From mitochondrial morphology to apoptosis: genetic analysis of OPA1 function and regulation

Mitochondria are essential organelles for life and death of the cell: they produce most of the cellular ATP (Danial et al., 2003), regulate cytosolic Ca2+ signalling (Rizzuto et al., 2000), and integrate and amplify different apoptotic stimuli (Green and Kroemer, 2004). Such a functional versatility is matched by a complex and dynamic morphology, both at the ultrastructural and at the cellular level (Griparic and van der Bliek). At the ultrastructural level, the mitochondrial cristae constitute a separate compartment connected to the thin intermembrane space by narrow tubular junctions (Frey and Mannella, 2000). In the cytosol, mitochondria are organized in a network of individual organelles that dynamically fuse and divide. Mitochondrial morphology results from the equilibrium between fusion and fission processes, controlled by a family of "mitochondria-shaping" proteins, many of which are dynamin-related proteins initially identified by genetic screens in buddying yeast (Dimmer et al., 2002; Shaw and Nunnari, 2002). Dynamins are ubiquitous mechano-enzymes that hydrolyze GTP to regulate fusion, fission, tubulation and elongation of cellular membranes (McNiven et al., 2000). In mammalians, mitochondrial fission is controlled by a cytosolic dynamin related protein DRP-1 (Smirnova et al., 2001) that translocates to sites of mitochondrial fragmentation where it binds to FIS1, its adapter in the outer membrane (Yoon et al., 2003) (James et al., 2003). Fusion is controlled by mitofusin-1 (MFN1) and-2 (MFN2), two large GTPases of the outer mitochondrial membrane, orthologues of S. cerevisiae Fzo1p (Rapaport et al., 1998). OPA1, the mammalian homologue of S. cerevisiae Mgm1p, is the only dynamin-related protein of the inner mitochondrial membrane (Olichon et al., 2002). Loss-of-function or dominant-negative mutations in Opa1 are associated with autosomal dominant optic atrophy (DOA), the leading cause of inherited optic neuropathy, characterized by retinal ganglion cells degeneration followed by ascending atrophy of the optic nerve (Alexander et al., 2000; Delettre et al., 2000). The aim of my PhD has been to generate, use and analyze genetic models in order to unravel the biological function of OPA1 as well as its regulation. In order to dissect the biological function of OPA1, we undertook a combination of genetics and imaging to address its role in regulating mitochondrial fusion/fission equilibrium. Imaging of wild type mouse embryonic fibroblasts (MEFs) cotransfected with a mitochondrially targeted cyan fluorescent protein (mtCFP) showed mitochondria as individual organelles, rod or round-shaped, with an average length of 3±0.34 µm along their major axis. Morphometric analysis confirmed that only 23% of the analyzed cells displayed elongated mitochondria, i.e. cells with axial length >5 µm and roundness index <0.5 in more than 50% of mitochondria. Cotransfection of OPA1 with mtCFP induced visible changes in the shape of the mitochondrial reticulum. The rod-shaped mitochondria appeared now to be interconnected in a branched network. Morphometric analysis confirmed this mitochondria-shaping effect of OPA1, with more than 50% of the cells analyzed showing elongated mitochondria. Furthermore, we analyzed the effect of pathogenic mutations of OPA1 on its ability to elongate mitochondria. A missense mutation in the GTPase domain (K301A) that reduces the GTPase activity of more than 80%, as well as a truncative one in the coiled coil domain (R905stop), which eliminates the C-terminal coiled-coil domain required in protein-protein interactions, abolished the ability of OPA1 to elongate mitochondria, indicating that it requires a functional GTPase and coiled-coil domain. To address the effect of reduced OPA1 levels on mitochondrial morphology we turned to stable, plasmid-generated RNA interference (RNAi). In cell clones where OPA1 was ablated, mitochondria appeared globular and fragmented as opposed to the rod, elongated organelles of the control clones. Tubulation induced by OPA1 is not the results of simple juxtaposition of mitochondria, but it represents the steady state appearance of increased mitochondrial fusion events, as substantiated by assays of mitochondrial fusion in polykarions induced by PEG treatment. Expression of OPA1 significantly speeded up mixing of matricial content, whereas its downregulation reduced mitochondrial fusion. In yeast, the pro-fusion activity of Mgm1p, the orthologue of OPA1, depends on the outer membrane mitochondria-shaping protein Fzo1p. We therefore wished to ascertain whether this paradigm was maintained in higher eukaryotes. We turned to a genetic approach, testing the ability of overexpressed OPA1 to promote mitochondrial tubulation in MEFs deficient for either Mfn1 or Mfn2. Expression of OPA1 induced mitochondrial tabulation and fusion in wt and in Mfn2-/- but not in Mfn1-/- cells. This defect was complemented by re-introduction of MFN1 but not MFN2, unequivocally identifying outer membrane MFN1 as an essential functional partner of OPA1. Moreover, MFN1 was unable to promote mitochondrial elongation if OPA1 had been ablated. Thus, OPA1 and MFN1 appear to functionally depend one on each other. To address whether Mfn1-/- MEFs displayed any defect in the preparatory events of mitochondrial juxtaposition and docking, we performed 4D-imaging of mitochondria, i.e. time series of z-stacks of mitochondrial images. The total number of contacts between mitochondria was not affected by OPA1 overexpression or by MFN deficiency. OPA1 facilitated fusion following contacts between wt and Mfn2-/- but not Mfn1-/- mitochondria. Taken together, our results suggested that OPA1 requires MFN1 to fuse the membranes of two juxtaposed mitochondria and not to produce inter-mitochondrial contacts. Our genetic analysis provided the first evidence of a functional diversity between MFN1 and MFN2, suggesting a functional axis between OPA1 and MFN1 (Cipolat et al., 2004). The discovery that OPA1 is a pro-fusion protein raised the question of whether this protein participated in the regulation of apoptosis, during which fusion is impaired. We therefore decided to genetically dissect the role of OPA1 in fusion and apoptosis. We could demonstrate that OPA1 has an antiapoptotic activity, controlling the cristae remodelling pathway of apoptosis, independently of mitochondrial fusion. OPA1 did not interfere with the activation of the core mitochondrial apoptotic pathway of BAX and BAK activation. Yet OPA1 inhibited the release of cytochrome c by preventing the remodelling of the cristae and the intramitochondrial redistribution of cytochrome c. Inactivating mutations in the GTPase domain of OPA1 impaired its anti-apoptotic activity, enhancing susceptibility to apoptosis induced by stimuli that recruit the mitochondrial pathway. While our results contributed to clarify the biological function of OPA1, they left open a number of questions. In particular, if the pro-fusion activity of OPA1 was dispensable for the inhibition of apoptosis, how was this function controlled? In yeast Mgm1p is processed by the inner mitochondrial membrane rhomboid protease Rbd1/Pcp1 into a short active form, responsible for the effects of Mgm1p on mitochondrial morphology (Herlan et al., 2003; McQuibban et al., 2003). The mammalian orthologue of Rbd1p, PARL, could similarly play a role in the regulation of one of the two biological functions we ascribed to OPA1, i.e. its effect in mitochondrial fusion and its anti-apoptotic activity. In order to address this issue, we decided to analyze the phenotype of a mouse model of Parl deletion. Parl-/- mice were born with normal Mendelian frequency and developed normally up to 4 weeks. From then on, mice displayed severe growth retardation and progressive atrophy in multiple tissues, leading to cachexia and death. The atrophy of Parl-/- tymi, spleens and muscular tissues was caused by an increased apoptosis of double-positive (CD4+CD8+) thymic lymphocytes, splenic B lymphocytes (B220+) and myoblasts, respectively. We investigated to what extent mitochondrial dysfunction and morphology dysregulation contributed to this multisystemic atrophy. PARL was not required for normal mitochondrial function: Parl-/- mitochondria did not display primary respiratory defects or latent mitochondrial dysfunction in hepatocytes, MEFs, primary myocytes and myotubes. Mitochondrial dysfunction therefore did not explain Parl-/- muscular atrophy and multisystem failure. Moreover Parl was not required for maintenance of mitochondrial shape and fusion, even in tissues severely affected by Parl ablation like muscle, and Parl was dispensable for regulation of mitochondrial dynamics by OPA1. We therefore investigated whether PARL regulates mitochondrial apoptotic machinery by analyzing apoptosis in MEFs treated with different intrinsic mitochondria utilizing stimuli. Parl-/- MEFs were more sensitive to all the stimuli tested as compared to their wt counterparts. Reintroduction of a catalytically active PARL showed that the defect was specific. PARL exerted its antiapoptotic effect at the mitochondrial level, since cytochrome c release and mitochondrial dysfunction following treatment with an apoptotic stimulus occurred faster in Parl-/- fibroblasts than in their relative wt counterparts. PARL did not regulate activation of the core BAX, BAK dependent apoptotic pathway, but it was required to keep in check the cristae remodelling pathway and to prevent mobilization of the cristae stores of cytochrome c during apoptosis. Since these results pointed to a role for PARL in the cristae remodelling pathway, regulated by OPA1, we ought to understand whether OPA1 required PARL to regulate apoptosis. OPA1 protected wt but not Parl-/- MEFs from apoptosis; furthermore, expression of OPA1 in Parl-/- MEFs did not reduce cytochrome c release, or mitochondrial depolarization following intrinsic stimuli. When Opa1 was silenced by siRNA in Parl-/- cells, they were no longer rescued by re-expression of PARL, demonstrating that PARL is genetically positioned upstream of OPA1. This genetic interaction was confirmed at multiple levels, since PARL and OPA1 interacted in a yeast two-hybrid and co-immunoprecipitation assays. PARL participated in the production of a soluble, IMS located, "anti-apoptotic" form of OPA1. The catalytic activity of PARL was required for the efficient production of soluble OPA1 and the re-introduction of a form of OPA1 in the IMS rescued the pro-apoptotic phenotype of Parl-/- cells. Thus, this IMS form resulted pivotal in controlling the pathway of cristae remodelling and cytocrome c redistribution. IMS and integral IM OPA1 indeed were both found to participate in the assembly of OPA1-containing oligomers that are early targets during cristae remodelling and greatly reduced in Parl /- mitochondria. The reduced level of OPA1 oligomers could account for the faster remodelling and cytochrome c mobilization observed in the absence of PARL. OPA1 affects complex cellular functions other than apoptosis, as substantiated in overexpression studies showing a role for this protein in movement of leukocytes (Campello et al., 2006) and formation of dendritic spines (Li et al., 2004). Furthermore, Opa1 knockout mice demonstrated that OPA1 is required for embryonic development. Homozygous mutant mice die in uterus at 13.5 dpc, with first notable developmental delay at E8.5 (Alavi et al., 2007). We therefore reasoned that levels of OPA1 are likely to affect development and function of multiple organs, by regulating mitochondrial fusion or apoptosis. In the last part of this Thesis, we therefore decided to study whether ablation of OPA1 influences differentiation of embryonic stem (ES) cells in vitro using a hanging-drop differentiation system. To this end, we analyzed an ES cell line where Opa1 had been gene trapped (Opa1gt), resulting in an Opa1+/- genotype. We compared the differentiation potential into cardiomyocytes and neurons of this Opa1gt ES cell line to its relative wt ES cell line. Opa1gt ES cells displayed a decreased capacity to differentiate into beating cardiomyocytes, while they retained a normal neuronal differentiation potential. These preliminary results indicate that OPA1 is a good candidate to regulate differentiation of ES cells in vitro. We now aim at understanding the molecular mechanism by which levels of OPA1 influence differentiation into cardiomyocytes. In conclusion, the data presented in this Thesis demonstrate genetically distinct roles of the mitochondrial dynamin related protein OPA1 in the regulation of organellar shape and apoptosis. The individuation that the functional axis between OPA1 and MFN1 (that regulates mitochondrial fusion) and the regulatory IMM network comprised of the couple substrate-protease Parl-Opa1 could perhaps even control embryonic differentiation opens novel, unexpected avenues to investigate the role of mitochondria in life and death of the cell.