Mitochondrial Pyruvate and Fatty Acid Flux Modulate MICU1-Dependent Control of MCU Activity

Abstract

The tricarboxylic acid (TCA) cycle converts the end products of glycolysis and fatty acid β-oxidation into the reducing equivalents NADH and FADH₂. Although mitochondrial matrix uptake of Ca²⁺ enhances ATP production, it remains unclear whether deprivation of mitochondrial TCA substrates alters mitochondrial Ca²⁺ flux. We investigated the effect of TCA cycle substrates on MCU-mediated mitochondrial matrix uptake of Ca²⁺, mitochondrial bioenergetics and autophagic flux. Inhibition of glycolysis, mitochondrial pyruvate transport, or mitochondrial fatty acid transport triggered expression of the MCU gatekeeper MICU1 but not the MCU core subunit. Knockdown of mitochondrial pyruvate carrier (MPC) isoforms or expression of the dominant negative mutant MPC1^R97W resulted in increased MICU1 protein abundance and inhibition of MCU-mediated mitochondrial matrix uptake of Ca²⁺. We also found that genetic ablation of MPC1 in hepatocytes and mouse embryonic fibroblasts resulted in reduced resting matrix Ca²⁺, likely because of increased MICU1 expression but resulted in changes in mitochondrial morphology. TCA cycle substrate–dependent MICU1 expression was mediated by the transcription factor early growth response 1 (EGR1). Blocking mitochondrial pyruvate or fatty acid flux was linked to increased autophagy marker abundance. These studies reveal a mechanism that controls the MCU-mediated Ca²⁺ flux machinery and that depends on TCA cycle substrate availability. This mechanism generates a metabolic homeostatic circuit that protects cells from bioenergetic crisis and mitochondrial Ca²⁺ overload during periods of nutrient stress.

One-Sentence Summary:

Nutrient stress triggers transcriptional changes that limit Ca²⁺ influx into mitochondria through the MCU.

Editor’s Summary:

Rheostat for mitochondrial Ca²⁺ uptake during stress

Ca²⁺ entering mitochondria through the mitochondrial Ca²⁺ uniporter (MCU) normally promotes the generation of ATP through the TCA cycle. Nemani et al. found that nutrient stress initiated a transcriptional program that restricted Ca²⁺ influx into mitochondria through the MCU to prevent changes in mitochondrial dynamics and bioenergetics. Pharmacological or genetic disruptions in glycolysis or uptake of pyruvate or fatty acids induced the transcription factor EGR1 to increase the expression of the gene encoding MICU1, a gatekeeper for the MCU, resulting in reduced mitochondrial Ca²⁺ uptake. The authors propose that this response prevents mitochondria from becoming overloaded with Ca²⁺ and limits bioenergetics crises during nutrient stress.

INTRODUCTION

Metabolic homeostasis is essential not only at the cellular level for processes from cell division, growth, proliferation, differentiation to cell death, but also for tissue and whole organismal development. The three major catabolic pathways of energy production stem from glucose, lipids and proteins. Mutations in components in these pathways present with disrupted energy metabolism in patients leading to organ system damage and sudden death (1). Oxidation of glucose-derived metabolites by mitochondria provide positive feedback mechanisms that subsequently increase uptake of Ca²⁺ into the matrix, which in turn activate mitochondrial metabolism, further enhancing ATP synthesis (2-4). Although pyruvate is generated through multiple routes in the cytoplasmic milieu, it is transported into mitochondria through a hetero-oligomeric complex of mitochondrial pyruvate carrier MPC1 and MPC2 proteins (5, 6). The critical role of MPCs in pyruvate transport is supported by the finding that a MPC1^R97W mutation perturbs pyruvate transport from the cytosol to the mitochondrial matrix, resulting in fatally severe hyper-lactic acidemia and hyper-pyruvicemia (7). Similar symptoms of elevated lactate levels and metabolic acidosis occur in neonates with mutations in the E1a subunit of pyruvate dehydrogenase (PDH) (8-10). Although genetic deletion of MPC results in embryonic lethality in mouse models, several normal and cancer cell lines survive the metabolic stress of loss of MPC through unknown mechanisms (11, 12). Conversely, forced ectopic expression of MPC suppresses tumor growth (13). Additionally, mutations in carnitine palmitoyl transferase (CPT1) that prevent conversion of carnitine to acyl-carnitine for transport into the mitochondria also result in severe metabolic crisis (14-16). Other mutations in mitochondrial carriers that cause deficient or absent transport of substrates into the mitochondrial matrix result in hyperglycemia, brain damage, coma, cardiac arrest and death (17). Metabolic response to nutrients is tightly regulated through nutrient and energy sensing mechanisms. Different organelles respond differently to metabolic stress. The ER has several mechanisms to detect nutrient stress (18-22). On the other hand, mitochondria employ two broad mechanisms of response to nutrient stress: altered mitochondrial Ca²⁺ dynamics and transcriptional and posttranslational modifications (23-25). Although aberrant energy production is a common denominator, the molecular mechanisms by which mitochondria sense and regulate their activity based on the availability of metabolic substrates remains elusive.

Ca²⁺ flux across the inner mitochondrial membrane (IMM) regulates cellular bioenergetics, intracellular cytoplasmic Ca²⁺ signals, and various cell death pathways (26, 27). Ca²⁺ uptake into the mitochondria is driven by the organelle’s highly negative membrane potential and occurs through a highly selective channel, the mitochondrial calcium uniporter (MCU) (28, 29). Catabolism of carbohydrates, proteins and fatty acids results in the production of substrates that energize mitochondria, facilitating ATP generation. Mitochondrial matrix calcium activates multiple components of the TCA cycle, facilitating the production of reducing equivalents that feed into the electron transport chain (ETC) (27). Thus Ca²⁺ and metabolites are both necessary to maintain a pool of healthy mitochondria that generate ATP for cellular processes. How Ca²⁺-dependent mechanisms modulate mitochondria functions during nutrient deficiency, however, are poorly defined.

Here, we showed that mitochondria responded to nutrient stress by transcriptional upregulation of the gene encoding the gatekeeper of the mitochondrial calcium uniporter, MICU1, which was mediated by early growth response 1 (EGR1). The resultant increase in MICU1 expression led to inhibition of MCU-mediated mitochondrial matrix Ca²⁺ (_mCa²⁺) uptake under nutrient deficient conditions, possibly preventing _mCa²⁺ overload and subsequent cellular death. Furthermore, substrate deficiency impaired mitochondrial bioenergetics and promotes cell survival by up-regulating autophagy.

RESULTS:

Deprivation of mitochondrial substrates by inhibiting metabolic flux induces MCU complex MICU1 protein abundance

Cytosolic Ca²⁺ signals determine mitochondrial reducing equivalents and cell survival by altering MCU-mediated _mCa²⁺ uptake (30). To understand whether altering mitochondrial substrates modulated MCU complex abundance, we treated primary cardiomyocytes with inhibitors of key metabolic steps—glycolysis (2-deoxyglucose, 2-DG), mitochondrial fatty acid transport (etomoxir), or mitochondrial pyruvate transport (UK5099). Glycolysis is halted by 2-deoxyglucose (2-DG), etomoxir blocks fatty acid transport by inhibiting carnitine acyltransferase, and UK5099 interrupts mitochondrial pyruvate transport (31-33). Western blot analysis revealed that 2-DG, etomoxir, and UK5099, alone or in combination, significantly increased MICU1 protein abundance without affecting MCU expression (Fig. 1A). Pharmacologic blockade of pyruvate metabolism in the HepG2 hepatocellular carcinoma cell line also increased MICU1 protein abundance without altering MCU protein expression levels after (Fig. 1B). Next, we utilized an shRNA-based strategy to target the pyruvate transporter proteins MPC1 and MPC2 and induce a state of nutrient stress. RNAi-mediated silencing decreased the abundance of MPC1 and MPC2 by at least 80% in primary cardiomyocytes, HEK 293T and HepG2 cells (fig. S1, A to C), leading to increased MICU1 protein expression in all three cell types (Fig. 1, C to E). To further verify the link between mitochondrial pyruvate metabolism and MICU1 expression, we stably reconstituted the human dominant negative MPC mutant (MPC1^R97W) in HEK293T cells (fig. S1C). Cells stably expressing MPC1^R97W exhibited increased MICU1 protein levels similar to those in MPC siRNA-transfected cells (Fig. 1E). Densitometric analysis revealed a 2- to 15-fold increase in MICU1 protein abundance due to pharmacological or genetic interference in pyruvate metabolism (Fig. 1, A to E). Together, these data reveal that preventing pyruvate and fatty acid entry into mitochondria results in MICU1 up-regulation.

Figure 1: Deprivation of TCA cycle metabolites increases MICU1 protein abundance.

A and B) Primary rat cardiomyocytes (A) and HepG2 cells (B) were exposed to either 2-DG (1mM), etomoxir (10μM), UK5099 (100μM), or combination of 2-DG + etomoxir and UK5099 + etomoxir overnight and immunoblotted for MICU1 and MCU. TOM20 was used as loading control. Bottom panel presents relative protein abundance quantified using Image J. Mean ± SEM from three independent experiments. *P<0.05, **P<0.01, n.s., not significant.

C) MPC1 knockdown (KD) and MPC2 KD HepG2 cells were immunoblotted for MICU1 and MCU. TOM20 was used as loading control. Bottom panel presents relative density of protein abundance quantified using Image J. Mean ± SEM from three independent experiments. **P<0.01, n.s., not significant.

D) Primary rat cardiomyocytes were transiently silenced for MPC1 and MPC2 using RNAi and immunoblotted for MICU1 and MCU. TOM20 was used as loading control. Bottom panel presents relative protein abundance quantified using Image J. Mean ± SEM from three independent experiments. ***P<0.001, n.s., not significant.

E) HEK293T control, MPC2 KD and stably expressing MPC1^R97W cells were immunoblotted for MICU1 and MCU. Tom20 was used as loading control. Bottom panel presents relative protein abundance quantified using Image J. Mean ± SEM from three independent experiments. *P<0.05, n.s., not significant.

Loss of MPC1- and MPC2-dependent MICU1 induction alters basal MCU-mediated _mCa²⁺ uptake

Next, we assessed if the elevated levels of MICU1 protein expression observed after blockade of mitochondrial pyruvate entry played a role in modulating MCU-mediated _mCa²⁺ uptake. MICU1 binds to the MCU and inhibits MCU-mediated Ca²⁺ uptake in resting cells; loss of MICU1 results in basal matrix _mCa²⁺ accumulation (34-38). To understand how nutrient deficiency alters mitochondrial function, HepG2 cells stably expressing MPC1 and MPC2 shRNA were generated, which exhibited over 80% knockdown efficiency at the mRNA and protein levels (fig. S2, A to D). From the six stable cell lines we observed two scenarios. First, MPC1.1 and MPC1.2 shRNA HepG2 clones expressed reduced levels of MPC1 and MPC2 mRNA, possibly due to a lack of isoform specificity of the shRNA. Second, consistent with previous reports, we observed that loss of one isoform in MPC1.3, MPC2.1, MPC2.2, and MPC2.3 shRNA HepG2 clones resulted in possible destabilization of MPC protein complex and subsequent protein degradation of heteromeric subunit (fig. S2, C and D) (5, 6, 12). For our experiments we utilized MPC1.3 and MPC2.1 stable cell lines. Because MICU1 was elevated under conditions of nutrient deficiency, we asked if these cells exhibited differences in basal matrix Ca²⁺. Digitonin-permeabilized control and MPC shRNA-expressing cells were exposed to thapsigargin (Tg), a SERCA inhibitor that prevents Ca²⁺ uptake by the ER, and subsequently with the protonophore CCCP, which depolarizes the mitochondrial membrane potential (ΔΨ_m) and causes release of matrix Ca²⁺. CCCP-induced release of matrix Ca²⁺ was significantly reduced in cells with MPC knockdown (fig. S2, E and F). To validate that the observed decrease in matrix Ca²⁺ was due to pyruvate deficiency-dependent increases in MICU1 expression, we performed rescue experiments. Reconstitution of MPC1 and MPC2 in MPC shRNA-expressing cells normalized basal matrix Ca²⁺, as expected (fig. S2, G and H). Because mitochondria shape cytosolic Ca²⁺ dynamics and loss of the pyruvate carrier resulted in decreased matrix Ca²⁺, we next tested if cytosolic Ca²⁺ dynamics were also affected by loss of MPC expression. We observed that ATP-induced _cCa²⁺ elevation was unaffected by knockdown of either MPC1 or MPC2 (fig. S2I and inset), despite significant decreases in _mCa²⁺ uptake (fig. S2, J and K). This loss of ATP-induced _mCa²⁺ uptake was reversed by MPC reconstitution, demonstrating the specificity of the relationship between mitochondrial pyruvate availability and the organelle’s _cCa²⁺ buffering capacity (fig. S2L). To demonstrate that genetic ablation of MPC1 lowered _mCa²⁺ through regulation of MCU activity, MPC1 was deleted using an adenoviral mediated Cre-recombinase (Ad iCre) delivered to MPC1^fl/fl hepatocytes (Fig. 2A). Hepatocytes with Cre-mediated deletion of MPC1 exhibited reduced mitochondrial Ca²⁺ uptake (Fig. 2B and 2C). Additionally, basal mitochondrial Ca²⁺ was significantly reduced, an observation that suggests that the MCU “gatekeeping” activity was increased, possibly due to upregulation of MICU1 (Fig. 2D). In permeabilized primary hepatocytes, normal ΔΨ_m was maintained after Cre-mediated loss of MPC1 (Fig. 2E). However, MPC1 deletion significantly reduced _mCa²⁺ uptake after exposure to a bolus of Ca²⁺ (Fig. 2F). Moreover, MPC1^fl/fl + Ad iCre hepatocytes had significantly reduced MCU activity, MCU-mediated Ca²⁺ uptake rates, and basal matrix [Ca²⁺] following extramitochondrial Ca²⁺ addition (Fig. 2, G to K). Mitochondria exhibited a unique morphology upon acute deletion of MPC1 (Fig. 2L). Analyses of mitochondrial length, area and perimeter showed a significant remodeling of mitochondria (Fig. 2M-O). We then asked whether these functional and morphologic changes were conserved across cell types. MPC^fl/fl mouse embryonic fibroblasts (MEFs) infected with Ad iCre demonstrated _mCa²⁺ dynamics similar to those found in MPC1 deleted primary hepatocytes (Fig. S3A). Live cell confocal analysis of _mCa²⁺ dynamics revealed that peak and basal [Ca²⁺]_m were significantly lower in MPC^fl/fl + Ad iCre MEFs (Fig. S3 B-C). As in primary hepatocytes, MPC deletion did not alter the ΔΨ_m, in mouse embryonic fibroblasts (fig. S3, D and E). Although mitochondrial morphology was also altered in the MPC^fl/fl MEFs after Cre-mediated MPC1 deletion, the remodeling was distinctive compared to primary hepatocytes; the average length, perimeter, and area were significantly reduced (fig. S3 F-H). These studies suggest that the mechanism behind the alteration of MCU-mediated _mCa²⁺ uptake during conditions of nutrient deficiency is conserved across cell types, although the resulting mitochondrial remodeling that occurs under these conditions may be affected by differences in cellular-specific metabolite sensitivities.

Figure 2. MPC1-deficient hepatocytes exhibit reduced MCU-mediated mitochondrial Ca²⁺ uptake.

A) MPC1 transcript was analyzed by RT-qPCR. Inset depicts genotyping of MPC1 floxed mice. Bottom panel shows the MPC1 protein abundance four days after viral Ad iCre infection. Data from three independent experiments.

B) Mean traces for [Ca²⁺]_m (GCaMP6-mt) fluorescence measured in Mpc1^fl/fl or Mpc1^fl/fl + Ad iCre hepatocytes. n = 3 independent experiments.

C) Quantification of peak mito-GCaMP6 fluorescence after ionomycin (2.5 μM) stimulation from traces shown in (B).

D) Quantification of basal mito-GCaMP6 fluorescence from traces shown in (B).

E) Quantification of TMRE fluorescence in Mpc1^fl/fl and Mpc1^fl/fl + Ad iCre hepatocytes.

F) Hepatocytes from Mpc1^fl/fl mice were isolated, infected with Ad iCre for five days and Permeabilized with digitonin (40 μg/ml) in intracellular-like media containing thapsigargin (2 μM) and [Ca²⁺] indicator Fura-2FF (1 μM). After reaching steady-state ΔΨ_m, a bolus Ca²⁺ (50 μM) pulse was added at the indicated time-point before the mitochondrial uncoupler CCCP (10 μM) was added. Representative traces of extramitochondrial Ca²⁺ ([Ca²⁺]_out) clearance by permeabilized hepatocytes (MPC1^fl/fl (black) and MPC1^fl/fl + Ad iCre (red)). n = 3 independent experiments.

G) Analysis of mitochondrial Ca²⁺ uptake rate from traces shown in (F).

H) Bar chart represents mitochondrial Ca²⁺ uptake rate from traces shown in (F).

I) Quantification of total mitochondrial Ca²⁺ uptake following extramitochondrial Ca²⁺ (50 μM) pulse addition from traces shown in (F).

J) Total Ca²⁺_m uptake after single Ca²⁺ pulse was determined in MPC^fl/fl (black) and MPC1^fl/fl + Ad iCre (red) hepatocytes by recording Ca²⁺ released by CCCP from traces shown in (F).

K) Quantification of basal matrix Ca²⁺ from traces shown in (F).

L) Assessment of hepatocyte mitochondrial morphology. 36 hours after mito-GCaMP6 transduction, hepatocytes were stained with TMRE and images were acquired using confocal imaging system. Inset shows higher magnification images. Scale bar, 10 μm. n = 3-6 independent experiments.

M-O) Analyses of mitochondrial length, area, and perimeter from data shown in (L). 50 mitochondria per condition. Mean ± SEM. *P<0.05, **P<0.01, ***P<0.001, n.s., not significant.

To further verify the role of MPC in _mCa²⁺ uptake, we measured basal _mCa²⁺ levels in MPC1^R97W-expressing cells. A point mutation in MPC1 transforming arginine to tryptophan (R97W), renders defects in mitochondrial pyruvate oxidation. Patients with MPC1^R97W mutation exhibit lactic acidosis and hyper-pyruvicemia (7). As expected, MPC1^R97W expression lowered mitochondrial matrix Ca²⁺ content without affecting ΔΨ_m (Fig. S4, A to D). Collectively, these data suggest that inhibition of mitochondrial pyruvate transport alters MCU-mediated _mCa²⁺ uptake possibly through nutrient stress-induced expression of MICU1.

EGR1 controls the MICU1 promoter and induces MICU1 expression during mitochondrial stress

A possible mechanism for MICU1 upregulation during periods of nutritional deficiency is through alteration of MICU1 transcriptional regulation. To test this notion, a proximal MICU1 promoter (−566 to −297 bp) was inserted into a luciferase vector. Silencing of MPC2 expression in HEK 293T cells led to an ~600-fold increase in MICU1 transcription promoter luminescence signal compared to cells expressing a control luciferase vector (Fig. 3A). Bioinformatic analysis of the MICU1 promoter region revealed several putative binding sites for the transcription factor EGR. Early growth response factors (EGRs) are first among the transcription factors to respond to various environmental conditions—such as growth factors, hormones, neurotransmitters, mitogens and metabolites—that functionally control cell proliferation in multiple cell types (39-41). The EGR family consists of four members (EGR1, EGR2, EGR3 and EGR4) with considerable overlap in their consensus DNA binding sequences (Fig. 3B). To test whether EGR1 regulated MICU1 transcription, we utilized primary endothelial cells (ECs) freshly isolated from EGR1^−/− mice. EGR1^−/− ECs displayed a constitutive loss of MICU1 mRNA expression with concomitant increases in basal _mCa²⁺ levels (Fig. 3C). Because MICU1 is a gatekeeper of MCU at low _cCa²⁺ (below 1.0-2.0 μM), we then tested if EGR1 knockout mitochondria cleared 1 μM Ca²⁺ through MCU. EGR1^−/− ECs but not WT and EGR4^−/− ECs rapidly took up Ca²⁺ after application of a 1 μM Ca²⁺ bolus without changes in ΔΨ_m (Fig. 3, D-F). Consistent with our MICU1 luciferase reporter results, EGR1^−/− ECs demonstrated an increase in release of _mCa²⁺ upon addition of CCCP (Fig. 3, E and F), suggesting that EGR1 likely induces MICU1 transcription during states of mitochondrial nutrient deficiency. To further confirm that MICU1 expression was regulated by EGR1 rather than EGR4, primary hepatocytes isolated from EGR1^−/− and EGR4^−/− mice were treated with UK5099 and etomoxir. WT and EGR4^−/− hepatocytes, but not EGR1^−/− hepatocytes, exhibited a significant upregulation of MICU1 protein that was similar to other cell types (Fig. 3G). Densitometric analysis revealed about 2 to 3-fold increase in the protein levels of MICU1 in the WT and EGR4^−/− but not in EGR1^−/− hepatocytes (Fig. 3H). RT-qPCR analyses showed cells treated with UK5099 exhibited a significant elevation in MICU1 mRNA transcript (fig. S5). Although MICU1 protein abundance was elevated by blockade of either pyruvate or fatty acid transport into the mitochondria (Fig. 1A to E), it is plausible that Etomoxir-induced MICU1 protein elevation is partly through MICU protein complex stabilization (42). Nevertheless, future studies are warranted to understand the fatty acid dependent MCU activity. Finally, we found that basal _mCa²⁺, was decreased in WT and EGR1^−/− hepatocytes (Fig. 3I). These observations reveal that EGR1 is required for nutrient stress-induced MICU1 upregulation for control of _mCa²⁺ uptake.

Figure 3: EGR1 but not EGR4 regulates MICU1 transcription under substrate deficiency.

A) Control and MPC2 siRNA knockdown 293T cells were transfected with MICU1 promoter-luciferase construct and analyzed for luciferase activity. Data represent mean ± SEM (n=3 independent experiments), ***P<0.001.

B) Schematic of the consensus DNA binding region in EGR1, EGR2, EGR3 and EGR4.

C) Left panel presents relative MICU1 mRNA expression in ECs from WT and EGR1^−/− animals. Right panel represents quantification of normalized basal matrix _mCa²⁺ in ECs from WT and EGR1^−/− animals (n=3 animals for each genotype).

D) Top panel. Representative traces of permeabilized (40 μg/ml digitonin) ECs from WT, EGR1^−/− and EGR4^−/− mice loaded with 2 μM thapsigargin (Tg) and the ratiometric ΔΨ_m fluorophore JC-1 and pulsed with 1μM Ca²⁺ then with 10 μM uncoupler CCCP at 700 s. Bottom panel. Representative traces of permeabilized (40 μg/ml digitonin) ECs from WT, EGR1^−/− and EGR4^−/− mice loaded with 2 μM thapsigargin and ratiometric Ca²⁺ indicator Fura2-FF then pulsed with 1μM Ca²⁺ followed by 10 μM CCCP at 700 s. Data representative of 3 independent experiments.

E) Magnified inset showing matrix _mCa²⁺ released upon stimulation with CCCP.

F) Top panel. Bar graph presents quantification of amount of matrix _mCa²⁺ released upon stimulation with CCCP. Bottom panel. Bar graph presents quantification of rate of _mCa²⁺ uptake with 1μM Ca²⁺ pulse. Data from 3 independent experiments.

G) Hepatocytes isolated from WT, EGR1^−/− and EGR4^−/− mice were exposed to either 2-DG (1mM), etomoxir (10μM), UK5099 (100 μM), or combination of 2-DG + etomoxir and UK5099 + etomoxir overnight and immunoblotted for MICU1 and MCU. TOM20 was used as loading control.

H) Bar graph presents relative protein abundance quantified using Image J. Data represent mean ± SEM from 3 independent experiments. **P<0.01, ns= not significant.

I) Bar graph presents basal _mCa²⁺ as measured by Rho-2 fluorescence in hepatocytes isolated from WT and EGR1^−/− animals exposed to either 2-DG, etomoxir, UK5099, or combination of 2-DG + etomoxir and UK5099 + etomoxir. Data represent mean ± SEM from 3 independent experiments *P<0.05, **P<0.01, ns= not significant.

Impaired mitochondrial pyruvate flux promotes lactate accumulation and perturbs bioenergetics

To examine the link between pyruvate flux and MICU1-mediated MCU complex function, we measured mitochondrial bioenergetic parameters. Basal and maximal oxygen consumption were significantly decreased after knockdown of either MPC1 or MPC2 (Fig. 4, A to C). Additionally, MPC knockdown decreased total cellular ATP levels and increased levels of mROS (Fig. 4, D and E). Because pyruvate metabolic flux generates NADH, we next assessed mitochondrial NADH synthesizing capabilities. To measure the dynamic change of NADH/NAD⁺ ratio, control and MPC2 KD HepG2 cells were transiently transfected with a NADH sensor, peredox-mCherry. Stimulation with the protonophore FCCP rapidly dissipates ΔΨ_m, leading to maximization of electron transport chain complex activity. As expected, FCCP rapidly induced an increase in NADH in control cells, but MPC2 knockdown greatly attenuated FCCP-induced NADH production (Fig. 4F and inset), consistent with lower pyruvate and reduced basal _mCa²⁺. MPC1^R97W mutant cells showed reduced cellular ATP levels and increased mitochondrial ROS (Fig. 4, G and H). Additionally, intracellular and extracellular lactate levels were elevated in MPC1^R97W-expressing cells compared to control cells (Fig. 4I), suggesting conversion of the accumulated cytosolic pyruvate to lactate. Together, these data suggest that the blockade of pyruvate entry into mitochondria results in bioenergetic crisis, possibly due to increased MICU1 expression. To demonstrate that MICU1 plays a role in MPC1/2 knockdown-mediated bioenergetic crisis, hepatocytes were harvested from control MPC^fl/fl mice and those injected with Ad iCre virus. siRNA targeting MICU1 was then added to the hepatocytes to create four conditions: −Cre/−siRNA, −Cre/+siRNA, +Cre/−siRNA, +Cre/+siRNA (Fig. 5A), and OCR was measured (Fig. 5, B-E). As expected, MPC1 deletion resulted in marked reduction of maximal OCR (Figure 5, B and D). Intriguingly, MICU1 silencing alone showed significant increase in OCR (Fig. 5, B and D). Silencing of MICU1 when MPC1 was deleted caused a partial recovery of maximal OCR (Fig. 5B and D). Together, these results support the notion that increased MICU1 expression under conditions of mitochondrial pyruvate deprivation controls mitochondrial bioenergetics.

Figure 4: Loss of mitochondrial pyruvate flux perturbs cellular bioenergetics.

A) Measurement of oxygen consumption rate (OCR) in WT, MPC1KD and MPC2 KD HepG2 cells. After basal OCR was measured, oligomycin (1 μM) (A), FCCP (1 μM) (B), and rotenone (100 μM) + Antimycin A (100 μM) (C) were added as indicated.

B and C) Bar graph presents mean basal and maximal OCR. Mean ± SEM, (n=10 biological replicates for each group); ***P<0.001, **P<0.01

D) Bar graph presents cellular ATP levels in WT, MPC1 KD and MPC2 KD HepG2 cells. Mean ± SEM (n=8 biological replicates for each group); ***P<0.001, **P<0.01

E) Quantification of MitoSOX Red fluorescence in WT, MPC1 KD and MPC2 KD HepG2 cells. Mean ± SEM (n=8 biological replicates for each group); ***P<0.001, **P<0.01

F) Mean traces of cellular NADH/NAD+ fluorescence in WT, MPC1 KD and MPC2 KD HepG2 cells. Inset: Peak NADH/NAD+ fluorescence. Mean ± SEM (n=4 biological replicates for each group); *P<0.05

G) Bar graph presents cellular ATP levels in HepG2 cells stably expressing FLAG-tagged empty vector and MPC1^R97W plasmids. Mean ± SEM (n=8 biological replicates for each group); **P<0.01

H) Quantification of MitoSOX Red fluorescence in HepG2 cells stably expressing FLAG-tagged empty vector and MPC1^R97W plasmids. Mean ± SEM (n=3 biological replicates for each group); **P<0.01

I) Quantification of intracellular and extracellular lactate levels in HepG2 cells stably expressing FLAG-tagged empty vector and MPC1^R97W plasmids. Mean ± SEM (n=3 biological replicates for each group); ***P<0.001, *P<0.05

Figure 5. MICU1 alters mitochondrial oxygen consumption rate in MPC1 deleted hepatocytes.

A) Assessment of MICU1 protein abundance in MPC1^fl/fl hepatocytes following Ad iCre transduction and MICU1 specific siRNA. Protein abundance assessed in the following four groups: Mpc1^fl/fl -siRNA, Mpc1^fl/fl + Ad iCre -siRNA, Mpc1^fl/fl + MICU1siRNA, Mpc1^fl/fl + Ad iCre + MICU1siRNA (n=3 independent experiments).

B) Measurement of oxygen consumption rate (OCR) in Mpc1^fl/fl -siRNA, Mpc1^fl/fl + Ad iCre -siRNA, Mpc1^fl/fl + MICU1siRNA, Mpc1^fl/fl + Ad iCre + MICU1siRNA hepatocytes. After basal OCR was measured, oligomycin (1 μM) (A), FCCP (2 μM) (B), and rotenone (0.5 μM) + Antimycin A (0.5 μM) were added as indicated (n=3 independent experiments).

C-E) Bar graphs present mean basal OCR (C), maximal OCR (D), and proton leak (E). Mean ± SEM (n=7 biological replicates for each group); *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, n.s. not significant.

MPC-mediated pyruvate flux controls autophagy

Having observed that loss of mitochondrial pyruvate transport machinery detrimentally affected cellular energy metabolism, we next examined if alternative modes of survival were being utilized by the cell. Autophagy is a conserved catabolic process of degrading selective dysfunctional cellular components and organelles under conditions of nutrient deprivation and energy crisis (43, 44). When energy demand exceeds supply, dysfunctional mitochondria are targeted to the autophagosome and degraded by lysosomes. Thus, eliminating damaged mitochondria is a survival mechanism during nutrient stress. Autophagy is a dynamic process that includes formation of autophagosomes, fusion of autophagosomes and lysosomes, and finally, degradation in lysosomes (45). To investigate whether autophagy was increased during perturbed mitochondrial pyruvate transport, autophagic flux analysis was performed in cells lacking functional MPCs and transduced with the tandem LC3-GFP-RFP reporter that allows for detection of each step of the autophagic flux process. Green puncta indicates autophagosome formation and red puncta indicates autophagolysosome formation. In the acidic environment of lysosomes, GFP signal is quenched. Under fed conditions, whereas nominal levels of LC3 puncta were observed in control cells, an increase in LC3 signal (red and green puncta) was observed in cells expressing either MPC2 shRNA or the MPC1^R97W mutant (Fig. 6A), suggesting that even under growth-optimal conditions, loss of MPC led to increased autophagic flux corresponding with reduced cellular energy production. Starvation led to increased levels of puncta in all the cell types (Fig. 6A). Quantification of the number of LC3 autophagosomes revealed a significant increase in the number of LC3 puncta in cells expressing either MPC2 shRNA or MPC1^R97W under basal conditions (Fig. 6, A and B). Starvation resulted in similar numbers of autophagosomes in all cell types (Fig. 6, A and B). Complementary to the imaging data, lysates from cells expressing MPC2 shRNA or MPC1^R97W showed enhanced LC3I processing to LC3II and increased degradation of p62 under normal conditions compared to control cells (Fig. 6, C to F). We next asked whether pharmacologic blockade of fatty acid flux by etomoxir enhanced autophagic flux. Similar to MPC2 knockdown and MPC1^R97W mutant overexpression conditions, the number of LC3 puncta were increased following etomoxir treatment in WT hepatocytes (Fig. 7, A and B). Finally, we examined whether genetic deletion of MPC1 and/or KD of MICU1 altered the autophagic flux process. Hepatocytes were infected with Ad iCre virus, transfected with MICU1 siRNA (generating −Cre/−siRNA, −Cre/+siRNA, +Cre/−siRNA, and +Cre/+siRNA cells), and infected with adenovirus expressing tandem fluorescent-tagged LC3 (mRFP-GFP-LC3) (Fig. 7C). mRFP-LC3 puncta were imaged and quantified to analyze the degree of autophagic flux within cells. +Cre cells had more puncta than −Cre cells, indicating a link between MPC deletion and induction of autophagy (Fig. 7D). The highest number of puncta observed was within the +Cre/−siRNA condition, in which MPC was deleted and MICU1 was unaltered. In contrast, MPC deletion and MICU1 silencing resulted in a significant reduction of LC3 puncta, suggesting that MPC deletion promotes autophagic signaling partly through MICU1 induction and MCU activity. Collectively, these results establish a link between alterations in MCU-mediated _mCa²⁺ uptake and increases autophagic flux during periods of metabolite scarcity and nutrient stress.

Figure 6: Loss of mitochondrial pyruvate flux upregulates autophagic flux.

A) Representative confocal images of mRFP-GFP tandem fluorescent-tagged LC3 (tfLC3) in WT, MPC2 KD and MPC1^R97W HeLa cells before and after 24 hours starvation (n=4 cells for each group). Scale bar, 20 μm.

B) Quantification of autophagy performed as normalized LC3 puncta. Mean ± SEM (n=4 cells for each group); ***P < 0.001

C) Western blot analysis of cell lysates from WT, MPC2 KD and MPC1^R97W HeLa cells before and after starved for 16h. The cell lysates were probed for LC3, p-AMPK, AMPK and p62 antibodies, β-actin served as the loading control. (n=3 independent experiments)

D) Bar graph presents quantification of LC3-I/LC3-I+LC3-II. Mean ± SEM (n=3 independent experiments); **P<0.01, *P<0.05

E) Bar graph presents quantification of p62. Mean ± SEM (n=3 independent experiments); **P<0.01

F) Bar graph presents quantification of p-AMPK/AMPK. Mean ± SEM (n=3 independent experiments); **P<0.01

Figure 7: Pharmacologic blockade of fatty acid flux and genetic ablation of MPC1 elicit autophagic flux.

A) Representative confocal images of mRFP-GFP tandem fluorescent-tagged LC3 (tfLC3) in WT hepatocytes treated with etomoxir (10 or 100 μM) for 24 hours. Scale bar, 10 μm.

B) Quantification of autophagy performed as normalized LC3 puncta. Mean ± SEM (n=3 independent experiments) **P<0.01, ***P<0.001, n.s. not significant.

C) Representative confocal images of mRFP-GFP tandem fluorescent-tagged LC3 (tfLC3) in Mpc1^fl/fl -siRNA, Mpc1^fl/fl + Ad iCre -siRNA, Mpc1^fl/fl + MICU1 siRNA, Mpc1^fl/fl + Ad iCre + MICU1 siRNA hepatocytes. Scale bar, 10 μm.

D) Quantification of autophagy performed as normalized LC3 puncta. Mean ± SEM (n= 3 independent experiments). LC3 puncta was quantified from 10 cells per group. ****P<0.0001, n.s. not significant.

DISCUSSION

The ability to sense and respond to stress conditions is a key feature at cellular level that drives evolution and facilitates growth (46). This study revealed an unexpected role of a _mCa²⁺ sensing protein in mediating responses to mitochondrial substrate availability. MICU1 protein abundance was induced by both blocking pyruvate and fatty acid transport into the mitochondria. Our analysis of MCU-mediated _mCa²⁺ uptake under conditions of metabolic stress indicated that EGR1-mediated MICU1 upregulation prevented basal _mCa²⁺ accumulation and reduced mitochondrial bioenergetics. Conversely, deletion of MPC caused an increase in cellular lactate and in oxidative stress. Functionally, we found that blockade of MPC-dependent pyruvate flux enhanced autophagy. Our findings suggest that substrate-dependent control of _mCa²⁺ uptake could determine metabolic switching and cell survival under conditions of nutritional scarcity.

ER and the mitochondria are major metabolite sensing organelles (18, 20, 47-49). Proper protein folding in the ER is an energy consuming process. Conditions of nutrient stress limit energy production, thus halting the protein folding process. The ER senses an accumulation of unfolded proteins and triggers an unfolded protein response (UPR) that induces the transcription of a large set of nuclear-encoded genes to either degrade the unfolded proteins or to increase the capacity for protein folding. Thus, the ER senses nutrient deprivation and upregulates the UPR accordingly (49, 50). The mitochondrion, on the other hand, responds to nutrient stress by undergoing numerous posttranslational modifications of proteins and altering its dynamics (47). Metabolic changes in mitochondria cause broad changes in nuclear gene expression through retrograde mitochondria-to-nuclear signaling referred to as the mitochondrial stress response. Our data describes a mechanism by which MICU1 protein levels were elevated upon mitochondrial nutrient deprivation in various cell types, so that MICU1 may serve as a sensor of mitochondrial substrate levels (Fig. 1A to E, ). MICU1 is characterized as the gatekeeper of the mitochondrial calcium uniporter complex, limiting Ca²⁺ entry into the mitochondria under resting conditions (34). Consequently, increased MICU1 levels coincide with decreased basal resting matrix Ca²⁺, thereby protecting the cells from Ca²⁺ overload under conditions of metabolite scarcity (Fig. 2, B and F). To dissect the mechanisms of the mitochondrial metabolite-Ca²⁺ homeostasis axis, we utilized genetic and pharmacologic tools to silence and inhibit the mitochondrial pyruvate carrier proteins. The mitochondrial pyruvate transport machinery is comprised of MPC1 and MPC2 which form a hetero-oligomeric complex. Loss of one isoform results in loss of the other subunit, halting pyruvate flux into mitochondria. Genetic deletion of either MPC1 or MPC2 in mice results in embryonic lethality at E11 - E13.5, likely due to decreased mitochondrial oxidative phosphorylation activity (11, 51). Our results showed that cells in which MPC2 was silenced had a more severe phenotype compared to cells in which MPC1 was silenced, possibly due to the homo-oligomeric nature of MPC2. This original observation has been confirmed by a study using purified recombinant human MPC1 and MPC2 reconstituted in lipid vesicles. Whereas MPC2 homo-oligomers efficiently transport pyruvate at a rate similar to that of MPC1-MPC2 hetero-oligomers, MPC1 homo-oligomers do not transport pyruvate into the proteoliposomes (52). Disease-associated mutations in the mitochondrial pyruvate carrier cause severe neonatal metabolic acidosis with growth retardation leading to sudden death (7). Our findings reveal that loss of pyruvate entry into the mitochondrial matrix either due to silencing of MPC or due to mutations in MPC (MPC1^R97W) led to elevated MICU1 protein levels. Luciferase reporter assays revealed MICU1 was transcriptionally elevated by EGR1 but not EGR4 (Fig. 3A and G).

Mitochondria rapidly adapt to changes in energy supply and demand and alter their capacity for and efficiency of ATP generation. Substrates such as pyruvate, succinate, and malate feed into the TCA cycle producing reducing equivalents that in turn supply the electron transport chain (ETC) for ATP generation. Loss of pyruvate flux into mitochondria led to a reduction in basal and maximal OCR and total cellular ATP levels and an increase in mROS levels. Production of reducing equivalents (NADH) was severely perturbed in cells with MPC knockdown compared to the control, revealing severe cellular bioenergetic crisis (Fig. 4 F). Additionally, we observed an increase in autophagy in the nutrient deprived cells, as well as in cells treated with pharmacologic inhibitors of TCA precursor oxidation and genetic deletion of MPC1 (Fig 6, A to F and 7A to D). In conclusion, we describe a mechanism by which mitochondria respond to metabolite deficiency through induction of MICU1, which controls _mCa²⁺ uptake and bioenergetics. Thus, increased MICU1 levels protect the cells from Ca²⁺-overload and cell death in periods of nutritional deficiency.

MATERIALS AND METHODS

Cell line culture

HeLa, HEK 293T and HepG2 cells were cultured using low-glucose DMEM (GIBCO, Life Technologies) containing 10% (v/v) fetal bovine serum and 1% (v/v) penicillin/ streptomycin, with or without 2μg/ml puromycin. HepG2 and HEK293T cells (5x10⁵ /well) grown in 6 well plates were transduced with lentiviruses expressing shRNA for NegShRNA, MPC1 (Sigma # SHCLNV-NM_016098) and MPC2 (Sigma #SHCLNV-NM_015415). HEK293T cells were transfected with FLAG-tagged MPC1^R97W. Two days post transduction, the cells were selected with puromycin (2 μg/ml) or blasticidin (2 μg/ml) for 6–10 d and expanded.

Mouse models and primary cell culture

MPC1^fl/fl mice originally were obtained from Eric Taylor (University of Iowa), where they were generated on a C57Bl/6N background, backcrossed through C57Bl/6J. MPC1^fl/fl mice were maintained through homozygous breeding under standard housing conditions. EGR1^−/− and EGR4^−/− animals were previously generated (53-56). All animal experiments were approved by UT Health San Antonio’s and Temple University’s IACUC and followed AAALAC guidelines.

Primary mouse hepatocytes were isolated using a two-step collagenase perfusion technique with slight modifications (57, 58). In brief, mouse liver was sequentially perfused with perfusion medium-I (DPBS containing 10mM HEPES, 0.05% w/v KCl, 5mM glucose, 200 μM EDTA, pH 7.4) and perfusion medium-II (DPBS containing 30mM HEPES, 0.05% w/v KCL, 5mM Glucose, 1mM CaCl₂, pH 7.4) containing collagenase D (400μg/ml). Liver lobes were dissected and dissociated, and crude hepatocyte preparation was passed through a gauze mesh filter (100μM diameter). The preparation was centrifuged at 50×g for 2 min to pellet parenchymal hepatocytes. The hepatocytes were washed five times with perfusion medium-II and plated in culture dishes in Williams E medium (supplemented with 1% (v/v) antibiotic-antimycotic solution (Gibco), 1% (v/v) 200mM L-glutamine, 1% (v/v) non-essential amino acids, and 10% (v/v) heat-inactivated fetal bovine serum). After four to six hours, the culture medium was replaced. MPC^fl/fl hepatocytes were subsequently treated with Ad iCre (Vector BioLabs, Cat#1045N) using an MOI of 10-20. Mouse pulmonary microvascular endothelial cells (MPMVECs) were isolated from the lungs of wild-type, EGR1^−/− and EGR4^−/− mice as described elsewhere (59). Briefly, freshly harvested mouse lungs were treated with collagenase, and endothelial cells were isolated with magnetic beads coated with monoclonal antibody to CD144 (BD Biosciences Pharmingen). MPMVECs were propagated in DMEM complete media supplemented with 15% (v/v) FBS, nonessential amino acids, and penicillin/streptomycin. Cells that were not bound to the beads during MPMVEC isolation were plated separately and grown as murine lung fibroblasts (MLFs). MLFs were transduced with Ad-Cre-GFP virus (50 MOI). The loss of EGR1 and EGR4 were confirmed by western blotting.

Ventricular cardiomyocytes from one to two-day-old rat hearts (NRVMs) were prepared as previously described (60, 61). NRVMs were cultured in Ham’s F-10 supplemented with 5% fetal bovine serum (FBS) and penicillin/streptomycin (100 U/ml) at 37°C in a 95% air/5% CO2 humidified atmosphere for 4 days. NRVMs were transfected with indicated siRNAs and experiments were performed after 72 hr.

MPC^fl/fl mouse embryonic fibroblasts (MEFs) were isolated by harvesting embryos from pregnant MPC^fl/fl mice 13-14 days post-coitum (62, 63). Pregnant mice were anesthetized with isoflurane and killed by cervical dislocation. Embryos were immediately dissected and removed from the uterine cavity. Harvested embryos were rinsed in 70% (v/v) ethanol and placed in separate covered petri dishes containing 10 mL of 0.25% trypsin-EDTA. Heads of embryos were removed and used for genotypic confirmation. Embryonic tissues were then minced by chopping with a razor blade, and the resulting solution was incubated at 37°C for 15 minutes. The cell suspension was then transferred to a conical tube and fresh MEF growth media (DMEM supplemented with 10% (v/v) FBS, 1% (v/v) 200 mM L-glutamine, 1% antibiotic-antimycotic solution) was added to deactivate the trypsin. Cells were pelleted at 1000 rpm for 5 min and resuspended in fresh MEF growth medium and plated in T-75 flasks overnight. The following day, media was changed, and cells were allowed to expand before trypsinization, replating, and use in experiments.

Plasmids and antibodies

Peredox-mCherry (64) was purchased. Antibodies specific for MPC1 (Cell Signaling Technology, Cat # 14462S; 1:1000), MPC2 (Sigma, Cat# HPA056091; 1:500) LC3 (Sigma, Cat# L7543; 1:10000), p62 (Cell Signaling Technology, Cat# 5114; 1:1000), P-AMPK (Cell Signaling Technology; 1:1,000), AMPK (Cell Signaling Technology; 1:1000), β-actin (Santa Cruz Biotechnology Inc, Cat# sc-47778 HRP; 1:10000), FLAG (1:2000; Monoclonal anti-FLAG M2-peroxidase, Sigma), Tom 20 (Santa Cruz Biotechnology Inc, Cat# sc-11415). Antibodies specific for MICU1 were generated in-house (anti-rabbit MICU1) (30, 58, 65, 66).

RT-qPCR analysis

The knockdown expression was assessed by RT-qPCR. Briefly, total RNA was isolated from HepG2 Neg shRNA, MPC1 KD and MPC2 KD cells, and primary hepatocytes using the RNeasy Mini Kit (Qiagen, Valencia, CA) in accordance with the manufacturer’s instructions. Total RNA (1 μg) was reverse transcribed with iScript cDNA synthesis kit (BioRad). Real-time qPCR was performed with the gene specific Taqman qPCR primers as per the manufacturer’s instructions. The relative mRNA abundance was normalized to WT . (7300 Real Time PCR system RQ study software ; Applied Biosystems, Carlsbad, CA).

Matrix Ca²⁺ and ΔΨ_m measurement in permeabilized cells

Mitochondrial Ca²⁺ uptake and ΔΨ_m were determined by simultaneous monitoring of cytosolic Ca²⁺ with Fura-2FF (0.5 μM; Life Technologies) and ΔΨ_m with the lipophilic cationic dye 5,5′,6,6′-tetrachloro-3,3′-tetraethylbenzimidazolcarbocyanine (JC-1; 800 nM; Life Technologies). After trypsinization and neutralization, cells were pelleted at 200 × g for 5 min, resuspended in 20 mL of phosphate-buffered saline, and pelleted at 150 × g for 2 min. 8 × 10⁶ cells were resuspended in 1.5 mL ICM buffer containing 40 μg/ml digitonin to permeabilize the cells, protease inhibitors (EDTA-free complete tablets; Roche Applied Science, Indianapolis, IN), 2 μM thapsigargin to block the SERCA pump, and 2 mM succinate to energize mitochondria (67). Fura-2FF (1 μM) was used as a cytosolic Ca²⁺ indicator. After 20 s of data recording, JC-1 was added. At 1200 s or 400 s CCCP, 10 μM, was added. Fluorescence was measured using a dual-wavelength spectrofluorometer (PTI) with 490-nm excitation and 535-nm emission for monomeric JC-1 and 570/595 nm for the J-aggregate. ΔΨ_m was calculated as the ratio of J-aggregate and the monomer (68).

Cytosolic and mitochondrial Ca²⁺ dynamics

HepG2 negative shRNA (NegshRNA), MPC1 knockdown and MPC2 knockdown cells were grown on 25-mm glass coverslips for 48 h and loaded with 2 μM Rhod-2 AM (50 min) and 5 μM Fluo-4 AM (30 min) (Life Technologies) in extracellular medium (69, 70). After 1 min of baseline recording, agonist (ATP, 10 μM) was added, and confocal images were recorded every 3 s (510 Meta; Carl Zeiss, Thornwood, NY) at 488- and 561-nm excitation using a 63× oil objective to simultaneously monitor cytoplasmic and mitochondrial Ca²⁺ dynamics. Images were analyzed and quantified using ImageJ (National Institutes of Health, Bethesda, MD) and custom-made software (Spectralyzer, Elmsford, NY).

Confocal analysis of mitochondrial Ca²⁺ and ΔΨ_m

Freshly harvested primary murine hepatocytes were cultured on collagen coated glass coverslips and infected with an adenovirus construct of mito-GCaMP6 for 24 hours. The next day, cells were incubated with TMRE (50 nM) for 30 min at 37°C. Images were acquired using a Leica SP8 confocal microscope using a 100× oil objective at excitation wavelengths of 488 nm and 556 nm. After 30 sec baseline recording, ionomycin (2.5 μM) was added and images collected every 3 seconds. Images were quantified for mito-GCaMP6 and TMRE (50 nM) fluorescence using LASX software (Leica) and plotted as arbitrary units in GraphPad Prism 6 software. Mitochondrial length, perimeter, and surface area were quantified using ImageJ and plotted in GraphPad Prism.

Immunoblotting

HepG2 cells, HEK 293T cells, cardiomyocytes, and hepatocytes (from WT, MPC^fl/fl, EGR1^−/− and EGR4^−/− mice) were treated with or without the inhibitors 2DG (1mM), etomoxir (10μM), UK5099 (100μM) or all three overnight and lysed using RIPA buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.25% deoxycholic acid, 1 mM EDTA, 1% NP-40, protease inhibitor cocktail (Complete; Roche and 1 mM PMSF)). Equal amounts of protein were separated on 4-12% Bis-Tris polyacrylamide gel, transferred to a PVDF membrane, and probed with corresponding antibodies as specified.

Mitochondrial ROS measurement

Mitochondrial superoxide was measured by using the mitochondrial oxygen free radical indicator MitoSOX Red (molecular probes; Invitrogen) as described previously (70, 71). Briefly, cells grown on 0.2% gelatin coated glass coverslips were loaded with 5 μM MitoSOX Red for 30 min, and coverslips were mounted in an open perfusion microincubator (PDMI-2; Harvard Apparatus) at 37°C and imaged. Confocal (510 Meta; Carl Zeiss, Inc.) images were obtained at 561 nm excitation by using a 63× oil objective. Images were analyzed, and the mean MitoSOX Red fluorescence was quantified by using Image J software (NIH).

Oxygen consumption rate

Oxygen consumption rate (OCR) was measured in MPC^fl/fl primary hepatocytes, HepG2 control, MPC1 knockdown and MPC2 knockdown at 37°C in an XF96 extracellular flux analyzer (Seahorse Bioscience). Cells (ECs; 2×10⁴, cardiomyocytes; 4× 10³, hepatocytes; 6 x 10⁴) were sequentially exposed to oligomycin, FCCP, and rotenone plus antimycin A (72).

Total cellular ATP

Total ATP abundance was assessed using CellTiter-Glo luminescent assay, and luminescence was measured using the Infinite M1000 PRO plate reader (, Tecan).

Luciferase assay

HEK293 cells transfected with scrambled or MPC2 siRNA grown on 6 well plates were transfected with either MICU1 promoter reporter plasmid or control vector using Mirus LT-1 transfection reagent (Mirus) as per the manufacturer’s instructions. 48 hr post-transfection, cells were lysed (Promega) and the luciferase activity was detected using Bright-Glo Luciferase Assay System (Promega) as per the manufacturer’s instruction. Values were normalized with untreated controls and the relative luciferase activity (RLU) was plotted using GraphPad PRISM software version 5.0 (GraphPad).

NADH measurement

HepG2 NegshRNA and MPC2 KD cells (1×10⁵ cells) were transiently transfected with peredox m-Cherry. Cells were imaged 48 hr post transfection at 37°C. Confocal images were acquired at 405 nm and 561 nm excitation every 3s using Carl Zeiss 710 Meta NLO. Images were analyzed using ZEN 2010 software and plotted with SigmaPlot. (73).

Autophagy

HeLa NegshRNA, MPC2 knockdown, MPC1^R97W and primary hepatocyges were plated on 0.2% gelatin coated cover slips. Cells were infected with adenovirus expressing mRFP-GFP-LC3. After 36 hours of infection confocal images were obtained at 561 and 488 nm excitation using a 40× oil objective (510 Meta; Carl Zeiss, Inc.) and analyzed and quantified using ZEN 2010 software. Autophagic flux was determined by evaluating the punctate pattern of GFPLC3/ mRFP as described earlier (74). For primary hepatocyte autophagic flux analysis, images were acquired using a Leica SP8 confocal microscope using a 100× oil objective at excitations of 488 nm and 556 nm. Images were analyzed using LASX software (Leica) and plotted in GraphPad Prism 6 software.

Statistical analysis

Data were expressed as the mean ± SEM. Statistical significance was evaluated via Student’s unpaired t test, one-way and two-way ANOVA. P <0.05 was considered statistically significant. All experiments were conducted at least three times unless otherwise specified. Data were plotted either with Sigma Plot 11.0 software or GraphPad Prism version 6 software.

Supplementary Material

Supplemental Material

Figure S1: MPC1 and MPC2 abundance following RNAi-mediated knockdown.

Figure S2: Silencing MPC alters mitochondrial Ca²⁺ uptake in a MICU1-dependent manner.

Figure S3. MPC1-deficient MEFs exhibit reduced MCU-mediated mitochondrial Ca²⁺ uptake and altered mitochondrial morphology.

Figure S4: Pyruvate carrier mutations result in decreased matrix Ca²⁺ in a MICU1-dependent manner.

Figure S5: Pharmacologic blockade of fatty acid and pyruvate flux by etomoxir and UK5099 enhances MICU1 mRNA abundance.