Mitofusins: Disease Gatekeepers and Hubs in Mitochondrial Quality Control by E3 Ligases

Abstract

Mitochondria are dynamic organelles engaged in quality control and aging processes. They constantly undergo fusion, fission, transport, and anchoring events, which empower mitochondria with a very interactive behavior. The membrane remodeling processes needed for fusion require conserved proteins named mitofusins, MFN1 and MFN2 in mammals and Fzo1 in yeast. They are the first determinants deciding on whether communication and content exchange between different mitochondrial populations should occur. Importantly, each cell possesses hundreds of mitochondria, with a different severity of mitochondrial mutations or dysfunctional proteins, which potentially spread damage to the entire network. Therefore, the degree of their merging capacity critically influences cellular fitness. In turn, the mitochondrial network rapidly and dramatically changes in response to metabolic and environmental cues. Notably, cancer or obesity conditions, and stress experienced by neurons and cardiomyocytes, for example, triggers the downregulation of mitofusins and thus fragmentation of mitochondria. This places mitofusins upfront in sensing and transmitting stress. In fact, mitofusins are almost entirely exposed to the cytoplasm, a topology suitable for a critical relay point in information exchange between mitochondria and their cellular environment. Consistent with their topology, mitofusins are either activated or repressed by cytosolic post-translational modifiers, mainly by ubiquitin. Ubiquitin is a ubiquitous small protein orchestrating multiple quality control pathways, which is covalently attached to lysine residues in its substrates, or in ubiquitin itself. Importantly, from a chain of events also mediated by E1 and E2 enzymes, E3 ligases perform the ultimate and determinant step in substrate choice. Here, we review the ubiquitin E3 ligases that modify mitofusins. Two mitochondrial E3 enzymes-March5 and MUL1-one ligase located to the ER-Gp78-and finally three cytosolic enzymes-MGRN1, HUWE1, and Parkin-were shown to ubiquitylate mitofusins, in response to a variety of cellular inputs. The respective outcomes on mitochondrial morphology, on contact sites to the endoplasmic reticulum and on destructive processes, like mitophagy or apoptosis, are presented. Ultimately, understanding the mechanisms by which E3 ligases and mitofusins sense and bi-directionally signal mitochondria-cytosolic dysfunctions could pave the way for therapeutic approaches in neurodegenerative, cardiovascular, and obesity-linked diseases.

Keywords: E3 ligases; ER; MFN1/MFN2; mitochondria; mitofusins; mitophagy; quality control; ubiquitin.

Figure 1

Ubiquitylation cascade. Ubiquitylation of substrates requires a cascade of events involving three enzymes: an E1 ubiquitin-activating enzyme, an E2 ubiquitin-conjugating enzyme, and an E3 ubiquitin ligase. First in this cascade, the E1 enzyme activates ubiquitin and transfers it to the E2 enzyme in an ATP-dependent manner with which ubiquitin is conjugated. Afterward, the ubiquitin molecule is transferred from the E2 enzyme to the specific target substrate by the E3 ligase enzymes, which either actively receives ubiquitin from E2 and then transfers it to the substrate or serves as a binding platform between the E2 and the substrate. Finally, on the target substrate, mono, mono-multi, or polyubiquitylation can occur.

Figure 2

Mitochondrial roles and post-translational modifications of mitochondrial dynamic proteins. Mitochondria are involved in several cellular processes other than ATP production, such as calcium (Ca²⁺) buffering, phospholipid biosynthesis, iron-sulfur clusters (Fe-S) assembly from iron (Fe²⁺), regulation of programmed cell death via cytochrome c (Cytc c) release, and mitophagy. Mitochondrial proteins known for their role in mitochondrial dynamics, such as Drp1, Mfn1/2, and Opa1, are a target for several post-translational modifications, which differently regulate their level and, hence, function. Drp1, on the OMM, can be modified by sumoylation, phosphorylation, acetylation, ubiquitylation, and S-nitrosylation. Mitofusins, also on the OMM, can be a target for phosphorylation, ubiquitylation, or acetylation. Opa1, in the IMM, can suffer processing or be modified through acetylation.

Figure 3

Structure and topology models of mitofusins. (A) Linear structure of mitofusin, with the GTPase domain locating at the N-terminal, one hydrophobic heptad repeat (HR1), the transmembrane anchor(s), and a second hydrophobic heptad repeat (HR2). (B) Crystal structure of MFN1 and MFN2 modeled on BDLP and mini-MFN1, according to the first topology proposed, with two transmembrane domains and both the N- and C-terminus facing the cytosol (Rojo et al., 2002; Low and Löwe, 2006; Low et al., 2009; Qi et al., 2016; Cao et al., 2017). (C) Structural scheme of MFN1 and MFN2 according to the second topology proposed with a single spanning-membrane domain, instead of two, and the C-terminus residing in the IMS and not facing the cytosol (Mattie et al., 2018).

Figure 4

Mitochondrial morphology upon knockout of Mitofusin 1 or 2 and its disease-associated roles. Although extremely similar in their sequence and structure, each mitofusin ablation leads to strikingly different mitochondrial morphologies. Mitofusin 1 knockout gives origin to a highly fragmented mitochondrial network composed of many small fragments, whereas depletion of Mitofusin 2 leads to a network where mitochondrial fragments are found enlarged and aggregated in clusters, commonly perinuclearly organized. Several diseases have been associated with knockout of Mitofusin 1 or 2. Homozygous knockout of Mitofusin 1 or 2 leads to embryonic defects such as defective giant cell layer and leads, ultimately, to lethality. While Mitofusin 1 only appears to have an effect at the embryonic level, knockout of its homologue protein, Mitofusin 2, has been shown to relate to several other defects. Animal models depleted for Mitofusin 2 display severe cardiac defects such as cardiomyocyte dysfunction, rapid progressive dilated cardiomyopathy, and final heart failure. Moreover, Mitofusin 2 mutations are the primary cause of the incurable neuropathy Charcot-Marie Tooth Type 2A for which no disease-underlying functions have been yet identified. Additionally, links with other neurodegenerative diseases such as Alzheimer’s and Parkinson’s diseases have been made, although the molecular mechanisms underlying it are not fully understood. Low levels of Mitofusin 2 were also shown to have a strong positive correlation with diabetes type 2 and obesity.

Figure 5

Roles of mitochondria-ER contacts. The physical contacts established between mitochondria and ER are responsible for several cellular processes such as mitochondrial fission, calcium buffering, and phospholipid synthesis. During mitochondrial fission, ER tubules are found in contact with mitochondria in the future fission sites. Drp1 is recruited to these sites and, together with ER tubules, promotes constriction and fission of mitochondria. Calcium (Ca²⁺) is transported from the ER to mitochondria via transporters in each membrane. On the ER membrane, Ca²⁺ is exported via the inositol 1,4,5-triphosphate receptor (I3PR) to the voltage-dependent anion-selective channel (VDAC) on the mitochondrial outer membrane. For the synthesis of some phospholipids, the required enzymes are found in the mitochondria and not in the endoplasmic reticulum (ER). This implies the transfer of precursor forms of phospholipids from the ER to mitochondria where they can be modified and then re-transferred to the ER. For example, the production of the mitochondrial phospholipid cardiolipin (CL) depends on precursor lipids present at the ER. Phosphatidic acid (PA) is transferred from the ER, across the IMS to the IMM, where it is enzymatically modified to form CL, which is further transported to the OMM. An example of a bidirectional movement of lipids between the ER and mitochondria is the production of phosphatidylethanolamine (PE) and phosphatidylcholine (PC). Phosphatidylserine (PS) is first produced on the ER membrane and then translocated to the OMM. On the OMM, PS is transferred to the IMM where it is enzymatically modified to form PE. Finally, in order to produce PC, the precursor PE must be transferred back to the ER where specific enzymes modify it into PC.

Figure 6

Apoptotic intrinsic pathway mediated by mitochondria. The programmed and regulated cell death, apoptosis, can occur via two different pathways—intrinsic or extrinsic—according to the origin of the death stimuli, whether it is intrinsic or extrinsic to the cell. Upon intrinsic death stimuli, such as, for example, DNA damage or oncogene activation, the intrinsic apoptotic pathway is activated, which is mediated by mitochondria. Intrinsic stimuli induce the oligomerization of a pro-apoptotic BcL-2 protein—BAX. These oligomers are able to permeabilize the mitochondrial membrane by pore formation on the OMM. Membrane permeabilization allows the release of pro-apoptotic molecules from the IMS, importantly, cytochrome c. In a complex together with other pro-apoptotic proteins, cytochrome c activates caspases, the effectors of apoptosis.

Figure 7

Mitochondrial clearance via the ubiquitin-mediated pathway. Upon mitochondrial depolarization, a cascade of events is initiated, which targets damaged mitochondria, or portions of it, for degradation by the autophagy machinery. In a first step, the kinase PINK1 accumulates at the mitochondrial outer membrane and initiates phosphorylation (P) and recruitment of the E3 ligase Parkin. Activated Parkin ubiquitylates (U) several outer mitochondrial membrane proteins such as Mitofusins 1 and 2. Additional phosphorylation of ubiquitin generates a positive feedback loop increasing Parkin recruitment and further ubiquitylation. The formation of ubiquitin chains on mitochondrial surface proteins promotes its binding to lipidated LC3, an autophagosome receptor, via the mitochondrial receptors (R) such as optineurin, NDP52, or p62. From this point, the mitochondria, or its fragments, meant to be degraded are surrounded by the autophagosome, which finally fuses with the lysosome for final destruction.

Figure 8

E3 ligases that modify mitofusins and cellular processes associated. MARCH5, Parkin, and Gp78 regulate both mitofusins, whereas MGRN1 affects MFN1 and HUWE1 and MUL1 affect MFN2.

Figure 9

Residues, E3 ligases, and processes regulating MFN1 (A) MFN2 (B). Representation of the triggers identified to modify mammalian mitofusins by ubiquitylation, phosphorylation, and acetylation. The enzymes evolved in each case, and the cellular outcome is also depicted. The vertical bar on each side of the structure denotes that the residues modified by ubiquitylation in MFN1 or MFN2 are not known. See text for details.

Figure 10

Pro- and anti-mitophagic roles of mitofusins. Mitofusins are reported to induce and inhibit mitophagy through a variety of processes. Pro-mitophagic roles of mitofusins can be promoted by (1) phosphorylation of MFN2 by PINK1, and subsequent recruitment of Parkin to mitochondria, (2) Gp78-mediated ubiquitylation of MFN1, and (3) SIRT1-mediated deacetylation of MFN2. Anti-mitophagic roles have been described for Mitofusin 2 via (1) MULAN-mediated ubiquitylation, (2) increase of ER-mitochondria contacts directly mediated by MFN2, and (3) mutations in MFN2 associated with Charcot-Marie-Tooth Type 2A (CMT2A) neuropathy.