related protein 1 (Drp-1, also known as Dlp1, Dvlp or Dymple) participates in mitochondrial fission.


Opa1 is a 120 kDa GTPase located on the inner mitochondrial membrane facing the intermembrane space (Olichon et al., 2002; Satoh et al., 2003). Opa1 has been identified as the gene mutated in familial autosomal dominant optic atrophy (ADOA), also known as type I Kjer disease, affecting mainly retinal ganglion cells and causing progressive blindness owing to the loss of these cells. The gene encoding for Opa1 is composed of 30 exons, and at least eight mRNA isoforms arising from alternative splicing and with tissue-specific expression have been identified (Delettre et al., 2001). A total of 83 family-specific mutations have been associated with ADOA so far, and their number seems to be growing (Alexander et al., 2000; Delettre et al., 2000, 2001, 2002; Pesch et al., 2001; Thiselton et al., 2001, 2002; Toomes et al., 2001; Marchbank et al., 2002; Shimizu et al., 2002; Amati-Bonneau et al., 2005; Ferre et al., 2005). About half of the mutations associated with the disease are located within the GTPase domain, suggesting that pathogenesis is related to loss of function of the protein and haploinsufficiency in patients with mutation on a single allele (Ferre et al., 2005). Indeed, ablation of Opa1 by short interfering RNA has a profound impact on cell viability, causing mitochondrial fragmentation, dissipation of mitochondrial membrane potential, ultrastructural changes and cell death (Olichon et al., 2003; Cipolat et al., 2004; Griparic et al., 2004; Chen et al., 2005). Expression of Opa1 promotes on the other hand mitochondrial fusion, provided that the GTPase and the C-terminal coiled-coil domain of the protein are not mutated (Cipolat et al., 2004). The requirement for the coiled-coil domain suggests that protein–protein interactions could be important for fusion mediated by Opa1. This could occur via Opa1 homo-oligomers, a situation similar to that of dynamin I (Zhang and Hinshaw, 2001). Alternatively, Opa1 could engage in hetero-oligomers with other mitochondria-shaping proteins. This is the case of Mfns, which heteroligomerize in the outer mitochondrial membrane (Chen et al., 2003). Whereas the former cannot be ruled out, genetic evidence supports interactions between Opa1 and Mfn1, which is required for Opa1-mediated fusion (Cipolat et al., 2004). Mitofusin 1 is an 84 kDa integral protein of the outer mitochondrial membrane, whose genetic ablation in the mouse results in embryonic lethality and marked reduction of mitochondrial fusion (Chen et al., 2003). Mitofusin 1 can engage in homo-oligomers, as substantiated by an elegant structural study by D Chan and colleagues: Mfn1 is required on adjacent mitochondria to mediate fusion via interactions of a heptad repeat region that mediates oligomerization of the protein (Koshiba et al., 2004). Mitofusin 1 shows similar topology and 81% sequence homology with Mfn2, and both proteins result necessary for apposing mitochondria during fusion (Koshiba et al., 2004). Deletion of Mfn2 in the mouse is also embryonic lethal, albeit it results in disruption of the placental trophoblast giant cell layer. Knocking down of both proteins blocks mitochondrial fusion, but embryonic fibroblasts lacking Mfn1 or Mfn2 display distinct types of fragmented mitochondria (Chen et al., 2003). Therefore, in spite of the high homology between the two proteins, their function may not be redundant. Mutations in Mfn2 are associated with Charcot–Marie–Tooth type 2A, an inherited peripheral neuropathy (Zuchner et al., 2004; Claeys et al., 2005). Moreover, levels of Mfn2 regulate oxidative metabolism in muscle mitochondria and are diminished in obese patients as well as in a rat model of type II diabetes (Bach et al., 2003); levels of Mfn2 can also control neointimal proliferation in models of hypertension (Chen et al., 2004). Other functional differences exist at the level of GTPase activity of the two proteins, as Mfn1 is much more efficient in hydrolyzing GTP (Ishihara et al., 2004). This has been confirmed by the finding that Mfn2 regulates fusion independently from its GTPase activity. A constitutively active mutant of Mfn2 showing enhanced rate of nucleotide exchange but a decreased GTP hydrolysis stimulates massive fusion (Neuspiel et al., 2005), raising the possibility that Mfn2 rather acts as a signalling GTPase.

Little is known about the molecular details of membrane fusion by Opa1 and mitofusins, but it is becoming clear that it differs from soluble N-ethylmaleimide-sensitive factor attachment protein receptor-(SNARE) mediated processes and is likely to have evolved independently (Meeusen et al., 2004; Meeusen and Nunnari, 2005). One aspect is that outer and inner membrane fusion seem to be separated events: fusion of the outer membrane relies on low levels of GTP hydrolysis and on the existence of a proton gradient across the inner membrane, while fusion of the inner membrane requires elevated levels of GTP hydrolysis (Malka et al., 2005; Meeusen and Nunnari, 2005). Whereas Opa1 is required for the fusion of the inner membrane, it is dispensable for fusion of the outer, which depends on the two mitofusins (Chen et al., 2005). Residual levels of inner membrane fusion exist in single mitofusin knockout cells, whereas double Mfn1-/-, Mfn2-/- cells completely lack mixing of matricial content, an indication of inner membrane fusion (Chen et al., 2003, 2005). This could mean that the two mitofusins converge on Opa1 to mediate fusion, suggesting that Opa1 could not distinguish between Mfn1 and Mfn2 to promote fusion. On the other hand, mitochondrial fusion is unaffected by Opa1 expression in Mfn1-/- but not Mfn2-/- cells, and Mfn2 but not Mfn1 cells promote mitochondrial elongation in cells where Opa1 has been knocked down (Cipolat et al., 2004). Whether mitofusins can be used unselectively by Opa1 to mediate fusion of the four membranes of two juxtaposing mitochondria remains to be clarified and a conclusive answer could come from the generation of mouse models of Opa1 ablation.

A further level of complexity in the regulation of fusion is expected to come from the analysis of the role of Parl, the mammalian orthologue of the rhomboid protease rbd1p of yeast. Rhomboids are intramembrane proteases that cleave and activate a variety of membrane-inserted proteins. They participate in crucial biological processes such as Notch signalling. Considerable interest comes from the discovery that one of the members of this family of proteins is targeted to the inner mitochondrial membrane. Its ablation in yeast results in mitochondrial dysfunction, as substantiated by the petite phenotype of the deleted strains, and abnormal morphology. Both phenotypes are caused by a defect in the processing and therefore activation of the yeast orthologue of Opa1, mgm1p. The mammalian orthologue of rbd1p, Parl, is a mitochondrial protein that shares considerable homology with the yeast rbd1p, except for its N-terminus, which is highly conserved among vertebrates but has no significant homology with Parl orthologues found in lower eukaryotes. Whereas this part of the protein is self-cleaved to generate small peptides involved in mitochondrion-to-nucleus signalling (Sik et al., 2004), whether Parl could also operate like rbd1p, activating Opa1 to maintain morphologically and functionally normal mitochondria remained to be addressed. To this end, we turned to a genetic model and we generated and analysed a Parl-/- mouse. The ablation of Parl did not result in embryonic lethality, but Parl-/- mice died between weeks 8 and 12 of muscular, thymic and splenic atrophy. Lack of Parl did not affect mitochondrial function and morphology in all the tissues analysed, but it resulted in increased degrees of apoptosis in situ and induced by intrinsic stimuli in fibroblasts derived from knockout embryos. Thus, at a major difference from yeast, the rhomboid Parl is not essential for mitochondrial fusion and function, but plays a role in controlling apoptosis (Cipolat et al., 2006). We will come later to the mechanism of Parl involvement in cell death.


Mitochondrial fission relies on Drp-1, a large cytosolic GTPase of about 82 kDa with similarities to dynamin (Labrousse et al., 1999; Smirnova et al., 2001). Dynamin-related protein 1 is cytosolic, and under circumstances that promote fission it translocates to mitochondria, where it docks to the outer membrane by interacting with hFis1, a 17 kDa protein with a tetratricopeptide repeat (TPR) region facing into the cytosol (Mozdy et al., 2000; James et al., 2003; Suzuki et al., 2003). The TPR domain seems to be responsible for interaction of hFis1 with Drp-1 (Yu et al., 2005). Knocking down hFis1 expression alters mitochondrial shape and distribution, indicating that hFis1 is required for Drp-1 function on mitochondria (James et al., 2003; Yoon et al., 2003). Furthermore, overexpression of hFis1 results in fragmented mitochondria, whereas blocking hFis1 by microinjected antibodies or by short interfering RNA results in elongated and collapsed mitochondria (James et al., 2003; Yoon et al., 2003). The Drp-1 homologue in yeast, Dnm1, causes mitochondrial fragmentation by accumulating in foci at sites of mitochondrial division and elongating in spirals of Dnm1 oligomers that span the size of mitochondria and constrict the membranes, breaking down the tubular organelle (McNiven et al., 2000; Ingerman et al., 2005). Similarly, dynamin I oligomerizes at the neck of nascent plasma membrane endocytotic vesicles that are undergoing fission. This suggests that all dynamin-related fission molecules may function by a common mechanism (Praefcke and McMahon, 2004). Mammalian Drp-1 has also been shown to form tubules (Yoon et al., 2001), but other molecules are required to organize and constrict mitochondria after formation of Drp-1 tubules. In particular, knocking down endophilin B1, a member of a family of proteins involved in endocytic vesicle formation, promotes the formation of Drp-1 tubules that elongate through the cytoplasm without breaking down the mitochondrial network (Karbowski et al., 2004b).

How fission of the two membranes (inner and outer) is coordinated and whether it always proceeds in concert remains to be elucidated. An elegant study by the group of M Rojo has examined fusion of the two membranes and concluded that both can proceed autonomously (Malka et al., 2005). It is conceivable that fission can also follow the same path, resulting for example in long tubuli of outer mitochondrial membrane that wrap around smaller mitochondrial units originated by fission of the inner membrane. The molecular mechanisms as well as the physiological consequences of this scenario remain to be determined.


Changes in mitochondrial shape during apoptosis: a summary

The detailed molecular events behind mitochondrial fusion and fission would not have gained tremendous interest if these were not associated with the relevant steps of mitochondrial death, namely release of cytochrome c and dysfunction of the organelle. Many other reviews have covered in detail the changes of mitochondrial shape during apoptosis (see e.p. Perfettini et al., 2005; Scorrano, 2005; Youle and Karbowski, 2005), so here we will summarize the two main morphological changes of mitochondria in cells committed to die. In most cells, mitochondria exist as a network of interconnected organelles, but early during apoptosis this network undergoes fragmentation (Frank et al., 2001). This apoptotic fission is causally related to cytochrome c release and dysfunction of the organelle, albeit the mechanism by which this occurs is still unclear. The importance of apoptotic fission is confirmed in the nematode Caenorhabditis elegans, where it appears so far the only mitochondrial change observed during developmental death (Jagasia et al., 2005).

Concurrently to their fission, mitochondria undergo ultrastructural changes that include opening of the narrow tubular cristae junction and fusion of individual cristae. This results in increased availability of the cristae-endowed cytochrome c to be released across the outer membrane (Scorrano et al., 2002). The two processes seem to be coordinated, as fission results in cristae remodelling by an yet uncharacterized mechanism (Germain et al., 2005). A cartoon summarizing these changes and the factors promoting them is depicted in Figure 1.

Figure 1.
Figure 1 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact or the author

Mitochondrial shape changes and apoptosis. Mitochondrial remodelling as well as fragmentation result in more cytochrome c available in the intermembrane space and released. Factors and proteins influencing these processes are shown.

Full figure and legend (97K)


The 'liaison dangereuse': mitochondria-shaping proteins and apoptosis

We have seen that mitochondrial shape changes may be relevant for the mitochondrial pathway of apoptosis. The climax of mitochondrial participation in apoptosis is the release of cytochrome c and other proapoptotic proteins that once in the cytosol coordinate the complete activation of effector caspases. This release is controlled by proteins of the Bcl-2 family, which includes both anti- and pro-apoptotic ones. In a widely accepted model, the 'BH3-only' subset of the proapoptotic members (like Bid, Bim, Bik) senses the death signal and conveys it to mitochondria, where these translocate to activate the 'multidomain' proapoptotic members like Bak and Bax (Danial and Korsmeyer, 2004). How mitochondria-shaping proteins interplay with the core apoptotic pathway remains to be elucidated, albeit some hints have started appearing. Circumstantial evidence of colocalization places Bax, Drp-1 and Mfn2 on the 'crime scene', at fission sites (Karbowski et al., 2002). In yeast, hFis1 blocks cell death induced by acetic acid. The mechanism by which hFis1 antagonizes apoptosis in S. cerevisiae remains unclear, but studies with recombinant protein have shown that hFis1 shares with Bcl-xL the ability to release small molecules from synthetic vesicles (Fannjiang et al., 2004). Whether mitochondria-shaping proteins physically interact and regulate pro- and antiapoptotic members of the Bcl-2 family is unclear. Overexpression studies have shown that mitofusins interfere with activation of Bax and Bak (Sugioka et al., 2004; Neuspiel et al., 2005), albeit these failed to show a direct interaction between these proteins. Ablation of Mfn1 does not influence multidomain activation and cytochrome c release, suggesting that Mfn2 has stronger regulatory effect on Bax and Bak (Arnoult et al., 2005a). In apoptosis experiments following genetic modulation of levels and function of mitochondria-shaping proteins, the strongest effect on cell death was associated with Opa1 (Olichon et al., 2003; Lee et al., 2004). As Opa1 is an inner mitochondrial membrane protein facing the intermembrane space (Olichon et al., 2002), its localization argues against the modulation of Bax, Bak activation, occurring on the outer membrane. Moreover, no interaction has been demonstrated so far with members of the Bcl-2 family. Thus, an alternative mechanism should be proposed to explain the effect of Opa1 on cell death. Given its localization, Opa1 is a natural candidate to regulate the shape of the cristae. Genetic ablation of Opa1 indeed disrupts mitochondrial ultrastructure, but it is unclear whether this occurs as a result of the activation of the apoptotic cascade or as a consequence of the inactivation of Opa1-specific function (Olichon et al., 2003). Moreover, no evidence existed for a role of Opa1 in the cristae remodelling pathway during apoptosis and its role as a pro-fusion protein further bamboozled the scenario. We therefore examined the role of Opa1 in apoptosis in detail. A genetic approach revealed that Opa1 blocks from apoptosis independently of mitofusins, and therefore of mitochondrial fusion. This effect is granted by the ability of Opa1 to keep in check the cristae during apoptosis induced by intrinsic stimuli, as confirmed by electron tomography and measurements of cristae junctions. How Opa1 maintains the normal 18–20 nm diameter of the cristae junction (Frey and Mannella, 2000) is unclear. Other dynamin-related proteins constrict membranes from outside, but Opa1 is located inside the tubule that it should constrict. To solve this discrepancy, we followed mitochondrial sublocalization of Opa1 and found a soluble fraction of it in the intermembrane space. Interestingly, this soluble form binds to membrane-bound Opa1 to form easily detectable trimers, tetramers and probably even higher molecular weight oligomers. Of note, following BID treatment, these oligomers are disrupted, even before the activation of multidomain proapoptotics (Frezza et al., 2006). How a soluble isoform of Opa1 was generated becomes clearer by analysing Opa1 sub-mitochondrial distribution in Parl-/- mouse. The ablation of this protease results in a quasi-complete lack of soluble, intermembrane space Opa1 (suggesting that other proteases participate in this process). Opa1 oligomers do not form, cristae display larger junctions and cytochrome c is more readily available for release, resulting in increased susceptibility to apoptosis. Expression of wild-type Opa1 cannot protect Parl-/- cells from apoptosis. Conversely, an Opa1 mutant constitutively targeted to the IMS exerts its antiapoptotic action also in Parl-/- cells (Cipolat et al., 2006; Frezza et al., 2006). Thus, the complex between soluble and membrane-bound isoforms of Opa1 could represent the structure controlling the diameter of the cristae junction during apoptosis. Whether this complex is a direct target for proapoptotic members of the Bcl-2 family remains to be elucidated, but it has been reported that a fraction of Opa1 is released during apoptosis and that this could represent the mechanism leading to inhibition of fusion in cells commited to die (Karbowski et al., 2004a; Arnoult et al., 2005a). On the other hand, soluble, intermembrane space Opa1, at a major difference from its membrane-bound counterpart, does not promote mitochondrial fusion (Cipolat et al., 2006) and it is therefore unlikely that its release could represent the only mechanism promoting mitochondrial fission during apoptosis.

Other signals should participate in the recruitment of pro-fission molecules on the surface of mitochondria. Candidates include second messengers as well as proteins that are released from mitochondria in response to a death stimulus.


How do they go from here to there?

New evidences are starting to elucidate the unresolved issue of how Drp-1 is recruited to mitochondria in the course of cell death. The migration of Drp-1 presumably requires two subroutines: one involving the generation of the signal that activates the translocation programme; the other providing the physical vector to transfer Drp-1 to mitochondria.

The first subroutine has to deal with core signalling machineries of the cell and we should spend some words to place mitochondria in the context of integrated cell signalling, especially during apoptosis. Mitochondria modulate calcium signalling by regulating cytosolic Ca2+ transients, impinging on Ca2+ re-uptake by the endoplasmic reticulum and regulating Ca2+ entry into the cell via capacitative Ca2+ entry (Rizzuto et al., 2000). The role of Ca2+ in apoptosis was an early discovery and many molecular details that elucidate pathways and mechanisms of the involvement of this second messenger in cell death have been added (for a review, see Orrenius et al., 2003). A widely accepted model anticipates that cytosolic Ca2+ levels increase during apoptosis by selected stimuli; and proteins that regulate apoptosis, like members of the Bcl-2 family, also influence Ca2+ homeostasis to control death. This model is supported for example by the finding that expression of antiapoptotic Bcl-2 and Bcl-xL as well as ablation of proapoptotic Bax and Bak reduces resting endoplasmic reticulum (ER) Ca2+ levels (Lam et al., 1994; Foyouzi-Youssefi et al., 2000; Pinton et al., 2000; Li et al., 2002; Scorrano, 2003; Bassik et al., 2004). Expression of Bax raises steady-state Ca2+ levels in the ER before cells are committed to die (Chami et al., 2004), whereas progression towards death is associated with depletion of ER stores and mitochondrial accumulation of Ca2+ (Nutt et al., 2002a, 2002b). Ca2+ homeostasis is at the centre stage of apoptosis modulation; but does it also regulate the association of Drp-1 with mitochondria? Recent work by Shore and colleagues provides a clue to this question. Activation of caspase 8 by death receptors cleaves BAP31, an ER protein, resulting in Ca2+ release from the ER and uptake into mitochondria, followed by recruitment of Drp-1 to the organelle and subsequent fragmentation. Ca2+-dependent fragmentation of the mitochondrial network leads to release of cytochrome c and activation of the mitochondrial pathway of apoptosis (Breckenridge et al., 2003). Thus, Ca2+ signals from the ER, following an apoptotic stimulus, promote mitochondrial fragmentation and activation of the mitochondrial pathway of death. This model is further supported by the finding that BIK, a BH3-only member of the Bcl-2 family anchored on the ER surface, induces Drp-1 recruitment to mitochondria by promoting Ca2+ release from the ER. This results in Drp-1-dependent cristae remodelling but not in Drp-1-mediated egress of cytochrome c from mitochondria, which occurs only after activation of the multidomain proapoptotics Bax and Bak by a 'second hit' conveyed for example by a different BH3-only protein (Germain et al., 2005). These evidences suggest that Ca2+ release from the ER can mediate the association of Drp-1 with mitochondria. At the same time, the possibility that Drp-1 moves following mitochondrial dysfunction and a second, yet-unknown signal coming from damaged mitochondria remains still viable. On the other hand, mitochondrial fission could also block progression of Ca2+ waves along the mitochondrial network, a quite crucial feature in propagation of the apoptotic signal in large cells such as cardiomyocytes (Pacher and Hajnoczky, 2001). This has been recently demonstrated to occur in cells that express high levels of Drp-1, where apoptosis driven by Ca2+-dependent signals is inhibited as a consequence of the inability of fragmented mitochondria to convey the deadly Ca2+ wave across the cytoplasm (Szabadkai et al., 2004). The complex relationship between Ca2+ and mitochondrial fragmentation during apoptosis is summarized in Figure 2.

Figure 2.
Figure 2 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact or the author

Ca2+ and mitochondrial shape changes: a bidirectional relationship. Ca2+ crosstalk between endoplasmic reticulum (ER) and mitochondria is required for a variety of cellular and mitochondrial functions. The network of fused mitochondria is crucial in ensuring rapid propagation of the Ca2+ wave and dynamin-related protein 1 (Drp-1)-dependent mitochondrial fission can impede this (a). On the other hand, aberrant release of Ca2+ from the ER induced by apoptotic stimuli can trigger mobilization of proapoptotic Bax and Bak to mitochondria, where these coalesce with mitofusin 2 and Drp-1 to fragment the mitochondrial network (b).

Full figure and legend (132K)

One appealing candidate to signal recruitment of Drp-1 to mitochondria could be the cytosolic phosphatase calcineurin, which plays an adaptor role in multiple cellular pathways, including endocytosis. We have therefore investigated whether calcineurin could play a role in Drp-1-dependent mitochondrial fission. Stimuli that induce mitochondrial dysfunction and Ca2+ overload increase calcineurin activity. Normally, phosphorylated Drp-1 resides in a cytosolic complex with calcineurin. Its activation results in dephosphorylation of Drp-1 and its dissociation from this complex and translocation to mitochondria. Inhibition of calcineurin blocks Drp-1-dependent fission following mitochondrial dysfunction and Ca2+ overload (Figure 3) (Cereghetti et al., 2006). Interestingly, calcineurin plays a similar role in recruiting dynamin I to the fission neck of nascent endocytic vesicles (Liu et al., 1994), in close resemblance to this novel calcineurin switch that contributes to the regulation of mitochondrial shape.

Figure 3.
Figure 3 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact or the author

Calcineurin-dependent translocation of dynamin-related protein 1 (Drp-1) to mitochondria. Mitochondrial dysfunction results in activation of cytosolic calcineurin that dephosphorylates Drp-1 and recruits it to mitochondria, inducing fragmentation of the network. See text for details.

Full figure and legend (91K)

If Ca2+ is the second messenger coordinating Drp-1 recruitment, the problem of specificity is raised. In other words, how can such a universal messenger promote fission during normal life of the cell and fragmentation of mitochondria during apoptosis? One possibility is that calcineurin is effectively activated only when Ca2+ rises are associated with mitochondrial dysfunction. Alternatively, recruitment of Drp-1 during apoptosis requires additional components, such as proteins released from mitochondria. TIMM8a, a mitochondrial intermembrane space protein involved in protein import and mutated in the Mohr–Tranebjaerg–Jensen deafness–dystonia–optic atrophy syndrome (Koehler et al., 1999), is co-released with cytochrome c into the cytoplasm, where it binds to and promotes the mitochondrial redistribution of Drp-1 (Arnoult et al., 2005b). Proteins released from mitochondria could represent an efficient feed-forward loop to precipitate mitochondrial fragmentation and augment downstream events such as Drp-1-dependent cristae remodelling during apoptosis.

Following the recruitment signal, Drp-1 must physically move to sites of fission on mitochondrial membranes. Two candidates have been reported to coordinate its migration to mitochondria. The first one is represented by the retrograde motor protein complex dynein/dynactin, associated with microtubules. Dynamin-related protein 1 interacts with the dynamitin subunit of dynactin and overexpression of dynamitin results in translocation of Drp-1 from mitochondria to cytosol and microsomes (Varadi et al., 2004). This model is substantiated by the fact that interaction between Drp-1 and dynein/dynactin takes place on microtubules, which provide the tracks for mitochondrial movement in mammalian cells (Morris and Hollenbeck, 1995). The second theory, developed by De Vos et al. (2005), suggests that F-actin facilitates transport of Drp-1 to mitochondria. Mitochondrial fragmentation induced by inhibitors of electron transport and ATPase is prevented by disruption of the actin cytoskeleton. One unresolved question is whether treatments that disrupt cytoskeleton can alter for example association of Drp-1 with dynein/dynamitin. Moreover, the two interacting cytoskeletal components, microfilaments and microtubules, can both cooperate in driving Drp-1 to mitochondria. Once associated with mitochondria, Drp-1 is stabilized by sumoylation. The Sumo ligases Ubc9 and Sumo1 specifically interact with Drp-1 and Sumo1 colocalizes with sites of mitochondrial fission, even after fission has occurred. Accordingly, overexpression of Sumo1 increases fission, by protecting Drp-1 from degradation (Harder et al., 2004). Thus, once translocated to mitochondria, Drp-1 is stabilized to complete its task and fragment mitochondria.


Can mitochondrial shape changes be forced to induce apoptosis of cancer cells?

No matter how it occurs, the complete release of cytochrome c in the cytosol could be crucial to insure adequate strength of caspase activation despite their blockade by endogenous inhibitors such as the inhibitors of activation of proteases , often upregulated in cancer cells (Fong et al., 2000; Liston et al., 2003). Interestingly, also Mfn1 is upregulated in non-small-cell adenocarcinoma of the lung and in other cancer cell lines, further substantiating a role for mitochondria-shaping proteins in transformation (Chung et al., 2001). Mitochondria-shaping proteins seem therefore to be an appealing target to modulate the mitochondrial phase of apoptosis in cancer cells. One unresolved issue complicates so far approaches to target them in order to enforce apoptosis of cancer cells: we must know how they regulate the crucial event in the mitochondrial pathway of apoptosis, that is, cytochrome c release. In particular, it is unclear if and why egress of cytochrome c should be favoured from fragmented mitochondria. This question has crucial implication if one would modulate mitochondrial morphological changes in order to influence apoptosis. We believe that data support a 'unified' theory of mitochondrial morphological changes during apoptosis. This theory, demonstrated at least in the case of the BH3-only protein BIK, implies that BH3-only proteins induce fragmentation driven by pro-fission members like Drp-1, which in turn results in cristae remodelling and increased cytochrome c release (Germain et al., 2005). To hold true, this model should pass some crucial experimental tests, dealing with the antiapoptotic effects of molecules such as mitofusins and Opa1. If Opa1 protects from apoptosis independently of its pro-fusion effects but impinging directly on cristae remodelling (Frezza et al., 2006), the question of whether mitofusins block cristae remodelling remains open: do these prevent activation of Drp-1, or do these act at different levels, perhaps crosstalking with members of the Bcl-2 family? Some evidence would argue that Mfn2 interferes with activation of Bax and Bak (Sugioka et al., 2004; Neuspiel et al., 2005), but much more work is needed to understand how this blockage occurs. This question is not an irrelevant one, as inhibitors of GTPase activity could have opposite effects on cell viability. If fission mediated by the GTPase Drp-1 is crucial to release cytochrome c, GTPase inhibitors should not have any beneficial effect in driving apoptosis of cancer cells. Conversely, should upregulation of mitochondria-shaping proteins such as mitofusins or Opa1 prove to be a hallmark of transformation, one could imagine that targeting them with specific inhibitors could provide a way to maximize cytochrome c release and hence apoptosis of cancer cells. Alternatively, one potential approach would be to enforce Drp-1 translocation to mitochondria by impinging on the pathways that control its translocation to the organelle and subsequent fission and release of cytochrome c.

In conclusion, mitochondria-shaping proteins are an appealing target to modulate death of cancer cells by directly attacking the mitochondrial pathway of cell death.



  1. Alexander C, Votruba M, Pesch UE, Thiselton DL, Mayer S, Moore A et al. (2000). Nat Genet 26: 211–215. | Article | PubMed | ISI | ChemPort |
  2. Amati-Bonneau P, Guichet A, Olichon A, Chevrollier A, Viala F, Miot S et al. (2005). Ann Neurol 58: 958–963. | Article | PubMed | ChemPort |
  3. Arnoult D, Grodet A, Lee YJ, Estaquier J, Blackstone C. (2005a). J Biol Chem 280: 35742–35750. | Article | PubMed | ChemPort |
  4. Arnoult D, Rismanchi N, Grodet A, Roberts RG, Seeburg DP, Estaquier J et al. (2005b). Curr Biol 15: 2112–2118. | Article | PubMed | ChemPort |
  5. Bach D, Pich S, Soriano FX, Vega N, Baumgartner B, Oriola J et al. (2003). J Biol Chem 278: 17190–17197. | Article | PubMed | ISI |
  6. Bassik MC, Scorrano L, Oakes SA, Pozzan T, Korsmeyer SJ. (2004). EMBO J 23: 1207–1216. | Article | PubMed | ChemPort |
  7. Breckenridge DG, Stojanovic M, Marcellus RC, Shore GC. (2003). J Cell Biol 160: 1115–1127. | Article | PubMed | ISI | ChemPort |
  8. Cereghetti GM, Stangherlin A, Scorrano L. (2006) Submitted.
  9. Chami M, Prandini A, Campanella M, Pinton P, Szabadkai G, Reed JC et al. (2004). J Biol Chem 279: 54581–54589. | Article | PubMed | ISI | ChemPort |
  10. Chen H, Chomyn A, Chan DC. (2005). J Biol Chem 280: 26185–26192. | Article | PubMed | ChemPort |
  11. Chen H, Detmer SA, Ewald AJ, Griffin EE, Fraser SE, Chan DC. (2003). J Cell Biol 160: 189–200. | Article | PubMed | ISI | ChemPort |
  12. Chen KH, Guo X, Ma D, Guo Y, Li Q, Yang D et al. (2004). Nat Cell Biol 6: 872–883. | Article | PubMed | ChemPort |
  13. Chung JG, Yeh KT, Wu SL, Hsu NY, Chen GW, Yeh YW et al. (2001). Cancer Res 61: 8873–8879. | PubMed | ChemPort |
  14. Cipolat S, de Brito OM, Dal Zilio B, Scorrano L. (2004). Proc Natl Acad Sci USA 101: 15927–15932. | Article | PubMed | ChemPort |
  15. Cipolat S, Rudka T, Hartmann D, Costa V, Serneels D, Craessaerts K et al. (2006) Submitted.
  16. Claeys K, Verhoeven K, Zuchner S, Schroder JM, Vance JM, Timmerman V et al. (2005). J Peripher Nerv Syst 10: 16–17.
  17. Danial NN, Korsmeyer SJ. (2004). Cell 116: 205–219. | Article | PubMed | ISI | ChemPort |
  18. De Vos KJ, Allan VJ, Grierson AJ, Sheetz MP. (2005). Curr Biol 15: 678–683. | Article | PubMed | ChemPort |
  19. Delettre C, Griffoin JM, Kaplan J, Dollfus H, Lorenz B, Faivre L et al. (2001). Hum Genet 109: 584–591. | Article | PubMed | ISI | ChemPort |
  20. Delettre C, Lenaers G, Griffoin JM, Gigarel N, Lorenzo C, Belenguer P et al. (2000). Nat Genet 26: 207–210. | Article | PubMed | ISI | ChemPort |
  21. Delettre C, Lenaers G, Pelloquin L, Belenguer P, Hamel CP. (2002). Mol Genet Metab 75: 97–107. | Article | PubMed | ISI | ChemPort |
  22. Fannjiang Y, Cheng WC, Lee SJ, Qi B, Pevsner J, McCaffery JM et al. (2004). Genes Dev 18: 2785–2797. | Article | PubMed | ISI | ChemPort |
  23. Ferre M, Amati-Bonneau P, Tourmen Y, Malthiery Y, Reynier P. (2005). Hum Mutat 25: 423–428. | Article | PubMed | ChemPort |
  24. Fong WG, Liston P, Rajcan-Separovic E, St Jean M, Craig C, Korneluk RG. (2000). Genomics 70: 113–122. | Article | PubMed | ISI | ChemPort |
  25. Foyouzi-Youssefi R, Arnaudeau S, Borner C, Kelley WL, Tschopp J, Lew DP et al. (2000). Proc Natl Acad Sci USA 97: 5723–5728. | Article | PubMed | ChemPort |
  26. Frank S, Gaume B, Bergmann-Leitner ES, Leitner WW, Robert EG, Catez F et al. (2001). Dev Cell 1: 515–525. | Article | PubMed | ISI | ChemPort |
  27. Frey TG, Mannella CA. (2000). Trends Biochem Sci 25: 319–324. | Article | PubMed | ISI | ChemPort |
  28. Frezza C, Cipolat S, de Brito OM, Micaroni M, Beznoussenko GV, Bartoli D et al. (2006) Submitted.
  29. Germain M, Mathai JP, McBride HM, Shore GC. (2005). EMBO J 24: 1546–1556. | Article | PubMed | ISI | ChemPort |
  30. Griparic L, van der Wel NN, Orozco IJ, Peters PJ, van der Bliek AM. (2004). J Biol Chem 279: 18792–18798. | Article | PubMed | ISI | ChemPort |
  31. Harder Z, Zunino R, McBride H. (2004). Curr Biol 14: 340–345. | Article | PubMed | ISI | ChemPort |
  32. Ingerman E, Perkins EM, Marino M, Mears JA, McCaffery JM, Hinshaw JE et al. (2005). J Cell Biol 170: 1021–1027. | Article | PubMed | ChemPort |
  33. Ishihara N, Eura Y, Mihara K. (2004). J Cell Sci 117: 6535–6546. | Article | PubMed | ISI | ChemPort |
  34. Jagasia R, Grote P, Westermann B, Conradt B. (2005). Nature 433: 754–760. | Article | PubMed | ISI | ChemPort |
  35. James DI, Parone PA, Mattenberger Y, Martinou JC. (2003). J Biol Chem 278: 36373–36379. | Article | PubMed | ISI | ChemPort |
  36. Karbowski M, Arnoult D, Chen H, Chan DC, Smith CL, Youle RJ. (2004a). J Cell Biol 164: 493–499. | Article | PubMed | ISI | ChemPort |
  37. Karbowski M, Jeong SY, Youle RJ. (2004b). J Cell Biol 166: 1027–1039. | Article | PubMed | ISI | ChemPort |
  38. Karbowski M, Lee YJ, Gaume B, Jeong SY, Frank S, Nechushtan A et al. (2002). J Cell Biol 159: 931–938. | Article | PubMed | ISI | ChemPort |
  39. Koehler CM, Leuenberger D, Merchant S, Renold A, Junne T, Schatz G. (1999). Proc Natl Acad Sci USA 96: 2141–2146. | Article | PubMed | ChemPort |
  40. Koshiba T, Detmer SA, Kaiser JT, Chen H, McCaffery JM, Chan DC. (2004). Science 305: 858–862. | Article | PubMed | ISI | ChemPort |
  41. Labrousse AM, Zappaterra MD, Rube DA, van der Bliek AM. (1999). Mol Cell 4: 815–826. | Article | PubMed | ISI | ChemPort |
  42. Lam M, Dubyak G, Chen L, Nunez G, Miesfeld RL, Distelhorst CW. (1994). Proc Natl Acad Sci USA 91: 6569–6573. | Article | PubMed | ChemPort |
  43. Lee YJ, Jeong SY, Karbowski M, Smith CL, Youle RJ. (2004). Mol Biol Cell 15: 5001–5011. | Article | PubMed | ISI | ChemPort |
  44. Li C, Fox CJ, Master SR, Bindokas VP, Chodosh LA, Thompson CB. (2002). Proc Natl Acad Sci USA 99: 9830–9835. | Article | PubMed | ChemPort |
  45. Liston P, Fong WG, Korneluk RG. (2003). Oncogene 22: 8568–8580. | Article | PubMed | ISI | ChemPort |
  46. Liu JP, Sim AT, Robinson PJ. (1994). Science 265: 970–973. | PubMed | ISI | ChemPort |
  47. Malka F, Guillery O, Cifuentes-Diaz C, Guillou E, Belenguer P, Lombes A et al. (2005). EMBO Rep 6: 853–859. | Article | PubMed | ChemPort |
  48. Marchbank NJ, Craig JE, Leek JP, Toohey M, Churchill AJ, Markham AF et al. (2002). J Med Genet 39: e47. | Article | PubMed | ChemPort |
  49. McNiven MA, Yoon Y, Pitts KR. (2000). Mol Biol Cell 11: 219A.
  50. Meeusen S, McCaffery JM, Nunnari J. (2004). Science 305: 1747–1752. | Article | PubMed | ISI | ChemPort |
  51. Meeusen SL, Nunnari J. (2005). Curr Opin Cell Biol 17: 389–394. | Article | PubMed | ChemPort |
  52. Morris RL, Hollenbeck PJ. (1995). J Cell Biol 131: 1315–1326. | Article | PubMed | ChemPort |
  53. Mozdy AD, McCaffery JM, Shaw JM. (2000). J Cell Biol 151: 367–380. | Article | PubMed | ISI | ChemPort |
  54. Neuspiel M, Zunino R, Gangaraju S, Rippstein P, McBride HM. (2005). J Biol Chem 280: 25060–25070. | Article | PubMed | ChemPort |
  55. Nutt LK, Chandra J, Pataer A, Fang B, Roth JA, Swisher SG et al. (2002a). J Biol Chem 277: 20301–20308. | Article | PubMed | ISI | ChemPort |
  56. Nutt LK, Pataer A, Pahler J, Fang B, Roth J, McConkey DJ et al. (2002b). J Biol Chem 277: 9219–9225. | Article | PubMed | ISI | ChemPort |
  57. Olichon A, Baricault L, Gas N, Guillou E, Valette A, Belenguer P et al. (2003). J Biol Chem 278: 7743–7746. | Article | PubMed | ISI | ChemPort |
  58. Olichon A, Emorine LJ, Descoins E, Pelloquin L, Brichese L, Gas N et al. (2002). FEBS Lett 523: 171–176. | Article | PubMed | ISI | ChemPort |
  59. Orrenius S, Zhivotovsky B, Nicotera P. (2003). Nat Rev Mol Cell Biol 4: 552–565. | Article | PubMed | ISI | ChemPort |
  60. Pacher P, Hajnoczky G. (2001). EMBO J 20: 4107–4121. | Article | PubMed | ISI | ChemPort |
  61. Perfettini JL, Roumier T, Kroemer G. (2005). Trends Cell Biol 15: 179–183. | Article | PubMed | ISI | ChemPort |
  62. Pesch UE, Leo-Kottler B, Mayer S, Jurklies B, Kellner U, Apfelstedt-Sylla E et al. (2001). Hum Mol Genet 10: 1359–1368. | Article | PubMed | ISI | ChemPort |
  63. Pinton P, Ferrari D, Magalhaes P, Schulze-Osthoff K, Di VF, Pozzan T et al. (2000). J Cell Biol 148: 857–862. | Article | PubMed | ISI | ChemPort |
  64. Praefcke GJ, McMahon HT. (2004). Nat Rev Mol Cell Biol 5: 133–147. | Article | PubMed | ISI | ChemPort |
  65. Rizzuto R, Bernardi P, Pozzan T. (2000). J Physiol 529(Part 1): 37–47, 37–47. | Article | PubMed | ISI | ChemPort |
  66. Satoh M, Hamamoto T, Seo N, Kagawa Y, Endo H. (2003). Biochem Biophys Res Commun 300: 482–493. | Article | PubMed | ISI | ChemPort |
  67. Scorrano L. (2003). Cell Death Differ 10: 1287–1289. | Article | PubMed | ChemPort |
  68. Scorrano L. (2005). J Bioenerg Biomembr 37: 165–170. | Article | PubMed | ISI | ChemPort |
  69. Scorrano L, Ashiya M, Buttle K, Weiler S, Oakes SA, Mannella CA et al. (2002). Dev Cell 2: 55–67. | Article | PubMed | ISI | ChemPort |
  70. Shaw JM, Nunnari J. (2002). Trends Cell Biol 12: 178–184. | Article | PubMed | ISI | ChemPort |
  71. Shimizu S, Mori N, Kishi M, Sugata H, Tsuda A, Kubota N. (2002). Jpn J Ophthalmol 46: 336–340. | Article | PubMed | ChemPort |
  72. Sik A, Passer BJ, Koonin EV, Pellegrini L. (2004). J Biol Chem 279: 15323–15329. | Article | PubMed | ChemPort |
  73. Smirnova E, Griparic L, Shurland DL, van der Bliek AM. (2001). Mol Biol Cell 12: 2245–2256. | PubMed | ISI | ChemPort |
  74. Sugioka R, Shimizu S, Tsujimoto Y. (2004). J Biol Chem 279: 52726–52734. | Article | PubMed | ISI | ChemPort |
  75. Suzuki M, Jeong SY, Karbowski M, Youle RJ, Tjandra N. (2003). J Mol Biol 334: 445–458. | Article | PubMed | ISI | ChemPort |
  76. Szabadkai G, Simoni AM, Chami M, Wieckowski MR, Youle RJ, Rizzuto R. (2004). Mol Cell 16: 59–68. | Article | PubMed | ISI | ChemPort |
  77. Thiselton DL, Alexander C, Morris A, Brooks S, Rosenberg T, Eiberg H et al. (2001). Hum Genet 109: 498–502. | Article | PubMed | ChemPort |
  78. Thiselton DL, Alexander C, Taanman JW, Brooks S, Rosenberg T, Eiberg H et al. (2002). Invest Ophthalmol Vis Sci 43: 1715–1724. | PubMed | ISI |
  79. Toomes C, Marchbank NJ, Mackey DA, Craig JE, Newbury-Ecob RA, Bennett CP et al. (2001). Hum Mol Genet 10: 1369–1378. | Article | PubMed | ISI | ChemPort |
  80. Varadi A, Johnson-Cadwell LI, Cirulli V, Yoon Y, Allan VJ, Rutter GA. (2004). J Cell Sci 117: 4389–4400. | Article | PubMed | ChemPort |
  81. Yaffe MP. (1999). Science 283: 1493–1497. | Article | PubMed | ISI | ChemPort |
  82. Yoon Y, Krueger EW, Oswald BJ, McNiven MA. (2003). Mol Cell Biol 23: 5409–5420. | Article | PubMed | ISI | ChemPort |
  83. Yoon Y, Pitts KR, McNiven MA. (2001). Mol Biol Cell 12: 2894–2905. | PubMed | ISI | ChemPort |
  84. Youle RJ, Karbowski M. (2005). Nat Rev Mol Cell Biol 6: 657–663. | Article | PubMed | ISI | ChemPort |
  85. Yu TZ, Fox RJ, Burwell LS, Yoon Y. (2005). J Cell Sci 118: 4141–4151. | Article | PubMed | ChemPort |
  86. Zhang P, Hinshaw JE. (2001). Nat Cell Biol 3: 922–926. | Article | PubMed | ISI | ChemPort |
  87. Zuchner S, Mersiyanova IV, Muglia M, Bissar-Tadmouri N, Rochelle J, Dadali EL et al. (2004). Nat Genet 36: 449–451. | Article | PubMed | ISI | ChemPort |


LS is an assistant Telethon scientist of the Dulbecco-Telethon Institute, and research in his laboratory is supported by Telethon Italy, AIRC Italy, Compagnia di San Paolo, Human Frontier Science Program Organization, UMDF. GMC was supported by a Fellowship for Perspective Researchers of the Swiss National Foundation.



These links to content published by NPG are automatically generated


Ahead of the curve: mitochondrial fusion and phospholipase D

Nature Cell Biology News and Views (01 Nov 2006)