MCUB and mitochondrial calcium uptake – modeling, function and therapeutic potential

1. Introduction

Mitochondria are essential cellular organelles for energy production and calcium (Ca²⁺) critically regulates their function. At physiological levels, mitochondrial calcium (_mCa²⁺) uptake stimulates dehydrogenases and promotes bioenergetics. In pathological conditions such as ischemia-reperfusion (IR) injury, heart failure, skeletal muscle dystrophy, and neurodegenerative diseases, cell signaling causes sustained elevations in _cCa²⁺ levels by many mechanisms¹. In these conditions, mitochondria become overloaded with Ca²⁺ resulting in metabolic derangement, activation of Ca²⁺-dependent proteases, and excessive reactive oxygen species (ROS) generation; all mechanisms that contribute to cellular demise. Mechanistically, Ca²⁺ and ROS are prominent sensitizers of the mitochondrial permeability transition pore (mPTP), a large channel at the inner mitochondrial membrane (IMM) which permits ions and solutes up to 1.5 kD to diffuse causing matrix swelling, membrane reorganization, mitochondrial membrane potential collapse, and loss of the electrochemical gradient, thus disrupting oxidative phosphorylation². Cristae remodeling and outer mitochondrial membrane permeability allows apoptogen release initiating the caspase-mediated apoptotic cascade^{3, 4}. Additionally, if ATP is depleted due to loss of mitochondrial function Ca²⁺ activation of mPTP results in necrotic cell death⁵. Thus, preventing _mCa²⁺ overload is an attractive therapeutic target for a wide range of diseases, yet to date, no entities targeting _mCa²⁺ exchange have entered clinical trials. This lag in therapeutic application can be explained by the fact that the molecular components mediating _mCa²⁺ exchange were only recently identified⁶. In this editorial, we will discuss current strategies to prevent _mCa²⁺ overload and highlight MCUB (mitochondrial calcium uniporter dominant negative β, formerly known as coiled-coil domain containing 109B (CCDC109B), as a molecular target to modulate _mCa²⁺ uptake and reduce _mCa²⁺ overload in the context of disease⁷.

2. Targeting _mCa²⁺ uptake – The Mitochondrial Calcium Uniporter Channel (mtCU)

The genetic identification of the mtCU quickly led to the realization of considerable complexity due to its macromolecular nature. The precise stoichiometry and molecular weight (MW) of the uniporter remain unknown, but in general blue native page and size-exclusion chromatography suggests a complex of ~600 – 800 kD in size^{7, 8}. MCU (Mitochondrial Calcium Uniporter), the pore-forming subunit necessary for Ca²⁺ permeation, likely predominantly exists as a tetramer in a ‘dimer of dimers’ configuration⁹. The N- and C-terminus face the matrix, with two transmembrane domains forming a hydrophilic channel containing a highly conserved DIME sequence necessary for Ca²⁺ selectivity. MICU1 and MICU2 (Mitochondrial Calcium Uptake 1/2) contain EF-hand motifs to mediate Ca²⁺ binding to regulate channel opening. They inhibit channel activity at low _cCa²⁺ concentrations and sense elevations in cytosolic Ca²⁺ levels to increase open probability (gatekeeping function)^10–12. EMRE, (Essential MCU Regulator) is another mtCU subunit necessary for uniporter function, which is a conduit for the interaction between MCU and MICU1^{13, 14}. MCUB, a gene paralog of the pore-forming subunit, was originally proposed to exert a dominant negative effect on uniporter function¹⁵. MCUB shares ~50% amino acid sequence identity with MCU and contains an identical DIME sequence shown to be critical for Ca²⁺ pore selectivity calling into question MCUBs true mechanism of action¹⁵.

The mtCU has been implicated in various disease conditions since it’s the primary conduit for _mCa²⁺ overload. In humans MICU1 loss-of-function mutations results in neurological disorders and skeletal muscle myopathies, presumably due to loss of gatekeeping and _mCa²⁺ overload¹⁶. Mutations in the mitochondrial m-AAA proteases, which are associated with spinocerebellar ataxia and hereditary spastic paraplegia, were recently proposed to result in the accumulation of hyperactive mtCUs resulting in _mCa²⁺ overload¹⁷. This work mechanistically linked the failure to degrade EMRE to the accumulation of constitutively active uniporter channels. This directly links mtCU stoichiometry and assembly as critical to both proper function and disease.

Recent work in several disease models has revealed that targeting mtCU can indeed be protective against stressors and disease states featuring _mCa²⁺ overload. Reducing _mCa²⁺ uptake by MCU knockdown prevented glutamate excitotoxicity in primary neurons¹⁸. mcu deletion prevented _mCa²⁺ overload and dopaminergic neuronal loss in zebrafish models of Parkinson’s¹⁹. Our group showed that conditional Mcu deletion in the adult mammalian heart (Mcu cKO) improved resistance to Ca²⁺ induced mitochondrial permeability transition and was sufficient to reduce infarct size and cardiac pathology following ischemia-reperfusion injury^{20, 21}. This strongly suggests that targeting the mtCU to reduce _mCa²⁺ overload has therapeutic potential. However, loss of MCU did diminish contractile responsiveness to isoproterenol (adrenergic signaling), suggesting mtCU-dependent Ca²⁺ uptake is necessary for cellular bioenergetics during stress²⁰. This indicates that targeting MCU expression itself, or completely blocking _mCa²⁺ uptake, is likely to come with negative consequences. Given the presumed function of MCUB to negatively regulate the mtCU, it presented as an attractive target to reduce _mCa²⁺ overload without completely ablating the physiological/signaling function of matrix Ca²⁺. To discern how MCUB modulates the mtCU and examine if it was amendable to therapeutic manipulation, we recently generated gain- and loss-of-function models to explore MCUB’s mechanism of action and relevance to cardiac physiology⁷.

3. New insights regarding MCUB function in physiology and disease

Our first evidence that mechanisms were at play to modulate the assembly of the mtCU was that FPLC size-exclusion chromatography in MCUB^−/− cells revealed an overall increase in high-molecular weight (MW) complexes containing MCU, MICU1, and MICU2⁷. Conversely, overexpression of MCUB in cardiomyocytes decreased the levels of MCU, MICU1, and MICU2 in the mtCU complex⁷. These findings, coupled with the discovery that unlike MCU, MCUB does not directly interact with MICU1 or MICU2 suggests that the incorporation of MCUB into the uniporter channel displaces MCU and MCU-bound MICU1/2⁷. Physiologically, the incorporation of MCUB appears to be a stress-responsive mechanism to alter mtCU gating and limit cooperative activation of the channel. In support of this notion, we found that MCUB was not expressed in the high-MW uniporter complex of the heart at baseline, but only after stress or ischemic injury⁷. Functionally, we found that loss of MCUB increased channel capacitance and amplified cooperative activation resulting in enhanced _mCa²⁺ uptake at a given cytosolic Ca²⁺ level⁷. Conversely, forced expression of MCUB reduced _mCa²⁺ uptake and mitochondrial swelling in response to calcium challenge⁷. Altogether, these results paint a picture that the induction of MCUB is a compensatory mechanism to reduce _mCa²⁺ uptake and overload by altering the stoichiometry of regulatory subunits in the uniporter complex.

The controlled expression of MCUB in cardiomyocytes (flox-stop transgenic approach) in the adult heart revealed some interesting physiological findings. First, MCUB expression was sufficient to reduce _mCa²⁺ transients on a beat-to-beat basis in slowly paced cardiomyocytes⁷. It should be noted that we employed a very slow pacing frequency (0.1 Hz) to allow the assessment individual transients, as our group²² and others^{23, 24} have shown a summation of transients at higher pacing frequency due to slower efflux, such that resolving and interpreting uptake dynamics becomes difficult. The MCUB-mediated reduction in _mCa²⁺ uptake caused a huge reduction in contractile responsiveness to the beta-adrenergic agonist, isoproterenol⁷. Surprisingly, this effect was even greater than what we observed with deletion of Mcu^{20, 21}, which we think is due to temporal changes in compensatory mechanisms that will be addressed later. Our data suggested that contractile deficiencies were likely due to a loss of Ca²⁺-dependent activation of mitochondrial dehydrogenases, since we observed a significant increase in phosphorylated-pyruvate dehydrogenase (p-PDH), both at baseline and during adrenergic stimulation⁷. (Dephosphorylation of PDH at the E1α subunit is mediated by matrix Ca²⁺ activation of PDH phosphatase and results in increased PDH activity). In agreement, the assessment of oxygen consumption rates of adult cardiomyocytes revealed a substantial decrease in respiratory reserve capacity suggesting an overall impairment in cellular bioenergetics⁷. This energetic compromise was so significant that our first experiments of in vivo ischemia-reperfusion (IR) one week after the induction of MCUB expression resulted in lethality within the first 10 minutes of ischemia⁷. Our interpretation is that mtCU-Ca²⁺ uptake is needed to activate mitochondrial energetics in the non-ischemic/remote zone to maintain contractility and cardiac output in the severely compromised heart. Further investigation found that the expression of MCUB alone, elicited a transient decrease in cardiac function one week following MCUB expression, which was resolved three weeks later⁷. Our data suggested compensatory changes in mitochondrial energetics one month following MCUB expression was sufficient to restore oxidative reserve capacity and contractile function in response to adrenergic stimulation. We hypothesize that MCUB-expressing mitochondria activated calcium-independent mechanisms to regulate energetics since the phosphorylation of PDH at baseline and in response to isoproterenol was restored to control levels, even though _mCa²⁺ uptake was still blunted. In fact, a second IR study one month after MCUB expression found a significant reduction in infarct size and Ca²⁺-overload induction of mPTP and mitochondrial swelling⁷. The contrasting temporal phenotype of MCUB expression highlights the strong compensatory mechanisms that are induced in response to impaired _mCa²⁺ uptake and may explain the divergent phenotypes in various mtCU genetic models, particularly those with constitutive or germline alterations in expression^25–27. It’s likely that key mitochondrial calcium control points are rewired to maintain energetics by other means. For example, PDH activity is not only regulated by Ca²⁺ activation of the phosphatase, but also by Acetyl-CoA, NAD(H), pyruvate and fructose-1,6-bisphosphate levels. Thereby, compensatory alterations in the generation and consumption of metabolites is likely employed to counteract the loss of _mCa²⁺ uptake.

No effective therapeutics exist to prevent or reduce _mCa²⁺ overload. The only current strategy to block _mCa²⁺ uptake is Ruthenium Red derivatives, such as Ru360, which directly bind MCU to inhibit channel function^{28, 29}. Ruthenium Red derivatives lack specificity^30–32 and Ru360 is not very cell permeable²⁹ making these compounds unsuitable for clinical translation. However, significant advances in drug design are generating more selective, cell permeable MCU inhibitors that may have in vivo potential³³. Further, the tissue-dependent differences in mtCU density and composition^{34, 35} suggest the use of a general MCU inhibitor may have substantial on-target deleterious effects. The complexity of channel formation, stoichiometry and regulation suggests more basic research is needed to define the mechanistic targets with greatest therapeutic potential. MCUB represents an exciting alternative for regulating mtCU function to limit _mCa²⁺ overload. MCUB contains the exact same DIME sequence as MCU, which is critical for creation of the Ca²⁺ permeable pore. This calls into question exactly how MCUB limits _mCa²⁺ uptake. Does MCUB alter the structure of the pore or is the majority of regulation by way of altering channel composition and gating as we propose here? Previous work by the Rizzuto group¹⁵ suggests that MCUB cannot form a Ca²⁺ permeable channel independent of MCU, but the evidence for this negative result is somewhat limited. In fact, the addition of excess MCUB did not alter the electrophysiological properties of the uniporter and had a negligible effect on channel open probability in their experiments¹⁵. Further, MCUB alone could give rise to a functional channel permissive to Na⁺ conduction.

4. Expert Opinion on MCUB

We used Modeller³⁶ to align human MCU to the fungal MCU crystallography data reported by Fan et al.³⁷ and used this theoretical framework for structural prediction of channel formation by human MCUB (Figure 1–2). It should be noted that using Modeller has the caveat of lacking significant energy minimization, so the modelled structures may not represent the lowest energy state found in cells. Interestingly, there is striking similarity between the predicted MCU and MCUB structure both at the monomer (Figure 1A) and tetramer (Figure 1B) configurations particularly in the configuration of the pore where minor changes in pore size are observed (Figure 1C–D). It can also be appreciated that structurally the N- and C- termini of MCU and MCUB are more divergent. Altogether, this argues against the mechanism of action being solely due to alterations in Ca²⁺ permeation. Perhaps the mtCU is still capable of some minimal Ca²⁺ conduction when MCUB is present. It is even possible that MCUBs mechanism of action is independent from modulating the actual channel or pore region and is a direct result of altered gating/MICU interacting domains.

Figure 1. Structural modeling of MCUB and the mitochondrial calcium uniporter (mtCU).

Using Modeller³⁶, a program for comparative three dimensional protein structure analysis³⁸, we uploaded the structural data for fungal MCU (PDB 6D7W) reported in Fan et al.³⁷ and predicted the structure of human MCU followed by subsequent prediction of human MCUB. A) Comparison of monomeric structure. B) Comparison of human MCU and MCUB homo-oligomeric tetrameric structure. C) Zoom view of the mouth of the hydrophilic ionic pore formed by MCU tetramer, MCUB homo-oligomeric tetramer and a hetero-oligomeric tetramer (1 MCUB : 3 MCU). Red sphere represents a calcium (Ca²⁺) ion and purple and orange highlight the DIME motif. The dashed white line is a theoretic measurement in angstroms (A) between Aspartate-261 MCU residues or Aspartate-322 MCUB residues. D) Zoom of the sideview of opposing DIME motifs of MCU, MCUB, and MCU/MCUB hetero-oligomeric structure.

Figure 2. mitochondrial calcium uniporter (mtCU) electrostatic interactions and transmembrane domains.

A) Electrostatic map of MCU and MCUB tetramers and monomers. B) MCU and MCUB transmembrane domain interactions of MCU homo-oligomeric tetramer vs. hetero-oligomeric tetramer. Critical amino acid residues likely mediating hydrophobic interactions are highlighted.

Notably, Raffaello et al. identified two key amino acid residues within MCUB that differ from MCU, (MCU 251R, 256E vs. MCUB 251W, 256V), which they proposed may influence Ca²⁺ permeation of the pore¹⁵. Our structural models (Fig 2) indicate MCU and MCUB transmembrane domains vary in several amino acid residues which may alter Ca²⁺ binding capacity and pore structure, assuming a tetramer for discussion purposes. Transmembrane domain-1 (TM1) of MCU contains the hydrophobic residues Leu, Ala, Phe, Ala, and Trp, which interact with hydrophobic and polar-neutral amino acids Phe, Thr, Met and Tyr of TM2 in the neighboring MCU subunit. The insertion of MCUB into the mtCU, may disrupt these interactions, since TM1 of MCUB contains a polar-neutral Ser and a Gly in place of the MCU hydrophobic Ala and Phe residues. TM2 of MCUB contains a hydrophobic aromatic Phe residue in place of MCU’s polar neutral Thr. Altogether, our minimalist modeling suggests that these differing interactions result in a slight widening of the initial opening of the ionic pore.

Our proteomic data, indicate that MCUB does not interact with MICU1. Interestingly, the newly discovered DIME interacting-domain of MCU is proposed to be a critical interface of MICU1 and MCU¹⁴. The original amino acids Raffaello et al.¹⁵ identified in MCUB lie in very close proximity to this DIME interacting-domain and it is of significant interest to investigate if these amino acid differences are not only responsible for MCUBs lack of interaction with MICU1, but also for MCUBs mechanism of action.

Our study suggests a major molecular mechanism for MCUB regulation appears to be alterations in overall mtCU assembly. This is unique because it suggests alternatives to drug strategies that only target the pore and instead proposes that targeting the assembly or subunit stoichiometry is a new strategy to reduce _mCa²⁺ overload (Figure 3). Further research is needed to determine the exact amino acids responsible for MCUBs differential function in comparison to MCU. This insight may guide the generation of small peptides to modulate mtCU function in a cell-specific manner. In summary, even with the plethora of new studies on mtCU structure and function, much remains to be discovered about the molecular mechanisms governing the mtCU to aid the development of novel therapeutics to prevent the _mCa²⁺ overload that occurs in numerous disease states.

Figure 3. Mitochondrial calcium uniporter (mtCU) structure and function amendable to therapeutic targeting.

A) Ruthenium red derivatives, such as Ru360, currently are the only major class of inhibitors of mitochondrial calcium (_mCa²⁺) uptake. B) The mitochondrial AAA+ proteases are localized to the inner mitochondrial membrane and functionally target proteins for degradation. Lack of Afgl32 and Spg7 lead to excessive EMRE accumulation resulting in overactive MCU, mitochondrial calcium overload, and neurodegeneration which is implicated in the conditions of spinocerebellar ataxia and hereditary spastic paraplegia. This directly implicates the regulation of mtCU assembly in health and disease and suggests protease activity may be a potential therapeutic target. C) MICU1/2/3 (mitochondrial calcium uptake 1/2/3) are paralogs with functional distinction, as each member sets a different calcium threshold for entry and uniquely influences cooperative activation of mtCU. Therefore, modulating the expression or localization of MICUs may present a novel way to mitigate _mCa²⁺ overload. D) MCUB displaces MCU and MCU-bound MICU1/2 from mtCU, thus altering stoichiometry and assembly. Functionally, MCUB reduces _mCa²⁺ uptake and critically can prevent _mCa²⁺ overload during stress. E) Post-translational modifications such as phosphorylation^{39, 40} or glutathionylation⁴¹ have been reported to modulate the uniporter, opening the door to therapeutic targeting of the kinome, etc.