Introduction

Nitrate is the major nitrogen source for most plants, especially those grown in aerobic soil conditions1. Once taken up from soil, nitrate is either assimilated or stored in vacuoles. As the largest organelle in fully expanded plant cells, vacuoles are identified as the major nitrate storage pools and contain up to 90% of the total cellular nitrate2, 3. However, vacuolar nitrate is not readily accessible to NR (nitrate reductase), thus it has to be reallocated for metabolic use when necessary4, 5. Vacuolar nitrate release helps to maintain the relative steady level of cytosolic nitrate when external nitrogen supply was limited6, 7. During the dark-to-light transition, nitrate remobilization from vacuoles was also observed to comply with the new steady state caused by the increased NR activity4. Therefore, vacuolar nitrate and its remobilization are important for the regulation of nitrogen assimilation and nitrogen use efficiency5, 8, 9.

Transport across the tonoplast is energized by the vacuolar H+-ATPase (V-ATPase) and the vacuolar H+-pyrophosphatase (V-PPase), which create the proton gradient and the membrane potential10,11,12. For nitrate, its accumulation in vacuoles is probably mediated by the nitrate/proton antiport machinery13,14,15,16 and the nitrate/proton symport system may serve to remobilize vacuolar nitrate16,17,18. However, only a few tonoplast localized nitrate transporters have been identified up to date. AtCLCa and AtNRT2.7 are two transporters responsible for nitrate accumulation in vacuoles19,20,21. AtCLCa was a tonoplast localized 2NO3 −/1 H+ antiporter expressed in both shoots and roots20, 22. Disruption of AtCLCa led to approximately 50% decrease of vacuolar nitrate, suggesting an important role for AtCLCa in vacuolar nitrate accumulation19, 20. AtNRT2.7, however, was a tonoplast localized transporter expressed exclusively in seeds, which regulated the kinetics of seed germination by affecting nitrate storage in seed vacuoles21. AtCLCc was also supposed to be involved in vacuolar nitrate accumulation, because it was tonoplast localized and the related mutants showed lower nitrate contents22, 23. Regarding nitrate efflux from vacuoles, however, indirect evidences imply that AtCLCb and OsNPF7.2 might get involved, as they both were tonoplast-localized, and heterologous expression in Xenopus laevis oocytes indicated that they mediated nitrate uptake, but no evidence showed that functional disruption of these genes led to nitrate accumulation in vacuoles24, 25. AtCLCa was also implied to get involved in vacuolar nitrate efflux, because it mediated anion homeostasis in stomata movement26, while nitrate is one of the anions contributing to stomatal movement27, 28.

In the current study, three tonoplast-localized NRT1/NPF family members NPF5.11, NPF5.12 and NPF5.16 were identified by bioinformatics analysis, and functional characterization was performed. Our data suggested that these three transporters were all tonoplast localized, and mediated nitrate uptake in a pH-dependent low-affinity manner when heterologously expressed in oocytes. Further analysis indicated that they possibly modulated nitrate allocation between roots and shoots via vacuolar nitrate release.

Results

Tonoplast Localization of NPF5.11, NPF5.12 and NPF5.16

Based on previous studies about vacuole proteome29, we targeted NPF5.12 and its close homologs NPF5.11 and NPF5.16, members of NRT1/NPF family30, as candidate transporters for vacuolar nitrate transport. NPF5.11, NPF5.12 and NPF5.16 were predicted to contain 11, 12, 10 transmembrane domains (http://www.cbs.dtu.dk/services/TMHMM/), respectively. To investigate the subcellular localization of NPF5.11, NPF5.12 and NPF5.16, NPF5.11-EYFP, NPF5.12-EYFP and NPF5.16-EYFP driven by the cauliflower mosaic virus 35 S promoter were transiently expressed in Arabidopsis mesophyll protoplast. The yellow fluorescence signals of NPF5.11-EYFP (Fig. 1a–c), NPF5.12-EYFP (Fig. 1d–f) and NPF5.16-EYFP (Fig. 1g–i) were detected in the membrane around the large central vacuole. Similar results were obtained by transiently expressing these fusion proteins in onion epidermal cells, verifying that NPF5.11, NPF5.12 and NPF5.16 were tonoplast localized (Supplementary Fig. S1).

Figure 1
figure 1

Subcellular localization of NPF5.11, NPF5.12 and NPF5.16. NPF5.11-EYFP (a–c), NPF5.12-EYFP (d–f), NPF5.16-EYFP (g–i) or EYFP (j–l) was driven by the cauliflower mosaic virus 35 S promoter and transiently expressed in Arabidopsis mesophyll protoplasts. Overlap images of EYFP (green) and chlorophyll (red) fluorescence (a,d,g,j), EYFP fluorescence (b,e,h,k), and bright-field (c,f,i,l) images are shown. Bars = 20 μm.

Full size image

NPF5.11, NPF5.12 and NPF5.16 are pH-dependent Low-Affinity Nitrate Transporters

Given that Xenopus laevis oocytes did not contain vacuoles, we firstly tested the expression and localization of NPF5.11, NPF5.12 and NPF5.16 in oocytes. As a well-documented nitrate transporter31, NRT1.8 fused with GFP was used as a positive control and its fluorescence was detected at the rim of oocytes (Supplementary Fig. S2). Likewise, NPF5.11-EYFP, NPF5.12-EYFP and NPF5.16-EYFP fusion proteins could express in plasma membrane of oocytes though they were tonoplast localized transporters in Arabidopsis, indicating that we could use oocytes to explore the function of NPF5.11, NPF5.12 and NPF5.16 (Supplementary Fig. S2).

Electrophysiological analysis using cRNA-injected oocytes were performed to test whether NPF5.11, NPF5.12 and NPF5.16 could use nitrate as substrate. After 2 days of incubation, oocytes were voltage clamped at −60 mV and perfused with 10 mM nitrate at pH 5.5. Compared with water-injected oocytes (Fig. 2a), a larger inward current was induced by CHL1-injected oocytes (Fig. 2b), as reported before32. NPF5.11-, NPF5.12- or NPF5.16-injected oocytes also induced inward currents (Fig. 2c–e), indicating that they were electrogenic transporters using nitrate as the substrate.

Figure 2
figure 2

Functional characterization of NPF5.11, NPF5.12 and NPF5.16 in oocytes. (a–e) Currents elicited in oocytes injected with H2O (a), CHL1 cRNA (b), NPF5.11 cRNA (c), NPF5.12 cRNA (d) or NPF5.16 cRNA (e). Oocytes were voltage clamped at −60 mV and representative inward currents elicited by 10 mM NO3 − at pH 5.5 were recorded. (f–h) Nitrate uptake activity in oocytes injected with H2O, NPF5.11 cRNA, NPF5.12 cRNA, NPF5.16 cRNA, NRT1.8 cRNA or CHL1 cRNA. Oocytes were incubated with 10 mM 15NO3 − at pH 5.5 (f), 0.25 mM 15NO3 − at pH 5.5 (g) or 10 mM 15NO3 − at pH 7.4 (h) for 12 h. Values are means ± SD (n = 8–12). Asterisks indicate difference at P < 0.01 (**) compared with the H2O-injected oocytes by Student’s t-test. (i–k) Uptake kinetics of NPF5.11 (i), NPF5.12 (j) and NPF5.16 (k). Oocytes injected with NPF5.11 cRNA (i), NPF5.12 cRNA (j) or NPF5.16 cRNA (k) were incubated with indicated concentrations of 15NO3 − at pH 5.5 for 1.5 h, and the 15N contents were determined. Values are means ± SD (n = 6–12). The Km was calculated by fitting to the Michaelis-Menten equation using a nonlinear least squares method in the SigmaPlot program. The Km was 2.57 mM, 4.84 mM, or 2.91 mM for NPF5.11, NPF5.12 or NPF5.16, respectively.

Full size image

Nitrate transport activities of NPF5.11, NPF5.12 and NPF5.16 were further confirmed by analyzing 15NO3 − uptake activity. NPF5.11-, NPF5.12-, NPF5.16- or NRT1.8-injected oocytes showed enhanced 15NO3 − uptake activity when incubated with 10 mM 15NO3 − at pH 5.5, compared with water-injected oocytes (Fig. 2f). However, NPF5.11-, NPF5.12- or NPF5.16-injected oocytes almost did not uptake 15NO3 − when assayed with 0.25 mM 15NO3 − at pH 5.5, while CHL1-injected oocytes still showed high uptake activity (Fig. 2g). In addition, as expected for proton-coupled transporters, 15NO3 − uptake activities of NPF5.11-, NPF5.12- or NPF5.16-injected oocytes at pH 7.4 were much lower than those at pH 5.5, comparable with the negative control (Fig. 2h). It is worth mentioning that NPF5.11, NPF5.12 and NPF5.16 did not efflux nitrate from oocytes under pH 5.5 or pH 7.4 (Supplementary Fig. S3).

To further determine the uptake affinity of NPF5.11, NPF5.12 and NPF5.16, uptake activity of NPF5.11-, NPF5.12- or NPF5.16-injected oocytes at pH 5.5 was measured using different concentrations of 15NO3 − ranging from 0.25 mM to 30 mM as substrates. The K m for nitrate was calculated by fitting to the Michaelis-Menten equation, and was estimated as 2.57 mM, 4.84 mM, 2.91 mM respectively for NPF5.11, NPF5.12 and NPF5.16 (Fig. 2i–k). Taken together, these results suggested that NPF5.11, NPF5.12 and NPF5.16 were pH-dependent low-affinity nitrate transporters.

NPF5.11, NPF5.12 and NPF5.16 are mainly expressed in vascular stele of roots and leaves

The tissue localization of genes could provide hint for their physiological role. To elucidate the expression pattern of NPF5.11, NPF5.12 and NPF5.16, promoter-GUS (β-glucuronidase) reporter analysis was performed. The promoter region of NPF5.11, NPF5.12 and NPF5.16 were used for driving the expression of GUS in Columbia (Clo-0). As shown in Fig. 3, NPF5.11, NPF5.12 and NPF5.16 had a similar expression pattern, expressing in both shoots and roots. In shoots, they were mainly expressed in leaf veins while the mesophyll cells were also stained (Fig. 3a,e,i). In roots, GUS activity was detected in root vascular stele (Fig. 3b,f,j). Cross-sections of young seedling roots showed that NPF5.11 pro ::GUS, NPF5.12 pro ::GUS and NPF5.16 pro ::GUS were expressed in pericycle cells and parenchyma cells, and NPF5.11 pro ::GUS was also expressed in the phloem (Fig. 3c,d,g,h,k,l).

Figure 3
figure 3

NPF5.11, NPF5.12 and NPF5.16 are preferentially expressed in vascular tissues. Histochemical localization of GUS activity in NPF5.11 pro ::GUS transgenic plants (a–d), NPF5.12 pro ::GUS transgenic plants (e–h) and NPF5.16 pro ::GUS transgenic plants (i–l). The expression patterns of NPF5.11, NPF5.12 and NPF5.16 were determined in whole-mount seedlings (a,e,i), seedling roots (b,f,j) or cross-sectioned seedling roots (c,d,g,h,k,l). (m,n,o) Transcript expression of NPF5.11 (m), NPF5.12 (n) and NPF5.16 (o) in 28 d old plants. 1–8 indicated leaf positions arranged according to leaf ages (old to young); R, root; F, flower; S, stem. Data were normalized to that of SAND. Values are means ± SD, n = 3. Bars = 10 μm.

Full size image

The expression patterns of these three genes in adult plant were further investigated by qRT-PCR analysis. NPF5.11, NPF5.12 and NPF5.16 all showed high expression in root while the expression in flower and stem was quite low (Fig. 3m,n,o). The expression of NPF5.12 in old leaves was higher than that in young leaves, while NPF5.16 was preferentially expressed in young leaves (Fig. 3n,o).

More nitrate is translocated to shoots in triple mutant

Considering the tonoplast localization (Fig. 1) and pH-dependent nitrate uptake (Fig. 2), we proposed that NPF5.11, NPF5.12 and NPF5.16 might be responsible for uptaking nitrate from vacuole (pH 5.533) to cytoplasm in Arabidopsis. To test this hypothesis, we generated several lines of their single, double, and even triple mutants (Supplementary Figs S4,S6,S7). Note that only double mutant lines of npf5.12 npf5.16 were generated because NPF5.11 and NPF5.12 were tightly linked in Arabidopsis genome and npf5.11 mutant lines were in Ws background. Given vacuolar nitrate efflux is supposed to be enhanced when nitrogen is limited, we firstly analyzed the nitrate contents in leaves and roots in these mutants under both control condition and nitrogen-starved condition. As shown in Supplementary Figs S5, S6 and S7, no obvious difference was observed between the wild type control and all the mutants. However, when they were fed with 15NO3 −, the ratio of 15N concentration in shoots against that in roots (shoot/root) was higher in triple mutant lines than in the wild type (Fig. 4a, Supplementary Fig. S8), while no significant difference was observed between the single mutant lines and the wild type (Supplementary Fig. S9). These results suggested that more 15NO3 − was translocated to shoots in triple mutants, while our data also indicated that the root uptake capacity of triple mutant lines was not affected (Fig. 4b).

Figure 4
figure 4

Root-to-shoot nitrate transport enhanced in the triple mutant plants npf5.11 npf5.12 npf5.16. Plants were grown in hydroponics for 28 days and treated with 2.25 mM K15NO3 for 30 min. 15N contents in shoots and roots were analyzed. 15N concentration ratio between shoots and roots (S/R ratio, a) and root uptake activity (b) were determined. Values are means ± SD, n = 3. Asterisks indicate difference between wild type and triple mutant lines at P < 0.05 (*) by Student’s t-test.

Full size image

Root nitrate content is reduced in NPF5.12 overexpression lines

To further investigate the function of NPF5.11, NPF5.12 and NPF5.16, the overexpression lines of NPF5.11, NPF5.12 and NPF5.16 under the control of 35 S promoter were generated (Fig. 5a,b) and the nitrate content was analyzed (Fig. 5c, Supplementary Figs S10, S11). The result showed that nitrate contents in roots of NPF5.12 overexpression lines were lower than that of wild type under nitrogen-starved condition (Fig. 5c). This observation was not found in NPF5.11 and NPF5.16 overexpression lines (Supplementary Fig. S11). One explanation could be that NPF5.11 and NPF5.16 might require other components to function properly in planta.

Figure 5
figure 5

Decreased nitrate accumulation in roots of NPF5.12 overexpression lines. (a,b) Identification of NPF5.12 overexpression lines by RT-PCR (a) and quantitative PCR analysis (b). (c) 24 days old hydroponically grown plants were subjected to nitrogen-starvation for 30 h, then roots were sampled and nitrate contents were determined by HPLC. 12OE1 and 12OE2 were two independent NPF5.12 overexpression lines. Values are means ± SD, n = 5–7. Asterisks indicate difference between wild type and overexpression lines at P < 0.05 (*) and P < 0.01 (**) by Student’s t-test.

Full size image

Discussion

Significant progresses have been made in clarifying the nitrate uptake and transport in Arabidopsis by the characterization of the transporters in NRT1/NPF, NRT2, CLC and SLAC/SLAH families34. However, our current knowledge about nitrate transport across the tonoplast is quite limited though the significance of this process is widely recognized.

The transporters responsible for vacuolar nitrate efflux should be tonoplast-localized and uptake nitrate toward cytoplasm. Our data suggested that NPF5.11, NPF5.12 and NPF5.16 were localized in tonoplast in Arabidopsis (Fig. 1) and plasma membrane in Xenopus laevis oocytes (Supplementary Fig. S2), and in oocytes they could elicit inward currents by external nitrate and uptake nitrate in a pH dependent way (Fig. 2). When these results were assigned to the topology of the plant tonoplast, the inward currents might represent NO3 −/H+ efflux from the vacuole to the cytoplasm24, because the external medium for oocytes corresponds to the vacuole in planta35, 36 and the pH of vacuole in Arabidopsis is about 5.533, 37. Thus we speculated that NPF5.11, NPF5.12 and NPF5.16 were responsible for vacuolar nitrate release in Arabidopsis.

Considering that NPF5.11, NPF5.12 and NPF5.16 are predominantly expressed in vacuole membrane of pericycle cells and xylem parenchyma cells in roots (Figs 1, 3). We proposed that they might be involved in the regulation of nitrate long-distance transport by modulating the vacuolar sequestration capacity (VSC) of nitrate in roots. Relationship between VSC and long-distance transport of metals in plant have been well documented38, and accumulating evidences indicated that VSC of essential anions including sulfate and nitrate also regulated their long-distance transport39, 40. The vacuolar sulfate efflux transporters SULTR4;1 and SULTR4;2 played an essential role in delivering sulfate to the xylem vessels by balancing storage and turnover of sulfate in the root vacuoles39. While Han et al., found that the decreased VSC of nitrate in roots would enhance nitrate transport to shoots and contribute to a higher nitrogen use efficiency (NUE)40. Our hypothesis about the physiological role of NPF5.11, NPF5.12 and NPF5.16 was supported by the observation that more proportion of 15NO3 − was translocated to shoots in triple mutant lines (Fig. 4) and overexpression of NPF5.12 resulted in a lower nitrate contents in roots (Fig. 5c). In the triple mutant lines, root VSC increased but not too much was available to newly absorbed nitrate due to the impaired nitrate efflux from vacuoles, resulting in less vacuolar nitrate sequestration and the consequent enhancement of 15NO3 − long-distance transport to shoots when fed with 15NO3 − for a short time (30 min). In NPF5.12 overexpression lines, the overall nitrate contents in roots decreased because of the lower VSC of nitrate, thus leading to the higher S/R ratios in the overexpression lines (Fig. 5c, Supplementary Fig. S10b).

No significant difference of nitrate contents was detected between all the mutants and wild type under various growth conditions we tested (Supplementary Figs S5, S6, S7). We proposed that there might be other transporters or channels that function redundantly with NPF5.11, NPF5.12 and NPF5.16, as the reutilization of vacuolar nitrate is crucial to environmental adaption for plants. The speculation is according with the observation that more 15NO3 − was translocated to shoots in triple mutant but not in single mutant. Similarly, no obvious changes in nitrate allocation was observed in mutants of AtCLCb or OsNPF7.2 24, 25. In addition, considering their specific tissue localization in roots (Fig. 3), we speculated that physiological effect of NPF5.11, NPF5.12 and NPF5.16 in nitrate allocation might be more noticeable specifically in pericycle cells and parenchyma cells. Thus more definitive evidences are needed in the future to demonstrate the working model for vacuolar nitrate efflux.

Methods

Plant Materials and Growth Conditions

Arabidopsis (Arabidopsis thaliana) ecotype Col-0 or Ws was used as the wild-type control. The Arabidopsis T-DNA mutant lines npf5.11-1 (FLAG_493A07) and npf5.11-2 (FLAG_592C02) were ordered from INRA (National Institute for Agricultural Research)41; npf5.12-1 (GABI_810C10) was ordered from NASC (European Arabidopsis Stock Centre)42; npf5.12-2 (CS871745), npf5.16-1 (SALK_152449 C) and npf5.16-2 (SALK_200474 C) were ordered from ABRC (Arabidopsis Biological Resource Center)43, 44. Homozygous mutant plants were screened by PCR45. The npf5.12 npf5.16 double mutant lines of were generated by crossing npf5.12 and npf5.16 and identified by PCR. The double mutant npf5.12 npf5.16 was further transformed with CRISPR/Cas9 system to generate the triple mutant npf5.11 npf5.12 npf5.16, using two different target sequences46. The double mutant lines used for transformation were: tri1, npf5.12-1 npf5.16-2; tri2, npf5.12-1 npf5.16-1; tri3, npf5.12-2 npf5.16-1. The CRISPR-Cas9 T-DNA was not existent in triple mutants by Cas9 PCR confirming47. The primers used in these assays are listed in Supplemental Table S1.

Arabidopsis plants were grown in quarter-strength hydroponic solution at 22 °C with 16-h-light/8-h-dark cycles as described48. Plants were grown to 3-4 weeks old and then were treated with nitrogen-starved nutrient solution by replacing KNO3 and Ca(NO3)2 with KCl and CaCl2 as indicated time.

Functional Analysis of NPF5.11, NPF5.12 and NPF5.16 in Xenopus laevis Oocytes

cDNA fragments of targeted genes were recovered by restriction digestion and then subcloned into the oocyte expression vector pOO249. cRNA was synthesized using the Ambion mMessage mMachine kit according to the manufacturer’s manual. Oocytes were isolated and injected with 50 ng cRNA as described previously50. CHL1 cRNA or NRT1.8 cRNA injected oocytes were used as positive control and water-injected oocytes were used as negative control. Oocytes were incubated in a ND-96 Ringer solution for 2 days as described31. Voltage clamp recordings were initiated in a bath solution containing 230 mM mannitol, 0.15 mM CaCl2, and 10 mM MES/Tris, pH 5.551. Nitrate uptake or efflux assays with 15NO3 − were performed as described50, 52, 53 using a continuous-flow isotope ratio mass spectrometer coupled to a carbon nitrogen elemental analyzer (Vario EL III/Isoprime; Elementar). Uptake kinetics assays were performed as described54.

EYFP Fusion and Subcellular Localization

The amplified NPF5.11, NPF5.12 and NPF5.16 cDNA fragments were cloned in frame in front of EYFP in the vector 35 S::EYFP/PA7, resulting in the NPF5.11-EYFP, NPF5.12-EYFP and NPF5.16-EYFP constructs under the control of the 35 S promoter. The resulted constructs were transiently expressed in Arabidopsis protoplast using the polyethylene glycol-mediated transformation method55. Alternatively, these constructs were also transiently expressed in onion epidermal cells using a particle gun–mediated system (PDS-1000/He; Bio-Rad). The transformed protoplasts and bombarded cells were held in the dark at 22 °C for more than 30 h followed by EYFP imaging using confocal microscopy (Olympus-FV1000).

For the GFP- or EYFP- fusion proteins expression in oocytes assay, the constructs were generated by introducing NRT1.8-GFP, NPF5.11-EYFP, NPF5.12-EYFP or NPF5.16-EYFP into the vector pOO2. The cRNA was synthesized and injected into oocytes. After cultivating 2 days, fluorescence was observed using confocal microscope (Olympus-FV1000).

Histochemical Analysis and Tissue Sectioning

A 1679-bp, a 1314-bp or a 1695-bp genomic fragment immediately upstream from the NPF5.11, NPF5.12 or NPF5.16 start codon, respectively, was amplified using primers listed in Supplemental Table S1. After sequencing, the fragments were cloned into the binary vector GUS /pCambias1300 and were then transformed into Col-0 as described31. GUS staining was performed overnight as described31. Semithin sections (4μm) were cut, mounted on glass slides, and visualized on Leica-DM6000.

RT-PCR and Quantitative RT-PCR

Plants were grown to 28 days old in hydroponics, and then were sampled as indicated. Total RNA was extracted using TRIzol reagent (Invitrogen). First-strand cDNA synthesis and RT-PCR were performed as described31. Quantitative RT-PCR was performed on a Corbett Research Rotor-Gene 3000 thermal cycler using SYBR Premix Ex-Taq (TaKaRa) according to the manufacturer’s protocol. The primers used in these assays are listed in Supplemental Table S1, and the expression levels were normalized to those of the SAND or Actin2 control.

Nitrate Content Determination by HPLC

Plants were grown to 3–4 weeks old and were treated with nitrogen-starved nutrient solution for indicated time. Leaves and roots were harvested and washed at least three times by ultrapure water for 5 min and then extracted nitrate as described56.

Analysis of Root-to-Shoot Nitrate Transport Using 15NO3 −

Wild type and triple mutant plants were grown in hydroponics for 28 days old and then were transferred to 0.1 mM CaSO4 for 1 min, labeled in quarter-strength hydroponics medium with 2.25 mM K15NO3 with 99% atom excess of 15N for 30 min. At the end of labeling, plants were washed in 0.1 mM CaSO4 for 1 min and the shoots and roots were separated. The shoots and roots were sampled and detected as described57.

Statistical Analysis

Two-tailed Student’s t tests were performed. Differences were deemed significant (*) at P < 0.05 and extremely significant (**) at P < 0.01.

Data Availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.