Mitochondrial division, fusion and degradation

Abstract

The mitochondrion is an essential organelle for a wide range of cellular processes, including energy production, metabolism, signal transduction and cell death. To execute these functions, mitochondria regulate their size, number, morphology and distribution in cells via mitochondrial division and fusion. In addition, mitochondrial division and fusion control the autophagic degradation of dysfunctional mitochondria to maintain a healthy population. Defects in these dynamic membrane processes are linked to many human diseases that include metabolic syndrome, myopathy and neurodegenerative disorders. In the last several years, our fundamental understanding of mitochondrial fusion, division and degradation has been significantly advanced by high resolution structural analyses, protein-lipid biochemistry, super resolution microscopy and in vivo analyses using animal models. Here, we summarize and discuss this exciting recent progress in the mechanism and function of mitochondrial division and fusion.

Mitochondria are dynamic organelles that grow, divide and fuse in essentially all types of eukaryotic cells. These dynamic processes govern the size, number and shape of mitochondria, which in turn control the function, distribution and turnover of this essential organelle (1–8). Mitochondrial division and fusion are mediated by three dynamin-related GTPases. Dynamin-related protein 1 (Drp1) mediates mitochondrial division, while Optic atrophy 1 (Opa1) and mitofusin 1/2 mediate mitochondrial fusion. In addition to these proteins, another dynamin superfamily protein, dynamin-2, has recently been shown to mediate mitochondrial division along with Drp1. Drp1 is a soluble protein and is recruited to mitochondria via interactions with multiple receptor proteins that are integral to the mitochondrial outer membrane, such as mitochondrial fission factor (Mff), mitochondrial dynamics protein of 49 and 51 (Mid49/51) and mitochondrial fission 1 protein (Fis1) (1–4, 9). These receptor proteins appear to be structurally different and may regulate Drp1 in distinct ways. In contrast to Drp1, Opa1 and mitofusin 1/2 have transmembrane domains and are inserted into the inner membrane and the outer membrane, respectively. Opa1 mainly acts on inner membrane fusion while mitofusins predominantly function in outer membrane fusion. All of the three GTPases and Drp1 receptor Mff are mutated in human diseases, which indicates the importance of mitochondrial division and fusion to human health (10). Mutations in Drp1 cause neurodevelopmental and degenerative disorders as well as hypotonia (11–18), while mutations in Mff are associated with early-onset Leigh-like basal ganglia disease and epileptic encephalopathy (19, 20). Mitofusin 2 is mutated in Charcot-Marie-Tooth type 2 A and Opa1 is mutated, as its name suggests, in dominant optic atrophy (21, 22). In addition to these inherited diseases, pathological changes in mitochondrial division and fusion have been linked to other diseases and normal aging processes. Here, we discuss recent progress in understanding the mechanisms and function of mitochondrial division and fusion. In addition, we will review the emerging evidence that supports a role for mitochondrial division and fusion in the degradation of mitochondria through mitophagy, in vivo.

Mitochondrial Division

The core reaction of mitochondrial division is the constriction and scission of the two mitochondrial membranes. This reaction is driven by the recruitment of Drp1 to mitochondria by its receptor proteins that are located in the outer membrane (1–4, 9) (Fig. 1). Drp1 oligomerizes into filaments that wrap around mitochondria. Using GTP hydrolysis, the conformational changes of Drp1 oligomers constrict and cut the mitochondrial membrane. In recent years, additional critical components that regulate mitochondrial division, besides Drp1 and its receptor proteins, have been identified. These include another dynamin GTPase, dynamin-2, the actin cytoskeleton and lipids in the mitochondrial membrane. Below, we discuss these new players in the membrane remodelling process.

Fig. 1.

Mechanism and regulation of mitochondrial division. Mitochondrial division is mediated by the mechano-chemical GTPase Drp1, which is recruited to the mitochondrial outer membrane by its receptor proteins. Drp1 is also recruited to the ER by Mff and then transferred to the mitochondria. The actin system, which consists of actin filaments, myosin II and the ER-associated formin INF2, also controls Drp1 recruitment and membrane deformation at ER-mitochondrial contacts. After Drp1 oligomers are assembled on mitochondria, phospholipids, cardiolipin and phosphatidic acid balance the activation and inhibition of mitochondrial division.

Drp1 functions as a critical mechano-chemical GTPase during mitochondrial division. The loss of Drp1 blocks mitochondrial division and elongates or enlarges mitochondria. Supporting its direct role in mitochondrial division, Drp1 is located at sites of mitochondrial division. However, until very recently, it was unclear whether Drp1 alone is sufficient for membrane scission. Structural analysis of Drp1 has shown that the diameter of the Drp1 spiral is larger than that of the endocytic dynamins and appears to be insufficient to cut the membranes by itself (23, 24). Importantly, a recent study by Lee et al. (2016) (25) reported that dynamin-2, which is known to function in severing the plasma membrane during endocytosis, is recruited to mitochondrial division sites. Knockdown of dynamin-2 was found to inhibit mitochondrial division and apoptosis, revealing its functional importance for division. However, two subsequent studies have challenged the role of dynamin-2 in mitochondrial division. Kamerkar et al. (2018) (26) showed that although cells that lack dynamin-2 had dense mitochondrial networks, when the connectivity of the mitochondrial matrix was directly measured using FRAP, dynamin-2 knockout cells did not have an increase in connectivity, which is different from Drp1 knockout cells. Along the same lines, Fonseca et al. (2019) (27) examined knockout cells that lack dynamin-1, 2 and 3. In the triple knockout cells, mitochondria normally divided, suggesting that these dynamin proteins are not essential for mitochondrial division. If Drp1 directly mediates mitochondrial division, can this protein sever membranes? Previous studies have failed to directly show this activity for Drp1. The same study by Kamerkar et al. (2018) (26) used membrane tubules that are similar in diameter to mitochondrial tubules and showed that purified Drp1 cuts membranes in vitro. Supporting this fission activity of Drp1, Kalia et al. (2018) (28) reported that after dissociation from its receptor protein Mid49, Drp1 forms a ring structure with a diameter of 16 nm, which is smaller than was previously reported using cryo-electron microscopy. In this reaction, GTP hydrolysis dissociates Drp1 from co-polymerized Mid49 to trigger a conformation change into the narrow ring structure. In future studies, determining the exact role of Drp1 and dynamin-2 in mitochondrial division and their functional relationship will be important.

Drp1 has previously been shown to assemble on mitochondria through interactions with Drp1 receptor proteins located in the outer membrane, such as Mff, Mid49/51 and Fis1 (1–4, 9) (Fig. 1). Surprisingly, a recent study by Ji et al. (2017) (29) reported that a subset of Mff is also located in the ER membrane and recruits Drp1 to the ER. On the ER, Drp1 is oligomerized and then transferred to the mitochondrial outer membrane, likely through ER-mitochondria contact sites, to mediate mitochondrial division. Supporting the functional importance of the newly identified ER localization, artificially tethering Mff to the ER membrane was found to stimulate mitochondrial division (29). It is now important to delineate the extent to which these two populations of Drp1, which are assembled on the ER and mitochondria, contribute to mitochondrial division. Another key question is whether and how these assembly sites for Drp1 are regulated or selected by different types of cellular, physiological or pathological cues.

At the ER-mitochondria contact sites, the actomyosin system assembles and promotes mitochondrial division (1, 4) (Fig. 1). Non-muscle myosin II has been shown to recruit Drp1 to ER-mitochondrial contact sites along with ER-associated, actin nucleation factor formin, INF2, which induces the formation of actin filaments (30–33). A recent study by Yang and Svitkina (2019) (34) examined the ultrastructure of the actomyosin system at mitochondrial constriction sites using platinum replica electron microscopy. The authors found that actin filaments form an interstitial cytoskeletal network with a criss-cross array of long filaments at ER-mitochondria contact sites. Myosin II is also located near mitochondria and is recruited to constriction sites when mitochondrial division is induced. This study suggested that myosin II produces deformations of the interstitial actin network, which physically pushes mitochondria and thus, initiates curvature-sensing mechanisms that drive scission of mitochondria. Interestingly, live-cell imaging of actin filaments by Moore et al. (2016) (35) showed that actin cycles through mitochondria in cells and its focal assembly on mitochondria correlates with mitochondrial division. This actin cycling is dependent on both Arp2/3 and formin. It would be interesting to further test how these actin dynamics orchestrate the recruitment of Drp1 to ER-mitochondria contact sites through its receptor proteins and the conformational changes of Drp1 and the deformation of mitochondrial membranes in order to control the time and place of mitochondrial division.

Drp1 is located on the surface of the mitochondrial outer membrane. It remains to be determined whether the inner membrane is passively divided by Drp1 following the outer membrane division, or whether the inner membrane has separate machinery that works together with Drp1-mediated outer membrane division. Studies by Lee and Yoon (2014) (36), Chakrabarti et al. (2018) (37) and Cho et al. (2017) (38) reported that the mitochondrial inner membrane constricts at ER-mitochondria contact sties independently of the Drp1-dependent constriction of the outer membrane. This inner membrane constriction is also controlled by actin filaments formed by INF2 (37). The actin system at ER-mitochondria contact sites increases levels of calcium in the mitochondrial matrix through the mitochondrial calcium uniporter (37). Mitochondrial calcium in turn stimulates the constriction of the inner membrane. The inner membrane constriction precedes the outer membrane constriction and is regulated by Opa1, another dynamin-related protein that controls mitochondrial fusion and inner membrane morphology. It has been proposed that Opa1, which is cleaved by the mitochondrial Oma1 protease, interacts with the MICOS complex that connects the outer membrane and inner membrane (38). This binding of Opa1 inhibits the MICOS complex and dissociates these two membranes, facilitating the inner membrane constriction (38). The activation of Oma1 is stimulated by decreases in the membrane potential across the inner membrane. Increased levels of calcium in the matrix likely opens calcium-dependent potassium channels and dissipates the membrane potential (38). Interestingly, in the absence of Drp1, the inner membrane constriction is enhanced, suggesting that cells sense the absence of mitochondrial division and attempt to compensate by increasing the inner membrane constriction (38).

In addition to its receptor proteins and the actin cytoskeleton, mitochondrial lipids also regulate mitochondrial division (2) (Fig. 1). It has been shown that phospholipids in the mitochondrial membrane control the activity of Drp1 in stimulatory and inhibitory manners. A mitochondria-generated phospholipid, cardiolipin, binds Drp1 and stimulates mitochondrial division by promoting its assembly and thereby, assembly-stimulated GTPase activity (39–44). Cardiolipn has also been shown to control Drp1 in apoptosis (39). In contrast to the stimulatory role of cardiolipin, phosphatidic acid, which is a precursor lipid of many phospholipids, including cardiolipin, interacts with Drp1 and restrains its function (45, 46). Interactions with phosphatidic acid inhibit, after Drp1 is oligomerized, on the mitochondrial outer membrane (45). Therefore, phosphatidic acid likely creates a priming step for mitochondrial division at which Drp1 is stalled and after it is assembled into the division machinery. On mitochondria, the pre-assembled division machinery may rapidly induce division in response to cellular and physiological stimulations, which remove phosphoatidic acid in the outer membranes. As a potential mechanism that controls phosphatidic acid levels in the vicinity of the mitochondrial division machinery, Drp1 forms a complex with a mitochondrial phospholipase D, MitoPLD (47), which creates phosphatidic acid from cardiolipin (45). MitoPLD could function as a key regulator that converts stimulatory cardiolin and inhibitory phosphatidic acid in the outer membrane. It would be exciting to further decipher how the Drp1 receptors, the actin cytoskeleton and phospholipids are coordinated to control mitochondrial division.

Drp1 also divides peroxisomes (48). Mff is the major Drp1 receptor on peroxisomes. The loss of Drp1 or Mff elongates peroxisomes, similar to mitochondria. Beside these intracellular organelles, Li et al. (2013) (49) has shown that Drp1 and Mff control the formation of endocytic vesicles at the plasma membrane in the presynaptic terminus. The loss of Drp1 leads to the formation of abnormally enlarged endocytic vesicles. Furthermore, Itoh et al. (2018 and 2019) (50, 51) identified a brain-specific isoform of Drp1. The gene encoding Drp1 has four alternative exons and creates multiple isoforms. The longest isoform that contains all of the alternative exons is specifically expressed in the brain and its expression is induced during postnatal brain development. This isoform is enriched in the postsynaptic terminus and controls clathrin-mediated endocytosis at the postsynaptic terminus, independently of mitochondrial or peroxisome division (51). The loss of this isoform accumulates clathrin-coated pits during early stages, suggesting that this Drp1 isoform regulates the progression of endocytic vesicle maturation, rather than the final scission process (51). Although its specific cargos remain to be identified, the loss of this Drp1 isoform creates more dendrites in multiple types of neurons in vivo and in vitro, in parallel to increases in the ability of sensorimotor gating behaviour in mice (51). When this Drp1 isoform is expressed in fibroblasts, it is associated with the plasma membrane, endocytic vesicles and lysosomes, consistent with its function in neurons (50). It is now important to understand which exact step in endocytosis is regulated by Drp1 in neurons.

Mitochondrial Fusion

Mitochondrial fusion is mediated by Opa1 and mitofusin 1 and 2. Unlike Drp1 and dynamin-2, these GTPases are integral to the mitochondrial membrane with one or two transmembrane segments (1–4, 9). While Opa1 is located in the inner membrane, mitofusin 1/2 are located in the outer membrane. Despite their distinct membrane localization, Opa1 and mitofusin 1/2 form protein complexes and together achieve stable fusion of the mitochondrial membranes (45, 52). Similar to mitochondrial division, which often takes place at ER-mitochondrial contact sites (53, 54), a recent study by Guo et al. (2018) (55) revealed that mitochondrial fusion also occurs at ER-mitochondria contact sites using multi-colour super resolution microscopy called grazing incidence structured illumination microscopy. These data are consistent with a previous report that mitochondrial division and fusion frequently happen at the same region in cells, creating hot spots (56). It would be interesting to determine how the ER manages to control the two opposite reactions, mitochondrial division and fusion.

Although both Opa1 and mitofusins belong to the same dynamin superfamily, the mechanisms by which these proteins merge membranes are fundamentally different. While mitofusins mediate membrane fusion through homotypic protein–protein interaction (e.g. interactions between mitofusin in the opposite membranes), Ban et al. (2017) (57) showed that Opa1 fuses membranes through heterotypic Opa1-cardiolipin interaction (e.g. Opa1 in one membrane and cardiolipin in the other membrane). This Opa1-mediated membrane fusion likely involves the insertion of a part of Opa1 into the other membrane. In virus-host fusion events during infection, viral fusogenic proteins insert its small portion, called a fusion peptide, into the host membrane to drive membrane fusion (58). Although an equivalent fusion peptide has not been identified in Opa1, viral fusion proteins and Opa1 may share a common mechanistic principle to fuse membranes. Rather than fusion, homotypic interactions between Opa1 have been proposed to create the crista structure of the mitochondrial inner membrane (57).

Mitofusins 1 and 2 are homologous proteins, both of which mediate mitochondrial fusion. Compared to the individual loss of either mitofusin, the concomitant loss of both leads to stronger defects in mitochondrial fusion (59). Mitofusin 2, but not mitofusin 1, is mutated in Charcot-Marie-Tooth type 2 A (CMT2A), which causes degeneration of peripheral nerves (21, 60). Although its underlying disease mechanism is currently unknown. Recent studies by Franco et al. (2016) (61) and Rocha et al. (2018) (62) reported that a short peptide and a small chemical, both of which modulate the conformation of mitofusin 2 and stimulate its fusion activity, rescued defects in mitochondrial morphology from fragments to tubules in fibroblasts and neurons carrying CMT2A mutations. In addition to mitochondrial fusion, mitofusin 2 mediates the intracellular transport of mitochondria by connecting mitochondria to the Miro/Milton/kinesin motor complex (63). Neurons carrying CMT2A disease mutations in mitofusin 2 decrease mitochondrial motility. Interestingly, the small molecule that promotes mitochondrial fusion through mitofusin 2 was also found to restore axonal trafficking of mitochondria (62). These exciting studies open a new research avenue that may lead to the development of therapeutic interventions for CMT2A neuropathy. The molecular mechanism for these treatments has been shown to involve the release of intramolecular interactions between two coiled-coil heptad-repeat regions 1 and 2 (HR1 and 2) and leading to the opening of the mitofusin 2 conformation. However, using bioinformatic and biochemical analyses, Mattie et al. (2018) (64) reported a new topology of mitofusin 1 and 2—HR1 is located in the cytosol while HR2 is present in the intermembrane space of the mitochondria in both mitofusins 1 and 2 and therefore, HR1 and HR2 do not interact. It is now important to further test whether the topology of mitofusin 2 are regulated and altered to control mitochondrial fusion and transport.

In addition to mitochondrial fusion and transport, mitofusin 2 has a function in tethering mitochondria to the ER to create inter-organelle contacts though this functions has been controversial (65–68). Two studies recently reported new roles of mitofusin 2 through this ER-mitochondria tethering in human diseases. Wang et al. (2018) (69) found that mitofusin 2 controls the transport of calpastatin via its interaction with mitofusin 2-mediated ER-mitochondria contacts. Calpastatin is an inhibitor of calcium-dependent cysteine protease calpain and its transport to axonal termini is important for preventing degradation of neuromuscular junctions (69). Suggesting mitofusin 2 to be a potential target for therapeutic interventions of amyotrophic lateral sclerosis (ALS) which is associated with degeneration of neuromuscular junctions, overexpression of mitofusin 2 protected neuromuscular junctions from degeneration and skeletal muscle atrophy in mouse models of ALS (69). In addition to the function of mitofusin 2 in neurons, Hernandez-Alvarez et al. (2019) (70) have shown that mitofusin 2 is important for lipid transfer between the ER and mitochondria through their contact sites in the liver. Levels of mitofusin 2 were found to be decreased in both human patients with non-alcoholic steatohepatitis (NASH) and mouse NASH models (70). The authors showed that mitofusin 2 binds a phospholipid, phosphatidylserine and extracts this lipid from liposomal membranes in vitro (70). Liver-specific knockout of mitofusin 2 in mice was shown to result in lipid transport defects and ER stress, leading to NASH-like phenotypes and liver cancers.

Role of mitochondrial division and fusion in mitophagy

One of the key functions of mitochondrial division and fusion is to control the size of mitochondria. Previous studies have shown that the size of mitochondria is critical for the degradation of mitochondria via mitophagy (71–76). As described below, recent publications have reported that mitophagy controls the maintenance and inheritance of mitochondrial DNA (mtDNA) in the maternal germline and embryos. It is thought that oocytes have a mechanism that maintains healthy mitochondria across many generations because mtDNA are subjected to high levels of reactive oxygen species that produce frequent mutations and deletions, due to oxidative phosphorylation in mitochondria (77, 78). Lieber et al. (2019) (79) showed that mitochondria that carry mutated mtDNA are selectively removed by mitophagy in the female germline of Drosophila. This selective removal of mutated mtDNA is critical for the maintenance of wild type mtDNA in oocytes, and this mitophagy is driven by the fragmentation of mitochondria that is caused by decreased levels of mitofusin. In addition to mitofusin, the ATG1 kinase and the mitophagy receptor protein BNIP3L are required for this degradation mechanism (79). To trigger mitophagy, decreases in mitochondrial ATP levels (but not changes in the mitochondrial membrane potential) is critical (79). Because fragmented mitochondria that contain mutated mtDNA likely have decreased ATP levels, the mechanism that senses mitochondrial ATP levels could selectively mark dysfunctional mitochondria. Interestingly, not only necessary, but also induced fragmentation of mitochondria were found to be sufficient to initiate this mitophagy mechanism against mitochondria carrying mutant mtDNA. Therefore, the regulation of mitochondrial size by fusion enables oocytes to sustain and inherit healthy populations of mitochondria in the next generations. It is important that we now test and determine whether similar mechanisms work in mammals, especially humans.

When oocytes fuse with sperm during fertilization, paternal mitochondria are eliminated and maternal mitochondria are maintained. Sato and Sato (2011) and Al Rawi et al. (2011) (80, 81) and Sato et al. (2018) (82) have shown that paternal mitochondria are removed by mitophagy in embryos of Caenorhabditis elegans, while Rojansky et al. (2016) (83) showed a similar mitophagy in mice. In C. elegans, autophagosomes are recruited to mitochondria, which is indicated by the association of the autophagosomal protein LGG-1 (a homolog of LC3) with mitochondria (80). Demonstrating its functional importance, LGG-1 mutants were found to have decreased degradation of paternal mitochondrial in embryos. Different from the Drosophila oocytes described above, the mechanism of the degradation of C. elegans embryos is accompanied by the loss of the membrane potential in paternal mitochondria (82). Paternal mitochondria with decreased membrane potential undergo ubiquitination, and associates with LGG-1 and is subjected to degradation. To promote these processes, a protein kinase, IKKE-1, which is a homolog of mammalian TBK1, phosphorylates and activates an adaptor protein, ALLO-1, which binds to LGG-1 (81). In mouse embryos, two E3 ligases, parkin and MUL1, play redundant roles in the ubiquitination of paternal mitochondria (83). Similar to the mitochondria of C. elegans, paternal mitochondria in mice lose the membrane potential, which likely stimulates the recruitment and/or activation of these E3 ligases (83). To connect ubiquitinated mitochondria with autophagosomes, in mice, p62/SQSTM1 functions as an adaptor protein in this mitophagy mechanism (83). Because the loss of the mitochondrial membrane potential is a critical upstream event in C. elegans and mice, it is important to determine how paternal mitochondria selectively lose their membrane potential after fertilization.

In human patients with NASH, enlarged mitochondria, termed megamitochondria, have been observed in hepatocytes in the liver (84–86) (Fig. 2). Yamada et al. (2018 and 2019) (75, 87) have found that these megamitochondria are similar to large mitochondria produced by the lack of mitochondrial division in Drp1 knockout hepatocytes. The enlarged mitochondria in Drp1 knockout hepatocytes slowed mitophagy due to the extreme size of the mitochondria, but not the absence of mitochondrial division. Additional inhibition of mitochondrial fusion by Opa1 knockout reduced mitochondrial size and restored mitophagy in Drp1-Opa1 double knockout (75). The reduced rate of mitophagy in Drp1 knockout hepatocytes leads to the accumulation of mitophagy intermediates, which are associated with ubiquitin, the adaptor protein p62 and LC3 (74, 86). The large size of mitochondria likely lowered the efficiency of engulfment of mitochondria by autophagosomes, and revealed otherwise undetected intermediates. Hepatocytes of NASH mouse models generated megamitochondria, similar to hepatocytes of human NASH patients and hepatocytes of Drp1 knockout livers. These megamitochondria also built-up ubiquitin, p62 and LC3 (74, 86). Importantly, decreasing the size of megamitochondria cleared the accumulation of mitophagy intermediates and rescued liver damage in NASH mouse models (75). These findings suggest that re-initiating mitophagy by blocking mitochondrial fusion provides a potential medical treatment for NASH in humans. As described above, p62 has been suggested to function as an adaptor protein in mitophagy by connecting ubiquitinated mitochondria and autophagosomes through direct interactions with both ubiqutin and LC3 (88, 89). However, surprisingly, p62 was found to be required for ubiquitination of megamitochondria; p62 recruited a ubiquitin ligase complex containing the scaffold protein KEAP1 and the E3 ligase Rbx1 to mitochondria through direct interactions with KEAP1 (74). The p62-KEAP1-Rbx1 complex provides a new mechanism that labels damaged mitochondria for mitophagy. It should be noted that this ubiquitination mechanism is independent of the E3 ligase parkin and its activating kinase PINK1 (75, 87). Instead, parkin and PINK1 were found to specifically control the degradation of mitofusin 1 and 2 when megamitochondria are formed in Drp1 knockout hepatocytes (75, 87). Decreasing levels of mitofusins may decrease mitochondrial fusion. It appears that cells attempted to oppose mitochondrial fusion when mitochondrial division is blocked in order to readjust the balance of mitochondrial dynamics using the parkin-PINK1 mechanism. It is important to delineate how p62 senses mitochondrial defects in megamitochondria.

Fig. 2.

The formation of megamitochondria in the liver. In wild-type mitochondria, mitochondrial shape is regulated by the balance between mitochondrial division and fusion. Drp1 loss or NASH induces the formation of megamitochondria. Megamitochondria are suppressed by decreasing mitochondrial fusion via Opa1 loss. Restoring the mitochondrial size to normal levels rescues liver damage.

Although defects in parkin or PINK1 have been linked to both sporadic and genetic Parkinson’s disease, in vivo function of these proteins in mitophagy are largely unknown (74, 75, 90–92). Recent studies by Matheoud et al. (2016 and 2019) (93, 94) have shown that parkin and PINK1 regulate the formation of mitochondria-derived vesicles that are produced by budding off the mitochondrial membranes. Interestingly, parkin and PINK1 promote MDVs that are transported to peroxisomes and lysosomes, while they inhibit mitochondria-derived vesicles targeted to endosomes for mitochondrial antigen presentation in response to bacterial lipopolysaccharides that induce inflammation (93, 95–97). Mitochondrial antigen presentation brings mitochondrial proteins to major histocompatibility complex class I molecules on the surface of antigen-presenting cells (98). Mice lacking PINK1 were found to have increased mitochondrial antigen presentation upon intestinal infection by Gram-negative bacteria and showed Parkinson disease-like phenotypes such as motor impairment and dopaminergic axonal varicosities (94). Parkinson’s disease is associated with inflammation, however, the underlying mechanisms are mostly unknown. These recent studies provide a new mechanistic link of the autoimmune responses to the pathogenesis of Parkinson’s disease. Another study by Sliter et al. (2018) (99) also showed an important connection between inflammation and Parkinson’s disease through parkin and PINK1. These authors found that the lack of parkin or PINK1 increased STING-mediated inflammation in muscles when mice were subjected to exhaustive exercise. Because parkin and PINK1 knockout mice do not show clear phenotypes related to Parkinson’s disease under normal physiological conditions (100), the additional stress may reveal critical insights into the function of these proteins and their defects in Parkinson’s disease.

Concluding Remarks

In recent years, new and exciting mechanisms underlying the regulation and function of mitochondrial division and fusion have been revealed. These findings have helped us to integrate mitochondrial division and fusion in different cellular, physiological and pathological contexts. Mitochondrial division and fusion play important roles in several areas of the life sciences that range from brain development, neuronal plasticity, immune responses, both in the brain and muscle and lipid metabolism in the liver. As we have discussed above, these studies raise further important questions regarding these dynamic membrane processes. We expect that understanding the fundamental biology of mitochondrial division and fusion will provide novel insight into human health and disease.