Mitochondrial Dynamics Proteins As Emerging Drug Targets

Abstract

The importance of mitochondrial dynamics, the physiological process of mitochondrial fusion and fission, in regulating diverse cellular functions and cellular fitness has been well-established. Several pathologies are associated with aberrant mitochondrial fusion or fission that is often a consequence of deregulated mitochondrial dynamics proteins, however, pharmacological targeting of these proteins has been lacking and challenged by complex molecular mechanisms. Recent studies have advanced our understanding in this area and have enabled rational drug design and chemical screening strategies. We provide an updated overview of the regulatory mechanisms of fusion and fission proteins, their structure–function relationships, and the discovery of pharmacological modulators demonstrating their therapeutic potential. These advances provide exciting opportunities for the development of prototype therapeutics for various diseases.

Drugging mitochondrial dynamics

Mitochondrial structure varies significantly between different types of cells, phases of the cell cycle and stimuli of stress. Mitochondria are highly dynamic organelles and alter their morphology using the physiological process of fusion and fission, respectively (Figure 1A). Fusion and fission are counteracting processes that form a dynamic equilibrium known as mitochondrial dynamics that regulates diverse cellular processes (Box 1). Several reports have highlighted that mitochondrial dynamics are required for healthy development, self-renewal and differentiation of stem cells [1], and overall cellular fitness [2]. Furthermore, several studies using genetic models have established that restoration of deregulated mitochondrial dynamics can decrease the progression of several pathologies or prevent their manifestation. Increasing evidence about the role of mitochondrial dynamics in pathologies revealed that key effector proteins mediating fusion/fission are a novel class of drug targets, triggering screening campaigns for the discovery of pharmacological molecules that modulate their activity. While, pharmacological targeting of mitochondrial dynamics proteins has recently emerged, here we will provide a timely discussion of the recent literature.

Figure 1. Mitochondrial dynamics regulate mitochondrial morphology.

(A) Schematic representation of distinct mitochondrial networks and their implications in mitochondrial fitness. Fused mitochondrial networks increase from left to right, while fragmented mitochondrial increase right to left. (B) Schematic representation of the molecular mechanism that regulates mitochondrial fusion and fission. Oligomerization of MFNs initiates the first step of mitochondrial fusion, which is the fusion of the OMM. Next, the unprocessed form of OPA1 (L-OPA1) regulates the fusion of the IMM and completes mitochondrial fusion by forming homotypic and heterotypic complexes with the processed form of OPA1 (S-OPA1) and cardiolipins. Mitochondrial fission requires focal accumulation of the DRP1 mitochondrial receptors: MFF, FIS1, MID49, and MID51, and the translocation of DRP1 from the cytosol to the OMM. Oligomerization of DRP1 on the OMM promotes the ring-shaped DRP1 polymers that constrict mitochondria and mediate mitochondrial severing.

BOX 1: The multifaceted role of mitochondrial dynamics in cellular physiology.

Mitochondrial dynamics constitute an integral part of the homeostatic mechanism that is commonly used to regulate ATP production in response to alteration in the nutrient availability or energy demands. Previous studies have shown that fragmented mitochondria of proliferating cells are associated with high nutrient availability conditions and reduced capacity to produce ATP, while under low nutrient availability conditions mitochondria need to fuse to sustain ATP production by limiting mitochondrial clearance through mitophagy [65,66]. Interestingly, cancer cells often present a fragmented mitochondrial network and meet their high energetic demands by up-regulating glycolysis, a metabolic pathway capable of producing ATP faster than the mitochondrial OXPHOS, albeit less efficient [2,67]. Importantly, this metabolic switch can uncouple the TCA cycle from OXPHOS and ATP production and allow the use of several metabolites of the TCA cycle in anabolic reactions that are required for cell proliferation [68].

Beyond metabolism, mitochondrial dynamics play a role in other cellular processes such as cell death pathways [69], calcium signaling [70], and antiviral response [28]. Mitochondrial structure is a critical determinant of how mitochondria mediate protein-protein interactions at the OMM, uptake and release of proteins, nucleic acids, regulate the levels of metabolites that are required for biochemical reactions, and mitochondrial quality control [28,65,66,69,71,72]. In respect to cell death, a wealth of reports have shown that mitochondrial fission is required for the execution of early apoptotic events such as cytochrome c release [73]. Moreover, we and others have established that suppression of fusion decreases the threshold for the induction of mitochondrial outer membrane permeabilization (MOMP), a critical event that precedes cytochrome c release and is required for the initiation of apoptosis [2,74]. Although the process of fission positively regulates apoptosis, it has been shown that aberrant mitochondrial fragmentation significantly reduces the capacity of BAX to translocate and oligomerize on the OMM, a crucial event for the induction of MOMP [69]. Taken all together it is tempting to speculate that activation of fission supports the progression of cell death when it is stimulated by the induction of apoptosis, while hyperactivation of fission in advance of apoptosis induction from non-lethal stimuli can attenuate cell death. In respect to antiviral response, mitochondrial fusion is increased upon viral infection and mediates mitochondrial antiviral-signaling protein (MAVS) aggregation at the OMM, a process that is critical for the antiviral signal transduction [28].

This review will provide an updated overview on the molecular mechanisms that mediate mitochondrial fusion and fission with a focus on structural insights. Different classes of chemical probes and pharmacological molecules that alter mitochondrial dynamics will be reviewed with an emphasis on molecules that directly modulate the mitochondrial fusion or fission machinery. As we discuss the pharmacologic strategies that have already been used to target the effector molecules of fusion/fission we will give an outlook on the rising approaches to target those molecules using the current structural insights and/or molecular mechanisms that regulate the activity of these proteins.

Overview of mitochondrial dynamics

Mitochondrial fusion

Fusion of two mitochondria requires two steps as mitochondria have two membranes (Figure 1B). The first step is the fusion of the outer mitochondrial membrane OMM, which is mediated by mitofusin (MFN) 1 and 2 [3], two large GTPases that reside on the OMM [4]. MFNs from adjacent mitochondria undergo conformational activation to oligomerize and form homotypic (MFN1-MFN1 or MFN2-MFN2) or heterotypic (MFN1-MFN2) complexes and subsequently promote the tethering of the OMMs from adjacent mitochondria [5,6]. Knockout of MFNs results in excessive mitochondrial fragmentation [3]. Although MFN1 and MFN2 share high sequence homology, mitochondria from cells that lack MFN1 appear as small spheres of similar size [5]. While mitochondria from cells that lack MFN2 appear as small spheres of variable size or in oval shape [5]. Interestingly, knockout of MFN2 but not MFN1 reduces axon integrity and promotes cell death in dopamine producing neurons, a phenotype that highlights that MFN1 and MFN2 may have different roles in cells and the dependency of cells to these proteins can vary between different cell types [7]. Overexpression of MFNs induces the formation of mitochondrial clusters around the nucleus [8]. On the contrary, overexpression of MFN1 in MFN2 knockout cells or MFN2 in MFN1 knockout cells restores mitochondrial morphology [5,8]. Importantly, post translational modifications (PTMs) of MFNs (e.g. ubiquitylation and phosphorylation) have recently received increasing attention and emerged as a regulatory mechanism that controls fusion [9-11]. In addition to the fusogenic activity of MFNs, PTMs can regulate functions of MFNs non-related to fusion. For instance, ubiquitylation of MFN2 initiates mitochondrial clearance through mitophagy [12].

Fusion of the inner mitochondrial membrane IMMs is mediated by optic atrophy 1 (OPA1), a large GTPase that belongs to the dynamin family of proteins [13] (Figure 1B). Studies have shown that OPA1 forms homotypic (OPA1-OPA1) and heterotypic complexes (OPA1-cardiolipin) complexes upon the fusion of the OMM to mediate fusion of the IMM [14-16]. OPA1 has eight mRNA splice variants in humans and four in mice that are expressed in different tissues and two isoforms that differ in length, a short one S-OPA1 and a long one L-OPA1 [17]. OPA1 has two cleavage sites, namely S1 and S2, that are processed by the mitochondrial metalloproteases OMA1 and YME1L [18]. L-OPA1 is anchored on the IMM, while the S-OPA1 resides on the intermembrane space (IMS), within the cristae. Although the L-OPA1 can mediate the fusion of IMM under cellular stress, recent studies have shown that fusion is more efficient when both L-OPA1 and S-OPA1 are present [13,19]. Knockout of OPA1 results in aberrant mitochondrial fragmentation, while overexpression of OPA1 results in a tubulated mitochondrial network in a MFN1 dependent manner [3,19-21]. Aside from the regulation of IMM fusion, OPA1 has been shown to regulate cristae remodeling during apoptosis and open the “gate” for the release of the majority of cytochrome c that resides within the cristae [22].

Mitochondrial fission

Mitochondrial fission occurs at the contact sites of mitochondria with the endoplasmic reticulum (ER), and is primarily regulated by the cytosolic GTPase dynamin-related protein 1 (DRP1), which translocates to the OMM to sever mitochondria [23,24] (Figure 1B). There, DRP1 can bind to four receptor proteins: (i) mitochondrial fission factor (MFF), (ii,iii) mitochondrial dynamics proteins of 49 kDa (MID49) and 51 kDa (MID51), and (iv) mitochondrial fission 1 protein (FIS1), where MFF is the primary receptor that docks DRP1 on the OMM [25,26]. Focal accumulation of DRP1 receptor proteins at mitochondria–ER contact sites allows DRP1 to be docked to the OMM where it subsequently forms ring-shaped oligomers around mitochondria. Oligomeric forms of DRP1 constrict mitochondria upon GTP hydrolysis to complete the severing of mitochondria [27]. Loss of DRP1 induces mitochondrial elongation [28]. Notably, knockout of MFF results in a significantly milder phenotype, and this has been attributed to the compensatory role in DRP1 docking by other receptors [29]. Interestingly, knockout of FIS1 does not have a significant impact on mitochondrial morphology, whereas overexpression of FIS1 induces mitochondrial fragmentation [25,30].

Crosstalk between the fusion and fission machinery

Recent reports suggest that the components of the fusion and fission machinery interact with each other, and such interaction regulates balance between fusion and fission. Interestingly, overexpression of human FIS1 was found to induce mitochondrial fragmentation even in the absence of DRP1 by directly binding and reducing the GTPase activity of MFN1/2 and OPA1. Such association inhibits mitochondrial fusion and suggests that FIS1 can directly control mitochondrial fusion by interacting with the fusion effector proteins [31]. In line with this idea, overexpression of MID49 and MID51 has been shown to promote fusion in a DRP1 independent manner by decreasing the association of FIS1 with MFN1/2 and OPA1 [32]. These are exciting findings and challenge the current dogma that strictly places mitochondrial shaping proteins in either the group of mitochondrial fusion or fission machinery. Future studies are required to decipher if fusion and fission operate independently and antagonize each other or if they cooperate under a unifying mechanism to modulate mitochondrial morphology.

Structure-function relationships of mitochondrial fusion and fission proteins

MFNs

MFNs consist of a large GTPase domain, two heptad repeat (HR1 and HR2) domains that interact intramolecularly to form a V-shaped coiled coil structural motif, and a transmembrane domain that is responsible for anchoring MFNs on the OMM [33,34]. Several models have been proposed for the molecular mechanism that mediates MFNs oligomerization, but the exact mechanism remains elusive. Those models have been extensively reviewed elsewhere [35]. We previously provided structural models and supporting data proposing that oligomerization of MFNs requires the transition of MFNs from an anti-tethering to a pro-tethering conformation (Figure 2A) [6]. Such transition loosens the intramolecular HR1-HR2 interactions and exposes the HR2 domain to the cytosol where it can interact with the HR2 domain of another activated MFN from adjacent mitochondria leading to a trans dimerization [6]. Oligomerization of MFN2 requires dimerization of the GTPase domains which may occur in both trans or cis mode. [34]. Model building and molecular dynamics studies favor an oligomerization assembly mechanism compatible with cis dimerization of GTPase domain between different MFN molecules which also use trans dimerization through the HR2 domain [36]. Such model is consistent with all available biochemical data.

Figure 2. Structural analyses of mitochondrial fusion and fission machinery.

(A) Structural model of the MFN2 in the anti-tethering conformation. The structural model for the anti-tethering conformation of MFN2 was based on bacterial dynamin like protein (BDLP; PDB ID: 2J69). (B) Structural overview of MGM1 (PDB ID: 6QL4). MGM1 is the yeast homologue of OPA1. MGM1 consists of a GTPase domain (head), a BSE domain (neck), a stalk domain, and paddle domain (tail). (C) Structural overview of FIS1 (PDB ID: 1PC2). FIS1 consists of a TRP-like domain and a transmembrane domain. (D) Structural comparison of monomeric DRP1 in nucleotide free and bound conformations (PDB ID: 4BEJ, 5WP9). Binding of the nucleotide to the GTPase domain of DRP1 induces a conformational change that pushes away the BSE domain from the GTPase domain and generates an elongated DRP1 structure.

OPA1

OPA1 contains the following domains: an N-terminal transmembrane domain proximal to the GTPase domain (head), a bundle signaling element (BSE) domain (neck), a stalk domain, and a paddle domain (tail) (Figure 2B) [37]. OPA1 undergoes a partial proteolytic processing generating a transmembrane-containing long form (L-OPA1) and a soluble short form (S-OPA1) at the intermembrane space, and both are necessary for IMM fusion [13,18]. Truncated s-Mgm1 (s-OPA1 homologue in fungi) isoform was found to assemble, independently of nucleotide binding, in dimers mediated by stalk-stalk interactions, which further assemble into tetramers mediated from stalk-stalk and stalk- BSE interactions [37]. These ‘back-to-back’ oligomers in purified s-MGM-1 and S-OPA1 are able to tubulate IMM-like membranes by forming helical filaments wrapped around the IMM-like liposomes using also “head-to-head” (GTPase-GTPase) interactions upon nucleotide binding. These spiral oligomers can assemble to connect to opposing IMMs and stabilize the membrane curvature at the fusion site. Another structural study of yeast MGM1 in complex with GDP, revealed a different trimeric assembly using “head-to-tail” interactions [38]. Interestingly this trimer formation is inhibited by GTP binding suggesting that “head-to-head” interactions are not contributing to the IMM fusion but are required for oligomerization in helical filaments.

FIS1

FIS1 is anchored on the OMM by a transmembrane domain that is located on the C-terminus, while the cytosolic part of the protein is composed by six α-helices that are connected with short loops and form a single core domain (Figure 2C) [39]. Within the core domain, the central four helices are organized in tandem arrays of helix-loop-helix motifs similar to tetratricopeptides (TPR) motifs and form a twisted superhelical structure [39]. While in yeast FIS1 uses adaptor proteins such as MDV1 or CAF4 to dock DRP1 to mitochondria, the mammalian FIS1 is capable of recruiting DRP1 to mitochondria irrespective of adaptor proteins [40,41]. Recent structural studies have highlighted the role of the human FIS1 'arm' (N-terminal residues 1–8) in stabilizing intramolecular interactions with the TPR core and regulating FIS1 activity because its deletion impaired DRP1 recruitment to mitochondria and mitochondrial fragmentation [42]. Interestingly, the same arm region in yeast FIS1 mediates binding to recombinant yeast DRP1 homologue, Dnm1p.

DRP1

DRP1 is composed by four major components: a) an N-terminal GTPase domain; b) a variable domain (VD) that can interact with lipids; c) a stalk domain that is responsible for the homo-oligomerization of the DRP1; d) a bundle signaling element (BSE) that connects the GTPase domain with the stalk domain and regulates length of DRP1 (Figure 2D) [27]. Comparison of the nucleotide bound DRP1-GMPPCP complex structure with the apo-DRP1 structure uncovered that binding of the nucleotide elongates DRP1 by pushing away the BSE from the GTPase domain and exposes surfaces of DRP1 that interact with the mitochondrial receptor proteins (e.g. MID49) (Figure 2D) [27]. Further structural analysis on DRP1 oligomerization revealed that MID49 stabilized the nucleotide-bound DRP1 tetramers and induced the formation of MID49-DRP1 polymers in the shape of a linear filament [27]. Hydrolysis of GTP firstly induces the dissociation of MID49 from linear filament and subsequently the formation of ring-shaped DRP1 polymers capable of constricting liposomes to membrane tubules [27].

Pharmacological modulation of mitochondrial dynamics

Implications from the imbalance of mitochondrial dynamics contribute to the progression of neurodegeneration, cancer and cardiac diseases (Box 2), and other pathologies [43]. Thus, there is a novel opportunity to develop therapeutics for these diseases by targeting mitochondrial dynamics. To date several chemical probes and pharmacological compounds have been reported to modulate mitochondrial dynamics and are classified in four categories: a) modulators of protein oligomerization; b) GTPase inhibitors; c) compounds that regulate the transcription of the fusion/fission machinery; d) compounds that indirectly modulate PTMs on the fusion/fission machinery (Table 1). Here we will discuss the pharmacological strategies that were used to target the fusion/fission machinery and allude on the therapeutic potential of the discovered chemical probes.

Box 2. Imbalances in mitochondrial dynamics contribute to several pathologies.

Cancer:

oncogenes often tip the balance between fusion and fission to support survival and transformation [75]. For instance, DRP1-mediated mitochondrial fission is required for KRAS induced transformation in pancreatic cells and supports tumor growth [76,77], while MYC promotes mitochondrial fusion by up-regulating the genes that encode the mitochondrial fusion machinery [75,78]. Interestingly, several reports linked mitochondrial dynamics with angiogenesis and metastasis in cancer. Particularly, overexpression of MFN2 in endothelial cells has been shown to inhibit angiogenesis by reducing the expression of angiogenic factors [79]. Consistently, phosphorylation of FIS1 induced mitochondrial fragmentation and subsequently the formation of invadopodia or lamellipodia to mediate metastasis in hepatocellular carcinoma [80]. On the contrary, deletion of OPA1 attenuates metastasis and tumor growth in vivo [53].

Neurodegenerative diseases:

Aβ oligomers induce a cognitive decline and neuronal death in Alzheimer’s disease by promoting a network of hyper-fragmented mitochondria through downregulation of the Mfn2 gene and increased interactions between FIS1 and DRP1 [51,81,82]. Whether mitochondrial dysfunction has causative role to Alzheimer’s disease, or it is a consequence from the progression of the pathology has not been determined yet. Interestingly, overexpression of MFN2 has been shown to reduce neuronal death that is induced by Aβ oligomers [83]. Several loss of function mutations on Mfn2 gene that are found in Charcot Marie Tooth disease Type 2A (CMT2A) [81,84] reduce mitochondrial fitness and motility in neurons leading to neuronal degeneration. Moreover, missense mutations on the Opal gene that result in abnormal bioenergetics, calcium homeostasis and increased sensitivity to apoptosis are commonly found in autosomal dominant optic atrophy, peripheral neuropathy, and hearing loss [85-87].

Cardiovascular diseases:

OPA1 has been found to be downregulated in murine models of MI [88,89]. Interestingly, depending on the origin of mitochondrial stress cleavage of OPA1 by OMA1 has been shown to either protect or exacerbate the effects of stress in the heart. For instance, activation of OMA1 upon deletion of YME1L increased cleavage of L-OPA1, which was detrimental for the cardiac function and induced mitochondrial fragmentation, dilated cardiomyopathy and heart failure [90]. On the other hand, OMA1 has been shown to have cardioprotective effects upon activation of the mitochondrial unfolded protein response in the cristae [91]. Moreover, cardiac ablation of Mfn2 caused mitochondrial fragmentation and resulted in dilated cardiomyopathy in mice [92].

Table 1.

Chemical probes and pharmacological compounds that modulate mitochondrial dynamics

Chemical probe	Target	Effects on mitochondrial dynamics	Activity in cells/in vivo	References
	MFN1/2	Inhibits fusion	Decreases mitochondrial functionality. Induces minority MOMP and sensitizes to apoptosis	[2]
398-318Gly: QDRLKFIDKQGELLAQDYKLR	MFN1/2	Inhibits fusion	Decreases mitochondrial functionality	[6]
	MFN1/2	Promotes fusion	Improves mitochondrial functionality	[2]
	MFN1/2	Promotes fusion	Restores mitochondrial motility in neurons modeling CMT2A	[45]
	MFN1/2	Promotes fusion	Restores mitochondrial motility in neurons modeling CMT2A	[93]
367-384 Gly: DIAEAVRGIMDSLHMAAR	MFN1/2	Promotes fusion	Improves mitochondrial functionality	[6]
	MFN1/2	Promotes fusion	Improves mitochondrial functionality and transport in sciatic nerves	[94]
	MFN1/2	Promotes fusion	Improves mitochondrial functionality and transport in sciatic nerves	[94]
	OPA1	Inhibits fusion	Induces apoptosis and inhibits tumor growth in vivo	[53]
	DRP1/complex I	Inhibits fission	Protective effects in vivo in brain ischemia reperfusion injury	[56]
	DRP1	Inhibits fission	N/A	[55]
	DRP1	Inhibits fission	Inhibits tumor growth in vivo	[55]
P110 amino acid sequence: DLLPRGT	DRP1/FIS1 interaction	Inhibits fission	Protective effects in vivo against ALS and Alzheimer’s disease	[48]
	USP30	Promotes fusion	Improves mitochondrial functionality	[63]
	N/A	Promotes fusion	N/A	[95]
SAMβA amino acid sequence: NAENF	MFN1/βIIPKC	Promotes fusion	Improves heart failure in rats	[11]
	CK2	Promotes fusion	Protects in vivo against ischemic stroke in brain	[62]
	DHODH	Promotes fusion	N/A	[61]

Modulating protein oligomerization

Protein oligomerization is at the heart of the molecular mechanism that mediates fusion/fission. As discussed above, fusion and fission of mitochondrial membranes requires a coordinated effort of several mechanoenzymes whose strength grows in numbers and their activity is often regulated by adaptor proteins. Thus, manipulating the capacity of those molecules to participate in protein-protein interactions represents a powerful approach to influence the course of fusion/fission. In line with this idea, compounds that modulate MFNs oligomerization and the association of DRP1 to FIS1 have already been reported with remarkable effects on mitochondrial dynamics.

MFN1/2

Based on the intramolecular HR1-HR2 interactions in the anti-tethering MFN2 conformation and the intermolecular HR2-HR2 interactions in the pro-tethering MFN2 conformation, fragments of HR1 domain, peptides comprising HR1 residues 367-384 and 398-418 were designed and found to activate or inhibit the fusogenic activity of MFNs, respectively [6]. Mapping of the critical interactions between the HR1 and the HR2 domain led to generation of two distinct pharmacophore models that were used to screen in silico small molecules for engaging specific residues of HR2 domain of MFNs and mimic the HR1 peptides binding. Small molecule MFNs activators and inhibitors were identified through biochemical, biophysical and cellular studies [2,44].

Particularly, activator MASM7 was found to bind to the HR2 domain, promoted MFNs conformational activation and increased oligomerization leading to increased mitochondrial fusion (Figure 3A). For clarity, compound MASM7 was reported elsewhere as compound B, and MFN agonist [45]. MASM7 treatment also led to improved mitochondrial functionality as determined by increased membrane potential, respiration, and ATP production [2]. A series of other MFN activators have been developed using the same strategy and showed to restore mitochondrial function and motility of motor neurons directly reprogrammed from CMT2A patient fibroblasts and murine sciatic nerve axons ex vivo [45-47]. Treatment of MFN activators in mice expressing CMT2A-associatted MFN2 mutations in motor neurons, induced neuronal regrowth and gross neuromuscular function and integrity [46].

Figure 3. Pharmacological strategies to target mitochondrial fusion and fission machinery.

(A) MFNs agonists firstly promote the pro-tethering conformation of MFNs by disrupting the intramolecular HR1-HR2 interactions (step 1) and secondly promote MFNs oligomerization through the intermolecular interactions of the HR2 domains of MFNs from adjacent mitochondria (step 2). (B) MFNs antagonists inhibits mitochondrial fusion by inhibiting MFNs oligomerization through the disruption of HR2-HR2 the intermolecular interactions. (C) P110 rescues mitochondrial fusion by inhibiting DRP1-FIS1 association, and subsequently the translocation of DRP1 to mitochondria. (D) MYLS22 inhibit the GTPase activity of OPA1 and subsequently OPA1 oligomerization and fusion. (E) DRP1i inhibit the GTPase activity of DRP1 and subsequently DRP1 translocation and fission.

Interestingly, the small-molecule MFN inhibitor, MFI8, was found to bind to a different site on the HR2 domain compared to the activator MASM7, and decreased MFN oligomerization, leading to inhibition of mitochondrial fusion and induction of mitochondrial fragmentation (Figure 3B) [2]. Furthermore, inhibition of MFNs by MFI8 led to decreased mitochondrial functionality and induction of mitochondrial outer membrane permeabilization (MOMP) and cytochrome c release that increased sensitivity to apoptosis and led to synergistic cell death with a pro-apoptotic SMAC mimetic [2].

FIS1

Despite the perplexing role of FIS1 in the modulation of mitochondrial dynamics much attention has been given in controlling the DRP1–FIS1 interaction from a pharmacological perspective. Interestingly, DRP1 and FIS1 share three solvent-exposed regions with high sequence homology, a common feature of unrelated proteins that interact with each other [48]. Synthesis of peptides that correspond to these homologous amino acid sequences identified P110, a peptide comprising residues ⁵⁰DLLRPGT⁵⁶ of the DRP1 GTPase domain. P110 peptide can disrupt the DRP1-FIS1 complex formation in vitro and in cells upon treatment with mitochondrial stressors (Figure 3C) [48]. Notably, P110 reduced DRP1 translocation and rescued mitochondrial morphology only in neuronal cells that were stimulated with mitochondrial stressors, providing further support to the idea that FIS1 mediates mitochondrial fission when mitochondria are under stress and not at basal state [48,49].

The DRP1-FIS1 axis is activated in several neurological and cardiovascular pathologies to induce a network of hyper-fragmented mitochondria, oxidative stress, and mitochondrial dysfunction. Treatment of P110 to neuronal cells stimulated with Aβ₄₂ and to fibroblasts from amyotrophic lateral sclerosis (ALS) patients restored mitochondrial connectivity and rescued mitochondrial functionality [50,51]. Importantly, treatment of P110 in murine models of ALS and Alzheimer’s disease improved behavioral defects associated with the disease progression [50,51]. In a similar fashion, P110 rescued mitochondrial fragmentation, improved cardiac function, and decreased mortality in a murine model of cardiac dysfunction induced by lipopolysaccharides (LPS) [52].

Targeting the GTPase activity

The mechanoenzymes responsible for shaping mitochondrial morphology (OPA1, MFNs, and DRP1) rely on GTP hydrolysis for the fusion or constriction of mitochondrial membranes. Importantly, several studies have recently elucidated the structural topology of the GTPase domain in these proteins and established catalytic activity–structure relationships. These insights into the structure of the GTPase domain have paved the way for structure-guided drug design and have led to the development of compounds that inhibit the GTPase activity of OPA1 and DRP1.

OPA1

Despite the need for further validation of the oligomerization models and interactions required for the IMM fusion, the GTPase activity of OPA1 is considered to be crucial for fusion of IMMs to proceed. A high-throughput screening assay evaluated a chemical library for inhibition of recombinant OPA1 GTPase activity, leading to a specific and first-in-class OPA1 GTPase inhibitor, MYLS22 (Figure 3D) [53]. Interestingly, MYLS22 promoted mitochondrial fragmentation, increased cancer cell sensitivity to apoptosis and reduced tumor growth in a xenograft melanoma model [53]. Similarly, another study on triple-negative breast cancer cell lines and patient-derived xenograft models found that MYLS22 could suppress mitochondrial fusion and oxidative phosphorylation (OXPHOS), and limit tumor growth [54].

DRP1

Resolution of the DPR1 crystal structure paved the way for the rational design of DRP1 antagonists. A structure-guided drug design using DRP1 GTPase domain in conjunction with in silico screens identified Drpitor, an orthosteric DRP1 inhibitor that binds to the GTPase domain and inhibits its catalytic activity [55]. Further optimization of this compound led to the development of a more potent DRP1 inhibitor, namely Drpitor1a. Drpitor1a prevented mitochondrial fission, induced apoptosis in lung cancer cells and inhibited tumor growth in a mouse xenograft model of lung cancer [55]. Interestingly, Drpitor1a improved right ventricular diastolic dysfunction in a cardiac ischemia reperfusion injury model [55].

Previous reports identified mdivi-1 as a DNM1 (yeast ortholog of DRP1) inhibitor. A high throughput screening assay in yeast was performed to evaluate cellular growth under conditions where heat-induced mitochondrial fragmentation is detrimental for the survival of the cells [56]. Subsequent analysis, identified mdivi-1 as a potent compound in increasing mitochondrial connectivity by allosteric inhibition of the DNM1 GTPase activity and blocking of DNM1 oligomerization in vitro [56]. In the same study mdivi-1 was found to increase mitochondrial connectivity in mammalian cells and overexpression of human DRP1 reduced the capacity of mdivi-1 to promote mitochondrial fusion. Bordt et al. found that mdivi-1 was not able to inhibit the GTPase activity of human DRP1 or increase the mitochondrial connectivity of treated neuronal cells [57]. Instead, mdivi-1 reversibly inhibited complex I activity, reduced mitochondrial respiration, and increase the production of reactive oxygen species (ROS) [57]. Nevertheless, mdivi-1 has been shown to have protective effects against traumatic brain injury and brain ischemia-reperfusion injury in vivo [58,59]. Such effects could be attributed to the reduced interaction of DRP1 with BAX that positively regulates apoptosis [60].

Modulating protein levels of the fusion/fission machinery

Altering the transcription of genes encoding mitochondrial fusion/fission effector molecules is an alternative way to modulate mitochondrial fusion/fission. Leflunomide, an FDA-approved drug for the treatment of rheumatoid arthritis, was found to increase the expression of MFNs and promote mitochondrial fusion by blocking pyrimidine biosynthesis through dihydroorotate dehydrogenase (DHODH) inhibition [61]. Moreover, the natural product echinacoside increased the expression of MFN2 by binding to casein kinase 2 and facilitating the nuclear translocation of β-catenin [62]. As mentioned previously, the dynamic interplay between the L-OPA1 and S-OPA1 protein levels is regulated by OMA1 and YME1L, and has been shown to modulate the capacity of OPA1 to mediate IMM fusion. Thus, it may be possible to rescue the proteolytic cleavage of L-OPA1 and promote mitochondrial fusion by developing inhibitors of OMA1 and YMEL1.

Controlling PTMs of the fusion/fission machinery

Recent reports have shed light on the role of PTMs in the function of fusion/fission machinery. For instance, the GTPase activity of MFN1 is reduced when the molecule is phosphorylated by 1-βIIPKC [11]. This PTM inhibits mitochondrial fusion and promotes mitochondrial fragmentation. A rational design approach, similar to that of P110 that was described above, identified SAMβA, a peptide that mitigates the phosphorylation of MFN1 by reducing the association of MFN1 with 1-βIIKPC [11]. Treatment of SAMβA improved cardiac contractility in rats with heart failure. On the contrary, non-degradative ubiquitination of MFN1/2 has been reported to promote fusion by increasing MFN1/2 activity [63]. Moreover, ubiquitinated MFN1/2 were identified as substrates of the mitochondrial deubiquitinase, USP30, that negatively regulates mitochondrial fusion. Based on this finding, a small molecule, namely S3, was identified to promote fusion by rescuing MFN1/2 ubiquitination through direct inhibition of USP30 [63].

Concluding remarks and future perspectives

Remarkable progress has been made in deciphering the mechanisms underlying mitochondrial dynamics and in understanding their impact on other cellular processes, as well as in clarifying the role of key mitochondrial dynamics proteins in the development and progression of several human diseases. In addition, recent studies have elucidated structure–function relationships for the majority of the effector proteins that mediate fusion/fission and have provided crucial insights into the protein–protein interactions that are necessary for fusion and fission. Hence, several drug screening and chemical biology campaigns have been initiated with the aim of developing pharmacological agents that modulate the activity of these proteins and mitochondrial fusion/fission. These studies have further validated mitochondrial dynamics proteins as potential drug targets through numerous genetic approaches. More fundamental discoveries and drug development projects will surely emerge from this exciting area of mitochondrial biology.

Despite the significant progress in this field, there are still several questions (see outstanding questions) to be addressed for translating mechanisms and structural insights to therapeutics, which will need input from structural biology, biochemistry, cell biology, computational biology and chemistry. The aims for the next five years should be to continue deepen our understanding about the full-length structures of mitochondrial proteins, their conformational changes required to transition to fusion/fission permissive structures and oligomers and what specific interactions mediate these conformational changes and assemblies. In addition, the roles of the GTPase domains of DRP1, OPA1, and MFN1/2 in either fusion or fission at the structural and functional level remain unclear, and what nucleotide (GDP/GTP) binding accomplishes is also uncertain. Furthermore, the field should identify so far unknown effectors including PTMs that modulate protein structure, activity, and protein–protein interactions involved in monomer to oligomer transitions. Effector proteins that are necessary for fusion and fission, even if they are only present in specific cellular contexts and tissues, may also provide alternative drug targets for fusion/fission modulation. Concurrently with these efforts, recent success with small molecule targeting of mitochondrial dynamics proteins should inspire further chemical screening and drug discovery campaigns. Given the notable examples described for MFN1/2 and FIS1–DRP1, it is possible that similar approaches can be used to develop chemical probes and therapeutics to promote or inhibit protein–protein interactions that are crucial for fusion/fission or the formation of oligomers of DRP1, OPA1, and MFN1/2. In addition, because the role GTPase domains of OPA1, MFN1/2 and DRP1 are crucial for mitochondrial fusion and fission, selective inhibitors for each GTPase (either orthosteric or allosteric) would have significant potential for modulating fusion and fission. Toward these aims, advances in structural, biophysical, imaging, and single molecule studies for tracking conformational changes and oligomerization will facilitate the elucidation of molecular mechanisms and discovery of small molecule modulators [64].

Outstanding Questions.

• Can we provide a more complete picture of the full length structures and conformational changes involved with the membrane anchoring and oligomerization of mitochondrial dynamics proteins in fusion and fission.

• Can we provide a better understanding about the structural and functional role of GTPases of mitochondrial dynamics proteins in fusion and fission.

• What other unknown effector proteins modulate protein-protein interactions of the mitochondrial dynamics machinery that are necessary for fusion and fission.

• Can we develop chemical probes and therapeutics that will induce or inhibit oligomerization of e.g. DRP1, OPA1 and protein-protein interactions e.g. FIS1-MFN1/2, FIS1-OPA1, DRP1-BAX, MFN2-BAX to modulate fusion and fission?

• Can we develop selective orthesteric or allosteric inhibitors of GTPase domains of DRP1, MFN1, MFN2, OPA1 to modulate fusion and fission?

Resolving additional mysteries and pharmacological targeting of mitochondrial dynamics proteins promise exciting opportunities for the identification of prototype therapeutics for several unmet diseases where mitochondrial defects drive pathogenesis.

Highlights.

• Recent advances highlight the role of mitochondrial dynamics proteins and mitochondrial fusion/fission in cell physiology and diverse human pathologies.

• Significant progress in structural and biochemical investigations has illuminated the molecular mechanisms of mitochondrial fusion and fission that will inform rational strategies for drug targeting.

• Small molecules targeting mitochondrial dynamics proteins have demonstrated promising pharmacological activity and highlight the druggability of these proteins.

• Pharmacological restoration of deregulated mitochondrial dynamics decreased the progression of various pathologies.

Glossary:

Outer membrane

is the outer portion of mitochondria that allows communication with the cytosol and other organelles. It is impermeable for proteins but most small molecules, metabolites, cations can pass freely through passive diffusion. It contains proteins that regulate mitochondrial import, such as translocases and membrane pores, or they are involved in contacts with organelles and apoptosis.

Intermembrane space

This is the area between the inner and outer membranes. It contains a small number of proteins required for mitochondrial biogenesis or energy metabolism. For example, cytochrome c, which plays a role in apoptosis but also in shuttling electrons between complexes III and IV.

Inner membrane

is the membrane which separates the mitochondrial matrix from the outer membrane and holds proteins that have several functions. This is where the electron transport chain complexes reside. It is impermeable to most molecules and crossing the inner membrane by molecules requires special membrane transporters.

Cristae

are the folds of the inner membrane which are dynamic in shape and dimensions. They increase the surface area of the membrane and space available for chemical reactions and modulate the structure of protein complexes and the kinetics of chemical reactions

OXPHOS

oxidative phosphorylation is the metabolic pathway that mitochondria use to produce energy in the form of ATP by using nutrient oxidation as a driving force to transfer electrons through a series of redox reaction coupled with mitochondrial respiration.

Glycolysis

is the metabolic pathway that cells use to break down glucose to produce energy in the form of ATP and pyruvate, which can fuel the TCA cycle and produce more molecules of ATP.

Pharmacophore model

is the defined spatial orientation of steric and electronic features of a ligand that are required for the molecular recognition and specific interaction of the ligand with its protein target. Such steric and electronic features include hydrophobic centroids, aromatic rings, hydrogen bond donor/acceptor, anions and cations.

Invadopodia

are protrusions of the cellular membrane that are rich in actin and metalloproteases that appear in cancer cells during metastasis. Their role is to clear the path for cancer cell migration during metastasis by secreting proteases and degrading the extracellular matrix.

Lamellipodia

are thin sheet-like protrusions of the cell membrane and are composed by branched filaments of actin that form a mesh structure. Their biological role is to promote cell motility