OPA1, a mitochondrial pro-fusion protein, regulates the cristae remodelling pathway during apoptosis

Mitochondria are essential organelles for the life of the cells since it is the major source of ATP, key molecule for many endoergonic reaction. Recently it has been demonstrated that mitochondrial play a key role in many other cellular processes like Ca2+ signaling and programmed cell death. Following an apoptotic insult mitochondria release cytochrome c and other proteins required in the cytosol for the activation of the effector caspases required for cell demise. What is remarkable about cytochrome c release is that is fast, complete and usually is not associated with mitochondrial swelling. Thanks to the advances in 3D electron microscopy it has been demonstrated that cristae are not just invagination of the inner mitochondrial membrane (IMM) as previously depicted by Palade (Palade, 1952) but rather distinct compartments of it, separated from the inter membrane space (IMS) by tubular narrow cristae junctions. The majority of cytochrome c and the other respiratory chain components are restricted in this compartment. To reach a complete cytochrome c release in the absence of mitochondrial swelling mitochondria remodel their internal structure: individual cristae fuse and tubular narrow cristae junctions widen; this process, defined cristae remodeling is associated with the mobilization of cytochrome c towards the IMS for its subsequent release across the outer mitochondrial membrane (OMM) (Scorrano et al., 2002). The molecular mechanism beyond this dynamic process is not well understood and in the laboratory where I did my doctoral Thesis it has been hypothesized that OPA1, the only dynamin related protein of the IMM (Alexander et al., 2000; Delettre et al., 2000) could control cristae remodeling. Dynamin related proteins are regulators of mitochondrial morphology promoting mitochondrial fusion and fission. To this family belong Mitofusins (MFN) 1 and 2 in the OMM and OPA1 that resides in the IMM. OPA1 is a large GTPase anchored in the IMM, facing the IMS (Olichon et al., 2002; Satoh et al., 2003); it has been shown that in yeast, its orhologue Mgm1p is required for fusion competent mitochondria by the cooperation with a protein of the same family on the OMM called Fzo1p. In our laboratory it has been demonstrated that in mammalian cells OPA1 promotes mitochondrial fusion through one of the two mammaliam orthologue of Fzo1p called MFN1. In 2000 two distinct laboratories demonstrated that mutations in OPA1 gene are the cause of dominant optic atrophy (ADOA), the leading case of inherited blindness in human, characterized by selective death of retinal ganglion cell (RGC) (Alexander et al., 2000; Delettre et al., 2000). The fact the mutation in a mitochondrial protein involved in mitochondrial morphology caused cell death opened a new scenario that corroborates the central position of mitochondria in regulating apoptotic signaling. The aim of my thesis was to analyze the role of OPA1 in mitochondria-dependent apoptosis. We started with a brute force approach by overexpressing OPA1 in murine embryonic fibroblasts (MEFs) and measuring cells viability in response to intrinsic apoptotic stimuli that specifically trigger apoptosis through the mitochondrial pathway. Overexpression of wt OPA1 but not of mutant in the GTPase domain (OPA1K301A) or a truncated mutant in the coiled coil domain (OPAR905*) is able to prevent from apoptosis induced by hydrogen peroxide, staurosporine, etoposide and overexpression of tBID, a BH3 only protein of the Bcl-2 family that promotes cristae remodeling. To confirm that OPA1 antiapoptotic activity was exerted at the mitochondrial level we analyzed two aspects of the mitochondrial dysfunction: cytochrome c release and mitochondrial depolarization. To this aim we overexpressed a mitochondrially targeted red fluorescent protein (mtRFP) as marker of the mitochondrial network and then we immunodecorated cytochrome c with a FITC-conjugated secondary antibody. OPA1 overexpression prevented cytochrome c release in response to intrinsic stimuli while its inactive mutant OPAK301A aggravated cytochrome c release kinetic. We then analyzed another aspect of the mitochondrial dysfunction: mitochondrial depolarization, taking advantage of the potentiometric probe tetramethylrhodamine-methyl ester (TMRM) which mitochondrial fluorescence is proportional to mitochondrial potential. Overexpression of OPA1, but not of its inactive K301A mutant, prevented mitochondrial depolarization induced by intrinsic stimuli, confirming that OPA may prevent from apoptosis at the mitochondrial level by reducing cytochrome c release and mitochondrial depolarization. How can a dynamin related protein prevent from apoptosis? We asked this because when our study was ongoing an intriguing hypotesys emerged: during apoptosis mitochondrial network undergoes irreversible massive fragmentation; this event and apoptotic cristae remodeling are required for complete cytochrome c release. In principle, OPA1 could prevent apoptosis at both of these levels either counteracting mitochondrial fragmentation thanks to its pro-fusion activity or by the regulation of cristae remodeling. To understand at which of these levels OPA1 was exerting its antiapototic activity, we started a genetic approach, overexpressing OPA1 in Mfn1-/-, where OPA1 pro-fusion activity was prejudiced. Overexpression of OPA1 in these cells prevented from apoptosis induced by intrinsic stimuli; in view of the fact that a residual pro-fusion activity of OPA1 could be mediated by the presence of MFN2 we repeated the same experiments in cells in which both mitofusins were ablated (DMF). Also in this conditions OPA1 prevented from apoptosis at the mitochondrial level, slowing down cytochrome c release kinetic. OPA1 has an antiapoptotica function that is independent of its pro-fusion activity on the mitochondrial network. At this point we asked whether OPA1 may have a role on apoptotic cristae remodeling. We generated stable cell lines that stably overexpressed OPA1 and its K301A mutant both in wt and in Mfn1-/- cells and a cell line depleted of OPA1 by short hairpin RNA interference (shOPA1RNAi). We then isolated mitochondria and measured cytochrome c release induced by recombinant caspase 8 cleaved BID (cBID) using a specific ELISA immunoassay. Stable overexpression of OPA1 is able to prevent cytochrome c relase independently of MFN1 while its downregulation dramatically increases its release. Using a specific assay we observed that OPA1 is also able to prevent cytochrome c mobilization from the cristae independently of MFN. These results were confirmed by the fact that overexpression of the OPA1K301A mutant increased cytochrome c mobilization that was almost complete when OPA1 levels were depleted by RNAi. A thorough morphometric analysis of isolated mitochondria from these cell lines, associated with 3D reconstruction of electron microscopy tomography, showed that OPA1 controls cristae morphology and prevents cristae junction widening in response to cBID. To better understand the molecular mechanism through which OPA1 controls cristae remodeling and cristae junctions diameter we based our hypothesis on the possible analogy with vesciculation processes regulated by cytosolic dynamin, where GTPase activity of it mediated mechanoenzimatic constriction of the vesicle collar. Despite this analogy, we should mention that OPA1, unlike dynamin, is located on the inner side of the membrane to be constricted and not on the outside as dynamin complicating the model. First, we analyzed biochemical characteristic of OPA1: gel filtration studies showed that OPA1 is eluted at very high molecular weight fractions (>600 KDa) and in response to cBID incubation it is retrieved in low molecular weight fractions. Parallel studies in our laboratory demonstrated that OPA1 is processed by a rhomboid protease, PARL, into a short form found soluble in the IMS that is responsible for the antiapototic but not of the pro-fusion activity of OPA1. We therefore reasoned that OPA1 could organize into high molecular weight complexes made up at least by the PARL generated soluble form and the membrane bound form of OPA1. To confirm this hypothesis we crosslinked this complex and confirmed the presence of a high molecular weight immunoreactive band for OPA1 that disappear following the mechanical expansion of the cristae induced by osmotic swelling. These crosslinker-stabilized oligomers contain both the soluble and the membrane bound forms of OPA1 as demonstrated by their immunoreactivity for properly tagged and co-expressed forms. The OPA1-containing oligomers is targeted by cBID in a time dependent manner and OPA1 overexpression stabilizes these complexes. We can conclude that OPA1 controls cytochrome c mobilization and cristae remodeling that occurs during apoptosis. This function of OPA1 is independent of MFNs and is correlated to the formation of high molecular weight complexes. The data collected so far on OPA1 antiapoptotic function open a new scenario. First we need to investigate on the molecular composition of these complexes in normal and apoptotic conditions. To this aim we started a biochemical study on OPA1-containing complexes in mitochondria isolated from different genetic background in normal and apoptotic conditions. The proteomic analysis of the proteins eventually found in complex with OPA1 will allow us to comprehend the function and regulation of OPA1 oligomers before and after cell death induction. OPA1 appears as a crucial protein in the apoptotic process; as a confirmation of this, it has been found that OPA1 is highly overexpressed in some lung cancer (Dean Fennel, personal communication); we then asked whether OPA1 could be a target for the development of new drugs that enhance apoptosis in tumor cells. To this aim, we started a collaboration with Stefano Moro from the Department of Medicinal Chemistry of the University of Padova, to generate a library of candidate inhibitors of OPA1 performing a virtual screening of compounds targeted to the GTPase pocket of OPA1 obtained following an homology modeling on the Dyctiostelium Discoideum GTPase domain of Dynamin A. In conclusion, the data presented in this doctoral thesis show that mitochondrial protein OPA1 participates in the regulation of cytochrome c mobilization and cristae remodeling during apoptosis. We demonstrated that OPA1 organizes into high molecular weight complexes which disruption correlates with cristae junction widening. This function is distinct from its role in mitochondrial morphology and this suggest a bifurcation and specialization of OPA1 function during evolution.