In addition to respiration, plant mitochondria play a role in the metabolism of diverse amino acids, vitamins and lipids essential for organellar biogenesis and maintenance. The relative importance of these processes depends on the cell type and developmental stage. Studying the responses of plant mitochondria to metal stress implies the need to fractionate the responses in different plant organs. In leaves, the cellular environment of plant mitochondria is rather distinctive as compared to animal cells due to the presence of photosynthesis-derived O₂ and carbohydrates. In plant roots however, a different cellular environment as compared to the leaves will definitely influence the response of mitochondria to excess metals [16,17]. It is noteworthy to mention that plant mitochondria differ from their animal counterparts in that the former function in distinct processes, such as photorespiration, and that they possess unique components (alternative oxidase (AOX), alternative NAD(P)H dehydrogenases (NDs) and uncoupling proteins (UCPs) (Figure 1b) [16,17,20]. Electrons can pass through either the “standard” cytochrome pathway to complex IV (cytochrome c oxidase) (Figure 1a) or through the alternative pathway to the cyanide-insensitive AOX (Figure 1b). Although this alternative route does not contribute to ATP synthesis, AOX is currently considered both a target and regulator of stress responses in plants (Sections 3.1 and 4.3) [21,22].

Under normal conditions, the mitochondrial ETC and ATP synthesis are tightly coupled. However, plants suffering from biotic or abiotic stress often show an over-reduction of electron carriers such as ubiquinone, causing electron leakage from the system. These electrons possess a sufficient amount of free energy to directly reduce molecular O₂, with increased production of ROS such as superoxide (O₂°⁻) and hydrogen peroxide (H₂O₂) as unavoidable byproducts of aerobic metabolism [17,21,23–26].

The known sites of ROS production in the mitochondrial ETC are complexes I and III (for reviews, see [21,23]), where O₂°⁻ is formed and subsequently dismutated to H₂O₂. The uncharged H₂O₂ molecule is able to penetrate membranes and has a longer half-life as compared to O₂°⁻. Both properties make H₂O₂ an ideal candidate for ROS signaling from mitochondria to other organelles [17]. However, H₂O₂ can also react with reduced Fe²⁺ and Cu⁺ in the mitochondrion itself to produce highly toxic hydroxyl radicals and cause oxidative damage to mitochondrial proteins, lipids and DNA [21,25,27].

In mitochondria, transition metals such as Cu, Fe and Zn are crucial for a proper functioning of several enzymes involved in the TCA cycle, electron transport, synthesis of ATP and antioxidative defense [19,27]. However, the presence of free (redox-active) metal cations can initiate or propagate oxidative stress (Table 1). Results of numerous studies investigating plant metal stress responses point toward the generation of oxidative stress and mitochondrial dysfunction as determinants in metal-induced cytotoxicity. In several plant species, metal stress enhances mitochondrial ROS generation mainly by affecting respiratory gas exchange rates [28]. An excess of redox-active metals such as Fe and Cu influences mitochondrial respiration activity (Table 1) [29,30], which could be related to their direct potential to increase mitochondrial ROS production via Fenton and Haber-Weiss reactions. However, also non-redox-active metals negatively affect respiration processes (Table 1). Dixit et al. [31] studied the in vivo effects of chromium (Cr) in pea root mitochondria and observed an inactivated electron transport and enhanced generation of O₂° ⁻. The inhibition of root elongation after exposing pea to Al was attributed to metal-induced ROS production, inhibition of respiration and ATP depletion [32]. However, as adenylates show a drastic decrease in response to phosphate deficiency [33], which commonly coincides with Al exposure due to restricted phosphorus uptake [34], the indicative value of the observed ATP depletion is rather low. Instead, one should determine the ATP/ADP ratio to adequately rationalize the effects of metals on the efficiency of the mitochondrial respiratory chain.

Similar to animals [35], the plant mitochondrial ETC is also considered to be an important target of Cd toxicity [36–39]. Heyno et al. [36] demonstrated a fast Cd-induced stimulation of ROS generation inside root cells, mainly originating from the mitochondrial ETC. Exposure to Cd impairs proper mitochondrial functioning partly by affecting the organellar redox balance as shown by Smiri et al. [39]. Bi et al. [37] confirmed ROS production in mitochondria prior to chloroplasts in combination with altered mitochondrial distribution and mobility patterns in Cd-exposed Arabidopsis thaliana protoplasts (Table 1).

In addition to ROS, plant mitochondria are also able to produce nitric oxide (NO) under low oxygen levels that affects the activity of mitochondrial ETC and matrix enzymes and transcriptionally upregulates AOX. The role NO plays during oxidative signaling falls out of the scope of this review, but was reviewed in detail elsewhere [25,40]. A potential role for NO during metal stress responses in plants is suggested by the results of De Michele et al. [41], who demonstrated the involvement of NO in Cd-induced programmed cell death of Arabidopsis thaliana suspension cultures. Arasimowicz-Jelonek et al. [42] recently discussed the mode of action of NO during Cd stress in plants, with a potential intense relationship between NO and ROS signaling during metal stress as was also shown to occur in plant responses to biotic stress.

Stress-induced over-reduction of ETC components and resulting electron leakage from the system is a principal cause of mitochondrial ROS production. Plant mitochondria contain energy-dissipating systems able to regulate the mitochondrial membrane potential, thereby decreasing mitochondrial ROS production due to ETC over-reduction. However, one must keep in mind that these systems are not able to prevent damage by cytosolar ROS diffusing into the mitochondria [56]. Van Dongen et al. [57] recently reviewed the important role of alternative pathways regulating plant respiration. These pathways ensure metabolic adaptation to hypoxia or altered O₂ availability, which could also indicate their importance during metal stress conditions in plants.

Plant mitochondria contain several proteins present in the vicinity of the “classical” ETC components, which can alleviate the degree of coupling between electron transport and ATP synthesis. They bypass ETC complexes by diverting electrons from the primary cytochrome c pathway, while energy is dissipated as heat. The first enzyme to be discussed in the context of this alternative pathway is AOX, a terminal oxidase accepting electrons directly from ubiquinone and thereby bypassing complexes III and IV (Figure 1b). Since it was first suggested by Purvis and Shewfelt [58], several studies confirmed that AOX prevents mitochondrial oxidative stress (Figure 2). Maxwell et al. [59] demonstrated a direct link between functional AOX levels and mitochondrial ROS production in Arabidopsis cells, thereby confirming the ability of AOX to reduce mitochondrial ROS levels. Exposure to several metals has been shown to affect the alternative respiratory pathway mediated by AOX at different levels (Table 1). Excess Cu induced AOX at transcriptional and protein levels in sycamore cells [30]. In the protist Euglena gracilis, Cd stress led to an increased AOX content and capacity [60]. This was also demonstrated in barley plants, where Cd exposure altered the contribution of the alternative respiratory pathway to total respiration as measured by O₂ uptake in the presence of the AOX inhibitor salicylhydroxamic acid (SHAM). The authors have shown strong effects of high Cd concentrations on total and alternative respiratory rate and suggested the activated alternative respiration to act as a homeostasis mechanism in Cd-stressed root cells [50]. In addition, Cr exposure enhanced the alternative respiratory rate in Salvinia leaves [51]. Pea root tissues exposed to Pb showed a dose-dependent increase in the expression of genes coding for AOX, which was immunologically confirmed [45]. Also, the transcript level of the major AOX isoform in Arabidopsis (AOX1a) increased in Al-stressed protoplasts in a time-dependent way [44]. From these observations, it is clear that AOX activation could be involved in avoiding metal-induced oxidative stress in plants.

In addition to AOX, plant mitochondria contain alternative NDs able to oxidize cytosolic or matrix NADH/NADPH, thereby bypassing complex I and reducing ubiquinone without pumping protons across the inner membrane (Figure 1b) [61]. Although co-regulated expression patterns for several members of the AOX and alternative ND families were detected under multiple stress conditions affecting mitochondrial respiration [62,63], more research is needed to explore the potential role of alternative NDs during metal stress in plants. Results of our work indicate a possible co-regulation of AOX and alternative ND transcription in Cd-exposed Arabidopsis roots and leaves [64]. This suggests the formation of an abridged but functional alternative respiratory chain with electrons from the alternative NDs passing to AOX via ubiquinone (Figure 1b) under conditions comprising the mitochondrial ETC such as Cd exposure.

Further, mitochondrial UCPs catalyze a proton leak that dissipates the proton electrochemical gradient over the inner mitochondrial membrane, thereby shortcutting the ATP synthase complex and thus oxidative phosphorylation (Figure 1b) [65]. A number of observations suggest a role for UCPs mediating cellular tolerance to oxidative stress [66]. Superoxide [67] and lipid peroxidation products [68] stimulate UCP activity in plant mitochondria. The expression of genes coding for UCP is induced by low temperature conditions [69] and treatment with H₂O₂ [70,71]. Overexpressing a gene coding for UCP conferred tolerance to oxidative stress in rice [72] and lack of UCP induced localized oxidative stress in Arabidopsis [73]. In durum wheat mitochondria, it was shown that drought stress induced ROS production, which further promoted UCP activity. Thereby, the authors validated a feedback link between stress-induced mitochondrial ROS production and UCP-mediated inhibition of further ROS accumulation [25,74]. As metals induce mitochondrial oxidative stress and ROS production (Section 2.2), a role for UCP is plausible and further research should be conducted in this area. Promising results were published by Yin et al. [75], who demonstrated increased lipid peroxidation products shown to activate UCP [68] in Al-stressed tobacco roots. Although both AOX and UCP activity confer dissipation of energy as heat, they respond to different environmental stimuli. Rasmusson et al. [76] suggested AOX activity to be involved in the acute response to ETC over-reduction, while UCP could become important during prolonged mitochondrial oxidative stress based on their direct versus indirect effect on the transfer of electrons. In addition, the AOX enzyme can be inhibited by lipid peroxidation products, while UCP is stimulated by O₂°⁻ [63,77], thereby signifying the role both enzymes can play in metal-induced oxidative stress as mentioned above.

Once formed, O₂°⁻ radicals are rapidly dismutated to H₂O₂ via the manganese superoxide dismutase (MnSOD) enzyme present in the mitochondrial matrix [56,78]. It was shown that transition metals such as Cu, Fe and Zn do not affect MnSOD activity in tobacco cell cultures [79]. However, MSD (MnSOD) transcript levels were slightly affected by Cd and Cu exposure in Arabidopsis thaliana [52] and the MnSOD activity in pea root mitochondria increased after Pb exposure [46]. These apparent conflicting results (Table 1) can be attributed to the chosen experimental setup. The observed effects can differ when measured on whole plants versus cell cultures and depending on the used metal concentrations, which greatly hinders accurate interpretation of the available experimental data. However, a role for MnSOD in the adaptation to Al stress is suggested by experimental data using transgenic plants overexpressing MnSOD. Aluminium-induced inhibition of root growth, lipid peroxidation and callose accumulation—a marker for Al injury—decreased in homozygous transgenic overexpressor plants as compared to wildtypes. In conclusion, the authors suggested an improved resistance to Al toxicity mediated by MnSOD in Brassica napus [49].

In order to provide optimal defense against O₂°⁻-derived H₂O₂ production, MnSOD must act in concert with H₂O₂-scavenging components. An important system removing H₂O₂ is the ascorbate-glutathione cycle comprised of four enzymes (ascorbate peroxidase (APX), dehydroascorbate reductase (DHAR), monodehydroascorbate reductase (MDHAR) and glutathione reductase (GR)) and two metabolites (ascorbate (AsA) and glutathione (GSH)) that are regenerated using NADPH equivalents [80]. Jiménez et al. [81] demonstrated the presence of this cycle in plant mitochondria and several cycle enzymes were shown to be dually targeted to mitochondria and chloroplasts [82]. In general, several reports demonstrated that metal stress affects the plant AsA-GSH cycle at both enzyme and metabolite levels (Table 1; Seth et al. [14] and references therein) and mitochondria show an interesting link to this cycle as they harbor the last enzyme in the AsA biosynthesis pathway. In this final step, l-galactono-γ-lactone (GL) is converted to AsA by the membrane-bound l-galactono-γ-lactone dehydrogenase (GLDH), which uses cytochrome c as an electron acceptor and thereby donates electrons to the ETC [25,83]. Zhao et al. [84] have demonstrated a protective role for GL during Cd stress in winter wheat. Application of this AsA precursor lowered Cd-induced H₂O₂ production and increased peroxidase activities [84]. In addition to AsA biosynthesis, it was recently shown that AsA regeneration from dehydroascorbate is also coupled to the plant mitochondrial ETC at the level of complex II, i.e., succinate dehydrogenase [25,85]. As metals strongly affect this complex [50,86], they can also influence the rate of AsA reduction under stress conditions when AsA functions as an important antioxidant.

Plant mitochondria also contain peroxiredoxin, thioredoxin and glutaredoxin enzyme systems capable of scavenging H₂O₂. These enzymes became a topic of great interest over the recent years (for a review see [56]). In Arabidopsis thaliana cell cultures, Cd application increased the mitochondrial peroxiredoxin content (Table 1) [47]. Finkemeier et al. [87] demonstrated a principal role for the mitochondrial peroxiredoxin isoform F (PrxII F) in antioxidant defense and potential redox signaling in plant cells using knockout (KO)-AtPrxII F Arabidopsis seedlings. Under CdCl₂ exposure and after SHAM-administration, the root growth of KO seedlings was more compromised as compared to wildtype seedlings, thereby signifying the involvement of this mitochondrial peroxiredoxin in Cd detoxification [87]. Gelhaye et al. [88] have demonstrated the presence of a mitochondrial thioredoxin isoform in plant mitochondria capable of AOX regulation. This suggests its involvement in metal-induced mitochondrial responses via AOX-mediated modulation of ROS formation. Lastly, Smiri et al. [39] observed a Cd-evoked decrease in glutaredoxin activity measured in burst mitochondria extracts from germinating pea seeds. Although these results point towards the possible involvement of peroxiredoxin, thioredoxin and glutaredoxin enzyme systems in mitochondrial stress responses, the exact mechanisms under metal exposure need to be resolved.

Once ROS are formed in mitochondria of metal-stressed plants, they can either diffuse out of the mitochondria to mediate signaling functions or induce protein, lipid and DNA damage in the organelle itself [21] (Figure 2). Compared to nuclear DNA, the sensitivity of mitochondrial DNA to oxidative stress-induced damage is much higher due to the lack of chromatin organization and lower DNA repair activity [91]. Lipid peroxidation mediated by the interaction between ROS and membrane lipids can distort mitochondrial membrane integrity as shown by Panda et al. [43] in Al-stressed tobacco cells, thereby restricting ETC function. To detoxify lipid peroxidation products, plant mitochondria contain amongst other mechanisms a glutathione-S-transferase (GST) that was strongly upregulated in Cd-exposed plant cells, further supporting its critical role during metal stress (Table 1) [47].

Furthermore, plant mitochondrial proteins are susceptible to metal-catalyzed oxidation, leading to the irreversible formation of reactive carbonyl groups on amino acid side chains and hence reduced protein function [27]. Substantial evidence of both TCA cycle and photorespiration pathway proteins as important targets for ROS or lipid peroxidation products produced under environmental stresses such as drought, high light and metal exposure was summarized by Taylor et al. [92]. Bartoli et al. [93] have demonstrated a higher carbonyl accumulation in the wheat mitochondrial proteome as compared to other ROS-producing organelles such as chloroplasts and peroxisomes during well-irrigated and drought stress conditions. Mitochondrial protein carbonylation is rather selective, with not all proteins evenly susceptible to this process. This was demonstrated by Kristensen et al. [94], who have shown distinct subpopulations of the mitochondrial matrix proteome to be carbonylated after Cu and H₂O₂ treatment. Mitochondrial aconitase seems particularly prone to oxidative damage [63] as it is either carbonylated after in vitro oxidation with Cu and H₂O₂ [94] or fragmented after oxidative stress [55]. In addition, its abundance was decreased in Cd-exposed Arabidopsis thaliana cells [47].

Although enhanced mitochondrial ROS levels may serve as monitors and signal the extent of environmental stress throughout plant cells, they may also lead to oxidative damage and programmed cell death (PCD) when mitochondrial and/or cellular antioxidative defense and repair systems are overwhelmed. Programmed cell death is an active and genetically controlled process essential for growth and development, as well as for adaptation to altered environmental conditions [95]. In animals, the essential role of mitochondria in the signaling pathway transducing specific signals into the execution of cell death is widely accepted [96]. A similar function was suggested for plant mitochondria [97], with H₂O₂ and other ROS as signals modulating plant PCD. The main event in plant PCD is the release of mitochondrial cytochrome c [98]. This was shown to occur in Al-stressed tobacco cells by Panda et al. [43]. In addition, ROS potentially derived from or interfering with mitochondrial functioning could be key mediators of metal-induced cell death as discussed in several research papers. Exposure to Zn induced cell death in rice roots, which could be mediated by ROS derived from the mitochondrial ETC [53]. Recently, Li and Xing [44] investigated possible mechanisms underlying Al-induced PCD in Arabidopsis mesophyll protoplasts using fluorescence techniques to monitor the in vivo behavior of plant mitochondria and caspase-3-like activity. After a quick ROS burst, mitochondrial swelling and loss of the transmembrane potential occurred prior to PCD. Application of AsA prior to Al-exposure slowed down but did not prevent these processes [44]. Mitochondrial O₂°⁻ production—rather than NADPH oxidase-derived extracellular H₂O₂—was shown to be a key event in Cd-induced cell death in tobacco cells [48]. In addition, NO and ROS may be co-involved in Cd-mediated PCD as shown by De Michele et al. [41] and recently discussed by Arasimowicz-Jelonek et al. [42]. Although our insights are currently increasing (Table 1), more research is required to fully unravel the importance of plant mitochondria and the specific mediators involved during metal-induced PCD, which could be a regulated defense response to improve plant survival under metal stress conditions.

Depending on the intensity of the stressor, metals are able to induce mitochondrial damage and/or signaling outside the mitochondria. An altered organellar redox state generates signals that are transmitted to the nucleus in a process called retrograde signaling. This process occurs between mitochondria, chloroplasts and the nucleus and can be mediated by ROS or oxidative stress-induced secondary signals. Recently, Suzuki et al. [99] reviewed the intense relationship between mitochondria and chloroplasts in stress-induced redox signaling throughout the cell. Galvez-Valdivieso et al. [100] summarized the involvement of ROS in chloroplastic retrograde signaling, for which more data are available as compared to mitochondria. Due to their ability to signal across organellar membranes, ROS are considered as key components transducing signals between organelles. The dynamics and specificity of ROS-induced signaling are still questioned, but were recently reviewed by Mittler et al. [101]. They suggest ROS signaling to be a dynamic process occurring within cells between different organelles, as well as over long distances between different cells. In addition, ROS-induced oxidative damage to mitochondrial components may produce secondary signals. Møller and Kristensen [102] reviewed the potential of ROS-mediated protein oxidation in plant mitochondria as a stress indicator. Oxidatively damaged proteins such as those functioning in the TCA cycle and antioxidative defense can either be degraded by proteases or serve as an alarm signal to initiate plant responses at the cellular level [102,103].

To date, no specific components of any mitochondrial retrograde signaling pathway have been identified [99]. However, a mechanism of retrograde signaling is suggested to be involved in plant stress responses to excess Al as AOX transcription is increased [44,104]. Therefore, the involvement of retrograde mechanisms in oxidative stress and/or signaling induced by other metals is plausible and deserves further attention. Recently, Van Aken et al. [105] demonstrated that plant mitochondria respond to a wide variety of abiotic stresses (e.g., salt and heat), chemical inhibitors (e.g., rotenone) and hormones (e.g., abscisic and salicylic acid). They also offered a range of stress-responsive genes as potential targets to study novel mitochondrial retrograde signaling pathways.

The most intensively studied model for retrograde signaling between the mitochondrion and nucleus resulting in acclimation to stress conditions is AOX (Section 3.1, Figure 2). This enzyme could play a pivotal role during metal-induced signaling in plant mitochondria since AOX transcription, protein content and/or activity/capacity are commonly increased upon metal exposure (Table 1). Induction of AOX correlates with an enhanced tolerance to ozone and Tobacco mosaic virus in tobacco plants [20], potentially by altering organellar and/or cellular ROS levels [16] which could also be important during metal stress. Van Aken et al. [22] suggest that AOX is not only a target, but also a regulator of diverse stress responses in plants, mainly based on the results of studies using transgenic plants either with increased or reduced AOX content. The absence of AOX in Arabidopsis resulted in acute sensitivity to combined drought and light stress, confirming a role for AOX in determining the steady-state cellular redox balance [106]. This further suggests the possible involvement of AOX during acclimation to metal exposure in plants. Arabidopsis protoplasts lacking the gene coding for the major isoform AOX1a showed a dramatically decreased viability during Al exposure as compared to wildtype protoplasts [44]. Conversely, AOX1a overexpression enhanced Al tolerance, confirming the protective role of AOX against Al-mediated PCD [44]. The possible function of AOX as a mitochondrial “survival” protein was suggested by Robson and Vanlerberghe [107] in transgenic tobacco cells lacking AOX exposed to H₂O₂, salicylic acid and the protein phosphatase inhibitor cantharidin. It is hypothesized that the survival function of AOX is based on its ability to continuously suppress mitochondrial ROS generation. This further prevents oxidative damage that could otherwise evoke disturbed gene expression and favor PCD. In addition, the maintenance of respiration in stressful conditions by the alternative route also contributes to the hypothesis of AOX as a survival protein [107]. Vanlerberghe et al. [108] recently reviewed the postulated metabolic and physiological roles of the alternative respiratory pathway, with AOX possibly maintaining a homeostatic mitochondrial signal during stress conditions. Interestingly, the AsA biosynthesis capacity increased in isolated mitochondria of plants overexpressing AOX [109], suggesting a link between this energy-dissipating enzyme and mitochondrial and cellular antioxidative metabolism. Further research using transgenic plants exposed to different metals will contribute to our insights into the role of AOX in metal-induced (retrograde) signaling. However, other proteins and/or metabolites should also be studied to fully unravel the mechanisms involved in the acclimation of plants growing on metal-contaminated soils.

In stress conditions compromising the phosphorylating cytochrome respiratory pathway, such as metal exposure, plant mitochondria use their alternative respiratory pathway to maintain the electron flux to O₂. As discussed earlier, this may reduce mitochondrial ROS production by electron leakage from the system, thereby decreasing metal-induced oxidative stress at organellar and possibly cellular levels. In addition, the sustained electron flux allows a continuously operating glycolysis and TCA cycle, thereby ensuring a great metabolic flexibility under stress conditions [33]. This provides an alternative for the non-phosphorylating respiratory bypasses to ameliorate metal stress—next to their proposed role in modulating ROS production. Indeed, an increased TCA cycle flux results in the production of organic acids in the mitochondrial matrix. These metabolites form a direct link between plant mitochondria and acclimation to metal stress as they are associated with metal hyperaccumulation and tolerance in several plant species. Citrate, malate and oxalate have been suggested as key cellular ligands for Cd, Ni and Zn, mediating metal transport through the xylem and vacuolar sequestration of metal-ligand complexes (for reviews see [33,110,111] and references therein). In Lycopersicon esulentum, Cd-induced secretion of oxalate from root apices was found to be associated with Cd resistance [112]. In addition, both citrate and malate were shown to be secreted from root apices of plants exposed to Al, thereby establishing an Al tolerance mechanism [113]. Singh and Chauhan [114] also discussed the potential of organic acids to detoxify absorbed Al in the cytosol, followed by vacuolar sequestration as an internal tolerance mechanism exploited by several Al accumulating plant species. Interestingly, Gray et al. [115] have demonstrated an increased AOX1 transcript level in tobacco cells incubated with organic acids such as citrate, malate and 2-oxoglutarate in a physiologically relevant concentration range. The authors postulate a plant mitochondrial retrograde signaling pathway for the regulation of AOX gene expression based on TCA cycle intermediates, which could function concomitantly with ROS signaling to the nucleus (Figure 2) [115].