Calcium signalling-dependent mitochondrial dysfunction and bioenergetics regulation in respiratory chain Complex II deficiency

Abstract

Despite advanced knowledge on the genetic basis of oxidative phosphorylation-related diseases, the molecular and/or cellular determinants for tissue-specific dysfunction are not completely understood. Here, we report the cellular events associated with mitochondrial respiratory Complex II deficiency occurring before cell death. Mutation or chronic inhibition of Complex II determined a large increase of basal and agonist-evoked Ca²⁺ signals in the cytosol and the mitochondria, in parallel with mitochondrial dysfunction characterized by membrane potential (Δψ_mit) loss, [ATP] reduction and increased reactive oxygen species production. Cytosolic and mitochondrial Ca²⁺ overload are linked to increased endoplasmic reticulum (ER) Ca²⁺ leakage, and to SERCA2b and PMCA proteasome-dependent degradation. Increased [Ca²⁺]_mit is also contributed by decreased mitochondrial motility and increased ER-mitochondria contact sites. Interestingly, increased intracellular [Ca²⁺] activated on the one hand a compensatory Ca²⁺-dependent glycolytic ATP production and determined on the second hand mitochondrial pathology. These results revealed the primary function for Ca²⁺ signalling in the control of mitochondrial dysfunction and cellular bioenergetics outcomes linked to respiratory chain Complex II deficiency.

Main

Mitochondria are the driving force behind life, as mitochondrial oxidative phosphorylation provides the main source of ATP in the cell. In addition to energy production, mitochondria have a crucial function in mediating amino-acid biosynthesis, fatty acid oxidation, intermediate metabolic pathways, free radicals production and Ca²⁺ homeostasis.¹

Mitochondria affect intracellular Ca²⁺ metabolism in two ways: (i) directly by regulating both the amplitude, duration, location and propagation of cytosolic Ca²⁺ elevations and the recycling of Ca²⁺ towards the endoplasmic reticulum (ER) and² (ii) indirectly by producing ATP, which is used by Ca²⁺-dependent ATPases to pump Ca²⁺ out of the cell (by the plasma-membrane Ca²⁺ ATPase: PMCA) or into intracellular stores (by the sarco-ER Ca²⁺ ATPase: SERCA). Conversely, Ca²⁺ entering to the mitochondrial matrix regulates mitochondrial metabolism through the activation of the Ca²⁺-dependent enzymes of the Krebs cycle.³

Leigh's syndrome is a rare but severe and fatal encephalopathy of early childhood that is most frequently associated with deficiencies in nucleus-encoded subunits of Complex I (NDUFS3 and NDUFS7),^{4, 5} Complex II (SDHA),⁶ Complex IV (SURF1)⁷ or pyruvate dehydrogenase.⁸ Despite the identification of the genetic origin of Leigh's syndrome, the molecular and cellular events associated with pathology are not completely understood.

Deregulations of intracellular Ca²⁺ signalling have been reported in different models of mitochondrial respiratory chain diseases,^{9, 10, 11} whereas data on Complex II deficiency are still lacking.

Complex II (succinate: ubiquinone (UQ) oxidoreductase) has a central function in oxidative metabolism, being an important enzyme of the citric acid cycle as well as a Complex of the mitochondrial respiratory chain.¹²

We report here the cellular events deriving from Complex II mutation associated with Leigh's syndrome. This study was conducted also in cellular models in which we mimicked Complex II deficiency by using Complex II inhibitors. We showed broad subcellular Ca²⁺ signalling alterations occurring before cell demise in Complex II-deficient cells. Interestingly, we showed that cytosolic and mitochondrial Ca²⁺ overload are linked to SERCA and PMCA degradation occurring in a proteasome-dependent manner. Our study also revealed a dual function of Ca²⁺ deregulation in the control of cellular bioenergetics outcome and mitochondrial pathology.

Results

We studied fibroblasts isolated from a patient with Leigh's syndrome harbouring a homozygous Arg554Trp substitution in the Fp subunit of the Complex II (SDHA_R554W).¹³ These fibroblasts were shown to present a reduction of Complex II activity to ∼30% of control values.¹³ Leigh's syndrome associated with Complex II mutations was described in very rare cases.⁶ To bypass this limitation, we conducted the whole study using two inhibitors of Complex II (3-NP (3-nitropropionic acid) or Atpenin A5^{14, 15}) in both fibroblasts derived from healthy individuals and the neuronal-derived cell line (SH-SY5Y). 3-NP is structurally similar to Complex II substrate succinate thus interacting with some residue in the active site of the enzyme.¹⁴ Atpenins were shown to share structural similarity with UQ thus blocking the electron transfer between the enzyme and UQ by binding to a region that partly overlaps with the physiological UQ-binding site.¹⁵ To note, we deliberately used (i) concentrations of 3-NP (300 μM) and Atpenin A5 (100 nM), which did not induce cell death at 20 h thus affording to study Complex II deficiency independently from cell death (Supplementary Figure S1a) and (ii) extended application (20 h) of the drugs to chronically inhibit Complex II (Supplementary Figure S1b). Importantly, under these conditions, Complex II activity was reduced to 15–40% of control values in both mutated cells and control fibroblasts and SH-SY5Y cells (Supplementary Figure S1b and Table S1). Application of 3-NP or Atpenin A5 at the same concentrations for 72 h or application of 1 mM 3-NP for 20 h determined cell death showing that Complex II inhibition is linked to cell death (Supplementary Figure S1c and 1d right).

To show that in our conditions we mimicked chronic Complex II inhibition on drug application, we measured Complex II activity at 6 and 20 h after treatment with 300 μM 3-NP. Interestingly, reduction of Complex II activity occurred on 6 h of treatment and was maintained to a similar level on 20 h of treatment (data not shown).

We assessed mitochondrial functionality on Complex II mutation or inhibition. Mitochondrial potential (Δψ_mit) was measured by using tetramethyl rhodamine methyl ester (TMRM) dye, which accumulates within the mitochondrial matrix as a function of Δψ_mit. As expected, we observed a significant decrease (∼70% versus control considered as 100%) of Δψ_mit on Complex II mutation or inhibition in both fibroblasts and SH-SY5Y cells (Figures 1a and b, respectively; Supplementary Table S2).

Figure 1

Mitochondrial function is altered on Complex II defect. (a, b) Δψ_mit, as assessed by TMRM probe in SDHA_R554W-mutated fibroblasts, and in control fibroblasts (a) and SH-SY5Y cells (b) left untreated (control) or treated with 300 μM 3-NP or 100 nM Atpenin A5 for 20 h. Basal TMRM intensity was normalized to the fluorescence drop (Δψ_mit collapse) on 10 μM FCCP application. Representative traces show normalized f/f₀ TMRM intensity, where f is the basal TMRM intensity, and f₀ the TMRM intensity after FCCP application. Graphs show normalized TMRM (f/f₀) intensity in percentage versus control considered as 100%. Images are representative of TMRM loading in mutated fibroblasts and controls left untreated or treated as detailed before. The mean of normalized TMRM signal±S.E.M., normalized TMRM signal (%) and the number of experiments are represented in Supplementary Table S2. (c) Superoxide content as determined with MitoSOX dye in SDHA_R554W fibroblasts, and in control fibroblasts left untreated (control) or treated as in (a). Representative images of MitoSOX loading are shown. The mean of MitoSOX intensity±S.E.M. and the number of experiments are represented in Supplementary Table S3. (d, e) Mitochondrial basal ATP level normalized to the maximum luciferase signal in SDHA_R554W and control fibroblasts (d) and in SH-SY5Y cells left untreated (control) or treated as in (a). (e) Data are expressed as percentage of luciferase light output (CPS: counts per second) with respect to the value obtained after cell permeabilization in the presence of maximal [ATP] (see Materials and methods for details). Representative traces of normalized luciferase signal (% CPS) are shown in the insets. The mean of basal ATP level±S.E.M. and the number of experiments are represented in Supplementary Table S4. (a–e) Data are representative of three independent experiments. ^*P⩾0.001, ^**P< 0.001 versus control

The mitochondrial respiratory chain is known to be the major source of reactive oxygen species (ROS) within the cell.¹⁶ We measured mitochondrial production of superoxide (O₂^•−) by using MitoSOX red-mitochondrial superoxide indicator. Complex II-deficient cells displayed an increase in mitochondrial O₂^•− production of almost 25% as compared with control (Figure 1c; Supplementary Table S3). The specificity of the test was confirmed by using the ROS scavenger pyrrolidine dithiocarbamate (PDCT) (Figure 1c; Supplementary Table S3). This result is consistent with earlier reported data, which show increased activities of both mitochondrial and cytosolic inducible superoxide dismutase in SDHA_R554W cells.¹⁷

Next, we analysed the impact of the mitochondrial Complex II defect on ATP level. Basal-mitochondrial ATP concentration ([ATP]_mit) was measured using a luciferase probe targeted to the mitochondria (mitLuc). We observed a decreased basal [ATP]_mit in SDHA_R554W-mutated fibroblasts (74% of that in control cells) as well as in SH-SY5Y cells on Complex II inhibition (83% in 3-NP and 77% in Atpenin A5 of that in control cells) (Figures 1d and e; Supplementary Table S4).

Ca²⁺ is an important regulator of mitochondrial function and acts at several levels within the organelle to stimulate OXPHOS.³ We measured subcellular Ca²⁺ signalling in fibroblasts and SH-SY5Y cells on Complex II mutation or inhibition. Measurement of basal [Ca²⁺]_cyt, by using the Fluo-4, AM dye, revealed a significant increase (∼2–4-fold) of basal [Ca²⁺]_cyt in SDHA_R554W-mutated fibroblasts and in control fibroblasts on Complex II inhibition as compared with non-treated ones (Figure 2a; Supplementary Table S5). Similarly, we also observed an increase of basal [Ca²⁺]_mit in SDHA_R554W-mutated fibroblasts and in control fibroblasts on Complex II inhibition (∼1.5–2-fold) as compared with non-treated cells as measured by X-Rhod-1, AM probe (Figure 2b; Supplementary Table S6).

Figure 2

Complex II defect leads to cytosolic and mitochondrial Ca²⁺ signalling deregulation. (a, b) Basal cytosolic [Ca²⁺] (calibrated Fluo-4, AM intensity, see Materials and methods for details) (a) and basal-mitochondrial [Ca²⁺] (calibrated X-Rhod-1, AM intensity, see Materials and Methods for details) (b) in SDHA_R554W-mutated fibroblasts and in control fibroblasts left untreated (control) or treated as in Figure 1a. Representative images for Fluo-4, AM and X-Rhod-1, AM loading are shown. The mean value±S.E.M. and the number of experiments are indicated in Supplementary Tables S5 and S6. (c, d) Representative curves of agonist-evoked cytosolic Ca²⁺ signal ([Ca²⁺]_cyt) (c) and mitochondrial Ca²⁺ signal ([Ca²⁺]_mit) (d) in control and SDHA_R554W fibroblasts (left panel) and in SH-SY5Y cells (right panel) left untreated (control) or treated as in Figure 1a. The mean values±S.E.M. and the number of experiments are indicated in Supplementary Tables S7 and S8. (a–d) Data are representative of three independent experiments. ^*P⩾0.001, ^**P< 0.001 versus control

We then analysed agonist-evoked cytosolic and mitochondrial Ca²⁺ signals using adenovirus-delivered recombinant aequorin Ca²⁺ probes (cytAEQ, mitAEQ, respectively).

By using the cytosolic aequorin probe, we showed a significant increase (∼1.5–1.7-fold) of the agonist-evoked cytosolic Ca²⁺ signal ([Ca²⁺]_cyt) in SDHA_R554W-mutated fibroblasts and SH-SY5Y cells on Complex II inhibition as compared with controls (Figure 2c; Supplementary Table S7).

By using the mitochondrial aequorin probe, we noticed a significant increase of agonist-evoked mitochondrial Ca²⁺ signal ([Ca²⁺]_mit) in SDHA_R554W-mutated fibroblasts (∼3-fold increase) and SH-SY5Y cells on Complex II inhibition (∼1.5-fold increase) as compared with controls (Figure 2d; Supplementary Table S8).

Mitochondrial Ca²⁺ uptake and Δψ_mit are tightly cross-regulated.¹ On the one hand, the main driving force for Ca²⁺ accumulation across the inner mitochondrial membrane is the electrochemical gradient (ΔμH) established and maintained by the respiratory chain, whereas, on the other hand, depolarization occurs transiently after Ca²⁺ uptake in physiological conditions.¹ We measured the kinetics of the Ca²⁺ and potential changes occurring in mitochondria of Complex II-deficient cells. Interestingly, we observed an increased agonist-evoked mitochondrial Ca²⁺ response as early as after 4 h of treatment with Complex II inhibitors (Figure 3a; Supplementary Table S8). This result was observed with both inhibitors 3-NP and Atpenin A5 and in both fibroblasts and SH-SY5Y cells (Figure 3a; Supplementary Table S8). In parallel, no modification of Δψ_mit was noticed in cells treated in the same conditions (Figure 3b; Supplementary Table S2). This result suggests that increase of mitochondrial Ca²⁺ uptake precedes the loss of Δψ_mit in Complex II-deficient cells.

Figure 3

Mitochondrial Ca²⁺ overload occurs early on Complex II deficiency. (a) Agonist-evoked mitochondrial Ca²⁺ signal ([Ca²⁺]_mit) in control fibroblasts and SH-SY5Y cells untreated (control) or treated with 300 μM 3-NP or 100 nM Atpenin A5 for 2, 4 or 6 h. The mean values±S.E.M. and the number of experiments are indicated in Supplementary Table S8. (b) Δψ_mit, as performed in Figures 1a and b in control fibroblasts and SH-SY5Y cells untreated (control) or treated as in (a). Graphs show normalized TMRM intensity (f/f₀). The mean f/f₀ values±S.E.M. and the number of experiments are represented in Supplementary Table S2. Data are representative of three independent experiments. ^*P⩾0.001, ^**P< 0.001 versus control

We then investigated the mechanisms underlying increased [Ca²⁺] associated with Complex II deficiency. As cytosolic and mitochondrial Ca²⁺ signals are largely dependent of ER Ca²⁺ store, we measured the state of ER Ca²⁺ load by using the ER-targeted aequorin probe. SDHA_R554W mutation determined a decrease of steady state ER-free Ca²⁺ ([Ca²⁺]_ER) by almost ∼40% as compared with three different controls (Figure 4a; Supplementary Table S9). A slighter but significant decrease (∼5%) of ER steady state was also obtained on Complex II inhibition by 3-NP and Atpenin A5 (Supplementary Figure S2a and Table S9). We then showed that ER Ca²⁺ depletion is associated with decreased ER Ca²⁺ uptake rate, and increased passive ER Ca²⁺ leakage (Figures 4b and c, respectively). No modification of agonist-evoked Ca²⁺ release from the store was noticed between control and Complex II-deficient cells (Supplementary Figure S2b).

Figure 4

Increased intracellular Ca²⁺ signals on Complex II deficiency is associated with deregulated ER Ca²⁺ homeostasis without a modification of calcium entry through plasma membrane. (a) Representative curves of ER Ca²⁺ concentration ([Ca²⁺]_ER) in control and SDHA_R554W fibroblasts. Graph: mean±S.E.M. The mean values±S.E.M. and the number of experiments are indicated in Supplementary Table S9. (b) Dependence on [Ca²⁺] of the ER Ca²⁺ uptake rate analysed on application of CaCl₂ solution (see Materials and methods for details). (c) Dependence on [Ca²⁺] of the ER Ca²⁺ leak rate analysed on application of the SERCA inhibitor TBuBHQ (see Materials and Methods for details). (d) Capacitative Ca²⁺ influx analysis in control and SDHA_R554W fibroblasts challenged first with 100 μM histamine in 1 mM EGTA to empty intracellular Ca²⁺ pools and then with 100 μM histamine in 1 mM CaCl₂ to activate Ca²⁺ entry. (a–d) Data are representative of three independent experiments. ^**P<0.001 versus control

Earlier data have shown that ER Ca²⁺ depletion can be linked to the induction of the unfolded protein response (UPR).¹⁸ However, our data revealed no induction of UPR in SDHA_R554W cells (Supplementary Figure S3).

Finally, no clear modification of capacitative Ca²⁺ entry was noted in SDHA_R554W-mutated cells as compared with controls (Figure 4d) ruling out a change in Ca²⁺ influx in the observed [Ca²⁺]_cyt and [Ca²⁺]_mit increase.

These results may indicate that the increased cytosolic and mitochondrial agonist-evoked Ca²⁺ signals are largely linked to increased basal Ca²⁺ levels. This late event is likely due, at least in part, to decreased ER Ca²⁺ store load and increased ER Ca²⁺ emptying through increased Ca²⁺ leak.

To investigate the molecular mechanisms underlying the intracellular Ca²⁺ overload related to Complex II deficiency, we analysed SERCA and PMCA expression in Complex II-deficient cells by western blotting. SH-SY5Y cells are known to express mostly SERCA2b and PMCA2 isoforms. As shown in Figure 5a, a significant decrease of SERCA2b and PMCA2 expression was observed in SDHA_R554W cells and in SH-SY5Y cells on Complex II inhibition (Figure 5a). Interestingly, a similar result was observed on inhibition of mitochondrial ATP synthase by Oligomycin (Figure 5a).

Figure 5

Complex II deficiency is associated with down-regulation of SERCA2b and PMCA expression. (a) Western-blot analyses of microsomal fraction (SERCA2b and GRP94) and of plasma-membrane fraction (PMCA, and transferrin (TFN)) of control and SDHA_R554W fibroblasts and SH-SY5Y cells left untreated (control) or treated with 300 μM 3-NP, 100 nM Atpenin A5 or 5 μM Oligomycin for 20 h. Quantification of SERCA2b protein level was normalized to GRP94 signal and PMCA to TFN signal and is represented in arbitrary units (AU). (b) SERCA and PMCA expression down-regulation in Complex II-deficient cells is not linked to reduced mRNA levels. Semi-quantitative RT-PCR analysis of SH-SY5Y cells left untreated (control) or treated as in (a). The blots show the results of different cycles of the PCR of SERCA2b (24, 26 and 28 cycles), PMCA2 (32, 34 and 36 cycles) and GAPDH used as control (22, 24, 26 cycles). Quantification of three independent experiments is shown in graphs as mean±S.E.M., non-significant (ns) versus control. Representative blots are shown. (c) Rapid degradation of SERCA2b and PMCA2 on Complex II inhibition by 3-NP (300 μM) or ATP synthase inhibition by Oligomycin (5 μM). SH-SY5Ycells were treated with 100 μM cyclohexemide (CHX) alone or in the presence of 3-NP or Oligomycin for the indicated time points. α-tubulin was used as loading control. Quantification of three independent experiments is shown in graphs as mean±S.E.M., ns, non-significant, (^*) P⩾0.01 versus untreated cells (control) considered as 1. Representative blots are shown. (d) Effect of proteasome inhibitor MG132 on SERCA2b and PMCA2 expression. After 2 h pre-incubation of SH-SY5Y cells with or without MG132 (5 μM), cells were treated for 1 h with 100 μM cycloheximide in the presence or not of 300 μM 3-NP or 5 μM Oligomycin. α-tubulin was used as loading control

SERCA and PMCA down-regulation could be due to reduced transcription level or increased rate of degradation. By using semi-quantitative RT-polymerase chain reaction (PCR), we showed that SERCA2b and PMCA2 transcription level is not altered on Complex II or ATP synthase inhibition (Figure 5b). We then investigated SERCA2b and PMCA2 degradation. SH-SY5Y cells were treated with the protein synthesis inhibitor cyclohexemide (CHX) in the presence or not of 3-NP or Oligomycin. Kinetic analyses revealed a rapid degradation of both SERCA2b and PMCA2 on Complex II or ATP synthase (Figure 5c) inhibition. To determine whether SERCA2b and PMCA are degraded by the proteasome, we treated the cells with the proteasome inhibitor MG132 for 2 h and then applied CHX alone or in the presence of 3-NP and Oligomycin for 1 h. As shown in Figure 5d, treatment of the cells with MG132 prevented SERCA2b and PMCA2 degradation on Complex II or ATP synthase inhibition (Figure 5d).

These data reveal the involvement of the proteasome-dependent degradation pathway in SERCA2b and PMCA2 down-regulation. The result obtained on ATP synthase blockade by Oligomycin may reveal that rapid degradation of SERCA2b and PMCA2 is an ATP-dependent mechanism. Overall, these data suggests that cytosolic and mitochondrial Ca²⁺ overload on Complex II deficiency are a consequence of suppressed cytosolic Ca²⁺ extrusion through the plasma membrane and accumulation in the ER by PMCA and SERCA, respectively.

As mitochondrial Ca²⁺ overload is a major actor triggering mitochondrial structure alteration and swelling, we analysed the mitochondrial structure in the context of Complex II defect. High-resolution imaging revealed no clear mitochondrial structure modification in Complex II-deficient cells as compared with controls (Supplementary Figure S4a). Accordingly, mitochondrial volume was also not modified on Complex II mutation or inhibition in both fibroblasts and SH-SY5Y cells (Supplementary Figure S4b). These data are in accordance with the absence of apoptosis in both models and showed also that increased mitochondrial Ca²⁺ content is not associated with increased mitochondrial biogenesis (as estimated by unchanged volume).

It has been shown that elevated Ca²⁺ signals in micro-domains between mitochondria and ER Ca²⁺ release sites immobilize mitochondria close to the ER, thus increasing Ca²⁺ transfer between the organelles.¹⁹ As we showed that Complex II mutation induced on the one hand an increased ER Ca²⁺ leak and on the other hand an elevation of mitochondrial Ca²⁺ load (Figures 2b, d and 4), we tested whether these modifications are associated with mitochondrial movement inhibition and ER-mitochondria increased contact sites. High-resolution time-lapse imaging of the mitochondrial network revealed a significant reduction of mitochondrial movements on Complex II mutation or inhibition in both fibroblasts and SH-SY5Y cells (Figure 6a; Supplementary Table S10). Accordingly, high-resolution imaging revealed an increase of the extent of association between ER and mitochondria on Complex II deficiency (Figure 6b; Supplementary Table S11).

Figure 6

Mitochondrial motility and ER-mitochondria contact sites are modified on Complex II defect. (a) Mitochondrial movement index (MI) was analysed in the same conditions as in Figures 1a and b by using Mitotracker red dye. Images converted to binary files were sequentially subtracted to generate ΔF. MI corresponds to the ratio of ΔF on total mitochondrial area. Data are presented as MI versus control considered as 1. MI±S.E.M. are shown in Supplementary Table S10. (b) ER-mitochondria contact sites were analysed in cells transfected with erGFP and loaded with MitoTracker Deep Red in the same conditions as in (a). Representative images of ER (erGFP, green), mitochondria (Mito Deep Red, red) and overlay images (Merge) are shown. Graph: quantitative analysis of the colocalization of ER and mitochondria (% of mitochondrial volume in contact with ER).% of ER-mitochondria contact sites±S.E.M. is presented in Supplementary Table S11. (a, b) Data are representative of three independent experiments. ^*P⩾0.001, ^**P<0.001 versus control

We then investigated the impact of Ca²⁺ deregulation on cellular bioenergetics and mitochondrial dysfunction related to Complex II deficiency.

Measurement of total ATP level on total cell homogenates revealed increased total [ATP] in SDHA_R554W fibroblasts and in control fibroblasts treated with Atpenin A5 (∼+15% versus control) (Figure 7a; Supplementary Table S12). Similar results were obtained in SH-SY5Y cells on Complex II inhibition (Figure 7b; Supplementary Table S12). Besides the mitochondrial respiratory chain, glycolysis constitutes a second source of cellular ATP. As we showed a decrease of [ATP]_mit (Figures 1d and e), the increased total [ATP] is expected to be due to glycolysis activation. Indeed, glycolysis inhibition with 2-deoxy-D-glucose (a non-metabolizable analogue of glucose) leads to a decrease of total [ATP] (Figure 7a and b; Supplementary Table S12). In agreement with this result, hyperlactataemia derived from increased glycolysis has been reported in SDHA_R554W fibroblasts.²⁰ It has been recently shown that intracellular Ca²⁺ may activate glycolytic ATP synthesis.²¹ Accordingly, we showed that Ca²⁺ chelation by BAPTA, AM reduced total [ATP] in control and Complex II-deficient cells (Figure 7c; Supplementary Table S13). Interestingly, the extent of ATP reduction was larger in Complex II-deficient cells as compared with controls (72% in control, 60% in SDHA_R554W and 65% in Atpenin A5) (Figure 7c; Supplementary Table S13).

Figure 7

Mitochondrial Ca²⁺ overload is implicated in the modulation of cellular bioenergetics and mitochondrial dysfunction on Complex II deficiency. (a, b) Total ATP level in SDHA_R554W-mutated fibroblasts and in control fibroblasts left untreated (control) or treated with Atpenin A5 as in Figure 1a (a) and in SH-SY5Y cells left untreated (control) or treated as in Figure 1b (b) in the presence or not of 10 mM 2-deoxy-D-glucose (2-DG).% of total ATP versus control±S.E.M. and number of experiments are represented in Supplementary Table S12. (c) Total ATP level in SDHA_R554W-mutated fibroblasts and in control fibroblasts left untreated (control) or treated with 100 nM Atpenin A5 in the presence or not of 5 μM BAPTA, AM for 20 h.Percentage of total ATP versus control±S.E.M. and number of experiments are represented in Supplementary Table S13. (d) Δψ_mit analyses in control and SDHA_R554W fibroblasts left untreated or treated with BAPTA, AM as in (d). Percentage of TMRM intensity versus control±S.E.M. and number of experiments are represented in Supplementary Table S14. (e) Mitochondrial superoxide content quantification performed as in Figure 1c in control and SDHA_R554W fibroblasts left untreated or treated with BAPTA-AM as in (d). The mean of MitoSOX intensity in percentage versus control±S.E.M. and number of experiments are represented in Supplementary Table S15. (d, e) Representative images are shown. (a–e) Data are representative of three independent experiments. ^*P⩾0.001, ^**P<0.001 non-significant (ns) versus control or versus untreated conditions (see Supplementary Tables S13-15)

The data presented in Figure 1 showed an overt mitochondrial pathology on Complex II deficiency. Different laboratories have reported the complex cross-interactions between Δψ_mit, Ca²⁺ and ROS production.¹ It was also shown that mitochondrial depolarization requires and is maintained by the high [Ca²⁺]_cyt.²² We thus investigated the role of cytosolic Ca²⁺ overload on Δψ_mit loss linked to Complex II deficiency. As shown in Figure 7d, application of BAPTA, AM rescued mitochondrial potential in SDHA_R554W cells (Figure 7d; Supplementary Table S14). We assume that mitochondrial potential recovery observed in SDHA_R554W is due to reduced mitochondrial Ca²⁺ uptake as a consequence of removal of [Ca²⁺]_cyt by BAPTA, AM. Indeed, BAPTA, AM application reduced mitochondrial basal and agonist-evoked [Ca²⁺] (data not shown). In accordance with this result, cell treatment with BAPTA, AM also revealed a significant reduction of [O₂^•−] in SDHA_R554W cells (Figure 7e; Supplementary Table S15). These findings suggest that Ca²⁺ overload participated to mitochondrial pathology observed in Complex II-deficient cells.

We finally addressed the question about the consequences of mitochondrial pathology linked to Complex II deficiency on cellular fate. In accordance with earlier reported data,^{23, 24} we showed that complete Complex II inhibition (1 mM 3-NP treatment for 20 h) (Supplementary Figure S1d, right panel) determined cell death (Supplementary Figure S1d, left panel), mitochondrial structure alteration (Supplementary Figure S5a) and reduction of mitochondrial Ca²⁺ load (Supplementary Figure S5b). These processes are known as markers of mitochondrial outer membrane permeabilization and swelling accompanying cell death.

Discussion

By studying fibroblasts isolated from a patient affected by Leigh's syndrome harbouring respiratory chain Complex II (SDHA_R554W) mutation¹³ and fibroblasts and neuronal-derived cells chronically treated with Complex II inhibitors, we reported subcellular Ca²⁺ deregulations occurring before cell death and showed that Complex II deficiency induced a deep regulation of Ca²⁺ signalling-dependent bioenergetics and mitochondrial dysfunction, which control cellular fate.

Only few studies have been focusing on the study of Complex II inhibition and none has provided a global assessment of subcellular Ca²⁺ signalling before cell death. These studies reported that inhibition of Complex II by 3-NP enhances basal and NMDA-induced intracellular [Ca²⁺] in neurons,^{23, 25, 26} potentially leading to Δψ_mit loss, mitochondrial structure alteration, ROS production and cell death.^{23, 26, 27}

The parallel and thorough investigation of subcellular Ca²⁺ and bioenergetics alterations on Complex II mutation or chronic inhibition allowed us to reveal the following pattern (Figure 8): Complex II deficiency is associated with the proteasome-dependent degradation of the two important Ca²⁺ ATPases, SERCA and PMCA thus leading to increase cytosolic Ca²⁺ signals. Increased [Ca²⁺]_cyt and ER Ca²⁺ leak likely caused increased [Ca²⁺]_mit through increased ER-mitochondria contact sites and mitochondrial immobilization. As 99.9% of the total matrix Ca²⁺ content is in bound form,²⁸ we cannot exclude a function for an increased mitochondrial retention capacity in the observed mitochondrial Ca²⁺ overload (e.g. Ca²⁺ binding to cardiolipin, anionic phospholipids, the carboxy-anion-containing metabolites of the Krebs cycle (citrate, oxaloacetate) and inorganic phosphate ²⁸).

Figure 8

Schematic model representing subcellular Ca²⁺ deregulation, mitochondrial dysfunction and bioenergetics outcomes on Complex II mutation or inhibition. Complex II deficiency induces mitochondrial function alteration (reduced Δψ_mit and [ATP]_mit and increased [O₂⁻]), and subcellular Ca²⁺ deregulation. The increased [Ca²⁺]_cyt is mediated by: (i) increased ER Ca²⁺ leak; and (ii) degradation of PMCA and SERCA2b. Mitochondrial Ca²⁺ overload is facilitated by mitochondrial movement inhibition and increased ER-mitochondria contact sites. This leads on the one hand to mitochondrial Ca²⁺ overload and on the other hand to glycolysis-dependent ATP production. RCC: respiratory chain complexes. CLC: Ca²⁺ leak channels. U: uniporter

The situation in Complex II-deficient cells is particular as a large number of studies have reported reduced Ca²⁺ signals in different models of mitochondrial respiratory chain diseases (i.e. in MELAS: mitochondrial encephalopathy, lactic acidosis and stroke-like episodes,¹⁰ and MERRF: myoclonic epilepsy and ragged-red fibres,⁹ and in Complex I-mutated fibroblasts¹¹). To note, our study was conducted in models before cell death thus revealing primary Ca²⁺ deregulation events occurring on Complex II deficiency.

We showed that cytosolic and mitochondrial Ca²⁺ overload linked to Complex II deficiency are likely to be due to rapid degradation of SERCA2b and PMCA2 occurring in a proteasome-dependent manner. This phenomenon is predictably related to the decreased mitochondrial ATP production, as SERCA2b and PMCA2 degradation was also observed on ATP synthase inhibition. Complex II-dependent SERCA2b and PMCA2 degradation occurs despite the slight increase of glycolysis-dependent, cytosolic [ATP] pointing out a primary function of mitochondrial ATP in the degradation process. Indeed, mitochondrial ATP is considered the major cellular ATP source. Alternatively, and not exclusively, the proximity between mitochondria and plasma and/or ER membranes could explain the sensitivity of SERCA2b and PMCA2 to mitochondrial ATP production. Accordingly, the existence of ATP micro-domains beneath the plasma membrane and in ER-mitochondria contact sites has been already shown in some cell models.^{29, 30}

Proteasome-dependent degradation of SERCA2a isoform was reported in an earlier work by Ihara et al.³¹ Ubiquitination of PMCA1 isoform has also been detected after preconditioning ischaemia in rat cortical neuronal cultures.³² As polyubiquitylated proteins are the preferred proteasome 26S substrates, we postulate that SERCA2b and PMCA2 undergo proteasome-dependent degradation after their ubiquitination.

It is known that ATP/Mg²⁺ regulates the 26S proteasome activity, its assembly and stability.³³ However, Geng et al.³⁴ showed that a subset of 26S proteasome is activated as ATP levels decline. Thus, in the context of Complex II deficiency, mitochondrial ATP depletion may have favoured the activation of a subset of 26S proteasome leading to SERCA2b and PMCA2 degradation.

Increased total ATP production has been recently reported in MELAS syndrome linked to Complex I ND5 gene mutation.³⁵ In this study, glycolytic ATP has been shown to be consumed by mitochondria to maintain Δψ_mit. In addition, Ca²⁺-dependent activation of anaerobic glycolysis and increased cytosolic ATP have been recently described during apoptotic cell death.²¹ As our study was conducted before cell death, increased glycolysis-dependent [ATP] may be a compensatory mechanism to the decreased oxidative phosphorylation-dependent ATP production. This phenomenon is Ca²⁺ dependent as we showed that Ca²⁺ chelation by BAPTA, AM reversed glycolytic-dependent ATP production. Thus, the increase of intracellular Ca²⁺ signals may be considered as a first attempt to delay cellular pathology through the activation of glycolytic-dependent ATP production.

Our study revealed a double hint of Ca²⁺ signalling deregulation in Complex II-deficient cells as besides increasing glycolytic ATP production, Ca²⁺ overload favoured mitochondrial pathology. Thus, depending on their energetic needs, cells with Complex II defect may thus undergo a progressive mitochondrial dysfunction, characterized by Δψ_mit loss, Ca²⁺ overload and increased ROS, eventually leading to cell death (Supplementary Figure S5). As a matter of fact, neuronal loss was reported in the brain of the patient harbouring SDHA_R554W mutation.³⁶

Materials and Methods

Chemicals

Culture material was obtained from Invitrogen (SARL, Cergy Pontoise, France), coelenterazine, and fluorescent dyes from Molecular Probes (Invitrogen, SARL), and Atpenin A5 from Coger (eMarketLabo.com). All the other reagents and chemicals were from Sigma-Aldrich (Sigma-Aldrich, SARL, Lyon, France).

Cell culture

Fibroblasts were grown as already described.¹⁷ Human SH-SY5Y neuroblastoma cells (CRL-2266: ATCC, Manassas, VA, USA) were cultured following manufacturer's instructions.

Aequorin measurements

To bypass low transfection efficiency of fibroblasts and SH-SY5Y cells, we used the earlier described adenoviral system expressing cytosolic (AdCMVcytAEQ)-, mitochondrial (AdCMVmitAEQmut and AdCMVmitAEQwt)-, and ER (AdCMVerAEQ)-targeted aequorin probes.^{37, 38, 39} For infection, ∼20000 cells (fibroblasts) and 150000 cells (SH-SY5Y) were spotted on 13-mm coverslips, and placed 24 h later in contact with the appropriate virus (10⁵ infectious particles/ml). Aequorin measurements were performed 48 h post-infection. Specific targeting of adenoviral aequorin probes to desired organelles was verified by immunohistochemistry (Supplementary Figure S6).

MitAEQ and cytAEQ were reconstituted with 5 μM coelenterazine for 2 h in Krebs–Ringer-modified buffer (KRB: 125 mM NaCl, 5 mM KCl, 1 mM Na₃PO₄, 1 mM MgSO₄, 5.5 mM glucose and 20 mM Hepes, pH 7.4) supplemented with 1 mM CaCl₂ (KRB/CaCl₂) at 37°C. Cytosolic and mitochondrial signals were obtained on application of 100 μM Histamine (fibroblasts) or 500 μM Carbachol (SH-SY5Y).

Capacitative Ca²⁺ entry analyses were performed by using cytosolic (AdCMVcytAEQ)-aequorin probe. Cells transferred to the luminometer were treated as specified in the legend of the figure.

For reconstitution with high efficiency of the erAEQ, the luminal [Ca²⁺] of this compartment was first reduced. This was achieved by incubating the cells for 1 h at 4°C in KRB supplemented with 5 μM n-coelenterazine, 1 μM Ca²⁺ ionophore ionomycin and 600 μM EGTA. After this incubation, cells were extensively washed with KRB supplemented with 2% bovine serum albumin before the luminescence measurement was initiated. The ER was refilled by exposing the cells to extracellular 1 mM CaCl₂. After reaching the steady state value, agonist-evoked Ca²⁺ release was initiated by application of agonists. The SERCA blocker, TBuBHQ (2, 5 -Di (tert-butyl) - 1, 4-benzohydroquinone) was used to initiate the release of stored Ca²⁺ and thus to analyse ER Ca²⁺ leak. All aequorin measurements were carried out in a purpose built luminometer. The experiments were terminated by lysis of the cells with 100 μM digitonin in a hypotonic Ca²⁺-rich solution (10 mM CaCl₂ in H₂O), thus discharging the remaining aequorin pool. The light signal was collected and calibrated into [Ca²⁺] values, as earlier described.¹⁸ On the basis of the experimental traces, the maximum rates of ER Ca²⁺ uptake were measured from the first derivative of the ascending refilling phase obtained after CaCl₂ addition, and the maximum rates of ER Ca²⁺ leak were measured from the first derivative of the curve on addition of TBuBHQ using Origin 5 software. Fitting of the curve was performed using Microsoft Excel software.

ATP measurements

Mitochondrial [ATP] was measured after transient transfection of mitochondrial-targeted luciferase probe (mit-LUC). Briefly, cell luminescence was measured in the same luminometer used for aequorin measurements. Cells were perfused with KRB, supplemented with 1 mM CaCl₂ and then shifted to the same buffer supplemented with 20 μM luciferin. At the end of each experiment, digitonin-permeabilized cells were perfused with an intracellular buffer (IB) (140 mM KCl, 5 mM NaCl, 1 mM K₃HPO₄, 2 mM MgSO₄, 20 mM Hepes, pH 7.05 at 37°C) supplemented with 100 μM EGTA, 20 μM luciferin and 5 mM of ATP/MgCl₂. Mitochondrial basal ATP level was normalized to the maximum luciferase signal obtained after cell permeabilization with digitonin and considered as 100%. For total ATP level, we used ATP Bioluminescence Assay kit CL II (Roche, Diagnostics, Meylan, France).

Measurements of cytosolic and mitochondrial basal [Ca²⁺] concentrations

Cells spotted on 24-mm coverslips were loaded with 5 μM Fluo-4, AM (cytosolic Ca²⁺ probe) or 1 μM X-Rhod-1, AM (mitochondrial Ca²⁺ probe) prepared in a KRB/1 mM CaCl₂ at 37°C for 30 min. After a brief washout, time-lapse images were acquired on a Zeiss LSM 510 confocal microscope (Carl Zeiss, S.A.S. LePecq, France). We calibrated Fluo-4, AM and X-Rhod-1, AM probes as recently described by our group¹⁸ and provided absolute basal Ca²⁺ values in the cytosol and the mitochondria. After a control period (providing the basal fluorescent intensity, F), F_max was obtained on application of ionomycin/Ca²⁺ solution (ionomycin: 10 μM and Ca²⁺: 50 mM). Then, the F_min value was obtained on addition of 10 mM EGTA solution in the same bath. We used the equation provided by the manufacturer [Ca²⁺]_free=K_d^*((F-F_min)/(F_max-F)) and the K_d value of 700 nM for X-Rhod-1, AM and of 345 nM for Fluo-4, AM.

Measurements of mitochondrial superoxide concentration

MitoSOX red is a novel fluorogenic dye for highly selective detection of superoxide in the mitochondria of living cells. Cells spotted on 24-mm coverslips were loaded with 5 μM MitoSOX red prepared in a KRB/1 mM CaCl₂ at 37°C for 30 min. After a brief washout, Z-stack images were acquired on a Zeiss LSM 510 confocal microscope (Carl Zeiss). Dye intensity was quantified on maximal projection of Z-stack images after thresholding, using the Zeiss LSM and ImageJ softwares. We presented data as mean fluorescence intensity of oxidized MitoSOX. We assume that MitoSOX probe give a semi-quantitative detection and not an absolute estimation of mitochondrial superoxide value. Importantly, as control, we used the ROS scavenger PDCT to show the specificity of this probe.

Imaging analyses

For the measurement of mitochondrial movements or mitochondrial potential (Δψ_mit), images were acquired on a Zeiss LSM 510 confocal microscope after loading the cells with 10 nM Mitotracker red (movement analysis) or 20 nM TMRM (Δψ_m analysis) in KRB/1 mM CaCl₂ at 37 °C for 30 min.

For mitochondria movement analysis, time series images were taken with a time interval of 5 s between each image. Images were 2D-deconvolved, median filtered, thresholded and clipped to 8-bit binary images using Metamorph (Universal Imaging) software. Using the stack-T-functions/Delta-F-down plugin of the WCIF ImageJ software (http://www.uhnres.utoronto.ca/facilities/wcif/), pixels in each frame were subtracted from the next frame. The resulting images were quantified by measuring the total area of object on the binary and ΔF images by the integrated morphometry analysis function of the Metamorph software. Data were normalized as the ratio of ΔF area values over the total area of the original binary images for each cell. This parameter is referred to as the mitochondrial movement index.¹⁸ Binarization and normalization were applied to avoid artefacts eventually arising from changes of intensity and focal plane.

To obtain normalized TMRM fluorescence signal, time-lapse images were acquired on a Zeiss LSM 510 confocal microscope. Images were acquired before and after application of 10 μM mitochondrial uncoupler FCCP. To show specific TMRM binding, measurements were routinely corrected for residual TMRM fluorescence after full Δψ_mit collapse with FCCP. Basal TMRM signal was normalized to the remaining signal obtained on FCCP application and presented as f/f₀ where f is the mean of basal TMRM intensity along 10 s in the plateau phase and f₀ is the mean of TMRM intensity along the baseline after FCCP application.

For analyses of ER-mitochondria contact sites, cells were transfected with erGFP and loaded with Mitotracker Deep Red. The spectral properties of these two fluorochromes allow specific identification of the two compartments and quantification of contact sites. 16-bit images were acquired with a pixel size equal or lower to the maximal resolution of the 63X objective to allow optimal detection of colocalized pixels on the stage of a Leica SP5 confocal microscope. ER and mitochondria colocalization was calculated as the number of voxels (volume pixels) occupied by both signals (i.e. erGFP and Mitotracker Deep Red) over all voxels occupied by the mitochondria (Mitotracker Deep Red signal) in thresholded images using Imaris software.

Protein preparation and western-blot analyses

Plasma membrane and microsomal fractions were isolated as already described.^{18, 40} SDS-PAGE was then performed. The signal was revealed using ECL Plus western-blot detection reagent (Amersham Biosciences, Orsay, France).

For the experiments shown in Figures 5c and d, membrane fraction obtained as already described¹⁸ was used to reveal both SERCA and PMCA. Protein synthesis was inhibited using 100 μM CHX alone or in the presence of 300 μM 3-NP or 5 μM Oligomycin for the indicated time points. Proteasome-dependent degradation of SERCA2b and PMCA2 was assessed using proteasome inhibitor MG132. After 2 h pre-incubation of SH-SY5Y cells with or without MG132 (5 μM), cells were treated for 1 h with 100 μM CHX in the presence or not of 300 μM 3-NP or 5 μM Oligomycin. Loading control was assessed using α-tubulin, GRP94 as a marker of a resident ER protein and cadherin as a marker of a resident plasma-membrane protein. Similar result was obtained using these loading controls. Only the result of α-tubulin is shown.

RNA extraction and RT-PCR analysis

Total RNA was extracted using RNeasy Mini kit (Qiagen, Courtaboeuf, France); 1 μg of total RNA was used for reverse transcription using High-Capacity cDNA Archive kit (Applied Biosystems, Courtaboeuf, France) with random hexamer primers according to the manufacturer's protocol. The final reaction volume was 25 μl. The reaction tubes were incubated at 25 °C for 10 min, 120 min at 37 °C and 5 min at 85°C. The primers for PCR for PMCA2 transcript were forward: TCCCGACCAGCAGACTCAAGTT, reverse: CGAGTTCTGCTTGAGCGCGG (fragment size: 350 bp). For SERCA 2b transcript, the following primers were used: forward: TTTCCTACAGTGTAAAGAGGACAACC, reverse: AGACCAGAACATATCGCTAAAGTT (fragment size: 514 bp). PMCA and SERCA2b transcript levels were normalized to GAPDH transcript level amplified using forward: CTTCATTGACCTCAACTACATGGT, reverse: CTCAGTGTAGCCCAGGATGCCCTT (fragment size: 727 bp). PCR was performed using 0.5 μl of the reverse transcription mixture in a final volume of 50 μl containing: 1 × Green GoTaq buffer, primers (10 μM each) and Gren Go Taq Hot start polymerase (2.5 U) and dNTP (10 mM each). PCR amplification cycles were run using a thermal cycler (PerkinElmer, Courtaboeuf, France) as following: denaturation at 95°C, 5 min; annealing at 60°C, 1 min; extension at 72°C, 1 min, and heating at 72°C, 10 min. For semi-quantitative PCR, three different amplification cycles were used; see figure legend. Relative signal intensities were quantified using NIH Image J software and are shown in graphs.

Statistical analyses

The results are reported from at least three different experiments. Statistical analyses were performed using ANOVA. Bonferroni's post hoc analysis was subsequently performed on ANOVA results to determine significance. Bonferroni's t-test was calculated versus respective control (Supplementary Tables S1-11) and graphs in Figure 5 or all pairwise (Supplementary Tables S12-15).