Mitochondrial Composition, Function and Stress Response in Plants

Authors

  • Richard P. Jacoby,

    1. ARC Centre of Excellence in Plant Energy Biology
    2. Centre for Comparative Analysis of Biomolecular Networks (CABiN), MCS Building M316, The University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Western Australia, Australia
    Search for more papers by this author
  • Lei Li,

    1. ARC Centre of Excellence in Plant Energy Biology
    2. Centre for Comparative Analysis of Biomolecular Networks (CABiN), MCS Building M316, The University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Western Australia, Australia
    Search for more papers by this author
  • Shaobai Huang,

    1. ARC Centre of Excellence in Plant Energy Biology
    2. Centre for Comparative Analysis of Biomolecular Networks (CABiN), MCS Building M316, The University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Western Australia, Australia
    Search for more papers by this author
  • Chun Pong Lee,

    1. Department of Plant Sciences, University of Oxford, South Parks Road, Oxford OX1 3RB, United Kingdom
    Search for more papers by this author
  • A. Harvey Millar,

    Corresponding author
    1. ARC Centre of Excellence in Plant Energy Biology
    2. Centre for Comparative Analysis of Biomolecular Networks (CABiN), MCS Building M316, The University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Western Australia, Australia
    Search for more papers by this author
  • Nicolas L. Taylor

    1. ARC Centre of Excellence in Plant Energy Biology
    2. Centre for Comparative Analysis of Biomolecular Networks (CABiN), MCS Building M316, The University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Western Australia, Australia
    Search for more papers by this author

E-mail: harvey.millar@uwa.edu.au

Abstract

The primary function of mitochondria is respiration, where catabolism of substrates is coupled to ATP synthesis via oxidative phosphorylation. In plants, mitochondrial composition is relatively complex and flexible and has specific pathways to support photosynthetic processes in illuminated leaves. This review begins with outlining current models of mitochondrial composition in plant cells, with an emphasis upon the assembly of the complexes of the classical electron transport chain (ETC). Next, we focus upon the comparative analysis of mitochondrial function from different tissue types. A prominent theme in the plant mitochondrial literature involves linking mitochondrial composition to environmental stress responses, and this review then gives a detailed outline of how oxidative stress impacts upon the plant mitochondrial proteome with particular attention to the role of transition metals. This is followed by an analysis of the signaling capacity of mitochondrial reactive oxygen species, which studies the transcriptional changes of stress responsive genes as a framework to define specific signals emanating from the mitochondrion. Finally, specific mitochondrial roles during exposure to harsh environments are outlined, with attention paid to mitochondrial delivery of energy and intermediates, mitochondrial support for photosynthesis, and mitochondrial processes operating within root cells that mediate tolerance to anoxia and unfavorable soil chemistries.

  • image

[ A. Harvey Millar (Corresponding author)]

Introduction

The ATP needed for cellular maintenance and growth in organisms comes from respiration. It is the fundamental energy-conserving process that couples the transfer of potential energy from the oxidation of reduced organic matter to high-energy intermediates and heat. In aerobic respiration, which yields the highest efficiency of conversion to high-energy intermediates, mitochondria carry out the final steps to generate the bulk of the ATP through oxidative phosphorylation driven by oxidation of organic acids, to release CO2 and reduce O2 to water. However, mitochondria also play roles in a variety of important cellular processes associated with carbon, nitrogen, phosphorus and sulfur metabolism in plants. In photosynthetic tissues, mitochondria function is indispensable for chloroplast function. Mitochondria are key agents in how plants respond to oxidative stress, and plant mitochondria possess unique respiratory properties to enable these processes. Understanding the control and regulation of the respiratory processes is vital to alter the rate of plant biomass production and to explain plant growth and its variability in different environmental conditions. An exhaustive analysis of all the elements involved in these processes is not possible here, but by using specific examples and recent discoveries we can highlight some key elements in the structure, mechanism and regulation of this process. Firstly, we will consider the composition of the key processes in mitochondrial respiration, our understanding of the mechanism and regulation of the assembly process that builds the machinery and the differential steady-states and roles of these functions in different plant tissue types. Secondly, we will review our understanding of the regulatory changes induced by internal factors (signalling processes, redox control and oxidative stress) and by the environment (salinity, osmotic stress and nutrient deprivation).

Plant Mitochondrial Composition and Assembly

The functional steps of the respiratory apparatus in plant mitochondria can be framed as a sequential set of processes, involving the transport of reduced glycolytic products from the cytosol into the mitochondrion, and then encompassing a series of reactions leading to the release of CO2 and reduction of O2 to water. Firstly, a set of carriers and channels allow substrates and cofactors from the cytosol to enter the mitochondria and also facilitate the release of the products of respiration to the rest of the cell. Next, the tricarboxylic acid cycle (TCAC) and associated enzymes undertake the oxidative decarboxylation of organic acids to reduce NAD(P)+ and FAD+ to NAD(P)H and FADH2, respectively, and to drive substrate level phosphorylation of ADP to ATP. Thirdly, the classical OXPHOS electron transport chain (ETC) couples the oxidation of NAD(P)H and FADH2 to the reduction of O2 and the co-committed translocation of protons used to build an electrical gradient to drive oxidative phosphorylation. Finally, non-phosphorylating bypasses of the electron transport chain, the alternative oxidase and rotenone-insensitive NAD(P)H dehydrogenases, can alter the gearing between the TCA cycle and OXPHOS to facilitate the anaplerotic function of plant mitochondria for organic acid provision to cellular biosynthetic pathways without the full TCA cycle. This machinery and the regulation of pathways to assemble it define the primary functional composition of mitochondria.

Respiratory metabolite transporters

The two membranes of mitochondria have very different permeability properties. The outer membrane allows relatively non-specific transport of small molecules from the cytosol into the inter-membrane space (Mannella 1992; Mannella et al. 2001). The inner membrane contains very selective transporters for small molecules to the matrix space. This allows a complex set of inner membrane carrier functions to have a large influence on the functions of mitochondria (Laloi 1999). Transport across the outer membrane is largely via the voltage dependent anion channels (VDAC), that form β-barrel pores for the movement of respiratory substrates and products up to 1000 Da (Mannella and Tedeschi 1987; Robert et al. 2012). A family of related mitochondrial inner membrane carriers operate for the transport of organic acids, amino acids, inorganic phosphate and nucleotides. Complementation assays have defined the mitochondrial inorganic phosphate carriers (Hamel et al. 2004), and adenine di- and tri-nucleotide carriers (Palmieri et al. 2008) in plants. A general carrier able to transport a variety or both di- and tricarboxylic acids is likely to carry the bulk of organic acid traffic (Picault et al. 2002). Basic amino acid carriers have been identified that transport arginine, ornithine, lysine and histidine (Catoni et al. 2003a; Hoyos et al. 2003), a succinate-fumarate carrier has been identified (Catoni et al. 2003b) and the cofactor NAD+ has a specific carrier in plant mitochondria (Palmieri et al. 2009).

Tricarboxylic acid cycle

The nine enzymes of the TCAC represent the major carbon metabolising machinery present in plant mitochondria. Pyruvate is directly transported across the inner membrane or generated in the matrix from malate by the action of malic enzyme (ME). It is then oxidised by the pyruvate dehydrogenase complex (PDC) to form acetyl-CoA. PDC comprises three enzymes E1 (2-oxo acid dehydrogenase), E2 (acyltransferase) and E3 (lipoamide dehydrogenase) (Guan et al. 1995; Luethy et al. 1995). This complex is regulated by phosphorylation of E1, lowering PDC function in the day and increasing PDC function at night (Thelen et al. 2000). Citrate synthase (CS) catalyses the condensation of acetyl CoA with the dicarboxylate oxaloacetate, yielding citrate and releasing the CoA cofactor (La Cognata et al. 1996). Over-expression of citrate synthase in Arabidopsis enhances growth under low phosphorous conditions due to enhanced citrate excretion from the roots to increase inorganic phosphate availability (Koyama et al. 2000). Citrate is converted to isocitrate via aconitase (ACO), and isocitrate dehydrogenase (IDH) oxidises isocitrate to form 2-oxoglutarate. 2-oxoglutarate dehydrogenase, succinyl CoA ligase, succinate dehydrogenase (complex II see below), fumarase and malate dehydrogenase compete the cycle by reforming oxaloacetate. Recent studies of TCA cycle mutants have shown the wide impact these enzymes have not only in TCA cycle function but as steps for the anaplerotic delivery of organic acids for other processes in plant cells such as photosynthetic performance, plant biomass, photorespiration, nitrogen assimilation and amino acid metabolism, and even stomatal function. Antisense mutants of malate dehydrogenase (MDH) and aconitase in tomato exhibit faster rates of photosynthetic CO2 assimilation rates and higher ascorbate levels (Nunes-Nesi et al. 2005). Antisense of fumarase leads to substantial inhibition of photosynthetic performance and stomatal function (Nunes-Nesi et al. 2007). Knockdown of succinate dehydrogenase can alter stomatal aperture, change nitrogen use efficiency and alter disease signalling (Araujo et al. 2011; Fuentes et al. 2011; Gleason et al. 2011).

Oxidative phosphorylation (OXPHOS) apparatus

The so-called classical ETC is comprised of four large protein complexes (I, II, III, IV) that interact with each other via the small lipid ubiquinone (UQ) and the small protein cytochrome c. Electron flow from NADH to oxygen is coupled to proton translocation out of the matrix, to drive phosphorylation of ADP to form ATP by the F1FO ATP synthase (Complex V).

Complex I (CI) –  NADH-UQ oxidoreductase, catalyses the oxidation of matrix NADH to reduce ubiquinone (UQ) in the inner mitochondrial membrane. In plants, 49 subunits can be resolved from CI by electrophoretic separations (Heazlewood et al. 2003a; Klodmann et al. 2010) (Figure 1). Direct comparisons of CI subunit composition across taxa have revealed divergences between plant CI versus mammalian CI, with eight nuclear encoded plant CI subunits being plant-specific, and several others being common between plants and non-mammalian eukaryotes but absent in mammals (Meyer et al. 2008; Klodmann et al. 2010; Cardol 2011). Studies of mutations of CI subunits have shown that plants can survive without CI due to the activity of alternative NAD(P)H dehydrogenases (see below). Such mutants have a variety of interesting phenotypes including viral infection tolerance, prolonged hydration under water-deficient conditions and altered organic and amino acid concentrations (Dutilleul et al. 2003; Meyer et al. 2009). In Neurospora crassa, CI assembly analysis using radio-labelled pulse chase in mutants has revealed that the matrix and membrane arms assemble independently via separate pathways (Tuschen et al. 1990; Kuffner et al. 1998; Schulte 2001; Videira and Duarte 2002; Mimaki et al. 2012). By using a combination of radio-labelled pulse-chase experiments, in vitro mitochondrial import and monitoring of tagged CI subunits, several CI assembly models have also been proposed in human cells in assembly disturbed systems (Ugalde et al. 2004; Lazarou et al. 2007; Vogel et al. 2007; Mimaki et al. 2012). In plants, controlled dissociation of CI using low concentrations of SDS followed by BN-PAGE and peptide mass spectrometry enabled the visualisation and compositional analysis of 10 subcomplexes between 550 – 85 kDa in size, giving detailed insights into the internal architecture of CI (Klodmann et al. 2010) (Figure 1). Using Arabidopsis CI subunit knockout mutants, 200, 400, 450 and 650 kDa membrane arm subcomplexes have been identified using BN-PAGE and antibodies. It is proposed that these subcomplexes are assembly intermediates during CI formation, which accumulate when specific subunits are absent (Meyer et al. 2011). The first two assembly factors known for CI, CIA40 and CIA84, were discovered in N. crassa (Kuffner et al. 1998). Nine assembly factors including NDUFAF2/B17.2, NDUFAF1/CIA30, C20orf7, C80orf38, NDUFAF4, NDUFAF3, NUBPL, FOXRED1 and ACAD9 have been found in humans and deficiency can impair CI assembly and lead to clinical phenotypes in patients (Nouws et al. 2012). Little is known about assembly factors in plants, with only L-galactono-1,4-lactone dehydrogenase (GLDH) described as a potential assembly factor in Arabidopsis (Pineau et al. 2008) (Figure 1). Given the conserved core subunits but divergence of accessory subunits amongst eukaryotes, it is not yet clear whether the accessory subunits play different roles in complex I assembly and thus whether or not the assembly of CI follows the same pathway in different organisms.

Figure 1.

Composition and assembly of OX PHOS protein complexes in plant mitochondria. 

Numerous investigations have given insights into the composition and assembly of the large, multi-subunit complexes that constitute the classical electron transport chain (ETC). This figure presents current knowledge of how ETC complex assembly is sequentially organized through intermediate subcomplexes, and the assembly factors that mediate these processes.

Complex II (CII) –  Succinate dehydrogenase, is an enzyme of both the TCAC and the respiratory ETC. In all organisms, it is made from four core subunits: a flavoprotein (SDHI), an iron-sulphur subunit (SDH2) and two membrane anchor subunits (SDH3 and SDH4). Purification of the complex using BN-PAGE has revealed the common core subunits, but also four proteins of unknown function that co-migrate with the complex (Eubel et al. 2003; Millar et al. 2004a) (Figure 1). In Arabidopsis, all SDH subunits are encoded in the nuclear genome. Knockout mutants of the SDH1 gene are embryo lethal (Leon et al. 2007), but knockdown of SDH1 and SDH2 lead to phenotypes associated with altered stomatal aperture, altered mitochondrial ROS production and altered nitrogen use efficiency (Fuentes et al. 2011; Gleason et al. 2011). Several proteins assisting CII assembly have been described in yeast and mammalian cells, but only SDHAF1 and SDHAF2 are considered to be real assembly factors that directly and specifically aid CII assembly (Ghezzi et al. 2009; Hao et al. 2009; Rutter et al. 2010). In Arabidopsis, knockdown of the SDHAF2 homolog lowers SDH assembly and markedly reduces root growth (Huang et al. 2012) (Figure 1).

Complex III (CIII) –  Ubiquinone-cytochrome c oxidoreductase, contains 10 subunits including the bifunctional core proteins that act both in CIII function and as the matrix processing peptidase, removing presequences from imported matrix proteins (Figure 1). Only one subunit of this complex, cytochrome b, is encoded by the plant mitochondrial genome (Unseld et al. 1997), with the remaining nine all encoded by the nuclear genome. In BN-PAGE separations from Arabidopsis mitochondria, all of these subunits have been identified and linked back to a set of mostly single copy genes (Werhahn and Braun 2002; Meyer et al. 2008). Yeast provides an ideal model system to study CIII, due to its ability to survive by fermentation in the absence of the complex, making gene knock-out and mutagenesis possible (Smith et al. 2012). CIII assembly follows a modular assembly model including, early core subcomplex, late core subcomplex and a dimeric CIII states (Smith et al. 2012). There have been 13 assembly factors implicated in aiding the different stages of CIII assembly in yeast. Two of these, BCS1L and TTC19, were also found to have functional homologs in mammalian CIII assembly (Diaz et al. 2011; Smith et al. 2012). Little is known about CIII assembly or functional assembly factors in plants.

Complex IV (CIV) –  Cytochrome c oxidase, is the terminal oxidase of the classical ETC. Purification of CIV in plants originally found only seven or eight subunits (Peiffer et al. 1990), but more recently, a CIV complex containing 14 protein bands was separated from Arabidopsis (Millar et al. 2004a) (Figure 1). Eight proteins homologous to known CIV subunits from other organisms, together with a further six proteins that may represent plant specific CIV subunits, were identified. Analysis of human CIV via BN-PAGE separation has revealed an assembly pathway characterized by the sequential incorporation of CIV subunits, initiated by subunit 1 and subsequently progressing through several discrete assembly intermediates (Barrientos et al. 2009). Studies in yeast have revealed over 40 assembly factors that aid different stages of CIV assembly, but only a few homologs for these factors have been defined in humans (Barrientos et al. 2009; Diaz et al. 2011). A plant homolog of yeast assembly factor COX19 has been studied and found capable of complementing the yeast cox19 null mutant and might play a role in the biogenesis of plant cytochrome c oxidase to replace damaged forms of the enzyme (Attallah et al. 2007) (Figure 1). However, it seems evident that our knowledge about the assembly of CIV in plants is still incomplete.

Complex V (CV) –  ATP synthase is a membrane-bound F1F0 type H+-ATP synthase that catalyses the terminal step in oxidative phosphorylation through which ATP is produced. It is composed of a hydrophilic F1 component which catalyses ATP formation and protrudes into the matrix and a hydrophobic F0 component which channels protons through the membrane while also anchoring the whole complex to the mitochondrial inner membrane (Senior 1990; Hamasur and Glaser 1992; Velours and Arselin 2000; Heazlewood et al. 2003b). The general structure and the core subunits of the enzyme are highly conserved in both prokaryotic and eukaryotic organisms (Millar et al. 2011). In plant, most of mitochondrial F1 ATP synthase subunits are encoded in the nucleus and translated in the cytosol before being imported into the mitochondria (β, γ, δ and ɛ), while most of the F0 subunits are encoded in the plant mitochondrial genome and translated in the mitochondrial matrix (a, b, c and A6L) (Jansch et al. 1996; Heazlewood et al. 2003c; Sabar et al. 2003; Sabar et al. 2005) (Figure 1). In plants, the F1α subunit is encoded in the mitochondrial genome in most species. Alterations of mitochondrial-encoded subunits of the F1F0-ATP synthase are frequently associated with cytoplasmic male sterility (CMS) in plants, presumably due to the high ATP demand of floral tissues (Xu et al. 2008). While knockouts of ATP synthase core subunits are lethal in plants, inducible knockdown with a dexamethasone-inducible promoter has enabled investigations into the tissue-specific phenotypes incurred by slowing the rates of mitochondrial ADP:ATP cycling across a range of developmental stages (Robison et al. 2009). Induction of the knockdown during germination in the light leads to seedling lethality. Other phenotypes include the stunting of dark-grown (etiolated) seedlings, downward curling or wavy-edged leaf margins of light-grown plants, and ball-shaped unexpanded flowers (Robison et al. 2009), highlighting the high energetic demand of key growth stages.

The subunits that form the F1 component are kept in tight stoichiometry in prokaryotic and eukaryotic organisms through regulation of the assembly process (Senior 1990; Hamasur and Glaser 1992; Velours and Arselin 2000; Li et al. 2012). Models of yeast mitochondrial F1F0 ATP synthase assembly involve two separate but coordinately regulated pathways, where two separate subcomplexes are assembled in parallel, before converging to form functional F1F0 (Rak and Tzagoloff 2009; Rak et al. 2011). Recent research in Arabidopsis using progressive 15N labeling has measured differential rates of turnover between different subpopulations of the F1 subcomplex. Intriguingly, the same subunits of F1 can exhibit faster or slower turnover rates depending upon the intra-mitochondrial localization they are found in, or upon the quaternary structure of the F1 subcomplex that they constitute. For instance, subunits of F1 that were detected within the matrix-localized and membrane-associated F1 subcomplexes both exhibited faster turnover rates compared to those same subunits detected within the intact, membrane-spanning F1F0 complex (Li et al. 2012). The proposed assembly model for plant CV comprises three steps, the first being the formation of a rapidly turned over F1 subcomplex in the matrix, then an intermediate stage where F1 associates with the inner membrane and still turns over at a fast rate, and then a final unison of F1 with FO to form functional CV (Li et al. 2012) (Figure 1). This model of CV assembly was corroborated by in vitro import assays where radiolabelled CV subunits were incubated with isolated mitochondria, and assembly intermediates visualised by scintillation counting of BN-PAGE separations (Li et al. 2012). ATP synthase assembly factors including Atp10, Atp11, Atp12, Atp22, Atp23 and Fmc1 have been discovered in yeast (Pickova et al. 2005; Osman et al. 2007). Atp11, Atp12 and Fmc1 mediate the formation of the F1 subcomplex while Atp10, Atp22 and Atp23 are essential for the formation of F0 (Pickova et al. 2005; Osman et al. 2007). Hsp60 and Hsp70 also contribute to efficient CV assembly (Osman et al. 2007). A phylogenetic analysis of ATP synthase assembly factors has found Atp11 and Atp12 are preserved in almost all eukaryotic organisms, including plants, while the other assembly factors show evidence of divergent evolution across taxa (Pickova et al. 2005) (Figure 1). However, a detailed study of the presence and conservation of CV assembly factors across sequenced plant genomes has not been undertaken to our knowledge.

Alternative electron transport pathways

In addition to the classical OXPHOS machinery, plant mitochondria contain non-phosphorylating respiratory bypasses of electron transport and of proton-coupled ATP synthesis. These pathways were first identified by the ability of plant mitochondria to respire in the presence of cyanide and rotenone, potent inhibitors of CIV and CI, respectively, and to exhibit natively uncoupled respiration in the absence of an ADP source.

Alternative oxidase (AOX) –  The cyanide-insensitive respiration is catalysed by the alternative oxidase (AOX), a diiron quinol oxidase that branches from the respiratory chain at UQ and reduces oxygen to water without proton translocation. AOX appears to play an antioxidant role in plant mitochondria, is actively induced by oxidative stress (Van Aken et al. 2009) and the different genes for the oxidase have been shown to be both tissue- and development-specific in their expression patterns (Saisho et al. 2001; Thirkettle-Watts et al. 2003). Knockout of AOX leads to anthocyanin and ROS accumulation in the leaves under the combination of high light and drought stress (Giraud et al. 2008).

Alternative NADH dehydrogenases –  These type II NAD(P)H dehydrogenases are found on both sides of the inner mitochondrial membrane. External or cytosolic NADH and NADPH can be oxidised via these dehydrogenases which are insensitive to the CI inhibitor rotenone. These pathways operate without the translocation of protons (Finnegan et al. 2004; Rasmusson et al. 2004). The Arabidopsis genome contains seven genes encoding these Type II NAD(P)H dehydrogenases (Michalecka et al. 2003; Moore et al. 2003), falling into three subgroups: Atnda (1 and 2), Atndb (1 – 4) and Atndc1. A further complication in Arabidopsis is the dual localisation of several of the alternative dehydrogenases in subcellular compartments other than mitochondria (Carrie et al. 2008).

Uncoupling proteins (UCPs) –  UCPs are members of the mitochondrial carrier family of proteins and have been the focus of considerable study as pathways for non-phosphorylating/uncoupled respiration by virtue of their ability to transport H+ back across the inner membrane, dissipating the electrical potential built by the ETC. UCPs can be activated by reactive oxygen species (ROS) and this effect may indicate an important biochemical control mechanism for the engagement of this pathway in vivo (Considine et al. 2003). Analysis of knockouts of UCP (AtUCP1) showed that its absence led to localized oxidative stress but did not impair the ability of the plant to withstand a wide range of abiotic stresses. However, knockout of UCP1 limited the photorespiration rate of plants and reduced the photosynthetic carbon assimilation rate (Sweetlove et al. 2006). This suggests that the main role of UCP1 in leaves is to maintain the redox poise of the mitochondrial ETC to facilitate photosynthesis (Sweetlove et al. 2006).

Plant Mitochondrial Composition Variation in Different Tissues

In response to alterations in cellular metabolic and energy demands, mitochondria often undergo changes in their morphology and respiratory capacity by regulating the composition and abundance of the protein machinery that has been outlined above. In this way, mitochondria are dynamically tuned to meet the specific need for energy in different tissue types or in response to the environment. These differences, or heterogeneity of mitochondria, have been observed through reports of tissue selective phenotypes of mutants, through evidence of transcriptional programming of mitochondrial functions and through examples of steady-state differences in organelle composition and post-translational differentiation of mitochondrial function in different tissues.

Mutations of nuclear genes encoding mitochondrial proteins have been reported to yield organ-specific plant phenotypes in a number of recent reverse-genetics studies. These include delayed development and flowering by loss of PPR proteins (de Longevialle et al. 2007; Sosso et al. 2012), altered leaf morphology and/or photosynthetic capacity by loss of CI, CII or mitochondrial malate dehydrogenase (Meyer et al. 2009; Tomaz et al. 2010; Fuentes et al. 2011), and alteration in root morphology and respiratory rate and inhibition of stomatal function by loss of fumarase (Nunes-Nesi et al. 2007; van der Merwe et al. 2009). These observed phenotypes could be explained by: (i) the inability of mitochondria to meet energy demands in a particular tissue, and/or (ii) the incompatibility of a mutation in the mitochondrial proteome that requires the expression of particular isoforms of proteins, the assembly of particular complexes, and/or the stoichiometry of different components in pathways for tissue-specific functions.

A number of nuclear-encoded mitochondrial respiratory components have been shown to be co-regulated in various vegetative and reproductive organs at the transcriptional level (Gonzalez et al. 2007; Lee et al. 2011). Promoter analyses of the co-regulated components have uncovered common site II motifs in the proximal promoter of these genes that may direct organ-specific, metabolic, environmental and developmental responses (Welchen and Gonzalez 2005; Gonzalez et al. 2007). Analysis of broader functional categories of genes has revealed that components of CI and CV are constitutively expressed, whereas genes encoding for mitochondrial photorespiratory machinery and heat shock proteins are expressed selectively across the plant tissues examined (Lee et al. 2011). While there are a number of examples where there is a strong correlation between transcript abundance and protein abundance/activity across the tissues examined, there are many cases that show otherwise, notably for NAD-malic enzyme, aldehyde dehydrogenase and thioredoxin reductase (Lee et al. 2012). Therefore, caution has to be taken when interpreting tissue-specific differences in the activity of enzymatic steps based on differences in transcript data alone.

To analyze the specialized role of mitochondria during plant development, extensive mitochondrial proteomic comparisons of vegetative (cell culture, root and shoot) and reproductive (silique, stem and flower) phases of development have been recently reported in Arabidopsis (Lee et al. 2012). Using differential 2-D gel electrophoresis, a total of 83 non-redundant proteins consisting of components of the TCA cycle and photorespiration as well as enzymes that depend on the supply of intermediates from these metabolic pathways were identified. While the abundance of individual subunits in the ETC generally remains unchanged across the vegetative tissue types compared, the respiratory capacity alters depending on the substrate choice and/or availability of the substrate in that particular tissue/cell type (Lee et al. 2008; Lee et al. 2011). Determining differences in the abundance of a protein can allow prediction of the degree of variation in metabolic flux between different organ/cell types (Johnson et al. 2007). By mapping these changes on a predesigned scheme of mitochondrial metabolism, the specific enzymatic steps which are regulated due to tissue specialization can be pinpointed (Figure 2). Functional analysis of the vegetative organs/cell reveals specific differences in the central carbon metabolism e.g. shoot mitochondria has a specialized role in glycine cleavage via photorespiration, cell culture mitochondria mainly utilize citrate from the TCA cycle and peroxisomal β-oxidation to drive the decarboxylating reactions of the TCA cycle and fuel ATP formation, while root mitochondria have a higher capacity for converting 2-oxoglutarate into fumarate for energy production via CII (Figure 2). Mitochondria from organs in the reproductive phase tend to have a specialized role in metabolism other than the TCA cycle, such as the maintenance of mitochondrial redox environment in flowers, and nitrogen (glutamate) metabolism in stems. These mitochondrial specializations generally coincide with the main physiological role of each corresponding tissue type. For example, the up-regulation of malate dehydrogenase in the shoot provides evidence for its role in regulating the redox poise which is crucial for mediating photosynthesis and respiration in the light (Hanning and Heldt 1993; Raghavendra and Padmasree 2003; Tomaz et al. 2010).

Figure 2.

Heterogeneity of mitochondrial protein composition in key areas of metabolism. 

Mitochondrial metabolism is tuned to perform specialized roles across a range of plant tissues by modulations to protein abundance and enzyme activity. GDC, Glycine decarboxylase; SHMT, Serine hydroxymethyltransferase; Trx, Thioredoxin; NTR, NADPH thioredoxin reductase; ARG1, Arginase 1; ARG2, Arginase 2; AlaAT, Alanine aminotransferase; PDC, Pyruvate dehydrogenase; CS, Citrate synthase; ACON, Aconitase; IDH, Isocitrate dehydrogenase; GDH, Glutamate dehydrogenase; AspAT, Aspartate aminotransferase; OGDC, 2-Oxoglutarate dehydrogenase; S-CoA, Succinyl-CoA synthetase; SDH, Succinate dehydrogenase; Fum, Fumarase; MDH, Malate dehydrogenase.

Most mitochondrial proteomic studies have focus on the functional impact of tissue-specific nuclear-regulated transcriptional and post-transcriptional events by comparing the total abundance of a particular protein on the gel. While the abundance of a protein sometimes correlates with its maximal catalytic activity, this approximation cannot be applied to all enzymes, due to differential abundance of isoforms of certain mitochondrial proteins across tissues, as these differences in primary amino acid sequence could manifest in altered kinetics (Lee et al. 2008; Lee et al. 2011; Lee et al. 2012). For example, isoform 1 (AT4G08900) of arginase appears to be more highly expressed in vegetative tissue, whereas isoform 2 (AT4G08870) is more abundant in reproductive organs (Lee et al. 2012). Isoform-specific differences in vegetative and reproductive development have also been observed when each of the four voltage-dependent anion channels (VDAC) isoforms were disrupted (Tateda et al. 2011). Post-translational events also play a pivotal role in regulating enzyme activity and thus the flux of a metabolic pathway through modifications of proteins and the assembly of enzyme complexes. Some of the protein modifications observed on 2-D gels, especially truncated products and pI-shifted proteins, have often been perceived as artefacts introduced during sample preparation. However, in a recent survey of the mitochondrial proteome from different organ/cell types (Lee et al. 2012), many proteins in the TCA cycle, ETC and photorespiration undergo post-translational modifications in a tissue specific fashion that are highly reproducible across biological replicates, suggesting that these changes are not random. The functional implications of specialized differences in post-translational modifications on the contribution of mitochondrial metabolism in different tissues remains to be explored.

Plant Mitochondrial Oxidative Stress and Cellular Signalling

Environmental, biotic, abiotic and chemical stresses applied to plants are well known to induce oxidative stress in plant cells. These stresses alter plant metabolism, growth and development and, at their extremes, can lead to death. Recently, a number of studies have begun to examine the changes that occur within plant mitochondria following the induction of oxidative stress. The accumulation of ROS, ROS induced lipid peroxidation, changes in metal content, changes in protein abundance and their interactions in mitochondria following exposure to external stress, and the role of these changes in signaling beyond mitochondria, combine to define the importance of mitochondria as environmental sensors.

Accumulation of ROS in mitochondria

Mitochondria contain two terminal oxidases that reduce oxygen to water and the entire ETC is known to be a significant source of ROS under normal conditions. However, under steady state conditions, this ROS production is dealt with by antioxidant enzymes and small molecules to limit cellular damage. However, under some conditions, these defenses are overwhelmed and ROS accumulate (Figure 3). Superoxide is produced in mitochondria by peripheral single electron transfers from reduced components in the ETC to oxygen (Moller 2001). Classically, the ubiquinone pool and components in CI and CIII have been implicated, however recently CII has also been shown to produce significant superoxide (Quinlan et al. 2012). Measurements suggest that 2–5% of oxygen consumption by mitochondria is due to single electron superoxide formation, while the majority of oxygen consumption occurs at the terminal oxidases by four electron reduction of oxygen to water. The rate of superoxide production by mitochondria depends on the concentration of oxygen and on the redox poise of ETC components. Therefore, ROS production by mitochondria is low during hypoxic conditions (Noctor et al. 2007), is elevated when respiration inhibitors block the ETC and cause over-reduction of earlier ETC components (Maxwell et al. 1999), and can be altered by environmental factors and chemicals that alter the rate of these peripheral electron transfer reactions (Moller 2001; Moller et al. 2007; Noctor et al. 2007). Notably, nitric oxide is a potent inhibitor of the mitochondrial ETC and its generation during plant stress may be critical in the elevation of ROS production from mitochondria in plants (Millar and Day 1996; Yamasaki et al. 2001; Zottini et al. 2002). ROS have been shown to have a direct inhibitory effect on a number of mitochondrial enzymes including components of the ETC. Most notably H2O2 can inhibit the TCA cycle enzyme aconitase by modification of its 4Fe-4S cluster (Verniquet et al. 1991).

Figure 3.

Proposed scheme of the interaction of ROS, proteins, metals and lipid peroxidation end products in plant mitochondria exposed to stresses. 

When a plant is exposed to environmental stress this must be sensed either by a receptor on the cell wall/plasma membrane or directly inside the cell. This is then likely signalled to the nucleus and the mitochondria. Inside the organelles an accumulation of ROS can occur when antioxidants and antioxidant proteins (AP) are overwhelmed and this leads to the production of lipid peroxidation end products (LPEP), damaged proteins (DP), unfolded proteins (UP) and release of transition metals (TM). The accumulated ROS can directly inhibit proteins, while accumulated LPEP can modify proteins with lipoic acid cofactors (ML) directly amino acid (ML). Accumulated transition metals can facilitate metal catalyzed oxidation leading to the formation or carbonyl groups (CB). The organelles may also signal the accumulation of ROS to the nucleus. Either the external sensing of stress or organellar signalling of ROS leads to the production of new proteins by the nucleus, including replacement proteins sHSPs, proteases and antioxidant proteins (AP). These are then involved in the refolding of unfolded proteins (UP) or the degradation of damaged proteins (DP). AP, Antioxidant proteins; CB, Carbonyl modified amino acids; DP, Damaged proteins; LPEP, Lipid peroxidation end-products; MA, Michael adducts formation on amino acids by LPEP; ML, Michael adducts formation on lipoic acid by LPEP; sHSP, Small heat shock proteins; TM, Transition metals; UP, Unfolded proteins.

ROS induced lipid peroxidation in mitochondria

Lipid peroxidation in a mitochondrial context refers to free radical autoxidation of polyunsaturated fatty acids of membrane lipids such as linoleic acid, linolenic acid, arachidonic acid and hexadecatrienoic acid to yield various cytotoxic aldehydes, alkenals and hydroxyalkenals. The interaction of the hydroxyl radicle (OH ) with polyunsaturated fatty acids initiates lipid peroxidation that by a sequential series of reactions leads to a number of toxic lipid peroxidation end products (LPEP) by a non-enzymatic, metal ion enhanced process (Noordermeer et al. 2000) (Figure 3). Probably the most cytotoxic and studied LPEP is 4-hydroxy-2-nonenal (HNE). HNE is potentially able to undergo a number of reactions with proteins, phospholipids and nucleic acids. It has been shown to accumulate in plants during the oxidative burst (Deighton et al. 1999), biotic stresses (Montillet et al. 2002) and during exposure to chemical stresses (Winger et al. 2005). HNE has been shown to inhibit the activities of mitochondrial pyruvate dehydrogenase, 2-oxoglutarate dehydrogenase and glycine decarboxylase via the modification of the lipoic acid residues found on the E2 subunits of these enzymes (Taylor et al. 2002). This modification by HNE results in the formation of HNE-Michael adducts, which no longer allow the normal function of the essential E2 catalytic subunit. Further research has also identified a wider range of proteins that are damaged or inhibited by HNE. In some cases lipoic acid moieties are not involved and HNE acts directly by covalent modification of amino acid residues such as Cys, Lys, His, Ser and Tyr (Esterbauer et al. 1991). The alternative oxidase (AOX) and NAD-malic enzyme are inhibited by HNE (Winger et al. 2005), and it has been proposed in both enzymes that modification of a critical cysteine residue near the active site might be responsible (Millar and Leaver 2000; Winger et al. 2005). The likely pathways for the detoxification of lipid peroxidation end-products such as HNE are yet to be elucidated in plants, but it is likely to involve glutathione-s-transferase (GST) conjugation of modified peptides and may also involve aldehyde dehydrogenases and/or aldose reductases.

Metallome changes during oxidative stress

Plant mitochondria contain the transition metals Fe, Cu, Zn and Mn as well as trace levels of Co and Mo (Tan et al. 2010). The redox cycling metals Cu and Fe tend to be concentrated to the integral membrane proteome likely due to the abundance of Cu- and Fe-containing ETC components and account for approximately 75% of the mitochondrial metallome (Tan et al. 2010). Treatment of cultured cells with chemicals known to induce oxidative stress induce a reduction of peripheral membrane Fe and integral membrane Cu content, suggesting damage to membrane-associated ferro-proteins and membrane-embedded cupro-proteins (Tan et al. 2010). Significant losses have also been seen for soluble fraction Fe, Cu and Mn suggesting damage to metallo-matrix proteins and release of Fe, Cu and Mn. These labile transition metals, in particular redox active copper and iron ions, typically react with hydrogen peroxide in Fenton type reactions to catalyze the formation of OH . In addition, other redox-cycling reactions within the mitochondria exist and they are capable of eliciting metal-catalyzed oxidation (MCO) (Figure 3). Metal-catalyzed oxidation of proteins results in the oxidation of susceptible amino acids such as arginine, lysine, proline and histidine (Stadtman 1993), among a plethora of other poorly characterized consequences. One of the major by-products of MCO of proteins is the irreversible formation of carbonyl derivatives. These carbonyls are highly reactive and may cause protein aggregation if the damaged proteins are not degraded. These reactive carbonyls are often studied as markers of oxidative stress in both plants and animals. However, the impacts of protein carbonylation in plants at a subcellular level remains poorly understood.

Proteome changes during oxidative stress

A number of studies have revealed global changes in protein abundance of mitochondrial proteins following conditions that induce oxidative stress in a wide range of plant species (Sweetlove et al. 2002; Taylor et al. 2005; Chen et al. 2009; Taylor et al. 2009; Jacoby et al. 2010; Huang et al. 2011; Komatsu et al. 2011; Hossain et al. 2012; Tan et al. 2012). Recently it has also been shown that the large respiratory subunits of the ETC also coordinate protein changes to alter respiration in response to oxidative stress conditions (Tan et al. 2012). In addition to these changes are changes in proteins responding to ROS and the damage they cause. For example the mitochondrion is protected by a number of antioxidant enzymes that detoxify ROS and many have been observed to vary in abundance during oxidative stress including: Mn-superoxide dismutase (Jiang et al. 2007), ascorbate peroxidase (Dooki et al. 2006) monodehydroascorbate reductase (Sarry et al. 2006), glutathione peroxidase (Jiang et al. 2007) and peroxiredoxins (Sweetlove et al. 2002; Sarry et al. 2006). The importance of these organellar antioxidant defense mechanisms in plant stress tolerance has been highlighted by transgenic manipulation of the expression of these antioxidant enzymes (Allen et al. 1997). Changes in abundance of GST proteins are detected in almost every stress proteome study, sometimes accompanied by aldehyde dehydrogenases (Cui et al. 2005; Ndimba et al. 2005; Sarry et al. 2006), both of which may be involved in the detoxification of lipid-peroxidation end-products. Many metalloproteins have been shown to change in abundance following exposure to stress including Mn-SOD (Jiang et al. 2007), the Fe-S center containing CIII UCR1 (Tan et al. 2012), CI 75 kDa subunit (Taylor et al. 2005) and the copper interacting CI subunit B16 (Tan et al. 2012). In addition to these direct protein changes, other proteins have been observed to increase in abundance including mitochondrial class I and mitochondrial class II small heat shock proteins (sHsps) (Siddique et al. 2008) (Figure 3). It is generally accepted that sHsps alleviate the deleterious effects of stresses by preventing protein denaturation and aggregation, as well as facilitating the correct refolding of denatured proteins. Similarly increases in constitutive serine protease activity can be induced by oxidative stress in mitochondria (Sweetlove et al. 2002) although the specific Clp and FtsH serine proteases responsible for this remain unresolved. Mitochondria also contain Lon metalloproteases (Sarria et al. 1998; Rigas et al. 2009) and together with the serine proteases it seems likely that these proteins are responsible for the degradation of oxidatively damaged proteins (Figure 3).

Regulation of downstream gene expression such as GSTs and HSPs

Due to the complexity of the interconnected ROS signalling network that operates within plant cells (Mittler et al. 2004), it is difficult to attribute specific cellular responses to mtROS signals alone (Møller and Sweetlove 2011). However, by analysing the gene expression patterns elicited by respiratory inhibitors and the transcriptional anomalies exhibited by plants carrying mutations to mitochondrial genes, it appears that the expression patterns of certain gene clusters could be key indicators of mtROS signals. Here we briefly outline how the molecular signatures generated by mtROS signals could potentially be deduced by analysing transcriptional upregulation within subsets of the glutathione-s-transferases (GSTs) and heat shock proteins (HSPs). In maize, expression patterns of gstIII and gstI genes were similar between NCS4 (mitochondrial ribosome mutant affecting translation) and NCS6 (CIV mutant) compared to their respective wild type (Karpova et al. 2002). However, both gstIII and gstI were increased in the NCS2 (CI mutant) (Karpova et al. 2002). Therefore, dysfunction of CI (NCS2), but not dysfunction of either CIV (NCS6) or mitochondrial translation (NCS4), induces these GSTs. Interestingly, treatment with exogenous H2O2 can also induce expression of gstIII and gstI genes in maize leaves (Karpova et al. 2002), perhaps suggesting that the NCS2 mutation (CI) exerts its signaling effect via H2O2 signals derived from mitochondrion, whereas the NCS4 (ribosome) and NCS6 (CIV) mutations communicate mitochondrial dysfunction via a different route. These mechanisms of ROS signaling could be conserved between species, as mutations to CI in Arabidopsis induce accumulation of ROS (Lee et al. 2002; Meyer et al. 2009), and higher expression levels of GSTF3 and GSTF6 transcripts (Meyer et al. 2009). The activity of the Arabidopsis GSTF8 promoter has been defined as a marker for defense response during the early stages of stress exposure (Sappl et al. 2009), and recently, a forward genetics approach that employed GSTF8 promoter activity to define mutants with altered stress responses identified a novel CII mutant, dsr1 (Gleason et al. 2011). Analysis of GSTF8 promoter activity upon application of exogenous chemicals revealed that the dsr1 mutant could not mediate GSTF8 promoter activity in response to salicylic acid (SA), despite this treatment eliciting strong GSTF8 promoter activity in a wild type background. However, external supply of H2O2 recovered the induction of GSTF8 promoter activity in dsr1. These results suggest a stress signaling model whereby upstream SA signals are converted into mtROS signals via H2O2 production at CII. This signal is then transmitted to the nucleus to elicit downstream responses such as induction of a certain subset of GST transcripts.

Numerous studies have measured increased expression of genes encoding heat shock proteins upon disruption of the ETC via genetic mutation or treatment with respiratory inhibitors. For example, higher abundance of HSP transcripts was observed in Arabidopsis CI mutant lines (Meyer et al. 2009), in cultured Arabidopsis cells treated with CI inhibitors (Garmier et al. 2008), and in maize NCS mutants including NCS2 (CI), NCS4 (ribosomal translation) and NCS6 (CIV) (Kuzmin et al. 2004). Further dissection of the triggers that can elicit higher expression of sHSP22A suggest that its promoter activation depends upon a decrease in the potential difference across the mitochondrial inner membrane, independent of Ca2+ signals (Kuzmin et al. 2004). A study in C. elegans suggests that the induction of mitochondrial heat shock proteins is regulated by the proteolytic cleaving of specific mitochondrial proteins to generate messenger peptides which are exported to the nucleus, with the divergent amino acid sequences across a range of different messenger peptides providing the specificity required to elicit the subsequent induction of a specific set of genes (Haynes et al. 2010; Moller and Sweetlove 2011). This offers a tempting explanation for the link between sHSP22A induction and lower potential difference across the mitochondrial inner membrane in maize (Kuzmin et al. 2004), as the depolarization of mitochondrial membrane potential could permit the export of signaling peptides. It can be posited that different mitochondrial stress signals can activate divergent pathways of mitochondrial retrograde signaling, as within the aforementioned suite of maize mutants, there are significant overlaps amongst HSP gene family expression but divergent transcriptional responses within the set of GSTs, with higher expression of GSTI and GSTIII only being elicited by CI mutation. Molecular models of specific mitochondrial retrograde signaling pathways are maturing and detailed analysis of expression patterns within and between sets of stress responsive genes can further our understanding of the molecular signaling processes communicated by mtROS. This opportunity is particularly timely as well characterised suites of ETC mutants are now available across a range of species.

Plant Mitochondrial Roles in Harsh Environments

Insights into how the respiratory roles of mitochondria in cells operate under harsh environmental conditions encountered by intact plants have been derived by analyzing phenotypes across a wide range of physical scales (Figure 4). Ecosystem studies have shown that plant respiration rates were a major factor underpinning slower rates of forest and crop productivity across Europe during a 2003 heatwave, with prolonged heat and drought temporarily switching certain European forests into sources of atmospheric carbon, rather than sinks (Ciais et al. 2005). At an individual plant level, physiological studies have defined that respiratory rates are a key determinant of growth reductions under a range of stresses such as extreme temperatures, drought and salinity (Atkin et al. 2005; Atkin and Macherel 2009; Jacoby et al. 2011). The most powerful theoretical framework analysing these responses is the process of carbon balance, where growth rate is positioned as the sum total of carbon captured via photosynthesis minus carbon expended by respiration (McCree 1986; Amthor 2000). Tissue-level experiments have shown that some stresses, such as temperature, can induce rapid changes to respiratory rates of excised tissue, probably mediated through thermal effects on the kinetics of the ETC (Kurimoto et al. 2004; Armstrong et al. 2006), whereas exposing excised tissue to external NaCl has little impact on respiratory rates in the short term, probably because salinity exerts its effects on respiration rates through alterations to substrate provision and cellular energetic demand (Flowers 1972). Oxygen uptake rates of isolated mitochondria exposed to harsh conditions in vitro have defined the routes of mitochondrial electron transport that can withstand high concentrations of toxic substances such as NaCl, lipid peroxidation products and cadmium (Miller et al. 1973; Hamilton and Heckathorn 2001; Winger et al. 2007), while analyses of ETC biochemistry in mitochondria isolated from plants exposed to harsh stresses have repeatedly shown that the alternative respiratory pathways display increased activities in response to a wide range of stresses (McDonald 2008). In terms of individual enzymes, activity assays have shown that certain components of key metabolic pathways are sensitive to cellular products of stress, such as hydrogen peroxide, metal cations and HNE (Verniquet et al. 1991; Millar and Leaver 2000; Tan et al. 2010), so metabolic flexibility must be required to reshuffle TCA cycle flux around these blockages during environmental stress (Sweetlove et al. 2010). Recently, these characterisations of mitochondrial function are being complemented by high-throughput ‘omics studies that define the specific mitochondrial transcripts and proteins which underpin these tolerance strategies in commercially important crop species to provide mechanistic targets in respiration for crop improvement (Abe et al. 2002; Taylor et al. 2005; Jacoby et al. 2010). These can be considered in three main areas related to mitochondrial delivery of energy and intermediates, mitochondrial support for photosynthesis, and mitochondrial contributions to root cell homeostasis and transport processes.

Figure 4.

Plant mitochondrial roles in harsh environments. 

The left of this figure presents a schematic outline of three major mitochondrial processes that mediate plant growth and survival during stress. First is mitochondrial support for photosynthesis (green), second is the mitochondrial delivery of energy and intermediates (brown), and third are mitochondrial contributions to root cell homeostasis and transport processes (yellow). The right of this figure shows that mitochondrial processes such as respiration, metabolism and signaling can be analysed across a wide range of physical scales, demonstrating that mitochondrial stress responses are relevant to a broad range of researchers.

Cell survival under stress fuelled by mitochondrial provision of carbon skeletons, ATP and reducing equivalents

Experiments have described a number of cellular responses that mediate tolerance mechanisms in stress treated plants, such as altered rates of protein turnover, rebalancing of cellular metabolite pools, altered abundance of ROS species and changes in the redox ratios within pools of reducing equivalents. Each of these phenomena can be linked to mitochondrial processes through alterations to central metabolic pathways, or by their demand for fast rates of ADP:ATP cycling mediated by respiratory oxidative phosphorylation (Figure 4). For instance, plants display differential rates of protein turnover relative to unstressed controls in response to a wide range of stresses such as drought, osmotic stress and heat stress (Dungey and Davies 1982; Zagdanska 1995; Huang et al. 2012), and it has been well established that this protein quality control network incurs a significant energetic cost, as ATP is hydrolysed to fuel the refolding and degradation of damaged proteins, and also the synthesis of new replacement proteins (Moller et al. 2007). Although the plastidic ETC produces a large amount of cellular ATP in illuminated leaves, this is all consumed by the process of photosynthetic carbon reduction; therefore, mitochondrial electron transport provides the ATP during day and night in both roots and shoots for all other cellular operations. As a consequence, robust mitochondrial function in all tissues is crucial to plant survival under stress conditions. Another molecular signature of harsh environmental conditions is the accumulation of high concentrations of particular metabolites, such as proline, glycine betaine (GB) and GABA (Hare et al. 1998). At a molecular level, these molecules can stabilize proteins, scavenge ROS, and serve as alternative energy sources when classical metabolic pathways are substrate-limited or biochemically-inhibited (Arakawa and Timasheff 1985; Verslues and Sharp 1999; Chen and Dickman 2005). The mitochondrial role in regulation of proline and GABA concentrations comes through the abundance and activity of catabolic proteins such as ProDH, P5CDH and GABA-T that are located in the mitochondrial matrix (Miller et al. 2009; Renault et al. 2010), while a mitochondrial role in regulating cellular concentrations of GB comes through the provision of photorespiratory serine by the reactions of GDC and SHMT in mitochondria (Bhuiyan et al. 2007). Further links between mitochondrial function and cellular metabolic status come through the provision of 2-OG via the TCA cycle, as 2-OG is a precursor for many N-containing metabolites in plant cells, and exposure to abiotic stress can alter the rate of TCA cycle flux (Baxter et al. 2007; Sweetlove et al. 2010). Increased ROS abundance is a convergence point for a wide range of stress treatments, and stress treated plants commonly exhibit shifts in the redox poise of ascorbate and glutathione to more oxidized states following stress treatments (Foyer and Noctor 2011). Mitochondrial regulation of these phenomena has been illustrated by transgenic studies showing that manipulation of mitochondrial enzymes can alter whole-plant redox balance and stress tolerance (Dutilleul et al. 2003; Morgan et al. 2008; Tomaz et al. 2010).

Mitochondria support for photosynthesis during environmental challenge

Photosynthetic carbon capture is a coordinated process that requires mutual cooperation between organelles, and mitochondrial functions are particularly important in sustaining photosynthesis under high light. It is now widely accepted that fast photosynthetic rates and the avoidance of photoinhibition are dependent upon mitochondrial function being configured to rapidly transfer electrons from NADH into water through the non-phosphorylating bypasses of the classical ETC (Millar et al. 2011) (Figure 4). This framework is supported by several lines of evidence, such as measurements showing that the abundance and activity of this particular set of mitochondrial enzymes are induced by high light (Noguchi and Yoshida 2008), while toxic inhibition of the mitochondrial ETC leads to decreased photosynthetic rates (Saradadevi and Raghavendra 1992), and compellingly evidence from knockouts of genes encoding mitochondrial proteins that result in lower photosynthetic rates and acute sensitivity to high light coupled to drought (Sweetlove et al. 2006; Giraud et al. 2008). Rates of plant growth are largely determined by the balance between photosynthesis and respiration, with more productive, fast-growing plants generally allocating a smaller fraction of their daily fixed carbon to respiratory CO2 production. This reserves a larger fraction of fixed carbon to allocate into synthesising new tissue, which is primarily constructed by the accumulation of carbohydrates (Poorter et al. 1990; De Block and Van Lijsebettens 2011). However, there appears to be a limit to the productivity increases that can be acquired through slower respiratory rates, with canola experiments showing that plants selected for slightly slower respiration rates displaying increased biomass production, but plants where respiration rates had fallen below a certain threshold displayed dramatically slower growth rates probably because cellular energy supply cannot match baseline demand (Hauben et al. 2009). Although this study did not investigate a stress condition, the results identify an important limitation in strategies that aim to enhance biomass accumulation through selecting for uniformly lower respiration rates, as slow respiration would likely be a disadvantage during transient stressful periods which require increased rates of mitochondrial energy production to fuel energetically costly cellular defense processes. Respiratory elasticity is likely to be the most useful trait, enabling slow respiration rates to promote growth during optimal conditions, but faster respiration rates to fuel defense during transient stress periods.

Root-specific mitochondrial processes mediating tolerance to unfavourable soil conditions

Oxidative phosphorylation in mitochondria is the main source of ATP in root tissue, and mitochondrial processes are also involved in the tolerance of plants to root-specific stresses, such as low oxygen and toxic soil conditions (Figure 4). Root tissue is prone to dramatic fluctuations in cellular oxygen concentrations, owing to the low solubility of oxygen in water coupled with hydrological flood-drain cycles imposed by variations in rainfall or soil drainage, in both rainfed and irrigated agricultural systems. Under low oxygen, mitochondrial metabolism shifts away from the classical TCA cycle and ETC, towards other mitochondrial processes such as amino acid metabolism (Millar et al. 2004b; Taylor et al. 2010). Mitochondrial ROS defenses such as MnSOD accumulate under anoxia, presumably in anticipation of the forthcoming ROS burst that will occur upon re-oxygenation (Millar et al. 2004b; Shingaki-Wells et al. 2011). Large areas of the earth's surface are covered by soils with chemical properties that are sub-optimal for plant growth, due to high concentrations of salts or heavy metals, insufficient bioavailability of essential nutrients like iron, phosphate, sulphur and nitrogen, as well as unfavourably acidic or alkali pH. The perception of heavy metal toxicity has been shown to involve mitochondrial ROS signals (Garnier et al. 2006), and the key role of mitochondrial ROS defenses within root cells has been pinpointed by the dramatic root growth reductions elicited by exposing mitochondrial peroxiredoxin mutants to high cadmium concentrations (Finkemeier et al. 2005). Exudation of TCA cycle intermediates have been linked to iron absorption from soils where neutral pHs render the iron insoluble, as these acidic metabolites acidify the soil, thus solubilising sequestered iron to increase bioavailability (Vigani 2012). Conversely, aluminium toxicity can be alleviated by exudation of citrate and isocitrate, as these TCA cycle intermediates have chelating properties that decrease bioavailability of the toxic Al3+ ion (Ma et al. 2001; Fujii et al. 2012). Sodium exclusion is a well defined mechanism of plant NaCl tolerance, and strong links between the degree of sodium exclusion and the rate of root respiration have been demonstrated by the collapse of sodium exclusion following anoxia treatment of roots (Drew and Lauchli 1985). Radioisotope tracer studies have shown that rates of root respiration in saline media are positively correlated to sodium exclusion capacity between rice varieties (Malagoli et al. 2008), further illustrating this relationship.

Future Perspectives

As the powerhouses of eukaryotic cells, mitochondria represent an ancient but flexible factory in cells that enable cell function, growth and division through energy metabolism. While our knowledge of how plant mitochondria work is rapidly increasing, much is still being extrapolated from yeast and mammalian systems without direct evidence in plants. Detailed insights into the assembly of mitochondrial machinery, the signalling by mitochondrial of oxidative stress and the regulation of respiratory rate are still needed in order to maximise respiration for plant protection in harsh environments and to minimize respiratory losses to enhance plant yields.

(Co-Editor: Jianping Hu)

Acknowledgements

This work was supported by the Australian Research Council (ARC) ARC Centre of Excellence for Plant Energy Biology (CE0561495). RPJ is supported by a Grains Research and Development Corporation (GRDC) PhD scholarship, LL was funded by Scholarship International Research Fees (SIRF), University International Stipend (UIS) and a Top Up Scholarship for UIS. AHM is supported by the Australian Research Council (ARC) as an ARC Future Fellow.

Ancillary