Alternative Oxidase Alleviates Mitochondrial Oxidative Stress during Limited Nitrate Reduction in Arabidopsis thaliana

The conversion of nitrate to ammonium, i.e., nitrate reduction, is a major consumer of reductants in plants. Previous studies have reported that the mitochondrial alternative oxidase (AOX) is upregulated under limited nitrate reduction conditions, including no/low nitrate or when ammonium is the sole nitrogen (N) source. Electron transfer from ubiquinone to AOX bypasses the proton-pumping complexes III and IV, thereby consuming reductants efficiently. Thus, upregulated AOX under limited nitrate reduction may dissipate excessive reductants and thereby attenuate oxidative stress. Nevertheless, so far there is no firm evidence for this hypothesis due to the lack of experimental systems to analyze the direct relationship between nitrate reduction and AOX. We therefore developed a novel culturing system for A. thaliana that manipulates shoot activities of nitrate reduction and AOX separately without causing N starvation, ammonium toxicity, or lack of nitrate signal. Using shoots processed with this system, we examined genome-wide gene expression and growth to better understand the relationship between AOX and nitrate reduction. The results showed that, only when nitrate reduction was limited, AOX deficiency significantly upregulated genes involved in mitochondrial oxidative stress, reductant shuttles, and non-phosphorylating bypasses of the respiratory chain, and inhibited growth. Thus, we conclude that AOX alleviates mitochondrial oxidative stress and sustains plant growth under limited nitrate reduction.


Introduction
Plants use nitrate and ammonium as main sources of nitrogen (N).The conversion of nitrate to ammonium, i.e., nitrate reduction, requires eight electrons per nitrate, which accounts for about half of the energy required for protein synthesis from nitrate [1].Thus, ammonium is an energetically superior N source, but most crops prefer nitrate [2].In herbaceous plants, nitrate reduction generally occurs in the shoots [3][4][5], and acts as an electron sink for reductants such as NAD(P)H and reduced ferredoxin that are generated in the cytosol and chloroplasts [6].Hence, when the nitrate supply is limited, excessive reductants may accumulate within illuminated photosynthetic cells.Recent studies have shown that the NAD(P)H/NAD(P) + ratio and the levels of hydrogen peroxide are increased in A. thaliana leaves of ammonium-grown plants compared with nitrate-grown plants [6][7][8].
Moreover, nitrate deficiency treatments have been found to increase the levels of superoxide, hydrogen peroxide, and malondialdehyde in tobacco leaves [9].These findings suggest that a decrease in nitrate reduction could lead to oxidative stress in plants.
It is widely accepted that nitrate reduction significantly alters the action of the plant mitochondrial electron transport chain (mETC) [6].In fact, the number of electrons consumed in nitrate reduction appears to be comparable to that in mETC [6,10].In photosynthetic tissues, the expression and activities of several enzymes that bypass energy conservation steps in mETC are induced under limited nitrate reduction conditions, i.e., if the plant is experiencing N starvation or if ammonium is the sole N source [11][12][13][14][15].One enzyme is the alternative oxidase (AOX), which allows direct electron transfer from ubiquinone to molecular oxygen, bypassing the proton-pumping complexes III and IV [16].Therefore, upregulation of AOX under limited nitrate reduction may dissipate excessive reductants without being limited by steep proton gradients across the mitochondrial inner membrane, thereby attenuating reactive oxygen/nitrogen species (ROS/RNS) production and oxidative stress [17,18].Indeed, in Arabidopsis grown under low nitrate conditions, the shoot expression of antioxidant enzyme genes was induced by disrupting the major isoform AOX1a [14].Antisense suppression of Arabidopsis AOX1a was found to dramatically decrease the reducing state of ascorbate as an antioxidant under ammonium but not under nitrate [8].Moreover, in tobacco cell cultures subjected to N starvation, antisense suppression of AOX1 caused carbohydrate accumulation [19].This suggests that, under limited nitrate reduction, NADH oxidation via AOX instead of nitrate reductase (NR) could replenish NAD + to drive the glycolysis.Meanwhile, Arabidopsis AOX1a deficiency enhanced foliar nitrate assimilation under nitrate-replete conditions, implying competition between AOX and nitrate reduction for reductants [20].
The above studies suggest that AOX is tightly linked to nitrate reduction.However, these studies manipulated nitrate reduction activity by transferring plants to growth conditions that included no/low nitrate or the use of ammonium as the sole N source.Reduced nitrate reduction may therefore be accompanied by N starvation or ammonium toxicity [21], causing plant growth suppression and/or the initiation of stress responses.Moreover, since nitrate acts as a signal to alter genome-wide gene expression [22], a decrease in the nitrate supply also depletes the nitrate signal.For these reasons, it is impossible to distinguish whether the above-mentioned effects of AOX deficiency are caused by reduced nitrate reduction, N starvation, ammonium toxicity, or the lack of nitrate signal.To solve this problem, we developed a novel culturing system to manipulate the degree of nitrate reduction and AOX activity without causing N starvation, ammonium toxicity, or lack of nitrate signal.Using this system, we examined genome-wide gene expression and plant growth to better understand the relationship between AOX and nitrate reduction.Our results clearly support that AOX alleviates mitochondrial oxidative stress and sustains plant growth under limited nitrate reduction.

Determination of In Vitro Nitrate Reductase Activities
We assessed in vitro NR activity as per a previously published method [25] with slight modification.Frozen samples were ground in a multi-bead shocker (Yasui Kikai, Osaka, Japan) using zirconia beads.The resulting powder was then mixed with 10 vol. of extraction buffer (50 mM HEPES-KOH, pH 7.6, 1 mM EDTA, 7 mM cysteine) and the extracts were centrifuged at 20,400× g at 4 • C for 10 min.Next, a 30 µL aliquot of the supernatant was added to 90 µL of assay buffer (50 mM HEPES-KOH, pH 7.6, 133 µM NADH, 2 mM EDTA, 6.67 mM KNO 3 ).After incubation at 30 • C for 2 and 17 min, 24 µL of 1% (w/v) sulfanilamide solution in 1 N HCl and 24 µL of 0.02% (w/v) N-(1-naphthyl)ethylenediamine dihydrochloride solution were added to 48 µL aliquots of reaction mixture.Finally, nitrite content was determined based on absorbance readings at 540 nm.In vitro NR activity was calculated by determining the nitrite amount produced over a 15 min period.

Determination of In Vivo Nitrate Reductase Activities
We submerged fresh shoots in 50 vols.of reaction buffer (i.e., 100 mM sodium phosphate buffer, pH 7.4, 10 mM KNO 3 , 4% (v/v) n-propanol), followed by vacuum infiltration.The reaction mixture was then incubated at 30 • C for 1 h in the dark.The amount of nitrite produced was then visualized by mixing the supernatant with 1% (w/v) sulfanilamide solution in 1 N HCl, and 0.02% (w/v) N-(1-naphthyl)ethylenediamine dihydrochloride solution in a 2:1:1 ratio.

Determination of Nitrate Concentration
Nitrate content was quantified via a previously published protocol [26] with slight modifications.Nitrate was extracted with 10 vol. of deionized water at 100 • C for 20 min.Next, 10 µL of the supernatant was mixed with 40 µL of 5% (w/v) salicylic acid in concentrated sulfuric acid, and the resulting mixture was then incubated at room temperature for 20 min.A mock treatment of 40 µL of concentrated sulfuric acid was also produced.Finally, 1 mL of 8% (w/v) NaOH solution was added to the mixture, and nitrate was determined based on absorbance at 410 nm.

Determination of Total Protein
Total protein was determined as previously described [25].Frozen samples were homogenized with a multi-bead shocker (Yasui Kikai) using zirconia beads.Total proteins were then extracted with 10 vol. of sample buffer [2% (w/v) SDS, 62.5 mM Tris-HCl (pH 6.8), 10% (v/v) glycerol, and 0.0125% (w/v) bromophenol blue] and HaltTM protease inhibitor cocktail (ThermoFisher Scientific, Tokyo, Japan), followed by incubation at 95 • C for 5 min.Extracts were then centrifuged at 20,400× g at 8 • C for 10 min and 10 µL aliquots were suspended in 500 µL of deionized water.Next, 100 µL of 0.15% (w/v) sodium deoxycholate was added, and the mixture was incubated at room temperature for 10 min.Then, 100 µL of 72% (v/v) trichloroacetic acid was added, followed by incubation at room temperature for 15 min and centrifugation at 20,400× g for 10 min.Precipitates were then air-dried and suspended in 25 µL of deionized water.This suspension was used to determine total protein concentrations using Takara BCA Protein Assay Kits (TaKaRa, Kusatsu, Japan).

RNA Extraction
Frozen samples were ground in a multi-bead shocker (Yasui Kikai) using zirconia beads.Next, total RNA was extracted using an RNeasy Plant Mini Kit (Qiagen, Tokyo, Japan) according to the manufacturer's instruction.For RNA-seq library preparation, RNA was purified using on-column DNase digestion (Qiagen).

RNA-Seq
RNA quality was evaluated using a Qubit RNA IQ assay kit (ThermoFisher Scientific).RNA samples with RNA IQs 8.7-10.0 were used for library preparation.cDNA libraries were constructed using a NEBNext Ultra II RNA Library Prep Kit with sample purification beads (New England Biolabs, Tokyo, Japan), a NEBNext Poly(A) mRNA Magnetic Isolation Module (New England Biolabs), and a NEBNext Multiplex Oligos for Illumina (New England Biolabs).cDNA libraries were then sequenced using a NextSeq 500 (Illumina, Tokyo, Japan), and resulting bcl files were converted to fastq files using bcl2fastq (Illumina).The RNA-seq raw data are available in the ArrayExpress database under accession number E-MTAB-14027.The reads were analyzed according to the method described by Notaguchi et al. [27] and mapped to the Arabidopsis reference (TAIR10) using Bowtie [28] with the following options: "-all-best-strata".Finally, obtained reads were analyzed using iDEP version 0.96 [29], Metascape [30], and GeneCloud [31] using the default settings.

RT-qPCR
Reverse transcription (RT) was performed using a ReverTraAce qPCR RT Master Mix with gDNA Remover (Toyobo, Osaka Japan).Synthesized cDNA was then diluted tenfold with water and used for quantitative PCR (qPCR).RT-qPCR was performed by a QuantStudio 1 (ThermoFisher Scientific) with KOD SYBR qPCR Mix (Toyobo).Relative transcript levels were calculated using the comparative cycle threshold method with ACTIN3 as an internal standard [21].Primer sequences are shown in Table S1.

Statistical Analysis
The Tukey-Kramer multiple comparison test was conducted using R software v.2.15.3.

Manipulation of Nitrate Reduction and AOX Activities
We manipulated the activities of nitrate reduction and AOX via the transfer experiment (Figure 1A) using Col-0 and mutants deficient in either or both NR and AOX activities.First, plants were grown for 18 days with ammonium as the sole N source (Condition 1 in Figure 1A) to ensure uniform growth independent of NR activity.Higher medium pH and lower light intensity were used to reduce ammonium toxicity and AOX expression [11,13,21,32], allowing uniform growth regardless of AOX activity.Second, plants were subjected to N starvation at pH 5.7 for 24 h (Condition 2) to induce AOX expression [11,13].Finally, plants were transferred to a medium containing adequate nitrate and incubated under moderate light intensity (Condition 3), which induced NR expression and filled photosynthetically-derived reductants to the cell [16,22,33].The NR activity of Col-0 increased rapidly from 0 to 7 h after nitrate supply and slightly more from 7 to 24 h (Figure S1A).Both in vitro and in vivo NR activities were increased 7 h after nitrate supply in the shoots of Col-0 and aox1a-1 but not in nr and aox1a-1 nr (Figure 1B,C).Thus, we focused on the 7 h after nitrate supply to identify the early effects of nitrate reduction for subsequent experiments.We found that nitrate accumulated in the shoots of all lines after nitrate supply, with concentrations ranging from 12.1 to 22.4 µmol g −1 (Figure 1D), indicating an adequate supply of nitrate for signaling.Indeed, the expression of NIA2 encoding the major NR isoform in leaves [34], which is inducible by nitrate signal [22,35], was strongly induced 7 h after nitrate supply in Col-0 and aox1a-1 but not in nr and aox1a-1 nr (Figure S1B).Higher nitrate concentrations in shoots of nr and aox1a-1 nr than Col-0 and aox1a-1 would reflect a deficiency in nitrate reduction (Figure 1D).Shoot protein concentrations were comparable among all lines before and after nitrate supply, ranging from 27.0 to 29.7 mg g −1 (Figure 1E), suggesting that no N starvation occurred.RT-qPCR and Western blot analyses confirmed that the signals corresponding to AOX1a and AOX were negligible in aox1a-1 and aox1a-1 nr (Figure S1C,D), suggesting that the knockout of AOX1a was sufficient to diminish AOX, as reported earlier [13,14,23].Taken together, these culturing conditions permit comparisons of plants' differing nitrate reduction and AOX activity without causing ammonium toxicity, N starvation, and lack of nitrate signaling.

Manipulation of Nitrate Reduction and AOX Activities
We manipulated the activities of nitrate reduction and AOX via the transfer experiment (Figure 1A) using Col-0 and mutants deficient in either or both NR and AOX activities.First, plants were grown for 18 days with ammonium as the sole N source (Condition 1 in Figure 1A) to ensure uniform growth independent of NR activity.Higher medium pH and lower light intensity were used to reduce ammonium toxicity and AOX expression [11,13,21,32], allowing uniform growth regardless of AOX activity.Second, plants were subjected to N starvation at pH 5.7 for 24 h (Condition 2) to induce AOX expression [11,13].Finally, plants were transferred to a medium containing adequate nitrate and incubated under moderate light intensity (Condition 3), which induced NR expression and filled photosynthetically-derived reductants to the cell [16,22,33].The NR activity of Col-0 increased rapidly from 0 to 7 h after nitrate supply and slightly more from 7 to 24 h (Figure S1A).Both in vitro and in vivo NR activities were increased 7 h after nitrate supply in the shoots of Col-0 and aox1a-1 but not in nr and aox1a-1 nr (Figure 1B,C).Thus, we focused on the 7 h after nitrate supply to identify the early effects of nitrate reduction for subsequent experiments.We found that nitrate accumulated in the shoots of all lines after nitrate supply, with concentrations ranging from 12.1 to 22.4 µmol g −1 (Figure 1D), indicating an adequate supply of nitrate for signaling.Indeed, the expression of NIA2 encoding the major NR isoform in leaves [34], which is inducible by nitrate signal [22,35], was strongly induced 7 h after nitrate supply in Col-0 and aox1a-1 but not in nr and aox1a-1 nr (Figure S1B).Higher nitrate concentrations in shoots of nr and aox1a-1 nr than Col-0 and aox1a-1 would reflect a deficiency in nitrate reduction (Figure 1D).Shoot protein concentrations were comparable among all lines before and after nitrate supply, ranging from 27.0 to 29.7 mg g −1 (Figure 1E), suggesting that no N starvation occurred.RT-qPCR and Western blot analyses confirmed that the signals corresponding to AOX1a and AOX were negligible in aox1a-1 and aox1a-1 nr (Figure S1C,D), suggesting that the knockout of AOX1a was sufficient to diminish AOX, as reported earlier [13,14,23].Taken together, these culturing conditions permit comparisons of plants' differing nitrate reduction and AOX activity without causing ammonium toxicity, N starvation, and lack of nitrate signaling.B-E) Two plants of each line (eight in total) per plate were grown, and two shoots were pooled as one biological replicate.Data: mean ± SD (n = 3 (B), n = 5 (0 h) and 9 (7 h) (D), n = 4 (E)).For (C), color intensity is proportional to in vivo NR activity.Different lowercase letters indicate significant differences determined via Tukey-Kramer tests at p < 0.05.ND and NS denote "not detected" and "not significant".FW denotes "fresh weight".

AOX1a Deficiency Induces Genes Related to Mitochondrial Oxidative Stress under Limited Nitrate Reduction
Next, to dissect the role of AOX in limited nitrate reduction, we performed three independent RNA-seq analyses using shoots from Col-0, aox1a-1, aox1a-2, nr, aox1a-1 nr, and aox1a-2 nr 7 h after nitrate supply (Table S2).A k-means clustering analysis classified the transcripts into four groups according to their expression patterns (Figure 2A, Table S3).
In cluster D, AOX1a deficiency consistently induced gene expression in the nr background but not in the Col-0 background (Figure 2A,B).Moreover, 42 of 141 genes in cluster D were significantly upregulated in aox1a-1 nr and aox1a-2 nr relative to nr (Figures 2C and S2, Tables S4 and S5).Further enrichment analyses of the 42 genes identified significant overrepresentation of the terms "toxin catabolic process", "glutathione metabolism" (Figure 2D), "interpro-ipr004046 (glutathione S-transferase, C-terminal)", and "gst" (Figure 2E).These terms were derived from the glutathione S-transferase (GST) genes AT2G29460 (GSTU4), AT1G17170 (GSTU24), AT1G17180 (GSTU25), and AT1G02920 (GSTF7) (Table S5).GSTU4, a plant-specific tau class GST, may contribute to hydrogen peroxide degradation by using glutathione as an electron donor [36].The terms "response to hypoxia" and "response to oxidative stress" were also enriched in these 42 genes (Figure 2D).The genes induced by hypoxia [37] and H 2 O 2 treatments [38] were significantly upregulated in nr, and their induction was enhanced by AOX1a deficiency (Figure 2F,G and Tables S6 and S7).The NAC transcription factor ANAC017 mediates ROS-related mitochondrial retrograde signaling, thereby activating mitochondrial dysfunction stimulon genes including AOX1a, UPOX, and ANAC013 [39,40].The expression of ANAC017-inducible genes [39] and mitochondrial dysfunction stimulon genes [40] reached a maximum level in aox1a-1 nr and aox1a-2 nr (Figure 2H,I and Tables S8 and S9).Further RT-qPCR analyses revealed that, after nitrate supply, the expression of hypoxia-inducible genes (Figure 2J-M) and oxidative stress marker genes (Figure 2N-Q) was induced in nr, which was enhanced by AOX1a deficiency.Meanwhile, before nitrate supply, little difference was observed among all lines (Figure 2J-Q).It should be noted that hypoxia-inducible genes were induced in aox1a-1 nr and aox1a-2 nr (Figure 2F,J-M) despite no hypoxia treatment.Since hypoxia generally elevates cytosolic NADH/NAD + ratio [41], the hypoxia-inducible genes may be upregulated by reductant accumulation.Indeed, PDC1, whose protein catalyzes the rate-limiting step of NADH-oxidizing alcohol fermentation in the cytosol [42,43], was dramatically induced in the shoots lacking both NR and AOX (Figure 2L), suggesting an accumulation of excessive reductants.Together, our transcriptome analysis suggests that AOX dissipates excessive reductants and mitigates oxidative stress under limited nitrate reduction.

AOX1a Deficiency Induces Genes for Respiratory Bypasses and Reductant Shuttles under Limited Nitrate Reduction
The plant mETC possesses type II NAD(P)H dehydrogenases located on the cytosol side (ND ex ) or matrix side (ND in ) of the inner mitochondrial membrane [6].Since these dehydrogenases transfer electrons to ubiquinone and bypass the proton-pumping complex I, they can support the dissipation of excessive reductants.Of these, the expression of Arabidopsis NDB2, encoding the primary ND ex contributor to NADH oxidation [44], was significantly induced in nr after nitrate supply, which was intensified by AOX1a deficiency (Figure 3A,B).Since cytosolic NR has a much lower K m for NADH than ND ex [11,44], NDB2 may operate only when nitrate reduction is limited.NDA2, encoding the NADHoxidizing ND in [6], also showed an expression pattern similar to NDB2 after nitrate supply (Figure 3A,C).The compensated induction of NDB2 and NDA2 by AOX1a deficiency suggests a contribution of AOX1a-NDB2 and AOX1a-NDA2 modules to balance cellular redox under limited nitrate reduction.The co-function of AOX1a, NDB2, and NDA2 has already been suggested in plants subjected to various environmental stresses [44][45][46][47].Also, the STRING database ver.12.0 integrating protein-protein interactions [48] confirmed tight functional connections between AOX1a, NDB2, and NDA2 (Figure S3).Meanwhile, the expression of the uncoupling proteins genes (UCPs), whose proteins act as an uncoupler and/or as an aspartate/glutamate exchanger across the inner mitochondrial membrane [49], was little changed among the lines (Figure 3A).expression of Arabidopsis NDB2, encoding the primary NDex contributor to NADH oxidation [44], was significantly induced in nr after nitrate supply, which was intensified by AOX1a deficiency (Figure 3A,B).Since cytosolic NR has a much lower Km for NADH than NDex [11,44], NDB2 may operate only when nitrate reduction is limited.NDA2, encoding the NADH-oxidizing NDin [6], also showed an expression pattern similar to NDB2 after nitrate supply (Figure 3A,C).The compensated induction of NDB2 and NDA2 by AOX1a deficiency suggests a contribution of AOX1a-NDB2 and AOX1a-NDA2 modules to balance cellular redox under limited nitrate reduction.The co-function of AOX1a, NDB2, and NDA2 has already been suggested in plants subjected to various environmental stresses [44][45][46][47].Also, the STRING database ver.12.0 integrating proteinprotein interactions [48] confirmed tight functional connections between AOX1a, NDB2, and NDA2 (Figure S3).Meanwhile, the expression of the uncoupling proteins genes (UCPs), whose proteins act as an uncoupler and/or as an aspartate/glutamate exchanger across the inner mitochondrial membrane [49], was little changed among the lines (Figure 3A).The intracellular redox balance is tuned through the reductant shuttle systems across different cellular compartments [50][51][52][53][54]; electrons from membrane-impermeable NAD(P)H are temporarily stored in membrane-permeable compounds (e.g., malate, proline) through biochemical interconversions.The compounds are then transported across the membrane, followed by reconstitution of NAD(P)H.To reveal whether the redox perturbation due to deficiency of NR and/or AOX stimulates the shuttle systems, we surveyed transcriptional changes in the relevant genes (Figure 3A).DiT1/OMT1, which encodes the 2-oxogutarate/malate and oxaloacetate/malate translocator on the chloroplast inner membranes [51], was upregulated almost equally in nr, aox1a-1 nr, and aox1a-2 nr after nitrate supply (Figure 3A,D).In A. thaliana, the conversion of nitrite to ammonium by nitrite reductase occurs only in chloroplasts and requires six electrons (as six reduced ferredoxins) per nitrite.Hence, under limited nitrate reduction, DiT1/OMT1 as the malate valve [51][52][53] may export excessive reductants from chloroplasts, avoiding photo-oxidative stress.Meanwhile, DIC3, whose protein can transport malate from the mitochondria to the cytosol [52], was induced in nr after nitrate supply, and this induction was intensified by AOX1a deficiency (Figure 3A,E).The conversion of malate to oxaloacetate by cytosolic malate dehydrogenases is accompanied by the production of NADH [52,53].A similar trend to DIC3 was observed with PRODH1 and PRODH2 (Figure 3A,F,G), which encode proline dehydrogenases localized in mitochondria.Proline is produced in the cytosol by accepting electrons from NAD(P)H and can then be transported into the mitochondria [54].PRODH1/2 catalyze the direct electron transfer from proline to ubiquinone via FAD as a cofactor in the mitochondria, bypassing complex I. Thus, in shoots lacking NR and AOX, the upregulation of DIC3 and PRODH1/2 would lead to dissipation of excessive reductants, implying over-reduction of mETC.These also support the hypothesis that AOX consumes excessive reductants under limited nitrate reduction.

AOX1a Deficiency Inhibits Shoot Growth under Limited Nitrate Reduction
Finally, we analyzed shoot growth parameters following nitrate supply.Over 6 days, we found that AOX1a deficiency significantly reduced (−15%) shoot fresh weight in the nr background but not in the Col-0 background (Figure 4A).Meanwhile, rosette diameter was little affected by AOX1a deficiency (Figure 4B).When plants were grown in nutrient-rich soil, shoot appearance and shoot fresh weight were decreased in aox1a-1 nr and aox1a-2 nr relative to the others (Figure 4C-E).Together, these suggest that AOX is crucial for sustaining plant growth under limited nitrate reduction.

Conclusions
We have successfully developed a cultural system to manipulate activities of nitrate reduction and AOX without causing ammonium toxicity, N starvation, or lack of nitrate signaling.Analyses using this system suggest that AOX alleviates mitochondrial oxidative stress and sustains plant growth under limited nitrate reduction.Our transcriptional dissection indicates redox/metabolic perturbation in plants lacking either or both NR and AOX.Further analysis of physiological and biochemical details using this system is awaited to better understand the relationship between nitrogen metabolism and respiration.

Supplementary Materials:
The following supporting information can be downloaded at: www.mdpi.com/xxx/s1, Figure S1: Manipulation of activities of nitrate reduction and AOX. Figure S2: Normalized expression of 42 overlapped genes.Figure S3: Outputs from STRING database ver.12.0 using AOX1a (AT3G22370) as query.Table S1: Primer sequences.Table S2: Read counts data

Conclusions
We have successfully developed a cultural system to manipulate activities of nitrate reduction and AOX without causing ammonium toxicity, N starvation, or lack of nitrate signaling.Analyses using this system suggest that AOX alleviates mitochondrial oxidative stress and sustains plant growth under limited nitrate reduction.Our transcriptional dissection indicates redox/metabolic perturbation in plants lacking either or both NR and AOX.Further analysis of physiological and biochemical details using this system is awaited to better understand the relationship between nitrogen metabolism and respiration.

Figure 1 .
Figure 1.Manipulation of activities of nitrate reduction and AOX without causing N starvation, ammonium toxicity, or lack of nitrate signal.(A) Schematic diagram of experiment.(B) In vitro NR activity, (C) in vivo NR activity, (D) nitrate concentration, and (E) total protein concentration in plant shoots before and 7 h after nitrate supply.(B-E) Two plants of each line (eight in total) per

Figure 1 .
Figure 1.Manipulation of activities of nitrate reduction and AOX without causing N starvation, ammonium toxicity, or lack of nitrate signal.(A) Schematic diagram of experiment.(B) In vitro NR activity, (C) in vivo NR activity, (D) nitrate concentration, and (E) total protein concentration in plant shoots before and 7 h after nitrate supply.(B-E)Two plants of each line (eight in total) per plate were grown, and two shoots were pooled as one biological replicate.Data: mean ± SD (n = 3 (B), n = 5 (0 h) and 9 (7 h) (D), n = 4 (E)).For (C), color intensity is proportional to in vivo NR activity.Different lowercase letters indicate significant differences determined via Tukey-Kramer tests at p < 0.05.ND and NS denote "not detected" and "not significant".FW denotes "fresh weight".

Figure 2 .
Figure 2. Transcriptomic alteration by AOX1a deficiency under limited nitrate reduction conditions.(A-I) Shoots from plants 7 h after nitrate supply were subjected to RNA-seq.One plant from each line (six in total) per plate was grown.Six shoots were pooled as one biological replicate.(A) Heat map from a k-means clustering of normalized transcript levels.Magenta and green represent higher and lower expression levels, respectively.Numbers (1st, 2nd, 3rd) indicate the order of independent RNA-seq experiments.(B) Plots of normalized transcript levels in cluster D. Normalized transcript levels of splice variants were averaged for each gene.(C) Venn diagram showing the number of

Figure 2 .
Figure 2. Transcriptomic alteration by AOX1a deficiency under limited nitrate reduction conditions.(A-I) Shoots from plants 7 h after nitrate supply were subjected to RNA-seq.One plant from each line (six in total) per plate was grown.Six shoots were pooled as one biological replicate.(A) Heat map from a k-means clustering of normalized transcript levels.Magenta and green represent higher and lower expression levels, respectively.Numbers (1st, 2nd, 3rd) indicate the order of independent RNA-seq experiments.(B) Plots of normalized transcript levels in cluster D. Normalized transcript

Figure 4 .
Figure 4. Effects of AOX1a deficiency on shoot growth under limited nitrate reduction conditions.(A) Shoot fresh weight and (B) rosette diameter of plants 6 days after nitrate supply.One plant of each line (four in total) per plate were grown and one shoot was regarded as one biological replicate.(C) Shoot appearance of 24-day-old plants and shoot fresh weights of (D) 14-day-old and (E) 24day-old plants grown in pots containing nutrient-rich soil.Data: mean ± SD (n = 17 (A,B), n = 18 (D), n = 7 (E)).

Figure 4 .
Figure 4. Effects of AOX1a deficiency on shoot growth under limited nitrate reduction conditions.(A) Shoot fresh weight and (B) rosette diameter of plants 6 days after nitrate supply.One plant of each line (four in total) per plate were grown and one shoot was regarded as one biological replicate.(C) Shoot appearance of 24-day-old plants and shoot fresh weights of (D) 14-day-old and (E) 24-dayold plants grown in pots containing nutrient-rich soil.Data: mean ± SD (n = 17 (A,B), n = 18 (D), n = 7 (E)).