The lack of alternative oxidase at low temperature leads to a disruption of the balance in carbon and nitrogen metabolism, and to an up-regulation of antioxidant defence systems in Arabidopsis thaliana leaves

Authors

  • CHIHIRO K. WATANABE,

    Corresponding author
    1. Department of Biological Sciences, Graduate School of Science, Osaka University, 1-1 Machikaneyama-cho, Toyonaka, Osaka, 560-0043 and
    2. Department of Biological Sciences, Graduate School of Science, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-0033, Japan
      C. K. Watanabe. Fax: +81 3 5841 4465; e-mail: chihiro-nabe@biol.s.u-tokyo.ac.jp
    Search for more papers by this author
  • TAKUSHI HACHIYA,

    1. Department of Biological Sciences, Graduate School of Science, Osaka University, 1-1 Machikaneyama-cho, Toyonaka, Osaka, 560-0043 and
    2. Department of Biological Sciences, Graduate School of Science, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-0033, Japan
    Search for more papers by this author
  • ICHIRO TERASHIMA,

    1. Department of Biological Sciences, Graduate School of Science, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-0033, Japan
    Search for more papers by this author
  • KO NOGUCHI

    1. Department of Biological Sciences, Graduate School of Science, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-0033, Japan
    Search for more papers by this author

C. K. Watanabe. Fax: +81 3 5841 4465; e-mail: chihiro-nabe@biol.s.u-tokyo.ac.jp

ABSTRACT

Alternative oxidase (AOX) catalyses the ATP-uncoupling cyanide (CN)-resistant pathway. In this study, our aim was to clarify the physiological role of AOX at low temperature. We examined the effect of low-temperature treatment on CN-resistant respiration (CN-resistant R) and on the transcription of respiratory components in wild-type (WT) and aox1a knock-out transgenic (aox1a) Arabidopsis thaliana plants. In WT leaves, the expression of AOX1a mRNA was strongly induced by the low-temperature treatment, and thus CN-resistant R increased during low-temperature treatment. In aox1a, the CN-sensitive respiration, and the expression of NDB2 and UCP1 were increased compared with WT. We compared several physiological parameters between WT and aox1a. Low-temperature treatment did not result in a visible phenotype to distinguish aox1a from WT. In aox1a, several antioxidant defence genes were induced, and the malondialdehyde content was lower than in WT. Starch content and a ratio of carbon to nitrogen were higher in aox1a than in WT. Our results indicate that a lack of AOX was linked to a difference in the carbon and nitrogen balance, and an up-regulation of the transcription of antioxidant defence system at low temperature. It is likely that AOX is a necessary component in antioxidant defence mechanisms and for the control of a balanced metabolism.

INTRODUCTION

The mitochondrial respiratory chain in higher plants consists of the ATP-coupling cytochrome pathway (CP) and the cyanide (CN)-resistant respiratory pathway. The latter is catalysed by the alternative oxidase (AOX) at the surface of the mitochondrial inner membrane. Because AOX is not coupled to the generation of a H+ motive force, and thereby ATP production, the electron flux to AOX apparently wastes energy. Expression of AOX genes and AOX protein is induced in response to diverse biotic and abiotic stresses (Vanlerberghe & Ordog 2002; Finnegan, Soole & Umbach 2004). It has been suggested that stress-induced AOX plays a role in preventing the production of reactive oxygen species (ROS) (Maxwell, Wang & McIntosh 1999; Vanlerberghe, Robson & Yip 2002). In addition, the non-energy-conserving nature of the AOX pathway, along with complex biochemical pathways that regulate electron flux to AOX, provides plants with a considerable metabolic flexibility (Lambers 1982; Plaxton & Podestá 2006). Under low-nitrogen conditions, the AOX protein levels and CN-resistant respiration (CN-resistant R) capacity increased in tobacco cells (Sieger et al. 2005) and in spinach leaves (Noguchi & Terashima 2006). In both papers, it was suggested that AOX consumes excess sugars, and thus, exerts a control over the carbon (C) and nitrogen (N) balance in plants.

Exposure of plants to low temperatures induces AOX protein expression and increases the rate of CN-resistant R (Stewart et al. 1990; Vanlerberghe & McIntosh 1992; Gonzàlez-Meler et al. 1999). In leaves of five ruderal species (Collier & Cummins 1990) and in leaf mitochondria of potato (Covey-Crump et al. 2007), CP has been shown to be more sensitive to low-temperature stress compared with CN-resistant R. Therefore, AOX is thought to contribute to the acclimation of respiration to low temperatures. Using the oxygen isotope fractionation technique, Gonzàlez-Meler et al. (1999) observed that in hypocotyls and leaves of mung bean, low-temperature treatment resulted in an increase in AOX protein accumulation and in electron partitioning to the AOX. Using AOX1a antisense and over-expressing Arabidopsis thaliana plants, Fiorani, Umbach & Siedow (2005) showed that AOX is involved in shoot growth at low temperature at relatively early growth stages. However, the physiological role of AOX under low-temperature stress is still ambiguous.

Plant mitochondria have at least two other ATP-uncoupling pathways in the respiratory chain. (1) Bypassing the H+-pumping complex I, type II NAD(P)H dehydrogenases (NDs) are located at the outer (NDex) or inner (NDin) surfaces of the inner membrane (Michalecka et al. 2003; Rasmusson, Soole & Elthon 2004). (2) Uncoupling protein (UCP) catalyses H+ conductance from the inter-membrane space to the matrix, and thus dissipates the H+ gradient (Vercesi et al. 2006). These pathways may play a role in regulation of the redox balance and metabolism in mitochondria (Fernie, Carrari & Sweetlove 2004), and thus function cooperatively with AOX under some conditions. Recently, microarray analyses indicated that AOX1a and NDB2 in A. thaliana are co-expressed under various stress conditions (Clifton et al. 2005). When A. thaliana seedlings grown at 22 °C were transferred to 4 °C, UCP1 expression was induced (Maia et al. 1998). Similarly, A. thaliana plants transferred from 22 to 4 °C showed an increase in UCP4 and UCP5 transcripts after 3 h, and an increase in AOX1a after 12 h (Boreckýet al. 2006). However, these results contrast with a study by Clifton et al. (2005) who found no clear induction of AOX1a, 1c, NDs and UCP1, 2 genes in cultured cells of A. thaliana transferred from 22 to 10 °C. Thus, it remains unclear whether there is a complementary relationship in gene expressions between AOX, NDs and UCP at low temperatures.

Our aim in this study was to clarify the physiological role of AOX at low temperature in leaves of A. thaliana. Firstly, we transferred wild-type (WT) and aox1a knock-out transgenic plants (aox1a) to low-temperature conditions and examined changes in the CN-resistant respiratory rate. In addition, to elucidate the relationship between AOX, ND and UCP genes, we analysed the transcript levels of genes encoding for these respiratory components. Furthermore, we compared several physiological properties between WT and aox1a treated at 4 °C, such as chlorophyll fluorescence, malondialdehyde (MDA) content and the transcript levels of antioxidant defence genes. To investigate the C and N balance in aox1a, we compared a ratio of C with N, the content of non-structural carbohydrates and amino acids between WT and aox1a treated at 4 °C.

MATERIALS AND METHODS

Plant material and growth conditions

Arabidopsis thaliana (L.) Heynh. accession Columbia was used. Seeds were sown in Metro Mix 350 (Sun Gro Horticulture, Bellevue, WA, USA) and vermiculite (Vern-piece; Hakugen, Tokyo, Japan) (1:1) in 7 cm plastic pots. After seedlings emerged, the plants were irrigated with 30 mL of a 1/1000-strength mineral nutrient solution (Hyponex; Hyponex Japan, Osaka, Japan) once a week. The plants were grown in an air-conditioned room at 23–25 °C, and at 60–70% relative humidity. The day/night cycle was 14 h (1000–2400 h)/10 h, and the photosynthetic photon flux density (PPFD) was between 65 and 80 µmol photons m−2 s−1.

Identification of the T-DNA insertion in the AOX1a gene

The SALK_084897 T-DNA insertion line of the Salk Institute (Alonso et al. 2003) was obtained from the Arabidopsis Biological Resource Center (ABRC). The T-DNA insertion within the AOX1a gene was verified by a genomic PCR with a T-DNA-specific primer (LB-2R 5′-GAC CGC TTG CTG CAA CTC TCT CA-3′), and AOX1a-specific primers (fwd 5′-AAG GCG GCG AAA TCG CTG TT-3′ and rev 5′-CCA AGT ATG GCT TAA GCA GAG GTG A-3′). A homozygous T-DNA insertion plant was identified from the next generation by a genomic PCR and RT-PCR with total RNA. For RT-PCR, AOX1a-specific primers (fwd 5′-ATG TTC CAA CGA CGT TTC TTG-3′ and rev 5′-CCA AGT ATG GCT TAA GCA GAG GTG A) were used.

Low-temperature treatment and sampling

Four-week-old plants were transferred to a growth chamber with a temperature of 4 or 10 °C. In the growth chamber, the photoperiod was 24 h, and the PPFD was again between 65 and 80 µmol photons m−2 s−1. Rosette leaves and roots treated at low temperature for 3, 6, 12, 24, 72 and 120 h were sampled. Before the transfer, leaves and roots were sampled as control samples (0 h). The samples were rapidly harvested, weighed and frozen in liquid nitrogen.

Measurements of the respiratory rate

The respiratory rate of rosette leaves was measured as oxygen consumption with a Clark-type oxygen electrode cuvette (Rank Brothers, Cambridge, UK) at 25 °C and at the treatment temperature (10 or 4 °C) in 4 mL of a respiration buffer containing 50 mm HEPES, 10 mm MES (pH 6.6) and 0.2 mm CaCl2. Oxygen consumption rates were calculated assuming that the concentration of oxygen in the air-saturated buffer was 253, 341 and 396 µm at 25, 10 and 4 °C, respectively. To inhibit the CP, KCN was added at a concentration of 2 mm.

Analysis of respiratory acclimation

To estimate the degree of respiratory acclimation, we calculated Q10 values and H. The Q10 value indicates temperature sensitivity of respiration, and was defined as:

image(1)

where R1 and R2 are the mean respiratory rates measured at 25 °C and the treatment temperature (10 or 4 °C). T1 is 25 °C and T2 is the measurement temperature.

H represents the degree of respiratory acclimation (Loveys et al. 2003). H was calculated as:

image(2)

where Rcont is the mean respiratory rate of non-treated plants measured at 25 °C.

Extraction of total RNA and RT-PCR

Total RNA was extracted from the frozen rosette leaves and roots using the TRIzol Reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instruction for preventing contamination of proteoglycan and polysaccharide. cDNA was amplified from extracted RNA using a ThermoScript RT-PCR System (Invitrogen) with Oligo(dT)20 as a primer.

Real-time PCR

Transcript levels were measured using a 7300 Real-time PCR System and 7000 Sequence Detection System (Applied Biosystems, Foster City, CA, USA). Then, 1 µL of cDNA was amplified in the presence of 12.5 µL of 2× Power SYBR Green PCR Master Mix (Applied Biosystems), 0.5 µL of specific primers (final concentration, 0.2 µm) and 10.5 µL of sterilized water. PCR conditions were 50 °C for 2 min, 95 °C for 10 min and 40 cycles of 95 °C for 15 s followed by 60 °C for 1 min. Relative transcript levels were calculated by the comparative cycle threshold method. As an internal control, RPS15aA was used. The primer sequences used for the real-time PCR are listed in Supplementary Table S1.

Measurements of chlorophyll fluorescence

Chlorophyll fluorescence was measured with a pulse amplitude modulation (PAM) fluorometer (PAM101/102/103; Waltz, Effeltrich, Germany). After 30 min dark incubation at a room temperature (25–27 °C) or 4 °C, leaves were given the saturating pulse (SP) at 10 000 µmol photon m−2 s−1 (KL 1500; Schott, Cologne, Germany), and Fv/Fm was measured. Then, actinic light (AL) at 70 µmol photon m−2 s−1 was provided (KL 1500; Schott). An SP for the estimate of Fm′ was given after AL irradiation for 5 min at 25 °C, and 10 min at 4 °C, after which the chlorophyll fluorescence attained steady-state levels.

Measurements of MDA content

Malondialdehyde content was measured with modified TBARS assay (Hodges et al. 1999). Frozen leaves weighed about 100 mg [fresh weight (FW)] were ground in 2.5 mL of 80% (v/v) ethanol and centrifuged at 10 000 g for 10 min. The supernatant was divided into two aliquots. To one aliquot, an equal volume of +TBA solution [20% (w/v) trichloroacetic acid (TCA), 0.01% (w/v) butylated hydroxytoluene (BHT) and 0.65% thiobarbituric acid (w/v)] was added, and subsequently incubated at 95 °C for 25 min. To the other aliquot, −TBA solution [20% (w/v) TCA and 0.01% (w/v) BHT] was added and treated in the same way. The samples were cooled on ice and centrifuged at 10 000 g for 10 min. The absorbance of the supernatant was measured at 440, 532 and 600 nm. MDA content was calculated according to Hodges et al. (1999).

Measurements of carbon and nitrogen contents

Carbon and nitrogen contents were measured with a CN analyser (Vario EL III; Elementar Analysensysteme GmbH, Hanau, Germany). After the measurement of the respiratory rate, leaves were dried at 80 °C for 3 d, and then powdered. The powdered dry samples were analysed.

Measurements of non-structural carbohydrate contents

Contents of non-structural carbohydrates (starch, glucose and sucrose) were measured according to Ono, Terashima & Watanabe (1996). Frozen leaves were ground with Multi-Beads Shocker (Yasui Kikai, Osaka, Japan) using zirconia beads at 2000 rpm at 4 °C for 5 min in 80% ethanol. The suspension was incubated at 80 °C for 10 min, and centrifuged at 1500 g at 4 °C for 10 min. The precipitation was used for determination of starch. The supernatant was evaporated to remove ethanol with a centrifugal concentrator (CC-105; Tomy, Tokyo, Japan). The same volumes of distilled water and chloroform were added to the concentrated supernatant and mixed well. The mixture was centrifuged at 10 000 g at 4 °C for 10 min, and the upper aqueous phase was used for determination of glucose and sucrose. The precipitation of starch was suspended with distilled water and boiled for 60 min at 100 °C. An equal volume of amyloglucosidase was added to the boiled suspension and incubated for 60 min at 55 °C. The mixture was centrifuged at 10 000 g at 4 °C for 10 min, and the upper aqueous phase was used for determination of starch. Glucose-C2-test (Wako, Osaka, Japan) was used for detection.

Measurements of total amino acid content

Amino acid content was measured according to Rosen (1957). Leaf extracts that were prepared for determination of glucose were used for detection of amino acids. Leucine was used as a standard.

Statistical analyses

All statistical analyses were conducted with STATVIEW (version 5.0J; SAS, Cary, NC, USA).

RESULTS

Isolation of an AOX1a T-DNA insertion line

To examine whether AOX plays a role in the acclimation of respiration to low temperature, we set out to isolate an aox1a knock-out transgenic line.

We obtained lines with T-DNA insertions from the ABRC. From the SALK_084897 insertion line, we isolated a homozygous line for the T-DNA insertion. In this line, we checked for the presence of the T-DNA insertion by a genomic PCR with T-DNA-specific and AOX1a-specific primers (Fig. 1b). In contrast to WT, a PCR product of AOX1a was not amplified with AOX1a-specific primers in the aox1a line (lane 2 in Fig. 1b). We confirmed by sequencing that T-DNA was located at 918 nucleotides downstream of the start codon in an exon of the AOX1a gene (Fig. 1a). Figure 1c shows the result of an RT-PCR with AOX1a-specific primers. In the lane of the aox1a line, a pale band was found. However, the size was slightly smaller than that of WT. We performed a sequencing analysis of this PCR product and confirmed that it was a part of AOX1a cDNA (data not shown). Therefore, we conclude that the aox1a line lacks normal AOX1a expression. At 25 °C, total respiration rate (total R) of this aox1a line was similar to WT, but we found a clear reduction in the CN-resistant R in seedlings (data not shown) and in rosettes (Fig. 2c, 0 h) of this aox1a line.

Figure 1.

Confirmation of the T-DNA insertion in the AOX1a gene. (a) The position of the T-DNA insertion in the AOX1a gene of the SALK_084897 line. The AOX1a gene has four exons (represented as boxes), and T-DNA is located at 918 nucleotides downstream of the start codon. The orientation of the left border (LB) is indicated. (b) Genomic PCR revealed a homozygous T-DNA insertion in the aox1a line. Using a T-DNA-specific primer and AOX1a-specific primers, a PCR product of about 900 bp was amplified from the genomic DNA of the aox1a line (lane 4), but not from a wild-type (WT) control (lane 3). On the other hand, using AOX1a-specific primers, a 1608 bp PCR product was amplified from WT (lane 1), but not from the aox1a line (lane 2). (c) RT-PCR showed the absence of AOX1a transcript in the aox1a line. Using AOX1a-specific primers spanning the T-DNA insertion site, a 734 bp PCR product was amplified from the WT cDNA, but not from aox1a cDNA.

Figure 2.

Changes in total respiration rate (total R) and CN-resistant respiration (CN-resistant R) in the leaves of wild type (WT)–10 °C, WT–4 °C and aox1a–4 °C measured at 25 °C (a–c) and at each treatment temperature (d–f). (g–i) Changes in the degree of homeostasis (H) of total R and CN-resistant R in WT–10 °C, WT–4 °C and aox1a–4 °C after low-temperature treatment. Closed and open circles represent the total R and CN-resistant R, respectively. Circle and triangle symbols represent the WT and aox1a lines, respectively. The CN-resistant R was measured in the presence of 2 mm KCN. Means ± SEM are shown (n = 3–11). *, ** and *** denote statistically significant difference between WT–4 °C and aox1a–4 °C at each treatment time using Student's t-test (P < 0.05, P < 0.01 and P < 0.001).

Respiratory rates of leaves after low-temperature treatments

Total R and CN-resistant R of WT leaves treated at two different low temperatures (WT–10 °C or WT–4 °C) were measured at 25, 10 and 4 °C. Similarly, leaves of aox1a treated at 4 °C (aox1a–4 °C) were measured at 25 and 4 °C. Figure 2 shows the respiratory rate of WT–10 °C leaves (Fig. 2a,d), WT–4 °C leaves (Fig. 2b,e) and aox1a–4 °C leaves (Fig. 2c,f). Both total R and CN-resistant R of WT–10 °C leaves gradually increased [Fig. 2a,d; P < 0.001, one-way analysis of variance (anova), see Supplementary Table S2]. The total R of WT–4 °C increased more slowly than that of WT–10 °C (Fig. 2b,e). On the other hand, the CN-resistant R of WT–4 °C significantly increased in a similar slope to that of WT–10 °C when measured at 25 °C. The CN-resistant R of WT–10 °C and WT–4 °C was inhibited by 20 mm salicylhydroxamic acid (SHAM; an AOX inhibitor, data not shown). The total R of aox1a–4 °C increased during the low-temperature treatment (Fig. 2c,f). When measured at 25 °C, the total R of aox1a–4 °C was similar to that of WT–4 °C (Fig. 2b,c). The CN-resistant R of WT–4 °C increased by the treatment, but was much lower in aox1a–4 °C. After 12 h low-temperature treatment, the total R of aox1a–4 °C measured at 4 °C was significantly higher than that of WT–4 °C (Fig. 2e,f; P < 0.001, two-way anovain Supplementary Table S3). At 12 and 24 h after the treatment, the CN-resistant R of aox1a–4 °C slightly increased. However, because this increased respiration rate was not inhibited by 20 mm SHAM, it could not be ascribed to AOX (data not shown). We consider that it is a residual O2 consumption separate from respiration.

To quantify the level of temperature acclimation, we calculated Q10 values (proportional change in respiration per 10 °C). A high Q10 value indicates a strong temperature sensitivity of respiration. The Q10 values for CN-resistant R were consistently lower than those of total R throughout the experimental period, both in WT–10 °C and in WT–4 °C (Table 1). This suggests that AOX was less sensitive to low temperatures than CP. In addition, we calculated a parameter for homeostasis (H) according to Loveys et al. (2003), representing the degree of respiratory acclimation. In WT–10 °C, H of both total R and CN-resistant R increased by the treatment (Fig. 2g). In WT–4 °C, H of total R increased slowly (Fig. 2h), but H of CN-resistant R increased similar to that of WT–10 °C. These results suggest that AOX activity was promptly induced, and had a high ability to acclimate to low temperature. H of total R in the aox1a–4 °C was similar to that of WT–4 °C (Fig. 2i). Because we previously showed that the CN-resistant R in aox1a–4 °C was not a result of AOX activity, corresponding values of Q10 and H were not shown.

Table 1.  The Q10 values of total respiratory rate (total-R) and CN-resistant respiratory rate (CN-resistant R) in leaves of wild type (WT)–10 °C, WT–4 °C and aox1a–4 °C
 Treatment time
0 h3 h6 h12 h24 h72 h120 h
  1. CN-resistant R in aox1a–4 °C was not determined because there was almost no CN-resistant R in aox1a line.

  2. –, not determined.

WT–10 °C
 Total R2.712.393.303.462.68
 CN-resistant R2.212.032.732.312.30
WT–4 °C
 Total R2.322.822.603.323.053.182.53
 CN-resistant R1.831.921.701.702.092.472.21
aox1a–4 °C
 Total R2.262.232.572.262.192.152.37

Transcriptional changes in components of the respiratory pathway in leaves treated at low temperature

In A. thaliana, five AOX genes (AOX1a, 1b, 1c, 1d and AOX2) have been identified (Saisho et al. 1997; Considine et al. 2003; Thirkettle-Watts et al. 2003), and the genes encoding for NDin (NDA1, 2 and NDC1) and for NDex (NDB1, 2, 3, 4) have been identified (Rasmusson et al. 2004). Six UCP genes (UCP1, 2, 3, 4, 5, 6) have been identified (Maia et al. 1998; Watanabe et al. 1999; Boreckýet al. 2006), although activities of some UCP gene products have not been confirmed (Hourton-Cabassa et al. 2004). Transcript levels of AOX1a-1d, AOX2, NDB2, UCP1-5 and cytochrome c oxidase 6b subunit gene (COX6b) were measured using a real-time PCR system. As an internal control, we used 40S ribosomal protein subunit 15aA gene (Rps15aA) and elongation factor 1-alpha gene (EF-1-a; data not shown). The transcript levels were calculated from the difference in the threshold cycle (Ct) between Rps15aA and each gene. Figure 3 shows time-dependent changes in transcript levels. At 3 h after the start of the 10 °C treatment, the transcript level of AOX1a in WT–10 °C transiently and significantly increased three- to fourfolds (Fig. 3, Dunnett multiple comparison test in Supplementary Table S2). The expression of UCP5 also increased transiently at 3 h. In WT–4 °C, the transcript level of AOX1a peaked at 12 h after the start of the 4 °C treatment (Fig. 3). The expression of NDB2 and UCP5 increased in a similar manner, although the increase was only marginally significant in the case of NDB2 (P = 0.0776). UCP1 and COX6b were constitutively expressed throughout the treatment in both WT–10 °C and WT–4 °C. The transcript levels of AOX1b-1d, AOX2 and UCP2, 3 were very low (Fig. 3), although a small increase could be observed during the treatments. Relative transcriptional changes in each gene during low-temperature treatment are shown in Supplementary Table S4. The transcript levels of the respiratory components in roots did not change in both WT–10 °C (Supplementary Table S5) and WT–4 °C (data not shown). AOX1a expression in roots was about 10% of that in leaves.

Figure 3.

Changes in transcript levels of AOXs, NDB2, UCPs and COX6b genes in leaves of wild type (WT)–10 °C, WT–4 °C and aox1a–4 °C after low-temperature treatments. The transcript levels were calculated from the difference in the threshold cycle (Ct) between Rps15aA and each gene of interest. Means ± SEM are shown (n = 2–7). * denotes statistically significant differences between WT–4 °C and aox1a–4 °C at each treatment time using Student's t-test (P < 0.05).

We also examined the respiration rate and transcript level of respiratory components in WT grown for 120 h at 23–25 °C and 14 h light period (C-120 h), and also in WT plants grown for 24, 72 and 120 h at 23–25 °C under continuous light conditions (L-24, 72 and 120 h). There were only small differences in respiration or transcription between plants of this experiment and the WT plants measured before the start of the low-temperature treatment. In C-120 h and L-120 h, the total R decreased about 79 and 75%, and the CN-resistant R decreased 69 and 98%, respectively. AOX1a expression in C-120 h decreased 86% and in L-120 h increased 157%. These results indicate that ageing and continuous irradiation have only marginal effects on the data presented in Figs 2 and 3.

Next, we compared the transcript levels of respiratory components between WT–4 °C and aox1a–4 °C after low-temperature treatment. The transcript levels of AOX1a-1d, AOX2, NDB2, UCP1-5 and COX6b were measured. In the case of aox1a–4 °C, we used a different set of primers to amplify the modified aox1a transcript. The transcript level of AOX1a in aox1a–4 °C was very low, and the result of non-specific amplification (data not shown). Similar to WT–4 °C, the transcript levels of other AOX isoforms were low in aox1a–4 °C. The expression pattern of NDB2 was quite different from that of in WT–4 °C. NDB2 transcripts in aox1a–4 °C gradually increased during the low-temperature treatment, and the level at 120 h was about 10 times the level at 0 h. NDA1 expression decreased in aox1a–4 °C during the 4 °C treatment similar to what was observed for WT–4 °C. The transcription of UCP2-5 was similar to that of WT–4 °C. The transcript level of UCP1 gradually increased in aox1a–4 °C, and was significantly higher than in WT–4 °C at 24 and 72 h after the start of the treatment. There was a slight difference in the expression pattern of COX6b between WT–4 °C and aox1a–4 °C. The transcript level of COX6b in WT–4 °C decreased, whereas that in aox1a–4 °C remained almost constant, resulting in a significant difference in expression at 120 h after the start of the treatment.

Chlorophyll fluorescence and MDA content of leaves in WT and aox1a after low-temperature treatment

There was no visible phenotypic difference between WT and aox1a plants, even after 120 h of low-temperature treatment (data not shown). A number of physiological properties were analysed in WT–4 °C and aox1a–4 °C. We measured the maximal efficiency of photosystem II (Fv/Fm), the operating efficiency of photosystem II (ΦPSII) and non-photochemical quenching under 70 µmol photon m−2 s−1 (Fig. 4a–c). There were no significant differences between WT–4 °C and aox1a–4 °C in all three parameters either at room temperature or at 4 °C (see Supplementary Table S3).

Figure 4.

Changes in chlorophyll fluorescence and malondialdehyde (MDA) content of leaves in wild type (WT)–4 °C and aox1a–4 °C after low-temperature treatment. We measured the maximal efficiency of photosystem II (Fv/Fm) (a), the operating efficiency of photosystem II (ΦPSII) (b) and non-photochemical quenching (NPQ) (c). Solid lines represent values measured at 25 °C; dotted lines represent values measured at 4 °C. (d) MDA content. Means ± SEM are shown (a–c; n = 4–7, d; n = 4). * denotes statistically significant differences between WT and the aox1a line at each time-point using a Student's t-test (P < 0.05).

To assess the degree of oxidative stress in both lines, we measured MDA content in the leaves treated at 4 °C using a modified TBARS assay (Hodges et al. 1999). The MDA content in WT–4 °C initially increased, but gradually decreased again after 24 h (Fig. 4d). Finally, at 120 h, the MDA content was lower than that at 0 h. The change in the MDA content of aox1a–4 °C showed a pattern very similar to that of WT–4 °C. However, the level of MDA in aox1a–4 °C was significantly lower compared with that in WT–4 °C throughout the whole treatment period (P < 0.001, two-way anova, Supplementary Table S3).

A transcriptional analysis of several antioxidants after low-temperature treatment

Amirsadeghi et al. (2006) revealed that expression of antioxidant defence enzymes was up-regulated in tobacco leaves lacking AOX compared with WT.

We presumed that genes encoding antioxidant defences were up-regulated in aox1a–4 °C in which the MDA content was constitutively lower than in WT–4 °C, and thus investigated the expression of a number of antioxidant defence enzymes. We analysed those antioxidants that are known to be localize in the mitochondria [manganese superoxide dismutase 1 (MSD1), glutathione peroxidase 6 (GPX6), glutathione S-transferase class phi 6 (GSTF6)]. In addition, we determined the expression of antioxidants targeted to the cytosol [GPX2, ascorbate peroxidase 1 (APX1), APX2 and copper–zinc superoxide dismutase 1 (CSD1)], and to the chloroplast [iron superoxide dismutase 1 (FSD1)]. We also determined the expression of catalase 1 (CAT1) targeted to the peroxisome or mitochondrion (McClung 1997), and of 2-alkenal reductase (AER) targeted to the cytosol and nuclei (Mano et al. 2005). Some ROS scavengers, such as GPX6, CAT1, MSD1 and FSD1, were more strongly induced in aox1a–4 °C during the treatment (Fig. 5; Supplementary Table S3). The expression of CSD1 decreased with the treatment time in WT–4 °C, but not in aox1a–4 °C (P < 0.05, two-way anova, Supplementary Table S3). The transcript level of APX1 in aox1a–4 °C was higher than that in WT–4 °C after 24 h, but not significant. The situation was reversed in the case of APX2 that was higher in WT–4 °C than in aox1a–4 °C, but the APX2 transcript level was very low compared with APX1. The expression of GPX2 was similar between the two genotypes, except for a lower expression in the aox1a line at 24 h after the treatment. In addition, we examined the transcription of AER and GSTF6. These proteins are thought to detoxify lipid peroxidation products (Sweetlove et al. 2002; Mano et al. 2005). There was no significant difference in the expression of AER, but the expression of GSTF6 in WT–4 °C was significantly higher than that in aox1a–4 °C.

Figure 5.

Changes in the transcript levels of antioxidant genes in leaves of wild type (WT)–4 °C (circles) and aox1a–4 °C (triangles) after low-temperature treatment. The transcript levels were calculated from the difference in threshold cycle (Ct) between Rps15aA and each gene. Means ± SEM are shown (n = 4). * and ** denote statistically significant differences between WT and the aox1a line at each time-point using a Student's t-test (P < 0.05 and P < 0.01, respectively).

Changes in carbon and nitrogen content of leaves after low-temperature treatment

Previous studies suggested that AOX contributes to the maintenance of the C and N balance in cell culture (Sieger et al. 2005) and in spinach leaves (Noguchi & Terashima 2006) under N-limiting conditions. To examine whether AOX contributes to maintaining the C and N balance under low-temperature stress, we measured C and N contents in both WT and aox1a after low-temperature treatment. In Both lines, C and N contents on an FW basis increased during the treatment (Fig. 6a,b). The N content in aox1a–4 °C did not differ from that of WT–4 °C, except at 120 h, when the N content in aox1a–4 °C was lower than in WT–4 °C. On the other hand, the C content in aox1a–4 °C was slightly higher at 72 and 120 h compared with WT–4 °C. There were significant differences between the two genotypes in both C and N content (Supplementary Table S3). In both lines, the C/N ratio increased during low-temperature treatment (Fig. 6c). In aox1a–4 °C, the C/N ratio was significantly higher at 72 and 120 h than that of WT–4 °C. These results suggest that there is difference in C and N metabolism between WT and aox1a at low temperature.

Figure 6.

Changes of C (a) and N content on a fresh weight (FW) basis (b), and the C/N ratio (c) in leaves of wild type (WT) (circles) and aox1a (triangles) after low-temperature treatment (4 °C). Means ± SEM are shown (n = 4). * denotes a statistically significant difference between WT and aox1a at each time-point using a Student's t-test (P < 0.05).

Changes in the amount of non-structural sugars and amino acids after low-temperature treatment

It is known that low-temperature treatment can lead to an accumulation of non-structural sugars (Wanner & Junttila 1999). Because a higher carbohydrate content could provide an explanation for the increased C/N ratio of the aox1a line, we examined the amount of starch, glucose and sucrose in both genotypes. The starch content increased linearly during the treatment in WT–4 °C and aox1a–4 °C (Fig. 7a). At 72 and 120 h after the start of the treatment, the starch content in aox1a–4 °C was higher than in WT–4 °C. The glucose and sucrose contents were similar in WT–4 °C and aox1a–4 °C, and increased about fourfolds until 72 h, after which a gradual decrease was observed (data not shown).

Figure 7.

Changes in the content of non-structural sugars (a) and amino acids (b) in leaves of wild type (WT) (circles) and aox1a (triangles) after low-temperature treatment. Means ± SEM are shown (n = 4). * denotes a statistically significant difference between WT and aox1a at each time-point using a Student's t-test (P < 0.05).

We also measured total amino acids and soluble proteins. The amino acid content increased during low-temperature treatment in both WT–4 °C and aox1a–4 °C (Fig. 7b). The total amount of amino acids was slightly higher in aox1a–4 °C than in WT–4 °C, but we found no significant difference between both lines. The soluble protein content was constant throughout the treatment in both lines, but slightly lower in the case of aox1a–4 °C (data not shown).

DISCUSSION

We showed that at low temperatures, AOX1a gene expression is induced within 12 h, and thus, CN-resistant R increased in A. thaliana leaves. However, in this study, we found that respiratory rate acclimated to low temperature under AOX deficit condition. The induced AOX may contribute to the maintenance of the C and N balance at low temperature. The role of AOX in the prevention of ROS production can be complemented by antioxidant defence systems at low temperature. The induction of AOX1a expression by low temperature was accompanied by an increase in the expression of NDB2 and UCP5. Therefore, these genes may function cooperatively at low temperature.

Acclimation of leaf respiration to low temperature

The CN-resistant R was significantly increased by low-temperature treatments of WT plants (Fig. 2). This increase correlated to the rise in AOX protein amount (data not shown). The Q10 values of CN-resistant R were lower than those of total R, and homeostasis of CN-resistant R increased by low temperatures (Fig. 2g,h). These results suggest that AOX has a high ability for low-temperature acclimation. It is consistent with previous reports in other species (Collier & Cummins 1990; Stewart et al. 1990; Vanlerberghe & McIntosh 1992; Gonzàlez-Meler et al. 1999). The total R in aox1a was higher than that of WT, when measured at 4 °C (Fig. 2f). In aox1a–4 °C, the expressions of NDB2 and UCP1 were induced (Fig. 3). As NDB transfers an electron from the cytosol to ubiquinone, we conceive that the increase of NDB2 could raise the respiratory rate. UCP dissipates the H+ gradient, which increases the electron transport. It is indicated that AOX1a and NDB2 genes are co-expressed as a result of having common elements in the promoter regions (Clifton et al. 2005). We found that the expression of the AOX1a mRNA sequence upstream of the T-DNA insertion point was induced in aox1a–4 °C (data not shown). Therefore, it is possible that, in the aox1a line, a higher expression of NDB2 and UCP1 protein leads to an increase of the respiratory rates through the CP and complements the lack of AOX.

The role of AOX in leaves at low temperature

Although there were no significant differences in chlorophyll fluorescence parameters between WT–4 °C and aox1a–4 °C (Fig. 4), the C/N ratio was higher in aox1a–4 °C than in WT–4 °C (Fig. 6). The increase in the C/N ratio was mainly the result of starch accumulation (Fig. 7a). The increment in starch from 0 to 120 h accounted for 30% of that in C content. In aox1a–4 °C, the starch was more accumulated than in WT–4 °C, although the total R was higher than in WT–4 °C when measured at 4 °C (Fig. 2f). We think that effect of respiration on starch content is not so large, in particular at low temperature. Although it is hard to say what makes a big influence to the starch content in aox1a line, this accumulation in aox1a–4 °C may be attributed to a perturbation in the C and N metabolism. It has been suggested that AOX consumes excess sugars, and thereby maintains the C/N ratio (Sieger et al. 2005; Noguchi & Terashima 2006). However, our results indicate that AOX only indirectly influences the C/N ratio, because the total R in aox1a–4 °C was higher than in WT–4 °C, and carbohydrates would be consumed by the respiratory system in WT–4 °C more than in aox1a–4 °C. Dutilleul et al. (2005) have examined the NAD-dependent metabolic pathways in a cytoplasmic male sterility mutant which lacks the mitochondrial gene nad7 and a functional complex I. They suggested that modulation of NADH levels by the mitochondrial respiratory chain has far-reaching repercussions for the integration of C and N metabolisms (Dutilleul et al. 2005). We assumed that the lack of AOX1a would cause a modulation of NADH and thereby of TCA cycle intermediates, which lead to the alteration of the C and N metabolism. The induction of NDB2 and thereby the increase of ND activity might lead to the modulation of NADH levels.

Malondialdehyde is one of the lipid peroxidative products (Møller, Jensen & Hansson 2007) and can be used as an index for oxidative stress (Hodges et al. 1999; Fiorani et al. 2005). The MDA content in A. thaliana cells increased by hydrogen peroxide treatment (Sweetlove et al. 2002). In contrast to our expectation, the MDA content in aox1a–4 °C was lower than that in WT–4 °C even in untreated plants (Fig. 4d). It is probable that basal activities of antioxidant defence systems in aox1a plants are higher than in WT, although the transcript levels of measured antioxidant genes were not higher in aox1a–4 °C before the start of the treatment (Fig. 5). In tobacco leaves lacking AOX, genes encoding antioxidant defences were up-regulated compared with WT (Amirsadeghi et al. 2006). Recently Giraud et al. (2008) reported that expressions of FSD2, 3 and other antioxidant enzymes were up-regulated in AOX1a T-DNA insertion lines, one of which was identical with that we used, aox1a. These results suggested that, at low temperature, AOX may directly prevent ROS production, but this function can be complemented by other antioxidant defence systems. We presume that the expression of antioxidant defence systems was induced by the ROS generation which was caused by the lack of AOX, and thereby the basal activities of antioxidant defence systems were higher and the lipid peroxidation was lower in aox1a. Taken together, the lack of AOX can influence the C and N metabolism via imbalance in the respiratory chain, and the antioxidant defence systems via resulting generation of ROS. AOX may be a necessary component for the maintenance of a balanced metabolism and antioxidant defence system.

Fiorani et al. (2005) examined differences in MDA content of WT, aox1a antisense and AOX1a over-expressing plants of A. thaliana grown at 12 °C. They showed that the MDA content in the aox1a antisense line was higher than in WT (Fiorani et al. 2005). In addition, they found that anthocyanin content decreased in the aox1a antisense line grown at 12 °C (Fiorani et al. 2005). In our study, there was no apparent difference in the leaf colour between WT–4 °C and aox1a–4 °C (data not shown). It may be that the discrepancy between these results is because of the difference in the duration and intensity of the low-temperature treatment and in the remaining CN-resistant respiratory rate. AOX1a antisense plants which Fiorani et al. (2005) used remained about 27% of CN-resistant R (Umbach, Fiorani & Siedow 2005; Supplementary Table S1), although there was hardly remaining of the CN-resistant R in aox1a plants used in this study.

Transcriptional cooperation between AOX1a, NDB2 and UCP5

The transcript levels of AOX1a and UCP5 increased 3 h after the start of the 10 °C treatment (Fig. 3). In WT–4 °C, the AOX1a transcript peaked at 12 h, and the transcription patterns of NDB2 and UCP5 were similar to that of AOX1a (Fig. 3). These results suggest that mRNAs of AOX1a, NDB2 and UCP5 are simultaneously induced by low-temperature stress, and are in good agreement with the finding that AOX1a and NDB2 genes are co-expressed (Clifton et al. 2005). The peaks in the expression of these genes were occurring later during the 4 °C treatment compared with the situation in the 10 °C treatment. This lag may be the result of a slow-down in reaction kinetics at 4 °C. In both treatments, UCP1 was constitutively expressed and not induced by low temperature. This result is in disagreement with Maia et al. (1998), who showed that expression of UCP1 in A. thaliana plants was induced after 48 and 72 h after the start of the 4 °C treatment. On the other hand, the transcript levels of NDA1 decreased during the 4 °C treatment in both WT–4 °C and aox1a–4 °C, confirming previous results by Svensson et al. (2002).

In aox1a–4 °C, the other AOX isoforms did not complement the lack of AOX1a (Fig. 3). The transcript levels of AOX isoforms differ among organs or stress treatments (Clifton, Millar & Whelan 2006). Even if there is no AOX1a expression and thus AOX protein, the other isoforms may not be induced by low-temperature stress. Instead, the transcript levels of NDB2 and UCP1 were increased (Fig. 3). The increase in these components may contribute to the increase in the total R in aox1a–4 °C. We found that the expression of the AOX1a mRNA sequence upstream of the T-DNA insertion point was induced (data not shown), suggesting that common promoter regions of AOX1a and NDB2 are driven by signals deriving from low temperature.

In this study, using an aox1a knock-out line, we demonstrated that although AOX has high ability for acclimation to low temperature, other respiratory components may complement the total rate of respiration at low temperature. Because, at low temperature, the lack of AOX results in a disruption of C and N balance, we suggest that AOX may function to maintain the balance between C and N metabolism. The up-regulation of antioxidant defence systems in aox1a complemented the antioxidant function of AOX. The other ATP-uncoupling components, NDB and UCP, may function cooperatively and complementary with AOX at low temperatures.

ACKNOWLEDGMENTS

The ABRC provided the SALK_084897 T-DNA insertion line of Salk Institute. We are grateful to Profs. K. Takimoto and K. Matsui, and Dr. T. Ueda for their advice in molecular techniques; Dr. M. Kawaguchi for the use of real-time PCR instrument; Dr. J. Mano for the information about antioxidant enzymes and aldehydes; and Dr. D. Tholen for the critical reading of this manuscript. We also thank laboratory members for technical support, advice and encouragement. This study was supported by grants from the Ministry of Education, Science, Sports, and Culture of Japan (no. 17051019, 19039009).

Ancillary