Mitochondrial uncoupling protein contributes to the regulation of carbon and nitrogen metabolism and seed yield under low nitrogen stress in Arabidopsis

Xinyan Qiao Lanzhou University Tao Yu Lanzhou University Mengjiao Ruan Lanzhou University Chaiyan Cui Lanzhou University Cuiyun Chen Cold and Arid Regions Environmental and Engineering Research Institute: Northwest Institute of EcoEnvironment and Resources Yuanzhi Zhu Lanzhou University Fanglin Li Lanzhou University Shengwang Wang Lanzhou University Xiaofan Na Lanzhou University Xiaomin Wang Lanzhou University Yurong Bi (  yrbi@lzu.edu.cn ) Lanzhou University https://orcid.org/0000-0003-2412-3673


Introduction
The availability of nitrogen (N) in soil is essential for plant growth and defense [1][2][3]. However, plants often encounter limited N sources under nature environmental conditions, which leads to chlorophyll degradation, decreased photosynthesis and protein synthesis [4,5]. Moreover, application of N fertilizers results in contamination of groundwater and production of nitrogenous greenhouse gases [6,7].
Ammonium (NH 4 + ) and Nitrate (NO 3 − ) are two major forms of useable N for plants. NH 4 + can be directly assimilated into amino acids by glutamine synthetase and glutamate synthase [8]. However, NO 3 − is rst reduced to nitrite (NO 2 − ) by nitrate reductase (NR) in cytoplasm using NAD(P)H as the electron donor; then NO 2 − is reduced to NH 4 + by nitrite reductase (NiR). The reduction of NO 2 − and assimilation of NH 4 + occur in plastids/chloroplasts and the reduced ferriredoxin is used as the reductant [8]. Inhibition of the reductant output via the malate-oxaloacetic acid (Mal-OAA) shuttle from chloroplasts decreases the NO 3 − assimilation rate [9,10]. Therefore, plants need to balance the photosynthetic energy partition between carbon (C) and N assimilation at different development stages.
N assimilation is in closely related to C metabolism. The balance of N and C in plants is essential for the optimal growth and development [5]. NO 3 − availability affects transcriptions of a variety of genes involved in C metabolism, while C depletion decreases protein biosynthesis and alters N metabolism [11,12]. It is well accepted that low N inevitably results in low photosynthetic capacity and yield in plants [5]. Photosynthesis and respiration provide not only the reducing power but also C skeletons (2-oxoglutarate, citrate, and isocitrate) for N assimilation [9,13]. Many reports have con rmed that photosynthesis and respiration rates highly correlate to the intracellular N level [5,9,10]. Glycolytic pathway and the tricarboxylic acid cycle could be up-regulated in plant adaptation to low N condition [11,14,15]. Additionally, the capacity of alternative electron transport pathways (aETPs) in the respiratory chain also exhibits a strong correlation with the intracellular NO 3 − level [9,10,16]. aETPs are essential for plant adaptation to various stresses [17][18][19]. Particularly, they play key roles in maintaining a high photosynthetic capacity by dissipating the excess reductant from chloroplasts [10,20]. However, it is not well understood how aETPs coordinate and contribute to optimizing the C metabolism and N assimilation under low-N stress in plants.
Alternative pathway (AP) and uncoupling pathway are two major pathways of aETPs, which are mediated by alternative oxidase (AOX) and the uncoupling protein (UCP), respectively. AOX can directly transport electron from ubiquinone to O 2 , but it did not transfer H + across the mitochondrial inner membrane, thus dissipating energy as heat [21]. Comparatively, UCP increases the permeability of H + across the mitochondrial inner membrane, which destroys the H + gradient and decreases the ATP formation [22]. In Arabidopsis, ve AOX genes (AOX1a, AOX1b, AOX1c, AOX1d, AOX2) and six UCP genes (UCP1-6) have been reported [23]. Recently, it was found that the AOX1a or UCP1 mutation can alter foliar N and C assimilation rates in Arabidopsis fed with NO 3 − or NH 4 + [9]. Moreover, the transcripts of AOX1a and UCP1 are signi cantly up-regulated by low-N stress [10,16]. Excessive carbohydrates are preferentially respired by AOX, which suppresses the increase of the C/N ratio upon low-N condition [15,24]. However, it is indicated that AOX induced by low-N condition does not play important roles on C/N ratio regulation [16].
Sweetlove et al. [20] reported that the UCP1 mutation markedly decreases the AOX protein content, but does not affect other aETPs in ucp1 plants. Moreover, UCP1 is essential for the e cient photosynthesis mainly through maintaining photorespiratory rate in Arabidopsis [20]. However, it is still unknown whether UCP1 participates in regulating C/N ratio upon low-N condition at the whole plant level; and the respective contribution and coordination of UCP1 and AOX in Arabidopsis adaptation to low-N stress remain unclear.
Although above promising results have been reported, the authors rarely analyzed the effects of AOX1a and UCP1 mutation on the C/N ratio and yield of seeds as well as their respective roles in plant adaptation to low-N condition. The aim of this study is to clarify the functions of AOX1a and UCP1 in Arabidopsis tolerance to low-N stress by using the aox1a, ucp1 and aox1a/ucp1 mutants. In particular, we focused on the effects of AOX1a and UCP1 mutation on C and N assimilation processes, photosynthetic capacity, C/N ratio, and seed yield.

Materials And Methods
Plant materials and growth conditions Generation of proAOX1a: GUS and proUCP1: GUS transgenic lines The promoter sequences of AtAOX1a and AtUCP1 were ampli ed by using the primer pairs (proAOX1a-FP and proAOX1a-RP; proUCP1-FP and proUCP1-RP), respectively. Then, they were cloned into the expression vector pBIB-GUS by using the Gateway technology. Agrobacterium-mediated transformation was used to deliver the recombinant plasmids into Col-0 by oral-dipping method. The T 0 seeds were screened for positive lines by using the herbicide Basta.

Low nitrogen treatment
For the low nitrogen experiment at the seedling stage (0-10 d), NH 4 NO 3 and KNO 3 were removed from the 1/2 MS medium, then KNO 3 as the only nitrogen source was added at the concentrations of 0, 0.1 mM, and 1 mM, which were denoted as 0 N, 0.1 N, and 1 N medium, respectively, in this study. The depleted K + was supplied with KCl. The seeds were germinated and grew in 0 N, 0.1 N, or 1 N medium for treatments.

Measurement of photosynthetic uorescence parameters
Photosynthetic uorescence parameters were analyzed using the LI-6400XT photosynthesis system (LI-COR). The leaf temperature was maintained at 22°C during measurements. Photosynthesis was determined under ambient growth condition by using 300 PAR light intensity and 400 ppm CO 2 . The photosynthetic light-response curve was determined at 0-1200 PAR light intensity and 400 ppm CO 2 .

Measurement of C and N contents
The contents of C and N were analyzed by using a C/N element analyzer (Elementar vario EL cub). The dry samples were powdered and used for analysis.
GUS staining assay GUS staining assay was performed as described by Pelagio-Flores et al. [32]. Brie y, the proAOX1a::GUS and proUCP1:GUS transgenic lines were incubated in the staining buffer containing 0. Determination of nitrate reductase (NR) and nitrite reductase (NiR) activity The NR activity was analyzed as described by Du et al. [33]. The fresh tissues were homogenized with the buffer [0.1 M HEPES-KOH (pH 7.5), 3 % polyvinylpolypyrrolidone, 1 mM EDTA and 7 mM cysteine], and centrifuged at 10,000 g for 10 min. The supernatant was used for NR activity analysis.
The activity of NiR was determined according to Datta and Sharma [34]. Leaves were homogenized with the buffer [50 mM phosphate buffer (pH 8.8), 3% BSA, 1 mM EDTA and 25 mM cysteine]. Homogenate was centrifuged at 13,000 g for 15 min at 4℃. The supernatant was used directly for NiR activity measurement.

Determination of nitrate uptake and transport
The uptake and transport of nitrate were analyzed by using 15 N-KNO 3 (Sigma-Aldrich 57654-83-8) according to Gandi et al. [9]. 14-day-old seedlings were treated for 10 d on low-N (1 N) condition and then 24 h on 0 N condition, then transferred to the growth medium containing 10 mM 15 N-KNO 3 (CK) or 1 mM 15 N-KNO 3 (1 N) for 6 h. After 6 h of labelling, samples were washed and oven-dried, and then ground to ne powder. Total 15 N enrichment was measured using a gas stable isotopic ratio mass spectrometer (IRMS).

RNA-Seq analysis
The 14-day-old hydroponically grown seedlings were treated under control (11 N) or low nitrogen (1 N) conditions for 16 d. The shoots were collected and used for the high-throughput sequencing in Bai Mike Company (http://www.biocloud.net/). Gene expression levels were estimated by fragments per kilobase of transcript per million fragments (FPKM) mapped. Differential expression analysis of two groups was performed using the DESeq2 [35]. The p values were adjusted using the Benjamini and Hochberg's approach for controlling the false discovery rate (FDR). Genes with an adjusted p < 0.01 found by DESeq2 were assigned as differentially expressed. The FDR < 0.01 and Log 2 (FC)|≥ 2 were set as the threshold for signi cantly differentially expressed genes (DEGs). The KOBAS software was used to test the statistical enrichment of DGEs in KEGG pathways [36].

Statistical analysis
The data were expressed as means ± SD (n ≥ 3). All statistical analysis were conducted via one-way ANOVA and SPSS17.0. In the gures, the lowercase letters indicate the signi cant difference at P<0.05.
To explore the function and relationship of AOX1a and UCP1 in Arabidopsis tolerance to N de ciency, the aox1a and ucp1 were crossed to generate the double mutant aox1a/ucp1. PCR analysis con rmed the screened homozygous lines (Fig. S3). Phenotypic analysis of 8-day-old single and cluster seedlings showed that there was no difference among Col-0, aox1a, ucp1 and aox1a/ucp1 on 1/2 MS and 1 N conditions (Fig. 1a). However, under 0.1 N and 0 N conditions, the primary root length of aox1a/ucp1 seedlings was shorter than that of Col-0 seedlings (Fig. 1a, d), and the difference reached the maximum at day 8 (Fig. S4). Under 0.1 N condition for 8 d, the primary root length of aox1a/ucp1 was reduced by 25.4% compared to Col-0 (Fig. 1d). Moreover, the anthocyanin content and the purple rate of aox1a/ucp1 cotyledons were higher than that of Col-0 and single mutants under 0.1 N condition (Fig. 1b, c, e). The cotyledon purple rate of Col-0, aox1a, ucp1 and aox1a/ucp1 seedlings was 60.2%, 58.4%, 61.1%, and 82.8%, respectively (Fig. 1e). Thus, AOX1a and UCP1 might be collaboratively involved in the adaptation of Arabidopsis seedlings to low-N stress.
UCP1 plays more important roles in regulating photosynthesis than AOX1a upon low-N condition To further investigate the mechanism of AOX1a and UCP1 involvement in Arabidopsis adaption to low-N stress, we analyzed the rosette leaf area, shoot biomass and photosynthetic capacity of Col-0, aox1a, ucp1 and aox1a/ucp1 seedlings. In order to accurately control the N concentration in the medium, seedlings were grown hydroponically. Results showed that the rosette leaf area, fresh weight (FW) and dry weight (DW) of ucp1 and aox1a/ucp1 seedlings were markedly lower than that of Col-0 under moderate-N (5 N) and low-N (1 N) conditions, while aox1a had no signi cant difference compared to Col-0 (Fig. 2). Under low-N condition, the rosette leaf area, FW and DW in ucp1 were reduced by 24.4%, 29.0% and 27.7%, respectively, and by 44.6%, 26.3%, and 27.3% in aox1a/ucp1, respectively, compared to Col-0 ( Fig. 2b-d). These results indicate that loss of UCP1 function results in more signi cant inhibition of growth than loss of AOX1a function under low-N condition.
The total chlorophyll content (chlt) in ucp1 and aox1a/ucp1 plants were decreased by 7.4% and 9.2%, respectively, in comparison with that in Col-0 upon low-N condition. However, in aox1a, it did not change (Fig. 3a). Under the control condition, the net photosynthetic rate (A) and electron transfer rate (ETR) in ucp1 were signi cantly lower than that in Col-0. However in aox1a and aox1a/ucp1 plants, A and ETR were not different from that in Col-0 (Fig. 3b, c). Under low-N condition, A and ETR in aox1a was similar to that in Col-0. In contrast, they were signi cantly lower in ucp1 and aox1a/ucp1 than that in Col-0. A and ETR were reduced by 27.4% and 31.0%, respectively, in ucp1, and by 29.8% and 26.5%, respectively, in aox1a/ucp1 (Fig. 3b, c). The analysis result of light response curve showed that the light saturation point in ucp1 was markedly lower than that in Col-0 when the light intensity exceeded 200 µmol photons m − 2 s − 1 (PAR) under the control condition (Fig. 3d). Under low-N condition, the light saturation point in both ucp1 and aox1a/ucp1 plants were signi cantly lower than that in Col-0 and aox1a, but there was no difference between ucp1 and aox1a/ucp1 (Fig. 3e). These results indicate that aox1a/ucp1 and ucp1 plants have low photosynthetic capacity under low-N stress. UCP1 plays more important roles in regulating photosynthesis than AOX1a does in Arabidopsis tolerance to low-N stress.
Mutation of AOX1a and UCP1 decreases C and N contents in the shoots of aox1a/ucp1 Above results con rmed that mutation of AOX1a and UCP1 affects the photosynthetic capacity of aox1a/ucp1. We subsequently analyzed the changes of C and N contents under low-N stress. Upon low-N and moderate-N conditions, the percentage of C in Col-0, aox1a, ucp1, and aox1a/ucp1 shoots was decreased. Interestingly, there was a marked difference only between ucp1 and Col-0 under 1 N condition (Fig. 4a). Low N similarly decreased the percentage of N in all plants (Fig. 4b). The total C and N contents in ucp1 and aox1a/ucp1 shoots were signi cantly lower than that in Col-0 upon low-N condition (Fig. 4c,  d), which were reduced by 21.6% and 26.1%, respectively, in ucp1 and by 24.2% and 23.11%, respectively, in aox1a/ucp1 under 5 N condition. Under 1 N condition, the total C and N contents were reduced by 29.6% and 28.6%, respectively, in ucp1 and by 27.0% and 27.9%, respectively, in aox1a/ucp1. However, in aox1a, the total C and N contents were similar to that in Col-0 under all N conditions. Thus, the photosynthetic C xation and N assimilation capacity in ucp1 and aox1a/ucp1 were lower than that in Col-0 and axo1a plants under low-N condition, and this explained the decreased biomass in ucp1 and aox1a/ucp1 plants.

Expression patterns of AOX1a and UCP1 and changes of respiratory rates upon low-N condition
To investigate the expression patterns of AOX1a and UCP1 upon low-N condition, the proAOX1a:GUS and proUCP1:GUS transgenic lines were generated. GUS staining results showed that 0.1 N treatment induced marked increase of AOX1a and UCP1 expression in 6-day-old seedlings. AOX1a was expressed in the whole seedling, whereas UCP1 was mainly expressed in cotyledons and the maturation zone of primary roots (Fig. 5a-d). Under low-N condition, the AOX1a expression was higher than UCP1 in rosette leaves (Fig. 5e-h). At the reproductive stage, low N also increased the expression of AOX1a and UCP1 in owers and stamens (Fig. 5i-p). Speci cally, the expression of AOX1a was mainly increased in the anthers of stamens (Fig. 5n), while UCP1 was mainly expressed in the laments of stamens (Fig. 5p).
qRT-PCR results reported that low N induced the expression of AOX1a and UCP1 at almost every developmental stage. When Arabidopsis seeds were germinated and grown for 6 d in 0.1 N medium, the expression of AOX1a and UCP1 were enhanced by 9.3-fold and 1.8-fold, respectively (Fig. 5q). In rosette leaves, the response of AOX1a expression to low-N condition was higher than that of UCP1 (Fig. 5r). However, in owers and siliques, the AOX1a expression increased about 1.5-fold and 2.1-fold upon low-N condition, respectively, while the UCP1 expression was increased by 1.7-fold and 2.3-fold (Fig. 5s, t).
To clarify the relationship between AOX1a and UCP1 as well as the effect of low N on the alternative pathway capacity (V alt ) and UCP pathway activity (V ucp ), the respiratory rates were analyzed. Fig. S5 showed that low N signi cantly increased V alt and V ucp in Col-0. Under high-N (control) condition, V alt in ucp1 was decreased by 84.3% and V ucp in aox1a was decreased by 42.1% compared to Col-0.
Interestingly, the mutation of UCP1 in ucp1 stimulated the V alt by 63.5% upon low-N condition (Fig. S5a).
Similarly, V ucp was also increased in aox1a and Col-0 upon low-N condition. Thus, there should be the close relationship between AOX1a and UCP1 under normal growth condition.

Mutation of AOX1a and UCP1 affects N metabolism upon low-N condition
To investigate the effects of AOX1a and UCP1 mutations on N metabolism in Arabidopsis in response to low N, the differences of the uptake, assimilation and transport of N were analyzed in Col-0, aox1a, ucp1, aox1a/ucp1 plants. The results showed that low N decreased the NO 3 − content and NR activity, but increased NiR activity in Col-0, aox1a, ucp1 and aox1a/ucp1 shoots (Fig. 6a-c). Compared to Col-0 under 1 N condition, the NO 3 − content, NR and NiR activities were reduced by 11.5%, 9.5%, 15.0%, respectively, in ucp1 and by 34.0%, 21.8%, 13.3%, respectively, in aox1a/ucp1; however, there was no signi cant difference between aox1a and Col-0.
To analyze the uptake rate of NO 3 − and the transport from root to shoot, Arabidopsis seedlings were starved under 0 N condition for 24 h, and then transferred to the growth solution containing 10 mM 15 N-KNO 3 (control) or 1 mM 15 N-KNO 3 (1 N) for 6 h. Results showed that low N decreased the uptake rate of NO 3 − in Col-0, aox1a, ucp1 and aox1a/ucp1 roots (Fig. 6d). However, compared to Col-0, the uptake rate of NO 3 − in aox1a, ucp1 and aox1a/ucp1 were elevated by 15.4%, 16.7% and 21.6%, respectively, under the control condition, and by 6.9%, 12.0% and 21.2%, respectively, under 1 N treatment (Fig. 6d). In contrast, the transport activity of NO 3 − from root to shoot in aox1a, ucp1 and aox1a/ucp1 was signi cantly decreased compared to Col-0 under both control and low-N conditions (Fig. 6e). Based on these results, although mutations of AOX1a and UCP1 resulted in acceleration of NO 3 − uptake under low-N stress, the low assimilation rate and transport activity of N in aox1a/ucp1 seedlings impaired the N metabolism.
Mutation of AOX1a and UCP1 decreases the seed yield and the C/N content in aox1a/ucp1 under low-N stress N de ciency inevitably limits the reproductive growth of plants. Our results showed that low N signi cantly reduced the silique number and seed yield of Col-0, aox1a, ucp1 and aox1a/ucp1 plants (Fig. 7a, b). Loss of function of AOX1a and UCP1 resulted in more decrease of silique number and seed yield in aox1a/ucp1 (by 22.3% and 23.6%, respectively) than that in Col-0. A similar trend was also found in ucp1 mutant. However, there was no difference between aox1a and Col-0 (Fig. 7a, b). Figure 7c, d showed that the percentage of N was decreased in seeds of four genotypes upon low-N condition. However, the percentage of C in seeds had no signi cant changes compared to the control. Further result con rmed that the AOX1a mutation alone in aox1a did not affect the total C and N contents in seeds, but in ucp1 and aox1a/ucp1 seeds, the total C content was reduced by 21.1% and 21.8%, respectively, and the total N content was reduced by 27.4% and 34.1%, respectively (Fig. 7e, f). Thus, the decrease of C and N distribution in seeds accounted for the low seed yield in both ucp1 and aox1a/ucp1 plants.
Distinct transcript pro les in Col-0, aox1a, ucp1 and aox1a/ucp1 plants under low-N stress The different physiological response of Col-0, aox1a, ucp1, and aox1a/ucp1 seedlings to low-N stress might be caused by differential expressions of downstream genes involved in different metabolic pathways. To test the hypothesis and elucidate the molecular metabolism of UCP1 and AOX1a in Arabidopsis tolerance to low N, we compared gene expression pro les in Col-0, aox1a, ucp1, and aox1a/ucp1 shoots by using RNA sequencing. Compared to Col-0, transcriptomes of aox1a, ucp1, and aox1a/ucp1 had marked changes (Fig. S6). With thresholds of log 2 (FC) ≥ 2 and FDR < 0.01, there were 1605 up-regulated genes and 1559 down-regulated genes in Col-0 upon low-N condition. The up-and down-regulated genes were 1656 and 1153, respectively, in aox1a, 687 and 1378, respectively, in ucp1, and 2321 and 1796, respectively, in aox1a/ucp1 (Fig. S6a-d). KEGG functional enrichment analysis indicated that the DEGs in Col-0 were mainly enriched in ribosome, plant hormone signal transduction, biosynthesis of amino acids, and plant-pathogen interaction. DEGs were mainly enriched in ribosome, biosynthesis of amino acids and carbon metabolism in aox1a; in plant hormone signal transduction, plant-pathogen interaction and photosynthesis in ucp1; and in carbon metabolism, biosynthesis of amino acids, and photosynthesis in aox1a/ucp1 (Fig. S6e-f). Thus, mutation of UCP1 and AOX1a in aox1a/ucp1 markedly disturbed the photosynthesis and carbon metabolism pathways upon low-N condition.
To further investigate the mechanism of decreased resistance to low N in aox1a/ucp1, DEGs in Col-0 and aox1a/ucp1 were analyzed. Venn diagram showed that 1233 and 923 genes were co up-and downregulated, respectively, in Col-0 and aox1a/ucp1 (Fig. 8a). KEGG results showed that DEGs in both Col-0 and aox1a/ucp1 were mainly enriched in ribosome, biosynthesis of amino acids, photosynthesis and N metabolism. DEGs in Col-0 only (1961 genes) were mainly enriched in ribosome, plant hormone signal transduction, plant-pathogen interaction. DEGs only in aox1a/ucp1 (1008 genes) were mainly enriched in C metabolism, amino acid synthesis, C xation in photosynthesis organisms, and glycolysis ( Fig. 8b-d).
The numbers of DEGs involved in the N metabolism, C metabolism, photosynthesis, por and chl metabolism and ribosome were 4, 44, 11, 6, 12, respectively, in aox1a/ucp1, but were 2, 3, 1, 2, 57, respectively, in Col-0 (Fig. 8e, f). Figure 8g further showed that most DEGs in both Col-0 and aox1a/ucp1 in the aforementioned metabolism pathways were generally down-regulated. However, the decrease was more signi cant in aox1a/ucp1 than in Col-0 under low-N stress (Fig. 8g). qRT-PCR was performed to validate some of the DEGs in Col-0 and aox1a/ucp1. Among the six genes examined, they all showed similar expression patterns as the RNA-seq results (Fig. 8h). Thus, the decreased expressions of genes related to C metabolism, photosynthesis, N metabolism in aox1a/ucp1 could be one of the reasons for the low photosynthetic capacity and C/N level in seeds.

Discussion
Nitrogen (N) assimilation is essential for leaf metabolism and seed yield. In plants, carbon (C) metabolism are closely related to the N levels and its assimilation capacity [5,[37][38]]. The energy generated by light reaction not only supply C xation, but also N assimilation [8,9]. Therefore, plants must balance the energy distribution between C and N metabolism. Many reports have indicated that the alternative pathway and the uncoupling protein (UCP) pathway play key roles in decreasing photoxidation damage and maintaining photosynthesis rate by dissipating excessive reductants in chloroplasts under stresses [10,20]. Moreover in plants, the dysfunction of AOX1a or UCP1 disturbs the N assimilation and C metabolism [9,10]. In the study, we investigated the functions of AOX1a and UCP1 involved in regulating the balance of C and N metabolism in Arabidopsis adaptation to low-N stress.
Dysfunction of AOX1a and UCP1 decreases the tolerance of Arabidopsis to low-N stress at the seedling stage Alternative pathway is extensively involved in plant tolerance to environmental stresses, i.e., cold, salt, drought, low-N and so on [10,16,30,39]. qRT-PCR results showed that low-N signi cantly induces the expressions of AOX1a, AOX1c and AOX2, but decreases the expressions of AOX1b and AOX1d. It has been reported that the expression of AOX1a is tightly related to the N level in plants [10,16] and is involved in regulating N metabolism and the intensity of mitochondrial stress signal pathway [40]. However, the primary root length and fresh weight of seedlings of AOX mutants exhibit no signi cant difference to Col-0 upon low-N condition. Watanabe et al. [16] reported that low N can stimulate AOX1a expression, however, dysfunction of AOX1a did not affect the biomass of Arabidopsis. Thus, there should be additional compensation mechanisms when the alternative pathway is affected in plant adaptation to low N.
The functions of AOX and UCP are closely coordinated, as demonstrated in plant heat generation, cell energy balance and adversity defenses [22,30,41]. The effect of UCP1 dysfunction on N metabolism is more signi cant than AOX1a de ciency [16,18]. qRT-PCR showed that low N signi cantly increases UCP1, UCP2 and UCP3 expression. However, the phenotypes of 8-day-old UCP mutants are also similar to that of Col-0 under low-N condition. Double mutation of AOX1a and UCP1 in aox1a/ucp1 results in reduced primary root length and accumulation of anthocyanin compared to Col-0 and the single mutants. It is well known that increase of the root/shoot ratio, the accumulation of anthocyanin and the recycling of N from old leaves to young leaves are common features during plant tolerance to low-N condition [4,42]. pap1, a mutant defective in anthocyanin biosynthesis, shows decreased tolerance to low-N stress [43]. Our results suggest that aox1a/ucp1 has lower tolerance to low N than Col-0 and the single mutants. The high anthocyanin level and purple ratio of cotyledons could facilitate aox1a/ucp1 seedlings adaptation to low-N stress. In the process, AOX1a and UCP1 collaboratively function in the adaptation of Arabidopsis seedlings to low-N stress.
Mutation of UCP1 leads to low photosynthetic capacity as well as C and N content in aox1a/ucp1 under low-N stress It has been reported that loss of UCP1 function leads to photorespiration limitation and reduction of C assimilation rate under high light condition, but not loss of AOX1A function [9,20]. Under the control condition (high N), A and ETR in ucp1 are markedly decreased compared to that in Col-0. Furthermore, the light saturation point in ucp1 is signi cantly lower than that in Col-0, but not in aox1a and aox1a/ucp1. Interestingly, UCP1 mutation results in low AOX protein content under normal growth condition, but does not affect other mETPs [20]. The present result further showed that UCP1 mutation signi cantly decreases V alt under high-N condition. Similarly, AOX1a mutation also signi cantly decreases the activity of the UCP pathway. Under low-N stress, higher V alt is induced in ucp1 seedlings. These results indicate that there is a close correlation between the alternative pathway and the UCP pathway in Arabidopsis.
The photosynthetic capacity in plants is positively related to N levels in leaves [5,44]. The N availability in plant affects leaf growth and photosynthetic area through in uencing protein synthesis [37]. Under moderate and low-N conditions, the total rosette leaf area of ucp1 and aox1a/ucp1 is lower than that of Col-0. N mainly exists in photosynthetic enzymes and chlorophyll in plant shoots, so its level will affect the function and the number of chloroplasts [45,46]. The chlorophyll content, A and ETR in ucp1 and aox1a/ucp1 are signi cantly reduced under low-N condition. However, AOX1a mutation does not affect the photosynthesis capacity and leaf area. Moreover, there is a similar trend in above parameters between ucp1 and aox1a/ucp1 under low-N stress, indicating that the dysfunction of UCP1 is responsible for the low photosynthetic capacity and biomass in aox1a/ucp1 under low-N condition.
It has been reported that mutation of AOX1a and UCP1 affects the assimilation rate of C and N in Arabidopsis leaves [9]. Under low-N condition, the relative C and N contents in aox1a, ucp1, aox1a/ucp1 and Col-0 shoots are decreased. However, the total C and N contents in ucp1 and aox1a/ucp1 shoots are lower than that in aox1a and Col-0 shoots. Moreover, accumulation of shoot biomass shows a similar trend to the total C and N contents as well as the photosynthetic capacity under low-N condition.
Watanabe et al., [16] reported that the balance of C and N in plants might be strictly modulated by pathways other than AP upon low-N condition. Thus, the low photosynthetic capacity in ucp1 and aox1a/ucp1 under low-N condition results in decreased total C levels and biomass in shoots; in the process, UCP1 plays a more essential role than AOX1a.
Mutation of UCP1 disturbs the C/N ratio and leads to low seed yield in aox1a/ucp1 upon low-N condition The seed yield and pod setting rate in plants are closely related to N level and its assimilation capacity [5,47,48]. 15 N tracing results indicate that mutation of AOX1a and UCP1 accelerates the NO 3 − acquisition in aox1a/ucp1 roots under the control and low-N conditions. However, the transport activity of NO 3 − from roots to shoots is lower in aox1a/ucp1 than that in Col-0, which eventually results in a signi cant decrease of NO 3 − level in aox1a/ucp1 leaves upon low-N condition; moreover, the NR activity in leaves of Col-0, aox1a, ucp1, and aox1a/ucp1 is signi cantly decreased, but the activity decreases the most in aox1a/ucp1. NR activity has a signi cant positive correlation with protein content in leaves, and it can be used as a biochemical indicator to weigh grain yield and protein content [46,49]. Under low-N stress, the seed yield and silique number were decreased more in ucp1 and aox1a/ucp1 than that in Col-0 and aox1a. Gandi et al. [9] con rmed that UCP1 plays a more key role than AOX1a in the N and C metabolism under normal and high light conditions in Arabidopsis. Based on our results, we propose that loss of UCP1 function, not AOX1a, leads to the low seed yield in aox1a/ucp1 upon low-N condition.
The total C and N levels in seeds are positively correlated with the protein and fat contents in seeds, which also determines the quality and yield of seeds [50,51]. Loss of function of UCP1 and AOX1a in aox1a/ucp1 did not affect the C content in seeds, but signi cantly decreased the N content under low-N stress, which resulted in a high C/N ratio in aox1a/ucp1 seeds. Comparatively, the total C and N content in ucp1 and aox1a/ucp1 seeds were signi cantly lower than those in Col-0 and aox1a. Watanabe et al.
[16] indicated that AOX induced by low-N condition does not play an important role on the C/N ratio, instead it could be controlled by pathways other than alternative pathway. In short, the low N utilization e ciency in aox1a/ucp1 disturbes the metabolism of C and N, and eventually results in low seed yield and high C/N ratio under low-N stress; and UCP1 plays a key role for the maintenance of C/N ratio.
RNA-Seq analysis of Col-0 and aox1a/ucp1 shoots reveals the molecular mechanism for their differences in C and N metabolism as well as seed yield upon low-N condition. KEGG results indicate that low N disturbs ribosomes, photosynthesis, N metabolism, C metabolism, amino acid synthesis, porphyrin (por) and chlorophyll (chl) metabolism in Col-0 and aox1a/ucp1. Noticeably, by comparing different DEGs between Col-0 and aox1a/ucp1, we found that aox1a/ucp1 has more DEGs enriched in photosynthesis, N metabolism, C metabolism, and por and chl metabolism than Col-0; and most of their expressions are down-regulated. Moreover, the expression of common DEGs in photosynthesis and C/N metabolism is more down regulated in aox1a/ucp1 than that in Col-0 upon low-N condition. Thus, the decreased expressions of genes related to C metabolism, photosynthesis, and N metabolism in aox1a/ucp1 might be responsible for the low photosynthetic capacity as well as C and N levels in seeds.
Taken together, 8-day-old aox1a/ucp1 seedlings showed higher sensitivity to low-N stress in comparison with Col-0 and single mutant seedlings, suggesting that AOX1a and UCP1 are both involved in the Arabidopsis tolerance to low N at the seedling stage. The photosynthetic rate and the N assimilation capacity in ucp1 and aox1a/ucp1 plants are markedly weakened, leading to decreased leaf area, total C and N content, shoot biomass as well as seed yield compared to Col-0 and aox1a. These results indicate UCP1 plays more essential roles than AOX1a in Arabidopsis adaptation to low N at the vegetative and reproductive stages. RNA-seq analysis cos rmd that AOX1a and UCP1 mutation leads to more signi cant down-regulation of genes in photosynthesis and C/N metabolism in aox1a/ucp1 shoots in comparison with Col-0, which provides powerful evidence for the low seed yield, C and N levels in aox1a/ucp1. Noticeably, compared to AOX1a, UCP1 plays a more important role in regulating C/N ratio and yield of seeds in Arabidopsis adaptation to low-N stress.