Mitochondrial Mg2+ homeostasis decides cellular energy metabolism and vulnerability to stress

Cellular energy production processes are composed of many Mg2+ dependent enzymatic reactions. In fact, dysregulation of Mg2+ homeostasis is involved in various cellular malfunctions and diseases. Recently, mitochondria, energy-producing organelles, have been known as major intracellular Mg2+ stores. Several biological stimuli alter mitochondrial Mg2+ concentration by intracellular redistribution. However, in living cells, whether mitochondrial Mg2+ alteration affect cellular energy metabolism remains unclear. Mg2+ transporter of mitochondrial inner membrane MRS2 is an essential component of mitochondrial Mg2+ uptake system. Here, we comprehensively analyzed intracellular Mg2+ levels and energy metabolism in Mrs2 knockdown (KD) cells using fluorescence imaging and metabolome analysis. Dysregulation of mitochondrial Mg2+ homeostasis disrupted ATP production via shift of mitochondrial energy metabolism and morphology. Moreover, Mrs2 KD sensitized cellular tolerance against cellular stress. These results indicate regulation of mitochondrial Mg2+ via MRS2 critically decides cellular energy status and cell vulnerability via regulation of mitochondrial Mg2+ level in response to physiological stimuli.

However, it is not clear, in cells, how the changes of [Mg 2+ ] mito comprehensively affect the cellular energy metabolism in detail.
Although regulation of [Mg 2+ ] mito has not been elucidated in detail, mitochondrial Mg 2+ channel MRS2 is known to be a molecular machinery associated with Mg 2+ influx into mitochondria [31][32][33][34] . The rats with functional inactivation of mutated MRS2 have major mitochondrial deficits with a reduction in ATP, and increased numbers of mitochondria in oligodendrocytes 35 . Mg 2+ uptake into mitochondria via MRS2 is essential for the maintenance of respiratory chain and cell viability 5 .
In this study, we investigate how dysregulation of mitochondrial Mg 2+ homeostasis affects cellular energy maintenance and viability using single-cell fluorescence imaging and metabolomics analysis in Mrs2 knockdown (KD) cells. Mrs2 KD induces disruption of mitochondrial Mg 2+ homeostasis, which results in supression of mitochondrial ATP production and increased cellular stress susceptibility. These findings suggest that mitochondrial Mg 2+ plays important roles to maintain energy supply in cells, and its dysregulation causes cellular malfunction and multiple diseases.

Results
RNAi-mediated Mrs2 KD in HeLa cells. We investigated the importance of mitochondrial Mg 2+ homeostasis by an RNAi-mediated Mrs2 KD in HeLa cells. The best miRNA to knockdown MRS2 expression was selected by comparing loss of Mrs2 mRNA expression in HeLa cells with quantitative real time RT-PCR after 3 days of transfection of miR expression vector (Fig. S1). The miR expression vector #1 was optimal for Mrs2 KD in HeLa cells, and it was used for Mrs2 KD.
Effects of Mrs2 KD on Intracellular Mg 2+ Homeostasis. MRS2 is primary Mg 2+ uptake machinery in mitochondria 31,32,34 . To assess the effects of Mrs2 KD on [Mg 2+ ] mito homeostasis, [Mg 2+ ] mito was compared by using a ratiometric Mg 2+ indicator Mag-Fura-2. The cell membrane permeabilization protocol was used for the quantification of the [Mg 2+ ] mito 36 . Briefly, after loading Mag-Fura-2, cytosolic Mag-Fura-2 was washed out by cell membrane permeabilization with a detergent digitonin. While Mag-Fura-2 is normally used for the measurement of [Mg 2+ ] cyto (Fig. 1a-c), the co-localization of Mag-Fura-2 and a mitochondrial marker MitoFluor Red signals was observed after Mag-Fura-2 wash out from cytosol by digitonin treatment (Fig. 1d-f), indicating that this cell membrane permeabilization protocol by using Mag-Fura-2 enables mitochondrial Mg 2+ measurement.
[Mg 2+ ] mito in Mrs2 KD cells was lower than that in normal cells (Fig. 1g). Mrs2 KD cells were identified by expression of emGFP also coded in miRNA expression vector. [Mg 2+ ] cyto was also estimated with Mag-Fura-2 in normal usage ( Fig. 1h-j). The comparative analysis revealed the Mrs2 KD induced an increase in [Mg 2+ ] cyto (Fig. 1k) (Fig. 2j, left). During 6-7 min, the averaged amplitudes of FCCP-induced [Mg 2+ ] mito decrease in Mrs2 KD cells was larger than that in control cells (Fig. 2j, right). These results also indicate that, after the initial phase of FCCP-induced mitochondrial Mg 2+ release, mitochondria reuptake the released cytosolic Mg 2+ through MRS2, and mitochondria in Mrs2 KD cells failed it. Taken together, MRS2 plays a role as the mitochondrial Mg 2+ uptake systems in mammalian cells, and Mrs2 KD causes the deletion of mitochondrial Mg 2+ uptake in both steady state and Mg 2+ mobilization.

Suppression of TCA Cycle Induced by Disruption of Mitochondrial Mg 2+ Homeostasis.
To directly assess the effect of Mrs2 KD-induced homeostatic malfunction of mitochondrial Mg 2+ on global metabolism and, in particular, mitochondrial energy generation, the metabolomics of Mrs2 KD and control cells were investigated using capillary electrophoresis mass spectrometry (CE-MS) technique. CE-MS comprehensively quantify metabolites in biological samples 37 . We examined the differences in metabolite levels between control and Mrs2 KD cells. We quantitatively identified a total of 133 metabolites (control and Mrs2 KD samples, n = 6 respectively), and 24 metabolites were found to significantly differ between control and Mrs2 KD samples, which are overviewed in Fig. 3a (pathway information was obtained by reference to KEGG [http://www.genome. jp/kegg/pathway.html] and a previous report 38 ). Especially, most metabolites of TCA cycle such as malate, citrate, cis-aconitate and succinate, were reduced in Mrs2 KD cells (Fig. 3b). These results suggest that dysregulation of mitochondrial Mg 2+ homeostasis causes suppression of TCA cycle turnover in mitochondria. Despite Scientific RepoRts | 6:30027 | DOI: 10.1038/srep30027 suppression of TCA cycle activity, NADH/NAD + was not affected in Mrs2 KD cells (Fig. 3b). No effect on product of TCA cycle suggests that the direct impact of Mrs2 KD is downstream of TCA cycle in mitochondrial energy production.
Mrs2 KD Disrupt the Mitochondrial Membrane Potential. Using electron released from NADH and FADH 2 produced in TCA cycle, proton gradient is generated across mitochondrial inner membrane via electron-transport chain, which result in generation of mitochondrial membrane potential (Δ Ψ ). To assess the Δ Ψ , we quantified the ratio (red to green) of average mitochondrial fluorescence intensity of 5, 5′ , 6, 6′ -terachloro-1, 1′ , 3, 3′ -tetraethylbenzimidazolylcarbocyanine iodide (JC-1) per cells (Fig. 4a). In Mrs2 KD HeLa cells, the red-to-green ratio of JC-1 fluorescence, which indicates the mitochondrial inner membrane potential, was lower than control cells (Fig. 4b). This result suggests the mitochondrial Mg 2+ uptake through MRS2 is crucial for maintaining the Δ Ψ .
Mrs2 KD Disrupt the Energy Metabolism in Mitochondria. In mitochondria, ATP is made from ADP and phosphate by ATP synthase using an electrochemical gradient of protons across the mitochondrial inner membrane. Therefore, inside-negative Δ Ψ across the mitochondrial inner membrane is crucial for maintaining the physiological function of the oxidative phosphorylation to generate ATP. To access the intracellular energy status, we performed live imaging of cells expressing ATeam, which is genetically-encoded fluorescent resonance energy transfer (FRET)-based ATP indicator 39 . In cytosol and nucleus, the ATeam ratio, which indicates ATP level, were lower in Mrs2 KD cells than that in control cells (Fig. 4c,e). In contrast to the decrease in extra-mitochondrial ATP levels, in Mrs2 KD cells, ATP level increased in mitochondria (Fig. 4d) despite decrease of TCA cycle turnover (Fig. 3) and Δ Ψ m (Fig. 4a,b), suggesting that dysregulation of mitochondrial Mg 2+ causes the imbalances of ATP exports from mitochondria.

Dysregulation of Mitochondrial Mg 2+ Affect the Mitochondrial Morphology.
Mitochondria are highly dynamic organelle 12,40,41 , and their morphological changes regulate cellular metabolic processes and vice versa 12 . We assessed whether the malfunction of mitochondrial Mg 2+ regulatory system affects mitochondrial morphology. To obtain the exact mitochondrial morphology, mitochondrion selective probe insensitive to Δ Ψ , Mito Tracker Green FM, was loaded into both control and Mrs2 KD cells (Fig. 5b,c). To distinguish Mrs2 KD cells from control cells, Mrs2 miR expression vector coding tagBFP was used (Fig. 5a,c). Whereas the normal morphology of mitochondria in cells was tubular 10 (Fig. 5d), the increased accumulation of rounding mitochondria was observed in Mrs2 KD cells (Fig. 5e). In stress condition, the abnormal accumulation of large swollen mitochondria is reported in previous studies 42,43 . To quantitatively analyze the morphological difference of mitochondria between in control and in Mrs2 KD cells, a morphological feature was evaluated by computer-assisted image processing. The acquired images ( Fig. 5f) were processed to generate a mitochondria-specific binary image (Fig. 5g) allowing the quantification of mitochondrial shape (Fig. 5h) as previously described 44 . From quantitative morphological analysis of mitochondria, the aspect ratio (the ratio between the major and minor axes of the ellipse equivalent to the mitochondrial object) was calculated for each mitochondrion. The mitochondria in Mrs2 KD cells had lower aspect ratio than that in control cells (Fig. 5i), indicating that Mrs2 KD induced mitochondrial rounding.
Cell Vulnerability to Cellular Stress enhanced by Mitochondrial Mg 2+ dysregulation. We revealed the dysregulation of mitochondrial Mg 2+ results in the reduction of extra-mitochondrial ATP concentration. The negative effects of low ATP levels on cellular vulnerability are suggested 45 . To investigate the physiological effects of inaccessibility to Mg 2+ induced energy imbalance on cellular vulnerability to stress, the viabilities were compared under cellular stress conditions between control and Mrs2 KD cells by MTT assay. In Mrs2 KD cells, cell viabilities were lower than in control cells in the condition of 20 ng/mL TNF-α and 1 μ g/mL cycloheximide (CHX) (Fig. 6a) and also 1 mM H 2 O 2 conditions (Fig. 6b), respectively. These results indicate that dysregulation of mitochondrial Mg 2+ homeostasis causes increased susceptibility under cellular stress conditions. Metabolome analysis revealed that, under H 2 O 2 condition, ATP level in Mrs2 KD cells was lower than that in control cells (Fig. 6c), suggesting Mrs2 KD-induced lower ATP level cause cellular vulnerability against oxidative stress. Next, to investigate the effects of mitochondrial Mg 2+ dysregulation on cellular metabolome under stress condition, the overall impact of Mrs2 KD was compared using principal component analysis (PCA) for all detected 104 metabolites among 4 conditioned samples; control and Mrs2 KD cells with/without H 2 O 2 . PCA revealed that the clusters of control and Mrs2 KD cells under H 2 O 2 condition were more separated than in normal condition, indicating that Mrs2 KD-induced defection of mitochondrial ATP synthesis had larger impacts under stress conditions (Fig. 6d).

Discussion
We demonstrated that deficit of MRS2, which is a mitochondrial primary Mg 2+ regulatory system, causes disruption of mitochondrial energy metabolism and cellular sensitization against cellular stress. First, we confirmed that Mrs2 KD causes malfunction of mitochondrial Mg 2+ uptake (Fig. 2) and reduction of Mg 2+ stored in mitochondria (Fig. 1). Second, we revealed that Mrs2 KD induces decreases in substrates of TCA cycle (Fig. 3), Δ Ψ and extra-mitochondrial ATP levels (Fig. 4) in contrast increased mitochondrial ATP level (Fig. 4). In addition, we observed that mitochondria in Mrs2 KD cells have abnormal morphology (Fig. 5). Lastly, we showed that the effect of Mrs2 KD was noticeable under stress conditions, which sensitized cells to cellular stress (Fig. 6). These results indicate that mitochondrial Mg 2+ regulate the cellular energy status, and it changes mitochondrial morphology and affects cell vulnerability against biological stress.

Physiological Significance of mitochondrial Mg 2+ homeostasis. Although partial information
has been accumulated about Mg 2+ -dependent regulation of energy metabolism, the comprehensive effects of mitochondrial Mg 2+ on metabolic status in living cells have not been elucidated. As far as we know, this is the first work to demonstrate a series of relationships between Mg 2+ regulatory system and cellular energy metabolism. In Mrs2 KD cells, in contrast to decreased substrates of TCA cycle (Fig. 3), collapse of Δ Ψ m and decreased extra-mitochondrial ATP levels (Fig. 4), mitochondrial ATP level are increased (Fig. 4). These results indicate the mitochondrial dis-accessibility to Mg 2+ suppress ATP efflux from mitochondria, which is possibly mediated by ATP-Mg/P i carrier 6 . Mitochondrial ATP accumulation inhibits many enzymatic processes in TCA cycle 46 and electron transport chain activities 5 in a negative feedback manner. Suppression of TCA cycle and electron transport chain activities would result in reduced ATP production in mitochondria. Consequently, mitochondrial Mg 2+ regulates coupled reactions in mitochondrial energy metabolism, i.e. TCA cycle, electron transport chain and ADP/ATP translocation. In addition, morphological changes of mitochondria were also observed in Mrs2 KD cells (Fig. 5). Abnormal large and round mitochondria are also observed under pharmaceutically ATP synthesis-inhibited condition 47 . These are consistent with the idea that mitochondrial morphology is controlled by energy metabolism 48 . Abnormal mitochondrial morphology is associated with cancer 11 obesity, type 2 diabetes 12 , and neurodegenerative disorders [12][13][14] . It may be explained by the idea that metabolic impairment induces cellular vulnerability 45 . Actually, sensitization against cellular stress by mitochondrial Mg 2+ dysregulation was observed in Mrs2 KD cells (Fig. 6). In contrast, in a cellular model experiments of Parkinson's disease, increase in [Mg 2+ ] cyto , which probably links to [Mg 2+ ] mito increase, protects cells from neurodegeneration by maintaining cellular ATP concentration and suppressing ROS production 28 . In summary, mitochondrial Mg 2+ regulate the cellular metabolic process via shift of mitochondrial energy metabolism, and it changes mitochondrial morphology and affects the cell viability through changing stress susceptibility.
In normal cells, mitochondrial Mg 2+ would play a role as a regulator of metabolic state under physiological condition. A wide variety of hormonal regulations of intracellular Mg 2+ homeostasis has been reported 49 . In human, circadian rhythm for the serum Mg 2+ level with the peak around noon are reported 50 , which is corresponding to circadian Mg 2+ excretory rhythm with the peak at night 51

HeLa cells culture. HeLa cells were seeded in Dulbecco's modified Eagle's medium (DMEM) supplemented
with 10% heat-inactivated fetal bovine serum (FBS) and 50 U/mL penicillin and 50 mg/mL streptomycin (Life technologies, Carlsbad, CA, USA) and cultured at 37 °C in a humidified atmosphere containing 5% CO 2 . The cells were plated on glass-bottomed dishes (Iwaki, Tokyo, Japan), 100 mm cell culture dishes (Thermo Fisher Scientific, Waltham, MA, USA), or 48 well dishes (Thermo Fisher Scientific) for fluorescence imaging, metabolomics analysis or MTT assay, respectively. The medium was changed in every other day.
To identify the cells transfected with microRNA expression vector for Mrs2 knockdown, the sequence coding EmGFP or tagBFP were incorporated into the vectors. The miR RNAi vectors were transfected into cells 3 days prior to the experiments using Lipofectamine LTX (Life technologies) in the experiments except preparation for metabolome analysis.

Real time RT-PCR.
To determine Mrs2 mRNA levels in miR transfected cells, total RNA from HeLa cells were isolated and purified by using the RNeasy mini kit (QIAGEN, Tokyo, Japan). Total RNA was reverse-transcripted using SuperScript VILO (Life technologies), and the generated cDNA was used as template for quantitative real time PCR amplification with SYBR GreenER TM (Life technologies). The Mrs2 mRNA levels were normalized to GAPDH signals as an internal standard. Quantitative real time PCR were performed using the following primers: Mrs2 forward 5′-CAGTTGTCTGGAGAGGGTCAA-3′, Mrs2 reverse 5′ -AAGGATCAGTGGCTGCAAAA-3′, GAPDH forward 5′-CACCCACTCCTCCACCTTTG-3′ and GAPDH reverse 5′-CATGAGGTCCACCACCCTGT-3′.
Scientific RepoRts | 6:30027 | DOI: 10.1038/srep30027 Fluorescence imaging of the cytosolic or mitochondrial Mg 2+ levels. For the quantification of cytosolic Mg 2+ level in cells, HeLa cells were loaded with Mag-Fura-2-AM (Life technologies). For dye loading, HeLa cells were incubated in medium with 1 μ M and 0.02% F-127 for 45 min at 37 °C in a humidified atmosphere containing 5% CO 2 . The cells were gently washed twice with 1.0 mL of Hanks' balanced salt solution (HBSS) at pH7.4 (adjusted with NaOH) that consisted of (in mM) 137 NaCl, 5.4 KCl, 1.3 CaCl 2 , 0.5 MgCl 2 , 0.4 MgSO 4 , 0.3 Na 2 HPO 4 , 0.4 KH 2 PO 4 , 4.2 NaHCO 3 , 5.6 D-glucose, 5.0 HEPES. Then, further incubation was carried out for 15 min to allow for complete de-esterification of AM esters. For the measurement of the emitted fluorescence values at the excitation of 340 nm and 380 nm, a fluorescence microscope ECLIPSE TE300 (Nikon, Tokyo, Japan) equipped with a × 20 objective (S Fluor, Nikon) was used. A Xe lamp (150 W) with a monochromator unit was used for 340 nm and 380 nm excitation, and fluorescence was measured with a CCD camera (HiSCA, Hamamatsu Photonics, Shizuoka, Japan). The ratios of the fluorescence with excitation at 340 nm to that at 380 nm were calculated as an indicator of the cytosolic Mg 2+ levels. For the detection of Mrs2 KD cells, emGFP was illuminated with the excitation at 488 nm. Region of interest (ROI) was located on the respective cell areas. In each ROI, spatial averaged ratios of the emitted fluorescence with excitation at 340 nm to that at 380 nm was calculated as an indicator of the Mg 2+ levels.
For the quantification of mitochondrial Mg 2+ level in cells using ratiometric Mg 2+ indicator Mag-Fura-2, the cell permeabilization method using digitonin was performed as described previously 36 . Briefly, cells were stained with Mag-Fura-2 as described above. Cells loaded with Mag-Fura-2 were permeabilized with 20 μ g/mL digitonin in intracellular-like medium (ICM; 120 mM KCl, 10 mM NaCl, 1 mM KH 2 PO 4 , 20 mM HEPES-Tris at pH7.2 and 2 mM MgATP) buffer for 3 min, followed by washout of the released cytosolic Mag-Fura-2 with ICM. For the confirmation of mitochondrial localization and leakage from cytosol of Mag-Fura-2, fluorescence imaging was conducted with a confocal laser scanning microscope system (FluoViewFV1000; Olympus, Tokyo, Japan) mounted on an inverted microscope (IX81; Olympus) equipped with a × 60 oil objective. For the measurements Mag-Fura-2 signals, the cells were illuminated with the excitation wavelength at 405 nm from diode laser, and the signals from Mag-Fura-2 and Mito Fluor Red were separated using a 560 nm dichroic mirror. The fluorescence images were obtained by detecting the signals at 500-550 nm for Mag-Fura-2 and at 570-670 nm for Mito Fluor Red, respectively. For the measurement of the emitted fluorescence values at the excitation of 340 nm and 380 nm, a fluorescence microscope ECLIPSE TE300 equipped with a × 20 objective (S Fluor) was used. The procedure of Mag-Fura-2 detection was mentioned above. A ROI is located on the all spotted area with strong intensities in each cell (Fig. 1D).
Real time imaging of cytosolic Mg 2+ dynamics. For real time imaging of cytosolic Mg 2+ dynamics, highly selective Mg 2+ fluorescent dye KMG-104AM was used 24 . For loading KMG-104, cells were incubated with 5 μ M KMG-104AM and 0.02% F-127 in pH adjusted HBSS for 30 min at 37 °C. Then, the cells were washed twice with HBSS and incubated in HBSS for 15 min at 37 °C in a humidified atmosphere containing 5% CO 2 to allow for complete hydrolysis of the acetoxymethyl ester form. Fluorescence imaging was conducted with a confocal laser scanning microscope system, FV1000, equipped with a × 40 oil objective. For the measurements of KMG-104 signals, the cells loaded with KMG-104 were illuminated with the excitation wavelength at 488 nm from argon (Ar) laser. The fluorescence was obtained by detecting the signals at 500-600 nm. For the detection of tagBFP-labeled Mrs2 KD cells, the cells were illuminated with the excitation wavelength at 405 nm from diode laser, and its fluorescence was obtained by detecting the signals at 425-475 nm.
Real time imaging of mitochondrial Mg 2+ dynamics. For real time imaging of mitochondrial Mg 2+ dynamics, HeLa cells were stained with 20 μ M highly selective mitochondrial Mg 2+ fluorescent dye KMG-301AM in pH adjusted HBSS for 10 min on ice, so that hydrolysis of the acetoxymethyl ester by esterase present in the cytosol would be avoided 22 . Then, the cells were washed twice with HBSS and incubated in HBSS for 15 min at 37 °C in a humidified atmosphere containing 5% CO 2 to allow for complete hydrolysis of the acetoxymethyl ester form in mitochondria. The detailed property and usage of KMG-301AM was described in our previous report 22 . Fluorescence imaging was conducted with a confocal laser scanning microscope, FV1000, equipped with a × 40 oil objective. The cells loaded with KMG-301 were illuminated with the excitation wavelength at 559 nm from diode laser, and its fluorescence was obtained by detecting signals at 570-670 nm. For the detection of tagBFP-labeled Mrs2 KD cells, the cells were illuminated with the excitation wavelength at 405 nm from diode laser, and its fluorescence was obtained by detecting the signals at 425-475 nm.
Fluorescence imaging of the intracellular ATP levels. For the quantification of intracellular ATP levels in living cells, ATeam1.03 39 was transfected into HeLa cells using Lipofectamine LTX. The transfection was conducted the day before observation. Fluorescent imaging was conducted with a confocal laser scanning microscope system equipped with a × 20 air objective. For the measurements of ATeam signals, the cells were illuminated with the excitation wavelength at 440 nm from diode laser, and the signals from mseCFP and cp173-mVenus were separated using a 510 nm dichroic mirror and obtained at 480-510 nm for mseCFP and at 535-565 nm for cp173-Venus, respectively. The ratios of cp173-mVenus to mseCFP signals were calculated as an indicator of the cytosolic ATP levels.
Fluorescence imaging of the mitochondrial membrane potentials. For the quantification of mitochondrial membrane potential in living cells, HeLa cells were loaded with mitochondrial membrane potential sensitive dye 5, 5′ , 6, 6′ -terachloro-1, 1′ , 3, 3′ -tetraethylbenzimidazolylcarbocyanine iodide (JC-1; Life technologies). For dye loading, HeLa cells were incubated in pH adjusted HBSS with 10 μ g/mL JC-1 for 15 min at 37 °C in a humidified atmosphere containing 5% CO 2 . The cells were gently washed twice with 1.0 mL of HBSS. Fluorescent imaging was conducted with a confocal laser scanning microscope system, FV1000, equipped with a × 20 air objective. For the measurements of JC-1 signals, the cells were illuminated with the excitation wavelength at 488 nm from Ar laser, and the signals from JC-1 were separated using a 560 nm dichroic mirror and obtained by detecting the signals at 520-560 nm for green channel and at 575-620 nm for red channel, respectively. The ratio of signals from red channels to that from green channels was calculated as an indicator of the mitochondrial membrane potential. For the detection of tagBFP-labeled Mrs2 KD cells, the cells were illuminated with the excitation wavelength at 405 nm from diode laser, and its fluorescence images were obtained by detecting the signals at 425-475 nm.
Measurement of mitochondrial morphology. For assessment of mitochondrial morphology in living cells, mitochondria were stained with Mito Tracker Green FM and their morphology were quantified by handmade digital image processing software using MATLAB (MathWorks, Cambridge, UK). The algorithm for quantification of morphological feature is previously described 44 . Briefly, mitochondrial binary images were obtained from Mito Tracker Green FM-stained images, and aspect ratio of each mitochondrion was calculated as the ratio between the major and minor axes of the ellipse equivalent to the mitochondrial object.
Measurement of cell viability. For quantification of vulnerability against cellular stress, cell viabilities under the stress inducer H 2 O 2 -or TNFα /CHX-treated condition were measured using MTT assay. Control and Mrs2 KD cells were treated with H 2 O 2 (1 mM) or TNFα (20 ng/mL) plus cycloheximide (CHX; 1 μ g/mL) for 24 h. Then, the cells were incubated in the medium containing 0.5 mg/mL of MTT for 2 h at 37 °C in a humidified atmosphere containing 5% CO 2 . Then, the medium was removed, and 100 μ L of dimethylsulfoxide (DMSO, nacalai tesque, Kyoto, Japan) was added in each well to dissolve the precipitation. The absorbance at 570 nm under stress condition (ABS stress ) and normal condition (ABS control ) were measured using a microplate reader (Fluoroskan Ascent FL, Thermo Fisher Scientific). Cell viability was defined as the ABS stress /ABS control .
Image analysis and statistics. The fluorescence was calculated as the mean intensity over a ROI on the cell body of each cell using the software package, FluoView (Olympus), Aquacosmos (Hamamatsu Photonics) or handmade software by MATLAB.
Metabolome analysis. The cells were transfected with a plasmid for Mrs2 knockdown by electroporation using Neon (Life Technologies). The cells were plated on a 100 mm dish, and grown for 3 days at 37 °C in a humidified atmosphere containing 5% CO 2 . Culture medium was changed every day. Procedure for sample preparation of metabolome analysis was previously described 37 . Briefly, after washing cells twice with ice-cold 5% mannitol, metabolites were extracted by 1 mL ice-cold methanol containing internal standards (25 μ M each of methionine sulfone (MetSul; Wako, Osaka, Japan), 2-(N-morpholino)ethanesulfonic acid (MES; wako), D-Camphor-10-sulfonic acid (CSA; Wako). 400 μ L of collected extracts were transferred into another tube, mixed with 400 μ L chloroform and 200 μ L Milli-Q water, and centrifuged at 10,000 × g for 3 min at 4 °C. A 400 μ L aliquot of the aqueous layer was centrifugally filtered through a 5 kDa cutoff membrane (UltrafreeMC-PLHCC for Metabolome Analysis; Human Metabolome Technologies, Yamagata, Japan) to remove proteins from samples, followed by the centrifugal-concentration at 42 °C. CE-MS experiments were performed using Agilent CE Capillary Electrophoresis System.