The Impact of Changes in Mitochondrial Malate Dehydrogenase Abundance on the Respiratory Metabolism of Arabidopsis Thaliana Mutants and Ecotypes

The Impact of Changes in Mitochondrial Malate Dehydrogenase Abundance on the Respiratory Metabolism of Arabidopsis thaliana Mutants and Ecotypes

This thesis is presented for the degree of Doctor of Philosophy at The University of Western Australia

Yun Shin Sew

2015

ARC Centre of Excellence in Plant Energy Biology Centre for Comparative Analysis of Biomolecular Networks School of Chemistry and Biochemistry Declaration

The research presented in this thesis is my own work unless otherwise stated. This work was carried out in the Australian Research Council Centre of Excellence in Plant Energy Biology and Centre for Comparative Analysis of Biomolecular Networks at the University of Western Australia. The material presented in this thesis has not been submitted for any other degree.

Yun Shin Sew (Michelle)

Acknowledgements

There are many people that I would like to thank, who in diverse ways made this thesis possible.

First and foremost, I wish to express my sincerest gratitude to my two supervisors, Harvey Millar and Elke Stroeher for their brilliant suggestions, useful advices, and dedicated guidance throughout my PhD study. Many thanks to their continuous supports and patience in editing thesis drafts.

I am very thankful to the people who had provided me the technical assistance, Ricarda Fenske and Julia Grassl for their helps in conducting the multiple reaction monitoring assays and analyses, and Dorothee Hahne for her help in the metabolite analysis.

My heart-felt thanks go to the people in the Millar’s lab (2011 to 2014): Adriana, Alex, Christian, Clarke, Cornelia, Ellen, Fazilah, Hafiz, Jakob, James, Katharina, Lei, Leila, Nicholas, Rachel, Rali, Richard, Sandi, ShaoBai, Tiago and Yan. I appreciate the helpful advices and good friendship that had brightened up my PhD life in Perth from each of you.

I gratefully acknowledge the receipt of a PhD scholarship from Malaysian Agricultural Research and Development Institute (MARDI) and a PhD top-up scholarship funded by ARC Plant Energy Biology (PEB).

Finally, special thanks go to my husband and mum in law for their understanding and supports. A deep sense of gratitude to my parents and siblings for their loves and moral supports, especially my dear mum who always encourage me to achieve higher education.

Publications

Part of this thesis has been published in the following refereed journal article and methodological paper:

1. Sew, Y. S., Ströher, E., Holzmann, C., Huang, S., Taylor, N. L., Jordana, X., & Millar, A. H. (2013). Multiplex micro‐respiratory measurements of Arabidopsis tissues. New Phytologist, 200(3), 922-932.

2. Sew, Y. S., Millar, A. H., & Stroeher, E. (2015). Micro-respiratory measurements in plants. Plant Mitochondria: Methods and Protocols, 187- 196.

List of Abbreviation

1D one-dimensional ACO aconitase ADP adenosine diphosphate ATP adenosine triphosphate ANOVA analysis of variance AOX alternative oxidase BCAA branched-chain amino acid BSA bovine serum albumin CDT controlled deterioration test CMDH cytosolic malate dehydrogenase CHMDH chroloplastic malate dehydrogenase CS citrate synthase Col ecotype Columbia DIGE differential in gel electrophoresis DLD dihydrolipoamide dehydrogenase DTT dithiothreitol ECL enhanced chemiluminescence EDTA ethylenediaminetetraacetic acid eFP electronic fluorescent pictograph FAD+ flavin adenine dinucleotide

FADH2 reduced flavin adenine dinucleotide Fe-S iron sulphur FUM fumarase FW fresh weight GABA gamma-aminobutyric acid GC-MS gas chromatography-mass spectrometer GDC glycine decarboxylase

H2O2 hydrogen peroxide

- HCO3 bicarbonate HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

HPR hydroxypyruvate reductase IAA iodoacetamide IDH isocitrate dehydrogenase Ile Isoleucine IMM inner mitochondrial membrane IVD isovalery CoA dehydrogenase kDa kilodalton KCN potassium cyanide LC-MS/MS liquid chromatography-tandem mass spectrometry Leu leucine MDH malate dehydrogenase ME NAD-dependent malic enzyme MES 2-(N-morpholino)ethanesulfonic acid methionine mETC mitochondrial electron transport chain MeV multiexperiment viewer MLR multiple linear regression mMDH or MMDH mitochondrial malate dehydrogenase MRM multiple reaction monitoring MS mass spectrometry MW molecular weight m/z mass to charge ratio NAD+ nicotinamide adenine dinucleotide oxidized form NADH nicotinamide adenine dinucleotide reduced form NADPH reduced nicotinamide adenine dinucleotide phosphate

NDex external NAD(P)H dehydrogenase

NDin internal NAD(P)H dehydrogenase

NH4 ammonium NaOAC sodium acetate nmol nanomole OXPHOS oxidative phosphorylation apparatus OAA oxaloacetate OCR oxygen consumption rate

OGDH 2-oxoglutarate dehydrogenase OMM outer mitochondrial membrane PAGE polyacrylamide gel electrophoresis PCA principal component analysis PDC pyruvate dehydrogenase complex PDH pyruvate dehydrogenase PEPC phosophoenolpyruvate carboxylase PAGE polyacrylamide gel electrophoresis PCR polymerase chain reaction PMDH peroxisomal malate dehydrogenase pmol picomole PPFD photosynthetic photon flux density PVP polyvinylpyridine qPCR quantitative polymerase chain reaction R correlation coefficient R2 correlation RH relative humidity rpm revolutions per minute of rotor ROS reactive oxygen species RuBisco ribulose-1,5-bisphosphate carboxylase oxygenase SDH succinate dehydrogenase SDS sodium dodecyl sulphate SGT serine‐glyoxylate aminotransferase SHMT serine hydroxymethyltransferase SEM standard error of the mean SUC succinyl-CoA ligase TBS tris buffered saline TBST tris buffered saline plus tween TCA cycle tricarboxylic acid cycle Tris tris(hydroxymethyl)aminomethane TTC triphenyl tetrazolium chloride TZ tetrazolium

UTR untranslated region VDAC voltage dependent ion channel vol volume w/v weight per volume

Table of contents

Declaration ...... 2

Acknowledgements ...... 3

Publications ...... 4

List of Abbreviation ...... 5

Table of contents ...... 9

Abstract ……………………………………………………………………………………………………………12

Chapter 1: General Introduction ...... 15

1.1 Arabidopsis thaliana as a model plant ...... 16

1.2 The central role of plant mitochondria in primary metabolism ...... 17

1.3 Tricarboxylic acid (TCA) cycle ...... 18

1.4 Mitochondrial electron transport chain (mETC) and oxidative ...... phosphorylation ...... 27

1.5 New insights into Arabidopsis thaliana ecotypes ...... 30

1.6 Aims and approaches of the study ...... 32

Chapter 2: Multiplex micro-respiratory measurements of Arabidopsis tissues .... 35 Foreword to Study I ...... 36 Abstract ...... 39 Introduction ...... 40 Materials and methods ...... 42 Results ...... 47 Discussion ...... 59 Conclusions ...... 65 Chapter 3: Impact of mmdh loss on metabolic enzyme networks in Arabidopsis ...... leaves ...... 67 Foreword to Study II ...... 68 Abstract ...... 71 Introduction ...... 72

Materials and Methods ...... 76 Results ...... 84 Discussion ...... 106 Conclusion ...... 120 Chapter 4: Impact of mMDH loss on seeds and roots of mMDH mutants ...... 123 Foreword to Study III ...... 124 Abstract ...... 127 Introduction ...... 129 Materials and methods ...... 136 Results ...... 142 Discussion ...... 168 Conclusion ...... 183 Chapter 5: Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates ...... 185 Foreword to Study IV ...... 186 Abstract ...... 189 Introduction ...... 190 Materials and methods ...... 193 Results ...... 204 Discussion ...... 238 Conclusion ...... 249 Chapter 6: General Discussion ...... 251 6.1 Mitochondrial malate dehydrogenase as a significant regulator of plant respiration ...... 253 6.2 Relative contribution of the two mMDH genes in respiratory metabolism ...... 253 6.3 Functional roles of mMDHs in plant respiratory metabolism ...... 254 6.4 Relationships between mMDH and NAD-MDH isoforms in other subcellular compartments ...... 255 6.5 Relationships between TCA cycle enzyme levels in leaf respiratory metabolism in response to perturbed and naturally altered mMDH protein levels ...... 257

6.6 Multiple roles of mMDH in energy metabolism and homeostasis contribute towards energy efficiency during plant development ...... 260

Future direction ...... 270

References ...... 273

Abstract

Mitochondrial respiration releases carbon dioxide that was previously fixed by photosynthesis, so its rate influences the growth, productivity and energy use efficiency of plants. The tri-carboxylic acid (TCA) cycle is a critical intersection between catabolic and anabolic processes and it plays a central role in mitochondrial respiratory metabolism. Mitochondrial malate dehydrogenase (NAD- MDH) (EC 1.1.1.37) is involved in the final step of the TCA cycle, and is responsible for the inter-conversion of malate and oxaloacetate, concomitant with the reduction or oxidation of reducing equivalents. This thesis aimed to gain a better understanding of the role of mitochondrial MDH (mMDH) in respiratory metabolism during growth and development of Arabidopsis thaliana. This was undertaken by transcript, protein and metabolite analysis on different organs (leaf, seed and root) at distinct developmental stages using single (mmdh1-2 and mmdh2-1) and double (mmdh1-2mmdh2-1) T-DNA knockout lines, as well as in a complemented line (mmdh1mmdh2 35S: MMDH1), in the A. thaliana (Col-0) genetic background. The correlation between naturally varied mMDH levels and leaf respiratory metabolism of A. thaliana ecotypes were also investigated. Ecotype Col-0 was used as a reference for data integration between mMDH mutant experiments and ecotype assessments.

Both findings from mMDH mutants and ecotypes supported the hypothesis that mMDH plays a significant role in modulating leaf respiration in plants as lower protein abundance of mMDH was consistently, significantly and inversely correlated with plant respiration rates. MMDH1 (At1g53240) appeared to be the predominant mitochondrial isoform of NAD-MDH when compared to MMDH2 (At3g15020). This was inferred from higher transcript and protein abundances of MMDH1, and a larger impact of single MMDH1 mutation on plant phenotype, MDH expression compensation and metabolism. Functional redundancy between mMDH isoforms was evident in leaf respiratory metabolism as complementation of MMDH1 cDNA in the mMDH double mutant background restored wild-type growth phenotypes. However, the lack of MMDH2 could not be fully substituted by MMDH1 in heterotrophic organs (seeds and roots) and during certain plant developmental

12 stages, suggesting either some unknown specific roles of MMDH2 or the importance of the native promoter in complementation. Transcript analyses from mMDH mutants and ecotypes suggested that NAD-MDH isoforms with a higher degree of functional redundancy (e.g. MMDH1 and MMDH2) and those with relatively greater participation in the biochemical NAD-MDH network (e.g. MMDH1, CMDH1, CHMDH and PMDH2) showed stronger correlation with each other in expression level. These relationships symbolise a cooperative model of the MDH network in executing global cellular redox homeostasis, probably facilitated by malate-oxaloacetate shuttling system across subcellular compartments. In both mMDH mutants and ecotypes, MMDH1 and SDH1 proteins (At5g66760& At2g18450) were significantly correlated with respiration rates, demonstrating negative and positive relationship with respiration rates respectively. It is therefore suggested that MDH is a negative and SDH is a positive marker for plant respiration rates.

Loss of mitochondrial MDH had significant effects on the overall energy status of Arabidopsis plants. The important roles of mMDH in plant carbon partitioning and carbon use efficiency was evident from the delayed growth, significantly reduced biomass of both autotrophic and heterotrophic organs and seed defects in the mmdh1-2mmdh2-1 double mutant, concomitant with elevated respiration rates throughout Arabidopsis developmental stages. Complete loss or decreased mMDH protein levels in Arabidopsis resulted in a marked accumulation of leaf glycine, consistent with a slowed photorespiratory process. Defective nitrogen metabolism was also consistent with lower abundances of leaf 2-oxoglutarate dehydrogenase subunits (At3g55410 and At5g65750) in the mmdh1-2mmdh2-1 double mutant. Free amino acids, particularly branched-chain amino acids, accumulated in leaves of fast-respiring mmdh1-2mmdh2-1 double mutants and also in low-MDH ecotypes. This was observed to an even greater extent in seeds and roots of mmdh1- 2mmdh2-1, consistent with increased catabolism of amino acids in response to lower mMDH protein levels. Collectively, the present study has demonstrated that a deficiency of mMDH proteins affects nitrogen and carbon assimilation as well as the efficiency of carbon usage across Arabidopsis organs and their development and

13 that natural variation in MDH levels may be an important factor in defining respiration rates.

Chapter 1:

General Introduction

Chapter 1. General introduction

1.1 Arabidopsis thaliana as a model plant

Arabidopsis thaliana is a small dicotyledonous plant in the mustard family (Cruciferae or Brassicacea). Arabidopsis has a rapid entire life cycle of 6-8 weeks which include seed germination, formation of a rosette plant, bolting of the main stem, flowering and maturation of the first seeds. Arabidopsis plant can be easily grown in the laboratory environment, such as in a petri dish or maintained in pots located in a greenhouse or under controlled fluorescent lighting in a growth chamber. Since 1970s, A. thaliana has been seen as a genetic tool (Redei, 1975) and it has progressively developed as a model organism in the studies of genes related to physiological functions of plants (Meinke et al., 1998). It possesses one of the smallest genomes among all higher plants; approximately 146 megabases divided over 5 chromosomes (Bevan and Walsh, 2005). According to the data obtained from genome sequencing of A. thaliana which was completed in late 2000, its genome was predicted to contain approximately 26000 protein-coding genes that encoded for around 35000 protein isoforms with many of their functions unknown (Bevan and Walsh, 2005). Generation of Arabidopsis mutants via T-DNA insertion is one of the common approaches taken by scientists for functional studies of Arabidopsis genes. In this regards, the GABI-KAT, SAIL, Salk, WISC, FLAG, and SK sets were generated and sequence-indexed by various laboratories worldwide, made up of at least 325000 publicly available Arabidopsis T-DNA insertion lines (O'Malley and Ecker, 2010). To date, there is a spectrum of sophisticated Arabidopsis analysis, visualization tools and even sequence database that have been developed and are updated intermittently by Arabidopsis research community. For instance, the Arabidopsis Information Resource (TAIR) (http://Arabidopsis.org) (Rhee et al., 2003), Arabidopsis electronic Fluorescent Pictograph (eFP) browser (http://bar.utoronto.ca/efp/cgi-bin/efpWeb.cgi) (Winter et al., 2007), SIGnAL T- DNA express: Arabidopsis gene mapping tool (http://signal.salk.edu/cgi- bin/tdnaexpress), SIGnAL Arabidopsis transcriptome genomic express database (http://signal.salk.edu/cgi-bin/atta) and SUBcellular location of proteins in Arabidopsis (SUBA) (http://suba.plantenergy.uwa.edu.au/) (Tanz et al., 2013;

Chapter 1. General introduction

Hooper et al., 2014). These publicly available resources provide a convenient, powerful and efficient basis to investigate the gene expression and regulation, subcellular location, metabolic pathways and a broad range of interaction networks.

1.2 The central role of plant mitochondria in primary metabolism

Plants need energy to provide basic building blocks such as nucleic acids and amino acids which are the fundamental elements for growth, development, reproduction and maintenance. This energy is primarily supplied by respiration, an indispensable biological process that occurs in mitochondria. This organelle was first discovered in 1890 by Altmann and it was named ‘bioblast’ and shown to be an elementary component that carries out vital functions in living cells (Altmann, 1890). The name mitochondrion was introduced in 1898 and originates from the Greek "mitos" (thread) and "chondros" (granule), referring to the appearance of these structures during spermatogenesis. This organelle was first associated with cell respiration in the early 1910s in a study conducted by Battelli and Stern (Battelli and Stern, 1912). The primary function of plant mitochondria is oxidation of respiratory substrates and the transfer of electrons to oxygen via the respiratory electron transport. Besides that, mitochondria are also involved in the catabolic and biosynthetic activities such as synthesis of nucleotides, metabolism of amino acids and lipids, and synthesis of vitamins and cofactors that are crucial for growth and development in living organism. In plants, mitochondria participate in the photorespiratory pathway (Sarojini and Oliver, 1983), and the export of organic acid intermediates and carbon skeletons for nitrogen assimilation (Foyer et al., 2011). It has been suggested that plant mitochondria are generally responsible for the production of most of the adenosine-5’-triphosphate (ATP) in darkened green tissue and non-photosynthetic tissues, whereas chloroplasts are the main source of energy formation in the light (Hoefnagel et al., 1998). However, mitochondria in these tissues also contributed to the ATP production even in illuminated green tissue for sucrose synthesis, C4 metabolic pathway steps, metabolic transport and

Chapter 1. General introduction protein synthesis. Multiple lines of evidence demonstrated that mitochondria contribute considerably to high rates of photosynthesis in chloroplasts, by removing the excess reducing equivalents generated in the chloroplast via the malate valve, which are subsequently oxidized by the mitochondrial electron transport chain to generate ATP in the cytosol (Padmasree et al., 2002; Scheibe et al., 2005; Noctor et al., 2007).

1.3 Tricarboxylic acid (TCA) cycle

The tricarboxylic acid (TCA) cycle, also known as the citric acid cycle or the Krebs cycle was first proposed in 1937 (Krebs and Johnson, 1937). The TCA cycle represents a nearly universal feature of aerobic respiratory metabolism in organisms and is an important part of the respiratory machinery besides glycolysis and mitochondrial electron transport chain. In the mitochondrial matrix, there are eight enzymes that constitute the TCA cycle namely; aconitase, citrate synthase, isocitrate dehydrogenase, 2-oxoglutarate dehydrogenase, succinyl-CoA synthetase, succinate dehydrogenase, fumarase and malate dehydrogenase (Figure 1). Acetyl CoA is the pre-requisite input for the TCA cycle, which is derived from the end product of glycolysis (in plants and animals), β-oxidation of fatty acids (in animals), as well as from degradation of ketogenic amino acids depending on the type of cell and organism (Sweetlove et al., 2010). It is essential to maintain a working concentration of the organic acids that are substrates and products of these enzymes in order to regulate a flux within the cycle. These organic acids are often termed ‘TCA cycle intermediates’ as a result. This cyclic flux enables organic carbon substrate oxidation and the successive generation of reducing equivalents (NADH and FADH2) that are used for ATP synthesis by oxidative phosphorylation through the electron transport chain (Sweetlove et al., 2010). Nevertheless, with the exception of reactions catalysed by succinyl-CoA synthetase and succinate dehydrogenase, it has been suggested that all the other mitochondrial TCA cycle reactions can be bypassed by reactions catalysed by related enzymes in other

Chapter 1. General introduction subcellular compartments, showcasing the plasticity of carbon metabolism in plants (Sweetlove et al., 2010; Millar et al., 2011; Nunes-Nesi et al., 2013).

Figure 1: Schematic representation of processes in the tricarboxylic acid (TCA) cycle. Pyruvate from glycolysis in cytoplasm enters the mitochondrial matrix to be converted into acetyl-coA via pyruvate dehydrogenase which then becomes the substrate for the first enzymatic reaction in the TCA cycle. The successive enzymatic reactions are carried out sequentially by eight TCA cycle enzymes, with concomitant production of reducing equivalents NADH and FADH2, which are used by processes along the mitochondrial electron transport chain (mETC) to build up a proton gradient and synthesize ATP via oxidative phosphorylation.

Chapter 1. General introduction

1.3.1 Pyruvate dehydrogenase complex (PDC)

Even though PDC is not strictly a part of the TCA cycle, no discussion of the cycle is complete without considering its role as the primary source of acetyl-CoA that begins the cycle. Plants contain two distinct, spatially localized PDCs, one is found within the mitochondrial matrix and the other in the plastid (Rahmatullah et al., 1989). PDC is a multi-enzyme complex that catalyse the oxidative decarboxylation of pyruvate to yield acetyl-CoA and NADH by the sequential reactions of its three components: pyruvate dehydrogenase (E1; EC 1.2.4.1), dihydrolipoamide transacetylase (E2; EC 2.3.1.12) and dihydrolipoamide dehydrogenase (E3; EC 1.8.1.4) (Millar et al., 1998). Pyruvate dehydrogenase (E1) catalyses the decarboxylation of pyruvate followed by reductive acetylation of lipoyl moieties covalently linked to the dihydrolipoamide acetyltransferase (E2), the second catalytic component of the complex. E2 is then catalyses the formation of acetyl- CoA; and dihydrolipoamide dehydrogenase (E3) re-oxidises the reduced lipoyl moieties of E2 with the consequent reduction of NAD+ to NADH (Luethy et al., 1996). In Arabidopsis, the mtPDC is encoded by two genes for E1α (At1g59900 and At1g24180) (Luethy et al., 1995; Quint et al., 2009), one gene for E1β (At5g50850) (Luethy et al., 1994), three genes for E2 (At3g52200, At3g13930, and At1g54220) (Guan et al., 1995; Thelen et al., 1999; Taylor et al., 2004), and two genes for E3 (At1g48030 and At3g17240) (Lutziger and Oliver, 2001). Mitochondrial PDC has two associated regulatory enzymes known as pyruvate dehydrogenase kinase and phospho-pyruvate dehydrogenase phosphatase (Luethy et al., 1996). Dihydrolipoamide dehydrogenase (E3) is part of all the α-ketoacid dehydrogenase complexes (Hakozaki and Honda, 1990; Mooney et al., 2002); the pyruvate dehydrogenase complex (Luethy et al., 1996), the α-ketoglutarate dehydrogenase complex (Rex Sheu and Blass, 1999), and the branched-chain α-keto acid dehydrogenase complex, as well as the glycine decarboxylase complex (Oliver, 1994). The mitochondrial PDC is the crucial entry point for carbon into TCA cycle in respiratory metabolism (Lernmark and Gardestrom, 1994; Randall et al., 1996). Mutation in the E2 subunit of mitochondrial PDC was shown to reduce overall plant

Chapter 1. General introduction organ size and accumulation of amino acids and TCA cycle intermediates in Arabidopsis (Yu et al., 2012). In contrast, there was no obvious morphological phenotype demonstrated in the T-DNA knockout mutant of one E3 gene in Arabidopsis which suggested a functional redundancy between the two mitochondrial E3 genes (Lutziger and Oliver, 2001).

1.3.2 Citrate synthase (CS)

In the first reaction of the TCA cycle, citrate synthase (CS; EC 4.1.3.7) uses oxaloacetate and acetyl CoA to produce citrate. Plant mitochondrial citrate synthase was first characterized in mature leaves and young flower buds of potato (Landschutze et al., 1995) and it was suggested that it has an important role in floral development (Landschutze et al., 1995). This enzyme was also believed to act as a carbon skeleton source for nitrogen assimilation (Sienkiewicz-Porzucek et al., 2008). Citrate synthase has been long associated with root exudation where citrate secreted by the root acts as an important mediator of both uptake of phosphate and tolerance to aluminium in the soil (delaFuente et al., 1997; Kochian et al., 2004). Eucalyptus plants with over-expressed mitochondrial citrate synthase have been shown to have improved plant growth under low phosphorus stress, most likely due to the increased phosphate-acquisition caused by enhanced citrate excretion (Suzuki et al., 2004).

1.3.3 Aconitase (ACO)

Citrate is converted to isocitrate via a reversible hydration of cis-aconitate, a reaction catalysed by aconitase (ACO; EC 4.2.1.3). The active enzyme contains an iron-sulfur cluster (4Fe-4S) that is formed with three cysteine residues. In animals, this cluster was shown to be sensitive to conditions where iron concentration is low, oxidative stress is occurring or nitric oxide is present, causing a loss of its enzymatic function and a change to become an RNA-binding protein (Hentze and Kuhn, 1996). To date, there are three aconitase isoforms encoded in the

Chapter 1. General introduction

Arabidopsis genome. Two mitochondrial isoforms namely mACO2 (At4g26970) and mACO3 (At2g05710) have been previously identified and isolated from mitochondrial extracts (Sweetlove et al., 2002; Heazlewood et al., 2004) while the third isoform resides in the cytosol (Arnaud et al., 2007; Bernard et al., 2009; Ito et al., 2011). Tomato plants with lowered cytosolic and mitochondrial aconitase protein expression and enzyme activity exhibited increased photosynthetic sucrose synthesis and fruit yield (Carrari et al., 2003). Arabidopsis mitochondrial and cytosolic aconitases knockout plants and tobacco plants with decreased activities of these enzymes demonstrated significantly less chlorosis when subjected to superoxide generating compound such as paraquat treatment, suggesting these mutant plants are more tolerant to oxidative stress (Moeder et al., 2007). Very recently, aconitase was shown to exert respiratory control in animal tissues and its

- inactivation may provide a control mechanism to prevent superoxide (O 2) and hydrogen peroxide (H2O2) formation by the respiratory chain (Scandroglio et al., 2014).

1.3.4 Isocitrate dehydrogenase (IDH)

NAD-dependent isocitrate dehydrogenase (IDH; EC 1.1.1.41) reduces isocitrate to 2- oxoglutarate. In plants such as Arabidopsis thaliana and tobacco, IDH has been shown to be a heteromeric enzyme comprised of at least one catalytic and one regulatory subunit (Lancien et al., 1998; Lemaitre and Hodges, 2006). The Arabidopsis genome contains five genes encoding different IDH subunits (Lin et al., 2004; Lemaitre and Hodges, 2006). Earlier research in tobacco suggested that IDH could provide carbon skeletons for ammonium assimilation (Lancien et al., 1999). Also it has been proposed that there could be a role of NAD-IDH in N-assimilation due to the co-localization of NAD-IDH protein with NADH-GOGAT in the epidermis and the exodermis of rice root supplied with ammonium (Abiko et al., 2005). However, IDH-II knockout mutant in Arabidopsis with 55% reduction on IDH activity did not differ to wild type in terms of mitochondrial respiration or leaf photosynthesis, most likely due to functional substitution by the other isoform (Lin

Chapter 1. General introduction et al., 2004). In another study in Arabidopsis, different IDH genes were mutated giving idhv, idhi, and idhii mutants with leaf IDH activities reduced by 92%, 60%, and 43% respectively. These mutants were found to be invariant beside from reduction of certain free amino acid in comparison to wild type (Lemaitre et al., 2007). Mild reductions of mitochondrial NAD-dependent IDH activity in transgenic tomato plants result in a significant increased nitrate and protein levels but no implications for plant growth (Sienkiewicz-Porzucek et al., 2010). Recently, it was reported that a substantial decrease in Arabidopsis leaf NAD-and NADP-dependent IDH had not resulted in any changes in respiratory flux but instead in a proposed metabolic bypass for 2-oxoglutarate production involving lysine synthesis and degradation (Boex-Fontvieille et al., 2014). Thus, these studies collectively showed that NAD- dependent isocitrate dehydrogenase does not appear to be limiting for nitrogen assimilation in plants.

1.3.5 2-oxoglutarate dehydrogenase (OGDH)

The enzyme complex of 2-oxoglutarate dehydrogenase is responsible for the conversion of 2-oxoglutarate to succinyl-CoA, producing NADH and CO2 as by- products. Similarly to pyruvate dehydrogenase, this enzyme complex contains three enzymatic components namely 2-oxoglutarate dehydrogenase (E1; EC 1.2.4.2), dihydrolipoamide succinyltransferase (E2; EC 2.3.1.61) and lipoamide dehydrogenase (E3; 1.8.1.4) (Millar et al., 1999). The E1 subunit catalyses the decarboxylation of 2-oxoglutarate, the E2 subunit transfers the succinyl group to CoA via a lipoic acid moiety and, the E3 subunit of OGDH catalyses the reduction of NAD+ to NADH using electrons gathered from the lipoamide cofactor of the E2 subunit of OGDH. The E1 component is the initial, substrate-specific and irreversible reaction and was believed to limit the overall reaction rate of the enzyme complex due to its lower catalytic activity, in means of turnover rate, in comparison to the other two components (Waskiewicz and Hammes, 1984). This is consistent with the antisense inhibition of the E1 subunit of this enzyme complex in tomato plants that revealed substantial reduction in respiration rate, early flowering with expedited

Chapter 1. General introduction fruit ripening, and the earlier onset of leaf senescence (Araujo et al., 2012). Previously, this enzyme was shown to be a limiting factor for respiration and to have an essential role in nitrogen assimilation in potato tubers (Araujo et al., 2008).

1.3.6 Succinyl-CoA synthetase (SUC)

Succinyl-CoA synthetase is also known as succinyl-CoA ligase (SUC; EC 6.3.1.5). This enzyme catalyses the inter-conversion of succinyl-CoA to succinate, concomitantly with the synthesis of ATP from ADP and Pi. Succinyl-CoA synthetase has α and β subunits which are believed to be catalytic and regulatory subunits, respectively (Ryan et al., 1997; Lambeth et al., 2004). In tomato plants, there was a subtle decrease in photosynthesis and respiration rate in tomato leaves with succinyl-CoA ligase enzyme activities of 26%, 50% and 91% in antisense transgenic line and two RNAi mutants, respectively (Studart-Guimaraes et al., 2007). The metabolite and transcript analyses of those transgenic lines suggested the up-regulation of GABA shunt as an alternative pathway for succinate production to compensate for the decreased activity of succinyl-CoA. Aside from this, increased succinyl-CoA ligase abundance has been associated with its role in providing energy via ATP production in the TCA cycle to grapevine responding to water deficiency (Cramer et al., 2013).

1.3.7 Succinate dehydrogenase (SDH)

Succinate dehydrogenase (commonly referred to as Complex II) (SDH; EC 1.3.5.1) has a central role in mitochondrial metabolism, as a component in both the TCA cycle and mitochondrial electron transport chain (mETC). This enzyme catalyses oxidation of succinate to fumarate in the TCA cycle and converts ubiquinone to ubiquinol in the mETC, concomitantly with the reduction of FAD to FADH2. In eukaryotes, SDH contains four classical subunits (subunit 1-4) but an additional 4 hydrophilic subunits (subunit 5-8) are found in the plant mitochondrial SDH complex (Huang and Millar, 2013). SDH1 is a flavoprotein containing a covalently bound FAD, SDH2 is an iron-sulfur (Fe-S) protein, SDH3 and SDH4 are two integral

Chapter 1. General introduction membrane proteins that anchor the SDH1-SDH2 sub-complex to the matrix side of the inner membrane (Yankovskaya et al., 2003). The functions of the additional plant specific subunits of SDH (SDH 5-8) remain unknown but collectively they have been hypothesised to provide a secondary peripheral activity of the SDH complex in plants (Millar et al., 2004). Gene knockout mutants of SDH1-1 (At5g66760) showed disrupted gametophyte development while knockout of SDH1-2 showed no obvious phenotype on Arabidopsis growth and development (Leon et al., 2007). A heterozygous SDH1-1/sdh1-1 plant and partially silenced SDH1 mutant plant using

RNA interference in Arabidopsis showed higher CO2 assimilation rates and enhanced growth for both mutants compared to wild-type plant (Fuentes et al., 2011). Knockdown of an assembly factor of SDH1 named SDHAF2 (At5g51040) showed reduced SDH activity, retarded root growth but did not affect leaf growth with unaltered photosynthetic rate and stomatal conductance (Huang et al., 2013). The sdhaf2 mutant demonstrated decreased root elongation rate and root tip respiration rate as well as seed abortion during early seed development (Huang et al., 2013). SDH2-1 and SDH2-2 knockout mutants in Arabidopsis did not give any phenotype suggesting these genes are functionally redundant (Roschzttardtz et al., 2009). However, mutation on the third SDH2 gene isoform (SDH2-3) caused delayed seed germination, suggesting its specific role during embryo development and early germination (Roschzttardtz et al., 2009).

1.3.8 Fumarase (FUM)

Fumarase (FUM; EC 4.2.1.2) (also known as fumarate hydratase) catalyses the reversible hydration of fumarate to malate. Two fumarase genes (FUM1 and FUM2) were reported which encode for a mitochondrial (At2g47510) (Heazlewood and Millar, 2005) and cytosolic (At5g50950) isoenzyme (Pracharoenwattana et al., 2010). It was suggested by Chia et al. (2000) that fumarate, which often accumulates in Arabidopsis leaves, which could act as a carbon sink similar to starch and soluble sugar, being used as an energy source and carbon skeleton for other plant biosynthesis processes. Antisense inhibition of mitochondrial fumarase

Chapter 1. General introduction

(FUM1) in tomato reduced the total cellular activity of fumarase and result in reduced transpiration rate and stomatal conductance, impaired photosynthesis and decreased biomass on a whole plant basis (Nunes-Nesi et al., 2007). Transgenic tomato plant exhibiting reduced expression of mitochondrial fumarase demonstrated reduced root dry mass and respiration rate (Van der Merwe et al., 2009).

1.3.9 Malate dehydrogenase (MDH)

The last step of the TCA cycle is inter-conversion of malate to oxaloacetate (OAA) with concomitant reduction of NAD+ to NADH catalysed by malate dehydrogenase (MDH) (MDH; EC 1.1.1.37). The in vitro activity of mitochondrial MDH was found to favour the reduction of OAA to malate than the oxidation of malate to OAA in watermelon (Walk et al., 1977) and maize (Hayes et al., 1991). Mitochondrial MDH was also shown to be involved in photorespiration, by providing a supply of NAD+ for glycine decarboxylase for the conversion of glycine to serine during the reduction of OAA to malate in the TCA cycle (Journet et al., 1981). In plants, OAA is reduced to malate via the mitochondrial MDH reaction prior to a decarboxylation step by NAD-malic enzyme to form pyruvate and supply CO2 for fixation in bundle sheath cells of chloroplast (Hatch and Osmond, 1976). Therefore, it is reasoned that MDH has a central role in partitioning of carbon and energy, linking respiration with photosynthesis and photorespiration in plants, in addition to its classical role as the final step of the TCA cycle (Cousins et al., 2008; Tomaz et al., 2010). Further evidence for this is a much higher catalytic constant of mitochondrial MDH towards OAA reduction (3000 s-1) compared to the other TCA cycle enzyme PDC (approximately 5 s-1) (Igamberdiev et al., 2014) and photorespiratory enzyme GDC (less than 10 s-1) (Bykova et al., 2014) in photorespiring tissues. The high activity of mitochondrial MDH would enable a fast equilibrium of the mitochondrial NADH/NAD+ (Igamberdiev et al., 2014). Two mitochondrial MDH isoforms have been identified in Arabidopsis mitochondria, and are referred to as mMDH1 (At1g53240) and mMDH2 (At3g15020) (Heazlewood et al., 2004; Tomaz et al.,

Chapter 1. General introduction

2010). Apart from these two mitochondrial isoforms, MDH1 and MDH2, there are six other MDH gene isoforms present in in Arabidopsis; one plastidial, three cytosolic and two peroxisomal. Mitochondrial MDHs have important roles in oxidation reaction of NADH within the TCA cycle and exchange of reducing equivalents between metabolic pathways in different subcellular compartments (Scheibe et al., 2005; Pracharoenwattana et al., 2007). Mitochondrial MDH serves as an important component of the malate transport system mediated by malate- OAA and malate-aspartate shuttles for the exchange of substrates and reducing equivalents across the mitochondrial membrane (Scheibe, 2004; Nunes-Nesi and Fernie, 2007). Antisense-repression of MDH in transgenic tomato plants reduced the activity of the enzyme and demonstrated enhanced leaf photosynthetic activity and decreased root dry mass (Nunes-Nesi et al., 2005; Van der Merwe et al., 2009). It was previously shown that the disruption of both mitochondrial MDH gene isoforms dramatically reduced total mMDH activity and led to a significantly elevated leaf respiration rate and altered mitochondrial proteome in Arabidopsis plants (Tomaz et al., 2010).

1.4 Mitochondrial electron transport chain (mETC) and oxidative phosphorylation

The reductive potential in the form of NADH and FADH2 obtained from the oxidation of organic compounds in the TCA cycle is used to generate usable ATP via the mitochondrial respiratory electron transport chain (mETC). This involves translocating protons from matrix to the mitochondrial intermembrane space and then coupling their return to the matrix with ATP synthesis (Mitchell, 1961). The mETC and ATP synthesis machinery is made of multi-subunits protein complexes: NADH dehydrogenase (Complex I), succinate dehydrogenase (Complex II), cytochrome c reductase (Complex III), cytochrome c oxidase (Complex IV) and ATP synthase (Complex V). Electrons are transferred from the coenzyme NADH and

FADH2 along the oxidoreductase complexes (Complex I, III and IV) to O2 concomitant with proton translocation across the inner mitochondrial membrane

Chapter 1. General introduction

and the formation of H2O. A proton gradient is eventually formed and used to generate ATP via phosphorylation of ADP by the ATP synthase complex (Complex V). Additional components serve as alternative non-electron pumping or proton- gradient dissipating components in the plant mETC system; including alternative NAD(P)H dehydrogenases, alternative oxidase (AOX) and uncoupling proteins (UCP). Complex I (NADH: ubiquinone oxidoreductase, EC 1.6.5.3) catalyses the first step of respiratory electron transport chain where the oxidation of NADH molecules provides two electrons for the reduction of ubiquinone to ubiquinol (Hirst et al., 2003). This complex has a large hydrophobic arm integral to the inner mitochondrial membrane and a hydrophilic arm that protrudes into the matrix and contains the NADH-binding sites and most of the Fe-S clusters (Remacle et al., 2008). Plant mitochondrial Complex I has an additional domain compared to its homolog from bacteria and eukaryotic lineages which includes carbonic anhydrase-like proteins and is hypothesised to facilitate the provision of mitochondrial CO2 for carbon fixation in chloroplasts (Braun and Zabaleta, 2007). Inhibition of Complex I activity following rotenone treatment did not induce oxidative stress or cell death but resulted in alteration of specific components of mitochondria which affected the growth of Arabidopsis cell culture (Garmier et al., 2008).

Complex II, alternatively known as succinate dehydrogenase, is also part of the citric acid cycle. The feature and functions of this complex are as described in the previous section. Complex III, ubiquinone-cytochrome c oxidoreductase, also known as the cytochrome bc1 complex, contains 11 subunits, 3 respiratory subunits (cytochrome B, cytochrome C1, Rieske protein), 2 core proteins and 6 low- molecular weight proteins. The Complex III catalyses the ubiquinone (Q) cycle, where electrons flow from ubiquinol to cytochrome c concomitant with the formation of a proton gradient across the membrane. Complex VI has cytochrome c oxidase activity and catalyses the last step of the electron transport chain and is regarded as a major regulatory site for oxidative phosphorylation in eukaryotes (Kadenbach et al., 2000). It receives electrons from each of the four cytochrome c molecules and converts molecular oxygen to two molecules of water. It is thought

Chapter 1. General introduction that plant Complex IV consists of 14 subunits, 8 are homologous to subunits obtained from eukaryotes and the other 6 are plant-specific subunits (Millar et al., 2004). Respiratory regulation via Complex IV is often known as cyanide-sensitive cytochrome pathway due to its inhibition by cyanide. Diminished cytochrome c oxidase activity by RNAi gene silencing in mouse cell lines in galactose medium resulted in cell death with a compromised mitochondrial membrane potential and decreased ATP level. This suggested a tight respiratory control of Complex IV subunits over oxidative phosphorylation (Li et al., 2006). Knockout of both cytochrome c (CYTc) genes in Arabidopsis produced embryos with arrested development. In addition, CYTC mutants with reduction in CYTC genes expression levels displayed reduced rosette size and lower respiration rate but increased alternative respiration rate and an overall delayed development in Arabidopsis

+ (Welchen et al., 2012). Complex V is the membrane-bound F1FO type H -ATP synthase in mitochondria that catalyses ATP synthesis using the proton-motive force generated by the substrate-driven electron transfer chain. FO components are made up of 3 α-subunits and 3 β-subunits arranged alternately around a central stalk consists of single copies of γ, δ, and ϵ-subunits (Rees et al., 2009). Dysfunction of ATP synthase can be linked to cytoplasmic male sterility in sunflower as the activity of F1FO-ATP synthase was significantly reduced in a sterile line in comparison to the fertile line (Sabar et al., 2003).

Plant mitochondria contain both alternative internal (NDin) and external (NDex) NAD(P)H dehydrogenases, located in the mitochondrial matrix and intermembrane space respectively. NAD(P)H dehydrogenases oxidise NADH or NADPH and reduce ubiquinone independently from proton pumping (Rasmusson and Agius, 2001; Rasmusson et al., 2004). Uncoupling protein (UCP) is a transport protein residing in the inner mitochondrial membrane which is able to decouple electron transport from oxidative phosphorylation, modulating the electrochemical gradient and dissipating excess energy as heat (Borecky and Vercesi, 2005; Krauss et al., 2005; Vercesi et al., 2006). The non-phosphorylating alternative pathway diverts the electron flow from cytochrome pathway at the ubiquinone pool to the alternative

Chapter 1. General introduction oxidase (AOX) which reduces molecular oxygen to water in a single four-electron transfer step (Day et al., 1991; Day et al., 1995). Succinate and malate were shown to stimulate AOX activity in isolated mitochondria (Wagner et al., 1995) which is due to the organic acids are being converted to pyruvate by malic enzyme, indicating an essential role of pyruvate for AOX activity (Millar et al., 1996). Enhanced respiration via the AOX pathway is often associated with abiotic and biotic stresses, maintaining the mitochondrial signalling homeostasis via regulation of signalling molecule levels (superoxide, nitric oxide and other redox couples) (Vanlerberghe, 2013).

Figure 2: Schematic representation of mitochondrial electron transport chain (mETC) in plant which consists of Complex I (NADH dehydrogenases), Complex II (succinate dehydrogenase), Complex III (ubiquinol-cytochrome c oxidoreductase), Complex IV (cytochrome c oxidase), Complex V (ATP synthase), cytochrome c (Cyc C), ubiquinone pool (UQ), external NAD(P)H dehydrogenase (NDex), internal AD(P)H dehydrogenase (NDin), alternative oxidase (AOX) and uncoupling protein (UCP).

1.5 New insights into Arabidopsis thaliana ecotypes

Ecotypes or natural accessions are generally defined as distinct races of a species genetically adapted to particular habitats (Briggs and Walters, 1997). Due to a selfing capability of Arabidopsis thaliana, their ecotypes are normally homozygous and represent inbred lines in nature (Lawrence, 1976). Wild Arabidopsis thaliana occurs naturally in many different habitats throughout Eurasia, and they vary in

Chapter 1. General introduction traits such as leaf shape, flowering time, disease resistance and seed dormancy (Wyatt and Ballard, 2007). Due to its wide geographical distribution, a considerably variability and plasticity in genes and physiological responses for adaptation to different climatic conditions can be expected (Hoffmann, 2002). Therefore they are well-suited for natural genetic variation studies, serve as an important source of genetic variation which provides unique knowledge from functional, ecological, and evolutionary perspectives as well as an insight into control of important plant processes (Koornneef et al., 2004). A 1001 Genome Project to sequence geographically diverse A. thaliana ecotypes was initiated in 2010, aiming to identify the genetic variation contributing to adaptation to diverse environments (Cao et al., 2011). Natural variation of A. thaliana has been exemplified and exploited as an ideal model for several studies like flowering time (Stinchcombe et al., 2004; Balasubramanian et al., 2006), root exudation patterns (Micallef et al., 2009), response to organic acids and minerals (Agrawal et al., 2012; Prasch and Sonnewald, 2013), response to cold and heat stresses (Barah et al., 2013; Barah et al., 2013), abiotic and biotic stresses (Prasch and Sonnewald, 2013), response to ultraviolet radiation (Biswas and Jansen, 2012). Chevalier et al. (2004) reported that there were distinct proteomes in eight Arabidopsis thaliana ecotypes which were probably due to the expression of ecotype-specific protein isoforms. Interestingly, those differentially expressed MDH protein isoforms among ecotypes include mitochondrial MDH (At1g53240), cytosolic MDH (At1g04410) and chloroplast MDH (At3g47520). Recently a transcriptomics study on salt-stressed Arabidopsis ecotypes revealed a higher survival rate and lower electrolyte leakage observed in the Sha ecotype compared to the other two ecotypes, Ler and Col. Genes encoding for TCA pathway, hormone metabolism and development as well as stress and defence- related were found to be enriched in the Sha ecotype under saline conditions (Wang et al., 2013). Thus it is likely that Arabidopsis ecotypes exhibiting ecotypic variability in many aspects related to plant growth and development, including TCA cycle respiratory and energy metabolism.

Chapter 1. General introduction

1.6 Aims and approaches of the study

Former studies have provided evidence that the simultaneous knockout of both mitochondrial MDH gene isoforms raises leaf respiration and limits photorespiration (Tomaz et al., 2010). Furthermore, mitochondrial MDH was shown to exert the highest degree of TCA cycle flux coefficient of all eight resident enzymes (Araujo et al., 2012). However there is still a lack of knowledge on the spatial and temporal changes in respiratory metabolism modulated by mitochondrial MDH which are implied by reports based on reverse genetics and study of natural variation in Arabidopsis thaliana. The experiments described in this thesis were carried out to fulfil the following objectives: 1. To determine the impacts of mMDH loss on respiration and the metabolic network in different organs of Arabidopsis under varying growth and development conditions 2. To distinguish if there are different functions of mMDH isoforms, termed MMDH1 and MMDH2 3. To study the correlation between the abundance of mMDH, other TCA cycle enzymes and respiratory rates in respiratory mutants and across natural variation in Arabidopsis thaliana 4. To study the broader impact of mMDH loss on cellular metabolism by integrating omics data gathered in this study

The first two objectives used single and double knockout mutants of mMDH generated by T-DNA insertional mutagenesis in Arabidopsis thaliana ecotype Columbia. Arabidopsis thaliana ecotypes served as a tool to study the natural variation of respiration as outlined in the third project objective. Respiration rates were measured from different organs (leaf, seed and root) of mMDH mutants to examine the impacts of mMDH loss in both autotrophic (Chapter 3) and heterotrophic organs (Chapter 4) of Arabidopsis during different growth and developmental stages. Similarly, in order to explore the natural variation of respiration and associated pathways in Arabidopsis thaliana, respiration rates were

Chapter 1. General introduction measured from 49 Arabidopsis ecotypes to identify significantly slow and fast respiring ecotypes (Chapter 5). Alongside the respiratory data collection, multi- omics studies encompassing transcriptomics, proteomics and metabolomics were conducted. Transcript profiling on MDH gene isoforms in mMDH mutants was carried out using quantitative real-time PCR (qPCR), confirming the gene expression of mMDH isoforms in the corresponding mutants, but also to investigate functional redundancy and compensatory mechanism between isoforms across subcellular compartments. Multiple reactions monitoring (MRM) assays using a triple quadrupole (QQQ) mass spectrometer were exploited in this study to estimate abundance of TCA cycle enzymes in mMDH mutants and Arabidopsis ecotypes. This approach enabled identification and determination of significantly up- and down- regulation of TCA cycle enzymes from mMDH mutants, and helped in deciphering the intricate relationships between mMDH and the rest of the TCA cycle enzymes. Metabolomic profiling of mutants and ecotypes which used gas chromatography mass spectrometry (GC/MS) had generated important information regarding relative changes in primary metabolite abundance across mMDH mutants and Arabidopsis ecotypes for comparative studies. The ultimate aim of this thesis is to consolidate all the physiological and omics data gathered from mMDH mutants and Arabidopsis ecotypes to elucidate the broader functions of mMDH in respiratory metabolism, and to define the role of mMDH in other important plant metabolic pathways such as photorespiration, carbon and nitrogen metabolism as well as mitochondrial redox homeostasis.

Chapter 1. General introduction

Chapter 2:

Multiplex micro-respiratory measurements

of Arabidopsis tissues

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues

Foreword to Study I

Assessment of mitochondrial respiration rate is a commonly used approach by researchers to gain an insight into the metabolic and physiological state of living tissues. During my PhD, Arabidopsis thaliana T-DNA knockout lines and ecotypes had been used to study the role of mitochondrial malate dehydrogenase in respiratory metabolism during plant growth and development. However, doing respiration measurements in this model organism often faced some limitations due to the small-size of its organs and tissues. For instance, the conventional respiration measurement using a polarographic Clark-type oxygen electrode requires a minimum sample fresh weight (20 to 50 mg) to overcome the drift value created in the micro-chamber by the oxygen sonsumption of the electrode itself when a voltage is applied. Therefore, pooling of samples to meet this weight requirement is inevitable and this could create spurious results, mask biological variation within the pooled samples, and limit the scope of studies in Arabidopsis. Therefore, I worked with colleagues to develop a micro-respiratory assay by adapting a commercial technology used for mammalian cell respiration analysis to suit the small-size of Arabidopsis organs and tissues. We have demonstrated that the respiration measurement with this newly adapted method could be performed on a single leaf disc, root section or seed of Arabidopsis. The data presented in this study exemplified assessment of a more defined respiratory rate in a single tissue specimen and minimised bias from pooled samples.

Author contributions: The work presented in this published paper arose from dicussions with my supervisors, Millar A.H. and Ströher E. and involved the adaptation and modification of methods to use this respiratory system by myself and the other authors. I conducted and analysed the dark respiration rate measurements during leaf development using both XF96 analyser and Clark-type oxygen electrodes (Figure 1 and 2; Supplementary figure 1-4); S. Huang measured and analysed the spatial root respiration rate on root tips and expanded regions

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues using XF96 analyser (Figure 3); while measurements and analysis of respiratory kinetics during seed germination were carried out by C. Holzmann (student of X. Jordana (Figure 4). N. Taylor developed a program in 96-well robotic liquid handling station (Bravo; Agilent Technologies, Mulgrave, VIC, Australia) for convenient and uniform XF 96-well microtiter plate preparation. S. Huang, N. Taylor, C. Holzmann and myself contributed and wrote the respective experimental findings in the manuscript. The manuscript was revised by Ströher E. and finalised by Millar A.H.

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues

Multiplex micro-respiratory measurements of Arabidopsis tissues

Yun Shin Sew 1,2, Elke Ströher 1,2, Cristian Holzmann 1,3, Shaobai Huang 1,2, Nicolas L. Taylor1,2, Xavier Jordana3, A. Harvey Millar 1,2

1ARC Centre of Excellence in Plant Energy Biology and 2Centre for Comparative Analysis of Biomolecular Networks (CABiN), Bayliss Building M316, The University of Western Australia, 35 Stirling Highway, Crawley WA 6009, Western Australia, Australia. 3Millenium Nucleus in Plant Functional Genomics, Departamento de Genética Molecular y Microbiología, Facultad de Ciencias Biológicas, Pontificia Universidád Católica de Chile, Casilla 114-D, Santiago, Chile.

*Corresponding author: A. Harvey Millar

ARC Centre of Excellence in Plant Energy Biology and Centre for Comparative Analysis of Biomolecular Networks, The University of Western Australia (M316) 35 Stirling Highway, Crawley, WA, 6009, Australia

Tel: +61 8 6488 7245 Fax: +61 8 6488 4401

e-mail: [email protected]

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues

Abstract

Researchers often want to study the respiratory properties of individual plants or parts of plants, in response to a range of treatments. Arabidopsis is an obvious model for much of this work, however, due to its size it represents a challenge for gas exchange measurements of respiration. The combination of micro-respiratory technologies with multiplex assays has the potential to bridge this gap, and make measurement possible in this model plant species. We show the adaptation of the commercial technology used for mammalian cell respiration analysis to study three critical tissues of interest; leaf sections, root tips and seeds. The measurement of respiration in single leaf discs has allowed the age dependence of the respiration rate in Arabidopsis leaves across the rosette to be observed. The oxygen consumption of single root tips from plate-grown seedlings shows the enhanced respiration of root tips and their time-dependent susceptibility to salinity. The monitoring of single Arabidopsis seeds shows the kinetics of respiration over 48h post-imbibition, and the effect of phytohormones gibberellic acid (GA3) and abscisic acid (ABA) on respiration during seed germination. These studies highlight the potential for multiplexed micro-respiratory assays to study oxygen consumption in Arabidopsis tissues, and opens up new possibilities to screen and to study mutants and to identify differences in ecotype or populations of different plant species.

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues

Introduction

Plant cells rely on mitochondrial respiration for ATP, carbon skeletons for amino acid assimilation and organic acid building blocks for biosynthetic pathways.

Respiration is the principal component in CO2 loss from cells and is a key factor in the assessment of the carbon balance of plants and in defining the factors influencing the plant growth rate (Amthor, 1989). The assessment of the cellular respiration rate therefore provides an important insight into the metabolic activity and physiological state of plant tissues (Lambers, 1985). The respiration rate can be measured noninvasively as gas exchange from the surface of tissues via the monitoring of the rate of O2 consumption or CO2 production. O2 consumption measurements have relied on low throughput and time-consuming gas- or liquid- phase analysis of O2 concentration by polarographic Clark-type oxygen electrodes in closed systems (Walker and Walker, 1987; Hunt, 2003). CO2 production has been measured using gas-phase infra-red gas analysers in closed systems or in differential open system configurations (Hill and Powell, 1968; Hunt, 2003). Micro- electrodes based on polarographic methods have also been used to monitor O2 concentrations inside seeds and siliques (Porterfield et al., 1999) and in root tissues (Armstrong et al., 2000). Recently microelectrodes have even been adapted to measure respiration inside single photosynthetic cells (Bai et al., 2011). However, these miniaturized methods are highly technical, low throughput, require substantial specialization and often involve painstaking adaptation for use on specific tissues of the target plant species.

Arabidopsis has now become the key model for understanding the molecular components of respiration in plants. Most of our recent advances in the understanding of the biogenesis of mitochondria and the retrograde regulation of respiration by intracellular signalling processes has originated from studies in this species (Millar et al., 2011). However, reports of the measurement of the respiration rate of Arabidopsis, and how it is altered when mitochondrial functions are changed, have been limited as a result of two key constraints. First, the small

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues size of many Arabidopsis tissues has limited the options for the use of many conventional gas exchange systems to measure respiration rates (and micro- respirometry, such as that reported in Arabidopsis siliques (Porterfield et al., 1999), is a very specialized field). Second, the lack of high-throughput assay systems has limited the full use of the resources in Arabidopsis biology to assess respiratory phenotypes through the access of a wide range of mutants, ecotypes and tissue types.

The development of analyte-selective fluorophores, which monitor the partial pressure of oxygen, coupled to fibre-optic cables to monitor their fluorescent properties, has opened up new opportunities in respiratory measurements. Fluorophore-based micro-oxygen sensors have been used to monitor oxygen levels inside plant seeds (Borisjuk and Rolletschek, 2009; Ast et al., 2012) and in the root rhizosphere (Rudolph et al., 2012) to study hypoxia. These measurements are of oxygen concentration, not respiration rate, and so diffusion of gases for specific tissues needs to be calculated or standardized for respiration rates to be deduced by time series measurements of oxygen concentration (Rudolph et al., 2012). Coupling fluorophore-based micro-oxygen sensors to microtiter plate assays, for which standardized diffusion rates can be calculated, has allowed the high- throughput analysis of the respiration rate in milligrams of tissue in microliter volumes (Ferrick et al., 2008; Gerencser et al., 2009). Such systems have been commercialised and are now being used to measure cellular respiration rates and cellular bioenergetics of isolated mitochondria and cells from mammalian tissues (Beeson et al., 2010; Rogers et al., 2011; Zhang et al., 2011; Zhang et al., 2012). However, to our knowledge, the adaptation and use of such systems for intact plant tissues has not been tested systematically.

Here, we present optimized methods to adapt the use of commercial microplate assays of oxygen consumption by analyte selective fluorophores to measure the respiration rates of Arabidopsis leaf, root and seed samples. We show that this approach allows high-throughput measurements of the respiration rate in leaf

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues laminar and vascular regions of a single leaf, the respiration rate of single root tips and even the respiration of single imbibed seeds. We illustrate that biological changes in respiration associated with leaf development, leaf age, root segments and hormone-dependent changes in seed germination can be measured and compared. These developments, and the use of commercial systems and consumable packs already optimized and available to researchers, open up opportunities for the in-depth analysis respiratory phenotypes and their relation to developmental processes in small tissue samples from a variety of plants.

Materials and methods

Extracellular Flux Analyser XF96 and 96-well plate set-up

Seahorse XF96 Extracellular Flux Analyser measurement is based on the fluorimetric detection of O2 levels via fluorophores in a commercial sensor cartridge. Oxygen quenches the fluorescence of a fluorescein complex, the fluorescence is detected by a fibre-optic waveguide and converted into the basal oxygen consumption rate (OCR). During the ‘measurement’ phase, the concentrations are measured continuously until the rate of change is linear, and then OCR is determined from the slope. The probes lift whilst in the ‘mixing’ and ‘waiting’ steps to allow the larger medium above to mix with the medium in the transient microchamber, re- oxygenating the solution and thus restoring the oxygen concentration values to baseline. The XF96 data can be visualized and analysed in both XF96 Analyser software and an Excel-based data viewer. For the underlying calculations, the reader is referred to the literature on the development of this system (Ferrick et al., 2008; Gerencser et al., 2009). Respiration measurements were performed in an XF96 Extracellular Flux Analyser (Seahorse Bioscience, Billerica, MA, USA) to obtain the OCR of plant tissues. The 96-well sensor cartridge was hydrated in 200 µl per well of XF Calibrant Solution (Seahorse Bioscience) overnight at 37°C before the assay. Several hours before the measurement commenced, the heater of the

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues instrument was turned off to obtain a stable internal measurement temperature in the machine at c. 28°C. Plates (and injection ports when indicated) were filled using multichannel pipettes or en masse by a 96-well robotic liquid handling station (Bravo; Agilent Technologies, Mulgrave, VIC, Australia) using in-house-developed device and protocol programs.

Leaf respiration rate measurements by XF96

Wild-type seeds of Arabidopsis thaliana (L.) Heynh (ecotype Columbia) were placed on wet filter paper and incubated at 4°C for 3 days. The imbibed seeds were transferred to individual pots containing a 1:3 perlite: soil mix and covered with a transparent acrylic hood to maintain humidity. The seedlings were grown in a controlled environment growth chamber maintaining a short-day photoperiod (8 h : 16 h, light : dark), a photon flux of 150 µmol photons m-2s-1, a relative humidity of 75% and a temperature cycle of 22°C : 17°C, day : night temperature regime. When the seedlings were established, the acrylic hood was removed and the plants were subsequently grown with regular watering. At an age of 4–6 week, as indicated, the plants were used for the measurements.

Single leaf discs were immobilized in wells with either Cell-Tak (BD Bioscience, North Ryde, NSW, Australia) or a commercial skin adhesive Leukosan® (BSN Medical, Mount Waverley, B.C., VIC, Australia) mixed with agarose. For Cell-Tak adhesion, 16 µl of the Cell-Tak mixture, pH 7 (5% (v/v) Cell-Tak, 45 mM sodium bicarbonate, pH 8.0), was used to coat the bottom of each well of the microtiter plate. The absorption of Cell-Tak to the well bottom was allowed for 20 min at room temperature, after which the Cell-Tak mixture was discarded by aspiration before rinsing with distilled water. Single 2.5 mm diameter leaf discs, which had been freshly cut with a leaf punch, were then placed at the center of each well and gently pressed to the well bottom using a cotton bud. An even contact between the leaf disc and the Cell-Tak-coated layer on the bottom of the well was required for optimal adhesion. Adhesion was allowed for 30 min before 200 µl of respiration

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues

buffer (10 mM HEPES, 10 mM MES and 2 mM CaCl2, pH 7.2 (Atkin et al., 1993; Armstrong et al., 2006)) was added to the wells. For the skin-glue adhesive and agarose mixture, a combination of 2.5% (v/v) Leukosan® adhesive in 0.25% (w/v) agarose was prepared and kept above 60°C to avoid solidification. For each well, 1 µl of the adhesive mixture was pipetted onto the center of the well bottom. Then, 2.5 mm diameter leaf discs were positioned on top of the mixture before gentle pressure with a cotton bud. As the adhesive mixture sets in c. 2 min, sequential handling of the samples is required if large numbers of leaf discs are used. After 2 min, respiration buffer can be added on top of the leaf discs to avoid dehydration. A full plate of 96 leaf discs could be manually adhered in c. 45 min. Once leaf adhesion had been achieved, wells were filled with 200 µl of leaf respiration buffer and loaded into the plate reader after the calibration steps. The time events for both basal respiration measurement and injection were mixing (3 min), waiting (4 min) and measurement (5 min). The method allowed for 10 cycles of mixing, waiting and measurement. The OCR of single leaf discs was recorded by Seahorse XF Acquisition and Analysis Software (Version 1.3; Seahorse Bioscience).

Root respiration measurements by XF96

Seeds of Arabidopsis (A. thaliana) ecotype (Columbia-0) were sown on half-strength Murashige and Skoog (MS) Gamborg B5 plates containing 0.8% (w/v) agar, 1% (w/v) sucrose, 1.8 mM MES at pH 5.8 adjusted by KOH. The plates were placed at 4°C in the dark for 2 d and then transferred to a growth room with a photoperiod of 16 h : 8 h, light : dark at a light intensity of 200 µmol photons m-2s-1, relative humidity of 70% and temperature cycle of 22°C : 17°C, day : night. The plates were set in a vertical position. After 7 d of growth, c. 5 mm of the expanded section or elongating root tip were cut for respiration assay with eight replicates for each treatment. The 96-well sensor cartridge was hydrated in 200 µl per well XF Calibrant Solution (Seahorse Bioscience) as mentioned above. After calibration, the 96-well utility plate was filled with 100 µl of respiration buffer containing 0, 100, 200 or 400 mM

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues

NaCl. In each well, a single root tip (tip; c. 5 mm) or root expanded section (EXP; c. 5 mm) was added to the bottom of the well. The time events for both basal respiration measurement and injection were mixing (2 min), waiting (3 min) and measurement (5 min). Seven cycles of mixing, waiting and measurement were applied for time course measurements. The OCR of the single root tip or expanded section was recorded by Seahorse XF Acquisition and Analysis Software (Version 1.3; Seahorse Bioscience).

Seed respiration measurements by XF96

For multiple seed measurements, intact seeds (c. 1 mg) were placed in a 96-well plate and surface sterilized by soaking for 7 min in 12.5% (w/v) NaClO and 0.1%

(v/v) Tween20, followed by two washing steps with distilled H2O. After this, wells were filled with 200 µl of seed respiration medium (5 mM KH2PO4, 10 mM TES, 10 mM NaCl, 2 mM MgSO4, pH 7.2) and loaded into the plate reader after the calibration steps using the Bravo liquid handling station (Agilent Technologies). Where indicated, inhibitors were added to the medium with a final concentration of 2 µM for KCN or 5 mM for salicylhydroxamic acid (SHAM). Oxygen concentrations before and after inhibitor injection were determined by 11 cycles of mixing (3 min), waiting (4 min) and measurement (5 min). The OCR of seeds was recorded by Seahorse XF Acquisition and Analysis software (Version 1.3; Seahorse Bioscience), and each well was normalised by the milligram weight of seeds used.

For single seed measurements, a sterile solution of 0.25% (w/v) agarose was used, and kept at 65°C to avoid solidification. The agarose solution was pipetted (1 µl) into the center of each well bottom. Seeds were sterilized by overnight incubation with chlorine gas (100 ml of 12% NaOCl and 3 ml of 37% HClO) in a closed vessel. Each single seed was placed with a sterile toothpick, making sure the adhesion of each seed was in the center of the well. Then, the wells were filled with 200 µl of seed respiration medium and loaded into the plate reader after the calibration steps. Where indicated, hormones were added to the respiration medium with a

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues final concentration of 2.4 µM for abscisic acid (ABA; PhytoTechnology Laboratories,

Shawnee Mission, KS, USA) and 1.2 mM for gibberellic acid (GA3; Sigma-Aldrich). The respiration measurements were made by mixing (3 min), waiting (4 min) and measurement (60 min). The method was run for 48 cycles, achieving a total of 50 h of measurements. The OCR of single seeds was recorded by Seahorse XF Acquisition and Analysis Software (Version 1.3; Seahorse Bioscience).

Leaf respiration by Clark-type oxygen electrode

Plants were grown under the conditions described for XF96 above. The OCR of leaf discs was measured using a liquid-phase Oxygraph system (Hansatech Instruments, Pentney, Norfolk, UK). Before the measurement, the electrode was calibrated at 25°C by the addition of sodium dithionite to 1 ml of aerated autoclaved water to completely deplete oxygen. Leaf discs totalling 40–60 mg fresh weight (FW) of 7mm diameter leaf discs were immersed in leaf respiration buffer and incubated in the dark for 30 min. Leaf respiration was performed in a 2 ml volume for at least 15 min at 25°C in a darkened electrode chamber. The amount of oxygen being consumed by the leaf discs was recorded using Oxygraph Plus v1.02 software (Hansatech Instruments), and the OCR (nmol min-1 g-1 FW) was calculated accordingly to the FW of the leaf discs.

Statistical analysis

The statistical software package IBM SPPS Statistics 19 (IBM Australia, St Leonards, NSW, Australia) was used for data analysis where indicated. An analysis of variance, followed by multiple comparison using post hoc tests and Tukey’s honestly significant difference (HSD) mean separation test, was performed to determine the statistical significance of differences of the mean values at P≤0.05.

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues

Results

Adhesion of leaf discs for respiratory measurements

Making OCR measurements in microtiter plates of the Seahorse XF96 requires that the tissues remain at the bottom of the well and do not move during the cycles of mixing and measurement. This requirement is not present when using oxygen electrode or infra-red gas analysis techniques, and is much less of a problem when using mammalian tissues as they are not buoyant structures. To develop suitable adhesion techniques, we trialled two different methods: one using mixtures containing 5% (v/v) Cell-Tak (BD Bioscience) and another using 2.5% (v/v) Leukosan® adhesive in 0.25% (w/v) agarose. Both adhesion methods were found to immobilize leaf discs submerged in buffer for several hours. However, the investigation of the effectiveness of the two adhesion mixtures during the course of the mixing assays showed that the Leukosan® adhesive treatment produced far fewer leaf disc detachment events and a lower standard error for OCR

(Supplementary figure 1A ad 1B). Analysis showed that an OCR of 143 ± 11 pmol O2 min-1 (for a leaf disc of c. 0.7 mg FW) could be consistently measured. Replicate leaf discs from the same leaf gave more consistent results than leaf-to-leaf comparisons, suggesting some variability of OCR between leaves (Supplementary figure 1A and 1B). To test the effect of the adhesive on OCR, we performed similar measurements using 7 mm diameter leaf discs in a Clark-type oxygen electrode (Oxygraph; Hansatech Instruments). The mean OCR g-1 FW of leaf discs did not change with increasing amount of leaf discs adhered together during the analysis, indicating no substantial effect of the adhesive on oxygen diffusion that could slow the respiration rate (Supplementary figure 1C). Calculations based on these measurements showed that c. 40 times more leaf tissue is required for an accurate OCR measurement in the typical 1 ml Clark-type oxygen electrode than in the microtiter plate fluorescence assay. All further experiments were performed using the Leukosan®/agarose mixture.

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues

As a result of the need to fix the leaf discs in the wells, the preparation of a full 96- well plate takes c. 45 min. To test whether the order in which the leaf discs are laid down influences the reading, we used different leaf developmental stages, including slow and fast respiring stages, from two plants. Leaves and cotyledons were selected and the two sets of discs were fixed in the wells with a c. 30 min time difference between the sets (Supplementary figure 2). Similar differences in respiration rate between the leaf stages were recorded. To test the dynamic range of the Seahorse XF96 instrument, an experiment was performed using different amounts of leaf tissue. As the leaf discs must be fixed to the bottom of the well, the maximal size of the leaf disc is limited by the diameter of the well, and only one leaf disc can be used. In addition to the leaf disc size used for all the other experiments described here (0.7 mg FW), four additional sizes were employed (Supplementary figure 3A). The graph shows that the OCR increases linearly with increasing tissue amount (R2 = 0.873). In separate experiments using over 230 large leaf discs (1.6

-1 mg FW), individual leaf disc values up to 500 pmol O2 min were measured, the distribution of rates closely resembling a normal distribution (Supplementary figure 3B). As the leaf discs used here are small and have a significant cut surface area to total surface area, an experiment was performed to test for a possible wounding- induced oxygen consumption effect on the readings. The standard procedure to reduce this effect by dark incubation was performed (Azcon-Bieto et al., 1983; Azcon-Bieto et al., 1983; Day et al., 1985). Leaf discs were excised from three individual plants and incubated for 30 min in the dark in respiration buffer, before the measurements were performed (Supplementary figure 4). No significant difference could be detected. All further experiments presented were performed without the 30 min dark incubation before adhering discst to he wells.

Respiration rate across Arabidopsis leaf surfaces

The ability to measure respiration in small leaf discs allowed us to survey the respiration rate of different regions across single Arabidopsis leaves. Nine 2.5 mm diameter leaf discs were excised from three independent mature leaves of 4-week

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues old A. thaliana plants to assess the respiration rate of the lamina left (L), lamina right (R) and mid-rib (M) positions on the leaf blade (Figure 1). The mean OCR of each leaf disc position was assessed by averaging the mean OCR from three different leaves. On a leaf area basis, mid-ribs (M1–3) constantly showed a higher mean OCR than laminar positions, left (L1–3) and right (R1–3; Figure 1B). Comparison of the mean OCR values showed that there were significant differences between mid-ribs (M2 and M3) and both laminar left (L1–3) and right (R1–3; P ≤ 0.05) positions. On a weight/volume basis, mid-ribs and lamina discs varied significantly, with c. 1.4-fold higher average FW of mid-rib leaf discs. As a result, mid-ribs exhibited a lower mean OCR than laminar leaf discs on a weight basis (Figure 1C). Statistically significant differences between laminar left (L1–3) and mid- rib (M1 and M2) disc positions (P≤0.05) were apparent in the data. This indicates that, where a leaf disc is cut across the Arabidopsis leaf surface, this can influence the OCR measured. The data also showed the consistency of measurements along the leaf blade for laminar and vascular regions.

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues

Supplementary figure 1. Oxygen consumption rate (OCR) of single leaf discs adhered to the bottom of a XF96 cell culture microplate well using (A) 5% (v/v) Cell- Tak and (B) 2.5% (v/v) Leukosan® Adhesive and 0.25% (w/v) agarose. Measurements made in a Seahorse XF96 analyser were compared. Values represent the mean OCR from 10 successive measurements of OCR for individual leaf discs in a single assay. The standard error (±SE) within the technical replication of OCR is given. A total of 22 independent leaf discs were tested for their OCR using each adhesion method. Leaf discs depicted as ‘a’ and ‘b’ were from the same leaf and the positions were close to each other. When leaf discs detached from the bottom of the well, measurement was disrupted and is indicated by nd. In the column. (C) The effect of leaf disc adhesion with 2.5% (v/v) Leukosan® Adhesive and 0.25% (w/v) agarose on the OCR by using an Oxygraph Clark Type electrode (Hansatech). Mean OCR ± SE is shown (n=4) on a leaf FW basis is shown. Each measurement was made with four leaf discs in the Clark type electrode. The effect of gluing none, two, three and all four discs together was evaluated.

Supplementary figure 2. Investigation of differences in respiration rates of leaves from two 6-week old Arabidopsis thaliana plants when the tie discs were adhered to the plate differed by 30 min. Leaf discs were excised from cotyledon, mature and young leaves of two plants (A and B) and adhered using mixture of 2.5% (v/v)

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues

Leukosan® adhesive in 0.25% (w/v) agarose to a XF96 microtiter plate. The value represent mean OCR ± SE (n=6). All samples were analysed together at the same tie on the one plate. In each case similar differences in respiration rate were found between the three leaf ages.

Supplementary figure 3. (A) Respiration rates of different sizes of leaf discs excised from mature leaves from 7.5-week old Arabidopsis thaliana plants. Leaf discs were adhered using a mixture of 2.5% (v/v) Leukosan® adhesive in 0.25% (w/v) agarose to the bottom of the XF96 microtiter plate. The value represents mean OCR (n=6, mean±SE). A linear correlation between respiration rates and fresh weight of leaf discs was performed and R2 is reported. (B) Distribution of a series of XF96 leaf respiration data from different aged leaves (solid line) and normal distribution of random data generated within the same parameters (dashed line). Comparison of both sets of data indicated that the measured respiration data are close to being normally distributed.

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues

Supplementary figure 4. Analysis of respiration rates on mature leaves of 4 week old Arabidopsis thaliana plant with or without dark incubation prior to the measurement. Three pairs of leaf discs were excised. Three sets of paired discs (1-3) were excised from positions close to each other in separate leaves. Paired discs were separated and either incubated in leaf respiration buffer in the dark for 30 min (a) or adhered without dark incubation (b). All leaf discs were adhered to the bottom of wells in the XF96 microtiter plate using a mixture of 2.5% (v/v) Leukosan® adhesive in 0.25% (w/v) agarose. The bars represent mean OCR from the first 30 min of measurement (grey bar) and from the second 30 min of measurement (dark grey bar) (n=3, mean±SE). There was no consistent significant difference of means between the control and dark incubation treatments (p>0.05).

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues

Figure 1. Survey of Arabidopsis leaf blade respiration rate excised from individual mature leaves of 4-week old Arabidopsis thaliana plants. (A) The disc positions tested are depicted in the vertical (1, 2 and 3) and horizontal (L, left; M, mid-ribs; R, right) axes. (B) Respiration rates on a leaf area basis. (C) Respiration rates on a leaf weight basis. The values represent the mean oxygen consumption rate (OCR; n=3; mean±SE). *, Significant difference (P ≤ 0.05) between M1–3 and the L1–3 and R1–3 bars.

Respiration rates in Arabidopsis leaves of different sizes and ages

To gain further insight into the effect of leaf age and leaf size on leaf OCR, assays on leaves across the rosette of 4- and 6-week-old plants were performed (Figure 2). The growth of A. thaliana plants was observed from when the cotyledons first started to expand. The sequence of subsequent leaf development was systematically recorded and all leaves were tagged for the final analysis phase. The OCR from each leaf was measured simultaneously in the microtiter plate assays to avoid any differences associated with time of day or time from leaf harvest. The data showed that OCR increased gradually from mature to immature leaves (linear R2 = 0.81, polynomial R2 = 0.85 at 4 week; and linear R2 = 0.63, polynomial R2 = 0.84 at 6-week), although there were also some fluctuations spanning across leaf age.

-1 -1 The median OCRs were 155 pmolO2 min per disc and 233 pmolO2 min per disc for 4- and 6-week old plants, respectively. Interestingly, the peak OCR in leaf 13, initially noted in 4-week old plants, was maintained at 6 week. After this point in development, new leaves appear to retain the same higher rate of respiration as leaf 13.

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues

Figure 2. The effects of development, leaf age and leaf size on the oxygen consumption rate (OCR) of Arabidopsis thaliana leaves: (A) 4-week old plant and (B) 6-week old plant grown under short-day conditions. The values represent the mean OCR (n=4; mean± SE). The yellow lines indicate the calculated median OCR and a colour scale were created on the basis of the median for each plant age. The plant rosette and the size of each leaf are shown in the images marked with leaf numbers. A colour scale assigned on the basis of the calculated median aids the visualization and comparison between the size, developmental stage and rosette position of each leaf and its OCR value. Linear and polynomial lines of best fit are shown and R2 values are reported (linear, blue; polynomial, red).

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues

Respiration rates of root tips and expanding regions

Root growth on plates is commonly measured as a phenotype of Arabidopsis mutants and in assays analysing chemical effectors and nutritional responses (Migliaccio and Piconese, 2001; Oliva and Dunand, 2007). However, the very small mass of Arabidopsis roots often precludes biochemical measurements at the single root level. The differential rate of respiration in the growing tip and in the previously expanded regions is of interest, as it is considered to be an important factor in determining the root growth rate (Hanbury and Atwell, 2005). The OCRs of single root tips and single 5 mm sections of expanded roots were found to be sufficient to make accurate measurements using micro-respiratory techniques (Figure 3A). The data showed that OCR was three times higher in root tips than in expanded root regions (Figure 3B). The treatment of plants with NaCl has been reported to stimulate or inhibit the respiration of roots depending on the species studied (Jacoby et al., 2011). Treatments for only 10 min with 100 mM or 200 mM NaCl led to no significant change in respiration rate in our assays. By contrast, 400 mM NaCl for 10 min halved the respiration rate of single root tips (Figure 3B). However, a time course of the respiratory response showed that 200 mM NaCl lowered the respiration rate over the first hour of treatment, whereas 400 mM NaCl stopped the respiration rate in root tips in the same time frame (Figure 3C). This shows that, for the lower salt concentrations, time-dependent effects can be monitored using this respiration assaying system.

Respiration rates of Arabidopsis seeds and the respiratory response during germination and hormone treatments

The kinetics of respiration in seeds during germination has been studied in a variety of species, but is difficult in Arabidopsis because of seed size. Using 1 mg of Arabidopsis seeds, we measured the initiation of respiration during the first 60 min post-imbibition, and recorded a four-fold rise in OCR (Figure 4A). The respiration of seeds could be inhibited significantly by the simultaneous injection of the

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues respiratory poison KCN into the microtiter plate assays. The addition of the alternative respiratory pathway inhibitor SHAM failed to further inhibit OCR. This could either be a result of the difficulty of this compound in entering seeds or a lack of a significant alternative pathway rate early in the seed germination process. Previous studies have shown that alternative oxidase is induced during the second 24 h post-imbibition in Arabidopsis seeds (Narsai et al., 2011). To confirm that the OCR rise observed during this first hour is the initiation of respiration, we performed a study of control seeds and two seed treatments, one treatment involving pre-imbibition for 100 min and the other a 100°C heat treatment for 1 h (Figure 4B). Pre-imbibed seeds immediately attained an OCR similar to the maximal rate over the 120 min of the experiment. Control seed OCR rose to this value over the first 40 min. Heat-treated seeds did not respire during the 120 min period.

By extending the time period for each respiratory measurement from 5 to 60 min (as outlined in Materials and Methods), we were able to modify the OCR assay to allow the measurement of the OCR for single seeds throughout the first 48 h post- imbibition. These assays showed that there are several phases of OCR during this 48 h period, beginning with a steady rise over the first 24 h, followed by a slowing of the rate of acceleration of OCR, and a subsequent rise in rate between 30 and 40 h post-imbibition (Figure 4C). The addition of the germination-stimulating hormone

GA3 increased the respiration rate during this 48-h period, but without any clear change in respiration kinetics. To determine whether abscisic acid (ABA) had a contrasting impact, we repeated this 48-h study and compared control seeds with ABA-treated seeds. Respiration of ABA-treated seeds was similar to that of untreated seeds for the first 2–3 h; OCR then remained constant until 12 h post- imbibition, but finally declined over the remaining time in the assay (Figure 4D). These ABA-treated seeds did not visibly germinate in the 96-well plates, whereas the seeds that were not treated germinated normally during the measurement.

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues

Figure 3. Respiration rates of single expanded region (EXP) and tip (TIP) of a 7-day old root of Arabidopsis thaliana seedlings. Plants were grown on agar plates under long-day conditions. (A) Single c. 5 mm sections of the root expanded region and root tip were used for each respiration assay. (B) Respiration rates of single root expanded region and root tip with or without different NaCl treatments for 10 min (n=5–8; mean±SE). (C) Time course of respiration rates of root tips treated with different NaCl concentrations (n=5–8; mean±SE).

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues

Figure 4. Respiration rates of Arabidopsis thaliana seeds. (A) Respiration rate of the first 110 min post-imbibition of 1 mg of Col-0 seeds. Vertical lines indicate the time of addition of KCN (2 µM) and salicylhydroxamic acid (SHAM) (5 mM; n=8; mean±SE). (B) Respiration rate of 1 mg of Col-0 seeds which were untreated and

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues assayed directly on imbibition (control), incubated in buffer at room temperature for 100 min before measurements (pre-treated) or heated at 100°C for 1 h in a buffer solution before measurement (heat-treated; n = 8; mean±SE). (C) Respiration rate of single untreated Arabidopsis seeds and seeds incubated in 1.2 mM gibberellic acid (GA3). Each seed was fixed to the center of the well with 0.25% (w/v) agarose (n=14, mean±SE). (D) Respiration rate of single untreated Arabidopsis seeds and seeds incubated in 4 µM abscisic acid (ABA) Each seed was fixed to the center of the well with 0.25% (w/v) agarose (n=20, mean±SE).

Discussion

Technical limitations and advances for OCR measurements of plant tissues

For decades, researchers have been using Clark-type oxygen electrodes or infra-red gas analysers to measure the respiration rate from Arabidopsis cells and tissues (Noren et al., 1999; Hunt, 2003; Williams et al., 2008; Tomaz et al., 2010; Yang et al., 2011). In order to overcome the impact of the baseline drift value (c. 0.2 nmol min-1 in a typical 1 ml Clark-type oxygen electrode) and the differential needed between reference and sample gas streams in infra-red gas analyser measurements

(>5 ppm CO2 for accurate respiratory measurements), a minimum of 20–50 mg of plant tissue is normally needed for a single assay to avoid spurious results (Hunt, 2003; Meyer et al., 2009; Tomaz et al., 2010). As Arabidopsis tissues are much smaller in size than many other species used in plant research, the pooling of samples from different biological replicates has usually been required for respiratory measurements. This is not ideal and has limited the accuracy of studies that have focused on specific tissues at certain developmental stages. Because of this, it is not surprising that many reports find little if any differences in whole- tissue OCR between genotypes and/or treatments of Arabidopsis plants.

Fluorescence-based dispersed measurement of OCR in multi-well plate format offers high-throughput respirometry with a greatly decreased sample size requirement for each assay. We have shown that c. 40-fold less leaf tissue (FW c. 1 mg) can be used in a similar time frame to other assays (<60 min). Through an extension of the time of methods, even single seeds can be assayed for their OCR.

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues

This approach allows for high sensitivity in OCR detection, a greater number of respiratory data points and extremely low sample mass requirements, which will be especially useful for respiratory studies of scarce biological samples from plants.

A significant issue for the use of the microtiter plate OCR assays in the Seahorse XF96 is the need to secure material during the mixing and measurement phases. This is especially problematic for plant leaves as they are gas-filled structures, and so their buoyancy needs to be overcome for an extended period of time and during the addition and mixing phases of the assays. Two different methods were tested to immobilize leaf discs onto microtiter plate bases with differing success. Cell-Tak (BD Bioscience) is a formulation of multiple polyphenolic proteins extracted from the blue mussel Mytilus edulis (Silverman and Roberto, 2007). Researchers have been using this adhesive protein mixture to immobilize animal cells and tissues for microplate assays for a number of years (Choi et al., 2010; Zhang et al., 2011; Robinson et al., 2012). However, Cell-Tak is expensive and we found that it took c. 30 min to adhere, leading to dehydration of leaf tissues which is undesirable. Cell- Tak also had a significant failure rate across wells in securing leaf tissues (c. 20% failure, Figure S1). A much lower cost and more rapid solution was the use of medical-grade skin-glue (Leukosan®), which is nontoxic, sets in c. 2 min and, when mixed with agarose, provided an excellent adhesive for leaf tissues to plastic surfaces (<5% failure, Figure S1). The agarose also provided aeration on the side of the leaf disc in contact with the plastic, as agarose has a gas-permeable macroporous structure with pore sizes of 100–300 nm (Plieva et al., 2009). Larger scale multiplex assays using most or all of the 96 positions on a plate could be adhered, covered with respiration buffer and ready for assay by the Seahorse XF96 in c. 45 min using the agarose plus skin-glue method. The respiration rate was not greatly influenced by the order in which the samples were loaded (Figure S2), or by wounding effects (Figure S4), and it could be conducted over a dynamic range of c.

-1 20 to 500 pmolO2 min . Direct comparison of the readings for leaf discs from the Clark-type oxygen electrode and the Seahorse XF96 revealed overall higher values from the micro-respiratory technology in our hands. The discrepancy can be

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues explained by various factors. The Clark-type oxygen electrode is a closed system, whereas theSeahorse technology is based on a semi-closed measuring environment which requires a range of diffusion calculations to be undertaken (Gerencser et al., 2009). As this device was developed for mammalian cell lines, it is equipped with a heater to ensure an optimal temperature of 37°C. Cooling is not possible and the lowest possible temperature in room temperature conditions is reported by the device as c. 28°C. A higher temperature leads to an increased respiration rate and could also contribute to the differences noted. Based on our experiments, we recommend the use of this technology to detect relative changes within a single plate or different plates using the same method. Comparisons between plates using different methods (e.g. measurement time) and between fluorescence-based micro-respiratory and Clark-type electrode assays tend to yield differences in absolute rate which are difficult to account for precisely, but show similar relative differences between biological samples.

Variations in leaf respiration rate across development

The architecture of leaf structures is closely related to their function, and thus is an important determinant of the primary productivity of plants (Fosket, 1994). Our results revealed that the OCR of the mid-rib vascular region is different from that of the lamina of Arabidopsis leaves on both a leaf area and leaf weight basis. The key physiological and structural differences between the lamina and mid-rib have been well addressed in leaves (Sylvester et al., 1996; Nelson and Dengler, 1997). Most fundamentally, this has shown that the ratio of spongy mesophyll to palisade is greatest in the mid-rib portion of the leaf and steadily decreases towards the leaf margin. Comparative data analysis of mitochondrial density in Arabidopsis tissue has shown that there is approximately half the mitochondrial volume (µm3 µm-3 tissue) in spongy mesophyll tissue than in palisade tissue (Armstrong et al., 2006). A relatively sparse distribution of mitochondrial number in a higher cell volume could explain the mid-rib to lamina differences in OCR observed here. Tschiersch et al.

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues

(2012) used fluorescence measurements of oxygen concentration to image leaves, and noted that the concentration in intercostal regions of the leaf blade declined faster than in veins, and concluded that oxygen distribution was aligned to the structure in the leaf. This could be interpreted to mean that OCRs were faster in intercostal areas of the leaf (similar to our lamina leaf discs) relative to the veins (similar to our mid-rib region leaf discs); therefore, findings from both leaf discs and leaf imaging are in agreement. Our data were consistent with a general trend of an increase in respiration rate from mature to immature leaves, independent of leaf size. Regression analyses indicated a relatively strong correlation between the two sets of variables in the plants tested (R ≥ 0.60). These data suggest that leaf aging changes the respiration rate in Arabidopsis. Jeong et al. (2004) showed this in aspen leaves, where OCR decreased by >50% from young leaves to mature leaves. In Arabidopsis, immature, partially expanded leaves have been reported to show significantly higher rates of respiration compared with mature fully expanded leaves (Armstrong et al., 2006b). Our data provide a high-definition dataset showing the timing and extent of this phenomenon across the rosette. The reason for this difference most probably resides in a combination of mitochondrial number in leaves and metabolic demands in different leaves. The respiratory process is thought to assimilate nearly half of the total carbon gained from the photosynthesis process (Mogensen, 1977; Lambers, 1985; Amthor, 1989) and its consequence losses are equally shared between growth and maintenance processes during developmental stages (Amthor, 1984; Lambers, 1985). Growth respiration provides energy for the synthesis of new tissue throughout the developmental process, whereas maintenance respiration generates energy to be used for the synthesis of essential substances for existing tissues and metabolites for the survival and adaptation of plants under various environmental conditions (Lambers, 1985; Amthor, 1989). Previous findings have shown that the cost of maintenance respiration is comparable with the cost of growth in herbaceous plants, such as Arabidopsis. Once plant tissues reach maturation, the growth rate and respiration

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues slow, and energy obtained from respiration mainly goes towards maintenance and transport processes (Amthor, 1984).

Spatial variation in root respiration rate

In this study, we showed that the small root tips of Arabidopsis have a nearly three- fold higher OCR when compared with a section of expanded root (Figure 3). This is consistent with the expected higher energy demand in root tips, required for elongation, than in the expanded region of roots, or could relate to smaller vacuoles in the root tips. In Arabidopsis, mitochondrial mutants in the Lon1 protease (Solheim et al., 2012), in the membrane chaperone prohibitin (Van Aken et al., 2007) and in Complex I subunits (de Longevialle et al., 2007; Meyer et al., 2009) have short roots. To our knowledge, there is no precise information on the rate of respiration required to maintain root growth in Arabidopsis. However, we have reported recently that succinate dehydrogenase assembly factor 2 (sdhaf2) is needed for the assembly and activity of mitochondrial Complex II and for normal root elongation in Arabidopsis (Huang et al., 2013). Whole-root respiratory assays showed no difference between wild type and sdhaf2, but micro-respiratory measurements of root tips showed low oxygen consumption in sdhaf2, suggesting that a metabolic deficit is responsible for the decreased growth of the root tip (Huang et al., 2013), 2013). Micro-respiratory techniques could allow the measurement of root respiration in a range of mutants to determine whether root tip respiration is a major controller of root growth rate in Arabidopsis. Studies of the response of whole root systems to NaCl treatments have shown stimulatory (Shone and Gale, 1983; Burchett et al., 1984; Cramer et al., 1995) and inhibitory (Hwang and Morris, 1994; Epron et al., 1999) effects and, in some cases, no consistent response in respiration rate (Blacquiere and Lambers, 1981; Malagoli et al., 2008). Here, we found a consistent inhibition of OCR by increasing NaCl concentration and increasing time of exposure. Mixed respiratory responses to NaCl treatments in the variety of plant species studied may indicate that OCR in distinct regions of roots responds differently to salt (Jacoby et al., 2011). Dissection of the

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues respiratory response of root tissues is evidently required to better understand the impact of saline conditions on the root system. The future use of micro-respiratory measurements to calculate root respiration and its response to combinations of different substrates or chemicals will aid our understanding of the physiological importance of respiration in defining root growth.

Kinetics of seed respiration during germination

The Arabidopsis seed OCR shown here has two phases during the germination process. One phase is seen from the onset of imbibition until 10–20 h post- imbibition, and most probably represents the physical hydration process. This first phase is followed by a short lag and then another phase of increasing respiration rate starting 20–30 h post-imbibition. This two-step phenomenon and its timing are consistent with the phases of metabolic initiation and mitochondrial biogenesis reported from Arabidopsis seed transcript profiling over the first 48 h post- imbibition (Narsai et al., 2011). We found that OCR of Arabidopsis seed was inhibited by >70% by the respiratory poison KCN. This suggests that most of the respiration flux occurs via the cytochrome pathway in Arabidopsis mitochondria. The low level of participation of the alternative pathway of respiration may be supported by the lack of effect of the alternative pathway inhibitor SHAM (Lambers, 1985). The predominance of the cytochrome pathway during germination has also been reported in pea seeds and maize embryos, suggesting that this could be a conserved feature of respiration in a range of plant seeds (Alscherherman et al., 1981; Ehrenshaft and Brambl, 1990; Logan et al., 2001).

The phytohormones ABA and GA3 elicit a series of signal transduction pathways and normally show an antagonistic interaction. ABA controls dormancy maintenance, with ABA synthesis increasing to arrest germination until conditions are favorable for germination (Lopez-Molina et al., 2001; Reyes and Chua, 2007). By contrast, the synthesis of gibberellins is linked to germination initiation (Weitbrecht et al., 2011).

In our experiments, the treatment of Arabidopsis seeds with GA3 increased the

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues respiration rate significantly in the latter stages of the germination process. ABA treatments did not show an increase in OCR during the early stages after imbibition associated with the physical imbibition phase. However, ABA treatment showed a dramatic reduction in the OCR associated with the rest of the germination process. The suppression of OCR might be one of the mechanisms to regulate the germination process during hormonally regulated checkpoints. The capacity of this 96-well microtiter plate system to measure OCR of single Arabidopsis seeds over days, and their response to phytohormones, would allow the survey of seed OCR in libraries of Arabidopsis seeds during the germination process. As seeds germinate and survive the assay, this is a physiological, but nondestructive assay system. This would make the micro-respiratory technique a powerful tool to develop phenotype screens of mutant and ecotype populations to help define regulators of the kinetics of respiration initiation during germination.

Conclusions

The adaptation of commercial, 96-well microtiter plate systems that measure OCR of plant tissues provides new opportunities for respiratory research. The small volume limit in the measurements in these instruments actually facilitates the analysis of key Arabidopsis tissues, and other small tissue samples from any plant species, that have often been particularly challenging in the past. By showing the dynamics of measurements made on leaves, root tips and seeds, we hope to stimulate research using these new tools. The potential for multiplexed micro- respiratory assays of up to 96 samples simultaneously means that the assay of mutant populations, phenotypic screens and wider ecotype comparisons in Arabidopsis may be possible in the future. This could provide new ways of combining molecular and physiological studies of respiration in plants.

Chapter 2. Multiplex micro-respiratory measurements of Arabidopsis tissues

Chapter 3:

Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

Foreword to Study II

A common way to study or characterise gene(s) in Arabidopsis thaliana is to disrupt the function of the targeted gene(s) via a reverse genetic approach. The study of mitochondrial malate dehydrogenase (mMDH) in Arabidopsis thaliana began by analysis of mmdh1-2 and mmdh2-1 T-DNA knockout lines. When a null allele of both mMDH isoforms was formed, a double mutant (mmdh1-2mmdh2-1) created from the above single mutant lines, the plant appeared to be slow growing with a small sized rosette compared to wild type plants. This mmdh1-2mmdh2-1 exhibited significantly high leaf respiration rate compared to wild type. One of my PhD research aims is to build upon those previous findings and further explore the changes of respiratory metabolism in mmdh1-2, mmdh2-1 and mmdh1-2mmdh2-1 using physiological and molecular studies. In this chapter, I will discuss the results obtained using respiration analysis, transcript analysis of MDH isoforms and multiple reaction monitoring assays of other TCA cycle proteins using samples of fully expanded leaves from wild type, single and double knockout lines of mMDH as well as a complemented line. Given that leaf respiration rates correlate with developmental stages in wild type plants of Arabidopsis (as discussed in Chapter 2), the changes in respiratory rates during leaf development in mMDH mutants were also examined. Together with the observations of TCA cycle protein abundance changes between old and young leaf age from the above genotypes, a more defined role of mMDH in respiratory metabolism of the overall leaf development is proposed.

Author contributions: The ideas and experiment designs were finalised by myself after extensive discussions with my supervisors Millar, A.H. and Stroeher, E. I conducted all the experimental works including the growth of wild type and mMDH mutant plants, leaf respiration measurements using Hansatech Oxytherm system and Seahorse XF96 instrument, qPCR analysis on the MDH isoforms transcript levels, preparation of protein samples from mMDH mutant plants using high speed

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

centrifugation method. With assistance from Fenske, R. and Grassl, J., protein abundances of TCA cycle enzymes in wild type and mMDH mutants were quantitated using Multiple Reaction Monitoring (MRM) assays conducted in a triple quadrupole LC-MS. I performed all the analysis and data integration for biological interpretation. With guidance from Grassl, J. the workflow of MRM assay analysis for my data was finalised. This manuscript was written by myself and it was revised by Stroeher, E. and Millar, A.H.

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

Impact of mMDH loss on metabolic enzyme networks in leaves

Yun Shin Sew 1,2, Elke Ströher 1,2, Ricarda Fenske 1,2, Julia Grassl 1,2, A. Harvey Millar1,2*

*Corresponding author: A. Harvey Millar

Tel: +61 8 6488 7245 Fax: +61 8 6488 4401 e-mail: [email protected]

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

Abstract

Mitochondrial malate dehydrogenase (mMDH) is a NAD-dependent dehydrogenase which catalyses inter-conversion of malate to oxaloacetate (OAA) via the reduction of NAD+ to NADH in the final step of reaction in tricarboxylic acid (TCA) cycle. Single and double mutant lines of mMDH and a complemented line were used to explore the role of MDH in fully expanded leaf respiratory metabolism of Arabidopsis thaliana. Mitochondrial MDH isoforms, MMDH1 and MMDH2 showed functional redundancy as single mutants of mMDH (mmdh1-2 and mmdh2-1) did not differ in their leaf respiration rates. However, the mMDH double mutant (mmdh1-2mmdh2- 1) respired at a significantly high rate than wild type. Quantitative PCR assays showed that transcripts encoding cytosolic, chloroplastic and peroxisomal MDH were significant up-regulated in mmdh1-2mmdh2-1, implying these isoforms exhibit some compensatory mechanisms of metabolism, probably via malate and OAA transport between subcellular compartments. The abundance of an array of TCA cycle enzymes increased significantly upon the loss of mMDH proteins, consistent with a redox homeostasis occurring to balance the perturbed redox poise in mitochondrial matrix due to the loss of this dehydrogenase. In many of the observations made, with the exception of respiration rate, the effect of losing the MMDH1 isoform was as great as losing both MMDH1 and MMDH2, implying a pronounced role of MMDH1 in leaf metabolism. Sole complementation with the MMDH1 gene in mmdh1mmdh2 35S: MMDH1 could revert most of the aberrant physiological and underlying molecular changes in mmdh1-2mmdh2-1. An inverse relationship was observed between respiration rate and leaf developmental stage, with significant decreases in respiration rate measured as leaves aged in different mMDH genetic backgrounds. Significantly high respiration rates were consistently observed in mmdh1-2mmdh2-1 at all stages of leaf development. These data elucidated the crucial role of mMDH in governing leaf respiratory metabolism and also show its impact on primary metabolism associated with photorespiration, carbon and nitrogen assimilation during the growth and development in plants.

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

Introduction

Mitochondrial respiration is an indispensable metabolic reaction for all living cells in nature as it is a primary source of energy in the form of adenosine triphosphate (ATP) needed to maintain other biochemical processes. The respiratory pathway is made of 3 components: glycolysis, the tricarboxylic acid (TCA) cycle and the mitochondrial electron transport chain. In all eukaryotic cells, the TCA cycle takes place in the mitochondrial matrix and contains an interconnected series of enzymatic reactions. The TCA cycle is responsible for oxidation of respiratory substrates to generate the reducing equivalents (NADH and FADH2) fed into the electron transport chain that is coupled to ATP synthesis by oxidative phosphorylation. Mitochondrial malate dehydrogenase (mMDH) is a NAD- dependent dehydrogenase, which catalyses the inter-conversion of malate to oxaloacetate (OAA) and simultaneously reduces NAD+ to NADH. The operation of NAD-MDH activity either towards the malate oxidation or oxaloacetate reduction depends on the malate: OAA ratio and the redox state in the mitochondrial matrix.

There are two malate dehydrogenase isoforms in Arabidopsis thaliana mitochondria which are designated as MMDH1 and MMDH2. These MDH genes share 84% of sequence homology in their gene coding region while their protein sequences show 89% identity and 95% similarity. Aside from the mitochondrial form of MDH, plant tissues contain other MDH isoforms which reside in different subcellular compartments, namely the cytosol, peroxisome and chloroplast. Each of these is transcribed from separate MDH genes. Every single MDH isoform is believed to have unique kinetic properties, subcellular targeting and physiological functions (Gietl, 1992). Gene expression profiles of the eight NAD-MDH isoforms from the fully-expanded leaves of Arabidopsis thaliana retrieved from Affymetrix ATH1 arrays data sets (Schmid et al., 2005), available at AtGenExpress website (http://www.weigelworld.org/resources/microarray/AtGenExpress) are depicted in Figure 1. Three MDH gene isoforms namely PMDH2, CMDH1 and MMDH1 were shown to dominant gene expression values indicating these are the major MDH

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

isoforms in Arabidopsis leaves. The gene expression value of MMDH1 is found to be half of the expression value of PMDH1. Between the mitochondrial MDH isoforms, MMDH1 has a 10-fold higher expression value than MMDH2, and is considered as the dominant form of MDH in leaf mitochondria. This is consistent with findings at the protein level between the two mitochondrial MDH protein isoforms, MMDH1 dominants the MDH protein in mitochondrial extracts of Arabidopsis tissues (Lee et al., 2008).

Figure 1. Comparison of gene expression profile of NAD-MDH isoforms from 17-day old rosette leaf number 8 Arabidopsis thaliana plant (Columbia sp.) grown on soil and under continuous light conditions. Gene expression data is obtained from publicly available microarray data deposited at AtGenExpress website. Mitochondrial MDH [MMDH1 (At1g53240), MMDH2 (At3g15020)], cytosolic MDH [CMDH1 (At1g04410), CMDH2 (At5g43330) and CMDH3 (At5g56720)], chloroplastic MDH [CHMDH (At3g47520)] and peroxisomal MDH [PMDH1 (At2g22780), PMDH2 (At5g09660)].

The plant mitochondrial inner membrane contains a malate and OAA translocator. It is suggested that in a malate-OAA shuttle, reducing equivalents are transferred from mitochondrial matrix to cytosol, the efflux and influx of oxaloacetate between matrix and cytosol are governed by the oxaloacetate level in order to maintain a redox gradient on both sides of mitochondrial membrane (Neuburger and Douce,

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

1980; Zoglowek et al., 1988). Mitochondrial malate dehydrogenase is proposed to be an essential component for both malate-OAA and malate-aspartate shuttles, involved in the exchange of substrate and reducing equivalents across the mitochondrial membrane (Scheibe, 2004; Nunes-Nesi et al., 2005; Nunes-Nesi and Fernie, 2007). Malate-OAA translocators are also present in other subcellular compartments and are involved in the exchange of malate with oxaloacetate between compartments, linking NAD-dependent MDH isoforms present in the mitochondria, cytosol, chloroplast and peroxisomes and NADP-dependent MDH in chloroplasts into a MDH cellular network (Gietl, 1992). These MDH protein isoforms play essential roles in balancing the ratio of ATP/ NADH or ATP/ NADPH and thus enabled redox reactions to cooperate across cell compartments in various plant metabolic schemes (Krömer, 1995).

It has been long established that mitochondrial MDH is involved in multiple biochemical pathways in plants. The classical role of mitochondrial MDH, together with other NAD-linked dehydrogenases such as isocitrate dehydrogenase and 2- oxoglutarate dehydrogenase, is to provide NADH to be oxidised by the respiratory electron transport chain (Musrati et al., 1998). Mitochondrial MDH is also involved in the C4 pathway, where OAA is reduced to malate prior to a decarboxylation step by NAD-malic enzyme to form pyruvate and supply CO2 for fixation in bundle sheath cells of chloroplast (Hatch and Osmond, 1976). Finally, mitochondrial MDH facilitates the equilibrium of NADH/NAD+ in mitochondria during photorespiration. It has been proposed that NADH generated by glycine oxidation during photorespiration is oxidised by mitochondrial MDH with the simultaneous reduction of oxaloacetate to malate (Journet et al., 1981; Wiskich et al., 1990).

Previous studies reported that double knockout of mitochondrial MDH in Arabidopsis showed delayed seed germination, failed to survive in outdoor conditions, but survived with small and slow growing phenotypes under optimal growth chamber conditions (Tomaz et al., 201 0). The disruption of mitochondrial MDH was found to significantly elevate respiration rate and adversely impacted photorespiration and plant growth, while complementation with a sole MMDH1

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

cDNA expression was able to revert the aberrant phenotypes and restored the respiration rate to a wild type level. In other plant species, antisense studies in tomato revealed an approximately 60% reduction in mitochondrial MDH activity was able to increase photosynthetic activity, carbon assimilation and biomass on the whole plant basis. The transgenic tomato plant showed an accumulation of carbohydrate and increased ascorbate level in their leaves (Nunes-Nesi et al., 2005). These antisense tomato plants were reported to demonstrate a marked reduction in their root biomass and decreased respiratory activity when total mMDH activity was down to an approximately 40% relative to wild type (Van der Merwe et al., 2009). In a more recent study, the value of the flux control coefficient of mitochondrial MDH for respiration in tomato plant was deemed to be the highest (1.76) among all other TCA cycle enzymes (Araujo et al., 2012). The flux control coefficient is used to quantify the effect of an enzyme on the overall steady-state flux in the analysis of metabolic regulation (Delgado and Liao, 1992). All the above indicate that MDH is important for plant respiratory metabolism.

We had previously explored the relationship between dark respiration rates and leaf development using a micro-respiratory measurement method and showed that they were inversely correlated, indicated by a gradual decrease in respiration rate with ascending leaf age (Sew et al., 2013). The link between malate dehydrogenase and leaf development had been previously explored in several studies. For instance, it was found that the total amount of malate dehydrogenase activity extractable from leaf blade of cotton plant increased significantly from the first leaf (old leaf) to the eighth leaf (matured leaf) (O'Sullivan and Wedding, 1972). This was consistent with the observation of more intense protein bands of MDH in matured leaves extract compared to the older leaves in Xanthium (Chen et al., 1970). Interestingly, the two mitochondrial MDH gene isoforms demonstrate a different pattern of temporal gene expression (Arabidopsis electronic Fluorescent Pictograph (eFP) browser). The gene expression of MMDH1 in true leaves gradually from young to adult stage but declined during leaf senescence while MMDH2 displayed a lower but a constant expression pattern during leaf development (Winter et al., 2007). In

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

view of the significantly high respiration rate and an overall defective plant development in mmdh1-2mmdh2-1 plant, it is of interest to investigate the changes in the respiratory metabolism caused by different mMDH gene constituents during leaf development. Here, a detailed investigation focusing on the respiratory metabolism changes of the Arabidopsis mMDH null mutant together with its complemented line at various leaf developmental stages was performed. We evaluated the genotypic differences regarding their respiration rates, gene expression of MDH isoforms and TCA cycle enzymes abundances. This integrative study, linking both physiological and molecular evidence, enabled a more comprehensive insight into the impacts on overall respiratory metabolism that result from the loss of mMDH gene expression during leaf development of Arabidopsis.

Materials and Methods

Plant materials and growth conditions

Surface sterilised Arabidopsis thaliana seeds of wild type (WT) cv. Columbia, single mMDH mutants (GABI_540F04 for mmdh1-2 and SALK_126994 for mmdh2-1, both T-DNA insertion lines), double mMDH mutant (mmdh1-2mmdh2-1) as well as complemented mMDH mutant (mmdh1mmdh2 35S: MMDH1) were sown onto agar plates (½ strength Gamborg B5 basal salt medium, 2 mM MES, 1% [w/v] % sucrose, 0.8 [w/v] % agar, pH 5.7). Plates were placed in cold for 3 days for seed stratification before they were moved to a growth chamber and grown in a short day photoperiod (8 h light/16 h dark), an irradiance of 150 µmol m-2 s-1 PPFD, a relative humidity of 75% and a temperature cycle of 22°C day/17°C night) for about a week. When WT and mMDH mutant seedlings had established, they were transferred onto soil mixture containing compost, perlite and vermiculite in a ratio of 3:1:1 in trays before they continued to grow under short day growing conditions as described above.

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

Genotyping of Arabidopsis mutants

All MMDH mutants were subjected to PCR genotyping for confirmation of the T- DNA insertions in single and double MMDH mutants, and the presence of a MMDH1 cDNA fragment in mmdh1mmdh2 35S: MMDH1 mutant. Firstly, approximately 50 mg of leaf tissue was obtained from each approximately 2-week old mutant plantlet and placed into a 2 mL safe lock micro-centrifuge tube before 500 µL of genomic DNA extraction buffer and a 5 mm grinding steel bead was added. The mixtures were then homogenised in a Retsch mixer mill (Type MM 300, Qiagen, Hilden, Germany) prior to adding 66 µL of 10% [w/v] SDS and mixing by inversion. Then 166 µL of 5 M potassium acetate was added into the mixture and inverted for several times before a centrifugation took place at 4°C for 15 min. Approximately 600 µL clear supernatant was recovered and 0.7 vol of chilled absolute isopropanol was added before overnight DNA precipitation at -20°C. The next day, the genomic DNA pellet was recovered by centrifugation at 13000 rpm at 4°C for 20 min and the pellet was then washed with 500 µL of chilled 70% [v/v] ethanol. Then, the DNA pellet was air dried before reconstitution in 100 µL milli-Q grade water. PCR genotyping using extracted genomic DNA from each mMDH mutant was carried out with gene-specific primers designed flanking the mMDH genes. These gene specific primers sequences (5’-3’) were as follows: MMDH1-2-F (CGGGATTCAAATTGTGATC AC), MMDH1-2-R (GATTGCTTCAGATTCGTCAGC), MMDH2-1-F (ATGACCACCAACA ACTGGAAC), MMDH2-1-R (AATCCCATAACTTTCCCAACG), MMDH1 cDNA-F (CAG CCTCTTGCTCTCCTCAT), MMDH1 cDNA-R (TCAAGAACCTCCTCCACACC), GABI-Kat (o8409: ATATTGACCATCATACTCATTGC) or SALK (LBa1: TGGTTCACGTAGTGGGCC ATCG) T-DNA left border primers. PCR products were visualised on an agarose gel stained with ethidium bromide.

Dark respiration measurement using Clark-type oxygen electrode

Leaf discs of approximately 7 mm diameter were prepared using a cork borer to yield 30-50 mg fresh weight from each genotype. Detached leaf discs were incubated in leaf respiration buffer (10 mM HEPES, 10 mM MES, and 2 mM CaCl2,

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

pH 7.2) in the dark for 30 min prior to the OCR measurement in 2 mL volume for at least 15 min at 25°C in a darkened electrode chamber. OCR was recorded using the Oxygraph Plus v1.02 software (Hansatech Instruments) and adjusted to fresh

-1 -1 weight to obtain OCR per gram fresh weight (nmolO2 min g FW ) of leaf tissue.

Isolation of total RNA

Leaf tissues with a total fresh weight (80-100 mg) were harvested, snap-frozen in a 2 mL safe lock microcentrifuge tube and subsequently ground to homogeneity, using a 5 mm steel bead (Qiagen, Clifton Hill, Australia) in a Retsch mixer mill (Type MM 300, Retsch, Düsseldorf, Germany) with the frequency set at 18 Hz, for 2 times with 60 seconds each. Total RNA was then isolated using the RNeasy Plant Mini Kit (Qiagen, Hilden, Germany) following the manufacturer’s instruction. The total RNA was eluted in 40 µL RNAase-free water (pre-warmed at 50°C) and stored in -80°C for later use. For purification of extracted total RNA, Turbo™ DNAase (Ambion, Victoria, Australia) was first added to the RNA to remove contaminating genomic DNA according to the manufacturer’s instructions with minor modifications. The modified experimental procedure was to precipitate of treated total RNA using 0.1 vol of 3M Sodium acetate (3M NaOAc, pH 5.2) and 2.5 vol of chilled absolute ethanol for overnight at -20°C instead of using a DNase inactivation reagent. The total RNA was then recovered by centrifugation at 15000 rpm for 30 min at 4°C followed by a washing step with 500 µL of chilled 70% [v/v] ethanol. The total RNA was resuspended in 20 µL of RNAse-free water after it was air-dried for about 10 min at room temperature. The integrity of purified total RNA was checked with a 1% formaldehyde agarose denaturing gel electrophoresis and a Nanodrop ND-1000 spectrophotometer (Thermo Scientifics, Wilmington, USA) was used to measure the concentration of RNA and check for impurities.

Quantitative real-time (qPCR) assay and analysis

First strand cDNA synthesis with priming reaction to oligo-dT (Invitrogen, Carlsbad, CA) was performed using a total of one microgram of DNA-free total RNA according

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

to manufacturer instruction. A 5 µL final reaction mixture volume per well was prepared for qPCR assay which consisted of 0.5 µM primer pair, 1X LightCycler® 480 SYBR Green I Master [Roche] and 0.5 µL of diluted cDNA template [1:20 dilution factor] before the mixture was loaded into 384-well PCR plate. Genes encoding for Arabidopsis housekeeping genes which include Clathrin (At5g46630), PPase (At1g13320) and YLS8 (At5g08290) and MDH gene isoforms consisting of two mitochondrial MDH (MMDH1_At1g53240 and MMDH2_At3g15020), two cytosol MDH (CMDH1 _At1g04410, CMDH2_At5g43330 and CMDH3_At5g56720), a chloroplatic MDH (CHMDH_At3g47520) and two peroxisomal MDH (PMDH1_ At2g22780 and PMDH2_At5g09660) were chosen for gene-specific primer design using QuantPrime online software (http://www.quantprime. de/) (Arvidsson et al., 2008). Selection of gene-specific primers for both housekeeping genes and MDH gene isoforms was based on melting curve analysis and presence of a single PCR amplicon using cDNA of ecotype Col. The sequences of gene-specific primers (5’-3’) were as follows: Clathrin-F (TCGATTGCTTGGTTTGGAAGAT), Clathrin-R (GCACTTAGC GTGGACTCTGTTTG), PPase-F (TAACGTGGCCAAAATGATGC), PPase-R (GTTCTCCACAA CCGCTTGGT), YLS8-F (GGGATGAGACCTGTATGCAGATGGA), YLS8-R (GCTCGTACAT GGTGTTGAAGTCTGG), MMDH1-F (AATGTTCCGGTGATTGGTGGTC), MMDH1-R (TTGGCTTGAGGAGTTGCCTGAG), MMDH2-F (GCCAAGTATTGCCCACAAGCAC), MMDH2-R (TCAGCTGCAATTGGAACAGTGGAG), CMDH1-F (TGCTTGTGACCACATCCGT GAC), CMDH1-R (CCATGGAAACGAACGTACCCTCTG), CMDH2-F (GCTGCACCAAAC TGCAAGGTTC), CMDH2-R (TGTTGTGGTCAAGCCTGGTCAAG), CMDH3-F (ACCGGTG CAGCAGGAAACATAG) and CMDH3-R (TCATGGGTTGATCTGGACCTAGC), CHMDH-F (GCTCACTGTTAGGATTCAGAACGC), CHMDH-R (AACCTGCACCTGCCTTAGCATC), PMDH1-F (TGACTGAGCTTCCCTTCTTCGC), PMDH1-R (TGCCTTTCTAATCCCATCCTCTC), PMDH2-F (TTCGTGGAGATGCCAACCAGAG) and PMDH2-R (ACAGAGTTCTTGGCCTCCA TCTG). Subsequently a qPCR assay was performed in a LightCycler® 480 instrument II (Roche) with the following cycling protocol: denaturation (10 min, 95 °C), 40 amplification cycles (95 °C for 10 sec; 60 °C for 10 sec; 72 °C for 10 sec), melting curve analysis (95 °C for 10 sec; 65°C for 60 sec) and followed by a transition rate of

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

+0.1 °C per cycle and continuous data acquisition. Data obtained was then analysed using the LightCycler® data analysis software (Roche). Mean transcript level for each biological replicate was calculated by averaging the 3 technical replicates. To account for variations between cDNA samples, the mean transcript levels of each MDH isoform were normalised with their respective mean ratio of transcript level of housekeeping genes mutant to wild type. This was done by first calculating a ratio of mean transcript level from individual housekeeping gene between mutant and wild type and the mean ratio from all the housekeeping genes was then calculated. Subsequently, the mean normalised transcript level of each MDH isoform was calculated from 3 biological replicates before a final relative value of individual MDH isoform transcript level in mutant vs. wild type was obtained in log2 scale.

Multiple reaction monitoring (MRM) sample preparation and analysis

Protein extractions of Arabidopsis wild type and mutant samples were performed after completing a high-speed centrifugation protocol. First, leaf tissue of five-week old plants was harvested and homogenised in a cold room using a pre-chilled mortar and pestle in the presence of chilled grinding buffer (0.3 M sucrose, 25 mM

Na4P2O7, 10 mM K4P2O7, 2 mM EDTA, 1% [w/v] PVP-40, 1% [w/v] BSA, 20 mM sodium-ascorbic acid, 20 mM L-cysteine, pH 7.5) in a ratio of 1 g leaf tissue fresh weight (FW) to 5 mL of grinding buffer. The homogenate was filtered through 4 layers of pre-wet gauzes and then 1 layer of pre-wet fine Miracloth (Calbiochem) before the filtrate was centrifuged at 2500 g for 5 min at 4°C. The supernatant was then collected and centrifuged at 17400 g for 20 min at 4°C. The resulting pellet was then resuspended in 1 mL buffer (0.3 M sucrose, 10 mM TES, pH7.5). Protein concentration was determined using a Bradford assay with spectrophotometric measurement at 595 nm. BSA was used as a protein standard. For MRM sample preparation, 200 µg soluble protein was precipitated with 5X volume of chilled absolute concentrated acetone overnight at -20°C. On the following day the precipitated proteins were pelleted at 20000 g for 20 min at 4°C. This was followed

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

by two washing steps using chilled absolute acetone before the precipitated protein was resuspended in 200 µL of buffer (8 M urea, 50 mM NH4HCO3, and 5 mM DTT) and incubated at 37°C for an hour with gentle shaking at 300 rpm. The sample was then treated with 10 mM iodoacetamide (IAA) for 30 min at room temperature in the dark prior to a 7 times dilution with 50 mM NH4HCO3 to obtain a final concentration of ≤1 M urea in the mixture. Each protein mixture was then digested with 10 µg trypsin (trypsin powder dissolved in 0.01% [v/v] trifluoroacetic acid to a concentration of 1 mg mL-1) overnight at 37°C. The samples were then acidified to 1% [v/v] with formic acid. For each solid phase extraction Silica C18 Macrospin column (The Nest Group, Massachusetts, USA), 750 µL of 70% [v/v] acetonitrile and 0.1% [v/v] formic acid was first used to equilibrate the column and followed by 750 µL of 5% [v/v] acetonitrile and 0.1% [v/v] formic acid for charging the column before sample loading. After the column was loaded with sample, it was centrifuged for 3 min at 150 g at room temperature and followed by two washes with 750 µL of 5% [v/v] acetonitrile and 0.1% [v/v] formic acid. The purified protein was eluted twice with 750 µL of 70% [v/v] acetonitrile and 0.1% [v/v] formic acid. The eluates of each sample were combined and then dried in vacuum centrifuge for 4 h at room temperature before it was resuspended in 5% [v/v] acetonitrile and 0.01% [v/v] formic acid to a final concentration of 1 µg µL-1 of purified protein. For the multiple reaction monitoring (MRM) assay, 1 µL of each sample (final concentration of 1 µg µL-1) was injected into an Agilent 6430 QqQ mass spectrometer with an HPLC Chip Cube source (Agilent Technologies). Each sample was injected 3 times in total for technical replication. Detailed methods of MRM run, optimization of collision energy (CE) and selection of candidate MRM transitions are as described in Taylor et al. (2014). Basically for each unique peptide, three transitions, one quantifier, and two qualifiers were chosen to validate the quantifier. A total of three peptides per protein were optimised. The peptides were designed in such a way that they do not contain any missed cleavage and cysteine residue which could potentially susceptible to carbamidomethylation and oxidation (Liebler and Zimmerman, 2013). Integration of MRM data was performed using MassHunter Workstation

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

software (Quantitative Analysis version B.06.00 for QQQ, Agilent Technologies). The signal response of a targeted peptide was subjected to normalization (targeted signal responses over the total signal responses of all peptides measured within a protein sample). A mean value of each peptide was then calculated from 3 technical replicates prior to averaging each protein from their corresponding peptides. Principal component analysis (PCA) in R conductor program was then used to detect any outlier which could potentially result a spurious MRM dataset. Subsequently the mean abundance value of each protein was obtained by averaging their corresponding peptides abundance values from 3 independent biological replicates. The change of abundance of each protein in a mutant was compared to wild type by calculating the ratio of mean protein abundance in mutant to wild type respectively. Finally, relative protein abundance values in log2 scale were loaded into MultiExperiment Viewer (MeV) (Saeed et al., 2003) for heat map generation before mapping to the TCA cycle and for hierarchical clustering analysis.

Multiplex micro-respiratory measurement

Arabidopsis wild type (WT), mMDH double mutant (mmdh1-2mmdh2-1) and complemented mMDH double mutant plants (mmdh1mmdh2 35S: MMDH1) were grown for a 6-week period under short-day growth conditions. Upon expansion of the cotyledons, the order in which the subsequent true leaves emerged was recorded for each plant. In general, leaves that emerged within the first 10 days were categorised as old leaves, from day 11 to day 31 day were categorised as mature leaves and from day 31 onwards were categorised as young leaves. Additionally, the size and the shape of each leaf were used as guidelines for leaf group identification. For leaf OCR measurement, a single leaf disc was excised from a relatively similar position at the lamina of each leaf (cotyledons were excluded). The leaf respiration measurements across different developmental stages from individual plant rosette were conducted using a XF96 Extracellular Flux Analyzer (Seahorse Bioscience, Billerica, MA) which is more suitable for respiration measurement on single tiny-sized leaf discs (Sew et al., 2013). The 96-well sensor

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

cartridge was hydrated according to the manufacturer’s recommendations with 200 µL/well XF calibrant solution (Seahorse Bioscience, Billerica, MA) overnight at 37°C prior to the assay. On the day of respiration assay, the heating controller of XF analyzer was deactivated for at least an hour to allow the internal measurement temperature in the machine to reach equilibrium at room temperature (approximately 26°C to 28°C). For sample preparation, a single leaf disc of 2.5 mm in diameter was immobilised to the well bottom with Leukosan® adhesive mixture in a XF96 cell culture microplate. The wells were then filled with 200 µL of leaf respiration medium (10 mM HEPES, 10 mM MES, and 2 mM CaCl2, pH 7.2) before loading the plate into the instrument after the calibration steps. The protocol for basal respiration measurement was 10 loops of mixing (3 min), waiting (4 min) and measuring (5 min). The oxygen consumption rates of a single leaf disc were recorded by the Seahorse XF Acquisition and Analysis Software (Version 1.3; Seahorse Bioscience, Billerica, MA). In the second respiratory assay, leaf developmental stages were divided into three distinct age groups namely old, mature and young leaf. Oxygen consumption rate (OCR) of a single leaf disc was measured from 4 representative leaves per individual age group per genotype. A mean OCR value of each leaf disc was obtained from the average of technical replicates while a mean group OCR values were then calculated by combining the OCR of four representative leaves per individual age group (old, mature and young) per genotype. The final mean OCR values were then calculated from measurements of independent plants (4 plants each for WT and mmdh1-2mmdh2-1 whereas 3 plants for mmdh1mmdh2 35S: MMDH1). For comparison across genotypes, the mean OCR values were normalised with the fresh weight of leaf discs of the corresponding genotype.

Statistical analysis

For leaf respiration data analysis, the Student’s t-Test was used to analyse the significant differences between the different mutant lines and wild type at P<0.05 and P<0.01. For MRM protein abundance data analysis, one-way ANOVA was

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

performed, followed by Tukey’s honestly significant differences (HSD) multiple comparison test for significant differences between the mean values by the least significant difference (LSD) test at P<0.05 and P<0.01. The ANOVA analyses were performed using IBM SPPS Statistics 19. A heat map of relative abundance of TCA cycle proteins was generated using MultiExperiment Viewer (MeV, version 4.9) (Saeed et al., 2003). In addition, hierarchical clustering analysis was performed using a Pearson correlation method in the MeV software.

Results

Validation of mMDH mutants via PCR genotyping

There were a total of four mMDH mutant lines used in this study which include two single mMDH mutants (mmdh1-2 and mmdh2-1), a mMDH double mutant (mmdh1- 2mmdh2-1) and a complemented line (mmdh1mmdh2 35S:MMDH1). Figure 1 (A and B) depicts the T-DNA insertion sites of single mMDH mutants, mmdh1-2 (GABI_540F04) and mmdh2-1 (SALK_126994) at the first exonic and intronic region of MMDH1 and MMDH2 gene respectively. The mmdh1-2mmdh2-1 mutant was generated by crossing both homozygous mmdh1-2 and mmdh2-1 single mutant lines. The complemented line, mmdh1mmdh2 35S:MMDH1 has the genetic background as the mMDH double mutant but was complemented with a MMDH1 cDNA fragment, and its expression was driven by a double Cauliflower Mosaic Virus (CaMV) 35S promoter.

Observations of plant growth and phenotypic appearance of 4-week old single mutant lines revealed a close resemblance between those mutants and wild type except for a slightly larger rosette diameter in mmdh2-1 mutant compared to wild type (Figure 2). However, mmdh1-2mmdh2-1 showed a stunted growth phenotype, which was likely to account for marked reduction in leaf number, rosette diameter, and rosette biomass and seed production compared to wild type under both short- day and long-day lighting regimes. Insertion of a MMDH1 gene fragment in mmdh1-

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

2mmdh2-1 could revert the major growth anomalies to wild type phenotypes. These results were consistent with those reported previously in Tomaz et al. (2010). PCR genotyping using gene specific primers was conducted to validate the homozygosity and identity of each mutant line. Due to the location of inserted T- DNA element at the first exonic and intronic region of MMDH1 and MMDH2 gene respectively for mmdh1-2 (GABI_540F04) and mmdh2-1 (SALK_126994) mutants, appropriate primer combinations were used to amplify the DNA fragments flanking mMDH genes and/or T-DNA insertion sites from individual mutants. Subsequently, the confirmation of individual genotype was based on the presence and the size of amplified DNA fragments corresponding to mMDH alleles and the inserted T-DNA elements as depicted in Figure 3.

Splice variant

Figure 1. Characterisation of mMDH T-DNA insertion lines. (A) Genomic structure of single knockout line, mmdh1-2 (GABI_540F04) showing the position of the T-DNA insertion at the first exon of AtMMDH1 gene. (B) Location of T-DNA insertion for single knockout line, mmdh2-1 (SALK_126994) at the intragenic region of AtMMDH2 gene. Dark grey bars represent coding regions, lines represent introns, and white bars indicate untranslated regions of individual gene. Transcription start and end sites are indicated as ATG and TAA respectively. T-DNA insertions are indicated by triangles and arrows within the bars point towards the T-DNA left border.

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

Figure 2. Phenotypic appearances of 4-week old Arabidopsis wild type and mMDH knockout line plants grown under short-day growth conditions. Single knockout lines (mmdh1-2 and mmdh2-1), double knockout line (mmdh1-2mmdh2-1 with both MMDH1 and MMDH2 genes deleted) and complemented line (mmdh1mmdh2 35S: MMDH1) which has the same genetic background as mmdh1-2mmdh2-1 and the MMDH1 cDNA inserted under expression of a double Cauliflower Mosaic Virus (CaMV) 35S promoter.

Figure 3. PCR genotyping results of homozygous MMDH mutants. Genomic DNA was extracted from individual plants of wild type (WT), mmdh1-2 (GABI_540F04), mmdh2-1 (SALK_126994), mmdh1-2mmdh2-1 and mmdh1mmdh2 35S: MMDH1. Depicted are PCR products with estimated amplicon sizes resulting from PCR amplification of intact gene segments using gene-specific primer pairs (MMDH1-2-F, MMDH1-2-R, MMDH2-1-F, MMDH2-1-R, MMDH1 cDNA–F and MMDH1 cDNA–R) and amplification of T-DNA inserts using T-DNA left border primers (GABI-Kat o8409 and SALK LBa1) together with a gene specific right border primer.

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

Leaf dark respiration analysis of Arabidopsis mMDH mutants

Tomaz et al. (2010) reported that perturbation of one of the mMDH genes in Arabidopsis plants (mmdh1-2 and mmdh2-1) resulted in a comparable leaf dark respiration rates to wild type. However, when both mMDH genes were absent there was almost a 3-fold increase in leaf oxygen uptake rate in the double mutant compared to the wild type. The significantly high leaf respiration rate in mmdh1-2 mmdh2-1 mutant was restored to wild type level by complementing a single MMDH1 cDNA in the mmdh1-2mmdh2-1 background. In this study, oxygen uptake rates of the darkened leaf discs from mature leaves from mMDH mutants were measured using a liquid phase Clark-type oxygen electrode system. There were a total of 22 individual plants for mmdh1-2, mmdh2-1 and mmdh1mmdh2 35S: MMDH1 were used for measurements except for mmdh1-2mmdh2-1 where 11 individual plants were used. Figure 4 demonstrates the oxygen consumption rates of darkened leaf in wild type and the corresponding mMDH mutants. It was noticed that the leaf respiration rate demonstrated by both single mutants of mmdh1-2 and mmdh2-1 did not differ from wild type significantly, notwithstanding a slightly lower rate seen in the mmdh2-1 mutant. By contrast, mMDH double mutant (mmdh1- 2mmdh2-1) respired at a significantly high rate (almost 1.5-fold change) compared to wild type (P<0.01). A single MMDH1 cDNA fragment in mmdh1mmdh2 35S: MMDH1 successfully restored the mmdh1-2mmdh2-1 respiration rate to wild type levels. These findings are in agreement with Tomaz et al. (2010) and confirm a significant elevation of leaf respiration rate in Arabidopsis mutant plants in the absence of both mMDH gene isoforms.

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

Figure 4. Leaf dark respiration analysis of 5-week old leaves of wild type (WT) and mMDH mutants grown under short-day conditions. OCR measurement was performed in a liquid phase Clark-type oxygen electrode system. Oxygen consumption per gram fresh weight (gFW) of leaf discs was obtained (mean ±S.E., n=22 for WT and mutants except n=11 for mmdh1-2mmdh2-1). Significantly different mean OCR from WT is marked with asterisks (**) as shown in Student’s t- Test analysis (P<0.01).

Quantitative PCR assay of mMDH mutants

It is of interest to uncover the underlying molecular changes in mMDH double mutant plants to understand their aberrant physiological changes such as the marked reduction in biomass and significantly higher leaf respiration rate. One such change could be an underlying compensatory response in cellular MDH isoforms in other cellular compartments when both mitochondrial MDH isoforms are absent from the mitochondrial matrix. To test this, a quantitative PCR assay was conducted to examine the changes in MDH isoforms gene expression profiles of all MDH isoforms in the different mMDH mutant backgrounds. With a reduction of MMDH1 transcript level of at least 100-fold in both mmdh1-2 and mmdh1-2mmdh2-1 mutant lines compared to wild type, the MMDH1 gene expression was confirmed to be impeded in these mutants (Figure 5). In contrast, reconstituting MMDH1 cDNA fragment in the complemented line transformant demonstrated a significantly high

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

MMDH1 transcript level (approximately 2.3-fold) compared to wild type. The gene 35S expression of MMDH1 in the complemented line was driven by a double CaMV promoter, which is likely to enhance the overall MMDH1 gene expression. The MMDH2 transcript level was found to be reduced to a lesser degree than the MMDH1 transcript level, with an approximately 6-fold reduction in both mmdh2- 1and mmdh1-2mmdh2-1 mutants. However there was a mild but significant increase in MMDH2 transcript level (1.4-fold) in mmdh1-2 indicating a possible compensation of MMDH1 transcript loss by MMDH2 expression. A significant 2-fold reduction of MMDH2 transcript level was found in the complemented line. The observed less strong reduction compared to the parent line could be due to some feedback mechanism carried out by the overexpressed MMDH1. However whether this mechanism exists, further investigation is needed.

The two cytosolic MDH isoforms CMDH1 and CMDH2, showed similar expression profiles across MDH mutants except that the CMDH2 transcript level was significantly lower in mmdh1-2 mutant than the CMDH1 expression in the same mutant. There was a significantly increased chloroplastic MDH (CHMDH) transcript level in mmdh1-2 and mmdh2-1 mutants and to a greater extent in mmdh1- 2mmdh2-1 mutant. However, the CHMDH transcript level in mmdh1mmdh2 35S: MMDH1 was restored. Whereas the transcript levels of PMDH1 and PMDH2 were in opposite trend across mMDH mutants. A subtle up-regulation of PMDH1 transcripts were observed in contrast to a significant down-regulation of PMDH2 expression across mutants. Interestingly, mmdh1mmdh2 35S: MMDH1 failed to restore the transcript level of PMDH2 to wild type level instead of a further decrease in the PMDH2 transcript level after complementation.

Analysis of the TCA cycle proteome of mMDH mutants The consistent findings of elevated respiration rate in mMDH double mutant possibly indicate other changes were occurring in the TCA cycle of this mutant. A targeted proteomics approach was thus designed to test this by using the multiple

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

Figure 5. MDH gene expression profiles in mature leaves of mMDH mutants. The data represent relative mean (±SE) transcript level in log2 scale from three independent biological replicates of mmdh1-2, mmdh2-1, mmdh1-2mmdh2-1 and mmdh1mmdh2 35S: MMDH1 mutants for transcript level of mitochondrial MDH1 (MMDH1) (A), mitochondrial MDH2 (MMDH2) (B), cytosolic MDH1 (CMDH1) (C), cytosolic MDH2 (CMDH2) (D), cytosolic MDH3 (CMDH3) (E), chloroplastic MDH

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

(CHMDH) (F), peroxisomal MDH1 (PMDH1) (G) and peroxisomal MDH2 (PMDH2) (H) to wild type. Mutants that showed significant difference from wild type by Student’s t-Test analysis are marked with asterisk(s) indicating for P<0.05 (*) and P<0.01 (**) respectively. reaction monitoring (MRM) assay in a triple quadrupole (QqQ) mass analyser for estimating the relative abundance of TCA cycle enzymes in mMDH mutants and wild type. MRM experiments allow high specificity and sensitivity for absolute or relative quantitation of targeted compounds such as peptides and metabolites (Kitteringham et al., 2009). MRM assays are peptide sequence specific and have the capacity to detect a dozen to hundreds of peptides simultaneously (Liebler and Zimmerman, 2013). In this current study, a previously developed and established MRM protein assay method reported by Taylor et al. (2014) was adapted for relative quantification of individual Arabidopsis TCA cycle enzymes protein abundances in mMDH mutants leaf samples. Those enzymes were pyruvate decarboxylase complex (PDC), aconitase (ACO), citrate synthase (CS), isocitrate dehydrogenase (IDH), 2-oxoglutarate dehydrogenase complex (OGDC), 2- oxoglutarate dehydrogenase (OGDH), dihydrolipoyl dehydrogenase (DLD), succinyl- CoA synthetase (SUC), fumarase (FUM) and malate dehydrogenase (MDH). Three unique peptides were analysed for each TCA cycle enzyme. Each unique peptide consisted of a precursor ion with 3 transitions for each precursor (one quantifier and two qualifiers). There were a total of 101 peptides targeted to 33 isoforms encoding these 9 TCA cycle enzymes that were tracked by MRM assays in mMDH mutants, including several peptides that were common between protein isoforms. Those common peptides were for OGDC E2 components (At4g26910&At5g55070), DLD E3 components (At1g48030&At3g17240) and SDH subunit 1 components (At5g66760& At2g18450). The quantitative data for 35 peptides from 23 different proteins which gave reliable results with high signal-to-noise ratios across all the leaf protein samples are listed in Table 1.

In order to visualise protein abundance changes across mMDH mutants, a heat map representing relative protein abundance within each mMDH mutant compared to

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

wild type in log2 scale was constructed (Figure 6). Measured protein abundance of MMDH1 in mmdh1-2 and mmdh1-2mmdh2-1 mutants was reduced by 34-fold and 55-fold, respectively. These extremely low signals for MMDH1 peptides are consistent with the loss of MMDH1 function in those mutants. The 2.2-fold increase of MMDH1 protein in complemented line was expected and is consistent with the 2.3-fold MMDH1 transcript level increase in this complemented line. Intriguingly, the MMDH1 gene expression which was driven by a double CaMV 35S promoter was enhanced resulted in two-fold higher protein abundance. Interestingly, the protein abundances of isocitrate dehydrogenases (IDHs), succinyl-CoA synthetase (SUC) and succinate dehydrogenases (SDHs) were synchronously and significantly up-regulated in both mmdh1-2 and mmdh1-2mmdh2-1 mutants, by approximately 1.1-fold to 1.7-fold compared to wild type. Surprisingly, impacts on the above- mentioned protein abundances of IDHs, SUC and SDHs did not differ between mmdh1-2 and mmdh1-2mmdh2-1. These findings imply that the impact of MMDH1 protein depletion on other TCA cycle enzymes was relatively greater that the loss of the MMDH2 protein. This is consistent with MMDH1 having the more dominant role in regulating TCA cycle respiratory metabolism. The opposite trend in protein abundances of IDHs, SUC and SDHs was seen in the complemented line, with a 1.8 to 3.2-fold decrease compared to wild type. This is most likely to be linked to the over-expression of the MMDH1 protein in the complemented line.

A hierarchical clustering analysis of TCA cycle protein abundance across mMDH mutants was performed using protein abundance relative to wild type in log2 scale and the Pearson correlation method (Figure 7). Clustering analysis showed that protein abundances generally fitted into two main nodes across mMDH mutants. The first node was largely comprised of isocitrate dehydrogenases, succinyl-CoA synthetase and succinate dehydrogenases which had high consistency of expression profile among the mutants. Whereas the second node of TCA cycle enzymes demonstrated a more diverse protein abundance profiles across mutants. Those proteins were mainly involved in the other half of the TCA cycle from fumarase, pyruvate dehydrogenases, citrate synthase and aconitases. The protein abundance

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

ss ss the the cted cted 2 ion

protein protein name;

, peptide product, peptide ct ion 3 (qualifier) Ratio Pro 1/Pro 2, ratio 2, Pro1/Pro Ratio ratio

Genome Initiative identifier; Protein,

Arabidopsis

based assay. AGI, -

Table 1. Signature peptides targeted for a total of 23 TCA mMDH cycle mutants proteins and which wild gave desirable type protein in signal our responses acro MRM Sequence, peptide sequence; Precursor m/z, peptide precursor ion mass/charge ratio; Precursor m/z (Quant.), peptide precursor ion mass (quantifier); Pro 1 m/z product(Quant.), peptide ion mass/charge1 (quantifier) ratio;1 ion Pro (Quant.) ion 1 (quantifier) fragmentation series location; Pro 2 m/z (Qual.), peptide product ion 2 (qualifier) mass/charge ratio; (Qual.), Pro peptide product ion 2 (qualifier) mass/charge fragmentation ratio; series Pro 3 location; ion (Qual.), Pro peptide 3 product ion m/z 3 CE, peptide predictedon fromPred. (MacLean column; CE CE; al., Skyline et Area2010); CE, Opt. optimised (qualifier) (Qual.), fragmentation peptide series location; produ RT, retention time of of the area of extracted ion chromatograph of product ion 1/product ion 2; Area Ratio Pro 1/Pro 3, ratio of the area of extra 3. ion product ion 1/product ion of chromatograph

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

0.7

0.1

0.7

0.5

0.3

0.2

0.6

0.5

0.4

0.3

0.2

0.3

0.8

0.1

0.3

0.1

0.4

0.8

0.1 Pro

Pro

ratio ratio

Area

1/ 1/

0.7

0.8

0.2

0.8

0.9

0.4

0.3

0.8

0.5

0.8

0.6

0.3

0.8

0.7

0.8

0.5

0.3

0.6

0.4

0.6

0.1

Pro

ratio ratio

Area

CE CE

(V)

9.4

6.8

7.3

7.0

26.2

18.7

16.3

18.8

21.5

11.5

12.1

23.5

19.9

12.9

10.2

16.0

13.2

20.1

18.9

18.0

16.6

25.9

10.7

Opt. Opt.

CE CE

(V)

5.4

6.8

8.9

6.2

7.3

7.0

9.2

30.2

22.7

16.3

18.8

25.5

11.5

12.1

27.5

19.9

20.0

20.1

16.0

22.9

14.0

12.6

21.9

14.7 Pred. Pred.

RT RT

9.0

9.3

9.4

9.2

8.8

9.4

9.0

8.4

9.7

15.4

14.1

10.0

11.5

11.9

10.0

11.4

10.8

11.7

10.7

16.1

10.4

12.8 12.1

(min)

10 19

ion ion

Pro 3 Pro3

(Qual.)

m/z m/z

1063

1090

1012

1173

Pro3

785.5

863.4

727.5

660.3

404.2

953.5

403.2

761.4

490.2

674.4

515.3

682.5

697.4

446.3

447.2

474.3

480.7

941.5

759.5 1099.0

(Qual.)

18 14

ion ion

Pro2

(Qual.)

m/z m/z

1198

1162

1064

Pro2

732.4

614.4

921.4

831.4

676.4

589.3

690.3

676.3

991.4

644.3

583.4

598.3

656.4

488.7

578.3

452.7

551.3

960.5

844.4

684.9 985.6

(Qual.)

16 ion ion

Pro1

(Quant.)

m/z m/z

Pro1

970.6

486.3

418.2

902.5

506.3

660.3

874.5

805.4

743.4

484.3

435.3

328.7

976.4

725.3

904.4

650.4

985.5

529.3

471.2

902.5

493.3

1083.5

1014.5 1163.5

(Quant.)

m/z m/z

charge charge (Quant.)

Precursor Precursor

m/z

897.0

749.4

409.6

625.3

673.3

804.4

437.7

529.8

542.2

844.4

695.3

479.2

426.8

449.2

442.8

805.7

485.2

807.4

619.3

882.4

578.8

552.8

735.4 593.3

Precursor Precursor

Peptide sequence Peptide

VAILGAAGGIGQPLALLMK

SEVVGYMGDDNLAK

EGLEALKPELK

SLQNFEIGGER

AIMQAAQEVAEGK

AVIELENYGLPFSR

TIAWLDR

ALLAEDASLR

FMEWWER

AIQDVFPNESELVVK

LITADDLDDAAEK

DVVEDPQR

GLVVPVIR

EAVYFLR

LDPLGLEK

YLQMSDDNPYVIPDMEPTMR

FPFMANSR

LGSEVTVVEFAGDIVPSMDGEIR

LIDDMVAYAVK

VFTTAGVPIEWEEHYVGTEIDPR

VPPEVMESIR

YPEIYYEK

TPVGGGVSSLNVNLR

SLPEGLLESIK

AGI

At1g53240

At2g47510

At5g66760&At2g18450

At5g66760

At3g47833

At1g08480

At2g20420

At4g26910&At5g55070

At5g65750

At3g55410

At1g48030&At3g17240

At1g48030

At5g14590

At5g03290

At4g35260

At3g09810

At2g17130

MDH

FUM

SDH

SUC

OGDC

OGDH

DLD

IDH Protein

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

profile of 2-oxoglutarate dehydrogenase subunits appeared to be exceptional from other TCA cycle enzymes with their constant decrease in abundances across mMDH mutants. Notably, similar TCA cycle protein abundance profiles between mmdh1- 2mmdh2-1 and mmdh1-2 linked them closely, was most likely due to the loss of MMDH1 protein in both these mutants.

Figure 6. Heat map of relative TCA cycle protein abundance in mMDH mutants relative to wild type. Relative mean values of protein abundance of 3 individual biological replicates in log2 scale were shown in blue to red colour scheme which denotes for low to high protein abundances. Protein abundance of mutants was deemed to be significantly different from wild type using one-way ANOVA with post-hoc analysis at P<0.05 (*) and P<0.01 (**) respectively.

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

Figure 7. Hierarchical cluster of TCA cycle protein abundance profiles in mMDH mutants compared to wild type. Heat map in blue to red colour scheme represent low to high protein abundance ratio in log2 scale. Clustering of proteins was performed using Pearson correlation method in MeV software.

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

Dark respiration analysis of mMDH mutant leaves of different developmental stages

To test if aberrant phenotypes and biological defects such as slow growth were related to high respiration rates, we next examined respiratory rates and respiratory metabolism of the TCA cycle during leaf development in mmdh1mmdh2 and its complemented transformant. The OCR was assessed from a single leaf disc detached from individual leaf across the rosette of 6-week old wild type, mmdh1- 2mmdh2-1 and mmdh1mmdh2 35S: MMDH1 plants grown under short-day conditions using multiplex micro-respiratory method (Sew et al., 2013; Sew et al., 2015). In comparison to wild type, the rosette of mmdh1-2mmdh2-1 mutant showed a significant reduction in rosette diameter and contained only half of the total true leaf number (13-15 leaves excluding cotyledons). Conversely, mmdh1mmdh2 35S: MMDH1 plants were comparable to WT in terms of the rosette size and number of leaves. The OCR data showed an oscillating pattern with gradually increasing respiration rates from the oldest leaf to the youngest leaf in individual rosette notwithstanding the plant genotype (Figure 9). The relationship between OCR and developmental stages appeared to follow a polynomial model (non-linear relationship), and therefore the prediction of the correlation was best suited with polynomial regression. In addition, the relationship was also predicted with a linear regression for comparison purposes. Analysis showed that there were good correlations of mean OCR data across leaf developmental stages indicated by polynomial correlation values, R= 0.742, 0.766, 0.850 as well as linear correlation values, R2= 0.667, 0.674 and 0.797 for WT, mmdh1-2mmdh2-1 and mmdh1mmdh2 35S: MMDH1 respectively. This implies that leaf respiration rates are well correlated with their developmental stage. These findings were consistent with our previous report that younger leaves in general respire at higher rates than older leaves (Sew et al., 2013).

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

WT mmdh1-2mmdh2-1 mmdh1mmdh2 35S: MMDH1

Figure 8. Vegetative phenotypes of 6-week old representative wild type (WT), mmdh1-2mmdh2-1 and mmdh1mmdh2 35S: MMDH1 plants grown under short-day growing conditions. Notably, WT and mmdh1mmdh2 35S: MMDH1 showed comparable numbers of true leaves with approximately 23-25 leaves in contrast to mmdh1-2mmdh2-1 which demonstrated developmentally delayed growth compared to WT with approximately 13-15 true leaves.

W T

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

Figure 9. Changes in phenotypic appearances and respiration rates of wild type (WT), mMDH double mutant (mmdh1-2mmdh2-1) and complemented line (mmdh1mmdh2 35S: MMDH1) in individual leaves in the whole rosette. Representative line-up images of all leaves within a rosette of a 6-week old WT (A), mmdh1mmdh2 (B) and mmdh1mmdh2 35S: MMDH1 (C) grown under short-day conditions and leaf oxygen consumption rate (OCR) corresponding to their age for individual genotypes (a-c) respectively. The values represent mean OCR (n=8; mean±SD). The numbers which are in ascending orders indicate the leaf developmental stage from the oldest to the youngest true leaves. The blue dotted lines separate three age groups designated as old, mature and young leaves of individual plants (A-C). Linear and polynomial lines of best-fit are depicted as black and red dotted lines (a-c) respectively and their corresponding correlation values, R2 are reported.

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

In subsequent assays, leaf age was classified into three specific groups: old, mature and young and the effect of underlying mMDH genetic backgrounds and leaf developmental stages on respiration rate were investigated. OCR of the single leaf disc was measured from 4 representative leaves of individual age groups in genotype. Mean OCR values were then calculated from four WT and mmdh1- 2mmdh2-1 plants and three mmdh1mmdh2 35S: MMDH1 plants (Figure 10). In general, a decreasing respiration rate was observed when the leaf age increases notwithstanding the plant genotype. For wild type, the respiration rate was found

-1 -1 to be the highest in the young leaf (mean OCR= 319 pmolO2 min gFW ) and it decreased markedly to approximately 42% when the leaf reached the mature stage and reduced slightly during senescence. Leaf respiration rates of mmdh1-2mmdh2- 1 were significantly elevated compared to wild type (P<0.01), with approximately 1.6-fold, 1.8-fold an 1.4-fold differences to wild type for the old, mature and young leaf age group, respectively. This suggests that the absence of mMDH genes resulted in increased leaf respiration rates in mmdh1-2mmdh2-1 regardless of the leaf age. The complemented line failed to restore leaf respiration rate at the young leaf stage, where the rate was as high as the double mutant mmdh1mmdh2 (458

-1 -1 pmolO2 min gFW ). However the respiration rates of mmdh1mmdh2 35S: MMDH1 declined to a level which was slightly higher than WT at the mature leaf stage and rate was subtly lower than wild type when leaf started to senescence. These findings indicated that by reconstituting the major isoform of mMDH in mmdh1- 2mmdh2-1, the significantly high respiration rate to wild type levels were restored once the leaf became mature.

100

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

Figure 10. Leaf oxygen consumption rate at different developmental stages (old mature and young) of WT, mmdh1-2mmdh2-1 and mmdh1mmdh2 35S: MMDH1. The data represent mean OCR (n=16 for WT and mmdh1-2mmdh2-1, n=12 for mmdh1mmdh2 35S: MMDH1, mean±SE). Values that are significantly different from WT in their corresponding age group are marked as black asterisk(s) (*) and (**) at P<0.05 and P<0.01 respectively while a significant difference from mmdh1mmdh2 35S: MMDH1 are marked as orange asterisks (**) at P<0.01.

Assessing the abundance of TCA cycle proteins from old and young leaves of mMDH mutants

In order to complement the results of TCA cycle protein abundance changes in mature leaf of mMDH mutants, another MRM assay on young and old leaves of those mutants was conducted. It was anticipated that these distinct leaf age groups, old vs. young, may provide a contrasting protein abundance profile among genotypes, simultaneously obtaining a more comprehensive developmental insight into the result of the deletion or overexpression of mMDH gene(s). Figure 11 depicts the changes of relative TCA cycle protein abundances between old and young leaf of mmdh1-2mmdh2-1 and mmdh1mmdh2 35S: MMDH1, respectively to wild type. Similar protein abundance profiles were observed for TCA cycle enzyme clusters of isocitrate dehydrogenases (IDHs), succinyl-CoA synthetase (SUC) and

101

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

succinate dehydrogenases (SDHs) which showed consistently higher abundances in mmdh1mmdh2 but lower abundance in complemented line regardless of age, except for SUC (At5g08300) and SDH (At3g27380&At5g40650). While for other TCA cycle enzymes, in many cases the protein abundances appeared to be varying between their subunits, particularly obvious for pyruvate dehydrogenases and 2- oxoglutarate dehydrogenases.

Figure 11. Relative TCA cycle protein abundances in mmdh1-2mmdh2-1 and mmdh1mmdh2 35S: MMDH1 compared to wild type (WT) according to their leaf developmental stages. The heat map is shown in a log2 scale with values representing ratios of mean TCA cycle protein abundances obtained from 3 independent biological replicates of each mutants and wild type, in old and young leaf groups. Relative abundance of proteins from low to high is indicated in blue to red colour scheme. Proteins that demonstrated significant difference to wild type by one-way ANOVA with post-hoc analysis are marked with one and two asterisks corresponding to P<0.05 and P<0.01 respectively.

102

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

Hierarchical clustering analysis of TCA cycle protein abundances using Pearson correlation method revealed three major nodes with the protein changes dependent to genotype or age (Figure 12). Group I and III proteins are both classified as age-dependent TCA cycle proteins as the protein abundances decrease when the leaf ages in Group I and vice versa for Group III. Group I proteins included pyruvate dehydrogenase components such as PDH-E1-alpha component (At1g59900), PDC E2-1 component (At3g52200), PDH complex E3 components or dihydrolipoamide dehydrogenases (DLD) (At1g48030&At3g17240) as well as OGDH E1 component (At3g55410). The components of the pyruvate dehydrogenase complex catalyse the oxidative decarboxylation of pyruvate to yield acetyl-CoA and are the primary entry point of glucose-derived pyruvate into the TCA cycle. Likewise, 2-oxoglutarate dehydrogenase is thought to be the rate-limiting step for respiration and key control point of carbon skeleton entry to the nitrogen assimilation pathway (Araujo et al., 2008). This is supported by the fact that both PDH (At1g59900) and OGDH (At3g55410) are the major protein isoforms in Arabidopsis thaliana adult leaf according to a transcript expression survey from AtGenExpress microarray database implying an important role of these enzymes in the above mentioned metabolic processes. Our findings coincide with an early study, where the activity of pyruvate dehydrogenase complex (PDC) was the highest in the youngest leaves and decline drastically when leaves matured and became photosynthetically competent, and dropped to zero activity when leaves started to senescence (Luethy et al., 2001). Additionally, the same study reported a good agreement between the total PDC activity and mitochondrial PDC subunit protein levels using immunoblotting. On the contrary, TCA cycle proteins clustered in Group III demonstrated an increase in protein abundances in parallel with the leaf developmental process. Among those proteins, abundances of citrate synthase (At2g44350) and aconitase (At4g26970) were detected to be significantly elevated in old leaves of mmdh1-2mmdh2-1 (P<0.01). Both these proteins are the major protein isoforms of these enzymes in Arabidopsis leaves.

103

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

Proteins clustered in Group II are considered to be genotype-dependent as they displayed contrasting abundance profiles between mmdh1mmdh2 and mmdh1mmdh2 35S: MMDH1 transformants. Notably, a majority of the proteins detected in our MRM experiment (approximately 60%) displayed genotypic differences rather than leaf ageing differences. Among those, pyruvate At3g09810 and At5g03290) and succinate dehydrogenases (At3g47833, At5g66760& At2g18450) displayed the statistically significant contrasting dehydrogenase abundance differences between mmdh1-2mmdh2-1 and mmdh1mmdh2 35S: MMDH1 (P<0.05). A common feature for those 3 TCA enzymes is that they are (At1g24180), isocitrate dehydrogenases (At4g35260), known as NAD-linked or FAD- linked dehydrogenases and are involved in oxidoreductase reactions of TCA cycle. Hence, the results presented here suggested that Group II proteins are highly responsive to a perturbed redox status of the mitochondrial matrix caused by removal of NAD-MDH activity. It was interesting to note that, a couple of proteins in Group II were not only responding to genotype defect but also showed significant high abundance in old leaves compared to young leaves in mmdh1mmdh2 such as PDH (At5g50850) (P<0.05), SDH (At1g08480) (P<0.05) and SDH (At5g66760& At2g18480) (P<0.01). Therefore these enzymes seemed to be affected by both ageing and the perturbed mMDH level. This may be due to the fact that these protein isoforms are the major expressed isoforms of these enzymes in leaf tissue of Arabidopsis according to publicly available microarray data at AtGenExpress website tool.

104

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

Group I

Group II

Group III

Figure 12. Hierarchical clustering analysis of TCA cycle protein abundance measured in mMDH mutants (mmdh1-2mmdh2-1 and mmdh1mmdh2 35S: MMDH1) relative to wild type according to different leaf developmental stages (young and old leaf). Pearson correlation method was used to generate protein clusters. The TCA cycle enzymes were further grouped in 3 major nodes, Group I contains age-dependent TCA cycle proteins which showed decreased abundance in old leaves (boxed in green), Group II consists of genotypic-dependent TCA cycle proteins which showed higher abundance irrespective of leaf age in mmdh1-2mmdh2-1 and mmdh1mmdh2 35S: MMDH1 (boxed in purple), Group III comprised of age-dependent TCA cycle proteins which showed increased abundance in old leaf (boxed in orange). The differences in protein abundance was deemed to be significant at P<0.05(*) and P<0.01 (**) respectively using Student’s t-Test analysis.

105

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

Discussion mMDH is essential for leaf mitochondrial respiration and plant development

Under short-day light regimes, mmdh1-2, mmdh2-1 and mmdh1mmdh2 35S: MMDH1 mutant lines exhibited phenotypes comparable to wild type plants in term of rosette diameter and biomass. In contrast, the mmdh1-2mmdh2-1 mutant demonstrated marked reduction in rosette diameter and tissue fresh weight which were most likely attributed to its delayed seed germination and slow growth. Single mutants of mMDH (mmdh1-2 and mmdh2-1) did not appear to differ in their leaf respiration rates compared to wild type, but the double mutant exhibited an elevated rate of approximately 1.5-fold (P<0.01) indicating the null allele of mMDH led to significant altered leaf mitochondrial respiration. Tomaz and co-workers previously reported that the leaf respiration rate of mmdh1mmdh2 double mutant was significantly elevated (Tomaz et al., 2010). Nevertheless the double mutant demonstrated declined net carbon assimilation rate but elevated maximum photosynthetic electron transport chain activity, presumably carbon and energy resources were directed away to sustain the significant high respiratory rate in this mutant (Tomaz et al., 2010). The same study also reported that the aberrant physiological changes including leaf respiration rate were successfully restored with a regain of MMDH1 cDNA in the complemented line. Herein my data are consistent with their findings, and thus the important functional role of mMDH genes in respiratory and carbon metabolism during plant development is validated and reemphasised. The restoration of function by a sole MMDH1 cDNA in double mutant background as exemplified here and in Tomaz et al. (2010) is most likely due to the fact that MMDH1 is the dominant mMDH isoform in Arabidopsis leaf tissue. This is consistent with transcript data of MDH isoforms available at AtGenExpress website and protein data from Arabidopsis mitochondrial extracts as reported previously (Millar et al., 2001; Lee et al., 2008).

106

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

Compensation by other MDH gene isoforms upon the loss of mMDH genes

However, Tomaz et al (2010) did not directly assess whether transcript levels of other MDH genes altered in the different mutant genotypes. My qPCR assay results has confirmed the undetectable levels of MMDH1 and MMDH2 transcript in mmdh1-2 and mmdh2-1 mutants, respectively, compared to wild type and a complete loss of both mMDH isoform transcripts in mmdh1-2mmdh2-1 mutant. While mmdh1mmdh2 35S: MMDH1 showed a 2.3-fold increase of MMDH1 transcript abundance relative to wild type indicating an over-expression of MMDH1 upon gene complementation. This presumably a consequence of enhanced MMDH1 gene expression driven by a double 35S CaMV promoter to a greater extent than wild type level. The depletion of MMDH1 in the mmdh1-2 mutant resulted in an up- regulation of other cellular MDH isoform transcript levels such as MMDH2, CMDH1, CHMDH and PMDH1. This indicated that the loss of MMDH1 isoform has elicited compensatory mechanisms from other MDH isoforms. It had been suggested that compensation by other genes with similar function or by the redundancy of the related genetic network can occur when a functional gene is absent (Gu, 2003). The 1.4-fold increased MMDH2 transcript level in the mmdh1-2 single mutant compared to wild type exemplifies a direct compensatory mechanism by a gene isoform with functional redundancy within the same subcellular compartment. However, there was no change in the transcript level of MMDH1 when MMDH2 gene isoform is absent from a plant. This implies that although gene isoforms from the same subcellular compartment are functionally redundant, they might not always compensate for each other, it rather depends on the dominance of expression of that particular gene isoform within a coordinated network (i.e. need) and transcriptional regulatory processes (i.e. mechanism). It is therefore anticipated that a compensatory mechanism elicited by gene perturbation of a dominant MMDH1 isoform would be greater compared to its gene counterpart MMDH2 in an Arabidopsis leaf.

The transcript levels of CMDH1, CHMDH and PMDH1 were significantly increased in mmdh1-2 single mutant and mmdh1-2mmdh2-1 double mutant, implying a similar

107

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

compensatory mechanism due to the absence of the dominant isoform MMDH1 in these mutants. These significantly increased transcript levels of other MDH isoforms show for the first time that MDH gene isoforms from different subcellular compartments can compensate for each other. The malate-OAA shuttle makes possible a movement of substrates between MDH isoforms across different subcellular compartments, concomitantly transferring redox equivalents across compartments (Zoglowek et al., 1988; Pastore et al., 2003). Therefore when MMDH1 was depleted as in the mmdh1-2 mutant, it is anticipated that an excess of malate in mitochondria could elicit a higher malate-oxaloacetate shuttling activity to prevent the build-up of excess reducing equivalents in mitochondria. This could have signaled an up-regulated CMDH1 transcript level in the cytosol to meet the increased malate oxidation required. Compensation by CMDH1 in mmdh1- 2mmdh2-1 was at a relatively higher transcript level than in mmdh1-2 mutant (1.32- fold for mmdh1-2 mutant and 1.49-fold for mmdh1-2mmdh2-1 respectively) indicating a greater compensatory mechanism by cytosolic MDH exerted in mMDH double mutant. Both cytosolic MDH isoform (CMDH1 and CMDH2) transcript levels were also increased and could counteract the possibly increased flow of malate into the cytosol from mitochondria to respond to a more severely perturbed redox status in DKOB. This is further supported by a greater increase in fold change of CHMDH transcript level observed in mmdh1-2mmdh2-1 mutant (1.69-fold) compared to mmdh1-2 (1.18-fold) and mmdh2-1 (1.31-fold). It seems from results here that a highly correlated compensatory mechanism exist between mitochondrial and chloroplastic MDH isoforms, where an increasing transcript level of CHMDH corresponds to the total transcript level of mMDH isoforms. This is further confirmed by the down-regulated CHMDH transcript level in mmdh1mmdh2 35S: MMDH1, which have an overexpressed MMDH1 transcript level. Chloroplastic MDH is an essential component of the malate valve in darkened chloroplasts and non-green plastids where it is thought to operate by exporting malate into the cytosol concomitantly transferring reducing equivalents to regulate cellular redox homeostasis (Scheibe, 2004). It has been suggested that the malate valve could

108

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

enable multiple MDH isoforms from different subcellular compartments to maintain a balanced redox state during photosynthesis under high-light stress in Arabidopsis wild type plants (Hebbelmann et al., 2012), however the data shown here prove that compensation involves transcriptional regulation across the MDH gene network.

A consistent up-regulated PMDH1 gene transcript levels was also noticed in both mMDH single and double mutant while a restored transcript level was observed in the complemented line, implying an additional compensatory mechanism operated by PMDH1 gene in response to the transcript deficiency of mMDH isoforms. However in contrast to the effects on CMDH and CHMDH, PMDH2 seemed to be down-regulated consistently in all the mMDH mutant lines. Cousin and co-workers pointed out that the PMDH-mediated redox-shuttle oxidises malate to oxaloacetate concomitantly with the reduction of hydroxypyruvate to glycerate via the peroxisomal hydroxypyruvate reductase (HPR1) (Cousins et al., 2008; Cousins et al., 2011). Evidence from previous mMDH knockout studies revealed that the loss of mMDH gene isoforms affected photorespiration by reduction of the glycine decarboxylase activity thereby limiting photorespiratory carbon flux (Tomaz et al., 2010). Therefore it can be anticipated that the activities of serine‐glyoxylate aminotransferase (SGT) and HPR1 as well as PMDH are reduced due to an insufficient supply of serine from a defected photorespiratory cycle. This could be an appropriate explanation for the above-mentioned up-regulated CMDH transcript levels, in the provision of malate-OAA shuttle mediated by the peroxisomal MDHs which allow for the exchange and regeneration of reducing equivalents across compartment to maintain redox homeostasis (Pracharoenwattana et al., 2007; Graham, 2008). Reduction in PMDH2 transcript levels could plausibly be a compensatory strategy necessitating an adjustment of the PMDH2-mediated redox shuttle for a constant cellular NAD+/NADH ratios. There are a couple of NAD+ related redox reactions in the peroxisomal metabolism such as photorespiration, β- oxidation and H2O2 detoxification (Cousins et al., 2008; Graham, 2008) that are believed to work in concert with PMDH2 in maintaining the redox poise.

109

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

Considering that PMDH2 is the most dominant MDH gene isoform as shown in microarray data of AtGenExpress, it is thus hypothesised that PMDH2 would have a pivotal role in the global cellular redox homeostasis, cooperating with other MDH gene isoforms from different subcellular compartments to form a coordinated compensatory mechanism.

Changes in TCA cycle respiratory component abundance following defective mMDH protein expression

Perturbation of both mMDH genes resulted in a significantly high leaf respiration rate and altered MDH isoform transcript levels indicating significant changes in respiratory metabolism. Therefore, a targeted proteomics was applied in order to determine the effect on protein levels in the TCA cycle. MRM assays are a targeted proteomic technology platform which enables selective proteome profiling of an organism, simultaneously estimating the abundance of specific proteins. Herein, a MRM assay methodology previously established and optimised in our lab as described in Taylor et al. (2014) was used to assess the relative protein abundance of the enzymes that constitute the entire TCA cycle. Our MRM data demonstrated that MMDH1 protein abundance was dramatically reduced by 34-fold and 55-fold relative to wild type in mmdh1-2 and mmdh1-2mmdh2-1 respectively, probably indicating a complete loss of MMDH1 protein. In contrast, a 2.2-fold increase of MMDH1 protein abundance in the complemented line (mmdh1mmdh 35S: MMDH1) indicated a greater protein expression of an MMDH1 cDNA driven by a double 35S promoter in the background of the mMDH double mutant line. The transcript data showed a 100-fold down-regulation of MMDH1 transcript level in those mutants is in good agreement with the findings at the protein levels. Nevertheless, it is necessary to note that a majority of MMDH2 unique peptide abundances across samples including wild type was too low to be detected in whole leaf assays and showed a low signal to noise ratio. Therefore MMDH2 protein was excluded from MRM analysis to avoid spurious results. However, MMDH2 transcript levels were 6-fold down-regulated in both mmdh2-1 and mmdh1-2mmdh2-1

110

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

mutants. This implies the transcription of MMDH2 was disrupted resulting in an impaired MMDH2 protein expression in those mutants. The protein abundances of MMDH1 and MMDH2 were found to be down-regulated as much as 167-fold and 2.3-fold respectively from a differential in-gel electrophoresis (DIGE) analysis of leaf mitochondria in mMDH double mutant (Tomaz et al., 2010). In another DIGE analysis, it was revealed that MMDH1 and MMDH2 protein abundances in the single mutant lines, mmdh1-2 and mmdh2-1, were down-regulated by 48.2-fold and 1.7-fold respectively (Tomaz, 2012).

The relative changes in abundance of other TCA cycle proteome components in mMDH mutants and wild type were compared to gain an insight into the altered respiratory metabolism. The TCA cycle proteome of mmdh1-2 and mmdh1- 2mmdh2-1 were very similar but markedly different from wild type with consistently higher abundances of isocitrate dehydrogenases (IDHs), succinyl-CoA synthetase (SUC) and succinate dehydrogenases (SDHs). In contrast, subtle differences of TCA cycle enzyme protein abundances were found between mmdh2- 1 mutant and wild type (Figure 6). This clearly indicates that knocking out MMDH2 gene is not as significant as the absence of its counterpart MMDH1 to the TCA cycle respiratory proteome as a whole.

Reconstituting only MMDH1 cDNA into mmdh1-2mmdh2-1 mutant could successfully restore a wild type like respiration rate and phenotypic appearances as discussed earlier. From the analysis of MRM assays, a trend of decreased IDHs, SUC and SDHs protein abundances (a reduction by 1.8-fold to 3.2-fold) was detected when MMDH1 protein was expressed in the complemented line. Multiple explanations for this altered TCA cycle proteome following genetic perturbation and restoration of mMDH could be made. In darkened Arabidopsis wild type leaves, Lee and co-workers proposed that there is a malate influx into mitochondria matrix and mitochondrial NAD-MDH works in a forward direction by converting malate to oxaloacetate, probably resulting in a higher abundance of pyruvate dehydrogenase complex, citrate synthase, aconitase and isocitrate dehydrogenase, leading to a stronger respiratory flux of the first half of the TCA cycle (Lee et al., 2010). It is

111

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

therefore anticipated that when both mMDH proteins are depleted, malate is likely to accumulate in the mitochondrial matrix following the disruption of malate oxidation reaction. This is supported by the reported two-fold increase in succinate, fumarate and malate levels relatively to wild type in the darkened leaf of mMDH double mutant (Tomaz, 2012). It can be hypothesised that up-regulated CMDH1 and CMDH2 transcript levels in mmdh1-2mmdh2-1 mutant could have increased the activity of malate oxidation in cytosol, possibly representing a feedback mechanism in response to the accelerated export of malate from mitochondria into the cytosol in the absence of mitochondrial NAD-MDH activity. Due to the proposal that the respiratory flux through the TCA cycle in the dark is always in a clockwise direction (Lee et al., 2010) the cytosolic oxaloacetate concentration would be increasing. In return, this could result in an enhanced influx of oxaloacetate into mitochondria leading to an accelerated TCA cycle respiratory metabolism. This is indicated by a number of significant increases in TCA cycle enzyme abundance, particularly for enzymes involved in the second half of the cycle such as IDHs, SUC and SDHs. Hence, this putative compensatory strategy which uses cytosolic malate oxidation to bypass the disrupted mitochondrial malate oxidation reaction had significantly increased leaf dark respiration rate in the mMDH double mutant as discussed in the earlier section.

The importance of mMDH was evident previously from the fact that complete disruption of mitochondrial NAD-MDH activity lead to at least a 40% reduction of the total cellular MDH activity (Tomaz et al., 2010). It can be anticipated that oxidation/reduction reaction using the mitochondrial NAD+ pool would be significantly impacted in the mMDH double mutant, leading to major alterations of NAD+/NADH ratio in the mitochondrial matrix and potentially a shift in global cellular redox poise. Presumably, there are multiple strategies to maintain cellular redox homeostasis in response to a perturbed NAD+/NADH ratio. One of them would be the compensatory mechanisms mediated by other MDH isoforms (particularly CMDH1, CMDH2 and PMDH2) as discussed earlier. Nevertheless, based on our MRM assay results, potential intrinsic redox homeostasis mechanisms were

112

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

observed responding to the increased NAD+ pool in mitochondria matrix when malate is not converted to oxaloacetate. Due to the pivotal roles of NADH and NADPH levels within the mitochondrial respiratory metabolism (Noctor et al., 2007), it is essential that the free NADH concentration is more or less maintained at a constant level under diverse metabolic conditions (Kasimova et al., 2006). It is therefore anticipated that in response to a perturbed NAD+/NADH ratio, readjustment of cellular redox homeostasis could be accomplished via up-regulation of other TCA cycle enzymes. The increased protein abundances of NAD-linked and FAD-linked dehydrogenases such as IDHs and SDHs shown here could indicate an increase in high-energy electron carriers for NADH and FADH2 in vivo. This could be linked to a significantly elevated respiration rate in mmdh1-2mmdh2-1, as a higher ATP production via oxidative phosphorylation would be expected from a collectively increased concentration of NADH and FADH2. This is supported by evidence from several studies that a high cellular NAD+ level can act as an effector for isocitrate dehydrogenase activity (Behal et al., 1996; Falk et al., 1998; Igamberdiev and Gardestrom, 2003).

The application of the MRM approach facilitated comparative proteomic study between mMDH mutants and wild type by providing protein abundance information for the enzymes directly associated with the generation or use of substrates of MDH. Protein abundance provides unique information, distinct from the evolutionary, structural as well as functional characteristics of a protein (Zhong et al., 2012). There is a good correlation found between estimated protein abundance of an enzyme and its activity (Yoshida et al., 1995) probably better than with its mRNA abundance (Gygi et al., 1999). Nevertheless, consideration must be made that post-translational modifications might change the conformation of a protein and affect its function (Feng et al., 2014) and thus protein abundance can vary from activity. Given this, further investigation by measuring the kinetic activity of the TCA cycle enzymes is necessary to validate the protein abundance data of mMDH mutants for a more conclusive finding.

113

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

Different mMDH genetic backgrounds caused distinct changes in respiration rates across leaf developmental stages

Consistently high respiration rates were observed in mature leaves of mMDH double mutants in this current study, confirming the previous study by Tomaz et al. (2010). Using a micro-respiratory assay, we have illustrated an inverse relationship between leaf dark respiration rates and developmental stages in Arabidopsis thaliana wild type plants (ecotype Columbia) where oxygen consumption rates gradually decreased with leaf development (Sew et al., 2013). In this current study, a similar approach was applied to investigate the effects of different mMDH genetic makeup on leaf respiration during leaf development. There were obvious growth phenotype differences between the mMDH double mutant and wild type, where mmdh1-2mmdh2-1 mutant only contained about half of the number of leaves compared to wild type. This clearly show that the leaf developmental process was affected in mMDH double mutant, possibly due to a delay in seed germination (approximately 4-7 days delayed compared to wild type) and the slow growing characteristics of this mutant plant. Our respiration analysis demonstrated that leaf respiration rates exhibited an oscillating pattern with gradual decline during aging, which is consistent with the findings we reported previously in Sew et al. (2013). There was no difference between mMDH single mutants and wild type, good correlations were generally found between the leaf respiration rates and leaf developmental stages as indicated by linear and polynomial correlation values with R2>0.66 and R2>0.74 respectively. These findings implied that respiratory physiology is associated with leaf development, even in the mMDH mutants. Several published studies also point out that dark respiration is positively correlated with plant growth rate (Hay, 1989; Smith et al., 1995). The respiration rate of three distinct leaf age groups (old, mature and young) showed that young leaf respired at the highest rate in all individual genotypes. In wild type, young leaves respired at 1.7- fold and 2-fold faster than mature and old leaves respectively. In the two- component respiration model, biomass is correlated with respiratory cost composed of growth and maintenance respiration rates (Amthor, 1989). The higher

114

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

rate of growth respiration is involved in biogenesis of cell structures related to growth, nutrient uptake and phloem transport in plants while the lower rate of maintenance respiration is more associated with regulation of ion gradients and concentration and the process of protein and membrane turnover (Amthor, 1989; Thornley and Cannell, 2000). Thus, it is anticipated that growth respiration and overall respiratory cost invested in young tissues would be higher to provide both the energy and the carbon skeletons for biomass production and amino acid biosynthesis.

In both conventional Clark-type liquid phase O2 electrode assay and multiplex micro-respiratory assay, mature leaves of mmdh1-2mmdh2-1 exhibited a significantly high respiration rate compared to wild type which is consistent with the finding reported in Tomaz et al. (2010). The respiration rates of mmdh1- 2mmdh2-1 decreased from young to mature and old leaf stages but were consistently higher than the rates exhibited by wild type at corresponding leaf developmental stages (P<0.01). This clearly indicated that the absence of mMDH gene isoforms resulted in significantly elevated respiration rates regardless of the leaf age. The defected respiratory metabolism across leaf development in the mMDH double mutant implies a profound role of mMDH in modulating leaf respiration rate during both growth and maintenance.

Surprisingly, young leaves of mmdh1mmdh2 35S: MMDH1 were found to respire at a comparable rate to that observed in mmdh1-2mmdh2-1 but reduced drastically to a rate in between mmdh1-2mmdh2-1 and wild type levels at the mature leaf stage. The old leaves of mmdh1mmdh2 35S: MMDH1 respired at a rate slightly lower than wild type. These findings collectively imply that complementation with MMDH1 gene was not sufficient to restore respiration rate exhibited in young growing leaves but recovery was progressively improved at the successive leaf developmental stages as respiration shifted to a maintenance roles (partial and complete complementation at mature and old leaf stages, respectively). Although there was over-expression of MMDH1 transcript and protein in the complemented line as mentioned earlier, this transformant is still lacking the MMDH2 isoform.

115

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

Intriguingly, MMDH2 might play an essential role in the early stage of leaf growth and deficiency of MMDH2 attributed to a significantly high respiration rate in the young leaf. Further experimental investigation is needed to gain a better understanding of the role of MMDH2 gene during leaf development as loss of MMDH2 in single mutants did not lead to a young leaf respiratory phenotype that might otherwise have been anticipated.

Differential changes of other TCA cycle respiratory components in mMDH mutants at distinct leaf developmental stages

A comparative analysis of the TCA cycle proteome of young and old leaves of mMDH mutants was performed using a targeted proteomic MRM assay. In general, the changes in TCA cycle enzyme abundances of mMDH mutants were associated with either the genotype or age. Clustering of relative protein abundance of TCA cycle enzymes revealed unique protein expression patterns between three distinct groups. Group I and III were classified as age-dependent TCA cycle enzymes as there were contrasting protein abundances between old and young leaf in both mmdh1- 2mmdh2-1 and mmdh1mmdh2 35S: MMDH1 mutants. TCA cycle enzymes clustered in Group I collectively showed increase in abundance at the early stage of leaf development. The significant increased abundance of pyruvate dehydrogenase (At1g59900) in young leaf tissues may indicate a stronger respiratory flux through the whole TCA cycle in rapidly growing leaves. It has been suggested that because PDH and its complex PDC, is the enzyme complex immediately upstream of the TCA cycle, its activity is the primary control point for carbon entry into the TCA cycle (Denton and Pask, 1975; Tovar-Mendez et al., 2003). Several studies revealed promising roles of pyruvate dehydrogenase (PDH) in leaf development where a maximum PDH activity was observed in the region of actively dividing and expanding plant tissue implying its role in meeting high metabolic demands of those immature tissues (Owen et al., 1986). In addition to PDC, a significant higher abundance of 2-oxoglutarate dehydrogenase (At3g55410) and dihydrolipoamide dehydrogenase components (At1g48030&At3g17240) in younger leaf tissues were

116

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

also evident from our MRM data. Dihydrolipoamide dehydrogenase (DLD) is a component not only of pyruvate dehydrogenase complex, but also of the 2- oxoglutarate dehydrogenase complex and branched-chain dehydrogenase complexes and the L-protein of the glycine cleavage system (Bourguignon et al., 1996; Sheu and Blass, 1999). DLD is a flavoprotein, which catalyses the two-electron transfer of reducing equivalents from E2 component to produce NADH and H+. Evidence of DLD activity in photosynthetic tissue is rarely reported, however it was found to be 53% higher in young roots than in older ones in Arabidopsis (Lutziger and Oliver, 2001). Intriguingly, the higher abundance of DLD in actively growing leaf tissues could be explained by a greater demand of these enzymes to meet their multiple roles for energy production, carbon and nitrogen assimilation, amino acid biosynthesis as well as photorespiration. This could be linked to studies done on tomato where antisense of OGDH demonstrated an earlier onset of leaf senescence reflecting the importance of this enzyme in both photosynthetic and respiratory metabolism where carbon-nitrogen interactions are apparent (Araujo et al., 2012).

In contrast, Group III proteins were also classified as age-dependent TCA cycle enzymes but showed the inverse pattern to Group I proteins. These proteins are notably closely associated with leaf maturation with their abundances increased with leaf age. It was observed that there were significant differences between old and young leaves when mMDH was knockout in the mmdh1-2mmdh2-1 plants particularly for citrate synthase (At2g44350) and aconitase (At4g26870) (P<0.01). In potato studies, mitochondrial citrate synthase was shown to have higher mRNA expression as well as increased activity when the leaf aged, implying its expression is developmentally regulated (Landschutze et al., 1995). Mitochondrial aconitase has been suggested as the main functional target of aging in the TCA cycle of animal tissues and its activity decreases during ageing process (Delaval et al., 2004; Yarian et al., 2006). Intriguingly, a higher abundance of aconitase (At4g26970) in older mmdh1-2mmdh2-1 leaves could be associated with accumulation of non-functional aconitase due to inactivation of its iron sulfur (Fe-S) cluster by oxidants in aging tissues (Yarian et al., 2006; Scandroglio et al., 2014). Interestingly, SDH subunit 2

117

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

(At3g27380&At5g40650) which is also a protein containing Fe-S clusters was also found in Group III, and behaves in a similar way as aconitase.

By combining the comparative analysis of respiratory and proteomic data of mMDH mutants, some inferences could be made. Both Group I and III TCA cycle enzymes are characterised as age-dependent enzymes, in which the protein abundance changes are intimately linked with the leaf developmental stage. These two groups of enzymes could plausibly account for growth respiration rates in plant leaves. When compared to wild type, consistent up-regulated abundances of IDHs, SUC and SDHs (clustered under Group II TCA cycle enzyme) across leaf developmental stages were shown in mmdh1-2mmdh2-1. Group II TCA cycle enzymes are characterised as genotypic-dependent enzymes, with a large number of these enzymes consistently and significantly up-regulated throughout the entire leaf development course, possibly to compensate the altered redox poise resulting from depletion of mMDH genes in mmdh1-2mmdh2-1 mutant. Therefore, it is plausible that Group II TCA cycle enzymes are more important in maintenance respiration during leaf development of Arabidopsis.

Loss of mMDH leads to high leaf respiration rates, low carbon use efficiency and defective photorespiration and nitrogen assimilation

It was evident from this current study that perturbation of mMDH gene isoforms in Arabidopsis increased the protein abundances of an array of TCA cycle enzymes which potentially resulted in an overall increase in TCA cycle respiratory flux. This is a plausible factor that can account for the significantly higher leaf respiration rate observed in this mutant as evaluated in both current and previous studies. Based on the quantitative relationship established between respiratory energy metabolism and cellular processes, the balance of growth respiration and maintenance respiration make up the plant total respiratory cost (Lambers, 1985). In a similar manner, the carbon use efficiency of plants can be described as a combinatory effect of growth respiration, maintenance respiration and relative growth rate (Van Iersel, 2003). It was suggested that a decrease in relative growth rate will result in

118

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

lower carbon use efficiency but increase the importance of maintenance respiration in the carbon balance of the plant. Remarkably high leaf respiration rates demonstrated by mMDH double mutants across leaf development are indicative of a consistent high supply of carbon skeletons and energy to meet the cellular metabolic demand in this mutant. However, the increased respiratory metabolism in the mMDH double mutant, does not enhance the overall growth and development. Instead, the plant appeared to be small in size with marked reduction of rosette diameter and exhibited slow growing characteristic as well as a visible decrease in biomass. The majority of energy generated from increased respiratory metabolism in mMDH mutant could have been consumed for cellular maintenance processes in order to keep cells in a viable state and this presumably happens at all stages of leaf development. This appears linked to a constant up-regulation of the TCA cycle enzyme profile (cluster of genotypic-dependent TCA cycle enzymes in Group II) throughout leaf developmental stages. It can be hypothesised that some kind of readjustment of perturbed redox poise in the mitochondrial matrix by the concerted reaction of those enzymes is increasing maintenance respiration functions or making them less efficient in the mMDH double mutant.

It has previously been highlighted that there was a marked accumulation of glycine (approximately 27-fold) in mmdh1-2mmdh2-1 supported by significant reduction of photorespiration rate, implying a defective photorespiration process in this double mutant (Tomaz et al., 2010). This could be linked to the absence of mitochondrial MDH which lead to a failure to recycle NADH formed in the glycine decarboxylase (GDC) reaction in the mitochondrial matrix during photorespiration. It is worth noting that, glycine oxidation is a rapid reaction and is thought to be the primary metabolic reaction in mitochondria of illuminated leaves (Oliver et al., 1990). High carbon flux from glycine into the mitochondrial matrix has been associated with production of ammonia that needs to be re-assimilated in the chloroplast (Keys et al., 1978; Keys, 2006). Knockout of GDC inhibits photosynthetic CO2 fixation (Artus et al., 1994), increases respiration rate and increased NADP/NAD-dependent malate dehydrogenases activities (Igamberdiev et al., 2001). Thus, disruption of the GDC

119

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

reaction as a consequence of mMDH genes deletion is likely to have a negative impact on photorespiration where it limits the carbon flow and nitrogen assimilation in C3 plant. This is supported by decreased protein abundances of 2- oxoglutarate dehydrogenase (OGDH) E1 subunits (At3g55410 and At5g65750) in mMDH mutants presented in our data. Inhibition studies of the OGDH complex using chemical inhibitors in potato tubers demonstrated that alterations in the TCA cycle intermediates levels and amino acids are important for nitrate assimilation (Araujo et al., 2008). More convincingly, antisense studies on 2-oxoglutarate dehydrogenase E1 subunit in tomato plant revealed that nitrate, amino acid and protein content declined drastically in the transformants indicating a substantial alteration of nitrogen metabolism in those antisense lines (Araujo et al., 2012). In addition, using DIGE analysis it was found that there was a significant 2-fold decrease in protein abundance of glutamate dehydrogenase in the mMDH double mutant mitochondrial proteome (Tomaz, 2012). Collectively, all the above findings imply that while respiration rate is increased, there is significant lowering in photorespiration, carbon and nitrogen metabolism as a consequence of mMDH loss in Arabidopsis plants.

Conclusion

This current study had witnessed and validated the significantly higher respiration rate exhibited in mMDH double mutant and investigated the underlying altered respiratory metabolism by focusing on the TCA cycle enzyme proteome of mMDH mutants. We provided evidence of multiple compensatory strategies including altered expression of other MDH gene isoforms, the malate-OAA shuttle system and induction of TCA cycle enzymes (IDHs, SUC and SDHs) in re-adjusting the perturbed redox poise in mitochondrial matrix to circumvent the lack of mitochondrial malate dehydrogenases. These strategies to circumvent mMDH loss during leaf development were persistent at all stages of leaf development, leading to the high respiratory rate phenotype of the whole rosette. Overall, the data

120

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

presented in this study has advanced our knowledge of the critical function of mitochondrial MDH in modulating leaf mitochondrial respiration and its far- reaching roles in coordinating respiratory metabolism with leaf development, photorespiration, carbon and nitrogen assimilation in Arabidopsis. In view of the above, mitochondrial MDH can be envisioned as a possible target for genetic engineering to enhance plant growth rate and biomass.

121

Chapter 3. Impact of mMDH loss on metabolic enzyme networks in Arabidopsis leaves

122

Chapter 4:

Impact of mMDH loss on seeds and roots of mMDH mutants

123

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

Foreword to Study III

In Chapter 3, the impact of the loss of both mitochondrial malate dehydrogenase (mMDH) gene isoforms on photosynthetic tissue e.g. leaf tissue had been documented, in particular a significantly elevated leaf respiration rate and an altered leaf respiratory metabolism were evident. While propagating with this mMDH double mutant (mmdh1-2mmdh2-1), a significant reduction in seed germination rate (up to 80% reduction compared to wild type) and a strikingly low seed viability percentage were observed. This was consistent with the previous findings where low seed yield was obtained from mmdh1-2 single mutant and the mmdh1-2mmdh2-1 double mutant plant failed to survive when grown outdoors (Tomaz et al., 2010). Considering that mMDH has a pivotal role in regulating leaf respiration, we hypothesized that there would also be a change in seed respiration of mMDH mutants. Thus, assessment of seed respiration was carried out using the multiplex micro-respiratory method, developed in Chapter 2 coupled with seed metabolite profiling of wild type, mmdh1-2mmdh2-1 and a complemented line (mmdh1mmdh2 35S: MMDH1) at three selected maturation stages (green, green ripe and ripe stage). Due to the extremely low seed viability rate, it was speculated that seeds of mmdh1-2mmdh2-1 would have a shorter longevity. To test this, a seed ageing assay was conducted on this mutant together with a wild type and complemented line for comparison. Aside from that, it was important to understand the post-germinative events taking place in the mMDH double mutant that lead to its albeit small and slow growing phenotype (as discussed in Chapter 3). Therefore a root respiration assay was performed on mMDH mutants using two different measurement methods: the pre-established root micro-respiratory method (as discussed in Chapter 2) for root tips and root expanded regions and liquid-phase Clark-type oxygen electrode for whole root respiration assessment of mutants and wild type. The combined outcome of the study presented in this chapter illustrated an unprecedented role of mMDH genes in regulating respiratory metabolism of Arabidopsis seeds and roots, indicating that mMDH is an essential

124

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants element for production of good quality seed and healthy root growth and development in Arabidopsis.

Author contributions: Experiments conducted in this study were finally defined by me following discussions and advice on design from my supervisors, Millar, A.H. and Stroeher, E. I also modified methods and adapted protocols as needed to conduct this research. All the experimental works including the growth of wild type and mMDH mutant plants, seeds harvesting, seed germination assay, seed viability test, seed and primary root respiration measurements using Hansatech Oxytherm system and Seahorse XF96 instrument, seed ageing assays, primary root length and root width analysis were conducted by me. The seed metabolite profiling was outsourced to Metabolomic Australia and I performed the analysis of seed metabolite data with analysis guidance from Fenske R. All the analysis and data integration for biological interpretation was performed by me. This manuscript was written by me and it was revised by Stroeher, E. and Millar, A.H.

125

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

Impact of mMDH loss on seeds and roots of mMDH mutants

Yun Shin Sew 1,2, Elke Ströher 1,2, Ricarda Fenske 1,2, A. Harvey Millar1,2*

*Corresponding author: A. Harvey Millar

Tel: +61 8 6488 7245 Fax: +61 8 6488 4401 e-mail: [email protected]

126

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

Abstract

Mitochondrial malate dehydrogenase (mMDH) (EC 1.1.1.37) catalyses the inter- conversion of oxaloacetate and malate in the final enzymatic step of the tricarboxylic acid cycle. The role of mMDH in Arabidopsis heterotrophic organs was investigated on seed and root tissues of Arabidopsis mMDH single (mmdh1-1, mmdh1-2 and mmdh2-1) and double mutant (mmdh1-2mmdh2-1) together with a complemented line (mmdh1mmdh2 35S: MMDH1). A large proportion of mmdh1- 2mmdh2-1 mutant seeds developed only to torpedo-shaped embryos, indicative of arrested seed embryo growth during embryogenesis. The developing seed of mmdh1-2mmdh2-1 mutant showed a paler green phenotype when compared to wild type and was believed to be photosynthetically incompetent, resulting in a deficit in reserve accumulation and reduced seed biomass. Moreover, respiration rate was constantly and significantly elevated in maturing seeds of mmdh1- 2mmdh2-1, together with a consistently high content of free amino acids (branched-chain amino acids, alanine, serine, glycine, proline and threonine). The impaired seed maturation process is implicated in the successive growth and development of seeds and roots of this mmdh1-2mmdh2-1 plant such as an expedited seed ageing process, strong reduction in seed viability and germination rate as well as a retarded post-germinative growth, significantly higher root respiration rate, shorter root length and decreased root biomass. Complementation by re-constituting a sole MMDH1 cDNA in complemented line failed to fully restore the anomalies observed in mmdh1-2mmdh2-1 seeds, indicating sub- functionalisation of MMDH1 and MMDH2 gene isoforms in regulating respiratory metabolism of heterotrophic tissue in A. thaliana. High respiration rates at the root tips compared to the expanded regions, with a concomitant higher gene expression of MMDH1 inferred from transcript studies, suggests that MMDH1 participates more in active zone root respiration than its gene counterpart MMDH2. Collectively, this present study provides evidence of an important role of mMDH

127

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants genes in respiratory metabolism of Arabidopsis seed and root, orchestrating growth and development in these non-photosynthetic organs of plant.

128

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

Introduction

All living organisms require energy to fuel their metabolic processes for growth, development, reproduction and maintenance. Energy is mainly conserved in the form of ATP and the energy requirement in a particular organism generally varies according to the developmental stage and metabolic state in vivo. In plant such as Arabidopsis sp., mitochondria and chloroplast are the predominant sites for ATP production via mitochondrial oxidative phosphorylation and chloroplastic photophosphorylation, respectively. Heterotrophic and non-photosynthetic plant tissues such as seeds and roots respectively are largely depending on mitochondrial respiration as an energy resource. The tricarboxylic acid cycle is one of the important components of mitochondrial respiration has been shown to play an essential role in providing carbon skeletons, ATP and reductant for various biochemical reactions including carbon and nitrogen metabolism and biosynthesis of cell structures, in order to meet high energy demand of fast-growing heterotrophic organs (Pradet and Raymond, 1983; Dieuaide-Noubhani et al., 1997; Stasolla et al., 2003).

Mitochondrial malate dehydrogenase (mMDH) is known to be one of the key components of the TCA cycle of mitochondrial respiratory machinery. This enzyme catalyzes the reversible oxidation of malate to oxaloacetate coupled to the reduction of the NAD pool. Besides its role as a NAD+-dependent dehydrogenase in the final step of TCA cycle as mentioned above, mMDH is involved in the photorespiratory pathway by supplying NAD+ to glycine decarboxylase for conversion of glycine to serine via MDH’s reverse reaction, reduction of oxaloacetate to malate (Journet et al., 1981). Lastly, mMDH mediates the supply of

+ malate to pyruvate via NAD malic enzyme reaction to provide CO2 for fixation in bundle sheath cells in chloroplasts in the C4 pathway (Hatch and Osmond, 1976). In Arabidopsis thaliana, there are multiple isoforms of malate dehydrogenase (MDH) that possess NAD+-dependent dehydrogenase activity and they are localized in different subcellular compartments such as mitochondria, chloroplasts,

129

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants peroxisomes and the cytosol. Two MDH isoforms exist in Arabidopsis mitochondria namely MMDH1 (At1g53240) and MMDH2 (At3g15020). It has been long established that as the enzymatic reactions and kinetics, substrates (malate and oxaloacetate) and reducing equivalents (NAD+ and NADH) are shared among MDH isoforms, the malate-oxaloacetate shuttle could transfer substrates and products to cooperate in redox reactions and metabolism across membranes (Krömer, 1995).

Arabidopsis plants lacking both mMDH isoforms namely mmdh1-2mmdh2-1 had been characterized as viable, albeit small, slow growing and demonstrate significantly higher leaf respiration rate in the dark and light compared to wild type

(Tomaz et al., 2010). In addition, low net CO2 assimilation and inefficiency of photorespiration were reported as major drawbacks in mmdh1-2mmdh2-1. This double mutant plants failed to survive in outdoor conditions however single mMDH mutant (mmdh1-2) showed a significant 60% reduction in seed production (Tomaz et al., 2010). Recently, knockout of the plastidial NAD-MDH (plNAD–MDH) in Arabidopsis found that only heterozygous plants survived and the mutant plant produced both heterozygous and homozygous mutant seeds with green and white phenotypes, respectively. The homozygous mutant seeds were arrested at the globular stage with embryos that appeared to be tiny and wrinkled (Selinski et al., 2014). Similarly, another gene silencing study of plastidial NAD-MDH (pdnad-mdh) in Arabidopsis showed that the homozygous mutant plant were non-viable whereas mutant seeds produced by heterozygous plant gave rise to embryos arrested in the globular-to-heart transition stage (Beeler et al., 2014). Taken together the above findings imply that both mitochondrial and plastidial malate dehydrogenases have indispensable roles in Arabidopsis plant and seed development.

In the life cycle of Arabidopsis thaliana or other higher plants of spermaphyta, seed is the crucial element for the continuity of plant and dispersion of species. Seed development and maturation are the key processes that determine the quality of genetic material and nutrients required for the next sporophytic generations through seed propagation (Baud et al., 2008). In general, seed development could be divided into two phases termed embryo morphogenesis and maturation. In the

130

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants first phase usually 0-7 days after fertilization (DAF), a series of programmed cell divisions take place upon formation of zygote and endosperm where embryos acquire the basic architecture of the plant (Mayer and Jurgens, 1998). The chronological changes of embryo architecture during seed embryogenesis encompasses octant, globular and ends at heart-shaped or torpedo-shaped embryo (Mayer et al., 1991). The maturation phase (8-20 DAF) is divided into 3 main stages known as early, intermediate and late maturation. Embryos undergo a rapid growth period after entering the early maturation phase (Goldberg et al., 1994). Throughout the maturation stages, the embryo fills the seed concurrently with the resorption of the endosperm (Mansﬁeld and Briarty, 1993). The endosperm mediates the transfer of nutrients from the mother plant to the embryo and it is an important site of genetic imprinting to prevent parthenogenesis in flowering plants (Huh et al., 2007). The maturation phase involves the accumulation of reserve compounds, acquirement of tolerance to desiccation, arrest of growth and germination which breaks the transient dormancy period (Goldberg et al., 1994; Harada, 1997). More specifically, maturation is characterized as a marked increase in seed dry weight and large quantities of stored fatty acids and proteins are found in the embryo, which could account for approximately 40% of dry matter in a matured Arabidopsis seed (Baud et al., 2002). Lipids are accumulated in the form of triacylglycerols which are then stored in oil bodies, and serve as the major carbon and energy reserves to fuel seed germination and growth of the young seedling (Mansfield and Briarty, 1992). Other pronounced metabolic changes observed in Arabidopsis seeds which undergo transition from reserve accumulation to seed desiccation include accumulation of distinct sugars, organic acids, nitrogen-rich amino acids, and shikimate-derived metabolites (Fait et al., 2006). In maturing pea seeds, it was evident that gas exchange level declined, concomitantly with a decrease in succinate dehydrogenase and malate oxidase activity, coupled with a drop in the respiratory control and phosphorylation efficiency when mitochondria gradually loses cristae upon water loss (Kolloffe.C, 1970). In developing sunflower seeds, high rates of oxygen uptake during embryogenesis were associated with a

131

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants rapid cell division process happened in embryos. However, the respiratory control coefficient decreased progressively upon starting of seed reverse accumulation (Zaitseva et al., 2002). Following the period of reserve accumulation, seeds undergo a last stage of maturation phase, known as maturing drying, allowing a major loss of seed moisture before dry seed becomes quiescent (Kermode et al., 1985). Similar to most of the bean seeds, Arabidopsis seeds are known as orthodox seeds which are tolerant to desiccation, meaning that they can withstand a loss of water content up to 95% at the final stage of desiccation (Rangel-Fajardo et al., 2011). Mitochondria of dry maize seed embryo under desiccation stress appeared to be poorly developed with their internal membrane structure contained only a few cristae as observed from electron micrographs (Logan et al., 2001). This is consistent with an undetectable respiration rate of shrubs tree seeds at the end of desiccation (Lin and Chen, 1995). The shutdown of respiration and metabolic inactivity of quiescent dry seed enable seeds to be stored for a long period of time until germination is initiated under favourable conditions.

The primary causes contributing to the differences in seed vigour and germination are known to be seed ageing and physiological deterioration (Powell and Matthews, 1981; Powell et al., 1984; Hampton and Tekrony, 1995). Ageing in seeds involves a progressive deterioration of structures and functions of the seed over time (Mohamed-Yasseen, 1991) and this process is an irreversible degenerative change which is generally considered to signify the death of the seed (Roberts, 1972). Seed deterioration could also be associated with environmental factors during handling, shipment and storage of seeds particularly high moisture (Robertson et al., 1973; Nakayama et al., 1981), high temperature (Alhamdan et al., 2011) and infection by seedborne diseases (microbial and fungal infection) (Tsror et al., 1999) which can exacerbate the deterioration of seeds. In order to shorten the often long periods of natural aging, artificial aging or a controlled deterioration test (CDT) was developed and commonly used to assess storage potential and longevity in seeds (Powell et al., 1984; Tesnier et al., 2002). CDT may create artificial ageing effects to the treated seeds by accelerating the production of proteases and subsequently breaking down

132

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants the stored proteins, impairing metabolism and energy supply and eventually resulting in seed deterioration (Xin et al., 2011). Seeds may deteriorate both on the mother plant and in storage and the rate of deterioration is highly dependent upon the environmental conditions, particularly temperature and relative humidity (RH).

Seed germination is a complex process that is governed by various external factors such as temperature, oxygen, light and the water potential of the growing medium and also the internal cues of the seeds themselves, such as hormones (Bewley, 1997). Germination occurs after a quiescent dry seed is imbibed until the emergence of radicle through the enclosing structures such as micropylar endosperm and testa (Bewley and Black, 1978). The number of cristae in the mitochondria as well as the number of mitochondria increased markedly during seed germination (Breidenb et al., 1966; Ueda and Tsuji, 1971). This is accompanied by a series of seed metabolic events which include changes in subcellular structure, respiration, syntheses of macromolecule and elongation of cells (Bewley, 2001). During the onset of seed germination, reserves stored in the seeds are degraded and mobilized to fuel metabolic processes (Bewley, 1997; Borisjuk et al., 2004; Penfield et al., 2005). A successful establishment of seedling hence relies largely on two factors; how effective the mobilization of storage reserve is and the competency of the seed itself in storing reserve accumulation during the maturation process through the partitioning of carbon and nitrogen in the developing seed (Eastmond and Rawsthorne, 2000; Eastmond and Graham, 2001). The energy to power reserve mobilization during seed germination is supported by the respiratory process which is initiated immediately upon rehydration (Ehrenshaft and Brambl, 1990). It has been suggested that a water influx into the cells of imbibed seed could lead to rapid leakage of solutes and metabolites and be followed by resumption of metabolic activity in quiescent dry seeds (Bewley, 1997). Transcripts related to energy metabolism (glycolysis, TCA cycle, oxidative phosphorylation and mitochondrial electron transport) were found to be up- regulated significantly in germinating seeds of Arabidopsis (Narsai et al., 2011) and Chinese bread wheat (Yu et al., 2014), indicated that respiration is one of the

133

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants essential processes to power seed germination. An elevated respiration rate up to several hundredfold was demonstrated in mitochondria isolated from excised rice seed embryos after 48 h post imbibition under aerobic conditions further showing the prominent role of mitochondrial function during germination (Taylor et al., 2010). An earlier study on rice germination under aerobic conditions found that mitochondria exhibited morphological differentiation with reduced levels of components of the protein import apparatus, high levels of TCA cycle constituents and components of electron transport chain as well as increased capacity in the general mitochondrial import pathway (Howell et al., 2007). It was suggested that step-wise mitochondrial biogenesis generated energetically active organelles and strengthened their energetic functions; as revealed by temporal gene expression pattern of Arabidopsis seeds across different germination stages (Narsai et al., 2011). Taken together all the reports above, mitochondria clearly exhibit respiratory control in seeds and play a crucial role in coordinating successful seed germination events. In view of the above, it was interesting to investigate the role of mMDH, as one of important mitochondrial respiratory genes, during seed development and germination events.

Roots represent an important component for successful seedling establishment in plants. A major source of carbon for energy generation to support root growth processes like ion uptake, maintenance and turnover of existing and biosynthesis is presumably obtained from photosynthetic tissues (Farrar, 1985). Root system architecture is an important developmental and agronomic trait, as an indicator of plant developmental processes such as the overall plant architecture, growth rate and yield, abiotic stress resistance, nutrient uptake, and developmental plasticity in response to environmental changes (Jung and McCouch, 2013). Hence, Arabidopsis roots are often a model system for studying growth, development and response to environmental stresses or gene knockout effects. It was suggested that respiration rate in root is strongly correlated with the root section length but a weak relationship with their growth rate (Audus and Garrard, 1953). Respiration studies on root of soybean seedlings revealed that 4-day old young root respiration was

134

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants almost fully contributed via cytochrome c oxidase but shifted to a larger proportion of alternative oxidase respiratory pathway (>50%) in root of 17-day old seedling, with a decrease in the whole root respiration as seedlings aged (Millar et al., 1998). Survey of Arabidopsis microarray transcript data deposited in the Genevestigator database (Hruz et al., 2008) revealed that in Arabidopsis root tissues, MMDH1 transcript level was generally seven times higher than MMDH2. Transgenic tomato roots with down-regulated mMDH activity were shown to have several alterations in the root tips. Those aberrant phenotypes included a markedly reduced root area and cell width, significantly shorter root length, overall reduction in root dry mass and significant decline in root respiration compared to wild type (van der Merwe et al., 2009).

To date, the impact of losing both mMDH isoforms in Arabidopsis seed and root has not been investigated and reported in the literature. In this report we characterized seeds of Arabidopsis mMDH mutants (mMDH double mutant and complemented line) from different perspectives including their efficiency in germination, respiration rate, metabolites composition at different seed maturation stages, the competency of mutant seed to withstand the ageing process as well as their post- germinative growth performance upon seedling establishment. Comparative studies of mMDH mutants with wild type by associating the respective phenotypes, physiological and metabolic changes in a spatiotemporal approach could advance our understanding of the role of mMDH in the early phase of the Arabidopsis life cycle.

135

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

Materials and methods

Seed germination assay

Seeds of Arabidopsis thaliana wild type (ecotype Columbia), single mMDH mutants (mmdh1-1, mmdh1-2 and mmdh2-1), double mMDH mutants (mmdh1-1mmdh2-1, mmdh1-2mmdh2-1 and a complemented line (re-introduced a sole MMDH1 cDNA into mmdh1-2mmdh2-1 mutant in which expression of cDNA was driven by a double 35S promoter, namely mmdh1mmdh2 35S: MMDH1) were subjected for seed germination assay. Seeds were sterilized with gas chlorine overnight and were sown on half strength Murashige and Skoog (MS) plates containing 0.8% [w/v] agar, 1% [w/v] sucrose, 1.8 mM MES at pH 5.8 adjusted by KOH. The agar plates were then incubated at 4°C in the dark for 4 days before transferring to a growth room with a short-day light regime (8 h : 16 h, light : dark) at a light intensity of 150 µmol m-2s-1, relative humidity of 70% and temperature cycle of 22°C: 17°C, day: night. The agar plates were placed in a vertical orientation. Examination for seed germination was performed using light microscope and germination was defined as the rupture of testa concomitantly with the protrusion of the radicle. Percentage of seed germination was determined daily for a total of 19 days and accumulative germination percentage was scored accordingly.

Tetrazolium assay to determine seeds viability

Arabidopsis thaliana seeds of wild type, mMDH double mutant (mmdh1-2mmdh2- 1) and complemented line (mmdh1mmdh2 35S: MMDH1) were placed on water- saturated filter paper in a petri dish before they were incubated in cold for 2 days. As a negative control, wild type seeds were heated in distilled water at 99°C for an hour before the cold incubation step. The seed coats were then removed from embryos prior to incubation in 1% (w/v) triphenyl tetrazolium chloride (TTC) solution at 30°C for 3 hours in the dark. The viable seed embryos were observed under light microscope after incubation, indicated as light pink to red stained

136

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants embryos. The percentage of seed viability was subsequently scored by counting the number of viable seed embryos compared to total seed embryos.

Seed maturity identification and collection on the mother plant

Arabidopsis thaliana seeds of wild type (ecotype Columbia), mMDH double mutant (mmdh1-2mmdh2-1) and complemented line (mmdh1mmdh2 35S: MMDH1) were germinated on agar plates (half-strength Gamborg B5 basal salt medium, 2 mM MES, 1% [w/v] sucrose, 1% [w/v] agar, pH 5.7) for approximately 1 week under short-day photoperiod (8 h: 16 h, light: dark), a photon ﬂux of 150 µmol photons m- 2s-1, a relative humidity of 75% and a temperature cycle of 22°C: 17°C, day: night temperature regime. Considering the characteristic of delayed seed germination and slow growth of mmdh1-2mmdh2-1, the growing of mmdh1-2mmdh2-1 seedlings was conducted one week earlier than wild type and complemented line. This was to synchronize the flowering time and discrepancies of seed maturation between mmdh1-2mmdh2-1 and wild type or mmdh1mmdh2 35S: MMDH1. The transferred seedlings were further grown under short-day conditions for about 8 weeks. Three seed maturity stages were of interest in this study, they were designated as green (G), green ripe (GR) and ripe (R) based on the colour indices of the siliques on their mother plant. Full green stage was defined as the green stage of seed excised from green silique, green ripe stage was defined as 40-60% greenish-yellow silique and the ripe seed stage was the stage where silique turned to a full yellowish colour. The seeds from different maturity stages as described above were collected from individual genotype and immediately used for respiration assay or frozen in liquid nitrogen and stored at -80°C for metabolite analysis.

Micro-respiratory assay of mMDH mutant seeds

The 96-well sensor cartridge was hydrated in 100 µL per well XF Calibrant Solution (Seahorse Bioscience) overnight at 37°C the day before the respiration assay. The heater of XF96 Extracellular Analyzer (Seahorse Bioscience) was turned off several

137

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants hours before the commencement and during the course of seed respiration measurement. The 96-well sensor cartridge was calibrated with fresh 200 µL per well XF Calibrant Solution (Seahorse Bioscience). A total of 10 fresh seeds per genotype of different maturity stages collected from individual genotype (wild type, mmdh1-2mmdh2-1 and mmdh1mmdh2 35S: MMDH1) were placed into each well of a 96-well microtiter plate containing 200 µL of seed respiration buffer (5 mM

KH2PO4 , 10 mM TES, 10 mM NaCl, 2 mM MgSO4, pH 7.2). After calibration completed for the sensor cartridge, the 96-well microtiter plate was loaded into the plate reader. The cycling protocol was programmed as 15 loops of 3 min mixing and 5 min measurement steps. The OCR of seeds was recorded by Seahorse XF Acquisition and Analysis Software (Version 1.3; Seahorse Bioscience). The seed viability test was performed on the seeds that were used in respiration assay. Normalised OCR values were obtained after the seed respiration rates were adjusted with seed viability percentage of each genotype.

Seed metabolites extraction for gas chromatography-mass spectrometry (GC-MS) analysis

To approximately 10 mg of seeds per replicate in a 2 mL safe lock micro-centrifuge tube, 100 µL cold metabolite extraction buffer (85% [w/v] HPLC-grade methanol, 15% [w/v] untreated MilliQ water, and 100 ng µL-1 ribitol) was added prior to incubation at 65°C for 30 min in a thermomixer. Then the seeds were spun down by centrifuging at 20000 g for 10 min at 4°C. A volume of 40 µL extracted metabolites solution was transferred into glass insert (Agilent, CA, USA) and dried in a vacuum centrifuge for approximately 3 hours. Subsequently twenty microliters of 20 mg µL-1 methoxylamine-HCl (98% purity; Sigma) was added to each of the dried samples. Samples were then shaken at 1400 rpm for 90 min at 30°C. To each sample, 30 µL of N-methyl-N-(trimethylsilyl)-trifluoroacetamide (derivatization grade; Sigma) was added, followed by shaking again at 1400 rpm for 30 min at 37°C. After this, 10 µL of n-alkane retention index markers (0.029% [v/v] n-dodecane, 0.029% [v/v] n- pentadecane, 0.029% [w/v] n-nonadecane, 0.029% [w/v] n-docosane, 0.029% [w/v]

138

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants n-octacosane, 0.029% [w/v] n-dotriacontane, and 0.029% [w/v] n-hexatriacontane dissolved in anhydrous pyridine) was added and vortexed. The reaction mixture was incubated for 30 min at room temperature prior to injecting 1 µL of the reaction mixture into GC-MS instrument (Agilent, CA, USA). GC-MS data were collected and analysed using Quantitative Analysis (MS) Software (Agilent, CA, USA). The abundance of each metabolite contained in the individual sample was normalised with internal ribitol abundance. Subsequently ratio metabolite abundance value of mutants to wild type was calculated by dividing the normalised metabolite value with mean abundance value of the corresponding metabolite from wild type. These normalised relative metabolites abundance values were then loaded into SPSS statistical software for one way ANOVA with Post-hoc analysis. For metabolite data visualization, the values obtained from log2 transformed ratio metabolite of mutants to wild type were used to generate a heat map for metabolic pathways mapping and for hierarchical clustering analysis in MeV software.

Controlled deterioration test of mMDH mutant seeds

The controlled deterioration test (CDT) was conducted on seeds of wild type, double mMDH mutant (mmdh1-2mmdh2-1) and complemented line (mmdh1mmdh2 35S: MMDH1). A total of 1 mg per genotype (approximately 50 seeds) for wild type and complemented line seeds, while 2 mg per replicate for mmdh1-2mmdh2-1 seeds sealed in a fine poly mesh bag. Approximately 50 replicates (50 bags) for each genotype were prepared for collection at multiple time points during the length of the experiment. They were placed in a petri dish over a

-1 47% relative humidity (RH) lithium chloride solution (385 g LiCl L H2O) within a sealed box and place at room temperature (approximately 20°C) for 7 days to pre- equilibrate the seeds. The initial seed viability of each genotype before ageing treatment (Day 0 ageing day) was determined using a Tetrazolium assay. Heat-killed seeds were used as negative control. For the ageing assay, seeds were transferred to a second sealed box containing 60% relative humidity (RH) lithium chloride

-1 solution (300 g LiCl L H2O) incubated at 45°C in the dark. One seed bag of each

139

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants genotype at a time was removed periodically throughout ageing and sown onto 1% [w/v] agar plates before they were grown under long-day conditions. The emergence of seed radicle was observed after 7 days of growth under a light microscope in order to obtain the seed germination rate. Two batches of ageing assays were performed in this study where the first assay included wild type and mmdh1-2mmdh2-1 seeds while second assay included the complemented line. The ageing durations were 60 days and 45 days respectively for the two different batches of seeds. Probit analysis (Finney, 1982) was performed using excel spreadsheet to create a seed survival curve by plotting seed viability (percentage of seed germination) against the ageing period (days) for each genotype. The viability equation for Probit analysis is as below which has been described previously in Long et al. (2008): Germination (%) = (100-α) / [1+ e –β(t-c)] where α is the fitted initial germination (percentage) at Day 0 ageing, β is the rate of loss of viability (declining log phase of the curve), t is the accumulated ageing period

(expressed in days), and c is the P50 value where viability declines to 50%, the initial germination rate for mutant seeds was normalised to 100% to standardize the discrepancies of initial rate among genotypes for comparison. In each batch of CDT, seed moisture content for Phase 1 (47% RH) and Phase 2 CDT assay (60% RH) of each genotype was determined gravimetrically by weighing seed samples before and after drying in an oven at 105°C for 24 hours. Seed moisture content (%) = (Wet weight of seed-Oven dry weight of seed) x 100 Seed wet weight

Root length assay of mMDH mutants

Seeds of wild type, mmdh1-2mmdh2-1 and mmdh1mmdh2 35S: MMDH1 were surface sterilized via fumigation with chlorine gas overnight and sown onto agar plates (half strength Gamborg B5 basal salt medium, 2 mM MES, 1% [w/v] sucrose, 1.0% agar, pH 5.7). The seeds were subjected to cold stratification for 3 days prior to growing the seedling vertically on agar plates under short-day light regime (8 h

140

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants light/16 h dark), a temperature cycle of 22°C day/ 17°C night, a relative humidity of 75% and an irradiance of 250 µmol m-2 s-1 photosynthetic photon flux density (PPFD). The images of roots in the agar plates were captured and the root length was analysed using ImageJ software (Schneider et al., 2012).

Micro-respiratory assay of mMDH mutants primary root tips and expanded regions

Approximately 3 mm in length root tips or root expanded regions were excised from the primary root of 10-day old seedlings of Arabidopsis wild type (Col-0), mMDH single mutants (mmdh1-2, mmdh2-1), mMDH double mutant (mmdh1- 2mmdh2-1) and complemented line (mmdh1mmdh2 35S: MMDH1) grown in agar plates under short day conditions. For oxygen uptake measurements, a pair of excised root tips (TIP) or root expanded regions (EXP) were placed at the bottom of each well of a 96-well microtiter plate containing 200 µL of respiration buffer (10 mM HEPES, 10 mM MES and 2 mM CaCl2, pH 7.2). The microtiter plate was loaded into XF96 Extracellular Flux Analyzer (Seahorse Bioscience, Billerica, MA, USA) after the calibration steps of sensor cartridge. The time events for both basal respiration measurement and injection were mixing (2 min), waiting (3 min) and measurement (5 min). Fifteen cycles of mixing, waiting and measurement were applied for time course measurements. The OCR of the root tips or expanded regions was recorded by Seahorse XF Acquisition and Analysis Software (Version 1.3; Seahorse Bioscience). The root OCR was adjusted with the calculated root volume in mm3 of the individual genotype. Root volume was calculated as V=πr2h, where V is the volume of a root cylinder, r is the root radius and h is the length of excised root

-1 section. The final mean adjusted root OCR values were expressed as pmolO2 min mm-3.

141

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

Whole root respiration measurement

Whole root oxygen uptake measurement was conducted using the liquid-phase Clark-type Oxygraph system (Hansatech Instruments). A total of 30-40 mg fresh weight of the whole root of wild type, mmdh1-2mmdh2-1 and mmdh1mmdh2 35S: MMDH1 was excised from their corresponding 3-week old seedlings grown on agar plates (half-strength Gamborg B5 basal salt medium, 2 mM MES, 1% [w/v] sucrose, 1% [w/v] agar, pH 5.7) under short day conditions. Before the measurement, the electrode was calibrated at 25°C by adding sodium dithionite to 1 mL of aerated respiration buffer (10 mM HEPES, 10 mM MES and 2 mM CaCl2, pH 7.2) (Atkin et al., 1993; Armstrong et al., 2006). The excised roots were placed into the electrode chamber loaded with 2 mL respiration buffer and OCR measurement was performed for 30 min. The OCR was recorded using the Oxygraph Plus v1.02 software (Hansatech Instruments) and adjusted to fresh weight to obtain OCR per

-1 -1 gram fresh weight (nmol O2 min g FW ) of root tissue.

Results

Consequences of the loss of mMDH on seed germination

All mMDH mutants have been previously characterized as reported in Tomaz et al. (2010) and Tomaz (2012). It is of interest to investigate the impacts of mMDH on seed development and quality in Arabidopsis. A germination assay was performed on wild type seeds of A. thaliana ecotype Columbia, single mutants of mMDH (mmdh1-1, mmdh1-2 and mmdh2-1), double mutants of mMDH (mmdh1- 1mmdh2-1 and mmdh1-2mmdh2-1) and complemented line (mmdh1mmdh2 35S: MMDH1). The seeds were grown on 1% sucrose supplemented MS media agar plates under short-day light regimes for 19 days and seed germination was observed daily using light microscope. The germination was defined as the rupture of the seed coat concomitantly with emergence of the radicle. Representative images captured for seed germination assay at day-0, day-3, day-5, day-7, day-9

142

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants and day-19 were shown in Figure 1. The examination on the agar plates showed that wild type and the majority of the mMDH mutants had started seed germination with visible protrusion of radicles from seeds at day-3 and emergence of cotyledons at day-5 of the assay except for mmdh1-2mmdh2-1. This coincided with detailed microscopic observation where wild type and most of the mMDH mutants achieved at least approximately 50% germination rate at day-2 and full germination rate at day-5 except for both mMDH double mutants (mmdh1- 1mmdh2-1 and mmdh1-2mmdh2-1) (Figure 2). These findings implied that the deletion of a single mMDH isoform did not affect the seed germination rate. In comparison to other mMDH mutants, which had started to germinate at day-1, mmdh1-2mmdh2-1 demonstrated delayed germination with the first emergence of radicle observed only at day-4 and extremely slow germination rate during the course of assay.

Figure 1. Seed germination assay of mMDH mutants. Representative images of seeds growth of mMDH mutants on agar plates at five different time points (Day 0,

143

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

Day 3, Day 5, Day 7, Day 9 and Day 19) under short-day light regimes. Seeds of Arabidopsis thaliana in clockwise orientation of agar plate ecotype Columbia (WT), complemented line (mmdh1mmdh2 35S: MMDH1), double mMDH mutants (mmdh1-1mmdh2-1 and mmdh1-2mmdh2-1), single mMDH mutants (mmdh1-1, mmdh1-2 and mmdh2-1).

Figure 2: Seed germination analysis of mMDH mutants. Single mutants of mMDH (mmdh1-1, mmdh1-2 and mmdh2-1), mMDH double mutants (mmdh1-1mmdh2-1 and mmdh1-2mmdh2-1) and complemented line (mmdh1mmdh2 35S: MMDH1) seeds were grown on agar plates for 19 days under short-day conditions (as shown in Figure 1). The seed germination rate of mMDH mutants were calculated as percentage of radicle emergence (n=20-50) observed under a dissecting microscope. The cumulative percentage of seed germination was shown during the intervals of the germination assay.

Notably, most of the seeds achieved full germination by day-5 with the exception of double mutant seeds. Although, there was lag period for germination of a small number of mmdh1-1mmdh2-1 from day 3 to day-7, a full germination was achieved for this double mutant seed at day-8. However, mmdh1-2mmdh2-1 seeds gave a strikingly low germination percentage only reaching 20% at the end of the assay. Indeed, the fact that seed germination was delayed when both MMDH genes were absent as observed in mmdh1-1mmdh2-1 and mmdh1-2mmdh2-1 with

144

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants a more severe suppression of seed germination demonstrated by mmdh1- 2mmdh2-1. This implies that the mmdh1-2mmdh2-1 mutant carries the null allele of mMDH isoforms. Therefore mmdh1-2mmdh2-1 and its parental single mutant lines namely mmdh1-2 and mmdh2-1 were selected for further experiments. The complementation with the MMDH1 gene in the mmdh1-2mmdh2-1 mutant background could revert the seed germination effectively to wild type levels.

Two possible reasons for the non-germination of a large number of mmdh1- 2mmdh2-1 mutant seeds were considered, firstly it could be due to seed dormancy or secondly a lack of seed viability. Therefore, a seed viability test based on tetrazolium (TZ) staining was performed on excised embryos of non- germinating mmdh1-2mmdh2-1 mutant seeds in parallel to wild type and heat- killed wild type seeds as positive and negative controls, respectively (Figure 3). The TZ test measures the activity of dehydrogenase enzymes used in respiration, which produce free H+ ions and convert 2,3,5 triphenyl tetrazolium chloride (TTC) to formazan, an insoluble red dye (Mosmann, 1983). Results showed that full viability rate for red-stained wild type seed embryos after TZ staining, while non- germinating mmdh1-2mmdh2-1 seed embryos remained unstained as did the negative control. Hence, it was confirmed that the non-germinating mmdh1- 2mmdh2-1 seeds were no longer viable, resulting in a failure to germinate. It is also worth mentioning that dissected seed embryos of mmdh1-2mmdh2-1 revealed a deformed shape with noticeable torpedo and bent cotyledon appearances compared to the enlarged mature seed embryos observed in wild type. These aberrant phenotypes however were not shown in the complemented line mmdh1mmdh2 35S: MMDH1 seed embryos, indicating there was restoration upon reconstitution of MMDH1 cDNA in the complemented line transformants.

145

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

Figure 3. Seed viability assay of wild type, mmdh1-2mmdh2-1 mutant and heat- killed wild type. The embryos excised from wild type seeds (A), non-germinating seeds of mmdh1-2mmdh2-1 mutant from the germination assay (B) and heat-killed wild type seeds (C) were stained with tetrazolium chloride solution and examined under a light microscope. Wild type seeds, which served as positive control in the assay, achieved a full viability rate and heat-killed wild type seeds as negative controls remained colourless. The embryos from non-germinating mmdh1- 2mmdh2-1 seeds were found not viable as no colour change was observed. Notably seed development of the mmdh1-2mmdh2-1 mutant was arrested between torpedo shaped embryos (circled in red) and bent cotyledon (circled in blue) stages.

Respiration analysis of mMDH mutant seeds at different maturation stages

Seed germination potential was greatly reduced to approximately 20% in the mMDH double mutant as evidenced from the above-mentioned germination assay implying an obligatory role of mMDH for producing healthy seeds. A critical role of mMDH in governing leaf respiration rate, where the loss of both mMDH gene isoforms in Arabidopsis plant had elevated the leaf respiration rate significantly, had been documented previously (Tomaz et al., 2010). Taken together with the defective seed germination rate observed in mMDH double mutant, it was hypothesized that seed respiration rate of this mutant was mostly likely to be affected at an early stage. We had previously developed a multiplex micro- respiratory assay for Arabidopsis seeds (Sew et al., 2013; Sew et al., 2015). In this current study, the application of multiplex micro-respiratory assay was used to investigate mMDH mutant seeds and changes in seed respiration. The seeds of individual genotype were harvested from siliques at three designated maturation stages of green (G), transition of green to ripe stage (GR) and ripe (R) as shown in Figure 4. The green and green ripe stages represent early and mid-maturation

146

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants stages of seed, respectively. While the ripe stage is defined as the late maturation stage in which seeds enter a phase of desiccation. The seeds of mmdh1-2mmdh2-1 were smaller in size at all maturation stages compared to wild type. More prominently, a noticeable whitish to pale green phenotype for mmdh1-2mmdh2-1 seeds in the actual green stage was observed which most likely indicates a reduction in chlorophyll amount. In addition, a shrivelled and wrinkled seed coat phenotype seen in mmdh1-2mmdh2-1 seed at the ripe stage could imply a deficit in reserve accumulation. Complemented line seeds were slightly paler green than wild type seeds in the green stage and turned to brown colour in the green ripe stage which might indicate an earlier cessation of photosynthesis.

A B

147

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

Figure 4. Representative images of visible phenotypes of developing siliques and seeds of mMDH mutants at different developmental stages. Siliques (A) and seeds contained in those siliques (B) from wild type, mmdh1-2mmdh2-1 and mmdh1mmdh2 35S: MMDH1 at developmental stages of green, green ripe and ripe.

The seed respiration rate of wild type, mmdh1-2mmdh2-1 and mmdh1mmdh2 35S:mmdh1 at designated maturation stages is shown as mean oxygen consumption rates (OCR) of 10 seeds from individual genotypes in Figure 5A. There were at least 14 independent biological replicates with 10 seeds per replicate for respiration rate assessment in each genotype. For wild type, it was shown that

-1 -1 green seed respired at the highest rate (5.8 pmol O2 min seed ) and declined to

-1 -1 almost half of the initial rate at green ripe and ripe stage (4.1 pmol O2 min seed

-1 -1 and 3.7 pmol O2 min seed respectively) (P<0.01). By contrast, a strikingly high respiration rate was demonstrated by mmdh1-2mmdh2-1 green seeds with

-1 -1 approximately 17.2 pmol O2 min seed which is 2.9 times higher than the wild type rate (P<0.01). Although the respiration rates in mmdh1-2mmdh2-1 seeds declined approximately 2-fold at the two successive maturation stages, the rates remained significantly higher than wild type across all stages (P<0.05). Seeds of mmdh1- 2mmdh2-1 displayed a similar trend to wild type with progressively reduced

148

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants respiration rates as they matured. In comparison, respiration rates of maturing mmdh1mmdh2 35S: MMDH1 seeds were invariant from wild type except for a significantly higher rate at the ripe stage, indicating only a partial restoration to wild type in this complemented line. It is worth noting that different viability among seeds (particularly for mmdh1-2mmdh2-1 based on the previous observations) could complicate the comparison between genotypes. To address this issue, seeds used for micro-respiratory assay were retrieved and subjected to a viability test. The majority of the seeds showed comparable viability rate except for mmdh1- 2mmdh2-1 green seeds (a relatively low viability of 63%) (Table 1). Taking this into account, and assuming non-viable seeds would not respire, seed respiration rates of individual genotype were then normalised by their corresponding viability percentage (Figure 5B). Following normalisation, respiration rates of mmdh1- 2mmdh2-1 seeds were substantially higher across all the maturation stages, contrasting to the subtle differences in wild type and the complemented line. Notably, normalised respiration rates of mmdh1-2mmdh2-1 were approximately 3.7-fold that of wild type in the green seeds while a 2.5-fold increment relative to wild type was observed in the green ripe and ripe seeds. The normalised data could provide a more accurate respiration rates attributed by the viable seeds, allowing better comparative analysis between genotypes. Importantly, it was evident that from both analysis (with and without normalisation based on viability), mmdh1- 2mmdh2-1 seeds respired at consistently high respiration rates throughout the seeds maturation stages.

149

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

B B

Figure 5. Seed respiration analysis of mMDH double mutant (mmdh1-2mmdh2-1) and complemented line (mmdh1mmdh2 35S: MMDH1) in comparison to wild type at different developmental stages. Plants of each line were grown in short-day conditions until siliques formed and seed embryos were dissected from siliques at green (G), transition from green to ripe (GR) and ripe (R) stage. Values represent the mean OCR per seed (A) and mean OCR per seed normalised by the percentage of seed viability in individual genotype (B) (mean±SE, n=14-23). Student’s t-Test showing significant difference between mmdh1-2mmdh2-1 or mmdh1mmdh2 35S: MMDH1 and wild type is marked as black asterisk while the significant difference between mmdh1-2mmdh2-1 and mmdh1mmdh2 35S: MMDH1 is indicated as orange asterisk. Single asterisk (*) indicates significance at P<0.05 and double asterisks (**) at P<0.01.

150

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

Table 1. Percentage of wild type and mMDH mutants seed viability at different maturation stages

Changes of metabolite abundances in wild type and mMDH mutants across seed maturation stages

Following the seed respiration analysis as described above, it was of interest to examine the alteration in primary metabolism which had led to significantly high respiration rate in mMDH double mutant. Therefore, metabolite profiling using GC- MS approach was subsequently conducted on seeds of different maturation stages from wild type, mmdh1-2mmdh2-1 and mmdh1mmdh2 35S: MMDH1. Integration of seed respiration data and metabolite data could improve our knowledge of the role of mMDH in the overall seed maturation process. Figure 6 (A-C) depicts the metabolite profile across different seed maturation stages of mature green, green ripe and ripe stage of wild type, mmdh1-2mmdh2-1 and mmdh1mmdh2 35S: MMDH1, respectively. Data represent the metabolite level determined as relative abundance of each metabolite to internal ribitol of the corresponding sample. The significant changes of metabolites across maturation stages were determined using Student’s t-Test where metabolites at the middle and late maturation stages were

151

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants compared to early stage in individual genotype. For wild type, the level of carbohydrates such as sucrose, glucose, galactose and fructose showed a different abundance pattern throughout the seed maturation stages (Figure 6A). Sucrose levels were comparable across maturation stages, while glucose and galactose increased marginally at the green ripe stage but were not significantly different from in the green stage. However, fructose levels seemed to decline markedly with seed maturity, and were significantly different from the green stage (P<0.05). Glycolysis may be decreased in green ripe and ripe stage because there was a significant reduction of almost 2-fold in the abundance of both glucose 6-phosphate and fructose 6-phosphate compared to the green stage and both were below detectable levels in the ripe stage (P<0.01). Aside from that, there were synchronous metabolite profiles changes observed for a majority of TCA cycle intermediates such as citrate, isocitrate, 2-oxoglutarate and malate across seed maturation stages except for succinate and fumarate. These metabolites were found to be significantly higher in abundance at the green stage and decreased markedly as seed maturation progressed (P<0.05). This was in agreement with the profiles shown by glucose 6-phosphate and fructose 6-phosphate as mentioned above which could mark a decrease in metabolic flux from glycolysis through the TCA cycle during the later stages of seed maturation. Amino acids such as alanine, asparagine and threonine peaked in abundance at the green stage and decreased significantly thereafter at the later stages in wild type seed. In contrast, branched chain amino acids such as leucine, valine and isoleucine were not detected at the green stage of wild type but their abundances increased markedly from middle to the late stage of seed maturation which could be an indication of higher demand for free amino acids when seeds mature. Similarly, shikimate accumulated significantly in the middle and late stages of seed maturation which was most likely accounted for by the significantly increased biosynthesis of its downstream metabolites (tryptophan, phenylalanine and tyrosine) during the seed desiccation stage (P<0.01).

152

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

In contrast to wild type, hexoses of mmdh1-2mmdh2-1 seed such as sucrose, glucose, galactose and fructose were shown to decrease significantly at green ripe and ripe stage (P<0.01) which might reflect a consistently down-regulated sugar metabolism in mmdh1-2mmdh2-1 towards maturity. Similar to wild type, metabolites of mmdh1-2mmdh2-1 seed in glycolysis and the TCA cycle showed substantial accumulation at the green stage but were significantly reduced at the successive maturation stages (P<0.01) implying metabolic flux along respiratory pathways may be decreased with maturity. Notably, all the TCA cycle intermediates in mmdh1-2mmdh2-1 seeds uniformly showed reductions by at least 50%, including succinate and fumarate (P<0.05) Likewise, alanine, serine and its derivative glycine decreased in abundance, which contrasted to an increased abundance of branched chain amino acids (leucine, valine and isoleucine) across the seed maturation stages. Accumulation of threonine, proline and myo-inositol, mannitol and sorbitol, which peaked at the green ripe stage were mostly likely the result of the onset of seed desiccation.

A general feature projected by mmdh1mmdh2 35S: MMDH1 was that a majority of the measured metabolites demonstrated the greatest abundance at the green stage with the exception of asparagine, myo-inositol and metabolites involved in shikimate biosynthesis pathway (shikimate, tryptophan, phenylalanine and tyrosine) as well as mannitol and sorbitol. In other words, almost all the amino acids including branched chain amino acids measured in mmdh1mmdh2 35S: MMDH1 seeds peaked at the early maturation stage and then decreased sharply at the succeeding developmental stages. The contrasting profile of branched chain amino acids observed between mmdh1mmdh2 35S: MMDH1 and mmdh1-2mmdh2-1 as well as wild type, may resulted from the high level of MMDH1 gene products in mmdh1mmdh2 35S: MMDH1 as MMDH1 cDNA complementation is driven by 35S promoters.

AB C

153

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

154

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

Figure 6. The changes of seed metabolites during seed maturation process in wild type (A), mmdh1-2mmdh2-1 (B) and mmdh1mmdh2 35S: MMDH1 (C). For each metabolite, the bars represent changes of metabolites abundance over the maturation stages (green, green ripe and ripe stages in green-orange-brown colour scheme) (mean±SE, n=3-5) where normalised values of each metabolite (normalised by internal ribitol and average total response value across maturation stages) were used. Gray boxes indicate metabolites which were not measured while n.d. marked for metabolites that were not detected in certain maturation stage. Significant differences of metabolites between green ripe or ripe stage and green stage at P<0.05 are marked as * while at P<0.01 are marked as **.

Altered primary metabolism in mMDH mutants seed across developmental stages In the earlier sections, the changes of metabolites from individual genotype across maturation stages were analysed. It is of interest to investigate the impact of disruption of both MMDH gene isoforms on metabolite content of Arabidopsis. Therefore, a comparative analysis of seed metabolite profile between mMDH mutants and wild type at the respective developmental stage was subsequently

155

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants performed to gain an insight into the perturbed primary metabolism in the different mMDH genetic backgrounds. Log2 transformed values of metabolite abundance measured in mutants’ ratio to wild type at the respective maturation stage were used to generate a heat map as depicted in Figure 7. It is important to note that there were a number of amino acids either not detected by GC-MS or they gave diminutive response in wild type at the respective seed maturation stage. As a result, comparison of fold change and the subsequent significance analysis test using Student’s t-Test between mMDH mutants and wild type could not be performed and the respective metabolites are marked as yellow boxes. Results revealed varying abundance profiles for hexoses during seed maturation between mmdh1-2mmdh2-1 and mmdh1mmdh2 35S: MMDH1 compared to WT. Whereas sucrose levels in mutants were comparable to wild type at all stages of seed maturation, other hexoses such as glucose, galactose, and fructose were initially higher than wild type at the early maturation but declined to lower levels than wild type at the succeeding maturation stages. While metabolites along the glycolysis such as glucose 6-phosphate and fructose 6-phosphate were approximately 2-fold and 4-fold higher in both mutants than wild type at green and green ripe stage, respectively. Conversely, both mutants showed collectively and significantly decrease in TCA cycle intermediate abundance ratios to WT at all stages of maturation. More prominently, there was massive and significant decrease in the relative abundance of 2-oxoglutarate by 4 to 6-fold across maturation stages in mmdh1-2mmdh2-1 seeds compared to WT seeds. Interestingly, the relative abundance of the derivatives of 3-phosphoglycerate such as serine, glycine in mmdh1-2mmdh2-1 mutant were found to be higher during maturation stages (P<0.05), particularly in the green stage where approximately 22-fold change relative to wild type. Interestingly, a consistently and significantly high abundance of free amino acid such as isoleucine, valine, leucine, alanine, proline and threonine at all stages of maturation of mmdh1-2mmdh2-1 mutant (P<0.05) compared to WT. The profiles of these free amino acids in mmdh1mmdh2 35S: MMDH1 seemed to be more dynamic and responses were not as strong as in mmdh1-2mmdh2-1.

156

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

Importantly, the complemented line seemed to be able to alleviate the high relative abundance of a majority of the metabolites seen in mmdh1-2mmdh2-1.

Figure 7. Seed metabolite changes in mMDH mutants relative to wild type during maturation stages. Relative mean ribitol normalised metabolite levels to wild type (n=3-5) of the corresponding stages in log2 scale were shown in blue to red colour scheme which denotes for lower to higher metabolite ratios. Boxes with yellow borders indicate that metabolites were detected in a mutant but not in the wild type at the respective stage, therefore a fold change could not be determined. Light grey boxes denote for metabolites which were not measured in the study, while dark grey boxes indicate undetected metabolites at certain maturation stage. Metabolite levels of mutants which differed significantly from wild type at the corresponding maturation stages were marked as single asterisk (*) and double asterisks (** ) for P<0.05 and P<0.01 respectively.

157

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

Seed ageing assay on mMDH mutants with the controlled deterioration test

In regards to the defective seed viability and germination rate exhibited by double mutant seed, a key question is whether the loss of MMDH gene had accelerated seed ageing effects that accounted for a decrease in seed storage potential. Therefore we next conducted controlled deterioration tests (CDT) on mMDH mutants to further investigate the effects of losing mMDH expression in Arabidopsis seeds. Two independent CDT assays were conducted in this current study, the first CDT assay consisted of seeds of wild type and mMDH double mutant (mmdh1- 2mmdh2-1) and the second assay included a complemented line (mmdh1mmdh2 35S: MMDH1) to prove that the phenotype was linked to the decreased mMDH protein amount. CDT can be divided into three phases: Phase 1 is known as rehydration step of dry seeds in which seeds were subjected to homogenous moisture in a sealed container with 45% relative humidity (RH) for a week at room temperature to minimise the change in seed moisture content before switching to artificial ageing treatment in Phase 2. The ageing condition for Phase 2 is RH=60% for 40-60 days at 45°C where the increase in temperature and humidity could accelerate the seed ageing process. Due to the fact that the viability of a seed lot at a certain time point is not only affected by ageing but also relies on the initial germination rate of the seed lot, the tetrazolium assay for seed viability was carried out to obtain the percentage of viable seeds at the beginning and the end of ageing assay. Images shown in Figure 8 are aged seed embryos of individual genotype including a heat killed wild type as control after tetrazolium staining conducted as part of the first seed ageing assay (A-G) and second seed ageing assay (a-h) respectively. In the first seed ageing assay, the initial germination rate for wild type

B and mmdh1-2mmdh2-1 seeds at Day 0 ageing were 95% and 77.3% respectively. In contrast, after 45 days of ageing at 45°C and 60% relative humidity, the seeds for both mutant and wild type were no longer viable. Whereas in the second ageing assay, at Day 0 of ageing test, the percentage of germination of WT, mmdh1mmdh2 35S: MMDH1 and mmdh1-2mmdh2-1 were 89%, 83% and 43% respectively and none of the individual genotype seeds was viable after 35 days of ageing.

158

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

In order to construct seed survival curves, the germination rates of aged seeds in the first and second CDT assay were scored and analysed using the Probit regression model. Probit analysis is a linear regression method which takes into account the varying precision of data points produced by the log transformation. For a fair comparison between genotypes with different initial germination rates, the percentage of seed germination was normalised with the measured viability rates prior to constructing the seed survival curve for each ageing assay. When the normalised percentage of seed germination was plotted against storage period or ageing day, a sigmoidal survival curve that represents a negative cumulative normal distribution (Ellis and Roberts, 1980) was obtained as depicted in Figure 9. Based on the seed survival curve of first seed ageing assay, it was predicted that the longevity of wild type and mMDH double mutant (mmdh1-2mmdh2-1) seeds were approximately 55 and 30 days respectively, as indicated by the intercepts between the survival curves with x-axis (day of seed storage period). The loss of MMDH genes led to a shortage of 25 days longevity (approximately 45% reduction) compared to wild type. Notably, mmdh1-2mmdh2-1 seeds showed a sharp decline in germination rate after 10 days of ageing in contrast to wild type seeds, with a rapid decrease in germination rate only after Day 22 of the ageing test. In second seed ageing assay, the interval of ageing was much shorter due to a generally lower initial percentage of seed germination of the individual genotype as compared to first ageing assay. It was noticed that mmdh1-2mmdh2-1 seeds demonstrated the steepest decline, and this was followed by mmdh1mmdh2 35S: MMDH1 and wild type seeds. The rate of seed germination in mmdh1-2mmdh2-1 reduced very significantly at the early time points (Day 1 to Day 5 of ageing test) just after entering the Phase 2 of CDT, contrasting with wild type seeds that deteriorated gradually and constantly throughout the course of ageing test. It was interesting to note that mmdh1mmdh2 35S: MMDH1 seeds declined moderately with a rate between WT and mmdh1-2mmdh2-1. The predicted longevity of mmdh1mmdh2 35S: MMDH1 seeds as indicated by seed survival curve was approximately 25 days which was 3 times more than mmdh1-2mmdh2-1 but 1.6 times less than wild type.

159

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

These results implied that was a partial restoration to wild type-like longevity observed in mmdh1mmdh2 35S: MMDH1 seeds upon reconstituting a sole, constitutively expressed MMDH1 gene in that complemented line.

Figure 8. Light microscope observation of seed viability of mMDH mutants using the tetrazolium (TZ) staining assay. Images of representative seed embryos of wild type (A), mmdh1-2mmdh2-1 (B) and heat-killed wild type (C) at Day 0 ageing of the first CDT test after TZ staining assay. The remaining seeds were aged at 45°C and 60% relative humidity for 45 days followed by a second TZ staining assay conducted on wild type (D), mmdh1-2mmdh2-1 (E) and heat-killed wild type (F). While images a-h show TZ staining results for the second CDT test, wild type (a), mmdh1-2mmdh2-1 (b), mmdh1mmdh2 35S: MMDH1 (c) and heat-killed wild type (d) at Day 0 ageing and wild type (e), mmdh1-2mmdh2-1 (f), mmdh1mmdh2 35S: MMDH1 (g) and heat- killed wild type (h) after 35 days of ageing treatment, respectively.

160

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

Figure 9. Seed survival curves for wild type, mMDH double mutant (mmdh1- 2mmdh2-1), complemented line (mmdh1mmdh2 35S: MMDH1). Seeds were aged at 60% RH and 45°C for 45 days (A) and 35 days (B) prior to seed germination assays in long -day lighting regimes. The percentage of seed germination was determined at day 7 after seeds sown on agar plate and indicated as dots in the chart for WT (red), mmdh1-2mmdh2-1 (blue) and mmdh1mmdh2 35S: MMDH1 (green). The fitted seed survival curve is plotted according to Probit analysis based on cumulative normalised seed germination data of each genotype.

Analysis of root growth in mMDH mutants

Implication of the loss of MMDH gene(s) on Arabidopsis root growth has not been well defined to date. Data gathered from various perspectives of mMDH mutants as e.g. their respiration, germination and metabolism showed significant alterations from wild type levels and had led us to further investigate the post-germinative growth of mMDH mutants. Therefore, root length assays were conducted on

161

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants mMDH single and double mutant lines (mmdh1-1, mmdh1-2, mmdh2-1 and mmdh1-2mmdh2-1), together with a mMDH complemented line (mmdh1mmdh2 35S: MMDH1) and wild type (Col-0). The seedlings of each line were grown vertically on half-strength MS medium under short-day conditions. The representative images of mMDH mutant seedlings on agar plates grown under short-day conditions for 14 days are shown in Figure 10.

(A) (B)

1 1 cm cm

Figure 10. Root length assay of 14-day old mMDH mutants seedlings. Representative images of root growth of mmdh1-2 (A), mmdh2-1 (B), mmdh1- 2mmdh2-1 (C) and mmdh1mmdh2 35S: MMDH1 (D) compared to wild type. Seeds of each genotype were sown on a vertically oriented agar plate (n=9 except mmdh1-2mmdh2-1 n=20) and seedlings were grown under short-day conditions.

In general, there were distinctly shorter root lengths and noticeable smaller root diameters between mmdh1-2mmdh2-1 and wild type seedlings. The measurement of root length was performed on Day-10 and Day-16 old seedlings of those mutant

162

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants lines using ImageJ software. Analysis of both Day-10 and Day-16 root length measurements showed that mmdh1-2mmdh2-1 (double knockout line of MMDH isoforms) gave the shortest root length among all the mutant lines (Figure 11).

Figure 11. Root length analysis of mMDH mutants seedlings. Values represent primary root length of seedlings in centimetre (mean ± SE, n=24-72) of mMDH single mutants (mmdh1-2 and mmdh2-1), a mMDH double mutant (mmdh1- 2mmdh2-1) and a complemented line (mmdh1mmdh2 35S: MMDH1) compared to wild type (WT) grown on vertically oriented agar plates for 10 days (A) and 16 days (B) under short-day conditions. Student’s t-Test analysis showed root length that differed significantly between mutants and wild type on the same plate at P<0.01 are marked with double asterisks (**).

163

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

Convincingly, there were significant differences with 62% and 65% reduction in root length of 10-day and 16-day old mmdh1-2mmdh2-1 seedlings respectively as compared to wild type (P<0.01). While mMDH single mutants demonstrated varying root lengths, a slight but significant reduction in root length by 6% and 10% relative to wild type seen in 10-day and 16-day old mmdh1-2 mutant seedlings respectively (P<0.01) notwithstanding a comparable root length to wild type as observed in mmdh2-1 mutant. The replacement of a sole MMDH1 cDNA in double mutant background largely restored the stunted root growth of mmdh1- 2mmdh2-1 seedlings, but they were still approximately 15-18% reduced in length compared to wild type (P<0.01). The above findings indicated that there were restrictions in root growth in Arabidopsis when MMDH was withdrawn and more prominently with the absence of both MMDH isoforms.

Root respiration of mMDH mutants

The fact that disruption of both mMDH genes in Arabidopsis leaf caused significantly elevated respiration rate established an important link between the functional role of mitochondrial malate dehydrogenase and mitochondrial respiration in Arabidopsis (Tomaz et al., 2010). In view of that, root respiration of mMDH mutants is another interesting aspect to be explored and these data could potentially bridge the root growth phenotypes in mMDH mutants and the exclusive role of MMDH gene in root mitochondrial respiration. Multiplex micro- respiratory assay performed previously on root parts of Arabidopsis wild type (Col-0) seedling showed that there were different respiration rates between root tips and expanded regions of root (Sew et al., 2013). Therefore, in this current study, this approach was used to investigate the effects of mMDH loss on root respiration rate in the mMDH mutants. Approximately 3 mm root tip (TIPS) or root expanded regions (EXPS) were excised from 12-day old short-day grown roots of wild type and mMDH mutants seedlings and placed in a 96-well microtiter plate prior to respiratory measurement using the XF96 Extracellular Analyser (Seahorse Bioscience, Billerica, MA). Those mutants lines were single mMDH knockout lines (mmdh1-2 and

164

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants mmdh2-1), a double mMDH knockout line (mmdh1-2mmdh2-1) and a complemented line (mmdh1mmdh2 35S: MMDH1) and the respiratory rates were measured as oxygen consumption rates (OCR) of two root sections (TIPS or EXPS) per well. Based on ImageJ analysis of the primary root width of mMDH mutants, it was found that their root diameters were significantly different from wild type (P<0.05 for mmdh1-2mmdh2-1 and P<0.01 for other mMDH mutants) (Figure 12). However there was no noticeable variation between root tip and expanded root region in individual genotype according to Student’s t-Test analysis. Significant differences of root width presumably resulted in marked variation of root volumes across genotypes which could potentially lead to discrepancies in the comparison of root OCR among genotypes. Hence, mean OCR data of individual lines were adjusted with respective root volume as shown in Figure 13.

Figure 12: Variations of root sections width among mMDH mutants. Values represent width of expanded region (EXP) and tip (TIP) from the primary root in millimetre (mean±SE, n=10-35) measured from 12-day old A. thaliana seedlings of wild type (WT), mMDH single (mmdh1-2 and mmdh2-1), mMDH double mutant (mmdh1-2mmdh2-1) and complemented line (mmdh1mmdh2 35S: MMDH1). Root width of mMDH mutants that significantly different from wild type based on Student’ s t-Test analysis is marked as single asterisk (*) or double asterisks (**) for P<0.05 and P<0.01, respectively.

165

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

Figure 13. Oxygen consumption rates of root expanded region (EXP) and root tip region (TIP) excised from 12-day old seedlings of A. thaliana wild type, mMDH single (mmdh1-2 and mmdh2-1) and mMDH double mutant (mmdh1-2mmdh2-1) and complemented line (mmdh1mmdh2 35S: MMDH1). The seedlings were grown on vertically oriented agar plates under short-day light regimes. Values represent the adjusted OCR with root volume in mm3 (mean±SE, n=8-21). Student’s t-Test analysis shows significant differences between mutant OCR and wild type of their respective root part at P<0.05 (marked as *) and P<0.01 (marked as **), while significant differences between mutants and complemented line at P<0.05 (marked as *) and P<0.01 (marked as **).

Results revealed that oxygen uptake rates by expanded root regions (EXPS) were generally lower compared to root tips (TIPS) across different genotypes. The observed changes in OCR ranged from 1.8-fold to 2.6-fold increase in root tips compared to EXPS of the corresponding genotype (P<0.01). These results reflected a general higher oxygen demand in root tips compared to expanded root regions. This current results is consistent with our previous report where the oxygen uptake of wild type root tips was significantly higher than of expanded regions (Sew et al., 2013). Both EXPs and TIPs of mmdh1-2mmdh2-1 mutants respired at significantly higher rates compared to wild type (P<0.01). Respiration rate of single mMDH mutant, mmdh1-2 was significantly higher in root tips compared to wild type (P<0.01) while expanded region of mmdh2-1 root respired at a significantly and relatively lower rate than wild type (P<0.05). The slight difference between mmdh1- 2 mutant and wild type in EXP regions was not statistically significant. However, the

166

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

TIP region of this single mutant exhibited significantly high respiration rate than wild type with approximately a 1.5-fold higher rate (P<0.01). Similarly, the complemented line was found to respire at a comparable rate to wild type in the root EXP region but not in the TIP region in which a significant 1.3-fold higher rate was measured. These results imply that mmdh1mmdh2 35S: MMDH1 was not able to fully restore the abnormally high rate of mmdh1-2mmdh2-1 to wild type levels at the root tip region but was able to in the expanded region.

In view of the significant elevated primary root respiration rates in both TIP and EXP of mMDH double mutant, we decided to investigate the alteration or difference in the entire root respiration of this mutant. Therefore, we next conducted a respiration assay using the whole root of 3-week old seedlings of wild type, mmdh1-2mmdh2-1 and mmdh1mmdh2 35S: MMDH1 mutants grown on vertical agar plates under short-day conditions in a conventional liquid phase Clark-type oxygen electrode. It was noticed that the whole root respiration rate was 1.5-fold higher in mmdh1-2mmdh2-1 compared to wild type (P<0.01) whereas mmdh1mmdh2 35S: MMDH1 showed a slightly lower rate than mmdh1-2mmdh2-1 but significantly different to wild type (increased by 1.4-fold) (P<0.05) (Figure 14). These data showed that mMDH double mutant, mmdh1-2mmdh2-1 gave significantly high whole root respiration and complementation with a sole MMDH1 gene in complemented line was not able to fully restore the aberrant respiration rate of mmdh1-2mmdh2-1.

167

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

Figure 14: Comparison of whole root respiration rates among wild type, mmdh1- 2mmdh2-1 and mmdh1mmdh2 35S: MMDH1. Values represent the mean OCR (mean±SE, n=4) expressed in gram per fresh weight of the whole root tissues which were excised from 3-week old seedlings of the corresponding genotype grown in agar plates under short-day conditions. Differences of respiration rates between mmdh1-2mmdh2-1 or mmdh1mmdh2 35S: MMDH1 and wild type were deemed to be significant at P<0.05 (*) and at P<0.01 (**).

Discussion

Arrested seed development leads to defective germination in mmdh1-2mmdh2-1

Germination rates of mMDH single mutants (mmdh1-1, mmdh1-2 and mmdh2-1) were found to be comparable to wild type implying there might be functional redundancy in mMDH gene isoforms to compensate for the loss of a gene counterpart during seed germination. In contrast, both mMDH double mutants (mmdh1-1mmdh2-1 and mmdh1-2mmdh2-1) showed some delay in seed germination notwithstanding there was a distinct difference in the final percentage of germinated seed between the two mutant lines. Clearly, a more severe suppression of seed germination was observed in mmdh1-2mmdh2-1 with approximately 80% reduction in germination rate in comparison to DKOA which achieved a full germination when the growth period on agar plates was extended. These results imply that mmdh1-2mmdh2-1 mutant is a better candidate for

168

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants mMDH loss-of-function studies. These findings were consistent with the previous observations that zero survival rate for mmdh1-2mmdh2-1 plant grown under outdoor conditions (Tomaz et al., 2010), which suggested that the loss of both mMDH gene isoforms had severely reduced seed production in Arabidopsis.

A detailed microscopic observation of non-germinating seeds of mmdh1-2mmdh2- 1 confirmed that the subset of seeds that don’t germinate were truly non-viable showing deformed embryo shapes. It can be suggested that early embryo morphogenesis which involves cell divisions in seed, ends at heart shaped or torpedo-shaped embryo (approximately 7-8 DAF) where the basic body pattern of the embryo is usually formed (Mayer et al., 1991; Baud et al., 2002). As seeds like those in Arabidopsis are known to be heterotrophic organs, their growth and development generally rely on the nutrient loading from the mother plant (Zhang et al., 2007). This process is particularly crucial during early embryo development, where nutrients are delivered to sink tissues from the maternal source tissues (Hua et al., 2012). Tomaz and co -workers found that mMDH double mutant plants exhibited a significant high leaf respiration rate, low net CO2 assimilation and photorespiratory defects with an albeit small rosette and slow growing phenotypes (Tomaz et al., 2010). It was suggested that carbon and energy were directed away by the high respiration rate exhibited by the mMDH double mutant. In addition, it was reported elsewhere that efficiency of carbon fixation and recovery as well as nitrogen assimilation were severely affected in other photorespiratory mutants (Somerville and Ogren, 1980, 1981; Collakova et al., 2008). It can therefore be hypothesised that there might be insufficient supply of carbon and nitrogen sources from the mother plant to the seed embryos for growth and development in the mMDH double knockout plant.

In general, a significant positive correlation between oxygen consumption rate during seed imbibition and later stages of germination and seedling establishment has been documented (Woodstock and Grabe, 1967). Respiration and metabolic activities in seeds are constitutively resumed upon rehydration of seeds, and energy generation from respiration is used to mobilize stored reserve and power seed

169

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants germination. Although it was proposed that respiratory control was predominantly associated with the alternative respiratory pathway (cyanide-insensitive) at the earliest stages, there is a transitory changes to a cytochrome oxidase (cyanide- sensitive) respiration at the successive stages in germinating seeds (Yentur and Leopold, 1976). This is in line with the observation that the resumption of a normal cytochromic pathway for ATP synthesis during the first minutes and hours after the imbibition of lettuce seeds (Hourmant and Pradet, 1981). Previous study has shown that mMDH contributed the highest control coefficient flux for respiration among all other TCA cycle enzymes in tomato plants, implying a pronounced effect of mMDH in plant respiratory metabolism (Araujo et al., 2012). This is supported by the findings on mitochondrial MDH activity which was found to contribute about 16% of the 7-fold increased total MDH activity extracted from the cotyledons after day 4 of seed germination in watermelon (Walk and Hock, 1976). Proteomic analysis has revealed that there were a total of 95 significantly accumulated proteins during Arabidopsis seed germination (proteins involved in carbohydrate, energy and amino acid metabolism), mMDH (At1g53240) was one of identified proteins (Fu et al., 2005). Also, a comprehensive Arabidopsis seed transcriptome analysis documented that transcript abundance of mMDH was up-regulated by 2- fold at the early stage seed germination (Weitbrecht et al., 2011). All the above findings together with the data presented here that mutation on both mMDH genes in Arabidopsis resulted a marked decline in seed germination rate, clearly reflecting an obligatory role of these genes in coordinating a successful and timely germination event.

Physiological and metabolic changes of Arabidopsis wild type seeds during maturation stages

From the results of micro-respiratory assays of Arabidopsis wild type seeds, the OCR was highest in the early maturation stage (green stage) and decreasing to almost half of the initial rate at the late maturation stage (green ripe stage) and desiccation

170

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants stage (ripe stage). Seed respiration as a more direct measure of metabolism, signifies the energy demand and gross metabolic state of seed at a particular developmental or maturation stage. This distinct decline in seed respiratory capacity as seeds mature was most likely to be linked to the reduction in seed moisture content as maturation progressed. In general, the moisture content of seed gradually decreases to 5-15% (fresh weight basis) during seed developmental stages depending on the species (Macherel et al., 2007), switching from soggy to a completely dry state. Previous findings have pointed out that changes in physiological activities such as oxygen uptake, photosynthetic electron transport, dormancy breaking and photochemical reactions in seeds are associated with the moisture content of seeds (Leopold and Vertucci, 1989; Vertucci, 1989). More specifically, varying oxygen uptake rates were shown to correspond to discrete hydration levels in soybean and mitochondrial respiration was suggested only to be effective at the highest level where water content was increased above 25% (Vertucci, 1989). Respiratory activities of shrub seeds (Machilus thunbergia) were found to decline during the maturation stage and reached zero at the end of desiccation suggesting a positive relationship between seed respiration and oxygen demand corresponding to a gradual loss in seed moisture content during the maturation stages (Lin and Chen, 1995). Taken together with the above findings, respiration is clearly a dynamic and developmental dependent process in seed.

Growing seed embryos are generally heterotrophic and rely on mitochondrial respiration. However, as seeds develop and mature, a gradual reduction of oxygen concentration from the seed coat towards the centre, developing an internal hypoxic environment within the seed (Rolletschek et al., 2002; Borisjuk and Rolletschek, 2009; Caccere et al., 2013). Seeds are thus subjected to a more severe oxygen limiting condition as maturation progresses, exacerbated by declination in its photosynthetic activity during the process. Studies have pointed out that seeds of several plant species in the early maturation stages are rich in chloroplasts and photosynthetically active (Ruuska et al., 2002; Rolletschek et al., 2003; Ruuska et al., 2004; Rolletschek et al., 2005). Moreover, a good correlation was found between

171

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants the chlorophyll content and photosynthesis-dependent oxygen evolution meaning that embryonic photosynthetic activity is responsible for maintaining the internal oxygen level in the seed (Eastmond et al., 1996). In addition, it was evident that the chlorophyll content, photosynthetic oxygen production and greening pattern conformed to the ATP distribution pattern within embryos of faba bean (Vicia faba) in both a spatial and temporal manner (Borisjuk et al., 2003). Recently, photosynthetic capacity acquired by the seed embryo at the green stage was shown to improve the efficiency of seed respiration as a considerable amount of oxygen is being supplied and can lower the formation of lactic acid (Caccere et al., 2013). Therefore, seed mitochondrial respiratory capacity is presumably increased during the greening process of the seed due to oxygen availability and ATP need, consistent with the highest rate of respiration observed in green seed of wild type (also in mutants). While a cessation of photosynthesis activity at the green ripe and ripe stage resulted decreased oxygen availability, less energy production and thus leading to a sharp decrease of seed respiration. Results of both seed respiration rate and metabolites abundances corresponded well to each other where seed respiration rates significantly declined concomitant with a marked reduction in metabolites in respiratory pathways from glycolysis to TCA cycle at the mid- and late maturation phases of seed. This is explained by an initiation of the metabolic quiescence in seeds, resulting in a drastic drop in energy demand for metabolic cellular processes as seeds become mature. The data presented here are in line with previous findings that metabolites of glycolysis and TCA cycle declined with seed maturity in Arabidopsis (Fait et al., 2006), and a diminished flux through the TCA cycle reactions at the late maturation stages of B. napus seed (Schwender et al., 2006).

It was interesting to note that branched-chain amino acids (BCAAs) such as leucine, valine and isoleucine which were not detected at the green stage of wild type, increased substantially and constantly at the later stages of seed maturation. This contrasting abundance pattern demonstrated by BCAAs, may indicate a higher use of these amino acids for energy generation in respiratory chain towards the end of

172

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants seed desiccation. Several studies have shown that significant accumulations of BCAAs and isovaleryl-CoA (an intermediate in BCAAs metabolism) during dark- induced senescence or extended darkness (Ishizaki et al., 2005; Ishizaki et al., 2006; Araujo et al., 2010). Under such carbon starvation conditions, protein degradation followed by a successive breakdown of amino acids could provide electrons directly for the electron transport chain (ETC) via electron-transfer flavoprotein (ETF) complex or alternatively products from amino acid catabolism could be fed into the TCA cycle to drive respiration (Araujo et al., 2011). This is reasonably explained by the fact that seeds are starved of limited carbon sources when the photosynthesis ceases and a decrease in sucrose supply occurs from mother plant. However some degrees of energy level have to be sustained for the ongoing metabolic processes such as reserve accumulation and the lignification process in the drying seed. Degradation of free amino acids can be seen as an alternative carbon for metabolic process and ATP source via the respiratory chain to circumvent a number of limitations as mentioned earlier as well as the reducing availability of oxygen and dehydration process in maturing seed. Notably, shikimate showed significantly higher abundances from the onset of desiccation, and in concordance with its derivative aromatic amino acids (tryptophan, phenylalanine and tyrosine) which showed substantial accumulations only in the desiccating seeds. These findings reflected that there is a particular need for these amino acids during the seed drying process. It was formerly suggested that accumulation of phenolic compounds is associated with their role in strengthening cell walls during seed desiccation contributing to the establishment of the hard brown seed coat in bean (Yeung and Cavey, 1990). This is consistent with the findings that two types of flavonoid namely proanthocyanidins (PAs) or condensed tannins, and flavonols were accumulated during Arabidopsis seed development as reported previously (Routaboul et al., 2006). While a laccase was found to oxidise the colourless of PAs and thus browning of the seed coat in Arabidopsis (Pourcel et al., 2005). More recently, it was shown that phenylalanine is involved in the biosynthesis of lignin

173

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants which is a phenylpropanoid polymer in secondary cell walls in the seed coat of plants (Chen et al., 2012).

Elevated seed respiration, perturbed metabolic processes and expedited seed deterioration occurs in Arabidopsis mMDH mutants

The seed respiration analysis revealed distinct seed respiration patterns between mMDH mutants and wild type corresponding to the seed maturation process. The Arabidopsis mMDH double mutant (mmdh1-2mmdh2-1) demonstrated significantly higher respiration rates at all stages of seed maturation indicating the importance of functional mMDH gene isoforms for seed respiratory metabolism. Complemented line transformant seeds were found to respire at rates that were higher than wild type across maturation stages, with the exception of the ripe stage. This implies that complementation of the null mMDH background with a sole MMDH1 cDNA as in mmdh1mmdh2 35S: MMDH1 was not fully sufficient to restore seed respiration rates to wild type levels. Furthermore, the perturbed respiratory metabolism of both mMDH mutants compared to wild type could be seen from uniform metabolite profiles of intermediates in respiratory pathways of those mutants. They were consistently up-regulated in the intermediate steps of glycolysis, but showed an opposite profile of an overall decrease in TCA cycle intermediates across maturation stages in both mutants. It seems as if the metabolic fluxes from increased abundance of glycolysis intermediates were being diverted to other metabolic pathways such as amino acid biosynthesis, plant structural and secondary metabolism, rather than through the TCA cycle. Considering respiration via the TCA cycle involves a series of respiratory substrate oxidations which requires a generous amount of oxygen, intriguingly, mMDH mutants exhibit a lower rate of aerobic respiration and need to be adapted to respire in more hypoxic surroundings. This is supported by the physical appearance of mMDH double mutant seeds that have lower amounts of chlorophyll at the green stage, and thus lower photosynthetic activity, likely resulting in deficient internal oxygen compared to wild type. Moreover, an earlier seed desiccation and

174

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants browning event in both mMDH mutants could reflect a shorter period of photosynthesis than wild type. Taking into account all the above, seeds of mMDH mutants are most likely suffering from limiting oxygen availability and therefore respire both aerobically and anaerobically throughout the maturation. While seeds of mMDH mutants may be exposed to a greater anoxic stress than wild type, several other sources such as sugar, protein and amino acids could support anaerobic metabolism. Sugar is considered to be the immediate carbon source through anaerobic glycolysis to sustain energy production (Rodrigues et al., 2006). This is in agreement with a number of sugar-related compounds such as sucrose, glucose and fructose (hexoses), mannitol and sorbitol (sugar alcohol), glucose-6- phosphate and fructose-6-phosphate (glycolysis intermediates) which were constantly maintained at a relatively high levels in mutants over the entire period of seed maturation.

There are several levels of explanation for a significantly increased amino acid pool consisting of alanine, serine, glycine, threonine, proline and BCAAs particularly during the seed maturation stages of mmdh1-2mmdh2-1. Firstly, protein degradation leading to an increase of free amino acid pool is known to be a response elicited by stresses, which helps to eliminate harmful aggregates or proteins from accumulating in the cells (Flick and Kaiser, 2012). Under low moisture environments, degradation of protein works to release free amino acids and thus could maintain osmotic potential of tissue (Takeba, 1980, 1980). While in a low oxygen tension such as hypoxia, free amino acids could accumulate as a result of a decline in protein synthesis rate under shortage of energy (Bertani and Brambilla, 1982; Reggiani et al., 1988). Also, catabolism of BCAAs specifically has been shown as an oxidative phosphorylation energy source that stimulates mitochondrial oxygen consumption (Taylor et al., 2004). This is in agreement with other studies that catabolism of amino acids serve as an energy compensatory mechanism which allows for ATP production under energy deprivation or stress-related conditions (Araujo et al., 2011; Galili, 2011). Aside from that, a plausible link between accumulation of osmolytes attributed by free amino acids and soluble sugars has

175

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants been reported as a mechanism to maintain osmotic homeostasis in maize seedling under salinization conditions (Chyzhykova and Palladina, 2006). In comparison with the mMDH double mutant, it can be hypothesized that the oxygen tension in mmdh1mmdh2 35S: MMDH1 seeds could be alleviated as the fold change of a majority of amino acids was relatively lower.

The significantly up-regulated amino acid metabolism in mMDH mutants seems to have increased the conversion of phosphoenolpyruvate to pyruvate, resulting in reduced abundances of shikimic acid pathway derivatives. However, phenylalanine and tryptophan contents in desiccating seeds of those mutants were relatively high compared to wild type, particularly a significant increased level of phenylalanine was evident in mmdh1-2mmdh2-1 seeds. Intriguingly, knockout of both mMDH isoforms could have resulted in a higher demand for phenylalanine which is the substrate of cinnamic acid via phenylalanine ammonia-lyase (PAL) for lignin biosynthesis pathways (Weng and Chapple, 2010). This is supported by findings that an increase in phenylalanine correlated with a higher lignin content in maize pulvini (Zhang et al., 2011). A higher lignin content may be required for mMDH double mutant desiccating seeds which helps to reduce seed pore size in respond to water stress, however this could account for a reduced water and solute permeability into the seed during imbibition (Liang et al., 2006). Seed coat permeability is one of the crucial factors for seed germination, poor quality of seed coat with low permeability was shown to cause slow germination or low germination rate in seed (Calero et al., 1981; Nooden et al., 1985; Tezuka et al., 2012). This might be a plausible explanation for a noticeable lagging in the germination and possibly decreased germination rate of mMDH double mutant seeds due to a higher lignin composition in their seed coats that hampers water influx into the seed during imbibition.

Use of the controlled deterioration test (CDT) on mMDH mutants and wild type seeds demonstrated differential longevity profiles corresponding to their genetic and physiological storage potential. Both, the first and second CDT experiments revealed an accelerated seed ageing process with up to a 5-fold reduction of mMDH double mutant longevity compared to wild type. Complementation with MMDH1

176

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants cDNA was not sufficient to fully restore the altered longevity in mMDH double mutant, implying an important role of mMDH2, or timing of MMDH1 expression, which need further investigation. Interestingly, seed respiration defects was shown to be related to the primary lesions in mitochondria resulted from the pre-harvest deterioration following a high water content and high temperature storage condition in soybean (Amable and Obendorf, 1986). Nevertheless, alterations in the membrane integrity of mitochondria as well as mitochondrial functions were severely affected with reduced oxidation rates and respiratory controls in fast respiring, imbibed aged pea seeds as reported previously (Benamar et al., 2003). The damaged mitochondria were believed to be responsible for high oxygen consumption and impaired ATP production in the aged seeds that ultimately resulted in the slow rate and low percentage of seed germination. In this current study, we found consistently high respiring mMDH double mutant seeds during the pre-harvest period, showing that this genetic dependent respiratory defect was persisted in prolonged post-harvest storage. This phenomenon needs further investigation and it could potentially explain the marked reduction of seed germination rate in the mMDH double mutant.

The above-mentioned findings related to oxidative property changes in aged seed mitochondria could be plausibly linked with oxidative stress in deteriorating seeds. It has been well established that oxidative stress is an important cause of seed deterioration during natural or artificial ageing of many plant species (Sun and Leopold, 1995; Goel et al., 2003; Spano et al., 2011; Chen et al., 2013). Recently, it was found that the level of reactive oxygen species increased in aged seeds of soybean under artificial ageing conditions and this was shown to reduce mitochondrial and ascorbate-glutathione cycle activity (Xin et al., 2014). In addition, there was an increase in protein oxidation during CDT which leads to alteration of the functional properties of seed proteins and enzymes as well as increasing their susceptibility to protein degradation (Rajjou et al., 2008). Besides proteins, free amino acids were found to be susceptible to oxidation by reactive oxygen species (ROS) (Stadtman and Levine, 2003). This ROS mediated chemical amino acid

177

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants oxidation reaction could impact mMDH double mutant seeds as a significantly high free amino acid pool was found in the maturing and desiccating seeds of this mutant. These readily available and abundant free amino acids can be immediately oxidized by ROS, and thus expedited the seed deterioration process. As mentioned in the earlier sections, amino acids can be alternative respiratory substrates for energy supply under low oxygen condition. Intriguingly chemically oxidized amino acids could trigger an autophagy mechanism that breaks down oxidized amino acids to support respiration in ageing seeds (Izumi et al., 2013). Both BCAAs and α-keto acids (by product of amino acid catabolism) had been shown to be cytotoxic compounds that can induce apoptosis, leading to cell death in mammals (Eden and Benvenisty, 1999). Therefore, such a high BCAAs content in mMDH double mutant seeds presumably increases cellular toxicity that could provoke early cell death, thus causing a shortened longevity. Lipid peroxidation is another form of oxidation reaction associated with seed ageing which disrupts the integrity of membrane phospholipids and peroxidation of storage lipids responding to oxidative stress under unfavourable environmental conditions (Stewart and Bewley, 1980; Wilson and McDonald, 1986; Al-Maskri, 2002). These further result in seed deterioration leading to a decline in germinability and loss of viability (Al-Maskri, 2002). It is therefore anticipated that the removal of mMDH genes might have increased the susceptibility of seed cells to the above-mentioned molecular events leading to a significant shortened longevity in mMDH double mutant ageing seeds. However, further investigation of the biochemical properties described above is required to uncover more of the underlying causes to account for the aberrant phenotypes of the aged seeds of this mMDH double mutant.

Arabidopsis root growth and respiration rate are determined by spatial expression of mMDH gene isoforms

The significantly low germination rate observed in mMDH double mutant seed led us to investigate and characterize the post-germinative growth of this mutant. The behaviour and performance of root growth in mMDH double mutant in parallel with

178

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants single mMDH mutants, complemented line and wild type were evaluated based on the total length and respiration rate of their primary root. With the absence of MMDH1 gene in Arabidopsis root, the root length was significantly reduced, contrary to the MMDH2 gene knockout line which did not differ from wild type, implying more impact is by the loss of MMDH1 gene. The root growth defect was more prominent in mMDH double mutant with approximately 60% reduction in root length compared to wild type. These results imply that mutating both MMDH1 and MMDH2 gene isoforms has a synergic effect and further impacted root growth. Slow root growth of seedlings derived from aged maize seeds could be related to decreased cell division and cell expansion in the growing zones of seedling roots (Bingham and Merritt, 1999). It is therefore plausible that the expedited seed ageing deterioration process of mmdh1-2mmdh2-1 seeds during storage is the primary reason for the observed reduction in root length and slow-growing root characteristics. The complementation with MMDH1 gene aided root growth but failed to fully restore root length to wild type level, as well as their respiration rate was significantly different to wild type. This clearly indicated a likely role of MMDH2 gene in normal root development of Arabidopsis. Assessment of root width of individual mMDH mutants revealed those roots were significantly thicker than wild type except for mMDH double mutant, notably the mmdh2-1 mutant exhibited the largest root width among the other mutants. The above findings collectively indicated a specific role of MMDH gene isoforms in regulating the root growth of Arabidopsis. This is in agreement with a differential gene expression pattern between MMDH1 and MMDH2 in the Arabidopsis thaliana primary root shown in the Arabidopsis eFP browser (http://bbc.botany.uroto.ca/efp/cgi-bin/efpweb.cgi) (Winter et al., 2007) (Figure 15) based on the report by Birnbaum and co-workers (Birnbaum et al., 2003). The developmental stages of Arabidopsis root were divided into three stages in which stage I-III represent the meristematic, transient and elongation zone of the root. Notably, the gene expression ratios of MMDH1/MMDH2 were in a range of 1.15-1.8 in the root apex regions while the ratios were 0.34-0.54 in the expanded regions, indicating that a higher MMDH1

179

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants expression at the root tip region while MMDH2 is expressed more strongly in the root elongation zone. These differential gene expression between mMDH gene isoforms exemplify a coordinated spatial expression, associating activity of genes with specific cell fate and tissue specialization in the Arabidopsis root (Birnbaum et al., 2003). This further suggested that the early root growth is largely contributed by the gene expression of MMDH1, while MMDH2 is progressively expressed at the successive root development stages with its role presumably involved in the expansion and elongation of the root. Notably, differential gene expression of mMDH in root parts correspond well with the root length data of mMDH mutants: the perturbed MMDH1 gene expression at the early stage of root development is reflected in a significant reduction in root length of mmdh1-2 mutant (indicated by 10-day old seedlings); whereas altered gene expression of MMDH2 became more noticeable with a progressively decline in the growth rate of root at later stages of development. This was represented by a relatively shorter root length than expected in the 16-day old seedlings of mmdh2-1 mutant.

180

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

Figure 15: Relative gene expression of mMDH isoforms in primary root regions of a 6-day old Arabidopsis thaliana ecotype Col-0. The image shown is adapted from Arabidopsis electronic fluorescent pictograph (eFP) browser. Log2 ratios of gene expression level of MMDH1 (At1g53240) to MMDH2 (At3g15020) in meristematic (stage I), transient region (stage II) and elongation region (stage III) are depicted in red-yellow-blue colour scheme. Note that the relative gene expression of MMDH1 over MMDH2 is gradually decreasing from root tip to maturing regions.

An overall higher respiratory activity in the root tip region compared to the root expanded region of all plant genotypes (wild type and mMDH mutants) implies that there is a higher energy demand for the root apex region containing actively dividing cells. The current findings are consistent with a study published earlier where root tips showed a marked increased oxygen consumption rate compared to the root expanded region of A. thaliana (ecotype Columbia) seedlings (Sew et al., 2013). One explanation for these tissue dependent differences could be the carbohydrate metabolism during root development. The root apex has an active carbon flow through glycolysis and the TCA cycle in order to meet the demand for ATP and carbon skeletons for cell division and many other biosynthesis processes (Horecker, 1962). Additionally, highly respiring root cells available particularly in the root cap had shown to be correlated with highly specialized patterns of metabolism (Street et al., 1976). In a similar study, A. belladonna (perennial herbaceous plant) roots revealed oxygen requirements for root tips were 3.3–11.5 times higher than for the remainder of the root (Williams and Doran, 1999). Recently, it was demonstrated that the cell cycle in the elongation zone is slower and longer than that in the apical meristem zone of A. thaliana root, and that the elongation zone is sequentially transformed into a differentiation zone as the root development progresses (Hayashi et al., 2013). Aside from that, based on positive correlation between oxygen uptake rates of P. persica root system and root meristem growth rate, Bidel and co-workers proposed that respiration rate and root growth rate is linearly correlated to each other (Bidel et al., 2000). All the above signify a greater respiratory quotient in root tips compared to root expanded regions.

181

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

Interestingly, it was demonstrated later that among the TCA cycle enzyme activities in the proximal meristem of maize root apical tissues, MDH exhibited the highest activity indicating a dominant role in root respiratory metabolism (Jiang et al., 2006). The profound role of mMDH These results conformed with a distinctly high gene expression of MMDH1 compared to MMDH2 in root tips region as discussed earlier (Birnbaum et al., 2003) in root development was evidenced from the root respiration data of mMDH mutants. For root tips, the oxygen uptake rate was significantly elevated when MMDH1 gene was absent and to a greater extent upon loss of both mMDH gene isoforms. Likewise, the root expanded regions of the mMDH double mutant respired at the highest rate. When all taken together, this points to a predominant role of MMDH1 in regulating mMDH activity in the root tip and possibly a relatively higher importance in the overall mMDH activity in both tip and expanded region of Arabidopsis root as compared to its gene counterpart, MMDH2.

The respiration rate of the whole root system in mMDH mutants revealed an approximately 50% increment in oxygen consumption rate in the mMDH double mutant while the complemented line respired at a rate of 40% more than wild type. Notably, complemented line demonstrated a gradual increase in the whole root respiration rate as root development progressed. This again could be plausibly linked to an increasing importance of MMDH2 gene in the successive root development as indicated by a higher MMDH2 gene expression in the root expanded region and matured region. Given that root respiration could be separated into growth and maintenance respiration (Lambers et al., 1983), and the participation of each is presumably based on the proportion of metabolically active meristematic and non-meristematic tissues in roots, corresponding to higher respiration rates in fine roots than in coarse roots (Pregitzer et al., 1998; Desrochers et al., 2002). All of the above collectively leads to a hypothesis that MMDH1 is predominantly involved in regulating the growth respiratory metabolism of rapidly dividing root meristem cells while MMDH2 is potentially important for maintenance respiration of differentiating and maturing root cells.

182

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

Importantly, the root length and respiration data presented in this study confirmed a role of mMDH genes in root development by consistently showing retarded root growth and significantly elevated root respiration rate as consequences of the gene perturbation of mMDH isoforms in Arabidopsis seedlings. It was evident that down-regulation of mMDH activity as much as 39% relative to wild type in antisense tomato plants had resulted significant reductions of the root area and dry mass as well as the respiratory rates (Van der Merwe et al., 2009). These findings consistent with the data presented here, clearly indicating an essential role of mMDH in growth and development of heterotrophic organ of Arabidopsis.

Conclusion

This study has shown mitochondrial malate dehydrogenase as an important respiratory element in the A. thaliana heterotrophic organs, seed and root. Perturbation of both mMDH genes in Arabidopsis seeds significantly elevated respiration rate at all stages of maturation, markedly reduced reserve accumulation and accelerated the ageing process in seed which collectively resulted in a decline in seed biomass, viability and germination potential. Arabidopsis primary root growth was retarded with dramatic reduction in length and exhibited a significant high respiration rate upon loss of both mMDH genes. A spatially regulated function of mMDH gene isoforms in root respiratory metabolism of Arabidopsis was hypothesized based on the coordinated changes between respiration rate, root length of mMDH single mutants and a pre-defined differentially expressed mMDH gene isoforms in root tissues (from Arabidopsis eFP browser). Overall, this study illustrates the potential of both mMDH genes as possible targets for improvement of Arabidopsis seed quality, shelf life and germination; enhancing the post- germinative growth which includes the establishment of seedlings, improved root growth as well as increased biomass for application in plant genetic engineering.

183

Chapter 4. Impact of mMDH loss on seeds and roots of mMDH mutants

184

Chapter 5:

Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

185

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Foreword to Study IV

In the previous chapters, I have discussed the knockout effects of mMDH genes via a reverse genetic approach in a commonly used laboratory Arabidopsis thaliana ecotype namely Columbia (Col). I presented evidence of both mMDH genes deletion leading to significant alteration of respiratory metabolism in the mutated Arabidopsis plants. The prominent observations included significantly higher leaf, seed and root respiration rates. In addition, the mMDH null mutation in Arabidopsis demonstrated late flowering phenotypes, delayed seeds germination and significant reduction in the root length. In this chapter, as an alternative to chemical-induced mutation, we exploited the natural occurring genetic variation among Arabidopsis thaliana accessions to characterise the mMDH genes from a wider perspective so as to gain a global insight into the regulatory networks of the Arabidopsis respiratory system. We hypothesised these mMDH genes have an important role in regulating the respiration rates among ecotypes underpinned by genetic diversity in their levels of expression.

Author contributions: Experiments were designed by myself, Millar, A.H. and Stroeher, E. I conducted the experiments including the growing of Arabidopsis thaliana ecotypes plants, respiration measurements using the Hansatech Oxytherm system, western blots analysis with anti-Maize and Arabidopsis porin antibodies, qPCR analysis on MDH isoforms transcript levels, preparation of protein samples from ecotypes using a high speed centrifugation method and extraction of metabolites from ecotypes samples. The protein abundances of TCA cycle enzymes protein were determined by Multiple Reaction Monitoring assays and were conducted on a triple quadrupole LC- MS with the technical assistance from Fenske R. and analysis carried out with guidance from Grassl, J. Metabolites levels from ecotypes samples were determined using GC-MS by outsourcing the measurements to Metabolomics Australia. I analysed the metabolite data with the help from

186

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Dorothee Hahne from Metabolomics Australia. I wrote this manuscript and it was revised by Stroeher, E. and Millar, A.H.

187

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Yun Shin Sew 1,2, Elke Ströher 1,2, Julia Grassl 1,2, Ricarda Fenske 1,2, A. Harvey Millar1,2* 1ARC Centre of Excellence in Plant Energy Biology and 2Centre for Comparative Analysis of Biomolecular Networks (CABiN), Bayliss Building M316, The University of Western Australia, 35 Stirling Highway, Crawley WA 6009, Western Australia, Australia.

*Corresponding author: A. Harvey Millar

Tel: +61 8 6488 7245 Fax: +61 8 6488 4401 e-mail: [email protected]

188

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Abstract

The natural genetic variation of Arabidopsis thaliana ecotypes is an invaluable resource for providing biological information encompassing functional, ecological and evolutionary aspects of metabolism. Herein, we exploited forty-nine Arabidopsis thaliana ecotypes for their natural variation of respiratory metabolism using multi-omics studies. We aimed to investigate the role of mitochondrial malate dehydrogenases (mMDHs) and other important components in the respiratory metabolism of ecotypes underpinned by their genetic variation. Systematic screenings of leaf dark respiration rates in 49 ecotypes led to the selection of a subset of 10 ecotypes with significantly low and high respiration rates which were categorised into two distinct groups. Quantitative PCR assay revealed there was natural variation of gene expression of MDH isoforms among ecotypes but there was no significant change of MDH gene expression between the two distinct groups of ecotype based on respiratory rate. Targeted proteomics using a multiple reaction monitoring method showed prominent changes of relative TCA cycle protein abundances among ecotypes. Interestingly, MMDH1 (At1g53240) was consistently found to be one of the respiratory components that showed a significant negative association with respiration rate both by Student’s t-Test between the two distinct ecotype groups and by multiple regression analysis with ecotype respiration rates. Compared to TCA cycle respiratory protein abundance data, ecotypes metabolites data showed a relatively weak correlation with ecotype respiratory data, however a plausible link to natural variation of water use efficiency among ecotypes had been demonstrated in this study. Overall, the role of MMDH1 (At1g53240) in governing the natural variation of ecotype respiratory metabolism has been highlighted and the significant impact of this enzyme in regulating the plant respiratory system in artificially generated genetic variation of mMDH genes in Arabidopsis ecotype Columbia has been validated and reemphasised as an ecologically relevant phenomenon.

189

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Introduction

Arabidopsis thaliana originates from the Cruciferae (family Brassicaceae) and it is well recognised as a self-pollinating plant species which can occupy a wide range of habitat types and is distributed all over the world. The ecotypes or accessions of Arabidopsis thaliana collected from the wild generally represent homozygous genotypes and their native range is considered to be continental Eurasia and North Africa (Al-Shehbaz and O'Kane, 2002). These Arabidopsis ecotypes exhibit genetic and phenotypic variations and have become one of the most significant model plant species for the study of genetic variation (Somerville and Koornneef, 2002). For instance the phenotypic differences across worldwide Arabidopsis ecotypes are observed in the diameter of rosette and leaf shape (Perez-Perez et al., 2002), size (Li et al., 1998), fresh or dry weight (Aarssen and Clauss, 1992; Loudet et al., 2003), plant height (Pigliucci and Schlichting, 1995) and flowering time (Sanda et al., 1997; Stratton, 1998; Gazzani et al., 2003). Studies have shown that Arabidopsis thaliana ecotypes have remarkable differences in their genetic pool and therefore exhibit a huge range of genetic divergence; for instance in starch biosynthesis (Shane et al., 2004), metabolism (Chevalier et al., 2004; Cross et al., 2006; Juszczak et al., 2012; Gordes et al., 2013; Sulpice et al., 2013), response to environmental factors (Li et al., 2006; Agrawal et al., 2012) and response to stresses (Katori et al., 2010; Barah et al., 2013; Chan et al., 2013). Such intraspecific natural variations are believed to result from spontaneous naturally occurring mutations and exist within-species by evolutionary processes (Alonso-Blanco et al., 2009). Arabidopsis thaliana accessions serves as the richest resources for functional, ecological, and evolutionary studies in a model plant (Koornneef et al. 2004).

To date there are at least 1049 different ecotypes have been registered in the Arabidopsis thaliana collection database (http://1001genomes.org/accessions. html). With the initiatives of genome sequencing and analysis of some of the ecotypes, they have increasing importance and research potential (Zeller, 2005; Cao

190

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

et al., 2011; Gan et al., 2011). A survey on the genetic diversity of 95 Arabidopsis ecotypes revealed approximately 21% of the single nucleotide polymorphisms (SNPs) discovered were observed in a single ecotype and 45% of the SNPs found in the exon regions could lead to an amino acid change (McKhann et al., 2004). Evidence has shown that the amino acid polymorphisms of light receptor phytochrome B among Arabidopsis ecotypes cause differential responses to light (Filiault et al., 2008). Single amino acid substitution in adenosine 5’-phosphosulfate reductase of sulfate reduction pathway between Arabidopsis ecotypes Bay-0 and Shahdara has resulted a reduction of its enzyme activity leading to sulfate accumulation in Shahdara (Loudet et al., 2007). Exploring the natural variation of previously characterised genes as alternative source of new alleles is complementary to reverse genetics studies and could help us to gain insights into the regulation of vital plant biological processes underpinning their genetic variation.

There is a general assumption that genetic diversity across ecotypes is highly associated with their natural habitats and ecotype traits are presumably ecological important for different environmental adaptations (Alonso-Blanco and Koornneef, 2000). Previous findings have shown environmental changes are responsible for the variation of eco-physiological traits in plant species (Chapin and Oechel, 1983; Saldana et al., 2007; Gianoli et al., 2012; Feng and Dietze, 2013). For instance, respiration rates of highland and lowland grown Japanese knotweed ecotypes were found to be significantly different from each other implying ecotypic differences in the performance of maintenance respiration in varying temperate climates (Mariko and Koizumi, 1993). This might be a good indication for Arabidopsis thaliana ecotypes to exhibit natural variations of respiration rates in relation to their origins with varying altitudinal and/or latitudinal gradients.

In plants, malate dehydrogenases exist in multiple isoforms and isoenzymes are localised in different subcellular compartments. Malate dehydrogenases take part

191

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

in several important plant metabolic pathways including the tricarboxylic acid cycle, glyoxylate bypass, amino acid synthesis, gluconeogenesis and exchange of metabolic compounds across different subcellular organelles (Musrati et al., 1998). There are two isoforms of mitochondrial malate dehydrogenase (mMDH) which catalyse a reversible reaction of malate to oxaloacetate (OAA) coupled to the reduction of the NAD pool in the tricarboxylic acid (TCA) cycle of Arabidopsis thaliana. The importance of mitochondrial malate dehydrogenase as one of the key respiratory enzymes in plants has been reported (Journet et al., 1981; Nunes-Nesi et al., 2005; Van der Merwe et al., 2009; Tomaz et al., 2010; Araujo et al., 2012). A survey using the publicly available Arabidopsis tool known as the Arabidopsis eFP Browser (http://bar.utoronto.ca/efp_Arabidopsis/cgi-bin/efpWeb.cgi) revealed differential gene expression of mMDH isoforms across 34 A. thaliana ecotypes. In addition, there were general distinct proteome profiles inclusive of a distinguishable MDH protein expression across eight Arabidopsis thaliana ecotypes tested (Chevalier et al., 2004). These findings not only imply the presence of ecotype-specific protein expression but also that MDH could be a key contributor in shaping the genetic variation of respiratory metabolism among Arabidopsis ecotypes.

Dark respiration analysis is one of the common approaches used in plant physiology studies. In Arabidopsis research, laboratory-preferred ecotypes such as Columbia (Col-0), Landsberg erecta (Ler) and Cape Verde Islands (Cvi-0) are the choices for most of the studies. However, none of the studies has investigated the natural variation of Arabidopsis ecotypes from the perspective of their dark respiration rate measurements and comparative omics studies. Hence, in this present study we explored the distinct respiration rates across a total of 49 Arabidopsis thaliana ecotypes and combined the powerful multi-omics approaches encompassing transcript analysis, proteomics and metabolomics to provide a view of the

192

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

underlying molecular factors influencing the natural genetic variation of respiratory capacity and metabolism amongst Arabidopsis ecotypes.

Materials and methods

Arabidopsis thaliana ecotypes used as plant material

A nested core collection of 48 Arabidopsis thaliana ecotypes as previously reported in McKhann et al. (2004) and an additional reference accession, ecotype Columbia (Col-0) were selected for this study. This nested core collection of ecotypes represents the majority of the genetic diversity of Arabidopsis thaliana from an original collection of 265 germplasms. The seeds of those Arabidopsis ecotypes were obtained from Versailles Arabidopsis Stock Centre of the National Institute for Agricultural Research (INRA) and their identities are described in Table 1. For cultivation of Arabidopsis thaliana ecotypes, the seeds were sowed on moist soil mix containing compost, perlite and vermiculite in the ratio of 3:1:1 in trays and cold treated in the dark for 3 days (stratification process) in order to synchronise seed germination. Then the plants were grown in short day lighting conditions (8 hr light/ 16 hr dark) with light intensity of 120- 140 µmolm-2s-1 and a temperature cycle of 22°C day/17°C night.

193

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Table 1. List of the 49 ecotypes used for this study and their origins.

Ecotype Stock centre number in Name number or Country Versailles number Latitude this study reference 1 Akita 1 Japan 252 AV N 39°43' 2 Alc-0 N1656 Spain 178 AV N 40°29' 3 Bl-1 N968 Italy 42 AV N 44°29' 4 Bla-1 N970 Spain 76 AV N 41°41' 5 Blh-1 N1030 Czech Republic 180 AV N 48°49' 6 Bur-0 N1028 Eire 172 AV N 53°07' 7 Can-0 N1064 Canary Islands 163 AV N 28°00' 8 Ct-1 N1094 Italy 162 AV N 37°3' 9 Cvi-0 N902 Cape Verde Islands 166 AV N 16°00' 10 Edi-0 N1122 United Kingdom 83 AV N 50°57' 11 Enkheim-T N921 Germany 197 AV N 50°08' 12 Ge-0 N1186 Switzerland 101 AV N 46°12' 13 Gre-0 N1210 USA 200 AV N 43°11' 14 Ishikawa 1 Japan 253 AV N 36°38' 15 Ita-0 N1244 Morocco 157 AV N 34°04' 16 JEA 2.3 France 25 AV N 43°41' 17 Jm-0 N1258 Czech Republic 206 AV N 49°04' 18 Kn-0 N1286 Lithuania 70 AV N 54°54' 19 Kondara N916 Tadjikistan 190 AV N 38°48 20 Lip-0 N1336 Poland 63 AV N 50°09' 21 Mh-0 N904 Poland `175 AV N 53°31' 22 Ms-0 N905 Russia 93 AV N 55°45' 23 Mt-0 N1380 Libya 94 AV N 32°34' 24 N13 N22491 Russia 266 AV N 62°07' 25 N14 N22492 Russia 267 AV N 62.2 26 N6 N22484 Russia 262 AV N 62°00' 27 N7 N22485 Russia 263 AV N 61°51' 28 Nok-1 N1400 Netherlands 95 AV N 52°14' 29 Oy-0 N1436 Norway 224 AV N 60°23'

194

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

30 Pa-1 N1438 Italy 50 AV N 38°07' 31 Pi-0 N1454 Austria 40 AV N 47°04' 32 Pyl-1 2.3 France 8 AV N 44°39' 33 Ran 2.3 France 21 AV N 48°3' 34 Ri-0 N1492 Canada 160 AV N 49°1' 35 Rld-2 N1641 Russia 229 AV N 56°15' 36 Rubezhnoe-1 N927 Ukraine 231 AV N 48°13' 37 Sah-0 N1500 Spain 233 AV N 38°52' 38 Sakata 1 Japan 257 AV N 33°55' 39 Sap-0 N1506 Czech Republic 234 AV N 49°49' 40 Sav-0 N1514 Czech Republic 235 AV N 49°49' 41 Shakdara N929 Tadjikistan 236 AV N 37°29' 42 Sp-0 N1530 Germany 53 AV N 52°30' 43 St-0 N1534 Sweden 62 AV N 59°19' 44 Stw-0 N1538 Russia 92 AV N 52°57' 45 Ta-0 N1548 Czech Republic 56 AV N 49°25' 46 Te-0 N1550 Finland 68 AV N 60°04' 47 Tsu-0 N1564 Japan 91 AV N 34°19' 48 Yo-0 N1622 USA 250 AV N 37°45' 49 Col N1092 Poland 186 AV N 52°44'

Dark respiration measurements of Arabidopsis ecotype leaf samples

Approximately 7 mm diameter leaf discs from mature leaves of 6-week old plant were prepared using a cork borer and a total of 30-50 mg fresh weight of leaf discs were pooled (varying fresh weight depending on the leaf size of the ecotype) and leaf respiration buffer (10 mM HEPES, 10 mM MES, and 2 mM CaCl2, pH 7.2) was added prior to dark incubation for 30 min. Leaf oxygen consumption rate (OCR) measurements were performed using a liquid-phase Clark-type Oxygraph system (Hansatech Instruments) in a 2 mL volume for at least 15 min at 25°C in a darkened electrode chamber. The OCR was recorded using the Oxygraph Plus v1.02 software

195

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

(Hansatech Instruments) and adjusted to fresh weight to obtain OCR per gram fresh

-1 -1 weight (nmolO2 min g FW ) of leaf tissue.

Leaf total protein extraction

Approximately 100 mg fresh weight of five-week-old leaf tissues from Arabidopsis ecotypes was snap frozen and ground in liquid nitrogen using 5 mm steel beads (Qiagen, Australia). The ground tissue was homogenised in 200 µL of protein extraction buffer [1x Phosphate Buffered Saline (PBS), 1 mM EDTA and 1x protease inhibitor cocktail (Roche), pH 7.5] and then kept on ice while mixed vigorously for 5 min. The homogenate was then centrifuged at 2000 g for 5 min at 4°C to remove large cell debris. The soluble proteins were obtained from the supernatant. Subsequently, the concentration of the extracted total soluble proteins was quantitated using the Amido black protein assay (DieckmannSchuppert and Schnittler, 1997). BSA was used as a protein standard.

Western blotting and immunodetection of Arabidopsis leaf porin abundance

The protein samples were added to 5X sample loading buffer (250 mM Tris-HCl pH 6.8, 10% [w/v] SDS, 50% [v/v] glycerol, 0.01% [w/v] bromophenol blue and 100 mM DTT) prior to heating at 65°C for 10 min. The denatured protein samples were loaded with an equal volume in each lane of Bio-Rad Criterion precast gels (10%– 20% [w/v] acrylamide, Tris-HCl, 1 mm thick, 18-well comb) and gel electrophoresis was performed at 180 V per gel for 1 hour. Polyacrylamide gels, filter papers and Hybond-C Extra nitrocellulose membrane (GE Healthcare) were soaked in transfer solutions (40 mM glycine, 50 mM Tris, 0.04% [w/v] SDS, 20% [v/v] methanol) for 10 min. The protein were transferred from the gel onto a Hybond-C Extra nitrocellulose membrane (GE Healthcare) using a Hoefer SemiPhor (GE Healthcare) instrument according to the manufacturer’s instructions. The transferred proteins were visualised on the membrane by staining with Ponceau S solution (0.1% [w/v]

196

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Ponceau S and 5% [w/v] acetic acid) and then destained with 1X TBS-Tween (TBS-T; 0.15 M NaCl, 10 mM Tris-HCl, 0.1% [v/v] Tween-20, pH 7.4). The membrane was then blocked with 1% [v/v] blocking solution (Roche Diagnostics) in 1X TBS solution for 1 hour at room temperature with gently shaking. The membrane was probed with two different primary antibodies; anti-Maize porin (1:5000) and anti- Arabidopsis thaliana VDAC1 (diluted to 1:5000) (AS 07212, Agrisera, Vännäs, Sweden) overnight at 4°C. After washing the membrane 4 times with 1X TBST solution for 5 min each wash, the secondary anti-mouse or anti-rabbit antibodies (diluted to 1:20000) was applied. The immunodetection was performed using the ECL Plus Western Blotting Detection Kit (Amersham, Braunschweig, Germany) based on chemiluminescent signals derived from the horseradish peroxidase which is coupled to the secondary antibodies. Quantitative light emission was recorded using ImageQuant-RT ECL (Amersham Biosciences). Analysis of porin/ VDAC signal intensities were performed using ImageJ software (Schneider et al., 2012).

Isolation of total RNA

Total RNA was prepared from 5-week old Arabidopsis ecotype plants grown in short-day conditions. A 5 mm steel bead was added to <100 mg of frozen leaf tissue in a 2 mL microcentrifuge tube and ground to homogeneity using a Retsch Mixer Mill (Type MM 300, Qiagen, Hilden, Germany) for 2x 60 seconds with frequency set at 18 Hz in the presence of liquid nitrogen. Total RNA extraction was then performed using the RNeasy Plant Mini Kit (Qiagen, Hilden, Germany) following the manufacturer’s instruction. The total RNA was eluted in a 40 µL volume using RNAase-free water (pre-warmed at 50°C).

Purification of total RNA

The total RNA was then treated with Turbo™ DNAase (Ambion, Victoria, Australia) to remove contaminating genomic DNA according to the manufacturer’s

197

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

instructions for standard digestion, except that samples were incubated at 37°C for 45 min. The treated total RNA was then precipitated with 0.1 vol of 3M sodium acetate (3M NaOAc, pH 5.2) and 2.5 vol of chilled absolute ethanol, overnight at -20°C instead of using a DNAse inactivation reagent as suggested in the manufacturer’s instruction. The total RNA was pelleted by centrifugation at 15,000 g for 30 min at 4°C followed by a washing step with 500 µL of chilled 70% [v/v] ethanol. After air drying for about 10 min at room temperature the total RNA was resuspended in 20 µL of RNAse-free water. The integrity of purify total RNA was checked by running a 1% formaldehyde agarose denaturing gel electrophoresis and a Nanodrop ND-1000 spectrophotometer (Thermo Scientifics, Wilmington, USA) was used to measure its concentration and check for impurities.

Quantitative real-time (qPCR) assay and analysis

One microgram of total RNA after RNA quantitation was used for first strand cDNA synthesis with a priming reaction to oligo(dT) (Invitrogen, Carlsbad, CA) according to manufacturer’s instruction. For qPCR assay preparation, a 5 µL final reaction mixture volume (0.5 µM primer pair, 2.5 µL 1x LightCycler® 480 SYBR Green I Master [Roche]; 0.5 µL of diluted cDNA template [diluted to 1:20] were loaded into a 384-well PCR plate. The qPCR assay was performed in a LightCycler® 480 instrument II (Roche) in which the cycles were programmed with the following steps: denaturation (10 min, 95 °C), 40 amplification cycles (95 °C for 10 sec; 60 °C for 10 sec; 72 °C for 10 sec), melting curve analysis (95 °C for 10 sec; 65°C for 60 sec) and followed by a transition rate of +0.1 °C per cycle and continuous data acquisition. The qPCR data were then analysed using the LightCycler® data analysis software (Roche). Gene-specific primers were used for qPCR assay which were designed for genes encoding Arabidopsis housekeeping genes; Clathrin (Clathrin_At5g46630), alkaline pyrophosphatase (PPase_At1g13320) and yellow- leaf-specific protein 8 (YLS8_At5g08290) as well as MDH isoforms; consisting of

198

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

mitochondrial MDHs (MMDH1_At1g53240 and MMDH2_At3g15020), cytosol MDHs (CMDH1_At1g04410, CMDH2_At5g43330, CMDH3_At5g56720), chloroplastic MDH (CHMDH_At3g47520) and peroxisomal MDHs (PMDH1_At2g22780 and PMDH2_ At5g09660) using QuantPrime online software (http://www.quantprime.de/) (Arvidsson et al., 2008). Selection of gene specific primers for both housekeeping genes and MDH isoforms was based on melting curve analysis and the presence of a single PCR amplicon using cDNA of ecotype Col. Those selected gene-specific primers sequences (5’-3’) were clathrin-F (TCGATTGCTTGGTTTGGAAGAT), clathrin-R (GCACTTAGCGTGGACTCTGTTTG), PPase-F (TAACGTGGCCAAAATGATGC), PPase-R (GTTCTCCACAACCGCTTGGT), YLS8-F (GGGATGAGACCTGTATGCAGATGGA), YLS8-R (GCTCGTACATGGTGTTGAAGTCTGG), MMDH1-F (AATGTTCCGGTGATTGGTGGTC), MMDH1-R (TTGGCTTGAGGAGTTGCCTGAG), MMDH2-F (GCCAAGTATTGCCCACAAGC AC), MMDH2-R (TCAGCTGCAATTGGAACAGTGGAG), CMDH1-F (TGCTTGTGACCAC ATCCGTGAC), CMDH1-R (CCATGGAAACGAACGTACCCTCTG), CMDH2-F (GCTGCACC AAACTGCAAGGTTC), CMDH2-R (TGTTGTGGTCAAGCCTGGTCAAG), CMDH3-F (ACC GGTGCAGCAGGAAACATAG) and CMDH3-R (TCATGGGTTGATCTGGACCTAGC), CHMDH-F (GCTCACTGTTAGGATTCAGAACGC), CHMDH-R (AACCTGCACCTGCCTTAG CATC), PMDH1-F (TGACTGAGCTTCCCTTCTTCGC), PMDH1-R (TGCCTTCTCTAATC CCATCCTCTC), PMDH2-F (TTCGTGGAGATGCCAACCAGAG) and PMDH2-R (ACAGAG TTCTTGGCCTCCATCTG). Mean transcript level of each targeted gene for each biological replicate was calculated from 3 technical replicates. The mean value of each MDH isoform transcript level was then normalised using the mean transcript level of the housekeeping genes within the same biological replicate. Finally a mean transcript level of each MDH isoform was obtained from 3 independent biological replicates. For comparison of each MDH isoform across different ecotypes, the relative transcript level of the MDH isoform in each ecotype was calculated by dividing their transcript abundance value with the average of the corresponding MDH transcript level across all the ecotypes.

199

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Multiple reaction monitoring (MRM) sample preparation and analysis

For protein extraction of Arabidopsis ecotypes samples, tissue of six-week-old plants were harvested and homogenised in the cold room using a pre-chilled mortar and pestle in the presence of chilled grinding buffer (0.3 M sucrose, 25 mM

Na4P2O7, 10 mM K4P2O7, 2 mM EDTA, 1% [w/v] PVP-40, 1% [w/v] BSA, 20 mM sodium-ascorbic acid, 20 mM L-cysteine, pH 7.5) according to a ratio of 1 g FW: 5 mL buffer. The homogenate was filtered through 4 layers of gauze and 1 layer of fine pre-wet Miracloth (Calbiochem) and centrifuged at 2500 g for 5 min at 4°C. Subsequently the supernatant was centrifuged at high speed 17400 g for 20 min at 4°C. The resulting protein pellet was then resuspended with 1 mL protein buffer (0.3 M sucrose, 10 mM TES, pH7.5). The Bradford assay was used to quantitate the protein concentration and spectrophotometric measurements were performed at a wavelength of 595 nm using BSA as a protein standard. For MRM sample preparation, 200 µg soluble proteins were precipitated with 5X volume of chilled absolute concentrated acetone for overnight at -20°C. On the next day the precipitated proteins were pelleted at 20000 g for 20 min at 4°C. This was followed by 2X washing steps using chilled absolute acetone before the precipitated soluble protein was resuspended in 200 µL of buffer (8 M urea, 50 mM NH4HCO3, and 5 mM DTT) and incubated at 37°C for an hour with gentle shaking at 300 rpm. The sample was then treated with 10 mM iodoacetamide (IAA) for 30 min at room temperature in the dark prior to adding 1 M urea with 50 mM NH4HCO3 to the sample (diluted to 1:6). Each protein mixture was then digested by adding 10 µg trypsin (trypsin powder dissolved in 0.01% [v/v] trifluoroacetic acid to a concentration of 1 mg mL-1) overnight at 37°C. The samples were then acidified to 1% [v/v] with formic acid. Before loading each sample into solid phase extraction Silica C18 Macrospin column (The Nest Group, Massachusetts, USA), each column was equilibrated with 750 µL of 70% [v/v] acetonitrile and 0.1% [v/v] formic acid and charged with 750 µL of 5% [v/v] acetonitrile and 0.1% [v/v] formic acid. The sample was loaded onto the

200

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

column and centrifuged for 3 min at 150 g at room temperature and followed by two washes with 750 µL of 5% [v/v] acetonitrile and 0.1% [v/v] formic acid. The purified protein was eluted twice with 750 µL of 70% [v/v] acetonitrile and 0.1% [v/v] formic acid. The eluate was dried in vacuum centrifuge for 4 hours at room temperature before it was resuspended in 5% [v/v] acetonitrile and 0.01% [v/v] formic acid to a final concentration of 1 µg µL-1. For the multiple reaction monitoring (MRM) assay, 1 µL of each sample (final concentration of 1 µg µL-1) was injected into an Agilent 6430 QqQ mass spectrometer with an HPLC Chip Cube source (Agilent Technologies). For each sample three injections were separately run for technical replication. Detailed methods of the MRM run, optimization of collision energy (CE) and selection of candidate MRM transitions were as described in (Taylor et al., 2014). For each unique peptide, three transitions, one quantifier, and two qualifiers were chosen to validate the quantifier. Three peptides per protein were optimised. The peptides were designed in such a way that they do not contain any missed cleavage and cysteine residue which could potentially susceptible to carbamidomethylation and oxidation (Liebler and Zimmerman, 2013). Integration of MRM data was performed using MassHunter Workstation software (Quantitative Analysis version B.06.00 for QQQ, Agilent Technologies). The signal response of a targeted peptide was subjected to normalization (targeted signal over the total signal responses of all peptides measured within an ecotype sample). A mean value of each peptide was then calculated from 3 technical replicates prior to averaging each protein from their corresponding peptides. Principal component analysis (PCA) in the R conductor program was then used to detect any outlier from average protein values of each biological replicate which could skew the MRM dataset. In order to compare relative protein abundance across ecotypes, each protein value was divided by mean of the corresponding protein value across ecotypes. These values were subsequently subjected to multiple regression analysis using the SPSS program. For mapping of the data onto the TCA cycle pathway, a mean value of each protein was calculated from 3

201

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

independent biological replicates and then relative protein abundance values were loaded into MultiExperiment Viewer (MeV, version 4.9) (Saeed et al., 2003) for heat map generation and protein distance matrices analyses.

Metabolites extraction for preparation of GC-MS samples and data analysis

Approximately 30 mg of frozen leaf tissue of Arabidopsis ecotypes were ground in a 2 mL microcentrifuge tube with steel beads using the Retsch Mixer Mill (Type MM 300, Qiagen, Hilden, Germany). After adding 500 µL of cold extraction buffer (0.5 mL; 85% [v/v] HPLC-grade methanol, 15% [v/v] untreated MilliQ water, and 100 ng µL-1 ribitol), the tube was vigorously vortexed, and then continuously shaken at 1,400 rpm for 20 min at 65°C in a Eppendorf thermo block. The cell debris was spun down by centrifuging at 3,000 g for 3 min. A volume of 60 µL extracted metabolites solution was transferred into glass insert (Agilent, CA, USA) and dried in a vacuum centrifuge for approximately 3 hours. Subsequently twenty microliters of 20 mg µL-1 methoxylamine-HCl (98% purity; Sigma) was added to each dried samples. Samples were then shaken at 1400 rpm for 90 min at 30°C. To each sample, 30 µL of N- methyl-N-(trimethylsilyl)-trifluoroacetamide (derivatization grade; Sigma) was added, followed by shaking again at 1400 rpm for 30 min at 37°C. After this, 10 µL of n-alkane retention index markers (0.029% [v/v] n-dodecane, 0.029% [v/v] n- pentadecane, 0.029% [w/v] n-nonadecane, 0.029% [w/v] n-docosane, 0.029% [w/v] n-octacosane, 0.029% [w/v] n-dotriacontane, and 0.029% [w/v] n-hexatriacontane dissolved in anhydrous pyridine) was added and vortexed. The reaction mixture was incubated for 30 min at room temperature prior to injecting 1 µL of the reaction mixture into GC-MS instrument (Agilent, CA, USA). GC-MS data were collected and analysed using Quantitative Analysis (MS) Software (Agilent, CA, USA). Relative abundance of each metabolite across ecotypes was calculated by dividing the ribitol normalised metabolite value with the average value across 10 ecotypes of the corresponding metabolites. These normalised relative metabolites abundance

202

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

values were then loaded into SPSS statistical software for Pearson correlation and multiple regression analysis. For mapping to metabolic pathways, a mean value of relative metabolite abundance was calculated by averaging all the independent biological replicates prior to loading into MultiExperiment Viewer (MeV, version 4.9) software for heat map and hierarchical clustering analysis.

Statistical analysis Correlation analysis was performed using the Pearson correlation coefficients method. One-way ANOVA was carried out followed by Tukey’s honestly significant differences (HSD) multiple comparison test for significant difference from the mean values by the least significant difference (LSD) test at a 0.05 and 0.01 probability level. Both correlation and ANOVA analysis were performed using IBM SPPS Statistics 19. Heat maps of the relative abundance of transcripts, proteins and metabolites were generated using MultiExperiment Viewer (MeV, version 4.9). Protein distance matrices and hierarchical clustering analysis were performed using the Pearson correlation method in the MeV software. Multiple linear regression with backward elimination selection with a cut-off value at P>0.05 was performed using IBM SPPS Statistics 19 to reduce the data and to predict the simplest set of relationships between protein abundance of TCA cycle enzymes and respiration rates across the final 10 Arabidopsis thaliana ecotypes.

203

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Results

Leaf dark respiration analysis of Arabidopsis ecotypes

The first objective of the present study was to examine the respiration rate across 49 Arabidopsis thaliana ecotypes; including a commonly used laboratory accession Columbia (Col) for comparison. The dark respiration measurements of Arabidopsis ecotypes were performed in 4 sequential screenings. The first screening was carried out on the full set of 49 ecotypes with observed OCR ranging from 68 to 145

-1 -1 nmolO2 min gFW (Figure 1). Statistical analysis by Student’s t-Test (P<0.1 and P<0.05) was used to identify ecotypes that were significant different from the ecotype 49 (Col) which was the reference ecotype in our study. In addition, ecotypes with significant low and high OCR were determined using Student’s t-Test (P<0.05) against ecotype 47, which was positioned in the middle of the sorted OCR ranking. Combining the results of Student’s t-Test analysis, a subset of 16 ecotypes (9 and 7 ecotypes with significant high and low respiration rate respectively) was selected for a second screening of respiration rate. A similar approach was used for successive respiration analysis where another subset of 10 ecotypes was selected for the third level of screening. In order to increase the confidence level of statistical analysis of respiratory data, the number of independent biological replicate for each ecotype was increased from 8 to 15 plants in the third set of measurements. The same set of 10 ecotypes was used in a fourth set of measurements to validate the previous respiration data where 12 plants per ecotype were used. Images of the selected final 10 Arabidopsis ecotypes are shown in Figure 2. The respiration data from all four sets of respiration measurements was combined for a more comprehensive analysis of rate differences. Mean OCR of the

-1 -1 final 10 Arabidopsis ecotypes ranged from 73 to 108 nmolO2 min gFW (Figure 3). Ecotype Col was found in a middle position within the respiration data range with

-1 -1 approximately 84 nmolO2 min gFW . Among the 10 ecotypes, there were 6 ecotypes with an OCR significantly different from ecotype Col after Student’s t-Test

204

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

analysis (P<0.05). Additional Student’s t-Test analyses were performed against ecotype 6 and ecotype 30. With all the Student’s t-Test results, it was found that these 10 ecotypes could be clustered into 2 distinct groups based on the significant differences of their respiration rates to one another. Similar results were obtained by a one-way ANOVA with Tukey post-hoc multiple comparison followed by homogenous subset analysis using the SPSS statistical package where there were 2 significant different groups of ecotypes regarding their respiration rates. We thus named the two distinct groups as bottom (ecotype 38, 39, 6, 2 and 49) and top performers (ecotype 44, 46, 30, 25 and 20) with significant low and high respiration rates respectively.

Survey of porin/ VDAC level in Arabidopsis ecotypes

As the primary objective of this study was to define the control of respiration rates of Arabidopsis ecotypes, investigation of mitochondrial mass was crucial to exclude variation of this as the underlying reason for natural variation in respiration rate. It has been shown that mitochondrial mass and porin level are well correlated from a number of previous studies (Shane et al., 2004; Noguchi et al., 2005). A total of 16 selected Arabidopsis ecotypes (candidates for the 2nd respiration screening) were subjected to Western blot analysis of their porin/ VDAC protein level. Proteins from the selected 16 ecotypes leaf tissues with 3 independent biological replicates each were extracted using the same ratio of leaf tissue fresh weight to extraction buffer volume. After protein quantification, an equal volume of 3.5 µL of protein extract of each ecotype was loaded onto one dimensional SDS-PAGE gel. This equal volume of protein loading approach (same ratio leaf fresh weight to buffer volume) was used as it is suitable for comparison with normalised respiration rate with fresh weight of leaf tissues of ecotype.

205

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

screening screening

day conditions. Values -

) with significant difference to ecotype 47 at P<0.05 ecotypes grown under short

Arabidopsis

week week old 49 -

Test. The circled ecotypes with low OCR (blue) and high OCR (red) were selected for the 2 -

tudent’s t an OCR (n=4; mean ±SE) sorted by increasing OCR value order. The (*) and (**) indicate ecotypes with significant Leaf dark respiration analysis of 5

Figure 1 represent me difference to ecotype 49 (Col) at P<0.1 and P<0.05 respectively whilst ( identified by S test. 206

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

E cotype 2 Ecotype 6 Ecotype 20 Ecotype 25 Ecotype 30

E cotype 38 Ecotype 39 Ecotype 44 Ecotype 46 Ecotype 49

Figure 2. The phenotypic appearance of the selected final 10 Arabidopsis thaliana ecotype rosettes of 4-week old plants. Note that the identities of these ecotypes are presented in numbers. Detailed descriptions of each ecotype are available in Table 1.

Figure 3. Combined analysis of leaf dark respiration data of the selected final 10 Arabidopsis ecotypes. Respiration rate is presented in mean OCR ±SE, n=27-39. Student’s t-Test gave significant differences to ecotype 49 (Col) (P<0.05), ecotypes 6 (P<0.05) and ecotype 30 (P<0.05) which were indicated with black, red and blue asterisks respectively. The Student’s t-Test results were validated using one-way ANOVA with a multiple comparison test in the SPSS statistical software. Thus the ecotypes were distinguished in two main clusters, bottom performers with significant low respiration rates (boxed in blue) and top performers with significant high respiration rates (boxed in red).

207

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

For Western blot analysis, a monoclonal anti-Maize porin antibody and polyclonal Arabidopsis thaliana voltage-dependent anion channel (VDAC1) antibodies (Agrisera, Vännäs, Sweden) were used. Protein sequence analysis showed that anti-Maize porin (Uniprot ID: Q9SPD7) and Arabidopsis VDAC1 (Uniprot ID: Q9SRH5) share 74% protein similarity and 57% protein identity. To represent the porin/VDAC protein level in each ecotype the mean of the porin signal intensities from the western blots of three independent biological replicates was calculated using the ImageJ software (National Institutes of Health, Bethesda, MD). Western blots analysis showed that porin signals of Arabidopsis ecotypes were observed at approximately 29 kDa with both anti-Maize porin and anti-AtVDAC1 antibodies as shown in Figure 4A and 4B respectively. Comparison of the mean percent of total porin signal detected by both antibodies displayed a similar trend with good correlation, R²=0.580 (P<0.05) (Figure 5). This implies porin in Arabidopsis ecotypes was detected by both antibodies and gave similar results. Porin content varied between the ecotypes. Based on the assumption that the porin content linked closely with mitochondrial mass, the findings here showed that there is natural variation of mitochondrial mass. However there was no specific trend of percent porin observed with their ascending respiration rates (Figure 6) and low correlations were found between respiration rates and porin level with R²=0.312 (P<0.05) and R²=0.304 (P<0.05) for anti-Maize porin and anti-AtVDAC1 antibodies respectively (Figure 7A and 7B). The natural variation of respiration rate observed across Arabidopsis ecotypes was not well associated with the varying amount of mitochondrial mass. Differences seen in respiration rates of Arabidopsis ecotypes are possibly due to the genetic variation of primary metabolic composition rather than their mitochondrial number or mass.

208

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Mito 2 6 9 14 20 22 25 30 32 33 38 39 44 46 48 49 maize porin,

29kDa A rubisco large subunit, 53kDa Mito 2 6 9 14 20 22 25 30 32 33 38 39 44 46 48 49 AtVDAC1,

B 29kDa

rubisco large subunit, 53kDa Figure 4. Western blot analysis of porin level in 16 selected Arabidopsis ecotypes. The porin signals were detected by anti-Maize porin (A) and anti-AtVDAC1 antibodies (B) with a protein molecular weight of approximately 29kDa. The protein extracts were loaded in equal volume (same ratio fresh weight leaf tissue to buffer volume) in both analyses. Isolated mitochondrial protein (Mito) was loaded onto SDS-PAGE gels as a control. Rubisco large subunit from ecotype leaf extract was observed at approximately 53kDa on nitrocellulose membrane stained with amido black.

Figure 5. A scatter plot of the mean percent porin detected by anti-Maize porin and anti-AtVDAC1 antibodies. Line indicates the linear relationship between the antibodies and a trendline was predicted with Pearson correlation value of R2=0.5801 and significant value P<0.05 which was validated using the SPSS statistical software.

209

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Figure 6. Comparison of Arabidopsis ecotype porin level detected by anti-Maize porin and anti-AtVDAC1 antibodies. Values represent calculated mean percent porin content of Arabidopsis ecotypes (mean ±SE, n=3) from Western blot analysis using anti-Maize porin (shaded in light grey) and anti-AtVDAC1 (shaded in dark grey) antibodies. The ecotype positions were sorted in ascending order according to their mean oxygen consumption rate order calculated from the combined respiration analysis of ecotype.

A A B B

Figure 7. Relationship between respiration rate and porin level in 16 selected Arabidopsis ecotypes. Pearson correlation coefficients of R2=0.3121 and R2=0.3038 between respiration rate and porin level detected by anti-Maize porin (A) and anti- AtVDAC1 (B) antibodies (P<0.05 as validated in SPSS statistical software).

210

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Quantitative PCR analysis of MDH isoforms

Evidence has shown that the loss of both mMDH isoforms resulted in a significant high leaf respiration rate in Arabidopsis plant (Tomaz et al., 2010). In this study, we aimed to further investigate the role of mMDHs (MMDH1 and MMDH2) in regulating the here established natural occurring variation of leaf respiration rate in Arabidopsis thaliana ecotypes. Thus quantitative PCR assays were performed to obtain the mMDHs gene expression profiles as well as of other MDH isoforms across the final 10 Arabidopsis ecotypes using mature leaf samples from those ecotypes. It was hypothesised that the fast-respiring ecotypes could have differential gene expression of mMDHs or other MDH isoforms compared to slow respiring ecotypes. Firstly, a survey of the gene expression of MDH isoforms from Arabidopsis thaliana ecotype Columbia microarray data (ATH1: 22K array) was carried out using Genevestigator software (https://www.genevestigator.com/) (Hruz et al., 2008). The hierarchical clustering of MDH isoforms based on the percent of their expression potential throughout various developmental stages is shown in Figure 8. The gene expression of MDH isoforms in developed rosettes was of our interest as it best matched with the leaf age of ecotypes used for the current qPCR assay. Note that the survey results demonstrated that MMDH1, CMDH1, CHMDH and PMDH2 give relatively higher expression (>50% of expression potential) than MMDH2, CMDH2, CMDH3 and PMDH1 in most of developmental stages of Arabidopsis thaliana ecotype Columbia. These overall transcript abundances of MDH isoforms in Arabidopsis leaf by using a laboratory reference ecotype Columbia provided some fundamental information before exploring the broader range of ecotypes.

211

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Figure 8. Hierarchical cluster analysis of MDH isoforms gene expression in different developmental stages of the Arabidopsis thaliana ecotype Columbia using the Genevestigator software. Heat map with red-white colour coding represented the percent of a given MDH gene expression level across all measurements available in the dataset expressed in log10 scale. Results for the developed leaf are highlighted (boxed in blue).

Three housekeeping genes namely clathrin, alkaline pyrophosphatase (PPase) and yellow-leaf-specific protein 8 (YLS8) (Hong et al., 2010) were found suitable for the qPCR assay as those primer sets showed a high consistency and similar gene expression pattern across the selected final 10 Arabidopsis ecotypes (data not shown). Quantitative PCR results in general showed that MMDH1 and PMDH2 gave the highest transcript abundance compared to the rest of the MDH genes (Figure 9).

212

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

The results of the qPCR assays were comparable to the fore-mentioned survey results using the Genevestigator software. Both studies found that based on gene expression data PMDH2, MMDH1, CHMDH and CMDH1 are the major isoforms of MDH genes in mature leaves of Arabidopsis. However there was no specific trend in any of the MDH isoform transcript level observed when ecotypes order was sorted in ascending respiration rates. The correlations between mean respiration rates and individual mean MDH isoform transcript levels in the selected final 10 Arabidopsis ecotypes were analysed using scatter plot and Pearson correlation method (Figure 10). It was observed that there was a weak relationship of respiration rates with any of the MDH isoform transcript levels, R2=0.0185, R2=0.0191, R2=0.00006, R2=0.0014, R2=0.0102, R2=0.0422, R2=0.1281 and R2=0.0271 for MMDH1, MMDH2, CMDH1, CMDH2, cMDH3, CHMDH, PMDH1 and PMDH2, respectively. It was noticed that the majority of the MDH isoforms transcript levels correlated positively with respiration rate except for CMDH2 and PMDH1 which showed negative relationships with respiration rates. These results indicated that none of the MDH isoforms transcript level could be significantly linked to the respiration rates in selected final 10 Arabidopsis ecotypes.

A hierarchical clustering analysis of the normalised relative MDH transcript levels across ecotypes was performed using average linkage and Pearson correlation method using the MultiExperiment Viewer (MeV) software (Figure 11). The order of ecotypes in the clustering analysis was sorted according to increasing respiration rates of ecotypes after combining all four independent measurements as discussed earlier. It was noticed that most of the MDH isoforms clustered well together except PMDH1 and CMDH3.

213

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Figure 9. Gene expression of MDH isoforms across the selected final 10 Arabidopsis ecotypes. Mean transcript level of MMDH1 (A), MMDH2 (B), CMDH1 (C), CMDH2 (D), CMDH3 (E), CHMDH (F), PMDH1 (G) and PMDH2 (H) (mean ±SE, n=3). The order of ecotypes was sorted according to ascending respiration rates.

214

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Figure 10. Relationship between respiration rates and MDH isoforms transcript levels in Arabidopsis ecotypes. Scatter plot of mean respiration rates, (n=27-39) and mean transcript levels of MMDH1 (A), MMDH2 (B), CMDH1 (C), CMDH2 (D), CMDH3 (E), CHMDH (F), PMDH1 (G) and PMDH2 (H) (n=3) from the selected final 10 Arabidopsis ecotypes are shown. Pearson correlation values, R2, for each respective relationship was shown.

215

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Figure 11. Hierarchical clustering of MDH gene expression in the selected final 10 Arabidopsis ecotypes. The heat map with a colour scheme from green to red indicates low to high relative abundance of MDH transcripts. Gene clustering was performed using the mean values of relative MDH isoform transcript levels across the ecotypes based on the average linkage and Pearson correlation method (scale shown as Pearson correlation coefficient).

In general the results of our qPCR assays are in line with the fore-mentioned survey results from Genevestigator where the major isoforms of MDH genes were also well clustered together. A closer look at the pattern of relative transcript level of mMDH isoforms across ecotypes revealed that MMDH1 and MMDH2 gave comparable relative abundance across ecotypes and similar to CMDH1, CMDH2 and CHMDH. There were only subtle changes in the relative abundance of PMDH2 across ecotypes compared to other MDH isoforms. This might indicate that the expression of PMDH2 is fairly consistent and exhibited the least transcript expression variation driven by the genetic differences among ecotypes. A Pearson correlation analysis of

216

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

the relative abundance of MDH isoforms across ecotypes was further performed using SPSS statistical software as shown in together except PMDH1 and CMDH3. In general the results of our qPCR assays are in line with the fore-mentioned survey results from Genevestigator where the major isoforms of MDH genes were also well clustered together. A closer look at the pattern of relative transcript level of mMDH isoforms across ecotypes revealed that MMDH1 and MMDH2 gave comparable relative abundance across ecotypes and similar to CMDH1, CMDH2 and CHMDH. There were only subtle changes in the Table 2. It was shown that MMDH1 significantly correlates with many other MDH isoforms such as MMDH2, CMDH1, CMDH2, CHMDH and PMDH2 with a correlation coefficient of R>0.75 (P<0.05) and similarly MMDH2 has a correlation coefficient of R>0.73 (P<0.05). Notably, the correlation coefficients were found to be higher for isoforms which belong to the same subcellular compartment; for instance between MMDH1 and MMDH2 and between CMDH1 and CMDH2 (R>0.80). Additionally, MDH isoforms that located at the neighbouring subcellular compartments were presumably highly correlated (R>0.80) particularly between mitochondrial and cytosolic MDHs as well as between cytosolic and chloroplastic MDH.

Table 2. Pearson correlation coefficient matrix of MDH isoform gene transcript abundances in the final 10 Arabidopsis thaliana ecotypes. The correlations are deemed to be significant at P<0.05 and P<0.01 marked by one asterisk (*) and two asterisks (**) respectively by two-tailed tests.

217

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Multiple reaction monitoring for robust quantitative of TCA cycle enzymes protein abundances in Arabidopsis ecotypes

In the recent development of plant proteomics research, the use of Selected Reaction Monitoring (SRM) and Multiple Reaction Monitoring (MRM) methods has been highlighted as an emerging approach to complement untargeted shotgun proteomics methods. These methods provide higher sensitivity and specificity in quantifying alternative forms of protein for instance protein isoforms, splice variants, amino acid polymorphism as well as proteins containing posttranslational modifications (Picotti et al., 2013). Evidences have shown these methods effectively estimate both high and low protein abundances of enzymes from various plant biological pathways using either absolute or relative quantitative approaches (Wienkoop and Weckwerth, 2006; Taylor et al., 2014). More specifically, MRM has been commonly used to quantify protein expression which provides high selectivity by monitoring chromatographic co-elution of multiple transitions for a given peptide. In provision of the above advantages, herein we have applied this technique to uncover changes of TCA cycle enzyme abundance in relation to the natural variation of ecotype respiration rates. The MRM protein assays for the individual Arabidopsis TCA cycle enzymes have been systematically developed and tested in Arabidopsis mitochondrial protein extracts as previously described in Taylor et al. (2014). The unique peptides including a precursor ion with 3 transitions each precursor (one quantifier and two qualifiers) and 3 peptides per protein were developed and optimised for each TCA cycle enzyme in Arabidopsis thaliana ecotype Columbia. Overall, a total of 101 peptides targeted to 33 isoforms encoding for 9 TCA cycle enzymes were analysed for their uses in this study. Those enzymes include pyruvate decarboxylase (PDC), pyruvate dehydrogenase (PDH), aconitase (ACO), citrate synthase (CS), isocitrate dehydrogenase (IDH), 2-oxoglutarate dehydrogenase complex (OGDC), 2-oxoglutarate dehydrogenase (OGDH), dihydrolipoyl dehydrogenase (DLD), succinyl-CoA synthetase (SUC), fumarase (FUM)

218

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

and malate dehydrogenase (MDH). In some cases peptides could detect multiple isoforms of a particular TCA cycle enzyme due to common peptide sequences shared between those protein isoforms. For instance, there are peptides that identify both protein isoforms of OGDC E2 components (At4g26910&At5g55070), DLD E3 components (At1g48030&At3g17240) and SDH subunit 1 components (At5g66760&At2g18450). At the end of the analysis of MRM assays, it was found that 34 peptides from 23 different proteins of the TCA cycle (Table 3) produced good MRM signals (high signal-to-noise ratios) across leaf protein samples extracted from the 10 selected Arabidopsis ecotypes. Taking into consideration that some unique peptides could potentially fail to be identified because of a non-synonymous nucleotide substitution in a certain ecotype, a mean value of TCA cycle protein abundances was obtained from their corresponding unique peptides before the relative protein abundance values across ecotypes were determined (detailed analysis workflow has been described in the methods and materials).

The changes in TCA cycle protein abundance across ecotypes are presented in heat map format in Figure 12. The blue-white-red colour scheme of the heat map denotes for low to high relative protein abundances of the corresponding TCA cycle protein, which further represent the natural variation of TCA cycle proteins among Arabidopsis thaliana ecotypes. When the position of ecotypes was sorted according to their increasing respiration rates, protein abundance of TCA cycle enzymes displayed a variety of clusters. Distinct gradual increases in protein abundances were noticeable for TCA cycle enzymes such as OGDC (At4g26910&At5g55070), SUC (At2g20420) and SDH (At5g66760&At2g18450) from slow to fast-respiring ecotypes. Conversely, in the same order of ecotypes, protein abundances of DLD (At1g48030), CS (At2g44530) and MDH (At1g53240) were progressively decreased implying an opposite trend in the latter group of enzymes. As mentioned earlier, there were two distinct groups of ecotypes that could be classified from the leaf respiration analysis, slow-respiring and fast-respiring (bottom and top performers) ecotypes.

219

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

on on

(Qual.), (Qual.),

TCA cycle enzymes which gave good signal Arabidopsis thaliana Arabidopsis ) mass/charge ratio; Pro 1 ion (Quant.), peptide product ion 1 Genome Genome Initiative identifier; Protein, protein name; Sequence,

ation; Pro 3 m/z (Qual.), peptide product ion 3 (qualifier) mass/charge al., 2010); Opt. CE, optimised CE; Area Ratio Pro 1/Pro 2, ratio of the area Arabidopsis

ecotypes. AGI,

Arabidopsis

Table 3. List MRM signature peptides developed and optimised for responses across final 10 peptide sequence; Precursor m/z, peptide precursor ion mass/charge ratio; Precursor m/z (Quant.), peptide (quantifier); precursor Pro ion 1 mass m/z (Quant.), peptide product ion 1 (quantifier (quantifier) fragmentation series location; Pro 2 m/z (Qual.), peptide product ion 2 (qualifier) mass/charge ratio; Pro 2 ion peptide product ion 2 (qualifier) fragmentation series loc ratio; Pro 3 ion (Qual.), peptide product ion column; Pred. CE, predicted CE from Skyline (MacLean et 3 (qualifier) fragmentation series location; RT, retention of time extracted of ion the chromatograph peptide of product ion product ion 3. ion 1/product chromatograph of 1/product ion 2; Area Ratio Pro 1/Pro 3, ratio of the area of extracted ion 220

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

221

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

r ch

05)

ositive o for for whi relation is

ecotypes. The order of ecotypes is

Arabidopsis red red colour scheme indicates low to high abundance of the - white -

Heat Heat map of relative protein abundance in the respiratory TCA cycle across

the 2 distinct groups of ecotypes with significant low respiration rate (ecotype 38, 39, 6, 2 and 49) and significant low

Figure 12. sorted according to increasing respiration rates (as shown in top left bar chart).AGI numbers indicate the protein identities MRM peptide markers gave good corresponding TCA signal cycle protein. responses. TCA cycle Blue proteins boxed in between red denote for significant differential expressed proteins (P<0. respiration rate (ecotype 44, 46,30, 25 and 20). Multiple linear regression analysis led to the prediction of proteins with p negative correlation with increasing respiration rates. They are highlighted in bold in red and blue respectively and the cor **)*) and withrespectively. P<0.01 (marked with P<0.05 at(marked tobe deemed significant 222

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Therefore a Student’s t-Test analysis was performed using the means of relative abundance of the TCA cycle proteins across ecotypes from 3 independent biological replicates to distinguish TCA cycle proteins that were significantly different for the two groups of ecotypes. MDH (At1g53240) and SDH (At5g66760&At2g18450) were significantly differentially expressed proteins between slow and fast-respiring ecotypes (P=0.018 and P=0.025 respectively). There was a contrast in the protein abundance profile for these enzymes, with decreasing and increasing abundance profiles for MDH and SDH, respectively. Hence, it is evident that the significant low or high respiration rates exhibited by Arabidopsis ecotypes best correlated with the genetic variation within the TCA cycle of protein expression of these two enzymes.

Next the relationship between TCA cycle enzymes was examined in an attempt to understand on how the stoichiometry of these 23 TCA cycle enzymes are set, particularly the changes of other TCA cycle enzymes stoichiometries in response to altered MDH protein expression in Arabidopsis ecotypes. A protein distance matrix was then constructed using the relative protein abundance across ecotypes and the Pearson correlation method within the MeV software (Figure 13). This distance matrix analysis gives an intuitive and comprehensive view of the distance (or similarity) between any two proteins in a coloured gradient matrix. It provides a general picture of the correlations between TCA cycle protein abundances through which potentially co-regulated proteins within the respiratory network could be predicted. Malate dehydrogenase (At1g53240) showed good correlation with IDHs (At5g03290 and At5g14590), OGDH (At4g26910&At5g55070) and a number of SDHs (At3g47833, At5g66760 and At2g18450) and FUM (At2g47510). A trend displayed in this protein distance matrix analysis is that neighbouring TCA cycle proteins are correlated well to each other. This could be seen from the protein pairs between FUM and MDH, ACO and IDHs as well as SUC and SDHs.

223

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Figure 13. Distance matrix analysis of TCA cycle proteins of Arabidopsis ecotypes. The matrices between protein pairs were computed using a Pearson correlation method. The dark red to bright yellow colour scheme denotes low to high similarity of the corresponding protein pairs in terms of their protein abundance pattern across ecotypes.

224

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

The distance matrix analysis of TCA cycle protein abundances provides information on how strongly the two proteins are correlated, however it does not show whether they are positively or negatively correlated. Therefore, the relationship of TCA cycle proteins was further analysed using Pearson correlation. This provides the nature of the relationship i.e. positive or negative correlation, in addition to correlation coefficient values, indicating the strength and significance level of the relationship. A number of significant relationships between TCA cycle enzymes based on their protein abundances were predicted as shown in Table 4. For simplicity, the relationships between MDH and other TCA cycle enzymes are discussed as the primary focus of this study. It was predicted that MDH (At1g53240) significantly and positively correlated with DLD (At1g48030) (R=0.399, P<0.05), which is also known as PDC E3 subunit 1. Conversely, a significant negative correlation was found between MDH (At1g53240) and IDHs (At5g03290 and At5g14590) with correlation coefficients of R=-0.370 (P<0.05) and R=-0.470 (P<0.01) respectively. It was also found that there were several TCA cycle enzymes which showed more than 2-fold differences in protein abundance (difference between minimum and maximum relative protein abundance values) across the final 10 Arabidopsis ecotypes. Those enzymes include IDHs (At5g03290 and At5g14590) (2.1 and 2.4-fold), FUM (At2g47510) (2.2-fold), DLD (At1g48030) (2.9-fold) and PDC E2-3 (At1g54220) (3.1- fold). These large deviations of relative protein abundances of TCA cycle enzymes across ecotypes which could result in ecotype-specific respiratory enzyme expression and should be investigated further.

225

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

(At1g53240) MDH

tailed

(At2g47510) FUM .053 -

(At5g66760&At2g18450) SDH .183 .268 -

ecotypes. Values (At5g66760) SDH .344 .119 .109 -

(At3g47833) SDH .286 .322 .022 .154 -

(At1g08480) SDH

.249 .286 .126 .255 - .034 Arabidopsis Arabidopsis

(At2g20420) SUC .226 .211 .458* .205 .018 - - - .136 - - across

(At1g48030&At3g17240) DLD .130 .114 - .120 .228 .374* .090 .068 -

(At1g48030) DLD .191 .099 .328 .287 .358 - .004 - - - - .088 .399*

(At4g26910&At5g55070) OGDC .341 .165 .230 .148 .194 - .120 - - .533** .344 - .063 -

) At5g65750 ( OGDH .339 .260 .294 .362* .240 .023 - .087 - .146 - - - - .256 .051

) At3g55410 ( OGDH .232 .080 .232 .117 .209 .037 .295 - .108 .263 .023 - - - - - .178

0**

(At5g14590) IDH .298 .063 .161 .302 .251 .47 - - .120 .144 - - .047 .152 .514** - .105 -

(At5g03290) IDH .280 .215 .085 .249 .370* .329 - .126 .293 .022 - - - .352 .182 .113 .429* -

(At4g35260) IDH .226 .156 .271 .072 .146 .199 .135 .212 .369* - - - .067 - - - - .001 - .105 .047 - .343

(At2g17130) IDH .398* .031 .053 .190 .291 .305 .350 - .347 .019 .117 .206 - - - .408* - .051 - .021 .067 -

(At4g26970) ACO .231 .346 .155 .038 .526** .031 .163 .257 .128 - .130 - - .047 - - - .246 - .522** - .106 .361* - .218

(At2g05710) ACO .075 .327 .362* .128 .145 .278 .271 .027 .132 .068 .168 .132 - - - .541** .102 - - .552** .119 - - - - - .119

(At2g44350) CS .001 .227 .153 .053 .301 .517** .158 .469** .349 .097 .412* .050 .180 - - .110 .281 .207 - .110 - .197 ------

sks sks (**) indicate the significant level of corresponding correlation at P<0.05 and P<0.01 (two (At5g50850) PDH .078 .086 .227 .240 .075 .109 .220 .329 .280 .440* .319 .151 - .230 .046 - - .255 .008 - .273 - .412* - - - - - .140

ation ation coefficient matrix of relative abundance of TCA cycle proteins

PDC (At3g13930) PDC 5 .135 .210 .379* .238 .250 .015 .002 .241 .278 .340 .187 .214 .285 - - .106 - - .16 - .038 .130 .320 .174 - - - - - .191

(At1g54220) PDC .208 .061 .019 .184 .132 .159 .086 .317 .055 - - - .008 - .237 - .472** .212 .009 .554** .056 .033 - - - .057 - .019 .625** .010

) )

respectively.

At3g55410 At5g65750 ( (

H (At4g35260)H (At1g54220)PDC (At3g13930)PDC PDH(At5g50850) (At2g44350) CS (At2g05710)ACO (At4g26970)ACO (At2g17130)IDH ID (At5g03290)IDH (At5g14590)IDH OGDH OGDH (At4g26910&At5g55070)OGDC (At1g48030)DLD (At1g48030&At3g17240)DLD (At2g20420)SUC SDH(At1g08480) SDH(At3g47833) SDH(At5g66760) SDH(At5g66760&At2g18450) (At2g47510)FUM (At1g53240)MDH Table 4. Pearson correl highlighted in orange and green denote for positive and negative correlation of the paired proteins. Values marked with one asterisk (*) and two asteri test)

226

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

The relationships between ecotypes respiration rates and their underlying TCA cycle protein abundances were next investigated using multiple linear regression analysis with backward elimination selection (Table 5). This was undertaken because our primary interest was to decipher the minimal subset of TCA cycle enzymes that could be used to explain the respiration rate of Arabidopsis ecotypes. Multiple linear regression (MLR) is often used to predict the relationships between the predictors (in this case relative protein abundance of TCA cycle enzymes estimated from ecotypes) and criterion variables (in this case respiration rates of ecotypes). A backward elimination selection was chosen so that the weakest predictor variables were removed and regression re-calculated. The removal and recalculation steps were repetitive until only useful predictor variables remained in the final regression model. The R square is the percentage of the response variable variation that is explained by a linear model while the adjusted R square is value of R square adjusted for the number of explanatory terms in a model relative to the number of data points. Standardised coefficient of predictor gives a measure of the contribution of each variable to the model. A large value indicates that a unit change in this predictor variable has a large effect on the criterion variable. The t statistic is the coefficient divided by its standard error. The t and Sig. (P) values give a rough indication of the impact of each predictor variable for instance a big absolute t value and small P value suggests that a predictor variable is having a large impact on the criterion variable.

Mean respiration rates of ecotypes were used as dependent variables (criterion variables) while individual relative abundance of TCA cycle proteins across the 10 ecotypes from all 3 biological replicate were used as independent variables (predictors) in the regression model. The best predicted regression model revealed an adjusted R square, R=0.379, between the predicted set of 23 TCA cycle enzymes protein abundances and respiration rates of ecotypes. Only 1 protein was predicted to correlate positively (P<0.05) while 3 proteins correlated negatively (P<0.01) with

227

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

increasing respiration rates in ecotypes. OGDC (At4g26910&At5g55070) was the sole protein which showed significant positive correlation with respiration rates (Standardised coefficient, Beta=0.423; t=2.409; P=0.024). It is interesting to note that MDH (At1g53240) demonstrated the strongest impact on ecotype respiration rates by showing a negative correlation at the highest t value but at the lowest P value (Beta=-0.592; t=-3.70; P=0.001). Similarly, SDH (At3g47833) and CS (At2g44350) correlated significantly in an opposite direction with respiration rates with t=-3.117 and t=-2.888 (P<0.01).

The relationships between the above mentioned ecotype relative TCA cycle enzyme relative protein abundances and their respiration rates were further analysed using scatter plots coupled with Pearson correlation analysis. Relative protein abundance values of individual TCA cycle enzyme of MDH (At1g53240), CS (At2g44350), SDH (At3g47833) and OGDC (At4g26910&At5g55070) from 3 independent biological replicates against respiration rates were used to construct scatter plots as shown in Figure 14 (A-D). The predicted correlation values between MDH (At1g53240), CS (At2g44350), SDH (At3g47833), OGDC (At4g26910& At5g55070) and respiration rates were R2=-0.1855, R2=-0.0338, R2=-0.00005 and R2=0.0674 respectively. Obviously, MDH (At1g53240) showed the highest negative correlation value with respiration rate indicating its decrease in protein abundance when ecotype respiration rates were elevated. Both multiple linear regression and Pearson correlation with scatter plot analyses consistently demonstrated that MDH (At1g53240) protein was inversely correlated to Arabidopsis ecotype respiration rates. These findings imply that amongst the TCA cycle, MDH protein expression level is a major determinant for the natural variation of respiration rates in Arabidopsis ecotypes.

228

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Table 5. Modelling of relationship between Arabidopsis ecotypes relative TCA cycle protein abundances and their respiration rate using multiple regression analysis with backward elimination method (cut-off value at P>0.05). Respiration rates of final 10 ecotypes were used as dependent variables and relative protein abundance across ecotypes (n=3) as independent variables. The table listed the significant (Sig.) proteins (P<0.05) for respiration rate in the best model. Beta indicates the standardised coefficient regression weight while t value denotes for the impact of individual protein on respiration rate. R square and adjusted R square indicate the correlation of proteins collectively with respiration rates in the best predicted model.

Figure 14. Scatter plot and Pearson correlation analyses of TCA cycle enzymes and natural variation of respiration rates in Arabidopsis ecotypes. Relative protein abundance of individual TCA cycle enzyme of MDH (At1g53240) (A), CS (At2g44350) (B), SDH (At3g47833) (C), OGDC (At4g26910& At5g55070) (D) correlate to respiration rates with R2=0.1855, R2=-0.0338, R2=-0.00005 and R2=0.0674 respectively.

229

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Metabolites variation underpinned by distinct respiratory differences among Arabidopsis ecotypes identified

Information on the varying metabolites levels across Arabidopsis ecotypes could provide us with important clues to the observed natural variation of respiratory metabolism. This is supported by several lines of evidences pointing out that there are different metabolic traits in Arabidopsis accessions (Cross et al., 2006; Agrawal et al., 2012; Bai et al., 2012; Sulpice et al., 2013). However no link between changes in metabolite levels and varying respiration rates in Arabidopsis ecotypes have been established yet. In this current study, the primary metabolite levels were measured using GC-MS in the selected final 10 Arabidopsis ecotypes with 6 biological replicates each. Relative metabolite level values were used to construct heat maps and the order of ecotypes was sorted according to increasing respiration rates. An individual metabolite heat map was recapitulated for the primary metabolic pathway as shown in Figure 15. It was observed that approximately 70% of the metabolites studied showed subtle changes across ecotypes except for a couple of metabolites which showed great variations (>3-fold difference). There were great differences seen mainly in the level of carbohydrates among ecotypes such as fructose, galactose and glucose with 25.1-fold, 20.5-fold and 11.8-fold differences respectively. Glycine and serine which are involved in photorespiration also displayed a considerable variation among ecotypes, with 10.7-fold and 3.43-fold differences, respectively. While amino acids which include glutamine, threonine and isoleucine were found to differ as much as 6.0-fold, 3.7-fold and 3.6-fold respectively from our experimental ecotypes. Aside from that, other compounds and organic acids with substantial variations in ecotypes were gamma aminobutyric acid (GABA), malate and fumarate with 4.1-fold, 3.93-fold and 3.2-fold differences, respectively. From Student’s t-Test results, the levels of 8 metabolites were found to be significantly different between bottom and top performers in ecotype

230

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

respiration groups. Those metabolites were serine, threonate, isoleucine, threonine, fructose, galactose, glucose and myo-inositol with the first two metabolites significantly different at P<0.05 and the rest showed a higher significance level of P<0.01.

A closer examination by correlating the detected 26 primary metabolites among ecotypes using a hierarchical clustering approach coupled with Pearson correlation analysis revealed the potential of natural genetic variation of ecotypes in shaping the metabolite correlation networks (Figure 16). It was clearly seen that metabolites involved in sugar metabolism such as fructose, galactose, glucose and sucrose were clustered together and its neighbouring sub-clusters comprised of amino acids such as serine, isoleucine, threonine as well as GABA. Ecotype 25 (the top performer of the high respiration group) consistently showed the greatest abundance of hexoses (fructose, galactose and glucose) conversely to the profile displayed by ecotype 39 (the bottom performer of low respiration group). All the organic acids from TCA cycle intermediates namely succinate, fumarate, malate, citrate, isocitrate were found closely connected to each other and clustered under the same node as glutamine which is the downstream product of glutamate and 2-oxoglutarate (both the metabolites were under the detection limit in the current study). Aside from that, amino acids such as valine and glycine had comparable abundance profiles where ecotypes with significant higher respiration rates (ecotype 20, 25 and 30) gave markedly high abundances notwithstanding the discordance seen in ecotype 38 (ecotype of which respire at the slowest rate). There were 2 nodes of metabolites which demonstrated subtle deviations of their levels across ecotypes. Seemingly the production of those metabolites in ecotypes is fairly steady with the least effects from natural genetic variation among ecotypes.

231

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

red colour scheme denotes for low to -

white -

es of ecotypes respectively using multiple linear regression regression analysis. linear multiple using es of ecotypes respectively 2 and 49) and top performers (ecotype 44, 46, 30, 25 and 20) in the context of their respiration rates at

ndance for the corresponding metabolite respectively. Gray boxes indicate that metabolites could not be detected or the the TCA cycle in Arabidopsis ecotypes. The heat map represents the relative metabolite levels across ecotypes sorted by

Figure 15. The effects of natural variation in respiration rates on carbohydrate metabolism, and glycolysis, amino acid metabolis increasing respiration rates (results as demonstrated in top left bar chart). Green high abu were not measured. Metabolites with ecotype colour boxes framed in red denotes (ecotype for 38, significant 39, difference 6, between bottom P<0.05. The red and blue boxed metabolites indicate increasingrespiration rat significant positive and negative correlation (P<0.05 and P<0.01) with 232

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Metabolite level of Arabidopsis ecotype

Figure 16. Hierarchical clustering and heat map of metabolite profiles of Arabidopsis ecotypes sorted by increasing respiration rates. Mean relative metabolite level is demonstrated with mean ratio of normalised ribitol metabolite response values to mean response value of the corresponding metabolite calculated from 5-6 individual biological replicates. The heat map presented in dark green to bright red colour scale denotes for low to high metabolite level in ecotypes. The clustering analysis was performed using the Pearson correlation method.

233

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

The relationship between metabolites was further investigated using the Pearson correlation coefficient matrix output from the SPSS statistical package as shown in Table 6. Overall, it was found that there were many significant relationships between metabolite pairs (P<0.05) and the majority of them were positively correlated (cells shaded in orange). Conversely, a few metabolites such as valine, glycerol, isoleucine and GABA showed negative relationships with their paired metabolites. For instance, valine was consistently and negatively correlated with citrate, isocitrate, succinate and fumarate in the TCA cycle and amino acids synthesised from TCA cycle intermediate such as glutamine and putrescine. Similarly, glycerol and isoleucine showed significant negative correlations with their paired metabolites which were primarily from TCA cycle intermediates. Hexoses such as fructose, galactose and glucose were shown to correlate very strongly with each other (R>0.97). There were also strong positive correlations between TCA cycle intermediates; notably neighbouring organic acids displayed stronger correlation coefficients than others. For instance, the correlation coefficient between malate and fumarate was found to be stronger than between malate and citrate as well as between malate and isocitrate (R=0.840, R=0.765, R=0.618 respectively). These results were in line with the hierarchical clustering findings as previously discussed.

The relationships between individual metabolites and ecotypes respiration rates was subsequently explored using multiple linear regression analysis with the backwards elimination method using the SPSS statistical package (Table 7). The best predicted regression model contained 13 metabolites with significant correlations (P<0.01). Collectively these metabolites were closely linked to respiration rates with an overall adjusted correlation, R2=0.937. There were 5 metabolites predicted to be positively, while 8 metabolites were negatively, correlated with respiration rates. Among those metabolites, isoleucine and glycerol displayed the strongest positive effects on respiration rates, t=10.246 and t=8.266 respectively. Conversely, serine exhibited a strong and negative correlation with respiration rates (t=-10.727). All

234

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

other metabolites such as hexoses (galactose, fructose and sucrose), organic acids (succinate and fumarate), amino acids (valine, threonine, beta alanine and tyrosine) and shikimate showed relatively weaker correlations when respiration rates, with estimated t values of less than 7.

235

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

tailed tailed -

ecotypes primary metabolites. Values

Arabidopsis Arabidopsis

e and green denote for positive and negative correlation between the corresponding metabolites,

Table 6. Pearson correlation coefficient matrix of relative abundance of highlighted in orang respectively. Values with one asterisk (*) or two asterisks (**) are deemed to be significant at P<0.05 and P<0.01 (two test) respectively.

236

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Table 7. Relationship between the metabolite level and respiration rate in final 10 selected Arabidopsis ecotypes using a multiple regression analysis with backward elimination method (cut-off value at P>0.05). Respiration rates were used as dependent variables and mean relative metabolite level across ecotypes obtained from 5-6 individual biological replicates as independent variables. Note that the best predicted model is shown in table. Beta value indicates standardised coefficient regression weight while R square and adjusted R square values indicate the correlation of the best predicted model.

Standardized R Adjusted Model Coefficients t Sig. Square R Square Beta (Constant) 14.125 0.000 0.970 0.937 Valine -0.809 -6.850 0.000 Serine -2.202 -10.727 0.000 Glycerol 0.737 8.266 0.000 Isoleucine 3.919 10.246 0.000 Threonine -1.395 -4.735 0.000 Succinate -1.176 -6.226 0.000 Fumarate 1.094 5.835 0.000 Beta alanine -0.400 -3.727 0.003 Shikimate -0.442 -4.399 0.001 Fructose -2.056 -4.921 0.000 Galactose 2.522 4.926 0.000 Tyrosine 0.304 5.277 0.000 Sucrose -0.352 -3.143 0.008

237

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Discussion

Natural variation of Arabidopsis ecotypes respiration rates is independent of mitochondrial mass

Natural variation of respiration rates among Arabidopsis thaliana ecotypes has not been reported elsewhere and therefore my study has taken the initiative to explore its potential to understand respiration as a complementary approach to classical reverse genetics of respiratory components. Both Student’s t-Test and one-way ANOVA distinguished two distinct ecotype groups where significant differences in respiration data between bottom performers (ecotypes 38, 39, 6, 2 and 49) and top performers (ecotype 44, 46, 30, 25 and 20) for significant low and high respiration rates respectively. In order to examine whether the natural variation of respiration rates in ecotypes is due to the differences in their mitochondrial mass, the porin amounts from those ecotypes were subsequently determined. Western blot analysis using two different porin antibodies, anti-Maize porin and anti-AtVDAC1 antibodies showed good consistency with a correlation of R=0.580 (P<0.05). However a weak correlation was found between porin content of Arabidopsis ecotypes and their respiration rates with a Pearson correlation of R2=0.312 and R2=0.304 (P<0.05) for anti-Maize porin and anti-AtVDAC1 antibodies respectively. These findings imply that the varying respiration rates seen in Arabidopsis ecotypes is weakly connected to the differences of their mitochondrial mass. This indicates that natural variation of ecotypes respiration rates could be due to the underlying genetic differences leading to a specifically modified respiratory capacity and metabolism in the different ecotypes.

There are an increasing number of Arabidopsis ecotypes available from a broad, world-wide distribution over a range of geographical locations. It has been long established that genetic variation in a plant species is structured in space and time (Loveless and Hamrick, 1984). The adaptation of these ecotypes to their varying

238

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

natural growing environments and conditions has been associated with the evolution of Arabidopsis genes. Therefore, a large pool of naturally adaptive traits in A. thaliana has made this model organism excellent for studying the genetics of natural variation (Bakker et al., 2006; Weigel and Mott, 2009). Evidence relating to the evolutionary ecology perspectives and the resulting genetic traits of Arabidopsis ecotypes has been previously shown (Mitchell-Olds, 2001; Shindo et al., 2007; Bergelson and Roux, 2010). Previous findings in sugar maple populations have shown that populations originating from high latitude and altitudes often respire faster in comparison to their counterparts from lower latitudes and altitude, when grown under the same conditions (Ledig and Korbobo, 1983).

Significant positive correlations between altitudes and respiration rates had been found in Norway spruce tree (Picae abies) (Oleksyn et al., 1998). These studies imply natural genetic variation of respiratory trait could potentially improve the physiology, growth and development of ecotypes from a plant species under changing environments. In an earlier section, we reported that our set of Arabidopsis ecotypes which originated from geographically regions of different latitudes exhibited varying respiration rates in a uniform laboratory environment. Further analysis on the relationship between mean respiration rates of 49 Arabidopsis ecotypes obtained from the first batch of respiration measurement and their corresponding latitudinal origins (as shown in Table 1) revealed a significant correlation coefficient value, R=0.336 (P=0.018) and R=0.359 (P=0.011) using Pearson and Spearman’s rho correlation method respectively (Table 8). Taken together with the literature, this implies that a plausible link between important plant physiology processes such as respiration rate and the latitudinal gradient of Arabidopsis ecotypes may exist.

239

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Gene expression of MDH isoforms among Arabidopsis ecotypes

As western blot analysis showed weak correlation between natural variation of respiration rates and their varying mitochondria mass, there are likely other underlying mechanisms that modulate the significantly high leaf respiration rate in some Arabidopsis ecotypes. Previous findings showed a null allele for mitochondrial MDH isoforms had significantly elevated leaf respiration rate in Arabidopsis plant (Tomaz et al., 2010). I therefore hypothesised that the varying expression level of mitochondrial MDH gene and protein isoforms could be a key factor in determining the natural variation of respiration rates of ecotypes. Hence, the measurements of gene expression level of MDH isoforms as well as other MDH isoforms were performed from the final 10 Arabidopsis ecotypes, considering that there could have been some compensatory mechanisms involving other MDH isoforms from different subcellular compartments.

Results of quantitative PCR assays revealed noticeable differential transcript levels of eight MDH isoforms in the final 10 Arabidopsis ecotypes. However, no specific trend of transcript level (decreasing or increasing in transcript level) was observed in either mitochondrial MDH isoforms or any of the MDH isoform when ecotype position is sorted according to their mean respiration rates. Comparison of transcript profiles between neighbouring MDH isoforms (isoforms located in the same or adjacent subcellular compartment) revealed none of the MDH isoform exhibited obvious compensatory mechanism. These could imply that unlike the MDH gene perturbation study, gene expression of MDH isoforms are in steady-state conditions under natural genetic variation of Arabidopsis ecotypes. There were only weak correlations between respiration rates and the MDH isoforms in the selected final 10 Arabidopsis ecotypes (R2<0.1281). No significant difference was found for any of the MDH isoform expression level between the distinct ecotype respiratory groups implying respiration rate was not correlated with MDH isoforms transcript levels. The above findings could be explained by a possibility of ecotype-specific

240

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

gene expressions among MDH isoforms which have complicated the prediction of relationship between respiration rates and their MDH isoform transcript levels. The current MDH isoforms transcript data solely may also not sufficient to reveal a promising relationship between them. Transcript abundances for the 4 MDH isoforms MMDH1, CMDH1, CHMDH and PMDH2 were to be found particularly higher than the rest of MDH isoforms. This finding is in agreement with the MDH genes expression profiles in mature rosette of Arabidopsis ecotype Columbia obtained from Genevestigator software (Figure 8) where those MDH isoforms are the major MDH gene isoforms in the Arabidopsis ecotype Columbia based on microarray hybridization signal intensity.

Pearson correlation analysis using SPSS statistical software showed that the correlation coefficient value of a particular paired MDH isoforms should be more than 0.6 in order for a significant relationship to exist at P<0.05. When strong and significant correlations (R>0.6, P<0.05) between MDH isoforms exist, those relationships in general display the following features: the major isoforms of MDH involved display the following features: 1) they share the same enzymatic reaction within the same subcellular compartment for instance MMDH1 and MMDH2 gene (R>0.90) and 2) they have higher transcript abundance which largely encodes for MDH enzymes involved in a sequential biochemical reactions in a neighboring subcellular compartment (such as the major isoforms of MDH: MMDH1, CMDH1, CMDH2, CHMDH and PMDH2; R>0.80). This is in concordance with the previous reports that correlations between transcripts are often stronger for genes encoding for enzymes that catalyse nearby reactions than metabolically distant reactions (Kharchenko et al., 2005).

241

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Table 8. Relationship between A. thaliana ecotypes latitudinal origins and their respiration rates. Correlation analysis was performed using mean respiration rates of 49 Arabidopsis ecotypes obtained from the first batch of respiration measurement and their respective latitudinal origins as listed in Table 1. The correlation coefficient values were deemed to be significant at P<0.05 for both Pearson and Spearman’s rho correlation methods and marked as (*).

Abundance profile of TCA cycle enzymes and metabolites and their intra- relationships in Arabidopsis ecotypes

While quantitative PCR assay results revealed a minimal connection between mitochondrial MDH and respiration rates, there were highly varying MDHs transcript abundances among ecotypes. The protein expression levels of mitochondrial MDHs in ecotypes were then subsequently measured to further test the relationship. There are no commercial antibodies available for mitochondrial MDHs to differentially assess their protein abundances using Western blotting, so a mass spectrometry based quantification approach was chosen in this study instead. Single reaction monitoring (SRM) has been established and optimised in our laboratory to assess the abundance of specific proteins as reported previously (Taylor et al., 2014). It is of great advantage using MRM assays for concurrent examination of protein abundances changes in both MDH and other TCA cycle enzymes among Arabidopsis ecotypes.

In a proteomics study on 8 Arabidopsis ecotypes using 2-D gel electrophoresis and peptide mass fingerprinting, there was ecotype-specific protein expression

242

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

identified for 25 major protein spots (Chevalier et al., 2004). The most varied major proteins were found to be involved in energy and carbon metabolism (glycolysis, TCA cycle and mitochondrial respiration) implying the greatest variation of proteins was responsible for important functions in those ecotypes. We report a pilot study in exploring the natural variation of ecotype-specific protein expression focusing on respiratory metabolism. From the results a total of 23 proteins encoding for 9 TCA cycle enzymes revealed noticeable protein expression variations ranging from 1.20- fold to 3.08-fold differences among the ecotypes tested (Figure 12) which is in agreement with the previous report by Chevalier et al. (2004). By sorting the ecotype positions according to their respiration rates from the lowest to the highest rate, it was noticed that approximately 35% of the TCA cycle enzymes gave correlative trends of protein abundances with respiration rates. Interestingly, we found that MDH (At1g53240) protein expression showed a gradual decreasing trend in its protein abundances as the respiration rates of ecotypes increased. This is preliminary evidence for our hypothesis that reduced MDH gene expression can elevate leaf respiration rate. Other TCA cycle enzymes which show decreasing trend in their protein abundances include DLD (At1g48030) and CS (At2g44530) in contrast to OGDCs (At4g26910&At5g55070), SUC (At2g20420) and SDHs (At5g66760&At2g18450) which gradually increased in their protein abundances.

Some of the TCA cycle enzymes such as PDH (At1g54220), ACO (At4g26070) and IDH (At5g03290) demonstrated ecotype-specific protein abundance. Considering TCA cycle enzymes work collectively in the respiratory metabolism, it is likely that some coordinated changes in protein expression between TCA cycle enzymes would occur. To test this, Pearson correlation analysis was used to calculate correlations between TCA cycle enzymes abundances in Arabidopsis ecotypes (Figure 13 and Table 4). Pearson correlation method has been commonly used to predict relationships between biological data (data of transcripts, proteins and metabolites) within an interactive network (Chen et al., 2010; Mutwil et al., 2010). Results of

243

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Pearson correlation analysis presented as a protein distance matrix (Figure 12) and a correlation coefficient matrix (Table 4) gave comparable outcomes. The protein distance matrix provided a graphical representation of the strength of relationship between protein pairs while the correlation matrix table provided the correlation values, significant levels and directions of the relationship between the protein pairs (i.e. positively or negatively correlated). There are several proteins pairs which showed strong and significant relationships (R>0.5, P<0.01) indicating these proteins work in a highly coordinated manner. An overall interpretation of the relationships between TCA cycle enzymes is that two distinct protein groups which are either co-expressed or differently expressed based on their interaction network have been identified. TCA cycle enzymes which are potentially co-expressed in Arabidopsis ecotypes alongside MDH (At1g53240) are PDH complex E2-2 component (At3g13930), PDH E1 component (At5g50850), PDC E3-1 (At1g48030), CS (At2g44350), ACO (At2g05710), IDH (At4g35260) and OGDH E1 subunit (At3g55410). Conversely, the identified differently expressed proteins with respect to MDH (At1g53240) are PDH E2-3 component (At1g54220), IDHs (At5g03290 and At5g14590), OGDC E2-2 and E2-1 components (At4g26910& At5g55070), SDHs (At5g66760, At3g47833 and At5g66760& At2g18450) and FUM (At2g47510). Considering there are post-translational modifications that modulate the activity and macromolecular interactions of proteins, these apparent groups need to be further investigated by new experiments such as protein-protein interaction assays which could characterise protein relationships within a biological network (Chen and Xu, 2003). Given the importance of TCA cycle enzymes in respiratory metabolism, understanding the co-expression and differential expression of TCA cycle enzymes could facilitate us the selection of gene candidates for metabolic engineering to improve the growth, development and/or stress response of plant.

Leaf metabolite profiling was conducted on the final 10 Arabidopsis ecotypes as an alternative way to further investigate the natural variation of respiratory

244

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

metabolism in ecotypes. Although the changes of a majority of the metabolite levels were fairly constant across ecotypes, we found the greatest variations in hexoses (fructose, galactose and glucose) which accounted for >11-fold differences among ecotypes. Ecotype 25 which was one of the top performers in respiration rates demonstrated strikingly high hexoses levels in leaf in contrast to ecotype 39 and 2 (bottom performers). Moreover, there were significant differences of hexoses levels between slow-respiring and fast-respiring ecotypes according to Student’s t-Test analysis (P<0.01) indicating hexoses could be potential metabolite markers of ecotype respiratory trait. The second group of metabolites significantly different between the two ecotype respiratory groups were amino acids (serine, threonate, isoleucine and threonine) (P<0.05). Taken together, these findings imply that natural variation of respiratory metabolism in Arabidopsis ecotypes is believed to be associated with their carbohydrate steady-state and their amino acid metabolism.

The leaf metabolites revealed many significant correlations between them, with stronger connections seen between metabolites that belong to the same functional categories e.g. aliphatic amino acids (valine and isoleucine) (R>0.5, P<0.01), organic acids (succinate, fumarate, malate, citrate and isocitrate) (R>0.6, P<0.01) and carbohydrates (fructose, galactose, glucose) (R>0.9, P<0.01). These suggest that neighbouring metabolites or metabolites in the same pathway are highly correlated with each other which are consistent with the findings reported recently (Sulpice et al., 2013). Notably, TCA respiratory intermediates (succinate, fumarate, malate, citrate and isocitrate) showed positive correlation coefficient with glutamine (R>0.6, P<0.01) and strongly correlated with citrate and isocitrate (R>0.96, P<0.01), indicating a high connectivity between respiratory metabolism and nitrogen assimilation in ecotypes. Besides that, amino acids (serine, isoleucine, threonine and glycine) showed many significant positive correlations with carbohydrates (fructose, galactose, glucose and sucrose) (R>0.3, P<0.05) reflecting coordinated changes between amino acid metabolism and sugar metabolism in Arabidopsis

245

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

ecotypes. This was evident that there were considerably strong relationships between enzymes involved in amino acids metabolism and sucrose breakdown pathway among 129 Arabidopsis ecotypes (Sulpice et al., 2013). However, valine was found consistently to be correlated significantly in an opposite direction with TCA cycle intermediates (R<-0.2, P<0.05). Contrarily, valine showed significant positive relationships with isoleucine, serine, glycine and GABA in the ecotypes set of this current study (R>0.3, P<0.01). Intriguingly, valine could be an alternative respiratory substrate when there are insufficient abundances of TCA cycle intermediates in ecotypes under adverse environmental conditions. Degradation of branched-chain amino acids (valine, leucine and isoleucine) by isovaleryl-coenzyme A dehydrogenase (IVD) in mitochondria (Daschner et al., 2001) has been suggested as an alternative way of providing electrons to respiratory electron transfer chain for energy generation (Ishizaki et al., 2005; Ishizaki et al., 2006; Araujo et al., 2010), and to support alternative respiration under adverse environmental conditions. Moreover, the enhanced conversion of glycine to serine by glycine decarboxylase (GDC) could increase NADH formation for energy generation. In addition, accumulation of GABA has been previously linked to cold acclimation and frost tolerance in barley and wheat (Mazzucotelli et al., 2006). The above-mentioned accumulation of free amino acids and GABA could be plausibly explained as genetic acclimation to low temperature growing origin particularly for significantly fast- respiring ecotypes.

Selective primary TCA cycle respiratory components and metabolites significantly correlated with natural variation of respiration rates among Arabidopsis ecotypes

The ultimate aim of this current study was to identify the key components responsible for natural genetic variation of respiratory metabolism in Arabidopsis ecotypes by exploiting their varying respiration rates. Apart from performing the

246

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

Student’s t-Test and correlation analysis to identify significantly expressed TCA cycle proteins and metabolites between distinct ecotypes respiratory groups (as discussed earlier), I have conducted multiple regression analysis to predict proteins and metabolites in which their significant changes statistically account for natural variation of respiration rates among the selected final 10 Arabidopsis ecotypes. The latter analysis creates a regression model that explains how many variables (abundance of proteins or metabolites in ecotype) can be reduced to a minimal set needed to predict an outcome (natural variation of ecotype respiration rates).

From the previous reverse genetic studies of mitochondrial MDH in Arabidopsis ecotype Columbia, we know that respiration rate is significantly elevated with the complete absence of expression of both mitochondrial MDH genes (Tomaz et al., 2010). I have demonstrated that MDH (At1g53240) has a significantly different protein expression profile between two distinct sets of ecotypes based on respiration rates and also that MDH gave the most pronounced negative and significant relationship with ecotype respiration rates, t=-3.70 (P=0.001) in multiple regression analysis. This implies that MDH (At1g53240) represents a negative driver in the governance of the natural variation of ecotype respiration rates.

A reason why ecotypes respire faster when their MDH protein abundance is low might be to increase the shuttling of malate-oxaloacetate between mitochondrial and cytosol; exporting malate into the cytoplasm in exchange for oxaloacetate. Our findings in Arabidopsis ecotypes investigating the relationship between natural variation of respiration rates and their MDH protein expression level are consistent with our hypothesis and the previous findings in MDH genes knockout studies by Tomaz et al. (2010). Furthermore, other studies reported that MDH demonstrated the greatest metabolic control coefficient flux for plant respiration compared to other TCA cycle enzymes (Araujo et al., 2012). Taken together the previous findings in mitochondrial MDH knockout studies and this current evidence from Arabidopsis

247

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

ecotypes, there is now sufficient evidence to confirm an important role of MDH in modulating leaf respiration rate in Arabidopsis.

Succinate dehydrogenase subunit 7 (At3g47833) was predicted to have the second strongest negative relationship with ecotype respiration rates (t=-3.117, P=0.005). Unfortunately the knowledge of this plant specific subunit of the SDH enzyme is still in infancy and its participation in the plant respiratory metabolism remains unknown. It would be interesting to further characterise this SDH subunit 7 in view of the correlation of its abundance to leaf respiration rate. The last predicted protein that correlate negatively and significantly with ecotypes respiration rate was CS (At2g44350) with an estimated t value=-2.888 (P=0.008). Citrate synthase catalyses the condensation reaction of oxaloacetate and acetyl-CoA to form citrate which is the entry point of carbon into the TCA cycle. Previously, mitochondrial citrate synthase gene knockdown studies in tomato which led to a mild reduction of the total cellular activity of this enzyme and has been linked to an increased rate of respiration (Sienkiewicz-Porzucek et al., 2008). However, the levels of other TCA cycle intermediates were found to be decreased slightly in citrate synthase knockdown plants and they had impaired nitrogen assimilation. Another knockdown study of mitochondrial citrate synthase gene in green algae, which is a homolog to mitochondrial CS (At2g44350) in Arabidopsis, led to a significant increase in triacylglycerols (TAGs) in the cell (Deng et al., 2013). The only significant positive relationship predicted by multiple regression analysis was found between respiration rate and OGDC complexes (At4g26910&At5g55070) where t=2.409 (P=0.024). These genes encode for dihydrolipoamide succinyltransferases 2- oxoglutarate dehydrogenase complex E2-2 and E2-1 component respectively. Previous reports of inhibition of 2-oxoglutarate dehydrogenase (OGDH) suggested that this enzyme had a rate-limiting effect on respiration in potato tuber (Araujo et al., 2008).

248

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

In examining the predicted relationships between metabolite and respiration rates among Arabidopsis ecotypes, it is important to note that a single metabolite could be involved in multiple pathways and the levels of each metabolite are determined by interactions between many enzymes. Moreover, the natural genetic variations of Arabidopsis ecotypes contain allelic diversity for large numbers of genes and hence resulted in alteration in some of the enzymes activities that account for metabolites changes. These features make inferences about the relationships predicted between metabolites and respiration rate among ecotypes complicated. According to the results of multiple regression analysis, the predicted best model contains metabolites that collectively correlate with natural variation of respiration rate in Arabidopsis ecotypes. Among those, serine showed the highest regression t-value (t=-10.727, P<0.001). Intriguingly, serine level is predicted to be decreased in ecotypes that respire at higher rates. Serine is the product of glycine decarboxylase which is the main respiratory dehydrogenase in the light for photorespiration. Even though the measurements here were made of dark respiration rate, metabolites were sampled from lighted plants during the day, so a photorespiratory trait of low serine could be linked to general mitochondrial function observed in the dark respiration assays.

Conclusion

The wide geographical distribution of Arabidopsis thaliana ecotypes has exposed this plant species to diverse environmental growing conditions and stresses. Individual Arabidopsis ecotypes have each developed their own fitness strategy by altering physiological processes and metabolic contents in order to be survived and compete in their unique habitat. In this study, we exploited the natural genetic variation among 49 Arabidopsis ecotypes to identifying 10 ecotypes with significantly low and high leaf respiration rates and further investigated the respiratory metabolism of those 10 ecotypes using studies encompassing

249

Chapter 5. Investigating the correlation of mMDH and related enzyme networks with natural variation in Arabidopsis thaliana respiratory rates

transcriptomics, targeted proteomics and metabolomics. Correlation analysis enabled us to identify set of components that correlate positively or negatively to each other and to respiratory rate. It is worth highlighting that among the omics studies which had been performed in this study, the targeted proteomics approach of measuring relative TCA cycle protein abundances across was the most successful in capturing biological information linked to respiratory rate. Both Student’s t-Test analysis between distinct ecotype respiratory groups and multiple regression analysis of protein abundance data of TCA cycle enzymes had uncovered potentially important regulatory components of respiratory metabolism among Arabidopsis ecotypes. It was consistently shown that protein abundance of mitochondrial MDH (At1g53240) significantly decreased, when respiration rate increased among Arabidopsis ecotypes. This negative relationship between MDH protein expression and leaf respiration rate in Arabidopsis ecotypes agrees with the previous findings that knockout of both mitochondrial MDH genes gave significantly high leaf respiration rate. Reverse genetics uses laboratory-induced mutations approach to characterise the function of a targeted gene while natural genetic variation offers a powerful tool to estimate the relative fitness of phenotypic and allelic variants of complex traits among wild accessions. Hence, the reproducible findings from both reverse genetics and natural genetic variation studies strengthen the proposition that mitochondrial MDH (At1g53240) has a pronounced role as a TCA cycle enzyme in the rate of leaf respiratory metabolism in Arabidopsis thaliana.

250

Chapter 6:

General Discussion

251

Chapter 6. General Discussion

This thesis provides new insights into the role of mitochondrial malate dehydrogenase on the respiratory metabolism of Arabidopsis thaliana at various developmental stages. The functional roles of mitochondrial MDH in plant respiration were investigated from two different perspectives. The first approach used T-DNA knockout lines for mMDH genes in the commonly used laboratory ecotype named Columbia. This approach provides a straight forward genetic cause and effect study, i.e. changes in respiratory rate and metabolism were studied in the presence or absence of a certain mMDH gene isoform in a common genetic background. This analysis of function has been carried out spatially and temporary, i.e. in different Arabidopsis organs (e.g. leaf, seed and root) and at various developmental stages (e.g. old, mature and young leaf and seed stages) (Chapter 3 and 4). The second approach investigated the correlation of altered mMDH gene expression levels with naturally varied leaf respiratory rates and metabolism from a broader set of 49 Arabidopsis thaliana ecotypes (Chapter 5). This aimed to determine if natural variation in mMDH abundance was consistent with its control of respiration rate observed in genetic manipulation experiments. The ecotype Columbia was one of the selected Arabidopsis ecotypes for the ease of data comparison between the two approaches.

Alongside these biological aims, I have established a multiplexed micro-respiratory assay method for Arabidopsis by adapting a commercial technology used for mammalian cell respiration analysis (Chapter Two). This modified method has facilitated detailed respiratory studies of the small-sized tissues of Arabidopsis and importantly made respiratory measurement on single seeds from this model plant possible. I have demonstrated the success of this method in identifying differential effects on respiration in mmdh mutants using a single leaf disc (Chapter 3), root tip and expanded region of seedlings as well as maturing seeds (Chapter 4). This general discussion integrates the key results from Chapter 3 to 5 in order to outline a conceptual framework to understand the role of mitochondrial malate dehydrogenase in regulating respiration rates and metabolic readjustment spatially and temporally during Arabidopsis development. I highlight the functions of mMDH

252

Chapter 6. General Discussion in growth and development of Arabidopsis plants, inferred from findings on physiological and molecular aspects of central metabolism in 6 sections as outlined below:

6.1 Mitochondrial malate dehydrogenase as a significant regulator of plant respiration

A hallmark of mMDH genes deficiency in Arabidopsis is a significantly elevated respiration rate in autotrophic (leaves) and heterotrophic (seeds and roots) organs in Arabidopsis at all stages of development tested to date. These findings are in agreement with previously reported data on significantly elevated leaf respiration rate in the mMDH double mutant (Tomaz et al., 2010) and evidence that mMDH has the highest control coefficient in the TCA cycle for respiratory flux in tomato (Araujo et al., 2012). From an ecological perspective, naturally varied leaf respiration rates of 49 A. thaliana ecotypes showed a significant negative correlation with mMDH protein expression levels. The relative relationship between MMDH1 protein abundance and respiration rate of ecotypes was found to be the strongest (positive or negative) among all the TCA cycle enzymes. Taking the findings of both mutant and ecotype studies together it is clear that the rates of respiration are consistently inversely related to the mMDH protein levels in Arabidopsis. It can be concluded that mMDHs have a promising role in modulating plant respiration rate.

6.2 Relative contribution of the two mMDH genes in respiratory metabolism

Quantitative analysis of leaf transcript data from Arabidopsis wild type and ecotype plants posited that MMDH1 (At1g53240) is the predominantly expressed gene isoform in mature leaves of Arabidopsis compared to its counterpart, MMDH2 (At3g15020). This is consistent with the previous findings on Arabidopsis microarray gene expression (Schmid et al., 2005). It is also the case at the protein level, MMDH1 protein gave a comparatively good signal to noise response in the MRM assay in contrast to low protein levels of MMDH2 among Arabidopsis ecotypes,

253

Chapter 6. General Discussion consistent with MMDH1 protein being the major protein identified in the Arabidopsis leaf mitochondrial proteome (Millar et al., 2001; Lee et al., 2008; Taylor et al., 2011). This finding is also in line with an earlier report on mMDH mutants where MMDH1 showed a greater contribution to the total MDH activity in leaf protein extracts than MMDH2 (Tomaz et al., 2010). Therefore it is consistently evident that MMDH1 has a relatively greater contribution than MMDH2 to leaf respiratory metabolism.

6.3 Functional roles of mMDHs in plant respiratory metabolism

Notwithstanding there are different genes and protein expression levels between mMDH isoforms, in some circumstances these isoforms may demonstrate functional redundancy. For instance, single mMDH mutants demonstrated wild- type like leaf respiration rates and similar seed germination phenotypes. In these scenarios, the effects of lacking one functional mMDH were seemingly masked by the function of its gene counterpart. Functional redundancy appeared to be one of the common features observed in studies of gene isoforms in plants as well as in other organisms (Bayer et al., 1997; Konopka and Bednarek, 2008; Blanchoin and Staiger, 2010). However, mMDH isoforms may not always functionally complement each other during certain plant development stages. This is inferred when the sole expression of MMDH1 cDNA failed to fully restore the respiration rates of mature and young leaves (albeit the gene expression was driven by a double 35S promoter), indicating a potential role of MMDH2, or possibly the promoter characteristics of MMDH1, in developmental-dependent respiratory metabolism. In addition, only expressing MMDH1 by 35S in the complemented line did not significantly restore some metabolites compared to wild type in maturing seeds, as well as only yielding a partial restoration of wild type-like seed ageing behaviour. More studies of the mmdh2-1 plant line under these conditions will be required to determine if isoform or promoter is the defining feature of the effects observed.

A functional specificity between mMDH isoforms might also be inferred from a spatial gene expression pattern of mMDH isoforms in Arabidopsis root parts

254

Chapter 6. General Discussion

(Birnbaum et al., 2003). The gradual decreasing respiration rate from the root tips to expanded regions in mmdh1-2 and mmdh1-2mmdh2-1 mutants (Figure 13 of Chapter 4) correlate with the relatively higher MMDH1 gene expression observed at the tip region (Birnbaum et al., 2003), indicating a greater energy demand for actively growing cells at the root tips could be linked to MMDH1. In contrast, MMDH2 may be more involved in maintenance respiration which is needed to keep the existing root cells viable in the mature root zones.

6.4 Relationships between mMDH and NAD-MDH isoforms in other subcellular compartments

The simple interpretation of mMDH mutant analysis refers to a cause-and-effect relationship between MDH isoforms and Arabidopsis phenotypes, and correlating the innate relationship between MDH gene isoforms in Arabidopsis ecotypes with respiration, ignores the possible effect of MDHs working across compartments as a network. A collaborative MDH gene network may exist based on the analysis of both mMDH mutants and ecotypes transcript data. Cooperative relationships between MDH isoforms would be aided by the versatile feature of malate and oxaloacetate in plant metabolism, namely that both substrates can be easily transported across different cellular membranes and various compartments via the malate-OAA shuttle (Heber, 1974; Lance and Rustin, 1984). Mitochondrial MDH isoforms showed the closest relationship in the provision of common enzymatic properties allowing functional substitution between these two isoforms. Cytosolic MDHs appeared to be the second cluster of isoforms which correlate well with mitochondrial MDHs, followed by chloroplastic and peroxisomal isoforms of MDHs (Table 2 of Chapter 5). In many cases as evident from other gene expression studies, gene isoforms exhibit functional redundancy and compensation within their metabolic network (Lisenbee et al., 2005; Yi et al., 2006; Konopka and Bednarek, 2008; Schmitz et al., 2009). It has been suggested that the degree of redundancy represents the degree of correlation among genes contributing to a single function (i.e. a set of gene isoforms) (Deisboeck and Kresh, 2007). In the case

255

Chapter 6. General Discussion of MDH, the major forms of MDHs such as MMDH1, CMDH1, CHMDH and PMDH2 showed higher correlations to each other, implying that the relativeness participation of MDH within the network may determine the degree of their relationships (Table 2 of Chapter 5).

It is anticipated that in the circumstances such as the complete loss of mMDH protein (e.g. mmdh1-2mmdh2-1) and lower mMDH transcript levels (in ecotypes), the deficiency of mitochondrial MDH enzyme activity is likely to result in malate accumulation and a consequent disequilibrium of NADH/NAD+ ratio in the mitochondrial matrix. Both NAD(H) and NADP(H) are known as central mediators of reductant levels in metabolic processes and NADH is critical to the catabolic process of respiration (Rasmusson and Wallstrom, 2010). It is thus crucial to maintain a cellular redox balance by equilibrating the malate/oxaloacetate ratios in the mitochondrial matrix and the cytosol (rather than a secondary equilibration mechanism achieved via the malate-aspartate shuttle) (Kromer and Heldt, 1991). This could plausibly explain a coordinated increase in transcript levels of neighbouring NAD-MDH genes (CMDH, CHMDHMDH and P ) to readjust the NAD+ pool in their respective compartments and to regain a balanced global cellular poise and maintain organic acid metabolism (Figure 1). This idea is in line with a cooperative strategy reported earlier between NAD-MDH and NADP-MDH isoforms in performing cellular redox homeostasis (Hebbelmann et al., 2012; Taniguchi and Miyake, 2012). Overall, MDH transcript data presented in this thesis is consistent with the malate-OAA shuttling system adapting to mMDH abundance changes and depict a defined relationship between NAD-MDH isoform levels in Arabidopsis mutants and ecotypes.

256

Chapter 6. General Discussion

Figure 1: Schematic representation of cooperative interaction between NAD-MDH gene isoforms when mitochondrial MDH isoforms (MMDH) are absent (in mMDH double mutant) or MMDH transcripts are at low abundance (in ecotypes). It is anticipated that the malate-oxaloacetate (malate-OAA) translocator in mitochondria transports malate out of mitochondrial matrix concomitantly with an influx of OAA into the matrix, and this elicits increases in transcript abundances of cytosolic MDH (CMDH), chloroplastic MDH (CHMDH) and peroxisomal MDH (PMDH) localized in other subcellular compartments (i.e. cytosol, chloroplast and peroxisome) to re-establish cellular redox balance. MDH genes are also involved in photorespiration by recycling NADH and to provide NAD+ for the conversion of glycine to serine catalysed via the glycine decarboxylase (GDC) by MMDH, and providing NADH for the conversion of hydroxypyruvate to glycerate via hydroxypyruvate reductase by PMDH.

6.5 Relationships between TCA cycle enzyme levels in leaf respiratory metabolism in response to perturbed and naturally altered mMDH protein levels

There appeared to be common coordinated changes in abundance between TCA cycle enzymes in response to changes in mMDH protein abundances, as deciphered from MRM data of both mMDH mutants and ecotypes. Following perturbed or altered mMDH protein levels, shuttling of excess malate from mitochondrial matrix

257

Chapter 6. General Discussion to cytosol in exchange with oxaloacetate is likely to increase in order to persevere a respiratory flux through the TCA cycle. This would represent an alternative route to circumvent the lack of mitochondrial malate dehydrogenase activity. This is associated with significantly high leaf respiration rates and increased abundance in a number of TCA cycle enzymes in both the mMDH double mutant and fast- respiring ecotypes. mMDH was found to have the highest flux control coefficient (1.76) for respiration among all other TCA cycle enzymes in tomato, as compared to CS (-0.4), ACO (0.964), IDH (-0.123), SUC (0.0008), SDH (0.289) and fumarase (0.601) (Araujo et al., 2012). Thus it is plausible that mMDH has a major role in maintaining the NAD+/NADH ratio and mitochondrial redox homeostasis.

However, despite a similar link to increasing respiration rate, there were also some contrasting TCA cycle protein abundance patterns between mMDH mutants and ecotypes. There was a uniform increase in protein abundances for the whole cluster of IDHs and SDHs in mMDH double mutant plants in the Columbia background. While a more complex profile of TCA cycle enzyme abundances were observed across Arabidopsis ecotypes tested. Statistical analyses revealed that only certain isoforms of MDH, SDH, CS and OGDC are predicted to have significant relationships with ecotype respiration rates. Our data showed that only SDH1-1 and SDH1-2 proteins (At5g66760&At2g18450) are commonly and significantly increased in mMDH mutants (mmdh1-2 and mmdh1-2mmdh2-1) and ecotypes, in both cases these proteins correlated inversely with mMDH levels. The protein sequence of SDH1 is highly conserved interspecies although some differences exists across kingdoms (Burger et al., 1996; Huang et al., 2010). SDH1-1 plays a key role in stimulating the formation of mitochondrial H202 and ROS leading to an antifungal and antibacterial defence system in stressed plants (Gleason et al., 2011). High respiration rate has been shown to increase ROS production under the constraint of electron transport chain reduction levels, which could lead to oxidative damage in cells (Tiwari et al., 2002). Therefore, collectively, it can be concluded that the abundance of SDH1 could be a prominent indicator for plant respiration rates.

258

Chapter 6. General Discussion

An evolutionary study on TCA cycle genes suggested that the enzymes of the final steps of TCA cycle from succinate to oxaloacetate are the most highly conserved in sequence, whereas genes for initial steps from acetyl-CoA to 2-oxoglutarate have the least conservation in sequence across many different genomes (Huynen et al., 1999). This indicates a higher frequency of genetic variations among enzymes involved in the first half of TCA cycle. Sequence variation thus shows a pattern similar to protein abundance, with a more synchronised increasing abundance of SDH1 proteins among fast-respiring ecotypes, while a more random abundance profile for IDH and ODGH proteins across ecotypes. Aside from genetic variation in promoter activity, protein synthesis or protein stability as possible factors affecting ecotype TCA cycle proteome, one could also consider the existence of epigenetic variation among ecotypes. Epigenetics can produce differential transcriptional outcomes with no alteration on genomic sequence (Hitchler and Domann, 2007), and it thus represents a crucial mechanism which integrates genetic and environmental stimuli and translates them into phenotypic outcomes (Cyr and Domann, 2011). The TCA cycle is posited to broadly impact epigenetic events through its intermediates including NAD+, S-adenosylmethionine and 2- oxoglutarate which are all associated with epigenetic mechanisms and signalling, that can result in measurable changes in gene expression (Cyr and Domann, 2011). Wild ecotypes of Arabidopsis perceive a wide range of environmental stress in their native habitats, which in turn could give rise to unique phenotypes or ecotypic variability in respiratory metabolism via epigenetic events. The modified traits could be passed to the following generations (transgenerational epigenetics) as inherited epigenetic effects on phenotypes have been documented in various organisms including plants (Jablonka and Raz, 2009; Whittle et al., 2009). Further investigation is now required to identify the phenotypic variations among ecotypes and whether they can be attributed to genetic differences or epi-alleles or the combination of both.

259

Chapter 6. General Discussion

6.6 Multiple roles of mMDH in energy metabolism and homeostasis contribute towards energy efficiency during plant development

6.6.1 Altered metabolite profiles during photorespiration in Arabidopsis plants lacking mMDH

There were large increases of leaf glycine content in both the mMDH double mutant (27-fold higher than wild type level) (Tomaz et al., 2010) and fast-respiring Arabidopsis ecotype (approximately 10-fold higher than the slow-respiring ones) from the data presented in this thesis. Mitochondrial MDH is thought to cooperate with GDC and play an important role in the photorespiration process, by recycling of NADH formed in the GDC reaction during the OAA to malate reaction of MDH in the light (Figure 1) (Wiskich and Dry, 1985; Wiskich et al., 1990). The large accumulation of glycine could be explained by the deficiency or loss of mMDH activity and account for a limited photorespiratory cycle. This notion is supported by a lower postillumination burst, alterations in CO2 assimilation/intercellular CO2 curves at low CO2, and the light-dependent elevated concentration of photorespiratory metabolites (Tomaz et al., 2010).

6.6.2 Impact of mMDH abundance on plant growth rate and carbon metabolism

Fast-respiring ecotypes and mmdh1-2mmdh2-1 mutants varied in their carbon metabolism due to their different genetic architecture notwithstanding their significantly high respiration rates and lower levels of mMDH proteins. The mMDH double mutant showed a slow growth and reduction in biomass across different organs (leave, seed and root) and throughout plant developmental stages. In addition, the mMDH double mutant plants showed delayed flowering and produced shorter siliques and significantly lower seed yield compared to wild type. On the contrary, there was no specific trend of phenotype changes (in terms of the leaf size and rosette diameter) between slow and fast-respiring ecotypes.

260

Chapter 6. General Discussion

The levels of galactose and glucose were found to be significantly lower in 30-mins darkened leaves of mmdh1-2mmdh2-1 (Tomaz, 2012). This might be due to a rapid consumption of hexoses to support its significantly high respiration rate after the transition from light to dark. In contrast, a significantly higher abundance of hexoses i.e. galactose, glucose and fructose were found in fast-respiring ecotypes compared to the slower ones. Intriguingly, higher sugar content in fast-respiring ecotypes might be an adaptive strategy in providing more carbon skeleton and energy to maintain their high respiration activities. It has been reported that higher sugar content could contribute directly to a higher rate of respiration in wheat leaves (Azconbieto et al., 1983). Notably, the fast respiring ecotype 25 (N14) contains exceptionally high level of hexoses (11-fold higher than other fast-respiring ecotypes), seemingly a unique sugar profile which would be worth exploring further.

The phenotype of the mMDH double mutant is believed to be due to a defective carbon metabolism upon the disruption of both mMDH functional genes, supported by a substantial reduction of net carbon assimilation as much as 32% in leaves of mmdh1-2mmdh2-1 compared to wild type (Tomaz et al., 2010). The fast respiring activity of the mMDH double mutant is believed to direct away the energy and carbon skeletons needed for various plant biosynthesis processes leading to decreased carbon assimilation in this mutant. While leaf respiration rate appears to be only a fraction of the photosynthetic rate in leaves in the light, the losses by respiration in all tissues during both day and night mean that over a plant’s life cycle, loss of carbon by respiration may be up to 50% of the net carbon gain from photosynthesis (Amthor, 1989). Also, a limitation in photorespiratory rate in the mMDH double mutant could account for a considerable loss of fixed carbon. This is because photorespiration allows the regaining of three moles of carbon (as 3PGA) into the C3 carbon assimilation cycle (Berry et al., 1978) for every single respired

CO2 molecule in the light (Ludwig and Canvin, 1971). There is no data on the carbon assimilation rates of the ecotype in the conditions used in this study, owing to time limitation, but it would be expected that the carbon metabolisms of ecotypes are

261

Chapter 6. General Discussion likely to differ from one to another. It is reasonable to anticipate that some kind of feedback mechanisms could be elicited upon the shortage of carbon in the mmdh1- 2mmdh2-1 mutant, supported by significantly altered protein level of enzymes involved in amino acid catabolic pathway in mMDH knockout lines (Tomaz, 2012). Fast-respiring ecotypes (the top 3 performers), showed an increase in a number of amino acid abundances like valine, isoleucine and threonine. Degradation of branched chain amino acids have been suggested to support respiration in tomato fruits (Kochevenko et al., 2012) and Arabidopsis leaves during senescence (Kleessen et al., 2012). From both experimental approaches using ecotypes and mutants, the engagement of BCAA catabolism is presumably required to provide carbon skeletons and energy, as an alternative source to account for the significant carbon loss via respiratory activities in these plants.

Arabidopsis plants depleted in mMDH are likely to be subjected to some kinds of metabolic stresses, with evidence that AOX protein abundance was 2-fold higher than wild type across all the single and double mMDH mutants (Tomaz, 2012). In view of the increased abundances of IDHs, SUC and SDHs in mutant leaves lacking MMDH1 protein (Figure 6 of Chapter 3), it would be anticipated that a significantly higher level of intra-mitochondrial NADH and FADH2 is maintained. This in turn will result in higher redox poise of components in the cytochrome pathway and a subsequent activation of AOX respiratory pathway (Azcón-Bieto et al., 1983). This also plausibly linked to the significantly high respiration rate in the mMDH double mutant. Tight metabolic controls between the redox status and the central carbon metabolism of cells normally determine the energy use efficiency in plant under certain conditions (Wilson et al., 2006). The role of AOX in modulating carbon use efficiency in plant has been well established in a range of studies, where AOX constitutes a non-energy conserving branch of mETC and therefore decrease the efficiency while carbon is utilised for growth (Vanlerberghe et al., 1997; Sieger et al., 2005). This is reflected in an important role of mMDH in regulating carbon partitioning and determining the carbon use efficiency of the plant. In the ecotypes, there is no evidence for naturally varied AOX transcript or protein expression in

262

Chapter 6. General Discussion

Arabidopsis ecotypes that links to respiration rate. However, partitioning of this protein to the alternative pathway is varied over a seasonal timescales and altitudinal gradients (Searle, 2010), consistent with earlier findings that showed a temperature-dependent protein expression of AOX and increased abundance of AOX with low temperatures (Vanlerberghe and McIntosh, 1992; Gonzàlez-Meler et al., 1999; Mizuno et al., 2008). In conjunction to this, remarkable increases in the GABA accumulation among the top 3 fast-respiring ecotypes raise the possibility that their significantly high respiration rates associated with high levels of AOX protein expression. The increased level of GABA or AOX are known to be metabolic markers upon exposure to oxidative stress (Arnholdt-Schmitt et al., 2006; Obata and Fernie, 2012). These correlations need to be tested directly in both ecotypes and MDH mutants.

6.6.3 A role of mMDH in coordinating nitrogen metabolism

Mitochondrial respiration provides the redox equivalents and carbon skeletons necessary for nitrogen assimilation in the light (Atkin et al., 2000). OGDH has been suggested to play a major role in nitrogen assimilation and the successive amino acid production (Hodges, 2002; Araujo et al., 2008). Nitrogen assimilation is likely to be affected in mMDH single and double mutant leaves as 2-oxoglutarate dehydrogenase subunits (At3g55410 and At5g65750) are significantly lower in abundance compared to wild type. From a DIGE analysis of shoot mitochondria of the mmdh1-2mmdh2-1 mutant, there were also significantly lower abundance of glutamate dehydrogenases (GDH1_At5g18170 and GDH2_At5g07440) (Tomaz, 2012). The mitochondrial GDH catalyses the reductive amination of 2-oxoglutarate to glutamate and the oxidative deamination of glutamate to 2-oxoglutarate (Aubert et al., 2001). A loss-of-function study on GDH genes showed an increased susceptibility of the gdh double mutant to dark-induced carbon starvation, suggesting a role of GDH as a gatekeeper for providing alternative carbon source to the respiratory pathway via amino acid catabolism during sugar starvation (Miyashita and Good, 2008). The above appeared to correspond to the observation

263

Chapter 6. General Discussion of increased abundance of free amino acids like valine, alanine and beta-alanine in mmdh1-2mmdh2-1 (Tomaz, 2012). Furthermore it is possible that re-assimilation of ammonia into glutamine (known as photorespiratory nitrogen cycle) (Keys et al., 1978; Keys, 2006) would be altered due to defective photorespiration in mmdh1- 2mmdh2-1. Photorespiratory mutants originally isolated in a wide range of plant species demonstrate a metabolic defect related to the photorespiratory nitrogen cycle (Somerville, 1986; Leegood et al., 1995; Wingler et al., 2000).

As mentioned above, there was a differential amino acids content between slow and fast-respiring Arabidopsis ecotypes. For example, branched-chain amino acids (valine and isoleucine) and other amino acids (threonine, serine and glycine) showed substantially higher accumulation in fast-respiring ecotypes. A large number of those amino acids (valine, isoleucine, serine and threonine) were found by multiple linear regression analysis to be sufficient to predict ecotypes respiration rate. Therefore there is a preliminary case to propose that fast-respiring Arabidopsis ecotypes (particularly the top 3 performers) could behave similarly to mMDH mutants by catabolising amino acids to counterbalance the loss of carbon skeletons during their high respiratory activities. Catabolism of branched chain amino acids such as isoleucine, valine and leucine has been reported to be associated with carbon starvation in plant under stress conditions or extended darkness (Ishizaki et al., 2005; Araujo et al., 2011; Kleessen et al., 2012). The nitrogen metabolism of the current set of Arabidopsis ecotypes should be investigated top test this link, there is substantial evidence from Arabidopsis ecotype research of diversity of amino acids (Gördes et al., 2013), genetically varied nitrogen use efficiency traits (Loudet et al., 2003), differential responses to nitrogen supply (Ikram et al., 2011), nitrate uptake and nitrogen use efficiency under varying nitrogen supply (Chardon et al., 2010).

264

Chapter 6. General Discussion

6.6.4 A role for mMDH in ensuring energy reserves in heterotrophic organs of Arabidopsis

The loss of mMDH proteins has negatively impacted the energy balance in respect to the sink-to-source relationship in the plant. Low seed production, arrested early embryogenesis (torpedo-shaped embryos) and impaired seed filling process (reduced seed dry weight and size) were most likely to be caused by an insufficient maternal nutrient supply during seed growth and development in mmdh1- 2mmdh2-1. It has been shown that maternal ovule integument and filial compartments (endosperm and embryo) are dynamically interacting with each other during seed development, and seed size of Arabidopsis thaliana is regulated by both seed coat cell elongation under maternal control and endosperm growth which is under zygotic control (Garcia et al., 2005). Impaired seed production and low seed quality has previously been seen in Arabidopsis transgenic plants with mutated genes involved in carbon and nitrogen metabolism (Andriotis et al., 2012; Guan et al., 2015). Moreover, the developing mmdh1-2mmdh2-1 seeds are photosynthetically less competent as depicted from their whitish to pale green seed phenotype, which may results in a more internal hypoxic condition compared to wild type. Embryogenic photosynthesis can help green developing seeds to be relieved from anoxia during the day where a limited amount of oxygen can be diffused from the outside due to low gas permeability of seed integuments (Borisjuk and Rolletschek, 2009). Presumably, a lower photosynthetic capacity could result in insufficient reserve accumulation in mmdh1-2mmdh2-1 seeds as photosynthesis is a significant ATP provider in maturing seeds and the costs of protein and fatty acid synthesis are high (Ruuska et al., 2004; Goffman et al., 2005). Indeed, impaired embryonic photosynthesis has been reported to account for the lack of accumulation of key metabolites and storage compounds, and thus a slower seed germination (Westoby et al., 1992; Allorent et al., 2015).

When the oxygen concentration falls due to demand outstripping supply rate, plants can down-regulate a suite of energy and inhibit oxygen consuming metabolic pathways such as mitochondrial respiration to avoid internal anoxia (Geigenberger,

265

Chapter 6. General Discussion

2003; Gupta et al., 2009). This is consistent with significantly lower abundances of TCA cycle intermediates throughout the maturation stages in mmdh1-2mmdh2-1 seeds as compared to wild type. It is anticipated that this would result in pyruvate accumulation in the cytosol, which is supported by an increase of alanine content in mmdh1-2mmdh2-1 seeds. Under limited oxygen concentration, a high enzymatic rate of alanine aminotransferase was evident and excess pyruvate is converted to alanine (Loreti et al., 2005).

As a results of both insufficient energy conservation and high respiratory activities as mentioned above, it is mostly likely that seeds of mmdh1-2mmdh2-1 mutant are subjected to both an energy deficit and carbon deprivation. Plant cells utilise alternative metabolic pathways such as glycolytic pathway, protein degradation and catabolism of amino acids to produce ATP as a metabolic adaptation under oxidative stress conditions (Baxter et al., 2007). This was manifested in mmdh1- 2mmdh2-1 mutant seeds where marked increase in abundances of glycolysis intermediates, and branched chain amino acids (valine, leucine and isoleucine) in mmdh1-2mmdh2-1 seeds, was consistent throughout the maturation stages. This is in good agreement with increased valine and altered BCAA metabolism in mmdh1- 2mmdh2-1 leaves as documented earlier in Tomaz (2012). Although water deprivation is encountered in wild type Arabidopsis seeds at the late maturation stages, the loss of mMDH may increase the susceptibility of mmdh1-2mmdh2-1 seeds to osmotic stress. This could be inferred from significantly increased level of an array of osmolytes i.e. proline, sorbitol, mannitol, glycine and serine (Hare et al., 1998) in double mutant seeds compared to wild type starting from the early maturation stage. This is also coincident with an earlier onset of seed desiccation and noticeable shrivelled seed phenotype in double mutant relative to wild type at mid-phase of maturation. This early termination of the maturation stages could be a metabolic adaptation strategy for mmdh1-2mmdh2-1 seeds to enter a quiescent stage for optimum energy reservation.

The above phenomenon collectively attributed to a significantly shortened shelf-life and longevity of mmdh1-2mmdh2-1 mutant seeds (4 times shorter than wild type).

266

Chapter 6. General Discussion

A broad range of environmental stresses like cold, salinity, drought and highlight could lead to an elevated production of ROS (Maurino and Flügge, 2008). Therefore the oxidative and osmotic stresses could increase the ROS level in mmdh1- 2mmdh2-1 seeds, which might lead to lipid peroxidation, particularly during seed storage. Damages to the integrity of seed coat by lipid peroxidation would account for a lower seed germination rate as well as affecting the vigour and longevity potential of seeds (Souza and Marcos-Filho, 2001). The insufficient storage reserve in mmdh1-2mmdh2-1 mutant seeds may limit the reserve mobilization to fuel metabolic resumption during seed germination and the successive post-germinative growth. The transcript and protein levels of mMDH increased markedly upon the metabolic resumption in imbibed seeds (Sugimoto and Morohashi, 1989; Narsai et al., 2011), an indicative of its essential role during seed germination. High respiration rates persist through the post-germinative growth of mmdh1-2mmdh2- 1 mutant, notably for the root tip of the primary root as well as in the whole root system. It have been reported that under low nutrient supply and ambient CO2 concentration conditions, carbon demand for root respiration (expressed as a fraction of the daily assimilation) would be higher than that under high nutrient supply (Lambers et al., 1996). Thus, despite a deficiency of storage reserve in mmdh1-2mmdh2-1 mutant seeds, a large carbon pool is anticipated to be devoted for sustaining their high root oxygen uptake rate. Retarded root growth, slowed root elongation and significantly shorter root length of mmdh1-2mmdh2-1 mutant than wild type symbolise a state of root carbon and energy deprivation in this double mutant. Roots of transgenic tomato with decreased mMDH activity also demonstrated similar phenotypes, notably a stunted root growth and reduced root area (Van der Merwe et al., 2009). Therefore, it is evident again the importance of mMDH for seedling establishment, root growth and development.

From both mMDH mutants and ecotypes data, it was evident that mitochondrial MDH genes and cytosolic MDH genes are correlated and work in a coordinated manner. The relationship between mitochondrial and cytosolic MDH has been studied before in roots subjected to aluminium stress. A body of literature

267

Chapter 6. General Discussion suggested that malate exudates at the root tips confers aluminium tolerance in soil (Delhaize et al., 1993; Ryan et al., 1995; Tang et al., 2002). While an increase in cytosolic MDH (homolog to Arabidopsis CMDH1 gene) have shown to enhance malate exudates in root apices of transgenic alfafa (Tesfaye et al., 2001). However, more recently, mitochondrial MDH mRNAs abundance was found to decrease in response to aluminium stress in rye roots (Abd El-Moneim et al., 2015). This repression was believed to be a feedback mechanism of increased root exudation of organic acids such as malate in rye roots, consistent with increased malate abundance in root exudates of antisense-repressed mitochondrial MDH transgenic tomato (Van der Merwe et al., 2009). Taken together, the above findings infer an inverse relationship between mitochondrial and cytosolic MDH, resembling the compensatory strategy exhibited in mMDH mutants. Also, considering a high level of gene expression of MMDH1 at the root apices of Arabidopsis (as discussed earlier in section 6.2), the MMDH1 gene would be anticipated to have some potential role in aluminium tolerance of the plant root system which is worthy of further investigation. The major changes in the developmental stages of mmdh1-2mmdh2- 1 mutant in comparison to wild type plant are summarised in Figure 2.

268

Chapter 6. General Discussion

Figure 2. Schematic representation of major alteration of developmental stages of Arabidopsis mmdh1-2mmdh2-1 mutant relative to the wild type plant.

269

Future direction

The data presented in this thesis comprises a wide study on mitochondrial MDHs in Arabidopsis using T-DNA knockout and natural variation studies as analysis tools. In this study one of the key observations is the complementation with a sole MMDH1 DNA into mmdh1-2mmdh2-1 double mutant with expression driven by a double 35S promoter did not always restore the wild type phenotype. This could signify unknown roles of the MMDH2 gene during the growth and development of Arabidopsis plants particularly in heterotrophic organs (seeds and roots). It would be exciting to conduct a similar experiment in order to investigate the functional diversity of the MMDH2 gene. This could be done by using a complementation line with a only MMDH2 cDNA into the double mutant background. The expression of complemented MMDH2 cDNA could be driven by a double 35S promoter, similar to the approach used for MMDH1 cDNA in the present study. Ideally, a native promoter of individual mMDH isoforms could be used to drive the optimal expression of MMDH1 or MMDH2 cDNA in their respective complemented line. The outcome of the study could lead us to a better understanding of the functional role of MMDH2 gene in respiratory metabolism in those specific tissues.

Based on the findings of mMDH mutants and ecotypes as described in the earlier chapters, there is a promising relationship between MDH isoforms in exerting compensatory mechanisms under externally manipulated genotypes i.e. loss-of- function mMDH mutants and genotypes with naturally varied mMDH gene expression. The eight different MDH isoforms are located in different subcellular compartments and seemingly have unique functional roles in regulating plant growth and development, despite the use of a shuttling mechanism like malate- OAA and malate valves that enable cellular redox homeostasis between compartments. For instance, the plastidial isoform of NAD-MDH is associated with autotrophic metabolism, indicated by lower chlorophyll content, photosynthetic rate, and daytime carbohydrate levels, and disordered chloroplast ultrastructure, particularly in developing leaves in a microRNA suppression study (Beeler et al.,

270

Future direction

2014). In contrast, the plant peroxisomal MDH isoform is mainly involved in fatty acid β-oxidation (Pracharoenwattana et al., 2007), the glyoxylate cycle (Kunze et al., 2006) and supplying reductants for the hydroxypyruvate reductase (HPR1) reaction in leaves (Hu et al., 2012). Nevertheless, the data presented in this thesis revealed a multifaceted role of mitochondrial MDH in mitochondrial respiratory metabolism, photorespiration, carbon and nitrogen metabolism of Arabidopsis. It would be interesting to elucidate a more defined relationship between these MDH isoforms in mMDH mutants as well as ecotype plants under some laboratory controlled growing conditions such as abiotic stresses. The changes of MDH isoform expression could be assessed in terms of their transcript levels or their protein abundances using the MRM approach.

Acid soil is an unresolved problem for world crop production, as an approximately 30% of the world’s total land area are acidic, and over 50% of the world’s potential arable lands are acidic (Vonuexkull and Mutert, 1995). A significant amount of research have been undertaken to improve crop yield on acidic soil caused by high aluminium content, including over-expressing a nodule MDH in transgenic alfafa root (homologue to Arabidopsis cytosolic MDH) which has demonstrated positive impacts to confer aluminium toxicity (Tesfaye et al., 2001). But none of the studies has been performed using mitochondrial MDH genes. In view of a plausible connection between the expressions of mMDH genes and the production of root exudates to confer soil aluminium toxicity, it could be interesting to design experiments in this direction. A straight forward approach would be examining the root growth of the single mMDH mutant (mmdh1-2 and mmdh2-1), mMDH double mutant (mmdh1-2mmdh2-1) and complemented line (mmdh1-2mmdh2-1: 35S MMDH1) on agar supplemented with different aluminium concentrations. The underlying mechanism of action and compensatory relationship between mMDH isoforms under such circumstances are worthy to be investigated via omics studies.

Research on Arabidopsis ecotypes is an exciting field, as there is an increasing awareness of how to harness this valuable resource from natural variation of Arabidopsis amongst the plant research community. The findings on A. thaliana

271

Future direction ecotypes as described in this thesis serve as groundwork for further investigation, for instance more targeted experiments could be performed on significantly high and low respiration rate ecotypes. In this regards, drought, high temperature, highlight, cold stress and salinity could be the subject of laboratory induced environmental stress for these ecotypes to see the effect on respiration rates. Subsequent screening for corresponding respiratory traits to specific stresses from these ecotypes could help to untangle the complexity of respiratory modulation event in ecotype adaptation to these stresses. The measurement of TCA cycle enzymes changes could be expanded to enzymatic activities analyses or protein turnover study in addition to assessment of protein abundances under varying growing conditions. It is anticipated that the potential benefit of respiratory traits could be exploited for future agronomy improvement of important crops targeted to specific global climate or environmental conditions.

272

References

Aarssen LW, Clauss MJ (1992) Genotypic variation in fecundity allocation in Arabidopsis thaliana. Journal of Ecology 80: 109-114 Abd El-Moneim D, Contreras R, Silva-Navas J, Gallego FJ, Figueiras AM, Benito C (2015) On the consequences of aluminium stress in rye: repression of two mitochondrial malate dehydrogenase mRNAs. Plant Biology 17: 123-133 Abiko T, Obara M, Ushioda A, Hayakawa T, Hodges M, Yamaya T (2005) Localization of NAD-isocitrate dehydrogenase and glutamate dehydrogenase in rice roots: Candidates for providing carbon skeletons to NADH-glutamate synthase. Plant and Cell Physiology 46: 1724-1734 Agrawal B, Lakshmanan V, Kaushik S, Bais H (2012) Natural variation among Arabidopsis accessions reveals malic acid as a key mediator of nickel (Ni) tolerance. Planta 236: 477-489 Al-Maskri A (2002) Effect of accelerated aging on lipid peroxidation, leakage and seedling vigor (RGR) in cucumber (Cucumis sativus L.) seeds. Pakistan journal of agricultural sciences 39: 330-337 Al-Shehbaz IA, O'Kane SL, Jr. (2002) Taxonomy and phylogeny of Arabidopsis (Brassicaceae). The Arabidopsis book / American Society of Plant Biologists 1: e0001 Alhamdan AM, Alsadon AA, Khalil SO, Allah MA, Nagar ME, Ibrahim AA (2011) Influence of storage conditions on seed quality and longevity of four vegetable crops. American-Eurasian Journal Agricultural & Environmental Science 11 353-359 Allorent G, Osorio S, Ly Vu J, Falconet D, Jouhet J, Kuntz M, Fernie AR, Lerbs‐Mache S, Macherel D, Courtois F (2015) Adjustments of embryonic photosynthetic activity modulate seed fitness in Arabidopsis thaliana. New Phytologist 205: 707-719 Alonso-Blanco C, Aarts MGM, Bentsink L, Keurentjes JJB, Reymond M, Vreugdenhil D, Koornneef M (2009) What has natural variation taught us about plant development, physiology, and adaptation? Plant Cell 21: 1877-1896 Alonso-Blanco C, Koornneef M (2000) Naturally occurring variation in Arabidopsis: an underexploited resource for plant genetics. Trends in Plant Science 5: 22-29 Alscherherman R, Musgrave M, Leopold AC, Khan AA (1981) Respiratory changes with stratification of pear seeds. Physiologia Plantarum 52: 156-160 Altmann R (1890) Die Elementarorganismen und ihre beziehungen zu den zellen. Verlag yon Veit & Comp, Leipzig Amable RA, Obendorf RL (1986) Soybean seed respiration during simulated preharvest deterioration. Journal of Experimental Botany 37: 1364-1375 Amthor JS (1984) The role of maintenance respiration in plant growth. Plant Cell and Environment 7: 561-569 Amthor JS (1989) Respiration and Crop Productivity. Springer-Verlag, New York, NY, USA Andriotis VME, Pike MJ, Schwarz SL, Rawsthorne S, Wang TL, Smith AM (2012) Altered starch turnover in the maternal plant has major effects on Arabidopsis fruit growth and seed composition. Plant Physiology 160: 1175-1186 Araujo WL, Ishizaki K, Nunes-Nesi A, Larson TR, Tohge T, Krahnert I, Witt S, Obata T, Schauer N, Graham IA, Leaver CJ, Fernie AR (2010) Identification of the 2- hydroxyglutarate and isovaleryl-CoA dehydrogenases as alternative electron donors linking lysine catabolism to the electron transport chain of Arabidopsis mitochondria. Plant Cell 22: 1549-1563

273

References

Araujo WL, Nunes-Nesi A, Nikoloski Z, Sweetlove LJ, Fernie AR (2012) Metabolic control and regulation of the tricarboxylic acid cycle in photosynthetic and heterotrophic plant tissues. Plant Cell and Environment 35: 1-21 Araujo WL, Nunes-Nesi A, Trenkamp S, Bunik VI, Fernie AR (2008) Inhibition of 2- oxoglutarate dehydrogenase in potato tuber suggests the enzyme is limiting for respiration and confirms its importance in nitrogen assimilation. Plant Physiology 148: 1782-1796 Araujo WL, Tohge T, Ishizaki K, Leaver CJ, Fernie AR (2011) Protein degradation - an alternative respiratory substrate for stressed plants. Trends in Plant Science 16: 489-498 Araujo WL, Tohge T, Osorio S, Lohse M, Balbo I, Krahnert I, Sienkiewicz-Porzucek A, Usadel B, Nunes-Nesi A, Fernie AR (2012) Antisense inhibition of the 2- oxoglutarate dehydrogenase complex in tomato demonstrates its importance for plant respiration and during leaf senescence and fruit maturation. Plant Cell 24: 2328-2351 Armstrong AF, Logan DC, Atkin OK (2006) On the developmental dependence of leaf respiration: Responses to short- and long-term changes in growth temperature. American Journal of Botany 93: 1633-1639 Armstrong AF, Logan DC, Tobin AK, O'Toole P, Atkin OK (2006b) Heterogeneity of plant mitochondrial responses underpinning respiratory acclimation to the cold in Arabidopsis thaliana leaves. Plant Cell and Environment 29: 940-949 Armstrong W, Cousins D, Armstrong J, Turner DW, Beckett PM (2000) Oxygen distribution in wetland plant roots and permeability barriers to gas-exchange with the rhizosphere: a microelectrode and modelling study with Phragmites australis. Annals of Botany 86: 687-703 Arnaud N, Ravet K, Borlotti A, Touraine B, Boucherez J, Fizames C, Briat JF, Cellier F, Gaymard F (2007) The iron-responsive element (IRE)/ iron-regulatory protein 1 (IRP1)-cytosolic aconitase iron-regulatory switch does not operate in plants. Biochemical Journal 405: 523-531 Arnholdt-Schmitt B, Costa JH, de Melo DF (2006) AOX–a functional marker for efficient cell reprogramming under stress? Trends in plant science 11: 281-287 Artus NN, Naito S, Somerville CR (1994) A mutant of Arabidopsis thaliana that defines a new locus for glycine decarboxylation. Plant and Cell Physiology 35: 879-885 Arvidsson S, Kwasniewski M, Riano-Pachon DM, Mueller-Roeber B (2008) QuantPrime - a flexible tool for reliable high-throughput primer design for quantitative PCR. BMC Bioinformatics 9 Ast C, Schmaelzlin E, Loehmannsroeben HG, van Dongen JT (2012) Optical oxygen micro- and nanosensors for plant applications. Sensors 12: 7015-7032 Atkin OK, Cummins WR, Collier DE (1993) Light induction of alternative pathway capacity in leaf slices of Belgium endive. Plant Cell and Environment 16: 231-235 Atkin OK, Millar AH, Gardeström P, Day DA (2000) Photosynthesis, carbohydrate metabolism and respiration in leaves of higher plants. In R Leegood, T Sharkey, S von Caemmerer, eds, Photosynthesis, Vol 9. Springer Netherlands, pp 153-175 Aubert S, Bligny R, Douce R, Gout E, Ratcliffe RG, Roberts JKM (2001) Contribution of glutamate dehydrogenase to mitochondrial glutamate metabolism studied by 13C and 31P nuclear magnetic resonance. Journal of Experimental Botany 52: 37-45 Audus LJ, Garrard A (1953) Studies on the Growth and Respiration of Roots: I. The effects of stimulatory and inhibitory concentrations of β-indolylacetic acid on root sections of Pisum sativum. Journal of Experimental Botany 4: 330-348

274

References

Azcon-Bieto J, Day DA, Lambers H (1983) The regulation of respiration in the dark in wheat leaf slices. Plant Science Letters 32: 313-320 Azcon-Bieto J, Lambers H, Day DA (1983) Effect of photosynthesis and carbohydrate status on respiratory rates and the involvement of the alternative pathway in leaf respiration. Plant Physiology 72: 598-603 Bai B, Sikron N, Gendler T, Kazachkova Y, Barak S, Grafi G, Khozin-Goldberg I, Fait A (2012) Ecotypic variability in the metabolic response of seeds to diurnal hydration- dehydration cycles and its relationship to seed vigor. Plant and Cell Physiology 53: 38-52 Bai SJ, Ryu WH, Fasching RJ, Grossman AR, Prinz FB (2011) In vivo O2 measurement inside single photosynthetic cells. Biotechnology Letters 33: 1675-1681 Bakker EG, Stahl EA, Toomajian C, Nordborg M, Kreitman M, Bergelson J (2006) Distribution of genetic variation within and among local populations of Arabidopsis thaliana over its species range. Molecular Ecology 15: 1405-1418 Balasubramanian S, Sureshkumar S, Agrawal M, Michael TP, Wessinger C, Maloof JN, Clark R, Warthmann N, Chory J, Weigel D (2006) The PHYTOCHROME C photoreceptor gene mediates natural variation in flowering and growth responses of Arabidopsis thaliana. Nature Genetics 38: 711-715 Barah P, Jayavelu ND, Mundy J, Bones AM (2013) Genome scale transcriptional response diversity among ten ecotypes of Arabidopsis thaliana during heat stress. Frontiers in Plant Science 4 Barah P, Jayavelu ND, Rasmussen S, Nielsen HB, Mundy J, Bones AM (2013) Genome-scale cold stress response regulatory networks in ten Arabidopsis thaliana ecotypes. BMC Genomics 14 Battelli F, Stern L (1912) Die Oxydationsfermente. Ergebnisse der Physiologie 12: 96-268 Baud S, Boutin JP, Miquel M, Lepiniec L, Rochat C (2002) An integrated overview of seed development in Arabidopsis thaliana ecotype WS. Plant Physiology and Biochemistry 40: 151-160 Baud S, Dubreucq B, Miquel M, Rochat C, Lepiniec L (2008) Storage reserve accumulation in Arabidopsis: metabolic and developmental control of seed filling. The Arabidopsis book / American Society of Plant Biologists 6: e0113 Baxter CJ, Redestig H, Schauer N, Repsilber D, Patil KR, Nielsen J, Selbig J, Liu J, Fernie AR, Sweetlove LJ (2007) The metabolic response of heterotrophic Arabidopsis cells to oxidative stress. Plant Physiology 143: 312-325 Bayer CA, vonKalm L, Fristrom JW (1997) Relationships between protein isoforms and genetic functions demonstrate functional redundancy at the Broad-Complex during Drosophila metamorphosis. Developmental Biology 187: 267-282 Beeler S, Liu H-C, Stadler M, Schreier T, Eicke S, Lue W-L, Truernit E, Zeeman SC, Chen J, Koetting O (2014) Plastidial NAD-dependent malate dehydrogenase is critical for embryo development and heterotrophic metabolism in Arabidopsis. Plant Physiology 164: 1175-1190 Beeson CC, Beeson GC, Schnellmann RG (2010) A high-throughput respirometric assay for mitochondrial biogenesis and toxicity. Analytical Biochemistry 404: 75-81 Behal RH, Neal W, Oliver DJ (1996) Biochemical and molecular characterization of NAD(+)- dependent isocitrate dehydrogenase and fumarase from oil seed plants. Plant Physiology 111: 601-601 Benamar A, Tallon C, Macherel D (2003) Membrane integrity and oxidative properties of mitochondria isolated from imbibing pea seeds after priming or accelerated ageing. Seed Science Research 13: 35-45

275

References

Bergelson J, Roux F (2010) Towards identifying genes underlying ecologically relevant traits in Arabidopsis thaliana. Nature Reviews Genetics 11: 867-879 Bernard DG, Cheng Y, Zhao Y, Balk J (2009) An allelic mutant series of ATM3 reveals its key role in the biogenesis of cytosolic iron-sulfur proteins in Arabidopsis Plant Physiology 151: 2187-2187 18 Berry JA, Osmond CB, Lorimer GH (1978) Fixation of O2 during photorespiration kinetic and steady-state studies of the photorespiratory carbon oxidation cycle with intact leaves and isolated chloroplasts of C3 plants. Plant physiology 62: 954-967 Bertani A, Brambilla I (1982) Effect of decreasing oxygen concentration on some aspects of protein and amino acid metabolism in rice roots. Zeitschrift Fur Pflanzenphysiologie 107: 193-200 Bevan M, Walsh S (2005) The Arabidopsis genome: a foundation for plant research. Genome Research 15: 1632-1642 Bewley JD (1997) Seed germination and dormancy. Plant Cell 9: 1055-1066 Bewley JD (2001) Seed germination and reserve mobilization. In eLS. John Wiley & Sons, Ltd Bewley JD, Black M (1978) Physiology and biochemistry of seeds in relation to germination, Vol Vol. 1. Springer Verlag., Berlin Bidel LPR, Renault P, Pages L, Riviere LM (2000) Mapping meristem respiration of Prunus persica (L.) Batsch seedlings: potential respiration of the meristems, O2 diffusional constraints and combined effects on root growth. Journal of Experimental Botany 51: 755-768 Bingham IJ, Merritt GJ (1999) Effects of seed ageing on early post-germination root extension in maize: a spatial and histological analysis of the growth-zone. Seed Science and Technology 27: 151-162 Birnbaum K, Shasha DE, Wang JY, Jung JW, Lambert GM, Galbraith DW, Benfey PN (2003) A gene expression map of the Arabidopsis root. Science 302: 1956-1960 Biswas DK, Jansen MAK (2012) Natural variation in UV-B protection amongst Arabidopsis thaliana accessions. Emirates Journal of Food and Agriculture 24: 621 Blacquiere T, Lambers H (1981) Growth, photosynthesis and respiration in Plantago coronopus as affected by salinity. Physiologia Plantarum 51: 265-268 Blanchoin L, Staiger CJ (2010) Plant formins: diverse isoforms and unique molecular mechanism. Biochimica Et Biophysica Acta-Molecular Cell Research 1803: 201-206 Boex-Fontvieille ERA, Gauthier PPG, Gilard F, Hodges M, Tcherkez GGB (2014) A new anaplerotic respiratory pathway involving lysine biosynthesis in isocitrate dehydrogenase-deficient Arabidopsis mutants (vol 199, pg 673, 2013). New Phytologist 201: 1509-1509 Borecky J, Vercesi AE (2005) Plant uncoupling mitochondrial protein and alternative oxidase: energy metabolism and stress. Bioscience Reports 25: 271-286 Borisjuk L, Rolletschek H (2009) The oxygen status of the developing seed. New Phytologist 182: 17-30 Borisjuk L, Rolletschek H, Radchuk R, Weschke W, Wobus U, Weber H (2004) Seed development and differentiation: A role for metabolic regulation. Plant Biology 6: 375-386 Borisjuk L, Rolletschek H, Walenta S, Panitz R, Wobus U, Weber H (2003) Energy status and its control on embryogenesis of legumes: ATP distribution within Vicia faba embryos is developmentally regulated and correlated with photosynthetic capacity. Plant Journal 36: 318-329

276

References

Bourguignon J, Merand V, Rawsthorne S, Forest E, Douce R (1996) Glycine decarboxylase and pyruvate dehydrogenase complexes share the same dihydrolipoamide dehydrogenase in pea leaf mitochondria: Evidence from mass spectrometry and primary-structure analysis. Biochemical Journal 313: 229-234 Braun HP, Zabaleta E (2007) Carbonic anhydrase subunits of the mitochondrial NADH dehydrogenase complex (Complex I) in plants. Physiologia Plantarum 129: 114-122 Breidenb RW, Castelfr P, Peterson C (1966) Biogenesis of mitochondria in germinating peanut cotyledons. Plant Physiology 41: 803-809 Briggs D, Walters SM (1997) Plant variation and evolution. Cambridge University Press Burchett MD, Field CD, Pulkownik A (1984) Salinity, growth and root respiration in the grey mangrove, Avicennia marina. Physiologia Plantarum 60: 113-118 Burger G, Lang BF, Reith M, Gray MW (1996) Genes encoding the same three subunits of respiratory Complex II are present in the mitochondrial DNA of two phylogenetically distant eukaryotes. Proceedings of the National Academy of Sciences of the United States of America 93: 2328-2332 Bykova NV, Møller IM, Gardeström P, Igamberdiev AU (2014) The function of glycine decarboxylase complex is optimized to maintain high photorespiratory flux via buffering of its reaction products. Mitochondrion 19: 357-364 Caccere R, Teixeira SP, Centeno DC, Figueiredo-Ribeiro RdCL, Braga MR (2013) Metabolic and structural changes during early maturation of Inga vera seeds are consistent with the lack of a desiccation phase. Journal of Plant Physiology 170: 791-800 Calero E, West SH, Hinson K (1981) Water absorption of soybean seeds and associated causal factors. Crop Science 21: 926-933 Cao J, Schneeberger K, Ossowski S, Gunther T, Bender S, Fitz J, Koenig D, Lanz C, Stegle O, Lippert C, Wang X, Ott F, Muller J, Alonso-Blanco C, Borgwardt K, Schmid KJ, Weigel D (2011) Whole-genome sequencing of multiple Arabidopsis thaliana populations. Nat Genet 43: 956-963 Carrari F, Nunes-Nesi A, Gibon Y, Lytovchenko A, Loureiro ME, Fernie AR (2003) Reduced expression of aconitase results in an enhanced rate of photosynthesis and marked shifts in carbon partitioning in illuminated leaves of wild species tomato. Plant Physiology 133: 1322-1335 Chan ZL, Loescher W, Grumet R (2013) Transcriptional variation in response to salt stress in commonly used Arabidopsis thaliana accessions. Plant Physiology and Biochemistry 73: 189-201 Chapin FS, Oechel WC (1983) Photosynthesis, respiration, and phosphate absorption by Carex aquatilis ecotypes along latitudinal and local environmental gradients. Ecology 64: 743-751 Chardon F, Barthélémy J, Daniel-Vedele F, Masclaux-Daubresse C (2010) Natural variation of nitrate uptake and nitrogen use efficiency in Arabidopsis thaliana cultivated with limiting and ample nitrogen supply. Journal of experimental botany 61: 2293-2302 Chen F, Tobimatsu Y, Havkin-Frenkel D, Dixon RA, Ralph J (2012) A polymer of caffeyl alcohol in plant seeds. Proceedings of the National Academy of Sciences of the United States of America 109: 1772-1777 Chen H, Osuna D, Colville L, Lorenzo O, Graeber K, Kuester H, Leubner-Metzger G, Kranner I (2013) Transcriptome-wide mapping of pea seed ageing reveals a pivotal role for genes related to oxidative stress and programmed cell death. Plos One 8 Chen SL, Towill LR, Loewenbe JR (1970) Isoenzyme patterns in developing Xanthium leaves. Physiologia Plantarum 23: 434-443

277

References

Chen WH, de Meaux J, Lercher MJ (2010) Co-expression of neighbouring genes in Arabidopsis: separating chromatin effects from direct interactions. BMC Genomics 11 Chen Y, Xu D (2003) Computational analyses of high-throughput protein-protein interaction data. Current Protein & Peptide Science 4: 159-180 Chevalier F, Martin O, Rofidal V, Devauchelle AD, Barteau S, Sommerer N, Rossignol M (2004) Proteomic investigation of natural variation between Arabidopsis ecotypes. Proteomics 4: 1372-1381 Chia DW, Yoder TJ, Reiter WD, Gibson SI (2000) Fumaric acid: an overlooked form of fixed carbon in Arabidopsis and other plant species. Planta 211: 743-751 Choi BH, Choi YS, Kane DG, Kim BJ, Song YH, Cha HJ (2010) Cell behavior on extracellular matrix mimic materials based on mussel adhesive protein fused with functional peptides. Biomaterials 31: 8980-8988 Chyzhykova OA, Palladina TO (2006) The role of amino acids and sugars in supporting of osmotic homeostasis in maize seedlings under salinization conditions and treatment with synthetic growth regulators. Ukrainskii biokhimicheskii zhurnal 78: 124-129 Collakova E, Goyer A, Naponelli V, Krassovskaya I, Gregory JF, Hanson AD, Shachar-Hill Y (2008) Arabidopsis 10-formyl tetrahydrofolate deformylases are essential for photorespiration. Plant Cell 20: 1818-1832 Cousins AB, Pracharoenwattana I, Zhou WX, Smith SM, Badger MR (2008) Peroxisomal malate dehydrogenase is not essential for photorespiration in Arabidopsis but its absence causes an increase in the stoichiometry of photorespiratory CO2 release. Plant Physiology 148: 786-795 Cousins AB, Walker BJ, Pracharoenwattana I, Smith SM, Badger MR (2011) Peroxisomal hydroxypyruvate reductase is not essential for photorespiration in Arabidopsis but its absence causes an increase in the stoichiometry of photorespiratory CO2 release. Photosynthesis Research 108: 91-100 Cramer GR, Van Sluyter SC, Hopper DW, Pascovici D, Keighley T, Haynes PA (2013) Proteomic analysis indicates massive changes in metabolism prior to the inhibition of growth and photosynthesis of grapevine (Vitis vinifera L.) in response to water deficit. BMC Plant Biology 13 Cramer MD, Schierholt A, Wang YZ, Lips SH (1995) The influence of salinity on the utilization of root anaplerotic carbon and nitrogen-metabolism in tomato seedlings. Journal of Experimental Botany 46: 1569-1577 Cross JM, von Korff M, Altmann T, Bartzetko L, Sulpice R, Gibon Y, Palacios N, Stitt M (2006) Variation of enzyme activities and metabolite levels in 24 Arabidopsis accessions growing in carbon limited conditions. Plant Physiology 142: 1574-1588 Cyr AR, Domann FE (2011) The redox basis of epigenetic modifications: from mechanisms to functional consequences. Antioxidants and Redox Signaling 15: 551-589 Daschner K, Couee I, Binder S (2001) The mitochondrial isovaleryl-coenzyme A dehydrogenase of Arabidopsis oxidizes intermediates of leucine and valine catabolism. Plant Physiology 126: 601-612 Day DA, Devos OC, Wilson D, Lambers H (1985) Regulation of respiration in the leaves and roots of two Lolium perenne populations with contrasting mature leaf respiration rates and crop yields. Plant Physiology 78: 678-683 Day DA, Dry IB, Soole KL, Wiskich JT, Moore AL (1991) Regulation of alternative pathway activity in plant-mitochondria -deviations from Q-pool behavior during oxidation of NADH and quinols. Plant Physiology 95: 948-953

278

References

Day DA, Whelan J, Millar AH, Siedow JN, Wiskich JT (1995) Regulation of the alternative oxidase in plants and fungi. Australian Journal of Plant Physiology 22: 497-509 de Longevialle AF, Meyer EH, Andres C, Taylor NL, Lurin C, Millar AH, Small ID (2007) The pentatricopeptide repeat gene OTP43 is required for trans-splicing of the mitochondrial nad1 intron 1 in Arabidopsis thaliana. Plant Cell 19: 3256-3265 Deisboeck T, Kresh JY (2007) Complex systems science in biomedicine. Springer Science and Business Media delaFuente JM, RamirezRodriguez V, CabreraPonce JL, HerreraEstrella L (1997) Aluminum tolerance in transgenic plants by alteration of citrate synthesis. Science 276: 1566- 1568 Delaval E, Perichon M, Friguet B (2004) Age related impairment of mitochondrial matrix aconitase and ATP stimulated protease in rat liver and heart. European Journal of Biochemistry 271: 4559-4564 Delgado J, Liao JC (1992) Determination of flux control coefficients from transient metabolite concentrations. Biochemical Journal 282: 919-927 Delhaize E, Craig S, Beaton CD, Bennet RJ, Jagadish VC, Randall PJ (1993) Aluminum tolerance in wheat (Triticum aestivum L.). I. Uptake and distribution of aluminum in root apices. Plant Physiology 103: 685-693 Deng XD, Cai JJ, Fei XW (2013) Effect of the expression and knockdown of citrate synthase gene on carbon flux during triacylglycerol biosynthesis by green algae Chlamydomonas reinhardtii. BMC Biochemistry 14 Denton RM, Randle, P.J., Bridges, B.J., Cooper, R.H., Kerbey, A.L., , Pask HT, Severson, D.L., Stansbie, D., Whitehouse, S. (1975) Regulation of mammalian pyruvate dehydrogenase. Molecular and Cellular Biochemistry 9: 27-53 Desrochers A, Landhausser SM, Lieffers VJ (2002) Coarse and fine root respiration in aspen (Populus tremuloides). Tree Physiology 22: 725-732 DieckmannSchuppert A, Schnittler HJ (1997) A simple assay for quantification of protein in tissue sections, cell cultures, and cell homogenates, and of protein immobilized on solid surfaces. Cell and Tissue Research 288: 119-126 Dieuaide-Noubhani M, Canioni P, Raymond P (1997) Sugar starvation induced changes of carbon metabolism in excised maize root tips. Plant Physiology 115: 1505-1513 Eastmond P, Kolacna L, Rawsthorne S (1996) Photosynthesis by developing embryos of oilseed rape (Brassica napus L). Journal of Experimental Botany 47: 1763-1769 Eastmond PJ, Graham IA (2001) Re-examining the role of the glyoxylate cycle in oilseeds. Trends in Plant Science 6: 72-77 Eastmond PJ, Rawsthorne S (2000) Coordinate changes in carbon partitioning and plastidial metabolism during the development of oilseed rape embryo. Plant Physiology 122: 767-774 Eden A, Benvenisty N (1999) Involvement of branched-chain amino acid aminotransferase (Bcat1/Eca39) in apoptosis. FEBS Letters 457: 255-261 Ehrenshaft M, Brambl R (1990) Respiration and mitochondrial biogenesis in germinating embryos of maize. Plant Physiology 93: 295-304 Ellis RH, Roberts EH (1980) Improved equations for the prediction of seed longevity. Annals of Botany 45: 13-30 Epron D, Toussaint ML, Badot PM (1999) Effects of sodium chloride salinity on root growth and respiration in oak seedlings. Annals of Forest Science 56: 41-47 Fait A, Angelovici R, Less H, Ohad I, Urbanczyk-Wochniak E, Fernie AR, Galili G (2006) Arabidopsis seed development and germination is associated with temporally distinct metabolic switches. Plant Physiology 142: 839-854

279

References

Falk KL, Behal RH, Xiang CB, Oliver DJ (1998) Metabolic bypass of the tricarboxylic acid cycle during lipid mobilization in germinating oilseeds - Regulation of NAD(+)- dependent isocitrate dehydrogenase versus fumarase. Plant Physiology 117: 473- 481 Farrar JF (1985) The respiratory source of CO2. Plant Cell and Environment 8: 427-438 Feng XH, Dietze M (2013) Scale dependence in the effects of leaf ecophysiological traits on photosynthesis: Bayesian parameterization of photosynthesis models. New Phytologist 200: 1132-1144 Feng Y, De Franceschi G, Kahraman A, Soste M, Melnik A, Boersema PJ, de Laureto PP, Nikolaev Y, Oliveira AP, Picotti P (2014) Global analysis of protein structural changes in complex proteomes. Nature Biotechnology 32 Ferrick DA, Neilson A, Beeson C (2008) Advances in measuring cellular bioenergetics using extracellular flux. Drug Discovery Today 13: 268-274 Filiault DL, Wessinger CA, Dinneny JR, Lutes J, Borevitz JO, Weigel D, Chory J, Maloof JN (2008) Amino acid polymorphisms in Arabidopsis phytochrome B cause differential responses to light. Proceedings of the National Academy of Sciences of the United States of America 105: 3157-3162 Finney DJ (1982) Citation classic - probit analysis - a statistical treatment of the sigmoid response curve. Current Contents/Life Sciences: 18-18 Flick K, Kaiser P (2012) Protein degradation and the stress response. Seminars in Cell and Developmental Biology 23: 515-522 Fosket DE (1994) Plant growth and development: a molecular approach. Academic Press Inc. Foyer CH, Noctor G, Hodges M (2011) Respiration and nitrogen assimilation: targeting mitochondria-associated metabolism as a means to enhance nitrogen use efficiency. Journal of Experimental Botany 62: 1467-1482 Fu Q, Wang BC, Jin X, Li HB, Han P, Wei KH, Zhang XM, Zhu YX (2005) Proteomic analysis and extensive protein identification from dry, germinating Arabidopsis seeds and young seedlings. Journal of Biochemistry and Molecular Biology 38: 650-660 Fuentes D, Meneses M, Nunes-Nesi A, Araujo WL, Tapia R, Gomez I, Holuigue L, Gutierrez RA, Fernie AR, Jordana X (2011) A deficiency in the flavoprotein of Arabidopsis mitochondrial Complex II results in elevated photosynthesis and better growth in nitrogen-limiting conditions. Plant Physiology 157: 1114-1127 Galili G (2011) The aspartate-family pathway of plants: linking production of essential amino acids with energy and stress regulation. Plant Signaling and Behavior 6: 192- 195 Gan XC, Stegle O, Behr J, Steffen JG, Drewe P, Hildebrand KL, Lyngsoe R, Schultheiss SJ, Osborne EJ, Sreedharan VT, Kahles A, Bohnert R, Jean G, Derwent P, Kersey P, Belfield EJ, Harberd NP, Kemen E, Toomajian C, Kover PX, Clark RM, Ratsch G, Mott R (2011) Multiple reference genomes and transcriptomes for Arabidopsis thaliana. Nature 477: 419-423 Garcia D, Fitz Gerald JN, Berger F (2005) Maternal control of integument cell elongation and zygotic control of endosperm growth are coordinated to determine seed size in Arabidopsis. Plant Cell 17: 52-60 Garmier M, Carroll AJ, Delannoy E, Vallet C, Day DA, Small ID, Millar AH (2008) Complex I dysfunction redirects cellular and mitochondrial metabolism in Arabidopsis. Plant Physiology 148: 1324-1341 Gazzani S, Gendall AR, Lister C, Dean C (2003) Analysis of the molecular basis of flowering time variation in Arabidopsis accessions. Plant Physiology 132: 1107-1114

280

References

Geigenberger P (2003) Response of plant metabolism to too little oxygen. Current Opinion in Plant Biology 6: 247-256 Gerencser AA, Neilson A, Choi SW, Edman U, Yadava N, Oh RJ, Ferrick DA, Nicholls DG, Brand MD (2009) Quantitative microplate-based respirometry with correction for oxygen diffusion. Analytical Chemistry 81: 6868-6878 Gianoli E, Saldana A, Jimenez-Castillo M (2012) Ecophysiological traits may explain the abundance of climbing plant species across the light gradient in a temperate rainforest. Plos One 7: e38831 Gietl C (1992) Malate dehydrogenase isoenzymes - cellular locations and role in the flow of metabolites between the cytoplasm and cell organelles. Biochimica Et Biophysica Acta 1100: 217-234 Gleason C, Huang S, Thatcher LF, Foley RC, Anderson CR, Carroll AJ, Millar AH, Singh KB (2011) Mitochondrial Complex II has a key role in mitochondrial-derived reactive oxygen species influence on plant stress gene regulation and defense. Proceedings of the National Academy of Sciences of the United States of America 108: 10768- 10773 Goel A, Goel AK, Sheoran IS (2003) Changes in oxidative stress enzymes during artificial ageing in cotton (Gossypium hirsutum L.) seeds. Journal of Plant Physiology 160: 1093-1100 Goffman FD, Alonso AP, Schwender J, Shachar-Hill Y, Ohlrogge JB (2005) Light enables a very high efficiency of carbon storage in developing embryos of rapeseed. Plant Physiology 138: 2269-2279 Goldberg RB, Depaiva G, Yadegari R (1994) Plant embryogenesis - zygote to seed. Science 266: 605-614 Gonzàlez-Meler MA, Ribas-Carbo M, Giles L, Siedow JN (1999) The effect of growth and measurement temperature on the activity of the alternative respiratory pathway. Plant Physiology 120: 765-772 Gördes D, Koch G, Thurow K, Kolukisaoglu Ü (2013) Analyses of Arabidopsis ecotypes reveal metabolic diversity to convert D-amino acids. SpringerPlus 2: 1-11 Graham IA (2008) Seed storage oil mobilization. In Annual Review of Plant Biology, Vol 59. Annual Reviews, Palo Alto, pp 115-142 Gu X (2003) Evolution of duplicate genes versus genetic robustness against null mutations. Trends in Genetics 19: 354-356 Guan M, Moller IS, Schjoerring JK (2015) Two cytosolic glutamine synthetase isoforms play specific roles for seed germination and seed yield structure in Arabidopsis. Journal of Experimental Botany 66: 203-212 Guan YH, Rawsthorne S, Scofield G, Shaw P, Doonan J (1995) Cloning and characterization of a dihydrolipoamide acetyltransferase (E2) subunit of the pyruvate dehydrogenase complex from Arabidopsis thaliana. Journal of Biological Chemistry 270: 5412-5417 Gupta KJ, Zabalza A, van Dongen JT (2009) Regulation of respiration when the oxygen availability changes. Physiologia Plantarum 137: 383-391 Gygi SP, Rochon Y, Franza BR, Aebersold R (1999) Correlation between protein and mRNA abundance in yeast. Molecular and Cellular Biology 19: 1720-1730 Hakozaki M, Honda K (1990) Specificity of lipoamide dehydrogenase for alpha-keto acid dehydrogenase complexes and the glycine cleavage system. Fukushima Journal of Medical Science 36: 71-81 Hampton JG, Tekrony DM (1995) Handbook of vigour test methods. International Seed Testing Association, Zurich

281

References

Hanbury CD, Atwell BJ (2005) Growth dynamics of mechanically impeded lupin roots: Does altered morphology induce hypoxia? Annals of Botany 96: 913-924 Harada JJ (1997) Seed maturation and control of germination. Cellular and Molecular Biology of Plant Seed Development 4: 545-592 Hare PD, Cress WA, Van Staden J (1998) Dissecting the roles of osmolyte accumulation during stress. Plant Cell and Environment 21: 535-553 Hatch MD, Osmond CB (1976) Compartmentation and transport in C4 photosynthesis. Springer Berlin Heidelberg Hay R, K.M., Walker, Andrew, J. (1989) An introduction to the physiology of crop yield. Longman Scientific and Technical. New York Hayashi K, Hasegawa J, Matsunaga S (2013) The boundary of the meristematic and elongation zones in roots: endoreduplication precedes rapid cell expansion. Scientific Reports 3 Hayes MK, Luethy MH, Elthon TE (1991) Mitochondrial malate dehydrogenase from corn - purification of multiple forms. Plant Physiology 97: 1381-1387 Heazlewood JL, Millar AH (2005) AMPDB: the Arabidopsis mitochondrial protein database. Nucleic Acids Research 33: D605-D610 Heazlewood JL, Tonti-Filippini JS, Gout AM, Day DA, Whelan J, Millar AH (2004) Experimental analysis of the Arabidopsis mitochondrial proteome highlights signaling and regulatory components, provides assessment of targeting prediction programs, and indicates plant-specific mitochondrial proteins. Plant Cell 16: 241- 256 Hebbelmann I, Selinski J, Wehmeyer C, Goss T, Voss I, Mulo P, Kangasjarvi S, Aro E-M, Oelze M-L, Dietz K-J, Nunes-Nesi A, Do PT, Fernie AR, Talla SK, Raghavendra AS, Linke V, Scheibe R (2012) Multiple strategies to prevent oxidative stress in Arabidopsis plants lacking the malate valve enzyme NADP-malate dehydrogenase. Journal of Experimental Botany 63: 1445-1459 Heber U (1974) Metabolite exchange between chloroplasts and cytoplasm. Annual Review of Plant Physiology and Plant Molecular Biology 25: 393-421 Hentze MW, Kuhn LC (1996) Molecular control of vertebrate iron metabolism: mRNA- based regulatory circuits operated by iron, nitric oxide, and oxidative stress. Proceedings of the National Academy of Sciences of the United States of America 93: 8175-8182 Hill DW, Powell T (1968) Non-dispersive infra-red gas analysis in science, medicine and industry. Hilger, London Hirst J, Carroll J, Fearnley IM, Shannon RJ, Walker JE (2003) The nuclear encoded subunits of Complex I from bovine heart mitochondria. Biochimica Et Biophysica Acta- Bioenergetics 1604: 135-150 Hitchler MJ, Domann FE (2007) An epigenetic perspective on the free radical theory of development. Free Radical Biology and Medicine 43: 1023-1036 Hodges M (2002) Enzyme redundancy and the importance of 2-oxoglutarate in plant ammonium assimilation. Journal of Experimental Botany 53: 905-916 Hoefnagel MHN, Atkin OK, Wiskich JT (1998) Interdependence between chloroplasts and mitochondria in the light and the dark. Biochimica Et Biophysica Acta-Bioenergetics 1366: 235-255 Hoffmann MH (2002) Biogeography of Arabidopsis thaliana (L.) Heynh. (Brassicaceae). Journal of Biogeography 29: 125-134 Hong SM, Bahn SC, Lyu A, Jung HS, Ahn JH (2010) Identification and testing of superior reference genes for a starting pool of transcript normalization in Arabidopsis. Plant and Cell Physiology 51: 1694-1706

282

References

Hooper CM, Tanz SK, Castleden IR, Vacher MA, Small ID, Millar AH (2014) SUBAcon: a consensus algorithm for unifying the subcellular localization data of the Arabidopsis proteome. Bioinformatics 30: 3356-3364 Horecker BL (1962) Interdependent pathways of carbohydrate metabolism. Harvey Lectures: 35-61 Hourmant A, Pradet A (1981) Oxidative phosphorylation in germinating lettuce seeds (Lactuca sativa) during the 1st hours of imbibition. Plant Physiology 68: 631-635 Howell KA, Millar AH, Whelan J (2007) Building the powerhouse: what are the signals involved in plant mitochondrial biogenesis? Plant Signaling and Behavior 2: 428-430 Hruz T, Laule O, Szabo G, Wessendorp F, Bleuler S, Oertle L, Widmayer P, Gruissem W, Zimmermann P (2008) Genevestigator v3: a reference expression database for the meta-analysis of transcriptomes. Advances in Bioinformatics 2008: 420747 Hu J, Baker A, Bartel B, Linka N, Mullen RT, Reumann S, Zolman BK (2012) Plant peroxisomes: biogenesis and function. The Plant Cell Online 24: 2279-2303 Hua W, Li R-J, Zhan G-M, Liu J, Li J, Wang X-F, Liu G-H, Wang H-Z (2012) Maternal control of seed oil content in Brassica napus: the role of silique wall photosynthesis. Plant Journal 69: 432-444 Huang S, Millar AH (2013) Succinate dehydrogenase: the complex roles of a simple enzyme. Current Opinion in Plant Biology 16: 344-349 Huang S, Taylor NL, Narsai R, Eubel H, Whelan J, Millar AH (2010) Functional and composition differences between mitochondrial Complex II in Arabidopsis and rice are correlated with the complex genetic history of the enzyme. Plant Molecular Biology 72: 331-342 Huang S, Taylor NL, Stroeher E, Fenske R, Millar AH (2013) Succinate dehydrogenase assembly factor 2 is needed for assembly and activity of mitochondrial Complex II and for normal root elongation in Arabidopsis. Plant Journal 73: 429-441 Huang SB, Taylor NL, Stroher E, Fenske R, Millar AH (2013) Succinate dehydrogenase assembly factor 2 is needed for assembly and activity of mitochondrial Complex II and for normal root elongation in Arabidopsis. Plant Journal 73: 429-441 Huh JH, Bauer MJ, Hsieh T-F, Fischer R (2007) Endosperm gene imprinting and seed development. Current Opinion in Genetics and Development 17: 480-485 Hunt S (2003) Measurements of photosynthesis and respiration in plants. Physiologia Plantarum 117: 314-325 Huynen MA, Dandekar T, Bork P (1999) Variation and evolution of the citric acid cycle: a genomic perspective. Trends in Microbiology 7: 281-291 Hwang YH, Morris JT (1994) Whole-plant gas-exchange responses of Spartina alterniflora (Poaceae) to a range of constant and transient salinities. American Journal of Botany 81: 659-665 Igamberdiev AU, Bykova NV, Lea PJ, Gardestrom P (2001) The role of photorespiration in redox and energy balance of photosynthetic plant cells: A study with a barley mutant deficient in glycine decarboxylase. Physiologia Plantarum 111: 427-438 Igamberdiev AU, Gardestrom P (2003) Regulation of NAD- and NADP-dependent isocitrate dehydrogenases by reduction levels of pyridine nucleotides in mitochondria and cytosol of pea leaves. Biochimica Et Biophysica Acta-Bioenergetics 1606: 117-125 Igamberdiev AU, Lernmark U, Gardeström P (2014) Activity of the mitochondrial pyruvate dehydrogenase complex in plants is stimulated in the presence of malate. Mitochondrion 19: 184-190 Ikram S, Bedu M, Daniel-Vedele F, Chaillou S, Chardon F (2011) Natural variation of Arabidopsis response to nitrogen availability. Journal of experimental botany: err244

283

References

Ishizaki K, Larson TR, Schauer N, Fernie AR, Graham IA, Leaver CJ (2005) The critical role of Arabidopsis electron transfer flavoprotein: ubiquinone oxidoreductase during dark induced starvation. Plant Cell 17: 2587-2600 Ishizaki K, Schauer N, Larson TR, Graham IA, Fernie AR, Leaver CJ (2006) The mitochondrial electron transfer flavoprotein complex is essential for survival of Arabidopsis in extended darkness. Plant Journal 47: 751-760 Ito J, Batth TS, Petzold CJ, Redding-Johanson AM, Mukhopadhyay A, Verboom R, Meyer EH, Millar AH, Heazlewood JL (2011) Analysis of the Arabidopsis cytosolic proteome highlights subcellular partitioning of central plant metabolism Journal of Proteome Research 10: 3309-3309 Izumi M, Hidema J, Makino A, Ishida H (2013) Autophagy contributes to night time energy availability for growth in Arabidopsis. Plant Physiology 161: 1682-1693 Jablonka E, Raz G (2009) Transgenerational epigenetic inheritance: prevalence, mechanisms, and implications for the study of heredity and evolution. Quarterly Review of Biology 84: 131-176 Jacoby RP, Taylor NL, Millar AH (2011) The role of mitochondrial respiration in salinity tolerance. Trends in Plant Science 16: 614-623 Jeong ML, Jiang HY, Chen HS, Tsai CJ, Harding SA (2004) Metabolic profiling of the sink-to- source transition in developing leaves of quaking aspen. Plant Physiology 136: 3364-3375 Jiang KN, Ballinger T, Li D, Zhang SB, Feldman L (2006) A role for mitochondria in the establishment and maintenance of the maize root quiescent center. Plant Physiology 140: 1118-1125 Journet EP, Neuburger M, Douce R (1981) Role of glutamate-oxaloacetate transaminase and malate-dehydrogenase in the regeneration of NAD+ for glycine oxidation by spinach leaf mitochondria. Plant Physiology 67: 467-469 Jung JKH, McCouch S (2013) Getting to the roots of it: genetic and hormonal control of root architecture. Frontiers in Plant Science 4 Juszczak I, Rudnik R, Pietzenuk B, Baier M (2012) Natural genetic variation in the expression regulation of the chloroplast antioxidant system among Arabidopsis thaliana accessions. Physiologia Plantarum 146: 53-70 Kadenbach B, Huttemann M, Arnold S, Lee I, Bender E (2000) Mitochondrial energy metabolism is regulated via nuclear-coded subunits of cytochrome c oxidase. Free Radical Biology and Medicine 29: 211-221 Kasimova MR, Grigiene J, Krab K, Hagedorn PH, Flyvbjerg H, Andersen PE, Moller IM (2006) The free NADH concentration is kept constant in plant mitochondria under different metabolic conditions. Plant Cell 18: 688-698 Katori T, Ikeda A, Iuchi S, Kobayashi M, Shinozaki K, Maehashi K, Sakata Y, Tanaka S, Taji T (2010) Dissecting the genetic control of natural variation in salt tolerance of Arabidopsis thaliana accessions. Journal of Experimental Botany 61: 1125-1138 Kermode AR, Gifford DJ, Bewley JD (1985) The role of maturation drying in the transition from seed development to germination: III. Insoluble protein synthetic pattern changes within the endosperm of Ricinus communis l. seeds. Journal of Experimental Botany 36: 1928-1936 Keys AJ (2006) The re-assimilation of ammonia produced by photorespiration and the nitrogen economy of C3 higher plants. Photosynthesis Research 87: 165-175 Keys AJ, Bird IF, Cornelius MJ, Lea PJ, Wallsgrove RM, Miﬂin BJ (1978) Photorespiratory nitrogen cycle. Nature 275: 741–743 Kharchenko P, Church GM, Vitkup D (2005) Expression dynamics of a cellular metabolic network. Molecular Systems Biology 1

284

References

Kitteringham NR, Jenkins RE, Lane CS, Elliott VL, Park BK (2009) Multiple reaction monitoring for quantitative biomarker analysis in proteomics and metabolomics. Journal of Chromatography B-Analytical Technologies in the Biomedical and Life Sciences 877: 1229-1239 Kleessen S, Araújo WL, Fernie AR, Nikoloski Z (2012) Model-based confirmation of alternative substrates of mitochondrial electron transport chain. Journal of Biological Chemistry 287: 11122-11131 Kochevenko A, Araújo WL, Maloney GS, Tieman DM, Do PT, Taylor MG, Klee HJ, Fernie AR (2012) Catabolism of branched chain amino acids supports respiration but not volatile synthesis in tomato fruits. Molecular Plant 5: 366-375 Kochian LV, Hoekenga OA, Pineros MA (2004) How do crop plants tolerate acid soils? - Mechanisms of aluminum tolerance and phosphorous efficiency. Annual Review of Plant Biology 55: 459-493 Kolloffe C (1970) Oxidative and phosphorylative activity of mitochondria from pea cotyledons during maturation of seed. Planta 91: 321-328 Konopka CA, Bednarek SY (2008) Comparison of the dynamics and functional redundancy of the Arabidopsis dynamin-related isoforms DRP1A and DRP1C during plant development. Plant Physiology 147: 1590-1602 Koornneef M, Alonso-Blanco C, Vreugdenhil D (2004) Naturally occurring genetic variation in Arabidopsis thaliana. Annual Review of Plant Biology 55: 141-172 Krauss S, Zhang CY, Lowell BB (2005) The mitochondrial uncoupling-protein homologues. Nature Reviews Molecular Cell Biology 6: 248-261 Krebs HA, Johnson WA (1937) Metabolism of ketonic acids in animal tissues. Biochemical Journal 31: 645-660 Krömer S (1995) Respiration during photosynthesis. Annual Review of Plant Biology 46: 45– 70 Kromer S, Heldt HW (1991) Respiration of pea leaf mitochondria and redox transfer between the mitochondrial and extramitochondrial compartment. Biochimica Et Biophysica Acta 1057: 42-50 Kunze M, Pracharoenwattana I, Smith SM, Hartig A (2006) A central role for the peroxisomal membrane in glyoxylate cycle function. Biochimica Et Biophysica Acta- Molecular Cell Research 1763: 1441-1452 Lambers H (1985) Respiration in intact plants and tissues: Its regulation and dependence on environmental factors, metabolism and invaded organisms. In R Douce, D Day, eds, Higher Plant Cell Respiration. Springer Berlin Heidelberg, pp 418-473 Lambers H, Stulen I, vanderWerf A (1996) Carbon use in root respiration as affected by elevated atmospheric O2. Plant and Soil 187: 251-263 Lambers H, Szaniawski RK, Devisser R (1983) Respiration for growth, maintenance and ion uptake - an evaluation of concepts, methods, values and their significance. Physiologia Plantarum 58: 556-563 Lambeth DO, Tews KN, Adkins S, Frohlich D, Milavetz BI (2004) Expression of two succinyl- CoA synthetases with different nucleotide specificities in mammalian tissues. Journal of Biological Chemistry 279: 36621-36624 Lance C, Rustin P (1984) The central role of malate in plant metabolism. Physiologie Vegetale 22: 625-641 Lancien M, Ferrario-Méry S, Roux Y, Bismuth E, Masclaux C, Hirel B, Gadal P, Hodges M (1999) Simultaneous expression of NAD-dependent isocitrate dehydrogenase and other krebs cycle genes after nitrate resupply to short term nitrogen starved tobacco. Plant Physiology 120: 717-726

285

References

Lancien M, Gadal P, Hodges M (1998) Molecular characterization of higher plant NAD- dependent isocitrate dehydrogenase: evidence for a heteromeric structure by the complementation of yeast mutants. Plant Journal 16: 325-333 Landschutze V, Mullerrober B, Willmitzer L (1995) Mitochondrial citrate synthase from potato - predominant expression in mature leaves and young flower buds. Planta 196: 756-764 Landschutze V, Willmitzer L, Mullerrober B (1995) Inhibition of flower formation by antisense repression of mitochondrial citrate synthase in transgenic potato plants leads to a specific disintegration of the ovary tissues of flowers. EMBO Journal 14: 660-666 Ledig FT, Korbobo DR (1983) Adaptation of sugar maple populations along altitudinal gradients - photosynthesis, respiration, and specific leaf weight. American Journal of Botany 70: 256-265 Lee CP, Eubel H, Millar AH (2010) Diurnal changes in mitochondrial function reveal daily optimization of light and dark respiratory metabolism in Arabidopsis. Molecular and Cellular Proteomics 9: 2125-2139 Lee CP, Eubel H, O'Toole N, Millar AH (2008) Heterogeneity of the mitochondrial proteome for photosynthetic and non-photosynthetic Arabidopsis metabolism. Molecular and Cellular Proteomics 7: 1297-1316 Leegood RC, Lea PJ, Adcock MD, Häusler RE (1995) The regulation and control of photorespiration. Journal of Experimental Botany 46: 1397-1414 Lemaitre T, Hodges M (2006) Expression analysis of Arabidopsis thaliana NAD-dependent isocitrate dehydrogenase genes shows the presence of a functional subunit that is mainly expressed in the pollen and absent from vegetative organs. Plant and Cell Physiology 47: 634-643 Lemaitre T, Urbanczyk-Wochniak E, Flesch V, Bismuth E, Fernie AR, Hodges M (2007) NAD- dependent isocitrate dehydrogenase mutants of Arabidopsis suggest the enzyme is not limiting for nitrogen assimilation. Plant Physiology 144: 1546-1558 Leon G, Holuigue L, Jordana X (2007) Mitochondrial Complex II is essential for gametophyte development in Arabidopsis. Plant Physiology 143: 1534-1546 Leopold AC, Vertucci CW (1989) Moisture as a regulator of physiological reaction in seeds. In PC Stanwood, MB McDonald, eds, Seed Moisture. Crop Science Society of America, pp 51-67 Lernmark U, Gardestrom P (1994) Distribution of pyruvate dehydrogenase complex activities between chloroplasts and mitochondria from leaves of different species. Plant Physiology 106: 1633-1638 Li B, Suzuki JI, Hara T (1998) Latitudinal variation in plant size and relative growth rate in Arabidopsis thaliana. Oecologia 115: 293-301 Li PH, Sioson A, Mane SP, Ulanov A, Grothaus G, Heath LS, Murali TM, Bohnert HJ, Grene R (2006) Response diversity of Arabidopsis thaliana ecotypes in elevated CO2 in the field. Plant Molecular Biology 62: 593-609 Li Y, Park JS, Deng JH, Bai Y (2006) Cytochrome c oxidase subunit IV is essential for assembly and respiratory function of the enzyme complex. Journal of Bioenergetics and Biomembranes 38: 283-291 Liang M, Davis E, Gardner D, Cai X, Wu Y (2006) Involvement of AtLAC15 in lignin synthesis in seeds and in root elongation of Arabidopsis. Planta 224: 1185-1196 Liebler DC, Zimmerman LJ (2013) Targeted quantitation of proteins by mass spectrometry. Biochemistry 52: 3797-3806

286

References

Lin M, Behal RH, Oliver DJ (2004) Characterization of a mutation in the IDH-II subunit of the NAD(+)-dependent isocitrate dehydrogenase from Arabidopsis thaliana. Plant Science 166: 983-988 Lin TP, Chen MH (1995) Biochemical characteristics associated with the development of the desiccation-sensitive seeds of Machilus thunbergii Sieb. and Zucc. Annals of Botany 76: 381-387 Lisenbee CS, Lingard MJ, Trelease RN (2005) Arabidopsis peroxisomes possess functionally redundant membrane and matrix isoforms of monodehydroascorbate reductase. Plant Journal 43: 900-914 Logan DC, Millar AH, Sweetlove LJ, Hill SA, Leaver CJ (2001) Mitochondrial biogenesis during germination in maize embryos. Plant Physiology 125: 662-672 Long RL, Panetta FD, Steadman KJ, Probert R, Bekker RM, Brooks S, Adkins SW (2008) Seed persistence in the field may be predicted by laboratory-controlled aging. Weed Science 56: 523-528 Lopez-Molina L, Mongrand S, Chua NH (2001) A postgermination developmental arrest checkpoint is mediated by abscisic acid and requires the AB15 transcription factor in Arabidopsis. Proceedings of the National Academy of Sciences of the United States of America 98: 4782-4787 Loreti E, Poggi A, Novi G, Alpi A, Perata P (2005) A genome-wide analysis of the effects of sucrose on gene expression in Arabidopsis seedlings under anoxia. Plant Physiology 137: 1130-1138 Loudet O, Chaillou S, Merigout P, Talbotec J, Daniel-Vedele F (2003) Quantitative trait loci analysis of nitrogen use efficiency in Arabidopsis. Plant Physiology 131: 345-358 Loudet O, Saliba-Colombani V, Camilleri C, Calenge F, Gaudon V, Koprivova A, North KA, Kopriva S, Daniel-Vedele F (2007) Natural variation for sulfate content in Arabidopsis thaliana is highly controlled by APR2. Nature Genetics 39: 896-900 Loveless MD, Hamrick JL (1984) Ecological determinants of genetic-structure in plant- populations. Annual Review of Ecology and Systematics 15: 65-95 Ludwig LJ, Canvin DT (1971) Rate of photorespiration during photosynthesis and relationship of substrate of light respiration to products of photosynthesis in sunflower leaves. Plant Physiology 48: 712-719 Luethy MH, Gemel J, Johnston ML, Mooney BP, Miernyk JA, Randall DD (2001) Developmental expression of the mitochondrial pyruvate dehydrogenase complex in pea (Pisum sativum) seedlings. Physiologia Plantarum 112: 559-566 Luethy MH, Miernyk JA, David NR, Randall DD (1996) Plant pyruvate dehydrogenase complexes. In M Patel, T Roche, R Harris, eds, Alpha-Keto Acid Dehydrogenase Complexes. Birkhäuser Basel, pp 71-92 Luethy MH, Miernyk JA, Randall DD (1994) The nucleotide and deduced amino acid sequences of a cDNA encoding the E1-beta subunit of the Arabidopsis thaliana mitochondrial pyruvate dehydrogenase complex. Biochimica Et Biophysica Acta- Bioenergetics 1187: 95-98 Luethy MH, Miernyk JA, Randall DD (1995) The mitochondrial pyruvate dehydrogenase complex: nucleotide and deduced amino acid sequences of a cDNA encoding the Arabidopsis thaliana E1-alpha subunit. Gene 164: 251-254 Lutziger I, Oliver DJ (2001) Characterization of two cDNAs encoding mitochondrial lipoamide dehydrogenase from Arabidopsis. Plant Physiology 127: 615-623 Macherel D, Benamar A, Avelange-Macherel M-H, Tolleter D (2007) Function and stress tolerance of seed mitochondria. Physiologia Plantarum 129: 233-241

287

References

Malagoli P, Britto DT, Schulze LM, Kronzucker HJ (2008) Futile Na+ cycling at the root plasma membrane in rice (Oryza sativa L.): kinetics, energetics, and relationship to salinity tolerance. Journal of Experimental Botany 59: 4109-4117 Mansﬁeld SG, Briarty LG (1993) Endosperm development, Ed (Bowman, J., ed). Springer- Verlag, Berlin Mansfield SG, Briarty LG (1992) Cotyledon cell development in Arabidopsis thaliana during reserve deposition. Canadian Journal of Botany-Revue Canadienne De Botanique 70: 151-164 Mariko S, Koizumi H (1993) Respiration for maintenance and growth in Reynoutria japonica ecotypes from different altitudes on Mt Fuji. Ecological Research 8: 241-246 Maurino VG, Flügge U-I (2008) Experimental systems to assess the effects of reactive oxygen species in plant tissues. Plant Signaling and Behavior 3: 923-928 Mayer U, Jurgens G (1998) Pattern formation in plant embryogenesis: a reassessment. Seminars in Cell and Developmental Biology 9: 187-193 Mayer U, Ruiz RAT, Berleth T, Misera S, Jurgens G (1991) Mutations affecting body organization in the Arabidopsis embryo. Nature 353: 402-407 Mazzucotelli E, Tartari A, Cattivelli L, Forlani G (2006) Metabolism of gamma-aminobutyric acid during cold acclimation and freezing and its relationship to frost tolerance in barley and wheat. Journal of Experimental Botany 57: 3755-3766 McKhann HI, Camilleri C, Berard A, Bataillon T, David JL, Reboud X, Le Corre V, Caloustian C, Gut IG, Brunel D (2004) Nested core collections maximizing genetic diversity in Arabidopsis thaliana. Plant Journal 38: 193-202 Meinke DW, Cherry JM, Dean C, Rounsley SD, Koornneef M (1998) Arabidopsis thaliana: a model plant for genome analysis. Science 282: 662-682 Meyer EH, Tomaz T, Carroll AJ, Estavillo G, Delannoy E, Tanz SK, Small ID, Pogson BJ, Millar AH (2009) Remodeled respiration in ndufs4 with low phosphorylation efficiency suppresses Arabidopsis germination and growth and alters control of metabolism at night. Plant Physiology 151: 603-609 Micallef SA, Shiaris MP, Colon-Carmona A (2009) Influence of Arabidopsis thaliana accessions on rhizobacterial communities and natural variation in root exudates. Journal of Experimental Botany 60: 1729-1742 Migliaccio F, Piconese S (2001) Spiralizations and tropisms in Arabidopsis roots. Trends in Plant Science 6: 561-565 Millar AH, Atkin OK, Menz RI, Henry B, Farquhar G, Day DA (1998) Analysis of respiratory chain regulation in roots of soybean seedlings. Plant Physiology 117: 1083-1093 Millar AH, Eubel H, Jansch L, Kruft V, Heazlewood JL, Braun HP (2004) Mitochondrial cytochrome c oxidase and succinate dehydrogenase complexes contain plant specific subunits. Plant Molecular Biology 56: 77-90 Millar AH, Hill SA, Leaver CJ (1999) Plant mitochondrial 2-oxoglutarate dehydrogenase complex: purification and characterization in potato. Biochemical Journal 343: 327- 334 Millar AH, Hoefnagel MHN, Day DA, Wiskich JT (1996) Specificity of the organic acid activation of alternative oxidase in plant mitochondria. Plant Physiology 111: 613- 618 Millar AH, Knorpp C, Leaver CJ, Hill SA (1998) Plant mitochondrial pyruvate dehydrogenase complex: purification and identification of catalytic components in potato. Biochemical Journal 334: 571-576 Millar AH, Sweetlove LJ, Giege P, Leaver CJ (2001) Analysis of the Arabidopsis mitochondrial proteome. Plant Physiology 127: 1711-1727

288

References

Millar AH, Whelan J, Soole KL, Day DA (2011) Organization and regulation of mitochondrial respiration in plants. In SS Merchant, WR Briggs, D Ort, eds, Annual Review of Plant Biology, Vol 62, Vol 62. Annual Reviews, Palo Alto, pp 79-104 Mitchell-Olds T (2001) Arabidopsis thaliana and its wild relatives: a model system for ecology and evolution. Trends in Ecology and Evolution 16: 693-700 Mitchell P (1961) Coupling of phosphorylation to electron and hydrogen transfer by a chemi-osmotic type of mechanism. Nature 191: 144-148 Miyashita Y, Good AG (2008) Glutamate deamination by glutamate dehydrogenase plays a central role in amino acid catabolism in plants. Plant Signaling & Behavior 3: 842- 843 Mizuno N, Sugie A, Kobayashi F, Takumi S (2008) Mitochondrial alternative pathway is associated with development of freezing tolerance in common wheat. Journal of Plant Physiology 165: 462-467 Moeder W, del Pozo O, Navarre DA, Martin GB, Klessig DF (2007) Aconitase plays a role in regulating resistance to oxidative stress and cell death in Arabidopsis and Nicotiana benthamiana. Plant Molecular Biology 63: 273-287 Mogensen VO (1977) Field measurements of dark respiration rates of roots and aerial parts in italian ryegrass and barley. Journal of Applied Ecology 14: 243-252 Mohamed-Yasseen Y (1991) Onion seed aging and plant regeneration in vitro. PhD dissertation. University of Illinois, Urbana-Champaign Mooney BP, Miernyk JA, Randall DD (2002) The complex fate of alpha-ketoacids. Annual Review of Plant Biology 53: 357-375 Mosmann T (1983) Rapid colorimetric assay for cellular growth and survival - application to proliferation and cyto-toxicity assays. Journal of Immunological Methods 65: 55-63 Musrati RA, Kollarova M, Mernik N, Mikulasova D (1998) Malate dehydrogenase: distribution, function and properties. General Physiology and Biophysics 17: 193- 210 Mutwil M, Usadel B, Schutte M, Loraine A, Ebenhoh O, Persson S (2010) Assembly of an interactive correlation network for the Arabidopsis genome using a novel heuristic clustering algorithm. Plant Physiology 152: 29-43 Nakayama Y, Saio K, Kito M (1981) Decomposition of phospholipids in soybeans during storage. Cereal Chemistry 58: 260-264 Narsai R, Law SR, Carrie C, Xu L, Whelan J (2011) In-depth temporal transcriptome profiling reveals a crucial developmental switch with roles for RNA processing and organelle metabolism that are essential for germination in Arabidopsis. Plant Physiology 157: 1342-1362 Nelson T, Dengler N (1997) Leaf vascular pattern formation. Plant Cell 9: 1121-1135 Neuburger M, Douce R (1980) Effect of bicarbonate and oxaloacetate on malate oxidation by spinach leaf mitochondria. Biochimica Et Biophysica Acta 589: 176-189 Noctor G, De Paepe R, Foyer CH (2007) Mitochondrial redox biology and homeostasis in plants. Trends in Plant Science 12: 125-134 Nooden LD, Blakley KA, Grzybowski JM (1985) Control of seed coat thickness and permeability in soybean - a possible adaptation to stress. Plant Physiology 79: 543- 545 Noren H, Svensson P, Andersson B (1999) Auxiliary photosynthetic functions of Arabidopsis thaliana - studies in vitro and in vivo. Bioscience Reports 19: 499-509 Nunes-Nesi A, Araujo WL, Obata T, Fernie AR (2013) Regulation of the mitochondrial tricarboxylic acid cycle. Current Opinion in Plant Biology 16: 335-343

289

References

Nunes-Nesi A, Carrari F, Gibon Y, Sulpice R, Lytovchenko A, Fisahn J, Graham J, Ratcliffe RG, Sweetlove LJ, Fernie AR (2007) Deficiency of mitochondrial fumarase activity in tomato plants impairs photosynthesis via an effect on stomatal function. Plant Journal 50: 1093-1106 Nunes-Nesi A, Carrari F, Lytovchenko A, Smith AMO, Loureiro ME, Ratcliffe RG, Sweetlove LJ, Fernie AR (2005) Enhanced photosynthetic performance and growth as a consequence of decreasing mitochondrial malate dehydrogenase activity in transgenic tomato plants. Plant Physiology 137: 611-622 Nunes-Nesi A, Fernie AR (2007) Mitochondrial metabolism. In Annual Plant Reviews Volume 31: Plant Mitochondria. John Wiley & Sons, Inc., pp 212-277 O'Malley RC, Ecker JR (2010) Linking genotype to phenotype using the Arabidopsis unimutant collection. Plant Journal 61: 928-940 O'Sullivan SA, Wedding RT (1972) Malate dehydrogenase isoenzymes in cotton leaves of different ages. Physiologia Plantarum 26: 34-38 Obata T, Fernie AR (2012) The use of metabolomics to dissect plant responses to abiotic stresses. Cellular and Molecular Life Sciences 69: 3225-3243 Oleksyn J, Modrzynski J, Tjoelker MG, Zytkowiak R, Reich PB, Karolewski P (1998) Growth and physiology of Picea abies populations from elevational transects: common garden evidence for altitudinal ecotypes and cold adaptation. Functional Ecology 12: 573-590 Oliva M, Dunand C (2007) Waving and skewing: how gravity and the surface of growth media affect root development in Arabidopsis. New Phytologist 176: 37-43 Oliver DJ (1994) The glycine decarboxylase complex from plant mitochondria. Annual Review of Plant Physiology and Plant Molecular Biology 45: 323-337 Oliver DJ, Neuburger M, Bourguignon J, Douce R (1990) Glycine metabolism by plant mitochondria. Physiologia Plantarum 80: 487-491 Owen JH, Laybournparry JEM, Wellburn AR (1986) Leaf respiration during early plastidogenesis in light-grown barley seedlings. Physiologia Plantarum 68: 100-106 Padmasree K, Padmavathi L, Raghavendra AS (2002) Essentiality of mitochondrial oxidative metabolism for photosynthesis: optimization of carbon assimilation and protection against photoinhibition. Critical Reviews in Biochemistry and Molecular Biology 37: 71-119 Pastore D, Di Pede S, Passarella S (2003) Isolated durum wheat and potato cell mitochondria oxidize externally added NADH mostly via the malate/oxaloacetate shuttle with a rate that depends on the carrier-mediated transport. Plant Physiology 133: 2029-2039 Penfield S, Graham S, Graham I (2005) Storage reserve mobilization in germinating oilseeds: Arabidopsis as a model system. Biochemical Society Transactions 33: 380– 383 Perez-Perez JM, Serrano-Cartagena J, Micol JL (2002) Genetic analysis of natural variations in the architecture of Arabidopsis thaliana vegetative leaves. Genetics 162: 893-915 Picotti P, Bodenmiller B, Aebersold R (2013) Proteomics meets the scientific method. Nature Methods 10: 24-27 Pigliucci M, Schlichting CD (1995) Reaction norms of Arabidopsis (Brassicaceae) .3. Response to nutrients in 26 populations from a worldwide collection. American Journal of Botany 82: 1117-1125 Plieva F, Galaev I, Mattiasson B (2009) Macroporous polysaccharide gels. Macroporous Polymers: Production Properties and Biotechnological/Biomedical Applications, CRC Press, USA: 131-154

290

References

Porterfield DM, Kuang AX, Smith PJS, Crispi ML, Musgrave ME (1999) Oxygen-depleted zones inside reproductive structures of Brassicaceae: implications for oxygen control of seed development. Canadian Journal of Botany-Revue Canadienne De Botanique 77: 1439-1446 Pourcel L, Routaboul JM, Kerhoas L, Caboche M, Lepiniec L, Debeaujon I (2005) TRANSPARENT TESTA10 encodes a laccase-like enzyme involved in oxidative polymerization of flavonoids in Arabidopsis seed coat. Plant Cell 17: 2966-2980 Powell AA, Don R, Haigh P, Phillips G, Tonkin JHB, Wheaton OE (1984) Assessment of the repeatability of the controlled deterioration vigor test both within and between laboratories. Seed Science and Technology 12: 421-427 Powell AA, Matthews S (1981) Evaluation of controlled deterioration, a new vigor test for small seeded vegetables. Seed Science and Technology 9: 633-640 Pracharoenwattana I, Cornah JE, Smith SM (2007) Arabidopsis peroxisomal malate dehydrogenase functions in beta-oxidation but not in the glyoxylate cycle. Plant Journal 50: 381-390 Pracharoenwattana I, Zhou WX, Keech O, Francisco PB, Udomchalothorn T, Tschoep H, Stitt M, Gibon Y, Smith SM (2010) Arabidopsis has a cytosolic fumarase required for the massive allocation of photosynthate into fumaric acid and for rapid plant growth on high nitrogen. Plant Journal 62: 785-795 Pradet A, Raymond P (1983) Adenine nucleotide ratios and adenylate energy charge in energy metabolism. Annual Review of Plant Physiology and Plant Molecular Biology 34: 199-224 Prasch CM, Sonnewald U (2013) In silico selection of Arabidopsis thaliana ecotypes with enhanced stress tolerance. Plant Signaling and Behavior 8: e26364 Pregitzer KS, Laskowski MJ, Burton AJ, Lessard VC, Zak DR (1998) Variation in sugar maple root respiration with root diameter and soil depth. Tree Physiology 18: 665-670 Quint M, Barkawi LS, Fan KT, Cohen JD, Gray WM (2009) Arabidopsis IAR4 modulates auxin response by regulating auxin homeostasis. Plant Physiology 150: 748-758 Rahmatullah M, Gopalakrishnan S, Radke GA, Roche TE (1989) Domain structures of the dihydrolipoyl transacetylase and the protein-x components of mammalian pyruvate dehydrogenase complex - selective cleavage by protease ARG C. Journal of Biological Chemistry 264: 1245-1251 Rajjou L, Lovigny Y, Groot SPC, Belghaz M, Job C, Job D (2008) Proteome-wide characterization of seed aging in Arabidopsis: a comparison between artificial and natural aging protocols. Plant Physiology 148: 620-641 Randall DD, Miernyk JA, David NR, Gemel J, Luethy MH (1996) Regulation of leaf mitochondrial pyruvate dehydrogenase complex activity by reversible phosphorylation. Clarendon Press, Oxford Rangel-Fajardo MA, Delgado-Alvarado A, Cordova-Tellez L, Zavaleta-Mancera HA, Villegas-Monter A, Lopez-Andrade AP, Vidales-Fernandez I (2011) Accumulation of sugars in cacao (Theobroma cacao L.) seeds of three genetic origins and its relationship to desiccation tolerance. Seed Science and Technology 39: 1-11 Rasmusson AG, Agius SC (2001) Rotenone-insensitive NAD(P)H dehydrogenases in plants: Immunodetection and distribution of native proteins in mitochondria. Plant Physiology and Biochemistry 39: 1057-1066 Rasmusson AG, Soole KL, Elthon TE (2004) Alternative NAD(P)H dehydrogenases of plant mitochondria. Annual Review of Plant Biology 55: 23-39 Rasmusson AG, Wallstrom SV (2010) Involvement of mitochondria in the control of plant cell NAD(P)H reduction levels. Biochemical Society Transactions 38: 661-666

291

References

Redei GP (1975) Arabidopsis as a genetic tool. Annual Review of Genetics 9: 111-127 Rees DM, Leslie AGW, Walker JE (2009) The structure of the membrane extrinsic region of bovine ATP synthase. Proceedings of the National Academy of Sciences of the United States of America 106: 21597-21601 Reggiani R, Cantu CA, Brambilla I, Bertani A (1988) Accumulation and interconversion of amino-acids in rice roots under anoxia. Plant and Cell Physiology 29: 981-987 Remacle C, Barbieri MR, Cardol P, Hamel PP (2008) Eukaryotic Complex I: functional diversity and experimental systems to unravel the assembly process. Molecular Genetics and Genomics 280: 93-110 Rex Sheu KF, Blass JP (1999) The α-ketoglutarate dehydrogenase complex. Annals of the New York Academy of Sciences 893: 61-78 Reyes JL, Chua NH (2007) ABA induction of miR159 controls transcript levels of two MYB factors during Arabidopsis seed germination. Plant Journal 49: 592-606 Rhee SY, Beavis W, Berardini TZ, Chen GH, Dixon D, Doyle A, Garcia-Hernandez M, Huala E, Lander G, Montoya M, Miller N, Mueller LA, Mundodi S, Reiser L, Tacklind J, Weems DC, Wu YH, Xu I, Yoo D, Yoon J, Zhang PF (2003) The Arabidopsis Information Resource (TAIR): a model organism database providing a centralized, curated gateway to Arabidopsis biology, research materials and community. Nucleic Acids Research 31: 224-228 Roberts EH (1972) Loss of viability and crop yields. Chapman and Hall, London. Robertson JA, Morrison WH, Burdick D (1973) Chemical evaluation of oil from field damaged and storage damaged soybeans. Journal of the American Oil Chemists Society 50: 443-445 Robinson GL, Dinsdale D, MacFarlane M, Cain K (2012) Switching from aerobic glycolysis to oxidative phosphorylation modulates the sensitivity of mantle cell lymphoma cells to TRAIL. Oncogene 31: 4996-5006 Rodrigues F, Ludovico P, Leao C (2006) Sugar metabolism in yeasts: an overview of aerobic and anaerobic glucose catabolism. Biodiversity and Ecophysiology of Yeasts: 101- 121 Rogers GW, Brand MD, Petrosyan S, Ashok D, Elorza AA, Ferrick DA, Murphy AN (2011) High throughput microplate respiratory measurements using minimal quantities of isolated mitochondria. Plos One 6 Rolletschek H, Borisjuk L, Koschorreck M, Wobus U, Weber H (2002) Legume embryos develop in a hypoxic environment. Journal of Experimental Botany 53: 1099-1107 Rolletschek H, Radchuk R, Klukas C, Schreiber F, Wobus U, Borisjuk L (2005) Evidence of a key role for photosynthetic oxygen release in oil storage in developing soybean seeds. New Phytologist 167: 777-786 Rolletschek H, Weber H, Borisjuk L (2003) Energy status and its control on embryogenesis of legumes. Embryo photosynthesis contributes to oxygen supply and is coupled to biosynthetic fluxes. Plant Physiology 132: 1196-1206 Roschzttardtz H, Fuentes I, Vasquez M, Corvalan C, Leon G, Gomez I, Araya A, Holuigue L, Vicente-Carbajosa J, Jordana X (2009) A nuclear gene encoding the iron-sulfur subunit of mitochondrial Complex II is regulated by B3 domain transcription factors during seed development in Arabidopsis. Plant Physiology 150: 84-95 Routaboul JM, Kerhoas L, Debeaujon I, Pourcel L, Caboche M, Einhorn J, Lepiniec L (2006) Flavonoid diversity and biosynthesis in seed of Arabidopsis thaliana. Planta 224: 96- 107 Rudolph N, Esser HG, Carminati A, Moradi AB, Hilger A, Kardjilov N, Nagl S, Oswald SE (2012) Dynamic oxygen mapping in the root zone by fluorescence dye imaging combined with neutron radiography. Journal of Soils and Sediments 12: 63-74

292

References

Ruuska SA, Girke T, Benning C, Ohlrogge JB (2002) Contrapuntal networks of gene expression during Arabidopsis seed filling. Plant Cell 14: 1191-1206 Ruuska SA, Schwender J, Ohlrogge JB (2004) The capacity of green oilseeds to utilize photosynthesis to drive biosynthetic processes. Plant Physiology 136: 2700-2709 Ryan DG, Lin TW, Brownie E, Bridger WA, Wolodko WT (1997) Mutually exclusive splicing generates two distinct isoforms of pig heart succinyl-CoA synthetase. Journal of Biological Chemistry 272: 21151-21159 Ryan PR, Delhaize E, Randall PJ (1995) Malate efflux from root apices and tolerance to aluminum are highly correlated in wheat. Australian Journal of Plant Physiology 22: 531-536 Sabar M, Gagliardi D, Balk J, Leaver CJ (2003) ORFB is a subunit of F1FO-ATP synthase: insight into the basis of cytoplasmic male sterility in sunflower. EMBO Reports 4: 381-386 Saeed AI, Sharov V, White J, Li J, Liang W, Bhagabati N, Braisted J, Klapa M, Currier T, Thiagarajan M, Sturn A, Snuffin M, Rezantsev A, Popov D, Ryltsov A, Kostukovich E, Borisovsky I, Liu Z, Vinsavich A, Trush V, Quackenbush J (2003) TM4: A free, open-source system for microarray data management and analysis. Biotechniques 34: 374 Saldana A, Lusk CH, Gonzales WL, Gianoli E (2007) Natural selection on ecophysiological traits of a fern species in a temperate rainforest. Evolutionary Ecology 21: 651-662 Sanda S, John M, Amasino R (1997) Analysis of flowering time in ecotypes of Arabidopsis thaliana. Journal of Heredity 88: 69-72 Sarojini G, Oliver DJ (1983) Extraction and partial characterization of the glycine decarboxylase multienzyme complex from pea leaf mitochondria. Plant Physiology 72: 194-199 Scandroglio F, Tortora V, Radi R, Castro L (2014) Metabolic control analysis of mitochondrial aconitase: influence over respiration and mitochondrial superoxide and hydrogen peroxide production. Free Radical Research 48: 684-693 Scheibe R (2004) Malate valves to balance cellular energy supply. Physiologia Plantarum 120: 21-26 Scheibe R, Backhausen JE, Emmerlich V, Holtgrefe S (2005) Strategies to maintain redox homeostasis during photosynthesis under changing conditions. Journal of Experimental Botany 56: 1481-1489 Schmid M, Davison TS, Henz SR, Pape UJ, Demar M, Vingron M, Scholkopf B, Weigel D, Lohmann JU (2005) A gene expression map of Arabidopsis thaliana development. Nature Genetics 37: 501-506 Schmitz AJ, Glynn JM, Olson BJSC, Stokes KD, Osteryoung KW (2009) Arabidopsis FtsZ2-1 and FtsZ2-2 are functionally redundant, but FtsZ-based plastid division is not essential for chloroplast partitioning or plant growth and development. Molecular Plant 2: 1211-1222 Schneider CA, Rasband WS, Eliceiri KW (2012) NIH Image to ImageJ: 25 years of image analysis. Nature Methods 9: 671-675 Schwender J, Shachar-Hill Y, Ohlrogge JB (2006) Mitochondrial metabolism in developing embryos of Brassica napus. Journal of Biological Chemistry 281: 34040-34047 Searle SYY (2010) The temperature dependence of plant alternative oxidase and its impact on respiration rates in nature. Selinski J, Koenig N, Wellmeyer B, Hanke GT, Linke V, Neuhaus HE, Scheibe R (2014) The plastid-localized NAD-dependent malate dehydrogenase is crucial for energy homeostasis in developing Arabidopsis thaliana seeds. Molecular Plant 7: 170-186

293

References

Sew YS, Millar AH, Stroeher E (2015) Micro-respiratory measurements in plants. Plant Mitochondria: Methods and Protocols: 187-196 Sew YS, Stroeher E, Holzmann C, Huang S, Taylor N, Jordana X, Millar AH (2013) Multiplex micro-respiratory measurements of Arabidopsis tissues. New Phytologist 200: 922- 932 Shane MW, Cramer MD, Funayama-Noguchi S, Cawthray GR, Millar AH, Day DA, Lambers H (2004) Development physiology of cluster-root carboxylate synthesis and exudation in harsh hakea. Expression of phosphoenolpyruvate carboxylase and the alternative oxidase. Plant Physiology 135: 549-560 Sheu KFR, Blass JP (1999) The alpha-ketoglutarate dehydrogenase complex. In JP Blass, FH McDowell, eds, Oxidative/Energy Metabolism in Neurodegenerative Disorders, Vol 893. New York Acad Sciences, New York, pp 61-78 Shindo C, Bernasconi G, Hardtke CS (2007) Natural genetic variation in Arabidopsis: tools, traits and prospects for evolutionary ecology. Annals of Botany 99: 1043-1054 Shone MGT, Gale J (1983) Effect of sodium-chloride stress and nitrogen-source on respiration, growth and photosynthesis in lucerne (Medicago sativa L.). Journal of Experimental Botany 34: 1117-1125 Sieger SM, Kristensen BK, Robson CA, Amirsadeghi S, Eng EWY, Abdel-Mesih A, Møller IM, Vanlerberghe GC (2005) The role of alternative oxidase in modulating carbon use efficiency and growth during macronutrient stress in tobacco cells. Journal of Experimental Botany 56: 1499-1515 Sienkiewicz-Porzucek A, Nunes-Nesi A, Sulpice R, Lisec J, Centeno DC, Carillo P, Leisse A, Urbanczyk-Wochniak E, Fernie AR (2008) Mild reductions in mitochondrial citrate synthase activity result in a compromised nitrate assimilation and reduced leaf pigmentation but have no effect on photosynthetic performance or growth. Plant Physiology 147: 115-127 Sienkiewicz-Porzucek A, Sulpice R, Osorio S, Krahnert I, Leisse A, Urbanczyk-Wochniak E, Hodges M, Fernie AR, Nunes-Nesi A (2010) Mild reductions in mitochondrial NAD- dependent isocitrate dehydrogenase activity result in altered nitrate assimilation and pigmentation but do not impact growth. Molecular Plant 3: 156-173 Silverman HG, Roberto FF (2007) Understanding marine mussel adhesion. Marine Biotechnology 9: 661-681 Smith BN, Lytle CM, Hansen LD (1995) Predicting plant growth rates from dark respiration rates: an experimental approach. Proceedings: Wildland Shrub and Arid Land Restoration Symposium 315: 243-245 Solheim C, Li L, Hatzopoulos P, Millar AH (2012) Loss of Lon1 in Arabidopsis changes the mitochondrial proteome leading to altered metabolite profiles and growth retardation without an accumulation of oxidative damage. Plant Physiology 160: 1187-1203 Somerville C (1986) Analysis of photosynthesis with mutants of higher plants and algae. Annual review of plant physiology 37: 467-506 Somerville C, Koornneef M (2002) Timeline - a fortunate choice: the history of Arabidopsis as a model plant. Nature Reviews Genetics 3: 883-889 Somerville CR, Ogren WL (1980) Inhibition of photosynthesis in Arabidopsis mutants lacking leaf glutamate synthase activity. Nature 286: 257-259 Somerville CR, Ogren WL (1981) Photorespiration deficient mutants of Arabidopsis thaliana lacking mitochondrial serine transhydroxymethylase activity. Plant Physiology 67: 666-671 Souza FHDDE, Marcos-Filho J (2001) The seed coat as a modulator of seed-environment relationships in Fabaceae. Brazilian Journal of Botany 24: 365-375

294

References

Spano C, Bottega S, Lorenzi R, Grilli I (2011) Ageing in embryos from wheat grains stored at different temperatures: oxidative stress and antioxidant response. Functional Plant Biology 38: 624-631 Stadtman ER, Levine RL (2003) Free radical mediated oxidation of free amino acids and amino acid residues in proteins. Amino Acids 25: 207-218 Stasolla C, Loukanina N, Ashihara H, Yeung EC, Thorpe TA (2003) Pyrimidine deoxyribonucleotide metabolism during maturation and germination of white spruce (Picea glauca) somatic embryos: metabolic fate of C-labelled cytidine, deoxycytidine and thymidine. Physiologia Plantarum 118: 499-506 Stewart RRC, Bewley JD (1980) Lipid peroxidation associated with accelerated aging of soybean axes. Plant Physiology 65: 245-248 Stinchcombe JR, Weinig C, Ungerer M, Olsen KM, Mays C, Halldorsdottir SS, Purugganan MD, Schmitt J (2004) A latitudinal cline in flowering time in Arabidopsis thaliana modulated by the flowering time gene FRIGIDA. Proceedings of the National Academy of Sciences of the United States of America 101: 4712-4717 Stratton DA (1998) Reaction norm functions and QTL-environment interactions for flowering time in Arabidopsis thaliana. Heredity 81: 144-155 Street HE, Elliot MC, Fowler MW, eds (1976) The physilogy of root. Elsevier, Amsterdam Studart-Guimaraes C, Fait A, Nunes-Nesi A, Carrari F, Usadel B, Fernie AR (2007) Reduced expression of succinyl-coenzyme A ligase can be compensated for by up-regulation of the gamma-aminobutyrate shunt in illuminated tomato leaves. Plant Physiology 145: 626-639 Sugimoto M, Morohashi Y (1989) Comparison of the developmental patterns of mitochondrial malate dehydrogenase between mung bean and cucumber cotyledons during and following germination. Physiologia Plantarum 77: 238-243 Sulpice R, Nikoloski Z, Tschoep H, Antonio C, Kleessen S, Larhlimi A, Selbig J, Ishihara H, Gibon Y, Fernie AR, Stitt M (2013) Impact of the carbon and nitrogen supply on relationships and connectivity between metabolism and biomass in a broad panel of Arabidopsis accessions. Plant Physiology 162: 347-363 Sun WQ, Leopold AC (1995) The maillard reaction and oxidative stress during aging of soybean seeds. Physiologia Plantarum 94: 94-104 Suzuki Y, Kawazu T, Tsuyama M, Wada T, Kondo K, Mizuno R, Hara T, Koyama H (2004) Characteristics of transgenic Eucalyptus hybrids with an overexpression of a plant mitochondrial citrate synthase. Plant and Cell Physiology 45: S34-S34 Sweetlove LJ, Beard KFM, Nunes-Nesi A, Fernie AR, Ratcliffe RG (2010) Not just a circle: flux modes in the plant TCA cycle. Trends in Plant Science 15: 462-470 Sweetlove LJ, Heazlewood JL, Herald V, Holtzapffel R, Day DA, Leaver CJ, Millar AH (2002) The impact of oxidative stress on Arabidopsis mitochondria. Plant Journal 32: 891- 904 Sylvester AW, Smith L, Freeling M (1996) Acquisition of identity in the developing leaf. Annual Review of Cell and Developmental Biology 12: 257-304 Takeba G (1980) Changes revealed by a tracer technique in the amino acid metabolism of thermodormant and non-dormant New York lettuce seeds. Plant and Cell Physiology 21: 1627-1638 Takeba G (1980) Phytochrome-mediated accumulation of free amino acids in embryonic axes of New York lettuce seeds. Plant and Cell Physiology 21: 1651-1656 Tang Y, Garvin DF, Kochian LV, Sorrells ME, Carver BF (2002) Physiological genetics of aluminum tolerance in the wheat cultivar Atlas 66. Crop Science 42: 1541-1546

295

References

Taniguchi M, Miyake H (2012) Redox-shuttling between chloroplast and cytosol: integration of intra-chloroplast and extra-chloroplast metabolism. Current Opinion in Plant Biology 15: 252-260 Tanz SK, Castleden I, Hooper CM, Vacher M, Small I, Millar AH (2013) SUBA3: a database for integrating experimentation and prediction to define the SUBcellular location of proteins in Arabidopsis. Nucleic Acids Research 41: D1185-D1191 Taylor NL, Fenske R, Castleden I, Tomaz T, Nelson CJ, Millar AH (2014) Selected reaction monitoring to determine protein abundance in Arabidopsis using the Arabidopsis proteotypic predictor. Plant Physiology 164: 525-536 Taylor NL, Heazlewood JL, Day DA, Millar AH (2004) Lipoic acid-dependent oxidative catabolism of alpha-keto acids in mitochondria provides evidence for branched- chain amino acid catabolism in Arabidopsis. Plant Physiology 134: 838-848 Taylor NL, Heazlewood JL, Millar AH (2011) The Arabidopsis thaliana 2-D gel mitochondrial proteome: refining the value of reference maps for assessing protein abundance, contaminants and post-translational modifications. Proteomics 11: 1720-1733 Taylor NL, Howell KA, Heazlewood JL, Tan TYW, Narsai R, Huang S, Whelan J, Millar AH (2010) Analysis of the rice mitochondrial carrier family reveals anaerobic accumulation of a basic amino acid carrier involved in arginine metabolism during seed germination. Plant Physiology 154: 691-704 Tesfaye M, Temple SJ, Allan DL, Vance CP, Samac DA (2001) Overexpression of malate dehydrogenase in transgenic alfalfa enhances organic acid synthesis and confers tolerance to aluminum. Plant Physiology 127: 1836-1844 Tesnier K, Strookman-Donkers HM, Van Pijlen JG, Van der Geest AHM, Bino RJ, Groot SPC (2002) A controlled deterioration test for Arabidopsis thaliana reveals genetic variation in seed quality. Seed Science and Technology 30: 149-165 Tezuka T, Yokoyama H, Tanaka H, Shiozaki S, Oda M (2012) Seed and embryo germination in Ardisia crenata. Journal of Botany 2012: 7 Thelen JJ, Muszynski MG, David NR, Luethy MH, Elthon TE, Miernyk JA, Randall DD (1999) The dihydrolipoamide S-acetyltransferase subunit of the mitochondrial pyruvate dehydrogenase complex from maize contains a single lipoyl domain. Journal of Biological Chemistry 274: 21769-21775 Thornley JHM, Cannell MGR (2000) Modelling the components of plant respiration: Representation and realism. Annals of Botany 85: 55-67 Tiwari BS, Belenghi B, Levine A (2002) Oxidative stress increased respiration and generation of reactive oxygen species, resulting in ATP depletion, opening of mitochondrial permeability transition, and programmed cell death. Plant Physiology 128: 1271-1281 Tomaz T (2012) Mitochondrial control of global REDOX balance in Arabidopsis: insights gained from manipulation of the TCA cycle and ascorbate biosynthesis pathway. University of Western Australia Tomaz T, Bagard M, Pracharoenwattana I, Linden P, Lee CP, Carroll AJ, Stroeher E, Smith SM, Gardestrom P, Millar AH (2010) Mitochondrial malate dehydrogenase lowers leaf respiration and alters photorespiration and plant growth in Arabidopsis. Plant Physiology 154: 1143-1157 Tovar-Mendez A, Miernyk JA, Randall DD (2003) Regulation of pyruvate dehydrogenase complex activity in plant cells. European Journal of Biochemistry 270: 1043-1049 Tschiersch H, Liebsch G, Borisjuk L, Stangelmayer A, Rolletschek H (2012) An imaging method for oxygen distribution, respiration and photosynthesis at a microscopic level of resolution. New Phytologist 196: 926-936

296

References

Tsror L, Aharon M, Erlich O (1999) Survey of bacterial and fungal seedborne diseases in imported and domestic potato seed tubers. Phytoparasitica 27: 215-226 Ueda K, Tsuji H (1971) Ultrastructural changes of organelles in coleoptile cells during anaerobic germination of rice seeds. Protoplasma 73: 203-215 Van Aken O, Pecenkova T, van de Cotte B, De Rycke R, Eeckhout D, Fromm H, De Jaeger G, Witters E, Beemster GTS, Inze D, Van Breusegem F (2007) Mitochondrial type-I prohibitins of Arabidopsis thaliana are required for supporting proficient meristem development. Plant Journal 52: 850-864 Van der Merwe MJ, Osorio S, Moritz T, Nunes-Nesi A, Fernie AR (2009) Decreased mitochondrial activities of malate dehydrogenase and fumarase in tomato lead to altered root growth and architecture via diverse mechanisms. Plant Physiology 149: 653-669 Van Iersel MW (2003) Carbon use efficiency depends on growth respiration, maintenance respiration, and relative growth rate. A case study with lettuce. Plant Cell and Environment 26: 1441-1449 Vanlerberghe GC (2013) Alternative oxidase: a mitochondrial respiratory pathway to maintain metabolic and signaling homeostasis during abiotic and biotic stress in plants. International Journal of Molecular Sciences 14: 6805-6847 Vanlerberghe GC, McIntosh L (1992) Lower growth temperature increases alternative pathway capacity and alternative oxidase protein in tobacco. Plant Physiology 100: 115-119 Vanlerberghe GC, Vanlerberghe AE, McIntosh L (1997) Molecular genetic evidence of the ability of alternative oxidase to support respiratory carbon metabolism. Plant Physiology 113: 657-661 Vercesi AE, Borecky J, Godoy Maia ID, Arruda P, Cuccovia IM, Chaimovich H (2006) Plant uncoupling mitochondrial proteins. In Annual Review of Plant Biology, Vol 57. Annual Reviews, Palo Alto, pp 383-404 Vertucci CW (1989) The effects of low water contents on physiological activities of seeds. Physiologia Plantarum 77: 172-176 Vonuexkull HR, Mutert E (1995) Global extent, development and economic-impact of acid soils. Plant and Soil 171: 1-15 Wagner AM, Vandenbergen CWM, Wincencjusz H (1995) Stimulation of the alternative pathway by succinate and malate. Plant Physiology 108: 1035-1042 Walk RA, Hock B (1976) Mitochondrial malate dehydrogenase of watermelon cotyledons: Time course and mode of enzyme activity changes during germination. Planta 129: 27-32 Walk RA, Michaeli S, Hock B (1977) Glyoxysomal and mitochondrial malate-dehydrogenase of watermelon (Citrullus vulgaris) cotyledons. 1. Molecular-properties of purified isoenzymes. Planta 136: 211-220 Walker D, Walker R (1987) The use of the oxygen electrode and fluorescence probes in simple measurements of photosynthesis. Wang Y, Yang L, Zheng Z, Grumet R, Loescher W, Zhu JK, Yang P, Hu Y, Chan ZL (2013) Transcriptomic and physiological variations of three Arabidopsis ecotypes in response to salt stress. Plos One 8: e69036 Waskiewicz DE, Hammes GG (1984) Elementary steps in the reaction mechanism of the alpha-ketoglutarate dehydrogenase multienzyme complex from Escherichia coli: kinetics of succinylation and desuccinylation. Biochemistry 23: 3136-3143 Weigel D, Mott R (2009) The 1001 genomes project for Arabidopsis thaliana. Genome Biology 10: 107

297

References

Weitbrecht K, Muller K, Leubner-Metzger G (2011) First off the mark: early seed germination. Journal of Experimental Botany 62: 3289-3309 Welchen E, Hildebrandt TM, Lewejohann D, Gonzalez DH, Braun HP (2012) Lack of cytochrome c in Arabidopsis decreases stability of Complex IV and modifies redox metabolism without affecting Complexes I and III. Biochimica Et Biophysica Acta- Bioenergetics 1817: 990-1001 Weng JK, Chapple C (2010) The origin and evolution of lignin biosynthesis. New Phytologist 187: 273-285 Westoby M, Jurado E, Leishman M (1992) Comparative evolutionary ecology of seed size. Trends in Ecology & Evolution 7: 368-372 Whittle CA, Otto SP, Johnston MO, Krochko JE (2009) Adaptive epigenetic memory of ancestral temperature regime in Arabidopsis thaliana. Botany-Botanique 87: 650- 657 Wienkoop S, Weckwerth W (2006) Relative and absolute quantitative shotgun proteomics: targeting low-abundance proteins in Arabidopsis thaliana. Journal of Experimental Botany 57: 1529-1535 Williams GRC, Doran PM (1999) Investigation of liquid-solid hydrodynamic boundary layers and oxygen requirements in hairy root cultures. Biotechnology and Bioengineering 64: 729-740 Williams TCR, Miguet L, Masakapalli SK, Kruger NJ, Sweetlove LJ, Ratcliffe RG (2008) Metabolic network fluxes in heterotrophic Arabidopsis cells: Stability of the flux distribution under different oxygenation conditions. Plant Physiology 148: 704-718 Wilson DO, McDonald MB (1986) The lipid peroxidation model of seed aging. Seed Science and Technology 14: 269-300 Wilson KE, Ivanov AG, Öquist G, Grodzinski B, Sarhan F, Huner NP (2006) Energy balance, organellar redox status, and acclimation to environmental stress. Botany 84: 1355- 1370 Wingler A, Lea PJ, Quick WP, Leegood RC (2000) Photorespiration: metabolic pathways and their role in stress protection. Philosophical Transactions of the Royal Society B: Biological Sciences 355: 1517-1529 Winter D, Vinegar B, Nahal H, Ammar R, Wilson GV, Provart NJ (2007) An "Electronic Fluorescent Pictograph" browser for exploring and analyzing large scale biological data sets. Plos One 2: e718 Wiskich JT, Bryce JH, Day DA, Dry IB (1990) Evidence for metabolic domains within the matrix compartment of pea leaf mitochondria - implications for photorespiratory metabolism. Plant Physiology 93: 611-616 Wiskich JT, Dry IB (1985) The tricarboxylic acid cycle in plant mitochondria: its operation and regulation. In R Douce, D Day, eds, Higher Plant Cell Respiration, Vol 18. Springer Berlin Heidelberg, pp 281-313 Woodstock LW, Grabe DF (1967) Relationships between seed respiration during imbibition and subsequent seedling growth in Zea mays l. Plant Physiology 42: 1071-1076 Wyatt S, Ballard HE (2007) Arabidopsis ecotypes: a model for course projects in organismal plant biology and evolution. American Biology Teacher 69: 477-481 Xin X, Lin X-H, Zhou Y-C, Chen X-L, Liu X, Lu X-X (2011) Proteome analysis of maize seeds: the effect of artificial ageing. Physiologia Plantarum 143: 126-138 Xin X, Tian Q, Yin G, Chen X, Zhang J, Ng S, Lu X (2014) Reduced mitochondrial and ascorbate-glutathione activity after artificial ageing in soybean seed. Journal of Plant Physiology 171: 140-147

298

References

Yang XY, Chen ZW, Xu T, Qu Z, Pan XD, Qin XH, Ren DT, Liu GQ (2011) Arabidopsis kinesin KP1 specifically interacts with VDAC3, a mitochondrial protein, and regulates respiration during seed germination at low temperature. Plant Cell 23: 1093-1106 Yankovskaya V, Horsefield R, Tornroth S, Luna-Chavez C, Miyoshi H, Leger C, Byrne B, Cecchini G, Iwata S (2003) Architecture of succinate dehydrogenase and reactive oxygen species generation. Science 299: 700-704 Yarian CS, Toroser D, Sohal RS (2006) Aconitase is the main functional target of aging in the citric acid cycle of kidney mitochondria from mice. Mechanisms of Ageing and Development 127: 79-84 Yentur S, Leopold AC (1976) Respiratory transition during seed germination. Plant Physiology 57: 274-276 Yeung EC, Cavey MJ (1990) Developmental changes in the inner epidermis of the bean seed coat. Protoplasma 154: 45-52 Yi H, Sardesai N, Fujinuma T, Chan C-W, Veena, Gelvin SB (2006) Constitutive expression exposes functional redundancy between the Arabidopsis histone H2A gene HTA1 and other H2A gene family members. Plant Cell 18: 1575-1589 Yoshida K, Chen RM, Tanaka A, Teramoto H, Tanaka R, Timko MP, Tsuji H (1995) Correlated changes in the activity, amount of protein, and abundance of transcript of NADPH - protochlorophyllide oxidoreductase and chlorophyll accumulation during greening of cucumber cotyledons. Plant Physiology 109: 231-238 Yu H, Du X, Zhang F, Zhang F, Hu Y, Liu S, Jiang X, Wang G, Liu D (2012) A mutation in the E2 subunit of the mitochondrial pyruvate dehydrogenase complex in Arabidopsis reduces plant organ size and enhances the accumulation of amino acids and intermediate products of the TCA Cycle. Planta 236: 387-399 Yu Y, Guo G, Lv D, Hu Y, Li J, Li X, Yan Y (2014) Transcriptome analysis during seed germination of elite Chinese bread wheat cultivar Jimai 20. BMC Plant Biology 14 Zaitseva MG, Kasumova IV, Kasumov EA, Borisova MA, Il'chishina NV (2002) Respiration of mitochondria in developing sunflower seeds. Biology Bulletin 29: 555-558 Zeller G (2005) Resequencing data of 20 Arabidopsis ecotypes. Universitat Tubingen Zhang J, Nuebel E, Wisidagama DRR, Setoguchi K, Hong JS, Van Horn CM, Imam SS, Vergnes L, Malone CS, Koehler CM, Teitell MA (2012) Measuring energy metabolism in cultured cells, including human pluripotent stem cells and differentiated cells. Nature Protocols 7: 1068-1085 Zhang Q, Pettolino FA, Dhugga KS, Rafalski JA, Tingey S, Taylor J, Shirley NJ, Hayes K, Beatty M, Abrams SR, Zaharia LI, Burton RA, Bacic A, Fincher GB (2011) Cell wall modifications in maize pulvini in response to gravitational stress. Plant Physiology 156: 2155-2171 Zhang WH, Zhou Y, Dibley KE, Tyerman SD, Furbank RT, Patrick JW (2007) Nutrient loading of developing seeds. Functional Plant Biology 34: 314-331 Zhang Z, Wakabayashi N, Wakabayashi J, Tamura Y, Song WJ, Sereda S, Clerc P, Polster BM, Aja SM, Pletnikov MlV, Kensler TW, Shirihai OS, Iijima M, Hussain MA, Sesaki H (2011) The dynamin-related GTPase Opa1 is required for glucose-stimulated ATP production in pancreatic beta cells. Molecular Biology of the Cell 22: 2235-2245 Zhong F, Yang D, Hao YW, Lin CZ, Jiang Y, Ying WT, Wu SF, Zhu YP, Liu SQ, Yang PY, Qian XH, He FC (2012) Regular patterns for proteome-wide distribution of protein abundance across species. Plos One 7: e32423 Zoglowek C, Kromer S, Heldt HW (1988) Oxaloacetate and malate transport by plant mitochondria. Plant Physiology 87: 109-115

299