Pivotal role of mitochondrial Na⁺–Ca²⁺ exchange in antigen receptor mediated Ca²⁺ signalling in DT40 and A20 B lymphocytes

Abstract

Non-technical summary

Elevation of cytoplasmic Ca²⁺ is one of the early responses of lymphocytes upon the antigen recognition by the surface receptor. We show here that the Ca²⁺ response is maintained by Ca²⁺ transfer from mitochondria to endoplasmic reticulum through mitochondrial Na⁺–Ca²⁺ exchange. The result helps us understanding how lymphocyte responses to antigen are organized at organelle level.

Abstract

Cytoplasmic Ca²⁺ concentration (µCa²⁺½_i) increases upon activation of antigen-receptor in lymphocytes. Mitochondria have been suggested to regulate the µCa²⁺½_i response, but the molecular mechanisms and the roles are poorly understood. To clarify them, we carried out a combination study of mathematical simulations and knockout or knockdown of NCLX, a gene candidate for the mitochondrial Na⁺–Ca²⁺ exchanger (NCX_mit), in B lymphocytes. A mathematical model of Ca²⁺ dynamics in B lymphocytes demonstrated that NCX_mit inhibition reduces basal Ca²⁺ content of endoplasmic reticulum (ER) and suppresses B-cell antigen receptor (BCR)-mediated µCa²⁺½_i rise. The predictions were validated in DT40 B lymphocytes of heterozygous NCLX knockout (NCLX^+/−). In NCLX^+/− cells, mitochondrial Ca²⁺ efflux via NCX_mit was strongly decelerated, suggesting NCLX is a gene responsible for NCX_mit in B lymphocytes. Consistent with the predictions, ER Ca²⁺ content declined and µCa²⁺½_i hardly rose upon BCR activation in NCLX^+/− cells. ER Ca²⁺ uptake was reduced to ∼58% of the wild-type (WT), while it was comparable to WT when mitochondrial respiration was disturbed. Essentially the same results were obtained by a pharmacological inhibition or knockdown of NCLX by siRNA in A20 B lymphocytes. Unexpectedly, ER Ca²⁺ leak was augmented and co-localization of mitochondria with ER was lower in NCLX^+/− and NCLX silenced cells. Taken together, we concluded that NCLX is a key Ca²⁺ provider to ER, and that NCLX-mediated Ca²⁺ recycling between mitochondria and ER is pivotal in B cell responses to antigen.

Introduction

Ca²⁺ is an important second messenger in the lymphocyte activation by antigen (Feske, 2007; Scharenberg et al. 2007; Vig & Kinet, 2009)). The molecular mechanisms underlying the antigen receptor mediated µCa²⁺½_i rise has been intensively studied. Upon the antigen binding to the surface receptor of lymphocytes, InsP₃ increases and facilitates Ca²⁺ release from the InsP₃ receptor (IP₃R) on the ER membrane. The Ca²⁺ release from ER results in the initial µCa²⁺½_i increase after the receptor activation. The subsequent Ca²⁺ depletion of ER causes translocation of the stromal interaction molecule 1 (STIM1) to the vicinity of plasmalemma, inducing a sustained and oscillatory µCa²⁺½_i increase by the activation of store-operated Ca²⁺ entry (SOCE) through Ca²⁺ release-activated Ca²⁺ channels encoded by ORAI1 (Feske, 2007; Vig & Kinet, 2009)). Mutations of STIM1 or ORAI1 cause hereditary immunodeficiency diseases in human (Feske, 2007; Vig & Kinet, 2009)). After the µCa²⁺½_i increase, lymphocytes undergo rapid proliferation and differentiation or are subjected to apoptosis, depending on the differentiation stage (Scharenberg et al. 2007)).

Mitochondria have been known as intracellular Ca²⁺ stores as well as ATP-producing factories in various cells (Celsi et al. 2009)). They have been suggested to regulate the µCa²⁺½_i response, but the molecular mechanisms and the roles in the antigen receptor mediated Ca²⁺ signalling are poorly understood. Ca²⁺ enters mitochondria through a Ca²⁺ selective channel, the Ca²⁺ uniporter (CaUni), according to the large negative membrane potential (Kirichok et al. 2004; Perocchi et al. 2010; Baughman et al. 2011)), and is extruded by the H⁺–Ca²⁺ exchanger (HCX_mit) and/or the Na⁺–Ca²⁺ exchanger (NCX_mit) (Castaldo et al. 2009; Celsi et al. 2009; Jiang et al. 2009; Palty et al. 2010)). Ca²⁺ extrusion by NCX_mit depends on the mitochondrial membrane potential, being facilitated by the negative potential (Kim & Matsuoka, 2008)), and HCX_mit also depends on this (Bernardi, 1999; Jiang et al. 2009)). The released Ca²⁺ from ER or sarcoplasmic reticulum (SR) enters mitochondria and the subsequent rise of mitochondrial Ca²⁺ activates several mitochondrial dehydrogenases (Jo et al. 2006; Csordas & Hajnoczky, 2009)). The mitochondrial Ca²⁺ sequestration and/or the mitochondrial metabolites have been reported to fine-tune the amplitude of SOCE (Hoth et al. 1997; Zablocki et al. 2005; Parekh, 2008; Schwindling et al. 2010)). Interestingly, mitochondria accumulate in the vicinity of immunological synapses in Jurkat T cells upon T cell receptor activation (Quintana et al. 2007)). The accumulation of mitochondria was suggested to support sustaining of the µCa²⁺½_i elevation by facilitating SOCE. However, except for the involvements in SOCE, roles of mitochondria in antigen receptor mediated Ca²⁺ signalling are not understood at all. Especially, it has not been clarified how Ca²⁺ flux from mitochondria contributes to the Ca²⁺ signalling.

NCLX/NCKX6 was first cloned as a gene for a subtype of K⁺-dependent or -independent Na⁺–Ca²⁺ exchanger that was suggested to be located at the ER or plasma membrane (Cai & Lytton, 2004; Palty et al. 2004)). Recently, Palty et al. (2010) reported that NCLX/NCKX6 is a gene candidate for NCX_mit. In this study, we found that NCLX/NCK6 is a gene responsible for NCX_mit also in B lymphocytes. We investigated the roles of NCX_mit in the B-cell antigen receptor (BCR) mediated Ca²⁺ signalling with a study combining mathematical modelling and knockout or knockdown of NCLX in DT40 and A20 B lymphocytes. It is demonstrated that NCLX (NCX_mit) encodes a key Ca²⁺ provider to ER and that NCLX mediated Ca²⁺ refilling of ER is essential for BCR-mediated Ca²⁺ signalling.

Methods

Computer simulation

A computer model of BCR-mediated Ca²⁺ dynamics was created using Delphi (Embarcadero Technologies, Inc., San Francisco, CA, USA) and its details are described in online Supplementary Material.

Cell culture

Two types of B lymphocyte were used in this study: chicken DT40 B lymphocytes expressing an IgM isotype (Baba et al. 1985)) and murine A20 B lymphocytes expressing an IgG isotype (Kim et al. 1979)) BCR at plasma membrane. DT40, A20, and NIH/3T3cells were maintained in RPMI 1640 (Invitrogen) medium supplemented with 10% fetal bovine serum (Nichirei Bioscience Inc., Tokyo, Japan), 2 mm sodium pyruvate (Lonza, Basel, Switzerland), and 50 μm 2-mercaptoethanol (Wako, Osaka, Japan) at 37°C in a humidified incubator with 5% CO₂ (Baba et al. 2006)).

Construction of NCLX-deficient DT40

Short and long arms were generated by PCR amplification using primer sets shown in Table S1 and were cloned into pBluescript KS(+). Then the puromycin cassette was cloned by PCR from pCMVpuro (Clontech) and subsequently inserted to generate the targeting construct. 1 × 10⁷ DT40 cells were transfected with 25 μg of NotI-linearlized targeting construct by an electroporation, using Gene Pulser apparatus (Bio-Rad) at 550 V, 25 μF. After 24 h culture, cells were selected by culturing in the presence of 0.5 μg ml⁻¹ puromycin for 7 days. Obtained puromycin-resistant colonies were expanded and checked for the homologous recombination by PCR. The homozygous knockout of NCLX was lethal in DT 40 cells so that we could not carry out a study on it.

Western blots

An affinity-purified antibody to NCLX was prepared at Scrum Inc., Tokyo, Japan by immunizing rabbits with a synthetic peptide (CPTDAEEQESSGTN) corresponding to the amino acid residues 268–280. 4 × 10⁶ cultured cells were lysed with M-PER Mammalian Protein Extraction Reagent (Thermo) and were resolved by SDS-PAGE using NuPAGE 4–12% Bis-Tris gel (Invitrogen). The gels were transferred to a PVDF membrane, and were then blocked for 30 min with Blocking One (Nakalai Tesque, Kyoto, Japan) followed by incubation with 1000× diluted primary antibody for 1 h at room temperature. After washing, membranes were incubated with 10,000× diluted goat anti-rabbit-IgG(H+L)-HRP (Thermo) for 30 min. The image was developed with Pierce Western Blotting Substrate (Thermo) and resolved by LAS-4000 mini (Fujifilm).

Cloning of NCLX

Chicken and mouse NCLXs were cloned from DT40 and A20 cells, respectively. Total RNA was isolated from DT40 and A20 cells with RNeasy Plus Mini kit (Qiagen GmbH, Hiden, Germany), and was then reverse-transcribed using Transcriptor First Strand cDNA Synthesis Kit (Roche Diagnostics). The open reading frame cDNA fragments of chicken (cNCLX) and mouse (mNCLX) NCLXs were generated by PCR amplification, and were then inserted into mammalian expression vector pENTR-SD/D-TOPO (Invitrogen) and pME18SFL3 (Toyobo, Osaka, Japan), respectively. The sequences of the isolated clones were identical to the published cNCLX (XM_415321.2) and mNCLX (NM_133221.2) sequences. A cDNA fragment encoding EGFP was first inserted into the 3′ end of cNCLX cDNA in the pENTR-SD/D-TOPO vector. The gateway conversion cassette Frame A (Invitrogen) was appropriately inserted between blunted NotI and BamHI sites of pQCXIP vector (Clontech) to prepare the destination vector pQCGIP. The expression vector cNCLX-EGFP/pQCGIP was generated by LR-reaction between cNCLX-EGFP/pENTR and pQCGIP using LR clonase (Invitrogen).

Human NCLX (hNCLX) cDNA, hNCLX/pME18SFL3, was purchased from Toyobo.

Transfection of NCLX siRNA and plasmids

siRNA for mNCLX was obtained from Ambion (Austin, TX, USA) (Silencer Select siRNA).

2 × 10⁶ A20 cells were transfected with either 2 μg of mNCLX/pME18SFL3, 2 μg of hNCLX/pME18SFL3, 10 pmol mNCLX siRNA, or 10 pmol control siRNA (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA) by an electroporation, using an Amaxa Cell Line Nucleofector Kit V (Lonza) with Nucleofector II Device (Lonza), according to the manufacturer's protocol.

NIH/3T3 cells were transiently transfected with cNCLX-EGFP/pQCGIP using FuGENE6 (Roche Diagnostics).

Cells were used after 24 h of transfection.

Localization of EGFP-labelled cNCLX, mitochondria and ER

Cell images were acquired at room temperature using a laser scanning confocal microscope (FV500 Olympus) equipped with a 150× oil objective lens (NA 1.45). Mitochondria and ER were stained with MitoTracker Green or Orange and ERTracker Red, respectively. EGFP and MitoTracker Green were excited at 488 nm and the fluorescence at 505–525 nm was collected. MitoTracker Orange and ERTracker Red were excited at 543 nm and the fluorescence at >560 nm was collected. All confocal images were processed by Autoquant X2.2 (Media Cybernetics Inc., Bethesda, MD, USA) to obtain 2D non-blind deconvolution images using theoretical point spread function. Co-localization of mitochondria with ER was quantified by the Pearson's co-localization coefficient (Adler & Parmryd, 2010)) and the Manders co-localization coefficient (Zinchuk et al. 2007)) using a plugin (JACoP) of ImageJ. The offset of each image was set automatically to avoid arbitrary judgment.

Measurement of µCa²⁺½_i in single cells

Cells were plated on poly-l-lysine-coated coverslips and centrifuged at 300 g for 2 min to facilitate attachment to the coverslips. The cells were incubated with 5 μm Fura-2 AM (Dojindo, Kumamoto, Japan) for 20 min at 37°C, and then washed with a physiological salt solution (PSS). The coverslip was transferred to a temperature-controlled recording chamber (Diamedical, Tokyo, Japan), ∼35°C, on a fluorescence microscope (Eclipse Ti, Nikon) equipped with a 20× objective lens (NA 0.75). Fluorescence images of the cells were recorded using an EM-CCD camera (ImagEM, Hamamatsu Photonics) and analysed with AQUACOSMOS software (Hamamatsu Photonics, Hamamatsu, Japan). Cells were alternately excited at 340 and 380 nm every 5 s, with emission signals being recorded at 500–530 nm. Calibration was carried out according to Williams & Fay (1990). The PSS contained (mm): 150 NaCl, 4 KCl, 2 CaCl₂, 1 MgCl₂, 5 Hepes, and 5.6 glucose (pH 7.4 with NaOH).

Measurement of cytoplasmic Na⁺ ()-dependent Ca²⁺ efflux

µCa²⁺½_mit was measured as describe previously (Kim & Matsuoka, 2008)). After incubation with 5 μm Rhod-2 AM (Invitrogen) in PSS for 30 min at 37°C, the cells were transferred to a perfusion chamber on a laser scanning confocal microscope (FV500 Olympus) equipped with a 60× objective lens (NA 0.9). The plasma membrane was permeabilized by perfusing the cells with a Ca²⁺-free cytosol-like medium (CLM_mit) containing 0.1 mg ml⁻¹ saponin for 5 min. Then mitochondria were loaded with Ca²⁺ by perfusing the cells with the CLM_mit containing free 10 μm Ca²⁺ and no Na⁺, and a -dependent µCa²⁺½_mit decrease (NCX_mit activity) was initiated by removal of and addition of 10 mm. Rhod-2 images were obtained with 543 nm excitation and 560–600 nm emission every 20 s at 35°C. The initial velocity was measured by fitting a linear function to the initial three points (40 s). Under the above experimental conditions, the most part of Ca²⁺ efflux from mitochondria depended on as shown later in Figs 2, 6 and 7. However, when mitochondria were loaded with lower concentration of Ca²⁺, a substantial part of the Ca²⁺ efflux was independent of and inhibited by ruthenium red (RR) (Fig. S1)), suggesting that the -independent Ca²⁺ eflux is mediated by HCX_mit (Letm1, Jiang et al. 2009)). In this study, we did not study the HCX_mit in detail and the higher Ca²⁺ loading was used to preferentially measure NCX_mit activity. It should be noted that NCLX knockout almost abolished NCX_mit activity, but did not affect the RR-sensitive -independent component or HCX_mit (Fig. S1)).

Figure 2. Properties of NCLX^+/− DT40 B lymphocytes.

A, expression of NCLX protein in WT and NCLX^+/− DT40 cells (Western blot analysis, upper panel). Total IgM was used as reference. Expression of surface antigen receptor (IgM) was not altered in NCLX^+/− DT40 cells (lower panel). B, NCX_mit activity. The activity was measured in WT with 10 mm (N= 7), WT without (N= 7), WT with 10 mm+ 20 μm CGP (N= 4) and NCLX^+/− with 10 mm (N= 4). n= 1–4 for each experiment. C, localization of EGFP-labelled NCLX (green) in NIH/3T3 fibroblasts. In the upper panel, mitochondria were stained with a mitochondria-specific dye (MitoTracker Orange, red), and a merged image is shown on the right. In the lower panel, ER was stained with an ER specific dye (ERTracker Red, magenta). Scale bars, 5 μm. P: plasmalemma, N: nucleus. D, BCR-mediated µCa²⁺½_i rise. Anti-IgM antibody at 3 μg ml⁻¹ was applied to stimulate BCR-mediated signalling. Seven representative data are shown for WT and NCLX^+/− DT40 cells. The percentage of cells whose µCa²⁺½_i increased more than three times is summarized at the right panel (WT: N= 7, NCLX^+/−: N= 10, n= 24–100 for each experiment). Data in B and the right panel of D are means ± SEM of independently recorded averaged responses.

Figure 6. Overexpression and silencing mNCLX in A20 cells.

A, NCX_mit activity. 10 mm-dependent Ca²⁺_mit decay was measured in A20 cells transfected with control siRNA (C_siRNA, N= 8), mNCLX siRNA (NCLX_siRNA, N= 4), and mNCLX cDNA (+mNCLX, N= 8). Red circles indicate data of C_siRNA without (C_siRNA:0 Na⁺, N= 4). n= 1–4 for each group. Initial velocity of µCa²⁺½_mit decay is shown at right for C_siRNA and +mNCLX (N= 8 for each group). *P < 0.001. B, BCR-mediated µCa²⁺½_i rises. Fura-2-loaded A20 cells were stimulated by the 5 μg ml⁻¹ anti-mouse IgG antibody. Averaged traces are presented for C_siRNA, +mNCLX, and NCLX_siRNA. N= 4 for each group and n= 50–100 for each experiment. Initial µCa²⁺½_i peak and time constant (τ) of µCa²⁺½_i decay are plotted at middle and right, respectively. Time constants were obtained by fitting an exponential function to data after the peak. *P < 0.001. C, ER Ca²⁺ content. The same protocol as Fig. 4A was used in A20 cells (C_siRNA, +mNCLX, NCLX_siRNA, and C_siRNA + 1 μm TG). Initial µCa²⁺½_i peak is shown at right (N= 4 for each group and n= 19–85 for each experiment). *P < 0.05. All the data are presented as means ± SEM of independent recordings.

Figure 7. Rescue of A20 cells transfected with mNCLX siRNA by co-expressing hNCLX.

A, NCX_mit activity. decay was measured in A20 cells transfected with control siRNA (C_siRNA), mNCLX siRNA (NCLX_siRNA), or hNCLX cDNA and mNCLX siRNA (+hNCLX & NCLX_siRNA). N= 4 for each group and n= 2–5 for each experiment. B, BCR-mediated µCa²⁺½_i rise. Data are for C_siRNA, NCLX_siRNA, and +hNCLX & NCLX_siRNA. N= 4 for each group and n= 40–63 for each experiment. All the data are means ± SEM of independent recordings.

As shown later in Fig. 4A, an analysis with a dye sensitive to mitochondrial membrane potential, JC-1, revealed the increase of cell population in NCLX^+/− cells whose mitochondria depolarized in NCLX^+/− cells. The partial depolarization of mitochondria might affect the measurement of -dependent Ca²⁺ efflux with Rhod-2 in NCLX^+/− DT40 cells. However, as demonstrated in Fig. S2, no significant difference was found in the extent of Rhod-2 and Ca²⁺ loads into mitochondria between WT and NCLX^+/− cells. Of course, cells in which Rhod-2 could not be loaded were eliminated in the experiments. These cells might reflect fully mitochondria-depolarized and/or apoptotic cells, in other word, almost non-viable cells. We do not think the elimination of these cells affected the characterization of NCLX^+/− cells.

Figure 4. Progress of apoptosis in NCLX^+/− DT40 cells.

A, mitochondrial membrane potential measured by JC-1 staining. The bar graph shows the percentage of green/red ratio more than 10 (N= 3). B, DNA content was measured by PI staining. The bar graph shows the percentage of subG₁ (N= 3). Data are means ± SEM of independent recordings. n= 2 × 10⁴ for each experiment. *P < 0.001.

The CLM_mit contained (mm): 118 KCl, 10 EGTA, 10 Hepes, 3 K₂ATP, 2 potassium pyruvate, 1 K₂HPO₄, 2 succinate, 0.1 K-ADP, 2 malate, and 2 potassium glutamate (pH 7.2 with KOH). Calculated free Mg²⁺ and Ca²⁺ concentrations (Patton et al. 2004)) were 1 mm and 10 μm, respectively. The CLM_mit containing 10 mm Na⁺ was prepared by replacing KCl with equimolar NaCl.

Measurement of ER Ca²⁺

ER Ca²⁺ (µCa²⁺½_er) was measured according to Tovey et al. (2006). The cells were incubated with 20 μm Mag Fluo-4 AM (Invitrogen, a Ca²⁺-sensitive dye with low Ca²⁺ binding affinity, K_d≍ 22 μm) in a Hepes-buffered saline for 60 min at 20°C, then washed with a cytosol-like medium (CLM_er). The cells were incubated with 30 μmβ-escin for 4 min to permeabilize the plasma membrane, washed twice, and were dispensed into a 96-well assay plate which was coated with poly l-lysine and centrifuged at 300 g for 2 min at 22°C. The Mag Fluo-4 fluorescence was measured by FlexStation (Molecular Devices) with excitation 490 nm and emission 525 nm every 1.5 s at room temperature. The Hepes-buffered saline contained (mm): 135 NaCl, 5.9 KCl, 11.6 Hepes, 1.5 CaCl₂, 11.5 glucose, and 1.2 MgCl₂ (pH 7.3 with NaOH). The CLMer contained (mm): 140 KCl, 20 NaCl, 1 EGTA and 20 Pipes (pH 7.0 with KOH). Mitochondria substrates (2 potassium pyruvate, 1 K₂HPO₄, 2 succinate, 0.1 K-ADP, 2 malate, and 2 potassium glutamate) were added to CLMer to keep mitochondria function intact. Carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone (FCCP), an uncoupler, was added and the mitochondria substrates were removed to disturb mitochondria function. Free Ca²⁺ concentrations were calculated using winmaxc software (Patton et al. 2004)) and adjusted to 100 nm. Ca²⁺ uptake of ER was activated by applying 0.1 mm MgATP and 100 nm Ca²⁺ to the permeabilized cells. The initial velocity of Ca²⁺ uptake was obtained by fitting a linear function to the initial four data (4.5 s). For evaluation of Ca²⁺ leak from ER, Ca²⁺ leak was initiated by inhibition of SERCA with 1 μm thapsigargin (TG) and 1.5 μm cyclopiazonic acid (CPA). The fluorescence decay speed during the initial 1 min (DT40) or 3 min (A20) was measured.

Analysis of surface IgM

WT and NCLX^+/− DT40 cells were treated with or without anti-chicken IgM antibody and stained with FITC-conjugated anti-mouse IgM anitibody. The fluorescence was analysed with a flow cytometry (LSR; BD Biosciences, Franklin Lakes, NJ, USA). Data acquisition and analysis were performed with the CellQuest (BD Biosciences) and FlowJo softwares (Tree Star, Inc., Ashland, OR, USA), respectively.

Analysis of DNA fragmentation

1 × 10⁶ cells were suspended in 200 μl of 0.2% Triton X-100/PBS, and were then incubated in the PBS containing 30 μg of RNaseA for 5 min at room temperature. The cells were analysed, after the addition of 800 μl PBS including 10 μg ml⁻¹ of propidium iodide (PI), with a flow cytometer (LSR).

Measurement of mitochondrial membrane potential

5 × 10⁶ cells were incubated in 0.5 ml JC-1 (a mitochondria membrane potential sensitive dye) solution at 37°C for 30 min, washed, and then resuspended in 0.5 ml cell culture medium (JC-1 Kit, Cell Technology Inc., Mountain View, CA, USA). The cells were analysed with a flow cytometer (LSR).

Real time PCR

Total RNA was isolated from three different batches each of WT and NCLX^+/− DT40 cells with RNeasy Plus Mini kit (Qiagen GmbH), and then was reverse-transcribed using Transcriptor First Strand cDNA Synthesis Kit (Roche). For the A20 cells transfected with control or mNCLX siRNAs, viable cells were sorted with a FACSAria III cell sorter (BD Biosciences) 24 h after the transfection. Real-time PCR was performed with the SYBR green dye technique on a Light Cycler 480 Instrument (Roche Diagnostics). The reaction conditions were 95°C for 10 min, followed by 45 cycles of 95°C for 10 s, 55°C for 20 s and 72°C for 10 s, using specific primers listed in Table S1.

Solutions and drugs

Ru360 (a CaUni inhibitor), thapsigargin (TG, a SERCA inhibitor), CPA (a SERCA inhibitor), and FCCP (a protonophore) were purchased from Sigma-Aldrich. MitoTracker Green FM, MitoTracker Orange CMTMRos, and ERTracker Red were from invitrogen. CGP-37157 (CGP, a NCX_mit inhibitor, Cox et al. 1993)) was from Tocris Tocris Bioscience, Minneapolis, USA; RR (an HCX_mit inhibitor, Jiang et al. 2009)) was from Wako; and ionomycin was from Calbiochem. The stock solutions for Ru360, CGP, TG, CPA, ionomycin and FCCP were prepared with DMSO, the final concentration of which was 0.01–0.5%.

Statistical analysis

All data are presented as means ± SEM of independent recordings. Number of the independent experiments and that of cells per recording are presented as N and n, respectively. In the measurements of cell responses under the microscope, the responses from n individual cells were averaged for each recording. Then the statistical evaluation was performed on these averaged responses from N independent recordings. An independent experiment was repeated more than three times. Statistical analyses were performed by one-way ANOVA multiple comparisons (SigmaPlot, Systat Software Inc., San Jose, CA, USA). Multiple and two-group comparisons were performed according to Student–Newman–Keul's method and paired or unpaired two-tailed Student's t test, respectively. P < 0.05 was considered significant.

Results

To study how the mitochondrial Ca²⁺ handling proteins participate in the BCR-mediated Ca²⁺ signalling, we first constructed a computer model of intracellular Ca²⁺ dynamics in lymphocytes, based on our Ca²⁺ measurements using DT40 B lymphocytes and previous cell models (Houart et al. 1999; Matsuoka et al. 2003; Kuzumoto et al. 2008)). A scheme illustrating the model is shown in Fig. 1A. The model cell has four compartments: extracellular space, cytoplasm, ER and mitochondria. It well reproduces BCR mediated InsP₃ increase, SOCE, and changes of µCa²⁺½_er and µCa²⁺½_i (Fig. 1B, black lines). See Supplemental Material for details of the model. The model analysis demonstrates that the suppression of NCX_mit, or reducing an amplitude factor of NCX_mit (A_NCXmit), inhibits the BCR-mediated initial µCa²⁺½_i rise which is caused by a massive Ca²⁺ release from ER, and supresses subsequent oscillatory µCa²⁺½_i rise (Fig. 1B, blue and red lines). The suppression of µCa²⁺½_i rise is due to the reduction of basal Ca²⁺ content of ER (right panel of Fig. 1B, blue and red lines). Although the model demonstrates only a representative pattern of Ca²⁺ changes in B lymphocytes and the details may be somewhat different from real cells, the simulation results prompted us to hypothesize that NCX_mit functions as a bridge for Ca²⁺ movement between mitochondria and ER that is important in lymphocyte responses to antigen. This hypothesis was examined in the following experiments.

Figure 1. Model simulation of BCR mediated Ca²⁺ dynamics.

A, a model scheme. B, simulation of BCR-mediated µCa²⁺½_i (left) and µCa²⁺½_er (right) responses. Results with a control set of parameters are shown in black. Reduction of NCX_mit amplitude factor (A_NCXmit) reduced the BCR-mediated µCa²⁺½_i rise and basal µCa²⁺½_er. A_NCXmit= 0.003 (blue) and 0 (red). BCR was stimulated 30 min after changing the parameter.

Recently, a gene called NCLX/NCKX6 was reported to be a candidate for NCX_mit (Palty et al. 2010)). Here we found that NCLX protein expresses and functions as NCX_mit also in DT40 B lymphocytes. The heterozygous knockout of NCLX (NCLX^+/−) significantly reduced the protein expression in DT40 B lymphocytes, without affecting total as well as surface IgM (BCR) contents (Fig. 2A)). In wild-type DT40 cells (WT) whose plasma membrane was permeabilized, the removal induced the decline in the presence of , shown as the clear fluorescence decrease of a indicator, Rhod-2 (Fig. 2B, black circles). The NCX_mit activity was almost completely abolished in NCLX^+/− cells (green circles), where the extent of decline was similar to that in the absence of (white circles) or in the presence of a NCX_mit inhibitor, 20 μm CGP (red circles) in WT cells. These data indicate that NCLX is associated with -dependent Ca²⁺ efflux from mitochondria in DT40 B lymphocytes. Cellular localization of NCLX was studied by expressing EGFP-labelled cNCLX in NIH/3T3 fibroblasts, which have larger cytosolic space than DT40 cells. NCLX nearly exclusively localized in mitochondria, which were stained with a mitochondria-specific dye (Fig. 2C, upper panel, Pearson's co-localization coefficient = 0.75 ± 0.03, N= 3, n= 1–2), and a small portion of the fluorescence was overlapped with ER (Fig. 2C, lower panel) probably due to close localization of ER and mitochondria, examined later in detail. Although expression of the EGFP-labelled cNCLX was less than 10%, mitochondrial localization of cNCLX is in good agreement with the result by Palty et al. (2010).

Consistent with the model simulation, NCX_mit (NCLX) was found to be pivotal in BCR mediated Ca²⁺ signalling (Fig. 2D)). The BCR stimulation in WT cells by anti-IgM antibody induced an initial large increase in µCa²⁺½_i due to the Ca²⁺ release from ER, followed by a sustained and oscillatory increase in µCa²⁺½_i. Although the expression of BCR was comparable to that in WT (Fig. 2A)), the µCa²⁺½_i increase was almost abolished in NCLX^+/− cells (middle panel of Fig. 2D)). Only ∼10% of NCLX^+/− cells responded, though slightly, while ∼95% did in WT (right panel of Fig. 2D)). Doubling the concentration of anti-IgM antibody or extracellular Ca²⁺ still did not help NCLX^+/− cells respond to BCR stimulation (data not shown). Basal µCa²⁺½_i without the BCR stimulation was slightly higher in NCLX^+/− cells (33.5 ± 2.9 nm in WT and 55.0 ± 4.8 nm in NCLX^+/−, N= 7 and 10, respectively. n= 24–100, P < 0.001).

The absence of BCR-mediated µCa²⁺½_i rise in NCLX^+/− cells was suggestive of a decrease in ER Ca²⁺ content, as predicted by the model simulation. ER Ca²⁺ content was directly assessed in Fig. 3A, where DT40 cells were superfused with a nominally Ca²⁺-free solution to minimize Ca²⁺ entry from the extracellular space. Ionomycin at 1 μm was applied to induce Ca²⁺ release from intracellular Ca²⁺ stores. The ionomycin-induced µCa²⁺½_i elevation reflects ER Ca²⁺ content because 1 μm TG, a selective blocker of SERCA, markedly reduced the µCa²⁺½_i rise. In NCLX^+/− cells, the average of peak µCa²⁺½_i was ∼30% of that in WT. ER Ca²⁺ content indeed significantly declined in NCLX^+/− cells. However, expressions of major Ca²⁺ handling proteins in ER and plasma membrane were unlikely to be affected in NCLX^+/− cells because the mRNA levels were not significantly altered (Fig. S3)). Therefore, the data suggest that NCX_mit or NCLX on mitochondria supports ER Ca²⁺ filling.

Figure 3. Contribution of NCX_mit to ER Ca²⁺ handling.

A, ER Ca²⁺ content was measured as 1 μm ionomycin-induced Ca²⁺ release in Fura-2-loaded DT40 cells superfused with a nominally Ca²⁺ free PSS. Note that the Ca²⁺ release was almost suppressed by 1 μm TG treatment. N= 3 for each group and n= 21–33 for each experiment. Data are means ± SEM of independently recorded averaged responses. B, C and D, ER Ca²⁺ uptake in WT and NCLX^+/− DT40 cells. ER Ca²⁺ uptake were activated by applying 0.1 mm MgATP and 100 nm Ca²⁺ in Mag Fluo-4-loaded and permeabilized cells under the conditions that mitochondrial respiration was intact (B) or disturbed (C). Control data are presented with data in the presence of 20 μm CGP or 1 μm TG. D, bar graphs summarizing the initial velocity of µCa²⁺½_er increase (left) and steady state µCa²⁺½_er (right). Data are means ± SEM of independent recordings. N= 29, 12, 27, 12, 10, 10, 10 and 10 from left to right bars in D, respectively. The same N is appreciable to B and C. n= 1.2 × 10⁶ for each experiment. *P < 0.001, #P= 0.046, †P= 0.019.

To quantitatively examine the contribution of NCX_mit to ER Ca²⁺ handling, Ca²⁺ uptake of ER was measured under the conditions that mitochondrial respiration was intact (Fig. 3B)) or disturbed (Fig. 3C)). Under the condition of disturbed mitochondria, it was expected that Ca²⁺ efflux via NCX_mit would be suppressed or NCX_mit reversed to the Ca²⁺ influx mode and that Ca²⁺ influx via CaUni would be suppressed, because of mitochondrial depolarization (Kirichok et al. 2004; Kim & Matsuoka, 2008)). In NCLX^+/− DT40 cells, the Ca²⁺ uptake via SERCA, measured as the initial velocity of µCa²⁺½_er rise, and the steady state ER Ca²⁺ level were reduced to 57.8% and 54.0% of WT, respectively, when mitochondrial respiration was intact (Fig. 3D)). NCX_mit inhibition by 20 μm CGP attenuated the Ca²⁺ uptake and the steady state ER Ca²⁺ level by 31.5% and 58.0% in WT, respectively, while the effect was smaller in NCLX^+/− cells. However, when mitochondria were disturbed (Fig. 3C)), no significant difference was found both in Ca²⁺ uptake and in steady state ER Ca²⁺ level between WT and NCLX^+/− DT40 cells. And the effects of CGP were significantly reduced (Fig. 3D)). The data demonstrated that the reduction of NCX_mit function by NCLX knockout or by the inhibitor CGP affects the ER Ca²⁺ uptake via SERCA. It was unlikely that the expression of SERCA per se was altered in NCLX^+/− cells because the initial velocity of ER Ca²⁺ uptake was similar to that of WT when mitochondrial respiration was disturbed. Namely, the data indicated that NCLX is a key Ca²⁺ provider to ER.

Elevation of µCa²⁺½_mit caused by the disturbance of Ca²⁺ efflux through NCX_mit and/or ER stress caused by the partial depletion of ER Ca²⁺ in NCLX^+/− cells may lead the cells to apoptosis (Breckenridge et al. 2003; Pinton et al. 2008; Celsi et al. 2009; Rodriguez et al. 2011)). Indeed in NCLX^+/− cells, an analysis with a dye sensitive to mitochondrial membrane potential, JC-1, revealed an increase in cell population with mitochondria depolarized (Fig. 4A)), which is one of the hallmarks of early apoptosis (Pinton et al. 2008; Celsi et al. 2009)). Apoptotic NCLX^+/− cells increased by twofold as demonstrated by an increase in DNA fragmentation (subG₁) (Fig. 4B)). However, the spontaneous progress of apoptosis cannot account for the unresponsiveness of µCa²⁺½_i to BCR stimulation in the NCLX^+/− cells, because the population of apoptotic cells was limited whereas almost none of the cells responded to BCR stimulation.

DT40 cells might adapt to the disruption of the NCLX gene by modifying gene expressions. However, key phenotypes of NCLX^+/− cells, namely the decline of ER Ca²⁺ content and the absence of BCR-mediated µCa²⁺½_i elevation, were not likely to be caused by secondary adaptations to NCLX knockout, because both short-term pharmacological blockade of NCX_mit by CGP and silencing NCLX by siRNA caused essentially the same effects as NCLX^+/− cells (Figs 5 and 6)). Treating the WT DT40 cells with CGP for 15 min reduced both the initial peak and steady state µCa²⁺½_i levels after the BCR stimulation in a dose-dependent manner (Fig. 5A)). CGP applied after the BCR stimulation also significantly reduced the µCa²⁺½_i elevation (Fig. 5B)). ER Ca²⁺ content remarkably declined in the CGP-treated DT40 cells (Fig. 5C)).

Figure 5. NCX_mit inhibition by CGP.

A, BCR-mediated µCa²⁺½_i rise. WT DT40 cells were stimulated by 3 μg ml⁻¹ anti-IgM antibody. Cells were treated with 0, 2, 20 or 50 μm CGP. N= 6 for each group and n= 10–15 for each experiment. The CGP treatment started 15 min before the recording. Bar graphs present initial µCa²⁺½_i peak (middle) and µCa²⁺½_i level 10 min after the BCR stimulation (right). *P < 0.01 vs. control. B, blockade of NCX_mit after BCR stimulation. 20 μm CGP was applied after the BCR stimulation by 3 μg ml⁻¹ anti-IgM antibody. µCa²⁺½_i traces of seven representative cells are shown (left). The bar graph presents µCa²⁺½_i before and after the addition of CGP. N= 3 and n= 10–20 for each experiment. *P < 0.001. C, ER Ca²⁺ content. Fura-2-loaded DT40 cells were superfused with a nominally Ca²⁺ free PSS, followed by that containing 1 μm ionomycin to induce ER Ca²⁺ release. CGP at 50 μm was applied concurrently with the superfusion of nominally Ca²⁺ free PSS. The bar graph presents initial µCa²⁺½_i peak. N= 6 for each group and n= 5–11 for each experiment. *P < 0.001. All the data are means ± SEM of independent recordings except for the left panel of B.

Mouse B lymphocytes, A20, were used for RNA interference experiments because of the low transfection efficacy in DT40 B lymphocytes (Fig. 6)). The transfection of NCLX siRNA in A20 lymphocytes caused 52.0 ± 9.4% (N= 3) reduction of NCLX mRNA expression. Knockdown of NCLX by siRNA (NCLX_siRNA) markedly slowed -dependent Ca²⁺ efflux from mitochondria to the level similar to that of control siRNA without (C_siRNA: 0Na⁺) (Fig. 6A)). In accord with NCLX^+/− DT 40 cells, the NCLX knockdown significantly reduced the BCR-mediated µCa²⁺½_i rise at initial peak (Fig. 6B)) and steady state (39 ± 3 nm in NCLX_siRNA vs. 58 ± 4 nm in C_siRNA, N= 4 for each group and n= 62–123 for each experiment, P= 0.009). ER Ca²⁺ content was also reduced (Fig. 6C)). In contrast, overexpression of mNCLX (+mNCLX) accelerated the rate of Ca²⁺ efflux from mitochondria twofold (right panel of Fig. 6A)). It also broadened the initial µCa²⁺½_i increase as indicated by a decay time constant (τ, right panel of Fig. 6B)) and increased the steady state µCa²⁺½_i (98 ± 5 nm in +mNCLX, N= 4 and n= 50–85 for each experiment, P < 0.001 vs. C_siRNA), although ER Ca²⁺ content was not altered (Fig. 6C)). The ER refilling role of NCLX might be almost saturated in the control cells as suggested by the model simulation later in Fig. 9, although the broad µCa²⁺½_i increase is suggestive of larger Ca²⁺ efflux from mitochondria.

Figure 9. Ca²⁺ recycling between mitochondria and ER.

A, a scheme of Ca²⁺ recycling between mitochondria and ER. B, simulation study on amplitudes of NCX_mit (A_NCXmit) and CaUni (A_CaUni). A_NCXmit and A_CaUni were systematically changed from 10⁻³ to 10¹. BCR was stimulated 30 min after changing the parameters. µCa²⁺½_er before the BCR stimulation (upper) and mean µCa²⁺½_i for 5 min after the 20 min BCR stimulation (lower) are presented.

The knockdown of NCLX by siRNA was successfully compensated by concurrent expression of hNCLX whose corresponding regions are not complementary to siRNA (+hNCLX and NCLX_siRNA) as shown in Fig. 7. The rate of decline (Fig. 7A)) and the BCR-mediated initial µCa²⁺½_i peak (Fig. 7B)) in +hNCLX and NCLX_siRNA cells were comparable to those of cells transfected with control siRNA (C_siRNA). These data with A20 B lymphocytes further demonstrated that NCLX is a gene responsible for NCX_mit and that this functions as a Ca²⁺ provider to ER.

The data above indicated that the NCLX on mitochondria is functionally pivotal for ER Ca²⁺ refilling. Close localization of mitochondria and ER, as suggested in several types of cell (Pinton et al. 2008; Csordas & Hajnoczky, 2009)), may enable this tight functional coupling. High Pearson's co-localization coefficient indicated that mitochondria and ER co-localize closely in both DT40 and A20 B lymphocytes (Fig. 8A)). However, the coefficient value was lower in NCLX-knockout (NCLX^+/−) and -knockdown cells (NCLX_siRNA) compared to the control cells. Further analysis with Manders's co-localization coefficient indicated that about 40% of ER co-localized with mitochondria and about 90% of mitochondria co-localized with ER in control cells of the two B lymphocytes (bottom right panel in Fig. 8A)). Closer inspection revealed that ER to where mitochondria were adjacent was significantly lower in NCLX-knockout and -knockdown cells. Co-localization of mitochondria to ER was lower by ∼30% in these cells, while that of ER to mitochondria was unaltered. Interestingly, treating WT DT40 cells or A20 cells with 20 μm CGP for 24 h did not alter the co-localization. Reduction of NCLX expression might disturb the structural coordination of the two organelles. Consistent with those two co-localization analyses, spontaneous Ca²⁺ leak from ER was larger by 31 and 37% in NCLX^+/− and NCLX_siRNA cells than WT and control siRNA-transfected A20 cells (C_siRNA), respectively, while the CGP treatment had no effect (Fig. 8B)).

Figure 8. Structural coordination of ER and mitochondria in DT40 and A20 cells.

A, co-localization of mitochondria and ER. WT and NCLX^+/− DT40 cells were stained with MitoTracker Orange (left panel) and ERTracker Red (middle panel). The merged images are shown in the right panel. Bottom graphs show Pearson's co-localization coefficient (PCC) and Manders's co-localization coefficient (MCC). WT: N= 59, WT+CGP: N= 34, NCLX^+/−: N= 37, C_siRNA: N= 31, and NCLX_siRNA: N= 41. n= 1–2 for each image. Scale bars, 1 μm. *P < 0.001. B, Ca²⁺ leak from ER was evaluated in Mag Fluo-4 loaded permeabilized cells. Left panel presents pooled data of ER Ca²⁺ leak induced by 1 μm TG and 1.5 μm CPA in WT, WT+20 μm CGP, and NCLX^+/− DT40 cells. The initial leak speed is summarized in the middle and right panels. WT: N= 64, WT+CGP: N= 29, NCLX^+/−: N= 33, C_siRNA: N= 24, and NCLX_siRNA: N= 24. n= 1.2 × 10⁶ for each group. Data are means ± SEM of independent recordings. *P < 0.001.

Discussion

NCX_mit was discovered in 1974 by Carafoli et al. and its significance in regulating is widely recognized in various cells (Castaldo et al. 2009; Celsi et al. 2009)). Recently, Palty et al. (2010) proposed that NCLX/NCKX6 is a gene encoding the mitochondrial Na⁺–Ca²⁺ exchanger in HEK-293, SHSY-5Y or CHO cells. Our experimental data not only support Palty's findings but also demonstrate for the first time the importance of NCLX in the BCR-mediated Ca²⁺ signalling of B lymphocytes. NCLX/NCKX6 was first cloned as a new subtype of K⁺-dependent or -independent Na⁺–Ca²⁺ exchanger and was suggested to be located at the ER or plasma membrane (Cai & Lytton, 2004; Palty et al. 2004)). Our experiments with EGFP-labelled NCLX in NIH/3T3 fibroblasts (Fig. 2C)) suggested that NCLX exclusively expresses in mitochondria. We have no clear explanation for the different intracellular localization of NCLX from the previous studies. However, significant inhibition of -dependent decline by NCLX-knockout or -knockdown indicates that NCLX is responsible for NCX_mit function.

B lymphocytes apparently have two Ca²⁺ extrusion pathways in mitochondria: Na⁺-dependent and -independent Ca²⁺ extrusion systems. The former is fast and mediated by NCX_mit or NCLX, and the latter is slow and probably mediated by HCX_mit or Letm1. High Ca²⁺ loading into mitochondria reduced the contribution of HCX_mit in Ca²⁺ extrusion because the activity saturates at low Ca²⁺ loads (Bernardi, 1999)). NCLX knockout almost abolished the NCX_mit activity while RR-sensitive HCX_mit activity was not affected (Fig. S1)). The dual Ca²⁺ extrusion system might allow B lymphocytes to survive the disturbance of NCX_mit.

Our major finding is that NCX_mit (NCLX) is the key Ca²⁺ provider from mitochondria to ER in B lymphocytes. Our results are in line with previous studies using the pharmacological reagent (CGP) in HeLa cells, endothelial cells and vascular smooth muscle cells (Arnaudeau et al. 2001; Malli et al. 2005; Poburko et al. 2009)), though the specificity of the drug is uncertain. For example, CGP is known to suppress the activity of plasmalemmal NCX and L-type Ca²⁺ channels (Omelchenko et al. 2003; Thu le et al. 2006)). In the above studies, CGP only moderately attenuated ER Ca²⁺ uptake during agonist stimulation. However, the targeted knockout or knockdown of NCLX in the present study caused a drastic decrease of ER Ca²⁺ content, directly indicating the pivotal role of NCX_mit (NCLX) in Ca²⁺ refilling of ER in DT40 and A20 B lymphocytes. Ca²⁺ transfer from ER to mitochondria was also significant in the B lymphocytes (see Fig. S4 demonstrating the simultaneous measurements of µCa²⁺½_i and µCa²⁺½_mit in DT40 B lymphocytes), consistent with previous reports using various types of cell (Csordas & Hajnoczky, 2009)). Therefore, a considerable amount of Ca²⁺ should recycle between mitochondria and ER, and the inter-organelle Ca²⁺ recycling must be pivotal for maintaining ER Ca²⁺ content and BCR-mediated Ca²⁺ signalling (Fig. 9A)). The contribution of NCX_mit to the ER refilling is possibly dependent on cell type, because the tissue distribution of NCLX has a unique pattern (high in pancreas, skeletal muscle and stomach, and low in kidney and lung (Palty et al. 2004)). Our computer simulation predicts that the contribution depends on the expression and/or activity balance between CaUni and NCX_mit (Fig. 9B)). Further analysis is necessary to clarify the contribution of NCLX in each cell type.

Close contact of SR/ER to mitochondria creates narrow interorganellar spaces. Our computer simulation suggests that large Ca²⁺ fluxes via NCX_mit and CaUni, which are comparable to that of SOCE (see supplemental Fig. 3)), are necessary to induce a significant Ca²⁺ refilling role of NCX_mit. As shown in Fig. 9B, reducing CaUni amplitude (A_CaUni) in our model offsets the effects of NCX_mit reduction and increases µCa²⁺½_i without affecting ER Ca²⁺ content at standard A_NCXmit. Since Ca²⁺ concentration in the interorganellar space is expected to rise higher than in bulk cytosol (Csordas et al. 2010)), such large Ca²⁺ flux would be possible in vivo. Interestingly, the voltage-dependent anion channel of outer mitochondrial membrane and IP₃R of ER were suggested to be located in the interorganellar regions and create a Ca²⁺ pathway from ER to mitochondria (Mendes et al. 2005; Szabadkai et al. 2006)). Our study provides new evidence for the reverse Ca²⁺ flow from mitochondria to ER through NCX_mit, interorganellar spaces and SERCA. The functional linkage between NCLX and SERCA suggests the two molecules face one another across the narrow interorganellar spaces.

Some of the experimental results would be explained by assuming that NCLX expresses in ER and regulates directly ER Ca²⁺ contents. However, the possibility is unlikely because the expression of EGFP-labelled NCLX did not completely overlap with ER (Fig. 2C)). Furthermore, ER Ca²⁺ uptake was affected neither by CGP (Fig. 4D)) nor by removal of cytoplasmic Na⁺ (Fig. S5)) when mitochondrial respiration was disturbed. Thus, we concluded that NCLX locates mainly on mitochondria and functions as a Ca²⁺ provider to ER.

One might assume that blocking CaUni, a mediator of Ca²⁺ influx into mitochondria, results in the reduction of Ca²⁺ efflux from mitochondria. However, CaUni does not much affect ER Ca²⁺ refilling. BCR-mediated µCa²⁺½_i elevation was little affected by Ru360, a selective blocker of CaUni, though the initial Ca²⁺ peak slightly increased (Fig. S6A)). Steady state ER Ca²⁺ level was also unaffected by Ru360 (Fig. S6B)). These data underline the importance of NCX_mit or NCLX, rather than CaUni, in determining and regulating the ER Ca²⁺ content of B lymphocytes.

Surprisingly, the structural coordination between ER and mitochondria was disrupted by the NCLX knockout or knockdown (Fig. 8)). Although the cause remains unresolved, NCLX may be associated with the tethering proteins connecting ER and mitochondria (Giorgi et al. 2008; Pinton et al. 2008; Csordas & Hajnoczky, 2009)) and the reduction of NCLX may weaken ER–mitochondria interactions. Alternatively, mitochondrial depolarization induced by the NCLX knockout or knockdown as shown in Fig. 4 might result in elimination of the impaired mitochondria by mitophagy (Youle & Narendra, 2011)). Since the ER–mitochondria communication is probably a mechanism common to every type of cell, our finding may be applicable to a wide range of cell functions, e.g. excitation, secretion and cell death. Mutations of one of the tethering proteins, mitofusin 2, loosen ER–mitochondria interactions, being suggested as pathogenesis of an inherited motor neuropathy (Charcot-Marie-Tooth type IIa) (de Brito & Scorrano, 2008)). NCX_mit function might be altered in cells that the tethering proteins are abnormally functioning. Recently, it was reported that one type of Parkinson's disease is related to cell death due to the impairment of NCX_mit function (Gandhi et al. 2009)). Suppression of NCX_mit may induce mitochondrial Ca²⁺ elevation and/or ER stress via partial depletion of ER Ca²⁺, resulting in the opening of the permeability transition pore and apoptosis (Breckenridge et al. 2003; Pinton et al. 2008; Celsi et al. 2009; Rodriguez et al. 2011)). Taken together, these findings suggest a wide range of roles of NCX_mit in regulating cell functions.

Glossary

Abbreviations

BCR

B-cell antigen receptor

Ca²⁺ in endoplasmic reticulum

mitochondrial Ca²⁺

CaUni

mitochondrial Ca²⁺ uniporter

CGP

CGP-37157

CLM_mit

cytosol-like medium for measurement of mitochondrial Ca²⁺

CLMer

cytosol-like medium for measurement of

cNCLX

chicken NCLX

CPA

cyclopiazonic acid

ER

endoplasmic reticulum

HCX_mit

mitochondrial H⁺–Ca²⁺ exchanger

hNCLX

human NCLX

IP₃R

InsP₃ receptor

mNCLX

mouse NCLX

cytoplasmic Na⁺

NCLX^+/−

heterozygous knockout of NCLX

NCX_mit

mitochondrial Na⁺–Ca²⁺ exchanger

PSS

physiological salt solution

PI

propidium iodide

RR

ruthenium red

SERCA

sarco/endoplasmic reticulum ATPase

SOCE

store-operated Ca²⁺ entry

SR

sarcoplasmic reticulum

STIM1

stromal interaction molecule 1

TG

thapsigargin

WT

wild-type

Author contributions

M.H. and S.M. contributed to the conception and design of the experiments. B.K., A.T., O.K., M.H. and S.M. analysed and interpreted the data. B.K., A.T. and S.M. drafted the article or revised it critically for important intellectual content. All authors discussed the results and approved the final version of the manuscript.