Differential induction of mitochondrial machinery by light intensity correlates with changes in respiratory metabolism and photorespiration in rice leaves


  • Shaobai Huang,

    1. ARC Centre of Excellence in Plant Energy Biology and Centre for Comparative Analysis of Biomolecular Networks (CABiN), The University of Western Australia, Crawley, WA, Australia
    Search for more papers by this author
  • Richard P. Jacoby,

    1. ARC Centre of Excellence in Plant Energy Biology and Centre for Comparative Analysis of Biomolecular Networks (CABiN), The University of Western Australia, Crawley, WA, Australia
    Search for more papers by this author
  • Rachel N. Shingaki-Wells,

    1. ARC Centre of Excellence in Plant Energy Biology and Centre for Comparative Analysis of Biomolecular Networks (CABiN), The University of Western Australia, Crawley, WA, Australia
    Search for more papers by this author
  • Lei Li,

    1. ARC Centre of Excellence in Plant Energy Biology and Centre for Comparative Analysis of Biomolecular Networks (CABiN), The University of Western Australia, Crawley, WA, Australia
    Search for more papers by this author
  • A. Harvey Millar

    Corresponding author
    • ARC Centre of Excellence in Plant Energy Biology and Centre for Comparative Analysis of Biomolecular Networks (CABiN), The University of Western Australia, Crawley, WA, Australia
    Search for more papers by this author

Author for correspondence:

A. Harvey Millar

Tel: +61 8 6488 7245

Email: harvey.millar@uwa.edu.au


  • The light responsiveness of mitochondrial function was investigated through changes in mitochondrial composition and metabolism in rice (Oryza sativa) shoots.
  • The mitochondrial proteome and metabolite abundances under low light, (LL, 100 μmol m−2 s−1), and high light (HL, 700 μmol m−2 s−1) were measured along with information on shoot photosynthetic, respiratory and photorespiratory activity.
  • Specific steps in mitochondrial tricarboxylic acid (TCA) cycle metabolism were decreased under HL, correlating with lower respiration rate under HL. The abundance of mitochondrial enzymes in branch chain metabolism was reduced under HL/LL, and correlated with a decrease in the abundance of a range of amino acids in the HL/LL. Mitochondrial nucleoside diphosphate kinase was increased under LL/HL treatments. Significant accumulation of glycine decarboxylase P, T subunits and serine hydroxymethyltransferase occurred in response to light. The abundance of the glycine decarboxylase (GDC) H subunit proteins was not changed by HL/LL treatments, and the abundance of GDC L subunit protein was halved under HL, indicating a change in the stoichiometry of GDC subunits, while photorespiration was fourfold higher in LL- than in HL-treated plants.
  • Insights into these light-dependent phenomena and their importance for understanding the initiation of photorespiration in rice and adaptation of mitochondria to function in photosynthetic cells are discussed.


Photosynthesis and respiration in the plant cell are intimately linked processes; they jointly participate in the exchange of carbon dioxide and oxygen and, as a result, chloroplasts and mitochondria rely on each other for the provision of sugars and ATP, respectively (Nunes-Nesi et al., 2011). Additionally, photorespiratory oxygenation of Ribulose-5-P in chloroplasts leads to enhanced mitochondrial respiration to regenerate substrates for the Calvin cycle (Maurino & Peterhansel, 2010). It has been estimated that reduction of the net photosynthetic CO2 assimilation rate as a result of photorespiration at high temperature in C3 plants like rice can be as high as 25–35% (Sage, 2001). The subtraction of the respiration rate from the photosynthetic rate determines the biomass production in plants and is an important agricultural index for crop production (Amthor, 1989). Efforts to engineer C4 photosynthesis or bypass the role of mitochondria in Ribulose-5-P regeneration in order to eliminate photorespiration in C3 plants are continuing in several cereal crops (Long et al., 2006; Hibberd et al., 2008). However, despite the extensive literature on the differentiation of the plastid from an etioplast to a chloroplast in response to light (Kleffmann et al., 2007; Schattat et al., 2012), and an in-depth understanding of the processes that define the photorespiratory pathway in C3 plants (von Caemmerer & Evans, 2010; Maurino & Peterhansel, 2010), insights are still lacking into the differentiation of mitochondria during the development of photosynthetic cells of plants and into how conserved these processes are between the major C3 dicot crop species and the C3 Poaceae.

In green plant tissues, mitochondria have long been known to undertake different metabolism during dark and light periods. In the dark and at night, organic acids from breakdown of carbohydrates or proteins (Journet et al., 1986; Brouquisse et al., 1998; Hoefnagel et al., 1998) are major substrates for leaf respiration. In the day or in the light, the rate of tricarboxylic acid (TCA) cycle-linked mitochondrial respiration is lowered (Kromer et al., 1993; Atkin et al., 1998; Griffin et al., 2001; Tcherkez et al., 2005, 2008) and oxidation of the amino acid glycine, generated by peroxisomes from chloroplast-derived photorespiratory products, is the major substrate for mitochondrial respiration in C3 plants (Arron & Edwards, 1980; Walker & Oliver, 1986; Lernmark et al., 1990; Padmasree et al., 2002). While this switch between day and night or light and dark function is linked to substrate availability, it is also regulated at the transcriptional and post-translational levels. The expression of genes encoding mitochondrial glycine decarboxylase (GDC) subunits and serine hydroxymethyltransferase (SHMT) in pea leaf has been shown to be highly up-regulated by light in classical etiolated shoot greening experiments (Macherel et al., 1990; Turner et al., 1993). This effect is reportedly regulated by phytochrome-mediated transcriptional control of the promoters of photorespiratory components (Turner et al., 1993; Tepperman et al., 2001). Post-translationally, the phosphorylation and inactivation of the mitochondrial pyruvate dehydrogenase complex, PDC, occurs in light (Tovar-Mendez et al., 2003), and an unknown post-translational inactivation of GDC function occurs in darkness (Lee et al., 2010). Mitochondrial proteins associated with photorespiratory glycine oxidation reactions are always found in highest abundance in photosynthetic tissues and are much lower in abundance in etiolated tissues and in nongreen tissues (Sachlstrom & Ericson, 1984; Newton & Walbot, 1985; Remy et al., 1987; Lind et al., 1991; des Francs-Small et al., 1992; Bardel et al., 2002; Lee et al., 2008, 2011).

Most of what is currently known about plant mitochondria responses to light is derived from studies of C3 dicotyledonous plants such as pea, potato, spinach and Arabidopsis. However, C3 Poaceae represent a larger proportion of crops and the international focus on C4 engineering to decrease photorespiration has more recently turned to work on these grasses, most notably wheat and rice (von Caemmerer et al., 2012; Reynolds et al., 2012). Early studies on the wheat mitochondrial proteome response to light showed significant induction of GDC subunits in isolated mitochondria (Rios et al., 1991), but this was before detailed studies of the specific proteins involved could be performed. In rice, our previous analysis of biogenesis and heterogeneity of mitochondrial proteome using available microarray data revealed that the transcripts for genes encoding mitochondrial photorespiration (GDC and SHMT) have selectively higher steady-state levels in leaf tissues (Huang et al., 2009). However, analysis of the 52 genes in rice predicted to encode enzymes involved in eight steps of the photorespiration pathway, including genes encoding mitochondrial GDC and SHMT subunits, revealed that only one or two genes for each step had higher absolute expression than that of other homolog genes in the same step in response to light, suggesting very selective expression of particular photorespiratory genes in the light in rice (Jung et al., 2008). It was still largely unknown whether these transcriptional changes lead to changes in the abundance of mitochondrial photorespiratory proteins, photorespiration rate or other changes in mitochondrial function in rice. Previous proteomics studies on the response of etiolated rice leaves to light have failed to identify changes in mitochondrial enzymes as a result of the low relative abundance of mitochondria in whole leaf tissue for analysis (Komatsu et al., 1999; Yang et al., 2007). The induction of the photorespiratory apparatus in rice leaves in response to greening in the light not only is a model for studying the initiation of photorespiration, but is also physiologically important because of the practice of low light germination, transplantation, and greening of rice seedlings in the field in many rice-growing regions of the world (De Datta, 1981). Even in case of direct sowing, rice germinates in low-light conditions in muddy paddy fields or underwater and then adapts to higher light intensities when the seedling grows tall enough to reach air and full light intensity (De Datta, 1981).

To better understand the light responsiveness of the protein machinery in rice mitochondria, we investigated changes of metabolite profiles and mitochondrial proteome in 10-d-old etiolated rice plants in response to two intensities of white light (low light (LL), 100 μmol m−2 s−1, and high light (HL), 700 μmol m−2 s−1) for 24 h. Our data revealed that changes in mitochondrial metabolic machinery in response to light correlate with decreased respiration rate and lowered free amino acid content of leaves. Mitochondrial machinery for photorespiration was significantly but differentially induced by light which correlated with a higher photorespiratory capacity in LL- than in HL-treated plants, measured as the postillumination burst (PIB) in respiration.

Materials and Methods

Growth of rice seedlings

Batches of 200 g of rice (Oryza sativa L. cv ‘Amaroo’) seeds were washed in 3% (v/v) bleach for 10 min and rinsed in distilled water. Plants were grown in the dark in vermiculite trays at a constant temperature of 30°C with humidity at 70% and ambient CO2 (c. 390 ppm), and watered daily. After 10 d of growth, the rice plants were exposed to light at an intensity of 100 (LL) and 700 (HL) μmol m−2 s−1 for 24 h. For the continuous light (CL) treatment, the rice plants were grown at a light intensity of 100 μmol m−2 s−1 continuously for 10 d.

Chlorophyll content measurement

Rice shoots (c. 200 mg) were ground on ice under low room light (c. 10 μmol m−2 s−1) and extracted with 10 ml of 80% (v/v) acetone. The homogenate was centrifuged at 1500 g for 2 min. The supernatant was retained and pellet was re-extracted twice with 5 ml of 80% (v/v) acetone. All the supernatants were then pooled together. The absorbance of each extract was measured at 663 and 645 nm with a spectrophotometer. The concentrations of Chla, b and total Chl were calculated using Arnon's equations (Arnon, 1949).

Photosynthesis and PIB measurements

Rates of photosynthesis and postillumination respiratory burst were measured using a LI-6400 XT infrared gas analyzer (Li-Cor, Lincoln, NE, USA). Four leaves from a midleaf region (c. 2 cm long) were laid side by side in a 6 cm2 leaf chamber under conditions with relative humidity of 60–70%, temperature of 30°C and reference CO2 concentration of 400 ppm. After the leaves had acclimatized to chamber conditions, gas-exchange parameters were recorded across a series of light intensities (0, 50, 100, 300, 500, 700 and 900 μmol m−2 s−1). CO2 assimilation rate was recorded in four independent experiments. For the PIB assays, four leaves were laid side by side in a 6 cm2 chamber with a light intensity of 1000 μmol m−2 s−1, a reference CO2 concentration of 100 ppm, a temperature of 30°C and relative humidity of 60–70%. CO2 assimilation rate was monitored for 4 min before the light intensity was switched to zero and CO2 exudation rate was monitored for 4 min. Data were logged every 2 s. Three independent biological replicates were conducted.

Tissue oxygen uptake measurements

Oxygen consumption of c. 2 cm midleaf regions of rice leaves of c. 150 mg FW was measured by a computer-controlled Clark-type O2 electrode unit. In the sealed chamber, measurements were conducted in 2 ml of air-saturated buffer composed of 5 mM KH2PO4, 10 mM TES, 10 mM NaCl, and 2 mM MgSO4, pH 7.2.

Isolation of rice mitochondria and plastid using gradient centrifugation

Shoot mitochondria were isolated from shoots cooled to 4°C using a method previously described (Huang et al., 2009). The purity of isolated mitochondria using this method was established as > 95% based on calculation of the ratio of contaminant protein peptides to total mitochondrial protein peptides detected by a nongel-based Liquid chromatography-tandem mass spectrometry (LC-MS/MS) method (Huang et al., 2009). Approximately 100 g of shoot material was homogenized with a Polytron blender (Kinematica, Kriens, Switzerland) in 300 ml of cold grinding medium (0.3 M sucrose, 25 mM tetrasodium pyrophosphate, 1% (w/v) PVP-40, 2 mM EDTA, 10 mM KH2PO4, 1% (w/v) BSA, 20 mM ascorbic acid, pH 7.5) for 10 s, twice, with 5–10 s intervals between bursts. The homogenate was filtered through four layers of Miracloth and centrifuged at 1500 g for 5 min, and the resulting supernatant was then centrifuged at 24 000 g for 15 min. The organelle pellet was washed by repeating the 1500 and 24 000 g centrifugation steps twice in sucrose wash medium. The resulting pellet of crude organelles was carefully resuspended in sucrose wash medium and gently layered over a 35 ml continuous 28% Percoll density gradient containing 0–4.4% PVP-40. The gradients were then centrifuged at 40 000 g for 45 min. The mitochondrial band was seen as a yellow-brown band near the bottom of the tube. The upper layer contained crude plastid (see Fig. 2). The chloroplast and mitochondrial band were collected. The chloroplast and mitochondrial fractions were diluted approximately fivefold with sucrose wash buffer and centrifuged at 24 000 g for 10 min. The washed mitochondria fraction was further purified with a second Percoll density centrifugation as described for the first gradient. The final mitochondrial band was collected and washed three times by dilution and centrifugation.

Differential in-gel electrophoresis (DIGE) two-dimensional isoelectric focusing/sodium dodecyl sulfate polyacrylamide gel electrophoresis (IEF/SDS-PAGE)

Samples (50 μg) of mitochondria protein isolated from dark- and light-treated shoots, as well as 50 μg of mixture of both samples (1 : 1), were acetone-precipitated, resolubilized in lysis buffer (7 M urea, 2 M thiourea, 4% (w/v) CHAPS (3-[(3-Cholamidopropyl)dimethylammonio]-1-propanesulfonate), and 40 mM Tris base, pH 8.5), and individually labeled with 400 pmol of weight- and pI (isoelectric point)-matched fluorescent dyes known as Cy2, Cy3, and Cy5 (GE Healthcare, Buckinghamshire, UK). Samples were then combined and separated on IEF strips pH 3-10NL (24 cm; GE Healthcare) according to the manufacturer's instructions. After the first dimension running, the strips were transferred to an equilibration buffer consisting of 50 mM Tris-HCl (pH 8.8), 6 M urea, 2% (w/v) SDS, 0.001% (w/v) bromophenol blue, and 65 mM DTT or 130 mM iodoacetamide incubated for 15 min at room temperature with rocking. The strips were transferred to 12% (w/v) acrylamide Gly gels and covered with 1.2% agarose in gel buffer. For the second dimension, gels were run at 50 mA per gel for 6 h. Proteins were visualized on a Typhoon laser scanner (GE Healthcare), and image comparison was performed using the DECYDER software package (version 6.5; GE Healthcare). Three independent experiments were performed, and each of the resulting three gel sets was first analyzed using the differential in-gel analysis mode in DECYDER before a comprehensive biological variance analysis including all three gel sets. Gel spots were filtered according to their presence and average abundance ratio. Gel images were electronically overlaid using ImageQuant TL software (GE Healthcare). Both DIGE and preparative gels were Coomassie Brilliant Blue-stained and destained for further mass spectrometry analysis.

Identification of spots using MALDI TOF/TOF

Peptides were analyzed with an UltraFlex III matrix-assisted laser desorption ionization–time-of-flight/time-of-flight (MALDI-TOF/TOF) mass spectrometer (Bruker Daltonics, Billlerica, MA, USA) using a method described previously (Li et al., 2012). In-gel digested peptides were reconstituted in 5 μl of 5% acetonitrile (ACN) (v/v), 0.1% (v/v) trifluoroacetic acid (TFA). Two microliters of each sample were spotted onto a MTP 384 MALDI target plate and mixed with 2 μl of spotting matrix (90% ACN (v/v), 10% (v/v) saturated α-cyano-4-hydroxycinnamic acid in TA90 solution (90% ACN, 0.01% TFA)). Dried spots were overlaid with 10 μl of cold washing buffer (10 mM NH4H2PO4, 0.1% TFA) and allowed to stand for 10 s before removal by pipette. The spots were analyzed at 50–85% laser intensity with up to 1200 shots for MS analysis per spot. Ions between 700 and 4000 m/z were selected for MS/MS experiments using 3% additional laser power. Masses corresponding to trypsin autolysis were excluded from analysis. Tandem MS data were analyzed using Biotools (Bruker Daltonics) and an in-house Rice database comprising (Osa 5) from the Rice Genome annotation project (http://rice.plantbiology.msu.edu/), using the Mascot search engine version 2.1.04 and utilizing error tolerances of ± 1.2 Da for MS and ± 0.6 Da for MS/MS; ‘Max Missed Cleavages’ set to 1; variable modifications of oxidation (Met), carbamidomethyl (Cys), and deaminated (Asn and Gln). Only protein matches with more than two peptides and with ion scores > 37 were used for analysis (< 0.05). For all spots, both MS and MS/MS were analyzed for protein identifications.

Metabolite extraction and GC-MS analysis

Metabolites were extracted from snap-frozen 10-d-old etiolated rice shoots grown in the dark, or plants switched to 4 or 24 h of LL or HL, according to a method adapted from that described by Shingaki-Wells et al. (2011). The derivatized metabolite samples were analyzed on an Agilent GC/mass selective detector (MSD) system (Agilent Technologies, http://www.home.agilent.com). Raw GC-MS data preprocessing and statistical analysis were performed using METABOLOME-EXPRESS software (version 1.0, http://www.metabolome-express.org). Detailed methods are provided in Carroll et al. (2010).


Responses of plant growth and shoot Chl content to LL and HL

Rice seedlings cultivated under light from seed yielded a maximum net photosynthetic rate of 11 μmol CO2 m−2 s−1 at 900 μmol m−2 s−1 at 400 ppm CO2 (Fig. 1a). Light intensities of 700 and 100 μmol m−2 s−1 provided 90 and 30% of maximum photosynthetic rate, respectively (Fig. 1a), and presented the low and high range of light intensities experienced by rice in field measurements in different parts of the world (De Datta, 1981). Etiolated rice plants propagated in the dark for 10 d (D) were exposed to low white light (LL, 100 μmol m−2 s−1) and high white light (HL, 700 μmol m−2 s−1) for 24 h (Fig. 1b). After 24 h of light treatment, the plants turned green with a significantly greater visible response under LL than under HL (Fig. 1b). We monitored Chl content after 4 and 24 h of LL and HL treatment. After 4 h, LL but not HL plants had significantly higher Chl content in the shoot (Table 1). Consistent with the visible difference in color change after 24 h of treatment, the total Chl content in HL-treated shoots was 39% of that measured in LL-treated shoots (Table 1), indicating less Chl synthesis under HL than under LL. Amounts of Chla and b in HL-treated shoots were 44 and 25% of those in LL-treated shoots, respectively (Table 1). This indicated relatively less Chlb accumulation in rice shoots under the HL treatment. Consistent with these differences in Chl content, we observed lower photosynthetic rate in HL than in LL plants in light response curves and also that higher light intensities yielded photoinhibition of CO2 assimilation in LL plants compared with plants grown in continuous light (CL; Fig. 1c). Steady-state respiratory rates of this tissue in the dark showed a 10% lower respiratory rate in the shoots of HL-treated plants (Fig. 1d).

Table 1. Changes of Chl content in 10-d-old rice (Oryza sativa) shoots in response to light
Light (μmol m−2 s−1)Time (h)Total Chl (μg g−1 FW)Chla (μg g−1 FW)Chlb (μg g−1 FW)
  1. Rice plants were grown in the dark for 10 d (0 h) and then exposed to low light (100 μmol m−2 s−1) and high light (700 μmol m−2 s−1) for 4 and 24 h. The shoot samples were extracted with 80% acetone for Chl content measurements. Errors are given as ±SE.

0012.0 ± 3.27.6 ± 0.54.5 ± 3.6
100486.9 ± 2.665.5 ± 4.221.4 ± 1.5
10024452.2 ± 18.4353.8 ± 17.498.5 ± 1.6
700410.0 ± 1.48.5 ± 1.01.5 ± 0.6
70024178.8 ± 31.5154.4 ± 27.724.4 ± 4.3
Figure 1.

Rice (Oryza sativa) leaf photosynthesis and respiration rates in response to light. (a) Net photosynthetic rate in rice leaves responding to light. CO2 assimilation was measured using AN LI-6400 XT infrared gas analyzer (Li-Cor) with CO2 at 400 ppm and temperature at 30°C. (b) Response of 10-d-old etiolated rice shoots treated with light for 24 h. Rice plants were grown in the dark for 10 d (left, etiolated) and then exposed to low light (LL, middle, 100 μmol m−2 s−1) and high light (HL, right, 700 μmol m−2 s−1) for 24 h. The photograph inserts show a closer view of representative plants. (c) Light response of photosynthesis in rice leaves after LL, HL and continuous light (CL) treatments. CO2 assimilation was measured using the LI-6400 XT infrared gas analyzer (Li-Cor) with CO2 at 400 ppm and temperature at 30°C. (d) Effect of dark (D), LL and HL on respiration of rice shoots. The respiration was measured using a Clark O2 electrode at 30°C. The asterisk (*) indicates a P-value < 0.05. The error bars in panels (c) and (d) are ± SE.

Responses of the rice shoot mitochondria proteome to the light

To investigate what changes in metabolic machinery were accompanying or were determined by changes in photosynthetic and respiratory rates (Fig. 1c,d), the mitochondrial and plastid proteome of shoots transferred from dark to LL and HL were investigated. We isolated mitochondria and plastids by centrifugation to yield crude organelle pellets followed by separation using 0–4.4% PVP in self-forming Percoll gradients (Supporting Information, Fig. S1A). Consistent with the changes of shoot color and the chlorophyll contents, the upper band containing plastid from the LL-treated shoot samples was greener than from the HL-treated shoot samples (Fig. S1A). We analyzed the changes in the protein patterns from isolated mitochondria and plastids using one-dimensional SDS-PAGE (Fig. S1B). Dramatic changes in protein pattern induced by both LL and HL were found in the shoot plastid, but not in the shoot mitochondrial profile (Fig. S1B). Plastid proteins such as AAA-type ATPases and protochlorophyllide reductase A that were dominant in etiolated shoots disappeared within 24 h of LL and HL treatment (Fig. S1B, Table 2), while newly synthesized proteins such as ferredoxin-NADP reductase, oxygen-evolving enhancer protein 1, Chl A–B binding protein, photosystem II 10 kDa polypeptide and cytochrome b559 alpha chain substantially accumulated in both LL- and HL-treated shoots (Fig. S1, Table 2). These dramatic changes in protein pattern in the rice shoot plastid induced by light are consistent with an earlier study of the rice plastidic proteome in response to light (Kleffmann et al., 2007) and also whole etiolated rice leaf in response to light (Komatsu et al., 1999; Yang et al., 2007).

Table 2. Identification of selected plastidic proteins from rice (Oryza sativa) shoots in response to light
  1. Spot numbers are presented in Fig. S1. The predicted molecular mass in kD (MM) and pI (isoelectric point) of the match are shown along with the Mascot protein ion score (< 0.05, when peptide score > 37), number of peptides matched to MS/MS spectra, and the percentage coverage of the matched sequence.

1Os11g47970AAA-type ATPase family protein21752.1/5.59412.4
2Os04g58200Protochlorophyllide reductase A24931.1/8.83725.3
3Os04g58200Protochlorophyllide reductase A11431.1/8.8337.7
4Os06g01850Ferredoxin–NADP reductase11740.0/8.7229.7
5Os01g31690Oxygen-evolving enhancer protein 153634.8/6.10628.2
6Os07g37240Chl A–B binding protein35231.3/5.33823.8
7Os11g13890Chl A–B binding proteind27830.3/5.50422.3
8Os07g37550Chl A–B binding protein26128.8/5.82419.9
9Os04g38410Chl A–B binding protein32527.0/6.75420.2
10Os08g10020Photosystem II 10 kDa polypeptide14713.7/8.52321.4
11Osp1g00520Cytochrome b559 alpha chain2409.4/4.64437.3

To obtain a detailed and quantitative analysis of the rice shoot mitochondrial proteome in response to the light, we conducted DIGE (Blasing et al., 2005) by labeling mitochondrial proteins from etiolated (D), LL- and HL-treated shoots with different Cy dyes (Fig. 2). The changes in the mitochondrial proteome induced by LL and HL were similar, but with a more pronounced response induced by HL than by LL (Fig. 2, Table 3). We observed increases of photorespiration-related glycine decarboxylase P (GDCP) proteins (Os06g40940, Os01g51410), glycine decarboxylase T (GDC-T) protein (Os04g53230) and SHMT (Os03g52840) under both LL and HL conditions (Fig. 2, Table 3), which is consistent with up-regulated expression of these specific genes by light (Jung et al., 2008; see Table 3). Accumulation of GDC-P proteins was higher in LL-treated shoots than in HL-treated shoots, while the opposite pattern was observed in accumulation of SHMT (Os03g52840) and GDC-T protein (Os04g53230; Fig. 2, Table 3). This result suggests that distinct mitochondrial components of the photorespiratory machinery might be differentially regulated by light intensity. Surprisingly, we did not find any significant change in abundance of GDC-H proteins (Os02g07410, Os10g31780) under either LL or HL conditions (Fig. 2, Table 3), even though significant up-regulated expression of Os10g31780 by light has been previously reported (Jung et al., 2008). The finding that GDC-H proteins were not responsive to light in rice is surprising given this subunit's clear accumulation in other plants, such as Arabidopsis, pea and wheat. In fact, it has been one of the most intensively studied mitochondrial proteins induced by light via a light-responsive promoter (Walker & Oliver, 1986; Macherel et al., 1990, 1992; Rios et al., 1991; Turner et al., 1993; Oliver & Raman, 1995; Lee et al., 2010). The GDC-L protein is a subunit shared with the pyruvate dehydrogenase complex and 2-oxoglutarate dehydrogenase complex of the TCA cycle (Bourguignon et al., 1996; Mooney et al., 2002). In pea, a single gene exists and its product is found in all three complexes (Bourguignon et al., 1996). In Arabidopsis, only one of two GDC-L proteins is light-inducible, but both are found in all three complexes (Lutziger & Oliver, 2001). We detected four distinct protein spots for GDC-L (Os01g22520). Although they showed no response in LL, three of the four GDC-L spots decreased in abundance in HL (Fig. 2, Table 3), even though up-regulation of Os01g22520 transcript under light has been previously reported (Table 3).

Table 3. Selected protein spots identified from differential in-gel electrophoresis DIGE two-dimensional gels in Fig. 2
ClassSpotAccessionDescriptionRatio of protein abundanceExpression1ScoreMM/pIPeptidesCoverage
  1. TCA, tricarboxylic acid.

  2. Proteins whose spot numbers are indicated in Fig. 2 were identified by matrix-assisted laser desorption ionization–time-of-flight/time-of-flight (MALDI-TOF/TOF) MS. The predicted molecular mass in kD (MM) and pI (isoelectric point) of the match are shown along with the MOWSE score (< 0.05, when peptide score > 37), number of peptides matched to tandem mass spectra, and the % coverage of the matched sequence. The LL/D and HL/D columns represented ratios of protein abundance of low light (LL)-treated to etiolated rice (Oryza sativa) plants (D) and high light (HL)-treated to etiolated plants (D), respectively. Significant changes in abundance ratio (P < 0.05) are highlighted in bold. 1Gene expression data with ratios of light to dark were extracted from Jung et al. (2008). Cells with N/A (not available) represented an absence of data in Jung et al. (2008).

Photorespiration1Os06g40940Glycine decarboxylase (GDC P) 2.46 0.029 1.49 0.0305.09401111.4/6.35910.2
2Os06g40940Glycine decarboxylase (GDC P) 2.13 0.034 1.96 0.0215.09607111.4/6.351013.3
3Os01g51410Glycine decarboxylase (GDC P) 2.31 0.0391.390.1705.09362111.4/6.51911.1
4Os01g51410Glycine decarboxylase (GDC P) 2.77 0.0171.370.4105.09671111.4/6.511014.3
5Os04g53230Aminomethyltransferase (GDC T)1.280.120 1.39 0.0045.3734143.9/8.53422.5
10Os02g07410Glycine decarboxylase H protein−1.120.830−1.490.1800.665117.0/4.8215.7
11Os10g37180Glycine decarboxylase H protein1.210.320−1.060.2502.9627517.4/4.92547.6
12Os10g37180Glycine decarboxylase H protein1.080.610−1.200.3002.969817.4/4.92215.9
6Os03g52840Serine hydroxymethyltransferase 1.48 0.003 1.63 0.0313.5050656.4/8.40817.2
7Os03g52840Serine hydroxymethyltransferase 1.69 0.010 2.09 0.0033.5030356.4/8.40820.9
8Os03g52840Serine hydroxymethyltransferase 1.42 0.025 2.58 0.0013.5051856.4/8.40922.2
9Os03g52840Serine hydroxymethyltransferase1.210.064 2.44 0.0013.5046256.4/8.40922.2
Photorespiration and TCA cycle13Os01g22520Dihydrolipoyl dehydrogenase 1 (GDC L)1.160.1801.980.0501.9545452.6/7.21620.1
14Os01g22520Dihydrolipoyl dehydrogenase 1 (GDC L)1.540.0032.710.0061.9543852.6/7.21720.1
15Os01g22520Dihydrolipoyl dehydrogenase 1 (GDC L)−1.170.1101.520.0111.9514452.6/7.21414.5
16Os01g22520Dihydrolipoyl dehydrogenase 1 (GDC L)1.010.820−1.060.4801.9545152.6/7.21620.1
TCA cycle17Os08g42410Pyruvate dehydrogenase E1 β subunit−1.080.0781.250.004N/A14339.9/5.2548.6
18Os09g33500Pyruvate dehydrogenase E1 β subunit−1.030.7501.330.0010.8832640.2/5.36721.8
19Os03g04410Aconitate hydratase−1.010.8901.490.030N/A764106.2/6.45914.1
20Os07g38970Succinyl-CoA ligase subunit α-21.200.0461.750.043N/A60634.2/8.46726.9
21Os02g40830Succinyl-CoA ligase subunit β1.090.2101.590.039N/A17145.1/5.98515.4
22Os07g04240Succinate dehydrogenase flavoprotein1.070.3901.520.011N/A18868.8/6.61614.8
AA and branch chain metabolism23Os02g14110Aspartate aminotransferase−1.120.1601.540.000N/A7748.1/8.0929.5
26Os05g03480Isovaleryl-CoA dehydrogenase1.090.1501.690.010−4.0230644.5/6.52827.9
27Os07g09060Methylmalonate-semialdehyde DH1.120.1401.440.006N/A36757.2/5.98921.0
28Os07g09060Methylmalonate-semialdehyde DH1.030.7501.630.012N/A40957.2/5.98715.5
Others29Os09g31486Heat shock protein (70 kDa)1.120.300 1.85 0.000N/A21572.8/5.6668.8
30Os05g51700Nucleoside diphosphate kinase 1.41 0.042 2.70 0.0141.420525.9/8.88313.0
31Os03g63410Elongation factor Tu1.190.0021.440.001N/A42048.4/6.04819.4
Contaminants32Osp1g00420Rubisco large chain1.060.350 2.38 0.006N/A52852.8/6.221023.5
33Osp1g00420Rubisco large chain1.020.820 1.97 0.005N/A29052.8/6.22816.4
Figure 2.

Changes occurring in the mitochondrial proteome of rice (Oryza sativa) shoots treated under low light (LL) and high light (HL) conditions. Mitochondrial proteins from dark-treated (D) etiolated shoots (labeled with Cy3, green color) or from LL- and HL-treated shoots (labeled with Cy5, red color) were separated by two-dimensional isoelectric focusing/sodium dodecyl sulfate polyacrylamide gel electrophoresis (IEF/SDS-PAGE) (pI 3-10 NL). The top panels represent a gel image from each fluorescence signal, and the bottom panels are combined fluorescence images overlaid using ImageQuant TL software (GE Healthcare). Yellow spots represent proteins of equal abundance before and after light treatment. The numbered white arrows indicate protein spots with statistically significantly changes in abundance (= 3. < 0.05). The numbered yellow arrows indicate glycine decarboxylase (GDC) H subunits that did not significantly change in abundance but were selected for analysis. Evidence of protein identification of selected protein spots analyzed by matrix-assisted laser desorption ionization–time-of-flight/time-of-flight (MALDI-TOF/TOF) MS is presented in Table 3.

The TCA cycle and related enzymes, such as pyruvate dehydrogenase E1 subunits (Os08g42410; Os09g33500), aconitate hydratase (Os03g04410), succinyl-CoA ligase (S-CoA-L) β subunits (Os02g40830) and succinate dehydrogenase (SDH) flavoprotein (Os07g04240), were unchanged by LL treatment, but were all significantly decreased in abundance after 24 h of HL treatment (Table 3, Fig. 2). The decreased abundance of TCA cycle enzymes, including the S-CoA-L α2 subunit (Os07g38970) of the TCA cycle and GDC-L/lipoamide dehydrogenase (Os01g22520) under both light treatments, could slow the catalytic capacity of the TCA cycle, and consequently alter TCA cycle-related intermediates.

Mitochondrial proteins involved in amino acid metabolism, such as aspartate aminotransferase (Os02g14110), arginase (Os04g01590), and gamma-aminobutyrate transaminase (Os04g52450), were all decreased by HL treatment (Table 3). A similar pattern of changes was evident for proteins involved in branch chain amino acid catabolism such as isovaleryl-CoA dehydrogenase (Os05g03480) and methylmalonate-semialdehyde dehydrogenase (Os07g09060; Table 3). Interestingly, no significant changes in those enzymes under LL treatment were observed, with the exception of gamma-aminobutyrate transaminase (Os04g52450; Table 3). Several other mitochondria proteins whose abundance changed in response to light were also detected. The 70 kDa heat shock protein (Os09g31486) was significantly increased under HL treatment (Table 3), and nucleoside diphosphate kinase III (Os05g51700) was also induced in a light-dependent manner. Physical association of nucleoside diphosphate kinase (NDPK) and heat shock proteins, and their coinduction by heat and stress, have been reported in a range of plant species (Galvis et al., 2001). Roles for NDPK in light signaling in plants, fungi and insects have also been reported (Randazzo et al., 1991; Hasunuma et al., 2003; Im et al., 2004; Yoshida & Hasunuma, 2006).

Changes in organic acid and amino pools in shoots during light treatments

To determine whether or not the changes in mitochondrial enzyme abundances were associated with metabolite pools altered by light, we measured metabolite profiles of rice shoots using GC-MS after a transition from dark to 4 and 24 h of LL or HL treatment (Table S1, Fig. 3). After 4 h light treatment, there was no major change in the profiles of major metabolites in LL- or HL-treated shoots, with the exception of the significant accumulation of the mitochondrial TCA cycle intermediate succinic acid (1.9- to 2.1-fold) and the amino acid hydroxyproline (1.48- to 1.52-fold) under both light conditions (Table S1). The former is consistent with the lowering of the SDH enzyme abundance observed under HL. We have previously observed that limitation of SDH leads to rapid accumulation of succinate (Gleason et al., 2011). Slight increases in the abundance of specific carbohydrates, notably d-ribose, β-d-Fucose and cellobiose, were also observed, but only following HL treatment (Table S1).

Figure 3.

Effect of light on the abundance of metabolites in rice (Oryza sativa) shoots associated with carbohydrate metabolism, glycolysis, amino acid metabolism, and the tricarboxylic acid (TCA) cycle. Rice seeds were germinated and grown in the dark for 10 d (D) and then transferred to light conditions with intensities of 100 μmol m−2 s−1 (LL) and 700 μmol m−2 s−1 (HL) for 24 h. The red boxes represent metabolites higher in abundance under HL or LL conditions relative to D (P-value < 0.05). The green boxes represent metabolites significantly lower in abundance under HL or LL conditions relative to D (P-value < 0.05). The yellow boxes represent metabolites whose abundances are unchanged by light exposure. Mitochondrial enzymes whose abundances changed (Table 3) are also colored in this fashion. The numbers on the top left and right side of each box represent the response fold change value (RV) of the corresponding metabolite LL/D and HL/D in rice shoots, respectively. All data were extracted from Table 3 and Table S1. Amino acid abbreviations are listed in Table S1. Other abbreviations: AA, amino acid; G-3-P, glyceraldehyde-3-phosphate; PEP, phosphoenolpyruvate. Mitochondrial protein abbreviations are as follows: PDH, pyruvate dehydrogenase; CS, citrate synthase; IDH, isocitrate dehydrogenase; 2-OGDH, 2-oxoglutarate dehydrogenase; S-CoA-L, succinyl-CoA ligase; SDH, succinate dehydrogenase; FUM, fumarase; MDH, malate dehydrogenase; GDC, Gglycine decarboxylase; SHMT, serine hydroxymethyltransferase.

After 24 h of treatment, dramatic changes in the abundance of a number of organic acids were evident in both LL and HL treatments (Fig. 3, Table S1). Generally, whole cell abundance of organic acids that are also intermediates of the TCA cycle were significantly increased in abundance by light treatment for 24 h (Table S1). Most notably, succinic acid and isocitric acid increased by 1.6- to 2.6-fold under both LL and HL treatments (Fig. 3). Increased synthesis of isocitrate and succinate from the peroxisomal glyoxylate and beta-oxidation pathways, and inability of the TCA cycle to fully utilize these extra substrate pools are consistent with the changes seen. There was also fluctuation in the abundance of several carbohydrates in rice shoots after 24 h of LL or HL treatment: erythritol, rhamnose and d-ribose were increased, while d-glucose and d-galactose were decreased in both LL- and HL-treated shoots (Table S1).

Large reductions in the abundance of a range of amino acids were also observed, with the exception of hydroxyproline (Fig. 3, Table S1). LL treatment induced greater reduction than HL treatment in the abundance of most amino acids (Fig. 3). Interestingly, the branched chain amino acids l-valine and l-isoleucine halved in abundance in HL-treated shoots, but were lowered to only 10% in LL-treated shoots relative to etiolated shoots (Fig. 3). This correlates with the lowered abundance of branched chain degradation enzymes in HL but not LL (Table 3, Fig. 4). Notably, glycine, the substrate for the induced glycine decarboxylase subunits, was one of only a few amino acids that did not change in abundance after either LL or HL treatment (Fig. 3, Table S1), in contrast to the large depletion of many of the other amino acids.

Figure 4.

Effect of light exposure on photorespiration in rice (Oryza sativa) leaves. (a) The CO2 assimilation rate was determined from measurements taken using an LI-6400 XT infrared gas analyzer (Li-Cor) with light intensity of 1000 μmol m−2 s−1, CO2 concentration of 100 ppm and temperature of 30°C. The postillumination burst (PIB) was monitored after switching from light to dark in the first 20 s. The steady state of respiration in the dark (Rd) was reached after each PIB. Three independent measurements are presented in Fig. S2. (b) Summary of the average ± SE for the lit CO2 assimilation rates (A) and dark CO2 production during the PIB and during Rd are plotted for each light treatment.

Response of photorespiration to light intensity treatment

To determine whether there are biological consequences of the light treatments on photorespiration, we measured the PIB using an infrared gas analyzer (IRGA), in LL-, HL-, and CL-treated plants. Immediately following the switch to darkness, HL-treated leaves assumed a steady-state respiration rate in the dark that was maintained over 4 min (Fig. 4a). This contrasts with LL-treated leaves, where the shift to darkness elicited a sharp pulse of CO2 release in the first 20 s, before the respiration rate settled to a steady state (Fig. 4a). This PIB indicates an at least fourfold higher rapid respiratory oxidation of glycine following the shift to the dark (Fig. 4b), which rapidly dissipates, despite the same steady-state abundance of glycine in the light in both HL and LL shoots (Fig. 3). A similar PIB was also recorded in CL growth plants, indicating a robust photorespiratory rate could be initiated in 24 h by LL but not by HL treatment of etiolated rice shoots (Fig. 4b).


We previously analyzed the biogenesis and heterogeneity of rice mitochondrial proteome using available microarray data and revealed that the transcripts for genes encoding rice mitochondrial photorespiration (GDC and SHMT) have selectively higher steady-state levels in leaf tissues (Huang et al., 2009), much like those in various dicot species (Walker & Oliver, 1986; Macherel et al., 1990, 1992; Rios et al., 1991; Turner et al., 1993; Oliver & Raman, 1995; Lee et al., 2010). In this study, we directly investigated changes of mitochondrial proteome and metabolite profiles in 10-d-old etiolated rice plants in response to two intensities of white light (100 and 700 μmol m−2 s−1) for 24 h, representing the low and high range of light intensities experienced in rice fields (De Datta, 1981). Our findings suggest that mitochondrial composition in etiolated shoots respond differently to LL and HL, with some responses more pronounced under HL and others more evident under LL. Interestingly, chloroplast proteome development, Chl content and the initiation of photosynthesis all showed a greater response to LL than to HL. The majority of the variation in the plant mitochondrial proteome in response to light occurred in proteins involved in the mitochondrial photorespiratory machinery, the TCA cycle, amino acid, and branch chain amino acid metabolism. Two notable differences other than this major C/N machinery were the light-dependent increases in HSP70 and NDPKIII. We observed significant changes in metabolite concentrations in response to light, dominated by the general decrease in amino acid content and increase of organic acid content.

NDPKIII in rice

Mitochondrial NDPK was increased under both LL and HL conditions (Table 3). In Arabidopsis, mitochondrial NDPKIII is reported to be dual-targeted to chloroplast based on green fluorescent protein localization and western bolt experiments (Hammargren et al., 2007). In rice, NDPKIII is exclusively mitochondrial-targeted, while NDPKII is targeted to plastids (Kihara et al., 2011). Enzymatically NDPK is involved in nucleotide conversion (classically ATP + guanosine diphosphate (GDP) → ADP + guanosine triphosphate (GTP)), and a role for mitochondrial NDPK in phosphor transfer within nucleotide pools is often discussed (Johansson et al., 2004). However, NDPKs are more widely involved in signaling processes in a number of species, notably in light signaling, through phytochrome-mediated processes in plants (Im et al., 2004) or via singlet oxygen processes in fungi such as Neurospora crassa (Yoshida & Hasunuma, 2006). In these roles, NDPK directly converts bound GDP to GTP while the nucleotides are still bound to G proteins (Randazzo et al., 1991). Knowledge of NDPK binding partners in plant mitochondria is very limited; the only report is NDPK interaction with a 86 kDa heat shock protein in pea during heat stress (Galvis et al., 2001). In mammals, succinyl-CoA-ligase is a binding partner for NDPK in mitochondria (Kowluru et al., 2002) and two mitochondrial S-CoA-L proteins (Os07g38970, Os02g40830) were decreased in abundance in response to light in rice (Table 3). However, if NDPKIII in rice is in the intermembrane space as it is in Arabidopsis, pea and yeast (but not in mammals), then NDPK and S-CoA-L proteins will not be able to interact in vivo.

TCA cycle, amino acid and organic acid metabolism

We observed an apparent reduction of TCA cycle metabolism in response to light treatments, as supported by accumulation of TCA cycle substrates and reduction in abundance of TCA cycle-related enzymes, especially under HL treatment (Figs 2, 3). These results were consistent with previous studies showing that the rate of TCA cycle-linked mitochondrial respiration in the light is lower than in darkness (Kromer et al., 1993; Atkin et al., 1998; Griffin et al., 2001; Tcherkez et al., 2005, 2008). Our results support the proposition by Tcherkez et al. (2008) in leaves of Xanthium strumarium that the decrease focuses on the early and middle steps of the TCA cycle. In Arabidopsis shoots, a similar decrease in the abundance of enzymes in the early and middle steps of the TCA cycle occurred by analysis of diurnal changes in mitochondrial function (Lee et al., 2010). Protein abundance of mitochondrial enzymes involved in amino acid and branch chain metabolism in rice were also decreased by light (Table 3), which was in general agreement with diurnal changes in Arabidopsis mitochondrial amino acid metabolism (Lee et al., 2010). Given the sophisticated pathways of amino acid synthesis and degradation in plant cells, such a reduction in mitochondrial amino acid metabolism could contribute to the changes of amino acid abundances observed in the light (Fig. 3). The development of chloroplasts from etiolated plastids requires a considerable investment of amino acids for protein synthesis during the transition to the light. Such development is delayed by HL treatment compared with LL treatment, as indicated by Chl content, and correlates to the larger residual amino acid content in HL-treated shoots than in LL-treated shoots (Table 1, Fig. 3). Our analysis of Chl content in plants subjected to different light intensities is also consistent with the report that higher light treatment can lower Chl content, even though maximal growth occurs in higher light regimes (Fu et al., 2011), and that shade species maintain higher Chl content in order to maximize light-use efficiency (Boardman, 1977).

Photorespiration in rice

Glycine is one of relatively few amino acids in etiolated rice shoot that was not decreased in abundance by light (Fig. 3). Mitochondrial proteins involved in glycine metabolism, such as GDC subunits and SHMT, were significantly increased in abundance (Table 3, Fig. 3). The induction of glycine-dependent photorespiration proteins by light is widely reported in other plant species, such as Arabidopsis, pea and wheat (Rios et al., 1991; Rogers et al., 1991; Lee et al., 2010). However, the abundance of GDC-H proteins under both light conditions studied here were unchanged (Table 3, Fig. 2), which is in contrast to the common observation of light-induced GDC H proteins in other plant species (Rios et al., 1991). The transcript for GDC-H (Os10g31780) is also reported to respond to light in rice (Jung et al., 2008), so it is possible that this GDC-H protein might be regulated at the post-transcriptional/-translational level or via protein degradation. Previously, light was reported to induce similar gene expression patterns of GDC subunits and photosynthetic genes such as Rubisco small subunit RbsS, but GDC translation was clearly delayed compared with Rubisco (Oliver & Raman, 1995; Vauclare et al., 1996). GDC is also known to be susceptible to oxidative damage by ROS in mitochondria (Douce et al., 2001) and GDC-H protein lipoic acid is targeted by lipid peroxides, which could make GDC-H a target for degradation in mitochondria (Taylor et al., 2002). Under sudden HL treatment, rice mitochondria may be under enhanced light stress, as indicated by induction of heat shock protein 70 (Os09g31486). GDC-H could thus be turned over rapidly after its synthesis. Interestingly, the GDC-L subunit (Os01g22520), which is known to contain cysteine residues sensitive to damage, was reduced in abundance under HL (Table 3). The decrease in the GDC-L subunit under HL might explain, at least in part, why there is no apparent photorespiration rate in these HL-exposed shoots. Unlike the pyruvate dehydrogenase complex and the 2-oxoglutarate dehydrogenase complex of the TCA cycle that form large complexes with strict stoichiometry (Mooney et al., 2002), GDC forms transient interactions across the network of subunits to facilitate the oxidation of glycine to serine (Douce et al., 2001). The changes in GDC stoichiometry observed here, coupled with large differences in photorespiratory rate (Fig. 4), highlight the mitochondrial rate-limiting components contributing to photorespiratory capacity.

In conclusion, our data show that light-induced changes in the etiolated rice shoot led to a shift in its mitochondrial metabolic machinery from the TCA cycle, amino acid-related metabolism to photorespiration-related metabolism. The unexpected lack of light induction of GDC-H and -L subunit proteins in rice shoot mitochondria warrants further attention to understand the mechanism behind this phenomenon and its implications for photorespiratory development in light-exposed rice seedlings. Detailed analyses of light-induced greening of dark-germinated seedlings, kinetic responses to different light intensities between LL and HL, and diurnal changes in the rice leaf mitochondrial metabolism are required to better mimic nature field conditions and provide further insights into how mitochondrial metabolism adapts to its role in photosynthetic rice tissues.


This research was funded by support from the ARC Centre of Excellence in Plant Energy Biology (CE0561495) and the University of Western Australia and Agilent Technologies Australia through support for the Centre for Comparative Analysis of Biomolecular Networks to A.H.M. A.H.M. was funded as an ARC Australian Future Fellow (FT110100242). R.P.J. is supported by a Grains Research and Development Corporation (GRDC) PhD scholarship. R.N.S.W. is supported by GRDC and an Australian Postgraduate Award PhD scholarship. L.L. is supported by Scholarship International Research Fees, a University International Stipend, and a Top Up Scholarship from University International Stipend. We also thank Mr Matthew D. Goonatillake for technical help.