Mitochondrial Magnesium is the cationic rheostat for MCU-mediated mitochondrial Ca2+ uptake

Calcium (Ca2+) uptake by mitochondria is essential in regulating bioenergetics, cell death, and cytosolic Ca2+ transients. Mitochondrial Calcium Uniporter (MCU) mediates the mitochondrial Ca2+ uptake. MCU is a heterooligomeric complex with a pore-forming component and accessory proteins required for channel activity. Though MCU regulation by MICUs is unequivocally established, there needs to be more knowledge of whether divalent cations regulate MCU. Here we set out to understand the mitochondrial matrix Mg2+-dependent regulation of MCU activity. We showed Mrs2 as the authentic mammalian mitochondrial Mg2+ channel using the planar lipid bilayer recordings. Using a liver-specific Mrs2 KO mouse model, we showed that decreased matrix [Mg2+] is associated with increased MCU activity and matrix Ca2+ overload. The disruption of Mg2+dependent MCU regulation significantly prompted mitochondrial permeability transition pore opening-mediated cell death during tissue IR injury. Our findings support a critical role for mMg2+ in regulating MCU activity and attenuating mCa2+ overload.

Though many Ca 2+ channels, including InP 3 R 29 , RyRs 30 , and CRAC 31,32 , exhibit Ca 2+ -dependent feedback mechanisms, whether divalent cations regulate MCU activity is unclear. To understand in depth the regulation of MCU by divalent cations, we resolved the atomic structure of the conserved MCU matrix domain 33 . Our previous structural analysis of the amino-terminal domain of human MCU revealed a -grasplike fold containing the MCU regulating acidic patch (MRAP) that binds Mg 2+ /Ca 2+ with ~mM affinity. The binding of Mg 2+ /Ca 2+ to MRAP destabilizes MCU, shifts the self-association equilibrium to monomer, and attenuates mCa 2+ uptake 33 . A seminal work from Foskett's group showed the Ca 2+ that permeates through the channel pore to bind the MRAP region and close the channel despite Ca 2+ binding to MICU1/2 34 . Also, mutating the two key aspartic acid residues (D131 and D147) to alanine abolished matrix [Ca 2+ ] dependent channel inhibition, validating divalent cation binding site encompassing D131 and D147 could account for Ca 2+ (and Mg 2+ ) dependent MCU inhibition 34 . Though matrix Ca 2+ was shown to inhibit MCU channel activity, the weak binding affinity for divalent cation for MRAP is well suited to the high Mg 2+ levels of the mitochondrial matrix (0.5-1.5 mM). Because proteins structurally sensitive to cations have evolved affinities close to the concentration range in the compartment where they function, we anticipate mitochondrial matrix [Mg 2+ ] to regulate MCU channel activity tightly in addition to Ca 2+ .
Here we set out to understand the matrix [Mg 2+ ] dependent regulation of MCU-channel activity. Using biochemical, lipid bilayer, and patch-clamp recordings, we show Mrs2 as the authentic mammalian mitochondrial Mg 2+ channel. Using our newly developed liver-specific mouse model, we illustrate the loss of Mrs2 to drastically alter the matrix [Mg 2+ ]. The decreased matrix [Mg 2+ ] was associated with increased MCU activity and matrix Ca 2+ overload. The disruption of Mg 2+ -dependent MCU activity also significantly prompted mitochondrial permeability transition pore opening (mPTP). We employed a hanging-weight system for liver ischemia/reperfusion (IR) to assess the loss of Mg 2+ -dependent MCU regulation and mPTP-mediated cell death during tissue IR injury. Our findings support a critical role for mMg 2+ in regulating MCU activity and attenuating mCa 2+ overload.
Mitochondrial RNA splicing 2 (Mrs2) was initially considered the primary Mg 2+ transporter in Saccharomyces cerevisiae mitochondria 35,36 . The electrophysiological studies showed yeast Mrs2 (sMrs2) to form an Mg 2+ -selective channel (155 pS) and not permeable to Ca 2+ 36 . Though the core/pore component of the mMg 2+ influx machinery was defined in yeast, the physiologic relevance of this observation in higher-order systems remains understudied. A recent study from Madesh and colleagues showed mammalian Mrs2 (mMrs2) to form a selective pore for lactate-mediated mMg 2+ uptake 37 , authenticating Mrs2 as the conserved mMg 2+ transport machinery from yeast to mammals. Though mMg 2+ influx machinery was defined in mammals, the electrophysiological properties of mMrs2 are not established. To show whether mMrs2 forms an Mg 2+selective channel, we purified mMrs2 from HEK293 cells (Fig. 1d). The highly purified Mrs2 was used in planar lipid bilayer recordings for studying its single-channel activity. Our recordings show mMrs2 to form a voltagegated Mg 2+ selective channel with multiple subconductance states ( Fig. 1e-g). We recorded channels with the peak conductance activity varying from 80pS-250pS and mean conductance activity of ~150 pS ( Fig. 1e-1j).
Single-channel currents were recorded in the presence of Mg 2+ only in Na + -, K + -, Cl --, and Ca 2+ -free buffer and were shown to be inhibited by Co 2+ (Fig. 1h-j). The continuous ramp voltage recording of the Mrs2 channel from −100 mV to +100 mV shows that the channel is inhibited by Co 2+ at all voltages (Fig. 1h). Fig. 1i shows a continuous lipid bilayer recording of Mrs2 before and after adding Co 2+ .
After confirming, Mrs2 channel activity, we extracted RNA and proteins from a panel of adult mouse tissues to verify whether Mrs2 is ubiquitously expressed in all mammalian tissues. Though differentially expressed, qPCR and Western blot analysis show, Mrs2 to be ubiquitously present in all metabolically active tissues ( Fig. 2a and 2b; supplementary fig. 1a-1c). After confirming the ubiquitous distribution of Mrs2, we asked whether Mrs2 localizes explicitly to the mammalian mitochondria. To verify the mitochondrial distribution of Mrs2, we performed confocal microscopy and super-resolution structured illumination microscopy (SIM) imaging in HeLa cells expressing Mrs2-GFP and Cox8-mRFP. Confocal images at 488/561 nm and the intensity profile analysis overlap GFP and mRFP signals ( Fig. 2c and 2d). Additionally, the SIM images and the correlation analysis substantiate a mitochondrial localization of Mrs2 ( Fig. 2e and 2f). After confirming the localization of Mrs2 to the mitochondria, we performed a Western blot analysis to study the sub-organellar localization and orientation of Mrs2 12 . HEK 293 cells stably expressing Mrs2-Flag were permeabilized with digitonin (40 µg/ml). Digitonin-permeabilized cells were incubated with mastoparan or alamethicin to elicit outer or inner mitochondrial membrane permeabilization. OMM permeabilization (with mastoparan) was confirmed by the cytosolic appearance of cytochrome C (Fig. 2g). Permeabilization of both OMM and IMM (with alamethicin) was marked by the cytosolic appearance of cyclophilin D (matrix protein) (Fig. 2g). Mrs2 was observed in the membrane fraction in both mastoparan or alamethicin-treated cells, revealing Mrs2 as an integral membrane protein (Fig. 2g). The orientation of Mrs2 was evaluated by mitochondrial subfractionation 18 . Mitoplasts prepared from HEK293 cells stably expressing C-terminal Flag-tagged Mrs2 were subjected to proteinase K digestion followed by Western blot analysis with antibodies specific for Flag suggested Mrs2 to be enriched in mitoplasts (Fig. 2h). A mitochondrial matrix protein, CypD, was protected from proteinase K digestion (Fig. 2h). Similarly, the integral membrane protein Mrs2 remained stable with no loss of Flag-tag, suggesting the C-terminal end of Mrs2 to face the mitochondrial matrix. These results suggest that N and C termini face the mitochondrial matrix side, consistent with two transmembrane-spanning regions ( Fig. 2i).
Because we confirmed Mrs2 as the mMg 2+ influx machinery, we anticipate losing Mrs2 to decrease matrix [Mg 2+ ] and positively regulate MCU activity. To verify whether matrix [Mg 2+ ] regulates MCU activity, we generated a liver-specific Mrs2 KO mouse model. Loxp/loxp knockin mice (Mrs2 fl/fl ) were generated using CRISPR/Cas9 strategy. Mrs2 fl/fl mice were crossed with Albumin-Cre to allow germline deletion specifically in the liver (Mrs2 hep ) (Fig.3a). The gene targeting was confirmed by genotyping (Fig.3b). The loss of Mrs2 was confirmed using qPCR and Western blot analysis ( Fig.3c-3e). The loss of Mrs2 did not alter the expression of other mitochondrial proteins like the OXPHOS complex, MCU, MICU1, and MCUR1 ( Fig.3d-3e). We next asked whether loss of Mrs2 alters the mMg 2+ uptake. Hepatocytes isolated from Mrs2 fl/fl and Mrs2 hep were permeabilized, and mMg 2+ uptake and mitochondrial membrane potential ( m ) were measured simultaneously ( Fig. 3f and 3g) using a multi-wavelength-excitation dual wavelength-emission spectrofluorometer at 340/380 nm. The ratiometric dye, Mag-Fura-2 tetra potassium salt, was calibrated, and the mitochondrial [Mg 2+ ] was determined by 2 / where the dissociation constant Kd is 1.98 mM.
Mrs2 hep mitochondria showed no mMg 2+ uptake in response to extramitochondrial Mg 2+ pulse ( Fig. 3g-3i). The loss of mMg 2+ uptake was not due to changes in  m ( Fig. 3f and 3j; supplementary Fig. 2a and 2b). We next asked whether the loss of Mrs2 caused a change in the total matrix [Mg 2+ ] during the resting state.
To validate the role of mMg 2+ in regulating MCU-mediated mCa 2+ uptake, we used our permeabilized cell system to perform simultaneous measurements of mCa 2+ uptake and  m 27, 38 ( Fig. 4a and 4b).  Fig. 3a-3c) and decreased mMg 2+ uptake ( Supplementary Fig. 3d and 3e) in HEK293 cells to alter MCU activity (Supplementary Fig. 3f-3i). Previously we showed the binding of Mg 2+ to the MRAP region to destabilize and shift the self-association equilibrium of MCU to monomer 33 .
Consistent with our previous finding, FPLC analysis showed a heterogenous population of MCU complex (both low and high molecular weight) in HEK 293 NegShRNA cells, whereas MCU assembled as a supercomplex in Fig. 3j and 3k). Collectively, these results suggest the loss of mMg 2+ to stabilize the MCU complex and promote channel activity.

Mrs2 KD cells (Supplementary
We next asked whether increased MCU activity overloads matrix Ca 2+ at the resting state. Permeabilized cells loaded with Fura-FF tetra potassium salt were treated with Ru360 (to block any extramitochondrial Ca 2+ uptake) and CGP (to block the efflux of mitochondrial Ca 2+ ). After baseline measurement, an uncoupler CCCP was added to depolarize the mitochondrial membrane (Fig. 4h). Though MCU activity was increased in Mrs2 hep , we do not observe an increase in resting matrix [Ca 2+ ] ( Fig. 4h and   4j). Interestingly, when Ru360 was removed from the experimental condition described above, we observed the matrix [Ca 2+ ] significantly higher in Mrs2 hep than Mrs2 fl/fl ( Fig. 4i and 4j 39 . We used our permeabilized cell system to simultaneously measure mCa 2+ uptake and m (Supplementary Fig. 4a-4i). Mrs2 hep rapidly cleared the 1 M Ca 2+ pulse, whereas the channel gatekeeping function remained intact in Mrs2 fl/fl (Supplementary Fig. 4b, 4d-4e). Additionally, the bath Ca 2+ before adding Ca 2+ pulse was lower in Mrs2 hep compared to Mrs2 fl/fl (Supplementary Fig. 4f and   4g). Also, the bath Ca 2+ after CCCP addition was significantly higher in Mrs2 hep than Mrs2 fl/fl ( Supplementary   Fig. 4h and 4i Fig. 4j).
Since we saw increased MCU activity in the low-[Ca 2+ ]i regime despite stable MCU/MICU1 interaction, we asked whether the increased mCa 2+ uptake is strictly MCU-mediated. We stably expressed MCU DIME mutant in HEK293 NegShRNA and Mrs2KD cells. MCU DIME expression significantly decreased MCU-mediated mCa 2+ uptake in both control and KD cells (Fig. 4k-4n (Fig. 5a and 5b). The decrease in PDH phosphorylation can be attributed to increased matrix [Ca 2+ ], but the loss in complete PDH dephosphorylation marks the critical need for matrix Mg 2+ to activate PDHP.
Because we saw a trend in decreased PDH phosphorylation, we assessed the ATP levels in control and KO/KD cells. ATP was significantly decreased in KO/KD cells compared to control (Fig. 5c, Supplementary   Fig. 5a). The decrease in ATP levels was marked by an increase (not significant) in AMPK phosphorylation ( Fig. 5a and 5b).
Because we saw opposing results with PDH activation and ATP levels, we assessed the oxygen consumption rate (OCR) in control and Mrs2 KO/KD cells (Fig. 5d, 5e, Supplementary Fig. 5b, and 5c).
Basal and maximal respiration was increased in KO/KD cells compared to the control (Fig. 5d, 5e,   Supplementary Fig. 5b, and 5c). Further analysis of the OCR data showed increased proton leak in KO/KD cells with a concomitant decrease in ATP-coupled respiration (Fig. 5d, 5e, Supplementary Fig. 5b, and 5c).
Though the proton leak was increased in Mrs2 KO/KD cells, we do not see a decrease in m. Instead, a significant increase in m was observed (Fig. 3f, 3j According to the chemiosmotic theory, the m can be maintained either by the electron transport chain through respiration or by ATP hydrolysis via the F1-F0 ATPase. Because we observed increased proton leak and decreased ATP, we asked whether the reversal of F1-F0 ATPase contributes to increased m. Using our permeabilized cell system, we measured m in Mrs2 fl/fl and Mrs2 hep in the presence or absence of Oligomycin (ATP synthase and hydrolysis inhibitor) or BTB06584 (selective inhibitor of ATP hydrolysis). Both Oligomycin and BTB treatment depolarized/normalized the m in Mrs2 hep . Conversely, oligomycin treatment hyperpolarized the Mrs2 fl/fl mitochondria, and BTB had no effect (Supplementary Fig. 5d and 5e). Because the reversal of ATP synthase depletes cellular ATP and activates the mitochondrial permeability transition pore (mPTP) 41 , we next asked whether increased matrix Ca 2+ overload and ATP reversal make Mrs2 hep susceptible to mPTP opening. We simultaneously measured calcium retention capacity (CRC) and m in Mrs2 fl/fl and Mrs2 hep . Mrs2 hep exhibited decreased CRC associated with an early m collapse ( Fig. 5f-5j), indicating matrix Ca 2+ overload-induced activation of the mPTP. Increased Ca 2+ -induced mitochondrial swelling (Fig. 5k), early TMRM/calcein fluorescence loss ( Fig. 5l and 5m), and increased ionomycin-mediated cell death confirmed matrix Ca 2+ overload-induced mPTP opening in Mrs2 hep ( Fig. 5n and 5o).
Next, we used gain/loss-of-function Mrs2 mutants to examine whether the increased MCU activity and matrix Ca 2+ overload-induced activation of the mPTP are causal effects of decreased mMg 2+ uptake. It has been proposed that single amino acid substitutions in the G-M-N motif of Mrs2 are sufficient to abolish Mg 2+ transport or profoundly change the ion selectivity of channel [42][43][44] . Additionally, reports had shown eventually complete inhibition of CorA-driven Mg 2+ currents when the intracellular domain of CorA was exposed to mM concentration of Mg 2+ 45, 46 . This phenomenon could be due to a self-regulatory mechanism, where an increase in the local [Mg 2+ ] saturates the putative Mg 2+ binding site (MBS), triggering channel closure 46 . We generated loss-of-function (Mrs2 GMN ) and gain-of-function (Mrs2 MBS ) mutant constructs (Fig. 6a). Western blot confirms the ectopic expression of Flag-tagged Mrs2 WT , Mrs2 GMN , and Mrs2 MBS in HEK293 WT cells (Fig. 6b). The permeabilized cell system analysis ( Fig. 6c-6g) show abolished and increased mMg 2+ uptake in Mrs2 GMN and Mrs2 MBS , respectively ( Fig. 6d, 6f, and 6g). The expression of Mrs2 WT did not alter the mMg 2+ uptake, further authenticating mMg 2+ homeostasis to be maintained through a negative feedback loop 45,46 . We next sought to define whether altered mMg 2+ uptake modifies mCa 2+ uptake. Simultaneous measurements of m and mCa 2+ ( Fig. 6h-6l) show increased and decreased mCa 2+ uptake in Mrs2 GMN and Mrs2 MBS , respectively ( Fig. 6i, 6k, and 6l). Similar to Mrs2KO/KD cells, basal and maximal OCR was increased in Mrs2 GMN ( Fig. 6m and 6n) with a decrease in ATP levels (Fig. 6o). The decrease in ATP levels and increased m was due to increased proton leak (Fig. 6c, 6e, 6h, 6j, and 6n), further confirming the loss of mMg 2+ to result in ATP synthase reversal. Our results also show decreased CRC ( Fig. 6p-6s) and early calcein quenching ( Fig. 6t) The expression of Mrs2 MBS preserved the mitochondrial function ( Fig. 6m-6o) and delayed the Ca 2+ -mediated mPTP opening ( Fig. 6p-6t). These data show MCU activity and Ca 2+ -mediated mPTP opening to be fine-tuned by Mrs2-mediated mMg 2+ uptake.
To further confirm whether loss of Mrs2 potentiates MCU-mediated mCa 2+ uptake, overloads matrix Ca 2+ , and induces cell death, we performed an acute liver ischemic reperfusion (IR) injury by hanging weight method. Using this method, we saw the portal triad immediately occluded by hanging the weights over the poles, causing the blood supply to the left, median, and caudate lobes of the liver to be interrupted. Successful occlusion was confirmed by visual inspection of pale blanching in the ischemic lobes (i.e., a change in color from red to a pale color) ( Fig. 7a; left panel). In contrast, the change of color immediately disappeared when the hanging weights were removed from the poles, and the liver was reperfused (Fig. 7a, right panel).
Because elevated plasma ALT and AST are reliable markers for liver parenchymal cell membrane integrity and liver injury 47, 48 , we measured ALT and AST activity in plasma after 30 mins of ischemia followed by 1 h of reperfusion. The ALT and AST levels were increased in Mrs2 hep ( Fig. 7b and 7c). Our results also show increased levels of inflammatory cytokines in the plasma of Mrs2 hep (Fig. 7d and 7e) compared to sham and Mrs2 fl/fl IR injured mice. Histological analysis shows hepatocyte liver necrosis, mononucleated cell infiltration, enlarged central vein (cv), and increased congestion in Mrs2 hep compared to sham and Mrs2 fl/fl IR injured mice ( Fig. 7f). Taken together; these data show Mrs2-mediated mitochondrial Mg 2+ uptake is critical in maintaining MCU activity and protecting mitochondria from Ca 2+ overload mediated mPTP opening during IR injury.
Significantly Mg 2+ alters the electrophysiological properties of ion channels 55 (Fig. 1). Using biochemical assays, we show Mrs2 to be an integral membrane protein that localizes to the mitochondrial inner membrane with its N and C-termini within the matrix (Fig. 2). Because Ca 2+ activates the mitochondrial permeability transition pore, we asked whether loss of MCU regulation in Mrs2 KO cells results in early mPTP opening. We observed early mPTP opening in Mrs2 KO cells (Fig. 5), similar to that observed in HEK cells in which the matrix Ca 2+ -dependent MCU regulation is abolished 34 . Our results also show Mrs2 KO mice susceptible to liver ischemia-reperfusion (IR) injury (Fig. 7).
We anticipate the loss of Mrs2 to activate mPTP through two modes, direct and indirect regulation. Our results show that loss of Mrs2 positively regulates MCU activity, overloads matrix Ca 2+ , and triggers mPTP opening.
However, Mg 2+ is an inhibitor of mPTP opening 112 Fig. 5). We also anticipate a fall in the electrochemical proton gradient and membrane potential depolarization in Mrs2 KO hepatocytes during early stages of development. The compensatory pathways induced by the loss of Mrs2 during development possibly reversed the direction of ATP synthase rotation, resulting in ATP hydrolysis. We anticipate the reversal of ATP synthase to extrude protons into the intermembrane space and contribute to restoring the proton gradient and the membrane potential while simultaneously resulting in a net loss of ATP 115,116 . We used Oligomycin and BTB treatment to show the reverse activity of ATP synthase (Supplementary Fig. 5). Oligomycin blocks ATP hydrolysis and synthesis, whereas BTB is a specific ATP hydrolysis inhibitor. If the respiratory chain can keep up the membrane potential intact, then oligomycin treatment will hyperpolarize the mitochondria; however, if membrane potential is maintained by ATP hydrolysis, oligomycin/BTB will induce depolarization, a phenomenon termed as "oligomycin null-point" [117][118][119] . ATP hydrolysis can occur in extreme pathological conditions or normal or mildly compromised mitochondria, such as in humans with mitochondrial genetic disease and myopathies 120 . Our results demonstrate oligomycin null-point in Mrs2 KO mitochondria ( Supplementary Fig. 5

DECLARATION OF INTERESTS
The authors declare no competing interests.
Mrs2 fl/fl and Mrs2 hep mice were maintained in the Penn State College of Medicine animal facility following approval from the Institutional animal care and use committee. All mice were grouped according to sex, age, and genotype and used as required. Both male and female mice were used to isolate primary hepatocytes and mitochondria. For liver IR injury, male mice were used.
Acute liver ischemia and reperfusion injury: hanging weight system: We used portal triad occlusion and hanging weight system to induce acute liver ischemia-reperfusion injury as described previously 122 . The peritoneal cavity was exposed after a midline laparotomy and incision of the linea alba. The liver was kept wet and warm during surgery with a wet swab soaked with saline at 37°C. The stomach and duodenum were caudally displaced using a wet cotton tip swab to expose the portal triad and caudate lobe. The caudate lobe was gently separated from the left lobe, and the right lobe was then slightly shifted to clearly view the portal triad above the bifurcation of the right, median, and left lobes. Once visually identified, the needle, followed by a suture (7/0 nylon suture; Ethicon, Norderstedt, Germany), was placed under the portal triad, including the hepatic artery, hepatic vein, and common bile duct. The left end of the suture was then placed over the right pole, whereas the right end was placed over the left pole, and a weight of 1.5 grams was attached to each end.
The surgical wound was closed using continuous muscle walls and skin sutures. After surgery, mice were allowed to recover for 1 h of reperfusion under a heating lamp. Sham-operated mice served as the control and underwent anesthesia, laparotomy, and exposure to the portal triad without I/R. All animals survived the surgical procedure, and no complications were observed with portal triad occlusion using the hanging-weight system or in control mice.
Plasmids: Human Mrs2 full length (Mrs2 WT ) and its mutants (Mrs2p GMN and Mrs2p MBS ) were custom synthesized as gBlock gene fragments from IDT Inc. and cloned into pCMV6-Entry Cloning Vector (Origene) for expression in mammalian system. The plasmids/clones were confirmed by restriction enzyme analysis (REA) and sequencing before use.
Primary hepatocytes and cell lines: Primary adult mouse hepatocytes were isolated from 10-12 week-old male and female animals using the two-step collagenase perfusion technique with slight modifications 123   To test the orientation of Mrs2, mitochondria isolated from HEK293 cells stably expressing Mrs2 WT  NegShRNA and Mrs2KD was analyzed by qPCR using the PrimeTime predesigned probe for human Mrs2.
HPRT was used to normalize the mRNA levels.
Simultaneous measurement of mMg 2+ /mCa 2+  Plasma enzymatic and cytokine measurements: Plasma aspartate (AST) and alanine (ALT) aminotransferase activities were measured using a commercially available kit (Cayman) as per the manufacturer's instruction.
We used Mouse Cytokine Array Panel A (Proteome ProfilerTM Array; R&D systems) to determine the relative levels of 40 mouse cytokines.
Histological assessment of damage: The median and left liver lobes were harvested and fixed in 4% formalin.
Fixed tissues were subsequently sectioned, ten μm thick, collected on (+) charge slides, and stained with hematoxylin and eosin. Examination and scoring of each lobe were carried out by a pathologist who was blinded to the experimental group.
Quantification and statistical analysis: Data from multiple experiments (≥3) were quantified and expressed as Mean ± SE, and differences between groups were analyzed using the two-tailed paired Student's t-test or, when not normally distributed, a nonparametric Mann-Whitney U-test (Wilcoxon Rank-Sum Test) for two groups. The data were computed with GraphPad Prism version 9.0 or SigmaPlot 11.0 software. Differences in means among multiple datasets were analyzed using one-way ANOVA with Tukey correction performed. A P ≤ 0.05 was considered significant in all analyses.  The freshly prepared mitoplasts were exposed to Proteinase K for 10 min. The samples were probed using antibodies specific for Flag, MCU, CypD, Tom20, and actin. Enrichment of Mrs2 in the mitoplasts with no loss of Flag tag suggests Mrs2 as an integral membrane protein that localizes to the IMM with its N and C-termini facing the matrix. (i) Schematic representation of Mrs2 with its functional domains in the N-terminus, Transmembrane, and IMS loop.