ROS (reactive oxygen species) generated by NADPH oxidases play an important role in cellular signal transduction regulating cell proliferation, survival and differentiation. Nox4 (NADPH oxidase 4) induces cellular senescence in human endothelial cells; however, intracellular targets for Nox4 remained elusive. In the present study, we show that Nox4 induces mitochondrial dysfunction in human endothelial cells. Nox4 depletion induced alterations in mitochondrial morphology, stabilized mitochondrial membrane potential and decreased production of H2O2 in mitochondria. High-resolution respirometry in permeabilized cells combined with native PAGE demonstrated that Nox4 specifically inhibits the activity of mitochondrial electron transport chain complex I, and this was associated with a decreased concentration of complex I subunits. These data suggest a new pathway by which sustained Nox4 activity decreases mitochondrial function.
NADPH oxidases are membrane-bound enzymes, which reduce molecular oxygen to the superoxide radical anion by electron transfer. So far, five Nox (NADPH oxidase) genes have been identified in the human genome, termed NOX1–NOX5 . Whereas Nox1–Nox3 require the recruitment of cytosolic regulatory subunits for their activation , Nox4 requires only p22phox as a cofactor, as Nox4–p22phox complexes are constitutively active . Several Nox enzymes, including Nox4, contribute to the control of proliferation and survival of different tumour cell types . Nox4 is ubiquitously expressed in a variety of tissues, such as the kidney, lung, heart and the vascular system . Nox4 contains six transmembrane domains and a cytosolic C-terminal domain, which comprises binding sites for FAD and NADPH. Nox4-derived ROS (reactive oxygen species), in particular H2O2 , were shown to modulate several cellular signal transduction pathways. Signalling through receptor tyrosine kinases, e.g. the EGF (epidermal growth factor) receptor, is positively influenced by Nox4, and this activity has been linked to the ability of Nox4-derived ROS to inactivate protein tyrosine phosphatases, thereby potentiating receptor tyrosine kinase signalling . Nox4 is also required for the activation of several other kinases, such as MAPK (mitogen-activated protein kinase), p38 and JNK (c-Jun N-terminal kinase) , and regulates cellular calcium signalling . Nox4-derived ROS play an important role in controlling cell proliferation and cell survival [8,9]. In the human vasculature, SMCs (smooth muscle cells) are known to express Nox1, Nox2 and Nox4, the latter playing a key role in SMC migration , in combination with Poldip2 (polymerase δ-interacting protein 2) . Human and murine vascular endothelial cells express Nox2 at relatively low levels, but display a high expression of Nox4 , and Nox4 influences cell proliferation and apoptosis in human endothelial cells . The subcellular localization of Nox4 is subject to some debate and may vary between different cell types. Thus Nox4 was found in the endoplasmic reticulum , at focal adhesions , in mitochondria  and in the nucleus . Depletion of Nox4 was found to reduce oxidative damage to nuclear DNA and delay senescence both in human endothelial cells  and in H-RasV12-expressing human thyrocytes . Conceivably, nuclear DNA is damaged directly by Nox4-derived H2O2, e.g. H2O2 produced by Nox4–p22phox complexes in the perinuclear space . In an alternative model, Nox4 localizes to mitochondria where it potentiates mitochondrial production of ROS ; however, molecular mechanisms of Nox4 action in the mitochondria have remained elusive. In the present study, we have addressed this question and identified respiratory chain complex I as a previously unreported functional target for Nox4.
HUVECs (human umbilical vein endothelial cells) were isolated and propagated as described in . Suitable shRNA (short hairpin RNA) constructs directed to Nox4 were transferred into lentiviral vectors (pLKO, Open Biosystems) and used for gene silencing, as described in . Early-passage (defined herein as cells below passage 12) and late-passage cells (defined herein as cells at or after passage 20, where the senescent phenotype was observed) were used.
Protein detection by immunoblot analysis
Cellular protein lysates were prepared in RIPA buffer (50 mM Tris/HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100 and 0.1% SDS) and subjected to standard immunoblotting, using primary antibodies (Invitrogen) against complex I NDUFA9 [NADH dehydrogenase (ubiquinone) 1 α subcomplex 9] subunit, complex II 70 kDa subunit, complex III UQCRC1 (ubiquinol-cytochrome c reductase core protein I) subunit and complex IV subunit IV. Results from three independent experiments were evaluated by densitometry, using an Alpha Innotech FluorChem® HD2 instrument.
Mitochondrial membrane potential (Δψm)
To determine the electric potential of the inner mitochondrial membrane, cells were suspended in culture medium containing JC-1 (5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolylcarbocyanine iodide) fluorescent dye at 0.5 μg/ml , incubated for 30 min at 37°C, washed and resuspended in PBS containing 1 mM pyruvate and 5 mM glucose. Fluorescence was measured in 104 cells using a flow cytometer (FACS Canto II, Becton Dickinson), using a 488 nm laser for excitation and recording emission at wavelength of 530/590 nm. In healthy cells, JC-1 enters the negatively charged mitochondria where it aggregates and fluoresces red. When the mitochondrial potential drops, JC-1 exists as monomers in cytoplasm and fluoresces green. It allows the definition of two populations of cells, with high Δψm (high red and low green fluorescence) and low Δψm (low red and high green fluorescence). Pre-incubation of the cells with the mitochondrial uncoupler FCCP (carbonyl cyanide p-trifluoromethoxyphenylhydrazone) allowed the definition of the cell population with lowered Δψm. All cells with lower green and higher red fluorescence were counted as the cells with high Δψm (as a percentage of the population).
Assessment of ROS
The level of intracellular ROS was measured in situ using flow cytometry in cells stained with the DHE (dihydroethidium) or H2-DCFDA (2′,7′-dichlorodihydrofluorescein diacetate) fluorescent probes (Molecular Probes). The cells (2×105) were suspended in 1 ml of the culture medium containing 20 μM DHE or in 1 ml of HBSS (Hanks buffered salt solution) containing 10 μM H2-DCFDA. The cells were incubated for 30 min at 37°C, washed twice with PBS and resuspended in PBS. The fluorescence was measured using a flow cytometer (FACS Canto II). The level of ROS was estimated as a mean value of DHE or H2-DCFDA fluorescence in 104 cells. Mitochondrial complex I inhibitor rotenone (1 μM) was added along with DHE as a positive control.
Determination of mitochondrial ROS levels
MitoTracker Red CM-H2XRos (Molecular Probes) was used to monitor mitochondrial redox state as recommended by the manufacturer. Briefly, cells were incubated with 100 nM MitoTracker Red CM-H2XRos in culture medium for 15–30 min at 37°C, washed twice with PBS and measured using a flow cytometer (FACS Canto II) at excitation/emission maxima of 579/599 nm.
Determination of mitochondrial and cytosolic H2O2 levels
For the detection of mitochondrial or cytosolic H2O2 levels, the cells were transfected with control, pHyPer-dMito/pLenti6/V5-DEST or pHyPer-Cyto/pLenti6/V5-DEST Gateway lentiviral vector and after expansion analysed using confocal fluorescence microscopy. As a positive control, the cells were incubated for 18 h at 37°C with 200 units/ml CAT-PEG [catalase conjugated to poly(ethylene glycol)].
Analysis of mitochondrial respiratory function
For determination of respiratory activity in intact cells, cells were harvested, and 106 cells were resuspended in 3 ml of culture medium and subjected to high-resolution respirometry using an Oxygraph-2k™ machine (OROBOROS Instruments), as described in . After observing steady-state respiratory flux, the ATP synthase (complex V) inhibitor oligomycin (1 μg/ml) was added, followed by uncoupling of oxidative phosphorylation by stepwise titration of FCCP. Finally, respiration was inhibited by the complex I and complex III inhibitors rotenone (0.5 μM) and antimycin A (2.5 μM) respectively. For the detection of respiratory activity of different OxPhos (mitochondrial oxidative phosphorylation) system complexes, oxygen consumption was measured in permeabilized cells. Approximately 106 cells were resuspended in 3 ml of mitochondrial respiration medium MiRO5 (OROBOROS Instruments) (0.5 mM EGTA, 3 mM MgCl2·6H2O, 60 mM potassium lactobionate, 20 mM taurine, 10 mM KH2PO4, 20 mM Hepes, 110 mM sucrose and 1 g/l BSA, pH 7.1) and subjected to high-resolution respirometry using Oxygraph-2k™ (OROBOROS Instruments). After observing steady-state basal respiratory flux, the cells were permeablized with digitonin (1 mg/ml, 10 μl/106 cells) and the rate of respiration was determined after addition of 2 mM malate and 10 mM glutamate for complex I-coupled respiration, 2.5 mM ADP for complex I state III respiration, 10 mM succinate for complex I and complex II state III respiration, 2.5 μM FCCP for complex I and complex II-uncoupled respiration, 0.5 μM rotenone for complex II-uncoupled respiration and finally 2.5 μM antimycin A, 2 mM ascorbate and 20 μM TMPD (N,N,N′,N′-tetramethyl-p-phenylenediamine) for complex IV-uncoupled respiration. DatLab™ software (OROBOROS Instruments) was used for data analysis. The respirometry data were normalized to CS (citrate synthase) activity, a marker for mitochondrial biomass . Briefly, two portions of 300 μl of the sample were taken from the cell suspension and stirred in the oxygraph chamber before the chamber was closed for recording respiration. Samples were frozen in liquid nitrogen and stored at −80°C. Total cell lysate (100 μl) was added to 900 μl of medium containing 0.1 mM DTNB [5,5′-dithiobis-(2-nitrobenzoic acid)], 0.5 mM oxaloacetate, 50 μM EDTA, 0.31 mM acetyl-CoA, 5 mM triethanolamine hydrochloride and 0.1 M Tris/HCl (pH 8.1). The activity of CS was measured spectrophotometrically at 412 nm and 30°C.
Solubilization of membrane proteins
Mitochondria were solubilized for 30 min on ice with a ratio of 8 g of digitonin (Acros Organics)/g of protein at a final detergent concentration of 1% (w/v), in a buffer containing 30 mM Hepes/KOH (pH 7.4), 150 mM potassium acetate, 10% (v/v) glycerol and 0.5% protease inhibitor cocktail (Sigma, P8340).
BN (Blue native)-PAGE
The solubilized mitochondrial membrane proteins were separated on linear gradient gels at 4°C. Solubilized BHM (bovine heart mitochondria) prepared from tissue and stored at −80°C (3 g of digitonin/g of protein) were loaded as molecular-mass standards for OxPhos (super)complexes. Two types of gels were employed: small gels (10 cm×10.5 cm×0.15 cm) with a linear gradient of total acrylamide concentration from 3 to 13%, overlaid with a 3% stacking gel (Hoefer Mighty Small II system), and large gels (18 cm×16 cm×0.15 cm) with a linear gradient from 4 to 13%, overlaid with a 3.5% stacking gel (Hoefer SE 600 system). The electrophoresis was performed essentially as described in [24,25]. For visualization of protein bands, the gels were stained with CBG (Rotiblue, Roth A152.1). For two-dimensional BN/Tricine-SDS/PAGE, lanes of the large BN gel were excised and processed as described in [24,25] using a stacking gel (containing Tricine-SDS gel buffer) with a total acrylamide concentration of 5% and a separating Tricine-SDS gel with 13% (Hoefer SE 600 system), followed by electrophoresis. For visualization of protein spots, the gels were stained with silver. Mitochondrial proteins were identified by their characteristic migration behaviour in the BN-PAGE, the specific staining pattern of the activity assays and according to previous protein profiling of rat tissue and cell culture mitochondria by peptide mass fingerprint with MALDI–TOF-MS (matrix-assisted laser-desorption ionization–time-of-flight MS) . For in-gel activity assays, BN-PAGE lanes were incubated at room temperature (20°C) in the complex I or complex IV assay solution [50 mM trisodium phosphate, 5 mM reduced cytochrome c, 1.58 mM DAB (diaminobenzidine), pH 7.2] . The NADH dehydrogenase activity of complex I was probed by in-gel formazan precipitation using complex I assay solution. The reactions were stopped in 50% (v/v) methanol and 10% (v/v) acetic acid for 10 min. Images of the gel lanes were scanned using Epson Biostep. For quantification of complex I activity, images of the gel lane were scanned before (t=0 min) and during (t=80 min) incubation, followed by calculation of the relative activity .
Determination of mitochondrial morphology
Cell growth and live confocal imaging were performed in eight-well chambered coverglasses (Nalge Nunc International). The cells were stained for 15 min in 200 μl of culture medium containing 100 μM TMRM (tetramethylrhodamine methyl ester) fluorescent probe (Molecular Probes). After washing cells twice in HBSS, staining of the cells was analysed by confocal microscopy using a microlens-enhanced Nipkow disc-based confocal system UltraView RS (PerkinElmer) mounted on an Olympus IX-70 inverted microscope. Images were acquired using a ×100 oil-immersion objective (Olympus, PlanApo; with a numerical aperture of 1.4). For three-dimensional reconstructions, cells were incubated with TMRM for 20 min at 37°C, and z-stacks were taken with a Zeiss Observer.Z1 inverted microscope in combination with the UltraView Vox Confocal Imaging System (PerkinElmer).
Laser-induced collapse of the mitochondrial potential
The lipophilic cation TMRM accumulates selectively into mitochondria according to their Δψm. Owing to constant laser illumination of TMRM within the mitochondria, ROS are generated which provoke the mPT (mitochondrial permeability transition). Using this model, induction of the mPT is seen by collapse of the Δψm. This model represents a widely published and reliable way to reproducibly induce the loss of Δψm . For this purpose, we stained the cells with TMRM fluorescent probe as described above. Constant laser illumination at a wavelength of 561 nm was performed for 20 s using the aforementioned settings. The TMRM fluorescence corresponding to Δψm was constantly monitored using the UltraView software (PerkinElmer).
RESULTS AND DISCUSSION
Nox4 effects on mitochondrial morphology and stress resistance
Nox4 was knocked down in HUVECs by lentiviral shRNA vectors as described previously . At 3 weeks after infection, both Nox4 mRNA and protein levels were significantly decreased in cells infected with Nox4-specific shRNA vectors, but remained unchanged in cells infected with control shRNA vectors (Supplementary Figure S1 at http://www.biochemj.org/bj/452/bj4520231add.htm). When mitochondrial morphology was analysed by confocal microscopy using the Δψm-specific dye TMRM, a clear difference in mitochondrial morphology was evident. Thus, in Nox4-depleted HUVECs, mitochondria revealed a feature of highly separated, not interconnected, networks compared with shRNA vector-treated cells (Figure 1A). This was confirmed further by TMRM-based imaging using confocal microscopy, followed by three-dimensional reconstruction (Figure 1B). We also used the Volocity® 3D Image Analysis software (PerkinElmer) to score connectivity of mitochondrial structures. For quantification, mitochondrial networks were divided into four categories, depending on their connectivity, measured as the apparent volume of continuous mitochondrial tubes: large (apparent volume >400 μm3), intermediate (apparent volume 10–400 μm3), fragmented (apparent volume 2–10 μm3) and small (apparent volume <2 μm3) (Supplementary Figure S2 at http://www.biochemj.org/bj/452/bj4520231add.htm). Subsequently, the percentage of mitochondrial subnetworks of different size categories was determined by counting. In shRNA vector-treated cells, we regularly detected a single highly connected mitochondrial network, along with a few intermediate mitochondrial networks and a few smaller objects. In Nox4-depleted cells, we consistently failed to detect any highly connected mitochondrial networks; instead, such cells contained several intermediate networks and a larger percentage of smaller objects (Supplementary Figure S2). These findings revealed a significantly higher mitochondrial connectivity in shRNA vector-treated compared with Nox4-depleted cells.
Effects of Nox4 on mitochondrial morphology and stress resistance
To assess effects of Nox4 depletion on mitochondrial stress resistance, Nox4-depleted and shRNA vector-treated HUVECs were laser-irradiated at a wavelength of 561 nm; subsequently, changes in Δψm were monitored using TMRM. In these experiments, Δψm collapsed after 40 s of irradiation in shRNA vector-treated cells (Figure 1C). In contrast, mitochondria of Nox4-depleted cells were still able to maintain high Δψm even after 60 s of irradiation (Figure 1C) with a complete decrease in TMRM fluorescence after 180 s of irradiation. It seems possible that the decreased Δψm is due to photoreduction of nitrite to NO  which is known to impair mitochondrial function . However, we consider it unlikely that NO production contributes to the differences observed between shRNA vector-treated and Nox4-depleted cells, since the same light energy was applied in both cases.
The disconnection of the mitochondrial network in Nox4-depleted cells was confirmed further by the observation that, in shRNA vector-treated cells, the collapse of Δψm occurred in the whole mitochondrial network at once, whereas in Nox4-depleted HUVECs, the decrease in TMRM fluorescence occurred separately in time in different individual mitochondrial subnetworks (Figure 1C, circles). It is known that mitochondrial morphology depends on cell cycle phase and can change in response to various stimuli in a way that is incompletely understood . Mitochondrial fragmentation results from decreased mitochondrial fusion and/or increased mitochondrial fission, and a shift to mitochondrial fragmentation usually correlates with impaired mitochondrial function (reviewed in ), opposite to what we observed here for Nox4-depleted HUVECs. However, this correlation is not absolute, and decreases in mitochondrial fission, e.g. by depletion of the fission factor DLP (dynamin-like protein) in human cells, led to mitochondrial hyperfusion, accompanied by decreased mitochondrial function and decreased proliferation rate . The precise molecular alterations leading to altered morphology in Nox4-knockdown cells remain to be identified.
Nox4 activity induces mitochondrial dysfunction
Expression of Nox4 during extended passaging of human endothelial cells is essential for the timing of cellular senescence, since knocking down Nox4 by lentiviral shRNA vectors significantly delayed the senescent phenotype and extended cellular proliferation capacity . Apparently, prolonged activity of Nox4 in such cells leads to the accumulation of molecular damage, in particular DNA damage, which triggers the senescence response . To address molecular mechanisms by which DNA damage is generated, we compared early-passage cells (defined here as cells below passage 12) and late-passage cells (defined here as cells at or after passage 20, where the senescent phenotype was observed). Δψm was unaltered by Nox4 depletion in early passage after lentiviral infection (Figure 2A); at late passage, Nox4-depleted cells maintained higher Δψm relative to shRNA vector-treated cells (Figure 2B). Cellular redox state, probed with the redox-sensitive dyes DHE (Figures 3A and 3B) and H2-DCFDA (Figure 3C), steadily shifted to pro-oxidant conditions in shRNA vector-treated cells, but not Nox4-depleted cells, at late passage. Similarly, staining with MitoTracker Red CM-H2XRos, probing selectively the redox state in the mitochondria , was significantly decreased in late-passage Nox4-depleted relative to shRNA vector-treated cells (Figure 3D). Whereas collectively these data suggest that Nox4 stimulates ROS production in the mitochondria, other interpretations of these results are possible. Redox-sensitive fluorescence probes such as DHE, H2-DCFDA and MitoTracker Red CM-H2XRos do not allow the unambiguous identification of the ROS that are responsible for the observed change in fluorescence and discrimination between redox-regulated effects and effects of specific radicals, e.g. superoxide anion or H2O2. Hence the precise nature of changes detected by the fluorescence probes used in the present study is not known.
Effects of Nox4 on Δψm
Nox4 induces H2O2 production in mitochondria
Nox4 was found to localize within mitochondria in several cell types [16,34], and partial mitochondrial co-localization was also observed in HUVECs (results not shown), raising the possibility that Nox4 enhances the rate of ROS production in this organelle. As H2O2 is the main product of Nox4 activity , the presence of active Nox4 in mitochondria can be expected to raise mitochondrial H2O2 levels. This was confirmed in cells transfected with lentiviral vectors harbouring a mitochondrially expressed protein probe HyPer, allowing specific detection of mitochondrial H2O2 . In these experiments, late-passage Nox4-depleted cells revealed a significantly decreased HyPer fluorescence in both cytosol and the mitochondria when compared with shRNA vector-treated cells (Figures 3E and 3F). Collectively, these results suggest a significant decrease in H2O2 concentration in the mitochondria of Nox4-depleted cells. Alternatively, the signal output of HyPer sensors used could be influenced by other parameters, e.g. the status of the cellular NAD(P)H-dependent reducing systems. To address further the specific role of H2O2 in this process, cells were treated with CAT-PEG, which significantly decreased HyPer fluorescence (Figure 3G), indicating that Nox4 deficiency indeed leads to lower H2O2 levels in the mitochondria.
To characterize Nox4 effects on mitochondrial function further, respiratory activity was assessed by high-resolution respirometry. At early passage, mitochondrial respiratory activity was slightly decreased in Nox4-depleted relative to control cells (Figure 4A). The basal respiration rate decreased significantly with extended passaging of shRNA vector-treated HUVECs, whereas it was not affected by extended passaging in the absence of Nox4 (Figure 4B, and Supplementary Figure S3 at http://www.biochemj.org/bj/452/bj4520231add.htm). Similarly, state IV (after oligomycin addition) and maximal state III (after FCCP addition) respiration rates were strongly decreased in shRNA vector-treated HUVECs at late passage, but fully preserved in Nox4-depleted cells (Figure 4B). These findings suggest that sustained Nox4 activity decreased both basal respiration and the maximal respiratory capacity of mitochondria.
Nox4 reduces mitochondrial respiratory capacity
Mitochondrial activity is regulated by various extracellular and intracellular signalling pathways, including the insulin/IGF (insulin-like growth factor) pathway [36,37], and signalling through Src  and Akt  kinases. Since Nox4-derived ROS mediate signalling by an overlapping set of kinases, including Src , EGF receptor , MAPK, p38 and JNK , it is conceivable that Nox4 activity triggers specific signalling pathways resulting in decreased mitochondrial activity. However, signalling through JNK, ERK (extracellular-signal-regulated kinase) and p38 was not affected by Nox4 depletion in HUVECs , and we did not observe significant changes in Akt phosphorylation in Nox4-depleted relative to shRNA vector-treated cells (results not shown). We cannot, however, exclude the possibility that altered signalling through other pathways, so far not associated with Nox4 activity, may contribute to the differences in mitochondrial activity observed in Nox4-depleted and shRNA vector-treated cells.
Nox4 activity induces dysfunction of mitochondrial OxPhos complex I
To analyse Nox4 effects on the OxPhos system further, shRNA vector-treated and Nox4-depleted cells were permeabilized and subjected to high-resolution respirometry using specific substrates and inhibitors of different mitochondrial OxPhos complexes. Using this approach, we found significantly increased complex I state III as well as complex I-uncoupled respiratory activity in Nox4-depleted HUVECs when compared with shRNA vector-treated cells at late passage (Figure 5A, and also Supplementary Figure S4 at http://www.biochemj.org/bj/452/bj4520231add.htm). In Nox4-depleted cells, increased protein levels were found for complex I subunits, such as NDUFA9 (Figure 5B), suggesting that Nox4 reduces the abundance of mitochondrial complex I in late-passage HUVECs. Although respirometric data indicated increased complex II respiratory activity in Nox4-depleted cells, complex II protein level was not changed between samples (Figure 5). Neither complex IV activity nor protein levels were changed in Nox4-depleted compared with shRNA vector-treated HUVECs (Figure 5). Similarly, mitochondrial complex III protein levels were not affected by Nox4 depletion (Figure 5B). It is possible that Nox4 affects the synthesis and/or stability of complex I subunits encoded in the mitochondria. Using qRT-PCR (quantitative real-time PCR), we found no significant changes in mRNA concentration of complex I subunits (results not shown), suggesting that transcriptional regulation does not contribute to the altered abundance of complex I subunits in Nox4-depleted cells.
Nox4 reduces mitochondrial OxPhos complex I respiratory activity and protein level
Separation of mitochondrial OxPhos complexes under native conditions revealed a significant increase in both the protein level (Figure 6A) and activity (Figure 6B) of mitochondrial complex I-containing supercomplexes I1III2IV0-1 in Nox4-depleted HUVECs. Accordingly, our data indicate that Nox4 contributes to a decrease in the protein level of several subunits of the mitochondrial supercomplexes I1III2IV0-1 (Figure 6C, red circles). Neither complex IV subunit protein levels (Figure 6C, orange circles) nor complex IV in-gel activity (Figure 6D) were changed in these assays, confirming our observation that Nox4 has no influence on mitochondrial complex IV activity and protein level. Inactivation of complex I-containing supercomplexes I1III2 and I1III2IV in Nox4-expressing cells could reflect molecular damage of such complexes, e.g. by Nox4-derived H2O2. Alternatively, Nox4 may affect the assembly of individual complex I into supercomplexes that have much higher enzymatic activity [26,41]. Although more work is required to establish the detailed mechanism of Nox4-mediated inactivation of mitochondrial complex I, the data indicate that Nox4 activity decreases mitochondrial supercomplexes I1III2IV0-1 protein levels and activity, thereby contributing to mitochondrial dysfunction.
Effects of Nox4 on mitochondrial OxPhos complex I-containing supercomplexes
Several previous studies support a role of Nox4 to modulate mitochondrial function in vascular SMCs , murine cardiomyocytes  and prostate carcinoma cells  respectively; however, molecular mechanisms were not experimentally addressed in these studies. The association of electron transport chain complexes to higher-order structures (referred to as supercomplexes) is essential for mitochondrial function, and oxidative stress is known to disassemble supercomplexes, leading to cell damage and disease . In the present paper, we describe for the first time OxPhos complex I and supercomplexes containing complex I as functional targets of Nox4. Several studies suggest possible pathological implications due to changes in mitochondrial supercomplexes. For example, the absence of NDUFS4 [NADH dehydrogenase (ubiquinone) Fe–S protein 4] protein, an accessory subunit of mitochondrial complex I, results in decreased activity and stability of complex I leading to fatal mitochondrial encephalomyopathy in mice, whereas mutations in the NDUFS4 gene in humans cause a Leigh-like phenotype in humans . Moreover, a deficiency for NDUFA9, the complex I subunit shown in the present study to be targeted by Nox4, has also been described in patients with Leigh syndrome . Whereas these findings raise the possibility that increased Nox4 activity may play a role in Leigh syndrome and related pathologies, more work is required to address this question.
bovine heart mitochondria
catalase conjugated to poly(ethylene glycol)
epidermal growth factor
carbonyl cyanide p-trifluoromethoxyphenylhydrazone
Hanks balanced salt solution
human umbilical vein endothelial cell
insulin-like growth factor
c-Jun N-terminal kinase
mitogen-activated protein kinase
mitochondrial permeability transition
NADH dehydrogenase (ubiquinone) 1 α subcomplex 9
NADH dehydrogenase (ubiquinone) Fe–S protein 4
mitochondrial oxidative phosphorylation
reactive oxygen species
short hairpin RNA
smooth muscle cell
tetramethylrhodamine methyl ester
Rafał Kozieł, Haymo Pircher, Manuela Kratochwil, Barbara Lerner and Martin Hermann performed the experiments and evaluated data. Norbert Dencher, Rafał Kozieł and Pidder Jansen-Dürr designed the experiments and wrote the paper.
We thank Michael Neuhaus for excellent technical assistance.
This work was supported by the Austrian Science Fund (FWF) [grant number P23742] and the European Union [integrated project MiMAGE grant number LSHM-CT-2004- 512020 (to J.P.-D. and N.A.D.)]. M.K. and N.A.D. acknowledge support by the German Federal Ministry for Education and Research (BMBF) through the GerontoMitoSys project [grant number FKZ 0315584] as well as by the Deutsche Forschungsgemeinschaft Graduiertenkolleg (Ph.D. programme) ‘Molecular and cellular reaction on ionizing radiation’.