Mitochondrial composition and function under the control of hypoxia

Abstract

Hypoxia triggers several mechanisms to adapt cells to a low oxygen environment. Mitochondria are major consumers of oxygen and a potential source of reactive oxygen species (ROS). In response to hypoxia they exchange or modify distinct subunits of the respiratory chain and adjust their metabolism, especially lowering the citric acid cycle. Intermediates of the citric acid cycle participate in regulating hypoxia inducible factors (HIF), the key mediators of adaptation to hypoxia. Here we summarize how hypoxia conditions mitochondria with consequences for ROS-production and the HIF-pathway.

Graphical abstract

Highlights

• •Hypoxia provokes changes in mitochondrial morphology, metabolism, and respiration.
• •Hypoxia calls forth changes in redox signaling.
• •HIF-signaling is linked to mitochondrial metabolism and ROS formation.
• •Hypoxia adjusts ETC complex formation, activity, respiration, and ROS formation.

1. Introduction

Hypoxia imposes stress to cells and organisms, which occurs under both, pathological and non-pathological conditions. The lack of oxygen is linked to diseases like cancer, diabetes, or inflammation but constitutes also a challenge for people living at high altitude. Cells within organisms adapt to hypoxia by altering their metabolism. This is facilitated by changes in protein expression occurring at the transcriptional or translational level, mRNA- or protein stability as well as enzyme activity. Sensors and adaptors towards decreased oxygen availability are prolylhydroxylases (PHD), also known as Egl nine homolog 1 proteins (EGLN). Their loss of activity stabilizes the transcription factors hypoxia inducible factors (HIF) under hypoxia. HIFs enter the nucleus to enhance transcription of a variety of target genes, including mitochondrial components.

Major consumers of oxygen in the cell are mitochondria. Consequently, they are severely affected by decreased oxygen availability. Along those lines, hypoxia alters mitochondrial fusion and fission, mitophagy, and oxidative phosphorylation (OXPHOS). OXPHOS is adapted to hypoxia by remodeling the electron transport chain (ETC) as well as the activity of the TCA cycle. The mitochondrial respiratory chain was originally described as flavin- and cytochrome-containing proteins in the inner mitochondrial matrix [1]. This model proposed the four major complexes, i.e. NADH-coenzyme Q reductase (complex I), succinate-coenzyme Q reductase or succinate dehydrogenase (complex II or SDH), ubiquinol cytochrome c reductase (complex III), and cytochrome c oxidase (complex IV) of the respiratory chain randomly dispersed in the matrix, being connected by the redox active enzymes coenzyme Q (CoQ) and cytochrome c [2], [3]. This model was refined with complex I, III, and IV forming supercomplexes that allow an effective electron transport with a minimum of superoxide (O₂^-) production. Nevertheless, ROS (if not specified the term refers to both, superoxide and H₂O₂) from complex I, II, and III appear not only as an accidental escape of electrons from the ETC and their transfer to molecular oxygen, but are now considered as important mediators in physiological cell signaling. ROS production needs to be tightly controlled to avoid its overproduction, provoking damage of mitochondrial and extramitochondrial macromolecules and eliciting cell death [4]. Considering that ROS signaling is linked to the HIF system, it allows anticipating multiple layers of reciprocal interaction. Each ETC complex adapts to hypoxia by replacing distinct proteins, which alter the function of the complex. Using this system, it is not necessary to build an entire new complex to adapt, making these processes fast, reversible, and highly effective. Another adaptive response to hypoxia is the reduction of mitochondrial mass, by mitophagy [5]. This specialized form of autophagy involves factors such as nucleoporin p62, beclin1, microtubule-associated protein1A/1B-light chain 3 (LC3), BCL2/adenovirus E1B 19 kDa protein-interacting protein 3 (BNIP3), or BNIP3 ligant [6], [7]. These components are part of the phosphatidylinositide 3-kinase class III complex-mediated autophagosome formation. The autophagosome docks to lysosomes, followed by fusion and subsequent digestion of trapped macromolecules/organelles. Some proteins eliciting autophagosome formation like BNIP3 are HIF-regulated, suggesting that mitophagy under hypoxia is at least partly HIF-driven [8], [9], [10]. In this review, we address fundamental changes and adaptive mechanisms of mitochondria and mitochondrial ROS production to hypoxia and refer to the crosstalk between mitochondria and the HIF-system (Fig. 1).

Fig. 1.

Hypoxia and mitochondria. Hypoxia acts on mitochondria in multiple ways. It changes the protein composition of the electron transport chain, mitochondrial morphology, and mitochondrial mass. Consequences comprise an altered rate of respiration, changes in the mitochondrial intermediary metabolism, and an adjusted increased or decreased ROS-producing capacity. These changes determine mitochondrial biology with an outreaching impact on cell physiology. Abbreviations: mt: mitochondrial, ROS: reactive oxygen species.

2. Mitochondrial ROS production in brief

Besides NADPH oxidases, mitochondria emerged as a powerful source of ROS. Fig. 2 provides an overview towards ROS-production by the ETC and ROS-detoxifying enzymes. Deficiencies in assembling ETC components, mutations in distinct ETC subunits, or the specific inhibition of the ETC can increase ROS production. The molecular details of ROS production at distinct ETC complexes are elaborated in review articles [11], [12], [13], [14], [15], [16]. Individual targets, experiencing oxidative modifications in response to complex I and III ROS have been identified [17]. Complex I ROS oxidize TCA cycle associated proteins like subunits of pyruvate dehydrogenase or isocitrate dehydrogenase as well as 2-oxoglutarate dehydrogenase complex component E2 [17]. Apparently, complex I ROS are predominantly oxidizing proteins that are localized in the matrix. Complex III ROS in turn oxidize proteins such as voltage-dependent anion channel 3 and mitochondrial import inner membrane translocase subunit TIM50 but also NDUFB10 and NDUFS3, both components of complex I, which are located in the mitochondrial inner membrane [17]. The voltage-dependent anion channel 3 in the outer mitochondrial membrane was proposed as a marker for oxidative stress, occurring in the intermembrane space [18]. Conclusively, distinct sources of ROS modify divergent proteins and thus, alter discrete signaling pathways (Fig. 3).

Fig. 2.

Mitochondrial ROS. Complexes I, II, and III of the respiratory chain are shown to release O₂^- either into the matrix (CI, CII, CIII) or into the intermembrane space (CIII) [11]. To avoid damage of mitochondrial proteins, lipids, or DNA, effective defense mechanisms are in place. Mitochondrial or cytosolic SOD disproportionate O₂^- to molecular oxygen and H₂O₂. In turn, H₂O₂ is efficiently removed by catalase and/or the PRX/GPX system. Nevertheless, ROS may cause PTMs of redox-sensitive proteins, e.g. components of the ETC or transcriptional elements such as Nrf2 or PHDs. ROS cannot only be regarded as damaging agents, as they appear critical for physiological redox-signaling, e.g. during cytokine formation to protect cells from pathogen invasion. Interestingly, cancer cells were shown to protect mitochondria under hypoxic conditions by an increased expression of antioxidative proteins [28]. Thus, cytoplasmic and intermembrane space SOD1 and matrix SOD2 are transcriptionally upregulated under hypoxia, enhancing superoxide detoxification but producing H₂O₂. Although this review concentrates on ROS produced by the ETC, it should be mentioned that NAD-linked dehydrogenases of the TCA cycle such as 2-oxoglutarate dehydrogenase and pyruvate dehydrogenase are potential sites of electron leakage and thus, ROS formation [29]. Abbreviations: ETC: electron transport chain, GPX: glutathione peroxidase, Nrf2: Nuclear factor erythroid derived 2, IMS: mitochondrial intermembrane space, PHD: prolyl hydroxylase, PRX: peroxiredoxin, PTM: posttranslational modification, ROS: reactive oxygen species, SOD: superoxide dismutase, TCA: citric acid cycle.

Fig. 3.

ROS in cellular signaling. The picture of ROS has changed within the last decades from being considered a harmful substance to an important signaling molecule [30]. Especially mitochondrial-derived ROS are mediators of diverse signaling pathways to provoke immune response, elicit antioxidative signaling, initiate DNA damage responses, affect iron homeostasis, stimulate apoptosis, or signal towards cell survival and proliferation [31], [32], [33]. The apparently contradictory roles linked to ROS signaling can be explained by the distinct species, amount, and duration of ROS production as well as antioxidative capacities. For example, a high ROS production rate causes cell death, while moderate ROS levels provoke cytokine production (inflammation). Importantly, spacial and subcellular ROS production by mitochondria and NADPH oxidases create oxidative microenvironments within the cell, affecting distinct proteins or organells. This review focuses on mitochondrial ROS in connection with hypoxic adaptation. Fig. 3 shows several proteins (blue), which are influenced by ROS (inner heptagon), and their role in cellular signaling (outer ring). Abbreviations: Ask1: apoptosis signal-regulating kinase 1, ATM: ataxia telangiectasia mutated, Cyt c: cytochrome c, IRP: iron-related protein, NADPH: nicotinamide adenine dinucleotide phosphate, NDUFB10: NADH dehydrogenase [ubiquinone] 1 beta subcomplex subunit 10, NDUFS3: NADH dehydrogenase [ubiquinone] 1 iron-sulfur protein 3, NFkB: nuclear factor (NF) kappa B, Nrf2: Nuclear factor erythroid derived 2, p53: cellular tumor antigen p53, PI3K: phosphatidylinositol 3-kinase, PTP: protein-tyrosine phosphatase, ROS: reactive oxygen species, Shc: SHC-transforming protein, SOD: superoxide dismutase, TIM50: mitochondrial import inner membrane translocase subunit TIM50, VDAC: voltage-dependent anion channel.

3. Regulation of HIF

HIF proteins are key determinants of a cellular response to hypoxia. Once stabilized, they induce multiple target genes, thereby affecting intermediary metabolism e.g. by increasing the expression of proteins involved in glycolysis and decreasing oxygen-dependent pathways by altering the ETC complex structure and activity. The fundamental concepts of HIF regulation are outlined in Fig. 4.

Fig. 4.

Basics of the HIF-system. Under hypoxic conditions HIF is a master regulator for adaptation to low oxygen content and causes the expression of roughly 400 target genes [34]. The heterodimeric transcription factor consists of the constitutively expressed HIF-1β also known as ARNT and a labile α-subunit, belonging to the basic helix-loop-helix/Per ARNT SIM transcription factor family. Three HIF-α isoforms are known in mammals. HIF-1α and HIF-2α share high sequence homology and regulatory features, containing an oxygen-dependent degradation domain, with two conserved prolyl residues [35]. Under normoxia the α-subunit prolyl residues are continuously hydroxylated by PHD 1–3. This reaction is catalyzed by a set of non-haem Fe(II)- and 2-oxoglutarate-dependent dioxygenase. During catalysis, the splitting of molecular oxygen is coupled to the hydroxylation of HIF, while the oxidative decarboxylation of 2-oxoglutarate gives succinate and CO₂. PHDs are inhibited (Ⱶ) by the lack of oxygen or metabolites such as fumarate or succinate (further information in Fig. 6). Following hydroxylation, pVHL binds and marks HIF for ubiquitination and thus, proteasomal degradation [36], [37], [38]. Under hypoxic conditions, PHDs are inhibited, HIF-α gets stabilized, translocates to the nucleus, dimerizes with its corresponding β-subunit, and recognizes HREs within a promoter or enhancer of distinct target genes to initiate transcription by recruitment of cofactors like p300 or CBP [39], [40]. Inactivation of PHD enzymes may be promoted by ROS e.g. due to oxidation of the central Fe(II) to Fe(III), especially if the antioxidative capacity of the cell is low. It remains a matter of discussion to define concentrations, distinct ROS species and/or secondary metabolites that interfere with PHD activity. Recent studies revealed that PHDs are not only sensitive to iron oxidations but also contain redox-sensitive cysteines, which are subjected to multiple modifications. This highlights diverse possibilities of PHD activity regulation and thus HIF stabilization [41]. Abbreviations: ARNT: Aryl hydrocarbon receptor nuclear translocator, CBP: CREB-binding protein, HIF: hypoxia inducible factor, HRE: hypoxia response element, p300: histone acetyl transferase p300, PHD: prolyl hydroxylase, pVHL: tumor suppressor protein von Hippel-Lindau, ROS: reactive oxygen species.

4. Fusion and fission

Besides changes at the protein level, that affect the quarterly structure of protein aggregates, hypoxia impacts on the mitochondrial morphology, including cristae structure (Fig. 5) [19]. Under normoxia, mitochondria form tubular networks favoring OXPHOS and ATP production. Under hypoxia, mitochondria undergo fission and appear as single organelles, possibly to promote mitophagy, to keep ROS production at a physiological low level, and to maintain integrity by a decrease in respiratory activity.

Fig. 5.

Fusion and fission under hypoxia. A Fission: The dynamic mitochondrial network structure is controlled by fusion and fission. Fission is mediated by Drp1 and Fis1 [42]. Drp1 is a target of diverse posttranslational modifications. AKAP121 phosphorylates Drp1 in a PKA-dependent manner, inhibiting its function. Phosphorylation of Drp1 is indirectly attenuated by Siah2, which is expressed under severe hypoxic conditions, i.e., ischemia thereby promoting fission [43]. Interestingly, Siah2 is also described as a HIF-α-stabilizing protein by regulating PHD stability [44], arguing for a feed-forward amplification loop in regulating fission. Other reports demonstrated that Drp1 is upregulated under hypoxia to provoke migration of U251 cells [45]. Thus, decreased Drp1-mediated fission reduced migration of MDA-mb-231 cells, linking Drp1 to tumor progression and metastasis [46]. Furthermore, a knockdown (Ⱶ) of Drp1 enhanced fusion and reduced fission-mediated ROS production. ROS may foster mitophagy, indicating that the process of mitochondrial fusion is linked to physiological ROS generation [47]. B BNIP3 promotes fission: Opa1 is a dynamine of the inner mitochondrial membrane and interaction partner of Mfn1 that together promote fusion. Binding of the hypoxia-inducible protein BNIP3 to Opa1 promotes Opa1 oligomer disassembly, thereby fostering fission and further on, pushing cells towards apoptosis [48]. Besides, fission can be induced by ROS via the Akt- and NFκB-pathway [49]. However, at present it remains unclear, whether ROS generated by ETC defects induce fission under hypoxia. C HIGD1A promotes fusion: When the HIF-induced protein HIGD1A targets Opa1, it is protected from cleavage, promotes the interaction with Mfn1 and thus, fusion [50]. Logically, a loss of HIGD1A increases fission, cristae disorganization, and reduces cell growth. Thus, HIGD1A and BNIP3 exert opposing functions in fusion and fission under hypoxia. Besides Opa1-cleavage, a loss of ATP, likely due to an impaired ETC under hypoxia, is discussed to induce mitochondrial fission [51]. Abbreviations: AKAP121: A-kinase anchoring protein 121, Akt: RAC-alpha serine/threonine-protein kinase, BNIP3: BCL2/adenovirus E1B 19 kDa protein-interacting protein 3, Drp1: dynamin-related protein 1, ETC: electron transport chain, Fis1: fission 1, HIF: hypoxia inducible factor, HIGD1A: hypoxia inducible gene 1, Mfn1: mitofusin, NFkB: nuclear factor NF kappa B, Opa1: Optic atrophy 1, OXPHOS: oxidative phosphorylation, PHD: prolyl hydroxylase, PKA: protein kinase A ROS: reactive oxygen species, Siah2: E3 ubiquitin-protein ligase Siah2.

5. Crosstalk between HIF and mitochondria

Ongoing research in the areas of metabolomics, assembly, and structure of ETC components provide multiple links between mitochondria and the HIF pathway, as depicted in Fig. 6. HIF gets stabilized by a decreased PHD activity, an enzyme demanding metabolites of the mitochondrial TCA cycle for catalysis. In turn, HIF induces the expression of proteins, which impinge on both, metabolism and structure of mitochondria as described in detail in the following figures.

Fig. 6.

Crosstalk between mitochondria and HIF. A TCA cycle metabolites regulate PHD activity: Succinate, fumarate, pyruvate, oxalacetate, and isocitrate all attenuate PHD activity and thus, favor accumulation of HIF-α subunits [52], [53], [54]. In contrast, 2-oxoglutarate constitutes the cosubstrate for PHD activity. B HIF inhibits the TCA cycle: HIF-1 lowers the activity of the TCA cycle by inducing PDK1 [55], [56], [57]. PDK1 inhibits pyruvate dehydrogenase activity to reduce levels of acetyl CoA, which under normoxia is processed to citrate and fuels the TCA cycle. LDHA increases under hypoxia, followed by enhanced conversion of pyruvate to lactate [57], [58]. In turn, lactate stabilizes HIF, especially in tumor cells that are characterized by a shift from OXPHOS to glycolysis [59], [60], [61]. C Mitochondrial ROS interfere with PHD activity: Mitochondrial respiratory chain complexes are an established source of ROS. ROS, especially H₂O₂ inhibits PHD activity, thereby interfering with HIF-degradation [62]. A rapid increase in ROS production and thus, PHD inhibition was observed within the first minutes of hypoxic incubations, corroborating stabilization of HIF under acute hypoxic conditions [63]. A knockdown (Ⱶ) of GRIM-19 (NDUFA13), a subunit of complex I, induces ROS, with the consequence to stabilize HIF-1 and to enhance autophagy [64], [65]. Also complex II gained some interest. A loss of SDHB, the iron-sulfur cluster containing subunit of complex II, provoked a ROS-dependent stabilization of HIF and stimulated tumorigenesis [64], [65], [66]. A study by Guzy et al. documented that exclusively the loss of SDHB, but no other subunit of complex II, triggered ROS formation with the subsequent ability to stabilize HIF [64]. Also, complex III ROS were identified to stabilize HIF, which enhanced invasiveness in melanoma cells [67]. However, there is some controversy whether complex III ROS are a prerequisite to induce HIF even under hypoxia [68], [69]. A knockdown of the RISP of complex III reduced HIF-1 expression and was claimed to prove the connection between complex III ROS and HIF [70]. Diaz et al. showed that a knockdown of RISP also changes ETC supercomplex formation as well as the abundance of complex I and IV [71]. This leaves the possibility that a knockdown of RISP has consequences for the entire respiratory chain and possibly for mitochondrial metabolism in general, raising the question whether secondary effects explain the observations by Brunelle et al. Hypoxia also upregulates the mitochondrial protease Lon, which degrades mitochondrial proteins as part of the mitochondrial protein quality control system, thereby increasing ROS production followed by apoptosis [72]. Abbreviations: ETC: electron transport chain, HIF: hypoxia inducible factor, HRE: hypoxia response element, LDHA: lactate dehydrogenase A, Lon: Lon protease homolog, NDUFA13: NADH dehydrogenase [ubiquinone] 1 alpha subcomplex subunit 13, OXPHOS: oxidative phosphorylation, PDK1: pyruvate dehydrogenase kinase 1, PHD: prolyl hydroxylase, RISP: Rieske iron-sulfur protein, ROS: reactive oxygen species, SDHB: succinate dehydrogenase [ubiquinone] iron-sulfur subunit, TCA: citric acid cycle.

6. Adaptation of the respiratory chain to hypoxia

ROS generation is a tightly controlled process and has pivotal roles in patho-physiological signaling. Thus, adaption processes have evolved to maintain mitochondrial membrane potential and ATP production, at the same time avoiding uncontrolled ROS production. This becomes obvious when oxygen availability declines. A lack of oxygen induces reductive carboxylation, which was shown to increase ROS production [20], [21], [22], [23]. Moreover, hypoxia causes multiple changes in the composition of ETC complexes. These changes are required to keep mitochondria intact under low oxygen conditions and to prevent excessive ROS formation. Most of the changes in complex composition occur within complex I, a dominant acceptor of electrons, and complex IV, facilitating electron transport to molecular oxygen. Some of these alterations comprise the exchange of subunits within ETC complexes others modify complex structure, while subunit depletion also occurs. Remarkably, the amount of ROS formation or its reduction strongly depends on the subunit being affected. Fig. 7, Fig. 8 summarize distinct modifications of ETC complexes under hypoxia and outline consequences for ROS production. Besides hypoxia, reoxygenation appears as a trigger for ROS production influencing complex I. Thus, complex I subunits like NDUFV1, NDUFV2, NDUFS2, and NDUFS3 are oxidized after reoxygenation provoking a decrease in protein stability [24].

Fig. 7.

Hypoxic adaptation of complex I. A NDUFA4L2 decreases complex I activity: NDUFA4L2 was characterized by Tello et al. as a HIF-1 target gene, which reduced complex I activity under hypoxic conditions [73]. Induction of NDUFA4L2 under hypoxia decreased respiration, prevented an increase in ROS production and preserved the membrane potential. Lai et al. confirmed the HIF-1-dependent induction of NDUFA4L2 and connected its enhanced expression in hepatocellular carcinoma to poor patient survival [74]. Along those lines, a knockdown of NDUFA4L2 suppressed tumor growth as well as metastasis in vivo. B miR-210 decreases ISCU: miR-210 is an established HIF-target and one of the major regulators of metabolism under hypoxia [75], [76]. Furthermore, miR-210 reduces levels of complex I and IV together with their enzyme activity [77]. Thus, it is not surprising that miR-210 plays a central role in regulating ETC complex formation. Importantly, miR-210 attenuates ISCU [78], [79]. ISCU functions as a scaffold for maturation of iron-sulfur containing proteins, such as NDUFS in complex I. A defect in an iron-sulfur containing protein disrupts the electron flow within this complex. Thus, a miR-210-mediated decrease in ISCU induces ROS formation under hypoxia [80]. Both, NDUFA4L and ISCU are HIF- regulated proteins. While NDUFA4L2 is increased, ISCU is decreased by mR-210. Interestingly, they have opposing roles in ROS production. NDUFA4L2 induction decreases ROS, but a reduction of ISCU increases ROS formation. These findings underscore the relevance of the site for ROS production and highlight the difficulties in predicting how modifications or differently composed ETC complexes affect ROS production. C Hypoxia provokes complex I transition: Under hypoxia, respectively during ischemia, complex I changes its conformation from the active (A) to the dormant (D) form [81]. The D-form is considered to be silent in term of ROS formation and thus, does not create a burst in ROS following reoxygenation [82]. The precise mechanism is not fully understood, but the proteins NDUFA9, ND1, and ND3 at the connection of the hydrophilic and the membrane arm of complex I, apparently are involved [83]. After I/R, Cys39 of ND3 is exposed and can be S-nitrosated by different S-nitrosating species, maintaining complex I in the D-form and reducing ROS production upon reoxygenation [84]. Other proteins of the membrane arm, ND4 and ND5, were decreased under hypoxia at mRNA level, decreasing complex I activity [85]. Raising the question whether those subunits might also be involved in complex I transition. Abbreviations: ETC: electron transport chain, HIF: hypoxia inducible factor, HRE: hypoxia response element, I/R: ischemia/reperfusion, ISCU: iron sulfur cluster assembly enzyme, miR: micro RNA, ND1: NADH-ubiquinone oxidoreductase chain 1, ND3: NADH-ubiquinone oxidoreductase chain 3 ND4: NADH-ubiquinone oxidoreductase chain 4, ND5: NADH-ubiquinone oxidoreductase chain 5, NDUFA9: NADH dehydrogenase [ubiquinone] 1 alpha subcomplex subunit 9, NDUFA4L2: NADH dehydrogenase [ubiquinone] 1 alpha subcomplex subunit 4-like 2, ROS: reactive oxygen species.

Fig. 8.

Hypoxic adaptation of complex II and IV. A Complex II: As part of the TCA cycle and the respiratory chain complex II (succinate dehydrogenase) is integral to mitochondrial metabolism. HIF-1, via the induction of miR-210, modulates complex I, as described in Fig. 7[86]. In addition, miR-210 targets SDHD, one of the membrane bound and haem b containing subunits of complex II [87]. In A549 cells transfected with miR-210 SDHD expression decreased, accompanied by a lower complex II activity. Recent studies identified complex II, besides complex I and III, as a generator of ROS, opening the possibility that complex II serves as an additional regulator of ROS under hypoxia [88]. B miR-210 regulates complex IV abundance and activity: Besides ISCU and SDHD, the complex IV subunit COX10 was identified as a miR-210 target either by transfecting HCT116 cells with miR-210 under normoxia or with an antagomir under hypoxia [89]. These experiments linked miR-210 expression to increased ROS production. Moreover, a screening approach also identified NDUFA4 as a miR-210 regulated complex IV subunit [87], [90]. C Subunit exchange in complex IV: Complex IV is regulated by a HIF-dependent exchange of its subunits COX4-1 and COX4-2 under hypoxic conditions [91]. HIF-1 induces COX4-2 together with the Lon protease, which in turn degrades COX4-1. The complex subsequently incorporates COX4-2. This modification optimizes the efficiency of the complex to transfer electrons to molecular oxygen under low oxygen conditions and thus, to decrease ROS formation, maintain ATP production, and to preserve the integrity of complex IV. D HIGD1A enhances complex IV activity: As an early hypoxic inducible and HIF-dependent protein, HIGD1A apparently amalgamates many functions as already described for its function in fusion and fission. Among others, it ensures optimal performance of complex IV [92]. Hayashi et al. showed that HIGD1A binds to complex IV in vivo, while recombinant HIGD1A directly integrated into purified bovine complex IV. Precisely, HIGD1A exerts a protective function. It accumulated around the haem a containing active center of complex IV, a protein cluster, which is responsible for the proton pumping activity of complex IV. E COX5B decreases HIF-1α-dependently: In white adipose tissue COX5B decreases during aging due to HIF-1α mediated transcriptional repression [93]. The lack of COX5B decreases complex IV assembly and activity. Additionally, inhibition of COX5B facilitated lipid accumulation by a reduction in fatty acid oxidation, which in turn enlarged white adipocytes. Abbreviations: COX4: cytochrome c oxidase subunit 4, COX5B: cytochrome c oxidase subunit 5B, COX10: protoheme IX farnesyltransferase, ETC: electron transport chain, HIF: hypoxia inducible factor, HIGD1A: hypoxia inducible gene 1, HRE: hypoxia response element, Lon: Lon protease homolog, ISCU: iron sulfur cluster assembly enzyme, miR: micro RNA, NDUFA4: NADH dehydrogenase [ubiquinone] 1 alpha subcomplex subunit 4, PHD: prolyl hydroxylase, ROS: reactive oxygen species, SDHD: succinate dehydrogenase [ubiquinone] cytochrome b small subunit, TCA: citric acid cycle.

Another form of oxidative stress sensed and handled predominantly by mitochondria is ischemia and reperfusion (I/R). During I/R an oxidative burst occurs, which is generated by enhanced succinate dehydrogenase activity and Ca²⁺ influx, which opens the mitochondrial permeability transition pore. An overview of ROS generating mechanisms and consequences of I/R is depicted in Fig. 9.

Fig. 9.

Ischemia and Reperfusion. A I/R induces ROS: During I/R, the activity of ETC complexes is reduced by an extraordinary increase in ROS, which is a result of succinate accumulation under ischemia. Upon reperfusion, succinate is rapidly oxidized resulting in a burst of ROS, probably mediated by a reverse electron transport from complex II to complex I [94]. The increase of ROS together with an I/R-induced calcium influx into mitochondria opens the mitochondrial PTP, further increasing ROS formation, decreasing ETC activity, and finally provoking cell death [95]. B An open PTP decreases supercomplex formation: An open PTP causes supercomplex degradation after reperfusion. After I/R, supercomplex abundance was reduced but could be recovered either by the PTP inhibitor sanglifehrin or the ROS and electron scavenger XJB-5–131 [95]. C Consequences of mitochondrial ROS: Mitochondrial ROS production during I/R may activate survival mechanisms, i.e. mitogen-activated protein kinases such as JNK, p38, or PI3K, promote apoptosis-associated mechanisms such as PLC1- and PP2A-activation, or induce cell proliferation via ATM [96], [97], [98]. The distinct, in part contradictory, signaling qualities of ROS may be explained by dissimilar sources of ROS. Along those lines, complex I ROS are described to be harmful in cardiac I/R, while complex III ROS appeared to be cardio-protective [99]. Abbreviations: ATM: ataxia telangiectasia, ETC: electron transport chain, I/R: ischemia/reperfusion, JNK: c-Jun N-terminal kinase, p38: mitogen-activated protein kinase p38, PI3K: phosphatidylinositol 3-kinase, PP2A: Serine/threonine-protein phosphatase PP2A, PLC1: 1-phosphatidylinositol 4,5-bisphosphate phosphodiesterase, PTP: mitochondrial permeability transition pore, RET: reverse electron transport, ROS: reactive oxygen species.

7. Conclusion

Research over the last decades highlighted the impact of hypoxia on mitochondria and mitochondrial metabolism, including changes in ROS production and signaling. It is becoming apparent that ETC complexes adapt by posttranslational modification of distinct proteins in individual complexes or by altering subunit composition to keep ROS production at a low, physiological level (Fig. 10). However, cell type specific responses, the presence of various amounts of detoxifying components and ROS-facilitated oxidation in micro-compartments such as the inner membrane space versus the inner mitochondrial membrane or the mitochondrial matrix make it difficult to generalize observations. Another issue of controversy is stabilization of HIF by mitochondrial ROS and the production of ROS under hypoxia itself. Very likely, the different mechanisms that allow adaption of the ETC complexes to hypoxia evolved to keep ROS at a balanced level. Low physiological mitochondrial ROS are needed to guarantee cytokine formation during defense against invading pathogens, at the same time keeping mitochondrial integrity intact, e.g. avoiding destruction of mitochondrial DNA, lipids, or macromolecules [25]. Moreover, adaption may preserve other functions of mitochondria during intermediary metabolism, like amino acid metabolism or the delivery of building blocks needed for cell proliferation under hypoxia. The advantage of exchanging single subunits or modifying distinct proteins only may preserve the structure and composition of the entire ETC and guarantee a fast and high degree of reversibility upon the return to normoxia. What is missing at present, are detailed studies how mitochondria functionally, structurally, and ETC composition wise respond to acute and/or chronic hypoxia as well as the return to normoxia. It remains to be elucidated, whether mitophagy removes mitochondria, which failed to adapt, or diminishes mitochondria independently of their state of adaption as part of the metabolic adjustment to hypoxia. Techniques like complexome profiling coupled to 2-dimensional blue native electrophoresis, metabolomics, and cell respirometry will help to answer these question [26], [27].

Fig. 10.

Conclusion. Current concepts provide evidence for a complex interaction between a hypoxic mitochondrial biology and their ability to produce ROS. Mitochondrial ROS production cannot only be seen as an accidental escape of electrons from the ETC and its transfer to molecular oxygen with the production of superoxide and/or hydrogen peroxide but should also be considered an essential component of physiological cell communication. In a hypoxic environment, often found in tumors, mitochondrial biology is altered. It is generally accepted that OXPHOS is slowed down because reducing equivalents generated in the TCA cycle are decreasing. Nevertheless, hypoxia massively modulates ETC composition and activity. Interestingly, hypoxia, mostly via the HIF-transducing system, produces subtle changes by modifying individual proteins in complex I to IV or by replacing distinct proteins with either more effective (e.g. complex IV) or less efficient (e.g. complex I) variants. Energetically, this is more efficient rather than replacing entire ETC complexes and probably allows a much faster return to basal complex composition when returning to normoxia. Overall, these changes adjust the ability of ROS production under hypoxia, often reducing the ability to generate ROS. It can be speculated that under hypoxia, despite a reduced electron flow through the individual redox centers, mitochondria try to minimize ROS formation in order to lower the risk of macromolecular damage. One can also argue that changes are preventive in order to suppress a burst in ROS once mitochondria return from hypoxia back to normoxia. Along these lines, mitochondrial fission certainly may eliminate damaged mitochondria that often show increased ROS production. Fission observed during hypoxia may be considered as a preventive mechanism to lower ROS production and to exit to mitophagy once environmental conditions get worse or to fuse when mitochondria face normoxia again. We certainly have to learn more about the communicating ability and distinct targets of ROS under hypoxia and explore how a gradual and time-dependent decrease of oxygen affects mitochondrial biology and what happens upon the shift from hypoxia back to normoxia. Abbreviations: ETC: electron transport chain, HIF: hypoxia inducible factor, HIGD1A: hypoxia inducible gene 1, OXPHOS: oxidative phosphorylation, ROS: reactive oxygen species, TCA: citric acid cycle.

Conflict of interest

The authors declare no conflict of interest.

Acknowledgments

This work was supported by Deutsche Forschungsgemeinschaft (SFB 815, project A8).