VDAC Regulation of Mitochondrial Calcium Flux: From Channel Biophysics to Disease

Abstract

Voltage-dependent anion channel (VDAC), the most abundant mitochondrial outer membrane protein, is important for a variety of mitochondrial functions including metabolite exchange, calcium transport, and apoptosis. While VDAC’s role in shuttling metabolites between the cytosol and mitochondria is well established, there is a growing interest in understanding the mechanisms of its regulation of mitochondrial calcium transport. Here we review the current literature on VDAC’s role in calcium signaling, its biophysical properties, physiological function, and pathology focusing on its importance in cardiac diseases. We discuss the specific biophysical properties of the three VDAC isoforms in mammalian cells—VDAC 1, 2, and 3—in relationship to calcium transport and their distinct roles in cell physiology and disease. Highlighting the emerging evidence that cytosolic proteins interact with and regulate VDAC calcium permeability, we advocate for continued investigation into the VDAC interactome at the contact sites between mitochondria and organelles and its role in mitochondrial calcium transport.

Graphical Abstract

1. Introduction

Traditionally, the mitochondrial outer membrane (MOM) was regarded as just a barrier between the mitochondrial space and the cytosol. Emerging data has assigned more functions for the MOM, such as control over metabolite fluxes between the cytosol and mitochondria, tethering with the endoplasmic reticulum (ER), and lysosomes, and direct involvement in a cross-talk between mitochondria and proteins involved in metabolic and survival pathways. Some of these MOM functions are executed and regulated by the voltage-dependent anion channel (VDAC), the most abundant protein in the MOM. While VDAC’s role in conducting and regulating the MOM permeability to ions and small water-soluble metabolites is well-established [1–4], VDAC’s involvement in calcium (Ca²⁺) homeostasis is still unknown. In contrast to the mitochondria inner membrane (MIM), which has five known pathways for Ca²⁺ transport, the MOM does not have a specific Ca²⁺ transporter [5]. Tom40, the channel component of the translocase of the outer membrane (TOM complex) is a plausible candidate for a passive Ca²⁺ pathway [6]. The high cation selectivity of Tom40 [7] makes it an even more favorable candidate for Ca²⁺ transport than the weakly anion-selective VDAC. Indeed, Tom40 preference for cations matches positive charges of mitochondrial presequence matrix proteins [6, 8] and VDAC preference for anions matches a majority of negatively charged mitochondrial metabolites passing through it [4]. Such a straightforward comparison between two MOM major β-barrel channels raises the question of whether VDAC really plays an active role in Ca²⁺ regulation or it is merely an inert diffusion pore?

Initial studies on Ca²⁺ signaling using permeabilized cells relied on the assumption that the sheer number of highly conductive VDACs effectively constitutes no barrier to Ca²⁺, unlike the low copy number of Mitochondrial Calcium Uniporter (MCU) in the MIM, which regulates Ca²⁺ influx into the mitochondrial matrix (Figure 1) [9]. Though there are competing theories of the mechanism of Ca²⁺ flux through MCU (see discussion in [10]), it is well known that MCU shows positive cooperativity at elevated Ca²⁺ concentrations and therefore enhances Ca²⁺ flux at micromolar local Ca²⁺ concentrations [11]. In the cytosol, however, the global Ca²⁺ pool is generally around 100-500 nM [Ca²⁺] with peaks up to around 1-3 μM [12]. Therefore, rapid mitochondrial Ca²⁺ influx is achieved through tight contact sites between mitochondria and Ca²⁺ storage organelles such as ER, sarcoplasmic reticulum (SR), or lysosomes that are known as mitochondrial associated membranes (MAMs) [13]. MAMs have multi-protein complexes containing VDAC and Ca²⁺ channels that create highly localized Ca²⁺ microdomains for rapid Ca²⁺ transport across mitochondria [14] (Figure 1).

Figure 1. Ca²⁺ flux into the mitochondria through global cytoplasmic Ca²⁺ and Ca²⁺ microdomains.

Mitochondria interact with organelles such as the endoplasmic reticulum (ER) through contact sites generating Ca²⁺ microdomains. These Ca²⁺ microdomains (left half), contain a multi-protein complex linking IP3R (an ER Ca²⁺ channel in purple) to VDAC (3D structure representation) in the mitochondrial outer member (MOM). The microdomains enable rapid transfer of Ca²⁺ (Ca²⁺ sparks) from the ER to mitochondria through VDAC and the MCU in the inner membrane (MIM). Global or capacitive uptake of Ca²⁺ from the cytosol (right half) is slower due to the lower local concentration of Ca²⁺compared to microdomains. Ca²⁺enters the cytosol from extracellular space or ER stores.

In early studies, VDAC was interpreted as an inert diffusive pore to solutes and metabolites. However, recent research has uncovered a wide variety of biological processes that alter the accessibility of VDAC, reigniting the debate on whether VDAC could be an active regulator of mitochondrial Ca²⁺ flux [15–20]. The proposed mechanisms of VDAC regulation of Ca²⁺ range from the inherent gating of VDAC, to VDAC’s interaction with other proteins and the surrounding lipid environment.

In this review, we revisit the role of VDAC in translocation and regulation of mitochondrial Ca²⁺ fluxes. We discuss the biophysical basis of VDAC Ca²⁺ transport, the impact of VDAC cytosolic regulators on its Ca²⁺ permeability, and VDAC’s role in mitochondria-organelle Ca²⁺ crosstalk with particular attention to the contribution of the individual VDAC isoforms in Ca²⁺ signaling. Finally, we discuss the possible physiological implications of VDAC Ca²⁺ transport in diseases. It is our hope that a deeper understanding of the mechanisms of VDAC regulation will help the researchers in interpreting mitochondrial Ca²⁺ experiments in the cells, recognizing new pathways that could regulate mitochondrial Ca²⁺ flux through VDAC, and developing novel therapies targeting VDAC Ca²⁺ transport.

2. VDAC’s role in Ca²⁺ exchange between mitochondria and organelles

Ca²⁺ is an important signaling molecule that regulates cell function through spatial and temporal Ca²⁺ release patterns [21]. Importantly, cytoplasmic Ca²⁺ concentration is maintained at low levels against its concentration in the extracellular space and intracellular stores. These levels are maintained through signaling mechanisms that either pump Ca²⁺ out into the extracellular space or the intracellular stores in organelles such as the ER, SR, and mitochondria [22]. Mitochondria exchange Ca²⁺ with the ER, SR, and lysosomes at close contact sites (MAMs). These contact sites have multi-protein complexes linking VDAC in the MOM to Ca²⁺ channels such as the ryanodine receptors (RyRs) at the SR, inositol 1,4,5-trisphosphate receptors (IP₃Rs) at the ER, and TRPML1 at the lysosome [23, 24] (Figure 2). Such tight co-localization of multiple proteins involved in Ca²⁺ transduction is important for rapid Ca²⁺ transfer to mitochondria through a process known as Ca²⁺ “sparks” or “puffs”. Ca²⁺ sparks are localized microdomains of highly concentrated Ca²⁺ that result in transient micromolar concentration promoting rapid Ca²⁺ uptake across MCU [25]. Ca²⁺ sparks in the mitochondria mediate signaling pathways that control ATP supply, metabolism, and apoptosis in cell [26–28].

Figure 2. VDAC protein partners in Ca²⁺ regulation.

VDAC can regulate mitochondrial function through its interaction with multiple proteins such as various organelle Ca²⁺ channels and cytoplasmic and MOM proteins. VDAC facilitates efficient Ca²⁺ transfer to mitochondria by forming multi-protein complexes with Ca²⁺ channels in other organelles such as the IP3R-VDAC-GRP75-DJ-1 complex, the RyR2-VDAC, and TRPML1-VDAC on the ER, SR, and lysosome, respectively. ER-associated CKAP4 competes with the IP3R-GRP75 complex for interaction with VDAC. VDAC interacts with cytosolic free dimeric tubulin and PD-associated protein α-synuclein (αSyn). Tubulin and αSyn bind to the MOM and reversibly block VDAC by their disordered acidic C-terminal tails (shown in red). VDAC also interacts with TSPO, another MOM protein, which has been shown to affect Ca²⁺ uptake in cardiac disease.

Earlier studies on the role of MOM permeability in the uptake of Ca²⁺ focused on global or “capacitive” Ca²⁺ exposure whereupon an influx of Ca²⁺ into the cytoplasm saturates the mitochondria. For example, in HeLa cells, cytosolic to mitochondrial Ca²⁺ transfer, carried out by exposing permeabilized cells to extracellular Ca²⁺, was not altered by VDAC1 overexpression [17]. These experiments suggested that in basal conditions, flux across the VDAC is enough to saturate the MCU, and therefore overexpressing VDAC may not significantly alter flux. However, the same work showed that VDAC overexpression increases Ca²⁺ permeability at mitochondria-ER contact sites (MERCS). The authors further looked at the colocalization of VDAC-GFP at the MOM with the ER protein, calreticulin, and found that overexpression of VDAC did not alter the overlap between the ER and mitochondria. Nevertheless, using single-cell measurements of the ER to mitochondria Ca²⁺ transport, they found that overexpression of VDAC reduced the delay between ER Ca²⁺ release and subsequent mitochondrial uptake. These experiments suggest that VDAC overexpression does not increase the number of MERCS but does increase the permeability of MERCS to Ca²⁺. However, since colocalization methods have some obvious limitations, namely the diffraction limit of light precludes probing sub-250 nm structures such as the contact sites, novel methods that offer more accurate measurements of true MERCS such as split-GFP-based contact site sensor (SPLICS), FRET, and proximity labeling techniques should be used to further study these junctions [13, 29]. These data first indicated that the transfer of Ca²⁺ through the ER-MOM contacts critically depends on VDAC which forms a kind of “kinetic bottleneck” for the total mitochondrial Ca²⁺ homeostasis [17]. Altogether, these results indicate that VDAC is important for Ca²⁺ transfer from the ER and SR, where transient microdomains of Ca²⁺ flux propagate a wide variety of signaling responses.

Beyond mitochondria-ER contact sites, contacts between mitochondria and lysosomes were recently described (Figure 2). Wong et al. showed that the transient contacts between lysosomes and mitochondria induce mitochondrial fission [30]. Following up on these interesting results their group demonstrated that Ca²⁺ efflux from the lysosomes to the mitochondria is mediated by tethering of lysosomal-mitochondria contact sites through the transient receptor potential mucolipin 1 (TRPML1) [31]. Loss-of-function mutations in TRPML1 are known to cause mucolipidosis type IV (MLIV), a lysosomal storage disease that triggers severe neuronal impairment [32, 33]. Peng et al. showed that patient-derived fibroblasts harboring mutations in TRPML1 showed impaired mitochondrial Ca²⁺ uptake and this may be associated with mitochondrial damages found in MLIV [31]. In a coimmunoprecipitation pull-down for TRPML1, they found VDAC1, but not VDAC2 or VDAC3, in complex with TRPML1 [31]. They also found that mitochondrial Ca²⁺ uptake at mitochondria-lysosome contact sites is modulated by VDAC1 in the MOM and by MCU in the MIM. To test the role of VDAC1 in mitochondrial uptake of lysosomal Ca²⁺, the authors compared Ca²⁺ dynamics in cells expressing either wild-type human VDAC1 or its mutant E73Q with a single mutation in a putative Ca²⁺-binding site previously proposed by Israelson et al. [18]. The E73Q VDAC1 mutant showed a smaller increase in mitochondrial Ca²⁺ uptake than the wild-type. These results led authors to conclude that VDAC1 (but not VDAC2 or VDAC3) serves as a mediator of lysosomal Ca²⁺ uptake at mitochondria-lysosome contact sites, thus assigning a single residue, E73, a decisive role in this process. However, recent in vitro studies put in doubt the role of E73 as a unique VDAC1 Ca²⁺ binding site.

VDAC1 E73 residue was first suggested as a possible Ca²⁺ binding site [18, 19] (E72 according to the VDAC1 sequence used in these works) before the VDAC1 crystal structure was solved in 2008 [34–36]. which revealed that E73 is buried in the middle of the lipid membrane hydrocarbon core. Since then, this notorious residue has drawn attention from VDAC researchers [37] and especially in NMR spectroscopy and molecular dynamic (MD) simulation studies [36, 38–40]. Indeed, it is intuitively clear that the probability of a Ca²⁺ ion reaching the middle of the hydrocarbon core of the lipid bilayer is extremely low if not impossible. Interestingly, the same E73 residue was later identified as a specific binding site for cholesterol [35, 41, 42] and allopregnanolone [43, 44] by using NMR, photoaffinity labeling, and computational analysis. This raises a natural question of how the same residue could be a binding site for both a hydrophilic ion and a hydrophobic steroid. In addition, it should also be noted that E73 is not conserved among VDAC isoforms and species: mammalian VDAC3, VDAC from yeast [45], and fungi [37] have a non-charged glutamine Q73 (see also discussion in [46]). Recently, it was shown in experiments with wild-type mouse VDAC1 and mutants E73Q and E73A reconstituted into planar lipid membranes that E73 is not involved in VDAC1 gating [46], contrary to the early work prescribing a regulatory role for this residue [36, 38]. In the latest experiments with VDAC1 reconstituted into the planar membranes in pure CaCl₂ gradient, our group showed that VDAC1 E73Q mutant has the same Ca²⁺ selectivity as the wild-type [47], at odds with the proposed role of E73 as a Ca²⁺ binding site. E73Q VDAC1 mutation may disrupt the Ca²⁺ signaling complex at the contact site or affect the VDAC-lipid interface. We have previously shown that a hydrophobic molecule, by binding to the external hydrophobic surface of VDAC β-barrel at the protein-lipid interface, could directly affect VDAC gating properties [48, 49]. This may hint at an explanation of why the E73Q VDAC1 mutant caused a reduction of lysosomal Ca²⁺ uptake shown by Peng, Wong et al. [31].

3. VDAC isoforms in Ca²⁺ transport

There are three mammalian isoforms of VDAC - VDAC1, VDAC2, and VDAC3 - encoded by individual genes for proteins of ~ 32 kDa. Each VDAC isoform has a tissue-specific expression pattern, hinting at unique roles for VDAC isoforms in organ physiology [50]. In general, VDAC1 and VDAC2 show higher levels of expression (~90%) than VDAC3 which is far less abundant (~10%), with an exception in the testis [45, 51]. Of the three isoforms, only embryonic knockout of VDAC2 is either lethal according to earlier studies [52] or resulted in developmentally impaired neonatal mice according to a recent study [53]. Embryonic knockout of VDAC1 and VDAC3 results in mice with physiologic impairments, such as bioenergetic defects in VDAC1 and VDAC3 knockout mice [54, 55] and infertility in VDAC3 knockout [56]. One of the outstanding questions in VDAC research is determining the relationship between specific physiological function and biophysical channel properties of each isoform. Recent data suggest that one of the distinguishing features of each isoform is their differential regulation and interaction with protein partners [52, 53, 57, 58].

Another imperative question in mitochondrial physiology is defining the role of each VDAC isoform in mediating Ca²⁺ signaling. Szabadai et al. first showed that VDAC1 forms a complex with IP3R mediated by the chaperone protein glucose-regulated protein 75 (grp75) [14]. In addition, the loss of grp75 dissociated the IP3R-VDAC1 interaction and impaired mitochondrial Ca²⁺ uptake after Ca²⁺ release from ER. Later it was shown that the IP3R-grp75 complex is unique to the VDAC1 isoform [16]. However, the uniqueness of VDAC1 to the complex has recently been called into question by Harada et al., who found that VDAC2 can also form a complex with IP3R [59]. Liu et al. has determined that another protein, DJ-1, mutated in familial forms of Parkinson’s disease (PD) is also a component of the IP3R-GRP75-VDAC1 complex [60]. Ablation of DJ-1 dissociated the IP3R-GRP75-VDAC1 complex and reduced mitochondrial Ca²⁺ uptake from the ER. Altogether, these findings indicate that a large VDAC-containing complex exists at the MERCS that can regulate Ca²⁺ flux (Figure 2). Given that the complex is regulated by its interaction with DJ-1 and the evidence that mitochondrial Ca²⁺ signaling is altered in PD, it is possible that the IP3R-GRP75-VDAC1 super-complex functions as a regulator of Ca²⁺ in neuronal cells where rapid Ca²⁺ signaling is imperative to meet high energy demands during neuronal activity.

Like VDAC1, VDAC2 has also been found in complexes with Ca²⁺ channels. Kee-Min and colleagues performed a screen to look for interacting partners of the Ryanodine Receptor 2 (RyR2), which mediates Ca²⁺ transport from the SR to the mitochondria in the heart [61]. They found VDAC2 as an interacting partner of RyR2 and showed that the two proteins colocalize using immunofluorescence and immuno-EM methods. As mentioned above, Harada et al. found that VDAC2 also colocalized with the IP3R-GRP75 complex in HeLa cells. In addition, they found that another ER-associated protein, cytoskeleton-associated protein 4 (CKAP4), bound to VDAC2 and detached it from the IP3R-GRP75 complex [59]. The ablation of CKAP4 led to enhanced MERCS formation and subsequent mitochondrial dysfunction caused by Ca²⁺ influx [59]. While most of the studies have been focused on Ca²⁺ transport involving the two major isoforms, VDAC1 and VDAC2, there has yet to be a concerted effort to establish a role of the minor isoform VDAC3 in Ca²⁺ transport.

Following their original work on VDAC1’s role in Ca²⁺ transport [17], Rizzuto’s group dissected the role of each VDAC isoform. De Stefani and colleagues found that knockout of individual VDAC isoform impaired Ca²⁺ transport to the mitochondria after IP3R agonist release [16]. Conversely, overexpression of an individual isoform enhanced the ER to mitochondrial Ca²⁺ transport. Interestingly, overexpression or silencing of VDAC2 or VDAC3 had a slightly higher effect on mitochondrial Ca²⁺ uptake than overexpression of VDAC1. The authors concluded that this difference may arise from the different Ca²⁺ transport properties of each isoform. However, this hypothesis has not been explored further.

Recently our group showed that all three VDAC isoforms have slight, but significantly different Ca²⁺ permeability with VDAC1 being more anionic and consequently less Ca²⁺ selective among three isoforms and VDAC3 more cationic and consequently more Ca²⁺ selective [47]. Interestingly, these data obtained on recombinant VDAC isoforms reconstituted into the planar lipid membranes, support an early hypothesis by De Stefani et al. [16], where the authors assigned the slight difference in mitochondrial Ca²⁺ uptake caused by individual VDAC isoform overexpression to the isoform-specific Ca²⁺ permeability of each isoform. We can speculate even further suggesting that the enhanced permeability to Ca²⁺ by VDAC3 found in our in vitro study, may be related to sperm defects and infertility in VDAC3 knockout mice due to the high expression levels of VDAC3 in the testes and sperm [56].

4. VDAC Permeability to Ca²⁺ - Biophysical Studies

Typically, ion channels are studied in situ using the patch-clamp method. However, the study of ion channels from cell organelles by this method is an experimentally daunting endeavor, because applying a patch-clamp to ~1 μm organelles, such as the mitochondrion, presents understandable experimental challenges. Therefore, the method of reconstituting organellular ion channel proteins into planar lipid membranes is, so far, the best available method to achieve quantitative characterization of ion channel properties. Mitochondrial VDAC is no exception. The experimental setup for VDAC reconstitution is schematically shown in Figure 3A. When reconstituted in physiologically relevant 150 mM KCl, VDAC forms large, ~0.7 nS [46] anion-selective channels (Cl⁻/K⁺ ~ 3) permeable to non-charged polymers up to 2-5 kDa [62, 63] and ATP and other negatively charged respiratory metabolites [2, 3]. The characteristic property of reconstituted VDACs, and its namesake, is gating. Gating refers to the ability of VDAC to move under the applied voltages of > 30 mV from a unique high conducting or “open” state to a variety of low conducting or so-called “closed” states [4, 64, 65] (Figure 3B). As most of the VDAC channel properties have been obtained and quantified in monovalent KCl and NaCl solutions, it was unclear whether VDAC would behave similarly in the presence of Ca²⁺ or in pure CaCl₂ solutions. The answer to the former question can be found in the earlier works of Colombini’s group: most of the experiments with reconstituted VDAC were performed in the presence of non-physiological 2-5 mM Ca²⁺, which nonetheless did not measurably affect channel conductance, selectivity, or gating in 1 M KCl [15, 66, 67]. The answer to the latter question was obtained by Gincel et al. who were the first to demonstrate typical VDAC gating for channels reconstituted into pure 150/500 mM CaCl₂ solutions [68]. The authors also found that the selectivity of the VDAC open state measured in pure CaCl₂ gradient, is anionic with a permeability ratio between Cl⁻ and Ca²⁺, P_Cl− /P_Ca2+, equal to ~3. Later, Tan and Colombini measured VDAC selectivity in a lower concentration of CaCl₂ (in 80/20 mM CaCl₂ gradient) and found that in this condition, Cl⁻ permeability is ~25-fold higher than that of Ca²⁺ (Figure 4A) [15]. These results are not surprising considering the lower mobility of divalent Ca²⁺ than of monovalent Cl⁻, and that VDAC anionic selectivity increases with decreasing monovalent salt concentration [69]. Most importantly, Tan and Colombini showed that VDAC’s more cationic closed states have higher permeability for Ca²⁺ than the open state. They extrapolated their results to a physiologically low Ca²⁺ concentration of 1μM and found that a calculated Ca²⁺ flux through the closed states is about 4 to 10-fold higher than through the open state (Figure 4B). These results opened an important possibility that by gating, VDAC could regulate Ca²⁺ fluxes to and from mitochondria, thus providing evidence for an active role for VDAC in Ca²⁺ homeostasis rather than just being an inert pore. The immediate consequence of these findings is that VDAC in its open state facilitates fluxes of metabolites, but keeps Ca²⁺ flux low to maintain healthy mitochondrial respiration. In the closed states, VDAC diminishes metabolite fluxes but increases Ca²⁺ permeation. This state would attenuate respiration and accentuate apoptotic signals. The authors concluded that VDAC gating is not just a peculiar biophysical phenomenon but important physiological property to control MOM permeability [15].

Figure 3. Experimental assessment of VDAC properties.

(A) A schematic experimental setup for VDAC reconstitution to the planar lipid membrane. The experimental chamber for planar lipid membrane formation consists of two compartments, cis and trans. Potential is defined as positive when it is greater at the side of VDAC addition (cis side). The current recording is performed in a voltage-clamp mode shown by the model circuit for current amplification and registration. Adapted by permission from Rostovtseva and Bezrukov (2015). Copyright © Springer International Publishing Switzerland 2015. (B) A representative single-channel current trace obtained with recombinant mouse VDAC1 reconstituted into the planar membrane at different applied voltages. Under high applied voltage (± 60 mV) channel conductance moves from the unique high conducting “open” state to a variety of low-conducting “closed” states. Relaxing voltage to 0 mV reopens the channel. Adapted from Queralt-Martin et al. J. Gen. Physiol. (2019). Copyright with permission © Rockefeller University Press 2019.

Figure 4. VDAC voltage gating enhances Ca²⁺ permeability.

(A) Permeability of VDAC to Ca²⁺ and Cl⁻ as a function of the total conductance of the channel. (B) Comparison of the permeability of Ca²⁺ for the VDAC open and closed states. Adapted from Tan and Colombini. Biochim. Biophys. Acta. (2007). Copyright with permission Copyright © 2007 Elsevier.

The discussions about the physiological relevance of VDAC gating, are confounded by open questions about a quantitative value for the transmembrane potential across the MOM in vivo. Common sense suggests that this potential should be negligible, due to the high abundance of the highly conductive VDACs in the MOM (see discussions, e.g. in [4, 15]). Colombini suggested that the main source of the potential across the MOM is the so-called Donnan potential which arises from a high concentration of VDAC-impermeable polyanions and polycations (e.g., cytochrome c of +9 charges) in the intermembrane space and the cytosol [4]. An intriguing theoretical model of the possible source of MOM potential developed by Lemeshko identifies the VDAC-hexokinase complex as a potential-generating “battery” [70, 71]. Lemeshko has given estimates of the MOM potential as high as 50 mV, negative at the cytoplasmic side of the MOM. In his model, the Gibbs free energy of kinase reactions is the driving force for the potential generation across the MOM. The model gives a new physiological relevance not only to the VDAC voltage-gating but also to the channel’s interaction with hexokinases and other cytosolic regulators such as tubulin. This leads us to another key question, namely: how does VDAC’s interaction with protein partners influence its permeability to Ca²⁺?

5. Regulation of VDAC Ca²⁺ transport

VDAC is known to interact with a quite impressive number of various genetically, structurally, and physiologically unrelated cytosolic and mitochondrial membrane proteins (see reviews by [58, 72, 73]). The former include glycolytic enzymes, hexokinase, dimeric tubulin, and neuronal proteins intimately involved in different neurological diseases such as α-synuclein (αSyn), Aβ peptide, tau, and SOD1 [73–79]. The latter include the cholesterol transporter, translocator protein 18 kDa (TSPO) [80], and Bcl-2 family proteins that ultimately link VDAC to apoptosis [52, 53, 81, 82]. Potentially, all these proteins can affect VDAC Ca²⁺ permeability, thus influencing mitochondrial Ca²⁺ homeostasis. Here we focus on two well-studied cases of VDAC cytosolic regulators αSyn and tubulin, as well as VDAC’s neighbor in the MOM, TSPO.

αSyn is a small, 140-amino acids, intrinsically disordered protein highly expressed in neurons. The abnormal accumulations of αSyn are considered to be the pathological hallmark of PD [83]. It is the major component of the Lewy bodies found in substantia nigra of the post-mortem brains of the PD patients where it exists in the fibrillar form [84]. Based on these observations most of the studies on the role of αSyn in PD and other neurodegeneration were focused on the pathological role of the αSyn aggregates. Under normal conditions, αSyn constitutes up to 1% of the total cytosolic protein in neurons [85] where it exists in predominantly monomeric form. Small amounts of αSyn are also found in peripheral tissues [86]. The point mutations, genomic duplications, and triplications found in the gene encoding αSyn provide further genetic evidence of αSyn connection with familial and sporadic forms of PD [86]. Specific relationships between αSyn toxicity and mitochondria dysfunction are well established [87, 88]. The latest work has found intriguing relationships between a newly discovered mutation in the ITPKB kinase that is protective against sporadic PD, αSyn aggregation, and mitochondrial Ca²⁺ uptake [89]. Inhibition of ITPKB was shown to enhance ER-to-mitochondria Ca²⁺ uptake, subsequently exacerbating αSyn aggregation and PD pathology via the impairment of autophagy. However, the precise role of αSyn, especially its monomeric form, in neurodegeneration remains enigmatic. In particular, monomeric and oligomeric αSyn are found associated with both mitochondrial membranes [90, 91], causing impairment of the mitochondria respiratory complexes [92], an increase of oxidative stress [93] and fission [94]. Recently our group using in situ and in vitro experiments showed that αSyn translocates into the mitochondria through VDAC [49, 76, 95]. Neuronally differentiated SHSY5Y cells with overexpressed wild type αSyn have been chosen as an appropriate cell model of PD [96]. The involvement of VDAC1 in αSyn translocation into mitochondria was confirmed by VDAC1 knockdown in these cells with siRNA [49]. The question of whether the other two VDAC isoforms also serve as pathways f6or the αSyn entry into mitochondria remains open until further studies. The experiments with recombinant mammalian VDAC1 and VDAC3 reconstituted into planar lipid membranes showed that αSyn effectively blocks and translocates through both VDACs [57, 76] suggesting that most likely αSyn enters mitochondria through all VDAC isoforms, but with different efficiency.

Figure 5A shows a typical electrophysiological experiment with reconstituted VDAC in the presence of αSyn. The ion current through the single VDAC channel is quiet, and the channel does not gate at the relatively low applied voltage (25 mV as in the experiment shown in Figure 5A). The channel can remain open at these experimental conditions for up a few hours with occasional transitions between the open and lower conducting states. Addition of nanomolar concentration of αSyn (50 nM as in Figure 5A) changes the situation dramatically by inducing fast fluctuations of current between the open and a well-defined blocked state with ~ 40% of the open state conductance [76]. The time-resolved blockage events in the milliseconds range are clearly seen in the inset of Figure 5A. It should be noted that αSyn-induced blockages are in striking contrast with the voltage-induced VDAC closure which is characterized by much longer-lasting closures (10-100 s) with a variety of conductances (Figure 3B) [97]. αSyn blocks VDAC from both sides of the channel (added to the both sides of a planar membrane, Figure 3A) when a negative potential is applied from the side of αSyn addition [76]. When the sign of the potential is reversed, no blockage events are detected. This observation and the fact that αSyn with truncated C-terminal induces orders of magnitude less blockages than the wild-type [76], suggest that the highly negatively charged 45-residue C-terminal domain of αSyn is responsible for the channel blockage. On the other hand, a synthetic peptide of the C-terminus of αSyn does not measurably block the channel up to 0.5 mM [76]. Altogether, these observations led us to propose a molecular mechanism where the acidic C-terminus of αSyn is captured inside the net positive VDAC pore while being anchored to the membrane surface by its N-terminus, preventing free translocation through the pore (Figure 5C). A single polypeptide strand of αSyn fits comfortably into the ~2.5-2.7 nm diameter VDAC pore. The blockage time increases exponentially with applied voltage up to a certain voltage – a “turnover” potential – which amplitude depends on experimental conditions such as lipid composition and salt concentration [76, 98]. At potentials higher than the turnover potential, the blockage time decreases with voltage signifying translocation of αSyn through the pore [76]. Thus, under certain conditions, αSyn can translocate through the VDAC pore into the mitochondria and target the electron transport complexes in the MIM leading to their impairment and mitochondria dysfunction, the conjecture that was supported by the loss of mitochondria potential in cells with overexpressed wild-type αSyn [49, 96].

Figure 5. Regulation of VDAC Ca²⁺ permeability by α-synuclein.

(A) A representative experiment showing a current record through a single VDAC in 1M KCl before and after the addition of 50 nM of α-synuclein (αSyn). αSyn induces fast characteristic blockage events best seen in a finer time scale in Inset. (B) Current through a single VDAC in response to applied triangular voltage wave (50 mHz, ±60 mV) obtained in 150mM/30mM CaCl₂ gradient. Dashed blue and red lines indicate VDAC open and αSyn-blocked states, respectively. The intersection of these lines with the zero-current line (I = 0) corresponds to the reversal potential (Ψ_rev). Positive Ψ_rev corresponds to anionic selectivity of the open state and negative to cationic selectivity of the αSyn-blocked state. The voltage-induced closed states are indicated by the green arrow. Current records were digitally filtered using a 100 Hz filter. (C) A model of αSyn regulation of Ca²⁺ flux. Under applied negative voltage, acidic C-terminus of αSyn is captured inside the VDAC pore resulting in reversed VDAC selectivity and consequently in enhanced Ca²⁺ flux through the pore. When the channel is open, its slight anionic selectivity causes a reduced Ca²⁺ permeability and favors the fluxes of negatively charged metabolites such as ADP and ATP.

The fact most relevant for our current discussion is that when the highly negatively charged C-terminal domain of αSyn is transiently trapped inside the pore, the blocked state becomes cation selective [98, 99]. This leads to two important physiological consequences: impaired translocation of the negatively charged respiratory substrates such as ATP and ADP, which is due to the creation of electrostatic and steric barriers by the C-terminus inside the pore, and higher permeability for Ca²⁺ (Figure 5C). The increase in mitochondria uptake of Ca²⁺ released from the ER in response to αSyn overexpression in HeLa cells [100] supports the latter implication. The direct experimental evidence of the increased Ca²⁺ flux through the αSyn blocked state was obtained recently by our group in experiments with reconstituted VDAC [47]. The representative experiment in Figure 5B shows a single-channel trace obtained under the applied triangular voltage wave in a 150 mM /30 mM CaCl₂ gradient in the presence of 50 nM αSyn in the cis compartment. Fast fluctuations of current between the open (indicated by the blue dashed line in Figure 5B) and blocked (indicated by the red dashed line) states correspond to time-resolved individual blockage events induced by aSyn. It can be seen that the reversal potential, Ψ_rev, (the voltage corresponding to zero current) has an opposite sign for the open and blocked states. A positive Ψ_rev corresponds to anion and negative to cation selectivity (Figure 5B). These results demonstrate that VDAC blocked by αSyn is significantly more permeable for Ca²⁺ than the open channel [47]. Indeed, the calculated permeability ratio between Cl⁻ and Ca²⁺, P_Cl /P_Ca, reduces from 40.7 ± 9.8 for the open state to 0.58 ± 0.24 for the blocked state [47], which is a 70-fold change in favor of Ca²⁺ permeability in the blocked state. The cartoon in Figure 5C illustrates the proposed mechanism of αSyn regulation of Ca²⁺ transfer through VDAC. Therefore, αSyn interaction with VDAC might regulate mitochondrial Ca²⁺ uptake, but further genetic evidence is needed to conclusively surmise VDAC’s role in αSyn modulation of Ca²⁺ crosstalk between mitochondria and the ER [92, 100].

We speculate that another potent VDAC cytosolic regulator, dimeric tubulin [75, 95], can also regulate Ca²⁺ flux through VDAC by a similar mechanism as αSyn. It was shown that tubulin blocks the VDAC pore with its disordered negatively charged C-terminal tail and consequently reversed VDAC selectivity to cationic [62]. Therefore, we anticipate that the blockage of the VDAC pore by tubulin will increase Ca²⁺ permeability through the MOM.

VDAC is known to undergo post-translational modifications including phosphorylation [101, 102], ubiquitination [103, 104], and succination [105]. While none of these modifications have been shown to directly alter the Ca²⁺ transport properties of VDAC, they have been shown to affect VDAC interaction with other proteins that could in turn, indirectly affect Ca²⁺ signaling. For example, it was shown that phosphorylation of VDAC by glycogen synthase kinase-3β (GSK3β) in vitro, enhances VDAC blockage by tubulin [101]. Therefore, following the above speculations, VDAC phosphorylation could potentially affect its Ca²⁺ permeability via modulating VDAC interaction with tubulin. It was also suggested that depletion of the VDAC complex with GSK3ß and BCL-xL in response to reduced VDAC phosphorylation in steatosis may enhance mitochondrial permeability to Ca²⁺ and sensitize the cell to apoptosis[106]. Lastly, phosphorylation of VDAC was shown to displace the glycolytic enzyme Hexokinase2 (HK2) from VDAC1 [107, 108]. Concurrently the dissociation of HK2 from the MOM induces apoptosis via an influx of Ca²⁺ from the ER [109], thus letting us hypothesize that phosphorylation of VDAC may enhance mitochondrial Ca²⁺ uptake via HK2 displacement. Altogether, these observations open an intriguing possibility that other cytosolic proteins, yet to be discovered, could also regulate Ca²⁺ fluxes through VDAC. Future studies will reveal if this prediction is true.

Little is known about the mechanism of VDAC’s interaction with its neigboroughing membrane proteins in the crowded MOM environment. Apoptotic stimuli that evoke mitochondrial Ca²⁺ influx have been found to enhance VDAC expression and oligomer formation [110, 111]. VDAC’s natural potency to form ordered oligomeric arrays in MOM is well-known and was first demonstrated by electron micrographs by Mannella’s group [112] followed later by the work of Scheurung’s group [113] where supramolecular assembly of VDAC was demonstrated by atomic force microscopy in native MOM. Further, x-ray and NMR crystallography confirmed that VDAC could exist in dimeric conformation and form homo-oligomers of different sizes [114–116]. However, the important questions of how VDAC oligomerization affects Ca²⁺ homeostasis and whether oligomerization affects the Ca²⁺ permeability of the channel or regulates Ca²⁺ fluxes remain to be answered. Also, VDAC’s interaction and hetero-oligomerization with other MOM proteins and their potential role in Ca²⁺ regulation is also intriguing. For example, an increase in the expression of TSPO coincided with reduced mitochondrial Ca²⁺ uptake during heart failure, which was reversed in TSPO knockout mice [117]. The observed reduction in Ca²⁺ uptake could potentially be due to TSPO’s modulation of VDAC opening as hypothesized by Gatliff and Campanella [118]. Further studies on the interaction between TSPO and VDAC and its effect on VDAC Ca²⁺ transport are needed to understand their role in heart disease.

6. Role of VDAC in Ca²⁺ homeostasis in cardiac disease

Heart diseases are linked to mitochondrial dysfunction due to impaired Ca²⁺ uptake, bioenergetics, and mitophagy. For example, during ischemia-reperfusion injury, there is an increased translocation of GSK-3β to mitochondria which is known to modify VDAC by phosphorylation [119]. In addition, Gomez et al. showed that the inhibition of GSK-3β results in reduced Ca²⁺ overload and cell death [120]. Based on these two works we can hypothesize that VDAC phosphorylation may increase Ca²⁺ exchange through VDAC resulting in mitochondrial Ca²⁺ overload. Also, in a model of Duchenne muscular dystrophy, Viola et al. showed that the L-type Ca²⁺ channel regulates mitochondrial function by modulating VDAC function via cytoskeletal proteins [121]. Ca²⁺ influx through the L-type Ca²⁺ channel initiates contraction in the heart and increases mitochondrial membrane potential, ROS, NADH, and metabolic activity in a Ca²⁺-dependent manner [121]. The authors hypothesize that the absence of dystrophin results in the disarray of cytoskeletal proteins leading to the reduced metabolic activity due to the loss of communication between the L-type Ca²⁺ channel and VDAC through F-actin. However, perturbation of Ca²⁺-induced Ca²⁺ release (CICR) from the SR to mitochondria through the VDAC2-RyR2 channel complex due to cytoskeletal disarray may also affect energy production [122]

Given the lethality of VDAC2 knockout in mice models, it is of great interest to investigate the specific role of VDAC2 in cardiac function. During a heartbeat, the energy is required to translate the cardiac action potential to the physical contraction of cardiac myocytes by CICR to mitochondria through VDAC2-RyR2. Knockdown of VDAC2 increased intracellular Ca²⁺ levels upon electrical stimulation and decreased mitochondrial Ca²⁺ uptake after caffeine-induced activation of the RyR2 channels [61]. It was also found that VDAC2 knockdown modified mitochondrial Ca²⁺ uptake and the intensity and duration of SR Ca²⁺ sparks even though the autorhythmic and caffeine-induced Ca²⁺ transients were not affected [123]. Further evidence for VDAC2 as a modulator of rhythmic beating in the heart emerged when Shimizu and colleagues showed that overexpression of zebrafish VDAC2 (zfVDAC2) was antiarrhythmic in the zebrafish arrhythmia model [124]. They also found that a novel antiarrhythmic compound efsevin enhances mitochondrial uptake of Ca²⁺, which is mediated by its direct interaction with VDAC2, and the knockout of zfVDAC2 prevents efsevin’s protective effect. The authors proposed that efsevin enhanced zfVDAC2 permeability to Ca²⁺, thereby buffering aberrant Ca²⁺ signals in the heart. This was further confirmed by Wilting et al. who showed that efsevin binding to zfVDAC2 increases the channel’s cationic selectively, thus presumably increasing zfVDAC2 permeability to Ca²⁺ [125]. These studies show strong evidence for VDAC2’s specific role in Ca²⁺ homeostasis essential for proper cardiac function.

Taken together, the above studies tend to assign a specific role for VDAC, particularly VDAC2, in mitochondrial buffering of Ca²⁺ in cardiac physiology, suggesting that the lethality of VDAC2 knockout may be related to impaired cardiac function. The protective effect of efsevin in arrhythmia is promising, but further studies to discern the role of VDAC isoforms in ischemia/reperfusion and other heart diseases are needed to develop drugs to treat heart disease.

7. Conclusions and Perspectives

VDAC is well known for its role in shuttling metabolites between the cytosol and mitochondria. In this review, we have focused on its lesser-known function in Ca²⁺ transport and regulation of Ca²⁺ signaling at the mitochondria. VDAC is part of multi-protein complexes with specific organelle Ca²⁺ channels, which allows the generation of Ca²⁺ microdomains that facilitate rapid transfer of Ca²⁺ to mitochondria from the intracellular stores such as ER, SR, and lysosomes.

Biophysical studies of VDAC channel properties in the presence of Ca²⁺ suggest that VDAC could regulate Ca²⁺ fluxes to and from mitochondria by gating. Furthermore, these fluxes could be modulated by cytosolic VDAC regulators, such as neuronal protein αSyn. The distinct biophysical properties and physiological roles of VDAC isoforms in Ca²⁺ transport are especially interesting and warrant further study to understand the different effects of specific VDAC isoform knockouts, especially the lethality of VDAC2 knockout, in mice models. VDAC2 is also important in mitochondrial function in cardiac cells and could potentially be a target for drug development to treat heart diseases.

We hope that this review will help researchers to appreciate the contemporary state of research on VDAC’s role in Ca²⁺ homeostasis and promote further investigations into the isoform-specific role of VDACs and their regulation in Ca²⁺ signaling and mitochondrial function. As we move on from the traditional view of VDAC as an inert diffusion pore, we expect that future research will uncover the mechanisms by which VDAC controls mitochondrial Ca²⁺ fluxes.

Highlights.

• VDAC plays an important role in mitochondrial calcium transport, regulation, and signaling.

• VDAC is part of multi-protein complexes linking it to calcium channels in other organelles for efficient calcium exchange.

• VDAC calcium transport can be regulated by voltage gating and by its interaction with proteins such as α-synuclein.

• Calcium transport through VDAC is isoform specific.