Disruption of the mitochondrial alternative oxidase (AOX) and uncoupling protein (UCP) alters rates of foliar nitrate and carbon assimilation in Arabidopsis thaliana

Summary We demonstrated that mitochondrial respiration contributes to energy balancing between nitrogen and carbon metabolism by competing for reductant with de novo NO3 – assimilation and avoiding chloroplastic excess reductant.


Introduction
Nitrogen (N) availability is a major determinant of plant growth and productivity (Reich et al., 1997(Reich et al., , 2006a. However, N fertilizers also represent a large economical cost and source of ground water pollution in agriculture systems. In higher plants, nitrate (NO 3 -) and ammonium (NH 4 + ) are the two primary forms of assimilated inorganic N. The latter is typically assimilated directly into amino acids through the glutamine synthetase-glutamate synthase complex, whereas the former is first reduced to nitrite by cytosolic nitrate reductase and then to NH 4 + by plastidic nitrite reductase (Epstein and Bloom, 2005). The assimilation of NO 3 has a higher energetic requirement compared with NH 4 + assimilation (Noctor and Foyer, 1998;Scheurwater et al., 1999;Escobar et al., 2006) and in the leaves the cytosolic reduction of nitrate to nitrite may consume reductant exported from either the chloroplast or the mitochondria (Foyer et al., 2011). On the other hand, nitrite reduction and NH 4 + assimilation occur in the chloroplast stroma and consume reduced ferredoxin. Therefore, the complexity and compartmentalization of these pathways necessitates balancing photosynthetic energy supply and demand to optimize rates of NO 3 and CO 2 assimilation. The availability of N is known to affect rates of photosynthesis and respiration (Reich et al., 2006b;Atkinson et al., 2007), and there is a strong correlation between leaf N content and rates of respiration (Terashima and Evans, 1988;Makino and Osmond, 1991;Byrd et al., 1992;Lusk and Reich, 2000). For example, the tricarboxylic acid (TCA) cycle enzymes fumarase, NAD-dependent isocitrate dehydrogenase (Makino and Osmond, 1991), and NAD-dependent malic enzyme (Noguchi and Terashima, 2006) are up-regulated under low N. Additionally, the expression and activity of several glycolytic and TCA cycle enzymes were differentially influenced following NH 4 + or NO 3 feeding (Larsen et al., 1981;Scheible et al., 1997;Lancien et al., 1999;Stitt, 1999). These changes in respiration are linked to the supply of carbon skeletons (e.g. 2-oxoglutarate, isocitrate, and citrate) from the TCA cycle to maintain optimum N assimilation and amino acid biosynthesis (Larsen et al., 1981;Scheible et al., 1997;Lancien et al., 1999). Furthermore, the N source (NO 3 versus NH 4 + ) has been shown to change gene expression of the mitochondrial electron transport chain, particularly the alternative oxidase (AOX), type II NAD(P)H dehydrogenases, and uncoupling proteins (UCPs) of the mitochondrial alternative electron transport (mAET) (Escobar et al., 2006;Patterson et al., 2010). Additionally, the capacity and protein amount of AOX, a major component of mAET, increase under low NO 3 conditions (Sieger et al., 2005;Watanabe et al., 2010;Hachiya and Noguchi, 2011). The mAET bypasses one or more of the multiprotein complexes of the 'classic' electron transport chain, minimizing proton pumping across the inner membrane (Vanlerberghe and McIntosh, 1997;Rasmusson et al., 2004). Consequently, mAET oxidizes NAD(P)H uncoupled from ATP production and has been proposed to dissipate excess reductant (Raghavendra and Padmasree, 2003). For example, mAET plays an important role in the response to several environmental constraints such as cold (Armstrong et al., 2008;Watanabe et al., 2008), elevated CO 2 (Gandin et al., 2012), drought (Bartoli et al., 2005;Giraud et al., 2008), phosphate limitation (GonzàLez-Meler et al., 2001;Sieger et al., 2005;Escobar et al., 2006), high light stress (Ribas-Carbo et al., 2000;Sweetlove et al., 2006;Yoshida and Noguchi, 2009), and other reactive oxygen species-inducing stress conditions (Maxwell et al., 1999). Additionally, it has been suggested that the capacity of mAET responds to changes in NO 3 assimilation (Dutilleul et al., 2005;Escobar et al., 2006).
Reductant availability within the leaf cytoplasm probably often limits rates of de novo NO 3 assimilation. This is in part due to the low cytosolic NADH availability (0.3-0.7 μM), which is far below the 7 μM K m of nitrate reductase for NADH (Kaiser et al., 2000;Heineke et al., 2001). Therefore, conditions such as high light or perhaps high rates of photorespiration that increase cytosolic NADH concentrations have been suggested to increase rates of foliar NO 3 assimilation (Bloom et al., 2002;Searles and Bloom, 2003;Rachmilevitch et al., 2004;Guo et al., 2007). Alternatively, NO 3 assimilation decreases under conditions that restrict the export of reductant via the malate shuttle from the chloroplast. It has also been reported in the literature that changes in mitochondrial electron transport, particularly the alternative non-phosphorylating pathways, could influence the de novo assimilation of N (Watanabe et al., 2008). However, it remains unclear how the alternative non-phosphorylating pathways of the mitochondrial inner membrane influence foliar NO 3 assimilation. Additionally, it has been shown that both the alternative non-phosphorylating pathways of the mitochondrial inner membrane and N assimilation oxidize excess reductant produced by photosynthesis, photorespiratory glycine oxidation, and the TCA cycle (Sweetlove et al., 2006;Gandin et al., 2012). Therefore, the energy partitioning in the cell is probably balanced in part by both mitochondrial electron transport and NO 3 assimilation (Hachiya and Noguchi, 2011), avoiding chloroplast over-reduction and maintaining optimal rates of photosynthesis. The role of the mAET and NO 3 assimilation in energy consumption and dissipation is well accepted; however, their respective contribution and coordination remain unclear. Therefore, the aim of this research is to test the hypothesis that changes in mAET and NO 3 assimilation influence energy partitioning between N and carbon metabolism. To test this hypothesis, this study (i) investigated the influence of changes in mAET capacity on de novo NO 3 assimilation and (ii) determined the response of photosynthetic CO 2 assimilation and electron transport to changes in N source (NO 3 versus NH 4 + ) in plants with disrupted mAET [wild type (WT) versus aox1a and ucp1].

Materials and methods
Plant material and growth conditions WT and T-DNA insertion lines for AOX1a (SALK_084897) and UCP1 (SAIL_536G01) plants of Arabidopsis thaliana (Sweetlove et al., 2006;Giraud et al., 2008;Gandin et al., 2012) were grown hydroponically in a controlled environment growth chamber (Biochambers GC-16, Winnipeg, Manitoba, Canada) at a photosynthetic photon flux density (PPFD) of 160 μmol quanta m -2 s -1 at plant height, relative humidity of 50%, and air temperature of 23 °C and 18 °C during the day and night, respectively, with a 10 h day. The SAIL_563G01 line was obtained from the TAIR collection and homozygous lines were screened by PCR of genomic DNA using the GACGAAGATGTGAAGTAGACC/ TAGCATCTGAATTTCATAACCAATCTCGATACAC and GACGAAGATGTGAAGTAGACC/TCAGTTTCTTTTGGACG CATCG primer pairs. Homozygous lines were selfed and screened again by PCR using the same primer pair. Seeds were germinated on Rockwool cylinders (GroDan Cubes, Rockwool BV, Roemond, The Netherlands) for 7 d. Subsequently, seedlings were transferred to 14 litre containers filled with aerated nutrient solution containing 0.2 mM NH 4 Cl, 0.2 mM KNO 3 , 1.25 mM CaSO 4 , 0.75 mM MgSO 4 , 0.5 mM KH 2 PO 4 , 0.04 g l −1 FeDPTA, and micronutrients (Gibeaut et al., 1997). Nutrient solution was replaced every 2 d.
Nitrate uptake, content and assimilation WT, aox1a and ucp1 A. thaliana plants were grown as described above in a hydroponic solution containing 0.2 mM NO 3 at natural abundance 15/14 N. Subsequently, plants were grown for 24 h on a nutrient solution depleted of N (as above). Half of the plants were shifted to a nutrient solution supplemented with 0.2 mM NO 3 enriched 25% with 15 NO 3 and the other half were fed natural abundance 15 NO 3 as a control. In the first set of experiments, plants were labelled for up to 9 h to look at the time course of NO 3 uptake and assimilation (time effect, Supplementary Fig. S1 available at JXB online). Subsequent experiments were limited to 6 h of feeding (irradiance effect, Fig. 1). After labelling, plants were rinsed in ultra-pure water then separated into shoots and roots, oven-dried, and ground to a fine powder in a mortar and pestle. Total 15 N enrichment was measured using an elemental analyser (ECS 4010, Costech Analytical, Valencia, CA, USA) connected directly to a continuous flow isotope ratio mass spectrometer (Delta PlusXP, Thermofinnigan, Bremen, Germany) (Brenna et al., 1997). Isotopic reference materials are interspersed with samples for isotope ratio calibration and acetanilide was used in a multipoint correction for N%.
Free NO 3 was extracted from powder plant material, and 15 N enrichment of free NO 3 was estimated by converting NO 3 to N 2 O using the denitrifying Pseudomonas aureofaciens (Sigman et al., 2001;Casciotti et al., 2002). The headspace N 2 O was sampled with a two-holed needle mounted on an autosampler (GC-PAL, CTC Analytics, Switzerland) and connected to a GasBench II (ThermoFinnigan) interface. Samples were cleaned of water and volatile organic compounds (VOCs) through a liquid nitrogen/ethanol slush trap (-110 °C), and of CO 2 and H 2 O through an ascarite/ magnesium perchlorate trap. Further removal of VOCs was achieved with a Supelco type F trap following the slush trap. Purified samples were separated through a Poraplot Q GC column (Varian, 25 m×0.32 mm ID), run through a final nafion water trap (Permapure LLC, NJ) for trace water removal, and analysed by a continuous flow isotope ratio mass spectrometer (Delta PlusV, Thermofinnigan) (Brenna et al., 1997). NO 3 assimilation was thus estimated as NO 3 -assimilation=total 15 N-15 NO 3 -. Finally, total free NO 3 was quantified from powder using high-pressure liquid chromatography (Thayer and Huffaker, 1980).

Amino acid analysis
Leaf tissues were ground in 0.25 ml of 0.1 M HCl using a micropestle, and 400 μM aminobutyric acid was added as internal control. Samples were centrifuged for 20 min at 24 000 g at 4 °C and the supernatant was collected and the pellets re-extracted with 0.25 ml of 0.1 M HCl. Supernatants were combined and filtered through a 45 μm polyvinylidene difluoride (PVDF) filter and stored at -80 °C until analysis. Amino acids were derivatized using 4-fluoro-7-nitro-2,1,3-benzoxadiazole (NBD-F) according to Aoyama et al. (2004) by incubating 5 μl of the extract with 50 mM borate buffer pH 9.5 and 3 mM NDB-F reagent at 60 °C for 10 min. Reactions were terminated by addition of 333 mM tartrate buffer pH 2.0 and derivatized amino acids were separated using an Alliance ® HPLC System (2695 Separations Module, Waters, Milford, MA, USA) and fluorometrically detected at 540 nm with excitation at 470 nm (2475 Multi-Wavelength Fluorescence Detector, Waters).

Feeding system for gas exchange and chlorophyll fluorescence
Twenty-four hours before the gas exchange measurements, plants were transferred to nutrient solution depleted in N source (neither NH 4 + nor NO 3 -) for starvation. Subsequently, plants were individually transferred 12 h before measurements to stainless steel cuvettes sealed with Teflon caps. The root cuvettes were fed with a continuous flow of nutrient solution supplemented with either NO 3 or NH 4 + using a custom-built multiplant feeding system. Nutrient solution was equally distributed between each of the six cuvettes using electronically controlled solenoid valves (ASCO RedHat II 8262, Florham Park, NJ, USA). Each individual stainless steel cuvette was housed within a sealed PVC tube containing temperature-controlled circulating water to maintain the root system at 25 °C.

Gas exchange and chlorophyll fluorescence
Gas exchange and chlorophyll fluorescence measurements were made on fully expanded leaves using a LI6400 (LICOR Biosciences, Lincoln, NE, USA) leaf chamber (LI6400-40). Gas exchange measurements were made at a pO 2 of 18.6 kPa, a leaf temperature of 25 °C, a saturating light intensity of 1000 μmol quanta m −2 s −1 PAR (photosynthetically active radiation), and a CO 2 partial pressure of 37.2 Pa. Light-response curves were made by decreasing light from 2000 to 1500, 1200, 1000, 800, 500, 200, 100, 40, and 20 μmol quanta m −2 s −1 PAR. The O 2 response curves were made by modulating pO 2 inside the chamber using two mass flow controllers (Aalborg, Orangeburg, NY, USA) to mix N and O 2 gas proportional to 46.6, 32.6, 18.6, 9.3, and 1.9 kPa pO 2 . The order of pO 2 during the measurements was randomized. Simultaneously, chlorophyll fluorescence measurements were made using a LI6400-40 pulse-modulated uptake and assimilation, and free NO 3 content in wild-type, aox1a and ucp1 shoots of A. thaliana fed with 15 NO 3 for 6 h. Plants were exposed to either growth (A, 160 μmol quanta m -2 s -1 ) or saturating (B, 1000 μmol quanta m -2 s -1 ) light conditions during the feeding. Shown are the means ±SE of measurements made on five plants. An asterisk denotes a significant difference (P<0.05) between genotypes.
fluorometer and multiphase flash protocol (Loriaux et al., 2013). The quantum yield of photosystem II (ϕ PSII ) and photochemical quenching (qP) were determined as (F m '-F s )/F m ') and (F m '-F s )/(F m '-F o '), respectively. The rate of linear electron transport through PSII (J f ) was calculated from chlorophyll fluorescence measurements as J f =ϕ PSII ×Abs×I×0.48, where Abs is leaf absorption (=0.85), I is the incident irradiance, and assuming a relative excitation distribution to PSII of 0.48 (Laisk and Loreto, 1996). Furthermore, the rate of electron transport required to sustain the photosynthetic carbon reduction and photorespiratory cycles (J g ) was calculated from gas exchange measurements as J g =(A net +R d ) (4C c +8Г*)/(C c -Г*) where A net is net CO 2 assimilation rate, R d is dark-type respiratory rate, C c is the chloroplastic CO 2 partial pressure, and Г* is the CO 2 compensation point in the absence of dark-type respiration.

Leaf characteristics
Measurements of growth, leaf chlorophyll content, and Rubisco activity were made to characterize WT, aox1a, and ucp1 A. thaliana plants grown in hydroponic conditions. Shoot dry weight was 9 and 10% lower in ucp1 compared with the WT and aox1a, respectively (Table 1). However, leaf number, diameter of the rosette, LMA, and root biomass were similar between genotypes. Additionally, leaf chlorophyll content and Rubisco activity were similar, with an average chlorophyll a/b ratio of 1.8 and Rubisco activity of 48 μmol m -2 s -1 , suggesting similar photosynthetic capacity in all three genotypes (Table 1).

Nitrate uptake, accumulation, and assimilation
To estimate the impact of AOX1a and UCP1 function on de novo N uptake, rates of NO 3 uptake and assimilation were measured by feeding hydroponically grown plants 15 N-enriched NO 3 -. The uptake and accumulation of NO 3 were quantified from measurements of bulk leaf 15 N and free 15 NO 3 -, respectively, relative to plants fed natural abundance NO 3 -. The assimilation of NO 3 was estimated from differences between total leaf 15 N minus 15 NO 3 content. Rates of NO 3 uptake and assimilation were measured in WT plants at three time points after initiating feeding (3, 6, and 9 h) to determine the optimum time. Rates of shoot NO 3 uptake, assimilation, and accumulation were similar after 3, 6, and 9 h ( Supplementary Fig. S1 at JXB online). However, rates of root NO 3 uptake and assimilation were higher at 3 h compared with 6 h and 9 h. Therefore, NO 3 uptake, assimilation, and content were measured in WT, aox1a, and ucp1 plants after 6 h of 15 NO 3 feeding. Under 160 μmol quanta m -2 s -1 (growth irradiance), shoot NO 3 uptake was higher in ucp1 compared with aox1a and WT plants. However, shoot assimilation of NO 3 was higher in both aox1a and ucp1 compared with the WT, while shoot NO 3 content was similar between all three genotypes (Fig. 1). Regardless of the genotype, the uptake, assimilation, and content of NO 3 in the roots were similar ( Supplementary  Fig. S2 at JXB online). After 6 h under high light (1000 μmol quanta m -2 s -1 ), shoot NO 3 uptake and assimilation were higher in ucp1 compared with the WT and aox1a; however, NO 3 content was lower (Fig. 1). Under the high light treatment, the root NO 3 uptake, assimilation, and content were similar between all three genotypes ( Supplementary Fig. S2).

Amino acid analysis
Nineteen amino acids were quantified in leaves of A. thaliana WT, aox1a, and ucp1 fed either NO 3 or NH 4 + as sole N source under both growth (160 μmol quanta m -2 s -1 ) and high (1000 μmol quanta m -2 s -1 ) irradiance. Asparagine, aspartate, and lysine contents were higher in the aox1a mutant plants compared with the WT, regardless of N source or irradiance ( Fig. 2; Supplementary Table S1 at JXB online). However, only asparagine was significantly higher in ucp1 compared with the WT plants. In contrast, under growth irradiance, cysteine, glycine, and serine were lower in ucp1 compared with the WT, regardless of the N source. However, under saturating irradiance, glycine and serine were significantly decreased but cysteine was not in ucp1 compared with WT plants.

Response of photosynthesis
Measurements of A net , ϕ PSII , and the excitation pressure on the chloroplast electron transport chain (1-qP) were measured in WT, aox1a, and ucp1 plants fed either NO 3 or NH 4 + to test the impact of two major electron sinks on photosynthesis. In WT plants, A net and ϕ PSII increased while 1-qP decreased under NO 3 compared with NH 4 + feeding (Fig. 3). However, in the aox1a and ucp1 plants the change in N source had no effect on A net , ϕ PSII , or 1-qP. Under NO 3 and saturating irradiance, A net was lower in ucp1 compared with WT and aox1a plants; however, under non-saturating irradiances, A net was similar between genotypes. Furthermore, under NO 3 -, the measured ϕ PSII was lower and 1-qP higher in the ucp1 and aox1a plants compared with the WT plants, regardless of irradiance. Under NH 4 + feeding, A net , ϕ PSII , and 1-qP were not significantly different between ucp1, aox1a, and WT plants across all irradiances. There was a significant oxygen response of A net but not ϕ PSII and 1-qP for all three genotypes regardless of N form (Fig. 4). In WT plants A net and ϕ PSII were higher while 1-qP was lower under NO 3 compared with NH 4 + feeding across all pO 2 ; however, there was no difference in these parameters between N form in the ucp1 and aox1a plants (Fig. 4). Additionally, there was a significant difference in A net , ϕ PSII , and 1-qP between WT and both mutant lines (ucp1 and aox1a) at all pO 2 under NO 3 but not NH 4 + . The rate of linear electron transport through PSII estimated from chlorophyll fluorescence (J f ) was compared with the electron transport demand required to sustain rates of CO 2 assimilation and photorespiration (J g ). The relationship between J f and J g was linear in response to irradiance and did not differ between NO 3 and NH 4 + feeding for all three genotypes (Fig. 5A). Additionally, the relationship between J f and J g under NO 3 feeding was similar between all three genotypes in response to pO 2 ; however, under NH 4 + feeding, there was a significantly higher slope in the ucp1 plants (1.18) compared with aox1a (0.95) and WT (0.97) plants (Fig. 5B).

Discussion
The disruption of mAET had significant impacts on both foliar N and carbon metabolism in A. thaliana. For example, the loss of the uncoupling protein UCP1 or the alternative oxidase AOX1a increased rates of foliar NO 3 assimilation compared with WT plants. Additionally, A net was lower in ucp1 plants compared with the WT and aox1a lines in NO 3 --fed plants. However, under NH 4 + , rates of A net were not significantly different between genotypes. As discussed below, these data demonstrate that UCP1 and AOX1a are important for balancing the energy partitioning between N and carbon metabolism.

Disruption of alternative mitochondrial electron transport enhances foliar nitrate assimilation
The rates of foliar NO 3 assimilation were higher in the aox1a and ucp1 plants compared with the WT under the non-saturating growth light (160 μmol photons m -2 s -1 ) conditions (Fig. 1). Low light conditions would probably decrease the amount of reductant exported from the chloroplast, and the competition for reductant between NO 3 assimilation and mitochondrial electron transport would be high. Under these conditions, the availability of NADH for the cytosolic conversion of NO 3 to NO 2 and the chloroplastic reduction of NO 2 to NH 4 + by ferredoxin would be limiting. Therefore, the increase in NO 3 assimilation in the upc1 and aox1a plants under low light is probably attributed to an increased availability of cytosolic reductant because of decreased consumption by the mitochondria. The increased NO 3 assimilation in the ucp1 and aox1a plants also corresponded to a higher amino acid content (Fig. 2, Supplementary Table S1 at JXB online) and lower total leaf carbon to N ratio as previously observed in tobacco plants deficient in the mitochondrial complex I (CMS mutant) (Dutilleul et al., 2005) and in AOX1a (Watanabe et al., 2008). Alternatively, at saturating light (1000 μmol photons m -2 s -1 ), the export of excess reductant from the chloroplast into the cytosol will be high and will not limit rates for NO 3 assimilation. In fact under high light, there was only a small increase in NO 3 assimilation with the Fig. 2. Changes in the content of 19 different amino acids in aox1a and ucp1 leaves of A. thaliana fed with either NO 3 or NH 4 + as sole N source and exposed to either growth (160 μmol quanta m -2 s -1 ) or saturating (1000 μmol quanta m -2 s -1 ) irradiance for 6 h. Values represent the log2 of the mutant to WT ratio. Values in bold indicate differences statistically significantly different between the WT and mutants. (This figure is available in colour at JXB online.) loss of UCP1 and there was no change in NO 3 assimilation with the absence of AOX1a.
Generally, the effect of lack of UCP1 on N metabolism was more marked than that of the lack of AOX1a. The expression of UCP1 has been shown to be up-regulated in the aox1a mutant (Watanabee et al., 2008(Watanabee et al., , 2010. Given the potentially overlapping function of these two respiratory components, UCP1 may partly compensate for the loss of AOX1a, which would therefore diminish the impact of the genetic manipulation. However, AOX content has been shown to decrease in the ucp1 mutant (Sweetlove et al., 2006). The higher rates of foliar NO 3 assimilation in the ucp1 and aox1a plants also require an increase in carbon skeletons for the de novo synthesis of amino acids. The present results show a significant increase in aspartate and asparagine levels in mutants compared with the WT. The aspartate pathway drives the synthesis of several amino acids that may contribute to generate additional energy under stress conditions (Galili, 2011). This pathway has been suggested to operate in combination with the TCA cycle in inducing the catapleurotic fluxes with energy deprivation conditions. However, anapleurotic fluxes can also supply amino acid synthesis with carbon skeletons. In aox1a and ucp1 mutants, it is likely that asparagine family synthesis was driven by reductant accumulation and carbon skeleton availability leaking out of the TCA cycle. It has been demonstrated that the carbon needed for the synthesis of amino acids comes primarily from the partial operation of the TCA cycle (Chen and Gadal, 1990;Tcherkez et al., 2009;Sweetlove et al., 2010). Additionally, it has been reported that there is an increase in the anaplerotic production of carbon skeletons through an incomplete TCA cycle with increased NO 3 assimilation (Scheible et al., 1997;Stitt, 1999). Therefore, the increase in NO 3 assimilation and overall amino acid content observed in the ucp1 and aox1a lines probably increased the TCA production of carbon skeletons. This would further increase NADH production available via the malate shuttle for de novo NO 3 assimilation. Root NO 3 assimilation was not altered by the loss of UCP1 and AOX1a, suggesting a more important role for AOX1a and UPC1 in photosynthetic tissue. This is supported by 6-fold higher AOX1a transcripts in the shoot compared with the root in A. thaliana; however, the UCP1 expression level is similar between root and shoot tissues (Watanabe et al., 2010). Furthermore, the AOX1d transcript level is enhanced in the root of aox1a mutants, but not in the shoot, suggesting a potential compensatory effect Fig. 3. Net CO 2 assimilation rate (A net ), quantum yield (ϕ PSII ), and excitation pressure (1-qP) in response to light intensity in wild-type, aox1a, and ucp1 leaves of A. thaliana fed either NO 3 or NH 4 + as sole N source. Shown are the means of measurements made on three plants. The inset box within each panel presents the grand mean with the standard error, estimated from the MSE term in the ANOVA. Significant differences (P<0.05) were denoted by an asterisk either in the inset box when the genotype effect was significant or on the graph to indicate a significant genotype×irradiance interaction.
for the loss of AOX1a in the roots (Watanabe et al., 2010). Taken together, AOX1a and potentially UCP1 appears to have a more predominant role in shoot than root tissues and plays an important role in regulating de novo shoot assimilation of NO 3 -, particularly when the reductant availability is limiting.

Alternative mitochondrial electron transport and nitrate assimilation synergistically optimize photosynthesis
The present results demonstrate that both NO 3 assimilation and mAET optimize rates of photosynthetic CO 2 assimilation; however, independently neither is sufficient to influence A net . For example, A net was higher under NO 3 compared with NH 4 + feeding in WT plants but not in the aox1a and ucp1 mutants, despite increased NO 3 assimilation in these plants. Additionally, under NH 4 + feeding, there was not a significant difference in A net between the WT and the two mutant lines (ucp1 and aox1a). This suggests that the increase in A net under NO 3 is dependent on functional AOX1A and UCP1, and that NO 3 assimilation alone is insufficient to alter A net . Furthermore, the loss of AOX1A and UCP1 had no effect on A net or leaf photochemistry under NH 4 + .
In WT plants, A net and ϕ PSII were 12-17% higher under NO 3 --compared with NH 4 + -fed plants, respectively. An increase in A net and ϕ PSII under NO 3 versus NH 4 + feeding has been previously described in barley (Bloom et al., 1989), wheat (Bloom et al., 2002), tomato (Searles and Bloom, 2003), and maize (Cousins and Bloom, 2003). The foliar reduction of NO 3 to NH 4 + requires cytosolic NADH, typically generated from reductant exported from the chloroplast, and ferredoxin within the chloroplast. Therefore, NO 3 assimilation may compete for reductant with Rubisco-mediated assimilation of CO 2 (Bloom et al., 2002). However, under the conditions used here and as previously reported (Bloom et al., 1989(Bloom et al., , 2002Cousins and Bloom, 2003;Searles and Bloom, 2003), the rates of A net were higher in NO 3 --versus NH 4 + fed plants. This suggests that the consumption of reductant via NO 3 assimilation stimulates rates of net CO 2 assimilation and ϕ PSII , probably through optimizing the ATP/NADPH production within chloroplasts. Additionally, the present data indicate that NO 3 assimilation contributes to protect the chloroplast electron transport chain from over-reduction and therefore ensures optimal rates of CO 2 assimilation.
The measured A net was lower in NO 3 --fed ucp1 mutants compared with WT plants; however, there was no difference Fig. 4. Net CO 2 assimilation rate (A net ), quantum yield (ϕ PSII ), and excitation pressure (1-qP) in response to O 2 atmospheric partial pressure in wild-type, aox1a, and ucp1 leaves of A. thaliana fed with either NO 3 or NH 4 + as sole N source. Shown are the means ±SE of measurements made on three plants. The inset box within each panel presents the grand mean with the standard error, estimated from the MSE term in the ANOVA. An asterisk denotes a significant difference (P<0.05) between genotypes.
in A net between ucp1 and WT plants under NH 4 + feeding (Fig. 3). The decrease in A net seen under NO 3 feeding is consistent with these plants grown in soil as reported by Sweetlove et al. (2006). These authors attributed the decrease in A net in the ucp1 compared with the WT to a restricted flux through the photorespiratory pathway and an associated limited regeneration of ribulose-1,5-bisphosphate. A decrease in the glycine to serine conversion could affect methionine synthesis through C1 metabolism. Both serine and methionine are precursors of cysteine synthesis, which is at a low level in ucp1 compared with the WT. These data are similar to those of Sweetlove et al. (2006), who also showed a lower amount of glycine and serine in the ucp1 mutant compared with the WT (Fig. 2). However, it was found that the difference in A net between the ucp1 mutant and the WT was constant in response to O 2 availability (from 1.9 kPa to 46.6 kPa O 2 ). This suggests that UCP1 optimizes A net regardless of rates of photorespiration and probably plays an important role in balancing the energy supply with demand between N and carbon metabolism within the leaf. The N feeding experiments demonstrated that NO 3 assimilation could compensate for the lack of excess reductant consumption by the mitochondria in the absence of UCP1. However, A net was not enhanced in aox1a and ucp1 mutants fed with NO 3 as seen in WT plants, suggesting that the stimulation of CO 2 fixation by NO 3 assimilation is dependent on a fully functional mAET.
In the ucp1 mutants fed with NH 4 + , the electron production by the chloroplastic electron transport chain increased significantly with decreasing oxygen compared with the electron demand for A net and photorespiration (Fig. 5B). This shift between electron production and demand indicates extra electron transport to alternative chloroplastic sinks such as the Mehler reaction. In addition to the Mehler reaction, other alternative electron sinks such as the cyclic electron flux and chlororespiration have been reported to optimize ATP synthesis, balance the production of the chloroplastic ratio of ATP/ NADPH, and avoid over-reduction of the photosynthetic electron transport chain (Johnson, 2005). In ucp1 mutants fed NH 4 + under non-photorespiratory conditions, the change in J f /J g suggests that the Mehler reaction has a greater influence on linear electron flow compared with other genotypes and treatments. This would avoid excess NADPH accumulation within the stroma due to the loss of two other major electron sinks (mitochondria alternative pathways and NO 3 assimilation) while still maintaining rates of linear electron transport.

Conclusion
In summary the importance of mAET in foliar de novo NO 3 assimilation in A. thaliana, particularly under conditions that limit reductant availability (low light), is demonstrated Fig. 5. Correlation between the electron transport rate through PSII (J f ) and the electron transport rate required for CO 2 assimilation and photorespiration (J g ) in wild-type, aox1a, and ucp1 leaves of A. thaliana fed with either NO 3 or NH 4 + as sole N source. Correlation was established in response to light intensity (A; from 2000 to 20 μmol quanta m -2 s -1 ) and O 2 partial pressure (B; from 46.6 to 1.9 kPa pO 2 ).
in the present work. Additionally, the data show that both mAET and NO 3 assimilation influence rates of photosynthetic CO 2 assimilation and electron transport. mAET and NO 3 assimilation appear to function synergistically to avoid excess reductant accumulation and over-reduction of the chloroplast. Finally, the data demonstrate that mitochondrial respiration significantly contributes to the energy balancing between N and carbon metabolism.

Supplementary data
Supplementary data are available at JXB online. Figure S1. Rates of NO 3 uptake and assimilation, and free NO 3 content in wild-type shoots and roots of A. thaliana fed with 15 NO 3 for 3, 6, or 9 h. Table S1. Amino acid levels (nmol mg -1 DW) in shoots of wild-type, aox1a, and ucp1 Arabidopsis thaliana fed either NO 3 or NH 4 + as sole N source. Figure S2. Rates of NO 3 uptake and assimilation, and free NO 3 content in wild-type, aox1a, and ucp1 roots of A. thaliana fed with 15 NO 3 for 6 h.