This Article Abstract Full Text (PDF) Supplemental Data All Versions of this Article: 19/11/3760    most recent tpc.106.048173v1 Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via CrossRef Citing Articles via Google Scholar Google Scholar Articles by Segonzac, C. Articles by Gibrat, R. Search for Related Content PubMed PubMed Citation Articles by Segonzac, C. Articles by Gibrat, R. Agricola Articles by Segonzac, C. Articles by Gibrat, R.

Nitrate Efflux at the Root Plasma Membrane: Identification of an Arabidopsis Excretion Transporter^[W]

^a Biochimie et Physiologie Moléculaire des Plantes, Agro-M/Centre National de la Recherche Scientifique/Institut National de la Recherche Agronomique/Université Montpellier 2, Unité Mixte de Recherche 5004, F-34060 Montpellier Cedex 1, France
^b Proteomique, Institut National de la Recherche Agronomique, Unité de Recherche 1199, F-34060 Montpellier Cedex 1, France

² Address correspondence to gibrat{at}supagro.inra.fr.

	ABSTRACT

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

 
Root NO₃^– efflux to the outer medium is a component of NO₃^– net uptake and can even overcome influx upon various stresses. Its role and molecular basis are unknown. Following a functional biochemical approach, NAXT1 (for NITRATE EXCRETION TRANSPORTER1) was identified by mass spectrometry in the plasma membrane (PM) of Arabidopsis thaliana suspension cells, a localization confirmed using a NAXT1–Green Fluorescent Protein fusion protein. NAXT1 belongs to a subclass of seven NAXT members from the large NITRATE TRANSPORTER1/PEPTIDE TRANSPORTER family and is mainly expressed in the cortex of mature roots. The passive NO₃^– transport activity (K_m = 5 mM) in isolated root PM, electrically coupled to the ATP-dependant H⁺-pumping activity, is inhibited by anti-NAXT antibodies. In standard culture conditions, NO₃^– contents were altered in plants expressing NAXT-interfering RNAs but not in naxt1 mutant plants. Upon acid load, unidirectional root NO₃^– efflux markedly increased in wild-type plants, leading to a prolonged NO₃^– excretion regime concomitant with a decrease in root NO₃^– content. In vivo and in vitro mutant phenotypes revealed that this response is mediated by NAXT1, whose expression is upregulated at the posttranscriptional level. Strong medium acidification generated a similar response. In vitro, the passive efflux of NO₃^– (but not of Cl^–) was strongly impaired in naxt1 mutant PM. This identification of NO₃^– efflux transporters at the PM of plant cells opens the way to molecular studies of the physiological role of NO₃^– efflux in stressed or unstressed plants.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

 
Nitrate uptake by plant roots and its subsequent reduction and assimilation are essential for plant growth as well as for N input in many terrestrial trophic chains (Crawford and Glass, 1998; Daniel-Vedele et al., 1998; Williams and Miller, 2001). It results from the balance between an active influx mediated by nH⁺:mNO₃^– symporters (with n > m) and a passive efflux (i.e., an electrically driven uniport) (Crawford and Glass, 1998). Several uptake symporters have been characterized in the NITRATE TRANSPORTER1 (NRT1) and NRT2 gene families (Miller et al., 2007; Tsay et al., 2007), whereas the molecular basis of cellular efflux is still unknown.

In well-supplied and nonstressed plants, NO₃^– efflux can be high but remains lower than influx (Kronzucker et al., 1999), and long-term control of the uptake regime relies on the regulation of active influx transport systems (Lee, 1993). Upon certain biotic (Garcia-Brugger et al., 2006) or abiotic stresses, such as mechanical or transplant shocks (Pearson et al., 1981; Macduff and Jacksson 1992; Dehlon et al., 1995; Aslam et al., 1996) or medium acidification (Aslam et al., 1995), marked increases of NO₃^– efflux leading to (net) NO₃^– excretion were reported. The biological significance of this response remains obscure, as does more generally the physiological role of root NO₃^– efflux.

In vitro, it has long been established that the addition of NO₃^– to plasma membranes (PMs) isolated from a wide range of plant and fungal materials strongly stimulates H⁺-ATPase pumping activity by dissipating the membrane potential (E_m) generated by the pump (Vara and Serrano, 1982; Perlin et al., 1984; De Michelis and Spanswick, 1986). This so-called short-circuiting stimulation by NO₃^– provided evidence for the existence of a passive NO₃^– efflux system in isolated PMs. Its functional features indicated that it could be of biological significance, since, in particular, it displays K_m and V_max values compatible with root NO₃^– cytosolic concentration and efflux capacity, respectively (Pouliquin et al., 2000).

We designed a potentiometric method using the fluorescent oxonol VI dye to measure in vitro both E_m and passive NO₃^– flux on isolated PMs (Pouliquin et al., 1999; Gibrat and Grignon, 2003). On the other hand, we developed methods to solubilize and separate intrinsic membrane proteins in a native (functional) state by gel filtration (GFC) and/or ion-exchange chromatography (IEC). Associated with a final electrophoretic separation in denaturing conditions, GFC/IEC/SDS-PAGE made it possible to identify many membrane transporters by mass spectrometry (MS) (Szponarski et al., 2004, 2007; Delom et al., 2006).

Here, we report the identification of an Arabidopsis thaliana NO₃^– efflux transport protein through a biochemical approach correlating efflux activity and polypeptide abundance in chromatographic fractions of solubilized intrinsic PM proteins from suspension cells. This protein, designated NAXT1 (for NITRATE EXCRETION TRANSPORTER1), is a member of a subset of seven highly similar NAXT proteins belonging to the large NRT1/PEPTIDE TRANSPORTER (NRT1/PTR) family (Tsay et al., 2007). Besides NAXT1, one or several NAXT proteins are also involved in passive NO₃^– transport activity of isolated PMs and in the extent of shoot and root NO₃^– contents in plants grown in standard conditions. In vivo and in vitro mutant phenotypes provide evidence that NAXT1 is the PM efflux transporter responsible for the prolonged root NO₃^– excretion observed after acid load or acidification of the hydroponic medium. Unexpectedly, these treatments induce the accumulation of the NAXT1 protein but not of the transcript.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

 
A Functional Biochemical Approach Leads to the Identification of a Candidate Protein for PM NO₃^– Efflux
A functional biochemical strategy, summarized in Figure 1 , was launched on PMs isolated from tobacco (Nicotiana tabacum) BY2 and Arabidopsis suspension cells to identify polypeptide candidate(s) for the NO₃^– efflux activity. Intrinsic membrane proteins from BY2 cells were solubilized and separated in native conditions by IEC. In each IEC fraction, image analysis of the SDS-PAGE pattern was performed to determine the abundance of the different polypeptide bands (Figure 1A), and in parallel, the NO₃^– efflux activity was measured after reinsertion of the whole protein content into liposomes (Figure 1B). A correlation was then searched for between the abundance of each detected polypeptide band and the activity along successive IEC fractions. Two polypeptide bands of 42 and 17 kD (denoted B₄₂ and B₁₇), both present in the most active fraction, were selected (Figure 1C).

View larger version (43K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 1. Biochemical Strategy That Led to the Identification of NAXT1. The biochemical strategy was developed on PMs isolated first from tobacco BY2 cells and second from Arabidopsis suspension cells. After PM stripping and nondenaturing solubilization (see Methods), intrinsic proteins from tobacco BY2 cells were rapidly fractionated by IEC using various exchanger columns and salt gradients. Polypeptide abundance was estimated in successive IEC fractions by image analysis of SDS-PAGE patterns. NO3– efflux and permeability coefficients were determined in parallel by imposing K+ diffusion Em on proteoliposomes reconstituted from protein fractions. Data from different fractionation experiments are expressed on a relative basis and merged in order to search for a correlation between the facilitated passive transport activity and abundance of the different polypeptide bands detected by SDS-PAGE. Two tobacco polypeptides of 42 and 17 kD (denoted B42 and B17, respectively) were correlated, but attempts to obtain sequence data after excision and trypsin digestion of the bands were unsuccessful. B17 was successfully used to prepare rabbit polyclonal antibodies (ABY2), which made it possible to immunodetect essentially two polypeptides of 17 and 42 kD in PMs from both tobacco and Arabidopsis cells. After fractionation by GFC/IEC/SDS-PAGE, the immunodetected B42 of Arabidopsis was excised and the trypsin-digested peptides were analyzed by MS, leading to the identification of NAXT1. FPLC, fast protein liquid chromatography. (A) Example of SDS-PAGE patterns of successive active fractions obtained by IEC with a Protein Pack Q HR 10/10 column (Waters) eluted between 200 and 300 mM KCl with successive 25 mM steps. Rf, relative migration; a.u., arbitrary unit. (B) Permeability coefficient to NO3– (PNO3) of protein fractions from (A) reinserted into liposomes. The dashed line indicates the PNO3 of control liposomes, corresponding to the nonfacilitated passive diffusion of NO3–. (C) Activity versus abundance correlations for polypeptide bands of 42 kD (B42; r = 0.82) and 17 kD (B17; r = 0.94). The SDS-PAGE pattern of the fraction displaying the highest activity (F275) is shown. Stars indicate the positions of B42 and B17. (D) Immunoblot detection of proteins in PM isolated from tobacco BY2 and Arabidopsis cells by ABY2, the antibodies directed against B17. (E) Identification of the Arabidopsis 42-kD protein ortholog (star) in a GFC/IEC/SDS-PAGE fraction. After B42 excision and trypsin digestion, MS analysis allowed the identification of NAXT1.

Importantly, A_BY2 provided evidence for the existence of an ortholog of 42 kD in the PM fraction from Arabidopsis suspension cells (Figure 1D). Intrinsic PM proteins from Arabidopsis cells were then separated by GFC/IEC/SDS-PAGE (Figure 1E) (Szponarski et al., 2004, 2007), and MS analysis of the immunodetected B₄₂ led to the identification of an as yet uncharacterized 61-kD protein (NP_190151, encoded from locus At3g45650) herein designated NAXT1. In the presence of a broad-specificity plant protease inhibitor cocktail (AP2 instead of AP1; see Methods), A_BY2 clearly immunodetected a 61-kD band in the PMs isolated from roots of Arabidopsis (first lane in Figure 4A below). MS analysis of a NAXT-enriched GFC/IEC/SDS-PAGE fraction of Arabidopsis root PMs confirmed the presence of NAXT1 in the immunodetected 61-kD band (data not shown). This indicated that both the 17- and 42-kD polypeptides previously immunodetected by A_BY2 were probably proteolytic products derived from the full-length NAXT1 protein.

View larger version (23K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 4. ABY2 and ANAXT Antibodies Inhibit PM NO3– Efflux Activity in Vitro. (A) ABY2 and ANAXT antibodies immunodetect the same 61-kD NAXT band in Arabidopsis root PM proteins. The protease inhibitor cocktail AP2 (see Methods) was used throughout PM isolation. (B) Negative and positive controls performed on PM H+-ATPase activity. The maximum H+ pumping activity (VH) was measured in the presence of valinomycin. Antibodies directed against PM H+-ATPase (AATPase) efficiently inhibit the H+ pumping activity, whereas ANAXT does not. The IgG:protein ratio was 5:1 (w/w). Ct, control treatment (no IgG added). (C) Inhibition of NO3– efflux activity. ABY2 and ANAXT antibodies efficiently inhibit NO3– efflux activity (VNO3) assayed at 20 mM NO3– at the indicated IgG:protein ratios (w/w). PI, preimmune. Error bars in (B) and (C) indicate SE (n = 3).

View larger version (71K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 2. NAXT1 Belongs to a Subclass of the NRT1/PTR Family. (A) Unrooted phylogenetic tree of Arabidopsis proteins of the NRT1/PTR family. Members for which functional characterization is described in the literature are indicated in boldface. The NAXT protein subclass is highlighted in the gray oval. (B) Multiple sequence alignment of the seven NAXT proteins. Black-boxed residues are shared by at least four NAXT members. The consensus sequence showing residues conserved in all NAXT members is presented below each line of the alignment. The positions of the NAXT1-derived P1, P2, and P3 peptide sequences used to generate the ANAXT antibodies are indicated. The 11–amino acid long NAXT signature sequence is boxed.

View larger version (17K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 3. Nitrate Efficiently Abolishes the Em Generated by the H+-ATPase of Arabidopsis Root PM. Root PM was isolated from 6-week-old wild-type or mutant plants grown in standard hydroponics. Vesicles (5 µg protein/mL) were double-labeled with fluorescent probes for simultaneous monitoring of pH (ACMA; in [A] and [C]) and Em (oxonol VI; in [B] and [D]). The y axis refers to the following ratios of fluorescence intensity (F) of the dyes. ACMA response is expressed as the ratio of F (recorded during the experiment) to the basal value F0 (at pH = 0) recorded after the addition of the protonophore carbonylcyanide p-trifluoromethoxyphenylhydrazone (FCCP). Oxonol VI response is expressed as the ratio of the fluorescence increase F in response to Em to its basal value F0 (at zero Em), also recorded after FCCP addition. The assay medium contained 10 mM BTP-SO4, pH 6.5, 70 mM K2SO4, 0.5 M sorbitol, 3 mM ATP, 1 µM ACMA, 0.2 µM oxonol VI, and protease inhibitor and anti-phosphatase cocktails. Vesicles contained the same medium without ATP. At the indicated times (arrows), Mg2+ (3.5 mM) was added to activate the H+-ATPase, and then valinomycin (200 nM; [A] and [B]) or NO3– (20 mM; [C] and [D]) was added to collapse Em; finally, FCCP (200 nM) was added to collapse pH. Closed and open symbols in (A) to (D) correspond to two measurements in each condition. In (E), the initial rate of ACMA quenching (V) was measured just after activation of the H+-ATPase of wild-type (black symbols), naxt1-1 (white symbols), or naxt1-2 (gray symbols) root PM. V was determined in the presence of valinomycin (200 nM) but in the absence of NO3– (triangles), or in the absence of valinomycin but in the presence of NO3– (circles), at the indicated concentrations added just prior to H+-ATPase activation. Error bars indicate SE (n = 6). The curve results from a nonlinear fit of average values for the three genotypes, according to the equation V = V0 + (VNO3max x NO3–)/(Km + NO3–), giving Km = 4.8 ± 0.1 mM, V0 = 2.2 ± 0.1, and Vmax = 6.4 ± 0.6%·min–1·µg–1 protein.

The ability of A_BY2 and A_NAXT antibodies to specifically inhibit the NO₃^– efflux activity in vitro (V_NO3) was tested using a previously established functional assay (Gaymard et al., 1993). As a positive control, polyclonal anti-PM ATPase antibodies (Santoni et al., 1993), designated here A_ATPase, inhibited by >80% the maximum rate of H⁺ excretion by the H⁺-ATPase short-circuited with the K⁺ ionophore valinomycin (Figure 4B). Conversely, as expected, this activity was not affected significantly by A_NAXT at an identical 5:1 IgG:protein ratio. In the presence of A_NAXT and A_BY2 at a 1:1 IgG:protein ratio, V_NO3 decreased by 60 and 45%, respectively (Figure 4C), whereas no inhibition was observed with preimmune antibodies (A_PI). In the presence of A_NAXT and A_BY2 at a 5:1 IgG:protein ratio, inhibition levels reached 85 and 70%, respectively, relative to V_NO3 in the presence of A_PI. This indicates that most of the NO₃^– efflux activity in root PM vesicles is mediated by NAXT proteins.

RNA interference (RNAi) transgenic plants (reviewed in Brodersen and Voinnet, 2006; Mansoor et al., 2006) impaired in the expression of all NAXT members (naxt-RNAi) were obtained (Figures 5A and 5B ). Although no growth phenotype was observed, NO₃^– contents of naxt-RNAi plants grown in vitro on standard Murashige and Skoog medium were altered significantly compared with those of wild-type plants (30% higher in roots and 26% lower in shoots; Figure 5C). This indicates that one or several NAXT members are involved in the distribution of NO₃^– in plants.

View larger version (26K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 5. Root and Shoot NO3– Contents of Plants Grown in Vitro in Standard Conditions: Effect of the Suppression of the Expression of NAXT Genes. (A) Schematic representation of the RNAi cassette integrated in the genome of naxt-RNAi transgenic plants. The 296-bp-long NAXT1 ORF cDNA fragment (nucleotides 457 to 753) displays 80 to 93% identity to corresponding sequences of other NAXT members. This cDNA fragment was cloned in the pHannibal vector in the two opposite orientations (arrows) downstream and upstream of the PDK intron (see Methods). nt, nucleotide; PDK, pyruvate orthophosphate dikinase; Pro 35S, 35S promoter from Cauliflower mosaic virus; Ter OCS, octopine synthase terminator. (B) RT-PCR analysis of total root RNAs shows at least partial suppression of gene expression of all NAXT members in naxt-RNAi plants. The expression level of Elongation Factor (EF1) was used as an internal control. PCR was performed using 28 cycles, except for At3g45720 (35 cycles) and EF1 (20 cycles). At3g45690 was undetectable in either plant even using 35 PCR cycles (data not shown). Primer sequences are provided in Supplemental Table 1 online. (C) In standard conditions, naxt1-1 mutant plants display wild-type-like root and shoot NO3– contents, but naxt-RNAi plants show significantly altered NO3– contents in both roots and shoots. The T-DNA insertion line naxt1-1 is described in Supplemental Figure 2 online. DW, dry weight.

View larger version (94K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 6. The NAXT1-GFP Fusion Protein Is Expressed in the PM of Arabidopsis Cells. GFP constructs were expressed transiently in Arabidopsis protoplasts, and signals were examined with a laser confocal microscope under transmitted light ([A] and [C]) or under fluorescence after excitation at 530 nm ([B] and [D]). Bars = 10 µm. (A) and (B) Protoplast transformed with the control construct (P35S:GFP). (C) and (D) Protoplast transformed with pG611, a construct expressing NAXT1 fused to the N-terminal part of GFP (P35S:NAXT1-GFP). T, tonoplast.

View larger version (77K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 7. NAXT1 Is Expressed Mainly in the Cortex of Mature Roots. (A) The NAXT1 gene is more strongly expressed in roots than in leaves. RT-PCR was performed on 6-week-old hydroponic plants. The two primers used were F1-738 and R1-1457 (sequences provided in Supplemental Table 1 online). The expression of EF1 was used as a control. (B) to (G) Histochemical analyses of transgenic Arabidopsis plants expressing the PNAXT1:GUS construct. Plants were cultured in vitro ([B], [C], [F], and [G]) or in hydroponics ([D] and [E]). (B) and (C) Micrographs of 7-d-old (B) and 14-d-old (C) seedlings showing major GUS staining in roots. Bars = 1.2 mm. (D) Micrograph of part of the root system from a 6-week-old plant. No staining is detected in the oldest zone of the primary root. Bar = 1.2 mm. (E) No GUS staining is observed at the root tip. (F) Micrograph of the mature root of a 7-d-old seedling showing strong GUS staining in the cortex. Bar = 50 µm. (G) Cross section of the mature root of a 7-d-old seedling showing stained cortical cells. The epidermis is also stained, but less strongly. Bar = 50 µm.

These results show that NAXT1 is essentially expressed in the mature root cortex and are consistent with the hypothesis that NAXT1 is involved in root NO₃^– efflux to the external medium.

naxt1 T-DNA Insertion Mutants Grown in Standard Conditions Do Not Display an Apparent Phenotype
Light and plant N status, both known to regulate the gene expression of major NO₃^– uptake transporters such as NRT1:1 (Forde, 2000), did not affect NAXT1 gene expression in hydroponically grown plants (see Supplemental Figure 1 online). Phenotypes resulting from the alteration of NAXT1 gene expression were searched for in standard culture conditions using naxt1-1 and naxt1-2, two T-DNA insertion lines (insertions in exon 3 of NAXT1) in which the mRNA of NAXT1 was undetectable (see Supplemental Figures 2A and 2B online). Wild-type and naxt1 plants displayed no significant difference in root and shoot growth (data not shown), in root and shoot NO₃^– contents (Figure 5C; controls in Figures 8B and 10A below), in unidirectional root NO₃^– efflux and influx (Figure 8), and in passive NO₃^– transport activity in vitro (controls in Figures 9 and 10 below). Consequently, although some NAXT members are involved in passive NO₃^– transport and in root/shoot NO₃^– distribution, the contribution of NAXT1 seems negligible in plants grown in standard conditions.

View larger version (20K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 8. NAXT1 Is Required for the Acid Load–Dependent Root NO3– Excretion. Acid load was applied on 6-week-old plants grown on standard hydroponics by adding 10 mM propionate to the medium 30 min prior to the determination of root NO3– content or prior to plant transfer to measure unidirectional NO3– influx (OC) and efflux (CO). DW, dry weight. (A) The increase of CO in the wild type is abolished in naxt1-1 plants. The inset shows OC in the same plants. Error bars indicate SE (n = 8). (B) The decrease in root NO3– content observed in wild-type plants after acid load (gray bars) is abolished in naxt1 mutants and restored in the complemented naxt1-1 mutant. Control plants (no acid load) were transplanted (white bars) or not (black bars) prior to the determination of root NO3– content. Error bars indicate SE (n = 6).

View larger version (22K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 10. NAXT1-Mediated Responses in Plants Subjected to Medium Acidification. Plants were grown in standard hydroponics (pH 6) for 6 weeks. Prior to analysis, they were incubated for 30 min with nutrient medium brought to pH 3. (A) The decrease in root NO3– content observed in wild-type plants after medium acidification is abolished in naxt1 mutants. Black bars, control plants; white bars, plants subjected to medium acidification. Error bars indicate SE (n = 6). DW, dry weight. (B) NAXT1 protein level is upregulated, whereas NAXT1 transcript level remains stable, upon medium acidification. RT-PCR and protein gel blot analyses were performed as described for Figure 9. (C) In vitro, the passive NO3– flux (JNO3) in wild-type root PM displays an acidic optimum pH of 6.5. JNO3 was measured at 20 mM NO3– using the oxonol VI probe as detailed in Supplemental Figure 3 online. (D) The naxt1-1 mutant is strongly impaired in NO3– but not in Cl– passive flux. JNO3 or JCl in vesicles from wild-type and naxt1-1 root PMs were compared at pH 6.5 and 20 mM NO3– or 50 mM Cl–.

View larger version (23K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 9. NAXT1 Protein Accumulation Is Responsible for the Acid Load–Dependent Increase in NO3– Efflux Activity Measured in Vitro. Acid load was applied on 6-week-old plants grown on standard hydroponics by adding 10 mM propionate to the medium for the indicated durations prior to root PM isolation. Except where indicated, protein gel blot analysis was performed using ANAXT antibodies. (A) When the duration of the acid load treatment increases, NO3– efflux activity of isolated root PM vesicles (VNO3) is enhanced in wild-type plants but not in naxt1-1 or naxt1-2 plants. Acid load was applied or not (black bars) during 1 h (gray bars) or 3 h (white bars). VNO3 is expressed in percentage relative to that of PM vesicles from control wild-type plants. Error bars indicate SE (n = 5). (B) NAXT1 gene expression is not affected by acid load. NAXT1 RT-PCR was performed on total RNAs isolated from wild-type roots with the primers F1-738 and R1-1457 (see Supplemental Table 1 online). The expression of EF1 was used as a control. (C) The intensity of the NAXT band immunodetected in root PM is enhanced by acid load in wild-type plants but not in naxt1-1 or naxt1-2 plants. Plant roots were acid-loaded (+) or not (–) for 1 h. (D) Upon acid load, time-dependent NAXT accumulation demonstrated by protein gel blot on wild-type root PM is also observed on whole root membranes (P100K). (E) NAXT1 protein accumulates in acid-loaded roots of transgenic plants expressing NAXT1 constitutively. Protein gel blot analyses were performed on PMs from a complemented naxt1-1 plant constitutively expressing the NAXT1 coding sequence and from a transgenic plant constitutively expressing AcV5-tagged NAXT1. The NAXT1-AcV5 fusion protein was immunodetected using an anti-AcV5 monoclonal antibody.

Unexpectedly, _co largely exceeded _oc in wild-type and naxt1 unloaded controls (Figure 8A, inset), showing that even control plants underwent a NO₃^– excretion regime. However, root NO₃^– contents measured prior to, and 30 min after, plant transfer in the ¹⁵N assay medium were similar, indicating that the excretion regime was transient in the absence of acid load (Figure 8B). This transient excretion is likely attributable to the transplant shock subsequent to the transfer of plants to the ¹⁵N assay medium performed just prior to _co and _oc measurements (Delhon et al., 1995). Indeed, mechanical stress almost instantaneously triggers a NO₃^– excretion due to a strong increase in the efflux, which recovers its basal value within a few minutes, the influx being unaffected (Aslam et al., 1996). Here, NAXT1 does not seem to be involved in the transient excretion observed after plant transfer, because _co and NO₃^– contents of wild-type and naxt1-1 roots were very similar in unloaded controls.

By contrast, 30 min after the addition of propionate to the nutrient medium (containing 1 mM NO₃^–), the root NO₃^– content decreased by 30% in the wild type, whereas it was not affected in either the naxt1-1 or the naxt1-2 mutant (Figure 8B). Furthermore, it decreased by 45% in a naxt1-1 mutant complemented with the NAXT1 coding sequence expressed constitutively under the control of the cauliflower mosaic virus (CaMV) 35S promoter harboring a duplicated enhancer sequence (P70) (Kay et al., 1987). Thus, acid load is responsible for prolonged root NO₃^– excretion mediated by NAXT1, leading to a marked decrease in root NO₃^– content. This is the reason for the designation of this protein as NAXT (for NITRATE EXCRETION TRANSPORTER).

Prolonged acid load could not be easily used in vivo to investigate its effect on _co, because of the large and rapid decrease in root NO₃^– content (i.e., in NO₃^– potentially available to sustain _co). Therefore, NO₃^– efflux activity in vitro (V_NO3) was measured on root PM isolated from plants acid-loaded (or not) for 1 or 3 h. V_NO3 in wild-type, naxt1-1, and naxt1-2 controls (non-acid-loaded plants) was indistinguishable (Figure 9A), consistent with the absence of an efflux phenotype observed previously in planta.

By contrast, V_NO3 was higher by 50 and 80% in wild-type PM after 1 and 3 h of acid load, respectively. This increase was impaired in both naxt1-1 and naxt1-2 mutants, showing that the increase in V_NO3 observed in wild-type PM upon acid load resulted from the induction of a NAXT1-mediated activity.

NAXT1 Expression Is Regulated at the Posttranscriptional Level
The acid load–mediated functional response described above was not related to an upregulation of NAXT1 gene expression, because the NAXT1 transcript level remained stable upon acid load of wild-type roots (Figure 9B). Since no NAXT1-specific antibodies were available to directly investigate NAXT1 protein level in the wild type, protein gel blot signals of root PM from the wild type and naxt1 mutants were compared using A_NAXT. No significant difference was observed between these plants when grown in standard conditions (Figure 9C), in agreement with the absence of a mutant phenotype reported above. By contrast, the intensity of the NAXT band increased significantly after 1 h of acid load in the wild type but not in the two naxt1 mutants, suggesting that NAXT1 accumulates in the wild-type PM upon acid load.

The intensity of the NAXT band also increased in the whole membrane fraction (P100K) upon acid load (Figure 9D), showing that the NAXT1 accumulation in the PM is unlikely to result from its relocalization from another storage membrane compartment. The acid load–dependent accumulation of NAXT1 in the wild-type PM was then monitored in transformed plants constitutively expressing the NAXT1 coding sequence under the control of the P70 promoter (Figure 9E). In complemented naxt1-1 plants, the NAXT band (detected by A_NAXT) increased after acid load, suggesting that the accumulation of NAXT1 transcript and protein is uncoupled. In a transgenic line constitutively expressing NAXT1 tagged with the AcV5 epitope (Lawrence et al., 2003), the NAXT1-AcV5 fusion protein (detected by a monoclonal anti-AcV5 antibody) also accumulated after acid load. This demonstrates that the expression of NAXT1 is regulated at the posttranscriptional level.

NAXT1 Is the Efflux Transporter Responsible for the Root NO₃^– Excretion Induced by Plant Medium Acidification
Medium acidification can be considered a biologically more relevant treatment than acid load artificially generated with propionate. Acidification of the barley (Hordeum vulgare) root hydroponic medium was shown to have little effect on NO₃^– influx but markedly stimulated efflux (Aslam et al., 1995). In particular, a net excretion regime was observed in 2 mM MES buffer at pH 3. We hypothesized that medium acidification could be responsible for an acidic drift of the cytosolic pH and induce a response similar to that observed upon acid load. Indeed, membrane bilayers display a surprisingly high permeability coefficient to H⁺ (10^–6 M^–1 s^–1; Rossignol et al., 1982). Moreover, accounting for the large (negative) E_m of root cells, a strong passive entry of H⁺ within root cells is expected at pH 3.

Consistently, wild-type root NO₃^– content decreased by almost 50% after acidification of the plant nutrient medium for 30 min at pH 3, whereas it was not affected significantly in either naxt1-1 or naxt1-2 mutant roots (Figure 10A ). NAXT1 expression remained stable upon medium acidification, whereas protein gel blot analyses showed a marked increase in the intensity of the NAXT band in wild-type but not in naxt1-1 root PM (Figure 10B). These data indicate that, like acid load, medium acidification induces prolonged root NO₃^– excretion mediated by NAXT1, which is also upregulated at the posttranscriptional level.

Some intrinsic properties of the NO₃^– efflux were also determined on root PM of plants treated at pH 3, using a transport assay independent of the H⁺-ATPase (Pouliquin et al., 1999). As detailed in Supplemental Figure 3 online, the rate of passive NO₃^– flux (J_NO3) was determined from the rate of NO₃^–-dependent dissipation of imposed diffusion E_m monitored quantitatively with oxonol VI. J_NO3 in wild-type PM, assayed between pH 5.5 and 7.0, revealed an optimal pH at 6.5 (Figure 10C). J_NO3 in naxt1-1 was negligible compared with J_NO3 in the wild type at pH 6.5 (Figure 10D), showing that, essentially, NAXT1 operates at this optimal acidic pH. By contrast, naxt1-1 was not affected significantly in passive Cl^– flux, showing that NAXT1 is not competent for Cl^–.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

 
A Biochemical Strategy for the Identification of Secondary Transport Systems
The plant PM H⁺-ATPase, an archetype of the plant primary transporter, is one of the most electrogenic pumps, able to generate E_m down to –250 mV (Sanders and Slayman, 1989). Here, the H⁺ pump from Arabidopsis PM is short-circuited in vitro after the addition of a low concentration of NO₃^– (20 mM), as already observed in maize (Zea mays) root PM (Grouzis et al., 1997). By contrast, short-circuiting of the H⁺ pump by K⁺ requires the addition of a K⁺ ionophore and 10-fold higher K⁺ concentrations. In vitro, the root cell PM appears more conductive to NO₃^– than to K⁺. This conclusion is consistent with the large permeability coefficient to NO₃^– observed in vesicles from maize root PM (P_NO3 10^–9 M^–1 s^–1; Pouliquin et al., 1999). These observations made in vitro have implications in terms of the relative abundance of transport systems at the surface of plant cells, because isolated PM fractions are constituted of microscopic vesicles. After cell homogenization, 1 µm² of PM surface generates 100 vesicles (Pouliquin et al., 2000). PM H⁺-ATPase accounts for 5% of root PM proteins (Serrano, 1985; Sussman, 1994), which would approximately correspond to 500 H⁺ pump molecules per 1 µm². Therefore, it is expected that most PM vesicles are competent for H⁺ pumping. Similarly, the NO₃^– short-circuiting of the H⁺ pump observed in PM vesicles indicates that most vesicles are also competent for passive NO₃^– transport and thus that the transport is mediated by a relatively abundant PM transport protein. From these considerations, it was anticipated that such a transport system should be amenable to biochemical tracking based on an abundance versus activity correlation approach in native chromatographic fractions of intrinsic PM proteins, followed by MS identification of candidate proteins. This strategy led here to the identification of NAXT1, a transporter belonging to the large NRT1/PTR family.

Two types of antibodies were obtained in this study. A_BY2 antibodies were directed against a candidate polypeptide band correlated with the efflux activity in chromatographic fractions of PM proteins from tobacco. A_NAXT antibodies were raised against synthetic peptides to specifically detect NAXT homologs in Arabidopsis. Despite differences in their conception and origin, both antibodies immunodetected the same NAXT-containing 61-kD polypeptide band in Arabidopsis root PM and both specifically inhibited NO₃^– efflux activity measured in vitro. These results validate the candidate protein identified following the biochemical approach adopted here.

This study demonstrates the feasibility of a functional biochemical strategy to identify ion transporters. However, such an approach is probably not suitable for the identification of transport systems displaying poorly detectable activity at the isolated PM level and/or very low abundance. Ion channels, for example, are poorly abundant in membranes from plant tissues (typically, from 0.1 to 1 channel molecule per 1 µm²). Despite their high molecular transport activity, they would functionally remain almost undetectable at the vesicle level, since only one vesicle per 100 to 1000 should be channel-competent. Consistently, a K⁺ ionophore is required to short-circuit H⁺-ATPase in isolated root PM vesicles, although different K⁺ channels coexist in this membrane.

A Molecular Basis for Passive NO₃^– Efflux Activity at the PM of Plant Cells
Nitrate transport activity coupled to the PM H⁺-ATPase was long since demonstrated in vitro on root PM vesicles (Vara and Serrano, 1982; Perlin et al., 1984; De Michelis and Spanswick, 1986). Using antibodies directed against NAXT proteins (A_NAXT) in a functional inhibition assay, we show here that NAXT transporters are responsible for most of this activity in Arabidopsis root PM.

In vitro, two different spectroscopic methods applied to membrane vesicles made it possible to show that NAXT-mediated NO₃^– transport is driven by E_m and associated with the transfer of a net negative charge in the same direction. NAXT nH⁺:mNO₃^– cotransport mechanisms with m > n seem unlikely in the light of the comparison of the maximum acidification rates of the H⁺-ATPase short-circuited by valinomycin or by NO₃^– (Figure 3). Valinomycin is an ionophore mediating a passive K⁺ uniport, expected to enable the exchange of one K⁺ per pumped H⁺. Therefore, valinomycin-mediated K⁺ short-circuiting allows for the measurement of the true maximum initial acidification rate of the H⁺-ATPase. The initial acidification rate in the presence of 60 mM NO₃^– accounted for 85% of the value measured in the presence of valinomycin. If NO₃^– was transported according to an antiport or a symport mechanism, significantly lower or higher maximum initial acidification rates, respectively, should have been observed upon NO₃^– short-circuiting. Thus, the NAXT system likely mediates an E_m-driven uniport of one NO₃^– per H⁺ pumped by the H⁺-ATPase. In situ, the NAXT system is expected to mediate a passive NO₃^– efflux at the root PM, because it is functionally coupled to the root PM H⁺-ATPase and NO₃^– is transported in the same direction as the pumped H⁺. This conclusion is consistent with physiological studies indicating that root NO₃^– efflux, in contrast with NO₃^– influx, is a thermodynamically downhill process (reviewed in Crawford and Glass, 1998; Miller et al., 2007). Moreover, although the extent of the cytosolic NO₃^– pool is debated (Britto and Kronzucker, 2003), the K_m value of the NAXT system observed in vitro is consistent with the range of cytosolic NO₃^– values recently observed by different methods in well-supplied plants (Radcliffe et al., 2005; Ritchie, 2006).

Except for At3g45690, whose expression was undetectable in shoots or in roots, all NAXT genes are expressed in Arabidopsis roots (Tsay et al., 2007; Arabidopsis eFP Browser, http://bbc.botany.utoronto.ca/efp/cgi-bin/efpWeb.cgi; this work) and several NAXT proteins have been identified in wild-type root PM by MS (W. Szponarski, personal communication). The strong similarity between the NAXT proteins should be emphasized and suggests a recent duplication event from a common ancestor gene. Although it cannot be specified at present which members of the NAXT family actually contribute to the NO₃^– efflux activity in root PM from plants grown in standard hydroponics, it is likely that all members share a common transport function in PM.

In planta, the ¹⁵N method used here allowed the measurement of both root unidirectional efflux and influx (_co and _oc) on the same plants. This made it possible to show that NAXT1 is responsible for an acid load–induced increase in _co, which reaches a value 15-fold higher than _oc. In vitro, using two different spectroscopic methods, the increase in passive NO₃^– efflux measured in the PM from plant roots subjected to acid load or medium acidification was found to be mediated by NAXT1.

This study provides a molecular basis for passive anion efflux at the PM of plant cells and introduces a new transport function in the large NRT1/PTR family (50 members). To date, few Arabidopsis NRT1/PTR members have been characterized at the functional level (Tsay et al., 2007). Among them, At NRT1 proteins are involved in NO₃^– uptake (three members characterized to date: Huang et al., 1999; Liu et al., 1999; Chiu et al., 2004), whereas At PTR proteins are involved in oligopeptide uptake (three members characterized: Song et al., 1996; Dietrich et al., 2004; Karim et al., 2005). In addition, eight other Arabidopsis NRT1/PTR members were proven to encode NO₃^– transporters after their heterologous expression in Xenopus oocytes, according to unpublished data mentioned by Tsay et al. (2007).

The finding that passive NO₃^– efflux can be mediated by a subclass of NRT1/PTR transporters was rather unexpected with regard to the variety of channels mediating the passive transport of mineral ions in plants. In particular, various anion channels have been functionally characterized in roots, with different selectivities, gatings, and pharmacological signatures (reviewed in Roberts, 2006). In isolated PM vesicles, passive NO₃^– efflux was shown to be voltage-gated, and data were formally equivalent to a current mediated by a rectifying anion channel displaying an optimum voltage at –105 mV (Pouliquin et al., 2000). Nevertheless, transporters can also be voltage-gated, as exemplified by the 1 NO₃^–:2 H⁺ symport activity observed upon expression of the high-affinity active fungal transporter NrtA in Xenopus oocytes (Boyd et al., 2003). Furthermore, the prokaryotic CLCec anion channel, displaying an almost occluded aqueous pore, was shown to mediate a stoichiometrically fixed anion/proton antiport activity (reviewed in Miller, 2006). Such a transporter activity involving channel-like features was also recently reported for the plant vacuolar At CLCa (De Angeli et al., 2006). These data indicate that the distinction between channels and transporters is not clear-cut, although ion channels generally display high transport activities and low abundances relative to transporters. It is worth noting that anion channels could also be involved in the root NO₃^– efflux observed in planta, but their activity would likely remain undetectable in membrane vesicle preparations, as discussed above.

Finally, it should be emphasized that since all NAXT members are expressed in root tissues, confirmation of the molecular transport mechanisms and properties of the different members will require their functional analysis in a suitable heterologous system.

NAXT1 Protein Accumulation in PM and naxt1 Mutant Phenotypes
In standard culture conditions, naxt1 mutants displayed no apparent phenotype, although NAXT1 gene expression was detected and NAXT1 protein was identified by MS in wild-type root PM. In particular, PM isolated from wild-type and naxt1 roots displayed similar NO₃^– efflux activities. Since the intensity of the immunodetected NAXT band was not significantly affected by the disruption of the NAXT1 gene, it is likely that the NAXT1 protein accounts for only a minor part of the pool of NAXT proteins in the root PM of wild-type plants grown in standard conditions.

Acid load or medium acidification did not affect the NAXT1 transcript level. However, both induced an increase in the intensity of the immunodetected NAXT band in wild-type but not in naxt1 mutant root PM. This suggests that, essentially, the NAXT1 protein accumulates in wild-type PM upon these treatments and that this accumulation strongly increases the pool of NAXT proteins. The acid load–induced accumulation of a constitutively expressed AcV5-tagged NAXT1 protein confirmed that NAXT1 accumulation in the PM is controlled at the posttranscriptional level.

To our knowledge, two other examples of the uncoupling of mRNA and protein accumulation have been described for ion transporters of Arabidopsis PM, using similar 35S transgenic plants. When the high-affinity iron uptake transporter IRT1 (Connolly et al., 2002) or the boron exporter BOR1 (Takano et al., 2005) are constitutively expressed in transgenic lines under the control of the CaMV 35S promoter, both transporters accumulate in limiting substrate conditions but are degraded upon substrate resupply. In the presence of sufficient substrate, the turnover of IRT1 is proposed to be downregulated by ubiquitin-dependent proteolysis, whereas BOR1 is transferred to the vacuole via the endosomes for degradation.

Here, a posttranscriptional control was found to limit the PM accumulation level of NAXT1 in standard hydroponics. It is noteworthy that NO₃^– efflux increases the energetic cost of the net NO₃^– uptake, since a minimum expenditure of 2 mol of ATP is required per mole of absorbed NO₃^– (Briskin et al., 1995). Nitrate efflux has been reported to be responsible for higher respiratory costs of NO₃^– uptake in slow-growing grasses (Scheurwater et al., 1999). Furthermore, induction of NAXT1 accumulation was found here to promote a net excretion regime that is expected to have strong detrimental effects for N nutrition. In this context, the posttranscriptional control limiting the NAXT1 accumulation level in standard culture conditions seems essential.

Posttranscriptional mechanisms allowing variable expression of a protein while the mRNA remains constantly expressed are expected to ensure rapid functional responses. Consistently, naxt1 mutant plants display a clear root NO₃^– excretion phenotype after 30 min of acid load or medium acidification. This could reveal the importance of NO₃^– excretion, an energetically costly process consuming a nutritional resource, as an early response to marked acidic drift of the cytosolic pH. Because of the tight coupling of H⁺ and NO₃^– excretion observed in vitro (Vara and Serrano, 1982; De Michelis and Spanswick, 1986; Grouzis et al., 1997) and the known ability of acid load to stimulate H⁺ excretion by PM ATPase in vivo (Hager and Moser, 1985; Mathieu et al., 1986; Beffagna and Romani 1991), we hypothesized that cytosolic acidification of root cells leads to NO₃^– excretion (Pouliquin et al., 2000). The results obtained here support this hypothesis. In particular, NAXT1 displays optimal activity at pH 6.5. However, the biological significance of such NO₃^– excretion remains unclear. In principle, H⁺ excretion by H⁺-ATPase is expected to attenuate acidic drift in the cytosol, provided excreted H⁺ can be electrically balanced by an equivalent transport of a strong ion (Felle, 2005). Consequently, it is tempting to propose that coexcretion of H⁺ and NO₃^– in vivo would allow the attenuation of stress-generated acidic drifts.

Possible Biological Relevance of NAXT-Mediated Root NO₃^– Efflux
Root NO₃^– efflux to the external medium is a component of net NO₃^– uptake. It may account for 80% of root NO₃^– influx in well-supplied plants (Kronzucker et al., 1999). In standard culture conditions, naxt1 mutants displayed no efflux phenotype and their growth or NO₃^– content was similar to those of wild-type plants. Thus, NAXT1 is not likely to play a role in plant N nutrition in standard culture conditions. Nevertheless, in this condition, NO₃^– efflux in vitro was inhibited using an antibody directed against the whole NAXT protein family. Furthermore, partial suppression of gene expression of all NAXT members in naxt-RNAi plants led to a significant modification of both root and shoot NO₃^– contents. Together, these data support the hypothesis that the activity of one or several NAXT members (other than NAXT1) is involved in NO₃^– transport in plants in standard culture conditions. Two NAXT members (Atg45680 and Atg45710) expressed in the root epidermis or cortex (Arabidopsis eFP Browser, http://bbc.botany.utoronto.ca/efp/cgi-bin/efpWeb.cgi) could also be involved in NO₃^– efflux to the external medium.

Nitrate efflux at the PM is also involved in xylem loading. Several anion channels have been described at the electrophysiological level in xylem parenchyma cells (reviewed in Roberts, 2006). A NAXT member (Atg45700) is strongly expressed in the root stele (Arabidopsis eFP Browser, http://bbc.botany.utoronto.ca/efp/cgi-bin/efpWeb.cgi), raising the possibility that transporters from the NAXT family are also involved in this process. Interestingly, relative to wild-type plants, the NO₃^– content of a naxt-RNAi line was significantly higher in roots and lower in shoots, suggesting that root-to-shoot NO₃^– translocation is altered in this mutant.

Large increases in NO₃^– efflux associated with NO₃^– excretion were reported in stresses such as defoliation (Macduff and Jacksson, 1992), mechanical stress (Aslam et al., 1996), transplant shock (Dehlon et al., 1995), and pathogen attacks (Garcia-Brugger et al., 2006). Here, we show that NAXT1 is responsible for the prolonged root NO₃^– excretion regime induced by acid load or medium acidification, leading to a decrease in root NO₃^– content. With up to 80 mM NO₃^– in well-fed plants (Radcliffe et al., 2005), the vacuolar pool accounts for the major part of the cell NO₃^– content (Devienne et al., 1994) and is likely remobilized to sustain NO₃^– cellular excretion. Consistently, NAXT1 is not expressed in root tip meristematic cells (Figure 7E) known to display very small vacuoles (Patel et al., 1990) but is expressed preferentially in mature cortical cells displaying large differentiated vacuoles.

Strong decreases in cell NO₃^– content due to prolonged NO₃^– excretion (Wendehenne et al., 2002) and early cytosol acidification (Mathieu et al., 1996; Roos et al., 1998) have been reported upon pathogenic attacks or application of elicitors of defense reactions. Additionally, acidification of plant medium has been shown to stimulate root NO₃^– efflux to the outer medium (Aslam et al., 1995). It should be noted that considerable variations in soil pH may occur within a region, depending on the type of soil and seasonal changes in soil moisture, temperature, microbial activity, and plant growth. Highly to extremely acidic soil solutions are not rare situations in natural environments (Soil Survey Staff, 1999). These data provide possible biological significance for the NAXT1-mediated NO₃^– excretion observed here upon root cell acidification.

In conclusion, a transporter endowed with NO₃^– excretion activity has been identified. NAXT1 mutants and transformants will likely help to clarify the implication and role of NAXT1 in stresses generating cytosolic pH drifts and/or NO₃^– excretion. More generally, characterization of members of the NAXT family is expected to lead to a better understanding of the enigmatic role of NO₃^– efflux in plants.

	METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

 
Plant Cell Culture, Plant Growth Conditions, and Mutant Line Genotyping
Tobacco BY2 (Nicotiana tabacum cv Bright Yellow 2) cell suspensions (Nagata et al., 1992) were cultured in the dark at 26°C in Murashige and Skoog medium, pH 5.6, containing Murashige and Skoog salt (Murashige and Skoog, 1962), 1 mg/L thiamine-HCl, 0.2 mg/L 2,4-dichlorophenylacetic acid, 100 mg/L myo-inositol, 30 g/L sucrose, 200 mg/L KH₂PO₄, and 2 g/L MES. Cells were maintained by weekly dilution (1:40) into fresh medium. Arabidopsis thaliana ecotype Columbia cell suspensions were cultured under continuous light at 24°C in Jouanneau and Péaud-Lenoel medium as described (Axelos et al., 1992). They were subcultured with a 1:9 dilution factor every 7 d and used for membrane isolation at 7 d after transfer.

For GUS analyses of 7- and 14-d-old plants and to phenotype 14-d-old naxt-RNAi plants, Arabidopsis plants were grown in vitro in twofold-diluted Murashige and Skoog culture medium. For other analyses, Arabidopsis plants were grown for 6 weeks in hydroponic conditions as described (Lejay et al., 1999). Unless stated otherwise, the source of NO₃^– was 1 mM NH₄NO₃. Acid load treatment was performed by adding K-propionate (10 mM final concentration, pH 6.0) in the plant nutrient medium.