Complex I and II are Required for Normal Mitochondrial Ca²⁺ Homeostasis

Abstract

Cytosolic calcium (_cCa²⁺) entry into mitochondria is facilitated by the mitochondrial membrane potential (ΔΨm), an electrochemical gradient generated by the electron transport chain (ETC). Is has been assumed that as long as mutations that affect the ETC do not affect the ΔΨm, the mitochondrial Ca²⁺ (_mCa²⁺) homeostasis remains normal. We show that knockdown of NDUFAF3 and SDHB reduce ETC activity altering _mCa²⁺ efflux and influx rates while ΔΨm remains intact. Shifting the equilibrium toward lower [Ca²⁺]_m accumulation renders cells resistant to death. Our findings reveal an unexpected relationship between complex I and II with the _mCa²⁺ homeostasis independent of ΔΨm.

1. Introduction

The electron transport chain (ETC) consists of four protein complexes embedded in the inner mitochondrial membrane that sequentially transport electrons delivered by NADH (Marreiros et al., 2016). The free energy from redox reactions at complex I, III and IV is used to translocate protons across the inner mitochondrial membrane, forming an electrochemical gradient called the mitochondrial membrane potential (ΔΨ_m), which is used by the F0-F1-ATP synthetase to generate ATP (Saraste, 1999). In addition, ΔΨ_m allows the import and export of metabolites, proteins and ions to and from the mitochondrial matrix (Poburko and Demaurex, 2012; Zorova et al., 2017). In fact, the ΔΨ_m is considered to be the main driving force for the mitochondrial Ca²⁺ (_mCa²⁺) uptake (Nicholls and Budd, 2000) and its maintenance is essential to prevent the onset of cell death (Duchen, 2000).

The influx of Ca²⁺ into the mitochondria promotes the activity of at least seven different mitochondrial proteins including three rate-limiting dehydrogenases of the TCA cycle (Glancy and Balaban, 2012). Thus, Ca²⁺ plays a strategic task in coordinating cellular workload and generation of ATP (Denton, 2009).

Dysfunctions either in complex I, the larger enzyme of the ETC composed by 44 subunits; 14 highly conserved core subunits that harbor bioenergetic functions, and 30 accessory subunits (Hirst, 2010) or complex II (succinate-ubiquinone oxidoreductase), the only member of the ETC that also participates in the TCA cycle and is comprised of four subunits (SDHA, SDHB, SDHC, and SDHD) (Bezawork-Geleta et al., 2017) have been associated with several diseases that show severe alterations in Ca²⁺ homeostasis such as cancer (Bustos et al., 2017; Kluckova et al., 2013; Urra et al., 2017) and neurodegenerative conditions such as Parkinson’s (Cieri et al., 2017; Gatt et al., 2016; Haelterman et al., 2014; Keeney et al., 2006), Alzheimer’s disease (Kim et al., 2001; Muller et al., 2018; Zhang et al., 2015) and Leigh syndrome (Quinlan et al., 2012; Wasniewska et al., 2001). However, the relation between complex I or complex II dysfunction and Ca²⁺ dyshomeostasis has been poorly explored and only associated with loss of ΔΨm (Mbaya et al., 2010; Visch et al., 2004), which occurs in specific scenarios, while in most the ΔΨm remains constant thanks to the reversal of ATP-synthase activity (Chinopoulos, 2011).

Here, we show that knockdown of both complex I assembly factor NDUFAF3 and complex II catalytic subunit SDHA reduce mitochondrial Ca²⁺ fluxes and mitochondrial respiration without changing the ΔΨ_m. Moreover, we demonstrate that the reduction in _mCa²⁺ flux protects these cells from cell death induced by mitochondrial Ca²⁺ overload. Our findings reveal that complex I and complex II play a role in Ca²⁺ homeostasis beyond their role in the maintenance of the ΔΨ_m.

2. Materials and Methods

2.1. Materials

Antibodies: From Santa Cruz Biotechnology®: NDUFAF3 (sc-99317), β-actin (sc-47778). Abcam®: SDHA (ab14715), SDHB (ab14714), ATP5A (ab14748) NDUFB8, (ab110242). Reagents and inhibitors were purchased from Sigma-Aldrich unless specifically stated.

2.2. Cell culture and transfection.

MCF7 cell line, derived from human breast cancer, was purchased from ATCC (ATCC® Number: HTB-22™) and cultured in DMEM HG supplemented with 10% FBS. For NDUFAF3 knockdown (NDUFAF3-KD), MCF7 cells where transduced with lentiviral particles (MISSION® Lentiviral Transduction Particles – Sigma-Aldrich SHCLNV-NM_199069) and 24h after transduction selected with puromycin. For SDHA knockdown (SDHA-KD) cells, shRNA construct in retroviral untagged vector was used (Origene™ TR315539). Transfection was done with FuGENE® in reduced serum media for 24 hours and selected with puromycin.

2.3. Western Blotting.

Cell extracts were prepared with ice-cold 1X RIPA buffer (EMD Millipore®) supplemented with protease and phosphatase inhibitors (Complete EDTA-free, PhosSTOP, Roche). Equals amount of protein extracts were separated in NuPAGE 4–12% Bis-Tris protein gels (ThermoFisher Scientific) and transferred to PDVF membranes.

2.4. Immunopurification of complex I and II.

MCF7 Scramble, MCF7 NDUFAF3-KD and MCF7 SDHA-KD cells were lysed in lysis buffer [150 mM Tris-HCl, 150 mM NaCl (Merck), 1 mM EDTA (Chemix, 5 mM NaF (Sigma-Aldrich), 0.5% v/v NP-40 (Sigma-Aldrich), pH 7.4] containing protease inhibitors 1 mM phenylmethylsulfonyl fluoride (PMSF; Sigma-Aldrich) and protease inhibitor cocktail (PIC; Cytoskeleton, Inc.). The lysates were centrifuged at 11,000 g at 4 °C for 10 min. The supernatants were incubated with Complex I or Complex II immunocapture KIT (Abcam, ab109711 and ab109799, respectively) for 16 h at 4°C. After incubation, immunoprecipitation reaction products were washed ten times in buffer (150 mM NaCl, 1 mM EDTA, 50 mM Tris-HCl, pH 7.4, 5 mM NaF, 0.5% v/v NP-40). Immunocomplexes were eluted by boiling the samples in reducing sample buffer for 5 min and then, size-fractionated on SDS–PAGE.

2.5. Mitochondrial complex I assay.

The activity of mitochondrial complex I was determined using the complex I enzymatic activity assay kit (ab109721) as recommended by the manufacture. Briefly, MCF7 neg shRNA and -NDUFAF3-KD cells were lysated as described in section 2.4 and the complex I activity, define as the oxidation of NADH to NAD+, determine spectrophotometrically at 450 nm.

2.6. Mitochondrial Complex II assay.

The activity of mitochondrial complex II was determined using the complex II enzymatic activity assay kit (ab109908) as recommended by the manufacture. Briefly, MCF7 neg shRNA and MCF7 SDHA-KD cells were lysed as described in section 2.4 and the complex II immunocaptured within the wells of the microplate. The activity of complex II, was determine by the reduction of the dye DCPIP (2,6-dichlorophenolindophenol) during the production of ubiquinol spectrophotometrically at 600 nm.

2.7. Extracellular flux analysis.

Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were assessed in an extracellular flux analyzer XF^e96 (Agilent Technologies™, CA, USA) as described previously [1]. Briefly, MCF7 cell lines generated for this study were seeded on XFe96-well plates (15,000 cells per well) and pre-incubated overnight at 37°C in 5% CO2 atmosphere. The following day, the culture medium was replaced with assay media (unbuffered DMEM supplemented with glutaMAX®, 10 mM glucose and 1 mM pyruvate, pH 7.4) 1 hour prior to the assay and left for the duration of the experiment. After establishing the baseline OCR, cells were sequentially challenged with oligomycin (1 μM), FCCP (125 nM) and rotenone + Antimycin A (both 1 μM) to reveal basal, maximal, ATP-coupled respiration and proton leak. For ECAR, the same assay media was used and cells were sequentially challenged with glucose (10 mM), oligomycin (1 μM) and the hexokinase inhibitor 2-deoxyglucose (2-DG, 100 mM) to reveal the glycolytic capacity of the cells. The data were normalized by protein content.

2.8. Mitochondrial Ca²⁺ dynamics in permeabilized cells.

MCF7 cell lines generated for this study (7.5×10⁶ cells) were permeabilized using digitonin (40 μM) containing intracellular-like medium [ICM; permeabilization buffer: 120 mM KCl; 10 mM NaCl; 1 mM KH2PO4; 20 mM HEPES-Tris, pH 7.2; 2 μM thapsigargin, and protease inhibitors (Roche Applied Science, Minneapolis, MN, USA) supplemented with succinate (3 mM)]. Fura-2FF (ThermoFisher Scientific®) was added to measure extramitochondrial Ca²⁺ using a fluorimeter (Photon Technology International). After baseline recording, 10 μM Ca²⁺ was added to cells at 350s to measure mitochondrial Ca²⁺ uptake, followed by the injection of the MCU inhibitor Ru360 (1 μM) at 550s, the inhibitor of the Na⁺/Ca²⁺ exchanger CGP37157 (10 μM) at 610s, and FCCP (2 μM) at 750s.

2.9. Cell proliferation assay.

Cell lines were incubated for 0, 24, 48 and 72 h. At each time point, culture medium was removed, and cells were carefully washed twice with PBS. After washing, cells were incubated at room temperature for 20 minutes in 0.5% crystal violet 20% methanol staining solution. Next, the plate was washed carefully four times in a gentle stream of tap water, inverted on a filter paper to remove the remaining liquid and dried at room temperature for at least 1 hour. Finally, crystal violet was solubilized adding 200 μL of methanol per well. OD was measured at 570 nm.

2.10. Cell migration assay.

MCF7 cell lines with density of 1 × 10⁶ cells/well were seeded in 6-well plates and grown until a confluent monolayer of cells was achieved. Then, a cross-shaped scratch over the well using a sterile 200 μl tip was made. Wells were washed with sterile PBS to remove detached cells and debris and then cultured in 2 ml of DMEM high glucose media supplemented with 25 mM HEPES and 1% fetal bovine serum to avoid cell proliferation. Wells were photographed in multiple locations with 20x magnification in an Olympus CKX41 microscope using a Qimaging micropublisher 3.3 RTV camera. Images were analyzed with Image J and data expressed as percentage (%) of gap closure as previously described (Liang et al., 2007).

2.11. Cell viability assay.

MCF7 cell lines were seed in 24-well plates and incubated for 24 h. Then, the cells were either exposed to DMSO (control), 5 μM ionomycin, 40 μM C2-ceramide or 30 μM thapsigargin. After 24h of exposure, cells were trypsinized, washed with PBS, and re-suspended in PBS solution with 5 mg/mL of propidium iodide (PI, ThermoFisher Scientific). Incorporation of PI in dead cells was detected using a BD FACSAria III flow cytometer.

2.12. Measurement of Mitochondrial Superoxide

Mitochondrial superoxide was measured by using the mitochondrial oxygen free radical indicator MitoSOX Red (molecular probes; Invitrogen) as previously described (Dong et al., 2017). Briefly, cells were grown on coverslips and loaded with 5 μM MitoSOX Red for 30 min. Coverslips were mounted in a confocal microscope (510 Meta; Carl Zeiss, Inc.) and images were obtained at 561 nm excitation using a 63x oil objective. Images were analyzed, and MitoSOX Red fluorescence was quantified using ImageJ software (NIH). For flow cytometry experiments, 3 × 10⁴ cells were trypsinized and re-suspended in fresh medium without FBS and MitoSOX fluorescence detected using a BD FACSCanto flow cytometer.

2.13. Mitochondrial membrane potential.

MCF7 cell lines were loaded with tetramethylrhodamine methyl ester (TMRM, 5 nM, Life Technologies) for 30 min at 37°C and 5% CO2. Hoechst 33342 (Life Technologies) was used as nuclear counterstain. Images were acquired using a Nikon C2+ spectral confocal microscope using the same settings for all experiments. TMRM fluorescence was quantified with ImageJ software.

2.14. Statistics.

All data are summarized as mean ± SEM; significance of differences was assessed using unpaired t-tests. Differences were accepted as significant at the 95% level (p < 0.05).

3. Results

3.1. Knockdown of NDUFAF3 and SDHA reduces mitochondrial oxygen consumption and increases glycolysis in MCF7 cells.

NDUFAF3 and SDHA are essential for the assembly of complex I and complex II (van den Ecker et al., 2012; Zurita Rendon and Shoubridge, 2012) and their absence is presumed to disrupt mitochondrial respiration. Here, we experimentally determined for the first time the effect of knocking down NDUFAF3 and SDHA on oxygen consumption in MCF7 cells. Using two different shRNA against NDUFAF3 and three against SDHA we generated stable knockdown MCF7 cell lines. As observed in Figure 1A and 1B, NDUFAF3–271 and NDUFAF3–564 cell lines show a decrease in NDUFAF3 protein expression of 52% and 74.5% respectively (Fig. 1A and C). On the other hand, from the stable cell lines generated to knockdown SDHA, only SDHA-2 showed a significant reduction in the expression of SDHA (65%) (Fig. 1B, D). The level of complex V subunit ATP5A, which was used as a mitochondrial loading control, remained unchanged (Fig. 1A and B) similar to what is observed with the housekeeping protein actin. Given these results, we continued our studies in cell lines NDUFAF3–564 and SDHA-2 which hereafter we refer to as NDUFAF3 KD and SDHA KD respectively. To explore the assembly status of complex I and complex II in these cells, we determine whether subunits NDUFB8 and SDHB co-immunoprecipitated with complex I and complex II respectively. As expected, when complex I is immunoprecipitated in the control cell line (neg shRNA) the complex I structural subunit NDUFB8 is present indicating assembly of complex I. However, this subunit is not seen when complex I is immunoprecipitated in NDUFAF3 KD cells, confirming that NDUFAF3 assembly factor is required for complex I assembly (Barbi de Moura et al., 2012) (Fig. 1E). Similarly, the complex II structural subunit SDHB co-immunoprecipitated with complex II in the control cell line, but not in the SDHA KD cell line (Fig. 1F) confirming the absence of the SDHA catalytic subunit impairs complex II assembly (Saxena et al., 2016). Activity assays confirmed the reduced function of complex I and complex II in NDUFAF3 KD (Fig. 1G) and SDHA KD (Fig. 1H) cells, respectively.

Figure 1. Knockdown of NDUFAF3 and SDHA subunits affects global cellular bioenergetics.

A. Representative Western blots of NDUFAF3, ATP5A in total lysates from MCF7 cells stable expressing either a non-target shRNA or a shRNA against NDUFAF3 (clones 271 and 564). B. Representative Western blots of SDHA, and ATP5A in total lysate from MCF7 cells stable expressing either a non-target shRNA (neg) or a shRNA against SDHA (clones 2, 3 and 5). C. Analysis of NDUFAF3 expression normalized over mitochondrial marker ATP5A and expressed as average fold increase over control cells (neg). D. Analysis of SDHA levels normalized over mitochondrial marker ATP5A and expressed as average fold increase over control cells. E. Co-immunoprecipitation of complex I and NDUFB8. F. Co-immunoprecipitation of complex II and SDHB. G. Determination of complex I activity at 450 nm over time. At the straight portion of the progression curve, slope (mOD/min) was calculated. H. Determination of complex II activity at 600 nm over time. At the straight portion of the progression curve, slope (mOD/min) was calculated. I. Representative traces of oxygen consumption rate (OCR) in neg shRNA, NDUFAF3–564 and SDHA-2 MCF7 cells exposed sequentially to oligomycin (oligo, 1 μM), FCCP (0.125 μM) and antimycin A plus rotenone (A+R, 1 μM each). J. Basal respiration calculated by subtracting the non-mitochondrial OCR (after A+R) from the OCR measurement before oligomycin. K. Maximal respiration calculated by subtracting non-mitochondrial OCR (after A+R) to the OCR after FCCP. L. ATP coupled OCR calculated by subtracting OCR after and before oligomycin. M. Representative traces of extracellular acidification rate (ECAR) in cells exposed sequentially to glucose (10mM), oligomycin (oligo, 1 μM) and 2-deoxyglucose (2DG, 100 mM). N. Glycolysis calculated by subtracting ECAR measured before glucose from ECAR after 2-DG. O. Glycolytic capacity calculated by subtracting ECAR measured before 2-DG from ECAR measured after oligomycin. All data shown represent mean ± SEM of at least three independent experiments. * = p<0.05, ** = p <0.01, *** = p < 0.001, ns = not significant.

Next, to determine whether NDUFAF3 KD and SDHA KD cell lines have impaired mitochondrial respiration, we performed oxygen consumption measurements. Both NDUFAF3 KD and SDHA KD cell lines showed lower levels of basal and maximal oxygen consumption rate (OCR) than the control cell line (neg shRNA) (Fig. 1I, J and K), confirming a defect in the electron transport chain. In addition, the respiration coupled to the ATP generation, calculated as the difference in OCR before and after oligomycin injection was also significantly diminished (Fig. 1L).

It is well documented that mitochondrial electron transport chain inhibition induces a compensatory rise of glycolysis (Hill et al., 2012; Saxena et al., 2016). Thus, by measuring the extracellular acidification rate (ECAR) we determined that NDUFAF3 KD and SDHA KD cells present higher glycolytic flux than the control cells (Fig. 1M and N). In addition, by inhibiting ATP synthase with oligomycin, we were able to determine that NDUFAF3 and SDHA KD cells demonstrated a higher glycolytic capacity, suggesting a metabolic remodeling toward glycolysis (Fig. 1O).

3.2. Intact complex I or complex II are necessary to maintain _mCa²⁺ homeostasis.

Maintenance of the ΔΨ_m is essential for cellular Ca²⁺ homeostasis (Contreras et al., 2010). Inhibition of complex I activity has been associated with both increased and decreased ΔΨ_m (Abramov et al., 2010; Distelmaier et al., 2009a; Distelmaier et al., 2009b; Koopman et al., 2005; Valsecchi et al., 2012), however, in NDUFAF3 KD cells the ΔΨ_m remained unaffected (Figure 2A and B). On the other hand, complex II plays an important role in maintaining ΔΨ_m in normoxic and hypoxic conditions (Hawkins et al., 2010). Surprisingly, SDHA knockdown induced a small, but significant increase in ΔΨ_m (Figure 2A and B) rather than a decrease. Under these conditions we hypothesized that mitochondrial Ca²⁺ flux should remain unchanged or increased in the case of the SDHA knockdown cells. Thus, we examined cytosolic clearance, mitochondrial Ca²⁺ efflux and release of total mitochondrial Ca²⁺ by assessing free extramitochondrial Ca²⁺ concentration in permeabilized cells as described before (Doonan et al., 2014). First, we triggered Ca²⁺ uptake, by adding a pulse of Ca²⁺ to permeabilized neg-shRNA, NDUFAF3, and SDHA KD cells in presence of the Ca²⁺ indicator, Fura-2FF. Surprisingly, the Ca²⁺ uptake rate was diminished in NDUFAF3 and SDHA KD cells (Fig. 2C and D). Then, in order to visualize [Ca²⁺]_m efflux, we blocked mitochondrial Ca²⁺ uptake by adding Ru360, which revealed that both NDUFAF3 and SDHA KD cells have reduced efflux rate, suggesting that [Ca²⁺]_m steady state was lower in the KD cells (Fig. 2E and F). Mitochondrial efflux was stopped with the addition of NCLX inhibitor CGP-37157. Finally, to determine the total levels of mitochondrial Ca²⁺, the mitochondrial membrane potential was dissipated by adding the uncoupler carbonyl cyanide-4-(trifluoromethoxy)phenylhydrazone (FCCP), which revealed that NDUFAF3 and SDHA KD cells have a significantly lower [Ca²⁺]_m content compared with control cells (Fig. 2G and H). Altogether, these results confirm that the absence of either complex I or complex II activity and/or levels, reduce [Ca²⁺]_m efflux and influx rate, shifting the equilibrium of mitochondrial Ca²⁺ towards a lower [Ca²⁺]_m accumulation.

Figure 2. Effects of NDUFAF3 and SDHA knockdown on mitochondrial membrane potential and Ca2+ influx and efflux rates.

A. Representative confocal images of neg shRNA, NDUFAF3–564 and SDHA-2 MCF7 cell lines loaded with mitochondrial membrane potential indicator TMRE. B. Quantification of TMRE fluorescence C. Representative Ca²⁺ traces of neg shRNA, NDUFAF3–564 and SDHA-2 MCF7 cell lines permeabilized with 40 μg/mL of digitonin and loaded with the ratiometric Ca²⁺ indicator Fura 2-FF. A pulse of 10 μM Ca²⁺ was added at 350 s to measure mitochondrial Ca²⁺ uptake, followed by the addition of 1 μM Ru360 at 550 s, 10 μM CGP37157 at 610 s, and 2 μM uncoupler FCCP, at 750 s. D. Quantification of Ca²⁺ uptake rate. E. Magnification of Ca²⁺ efflux from panel A (between 525 and 650 s). F. Quantification of Ca²⁺ efflux rate after addition of RU360. G. Magnification of Ca²⁺ release after FCCP addition from panel A (between 700 and 900 s). H. Quantification of FCCP-induced release of mitochondrial Ca²⁺. All data shown represent mean ± SEM of three independent experiments. *p<0.05, **p<0.01, ***p<0.001. ns = not significant.

3.3. Effect of NDUFAF3 and SDHA knockdown in cellular proliferation and migration.

Inhibition of Ca²⁺ transfer to mitochondria lowers mitochondrial respiration reducing the clonogenic capacity of breast cancer cells (Cardenas et al., 2016). As the lack of complex I or complex II activity decreases respiration and the flux of Ca²⁺ to mitochondria, we assessed proliferation in NDUFAF3 and SDHA KD cells and no differences were found after 72h (Fig. 3A and B). Accordingly, no increase in cell death was observed in the KD cells compared to the control (Fig. 3C). Cellular migration can also be affected by a reduction of mitochondrial Ca²⁺ content (Tosatto et al., 2016), thus, we studied migrating behavior in NDUFAF3 and SDHA KD cells. Remarkably, NDUFAF3 KD cells migrate faster than control cells (3D and E), while SDHA KD cells behave as controls, suggesting that another mechanism independent of mitochondrial Ca²⁺ homeostasis operates during migration in complex I KD cells. Reactive oxygen species (ROS), which are known to increase with mutations in mitochondrial or nuclear DNA-encoded ETC proteins that diminish complex I or II (Sabharwal and Schumacker, 2014; Sullivan and Chandel, 2014), play an important role in migration and cell adhesion (Hurd et al., 2012). We therefore determined the levels of superoxide in both NDUFAF3 and SDHA KD cells by confocal microscopy, as done before by our group (Mallilankaraman et al., 2012b), and by flow cytometry. Surprisingly, mitochondrial superoxide levels in NDUFAF3 KD cells were similar to controls with both techniques (Fig. 3F–H), while SDHA KD cells show a small, but significant rise in mitochondrial superoxide levels by confocal microscopy than is not seen by flow cytometry (Fig. 3F–H).

Figure 3. Effects of NDUFAF3 and SDHA knockdown on proliferation, migration and mitochondrial ROS generation.

A. Representative images of proliferating cells stained with crystal violet at 0, 24, 48 and 72 hrs. B. Cell proliferation rates were determined by absorbance of crystal violet stain. C. Cell death determine as the percentage of cells that incorporated propidium iodide (PI). D. Representative images of a wound healing assay of neg shRNA, NDUFAF3–564 and SDHA-2 cells. E. Migration was determined as the percentage of cells covering the scratched area. F. Representative images of neg shRNA, NDUFAF3–564 and SDHA-2 cells stained with MITOSOX red. Scale bar represents 10 μm. G. Quantitation of MITOSOX red fluorescence. H. quantification of MITOSOX red fluorescence by flow cytometry. All data shown represent the mean ± SEM of three independent experiments. *p<0.05, **p<0.01, ***p<0.001.

3.4. The absence of complex I and II prevents cell death induced by _mCa²⁺ overload.

Given the shifting of the _mCa²⁺ equilibrium towards lower [Ca²⁺]_m accumulation in NDUFAF3 and SDHA KD cells, we determined whether these cells were more resistant to cell death induced by mitochondrial Ca²⁺ overload. Accordingly, we treated all cells with 5 μM ionomycin or 40 μM ceramide C2 for 24h and measured cell death using propidium iodide and flow cytometry. As shown in Figure 4, complex I and complex II KD cells were significantly more resistant to these challenges than control cells.

Figure 4. Effects of NDUFAF3 and SDHA knockdown on mitochondrial Ca2+ overload-induced cell death.

Quantification of dead cells was determined by propidium iodide incorporation measured with flow cytometry (10,000 events) in cells treated either with ionomycin or ceramide for 24 h. All data shown represent the mean ± SEM of three independent experiments. *p<0.05, **p<0.01, ***p<0.001.

4. Discussion

Although the regulation of the ETC complex and its influence on cell physiology has been studied at length (Guo et al., 2018; Lobo-Jarne and Ugalde, 2018), there are still aspects that have received little attention. For instance, Ca²⁺, a pivotal regulator of mitochondrial function (Cardenas et al., 2010), and key determinant of cell fate (Berridge et al., 2003) has been assumed to be regulated by the ETC indirectly through the ΔΨ_m. Here, we show that either complex I or II regulate _mCa²⁺ homeostasis independently of the ΔΨ_m. Downregulation of NDUFAF3, an early assembly factor of complex I, decreased basal, ATP-coupled and maximal oxygen consumption and the mitochondrial Ca²⁺ fluxes in the human cancer cell line MCF7. The ΔΨ_m was conserved to a similar extent as in control cells. The same characteristics were observed during complex II subunit SDHA knockdown. SDHA contains the prosthetic group that oxidizes succinate to fumarate. Mitochondrial internal membrane damage can cause low ETC activity (Brand and Nicholls, 2011), however, this possibility was discarded in our system since proton leak showed no difference between knockdown and control cells (data not shown), indicating that the internal mitochondrial membrane was intact.

The lack of ΔΨ_m depolarization observed in NDUFAF3 and SDHA KD cells was expected as it has been established that in either absence or low activity of complex I, II, III and IV, the ΔΨ_m is maintained by complex V (F₁F₀-ATP synthase) reverse activity, which pumps protons towards the mitochondrial intermembrane space hydrolyzing ATP as the energy source (Chinopoulos, 2011; Jonckheere et al., 2012). The maintenance of the ΔΨ_m is essential to avoid the activation of the intrinsic branch of apoptosis (Gottlieb et al., 2003). For example, in neurons carrying either a severe mutant form of complex I that reduced its activity to less than 10%, or a mutant form of complex IV that reduced its activity to 40%, the ΔΨ_m remained stable (Abramov et al., 2010). Similarly, primary isocortical neurons and astrocytes derived from a complex I deficient Ndufs4 knockout mouse showed no alteration in ΔΨ_m (Bird et al., 2014). However, in both cases the ΔΨ_m collapsed in response to oligomycin, a well-known inhibitor of complex V activity (Hong and Pedersen, 2008), demonstrating that the ΔΨ_m was maintained by complex V in reverse mode. Supporting the notion that the ATP synthase reverse activity could maintain the ΔΨ_m in NDUFAF3 and SDHA KD cells, both cell lines exhibited an increased glycolytic flux. Glycolysis and OXPHOS are tightly coupled by intermediary metabolites, as has beenrevealed in cancer cells (Vander Heiden et al., 2009). Complex I uses NADH as a source of electrons for proton pumping and for ubiquinone reduction. When complex I is inhibited for example by rotenone, NADH concentration increases, diminishing NAD⁺/NADH ratio, which in turn, increases the conversion of pyruvate to lactate. Therefore, increasing NADH concentration, increases glycolytic flux as well. Also, when OXPHOS activity is decreased due to low SDH activity, AMPK is activated, upregulating lactate dehydrogenase activity (Hou et al., 2018), which replenishes ATP levels necessary to maintain Complex V working in reverse and maintaining ΔΨm.

Strikingly, the ΔΨ_m in SDHA knockdown cells is even higher than in control cells. Inhibition of complex I can induce a mitochondrial hyperpolarization due to low activity of complex II, III, IV and V (Forkink et al., 2014), and therefore a similar mechanism could be expected to operate in these cells. Since ΔΨ_m is the main driving force for Ca²⁺ uptake into mitochondria, no changes were expected in the mitochondrial Ca²⁺ flow in NDUFAF3 and SDHA KD cells. Unexpectedly, we found that the Ca²⁺ influx to the mitochondrial matrix as well as the mitochondrial Ca²⁺ efflux was lower in both knockdown cell lines (Fig. 2), causing a lower overall mitochondrial Ca²⁺ content. As far as we know, this is the first time that a correlation between either complex I or II and mitochondrial Ca²⁺ homeostasis independent of ΔΨ_m has been established. In a previous report, fibroblasts from patients carrying isolated mutations in complex I exhibited a reduced _mCa²⁺ uptake/accumulation that impaired ATP production, but this occurred alongside a reduction in the ΔΨ_m (Visch et al., 2004). We, on the other hand, show that a reduction of either complex I or complex II in absence of an altered ΔΨ_m, establishes a new Ca²⁺ “set-point” (Mallilankaraman et al., 2012b; Nicholls, 2017). The entrance of Ca²⁺ to mitochondria is mainly through the mitochondrial calcium uniporter (MCU) complex, which is immersed in the inner mitochondrial membrane. It is composed by several functional and regulatory subunits, MCU being the pore-forming subunit (Baughman et al., 2011; Csordás et al., 2013; De Stefani et al., 2011; Mallilankaraman et al., 2012a; Mallilankaraman et al., 2012b; Patron et al., 2014; Perocchi et al., 2010; Plovanich et al., 2013; Sancak et al., 2013). The intake of Ca2+ through MCU is regulated by the MCUb (Raffaello et al., 2013) MICU1 (Patron et al., 2014) (Hoffman et al., 2013), MICU2 (Plovanich et al., 2013), MCUR1 (Mallilankaraman et al., 2012a) and EMRE (Sancak et al., 2013). MICU1 and MICU2 inhibit the pore-conformation of the complex, while MCUR1, regulates positively the pore activity (Woods et al., 2019). Since a decrease in MCU protein levels was not detected in our KD cell lines (data not shown), the possibility of the existence of a spatial and/or functional interaction between ETC proteins and MCU-regulatory subunits should be not dismissed. Additionally, the MCU complex is post-transcriptionally regulated by phosphorylation and chemical modification such S-glutathionylation (Pallafacchina et al., 2018), which may explain the phenomenon observed in NDUFAF3 and SDHA KD cells. New experiments are necessary to determine the exact mechanism responsible for the reduction of mCa2+ flux observed in these cells.

In many cellular contexts, a reduction in _mCa²⁺ uptake/accumulation allows cells to resist insults that would otherwise cause cell death mediated by activation of the mitochondrial permeability transition (MPT) (Angelova et al., 2019; Kon et al., 2017; Luongo et al., 2015; Oropeza-Almazan et al., 2017). Therefore, it is not surprising that NDUFAF3 and SDHA KD cells show resistance to death upon stimulation with known inductors of cell death mediated by _mCa²⁺ overload such as ionomycin (Nemani et al., 2018) and ceramide (Parra et al., 2013). Interestingly, a decrease of either complex I or II activity due to subunit mutations can generate cancer mainly by increasing resistance to apoptosis (Kluckova et al., 2013; Urra et al., 2017). Therefore, in the context of cancer mediated by either complex I or II impaired activity, it is possible to speculate that at least the resistance to apoptosis could be the result of low _mCa²⁺ uptake/accumulation.

The lack of mitochondrial respiration does not affect proliferation, as shown here for both NDUFAF3 and SDHA KD cell lines, as long the ΔΨ_m is sustained by an upregulation of glycolysis and the reverse activity of the F₀F₁-ATP synthetase. Martinez-Reyes et al., 2016, elegantly demonstrated that restoring the reverse activity of the F₀F₁-ATP synthetase by knocking out its endogenous physiological inhibitor ATPIF1, was sufficient to restore the ΔΨ_m and proliferative capability of cells lacking mtDNA. On the other hand, restoring TCA cycle oxidative metabolites by ectopic expression of S. cerevisae NDI1 and C. intestinalis, which does not recover the ΔΨm, was not sufficient to maintain proliferation (Martinez-Reyes et al., 2016), highlighting the importance of the ΔΨm in proliferation. Cell migration is a pivotal step in metastasis (Yilmaz and Christofori, 2010), and is responsible for most patient deaths from solid tumors (Steeg, 2016). Several reports show that down-modulation of certain complex I subunits by genetic or pharmacologic means generates an enhanced migratory behavior of cancer cells and metastasis (He et al., 2013; Ishikawa et al., 2008; Li et al., 2015; Yuan et al., 2015), through a ROS-mediated mechanism that involves modulation of extracellular matrix proteins. In NDUFAF3 KD cells we observed an increase in cell migration, however, this was independent of ROS generation. On the other hand, germline complex II mutations that cause pheochromocytoma/paraganglioma syndrome type 4 (PGL4), which is characterized by the presence of pheochromocytoma and paraganglioma, two types of cancer that very rarely metastasize (Angelousi et al., 2015), and gastrointestinal stromal tumors and renal tumors, which have the potential to metastasize, act through a mechanism mediated by ROS (Shanmugasundaram and Block, 2016; Shi et al., 2017). However, SDHA knockdown cell lines did not show any increase in their migrative behavior compared with control cells. These results that seem controversial remind us there are many factors that regulate complex cellular phenomena such as cellular migration.

5. Conclusions

Overall, our data demonstrate that decreasing complex I and II levels affects Ca²⁺ homeostasis, leading to a decrease in the _mCa²⁺ set-point independent of ΔΨ_m. This could have some interesting implications, as until now, the mitochondrial Ca²⁺ homeostasis has been overlooked as related to the array of diseases that involve mutations of the ETC that impair mitochondrial function without loss of the ΔΨ_m. Thus, the modulation of the _mCa²⁺ offers a new window of intervention in cancer and other diseases related to either complex I or II downregulation activity.

HIGHLIGHTS.

• Knockdown of NDUFAF3 and SDHB altered mitochondrial Ca²⁺ efflux and influx rates in absence of ΔΨm loss.

• Shifting the equilibrium toward lower concentration of mitochondrial Ca²⁺ accumulation renders cells resistant to death.

• Our findings reveal an unexpected relationship between complex I and II with the mitochondrial calcium homeostasis independent of ΔΨm.